From: Alex Williamson <alex.williamson@redhat.com>
To: "Cédric Le Goater" <clg@redhat.com>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/2] vfio/mtty: Fix eventfd leak
Date: Mon, 16 Oct 2023 14:16:35 -0600 [thread overview]
Message-ID: <20231016141635.74f8908e.alex.williamson@redhat.com> (raw)
In-Reply-To: <04d9af1d-e459-4431-bea3-679ade88f7d5@redhat.com>
On Mon, 16 Oct 2023 09:52:41 +0200
Cédric Le Goater <clg@redhat.com> wrote:
> On 10/13/23 21:56, Alex Williamson wrote:
> > Found via kmemleak, eventfd context is leaked if not explicitly torn
> > down by userspace. Clear pointers to track released contexts. Also
> > remove unused irq_fd field in mtty structure, set but never used.
>
> This could be 2 different patches, one cleanup and one fix.
Of course.
> > Fixes: 9d1a546c53b4 ("docs: Sample driver to demonstrate how to use Mediated device framework.")
> > Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
> > ---
> > samples/vfio-mdev/mtty.c | 28 +++++++++++++++++++++++-----
> > 1 file changed, 23 insertions(+), 5 deletions(-)
> >
> > diff --git a/samples/vfio-mdev/mtty.c b/samples/vfio-mdev/mtty.c
> > index 5af00387c519..0a2760818e46 100644
> > --- a/samples/vfio-mdev/mtty.c
> > +++ b/samples/vfio-mdev/mtty.c
> > @@ -127,7 +127,6 @@ struct serial_port {
> > /* State of each mdev device */
> > struct mdev_state {
> > struct vfio_device vdev;
> > - int irq_fd;
> > struct eventfd_ctx *intx_evtfd;
> > struct eventfd_ctx *msi_evtfd;
> > int irq_index;
> > @@ -938,8 +937,10 @@ static int mtty_set_irqs(struct mdev_state *mdev_state, uint32_t flags,
> > {
> > if (flags & VFIO_IRQ_SET_DATA_NONE) {
> > pr_info("%s: disable INTx\n", __func__);
> > - if (mdev_state->intx_evtfd)
> > + if (mdev_state->intx_evtfd) {
> > eventfd_ctx_put(mdev_state->intx_evtfd);
> > + mdev_state->intx_evtfd = NULL;
> > + }
> > break;
> > }
> >
> > @@ -955,7 +956,6 @@ static int mtty_set_irqs(struct mdev_state *mdev_state, uint32_t flags,
> > break;
> > }
>
> Shouln't mdev_state->intx_evtfd value be tested before calling
> eventfd_ctx() ?
The state of mtty interrupt handling is really quite atrocious, it's a
pretty significant overhaul to really make it comply with the SET_IRQS
ioctl. I'll see what I can do, but it's so broken that I hope you
won't insist on splitting out each fix. Thanks,
Alex
> > mdev_state->intx_evtfd = evt;
> > - mdev_state->irq_fd = fd;
> > mdev_state->irq_index = index;
> > break;
> > }
> > @@ -971,8 +971,10 @@ static int mtty_set_irqs(struct mdev_state *mdev_state, uint32_t flags,
> > break;
> > case VFIO_IRQ_SET_ACTION_TRIGGER:
> > if (flags & VFIO_IRQ_SET_DATA_NONE) {
> > - if (mdev_state->msi_evtfd)
> > + if (mdev_state->msi_evtfd) {
> > eventfd_ctx_put(mdev_state->msi_evtfd);
> > + mdev_state->msi_evtfd = NULL;
> > + }
> > pr_info("%s: disable MSI\n", __func__);
> > mdev_state->irq_index = VFIO_PCI_INTX_IRQ_INDEX;
> > break;
> > @@ -993,7 +995,6 @@ static int mtty_set_irqs(struct mdev_state *mdev_state, uint32_t flags,
> > break;
> > }
> > mdev_state->msi_evtfd = evt;
> > - mdev_state->irq_fd = fd;
> > mdev_state->irq_index = index;
> > }
> > break;
> > @@ -1262,6 +1263,22 @@ static unsigned int mtty_get_available(struct mdev_type *mtype)
> > return atomic_read(&mdev_avail_ports) / type->nr_ports;
> > }
> >
> > +static void mtty_close(struct vfio_device *vdev)
> > +{
> > + struct mdev_state *mdev_state =
> > + container_of(vdev, struct mdev_state, vdev);
> > +
> > + if (mdev_state->intx_evtfd) {
> > + eventfd_ctx_put(mdev_state->intx_evtfd);
> > + mdev_state->intx_evtfd = NULL;
> > + }
> > + if (mdev_state->msi_evtfd) {
> > + eventfd_ctx_put(mdev_state->msi_evtfd);
> > + mdev_state->msi_evtfd = NULL;
> > + }
> > + mdev_state->irq_index = -1;
> > +}
> > +
> > static const struct vfio_device_ops mtty_dev_ops = {
> > .name = "vfio-mtty",
> > .init = mtty_init_dev,
> > @@ -1273,6 +1290,7 @@ static const struct vfio_device_ops mtty_dev_ops = {
> > .unbind_iommufd = vfio_iommufd_emulated_unbind,
> > .attach_ioas = vfio_iommufd_emulated_attach_ioas,
> > .detach_ioas = vfio_iommufd_emulated_detach_ioas,
> > + .close_device = mtty_close,
> > };
> >
> > static struct mdev_driver mtty_driver = {
>
next prev parent reply other threads:[~2023-10-16 20:17 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-13 19:56 [PATCH 0/2] vfio/mtty: Add migration support Alex Williamson
2023-10-13 19:56 ` [PATCH 1/2] vfio/mtty: Fix eventfd leak Alex Williamson
2023-10-16 7:52 ` Cédric Le Goater
2023-10-16 20:16 ` Alex Williamson [this message]
2023-10-13 19:56 ` [PATCH 2/2] vfio/mtty: Enable migration support Alex Williamson
2023-10-16 14:29 ` Cédric Le Goater
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=20231016141635.74f8908e.alex.williamson@redhat.com \
--to=alex.williamson@redhat.com \
--cc=clg@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@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
all inboxes | Powered by JetHome®