| commit | 3ec7c02cc3e922b9364dc8cfd1d4546671b91003 | |
| author | Florian Weimer <fweimer@redhat.com> | |
| Fri, 23 Jun 2017 15:23:44 +0000 (23 17:23 +0200) | ||
| committer | Florian Weimer <fweimer@redhat.com> | |
| Fri, 23 Jun 2017 15:24:40 +0000 (23 17:24 +0200) | ||
| tree | 283d54448fe89359272093156316884e61992c9a | treesnapshot (tar.gz zip) |
| parent | 7fa1d9462baabc5a1058efc13a48444af4678acf | commitdiff |
x86-64: memcmp-avx2-movbe.S needs saturating subtraction [BZ #21662]
This code:
L(between_2_3):
/* Load as big endian with overlapping loads and bswap to avoid
branches. */
movzwl -2(%rdi, %rdx), %eax
movzwl -2(%rsi, %rdx), %ecx
shll $16, %eax
shll $16, %ecx
movzwl (%rdi), %edi
movzwl (%rsi), %esi
orl %edi, %eax
orl %esi, %ecx
bswap %eax
bswap %ecx
subl %ecx, %eax
ret
needs a saturating subtract because the full register is used.
With this commit, only the lower 24 bits of the register are used,
so a regular subtraction suffices.
The test case change adds coverage for these kinds of bugs.
This code:
L(between_2_3):
/* Load as big endian with overlapping loads and bswap to avoid
branches. */
movzwl -2(%rdi, %rdx), %eax
movzwl -2(%rsi, %rdx), %ecx
shll $16, %eax
shll $16, %ecx
movzwl (%rdi), %edi
movzwl (%rsi), %esi
orl %edi, %eax
orl %esi, %ecx
bswap %eax
bswap %ecx
subl %ecx, %eax
ret
needs a saturating subtract because the full register is used.
With this commit, only the lower 24 bits of the register are used,
so a regular subtraction suffices.
The test case change adds coverage for these kinds of bugs.
| ChangeLog | diffblobblamehistory | |
| string/test-memcmp.c | diffblobblamehistory | |
| sysdeps/x86_64/multiarch/memcmp-avx2-movbe.S | diffblobblamehistory |