Re: ukbd: short freeze when activating LEDs

From: Hans Petter Selasky <hselasky_at_c2i.net>
Date: Sat, 5 Sep 2009 19:41:02 +0200
On Saturday 05 September 2009 18:47:32 Stefan Ehmann wrote:
> On Saturday 05 September 2009 18:02:37 you wrote:
> > On Saturday 05 September 2009 16:32:55 Stefan Ehmann wrote:
> > > Whenever I press capslock/numlock, the system shortly (< 0.5 ms)
> > > freezes.
> >
> > How did you measure this?
>
> The mouse is not responding and sound also stops during that period.
> It's probably much shorter, more likely 0.1-0.2ms.
>
> > Are you able to figure out why it is hanging?
> >
> > Has this got anything to do with BIOS or microcode running on the CPU?
>
> No idea. top shows > 95% interrupt if I'm on the console, in xorg the cpu
> goes to system. Interestingly, the number of interrupts displayed by systat
> is very moderate (if I just move the mouse, it's much higher)
>
> > USB keyboard LEDs are set asynchronously. It should not block like you
> > explain.
>
> The keyboard works fine on my notebook. I think the problem is the USB
> controller on my PC, not the keyboard driver. dmesg reports it as
> uhci0: <VIA 83C572 USB controller> port 0xb400-0xb41f at device 16.0 on
> pci0 uhci0: [ITHREAD]
> uhci0: LegSup = 0x003a
> usbus0: <VIA 83C572 USB controller> on uhci0
>
> Historically, I've always had lots of trouble with USB on this particular
> PC. With the old USB code, I've got lots of issues. scanner didn't work at
> all, sometimes even problems with basic things like umass.
>
> With new USB things are much better, but not always perfect.
>
> If there's no easy way to debug, it's probably not worth the hassle. The PC
> is nearing the end of its life-time anyway.

I see. Your observation is interesting and valuable. It might look like some 
piece of hardware is excessively using the RAM. Maybe an USB recording off the 
USB cable will give the final answer. There are some relatively cheap USB 
analyzers from Beagle which you can buy.

--HPS
Received on Sat Sep 05 2009 - 15:40:42 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:54 UTC