| commit | 02b78ff749f0c88771713368dbb2a09b1979814f | |
| author | H.J. Lu <hjl.tools@gmail.com> | |
| Mon, 30 Jan 2017 18:59:15 +0000 (30 10:59 -0800) | ||
| committer | H.J. Lu <hjl.tools@gmail.com> | |
| Mon, 30 Jan 2017 18:59:31 +0000 (30 10:59 -0800) | ||
| tree | f65e7fc16ef78b3e4254dcf0dcdb1e98864c5576 | treesnapshot (tar.gz zip) |
| parent | f8bf15febcaf137bbec5a61101e88cd5a9d56ca8 | commitdiff |
Add VZEROUPPER to memset-vec-unaligned-erms.S [BZ #21081]
Since memset-vec-unaligned-erms.S has VDUP_TO_VEC0_AND_SET_RETURN at
function entry, memset optimized for AVX2 and AVX512 will always use
ymm/zmm register. VZEROUPPER should be placed before ret in
L(stosb):
movq %rdx, %rcx
movzbl %sil, %eax
movq %rdi, %rdx
rep stosb
movq %rdx, %rax
ret
since it can be reached from
L(stosb_more_2x_vec):
cmpq $REP_STOSB_THRESHOLD, %rdx
ja L(stosb)
[BZ #21081]
* sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S
(L(stosb)): Add VZEROUPPER before ret.
Since memset-vec-unaligned-erms.S has VDUP_TO_VEC0_AND_SET_RETURN at
function entry, memset optimized for AVX2 and AVX512 will always use
ymm/zmm register. VZEROUPPER should be placed before ret in
L(stosb):
movq %rdx, %rcx
movzbl %sil, %eax
movq %rdi, %rdx
rep stosb
movq %rdx, %rax
ret
since it can be reached from
L(stosb_more_2x_vec):
cmpq $REP_STOSB_THRESHOLD, %rdx
ja L(stosb)
[BZ #21081]
* sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S
(L(stosb)): Add VZEROUPPER before ret.
| ChangeLog | diffblobblamehistory | |
| sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S | diffblobblamehistory |