On 2018年01月03日 18:14, Dmitry Katsubo wrote: > On 2018-01-03 05:58, Qu Wenruo wrote: >> On 2018年01月03日 09:12, Dmitry Katsubo wrote: >>> Dear btrfs team, >>> >>> I send a kernel crash report which I have observed recently during btrfs scrub. >>> It looks like scrub itself has completed without errors. >> >> It's not a kernel crash (if I didn't miss anything), but just kernel >> warning. >> >> The warning is caused by the fact that your fs (mostly created by old >> mkfs.btrfs) has device with unaligned size. >> >> You could either resize the device down a little (e.g. -4K) and newer >> kernel (the one you're using should be new enough) could handle it well. >> >> Or you could update your btrfs-progs (I assume you're using Arch, which >> is already shipping btrfs-progs v4.14) and use "btrfs rescue >> fix-device-size" to fix other device related problems offline. >> (Not only the warning, but also potential superblock size mismatch) >> >> Thanks, >> Qu > > Thanks for reply! > > Why couldn't a warning message be issued as one-liner, e.g. with proper > description and without scaring stack trace? Because we're using WARN_ON() and the backtrace is in fact very useful for developers to locate the problem. For one-line message, the problem is we don't know who caused the problem if there are multiple callers. But since the cause of the problem is located, one-line output seems better. Thanks, Qu > > btrfs /dev/sda1 warning: device size is not aligned with FS (mostly created by old mkfs.btrfs), see https://btrfs.wiki.kernel.org/index.php/FAQ#... >
Attachment:
signature.asc
Description: OpenPGP digital signature
