Re: [PATCH v1 0/2] usb: gadget: uvc: stability fixes when stopping streams

[Avichal Rakesh <arakesh@google.com>]

On 9/15/23 18:35, Avichal Rakesh wrote:
> 
> 
> On 9/15/23 16:16, Michael Grzeschik wrote:
>> Hi Avichal
>>
>> On Thu, Sep 14, 2023 at 04:05:36PM -0700, Avichal Rakesh wrote:
>>> On Mon, Sep 11, 2023 at 9:19 PM Avichal Rakesh <arakesh@google.com> wrote:
>>>>
>>>> We have been seeing two main bugs when stopping stream:
>>>> 1. attempting to queue usb_requests on a disabled usb endpoint, and
>>>> 2. use-after-free problems for inflight requests
>>>>
>>>> Avichal Rakesh (2):
>>>>   usb: gadget: uvc: prevent use of disabled endpoint
>>>>   usb: gadget: uvc: prevent de-allocating inflight usb_requests
>>>>
>>>>  drivers/usb/gadget/function/f_uvc.c     | 11 ++++----
>>>>  drivers/usb/gadget/function/f_uvc.h     |  2 +-
>>>>  drivers/usb/gadget/function/uvc.h       |  5 +++-
>>>>  drivers/usb/gadget/function/uvc_v4l2.c  | 21 ++++++++++++---
>>>>  drivers/usb/gadget/function/uvc_video.c | 34 +++++++++++++++++++++++--
>>>>  5 files changed, 60 insertions(+), 13 deletions(-)
>>>>
>>>
>>> Bumping this thread up. Laurent, Dan, and Michael could you take a look?
>>
>> I tested the patches against my setup and it did not help.
> 
> Thank you for testing the patch, I really appreciate your help in testing the 
> patches!
> 
>>
>> In fact I saw two different issues when calling the streamoff event.
>>
>> One issue was a stalled pipeline after the streamoff from the host came in.
>> The streaming application did not handle any events anymore.
> 
> This sounds like uvc_function_setup_continue was never called, so no more control
> events were queued for the userspace to handle. This is a little bit of a shot in
> the dark: if you are not using the Laurent's uvc-gadget
> (https://git.ideasonboard.org/uvc-gadget.git) on the gadget, could you check that 
> your userspace application (on the gadget side) calls VIDIOC_STREAMOFF when 
> handling UVC_EVENT_STREAMOFF? 
> 
> This is similar to how it should called VIDIOC_STREAMON when handling 
> UVC_EVENT_STREAMON. Before my patch, I think UVC gadget driver is functionally fine
> with userspace application not calling VIDIOC_STREAMOFF. However, my patch prevents
> the host from making any more control requests until the gadget's userspace
> application calls VIDIOC_STREAMOFF, which looks like a stalled control ep.
> 
>>
>> The second issue was when the streamoff event is triggered sometimes the
>> following trace is shown, even with your patches applied.
>>
>>
>> [  104.202689] Unable to handle kernel paging request at virtual address 005bf43a692a5fd5
>> [  104.235122] Mem abort info:
>> [  104.238257]   ESR = 0x0000000096000004
>> [  104.242449]   EC = 0x25: DABT (current EL), IL = 32 bits
>> [  104.248391]   SET = 0, FnV = 0
>> [  104.251803]   EA = 0, S1PTW = 0
>> [  104.255313]   FSC = 0x04: level 0 translation fault
>> [  104.260765] Data abort info:
>> [  104.263982]   ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000
>> [  104.270114]   CM = 0, WnR = 0, TnD = 0, TagAccess = 0
>> [  104.275760]   GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
>> [  104.281698] [005bf43a692a5fd5] address between user and kernel address ranges
>> [  104.290042] Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP
>> [  104.297060] Dumping ftrace buffer:
>> [  104.300869]    (ftrace buffer empty)
>> [  104.304862] Modules linked in: st1232 hantro_vpu v4l2_vp9 v4l2_h264 uio_pdrv_genirq fuse [last unloaded: rockchip_vpu(C)]
>> [  104.312080] panfrost fde60000.gpu: Panfrost Dump: BO has no sgt, cannot dump
>> [  104.317137] CPU: 0 PID: 465 Comm: irq/46-dwc3 Tainted: G         C         6.5.0-20230831-2+ #5
>> [  104.317144] Hardware name: WolfVision PF5 (DT)
>> [  104.317148] pstate: 604000c9 (nZCv daIF +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
>> [  104.317154] pc : __list_del_entry_valid+0x48/0xe8
>> [  104.352728] lr : dwc3_gadget_giveback+0x3c/0x1b0
>> [  104.357893] sp : ffffffc08381bc60
>> [  104.361593] x29: ffffffc08381bc60 x28: ffffff80047d4000 x27: ffffff80047de440
>> [  104.369576] x26: 0000000000000000 x25: ffffffc08135b2d0 x24: ffffffc08381bd00
>> [  104.377559] x23: 00000000ffffff98 x22: ffffff8004204880 x21: ffffff80047d4000
>> [  104.385541] x20: ffffff800718dea0 x19: ffffff800718dea0 x18: 0000000000000000
>> [  104.393523] x17: 7461747320687469 x16: 7720646574656c70 x15: 6d6f632074736575
>> [  104.401504] x14: 716572205356203a x13: 2e3430312d207375 x12: 7461747320687469
>> [  104.409486] x11: ffffffc0815c98f0 x10: 0000000000000000 x9 : ffffffc0808f4fa0
>> [  104.417468] x8 : ffffffc082415000 x7 : ffffffc0808f4e2c x6 : ffffffc0823d0928
>> [  104.425450] x5 : 0000000000000282 x4 : 0000000000000201 x3 : d85bf43a692a5fcd
>> [  104.433431] x2 : ffffff80047d4048 x1 : ffffff800718dea0 x0 : dead000000000122
>> [  104.441413] Call trace:
>> [  104.444142]  __list_del_entry_valid+0x48/0xe8
>> [  104.449013]  dwc3_gadget_giveback+0x3c/0x1b0
>> [  104.453786]  dwc3_gadget_ep_cleanup_cancelled_requests+0xe0/0x170
>> [  104.460599]  dwc3_process_event_buf+0x2a8/0xbb0
>> [  104.465662]  dwc3_thread_interrupt+0x4c/0x90
>> [  104.470435]  irq_thread_fn+0x34/0xb8
>> [  104.474431]  irq_thread+0x1a0/0x290
>> [  104.478327]  kthread+0x10c/0x120
>> [  104.481933]  ret_from_fork+0x10/0x20
>>
>> The error path triggering these list errors are usually in the
>> dwc3 driver handling the cancelled or completed list.
> 
> It looks like we're still freeing un-returned requests :(. If you still have
> the setup can you pull the uvc logs to see if waiting for requests to be returned timed 
> out? I wonder if dwc3's interrupt handler is being scheduled too late. 500ms seemed
> like a reasonable time out to me, but this seems to prove otherwise.
> 


Hey Michael, were you able to look into the comments from the previous
email? 

Avi.