linux-kernel - Re: [PATCH v5 3/3] Add BUG_XX() debugging hard/soft lockup detection

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] [day] [month] [year] [list]

Message-ID: <CA+ekxPU8k=Uo79Mo0M0k2o=Cxo=hTJYfnj09RtFFL0kA-2sO+Q@mail.gmail.com>
Date:	Wed, 3 Feb 2016 19:48:56 -0700
From:	Jeffrey Merkey <jeffmerkey@...il.com>
To:	Don Zickus <dzickus@...hat.com>
Cc:	linux-kernel@...r.kernel.org, akpm@...ux-foundation.org,
	atomlin@...hat.com, cmetcalf@...hip.com, fweisbec@...il.com,
	hidehiro.kawai.ez@...achi.com, mhocko@...e.cz, tj@...nel.org,
	uobergfe@...hat.com
Subject: Re: [PATCH v5 3/3] Add BUG_XX() debugging hard/soft lockup detection

>> So, if you disable 'sysctl_hardlockup_all_cpu_backtrace' and enable
>> 'hardlockup_panic', you should be able to achieve what you want, no?
>>
>> But you mentioned you wanted to recover?  Hence avoiding the panic?
>>
>> Cheers,
>> Don
>>
>
> Yes.  Avoiding the panic is what I wanted to do.  These flags work at
> catching the bug as you suggest, however, it does not allow you to
> step out of the hard lockup detector into the offending code, which is
> what a breakpoint does for you here.  We almost need an explicit call
> to enter a debugger here.  I need to go off and rethink this some
> more.
>
> :-)
>
> Jeff
>

Hi Don,

I've decided to maintain this capability in the MDB debugger patch
series I distribute from the MDB Debugger tree on github.   All the
wonderful feedback and discussion has really helped me understand
what's required here and exposed some issues in native Linux for some
folks to think about.  I am not certain this type of change belongs in
native Linux since linux has no in-kernel debugger that provides
disassembly -- it has a device driver that talks to a remote user
space application called GDB and a headless debug dump tool called
kdb.

So unless there is a native in-kernel debugger in Linux such a change
is sort of pointless and from what I have gathered, the x86 folks are
not interested in fixing the entry code and other areas to allow a
full blown kernel debugger unless I just go and patch that myself in
my tree.

This has been very helpful in helping me to understand what types of
changes should be submitted here and are helpful.

Jeff