Subject: kern/24276: uhci halts on apm suspend
To: None <gnats-bugs@gnats.netbsd.org>
From: None <ljocha@ics.muni.cz>
List: netbsd-bugs
Date: 01/30/2004 14:23:26
>Number:         24276
>Category:       kern
>Synopsis:       uhci host controller halts on amp suspend
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Jan 30 13:31:00 UTC 2004
>Closed-Date:
>Last-Modified:
>Originator:     Ales Krenek
>Release:        NetBSD 1.6.1_STABLE
>Organization:
Masaryk University
>Environment:
System: NetBSD balkis.ics.muni.cz 1.6.1_STABLE NetBSD 1.6.1_STABLE (BALKIS) #3: Fri Jan 23 12:17:58 CET 2004 ljocha@hiram.ics.muni.cz:/usr/src/sys-1.6/arch/i386/compile/BALKIS i386
Architecture: i386
Machine: i386
>Description:
The machine is Dell Latitude C600, bios revision A22.
options PCIBIOS,PCIBIOS_BUS_FIXUP,PCIBIOS_INTR_FIXUP are set in the kernel
(its necessary due to cardbus).

When it wakes up from APM suspend to memory, I always get 

 uhci0: interrupt while not operating ignored

but from time to time (10-20%) USB stops working instead with the following message:

 uhci0: host controller halted
 uhci0 regs: cmd=0080, sts=0020, intr=000f, frnum=00d4, flbase=00014350, sof=0040, portsc1=01a5, portsc2=0080
 intrs=214233
 QH(0xd2f7bf80) at 00016f80: hlink=00016fa2 elink=00000001

I believe it happens just before the system goes to suspend but I'm not
completely sure.

I'd appretiate even a workaround, i.e. how to force the controller to reset.

>How-To-Repeat:
Not easy. I failed to elaborate a deterministic way to repeat. I suspect that
an active PCMCIA WiFi card increases the probability.
>Fix:
	
>Release-Note:
>Audit-Trail:
>Unformatted: