language-servers/roc
1
0
Fork 0
mirror of https://github.com/roc-lang/roc.git synced 2025-08-01 10:52:18 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 6

change to more specific .balign assembly attribute

Browse source
This commit is contained in:
Brendan Hansknecht 2023-06-02 12:51:25 -07:00
parent 60e70c6b1b
commit 7b8bd988d4
No known key found for this signature in database
GPG key ID: 0EA784685083E75B
1 changed files with 10 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

20
crates/compiler/builtins/bitcode/src/libc/folly/memcpy-x86_64.S
View file

@ -89,7 +89,7 @@ __folly_memcpy_short_{[prefetch]s}:
movw %r9w, -2(%rdi,%rdx)
ret
.align 2
.balign 2
.L_EQ1_{[prefetch]s}:
movb (%rsi), %r8b
movb %r8b, (%rdi)
@ -97,7 +97,7 @@ __folly_memcpy_short_{[prefetch]s}:
// Aligning the target of a jump to an even address has a measurable
// speedup in microbenchmarks.
.align 2
.balign 2
.L_GE4_LE7_{[prefetch]s}:
movl (%rsi), %r8d
movl -4(%rsi,%rdx), %r9d
@ -113,7 +113,7 @@ __folly_memcpy_short_{[prefetch]s}:
// stacktraces will show __folly_memcpy since that is the name of the function.
// This is intended to aid in debugging by making it obvious which version of
// memcpy is being used.
.align 64
.balign 64
.globl __folly_memcpy_{[prefetch]s}
# .type __folly_memcpy_{[prefetch]s}, @function not supported by windows
@ -147,7 +147,7 @@ __folly_memcpy_{[prefetch]s}:
.L_EQ0_{[prefetch]s}:
ret
.align 2
.balign 2
.L_GE17_LE32_{[prefetch]s}:
movdqu (%rsi), %xmm0
movdqu -16(%rsi,%rdx), %xmm1
@ -155,7 +155,7 @@ __folly_memcpy_{[prefetch]s}:
movdqu %xmm1, -16(%rdi,%rdx)
ret
.align 2
.balign 2
.L_GE193_LE256_{[prefetch]s}:
vmovdqu %ymm3, 96(%rdi)
vmovdqu %ymm4, -128(%rdi,%rdx)
@ -175,7 +175,7 @@ __folly_memcpy_{[prefetch]s}:
vzeroupper
ret
.align 2
.balign 2
.L_GE33_{[prefetch]s}:
vmovdqu (%rsi), %ymm0
vmovdqu -32(%rsi,%rdx), %ymm7
@ -238,7 +238,7 @@ __folly_memcpy_{[prefetch]s}:
// are unlikely.
ja .L_OVERLAP_{[prefetch]s}
.align 2
.balign 2
.L_NO_OVERLAP_{[prefetch]s}:
vmovdqu %ymm0, (%rdi)
vmovdqu %ymm1, 32(%rdi)
@ -266,7 +266,7 @@ __folly_memcpy_{[prefetch]s}:
cmp $32768, %rdx
jae .L_NON_TEMPORAL_LOOP_{[prefetch]s}
.align 2
.balign 2
.L_ALIGNED_DST_LOOP_{[prefetch]s}:
{[prefetch]s} 128(%rdi)
{[prefetch]s} 192(%rdi)
@ -298,7 +298,7 @@ __folly_memcpy_{[prefetch]s}:
vzeroupper
ret
.align 2
.balign 2
.L_NON_TEMPORAL_LOOP_{[prefetch]s}:
testb $31, %sil
jne .L_ALIGNED_DST_LOOP_{[prefetch]s}
@ -330,7 +330,7 @@ __folly_memcpy_{[prefetch]s}:
.L_OVERLAP_{[prefetch]s}:
.align 2
.balign 2
cmp %rdi, %rsi
jb .L_OVERLAP_BWD_{[prefetch]s} // %rsi < %rdi => backward-copy
je .L_RET_{[prefetch]s} // %rsi == %rdi => return, nothing to copy