[FFmpeg-cvslog] x86/tx_float: correctly load the transform length

Lynne git at videolan.org
Sun Jul 18 16:05:17 EEST 2021


ffmpeg | branch: master | Lynne <dev at lynne.ee> | Sun Jul 18 15:00:48 2021 +0200| [997f9bdb99de18b27f8f213f381c5e90536a1be9] | committer: Lynne

x86/tx_float: correctly load the transform length

The field is a standard field, yet we were loading it as if it was
a quadword. This worked for forward transforms by chance, but broke
when the transform was inverse.
checkasm couldn't catch that because we only test forward transforms,
which are identical to inverse transforms but with a different revtab.

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=997f9bdb99de18b27f8f213f381c5e90536a1be9
---

 libavutil/x86/tx_float.asm | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libavutil/x86/tx_float.asm b/libavutil/x86/tx_float.asm
index b71fb49592..4d2283fae1 100644
--- a/libavutil/x86/tx_float.asm
+++ b/libavutil/x86/tx_float.asm
@@ -884,7 +884,7 @@ ALIGN 16
 %macro FFT_SPLIT_RADIX_FN 1
 INIT_YMM %1
 cglobal split_radix_fft_float, 4, 8, 16, 272, lut, out, in, len, tmp, itab, rtab, tgt
-    mov lenq, [lutq + AVTXContext.m]
+    movsxd lenq, dword [lutq + AVTXContext.m]
     mov lutq, [lutq + AVTXContext.revtab]
     mov tgtq, lenq
 



More information about the ffmpeg-cvslog mailing list