OpenE2K
/
gcc
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

re PR target/81300 (-fpeephole2 breaks __builtin_ia32_sbb_u64, _subborrow_u64 on AMD64) 

2017-07-04  Uros Bizjak  <ubizjak@gmail.com>

	PR target/81300
	* config/i386/i386.md (setcc + movzbl/and to xor + setcc peepholes):
	Require dead FLAGS_REG at the beginning of a peephole.

testsuite/ChangeLog:

	PR target/81300
	* gcc.target/i386/pr81300.c: New test.

From-SVN: r249977
This commit is contained in:
Uros Bizjak 2017-07-04 22:52:32 +02:00
parent 1fcfac88de
commit 45bb7d5907
4 changed files with 98 additions and 53 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

99
gcc/ChangeLog
View File

@ -1,3 +1,9 @@
2017-07-04 Uros Bizjak <ubizjak@gmail.com>
PR target/81300
* config/i386/i386.md (setcc + movzbl/and to xor + setcc peepholes):
Require dead FLAGS_REG at the beginning of a peephole.
2017-07-04 Uros Bizjak <ubizjak@gmail.com>
PR target/81294
@ -258,7 +264,7 @@
2017-07-02 Jan Hubicka <hubicka@ucw.cz>
* tree-cfgcleanup.c (want_merge_blocks_p): New function.
(cleanup_tree_cfg_bb): Use it.
(cleanup_tree_cfg_bb): Use it.
* profile-count.h (profile_count::of_for_merging, profile_count::merge):
New functions.
* tree-cfg.c (gimple_merge_blocks): Use profile_count::merge.
@ -350,8 +356,8 @@
2017-07-02 Jan Hubicka <hubicka@ucw.cz>
* dumpfile.c: Include profile-count.h
* tree-cfg.c (gimple_duplicate_sese_tail): Drop UNUSED attributes; update
profile.
* tree-cfg.c (gimple_duplicate_sese_tail): Drop UNUSED attributes;
update profile.
(insert_cond_bb): Update profile.
* tree-cfg.h (insert_cond_bb): Update prototype.
* tree-chkp-opt.c (chkp_optimize_string_function_calls): Update.
@ -481,7 +487,7 @@
2017-07-01 Andi Kleen <ak@linux.intel.com>
* print-tree.c (print_node): Print all attributes.
* print-tree.c (print_node): Print all attributes.
2017-07-01 Jan Hubicka <hubicka@ucw.cz>
@ -678,7 +684,7 @@
* config/i386/i386.c (ix86_macro_fusion_pair_p): Push the check for
CC usage from generic code to here.
* sched-deps.c (sched_macro_fuse_insns): Move the condition for
CC usage into the target macros.
CC usage into the target macros.
2017-06-29 Maya Rashish <coypu@sdf.org>
@ -693,16 +699,16 @@
2017-06-29 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
* config/rs6000/rs6000.c (toc_relative_expr_p): Make tocrel_base
and tocrel_offset be pointer args rather than implicitly using
static versions.
(legitimate_constant_pool_address_p, rs6000_emit_move,
const_load_sequence_p, adjust_vperm): Add local tocrel_base and
tocrel_offset and use in toc_relative_expr_p call.
(print_operand, print_operand_address): Use static tocrel_base_oac
and tocrel_offset_oac.
(rs6000_output_addr_const_extra): Use static tocrel_base_oac and
tocrel_offset_oac.
* config/rs6000/rs6000.c (toc_relative_expr_p): Make tocrel_base
and tocrel_offset be pointer args rather than implicitly using
static versions.
(legitimate_constant_pool_address_p, rs6000_emit_move,
const_load_sequence_p, adjust_vperm): Add local tocrel_base and
tocrel_offset and use in toc_relative_expr_p call.
(print_operand, print_operand_address): Use static tocrel_base_oac
and tocrel_offset_oac.
(rs6000_output_addr_const_extra): Use static tocrel_base_oac and
tocrel_offset_oac.
2017-06-29 Maya Rashish <coypu@sdf.org>
@ -1038,12 +1044,12 @@
2017-06-29 Yury Gribov <tetra2005@gmail.com>
PR bootstrap/80565
* ipa-cp.c (allocate_and_init_ipcp_value): Add initialization
* ipa-cp.c (allocate_and_init_ipcp_value): Add initialization
code.
* ipa-inline.h
* ipa-inline.h
(edge_growth_cache_entry::edge_growth_cache_entry): New
function.
(reset_edge_growth_cache): Update to use constructor.
(reset_edge_growth_cache): Update to use constructor.
2017-06-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
@ -1238,7 +1244,7 @@
(ASM_OUTPUT_ALIGNED_BSS): Add #undef before #define.
2017-06-27 Jerome Lambourg <lambourg@adacore.com>
Olivier Hainque <hainque@adacore.com>
Olivier Hainque <hainque@adacore.com>
* config/i386/vxworks.h (DBX_REGISTER_NUMBER): Pick distinct
map for 64bits.
@ -1283,7 +1289,7 @@
* tree.c (builtin_structptr_types): Add four entries.
2017-06-27 Jerome Lambourg <lambourg@adacore.com>
Olivier Hainque <hainque@adacore.com>
Olivier Hainque <hainque@adacore.com>
* config/vxworks.h (VXWORKS_LIB_SPEC): Incorporate ...
(TLS_SYM): New local macro, forcing reference to __tls__ on
@ -1329,7 +1335,7 @@
condition.
2017-06-26 Jerome Lambourg <lambourg@adacore.com>
Olivier Hainque <hainque@adacore.com>
Olivier Hainque <hainque@adacore.com>
* config.gcc (*-*-vxworks*): Add TARGET_VXWORKS7=1 to tm_defines
for all vxworks7 targets.
@ -1694,7 +1700,7 @@
attribute type list for neon_multiply.
* config/arm/types.md (crypto_pmull): Add.
* config/arm/xgene1.md (xgene1_neon_pmull): Add crypto_pmull to
attribute type list.
attribute type list.
2017-06-20 Andreas Tobler <andreast@gcc.gnu.org>
@ -1950,9 +1956,9 @@
2017-06-16 Richard Earnshaw <rearnsha@arm.com>
* config/arm/t-linux-eabi (MULTILIB_EXCEPTIONS): Set to empty.
(MULTILIB_RESUE): Likewise.
(MULTILIB_MATCHES): Likewise.
(MULTLIB_REQUIRED): Likewise.
(MULTILIB_RESUE): Likewise.
(MULTILIB_MATCHES): Likewise.
(MULTLIB_REQUIRED): Likewise.
2017-06-16 Richard Earnshaw <rearnsha@arm.com>
@ -2325,28 +2331,28 @@
2017-06-16 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
* config/aarch64/aarch64.c (qdf24xx_prefetch_tune): Update
prefetch settings, and enable prefetching by default at -O3.
* config/aarch64/aarch64.c (qdf24xx_prefetch_tune): Update
prefetch settings, and enable prefetching by default at -O3.
2017-06-16 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
* config/aarch64/aarch64.c (aarch64_override_options_internal):
Set flag_prefetch_loop_arrays according to tuning data.
* config/aarch64/aarch64.c (aarch64_override_options_internal):
Set flag_prefetch_loop_arrays according to tuning data.
2017-06-16 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
* config/aarch64/aarch64-protos.h (struct cpu_prefetch_tune):
New tune structure.
(struct tune_params): Use cpu_prefetch_tune instead of cache_line_size.
[Unrelated to main purpose of the patch] Place the pointer field last
to enable type checking errors when tune structure are wrongly merged.
* config/aarch64/aarch64.c (generic_prefetch_tune,)
(exynosm1_prefetch_tune, qdf24xx_prefetch_tune,)
(thunderx2t99_prefetch_tune): New tune constants.
(tune_params *_tunings): Update all tunings (no functional change).
(aarch64_override_options_internal): Set PARAM_SIMULTANEOUS_PREFETCHES,
PARAM_L1_CACHE_SIZE, PARAM_L1_CACHE_LINE_SIZE, and PARAM_L2_CACHE_SIZE
from tunings structures.
* config/aarch64/aarch64-protos.h (struct cpu_prefetch_tune):
New tune structure.
(struct tune_params): Use cpu_prefetch_tune instead of cache_line_size.
[Unrelated to main purpose of the patch] Place the pointer field last
to enable type checking errors when tune structure are wrongly merged.
* config/aarch64/aarch64.c (generic_prefetch_tune,)
(exynosm1_prefetch_tune, qdf24xx_prefetch_tune,)
(thunderx2t99_prefetch_tune): New tune constants.
(tune_params *_tunings): Update all tunings (no functional change).
(aarch64_override_options_internal): Set PARAM_SIMULTANEOUS_PREFETCHES,
PARAM_L1_CACHE_SIZE, PARAM_L1_CACHE_LINE_SIZE, and PARAM_L2_CACHE_SIZE
from tunings structures.
2017-06-16 Jakub Jelinek <jakub@redhat.com>
@ -2478,7 +2484,7 @@
PR target/71663
* config/aarch64/aarch64.c (aarch64_expand_vector_init):
Improve vector initialization code gen for only variable case.
Improve vector initialization code gen for only variable case.
2017-06-14 Eric Botcazou <ebotcazou@adacore.com>
@ -3322,7 +3328,7 @@
PR sanitizer/80932
* fold-const.c (extract_muldiv_1) <case MINUS_EXPR>: Add
TYPE_OVERFLOW_WRAPS check.
TYPE_OVERFLOW_WRAPS check.
2017-06-07 Bin Cheng <bin.cheng@arm.com>
@ -3365,12 +3371,13 @@
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.
* config/aarch64/aarch64.c (aarch64_rtx_costs): Make sdiv more
expensive than udiv. Remove floating point cases from mod.
2017-06-07 Tamar Christina <tamar.christina@arm.com>
* config/arm/aarch-cost-tables.h (cortexa53_extra_cost): Increase idiv cost.
* config/arm/aarch-cost-tables.h (cortexa53_extra_cost):
Increase idiv cost.
2017-06-07 Tamar Christina <tamar.christina@arm.com>

12
gcc/config/i386/i386.md
View File

@ -11754,7 +11754,8 @@
(zero_extend (match_dup 1)))]
"(peep2_reg_dead_p (3, operands[1])
|| operands_match_p (operands[1], operands[3]))
&& ! reg_overlap_mentioned_p (operands[3], operands[0])"
&& ! reg_overlap_mentioned_p (operands[3], operands[0])
&& peep2_regno_dead_p (0, FLAGS_REG)"
[(set (match_dup 4) (match_dup 0))
(set (strict_low_part (match_dup 5))
(match_dup 2))]
@ -11775,7 +11776,8 @@
"(peep2_reg_dead_p (3, operands[1])
|| operands_match_p (operands[1], operands[3]))
&& ! reg_overlap_mentioned_p (operands[3], operands[0])
&& ! reg_set_p (operands[3], operands[4])"
&& ! reg_set_p (operands[3], operands[4])
&& peep2_regno_dead_p (0, FLAGS_REG)"
[(parallel [(set (match_dup 5) (match_dup 0))
(match_dup 4)])
(set (strict_low_part (match_dup 6))
@ -11797,7 +11799,8 @@
(and:SI (match_dup 3) (const_int 255)))
(clobber (reg:CC FLAGS_REG))])]
"REGNO (operands[1]) == REGNO (operands[3])
&& ! reg_overlap_mentioned_p (operands[3], operands[0])"
&& ! reg_overlap_mentioned_p (operands[3], operands[0])
&& peep2_regno_dead_p (0, FLAGS_REG)"
[(set (match_dup 4) (match_dup 0))
(set (strict_low_part (match_dup 5))
(match_dup 2))]
@ -11819,7 +11822,8 @@
"(peep2_reg_dead_p (3, operands[1])
|| operands_match_p (operands[1], operands[3]))
&& ! reg_overlap_mentioned_p (operands[3], operands[0])
&& ! reg_set_p (operands[3], operands[4])"
&& ! reg_set_p (operands[3], operands[4])
&& peep2_regno_dead_p (0, FLAGS_REG)"
[(parallel [(set (match_dup 5) (match_dup 0))
(match_dup 4)])
(set (strict_low_part (match_dup 6))

10
gcc/testsuite/ChangeLog
View File

@ -1,3 +1,8 @@
2017-07-04 Uros Bizjak <ubizjak@gmail.com>
PR target/81300
* gcc.target/i386/pr81300.c: New test.
2017-07-04 Uros Bizjak <ubizjak@gmail.com>
PR target/81294
@ -505,7 +510,7 @@
2017-06-24 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/52473
PR fortran/52473
* gfortran.dg/cshift_2.f90: New test.
2017-06-23 Jim Wilson <jim.wilson@linaro.org>
@ -536,8 +541,7 @@
2017-06-23 James Cowgill <James.Cowgill@imgtec.com>
* go.test/go-test.exp (go-set-goarch): Update MIPS architecture
names.
* go.test/go-test.exp (go-set-goarch): Update MIPS architecture names.
2017-06-22 Matt Turner <mattst88@gmail.com>

30
gcc/testsuite/gcc.target/i386/pr81300.c Normal file
View File

@ -0,0 +1,30 @@
/* PR target/81300 */
/* { dg-do run { target { ! ia32 } } } */
/* { dg-options "-O2" } */
int
__attribute__((noinline, noclone))
foo (void)
{
unsigned long long _discard = 0, zero = 0, maxull = 0;
unsigned char zero1 = __builtin_ia32_addcarryx_u64 (0, 0, 0, &_discard);
unsigned char zero2 = __builtin_ia32_addcarryx_u64 (zero1, 0, 0, &zero);
__builtin_ia32_sbb_u64 (0x0, 2, -1, &_discard);
unsigned char one = __builtin_ia32_sbb_u64 (0, zero, 1, &maxull);
unsigned long long x = __builtin_ia32_sbb_u64 (one, zero2, 0, &_discard);
unsigned long long z1 = 0;
__asm__ ("mov{q}\t{%1, %0|%0, %1}" : "+r" (z1) : "r" (x));
unsigned long long z2 = 3;
__asm__ ("mov{q}\t{%1, %0|%0, %1}" : "+r" (z2) : "r" (x));
return 1 - (z1 | z2);
}
int main ()
{
if (foo ())
__builtin_abort ();
return 0;
}