On 06/10/2015 10:34 PM, Laszlo Ersek wrote:
On 06/10/15 21:26, Marcel Apfelbaum wrote:
On 06/10/2015 08:07 PM, Laszlo Ersek wrote:
The PXB implementation doesn't allow firmware (SeaBIOS or OVMF) to boot
off devices behind the PXB. This happens because the
sysbus_get_fw_dev_path() function in "hw/core/sysbus.c" doesn't have
enough information to format a unique identifier for the PXB in question,
and consequently the OpenFirmware device path passed down to the guest
firmware in the "bootorder" fw_cfg file is unusable for identifying the
boot device.

For example, the command line fragment

    -device pxb,id=bridge1,bus_nr=4 \
    \
    -netdev user,id=netdev0 \
    -device e1000,netdev=netdev0,bus=bridge1,addr=2,bootindex=0

results in the following "bootorder" entry:

    /pci/pci-bridge@0/ethernet@2/ethernet-phy@0

The initial "pci" node is formatted by sysbus_get_fw_dev_path(), and the
resultant OpenFirmware device path is independent of bus_nr=4 -- and
therefore it is useless for identifying the device.

In this patch we insert a dummy bus between the main sysbus and the
TYPE_PXB_HOST device. Formatting child addresses is a bus class level
responsibility, which is exactly what we'll use here.

After the patch, the same command line fragment results in the following
OpenFirmware device path in the "bootorder" fw_cfg file:

    /extra-pci-roots/pxbhost@4/pci-bridge@0/ethernet@2/ethernet-phy@0

The original, initial "/pci" fragment has been replaced with
"/extra-pci-roots/pxbhost@4", which (a) looks better, (b) provides all
the
necessary information. sysbus_get_fw_dev_path() formats the first node
("extra-pci-roots") as always, and the new function
extra_pci_roots_bus_get_fw_dev_path() formats the second one
("pxbhost@4").

Here's the comparison ("diff -u -b -U28") between the "info qtree"
outputs, before and after (the hpet device is the first common line):

BTW, did you try to boot from it :) ?

No, not yet; I'll have to write additional OVMF code for that; but the
syntax is OK, and the information is there, so I'm fairly sure I can
write that code. :)

The SeaBIOS side I'll probably leave to you... /me ducks :)
I'll take it, sure, you are doing all the work anyway :)

Thanks,
Marcel

Thanks
Laszlo



Reply via email to