Received-SPF: pass (google.com: domain of linux-kernel+bounces-177005-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Sun, 12 May 2024 20:50:24 +0000
From: "Dr. David Alan Gilbert" <linux@treblig.org>
To: Muni Sekhar <munisekharrms@gmail.com>
Cc: kernelnewbies <kernelnewbies@kernelnewbies.org>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: Inquiry Regarding Handling of Kernel Crashes
Message-ID: <ZkErkG77vLkcVaUZ@gallifrey>
References: <CAHhAz+hcVRU8=SkkoHGRDN5zRvB2HY1-1Y2NQmNwR=07ObSFrQ@mail.gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
In-Reply-To: <CAHhAz+hcVRU8=SkkoHGRDN5zRvB2HY1-1Y2NQmNwR=07ObSFrQ@mail.gmail.com>
User-Agent: Mutt/2.2.12 (2023-09-09)

* Muni Sekhar (munisekharrms@gmail.com) wrote:
> Dear Linux Kernel Community,

Hi,

> I hope this email finds you well. I am currently engaged in testing
> device drivers in Linux kernel mode, and I have encountered various
> types of kernel crashes during my testing process.
> 
> Among these, some examples of kernel crashes include OOPS, lockups and others.
> 
> I have a few questions regarding the handling of kernel crashes during testing:
> 
> When encountering a kernel crash during testing, is it advisable to
> continue testing without rebooting the system? Or is it preferable to
> reboot the system after each kernel crash and then resume testing?

Rebooting is best.

> Can the first kernel crash, whether it is an OOPS,  or any other type
> crash, potentially lead to subsequent crashes of the same or different
> types? If so, should debugging efforts focus only on the first kernel
> crash, or should all subsequent crashes also be considered and
> addressed?

Yes - not all failures do that, but some will cause follow on crashes;
looking at the first crash normally gives the most reliable idea
of what went wrong. But keep all the logs, anything might help you figure
it out.

> In the event that the system needs to be rebooted after a kernel
> crash, how can user space test utilities be informed that a kernel
> crash has occurred? Additionally, how can the system be configured to
> automatically reboot in the event of a kernel crash?

See Documentation/admin-guide/kernel-parameters.txt  there are 
quite a few useful ones, in particular:
     oops=panic   will cause a panic after an oops
which when you combine it with
     panic=30

   means an oops will then cause a panic which causes a reboot.

You could also consider using a 'crash kernel' - on a panic
that lands in a fresh kernel that just saves a memory snapshot
that you can then try and debug.

Turning on a watchdog as well is good; some kernel bugs just hang
rather than giving a nice oops.

> I would greatly appreciate any insights or best practices you can
> share regarding the handling of kernel crashes during testing. Your
> expertise and guidance on this matter would be invaluable to my
> testing efforts.
> 
> Thank you very much for your time and assistance. I look forward to
> your response.

Good luck!

Dave

> 
> 
> -- 
> Thanks,
> Sekhar
> 
-- 
 -----Open up your eyes, open up your mind, open up your code -------   
/ Dr. David Alan Gilbert    |       Running GNU/Linux       | Happy  \ 
\        dave @ treblig.org |                               | In Hex /
 \ _________________________|_____ http://www.treblig.org   |_______/