This series makes the following test case succeed:
- Set the CODE env variable to the OVMF_CODE.fd file in the build
output.
- Set the TMPL env variable to the OVMF_VARS.fd file in the build
output.
- Set the ISO env variable to the pathname of a Linux LiveCD.
- Create a copy of the varstore template:
cp $TMPL vars.fd
- Create an empty disk image:
qemu-img create -f qcow2 test.qcow2 1G
- Start QEMU as follows.
qemu-system-x86_64 \
-m 2048 \
\
-machine q35,accel=kvm \
-device VGA \
\
-drive if=pflash,format=raw,file=$CODE,readonly \
-drive if=pflash,format=raw,file=vars.fd \
\
-drive id=cdrom,if=none,readonly,format=raw,file=$ISO \
-drive id=disk,if=none,format=qcow2,file=test.qcow2 \
\
-chardev file,id=debugfile,path=test.log \
-device isa-debugcon,iobase=0x402,chardev=debugfile \
\
-chardev stdio,id=char0,signal=off,mux=on \
-mon chardev=char0,mode=readline,default \
-serial chardev:char0 \
\
-device ioh3420,id=root_port,bus=pcie.0 \
\
-device x3130-upstream,id=upstream_port,bus=root_port \
\
-device xio3130-downstream,id=downstream_port1,bus=upstream_port,chassis=1 \
-device virtio-scsi-pci,id=scsi0,bus=downstream_port1 \
-device scsi-cd,bus=scsi0.0,drive=cdrom,bootindex=0 \
\
-device xio3130-downstream,id=downstream_port2,bus=upstream_port,chassis=2
This command line creates a PCIe root port, and cold-plugs a PCIe
switch into it. The PCIe switch has one upstream port and two
downstream ports. Into one of those downstream ports, a virtio-scsi
controller is cold-plugged, and the LiveCD is made available to the
guest as a SCSI CD-ROM on that virtio-scsi controller. The other
downstream port of the PCIe switch is left empty.
- After booting the LiveCD, enter [Ctrl-A C] to switch the terminal I/O
from the guest's serial port to the QEMU monitor, then hotplug a
virtio-block device with the following command into the second
downstream port:
device_add virtio-blk-pci,drive=disk,bus=downstream_port2
- Without the patches, the guest kernel will spew PCI resource
allocation errors to the syslog, and it might even hang eventually.
With the patches, the hotplugged disk can be used (/dev/vda).
Public branch: <https://github.com/lersek/edk2/commits/bridge_rsrc_pad>.
Cc: "Johnson, Brian J." <bjohn...@sgi.com>
Cc: Alex Williamson <alex.william...@redhat.com>
Cc: Andrew Fish <af...@apple.com>
Cc: Feng Tian <feng.t...@intel.com>
Cc: Jordan Justen <jordan.l.jus...@intel.com>
Cc: Liming Gao <liming....@intel.com>
Cc: Marcel Apfelbaum <mar...@redhat.com>
Cc: Michael D Kinney <michael.d.kin...@intel.com>
Cc: Ruiyu Ni <ruiyu...@intel.com>
Cc: Star Zeng <star.z...@intel.com>
Thanks
Laszlo
Laszlo Ersek (4):
MdePkg/IndustryStandard: introduce EFI_PCI_CAPABILITY_ID_SHPC
MdeModulePkg/PciBusDxe: look for the right capability in IsSHPC()
MdeModulePkg/PciBusDxe: recognize hotplug-capable PCIe ports
OvmfPkg: add PciHotPlugInitDxe
MdeModulePkg/Bus/Pci/PciBusDxe/PciHotPlugSupport.c | 73 +++-
MdeModulePkg/Bus/Pci/PciBusDxe/PciHotPlugSupport.h | 19 ++
MdePkg/Include/IndustryStandard/Pci22.h | 1 +
OvmfPkg/OvmfPkgIa32.dsc | 1 +
OvmfPkg/OvmfPkgIa32.fdf | 1 +
OvmfPkg/OvmfPkgIa32X64.dsc | 1 +
OvmfPkg/OvmfPkgIa32X64.fdf | 1 +
OvmfPkg/OvmfPkgX64.dsc | 1 +
OvmfPkg/OvmfPkgX64.fdf | 1 +
OvmfPkg/PciHotPlugInitDxe/PciHotPlugInit.c | 352 ++++++++++++++++++++
OvmfPkg/PciHotPlugInitDxe/PciHotPlugInit.inf | 46 +++
11 files changed, 496 insertions(+), 1 deletion(-)
create mode 100644 OvmfPkg/PciHotPlugInitDxe/PciHotPlugInit.inf
create mode 100644 OvmfPkg/PciHotPlugInitDxe/PciHotPlugInit.c
--
1.8.3.1
_______________________________________________
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
