2017-06-08 Richard Biener <rguenther@suse.de>
PR tree-optimization/80928
* gcc.dg/vect/slp-perm-8.c: Do not expect check loop to be vectorized.
From-SVN: r249004
2017-06-07 Tony Reix <tony.reix@atos.net>
Matthieu Sarter <matthieu.sarter.external@atos.net>
David Edelsohn <dje.gcc@gmail.com>
* simple-object-xcoff.c (simple_object_xcoff_find_sections):
Search symbol table for .go_export symbol and apply pfn if found.
Co-Authored-By: David Edelsohn <dje.gcc@gmail.com>
Co-Authored-By: Matthieu Sarter <matthieu.sarter.external@atos.net>
From-SVN: r248999
gcc/ChangeLog:
2017-06-07 Carl Love <cel@us.ibm.com>
* config/rs6000/rs6000-c: The return type of the following
built-in functions was implemented as int not long long. Fix sign
of return value for the unsigned version of vec_mulo and vec_mule.
vector unsigned long long vec_bperm (vector unsigned long long,
vector unsigned char)
vector signed long long vec_mule (vector signed int,
vector signed int)
vector unsigned long long vec_mule (vector unsigned int,
vector unsigned int)
vector signed long long vec_mulo (vector signed int,
vector signed int)
vector unsigned long long vec_mulo (vector unsigned int,
vector unsigned int)
* doc/extend.texi: Fix the documentation for the built-in
functions.
gcc/testsuite/ChangeLog:
2017-06-07 Carl Love <cel@us.ibm.com>
* gcc.target/powerpc/builtins-3.c: Fix vec_mule, vec_mulo test cases.
---
From-SVN: r248998
gcc/ChangeLog:
2017-06-07 Carl Love <cel@us.ibm.com>
PR target/80982
* config/rs6000/altivec.md (double<mode>2): Fix the implementation of
for BE.
From-SVN: r248997
gcc/ChangeLog:
2017-06-07 Carl Love <cel@us.ibm.com>
* config/rs6000/altivec.md: Fix argument swizzle in vec_doublel
support, Generate doublehv for signed int/float for BE case only.
From-SVN: r248996
PR libstdc++/81002
* include/bits/regex.h (basic_regex): Adjust call to __compile_nfa
so iterator type is deduced.
* include/bits/regex_compiler.h (__compile_nfa): Reorder template
parameters to allow iterator type to be deduced.
* testsuite/28_regex/basic_regex/ctors/basic/iter.cc: New.
From-SVN: r248989
Now rs6000_nonimmediate_operand is just nonimmediate_operand.
* config/rs6000/predicates.md (rs6000_nonimmediate_operand): Delete.
* config/rs6000/rs6000.md (*movsi_internal1, movsi_from_sf,
*mov<mode>_softfloat, and an anonymous splitter): Use
nonimmediate_operand instead of rs6000_nonimmediate_operand.
From-SVN: r248987
We can also remove the two other SPE registers.
* config/rs6000/darwin.h (REGISTER_NAMES): Delete the SPE_ACC and
SPEFSCR registers.
* config/rs6000/rs6000.c (rs6000_reg_names, alt_reg_names): Ditto.
(enum rs6000_reg_type): Delete SPE_ACC_TYPE and SPEFSCR_REG_TYPE.
(rs6000_debug_reg_global): Adjust.
(rs6000_init_hard_regno_mode_ok): Adjust.
(rs6000_dbx_register_number): Adjust.
* config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTER): Change to 115.
(FIXED_REGISTERS, CALL_USED_REGISTERS, CALL_REALLY_USED_REGISTERS):
Remove SPE_ACC and SPEFSCR.
(REG_ALLOC_ORDER): Ditto.
(FRAME_POINTER_REGNUM): Change to 111.
(enum reg_class): Remove the SPE_ACC and SPEFSCR registers.
(REG_CLASS_NAMES): Ditto.
(REG_CLASS_CONTENTS): Delete the SPE_ACC and SPEFSCR registers.
(REGISTER_NAMES): Ditto.
(ADDITIONAL_REG_NAMES): Ditto.
(rs6000_reg_names): Ditto.
* config/rs6000/rs6000.md: Renumber some register number
define_constants.
From-SVN: r248986
rs6000_cbranch_operator now is just comparison_operator, so just use
that directly.
* config/rs6000/predicated.md (rs6000_cbranch_operator): Delete.
* config/rs6000/rs6000.md: Replace rs6000_cbranch_operator by
comparison_operator.
From-SVN: r248977
This deletes -mfloat-gprs and the variables that go with it.
* config/rs6000/rs6000.c: Remove everything related to -mfloat-gprs.
* config/rs6000/rs6000.opt: Ditto.
* config/rs6000/t-rtems: Ditto.
From-SVN: r248976
Since rs6000 no longer supports SPE, TARGET_FPRS now always is true.
This makes TARGET_{SF,DF}_SPE always false. Many patterns in spe.md
can now be deleted; which makes it possible to merge e.g. negdd2 with
*negdd2_fpr.
Finally, e500.h is deleted (it isn't used).
* config/rs6000/darwin.md: Replace TARGET_FPRS by 1 and simplify.
* config/rs6000/dfp.md: Ditto.
(negdd2, *negdd2_fpr): Merge.
(absdd2, *absdd2_fpr): Merge.
(negtd2, *negtd2_fpr): Merge.
(abstd2, *abstd2_fpr): Merge.
* config/rs6000/e500.h: Delete file.
* config/rs6000/predicates.md (rs6000_cbranch_operator): Replace
TARGET_FPRS by 1 and simplify.
* config/rs6000/rs6000-c.c: Ditto.
* config/rs6000/rs6000.c: Ditto. Also replace TARGET_SF_SPE and
TARGET_DF_SPE by 0.
* config/rs6000/rs6000.h: Ditto. Delete TARGET_SF_SPE and
TARGET_DF_SPE.
* config/rs6000/rs6000.md: Ditto.
(floatdidf2, *floatdidf2_fpr): Merge.
(move_from_CR_gt_bit): Delete.
* config/rs6000/spe.md: Replace TARGET_FPRS by 1 and simplify.
(E500_CR_IOR_COMPARE): Delete.
(All patterns that require !TARGET_FPRS): Delete.
* config/rs6000/vsx.md: Replace TARGET_FPRS by 1 and simplify.
From-SVN: r248974
* include/bits/alloc_traits.h (__is_allocator, _RequireAllocator):
New trait and alias for detecting Allocator-like types.
* include/bits/basic_string.h (basic_string): Add deduction guide
from P0433.
* include/ext/alloc_traits.h (__gnu_cxx::__alloc_traits): Add template
parameter with default template argument that causes substitution
failures for types that cannot be allocators.
* testsuite/21_strings/basic_string/cons/char/deduction.cc: New.
* testsuite/21_strings/basic_string/cons/wchar_t/deduction.cc: New.
From-SVN: r248967
* graphds.c (add_edge): Intitialize edge's attached data.
(foll_in_subgraph, dfs_fst_edge, dfs_next_edge): New function
pointer parameter. Call pointed function on each edge during
graph traversing. Skip traversing the edge when the function
returns true.
(graphds_dfs, graphds_scc): Ditto.
(for_each_edge): New parameter. Pass the new parameter to callback
function.
* graphds.h (skip_edge_callback): New function pointer type.
(graphds_dfs, graphds_scc): New function pointer parameter.
(graphds_edge_callback, for_each_edge): New parameter.
From-SVN: r248964
* tree-vect-data-refs.c (vect_mark_for_runtime_alias_test): Factor
out code checking if runtime alias check is possible to below ...
Call the new function.
* tree-data-ref.c (runtime_alias_check_p): ... to new function.
* tree-data-ref.h (runtime_alias_check_p): New decalaration.
From-SVN: r248962
* tree-vect-loop-manip.c (vect_do_peeling): Don't skip vector loop
if versioning is required.
* tree-vect-loop.c (vect_analyze_loop_2): Merge niter check for loop
peeling with the check for versioning.
From-SVN: r248959
* tree-vectorizer.h (vect_build_loop_niters): New parameter.
* tree-vect-loop-manip.c (vect_build_loop_niters): New parameter.
Set true to new parameter if new ssa variable is defined.
(vect_gen_vector_loop_niters): Refactor. Set range information
for the new vector loop bound variable.
(vect_do_peeling): Ditto.
From-SVN: r248958
* tree-affine.c (ssa.h): Include header file.
(tree_to_aff_combination): Handle (T1)(X - CST) when inner type
has wrapping overflow behavior.
From-SVN: r248957
* tree-ssa-loop-ivopts.c (ivopts_estimate_reg_pressure): New
reg_pressure model function.
(ivopts_global_cost_for_size): Delete.
(determine_set_costs, iv_ca_recount_cost): Call new model function
ivopts_estimate_reg_pressure.
From-SVN: r248954
2017-06-07 Tamar Christina <tamar.christina@arm.com>
* config/aarch64/aarch64.c (aarch64_rtx_costs): Make sdiv more expensive than udiv.
Remove floating point cases from mod.
From-SVN: r248953
Do not use the RTEMS-specific task variables for the Ada run-time since
they are not available in RTEMS 4.12 or later. Use thread-local storage
instead.
gcc/ada
* Makefile.in (rtems): Use TLS implementation for s-tpopsp.adb.
* s-tpopsp-rtems.adb: Delete.
From-SVN: r248952
