Re: kern/50491: unkillable wait in usbd_transfer while using usmsc0 on raspberry pi 2

[Michael Plass <mfpnb%plass-family.net@localhost>]

The following reply was made to PR kern/50491; it has been noted by GNATS.

From: Michael Plass <mfpnb%plass-family.net@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: kern/50491: unkillable wait in usbd_transfer while using usmsc0 on raspberry pi 2
Date: Fri, 4 Dec 2015 19:24:21 -0800

 On Dec 4, 2015, at 10:20 AM, Nick Hudson wrote:
 >=20
 > Is 0xba526660 from a new lockup as I don't see it in the list of lwps =
 in=20
 > the PR?
 >=20
 Yes, it was a new instance. But I still have the ps list for that, and =
 it turns out to be:
 
 2193     1 3   3         0           ba526660              mdnsd usbxfer
 
 and we've seen the stack for that one before.
 
 I also had tracing on as before; the last little bit looks like:
 usbd_transfer: xfer=3D0xbacb9d40, flags=3D0x2, pipe=3D0xbafa0780, =
 running=3D1
 usbd_dump_queue: pipe=3D0xbafa0780
   xfer=3D0xbacb98c0
 usb_allocmem: use frag=3D0xc90acd80 size=3D4
 dwc2_device_ctrl_transfer:=20
 usb_insert_transfer: pipe=3D0xbafa0780 running=3D1 timeout=3D5000
 dwc2_device_start: xfer=3D0xbacb98c0 pipe=3D0xbafa0780
 dwctwo0: urb=3D0x81365588, xfer=3D0x0
 usb_schedsoftintr: polling=3D0
 usb_schedsoftintr: polling=3D0
 
 - Michael