On Fri, Aug 09, 2013 at 06:49:18PM +0300, Michael S. Tsirkin wrote:
On Fri, Aug 09, 2013 at 12:13:06AM -0400, Kevin O'Connor wrote:
I don't think SeaBIOS should continue to do the above once the tables are moved to QEMU. QEMU has all the info SeaBIOS has, so it can generate the tables correctly on its own.
In all practical situations, the PCI window should be at least an order of magnatude greater than the sum of the PCI bars in the system. If the bars are actually bigger than the window, then things are going to fail - the best the firmware can do is try to fail gracefully. I don't think it's worth the complexity to design mixed ownership and advanced interfaces just so we can fail slightly better.
If this is a real worry, QEMU can sum all the PCI bars and warn the user if they don't fit.
If we make it a rule that PCI is`setup before ACPI tables are read, then QEMU can do the patching itself when it detects BIOS reading the tables.
QEMU can do the patching regardless of the order of initialization. The PCI bar sizes are static - they have no relationship to the timing of the BIOS PCI initialization.