qemu_vfio_open_common() initializes s->lock only after passing s to
qemu_vfio_dma_map() via qemu_vfio_init_ramblock().
qemu_vfio_dma_map() tries to lock the uninitialized lock and crashes.
Fix by initializing s->lock first.
RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=1645840
Fixes: 418026ca43bc2626db092d7558258f9594366f28
Cc: qemu-stable@nongnu.org
Signed-off-by: Markus Armbruster <armbru@redhat.com>
---
Compile-tested only, as I lack suitable hardware. Cong Li (cc'ed) is
going to test it for real. It looks obvious enough to me to put it
into -rc3 without waiting for the test results. We can also wait and
put it into -rc4.
util/vfio-helpers.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c
index 1d9272efa4..cccc9cd42e 100644
--- a/util/vfio-helpers.c
+++ b/util/vfio-helpers.c
@@ -411,13 +411,13 @@ static int qemu_vfio_init_ramblock(const char *block_name, void *host_addr,
static void qemu_vfio_open_common(QEMUVFIOState *s)
{
+ qemu_mutex_init(&s->lock);
s->ram_notifier.ram_block_added = qemu_vfio_ram_block_added;
s->ram_notifier.ram_block_removed = qemu_vfio_ram_block_removed;
ram_block_notifier_add(&s->ram_notifier);
s->low_water_mark = QEMU_VFIO_IOVA_MIN;
s->high_water_mark = QEMU_VFIO_IOVA_MAX;
qemu_ram_foreach_block(qemu_vfio_init_ramblock, s);
- qemu_mutex_init(&s->lock);
}
/**
--
2.17.2
On Tue, Nov 27, 2018 at 09:41:43AM +0100, Markus Armbruster wrote: > qemu_vfio_open_common() initializes s->lock only after passing s to > qemu_vfio_dma_map() via qemu_vfio_init_ramblock(). > qemu_vfio_dma_map() tries to lock the uninitialized lock and crashes. > > Fix by initializing s->lock first. > > RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=1645840 > Fixes: 418026ca43bc2626db092d7558258f9594366f28 > Cc: qemu-stable@nongnu.org > Signed-off-by: Markus Armbruster <armbru@redhat.com> > --- > Compile-tested only, as I lack suitable hardware. Cong Li (cc'ed) is > going to test it for real. It looks obvious enough to me to put it > into -rc3 without waiting for the test results. We can also wait and > put it into -rc4. > > util/vfio-helpers.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
----- Original Message ----- > From: "Markus Armbruster" <armbru@redhat.com> > To: qemu-devel@nongnu.org > Cc: coli@redhat.com, fam@euphon.net, stefanha@redhat.com, pbonzini@redhat.com > Sent: Tuesday, November 27, 2018 4:41:43 PM > Subject: [PATCH for-3.1] vfio-helpers: Fix qemu_vfio_open_pci() crash > > qemu_vfio_open_common() initializes s->lock only after passing s to > qemu_vfio_dma_map() via qemu_vfio_init_ramblock(). > qemu_vfio_dma_map() tries to lock the uninitialized lock and crashes. > > Fix by initializing s->lock first. > > RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=1645840 > Fixes: 418026ca43bc2626db092d7558258f9594366f28 > Cc: qemu-stable@nongnu.org > Signed-off-by: Markus Armbruster <armbru@redhat.com> > --- > Compile-tested only, as I lack suitable hardware. Cong Li (cc'ed) is > going to test it for real. It looks obvious enough to me to put it > into -rc3 without waiting for the test results. We can also wait and > put it into -rc4. Smoke test of nvme:// pass with this patch. Ack. > > util/vfio-helpers.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c > index 1d9272efa4..cccc9cd42e 100644 > --- a/util/vfio-helpers.c > +++ b/util/vfio-helpers.c > @@ -411,13 +411,13 @@ static int qemu_vfio_init_ramblock(const char > *block_name, void *host_addr, > > static void qemu_vfio_open_common(QEMUVFIOState *s) > { > + qemu_mutex_init(&s->lock); > s->ram_notifier.ram_block_added = qemu_vfio_ram_block_added; > s->ram_notifier.ram_block_removed = qemu_vfio_ram_block_removed; > ram_block_notifier_add(&s->ram_notifier); > s->low_water_mark = QEMU_VFIO_IOVA_MIN; > s->high_water_mark = QEMU_VFIO_IOVA_MAX; > qemu_ram_foreach_block(qemu_vfio_init_ramblock, s); > - qemu_mutex_init(&s->lock); > } > > /** > -- > 2.17.2 > >
On Tue, 27 Nov 2018 at 08:42, Markus Armbruster <armbru@redhat.com> wrote: > > qemu_vfio_open_common() initializes s->lock only after passing s to > qemu_vfio_dma_map() via qemu_vfio_init_ramblock(). > qemu_vfio_dma_map() tries to lock the uninitialized lock and crashes. > > Fix by initializing s->lock first. > > RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=1645840 > Fixes: 418026ca43bc2626db092d7558258f9594366f28 > Cc: qemu-stable@nongnu.org > Signed-off-by: Markus Armbruster <armbru@redhat.com> > --- > Compile-tested only, as I lack suitable hardware. Cong Li (cc'ed) is > going to test it for real. It looks obvious enough to me to put it > into -rc3 without waiting for the test results. We can also wait and > put it into -rc4. > > util/vfio-helpers.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) Applied, thanks (for rc3). -- PMM
On Tue, 27 Nov 2018 09:41:43 +0100 Markus Armbruster <armbru@redhat.com> wrote: > qemu_vfio_open_common() initializes s->lock only after passing s to > qemu_vfio_dma_map() via qemu_vfio_init_ramblock(). > qemu_vfio_dma_map() tries to lock the uninitialized lock and crashes. > > Fix by initializing s->lock first. > > RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=1645840 > Fixes: 418026ca43bc2626db092d7558258f9594366f28 > Cc: qemu-stable@nongnu.org > Signed-off-by: Markus Armbruster <armbru@redhat.com> > --- > Compile-tested only, as I lack suitable hardware. Cong Li (cc'ed) is > going to test it for real. It looks obvious enough to me to put it > into -rc3 without waiting for the test results. We can also wait and > put it into -rc4. This looks obvious to me as well. > > util/vfio-helpers.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/util/vfio-helpers.c b/util/vfio-helpers.c > index 1d9272efa4..cccc9cd42e 100644 > --- a/util/vfio-helpers.c > +++ b/util/vfio-helpers.c > @@ -411,13 +411,13 @@ static int qemu_vfio_init_ramblock(const char *block_name, void *host_addr, > > static void qemu_vfio_open_common(QEMUVFIOState *s) > { > + qemu_mutex_init(&s->lock); > s->ram_notifier.ram_block_added = qemu_vfio_ram_block_added; > s->ram_notifier.ram_block_removed = qemu_vfio_ram_block_removed; > ram_block_notifier_add(&s->ram_notifier); > s->low_water_mark = QEMU_VFIO_IOVA_MIN; > s->high_water_mark = QEMU_VFIO_IOVA_MAX; > qemu_ram_foreach_block(qemu_vfio_init_ramblock, s); > - qemu_mutex_init(&s->lock); > } > > /** Reviewed-by: Cornelia Huck <cohuck@redhat.com>
© 2016 - 2024 Red Hat, Inc.