Philip Blundell writes:
> Yes, and my point is that it really doesn't matter. If you put an EBSA-285 in
> a PC then for it to be any use you must have a host driver -- if you don't
> want it to communicate with the host, it might as well be a standalone device.
> So, all the time the EBSA is running this driver will be loaded and the PCI
> configuration space will be stable.
You're still totally missing the point.
1. Power up PC with EBSA285 installed and Linux ROMMified
2. BIOS scans the PCI bus, and the PCI bus stops since EBSA285 sends
retry cycles
3. Linux boots on EBSA285, sets up DC21285, sets up BAR masks, and
reads BARs for virt_to_bus.
4. BIOS reads and possibly sets BARs (depending whether the BIOS is in PNP
mode, read l-k for more information on this).
5. PC boots into PNP OS.
6. PNP OS reassigns/assigns all BARs.
The net result is that Linux on the EBSA285 has out of date BAR information,
or the BARs could be setup quite late. How does the EBSA285 know when to
read the BARs for virt_to_bus? Does this make my point crystal clear?
_____
|_____| ------------------------------------------------- ---+---+-
| | Russell King [EMAIL PROTECTED] --- ---
| | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / |
| +-+-+ --- -+-
/ | THE developer of ARM Linux |+| /|\
/ | | | --- |
+-+-+ ------------------------------------------------- /\\\ |
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
