]> git.proxmox.com Git - mirror_qemu.git/blob - docs/igd-assign.txt
minikconf: do not include variables from MINIKCONF_ARGS in config-all-devices.mak
[mirror_qemu.git] / docs / igd-assign.txt
1 Intel Graphics Device (IGD) assignment with vfio-pci
2 ====================================================
3
4 IGD has two different modes for assignment using vfio-pci:
5
6 1) Universal Pass-Through (UPT) mode:
7
8 In this mode the IGD device is added as a *secondary* (ie. non-primary)
9 graphics device in combination with an emulated primary graphics device.
10 This mode *requires* guest driver support to remove the external
11 dependencies generally associated with IGD (see below). Those guest
12 drivers only support this mode for Broadwell and newer IGD, according to
13 Intel. Additionally, this mode by default, and as officially supported
14 by Intel, does not support direct video output. The intention is to use
15 this mode either to provide hardware acceleration to the emulated graphics
16 or to use this mode in combination with guest-based remote access software,
17 for example VNC (see below for optional output support). This mode
18 theoretically has no device specific handling dependencies on vfio-pci or
19 the VM firmware.
20
21 2) "Legacy" mode:
22
23 In this mode the IGD device is intended to be the primary and exclusive
24 graphics device in the VM[1], as such QEMU does not facilitate any sort
25 of remote graphics to the VM in this mode. A connected physical monitor
26 is the intended output device for IGD. This mode includes several
27 requirements and restrictions:
28
29 * IGD must be given address 02.0 on the PCI root bus in the VM
30 * The host kernel must support vfio extensions for IGD (v4.6)
31 * vfio VGA support very likely needs to be enabled in the host kernel
32 * The VM firmware must support specific fw_cfg enablers for IGD
33 * The VM machine type must support a PCI host bridge at 00.0 (standard)
34 * The VM machine type must provide or allow to be created a special
35 ISA/LPC bridge device (vfio-pci-igd-lpc-bridge) on the root bus at
36 PCI address 1f.0.
37 * The IGD device must have a VGA ROM, either provided via the romfile
38 option or loaded automatically through vfio (standard). rombar=0
39 will disable legacy mode support.
40 * Hotplug of the IGD device is not supported.
41 * The IGD device must be a SandyBridge or newer model device.
42
43 For either mode, depending on the host kernel, the i915 driver in the host
44 may generate faults and errors upon re-binding to an IGD device after it
45 has been assigned to a VM. It's therefore generally recommended to prevent
46 such driver binding unless the host driver is known to work well for this.
47 There are numerous ways to do this, i915 can be blacklisted on the host,
48 the driver_override option can be used to ensure that only vfio-pci can bind
49 to the device on the host[2], virsh nodedev-detach can be used to bind the
50 device to vfio drivers and then managed='no' set in the VM xml to prevent
51 re-binding to i915, etc. Also note that IGD is also typically the primary
52 graphics in the host and special options may be required beyond simply
53 blacklisting i915 or using pci-stub/vfio-pci to take ownership of IGD as a
54 PCI class device. Lower level drivers exist that may still claim the device.
55 It may therefore be necessary to use kernel boot options video=vesafb:off or
56 video=efifb:off (depending on host BIOS/UEFI) or these can be combined to
57 a catch-all, video=vesafb:off,efifb:off. Error messages such as:
58
59 Failed to mmap 0000:00:02.0 BAR <>. Performance may be slow
60
61 are a good indicator that such a problem exists. The host files /proc/iomem
62 and /proc/ioports are often useful for identifying drivers consuming ranges
63 of the device to cause such conflicts.
64
65 Additionally, IGD device are known to generate small numbers of DMAR faults
66 when initially assigned. It is believed that this is simply the IGD attempting
67 to access the reserved GTT space after reset, which it no longer has access to
68 when accessed from userspace. So long as the DMAR faults are small in number
69 and most importantly, not ongoing, these are not an indication of an error.
70
71 Additionally++, analog VGA output (as opposed to digital outputs like HDMI,
72 DVI, or DisplayPort) may be unsupported in some use cases. In the author's
73 experience, even DP to VGA adapters can be troublesome while adapters between
74 digital formats work well.
75
76 Usage
77 =====
78 The intention is for IGD assignment to be transparent for users and thus for
79 management tools like libvirt. To make use of legacy mode, simply remove all
80 other graphics options and use "-nographic" and either "-vga none" or
81 "-nodefaults", along with adding the device using vfio-pci:
82
83 -device vfio-pci,host=00:02.0,id=hostdev0,bus=pci.0,addr=0x2
84
85 For UPT mode, retain the default emulated graphics and simply add the vfio-pci
86 device making use of any other bus address other than 02.0. libvirt will
87 default to assigning the device a UPT compatible address while legacy mode
88 users will need to manually edit the XML if using a tool like virt-manager
89 where the VM device address is not expressly specified.
90
91 An experimental vfio-pci option also exists to enable OpRegion, and thus
92 external monitor support, for UPT mode. This can be enabled by adding
93 "x-igd-opregion=on" to the vfio-pci device options for the IGD device. As
94 with legacy mode, this requires the host to support features introduced in
95 the v4.6 kernel. If Intel chooses to embrace this support, the option may
96 be made non-experimental in the future, opening it to libvirt support.
97
98 Developer ABI
99 =============
100 Legacy mode IGD support imposes two fw_cfg requirements on the VM firmware:
101
102 1) "etc/igd-opregion"
103
104 This fw_cfg file exposes the OpRegion for the IGD device. A reserved
105 region should be created below 4GB (recommended 4KB alignment), sized
106 sufficient for the fw_cfg file size, and the content of this file copied
107 to it. The dword based address of this reserved memory region must also
108 be written to the ASLS register at offset 0xFC on the IGD device. It is
109 recommended that firmware should make use of this fw_cfg entry for any
110 PCI class VGA device with Intel vendor ID. Multiple of such devices
111 within a VM is undefined.
112
113 2) "etc/igd-bdsm-size"
114
115 This fw_cfg file contains an 8-byte, little endian integer indicating
116 the size of the reserved memory region required for IGD stolen memory.
117 Firmware must allocate a reserved memory below 4GB with required 1MB
118 alignment equal to this size. Additionally the base address of this
119 reserved region must be written to the dword BDSM register in PCI config
120 space of the IGD device at offset 0x5C. As this support is related to
121 running the IGD ROM, which has other dependencies on the device appearing
122 at guest address 00:02.0, it's expected that this fw_cfg file is only
123 relevant to a single PCI class VGA device with Intel vendor ID, appearing
124 at PCI bus address 00:02.0.
125
126 Footnotes
127 =========
128 [1] Nothing precludes adding additional emulated or assigned graphics devices
129 as non-primary, other than the combination typically not working. I only
130 intend to set user expectations, others are welcome to find working
131 combinations or fix whatever issues prevent this from working in the common
132 case.
133 [2] # echo "vfio-pci" > /sys/bus/pci/devices/0000:00:02.0/driver_override