Re: USB stops working with any of 2.4.22-pre's after 2.4.21

[John Wong]

acpi=off on 2.4.22-pre4 for the last 4 hours and haven't run into the
interrupt/usb problems.  Normally the problems arise not too soon into
the boot up, forcing me to boot up 2.4.21 if I ever wanted to log on
remotely.  Looks like all my problems are being caused by the newer ACPI
code base merged since pre1.

John

On Thu, Jul 10, 2003 at 11:58:13AM -0700, John Wong wrote:
> Sorry if this is a duplicate...I received an error bounce
> 
> ----
> ACPI isn't absolutely needed, but it comes in handy for issuing a
> shutdown.
> Though I suppose I can use apm to do this as well.
> 
> On a uni-processor system, isn't the ACPI tables used to setup the
> IO-APIC bits to have PCI IRQ routing so that I have more than 16
> interrupts?  This was another reason I had enabled ACPI.
> 
> When I do get those interrupt problems, my network still continues to
> function while the USB mouse fails.  I do agree it's not USB specific.
> It's more tied in to ACPI/(IO-)APIC.
> 
> When I get back home, I'll try booting it with acpi=off on pre4 and see
> how it goes.
> 
> John
> 
> On Thu, Jul 10, 2003 at 10:02:17AM -0700, Greg KH wrote:
> > On Wed, Jul 09, 2003 at 11:58:01PM -0700, John Wong wrote:
> > > Jul  9 23:39:44 gambit kernel: eth0: Resetting the Tx ring pointer.
> > > Jul  9 23:39:51 gambit kernel: usb-ohci.c: unlink URB timeout
> > > Jul  9 23:39:54 gambit kernel: NETDEV WATCHDOG: eth0: transmit timed out
> > 
> > Hm, looks like bad things are happening with your interrupts.
> > 
> > Do you need acpi to run this box?  What happens if you disable it?
> > 
> > As it looks like networking is also in trouble, I don't think this is a
> > USB specific problem for you.
> > 
> > Good luck,
> > 
> > greg k-h
> >