[x265] [PATCH 5/6] blockcopy8-sve.S: Fix branch target in cpy1Dto2D_shr_32x32_sve

George Steed george.steed at arm.com
Mon Jan 6 17:17:36 UTC 2025


The branch target for longer vector lengths was mistakenly written as
.vl_gt_48_cpy2Dto1D_shr_32x32 (note 2Dto1D rather than 1Dto2D). Fix the
label so that the code works as intended.
---
 source/common/aarch64/blockcopy8-sve.S | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/source/common/aarch64/blockcopy8-sve.S b/source/common/aarch64/blockcopy8-sve.S
index d5664af58..1d742a64c 100644
--- a/source/common/aarch64/blockcopy8-sve.S
+++ b/source/common/aarch64/blockcopy8-sve.S
@@ -1275,7 +1275,7 @@ function PFX(cpy1Dto2D_shr_32x32_sve)
     sub             w4, w3, #1
     dup             z1.h, w4
     cmp             x9, #48
-    bgt             .vl_gt_48_cpy2Dto1D_shr_32x32
+    bgt             .vl_gt_48_cpy1Dto2D_shr_32x32
     ptrue           p0.h, vl16
     mov             z2.h, #1
     lsl             z2.h, p0/m, z2.h, z1.h
-- 
2.34.1

-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0005-blockcopy8-sve.S-Fix-branch-target-in-cpy1Dto2D_shr_.patch
Type: text/x-diff
Size: 1313 bytes
Desc: not available
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20250106/22366088/attachment.patch>


More information about the x265-devel mailing list