On Sat, 2012-05-19 at 18:38 -0500, Jonathan Nieder wrote: > reopen 619381 > forwarded 619381 https://bugzilla.kernel.org/show_bug.cgi?id=31802 > tags 619381 + moreinfo > quit > > Jonathan Nieder wrote: > >>> On Wed, 2011-03-23 at 13:35 +0100, Michal Suchanek wrote: > > >>>> 1) Attach a SATA drive to a SATA/USB bridge. > >>>> 2) dd if=/dev/urandom of=/dev/sdf bs=4k > >>>> 3) detach the drive drom the bridge > >>>> > >>>> Results in kernel panic. > [...] > >>>> Pid: 15054, comm: dd Not tainted 2.6.38-1-amd64 #1 Dell Inc. OptiPlex > >>>> 960 /0Y958C > >>>> RIP: 0010:[<ffffffff8103b9ce>] [<ffffffff8103b9ce>] > >>>> task_rq_lock+0x37/0x7a > [...] > >>>> Code: [...] 89 dd 8b 40 18 <4c> 03 2c c5 d0 08 [...] > [...] > > forcemerge 631187 619381 > [...] > > Thanks again for a pleasant report. Closing optimistically. As usual, > > confirmation either way about the fix would be useful. > > Actually, these can't be the same bug, because 619381 was found using > 2.6.38-1 and 631187 was tested not to affect 2.6.38-5. > > What kernel are you using these days? Can you still reproduce this?
These bug reports involve different failure cases (soft-removal in #631187 and hard-removal in #619381). However, it's conceivable that the underlying bug was the same, exacerbated by some change in 2.6.39 so that soft-removal was also affected. Ben. -- Ben Hutchings All extremists should be taken out and shot.
signature.asc
Description: This is a digitally signed message part
