2.6.10 resuming laptop from suspension f*cks usb subsystem

2.6.10 resuming laptop from suspension f*cks usb subsystem

[Miguelanxo Otero Salgueiro]

Apart from the timer drift thingie, 2.6.10 brought some new features 
like usb devices (ramdomly) not working after resuming from suspend mode.

In 2.6.9 usb worked well, resuming from suspend just throws a bunch 
(near 20) messages like:

    drivers/usb/input/hid-core.c: input irq status -84 received

(getting form /var/log/syslog, just don't have time to switch back to 
2.6.9 and fiddle with it again). After those messages, the usb subsystem 
comes stable again and worked like a charm.

In 2.6.10, resuming from suspend mode just (randomly) crashes the USB 
subsystem, and I get the same messages (not sure about the whole message 
but the "-84" part really is there) over and over again until I reboot.

Re: 2.6.10 resuming laptop from suspension f*cks usb subsystem

[Paul Ionescu]

On Wed, 29 Dec 2004 15:53:46 +0100, Miguelanxo Otero Salgueiro wrote:

> Apart from the timer drift thingie, 2.6.10 brought some new features like
> usb devices (ramdomly) not working after resuming from suspend mode.
> 

Same problem here.
For me, the work around was to remove the uhci-hcd before suspending and
inserting it after resuming.

Re: 2.6.10 resuming laptop from suspension f*cks usb subsystem

[Pavel Machek]

Hi!

> Apart from the timer drift thingie, 2.6.10 brought some new features 
> like usb devices (ramdomly) not working after resuming from suspend mode.
> 
> In 2.6.9 usb worked well, resuming from suspend just throws a bunch 
> (near 20) messages like:
> 
>    drivers/usb/input/hid-core.c: input irq status -84 received
> 
> (getting form /var/log/syslog, just don't have time to switch back to 
> 2.6.9 and fiddle with it again). After those messages, the usb subsystem 
> comes stable again and worked like a charm.
> 
> In 2.6.10, resuming from suspend mode just (randomly) crashes the USB 
> subsystem, and I get the same messages (not sure about the whole message 
> but the "-84" part really is there) over and over again until I reboot.

Does it still happen with noapic? 2.6.10 has some interrupt related
problems with APIC...

								Pavel
-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!

Re: 2.6.10 resuming laptop from suspension f*cks usb subsystem

[Miguelanxo Otero Salgueiro]

>>In 2.6.10, resuming from suspend mode just (randomly) crashes the USB 
>>subsystem, and I get the same messages (not sure about the whole message 
>>but the "-84" part really is there) over and over again until I reboot.
>>    
>>
>
>Does it still happen with noapic? 2.6.10 has some interrupt related
>problems with APIC...
>
>								Pavel
>  
>
I have just rebooted 2.6.10 with this LILO command line

    LILO Boot: Linux-2.6.10 noapic

and if that disables APIC, then I've got the same problem. After 
suspending the laptop
two times, I get the same lines (described below) and the usb system 
goes nuts. After
removing & inserting uhci_hcd everything works fine again.

The lines are (endless loop):
drivers/usb/input/hid-core.c: input irq status -84 received

Re: 2.6.10 resuming laptop from suspension f*cks usb subsystem

[Pavel Machek]

Hi!

> >>In 2.6.10, resuming from suspend mode just (randomly) crashes the USB 
> >>subsystem, and I get the same messages (not sure about the whole message 
> >>but the "-84" part really is there) over and over again until I reboot.
> >>   
> >>
> >
> >Does it still happen with noapic? 2.6.10 has some interrupt related
> >problems with APIC...

> I have just rebooted 2.6.10 with this LILO command line
> 
>    LILO Boot: Linux-2.6.10 noapic
> 
> and if that disables APIC, then I've got the same problem. After 
> suspending the laptop
> two times, I get the same lines (described below) and the usb system 
> goes nuts. After
> removing & inserting uhci_hcd everything works fine again.
> 
> The lines are (endless loop):
> drivers/usb/input/hid-core.c: input irq status -84 received

I guess you need to ask on usb lists. If removing/inserting uhci_hcd
helps, it is likely that uhci_hcd needs to do a bit more in its
suspend/resume callbacks.
								Pavel
-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!