kernel 3.3 btrfs possible corrupt data

[Simon Jones]

Hi,

I wanted to see if anyone else has had or spotted a potential issue
with regard to some corruption of data.

I have a tv pc running Arch, and my laptop running Arch, over the
weekend I noticed some issues with my tv system and went to recompile
myth, this failed and caused a lock of the system with a segfault of
the btrfs module (all partitions are btrfs except boot)

I rebooted thinking it was a one off and re-tried, I currently have
issues with usb and my dvb-t2 tuner, so tried again and it did the
same with locking up and the segfault, so seeing the btrfs issue bug I
ran 3.3 to see if that was the issue, this then caused me to see more
data corruption in the mysql database for myth, trying to repair it
caused a lock up.

I was also seeing btrfs errors of checksum not matching etc, so to try
and repair I took the SSD out and put it in my laptop, I couldn't seem
to repair it wit the new btrfsck and it seems that's it's failing with
an assert.c line 147 error.

Reading up about SSD and the things they do, I believe because it's
using a sandforce controller, the deduplication of blocks has caused
an error that can not be repaired due to the additional blocks not
being available.

On top of this when the ssd was in my laptop, I couldn't unmount the
SSD and shutdown so had to poweroff, this was fine, but when I used
kernel 3.3 my boot system got damaged and I had to do a repair using a
live boot, this corruption only happens when running on kernel 3.3 and
not shutting down cleanly.

I hope this helps with anyone else who thinks they may be having issues.

Regards,

Simon
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/