http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46197
Summary: Bad split_stack_return pattern in i386.md
Product: gcc
Version: 4.6.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
AssignedTo: unassig...@gcc.gnu.org
ReportedBy: hjl.to...@gmail.com
CC: ubiz...@gmail.com
On Linux/x86, revision 166000 gave
FAIL: gcc.target/i386/avx-vzeroupper-10.c scan-assembler-times avx_vzeroupper 3
FAIL: gcc.target/i386/avx-vzeroupper-11.c scan-assembler-times avx_vzeroupper 3
FAIL: gcc.target/i386/avx-vzeroupper-12.c scan-assembler-times avx_vzeroupper 1
FAIL: gcc.target/i386/avx-vzeroupper-14.c scan-assembler-times avx_vzeroupper 1
FAIL: gcc.target/i386/avx-vzeroupper-7.c scan-assembler-times avx_vzeroupper 1
FAIL: gcc.target/i386/avx-vzeroupper-9.c scan-assembler-times avx_vzeroupper 1
The problem is
(define_insn "split_stack_return"
[(unspec_volatile [(match_operand:SI 0 "const_int_operand" "")]
UNSPEC_STACK_CHECK)]
vs.
(define_insn "avx_vzeroupper"
[(unspec_volatile [(match_operand 0 "const_int_operand" "")]
UNSPECV_VZEROUPPER)]
I generate avx_zeroupper. But it matches split_stack_return
since (int) UNSPEC_STACK_CHECK == (int) UNSPECV_VZEROUPPER.
