From: Tingmao Wang <m@maowtm.org>
To: Song Liu <song@kernel.org>, Dominique Martinet <asmadeus@codewreck.org>
Cc: v9fs@lists.linux.dev, "Eric Van Hensbergen" <ericvh@kernel.org>,
"Latchesar Ionkov" <lucho@ionkov.net>,
"Christian Schoenebeck" <linux_oss@crudebyte.com>,
"Mickaël Salaün" <mic@digikod.net>, bpf <bpf@vger.kernel.org>
Subject: Re: 9P change breaks bpftrace running in qemu+9p?
Date: Tue, 21 Oct 2025 01:54:11 +0100 [thread overview]
Message-ID: <086bb120-22eb-43ff-a486-14e8eeb7dd80@maowtm.org> (raw)
In-Reply-To: <CAHzjS_uEhozUU-g62AkTfSMW58FphVO8udz8qsGzE33jqVpY+g@mail.gmail.com>
On 10/20/25 22:52, Song Liu wrote:
> Hi Dominique,
>
> On Mon, Oct 20, 2025 at 2:32 PM Dominique Martinet
> <asmadeus@codewreck.org> wrote:
>>
>> Song Liu wrote on Mon, Oct 20, 2025 at 12:40:23PM -0700:
>>> I am running qemu 9.2.0 and bpftrace v0.24.0. I don't think anything is
>>> very special here.
>>
>> I don't reproduce either (qemu 9.2.4 and bpftrace v0.24.1, I even went
>> and installed vmtest to make sure), trying both my branch and a pristine
>> v6.18-rc2 kernel -- what's the exact commit you're testing and could you
>> attach your .config ?
>
> Attached, please find the config file.
>
> I tried to debug this, and found that the issue disappears when I remove
> v9fs_lookup_revalidate from v9fs_dentry_operations. But I couldn't figure
> out why d_revalidate() is causing such an issue.
I've compiled qemu 9.2.0 and download the binary build of bpftrace v0.24.0
from GitHub [1], and compiled kernel with your config, but unfortunately I
still can't reproduce it...
I do now get this message sometimes (probably unrelated?):
bpftrace (148) used greatest stack depth: 11624 bytes left
I don't really know how to proceed right now but I will have it run in a
loop and see if I can hit it by chance.
If you can reproduce it frequently and can debug exactly what is returning
-EIO in v9fs_lookup_revalidate that would probably be very helpful, or if
you can enable 9p debug outputs and see what's happening around the time
of error (CONFIG_NET_9P_DEBUG=y and also debug=5 mount options - I'm not
sure how to get vmtest to use a custom mount option but if it's
reproducible in plain QEMU that's also an option) that might also be
informative I think? I'm happy to take a deeper look (although I'm of
course less of an expert than Dominique so hopefully he can also give some
opinion).
I'm also curious if this can happen with just a usual `stat` or other
operations (not necessarily caused by dentry revalidation, and thus not
necessarily to do with my patch)
[1]: https://github.com/bpftrace/bpftrace/releases/tag/v0.24.0
>
> Thanks,
> Song
next prev parent reply other threads:[~2025-10-21 0:54 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-20 17:40 9P change breaks bpftrace running in qemu+9p? Song Liu
2025-10-20 18:40 ` Tingmao Wang
2025-10-20 19:40 ` Song Liu
2025-10-20 21:32 ` Dominique Martinet
2025-10-20 21:52 ` Song Liu
2025-10-21 0:54 ` Tingmao Wang [this message]
2025-10-21 6:49 ` Song Liu
2025-10-21 16:56 ` Alexei Starovoitov
2025-10-21 22:12 ` Dominique Martinet
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=086bb120-22eb-43ff-a486-14e8eeb7dd80@maowtm.org \
--to=m@maowtm.org \
--cc=asmadeus@codewreck.org \
--cc=bpf@vger.kernel.org \
--cc=ericvh@kernel.org \
--cc=linux_oss@crudebyte.com \
--cc=lucho@ionkov.net \
--cc=mic@digikod.net \
--cc=song@kernel.org \
--cc=v9fs@lists.linux.dev \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox