2001-02-14 17:24:45 +01:00
/ * params . def - Run - time parameters .
2017-01-01 13:07:43 +01:00
Copyright ( C ) 2001 - 2017 Free Software Foundation , Inc .
2001-02-14 17:24:45 +01:00
Written by Mark Mitchell < mark @ codesourcery . com > .
Makefile.in, [...]: replace "GNU CC" with "GCC".
* Makefile.in, alias.c, basic-block.h, bb-reorder.c, bitmap.c,
bitmap.h, builtin-types.def, builtins.c, builtins.def,
c-aux-info.c, c-common.c, c-common.def, c-common.h,
c-convert.c, c-decl.c, c-dump.c, c-dump.h, c-errors.c,
c-format.c, c-lang.c, c-lex.c, c-lex.h, c-parse.in,
c-pragma.c, c-pragma.h, c-semantics.c, c-tree.h, c-typeck.c,
caller-save.c, calls.c, collect2.c, collect2.h, combine.c,
conditions.h, config.gcc, configure.frag, configure.in,
conflict.c, convert.c, convert.h, cppspec.c, crtstuff.c,
cse.c, cselib.c, cselib.h, dbxout.c, dbxout.h, defaults.h,
dependence.c, df.c, df.h, diagnostic.c, diagnostic.h,
doloop.c, dominance.c, dwarf.h, dwarf2.h, dwarf2asm.c,
dwarf2asm.h, dwarf2out.c, dwarf2out.h, dwarfout.c,
emit-rtl.c, errors.c, errors.h, except.c, except.h,
exgettext, explow.c, expmed.c, expr.c, expr.h, final.c,
fixproto, flags.h, flow.c, fold-const.c, fp-test.c,
function.c, function.h, gbl-ctors.h, gcc.c, gcc.h, gcc.hlp,
gccspec.c, gcov-io.h, gcse.c, genattr.c, genattrtab.c,
gencheck.c, gencodes.c, genconfig.c, genemit.c,
genextract.c, genflags.c, gengenrtl.c, genmultilib,
genopinit.c, genoutput.c, genpeep.c, genrecog.c,
gensupport.c, gensupport.h, ggc-callbacks.c, ggc-common.c,
ggc-none.c, ggc-page.c, ggc-simple.c, ggc.h, global.c,
graph.c, graph.h, gthr-aix.h, gthr-dce.h, gthr-posix.h,
gthr-rtems.h, gthr-single.h, gthr-solaris.h, gthr-vxworks.h,
gthr-win32.h, gthr.h, haifa-sched.c, halfpic.c, halfpic.h,
hard-reg-set.h, hwint.h, ifcvt.c, input.h, insn-addr.h,
integrate.c, integrate.h, jump.c, lcm.c, libgcc2.c,
libgcc2.h, lists.c, local-alloc.c, loop.c, loop.h,
machmode.def, machmode.h, main.c, mbchar.c, mbchar.h,
mips-tdump.c, mips-tfile.c, mklibgcc.in, mkmap-flat.awk,
mkmap-symver.awk, optabs.c, output.h, params.c, params.def,
params.h, predict.c, predict.def, predict.h, prefix.c,
prefix.h, print-rtl.c, print-tree.c, profile.c, protoize.c,
read-rtl.c, real.c, real.h, recog.c, recog.h, reg-stack.c,
regclass.c, regmove.c, regrename.c, regs.h, reload.c,
reload.h, reload1.c, reorg.c, resource.c, resource.h, rtl.c,
rtl.def, rtl.h, rtlanal.c, sbitmap.c, sbitmap.h,
sched-deps.c, sched-ebb.c, sched-int.h, sched-rgn.c,
sched-vis.c, sdbout.c, sdbout.h, sibcall.c, simplify-rtx.c,
ssa-ccp.c, ssa-dce.c, ssa.c, ssa.h, stmt.c, stor-layout.c,
stringpool.c, system.h, timevar.c, timevar.def, timevar.h,
tlink.c, toplev.c, toplev.h, tree.c, tree.def, tree.h,
tsystem.h, unroll.c, unwind-dw2-fde.c, unwind-dw2-fde.h,
unwind-dw2.c, unwind-pe.h, unwind-sjlj.c, unwind.h,
unwind.inc, varasm.c, varray.c, varray.h, xcoffout.c,
xcoffout.h: replace "GNU CC" with "GCC".
From-SVN: r45105
2001-08-22 16:35:51 +02:00
This file is part of GCC .
2001-02-14 17:24:45 +01:00
Makefile.in, [...]: replace "GNU CC" with "GCC".
* Makefile.in, alias.c, basic-block.h, bb-reorder.c, bitmap.c,
bitmap.h, builtin-types.def, builtins.c, builtins.def,
c-aux-info.c, c-common.c, c-common.def, c-common.h,
c-convert.c, c-decl.c, c-dump.c, c-dump.h, c-errors.c,
c-format.c, c-lang.c, c-lex.c, c-lex.h, c-parse.in,
c-pragma.c, c-pragma.h, c-semantics.c, c-tree.h, c-typeck.c,
caller-save.c, calls.c, collect2.c, collect2.h, combine.c,
conditions.h, config.gcc, configure.frag, configure.in,
conflict.c, convert.c, convert.h, cppspec.c, crtstuff.c,
cse.c, cselib.c, cselib.h, dbxout.c, dbxout.h, defaults.h,
dependence.c, df.c, df.h, diagnostic.c, diagnostic.h,
doloop.c, dominance.c, dwarf.h, dwarf2.h, dwarf2asm.c,
dwarf2asm.h, dwarf2out.c, dwarf2out.h, dwarfout.c,
emit-rtl.c, errors.c, errors.h, except.c, except.h,
exgettext, explow.c, expmed.c, expr.c, expr.h, final.c,
fixproto, flags.h, flow.c, fold-const.c, fp-test.c,
function.c, function.h, gbl-ctors.h, gcc.c, gcc.h, gcc.hlp,
gccspec.c, gcov-io.h, gcse.c, genattr.c, genattrtab.c,
gencheck.c, gencodes.c, genconfig.c, genemit.c,
genextract.c, genflags.c, gengenrtl.c, genmultilib,
genopinit.c, genoutput.c, genpeep.c, genrecog.c,
gensupport.c, gensupport.h, ggc-callbacks.c, ggc-common.c,
ggc-none.c, ggc-page.c, ggc-simple.c, ggc.h, global.c,
graph.c, graph.h, gthr-aix.h, gthr-dce.h, gthr-posix.h,
gthr-rtems.h, gthr-single.h, gthr-solaris.h, gthr-vxworks.h,
gthr-win32.h, gthr.h, haifa-sched.c, halfpic.c, halfpic.h,
hard-reg-set.h, hwint.h, ifcvt.c, input.h, insn-addr.h,
integrate.c, integrate.h, jump.c, lcm.c, libgcc2.c,
libgcc2.h, lists.c, local-alloc.c, loop.c, loop.h,
machmode.def, machmode.h, main.c, mbchar.c, mbchar.h,
mips-tdump.c, mips-tfile.c, mklibgcc.in, mkmap-flat.awk,
mkmap-symver.awk, optabs.c, output.h, params.c, params.def,
params.h, predict.c, predict.def, predict.h, prefix.c,
prefix.h, print-rtl.c, print-tree.c, profile.c, protoize.c,
read-rtl.c, real.c, real.h, recog.c, recog.h, reg-stack.c,
regclass.c, regmove.c, regrename.c, regs.h, reload.c,
reload.h, reload1.c, reorg.c, resource.c, resource.h, rtl.c,
rtl.def, rtl.h, rtlanal.c, sbitmap.c, sbitmap.h,
sched-deps.c, sched-ebb.c, sched-int.h, sched-rgn.c,
sched-vis.c, sdbout.c, sdbout.h, sibcall.c, simplify-rtx.c,
ssa-ccp.c, ssa-dce.c, ssa.c, ssa.h, stmt.c, stor-layout.c,
stringpool.c, system.h, timevar.c, timevar.def, timevar.h,
tlink.c, toplev.c, toplev.h, tree.c, tree.def, tree.h,
tsystem.h, unroll.c, unwind-dw2-fde.c, unwind-dw2-fde.h,
unwind-dw2.c, unwind-pe.h, unwind-sjlj.c, unwind.h,
unwind.inc, varasm.c, varray.c, varray.h, xcoffout.c,
xcoffout.h: replace "GNU CC" with "GCC".
From-SVN: r45105
2001-08-22 16:35:51 +02:00
GCC is free software ; you can redistribute it and / or modify it under
the terms of the GNU General Public License as published by the Free
2007-07-26 10:37:01 +02:00
Software Foundation ; either version 3 , or ( at your option ) any later
Makefile.in, [...]: replace "GNU CC" with "GCC".
* Makefile.in, alias.c, basic-block.h, bb-reorder.c, bitmap.c,
bitmap.h, builtin-types.def, builtins.c, builtins.def,
c-aux-info.c, c-common.c, c-common.def, c-common.h,
c-convert.c, c-decl.c, c-dump.c, c-dump.h, c-errors.c,
c-format.c, c-lang.c, c-lex.c, c-lex.h, c-parse.in,
c-pragma.c, c-pragma.h, c-semantics.c, c-tree.h, c-typeck.c,
caller-save.c, calls.c, collect2.c, collect2.h, combine.c,
conditions.h, config.gcc, configure.frag, configure.in,
conflict.c, convert.c, convert.h, cppspec.c, crtstuff.c,
cse.c, cselib.c, cselib.h, dbxout.c, dbxout.h, defaults.h,
dependence.c, df.c, df.h, diagnostic.c, diagnostic.h,
doloop.c, dominance.c, dwarf.h, dwarf2.h, dwarf2asm.c,
dwarf2asm.h, dwarf2out.c, dwarf2out.h, dwarfout.c,
emit-rtl.c, errors.c, errors.h, except.c, except.h,
exgettext, explow.c, expmed.c, expr.c, expr.h, final.c,
fixproto, flags.h, flow.c, fold-const.c, fp-test.c,
function.c, function.h, gbl-ctors.h, gcc.c, gcc.h, gcc.hlp,
gccspec.c, gcov-io.h, gcse.c, genattr.c, genattrtab.c,
gencheck.c, gencodes.c, genconfig.c, genemit.c,
genextract.c, genflags.c, gengenrtl.c, genmultilib,
genopinit.c, genoutput.c, genpeep.c, genrecog.c,
gensupport.c, gensupport.h, ggc-callbacks.c, ggc-common.c,
ggc-none.c, ggc-page.c, ggc-simple.c, ggc.h, global.c,
graph.c, graph.h, gthr-aix.h, gthr-dce.h, gthr-posix.h,
gthr-rtems.h, gthr-single.h, gthr-solaris.h, gthr-vxworks.h,
gthr-win32.h, gthr.h, haifa-sched.c, halfpic.c, halfpic.h,
hard-reg-set.h, hwint.h, ifcvt.c, input.h, insn-addr.h,
integrate.c, integrate.h, jump.c, lcm.c, libgcc2.c,
libgcc2.h, lists.c, local-alloc.c, loop.c, loop.h,
machmode.def, machmode.h, main.c, mbchar.c, mbchar.h,
mips-tdump.c, mips-tfile.c, mklibgcc.in, mkmap-flat.awk,
mkmap-symver.awk, optabs.c, output.h, params.c, params.def,
params.h, predict.c, predict.def, predict.h, prefix.c,
prefix.h, print-rtl.c, print-tree.c, profile.c, protoize.c,
read-rtl.c, real.c, real.h, recog.c, recog.h, reg-stack.c,
regclass.c, regmove.c, regrename.c, regs.h, reload.c,
reload.h, reload1.c, reorg.c, resource.c, resource.h, rtl.c,
rtl.def, rtl.h, rtlanal.c, sbitmap.c, sbitmap.h,
sched-deps.c, sched-ebb.c, sched-int.h, sched-rgn.c,
sched-vis.c, sdbout.c, sdbout.h, sibcall.c, simplify-rtx.c,
ssa-ccp.c, ssa-dce.c, ssa.c, ssa.h, stmt.c, stor-layout.c,
stringpool.c, system.h, timevar.c, timevar.def, timevar.h,
tlink.c, toplev.c, toplev.h, tree.c, tree.def, tree.h,
tsystem.h, unroll.c, unwind-dw2-fde.c, unwind-dw2-fde.h,
unwind-dw2.c, unwind-pe.h, unwind-sjlj.c, unwind.h,
unwind.inc, varasm.c, varray.c, varray.h, xcoffout.c,
xcoffout.h: replace "GNU CC" with "GCC".
From-SVN: r45105
2001-08-22 16:35:51 +02:00
version .
2001-02-14 17:24:45 +01:00
Makefile.in, [...]: replace "GNU CC" with "GCC".
* Makefile.in, alias.c, basic-block.h, bb-reorder.c, bitmap.c,
bitmap.h, builtin-types.def, builtins.c, builtins.def,
c-aux-info.c, c-common.c, c-common.def, c-common.h,
c-convert.c, c-decl.c, c-dump.c, c-dump.h, c-errors.c,
c-format.c, c-lang.c, c-lex.c, c-lex.h, c-parse.in,
c-pragma.c, c-pragma.h, c-semantics.c, c-tree.h, c-typeck.c,
caller-save.c, calls.c, collect2.c, collect2.h, combine.c,
conditions.h, config.gcc, configure.frag, configure.in,
conflict.c, convert.c, convert.h, cppspec.c, crtstuff.c,
cse.c, cselib.c, cselib.h, dbxout.c, dbxout.h, defaults.h,
dependence.c, df.c, df.h, diagnostic.c, diagnostic.h,
doloop.c, dominance.c, dwarf.h, dwarf2.h, dwarf2asm.c,
dwarf2asm.h, dwarf2out.c, dwarf2out.h, dwarfout.c,
emit-rtl.c, errors.c, errors.h, except.c, except.h,
exgettext, explow.c, expmed.c, expr.c, expr.h, final.c,
fixproto, flags.h, flow.c, fold-const.c, fp-test.c,
function.c, function.h, gbl-ctors.h, gcc.c, gcc.h, gcc.hlp,
gccspec.c, gcov-io.h, gcse.c, genattr.c, genattrtab.c,
gencheck.c, gencodes.c, genconfig.c, genemit.c,
genextract.c, genflags.c, gengenrtl.c, genmultilib,
genopinit.c, genoutput.c, genpeep.c, genrecog.c,
gensupport.c, gensupport.h, ggc-callbacks.c, ggc-common.c,
ggc-none.c, ggc-page.c, ggc-simple.c, ggc.h, global.c,
graph.c, graph.h, gthr-aix.h, gthr-dce.h, gthr-posix.h,
gthr-rtems.h, gthr-single.h, gthr-solaris.h, gthr-vxworks.h,
gthr-win32.h, gthr.h, haifa-sched.c, halfpic.c, halfpic.h,
hard-reg-set.h, hwint.h, ifcvt.c, input.h, insn-addr.h,
integrate.c, integrate.h, jump.c, lcm.c, libgcc2.c,
libgcc2.h, lists.c, local-alloc.c, loop.c, loop.h,
machmode.def, machmode.h, main.c, mbchar.c, mbchar.h,
mips-tdump.c, mips-tfile.c, mklibgcc.in, mkmap-flat.awk,
mkmap-symver.awk, optabs.c, output.h, params.c, params.def,
params.h, predict.c, predict.def, predict.h, prefix.c,
prefix.h, print-rtl.c, print-tree.c, profile.c, protoize.c,
read-rtl.c, real.c, real.h, recog.c, recog.h, reg-stack.c,
regclass.c, regmove.c, regrename.c, regs.h, reload.c,
reload.h, reload1.c, reorg.c, resource.c, resource.h, rtl.c,
rtl.def, rtl.h, rtlanal.c, sbitmap.c, sbitmap.h,
sched-deps.c, sched-ebb.c, sched-int.h, sched-rgn.c,
sched-vis.c, sdbout.c, sdbout.h, sibcall.c, simplify-rtx.c,
ssa-ccp.c, ssa-dce.c, ssa.c, ssa.h, stmt.c, stor-layout.c,
stringpool.c, system.h, timevar.c, timevar.def, timevar.h,
tlink.c, toplev.c, toplev.h, tree.c, tree.def, tree.h,
tsystem.h, unroll.c, unwind-dw2-fde.c, unwind-dw2-fde.h,
unwind-dw2.c, unwind-pe.h, unwind-sjlj.c, unwind.h,
unwind.inc, varasm.c, varray.c, varray.h, xcoffout.c,
xcoffout.h: replace "GNU CC" with "GCC".
From-SVN: r45105
2001-08-22 16:35:51 +02:00
GCC is distributed in the hope that it will be useful , but WITHOUT ANY
WARRANTY ; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE . See the GNU General Public License
for more details .
2001-02-14 17:24:45 +01:00
You should have received a copy of the GNU General Public License
2007-07-26 10:37:01 +02:00
along with GCC ; see the file COPYING3 . If not see
< http : / / www . gnu . org / licenses / > . * /
2001-02-14 17:24:45 +01:00
/ * This file contains definitions for language - independent
2005-07-28 02:40:06 +02:00
parameters . The DEFPARAM macro takes 6 arguments :
2001-02-14 17:24:45 +01:00
2001-02-21 17:11:59 +01:00
- The enumeral corresponding to this parameter .
2001-02-14 17:24:45 +01:00
2009-11-25 11:55:54 +01:00
- The name that can be used to set this parameter using the
2001-02-14 17:24:45 +01:00
command - line option ` -- param < name >= < value > ' .
- A help string explaining how the parameter is used .
2001-05-01 20:51:06 +02:00
- A default value for the parameter .
2005-07-28 06:57:07 +02:00
- The minimum acceptable value for the parameter .
2005-07-28 02:40:06 +02:00
2005-07-28 06:57:07 +02:00
- The maximum acceptable value for the parameter ( if greater than
2005-07-28 02:40:06 +02:00
the minimum ) .
2015-10-13 12:08:59 +02:00
The DEFPARAMENUM < N > macro is similar , but instead of the minumum and maximum
arguments , it contains a list of < N > allowed strings , corresponding to
integer values 0 .. < N > - 1 . Note that the default argument needs to be
specified as one of the allowed strings , rather than an integer value .
2001-05-01 20:51:06 +02:00
Be sure to add an entry to invoke . texi summarizing the parameter . * /
2001-02-14 17:24:45 +01:00
optabs.c (expand_abs_nojump): Update BRANCH_COST call.
* optabs.c (expand_abs_nojump): Update BRANCH_COST call.
* fold-cost.c (LOGICAL_OP_NON_SHORT_CIRCUIT, fold_truthop): Likewise.
* dojump.c (do_jump): Likewise.
* ifcvt.c (MAX_CONDITIONAL_EXECUTE): Likewise.
(note-if_info): Add BRANCH_COST.
(noce_try_store_flag_constants, noce_try_addcc, noce_try_store_flag_mask,
noce_try_cmove_arith, noce_try_cmove_arith, noce_try_cmove_arith,
noce_find_if_block, find_if_case_1, find_if_case_2): Use compuated
branch cost.
* expr.h (BRANCH_COST): Update default.
* predict.c (predictable_edge_p): New function.
* expmed.c (expand_smod_pow2, expand_sdiv_pow2, emit_store_flag):
Update BRANCH_COST call.
* basic-block.h (predictable_edge_p): Declare.
* config/alpha/alpha.h (BRANCH_COST): Update.
* config/frv/frv.h (BRANCH_COST): Update.
* config/s390/s390.h (BRANCH_COST): Update.
* config/spu/spu.h (BRANCH_COST): Update.
* config/sparc/sparc.h (BRANCH_COST): Update.
* config/m32r/m32r.h (BRANCH_COST): Update.
* config/i386/i386.h (BRANCH_COST): Update.
* config/i386/i386.c (ix86_expand_int_movcc): Update use of BRANCH_COST.
* config/sh/sh.h (BRANCH_COST): Update.
* config/pdp11/pdp11.h (BRANCH_COST): Update.
* config/avr/avr.h (BRANCH_COST): Update.
* config/crx/crx.h (BRANCH_COST): Update.
* config/xtensa/xtensa.h (BRANCH_COST): Update.
* config/stormy16/stormy16.h (BRANCH_COST): Update.
* config/m68hc11/m68hc11.h (BRANCH_COST): Update.
* config/iq2000/iq2000.h (BRANCH_COST): Update.
* config/ia64/ia64.h (BRANCH_COST): Update.
* config/rs6000/rs6000.h (BRANCH_COST): Update.
* config/arc/arc.h (BRANCH_COST): Update.
* config/score/score.h (BRANCH_COST): Update.
* config/arm/arm.h (BRANCH_COST): Update.
* config/pa/pa.h (BRANCH_COST): Update.
* config/mips/mips.h (BRANCH_COST): Update.
* config/vax/vax.h (BRANCH_COST): Update.
* config/h8300/h8300.h (BRANCH_COST): Update.
* params.def (PARAM_PREDICTABLE_BRANCH_OUTCOME): New.
* doc/invoke.texi (predictable-branch-cost-outcome): Document.
* doc/tm.texi (BRANCH_COST): Update.
From-SVN: r139804
2008-08-30 16:19:01 +02:00
/ * When branch is predicted to be taken with probability lower than this
threshold ( in percent ) , then it is considered well predictable . * /
DEFPARAM ( PARAM_PREDICTABLE_BRANCH_OUTCOME ,
"predictable-branch-outcome" ,
2015-10-22 00:24:41 +02:00
"Maximal estimated outcome of branch considered predictable." ,
optabs.c (expand_abs_nojump): Update BRANCH_COST call.
* optabs.c (expand_abs_nojump): Update BRANCH_COST call.
* fold-cost.c (LOGICAL_OP_NON_SHORT_CIRCUIT, fold_truthop): Likewise.
* dojump.c (do_jump): Likewise.
* ifcvt.c (MAX_CONDITIONAL_EXECUTE): Likewise.
(note-if_info): Add BRANCH_COST.
(noce_try_store_flag_constants, noce_try_addcc, noce_try_store_flag_mask,
noce_try_cmove_arith, noce_try_cmove_arith, noce_try_cmove_arith,
noce_find_if_block, find_if_case_1, find_if_case_2): Use compuated
branch cost.
* expr.h (BRANCH_COST): Update default.
* predict.c (predictable_edge_p): New function.
* expmed.c (expand_smod_pow2, expand_sdiv_pow2, emit_store_flag):
Update BRANCH_COST call.
* basic-block.h (predictable_edge_p): Declare.
* config/alpha/alpha.h (BRANCH_COST): Update.
* config/frv/frv.h (BRANCH_COST): Update.
* config/s390/s390.h (BRANCH_COST): Update.
* config/spu/spu.h (BRANCH_COST): Update.
* config/sparc/sparc.h (BRANCH_COST): Update.
* config/m32r/m32r.h (BRANCH_COST): Update.
* config/i386/i386.h (BRANCH_COST): Update.
* config/i386/i386.c (ix86_expand_int_movcc): Update use of BRANCH_COST.
* config/sh/sh.h (BRANCH_COST): Update.
* config/pdp11/pdp11.h (BRANCH_COST): Update.
* config/avr/avr.h (BRANCH_COST): Update.
* config/crx/crx.h (BRANCH_COST): Update.
* config/xtensa/xtensa.h (BRANCH_COST): Update.
* config/stormy16/stormy16.h (BRANCH_COST): Update.
* config/m68hc11/m68hc11.h (BRANCH_COST): Update.
* config/iq2000/iq2000.h (BRANCH_COST): Update.
* config/ia64/ia64.h (BRANCH_COST): Update.
* config/rs6000/rs6000.h (BRANCH_COST): Update.
* config/arc/arc.h (BRANCH_COST): Update.
* config/score/score.h (BRANCH_COST): Update.
* config/arm/arm.h (BRANCH_COST): Update.
* config/pa/pa.h (BRANCH_COST): Update.
* config/mips/mips.h (BRANCH_COST): Update.
* config/vax/vax.h (BRANCH_COST): Update.
* config/h8300/h8300.h (BRANCH_COST): Update.
* params.def (PARAM_PREDICTABLE_BRANCH_OUTCOME): New.
* doc/invoke.texi (predictable-branch-cost-outcome): Document.
* doc/tm.texi (BRANCH_COST): Update.
From-SVN: r139804
2008-08-30 16:19:01 +02:00
2 , 0 , 50 )
2012-11-08 17:46:18 +01:00
DEFPARAM ( PARAM_INLINE_MIN_SPEEDUP ,
"inline-min-speedup" ,
2017-01-09 22:48:33 +01:00
"The minimal estimated speedup allowing inliner to ignore inline-insns-single and inline-insns-auto." ,
2017-02-11 22:49:51 +01:00
8 , 0 , 0 )
2012-11-08 17:46:18 +01:00
2002-04-27 20:59:40 +02:00
/ * The single function inlining limit . This is the maximum size
of a function counted in internal gcc instructions ( not in
real machine instructions ) that is eligible for inlining
by the tree inliner .
2005-03-22 21:53:40 +01:00
The default value is 450 .
2002-04-27 20:59:40 +02:00
Only functions marked inline ( or methods defined in the class
2005-03-22 21:53:40 +01:00
definition for C ++ ) are affected by this .
2002-04-27 20:59:40 +02:00
There are more restrictions to inlining : If inlined functions
call other functions , the already inlined instructions are
2009-11-25 11:55:54 +01:00
counted and once the recursive inline limit ( see
2002-04-27 20:59:40 +02:00
"max-inline-insns" parameter ) is exceeded , the acceptable size
gets decreased . * /
DEFPARAM ( PARAM_MAX_INLINE_INSNS_SINGLE ,
"max-inline-insns-single" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions in a single function eligible for inlining." ,
2009-07-06 11:37:01 +02:00
400 , 0 , 0 )
2002-04-27 20:59:40 +02:00
2003-03-02 22:18:16 +01:00
/ * The single function inlining limit for functions that are
inlined by virtue of - finline - functions ( - O3 ) .
This limit should be chosen to be below or equal to the limit
that is applied to functions marked inlined ( or defined in the
class declaration in C ++ ) given by the "max-inline-insns-single"
parameter .
2010-06-28 02:31:05 +02:00
The default value is 40 . * /
2003-03-02 22:18:16 +01:00
DEFPARAM ( PARAM_MAX_INLINE_INSNS_AUTO ,
"max-inline-insns-auto" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions when automatically inlining." ,
2010-06-28 02:31:05 +02:00
40 , 0 , 0 )
2003-03-02 22:18:16 +01:00
2004-05-13 08:41:07 +02:00
DEFPARAM ( PARAM_MAX_INLINE_INSNS_RECURSIVE ,
"max-inline-insns-recursive" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions inline function can grow to via recursive inlining." ,
2009-07-06 11:37:01 +02:00
450 , 0 , 0 )
2004-05-13 08:41:07 +02:00
DEFPARAM ( PARAM_MAX_INLINE_INSNS_RECURSIVE_AUTO ,
"max-inline-insns-recursive-auto" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions non-inline function can grow to via recursive inlining." ,
2009-07-06 11:37:01 +02:00
450 , 0 , 0 )
2004-05-13 08:41:07 +02:00
DEFPARAM ( PARAM_MAX_INLINE_RECURSIVE_DEPTH ,
"max-inline-recursive-depth" ,
2015-10-22 00:24:41 +02:00
"The maximum depth of recursive inlining for inline functions." ,
2004-12-01 17:46:25 +01:00
8 , 0 , 0 )
2004-05-13 08:41:07 +02:00
DEFPARAM ( PARAM_MAX_INLINE_RECURSIVE_DEPTH_AUTO ,
"max-inline-recursive-depth-auto" ,
2015-10-22 00:24:41 +02:00
"The maximum depth of recursive inlining for non-inline functions." ,
2004-12-01 17:46:25 +01:00
8 , 0 , 0 )
2004-05-13 08:41:07 +02:00
2005-07-28 23:45:27 +02:00
DEFPARAM ( PARAM_MIN_INLINE_RECURSIVE_PROBABILITY ,
"min-inline-recursive-probability" ,
2015-10-22 00:24:41 +02:00
"Inline recursively only when the probability of call being executed exceeds the parameter." ,
2005-07-28 23:45:27 +02:00
10 , 0 , 0 )
2009-05-15 19:11:00 +02:00
/ * Limit of iterations of early inliner . This basically bounds number of
nested indirect calls early inliner can resolve . Deeper chains are still
handled by late inlining . * /
DEFPARAM ( PARAM_EARLY_INLINER_MAX_ITERATIONS ,
"max-early-inliner-iterations" ,
2015-10-22 00:24:41 +02:00
"The maximum number of nested indirect inlining performed by early inliner." ,
2013-02-05 10:13:48 +01:00
1 , 0 , 0 )
2009-05-15 19:11:00 +02:00
2010-11-10 21:23:09 +01:00
/* Limit on probability of entry BB. */
DEFPARAM ( PARAM_COMDAT_SHARING_PROBABILITY ,
"comdat-sharing-probability" ,
2015-10-22 00:24:41 +02:00
"Probability that COMDAT function will be shared with different compilation unit." ,
2010-11-10 21:23:09 +01:00
20 , 0 , 0 )
2010-06-25 15:13:31 +02:00
/* Limit on probability of entry BB. */
DEFPARAM ( PARAM_PARTIAL_INLINING_ENTRY_PROBABILITY ,
"partial-inlining-entry-probability" ,
2015-10-22 00:24:41 +02:00
"Maximum probability of the entry BB of split region (in percent relative to entry BB of the function) to make partial inlining happen." ,
2017-06-22 13:40:00 +02:00
70 , 0 , 100 )
2010-06-25 15:13:31 +02:00
Makefile.in (loop-unroll.o): Add VARRAY_H dependency.
2004-10-18 Revital Eres <eres@il.ibm.com>
* Makefile.in (loop-unroll.o): Add VARRAY_H dependency.
* loop-unroll.c: Include varray.h.
(struct var_to_expand, struct opt_info): Rename split_ivs_info to
opt_info and expand it to support variable expansion.
(analyze_insns_in_loop): Rename analyze_ivs_to_split and
expand it to support variable expansion.
(pt_info_start_duplication): Rename si_info_start_duplication.
(apply_opt_in_copies): Rename split_ivs_in_copies and add support
to the variable expansion optimization.
(free_opt_info): Rename free_si_info.
(analyze_insn_to_expand_var, referenced_in_one_insn_in_loop_p,
expand_var_during_unrolling, insert_var_expansion_initialization,
combine_var_copies_in_loop_exit, release_var_copies,
get_expansion): New functions.
(peel_loop_completely, unroll_loop_constant_iterations,
unroll_loop_runtime_iterations, peel_loop_simple,
unroll_loop_stupid): Change uses of struct si_info
to struct opt_info
and add uses of fvariable-expansion-in-unroller flag.
* params.def: Add parameter to restrict the number of expansions.
* params.h: (MAX_VARIABLE_EXPANSIONS): New define to restrict
the number of expansions.
* common.opt: (fvariable-expansion-in-unroller): New flag.
* doc/invoke.texi: (fvariable-expansion-in-unroller): Document.
From-SVN: r89197
2004-10-18 07:31:28 +02:00
/ * Limit the number of expansions created by the variable expansion
optimization to avoid register pressure . * /
DEFPARAM ( PARAM_MAX_VARIABLE_EXPANSIONS ,
"max-variable-expansions-in-unroller" ,
2015-10-22 00:24:41 +02:00
"If -fvariable-expansion-in-unroller is used, the maximum number of times that an individual variable will be expanded during loop unrolling." ,
2004-12-01 17:46:25 +01:00
1 , 0 , 0 )
2009-11-25 11:55:54 +01:00
2007-01-14 13:42:40 +01:00
/* Limit loop autovectorization to loops with large enough iteration count. */
DEFPARAM ( PARAM_MIN_VECT_LOOP_BOUND ,
"min-vect-loop-bound" ,
2015-10-22 00:24:41 +02:00
"If -ftree-vectorize is used, the minimal loop bound of a loop to be considered for vectorization." ,
2007-09-12 07:05:04 +02:00
1 , 1 , 0 )
2007-01-14 13:42:40 +01:00
2001-02-21 17:11:59 +01:00
/ * The maximum number of instructions to consider when looking for an
instruction to fill a delay slot . If more than this arbitrary
number of instructions is searched , the time savings from filling
the delay slot will be minimal so stop searching . Increasing
values mean more aggressive optimization , making the compile time
increase with probably small improvement in executable run time . * /
DEFPARAM ( PARAM_MAX_DELAY_SLOT_INSN_SEARCH ,
"max-delay-slot-insn-search" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions to consider to fill a delay slot." ,
2004-12-01 17:46:25 +01:00
100 , 0 , 0 )
2001-02-21 17:11:59 +01:00
2001-02-23 18:47:52 +01:00
/ * When trying to fill delay slots , the maximum number of instructions
to consider when searching for a block with valid live register
information . Increasing this arbitrarily chosen value means more
aggressive optimization , increasing the compile time . This
parameter should be removed when the delay slot code is rewritten
to maintain the control - flow graph . * /
DEFPARAM ( PARAM_MAX_DELAY_SLOT_LIVE_SEARCH ,
"max-delay-slot-live-search" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions to consider to find accurate live register information." ,
2004-12-01 17:46:25 +01:00
333 , 0 , 0 )
2001-02-23 18:47:52 +01:00
2009-11-25 11:55:54 +01:00
/ * This parameter limits the number of branch elements that the
ChangeLog.0, [...]: Fix spelling errors.
* ChangeLog.0, ChangeLog.1, ChangeLog.2, ChangeLog.3, ChangeLog.4,
ChangeLog.5, ChangeLog, FSFChangeLog.10, FSFChangeLog.11,
c-common.c, c-common.h, c-parse.in, c-typeck.c, cfg.c, config.gcc,
configure, configure.in, except.c, except.h, flow.c, function.c,
gcc.c, gcse.c, genrecog.c, libgcc2.c, loop.c, loop.h, params.def,
predict.def, predict.h, reg-stack.c, regmove.c, sched-deps.c,
sched-int.h, sibcall.c, ssa.c, stringpool.c, toplev.c, tree.c,
unroll.c: Fix spelling errors.
From-SVN: r46595
2001-10-28 21:09:17 +01:00
scheduler will track anti - dependencies through without resetting
2009-11-25 11:55:54 +01:00
the tracking mechanism . Large functions with few calls or barriers
can generate lists containing many 1000 ' s of dependencies . Generally
2001-07-26 15:59:22 +02:00
the compiler either uses all available memory , or runs for far too long . * /
DEFPARAM ( PARAM_MAX_PENDING_LIST_LENGTH ,
"max-pending-list-length" ,
2015-10-22 00:24:41 +02:00
"The maximum length of scheduling's pending operations list." ,
2004-12-01 17:46:25 +01:00
32 , 0 , 0 )
2001-07-26 15:59:22 +02:00
haifa-sched.c (modulo_ii, [...]): New static variables.
* haifa-sched.c (modulo_ii, modulo_max_states, modulo_n_insns,
modulo_insns_scheduled, modulo_iter0_max_uid, modulo_backtracks_left,
modulo_last_stage): New static variables.
(set_modulo_params, discard_delay_pairs_above): New functions.
(struct delay_pair): New member stages.
(htab_i2_traverse, htab_i1_traverse): New static functions.
(record_delay_slot_pair): New arg stages. All callers changed.
Record it.
(pair_delay): Take stages into account.
(add_delay_dependencies): Don't do so for stage pairs.
(struct sched_block_state): New member modulo_epilogue.
(save_backtrack_point): Don't set SHADOW_P for stage pairs.
(unschedule_insns_until): Decrease modulo_insns_scheduled.
Set HARD_DEP without using or.
(resolve_dependencies): New static function.
(prune_ready_list): New arg modulo_epilogue_p. All callers changed.
If it is true, allow only insns with INSN_EXACT_TICK set.
(schedule_block): Return bool, always true for normal scheduling,
true or false depending on modulo scheduling success otherwise.
Add bookkeeping for modulo scheduling, and call resolve_dependencies
on everything left over after a modulo schedule.
(haifa_sched_init): Remove check_cfg call. Clear modulo_ii.
* sched-int.h (schedule_block, record_delay_slot_pair): Adjust
declarations.
(set_modulo_params, discard_delay_pairs_above): Declare.
* params.def (PARAM_MAX_MODULO_BACKTRACK_ATTEMPS): New.
* doc/invoke.texi (--param): Document it.
From-SVN: r179383
2011-09-30 15:33:54 +02:00
/ * This parameter limits the number of backtracking attempts when using the
haifa scheduler for modulo scheduling . * /
DEFPARAM ( PARAM_MAX_MODULO_BACKTRACK_ATTEMPTS ,
"max-modulo-backtrack-attempts" ,
2015-10-22 00:24:41 +02:00
"The maximum number of backtrack attempts the scheduler should make when modulo scheduling a loop." ,
haifa-sched.c (modulo_ii, [...]): New static variables.
* haifa-sched.c (modulo_ii, modulo_max_states, modulo_n_insns,
modulo_insns_scheduled, modulo_iter0_max_uid, modulo_backtracks_left,
modulo_last_stage): New static variables.
(set_modulo_params, discard_delay_pairs_above): New functions.
(struct delay_pair): New member stages.
(htab_i2_traverse, htab_i1_traverse): New static functions.
(record_delay_slot_pair): New arg stages. All callers changed.
Record it.
(pair_delay): Take stages into account.
(add_delay_dependencies): Don't do so for stage pairs.
(struct sched_block_state): New member modulo_epilogue.
(save_backtrack_point): Don't set SHADOW_P for stage pairs.
(unschedule_insns_until): Decrease modulo_insns_scheduled.
Set HARD_DEP without using or.
(resolve_dependencies): New static function.
(prune_ready_list): New arg modulo_epilogue_p. All callers changed.
If it is true, allow only insns with INSN_EXACT_TICK set.
(schedule_block): Return bool, always true for normal scheduling,
true or false depending on modulo scheduling success otherwise.
Add bookkeeping for modulo scheduling, and call resolve_dependencies
on everything left over after a modulo schedule.
(haifa_sched_init): Remove check_cfg call. Clear modulo_ii.
* sched-int.h (schedule_block, record_delay_slot_pair): Adjust
declarations.
(set_modulo_params, discard_delay_pairs_above): Declare.
* params.def (PARAM_MAX_MODULO_BACKTRACK_ATTEMPS): New.
* doc/invoke.texi (--param): Document it.
From-SVN: r179383
2011-09-30 15:33:54 +02:00
40 , 0 , 0 )
cgraph.c (cgraph_max_uid): New global variable.
* cgraph.c (cgraph_max_uid): New global variable.
(cgraph_node): Set uid field.
(create_edge): Keep inline flags consistent.
(dump_cgraph): Dump more info.
* cgraph.h (struct cgraph_local_info): Remove inline_many and
can_inline_once; add inlinable, disgread_inline_limits, and self_insn
(struct cgraph_global_info): Add insns, calls, cloned_times,
will_be_output.
(struct cgraph_node): Add uid.
(struct cgraph_edge): Add inline_call.
(cgraph_max_uid, cgraph_inline_p): Declare.
* cgraph.c: Include params.h and fibheap.h
(cgraph_mark_functions_to_inline_once): Kill.
(INSNS_PER_CALL): New constant.
(ncalls_inlined, nfunctions_inlined, initial_insns, overall_insns): New
static variables.
(cgraph_finalize_function): Do not analyze inlining.
(cgraph_finalize_compilation_unit): Set inlining attributes.
(cgraph_mark_functions_to_output): More consistency checks.
(cgraph_optimize_function): Set current_function_decl to NULL.
(cgraph_expand_function): Use new inline flags.
(cgraph_postorder): Expand from cgraph_expand_functions.
(INLINED_TIMES, SET_INLINED_TIMES): New macros.
(cgraph_inlined_into, cgraph_inlined_callees,
cgraph_estimate_size_after_inlining, cgraph_estimate_growth,
cgraph_mark_inline, cgraph_check_inline_limits,
cgraph_default_inline_p, cgraph_decide_inling_of_small_functions,
cgraph_decide_inlining, cgraph_inline_p): New functions.
* params.def (PARAM_LARGE_FUNCTION_INSNS, PARAM_LARGE_FUNCTION_GROWTH,
PARAM_INLINE_UNIT_GROWTH): New parameters.
* tree-inline.c (struct inline_data): New field current_decl.
(expand_call_inline): Avoid forward declarations; use
inlinable_function_p.
(optimize_inline_calls): Set id.current_decl.
Co-Authored-By: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
From-SVN: r69262
2003-07-12 03:07:40 +02:00
DEFPARAM ( PARAM_LARGE_FUNCTION_INSNS ,
"large-function-insns" ,
2015-10-22 00:24:41 +02:00
"The size of function body to be considered large." ,
2005-03-22 21:53:40 +01:00
2700 , 0 , 0 )
cgraph.c (cgraph_max_uid): New global variable.
* cgraph.c (cgraph_max_uid): New global variable.
(cgraph_node): Set uid field.
(create_edge): Keep inline flags consistent.
(dump_cgraph): Dump more info.
* cgraph.h (struct cgraph_local_info): Remove inline_many and
can_inline_once; add inlinable, disgread_inline_limits, and self_insn
(struct cgraph_global_info): Add insns, calls, cloned_times,
will_be_output.
(struct cgraph_node): Add uid.
(struct cgraph_edge): Add inline_call.
(cgraph_max_uid, cgraph_inline_p): Declare.
* cgraph.c: Include params.h and fibheap.h
(cgraph_mark_functions_to_inline_once): Kill.
(INSNS_PER_CALL): New constant.
(ncalls_inlined, nfunctions_inlined, initial_insns, overall_insns): New
static variables.
(cgraph_finalize_function): Do not analyze inlining.
(cgraph_finalize_compilation_unit): Set inlining attributes.
(cgraph_mark_functions_to_output): More consistency checks.
(cgraph_optimize_function): Set current_function_decl to NULL.
(cgraph_expand_function): Use new inline flags.
(cgraph_postorder): Expand from cgraph_expand_functions.
(INLINED_TIMES, SET_INLINED_TIMES): New macros.
(cgraph_inlined_into, cgraph_inlined_callees,
cgraph_estimate_size_after_inlining, cgraph_estimate_growth,
cgraph_mark_inline, cgraph_check_inline_limits,
cgraph_default_inline_p, cgraph_decide_inling_of_small_functions,
cgraph_decide_inlining, cgraph_inline_p): New functions.
* params.def (PARAM_LARGE_FUNCTION_INSNS, PARAM_LARGE_FUNCTION_GROWTH,
PARAM_INLINE_UNIT_GROWTH): New parameters.
* tree-inline.c (struct inline_data): New field current_decl.
(expand_call_inline): Avoid forward declarations; use
inlinable_function_p.
(optimize_inline_calls): Set id.current_decl.
Co-Authored-By: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
From-SVN: r69262
2003-07-12 03:07:40 +02:00
DEFPARAM ( PARAM_LARGE_FUNCTION_GROWTH ,
"large-function-growth" ,
2015-10-22 00:24:41 +02:00
"Maximal growth due to inlining of large function (in percent)." ,
2004-12-01 17:46:25 +01:00
100 , 0 , 0 )
2005-11-15 13:27:33 +01:00
DEFPARAM ( PARAM_LARGE_UNIT_INSNS ,
"large-unit-insns" ,
2015-10-22 00:24:41 +02:00
"The size of translation unit to be considered large." ,
2005-11-15 13:27:33 +01:00
10000 , 0 , 0 )
cgraph.c (cgraph_max_uid): New global variable.
* cgraph.c (cgraph_max_uid): New global variable.
(cgraph_node): Set uid field.
(create_edge): Keep inline flags consistent.
(dump_cgraph): Dump more info.
* cgraph.h (struct cgraph_local_info): Remove inline_many and
can_inline_once; add inlinable, disgread_inline_limits, and self_insn
(struct cgraph_global_info): Add insns, calls, cloned_times,
will_be_output.
(struct cgraph_node): Add uid.
(struct cgraph_edge): Add inline_call.
(cgraph_max_uid, cgraph_inline_p): Declare.
* cgraph.c: Include params.h and fibheap.h
(cgraph_mark_functions_to_inline_once): Kill.
(INSNS_PER_CALL): New constant.
(ncalls_inlined, nfunctions_inlined, initial_insns, overall_insns): New
static variables.
(cgraph_finalize_function): Do not analyze inlining.
(cgraph_finalize_compilation_unit): Set inlining attributes.
(cgraph_mark_functions_to_output): More consistency checks.
(cgraph_optimize_function): Set current_function_decl to NULL.
(cgraph_expand_function): Use new inline flags.
(cgraph_postorder): Expand from cgraph_expand_functions.
(INLINED_TIMES, SET_INLINED_TIMES): New macros.
(cgraph_inlined_into, cgraph_inlined_callees,
cgraph_estimate_size_after_inlining, cgraph_estimate_growth,
cgraph_mark_inline, cgraph_check_inline_limits,
cgraph_default_inline_p, cgraph_decide_inling_of_small_functions,
cgraph_decide_inlining, cgraph_inline_p): New functions.
* params.def (PARAM_LARGE_FUNCTION_INSNS, PARAM_LARGE_FUNCTION_GROWTH,
PARAM_INLINE_UNIT_GROWTH): New parameters.
* tree-inline.c (struct inline_data): New field current_decl.
(expand_call_inline): Avoid forward declarations; use
inlinable_function_p.
(optimize_inline_calls): Set id.current_decl.
Co-Authored-By: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
From-SVN: r69262
2003-07-12 03:07:40 +02:00
DEFPARAM ( PARAM_INLINE_UNIT_GROWTH ,
"inline-unit-growth" ,
2015-10-22 00:24:41 +02:00
"How much can given compilation unit grow because of the inlining (in percent)." ,
2015-04-04 23:07:46 +02:00
20 , 0 , 0 )
2008-08-24 22:09:32 +02:00
DEFPARAM ( PARAM_IPCP_UNIT_GROWTH ,
"ipcp-unit-growth" ,
2015-10-22 00:24:41 +02:00
"How much can given compilation unit grow because of the interprocedural constant propagation (in percent)." ,
2008-08-24 22:09:32 +02:00
10 , 0 , 0 )
2009-05-25 23:59:24 +02:00
DEFPARAM ( PARAM_EARLY_INLINING_INSNS ,
"early-inlining-insns" ,
2015-10-22 00:24:41 +02:00
"Maximal estimated growth of function body caused by early inlining of single call." ,
2015-01-16 04:08:33 +01:00
14 , 0 , 0 )
2006-11-28 11:53:16 +01:00
DEFPARAM ( PARAM_LARGE_STACK_FRAME ,
"large-stack-frame" ,
2015-10-22 00:24:41 +02:00
"The size of stack frame to be considered large." ,
2006-11-28 11:53:16 +01:00
256 , 0 , 0 )
DEFPARAM ( PARAM_STACK_FRAME_GROWTH ,
"large-stack-frame-growth" ,
2015-10-22 00:24:41 +02:00
"Maximal stack frame growth due to inlining (in percent)." ,
2006-11-28 11:53:16 +01:00
1000 , 0 , 0 )
cgraph.c (cgraph_max_uid): New global variable.
* cgraph.c (cgraph_max_uid): New global variable.
(cgraph_node): Set uid field.
(create_edge): Keep inline flags consistent.
(dump_cgraph): Dump more info.
* cgraph.h (struct cgraph_local_info): Remove inline_many and
can_inline_once; add inlinable, disgread_inline_limits, and self_insn
(struct cgraph_global_info): Add insns, calls, cloned_times,
will_be_output.
(struct cgraph_node): Add uid.
(struct cgraph_edge): Add inline_call.
(cgraph_max_uid, cgraph_inline_p): Declare.
* cgraph.c: Include params.h and fibheap.h
(cgraph_mark_functions_to_inline_once): Kill.
(INSNS_PER_CALL): New constant.
(ncalls_inlined, nfunctions_inlined, initial_insns, overall_insns): New
static variables.
(cgraph_finalize_function): Do not analyze inlining.
(cgraph_finalize_compilation_unit): Set inlining attributes.
(cgraph_mark_functions_to_output): More consistency checks.
(cgraph_optimize_function): Set current_function_decl to NULL.
(cgraph_expand_function): Use new inline flags.
(cgraph_postorder): Expand from cgraph_expand_functions.
(INLINED_TIMES, SET_INLINED_TIMES): New macros.
(cgraph_inlined_into, cgraph_inlined_callees,
cgraph_estimate_size_after_inlining, cgraph_estimate_growth,
cgraph_mark_inline, cgraph_check_inline_limits,
cgraph_default_inline_p, cgraph_decide_inling_of_small_functions,
cgraph_decide_inlining, cgraph_inline_p): New functions.
* params.def (PARAM_LARGE_FUNCTION_INSNS, PARAM_LARGE_FUNCTION_GROWTH,
PARAM_INLINE_UNIT_GROWTH): New parameters.
* tree-inline.c (struct inline_data): New field current_decl.
(expand_call_inline): Avoid forward declarations; use
inlinable_function_p.
(optimize_inline_calls): Set id.current_decl.
Co-Authored-By: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
From-SVN: r69262
2003-07-12 03:07:40 +02:00
2001-04-11 20:22:46 +02:00
/ * The GCSE optimization will be disabled if it would require
significantly more memory than this value . * /
DEFPARAM ( PARAM_MAX_GCSE_MEMORY ,
"max-gcse-memory" ,
2015-10-22 00:24:41 +02:00
"The maximum amount of memory to be allocated by GCSE." ,
2015-11-12 23:27:47 +01:00
128 * 1024 * 1024 , 0 , 0 )
2009-04-12 21:43:46 +02:00
2011-01-13 14:41:03 +01:00
/ * The GCSE optimization of an expression will avoided if the ratio of
insertions to deletions is greater than this value . * /
DEFPARAM ( PARAM_MAX_GCSE_INSERTION_RATIO ,
"max-gcse-insertion-ratio" ,
2015-10-22 00:24:41 +02:00
"The maximum ratio of insertions to deletions of expressions in GCSE." ,
2011-01-13 14:41:03 +01:00
20 , 0 , 0 )
common.opt: Add description of the new -fgcse-after-reload flag.
2004-03-03 Mostafa Hagog <mustafa@il.ibm.com>
* common.opt: Add description of the new -fgcse-after-reload flag.
* flags.h (flag_gcse_after_reload): Declaration of global variable.
* gcse.c (reg_used_on_edge ,reg_set_between_after_reload_p,
reg_used_between_after_reload_p, rtx get_avail_load_store_reg,
is_jump_table_basic_block, bb_has_well_behaved_predecessors,
get_bb_avail_insn, hash_scan_set_after_reload,
compute_hash_table_after_reload, eliminate_partially_redundant_loads,
gcse_after_reload, get_bb_avail_insn): New functions to implement
gcse-after-reload.
(gcse_after_reload_main): New function, the main entry point to
gcse-after-reload.
* rtl.h (gcse_after_reload_main): Declaration of the new function.
* opts.c (common_handle_option): Handle the -fgcse-after-reload flag.
* toplev.c (flag_gcse_after_reload): Initialization.
* passes.c (rest_of_handl_gcse2): Call gcse_after_reload_main.
* params.def (PARAM_GCSE_AFTER_RELOAD_PARTIAL_FRACTION,
PARAM_GCSE_AFTER_RELOAD_CRITICAL_FRACTION): New parameters for tuning
the gcse after reload optimization.
* params.h (GCSE_AFTER_RELOAD_PARTIAL_FRACTION,
GCSE_AFTER_RELOAD_CRITICAL_FRACTION): Two macros to access the tuning
parameters.
* doc/invoke.texi: Documentation for the new flag gcse-after-reload.
From-SVN: r78842
2004-03-03 17:32:45 +01:00
/ * This is the threshold ratio when to perform partial redundancy
elimination after reload . We perform partial redundancy elimination
when the following holds :
( Redundant load execution count )
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- - >= GCSE_AFTER_RELOAD_PARTIAL_FRACTION
( Added loads execution count ) * /
DEFPARAM ( PARAM_GCSE_AFTER_RELOAD_PARTIAL_FRACTION ,
"gcse-after-reload-partial-fraction" ,
2015-10-22 00:24:41 +02:00
"The threshold ratio for performing partial redundancy elimination after reload." ,
2004-12-01 17:46:25 +01:00
3 , 0 , 0 )
common.opt: Add description of the new -fgcse-after-reload flag.
2004-03-03 Mostafa Hagog <mustafa@il.ibm.com>
* common.opt: Add description of the new -fgcse-after-reload flag.
* flags.h (flag_gcse_after_reload): Declaration of global variable.
* gcse.c (reg_used_on_edge ,reg_set_between_after_reload_p,
reg_used_between_after_reload_p, rtx get_avail_load_store_reg,
is_jump_table_basic_block, bb_has_well_behaved_predecessors,
get_bb_avail_insn, hash_scan_set_after_reload,
compute_hash_table_after_reload, eliminate_partially_redundant_loads,
gcse_after_reload, get_bb_avail_insn): New functions to implement
gcse-after-reload.
(gcse_after_reload_main): New function, the main entry point to
gcse-after-reload.
* rtl.h (gcse_after_reload_main): Declaration of the new function.
* opts.c (common_handle_option): Handle the -fgcse-after-reload flag.
* toplev.c (flag_gcse_after_reload): Initialization.
* passes.c (rest_of_handl_gcse2): Call gcse_after_reload_main.
* params.def (PARAM_GCSE_AFTER_RELOAD_PARTIAL_FRACTION,
PARAM_GCSE_AFTER_RELOAD_CRITICAL_FRACTION): New parameters for tuning
the gcse after reload optimization.
* params.h (GCSE_AFTER_RELOAD_PARTIAL_FRACTION,
GCSE_AFTER_RELOAD_CRITICAL_FRACTION): Two macros to access the tuning
parameters.
* doc/invoke.texi: Documentation for the new flag gcse-after-reload.
From-SVN: r78842
2004-03-03 17:32:45 +01:00
/ * This is the threshold ratio of the critical edges execution count compared to
the redundant loads execution count that permits performing the load
redundancy elimination in gcse after reload . * /
DEFPARAM ( PARAM_GCSE_AFTER_RELOAD_CRITICAL_FRACTION ,
"gcse-after-reload-critical-fraction" ,
2015-10-22 00:24:41 +02:00
"The threshold ratio of critical edges execution count that permit performing redundancy elimination after reload." ,
2004-12-01 17:46:25 +01:00
10 , 0 , 0 )
2010-07-27 21:29:48 +02:00
/ * GCSE will use GCSE_COST_DISTANCE_RATION as a scaling factor
to calculate maximum distance for which an expression is allowed to move
from its rtx_cost . * /
DEFPARAM ( PARAM_GCSE_COST_DISTANCE_RATIO ,
"gcse-cost-distance-ratio" ,
2015-10-22 00:24:41 +02:00
"Scaling factor in calculation of maximum distance an expression can be moved by GCSE optimizations." ,
2010-07-27 21:29:48 +02:00
10 , 0 , 0 )
/ * GCSE won ' t restrict distance for which an expression with rtx_cost greater
than COSTS_N_INSN ( GCSE_UNRESTRICTED_COST ) is allowed to move . * /
DEFPARAM ( PARAM_GCSE_UNRESTRICTED_COST ,
"gcse-unrestricted-cost" ,
2015-10-22 00:24:41 +02:00
"Cost at which GCSE optimizations will not constraint the distance an expression can travel." ,
2010-07-27 21:29:48 +02:00
3 , 0 , 0 )
2010-07-27 21:48:15 +02:00
/ * How deep from a given basic block the dominator tree should be searched
for expressions to hoist to the block . The value of 0 will avoid limiting
the search . * /
DEFPARAM ( PARAM_MAX_HOIST_DEPTH ,
"max-hoist-depth" ,
2015-10-22 00:24:41 +02:00
"Maximum depth of search in the dominator tree for expressions to hoist." ,
2010-07-27 21:48:15 +02:00
30 , 0 , 0 )
2015-05-13 17:26:14 +02:00
/ * When synthesizing expnonentiation by a real constant operations using square
roots , this controls how deep sqrt chains we are willing to generate . * /
DEFPARAM ( PARAM_MAX_POW_SQRT_DEPTH ,
"max-pow-sqrt-depth" ,
2015-10-22 00:24:41 +02:00
"Maximum depth of sqrt chains to use when synthesizing exponentiation by a real constant." ,
2015-05-13 17:26:14 +02:00
5 , 1 , 32 )
2002-03-15 00:31:50 +01:00
/ * This parameter limits the number of insns in a loop that will be unrolled ,
2004-02-21 01:34:46 +01:00
and by how much the loop is unrolled .
2009-11-25 11:55:54 +01:00
2004-02-21 01:34:46 +01:00
This limit should be at most half of the peeling limits : loop unroller
decides to not unroll loops that iterate fewer than 2 * number of allowed
c-decl.c, [...]: Fix comment typos and formatting.
* c-decl.c, c-ppoutput.c, combine.c, cppfiles.c, dwarf2out.c,
expr.c, fold-const.c, gcc.c, haifa-sched.c, loop-iv.c,
params.def, read-rtl.c, rtl.c, rtlanal.c, toplev.c: Fix
comment typos and formatting. Follow spelling conventions.
From-SVN: r78555
2004-02-27 08:09:43 +01:00
unrollings and thus we would have loops that are neither peeled or unrolled
2004-02-21 01:34:46 +01:00
otherwise . * /
2002-03-15 00:31:50 +01:00
DEFPARAM ( PARAM_MAX_UNROLLED_INSNS ,
"max-unrolled-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions to consider to unroll in a loop." ,
2004-12-01 17:46:25 +01:00
200 , 0 , 0 )
cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* Makefile.in (loop-unroll.o): New.
* cfgloop.h (UAP_PEEL, UAP_UNROLL, UAP_UNROLL_ALL): New.
(unroll_and_peel_loops): Declare.
* alias.c (init_alias_analysis): Flag_unroll_loops renamed to
flag_old_unroll_loops.
* loop.c (loop_invariant_p): Ditto.
* unroll.c (unroll_loop): Flag_unroll_all_loops renamed to
flag_old_unroll_all_loops.
* flags.h (flag_unroll_loops): Renamed to flag_old_unroll_loops.
(flag_unroll_all_loops): Renamed to flag_old_unroll_all_loops.
* params.def (PARAM_MAX_UNROLLED_INSNS): Default value changed.
(PARAM_MAX_AVERAGE_UNROLLED_INSNS, PARAM_MAX_UNROLL_TIMES,
PARAM_MAX_PEELED_INSNS, PARAM_MAX_PEEL_TIMES,
PARAM_MAX_COMPLETELY_PEELED_INSNS, PARAM_MAX_COMPLETELY_PEEL_TIMES,
PARAM_MAX_ONCE_PEELED_INSNS): New.
* toplev.h (flag_old_unroll_loops, flag_old_unroll_all_loops): New.
(flag_unroll_loops, flag_unroll_all_loops): Used for new unroller
instead of old one.
(flag_peel_loops): New.
(lang_independent_options): The new flags added.
(rest_of_compilation): Call new unroller.
(process_options): Setup flags for coexistence of old and new unroller.
* doc/invoke.texi: Document new options.
* doc/passes.texi: Document new unroller pass.
From-SVN: r63462
2003-02-26 11:48:36 +01:00
/ * This parameter limits how many times the loop is unrolled depending
on number of insns really executed in each iteration . * /
DEFPARAM ( PARAM_MAX_AVERAGE_UNROLLED_INSNS ,
"max-average-unrolled-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions to consider to unroll in a loop on average." ,
2004-12-01 17:46:25 +01:00
80 , 0 , 0 )
cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* Makefile.in (loop-unroll.o): New.
* cfgloop.h (UAP_PEEL, UAP_UNROLL, UAP_UNROLL_ALL): New.
(unroll_and_peel_loops): Declare.
* alias.c (init_alias_analysis): Flag_unroll_loops renamed to
flag_old_unroll_loops.
* loop.c (loop_invariant_p): Ditto.
* unroll.c (unroll_loop): Flag_unroll_all_loops renamed to
flag_old_unroll_all_loops.
* flags.h (flag_unroll_loops): Renamed to flag_old_unroll_loops.
(flag_unroll_all_loops): Renamed to flag_old_unroll_all_loops.
* params.def (PARAM_MAX_UNROLLED_INSNS): Default value changed.
(PARAM_MAX_AVERAGE_UNROLLED_INSNS, PARAM_MAX_UNROLL_TIMES,
PARAM_MAX_PEELED_INSNS, PARAM_MAX_PEEL_TIMES,
PARAM_MAX_COMPLETELY_PEELED_INSNS, PARAM_MAX_COMPLETELY_PEEL_TIMES,
PARAM_MAX_ONCE_PEELED_INSNS): New.
* toplev.h (flag_old_unroll_loops, flag_old_unroll_all_loops): New.
(flag_unroll_loops, flag_unroll_all_loops): Used for new unroller
instead of old one.
(flag_peel_loops): New.
(lang_independent_options): The new flags added.
(rest_of_compilation): Call new unroller.
(process_options): Setup flags for coexistence of old and new unroller.
* doc/invoke.texi: Document new options.
* doc/passes.texi: Document new unroller pass.
From-SVN: r63462
2003-02-26 11:48:36 +01:00
/* The maximum number of unrollings of a single loop. */
DEFPARAM ( PARAM_MAX_UNROLL_TIMES ,
"max-unroll-times" ,
2015-10-22 00:24:41 +02:00
"The maximum number of unrollings of a single loop." ,
2004-12-01 17:46:25 +01:00
8 , 0 , 0 )
cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* Makefile.in (loop-unroll.o): New.
* cfgloop.h (UAP_PEEL, UAP_UNROLL, UAP_UNROLL_ALL): New.
(unroll_and_peel_loops): Declare.
* alias.c (init_alias_analysis): Flag_unroll_loops renamed to
flag_old_unroll_loops.
* loop.c (loop_invariant_p): Ditto.
* unroll.c (unroll_loop): Flag_unroll_all_loops renamed to
flag_old_unroll_all_loops.
* flags.h (flag_unroll_loops): Renamed to flag_old_unroll_loops.
(flag_unroll_all_loops): Renamed to flag_old_unroll_all_loops.
* params.def (PARAM_MAX_UNROLLED_INSNS): Default value changed.
(PARAM_MAX_AVERAGE_UNROLLED_INSNS, PARAM_MAX_UNROLL_TIMES,
PARAM_MAX_PEELED_INSNS, PARAM_MAX_PEEL_TIMES,
PARAM_MAX_COMPLETELY_PEELED_INSNS, PARAM_MAX_COMPLETELY_PEEL_TIMES,
PARAM_MAX_ONCE_PEELED_INSNS): New.
* toplev.h (flag_old_unroll_loops, flag_old_unroll_all_loops): New.
(flag_unroll_loops, flag_unroll_all_loops): Used for new unroller
instead of old one.
(flag_peel_loops): New.
(lang_independent_options): The new flags added.
(rest_of_compilation): Call new unroller.
(process_options): Setup flags for coexistence of old and new unroller.
* doc/invoke.texi: Document new options.
* doc/passes.texi: Document new unroller pass.
From-SVN: r63462
2003-02-26 11:48:36 +01:00
/* The maximum number of insns of a peeled loop. */
DEFPARAM ( PARAM_MAX_PEELED_INSNS ,
"max-peeled-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of insns of a peeled loop." ,
2012-11-16 18:27:30 +01:00
100 , 0 , 0 )
cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* Makefile.in (loop-unroll.o): New.
* cfgloop.h (UAP_PEEL, UAP_UNROLL, UAP_UNROLL_ALL): New.
(unroll_and_peel_loops): Declare.
* alias.c (init_alias_analysis): Flag_unroll_loops renamed to
flag_old_unroll_loops.
* loop.c (loop_invariant_p): Ditto.
* unroll.c (unroll_loop): Flag_unroll_all_loops renamed to
flag_old_unroll_all_loops.
* flags.h (flag_unroll_loops): Renamed to flag_old_unroll_loops.
(flag_unroll_all_loops): Renamed to flag_old_unroll_all_loops.
* params.def (PARAM_MAX_UNROLLED_INSNS): Default value changed.
(PARAM_MAX_AVERAGE_UNROLLED_INSNS, PARAM_MAX_UNROLL_TIMES,
PARAM_MAX_PEELED_INSNS, PARAM_MAX_PEEL_TIMES,
PARAM_MAX_COMPLETELY_PEELED_INSNS, PARAM_MAX_COMPLETELY_PEEL_TIMES,
PARAM_MAX_ONCE_PEELED_INSNS): New.
* toplev.h (flag_old_unroll_loops, flag_old_unroll_all_loops): New.
(flag_unroll_loops, flag_unroll_all_loops): Used for new unroller
instead of old one.
(flag_peel_loops): New.
(lang_independent_options): The new flags added.
(rest_of_compilation): Call new unroller.
(process_options): Setup flags for coexistence of old and new unroller.
* doc/invoke.texi: Document new options.
* doc/passes.texi: Document new unroller pass.
From-SVN: r63462
2003-02-26 11:48:36 +01:00
/* The maximum number of peelings of a single loop. */
DEFPARAM ( PARAM_MAX_PEEL_TIMES ,
"max-peel-times" ,
2015-10-22 00:24:41 +02:00
"The maximum number of peelings of a single loop." ,
2004-12-01 17:46:25 +01:00
16 , 0 , 0 )
2012-11-06 17:22:45 +01:00
/* The maximum number of peelings of a single loop that is peeled completely. */
DEFPARAM ( PARAM_MAX_PEEL_BRANCHES ,
"max-peel-branches" ,
2015-10-22 00:24:41 +02:00
"The maximum number of branches on the path through the peeled sequence." ,
2012-11-06 17:22:45 +01:00
32 , 0 , 0 )
cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* Makefile.in (loop-unroll.o): New.
* cfgloop.h (UAP_PEEL, UAP_UNROLL, UAP_UNROLL_ALL): New.
(unroll_and_peel_loops): Declare.
* alias.c (init_alias_analysis): Flag_unroll_loops renamed to
flag_old_unroll_loops.
* loop.c (loop_invariant_p): Ditto.
* unroll.c (unroll_loop): Flag_unroll_all_loops renamed to
flag_old_unroll_all_loops.
* flags.h (flag_unroll_loops): Renamed to flag_old_unroll_loops.
(flag_unroll_all_loops): Renamed to flag_old_unroll_all_loops.
* params.def (PARAM_MAX_UNROLLED_INSNS): Default value changed.
(PARAM_MAX_AVERAGE_UNROLLED_INSNS, PARAM_MAX_UNROLL_TIMES,
PARAM_MAX_PEELED_INSNS, PARAM_MAX_PEEL_TIMES,
PARAM_MAX_COMPLETELY_PEELED_INSNS, PARAM_MAX_COMPLETELY_PEEL_TIMES,
PARAM_MAX_ONCE_PEELED_INSNS): New.
* toplev.h (flag_old_unroll_loops, flag_old_unroll_all_loops): New.
(flag_unroll_loops, flag_unroll_all_loops): Used for new unroller
instead of old one.
(flag_peel_loops): New.
(lang_independent_options): The new flags added.
(rest_of_compilation): Call new unroller.
(process_options): Setup flags for coexistence of old and new unroller.
* doc/invoke.texi: Document new options.
* doc/passes.texi: Document new unroller pass.
From-SVN: r63462
2003-02-26 11:48:36 +01:00
/* The maximum number of insns of a peeled loop. */
DEFPARAM ( PARAM_MAX_COMPLETELY_PEELED_INSNS ,
"max-completely-peeled-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of insns of a completely peeled loop." ,
2014-11-22 20:02:12 +01:00
200 , 0 , 0 )
cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* Makefile.in (loop-unroll.o): New.
* cfgloop.h (UAP_PEEL, UAP_UNROLL, UAP_UNROLL_ALL): New.
(unroll_and_peel_loops): Declare.
* alias.c (init_alias_analysis): Flag_unroll_loops renamed to
flag_old_unroll_loops.
* loop.c (loop_invariant_p): Ditto.
* unroll.c (unroll_loop): Flag_unroll_all_loops renamed to
flag_old_unroll_all_loops.
* flags.h (flag_unroll_loops): Renamed to flag_old_unroll_loops.
(flag_unroll_all_loops): Renamed to flag_old_unroll_all_loops.
* params.def (PARAM_MAX_UNROLLED_INSNS): Default value changed.
(PARAM_MAX_AVERAGE_UNROLLED_INSNS, PARAM_MAX_UNROLL_TIMES,
PARAM_MAX_PEELED_INSNS, PARAM_MAX_PEEL_TIMES,
PARAM_MAX_COMPLETELY_PEELED_INSNS, PARAM_MAX_COMPLETELY_PEEL_TIMES,
PARAM_MAX_ONCE_PEELED_INSNS): New.
* toplev.h (flag_old_unroll_loops, flag_old_unroll_all_loops): New.
(flag_unroll_loops, flag_unroll_all_loops): Used for new unroller
instead of old one.
(flag_peel_loops): New.
(lang_independent_options): The new flags added.
(rest_of_compilation): Call new unroller.
(process_options): Setup flags for coexistence of old and new unroller.
* doc/invoke.texi: Document new options.
* doc/passes.texi: Document new unroller pass.
From-SVN: r63462
2003-02-26 11:48:36 +01:00
/* The maximum number of peelings of a single loop that is peeled completely. */
DEFPARAM ( PARAM_MAX_COMPLETELY_PEEL_TIMES ,
"max-completely-peel-times" ,
2015-10-22 00:24:41 +02:00
"The maximum number of peelings of a single loop that is peeled completely." ,
2004-12-01 17:46:25 +01:00
16 , 0 , 0 )
cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* cse.c (count_reg_usage): Fix handling of REG_EQUAL notes.
* Makefile.in (loop-unroll.o): New.
* cfgloop.h (UAP_PEEL, UAP_UNROLL, UAP_UNROLL_ALL): New.
(unroll_and_peel_loops): Declare.
* alias.c (init_alias_analysis): Flag_unroll_loops renamed to
flag_old_unroll_loops.
* loop.c (loop_invariant_p): Ditto.
* unroll.c (unroll_loop): Flag_unroll_all_loops renamed to
flag_old_unroll_all_loops.
* flags.h (flag_unroll_loops): Renamed to flag_old_unroll_loops.
(flag_unroll_all_loops): Renamed to flag_old_unroll_all_loops.
* params.def (PARAM_MAX_UNROLLED_INSNS): Default value changed.
(PARAM_MAX_AVERAGE_UNROLLED_INSNS, PARAM_MAX_UNROLL_TIMES,
PARAM_MAX_PEELED_INSNS, PARAM_MAX_PEEL_TIMES,
PARAM_MAX_COMPLETELY_PEELED_INSNS, PARAM_MAX_COMPLETELY_PEEL_TIMES,
PARAM_MAX_ONCE_PEELED_INSNS): New.
* toplev.h (flag_old_unroll_loops, flag_old_unroll_all_loops): New.
(flag_unroll_loops, flag_unroll_all_loops): Used for new unroller
instead of old one.
(flag_peel_loops): New.
(lang_independent_options): The new flags added.
(rest_of_compilation): Call new unroller.
(process_options): Setup flags for coexistence of old and new unroller.
* doc/invoke.texi: Document new options.
* doc/passes.texi: Document new unroller pass.
From-SVN: r63462
2003-02-26 11:48:36 +01:00
/* The maximum number of insns of a peeled loop that rolls only once. */
DEFPARAM ( PARAM_MAX_ONCE_PEELED_INSNS ,
"max-once-peeled-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of insns of a peeled loop that rolls only once." ,
2004-12-01 17:46:25 +01:00
400 , 0 , 0 )
2010-02-28 00:28:46 +01:00
/* The maximum depth of a loop nest we completely peel. */
DEFPARAM ( PARAM_MAX_UNROLL_ITERATIONS ,
"max-completely-peel-loop-nest-depth" ,
2015-10-22 00:24:41 +02:00
"The maximum depth of a loop nest we completely peel." ,
2010-02-28 00:28:46 +01:00
8 , 0 , 0 )
invoke.texi (-malign-double): Re-add lost warning.
* invoke.texi (-malign-double): Re-add lost warning.
* i386-protos.h (x86_output_mi_thunk): Declare.
* unix.h (ASM_OUTPUT_MI_THUNK): Move offline to ...
* i386.c (x86_output_mi_thunk): ... here; handle 64bits.
* dwarf2out.c (output_call_frame_info): Do not skip unwind info
when flag_asynchronous_unwind_tables is set.
* flags.h (flag_reorder_functions): Declare.
* function.c (prepare_function_start): Initialize frequnecy.
* params.def (HOT_BB_COUNT_FRACTION, HOT_BB_FREQUENCY_FRACTION): New paramters.
* Makefile.in (predict.o): Add dependency on target.h and params.h
* defaults.h (HOT_TEXT_SECTION_NAME,
UNLIKELY_EXECUTED_TEXT_SECTION_NAME): New macros.
* predict.c (choose_function_section): New function.
(estimate_bb_frequencies): Use it.
* toplev.c (flag_reorder_functions): New global variable.
(lang_independent_options): New.
(parse_options_and_default_flags): Set.
* varasm.c (assemble_start_function): Bypass functdion alignment
for never executed functions.
* invoke.texi (-freorder-blocks, -freorder-functions): Document.
(param hot-bb-count-fraction, hot-bb-frequency-fraction): New.
* tm.texi (HOT_TEXT_SECTION_NAME, UNLIKELY_EXECUTED_TEXT_SECTION_NAME):
Document.
Thu Jan 3 21:52:09 CET 2002 Jan Hubicka <jh@suse.cz>
* predict.c: Inlude profile.h
(MIN_COUNT): Rename to MIN_COUNT_FRACTION
(maybe_hot_bb_p, probably_cold_bb_p, probably_never_executed_bb_p):
Use the information about maximal counter in the program.
Thu Dec 20 22:14:00 CET 2001 Jan Hubicka <jh@suse.cz>
* basic-block.h (maybe_hot_bb_p, probably_cold_bb_p,
probably_never_executed_bb_p): New functions.
* cfgcleanup.c (outgoing_edges_match): Use them.
* predict.c (MIN_COUNT, MIN_FREQUENCY): New macros.
(maybe_hot_bb_p, probably_cold_bb_p,
probably_never_executed_bb_p): New functions.
* function.h (function): Add new field function_frequency.
* predict.c (compute_function_frequency): New function.
(estimate_probability): Call it.
From-SVN: r53478
2002-05-15 11:00:30 +02:00
cfgloop.h (fix_loop_placement, [...]): Declare.
* cfgloop.h (fix_loop_placement, can_duplicate_loop_p,
duplicate_loop_to_header_edge, loopify, remove_path, split_loop_bb):
Declare.
(DLTHE_FLAG_UPDATE_FREQ): New.
* cfgloopmanip.c (duplicate_loop, duplicate_subloops, copy_loops_to,
loop_redirect_edge, loop_delete_branch_edge, copy_bbs, remove_bbs,
rpe_enum_p, find_branch, alp_enum_p, add_loop, fix_loop_placements,
fix_bb_placement, fix_bb_placements, place_new_loop,
scale_loop_frequencies, scale_bbs_frequencies, record_exit_edges):
New static functions.
(fix_loop_placement, can_duplicate_loop_p,
duplicate_loop_to_header_edge, loopify, remove_path, split_loop_bb):
New functions.
* cfgloop.h (loop_optimizer_init, loop_optimizer_finalize,
unswitch_loops): Declare.
* loop-init.c: New file.
* loop-unswitch.c: New file.
* Makefile.in (loop-init.o, loop-unswitch.o): New.
* params.def (PARAM_MAX_UNSWITCH_INSNS, PARAM_MAX_UNSWITCH_LEVEL): New.
* toplev.c (DFI_loop2): New dump.
(flag_unswitch_loops): New.
(lang_independent_options): Add it.
(rest_of_compilation): Call new loop optimizer.
(parse_options_and_default_flags): Turn flag_unswitch_loops on with -O3.
From-SVN: r62578
2003-02-08 15:29:00 +01:00
/* The maximum number of insns of an unswitched loop. */
DEFPARAM ( PARAM_MAX_UNSWITCH_INSNS ,
"max-unswitch-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of insns of an unswitched loop." ,
2004-12-01 17:46:25 +01:00
50 , 0 , 0 )
cfgloop.h (fix_loop_placement, [...]): Declare.
* cfgloop.h (fix_loop_placement, can_duplicate_loop_p,
duplicate_loop_to_header_edge, loopify, remove_path, split_loop_bb):
Declare.
(DLTHE_FLAG_UPDATE_FREQ): New.
* cfgloopmanip.c (duplicate_loop, duplicate_subloops, copy_loops_to,
loop_redirect_edge, loop_delete_branch_edge, copy_bbs, remove_bbs,
rpe_enum_p, find_branch, alp_enum_p, add_loop, fix_loop_placements,
fix_bb_placement, fix_bb_placements, place_new_loop,
scale_loop_frequencies, scale_bbs_frequencies, record_exit_edges):
New static functions.
(fix_loop_placement, can_duplicate_loop_p,
duplicate_loop_to_header_edge, loopify, remove_path, split_loop_bb):
New functions.
* cfgloop.h (loop_optimizer_init, loop_optimizer_finalize,
unswitch_loops): Declare.
* loop-init.c: New file.
* loop-unswitch.c: New file.
* Makefile.in (loop-init.o, loop-unswitch.o): New.
* params.def (PARAM_MAX_UNSWITCH_INSNS, PARAM_MAX_UNSWITCH_LEVEL): New.
* toplev.c (DFI_loop2): New dump.
(flag_unswitch_loops): New.
(lang_independent_options): Add it.
(rest_of_compilation): Call new loop optimizer.
(parse_options_and_default_flags): Turn flag_unswitch_loops on with -O3.
From-SVN: r62578
2003-02-08 15:29:00 +01:00
/* The maximum level of recursion in unswitch_single_loop. */
DEFPARAM ( PARAM_MAX_UNSWITCH_LEVEL ,
"max-unswitch-level" ,
2015-10-22 00:24:41 +02:00
"The maximum number of unswitchings in a single loop." ,
2004-12-01 17:46:25 +01:00
3 , 0 , 0 )
cfgloop.h (fix_loop_placement, [...]): Declare.
* cfgloop.h (fix_loop_placement, can_duplicate_loop_p,
duplicate_loop_to_header_edge, loopify, remove_path, split_loop_bb):
Declare.
(DLTHE_FLAG_UPDATE_FREQ): New.
* cfgloopmanip.c (duplicate_loop, duplicate_subloops, copy_loops_to,
loop_redirect_edge, loop_delete_branch_edge, copy_bbs, remove_bbs,
rpe_enum_p, find_branch, alp_enum_p, add_loop, fix_loop_placements,
fix_bb_placement, fix_bb_placements, place_new_loop,
scale_loop_frequencies, scale_bbs_frequencies, record_exit_edges):
New static functions.
(fix_loop_placement, can_duplicate_loop_p,
duplicate_loop_to_header_edge, loopify, remove_path, split_loop_bb):
New functions.
* cfgloop.h (loop_optimizer_init, loop_optimizer_finalize,
unswitch_loops): Declare.
* loop-init.c: New file.
* loop-unswitch.c: New file.
* Makefile.in (loop-init.o, loop-unswitch.o): New.
* params.def (PARAM_MAX_UNSWITCH_INSNS, PARAM_MAX_UNSWITCH_LEVEL): New.
* toplev.c (DFI_loop2): New dump.
(flag_unswitch_loops): New.
(lang_independent_options): Add it.
(rest_of_compilation): Call new loop optimizer.
(parse_options_and_default_flags): Turn flag_unswitch_loops on with -O3.
From-SVN: r62578
2003-02-08 15:29:00 +01:00
2016-06-08 16:11:51 +02:00
/ * The maximum number of insns in loop header duplicated by he copy loop
headers pass . * /
DEFPARAM ( PARAM_MAX_LOOP_HEADER_INSNS ,
"max-loop-header-insns" ,
"The maximum number of insns in loop header duplicated by he copy loop headers pass." ,
20 , 0 , 0 )
2004-07-09 05:19:14 +02:00
/ * The maximum number of iterations of a loop the brute force algorithm
for analysis of # of iterations of the loop tries to evaluate . * /
DEFPARAM ( PARAM_MAX_ITERATIONS_TO_TRACK ,
"max-iterations-to-track" ,
2015-10-22 00:24:41 +02:00
"Bound on the number of iterations the brute force # of iterations analysis algorithm evaluates." ,
2004-12-01 17:46:25 +01:00
1000 , 0 , 0 )
2006-09-15 13:17:18 +02:00
/ * A cutoff to avoid costly computations of the number of iterations in
the doloop transformation . * /
DEFPARAM ( PARAM_MAX_ITERATIONS_COMPUTATION_COST ,
"max-iterations-computation-cost" ,
2015-10-22 00:24:41 +02:00
"Bound on the cost of an expression to compute the number of iterations." ,
2006-09-15 13:17:18 +02:00
10 , 0 , 0 )
2004-07-09 05:19:14 +02:00
ddg.h, [...]: New files.
* ddg.h, ddg.c, modulo-sched.c: New files.
* cfglayout.c (duplicate_insn_chain): Remove "static" and push
internals to "dupicate_insn".
(duplicate_insn): New function.
* cfglayout.h (duplicate_insn_chain, duplicate_insn): New
declarations.
* common.opt (fmodulo-sched): New flag.
* df.c (df_bb_regno_last_use_find, df_bb_regno_first_def_find):
Remove static and forward declaration.
(df_find_def, df_reg_used, df_bb_regno_last_def_find): New
functions.
* df.h (df_bb_regno_last_use_find, df_bb_regno_first_def_find,
df_bb_regno_last_def_find, df_find_def, df_reg_used): New
declarations.
* flags.h (flag_modulo_sched): New flag.
* opts.c (common_handle_option): Handle modulo-sched flag.
* params.def (max-sms-loop-number, sms-max-ii-factor,
sms-dfa-history, sms-loop-average-count-threshold): New
parameters.
* params.h (MAX_SMS_LOOP_NUMBER, SMS_MAX_II_FACTOR,
SMS_DFA_HISTORY, SMS_LOOP_AVERAGE_COUNT_THRESHOLD): New
parameters.
* passes.c ("sms", "sms-vcg"): New dumps.
(rest_of_handle_sched): Call sms_schedule.
* rtl.h (sms_schedule): New declaration.
* timevar.def (TV_SMS): New.
* toplev.c (flag_modulo_sched): Initialize.
(f_options): Handle -fmodulo-sched option.
* docs/invoke.texi: Document -fmodulo-sched & -dm options.
* docs/passes.texi: Document new SMS pass.f toplev.c doc/invoke.texi doc/passes.texi
Co-Authored-By: Mostafa Hagog <mustafa@il.ibm.com>
From-SVN: r82235
2004-05-25 14:55:06 +02:00
/* This parameter is used to tune SMS MAX II calculations. */
DEFPARAM ( PARAM_SMS_MAX_II_FACTOR ,
"sms-max-ii-factor" ,
2015-10-22 00:24:41 +02:00
"A factor for tuning the upper bound that swing modulo scheduler uses for scheduling a loop." ,
2004-12-01 17:46:25 +01:00
100 , 0 , 0 )
2011-04-17 16:32:24 +02:00
/* The minimum value of stage count that swing modulo scheduler will generate. */
DEFPARAM ( PARAM_SMS_MIN_SC ,
"sms-min-sc" ,
"The minimum value of stage count that swing modulo scheduler will generate." ,
2 , 1 , 1 )
ddg.h, [...]: New files.
* ddg.h, ddg.c, modulo-sched.c: New files.
* cfglayout.c (duplicate_insn_chain): Remove "static" and push
internals to "dupicate_insn".
(duplicate_insn): New function.
* cfglayout.h (duplicate_insn_chain, duplicate_insn): New
declarations.
* common.opt (fmodulo-sched): New flag.
* df.c (df_bb_regno_last_use_find, df_bb_regno_first_def_find):
Remove static and forward declaration.
(df_find_def, df_reg_used, df_bb_regno_last_def_find): New
functions.
* df.h (df_bb_regno_last_use_find, df_bb_regno_first_def_find,
df_bb_regno_last_def_find, df_find_def, df_reg_used): New
declarations.
* flags.h (flag_modulo_sched): New flag.
* opts.c (common_handle_option): Handle modulo-sched flag.
* params.def (max-sms-loop-number, sms-max-ii-factor,
sms-dfa-history, sms-loop-average-count-threshold): New
parameters.
* params.h (MAX_SMS_LOOP_NUMBER, SMS_MAX_II_FACTOR,
SMS_DFA_HISTORY, SMS_LOOP_AVERAGE_COUNT_THRESHOLD): New
parameters.
* passes.c ("sms", "sms-vcg"): New dumps.
(rest_of_handle_sched): Call sms_schedule.
* rtl.h (sms_schedule): New declaration.
* timevar.def (TV_SMS): New.
* toplev.c (flag_modulo_sched): Initialize.
(f_options): Handle -fmodulo-sched option.
* docs/invoke.texi: Document -fmodulo-sched & -dm options.
* docs/passes.texi: Document new SMS pass.f toplev.c doc/invoke.texi doc/passes.texi
Co-Authored-By: Mostafa Hagog <mustafa@il.ibm.com>
From-SVN: r82235
2004-05-25 14:55:06 +02:00
DEFPARAM ( PARAM_SMS_DFA_HISTORY ,
"sms-dfa-history" ,
2015-10-22 00:24:41 +02:00
"The number of cycles the swing modulo scheduler considers when checking conflicts using DFA." ,
2004-12-01 17:46:25 +01:00
0 , 0 , 0 )
ddg.h, [...]: New files.
* ddg.h, ddg.c, modulo-sched.c: New files.
* cfglayout.c (duplicate_insn_chain): Remove "static" and push
internals to "dupicate_insn".
(duplicate_insn): New function.
* cfglayout.h (duplicate_insn_chain, duplicate_insn): New
declarations.
* common.opt (fmodulo-sched): New flag.
* df.c (df_bb_regno_last_use_find, df_bb_regno_first_def_find):
Remove static and forward declaration.
(df_find_def, df_reg_used, df_bb_regno_last_def_find): New
functions.
* df.h (df_bb_regno_last_use_find, df_bb_regno_first_def_find,
df_bb_regno_last_def_find, df_find_def, df_reg_used): New
declarations.
* flags.h (flag_modulo_sched): New flag.
* opts.c (common_handle_option): Handle modulo-sched flag.
* params.def (max-sms-loop-number, sms-max-ii-factor,
sms-dfa-history, sms-loop-average-count-threshold): New
parameters.
* params.h (MAX_SMS_LOOP_NUMBER, SMS_MAX_II_FACTOR,
SMS_DFA_HISTORY, SMS_LOOP_AVERAGE_COUNT_THRESHOLD): New
parameters.
* passes.c ("sms", "sms-vcg"): New dumps.
(rest_of_handle_sched): Call sms_schedule.
* rtl.h (sms_schedule): New declaration.
* timevar.def (TV_SMS): New.
* toplev.c (flag_modulo_sched): Initialize.
(f_options): Handle -fmodulo-sched option.
* docs/invoke.texi: Document -fmodulo-sched & -dm options.
* docs/passes.texi: Document new SMS pass.f toplev.c doc/invoke.texi doc/passes.texi
Co-Authored-By: Mostafa Hagog <mustafa@il.ibm.com>
From-SVN: r82235
2004-05-25 14:55:06 +02:00
DEFPARAM ( PARAM_SMS_LOOP_AVERAGE_COUNT_THRESHOLD ,
"sms-loop-average-count-threshold" ,
2015-10-22 00:24:41 +02:00
"A threshold on the average loop count considered by the swing modulo scheduler." ,
2004-12-01 17:46:25 +01:00
0 , 0 , 0 )
ddg.h, [...]: New files.
* ddg.h, ddg.c, modulo-sched.c: New files.
* cfglayout.c (duplicate_insn_chain): Remove "static" and push
internals to "dupicate_insn".
(duplicate_insn): New function.
* cfglayout.h (duplicate_insn_chain, duplicate_insn): New
declarations.
* common.opt (fmodulo-sched): New flag.
* df.c (df_bb_regno_last_use_find, df_bb_regno_first_def_find):
Remove static and forward declaration.
(df_find_def, df_reg_used, df_bb_regno_last_def_find): New
functions.
* df.h (df_bb_regno_last_use_find, df_bb_regno_first_def_find,
df_bb_regno_last_def_find, df_find_def, df_reg_used): New
declarations.
* flags.h (flag_modulo_sched): New flag.
* opts.c (common_handle_option): Handle modulo-sched flag.
* params.def (max-sms-loop-number, sms-max-ii-factor,
sms-dfa-history, sms-loop-average-count-threshold): New
parameters.
* params.h (MAX_SMS_LOOP_NUMBER, SMS_MAX_II_FACTOR,
SMS_DFA_HISTORY, SMS_LOOP_AVERAGE_COUNT_THRESHOLD): New
parameters.
* passes.c ("sms", "sms-vcg"): New dumps.
(rest_of_handle_sched): Call sms_schedule.
* rtl.h (sms_schedule): New declaration.
* timevar.def (TV_SMS): New.
* toplev.c (flag_modulo_sched): Initialize.
(f_options): Handle -fmodulo-sched option.
* docs/invoke.texi: Document -fmodulo-sched & -dm options.
* docs/passes.texi: Document new SMS pass.f toplev.c doc/invoke.texi doc/passes.texi
Co-Authored-By: Mostafa Hagog <mustafa@il.ibm.com>
From-SVN: r82235
2004-05-25 14:55:06 +02:00
2012-11-23 09:49:43 +01:00
DEFPARAM ( HOT_BB_COUNT_WS_PERMILLE ,
"hot-bb-count-ws-permille" ,
"A basic block profile count is considered hot if it contributes to "
2015-10-22 00:24:41 +02:00
"the given permillage of the entire profiled execution." ,
2012-11-23 09:49:43 +01:00
999 , 0 , 1000 )
invoke.texi (-malign-double): Re-add lost warning.
* invoke.texi (-malign-double): Re-add lost warning.
* i386-protos.h (x86_output_mi_thunk): Declare.
* unix.h (ASM_OUTPUT_MI_THUNK): Move offline to ...
* i386.c (x86_output_mi_thunk): ... here; handle 64bits.
* dwarf2out.c (output_call_frame_info): Do not skip unwind info
when flag_asynchronous_unwind_tables is set.
* flags.h (flag_reorder_functions): Declare.
* function.c (prepare_function_start): Initialize frequnecy.
* params.def (HOT_BB_COUNT_FRACTION, HOT_BB_FREQUENCY_FRACTION): New paramters.
* Makefile.in (predict.o): Add dependency on target.h and params.h
* defaults.h (HOT_TEXT_SECTION_NAME,
UNLIKELY_EXECUTED_TEXT_SECTION_NAME): New macros.
* predict.c (choose_function_section): New function.
(estimate_bb_frequencies): Use it.
* toplev.c (flag_reorder_functions): New global variable.
(lang_independent_options): New.
(parse_options_and_default_flags): Set.
* varasm.c (assemble_start_function): Bypass functdion alignment
for never executed functions.
* invoke.texi (-freorder-blocks, -freorder-functions): Document.
(param hot-bb-count-fraction, hot-bb-frequency-fraction): New.
* tm.texi (HOT_TEXT_SECTION_NAME, UNLIKELY_EXECUTED_TEXT_SECTION_NAME):
Document.
Thu Jan 3 21:52:09 CET 2002 Jan Hubicka <jh@suse.cz>
* predict.c: Inlude profile.h
(MIN_COUNT): Rename to MIN_COUNT_FRACTION
(maybe_hot_bb_p, probably_cold_bb_p, probably_never_executed_bb_p):
Use the information about maximal counter in the program.
Thu Dec 20 22:14:00 CET 2001 Jan Hubicka <jh@suse.cz>
* basic-block.h (maybe_hot_bb_p, probably_cold_bb_p,
probably_never_executed_bb_p): New functions.
* cfgcleanup.c (outgoing_edges_match): Use them.
* predict.c (MIN_COUNT, MIN_FREQUENCY): New macros.
(maybe_hot_bb_p, probably_cold_bb_p,
probably_never_executed_bb_p): New functions.
* function.h (function): Add new field function_frequency.
* predict.c (compute_function_frequency): New function.
(estimate_probability): Call it.
From-SVN: r53478
2002-05-15 11:00:30 +02:00
DEFPARAM ( HOT_BB_FREQUENCY_FRACTION ,
"hot-bb-frequency-fraction" ,
2015-10-22 00:24:41 +02:00
"Select fraction of the maximal frequency of executions of basic block in function given basic block needs to have to be considered hot." ,
2004-12-01 17:46:25 +01:00
1000 , 0 , 0 )
2005-11-05 01:55:23 +01:00
2013-10-18 19:25:44 +02:00
DEFPARAM ( UNLIKELY_BB_COUNT_FRACTION ,
"unlikely-bb-count-fraction" ,
2015-10-22 00:24:41 +02:00
"The minimum fraction of profile runs a given basic block execution count must be not to be considered unlikely." ,
2013-10-18 19:25:44 +02:00
20 , 1 , 10000 )
2007-10-09 09:07:15 +02:00
DEFPARAM ( PARAM_ALIGN_THRESHOLD ,
"align-threshold" ,
2015-10-22 00:24:41 +02:00
"Select fraction of the maximal frequency of executions of basic block in function given basic block get alignment." ,
2013-04-11 18:20:26 +02:00
100 , 1 , 0 )
2007-10-09 09:07:15 +02:00
DEFPARAM ( PARAM_ALIGN_LOOP_ITERATIONS ,
"align-loop-iterations" ,
2016-06-20 13:12:07 +02:00
"Loops iterating at least selected number of iterations will get loop alignment.." ,
2007-10-09 09:07:15 +02:00
4 , 0 , 0 )
2005-11-05 01:55:23 +01:00
/ * For guessed profiles , the loops having unknown number of iterations
are predicted to iterate relatively few ( 10 ) times at average .
For functions containing one loop with large known number of iterations
and other loops having unbounded loops we would end up predicting all
c-typeck.c, [...]: Fix comment typos.
* c-typeck.c, config/i386/netware.h, config/m32c/cond.md,
config/ms1/ms1.h, config/rs6000/predicates.md,
config/s390/s390.c, params.def, postreload-gcse.c,
tree-flow-inline.h, tree-ssa-operands.c, tree-vectorizer.c,
tree-vrp.c, tree.c: Fix comment typos.
* doc/invoke.texi: Fix typos.
From-SVN: r106532
2005-11-05 21:24:18 +01:00
the other loops cold that is not usually the case . So we need to artificially
2009-11-25 11:55:54 +01:00
flatten the profile .
2005-11-05 01:55:23 +01:00
basic-block.h, [...]: Fix comment typos.
* basic-block.h, config/i386/winnt.c, config/pa/pa.c,
config/s390/s390.c, dfp.c, expr.c, fold-const.c, params.def,
reload.c, struct-equiv.c, tree-ssa-ccp.c, tree-ssa-pre.c,
tree-ssa-reassoc.c, tree-ssa-structalias.c: Fix comment typos.
* doc/invoke.texi, doc/tm.texi: Fix typos.
From-SVN: r108626
2005-12-16 07:09:31 +01:00
We need to cut the maximal predicted iterations to large enough iterations
2012-11-23 09:49:43 +01:00
so the loop appears important , but safely within maximum hotness
2005-11-05 01:55:23 +01:00
range . * /
DEFPARAM ( PARAM_MAX_PREDICTED_ITERATIONS ,
"max-predicted-iterations" ,
2015-10-22 00:24:41 +02:00
"The maximum number of loop iterations we predict statically." ,
2005-11-05 01:55:23 +01:00
100 , 0 , 0 )
2013-10-03 19:13:50 +02:00
/ * This parameter controls the probability of builtin_expect . The default
value is 90 % . This empirical value is obtained through the weighted
probability of FDO counters ( with the FDO count value as the weight )
2015-10-22 00:24:41 +02:00
in some real world programs :
2013-10-03 19:13:50 +02:00
( 1 ) Google performance test benchmarks : the probability is 0.9081 .
( 2 ) Linux 3.3 kernel running Google search workload : the probability
is 0.8717 . * /
DEFPARAM ( BUILTIN_EXPECT_PROBABILITY ,
"builtin-expect-probability" ,
"Set the estimated probability in percentage for builtin expect. The default value is 90% probability." ,
90 , 0 , 100 )
2002-06-01 23:31:42 +02:00
DEFPARAM ( TRACER_DYNAMIC_COVERAGE_FEEDBACK ,
"tracer-dynamic-coverage-feedback" ,
2015-10-22 00:24:41 +02:00
"The percentage of function, weighted by execution frequency, that must be covered by trace formation. Used when profile feedback is available." ,
2004-12-01 17:46:25 +01:00
95 , 0 , 100 )
2002-06-01 23:31:42 +02:00
DEFPARAM ( TRACER_DYNAMIC_COVERAGE ,
"tracer-dynamic-coverage" ,
2015-10-22 00:24:41 +02:00
"The percentage of function, weighted by execution frequency, that must be covered by trace formation. Used when profile feedback is not available." ,
2004-12-01 17:46:25 +01:00
75 , 0 , 100 )
2002-06-01 23:31:42 +02:00
DEFPARAM ( TRACER_MAX_CODE_GROWTH ,
"tracer-max-code-growth" ,
2015-10-22 00:24:41 +02:00
"Maximal code growth caused by tail duplication (in percent)." ,
2004-12-01 17:46:25 +01:00
100 , 0 , 0 )
2002-06-01 23:31:42 +02:00
DEFPARAM ( TRACER_MIN_BRANCH_RATIO ,
"tracer-min-branch-ratio" ,
2015-10-22 00:24:41 +02:00
"Stop reverse growth if the reverse probability of best edge is less than this threshold (in percent)." ,
2004-12-01 17:46:25 +01:00
10 , 0 , 100 )
2002-06-01 23:31:42 +02:00
DEFPARAM ( TRACER_MIN_BRANCH_PROBABILITY_FEEDBACK ,
"tracer-min-branch-probability-feedback" ,
2015-10-22 00:24:41 +02:00
"Stop forward growth if the probability of best edge is less than this threshold (in percent). Used when profile feedback is available." ,
2004-12-01 17:46:25 +01:00
80 , 0 , 100 )
2002-06-01 23:31:42 +02:00
DEFPARAM ( TRACER_MIN_BRANCH_PROBABILITY ,
"tracer-min-branch-probability" ,
2015-10-22 00:24:41 +02:00
"Stop forward growth if the probability of best edge is less than this threshold (in percent). Used when profile feedback is not available." ,
2004-12-01 17:46:25 +01:00
50 , 0 , 100 )
2002-11-12 01:27:31 +01:00
2003-02-15 22:06:16 +01:00
/* The maximum number of incoming edges to consider for crossjumping. */
DEFPARAM ( PARAM_MAX_CROSSJUMP_EDGES ,
"max-crossjump-edges" ,
2015-10-22 00:24:41 +02:00
"The maximum number of incoming edges to consider for crossjumping." ,
2004-12-01 17:46:25 +01:00
100 , 0 , 0 )
2003-02-15 22:06:16 +01:00
2004-08-06 19:20:53 +02:00
/* The minimum number of matching instructions to consider for crossjumping. */
DEFPARAM ( PARAM_MIN_CROSSJUMP_INSNS ,
"min-crossjump-insns" ,
2015-10-22 00:24:41 +02:00
"The minimum number of matching instructions to consider for crossjumping." ,
2013-04-08 09:44:33 +02:00
5 , 1 , 0 )
2004-08-06 19:20:53 +02:00
2005-10-05 21:11:42 +02:00
/* The maximum number expansion factor when copying basic blocks. */
DEFPARAM ( PARAM_MAX_GROW_COPY_BB_INSNS ,
"max-grow-copy-bb-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum expansion factor when copying basic blocks." ,
2005-10-05 21:11:42 +02:00
8 , 0 , 0 )
2005-02-01 11:03:15 +01:00
/* The maximum number of insns to duplicate when unfactoring computed gotos. */
DEFPARAM ( PARAM_MAX_GOTO_DUPLICATION_INSNS ,
"max-goto-duplication-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of insns to duplicate when unfactoring computed gotos." ,
2005-02-01 11:03:15 +01:00
8 , 0 , 0 )
2003-06-04 09:51:41 +02:00
/* The maximum length of path considered in cse. */
DEFPARAM ( PARAM_MAX_CSE_PATH_LENGTH ,
"max-cse-path-length" ,
2015-10-22 00:24:41 +02:00
"The maximum length of path considered in cse." ,
2013-04-09 16:56:59 +02:00
10 , 1 , 0 )
2005-11-05 01:55:23 +01:00
DEFPARAM ( PARAM_MAX_CSE_INSNS ,
2006-04-27 16:25:49 +02:00
"max-cse-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum instructions CSE process before flushing." ,
2005-11-05 01:55:23 +01:00
1000 , 0 , 0 )
2003-06-04 09:51:41 +02:00
2004-07-10 06:57:58 +02:00
/ * The cost of expression in loop invariant motion that is considered
expensive . * /
DEFPARAM ( PARAM_LIM_EXPENSIVE ,
"lim-expensive" ,
2015-10-22 00:24:41 +02:00
"The minimum cost of an expensive expression in the loop invariant motion." ,
2004-12-01 17:46:25 +01:00
20 , 0 , 0 )
2004-07-10 06:57:58 +02:00
tree-ssa-loop-ivopts.c: New file.
* tree-ssa-loop-ivopts.c: New file.
* Makefile.in (tree-ssa-loop-ivopts.c): Add.
* cfgloop.h (target_avail_regs, target_res_regs, target_small_cost,
target_pres_cost, target_spill_cost): Declare.
* cfgloopanal.c (avail_regs, res_regs, small_cost, pres_cost,
spill_cost): Renamed to ...
(target_avail_regs, target_res_regs, target_small_cost,
target_pres_cost, target_spill_cost): ... and exported.
(init_set_costs, global_cost_for_size): Work with renamed variables.
* common.opt (flag_ivopts): New flag.
* expr.c (expand_expr_real_1): Handle SSA_NAME case. Handle
REF_ORIGINAL.
* gimplify.c (struct gimplify_ctx): Add into_ssa field.
(internal_get_tmp_var, gimplify_modify_expr, gimplify_expr): Support
generating SSA form.
(force_gimple_operand): New function.
* timevar.def (TV_TREE_LOOP_IVOPTS): New timevar.
* tree-cfg.c (stmt_bsi): New function.
* params.def (PARAM_IV_CONSIDER_ALL_CANDIDATES_BOUND,
PARAM_IV_MAX_CONSIDERED_USES): New.
* tree-flow.h (stmt_bsi, tree_ssa_iv_optimize, split_loop_exit_edge,
bsi_insert_on_edge_immediate_loop. standard_iv_increment_position,
ip_end_pos, ip_normal_pos, force_gimple_operand): Declare.
* tree-gimple.c (is_gimple_formal_tmp_var): Accept ssa names.
* tree-nested.c (build_addr): Export.
* tree-optimize.c (init_tree_optimization_passes): Add
pass_iv_optimize.
* tree-pass.h (pass_iv_optimize): Declare.
* tree-ssa-loop-im.c (for_each_index): Handle REALPART_EXPR and
IMAGPART_EXPR.
* tree-ssa-loop-manip.c (create_iv): Force the base to be acceptable
as a phi node argument.
(split_loop_exit_edge, bsi_insert_on_edge_immediate_loop,
ip_end_pos, ip_normal_pos, standard_iv_increment_position): New
functions.
* tree-ssa-loop-niter.c (zero_p, unsigned_type_for): Export.
* tree-ssa-loop.c (tree_ssa_loop_ivopts, gate_tree_ssa_loop_ivopts,
pass_iv_optimize): New pass.
* tree-ssa-operands.c (get_indirect_ref_operands): Handle REF_ORIGINAL.
* tree-ssanames.c (release_ssa_name): Allow calling with var = NULL.
* tree.c (build_int_cst_type, cst_and_fits_in_hwi): New functions.
* tree.h (REF_ORIGINAL): New macro.
(build_int_cst_type, unsigned_type_for, zero_p,
cst_and_fits_in_hwi, build_addr): Declare.
* doc/invoke.texi (-fivopts): Document.
(PARAM_IV_CONSIDER_ALL_CANDIDATES_BOUND,
PARAM_IV_MAX_CONSIDERED_USES): Document.
* doc/passes.texi: Document induction variable optimizations pass.
* gcc.dg/tree-ssa/loop-2.c: New test.
* gcc.dg/tree-ssa/loop-3.c: New test.
* gcc.dg/tree-ssa/loop-4.c: New test.
* gcc.dg/tree-ssa/loop-5.c: New test.
From-SVN: r87100
2004-09-05 11:25:37 +02:00
/ * Bound on number of candidates for induction variables below that
all candidates are considered for each use in induction variable
optimizations . * /
DEFPARAM ( PARAM_IV_CONSIDER_ALL_CANDIDATES_BOUND ,
"iv-consider-all-candidates-bound" ,
2015-10-22 00:24:41 +02:00
"Bound on number of candidates below that all candidates are considered in iv optimizations." ,
2016-03-23 16:26:43 +01:00
40 , 0 , 0 )
tree-ssa-loop-ivopts.c: New file.
* tree-ssa-loop-ivopts.c: New file.
* Makefile.in (tree-ssa-loop-ivopts.c): Add.
* cfgloop.h (target_avail_regs, target_res_regs, target_small_cost,
target_pres_cost, target_spill_cost): Declare.
* cfgloopanal.c (avail_regs, res_regs, small_cost, pres_cost,
spill_cost): Renamed to ...
(target_avail_regs, target_res_regs, target_small_cost,
target_pres_cost, target_spill_cost): ... and exported.
(init_set_costs, global_cost_for_size): Work with renamed variables.
* common.opt (flag_ivopts): New flag.
* expr.c (expand_expr_real_1): Handle SSA_NAME case. Handle
REF_ORIGINAL.
* gimplify.c (struct gimplify_ctx): Add into_ssa field.
(internal_get_tmp_var, gimplify_modify_expr, gimplify_expr): Support
generating SSA form.
(force_gimple_operand): New function.
* timevar.def (TV_TREE_LOOP_IVOPTS): New timevar.
* tree-cfg.c (stmt_bsi): New function.
* params.def (PARAM_IV_CONSIDER_ALL_CANDIDATES_BOUND,
PARAM_IV_MAX_CONSIDERED_USES): New.
* tree-flow.h (stmt_bsi, tree_ssa_iv_optimize, split_loop_exit_edge,
bsi_insert_on_edge_immediate_loop. standard_iv_increment_position,
ip_end_pos, ip_normal_pos, force_gimple_operand): Declare.
* tree-gimple.c (is_gimple_formal_tmp_var): Accept ssa names.
* tree-nested.c (build_addr): Export.
* tree-optimize.c (init_tree_optimization_passes): Add
pass_iv_optimize.
* tree-pass.h (pass_iv_optimize): Declare.
* tree-ssa-loop-im.c (for_each_index): Handle REALPART_EXPR and
IMAGPART_EXPR.
* tree-ssa-loop-manip.c (create_iv): Force the base to be acceptable
as a phi node argument.
(split_loop_exit_edge, bsi_insert_on_edge_immediate_loop,
ip_end_pos, ip_normal_pos, standard_iv_increment_position): New
functions.
* tree-ssa-loop-niter.c (zero_p, unsigned_type_for): Export.
* tree-ssa-loop.c (tree_ssa_loop_ivopts, gate_tree_ssa_loop_ivopts,
pass_iv_optimize): New pass.
* tree-ssa-operands.c (get_indirect_ref_operands): Handle REF_ORIGINAL.
* tree-ssanames.c (release_ssa_name): Allow calling with var = NULL.
* tree.c (build_int_cst_type, cst_and_fits_in_hwi): New functions.
* tree.h (REF_ORIGINAL): New macro.
(build_int_cst_type, unsigned_type_for, zero_p,
cst_and_fits_in_hwi, build_addr): Declare.
* doc/invoke.texi (-fivopts): Document.
(PARAM_IV_CONSIDER_ALL_CANDIDATES_BOUND,
PARAM_IV_MAX_CONSIDERED_USES): Document.
* doc/passes.texi: Document induction variable optimizations pass.
* gcc.dg/tree-ssa/loop-2.c: New test.
* gcc.dg/tree-ssa/loop-3.c: New test.
* gcc.dg/tree-ssa/loop-4.c: New test.
* gcc.dg/tree-ssa/loop-5.c: New test.
From-SVN: r87100
2004-09-05 11:25:37 +02:00
/ * The induction variable optimizations give up on loops that contain more
induction variable uses . * /
DEFPARAM ( PARAM_IV_MAX_CONSIDERED_USES ,
"iv-max-considered-uses" ,
2015-10-22 00:24:41 +02:00
"Bound on number of iv uses in loop optimized in iv optimizations." ,
2004-12-01 17:46:25 +01:00
250 , 0 , 0 )
tree-ssa-loop-ivopts.c: New file.
* tree-ssa-loop-ivopts.c: New file.
* Makefile.in (tree-ssa-loop-ivopts.c): Add.
* cfgloop.h (target_avail_regs, target_res_regs, target_small_cost,
target_pres_cost, target_spill_cost): Declare.
* cfgloopanal.c (avail_regs, res_regs, small_cost, pres_cost,
spill_cost): Renamed to ...
(target_avail_regs, target_res_regs, target_small_cost,
target_pres_cost, target_spill_cost): ... and exported.
(init_set_costs, global_cost_for_size): Work with renamed variables.
* common.opt (flag_ivopts): New flag.
* expr.c (expand_expr_real_1): Handle SSA_NAME case. Handle
REF_ORIGINAL.
* gimplify.c (struct gimplify_ctx): Add into_ssa field.
(internal_get_tmp_var, gimplify_modify_expr, gimplify_expr): Support
generating SSA form.
(force_gimple_operand): New function.
* timevar.def (TV_TREE_LOOP_IVOPTS): New timevar.
* tree-cfg.c (stmt_bsi): New function.
* params.def (PARAM_IV_CONSIDER_ALL_CANDIDATES_BOUND,
PARAM_IV_MAX_CONSIDERED_USES): New.
* tree-flow.h (stmt_bsi, tree_ssa_iv_optimize, split_loop_exit_edge,
bsi_insert_on_edge_immediate_loop. standard_iv_increment_position,
ip_end_pos, ip_normal_pos, force_gimple_operand): Declare.
* tree-gimple.c (is_gimple_formal_tmp_var): Accept ssa names.
* tree-nested.c (build_addr): Export.
* tree-optimize.c (init_tree_optimization_passes): Add
pass_iv_optimize.
* tree-pass.h (pass_iv_optimize): Declare.
* tree-ssa-loop-im.c (for_each_index): Handle REALPART_EXPR and
IMAGPART_EXPR.
* tree-ssa-loop-manip.c (create_iv): Force the base to be acceptable
as a phi node argument.
(split_loop_exit_edge, bsi_insert_on_edge_immediate_loop,
ip_end_pos, ip_normal_pos, standard_iv_increment_position): New
functions.
* tree-ssa-loop-niter.c (zero_p, unsigned_type_for): Export.
* tree-ssa-loop.c (tree_ssa_loop_ivopts, gate_tree_ssa_loop_ivopts,
pass_iv_optimize): New pass.
* tree-ssa-operands.c (get_indirect_ref_operands): Handle REF_ORIGINAL.
* tree-ssanames.c (release_ssa_name): Allow calling with var = NULL.
* tree.c (build_int_cst_type, cst_and_fits_in_hwi): New functions.
* tree.h (REF_ORIGINAL): New macro.
(build_int_cst_type, unsigned_type_for, zero_p,
cst_and_fits_in_hwi, build_addr): Declare.
* doc/invoke.texi (-fivopts): Document.
(PARAM_IV_CONSIDER_ALL_CANDIDATES_BOUND,
PARAM_IV_MAX_CONSIDERED_USES): Document.
* doc/passes.texi: Document induction variable optimizations pass.
* gcc.dg/tree-ssa/loop-2.c: New test.
* gcc.dg/tree-ssa/loop-3.c: New test.
* gcc.dg/tree-ssa/loop-4.c: New test.
* gcc.dg/tree-ssa/loop-5.c: New test.
From-SVN: r87100
2004-09-05 11:25:37 +02:00
2004-12-18 21:22:52 +01:00
/ * If there are at most this number of ivs in the set , try removing unnecessary
ivs from the set always . * /
DEFPARAM ( PARAM_IV_ALWAYS_PRUNE_CAND_SET_BOUND ,
"iv-always-prune-cand-set-bound" ,
2015-10-22 00:24:41 +02:00
"If number of candidates in the set is smaller, we always try to remove unused ivs during its optimization." ,
2004-12-18 21:22:52 +01:00
10 , 0 , 0 )
2016-07-12 17:27:36 +02:00
DEFPARAM ( PARAM_AVG_LOOP_NITER ,
"avg-loop-niter" ,
"Average number of iterations of a loop." ,
10 , 1 , 0 )
2017-01-13 16:42:08 +01:00
DEFPARAM ( PARAM_DSE_MAX_OBJECT_SIZE ,
"dse-max-object-size" ,
"Maximum size (in bytes) of objects tracked bytewise by dead store elimination." ,
256 , 0 , 0 )
2005-04-21 10:48:55 +02:00
DEFPARAM ( PARAM_SCEV_MAX_EXPR_SIZE ,
"scev-max-expr-size" ,
2015-10-22 00:24:41 +02:00
"Bound on size of expressions used in the scalar evolutions analyzer." ,
2011-02-02 16:52:08 +01:00
100 , 0 , 0 )
DEFPARAM ( PARAM_SCEV_MAX_EXPR_COMPLEXITY ,
"scev-max-expr-complexity" ,
2015-10-22 00:24:41 +02:00
"Bound on the complexity of the expressions in the scalar evolutions analyzer." ,
2011-02-02 16:52:08 +01:00
10 , 0 , 0 )
2005-04-21 10:48:55 +02:00
2016-05-03 11:04:46 +02:00
DEFPARAM ( PARAM_MAX_TREE_IF_CONVERSION_PHI_ARGS ,
"max-tree-if-conversion-phi-args" ,
"Maximum number of arguments in a PHI supported by TREE if-conversion "
"unless the loop is marked with simd pragma." ,
4 , 2 , 0 )
tree-vectorizer.c (new_loop_vec_info): Initialize new field.
gcc/ChangeLog
* tree-vectorizer.c (new_loop_vec_info): Initialize new
field.
(destroy_loop_vec_info): Add call to VEC_free.
* tree-vectorizer.h (may_alias_ddrs): Define.
(LOOP_VINFO_MAY_ALIAS_DDRS): Define.
* tree-vect-analyze.c (vect_analyze_data_ref_dependence):
Change reporting to dump.
(vect_is_duplicate_ddr): New.
(vect_mark_for_runtime_alias_test): New.
(vect_analyze_data_ref_dependences) Add call to
vect_mark_for_runtime_alias_test.
(vect_enhance_data_refs_alignment): Define local variable
vect_versioning_for_alias_required, don't perform
peeling for alignment if versioning for alias is
required.
(vect_enhance_data_refs_alignment): Use
PARAM_VECT_MAX_VERSION_FOR_ALIGNMENT_CHECKS instead of
PARAM_VECT_MAX_VERSION_CHECKS.
* tree-vect-transform.c
(vect_create_cond_for_alias_checks): New.
(vect_transform_loop): Add call to
vect_create_cond_for_alias_checks.
(vect_vfa_segment_size): New.
* params.def (PARAM_VECT_MAX_VERSION_FOR_ALIGNMENT_CHECKS):
Rename.
(PARAM_VECT_MAX_VERSION_FOR_ALIAS_CHECKS): Define.
* gcc/doc/invoke.texi
(vect-max-version-for-alignment-checks): Document.
(vect-max-version-for-alias-checks): Document.
(vect-max-version-checks): Remove.
gcc/testsuite/ChangeLog
* gcc.dg/vect/vect-vfa-01.c: New.
* gcc.dg/vect/vect-vfa-02.c: New.
* gcc.dg/vect/vect-vfa-03.c: New.
* gcc.dg/vect/vect-vfa-04.c: New.
* gcc.dg/vect/vect-102a.c, gcc.dg/vect/vect-51.c,
gcc.dg/vect/pr29145.c, gcc.dg/vect/vect-43.c,
gcc.dg/vect/vect-61.c, gcc.dg/vect/vect-53.c,
gcc.dg/vect/vect-45.c, gcc.dg/vect/vect-101.c,
gcc.dg/vect/vect-37.c, gcc.dg/vect/vect-79.c,
gcc.dg/vect/vect-102.c, gcc.dg/vect/vect-dv-2.c,
gcc.dg/vect/vect-57.c, gcc.dg/vect/vect-49.c,
gfortran.dg/vect/pr19049.f90: Rename to start with
prefix no-vfa-.
* gcc.dg/vect/vect.exp: Disable versioning for alias
when test starts with no-vfa-.
* gfortran.dg/vect/vect.exp: Likewise.
From-SVN: r127559
2007-08-16 16:20:39 +02:00
DEFPARAM ( PARAM_VECT_MAX_VERSION_FOR_ALIGNMENT_CHECKS ,
"vect-max-version-for-alignment-checks" ,
2015-10-22 00:24:41 +02:00
"Bound on number of runtime checks inserted by the vectorizer's loop versioning for alignment check." ,
2005-09-06 12:59:58 +02:00
6 , 0 , 0 )
tree-vectorizer.c (new_loop_vec_info): Initialize new field.
gcc/ChangeLog
* tree-vectorizer.c (new_loop_vec_info): Initialize new
field.
(destroy_loop_vec_info): Add call to VEC_free.
* tree-vectorizer.h (may_alias_ddrs): Define.
(LOOP_VINFO_MAY_ALIAS_DDRS): Define.
* tree-vect-analyze.c (vect_analyze_data_ref_dependence):
Change reporting to dump.
(vect_is_duplicate_ddr): New.
(vect_mark_for_runtime_alias_test): New.
(vect_analyze_data_ref_dependences) Add call to
vect_mark_for_runtime_alias_test.
(vect_enhance_data_refs_alignment): Define local variable
vect_versioning_for_alias_required, don't perform
peeling for alignment if versioning for alias is
required.
(vect_enhance_data_refs_alignment): Use
PARAM_VECT_MAX_VERSION_FOR_ALIGNMENT_CHECKS instead of
PARAM_VECT_MAX_VERSION_CHECKS.
* tree-vect-transform.c
(vect_create_cond_for_alias_checks): New.
(vect_transform_loop): Add call to
vect_create_cond_for_alias_checks.
(vect_vfa_segment_size): New.
* params.def (PARAM_VECT_MAX_VERSION_FOR_ALIGNMENT_CHECKS):
Rename.
(PARAM_VECT_MAX_VERSION_FOR_ALIAS_CHECKS): Define.
* gcc/doc/invoke.texi
(vect-max-version-for-alignment-checks): Document.
(vect-max-version-for-alias-checks): Document.
(vect-max-version-checks): Remove.
gcc/testsuite/ChangeLog
* gcc.dg/vect/vect-vfa-01.c: New.
* gcc.dg/vect/vect-vfa-02.c: New.
* gcc.dg/vect/vect-vfa-03.c: New.
* gcc.dg/vect/vect-vfa-04.c: New.
* gcc.dg/vect/vect-102a.c, gcc.dg/vect/vect-51.c,
gcc.dg/vect/pr29145.c, gcc.dg/vect/vect-43.c,
gcc.dg/vect/vect-61.c, gcc.dg/vect/vect-53.c,
gcc.dg/vect/vect-45.c, gcc.dg/vect/vect-101.c,
gcc.dg/vect/vect-37.c, gcc.dg/vect/vect-79.c,
gcc.dg/vect/vect-102.c, gcc.dg/vect/vect-dv-2.c,
gcc.dg/vect/vect-57.c, gcc.dg/vect/vect-49.c,
gfortran.dg/vect/pr19049.f90: Rename to start with
prefix no-vfa-.
* gcc.dg/vect/vect.exp: Disable versioning for alias
when test starts with no-vfa-.
* gfortran.dg/vect/vect.exp: Likewise.
From-SVN: r127559
2007-08-16 16:20:39 +02:00
DEFPARAM ( PARAM_VECT_MAX_VERSION_FOR_ALIAS_CHECKS ,
"vect-max-version-for-alias-checks" ,
2015-10-22 00:24:41 +02:00
"Bound on number of runtime checks inserted by the vectorizer's loop versioning for alias check." ,
tree-vectorizer.c (new_loop_vec_info): Initialize new field.
gcc/ChangeLog
* tree-vectorizer.c (new_loop_vec_info): Initialize new
field.
(destroy_loop_vec_info): Add call to VEC_free.
* tree-vectorizer.h (may_alias_ddrs): Define.
(LOOP_VINFO_MAY_ALIAS_DDRS): Define.
* tree-vect-analyze.c (vect_analyze_data_ref_dependence):
Change reporting to dump.
(vect_is_duplicate_ddr): New.
(vect_mark_for_runtime_alias_test): New.
(vect_analyze_data_ref_dependences) Add call to
vect_mark_for_runtime_alias_test.
(vect_enhance_data_refs_alignment): Define local variable
vect_versioning_for_alias_required, don't perform
peeling for alignment if versioning for alias is
required.
(vect_enhance_data_refs_alignment): Use
PARAM_VECT_MAX_VERSION_FOR_ALIGNMENT_CHECKS instead of
PARAM_VECT_MAX_VERSION_CHECKS.
* tree-vect-transform.c
(vect_create_cond_for_alias_checks): New.
(vect_transform_loop): Add call to
vect_create_cond_for_alias_checks.
(vect_vfa_segment_size): New.
* params.def (PARAM_VECT_MAX_VERSION_FOR_ALIGNMENT_CHECKS):
Rename.
(PARAM_VECT_MAX_VERSION_FOR_ALIAS_CHECKS): Define.
* gcc/doc/invoke.texi
(vect-max-version-for-alignment-checks): Document.
(vect-max-version-for-alias-checks): Document.
(vect-max-version-checks): Remove.
gcc/testsuite/ChangeLog
* gcc.dg/vect/vect-vfa-01.c: New.
* gcc.dg/vect/vect-vfa-02.c: New.
* gcc.dg/vect/vect-vfa-03.c: New.
* gcc.dg/vect/vect-vfa-04.c: New.
* gcc.dg/vect/vect-102a.c, gcc.dg/vect/vect-51.c,
gcc.dg/vect/pr29145.c, gcc.dg/vect/vect-43.c,
gcc.dg/vect/vect-61.c, gcc.dg/vect/vect-53.c,
gcc.dg/vect/vect-45.c, gcc.dg/vect/vect-101.c,
gcc.dg/vect/vect-37.c, gcc.dg/vect/vect-79.c,
gcc.dg/vect/vect-102.c, gcc.dg/vect/vect-dv-2.c,
gcc.dg/vect/vect-57.c, gcc.dg/vect/vect-49.c,
gfortran.dg/vect/pr19049.f90: Rename to start with
prefix no-vfa-.
* gcc.dg/vect/vect.exp: Disable versioning for alias
when test starts with no-vfa-.
* gfortran.dg/vect/vect.exp: Likewise.
From-SVN: r127559
2007-08-16 16:20:39 +02:00
10 , 0 , 0 )
2013-09-24 22:44:03 +02:00
DEFPARAM ( PARAM_VECT_MAX_PEELING_FOR_ALIGNMENT ,
"vect-max-peeling-for-alignment" ,
2017-03-13 10:13:14 +01:00
"Maximum number of loop peels to enhance alignment of data references in a loop." ,
2013-09-24 22:44:03 +02:00
- 1 , - 1 , 64 )
2004-01-12 10:44:16 +01:00
DEFPARAM ( PARAM_MAX_CSELIB_MEMORY_LOCATIONS ,
"max-cselib-memory-locations" ,
2015-10-22 00:24:41 +02:00
"The maximum memory locations recorded by cselib." ,
2004-12-01 17:46:25 +01:00
500 , 0 , 0 )
2004-01-12 10:44:16 +01:00
2002-11-12 01:27:31 +01:00
# ifdef ENABLE_GC_ALWAYS_COLLECT
# define GGC_MIN_EXPAND_DEFAULT 0
2002-11-15 22:27:48 +01:00
# define GGC_MIN_HEAPSIZE_DEFAULT 0
2002-11-12 01:27:31 +01:00
# else
# define GGC_MIN_EXPAND_DEFAULT 30
2002-11-15 22:27:48 +01:00
# define GGC_MIN_HEAPSIZE_DEFAULT 4096
2002-11-12 01:27:31 +01:00
# endif
DEFPARAM ( GGC_MIN_EXPAND ,
"ggc-min-expand" ,
2015-10-22 00:24:41 +02:00
"Minimum heap expansion to trigger garbage collection, as a percentage of the total size of the heap." ,
2004-12-01 17:46:25 +01:00
GGC_MIN_EXPAND_DEFAULT , 0 , 0 )
2002-11-12 01:27:31 +01:00
DEFPARAM ( GGC_MIN_HEAPSIZE ,
"ggc-min-heapsize" ,
2015-10-22 00:24:41 +02:00
"Minimum heap size before we start collecting garbage, in kilobytes." ,
2004-12-01 17:46:25 +01:00
GGC_MIN_HEAPSIZE_DEFAULT , 0 , 0 )
2002-11-15 22:27:48 +01:00
# undef GGC_MIN_EXPAND_DEFAULT
# undef GGC_MIN_HEAPSIZE_DEFAULT
2002-11-12 01:27:31 +01:00
2004-01-11 16:56:26 +01:00
DEFPARAM ( PARAM_MAX_RELOAD_SEARCH_INSNS ,
"max-reload-search-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions to search backward when looking for equivalent reload." ,
2004-12-01 17:46:25 +01:00
100 , 0 , 0 )
2004-01-11 16:56:26 +01:00
2011-10-26 15:26:50 +02:00
DEFPARAM ( PARAM_SINK_FREQUENCY_THRESHOLD ,
"sink-frequency-threshold" ,
2015-10-22 00:24:41 +02:00
"Target block's relative execution frequency (as a percentage) required to sink a statement." ,
2011-10-26 15:26:50 +02:00
75 , 0 , 100 )
2004-03-22 23:05:05 +01:00
DEFPARAM ( PARAM_MAX_SCHED_REGION_BLOCKS ,
"max-sched-region-blocks" ,
2015-10-22 00:24:41 +02:00
"The maximum number of blocks in a region to be considered for interblock scheduling." ,
2004-12-01 17:46:25 +01:00
10 , 0 , 0 )
2004-03-22 23:05:05 +01:00
DEFPARAM ( PARAM_MAX_SCHED_REGION_INSNS ,
"max-sched-region-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of insns in a region to be considered for interblock scheduling." ,
2004-12-01 17:46:25 +01:00
100 , 0 , 0 )
2004-03-22 23:05:05 +01:00
sel-sched.h, [...]: New files.
2008-08-31 Andrey Belevantsev <abel@ispras.ru>
Dmitry Melnik <dm@ispras.ru>
Dmitry Zhurikhin <zhur@ispras.ru>
Alexander Monakov <amonakov@ispras.ru>
Maxim Kuvyrkov <maxim@codesourcery.com>
* sel-sched.h, sel-sched-dump.h, sel-sched-ir.h, sel-sched.c,
sel-sched-dump.c, sel-sched-ir.c: New files.
* Makefile.in (OBJS-common): Add selective scheduling object
files.
(sel-sched.o, sel-sched-dump.o, sel-sched-ir.o): New entries.
(SEL_SCHED_IR_H, SEL_SCHED_DUMP_H): New entries.
(sched-vis.o): Add dependency on $(INSN_ATTR_H).
* cfghooks.h (get_cfg_hooks, set_cfg_hooks): New prototypes.
* cfghooks.c (get_cfg_hooks, set_cfg_hooks): New functions.
(make_forwarder_block): Update loop latch if we have redirected
the loop latch edge.
* cfgloop.c (get_loop_body_in_custom_order): New function.
* cfgloop.h (LOOPS_HAVE_FALLTHRU_PREHEADERS): New enum field.
(CP_FALLTHRU_PREHEADERS): Likewise.
(get_loop_body_in_custom_order): Declare.
* cfgloopmanip.c (has_preds_from_loop): New.
(create_preheader): Honor CP_FALLTHRU_PREHEADERS.
Assert that the preheader edge will be fall thru when it is set.
* common.opt (fsel-sched-bookkeeping, fsel-sched-pipelining,
fsel-sched-pipelining-outer-loops, fsel-sched-renaming,
fsel-sched-substitution, fselective-scheduling): New flags.
* cse.c (hash_rtx_cb): New.
(hash_rtx): Use it.
* dbgcnt.def (sel_sched_cnt, sel_sched_region_cnt,
sel_sched_insn_cnt): New counters.
* final.c (compute_alignments): Export. Free dominance info after loop_optimizer_finalize.
* genattr.c (main): Output maximal_insn_latency prototype.
* genautomata.c (output_default_latencies): New. Factor its code from ...
(output_internal_insn_latency_func): ... here.
(output_internal_maximal_insn_latency_func): New.
(output_maximal_insn_latency_func): New.
* hard-reg-set.h (UHOST_BITS_PER_WIDE_INT): Define unconditionally.
(struct hard_reg_set_iterator): New.
(hard_reg_set_iter_init, hard_reg_set_iter_set,
hard_reg_set_iter_next): New functions.
(EXECUTE_IF_SET_IN_HARD_REG_SET): New macro.
* lists.c (remove_free_INSN_LIST_node,
remove_free_EXPR_LIST_node): New functions.
* loop-init.c (loop_optimizer_init): When LOOPS_HAVE_FALLTHRU_PREHEADERS,
set CP_FALLTHRU_PREHEADERS when calling create_preheaders.
(loop_optimizer_finalize): Do not verify flow info after reload.
* recog.c (validate_replace_rtx_1): New parameter simplify.
Default it to true. Update all uses. Factor out simplifying
code to ...
(simplify_while_replacing): ... this new function.
(validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): New.
* recog.h (validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): Declare.
* rtl.c (rtx_equal_p_cb): New.
(rtx_equal_p): Use it.
* rtl.h (rtx_equal_p_cb, hash_rtx_cb): Declare.
(remove_free_INSN_LIST_NODE, remove_free_EXPR_LIST_node,
debug_bb_n_slim, debug_bb_slim, print_rtl_slim): Likewise.
* vecprim.h: Add a vector type for unsigned int.
* haifa-sched.c: Include vecprim.h and cfgloop.h.
(issue_rate, sched_verbose_param, note_list, dfa_state_size,
ready_try, cycle_issued_insns, spec_info): Make global.
(readyp): Initialize.
(dfa_lookahead): New global variable.
(old_max_uid, old_last_basic_block): Remove.
(h_i_d): Make it a vector.
(INSN_TICK, INTER_TICK, QUEUE_INDEX, INSN_COST): Make them work
through HID macro.
(after_recovery, adding_bb_to_current_region_p):
New variables to handle correct insertion of the recovery code.
(struct ready_list): Move declaration to sched-int.h.
(rgn_n_insns): Removed.
(rtx_vec_t): Move to sched-int.h.
(find_insn_reg_weight): Remove.
(find_insn_reg_weight1): Rename to find_insn_reg_weight.
(haifa_init_h_i_d, haifa_finish_h_i_d):
New functions to initialize / finalize haifa instruction data.
(extend_h_i_d, init_h_i_d): Rewrite.
(unlink_other_notes): Move logic to add_to_note_list. Handle
selective scheduler.
(ready_lastpos, ready_element, ready_sort, reemit_notes,
find_fallthru_edge): Make global, remove static prototypes.
(max_issue): Make global. Add privileged_n and state parameters. Use
them.
(extend_global, extend_all): Removed.
(init_before_recovery): Add new param. Fix the handling of the case
when we insert a recovery code before the EXIT which has a predecessor
with a fallthrough edge to it.
(create_recovery_block): Make global. Rename to
sched_create_recovery_block. Update.
(change_pattern): Rename to sched_change_pattern. Make global.
(speculate_insn): Rename to sched_speculate_insn. Make global.
Split haifa-specific functionality into ...
(haifa_change_pattern): New static function.
(sched_extend_bb): New static function.
(sched_init_bbs): New function.
(current_sched_info): Change type to struct haifa_sched_info.
(insn_cost): Adjust for selective scheduling.
(dep_cost_1): New function. Move logic from ...
(dep_cost): ... here.
(dep_cost): Use dep_cost_1.
(contributes_to_priority_p): Use sched_deps_info instead of
current_sched_info.
(priority): Adjust to work with selective scheduling. Process the
corner case when all dependencies don't contribute to priority.
(rank_for_schedule): Use ds_weak instead of dep_weak.
(advance_state): New function. Move logic from ...
(advance_one_cycle): ... here.
(add_to_note_list, concat_note_lists): New functions.
(rm_other_notes): Make static. Adjust for selective scheduling.
(remove_notes, restore_other_notes): New functions.
(move_insn): Add two arguments. Update assert. Don't call
reemit_notes.
(choose_ready): Remove lookahead variable, use dfa_lookahead.
Remove more_issue, max_points. Move the code to initialize
max_lookahead_tries to max_issue.
(schedule_block): Remove rgn_n_insns1 parameter. Don't allocate
ready. Adjust use of move_insn. Call restore_other_notes.
(luid): Remove.
(sched_init, sched_finish): Move Haifa-specific initialization/
finalization to ...
(haifa_sched_init, haifa_sched_finish): ... respectively.
New functions.
(setup_sched_dump): New function.
(haifa_init_only_bb): New static function.
(haifa_speculate_insn): New static function.
(try_ready): Use haifa_* instead of speculate_insn and
change_pattern.
(extend_ready, extend_all): Remove.
(sched_extend_ready_list, sched_finish_ready_list): New functions.
(create_check_block_twin, add_to_speculative_block): Use
haifa_insns_init instead of extend_global. Update to use new
initialization functions. Change parameter. Factor out code from
create_check_block_twin to ...
(sched_create_recovery_edges) ... this new function.
(add_block): Remove.
(sched_scan_info): New.
(extend_bb): Use sched_scan_info.
(init_bb, extend_insn, init_insn, init_insns_in_bb, sched_scan): New
static functions for walking through scheduling region.
(sched_luids): New vector variable to replace uid_to_luid.
(luids_extend_insn): New function.
(sched_max_luid): New variable.
(luids_init_insn): New function.
(sched_init_luids, sched_finish_luids): New functions.
(insn_luid): New debug function.
(sched_extend_target): New function.
(haifa_init_insn): New static function.
(sched_init_only_bb): New hook.
(sched_split_block): New hook.
(sched_split_block_1): New function.
(sched_create_empty_bb): New hook.
(sched_create_empty_bb_1): New function.
(common_sched_info, ready): New global variables.
(current_sched_info_var): Remove.
(move_block_after_check): Use common_sched_info.
(haifa_luid_for_non_insn): New static function.
(init_before_recovery): Use haifa_init_only_bb instead of
add_block.
(increase_insn_priority): New.
* modulo-sched.c: (issue_rate): Remove static declaration.
(sms_sched_info): Change type to haifa_sched_info.
(sms_sched_deps_info, sms_common_sched_info): New variables.
(setup_sched_infos): New.
(sms_schedule): Initialize them. Call haifa_sched_init/finish.
Do not call regstat_free_calls_crossed.
(sms_print_insn): Use const_rtx.
* params.def (PARAM_MAX_PIPELINE_REGION_BLOCKS,
PARAM_MAX_PIPELINE_REGION_INSNS, PARAM_SELSCHED_MAX_LOOKAHEAD,
PARAM_SELSCHED_MAX_SCHED_TIMES, PARAM_SELSCHED_INSNS_TO_RENAME,
PARAM_SCHED_MEM_TRUE_DEP_COST): New.
* sched-deps.c (sched_deps_info): New. Update all relevant uses of
current_sched_info to use it.
(enum reg_pending_barrier_mode): Move to sched-int.h.
(h_d_i_d): New variable. Initialize to NULL.
({true, output, anti, spec, forward}_dependency_cache): Initialize
to NULL.
(estimate_dep_weak): Remove static declaration.
(sched_has_condition_p): New function. Adjust users of
sched_get_condition to use it instead.
(conditions_mutex_p): Add arguments indicating which conditions are
reversed. Use them.
(sched_get_condition_with_rev): Rename from sched_get_condition. Add
argument to indicate whether returned condition is reversed. Do not
generate new rtx when condition should be reversed; indicate it by
setting new argument instead.
(add_dependence_list_and_free): Add deps parameter.
Update all users. Do not free dependence list when
deps context is readonly.
(add_insn_mem_dependence, flush_pending_lists): Adjust for readonly
contexts.
(remove_from_dependence_list, remove_from_both_dependence_lists): New.
(remove_from_deps): New. Use the above functions.
(cur_insn, can_start_lhs_rhs_p): New static variables.
(add_or_update_back_dep_1): Initialize present_dep_type.
(haifa_start_insn, haifa_finish_insn, haifa_note_reg_set,
haifa_note_reg_clobber, haifa_note_reg_use, haifa_note_mem_dep,
haifa_note_dep): New functions implementing dependence hooks for
the Haifa scheduler.
(note_reg_use, note_reg_set, note_reg_clobber, note_mem_dep,
note_dep): New functions.
(ds_to_dt, extend_deps_reg_info, maybe_extend_reg_info_p): New
functions.
(init_deps): Initialize last_reg_pending_barrier and deps->readonly.
(free_deps): Initialize deps->reg_last.
(sched_analyze_reg, sched_analyze_1, sched_analyze_2,
sched_analyze_insn): Update to use dependency hooks infrastructure
and readonly contexts.
(deps_analyze_insn): New function. Move part of logic from ...
(sched_analyze): ... here. Also move some logic to ...
(deps_start_bb): ... here. New function.
(add_forw_dep, delete_forw_dep): Guard use of INSN_DEP_COUNT with
sel_sched_p.
(sched_deps_init): New function. Move code from ...
(init_dependency_caches): ... here. Remove.
(init_deps_data_vector): New.
(sched_deps_finish): New function. Move code from ...
(free_dependency_caches): ... here. Remove.
(init_deps_global, finish_deps_global): Adjust for use with
selective scheduling.
(get_dep_weak): Move logic to ...
(get_dep_weak_1): New function.
(ds_merge): Move logic to ...
(ds_merge_1): New static function.
(ds_full_merge, ds_max_merge, ds_get_speculation_types): New functions.
(ds_get_max_dep_weak): New function.
* sched-ebb.c (sched_n_insns): Rename to sched_rgn_n_insns.
(n_insns): Rename to rgn_n_insns.
(debug_ebb_dependencies): New function.
(init_ready_list): Use it.
(begin_schedule_ready): Use sched_init_only_bb.
(ebb_print_insn): Indicate when an insn starts a new cycle.
(contributes_to_priority, compute_jump_reg_dependencies,
add_remove_insn, fix_recovery_cfg): Add ebb_ prefix to function names.
(add_block1): Remove to ebb_add_block.
(ebb_sched_deps_info, ebb_common_sched_info): New variables.
(schedule_ebb): Initialize them. Use remove_notes instead of
rm_other_notes. Use haifa_local_init/finish.
(schedule_ebbs): Use haifa_sched_init/finish.
* sched-int.h: Include vecprim.h, remove rtl.h.
(struct ready_list): Delete declaration.
(sched_verbose_param, enum sched_pass_id_t,
bb_vec_t, insn_vec_t, rtx_vec_t): New.
(struct sched_scan_info_def): New structure.
(sched_scan_info, sched_scan, sched_init_bbs,
sched_init_luids, sched_finish_luids, sched_extend_target,
haifa_init_h_i_d, haifa_finish_h_i_d): Declare.
(struct common_sched_info_def): New.
(common_sched_info, haifa_common_sched_info,
sched_emulate_haifa_p): Declare.
(sel_sched_p): New.
(sched_luids): Declare.
(INSN_LUID, LUID_BY_UID, SET_INSN_LUID): Declare.
(sched_max_luid, insn_luid): Declare.
(note_list, remove_notes, restore_other_notes, bb_note): Declare.
(sched_insns_init, sched_insns_finish, xrecalloc, reemit_notes,
print_insn, print_pattern, print_value, haifa_classify_insn,
sel_find_rgns, sel_mark_hard_insn, dfa_state_size, advance_state,
setup_sched_dump, sched_init, sched_finish,
sel_insn_is_speculation_check): Export.
(struct ready_list): Move from haifa-sched.c.
(ready_try, ready, max_issue): Export.
(ebb_compute_jump_reg_dependencies, find_fallthru_edge,
sched_init_only_bb, sched_split_block, sched_split_block_1,
sched_create_empty_bb, sched_create_empty_bb_1,
sched_create_recovery_block, sched_create_recovery_edges): Export.
(enum reg_pending_barrier_mode): Export.
(struct deps): New fields `last_reg_pending_barrier' and `readonly'.
(deps_t): New.
(struct sched_info): Rename to haifa_sched_info. Use const_rtx for
print_insn field. Move add_block and fix_recovery_cfg to
common_sched_info_def. Move compute_jump_reg_dependencies, use_cselib ...
(struct sched_deps_info_def): ... this new structure.
(sched_deps_info): Declare.
(struct spec_info_def): Remove weakness_cutoff, add
data_weakness_cutoff and control_weakness_cutoff.
(spec_info): Declare.
(struct _haifa_deps_insn_data): Split from haifa_insn_data. Add
dep_count field.
(struct haifa_insn_data): Rename to struct _haifa_insn_data.
(haifa_insn_data_def, haifa_insn_data_t): New typedefs.
(current_sched_info): Change type to struct haifa_sched_info.
(haifa_deps_insn_data_def, haifa_deps_insn_data_t): New typedefs.
(h_d_i_d): New variable.
(HDID): New accessor macro.
(h_i_d): Change type to VEC (haifa_insn_data_def, heap) *.
(HID): New accessor macro. Rewrite h_i_d accessor macros through HID
and HDID.
(IS_SPECULATION_CHECK_P): Update for selective scheduler.
(enum SCHED_FLAGS): Update for selective scheduler.
(enum SPEC_SCHED_FLAGS): New flag SEL_SCHED_SPEC_DONT_CHECK_CONTROL.
(init_dependency_caches, free_dependency_caches): Delete declarations.
(deps_analyze_insn, remove_from_deps, get_dep_weak_1,
estimate_dep_weak, ds_full_merge, ds_max_merge, ds_weak,
ds_get_speculation_types, ds_get_max_dep_weak, sched_deps_init,
sched_deps_finish, haifa_note_reg_set, haifa_note_reg_use,
haifa_note_reg_clobber, maybe_extend_reg_info_p, deps_start_bb,
ds_to_dt): Export.
(rm_other_notes): Delete declaration.
(schedule_block): Remove one argument.
(cycle_issued_insns, issue_rate, dfa_lookahead, ready_sort,
ready_element, ready_lastpos, sched_extend_ready_list,
sched_finish_ready_list, sched_change_pattern, sched_speculate_insn,
concat_note_lists): Export.
(struct region): Move from sched-rgn.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN): Export.
(ebb_head, BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB, INSN_BB): Likewise.
(current_nr_blocks, current_blocks, target_bb): Likewise.
(dep_cost_1, sched_is_disabled_for_current_region_p, sched_rgn_init,
sched_rgn_finish, rgn_setup_region, sched_rgn_compute_dependencies,
sched_rgn_local_init, extend_regions,
rgn_make_new_region_out_of_new_block, compute_priorities,
debug_rgn_dependencies, free_rgn_deps, contributes_to_priority,
extend_rgns, deps_join rgn_setup_common_sched_info,
rgn_setup_sched_infos, debug_regions, debug_region, dump_region_dot,
dump_region_dot_file, haifa_sched_init, haifa_sched_finish): Export.
(get_rgn_sched_max_insns_priority, sel_add_to_insn_priority,
increase_insn_priority): Likewise.
* sched-rgn.c: Include sel-sched.h.
(ref_counts): New static variable. Use it ...
(INSN_REF_COUNT): ... here. Rewrite and move closer to uses.
(FED_BY_SPEC_LOAD, IS_LOAD_INSN): Rewrite to use HID accessor macro.
(sched_is_disabled_for_current_region_p): Delete static declaration.
(struct region): Move to sched-int.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
ebb_head): Define and initialize.
(RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN, debug_regions, extend_regions,
BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB): Move to
sched-int.h.
(find_single_block_region): Add new argument to indicate that EBB
regions should be constructed.
(debug_live): Delete declaration.
(current_nr_blocks, current_blocks, target_bb): Remove static qualifiers.
(compute_dom_prob_ps, check_live, update_live, set_spec_fed): Delete
declaration.
(init_regions): Delete declaration.
(debug_region, bb_in_region_p, dump_region_dot_file, dump_region_dot,
rgn_estimate_number_of_insns): New.
(too_large): Use estimate_number_of_insns.
(haifa_find_rgns): New. Move the code from ...
(find_rgns): ... here. Call either sel_find_rgns or haifa_find_rgns.
(free_trg_info): New.
(compute_trg_info): Allocate candidate tables here instead of ...
(init_ready_list): ... here.
(rgn_print_insn): Use const_rtx.
(contributes_to_priority, extend_regions): Delete static declaration.
(add_remove_insn, fix_recovery_cfg): Add rgn_ to function names.
(add_block1): Rename to rgn_add_block.
(debug_rgn_dependencies): Delete static qualifier.
(new_ready): Use sched_deps_info. Simplify.
(rgn_common_sched_info, rgn_const_sched_deps_info,
rgn_const_sel_sched_deps_info, rgn_sched_deps_info, rgn_sched_info): New.
(region_sched_info): Rename to rgn_const_sched_info.
(deps_join): New, extracted from ...
(propagate_deps): ... here.
(compute_block_dependences, debug_dependencies): Update for selective
scheduling.
(free_rgn_deps, compute_priorities): New functions.
(sched_rgn_init, sched_rgn_finish, rgn_setup_region,
sched_rgn_compute_dependencies): New functions.
(schedule_region): Use them.
(sched_rgn_local_init, sched_rgn_local_free, sched_rgn_local_finish,
rgn_setup_common_sched_info, rgn_setup_sched_infos):
New functions.
(schedule_insns): Call new functions that were split out.
(rgn_make_new_region_out_of_new_block): New.
(get_rgn_sched_max_insns_priority): New.
(rest_of_handle_sched, rest_of_handle_sched2): Call selective
scheduling when appropriate.
* sched-vis.c: Include insn-attr.h.
(print_value, print_pattern): Make global.
(print_rtl_slim, debug_bb_slim, debug_bb_n_slim): New functions.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_ALLOC_SCHED_CONTEXT, TARGET_SCHED_INIT_SCHED_CONTEXT,
TARGET_SCHED_SET_SCHED_CONTEXT, TARGET_SCHED_CLEAR_SCHED_CONTEXT,
TARGET_SCHED_FREE_SCHED_CONTEXT, TARGET_SCHED_GET_INSN_CHECKED_DS,
TARGET_SCHED_GET_INSN_SPEC_DS, TARGET_SCHED_SKIP_RTX_P): New target
hooks. Initialize them to 0.
(TARGET_SCHED_GEN_CHECK): Rename to TARGET_SCHED_GEN_SPEC_CHECK.
* target.h (struct gcc_target): Add them. Rename gen_check field to
gen_spec_check.
* flags.h (sel_sched_switch_set): Declare.
* opts.c (sel_sched_switch_set): New variable.
(decode_options): Unset flag_sel_sched_pipelining_outer_loops if
pipelining is disabled from command line.
(common_handle_option): Record whether selective scheduling is
requested from command line.
* doc/invoke.texi: Document new flags and parameters.
* doc/tm.texi: Document new target hooks.
* config/ia64/ia64.c (TARGET_SCHED_GEN_SPEC_CHECK): Define to ia64_gen_check.
(dfa_state_size): Do not declare locally.
* config/ia64/ia64.opt (msched-ar-data-spec): Default to 0.
* config/rs6000/rs6000.c (rs6000_init_sched_context,
rs6000_alloc_sched_context, rs6000_set_sched_context,
rs6000_free_sched_context): New functions.
(struct _rs6000_sched_context): New.
(rs6000_sched_reorder2): Do not modify INSN_PRIORITY for selective
scheduling.
(rs6000_sched_finish): Do not run for selective scheduling.
Co-Authored-By: Alexander Monakov <amonakov@ispras.ru>
Co-Authored-By: Dmitry Melnik <dm@ispras.ru>
Co-Authored-By: Dmitry Zhurikhin <zhur@ispras.ru>
Co-Authored-By: Maxim Kuvyrkov <maxim@codesourcery.com>
From-SVN: r139854
2008-09-01 10:57:00 +02:00
DEFPARAM ( PARAM_MAX_PIPELINE_REGION_BLOCKS ,
"max-pipeline-region-blocks" ,
2015-10-22 00:24:41 +02:00
"The maximum number of blocks in a region to be considered for interblock scheduling." ,
sel-sched.h, [...]: New files.
2008-08-31 Andrey Belevantsev <abel@ispras.ru>
Dmitry Melnik <dm@ispras.ru>
Dmitry Zhurikhin <zhur@ispras.ru>
Alexander Monakov <amonakov@ispras.ru>
Maxim Kuvyrkov <maxim@codesourcery.com>
* sel-sched.h, sel-sched-dump.h, sel-sched-ir.h, sel-sched.c,
sel-sched-dump.c, sel-sched-ir.c: New files.
* Makefile.in (OBJS-common): Add selective scheduling object
files.
(sel-sched.o, sel-sched-dump.o, sel-sched-ir.o): New entries.
(SEL_SCHED_IR_H, SEL_SCHED_DUMP_H): New entries.
(sched-vis.o): Add dependency on $(INSN_ATTR_H).
* cfghooks.h (get_cfg_hooks, set_cfg_hooks): New prototypes.
* cfghooks.c (get_cfg_hooks, set_cfg_hooks): New functions.
(make_forwarder_block): Update loop latch if we have redirected
the loop latch edge.
* cfgloop.c (get_loop_body_in_custom_order): New function.
* cfgloop.h (LOOPS_HAVE_FALLTHRU_PREHEADERS): New enum field.
(CP_FALLTHRU_PREHEADERS): Likewise.
(get_loop_body_in_custom_order): Declare.
* cfgloopmanip.c (has_preds_from_loop): New.
(create_preheader): Honor CP_FALLTHRU_PREHEADERS.
Assert that the preheader edge will be fall thru when it is set.
* common.opt (fsel-sched-bookkeeping, fsel-sched-pipelining,
fsel-sched-pipelining-outer-loops, fsel-sched-renaming,
fsel-sched-substitution, fselective-scheduling): New flags.
* cse.c (hash_rtx_cb): New.
(hash_rtx): Use it.
* dbgcnt.def (sel_sched_cnt, sel_sched_region_cnt,
sel_sched_insn_cnt): New counters.
* final.c (compute_alignments): Export. Free dominance info after loop_optimizer_finalize.
* genattr.c (main): Output maximal_insn_latency prototype.
* genautomata.c (output_default_latencies): New. Factor its code from ...
(output_internal_insn_latency_func): ... here.
(output_internal_maximal_insn_latency_func): New.
(output_maximal_insn_latency_func): New.
* hard-reg-set.h (UHOST_BITS_PER_WIDE_INT): Define unconditionally.
(struct hard_reg_set_iterator): New.
(hard_reg_set_iter_init, hard_reg_set_iter_set,
hard_reg_set_iter_next): New functions.
(EXECUTE_IF_SET_IN_HARD_REG_SET): New macro.
* lists.c (remove_free_INSN_LIST_node,
remove_free_EXPR_LIST_node): New functions.
* loop-init.c (loop_optimizer_init): When LOOPS_HAVE_FALLTHRU_PREHEADERS,
set CP_FALLTHRU_PREHEADERS when calling create_preheaders.
(loop_optimizer_finalize): Do not verify flow info after reload.
* recog.c (validate_replace_rtx_1): New parameter simplify.
Default it to true. Update all uses. Factor out simplifying
code to ...
(simplify_while_replacing): ... this new function.
(validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): New.
* recog.h (validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): Declare.
* rtl.c (rtx_equal_p_cb): New.
(rtx_equal_p): Use it.
* rtl.h (rtx_equal_p_cb, hash_rtx_cb): Declare.
(remove_free_INSN_LIST_NODE, remove_free_EXPR_LIST_node,
debug_bb_n_slim, debug_bb_slim, print_rtl_slim): Likewise.
* vecprim.h: Add a vector type for unsigned int.
* haifa-sched.c: Include vecprim.h and cfgloop.h.
(issue_rate, sched_verbose_param, note_list, dfa_state_size,
ready_try, cycle_issued_insns, spec_info): Make global.
(readyp): Initialize.
(dfa_lookahead): New global variable.
(old_max_uid, old_last_basic_block): Remove.
(h_i_d): Make it a vector.
(INSN_TICK, INTER_TICK, QUEUE_INDEX, INSN_COST): Make them work
through HID macro.
(after_recovery, adding_bb_to_current_region_p):
New variables to handle correct insertion of the recovery code.
(struct ready_list): Move declaration to sched-int.h.
(rgn_n_insns): Removed.
(rtx_vec_t): Move to sched-int.h.
(find_insn_reg_weight): Remove.
(find_insn_reg_weight1): Rename to find_insn_reg_weight.
(haifa_init_h_i_d, haifa_finish_h_i_d):
New functions to initialize / finalize haifa instruction data.
(extend_h_i_d, init_h_i_d): Rewrite.
(unlink_other_notes): Move logic to add_to_note_list. Handle
selective scheduler.
(ready_lastpos, ready_element, ready_sort, reemit_notes,
find_fallthru_edge): Make global, remove static prototypes.
(max_issue): Make global. Add privileged_n and state parameters. Use
them.
(extend_global, extend_all): Removed.
(init_before_recovery): Add new param. Fix the handling of the case
when we insert a recovery code before the EXIT which has a predecessor
with a fallthrough edge to it.
(create_recovery_block): Make global. Rename to
sched_create_recovery_block. Update.
(change_pattern): Rename to sched_change_pattern. Make global.
(speculate_insn): Rename to sched_speculate_insn. Make global.
Split haifa-specific functionality into ...
(haifa_change_pattern): New static function.
(sched_extend_bb): New static function.
(sched_init_bbs): New function.
(current_sched_info): Change type to struct haifa_sched_info.
(insn_cost): Adjust for selective scheduling.
(dep_cost_1): New function. Move logic from ...
(dep_cost): ... here.
(dep_cost): Use dep_cost_1.
(contributes_to_priority_p): Use sched_deps_info instead of
current_sched_info.
(priority): Adjust to work with selective scheduling. Process the
corner case when all dependencies don't contribute to priority.
(rank_for_schedule): Use ds_weak instead of dep_weak.
(advance_state): New function. Move logic from ...
(advance_one_cycle): ... here.
(add_to_note_list, concat_note_lists): New functions.
(rm_other_notes): Make static. Adjust for selective scheduling.
(remove_notes, restore_other_notes): New functions.
(move_insn): Add two arguments. Update assert. Don't call
reemit_notes.
(choose_ready): Remove lookahead variable, use dfa_lookahead.
Remove more_issue, max_points. Move the code to initialize
max_lookahead_tries to max_issue.
(schedule_block): Remove rgn_n_insns1 parameter. Don't allocate
ready. Adjust use of move_insn. Call restore_other_notes.
(luid): Remove.
(sched_init, sched_finish): Move Haifa-specific initialization/
finalization to ...
(haifa_sched_init, haifa_sched_finish): ... respectively.
New functions.
(setup_sched_dump): New function.
(haifa_init_only_bb): New static function.
(haifa_speculate_insn): New static function.
(try_ready): Use haifa_* instead of speculate_insn and
change_pattern.
(extend_ready, extend_all): Remove.
(sched_extend_ready_list, sched_finish_ready_list): New functions.
(create_check_block_twin, add_to_speculative_block): Use
haifa_insns_init instead of extend_global. Update to use new
initialization functions. Change parameter. Factor out code from
create_check_block_twin to ...
(sched_create_recovery_edges) ... this new function.
(add_block): Remove.
(sched_scan_info): New.
(extend_bb): Use sched_scan_info.
(init_bb, extend_insn, init_insn, init_insns_in_bb, sched_scan): New
static functions for walking through scheduling region.
(sched_luids): New vector variable to replace uid_to_luid.
(luids_extend_insn): New function.
(sched_max_luid): New variable.
(luids_init_insn): New function.
(sched_init_luids, sched_finish_luids): New functions.
(insn_luid): New debug function.
(sched_extend_target): New function.
(haifa_init_insn): New static function.
(sched_init_only_bb): New hook.
(sched_split_block): New hook.
(sched_split_block_1): New function.
(sched_create_empty_bb): New hook.
(sched_create_empty_bb_1): New function.
(common_sched_info, ready): New global variables.
(current_sched_info_var): Remove.
(move_block_after_check): Use common_sched_info.
(haifa_luid_for_non_insn): New static function.
(init_before_recovery): Use haifa_init_only_bb instead of
add_block.
(increase_insn_priority): New.
* modulo-sched.c: (issue_rate): Remove static declaration.
(sms_sched_info): Change type to haifa_sched_info.
(sms_sched_deps_info, sms_common_sched_info): New variables.
(setup_sched_infos): New.
(sms_schedule): Initialize them. Call haifa_sched_init/finish.
Do not call regstat_free_calls_crossed.
(sms_print_insn): Use const_rtx.
* params.def (PARAM_MAX_PIPELINE_REGION_BLOCKS,
PARAM_MAX_PIPELINE_REGION_INSNS, PARAM_SELSCHED_MAX_LOOKAHEAD,
PARAM_SELSCHED_MAX_SCHED_TIMES, PARAM_SELSCHED_INSNS_TO_RENAME,
PARAM_SCHED_MEM_TRUE_DEP_COST): New.
* sched-deps.c (sched_deps_info): New. Update all relevant uses of
current_sched_info to use it.
(enum reg_pending_barrier_mode): Move to sched-int.h.
(h_d_i_d): New variable. Initialize to NULL.
({true, output, anti, spec, forward}_dependency_cache): Initialize
to NULL.
(estimate_dep_weak): Remove static declaration.
(sched_has_condition_p): New function. Adjust users of
sched_get_condition to use it instead.
(conditions_mutex_p): Add arguments indicating which conditions are
reversed. Use them.
(sched_get_condition_with_rev): Rename from sched_get_condition. Add
argument to indicate whether returned condition is reversed. Do not
generate new rtx when condition should be reversed; indicate it by
setting new argument instead.
(add_dependence_list_and_free): Add deps parameter.
Update all users. Do not free dependence list when
deps context is readonly.
(add_insn_mem_dependence, flush_pending_lists): Adjust for readonly
contexts.
(remove_from_dependence_list, remove_from_both_dependence_lists): New.
(remove_from_deps): New. Use the above functions.
(cur_insn, can_start_lhs_rhs_p): New static variables.
(add_or_update_back_dep_1): Initialize present_dep_type.
(haifa_start_insn, haifa_finish_insn, haifa_note_reg_set,
haifa_note_reg_clobber, haifa_note_reg_use, haifa_note_mem_dep,
haifa_note_dep): New functions implementing dependence hooks for
the Haifa scheduler.
(note_reg_use, note_reg_set, note_reg_clobber, note_mem_dep,
note_dep): New functions.
(ds_to_dt, extend_deps_reg_info, maybe_extend_reg_info_p): New
functions.
(init_deps): Initialize last_reg_pending_barrier and deps->readonly.
(free_deps): Initialize deps->reg_last.
(sched_analyze_reg, sched_analyze_1, sched_analyze_2,
sched_analyze_insn): Update to use dependency hooks infrastructure
and readonly contexts.
(deps_analyze_insn): New function. Move part of logic from ...
(sched_analyze): ... here. Also move some logic to ...
(deps_start_bb): ... here. New function.
(add_forw_dep, delete_forw_dep): Guard use of INSN_DEP_COUNT with
sel_sched_p.
(sched_deps_init): New function. Move code from ...
(init_dependency_caches): ... here. Remove.
(init_deps_data_vector): New.
(sched_deps_finish): New function. Move code from ...
(free_dependency_caches): ... here. Remove.
(init_deps_global, finish_deps_global): Adjust for use with
selective scheduling.
(get_dep_weak): Move logic to ...
(get_dep_weak_1): New function.
(ds_merge): Move logic to ...
(ds_merge_1): New static function.
(ds_full_merge, ds_max_merge, ds_get_speculation_types): New functions.
(ds_get_max_dep_weak): New function.
* sched-ebb.c (sched_n_insns): Rename to sched_rgn_n_insns.
(n_insns): Rename to rgn_n_insns.
(debug_ebb_dependencies): New function.
(init_ready_list): Use it.
(begin_schedule_ready): Use sched_init_only_bb.
(ebb_print_insn): Indicate when an insn starts a new cycle.
(contributes_to_priority, compute_jump_reg_dependencies,
add_remove_insn, fix_recovery_cfg): Add ebb_ prefix to function names.
(add_block1): Remove to ebb_add_block.
(ebb_sched_deps_info, ebb_common_sched_info): New variables.
(schedule_ebb): Initialize them. Use remove_notes instead of
rm_other_notes. Use haifa_local_init/finish.
(schedule_ebbs): Use haifa_sched_init/finish.
* sched-int.h: Include vecprim.h, remove rtl.h.
(struct ready_list): Delete declaration.
(sched_verbose_param, enum sched_pass_id_t,
bb_vec_t, insn_vec_t, rtx_vec_t): New.
(struct sched_scan_info_def): New structure.
(sched_scan_info, sched_scan, sched_init_bbs,
sched_init_luids, sched_finish_luids, sched_extend_target,
haifa_init_h_i_d, haifa_finish_h_i_d): Declare.
(struct common_sched_info_def): New.
(common_sched_info, haifa_common_sched_info,
sched_emulate_haifa_p): Declare.
(sel_sched_p): New.
(sched_luids): Declare.
(INSN_LUID, LUID_BY_UID, SET_INSN_LUID): Declare.
(sched_max_luid, insn_luid): Declare.
(note_list, remove_notes, restore_other_notes, bb_note): Declare.
(sched_insns_init, sched_insns_finish, xrecalloc, reemit_notes,
print_insn, print_pattern, print_value, haifa_classify_insn,
sel_find_rgns, sel_mark_hard_insn, dfa_state_size, advance_state,
setup_sched_dump, sched_init, sched_finish,
sel_insn_is_speculation_check): Export.
(struct ready_list): Move from haifa-sched.c.
(ready_try, ready, max_issue): Export.
(ebb_compute_jump_reg_dependencies, find_fallthru_edge,
sched_init_only_bb, sched_split_block, sched_split_block_1,
sched_create_empty_bb, sched_create_empty_bb_1,
sched_create_recovery_block, sched_create_recovery_edges): Export.
(enum reg_pending_barrier_mode): Export.
(struct deps): New fields `last_reg_pending_barrier' and `readonly'.
(deps_t): New.
(struct sched_info): Rename to haifa_sched_info. Use const_rtx for
print_insn field. Move add_block and fix_recovery_cfg to
common_sched_info_def. Move compute_jump_reg_dependencies, use_cselib ...
(struct sched_deps_info_def): ... this new structure.
(sched_deps_info): Declare.
(struct spec_info_def): Remove weakness_cutoff, add
data_weakness_cutoff and control_weakness_cutoff.
(spec_info): Declare.
(struct _haifa_deps_insn_data): Split from haifa_insn_data. Add
dep_count field.
(struct haifa_insn_data): Rename to struct _haifa_insn_data.
(haifa_insn_data_def, haifa_insn_data_t): New typedefs.
(current_sched_info): Change type to struct haifa_sched_info.
(haifa_deps_insn_data_def, haifa_deps_insn_data_t): New typedefs.
(h_d_i_d): New variable.
(HDID): New accessor macro.
(h_i_d): Change type to VEC (haifa_insn_data_def, heap) *.
(HID): New accessor macro. Rewrite h_i_d accessor macros through HID
and HDID.
(IS_SPECULATION_CHECK_P): Update for selective scheduler.
(enum SCHED_FLAGS): Update for selective scheduler.
(enum SPEC_SCHED_FLAGS): New flag SEL_SCHED_SPEC_DONT_CHECK_CONTROL.
(init_dependency_caches, free_dependency_caches): Delete declarations.
(deps_analyze_insn, remove_from_deps, get_dep_weak_1,
estimate_dep_weak, ds_full_merge, ds_max_merge, ds_weak,
ds_get_speculation_types, ds_get_max_dep_weak, sched_deps_init,
sched_deps_finish, haifa_note_reg_set, haifa_note_reg_use,
haifa_note_reg_clobber, maybe_extend_reg_info_p, deps_start_bb,
ds_to_dt): Export.
(rm_other_notes): Delete declaration.
(schedule_block): Remove one argument.
(cycle_issued_insns, issue_rate, dfa_lookahead, ready_sort,
ready_element, ready_lastpos, sched_extend_ready_list,
sched_finish_ready_list, sched_change_pattern, sched_speculate_insn,
concat_note_lists): Export.
(struct region): Move from sched-rgn.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN): Export.
(ebb_head, BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB, INSN_BB): Likewise.
(current_nr_blocks, current_blocks, target_bb): Likewise.
(dep_cost_1, sched_is_disabled_for_current_region_p, sched_rgn_init,
sched_rgn_finish, rgn_setup_region, sched_rgn_compute_dependencies,
sched_rgn_local_init, extend_regions,
rgn_make_new_region_out_of_new_block, compute_priorities,
debug_rgn_dependencies, free_rgn_deps, contributes_to_priority,
extend_rgns, deps_join rgn_setup_common_sched_info,
rgn_setup_sched_infos, debug_regions, debug_region, dump_region_dot,
dump_region_dot_file, haifa_sched_init, haifa_sched_finish): Export.
(get_rgn_sched_max_insns_priority, sel_add_to_insn_priority,
increase_insn_priority): Likewise.
* sched-rgn.c: Include sel-sched.h.
(ref_counts): New static variable. Use it ...
(INSN_REF_COUNT): ... here. Rewrite and move closer to uses.
(FED_BY_SPEC_LOAD, IS_LOAD_INSN): Rewrite to use HID accessor macro.
(sched_is_disabled_for_current_region_p): Delete static declaration.
(struct region): Move to sched-int.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
ebb_head): Define and initialize.
(RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN, debug_regions, extend_regions,
BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB): Move to
sched-int.h.
(find_single_block_region): Add new argument to indicate that EBB
regions should be constructed.
(debug_live): Delete declaration.
(current_nr_blocks, current_blocks, target_bb): Remove static qualifiers.
(compute_dom_prob_ps, check_live, update_live, set_spec_fed): Delete
declaration.
(init_regions): Delete declaration.
(debug_region, bb_in_region_p, dump_region_dot_file, dump_region_dot,
rgn_estimate_number_of_insns): New.
(too_large): Use estimate_number_of_insns.
(haifa_find_rgns): New. Move the code from ...
(find_rgns): ... here. Call either sel_find_rgns or haifa_find_rgns.
(free_trg_info): New.
(compute_trg_info): Allocate candidate tables here instead of ...
(init_ready_list): ... here.
(rgn_print_insn): Use const_rtx.
(contributes_to_priority, extend_regions): Delete static declaration.
(add_remove_insn, fix_recovery_cfg): Add rgn_ to function names.
(add_block1): Rename to rgn_add_block.
(debug_rgn_dependencies): Delete static qualifier.
(new_ready): Use sched_deps_info. Simplify.
(rgn_common_sched_info, rgn_const_sched_deps_info,
rgn_const_sel_sched_deps_info, rgn_sched_deps_info, rgn_sched_info): New.
(region_sched_info): Rename to rgn_const_sched_info.
(deps_join): New, extracted from ...
(propagate_deps): ... here.
(compute_block_dependences, debug_dependencies): Update for selective
scheduling.
(free_rgn_deps, compute_priorities): New functions.
(sched_rgn_init, sched_rgn_finish, rgn_setup_region,
sched_rgn_compute_dependencies): New functions.
(schedule_region): Use them.
(sched_rgn_local_init, sched_rgn_local_free, sched_rgn_local_finish,
rgn_setup_common_sched_info, rgn_setup_sched_infos):
New functions.
(schedule_insns): Call new functions that were split out.
(rgn_make_new_region_out_of_new_block): New.
(get_rgn_sched_max_insns_priority): New.
(rest_of_handle_sched, rest_of_handle_sched2): Call selective
scheduling when appropriate.
* sched-vis.c: Include insn-attr.h.
(print_value, print_pattern): Make global.
(print_rtl_slim, debug_bb_slim, debug_bb_n_slim): New functions.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_ALLOC_SCHED_CONTEXT, TARGET_SCHED_INIT_SCHED_CONTEXT,
TARGET_SCHED_SET_SCHED_CONTEXT, TARGET_SCHED_CLEAR_SCHED_CONTEXT,
TARGET_SCHED_FREE_SCHED_CONTEXT, TARGET_SCHED_GET_INSN_CHECKED_DS,
TARGET_SCHED_GET_INSN_SPEC_DS, TARGET_SCHED_SKIP_RTX_P): New target
hooks. Initialize them to 0.
(TARGET_SCHED_GEN_CHECK): Rename to TARGET_SCHED_GEN_SPEC_CHECK.
* target.h (struct gcc_target): Add them. Rename gen_check field to
gen_spec_check.
* flags.h (sel_sched_switch_set): Declare.
* opts.c (sel_sched_switch_set): New variable.
(decode_options): Unset flag_sel_sched_pipelining_outer_loops if
pipelining is disabled from command line.
(common_handle_option): Record whether selective scheduling is
requested from command line.
* doc/invoke.texi: Document new flags and parameters.
* doc/tm.texi: Document new target hooks.
* config/ia64/ia64.c (TARGET_SCHED_GEN_SPEC_CHECK): Define to ia64_gen_check.
(dfa_state_size): Do not declare locally.
* config/ia64/ia64.opt (msched-ar-data-spec): Default to 0.
* config/rs6000/rs6000.c (rs6000_init_sched_context,
rs6000_alloc_sched_context, rs6000_set_sched_context,
rs6000_free_sched_context): New functions.
(struct _rs6000_sched_context): New.
(rs6000_sched_reorder2): Do not modify INSN_PRIORITY for selective
scheduling.
(rs6000_sched_finish): Do not run for selective scheduling.
Co-Authored-By: Alexander Monakov <amonakov@ispras.ru>
Co-Authored-By: Dmitry Melnik <dm@ispras.ru>
Co-Authored-By: Dmitry Zhurikhin <zhur@ispras.ru>
Co-Authored-By: Maxim Kuvyrkov <maxim@codesourcery.com>
From-SVN: r139854
2008-09-01 10:57:00 +02:00
15 , 0 , 0 )
DEFPARAM ( PARAM_MAX_PIPELINE_REGION_INSNS ,
"max-pipeline-region-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of insns in a region to be considered for interblock scheduling." ,
sel-sched.h, [...]: New files.
2008-08-31 Andrey Belevantsev <abel@ispras.ru>
Dmitry Melnik <dm@ispras.ru>
Dmitry Zhurikhin <zhur@ispras.ru>
Alexander Monakov <amonakov@ispras.ru>
Maxim Kuvyrkov <maxim@codesourcery.com>
* sel-sched.h, sel-sched-dump.h, sel-sched-ir.h, sel-sched.c,
sel-sched-dump.c, sel-sched-ir.c: New files.
* Makefile.in (OBJS-common): Add selective scheduling object
files.
(sel-sched.o, sel-sched-dump.o, sel-sched-ir.o): New entries.
(SEL_SCHED_IR_H, SEL_SCHED_DUMP_H): New entries.
(sched-vis.o): Add dependency on $(INSN_ATTR_H).
* cfghooks.h (get_cfg_hooks, set_cfg_hooks): New prototypes.
* cfghooks.c (get_cfg_hooks, set_cfg_hooks): New functions.
(make_forwarder_block): Update loop latch if we have redirected
the loop latch edge.
* cfgloop.c (get_loop_body_in_custom_order): New function.
* cfgloop.h (LOOPS_HAVE_FALLTHRU_PREHEADERS): New enum field.
(CP_FALLTHRU_PREHEADERS): Likewise.
(get_loop_body_in_custom_order): Declare.
* cfgloopmanip.c (has_preds_from_loop): New.
(create_preheader): Honor CP_FALLTHRU_PREHEADERS.
Assert that the preheader edge will be fall thru when it is set.
* common.opt (fsel-sched-bookkeeping, fsel-sched-pipelining,
fsel-sched-pipelining-outer-loops, fsel-sched-renaming,
fsel-sched-substitution, fselective-scheduling): New flags.
* cse.c (hash_rtx_cb): New.
(hash_rtx): Use it.
* dbgcnt.def (sel_sched_cnt, sel_sched_region_cnt,
sel_sched_insn_cnt): New counters.
* final.c (compute_alignments): Export. Free dominance info after loop_optimizer_finalize.
* genattr.c (main): Output maximal_insn_latency prototype.
* genautomata.c (output_default_latencies): New. Factor its code from ...
(output_internal_insn_latency_func): ... here.
(output_internal_maximal_insn_latency_func): New.
(output_maximal_insn_latency_func): New.
* hard-reg-set.h (UHOST_BITS_PER_WIDE_INT): Define unconditionally.
(struct hard_reg_set_iterator): New.
(hard_reg_set_iter_init, hard_reg_set_iter_set,
hard_reg_set_iter_next): New functions.
(EXECUTE_IF_SET_IN_HARD_REG_SET): New macro.
* lists.c (remove_free_INSN_LIST_node,
remove_free_EXPR_LIST_node): New functions.
* loop-init.c (loop_optimizer_init): When LOOPS_HAVE_FALLTHRU_PREHEADERS,
set CP_FALLTHRU_PREHEADERS when calling create_preheaders.
(loop_optimizer_finalize): Do not verify flow info after reload.
* recog.c (validate_replace_rtx_1): New parameter simplify.
Default it to true. Update all uses. Factor out simplifying
code to ...
(simplify_while_replacing): ... this new function.
(validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): New.
* recog.h (validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): Declare.
* rtl.c (rtx_equal_p_cb): New.
(rtx_equal_p): Use it.
* rtl.h (rtx_equal_p_cb, hash_rtx_cb): Declare.
(remove_free_INSN_LIST_NODE, remove_free_EXPR_LIST_node,
debug_bb_n_slim, debug_bb_slim, print_rtl_slim): Likewise.
* vecprim.h: Add a vector type for unsigned int.
* haifa-sched.c: Include vecprim.h and cfgloop.h.
(issue_rate, sched_verbose_param, note_list, dfa_state_size,
ready_try, cycle_issued_insns, spec_info): Make global.
(readyp): Initialize.
(dfa_lookahead): New global variable.
(old_max_uid, old_last_basic_block): Remove.
(h_i_d): Make it a vector.
(INSN_TICK, INTER_TICK, QUEUE_INDEX, INSN_COST): Make them work
through HID macro.
(after_recovery, adding_bb_to_current_region_p):
New variables to handle correct insertion of the recovery code.
(struct ready_list): Move declaration to sched-int.h.
(rgn_n_insns): Removed.
(rtx_vec_t): Move to sched-int.h.
(find_insn_reg_weight): Remove.
(find_insn_reg_weight1): Rename to find_insn_reg_weight.
(haifa_init_h_i_d, haifa_finish_h_i_d):
New functions to initialize / finalize haifa instruction data.
(extend_h_i_d, init_h_i_d): Rewrite.
(unlink_other_notes): Move logic to add_to_note_list. Handle
selective scheduler.
(ready_lastpos, ready_element, ready_sort, reemit_notes,
find_fallthru_edge): Make global, remove static prototypes.
(max_issue): Make global. Add privileged_n and state parameters. Use
them.
(extend_global, extend_all): Removed.
(init_before_recovery): Add new param. Fix the handling of the case
when we insert a recovery code before the EXIT which has a predecessor
with a fallthrough edge to it.
(create_recovery_block): Make global. Rename to
sched_create_recovery_block. Update.
(change_pattern): Rename to sched_change_pattern. Make global.
(speculate_insn): Rename to sched_speculate_insn. Make global.
Split haifa-specific functionality into ...
(haifa_change_pattern): New static function.
(sched_extend_bb): New static function.
(sched_init_bbs): New function.
(current_sched_info): Change type to struct haifa_sched_info.
(insn_cost): Adjust for selective scheduling.
(dep_cost_1): New function. Move logic from ...
(dep_cost): ... here.
(dep_cost): Use dep_cost_1.
(contributes_to_priority_p): Use sched_deps_info instead of
current_sched_info.
(priority): Adjust to work with selective scheduling. Process the
corner case when all dependencies don't contribute to priority.
(rank_for_schedule): Use ds_weak instead of dep_weak.
(advance_state): New function. Move logic from ...
(advance_one_cycle): ... here.
(add_to_note_list, concat_note_lists): New functions.
(rm_other_notes): Make static. Adjust for selective scheduling.
(remove_notes, restore_other_notes): New functions.
(move_insn): Add two arguments. Update assert. Don't call
reemit_notes.
(choose_ready): Remove lookahead variable, use dfa_lookahead.
Remove more_issue, max_points. Move the code to initialize
max_lookahead_tries to max_issue.
(schedule_block): Remove rgn_n_insns1 parameter. Don't allocate
ready. Adjust use of move_insn. Call restore_other_notes.
(luid): Remove.
(sched_init, sched_finish): Move Haifa-specific initialization/
finalization to ...
(haifa_sched_init, haifa_sched_finish): ... respectively.
New functions.
(setup_sched_dump): New function.
(haifa_init_only_bb): New static function.
(haifa_speculate_insn): New static function.
(try_ready): Use haifa_* instead of speculate_insn and
change_pattern.
(extend_ready, extend_all): Remove.
(sched_extend_ready_list, sched_finish_ready_list): New functions.
(create_check_block_twin, add_to_speculative_block): Use
haifa_insns_init instead of extend_global. Update to use new
initialization functions. Change parameter. Factor out code from
create_check_block_twin to ...
(sched_create_recovery_edges) ... this new function.
(add_block): Remove.
(sched_scan_info): New.
(extend_bb): Use sched_scan_info.
(init_bb, extend_insn, init_insn, init_insns_in_bb, sched_scan): New
static functions for walking through scheduling region.
(sched_luids): New vector variable to replace uid_to_luid.
(luids_extend_insn): New function.
(sched_max_luid): New variable.
(luids_init_insn): New function.
(sched_init_luids, sched_finish_luids): New functions.
(insn_luid): New debug function.
(sched_extend_target): New function.
(haifa_init_insn): New static function.
(sched_init_only_bb): New hook.
(sched_split_block): New hook.
(sched_split_block_1): New function.
(sched_create_empty_bb): New hook.
(sched_create_empty_bb_1): New function.
(common_sched_info, ready): New global variables.
(current_sched_info_var): Remove.
(move_block_after_check): Use common_sched_info.
(haifa_luid_for_non_insn): New static function.
(init_before_recovery): Use haifa_init_only_bb instead of
add_block.
(increase_insn_priority): New.
* modulo-sched.c: (issue_rate): Remove static declaration.
(sms_sched_info): Change type to haifa_sched_info.
(sms_sched_deps_info, sms_common_sched_info): New variables.
(setup_sched_infos): New.
(sms_schedule): Initialize them. Call haifa_sched_init/finish.
Do not call regstat_free_calls_crossed.
(sms_print_insn): Use const_rtx.
* params.def (PARAM_MAX_PIPELINE_REGION_BLOCKS,
PARAM_MAX_PIPELINE_REGION_INSNS, PARAM_SELSCHED_MAX_LOOKAHEAD,
PARAM_SELSCHED_MAX_SCHED_TIMES, PARAM_SELSCHED_INSNS_TO_RENAME,
PARAM_SCHED_MEM_TRUE_DEP_COST): New.
* sched-deps.c (sched_deps_info): New. Update all relevant uses of
current_sched_info to use it.
(enum reg_pending_barrier_mode): Move to sched-int.h.
(h_d_i_d): New variable. Initialize to NULL.
({true, output, anti, spec, forward}_dependency_cache): Initialize
to NULL.
(estimate_dep_weak): Remove static declaration.
(sched_has_condition_p): New function. Adjust users of
sched_get_condition to use it instead.
(conditions_mutex_p): Add arguments indicating which conditions are
reversed. Use them.
(sched_get_condition_with_rev): Rename from sched_get_condition. Add
argument to indicate whether returned condition is reversed. Do not
generate new rtx when condition should be reversed; indicate it by
setting new argument instead.
(add_dependence_list_and_free): Add deps parameter.
Update all users. Do not free dependence list when
deps context is readonly.
(add_insn_mem_dependence, flush_pending_lists): Adjust for readonly
contexts.
(remove_from_dependence_list, remove_from_both_dependence_lists): New.
(remove_from_deps): New. Use the above functions.
(cur_insn, can_start_lhs_rhs_p): New static variables.
(add_or_update_back_dep_1): Initialize present_dep_type.
(haifa_start_insn, haifa_finish_insn, haifa_note_reg_set,
haifa_note_reg_clobber, haifa_note_reg_use, haifa_note_mem_dep,
haifa_note_dep): New functions implementing dependence hooks for
the Haifa scheduler.
(note_reg_use, note_reg_set, note_reg_clobber, note_mem_dep,
note_dep): New functions.
(ds_to_dt, extend_deps_reg_info, maybe_extend_reg_info_p): New
functions.
(init_deps): Initialize last_reg_pending_barrier and deps->readonly.
(free_deps): Initialize deps->reg_last.
(sched_analyze_reg, sched_analyze_1, sched_analyze_2,
sched_analyze_insn): Update to use dependency hooks infrastructure
and readonly contexts.
(deps_analyze_insn): New function. Move part of logic from ...
(sched_analyze): ... here. Also move some logic to ...
(deps_start_bb): ... here. New function.
(add_forw_dep, delete_forw_dep): Guard use of INSN_DEP_COUNT with
sel_sched_p.
(sched_deps_init): New function. Move code from ...
(init_dependency_caches): ... here. Remove.
(init_deps_data_vector): New.
(sched_deps_finish): New function. Move code from ...
(free_dependency_caches): ... here. Remove.
(init_deps_global, finish_deps_global): Adjust for use with
selective scheduling.
(get_dep_weak): Move logic to ...
(get_dep_weak_1): New function.
(ds_merge): Move logic to ...
(ds_merge_1): New static function.
(ds_full_merge, ds_max_merge, ds_get_speculation_types): New functions.
(ds_get_max_dep_weak): New function.
* sched-ebb.c (sched_n_insns): Rename to sched_rgn_n_insns.
(n_insns): Rename to rgn_n_insns.
(debug_ebb_dependencies): New function.
(init_ready_list): Use it.
(begin_schedule_ready): Use sched_init_only_bb.
(ebb_print_insn): Indicate when an insn starts a new cycle.
(contributes_to_priority, compute_jump_reg_dependencies,
add_remove_insn, fix_recovery_cfg): Add ebb_ prefix to function names.
(add_block1): Remove to ebb_add_block.
(ebb_sched_deps_info, ebb_common_sched_info): New variables.
(schedule_ebb): Initialize them. Use remove_notes instead of
rm_other_notes. Use haifa_local_init/finish.
(schedule_ebbs): Use haifa_sched_init/finish.
* sched-int.h: Include vecprim.h, remove rtl.h.
(struct ready_list): Delete declaration.
(sched_verbose_param, enum sched_pass_id_t,
bb_vec_t, insn_vec_t, rtx_vec_t): New.
(struct sched_scan_info_def): New structure.
(sched_scan_info, sched_scan, sched_init_bbs,
sched_init_luids, sched_finish_luids, sched_extend_target,
haifa_init_h_i_d, haifa_finish_h_i_d): Declare.
(struct common_sched_info_def): New.
(common_sched_info, haifa_common_sched_info,
sched_emulate_haifa_p): Declare.
(sel_sched_p): New.
(sched_luids): Declare.
(INSN_LUID, LUID_BY_UID, SET_INSN_LUID): Declare.
(sched_max_luid, insn_luid): Declare.
(note_list, remove_notes, restore_other_notes, bb_note): Declare.
(sched_insns_init, sched_insns_finish, xrecalloc, reemit_notes,
print_insn, print_pattern, print_value, haifa_classify_insn,
sel_find_rgns, sel_mark_hard_insn, dfa_state_size, advance_state,
setup_sched_dump, sched_init, sched_finish,
sel_insn_is_speculation_check): Export.
(struct ready_list): Move from haifa-sched.c.
(ready_try, ready, max_issue): Export.
(ebb_compute_jump_reg_dependencies, find_fallthru_edge,
sched_init_only_bb, sched_split_block, sched_split_block_1,
sched_create_empty_bb, sched_create_empty_bb_1,
sched_create_recovery_block, sched_create_recovery_edges): Export.
(enum reg_pending_barrier_mode): Export.
(struct deps): New fields `last_reg_pending_barrier' and `readonly'.
(deps_t): New.
(struct sched_info): Rename to haifa_sched_info. Use const_rtx for
print_insn field. Move add_block and fix_recovery_cfg to
common_sched_info_def. Move compute_jump_reg_dependencies, use_cselib ...
(struct sched_deps_info_def): ... this new structure.
(sched_deps_info): Declare.
(struct spec_info_def): Remove weakness_cutoff, add
data_weakness_cutoff and control_weakness_cutoff.
(spec_info): Declare.
(struct _haifa_deps_insn_data): Split from haifa_insn_data. Add
dep_count field.
(struct haifa_insn_data): Rename to struct _haifa_insn_data.
(haifa_insn_data_def, haifa_insn_data_t): New typedefs.
(current_sched_info): Change type to struct haifa_sched_info.
(haifa_deps_insn_data_def, haifa_deps_insn_data_t): New typedefs.
(h_d_i_d): New variable.
(HDID): New accessor macro.
(h_i_d): Change type to VEC (haifa_insn_data_def, heap) *.
(HID): New accessor macro. Rewrite h_i_d accessor macros through HID
and HDID.
(IS_SPECULATION_CHECK_P): Update for selective scheduler.
(enum SCHED_FLAGS): Update for selective scheduler.
(enum SPEC_SCHED_FLAGS): New flag SEL_SCHED_SPEC_DONT_CHECK_CONTROL.
(init_dependency_caches, free_dependency_caches): Delete declarations.
(deps_analyze_insn, remove_from_deps, get_dep_weak_1,
estimate_dep_weak, ds_full_merge, ds_max_merge, ds_weak,
ds_get_speculation_types, ds_get_max_dep_weak, sched_deps_init,
sched_deps_finish, haifa_note_reg_set, haifa_note_reg_use,
haifa_note_reg_clobber, maybe_extend_reg_info_p, deps_start_bb,
ds_to_dt): Export.
(rm_other_notes): Delete declaration.
(schedule_block): Remove one argument.
(cycle_issued_insns, issue_rate, dfa_lookahead, ready_sort,
ready_element, ready_lastpos, sched_extend_ready_list,
sched_finish_ready_list, sched_change_pattern, sched_speculate_insn,
concat_note_lists): Export.
(struct region): Move from sched-rgn.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN): Export.
(ebb_head, BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB, INSN_BB): Likewise.
(current_nr_blocks, current_blocks, target_bb): Likewise.
(dep_cost_1, sched_is_disabled_for_current_region_p, sched_rgn_init,
sched_rgn_finish, rgn_setup_region, sched_rgn_compute_dependencies,
sched_rgn_local_init, extend_regions,
rgn_make_new_region_out_of_new_block, compute_priorities,
debug_rgn_dependencies, free_rgn_deps, contributes_to_priority,
extend_rgns, deps_join rgn_setup_common_sched_info,
rgn_setup_sched_infos, debug_regions, debug_region, dump_region_dot,
dump_region_dot_file, haifa_sched_init, haifa_sched_finish): Export.
(get_rgn_sched_max_insns_priority, sel_add_to_insn_priority,
increase_insn_priority): Likewise.
* sched-rgn.c: Include sel-sched.h.
(ref_counts): New static variable. Use it ...
(INSN_REF_COUNT): ... here. Rewrite and move closer to uses.
(FED_BY_SPEC_LOAD, IS_LOAD_INSN): Rewrite to use HID accessor macro.
(sched_is_disabled_for_current_region_p): Delete static declaration.
(struct region): Move to sched-int.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
ebb_head): Define and initialize.
(RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN, debug_regions, extend_regions,
BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB): Move to
sched-int.h.
(find_single_block_region): Add new argument to indicate that EBB
regions should be constructed.
(debug_live): Delete declaration.
(current_nr_blocks, current_blocks, target_bb): Remove static qualifiers.
(compute_dom_prob_ps, check_live, update_live, set_spec_fed): Delete
declaration.
(init_regions): Delete declaration.
(debug_region, bb_in_region_p, dump_region_dot_file, dump_region_dot,
rgn_estimate_number_of_insns): New.
(too_large): Use estimate_number_of_insns.
(haifa_find_rgns): New. Move the code from ...
(find_rgns): ... here. Call either sel_find_rgns or haifa_find_rgns.
(free_trg_info): New.
(compute_trg_info): Allocate candidate tables here instead of ...
(init_ready_list): ... here.
(rgn_print_insn): Use const_rtx.
(contributes_to_priority, extend_regions): Delete static declaration.
(add_remove_insn, fix_recovery_cfg): Add rgn_ to function names.
(add_block1): Rename to rgn_add_block.
(debug_rgn_dependencies): Delete static qualifier.
(new_ready): Use sched_deps_info. Simplify.
(rgn_common_sched_info, rgn_const_sched_deps_info,
rgn_const_sel_sched_deps_info, rgn_sched_deps_info, rgn_sched_info): New.
(region_sched_info): Rename to rgn_const_sched_info.
(deps_join): New, extracted from ...
(propagate_deps): ... here.
(compute_block_dependences, debug_dependencies): Update for selective
scheduling.
(free_rgn_deps, compute_priorities): New functions.
(sched_rgn_init, sched_rgn_finish, rgn_setup_region,
sched_rgn_compute_dependencies): New functions.
(schedule_region): Use them.
(sched_rgn_local_init, sched_rgn_local_free, sched_rgn_local_finish,
rgn_setup_common_sched_info, rgn_setup_sched_infos):
New functions.
(schedule_insns): Call new functions that were split out.
(rgn_make_new_region_out_of_new_block): New.
(get_rgn_sched_max_insns_priority): New.
(rest_of_handle_sched, rest_of_handle_sched2): Call selective
scheduling when appropriate.
* sched-vis.c: Include insn-attr.h.
(print_value, print_pattern): Make global.
(print_rtl_slim, debug_bb_slim, debug_bb_n_slim): New functions.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_ALLOC_SCHED_CONTEXT, TARGET_SCHED_INIT_SCHED_CONTEXT,
TARGET_SCHED_SET_SCHED_CONTEXT, TARGET_SCHED_CLEAR_SCHED_CONTEXT,
TARGET_SCHED_FREE_SCHED_CONTEXT, TARGET_SCHED_GET_INSN_CHECKED_DS,
TARGET_SCHED_GET_INSN_SPEC_DS, TARGET_SCHED_SKIP_RTX_P): New target
hooks. Initialize them to 0.
(TARGET_SCHED_GEN_CHECK): Rename to TARGET_SCHED_GEN_SPEC_CHECK.
* target.h (struct gcc_target): Add them. Rename gen_check field to
gen_spec_check.
* flags.h (sel_sched_switch_set): Declare.
* opts.c (sel_sched_switch_set): New variable.
(decode_options): Unset flag_sel_sched_pipelining_outer_loops if
pipelining is disabled from command line.
(common_handle_option): Record whether selective scheduling is
requested from command line.
* doc/invoke.texi: Document new flags and parameters.
* doc/tm.texi: Document new target hooks.
* config/ia64/ia64.c (TARGET_SCHED_GEN_SPEC_CHECK): Define to ia64_gen_check.
(dfa_state_size): Do not declare locally.
* config/ia64/ia64.opt (msched-ar-data-spec): Default to 0.
* config/rs6000/rs6000.c (rs6000_init_sched_context,
rs6000_alloc_sched_context, rs6000_set_sched_context,
rs6000_free_sched_context): New functions.
(struct _rs6000_sched_context): New.
(rs6000_sched_reorder2): Do not modify INSN_PRIORITY for selective
scheduling.
(rs6000_sched_finish): Do not run for selective scheduling.
Co-Authored-By: Alexander Monakov <amonakov@ispras.ru>
Co-Authored-By: Dmitry Melnik <dm@ispras.ru>
Co-Authored-By: Dmitry Zhurikhin <zhur@ispras.ru>
Co-Authored-By: Maxim Kuvyrkov <maxim@codesourcery.com>
From-SVN: r139854
2008-09-01 10:57:00 +02:00
200 , 0 , 0 )
2005-08-24 22:28:07 +02:00
DEFPARAM ( PARAM_MIN_SPEC_PROB ,
"min-spec-prob" ,
2015-10-22 00:24:41 +02:00
"The minimum probability of reaching a source block for interblock speculative scheduling." ,
2005-08-24 22:28:07 +02:00
40 , 0 , 0 )
2006-03-16 06:20:39 +01:00
DEFPARAM ( PARAM_MAX_SCHED_EXTEND_REGIONS_ITERS ,
"max-sched-extend-regions-iters" ,
2015-10-22 00:24:41 +02:00
"The maximum number of iterations through CFG to extend regions." ,
2006-06-06 11:44:17 +02:00
0 , 0 , 0 )
2006-03-16 06:20:39 +01:00
target.h (struct spec_info_def): New opaque declaration.
2006-03-16 Maxim Kuvyrkov <mkuvyrkov@ispras.ru>
* target.h (struct spec_info_def): New opaque declaration.
(struct gcc_target.sched): New fields: adjust_cost_2, h_i_d_extended,
speculate_insn, needs_block_p, gen_check,
first_cycle_multipass_dfa_lookahead_guard_spec, set_sched_flags.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_H_I_D_EXTENDED, TARGET_SCHED_SPECULATE_INSN,
TARGET_SCHED_NEEDS_BLOCK_P, TARGET_SCHED_GEN_CHECK,
TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC,
TARGET_SCHED_SET_SCHED_FLAGS): New macros to initialize fields in
gcc_target.sched.
(TARGET_SCHED): Use new macros.
* rtl.h (copy_DEPS_LIST_list): New prototype.
* sched-int.h (struct sched_info): Change signature of new_ready field,
adjust all initializations. New fields: add_remove_insn,
begin_schedule_ready, add_block, advance_target_bb, fix_recovery_cfg,
region_head_or_leaf_p.
(struct spec_info_def): New structure declaration.
(spec_info_t): New typedef.
(struct haifa_insn_data): New fields: todo_spec, done_spec, check_spec,
recovery_block, orig_pat.
(glat_start, glat_end): New variables declaraions.
(TODO_SPEC, DONE_SPEC, CHECK_SPEC, RECOVERY_BLOCK, ORIG_PAT):
New access macros.
(enum SCHED_FLAGS): New constants: SCHED_RGN, SCHED_EBB,
DETACH_LIFE_INFO, USE_GLAT.
(enum SPEC_SCHED_FLAGS): New enumeration.
(NOTE_NOTE_BB_P): New macro.
(extend_dependency_caches, xrecalloc, unlink_bb_notes, add_block,
attach_life_info, debug_spec_status, check_reg_live): New functions.
(get_block_head_tail): Change signature to get_ebb_head_tail, adjust
all uses in ddg.c, modulo-sched.c, haifa-sched.c, sched-rgn.c,
sched-ebb.c
(get_dep_weak, ds_merge): Prototype functions from sched-deps.c .
* ddg.c (get_block_head_tail): Adjust all uses.
* modulo-sched.c (get_block_head_tail): Adjust all uses.
(sms_sched_info): Initialize new fields.
(contributes_to_priority): Removed.
* haifa-sched.c (params.h): New include.
(get_block_head_tail): Adjust all uses.
(ISSUE_POINTS): New macro.
(glat_start, glat_end): New global variables.
(spec_info_var, spec_info, added_recovery_block_p, nr_begin_data,
nr_be_in_data, nr_begin_control, nr_be_in_control, bb_header,
old_last_basic_block, before_recovery, current_sched_info_var,
rgn_n_insns, luid): New static variables.
(insn_cost1): New function. Move logic from insn_cost to here.
(find_insn_reg_weight1): New function. Move logic from
find_insn_reg_weight to here.
(reemit_notes, move_insn, max_issue): Change signature.
(move_insn1): Removed.
(extend_h_i_d, extend_ready, extend_global, extend_all, init_h_i_d,
extend_bb): New static functions to support extension of scheduler's
data structures.
(generate_recovery_code, process_insn_depend_be_in_spec,
begin_speculative_block, add_to_speculative_block,
init_before_recovery, create_recovery_block, create_check_block_twin,
fix_recovery_deps): New static functions to support
generation of recovery code.
(fix_jump_move, find_fallthru_edge, dump_new_block_header,
restore_bb_notes, move_block_after_check, move_succs): New static
functions to support ebb scheduling.
(init_glat, init_glat1, attach_life_info1, free_glat): New static
functions to support handling of register live information.
(associate_line_notes_with_blocks, change_pattern, speculate_insn,
sched_remove_insn, clear_priorities, calc_priorities, bb_note,
add_jump_dependencies): New static functions.
(check_cfg, has_edge_p, check_sched_flags): New static functions for
consistancy checking.
(debug_spec_status): New function to call from debugger.
(priority): Added code to handle speculation checks.
(rank_for_schedule): Added code to distinguish speculative instructions.
(schedule_insn): Added code to handle speculation checks.
(unlink_other_notes, rm_line_notes, restore_line_notes, rm_other_notes):
Fixed to handle ebbs.
(move_insn): Added code to handle ebb scheduling.
(max_issue): Added code to use ISSUE_POINTS of instructions.
(choose_ready): Added code to choose between speculative and
non-speculative instructions.
(schedule_block): Added code to handle ebb scheduling and scheduling of
speculative instructions.
(sched_init): Initialize new variables.
(sched_finish): Free new variables. Print statistics.
(try_ready): Added code to handle speculative instructions.
* lists.c (copy_DEPS_LIST_list): New function.
* sched-deps.c (extend_dependency_caches): New function. Move logic
from create_dependency_caches to here.
(get_dep_weak, ds_merge): Make global.
* genattr.c (main): Code to output prototype for
dfa_clear_single_insn_cache.
* genautomata.c (DFA_CLEAR_SINGLE_INSN_CACHE_FUNC_NAME): New macros.
(output_dfa_clean_insn_cache_func): Code to output
dfa_clear_single_insn_cache function.
* sched-ebb.c (target_n_insns): Remove. Adjust all users to use
n_insns.
(can_schedule_ready_p, fix_basic_block_boundaries, add_missing_bbs):
Removed.
(n_insns, dont_calc_deps, ebb_head, ebb_tail, last_bb):
New static variables.
(begin_schedule_ready, add_remove_insn, add_block1, advance_target_bb,
fix_recovery_cfg, ebb_head_or_leaf_p): Implement hooks from
struct sched_info.
(ebb_sched_info): Initialize new fields.
(get_block_head_tail): Adjust all uses.
(compute_jump_reg_dependencies): Fixed to use glat_start.
(schedule_ebb): Code to remove unreachable last block.
(schedule_ebbs): Added code to update register live information.
* sched-rgn.c (region_sched_info): Initialize new fields.
(get_block_head_tail): Adjust all uses.
(last_was_jump): Removed. Adjust users.
(begin_schedule_ready, add_remove_insn, insn_points, extend_regions,
add_block1, fix_recovery_cfg, advance_target_bb, region_head_or_leaf_p):
Implement new hooks.
(check_dead_notes1): New static function.
(struct region): New fields: dont_calc_deps, has_real_ebb.
(RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB): New access macros.
(BB_TO_BLOCK): Fixed to handle EBBs.
(EBB_FIRST_BB, EBB_LAST_BB): New macros.
(ebb_head): New static variable.
(debug_regions, contributes_to_priority): Fixed to handle EBBs.
(find_single_block_regions, find_rgns, find_more_rgns): Initialize
new fields.
(compute_dom_prob_ps): New assertion.
(check_live_1, update_live_1): Fixed to work with glat_start instead of
global_live_at_start.
(init_ready_list): New assertions.
(can_schedule_ready_p): Split update code to begin_schedule_ready.
(new_ready): Add support for BEGIN_CONTROL speculation.
(schedule_insns): Fixed code that updates register live information
to handle EBBs.
(schedule_region): Fixed to handle EBBs.
(init_regions): Use extend_regions and check_dead_notes1.
* params.def (PARAM_MAX_SCHED_INSN_CONFLICT_DELAY,
PARAM_SCHED_SPEC_PROB_CUTOFF): New parameters.
* doc/tm.texi (TARGET_SCHED_ADJUST_COST_2, TARGET_SCHED_H_I_D_EXTENDED,
TARGET_SCHED_SPECULATE_INSN, TARGET_SCHED_NEEDS_BLOCK_P,
TARGET_SCHED_GEN_CHECK,
TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC,
TARGET_SCHED_SET_SCHED_FLAGS): Document.
* doc/invoke.texi (max-sched-insn-conflict-delay,
sched-spec-prob-cutoff): Document.
From-SVN: r112128
2006-03-16 06:27:03 +01:00
DEFPARAM ( PARAM_MAX_SCHED_INSN_CONFLICT_DELAY ,
"max-sched-insn-conflict-delay" ,
2015-10-22 00:24:41 +02:00
"The maximum conflict delay for an insn to be considered for speculative motion." ,
target.h (struct spec_info_def): New opaque declaration.
2006-03-16 Maxim Kuvyrkov <mkuvyrkov@ispras.ru>
* target.h (struct spec_info_def): New opaque declaration.
(struct gcc_target.sched): New fields: adjust_cost_2, h_i_d_extended,
speculate_insn, needs_block_p, gen_check,
first_cycle_multipass_dfa_lookahead_guard_spec, set_sched_flags.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_H_I_D_EXTENDED, TARGET_SCHED_SPECULATE_INSN,
TARGET_SCHED_NEEDS_BLOCK_P, TARGET_SCHED_GEN_CHECK,
TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC,
TARGET_SCHED_SET_SCHED_FLAGS): New macros to initialize fields in
gcc_target.sched.
(TARGET_SCHED): Use new macros.
* rtl.h (copy_DEPS_LIST_list): New prototype.
* sched-int.h (struct sched_info): Change signature of new_ready field,
adjust all initializations. New fields: add_remove_insn,
begin_schedule_ready, add_block, advance_target_bb, fix_recovery_cfg,
region_head_or_leaf_p.
(struct spec_info_def): New structure declaration.
(spec_info_t): New typedef.
(struct haifa_insn_data): New fields: todo_spec, done_spec, check_spec,
recovery_block, orig_pat.
(glat_start, glat_end): New variables declaraions.
(TODO_SPEC, DONE_SPEC, CHECK_SPEC, RECOVERY_BLOCK, ORIG_PAT):
New access macros.
(enum SCHED_FLAGS): New constants: SCHED_RGN, SCHED_EBB,
DETACH_LIFE_INFO, USE_GLAT.
(enum SPEC_SCHED_FLAGS): New enumeration.
(NOTE_NOTE_BB_P): New macro.
(extend_dependency_caches, xrecalloc, unlink_bb_notes, add_block,
attach_life_info, debug_spec_status, check_reg_live): New functions.
(get_block_head_tail): Change signature to get_ebb_head_tail, adjust
all uses in ddg.c, modulo-sched.c, haifa-sched.c, sched-rgn.c,
sched-ebb.c
(get_dep_weak, ds_merge): Prototype functions from sched-deps.c .
* ddg.c (get_block_head_tail): Adjust all uses.
* modulo-sched.c (get_block_head_tail): Adjust all uses.
(sms_sched_info): Initialize new fields.
(contributes_to_priority): Removed.
* haifa-sched.c (params.h): New include.
(get_block_head_tail): Adjust all uses.
(ISSUE_POINTS): New macro.
(glat_start, glat_end): New global variables.
(spec_info_var, spec_info, added_recovery_block_p, nr_begin_data,
nr_be_in_data, nr_begin_control, nr_be_in_control, bb_header,
old_last_basic_block, before_recovery, current_sched_info_var,
rgn_n_insns, luid): New static variables.
(insn_cost1): New function. Move logic from insn_cost to here.
(find_insn_reg_weight1): New function. Move logic from
find_insn_reg_weight to here.
(reemit_notes, move_insn, max_issue): Change signature.
(move_insn1): Removed.
(extend_h_i_d, extend_ready, extend_global, extend_all, init_h_i_d,
extend_bb): New static functions to support extension of scheduler's
data structures.
(generate_recovery_code, process_insn_depend_be_in_spec,
begin_speculative_block, add_to_speculative_block,
init_before_recovery, create_recovery_block, create_check_block_twin,
fix_recovery_deps): New static functions to support
generation of recovery code.
(fix_jump_move, find_fallthru_edge, dump_new_block_header,
restore_bb_notes, move_block_after_check, move_succs): New static
functions to support ebb scheduling.
(init_glat, init_glat1, attach_life_info1, free_glat): New static
functions to support handling of register live information.
(associate_line_notes_with_blocks, change_pattern, speculate_insn,
sched_remove_insn, clear_priorities, calc_priorities, bb_note,
add_jump_dependencies): New static functions.
(check_cfg, has_edge_p, check_sched_flags): New static functions for
consistancy checking.
(debug_spec_status): New function to call from debugger.
(priority): Added code to handle speculation checks.
(rank_for_schedule): Added code to distinguish speculative instructions.
(schedule_insn): Added code to handle speculation checks.
(unlink_other_notes, rm_line_notes, restore_line_notes, rm_other_notes):
Fixed to handle ebbs.
(move_insn): Added code to handle ebb scheduling.
(max_issue): Added code to use ISSUE_POINTS of instructions.
(choose_ready): Added code to choose between speculative and
non-speculative instructions.
(schedule_block): Added code to handle ebb scheduling and scheduling of
speculative instructions.
(sched_init): Initialize new variables.
(sched_finish): Free new variables. Print statistics.
(try_ready): Added code to handle speculative instructions.
* lists.c (copy_DEPS_LIST_list): New function.
* sched-deps.c (extend_dependency_caches): New function. Move logic
from create_dependency_caches to here.
(get_dep_weak, ds_merge): Make global.
* genattr.c (main): Code to output prototype for
dfa_clear_single_insn_cache.
* genautomata.c (DFA_CLEAR_SINGLE_INSN_CACHE_FUNC_NAME): New macros.
(output_dfa_clean_insn_cache_func): Code to output
dfa_clear_single_insn_cache function.
* sched-ebb.c (target_n_insns): Remove. Adjust all users to use
n_insns.
(can_schedule_ready_p, fix_basic_block_boundaries, add_missing_bbs):
Removed.
(n_insns, dont_calc_deps, ebb_head, ebb_tail, last_bb):
New static variables.
(begin_schedule_ready, add_remove_insn, add_block1, advance_target_bb,
fix_recovery_cfg, ebb_head_or_leaf_p): Implement hooks from
struct sched_info.
(ebb_sched_info): Initialize new fields.
(get_block_head_tail): Adjust all uses.
(compute_jump_reg_dependencies): Fixed to use glat_start.
(schedule_ebb): Code to remove unreachable last block.
(schedule_ebbs): Added code to update register live information.
* sched-rgn.c (region_sched_info): Initialize new fields.
(get_block_head_tail): Adjust all uses.
(last_was_jump): Removed. Adjust users.
(begin_schedule_ready, add_remove_insn, insn_points, extend_regions,
add_block1, fix_recovery_cfg, advance_target_bb, region_head_or_leaf_p):
Implement new hooks.
(check_dead_notes1): New static function.
(struct region): New fields: dont_calc_deps, has_real_ebb.
(RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB): New access macros.
(BB_TO_BLOCK): Fixed to handle EBBs.
(EBB_FIRST_BB, EBB_LAST_BB): New macros.
(ebb_head): New static variable.
(debug_regions, contributes_to_priority): Fixed to handle EBBs.
(find_single_block_regions, find_rgns, find_more_rgns): Initialize
new fields.
(compute_dom_prob_ps): New assertion.
(check_live_1, update_live_1): Fixed to work with glat_start instead of
global_live_at_start.
(init_ready_list): New assertions.
(can_schedule_ready_p): Split update code to begin_schedule_ready.
(new_ready): Add support for BEGIN_CONTROL speculation.
(schedule_insns): Fixed code that updates register live information
to handle EBBs.
(schedule_region): Fixed to handle EBBs.
(init_regions): Use extend_regions and check_dead_notes1.
* params.def (PARAM_MAX_SCHED_INSN_CONFLICT_DELAY,
PARAM_SCHED_SPEC_PROB_CUTOFF): New parameters.
* doc/tm.texi (TARGET_SCHED_ADJUST_COST_2, TARGET_SCHED_H_I_D_EXTENDED,
TARGET_SCHED_SPECULATE_INSN, TARGET_SCHED_NEEDS_BLOCK_P,
TARGET_SCHED_GEN_CHECK,
TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC,
TARGET_SCHED_SET_SCHED_FLAGS): Document.
* doc/invoke.texi (max-sched-insn-conflict-delay,
sched-spec-prob-cutoff): Document.
From-SVN: r112128
2006-03-16 06:27:03 +01:00
3 , 1 , 10 )
DEFPARAM ( PARAM_SCHED_SPEC_PROB_CUTOFF ,
"sched-spec-prob-cutoff" ,
"The minimal probability of speculation success (in percents), so that speculative insn will be scheduled." ,
40 , 0 , 100 )
2012-10-08 13:26:16 +02:00
DEFPARAM ( PARAM_SCHED_STATE_EDGE_PROB_CUTOFF ,
"sched-state-edge-prob-cutoff" ,
"The minimum probability an edge must have for the scheduler to save its state across it." ,
10 , 0 , 100 )
sel-sched.h, [...]: New files.
2008-08-31 Andrey Belevantsev <abel@ispras.ru>
Dmitry Melnik <dm@ispras.ru>
Dmitry Zhurikhin <zhur@ispras.ru>
Alexander Monakov <amonakov@ispras.ru>
Maxim Kuvyrkov <maxim@codesourcery.com>
* sel-sched.h, sel-sched-dump.h, sel-sched-ir.h, sel-sched.c,
sel-sched-dump.c, sel-sched-ir.c: New files.
* Makefile.in (OBJS-common): Add selective scheduling object
files.
(sel-sched.o, sel-sched-dump.o, sel-sched-ir.o): New entries.
(SEL_SCHED_IR_H, SEL_SCHED_DUMP_H): New entries.
(sched-vis.o): Add dependency on $(INSN_ATTR_H).
* cfghooks.h (get_cfg_hooks, set_cfg_hooks): New prototypes.
* cfghooks.c (get_cfg_hooks, set_cfg_hooks): New functions.
(make_forwarder_block): Update loop latch if we have redirected
the loop latch edge.
* cfgloop.c (get_loop_body_in_custom_order): New function.
* cfgloop.h (LOOPS_HAVE_FALLTHRU_PREHEADERS): New enum field.
(CP_FALLTHRU_PREHEADERS): Likewise.
(get_loop_body_in_custom_order): Declare.
* cfgloopmanip.c (has_preds_from_loop): New.
(create_preheader): Honor CP_FALLTHRU_PREHEADERS.
Assert that the preheader edge will be fall thru when it is set.
* common.opt (fsel-sched-bookkeeping, fsel-sched-pipelining,
fsel-sched-pipelining-outer-loops, fsel-sched-renaming,
fsel-sched-substitution, fselective-scheduling): New flags.
* cse.c (hash_rtx_cb): New.
(hash_rtx): Use it.
* dbgcnt.def (sel_sched_cnt, sel_sched_region_cnt,
sel_sched_insn_cnt): New counters.
* final.c (compute_alignments): Export. Free dominance info after loop_optimizer_finalize.
* genattr.c (main): Output maximal_insn_latency prototype.
* genautomata.c (output_default_latencies): New. Factor its code from ...
(output_internal_insn_latency_func): ... here.
(output_internal_maximal_insn_latency_func): New.
(output_maximal_insn_latency_func): New.
* hard-reg-set.h (UHOST_BITS_PER_WIDE_INT): Define unconditionally.
(struct hard_reg_set_iterator): New.
(hard_reg_set_iter_init, hard_reg_set_iter_set,
hard_reg_set_iter_next): New functions.
(EXECUTE_IF_SET_IN_HARD_REG_SET): New macro.
* lists.c (remove_free_INSN_LIST_node,
remove_free_EXPR_LIST_node): New functions.
* loop-init.c (loop_optimizer_init): When LOOPS_HAVE_FALLTHRU_PREHEADERS,
set CP_FALLTHRU_PREHEADERS when calling create_preheaders.
(loop_optimizer_finalize): Do not verify flow info after reload.
* recog.c (validate_replace_rtx_1): New parameter simplify.
Default it to true. Update all uses. Factor out simplifying
code to ...
(simplify_while_replacing): ... this new function.
(validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): New.
* recog.h (validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): Declare.
* rtl.c (rtx_equal_p_cb): New.
(rtx_equal_p): Use it.
* rtl.h (rtx_equal_p_cb, hash_rtx_cb): Declare.
(remove_free_INSN_LIST_NODE, remove_free_EXPR_LIST_node,
debug_bb_n_slim, debug_bb_slim, print_rtl_slim): Likewise.
* vecprim.h: Add a vector type for unsigned int.
* haifa-sched.c: Include vecprim.h and cfgloop.h.
(issue_rate, sched_verbose_param, note_list, dfa_state_size,
ready_try, cycle_issued_insns, spec_info): Make global.
(readyp): Initialize.
(dfa_lookahead): New global variable.
(old_max_uid, old_last_basic_block): Remove.
(h_i_d): Make it a vector.
(INSN_TICK, INTER_TICK, QUEUE_INDEX, INSN_COST): Make them work
through HID macro.
(after_recovery, adding_bb_to_current_region_p):
New variables to handle correct insertion of the recovery code.
(struct ready_list): Move declaration to sched-int.h.
(rgn_n_insns): Removed.
(rtx_vec_t): Move to sched-int.h.
(find_insn_reg_weight): Remove.
(find_insn_reg_weight1): Rename to find_insn_reg_weight.
(haifa_init_h_i_d, haifa_finish_h_i_d):
New functions to initialize / finalize haifa instruction data.
(extend_h_i_d, init_h_i_d): Rewrite.
(unlink_other_notes): Move logic to add_to_note_list. Handle
selective scheduler.
(ready_lastpos, ready_element, ready_sort, reemit_notes,
find_fallthru_edge): Make global, remove static prototypes.
(max_issue): Make global. Add privileged_n and state parameters. Use
them.
(extend_global, extend_all): Removed.
(init_before_recovery): Add new param. Fix the handling of the case
when we insert a recovery code before the EXIT which has a predecessor
with a fallthrough edge to it.
(create_recovery_block): Make global. Rename to
sched_create_recovery_block. Update.
(change_pattern): Rename to sched_change_pattern. Make global.
(speculate_insn): Rename to sched_speculate_insn. Make global.
Split haifa-specific functionality into ...
(haifa_change_pattern): New static function.
(sched_extend_bb): New static function.
(sched_init_bbs): New function.
(current_sched_info): Change type to struct haifa_sched_info.
(insn_cost): Adjust for selective scheduling.
(dep_cost_1): New function. Move logic from ...
(dep_cost): ... here.
(dep_cost): Use dep_cost_1.
(contributes_to_priority_p): Use sched_deps_info instead of
current_sched_info.
(priority): Adjust to work with selective scheduling. Process the
corner case when all dependencies don't contribute to priority.
(rank_for_schedule): Use ds_weak instead of dep_weak.
(advance_state): New function. Move logic from ...
(advance_one_cycle): ... here.
(add_to_note_list, concat_note_lists): New functions.
(rm_other_notes): Make static. Adjust for selective scheduling.
(remove_notes, restore_other_notes): New functions.
(move_insn): Add two arguments. Update assert. Don't call
reemit_notes.
(choose_ready): Remove lookahead variable, use dfa_lookahead.
Remove more_issue, max_points. Move the code to initialize
max_lookahead_tries to max_issue.
(schedule_block): Remove rgn_n_insns1 parameter. Don't allocate
ready. Adjust use of move_insn. Call restore_other_notes.
(luid): Remove.
(sched_init, sched_finish): Move Haifa-specific initialization/
finalization to ...
(haifa_sched_init, haifa_sched_finish): ... respectively.
New functions.
(setup_sched_dump): New function.
(haifa_init_only_bb): New static function.
(haifa_speculate_insn): New static function.
(try_ready): Use haifa_* instead of speculate_insn and
change_pattern.
(extend_ready, extend_all): Remove.
(sched_extend_ready_list, sched_finish_ready_list): New functions.
(create_check_block_twin, add_to_speculative_block): Use
haifa_insns_init instead of extend_global. Update to use new
initialization functions. Change parameter. Factor out code from
create_check_block_twin to ...
(sched_create_recovery_edges) ... this new function.
(add_block): Remove.
(sched_scan_info): New.
(extend_bb): Use sched_scan_info.
(init_bb, extend_insn, init_insn, init_insns_in_bb, sched_scan): New
static functions for walking through scheduling region.
(sched_luids): New vector variable to replace uid_to_luid.
(luids_extend_insn): New function.
(sched_max_luid): New variable.
(luids_init_insn): New function.
(sched_init_luids, sched_finish_luids): New functions.
(insn_luid): New debug function.
(sched_extend_target): New function.
(haifa_init_insn): New static function.
(sched_init_only_bb): New hook.
(sched_split_block): New hook.
(sched_split_block_1): New function.
(sched_create_empty_bb): New hook.
(sched_create_empty_bb_1): New function.
(common_sched_info, ready): New global variables.
(current_sched_info_var): Remove.
(move_block_after_check): Use common_sched_info.
(haifa_luid_for_non_insn): New static function.
(init_before_recovery): Use haifa_init_only_bb instead of
add_block.
(increase_insn_priority): New.
* modulo-sched.c: (issue_rate): Remove static declaration.
(sms_sched_info): Change type to haifa_sched_info.
(sms_sched_deps_info, sms_common_sched_info): New variables.
(setup_sched_infos): New.
(sms_schedule): Initialize them. Call haifa_sched_init/finish.
Do not call regstat_free_calls_crossed.
(sms_print_insn): Use const_rtx.
* params.def (PARAM_MAX_PIPELINE_REGION_BLOCKS,
PARAM_MAX_PIPELINE_REGION_INSNS, PARAM_SELSCHED_MAX_LOOKAHEAD,
PARAM_SELSCHED_MAX_SCHED_TIMES, PARAM_SELSCHED_INSNS_TO_RENAME,
PARAM_SCHED_MEM_TRUE_DEP_COST): New.
* sched-deps.c (sched_deps_info): New. Update all relevant uses of
current_sched_info to use it.
(enum reg_pending_barrier_mode): Move to sched-int.h.
(h_d_i_d): New variable. Initialize to NULL.
({true, output, anti, spec, forward}_dependency_cache): Initialize
to NULL.
(estimate_dep_weak): Remove static declaration.
(sched_has_condition_p): New function. Adjust users of
sched_get_condition to use it instead.
(conditions_mutex_p): Add arguments indicating which conditions are
reversed. Use them.
(sched_get_condition_with_rev): Rename from sched_get_condition. Add
argument to indicate whether returned condition is reversed. Do not
generate new rtx when condition should be reversed; indicate it by
setting new argument instead.
(add_dependence_list_and_free): Add deps parameter.
Update all users. Do not free dependence list when
deps context is readonly.
(add_insn_mem_dependence, flush_pending_lists): Adjust for readonly
contexts.
(remove_from_dependence_list, remove_from_both_dependence_lists): New.
(remove_from_deps): New. Use the above functions.
(cur_insn, can_start_lhs_rhs_p): New static variables.
(add_or_update_back_dep_1): Initialize present_dep_type.
(haifa_start_insn, haifa_finish_insn, haifa_note_reg_set,
haifa_note_reg_clobber, haifa_note_reg_use, haifa_note_mem_dep,
haifa_note_dep): New functions implementing dependence hooks for
the Haifa scheduler.
(note_reg_use, note_reg_set, note_reg_clobber, note_mem_dep,
note_dep): New functions.
(ds_to_dt, extend_deps_reg_info, maybe_extend_reg_info_p): New
functions.
(init_deps): Initialize last_reg_pending_barrier and deps->readonly.
(free_deps): Initialize deps->reg_last.
(sched_analyze_reg, sched_analyze_1, sched_analyze_2,
sched_analyze_insn): Update to use dependency hooks infrastructure
and readonly contexts.
(deps_analyze_insn): New function. Move part of logic from ...
(sched_analyze): ... here. Also move some logic to ...
(deps_start_bb): ... here. New function.
(add_forw_dep, delete_forw_dep): Guard use of INSN_DEP_COUNT with
sel_sched_p.
(sched_deps_init): New function. Move code from ...
(init_dependency_caches): ... here. Remove.
(init_deps_data_vector): New.
(sched_deps_finish): New function. Move code from ...
(free_dependency_caches): ... here. Remove.
(init_deps_global, finish_deps_global): Adjust for use with
selective scheduling.
(get_dep_weak): Move logic to ...
(get_dep_weak_1): New function.
(ds_merge): Move logic to ...
(ds_merge_1): New static function.
(ds_full_merge, ds_max_merge, ds_get_speculation_types): New functions.
(ds_get_max_dep_weak): New function.
* sched-ebb.c (sched_n_insns): Rename to sched_rgn_n_insns.
(n_insns): Rename to rgn_n_insns.
(debug_ebb_dependencies): New function.
(init_ready_list): Use it.
(begin_schedule_ready): Use sched_init_only_bb.
(ebb_print_insn): Indicate when an insn starts a new cycle.
(contributes_to_priority, compute_jump_reg_dependencies,
add_remove_insn, fix_recovery_cfg): Add ebb_ prefix to function names.
(add_block1): Remove to ebb_add_block.
(ebb_sched_deps_info, ebb_common_sched_info): New variables.
(schedule_ebb): Initialize them. Use remove_notes instead of
rm_other_notes. Use haifa_local_init/finish.
(schedule_ebbs): Use haifa_sched_init/finish.
* sched-int.h: Include vecprim.h, remove rtl.h.
(struct ready_list): Delete declaration.
(sched_verbose_param, enum sched_pass_id_t,
bb_vec_t, insn_vec_t, rtx_vec_t): New.
(struct sched_scan_info_def): New structure.
(sched_scan_info, sched_scan, sched_init_bbs,
sched_init_luids, sched_finish_luids, sched_extend_target,
haifa_init_h_i_d, haifa_finish_h_i_d): Declare.
(struct common_sched_info_def): New.
(common_sched_info, haifa_common_sched_info,
sched_emulate_haifa_p): Declare.
(sel_sched_p): New.
(sched_luids): Declare.
(INSN_LUID, LUID_BY_UID, SET_INSN_LUID): Declare.
(sched_max_luid, insn_luid): Declare.
(note_list, remove_notes, restore_other_notes, bb_note): Declare.
(sched_insns_init, sched_insns_finish, xrecalloc, reemit_notes,
print_insn, print_pattern, print_value, haifa_classify_insn,
sel_find_rgns, sel_mark_hard_insn, dfa_state_size, advance_state,
setup_sched_dump, sched_init, sched_finish,
sel_insn_is_speculation_check): Export.
(struct ready_list): Move from haifa-sched.c.
(ready_try, ready, max_issue): Export.
(ebb_compute_jump_reg_dependencies, find_fallthru_edge,
sched_init_only_bb, sched_split_block, sched_split_block_1,
sched_create_empty_bb, sched_create_empty_bb_1,
sched_create_recovery_block, sched_create_recovery_edges): Export.
(enum reg_pending_barrier_mode): Export.
(struct deps): New fields `last_reg_pending_barrier' and `readonly'.
(deps_t): New.
(struct sched_info): Rename to haifa_sched_info. Use const_rtx for
print_insn field. Move add_block and fix_recovery_cfg to
common_sched_info_def. Move compute_jump_reg_dependencies, use_cselib ...
(struct sched_deps_info_def): ... this new structure.
(sched_deps_info): Declare.
(struct spec_info_def): Remove weakness_cutoff, add
data_weakness_cutoff and control_weakness_cutoff.
(spec_info): Declare.
(struct _haifa_deps_insn_data): Split from haifa_insn_data. Add
dep_count field.
(struct haifa_insn_data): Rename to struct _haifa_insn_data.
(haifa_insn_data_def, haifa_insn_data_t): New typedefs.
(current_sched_info): Change type to struct haifa_sched_info.
(haifa_deps_insn_data_def, haifa_deps_insn_data_t): New typedefs.
(h_d_i_d): New variable.
(HDID): New accessor macro.
(h_i_d): Change type to VEC (haifa_insn_data_def, heap) *.
(HID): New accessor macro. Rewrite h_i_d accessor macros through HID
and HDID.
(IS_SPECULATION_CHECK_P): Update for selective scheduler.
(enum SCHED_FLAGS): Update for selective scheduler.
(enum SPEC_SCHED_FLAGS): New flag SEL_SCHED_SPEC_DONT_CHECK_CONTROL.
(init_dependency_caches, free_dependency_caches): Delete declarations.
(deps_analyze_insn, remove_from_deps, get_dep_weak_1,
estimate_dep_weak, ds_full_merge, ds_max_merge, ds_weak,
ds_get_speculation_types, ds_get_max_dep_weak, sched_deps_init,
sched_deps_finish, haifa_note_reg_set, haifa_note_reg_use,
haifa_note_reg_clobber, maybe_extend_reg_info_p, deps_start_bb,
ds_to_dt): Export.
(rm_other_notes): Delete declaration.
(schedule_block): Remove one argument.
(cycle_issued_insns, issue_rate, dfa_lookahead, ready_sort,
ready_element, ready_lastpos, sched_extend_ready_list,
sched_finish_ready_list, sched_change_pattern, sched_speculate_insn,
concat_note_lists): Export.
(struct region): Move from sched-rgn.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN): Export.
(ebb_head, BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB, INSN_BB): Likewise.
(current_nr_blocks, current_blocks, target_bb): Likewise.
(dep_cost_1, sched_is_disabled_for_current_region_p, sched_rgn_init,
sched_rgn_finish, rgn_setup_region, sched_rgn_compute_dependencies,
sched_rgn_local_init, extend_regions,
rgn_make_new_region_out_of_new_block, compute_priorities,
debug_rgn_dependencies, free_rgn_deps, contributes_to_priority,
extend_rgns, deps_join rgn_setup_common_sched_info,
rgn_setup_sched_infos, debug_regions, debug_region, dump_region_dot,
dump_region_dot_file, haifa_sched_init, haifa_sched_finish): Export.
(get_rgn_sched_max_insns_priority, sel_add_to_insn_priority,
increase_insn_priority): Likewise.
* sched-rgn.c: Include sel-sched.h.
(ref_counts): New static variable. Use it ...
(INSN_REF_COUNT): ... here. Rewrite and move closer to uses.
(FED_BY_SPEC_LOAD, IS_LOAD_INSN): Rewrite to use HID accessor macro.
(sched_is_disabled_for_current_region_p): Delete static declaration.
(struct region): Move to sched-int.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
ebb_head): Define and initialize.
(RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN, debug_regions, extend_regions,
BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB): Move to
sched-int.h.
(find_single_block_region): Add new argument to indicate that EBB
regions should be constructed.
(debug_live): Delete declaration.
(current_nr_blocks, current_blocks, target_bb): Remove static qualifiers.
(compute_dom_prob_ps, check_live, update_live, set_spec_fed): Delete
declaration.
(init_regions): Delete declaration.
(debug_region, bb_in_region_p, dump_region_dot_file, dump_region_dot,
rgn_estimate_number_of_insns): New.
(too_large): Use estimate_number_of_insns.
(haifa_find_rgns): New. Move the code from ...
(find_rgns): ... here. Call either sel_find_rgns or haifa_find_rgns.
(free_trg_info): New.
(compute_trg_info): Allocate candidate tables here instead of ...
(init_ready_list): ... here.
(rgn_print_insn): Use const_rtx.
(contributes_to_priority, extend_regions): Delete static declaration.
(add_remove_insn, fix_recovery_cfg): Add rgn_ to function names.
(add_block1): Rename to rgn_add_block.
(debug_rgn_dependencies): Delete static qualifier.
(new_ready): Use sched_deps_info. Simplify.
(rgn_common_sched_info, rgn_const_sched_deps_info,
rgn_const_sel_sched_deps_info, rgn_sched_deps_info, rgn_sched_info): New.
(region_sched_info): Rename to rgn_const_sched_info.
(deps_join): New, extracted from ...
(propagate_deps): ... here.
(compute_block_dependences, debug_dependencies): Update for selective
scheduling.
(free_rgn_deps, compute_priorities): New functions.
(sched_rgn_init, sched_rgn_finish, rgn_setup_region,
sched_rgn_compute_dependencies): New functions.
(schedule_region): Use them.
(sched_rgn_local_init, sched_rgn_local_free, sched_rgn_local_finish,
rgn_setup_common_sched_info, rgn_setup_sched_infos):
New functions.
(schedule_insns): Call new functions that were split out.
(rgn_make_new_region_out_of_new_block): New.
(get_rgn_sched_max_insns_priority): New.
(rest_of_handle_sched, rest_of_handle_sched2): Call selective
scheduling when appropriate.
* sched-vis.c: Include insn-attr.h.
(print_value, print_pattern): Make global.
(print_rtl_slim, debug_bb_slim, debug_bb_n_slim): New functions.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_ALLOC_SCHED_CONTEXT, TARGET_SCHED_INIT_SCHED_CONTEXT,
TARGET_SCHED_SET_SCHED_CONTEXT, TARGET_SCHED_CLEAR_SCHED_CONTEXT,
TARGET_SCHED_FREE_SCHED_CONTEXT, TARGET_SCHED_GET_INSN_CHECKED_DS,
TARGET_SCHED_GET_INSN_SPEC_DS, TARGET_SCHED_SKIP_RTX_P): New target
hooks. Initialize them to 0.
(TARGET_SCHED_GEN_CHECK): Rename to TARGET_SCHED_GEN_SPEC_CHECK.
* target.h (struct gcc_target): Add them. Rename gen_check field to
gen_spec_check.
* flags.h (sel_sched_switch_set): Declare.
* opts.c (sel_sched_switch_set): New variable.
(decode_options): Unset flag_sel_sched_pipelining_outer_loops if
pipelining is disabled from command line.
(common_handle_option): Record whether selective scheduling is
requested from command line.
* doc/invoke.texi: Document new flags and parameters.
* doc/tm.texi: Document new target hooks.
* config/ia64/ia64.c (TARGET_SCHED_GEN_SPEC_CHECK): Define to ia64_gen_check.
(dfa_state_size): Do not declare locally.
* config/ia64/ia64.opt (msched-ar-data-spec): Default to 0.
* config/rs6000/rs6000.c (rs6000_init_sched_context,
rs6000_alloc_sched_context, rs6000_set_sched_context,
rs6000_free_sched_context): New functions.
(struct _rs6000_sched_context): New.
(rs6000_sched_reorder2): Do not modify INSN_PRIORITY for selective
scheduling.
(rs6000_sched_finish): Do not run for selective scheduling.
Co-Authored-By: Alexander Monakov <amonakov@ispras.ru>
Co-Authored-By: Dmitry Melnik <dm@ispras.ru>
Co-Authored-By: Dmitry Zhurikhin <zhur@ispras.ru>
Co-Authored-By: Maxim Kuvyrkov <maxim@codesourcery.com>
From-SVN: r139854
2008-09-01 10:57:00 +02:00
DEFPARAM ( PARAM_SELSCHED_MAX_LOOKAHEAD ,
"selsched-max-lookahead" ,
2015-10-22 00:24:41 +02:00
"The maximum size of the lookahead window of selective scheduling." ,
sel-sched.h, [...]: New files.
2008-08-31 Andrey Belevantsev <abel@ispras.ru>
Dmitry Melnik <dm@ispras.ru>
Dmitry Zhurikhin <zhur@ispras.ru>
Alexander Monakov <amonakov@ispras.ru>
Maxim Kuvyrkov <maxim@codesourcery.com>
* sel-sched.h, sel-sched-dump.h, sel-sched-ir.h, sel-sched.c,
sel-sched-dump.c, sel-sched-ir.c: New files.
* Makefile.in (OBJS-common): Add selective scheduling object
files.
(sel-sched.o, sel-sched-dump.o, sel-sched-ir.o): New entries.
(SEL_SCHED_IR_H, SEL_SCHED_DUMP_H): New entries.
(sched-vis.o): Add dependency on $(INSN_ATTR_H).
* cfghooks.h (get_cfg_hooks, set_cfg_hooks): New prototypes.
* cfghooks.c (get_cfg_hooks, set_cfg_hooks): New functions.
(make_forwarder_block): Update loop latch if we have redirected
the loop latch edge.
* cfgloop.c (get_loop_body_in_custom_order): New function.
* cfgloop.h (LOOPS_HAVE_FALLTHRU_PREHEADERS): New enum field.
(CP_FALLTHRU_PREHEADERS): Likewise.
(get_loop_body_in_custom_order): Declare.
* cfgloopmanip.c (has_preds_from_loop): New.
(create_preheader): Honor CP_FALLTHRU_PREHEADERS.
Assert that the preheader edge will be fall thru when it is set.
* common.opt (fsel-sched-bookkeeping, fsel-sched-pipelining,
fsel-sched-pipelining-outer-loops, fsel-sched-renaming,
fsel-sched-substitution, fselective-scheduling): New flags.
* cse.c (hash_rtx_cb): New.
(hash_rtx): Use it.
* dbgcnt.def (sel_sched_cnt, sel_sched_region_cnt,
sel_sched_insn_cnt): New counters.
* final.c (compute_alignments): Export. Free dominance info after loop_optimizer_finalize.
* genattr.c (main): Output maximal_insn_latency prototype.
* genautomata.c (output_default_latencies): New. Factor its code from ...
(output_internal_insn_latency_func): ... here.
(output_internal_maximal_insn_latency_func): New.
(output_maximal_insn_latency_func): New.
* hard-reg-set.h (UHOST_BITS_PER_WIDE_INT): Define unconditionally.
(struct hard_reg_set_iterator): New.
(hard_reg_set_iter_init, hard_reg_set_iter_set,
hard_reg_set_iter_next): New functions.
(EXECUTE_IF_SET_IN_HARD_REG_SET): New macro.
* lists.c (remove_free_INSN_LIST_node,
remove_free_EXPR_LIST_node): New functions.
* loop-init.c (loop_optimizer_init): When LOOPS_HAVE_FALLTHRU_PREHEADERS,
set CP_FALLTHRU_PREHEADERS when calling create_preheaders.
(loop_optimizer_finalize): Do not verify flow info after reload.
* recog.c (validate_replace_rtx_1): New parameter simplify.
Default it to true. Update all uses. Factor out simplifying
code to ...
(simplify_while_replacing): ... this new function.
(validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): New.
* recog.h (validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): Declare.
* rtl.c (rtx_equal_p_cb): New.
(rtx_equal_p): Use it.
* rtl.h (rtx_equal_p_cb, hash_rtx_cb): Declare.
(remove_free_INSN_LIST_NODE, remove_free_EXPR_LIST_node,
debug_bb_n_slim, debug_bb_slim, print_rtl_slim): Likewise.
* vecprim.h: Add a vector type for unsigned int.
* haifa-sched.c: Include vecprim.h and cfgloop.h.
(issue_rate, sched_verbose_param, note_list, dfa_state_size,
ready_try, cycle_issued_insns, spec_info): Make global.
(readyp): Initialize.
(dfa_lookahead): New global variable.
(old_max_uid, old_last_basic_block): Remove.
(h_i_d): Make it a vector.
(INSN_TICK, INTER_TICK, QUEUE_INDEX, INSN_COST): Make them work
through HID macro.
(after_recovery, adding_bb_to_current_region_p):
New variables to handle correct insertion of the recovery code.
(struct ready_list): Move declaration to sched-int.h.
(rgn_n_insns): Removed.
(rtx_vec_t): Move to sched-int.h.
(find_insn_reg_weight): Remove.
(find_insn_reg_weight1): Rename to find_insn_reg_weight.
(haifa_init_h_i_d, haifa_finish_h_i_d):
New functions to initialize / finalize haifa instruction data.
(extend_h_i_d, init_h_i_d): Rewrite.
(unlink_other_notes): Move logic to add_to_note_list. Handle
selective scheduler.
(ready_lastpos, ready_element, ready_sort, reemit_notes,
find_fallthru_edge): Make global, remove static prototypes.
(max_issue): Make global. Add privileged_n and state parameters. Use
them.
(extend_global, extend_all): Removed.
(init_before_recovery): Add new param. Fix the handling of the case
when we insert a recovery code before the EXIT which has a predecessor
with a fallthrough edge to it.
(create_recovery_block): Make global. Rename to
sched_create_recovery_block. Update.
(change_pattern): Rename to sched_change_pattern. Make global.
(speculate_insn): Rename to sched_speculate_insn. Make global.
Split haifa-specific functionality into ...
(haifa_change_pattern): New static function.
(sched_extend_bb): New static function.
(sched_init_bbs): New function.
(current_sched_info): Change type to struct haifa_sched_info.
(insn_cost): Adjust for selective scheduling.
(dep_cost_1): New function. Move logic from ...
(dep_cost): ... here.
(dep_cost): Use dep_cost_1.
(contributes_to_priority_p): Use sched_deps_info instead of
current_sched_info.
(priority): Adjust to work with selective scheduling. Process the
corner case when all dependencies don't contribute to priority.
(rank_for_schedule): Use ds_weak instead of dep_weak.
(advance_state): New function. Move logic from ...
(advance_one_cycle): ... here.
(add_to_note_list, concat_note_lists): New functions.
(rm_other_notes): Make static. Adjust for selective scheduling.
(remove_notes, restore_other_notes): New functions.
(move_insn): Add two arguments. Update assert. Don't call
reemit_notes.
(choose_ready): Remove lookahead variable, use dfa_lookahead.
Remove more_issue, max_points. Move the code to initialize
max_lookahead_tries to max_issue.
(schedule_block): Remove rgn_n_insns1 parameter. Don't allocate
ready. Adjust use of move_insn. Call restore_other_notes.
(luid): Remove.
(sched_init, sched_finish): Move Haifa-specific initialization/
finalization to ...
(haifa_sched_init, haifa_sched_finish): ... respectively.
New functions.
(setup_sched_dump): New function.
(haifa_init_only_bb): New static function.
(haifa_speculate_insn): New static function.
(try_ready): Use haifa_* instead of speculate_insn and
change_pattern.
(extend_ready, extend_all): Remove.
(sched_extend_ready_list, sched_finish_ready_list): New functions.
(create_check_block_twin, add_to_speculative_block): Use
haifa_insns_init instead of extend_global. Update to use new
initialization functions. Change parameter. Factor out code from
create_check_block_twin to ...
(sched_create_recovery_edges) ... this new function.
(add_block): Remove.
(sched_scan_info): New.
(extend_bb): Use sched_scan_info.
(init_bb, extend_insn, init_insn, init_insns_in_bb, sched_scan): New
static functions for walking through scheduling region.
(sched_luids): New vector variable to replace uid_to_luid.
(luids_extend_insn): New function.
(sched_max_luid): New variable.
(luids_init_insn): New function.
(sched_init_luids, sched_finish_luids): New functions.
(insn_luid): New debug function.
(sched_extend_target): New function.
(haifa_init_insn): New static function.
(sched_init_only_bb): New hook.
(sched_split_block): New hook.
(sched_split_block_1): New function.
(sched_create_empty_bb): New hook.
(sched_create_empty_bb_1): New function.
(common_sched_info, ready): New global variables.
(current_sched_info_var): Remove.
(move_block_after_check): Use common_sched_info.
(haifa_luid_for_non_insn): New static function.
(init_before_recovery): Use haifa_init_only_bb instead of
add_block.
(increase_insn_priority): New.
* modulo-sched.c: (issue_rate): Remove static declaration.
(sms_sched_info): Change type to haifa_sched_info.
(sms_sched_deps_info, sms_common_sched_info): New variables.
(setup_sched_infos): New.
(sms_schedule): Initialize them. Call haifa_sched_init/finish.
Do not call regstat_free_calls_crossed.
(sms_print_insn): Use const_rtx.
* params.def (PARAM_MAX_PIPELINE_REGION_BLOCKS,
PARAM_MAX_PIPELINE_REGION_INSNS, PARAM_SELSCHED_MAX_LOOKAHEAD,
PARAM_SELSCHED_MAX_SCHED_TIMES, PARAM_SELSCHED_INSNS_TO_RENAME,
PARAM_SCHED_MEM_TRUE_DEP_COST): New.
* sched-deps.c (sched_deps_info): New. Update all relevant uses of
current_sched_info to use it.
(enum reg_pending_barrier_mode): Move to sched-int.h.
(h_d_i_d): New variable. Initialize to NULL.
({true, output, anti, spec, forward}_dependency_cache): Initialize
to NULL.
(estimate_dep_weak): Remove static declaration.
(sched_has_condition_p): New function. Adjust users of
sched_get_condition to use it instead.
(conditions_mutex_p): Add arguments indicating which conditions are
reversed. Use them.
(sched_get_condition_with_rev): Rename from sched_get_condition. Add
argument to indicate whether returned condition is reversed. Do not
generate new rtx when condition should be reversed; indicate it by
setting new argument instead.
(add_dependence_list_and_free): Add deps parameter.
Update all users. Do not free dependence list when
deps context is readonly.
(add_insn_mem_dependence, flush_pending_lists): Adjust for readonly
contexts.
(remove_from_dependence_list, remove_from_both_dependence_lists): New.
(remove_from_deps): New. Use the above functions.
(cur_insn, can_start_lhs_rhs_p): New static variables.
(add_or_update_back_dep_1): Initialize present_dep_type.
(haifa_start_insn, haifa_finish_insn, haifa_note_reg_set,
haifa_note_reg_clobber, haifa_note_reg_use, haifa_note_mem_dep,
haifa_note_dep): New functions implementing dependence hooks for
the Haifa scheduler.
(note_reg_use, note_reg_set, note_reg_clobber, note_mem_dep,
note_dep): New functions.
(ds_to_dt, extend_deps_reg_info, maybe_extend_reg_info_p): New
functions.
(init_deps): Initialize last_reg_pending_barrier and deps->readonly.
(free_deps): Initialize deps->reg_last.
(sched_analyze_reg, sched_analyze_1, sched_analyze_2,
sched_analyze_insn): Update to use dependency hooks infrastructure
and readonly contexts.
(deps_analyze_insn): New function. Move part of logic from ...
(sched_analyze): ... here. Also move some logic to ...
(deps_start_bb): ... here. New function.
(add_forw_dep, delete_forw_dep): Guard use of INSN_DEP_COUNT with
sel_sched_p.
(sched_deps_init): New function. Move code from ...
(init_dependency_caches): ... here. Remove.
(init_deps_data_vector): New.
(sched_deps_finish): New function. Move code from ...
(free_dependency_caches): ... here. Remove.
(init_deps_global, finish_deps_global): Adjust for use with
selective scheduling.
(get_dep_weak): Move logic to ...
(get_dep_weak_1): New function.
(ds_merge): Move logic to ...
(ds_merge_1): New static function.
(ds_full_merge, ds_max_merge, ds_get_speculation_types): New functions.
(ds_get_max_dep_weak): New function.
* sched-ebb.c (sched_n_insns): Rename to sched_rgn_n_insns.
(n_insns): Rename to rgn_n_insns.
(debug_ebb_dependencies): New function.
(init_ready_list): Use it.
(begin_schedule_ready): Use sched_init_only_bb.
(ebb_print_insn): Indicate when an insn starts a new cycle.
(contributes_to_priority, compute_jump_reg_dependencies,
add_remove_insn, fix_recovery_cfg): Add ebb_ prefix to function names.
(add_block1): Remove to ebb_add_block.
(ebb_sched_deps_info, ebb_common_sched_info): New variables.
(schedule_ebb): Initialize them. Use remove_notes instead of
rm_other_notes. Use haifa_local_init/finish.
(schedule_ebbs): Use haifa_sched_init/finish.
* sched-int.h: Include vecprim.h, remove rtl.h.
(struct ready_list): Delete declaration.
(sched_verbose_param, enum sched_pass_id_t,
bb_vec_t, insn_vec_t, rtx_vec_t): New.
(struct sched_scan_info_def): New structure.
(sched_scan_info, sched_scan, sched_init_bbs,
sched_init_luids, sched_finish_luids, sched_extend_target,
haifa_init_h_i_d, haifa_finish_h_i_d): Declare.
(struct common_sched_info_def): New.
(common_sched_info, haifa_common_sched_info,
sched_emulate_haifa_p): Declare.
(sel_sched_p): New.
(sched_luids): Declare.
(INSN_LUID, LUID_BY_UID, SET_INSN_LUID): Declare.
(sched_max_luid, insn_luid): Declare.
(note_list, remove_notes, restore_other_notes, bb_note): Declare.
(sched_insns_init, sched_insns_finish, xrecalloc, reemit_notes,
print_insn, print_pattern, print_value, haifa_classify_insn,
sel_find_rgns, sel_mark_hard_insn, dfa_state_size, advance_state,
setup_sched_dump, sched_init, sched_finish,
sel_insn_is_speculation_check): Export.
(struct ready_list): Move from haifa-sched.c.
(ready_try, ready, max_issue): Export.
(ebb_compute_jump_reg_dependencies, find_fallthru_edge,
sched_init_only_bb, sched_split_block, sched_split_block_1,
sched_create_empty_bb, sched_create_empty_bb_1,
sched_create_recovery_block, sched_create_recovery_edges): Export.
(enum reg_pending_barrier_mode): Export.
(struct deps): New fields `last_reg_pending_barrier' and `readonly'.
(deps_t): New.
(struct sched_info): Rename to haifa_sched_info. Use const_rtx for
print_insn field. Move add_block and fix_recovery_cfg to
common_sched_info_def. Move compute_jump_reg_dependencies, use_cselib ...
(struct sched_deps_info_def): ... this new structure.
(sched_deps_info): Declare.
(struct spec_info_def): Remove weakness_cutoff, add
data_weakness_cutoff and control_weakness_cutoff.
(spec_info): Declare.
(struct _haifa_deps_insn_data): Split from haifa_insn_data. Add
dep_count field.
(struct haifa_insn_data): Rename to struct _haifa_insn_data.
(haifa_insn_data_def, haifa_insn_data_t): New typedefs.
(current_sched_info): Change type to struct haifa_sched_info.
(haifa_deps_insn_data_def, haifa_deps_insn_data_t): New typedefs.
(h_d_i_d): New variable.
(HDID): New accessor macro.
(h_i_d): Change type to VEC (haifa_insn_data_def, heap) *.
(HID): New accessor macro. Rewrite h_i_d accessor macros through HID
and HDID.
(IS_SPECULATION_CHECK_P): Update for selective scheduler.
(enum SCHED_FLAGS): Update for selective scheduler.
(enum SPEC_SCHED_FLAGS): New flag SEL_SCHED_SPEC_DONT_CHECK_CONTROL.
(init_dependency_caches, free_dependency_caches): Delete declarations.
(deps_analyze_insn, remove_from_deps, get_dep_weak_1,
estimate_dep_weak, ds_full_merge, ds_max_merge, ds_weak,
ds_get_speculation_types, ds_get_max_dep_weak, sched_deps_init,
sched_deps_finish, haifa_note_reg_set, haifa_note_reg_use,
haifa_note_reg_clobber, maybe_extend_reg_info_p, deps_start_bb,
ds_to_dt): Export.
(rm_other_notes): Delete declaration.
(schedule_block): Remove one argument.
(cycle_issued_insns, issue_rate, dfa_lookahead, ready_sort,
ready_element, ready_lastpos, sched_extend_ready_list,
sched_finish_ready_list, sched_change_pattern, sched_speculate_insn,
concat_note_lists): Export.
(struct region): Move from sched-rgn.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN): Export.
(ebb_head, BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB, INSN_BB): Likewise.
(current_nr_blocks, current_blocks, target_bb): Likewise.
(dep_cost_1, sched_is_disabled_for_current_region_p, sched_rgn_init,
sched_rgn_finish, rgn_setup_region, sched_rgn_compute_dependencies,
sched_rgn_local_init, extend_regions,
rgn_make_new_region_out_of_new_block, compute_priorities,
debug_rgn_dependencies, free_rgn_deps, contributes_to_priority,
extend_rgns, deps_join rgn_setup_common_sched_info,
rgn_setup_sched_infos, debug_regions, debug_region, dump_region_dot,
dump_region_dot_file, haifa_sched_init, haifa_sched_finish): Export.
(get_rgn_sched_max_insns_priority, sel_add_to_insn_priority,
increase_insn_priority): Likewise.
* sched-rgn.c: Include sel-sched.h.
(ref_counts): New static variable. Use it ...
(INSN_REF_COUNT): ... here. Rewrite and move closer to uses.
(FED_BY_SPEC_LOAD, IS_LOAD_INSN): Rewrite to use HID accessor macro.
(sched_is_disabled_for_current_region_p): Delete static declaration.
(struct region): Move to sched-int.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
ebb_head): Define and initialize.
(RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN, debug_regions, extend_regions,
BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB): Move to
sched-int.h.
(find_single_block_region): Add new argument to indicate that EBB
regions should be constructed.
(debug_live): Delete declaration.
(current_nr_blocks, current_blocks, target_bb): Remove static qualifiers.
(compute_dom_prob_ps, check_live, update_live, set_spec_fed): Delete
declaration.
(init_regions): Delete declaration.
(debug_region, bb_in_region_p, dump_region_dot_file, dump_region_dot,
rgn_estimate_number_of_insns): New.
(too_large): Use estimate_number_of_insns.
(haifa_find_rgns): New. Move the code from ...
(find_rgns): ... here. Call either sel_find_rgns or haifa_find_rgns.
(free_trg_info): New.
(compute_trg_info): Allocate candidate tables here instead of ...
(init_ready_list): ... here.
(rgn_print_insn): Use const_rtx.
(contributes_to_priority, extend_regions): Delete static declaration.
(add_remove_insn, fix_recovery_cfg): Add rgn_ to function names.
(add_block1): Rename to rgn_add_block.
(debug_rgn_dependencies): Delete static qualifier.
(new_ready): Use sched_deps_info. Simplify.
(rgn_common_sched_info, rgn_const_sched_deps_info,
rgn_const_sel_sched_deps_info, rgn_sched_deps_info, rgn_sched_info): New.
(region_sched_info): Rename to rgn_const_sched_info.
(deps_join): New, extracted from ...
(propagate_deps): ... here.
(compute_block_dependences, debug_dependencies): Update for selective
scheduling.
(free_rgn_deps, compute_priorities): New functions.
(sched_rgn_init, sched_rgn_finish, rgn_setup_region,
sched_rgn_compute_dependencies): New functions.
(schedule_region): Use them.
(sched_rgn_local_init, sched_rgn_local_free, sched_rgn_local_finish,
rgn_setup_common_sched_info, rgn_setup_sched_infos):
New functions.
(schedule_insns): Call new functions that were split out.
(rgn_make_new_region_out_of_new_block): New.
(get_rgn_sched_max_insns_priority): New.
(rest_of_handle_sched, rest_of_handle_sched2): Call selective
scheduling when appropriate.
* sched-vis.c: Include insn-attr.h.
(print_value, print_pattern): Make global.
(print_rtl_slim, debug_bb_slim, debug_bb_n_slim): New functions.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_ALLOC_SCHED_CONTEXT, TARGET_SCHED_INIT_SCHED_CONTEXT,
TARGET_SCHED_SET_SCHED_CONTEXT, TARGET_SCHED_CLEAR_SCHED_CONTEXT,
TARGET_SCHED_FREE_SCHED_CONTEXT, TARGET_SCHED_GET_INSN_CHECKED_DS,
TARGET_SCHED_GET_INSN_SPEC_DS, TARGET_SCHED_SKIP_RTX_P): New target
hooks. Initialize them to 0.
(TARGET_SCHED_GEN_CHECK): Rename to TARGET_SCHED_GEN_SPEC_CHECK.
* target.h (struct gcc_target): Add them. Rename gen_check field to
gen_spec_check.
* flags.h (sel_sched_switch_set): Declare.
* opts.c (sel_sched_switch_set): New variable.
(decode_options): Unset flag_sel_sched_pipelining_outer_loops if
pipelining is disabled from command line.
(common_handle_option): Record whether selective scheduling is
requested from command line.
* doc/invoke.texi: Document new flags and parameters.
* doc/tm.texi: Document new target hooks.
* config/ia64/ia64.c (TARGET_SCHED_GEN_SPEC_CHECK): Define to ia64_gen_check.
(dfa_state_size): Do not declare locally.
* config/ia64/ia64.opt (msched-ar-data-spec): Default to 0.
* config/rs6000/rs6000.c (rs6000_init_sched_context,
rs6000_alloc_sched_context, rs6000_set_sched_context,
rs6000_free_sched_context): New functions.
(struct _rs6000_sched_context): New.
(rs6000_sched_reorder2): Do not modify INSN_PRIORITY for selective
scheduling.
(rs6000_sched_finish): Do not run for selective scheduling.
Co-Authored-By: Alexander Monakov <amonakov@ispras.ru>
Co-Authored-By: Dmitry Melnik <dm@ispras.ru>
Co-Authored-By: Dmitry Zhurikhin <zhur@ispras.ru>
Co-Authored-By: Maxim Kuvyrkov <maxim@codesourcery.com>
From-SVN: r139854
2008-09-01 10:57:00 +02:00
50 , 0 , 0 )
DEFPARAM ( PARAM_SELSCHED_MAX_SCHED_TIMES ,
"selsched-max-sched-times" ,
2015-10-22 00:24:41 +02:00
"Maximum number of times that an insn could be scheduled." ,
2017-02-17 15:47:08 +01:00
2 , 1 , 0 )
sel-sched.h, [...]: New files.
2008-08-31 Andrey Belevantsev <abel@ispras.ru>
Dmitry Melnik <dm@ispras.ru>
Dmitry Zhurikhin <zhur@ispras.ru>
Alexander Monakov <amonakov@ispras.ru>
Maxim Kuvyrkov <maxim@codesourcery.com>
* sel-sched.h, sel-sched-dump.h, sel-sched-ir.h, sel-sched.c,
sel-sched-dump.c, sel-sched-ir.c: New files.
* Makefile.in (OBJS-common): Add selective scheduling object
files.
(sel-sched.o, sel-sched-dump.o, sel-sched-ir.o): New entries.
(SEL_SCHED_IR_H, SEL_SCHED_DUMP_H): New entries.
(sched-vis.o): Add dependency on $(INSN_ATTR_H).
* cfghooks.h (get_cfg_hooks, set_cfg_hooks): New prototypes.
* cfghooks.c (get_cfg_hooks, set_cfg_hooks): New functions.
(make_forwarder_block): Update loop latch if we have redirected
the loop latch edge.
* cfgloop.c (get_loop_body_in_custom_order): New function.
* cfgloop.h (LOOPS_HAVE_FALLTHRU_PREHEADERS): New enum field.
(CP_FALLTHRU_PREHEADERS): Likewise.
(get_loop_body_in_custom_order): Declare.
* cfgloopmanip.c (has_preds_from_loop): New.
(create_preheader): Honor CP_FALLTHRU_PREHEADERS.
Assert that the preheader edge will be fall thru when it is set.
* common.opt (fsel-sched-bookkeeping, fsel-sched-pipelining,
fsel-sched-pipelining-outer-loops, fsel-sched-renaming,
fsel-sched-substitution, fselective-scheduling): New flags.
* cse.c (hash_rtx_cb): New.
(hash_rtx): Use it.
* dbgcnt.def (sel_sched_cnt, sel_sched_region_cnt,
sel_sched_insn_cnt): New counters.
* final.c (compute_alignments): Export. Free dominance info after loop_optimizer_finalize.
* genattr.c (main): Output maximal_insn_latency prototype.
* genautomata.c (output_default_latencies): New. Factor its code from ...
(output_internal_insn_latency_func): ... here.
(output_internal_maximal_insn_latency_func): New.
(output_maximal_insn_latency_func): New.
* hard-reg-set.h (UHOST_BITS_PER_WIDE_INT): Define unconditionally.
(struct hard_reg_set_iterator): New.
(hard_reg_set_iter_init, hard_reg_set_iter_set,
hard_reg_set_iter_next): New functions.
(EXECUTE_IF_SET_IN_HARD_REG_SET): New macro.
* lists.c (remove_free_INSN_LIST_node,
remove_free_EXPR_LIST_node): New functions.
* loop-init.c (loop_optimizer_init): When LOOPS_HAVE_FALLTHRU_PREHEADERS,
set CP_FALLTHRU_PREHEADERS when calling create_preheaders.
(loop_optimizer_finalize): Do not verify flow info after reload.
* recog.c (validate_replace_rtx_1): New parameter simplify.
Default it to true. Update all uses. Factor out simplifying
code to ...
(simplify_while_replacing): ... this new function.
(validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): New.
* recog.h (validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): Declare.
* rtl.c (rtx_equal_p_cb): New.
(rtx_equal_p): Use it.
* rtl.h (rtx_equal_p_cb, hash_rtx_cb): Declare.
(remove_free_INSN_LIST_NODE, remove_free_EXPR_LIST_node,
debug_bb_n_slim, debug_bb_slim, print_rtl_slim): Likewise.
* vecprim.h: Add a vector type for unsigned int.
* haifa-sched.c: Include vecprim.h and cfgloop.h.
(issue_rate, sched_verbose_param, note_list, dfa_state_size,
ready_try, cycle_issued_insns, spec_info): Make global.
(readyp): Initialize.
(dfa_lookahead): New global variable.
(old_max_uid, old_last_basic_block): Remove.
(h_i_d): Make it a vector.
(INSN_TICK, INTER_TICK, QUEUE_INDEX, INSN_COST): Make them work
through HID macro.
(after_recovery, adding_bb_to_current_region_p):
New variables to handle correct insertion of the recovery code.
(struct ready_list): Move declaration to sched-int.h.
(rgn_n_insns): Removed.
(rtx_vec_t): Move to sched-int.h.
(find_insn_reg_weight): Remove.
(find_insn_reg_weight1): Rename to find_insn_reg_weight.
(haifa_init_h_i_d, haifa_finish_h_i_d):
New functions to initialize / finalize haifa instruction data.
(extend_h_i_d, init_h_i_d): Rewrite.
(unlink_other_notes): Move logic to add_to_note_list. Handle
selective scheduler.
(ready_lastpos, ready_element, ready_sort, reemit_notes,
find_fallthru_edge): Make global, remove static prototypes.
(max_issue): Make global. Add privileged_n and state parameters. Use
them.
(extend_global, extend_all): Removed.
(init_before_recovery): Add new param. Fix the handling of the case
when we insert a recovery code before the EXIT which has a predecessor
with a fallthrough edge to it.
(create_recovery_block): Make global. Rename to
sched_create_recovery_block. Update.
(change_pattern): Rename to sched_change_pattern. Make global.
(speculate_insn): Rename to sched_speculate_insn. Make global.
Split haifa-specific functionality into ...
(haifa_change_pattern): New static function.
(sched_extend_bb): New static function.
(sched_init_bbs): New function.
(current_sched_info): Change type to struct haifa_sched_info.
(insn_cost): Adjust for selective scheduling.
(dep_cost_1): New function. Move logic from ...
(dep_cost): ... here.
(dep_cost): Use dep_cost_1.
(contributes_to_priority_p): Use sched_deps_info instead of
current_sched_info.
(priority): Adjust to work with selective scheduling. Process the
corner case when all dependencies don't contribute to priority.
(rank_for_schedule): Use ds_weak instead of dep_weak.
(advance_state): New function. Move logic from ...
(advance_one_cycle): ... here.
(add_to_note_list, concat_note_lists): New functions.
(rm_other_notes): Make static. Adjust for selective scheduling.
(remove_notes, restore_other_notes): New functions.
(move_insn): Add two arguments. Update assert. Don't call
reemit_notes.
(choose_ready): Remove lookahead variable, use dfa_lookahead.
Remove more_issue, max_points. Move the code to initialize
max_lookahead_tries to max_issue.
(schedule_block): Remove rgn_n_insns1 parameter. Don't allocate
ready. Adjust use of move_insn. Call restore_other_notes.
(luid): Remove.
(sched_init, sched_finish): Move Haifa-specific initialization/
finalization to ...
(haifa_sched_init, haifa_sched_finish): ... respectively.
New functions.
(setup_sched_dump): New function.
(haifa_init_only_bb): New static function.
(haifa_speculate_insn): New static function.
(try_ready): Use haifa_* instead of speculate_insn and
change_pattern.
(extend_ready, extend_all): Remove.
(sched_extend_ready_list, sched_finish_ready_list): New functions.
(create_check_block_twin, add_to_speculative_block): Use
haifa_insns_init instead of extend_global. Update to use new
initialization functions. Change parameter. Factor out code from
create_check_block_twin to ...
(sched_create_recovery_edges) ... this new function.
(add_block): Remove.
(sched_scan_info): New.
(extend_bb): Use sched_scan_info.
(init_bb, extend_insn, init_insn, init_insns_in_bb, sched_scan): New
static functions for walking through scheduling region.
(sched_luids): New vector variable to replace uid_to_luid.
(luids_extend_insn): New function.
(sched_max_luid): New variable.
(luids_init_insn): New function.
(sched_init_luids, sched_finish_luids): New functions.
(insn_luid): New debug function.
(sched_extend_target): New function.
(haifa_init_insn): New static function.
(sched_init_only_bb): New hook.
(sched_split_block): New hook.
(sched_split_block_1): New function.
(sched_create_empty_bb): New hook.
(sched_create_empty_bb_1): New function.
(common_sched_info, ready): New global variables.
(current_sched_info_var): Remove.
(move_block_after_check): Use common_sched_info.
(haifa_luid_for_non_insn): New static function.
(init_before_recovery): Use haifa_init_only_bb instead of
add_block.
(increase_insn_priority): New.
* modulo-sched.c: (issue_rate): Remove static declaration.
(sms_sched_info): Change type to haifa_sched_info.
(sms_sched_deps_info, sms_common_sched_info): New variables.
(setup_sched_infos): New.
(sms_schedule): Initialize them. Call haifa_sched_init/finish.
Do not call regstat_free_calls_crossed.
(sms_print_insn): Use const_rtx.
* params.def (PARAM_MAX_PIPELINE_REGION_BLOCKS,
PARAM_MAX_PIPELINE_REGION_INSNS, PARAM_SELSCHED_MAX_LOOKAHEAD,
PARAM_SELSCHED_MAX_SCHED_TIMES, PARAM_SELSCHED_INSNS_TO_RENAME,
PARAM_SCHED_MEM_TRUE_DEP_COST): New.
* sched-deps.c (sched_deps_info): New. Update all relevant uses of
current_sched_info to use it.
(enum reg_pending_barrier_mode): Move to sched-int.h.
(h_d_i_d): New variable. Initialize to NULL.
({true, output, anti, spec, forward}_dependency_cache): Initialize
to NULL.
(estimate_dep_weak): Remove static declaration.
(sched_has_condition_p): New function. Adjust users of
sched_get_condition to use it instead.
(conditions_mutex_p): Add arguments indicating which conditions are
reversed. Use them.
(sched_get_condition_with_rev): Rename from sched_get_condition. Add
argument to indicate whether returned condition is reversed. Do not
generate new rtx when condition should be reversed; indicate it by
setting new argument instead.
(add_dependence_list_and_free): Add deps parameter.
Update all users. Do not free dependence list when
deps context is readonly.
(add_insn_mem_dependence, flush_pending_lists): Adjust for readonly
contexts.
(remove_from_dependence_list, remove_from_both_dependence_lists): New.
(remove_from_deps): New. Use the above functions.
(cur_insn, can_start_lhs_rhs_p): New static variables.
(add_or_update_back_dep_1): Initialize present_dep_type.
(haifa_start_insn, haifa_finish_insn, haifa_note_reg_set,
haifa_note_reg_clobber, haifa_note_reg_use, haifa_note_mem_dep,
haifa_note_dep): New functions implementing dependence hooks for
the Haifa scheduler.
(note_reg_use, note_reg_set, note_reg_clobber, note_mem_dep,
note_dep): New functions.
(ds_to_dt, extend_deps_reg_info, maybe_extend_reg_info_p): New
functions.
(init_deps): Initialize last_reg_pending_barrier and deps->readonly.
(free_deps): Initialize deps->reg_last.
(sched_analyze_reg, sched_analyze_1, sched_analyze_2,
sched_analyze_insn): Update to use dependency hooks infrastructure
and readonly contexts.
(deps_analyze_insn): New function. Move part of logic from ...
(sched_analyze): ... here. Also move some logic to ...
(deps_start_bb): ... here. New function.
(add_forw_dep, delete_forw_dep): Guard use of INSN_DEP_COUNT with
sel_sched_p.
(sched_deps_init): New function. Move code from ...
(init_dependency_caches): ... here. Remove.
(init_deps_data_vector): New.
(sched_deps_finish): New function. Move code from ...
(free_dependency_caches): ... here. Remove.
(init_deps_global, finish_deps_global): Adjust for use with
selective scheduling.
(get_dep_weak): Move logic to ...
(get_dep_weak_1): New function.
(ds_merge): Move logic to ...
(ds_merge_1): New static function.
(ds_full_merge, ds_max_merge, ds_get_speculation_types): New functions.
(ds_get_max_dep_weak): New function.
* sched-ebb.c (sched_n_insns): Rename to sched_rgn_n_insns.
(n_insns): Rename to rgn_n_insns.
(debug_ebb_dependencies): New function.
(init_ready_list): Use it.
(begin_schedule_ready): Use sched_init_only_bb.
(ebb_print_insn): Indicate when an insn starts a new cycle.
(contributes_to_priority, compute_jump_reg_dependencies,
add_remove_insn, fix_recovery_cfg): Add ebb_ prefix to function names.
(add_block1): Remove to ebb_add_block.
(ebb_sched_deps_info, ebb_common_sched_info): New variables.
(schedule_ebb): Initialize them. Use remove_notes instead of
rm_other_notes. Use haifa_local_init/finish.
(schedule_ebbs): Use haifa_sched_init/finish.
* sched-int.h: Include vecprim.h, remove rtl.h.
(struct ready_list): Delete declaration.
(sched_verbose_param, enum sched_pass_id_t,
bb_vec_t, insn_vec_t, rtx_vec_t): New.
(struct sched_scan_info_def): New structure.
(sched_scan_info, sched_scan, sched_init_bbs,
sched_init_luids, sched_finish_luids, sched_extend_target,
haifa_init_h_i_d, haifa_finish_h_i_d): Declare.
(struct common_sched_info_def): New.
(common_sched_info, haifa_common_sched_info,
sched_emulate_haifa_p): Declare.
(sel_sched_p): New.
(sched_luids): Declare.
(INSN_LUID, LUID_BY_UID, SET_INSN_LUID): Declare.
(sched_max_luid, insn_luid): Declare.
(note_list, remove_notes, restore_other_notes, bb_note): Declare.
(sched_insns_init, sched_insns_finish, xrecalloc, reemit_notes,
print_insn, print_pattern, print_value, haifa_classify_insn,
sel_find_rgns, sel_mark_hard_insn, dfa_state_size, advance_state,
setup_sched_dump, sched_init, sched_finish,
sel_insn_is_speculation_check): Export.
(struct ready_list): Move from haifa-sched.c.
(ready_try, ready, max_issue): Export.
(ebb_compute_jump_reg_dependencies, find_fallthru_edge,
sched_init_only_bb, sched_split_block, sched_split_block_1,
sched_create_empty_bb, sched_create_empty_bb_1,
sched_create_recovery_block, sched_create_recovery_edges): Export.
(enum reg_pending_barrier_mode): Export.
(struct deps): New fields `last_reg_pending_barrier' and `readonly'.
(deps_t): New.
(struct sched_info): Rename to haifa_sched_info. Use const_rtx for
print_insn field. Move add_block and fix_recovery_cfg to
common_sched_info_def. Move compute_jump_reg_dependencies, use_cselib ...
(struct sched_deps_info_def): ... this new structure.
(sched_deps_info): Declare.
(struct spec_info_def): Remove weakness_cutoff, add
data_weakness_cutoff and control_weakness_cutoff.
(spec_info): Declare.
(struct _haifa_deps_insn_data): Split from haifa_insn_data. Add
dep_count field.
(struct haifa_insn_data): Rename to struct _haifa_insn_data.
(haifa_insn_data_def, haifa_insn_data_t): New typedefs.
(current_sched_info): Change type to struct haifa_sched_info.
(haifa_deps_insn_data_def, haifa_deps_insn_data_t): New typedefs.
(h_d_i_d): New variable.
(HDID): New accessor macro.
(h_i_d): Change type to VEC (haifa_insn_data_def, heap) *.
(HID): New accessor macro. Rewrite h_i_d accessor macros through HID
and HDID.
(IS_SPECULATION_CHECK_P): Update for selective scheduler.
(enum SCHED_FLAGS): Update for selective scheduler.
(enum SPEC_SCHED_FLAGS): New flag SEL_SCHED_SPEC_DONT_CHECK_CONTROL.
(init_dependency_caches, free_dependency_caches): Delete declarations.
(deps_analyze_insn, remove_from_deps, get_dep_weak_1,
estimate_dep_weak, ds_full_merge, ds_max_merge, ds_weak,
ds_get_speculation_types, ds_get_max_dep_weak, sched_deps_init,
sched_deps_finish, haifa_note_reg_set, haifa_note_reg_use,
haifa_note_reg_clobber, maybe_extend_reg_info_p, deps_start_bb,
ds_to_dt): Export.
(rm_other_notes): Delete declaration.
(schedule_block): Remove one argument.
(cycle_issued_insns, issue_rate, dfa_lookahead, ready_sort,
ready_element, ready_lastpos, sched_extend_ready_list,
sched_finish_ready_list, sched_change_pattern, sched_speculate_insn,
concat_note_lists): Export.
(struct region): Move from sched-rgn.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN): Export.
(ebb_head, BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB, INSN_BB): Likewise.
(current_nr_blocks, current_blocks, target_bb): Likewise.
(dep_cost_1, sched_is_disabled_for_current_region_p, sched_rgn_init,
sched_rgn_finish, rgn_setup_region, sched_rgn_compute_dependencies,
sched_rgn_local_init, extend_regions,
rgn_make_new_region_out_of_new_block, compute_priorities,
debug_rgn_dependencies, free_rgn_deps, contributes_to_priority,
extend_rgns, deps_join rgn_setup_common_sched_info,
rgn_setup_sched_infos, debug_regions, debug_region, dump_region_dot,
dump_region_dot_file, haifa_sched_init, haifa_sched_finish): Export.
(get_rgn_sched_max_insns_priority, sel_add_to_insn_priority,
increase_insn_priority): Likewise.
* sched-rgn.c: Include sel-sched.h.
(ref_counts): New static variable. Use it ...
(INSN_REF_COUNT): ... here. Rewrite and move closer to uses.
(FED_BY_SPEC_LOAD, IS_LOAD_INSN): Rewrite to use HID accessor macro.
(sched_is_disabled_for_current_region_p): Delete static declaration.
(struct region): Move to sched-int.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
ebb_head): Define and initialize.
(RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN, debug_regions, extend_regions,
BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB): Move to
sched-int.h.
(find_single_block_region): Add new argument to indicate that EBB
regions should be constructed.
(debug_live): Delete declaration.
(current_nr_blocks, current_blocks, target_bb): Remove static qualifiers.
(compute_dom_prob_ps, check_live, update_live, set_spec_fed): Delete
declaration.
(init_regions): Delete declaration.
(debug_region, bb_in_region_p, dump_region_dot_file, dump_region_dot,
rgn_estimate_number_of_insns): New.
(too_large): Use estimate_number_of_insns.
(haifa_find_rgns): New. Move the code from ...
(find_rgns): ... here. Call either sel_find_rgns or haifa_find_rgns.
(free_trg_info): New.
(compute_trg_info): Allocate candidate tables here instead of ...
(init_ready_list): ... here.
(rgn_print_insn): Use const_rtx.
(contributes_to_priority, extend_regions): Delete static declaration.
(add_remove_insn, fix_recovery_cfg): Add rgn_ to function names.
(add_block1): Rename to rgn_add_block.
(debug_rgn_dependencies): Delete static qualifier.
(new_ready): Use sched_deps_info. Simplify.
(rgn_common_sched_info, rgn_const_sched_deps_info,
rgn_const_sel_sched_deps_info, rgn_sched_deps_info, rgn_sched_info): New.
(region_sched_info): Rename to rgn_const_sched_info.
(deps_join): New, extracted from ...
(propagate_deps): ... here.
(compute_block_dependences, debug_dependencies): Update for selective
scheduling.
(free_rgn_deps, compute_priorities): New functions.
(sched_rgn_init, sched_rgn_finish, rgn_setup_region,
sched_rgn_compute_dependencies): New functions.
(schedule_region): Use them.
(sched_rgn_local_init, sched_rgn_local_free, sched_rgn_local_finish,
rgn_setup_common_sched_info, rgn_setup_sched_infos):
New functions.
(schedule_insns): Call new functions that were split out.
(rgn_make_new_region_out_of_new_block): New.
(get_rgn_sched_max_insns_priority): New.
(rest_of_handle_sched, rest_of_handle_sched2): Call selective
scheduling when appropriate.
* sched-vis.c: Include insn-attr.h.
(print_value, print_pattern): Make global.
(print_rtl_slim, debug_bb_slim, debug_bb_n_slim): New functions.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_ALLOC_SCHED_CONTEXT, TARGET_SCHED_INIT_SCHED_CONTEXT,
TARGET_SCHED_SET_SCHED_CONTEXT, TARGET_SCHED_CLEAR_SCHED_CONTEXT,
TARGET_SCHED_FREE_SCHED_CONTEXT, TARGET_SCHED_GET_INSN_CHECKED_DS,
TARGET_SCHED_GET_INSN_SPEC_DS, TARGET_SCHED_SKIP_RTX_P): New target
hooks. Initialize them to 0.
(TARGET_SCHED_GEN_CHECK): Rename to TARGET_SCHED_GEN_SPEC_CHECK.
* target.h (struct gcc_target): Add them. Rename gen_check field to
gen_spec_check.
* flags.h (sel_sched_switch_set): Declare.
* opts.c (sel_sched_switch_set): New variable.
(decode_options): Unset flag_sel_sched_pipelining_outer_loops if
pipelining is disabled from command line.
(common_handle_option): Record whether selective scheduling is
requested from command line.
* doc/invoke.texi: Document new flags and parameters.
* doc/tm.texi: Document new target hooks.
* config/ia64/ia64.c (TARGET_SCHED_GEN_SPEC_CHECK): Define to ia64_gen_check.
(dfa_state_size): Do not declare locally.
* config/ia64/ia64.opt (msched-ar-data-spec): Default to 0.
* config/rs6000/rs6000.c (rs6000_init_sched_context,
rs6000_alloc_sched_context, rs6000_set_sched_context,
rs6000_free_sched_context): New functions.
(struct _rs6000_sched_context): New.
(rs6000_sched_reorder2): Do not modify INSN_PRIORITY for selective
scheduling.
(rs6000_sched_finish): Do not run for selective scheduling.
Co-Authored-By: Alexander Monakov <amonakov@ispras.ru>
Co-Authored-By: Dmitry Melnik <dm@ispras.ru>
Co-Authored-By: Dmitry Zhurikhin <zhur@ispras.ru>
Co-Authored-By: Maxim Kuvyrkov <maxim@codesourcery.com>
From-SVN: r139854
2008-09-01 10:57:00 +02:00
DEFPARAM ( PARAM_SELSCHED_INSNS_TO_RENAME ,
"selsched-insns-to-rename" ,
2015-10-22 00:24:41 +02:00
"Maximum number of instructions in the ready list that are considered eligible for renaming." ,
sel-sched.h, [...]: New files.
2008-08-31 Andrey Belevantsev <abel@ispras.ru>
Dmitry Melnik <dm@ispras.ru>
Dmitry Zhurikhin <zhur@ispras.ru>
Alexander Monakov <amonakov@ispras.ru>
Maxim Kuvyrkov <maxim@codesourcery.com>
* sel-sched.h, sel-sched-dump.h, sel-sched-ir.h, sel-sched.c,
sel-sched-dump.c, sel-sched-ir.c: New files.
* Makefile.in (OBJS-common): Add selective scheduling object
files.
(sel-sched.o, sel-sched-dump.o, sel-sched-ir.o): New entries.
(SEL_SCHED_IR_H, SEL_SCHED_DUMP_H): New entries.
(sched-vis.o): Add dependency on $(INSN_ATTR_H).
* cfghooks.h (get_cfg_hooks, set_cfg_hooks): New prototypes.
* cfghooks.c (get_cfg_hooks, set_cfg_hooks): New functions.
(make_forwarder_block): Update loop latch if we have redirected
the loop latch edge.
* cfgloop.c (get_loop_body_in_custom_order): New function.
* cfgloop.h (LOOPS_HAVE_FALLTHRU_PREHEADERS): New enum field.
(CP_FALLTHRU_PREHEADERS): Likewise.
(get_loop_body_in_custom_order): Declare.
* cfgloopmanip.c (has_preds_from_loop): New.
(create_preheader): Honor CP_FALLTHRU_PREHEADERS.
Assert that the preheader edge will be fall thru when it is set.
* common.opt (fsel-sched-bookkeeping, fsel-sched-pipelining,
fsel-sched-pipelining-outer-loops, fsel-sched-renaming,
fsel-sched-substitution, fselective-scheduling): New flags.
* cse.c (hash_rtx_cb): New.
(hash_rtx): Use it.
* dbgcnt.def (sel_sched_cnt, sel_sched_region_cnt,
sel_sched_insn_cnt): New counters.
* final.c (compute_alignments): Export. Free dominance info after loop_optimizer_finalize.
* genattr.c (main): Output maximal_insn_latency prototype.
* genautomata.c (output_default_latencies): New. Factor its code from ...
(output_internal_insn_latency_func): ... here.
(output_internal_maximal_insn_latency_func): New.
(output_maximal_insn_latency_func): New.
* hard-reg-set.h (UHOST_BITS_PER_WIDE_INT): Define unconditionally.
(struct hard_reg_set_iterator): New.
(hard_reg_set_iter_init, hard_reg_set_iter_set,
hard_reg_set_iter_next): New functions.
(EXECUTE_IF_SET_IN_HARD_REG_SET): New macro.
* lists.c (remove_free_INSN_LIST_node,
remove_free_EXPR_LIST_node): New functions.
* loop-init.c (loop_optimizer_init): When LOOPS_HAVE_FALLTHRU_PREHEADERS,
set CP_FALLTHRU_PREHEADERS when calling create_preheaders.
(loop_optimizer_finalize): Do not verify flow info after reload.
* recog.c (validate_replace_rtx_1): New parameter simplify.
Default it to true. Update all uses. Factor out simplifying
code to ...
(simplify_while_replacing): ... this new function.
(validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): New.
* recog.h (validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): Declare.
* rtl.c (rtx_equal_p_cb): New.
(rtx_equal_p): Use it.
* rtl.h (rtx_equal_p_cb, hash_rtx_cb): Declare.
(remove_free_INSN_LIST_NODE, remove_free_EXPR_LIST_node,
debug_bb_n_slim, debug_bb_slim, print_rtl_slim): Likewise.
* vecprim.h: Add a vector type for unsigned int.
* haifa-sched.c: Include vecprim.h and cfgloop.h.
(issue_rate, sched_verbose_param, note_list, dfa_state_size,
ready_try, cycle_issued_insns, spec_info): Make global.
(readyp): Initialize.
(dfa_lookahead): New global variable.
(old_max_uid, old_last_basic_block): Remove.
(h_i_d): Make it a vector.
(INSN_TICK, INTER_TICK, QUEUE_INDEX, INSN_COST): Make them work
through HID macro.
(after_recovery, adding_bb_to_current_region_p):
New variables to handle correct insertion of the recovery code.
(struct ready_list): Move declaration to sched-int.h.
(rgn_n_insns): Removed.
(rtx_vec_t): Move to sched-int.h.
(find_insn_reg_weight): Remove.
(find_insn_reg_weight1): Rename to find_insn_reg_weight.
(haifa_init_h_i_d, haifa_finish_h_i_d):
New functions to initialize / finalize haifa instruction data.
(extend_h_i_d, init_h_i_d): Rewrite.
(unlink_other_notes): Move logic to add_to_note_list. Handle
selective scheduler.
(ready_lastpos, ready_element, ready_sort, reemit_notes,
find_fallthru_edge): Make global, remove static prototypes.
(max_issue): Make global. Add privileged_n and state parameters. Use
them.
(extend_global, extend_all): Removed.
(init_before_recovery): Add new param. Fix the handling of the case
when we insert a recovery code before the EXIT which has a predecessor
with a fallthrough edge to it.
(create_recovery_block): Make global. Rename to
sched_create_recovery_block. Update.
(change_pattern): Rename to sched_change_pattern. Make global.
(speculate_insn): Rename to sched_speculate_insn. Make global.
Split haifa-specific functionality into ...
(haifa_change_pattern): New static function.
(sched_extend_bb): New static function.
(sched_init_bbs): New function.
(current_sched_info): Change type to struct haifa_sched_info.
(insn_cost): Adjust for selective scheduling.
(dep_cost_1): New function. Move logic from ...
(dep_cost): ... here.
(dep_cost): Use dep_cost_1.
(contributes_to_priority_p): Use sched_deps_info instead of
current_sched_info.
(priority): Adjust to work with selective scheduling. Process the
corner case when all dependencies don't contribute to priority.
(rank_for_schedule): Use ds_weak instead of dep_weak.
(advance_state): New function. Move logic from ...
(advance_one_cycle): ... here.
(add_to_note_list, concat_note_lists): New functions.
(rm_other_notes): Make static. Adjust for selective scheduling.
(remove_notes, restore_other_notes): New functions.
(move_insn): Add two arguments. Update assert. Don't call
reemit_notes.
(choose_ready): Remove lookahead variable, use dfa_lookahead.
Remove more_issue, max_points. Move the code to initialize
max_lookahead_tries to max_issue.
(schedule_block): Remove rgn_n_insns1 parameter. Don't allocate
ready. Adjust use of move_insn. Call restore_other_notes.
(luid): Remove.
(sched_init, sched_finish): Move Haifa-specific initialization/
finalization to ...
(haifa_sched_init, haifa_sched_finish): ... respectively.
New functions.
(setup_sched_dump): New function.
(haifa_init_only_bb): New static function.
(haifa_speculate_insn): New static function.
(try_ready): Use haifa_* instead of speculate_insn and
change_pattern.
(extend_ready, extend_all): Remove.
(sched_extend_ready_list, sched_finish_ready_list): New functions.
(create_check_block_twin, add_to_speculative_block): Use
haifa_insns_init instead of extend_global. Update to use new
initialization functions. Change parameter. Factor out code from
create_check_block_twin to ...
(sched_create_recovery_edges) ... this new function.
(add_block): Remove.
(sched_scan_info): New.
(extend_bb): Use sched_scan_info.
(init_bb, extend_insn, init_insn, init_insns_in_bb, sched_scan): New
static functions for walking through scheduling region.
(sched_luids): New vector variable to replace uid_to_luid.
(luids_extend_insn): New function.
(sched_max_luid): New variable.
(luids_init_insn): New function.
(sched_init_luids, sched_finish_luids): New functions.
(insn_luid): New debug function.
(sched_extend_target): New function.
(haifa_init_insn): New static function.
(sched_init_only_bb): New hook.
(sched_split_block): New hook.
(sched_split_block_1): New function.
(sched_create_empty_bb): New hook.
(sched_create_empty_bb_1): New function.
(common_sched_info, ready): New global variables.
(current_sched_info_var): Remove.
(move_block_after_check): Use common_sched_info.
(haifa_luid_for_non_insn): New static function.
(init_before_recovery): Use haifa_init_only_bb instead of
add_block.
(increase_insn_priority): New.
* modulo-sched.c: (issue_rate): Remove static declaration.
(sms_sched_info): Change type to haifa_sched_info.
(sms_sched_deps_info, sms_common_sched_info): New variables.
(setup_sched_infos): New.
(sms_schedule): Initialize them. Call haifa_sched_init/finish.
Do not call regstat_free_calls_crossed.
(sms_print_insn): Use const_rtx.
* params.def (PARAM_MAX_PIPELINE_REGION_BLOCKS,
PARAM_MAX_PIPELINE_REGION_INSNS, PARAM_SELSCHED_MAX_LOOKAHEAD,
PARAM_SELSCHED_MAX_SCHED_TIMES, PARAM_SELSCHED_INSNS_TO_RENAME,
PARAM_SCHED_MEM_TRUE_DEP_COST): New.
* sched-deps.c (sched_deps_info): New. Update all relevant uses of
current_sched_info to use it.
(enum reg_pending_barrier_mode): Move to sched-int.h.
(h_d_i_d): New variable. Initialize to NULL.
({true, output, anti, spec, forward}_dependency_cache): Initialize
to NULL.
(estimate_dep_weak): Remove static declaration.
(sched_has_condition_p): New function. Adjust users of
sched_get_condition to use it instead.
(conditions_mutex_p): Add arguments indicating which conditions are
reversed. Use them.
(sched_get_condition_with_rev): Rename from sched_get_condition. Add
argument to indicate whether returned condition is reversed. Do not
generate new rtx when condition should be reversed; indicate it by
setting new argument instead.
(add_dependence_list_and_free): Add deps parameter.
Update all users. Do not free dependence list when
deps context is readonly.
(add_insn_mem_dependence, flush_pending_lists): Adjust for readonly
contexts.
(remove_from_dependence_list, remove_from_both_dependence_lists): New.
(remove_from_deps): New. Use the above functions.
(cur_insn, can_start_lhs_rhs_p): New static variables.
(add_or_update_back_dep_1): Initialize present_dep_type.
(haifa_start_insn, haifa_finish_insn, haifa_note_reg_set,
haifa_note_reg_clobber, haifa_note_reg_use, haifa_note_mem_dep,
haifa_note_dep): New functions implementing dependence hooks for
the Haifa scheduler.
(note_reg_use, note_reg_set, note_reg_clobber, note_mem_dep,
note_dep): New functions.
(ds_to_dt, extend_deps_reg_info, maybe_extend_reg_info_p): New
functions.
(init_deps): Initialize last_reg_pending_barrier and deps->readonly.
(free_deps): Initialize deps->reg_last.
(sched_analyze_reg, sched_analyze_1, sched_analyze_2,
sched_analyze_insn): Update to use dependency hooks infrastructure
and readonly contexts.
(deps_analyze_insn): New function. Move part of logic from ...
(sched_analyze): ... here. Also move some logic to ...
(deps_start_bb): ... here. New function.
(add_forw_dep, delete_forw_dep): Guard use of INSN_DEP_COUNT with
sel_sched_p.
(sched_deps_init): New function. Move code from ...
(init_dependency_caches): ... here. Remove.
(init_deps_data_vector): New.
(sched_deps_finish): New function. Move code from ...
(free_dependency_caches): ... here. Remove.
(init_deps_global, finish_deps_global): Adjust for use with
selective scheduling.
(get_dep_weak): Move logic to ...
(get_dep_weak_1): New function.
(ds_merge): Move logic to ...
(ds_merge_1): New static function.
(ds_full_merge, ds_max_merge, ds_get_speculation_types): New functions.
(ds_get_max_dep_weak): New function.
* sched-ebb.c (sched_n_insns): Rename to sched_rgn_n_insns.
(n_insns): Rename to rgn_n_insns.
(debug_ebb_dependencies): New function.
(init_ready_list): Use it.
(begin_schedule_ready): Use sched_init_only_bb.
(ebb_print_insn): Indicate when an insn starts a new cycle.
(contributes_to_priority, compute_jump_reg_dependencies,
add_remove_insn, fix_recovery_cfg): Add ebb_ prefix to function names.
(add_block1): Remove to ebb_add_block.
(ebb_sched_deps_info, ebb_common_sched_info): New variables.
(schedule_ebb): Initialize them. Use remove_notes instead of
rm_other_notes. Use haifa_local_init/finish.
(schedule_ebbs): Use haifa_sched_init/finish.
* sched-int.h: Include vecprim.h, remove rtl.h.
(struct ready_list): Delete declaration.
(sched_verbose_param, enum sched_pass_id_t,
bb_vec_t, insn_vec_t, rtx_vec_t): New.
(struct sched_scan_info_def): New structure.
(sched_scan_info, sched_scan, sched_init_bbs,
sched_init_luids, sched_finish_luids, sched_extend_target,
haifa_init_h_i_d, haifa_finish_h_i_d): Declare.
(struct common_sched_info_def): New.
(common_sched_info, haifa_common_sched_info,
sched_emulate_haifa_p): Declare.
(sel_sched_p): New.
(sched_luids): Declare.
(INSN_LUID, LUID_BY_UID, SET_INSN_LUID): Declare.
(sched_max_luid, insn_luid): Declare.
(note_list, remove_notes, restore_other_notes, bb_note): Declare.
(sched_insns_init, sched_insns_finish, xrecalloc, reemit_notes,
print_insn, print_pattern, print_value, haifa_classify_insn,
sel_find_rgns, sel_mark_hard_insn, dfa_state_size, advance_state,
setup_sched_dump, sched_init, sched_finish,
sel_insn_is_speculation_check): Export.
(struct ready_list): Move from haifa-sched.c.
(ready_try, ready, max_issue): Export.
(ebb_compute_jump_reg_dependencies, find_fallthru_edge,
sched_init_only_bb, sched_split_block, sched_split_block_1,
sched_create_empty_bb, sched_create_empty_bb_1,
sched_create_recovery_block, sched_create_recovery_edges): Export.
(enum reg_pending_barrier_mode): Export.
(struct deps): New fields `last_reg_pending_barrier' and `readonly'.
(deps_t): New.
(struct sched_info): Rename to haifa_sched_info. Use const_rtx for
print_insn field. Move add_block and fix_recovery_cfg to
common_sched_info_def. Move compute_jump_reg_dependencies, use_cselib ...
(struct sched_deps_info_def): ... this new structure.
(sched_deps_info): Declare.
(struct spec_info_def): Remove weakness_cutoff, add
data_weakness_cutoff and control_weakness_cutoff.
(spec_info): Declare.
(struct _haifa_deps_insn_data): Split from haifa_insn_data. Add
dep_count field.
(struct haifa_insn_data): Rename to struct _haifa_insn_data.
(haifa_insn_data_def, haifa_insn_data_t): New typedefs.
(current_sched_info): Change type to struct haifa_sched_info.
(haifa_deps_insn_data_def, haifa_deps_insn_data_t): New typedefs.
(h_d_i_d): New variable.
(HDID): New accessor macro.
(h_i_d): Change type to VEC (haifa_insn_data_def, heap) *.
(HID): New accessor macro. Rewrite h_i_d accessor macros through HID
and HDID.
(IS_SPECULATION_CHECK_P): Update for selective scheduler.
(enum SCHED_FLAGS): Update for selective scheduler.
(enum SPEC_SCHED_FLAGS): New flag SEL_SCHED_SPEC_DONT_CHECK_CONTROL.
(init_dependency_caches, free_dependency_caches): Delete declarations.
(deps_analyze_insn, remove_from_deps, get_dep_weak_1,
estimate_dep_weak, ds_full_merge, ds_max_merge, ds_weak,
ds_get_speculation_types, ds_get_max_dep_weak, sched_deps_init,
sched_deps_finish, haifa_note_reg_set, haifa_note_reg_use,
haifa_note_reg_clobber, maybe_extend_reg_info_p, deps_start_bb,
ds_to_dt): Export.
(rm_other_notes): Delete declaration.
(schedule_block): Remove one argument.
(cycle_issued_insns, issue_rate, dfa_lookahead, ready_sort,
ready_element, ready_lastpos, sched_extend_ready_list,
sched_finish_ready_list, sched_change_pattern, sched_speculate_insn,
concat_note_lists): Export.
(struct region): Move from sched-rgn.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN): Export.
(ebb_head, BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB, INSN_BB): Likewise.
(current_nr_blocks, current_blocks, target_bb): Likewise.
(dep_cost_1, sched_is_disabled_for_current_region_p, sched_rgn_init,
sched_rgn_finish, rgn_setup_region, sched_rgn_compute_dependencies,
sched_rgn_local_init, extend_regions,
rgn_make_new_region_out_of_new_block, compute_priorities,
debug_rgn_dependencies, free_rgn_deps, contributes_to_priority,
extend_rgns, deps_join rgn_setup_common_sched_info,
rgn_setup_sched_infos, debug_regions, debug_region, dump_region_dot,
dump_region_dot_file, haifa_sched_init, haifa_sched_finish): Export.
(get_rgn_sched_max_insns_priority, sel_add_to_insn_priority,
increase_insn_priority): Likewise.
* sched-rgn.c: Include sel-sched.h.
(ref_counts): New static variable. Use it ...
(INSN_REF_COUNT): ... here. Rewrite and move closer to uses.
(FED_BY_SPEC_LOAD, IS_LOAD_INSN): Rewrite to use HID accessor macro.
(sched_is_disabled_for_current_region_p): Delete static declaration.
(struct region): Move to sched-int.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
ebb_head): Define and initialize.
(RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN, debug_regions, extend_regions,
BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB): Move to
sched-int.h.
(find_single_block_region): Add new argument to indicate that EBB
regions should be constructed.
(debug_live): Delete declaration.
(current_nr_blocks, current_blocks, target_bb): Remove static qualifiers.
(compute_dom_prob_ps, check_live, update_live, set_spec_fed): Delete
declaration.
(init_regions): Delete declaration.
(debug_region, bb_in_region_p, dump_region_dot_file, dump_region_dot,
rgn_estimate_number_of_insns): New.
(too_large): Use estimate_number_of_insns.
(haifa_find_rgns): New. Move the code from ...
(find_rgns): ... here. Call either sel_find_rgns or haifa_find_rgns.
(free_trg_info): New.
(compute_trg_info): Allocate candidate tables here instead of ...
(init_ready_list): ... here.
(rgn_print_insn): Use const_rtx.
(contributes_to_priority, extend_regions): Delete static declaration.
(add_remove_insn, fix_recovery_cfg): Add rgn_ to function names.
(add_block1): Rename to rgn_add_block.
(debug_rgn_dependencies): Delete static qualifier.
(new_ready): Use sched_deps_info. Simplify.
(rgn_common_sched_info, rgn_const_sched_deps_info,
rgn_const_sel_sched_deps_info, rgn_sched_deps_info, rgn_sched_info): New.
(region_sched_info): Rename to rgn_const_sched_info.
(deps_join): New, extracted from ...
(propagate_deps): ... here.
(compute_block_dependences, debug_dependencies): Update for selective
scheduling.
(free_rgn_deps, compute_priorities): New functions.
(sched_rgn_init, sched_rgn_finish, rgn_setup_region,
sched_rgn_compute_dependencies): New functions.
(schedule_region): Use them.
(sched_rgn_local_init, sched_rgn_local_free, sched_rgn_local_finish,
rgn_setup_common_sched_info, rgn_setup_sched_infos):
New functions.
(schedule_insns): Call new functions that were split out.
(rgn_make_new_region_out_of_new_block): New.
(get_rgn_sched_max_insns_priority): New.
(rest_of_handle_sched, rest_of_handle_sched2): Call selective
scheduling when appropriate.
* sched-vis.c: Include insn-attr.h.
(print_value, print_pattern): Make global.
(print_rtl_slim, debug_bb_slim, debug_bb_n_slim): New functions.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_ALLOC_SCHED_CONTEXT, TARGET_SCHED_INIT_SCHED_CONTEXT,
TARGET_SCHED_SET_SCHED_CONTEXT, TARGET_SCHED_CLEAR_SCHED_CONTEXT,
TARGET_SCHED_FREE_SCHED_CONTEXT, TARGET_SCHED_GET_INSN_CHECKED_DS,
TARGET_SCHED_GET_INSN_SPEC_DS, TARGET_SCHED_SKIP_RTX_P): New target
hooks. Initialize them to 0.
(TARGET_SCHED_GEN_CHECK): Rename to TARGET_SCHED_GEN_SPEC_CHECK.
* target.h (struct gcc_target): Add them. Rename gen_check field to
gen_spec_check.
* flags.h (sel_sched_switch_set): Declare.
* opts.c (sel_sched_switch_set): New variable.
(decode_options): Unset flag_sel_sched_pipelining_outer_loops if
pipelining is disabled from command line.
(common_handle_option): Record whether selective scheduling is
requested from command line.
* doc/invoke.texi: Document new flags and parameters.
* doc/tm.texi: Document new target hooks.
* config/ia64/ia64.c (TARGET_SCHED_GEN_SPEC_CHECK): Define to ia64_gen_check.
(dfa_state_size): Do not declare locally.
* config/ia64/ia64.opt (msched-ar-data-spec): Default to 0.
* config/rs6000/rs6000.c (rs6000_init_sched_context,
rs6000_alloc_sched_context, rs6000_set_sched_context,
rs6000_free_sched_context): New functions.
(struct _rs6000_sched_context): New.
(rs6000_sched_reorder2): Do not modify INSN_PRIORITY for selective
scheduling.
(rs6000_sched_finish): Do not run for selective scheduling.
Co-Authored-By: Alexander Monakov <amonakov@ispras.ru>
Co-Authored-By: Dmitry Melnik <dm@ispras.ru>
Co-Authored-By: Dmitry Zhurikhin <zhur@ispras.ru>
Co-Authored-By: Maxim Kuvyrkov <maxim@codesourcery.com>
From-SVN: r139854
2008-09-01 10:57:00 +02:00
2 , 0 , 0 )
DEFPARAM ( PARAM_SCHED_MEM_TRUE_DEP_COST ,
"sched-mem-true-dep-cost" ,
2015-10-22 00:24:41 +02:00
"Minimal distance between possibly conflicting store and load." ,
sel-sched.h, [...]: New files.
2008-08-31 Andrey Belevantsev <abel@ispras.ru>
Dmitry Melnik <dm@ispras.ru>
Dmitry Zhurikhin <zhur@ispras.ru>
Alexander Monakov <amonakov@ispras.ru>
Maxim Kuvyrkov <maxim@codesourcery.com>
* sel-sched.h, sel-sched-dump.h, sel-sched-ir.h, sel-sched.c,
sel-sched-dump.c, sel-sched-ir.c: New files.
* Makefile.in (OBJS-common): Add selective scheduling object
files.
(sel-sched.o, sel-sched-dump.o, sel-sched-ir.o): New entries.
(SEL_SCHED_IR_H, SEL_SCHED_DUMP_H): New entries.
(sched-vis.o): Add dependency on $(INSN_ATTR_H).
* cfghooks.h (get_cfg_hooks, set_cfg_hooks): New prototypes.
* cfghooks.c (get_cfg_hooks, set_cfg_hooks): New functions.
(make_forwarder_block): Update loop latch if we have redirected
the loop latch edge.
* cfgloop.c (get_loop_body_in_custom_order): New function.
* cfgloop.h (LOOPS_HAVE_FALLTHRU_PREHEADERS): New enum field.
(CP_FALLTHRU_PREHEADERS): Likewise.
(get_loop_body_in_custom_order): Declare.
* cfgloopmanip.c (has_preds_from_loop): New.
(create_preheader): Honor CP_FALLTHRU_PREHEADERS.
Assert that the preheader edge will be fall thru when it is set.
* common.opt (fsel-sched-bookkeeping, fsel-sched-pipelining,
fsel-sched-pipelining-outer-loops, fsel-sched-renaming,
fsel-sched-substitution, fselective-scheduling): New flags.
* cse.c (hash_rtx_cb): New.
(hash_rtx): Use it.
* dbgcnt.def (sel_sched_cnt, sel_sched_region_cnt,
sel_sched_insn_cnt): New counters.
* final.c (compute_alignments): Export. Free dominance info after loop_optimizer_finalize.
* genattr.c (main): Output maximal_insn_latency prototype.
* genautomata.c (output_default_latencies): New. Factor its code from ...
(output_internal_insn_latency_func): ... here.
(output_internal_maximal_insn_latency_func): New.
(output_maximal_insn_latency_func): New.
* hard-reg-set.h (UHOST_BITS_PER_WIDE_INT): Define unconditionally.
(struct hard_reg_set_iterator): New.
(hard_reg_set_iter_init, hard_reg_set_iter_set,
hard_reg_set_iter_next): New functions.
(EXECUTE_IF_SET_IN_HARD_REG_SET): New macro.
* lists.c (remove_free_INSN_LIST_node,
remove_free_EXPR_LIST_node): New functions.
* loop-init.c (loop_optimizer_init): When LOOPS_HAVE_FALLTHRU_PREHEADERS,
set CP_FALLTHRU_PREHEADERS when calling create_preheaders.
(loop_optimizer_finalize): Do not verify flow info after reload.
* recog.c (validate_replace_rtx_1): New parameter simplify.
Default it to true. Update all uses. Factor out simplifying
code to ...
(simplify_while_replacing): ... this new function.
(validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): New.
* recog.h (validate_replace_rtx_part,
validate_replace_rtx_part_nosimplify): Declare.
* rtl.c (rtx_equal_p_cb): New.
(rtx_equal_p): Use it.
* rtl.h (rtx_equal_p_cb, hash_rtx_cb): Declare.
(remove_free_INSN_LIST_NODE, remove_free_EXPR_LIST_node,
debug_bb_n_slim, debug_bb_slim, print_rtl_slim): Likewise.
* vecprim.h: Add a vector type for unsigned int.
* haifa-sched.c: Include vecprim.h and cfgloop.h.
(issue_rate, sched_verbose_param, note_list, dfa_state_size,
ready_try, cycle_issued_insns, spec_info): Make global.
(readyp): Initialize.
(dfa_lookahead): New global variable.
(old_max_uid, old_last_basic_block): Remove.
(h_i_d): Make it a vector.
(INSN_TICK, INTER_TICK, QUEUE_INDEX, INSN_COST): Make them work
through HID macro.
(after_recovery, adding_bb_to_current_region_p):
New variables to handle correct insertion of the recovery code.
(struct ready_list): Move declaration to sched-int.h.
(rgn_n_insns): Removed.
(rtx_vec_t): Move to sched-int.h.
(find_insn_reg_weight): Remove.
(find_insn_reg_weight1): Rename to find_insn_reg_weight.
(haifa_init_h_i_d, haifa_finish_h_i_d):
New functions to initialize / finalize haifa instruction data.
(extend_h_i_d, init_h_i_d): Rewrite.
(unlink_other_notes): Move logic to add_to_note_list. Handle
selective scheduler.
(ready_lastpos, ready_element, ready_sort, reemit_notes,
find_fallthru_edge): Make global, remove static prototypes.
(max_issue): Make global. Add privileged_n and state parameters. Use
them.
(extend_global, extend_all): Removed.
(init_before_recovery): Add new param. Fix the handling of the case
when we insert a recovery code before the EXIT which has a predecessor
with a fallthrough edge to it.
(create_recovery_block): Make global. Rename to
sched_create_recovery_block. Update.
(change_pattern): Rename to sched_change_pattern. Make global.
(speculate_insn): Rename to sched_speculate_insn. Make global.
Split haifa-specific functionality into ...
(haifa_change_pattern): New static function.
(sched_extend_bb): New static function.
(sched_init_bbs): New function.
(current_sched_info): Change type to struct haifa_sched_info.
(insn_cost): Adjust for selective scheduling.
(dep_cost_1): New function. Move logic from ...
(dep_cost): ... here.
(dep_cost): Use dep_cost_1.
(contributes_to_priority_p): Use sched_deps_info instead of
current_sched_info.
(priority): Adjust to work with selective scheduling. Process the
corner case when all dependencies don't contribute to priority.
(rank_for_schedule): Use ds_weak instead of dep_weak.
(advance_state): New function. Move logic from ...
(advance_one_cycle): ... here.
(add_to_note_list, concat_note_lists): New functions.
(rm_other_notes): Make static. Adjust for selective scheduling.
(remove_notes, restore_other_notes): New functions.
(move_insn): Add two arguments. Update assert. Don't call
reemit_notes.
(choose_ready): Remove lookahead variable, use dfa_lookahead.
Remove more_issue, max_points. Move the code to initialize
max_lookahead_tries to max_issue.
(schedule_block): Remove rgn_n_insns1 parameter. Don't allocate
ready. Adjust use of move_insn. Call restore_other_notes.
(luid): Remove.
(sched_init, sched_finish): Move Haifa-specific initialization/
finalization to ...
(haifa_sched_init, haifa_sched_finish): ... respectively.
New functions.
(setup_sched_dump): New function.
(haifa_init_only_bb): New static function.
(haifa_speculate_insn): New static function.
(try_ready): Use haifa_* instead of speculate_insn and
change_pattern.
(extend_ready, extend_all): Remove.
(sched_extend_ready_list, sched_finish_ready_list): New functions.
(create_check_block_twin, add_to_speculative_block): Use
haifa_insns_init instead of extend_global. Update to use new
initialization functions. Change parameter. Factor out code from
create_check_block_twin to ...
(sched_create_recovery_edges) ... this new function.
(add_block): Remove.
(sched_scan_info): New.
(extend_bb): Use sched_scan_info.
(init_bb, extend_insn, init_insn, init_insns_in_bb, sched_scan): New
static functions for walking through scheduling region.
(sched_luids): New vector variable to replace uid_to_luid.
(luids_extend_insn): New function.
(sched_max_luid): New variable.
(luids_init_insn): New function.
(sched_init_luids, sched_finish_luids): New functions.
(insn_luid): New debug function.
(sched_extend_target): New function.
(haifa_init_insn): New static function.
(sched_init_only_bb): New hook.
(sched_split_block): New hook.
(sched_split_block_1): New function.
(sched_create_empty_bb): New hook.
(sched_create_empty_bb_1): New function.
(common_sched_info, ready): New global variables.
(current_sched_info_var): Remove.
(move_block_after_check): Use common_sched_info.
(haifa_luid_for_non_insn): New static function.
(init_before_recovery): Use haifa_init_only_bb instead of
add_block.
(increase_insn_priority): New.
* modulo-sched.c: (issue_rate): Remove static declaration.
(sms_sched_info): Change type to haifa_sched_info.
(sms_sched_deps_info, sms_common_sched_info): New variables.
(setup_sched_infos): New.
(sms_schedule): Initialize them. Call haifa_sched_init/finish.
Do not call regstat_free_calls_crossed.
(sms_print_insn): Use const_rtx.
* params.def (PARAM_MAX_PIPELINE_REGION_BLOCKS,
PARAM_MAX_PIPELINE_REGION_INSNS, PARAM_SELSCHED_MAX_LOOKAHEAD,
PARAM_SELSCHED_MAX_SCHED_TIMES, PARAM_SELSCHED_INSNS_TO_RENAME,
PARAM_SCHED_MEM_TRUE_DEP_COST): New.
* sched-deps.c (sched_deps_info): New. Update all relevant uses of
current_sched_info to use it.
(enum reg_pending_barrier_mode): Move to sched-int.h.
(h_d_i_d): New variable. Initialize to NULL.
({true, output, anti, spec, forward}_dependency_cache): Initialize
to NULL.
(estimate_dep_weak): Remove static declaration.
(sched_has_condition_p): New function. Adjust users of
sched_get_condition to use it instead.
(conditions_mutex_p): Add arguments indicating which conditions are
reversed. Use them.
(sched_get_condition_with_rev): Rename from sched_get_condition. Add
argument to indicate whether returned condition is reversed. Do not
generate new rtx when condition should be reversed; indicate it by
setting new argument instead.
(add_dependence_list_and_free): Add deps parameter.
Update all users. Do not free dependence list when
deps context is readonly.
(add_insn_mem_dependence, flush_pending_lists): Adjust for readonly
contexts.
(remove_from_dependence_list, remove_from_both_dependence_lists): New.
(remove_from_deps): New. Use the above functions.
(cur_insn, can_start_lhs_rhs_p): New static variables.
(add_or_update_back_dep_1): Initialize present_dep_type.
(haifa_start_insn, haifa_finish_insn, haifa_note_reg_set,
haifa_note_reg_clobber, haifa_note_reg_use, haifa_note_mem_dep,
haifa_note_dep): New functions implementing dependence hooks for
the Haifa scheduler.
(note_reg_use, note_reg_set, note_reg_clobber, note_mem_dep,
note_dep): New functions.
(ds_to_dt, extend_deps_reg_info, maybe_extend_reg_info_p): New
functions.
(init_deps): Initialize last_reg_pending_barrier and deps->readonly.
(free_deps): Initialize deps->reg_last.
(sched_analyze_reg, sched_analyze_1, sched_analyze_2,
sched_analyze_insn): Update to use dependency hooks infrastructure
and readonly contexts.
(deps_analyze_insn): New function. Move part of logic from ...
(sched_analyze): ... here. Also move some logic to ...
(deps_start_bb): ... here. New function.
(add_forw_dep, delete_forw_dep): Guard use of INSN_DEP_COUNT with
sel_sched_p.
(sched_deps_init): New function. Move code from ...
(init_dependency_caches): ... here. Remove.
(init_deps_data_vector): New.
(sched_deps_finish): New function. Move code from ...
(free_dependency_caches): ... here. Remove.
(init_deps_global, finish_deps_global): Adjust for use with
selective scheduling.
(get_dep_weak): Move logic to ...
(get_dep_weak_1): New function.
(ds_merge): Move logic to ...
(ds_merge_1): New static function.
(ds_full_merge, ds_max_merge, ds_get_speculation_types): New functions.
(ds_get_max_dep_weak): New function.
* sched-ebb.c (sched_n_insns): Rename to sched_rgn_n_insns.
(n_insns): Rename to rgn_n_insns.
(debug_ebb_dependencies): New function.
(init_ready_list): Use it.
(begin_schedule_ready): Use sched_init_only_bb.
(ebb_print_insn): Indicate when an insn starts a new cycle.
(contributes_to_priority, compute_jump_reg_dependencies,
add_remove_insn, fix_recovery_cfg): Add ebb_ prefix to function names.
(add_block1): Remove to ebb_add_block.
(ebb_sched_deps_info, ebb_common_sched_info): New variables.
(schedule_ebb): Initialize them. Use remove_notes instead of
rm_other_notes. Use haifa_local_init/finish.
(schedule_ebbs): Use haifa_sched_init/finish.
* sched-int.h: Include vecprim.h, remove rtl.h.
(struct ready_list): Delete declaration.
(sched_verbose_param, enum sched_pass_id_t,
bb_vec_t, insn_vec_t, rtx_vec_t): New.
(struct sched_scan_info_def): New structure.
(sched_scan_info, sched_scan, sched_init_bbs,
sched_init_luids, sched_finish_luids, sched_extend_target,
haifa_init_h_i_d, haifa_finish_h_i_d): Declare.
(struct common_sched_info_def): New.
(common_sched_info, haifa_common_sched_info,
sched_emulate_haifa_p): Declare.
(sel_sched_p): New.
(sched_luids): Declare.
(INSN_LUID, LUID_BY_UID, SET_INSN_LUID): Declare.
(sched_max_luid, insn_luid): Declare.
(note_list, remove_notes, restore_other_notes, bb_note): Declare.
(sched_insns_init, sched_insns_finish, xrecalloc, reemit_notes,
print_insn, print_pattern, print_value, haifa_classify_insn,
sel_find_rgns, sel_mark_hard_insn, dfa_state_size, advance_state,
setup_sched_dump, sched_init, sched_finish,
sel_insn_is_speculation_check): Export.
(struct ready_list): Move from haifa-sched.c.
(ready_try, ready, max_issue): Export.
(ebb_compute_jump_reg_dependencies, find_fallthru_edge,
sched_init_only_bb, sched_split_block, sched_split_block_1,
sched_create_empty_bb, sched_create_empty_bb_1,
sched_create_recovery_block, sched_create_recovery_edges): Export.
(enum reg_pending_barrier_mode): Export.
(struct deps): New fields `last_reg_pending_barrier' and `readonly'.
(deps_t): New.
(struct sched_info): Rename to haifa_sched_info. Use const_rtx for
print_insn field. Move add_block and fix_recovery_cfg to
common_sched_info_def. Move compute_jump_reg_dependencies, use_cselib ...
(struct sched_deps_info_def): ... this new structure.
(sched_deps_info): Declare.
(struct spec_info_def): Remove weakness_cutoff, add
data_weakness_cutoff and control_weakness_cutoff.
(spec_info): Declare.
(struct _haifa_deps_insn_data): Split from haifa_insn_data. Add
dep_count field.
(struct haifa_insn_data): Rename to struct _haifa_insn_data.
(haifa_insn_data_def, haifa_insn_data_t): New typedefs.
(current_sched_info): Change type to struct haifa_sched_info.
(haifa_deps_insn_data_def, haifa_deps_insn_data_t): New typedefs.
(h_d_i_d): New variable.
(HDID): New accessor macro.
(h_i_d): Change type to VEC (haifa_insn_data_def, heap) *.
(HID): New accessor macro. Rewrite h_i_d accessor macros through HID
and HDID.
(IS_SPECULATION_CHECK_P): Update for selective scheduler.
(enum SCHED_FLAGS): Update for selective scheduler.
(enum SPEC_SCHED_FLAGS): New flag SEL_SCHED_SPEC_DONT_CHECK_CONTROL.
(init_dependency_caches, free_dependency_caches): Delete declarations.
(deps_analyze_insn, remove_from_deps, get_dep_weak_1,
estimate_dep_weak, ds_full_merge, ds_max_merge, ds_weak,
ds_get_speculation_types, ds_get_max_dep_weak, sched_deps_init,
sched_deps_finish, haifa_note_reg_set, haifa_note_reg_use,
haifa_note_reg_clobber, maybe_extend_reg_info_p, deps_start_bb,
ds_to_dt): Export.
(rm_other_notes): Delete declaration.
(schedule_block): Remove one argument.
(cycle_issued_insns, issue_rate, dfa_lookahead, ready_sort,
ready_element, ready_lastpos, sched_extend_ready_list,
sched_finish_ready_list, sched_change_pattern, sched_speculate_insn,
concat_note_lists): Export.
(struct region): Move from sched-rgn.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN): Export.
(ebb_head, BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB, INSN_BB): Likewise.
(current_nr_blocks, current_blocks, target_bb): Likewise.
(dep_cost_1, sched_is_disabled_for_current_region_p, sched_rgn_init,
sched_rgn_finish, rgn_setup_region, sched_rgn_compute_dependencies,
sched_rgn_local_init, extend_regions,
rgn_make_new_region_out_of_new_block, compute_priorities,
debug_rgn_dependencies, free_rgn_deps, contributes_to_priority,
extend_rgns, deps_join rgn_setup_common_sched_info,
rgn_setup_sched_infos, debug_regions, debug_region, dump_region_dot,
dump_region_dot_file, haifa_sched_init, haifa_sched_finish): Export.
(get_rgn_sched_max_insns_priority, sel_add_to_insn_priority,
increase_insn_priority): Likewise.
* sched-rgn.c: Include sel-sched.h.
(ref_counts): New static variable. Use it ...
(INSN_REF_COUNT): ... here. Rewrite and move closer to uses.
(FED_BY_SPEC_LOAD, IS_LOAD_INSN): Rewrite to use HID accessor macro.
(sched_is_disabled_for_current_region_p): Delete static declaration.
(struct region): Move to sched-int.h.
(nr_regions, rgn_table, rgn_bb_table, block_to_bb, containing_rgn,
ebb_head): Define and initialize.
(RGN_NR_BLOCKS, RGN_BLOCKS, RGN_DONT_CALC_DEPS, RGN_HAS_REAL_EBB,
BLOCK_TO_BB, CONTAINING_RGN, debug_regions, extend_regions,
BB_TO_BLOCK, EBB_FIRST_BB, EBB_LAST_BB): Move to
sched-int.h.
(find_single_block_region): Add new argument to indicate that EBB
regions should be constructed.
(debug_live): Delete declaration.
(current_nr_blocks, current_blocks, target_bb): Remove static qualifiers.
(compute_dom_prob_ps, check_live, update_live, set_spec_fed): Delete
declaration.
(init_regions): Delete declaration.
(debug_region, bb_in_region_p, dump_region_dot_file, dump_region_dot,
rgn_estimate_number_of_insns): New.
(too_large): Use estimate_number_of_insns.
(haifa_find_rgns): New. Move the code from ...
(find_rgns): ... here. Call either sel_find_rgns or haifa_find_rgns.
(free_trg_info): New.
(compute_trg_info): Allocate candidate tables here instead of ...
(init_ready_list): ... here.
(rgn_print_insn): Use const_rtx.
(contributes_to_priority, extend_regions): Delete static declaration.
(add_remove_insn, fix_recovery_cfg): Add rgn_ to function names.
(add_block1): Rename to rgn_add_block.
(debug_rgn_dependencies): Delete static qualifier.
(new_ready): Use sched_deps_info. Simplify.
(rgn_common_sched_info, rgn_const_sched_deps_info,
rgn_const_sel_sched_deps_info, rgn_sched_deps_info, rgn_sched_info): New.
(region_sched_info): Rename to rgn_const_sched_info.
(deps_join): New, extracted from ...
(propagate_deps): ... here.
(compute_block_dependences, debug_dependencies): Update for selective
scheduling.
(free_rgn_deps, compute_priorities): New functions.
(sched_rgn_init, sched_rgn_finish, rgn_setup_region,
sched_rgn_compute_dependencies): New functions.
(schedule_region): Use them.
(sched_rgn_local_init, sched_rgn_local_free, sched_rgn_local_finish,
rgn_setup_common_sched_info, rgn_setup_sched_infos):
New functions.
(schedule_insns): Call new functions that were split out.
(rgn_make_new_region_out_of_new_block): New.
(get_rgn_sched_max_insns_priority): New.
(rest_of_handle_sched, rest_of_handle_sched2): Call selective
scheduling when appropriate.
* sched-vis.c: Include insn-attr.h.
(print_value, print_pattern): Make global.
(print_rtl_slim, debug_bb_slim, debug_bb_n_slim): New functions.
* target-def.h (TARGET_SCHED_ADJUST_COST_2,
TARGET_SCHED_ALLOC_SCHED_CONTEXT, TARGET_SCHED_INIT_SCHED_CONTEXT,
TARGET_SCHED_SET_SCHED_CONTEXT, TARGET_SCHED_CLEAR_SCHED_CONTEXT,
TARGET_SCHED_FREE_SCHED_CONTEXT, TARGET_SCHED_GET_INSN_CHECKED_DS,
TARGET_SCHED_GET_INSN_SPEC_DS, TARGET_SCHED_SKIP_RTX_P): New target
hooks. Initialize them to 0.
(TARGET_SCHED_GEN_CHECK): Rename to TARGET_SCHED_GEN_SPEC_CHECK.
* target.h (struct gcc_target): Add them. Rename gen_check field to
gen_spec_check.
* flags.h (sel_sched_switch_set): Declare.
* opts.c (sel_sched_switch_set): New variable.
(decode_options): Unset flag_sel_sched_pipelining_outer_loops if
pipelining is disabled from command line.
(common_handle_option): Record whether selective scheduling is
requested from command line.
* doc/invoke.texi: Document new flags and parameters.
* doc/tm.texi: Document new target hooks.
* config/ia64/ia64.c (TARGET_SCHED_GEN_SPEC_CHECK): Define to ia64_gen_check.
(dfa_state_size): Do not declare locally.
* config/ia64/ia64.opt (msched-ar-data-spec): Default to 0.
* config/rs6000/rs6000.c (rs6000_init_sched_context,
rs6000_alloc_sched_context, rs6000_set_sched_context,
rs6000_free_sched_context): New functions.
(struct _rs6000_sched_context): New.
(rs6000_sched_reorder2): Do not modify INSN_PRIORITY for selective
scheduling.
(rs6000_sched_finish): Do not run for selective scheduling.
Co-Authored-By: Alexander Monakov <amonakov@ispras.ru>
Co-Authored-By: Dmitry Melnik <dm@ispras.ru>
Co-Authored-By: Dmitry Zhurikhin <zhur@ispras.ru>
Co-Authored-By: Maxim Kuvyrkov <maxim@codesourcery.com>
From-SVN: r139854
2008-09-01 10:57:00 +02:00
1 , 0 , 0 )
Model cache auto-prefetcher in scheduler
* config/arm/arm-protos.h (struct tune_params): New field
sched_autopref_queue_depth.
* config/arm/arm.c (sched-int.h): Include header.
(arm_first_cycle_multipass_dfa_lookahead_guard,)
(TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD): Define hook.
(arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune,)
(arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune,)
(arm_cortex_a8_tune, arm_cortex_a7_tune, arm_cortex_a15_tune,)
(arm_cortex_a53_tune, arm_cortex_a57_tune, arm_xgene1_tune,)
(arm_cortex_a5_tune, arm_cortex_a9_tune, arm_cortex_a12_tune,)
(arm_v7m_tune, arm_cortex_m7_tune, arm_v6m_tune, arm_fa726te_tune):
Specify sched_autopref_queue_depth value. Enabled for A15 and A57.
* config/arm/t-arm (arm.o): Update.
* haifa-sched.c (update_insn_after_change): Update.
(rank_for_schedule): Use auto-prefetcher model, if requested.
(autopref_multipass_init): New static function.
(autopref_rank_for_schedule): New rank_for_schedule heuristic.
(autopref_multipass_dfa_lookahead_guard_started_dump_p): New static
variable for debug dumps.
(autopref_multipass_dfa_lookahead_guard_1): New static helper function.
(autopref_multipass_dfa_lookahead_guard): New global function that
implements TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD hook.
(init_h_i_d): Update.
* params.def (PARAM_SCHED_AUTOPREF_QUEUE_DEPTH): New tuning knob.
* sched-int.h (enum autopref_multipass_data_status): New const enum.
(autopref_multipass_data_): Structure for auto-prefetcher data.
(autopref_multipass_data_def, autopref_multipass_data_t): New typedefs.
(struct _haifa_insn_data:autopref_multipass_data): New field.
(INSN_AUTOPREF_MULTIPASS_DATA): New access macro.
(autopref_multipass_dfa_lookahead_guard): Declare.
From-SVN: r219789
2015-01-17 02:06:43 +01:00
DEFPARAM ( PARAM_SCHED_AUTOPREF_QUEUE_DEPTH ,
"sched-autopref-queue-depth" ,
"Hardware autoprefetcher scheduler model control flag. Number of lookahead cycles the model looks into; at '0' only enable instruction sorting heuristic. Disabled by default." ,
- 1 , 0 , 0 )
2005-01-19 10:31:16 +01:00
DEFPARAM ( PARAM_MAX_LAST_VALUE_RTL ,
"max-last-value-rtl" ,
2015-10-22 00:24:41 +02:00
"The maximum number of RTL nodes that can be recorded as combiner's last value." ,
2005-01-19 10:31:16 +01:00
10000 , 0 , 0 )
2014-07-23 11:10:37 +02:00
DEFPARAM ( PARAM_MAX_COMBINE_INSNS ,
"max-combine-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of insns combine tries to combine." ,
2014-07-23 11:10:37 +02:00
4 , 2 , 4 )
2004-08-19 12:36:07 +02:00
/ * INTEGER_CST nodes are shared for values [ { - 1 , 0 } .. N ) for
{ signed , unsigned } integral types . This determines N .
2012-08-17 09:53:05 +02:00
Experimentation shows 251 to be a good value that generates the
least amount of garbage for allocating the TREE_VEC storage . * /
2004-08-19 12:36:07 +02:00
DEFPARAM ( PARAM_INTEGER_SHARE_LIMIT ,
"integer-share-limit" ,
2015-10-22 00:24:41 +02:00
"The upper bound for sharing integer constants." ,
2012-08-17 09:53:05 +02:00
251 , 2 , 2 )
2004-08-19 12:36:07 +02:00
c-cppbuiltin.c (c_cpp_builtins): Add __SSP_ALL__ and __SSP__.
* c-cppbuiltin.c (c_cpp_builtins): Add __SSP_ALL__ and __SSP__.
* cfgexpand.c: Include params.h.
(has_protected_decls, has_short_buffer): New.
(expand_stack_vars): Take a predicate to determine what to expand.
(defer_stack_allocation): True when flag_stack_protect on.
(SPCT_HAS_LARGE_CHAR_ARRAY, SPCT_HAS_SMALL_CHAR_ARRAY): New.
(SPCT_HAS_ARRAY, SPCT_HAS_AGGREGATE): New.
(stack_protect_classify_type, stack_protect_decl_phase): New.
(stack_protect_decl_phase_1, stack_protect_decl_phase_2): New.
(add_stack_protection_conflicts, create_stack_guard): New.
(expand_used_vars): Add stack protection logic.
(tree_expand_cfg): Likewise.
* common.opt (Wstack-protector): New.
(fstack-protector, fstack-protector-all): New.
* function.c: Include predict.h.
(assign_parm_adjust_stack_rtl): Zap stack_parm when stack protect
wants to copy the parameter into the stack frame.
(stack_protect_prologue, stack_protect_epilogue): New.
(expand_function_end): Call stack_protect_epilogue. Do
sjlj_emit_function_exit_after after naked_return_label.
* function.h (struct function): Add stack_protect_guard.
* params.def (PARAM_SSP_BUFFER_SIZE): New.
* toplev.c (process_options): Disable flag_stack_protect and/or
warn_stack_protect based on FRAME_GROWS_DOWNWARD.
* tree.h (stack_protect_prologue): Declare.
* target-def.h (TARGET_STACK_PROTECT_GUARD): New.
(TARGET_STACK_PROTECT_FAIL): New.
(TARGET_INITIALIZER): Add them.
* target.h (struct gcc_target): Add stack_protect_guard and
stack_protect_fail.
* targhooks.c: Include ggc.h, gty header.
(stack_chk_guard_decl, default_stack_protect_guard): New.
(stack_chk_fail_decl, default_external_stack_protect_fail): New.
(default_hidden_stack_protect_fail): New.
* targhooks.h (default_stack_protect_guard): Declare.
(default_external_stack_protect_fail): Declare.
(default_hidden_stack_protect_fail): Declare.
* config/i386/i386.c (TARGET_STACK_PROTECT_FAIL): New.
* config/i386/i386.md (UNSPEC_SP_SET, UNSPEC_SP_TEST): New.
(trap): Use ud2.
(conditional_trap, conditional_trap_1): Remove.
(stack_protect_set, stack_protect_set_si, stack_protect_set_di): New.
(stack_protect_test, stack_protect_test_si, stack_protect_test_di): New.
* doc/md.texi (stack_protect_set, stack_protect_test): New.
* doc/tm.texi (TARGET_STACK_PROTECT_GUARD): New.
(TARGET_STACK_PROTECT_FAIL): New.
* libgcc-std.ver (GCC_4.1.0): New.
* libgcc.h (__stack_chk_guard): Declare.
(__stack_chk_fail, __stack_chk_fail_local): Declare.
* libgcc2.c (L_stack_chk, L_stack_chk_local): New.
* mklibgcc.in (lib2funcs): Add them.
From-SVN: r101348
2005-06-27 09:41:16 +02:00
DEFPARAM ( PARAM_SSP_BUFFER_SIZE ,
"ssp-buffer-size" ,
2015-10-22 00:24:41 +02:00
"The lower bound for a buffer to be considered for stack smashing protection." ,
c-cppbuiltin.c (c_cpp_builtins): Add __SSP_ALL__ and __SSP__.
* c-cppbuiltin.c (c_cpp_builtins): Add __SSP_ALL__ and __SSP__.
* cfgexpand.c: Include params.h.
(has_protected_decls, has_short_buffer): New.
(expand_stack_vars): Take a predicate to determine what to expand.
(defer_stack_allocation): True when flag_stack_protect on.
(SPCT_HAS_LARGE_CHAR_ARRAY, SPCT_HAS_SMALL_CHAR_ARRAY): New.
(SPCT_HAS_ARRAY, SPCT_HAS_AGGREGATE): New.
(stack_protect_classify_type, stack_protect_decl_phase): New.
(stack_protect_decl_phase_1, stack_protect_decl_phase_2): New.
(add_stack_protection_conflicts, create_stack_guard): New.
(expand_used_vars): Add stack protection logic.
(tree_expand_cfg): Likewise.
* common.opt (Wstack-protector): New.
(fstack-protector, fstack-protector-all): New.
* function.c: Include predict.h.
(assign_parm_adjust_stack_rtl): Zap stack_parm when stack protect
wants to copy the parameter into the stack frame.
(stack_protect_prologue, stack_protect_epilogue): New.
(expand_function_end): Call stack_protect_epilogue. Do
sjlj_emit_function_exit_after after naked_return_label.
* function.h (struct function): Add stack_protect_guard.
* params.def (PARAM_SSP_BUFFER_SIZE): New.
* toplev.c (process_options): Disable flag_stack_protect and/or
warn_stack_protect based on FRAME_GROWS_DOWNWARD.
* tree.h (stack_protect_prologue): Declare.
* target-def.h (TARGET_STACK_PROTECT_GUARD): New.
(TARGET_STACK_PROTECT_FAIL): New.
(TARGET_INITIALIZER): Add them.
* target.h (struct gcc_target): Add stack_protect_guard and
stack_protect_fail.
* targhooks.c: Include ggc.h, gty header.
(stack_chk_guard_decl, default_stack_protect_guard): New.
(stack_chk_fail_decl, default_external_stack_protect_fail): New.
(default_hidden_stack_protect_fail): New.
* targhooks.h (default_stack_protect_guard): Declare.
(default_external_stack_protect_fail): Declare.
(default_hidden_stack_protect_fail): Declare.
* config/i386/i386.c (TARGET_STACK_PROTECT_FAIL): New.
* config/i386/i386.md (UNSPEC_SP_SET, UNSPEC_SP_TEST): New.
(trap): Use ud2.
(conditional_trap, conditional_trap_1): Remove.
(stack_protect_set, stack_protect_set_si, stack_protect_set_di): New.
(stack_protect_test, stack_protect_test_si, stack_protect_test_di): New.
* doc/md.texi (stack_protect_set, stack_protect_test): New.
* doc/tm.texi (TARGET_STACK_PROTECT_GUARD): New.
(TARGET_STACK_PROTECT_FAIL): New.
* libgcc-std.ver (GCC_4.1.0): New.
* libgcc.h (__stack_chk_guard): Declare.
(__stack_chk_fail, __stack_chk_fail_local): Declare.
* libgcc2.c (L_stack_chk, L_stack_chk_local): New.
* mklibgcc.in (lib2funcs): Add them.
From-SVN: r101348
2005-06-27 09:41:16 +02:00
8 , 1 , 0 )
2013-10-09 23:07:54 +02:00
DEFPARAM ( PARAM_MIN_SIZE_FOR_STACK_SHARING ,
"min-size-for-stack-sharing" ,
"The minimum size of variables taking part in stack slot sharing "
2015-10-22 00:24:41 +02:00
"when not optimizing." ,
2013-10-09 23:07:54 +02:00
32 , 0 , 0 )
2005-11-04 21:09:25 +01:00
/ * When we thread through a block we have to make copies of the
statements within the block . Clearly for large blocks the code
duplication is bad .
PARAM_MAX_JUMP_THREAD_DUPLICATION_STMTS specifies the maximum number
of statements and PHI nodes allowed in a block which is going to
be duplicated for thread jumping purposes .
Some simple analysis showed that more than 99 % of the jump
threading opportunities are for blocks with less than 15
statements . So we can get the benefits of jump threading
without excessive code bloat for pathological cases with the
throttle set at 15 statements . * /
DEFPARAM ( PARAM_MAX_JUMP_THREAD_DUPLICATION_STMTS ,
"max-jump-thread-duplication-stmts" ,
2015-10-22 00:24:41 +02:00
"Maximum number of statements allowed in a block that needs to be duplicated when threading jumps." ,
2005-11-04 21:09:25 +01:00
15 , 0 , 0 )
2006-02-14 15:49:13 +01:00
/ * This is the maximum number of fields a variable may have before the pointer analysis machinery
2009-11-25 11:55:54 +01:00
will stop trying to treat it in a field - sensitive manner .
2006-02-14 15:49:13 +01:00
There are programs out there with thousands of fields per structure , and handling them
field - sensitively is not worth the cost . * /
DEFPARAM ( PARAM_MAX_FIELDS_FOR_FIELD_SENSITIVE ,
"max-fields-for-field-sensitive" ,
2015-10-22 00:24:41 +02:00
"Maximum number of fields in a structure before pointer analysis treats the structure as a single variable." ,
2008-04-15 17:41:52 +02:00
0 , 0 , 0 )
2006-04-14 02:07:48 +02:00
DEFPARAM ( PARAM_MAX_SCHED_READY_INSNS ,
"max-sched-ready-insns" ,
2015-10-22 00:24:41 +02:00
"The maximum number of instructions ready to be issued to be considered by the scheduler during the first scheduling pass." ,
2006-04-14 02:07:48 +02:00
100 , 0 , 0 )
2011-03-17 13:36:04 +01:00
/* This is the maximum number of active local stores RTL DSE will consider. */
DEFPARAM ( PARAM_MAX_DSE_ACTIVE_LOCAL_STORES ,
"max-dse-active-local-stores" ,
2015-10-22 00:24:41 +02:00
"Maximum number of active local stores in RTL dead store elimination." ,
2011-03-17 13:36:04 +01:00
5000 , 0 , 0 )
params.c (set_param_value): Initialize the "set" field.
* params.c (set_param_value): Initialize the "set" field.
* params.h (struct param_info): Add "set" field.
(PARAM_SET_P): New macro.
(PREFETCH_LATENCY, SIMULTANEOUS_PREFETCHES, L1_CACHE_SIZE,
L1_CACHE_LINE_SIZE): New macros.
* toplev.c (DEFPARAM): Initialize the "set" field.
* tree-ssa-loop-prefetch.c (PREFETCH_LATENCY,
SIMULTANEOUS_PREFETCHES): Removed.
(PREFETCH_BLOCK): Use L1_CACHE_LINE_SIZE.
(tree_ssa_prefetch_arrays): Dump the values of the parameters.
* config/sparc/sparc.c: Include params.h.
(sparc_override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sparc/sparc.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.c: Include params.h.
(k8_cost): Change default value for SIMULTANEOUS_PREFETCHES.
(override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sh/sh.h (SIMULTANEOUS_PREFETCHES): Removed.
(OPTIMIZATION_OPTIONS): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.c (ia64_optimization_options): Set
SIMULTANEOUS_PREFETCHES and L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.h (SIMULTANEOUS_PREFETCHES, PREFETCH_BLOCK):
Removed.
* params.def (PARAM_PREFETCH_LATENCY, PARAM_SIMULTANEOUS_PREFETCHES,
PARAM_L1_CACHE_SIZE, PARAM_L1_CACHE_LINE_SIZE): New params.
* doc/invoke.texi: Document new params.
From-SVN: r118728
2006-11-12 20:17:02 +01:00
/ * Prefetching and cache - optimizations related parameters . Default values are
usually set by machine description . * /
/* The number of insns executed before prefetch is completed. */
DEFPARAM ( PARAM_PREFETCH_LATENCY ,
"prefetch-latency" ,
2015-10-22 00:24:41 +02:00
"The number of insns executed before prefetch is completed." ,
params.c (set_param_value): Initialize the "set" field.
* params.c (set_param_value): Initialize the "set" field.
* params.h (struct param_info): Add "set" field.
(PARAM_SET_P): New macro.
(PREFETCH_LATENCY, SIMULTANEOUS_PREFETCHES, L1_CACHE_SIZE,
L1_CACHE_LINE_SIZE): New macros.
* toplev.c (DEFPARAM): Initialize the "set" field.
* tree-ssa-loop-prefetch.c (PREFETCH_LATENCY,
SIMULTANEOUS_PREFETCHES): Removed.
(PREFETCH_BLOCK): Use L1_CACHE_LINE_SIZE.
(tree_ssa_prefetch_arrays): Dump the values of the parameters.
* config/sparc/sparc.c: Include params.h.
(sparc_override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sparc/sparc.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.c: Include params.h.
(k8_cost): Change default value for SIMULTANEOUS_PREFETCHES.
(override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sh/sh.h (SIMULTANEOUS_PREFETCHES): Removed.
(OPTIMIZATION_OPTIONS): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.c (ia64_optimization_options): Set
SIMULTANEOUS_PREFETCHES and L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.h (SIMULTANEOUS_PREFETCHES, PREFETCH_BLOCK):
Removed.
* params.def (PARAM_PREFETCH_LATENCY, PARAM_SIMULTANEOUS_PREFETCHES,
PARAM_L1_CACHE_SIZE, PARAM_L1_CACHE_LINE_SIZE): New params.
* doc/invoke.texi: Document new params.
From-SVN: r118728
2006-11-12 20:17:02 +01:00
200 , 0 , 0 )
/* The number of prefetches that can run at the same time. */
DEFPARAM ( PARAM_SIMULTANEOUS_PREFETCHES ,
"simultaneous-prefetches" ,
2015-10-22 00:24:41 +02:00
"The number of prefetches that can run at the same time." ,
params.c (set_param_value): Initialize the "set" field.
* params.c (set_param_value): Initialize the "set" field.
* params.h (struct param_info): Add "set" field.
(PARAM_SET_P): New macro.
(PREFETCH_LATENCY, SIMULTANEOUS_PREFETCHES, L1_CACHE_SIZE,
L1_CACHE_LINE_SIZE): New macros.
* toplev.c (DEFPARAM): Initialize the "set" field.
* tree-ssa-loop-prefetch.c (PREFETCH_LATENCY,
SIMULTANEOUS_PREFETCHES): Removed.
(PREFETCH_BLOCK): Use L1_CACHE_LINE_SIZE.
(tree_ssa_prefetch_arrays): Dump the values of the parameters.
* config/sparc/sparc.c: Include params.h.
(sparc_override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sparc/sparc.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.c: Include params.h.
(k8_cost): Change default value for SIMULTANEOUS_PREFETCHES.
(override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sh/sh.h (SIMULTANEOUS_PREFETCHES): Removed.
(OPTIMIZATION_OPTIONS): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.c (ia64_optimization_options): Set
SIMULTANEOUS_PREFETCHES and L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.h (SIMULTANEOUS_PREFETCHES, PREFETCH_BLOCK):
Removed.
* params.def (PARAM_PREFETCH_LATENCY, PARAM_SIMULTANEOUS_PREFETCHES,
PARAM_L1_CACHE_SIZE, PARAM_L1_CACHE_LINE_SIZE): New params.
* doc/invoke.texi: Document new params.
From-SVN: r118728
2006-11-12 20:17:02 +01:00
3 , 0 , 0 )
invoke.texi (l1-cache-size): Update documentation.
* doc/invoke.texi (l1-cache-size): Update documentation.
(l2-cache-size): Document.
* params.h (L2_CACHE_SIZE): New macro.
* tree-ssa-loop-prefetch.c (L1_CACHE_SIZE_BYTES): Reflect
that L1_CACHE_SIZE is in kB now.
(L2_CACHE_SIZE_BYTES): New macro.
(tree_ssa_prefetch_arrays): Show size in kB.
* config/i386/i386.h (struct processor_costs): Add l1_cache_size
and l2_cache_size fields.
* config/i386/driver-i386.c (describe_cache): Detect cache size
in kB.
* config/i386/i386.c (size_cost, i386_cost, i486_cost,pentium_cost,
pentiumpro_cost, geode_cost, k6_cost, athlon_cost, k8_cost,
amdfam10_cost, pentium4_cost, nocona_cost, core2_cost,
generic64_cost, generic32_cost): Add l1_cache_size and l2_cache_size.
(override_options): Set l1-cache-size and l2-cache-size to default
values if not specified otherwise.
* params.def (PARAM_L1_CACHE_SIZE): Change to set in kB.
(PARAM_L2_CACHE_SIZE): New.
From-SVN: r127117
2007-08-01 12:57:03 +02:00
/* The size of L1 cache in kB. */
params.c (set_param_value): Initialize the "set" field.
* params.c (set_param_value): Initialize the "set" field.
* params.h (struct param_info): Add "set" field.
(PARAM_SET_P): New macro.
(PREFETCH_LATENCY, SIMULTANEOUS_PREFETCHES, L1_CACHE_SIZE,
L1_CACHE_LINE_SIZE): New macros.
* toplev.c (DEFPARAM): Initialize the "set" field.
* tree-ssa-loop-prefetch.c (PREFETCH_LATENCY,
SIMULTANEOUS_PREFETCHES): Removed.
(PREFETCH_BLOCK): Use L1_CACHE_LINE_SIZE.
(tree_ssa_prefetch_arrays): Dump the values of the parameters.
* config/sparc/sparc.c: Include params.h.
(sparc_override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sparc/sparc.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.c: Include params.h.
(k8_cost): Change default value for SIMULTANEOUS_PREFETCHES.
(override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sh/sh.h (SIMULTANEOUS_PREFETCHES): Removed.
(OPTIMIZATION_OPTIONS): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.c (ia64_optimization_options): Set
SIMULTANEOUS_PREFETCHES and L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.h (SIMULTANEOUS_PREFETCHES, PREFETCH_BLOCK):
Removed.
* params.def (PARAM_PREFETCH_LATENCY, PARAM_SIMULTANEOUS_PREFETCHES,
PARAM_L1_CACHE_SIZE, PARAM_L1_CACHE_LINE_SIZE): New params.
* doc/invoke.texi: Document new params.
From-SVN: r118728
2006-11-12 20:17:02 +01:00
DEFPARAM ( PARAM_L1_CACHE_SIZE ,
"l1-cache-size" ,
2015-10-22 00:24:41 +02:00
"The size of L1 cache." ,
invoke.texi (l1-cache-size): Update documentation.
* doc/invoke.texi (l1-cache-size): Update documentation.
(l2-cache-size): Document.
* params.h (L2_CACHE_SIZE): New macro.
* tree-ssa-loop-prefetch.c (L1_CACHE_SIZE_BYTES): Reflect
that L1_CACHE_SIZE is in kB now.
(L2_CACHE_SIZE_BYTES): New macro.
(tree_ssa_prefetch_arrays): Show size in kB.
* config/i386/i386.h (struct processor_costs): Add l1_cache_size
and l2_cache_size fields.
* config/i386/driver-i386.c (describe_cache): Detect cache size
in kB.
* config/i386/i386.c (size_cost, i386_cost, i486_cost,pentium_cost,
pentiumpro_cost, geode_cost, k6_cost, athlon_cost, k8_cost,
amdfam10_cost, pentium4_cost, nocona_cost, core2_cost,
generic64_cost, generic32_cost): Add l1_cache_size and l2_cache_size.
(override_options): Set l1-cache-size and l2-cache-size to default
values if not specified otherwise.
* params.def (PARAM_L1_CACHE_SIZE): Change to set in kB.
(PARAM_L2_CACHE_SIZE): New.
From-SVN: r127117
2007-08-01 12:57:03 +02:00
64 , 0 , 0 )
params.c (set_param_value): Initialize the "set" field.
* params.c (set_param_value): Initialize the "set" field.
* params.h (struct param_info): Add "set" field.
(PARAM_SET_P): New macro.
(PREFETCH_LATENCY, SIMULTANEOUS_PREFETCHES, L1_CACHE_SIZE,
L1_CACHE_LINE_SIZE): New macros.
* toplev.c (DEFPARAM): Initialize the "set" field.
* tree-ssa-loop-prefetch.c (PREFETCH_LATENCY,
SIMULTANEOUS_PREFETCHES): Removed.
(PREFETCH_BLOCK): Use L1_CACHE_LINE_SIZE.
(tree_ssa_prefetch_arrays): Dump the values of the parameters.
* config/sparc/sparc.c: Include params.h.
(sparc_override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sparc/sparc.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.c: Include params.h.
(k8_cost): Change default value for SIMULTANEOUS_PREFETCHES.
(override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sh/sh.h (SIMULTANEOUS_PREFETCHES): Removed.
(OPTIMIZATION_OPTIONS): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.c (ia64_optimization_options): Set
SIMULTANEOUS_PREFETCHES and L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.h (SIMULTANEOUS_PREFETCHES, PREFETCH_BLOCK):
Removed.
* params.def (PARAM_PREFETCH_LATENCY, PARAM_SIMULTANEOUS_PREFETCHES,
PARAM_L1_CACHE_SIZE, PARAM_L1_CACHE_LINE_SIZE): New params.
* doc/invoke.texi: Document new params.
From-SVN: r118728
2006-11-12 20:17:02 +01:00
/* The size of L1 cache line in bytes. */
DEFPARAM ( PARAM_L1_CACHE_LINE_SIZE ,
"l1-cache-line-size" ,
2015-10-22 00:24:41 +02:00
"The size of L1 cache line." ,
params.c (set_param_value): Initialize the "set" field.
* params.c (set_param_value): Initialize the "set" field.
* params.h (struct param_info): Add "set" field.
(PARAM_SET_P): New macro.
(PREFETCH_LATENCY, SIMULTANEOUS_PREFETCHES, L1_CACHE_SIZE,
L1_CACHE_LINE_SIZE): New macros.
* toplev.c (DEFPARAM): Initialize the "set" field.
* tree-ssa-loop-prefetch.c (PREFETCH_LATENCY,
SIMULTANEOUS_PREFETCHES): Removed.
(PREFETCH_BLOCK): Use L1_CACHE_LINE_SIZE.
(tree_ssa_prefetch_arrays): Dump the values of the parameters.
* config/sparc/sparc.c: Include params.h.
(sparc_override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sparc/sparc.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.h (PREFETCH_BLOCK, SIMULTANEOUS_PREFETCHES):
Removed.
* config/i386/i386.c: Include params.h.
(k8_cost): Change default value for SIMULTANEOUS_PREFETCHES.
(override_options): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/sh/sh.h (SIMULTANEOUS_PREFETCHES): Removed.
(OPTIMIZATION_OPTIONS): Set SIMULTANEOUS_PREFETCHES and
L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.c (ia64_optimization_options): Set
SIMULTANEOUS_PREFETCHES and L1_CACHE_LINE_SIZE parameters.
* config/ia64/ia64.h (SIMULTANEOUS_PREFETCHES, PREFETCH_BLOCK):
Removed.
* params.def (PARAM_PREFETCH_LATENCY, PARAM_SIMULTANEOUS_PREFETCHES,
PARAM_L1_CACHE_SIZE, PARAM_L1_CACHE_LINE_SIZE): New params.
* doc/invoke.texi: Document new params.
From-SVN: r118728
2006-11-12 20:17:02 +01:00
32 , 0 , 0 )
invoke.texi (l1-cache-size): Update documentation.
* doc/invoke.texi (l1-cache-size): Update documentation.
(l2-cache-size): Document.
* params.h (L2_CACHE_SIZE): New macro.
* tree-ssa-loop-prefetch.c (L1_CACHE_SIZE_BYTES): Reflect
that L1_CACHE_SIZE is in kB now.
(L2_CACHE_SIZE_BYTES): New macro.
(tree_ssa_prefetch_arrays): Show size in kB.
* config/i386/i386.h (struct processor_costs): Add l1_cache_size
and l2_cache_size fields.
* config/i386/driver-i386.c (describe_cache): Detect cache size
in kB.
* config/i386/i386.c (size_cost, i386_cost, i486_cost,pentium_cost,
pentiumpro_cost, geode_cost, k6_cost, athlon_cost, k8_cost,
amdfam10_cost, pentium4_cost, nocona_cost, core2_cost,
generic64_cost, generic32_cost): Add l1_cache_size and l2_cache_size.
(override_options): Set l1-cache-size and l2-cache-size to default
values if not specified otherwise.
* params.def (PARAM_L1_CACHE_SIZE): Change to set in kB.
(PARAM_L2_CACHE_SIZE): New.
From-SVN: r127117
2007-08-01 12:57:03 +02:00
/* The size of L2 cache in kB. */
DEFPARAM ( PARAM_L2_CACHE_SIZE ,
"l2-cache-size" ,
2015-10-22 00:24:41 +02:00
"The size of L2 cache." ,
invoke.texi (l1-cache-size): Update documentation.
* doc/invoke.texi (l1-cache-size): Update documentation.
(l2-cache-size): Document.
* params.h (L2_CACHE_SIZE): New macro.
* tree-ssa-loop-prefetch.c (L1_CACHE_SIZE_BYTES): Reflect
that L1_CACHE_SIZE is in kB now.
(L2_CACHE_SIZE_BYTES): New macro.
(tree_ssa_prefetch_arrays): Show size in kB.
* config/i386/i386.h (struct processor_costs): Add l1_cache_size
and l2_cache_size fields.
* config/i386/driver-i386.c (describe_cache): Detect cache size
in kB.
* config/i386/i386.c (size_cost, i386_cost, i486_cost,pentium_cost,
pentiumpro_cost, geode_cost, k6_cost, athlon_cost, k8_cost,
amdfam10_cost, pentium4_cost, nocona_cost, core2_cost,
generic64_cost, generic32_cost): Add l1_cache_size and l2_cache_size.
(override_options): Set l1-cache-size and l2-cache-size to default
values if not specified otherwise.
* params.def (PARAM_L1_CACHE_SIZE): Change to set in kB.
(PARAM_L2_CACHE_SIZE): New.
From-SVN: r127117
2007-08-01 12:57:03 +02:00
512 , 0 , 0 )
2007-07-11 15:50:13 +02:00
/ * Whether we should use canonical types rather than deep "structural"
type checking . Setting this value to 1 ( the default ) improves
compilation performance in the C ++ and Objective - C ++ front end ;
this value should only be set to zero to work around bugs in the
canonical type system by disabling it . * /
DEFPARAM ( PARAM_USE_CANONICAL_TYPES ,
"use-canonical-types" ,
2015-10-22 00:24:41 +02:00
"Whether to use canonical types." ,
2007-07-11 15:50:13 +02:00
1 , 0 , 1 )
2007-11-05 16:42:30 +01:00
DEFPARAM ( PARAM_MAX_PARTIAL_ANTIC_LENGTH ,
"max-partial-antic-length" ,
2015-10-22 00:24:41 +02:00
"Maximum length of partial antic set when performing tree pre optimization." ,
2007-11-05 16:42:30 +01:00
100 , 0 , 0 )
2007-12-13 15:22:23 +01:00
/ * The following is used as a stop - gap limit for cases where really huge
SCCs blow up memory and compile - time use too much . If we hit this limit ,
SCCVN and such FRE and PRE will be not done at all for the current
function . * /
DEFPARAM ( PARAM_SCCVN_MAX_SCC_SIZE ,
"sccvn-max-scc-size" ,
2015-10-22 00:24:41 +02:00
"Maximum size of a SCC before SCCVN stops processing a function." ,
2007-12-13 15:22:23 +01:00
10000 , 10 , 0 )
2012-08-22 15:17:26 +02:00
/ * The following is used as a stop - gap limit for cases where really huge
functions blow up compile - time use too much . It limits the number of
alias - queries we do for finding common subexpressions for memory loads and
stores . The number of alias - queries is otherwise limited by the number of
stores on paths to function entry . * /
DEFPARAM ( PARAM_SCCVN_MAX_ALIAS_QUERIES_PER_ACCESS ,
"sccvn-max-alias-queries-per-access" ,
2015-10-22 00:24:41 +02:00
"Maximum number of disambiguations to perform per memory access." ,
2012-08-22 15:17:26 +02:00
1000 , 0 , 0 )
[multiple changes]
2008-08-26 Vladimir Makarov <vmakarov@redhat.com>
* ira-build.c, ira-color.c, ira-costs.c, ira.h, ira-lives.c,
ira.c, ira-conflicts.c, ira-emit.c, ira-int.h: New files.
* doc/passes.texi: Describe IRA.
* doc/tm.texi (IRA_COVER_CLASSES,
IRA_HARD_REGNO_ADD_COST_MULTIPLIER): Describe the new macros.
* doc/invoke.texi (ira-max-loops-num): Describe the new parameter.
(-fira, -fira-algorithm, -fira-coalesce, -fno-ira-move-spills,
-fira-propagate-cost, -fno-ira-share-save-slots,
-fno-ira-share-spill-slots, -fira-verbose): Describe new options.
* flags.h (ira_algorithm): New enumeration.
(flag_ira_algorithm, flag_ira_verbose): New external variable
declarations.
* postreload.c (gate_handle_postreload): Don't do post reload
optimizations unless the reload is completed.
* reload.c (push_reload, find_dummy_reload): Use DF_LR_OUT for
IRA.
* tree-pass.h (pass_ira): New external variable declaration.
* reload.h: Add 2008 to the Copyright.
* cfgloopanal.c: Include params.h.
(estimate_reg_pressure_cost): Decrease cost for IRA optimization
mode.
* params.h (IRA_MAX_LOOPS_NUM): New macro.
* toplev.c (ira.h): New include.
(flag_ira_algorithm, flag_ira_verbose): New external variables.
(backend_init_target): Call ira_init.
(backend_init): Call ira_init_once.
(finalize): Call finish_ira_once.
* toplev.h (flag_ira, flag_ira_coalesce, flag_ira_move_spills,
flag_ira_share_save_slots, flag_ira_share_spill_slots): New
external variables.
* regs.h (contains_reg_of_mode, move_cost, may_move_in_cost,
may_move_out_cost): New external variable declarations.
(move_table): New typedef.
* caller-save.c: Include headers output.h and ira.h.
(no_caller_save_reg_set): New global variable.
(save_slots_num, save_slots): New variables.
(reg_save_code, reg_restore_code, add_stored_regs): Add
prototypes.
(init_caller_save): Set up no_caller_save_reg_set.
(init_save_areas): Reset save_slots_num.
(saved_hard_reg): New structure.
(hard_reg_map, saved_regs_num, all_saved_regs): New variables.
(initiate_saved_hard_regs, new_saved_hard_reg,
finish_saved_hard_regs, saved_hard_reg_compare_func): New
functions.
(setup_save_areas): Add code for sharing stack slots.
(all_blocks): New variable.
(save_call_clobbered_regs): Process pseudo-register too.
(mark_set_regs): Process pseudo-register too.
(insert_one_insn): Put the insn after bb note in a empty basic
block. Add insn check.
* global.c (eliminable_regset): Make it external.
(mark_elimination): Use DF_LR_IN for IRA.
(pseudo_for_reload_consideration_p): New.
(build_insn_chain): Make it external. Don't ignore spilled
pseudos for IRA. Use pseudo_for_reload_consideration_p.
(gate_handle_global_alloc): New function.
(pass_global_alloc): Add the gate function.
* opts.c (decode_options): Set up flag_ira. Print the warning for
-fira.
(common_handle_option): Process -fira-algorithm and -fira-verbose.
* timevar.def (TV_IRA, TV_RELOAD): New passes.
* regmove.c (regmove_optimize): Don't do replacement of output for
IRA.
* hard-reg-set.h (no_caller_save_reg_set, reg_class_subclasses):
New external variable declarations.
* local-alloc.c (update_equiv_regs): Make it external. Return
true if jump label rebuilding should be done. Rescan new_insn for
notes.
(gate_handle_local_alloc): New function.
(pass_local_alloc): Add the gate function.
* alias.c (value_addr_p, stack_addr_p): New functions.
(nonoverlapping_memrefs_p): Use them for IRA.
* common.opt (fira, fira-algorithm, fira-coalesce,
fira-move-spills, fira-share-save-slots, fira-share-spill-slots,
fira-verbose): New options.
* regclass.c (reg_class_subclasses, contains_reg_of_mode,
move_cost, may_move_in_cost, may_move_out_cost): Make the
variables external.
(move_table): Remove typedef.
(init_move_cost): Make it external.
(allocate_reg_info, resize_reg_info, setup_reg_classes): New
functions.
* rtl.h (init_move_cost, allocate_reg_info, resize_reg_info,
setup_reg_classes): New function prototypes.
(eliminable_regset): New external variable declaration.
(build_insn_chain, update_equiv_regs): New function prototypes.
* Makefile.in (IRA_INT_H): New definition.
(OBJS-common): Add ira.o, ira-build.o, ira-costs.o,
ira-conflicts.o, ira-color.o, ira-emit.o, and ira-lives.o.
(reload1.o, toplev.o): Add dependence on ira.h.
(cfgloopanal.o): Add PARAMS_H.
(caller-save.o): Add dependence on output.h and ira.h.
(ira.o, ira-build.o, ira-costs.o, ira-conflicts.o, ira-color.o,
ira-emit.o, ira-lives.o): New entries.
* passes.c (pass_ira): New pass.
* params.def (PARAM_IRA_MAX_LOOPS_NUM): New parameter.
* reload1.c (ira.h): Include the header.
(changed_allocation_pseudos): New bitmap.
(init_reload): Initiate the bitmap.
(compute_use_by_pseudos): Permits spilled registers in FROM.
(temp_pseudo_reg_arr): New variable.
(reload): Allocate and free temp_pseudo_reg_arr. Sort pseudos for
IRA. Call alter_reg with the additional parameter. Don't clear
spilled_pseudos for IRA. Restore original insn chain for IRA.
Clear changed_allocation_pseudos at the end of reload.
(calculate_needs_all_insns): Call IRA's mark_memory_move_deletion.
(hard_regno_to_pseudo_regno): New variable.
(count_pseudo): Check spilled pseudos. Set up
hard_regno_to_pseudo_regno.
(count_spilled_pseudo): Check spilled pseudos. Update
hard_regno_to_pseudo_regno.
(find_reg): Use better_spill_reload_regno_p. Check
hard_regno_to_pseudo_regno.
(alter_reg): Set up spilled_pseudos. Add a new parameter. Add
code for IRA.
(eliminate_regs_1): Use additional parameter for alter_reg.
(finish_spills): Set up pseudo_previous_regs only for spilled
pseudos. Call reassign_pseudos once for all spilled pseudos, pass
more arguments. Don't clear live_throughout and dead_or_set for
spilled pseudos. Use additional parameter for alter_reg. Call
mark_allocation_change. Set up changed_allocation_pseudos.
Remove sanity check.
(emit_input_reload_insns, delete_output_reload): Use additional
parameter for alter_reg. Call mark_allocation_change.
(substitute, gen_reload_chain_without_interm_reg_p): New
functions.
(reloads_conflict): Use gen_reload_chain_without_interm_reg_p.
* testsuite/gcc.dg/20080410-1.c: New file.
* config/s390/s390.h (IRA_COVER_CLASSES,
IRA_HARD_REGNO_ADD_COST_MULTIPLIER): Define.
* config/sparc/sparc.h (IRA_COVER_CLASSES): New macro.
* config/i386/i386.h (IRA_COVER_CLASSES): Ditto.
* config/ia64/ia64.h (IRA_COVER_CLASSES): Ditto.
* config/rs6000/rs6000.h (IRA_COVER_CLASSES): Ditto.
* config/arm/arm.h (IRA_COVER_CLASSES): Ditto.
* config/alpha/alpha.h (IRA_COVER_CLASSES): Ditto.
2008-08-24 Jeff Law <law@redhat.com>
* ira.c (setup_reg_class_intersect_union): Prefer smallest class
when ignoring unavailable registers.
2008-08-24 Jeff Law <law@redhat.com>
* ira-color.c (coalesced_pseudo_reg_slot_compare): Check
FRAME_GROWS_DOWNWARD and STACK_GROWS_DOWNWARD.
* ira.c (setup_eliminable_regset): Check stack_realign_needed.
* config/mn10300/mn10300.h (IRA_COVER_CLASSES): New macro.
2008-06-03 Steve Chamberlain <steve.chamberlain@gmail.com>
* ira-build.c (allocno_range_compare_func): Stabilize sort.
2008-05-29 Andy Hutchinson <hutchinsonandy@aim.com>
* config/avr/avr.h (IRA_COVER_CLASSES): New macro.
* reload1.c (find_reg): Process registers in register allocation order.
2008-05-10 Richard Sandiford <rsandifo@nildram.co.uk>
* toplev.c (backend_init_target): Move ira_init call from
here...
(lang_dependent_init_target): ...to here.
2008-05-10 Richard Sandiford <rsandifo@nildram.co.uk>
* ira.c (setup_class_subset_and_memory_move_costs): Don't
calculate memory move costs for NO_REGS.
2008-05-05 Kaz Kojima <kkojima@gcc.gnu.org>
* ira-color.c (ira_fast_allocation): Use no_stack_reg_p only if
STACK_REGS is defined.
2008-04-08 Andrew Pinski <andrew_pinski@playstation.sony.com>
* config/spu/spu.h (IRA_COVER_CLASSES): New macro.
2008-04-04 Bernd Schmidt <bernd.schmidt@analog.com>
* config/bfin/bfin.h (IRA_COVER_CLASSES): New macro.
2008-04-04 Kaz Kojima <kkojima@gcc.gnu.org>
* config/sh/sh.h (IRA_COVER_CLASSES): Define.
* config/sh/sh.md (movsicc_true+3): Check if emit returns a
barrier.
From-SVN: r139590
2008-08-26 14:39:58 +02:00
DEFPARAM ( PARAM_IRA_MAX_LOOPS_NUM ,
"ira-max-loops-num" ,
2015-10-22 00:24:41 +02:00
"Max loops number for regional RA." ,
2008-11-25 23:52:37 +01:00
100 , 0 , 0 )
[multiple changes]
2008-08-26 Vladimir Makarov <vmakarov@redhat.com>
* ira-build.c, ira-color.c, ira-costs.c, ira.h, ira-lives.c,
ira.c, ira-conflicts.c, ira-emit.c, ira-int.h: New files.
* doc/passes.texi: Describe IRA.
* doc/tm.texi (IRA_COVER_CLASSES,
IRA_HARD_REGNO_ADD_COST_MULTIPLIER): Describe the new macros.
* doc/invoke.texi (ira-max-loops-num): Describe the new parameter.
(-fira, -fira-algorithm, -fira-coalesce, -fno-ira-move-spills,
-fira-propagate-cost, -fno-ira-share-save-slots,
-fno-ira-share-spill-slots, -fira-verbose): Describe new options.
* flags.h (ira_algorithm): New enumeration.
(flag_ira_algorithm, flag_ira_verbose): New external variable
declarations.
* postreload.c (gate_handle_postreload): Don't do post reload
optimizations unless the reload is completed.
* reload.c (push_reload, find_dummy_reload): Use DF_LR_OUT for
IRA.
* tree-pass.h (pass_ira): New external variable declaration.
* reload.h: Add 2008 to the Copyright.
* cfgloopanal.c: Include params.h.
(estimate_reg_pressure_cost): Decrease cost for IRA optimization
mode.
* params.h (IRA_MAX_LOOPS_NUM): New macro.
* toplev.c (ira.h): New include.
(flag_ira_algorithm, flag_ira_verbose): New external variables.
(backend_init_target): Call ira_init.
(backend_init): Call ira_init_once.
(finalize): Call finish_ira_once.
* toplev.h (flag_ira, flag_ira_coalesce, flag_ira_move_spills,
flag_ira_share_save_slots, flag_ira_share_spill_slots): New
external variables.
* regs.h (contains_reg_of_mode, move_cost, may_move_in_cost,
may_move_out_cost): New external variable declarations.
(move_table): New typedef.
* caller-save.c: Include headers output.h and ira.h.
(no_caller_save_reg_set): New global variable.
(save_slots_num, save_slots): New variables.
(reg_save_code, reg_restore_code, add_stored_regs): Add
prototypes.
(init_caller_save): Set up no_caller_save_reg_set.
(init_save_areas): Reset save_slots_num.
(saved_hard_reg): New structure.
(hard_reg_map, saved_regs_num, all_saved_regs): New variables.
(initiate_saved_hard_regs, new_saved_hard_reg,
finish_saved_hard_regs, saved_hard_reg_compare_func): New
functions.
(setup_save_areas): Add code for sharing stack slots.
(all_blocks): New variable.
(save_call_clobbered_regs): Process pseudo-register too.
(mark_set_regs): Process pseudo-register too.
(insert_one_insn): Put the insn after bb note in a empty basic
block. Add insn check.
* global.c (eliminable_regset): Make it external.
(mark_elimination): Use DF_LR_IN for IRA.
(pseudo_for_reload_consideration_p): New.
(build_insn_chain): Make it external. Don't ignore spilled
pseudos for IRA. Use pseudo_for_reload_consideration_p.
(gate_handle_global_alloc): New function.
(pass_global_alloc): Add the gate function.
* opts.c (decode_options): Set up flag_ira. Print the warning for
-fira.
(common_handle_option): Process -fira-algorithm and -fira-verbose.
* timevar.def (TV_IRA, TV_RELOAD): New passes.
* regmove.c (regmove_optimize): Don't do replacement of output for
IRA.
* hard-reg-set.h (no_caller_save_reg_set, reg_class_subclasses):
New external variable declarations.
* local-alloc.c (update_equiv_regs): Make it external. Return
true if jump label rebuilding should be done. Rescan new_insn for
notes.
(gate_handle_local_alloc): New function.
(pass_local_alloc): Add the gate function.
* alias.c (value_addr_p, stack_addr_p): New functions.
(nonoverlapping_memrefs_p): Use them for IRA.
* common.opt (fira, fira-algorithm, fira-coalesce,
fira-move-spills, fira-share-save-slots, fira-share-spill-slots,
fira-verbose): New options.
* regclass.c (reg_class_subclasses, contains_reg_of_mode,
move_cost, may_move_in_cost, may_move_out_cost): Make the
variables external.
(move_table): Remove typedef.
(init_move_cost): Make it external.
(allocate_reg_info, resize_reg_info, setup_reg_classes): New
functions.
* rtl.h (init_move_cost, allocate_reg_info, resize_reg_info,
setup_reg_classes): New function prototypes.
(eliminable_regset): New external variable declaration.
(build_insn_chain, update_equiv_regs): New function prototypes.
* Makefile.in (IRA_INT_H): New definition.
(OBJS-common): Add ira.o, ira-build.o, ira-costs.o,
ira-conflicts.o, ira-color.o, ira-emit.o, and ira-lives.o.
(reload1.o, toplev.o): Add dependence on ira.h.
(cfgloopanal.o): Add PARAMS_H.
(caller-save.o): Add dependence on output.h and ira.h.
(ira.o, ira-build.o, ira-costs.o, ira-conflicts.o, ira-color.o,
ira-emit.o, ira-lives.o): New entries.
* passes.c (pass_ira): New pass.
* params.def (PARAM_IRA_MAX_LOOPS_NUM): New parameter.
* reload1.c (ira.h): Include the header.
(changed_allocation_pseudos): New bitmap.
(init_reload): Initiate the bitmap.
(compute_use_by_pseudos): Permits spilled registers in FROM.
(temp_pseudo_reg_arr): New variable.
(reload): Allocate and free temp_pseudo_reg_arr. Sort pseudos for
IRA. Call alter_reg with the additional parameter. Don't clear
spilled_pseudos for IRA. Restore original insn chain for IRA.
Clear changed_allocation_pseudos at the end of reload.
(calculate_needs_all_insns): Call IRA's mark_memory_move_deletion.
(hard_regno_to_pseudo_regno): New variable.
(count_pseudo): Check spilled pseudos. Set up
hard_regno_to_pseudo_regno.
(count_spilled_pseudo): Check spilled pseudos. Update
hard_regno_to_pseudo_regno.
(find_reg): Use better_spill_reload_regno_p. Check
hard_regno_to_pseudo_regno.
(alter_reg): Set up spilled_pseudos. Add a new parameter. Add
code for IRA.
(eliminate_regs_1): Use additional parameter for alter_reg.
(finish_spills): Set up pseudo_previous_regs only for spilled
pseudos. Call reassign_pseudos once for all spilled pseudos, pass
more arguments. Don't clear live_throughout and dead_or_set for
spilled pseudos. Use additional parameter for alter_reg. Call
mark_allocation_change. Set up changed_allocation_pseudos.
Remove sanity check.
(emit_input_reload_insns, delete_output_reload): Use additional
parameter for alter_reg. Call mark_allocation_change.
(substitute, gen_reload_chain_without_interm_reg_p): New
functions.
(reloads_conflict): Use gen_reload_chain_without_interm_reg_p.
* testsuite/gcc.dg/20080410-1.c: New file.
* config/s390/s390.h (IRA_COVER_CLASSES,
IRA_HARD_REGNO_ADD_COST_MULTIPLIER): Define.
* config/sparc/sparc.h (IRA_COVER_CLASSES): New macro.
* config/i386/i386.h (IRA_COVER_CLASSES): Ditto.
* config/ia64/ia64.h (IRA_COVER_CLASSES): Ditto.
* config/rs6000/rs6000.h (IRA_COVER_CLASSES): Ditto.
* config/arm/arm.h (IRA_COVER_CLASSES): Ditto.
* config/alpha/alpha.h (IRA_COVER_CLASSES): Ditto.
2008-08-24 Jeff Law <law@redhat.com>
* ira.c (setup_reg_class_intersect_union): Prefer smallest class
when ignoring unavailable registers.
2008-08-24 Jeff Law <law@redhat.com>
* ira-color.c (coalesced_pseudo_reg_slot_compare): Check
FRAME_GROWS_DOWNWARD and STACK_GROWS_DOWNWARD.
* ira.c (setup_eliminable_regset): Check stack_realign_needed.
* config/mn10300/mn10300.h (IRA_COVER_CLASSES): New macro.
2008-06-03 Steve Chamberlain <steve.chamberlain@gmail.com>
* ira-build.c (allocno_range_compare_func): Stabilize sort.
2008-05-29 Andy Hutchinson <hutchinsonandy@aim.com>
* config/avr/avr.h (IRA_COVER_CLASSES): New macro.
* reload1.c (find_reg): Process registers in register allocation order.
2008-05-10 Richard Sandiford <rsandifo@nildram.co.uk>
* toplev.c (backend_init_target): Move ira_init call from
here...
(lang_dependent_init_target): ...to here.
2008-05-10 Richard Sandiford <rsandifo@nildram.co.uk>
* ira.c (setup_class_subset_and_memory_move_costs): Don't
calculate memory move costs for NO_REGS.
2008-05-05 Kaz Kojima <kkojima@gcc.gnu.org>
* ira-color.c (ira_fast_allocation): Use no_stack_reg_p only if
STACK_REGS is defined.
2008-04-08 Andrew Pinski <andrew_pinski@playstation.sony.com>
* config/spu/spu.h (IRA_COVER_CLASSES): New macro.
2008-04-04 Bernd Schmidt <bernd.schmidt@analog.com>
* config/bfin/bfin.h (IRA_COVER_CLASSES): New macro.
2008-04-04 Kaz Kojima <kkojima@gcc.gnu.org>
* config/sh/sh.h (IRA_COVER_CLASSES): Define.
* config/sh/sh.md (movsicc_true+3): Check if emit returns a
barrier.
From-SVN: r139590
2008-08-26 14:39:58 +02:00
re PR rtl-optimization/38583 (huge test case makes register allocator run out of memory while constructing the conflict graph)
2009-01-05 Vladimir Makarov <vmakarov@redhat.com>
PR rtl-optimization/38583
* params.h (IRA_MAX_CONFLICT_TABLE_SIZE): New macro.
* params.def (ira-max-conflict-table-size): New.
* doc/invoke.texi (ira-max-conflict-table-size): Decribe.
* ira.h (ira_conflicts_p): New external definition.
* ira-conflicts.c (build_conflict_bit_table): Do not build too big
table. Report this. Return result of building.
(ira_build_conflicts): Use ira_conflicts_p. Check result of
building conflict table.
* ira-color.c (fast_allocation): Use num instead of
ira_allocnos_num.
(ira_color): Use ira_conflicts_p.
* global.c: Include ira.h.
(pseudo_for_reload_consideration_p, build_insn_chain): Use
ira_conflicts_p.
* Makefile.in (global.o): Add ira.h.
* ira-build.c (mark_all_loops_for_removal,
propagate_some_info_from_allocno): New.
(remove_unnecessary_allocnos): Call
propagate_some_info_from_allocno.
(remove_low_level_allocnos): New.
(remove_unnecessary_regions): Add parameter. Call
mark_all_loops_for_removal and remove_low_level_allocnos. Pass
parameter to remove_unnecessary_regions.
(ira_build): Remove all regions but root if the conflict table was
not built. Update conflict hard regs for allocnos crossing calls.
* ira.c (ira_conflicts_p): New global.
(ira): Define and use ira_conflicts_p.
* reload1.c (compute_use_by_pseudos, reload, count_pseudo,
count_spilled_pseudo, find_reg, alter_reg, finish_spills,
emit_input_reload_insns, delete_output_reload): Use
ira_conflicts_p.
From-SVN: r143112
2009-01-06 04:50:07 +01:00
DEFPARAM ( PARAM_IRA_MAX_CONFLICT_TABLE_SIZE ,
"ira-max-conflict-table-size" ,
2015-10-22 00:24:41 +02:00
"Max size of conflict table in MB." ,
2009-01-08 21:54:29 +01:00
1000 , 0 , 0 )
re PR rtl-optimization/38583 (huge test case makes register allocator run out of memory while constructing the conflict graph)
2009-01-05 Vladimir Makarov <vmakarov@redhat.com>
PR rtl-optimization/38583
* params.h (IRA_MAX_CONFLICT_TABLE_SIZE): New macro.
* params.def (ira-max-conflict-table-size): New.
* doc/invoke.texi (ira-max-conflict-table-size): Decribe.
* ira.h (ira_conflicts_p): New external definition.
* ira-conflicts.c (build_conflict_bit_table): Do not build too big
table. Report this. Return result of building.
(ira_build_conflicts): Use ira_conflicts_p. Check result of
building conflict table.
* ira-color.c (fast_allocation): Use num instead of
ira_allocnos_num.
(ira_color): Use ira_conflicts_p.
* global.c: Include ira.h.
(pseudo_for_reload_consideration_p, build_insn_chain): Use
ira_conflicts_p.
* Makefile.in (global.o): Add ira.h.
* ira-build.c (mark_all_loops_for_removal,
propagate_some_info_from_allocno): New.
(remove_unnecessary_allocnos): Call
propagate_some_info_from_allocno.
(remove_low_level_allocnos): New.
(remove_unnecessary_regions): Add parameter. Call
mark_all_loops_for_removal and remove_low_level_allocnos. Pass
parameter to remove_unnecessary_regions.
(ira_build): Remove all regions but root if the conflict table was
not built. Update conflict hard regs for allocnos crossing calls.
* ira.c (ira_conflicts_p): New global.
(ira): Define and use ira_conflicts_p.
* reload1.c (compute_use_by_pseudos, reload, count_pseudo,
count_spilled_pseudo, find_reg, alter_reg, finish_spills,
emit_input_reload_insns, delete_output_reload): Use
ira_conflicts_p.
From-SVN: r143112
2009-01-06 04:50:07 +01:00
params.def (PARAM_IRA_LOOP_RESERVED_REGS): New.
2009-09-26 Vladimir Makarov <vmakarov@redhat.com>
* params.def (PARAM_IRA_LOOP_RESERVED_REGS): New.
* params.h (IRA_LOOP_RESERVED_REGS): New
* tree-pass.h (pass_subregs_of_mode_init,
pass_subregs_of_mode_finish): Remove.
* passes.c (pass_subregs_of_mode_init,
pass_subregs_of_mode_finish): Remove.
(pass_reginfo_init): Move before loop optimizations.
* config/i386/i386.h (STACK_REG_COVER_CLASS): Define.
* common.opt (fira-loop-pressure): New.
* toplev.h (flag_ira_loop_pressure): New.
* rtl.h (init_subregs_of_mode, finish_subregs_of_mode): New
externals.
* reginfo.c (init_subregs_of_mode, finish_subregs_of_mode):
Make external and void type functions.
(gate_subregs_of_mode_init, pass_subregs_of_mode_init,
pass_subregs_of_mode_finish): Remove.
* ira-costs.c (init_costs): Call init_subregs_of_mode.
* regmove.c: Include ira.h.
(regmove_optimize): Call ira_set_pseudo_classes after IRA based
register pressure calculation in loops.
* loop-invariant.c: Include REGS_H and ira.h.
(struct loop_data): New members max_reg_pressure, regs_ref, and
regs_live.
(struct invariant): New member orig_regno.
(curr_loop): New variable.
(find_exits): Initialize regs_ref and regs_live.
(create_new_invariant): Initialize orig_regno.
(get_cover_class_and_nregs): New.
(get_inv_cost): Make aregs_needed an array. Use regs_needed as an
array. Add code for flag_ira_loop_pressure.
(gain_for_invariant): Make new_regs an array. Add code for
flag_ira_loop_pressure.
(best_gain_for_invariant): Ditto.
(set_move_mark): New parameter gain. Use it for debugging output.
(find_invariants_to_move): Make regs_needed and new_regs an array.
Add code for flag_ira_loop_pressure.
(move_invariant_reg): Set up orig_regno.
(move_invariants): Set up reg classes for pseudos for
flag_ira_loop_pressure.
(free_loop_data): Clear regs_ref and regs_live.
(curr_regs_live, curr_reg_pressure, regs_set, n_regs_set,
get_regno_cover_class, change_pressure, mark_regno_live,
mark_regno_death, mark_reg_store, mark_reg_clobber,
mark_reg_death, mark_ref_regs, calculate_loop_reg_pressure): New.
(move_loop_invariants): Calculate pressure. Initialize curr_loop.
* ira.c (ira): Call ira_set_pseudo_classes after IRA based
register pressure calculation in loops if new regs were added.
Call finish_subregs_of_mode.
* opts.c (decode_options): Set up flag_ira_loop_pressure.
* Makefile.in (loop-invariant.o): Add ira.h.
(regmove.o): Ditto.
* doc/invoke.texi (-fira-loop-pressure, ira-loop-reserved-regs):
Describe.
* doc/tm.texi (STACK_REG_COVER_CLASS): Describe.
From-SVN: r152770
2009-10-14 18:24:11 +02:00
DEFPARAM ( PARAM_IRA_LOOP_RESERVED_REGS ,
"ira-loop-reserved-regs" ,
2015-10-22 00:24:41 +02:00
"The number of registers in each class kept unused by loop invariant motion." ,
params.def (PARAM_IRA_LOOP_RESERVED_REGS): New.
2009-09-26 Vladimir Makarov <vmakarov@redhat.com>
* params.def (PARAM_IRA_LOOP_RESERVED_REGS): New.
* params.h (IRA_LOOP_RESERVED_REGS): New
* tree-pass.h (pass_subregs_of_mode_init,
pass_subregs_of_mode_finish): Remove.
* passes.c (pass_subregs_of_mode_init,
pass_subregs_of_mode_finish): Remove.
(pass_reginfo_init): Move before loop optimizations.
* config/i386/i386.h (STACK_REG_COVER_CLASS): Define.
* common.opt (fira-loop-pressure): New.
* toplev.h (flag_ira_loop_pressure): New.
* rtl.h (init_subregs_of_mode, finish_subregs_of_mode): New
externals.
* reginfo.c (init_subregs_of_mode, finish_subregs_of_mode):
Make external and void type functions.
(gate_subregs_of_mode_init, pass_subregs_of_mode_init,
pass_subregs_of_mode_finish): Remove.
* ira-costs.c (init_costs): Call init_subregs_of_mode.
* regmove.c: Include ira.h.
(regmove_optimize): Call ira_set_pseudo_classes after IRA based
register pressure calculation in loops.
* loop-invariant.c: Include REGS_H and ira.h.
(struct loop_data): New members max_reg_pressure, regs_ref, and
regs_live.
(struct invariant): New member orig_regno.
(curr_loop): New variable.
(find_exits): Initialize regs_ref and regs_live.
(create_new_invariant): Initialize orig_regno.
(get_cover_class_and_nregs): New.
(get_inv_cost): Make aregs_needed an array. Use regs_needed as an
array. Add code for flag_ira_loop_pressure.
(gain_for_invariant): Make new_regs an array. Add code for
flag_ira_loop_pressure.
(best_gain_for_invariant): Ditto.
(set_move_mark): New parameter gain. Use it for debugging output.
(find_invariants_to_move): Make regs_needed and new_regs an array.
Add code for flag_ira_loop_pressure.
(move_invariant_reg): Set up orig_regno.
(move_invariants): Set up reg classes for pseudos for
flag_ira_loop_pressure.
(free_loop_data): Clear regs_ref and regs_live.
(curr_regs_live, curr_reg_pressure, regs_set, n_regs_set,
get_regno_cover_class, change_pressure, mark_regno_live,
mark_regno_death, mark_reg_store, mark_reg_clobber,
mark_reg_death, mark_ref_regs, calculate_loop_reg_pressure): New.
(move_loop_invariants): Calculate pressure. Initialize curr_loop.
* ira.c (ira): Call ira_set_pseudo_classes after IRA based
register pressure calculation in loops if new regs were added.
Call finish_subregs_of_mode.
* opts.c (decode_options): Set up flag_ira_loop_pressure.
* Makefile.in (loop-invariant.o): Add ira.h.
(regmove.o): Ditto.
* doc/invoke.texi (-fira-loop-pressure, ira-loop-reserved-regs):
Describe.
* doc/tm.texi (STACK_REG_COVER_CLASS): Describe.
From-SVN: r152770
2009-10-14 18:24:11 +02:00
2 , 0 , 0 )
2014-02-25 21:34:44 +01:00
DEFPARAM ( PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS ,
"lra-max-considered-reload-pseudos" ,
2015-10-22 00:24:41 +02:00
"The max number of reload pseudos which are considered during spilling a non-reload pseudo." ,
2014-02-25 21:34:44 +01:00
500 , 0 , 0 )
2015-02-27 23:02:05 +01:00
DEFPARAM ( PARAM_LRA_INHERITANCE_EBB_PROBABILITY_CUTOFF ,
"lra-inheritance-ebb-probability-cutoff" ,
2015-10-22 00:24:41 +02:00
"Minimal fall-through edge probability in percentage used to add BB to inheritance EBB in LRA." ,
2015-02-27 23:02:05 +01:00
40 , 0 , 100 )
2008-07-01 10:54:18 +02:00
/ * Switch initialization conversion will refuse to create arrays that are
bigger than this parameter times the number of switch branches . * /
DEFPARAM ( PARAM_SWITCH_CONVERSION_BRANCH_RATIO ,
"switch-conversion-max-branch-ratio" ,
"The maximum ratio between array size and switch branches for "
2015-10-22 00:24:41 +02:00
"a switch conversion to take place." ,
2008-07-01 10:54:18 +02:00
8 , 1 , 0 )
2009-11-25 06:00:38 +01:00
/* Size of tiles when doing loop blocking. */
DEFPARAM ( PARAM_LOOP_BLOCK_TILE_SIZE ,
"loop-block-tile-size" ,
2015-10-22 00:24:41 +02:00
"size of tiles for loop blocking." ,
2009-11-25 06:00:38 +01:00
51 , 0 , 0 )
2010-03-13 18:34:38 +01:00
/* Maximal number of parameters that we allow in a SCoP. */
DEFPARAM ( PARAM_GRAPHITE_MAX_NB_SCOP_PARAMS ,
"graphite-max-nb-scop-params" ,
2015-10-22 00:24:41 +02:00
"maximum number of parameters in a SCoP." ,
2015-10-05 18:03:21 +02:00
7 , 0 , 0 )
2010-03-13 18:34:38 +01:00
2010-03-13 18:34:51 +01:00
/* Maximal number of basic blocks in the functions analyzed by Graphite. */
DEFPARAM ( PARAM_GRAPHITE_MAX_BBS_PER_FUNCTION ,
"graphite-max-bbs-per-function" ,
2015-10-22 00:24:41 +02:00
"maximum number of basic blocks per function to be analyzed by Graphite." ,
2010-03-13 18:34:51 +01:00
100 , 0 , 0 )
2015-10-21 23:18:17 +02:00
/* Maximal number of array references in a scop. */
DEFPARAM ( PARAM_GRAPHITE_MAX_ARRAYS_PER_SCOP ,
"graphite-max-arrays-per-scop" ,
2015-10-22 00:24:41 +02:00
"maximum number of arrays per scop." ,
2015-10-21 23:18:17 +02:00
100 , 0 , 0 )
Redesign Graphite scop detection
Redesign Graphite scop detection for faster compiler time and detecting more SCoPs.
Existing algorithm for SCoP detection in graphite was based on dominator tree
where a tree (CFG) traversal was required for analyzing an SESE. The tree
traversal is linear in the number of basic blocks and SCoP detection is
(probably) linear in number of instructions. That algorithm utilized a generic
infrastructure of SESE which does not directly represent loops. With regards to
graphite framework, we are only interested in subtrees with loops. The new
algorithm is geared towards tree traversal on loop structure. The algorithm is
linear in number of loops which is faster than the previous algorithm.
Briefly, we start the traversal at a loop-nest and analyze it recursively for
validity. Once a valid loop is found we find a valid adjacent loop. If an
adjacent loop is found and is valid, we merge both loop nests otherwise we form
a SCoP from the previous loop nest, and resume the algorithm from the adjacent
loop nest. The data structure to represent an SESE is an ordered pair of edges
(entry, exit). The new algoritm can extend a SCoP in both the directions. With
this approach, the number of instructions to be analyzed for validity reduces to
a minimal set. We start by analyzing those statements which are inside a loop,
because validity of those statements is necessary for the validity of loop. The
statements outside the loop nest can be just excluded from the SESE if they are
not valid.
This patch depends on: https://gcc.gnu.org/ml/gcc-patches/2015-09/msg02024.html
Passes (c,c++,fortran) regtest and bootstrap.
gcc/ChangeLog:
2015-09-27 Aditya Kumar <hiraditya@msn.com>
Sebastian Pop <s.pop@samsung.com>
* graphite-optimize-isl.c (optimize_isl):
* graphite-scop-detection.c (struct sese_l): New type.
(get_entry_bb): API for getting entry bb of SESE.
(get_exit_bb): API for getting exit bb of SESE.
(class debug_printer): New type. Simple printer in debug mode.
(trivially_empty_bb_p): New. Return true when BB is empty or
contains only debug instructions.
(graphite_can_represent_expr): Call scalar_evoution_in_region
instead of analyze_scalar_evolution. Pass in scop instead of only
the scop entry.
(stmt_has_simple_data_refs_p): Pass in scop instead of only the
scop entry.
(stmt_simple_for_scop_p): Same.
(harmful_stmt_in_bb): Same.
(graphite_can_represent_loop): Deleted.
(struct scopdet_info): Deleted.
(scopdet_basic_block_info): Deleted.
(build_scops_1): Deleted.
(bb_in_sd_region): Deleted.
(find_single_entry_edge): Deleted.
(find_single_exit_edge): Deleted.
(create_single_entry_edge): Deleted.
(sd_region_without_exit): Deleted.
(create_single_exit_edge): Deleted.
(unmark_exit_edges): Deleted.
(mark_exit_edges): Deleted.
(create_sese_edges): Deleted.
(build_graphite_scops): Deleted.
(canonicalize_loop_closed_ssa): Recompute all dominators at the
end.
(build_scops): Use the new scop_builder to build scops.
(dot_all_scops_1): Use the new pretty printer. Print loop father
as well.
(loop_body_is_valid_scop): New. Return true if loop body is a
valid scop.
(class scop_builder): New. Builds SCoPs for polyhedral
optimizatios.
(scop_builder): New. Constructor.
(static sese_l invalid_sese): sese_l with invalid edges.
(get_sese): Get an sese (from a loop) if possible, invalid_sese
otherwise.
(get_nearest_dom_with_single_entry): Get nearest dominator of a
basic_block with single entry. Return NULL if we get to the
beginning of a function.
(get_nearest_pdom_with_single_exit): Get nearest post-dominator of
a basic_block with single exit. Return NULL if we get to the
beginning of a function.
(print_sese): Pretty-print SESE.
(merge_sese): Merge two SESEs if possible and return the new SESE.
(build_scop_depth): Start building the SCoP within a loop nest.
(build_scop_breadth): Start building the SCoP at a single loop
depth. Merge adjacent SESEs if valid.
(can_represent_loop_1): Returns true if Graphite can represent
loop inside SCoP. Helper for can_represent_loop.
(can_represent_loop): Returns true if Graphite can represent LOOP
and all its nested loops in SCoP.
(loop_is_valid_scop): Returns true if LOOP and all its nests
constitute a valid SCoP.
(region_has_one_loop): Returns true of a region has only one loop.
(add_scop): Add SCoP to the list of valid scops. Removes an
already existing scop if it intersects with or subsumed by this
one.
(harmful_stmt_in_region): Returns true if SCoP has any statment
which cannot be represented by Graphite.
(subsumes): Returns true of SCoP S1 subsumes SCoP S2.
(remove_subscops): Remove any SCoP from the list of already found
SCoPs, if subsumed by S1.
(intersects): Return true if region bounded by SCoPs S1 and S2
intersect.
(remove_intersecting_scops): Remove any SCoP which intersects with
S1.
* graphite.c (print_graphite_scop_statistics):
(print_graphite_statistics): Print SCoP info while debugging.
(graphite_initialize): Early exit in case number of loops in a
function is less than PARAM_GRAPHITE_MIN_LOOPS_PER_FUNCTION or
basic blocks are more than PARAM_GRAPHITE_MAX_BBS_PER_FUNCTION.
(graphite_finalize):
* params.def: Add PARAM_GRAPHITE_MIN_LOOPS_PER_FUNCTION.
* sese.h (sese_loop_depth): Remove unnecessary gcc_assert.
(recompute_all_dominators): Recalculate POST_DOMINATORS.
* tree-cfg.c (print_loops): Print the function name while printing
loops.
gcc/testsuite/ChangeLog:
2015-09-27 Aditya Kumar <hiraditya@msn.com>
Sebastian Pop <s.pop@samsung.com>
* gcc.dg/graphite/block-1.c: Modified to match the pattern.
* gcc.dg/graphite/block-3.c: Same.
* gcc.dg/graphite/block-4.c: Same.
* gcc.dg/graphite/block-5.c: Same.
* gcc.dg/graphite/block-6.c: Same.
* gcc.dg/graphite/block-7.c: Same.
* gcc.dg/graphite/block-8.c: Same.
* gcc.dg/graphite/block-pr47654.c: Same.
* gcc.dg/graphite/interchange-0.c: Same.
* gcc.dg/graphite/interchange-1.c: Same.
* gcc.dg/graphite/interchange-10.c: Same.
* gcc.dg/graphite/interchange-11.c: Same.
* gcc.dg/graphite/interchange-12.c: Same.
* gcc.dg/graphite/interchange-13.c: Same.
* gcc.dg/graphite/interchange-14.c: Same.
* gcc.dg/graphite/interchange-15.c: Same.
* gcc.dg/graphite/interchange-3.c: Same.
* gcc.dg/graphite/interchange-4.c: Same.
* gcc.dg/graphite/interchange-5.c: Same.
* gcc.dg/graphite/interchange-6.c: Same.
* gcc.dg/graphite/interchange-7.c: Same.
* gcc.dg/graphite/interchange-8.c: Same.
* gcc.dg/graphite/interchange-9.c: Same.
* gcc.dg/graphite/interchange-mvt.c: Same.
* gcc.dg/graphite/pr35356-1.c (foo): Same.
* gcc.dg/graphite/pr35356-3.c: Same.
* gcc.dg/graphite/pr37485.c: Same.
* gcc/testsuite/gcc.dg/graphite/run-id-pr67700-1.c: New test case.
* gcc.dg/graphite/scop-1.c (int toto): Modified to match the pattern.
* gcc.dg/graphite/scop-11.c: Same.
* gcc.dg/graphite/scop-5.c: Same.
* gcc.dg/graphite/uns-block-1.c: Same.
* gcc.dg/graphite/uns-interchange-9.c: Same.
* gfortran.dg/graphite/block-1.f90: Same.
* gfortran.dg/graphite/interchange-3.f90: Same.
* gfortran.dg/graphite/pr14741.f90: Same.
From-SVN: r228215
2015-09-28 19:30:09 +02:00
/* Maximal number of basic blocks in the functions analyzed by Graphite. */
DEFPARAM ( PARAM_GRAPHITE_MIN_LOOPS_PER_FUNCTION ,
"graphite-min-loops-per-function" ,
2015-10-22 00:24:41 +02:00
"minimal number of loops per function to be analyzed by Graphite." ,
Redesign Graphite scop detection
Redesign Graphite scop detection for faster compiler time and detecting more SCoPs.
Existing algorithm for SCoP detection in graphite was based on dominator tree
where a tree (CFG) traversal was required for analyzing an SESE. The tree
traversal is linear in the number of basic blocks and SCoP detection is
(probably) linear in number of instructions. That algorithm utilized a generic
infrastructure of SESE which does not directly represent loops. With regards to
graphite framework, we are only interested in subtrees with loops. The new
algorithm is geared towards tree traversal on loop structure. The algorithm is
linear in number of loops which is faster than the previous algorithm.
Briefly, we start the traversal at a loop-nest and analyze it recursively for
validity. Once a valid loop is found we find a valid adjacent loop. If an
adjacent loop is found and is valid, we merge both loop nests otherwise we form
a SCoP from the previous loop nest, and resume the algorithm from the adjacent
loop nest. The data structure to represent an SESE is an ordered pair of edges
(entry, exit). The new algoritm can extend a SCoP in both the directions. With
this approach, the number of instructions to be analyzed for validity reduces to
a minimal set. We start by analyzing those statements which are inside a loop,
because validity of those statements is necessary for the validity of loop. The
statements outside the loop nest can be just excluded from the SESE if they are
not valid.
This patch depends on: https://gcc.gnu.org/ml/gcc-patches/2015-09/msg02024.html
Passes (c,c++,fortran) regtest and bootstrap.
gcc/ChangeLog:
2015-09-27 Aditya Kumar <hiraditya@msn.com>
Sebastian Pop <s.pop@samsung.com>
* graphite-optimize-isl.c (optimize_isl):
* graphite-scop-detection.c (struct sese_l): New type.
(get_entry_bb): API for getting entry bb of SESE.
(get_exit_bb): API for getting exit bb of SESE.
(class debug_printer): New type. Simple printer in debug mode.
(trivially_empty_bb_p): New. Return true when BB is empty or
contains only debug instructions.
(graphite_can_represent_expr): Call scalar_evoution_in_region
instead of analyze_scalar_evolution. Pass in scop instead of only
the scop entry.
(stmt_has_simple_data_refs_p): Pass in scop instead of only the
scop entry.
(stmt_simple_for_scop_p): Same.
(harmful_stmt_in_bb): Same.
(graphite_can_represent_loop): Deleted.
(struct scopdet_info): Deleted.
(scopdet_basic_block_info): Deleted.
(build_scops_1): Deleted.
(bb_in_sd_region): Deleted.
(find_single_entry_edge): Deleted.
(find_single_exit_edge): Deleted.
(create_single_entry_edge): Deleted.
(sd_region_without_exit): Deleted.
(create_single_exit_edge): Deleted.
(unmark_exit_edges): Deleted.
(mark_exit_edges): Deleted.
(create_sese_edges): Deleted.
(build_graphite_scops): Deleted.
(canonicalize_loop_closed_ssa): Recompute all dominators at the
end.
(build_scops): Use the new scop_builder to build scops.
(dot_all_scops_1): Use the new pretty printer. Print loop father
as well.
(loop_body_is_valid_scop): New. Return true if loop body is a
valid scop.
(class scop_builder): New. Builds SCoPs for polyhedral
optimizatios.
(scop_builder): New. Constructor.
(static sese_l invalid_sese): sese_l with invalid edges.
(get_sese): Get an sese (from a loop) if possible, invalid_sese
otherwise.
(get_nearest_dom_with_single_entry): Get nearest dominator of a
basic_block with single entry. Return NULL if we get to the
beginning of a function.
(get_nearest_pdom_with_single_exit): Get nearest post-dominator of
a basic_block with single exit. Return NULL if we get to the
beginning of a function.
(print_sese): Pretty-print SESE.
(merge_sese): Merge two SESEs if possible and return the new SESE.
(build_scop_depth): Start building the SCoP within a loop nest.
(build_scop_breadth): Start building the SCoP at a single loop
depth. Merge adjacent SESEs if valid.
(can_represent_loop_1): Returns true if Graphite can represent
loop inside SCoP. Helper for can_represent_loop.
(can_represent_loop): Returns true if Graphite can represent LOOP
and all its nested loops in SCoP.
(loop_is_valid_scop): Returns true if LOOP and all its nests
constitute a valid SCoP.
(region_has_one_loop): Returns true of a region has only one loop.
(add_scop): Add SCoP to the list of valid scops. Removes an
already existing scop if it intersects with or subsumed by this
one.
(harmful_stmt_in_region): Returns true if SCoP has any statment
which cannot be represented by Graphite.
(subsumes): Returns true of SCoP S1 subsumes SCoP S2.
(remove_subscops): Remove any SCoP from the list of already found
SCoPs, if subsumed by S1.
(intersects): Return true if region bounded by SCoPs S1 and S2
intersect.
(remove_intersecting_scops): Remove any SCoP which intersects with
S1.
* graphite.c (print_graphite_scop_statistics):
(print_graphite_statistics): Print SCoP info while debugging.
(graphite_initialize): Early exit in case number of loops in a
function is less than PARAM_GRAPHITE_MIN_LOOPS_PER_FUNCTION or
basic blocks are more than PARAM_GRAPHITE_MAX_BBS_PER_FUNCTION.
(graphite_finalize):
* params.def: Add PARAM_GRAPHITE_MIN_LOOPS_PER_FUNCTION.
* sese.h (sese_loop_depth): Remove unnecessary gcc_assert.
(recompute_all_dominators): Recalculate POST_DOMINATORS.
* tree-cfg.c (print_loops): Print the function name while printing
loops.
gcc/testsuite/ChangeLog:
2015-09-27 Aditya Kumar <hiraditya@msn.com>
Sebastian Pop <s.pop@samsung.com>
* gcc.dg/graphite/block-1.c: Modified to match the pattern.
* gcc.dg/graphite/block-3.c: Same.
* gcc.dg/graphite/block-4.c: Same.
* gcc.dg/graphite/block-5.c: Same.
* gcc.dg/graphite/block-6.c: Same.
* gcc.dg/graphite/block-7.c: Same.
* gcc.dg/graphite/block-8.c: Same.
* gcc.dg/graphite/block-pr47654.c: Same.
* gcc.dg/graphite/interchange-0.c: Same.
* gcc.dg/graphite/interchange-1.c: Same.
* gcc.dg/graphite/interchange-10.c: Same.
* gcc.dg/graphite/interchange-11.c: Same.
* gcc.dg/graphite/interchange-12.c: Same.
* gcc.dg/graphite/interchange-13.c: Same.
* gcc.dg/graphite/interchange-14.c: Same.
* gcc.dg/graphite/interchange-15.c: Same.
* gcc.dg/graphite/interchange-3.c: Same.
* gcc.dg/graphite/interchange-4.c: Same.
* gcc.dg/graphite/interchange-5.c: Same.
* gcc.dg/graphite/interchange-6.c: Same.
* gcc.dg/graphite/interchange-7.c: Same.
* gcc.dg/graphite/interchange-8.c: Same.
* gcc.dg/graphite/interchange-9.c: Same.
* gcc.dg/graphite/interchange-mvt.c: Same.
* gcc.dg/graphite/pr35356-1.c (foo): Same.
* gcc.dg/graphite/pr35356-3.c: Same.
* gcc.dg/graphite/pr37485.c: Same.
* gcc/testsuite/gcc.dg/graphite/run-id-pr67700-1.c: New test case.
* gcc.dg/graphite/scop-1.c (int toto): Modified to match the pattern.
* gcc.dg/graphite/scop-11.c: Same.
* gcc.dg/graphite/scop-5.c: Same.
* gcc.dg/graphite/uns-block-1.c: Same.
* gcc.dg/graphite/uns-interchange-9.c: Same.
* gfortran.dg/graphite/block-1.f90: Same.
* gfortran.dg/graphite/interchange-3.f90: Same.
* gfortran.dg/graphite/pr14741.f90: Same.
From-SVN: r228215
2015-09-28 19:30:09 +02:00
2 , 0 , 0 )
2015-09-09 06:20:47 +02:00
DEFPARAM ( PARAM_MAX_ISL_OPERATIONS ,
"max-isl-operations" ,
2015-12-17 19:50:05 +01:00
"maximum number of isl operations, 0 means unlimited" ,
2015-09-09 06:20:47 +02:00
350000 , 0 , 0 )
2012-01-27 14:47:41 +01:00
/* Avoid data dependence analysis on very large loops. */
DEFPARAM ( PARAM_LOOP_MAX_DATAREFS_FOR_DATADEPS ,
"loop-max-datarefs-for-datadeps" ,
2015-10-22 00:24:41 +02:00
"Maximum number of datarefs in loop for building loop data dependencies." ,
2012-01-27 14:47:41 +01:00
1000 , 0 , 0 )
2009-02-20 13:56:01 +01:00
/* Avoid doing loop invariant motion on very large loops. */
DEFPARAM ( PARAM_LOOP_INVARIANT_MAX_BBS_IN_LOOP ,
"loop-invariant-max-bbs-in-loop" ,
2015-10-22 00:24:41 +02:00
"Max basic blocks number in loop for loop invariant motion." ,
2009-02-20 13:56:01 +01:00
10000 , 0 , 0 )
2014-07-26 02:06:56 +02:00
/ * When the parameter is 1 , use the internal function id
to look up for profile data . Otherwise , use a more stable
external id based on assembler name and source location . * /
DEFPARAM ( PARAM_PROFILE_FUNC_INTERNAL_ID ,
"profile-func-internal-id" ,
2015-10-22 00:24:41 +02:00
"use internal function id in profile lookup." ,
2014-07-26 02:06:56 +02:00
0 , 0 , 1 )
2014-10-07 06:06:12 +02:00
/ * When the parameter is 1 , track the most frequent N target
addresses in indirect - call profile . This disables
indirect_call_profiler_v2 which tracks single target . * /
DEFPARAM ( PARAM_INDIR_CALL_TOPN_PROFILE ,
"indir-call-topn-profile" ,
2015-10-22 00:24:41 +02:00
"track topn target addresses in indirect-call profile." ,
2014-10-07 06:06:12 +02:00
0 , 0 , 1 )
2014-07-26 02:06:56 +02:00
passes.texi (Tree-SSA passes): Document SLP pass.
* doc/passes.texi (Tree-SSA passes): Document SLP pass.
* tree-pass.h (pass_slp_vectorize): New pass.
* params.h (SLP_MAX_INSNS_IN_BB): Define.
* timevar.def (TV_TREE_SLP_VECTORIZATION): Define.
* tree-vectorizer.c (timevar.h): Include.
(user_vect_verbosity_level): Declare.
(vect_location): Fix comment.
(vect_set_verbosity_level): Update user_vect_verbosity_level
instead of vect_verbosity_level.
(vect_set_dump_settings): Add an argument. Ignore user defined
verbosity if dump flags require higher level of verbosity. Print to
stderr only for loop vectorization.
(vectorize_loops): Update call to vect_set_dump_settings.
(execute_vect_slp): New function.
(gate_vect_slp): Likewise.
(struct gimple_opt_pass pass_slp_vectorize): New.
* tree-vectorizer.h (struct _bb_vec_info): Define along macros to
access its members.
(vec_info_for_bb): New function.
(struct _stmt_vec_info): Add bb_vinfo and a macro for its access.
(VECTORIZATION_ENABLED): New macro.
(SLP_ENABLED, SLP_DISABLED): Likewise.
(vect_is_simple_use): Add bb_vec_info argument.
(new_stmt_vec_info, vect_analyze_data_ref_dependences,
vect_analyze_data_refs_alignment, vect_verify_datarefs_alignment,
vect_analyze_data_ref_accesses, vect_analyze_data_refs,
vect_schedule_slp, vect_analyze_slp): Likewise.
(vect_analyze_stmt): Add slp_tree argument.
(find_bb_location): Declare.
(vect_slp_analyze_bb, vect_slp_transform_bb): Likewise.
* tree-vect-loop.c (new_loop_vec_info): Adjust function calls.
(vect_analyze_loop_operations, vect_analyze_loop,
get_initial_def_for_induction, vect_create_epilog_for_reduction,
vect_finalize_reduction, vectorizable_reduction,
vectorizable_live_operation, vect_transform_loop): Likewise.
* tree-data-ref.c (dr_analyze_innermost): Update comment,
skip evolution analysis if analyzing a basic block.
(dr_analyze_indices): Likewise.
(initialize_data_dependence_relation): Skip the test whether the
object is invariant for basic blocks.
(compute_all_dependences): Skip dependence analysis for data
references in basic blocks.
(find_data_references_in_stmt): Don't fail in case of invariant
access in basic block.
(find_data_references_in_bb): New function.
(find_data_references_in_loop): Move code to
find_data_references_in_bb and add a call to it.
(compute_data_dependences_for_bb): New function.
* tree-data-ref.h (compute_data_dependences_for_bb): Declare.
* tree-vect-data-refs.c (vect_check_interleaving): Adjust to the case
that STEP is 0.
(vect_analyze_data_ref_dependence): Check for interleaving in case of
unknown dependence in basic block and fail in case of dependence in
basic block.
(vect_analyze_data_ref_dependences): Add bb_vinfo argument, get data
dependence instances from either loop or basic block vectorization
info.
(vect_compute_data_ref_alignment): Check if it is loop vectorization
before calling nested_in_vect_loop_p.
(vect_compute_data_refs_alignment): Add bb_vinfo argument, get data
dependence instances from either loop or basic block vectorization
info.
(vect_verify_datarefs_alignment): Likewise.
(vect_enhance_data_refs_alignment): Adjust function calls.
(vect_analyze_data_refs_alignment): Likewise.
(vect_analyze_group_access): Fix printing. Skip different checks if
DR_STEP is 0. Keep strided stores either in loop or basic block
vectorization data structure. Fix indentation.
(vect_analyze_data_ref_access): Fix comments, allow zero step in
basic blocks.
(vect_analyze_data_ref_accesses): Add bb_vinfo argument, get data
dependence instances from either loop or basic block vectorization
info.
(vect_analyze_data_refs): Update comment. Call
compute_data_dependences_for_bb to analyze basic blocks.
(vect_create_addr_base_for_vector_ref): Check for outer loop only in
case of loop vectorization. In case of basic block vectorization use
data-ref itself as a base.
(vect_create_data_ref_ptr): In case of basic block vectorization:
don't advance the pointer, add new statements before the current
statement. Adjust function calls.
(vect_supportable_dr_alignment): Support only aligned accesses in
basic block vectorization.
* common.opt (ftree-slp-vectorize): New flag.
* tree-vect-patterns.c (widened_name_p): Adjust function calls.
(vect_pattern_recog_1): Likewise.
* tree-vect-stmts.c (process_use): Likewise.
(vect_init_vector): Add new statements in the beginning of the basic
block in case of basic block SLP.
(vect_get_vec_def_for_operand): Adjust function calls.
(vect_finish_stmt_generation): Likewise.
(vectorizable_call): Add assert that it is loop vectorization, adjust
function calls.
(vectorizable_conversion, vectorizable_assignment): Likewise.
(vectorizable_operation): In case of basic block SLP, take
vectorization factor from statement's type and skip the relevance
check. Adjust function calls.
(vectorizable_type_demotion): Add assert that it is loop
vectorization, adjust function calls.
(vectorizable_type_promotion): Likewise.
(vectorizable_store): Check for outer loop only in case of loop
vectorization. Adjust function calls. For basic blocks, skip the
relevance check and don't advance pointers.
(vectorizable_load): Likewise.
(vectorizable_condition): Add assert that it is loop vectorization,
adjust function calls.
(vect_analyze_stmt): Add argument. In case of basic block SLP, check
that it is not reduction, get vector type, call only supported
functions, skip loop specific parts.
(vect_transform_stmt): Check for outer loop only in case of loop
vectorization.
(new_stmt_vec_info): Add new argument and initialize bb_vinfo.
(vect_is_simple_use): Fix comment, add new argument, fix conditions
for external definition.
* passes.c (pass_slp_vectorize): New pass.
* tree-vect-slp.c (find_bb_location): New function.
(vect_get_and_check_slp_defs): Add argument, adjust function calls,
check for patterns only in loops.
(vect_build_slp_tree): Add argument, adjust function calls, fail in
case of multiple types in basic block SLP.
(vect_mark_slp_stmts_relevant): New function.
(vect_supported_load_permutation_p): Fix comment.
(vect_analyze_slp_instance): Add argument. In case of basic block
SLP, take vectorization factor from statement's type, check that
unrolling factor is 1. Adjust function call. Save SLP instance in
either loop or basic block vectorization structure. Return FALSE,
if SLP failed.
(vect_analyze_slp): Add argument. Get strided stores groups from
either loop or basic block vectorization structure. Return FALSE
if basic block SLP failed.
(new_bb_vec_info): New function.
(destroy_bb_vec_info, vect_slp_analyze_node_operations,
vect_slp_analyze_operations, vect_slp_analyze_bb): Likewise.
(vect_schedule_slp): Add argument. Get SLP instances from either
loop or basic block vectorization structure. Set vectorization factor
to be 1 for basic block SLP.
(vect_slp_transform_bb): New function.
* params.def (PARAM_SLP_MAX_INSNS_IN_BB): Define.
From-SVN: r147829
2009-05-24 10:44:56 +02:00
/* Avoid SLP vectorization of large basic blocks. */
DEFPARAM ( PARAM_SLP_MAX_INSNS_IN_BB ,
"slp-max-insns-in-bb" ,
2015-10-22 00:24:41 +02:00
"Maximum number of instructions in basic block to be considered for SLP vectorization." ,
passes.texi (Tree-SSA passes): Document SLP pass.
* doc/passes.texi (Tree-SSA passes): Document SLP pass.
* tree-pass.h (pass_slp_vectorize): New pass.
* params.h (SLP_MAX_INSNS_IN_BB): Define.
* timevar.def (TV_TREE_SLP_VECTORIZATION): Define.
* tree-vectorizer.c (timevar.h): Include.
(user_vect_verbosity_level): Declare.
(vect_location): Fix comment.
(vect_set_verbosity_level): Update user_vect_verbosity_level
instead of vect_verbosity_level.
(vect_set_dump_settings): Add an argument. Ignore user defined
verbosity if dump flags require higher level of verbosity. Print to
stderr only for loop vectorization.
(vectorize_loops): Update call to vect_set_dump_settings.
(execute_vect_slp): New function.
(gate_vect_slp): Likewise.
(struct gimple_opt_pass pass_slp_vectorize): New.
* tree-vectorizer.h (struct _bb_vec_info): Define along macros to
access its members.
(vec_info_for_bb): New function.
(struct _stmt_vec_info): Add bb_vinfo and a macro for its access.
(VECTORIZATION_ENABLED): New macro.
(SLP_ENABLED, SLP_DISABLED): Likewise.
(vect_is_simple_use): Add bb_vec_info argument.
(new_stmt_vec_info, vect_analyze_data_ref_dependences,
vect_analyze_data_refs_alignment, vect_verify_datarefs_alignment,
vect_analyze_data_ref_accesses, vect_analyze_data_refs,
vect_schedule_slp, vect_analyze_slp): Likewise.
(vect_analyze_stmt): Add slp_tree argument.
(find_bb_location): Declare.
(vect_slp_analyze_bb, vect_slp_transform_bb): Likewise.
* tree-vect-loop.c (new_loop_vec_info): Adjust function calls.
(vect_analyze_loop_operations, vect_analyze_loop,
get_initial_def_for_induction, vect_create_epilog_for_reduction,
vect_finalize_reduction, vectorizable_reduction,
vectorizable_live_operation, vect_transform_loop): Likewise.
* tree-data-ref.c (dr_analyze_innermost): Update comment,
skip evolution analysis if analyzing a basic block.
(dr_analyze_indices): Likewise.
(initialize_data_dependence_relation): Skip the test whether the
object is invariant for basic blocks.
(compute_all_dependences): Skip dependence analysis for data
references in basic blocks.
(find_data_references_in_stmt): Don't fail in case of invariant
access in basic block.
(find_data_references_in_bb): New function.
(find_data_references_in_loop): Move code to
find_data_references_in_bb and add a call to it.
(compute_data_dependences_for_bb): New function.
* tree-data-ref.h (compute_data_dependences_for_bb): Declare.
* tree-vect-data-refs.c (vect_check_interleaving): Adjust to the case
that STEP is 0.
(vect_analyze_data_ref_dependence): Check for interleaving in case of
unknown dependence in basic block and fail in case of dependence in
basic block.
(vect_analyze_data_ref_dependences): Add bb_vinfo argument, get data
dependence instances from either loop or basic block vectorization
info.
(vect_compute_data_ref_alignment): Check if it is loop vectorization
before calling nested_in_vect_loop_p.
(vect_compute_data_refs_alignment): Add bb_vinfo argument, get data
dependence instances from either loop or basic block vectorization
info.
(vect_verify_datarefs_alignment): Likewise.
(vect_enhance_data_refs_alignment): Adjust function calls.
(vect_analyze_data_refs_alignment): Likewise.
(vect_analyze_group_access): Fix printing. Skip different checks if
DR_STEP is 0. Keep strided stores either in loop or basic block
vectorization data structure. Fix indentation.
(vect_analyze_data_ref_access): Fix comments, allow zero step in
basic blocks.
(vect_analyze_data_ref_accesses): Add bb_vinfo argument, get data
dependence instances from either loop or basic block vectorization
info.
(vect_analyze_data_refs): Update comment. Call
compute_data_dependences_for_bb to analyze basic blocks.
(vect_create_addr_base_for_vector_ref): Check for outer loop only in
case of loop vectorization. In case of basic block vectorization use
data-ref itself as a base.
(vect_create_data_ref_ptr): In case of basic block vectorization:
don't advance the pointer, add new statements before the current
statement. Adjust function calls.
(vect_supportable_dr_alignment): Support only aligned accesses in
basic block vectorization.
* common.opt (ftree-slp-vectorize): New flag.
* tree-vect-patterns.c (widened_name_p): Adjust function calls.
(vect_pattern_recog_1): Likewise.
* tree-vect-stmts.c (process_use): Likewise.
(vect_init_vector): Add new statements in the beginning of the basic
block in case of basic block SLP.
(vect_get_vec_def_for_operand): Adjust function calls.
(vect_finish_stmt_generation): Likewise.
(vectorizable_call): Add assert that it is loop vectorization, adjust
function calls.
(vectorizable_conversion, vectorizable_assignment): Likewise.
(vectorizable_operation): In case of basic block SLP, take
vectorization factor from statement's type and skip the relevance
check. Adjust function calls.
(vectorizable_type_demotion): Add assert that it is loop
vectorization, adjust function calls.
(vectorizable_type_promotion): Likewise.
(vectorizable_store): Check for outer loop only in case of loop
vectorization. Adjust function calls. For basic blocks, skip the
relevance check and don't advance pointers.
(vectorizable_load): Likewise.
(vectorizable_condition): Add assert that it is loop vectorization,
adjust function calls.
(vect_analyze_stmt): Add argument. In case of basic block SLP, check
that it is not reduction, get vector type, call only supported
functions, skip loop specific parts.
(vect_transform_stmt): Check for outer loop only in case of loop
vectorization.
(new_stmt_vec_info): Add new argument and initialize bb_vinfo.
(vect_is_simple_use): Fix comment, add new argument, fix conditions
for external definition.
* passes.c (pass_slp_vectorize): New pass.
* tree-vect-slp.c (find_bb_location): New function.
(vect_get_and_check_slp_defs): Add argument, adjust function calls,
check for patterns only in loops.
(vect_build_slp_tree): Add argument, adjust function calls, fail in
case of multiple types in basic block SLP.
(vect_mark_slp_stmts_relevant): New function.
(vect_supported_load_permutation_p): Fix comment.
(vect_analyze_slp_instance): Add argument. In case of basic block
SLP, take vectorization factor from statement's type, check that
unrolling factor is 1. Adjust function call. Save SLP instance in
either loop or basic block vectorization structure. Return FALSE,
if SLP failed.
(vect_analyze_slp): Add argument. Get strided stores groups from
either loop or basic block vectorization structure. Return FALSE
if basic block SLP failed.
(new_bb_vec_info): New function.
(destroy_bb_vec_info, vect_slp_analyze_node_operations,
vect_slp_analyze_operations, vect_slp_analyze_bb): Likewise.
(vect_schedule_slp): Add argument. Get SLP instances from either
loop or basic block vectorization structure. Set vectorization factor
to be 1 for basic block SLP.
(vect_slp_transform_bb): New function.
* params.def (PARAM_SLP_MAX_INSNS_IN_BB): Define.
From-SVN: r147829
2009-05-24 10:44:56 +02:00
1000 , 0 , 0 )
2009-06-08 18:00:13 +02:00
DEFPARAM ( PARAM_MIN_INSN_TO_PREFETCH_RATIO ,
"min-insn-to-prefetch-ratio" ,
2009-11-18 12:09:50 +01:00
"Min. ratio of insns to prefetches to enable prefetching for "
2015-10-22 00:24:41 +02:00
"a loop with an unknown trip count." ,
2011-06-27 12:28:39 +02:00
9 , 0 , 0 )
2009-06-08 18:00:13 +02:00
DEFPARAM ( PARAM_PREFETCH_MIN_INSN_TO_MEM_RATIO ,
"prefetch-min-insn-to-mem-ratio" ,
2015-10-22 00:24:41 +02:00
"Min. ratio of insns to mem ops to enable prefetching in a loop." ,
2009-06-08 18:00:13 +02:00
3 , 0 , 0 )
2010-02-16 09:02:37 +01:00
/* Set maximum hash table size for var tracking. */
DEFPARAM ( PARAM_MAX_VARTRACK_SIZE ,
"max-vartrack-size" ,
2015-10-22 00:24:41 +02:00
"Max. size of var tracking hash tables." ,
2010-02-16 09:02:37 +01:00
50000000 , 0 , 0 )
2011-05-31 18:25:50 +02:00
/ * Set maximum recursion depth for var tracking expression expansion
and resolution . * /
DEFPARAM ( PARAM_MAX_VARTRACK_EXPR_DEPTH ,
"max-vartrack-expr-depth" ,
2015-10-22 00:24:41 +02:00
"Max. recursion depth for expanding var tracking expressions." ,
2011-07-21 00:30:06 +02:00
12 , 0 , 0 )
2011-05-31 18:25:50 +02:00
2013-01-03 09:52:10 +01:00
/ * Set maximum length of value location list for which var tracking
should add reverse operations . * /
DEFPARAM ( PARAM_MAX_VARTRACK_REVERSE_OP_SIZE ,
"max-vartrack-reverse-op-size" ,
2015-10-22 00:24:41 +02:00
"Max. size of loc list for which reverse ops should be added." ,
2013-01-03 09:52:10 +01:00
50 , 0 , 0 )
invoke.texi (-fvar-tracking-assignments): New.
gcc/ChangeLog:
* doc/invoke.texi (-fvar-tracking-assignments): New.
(-fvar-tracking-assignments-toggle): New.
(-fdump-final-insns=file): Mark filename as optional.
(--param min-nondebug-insn-uid): New.
(-gdwarf-@{version}): Mention version 4.
* opts.c (common_handle_option): Accept it.
* tree-vrp.c (find_assert_locations_1): Skip debug stmts.
* regrename.c (regrename_optimize): Drop last. Don't count debug
insns as uses. Don't reject change because of debug insn.
(do_replace): Reject DEBUG_INSN as chain starter. Take base_regno
from the chain starter, and check for inexact matches in
DEBUG_INSNS.
(scan_rtx_reg): Accept inexact matches in DEBUG_INSNs.
(build_def_use): Simplify and fix the marking of DEBUG_INSNs.
* sched-ebb.c (schedule_ebbs): Skip boundary debug insns.
* fwprop.c (forward_propagate_and_simplify): ...into debug insns.
* doc/gimple.texi (is_gimple_debug): New.
(gimple_debug_bind_p): New.
(is_gimple_call, gimple_assign_cast_p): End sentence with period.
* doc/install.texi (bootstrap-debug): More details.
(bootstrap-debug-big, bootstrap-debug-lean): Document.
(bootstrap-debug-lib): More details.
(bootstrap-debug-ckovw): Update.
(bootstrap-time): New.
* tree-into-ssa.c (mark_def_sites): Skip debug stmts.
(insert_phi_nodes_for): Insert debug stmts.
(rewrite_stmt): Take iterator. Insert debug stmts.
(rewrite_enter_block): Adjust.
(maybe_replace_use_in_debug_stmt): New.
(rewrite_update_stmt): Use it.
(mark_use_interesting): Return early for debug stmts.
* tree-ssa-loop-im.c (rewrite_bittest): Propagate DEFs into debug
stmts before replacing stmt.
(move_computations_stmt): Likewise.
* ira-conflicts.c (add_copies): Skip debug insns.
* regstat.c (regstat_init_n_sets_and_refs): Discount debug insns.
(regstat_bb_compute_ri): Skip debug insns.
* tree-ssa-threadupdate.c (redirection_block_p): Skip debug stmts.
* tree-ssa-loop-manip.c (find_uses_to_rename_stmt,
check_loop_closed_ssa_stmt): Skip debug stmts.
* tree-tailcall.c (find_tail_calls): Likewise.
* tree-ssa-loop-ch.c (should_duplicate_loop_header_p): Likewise.
* tree.h (MAY_HAVE_DEBUG_STMTS): New.
(build_var_debug_value_stat): Declare.
(build_var_debug_value): Define.
(target_for_debug_bind): Declare.
* reload.c (find_equiv_reg): Skip debug insns.
* rtlanal.c (reg_used_between_p): Skip debug insns.
(side_effects_p): Likewise.
(canonicalize_condition): Likewise.
* ddg.c (create_ddg_dep_from_intra_loop_link): Check that non-debug
insns never depend on debug insns.
(create_ddg_dep_no_link): Likewise.
(add_cross_iteration_register_deps): Use ANTI_DEP for debug insns.
Don't add inter-loop dependencies for debug insns.
(build_intra_loop_deps): Likewise.
(create_ddg): Count debug insns.
* ddg.h (struct ddg::num_debug): New.
(num_backargs): Pair up with previous int field.
* diagnostic.c (diagnostic_report_diagnostic): Skip notes on
-fcompare-debug-second.
* final.c (get_attr_length_1): Skip debug insns.
(rest_of_clean-state): Don't dump CFA_RESTORE_STATE.
* gcc.c (invoke_as): Call compare-debug-dump-opt.
(driver_self_specs): Map -fdump-final-insns to
-fdump-final-insns=..
(get_local_tick): New.
(compare_debug_dump_opt_spec_function): Test for . argument and
compute output name. Compute temp output spec without flag name.
Compute -frandom-seed.
(OPT): Undef after use.
* cfgloopanal.c (num_loop_insns): Skip debug insns.
(average_num_loop_insns): Likewise.
* params.h (MIN_NONDEBUG_INSN_UID): New.
* gimple.def (GIMPLE_DEBUG): New.
* ipa-reference.c (scan_stmt_for_static_refs): Skip debug stmts.
* auto-inc-dec.c (merge_in_block): Skip debug insns.
(merge_in_block): Fix whitespace.
* toplev.c (flag_var_tracking): Update comment.
(flag_var_tracking_assignments): New.
(flag_var_tracking_assignments_toggle): New.
(process_options): Don't open final insns dump file if we're not
going to write to it. Compute defaults for var_tracking.
* df-scan.c (df_insn_rescan_debug_internal): New.
(df_uses_record): Handle debug insns.
* haifa-sched.c (ready): Initialize n_debug.
(contributes_to_priority): Skip debug insns.
(dep_list_size): New.
(priority): Use it.
(rank_for_schedule): Likewise. Schedule debug insns as soon as
they're ready. Disregard previous debug insns to make decisions.
(queue_insn): Never queue debug insns.
(ready_add, ready_remove_first, ready_remove): Count debug insns.
(schedule_insn): Don't reject debug insns because of issue rate.
(get_ebb_head_tail, no_real_insns_p): Skip boundary debug insns.
(queue_to_ready): Skip and discount debug insns.
(choose_ready): Let debug insns through.
(schedule_block): Check boundary debug insns. Discount debug
insns, schedule them early. Adjust whitespace.
(set_priorities): Check for boundary debug insns.
(add_jump_dependencies): Use dep_list_size.
(prev_non_location_insn): New.
(check_cfg): Use it.
* tree-ssa-loop-ivopts.c (find-interesting_users): Skip debug
stmts.
(remove_unused_ivs): Reset debug stmts.
* modulo-sched.c (const_iteration_count): Skip debug insns.
(res_MII): Discount debug insns.
(loop_single_full_bb_p): Skip debug insns.
(sms_schedule): Likewise.
(sms_schedule_by_order): Likewise.
(ps_has_conflicts): Likewise.
* caller-save.c (refmarker_fn): New.
(save_call_clobbered_regs): Replace regs with saved mem in
debug insns.
(mark_referenced_regs): Take pointer, mark and arg. Adjust.
Call refmarker_fn mark for hardregnos.
(mark_reg_as_referenced): New.
(replace_reg_with_saved_mem): New.
* ipa-pure-const.c (check_stmt): Skip debug stmts.
* cse.c (cse_insn): Canonicalize debug insns. Skip them when
searching back.
(cse_extended_basic_block): Skip debug insns.
(count_reg_usage): Likewise.
(is_dead_reg): New, split out of...
(set_live_p): ... here.
(insn_live_p): Use it for debug insns.
* tree-stdarg.c (check_all_va_list_escapes): Skip debug stmts.
(execute_optimize_stdarg): Likewise.
* tree-ssa-dom.c (propagate_rhs_into_lhs): Likewise.
* tree-ssa-propagate.c (substitute_and_fold): Don't regard
changes in debug stmts as changes.
* sel-sched.c (moving_insn_creates_bookkeeping_block_p): New.
(moveup_expr): Don't move across debug insns. Don't move
debug insn if it would create a bookkeeping block.
(moveup_expr_cached): Don't use cache for debug insns that
are heads of blocks.
(compute_av_set_inside_bb): Skip debug insns.
(sel_rank_for_schedule): Schedule debug insns first. Remove
dead code.
(block_valid_for_bookkeeping_p); Support lax searches.
(create_block_for_bookkeeping): Adjust block numbers when
encountering debug-only blocks.
(find_place_for_bookkeeping): Deal with debug-only blocks.
(generate_bookkeeping_insn): Accept no place to insert.
(remove_temp_moveop_nops): New argument full_tidying.
(prepare_place_to_insert): Deal with debug insns.
(advance_state_on_fence): Debug insns don't start cycles.
(update_boundaries): Take fence as argument. Deal with
debug insns.
(schedule_expr_on_boundary): No full_tidying on debug insns.
(fill_insns): Deal with debug insns.
(track_scheduled_insns_and_blocks): Don't count debug insns.
(need_nop_to_preserve_insn_bb): New, split out of...
(remove_insn_from_stream): ... this.
(fur_orig_expr_not_found): Skip debug insns.
* rtl.def (VALUE): Move up.
(DEBUG_INSN): New.
* tree-ssa-sink.c (all_immediate_uses_same_place): Skip debug
stmts.
(nearest_common_dominator_of_uses): Take debug_stmts argument.
Set it if debug stmts are found.
(statement_sink_location): Skip debug stmts. Propagate
moving defs into debug stmts.
* ifcvt.c (first_active_insn): Skip debug insns.
(last_active_insns): Likewise.
(cond_exec_process_insns): Likewise.
(noce_process_if_block): Likewise.
(check_cond_move_block): Likewise.
(cond_move_convert_if_block): Likewise.
(block_jumps_and_fallthru_p): Likewise.
(dead_or_predicable): Likewise.
* dwarf2out.c (debug_str_hash_forced): New.
(find_AT_string): Add comment.
(gen_label_for_indirect_string): New.
(get_debug_string_label): New.
(AT_string_form): Use it.
(mem_loc_descriptor): Handle non-TLS symbols. Handle MINUS , DIV,
MOD, AND, IOR, XOR, NOT, ABS, NEG, and CONST_STRING. Accept but
discard COMPARE, IF_THEN_ELSE, ROTATE, ROTATERT, TRUNCATE and
several operations that cannot be represented with DWARF opcodes.
(loc_descriptor): Ignore SIGN_EXTEND and ZERO_EXTEND. Require
dwarf_version 4 for DW_OP_implicit_value and DW_OP_stack_value.
(dwarf2out_var_location): Take during-call mark into account.
(output_indirect_string): Update comment. Output if there are
label and references.
(prune_indirect_string): New.
(prune_unused_types): Call it if debug_str_hash_forced.
More in dwarf2out.c, from Jakub Jelinek <jakub@redhat.com>:
(dw_long_long_const): Remove.
(struct dw_val_struct): Change val_long_long type to rtx.
(print_die, attr_checksum, same_dw_val_p, loc_descriptor): Adjust for
val_long_long change to CONST_DOUBLE rtx from a long hi/lo pair.
(output_die): Likewise. Use HOST_BITS_PER_WIDE_INT size of each
component instead of HOST_BITS_PER_LONG.
(output_loc_operands): Likewise. For const8* assert
HOST_BITS_PER_WIDE_INT rather than HOST_BITS_PER_LONG is >= 64.
(output_loc_operands_raw): For const8* assert HOST_BITS_PER_WIDE_INT
rather than HOST_BITS_PER_LONG is >= 64.
(add_AT_long_long): Remove val_hi and val_lo arguments, add
val_const_double.
(size_of_die): Use HOST_BITS_PER_WIDE_INT size multiplier instead of
HOST_BITS_PER_LONG for dw_val_class_long_long.
(add_const_value_attribute): Adjust add_AT_long_long caller. Don't
handle TLS SYMBOL_REFs. If CONST wraps a constant, tail recurse.
(dwarf_stack_op_name): Handle DW_OP_implicit_value and
DW_OP_stack_value.
(size_of_loc_descr, output_loc_operands, output_loc_operands_raw):
Handle DW_OP_implicit_value.
(extract_int): Move prototype earlier.
(mem_loc_descriptor): For SUBREG punt if inner
mode size is wider than DWARF2_ADDR_SIZE. Handle SIGN_EXTEND
and ZERO_EXTEND by DW_OP_shl and DW_OP_shr{a,}. Handle
EQ, NE, GT, GE, LT, LE, GTU, GEU, LTU, LEU, SMIN, SMAX, UMIN,
UMAX, SIGN_EXTRACT, ZERO_EXTRACT.
(loc_descriptor): Compare mode size with DWARF2_ADDR_SIZE
instead of Pmode size.
(loc_descriptor): Add MODE argument. Handle CONST_INT, CONST_DOUBLE,
CONST_VECTOR, CONST, LABEL_REF and SYMBOL_REF if mode != VOIDmode,
attempt to handle other expressions. Don't handle TLS SYMBOL_REFs.
(concat_loc_descriptor, concatn_loc_descriptor,
loc_descriptor_from_tree_1): Adjust loc_descriptor callers.
(add_location_or_const_value_attribute): Likewise. For single
location loc_lists attempt to use add_const_value_attribute
for constant decls. Add DW_AT_const_value even if
NOTE_VAR_LOCATION is VAR_LOCATION with CONSTANT_P or CONST_STRING
in its expression.
* cfgbuild.c (inside_basic_block_p): Handle debug insns.
(control_flow_insn_p): Likewise.
* tree-parloops.c (eliminate_local_variables_stmt): Handle debug
stmt.
(separate_decls_in_region_debug_bind): New.
(separate_decls_in_region): Process debug bind stmts afterwards.
* recog.c (verify_changes): Handle debug insns.
(extract_insn): Likewise.
(peephole2_optimize): Skip debug insns.
* dse.c (scan_insn): Skip debug insns.
* sel-sched-ir.c (return_nop_to_pool): Take full_tidying argument.
Pass it on.
(setup_id_for_insn): Handle debug insns.
(maybe_tidy_empty_bb): Adjust whitespace.
(tidy_control_flow): Skip debug insns.
(sel_remove_insn): Adjust for debug insns.
(sel_estimate_number_of_insns): Skip debug insns.
(create_insn_rtx_from_pattern): Handle debug insns.
(create_copy_of_insn_rtx): Likewise.
* sel-sched-.h (sel_bb_end): Declare.
(sel_bb_empty_or_nop_p): New.
(get_all_loop_exits): Use it.
(_eligible_successor_edge_p): Likewise.
(return_nop_to_pool): Adjust.
* tree-eh.c (tre_empty_eh_handler_p): Skip debug stmts.
* ira-lives.c (process_bb_node_lives): Skip debug insns.
* gimple-pretty-print.c (dump_gimple_debug): New.
(dump_gimple_stmt): Use it.
(dump_bb_header): Skip gimple debug stmts.
* regmove.c (optimize_reg_copy_1): Discount debug insns.
(fixup_match_2): Likewise.
(regmove_backward_pass): Likewise. Simplify combined
replacement. Handle debug insns.
* function.c (instantiate_virtual_regs): Handle debug insns.
* function.h (struct emit_status): Add x_cur_debug_insn_uid.
* print-rtl.h: Include cselib.h.
(print_rtx): Print VALUEs. Split out and recurse for
VAR_LOCATIONs.
* df.h (df_inns_rescan_debug_internal): Declare.
* gcse.c (alloc_hash_table): Estimate n_insns.
(cprop_insn): Don't regard debug insns as changes.
(bypass_conditional_jumps): Skip debug insns.
(one_pre_gcse_pass): Adjust.
(one_code_hoisting_pass): Likewise.
(compute_ld_motion_mems): Skip debug insns.
(one_cprop_pass): Adjust.
* tree-if-conv.c (tree_if_convert_stmt): Reset debug stmts.
(if_convertible_stmt_p): Handle debug stmts.
* init-regs.c (initialize_uninitialized_regs): Skip debug insns.
* tree-vect-loop.c (vect_is_simple_reduction): Skip debug stmts.
* ira-build.c (create_bb_allocnos): Skip debug insns.
* tree-flow-inline.h (has_zero_uses): Discount debug stmts.
(has_single_use): Likewise.
(single_imm_use): Likewise.
(num_imm_uses): Likewise.
* tree-ssa-phiopt.c (empty_block_p): Skip debug stmts.
* tree-ssa-coalesce.c (build_ssa_conflict_graph): Skip debug stmts.
(create_outofssa_var_map): Likewise.
* lower-subreg.c (adjust_decomposed_uses): New.
(resolve_debug): New.
(decompose_multiword_subregs): Use it.
* tree-dfa.c (find_referenced_vars): Skip debug stmts.
* emit-rtl.c: Include params.h.
(cur_debug_insn_uid): Define.
(set_new_first_and_last_insn): Set cur_debug_insn_uid too.
(copy_rtx_if_shared_1): Handle debug insns.
(reset_used_flags): Likewise.
(set_used_flags): LIkewise.
(get_max_insn_count): New.
(next_nondebug_insn): New.
(prev_nondebug_insn): New.
(make_debug_insn_raw): New.
(emit_insn_before_noloc): Handle debug insns.
(emit_jump_insn_before_noloc): Likewise.
(emit_call_insn_before_noloc): Likewise.
(emit_debug_insn_before_noloc): New.
(emit_insn_after_noloc): Handle debug insns.
(emit_jump_insn_after_noloc): Likewise.
(emit_call_insn_after_noloc): Likewise.
(emit_debug_insn_after_noloc): Likewise.
(emit_insn_after): Take loc from earlier non-debug insn.
(emit_jump_insn_after): Likewise.
(emit_call_insn_after): Likewise.
(emit_debug_insn_after_setloc): New.
(emit_debug_insn_after): New.
(emit_insn_before): Take loc from later non-debug insn.
(emit_jump_insn_before): Likewise.
(emit_call_insn_before): Likewise.
(emit_debug_insn_before_setloc): New.
(emit_debug_insn_before): New.
(emit_insn): Handle debug insns.
(emit_debug_insn): New.
(emit_jump_insn): Handle debug insns.
(emit_call_insn): Likewise.
(emit): Likewise.
(init_emit): Take min-nondebug-insn-uid into account.
Initialize cur_debug_insn_uid.
(emit_copy_of_insn_after): Handle debug insns.
* cfgexpand.c (gimple_assign_rhs_to_tree): Do not overwrite
location of single rhs in place.
(maybe_dump_rtl_for_gimple_stmt): Dump lineno.
(floor_sdiv_adjust): New.
(cell_sdiv_adjust): New.
(cell_udiv_adjust): New.
(round_sdiv_adjust): New.
(round_udiv_adjust): New.
(wrap_constant): Moved from cselib.
(unwrap_constant): New.
(expand_debug_expr): New.
(expand_debug_locations): New.
(expand_gimple_basic_block): Drop hiding redeclaration. Expand
debug bind stmts.
(gimple_expand_cfg): Expand debug locations.
* cselib.c: Include tree-pass.h.
(struct expand_value_data): New.
(cselib_record_sets_hook): New.
(PRESERVED_VALUE_P, LONG_TERM_PRESERVED_VALUE_P): New.
(cselib_clear_table): Move, and implemnet in terms of...
(cselib_reset_table_with_next_value): ... this.
(cselib_get_next_unknown_value): New.
(discard_useless_locs): Don't discard preserved values.
(cselib_preserve_value): New.
(cselib_preserved_value_p): New.
(cselib_preserve_definitely): New.
(cselib_clear_preserve): New.
(cselib_preserve_only_values): New.
(new_cselib_val): Take rtx argument. Dump it in details.
(cselib_lookup_mem): Adjust.
(expand_loc): Take regs_active in struct. Adjust. Silence
dumps unless details are requested.
(cselib_expand_value_rtx_cb): New.
(cselib_expand_value_rtx): Rename and reimplment in terms of...
(cselib_expand_value_rtx_1): ... this. Adjust. Silence dumps
without details. Copy more subregs. Try to resolve values
using a callback. Wrap constants.
(cselib_subst_to_values): Adjust.
(cselib_log_lookup): New.
(cselib_lookup): Call it.
(cselib_invalidate_regno): Don't count preserved values as
useless.
(cselib_invalidate_mem): Likewise.
(cselib_record_set): Likewise.
(struct set): Renamed to cselib_set, moved to cselib.h.
(cselib_record_sets): Adjust. Call hook.
(cselib_process_insn): Reset table when it would be cleared.
(dump_cselib_val): New.
(dump_cselib_table): New.
* tree-cfgcleanup.c (tree_forwarded_block_p): Skip debug stmts.
(remove_forwarder_block): Support moving debug stmts.
* cselib.h (cselib_record_sets_hook): Declare.
(cselib_expand_callback): New type.
(cselib_expand_value_rtx_cb): Declare.
(cselib_reset_table_with_next_value): Declare.
(cselib_get_next_unknown_value): Declare.
(cselib_preserve_value): Declare.
(cselib_preserved_value_p): Declare.
(cselib_preserve_only_values): Declare.
(dump_cselib_table): Declare.
* cfgcleanup.c (flow_find_cross_jump): Skip debug insns.
(try_crossjump_to_edge): Likewise.
(delete_unreachable_blocks): Remove dominant GIMPLE blocks after
dominated blocks when debug stmts are present.
* simplify-rtx.c (delegitimize_mem_from_attrs): New.
* tree-ssa-live.c (remove_unused_locals): Skip debug stmts.
(set_var_live_on_entry): Likewise.
* loop-invariant.c (find_invariants_bb): Skip debug insns.
* cfglayout.c (curr_location, last_location): Make static.
(set_curr_insn_source_location): Don't avoid bouncing.
(get_curr_insn_source_location): New.
(get_curr_insn_block): New.
(duplicate_insn_chain): Handle debug insns.
* tree-ssa-forwprop.c (forward_propagate_addr_expr): Propagate
into debug stmts.
* common.opt (fcompare-debug): Move to sort order.
(fdump-unnumbered-links): Likewise.
(fvar-tracking-assignments): New.
(fvar-tracking-assignments-toggle): New.
* tree-ssa-dce.c (mark_stmt_necessary): Don't mark blocks
because of debug stmts.
(mark_stmt_if_obviously_necessary): Mark debug stmts.
(eliminate_unnecessary_stmts): Walk dominated blocks before
dominators.
* tree-ssa-ter.c (find_replaceable_in_bb): Skip debug stmts.
* ira.c (memref_used_between_p): Skip debug insns.
(update_equiv_regs): Likewise.
* sched-deps.c (sd_lists_size): Accept empty list.
(sd_init_insn): Mark debug insns.
(sd_finish_insn): Unmark them.
(sd_add_dep): Reject non-debug deps on debug insns.
(fixup_sched_groups): Give debug insns group treatment.
Skip debug insns.
(sched_analyze_reg): Don't mark debug insns for sched before call.
(sched_analyze_2): Handle debug insns.
(sched_analyze_insn): Compute next non-debug insn. Handle debug
insns.
(deps_analyze_insn): Handle debug insns.
(deps_start_bb): Skip debug insns.
(init_deps): Initialize last_debug_insn.
* tree-ssa.c (target_for_debug_bind): New.
(find_released_ssa_name): New.
(propagate_var_def_into_debug_stmts): New.
(propagate_defs_into_debug_stmts): New.
(verify_ssa): Skip debug bind stmts without values.
(warn_uninialized_vars): Skip debug stmts.
* target-def.h (TARGET_DELEGITIMIZE_ADDRESS): Set default.
* rtl.c (rtx_equal_p_cb): Handle VALUEs.
(rtx_equal_p): Likewise.
* ira-costs.c (scan_one_insn): Skip debug insns.
(process_bb_node_for_hard_reg_moves): Likewise.
* rtl.h (DEBUG_INSN_P): New.
(NONDEBUG_INSN_P): New.
(MAY_HAVE_DEBUG_INSNS): New.
(INSN_P): Accept debug insns.
(RTX_FRAME_RELATED_P): Likewise.
(INSN_DELETED_P): Likewise
(PAT_VAR_LOCATION_DECL): New.
(PAT_VAR_LOCATION_LOC): New.
(PAT_VAR_OCATION_STATUS): New.
(NOTE_VAR_LOCATION_DECL): Reimplement.
(NOTE_VAR_LOCATION_LOC): Likewise.
(NOTE_VAR_LOCATION_STATUS): Likewise.
(INSN_VAR_LOCATION): New.
(INSN_VAR_LOCATION_DECL): New.
(INSN_VAR_LOCATION_LOC): New.
(INSN_VAR_LOCATION_STATUS): New.
(gen_rtx_UNKNOWN_VAR_LOC): New.
(VAR_LOC_UNKNOWN_P): New.
(NOTE_DURING_CALL_P): New.
(SCHED_GROUP_P): Accept debug insns.
(emit_debug_insn_before): Declare.
(emit_debug_insn_before_noloc): Declare.
(emit_debug_insn_beore_setloc): Declare.
(emit_debug_insn_after): Declare.
(emit_debug_insn_after_noloc): Declare.
(emit_debug_insn_after_setloc): Declare.
(emit_debug_insn): Declare.
(make_debug_insn_raw): Declare.
(prev_nondebug_insn): Declare.
(next_nondebug_insn): Declare.
(delegitimize_mem_from_attrs): Declare.
(get_max_insn_count): Declare.
(wrap_constant): Declare.
(unwrap_constant): Declare.
(get_curr_insn_source_location): Declare.
(get_curr_insn_block): Declare.
* tree-inline.c (insert_debug_decl_map): New.
(processing_debug_stmt): New.
(remap_decl): Don't create new mappings in debug stmts.
(remap_gimple_op_r): Don't add references in debug stmts.
(copy_tree_body_r): Likewise.
(remap_gimple_stmt): Handle debug bind stmts.
(copy_bb): Skip debug stmts.
(copy_edges_for_bb): Likewise.
(copy_debug_stmt): New.
(copy_debug_stmts): New.
(copy_body): Copy debug stmts at the end.
(insert_init_debug_bind): New.
(insert_init_stmt): Take id. Skip and emit debug stmts.
(setup_one_parameter): Remap variable earlier, register debug
mapping.
(estimate_num_insns): Skip debug stmts.
(expand_call_inline): Preserve debug_map.
(optimize_inline_calls): Check for no debug_stmts left-overs.
(unsave_expr_now): Preserve debug_map.
(copy_gimple_seq_and_replace_locals): Likewise.
(tree_function_versioning): Check for no debug_stmts left-overs.
Init and destroy debug_map as needed. Split edges unconditionally.
(build_duplicate_type): Init and destroy debug_map as needed.
* tree-inline.h: Include gimple.h instead of pointer-set.h.
(struct copy_body_data): Add debug_stmts and debug_map.
* sched-int.h (struct ready_list): Add n_debug.
(struct deps): Add last_debug_insn.
(DEBUG_INSN_SCHED_P): New.
(BOUNDARY_DEBUG_INSN_P): New.
(SCHEDULE_DEBUG_INSN_P): New.
(sd_iterator_cond): Accept empty list.
* combine.c (create_log_links): Skip debug insns.
(combine_instructions): Likewise.
(cleanup_auto_inc_dec): New. From Jakub Jelinek: Make sure the
return value is always unshared.
(struct rtx_subst_pair): New.
(auto_adjust_pair): New.
(propagate_for_debug_subst): New.
(propagate_for_debug): New.
(try_combine): Skip debug insns. Propagate removed defs into
debug insns.
(next_nonnote_nondebug_insn): New.
(distribute_notes): Use it. Skip debug insns.
(distribute_links): Skip debug insns.
* tree-outof-ssa.c (set_location_for_edge): Likewise.
* resource.c (mark_target_live_regs): Likewise.
* var-tracking.c: Include cselib.h and target.h.
(enum micro_operation_type): Add MO_VAL_USE, MO_VAL_LOC, and
MO_VAL_SET.
(micro_operation_type_name): New.
(enum emit_note_where): Add EMIT_NOTE_AFTER_CALL_INSN.
(struct micro_operation_def): Update comments.
(decl_or_value): New type. Use instead of decls.
(struct emit_note_data_def): Add vars.
(struct attrs_def): Use decl_or_value.
(struct variable_tracking_info_def): Add permp, flooded.
(struct location_chain_def): Update comment.
(struct variable_part_def): Use decl_or_value.
(struct variable_def): Make var_part a variable length array.
(valvar_pool): New.
(scratch_regs): New.
(cselib_hook_called): New.
(dv_is_decl_p): New.
(dv_is_value_p): New.
(dv_as_decl): New.
(dv_as_value): New.
(dv_as_opaque): New.
(dv_onepart_p): New.
(dv_pool): New.
(IS_DECL_CODE): New.
(check_value_is_not_decl): New.
(dv_from_decl): New.
(dv_from_value): New.
(dv_htab_hash): New.
(variable_htab_hash): Use it.
(variable_htab_eq): Support values.
(variable_htab_free): Free from the right pool.
(attrs_list_member, attrs_list_insert): Use decl_or_value.
(attrs_list_union): Adjust.
(attrs_list_mpdv_union): New.
(tie_break_pointers): New.
(canon_value_cmp): New.
(unshare_variable): Return possibly-modified slot.
(vars_copy_1): Adjust.
(var_reg_decl_set): Adjust. Split out of...
(var_reg_set): ... this.
(get_init_value): Adjust.
(var_reg_delete_and_set): Adjust.
(var_reg_delete): Adjust.
(var_regno_delete): Adjust.
(var_mem_decl_set): Split out of...
(var_mem_set): ... this.
(var_mem_delete_and_set): Adjust.
(var_mem_delete): Adjust.
(val_store): New.
(val_reset): New.
(val_resolve): New.
(variable_union): Adjust. Speed up merge of 1-part vars.
(variable_canonicalize): Use unshared slot.
(VALUED_RECURSED_INTO): New.
(find_loc_in_1pdv): New.
(struct dfset_merge): New.
(insert_into_intersection): New.
(intersect_loc_chains): New.
(loc_cmp): New.
(canonicalize_loc_order_check): New.
(canonicalize_values_mark): New.
(canonicalize_values_star): New.
(variable_merge_over_cur): New.
(variable_merge_over_src): New.
(dataflow_set_merge): New.
(dataflow_set_equiv_regs): New.
(remove_duplicate_values): New.
(struct dfset_post_merge): New.
(variable_post_merge_new_vals): New.
(variable_post_merge_perm_vals): New.
(dataflow_post_merge_adjust): New.
(find_mem_expr_in_1pdv): New.
(dataflow_set_preserve_mem_locs): New.
(dataflow_set_remove_mem_locs): New.
(dataflow_set_clear_at_call): New.
(onepart_variable_different_p): New.
(variable_different_p): Use it.
(dataflow_set_different_1): Adjust. Make detailed dump
more verbose.
(track_expr_p): Add need_rtl parameter. Don't generate rtl
if not needed.
(track_loc_p): Pass it true.
(struct count_use_info): New.
(find_use_val): New.
(replace_expr_with_values): New.
(log_op_type): New.
(use_type): New, partially split out of...
(count_uses): ... this. Count new micro-ops.
(count_uses_1): Adjust.
(count_stores): Adjust.
(count_with_sets): New.
(VAL_NEEDS_RESOLUTION): New.
(VAL_HOLDS_TRACK_EXPR): New.
(VAL_EXPR_IS_COPIED): New.
(VAL_EXPR_IS_CLOBBERED): New.
(add_uses): Adjust. Generate new micro-ops.
(add_uses_1): Adjust.
(add_stores): Generate new micro-ops.
(add_with_sets): New.
(find_src_status): Adjust.
(find_src_set_src): Adjust.
(compute_bb_dataflow): Use dataflow_set_clear_at_call.
Handle new micro-ops. Canonicalize value equivalances.
(vt_find_locations): Compute total size of hash tables for
dumping. Perform merge for var-tracking-assignments. Don't
disregard single-block loops.
(dump_attrs_list): Handle decl_or_value.
(dump_variable): Take variable. Deal with decl_or_value.
(dump_variable_slot): New.
(dump_vars): Use it.
(dump_dataflow_sets): Adjust.
(set_slot_part): New, extended to support one-part variables
after splitting out of...
(set_variable_part): ... this.
(clobber_slot_part): New, split out of...
(clobber_variable_part): ... this.
(delete_slot_part): New, split out of...
(delete_variable_part): .... this.
(check_wrap_constant): New.
(vt_expand_loc_callback): New.
(vt_expand_loc): New.
(emit_note_insn_var_location): Adjust. Handle values. Handle
EMIT_NOTE_AFTER_CALL_INSN.
(emit_notes_for_differences_1): Adjust. Handle values.
(emit_notes_for_differences_2): Likewise.
(emit_notes_for_differences): Adjust.
(emit_notes_in_bb): Take pointer to set. Emit AFTER_CALL_INSN
notes. Adjust. Handle new micro-ops.
(vt_add_function_parameters): Adjust. Create and bind values.
(vt_initialize): Adjust. Initialize scratch_regs and
valvar_pool, flooded and perm.. Initialize and use cselib. Log
operations. Move some code to count_with_sets and add_with_sets.
(delete_debug_insns): New.
(vt_debug_insns_local): New.
(vt_finalize): Release permp, valvar_pool, scratch_regs. Finish
cselib.
(var_tracking_main): If var-tracking-assignments is enabled
but var-tracking isn't, delete debug insns and leave. Likewise
if we exceed limits or fail the stack adjustments tests, and
after all var-tracking processing.
More in var-tracking, from Jakub Jelinek <jakub@redhat.com>:
(dataflow_set): Add traversed_vars.
(value_chain, const_value_chain): New typedefs.
(value_chain_pool, value_chains): New variables.
(value_chain_htab_hash, value_chain_htab_eq, add_value_chain,
add_value_chains, add_cselib_value_chains, remove_value_chain,
remove_value_chains, remove_cselib_value_chains): New functions.
(shared_hash_find_slot_unshare_1, shared_hash_find_slot_1,
shared_hash_find_slot_noinsert_1, shared_hash_find_1): New
static inlines.
(shared_hash_find_slot_unshare, shared_hash_find_slot,
shared_hash_find_slot_noinsert, shared_hash_find): Update.
(dst_can_be_shared): New variable.
(unshare_variable): Unshare set->vars if shared, use shared_hash_*.
Clear dst_can_be_shared. If set->traversed_vars is non-NULL and
different from set->vars, look up slot again instead of using the
passed in slot.
(dataflow_set_init): Initialize traversed_vars.
(variable_union): Use shared_hash_*. Use initially NO_INSERT
lookup if set->vars is shared. Don't keep slot cleared before
calling unshare_variable. Unshare set->vars if needed. Adjust
unshare_variable callers. Clear dst_can_be_shared if needed.
Even ->refcount == 1 vars must be unshared if set->vars is shared
and var needs to be modified.
(dataflow_set_union): Set traversed_vars during canonicalization.
(VALUE_CHANGED, DECL_CHANGED): Define.
(set_dv_changed, dv_changed_p): New static inlines.
(track_expr_p): Clear DECL_CHANGED.
(dump_dataflow_sets): Set it.
(variable_was_changed): Call set_dv_changed.
(emit_note_insn_var_location): Likewise.
(changed_variables_stack): New variable.
(check_changed_vars_1, check_changed_vars_2): New functions.
(emit_notes_for_changes): Do nothing if changed_variables is
empty. Traverse changed_variables with check_changed_vars_1,
call check_changed_vars_2 on each changed_variables_stack entry.
(emit_notes_in_bb): Add SET argument. Just clear it at the
beginning, use it instead of local &set, don't destroy it at the
end.
(vt_emit_notes): Call dataflow_set_clear early on all
VTI(bb)->out sets, never use them, instead use emit_notes_in_bb
computed set, dataflow_set_clear also VTI(bb)->in when we are
done with the basic block. Initialize changed_variables_stack,
free it afterwards. If ENABLE_CHECKING verify that after noting
differences to an empty set value_chains hash table is empty.
(vt_initialize): Initialize value_chains and value_chain_pool.
(vt_finalize): Delete value_chains htab, free value_chain_pool.
(variable_tracking_main): Call dump_dataflow_sets before calling
vt_emit_notes, not after it.
* tree-flow.h (propagate_defs_into_debug_stmts): Declare.
(propagate_var_def_into_debug_stmts): Declare.
* df-problems.c (df_lr_bb_local_compute): Skip debug insns.
(df_set_note): Reject debug insns.
(df_whole_mw_reg_dead_p): Take added_notes_p argument. Don't
add notes to debug insns.
(df_note_bb_compute): Adjust. Likewise.
(df_simulate_uses): Skip debug insns.
(df_simulate_initialize_backwards): Likewise.
* reg-stack.c (subst_stack_regs_in_debug_insn): New.
(subst_stack_regs_pat): Reject debug insns.
(convert_regs_1): Handle debug insns.
* Makefile.in (TREE_INLINE_H): Take pointer-set.h from GIMPLE_H.
(print-rtl.o): Depend on cselib.h.
(cselib.o): Depend on TREE_PASS_H.
(var-tracking.o): Depend on cselib.h and TARGET_H.
* sched-rgn.c (rgn_estimate_number_of_insns): Discount
debug insns.
(init_ready_list): Skip boundary debug insns.
(add_branch_dependences): Skip debug insns.
(free_block_dependencies): Check for blocks with only debug
insns.
(compute_priorities): Likewise.
* gimple.c (gss_for_code): Handle GIMPLE_DEBUG.
(gimple_build_with_ops_stat): Take subcode as unsigned. Adjust
all callers.
(gimple_build_debug_bind_stat): New.
(empty_body_p): Skip debug stmts.
(gimple_has_side_effects): Likewise.
(gimple_rhs_has_side_effects): Likewise.
* gimple.h (enum gimple_debug_subcode, GIMPLE_DEBUG_BIND): New.
(gimple_build_debug_bind_stat): Declare.
(gimple_build_debug_bind): Define.
(is_gimple_debug): New.
(gimple_debug_bind_p): New.
(gimple_debug_bind_get_var): New.
(gimple_debug_bind_get_value): New.
(gimple_debug_bind_get_value_ptr): New.
(gimple_debug_bind_set_var): New.
(gimple_debug_bind_set_value): New.
(GIMPLE_DEBUG_BIND_NOVALUE): New internal temporary macro.
(gimple_debug_bind_reset_value): New.
(gimple_debug_bind_has_value_p): New.
(gsi_next_nondebug): New.
(gsi_prev_nondebug): New.
(gsi_start_nondebug_bb): New.
(gsi_last_nondebug_bb): New.
* sched-vis.c (print_pattern): Handle VAR_LOCATION.
(print_insn): Handle DEBUG_INSN.
* tree-cfg.c (remove_bb): Walk stmts backwards. Let loc
of first insn prevail.
(first_stmt): Skip debug stmts.
(first_non_label_stmt): Likewise.
(last_stmt): Likewise.
(has_zero_uses_1): New.
(single_imm_use_1): New.
(verify_gimple_debug): New.
(verify_types_in_gimple_stmt): Handle debug stmts.
(verify_stmt): Likewise.
(debug_loop_num): Skip debug stmts.
(remove_edge_and_dominated_blocks): Remove dominators last.
* tree-ssa-reasssoc.c (rewrite_expr_tree): Propagate into
debug stmts.
(linearize_expr): Likewise.
* config/i386/i386.c (ix86_delegitimize_address): Call
default implementation.
* config/ia64/ia64.c (ia64_safe_itanium_class): Handle debug
insns.
(group_barrier_needed): Skip debug insns.
(emit_insn_group_barriers): Likewise.
(emit_all_insn_group_barriers): Likewise.
(ia64_variable_issue): Handle debug insns.
(ia64_dfa_new_cycle): Likewise.
(final_emit_insn_group_barriers): Skip debug insns.
(ia64_dwarf2out_def_steady_cfa): Take frame argument. Don't
def cfa without frame.
(process_set): Likewise.
(process_for_unwind_directive): Pass frame on.
* config/rs6000/rs6000.c (TARGET_DELEGITIMIZE_ADDRESS): Define.
(rs6000_delegitimize_address): New.
(rs6000_debug_adjust_cost): Handle debug insns.
(is_microcoded_insn): Likewise.
(is_cracked_insn): Likewise.
(is_nonpipeline_insn): Likewise.
(insn_must_be_first_in_group): Likewise.
(insn_must_be_last_in_group): Likewise.
(force_new_group): Likewise.
* cfgrtl.c (rtl_split_block): Emit INSN_DELETED note if block
contains only debug insns.
(rtl_merge_blocks): Skip debug insns.
(purge_dead_edges): Likewise.
(rtl_block_ends_with_call_p): Skip debug insns.
* dce.c (deletable_insn_p): Handle VAR_LOCATION.
(mark_reg_dependencies): Skip debug insns.
* params.def (PARAM_MIN_NONDEBUG_INSN_UID): New.
* tree-ssanames.c (release_ssa_name): Propagate def into
debug stmts.
* tree-ssa-threadedge.c
(record_temporary_equivalences_from_stmts): Skip debug stmts.
* regcprop.c (replace_oldest_value_addr): Skip debug insns.
(replace_oldest_value_mem): Use ALL_REGS for debug insns.
(copyprop_hardreg_forward_1): Handle debug insns.
* reload1.c (reload): Skip debug insns. Replace unassigned
pseudos in debug insns with their equivalences.
(eliminate_regs_in_insn): Skip debug insns.
(emit_input_reload_insns): Skip debug insns at first, adjust
them later.
* tree-ssa-operands.c (add_virtual_operand): Reject debug stmts.
(get_indirect_ref_operands): Pass opf_no_vops on.
(get_expr_operands): Likewise. Skip debug stmts.
(parse_ssa_operands): Scan debug insns with opf_no_vops.
gcc/testsuite/ChangeLog:
* gcc.dg/guality/guality.c: New.
* gcc.dg/guality/guality.h: New.
* gcc.dg/guality/guality.exp: New.
* gcc.dg/guality/example.c: New.
* lib/gcc-dg.exp (cleanup-dump): Remove .gk files.
(cleanup-saved-temps): Likewise, .gkd files too.
gcc/cp/ChangeLog:
* cp-tree.h (TFF_NO_OMIT_DEFAULT_TEMPLATE_ARGUMENTS): New.
* cp-lang.c (cxx_dwarf_name): Pass it.
* error.c (count_non_default_template_args): Take flags as
argument. Adjust all callers. Skip counting of default
arguments if the new flag is given.
ChangeLog:
* Makefile.tpl (BUILD_CONFIG): Default to bootstrap-debug.
* Makefile.in: Rebuilt.
contrib/ChangeLog:
* compare-debug: Look for .gkd files and compare them.
config/ChangeLog:
* bootstrap-debug.mk: Add comments.
* bootstrap-debug-big.mk: New.
* bootstrap-debug-lean.mk: New.
* bootstrap-debug-ckovw.mk: Add comments.
* bootstrap-debug-lib.mk: Drop CFLAGS for stages. Use -g0
for TFLAGS in stage1. Drop -fvar-tracking-assignments-toggle.
From-SVN: r151312
2009-09-02 04:42:21 +02:00
/* Set minimum insn uid for non-debug insns. */
DEFPARAM ( PARAM_MIN_NONDEBUG_INSN_UID ,
"min-nondebug-insn-uid" ,
2015-10-22 00:24:41 +02:00
"The minimum UID to be used for a nondebug insn." ,
2017-02-23 09:38:35 +01:00
0 , 0 , 0 )
invoke.texi (-fvar-tracking-assignments): New.
gcc/ChangeLog:
* doc/invoke.texi (-fvar-tracking-assignments): New.
(-fvar-tracking-assignments-toggle): New.
(-fdump-final-insns=file): Mark filename as optional.
(--param min-nondebug-insn-uid): New.
(-gdwarf-@{version}): Mention version 4.
* opts.c (common_handle_option): Accept it.
* tree-vrp.c (find_assert_locations_1): Skip debug stmts.
* regrename.c (regrename_optimize): Drop last. Don't count debug
insns as uses. Don't reject change because of debug insn.
(do_replace): Reject DEBUG_INSN as chain starter. Take base_regno
from the chain starter, and check for inexact matches in
DEBUG_INSNS.
(scan_rtx_reg): Accept inexact matches in DEBUG_INSNs.
(build_def_use): Simplify and fix the marking of DEBUG_INSNs.
* sched-ebb.c (schedule_ebbs): Skip boundary debug insns.
* fwprop.c (forward_propagate_and_simplify): ...into debug insns.
* doc/gimple.texi (is_gimple_debug): New.
(gimple_debug_bind_p): New.
(is_gimple_call, gimple_assign_cast_p): End sentence with period.
* doc/install.texi (bootstrap-debug): More details.
(bootstrap-debug-big, bootstrap-debug-lean): Document.
(bootstrap-debug-lib): More details.
(bootstrap-debug-ckovw): Update.
(bootstrap-time): New.
* tree-into-ssa.c (mark_def_sites): Skip debug stmts.
(insert_phi_nodes_for): Insert debug stmts.
(rewrite_stmt): Take iterator. Insert debug stmts.
(rewrite_enter_block): Adjust.
(maybe_replace_use_in_debug_stmt): New.
(rewrite_update_stmt): Use it.
(mark_use_interesting): Return early for debug stmts.
* tree-ssa-loop-im.c (rewrite_bittest): Propagate DEFs into debug
stmts before replacing stmt.
(move_computations_stmt): Likewise.
* ira-conflicts.c (add_copies): Skip debug insns.
* regstat.c (regstat_init_n_sets_and_refs): Discount debug insns.
(regstat_bb_compute_ri): Skip debug insns.
* tree-ssa-threadupdate.c (redirection_block_p): Skip debug stmts.
* tree-ssa-loop-manip.c (find_uses_to_rename_stmt,
check_loop_closed_ssa_stmt): Skip debug stmts.
* tree-tailcall.c (find_tail_calls): Likewise.
* tree-ssa-loop-ch.c (should_duplicate_loop_header_p): Likewise.
* tree.h (MAY_HAVE_DEBUG_STMTS): New.
(build_var_debug_value_stat): Declare.
(build_var_debug_value): Define.
(target_for_debug_bind): Declare.
* reload.c (find_equiv_reg): Skip debug insns.
* rtlanal.c (reg_used_between_p): Skip debug insns.
(side_effects_p): Likewise.
(canonicalize_condition): Likewise.
* ddg.c (create_ddg_dep_from_intra_loop_link): Check that non-debug
insns never depend on debug insns.
(create_ddg_dep_no_link): Likewise.
(add_cross_iteration_register_deps): Use ANTI_DEP for debug insns.
Don't add inter-loop dependencies for debug insns.
(build_intra_loop_deps): Likewise.
(create_ddg): Count debug insns.
* ddg.h (struct ddg::num_debug): New.
(num_backargs): Pair up with previous int field.
* diagnostic.c (diagnostic_report_diagnostic): Skip notes on
-fcompare-debug-second.
* final.c (get_attr_length_1): Skip debug insns.
(rest_of_clean-state): Don't dump CFA_RESTORE_STATE.
* gcc.c (invoke_as): Call compare-debug-dump-opt.
(driver_self_specs): Map -fdump-final-insns to
-fdump-final-insns=..
(get_local_tick): New.
(compare_debug_dump_opt_spec_function): Test for . argument and
compute output name. Compute temp output spec without flag name.
Compute -frandom-seed.
(OPT): Undef after use.
* cfgloopanal.c (num_loop_insns): Skip debug insns.
(average_num_loop_insns): Likewise.
* params.h (MIN_NONDEBUG_INSN_UID): New.
* gimple.def (GIMPLE_DEBUG): New.
* ipa-reference.c (scan_stmt_for_static_refs): Skip debug stmts.
* auto-inc-dec.c (merge_in_block): Skip debug insns.
(merge_in_block): Fix whitespace.
* toplev.c (flag_var_tracking): Update comment.
(flag_var_tracking_assignments): New.
(flag_var_tracking_assignments_toggle): New.
(process_options): Don't open final insns dump file if we're not
going to write to it. Compute defaults for var_tracking.
* df-scan.c (df_insn_rescan_debug_internal): New.
(df_uses_record): Handle debug insns.
* haifa-sched.c (ready): Initialize n_debug.
(contributes_to_priority): Skip debug insns.
(dep_list_size): New.
(priority): Use it.
(rank_for_schedule): Likewise. Schedule debug insns as soon as
they're ready. Disregard previous debug insns to make decisions.
(queue_insn): Never queue debug insns.
(ready_add, ready_remove_first, ready_remove): Count debug insns.
(schedule_insn): Don't reject debug insns because of issue rate.
(get_ebb_head_tail, no_real_insns_p): Skip boundary debug insns.
(queue_to_ready): Skip and discount debug insns.
(choose_ready): Let debug insns through.
(schedule_block): Check boundary debug insns. Discount debug
insns, schedule them early. Adjust whitespace.
(set_priorities): Check for boundary debug insns.
(add_jump_dependencies): Use dep_list_size.
(prev_non_location_insn): New.
(check_cfg): Use it.
* tree-ssa-loop-ivopts.c (find-interesting_users): Skip debug
stmts.
(remove_unused_ivs): Reset debug stmts.
* modulo-sched.c (const_iteration_count): Skip debug insns.
(res_MII): Discount debug insns.
(loop_single_full_bb_p): Skip debug insns.
(sms_schedule): Likewise.
(sms_schedule_by_order): Likewise.
(ps_has_conflicts): Likewise.
* caller-save.c (refmarker_fn): New.
(save_call_clobbered_regs): Replace regs with saved mem in
debug insns.
(mark_referenced_regs): Take pointer, mark and arg. Adjust.
Call refmarker_fn mark for hardregnos.
(mark_reg_as_referenced): New.
(replace_reg_with_saved_mem): New.
* ipa-pure-const.c (check_stmt): Skip debug stmts.
* cse.c (cse_insn): Canonicalize debug insns. Skip them when
searching back.
(cse_extended_basic_block): Skip debug insns.
(count_reg_usage): Likewise.
(is_dead_reg): New, split out of...
(set_live_p): ... here.
(insn_live_p): Use it for debug insns.
* tree-stdarg.c (check_all_va_list_escapes): Skip debug stmts.
(execute_optimize_stdarg): Likewise.
* tree-ssa-dom.c (propagate_rhs_into_lhs): Likewise.
* tree-ssa-propagate.c (substitute_and_fold): Don't regard
changes in debug stmts as changes.
* sel-sched.c (moving_insn_creates_bookkeeping_block_p): New.
(moveup_expr): Don't move across debug insns. Don't move
debug insn if it would create a bookkeeping block.
(moveup_expr_cached): Don't use cache for debug insns that
are heads of blocks.
(compute_av_set_inside_bb): Skip debug insns.
(sel_rank_for_schedule): Schedule debug insns first. Remove
dead code.
(block_valid_for_bookkeeping_p); Support lax searches.
(create_block_for_bookkeeping): Adjust block numbers when
encountering debug-only blocks.
(find_place_for_bookkeeping): Deal with debug-only blocks.
(generate_bookkeeping_insn): Accept no place to insert.
(remove_temp_moveop_nops): New argument full_tidying.
(prepare_place_to_insert): Deal with debug insns.
(advance_state_on_fence): Debug insns don't start cycles.
(update_boundaries): Take fence as argument. Deal with
debug insns.
(schedule_expr_on_boundary): No full_tidying on debug insns.
(fill_insns): Deal with debug insns.
(track_scheduled_insns_and_blocks): Don't count debug insns.
(need_nop_to_preserve_insn_bb): New, split out of...
(remove_insn_from_stream): ... this.
(fur_orig_expr_not_found): Skip debug insns.
* rtl.def (VALUE): Move up.
(DEBUG_INSN): New.
* tree-ssa-sink.c (all_immediate_uses_same_place): Skip debug
stmts.
(nearest_common_dominator_of_uses): Take debug_stmts argument.
Set it if debug stmts are found.
(statement_sink_location): Skip debug stmts. Propagate
moving defs into debug stmts.
* ifcvt.c (first_active_insn): Skip debug insns.
(last_active_insns): Likewise.
(cond_exec_process_insns): Likewise.
(noce_process_if_block): Likewise.
(check_cond_move_block): Likewise.
(cond_move_convert_if_block): Likewise.
(block_jumps_and_fallthru_p): Likewise.
(dead_or_predicable): Likewise.
* dwarf2out.c (debug_str_hash_forced): New.
(find_AT_string): Add comment.
(gen_label_for_indirect_string): New.
(get_debug_string_label): New.
(AT_string_form): Use it.
(mem_loc_descriptor): Handle non-TLS symbols. Handle MINUS , DIV,
MOD, AND, IOR, XOR, NOT, ABS, NEG, and CONST_STRING. Accept but
discard COMPARE, IF_THEN_ELSE, ROTATE, ROTATERT, TRUNCATE and
several operations that cannot be represented with DWARF opcodes.
(loc_descriptor): Ignore SIGN_EXTEND and ZERO_EXTEND. Require
dwarf_version 4 for DW_OP_implicit_value and DW_OP_stack_value.
(dwarf2out_var_location): Take during-call mark into account.
(output_indirect_string): Update comment. Output if there are
label and references.
(prune_indirect_string): New.
(prune_unused_types): Call it if debug_str_hash_forced.
More in dwarf2out.c, from Jakub Jelinek <jakub@redhat.com>:
(dw_long_long_const): Remove.
(struct dw_val_struct): Change val_long_long type to rtx.
(print_die, attr_checksum, same_dw_val_p, loc_descriptor): Adjust for
val_long_long change to CONST_DOUBLE rtx from a long hi/lo pair.
(output_die): Likewise. Use HOST_BITS_PER_WIDE_INT size of each
component instead of HOST_BITS_PER_LONG.
(output_loc_operands): Likewise. For const8* assert
HOST_BITS_PER_WIDE_INT rather than HOST_BITS_PER_LONG is >= 64.
(output_loc_operands_raw): For const8* assert HOST_BITS_PER_WIDE_INT
rather than HOST_BITS_PER_LONG is >= 64.
(add_AT_long_long): Remove val_hi and val_lo arguments, add
val_const_double.
(size_of_die): Use HOST_BITS_PER_WIDE_INT size multiplier instead of
HOST_BITS_PER_LONG for dw_val_class_long_long.
(add_const_value_attribute): Adjust add_AT_long_long caller. Don't
handle TLS SYMBOL_REFs. If CONST wraps a constant, tail recurse.
(dwarf_stack_op_name): Handle DW_OP_implicit_value and
DW_OP_stack_value.
(size_of_loc_descr, output_loc_operands, output_loc_operands_raw):
Handle DW_OP_implicit_value.
(extract_int): Move prototype earlier.
(mem_loc_descriptor): For SUBREG punt if inner
mode size is wider than DWARF2_ADDR_SIZE. Handle SIGN_EXTEND
and ZERO_EXTEND by DW_OP_shl and DW_OP_shr{a,}. Handle
EQ, NE, GT, GE, LT, LE, GTU, GEU, LTU, LEU, SMIN, SMAX, UMIN,
UMAX, SIGN_EXTRACT, ZERO_EXTRACT.
(loc_descriptor): Compare mode size with DWARF2_ADDR_SIZE
instead of Pmode size.
(loc_descriptor): Add MODE argument. Handle CONST_INT, CONST_DOUBLE,
CONST_VECTOR, CONST, LABEL_REF and SYMBOL_REF if mode != VOIDmode,
attempt to handle other expressions. Don't handle TLS SYMBOL_REFs.
(concat_loc_descriptor, concatn_loc_descriptor,
loc_descriptor_from_tree_1): Adjust loc_descriptor callers.
(add_location_or_const_value_attribute): Likewise. For single
location loc_lists attempt to use add_const_value_attribute
for constant decls. Add DW_AT_const_value even if
NOTE_VAR_LOCATION is VAR_LOCATION with CONSTANT_P or CONST_STRING
in its expression.
* cfgbuild.c (inside_basic_block_p): Handle debug insns.
(control_flow_insn_p): Likewise.
* tree-parloops.c (eliminate_local_variables_stmt): Handle debug
stmt.
(separate_decls_in_region_debug_bind): New.
(separate_decls_in_region): Process debug bind stmts afterwards.
* recog.c (verify_changes): Handle debug insns.
(extract_insn): Likewise.
(peephole2_optimize): Skip debug insns.
* dse.c (scan_insn): Skip debug insns.
* sel-sched-ir.c (return_nop_to_pool): Take full_tidying argument.
Pass it on.
(setup_id_for_insn): Handle debug insns.
(maybe_tidy_empty_bb): Adjust whitespace.
(tidy_control_flow): Skip debug insns.
(sel_remove_insn): Adjust for debug insns.
(sel_estimate_number_of_insns): Skip debug insns.
(create_insn_rtx_from_pattern): Handle debug insns.
(create_copy_of_insn_rtx): Likewise.
* sel-sched-.h (sel_bb_end): Declare.
(sel_bb_empty_or_nop_p): New.
(get_all_loop_exits): Use it.
(_eligible_successor_edge_p): Likewise.
(return_nop_to_pool): Adjust.
* tree-eh.c (tre_empty_eh_handler_p): Skip debug stmts.
* ira-lives.c (process_bb_node_lives): Skip debug insns.
* gimple-pretty-print.c (dump_gimple_debug): New.
(dump_gimple_stmt): Use it.
(dump_bb_header): Skip gimple debug stmts.
* regmove.c (optimize_reg_copy_1): Discount debug insns.
(fixup_match_2): Likewise.
(regmove_backward_pass): Likewise. Simplify combined
replacement. Handle debug insns.
* function.c (instantiate_virtual_regs): Handle debug insns.
* function.h (struct emit_status): Add x_cur_debug_insn_uid.
* print-rtl.h: Include cselib.h.
(print_rtx): Print VALUEs. Split out and recurse for
VAR_LOCATIONs.
* df.h (df_inns_rescan_debug_internal): Declare.
* gcse.c (alloc_hash_table): Estimate n_insns.
(cprop_insn): Don't regard debug insns as changes.
(bypass_conditional_jumps): Skip debug insns.
(one_pre_gcse_pass): Adjust.
(one_code_hoisting_pass): Likewise.
(compute_ld_motion_mems): Skip debug insns.
(one_cprop_pass): Adjust.
* tree-if-conv.c (tree_if_convert_stmt): Reset debug stmts.
(if_convertible_stmt_p): Handle debug stmts.
* init-regs.c (initialize_uninitialized_regs): Skip debug insns.
* tree-vect-loop.c (vect_is_simple_reduction): Skip debug stmts.
* ira-build.c (create_bb_allocnos): Skip debug insns.
* tree-flow-inline.h (has_zero_uses): Discount debug stmts.
(has_single_use): Likewise.
(single_imm_use): Likewise.
(num_imm_uses): Likewise.
* tree-ssa-phiopt.c (empty_block_p): Skip debug stmts.
* tree-ssa-coalesce.c (build_ssa_conflict_graph): Skip debug stmts.
(create_outofssa_var_map): Likewise.
* lower-subreg.c (adjust_decomposed_uses): New.
(resolve_debug): New.
(decompose_multiword_subregs): Use it.
* tree-dfa.c (find_referenced_vars): Skip debug stmts.
* emit-rtl.c: Include params.h.
(cur_debug_insn_uid): Define.
(set_new_first_and_last_insn): Set cur_debug_insn_uid too.
(copy_rtx_if_shared_1): Handle debug insns.
(reset_used_flags): Likewise.
(set_used_flags): LIkewise.
(get_max_insn_count): New.
(next_nondebug_insn): New.
(prev_nondebug_insn): New.
(make_debug_insn_raw): New.
(emit_insn_before_noloc): Handle debug insns.
(emit_jump_insn_before_noloc): Likewise.
(emit_call_insn_before_noloc): Likewise.
(emit_debug_insn_before_noloc): New.
(emit_insn_after_noloc): Handle debug insns.
(emit_jump_insn_after_noloc): Likewise.
(emit_call_insn_after_noloc): Likewise.
(emit_debug_insn_after_noloc): Likewise.
(emit_insn_after): Take loc from earlier non-debug insn.
(emit_jump_insn_after): Likewise.
(emit_call_insn_after): Likewise.
(emit_debug_insn_after_setloc): New.
(emit_debug_insn_after): New.
(emit_insn_before): Take loc from later non-debug insn.
(emit_jump_insn_before): Likewise.
(emit_call_insn_before): Likewise.
(emit_debug_insn_before_setloc): New.
(emit_debug_insn_before): New.
(emit_insn): Handle debug insns.
(emit_debug_insn): New.
(emit_jump_insn): Handle debug insns.
(emit_call_insn): Likewise.
(emit): Likewise.
(init_emit): Take min-nondebug-insn-uid into account.
Initialize cur_debug_insn_uid.
(emit_copy_of_insn_after): Handle debug insns.
* cfgexpand.c (gimple_assign_rhs_to_tree): Do not overwrite
location of single rhs in place.
(maybe_dump_rtl_for_gimple_stmt): Dump lineno.
(floor_sdiv_adjust): New.
(cell_sdiv_adjust): New.
(cell_udiv_adjust): New.
(round_sdiv_adjust): New.
(round_udiv_adjust): New.
(wrap_constant): Moved from cselib.
(unwrap_constant): New.
(expand_debug_expr): New.
(expand_debug_locations): New.
(expand_gimple_basic_block): Drop hiding redeclaration. Expand
debug bind stmts.
(gimple_expand_cfg): Expand debug locations.
* cselib.c: Include tree-pass.h.
(struct expand_value_data): New.
(cselib_record_sets_hook): New.
(PRESERVED_VALUE_P, LONG_TERM_PRESERVED_VALUE_P): New.
(cselib_clear_table): Move, and implemnet in terms of...
(cselib_reset_table_with_next_value): ... this.
(cselib_get_next_unknown_value): New.
(discard_useless_locs): Don't discard preserved values.
(cselib_preserve_value): New.
(cselib_preserved_value_p): New.
(cselib_preserve_definitely): New.
(cselib_clear_preserve): New.
(cselib_preserve_only_values): New.
(new_cselib_val): Take rtx argument. Dump it in details.
(cselib_lookup_mem): Adjust.
(expand_loc): Take regs_active in struct. Adjust. Silence
dumps unless details are requested.
(cselib_expand_value_rtx_cb): New.
(cselib_expand_value_rtx): Rename and reimplment in terms of...
(cselib_expand_value_rtx_1): ... this. Adjust. Silence dumps
without details. Copy more subregs. Try to resolve values
using a callback. Wrap constants.
(cselib_subst_to_values): Adjust.
(cselib_log_lookup): New.
(cselib_lookup): Call it.
(cselib_invalidate_regno): Don't count preserved values as
useless.
(cselib_invalidate_mem): Likewise.
(cselib_record_set): Likewise.
(struct set): Renamed to cselib_set, moved to cselib.h.
(cselib_record_sets): Adjust. Call hook.
(cselib_process_insn): Reset table when it would be cleared.
(dump_cselib_val): New.
(dump_cselib_table): New.
* tree-cfgcleanup.c (tree_forwarded_block_p): Skip debug stmts.
(remove_forwarder_block): Support moving debug stmts.
* cselib.h (cselib_record_sets_hook): Declare.
(cselib_expand_callback): New type.
(cselib_expand_value_rtx_cb): Declare.
(cselib_reset_table_with_next_value): Declare.
(cselib_get_next_unknown_value): Declare.
(cselib_preserve_value): Declare.
(cselib_preserved_value_p): Declare.
(cselib_preserve_only_values): Declare.
(dump_cselib_table): Declare.
* cfgcleanup.c (flow_find_cross_jump): Skip debug insns.
(try_crossjump_to_edge): Likewise.
(delete_unreachable_blocks): Remove dominant GIMPLE blocks after
dominated blocks when debug stmts are present.
* simplify-rtx.c (delegitimize_mem_from_attrs): New.
* tree-ssa-live.c (remove_unused_locals): Skip debug stmts.
(set_var_live_on_entry): Likewise.
* loop-invariant.c (find_invariants_bb): Skip debug insns.
* cfglayout.c (curr_location, last_location): Make static.
(set_curr_insn_source_location): Don't avoid bouncing.
(get_curr_insn_source_location): New.
(get_curr_insn_block): New.
(duplicate_insn_chain): Handle debug insns.
* tree-ssa-forwprop.c (forward_propagate_addr_expr): Propagate
into debug stmts.
* common.opt (fcompare-debug): Move to sort order.
(fdump-unnumbered-links): Likewise.
(fvar-tracking-assignments): New.
(fvar-tracking-assignments-toggle): New.
* tree-ssa-dce.c (mark_stmt_necessary): Don't mark blocks
because of debug stmts.
(mark_stmt_if_obviously_necessary): Mark debug stmts.
(eliminate_unnecessary_stmts): Walk dominated blocks before
dominators.
* tree-ssa-ter.c (find_replaceable_in_bb): Skip debug stmts.
* ira.c (memref_used_between_p): Skip debug insns.
(update_equiv_regs): Likewise.
* sched-deps.c (sd_lists_size): Accept empty list.
(sd_init_insn): Mark debug insns.
(sd_finish_insn): Unmark them.
(sd_add_dep): Reject non-debug deps on debug insns.
(fixup_sched_groups): Give debug insns group treatment.
Skip debug insns.
(sched_analyze_reg): Don't mark debug insns for sched before call.
(sched_analyze_2): Handle debug insns.
(sched_analyze_insn): Compute next non-debug insn. Handle debug
insns.
(deps_analyze_insn): Handle debug insns.
(deps_start_bb): Skip debug insns.
(init_deps): Initialize last_debug_insn.
* tree-ssa.c (target_for_debug_bind): New.
(find_released_ssa_name): New.
(propagate_var_def_into_debug_stmts): New.
(propagate_defs_into_debug_stmts): New.
(verify_ssa): Skip debug bind stmts without values.
(warn_uninialized_vars): Skip debug stmts.
* target-def.h (TARGET_DELEGITIMIZE_ADDRESS): Set default.
* rtl.c (rtx_equal_p_cb): Handle VALUEs.
(rtx_equal_p): Likewise.
* ira-costs.c (scan_one_insn): Skip debug insns.
(process_bb_node_for_hard_reg_moves): Likewise.
* rtl.h (DEBUG_INSN_P): New.
(NONDEBUG_INSN_P): New.
(MAY_HAVE_DEBUG_INSNS): New.
(INSN_P): Accept debug insns.
(RTX_FRAME_RELATED_P): Likewise.
(INSN_DELETED_P): Likewise
(PAT_VAR_LOCATION_DECL): New.
(PAT_VAR_LOCATION_LOC): New.
(PAT_VAR_OCATION_STATUS): New.
(NOTE_VAR_LOCATION_DECL): Reimplement.
(NOTE_VAR_LOCATION_LOC): Likewise.
(NOTE_VAR_LOCATION_STATUS): Likewise.
(INSN_VAR_LOCATION): New.
(INSN_VAR_LOCATION_DECL): New.
(INSN_VAR_LOCATION_LOC): New.
(INSN_VAR_LOCATION_STATUS): New.
(gen_rtx_UNKNOWN_VAR_LOC): New.
(VAR_LOC_UNKNOWN_P): New.
(NOTE_DURING_CALL_P): New.
(SCHED_GROUP_P): Accept debug insns.
(emit_debug_insn_before): Declare.
(emit_debug_insn_before_noloc): Declare.
(emit_debug_insn_beore_setloc): Declare.
(emit_debug_insn_after): Declare.
(emit_debug_insn_after_noloc): Declare.
(emit_debug_insn_after_setloc): Declare.
(emit_debug_insn): Declare.
(make_debug_insn_raw): Declare.
(prev_nondebug_insn): Declare.
(next_nondebug_insn): Declare.
(delegitimize_mem_from_attrs): Declare.
(get_max_insn_count): Declare.
(wrap_constant): Declare.
(unwrap_constant): Declare.
(get_curr_insn_source_location): Declare.
(get_curr_insn_block): Declare.
* tree-inline.c (insert_debug_decl_map): New.
(processing_debug_stmt): New.
(remap_decl): Don't create new mappings in debug stmts.
(remap_gimple_op_r): Don't add references in debug stmts.
(copy_tree_body_r): Likewise.
(remap_gimple_stmt): Handle debug bind stmts.
(copy_bb): Skip debug stmts.
(copy_edges_for_bb): Likewise.
(copy_debug_stmt): New.
(copy_debug_stmts): New.
(copy_body): Copy debug stmts at the end.
(insert_init_debug_bind): New.
(insert_init_stmt): Take id. Skip and emit debug stmts.
(setup_one_parameter): Remap variable earlier, register debug
mapping.
(estimate_num_insns): Skip debug stmts.
(expand_call_inline): Preserve debug_map.
(optimize_inline_calls): Check for no debug_stmts left-overs.
(unsave_expr_now): Preserve debug_map.
(copy_gimple_seq_and_replace_locals): Likewise.
(tree_function_versioning): Check for no debug_stmts left-overs.
Init and destroy debug_map as needed. Split edges unconditionally.
(build_duplicate_type): Init and destroy debug_map as needed.
* tree-inline.h: Include gimple.h instead of pointer-set.h.
(struct copy_body_data): Add debug_stmts and debug_map.
* sched-int.h (struct ready_list): Add n_debug.
(struct deps): Add last_debug_insn.
(DEBUG_INSN_SCHED_P): New.
(BOUNDARY_DEBUG_INSN_P): New.
(SCHEDULE_DEBUG_INSN_P): New.
(sd_iterator_cond): Accept empty list.
* combine.c (create_log_links): Skip debug insns.
(combine_instructions): Likewise.
(cleanup_auto_inc_dec): New. From Jakub Jelinek: Make sure the
return value is always unshared.
(struct rtx_subst_pair): New.
(auto_adjust_pair): New.
(propagate_for_debug_subst): New.
(propagate_for_debug): New.
(try_combine): Skip debug insns. Propagate removed defs into
debug insns.
(next_nonnote_nondebug_insn): New.
(distribute_notes): Use it. Skip debug insns.
(distribute_links): Skip debug insns.
* tree-outof-ssa.c (set_location_for_edge): Likewise.
* resource.c (mark_target_live_regs): Likewise.
* var-tracking.c: Include cselib.h and target.h.
(enum micro_operation_type): Add MO_VAL_USE, MO_VAL_LOC, and
MO_VAL_SET.
(micro_operation_type_name): New.
(enum emit_note_where): Add EMIT_NOTE_AFTER_CALL_INSN.
(struct micro_operation_def): Update comments.
(decl_or_value): New type. Use instead of decls.
(struct emit_note_data_def): Add vars.
(struct attrs_def): Use decl_or_value.
(struct variable_tracking_info_def): Add permp, flooded.
(struct location_chain_def): Update comment.
(struct variable_part_def): Use decl_or_value.
(struct variable_def): Make var_part a variable length array.
(valvar_pool): New.
(scratch_regs): New.
(cselib_hook_called): New.
(dv_is_decl_p): New.
(dv_is_value_p): New.
(dv_as_decl): New.
(dv_as_value): New.
(dv_as_opaque): New.
(dv_onepart_p): New.
(dv_pool): New.
(IS_DECL_CODE): New.
(check_value_is_not_decl): New.
(dv_from_decl): New.
(dv_from_value): New.
(dv_htab_hash): New.
(variable_htab_hash): Use it.
(variable_htab_eq): Support values.
(variable_htab_free): Free from the right pool.
(attrs_list_member, attrs_list_insert): Use decl_or_value.
(attrs_list_union): Adjust.
(attrs_list_mpdv_union): New.
(tie_break_pointers): New.
(canon_value_cmp): New.
(unshare_variable): Return possibly-modified slot.
(vars_copy_1): Adjust.
(var_reg_decl_set): Adjust. Split out of...
(var_reg_set): ... this.
(get_init_value): Adjust.
(var_reg_delete_and_set): Adjust.
(var_reg_delete): Adjust.
(var_regno_delete): Adjust.
(var_mem_decl_set): Split out of...
(var_mem_set): ... this.
(var_mem_delete_and_set): Adjust.
(var_mem_delete): Adjust.
(val_store): New.
(val_reset): New.
(val_resolve): New.
(variable_union): Adjust. Speed up merge of 1-part vars.
(variable_canonicalize): Use unshared slot.
(VALUED_RECURSED_INTO): New.
(find_loc_in_1pdv): New.
(struct dfset_merge): New.
(insert_into_intersection): New.
(intersect_loc_chains): New.
(loc_cmp): New.
(canonicalize_loc_order_check): New.
(canonicalize_values_mark): New.
(canonicalize_values_star): New.
(variable_merge_over_cur): New.
(variable_merge_over_src): New.
(dataflow_set_merge): New.
(dataflow_set_equiv_regs): New.
(remove_duplicate_values): New.
(struct dfset_post_merge): New.
(variable_post_merge_new_vals): New.
(variable_post_merge_perm_vals): New.
(dataflow_post_merge_adjust): New.
(find_mem_expr_in_1pdv): New.
(dataflow_set_preserve_mem_locs): New.
(dataflow_set_remove_mem_locs): New.
(dataflow_set_clear_at_call): New.
(onepart_variable_different_p): New.
(variable_different_p): Use it.
(dataflow_set_different_1): Adjust. Make detailed dump
more verbose.
(track_expr_p): Add need_rtl parameter. Don't generate rtl
if not needed.
(track_loc_p): Pass it true.
(struct count_use_info): New.
(find_use_val): New.
(replace_expr_with_values): New.
(log_op_type): New.
(use_type): New, partially split out of...
(count_uses): ... this. Count new micro-ops.
(count_uses_1): Adjust.
(count_stores): Adjust.
(count_with_sets): New.
(VAL_NEEDS_RESOLUTION): New.
(VAL_HOLDS_TRACK_EXPR): New.
(VAL_EXPR_IS_COPIED): New.
(VAL_EXPR_IS_CLOBBERED): New.
(add_uses): Adjust. Generate new micro-ops.
(add_uses_1): Adjust.
(add_stores): Generate new micro-ops.
(add_with_sets): New.
(find_src_status): Adjust.
(find_src_set_src): Adjust.
(compute_bb_dataflow): Use dataflow_set_clear_at_call.
Handle new micro-ops. Canonicalize value equivalances.
(vt_find_locations): Compute total size of hash tables for
dumping. Perform merge for var-tracking-assignments. Don't
disregard single-block loops.
(dump_attrs_list): Handle decl_or_value.
(dump_variable): Take variable. Deal with decl_or_value.
(dump_variable_slot): New.
(dump_vars): Use it.
(dump_dataflow_sets): Adjust.
(set_slot_part): New, extended to support one-part variables
after splitting out of...
(set_variable_part): ... this.
(clobber_slot_part): New, split out of...
(clobber_variable_part): ... this.
(delete_slot_part): New, split out of...
(delete_variable_part): .... this.
(check_wrap_constant): New.
(vt_expand_loc_callback): New.
(vt_expand_loc): New.
(emit_note_insn_var_location): Adjust. Handle values. Handle
EMIT_NOTE_AFTER_CALL_INSN.
(emit_notes_for_differences_1): Adjust. Handle values.
(emit_notes_for_differences_2): Likewise.
(emit_notes_for_differences): Adjust.
(emit_notes_in_bb): Take pointer to set. Emit AFTER_CALL_INSN
notes. Adjust. Handle new micro-ops.
(vt_add_function_parameters): Adjust. Create and bind values.
(vt_initialize): Adjust. Initialize scratch_regs and
valvar_pool, flooded and perm.. Initialize and use cselib. Log
operations. Move some code to count_with_sets and add_with_sets.
(delete_debug_insns): New.
(vt_debug_insns_local): New.
(vt_finalize): Release permp, valvar_pool, scratch_regs. Finish
cselib.
(var_tracking_main): If var-tracking-assignments is enabled
but var-tracking isn't, delete debug insns and leave. Likewise
if we exceed limits or fail the stack adjustments tests, and
after all var-tracking processing.
More in var-tracking, from Jakub Jelinek <jakub@redhat.com>:
(dataflow_set): Add traversed_vars.
(value_chain, const_value_chain): New typedefs.
(value_chain_pool, value_chains): New variables.
(value_chain_htab_hash, value_chain_htab_eq, add_value_chain,
add_value_chains, add_cselib_value_chains, remove_value_chain,
remove_value_chains, remove_cselib_value_chains): New functions.
(shared_hash_find_slot_unshare_1, shared_hash_find_slot_1,
shared_hash_find_slot_noinsert_1, shared_hash_find_1): New
static inlines.
(shared_hash_find_slot_unshare, shared_hash_find_slot,
shared_hash_find_slot_noinsert, shared_hash_find): Update.
(dst_can_be_shared): New variable.
(unshare_variable): Unshare set->vars if shared, use shared_hash_*.
Clear dst_can_be_shared. If set->traversed_vars is non-NULL and
different from set->vars, look up slot again instead of using the
passed in slot.
(dataflow_set_init): Initialize traversed_vars.
(variable_union): Use shared_hash_*. Use initially NO_INSERT
lookup if set->vars is shared. Don't keep slot cleared before
calling unshare_variable. Unshare set->vars if needed. Adjust
unshare_variable callers. Clear dst_can_be_shared if needed.
Even ->refcount == 1 vars must be unshared if set->vars is shared
and var needs to be modified.
(dataflow_set_union): Set traversed_vars during canonicalization.
(VALUE_CHANGED, DECL_CHANGED): Define.
(set_dv_changed, dv_changed_p): New static inlines.
(track_expr_p): Clear DECL_CHANGED.
(dump_dataflow_sets): Set it.
(variable_was_changed): Call set_dv_changed.
(emit_note_insn_var_location): Likewise.
(changed_variables_stack): New variable.
(check_changed_vars_1, check_changed_vars_2): New functions.
(emit_notes_for_changes): Do nothing if changed_variables is
empty. Traverse changed_variables with check_changed_vars_1,
call check_changed_vars_2 on each changed_variables_stack entry.
(emit_notes_in_bb): Add SET argument. Just clear it at the
beginning, use it instead of local &set, don't destroy it at the
end.
(vt_emit_notes): Call dataflow_set_clear early on all
VTI(bb)->out sets, never use them, instead use emit_notes_in_bb
computed set, dataflow_set_clear also VTI(bb)->in when we are
done with the basic block. Initialize changed_variables_stack,
free it afterwards. If ENABLE_CHECKING verify that after noting
differences to an empty set value_chains hash table is empty.
(vt_initialize): Initialize value_chains and value_chain_pool.
(vt_finalize): Delete value_chains htab, free value_chain_pool.
(variable_tracking_main): Call dump_dataflow_sets before calling
vt_emit_notes, not after it.
* tree-flow.h (propagate_defs_into_debug_stmts): Declare.
(propagate_var_def_into_debug_stmts): Declare.
* df-problems.c (df_lr_bb_local_compute): Skip debug insns.
(df_set_note): Reject debug insns.
(df_whole_mw_reg_dead_p): Take added_notes_p argument. Don't
add notes to debug insns.
(df_note_bb_compute): Adjust. Likewise.
(df_simulate_uses): Skip debug insns.
(df_simulate_initialize_backwards): Likewise.
* reg-stack.c (subst_stack_regs_in_debug_insn): New.
(subst_stack_regs_pat): Reject debug insns.
(convert_regs_1): Handle debug insns.
* Makefile.in (TREE_INLINE_H): Take pointer-set.h from GIMPLE_H.
(print-rtl.o): Depend on cselib.h.
(cselib.o): Depend on TREE_PASS_H.
(var-tracking.o): Depend on cselib.h and TARGET_H.
* sched-rgn.c (rgn_estimate_number_of_insns): Discount
debug insns.
(init_ready_list): Skip boundary debug insns.
(add_branch_dependences): Skip debug insns.
(free_block_dependencies): Check for blocks with only debug
insns.
(compute_priorities): Likewise.
* gimple.c (gss_for_code): Handle GIMPLE_DEBUG.
(gimple_build_with_ops_stat): Take subcode as unsigned. Adjust
all callers.
(gimple_build_debug_bind_stat): New.
(empty_body_p): Skip debug stmts.
(gimple_has_side_effects): Likewise.
(gimple_rhs_has_side_effects): Likewise.
* gimple.h (enum gimple_debug_subcode, GIMPLE_DEBUG_BIND): New.
(gimple_build_debug_bind_stat): Declare.
(gimple_build_debug_bind): Define.
(is_gimple_debug): New.
(gimple_debug_bind_p): New.
(gimple_debug_bind_get_var): New.
(gimple_debug_bind_get_value): New.
(gimple_debug_bind_get_value_ptr): New.
(gimple_debug_bind_set_var): New.
(gimple_debug_bind_set_value): New.
(GIMPLE_DEBUG_BIND_NOVALUE): New internal temporary macro.
(gimple_debug_bind_reset_value): New.
(gimple_debug_bind_has_value_p): New.
(gsi_next_nondebug): New.
(gsi_prev_nondebug): New.
(gsi_start_nondebug_bb): New.
(gsi_last_nondebug_bb): New.
* sched-vis.c (print_pattern): Handle VAR_LOCATION.
(print_insn): Handle DEBUG_INSN.
* tree-cfg.c (remove_bb): Walk stmts backwards. Let loc
of first insn prevail.
(first_stmt): Skip debug stmts.
(first_non_label_stmt): Likewise.
(last_stmt): Likewise.
(has_zero_uses_1): New.
(single_imm_use_1): New.
(verify_gimple_debug): New.
(verify_types_in_gimple_stmt): Handle debug stmts.
(verify_stmt): Likewise.
(debug_loop_num): Skip debug stmts.
(remove_edge_and_dominated_blocks): Remove dominators last.
* tree-ssa-reasssoc.c (rewrite_expr_tree): Propagate into
debug stmts.
(linearize_expr): Likewise.
* config/i386/i386.c (ix86_delegitimize_address): Call
default implementation.
* config/ia64/ia64.c (ia64_safe_itanium_class): Handle debug
insns.
(group_barrier_needed): Skip debug insns.
(emit_insn_group_barriers): Likewise.
(emit_all_insn_group_barriers): Likewise.
(ia64_variable_issue): Handle debug insns.
(ia64_dfa_new_cycle): Likewise.
(final_emit_insn_group_barriers): Skip debug insns.
(ia64_dwarf2out_def_steady_cfa): Take frame argument. Don't
def cfa without frame.
(process_set): Likewise.
(process_for_unwind_directive): Pass frame on.
* config/rs6000/rs6000.c (TARGET_DELEGITIMIZE_ADDRESS): Define.
(rs6000_delegitimize_address): New.
(rs6000_debug_adjust_cost): Handle debug insns.
(is_microcoded_insn): Likewise.
(is_cracked_insn): Likewise.
(is_nonpipeline_insn): Likewise.
(insn_must_be_first_in_group): Likewise.
(insn_must_be_last_in_group): Likewise.
(force_new_group): Likewise.
* cfgrtl.c (rtl_split_block): Emit INSN_DELETED note if block
contains only debug insns.
(rtl_merge_blocks): Skip debug insns.
(purge_dead_edges): Likewise.
(rtl_block_ends_with_call_p): Skip debug insns.
* dce.c (deletable_insn_p): Handle VAR_LOCATION.
(mark_reg_dependencies): Skip debug insns.
* params.def (PARAM_MIN_NONDEBUG_INSN_UID): New.
* tree-ssanames.c (release_ssa_name): Propagate def into
debug stmts.
* tree-ssa-threadedge.c
(record_temporary_equivalences_from_stmts): Skip debug stmts.
* regcprop.c (replace_oldest_value_addr): Skip debug insns.
(replace_oldest_value_mem): Use ALL_REGS for debug insns.
(copyprop_hardreg_forward_1): Handle debug insns.
* reload1.c (reload): Skip debug insns. Replace unassigned
pseudos in debug insns with their equivalences.
(eliminate_regs_in_insn): Skip debug insns.
(emit_input_reload_insns): Skip debug insns at first, adjust
them later.
* tree-ssa-operands.c (add_virtual_operand): Reject debug stmts.
(get_indirect_ref_operands): Pass opf_no_vops on.
(get_expr_operands): Likewise. Skip debug stmts.
(parse_ssa_operands): Scan debug insns with opf_no_vops.
gcc/testsuite/ChangeLog:
* gcc.dg/guality/guality.c: New.
* gcc.dg/guality/guality.h: New.
* gcc.dg/guality/guality.exp: New.
* gcc.dg/guality/example.c: New.
* lib/gcc-dg.exp (cleanup-dump): Remove .gk files.
(cleanup-saved-temps): Likewise, .gkd files too.
gcc/cp/ChangeLog:
* cp-tree.h (TFF_NO_OMIT_DEFAULT_TEMPLATE_ARGUMENTS): New.
* cp-lang.c (cxx_dwarf_name): Pass it.
* error.c (count_non_default_template_args): Take flags as
argument. Adjust all callers. Skip counting of default
arguments if the new flag is given.
ChangeLog:
* Makefile.tpl (BUILD_CONFIG): Default to bootstrap-debug.
* Makefile.in: Rebuilt.
contrib/ChangeLog:
* compare-debug: Look for .gkd files and compare them.
config/ChangeLog:
* bootstrap-debug.mk: Add comments.
* bootstrap-debug-big.mk: New.
* bootstrap-debug-lean.mk: New.
* bootstrap-debug-ckovw.mk: Add comments.
* bootstrap-debug-lib.mk: Drop CFLAGS for stages. Use -g0
for TFLAGS in stage1. Drop -fvar-tracking-assignments-toggle.
From-SVN: r151312
2009-09-02 04:42:21 +02:00
2009-09-17 13:35:38 +02:00
DEFPARAM ( PARAM_IPA_SRA_PTR_GROWTH_FACTOR ,
"ipa-sra-ptr-growth-factor" ,
2009-11-18 12:09:50 +01:00
"Maximum allowed growth of size of new parameters ipa-sra replaces "
2015-10-22 00:24:41 +02:00
"a pointer to an aggregate with." ,
2009-09-17 13:35:38 +02:00
2 , 0 , 0 )
2011-11-08 12:13:41 +01:00
DEFPARAM ( PARAM_TM_MAX_AGGREGATE_SIZE ,
"tm-max-aggregate-size" ,
"Size in bytes after which thread-local aggregates should be "
"instrumented with the logging functions instead of save/restore "
2015-10-22 00:24:41 +02:00
"pairs." ,
2011-11-08 12:13:41 +01:00
9 , 0 , 0 )
2014-11-06 15:18:52 +01:00
DEFPARAM ( PARAM_SRA_MAX_SCALARIZATION_SIZE_SPEED ,
"sra-max-scalarization-size-Ospeed" ,
"Maximum size, in storage units, of an aggregate which should be "
2015-10-22 00:24:41 +02:00
"considered for scalarization when compiling for speed." ,
2014-11-06 15:18:52 +01:00
0 , 0 , 0 )
DEFPARAM ( PARAM_SRA_MAX_SCALARIZATION_SIZE_SIZE ,
"sra-max-scalarization-size-Osize" ,
"Maximum size, in storage units, of an aggregate which should be "
2015-10-22 00:24:41 +02:00
"considered for scalarization when compiling for size." ,
2014-11-06 15:18:52 +01:00
0 , 0 , 0 )
2011-07-19 00:03:39 +02:00
DEFPARAM ( PARAM_IPA_CP_VALUE_LIST_SIZE ,
"ipa-cp-value-list-size" ,
"Maximum size of a list of values associated with each parameter for "
2015-10-22 00:24:41 +02:00
"interprocedural constant propagation." ,
2010-08-05 15:23:07 +02:00
8 , 0 , 0 )
2011-07-19 00:03:39 +02:00
DEFPARAM ( PARAM_IPA_CP_EVAL_THRESHOLD ,
"ipa-cp-eval-threshold" ,
"Threshold ipa-cp opportunity evaluation that is still considered "
2015-10-22 00:24:41 +02:00
"beneficial to clone.." ,
2011-07-19 00:03:39 +02:00
500 , 0 , 0 )
2015-03-29 17:38:52 +02:00
DEFPARAM ( PARAM_IPA_CP_RECURSION_PENALTY ,
"ipa-cp-recursion-penalty" ,
"Percentage penalty the recursive functions will receive when they "
2015-10-22 00:24:41 +02:00
"are evaluated for cloning.." ,
2015-03-29 17:38:52 +02:00
40 , 0 , 100 )
DEFPARAM ( PARAM_IPA_CP_SINGLE_CALL_PENALTY ,
"ipa-cp-single-call-penalty" ,
2017-01-09 22:48:33 +01:00
"Percentage penalty functions containing a single call to another "
2015-10-22 00:24:41 +02:00
"function will receive when they are evaluated for cloning.." ,
2015-03-29 17:38:52 +02:00
15 , 0 , 100 )
2012-09-10 17:58:57 +02:00
DEFPARAM ( PARAM_IPA_MAX_AGG_ITEMS ,
"ipa-max-agg-items" ,
"Maximum number of aggregate content items for a parameter in "
2015-10-22 00:24:41 +02:00
"jump functions and lattices." ,
2012-09-10 17:58:57 +02:00
16 , 0 , 0 )
2012-11-07 16:55:54 +01:00
DEFPARAM ( PARAM_IPA_CP_LOOP_HINT_BONUS ,
"ipa-cp-loop-hint-bonus" ,
"Compile-time bonus IPA-CP assigns to candidates which make loop "
2015-10-22 00:24:41 +02:00
"bounds or strides known.." ,
2012-11-07 16:55:54 +01:00
64 , 0 , 0 )
2013-03-20 17:20:22 +01:00
DEFPARAM ( PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS ,
"ipa-cp-array-index-hint-bonus" ,
"Compile-time bonus IPA-CP assigns to candidates which make an array "
2015-10-22 00:24:41 +02:00
"index known.." ,
2013-03-20 17:20:22 +01:00
48 , 0 , 0 )
2014-05-23 17:52:20 +02:00
DEFPARAM ( PARAM_IPA_MAX_AA_STEPS ,
"ipa-max-aa-steps" ,
"Maximum number of statements that will be visited by IPA formal "
2015-10-22 00:24:41 +02:00
"parameter analysis based on alias analysis in any given function." ,
2014-05-23 17:52:20 +02:00
25000 , 0 , 0 )
2010-10-05 19:57:09 +02:00
/* WHOPR partitioning configuration. */
DEFPARAM ( PARAM_LTO_PARTITIONS ,
"lto-partitions" ,
2015-10-22 00:24:41 +02:00
"Number of partitions the program should be split to." ,
2011-03-21 14:42:27 +01:00
32 , 1 , 0 )
2010-10-05 19:57:09 +02:00
DEFPARAM ( MIN_PARTITION_SIZE ,
"lto-min-partition" ,
2015-10-22 00:24:41 +02:00
"Minimal size of a partition for LTO (in estimated instructions)." ,
2016-04-28 00:39:43 +02:00
10000 , 0 , 0 )
2010-12-07 14:46:29 +01:00
2016-04-27 12:46:16 +02:00
DEFPARAM ( MAX_PARTITION_SIZE ,
"lto-max-partition" ,
"Maximal size of a partition for LTO (in estimated instructions)." ,
1000000 , 0 , INT_MAX )
2010-12-07 14:46:29 +01:00
/* Diagnostic parameters. */
DEFPARAM ( CXX_MAX_NAMESPACES_FOR_DIAGNOSTIC_HELP ,
"cxx-max-namespaces-for-diagnostic-help" ,
"Maximum number of namespaces to search for alternatives when "
2015-10-22 00:24:41 +02:00
"name lookup fails." ,
2010-12-07 14:46:29 +01:00
1000 , 0 , 0 )
2011-03-24 09:23:39 +01:00
/* Maximum number of conditional store pairs that can be sunk. */
DEFPARAM ( PARAM_MAX_STORES_TO_SINK ,
"max-stores-to-sink" ,
2015-10-22 00:24:41 +02:00
"Maximum number of conditional store pairs that can be sunk." ,
2011-03-24 09:23:39 +01:00
2 , 0 , 0 )
2011-07-05 19:45:38 +02:00
/ * Override CASE_VALUES_THRESHOLD of when to switch from doing switch
statements via if statements to using a table jump operation . If the value
is 0 , the default CASE_VALUES_THRESHOLD will be used . * /
DEFPARAM ( PARAM_CASE_VALUES_THRESHOLD ,
"case-values-threshold" ,
"The smallest number of different values for which it is best to "
"use a jump-table instead of a tree of conditional branches, "
2015-10-22 00:24:41 +02:00
"if 0, use the default for the machine." ,
2011-07-05 19:45:38 +02:00
0 , 0 , 0 )
2011-07-27 15:22:09 +02:00
/* Data race flags for C++0x memory model compliance. */
DEFPARAM ( PARAM_ALLOW_STORE_DATA_RACES ,
"allow-store-data-races" ,
2015-10-22 00:24:41 +02:00
"Allow new data races on stores to be introduced." ,
2014-06-26 16:54:59 +02:00
0 , 0 , 1 )
2011-11-06 15:55:48 +01:00
2011-09-06 18:42:47 +02:00
/* Reassociation width to be used by tree reassoc optimization. */
DEFPARAM ( PARAM_TREE_REASSOC_WIDTH ,
"tree-reassoc-width" ,
"Set the maximum number of instructions executed in parallel in "
2015-10-22 00:24:41 +02:00
"reassociated tree. If 0, use the target dependent heuristic.." ,
2011-09-06 18:42:47 +02:00
0 , 0 , 0 )
re PR middle-end/43864 (Same basic blocks should be merged)
2011-09-27 Tom de Vries <tom@codesourcery.com>
PR middle-end/43864
* tree-ssa-tail-merge.c: New file.
(struct same_succ_def): Define.
(same_succ, const_same_succ): New typedef.
(struct bb_cluster_def): Define.
(bb_cluster, const_bb_cluster): New typedef.
(struct aux_bb_info): Define.
(BB_SIZE, BB_SAME_SUCC, BB_CLUSTER, BB_VOP_AT_EXIT): Define.
(gvn_uses_equal): New function.
(same_succ_print, same_succ_print_traverse, update_dep_bb)
(stmt_update_dep_bb, local_def, same_succ_hash)
(inverse_flags, same_succ_equal, same_succ_alloc, same_succ_delete)
(same_succ_reset): New function.
(same_succ_htab, same_succ_edge_flags)
(deleted_bbs, deleted_bb_preds): New var.
(debug_same_succ): New function.
(worklist): New var.
(print_worklist, add_to_worklist, find_same_succ_bb, find_same_succ)
(init_worklist, delete_worklist, delete_basic_block_same_succ)
(same_succ_flush_bbs, purge_bbs, update_worklist): New function.
(print_cluster, debug_cluster, update_rep_bb)
(add_bb_to_cluster, new_cluster, delete_cluster): New function.
(all_clusters): New var.
(alloc_cluster_vectors, reset_cluster_vectors, delete_cluster_vectors)
(merge_clusters, set_cluster): New function.
(gimple_equal_p, gsi_advance_bw_nondebug_nonlocal, find_duplicate)
(same_phi_alternatives_1, same_phi_alternatives, bb_has_non_vop_phi)
(deps_ok_for_redirect_from_bb_to_bb, deps_ok_for_redirect)
(find_clusters_1, find_clusters): New function.
(update_vuses, vop_phi, vop_at_entry, replace_block_by): New function.
(update_bbs): New var.
(apply_clusters): New function.
(update_debug_stmt, update_debug_stmts): New function.
(tail_merge_optimize): New function.
tree-pass.h (tail_merge_optimize): Declare.
* tree-ssa-pre.c (execute_pre): Use tail_merge_optimize.
* Makefile.in (OBJS-common): Add tree-ssa-tail-merge.o.
(tree-ssa-tail-merge.o): New rule.
* opts.c (default_options_table): Set OPT_ftree_tail_merge by default at
OPT_LEVELS_2_PLUS.
* tree-ssa-sccvn.c (vn_valueize): Move to ...
* tree-ssa-sccvn.h (vn_valueize): Here.
* timevar.def (TV_TREE_TAIL_MERGE): New timevar.
* common.opt (ftree-tail-merge): New switch.
* params.def (PARAM_MAX_TAIL_MERGE_COMPARISONS)
(PARAM_MAX_TAIL_MERGE_ITERATIONS): New parameter.
* doc/invoke.texi (Optimization Options, -O2): Add -ftree-tail-merge.
(-ftree-tail-merge, max-tail-merge-comparisons)
(max-tail-merge-iterations): New item.
From-SVN: r179275
2011-09-27 18:10:42 +02:00
DEFPARAM ( PARAM_MAX_TAIL_MERGE_COMPARISONS ,
"max-tail-merge-comparisons" ,
2015-10-22 00:24:41 +02:00
"Maximum amount of similar bbs to compare a bb with." ,
re PR middle-end/43864 (Same basic blocks should be merged)
2011-09-27 Tom de Vries <tom@codesourcery.com>
PR middle-end/43864
* tree-ssa-tail-merge.c: New file.
(struct same_succ_def): Define.
(same_succ, const_same_succ): New typedef.
(struct bb_cluster_def): Define.
(bb_cluster, const_bb_cluster): New typedef.
(struct aux_bb_info): Define.
(BB_SIZE, BB_SAME_SUCC, BB_CLUSTER, BB_VOP_AT_EXIT): Define.
(gvn_uses_equal): New function.
(same_succ_print, same_succ_print_traverse, update_dep_bb)
(stmt_update_dep_bb, local_def, same_succ_hash)
(inverse_flags, same_succ_equal, same_succ_alloc, same_succ_delete)
(same_succ_reset): New function.
(same_succ_htab, same_succ_edge_flags)
(deleted_bbs, deleted_bb_preds): New var.
(debug_same_succ): New function.
(worklist): New var.
(print_worklist, add_to_worklist, find_same_succ_bb, find_same_succ)
(init_worklist, delete_worklist, delete_basic_block_same_succ)
(same_succ_flush_bbs, purge_bbs, update_worklist): New function.
(print_cluster, debug_cluster, update_rep_bb)
(add_bb_to_cluster, new_cluster, delete_cluster): New function.
(all_clusters): New var.
(alloc_cluster_vectors, reset_cluster_vectors, delete_cluster_vectors)
(merge_clusters, set_cluster): New function.
(gimple_equal_p, gsi_advance_bw_nondebug_nonlocal, find_duplicate)
(same_phi_alternatives_1, same_phi_alternatives, bb_has_non_vop_phi)
(deps_ok_for_redirect_from_bb_to_bb, deps_ok_for_redirect)
(find_clusters_1, find_clusters): New function.
(update_vuses, vop_phi, vop_at_entry, replace_block_by): New function.
(update_bbs): New var.
(apply_clusters): New function.
(update_debug_stmt, update_debug_stmts): New function.
(tail_merge_optimize): New function.
tree-pass.h (tail_merge_optimize): Declare.
* tree-ssa-pre.c (execute_pre): Use tail_merge_optimize.
* Makefile.in (OBJS-common): Add tree-ssa-tail-merge.o.
(tree-ssa-tail-merge.o): New rule.
* opts.c (default_options_table): Set OPT_ftree_tail_merge by default at
OPT_LEVELS_2_PLUS.
* tree-ssa-sccvn.c (vn_valueize): Move to ...
* tree-ssa-sccvn.h (vn_valueize): Here.
* timevar.def (TV_TREE_TAIL_MERGE): New timevar.
* common.opt (ftree-tail-merge): New switch.
* params.def (PARAM_MAX_TAIL_MERGE_COMPARISONS)
(PARAM_MAX_TAIL_MERGE_ITERATIONS): New parameter.
* doc/invoke.texi (Optimization Options, -O2): Add -ftree-tail-merge.
(-ftree-tail-merge, max-tail-merge-comparisons)
(max-tail-merge-iterations): New item.
From-SVN: r179275
2011-09-27 18:10:42 +02:00
10 , 0 , 0 )
2016-10-28 16:18:50 +02:00
DEFPARAM ( PARAM_STORE_MERGING_ALLOW_UNALIGNED ,
"store-merging-allow-unaligned" ,
"Allow the store merging pass to introduce unaligned stores "
2017-03-07 10:36:44 +01:00
"if it is legal to do so." ,
2016-10-28 16:18:50 +02:00
1 , 0 , 1 )
DEFPARAM ( PARAM_MAX_STORES_TO_MERGE ,
"max-stores-to-merge" ,
2017-01-07 09:37:06 +01:00
"Maximum number of constant stores to merge in the "
2017-03-07 10:36:44 +01:00
"store merging pass." ,
2016-10-28 16:18:50 +02:00
64 , 2 , 0 )
re PR middle-end/43864 (Same basic blocks should be merged)
2011-09-27 Tom de Vries <tom@codesourcery.com>
PR middle-end/43864
* tree-ssa-tail-merge.c: New file.
(struct same_succ_def): Define.
(same_succ, const_same_succ): New typedef.
(struct bb_cluster_def): Define.
(bb_cluster, const_bb_cluster): New typedef.
(struct aux_bb_info): Define.
(BB_SIZE, BB_SAME_SUCC, BB_CLUSTER, BB_VOP_AT_EXIT): Define.
(gvn_uses_equal): New function.
(same_succ_print, same_succ_print_traverse, update_dep_bb)
(stmt_update_dep_bb, local_def, same_succ_hash)
(inverse_flags, same_succ_equal, same_succ_alloc, same_succ_delete)
(same_succ_reset): New function.
(same_succ_htab, same_succ_edge_flags)
(deleted_bbs, deleted_bb_preds): New var.
(debug_same_succ): New function.
(worklist): New var.
(print_worklist, add_to_worklist, find_same_succ_bb, find_same_succ)
(init_worklist, delete_worklist, delete_basic_block_same_succ)
(same_succ_flush_bbs, purge_bbs, update_worklist): New function.
(print_cluster, debug_cluster, update_rep_bb)
(add_bb_to_cluster, new_cluster, delete_cluster): New function.
(all_clusters): New var.
(alloc_cluster_vectors, reset_cluster_vectors, delete_cluster_vectors)
(merge_clusters, set_cluster): New function.
(gimple_equal_p, gsi_advance_bw_nondebug_nonlocal, find_duplicate)
(same_phi_alternatives_1, same_phi_alternatives, bb_has_non_vop_phi)
(deps_ok_for_redirect_from_bb_to_bb, deps_ok_for_redirect)
(find_clusters_1, find_clusters): New function.
(update_vuses, vop_phi, vop_at_entry, replace_block_by): New function.
(update_bbs): New var.
(apply_clusters): New function.
(update_debug_stmt, update_debug_stmts): New function.
(tail_merge_optimize): New function.
tree-pass.h (tail_merge_optimize): Declare.
* tree-ssa-pre.c (execute_pre): Use tail_merge_optimize.
* Makefile.in (OBJS-common): Add tree-ssa-tail-merge.o.
(tree-ssa-tail-merge.o): New rule.
* opts.c (default_options_table): Set OPT_ftree_tail_merge by default at
OPT_LEVELS_2_PLUS.
* tree-ssa-sccvn.c (vn_valueize): Move to ...
* tree-ssa-sccvn.h (vn_valueize): Here.
* timevar.def (TV_TREE_TAIL_MERGE): New timevar.
* common.opt (ftree-tail-merge): New switch.
* params.def (PARAM_MAX_TAIL_MERGE_COMPARISONS)
(PARAM_MAX_TAIL_MERGE_ITERATIONS): New parameter.
* doc/invoke.texi (Optimization Options, -O2): Add -ftree-tail-merge.
(-ftree-tail-merge, max-tail-merge-comparisons)
(max-tail-merge-iterations): New item.
From-SVN: r179275
2011-09-27 18:10:42 +02:00
DEFPARAM ( PARAM_MAX_TAIL_MERGE_ITERATIONS ,
"max-tail-merge-iterations" ,
2015-10-22 00:24:41 +02:00
"Maximum amount of iterations of the pass over a function." ,
re PR middle-end/43864 (Same basic blocks should be merged)
2011-09-27 Tom de Vries <tom@codesourcery.com>
PR middle-end/43864
* tree-ssa-tail-merge.c: New file.
(struct same_succ_def): Define.
(same_succ, const_same_succ): New typedef.
(struct bb_cluster_def): Define.
(bb_cluster, const_bb_cluster): New typedef.
(struct aux_bb_info): Define.
(BB_SIZE, BB_SAME_SUCC, BB_CLUSTER, BB_VOP_AT_EXIT): Define.
(gvn_uses_equal): New function.
(same_succ_print, same_succ_print_traverse, update_dep_bb)
(stmt_update_dep_bb, local_def, same_succ_hash)
(inverse_flags, same_succ_equal, same_succ_alloc, same_succ_delete)
(same_succ_reset): New function.
(same_succ_htab, same_succ_edge_flags)
(deleted_bbs, deleted_bb_preds): New var.
(debug_same_succ): New function.
(worklist): New var.
(print_worklist, add_to_worklist, find_same_succ_bb, find_same_succ)
(init_worklist, delete_worklist, delete_basic_block_same_succ)
(same_succ_flush_bbs, purge_bbs, update_worklist): New function.
(print_cluster, debug_cluster, update_rep_bb)
(add_bb_to_cluster, new_cluster, delete_cluster): New function.
(all_clusters): New var.
(alloc_cluster_vectors, reset_cluster_vectors, delete_cluster_vectors)
(merge_clusters, set_cluster): New function.
(gimple_equal_p, gsi_advance_bw_nondebug_nonlocal, find_duplicate)
(same_phi_alternatives_1, same_phi_alternatives, bb_has_non_vop_phi)
(deps_ok_for_redirect_from_bb_to_bb, deps_ok_for_redirect)
(find_clusters_1, find_clusters): New function.
(update_vuses, vop_phi, vop_at_entry, replace_block_by): New function.
(update_bbs): New var.
(apply_clusters): New function.
(update_debug_stmt, update_debug_stmts): New function.
(tail_merge_optimize): New function.
tree-pass.h (tail_merge_optimize): Declare.
* tree-ssa-pre.c (execute_pre): Use tail_merge_optimize.
* Makefile.in (OBJS-common): Add tree-ssa-tail-merge.o.
(tree-ssa-tail-merge.o): New rule.
* opts.c (default_options_table): Set OPT_ftree_tail_merge by default at
OPT_LEVELS_2_PLUS.
* tree-ssa-sccvn.c (vn_valueize): Move to ...
* tree-ssa-sccvn.h (vn_valueize): Here.
* timevar.def (TV_TREE_TAIL_MERGE): New timevar.
* common.opt (ftree-tail-merge): New switch.
* params.def (PARAM_MAX_TAIL_MERGE_COMPARISONS)
(PARAM_MAX_TAIL_MERGE_ITERATIONS): New parameter.
* doc/invoke.texi (Optimization Options, -O2): Add -ftree-tail-merge.
(-ftree-tail-merge, max-tail-merge-comparisons)
(max-tail-merge-iterations): New item.
From-SVN: r179275
2011-09-27 18:10:42 +02:00
2 , 0 , 0 )
2011-03-24 09:23:39 +01:00
2011-09-27 18:16:57 +02:00
/ * Maximum number of strings for which strlen optimization pass will
track string lenths . * /
DEFPARAM ( PARAM_MAX_TRACKED_STRLENS ,
"max-tracked-strlens" ,
"Maximum number of strings for which strlen optimization pass will "
2015-10-22 00:24:41 +02:00
"track string lengths." ,
2015-01-08 22:30:56 +01:00
10000 , 0 , 0 )
2011-09-27 18:16:57 +02:00
2012-06-06 11:12:58 +02:00
/* Keep this in sync with the sched_pressure_algorithm enum. */
DEFPARAM ( PARAM_SCHED_PRESSURE_ALGORITHM ,
"sched-pressure-algorithm" ,
2015-10-22 00:24:41 +02:00
"Which -fsched-pressure algorithm to apply." ,
2012-06-06 11:12:58 +02:00
1 , 1 , 2 )
2012-09-11 14:23:25 +02:00
/* Maximum length of candidate scans in straight-line strength reduction. */
DEFPARAM ( PARAM_MAX_SLSR_CANDIDATE_SCAN ,
"max-slsr-cand-scan" ,
"Maximum length of candidate scans for straight-line "
2015-10-22 00:24:41 +02:00
"strength reduction." ,
2012-09-11 14:23:25 +02:00
50 , 1 , 999999 )
2014-01-09 08:31:05 +01:00
DEFPARAM ( PARAM_ASAN_STACK ,
"asan-stack" ,
2015-10-22 00:24:41 +02:00
"Enable asan stack protection." ,
2014-01-09 08:31:05 +01:00
1 , 0 , 1 )
DEFPARAM ( PARAM_ASAN_GLOBALS ,
"asan-globals" ,
2015-10-22 00:24:41 +02:00
"Enable asan globals protection." ,
2014-01-09 08:31:05 +01:00
1 , 0 , 1 )
DEFPARAM ( PARAM_ASAN_INSTRUMENT_WRITES ,
"asan-instrument-writes" ,
2015-10-22 00:24:41 +02:00
"Enable asan store operations protection." ,
2014-01-09 08:31:05 +01:00
1 , 0 , 1 )
DEFPARAM ( PARAM_ASAN_INSTRUMENT_READS ,
"asan-instrument-reads" ,
2015-10-22 00:24:41 +02:00
"Enable asan load operations protection." ,
2014-01-09 08:31:05 +01:00
1 , 0 , 1 )
DEFPARAM ( PARAM_ASAN_MEMINTRIN ,
"asan-memintrin" ,
2015-10-22 00:24:41 +02:00
"Enable asan builtin functions protection." ,
2014-01-09 08:31:05 +01:00
1 , 0 , 1 )
DEFPARAM ( PARAM_ASAN_USE_AFTER_RETURN ,
"asan-use-after-return" ,
2015-10-22 00:24:41 +02:00
"Enable asan detection of use-after-return bugs." ,
2014-01-09 08:31:05 +01:00
1 , 0 , 1 )
2014-02-21 10:53:56 +01:00
2014-06-16 10:43:47 +02:00
DEFPARAM ( PARAM_ASAN_INSTRUMENTATION_WITH_CALL_THRESHOLD ,
"asan-instrumentation-with-call-threshold" ,
2014-08-11 08:12:12 +02:00
"Use callbacks instead of inline code if number of accesses "
2015-10-22 00:24:41 +02:00
"in function becomes greater or equal to this number." ,
2014-08-11 08:12:12 +02:00
7000 , 0 , INT_MAX )
2014-06-16 10:43:47 +02:00
2016-11-07 11:23:38 +01:00
DEFPARAM ( PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD ,
"use-after-scope-direct-emission-threshold" ,
2017-01-09 22:48:33 +01:00
"Use direct poisoning/unpoisoning instructions for variables "
2016-11-07 11:23:38 +01:00
"smaller or equal to this number." ,
256 , 0 , INT_MAX )
2014-02-21 10:53:56 +01:00
DEFPARAM ( PARAM_UNINIT_CONTROL_DEP_ATTEMPTS ,
"uninit-control-dep-attempts" ,
"Maximum number of nested calls to search for control dependencies "
2015-10-22 00:24:41 +02:00
"during uninitialized variable analysis." ,
2014-02-21 10:53:56 +01:00
1000 , 1 , 0 )
ipa-chkp.c: New.
gcc/
2014-11-05 Ilya Enkovich <ilya.enkovich@intel.com>
* ipa-chkp.c: New.
* ipa-chkp.h: New.
* tree-chkp.c: New.
* tree-chkp.h: New.
* tree-chkp-opt.c: New.
* rtl-chkp.c: New.
* rtl-chkp.h: New.
* Makefile.in (OBJS): Add ipa-chkp.o, rtl-chkp.o, tree-chkp.o
tree-chkp-opt.o.
(GTFILES): Add tree-chkp.c.
* mode-classes.def (MODE_POINTER_BOUNDS): New.
* tree.def (POINTER_BOUNDS_TYPE): New.
* genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
(POINTER_BOUNDS_MODE): New.
(make_pointer_bounds_mode): New.
* machmode.h (POINTER_BOUNDS_MODE_P): New.
* stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
(layout_type): Support POINTER_BOUNDS_TYPE.
* tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
* tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
* tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
(type_contains_placeholder_1): Likewise.
(build_common_tree_nodes): Initialize
pointer_bounds_type_node.
* tree.h (POINTER_BOUNDS_TYPE_P): New.
(pointer_bounds_type_node): New.
(POINTER_BOUNDS_P): New.
(BOUNDED_TYPE_P): New.
(BOUNDED_P): New.
(CALL_WITH_BOUNDS_P): New.
* gimple.h (gf_mask): Add GF_CALL_WITH_BOUNDS.
(gimple_call_with_bounds_p): New.
(gimple_call_set_with_bounds): New.
(gimple_return_retbnd): New.
(gimple_return_set_retbnd): New
* gimple.c (gimple_build_return): Increase number of ops
for return statement.
(gimple_build_call_from_tree): Propagate CALL_WITH_BOUNDS_P
flag.
* gimple-pretty-print.c (dump_gimple_return): Print second op.
* rtl.h (CALL_EXPR_WITH_BOUNDS_P): New.
* gimplify.c (gimplify_init_constructor): Avoid infinite
loop during gimplification of bounds initializer.
* calls.c: Include tree-chkp.h, rtl-chkp.h, bitmap.h.
(special_function_p): Use original decl name when analyzing
instrumentation clone.
(arg_data): Add fields special_slot, pointer_arg and
pointer_offset.
(store_bounds): New.
(emit_call_1): Propagate instrumentation flag for CALL.
(initialize_argument_information): Compute pointer_arg,
pointer_offset and special_slot for pointer bounds arguments.
(finalize_must_preallocate): Preallocate when storing bounds
in bounds table.
(compute_argument_addresses): Skip pointer bounds.
(expand_call): Store bounds into tables separately. Return
result joined with resulting bounds.
* cfgexpand.c: Include tree-chkp.h, rtl-chkp.h.
(expand_call_stmt): Propagate bounds flag for CALL_EXPR.
(expand_return): Add returned bounds arg. Handle returned bounds.
(expand_gimple_stmt_1): Adjust to new expand_return signature.
(gimple_expand_cfg): Reset rtx bounds map.
* expr.c: Include tree-chkp.h, rtl-chkp.h.
(expand_assignment): Handle returned bounds.
(store_expr_with_bounds): New. Replaces store_expr with new bounds
target argument. Handle bounds returned by calls.
(store_expr): Now wraps store_expr_with_bounds.
* expr.h (store_expr_with_bounds): New.
* function.c: Include tree-chkp.h, rtl-chkp.h.
(bounds_parm_data): New.
(use_register_for_decl): Do not registerize decls used for bounds
stores and loads.
(assign_parms_augmented_arg_list): Add bounds of the result
structure pointer as the second argument.
(assign_parm_find_entry_rtl): Mark bounds are never passed on
the stack.
(assign_parm_is_stack_parm): Likewise.
(assign_parm_load_bounds): New.
(assign_bounds): New.
(assign_parms): Load bounds and determine a location for
returned bounds.
(diddle_return_value_1): New.
(diddle_return_value): Handle returned bounds.
* function.h (rtl_data): Add field for returned bounds.
* varasm.c: Include tree-chkp.h.
(output_constant): Support POINTER_BOUNDS_TYPE.
(output_constant_pool_2): Support MODE_POINTER_BOUNDS.
(ultimate_transparent_alias_target): Move up.
(make_decl_rtl): For instrumented function use
name of the original decl.
(assemble_start_function): Mark function as global
in case it is instrumentation clone of the global
function.
(do_assemble_alias): Follow transparent alias chain
for identifier. Check if original alias is public.
(maybe_assemble_visibility): Use visibility of the
original function for instrumented version.
(default_unique_section): Likewise.
* emit-rtl.c (immed_double_const): Support MODE_POINTER_BOUNDS.
(init_emit_once): Build pointer bounds zero constants.
* explow.c (trunc_int_for_mode): Support MODE_POINTER_BOUNDS.
* target.def (builtin_chkp_function): New.
(chkp_bound_type): New.
(chkp_bound_mode): New.
(chkp_make_bounds_constant): New.
(chkp_initialize_bounds): New.
(load_bounds_for_arg): New.
(store_bounds_for_arg): New.
(load_returned_bounds): New.
(store_returned_bounds): New.
(chkp_function_value_bounds): New.
(setup_incoming_vararg_bounds): New.
(function_arg): Update hook description with new possible return
value CONST_INT.
* targhooks.h (default_load_bounds_for_arg): New.
(default_store_bounds_for_arg): New.
(default_load_returned_bounds): New.
(default_store_returned_bounds): New.
(default_chkp_bound_type): New.
(default_chkp_bound_mode): New.
(default_builtin_chkp_function): New.
(default_chkp_function_value_bounds): New.
(default_chkp_make_bounds_constant): New.
(default_chkp_initialize_bounds): New.
(default_setup_incoming_vararg_bounds): New.
* targhooks.c (default_load_bounds_for_arg): New.
(default_store_bounds_for_arg): New.
(default_load_returned_bounds): New.
(default_store_returned_bounds): New.
(default_chkp_bound_type): New.
(default_chkp_bound_mode); New.
(default_builtin_chkp_function): New.
(default_chkp_function_value_bounds): New.
(default_chkp_make_bounds_constant): New.
(default_chkp_initialize_bounds): New.
(default_setup_incoming_vararg_bounds): New.
* builtin-types.def (BT_BND): New.
(BT_FN_PTR_CONST_PTR): New.
(BT_FN_CONST_PTR_CONST_PTR): New.
(BT_FN_BND_CONST_PTR): New.
(BT_FN_CONST_PTR_BND): New.
(BT_FN_PTR_CONST_PTR_SIZE): New.
(BT_FN_PTR_CONST_PTR_CONST_PTR): New.
(BT_FN_VOID_PTRPTR_CONST_PTR): New.
(BT_FN_VOID_CONST_PTR_SIZE): New.
(BT_FN_VOID_PTR_BND): New.
(BT_FN_CONST_PTR_CONST_PTR_CONST_PTR): New.
(BT_FN_BND_CONST_PTR_SIZE): New.
(BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
(BT_FN_VOID_CONST_PTR_BND_CONST_PTR): New.
* chkp-builtins.def: New.
* builtins.def: include chkp-builtins.def.
(DEF_CHKP_BUILTIN): New.
* builtins.c: Include tree-chkp.h and rtl-chkp.h.
(expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_NARROW,
BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
(std_expand_builtin_va_start): Init bounds for va_list.
* cppbuiltin.c (define_builtin_macros_for_compilation_flags): Add
__CHKP__ macro when Pointer Bounds Checker is on.
* params.def (PARAM_CHKP_MAX_CTOR_SIZE): New.
* passes.def (pass_ipa_chkp_versioning): New.
(pass_early_local_passes): Renamed to pass_build_ssa_passes.
(pass_fixup_cfg): Moved to pass_chkp_instrumentation_passes.
(pass_chkp_instrumentation_passes): New.
(pass_ipa_chkp_produce_thunks): New.
(pass_local_optimization_passes): New.
(pass_chkp_opt): New.
* tree-pass.h (make_pass_ipa_chkp_versioning): New.
(make_pass_ipa_chkp_produce_thunks): New.
(make_pass_chkp): New.
(make_pass_chkp_opt): New.
(make_pass_early_local_passes): Renamed to ...
(make_pass_build_ssa_passes): This.
(make_pass_chkp_instrumentation_passes): New.
(make_pass_local_optimization_passes): New.
* passes.c (pass_manager::execute_early_local_passes): Execute
early passes in three steps.
(execute_all_early_local_passes): Renamed to ...
(execute_build_ssa_passes): This.
(pass_data_early_local_passes): Renamed to ...
(pass_data_build_ssa_passes): This.
(pass_early_local_passes): Renamed to ...
(pass_build_ssa_passes): This.
(pass_data_chkp_instrumentation_passes): New.
(pass_chkp_instrumentation_passes): New.
(pass_data_local_optimization_passes): New.
(pass_local_optimization_passes): New.
(make_pass_early_local_passes): Renamed to ...
(make_pass_build_ssa_passes): This.
(make_pass_chkp_instrumentation_passes): New.
(make_pass_local_optimization_passes): New.
* c-family/c.opt (fcheck-pointer-bounds): New.
(fchkp-check-incomplete-type): New.
(fchkp-zero-input-bounds-for-main): New.
(fchkp-first-field-has-own-bounds): New.
(fchkp-narrow-bounds): New.
(fchkp-narrow-to-innermost-array): New.
(fchkp-optimize): New.
(fchkp-use-fast-string-functions): New.
(fchkp-use-nochk-string-functions): New.
(fchkp-use-static-bounds): New.
(fchkp-use-static-const-bounds): New.
(fchkp-treat-zero-dynamic-size-as-infinite): New.
(fchkp-check-read): New.
(fchkp-check-write): New.
(fchkp-store-bounds): New.
(fchkp-instrument-calls): New.
(fchkp-instrument-marked-only): New.
(Wchkp): New.
* c-family/c-common.c (handle_bnd_variable_size_attribute): New.
(handle_bnd_legacy): New.
(handle_bnd_instrument): New.
(c_common_attribute_table): Add bnd_variable_size, bnd_legacy
and bnd_instrument. Fix documentation.
(c_common_format_attribute_table): Likewsie.
* toplev.c: include tree-chkp.h.
(process_options): Check Pointer Bounds Checker is supported.
(compile_file): Add chkp_finish_file call.
* ipa-cp.c (initialize_node_lattices): Use cgraph_local_p
to handle instrumentation clones properly.
(propagate_constants_accross_call): Do not propagate
through instrumentation thunks.
* ipa-pure-const.c (propagate_pure_const): Support
IPA_REF_CHKP.
* ipa-inline.c (early_inliner): Check edge has summary allocated.
* ipa-split.c: Include tree-chkp.h.
(find_retbnd): New.
(split_part_set_ssa_name_p): New.
(consider_split): Do not split retbnd and retval
producers.
(insert_bndret_call_after): new.
(split_function): Propagate Pointer Bounds Checker
instrumentation marks and handle returned bounds.
* tree-ssa-sccvn.h (vn_reference_op_struct): Transform opcode
into bit field and add with_bounds field.
* tree-ssa-sccvn.c (copy_reference_ops_from_call): Set
with_bounds field for instrumented calls.
* tree-ssa-pre.c (create_component_ref_by_pieces_1): Restore
CALL_WITH_BOUNDS_P flag for calls.
* tree-ssa-ccp.c: Include tree-chkp.h.
(insert_clobber_before_stack_restore): Handle
BUILT_IN_CHKP_BNDRET calls.
* tree-ssa-dce.c: Include tree-chkp.h.
(propagate_necessity): For free call fed by alloc check
bounds are also provided by the same alloc.
(eliminate_unnecessary_stmts): Handle BUILT_IN_CHKP_BNDRET
used by free calls.
* tree-inline.c: Include tree-chkp.h.
(declare_return_variable): Add arg holding
returned bounds slot. Create and initialize returned bounds var.
(remap_gimple_stmt): Handle returned bounds.
Return sequence of statements instead of a single statement.
(insert_init_stmt): Add declaration.
(remap_gimple_seq): Adjust to new remap_gimple_stmt signature.
(copy_bb): Adjust to changed return type of remap_gimple_stmt.
Properly handle bounds in va_arg_pack and va_arg_pack_len.
(expand_call_inline): Handle returned bounds. Add bounds copy
for generated mem to mem assignments.
* tree-inline.h (copy_body_data): Add fields retbnd and
assign_stmts.
* value-prof.c: Include tree-chkp.h.
(gimple_ic): Support returned bounds.
* ipa.c (cgraph_build_static_cdtor_1): Support contructors
with "chkp ctor" and "bnd_legacy" attributes.
(symtab_remove_unreachable_nodes): Keep initial values for
pointer bounds to be used for checks eliminations.
(process_references): Handle IPA_REF_CHKP.
(walk_polymorphic_call_targets): Likewise.
* ipa-visibility.c (cgraph_externally_visible_p): Mark
instrumented 'main' as externally visible.
(function_and_variable_visibility): Filter instrumentation
thunks.
* cgraph.h (cgraph_thunk_info): Add add_pointer_bounds_args
field.
(cgraph_node): Add instrumented_version, orig_decl and
instrumentation_clone fields.
(symtab_node::get_alias_target): Allow IPA_REF_CHKP reference.
(varpool_node): Add need_bounds_init field.
(cgraph_local_p): New.
* cgraph.c: Include tree-chkp.h.
(cgraph_node::remove): Fix instrumented_version
of the referenced node if any.
(cgraph_node::dump): Dump instrumentation_clone and
instrumented_version fields.
(cgraph_node::verify_node): Check correctness of IPA_REF_CHKP
references and instrumentation thunks.
(cgraph_can_remove_if_no_direct_calls_and_refs_p): Keep
all not instrumented instrumentation clones alive.
(cgraph_redirect_edge_call_stmt_to_callee): Support
returned bounds.
* cgraphbuild.c (rebuild_cgraph_edges): Rebuild IPA_REF_CHKP
reference.
(cgraph_rebuild_references): Likewise.
* cgraphunit.c: Include tree-chkp.h.
(assemble_thunks_and_aliases): Skip thunks calling instrumneted
function version.
(varpool_finalize_decl): Register statically initialized decls
in Pointer Bounds Checker.
(walk_polymorphic_call_targets): Do not mark generated call to
__builtin_unreachable as with_bounds.
(output_weakrefs): If there are both instrumented and original
versions, output only one of them.
(cgraph_node::expand_thunk): Set with_bounds flag
for created call statement.
* ipa-ref.h (ipa_ref_use): Add IPA_REF_CHKP.
(ipa_ref): increase size of use field.
* symtab.c (ipa_ref_use_name): Add element for IPA_REF_CHKP.
* varpool.c (dump_varpool_node): Dump need_bounds_init field.
(ctor_for_folding): Do not fold constant bounds vars.
* lto-streamer.h (LTO_minor_version): Change minor version from
0 to 1.
* lto-cgraph.c (compute_ltrans_boundary): Keep initial values for
pointer bounds.
(lto_output_node): Output instrumentation_clone,
thunk.add_pointer_bounds_args and orig_decl field.
(lto_output_ref): Adjust to new ipa_ref::use field size.
(input_overwrite_node): Read instrumentation_clone field.
(input_node): Read thunk.add_pointer_bounds_args and orig_decl
fields.
(input_ref): Adjust to new ipa_ref::use field size.
(input_cgraph_1): Compute instrumented_version fields and restore
IDENTIFIER_TRANSPARENT_ALIAS chains.
(lto_output_varpool_node): Output
need_bounds_init value.
(input_varpool_node): Read need_bounds_init value.
* lto-partition.c (add_symbol_to_partition_1): Keep original
and instrumented versions together.
(privatize_symbol_name): Restore transparent alias chain if required.
(add_references_to_partition): Add references to pointer bounds vars.
* dbxout.c (dbxout_type): Ignore POINTER_BOUNDS_TYPE.
* dwarf2out.c (gen_subprogram_die): Ignore bound args.
(gen_type_die_with_usage): Skip pointer bounds.
(dwarf2out_global_decl): Likewise.
(is_base_type): Support POINTER_BOUNDS_TYPE.
(gen_formal_types_die): Skip pointer bounds.
(gen_decl_die): Likewise.
* var-tracking.c (vt_add_function_parameters): Skip
bounds parameters.
* ipa-icf.c (sem_function::merge): Do not merge when instrumentation
thunk still exists.
(sem_variable::merge): Reset need_bounds_init flag.
* doc/extend.texi: Document Pointer Bounds Checker built-in functions
and attributes.
* doc/tm.texi.in (TARGET_LOAD_BOUNDS_FOR_ARG): New.
(TARGET_STORE_BOUNDS_FOR_ARG): New.
(TARGET_LOAD_RETURNED_BOUNDS): New.
(TARGET_STORE_RETURNED_BOUNDS): New.
(TARGET_CHKP_FUNCTION_VALUE_BOUNDS): New.
(TARGET_SETUP_INCOMING_VARARG_BOUNDS): New.
(TARGET_BUILTIN_CHKP_FUNCTION): New.
(TARGET_CHKP_BOUND_TYPE): New.
(TARGET_CHKP_BOUND_MODE): New.
(TARGET_CHKP_MAKE_BOUNDS_CONSTANT): New.
(TARGET_CHKP_INITIALIZE_BOUNDS): New.
* doc/tm.texi: Regenerated.
* doc/rtl.texi (MODE_POINTER_BOUNDS): New.
(BND32mode): New.
(BND64mode): New.
* doc/invoke.texi (-mmpx): New.
(-mno-mpx): New.
(chkp-max-ctor-size): New.
* config/i386/constraints.md (w): New.
(Ti): New.
(Tb): New.
* config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
* config/i386/i386-modes.def (BND32): New.
(BND64): New.
* config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
* config/i386/i386.c: Include tree-chkp.h, rtl-chkp.h, tree-iterator.h.
(regclass_map): Add bound registers.
(dbx_register_map): Likewise.
(dbx64_register_map): Likewise.
(svr4_dbx_register_map): Likewise.
(isa_opts): Add -mmpx.
(PTA_MPX): New.
(ix86_option_override_internal): Support MPX ISA.
(ix86_conditional_register_usage): Support bound registers.
(ix86_code_end): Add MPX bnd prefix.
(output_set_got): Likewise.
(print_reg): Avoid prefixes for bound registers.
(ix86_print_operand): Add '!' (MPX bnd) print prefix support.
(ix86_print_operand_punct_valid_p): Likewise.
(ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
UNSPEC_BNDLDX_ADDR.
(ix86_output_call_insn): Add MPX bnd prefix to branch instructions.
(ix86_class_likely_spilled_p): Add bound regs support.
(ix86_hard_regno_mode_ok): Likewise.
(x86_order_regs_for_local_alloc): Likewise.
(ix86_bnd_prefixed_insn_p): New.
(ix86_builtins): Add
IX86_BUILTIN_BNDMK, IX86_BUILTIN_BNDSTX,
IX86_BUILTIN_BNDLDX, IX86_BUILTIN_BNDCL,
IX86_BUILTIN_BNDCU, IX86_BUILTIN_BNDRET,
IX86_BUILTIN_BNDNARROW, IX86_BUILTIN_BNDINT,
IX86_BUILTIN_SIZEOF, IX86_BUILTIN_BNDLOWER,
IX86_BUILTIN_BNDUPPER.
(builtin_isa): Add leaf_p and nothrow_p fields.
(def_builtin): Initialize leaf_p and nothrow_p.
(ix86_add_new_builtins): Handle leaf_p and nothrow_p
flags.
(bdesc_mpx): New.
(bdesc_mpx_const): New.
(ix86_init_mpx_builtins): New.
(ix86_init_builtins): Call ix86_init_mpx_builtins.
(ix86_emit_cmove): New.
(ix86_emit_move_max): New.
(ix86_expand_builtin): Expand IX86_BUILTIN_BNDMK,
IX86_BUILTIN_BNDSTX, IX86_BUILTIN_BNDLDX,
IX86_BUILTIN_BNDCL, IX86_BUILTIN_BNDCU,
IX86_BUILTIN_BNDRET, IX86_BUILTIN_BNDNARROW,
IX86_BUILTIN_BNDINT, IX86_BUILTIN_SIZEOF,
IX86_BUILTIN_BNDLOWER, IX86_BUILTIN_BNDUPPER.
(ix86_function_value_bounds): New.
(ix86_builtin_mpx_function): New.
(ix86_get_arg_address_for_bt): New.
(ix86_load_bounds): New.
(ix86_store_bounds): New.
(ix86_load_returned_bounds): New.
(ix86_store_returned_bounds): New.
(ix86_mpx_bound_mode): New.
(ix86_make_bounds_constant): New.
(ix86_initialize_bounds):
(TARGET_LOAD_BOUNDS_FOR_ARG): New.
(TARGET_STORE_BOUNDS_FOR_ARG): New.
(TARGET_LOAD_RETURNED_BOUNDS): New.
(TARGET_STORE_RETURNED_BOUNDS): New.
(TARGET_CHKP_BOUND_MODE): New.
(TARGET_BUILTIN_CHKP_FUNCTION): New.
(TARGET_CHKP_FUNCTION_VALUE_BOUNDS): New.
(TARGET_CHKP_MAKE_BOUNDS_CONSTANT): New.
(TARGET_CHKP_INITIALIZE_BOUNDS): New.
(ix86_option_override_internal): Do not
support x32 with MPX.
(init_cumulative_args): Init stdarg, bnd_regno, bnds_in_bt
and force_bnd_pass.
(function_arg_advance_32): Return number of used integer
registers.
(function_arg_advance_64): Likewise.
(function_arg_advance_ms_64): Likewise.
(ix86_function_arg_advance): Handle pointer bounds.
(ix86_function_arg): Likewise.
(ix86_function_value_regno_p): Mark fisrt bounds registers as
possible function value.
(ix86_function_value_1): Handle pointer bounds type/mode
(ix86_return_in_memory): Likewise.
(ix86_print_operand): Analyse insn to decide abounf "bnd" prefix.
(ix86_expand_call): Generate returned bounds.
(ix86_setup_incoming_vararg_bounds): New.
(ix86_va_start): Initialize bounds for pointers in va_list.
(TARGET_SETUP_INCOMING_VARARG_BOUNDS): New.
* config/i386/i386.h (TARGET_MPX): New.
(TARGET_MPX_P): New.
(FIRST_PSEUDO_REGISTER): Fix to new value.
(FIXED_REGISTERS): Add bound registers.
(CALL_USED_REGISTERS): Likewise.
(REG_ALLOC_ORDER): Likewise.
(HARD_REGNO_NREGS): Likewise.
(VALID_BND_REG_MODE): New.
(FIRST_BND_REG): New.
(LAST_BND_REG): New.
(reg_class): Add BND_REGS.
(REG_CLASS_NAMES): Likewise.
(REG_CLASS_CONTENTS): Likewise.
(BND_REGNO_P): New.
(ANY_BND_REG_P): New.
(BNDmode): New.
(HI_REGISTER_NAMES): Add bound registers.
(ix86_args): Add bnd_regno, bnds_in_bt, force_bnd_pass and
stdarg fields.
* config/i386/i386.md (UNSPEC_BNDMK): New.
(UNSPEC_BNDMK_ADDR): New.
(UNSPEC_BNDSTX): New.
(UNSPEC_BNDLDX): New.
(UNSPEC_BNDLDX_ADDR): New.
(UNSPEC_BNDCL): New.
(UNSPEC_BNDCU): New.
(UNSPEC_BNDCN): New.
(UNSPEC_MPX_FENCE): New.
(UNSPEC_SIZEOF): New.
(BND0_REG): New.
(BND1_REG): New.
(type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
(length_immediate): Support mpxmov, mpxmk, mpxchk, mpxld, mpxst.
(prefix_rep): Check for bnd prefix.
(prefix_0f): Support mpxmov, mpxmk, mpxchk, mpxld, mpxst.
(length_nobnd): New.
(length): Use length_nobnd when specified.
(memory): Support mpxmov, mpxmk, mpxchk, mpxld, mpxst.
(BND): New.
(bnd_ptr): New.
(BNDCHECK): New.
(bndcheck): New.
(*jcc_1): Add MPX bnd prefix.
(*jcc_2): Likewise.
(jump): Likewise.
(*indirect_jump): Likewise.
(*tablejump_1): Likewise.
(simple_return_internal): Likewise.
(simple_return_internal_long): Likewise.
(simple_return_pop_internal): Likewise.
(simple_return_indirect_internal): Likewise.
(<mode>_mk): New.
(*<mode>_mk): New.
(mov<mode>): New.
(*mov<mode>_internal_mpx): New.
(<mode>_<bndcheck>): New.
(*<mode>_<bndcheck>): New.
(<mode>_ldx): New.
(*<mode>_ldx): New.
(<mode>_stx): New.
(*<mode>_stx): New.
move_size_reloc_<mode>): New.
* config/i386/predicates.md (address_mpx_no_base_operand): New.
(address_mpx_no_index_operand): New.
(bnd_mem_operator): New.
(symbol_operand): New.
(x86_64_immediate_size_operand): New.
* config/i386/i386.opt (mmpx): New.
* config/i386/i386-builtin-types.def (BND): New.
(ULONG): New.
(BND_FTYPE_PCVOID_ULONG): New.
(VOID_FTYPE_BND_PCVOID): New.
(VOID_FTYPE_PCVOID_PCVOID_BND): New.
(BND_FTYPE_PCVOID_PCVOID): New.
(BND_FTYPE_PCVOID): New.
(BND_FTYPE_BND_BND): New.
(PVOID_FTYPE_PVOID_PVOID_ULONG): New.
(PVOID_FTYPE_PCVOID_BND_ULONG): New.
(ULONG_FTYPE_VOID): New.
(PVOID_FTYPE_BND): New.
gcc/testsuite/
2014-11-05 Ilya Enkovich <ilya.enkovich@intel.com>
* gcc.target/i386/chkp-builtins-1.c: New.
* gcc.target/i386/chkp-builtins-2.c: New.
* gcc.target/i386/chkp-builtins-3.c: New.
* gcc.target/i386/chkp-builtins-4.c: New.
* gcc.target/i386/chkp-remove-bndint-1.c: New.
* gcc.target/i386/chkp-remove-bndint-2.c: New.
* gcc.target/i386/chkp-const-check-1.c: New.
* gcc.target/i386/chkp-const-check-2.c: New.
* gcc.target/i386/chkp-lifetime-1.c: New.
* gcc.dg/pr37858.c: Replace early_local_cleanups pass name
with build_ssa_passes.
From-SVN: r217125
2014-11-05 13:42:03 +01:00
DEFPARAM ( PARAM_CHKP_MAX_CTOR_SIZE ,
"chkp-max-ctor-size" ,
"Maximum number of statements to be included into a single static "
2015-10-22 00:24:41 +02:00
"constructor generated by Pointer Bounds Checker." ,
ipa-chkp.c: New.
gcc/
2014-11-05 Ilya Enkovich <ilya.enkovich@intel.com>
* ipa-chkp.c: New.
* ipa-chkp.h: New.
* tree-chkp.c: New.
* tree-chkp.h: New.
* tree-chkp-opt.c: New.
* rtl-chkp.c: New.
* rtl-chkp.h: New.
* Makefile.in (OBJS): Add ipa-chkp.o, rtl-chkp.o, tree-chkp.o
tree-chkp-opt.o.
(GTFILES): Add tree-chkp.c.
* mode-classes.def (MODE_POINTER_BOUNDS): New.
* tree.def (POINTER_BOUNDS_TYPE): New.
* genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
(POINTER_BOUNDS_MODE): New.
(make_pointer_bounds_mode): New.
* machmode.h (POINTER_BOUNDS_MODE_P): New.
* stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
(layout_type): Support POINTER_BOUNDS_TYPE.
* tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
* tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
* tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
(type_contains_placeholder_1): Likewise.
(build_common_tree_nodes): Initialize
pointer_bounds_type_node.
* tree.h (POINTER_BOUNDS_TYPE_P): New.
(pointer_bounds_type_node): New.
(POINTER_BOUNDS_P): New.
(BOUNDED_TYPE_P): New.
(BOUNDED_P): New.
(CALL_WITH_BOUNDS_P): New.
* gimple.h (gf_mask): Add GF_CALL_WITH_BOUNDS.
(gimple_call_with_bounds_p): New.
(gimple_call_set_with_bounds): New.
(gimple_return_retbnd): New.
(gimple_return_set_retbnd): New
* gimple.c (gimple_build_return): Increase number of ops
for return statement.
(gimple_build_call_from_tree): Propagate CALL_WITH_BOUNDS_P
flag.
* gimple-pretty-print.c (dump_gimple_return): Print second op.
* rtl.h (CALL_EXPR_WITH_BOUNDS_P): New.
* gimplify.c (gimplify_init_constructor): Avoid infinite
loop during gimplification of bounds initializer.
* calls.c: Include tree-chkp.h, rtl-chkp.h, bitmap.h.
(special_function_p): Use original decl name when analyzing
instrumentation clone.
(arg_data): Add fields special_slot, pointer_arg and
pointer_offset.
(store_bounds): New.
(emit_call_1): Propagate instrumentation flag for CALL.
(initialize_argument_information): Compute pointer_arg,
pointer_offset and special_slot for pointer bounds arguments.
(finalize_must_preallocate): Preallocate when storing bounds
in bounds table.
(compute_argument_addresses): Skip pointer bounds.
(expand_call): Store bounds into tables separately. Return
result joined with resulting bounds.
* cfgexpand.c: Include tree-chkp.h, rtl-chkp.h.
(expand_call_stmt): Propagate bounds flag for CALL_EXPR.
(expand_return): Add returned bounds arg. Handle returned bounds.
(expand_gimple_stmt_1): Adjust to new expand_return signature.
(gimple_expand_cfg): Reset rtx bounds map.
* expr.c: Include tree-chkp.h, rtl-chkp.h.
(expand_assignment): Handle returned bounds.
(store_expr_with_bounds): New. Replaces store_expr with new bounds
target argument. Handle bounds returned by calls.
(store_expr): Now wraps store_expr_with_bounds.
* expr.h (store_expr_with_bounds): New.
* function.c: Include tree-chkp.h, rtl-chkp.h.
(bounds_parm_data): New.
(use_register_for_decl): Do not registerize decls used for bounds
stores and loads.
(assign_parms_augmented_arg_list): Add bounds of the result
structure pointer as the second argument.
(assign_parm_find_entry_rtl): Mark bounds are never passed on
the stack.
(assign_parm_is_stack_parm): Likewise.
(assign_parm_load_bounds): New.
(assign_bounds): New.
(assign_parms): Load bounds and determine a location for
returned bounds.
(diddle_return_value_1): New.
(diddle_return_value): Handle returned bounds.
* function.h (rtl_data): Add field for returned bounds.
* varasm.c: Include tree-chkp.h.
(output_constant): Support POINTER_BOUNDS_TYPE.
(output_constant_pool_2): Support MODE_POINTER_BOUNDS.
(ultimate_transparent_alias_target): Move up.
(make_decl_rtl): For instrumented function use
name of the original decl.
(assemble_start_function): Mark function as global
in case it is instrumentation clone of the global
function.
(do_assemble_alias): Follow transparent alias chain
for identifier. Check if original alias is public.
(maybe_assemble_visibility): Use visibility of the
original function for instrumented version.
(default_unique_section): Likewise.
* emit-rtl.c (immed_double_const): Support MODE_POINTER_BOUNDS.
(init_emit_once): Build pointer bounds zero constants.
* explow.c (trunc_int_for_mode): Support MODE_POINTER_BOUNDS.
* target.def (builtin_chkp_function): New.
(chkp_bound_type): New.
(chkp_bound_mode): New.
(chkp_make_bounds_constant): New.
(chkp_initialize_bounds): New.
(load_bounds_for_arg): New.
(store_bounds_for_arg): New.
(load_returned_bounds): New.
(store_returned_bounds): New.
(chkp_function_value_bounds): New.
(setup_incoming_vararg_bounds): New.
(function_arg): Update hook description with new possible return
value CONST_INT.
* targhooks.h (default_load_bounds_for_arg): New.
(default_store_bounds_for_arg): New.
(default_load_returned_bounds): New.
(default_store_returned_bounds): New.
(default_chkp_bound_type): New.
(default_chkp_bound_mode): New.
(default_builtin_chkp_function): New.
(default_chkp_function_value_bounds): New.
(default_chkp_make_bounds_constant): New.
(default_chkp_initialize_bounds): New.
(default_setup_incoming_vararg_bounds): New.
* targhooks.c (default_load_bounds_for_arg): New.
(default_store_bounds_for_arg): New.
(default_load_returned_bounds): New.
(default_store_returned_bounds): New.
(default_chkp_bound_type): New.
(default_chkp_bound_mode); New.
(default_builtin_chkp_function): New.
(default_chkp_function_value_bounds): New.
(default_chkp_make_bounds_constant): New.
(default_chkp_initialize_bounds): New.
(default_setup_incoming_vararg_bounds): New.
* builtin-types.def (BT_BND): New.
(BT_FN_PTR_CONST_PTR): New.
(BT_FN_CONST_PTR_CONST_PTR): New.
(BT_FN_BND_CONST_PTR): New.
(BT_FN_CONST_PTR_BND): New.
(BT_FN_PTR_CONST_PTR_SIZE): New.
(BT_FN_PTR_CONST_PTR_CONST_PTR): New.
(BT_FN_VOID_PTRPTR_CONST_PTR): New.
(BT_FN_VOID_CONST_PTR_SIZE): New.
(BT_FN_VOID_PTR_BND): New.
(BT_FN_CONST_PTR_CONST_PTR_CONST_PTR): New.
(BT_FN_BND_CONST_PTR_SIZE): New.
(BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
(BT_FN_VOID_CONST_PTR_BND_CONST_PTR): New.
* chkp-builtins.def: New.
* builtins.def: include chkp-builtins.def.
(DEF_CHKP_BUILTIN): New.
* builtins.c: Include tree-chkp.h and rtl-chkp.h.
(expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_NARROW,
BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
(std_expand_builtin_va_start): Init bounds for va_list.
* cppbuiltin.c (define_builtin_macros_for_compilation_flags): Add
__CHKP__ macro when Pointer Bounds Checker is on.
* params.def (PARAM_CHKP_MAX_CTOR_SIZE): New.
* passes.def (pass_ipa_chkp_versioning): New.
(pass_early_local_passes): Renamed to pass_build_ssa_passes.
(pass_fixup_cfg): Moved to pass_chkp_instrumentation_passes.
(pass_chkp_instrumentation_passes): New.
(pass_ipa_chkp_produce_thunks): New.
(pass_local_optimization_passes): New.
(pass_chkp_opt): New.
* tree-pass.h (make_pass_ipa_chkp_versioning): New.
(make_pass_ipa_chkp_produce_thunks): New.
(make_pass_chkp): New.
(make_pass_chkp_opt): New.
(make_pass_early_local_passes): Renamed to ...
(make_pass_build_ssa_passes): This.
(make_pass_chkp_instrumentation_passes): New.
(make_pass_local_optimization_passes): New.
* passes.c (pass_manager::execute_early_local_passes): Execute
early passes in three steps.
(execute_all_early_local_passes): Renamed to ...
(execute_build_ssa_passes): This.
(pass_data_early_local_passes): Renamed to ...
(pass_data_build_ssa_passes): This.
(pass_early_local_passes): Renamed to ...
(pass_build_ssa_passes): This.
(pass_data_chkp_instrumentation_passes): New.
(pass_chkp_instrumentation_passes): New.
(pass_data_local_optimization_passes): New.
(pass_local_optimization_passes): New.
(make_pass_early_local_passes): Renamed to ...
(make_pass_build_ssa_passes): This.
(make_pass_chkp_instrumentation_passes): New.
(make_pass_local_optimization_passes): New.
* c-family/c.opt (fcheck-pointer-bounds): New.
(fchkp-check-incomplete-type): New.
(fchkp-zero-input-bounds-for-main): New.
(fchkp-first-field-has-own-bounds): New.
(fchkp-narrow-bounds): New.
(fchkp-narrow-to-innermost-array): New.
(fchkp-optimize): New.
(fchkp-use-fast-string-functions): New.
(fchkp-use-nochk-string-functions): New.
(fchkp-use-static-bounds): New.
(fchkp-use-static-const-bounds): New.
(fchkp-treat-zero-dynamic-size-as-infinite): New.
(fchkp-check-read): New.
(fchkp-check-write): New.
(fchkp-store-bounds): New.
(fchkp-instrument-calls): New.
(fchkp-instrument-marked-only): New.
(Wchkp): New.
* c-family/c-common.c (handle_bnd_variable_size_attribute): New.
(handle_bnd_legacy): New.
(handle_bnd_instrument): New.
(c_common_attribute_table): Add bnd_variable_size, bnd_legacy
and bnd_instrument. Fix documentation.
(c_common_format_attribute_table): Likewsie.
* toplev.c: include tree-chkp.h.
(process_options): Check Pointer Bounds Checker is supported.
(compile_file): Add chkp_finish_file call.
* ipa-cp.c (initialize_node_lattices): Use cgraph_local_p
to handle instrumentation clones properly.
(propagate_constants_accross_call): Do not propagate
through instrumentation thunks.
* ipa-pure-const.c (propagate_pure_const): Support
IPA_REF_CHKP.
* ipa-inline.c (early_inliner): Check edge has summary allocated.
* ipa-split.c: Include tree-chkp.h.
(find_retbnd): New.
(split_part_set_ssa_name_p): New.
(consider_split): Do not split retbnd and retval
producers.
(insert_bndret_call_after): new.
(split_function): Propagate Pointer Bounds Checker
instrumentation marks and handle returned bounds.
* tree-ssa-sccvn.h (vn_reference_op_struct): Transform opcode
into bit field and add with_bounds field.
* tree-ssa-sccvn.c (copy_reference_ops_from_call): Set
with_bounds field for instrumented calls.
* tree-ssa-pre.c (create_component_ref_by_pieces_1): Restore
CALL_WITH_BOUNDS_P flag for calls.
* tree-ssa-ccp.c: Include tree-chkp.h.
(insert_clobber_before_stack_restore): Handle
BUILT_IN_CHKP_BNDRET calls.
* tree-ssa-dce.c: Include tree-chkp.h.
(propagate_necessity): For free call fed by alloc check
bounds are also provided by the same alloc.
(eliminate_unnecessary_stmts): Handle BUILT_IN_CHKP_BNDRET
used by free calls.
* tree-inline.c: Include tree-chkp.h.
(declare_return_variable): Add arg holding
returned bounds slot. Create and initialize returned bounds var.
(remap_gimple_stmt): Handle returned bounds.
Return sequence of statements instead of a single statement.
(insert_init_stmt): Add declaration.
(remap_gimple_seq): Adjust to new remap_gimple_stmt signature.
(copy_bb): Adjust to changed return type of remap_gimple_stmt.
Properly handle bounds in va_arg_pack and va_arg_pack_len.
(expand_call_inline): Handle returned bounds. Add bounds copy
for generated mem to mem assignments.
* tree-inline.h (copy_body_data): Add fields retbnd and
assign_stmts.
* value-prof.c: Include tree-chkp.h.
(gimple_ic): Support returned bounds.
* ipa.c (cgraph_build_static_cdtor_1): Support contructors
with "chkp ctor" and "bnd_legacy" attributes.
(symtab_remove_unreachable_nodes): Keep initial values for
pointer bounds to be used for checks eliminations.
(process_references): Handle IPA_REF_CHKP.
(walk_polymorphic_call_targets): Likewise.
* ipa-visibility.c (cgraph_externally_visible_p): Mark
instrumented 'main' as externally visible.
(function_and_variable_visibility): Filter instrumentation
thunks.
* cgraph.h (cgraph_thunk_info): Add add_pointer_bounds_args
field.
(cgraph_node): Add instrumented_version, orig_decl and
instrumentation_clone fields.
(symtab_node::get_alias_target): Allow IPA_REF_CHKP reference.
(varpool_node): Add need_bounds_init field.
(cgraph_local_p): New.
* cgraph.c: Include tree-chkp.h.
(cgraph_node::remove): Fix instrumented_version
of the referenced node if any.
(cgraph_node::dump): Dump instrumentation_clone and
instrumented_version fields.
(cgraph_node::verify_node): Check correctness of IPA_REF_CHKP
references and instrumentation thunks.
(cgraph_can_remove_if_no_direct_calls_and_refs_p): Keep
all not instrumented instrumentation clones alive.
(cgraph_redirect_edge_call_stmt_to_callee): Support
returned bounds.
* cgraphbuild.c (rebuild_cgraph_edges): Rebuild IPA_REF_CHKP
reference.
(cgraph_rebuild_references): Likewise.
* cgraphunit.c: Include tree-chkp.h.
(assemble_thunks_and_aliases): Skip thunks calling instrumneted
function version.
(varpool_finalize_decl): Register statically initialized decls
in Pointer Bounds Checker.
(walk_polymorphic_call_targets): Do not mark generated call to
__builtin_unreachable as with_bounds.
(output_weakrefs): If there are both instrumented and original
versions, output only one of them.
(cgraph_node::expand_thunk): Set with_bounds flag
for created call statement.
* ipa-ref.h (ipa_ref_use): Add IPA_REF_CHKP.
(ipa_ref): increase size of use field.
* symtab.c (ipa_ref_use_name): Add element for IPA_REF_CHKP.
* varpool.c (dump_varpool_node): Dump need_bounds_init field.
(ctor_for_folding): Do not fold constant bounds vars.
* lto-streamer.h (LTO_minor_version): Change minor version from
0 to 1.
* lto-cgraph.c (compute_ltrans_boundary): Keep initial values for
pointer bounds.
(lto_output_node): Output instrumentation_clone,
thunk.add_pointer_bounds_args and orig_decl field.
(lto_output_ref): Adjust to new ipa_ref::use field size.
(input_overwrite_node): Read instrumentation_clone field.
(input_node): Read thunk.add_pointer_bounds_args and orig_decl
fields.
(input_ref): Adjust to new ipa_ref::use field size.
(input_cgraph_1): Compute instrumented_version fields and restore
IDENTIFIER_TRANSPARENT_ALIAS chains.
(lto_output_varpool_node): Output
need_bounds_init value.
(input_varpool_node): Read need_bounds_init value.
* lto-partition.c (add_symbol_to_partition_1): Keep original
and instrumented versions together.
(privatize_symbol_name): Restore transparent alias chain if required.
(add_references_to_partition): Add references to pointer bounds vars.
* dbxout.c (dbxout_type): Ignore POINTER_BOUNDS_TYPE.
* dwarf2out.c (gen_subprogram_die): Ignore bound args.
(gen_type_die_with_usage): Skip pointer bounds.
(dwarf2out_global_decl): Likewise.
(is_base_type): Support POINTER_BOUNDS_TYPE.
(gen_formal_types_die): Skip pointer bounds.
(gen_decl_die): Likewise.
* var-tracking.c (vt_add_function_parameters): Skip
bounds parameters.
* ipa-icf.c (sem_function::merge): Do not merge when instrumentation
thunk still exists.
(sem_variable::merge): Reset need_bounds_init flag.
* doc/extend.texi: Document Pointer Bounds Checker built-in functions
and attributes.
* doc/tm.texi.in (TARGET_LOAD_BOUNDS_FOR_ARG): New.
(TARGET_STORE_BOUNDS_FOR_ARG): New.
(TARGET_LOAD_RETURNED_BOUNDS): New.
(TARGET_STORE_RETURNED_BOUNDS): New.
(TARGET_CHKP_FUNCTION_VALUE_BOUNDS): New.
(TARGET_SETUP_INCOMING_VARARG_BOUNDS): New.
(TARGET_BUILTIN_CHKP_FUNCTION): New.
(TARGET_CHKP_BOUND_TYPE): New.
(TARGET_CHKP_BOUND_MODE): New.
(TARGET_CHKP_MAKE_BOUNDS_CONSTANT): New.
(TARGET_CHKP_INITIALIZE_BOUNDS): New.
* doc/tm.texi: Regenerated.
* doc/rtl.texi (MODE_POINTER_BOUNDS): New.
(BND32mode): New.
(BND64mode): New.
* doc/invoke.texi (-mmpx): New.
(-mno-mpx): New.
(chkp-max-ctor-size): New.
* config/i386/constraints.md (w): New.
(Ti): New.
(Tb): New.
* config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
* config/i386/i386-modes.def (BND32): New.
(BND64): New.
* config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
* config/i386/i386.c: Include tree-chkp.h, rtl-chkp.h, tree-iterator.h.
(regclass_map): Add bound registers.
(dbx_register_map): Likewise.
(dbx64_register_map): Likewise.
(svr4_dbx_register_map): Likewise.
(isa_opts): Add -mmpx.
(PTA_MPX): New.
(ix86_option_override_internal): Support MPX ISA.
(ix86_conditional_register_usage): Support bound registers.
(ix86_code_end): Add MPX bnd prefix.
(output_set_got): Likewise.
(print_reg): Avoid prefixes for bound registers.
(ix86_print_operand): Add '!' (MPX bnd) print prefix support.
(ix86_print_operand_punct_valid_p): Likewise.
(ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
UNSPEC_BNDLDX_ADDR.
(ix86_output_call_insn): Add MPX bnd prefix to branch instructions.
(ix86_class_likely_spilled_p): Add bound regs support.
(ix86_hard_regno_mode_ok): Likewise.
(x86_order_regs_for_local_alloc): Likewise.
(ix86_bnd_prefixed_insn_p): New.
(ix86_builtins): Add
IX86_BUILTIN_BNDMK, IX86_BUILTIN_BNDSTX,
IX86_BUILTIN_BNDLDX, IX86_BUILTIN_BNDCL,
IX86_BUILTIN_BNDCU, IX86_BUILTIN_BNDRET,
IX86_BUILTIN_BNDNARROW, IX86_BUILTIN_BNDINT,
IX86_BUILTIN_SIZEOF, IX86_BUILTIN_BNDLOWER,
IX86_BUILTIN_BNDUPPER.
(builtin_isa): Add leaf_p and nothrow_p fields.
(def_builtin): Initialize leaf_p and nothrow_p.
(ix86_add_new_builtins): Handle leaf_p and nothrow_p
flags.
(bdesc_mpx): New.
(bdesc_mpx_const): New.
(ix86_init_mpx_builtins): New.
(ix86_init_builtins): Call ix86_init_mpx_builtins.
(ix86_emit_cmove): New.
(ix86_emit_move_max): New.
(ix86_expand_builtin): Expand IX86_BUILTIN_BNDMK,
IX86_BUILTIN_BNDSTX, IX86_BUILTIN_BNDLDX,
IX86_BUILTIN_BNDCL, IX86_BUILTIN_BNDCU,
IX86_BUILTIN_BNDRET, IX86_BUILTIN_BNDNARROW,
IX86_BUILTIN_BNDINT, IX86_BUILTIN_SIZEOF,
IX86_BUILTIN_BNDLOWER, IX86_BUILTIN_BNDUPPER.
(ix86_function_value_bounds): New.
(ix86_builtin_mpx_function): New.
(ix86_get_arg_address_for_bt): New.
(ix86_load_bounds): New.
(ix86_store_bounds): New.
(ix86_load_returned_bounds): New.
(ix86_store_returned_bounds): New.
(ix86_mpx_bound_mode): New.
(ix86_make_bounds_constant): New.
(ix86_initialize_bounds):
(TARGET_LOAD_BOUNDS_FOR_ARG): New.
(TARGET_STORE_BOUNDS_FOR_ARG): New.
(TARGET_LOAD_RETURNED_BOUNDS): New.
(TARGET_STORE_RETURNED_BOUNDS): New.
(TARGET_CHKP_BOUND_MODE): New.
(TARGET_BUILTIN_CHKP_FUNCTION): New.
(TARGET_CHKP_FUNCTION_VALUE_BOUNDS): New.
(TARGET_CHKP_MAKE_BOUNDS_CONSTANT): New.
(TARGET_CHKP_INITIALIZE_BOUNDS): New.
(ix86_option_override_internal): Do not
support x32 with MPX.
(init_cumulative_args): Init stdarg, bnd_regno, bnds_in_bt
and force_bnd_pass.
(function_arg_advance_32): Return number of used integer
registers.
(function_arg_advance_64): Likewise.
(function_arg_advance_ms_64): Likewise.
(ix86_function_arg_advance): Handle pointer bounds.
(ix86_function_arg): Likewise.
(ix86_function_value_regno_p): Mark fisrt bounds registers as
possible function value.
(ix86_function_value_1): Handle pointer bounds type/mode
(ix86_return_in_memory): Likewise.
(ix86_print_operand): Analyse insn to decide abounf "bnd" prefix.
(ix86_expand_call): Generate returned bounds.
(ix86_setup_incoming_vararg_bounds): New.
(ix86_va_start): Initialize bounds for pointers in va_list.
(TARGET_SETUP_INCOMING_VARARG_BOUNDS): New.
* config/i386/i386.h (TARGET_MPX): New.
(TARGET_MPX_P): New.
(FIRST_PSEUDO_REGISTER): Fix to new value.
(FIXED_REGISTERS): Add bound registers.
(CALL_USED_REGISTERS): Likewise.
(REG_ALLOC_ORDER): Likewise.
(HARD_REGNO_NREGS): Likewise.
(VALID_BND_REG_MODE): New.
(FIRST_BND_REG): New.
(LAST_BND_REG): New.
(reg_class): Add BND_REGS.
(REG_CLASS_NAMES): Likewise.
(REG_CLASS_CONTENTS): Likewise.
(BND_REGNO_P): New.
(ANY_BND_REG_P): New.
(BNDmode): New.
(HI_REGISTER_NAMES): Add bound registers.
(ix86_args): Add bnd_regno, bnds_in_bt, force_bnd_pass and
stdarg fields.
* config/i386/i386.md (UNSPEC_BNDMK): New.
(UNSPEC_BNDMK_ADDR): New.
(UNSPEC_BNDSTX): New.
(UNSPEC_BNDLDX): New.
(UNSPEC_BNDLDX_ADDR): New.
(UNSPEC_BNDCL): New.
(UNSPEC_BNDCU): New.
(UNSPEC_BNDCN): New.
(UNSPEC_MPX_FENCE): New.
(UNSPEC_SIZEOF): New.
(BND0_REG): New.
(BND1_REG): New.
(type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
(length_immediate): Support mpxmov, mpxmk, mpxchk, mpxld, mpxst.
(prefix_rep): Check for bnd prefix.
(prefix_0f): Support mpxmov, mpxmk, mpxchk, mpxld, mpxst.
(length_nobnd): New.
(length): Use length_nobnd when specified.
(memory): Support mpxmov, mpxmk, mpxchk, mpxld, mpxst.
(BND): New.
(bnd_ptr): New.
(BNDCHECK): New.
(bndcheck): New.
(*jcc_1): Add MPX bnd prefix.
(*jcc_2): Likewise.
(jump): Likewise.
(*indirect_jump): Likewise.
(*tablejump_1): Likewise.
(simple_return_internal): Likewise.
(simple_return_internal_long): Likewise.
(simple_return_pop_internal): Likewise.
(simple_return_indirect_internal): Likewise.
(<mode>_mk): New.
(*<mode>_mk): New.
(mov<mode>): New.
(*mov<mode>_internal_mpx): New.
(<mode>_<bndcheck>): New.
(*<mode>_<bndcheck>): New.
(<mode>_ldx): New.
(*<mode>_ldx): New.
(<mode>_stx): New.
(*<mode>_stx): New.
move_size_reloc_<mode>): New.
* config/i386/predicates.md (address_mpx_no_base_operand): New.
(address_mpx_no_index_operand): New.
(bnd_mem_operator): New.
(symbol_operand): New.
(x86_64_immediate_size_operand): New.
* config/i386/i386.opt (mmpx): New.
* config/i386/i386-builtin-types.def (BND): New.
(ULONG): New.
(BND_FTYPE_PCVOID_ULONG): New.
(VOID_FTYPE_BND_PCVOID): New.
(VOID_FTYPE_PCVOID_PCVOID_BND): New.
(BND_FTYPE_PCVOID_PCVOID): New.
(BND_FTYPE_PCVOID): New.
(BND_FTYPE_BND_BND): New.
(PVOID_FTYPE_PVOID_PVOID_ULONG): New.
(PVOID_FTYPE_PCVOID_BND_ULONG): New.
(ULONG_FTYPE_VOID): New.
(PVOID_FTYPE_BND): New.
gcc/testsuite/
2014-11-05 Ilya Enkovich <ilya.enkovich@intel.com>
* gcc.target/i386/chkp-builtins-1.c: New.
* gcc.target/i386/chkp-builtins-2.c: New.
* gcc.target/i386/chkp-builtins-3.c: New.
* gcc.target/i386/chkp-builtins-4.c: New.
* gcc.target/i386/chkp-remove-bndint-1.c: New.
* gcc.target/i386/chkp-remove-bndint-2.c: New.
* gcc.target/i386/chkp-const-check-1.c: New.
* gcc.target/i386/chkp-const-check-2.c: New.
* gcc.target/i386/chkp-lifetime-1.c: New.
* gcc.dg/pr37858.c: Replace early_local_cleanups pass name
with build_ssa_passes.
From-SVN: r217125
2014-11-05 13:42:03 +01:00
5000 , 100 , 0 )
2014-12-06 20:19:37 +01:00
2016-01-27 20:19:47 +01:00
DEFPARAM ( PARAM_FSM_SCALE_PATH_STMTS ,
"fsm-scale-path-stmts" ,
"Scale factor to apply to the number of statements in a threading path when comparing to the number of (scaled) blocks." ,
2 , 1 , 10 )
2016-02-01 23:03:57 +01:00
DEFPARAM ( PARAM_FSM_MAXIMUM_PHI_ARGUMENTS ,
"fsm-maximum-phi-arguments" ,
"Maximum number of arguments a PHI may have before the FSM threader will not try to thread through its block." ,
100 , 1 , 999999 )
2016-01-27 20:19:47 +01:00
DEFPARAM ( PARAM_FSM_SCALE_PATH_BLOCKS ,
"fsm-scale-path-blocks" ,
"Scale factor to apply to the number of blocks in a threading path when comparing to the number of (scaled) statements." ,
3 , 1 , 10 )
2014-12-06 20:19:37 +01:00
DEFPARAM ( PARAM_MAX_FSM_THREAD_PATH_INSNS ,
"max-fsm-thread-path-insns" ,
2015-10-22 00:24:41 +02:00
"Maximum number of instructions to copy when duplicating blocks on a finite state automaton jump thread path." ,
2014-12-06 20:19:37 +01:00
100 , 1 , 999999 )
DEFPARAM ( PARAM_MAX_FSM_THREAD_LENGTH ,
"max-fsm-thread-length" ,
2015-10-22 00:24:41 +02:00
"Maximum number of basic blocks on a finite state automaton jump thread path." ,
2014-12-06 20:19:37 +01:00
10 , 1 , 999999 )
DEFPARAM ( PARAM_MAX_FSM_THREAD_PATHS ,
"max-fsm-thread-paths" ,
2015-10-22 00:24:41 +02:00
"Maximum number of new jump thread paths to create for a finite state automaton." ,
2014-12-06 20:19:37 +01:00
50 , 1 , 999999 )
2015-09-03 13:00:46 +02:00
DEFPARAM ( PARAM_PARLOOPS_CHUNK_SIZE ,
"parloops-chunk-size" ,
2015-10-22 00:24:41 +02:00
"Chunk size of omp schedule for loops parallelized by parloops." ,
2015-09-03 13:00:46 +02:00
0 , 0 , 0 )
Make tree_expr_nonnegative_warnv_p recurse into SSA names
The upcoming patch to move sqrt and cbrt simplifications to match.pd
caused a regression because the (abs @0)->@0 simplification didn't
trigger for:
(abs (convert (abs X)))
The simplification is based on tree_expr_nonnegative_p, which at
the moment just gives up if it sees an SSA_NAME.
This patch makes tree_expr_nonnegative_p recurse into SSA name
definitions, but limits the depth of recursion to a small number
for the reason mentioned in the comment (adapted from an existing
comment in gimple_val_nonnegative_real_p). The patch reuses code
in tree-vrp.c, moving it to gimple-fold.c. It also replaces calls
to gimple_val_nonnegative_real_p with calls to tree_expr_nonnegative_p.
A knock-on effect is that we can now prove _i_589 < 0 is false in
sequences like:
i_1917 = ASSERT_EXPR <i_1075, i_1075 == 0>;
_i_589 = (const int) i_1917;
_i_1507 = ASSERT_EXPR <_i_589, _i_589 < 0>;
This defeats an assert in tree-vrp.c that ASSERT_EXPR conditions
are never known to be false. Previously the assert only ever used
local knowledge and so would be limited to cases like x != x for
integer x. Now that we use global knowledge it's possible to prove
the assertion condition is false in blocks that are in practice
unreachable. The patch therefore removes the assert.
Bootstrapped & regression-tested on x86_64-linux-gnu. I didn't write
a specific test because this is already covered by the testsuite if
the follow-on patch is also applied.
gcc/
* params.def (PARAM_MAX_SSA_NAME_QUERY_DEPTH): New param.
* doc/invoke.texi (--param max-ssa-name-query-depth): Document.
* fold-const.h (tree_unary_nonnegative_warnv_p)
(tree_single_nonnegative_warnv_p, tree_call_nonnegative_warnv_p)
(tree_expr_nonnegative_warnv_p): Add depth parameters.
* fold-const.c: Include gimple-fold.h and params.h.
(tree_ssa_name_nonnegative_warnv_p): New function.
(tree_unary_nonnegative_warnv_p, tree_binary_nonnegative_warnv_p)
(tree_single_nonnegative_warnv_p, tree_call_nonnegative_warnv_p)
(tree_invalid_nonnegative_warnv_p, tree_expr_nonnegative_warnv_p):
Add a depth parameter and increment it for recursive calls to
tree_expr_nonnegative_warnv_p. Use tree_ssa_name_nonnegative_warnv_p
to handle SSA names.
* gimple-fold.h (gimple_val_nonnegative_real_p): Delete.
(gimple_stmt_nonnegative_warnv_p): Declare.
* tree-vrp.c (remove_range_assertions): Remove assert that condition
cannot be proven false.
(gimple_assign_nonnegative_warnv_p, gimple_call_nonnegative_warnv_p)
(gimple_stmt_nonnegative_warnv_p): Move to...
* gimple-fold.c: ...here. Add depth parameters and pass them
down to the tree routines. Accept statements that aren't
assignments or calls but just return false for them.
(gimple_val_nonnegative_real_p): Delete.
* tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
tree_expr_nonnegative_p instead of gimple_val_nonnegative_real_p.
Check HONOR_NANs first.
From-SVN: r228614
2015-10-08 18:49:24 +02:00
2015-10-13 12:08:59 +02:00
DEFPARAMENUM5 ( PARAM_PARLOOPS_SCHEDULE ,
"parloops-schedule" ,
"Schedule type of omp schedule for loops parallelized by "
2015-10-22 00:24:41 +02:00
"parloops (static, dynamic, guided, auto, runtime)." ,
2015-10-13 12:08:59 +02:00
static ,
static , dynamic , guided , auto , runtime )
Make tree_expr_nonnegative_warnv_p recurse into SSA names
The upcoming patch to move sqrt and cbrt simplifications to match.pd
caused a regression because the (abs @0)->@0 simplification didn't
trigger for:
(abs (convert (abs X)))
The simplification is based on tree_expr_nonnegative_p, which at
the moment just gives up if it sees an SSA_NAME.
This patch makes tree_expr_nonnegative_p recurse into SSA name
definitions, but limits the depth of recursion to a small number
for the reason mentioned in the comment (adapted from an existing
comment in gimple_val_nonnegative_real_p). The patch reuses code
in tree-vrp.c, moving it to gimple-fold.c. It also replaces calls
to gimple_val_nonnegative_real_p with calls to tree_expr_nonnegative_p.
A knock-on effect is that we can now prove _i_589 < 0 is false in
sequences like:
i_1917 = ASSERT_EXPR <i_1075, i_1075 == 0>;
_i_589 = (const int) i_1917;
_i_1507 = ASSERT_EXPR <_i_589, _i_589 < 0>;
This defeats an assert in tree-vrp.c that ASSERT_EXPR conditions
are never known to be false. Previously the assert only ever used
local knowledge and so would be limited to cases like x != x for
integer x. Now that we use global knowledge it's possible to prove
the assertion condition is false in blocks that are in practice
unreachable. The patch therefore removes the assert.
Bootstrapped & regression-tested on x86_64-linux-gnu. I didn't write
a specific test because this is already covered by the testsuite if
the follow-on patch is also applied.
gcc/
* params.def (PARAM_MAX_SSA_NAME_QUERY_DEPTH): New param.
* doc/invoke.texi (--param max-ssa-name-query-depth): Document.
* fold-const.h (tree_unary_nonnegative_warnv_p)
(tree_single_nonnegative_warnv_p, tree_call_nonnegative_warnv_p)
(tree_expr_nonnegative_warnv_p): Add depth parameters.
* fold-const.c: Include gimple-fold.h and params.h.
(tree_ssa_name_nonnegative_warnv_p): New function.
(tree_unary_nonnegative_warnv_p, tree_binary_nonnegative_warnv_p)
(tree_single_nonnegative_warnv_p, tree_call_nonnegative_warnv_p)
(tree_invalid_nonnegative_warnv_p, tree_expr_nonnegative_warnv_p):
Add a depth parameter and increment it for recursive calls to
tree_expr_nonnegative_warnv_p. Use tree_ssa_name_nonnegative_warnv_p
to handle SSA names.
* gimple-fold.h (gimple_val_nonnegative_real_p): Delete.
(gimple_stmt_nonnegative_warnv_p): Declare.
* tree-vrp.c (remove_range_assertions): Remove assert that condition
cannot be proven false.
(gimple_assign_nonnegative_warnv_p, gimple_call_nonnegative_warnv_p)
(gimple_stmt_nonnegative_warnv_p): Move to...
* gimple-fold.c: ...here. Add depth parameters and pass them
down to the tree routines. Accept statements that aren't
assignments or calls but just return false for them.
(gimple_val_nonnegative_real_p): Delete.
* tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
tree_expr_nonnegative_p instead of gimple_val_nonnegative_real_p.
Check HONOR_NANs first.
From-SVN: r228614
2015-10-08 18:49:24 +02:00
DEFPARAM ( PARAM_MAX_SSA_NAME_QUERY_DEPTH ,
"max-ssa-name-query-depth" ,
"Maximum recursion depth allowed when querying a property of an"
2015-10-22 00:24:41 +02:00
" SSA name." ,
2017-02-17 14:36:39 +01:00
3 , 1 , 10 )
2016-01-11 13:07:31 +01:00
DEFPARAM ( PARAM_MAX_RTL_IF_CONVERSION_INSNS ,
"max-rtl-if-conversion-insns" ,
"Maximum number of insns in a basic block to consider for RTL "
"if-conversion." ,
10 , 0 , 99 )
2016-01-19 11:35:10 +01:00
2016-07-21 17:37:50 +02:00
DEFPARAM ( PARAM_MAX_RTL_IF_CONVERSION_PREDICTABLE_COST ,
"max-rtl-if-conversion-predictable-cost" ,
"Maximum permissible cost for the sequence that would be "
"generated by the RTL if-conversion pass for a branch that "
"is considered predictable." ,
20 , 0 , 200 )
DEFPARAM ( PARAM_MAX_RTL_IF_CONVERSION_UNPREDICTABLE_COST ,
"max-rtl-if-conversion-unpredictable-cost" ,
"Maximum permissible cost for the sequence that would be "
"generated by the RTL if-conversion pass for a branch that "
"is considered unpredictable." ,
40 , 0 , 200 )
2016-01-19 11:35:10 +01:00
DEFPARAM ( PARAM_HSA_GEN_DEBUG_STORES ,
"hsa-gen-debug-stores" ,
"Level of hsa debug stores verbosity" ,
0 , 0 , 1 )
2016-03-30 09:47:40 +02:00
DEFPARAM ( PARAM_MAX_SPECULATIVE_DEVIRT_MAYDEFS ,
"max-speculative-devirt-maydefs" ,
"Maximum number of may-defs visited when devirtualizing "
"speculatively" , 50 , 0 , 0 )
2016-07-26 17:19:58 +02:00
DEFPARAM ( PARAM_MAX_VRP_SWITCH_ASSERTIONS ,
"max-vrp-switch-assertions" ,
"Maximum number of assertions to add along the default "
"edge of a switch statement during VRP" ,
10 , 0 , 0 )
2016-11-16 17:22:39 +01:00
DEFPARAM ( PARAM_VECT_EPILOGUES_NOMASK ,
"vect-epilogues-nomask" ,
"Enable loop epilogue vectorization using smaller vector size." ,
0 , 0 , 1 )
2001-02-14 17:24:45 +01:00
/ *
2014-01-09 08:31:05 +01:00
2001-02-14 17:24:45 +01:00
Local variables :
mode : c
2007-04-23 18:20:38 +02:00
End :
* /