When zero extending a 32-bit register, we emit a "mov", but currently
report the cost of the "mov" incorrectly.
In terms of speed, we currently say the cost is that of an extend
operation. But the cost of a "mov" is the cost of 1 instruction, so fix
that.
In terms of size, we currently say that the "mov" takes 0 instructions.
Fix it by changing it to 1.
Bootstrapped and tested on aarch64-none-elf.
gcc/
* config/aarch64/aarch64.c (aarch64_rtx_costs): Fix cost of zero extend.
From-SVN: r238820
This patch improves the readability of the prolog and epilog code by moving
some code into separate functions. There is no difference in generated code.
gcc/
* config/aarch64/aarch64.c (aarch64_pushwb_pair_reg): Rename.
(aarch64_push_reg): New function to push 1 or 2 registers.
(aarch64_pop_reg): New function to pop 1 or 2 registers.
(aarch64_expand_prologue): Use aarch64_push_regs.
(aarch64_expand_epilogue): Use aarch64_pop_regs.
From-SVN: r238818
gcc/
2016-07-28 Yuri Rumyantsev <ysrumyan@gmail.com>
PR tree-optimization/71734
* tree-ssa-loop-im.c (ref_indep_loop_p_1): Pass value of safelen
attribute instead of REF_LOOP and use it.
(ref_indep_loop_p_2): Use SAFELEN argument instead of REF_LOOP and
set it for Loops having non-zero safelen attribute.
(ref_indep_loop_p): Pass zero as initial value for safelen.
gcc/testsuite/
2016-07-28 Yuri Rumyantsev <ysrumyan@gmail.com>
PR tree-optimization/71734
* g++.dg/vect/pr70729-nest.cc: New test.
From-SVN: r238817
This patch reverts the change for PR 71902 since it causes 178.gagel
miscompile in spec2000 as reported in PR 71961 which was observed in
x86_64, aarch64, powerpc64.
gcc/fortran/ChangeLog:
2016-07-28 Renlin Li <renlin.li@arm.com>
Revert
2016-07-19 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/71902
* dependency.c (gfc_check_dependency): Use dep_ref. Handle case
if identical is true and two array element references differ.
(gfc_dep_resovler): Move most of the code to dep_ref.
(dep_ref): New function.
* frontend-passes.c (realloc_string_callback): Name temporary
variable "realloc_string".
gcc/testsuite/ChangeLog:
2016-07-28 Renlin Li <renlin.li@arm.com>
Revert
2016-07-19 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/71902
* gfortran.dg/dependency_47.f90: New test.
From-SVN: r238815
* g++.dg/vect/pr70944.cc: New test.
PR rtl-optimization/70944
* combine.c (make_compound_operation):
Do not allow make_compound_operation for vector mode
From-SVN: r238808
2016-07-27 Michael Meissner <meissner@linux.vnet.ibm.com>
* config/rs6000/vector.md (vec_extract<mode>): Change the calling
signature of rs6000_expand_vector_extract so that the element
number is a RTX instead of a constant integer.
* config/rs6000/rs6000-protos.h (rs6000_expand_vector_extract):
Likewise.
* config/rs6000/rs6000.c (rs6000_expand_vector_extract): Likewise.
(altivec_expand_vec_ext_builtin): Likewise.
* config/rs6000/altivec.md (reduc_plus_scal_<mode>): Likewise.
* config/rs6000/vsx.md (vsx_extract_<mode>): Fix spelling of the
MFVSRLD instruction.
From-SVN: r238799
For some reason I added make_location and some related functions to
tree.h/c, rather than to input.h/c. Move them there, so we can use them
without requiring tree, and add some selftest coverage.
gcc/ChangeLog:
* input.c (get_pure_location): Move here from tree.c.
(make_location): Likewise. Add header comment.
(selftest::test_accessing_ordinary_linemaps): Verify
pure_location_p, make_location, get_location_from_adhoc_loc and
get_range_from_loc.
* input.h (get_pure_location): Move declaration here from tree.h.
(get_finish): Likewise for inline function.
(make_location): Likewise for declaration.
* tree.c (get_pure_location): Move to input.c.
(make_location): Likewise.
* tree.h (get_pure_location): Move declaration to tree.h.
(get_finish): Likewise for inline function.
(make_location): Likewise for declaration.
libcpp/ChangeLog:
* include/line-map.h (source_location): Fix line numbers in
comment.
From-SVN: r238792
C++11 has a
static_assert (COND, MESSAGE)
which gives more readable error messages for STATIC_ASSERT than our
current implementation.
This patch makes us use it if __cplusplus >= 201103L
There's also a provisional static_assert (COND) in C++1z, but presumably
we should wait until that one is fully standardized before using it.
gcc/ChangeLog:
* system.h (STATIC_ASSERT): Use static_assert if building
with C++11 onwards.
From-SVN: r238786
gcc/testsuite/
* gcc.dg/torture/pr69352.c: Use __INTPTR_TYPE__ instead of
including stdint.h.
* gcc.dg/torture/pr71866.c: Use __UINTPTR_TYPE__ isntead of
including stdint.h.
From-SVN: r238780
[gcc]
2016-07-26 Michael Meissner <meissner@linux.vnet.ibm.com>
PR target/71869
* config/rs6000/rs6000.c (rs6000_generate_compare): Rework
__float128 support when we don't have hardware support, so that
the IEEE built-in functions like isgreater, first call __unordkf3
to make sure neither operand is a NaN, and if both operands are
ordered, do the normal comparison.
[gcc/testsuite]
2016-07-26 Michael Meissner <meissner@linux.vnet.ibm.com>
PR target/71869
* gcc.target/powerpc/float128-cmp.c: New test to make sure that
IEEE built-in functions handle quiet and signalling NaNs
correctly.
From-SVN: r238779
2016-07-22 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/71862
* class.c: Remove assert. Iterate over component only if non-null.
2016-07-22 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/71862
* gfortran.dg/pr71862.f90: New test.
From-SVN: r238774
gcc/ChangeLog:
* tree-vrp.c (dump_asserts_for): Print loc->expr instead of
name.
(extract_code_and_val_from_cond_with_ops): Verify that name is
either cond_op0 or cond_op1.
From-SVN: r238762
gcc/ChangeLog:
PR tree-optimization/18046
* genmodes.c (emit_mode_size_inline): Emit an assert that
verifies that mode is a valid array index.
(emit_mode_nuinits_inline): Likewise.
(emit_mode_inner_inline): Likewise.
(emit_mode_unit_size_inline): Likewise.
(emit_mode_unit_precision_inline): Likewise.
* tree-vrp.c: Include params.h.
(find_switch_asserts): Register edge assertions for the default
label which correspond to the anti-ranges of each case label.
* params.def (PARAM_MAX_VRP_SWITCH_ASSERTIONS): New.
* doc/invoke.texi: Document it.
gcc/testsuite/ChangeLog:
PR tree-optimization/18046
* gcc.dg/tree-ssa/ssa-dom-thread-6.c: Bump FSM count to 5.
* gcc.dg/tree-ssa/vrp103.c: New test.
* gcc.dg/tree-ssa/vrp104.c: New test.
From-SVN: r238761
2016-07-26 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
* gcc.target/powerpc/pr63354.c: Require lp64 since
-mprofile-kernel is not legal with -m32.
From-SVN: r238759
2016-07-26 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
* gimple-ssa-strength-reduction.c (slsr_process_phi): Remove dead
and unnecessary call to gimple_bb.
From-SVN: r238758
2016-07-26 Richard Biener <rguenther@suse.de>
PR rtl-optimization/71984
* simplify-rtx.c (simplify_subreg): Use GET_MODE_SIZE and prepare
for VOIDmode.
* gcc.dg/torture/pr71984.c: New testcase.
From-SVN: r238757
2016-07-26 Richard Biener <rguenther@suse.de>
PR middle-end/72517
* expmed.c (extract_bit_field_1): Constrain the vector mode
with element size matching the extraction mode size when
choosing a better vector mode to do the extraction from.
From-SVN: r238756
gcc/ChangeLog:
2016-07-26 Trevor Saunders <tbsaunde+gcc@tbsaunde.org>
* tree-outof-ssa.c (struct elim_graph): Change type of members
to auto_vec and auto_sbitmap.
(elim_graph::elim_graph): New constructor.
(delete_elim_graph): Remove.
(expand_phi_nodes): Adjust.
From-SVN: r238750