How to block pci config-reads during device self-test?

[Brian King]

I've been doing some talking with various hardware folks to see if there
  is a way to get this fixed and so far the answer I have been getting
is no. It is how the hardware works and there isn't much that can be
done about it in microcode.

To work around the problem we could add userspace pci config read and
write function pointers to the pci_dev struct which a device driver
could overload. These would be invoked for user initiated pci config
accesses. The device driver could then do the right thing for that
device if necessary.

I don't like the idea of "learning to live with this". People do run
into this problem and telling them it can't be fixed is not an
acceptable answer.

-Brian

linas at austin.ibm.com wrote:
>
> Am having trouble with PCI config-space reads ... I have a device
> (actualy Brian King has it) that can perform a built-in-self test
> (BIST). However, if anything does a PCI config-read during BIST, then
> the device does something crazy that makes the PCI controller chip
> take it offline.
>
> I'm not sure what's doing the config-spcae reads ... seems to be some
> user-space tool or daemon. I'm wondering if there is any practical way
> to block such reads to a given device until its self-test sequence is
> completed. I could try to modify the architecture-specific pci files
> to do this (arch/ppc64/kernel/pSeries_pci.c) but this seems a tad ugly
> ... is there another way? or do we have to just learn to live with
> this ahrdware?

--

** Sent via the linuxppc64-dev mail list. See http://lists.linuxppc.org/