Bug 205417
| Summary: | Files corruption ( fs/ext4/inode.c:3941 ext4_set_page_dirty+0x3e/0x50 [ext4] ) | ||
|---|---|---|---|
| Product: | File System | Reporter: | Ivan Baidakou (the.dmol) |
| Component: | ext4 | Assignee: | fs_ext4 (fs_ext4) |
| Status: | RESOLVED CODE_FIX | ||
| Severity: | normal | CC: | davidmnoriega, jack, tytso |
| Priority: | P1 | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Kernel Version: | 5.3.8 | Subsystem: | |
| Regression: | No | Bisected commit-id: | |
| Attachments: | /var/log/messages before I powered it off | ||
|
Description
Ivan Baidakou
2019-11-04 13:14:31 UTC
Can you send the complete /var/log/messages and/or /var/log/kern.log from say, a day before the hang up to the hang? These sorts of issues can be caused by I/O errors and/or other problems in the device driver and/or device mapper storage stack. So please also describe your hardware and storage configuration, and what distribution are you using. Also, is this a 5.3.8 upstream kernel which you compiled yourself or a distro-supplied kernel? If you compiled it yourself, please also send the kernel config Thanks! Created attachment 285781 [details]
/var/log/messages before I powered it off
> So please also describe your hardware and storage configuration, and what > distribution are you using. I use void linux distribution. My notebook is HP Probook 450 G5 ( https://support.hp.com/gb-en/document/c05682645 ). > Also, is this a 5.3.8 upstream kernel which you compiled yourself or a > distro-supplied kernel? If you compiled it yourself, please also send the > kernel config I use my distro kernel. I think their patches/config can be found at https://github.com/void-linux/void-packages/tree/master/srcpkgs/linux5.3/ Also, please note, that the reported issue (at least the reported stack trace) does not happen on 5.2.21 (again, shipped by void-linux). I currently us it. Hmm, so less than ten minutes after the system was booted, with no other interesting messages in /var/log/messages. Is this reproducible? If you boot 5.3.8 again, can you reliably reproduce the failure? And what sort of workload are you running when the system goes south? Thanks! Yes, the original message is reproduced every time I boot with 5.3.8 approximately after a few mins boot. I just launch my DE(awesome), Firefox, Telegram, Claws-mail, and terminal. Actually visually no buggy behavior can be observed, except the message in /var/log/messages. The buggy behavior was in Saturnday, when 1. I launched a WarThunder game, which does heavy I/O (I assume it calculates checksums of files, and if they mismatch, then download new asseets) - it just hanged. 2. Then I was sure that it is a bug in game, and removed it. A bit later I just spawned "make -j9" several times simultaneously for the same project (debug, release, and sanitizer-build) - and when it hanged, I finally payed attention to the original kernel message. I'm still not sure that the two events are linked: the kernel message and hangs on I/O a few hours later. I'm also seeing this message on Fedora 30, 5.3.8-200.fc30.x86_64 Could be the same as: https://bugs.freedesktop.org/show_bug.cgi?id=112012 https://bugzilla.redhat.com/show_bug.cgi?id=1758948 I believe this got tracked down to a problematic change in i915 driver in https://bugs.freedesktop.org/show_bug.cgi?id=111601 and should be fixed by now. Closing. |