Re: [PATCH 2.5.54] hermes: serialization fixes

[Alexander Hoogerhuis <alexh@ihatent.com>]

Stephen Evanchik <evanchsa@clarkson.edu> writes:

> On Friday 03 January 2003 12:47, you wrote:
> |  Why not put the spinlock/unlock inside hermes_bap_seek()?  Smaller, better
> |  contained and more readable.
> 
> That's the better solution, I'm trying to coordinate a bit with the 
> maintainer. The only reason I didn't do this in the first place is because 
> there is a (possibly unecessary) delay loop inside hermes_bap_seek that I 
> believe is trying to combat the same problem. I'm awaiting a response from 
> the maintainer since he knows a bit more about the hardware than I do.
> 

There is something not quite right about this patch. I used have a ton
of errors in my logs, and this patch seemed to clear this out nicely.

When I run with a patched driver now, I run for about 20 minutes with
various loads and sudenly the ksoftirqd_CPU0 process starts to hog my
CPU and not wanting to let go. As soon as I pull out the card, the
load returns to normal.

Is there any way I can provide more details on what is happening?

> 
> Stephen
> 

mvh,
A
-- 
Alexander Hoogerhuis                               | alexh@ihatent.com
CCNP - CCDP - MCNE - CCSE                          | +47 908 21 485
"You have zero privacy anyway. Get over it."  --Scott McNealy