Hi Alan,

ext Alan Stern wrote:
quoted text
> 
> I don't know, but it's a good start.  The IRQs for uhci-hcd and
> ehci-hcd are registered using the IRQF_DISABLED flag, which means that
> the handler routines uhci_irq() and ehci_irq() should always be called
> with interrupts disabled.
> 
> So that's the next thing to test.  Put a raw_irqs_disabled() test at
> the start of those two routines, just to make sure that interrupts
> don't somehow get enabled by mistake while the routine is running.  If 
> interrupts are already enabled when the routines are called then the 
> bug is somewhere else in the kernel.

Second try, slightly different tack: no dump_stack() in the _irq() 
functions, but disabled/enabled indication. Plus debugging code added in 
usb_hcd_add() to see the interrupt registration. Again too much data was 
printed for the bug to appear.

Looking at the collected data it looks like

  - ehci_irq() is always entered with interrupts enabled (WRONG).

  - uhci_irq() is entered sometimes with either interrupts enabled 
(WRONG) or disabled (OK).

But both interrupts are registered with IRQF_DISABLED...

Regards,

	Stefan

---
Stefan Becker
E-Mail: Stefan.Becker@nokia.com