2019-05-04 03:41:41

by Jia-Ju Bai

Subject: [PATCH] usb: host: xhci_debugfs: Fix a null pointer dereference in xhci_debugfs_create_endpoint()

In xhci_debugfs_create_slot(), kzalloc() can fail and
dev->debugfs_private will be NULL.
In xhci_debugfs_create_endpoint(), dev->debugfs_private is used without
any null-pointer check, and can cause a null pointer dereference.

To fix this bug, a null-pointer check is added in
xhci_debugfs_create_endpoint().

This bug is found by a runtime fuzzing tool named FIZZER written by us.

Signed-off-by: Jia-Ju Bai <[email protected]>
---
drivers/usb/host/xhci-debugfs.c | 3 +++
1 file changed, 3 insertions(+)

diff --git a/drivers/usb/host/xhci-debugfs.c b/drivers/usb/host/xhci-debugfs.c
index cadc01336bf8..7ba6afc7ef23 100644
--- a/drivers/usb/host/xhci-debugfs.c
+++ b/drivers/usb/host/xhci-debugfs.c
@@ -440,6 +440,9 @@ void xhci_debugfs_create_endpoint(struct xhci_hcd *xhci,
struct xhci_ep_priv *epriv;
struct xhci_slot_priv *spriv = dev->debugfs_private;

+ if (!spriv)
+ return;
+
if (spriv->eps[ep_index])
return;

--
2.17.0

2019-05-04 06:36:09

by Greg Kroah-Hartman

Subject: Re: [PATCH] usb: host: xhci_debugfs: Fix a null pointer dereference in xhci_debugfs_create_endpoint()

On Sat, May 04, 2019 at 11:37:48AM +0800, Jia-Ju Bai wrote:
> In xhci_debugfs_create_slot(), kzalloc() can fail and
> dev->debugfs_private will be NULL.
> In xhci_debugfs_create_endpoint(), dev->debugfs_private is used without
> any null-pointer check, and can cause a null pointer dereference.
>
> To fix this bug, a null-pointer check is added in
> xhci_debugfs_create_endpoint().
>
> This bug is found by a runtime fuzzing tool named FIZZER written by us.
>
> Signed-off-by: Jia-Ju Bai <[email protected]>

Very rare case, but nice fix. You should put "potential" in your
subject line as this is something that no one should ever hit :)

Anyway:

Reviewed-by: Greg Kroah-Hartman <[email protected]>

2019-05-04 07:32:46

by Jia-Ju Bai

Subject: Re: [PATCH] usb: host: xhci_debugfs: Fix a null pointer dereference in xhci_debugfs_create_endpoint()

On 2019/5/4 14:33, Greg KH wrote:
> On Sat, May 04, 2019 at 11:37:48AM +0800, Jia-Ju Bai wrote:
>> In xhci_debugfs_create_slot(), kzalloc() can fail and
>> dev->debugfs_private will be NULL.
>> In xhci_debugfs_create_endpoint(), dev->debugfs_private is used without
>> any null-pointer check, and can cause a null pointer dereference.
>>
>> To fix this bug, a null-pointer check is added in
>> xhci_debugfs_create_endpoint().
>>
>> This bug is found by a runtime fuzzing tool named FIZZER written by us.
>>
>> Signed-off-by: Jia-Ju Bai <[email protected]>
> Very rare case, but nice fix. You should put "potential" in your
> subject line as this is something that no one should ever hit :)

Okay, Greg, thanks for this advice :)


Best wishes,
Jia-Ju Bai

2019-05-06 11:15:02

by Mathias Nyman

Subject: Re: [PATCH] usb: host: xhci_debugfs: Fix a null pointer dereference in xhci_debugfs_create_endpoint()

On 4.5.2019 10.30, Jia-Ju Bai wrote:
>
>
> On 2019/5/4 14:33, Greg KH wrote:
>> On Sat, May 04, 2019 at 11:37:48AM +0800, Jia-Ju Bai wrote:
>>> In xhci_debugfs_create_slot(), kzalloc() can fail and
>>> dev->debugfs_private will be NULL.
>>> In xhci_debugfs_create_endpoint(), dev->debugfs_private is used without
>>> any null-pointer check, and can cause a null pointer dereference.
>>>
>>> To fix this bug, a null-pointer check is added in
>>> xhci_debugfs_create_endpoint().
>>>
>>> This bug is found by a runtime fuzzing tool named FIZZER written by us.
>>>
>>> Signed-off-by: Jia-Ju Bai <[email protected]>
>> Very rare case, but nice fix.  You should put "potential" in your
>> subject line as this is something that no one should ever hit :)
>
> Okay, Greg, thanks for this advice :)
>

Adding patch to queue, and added "potential" to subject line.

-Mathias