Fakephp to hotplug FPGAs

Fakephp to hotplug FPGAs

[Jay Miller]

Hi All,

I was hoping that someone could help me out of a hotplugging nightmare?

We're running Linux 2.6.6 on and embedded system with FPGAs on the other
side of a PCI-X bridge.  The FPGAs are programmed after boot and I have
tried to use fakephp to discover the newly programmed device without
rebooting.

This works in the sense that lspci reports a new device, but it doesn't
assign any resources for the device.  My assumption is that the bridge
doesn't have any resources to give because no device was detected on the
first bus scan.

I've searched for folks in a similar bind, and I did find John Carlson's
pci-rescan, but that appears to do the same thing as fakephp, and only
on bus 0.

I also used setpci to set the correct pci config space values for the
device and for the bridge, and tried using fakephp to re-discover the
device with the proper settings.

I even went as far as removing the "safety" in fakephp that prevents you
from removing a bridge to see if I'd get "lucky", and both the bridge
and device would be repopulated.  No dice.

I get the feeling that I'm missing something.  What am I doing wrong
(besides hacking up fakephp for my own purposes)?

Thanks,

Jay Miller
Actuality Systems, Inc.
jmiller@actuality-systems.com




-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Re: Fakephp to hotplug FPGAs

[Greg KH]

On Wed, Dec 08, 2004 at 04:25:16PM -0500, Jay Miller wrote:
> Hi All,
> 
> I was hoping that someone could help me out of a hotplugging nightmare?
> 
> We're running Linux 2.6.6 on and embedded system with FPGAs on the other
> side of a PCI-X bridge.  The FPGAs are programmed after boot and I have
> tried to use fakephp to discover the newly programmed device without
> rebooting.
> 
> This works in the sense that lspci reports a new device, but it doesn't
> assign any resources for the device.  My assumption is that the bridge
> doesn't have any resources to give because no device was detected on the
> first bus scan.

That is true.  You are going to have to figure out how to reserve a
resource range at boot time.  Hardware that supports PCI hotplug
controllers reserve this in their BIOS (or firmware in the case of PPC64
hardware).

It looks like you need to adapt the compact pci hotplug driver to do
this.  There were patches for the 2.4 kernel to accept ranges of
resources at boot time to ignore, and later on those ranges were
availble for the compact pci controller to dole out to any devices
attaching through it.  Scan the linux pci hotplug mailing list archives
for these patches.

Good luck,

greg k-h


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
http://productguide.itmanagersjournal.com/
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel