linux-kernel - Re: [PATCH 5.4 17/78] HID: Fix slab-out-of-bounds read in hid_field

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 5 Feb 2020 12:56:55 +0100
From:   peter enderborg <peter.enderborg@...y.com>
To:     Jiri Kosina <jikos@...nel.org>
CC:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Alan Stern <stern@...land.harvard.edu>
Subject: Re: [PATCH 5.4 17/78] HID: Fix slab-out-of-bounds read in
 hid_field_extract (Broken!)

On 2/5/20 10:54 AM, Jiri Kosina wrote:
> On Wed, 5 Feb 2020, Enderborg, Peter wrote:
>
>>>> This patch breaks Elgato StreamDeck.
>>> Does that mean the device is broken with a too-large of a report?
>> Yes.
> In which way does the breakage pop up? Are you getting "report too long" 
> errors in dmesg, or the device just doesn't enumerate at all?
>
> Could you please post /sys/kernel/debug/hid/<device>/rdesc contents, and 
> if the device is at least semi-alive, also contents of 
> /sys/kernel/debug/hid/<device>/events from the time it misbehaves?
>
Sure.

[   18.710500] hid-generic 0003:0FD9:0060.0005: report is too long
[   18.716511] hid-generic 0003:0FD9:0060.0005: item 0 1 0 9 parsing failed
[   18.723359] hid-generic: probe of 0003:0FD9:0060.0005 failed with error -22
[root@imx ~]# cat /sys/kernel/debug/hid/0003\:0FD9\:0060.0005/rdesc
05 0c 09 01 a1 01 09 01 05 09 19 01 29 10 15 00 26 ff 00 75 08 95 10 85 01 81 02 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 a0 81 02 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 a1 81 02 0a 00 ff 15 00 26 ff 00 75 08 96 fe 1f 85 02 91 02 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 03 b1 02 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 04 b1 02 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 05 b1 02 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 01 85 06 b1 02 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 07 b1 04 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 01 85 08 b1 04 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 09 b1 04 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 0a b1 04 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 0b b1 04 c0 c0

The rdesc is different in 5.3.16 and quite long where it works. The head is there:

[root@imx ~]# cat rdesc.5.3.16 | more
05 0c 09 01 a1 01 09 01 05 09 19 01 29 10 15 00 26 ff 00 75 08 95 10 85 01 81 02
 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 a0 81 02 0a 00 ff 15 00 26 ff 00 75 08 9
5 10 85 a1 81 02 0a 00 ff 15 00 26 ff 00 75 08 96 fe 1f 85 02 91 02 a1 00 0a 00
ff 15 00 26 ff 00 75 08 95 10 85 03 b1 02 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08
 95 10 85 04 b1 02 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 05 b1 02 c0 a
1 00 0a 00 ff 15 00 26 ff 00 75 08 95 01 85 06 b1 02 c0 a1 00 0a 00 ff 15 00 26
ff 00 75 08 95 10 85 07 b1 04 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 01 85 08
 b1 04 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08 95 10 85 09 b1 04 c0 a1 00 0a 00 f
f 15 00 26 ff 00 75 08 95 10 85 0a b1 04 c0 a1 00 0a 00 ff 15 00 26 ff 00 75 08
95 10 85 0b b1 04 c0 c0

  INPUT(1)[INPUT]
    Field(0)
      Application(Consumer.0001)
      Usage(17)
        Consumer.0001
        Button.0001
        Button.0002
        Button.0003
        Button.0004
        Button.0005
        Button.0006
        Button.0007
        Button.0008
        Button.0009
        Button.000a
        Button.000b
        Button.000c
        Button.000d
        Button.000e
        Button.000f
        Button.0010
      Logical Minimum(0)
      Logical Maximum(255)
      Report Size(8)
      Report Count(16)
      Report Offset(0)
      Flags( Variable Absolute )
  INPUT(160)[INPUT]
    Field(0)
      Application(Consumer.0001)
      Usage(16)
        Button.ff00
        Button.ff00
        Button.ff00
        Button.ff00

>>> Is it broken in Linus's tree?  If so, can you work with the HID
>>> developers to fix it there so we can backport the fix to all stable
>>> trees?
>> I cant see that there are any other fixes upon this so I dont think so. 
>> I can try.
>>
>>
>> Jiri is in the loop. I guess he is the "HID developers" ?
> Thanks,
>