Re: [PATCH] to drivers/parport/ieee1284_ops.c to fix timing dependent hang

[Tim Josling <tej@melbpc.org.au>]

Jeremy Fitzhardinge wrote:
> On Sat, 2003-03-22 at 01:23, Tim Josling wrote:
> 
>>According to my reading of the code, it should only happen in polled 
>>mode, but I have only one week of experience looking at kernel source.
> 
> 
> I'm wondering if a better fix might be to have something like:
> 
> 	if (wait * 2 > wait)
> 		wait *= 2;
> 
> at the bottom of the loop, so that the wrap-around doesn't happen.
> 

I tried
  wait *=2;
	if (wait > 10 * HZ)
		wait = 10 * HZ;
and that worked, so your theory would probably work. However to my mind 
it is a hack.

> 
>>So it should be a work-around, assuming interrupts work on the parallel 
>>port on your system :-). It is an very vexing problem, as I'm sure you know.
>>
>>By the way, LJ1100s tend to get page feeding problems about the time the 
>>warranty runs out, but HP has a free kit you can order to fix the problem.
> 
> 
> Yes, I just installed it.  It suddenly made the printer useful again, so
> I've printing more, and seeing the hangs.  I enabled interrupts, which
> seems to work OK.
> 
> 	J

Good. That (enabling interrupts fixes the problem) tends to add credence 
to my theory.

Happy printing!

Tim Josling