Re: pci config space
Re: pci config space
- Subject: Re: pci config space
- From: Andrew Gallatin <email@hidden>
- Date: Wed, 9 May 2007 14:51:48 -0400 (EDT)
Alexander von Below writes:
> If I may ask: Why do you need only the PCI configuration space?
A variety of reasons. Mainly to check for things like PCI parity
errors in large clusters, where running a gui app like Reggie is
extraordinarily cubmersome, and a command-line tool (like lspci) scales
much better. Also to provide a way to upgrade firmware in our NICs in
the field w/o forcing people to boot into some OS where config space
access is trivial (linux, solaris, windows, *bsd, basically anything
but OSX) or to move the NIC into a box already running such an OS.
> My suggestion would be to write a simple driver and communicate with
> it. That will free you from the availability of the CHUD API.
The problem with this is that the only thing I can visualize will work
only on x86 (io port access). I can't seem to figure out a way to get
"random access" to the PCI bus. Eg, I don't see a way specify a
bus/slot/function and access its PCI config space, even in the kernel.
Hence, leaching from Chud seemed appealing.. :)
Drew
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden