From: Wu Fengguang <fengguang.wu@intel.com>
To: Avi Kivity <avi@redhat.com>
Cc: "kvm@vger.kernel.org" <kvm@vger.kernel.org>
Subject: Re: [PATCH] Prevent trace call into unloaded module text
Date: Wed, 26 Nov 2008 19:59:06 +0800 [thread overview]
Message-ID: <20081126115906.GA15891@localhost> (raw)
In-Reply-To: <20081126114619.GA15568@localhost>
On Wed, Nov 26, 2008 at 07:46:19PM +0800, Wu Fengguang wrote:
> On Wed, Nov 26, 2008 at 01:17:55PM +0200, Avi Kivity wrote:
> > Wu Fengguang wrote:
> > > Add marker_synchronize_unregister() before module unloading.
> > > This prevents possible trace calls into unloaded module text.
> > >
> > > Signed-off-by: Wu Fengguang <wfg@linux.intel.com>
> > > ---
> > > diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> > > index a87f45e..64f38b3 100644
> > > --- a/virt/kvm/kvm_main.c
> > > +++ b/virt/kvm/kvm_main.c
> > > @@ -2102,5 +2102,6 @@ void kvm_exit(void)
> > > kvm_arch_exit();
> > > kvm_exit_debug();
> > > __free_page(bad_page);
> > > + marker_synchronize_unregister();
> > > }
> > > EXPORT_SYMBOL_GPL(kvm_exit);
> > >
> >
> > What about kvm-intel.ko and kvm-amd.ko? They also contain markers.
>
> vmx_exit and svm_exit() all calls into kvm_exit(), so they have been
> handled in a common way.
>
> > (and, why doesn't module unload do this automatically?)
>
> Maybe most modules don't need it for now?
OK I got a better answer: because marker_synchronize_unregister() must
be called after marker_probe_unregister() calls and the tear down of
any private data the trace functions rely on. So there are no
automatic way.
Below is the updated patch.
Thanks,
Fengguang
---
Prevent trace call into torn down text or data
Add marker_synchronize_unregister() immediately after probe unregisters.
This prevents possible trace calls into unloaded module text, or the
trace functions accessing invalidated data.
Signed-off-by: Wu Fengguang <wfg@linux.intel.com>
---
diff --git a/virt/kvm/kvm_trace.c b/virt/kvm/kvm_trace.c
index 41dcc84..f598744 100644
--- a/virt/kvm/kvm_trace.c
+++ b/virt/kvm/kvm_trace.c
@@ -252,6 +252,7 @@ void kvm_trace_cleanup(void)
struct kvm_trace_probe *p = &kvm_trace_probes[i];
marker_probe_unregister(p->name, p->probe_func, p);
}
+ marker_synchronize_unregister();
relay_close(kt->rchan);
debugfs_remove(kt->lost_file);
next prev parent reply other threads:[~2008-11-26 11:59 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-11-24 14:04 [PATCH] Prevent trace call into unloaded module text Wu Fengguang
2008-11-26 11:17 ` Avi Kivity
2008-11-26 11:46 ` Wu Fengguang
2008-11-26 11:59 ` Wu Fengguang [this message]
2008-11-27 13:32 ` Avi Kivity
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=20081126115906.GA15891@localhost \
--to=fengguang.wu@intel.com \
--cc=avi@redhat.com \
--cc=kvm@vger.kernel.org \
/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