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

Groups > linux.kernel > #1740784

Re: [PATCH] HID: usbhid: fix out-of-bounds bug

Show all headers | View raw

On Wed, 27 Sep 2017, Michel Hermier wrote:

> Le 27 sept. 2017 07:42, "Alan Stern" <stern@rowland.harvard.edu> a écrit :

> > -       for (n = 0; n < hdesc->bNumDescriptors; n++)
> > +       num_descriptors = min_t(int, hdesc->bNumDescriptors,
> > +                               (hdesc->bLength - 6) / 3);
> > +       for (n = 0; n < num_descriptors; n++)
> >                 if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
> >                         rsize = le16_to_cpu(hdesc->desc[n].
> wDescriptorLength);
> 
> Yes, this is a lot better.
> 
> 
> Is it possible to explicit the magic number 6 and 3 in the code. Currently,
> it looks like it comes from no where.

Yes, it is possible.  The 6 is equal to

	offsetof(struct hid_descriptor, desc)

and the 3 is equal to

	sizeof(struct hid_class_descriptor)

(at least, I think it is -- the structure is marked as packed so its 
size should be 3).

In this case I found the numbers to be more readable, but other people 
may have different opinions.

> I'm also wondering if this change will not affect some devices in the wild,
> by rejecting hid descriptors with num descriptors == 0 ?

It's possible, but I doubt it.  If such devices do exist, they should
never have worked in the first place.  Certainly they would generate
warnings or errors during enumeration because of their invalid
descriptors.

Alan Stern

Back to linux.kernel | Previous | Next — Next in thread | Find similar | Unroll thread

Thread

Re: [PATCH] HID: usbhid: fix out-of-bounds bug Alan Stern <stern@rowland.harvard.edu> - 2017-09-27 16:30 +0200
  Re: [PATCH] HID: usbhid: fix out-of-bounds bug Jaejoong Kim <climbbb.kim@gmail.com> - 2017-09-28 10:50 +0200
    [PATCH v2] HID: usbhid: fix out-of-bounds bug Jaejoong Kim <climbbb.kim@gmail.com> - 2017-09-28 12:20 +0200

csiph-web