Re: [PATCH] Move VMEnter and VMExit tracepoints closer to the actual event

From: Dario Faggioli
Date: Fri May 28 2021 - 13:03:59 EST

On Tue, 2021-05-25 at 16:20 +0000, Sean Christopherson wrote:
> On Fri, May 21, 2021, Dario Faggioli wrote:
> > >
> > Indeed. So, do you happen to have in mind what could be the best
> > place
> > and the best way for documenting this?
> I didn't have anything in mind, but my gut reaction is to add a new
> file dedicated
> to tracing/tracepoints in KVM, e.g.
>   Documentation/virt/kvm/tracepoints.rst or
> Documentation/virt/kvm/tracing.rst
Ok. Well, FWIW, this seems a good idea to me. :-)

> I'm sure there are all sorts of tips and tricks people have for using
> KVM's
> tracepoints, it would be nice to provide a way to capture and
> disseminate them.
> My only hesitation is that Documentation/virt/kvm/ might be too
> formal for what
> would effectively be a wiki of sorts.
Yeah, understand the concerns, I think. However, it seems to me that
how to interpret the kernel KVM tracepoint (i.e., this fact that they
mark the rather the beginning of the "logical" entry and exit sequences
rather than the actual instructions) does belong in the kernel's own
documentation, i.e., where you proposed above.

Surely when we'll have something like that, it seems natural that we'd
want to have more stuff there, and we'll have to judge what's best
suited for it and what should perhaps be somewhere else... But I think
it's worth a try, and I probably will try to put something together.

Thanks and Regards
Dario Faggioli, Ph.D
Virtualization Software Engineer
<<This happens because _I_ choose it to happen!>> (Raistlin Majere)

Attachment: signature.asc
Description: This is a digitally signed message part