Merge pull request #2049 from Celelibi/fix_crash_sgemm_sse_x64

Fix crash in sgemm SSE/nano kernel on x86_64
OpenMathLib · Mar 7, 2019 · 8d3d29e · 8d3d29e
2 parents db3dc9e + b7f59da
commit 8d3d29e
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/kernel/x86_64/gemm_kernel_4x8_nano.S b/kernel/x86_64/gemm_kernel_4x8_nano.S
@@ -135,7 +135,7 @@
 #endif
 
 	movq	%rsp, %rbx	# save old stack
-	subq	$128 + LOCAL_BUFFER_SIZE, %rsp
+	subq	$256 + LOCAL_BUFFER_SIZE, %rsp
 	andq	$-4096, %rsp	# align stack
 
 	STACK_TOUCHING

diff --git a/kernel/x86_64/gemm_kernel_8x4_sse.S b/kernel/x86_64/gemm_kernel_8x4_sse.S
@@ -383,7 +383,7 @@
 	EMMS
 
 	movq	%rsp, %rbx	# save old stack
-	subq	$128 + LOCAL_BUFFER_SIZE, %rsp
+	subq	$256 + LOCAL_BUFFER_SIZE, %rsp
 	andq	$-4096, %rsp	# align stack
 
 	STACK_TOUCHING