linux-kernel - Re: [tip:x86/irq] x86: Always use irq stacks

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Fri, 3 Sep 2010 17:00:41 +0200
From:	Ingo Molnar <mingo@...e.hu>
To:	mingo@...hat.com, hpa@...or.com, linux-kernel@...r.kernel.org,
	hch@....de, tglx@...utronix.de
Cc:	linux-tip-commits@...r.kernel.org
Subject: Re: [tip:x86/irq] x86: Always use irq stacks


* tip-bot for Christoph Hellwig <hch@....de> wrote:

> Commit-ID:  7974891db234467eaf1fec613ec0129cb4ac2332
> Gitweb:     http://git.kernel.org/tip/7974891db234467eaf1fec613ec0129cb4ac2332
> Author:     Christoph Hellwig <hch@....de>
> AuthorDate: Mon, 28 Jun 2010 14:15:54 +0200
> Committer:  Thomas Gleixner <tglx@...utronix.de>
> CommitDate: Tue, 29 Jun 2010 12:12:59 +0200
> 
> x86: Always use irq stacks

There is another (and different) kind of boot failure on 32-bit x86:

EXT3-fs (sda1): recovery required on readonly filesystem
EXT3-fs (sda1): write access will be enabled during recovery
EXT3-fs: barriers not enabled
BUG: unable to handle kernel NULL pointer dereference at 00000340
IP: [<4102dde1>] print_context_stack+0x81/0x110
*pde = 00000000 
Oops: 0000 [#1] PREEMPT 
last sysfs file: 

Pid: 0, comm: swapper Not tainted 2.6.35-rc3+ #33238 A8N-E/System Product Name
EIP: 0060:[<4102dde1>] EFLAGS: 00010016 CPU: 0
EIP is at print_context_stack+0x81/0x110
EAX: 00000000 EBX: 41f62d70 ECX: 00000000 EDX: 41f62000
ESI: 00000000 EDI: 41c5bc3c EBP: 41f62d54 ESP: 41f62d30
 DS: 007b ES: 007b FS: 0000 GS: 00e0 SS: 0068
Process swapper (pid: 0, ti=41f62000 task=41f64a80 task.ti=41f5c000)
Stack:
 41047a83 41f63ffc 41f62d84 ffffe000 41f62000 41f62d84 41f62d70 41c5bc3c
<0> 41f62000 41f62d84 4102ca48 41c5bc3c 7f9731b0 00000000 41f62d74 41f62d84
<0> 41047a83 00000000 7f9731b0 00000046 7f9731d4 41f62d9c 41033e50 00000000
Call Trace:
 [<41047a83>] ? enqueue_task+0x43/0x60
BUG: unable to handle kernel NULL pointer dereference at 00000340
IP: [<4102dde1>] print_context_stack+0x81/0x110
*pde = 00000000 
Oops: 0000 [#2] PREEMPT 
last sysfs file: 

Which crash i managed to bisect back to this commit:

 | 7974891db234467eaf1fec613ec0129cb4ac2332 is the first bad commit
 | commit 7974891db234467eaf1fec613ec0129cb4ac2332
 | Author: Christoph Hellwig <hch@....de>
 | Date:   Mon Jun 28 14:15:54 2010 +0200
 |
 |    x86: Always use irq stacks

I've attached the config and the full crashlog.

Checking out 7974891db2 and building+booting a kernel with this config 
causes shows the crash.

Checking out 7974891db2~1 and building+booting a kernel with this config 
gives a working system.

Note that tip:x86/irq has these commits currently:

 1813a68: x86: Move alloc_desk_mask variables inside ifdef
 2589737: x86-32: Align IRQ stacks properly
 dcfa726: x86: Remove CONFIG_4KSTACKS
 7974891: x86: Always use irq stacks

Although i didnt expect it to be relevant to the regression, i double 
checked that 1813a68 (i.e. all fixes applied) regresses too - it does.

Thanks,

	Ingo

View attachment "config" of type "text/plain" (65161 bytes)

View attachment "crash.log" of type "text/plain" (195447 bytes)