PR #22646 opened by mkver URL: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22646 Patch URL: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/22646.patch
It is only needed in the unlikely codepath. The ordinary one only uses six xmm registers. >From afc76ae1529cc3d45afdcd3023bb7df912c00ccd Mon Sep 17 00:00:00 2001 From: Andreas Rheinhardt <[email protected]> Date: Fri, 27 Mar 2026 22:10:03 +0100 Subject: [PATCH] avcodec/x86/h264_idct: Avoid spilling register unnecessarily It is only needed in the unlikely codepath. The ordinary one only uses six xmm registers. Signed-off-by: Andreas Rheinhardt <[email protected]> --- libavcodec/x86/h264_idct.asm | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/libavcodec/x86/h264_idct.asm b/libavcodec/x86/h264_idct.asm index 6ae8202748..0ced412a3e 100644 --- a/libavcodec/x86/h264_idct.asm +++ b/libavcodec/x86/h264_idct.asm @@ -647,7 +647,7 @@ RET %endmacro INIT_XMM sse2 -cglobal h264_luma_dc_dequant_idct, 3, 4, 7 +cglobal h264_luma_dc_dequant_idct, 3, 4, 6 movq m3, [r1+24] movq m2, [r1+16] movq m1, [r1+ 8] @@ -683,6 +683,7 @@ cglobal h264_luma_dc_dequant_idct, 3, 4, 7 RET .big_qmul: bsr t0d, t3d + WIN64_PUSH_XMM 7 add t3d, 128 << 16 mov t1d, 7 cmp t0d, t1d -- 2.52.0 _______________________________________________ ffmpeg-devel mailing list -- [email protected] To unsubscribe send an email to [email protected]
