-
Notifications
You must be signed in to change notification settings - Fork 11k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[regression] maybe a runtime issue related to D154953 #64339
Comments
|
If you're unable to use rsync to get the whole set of samples, you can manually download the specific one needed for this testcase - download http://fate-suite.ffmpeg.org/wmapro/Beethovens_9th-1_small.wma and place it in |
I think we missing a condition to check PowerC->isNonNegative()
|
I'll submit a commit similar to following, would you help me check the issue end to end? @mstorsjo
|
candidate MR: https://reviews.llvm.org/D156881 |
Thanks! The fix does seem to make all issues in this testsuite go away, when rebuilt with this patch. |
hi @mstorsjo, the fixing is land now, sorry for blocking your work |
Chromium also hit test failures after D154953, in BoringSSL. It appears that 68ea002 has fixed it, thanks! (For reference, our bug is https://crbug.com/1470223) |
Fixes llvm#64339 proofs: https://alive2.llvm.org/ce/z/yZ_I2a Reviewed By: goldstein.w.n Differential Revision: https://reviews.llvm.org/D156881
This is reported with @mstorsjo , so I'm checking:
This commit caused misoptimizations in the WMA decoder in ffmpeg, observed on all architectures. The misoptimization can be observed with https://martin.st/temp/wma-preproc.c, compiled with clang -target aarch64-linux-gnu -c -O3 wma-preproc.c -o libavcodec/wma.o
For a full runtime reproducible case:
The text was updated successfully, but these errors were encountered: