Message-ID: <CAAeHK+x=fJ0gAkEGyzWkEJoWGkv-8KWdq5eOtPOgD1mEo--3ug@mail.gmail.com>
Date:   Wed, 20 Sep 2017 23:28:35 +0200
From:   Andrey Konovalov <andreyknvl@...gle.com>
To:     Mike Isely at pobox <isely@...ox.com>
Cc:     Mauro Carvalho Chehab <mchehab@...nel.org>,
        linux-media@...r.kernel.org, LKML <linux-kernel@...r.kernel.org>,
        Dmitry Vyukov <dvyukov@...gle.com>,
        Kostya Serebryany <kcc@...gle.com>,
        syzkaller <syzkaller@...glegroups.com>
Subject: Re: usb/media/pvrusb2: warning in pvr2_send_request_ex/usb_submit_urb

On Wed, Sep 20, 2017 at 9:33 PM, Mike Isely <isely@...ly.net> wrote:

Hi Mike!

>
> What you have here is way beyond just feeding random crap in via the
> syscall interface.  To cause this you have to fake the presence of a
> pvrusb2 compatible *hardware* USB device and then lie about its endpoint
> configuration.  Is that really a concern here?  Are we now saying that
> any kernel driver which talks via USB must now also specifically verify
> the exact expected USB endpoint configuration?  Where does that end?
> How about the vendor-specific RPC protocol that the hardware actually
> implements over the bulk endpoint?  It's likely that the pvrusb2 driver
> may be making assumptions about the expected responses over that
> protocol.

The main assumption here is that an attacker has physical access to a
USB port on a machine. In such case a 100$ Facedancer21 board [1] or a
5$ Raspberry Pi Zero [2] in device mode can be used to emulate
arbitrary USB devices and exploit bugs in the kernel (to execute
arbitrary code or to leak data). USB device descriptors during
enumeration phase and all subsequently received from the device
packets (including vendor-specific protocols) should be considered
untrusted input and checked accordingly.

>
> Please realize that I'm not dismissing this.  I can see some merit in
> this.  But I'm just a bit surprised that now we're going this far.  Is
> this really the intention?  You're talking about code
> (pvrusb2_send_request_ex()) that hasn't changed in about 10 years.
> With this level of paranoia there's got to be a pretty target-rich
> environment over the set of kernel-supported USB devices.

Yes, the intention is to fuzz Linux kernel USB drivers (and USB core
code) by connecting random malformed USB devices and by sending
garbage during subsequent communication.

The fact that the code hasn't changed doesn't mean that it's not buggy :)

>
> To take this another step, wouldn't that same level of paranoia be a
> concern for any externally connected PCI-Express device?  Because that's
> another external way into the computer that involves very non-trivial
> and very hardware-centric protocols.  Thunderbolt devices would be an
> example of this.

At this point being able to connect a PCI-Express device usually leads
to being able to do a DMA attack. But sure, exploitable bugs in
PCE-Express device drivers would be a viable attack vector for systems
with proper IOMMU support. Same goes for any other hot-pluggable
externally accessible port/protocol.

>
>   -Mike

[1] https://int3.cc/products/facedancer21

[2] https://www.raspberrypi.org/products/raspberry-pi-zero/

Thanks!

>
>
> On Wed, 20 Sep 2017, Andrey Konovalov wrote:
>
>> Hi!
>>
>> I've got the following report while fuzzing the kernel with syzkaller.
>>
>> On commit ebb2c2437d8008d46796902ff390653822af6cc4 (Sep 18).
>>
>> There seems to be no check on endpoint type before submitting bulk urb
>> in pvr2_send_request_ex().
>>
>> usb 1-1: New USB device found, idVendor=2040, idProduct=7500
>> usb 1-1: New USB device strings: Mfr=0, Product=255, SerialNumber=0
>> usb 1-1: Product: a
>> gadgetfs: configuration #6
>> pvrusb2: Hardware description: WinTV HVR-1950 Model 750xx
>> usb 1-1: BOGUS urb xfer, pipe 3 != type 1
>> ------------[ cut here ]------------
>> WARNING: CPU: 1 PID: 2713 at drivers/usb/core/urb.c:449
>> usb_submit_urb+0xf8a/0x11d0
>> Modules linked in:
>> CPU: 1 PID: 2713 Comm: pvrusb2-context Not tainted
>> 4.14.0-rc1-42251-gebb2c2437d80 #210
>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>> task: ffff88006b7a18c0 task.stack: ffff880069978000
>> RIP: 0010:usb_submit_urb+0xf8a/0x11d0 drivers/usb/core/urb.c:448
>> RSP: 0018:ffff88006997f990 EFLAGS: 00010286
>> RAX: 0000000000000029 RBX: ffff880063661900 RCX: 0000000000000000
>> RDX: 0000000000000029 RSI: ffffffff86876d60 RDI: ffffed000d32ff24
>> RBP: ffff88006997fa90 R08: 1ffff1000d32fdca R09: 0000000000000000
>> R10: 0000000000000000 R11: 0000000000000000 R12: 1ffff1000d32ff39
>> R13: 0000000000000001 R14: 0000000000000003 R15: ffff880068bbed68
>> FS:  0000000000000000(0000) GS:ffff88006c600000(0000) knlGS:0000000000000000
>> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> CR2: 0000000001032000 CR3: 000000006a0ff000 CR4: 00000000000006f0
>> Call Trace:
>>  pvr2_send_request_ex+0xa57/0x1d80 drivers/media/usb/pvrusb2/pvrusb2-hdw.c:3645
>>  pvr2_hdw_check_firmware drivers/media/usb/pvrusb2/pvrusb2-hdw.c:1812
>>  pvr2_hdw_setup_low drivers/media/usb/pvrusb2/pvrusb2-hdw.c:2107
>>  pvr2_hdw_setup drivers/media/usb/pvrusb2/pvrusb2-hdw.c:2250
>>  pvr2_hdw_initialize+0x548/0x3c10 drivers/media/usb/pvrusb2/pvrusb2-hdw.c:2327
>>  pvr2_context_check drivers/media/usb/pvrusb2/pvrusb2-context.c:118
>>  pvr2_context_thread_func+0x361/0x8c0
>> drivers/media/usb/pvrusb2/pvrusb2-context.c:167
>>  kthread+0x3a1/0x470 kernel/kthread.c:231
>>  ret_from_fork+0x2a/0x40 arch/x86/entry/entry_64.S:431
>> Code: 48 8b 85 30 ff ff ff 48 8d b8 98 00 00 00 e8 ee 82 89 fe 45 89
>> e8 44 89 f1 4c 89 fa 48 89 c6 48 c7 c7 40 c0 ea 86 e8 30 1b dc fc <0f>
>> ff e9 9b f7 ff ff e8 aa 95 25 fd e9 80 f7 ff ff e8 50 74 f3
>> ---[ end trace 6919030503719da6 ]---
>>
>
> --
>
> Mike Isely
> isely @ isely (dot) net
> PGP: 03 54 43 4D 75 E5 CC 92 71 16 01 E2 B5 F5 C1 E8
>
> --
> You received this message because you are subscribed to the Google Groups "syzkaller" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to syzkaller+unsubscribe@...glegroups.com.
> For more options, visit https://groups.google.com/d/optout.

Powered by blists - more mailing lists