Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1599293 > unrolled thread

[PATCH 0/7] Input: fix NULL-derefs at probe

Back to article view | Back to linux.kernel

Contents

  [PATCH 0/7] Input: fix NULL-derefs at probe Johan Hovold <johan@kernel.org> - 2017-03-13 13:40 +0100
    [PATCH 7/7] Input: sur40 - fix NULL-deref at probe Johan Hovold <johan@kernel.org> - 2017-03-13 13:40 +0100
    Re: [PATCH 0/7] Input: fix NULL-derefs at probe Oliver Neukum <oneukum@suse.com> - 2017-03-13 16:20 +0100
      Re: [PATCH 0/7] Input: fix NULL-derefs at probe Johan Hovold <johan@kernel.org> - 2017-03-13 16:50 +0100
        Re: [PATCH 0/7] Input: fix NULL-derefs at probe Dmitry Torokhov <dmitry.torokhov@gmail.com> - 2017-03-16 23:40 +0100
          Re: [PATCH 0/7] Input: fix NULL-derefs at probe Johan Hovold <johan@kernel.org> - 2017-03-17 12:10 +0100
            Re: [PATCH 0/7] Input: fix NULL-derefs at probe Dmitry Torokhov <dmitry.torokhov@gmail.com> - 2017-03-17 22:10 +0100
              Re: [PATCH 0/7] Input: fix NULL-derefs at probe Johan Hovold <johan@kernel.org> - 2017-03-18 10:40 +0100

#1599293 — [PATCH 0/7] Input: fix NULL-derefs at probe

This series fixes a number of NULL-pointer dereferences due to missing
endpoint sanity checks that can be triggered by a malicious USB device.

Johan


Johan Hovold (7):
  Input: iforce - fix NULL-deref at probe
  Input: cm109 - fix NULL-deref at probe
  Input: ims-pcu - fix NULL-deref at probe
  Input: yealink - fix NULL-deref at probe
  Input: hanwang - fix NULL-deref at probe
  Input: kbtab - fix NULL-deref at probe
  Input: sur40 - fix NULL-deref at probe

 drivers/input/joystick/iforce/iforce-usb.c | 3 +++
 drivers/input/misc/cm109.c                 | 4 ++++
 drivers/input/misc/ims-pcu.c               | 4 ++++
 drivers/input/misc/yealink.c               | 4 ++++
 drivers/input/tablet/hanwang.c             | 3 +++
 drivers/input/tablet/kbtab.c               | 3 +++
 drivers/input/touchscreen/sur40.c          | 3 +++
 7 files changed, 24 insertions(+)

-- 
2.12.0

[toc] | [next] | [standalone]

#1599294 — [PATCH 7/7] Input: sur40 - fix NULL-deref at probe

Make sure to check the number of endpoints to avoid dereferencing a
NULL-pointer or accessing memory that lie beyond the end of the endpoint
array should a malicious device lack the expected endpoints.

Fixes: bdb5c57f209c ("Input: add sur40 driver for Samsung SUR40 (aka MS
Surface 2.0/Pixelsense)")
Cc: stable <stable@vger.kernel.org>	# 3.13
Cc: Florian Echtler <floe@butterbrot.org>
Cc: David Herrmann <dh.herrmann@gmail.com>
Cc: Henrik Rydberg <rydberg@euromail.se>

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 drivers/input/touchscreen/sur40.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/input/touchscreen/sur40.c b/drivers/input/touchscreen/sur40.c
index aefb6e11f88a..4c0eecae065c 100644
--- a/drivers/input/touchscreen/sur40.c
+++ b/drivers/input/touchscreen/sur40.c
@@ -527,6 +527,9 @@ static int sur40_probe(struct usb_interface *interface,
 	if (iface_desc->desc.bInterfaceClass != 0xFF)
 		return -ENODEV;
 
+	if (iface_desc->desc.bNumEndpoints < 5)
+		return -ENODEV;
+
 	/* Use endpoint #4 (0x86). */
 	endpoint = &iface_desc->endpoint[4].desc;
 	if (endpoint->bEndpointAddress != TOUCH_ENDPOINT)
-- 
2.12.0

[toc] | [prev] | [next] | [standalone]

#1599501

Am Montag, den 13.03.2017, 13:35 +0100 schrieb Johan Hovold:
> This series fixes a number of NULL-pointer dereferences due to
> missing
> endpoint sanity checks that can be triggered by a malicious USB
> device.
> 
At the risk of repeating myself, doesn't the sheer number of fixes
demonstrate the need for a more centralized check?

	Regards
		Oliver

[toc] | [prev] | [next] | [standalone]

#1599540

On Mon, Mar 13, 2017 at 04:15:18PM +0100, Oliver Neukum wrote:
> Am Montag, den 13.03.2017, 13:35 +0100 schrieb Johan Hovold:
> > This series fixes a number of NULL-pointer dereferences due to
> > missing
> > endpoint sanity checks that can be triggered by a malicious USB
> > device.
>
> At the risk of repeating myself, doesn't the sheer number of fixes
> demonstrate the need for a more centralized check?

No, I don't think that follows. These are plain bugs that needs to be
fixed (cf. not checking for allocation failures or whatever) and
backported to the stable trees.

I think I may have surveyed just about every USB driver this last week,
and there is no single pattern for how endpoints are verified and
retrieved that could easily be refactored into USB core.

Now there are certain patterns that could benefit from a few helpers,
and some obvious bugs could then be caught by declaring those helpers as
__must_check. But specifically, you'd still be checking the return
value from the helpers.

Then verifying the endpoint counts before calling driver probe,
typically only saves a bit of time while probing *malicious* devices
(and the occasional odd interface which cannot be matched on other
attributes).

That being said, we could still add a centralised sanity check for a
large class of drivers (e.g. that do not use altsettings and only need
minimum constraints) but it's not going to obviate the need for careful
driver implementations.

I'll be posting some more patches related to this shortly.

Thanks,
Johan

[toc] | [prev] | [next] | [standalone]

#1602892

On Mon, Mar 13, 2017 at 04:45:52PM +0100, Johan Hovold wrote:
> On Mon, Mar 13, 2017 at 04:15:18PM +0100, Oliver Neukum wrote:
> > Am Montag, den 13.03.2017, 13:35 +0100 schrieb Johan Hovold:
> > > This series fixes a number of NULL-pointer dereferences due to
> > > missing
> > > endpoint sanity checks that can be triggered by a malicious USB
> > > device.
> >
> > At the risk of repeating myself, doesn't the sheer number of fixes
> > demonstrate the need for a more centralized check?
> 
> No, I don't think that follows. These are plain bugs that needs to be
> fixed (cf. not checking for allocation failures or whatever) and
> backported to the stable trees.
> 
> I think I may have surveyed just about every USB driver this last week,
> and there is no single pattern for how endpoints are verified and
> retrieved that could easily be refactored into USB core.
> 
> Now there are certain patterns that could benefit from a few helpers,
> and some obvious bugs could then be caught by declaring those helpers as
> __must_check. But specifically, you'd still be checking the return
> value from the helpers.
> 
> Then verifying the endpoint counts before calling driver probe,
> typically only saves a bit of time while probing *malicious* devices
> (and the occasional odd interface which cannot be matched on other
> attributes).
> 
> That being said, we could still add a centralised sanity check for a
> large class of drivers (e.g. that do not use altsettings and only need
> minimum constraints) but it's not going to obviate the need for careful
> driver implementations.
> 
> I'll be posting some more patches related to this shortly.

There were some discussions about making and that would allow drivers
declare endpoints they want and have USB core ether fill them or not
even try to bind, but nothing concrete.

Anyway, I do not think we should be blocking this patch series on it; if
we come up with something clever we can always switch over.

Applied the lot.

Thanks.

-- 
Dmitry

[toc] | [prev] | [next] | [standalone]

#1603198

On Thu, Mar 16, 2017 at 03:37:28PM -0700, Dmitry Torokhov wrote:
> On Mon, Mar 13, 2017 at 04:45:52PM +0100, Johan Hovold wrote:
> > On Mon, Mar 13, 2017 at 04:15:18PM +0100, Oliver Neukum wrote:
> > > Am Montag, den 13.03.2017, 13:35 +0100 schrieb Johan Hovold:
> > > > This series fixes a number of NULL-pointer dereferences due to
> > > > missing endpoint sanity checks that can be triggered by a
> > > > malicious USB device.

> Applied the lot.

I noticed you dropped the Fixes tag from the patches that fix bugs which
predate git. While this is probably not much of an issue in this case, I
think it's generally a bad idea since we're loosing information this
way, and this specifically makes it harder for the stable maintainers to
figure out which tree to backport a fix to.

Thanks,
Johan

[toc] | [prev] | [next] | [standalone]

#1603620

On Fri, Mar 17, 2017 at 11:53:37AM +0100, Johan Hovold wrote:
> On Thu, Mar 16, 2017 at 03:37:28PM -0700, Dmitry Torokhov wrote:
> > On Mon, Mar 13, 2017 at 04:45:52PM +0100, Johan Hovold wrote:
> > > On Mon, Mar 13, 2017 at 04:15:18PM +0100, Oliver Neukum wrote:
> > > > Am Montag, den 13.03.2017, 13:35 +0100 schrieb Johan Hovold:
> > > > > This series fixes a number of NULL-pointer dereferences due to
> > > > > missing endpoint sanity checks that can be triggered by a
> > > > > malicious USB device.
> 
> > Applied the lot.
> 
> I noticed you dropped the Fixes tag from the patches that fix bugs which
> predate git. While this is probably not much of an issue in this case, I
> think it's generally a bad idea since we're loosing information this
> way, and this specifically makes it harder for the stable maintainers to
> figure out which tree to backport a fix to.

As far as I know the rule is: if no special markings then stable patch
should be applied as far as it can go.

There is no reason to say specify 2.6.12 commit, as in fact the
offending change is likely to be even earlier, so the annotation would
be effectively wrong.

Thanks.

-- 
Dmitry

[toc] | [prev] | [next] | [standalone]

#1603750

On Fri, Mar 17, 2017 at 02:03:15PM -0700, Dmitry Torokhov wrote:
> On Fri, Mar 17, 2017 at 11:53:37AM +0100, Johan Hovold wrote:
> > On Thu, Mar 16, 2017 at 03:37:28PM -0700, Dmitry Torokhov wrote:
> > > On Mon, Mar 13, 2017 at 04:45:52PM +0100, Johan Hovold wrote:
> > > > On Mon, Mar 13, 2017 at 04:15:18PM +0100, Oliver Neukum wrote:
> > > > > Am Montag, den 13.03.2017, 13:35 +0100 schrieb Johan Hovold:
> > > > > > This series fixes a number of NULL-pointer dereferences due to
> > > > > > missing endpoint sanity checks that can be triggered by a
> > > > > > malicious USB device.
> > 
> > > Applied the lot.
> > 
> > I noticed you dropped the Fixes tag from the patches that fix bugs which
> > predate git. While this is probably not much of an issue in this case, I
> > think it's generally a bad idea since we're loosing information this
> > way, and this specifically makes it harder for the stable maintainers to
> > figure out which tree to backport a fix to.
> 
> As far as I know the rule is: if no special markings then stable patch
> should be applied as far as it can go.

That's true for the stable tag itself, yes.

> There is no reason to say specify 2.6.12 commit, as in fact the
> offending change is likely to be even earlier, so the annotation would
> be effectively wrong.

It is still the first git commit which has the bug, and everyone
(dealing with code forensics) knows that 1da177e4c3f4
("Linux-2.6.12-rc2") is special.

Adding a Fixes-tag pointing to that initial commit, makes it clear that
bug has indeed been tracked as far back as reasonable. Omission of a
Fixes-tag could on the other hand be due to the submitter not bothering
to track the offending commit, thereby leaving it up to a stable
maintainer to do so (if only just be sure).

Thanks,
Johan

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web