Re: revert yenta free_irq on suspend

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Saturday, 30 of July 2005 22:54, Russell King wrote:
> On Sat, Jul 30, 2005 at 01:36:24PM -0700, Linus Torvalds wrote:
> > On Sat, 30 Jul 2005, Russell King wrote:
> > > 
> > > What this probably means is that we need some way to turn off interrupts
> > > from devices on suspend, and on resume, keep them off until drivers
> > > have had a chance to quiesce all devices, turn them back on, and then
> > > do full resume.
> > 
> > No, we just need to suspend and resume the interrupt controller properly.  
> > Which we had the technology for, and we actually used to do, but for some
> > (incorrect) reason ACPI people thought it should be up to individual
> > drivers.
> 
> I don't think so - I believe one of the problem cases is where you
> have a screaming interrupt caused by an improperly setup device.

This is happening in the real life, it seems (eg please see
http://bugzilla.kernel.org/show_bug.cgi?id=4416#c36).

> Consider the case where you have a shared interrupt line and you're
> partially through resuming devices, when one unresumed device (setup
> by the BIOS) suddenly starts asserting its interrupt.
> 
> The kernel then disables the source.  Unfortunately, that was the IRQ
> for your USB host, which has your USB keyboard and mouse attached.

Also, this has been discussed in this thread on the linux-pm list:
http://lists.osdl.org/pipermail/linux-pm/2005-May/000955.html

Greets,
Rafael
 

-- 
- Would you tell me, please, which way I ought to go from here?
- That depends a good deal on where you want to get to.
		-- Lewis Carroll "Alice's Adventures in Wonderland"