> > I found another problem --- present in 2.6.23.1, 2.6.25rc3, 2.6.25rc6
quoted text
> > 
> > --- when I run three threads concurrently reading raw disk partition and 
> > suspend, I get 100% reproducible failure. (with one thread running it 
> > usually succeeds, sometimes fail)
> 
> Are they userland threads or kernel threads?

Useland threads. Just dd if=/dev/hda of=/dev/null

quoted text
> > Either the on-going I/O will jam BIOS and I need to remove power to 
> > continue.
> > 
> > Or the machine suspends, wakes up and reports "hda: lost interrupt" 
> > (2.6.23.1 was able to recover from this condition, 2.6.25rc3,6 does not 
> > recover and it is not able to send any more disk IOs).
> > 
> > How is suspending disk IO supposed to work?
> 
> That depends on the driver, if I understand your question correctly.

The driver is normal IDE.

quoted text
> > And why it doesn't?
> 
> Hard to tell.  You didn't provide much information ...
> 
> Thanks,
> Rafael

Mikulas
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/