On Monday 14 April 2008 17:01:04 Larry Finger wrote:
> Michael Buesch wrote:
> > 
> > What do the target_idle_tssi from SPROM and the current_idle_tssi read
> > from the PHY look like? They should be almost equal. If not, this could
> > result in a complete TX breakage.
> 
> Would an idle_tssi problem also prevent DMA interrupts? That doesn't 
> seem likely to me.

I don't know. But I think this whole discussion is too much based on
theoretical speculations on what _could_ be going on.
The _only_ way I found out why stuff didn't work in the past
was by printing random values and registers. That's really the only
way you can get a _whole_ picture of what's going on.
So you have some assumptions that you know are right and then you check them.
In this case the idle_tssi. We know that abs(current - idle) <= 20, so we
should check whether that's the case. If it is, fine, go over to the next check.
This way we got several PCI, SSB, and PCMCIA chips working and I'm
pretty sure it's the only way to get a device working, until someone
proves me wrong. :)
Of course on regressions one can also compare some code. But in this case
the code changes are simply too big. It's basically a rewrite. So that
would not be the first thing I'd do. In fact, it would be the very last.

-- 
Greetings Michael.
_______________________________________________
Bcm43xx-dev mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev

Reply via email to