| |||||||||||||
Kernel prepatch 3.8-rc2
The second 3.8 prepatch is out. Linus
says: "It's a new year, people are getting back to work, and trying
desperately to forget the over-eating that has been going on for the last
two weeks. And hey, to celebrate, here's -rc2!" Perhaps as a result
of the aforementioned over-eating, this patch is relatively small; things
might be expected to pick up a bit by the time -rc3 comes out.
(Log in to post comments)
3.8-rc1+ transparent-huge-page zero-page split bugons Posted Jan 3, 2013 6:49 UTC (Thu) by Duncan (guest, #6647) [Link]
I've been seeing regular huge-page split kernel-bugons with 3.8-r1-g*, and don't see any fixes. For me, they seem to occur pretty regularly when I close my last firefox window.
I suspect it's the new transparent-huge-page zero-pages. Firefox apparently allocates some but never writes to them, then when it quits, I get the bugon apparently as the zero-huge-page is split back into its components to return to the pool. The code for splitting huge-pages apparently doesn't yet have an appropriate exception for the zero-huge-page split-and-return-to-pool case.
I just disabled transparent-huge-pages when I rebuilt for rc2. That should work around the problem for now, I hope.
I haven't bug-reported it yet. No lkm subscription so I usually file a bug on kernel bugzilla, but haven't on this one yet.
3.8-rc1+ transparent-huge-page zero-page split bugons Posted Jan 3, 2013 6:57 UTC (Thu) by ebiederm (subscriber, #35028) [Link]
lkml does not require a subscription to post.
Sending an email with backtraces to the developers of of that part of the kernel should get your issue looked at and with a little luck resolved quickly.
3.8-rc1+ transparent-huge-page zero-page split bugons Posted Jan 3, 2013 18:12 UTC (Thu) by Tara_Li (subscriber, #26706) [Link]
Yeah - if you can fairly reliably reproduce this - make the devs happy with one less heisenbug!
|
|||||||||||||