Kai Tietz [Wed, 25 Feb 2015 14:12:46 +0000 (15:12 +0100)]
backport: re PR tree-optimization/61917 (ICE on valid code at -O3 on x86_64-linux-gnu in vectorizable_reduction, at tree-vect-loop.c:4913)
2015-02-25 Richard Biener <rguenther@suse.de>
Kai Tietz <ktietz@redhat.com>
Backport from mainline
PR tree-optimization/61917
* tree-vect-loop.c (vectorizable_reduction): Allow
vect_internal_def without reduction to exit graceful.
ChangeLog testsuite/
2015-02-25 Kai Tietz <ktietz@redhat.com>
Backport from mainline
PR tree-optimization/61917
* gcc.dg/vect/vect-pr61917.c: New file.
Richard Biener [Mon, 23 Feb 2015 11:14:25 +0000 (11:14 +0000)]
Backport PRs 59354, 61634, 63844, 64909
2015-02-23 Richard Biener <rguenther@suse.de>
Backport from mainline
2014-11-19 Richard Biener <rguenther@suse.de>
PR tree-optimization/63844
* omp-low.c (fixup_child_record_type): Use a restrict qualified
referece type for the receiver parameter.
2014-11-27 Richard Biener <rguenther@suse.de>
PR tree-optimization/61634
* tree-vect-slp.c: Include gimple-walk.h.
(vect_detect_hybrid_slp_stmts): Rewrite to propagate hybrid
down the SLP tree for one scalar statement.
(vect_detect_hybrid_slp_1): New walker function.
(vect_detect_hybrid_slp_2): Likewise.
(vect_detect_hybrid_slp): Properly handle pattern statements
in a pre-scan over all loop stmts.
* gcc.dg/vect/pr61634.c: New testcase.
2015-01-14 Richard Biener <rguenther@suse.de>
PR tree-optimization/59354
* tree-vect-slp.c (vect_build_slp_tree_1): Treat loads from
groups larger than the slp group size as having gaps.
* gcc.dg/vect/pr59354.c: New testcase.
2015-02-10 Richard Biener <rguenther@suse.de>
PR tree-optimization/64909
* tree-vect-loop.c (vect_estimate_min_profitable_iters): Properly
pass a scalar-stmt count estimate to the cost model.
* tree-vect-data-refs.c (vect_peeling_hash_get_lowest_cost): Likewise.
* gcc.dg/vect/costmodel/x86_64/costmodel-pr64909.c: New testcase.
Kyrylo Tkachov [Fri, 20 Feb 2015 14:23:02 +0000 (14:23 +0000)]
[AArch64] Fix wrong-code bug in right-shift SISD patterns
Backport from mainline
2015-02-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
* config/aarch64/aarch64.md (*aarch64_lshr_sisd_or_int_<mode>3):
Mark operand 0 as earlyclobber in 2nd alternative.
(1st define_split below *aarch64_lshr_sisd_or_int_<mode>3):
Write negated shift amount into QI lowpart operand 0 and use it
in the shift step.
(2nd define_split below *aarch64_lshr_sisd_or_int_<mode>3): Likewise.
Georg-Johann Lay [Fri, 20 Feb 2015 10:46:05 +0000 (10:46 +0000)]
backport: re PR target/64452 (ICE in avr-gcc when passing struct member to varargs function)
gcc/
Backport from 2015-02-20 trunk r220847.
PR target/64452
* config/avr/avr.md (pushhi_insn): New insn.
(push<mode>1): Push virtual regs in one chunk using pushhi1_insn.
gcc/testsuite/
Backport from 2015-02-20 trunk r220847.
PR target/64452
* gcc.target/avr/torture/pr64452.c: New test.
Backport from mainline
2014-12-09 Richard Biener <rguenther@suse.de>
PR middle-end/64199
* fold-const.c (fold_binary_loc): Use TREE_OVERFLOW_P.
* gcc.dg/torture/pr64199.c: New testcase.
2015-01-14 Richard Biener <rguenther@suse.de>
PR tree-optimization/64493
PR tree-optimization/64495
* tree-vect-loop.c (vect_finalize_reduction): For double-reductions
assign the proper vectorized PHI to the inner loop exit PHIs.
* gcc.dg/vect/pr64493.c: New testcase.
* gcc.dg/vect/pr64495.c: Likewise.
2015-01-27 Richard Biener <rguenther@suse.de>
PR tree-optimization/56273
PR tree-optimization/59124
PR tree-optimization/64277
* tree-vrp.c (vrp_finalize): Emit array-bound warnings only
from the first VRP pass.
* g++.dg/warn/Warray-bounds-6.C: New testcase.
* gcc.dg/Warray-bounds-12.c: Likewise.
* gcc.dg/Warray-bounds-13.c: Likewise.
2015-02-19 Richard Biener <rguenther@suse.de>
Backport from mainline
2015-01-15 Richard Biener <rguenther@suse.de>
PR middle-end/64365
* tree-data-ref.c (dr_analyze_indices): Make sure that accesses
for MEM_REF access functions with the same base can never partially
overlap.
libgcc/
Backported from mainline
2015-02-17 Sandra Loosemore <sandra@codesourcery.com>
* config/arm/bpabi.S (test_div_by_zero): Make label names
consistent between thumb2 and arm mode cases. Separate the
signed comparison on the high word of the numerator from the
unsigned comparison on the low word.
* config/arm/bpabi-v6m.S (test_div_by_zero): Similarly separate
signed comparison.
gcc/testsuite/
Backported from mainline
2015-02-17 Sandra Loosemore <sandra@codesourcery.com>
pa.c (pa_secondary_reload): Request a secondary reload for all floading point loads and stores except those...
* config/pa/pa.c (pa_secondary_reload): Request a secondary reload
for all floading point loads and stores except those using a register
index address.
* config/pa/pa.md: Add new patterns to load a lo_sum DLT operand
to a register.
Mikael Morin [Fri, 13 Feb 2015 18:48:35 +0000 (18:48 +0000)]
Use the local name instead of the original name in the check for name conflicts...
Use the local name instead of the original name in the check for name conflicts
between a hosting program unit and use-associated symbols
in that program unit.
fortran/
PR fortran/63744
* module.c (check_for_ambiguous): Change argument type
from gfc_symbol to gfc_symtree. Check local (symtree) name
instead of original (symbol) name.
(read_module): Update caller.
constraints.md: Change "Q" and "T" constraints to memory constraints.
* config/pa/constraints.md: Change "Q" and "T" constraints to memory
constraints.
* config/pa/pa.c (pa_cannot_force_const_mem): Don't allow constant
symbolic references to data to be forced to constant memory on the
SOM target.
Paul Thomas [Thu, 12 Feb 2015 21:06:41 +0000 (21:06 +0000)]
re PR fortran/64932 (ICE in gfc_conv_descriptor_data_get for generated finalizer)
2015-02-12 Paul Thomas <pault@gcc.gnu.org>
PR fortran/64932
* trans-stmt.c (gfc_trans_deallocate): If a component array
expression is not a descriptor type and it is a derived type
that has allocatable components and is not finalizable, then
deallocate the allocatable components.
2015-02-12 Paul Thomas <pault@gcc.gnu.org>
PR fortran/64932
* gfortran.dg/finalize_28.f90: New test
* testsuite/libgomp.c/pr64824.c: New test.
* testsuite/libgomp.c/pr64868.c: New test.
* testsuite/libgomp.c++/pr64824.C: New test.
* testsuite/libgomp.c++/pr64868.C: New test.
Dominik Vogt [Mon, 9 Feb 2015 11:06:52 +0000 (11:06 +0000)]
extend.texi: s/390: Update documentation of hotpatch attribute.
2015-02-09 Dominik Vogt <vogt@linux.vnet.ibm.com>
* doc/extend.texi: s/390: Update documentation of hotpatch attribute.
* doc/invoke.texi (-mhotpatch): s/390: Update documentation of
-mhotpatch= option.
* config/s390/s390.opt (mhotpatch): s/390: Remove -mhotpatch and
-mno-hotpatch options. Change syntax of -mhotpatch= option.
* config/s390/s390.c (s390_hotpatch_trampoline_halfwords_default):
Renamed.
(s390_hotpatch_trampoline_halfwords_max): Renamed.
(s390_hotpatch_hw_max): New name.
(s390_hotpatch_trampoline_halfwords): Renamed.
(s390_hotpatch_hw_before_label): New name.
(get_hotpatch_attribute): Removed.
(s390_hotpatch_hw_after_label): New name.
(s390_handle_hotpatch_attribute): Add second parameter to hotpatch
attribute.
(s390_attribute_table): Ditto.
(s390_function_num_hotpatch_trampoline_halfwords): Renamed.
(s390_function_num_hotpatch_hw): New name.
Remove special handling of inline functions and hotpatching.
Return number of nops before and after the function label.
(s390_can_inline_p): Removed.
(s390_asm_output_function_label): Emit a configurable number of nops
after the function label.
(s390_option_override): Update -mhotpatch= syntax and remove -mhotpatch.
(TARGET_CAN_INLINE_P) Removed.
(TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): New.
2015-02-09 Dominik Vogt <vogt@linux.vnet.ibm.com>
* gcc/testsuite/gcc.target/s390/hotpatch-13.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-14.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-15.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-16.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-17.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-18.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-19.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-20.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-10.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-11.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-12.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-13.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-14.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-15.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-16.c: New testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-9.c: New
testcase.
* gcc/testsuite/gcc.target/s390/hotpatch-1.c: Testcase adjusted to
new -mhotpatch.
* gcc/testsuite/gcc.target/s390/hotpatch-10.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-11.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-12.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-2.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-3.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-4.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-5.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-6.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-7.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-8.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-9.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-1.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-2.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-3.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-4.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-5.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-6.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-7.c: Likewise.
* gcc/testsuite/gcc.target/s390/hotpatch-compile-8.c: Likewise.
re PR target/64580 (very high rs6000_stack_info() usage during LTO Firefox build on ppc64)
PR target/64580
Backport from mainline
* config.rs6000/rs6000.c (compute_vrsave_mask): Reverse loop order.
(rs6000_stack_info): Add assert.
(rs6000_output_savres_externs): New function, split off from...
(rs6000_output_function_prologue): ... here. Do not call it for
thunks.
Tim Shen [Tue, 3 Feb 2015 09:01:36 +0000 (09:01 +0000)]
re PR libstdc++/64680 (basic_regex::operator= does not reset flags)
PR libstdc++/64680
Backported from mainline
2015-01-22 Tim Shen <timshen@google.com>
* include/bits/regex.h (basic_regex<>::basic_regex,
basic_regex<>::operator=, basic_regex<>::imbue): Conform to the
standard interface.
* testsuite/28_regex/basic_regex/assign/char/cstring.cc: New testcase.
* config/i386/cpuinfo.c (processor_subtypes): Add
INTEL_COREI7_BROADWELL.
(get_intel_cpu): Support new Silvermont, Haswell and Broadwell
model numbers.
Jakub Jelinek [Sun, 1 Feb 2015 17:37:06 +0000 (18:37 +0100)]
backport: re PR rtl-optimization/61058 (ICE: RTL check: expected elt 3 type 'B', have '0' (rtx barrier) in distance_agu_use_in_bb, at config/i386/i386.c:16740 with __builtin_unreachable())
Backported from mainline
2015-01-27 Jakub Jelinek <jakub@redhat.com>
PR rtl-optimization/61058
* jump.c (cleanup_barriers): Update basic block boundaries
if BLOCK_FOR_INSN is non-NULL on PREV.
Jakub Jelinek [Sun, 1 Feb 2015 17:35:40 +0000 (18:35 +0100)]
backport: re PR c/64778 (ICE on invalid code on x86_64-linux-gnu: tree check: expected class ‘type’, have ‘exceptional’ (error_mark) in create_tmp_from_val, at gimplify.c:491)
Backported from mainline
2015-01-26 Jakub Jelinek <jakub@redhat.com>
PR c/64778
* c-typeck.c (convert_arguments): Return -1 if there are
error_args, even if we've diagnosed too many arguments.
Jakub Jelinek [Sun, 1 Feb 2015 17:33:19 +0000 (18:33 +0100)]
backport: re PR rtl-optimization/63637 (CSE on x86 asm()-s no longer working due to PR/60663 fix)
Backported from mainline
2015-01-23 Jakub Jelinek <jakub@redhat.com>
PR rtl-optimization/63637
PR rtl-optimization/60663
* cse.c (merge_equiv_classes): Set new_elt->cost to MAX_COST
if elt->cost is MAX_COST for ASM_OPERANDS.
(find_sets_in_insn): Fix up comment typo.
(cse_insn): Don't set src_volatile for all non-volatile
ASM_OPERANDS in PARALLELs, but just those with multiple outputs
or with "memory" clobber. Set elt->cost to MAX_COST
for ASM_OPERANDS in PARALLEL. Set src_elt->cost to MAX_COST
if new_src is ASM_OPERANDS and elt->cost is MAX_COST.
* gcc.dg/pr63637-1.c: New test.
* gcc.dg/pr63637-2.c: New test.
* gcc.dg/pr63637-3.c: New test.
* gcc.dg/pr63637-4.c: New test.
* gcc.dg/pr63637-5.c: New test.
* gcc.dg/pr63637-6.c: New test.
* gcc.target/i386/pr63637-1.c: New test.
* gcc.target/i386/pr63637-2.c: New test.
* gcc.target/i386/pr63637-3.c: New test.
* gcc.target/i386/pr63637-4.c: New test.
* gcc.target/i386/pr63637-5.c: New test.
* gcc.target/i386/pr63637-6.c: New test.
Jakub Jelinek [Sun, 1 Feb 2015 17:31:48 +0000 (18:31 +0100)]
backport: re PR debug/64663 (ICE at -O1 and above with -g enabled on x86_64-linux-gnu)
Backported from mainline
2015-01-20 Jakub Jelinek <jakub@redhat.com>
PR debug/64663
* dwarf2out.c (decl_piece_node): Don't put bitsize into
mode if bitsize <= 0.
(decl_piece_bitsize, adjust_piece_list, add_var_loc_to_decl,
dw_sra_loc_expr): Use HOST_WIDE_INT instead of int for bit
sizes and positions.
linux-atomic.c (__kernel_cmpxchg2): Change declaration of oldval and newval to const void *.
* config/pa/linux-atomic.c (__kernel_cmpxchg2): Change declaration of
oldval and newval to const void *. Fix typo.
(FETCH_AND_OP_2): Use __atomic_load_n to load value.
(FETCH_AND_OP_WORD): Likewise.
(OP_AND_FETCH_WORD): Likewise.
(COMPARE_AND_SWAP_2): Likewise.
(__sync_val_compare_and_swap_4): Likewise.
(__sync_lock_test_and_set_4): Likewise.
(SYNC_LOCK_RELEASE_2): Likewise.
Remove support for long long atomic operations.
Ian Lance Taylor [Fri, 30 Jan 2015 00:36:14 +0000 (00:36 +0000)]
compiler: Fix -fgo-prefix handling.
There was bug in the fix for PR 61880: it only worked fully
correctly for code compiled with -fgo-pkgpath. For code that
used -fgo-prefix, or that used neither option, the '.'
separating the prefix and the package name was converted to an
underscore, which did not happen before. This broke SWIG and
any other code that expected specific symbol names.
Fortunately all code compiled in libgo and all code compiled
by the go tool uses -fgo-pkgpath, so this probably did not
affect very many people.
This is an incomplete fix that does not modify the package
file format, for use on both mainline and the GCC 4.9 branch.
A follow on patch will fully fix the problem.
Paul Thomas [Tue, 27 Jan 2015 20:54:49 +0000 (20:54 +0000)]
re PR fortran/62044 (ICE in USE statement with RENAME for extended derived type)
2015-01-27 Paul Thomas <pault@gcc.gnu.org>
PR fortran/62044
* resolve.c (resolve_allocate_expr): If the default initializer
is NULL, keep the original MOLD expression so that the correct
typespec is available.
2015-01-27 Paul Thomas <pault@gcc.gnu.org>
PR fortran/62044
* gfortran.dg/allocate_with_mold_1.f90: New test
projects / thirdparty / gcc.git / log
