diff options
author | Łukasz Gieryk <lukasz.gieryk@linux.intel.com> | 2022-05-09 16:16:20 +0200 |
---|---|---|
committer | Klaus Jensen <k.jensen@samsung.com> | 2022-06-23 23:24:29 +0200 |
commit | 58660bfa3694a2d05eacb8dca7c28ffa08917578 (patch) | |
tree | e1e9df7427c7ed413afb3b6f3634c32e599d5c6a /hw/acpi/pcihp.c | |
parent | b7698b917abcbe673cbffc15ee41d95d3daa4af3 (diff) | |
download | qemu-58660bfa3694a2d05eacb8dca7c28ffa08917578.zip qemu-58660bfa3694a2d05eacb8dca7c28ffa08917578.tar.gz qemu-58660bfa3694a2d05eacb8dca7c28ffa08917578.tar.bz2 |
hw/acpi: Make the PCI hot-plug aware of SR-IOV
PCI device capable of SR-IOV support is a new, still-experimental
feature with only a single working example of the Nvme device.
This patch in an attempt to fix a double-free problem when a
SR-IOV-capable Nvme device is hot-unplugged in the following scenario:
Qemu CLI:
---------
-device pcie-root-port,slot=0,id=rp0
-device nvme-subsys,id=subsys0
-device nvme,id=nvme0,bus=rp0,serial=deadbeef,subsys=subsys0,sriov_max_vfs=1,sriov_vq_flexible=2,sriov_vi_flexible=1
Guest OS:
---------
sudo nvme virt-mgmt /dev/nvme0 -c 0 -r 1 -a 1 -n 0
sudo nvme virt-mgmt /dev/nvme0 -c 0 -r 0 -a 1 -n 0
echo 1 > /sys/bus/pci/devices/0000:01:00.0/reset
sleep 1
echo 1 > /sys/bus/pci/devices/0000:01:00.0/sriov_numvfs
nvme virt-mgmt /dev/nvme0 -c 1 -r 1 -a 8 -n 1
nvme virt-mgmt /dev/nvme0 -c 1 -r 0 -a 8 -n 2
nvme virt-mgmt /dev/nvme0 -c 1 -r 0 -a 9 -n 0
sleep 2
echo 01:00.1 > /sys/bus/pci/drivers/nvme/bind
Qemu monitor:
-------------
device_del nvme0
Explanation of the problem and the proposed solution:
1) The current SR-IOV implementation assumes it’s the PhysicalFunction
that creates and deletes VirtualFunctions.
2) It’s a design decision (the Nvme device at least) for the VFs to be
of the same class as PF. Effectively, they share the dc->hotpluggable
value.
3) When a VF is created, it’s added as a child node to PF’s PCI bus
slot.
4) Monitor/device_del triggers the ACPI mechanism. The implementation is
not aware of SR/IOV and ejects PF’s PCI slot, directly unrealizing all
hot-pluggable (!acpi_pcihp_pc_no_hotplug) children nodes.
5) VFs are unrealized directly, and it doesn’t work well with (1).
SR/IOV structures are not updated, so when it’s PF’s turn to be
unrealized, it works on stale pointers to already-deleted VFs.
The proposed fix is to make the PCI ACPI code aware of SR/IOV.
Signed-off-by: Łukasz Gieryk <lukasz.gieryk@linux.intel.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Diffstat (limited to 'hw/acpi/pcihp.c')
-rw-r--r-- | hw/acpi/pcihp.c | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/hw/acpi/pcihp.c b/hw/acpi/pcihp.c index bf65bbe..84d75e6 100644 --- a/hw/acpi/pcihp.c +++ b/hw/acpi/pcihp.c @@ -192,8 +192,12 @@ static bool acpi_pcihp_pc_no_hotplug(AcpiPciHpState *s, PCIDevice *dev) * ACPI doesn't allow hotplug of bridge devices. Don't allow * hot-unplug of bridge devices unless they were added by hotplug * (and so, not described by acpi). + * + * Don't allow hot-unplug of SR-IOV Virtual Functions, as they + * will be removed implicitly, when Physical Function is unplugged. */ - return (pc->is_bridge && !dev->qdev.hotplugged) || !dc->hotpluggable; + return (pc->is_bridge && !dev->qdev.hotplugged) || !dc->hotpluggable || + pci_is_vf(dev); } static void acpi_pcihp_eject_slot(AcpiPciHpState *s, unsigned bsel, unsigned slots) |