Providing machine info for reported bugs

Dmitry Vyukov


We've recently added a feature to collect some machine info on syzbot
and provide it with crashes, in the case it is useful for
debugging/understanding what happened.

Here are 2 example crashes, check out "VM info" column for crashes:

And here is a direct link with a sample of machine info:

Currently it includes only /proc/cpuinfo and some info from /sys/module/kvm*

This was a long-standing user feature request, in particular for
debugging KVM bugs (may be highly dependent on HW):

So please be aware of this new feature. Currently it's only shown on
the dashboard, we decided to not pollute emails with it yet.

What other info may be useful for debugging of bugs in your subsystem?
We don't want to dump just every bit of info in all sysfs/debugfs/dev
(can be gigs of text). But if there is some critical bit that you
really need to know for debugging of bugs, we will try to include it.
Basically: if you are receiving an external bug report, what info are
you generally asking for?
As far as I remember +Takashi asked for something in the past, like a
list of sound devices present, but I don't remember exactly what it

I've added kernelci@ because such a list may be useful for other
testing systems as well. Is there a prior art on this? LKML is not
read directly by lots of people, so if you know who may answer this
question for a particular subsystem, please CC.

Thank you