Re: Build-and-test run on FC6 x86_64 with newest kernel

[Kris Van Hees <kris dot van dot hees at oracle dot com>]

I'll post this to the bugzilla entry as well, but we're still seeing
hangs on both i386 and x86_64 FC6 systems, with the 2.6.20-1.2925.fc6
kernel.  The hangs are much more severe, with no messages being printed
to the kernel log due to hanging in an IRQ.

Here is the output on the console:
Stack:  ffff81000483b160 ffff810037e62500 ffff8100613820c0
000000000000000a
 0000000000000001 ffffffff802ccadc 0000000000000009 ffff8100b707a2c0
 ffff81000483f1a0 ffffffff80294fdb ffff81000483f2a0 0000000000000000
Call Trace:
 <IRQ>  [<ffffffff802ccadc>] kmem_rcu_free+0x19/0x38
 [<ffffffff80294fdb>] __rcu_process_callbacks+0x12d/0x1bc
 [<ffffffff8029508d>] rcu_process_callbacks+0x23/0x43
 [<ffffffff8028b742>] tasklet_action+0x53/0x9d
 [<ffffffff8025a23c>] call_softirq+0x1c/0x28
 [<ffffffff80211f97>] __do_softirq+0x55/0x2c
 [<ffffffff8025a23c>] call_softirq+0x1c/0x28
 <EOI>  [<ffffffff8028b630>] ksoftirqd+0x0/0xbf
 [<ffffffff8026758d>] do_softirq+0x2c/0x85
 [<ffffffff8028b68f>] ksoftirqd+0x5f/0xbf
 [<ffffffff802317ea>] kthread+0xd0/0xff
 [<ffffffff80259ec8>] child_rip+0xa/0x12
 [<ffffffff8023171a>] kthread+0x0/0xff
 [<ffffffff80259ebe>] child_rip+0x0/0x12


Code: 0f 0b eb fe 0f ba 33 07 48 83 c3 40 48 ff ca 48 83 fa ff 75
RIP  [<ffffffff802cab21>] kmem_freepages+0xef/0x132
 RSP <ffff8100df623ee8>
 <0>Kernel panic - not syncing: Aiee, killing interrupt handler!

On Thu, Mar 15, 2007 at 03:10:10PM -0700, Roland McGrath wrote:
> When you mention any kernel, please give a precise identifier,
> like the rpm version-release and architecture.
> 
> 
> Thanks,
> Roland