On 10/07/2012 12:13 PM, Avi Kivity wrote:
> On 10/05/2012 06:24 PM, Blue Swirl wrote:
>>
>> I'd suppose addressing devices in the bus could be implemented more
>> efficiently with better use of memory API, now some of it is
>> reimplemented. Maybe Avi can propose something?
>
> Luckily the low-order bits are used for offsets, and the high-order bits
> are used for selecting the sub-device.
>
> So you could easily have
>
> struct IPackDevice {
> DeviceState qdev;
> int32_t slot;
> /* IRQ objects for the IndustryPack INT0# and INT1# */
> qemu_irq *irq;
> MemoryRegion io_space;
> MemoryRegion id_space;
> MemoryRegion int_space;
> MemoryRegion mem8_space; /* for las3 */
> MemoryRegion mem16_space; /* for las2 */
> };
>
> The PCI device would then just map each space (with
> memory_region_add_subregion()) into las1/las2/las3 such that the high
> bits select the device/space. The low bits would automatically become
> the offset into the space.
Note: you can easily verify that the mapping is correct with 'info
mtree'. You should see something like:
pci:
...
xxxxxxxx-xxxxxxxx las1
xxxxxxxx-xxxxxxxxx ip0-io
xxxxxxxx-xxxxxxxxx ip0-id
xxxxxxxx-xxxxxxxxx ip0-int
xxxxxxxx-xxxxxxxxx ip1-io
xxxxxxxx-xxxxxxxxx ip1-id
xxxxxxxx-xxxxxxxxx ip1-int
xxxxxxxx-xxxxxxxx las2
xxxxxxxx-xxxxxxxx ip0-mem16
xxxxxxxx-xxxxxxxx ip1-mem16
with the addresses relative to the pci address space.
--
error compiling committee.c: too many arguments to function
