Fix Win64 ABI conformance when using xmm8-xmm11

Referring to https://docs.microsoft.com/en-US/cpp/build/stack-usage:

"All memory beyond the current address of RSP is considered volatile:
The OS, or a debugger, may overwrite this memory during a user debug
session, or an interrupt handler.  Thus, RSP must always be set before
attempting to read or write values to a stack frame."

Basically, if-- under extremely rare circumstances-- a context swap were
to occur between saving the values of xmm8-xmm11 and setting the new
value of rsp, the O/S might not preserve that area of the stack.  In
general, libjpeg-turbo should not be using xmm8-xmm11 before or after
the call to jsimd_huff_encode_one_block_sse2(), so this is probably a
non-issue, but it's still a good idea to fix it.

Based on
https://github.com/mayeut/libjpeg-turbo/commit/ff7d2030dd26c7b8c37ff540c594490198843f15

diff --git a/simd/jchuff-sse2-64.asm b/simd/jchuff-sse2-64.asm
index b1144d1cddef3195f92986f0edcddd6f4433cb77..0deb3e6fb420593089a5a593db805e22d89945e2 100644 (file)
--- a/simd/jchuff-sse2-64.asm
+++ b/simd/jchuff-sse2-64.asm
@@ -195,11 +195,11 @@ EXTN(jsimd_huff_encode_one_block_sse2):
         lea     rsp, [t2]
         collect_args
 %ifdef WIN64
-        movaps  XMMWORD [rsp-1*SIZEOF_XMMWORD], xmm8
-        movaps  XMMWORD [rsp-2*SIZEOF_XMMWORD], xmm9
-        movaps  XMMWORD [rsp-3*SIZEOF_XMMWORD], xmm10
-        movaps  XMMWORD [rsp-4*SIZEOF_XMMWORD], xmm11
         sub     rsp, 4*SIZEOF_XMMWORD
+        movaps  XMMWORD [rsp+0*SIZEOF_XMMWORD], xmm8
+        movaps  XMMWORD [rsp+1*SIZEOF_XMMWORD], xmm9
+        movaps  XMMWORD [rsp+2*SIZEOF_XMMWORD], xmm10
+        movaps  XMMWORD [rsp+3*SIZEOF_XMMWORD], xmm11
 %endif
         push rbx
 
@@ -343,10 +343,10 @@ EXTN(jsimd_huff_encode_one_block_sse2):
 
         pop rbx
 %ifdef WIN64
-        movaps  xmm11, XMMWORD [rsp+0*SIZEOF_XMMWORD]
-        movaps  xmm10, XMMWORD [rsp+1*SIZEOF_XMMWORD]
-        movaps  xmm9, XMMWORD [rsp+2*SIZEOF_XMMWORD]
-        movaps  xmm8, XMMWORD [rsp+3*SIZEOF_XMMWORD]
+        movaps  xmm8, XMMWORD [rsp+0*SIZEOF_XMMWORD]
+        movaps  xmm9, XMMWORD [rsp+1*SIZEOF_XMMWORD]
+        movaps  xmm10, XMMWORD [rsp+2*SIZEOF_XMMWORD]
+        movaps  xmm11, XMMWORD [rsp+3*SIZEOF_XMMWORD]
         add     rsp, 4*SIZEOF_XMMWORD
 %endif
         uncollect_args