From: Dario Faggioli <dfaggioli@suse.com> To: Steven Rostedt <rostedt@goodmis.org>, Tzvetomir Stoyanov <tz.stoyanov@gmail.com> Cc: Linux Trace Devel <linux-trace-devel@vger.kernel.org> Subject: Re: [PATCH v2] trace-cmd: Find PID of host-guest task from tracing vsock connection Date: Fri, 07 May 2021 14:48:25 +0200 [thread overview] Message-ID: <098a7c9a4af4c0e6364a29eeacbd74a7505f61c0.camel@suse.com> (raw) In-Reply-To: <20210507075834.43497e77@gandalf.local.home> [-- Attachment #1: Type: text/plain, Size: 2269 bytes --] On Fri, 2021-05-07 at 07:58 -0400, Steven Rostedt wrote: > On Fri, 7 May 2021 07:20:15 +0300 > Tzvetomir Stoyanov <tz.stoyanov@gmail.com> wrote: > > > > > Doesn't a VMExit always happen in the thread that runs the vCPU > > anyway? > > The problem is that we do not know what Tgid we are looking for. If > you > have 10 guests, you have 10 different Tgid's that are running those > guests. > How do you know which guest is attached to the cid you are looking > for? > There's currently no way to find that mapping. > Yes, it is true that this part is missing. I have, like, a thought in the back of my head that this may be doable in some way, but I can't actually come up with a working solution! What prompted my question was the fact that I was not understanding the focus on this side of the issue, when we had no link to the CID anyway. And that because... > In fact, that's why I posted > this patch: > > https://lore.kernel.org/kvm/20210505163855.32dad8e7@gandalf.local.home/ > ... I had missed this patch. :-) > So you trace the kvm exit, and all 10 guests were active at that > time, and > you have 100 threads that called kvm exit. Now how do you find the > pid of > the task that owns the cid you are looking for, especially if each > guest > has their own cid? > Yeah, as said, I see it now. > Oh, and the guest is run by a fork of qemu that has some other name! > So you > can not rely on looking at the thread's cmdline. > Yeah, that probably depends where you look. Plus, we know don't want to rely on QEMU doing things in a specific way. > Again, how else can you map the cid to the guest thread? That's the > missing > link that following the wake ups give you. > Yep, understood. And despite that weird feeling that there may be other/simpler ways, I can't name any, so we should go ahead with this approach I guess. :-) Thanks for taking the time to explain it, and sorry for the noise. :-D Regards -- Dario Faggioli, Ph.D http://about.me/dario.faggioli Virtualization Software Engineer SUSE Labs, SUSE https://www.suse.com/ ------------------------------------------------------------------- <<This happens because _I_ choose it to happen!>> (Raistlin Majere) [-- Attachment #2: This is a digitally signed message part --] [-- Type: application/pgp-signature, Size: 833 bytes --]
next prev parent reply other threads:[~2021-05-07 12:48 UTC|newest] Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-05-06 21:14 Steven Rostedt 2021-05-07 0:32 ` Dario Faggioli 2021-05-07 1:40 ` Steven Rostedt 2021-05-07 4:20 ` Tzvetomir Stoyanov 2021-05-07 11:58 ` Steven Rostedt 2021-05-07 12:48 ` Dario Faggioli [this message] 2021-05-07 13:13 ` Steven Rostedt
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=098a7c9a4af4c0e6364a29eeacbd74a7505f61c0.camel@suse.com \ --to=dfaggioli@suse.com \ --cc=linux-trace-devel@vger.kernel.org \ --cc=rostedt@goodmis.org \ --cc=tz.stoyanov@gmail.com \ --subject='Re: [PATCH v2] trace-cmd: Find PID of host-guest task from tracing vsock connection' \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).