-
Notifications
You must be signed in to change notification settings - Fork 22
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
Remove crashy unwrap()
in on_error
#1364
base: main
Are you sure you want to change the base?
Conversation
if f.frame_hdr | ||
.as_ref() | ||
.is_some_and(|frame_hdr| frame_hdr.refresh_context != 0) | ||
{ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not sure what the right behavior is in the None
case, but crashing is certainly not it
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Places where there is an unwrap should have been unchecked derefs in C, so let me check what C was doing here.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
So C is doing the same thing as we are here:
Line 3350 in c7d127e
if (f->frame_hdr->refresh_context) - https://code.videolan.org/videolan/dav1d/-/blob/master/src/decode.c?ref_type=heads#L3722
Can you check if you get an error/crash on dav1d
here or not? Thanks!
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Using https://crates.io/crates/dav1d 0.10.3 I do not get a crash, but I get an Invalid argument
error and some useful info printed:
Error parsing frame header
Error parsing OBU data
It should be noted though that dereferencing a null pointer is not necessarily a crash in C - it is Undefined Behavior. The compiler is free to assume the pointer is not null, and do whatever crazy shenanigans that implies.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hmm, it's possible UB could be doing something weird here, but I'd still expect a segfault for the most part. We likely diverged at some point, so we'll have to figure out where and see if that's something we need to fix or if it's okay to crash, just in a different place.
That said, this video should successfully decode with normal dav1d
, but doesn't? Is that a bug with dav1d
as well?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
or if it's okay to crash
I would prefer a rav1d
that never crashes on any input, but returns errors.
That said, this video should successfully decode with normal
dav1d
, but doesn't? Is that a bug withdav1d
as well?
I don't know if the video should successfully decode, or if it is actually a bad video.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would prefer a rav1d that never crashes on any input, but returns errors.
Yup, that would definitely be preferable.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
What input were you running this on?
I was running this video (the crash happens about half-way through it): https://storage.googleapis.com/rerun-test-assets/video/Sintel_1080_10s_av1.mp4 The crash happens during decode of the chunk starting at timestamp |
Hi @emilk, I have some time to look into this bug now. What specific command did you use on |
Sorry, I don't have a simple repro for you. I used the dav1d API added in #1362 |
Hi @emilk. I'm going to need a way to reproduce this to figure it out. I tried running this: > wget https://storage.googleapis.com/rerun-test-assets/video/Sintel_1080_10s_av1.mp4
> ffmpeg -i Sintel_1080_10s_av1.mp4 -c:v copy -f ivf Sintel_1080_10s_av1.ivf
> cargo build --release
> target/release/dav1d -i tests/large/Sintel_1080_10s_av1.ivf -o /dev/null
dav1d 966d63c1 - by VideoLAN
Decoded 223/300 frames (74.3%) - 968.60/30.00 fps (32.29x)Error decoding frame: Invalid argument
Error decoding frame: Invalid argument
Decoded 298/300 frames (99.3%) - 897.67/30.00 fps (29.92x)
> dav1d -i tests/large/Sintel_1080_10s_av1.ivf -o /dev/null
dav1d 1.5.0 - by VideoLAN
Decoded 223/300 frames (74.3%) - 889.49/30.00 fps (29.65x)Error decoding frame: Invalid argument
Error decoding frame: Invalid argument
Decoded 298/300 frames (99.3%) - 857.72/30.00 fps (28.59x) So I'm getting the exact same two |
I got a crash here, and I don't want a crash :)
It seems like this
unwrap
was introduced in #663 along with many others.May I suggest enabling the
clippy::unwrap_used
lint? Any.unwrap()
is a panic waiting to happen :/EDIT: The crash happens during decode of the chunk starting at timestamp
114176
of this video: https://storage.googleapis.com/rerun-test-assets/video/Sintel_1080_10s_av1.mp4