aboutsummaryrefslogtreecommitdiff
path: root/sysdeps/x86_64/multiarch/wcscpy-ssse3.S
diff options
context:
space:
mode:
Diffstat (limited to 'sysdeps/x86_64/multiarch/wcscpy-ssse3.S')
-rw-r--r--sysdeps/x86_64/multiarch/wcscpy-ssse3.S171
1 files changed, 171 insertions, 0 deletions
diff --git a/sysdeps/x86_64/multiarch/wcscpy-ssse3.S b/sysdeps/x86_64/multiarch/wcscpy-ssse3.S
index b7de092228..77889dd555 100644
--- a/sysdeps/x86_64/multiarch/wcscpy-ssse3.S
+++ b/sysdeps/x86_64/multiarch/wcscpy-ssse3.S
@@ -25,13 +25,27 @@ ENTRY (__wcscpy_ssse3)
mov %rsi, %rcx
mov %rdi, %rdx
+# ifdef __CHKP__
+ bndcl (%rdi), %bnd0
+ bndcl (%rsi), %bnd1
+ bndcu (%rsi), %bnd1
+# endif
cmpl $0, (%rcx)
jz L(Exit4)
+# ifdef __CHKP__
+ bndcu 4(%rcx), %bnd1
+# endif
cmpl $0, 4(%rcx)
jz L(Exit8)
+# ifdef __CHKP__
+ bndcu 8(%rcx), %bnd1
+# endif
cmpl $0, 8(%rcx)
jz L(Exit12)
+# ifdef __CHKP__
+ bndcu 12(%rcx), %bnd1
+# endif
cmpl $0, 12(%rcx)
jz L(Exit16)
@@ -40,10 +54,19 @@ ENTRY (__wcscpy_ssse3)
pxor %xmm0, %xmm0
mov (%rcx), %r9
+# ifdef __CHKP__
+ bndcu 7(%rdx), %bnd0
+# endif
mov %r9, (%rdx)
+# ifdef __CHKP__
+ bndcu (%rsi), %bnd1
+# endif
pcmpeqd (%rsi), %xmm0
mov 8(%rcx), %r9
+# ifdef __CHKP__
+ bndcu 15(%rdx), %bnd0
+# endif
mov %r9, 8(%rdx)
pmovmskb %xmm0, %rax
@@ -72,6 +95,10 @@ ENTRY (__wcscpy_ssse3)
jmp L(Shl12)
L(Align16Both):
+# ifdef __CHKP__
+ bndcu 16(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps (%rcx), %xmm1
movaps 16(%rcx), %xmm2
movaps %xmm1, (%rdx)
@@ -82,6 +109,10 @@ L(Align16Both):
test %rax, %rax
jnz L(CopyFrom1To16Bytes)
+# ifdef __CHKP__
+ bndcu 16(%rcx, %rsi), %bnd1
+ bndcu 15(%rdx, %rsi), %bnd0
+# endif
movaps 16(%rcx, %rsi), %xmm3
movaps %xmm2, (%rdx, %rsi)
pcmpeqd %xmm3, %xmm0
@@ -91,6 +122,10 @@ L(Align16Both):
test %rax, %rax
jnz L(CopyFrom1To16Bytes)
+# ifdef __CHKP__
+ bndcu 16(%rcx, %rsi), %bnd1
+ bndcu 15(%rdx, %rsi), %bnd0
+# endif
movaps 16(%rcx, %rsi), %xmm4
movaps %xmm3, (%rdx, %rsi)
pcmpeqd %xmm4, %xmm0
@@ -100,6 +135,10 @@ L(Align16Both):
test %rax, %rax
jnz L(CopyFrom1To16Bytes)
+# ifdef __CHKP__
+ bndcu 16(%rcx, %rsi), %bnd1
+ bndcu 15(%rdx, %rsi), %bnd0
+# endif
movaps 16(%rcx, %rsi), %xmm1
movaps %xmm4, (%rdx, %rsi)
pcmpeqd %xmm1, %xmm0
@@ -109,6 +148,10 @@ L(Align16Both):
test %rax, %rax
jnz L(CopyFrom1To16Bytes)
+# ifdef __CHKP__
+ bndcu 16(%rcx, %rsi), %bnd1
+ bndcu 15(%rdx, %rsi), %bnd0
+# endif
movaps 16(%rcx, %rsi), %xmm2
movaps %xmm1, (%rdx, %rsi)
pcmpeqd %xmm2, %xmm0
@@ -118,6 +161,10 @@ L(Align16Both):
test %rax, %rax
jnz L(CopyFrom1To16Bytes)
+# ifdef __CHKP__
+ bndcu 16(%rcx, %rsi), %bnd1
+ bndcu 15(%rdx, %rsi), %bnd0
+# endif
movaps 16(%rcx, %rsi), %xmm3
movaps %xmm2, (%rdx, %rsi)
pcmpeqd %xmm3, %xmm0
@@ -127,6 +174,10 @@ L(Align16Both):
test %rax, %rax
jnz L(CopyFrom1To16Bytes)
+# ifdef __CHKP__
+ bndcu 16(%rcx, %rsi), %bnd1
+ bndcu 15(%rdx, %rsi), %bnd0
+# endif
movaps %xmm3, (%rdx, %rsi)
mov %rcx, %rax
lea 16(%rcx, %rsi), %rcx
@@ -138,6 +189,10 @@ L(Align16Both):
.p2align 4
L(Aligned64Loop):
+# ifdef __CHKP__
+ bndcu (%rcx), %bnd1
+ bndcu 63(%rdx), %bnd0
+# endif
movaps (%rcx), %xmm2
movaps %xmm2, %xmm4
movaps 16(%rcx), %xmm5
@@ -168,6 +223,9 @@ L(Aligned64Leave):
pcmpeqd %xmm5, %xmm0
pmovmskb %xmm0, %rax
+# ifdef __CHKP__
+ bndcu -49(%rdx), %bnd0
+# endif
movaps %xmm4, -64(%rdx)
test %rax, %rax
lea 16(%rsi), %rsi
@@ -176,11 +234,17 @@ L(Aligned64Leave):
pcmpeqd %xmm6, %xmm0
pmovmskb %xmm0, %rax
+# ifdef __CHKP__
+ bndcu -33(%rdx), %bnd0
+# endif
movaps %xmm5, -48(%rdx)
test %rax, %rax
lea 16(%rsi), %rsi
jnz L(CopyFrom1To16Bytes)
+# ifdef __CHKP__
+ bndcu -17(%rdx), %bnd0
+# endif
movaps %xmm6, -32(%rdx)
pcmpeqd %xmm7, %xmm0
@@ -190,11 +254,17 @@ L(Aligned64Leave):
jnz L(CopyFrom1To16Bytes)
mov $-0x40, %rsi
+# ifdef __CHKP__
+ bndcu -1(%rdx), %bnd0
+# endif
movaps %xmm7, -16(%rdx)
jmp L(Aligned64Loop)
.p2align 4
L(Shl4):
+# ifdef __CHKP__
+ bndcu 12(%rcx), %bnd1
+# endif
movaps -4(%rcx), %xmm1
movaps 12(%rcx), %xmm2
L(Shl4Start):
@@ -206,6 +276,10 @@ L(Shl4Start):
jnz L(Shl4LoopExit)
palignr $4, %xmm1, %xmm2
+# ifdef __CHKP__
+ bndcu 28(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
movaps 28(%rcx), %xmm2
@@ -219,6 +293,10 @@ L(Shl4Start):
jnz L(Shl4LoopExit)
palignr $4, %xmm3, %xmm2
+# ifdef __CHKP__
+ bndcu 28(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
movaps 28(%rcx), %xmm2
@@ -232,6 +310,10 @@ L(Shl4Start):
jnz L(Shl4LoopExit)
palignr $4, %xmm1, %xmm2
+# ifdef __CHKP__
+ bndcu 28(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
movaps 28(%rcx), %xmm2
@@ -244,6 +326,9 @@ L(Shl4Start):
jnz L(Shl4LoopExit)
palignr $4, %xmm3, %xmm2
+# ifdef __CHKP__
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
lea 28(%rcx), %rcx
lea 16(%rdx), %rdx
@@ -258,6 +343,9 @@ L(Shl4Start):
.p2align 4
L(Shl4LoopStart):
+# ifdef __CHKP__
+ bndcu 12(%rcx), %bnd1
+# endif
movaps 12(%rcx), %xmm2
movaps 28(%rcx), %xmm3
movaps %xmm3, %xmm6
@@ -279,6 +367,9 @@ L(Shl4LoopStart):
lea 64(%rcx), %rcx
palignr $4, %xmm1, %xmm2
movaps %xmm7, %xmm1
+# ifdef __CHKP__
+ bndcu 63(%rdx), %bnd0
+# endif
movaps %xmm5, 48(%rdx)
movaps %xmm4, 32(%rdx)
movaps %xmm3, 16(%rdx)
@@ -287,6 +378,10 @@ L(Shl4LoopStart):
jmp L(Shl4LoopStart)
L(Shl4LoopExit):
+# ifdef __CHKP__
+ bndcu -4(%rcx), %bnd1
+ bndcu 11(%rdx), %bnd0
+# endif
movdqu -4(%rcx), %xmm1
mov $12, %rsi
movdqu %xmm1, -4(%rdx)
@@ -294,6 +389,9 @@ L(Shl4LoopExit):
.p2align 4
L(Shl8):
+# ifdef __CHKP__
+ bndcu 8(%rcx), %bnd1
+# endif
movaps -8(%rcx), %xmm1
movaps 8(%rcx), %xmm2
L(Shl8Start):
@@ -305,6 +403,10 @@ L(Shl8Start):
jnz L(Shl8LoopExit)
palignr $8, %xmm1, %xmm2
+# ifdef __CHKP__
+ bndcu 24(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
movaps 24(%rcx), %xmm2
@@ -318,6 +420,10 @@ L(Shl8Start):
jnz L(Shl8LoopExit)
palignr $8, %xmm3, %xmm2
+# ifdef __CHKP__
+ bndcu 24(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
movaps 24(%rcx), %xmm2
@@ -331,6 +437,10 @@ L(Shl8Start):
jnz L(Shl8LoopExit)
palignr $8, %xmm1, %xmm2
+# ifdef __CHKP__
+ bndcu 24(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
movaps 24(%rcx), %xmm2
@@ -343,6 +453,10 @@ L(Shl8Start):
jnz L(Shl8LoopExit)
palignr $8, %xmm3, %xmm2
+# ifdef __CHKP__
+ bndcu 24(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
lea 24(%rcx), %rcx
lea 16(%rdx), %rdx
@@ -357,6 +471,9 @@ L(Shl8Start):
.p2align 4
L(Shl8LoopStart):
+# ifdef __CHKP__
+ bndcu 8(%rcx), %bnd1
+# endif
movaps 8(%rcx), %xmm2
movaps 24(%rcx), %xmm3
movaps %xmm3, %xmm6
@@ -378,6 +495,9 @@ L(Shl8LoopStart):
lea 64(%rcx), %rcx
palignr $8, %xmm1, %xmm2
movaps %xmm7, %xmm1
+# ifdef __CHKP__
+ bndcu 63(%rdx), %bnd0
+# endif
movaps %xmm5, 48(%rdx)
movaps %xmm4, 32(%rdx)
movaps %xmm3, 16(%rdx)
@@ -386,6 +506,10 @@ L(Shl8LoopStart):
jmp L(Shl8LoopStart)
L(Shl8LoopExit):
+# ifdef __CHKP__
+ bndcu (%rcx), %bnd1
+ bndcu 7(%rdx), %bnd0
+# endif
mov (%rcx), %r9
mov $8, %rsi
mov %r9, (%rdx)
@@ -393,6 +517,9 @@ L(Shl8LoopExit):
.p2align 4
L(Shl12):
+# ifdef __CHKP__
+ bndcu 4(%rcx), %bnd1
+# endif
movaps -12(%rcx), %xmm1
movaps 4(%rcx), %xmm2
L(Shl12Start):
@@ -404,6 +531,10 @@ L(Shl12Start):
jnz L(Shl12LoopExit)
palignr $12, %xmm1, %xmm2
+# ifdef __CHKP__
+ bndcu 20(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
movaps 20(%rcx), %xmm2
@@ -417,6 +548,10 @@ L(Shl12Start):
jnz L(Shl12LoopExit)
palignr $12, %xmm3, %xmm2
+# ifdef __CHKP__
+ bndcu 20(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
movaps 20(%rcx), %xmm2
@@ -430,6 +565,10 @@ L(Shl12Start):
jnz L(Shl12LoopExit)
palignr $12, %xmm1, %xmm2
+# ifdef __CHKP__
+ bndcu 20(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
movaps 20(%rcx), %xmm2
@@ -442,6 +581,10 @@ L(Shl12Start):
jnz L(Shl12LoopExit)
palignr $12, %xmm3, %xmm2
+# ifdef __CHKP__
+ bndcu 20(%rcx), %bnd1
+ bndcu 15(%rdx), %bnd0
+# endif
movaps %xmm2, (%rdx)
lea 20(%rcx), %rcx
lea 16(%rdx), %rdx
@@ -456,6 +599,9 @@ L(Shl12Start):
.p2align 4
L(Shl12LoopStart):
+# ifdef __CHKP__
+ bndcu 4(%rcx), %bnd1
+# endif
movaps 4(%rcx), %xmm2
movaps 20(%rcx), %xmm3
movaps %xmm3, %xmm6
@@ -476,6 +622,9 @@ L(Shl12LoopStart):
lea 64(%rcx), %rcx
palignr $12, %xmm1, %xmm2
movaps %xmm7, %xmm1
+# ifdef __CHKP__
+ bndcu 63(%rdx), %bnd0
+# endif
movaps %xmm5, 48(%rdx)
movaps %xmm4, 32(%rdx)
movaps %xmm3, 16(%rdx)
@@ -484,6 +633,10 @@ L(Shl12LoopStart):
jmp L(Shl12LoopStart)
L(Shl12LoopExit):
+# ifdef __CHKP__
+ bndcu (%rcx), %bnd1
+ bndcu 3(%rdx), %bnd0
+# endif
mov (%rcx), %r9d
mov $4, %rsi
mov %r9d, (%rdx)
@@ -500,6 +653,9 @@ L(CopyFrom1To16Bytes):
jnz L(Exit4)
mov (%rcx), %rax
+# ifdef __CHKP__
+ bndcu 7(%rdx), %bnd0
+# endif
mov %rax, (%rdx)
mov %rdi, %rax
ret
@@ -510,6 +666,9 @@ L(ExitHigh):
jnz L(Exit12)
mov (%rcx), %rax
+# ifdef __CHKP__
+ bndcu 15(%rdx), %bnd0
+# endif
mov %rax, (%rdx)
mov 8(%rcx), %rax
mov %rax, 8(%rdx)
@@ -519,6 +678,9 @@ L(ExitHigh):
.p2align 4
L(Exit4):
movl (%rcx), %eax
+# ifdef __CHKP__
+ bndcu 3(%rdx), %bnd0
+# endif
movl %eax, (%rdx)
mov %rdi, %rax
ret
@@ -526,6 +688,9 @@ L(Exit4):
.p2align 4
L(Exit8):
mov (%rcx), %rax
+# ifdef __CHKP__
+ bndcu 7(%rdx), %bnd0
+# endif
mov %rax, (%rdx)
mov %rdi, %rax
ret
@@ -533,6 +698,9 @@ L(Exit8):
.p2align 4
L(Exit12):
mov (%rcx), %rax
+# ifdef __CHKP__
+ bndcu 11(%rdx), %bnd0
+# endif
mov %rax, (%rdx)
mov 8(%rcx), %eax
mov %eax, 8(%rdx)
@@ -542,6 +710,9 @@ L(Exit12):
.p2align 4
L(Exit16):
mov (%rcx), %rax
+# ifdef __CHKP__
+ bndcu 15(%rdx), %bnd0
+# endif
mov %rax, (%rdx)
mov 8(%rcx), %rax
mov %rax, 8(%rdx)
generated by cgit v1.2.3 (git 2.25.1) at 2024-06-13 09:58:15 +0000