Re: Strange issues with UAS URB cancellation

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Can I ask you to add some select traces on top of the patch?
They should show content of endpoint context:

as root:

I did that, however, even when mounting to /sys/kernel/debug/tracing, most of
the pseudo-files were actually at /sys/kernel/tracing and I had to fix the
path of the commands.

Sorry about that,

Ftrace control files used to be part of debugfs, located at
/sys/kernel/debug/tracing when mounting debugfs


           <idle>-0       [005] d.h2.  1809.524165: xhci_handle_command: CMD: Stop Ring Command: slot 9 sp 0 ep 11 flags C
           <idle>-0       [005] dNH2.  1809.524235: xhci_handle_command: CMD: Set TR Dequeue Pointer Command: deq 0000000001219f22 stream 5 slot 9 ep 11 flags C
           <idle>-0       [005] dNH2.  1809.524236: xhci_handle_cmd_set_deq: RS 00021 super-speed plus Ctx Entries 15 MEL 0 us Port# 1/0 [TT Slot 0 Port# 0 TTT 0 Intr 0] Addr 9 State configured
           <idle>-0       [005] dNH2.  1809.524237: xhci_handle_cmd_set_deq_ep: State stopped mult 1 max P. Streams 3 LSA interval 125 us max ESIT payload 0 CErr 3 Type Bulk IN HIDburst 15 maxp 1024 deq 00000000d9933400 avg trb len 0

I can't find anything wrong in the endpoint context after the failed 'Set TR Deq' command.
MaxPStreams is 3, so endpoint supports 16 streams (2^(maxPstreams + 1)), stream ID was 5 is in range.
LSA flag is set so endpoint uses Linear Stream Addressing, which matches the SCT value of 1.

Only reason left why 'Set TR Deq' would fail is if the stream context TR deq pointer
field is zero, see xhci spec 4.12.2.1 'Stream Array Bounds Checking'

xhci driver doesn't yet have tracing for stream contexts, I started writing those but
didn't finish them yet.

Thanks
Mathias




[Index of Archives]     [Linux Media]     [Linux Input]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Old Linux USB Devel Archive]

  Powered by Linux