* [PATCH] tpm: xen-tpmfront: fix use-after-free in backend_changed()
@ 2026-11-19 11:20 Suyuan Wang via B4 Relay
2026-11-19 15:20 ` Jarkko Sakkinen
0 siblings, 1 reply; 7+ messages in thread
From: Suyuan Wang via B4 Relay @ 2026-11-19 11:20 UTC (permalink / raw)
To: jgg, jarkko; +Cc: linux-kernel, jgross, sstabellini, linux-integrity, xen-devel
From: Suyuan Wang <stevewong@hust.edu.cn>
When the backend moves to Closing or Closed, backend_changed()
unregisters the device and then keeps using it:
device_unregister(&dev->dev);
xenbus_frontend_closed(dev);
The xenbus watch callback that calls backend_changed() does not hold a
reference to the device. The tpm_chip registered by xen-tpmfront has dev
as its parent and holds a reference to it, but tpmfront_remove() drops
that reference when it unregisters the chip. So device_unregister() can
release the last reference and free the xenbus_device, and
xenbus_frontend_closed(dev) then accesses freed memory.
Take a temporary reference before unregistering and drop it after
xenbus_frontend_closed(). This keeps dev alive for the last access and
keeps the existing order, in which TPM operations are stopped before
the frontend is marked Closed.
Found with a static analysis tool and checked by hand.
Fixes: e2683957fb26 ("drivers/tpm: add xen tpmfront interface")
Cc: stable@vger.kernel.org
Co-developed-by: Kaixuan Li <kaixuan.li@ntu.edu.sg>
Signed-off-by: Kaixuan Li <kaixuan.li@ntu.edu.sg>
Signed-off-by: Suyuan Wang <stevewong@hust.edu.cn>
---
Compile tested on x86_64 (W=1, no new warnings). Not run with a Xen vTPM
backend.
---
drivers/char/tpm/xen-tpmfront.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/char/tpm/xen-tpmfront.c b/drivers/char/tpm/xen-tpmfront.c
index 3c0cce57c4ae..1be925cca535 100644
--- a/drivers/char/tpm/xen-tpmfront.c
+++ b/drivers/char/tpm/xen-tpmfront.c
@@ -397,8 +397,11 @@ static void backend_changed(struct xenbus_device *dev,
case XenbusStateClosing:
case XenbusStateClosed:
+ /* device_unregister() can drop the last reference to dev */
+ get_device(&dev->dev);
device_unregister(&dev->dev);
xenbus_frontend_closed(dev);
+ put_device(&dev->dev);
break;
default:
break;
^ permalink raw reply related [flat|nested] 7+ messages in thread* Re: [PATCH] tpm: xen-tpmfront: fix use-after-free in backend_changed() 2026-11-19 11:20 [PATCH] tpm: xen-tpmfront: fix use-after-free in backend_changed() Suyuan Wang via B4 Relay @ 2026-11-19 15:20 ` Jarkko Sakkinen 2026-11-19 15:35 ` Jürgen Groß 0 siblings, 1 reply; 7+ messages in thread From: Jarkko Sakkinen @ 2026-11-19 15:20 UTC (permalink / raw) To: shyckless Cc: jgg, linux-kernel, jgross, sstabellini, linux-integrity, xen-devel On Thu, Nov 19, 2026 at 11:20:33AM +0000, Suyuan Wang via B4 Relay wrote: > From: Suyuan Wang <stevewong@hust.edu.cn> > > When the backend moves to Closing or Closed, backend_changed() > unregisters the device and then keeps using it: > > device_unregister(&dev->dev); > xenbus_frontend_closed(dev); > > The xenbus watch callback that calls backend_changed() does not hold a > reference to the device. The tpm_chip registered by xen-tpmfront has dev > as its parent and holds a reference to it, but tpmfront_remove() drops > that reference when it unregisters the chip. So device_unregister() can > release the last reference and free the xenbus_device, and > xenbus_frontend_closed(dev) then accesses freed memory. > > Take a temporary reference before unregistering and drop it after > xenbus_frontend_closed(). This keeps dev alive for the last access and > keeps the existing order, in which TPM operations are stopped before > the frontend is marked Closed. > > Found with a static analysis tool and checked by hand. > > Fixes: e2683957fb26 ("drivers/tpm: add xen tpmfront interface") > Cc: stable@vger.kernel.org > Co-developed-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> > Signed-off-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> > Signed-off-by: Suyuan Wang <stevewong@hust.edu.cn> > --- > Compile tested on x86_64 (W=1, no new warnings). Not run with a Xen vTPM > backend. > --- > drivers/char/tpm/xen-tpmfront.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/drivers/char/tpm/xen-tpmfront.c b/drivers/char/tpm/xen-tpmfront.c > index 3c0cce57c4ae..1be925cca535 100644 > --- a/drivers/char/tpm/xen-tpmfront.c > +++ b/drivers/char/tpm/xen-tpmfront.c > @@ -397,8 +397,11 @@ static void backend_changed(struct xenbus_device *dev, > > case XenbusStateClosing: > case XenbusStateClosed: > + /* device_unregister() can drop the last reference to dev */ > + get_device(&dev->dev); > device_unregister(&dev->dev); > xenbus_frontend_closed(dev); > + put_device(&dev->dev); Not a NAK but makes me wonder why this has not hit the wall ever. I.e. "feeling confused". > break; > default: > break; > > > > Br, Jarkko ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] tpm: xen-tpmfront: fix use-after-free in backend_changed() 2026-11-19 15:20 ` Jarkko Sakkinen @ 2026-11-19 15:35 ` Jürgen Groß 2026-11-19 16:30 ` Jarkko Sakkinen 0 siblings, 1 reply; 7+ messages in thread From: Jürgen Groß @ 2026-11-19 15:35 UTC (permalink / raw) To: Jarkko Sakkinen, shyckless Cc: jgg, linux-kernel, sstabellini, linux-integrity, xen-devel [-- Attachment #1.1: Type: text/plain, Size: 2430 bytes --] On 19.11.26 16:20, Jarkko Sakkinen wrote: > On Thu, Nov 19, 2026 at 11:20:33AM +0000, Suyuan Wang via B4 Relay wrote: >> From: Suyuan Wang <stevewong@hust.edu.cn> >> >> When the backend moves to Closing or Closed, backend_changed() >> unregisters the device and then keeps using it: >> >> device_unregister(&dev->dev); >> xenbus_frontend_closed(dev); >> >> The xenbus watch callback that calls backend_changed() does not hold a >> reference to the device. The tpm_chip registered by xen-tpmfront has dev >> as its parent and holds a reference to it, but tpmfront_remove() drops >> that reference when it unregisters the chip. So device_unregister() can >> release the last reference and free the xenbus_device, and >> xenbus_frontend_closed(dev) then accesses freed memory. >> >> Take a temporary reference before unregistering and drop it after >> xenbus_frontend_closed(). This keeps dev alive for the last access and >> keeps the existing order, in which TPM operations are stopped before >> the frontend is marked Closed. >> >> Found with a static analysis tool and checked by hand. >> >> Fixes: e2683957fb26 ("drivers/tpm: add xen tpmfront interface") >> Cc: stable@vger.kernel.org >> Co-developed-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> >> Signed-off-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> >> Signed-off-by: Suyuan Wang <stevewong@hust.edu.cn> >> --- >> Compile tested on x86_64 (W=1, no new warnings). Not run with a Xen vTPM >> backend. >> --- >> drivers/char/tpm/xen-tpmfront.c | 3 +++ >> 1 file changed, 3 insertions(+) >> >> diff --git a/drivers/char/tpm/xen-tpmfront.c b/drivers/char/tpm/xen-tpmfront.c >> index 3c0cce57c4ae..1be925cca535 100644 >> --- a/drivers/char/tpm/xen-tpmfront.c >> +++ b/drivers/char/tpm/xen-tpmfront.c >> @@ -397,8 +397,11 @@ static void backend_changed(struct xenbus_device *dev, >> >> case XenbusStateClosing: >> case XenbusStateClosed: >> + /* device_unregister() can drop the last reference to dev */ >> + get_device(&dev->dev); >> device_unregister(&dev->dev); >> xenbus_frontend_closed(dev); >> + put_device(&dev->dev); > > Not a NAK but makes me wonder why this has not hit the wall ever. > > I.e. "feeling confused". Is there any vtpm backend other than the stubdom based one? If not, then it is quite easy: this backend never switches state to XenbusStateClosing or XenbusStateClosed. Juergen [-- Attachment #2: OpenPGP digital signature --] [-- Type: application/pgp-signature, Size: 495 bytes --] ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] tpm: xen-tpmfront: fix use-after-free in backend_changed() 2026-11-19 15:35 ` Jürgen Groß @ 2026-11-19 16:30 ` Jarkko Sakkinen 2026-11-19 16:32 ` Jürgen Groß 0 siblings, 1 reply; 7+ messages in thread From: Jarkko Sakkinen @ 2026-11-19 16:30 UTC (permalink / raw) To: Jürgen Groß Cc: shyckless, jgg, linux-kernel, sstabellini, linux-integrity, xen-devel On Thu, Nov 19, 2026 at 04:35:02PM +0100, Jürgen Groß wrote: > On 19.11.26 16:20, Jarkko Sakkinen wrote: > > On Thu, Nov 19, 2026 at 11:20:33AM +0000, Suyuan Wang via B4 Relay wrote: > > > From: Suyuan Wang <stevewong@hust.edu.cn> > > > > > > When the backend moves to Closing or Closed, backend_changed() > > > unregisters the device and then keeps using it: > > > > > > device_unregister(&dev->dev); > > > xenbus_frontend_closed(dev); > > > > > > The xenbus watch callback that calls backend_changed() does not hold a > > > reference to the device. The tpm_chip registered by xen-tpmfront has dev > > > as its parent and holds a reference to it, but tpmfront_remove() drops > > > that reference when it unregisters the chip. So device_unregister() can > > > release the last reference and free the xenbus_device, and > > > xenbus_frontend_closed(dev) then accesses freed memory. > > > > > > Take a temporary reference before unregistering and drop it after > > > xenbus_frontend_closed(). This keeps dev alive for the last access and > > > keeps the existing order, in which TPM operations are stopped before > > > the frontend is marked Closed. > > > > > > Found with a static analysis tool and checked by hand. > > > > > > Fixes: e2683957fb26 ("drivers/tpm: add xen tpmfront interface") > > > Cc: stable@vger.kernel.org > > > Co-developed-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> > > > Signed-off-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> > > > Signed-off-by: Suyuan Wang <stevewong@hust.edu.cn> > > > --- > > > Compile tested on x86_64 (W=1, no new warnings). Not run with a Xen vTPM > > > backend. > > > --- > > > drivers/char/tpm/xen-tpmfront.c | 3 +++ > > > 1 file changed, 3 insertions(+) > > > > > > diff --git a/drivers/char/tpm/xen-tpmfront.c b/drivers/char/tpm/xen-tpmfront.c > > > index 3c0cce57c4ae..1be925cca535 100644 > > > --- a/drivers/char/tpm/xen-tpmfront.c > > > +++ b/drivers/char/tpm/xen-tpmfront.c > > > @@ -397,8 +397,11 @@ static void backend_changed(struct xenbus_device *dev, > > > case XenbusStateClosing: > > > case XenbusStateClosed: > > > + /* device_unregister() can drop the last reference to dev */ > > > + get_device(&dev->dev); > > > device_unregister(&dev->dev); > > > xenbus_frontend_closed(dev); > > > + put_device(&dev->dev); > > > > Not a NAK but makes me wonder why this has not hit the wall ever. > > > > I.e. "feeling confused". > Is there any vtpm backend other than the stubdom based one? > > If not, then it is quite easy: this backend never switches state to > XenbusStateClosing or XenbusStateClosed. > > > Juergen Should it be removed use the default-fallback? Br, Jarkko ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] tpm: xen-tpmfront: fix use-after-free in backend_changed() 2026-11-19 16:30 ` Jarkko Sakkinen @ 2026-11-19 16:32 ` Jürgen Groß 2026-11-19 19:31 ` Wong Steve 0 siblings, 1 reply; 7+ messages in thread From: Jürgen Groß @ 2026-11-19 16:32 UTC (permalink / raw) To: Jarkko Sakkinen Cc: shyckless, jgg, linux-kernel, sstabellini, linux-integrity, xen-devel [-- Attachment #1.1: Type: text/plain, Size: 2858 bytes --] On 19.11.26 17:30, Jarkko Sakkinen wrote: > On Thu, Nov 19, 2026 at 04:35:02PM +0100, Jürgen Groß wrote: >> On 19.11.26 16:20, Jarkko Sakkinen wrote: >>> On Thu, Nov 19, 2026 at 11:20:33AM +0000, Suyuan Wang via B4 Relay wrote: >>>> From: Suyuan Wang <stevewong@hust.edu.cn> >>>> >>>> When the backend moves to Closing or Closed, backend_changed() >>>> unregisters the device and then keeps using it: >>>> >>>> device_unregister(&dev->dev); >>>> xenbus_frontend_closed(dev); >>>> >>>> The xenbus watch callback that calls backend_changed() does not hold a >>>> reference to the device. The tpm_chip registered by xen-tpmfront has dev >>>> as its parent and holds a reference to it, but tpmfront_remove() drops >>>> that reference when it unregisters the chip. So device_unregister() can >>>> release the last reference and free the xenbus_device, and >>>> xenbus_frontend_closed(dev) then accesses freed memory. >>>> >>>> Take a temporary reference before unregistering and drop it after >>>> xenbus_frontend_closed(). This keeps dev alive for the last access and >>>> keeps the existing order, in which TPM operations are stopped before >>>> the frontend is marked Closed. >>>> >>>> Found with a static analysis tool and checked by hand. >>>> >>>> Fixes: e2683957fb26 ("drivers/tpm: add xen tpmfront interface") >>>> Cc: stable@vger.kernel.org >>>> Co-developed-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> >>>> Signed-off-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> >>>> Signed-off-by: Suyuan Wang <stevewong@hust.edu.cn> >>>> --- >>>> Compile tested on x86_64 (W=1, no new warnings). Not run with a Xen vTPM >>>> backend. >>>> --- >>>> drivers/char/tpm/xen-tpmfront.c | 3 +++ >>>> 1 file changed, 3 insertions(+) >>>> >>>> diff --git a/drivers/char/tpm/xen-tpmfront.c b/drivers/char/tpm/xen-tpmfront.c >>>> index 3c0cce57c4ae..1be925cca535 100644 >>>> --- a/drivers/char/tpm/xen-tpmfront.c >>>> +++ b/drivers/char/tpm/xen-tpmfront.c >>>> @@ -397,8 +397,11 @@ static void backend_changed(struct xenbus_device *dev, >>>> case XenbusStateClosing: >>>> case XenbusStateClosed: >>>> + /* device_unregister() can drop the last reference to dev */ >>>> + get_device(&dev->dev); >>>> device_unregister(&dev->dev); >>>> xenbus_frontend_closed(dev); >>>> + put_device(&dev->dev); >>> >>> Not a NAK but makes me wonder why this has not hit the wall ever. >>> >>> I.e. "feeling confused". >> Is there any vtpm backend other than the stubdom based one? >> >> If not, then it is quite easy: this backend never switches state to >> XenbusStateClosing or XenbusStateClosed. >> >> >> Juergen > > Should it be removed use the default-fallback? I don't think so. There might be other backends in future, so following the normal Xen PV-device state machine should be kept. Juergen [-- Attachment #2: OpenPGP digital signature --] [-- Type: application/pgp-signature, Size: 495 bytes --] ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] tpm: xen-tpmfront: fix use-after-free in backend_changed() 2026-11-19 16:32 ` Jürgen Groß @ 2026-11-19 19:31 ` Wong Steve 2026-11-19 20:15 ` Jarkko Sakkinen 0 siblings, 1 reply; 7+ messages in thread From: Wong Steve @ 2026-11-19 19:31 UTC (permalink / raw) To: Jürgen Groß Cc: Jarkko Sakkinen, jgg, linux-kernel, sstabellini, linux-integrity, xen-devel Hi Juergen, Jarkko, Thanks Juergen, that explains why nobody has run into it: with the stubdom backend the frontend never sees Closing or Closed, so this path is only reached with a backend that does close. I agree the normal state machine should stay. So with today's backend this is latent. Jarkko, if you prefer, I can send a v2 that says so in the commit message and drops the Cc: stable. Thanks, Suyuan Jürgen Groß <jgross@suse.com> 于2026年11月20日周五 00:32写道: > > On 19.11.26 17:30, Jarkko Sakkinen wrote: > > On Thu, Nov 19, 2026 at 04:35:02PM +0100, Jürgen Groß wrote: > >> On 19.11.26 16:20, Jarkko Sakkinen wrote: > >>> On Thu, Nov 19, 2026 at 11:20:33AM +0000, Suyuan Wang via B4 Relay wrote: > >>>> From: Suyuan Wang <stevewong@hust.edu.cn> > >>>> > >>>> When the backend moves to Closing or Closed, backend_changed() > >>>> unregisters the device and then keeps using it: > >>>> > >>>> device_unregister(&dev->dev); > >>>> xenbus_frontend_closed(dev); > >>>> > >>>> The xenbus watch callback that calls backend_changed() does not hold a > >>>> reference to the device. The tpm_chip registered by xen-tpmfront has dev > >>>> as its parent and holds a reference to it, but tpmfront_remove() drops > >>>> that reference when it unregisters the chip. So device_unregister() can > >>>> release the last reference and free the xenbus_device, and > >>>> xenbus_frontend_closed(dev) then accesses freed memory. > >>>> > >>>> Take a temporary reference before unregistering and drop it after > >>>> xenbus_frontend_closed(). This keeps dev alive for the last access and > >>>> keeps the existing order, in which TPM operations are stopped before > >>>> the frontend is marked Closed. > >>>> > >>>> Found with a static analysis tool and checked by hand. > >>>> > >>>> Fixes: e2683957fb26 ("drivers/tpm: add xen tpmfront interface") > >>>> Cc: stable@vger.kernel.org > >>>> Co-developed-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> > >>>> Signed-off-by: Kaixuan Li <kaixuan.li@ntu.edu.sg> > >>>> Signed-off-by: Suyuan Wang <stevewong@hust.edu.cn> > >>>> --- > >>>> Compile tested on x86_64 (W=1, no new warnings). Not run with a Xen vTPM > >>>> backend. > >>>> --- > >>>> drivers/char/tpm/xen-tpmfront.c | 3 +++ > >>>> 1 file changed, 3 insertions(+) > >>>> > >>>> diff --git a/drivers/char/tpm/xen-tpmfront.c b/drivers/char/tpm/xen-tpmfront.c > >>>> index 3c0cce57c4ae..1be925cca535 100644 > >>>> --- a/drivers/char/tpm/xen-tpmfront.c > >>>> +++ b/drivers/char/tpm/xen-tpmfront.c > >>>> @@ -397,8 +397,11 @@ static void backend_changed(struct xenbus_device *dev, > >>>> case XenbusStateClosing: > >>>> case XenbusStateClosed: > >>>> + /* device_unregister() can drop the last reference to dev */ > >>>> + get_device(&dev->dev); > >>>> device_unregister(&dev->dev); > >>>> xenbus_frontend_closed(dev); > >>>> + put_device(&dev->dev); > >>> > >>> Not a NAK but makes me wonder why this has not hit the wall ever. > >>> > >>> I.e. "feeling confused". > >> Is there any vtpm backend other than the stubdom based one? > >> > >> If not, then it is quite easy: this backend never switches state to > >> XenbusStateClosing or XenbusStateClosed. > >> > >> > >> Juergen > > > > Should it be removed use the default-fallback? > I don't think so. > > There might be other backends in future, so following the normal Xen PV-device > state machine should be kept. > > > Juergen ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] tpm: xen-tpmfront: fix use-after-free in backend_changed() 2026-11-19 19:31 ` Wong Steve @ 2026-11-19 20:15 ` Jarkko Sakkinen 0 siblings, 0 replies; 7+ messages in thread From: Jarkko Sakkinen @ 2026-11-19 20:15 UTC (permalink / raw) To: Wong Steve Cc: Jürgen Groß, jgg, linux-kernel, sstabellini, linux-integrity, xen-devel On Fri, Nov 20, 2026 at 03:31:53AM +0800, Wong Steve wrote: > Hi Juergen, Jarkko, > > Thanks Juergen, that explains why nobody has run into it: with the > stubdom backend the frontend never sees Closing or Closed, so this > path is only reached with a backend that does close. I agree the > normal state machine should stay. > > So with today's backend this is latent. Jarkko, if you prefer, I can > send a v2 that says so in the commit message and drops the > Cc: stable. > > Thanks, > Suyuan Given the feedback, go ahead. Br, Jarkko ^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2026-11-19 20:15 UTC | newest] Thread overview: 7+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2026-11-19 11:20 [PATCH] tpm: xen-tpmfront: fix use-after-free in backend_changed() Suyuan Wang via B4 Relay 2026-11-19 15:20 ` Jarkko Sakkinen 2026-11-19 15:35 ` Jürgen Groß 2026-11-19 16:30 ` Jarkko Sakkinen 2026-11-19 16:32 ` Jürgen Groß 2026-11-19 19:31 ` Wong Steve 2026-11-19 20:15 ` Jarkko Sakkinen
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox