Hotadd pci devices: Difference between revisions
No edit summary |
m (added categories) |
||
(2 intermediate revisions by one other user not shown) | |||
Line 10: | Line 10: | ||
for m in acpiphp pci_hotplug; do sudo modprobe ${m}; done | for m in acpiphp pci_hotplug; do sudo modprobe ${m}; done | ||
== Use QEMU monitor to hot add device == | |||
(qemu) drive_add 0 file=/storage/10G.img,format=qcow2,id=drive-virtio-disk1,if=none | |||
(qemu) device_add virtio-blk-pci,scsi=on,drive=drive-virtio-disk1 | |||
* check attached PCI devices | |||
(qemu) info pci | |||
* notice | * notice | ||
Line 88: | Line 96: | ||
[ 153.750637] sda: unknown partition table | [ 153.750637] sda: unknown partition table | ||
[ 153.759275] sd 2:0:0:0: [sda] Attached SCSI disk | [ 153.759275] sd 2:0:0:0: [sda] Attached SCSI disk | ||
[[Category:Docs]][[Category:HowTo]][[Category:Devices]] |
Latest revision as of 09:19, 31 May 2015
Using PCI Hotplug Support
PCI Hotplug support is now available in kvm.git, here is how to use it.
Install PCI Hotplug drivers in the guest (Linux)
I loaded the following modules to get pci hotplug working:
- acpiphp
- pci_hotplug
You can either add these to your distro's module list to load on boot, or run a command like this.
for m in acpiphp pci_hotplug; do sudo modprobe ${m}; done
Use QEMU monitor to hot add device
(qemu) drive_add 0 file=/storage/10G.img,format=qcow2,id=drive-virtio-disk1,if=none (qemu) device_add virtio-blk-pci,scsi=on,drive=drive-virtio-disk1
- check attached PCI devices
(qemu) info pci
- notice
- pci_add is old style, instead use device_add
Use QEMU monitor to hot add devices
The QEMU monitor has a new command, pci_add, which is used to hotadd the pci devices.
QEMU 0.9.1 monitor - type 'help' for more information (qemu) help pci_add pci_add bus nic|storage [[vlan=n][,macaddr=addr][,model=type]] [file=file][,if=type][,bus=nr]... -- hot-add PCI device
Add a NIC
The following command in the QEMU monitor hot-adds a nic to the guest.
- pci_add auto nic model=e1000
You can replace model=e1000 with either model=rtl8139, model=virtio, or model=ne2k_pci. Note that model=ne2000 isn't pci-based but rather ISA-based and cannot be used for hotplug.
(qemu) pci_add auto nic model=e1000 WARNING: PCI region size must be pow2 type=0x0, size=0x60000 OK bus 0, slot 5, function 0 (devfn 40)
In the guest we can check if the device was added by looking at dmesg.
[ 91.808265] acpiphp: Slot [30] registered [ 91.808276] acpiphp: Slot [31] registered [ 110.828130] PCI: Ignore bogus resource 6 [0:0] of 0000:00:02.0 [ 110.828333] decode_hpp: Could not get hotplug parameters. Use defaults [ 110.903259] Intel(R) PRO/1000 Network Driver - version 7.3.20-k2-NAPI [ 110.903264] Copyright (c) 1999-2006 Intel Corporation. [ 110.903489] ACPI: PCI Interrupt Link [LNKA] enabled at IRQ 10 [ 110.903492] ACPI: PCI Interrupt 0000:00:05.0[A] -> Link [LNKA] -> GSI 10 (level, high) -> IRQ 10 [ 111.238713] e1000: 0000:00:05.0: e1000_probe: (PCI:33MHz:32-bit) 52:54:00:12:34:57 [ 111.292550] e1000: eth1: e1000_probe: Intel(R) PRO/1000 Network Connection
Add a disk
Hot-adding a disk means hot-adding a PCI device, so you need PCI hotplug support in the guest for this to work.
The following command in the QEMU monitor will add a scsi disk to the guest:
* pci_add auto storage file=/home/rharper/work/images/empty.img,if=scsi
You can replace if=scsi with if=virtio, but IDE is not hotpluggable.
QEMU 0.9.1 monitor - type 'help' for more information
(qemu) pci_add auto storage file=/home/rharper/work/images/empty.img,if=scsi
OK bus 0, slot 5, function 0 (devfn 40)
In the guest we see via dmesg that the device was added.
[ 150.581752] PCI: Enabling device 0000:00:05.0 (0000 -> 0003) [ 150.582634] ACPI: PCI Interrupt Link [LNKA] enabled at IRQ 10 [ 150.582637] ACPI: PCI Interrupt 0000:00:05.0[A] -> Link [LNKA] -> GSI 10 (level, high) -> IRQ 10 [ 150.585099] sym0: <895a> rev 0x0 at pci 0000:00:05.0 irq 10 [ 150.594628] sym0: No NVRAM, ID 7, Fast-40, LVD, parity checking [ 150.607861] sym0: SCSI BUS has been reset. [ 150.607883] scsi2 : sym-2.2.3 [ 153.608433] scsi 2:0:0:0: Direct-Access QEMU QEMU HARDDISK 0.9. PQ: 0 ANSI: 3 [ 153.608462] target2:0:0: tagged command queuingenabled, command queue depth 16. [ 153.608484] target2:0:0: Beginning Domain Validation [ 153.609517] target2:0:0: Domain Validation skipping write tests [ 153.609522] target2:0:0: Ending Domain Validation [ 153.618880] scsi 2:0:0:0: Attached scsi generic sg1 type 0 [ 153.749967] Driver 'sd' needs updating - please use bus_type methods [ 153.750204] sd 2:0:0:0: [sda] 20971520 512-byte hardware sectors (10737 MB) [ 153.750253] sd 2:0:0:0: [sda] Write Protect is off [ 153.750255] sd 2:0:0:0: [sda] Mode Sense: 14 00 00 00 [ 153.750354] sd 2:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA [ 153.750478] sd 2:0:0:0: [sda] 20971520 512-byte hardware sectors (10737 MB) [ 153.750253] sd 2:0:0:0: [sda] Write Protect is off [ 153.750255] sd 2:0:0:0: [sda] Mode Sense: 14 00 00 00 [ 153.750354] sd 2:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA [ 153.750478] sd 2:0:0:0: [sda] 20971520 512-byte hardware sectors (10737 MB) [ 153.750534] sd 2:0:0:0: [sda] Write Protect is off [ 153.750536] sd 2:0:0:0: [sda] Mode Sense: 14 00 00 00 [ 153.750634] sd 2:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA [ 153.750637] sda: unknown partition table [ 153.759275] sd 2:0:0:0: [sda] Attached SCSI disk