Bug 11569 - Panic stop CPUs regression
Summary: Panic stop CPUs regression
Alias: None
Product: Other
Classification: Unclassified
Component: Other (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: Andi Kleen
Depends on:
Blocks: Regressions-2.6.26
  Show dependency tree
Reported: 2008-09-14 21:59 UTC by Rafael J. Wysocki
Modified: 2010-10-07 18:40 UTC (History)
3 users (show)

See Also:
Kernel Version: 2.6.27-rc
Regression: Yes
Bisected commit-id:


Description Rafael J. Wysocki 2008-09-14 21:59:44 UTC
Subject    : Don't complain about disabled irqs when the system has paniced
Submitter  : Andi Kleen <andi@firstfloor.org>
Date       : 2008-09-02 13:49
References : http://marc.info/?l=linux-kernel&m=122036356127282&w=4

This entry is being used for tracking a regression from 2.6.26.  Please don't
close it until the problem is fixed in the mainline.
Comment 1 Andi Kleen 2008-10-05 22:55:17 UTC
Updating description.

The problem is actually worse than described in the original report.

panic() calls smp_call_function() which nows allocates memory and complains
about disabled interrupts. Allocating memory or not allowing panic 
from interrupt off regionss definitely not a good idea.

I ran into this because it broke machine checks panics.
Comment 2 Rafael J. Wysocki 2008-10-24 14:32:39 UTC
Andi, has this problem been fixed in mainline?
Comment 3 Andi Kleen 2008-10-24 23:43:19 UTC
No it's still broken
Comment 4 Rafael J. Wysocki 2008-11-16 10:56:01 UTC
Any change?
Comment 5 Andi Kleen 2008-11-22 15:00:04 UTC
Sorry it's definitely still broken. There have been also reports about this
about users hitting the problem. 
I haven't gotten around to fix it yet, but it's on my todo list.
Comment 6 Alan 2010-01-19 20:09:10 UTC
Ping on status Andi ?
Comment 7 Florian Mickler 2010-10-07 18:40:07 UTC
I guess this is fixed by

commit 4ef702c10b5df18ab04921fc252c26421d4d6c75
Author: Andi Kleen <andi@firstfloor.org>
Date:   Wed May 27 21:56:52 2009 +0200

    x86: fix panic with interrupts off (needed for MCE)
    For some time each panic() called with interrupts disabled
    triggered the !irqs_disabled() WARN_ON in smp_call_function(),
    producing ugly backtraces and confusing users.

Please reopen, if this commit of yours does not fix your issue.

Note You need to log in before you can comment on or make changes to this bug.