I'm afraid the fs corruption thingie caused by the bttv driver doesn't
look finally solved to me. That's what happened:
I had already been used to getting tons of (mostly minor) errors on
e2fsck's. When I heard of Richard's bug report some weeks ago blaming the
bttv driver, I stopped using any v4l stuff and my filesystems stayed
clean. Over the weekend I installed 2.2.2-ac5 including your bttv fixes
and started playing around with some v4l stuff. Now guess what - yesterday
I spent a happy hour in lost+found... Of course I am unable to give any
hard evidence that the bttv driver is to blame but the apparent
correlation between use of bttv and fs corruption makes it a good main
suspect. (It was the only unusual kernel module I've been using over the
weekend.)
Talking to Richard tonight he told me that the test program he used in
identifying the bug showed corruption sometimes several seconds after the
close(), so the 0.1s delay you added might be far too small. Or there are
even more bug hiding...
For completeness: The system in question is a Dual-PII/350, 256MB,
IBM-DHEA w/DMA turned on, WinTV w/BT848, v4l version from 2.2.2-ac5.
Regards,
Daniel.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/