Re: [patch 3/2] m68k: Atari EtherNAT - add writew_be for data push

[Michael Schmitz <schmitz@biophys.uni-duesseldorf.de>]

Hi,

> > > case. Given that the lock handling of IDE may be dodgy, that scares
> me a bit.
> 
> > > Reasons enough for you?
> 
> Oops. It was not my intention to offend anyone. Excuses if I did.

No offense taken, I assure you :-) 

> > > I'll focus on getting SCSI fixed, first and foremost. First stab at
> this is
> > > attached (not fudging with the timeouts, but failing the request
> instead if the
> > > lock cannot be obtained).
> 
> That would be great because in that case I could start to test the
> installation procedure again.

I've come across a couple of other problems in the interaction between SCSI and 
IDE. Plus I'm seeing lost interrupts on IDE on occasion now, and I'm not sure if 
that is due to hardware trouble with my IDE disk, or SCSI hogging the interrupt 
for far too long. I have a hunch this may be at the core of the problem.

On the other hand, the SCSI driver tested OK when used without IDE I/O so 
something's improved...

I need to diagnose this in more detail really. I'll have to consult Stephen on 
how to replace modules in the install initrd image so kernel fixes can be 
tested without a full d-i rebuild. 

	Michael