linux-kernel - Re: a question about how to debug this case in ftrace

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: Tue, 25 Jun 2024 09:44:06 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: Andrew Halaney <ahalaney@...hat.com>
Cc: Song Chen <chensong_2000@....cn>, Derek Barbosa <debarbos@...hat.com>,
 pmladek@...e.com, john.ogness@...utronix.de, senozhatsky@...omium.org,
 linux-rt-users@...r.kernel.org, linux-kernel@...r.kernel.org,
 williams@...hat.com, jlelli@...hat.com
Subject: Re: a question about how to debug this case in ftrace

On Tue, 25 Jun 2024 08:20:15 -0500
Andrew Halaney <ahalaney@...hat.com> wrote:

> So I ended up doing a sequence like:
> 
>     tracing_off();
>     ftrace_dump(DUMP_ALL);

I've done that several times too.

> 
> in the softlockup code when it was detected. Ideally I wanted to look at
> the vmcore and look at the ftrace data in there (since debugging printk
> by using printk is a little confusing), but there was a makedumpfile bug
> I hit... so I went with the hacky route to prove to myself what was
> going on. I think since then that's been resolved. Hope that helps!

You may be interested in some work I'm doing that allows you to read
the ring buffer from a previous kernel after a crash.

https://lore.kernel.org/linux-trace-kernel/20240612231934.608252486@goodmis.org/

I also have a way to ask for any memory, that should be able to get the
same location most times, via a "reserve_mem=" kernel command line
parameter.

https://lore.kernel.org/linux-trace-kernel/20240613233415.734483785@goodmis.org/

They are both destined for the next merge window. After that, I have
one more patch that ties the two together, so that you can have a
kernel command line of:

 reserve_mem=12M:4096:trace trace_instance=bootmap@...ce

and then when you boot up, you would have a trace instance that would
be mapped to that memory. If your machine doesn't clear memory after a
crash, you can read the data from the crash on the next boot.

-- Steve