Re: [PULL 09/25] vfio-user: implement message receive infrastructure

[Peter Maydell <peter.maydell@linaro.org>]

On Tue, 28 Oct 2025 at 13:52, John Levon <levon@movementarian.org> wrote:
>
> On Tue, Oct 28, 2025 at 01:35:05PM +0000, Peter Maydell wrote:
>
> > On Thu, 10 Jul 2025 at 13:44, Peter Maydell <peter.maydell@linaro.org> wrote:
> > >
> > > On Thu, 26 Jun 2025 at 08:47, Cédric Le Goater <clg@redhat.com> wrote:
> > > >
> > > > From: John Levon <john.levon@nutanix.com>
> > > >
> > > > Add the basic implementation for receiving vfio-user messages from the
> > > > control socket.
> > > >
> > >
> > > Hi; Coverity suggests there are some issues with this code
> > > (CID 1611807, 1611808, 1611809):
> >
> > Hi; it looks like 1611807 and 1611808 (the resource leaks)
> > are still present in this code in current git; would somebody
> > like to have a look at this?
>
> Please see https://lore.kernel.org/qemu-devel/aG-4hkfLDEpDsqo6@movementarian.org/
>
> I believe them to be false positives.

That email only seems to talk about the locking issue (1611806,
1611809) which we've marked as false-positives in Coverity.
But 1611807 and 1611808 are "failed to free resources" issues:
we allocate memory into reqfds and msg, but in the error
exit path (e.g. if we "goto fatal" because qio_channel_read()
failed) it doesn't look like we ever free that memory.

thanks
-- PMM