Re: Weird PCI problem -Reply

Linux Lists (lists@cyclades.com)
Fri, 16 Apr 1999 09:17:21 -0700 (PDT)


On Thu, 15 Apr 1999, Mike Phillips wrote:
>
> I went a took a look at the cyclades code in my kernel source (2.2.5).
>
> You said the problem was with 2.0.36, well I'm not sure when 2.0.36 was
> released (early 98, if not earlier), but there have been a lot of
> changes to the cyclades driver in the last year.

I know, I made them ... :)

> You may want to try the board in a machine with a 2.2.x kernel.

This is a _production_ system in a _customer_ . I can't play _that_ much
with this system. If I could, I certainly would, believe me.

Anyhow, it shouldn't happen even in 2.0.36. Furthermore, I've never seen
it happen before. And we (Cyclades) have a very _very_ large Cyclom-Y
installed base using this HW version / kernel driver version, so I really
think this is a very weird and particular case.

> As for the interrupts themselves, they are actually implicitly turned on
> by the driver, so the driver may not be getting to that piece of code.

It is executing it, I checked that. Anyhow, I strongly believe that we
should _first_ solve the problem with the wrong allocation of the PCI
resource (I/O instead of memory address), and then probably everything
would be solved.

> The reason I mention bus mastering before was because I had a board not
> generating interrupts (although allowing itself to be setup), and
> turning on the bus master worked (I did it in the code).

A Cyclades card or something else ?!?!? If it was a Cyclades card, I'd
like to know more details about it (pieces of code, etc.).

Regards,
Ivan

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