Skip to content

R
riscv-gcc-1
Commit 431043bd by Uros Bizjak
Options

i386.exp (check_effective_target_3dnow): Fix test code. 

	* gcc.target/i386/i386.exp (check_effective_target_3dnow):
	Fix test code.
	(check_effective_target_xop): Remove __may_alias__ attribute.

From-SVN: r159621
parent feef2551
Showing with 37 additions and 28 deletions
gcc/ChangeLog
...@@ -130,20 +130,20 @@ ...@@ -130,20 +130,20 @@
2010-05-19 Christian Borntraeger <borntraeger@de.ibm.com> 2010-05-19 Christian Borntraeger <borntraeger@de.ibm.com>
* tree-ssa-loop-prefetch.c (mem_ref_group, ar_data): Change step * tree-ssa-loop-prefetch.c (mem_ref_group, ar_data): Change step
to tree. to tree.
(dump_mem_ref): Adopt debug code to handle a tree as step. This (dump_mem_ref): Adopt debug code to handle a tree as step. This
also checks for a constant int vs. non-constant but also checks for a constant int vs. non-constant but
loop-invariant steps. loop-invariant steps.
(find_or_create_group): Change the sort algorithm to only consider (find_or_create_group): Change the sort algorithm to only consider
steps that are constant ints. steps that are constant ints.
(idx_analyze_ref): Adopt code to handle a tree instead of a (idx_analyze_ref): Adopt code to handle a tree instead of a
HOST_WIDE_INT for step. HOST_WIDE_INT for step.
(gather_memory_references_ref): Handle tree instead of int and be (gather_memory_references_ref): Handle tree instead of int and be
prepared to see a NULL_TREE. prepared to see a NULL_TREE.
(prune_ref_by_self_reuse, prune_ref_by_group_reuse): Do not prune (prune_ref_by_self_reuse, prune_ref_by_group_reuse): Do not prune
prefetches if the step cannot be calculated at compile time. prefetches if the step cannot be calculated at compile time.
(issue_prefetch_ref): Issue prefetches for non-constant but (issue_prefetch_ref): Issue prefetches for non-constant but
loop-invariant steps. loop-invariant steps.
2010-05-18 Nathan Froyd <froydnj@codesourcery.com> 2010-05-18 Nathan Froyd <froydnj@codesourcery.com>
...@@ -272,7 +272,7 @@ ...@@ -272,7 +272,7 @@
* configure.ac (HAVE_AS_IX86_REP_LOCK_PREFIX): New test. * configure.ac (HAVE_AS_IX86_REP_LOCK_PREFIX): New test.
* configure: Regenerate. * configure: Regenerate.
* config.in: Regenerate. * config.in: Regenerate.
* config/i386/i386.c (print_operand) <case ;>: Also print ; if * config/i386/i386.c (print_operand) <case ;>: Also print ";" if
!HAVE_AS_IX86_REP_LOCK_PREFIX. !HAVE_AS_IX86_REP_LOCK_PREFIX.
Don't emit whitespace. Don't emit whitespace.
* config/i386/i386.md (*rep_movdi_rex64): Use {%;} after rep. * config/i386/i386.md (*rep_movdi_rex64): Use {%;} after rep.
...@@ -392,9 +392,9 @@ ...@@ -392,9 +392,9 @@
add BARRIER after previous bb if needed. add BARRIER after previous bb if needed.
2010-05-17 Nathan Froyd <froydnj@codesourcery.com> 2010-05-17 Nathan Froyd <froydnj@codesourcery.com>
* tree.c (build_function_type_list_1): Remove bogus assert condition. * tree.c (build_function_type_list_1): Remove bogus assert condition.
2010-05-17 Alan Modra <amodra@gmail.com> 2010-05-17 Alan Modra <amodra@gmail.com>
* config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Delete * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Delete
...@@ -430,9 +430,9 @@ ...@@ -430,9 +430,9 @@
2010-05-17 Christian Borntraeger <borntraeger@de.ibm.com> 2010-05-17 Christian Borntraeger <borntraeger@de.ibm.com>
* config/s390/s390.c: Define sane prefetch settings and activate * config/s390/s390.c: Define sane prefetch settings and activate
flag_prefetch_loop_arrays on -O3. flag_prefetch_loop_arrays on -O3.
* config/s390/s390.h: Declare that read can use write prefetch. * config/s390/s390.h: Declare that read can use write prefetch.
2010-05-17 Jakub Jelinek <jakub@redhat.com> 2010-05-17 Jakub Jelinek <jakub@redhat.com>
......
gcc/testsuite/ChangeLog
2010-05-20 Uros Bizjak <ubizjak@gmail.com>
* gcc.target/i386/i386.exp (check_effective_target_3dnow):
Fix test code.
(check_effective_target_xop): Remove __may_alias__ attribute.
2010-05-20 Hans-Peter Nilsson <hp@axis.com> 2010-05-20 Hans-Peter Nilsson <hp@axis.com>
PR target/44202 PR target/44202
...@@ -62,13 +68,13 @@ ...@@ -62,13 +68,13 @@
* gcc.target/i386/3dnowA-2.c: Likewise. * gcc.target/i386/3dnowA-2.c: Likewise.
* gcc.target/i386/pr42549.c: Likewise. * gcc.target/i386/pr42549.c: Likewise.
* gcc.target/i386/sse3-addsubpd.c: Use dg-require-effective-target sse3. * gcc.target/i386/sse3-lddqu.c: Use dg-require-effective-target sse3.
* gcc.target/i386/sse3-addsubpd.c: Likewise.
* gcc.target/i386/sse3-addsubps.c: Likewise. * gcc.target/i386/sse3-addsubps.c: Likewise.
* gcc.target/i386/sse3-haddpd.c: Likewise. * gcc.target/i386/sse3-haddpd.c: Likewise.
* gcc.target/i386/sse3-haddps.c: Likewise. * gcc.target/i386/sse3-haddps.c: Likewise.
* gcc.target/i386/sse3-hsubpd.c: Likewise. * gcc.target/i386/sse3-hsubpd.c: Likewise.
* gcc.target/i386/sse3-hsubps.c: Likewise. * gcc.target/i386/sse3-hsubps.c: Likewise.
* gcc.target/i386/sse3-lddqu.c: Likewise.
* gcc.target/i386/sse3-movddup.c: Likewise. * gcc.target/i386/sse3-movddup.c: Likewise.
* gcc.target/i386/sse3-movshdup.c: Likewise. * gcc.target/i386/sse3-movshdup.c: Likewise.
* gcc.target/i386/sse3-movsldup.c: Likewise. * gcc.target/i386/sse3-movsldup.c: Likewise.
...@@ -492,13 +498,13 @@ ...@@ -492,13 +498,13 @@
2010-05-17 Christian Borntraeger <borntraeger@de.ibm.com> 2010-05-17 Christian Borntraeger <borntraeger@de.ibm.com>
PR 44078 PR 44078
* gcc.dg/tree-ssa/prefetch-7.c: Change pattern to match only the old * gcc.dg/tree-ssa/prefetch-7.c: Change pattern to match only the old
debug messages but not the newly introduced one. debug messages but not the newly introduced one.
2010-05-16 Daniel Franke <franke.daniel@gmail.com> 2010-05-16 Daniel Franke <franke.daniel@gmail.com>
PR fortran/35779 PR fortran/35779
* gfortran.dg/initialization_25.f90: Commented testcase. * gfortran.dg/initialization_25.f90: Commented testcase.
* gfortran.dg/initialization_26.f90: New. * gfortran.dg/initialization_26.f90: New.
...@@ -624,8 +630,8 @@ ...@@ -624,8 +630,8 @@
2010-05-13 Daniel Franke <franke.daniel@gmail.com> 2010-05-13 Daniel Franke <franke.daniel@gmail.com>
PR fortran/35779 PR fortran/35779
* gfortran.dg/initialization_25.f90: New. * gfortran.dg/initialization_25.f90: New.
2010-05-13 Martin Jambor <mjambor@suse.cz> 2010-05-13 Martin Jambor <mjambor@suse.cz>
...@@ -799,8 +805,8 @@ ...@@ -799,8 +805,8 @@
2010-05-11 Daniel Franke <franke.daniel@gmail.com> 2010-05-11 Daniel Franke <franke.daniel@gmail.com>
PR fortran/43711 PR fortran/43711
* gfortran.dg/gomp/pr43711.f90: New. * gfortran.dg/gomp/pr43711.f90: New.
2010-05-11 Daniel Franke <franke.daniel@gmail.com> 2010-05-11 Daniel Franke <franke.daniel@gmail.com>
...@@ -1500,7 +1506,7 @@ ...@@ -1500,7 +1506,7 @@
* gcc.target/arm/thumb-stackframe.c: New test. * gcc.target/arm/thumb-stackframe.c: New test.
2010-04-27 Shujing Zhao <pearly.zhao@oracle.com> 2010-04-27 Shujing Zhao <pearly.zhao@oracle.com>
* gcc.dg/pr32207.c: Fix typo in expected warning messages. * gcc.dg/pr32207.c: Fix typo in expected warning messages.
* gcc.dg/misc-column.c: Likewise. * gcc.dg/misc-column.c: Likewise.
* gcc.dg/Walways-true-1.c: Likewise. * gcc.dg/Walways-true-1.c: Likewise.
...@@ -1542,7 +1548,7 @@ ...@@ -1542,7 +1548,7 @@
* obj-c++.dg/stubify-1.mm: Ditto. * obj-c++.dg/stubify-1.mm: Ditto.
* lib/objc-torture.exp: Do not require link success for * lib/objc-torture.exp: Do not require link success for
"trivial.m" in the runtime checks when dowhat = 'compile'. "trivial.m" in the runtime checks when dowhat = 'compile'.
* lib/dg-pch.exp: (dg-flags-pch): New Proc. * lib/dg-pch.exp (dg-flags-pch): New Proc.
* objc.dg/stubify-1.m: Restrict to ilp32 targets. Require * objc.dg/stubify-1.m: Restrict to ilp32 targets. Require
Darwin8/OSX10.4 - compatible code generation. Darwin8/OSX10.4 - compatible code generation.
* objc.dg/stubify-2.m: Ditto. * objc.dg/stubify-2.m: Ditto.
......
gcc/testsuite/gcc.target/i386/i386.exp
...@@ -40,9 +40,12 @@ proc check_effective_target_ms_hook_prologue { } { ...@@ -40,9 +40,12 @@ proc check_effective_target_ms_hook_prologue { } {
# Return 1 if 3dnow instructions can be compiled. # Return 1 if 3dnow instructions can be compiled.
proc check_effective_target_3dnow { } { proc check_effective_target_3dnow { } {
return [check_no_compiler_messages 3dnow object { return [check_no_compiler_messages 3dnow object {
void _m_pfadd (void) typedef int __m64 __attribute__ ((__vector_size__ (8)));
typedef float __v2sf __attribute__ ((__vector_size__ (8)));
__m64 _m_pfadd (__m64 __A, __m64 __B)
{ {
__builtin_ia32_pfadd(); return (__m64) __builtin_ia32_pfadd ((__v2sf)__A, (__v2sf)__B);
} }
} "-O2 -m3dnow" ] } "-O2 -m3dnow" ]
} }
...@@ -173,7 +176,7 @@ proc check_effective_target_fma4 { } { ...@@ -173,7 +176,7 @@ proc check_effective_target_fma4 { } {
# Return 1 if xop instructions can be compiled. # Return 1 if xop instructions can be compiled.
proc check_effective_target_xop { } { proc check_effective_target_xop { } {
return [check_no_compiler_messages xop object { return [check_no_compiler_messages xop object {
typedef long long __m128i __attribute__ ((__vector_size__ (16), __may_alias__)); typedef long long __m128i __attribute__ ((__vector_size__ (16)));
typedef short __v8hi __attribute__ ((__vector_size__ (16))); typedef short __v8hi __attribute__ ((__vector_size__ (16)));
__m128i _mm_maccs_epi16(__m128i __A, __m128i __B, __m128i __C) __m128i _mm_maccs_epi16(__m128i __A, __m128i __B, __m128i __C)
{ {
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment