This document covers the command line options which the Xen Hypervisor.
Most parameters take the form option=value
. Different
options on the command line should be space delimited. All options are
case sensitive, as are all values unless explicitly noted.
<boolean>
)All boolean option may be explicitly enabled using a
value
of > yes
, on
,
true
, enable
or 1
They may be explicitly disabled using a value
of >
no
, off
, false
,
disable
or 0
In addition, a boolean option may be enabled by simply stating its
name, and may be disabled by prefixing its name with
no-
.
####Examples
Enable noreboot mode > noreboot=true
Disable x2apic support (if present) > x2apic=off
Enable synchronous console mode > sync_console
Explicitly specifying any value other than those listed above is
undefined, as is stacking a no-
prefix with an explicit
value.
<integer>
)An integer parameter will default to decimal and may be prefixed with
a -
for negative numbers. Alternatively, a hexadecimal
number may be used by prefixing the number with 0x
, or an
octal number may be used if a leading 0
is present.
Providing a string which does not validly convert to an integer is undefined.
<size>
)A size parameter may be any integer, with a single size suffix
T
or t
: TiB (2^40)G
or g
: GiB (2^30)M
or m
: MiB (2^20)K
or k
: KiB (2^10)B
or b
: BytesWithout a size suffix, the default will be kilo. Providing a suffix other than those listed above is undefined.
Many parameters are more complicated and require more intricate configuration. The detailed description of each individual parameter specify which values are valid.
Some options take a comma separated list of values.
Some parameters act as combinations of the above, most commonly a mix of Boolean and String. These are noted in the relevant sections.
= force | ht | noirq | <boolean> | verbose
String, or Boolean to disable.
By default, Xen will scan the DMI data and blacklist certain systems
which are known to have broken ACPI setups. Providing
acpi=force
will cause Xen to ignore the blacklist and
attempt to use all ACPI features.
Using acpi=ht
causes Xen to parse the ACPI tables enough
to enumerate all CPUs, but will not use other ACPI features. This is not
common, and only has an effect if your system is blacklisted.
The acpi=noirq
option causes Xen to not parse the ACPI
MADT table looking for IO-APIC entries. This is also not common, and any
system which requires this option to function should be blacklisted.
Additionally, this will not prevent Xen from finding IO-APIC entries
from the MP tables.
Further, any of the boolean false options can be used to disable ACPI usage entirely.
Because responsibility for ACPI processing is shared between Xen and the domain 0 kernel this option is automatically propagated to the domain 0 command line.
Finally, acpi=verbose
will enable per-processor
information logging which may otherwise be too noisy in particular on
large systems.
= <integer>
Specify which ACPI MADT table to parse for APIC information, if more than one is present.
= <boolean>
Default:
false
Enforce checking that P-state transitions by the ACPI cpufreq driver actually result in the nominated frequency to be established. A warning message will be logged if that isn’t the case.
= <boolean>
Instruct Xen to ignore timer-interrupt override.
= s3_bios | s3_mode
s3_bios
instructs Xen to invoke video BIOS
initialization during S3 resume.
s3_mode
instructs Xen to set up the boot time (option
vga=
) video mode during S3 resume.
= <boolean>
Default:
false
Force boot on potentially unsafe systems. By default Xen will refuse to boot on systems with the following errata:
= <boolean>
Default:
false
Permit multiple copies of host p2m.
= bigsmp | default
Override Xen’s logic for choosing the APIC driver. By default, if
there are more than 8 CPUs, Xen will switch to bigsmp
over
default
.
= <boolean>
Default:
true
Permit Xen to use APIC Virtualisation Extensions. This is an optimisation available as part of VT-x, and allows hardware to take care of the guests APIC handling, rather than requiring emulation in Xen.
= verbose | debug
Increase the verbosity of the APIC code from the default value.
= <boolean>
Default:
true
Permit Xen to use “Always Running APIC Timer” support on compatible hardware in combination with cpuidle. This option is only expected to be useful for developers wishing Xen to fall back to older timing methods on newer hardware.
= List of [ <bool>, mac-permissive=<bool> ]
Controls for the Argo hypervisor-mediated interdomain communication service.
The functionality that this option controls is only available when Xen has been compiled with the build setting for Argo enabled in the build configuration.
Argo is a interdomain communication mechanism, where Xen acts as the central point of authority. Guests may register memory rings to recieve messages, query the status of other domains, and send messages by hypercall, all subject to appropriate auditing by Xen. Argo is disabled by default.
The mac-permissive
boolean controls whether wildcard
receive rings may be registered (mac-permissive=1
) or may
not be registered (mac-permissive=0
).
This option is disabled by default, to protect domains from a DoS by a buggy or malicious other domain spamming the ring.
= <boolean>
Default:
true
Permit Xen to use Address Space Identifiers. This is an optimisation which tags the TLB entries with an ID per vcpu. This allows for guest TLB flushes to be performed without the overhead of a complete TLB flush.
= <boolean>
Default:
false
Forces all CPUs’ full state to be logged upon certain fatal asynchronous exceptions (watchdog NMIs and unexpected MCEs).
= <boolean>
Default:
false
Permits Xen to set up and use PCI Address Translation Services. This is a performance optimisation for PCI Passthrough.
WARNING: Xen cannot currently safely use ATS because of its synchronous wait loops for Queued Invalidation completions.
= <size>
Default:
0
(no limit)
Specify a maximum amount of available memory, to which Xen will clamp the e820 table.
= List of [ <integer> | <integer>-<integer> ]
Specify that certain pages, or certain ranges of pages contain bad
bytes and should not be used. For example, if your memory tester says
that byte 0x12345678
is bad, you would place
badpage=0x12345
on Xen’s command line.
= idle | <boolean>
Default:
idle
Scrub free RAM during boot. This is a safety feature to prevent accidentally leaking sensitive VM data into other VMs if Xen crashes and reboots.
In idle
mode, RAM is scrubbed in background on all CPUs
during idle-loop with a guarantee that memory allocations always provide
scrubbed pages. This option reduces boot time on machines with a large
amount of RAM while still providing security benefits.
= <size>
Default:
128M
Maximum RAM block size chunks to be scrubbed whilst holding the page heap lock and not running softirqs. Reduce this if softirqs are not being run frequently enough. Setting this to a high value may cause boot failure, particularly if the NMI watchdog is also enabled.
= <size>
Default:
64M
Amount of memory reserved for the buddy allocator when colored
allocator is active. This option is available only when
CONFIG_LLC_COLORING
is enabled. The colored allocator is
meant as an alternative to the buddy allocator, because its allocation
policy is by definition incompatible with the generic one. Since the Xen
heap systems is not colored yet, we need to support the coexistence of
the two allocators for now. This parameter, which is optional and for
expert only, it’s used to set the amount of memory reserved to the buddy
allocator.
= List of [ <bool>, shstk=<bool>, ibt=<bool> ]
Applicability: x86
Controls for the use of Control-flow Enforcement Technology. CET is group a of hardware features designed to combat Return-oriented Programming (ROP, also call/jmp COP/JOP) attacks.
CET is incompatible with 32bit PV guests. If any CET sub-options are
active, they will override the pv=32
boolean to
false
. Backwards compatibility can be maintained with the
pv-shim mechanism.
An unqualified boolean is a shorthand for setting all suboptions at once.
The shstk=
boolean controls whether Xen uses Shadow
Stacks for its own protection.
The option is available when CONFIG_XEN_SHSTK
is
compiled in, and generally defaults to true
on hardware
supporting CET-SS. Specifying cet=no-shstk
will cause Xen
not to use Shadow Stacks even when support is available in hardware.
Some hardware suffers from an issue known as Supervisor Shadow Stack
Fracturing. On such hardware, Xen will default to not using Shadow
Stacks when virtualised. Specifying cet=shstk
will override
this heuristic and enable Shadow Stacks unilaterally.
The ibt=
boolean controls whether Xen uses Indirect
Branch Tracking for its own protection.
The option is available when CONFIG_XEN_IBT
is compiled
in, and defaults to true
on hardware supporting CET-IBT.
Specifying cet=no-ibt
will cause Xen not to use Indirect
Branch Tracking even when support is available in hardware.
= pit | hpet | acpi | tsc
If set, override Xen’s default choice for the platform timer. Having TSC as platform timer requires being explicitly set. This is because TSC can only be safely used if CPU hotplug isn’t performed on the system. On some platforms, the “maxcpus” option may need to be used to further adjust the number of allowed CPUs. When running on platforms that can guarantee a monotonic TSC across sockets you may want to adjust the “tsc” command line parameter to “stable:socket”.
= <integer>
Default:
2
Specify the event count threshold for raising Corrected Machine Check Interrupts. Specifying zero disables CMCI handling.
= <boolean>
Default:
false
Flag to indicate whether to probe for a CMOS Real Time Clock irrespective of ACPI indicating none to be there.
= <baud>[/<base-baud>][,[DPS][,[<io-base>|pci|amt][,[<irq>|msi][,[<port-bdf>][,[<bridge-bdf>]]]]]]
Both option com1
and com2
follow the same
format.
<baud>
may be either an integer baud rate, or the
string auto
if the bootloader or other earlier firmware has
already set it up.DPS
represents the number of data bits, the parity, and
the number of stop bits.
D
is an integer between 5 and 8 for the number of data
bits.P
is a single character representing the type of
parity:
n
Noo
Odde
Evenm
Marks
SpaceS
is an integer 1 or 2 for the number of stop
bits.<io-base>
is an integer which specifies the IO
base port for UART registers.<irq>
is the IRQ number to use, or 0
to use the UART in poll mode only, or msi
to set up a
Message Signaled Interrupt.<port-bdf>
is the PCI location of the UART, in
<bus>:<device>.<function>
notation.<bridge-bdf>
is the PCI bridge behind which is
the UART, in <bus>:<device>.<function>
notation.pci
indicates that Xen should scan the PCI bus for the
UART, avoiding Intel AMT devices.amt
indicated that Xen should scan the PCI bus for the
UART, including Intel AMT devices if present.A typical setup for most situations might be
com1=115200,8n1
In addition to the above positional specification for UART parameters, name=value pair specfications are also supported. This is used to add flexibility for UART devices which require additional UART parameter configurations.
The comma separation still delineates positional parameters. Hence, unless the parameter is explicitly specified with name=value option, it will be considered a positional parameter.
The syntax consists of com1=(comma-separated positional parameters),(comma separated name-value pairs)
The accepted name keywords for name=value pairs are:
baud
- accepts integer baud rate (eg. 115200) or
auto
bridge
- Similar to bridge-bdf in positional parameters.
Used to determine the PCI bridge to access the UART device. Notation is
xx:xx.x <bus>:<device>.<function>
clock-hz
- accepts large integers to setup UART clock
frequencies. Do note - these values are multiplied by 16.data-bits
- integer between 5 and 8dev
- accepted values are pci
OR
amt
. If this option is used to specify if the serial device
is pci-based. The io_base cannot be specified when dev=pci
or dev=amt
is used.io-base
- accepts integer which specified IO base port
for UART registersirq
- IRQ number to useparity
- accepted values are same as positional
parametersport
- Used to specify which port the PCI serial device
is located on Notation is xx:xx.x
<bus>:<device>.<function>
reg-shift
- register shifts required to set UART
registersreg-width
- register width required to set UART
registers (only accepts 1 and 4)stop-bits
- only accepts 1 or 2 for the number of stop
bitsThe following are examples of correct specifications:
com1=115200,8n1,0x3f8,4
com1=115200,8n1,0x3f8,4,reg-width=4,reg-shift=2
com1=baud=115200,parity=n,stop-bits=1,io-base=0x3f8,reg-width=4
= <size>
Default:
conring_size=16k
Specify the size of the console ring buffer.
= List of [ vga | com1[H,L] | com2[H,L] | pv | dbgp | ehci | xhci | none ]
Default:
console=com1,vga
Specify which console(s) Xen should use.
vga
indicates that Xen should try and use the vga
graphics adapter.
com1
and com2
indicates that Xen should use
serial ports 1 and 2 respectively. Optionally, these arguments may be
followed by an H
or L
. H
indicates that transmitted characters will have their MSB set, while
received characters must have their MSB set. L
indicates
the converse; transmitted and received characters will have their MSB
cleared. This allows a single port to be shared by two subsystems
(e.g. console and debugger).
pv
indicates that Xen should use Xen’s PV console. This
option is only available when used together with
pv-in-pvh
.
dbgp
or ehci
indicates that Xen should use
a USB2 debug port.
xhci
indicates that Xen should use a USB3 debug
port.
none
indicates that Xen should not use a console. This
option only makes sense on its own.
= none | date | datems | boot | raw
Default:
none
Can be modified at runtime
Specify which timestamp format Xen should use for each console line.
none
: No timestampsdate
: Date and time information
[YYYY-MM-DD HH:MM:SS]
datems
: Date and time, with milliseconds
[YYYY-MM-DD HH:MM:SS.mmm]
boot
: Seconds and microseconds since boot
[SSSSSS.uuuuuu]
raw
: Raw platform ticks, architecture and
implementation dependent
[XXXXXXXXXXXXXXXX]
For compatibility with the older boolean parameter, specifying
console_timestamps
alone will enable the date
option.
= <boolean>
Default:
false
Flag to indicate whether all guest console output should be copied into the console ring buffer.
= <switch char>[x]
Default:
conswitch=a
Can be modified at runtime
Specify which character should be used to switch serial input between Xen and dom0. The required sequence is CTRL-<switch char> three times.
The optional trailing x
indicates that Xen should not
automatically switch the console input to dom0 during boot. Any other
value, including omission, causes Xen to automatically switch to the
dom0 console during dom0 boot. Use conswitch=ax
to keep the
default switch character, but for xen to keep the console.
= power | performance
Default:
power
= arch_perfmon
If set, force use of the performance counters for oprofile, rather than detecting available support.
= none | {{ <boolean> | xen } { [:[powersave|performance|ondemand|userspace][,[<maxfreq>]][,[<minfreq>]]] } [,verbose]} | dom0-kernel | hwp[:[<hdc>][,verbose]]
Default:
xen
Indicate where the responsibility for driving power states lies. Note
that the choice of dom0-kernel
is deprecated and not
supported by all Dom0 kernels.
<maxfreq>
and <minfreq>
are
integers which represent max and min processor frequencies
respectively.verbose
option can be included as a string or also as
verbose=<integer>
for xen
. It is a
boolean for hwp
.hwp
selects Hardware-Controlled Performance States
(HWP) on supported Intel hardware. HWP is a Skylake+ feature which
provides better CPU power management. The default is disabled. If
hwp
is selected, but hardware support is not available, Xen
will fallback to cpufreq=xen.<hdc>
is a boolean to enable Hardware Duty
Cycling (HDC). HDC enables the processor to autonomously force physical
package components into idle state. The default is enabled, but the
option only applies when hwp
is enabled.There is also support for ;
-separated fallback options:
cpufreq=hwp;xen,verbose
. This first tries hwp
and falls back to xen
if unavailable. Note: The
verbose
suboption is handled globally. Setting it for
either the primary or fallback option applies to both irrespective of
where it is specified.
Note: grub2 requires to escape or quote ‘;’, so
"cpufreq=hwp;xen"
should be specified within double quotes
inside grub.cfg. Refer to the grub2 documentation for more
information.
= List of comma separated booleans
This option allows for fine tuning of the facilities Xen will use, after accounting for hardware capabilities as enumerated via CPUID.
Unless otherwise noted, options only have any effect in their negative form, to hide the named feature(s). Ignoring a feature using this mechanism will cause Xen not to use the feature, nor offer them as usable to guests.
Currently accepted:
The Speculation Control hardware features srbds-ctrl
,
md-clear
, ibrsb
, stibp
,
ibpb
, l1d-flush
and ssbd
are used
by default if available and applicable. They can all be ignored.
rdrand
and rdseed
have multiple
interactions.
For Special Register Buffer Data Sampling (SRBDS, XSA-320, CVE-2020-0543), RDRAND and RDSEED can be ignored.
Due to the absence of microcode to address SRBDS on IvyBridge client
hardware, the RDRAND feature is hidden by default for guests, unless
rdrand
is used in its positive form. Irrespective of the
setting here, VMs can use RDRAND if explicitly enabled in guest config
file, and VMs already using RDRAND can migrate in.
The RDRAND feature is disabled by default on AMD Fam15/16
systems, due to possible malfunctions after ACPI S3 suspend/resume.
rdrand
may be used in its positive form to override Xen’s
default behaviour on these systems, and make the feature fully
usable.
= fam_0f_rev_[cdefg] | fam_10_rev_[bc] | fam_11_rev_b
Applicability: AMD
If none of the other cpuid_mask_* options are given, Xen has a set of pre-configured masks to make the current processor appear to be family/revision specified.
See below for general information on masking.
Warning: This option is not fully effective on Family 15h processors or later.
= <integer>
Applicability: x86. Default:
~0
(all bits set)
The availability of these options are model specific. Some processors don’t support any of them, and no processor supports all of them. Xen will ignore options on processors which are lacking support.
These options can be used to alter the features visible via the
CPUID
instruction. Settings applied here take effect
globally, including for Xen and all guests.
Note: Since Xen 4.7, it is no longer necessary to mask a host to create migration safety in heterogeneous scenarios. All necessary CPUID settings should be provided in the VM configuration file. Furthermore, it is recommended not to use this option, as doing so causes an unnecessary reduction of features at Xen’s disposal to manage guests.
= <boolean>
= <boolean>
= <string>
Can be modified at runtime
Specify debug-key actions in cases of crashes. Each of the parameters
applies to a different crash reason. The <string>
is
a sequence of debug key characters, with +
having the
special meaning of a 10 millisecond pause.
crash-debug-debugkey
will be used for crashes induced by
the C
debug key (i.e. manually induced crash).
crash-debug-hwdom
denotes a crash of dom0.
crash-debug-kexeccmd
is an explicit request of dom0 to
continue with the kdump kernel via kexec. Only available on hypervisors
built with CONFIG_KEXEC.
crash-debug-panic
is a crash of the hypervisor.
crash-debug-watchdog
is a crash due to the watchdog
timer expiring.
It should be noted that dumping diagnosis data to the console can fail in multiple ways (missing data, hanging system, …) depending on the reason of the crash, which might have left the hypervisor in a bad state. In case a debug-key action leads to another crash recursion will be avoided, so no additional debug-key actions will be performed in this case. A crash in the early boot phase will not result in any debug-key action, as the system might not yet be in a state where the handlers can work.
So e.g. crash-debug-watchdog=0+0r
would dump dom0 state
twice with 10 milliseconds between the two state dumps, followed by the
run queues of the hypervisor, if the system crashes due to a watchdog
timeout.
Depending on the reason of the system crash it might happen that triggering some debug key action will result in a hang instead of dumping data and then doing a reboot or crash dump.
= <size>
Default:
4G
Specify the maximum address to allocate certain structures, if used in combination with the low_crashinfo command line option.
= <ramsize-range>:<size>[,...][{@,<}<offset>]
= <size>[{@,<}<offset>]
= <size>,below=offset
Specify sizes and optionally placement of the crash kernel
reservation area. The <ramsize-range>:<size>
pairs indicate how much memory to set aside for a crash kernel
(<size>
) for a given range of installed RAM
(<ramsize-range>
). Each
<ramsize-range>
is of the form
<start>-[<end>]
.
A trailing @<offset>
specifies the exact address
this area should be placed at, whereas <
in place of
@
just specifies an upper bound of the address range the
area should fall into.
< and below are synonyomous, the latter being useful for grub2 systems which would otherwise require escaping of the < option
= <integer>
= <integer>
= <integer>
Default:
10
Domains subject to a cap receive a replenishment of their runtime budget once every cap period interval. Default is 10 ms. The amount of budget they receive depends on their cap. For instance, a domain with a 50% cap will receive 50% of 10 ms, so 5 ms.
= <integer>
Default:
18
Specify the number of bits to use for the fractional part of the values involved in Credit2 load tracking and load balancing math.
= <integer>
Default:
30
Specify the number of bits to use to represent the length of the window (in nanoseconds) we use for load tracking inside Credit2. This means that, with the default value (30), we use 2^30 nsec ~= 1 sec long window.
Load tracking is done by means of a variation of exponentially weighted moving average (EWMA). The window length defined here is what tells for how long we give value to previous history of the load itself. In fact, after a full window has passed, what happens is that we discard all previous history entirely.
A short window will make the load balancer quick at reacting to load changes, but also short-sighted about previous history (and hence, e.g., long term load trends). A long window will make the load balancer thoughtful of previous history (and hence capable of capturing, e.g., long term load trends), but also slow in responding to load changes.
The default value of 1 sec
is rather long.
= cpu | core | socket | node | all
Default:
socket
Specify how host CPUs are arranged in runqueues. Runqueues are kept
balanced with respect to the load generated by the vCPUs running on
them. Smaller runqueues (as in with core
) means more
accurate load balancing (for instance, it will deal better with
hyperthreading), but also more overhead.
Available alternatives, with their meaning, are: * cpu
:
one runqueue per each logical pCPUs of the host; * core
:
one runqueue per each physical core of the host; * socket
:
one runqueue per each physical socket (which often, but not always,
matches a NUMA node) of the host; * node
: one runqueue per
each NUMA node of the host; * all
: just one runqueue shared
by all the logical pCPUs of the host
Regardless of the above choice, Xen attempts to respect
sched_credit2_max_cpus_runqueue
limit, which may mean more
than one runqueue for the all
value. If that isn’t
intended, raise the sched_credit2_max_cpus_runqueue
value.
= ehci[ <integer> | @pci<bus>:<slot>.<func> ]
= xhci[ <integer> | @pci<bus>:<slot>.<func> ][,share=<bool>|hwdom]
Specify the USB controller to use, either by instance number (when going over the PCI busses sequentially) or by PCI device (must be on segment 0).
Use ehci
for EHCI debug port, use xhci
for
XHCI debug capability. XHCI driver will wait indefinitely for the debug
host to connect - make sure the cable is connected. The
share
option for xhci controls who else can use the
controller: * no
: use the controller exclusively for
console, even hardware domain (dom0) cannot use it * hwdom
:
hardware domain may use the controller too, ports not used for debug
console will be available for normal devices; this is the default *
yes
: the controller can be assigned to any domain; it is
not safe to assign the controller to untrusted domain
Choosing share=hwdom
(the default) or
share=yes
allows a domain to reset the controller, which
may cause small portion of the console output to be lost.
The share=yes
configuration is not security
supported.
= <integer>
Default:
20
Limits the number lines printed in Xen stack traces.
= [cpu:]<size>
Default:
128
Specify the size of the console debug trace buffer. By specifying
cpu:
additionally a trace buffer of the specified size is
allocated per cpu. The debug trace feature is only enabled in debugging
builds of Xen.
= <boolean>
Default:
CONFIG_DIT_DEFAULT
Specify whether Xen and guests should operate in Data Independent Timing mode (Intel calls this DOITM, Data Operand Independent Timing Mode). Note that enabling this option cannot guarantee anything beyond what underlying hardware guarantees (with, where available and known to Xen, respective tweaks applied).
= <integer>
Specify the bit width of the DMA heap.
= List of [ pv | pvh, shadow=<bool>, verbose=<bool>,
cpuid-faulting=<bool>, msr-relaxed=<bool> ] (x86)
= List of [ sve=<integer> ] (Arm64)
Controls for how dom0 is constructed on x86 systems.
The pv
and pvh
options select the
virtualisation mode of dom0.
The pv
option is only available when
CONFIG_PV
is compiled in. The pvh
option is
only available when CONFIG_HVM
is compiled in. When both
options are compiled in, the default is PV.
In addition, the following requirements must be met:
The shadow
boolean allows dom0 to be explicitly
constructed using shadow paging. This option is unavailable when
CONFIG_SHADOW_PAGING
is disabled.
For PVH, dom0 defaults to using HAP on capable hardware, and falls back to shadow paging otherwise. A PVH dom0 cannot be used if Xen is compiled without shadow paging support, and the hardware lacks HAP support.
For PV, the use of dom0 shadow mode is only for development purposes. PV guests do no require any paging support by default.
The verbose
boolean is intended for diagnostics, and
prints out extra information during the dom0 build. It defaults to the
compile time choice of CONFIG_VERBOSE_DEBUG
.
The cpuid-faulting
boolean is an interim option, is
only applicable to PV dom0, and defaults to true.
Before Xen 4.13, the domain builder logic for guest construction depended on seeing host CPUID values to function correctly. As a result, CPUID Faulting was never activated for PV dom0’s, even on capable hardware.
In Xen 4.13, the domain builder logic has been fixed, and no longer has this dependency. As a consequence, CPUID Faulting is activated by default even for PV dom0’s.
However, as PV dom0’s have always seen host CPUID data in the past,
there is a chance that further dependencies exist. This boolean can be
used to restore the pre-4.13 behaviour. If specifying
no-cpuid-faulting
fixes an issue in dom0, please report a
bug.
The msr-relaxed
boolean is an interim option, and
defaults to false.
In Xen 4.15, the default behaviour for unhandled MSRs has been changed, to avoid leaking host data into guests, and to avoid breaking guest logic which uses #GP probing to identify the availability of MSRs.
However, this new stricter behaviour has the possibility to break
guests, and a more 4.14-like behaviour can be selected by specifying
dom0=msr-relaxed
.
If using this option is necessary to fix an issue, please report a bug.
Enables features on dom0 on Arm systems.
sve
integer parameter enables Arm SVE usage for
Dom0 and sets the maximum SVE vector length, the option is applicable
only to Arm64 Dom0 kernels. A value equal to 0 disables the feature,
this is the default value. Values below 0 means the feature uses the
maximum SVE vector length supported by hardware, if SVE is supported.
Values above 0 explicitly set the maximum SVE vector length for Dom0,
allowed values are from 128 to maximum 2048, being multiple of 128.
Please note that when the user explicitly specifies the value, if that
value is above the hardware supported maximum SVE vector length, the
domain creation will fail and the system will stop, the same will occur
if the option is provided with a positive non zero value, but the
platform doesn’t support SVE.= List of comma separated booleans
Applicability: x86
This option allows for fine tuning of the facilities dom0 will use, after accounting for hardware capabilities and Xen settings as enumerated via CPUID.
Options are accepted in positive and negative form, to enable or disable specific features. All selections via this mechanism are subject to normal CPU Policy safety and dependency logic.
This option is intended for developers to opt dom0 into non-default features, and is not intended for use in production circumstances. If using this option is necessary to fix an issue, please report a bug.
= List of [ passthrough=<bool>, strict=<bool>, map-inclusive=<bool>,
map-reserved=<bool>, none ]
Controls for the dom0 IOMMU setup.
The passthrough
boolean controls whether IOMMU
translation functionality is disabled for devices in dom0
(passthrough=1
) or whether the IOMMU is used to ensure that
dom0 can only DMA to its permitted areas of RAM
(passthrough=0
).
This option is only applicable to x86 PV dom0’s, and defaults to false.
Some older Intel VT-d hardware isn’t capable of disabling translation functionality on a per-device basis, and will cause this option to be ignored and assumed to be 0. Similar behaviour on such systems is only available by fully disabling all IOMMUs.
This option is hardwired to false for x86 PVH dom0’s (where a non-identity transform is required for dom0 to function), and is ignored for ARM.
The strict
boolean is applicable to x86 PV dom0’s
only and defaults to false. It controls whether dom0 can have IOMMU
mappings for all domain RAM in the system, or only for its allocated RAM
(and grant mappings etc.)
This option is hardwired to true for x86 PVH dom0’s (as RAM belonging to other domains in the system don’t live in a compatible address space), and is ignored for ARM.
The map-inclusive
boolean is applicable to x86 PV
dom0’s, and sets up identity IOMMU mappings for all non-RAM regions
below 4GB except for unusable ranges, and ranges belonging to Xen.
Typically, some devices in a system use bits of RAM for communication, and these areas should be listed as reserved in the E820 table and identified via RMRR or IVMD entries in the ACPI tables, so Xen can ensure that they are identity-mapped in the IOMMU. However, some firmware makes mistakes, and this option is a coarse-grain workaround for those errors.
Where possible, finer grain corrections should be made with the
rmrr=
, ivmd=
, ivrs_hpet[]=
, or
ivrs_ioapic[]=
command line options.
This option is disabled by default, and deprecated and intended for
removal in future versions of Xen. If specifying
map-inclusive
is the only way to make your system boot,
please report a bug.
The map-reserved
functionality is very similar to
map-inclusive
.
The differences from map-inclusive
are that
map-reserved
is applicable to both x86 PV and PVH dom0’s,
is enabled by default, and represents a subset of the correction by only
mapping reserved memory regions rather than all non-RAM
regions.
The none
option is intended for development purposes
only, and skips certain safety checks pertaining to the correct IOMMU
configuration for dom0 to boot.
Incorrect use of this option may result in a malfunctioning system.
= List of <hex>-<hex>
Specify a list of IO ports to be excluded from dom0 access.
= List of [ <integer> | <integer>-<integer> ]
Default:
All available LLC colors
Specify dom0 LLC color configuration. This option is available only
when CONFIG_LLC_COLORING
is enabled. If the parameter is
not set, all available colors are used.
Either:
= <integer>
.
The number of VCPUs to give to dom0. This number of VCPUs can be more than the number of PCPUs on the host. The default is the number of PCPUs.
Or:
= <min>-<max>
where<min>
and<max>
are integers.
Gives dom0 a number of VCPUs equal to the number of PCPUs, but always
at least <min>
and no more than
<max>
. Using <min>
may give more
VCPUs than PCPUs. <min>
or <max>
may be omitted and the defaults of 1 and unlimited respectively are used
instead.
For example, with dom0_max_vcpus=4-8
:
Number of PCPUs | Dom0 VCPUs 2 | 4 4 | 4 6 | 6 8 | 8 10 | 8
= <size>
Set the amount of memory for the initial domain (dom0). It must be greater than zero. This parameter is required.
= List of ( min:<sz> | max:<sz> | <sz> )
Set the amount of memory for the initial domain (dom0). If a size is positive, it represents an absolute value. If a size is negative, it is subtracted from the total available memory.
<sz>
specifies the exact amount of memory.min:<sz>
specifies the minimum amount of
memory.max:<sz>
specifies the maximum amount of
memory.If <sz>
is not specified, the default is all the
available memory minus some reserve. The reserve is 1/16 of the
available memory or 128 MB (whichever is smaller).
The amount of memory will be at least the minimum but never more than
the maximum (i.e., max
overrides the min
option). If there isn’t enough memory then as much as possible is
allocated.
max:<sz>
also sets the maximum reservation (the
maximum amount of memory dom0 can balloon up to). If this is omitted
then the maximum reservation is unlimited.
For example, to set dom0’s initial memory allocation to 512MB but
allow it to balloon up as far as 1GB use
dom0_mem=512M,max:1G
<sz>
is:<size> | [<size>+]<frac>%
<frac>
is an integer < 100
<frac>
specifies a fraction of host memory size
in percent.So <sz>
being 1G+25%
on a 256 GB host
would result in 65 GB.
If you use this option then it is highly recommended that you disable any dom0 autoballooning feature present in your toolstack. See the xl.conf(5) man page or Xen Best Practices.
This option doesn’t have effect if pv-shim mode is enabled.
= List of [ <integer> | relaxed | strict ]
Default:
strict
Specify the NUMA nodes to place Dom0 on. Defaults for vCPU-s created
and memory assigned to Dom0 will be adjusted to match the node
restrictions set up here. Note that the values to be specified here are
ACPI PXM ones, not Xen internal node numbers. relaxed
sets
up vCPU affinities to prefer but be not limited to the specified
node(s).
= <boolean>
Default:
false
Pin dom0 vcpus to their respective pcpus
= path [:options]
Default:
""
Specify the full path in the device tree for the UART. If the path
doesn’t start with /
, it is assumed to be an alias. The
options are device specific.
= <boolean>
Flag that specifies if RAM should be clipped to the highest cacheable MTRR.
Default:
true
on Intel CPUs, otherwisefalse
= <boolean>
Default:
false
Flag that enables verbose output when processing e820 information and applying clipping.
= off | on | skipmbr
Control retrieval of Extended Disc Data (EDD) from the BIOS during boot.
= no | force
Either force retrieval of monitor EDID information via VESA DDC, or disable it (edid=no). This option should not normally be required except for debugging purposes.
= List of [ rs=<bool>, attr=no|uc ]
Controls for interacting with the system Extended Firmware Interface.
The rs
boolean controls whether Runtime Services are
used. By default, Xen uses Runtime Services itself, and proxies certain
calls on behalf of dom0. Selecting rs=0
prohibits all use
of Runtime Services.
The attr=
string exists to specify what to do with
memory regions of unknown/unrecognised cacheability.
attr=no
is the default and will leave the memory regions
unmapped, while attr=uc
will map them as fully
uncacheable.
= List of [ ad=<bool>, pml=<bool>, exec-sp=<bool> ]
Applicability: Intel
Extended Page Tables are a feature of Intel’s VT-x technology, whereby hardware manages the virtualisation of HVM guest pagetables. EPT was introduced with the Nehalem architecture.
The ad
boolean controls hardware tracking of Access
and Dirty bits in the EPT pagetables, and was first introduced in
Broadwell Server.
By default, Xen will use A/D tracking when available in hardware,
except on Avoton processors affected by erratum AVR41. Explicitly
choosing ad=0
will disable the use of A/D tracking on
capable hardware, whereas choosing ad=1
will cause tracking
to be used even on AVR41-affected hardware.
The pml
boolean controls the use of Page
Modification Logging, which is also introduced in Broadwell Server.
PML is a feature whereby the processor generates a list of pages which have been dirtied. This is necessary information for operations such as live migration, and having the processor maintain the list of dirtied pages is more efficient than traditional software implementations where all guest writes trap into Xen so the dirty bitmap can be maintained.
By default, Xen will use PML when it is available in hardware. PML
functionally depends on A/D tracking, so choosing ad=0
will
implicitly disable PML. pml=0
can be used to prevent the
use of PML on otherwise capable hardware.
The exec-sp
boolean controls whether EPT superpages
with execute permissions are permitted. In general this is good for
performance.
However, on processors vulnerable CVE-2018-12207, HVM guest kernels can use executable superpages to crash the host. By default, executable superpages are disabled on affected hardware.
If HVM guest kernels are trusted not to mount a DoS against the system, this option can enabled to regain performance.
This boolean may be modified at runtime using
xl set-parameters ept=[no-]exec-sp
to switch between fast
and secure.
When switching from secure to fast, preexisting HVM domains will run at their current performance until they are rebooted; new domains will run without any overhead.
When switching from fast to secure, all HVM domains will immediately suffer a performance penalty.
Warning: No guarantee is made that this runtime option will be retained indefinitely, or that it will retain this exact behaviour. It is intended as an emergency option for people who first chose fast, then change their minds to secure, and wish not to reboot.
= [<domU number>][,<dom0 number>]
Default:
32,<variable>
Change the number of PIRQs available for guests. The optional first
number is common for all domUs, while the optional second number
(preceded by a comma) is for dom0. Changing the setting for domU has no
impact on dom0 and vice versa. For example to change dom0 without
changing domU, use extra_guest_irqs=,512
. The default value
for Dom0 and an eventual separate hardware domain is architecture
dependent. The upper limit for both values on x86 is such that the
resulting total number of IRQs can’t be higher than 32768. Note that
specifying zero as domU value means zero, while for dom0 it means to use
the default. Note further that the Dom0 setting has no useful meaning
for the PVH case; use of the option may have an adverse effect there,
though.
= <boolean>
Default :
true
Flag to enable or disable support for extended regions for Dom0 and Dom0less DomUs.
Extended regions are ranges of unused address space exposed to the guest as “safe to use” for special memory mappings. Disable if your board device tree is incomplete.
= permissive | enforcing | late | disabled
Default:
enforcing
Specify how the FLASK security server should be configured. This option is only available if the hypervisor was compiled with FLASK support. This can be enabled by running either: - make -C xen config and enabling XSM and FLASK. - make -C xen menuconfig and enabling ‘FLux Advanced Security Kernel support’ and ‘Xen Security Modules support’
permissive
: This is intended for development and is not
suitable for use with untrusted guests. If a policy is provided by the
bootloader, it will be loaded; errors will be reported to the ring
buffer but will not prevent booting. The policy can be changed to
enforcing mode using “xl setenforce”.enforcing
: This will cause the security server to enter
enforcing mode prior to the creation of domain 0. If an valid policy is
not provided by the bootloader and no built-in policy is present, the
hypervisor will not continue booting.late
: This disables loading of the built-in security
policy or the policy provided by the bootloader. FLASK will be enabled
but will not enforce access controls until a policy is loaded by a
domain using “xl loadpolicy”. Once a policy is loaded, FLASK will run in
enforcing mode unless “xl setenforce” has changed that setting.disabled
: This causes the XSM framework to revert to
the dummy module. The dummy module provides the same security policy as
is used when compiling the hypervisor without support for XSM. The
xsm_op hypercall can also be used to switch to this mode after boot, but
there is no way to re-enable FLASK once the dummy module is loaded.
= <height>
where height is8x8 | 8x14 | 8x16
Specify the font size when using the VESA console driver.
= <boolean>
Default:
false
Allow EPT to be enabled when VMX feature
VM_ENTRY_LOAD_GUEST_PAT
is not present.
Warning: Due to CVE-2013-2212, VMX feature
VM_ENTRY_LOAD_GUEST_PAT
is by default required as a
prerequisite for using EPT. If you are not using PCI Passthrough, or
trust the guest administrator who would be using passthrough, then the
requirement can be relaxed. This option is particularly useful for
nested virtualization, to allow the L1 hypervisor to use EPT even if the
L0 hypervisor does not provide VM_ENTRY_LOAD_GUEST_PAT
.
= List of [ max-ver:<integer>, transitive=<bool>, transfer=<bool> ]
Default (Arm):
gnttab=max-ver:1
Default (x86,PV):gnttab=max-ver:2,transitive,transfer
Default (x86,HVM):gnttab=max-ver:2,transitive
Control various aspects of the grant table behaviour available to guests.
max-ver
Select the maximum grant table version to offer
to guests. Valid version are 1 and 2.transitive
Permit or disallow the use of transitive
grants. Note that the use of grant table v2 without transitive grants is
an ABI breakage from the guests point of view.transfer
Permit or disallow the GNTTABOP_transfer
operation of the grant table hypercall. Note that disallowing
GNTTABOP_transfer is an ABI breakage from the guests point of view. This
option is only available on hypervisors configured to support PV
guests.The usage of gnttab v2 is not security supported on ARM platforms.
= <integer>
Default:
64
Can be modified at runtime
Specify the default upper bound on the number of frames which any domain may use as part of its grant table unless a different value is specified at domain creation.
Note this value is the effective upper bound for dom0.
= <integer>
Default:
1024
Can be modified at runtime
Specify the default upper bound on the number of frames which any domain may use as part of its maptrack array unless a different value is specified at domain creation.
Note this value is the effective upper bound for dom0.
= <boolean>
Applicability: x86
Default: true unless running virtualized on AMD or Hygon hardware
Control whether to use global pages for PV guests, and thus the need to perform TLB flushes by writing to CR4. This is a performance trade-off.
AMD SVM does not support selective trapping of CR4 writes, which means that a global TLB flush (two CR4 writes) takes two VMExits, and massively outweigh the benefit of using global pages to begin with. This case is easy for Xen to spot, and is accounted for in the default setting.
Other cases where this option might be a benefit is on VT-x hardware when selective CR4 writes are not supported/enabled by the hypervisor, or in any virtualised case using shadow paging. These are not easy for Xen to spot, so are not accounted for in the default setting.
= <level>[/<rate-limited level>]
where level isnone | error | warning | info | debug | all
Default:
guest_loglvl=none/warning
Can be modified at runtime
Set the logging level for Xen guests. Any log message with equal more more importance will be printed.
The optional <rate-limited level>
option instructs
which severities should be rate limited.
= <boolean>
Default:
true
Flag to globally enable or disable support for Hardware Assisted Paging (HAP)
= <boolean>
Default:
true
Flag to enable 1 GB host page table support for Hardware Assisted Paging (HAP).
= <boolean>
Default:
true
Flag to enable 2 MB host page table support for Hardware Assisted Paging (HAP).
= <domid>
Default:
0
Enable late hardware domain creation using the specified domain ID. This is intended to be used when domain 0 is a stub domain which builds a disaggregated system including a hardware domain with the specified domain ID. This option is supported only when compiled with XSM on x86.
= <boolean>
Default:
false
Control Xens use of the APEI Hardware Error Source Table, should one be found.
= <size>
Specify the memory boundary past which memory will be treated as highmem (x86 debug hypervisor only).
= <boolean>
Default :
false
Say yes at your own risk if you want to enable heterogenous computing (such as big.LITTLE). This may result to an unstable and insecure platform, unless you manually specify the cpu affinity of all domains so that all vcpus are scheduled on the same class of pcpus (big or LITTLE but not both). vcpu migration between big cores and LITTLE cores is not supported. See docs/misc/arm/big.LITTLE.txt for more information.
When the hmp-unsafe option is disabled (default), CPUs that are not identical to the boot CPU will be parked and not used by Xen.
= List of [ <bool> | broadcast=<bool> | legacy-replacement=<bool> ]
Applicability: x86
Controls Xen’s use of the system’s High Precision Event Timer. By
default, Xen will use an HPET when available and not subject to errata.
Use of the HPET can be disabled by specifying hpet=0
.
The broadcast
boolean is disabled by default, but
forces Xen to keep using the broadcast for CPUs in deep C-states even
when an RTC interrupt is enabled. This then also affects raising of the
RTC interrupt.
The legacy-replacement
boolean allows for control
over whether Legacy Replacement mode is enabled.
Legacy Replacement mode is intended for hardware which does not have an 8254 PIT, and allows the HPET to be configured into a compatible mode. Intel chipsets from Skylake/ApolloLake onwards can turn the PIT off for power saving reasons, and there is no platform-agnostic mechanism for discovering this.
By default, Xen will not change hardware configuration, unless the PIT appears to be absent, at which point Xen will try to enable Legacy Replacement mode before falling back to pre-IO-APIC interrupt routing options.
This behaviour can be inhibited by specifying
legacy-replacement=0
. Alternatively, this mode can be
enabled unconditionally (if available) by specifying
legacy-replacement=1
.
= <boolean>
Deprecated alternative of hpet=broadcast
.
= <integer>
The specified value is a bit mask with the individual bits having the following meaning:
Bit 0 - debug level 0 (unused at present) Bit 1 - debug level 1 (Control Register logging) Bit 2 - debug level 2 (VMX logging of MSR restores when context switching) Bit 3 - debug level 3 (unused at present) Bit 4 - I/O operation logging Bit 5 - vMMU logging Bit 6 - vLAPIC general logging Bit 7 - vLAPIC timer logging Bit 8 - vLAPIC interrupt logging Bit 9 - vIOAPIC logging Bit 10 - hypercall logging Bit 11 - MSR operation logging
Recognized in debug builds of the hypervisor only.
= <boolean>
Default:
false
Allow use of the Forced Emulation Prefix in HVM guests, to allow emulation of arbitrary instructions.
This option is intended for development and testing purposes.
Warning As this feature opens up the instruction emulator to arbitrary instruction from an HVM guest, don’t use this in production system. No security support is provided when this flag is set.
= <boolean>
Default:
true
Specify whether guests are to be given access to physical port 80 (often used for debugging purposes), to override the DMI based detection of systems known to misbehave upon accesses to that port.
= <integer>
= old | new
Default:
new
unless directed-EOI is supported
= List of [ <bool>, verbose, debug, force, required,
quarantine=<bool>|scratch-page,
sharept, superpages, intremap, intpost, crash-disable,
snoop, qinval, igfx, amd-iommu-perdev-intremap,
dom0-{passthrough,strict} ]
All sub-options are boolean in nature.
I/O Memory Memory Units perform a function similar to the CPU MMU (hence the name), but typically exist as a discrete device, integrated as part of a PCI Root Complex. The most common configuration is to have one IOMMU per package (for on-die PCIe devices and directly attached PCIe lanes), and one IOMMU covering the remaining I/O in the system.
The functionality in an IOMMU commonly falls into two orthogonal categories:
DMA remapping which uses a pagetable-like hierarchical structure and maps I/O Virtual Addresses (DFNs - Device Frame Numbers in Xen’s terminology) to System Physical Addresses (MFNs - Machine Frame Numbers in Xen’s terminology).
Interrupt Remapping, which controls incoming Message Signalled Interrupt requests, including their routing to specific CPUs.
IOMMU functionality can be used to provide a translation which the hardware device driver isn’t aware of (e.g. PCI Passthrough and a native driver inside the guest) and/or to enforce fine-grained control over the memory and interrupts which a device is attempting to access.
By default, IOMMUs are configured for use if they are available. An
overall boolean (e.g. iommu=no
) can override this and leave
the IOMMUs disabled.
The verbose
and debug
booleans can be
used to print additional diagnostic information. Neither are active by
default.
The force
and required
booleans are
synonymous and, when requested, will prevent Xen from booting if IOMMUs
aren’t discovered and enabled successfully.
The quarantine
option can be used to control Xen’s
behavior when de-assigning devices from guests. The default behaviour is
chosen at compile time, and is one of
CONFIG_IOMMU_QUARANTINE_{NONE,BASIC,SCRATCH_PAGE}
.
When a PCI device is assigned to an untrusted domain, it is possible for that domain to program the device to DMA to an arbitrary address. The IOMMU is used to protect the host from malicious DMA by making sure that the device addresses can only target memory assigned to the guest. However, when the guest domain is torn down, assigning the device back to the hardware domain would allow any in-flight DMA to potentially target critical host data. To avoid this, quarantining should be enabled. Quarantining can be done in two ways: In its basic form, all in-flight DMA will simply be forced to encounter IOMMU faults. Since there are systems where doing so can cause host lockup, an alternative form is available where accesses to memory will be directed to a scratch page. The implication here is that such accesses will go unnoticed, i.e. an admin may not become aware of the underlying problem.
Therefore, if this option is set to true (the default), Xen always quarantines such devices; they must be explicitly assigned back to Dom0 before they can be used there again. If set to “scratch-page”, still active DMA operations will additionally be directed to a “scratch” page. If set to false, Xen will only quarantine devices the toolstack has arranged for getting quarantined, and only in the “basic” form.
This option is only valid on builds supporting PCI.
The sharept
boolean controls whether the IOMMU
pagetables are shared with the CPU-side HAP pagetables, or allocated
separately. Sharing reduces the memory overhead, but doesn’t work in
combination with CPU-side pagefault-based features, e.g. dirty VRAM
tracking when a PCI device is assigned.
Due to implementation choices, sharing pagetables doesn’t work on AMD hardware, and this option is ignored. It is enabled by default on Intel systems.
This option is ignored on ARM, and the pagetables are always shared.
The superpages
boolean controls whether superpage
mappings may be used in IOMMU page tables. If using this option is
necessary to fix an issue, please report a bug.
This option is only valid on x86.
The intremap
boolean controls the Interrupt
Remapping sub-feature, and is active by default on compatible hardware.
On x86 systems, the first generation of IOMMUs only supported DMA
remapping, and Interrupt Remapping appeared in the second
generation.
This option is only valid on x86.
The intpost
boolean controls the Posted Interrupt
sub-feature. In combination with APIC acceleration (VT-x APICV, SVM
AVIC), the IOMMU can be configured to deliver interrupts from assigned
PCI devices directly into the guest, without trapping out into
hypervisor context.
This option depends on intremap
, and is disabled by
default due to some corner cases in the implementation which have yet to
be resolved.
This option is only valid on x86, and only builds of Xen with HVM support.
The crash-disable
boolean controls disabling IOMMU
functionality (DMAR/IR/QI) before switching to a crash kernel. This
option is inactive by default and is for compatibility with older kdump
kernels only. Modern kernels copy all the necessary tables from the
previous one following kexec which makes the transition transparent for
them with IOMMU functions still on.
The following options are specific to Intel VT-d hardware:
The snoop
boolean controls the Snoop Control
sub-feature, and is active by default on compatible hardware.
An incoming DMA request may specify Snooped (query the CPU caches for the appropriate lines) or Non-Snooped (don’t query the CPU caches). Non-Snooped accesses incur less latency, but behind-the-scenes hypervisor activity can invalidate the expectations of the device driver, and Snoop Control allows the hypervisor to force DMA requests to be Snooped when they would otherwise not be.
The qinval
boolean controls the Queued Invalidation
sub-feature, and is active by default on compatible hardware. Queued
Invalidation is a feature in second-generation IOMMUs and is a
functional prerequisite for Interrupt Remapping. Note that Xen
disregards this setting for Intel VT-d version 6 and greater as
Registered-Based Invalidation isn’t supported by them.
The igfx
boolean is active by default, and controls
whether IOMMUs in front of solely graphics devices get enabled or
not.
It is intended as a debugging mechanism for graphics issues, and to
be similar to Linux’s intel_iommu=igfx_off
option. If
specifying no-igfx
fixes anything, please report the
problem.
The following options are specific to AMD-Vi hardware:
The amd-iommu-perdev-intremap
boolean controls
whether the interrupt remapping table is per device (the default), or a
single global table for the entire system.
Using a global table is not security supported as it allows all devices to impersonate each other as far as interrupts as concerned (see XSA-36), but it is a workaround for SP5100 Erratum 28.
WARNING: The dom0-passthrough
and
dom0-strict
booleans are both deprecated, and superseded by
dom0-iommu={passthrough,strict} respectively - using both the
old and new command line options in combination is
undefined.
= <integer>
Default:
1000
Specify the timeout of the device IOTLB invalidation in milliseconds. By default, the timeout is 1000 ms. When you see error ‘Queue invalidate wait descriptor timed out’, try increasing this value.
= <boolean>
WARNING: This command line option is deprecated, and superseded by dom0-iommu=map-inclusive - using both options in combination is undefined.
= <integer>
Default:
32
Maximum number of guests any individual IRQ could be shared between, i.e. a limit on the number of guests it is possible to start each having assigned a device sharing a common interrupt line. Accepts values between 1 and 255.
= <integer>
= <start>[-<end>][=<bdf1>[-<bdf1'>][,<bdf2>[-<bdf2'>][,...]]][;<start>...]
Define IVMD-like ranges that are missing from ACPI tables along with the device(s) they belong to, and use them for 1:1 mapping. End addresses can be omitted when exactly one page is meant. The ranges are inclusive when start and end are specified. Note that only PCI segment 0 is supported at this time, but it is fine to specify it explicitly.
‘start’ and ‘end’ values are page numbers (not full physical addresses), in hexadecimal format (can optionally be preceded by “0x”).
Omitting the optional (range of) BDF spcifiers signals that the range is to be applied to all devices.
Usage example: If device 0:0:1d.0 requires one page (0xd5d45) to be reserved, and devices 0:0:1a.0…0:0:1a.3 collectively require three pages (0xd5d46 thru 0xd5d48) to be reserved, one usage would be:
ivmd=d5d45=0:1d.0;0xd5d46-0xd5d48=0:1a.0-0:1a.3
Note: grub2 requires to escape or quote special characters, like ‘;’ when multiple ranges are specified - refer to the grub2 documentation.
<hpet>
] (AMD)
=[<seg>:]<bus>:<device>.<func>
Force the use of
[<seg>:]<bus>:<device>.<func>
as
device ID of HPET <hpet>
instead of the one specified
by the IVHD sub-tables of the IVRS ACPI table.
<ioapic>
] (AMD)
=[<seg>:]<bus>:<device>.<func>
Force the use of
[<seg>:]<bus>:<device>.<func>
as
device ID of IO-APIC <ioapic>
instead of the one
specified by the IVHD sub-tables of the IVRS ACPI table.
= <boolean>
Force the use of use of the local APIC on a uniprocessor system, even if left disabled by the BIOS.
= <boolean>
= <boolean>
Default: false
This option is intended for debugging purposes only. Enable MSR_DEBUGCTL.LBR in hypervisor context to be able to dump the Last Interrupt/Exception To/From record with other registers.
= <boolean>
Default:
false
Flag to enable or disable LLC coloring support at runtime. This
option is available only when CONFIG_LLC_COLORING
is
enabled. See the general cache coloring documentation for more info.
= <integer>
Default:
Obtained from hardware
Specify the number of ways of the Last Level Cache. This option is
available only when CONFIG_LLC_COLORING
is enabled. LLC
size and number of ways are used to find the number of supported cache
colors. By default the value is automatically computed by probing the
hardware, but in case of specific needs, it can be manually set. Those
include failing probing and debugging/testing purposes so that it’s
possible to emulate platforms with different number of supported colors.
If set, also “llc-size” must be set, otherwise the default will be used.
Note that using both options implies “llc-coloring=on” unless an earlier
“llc-coloring=off” is there.
= <size>
Default:
Obtained from hardware
Specify the size of the Last Level Cache. This option is available
only when CONFIG_LLC_COLORING
is enabled. LLC size and
number of ways are used to find the number of supported cache colors. By
default the value is automatically computed by probing the hardware, but
in case of specific needs, it can be manually set. Those include failing
probing and debugging/testing purposes so that it’s possible to emulate
platforms with different number of supported colors. If set, also
“llc-nr-ways” must be set, otherwise the default will be used. Note that
using both options implies “llc-coloring=on” unless an earlier
“llc-coloring=off” is there.
= <integer>
Default:
lock-depth-size=64
Specifies the maximum number of nested locks tested for illegal
recursions. Higher nesting levels still work, but recursion testing is
omitted for those levels. In case an illegal recursion is detected the
system will crash immediately. Specifying 0
will disable
all testing of illegal lock nesting.
This option is available for hypervisors built with CONFIG_DEBUG_LOCKS only.
= <level>[/<rate-limited level>]
where level isnone | error | warning | info | debug | all
Default:
loglvl=info
Can be modified at runtime
Set the logging level for Xen. Any log message with equal more more importance will be printed.
The optional <rate-limited level>
option instructs
which severities should be rate limited.
= none | min | all
Default:
none
if not specified at all, or tomin
if low_crashinfo is present without qualification.
This option is only useful for hosts with a 32bit dom0 kernel, wishing to use kexec functionality in the case of a crash. It represents which data structures should be deliberately allocated in low memory, so the crash kernel may find find them. Should be used in combination with crashinfo_maxaddr.
= <size>
Default:
64M
Specify the threshold below which Xen will inform dom0 that the
quantity of free memory is getting low. Specifying 0
will
disable this notification.
= <integer>
Specify the maximum number of CPUs that should be brought up.
This option is ignored in pv-shim mode.
WARNING: On Arm big.LITTLE systems, when
hmp-unsafe
option is enabled, this command line option does
not guarantee on which CPU types will be used.
= <integer>[,<integer>]
Specify the deepest C-state CPUs are permitted to be placed in, and optionally the maximum sub C-state to be used used. The latter only applies to the highest permitted C-state.
= <integer>
Specifies the number of interrupts to be use for pin (IO-APIC or legacy PIC) based interrupts. Any higher IRQs will be available for use via PCI MSI.
= <integer>
Specifies the number of ARM GICv3 LPI interrupts to allocate on the host, presented as the number of bits needed to encode it. This must be at least 14 and not exceed 32, and each LPI requires one byte (configuration) and one pending bit to be allocated. Defaults to 20 bits (to cover at most 1048576 interrupts).
= <boolean>
Default:
true
Allows to disable the use of Machine Check Exceptions. Note that doing so may result in silent shutdown of the system in case an event occurs which would have resulted in raising a Machine Check Exception. Silent here is as far as Xen is concerned; firmware may offer to retrieve some collected data.
= <boolean>
Default:
false
Force broadcasting of Machine Check Exceptions, suppressing the use of Local MCE functionality available in newer Intel hardware.
= verbose
Specify verbose machine check output.
= <size>
Specify the maximum address of physical RAM. Any RAM beyond this limit is ignored by Xen.
= [<domU>][,[<ctldom>][,[<hwdom>][,<ptdom>]]]
x86 default:
9,18,12,12
ARM default:9,18,10,10
Change the maximum order permitted for allocation (or allocation-like) requests issued by the various kinds of domains (in this order: ordinary DomU, control domain, hardware domain, and - when supported by the platform - DomU with pass-through device assigned).
= <boolean>[,amd-fam10]
Default:
1
Specify if the MMConfig space should be enabled.
= <boolean> | all
Default:
false
By default, domains may not create cached mappings to MMIO regions.
This option relaxes the check for Domain 0 (or when using
all
, all PV domains), to permit the use of cacheable MMIO
mappings.
= <boolean>
Default:
true
Force Xen to (not) use PCI-MSI, even if ACPI FADT says otherwise.
= <boolean>
Default:
false
Print boot time MTRR state.
= <boolean>
Default:
true
Use the MWAIT idle driver (with model specific C-state knowledge) instead of the ACPI based one.
= ignore | dom0 | fatal
Default:
fatal
for a debug build, ordom0
for a non-debug build
Specify what Xen should do in the event of an NMI parity or I/O
error. ignore
discards the error; dom0
causes
Xen to report the error to dom0, while ‘fatal’ causes Xen to print
diagnostics and then hang.
Instruct Xen to ignore any IOAPICs that are present in the system, and instead continue to use the legacy PIC. This is not recommended with pvops type kernels.
Because responsibility for APIC setup is shared between Xen and the domain 0 kernel this option is automatically propagated to the domain 0 command line.
= <boolean>
Default:
true
By default, Xen will use the INVPCID instruction for TLB management if it is available. This option can be used to cause Xen to fall back to older mechanisms, which are generally slower.
= <integer>
The minimum interval between load balancing events on a given pcpu, in microseconds. A value of ‘0’ will disable rate limiting. Maximum value 1 second. At the moment only credit honors this parameter. Default 1ms.
= <boolean>
Disable software IRQ balancing and affinity. This can be used on systems such as Dell 1850/2850 that have workarounds in hardware for IRQ routing issues.
= <boolean>
Default:
false
Ignore the local APIC on a uniprocessor system, even if enabled by the BIOS.
= <boolean>
Do not execute real-mode bootstrap code when booting Xen. This option should not be used except for debugging. It will effectively disable the vga option, which relies on real mode to set the video mode.
= <boolean>
Do not automatically reboot after an error. This is useful for catching debug output. Defaults to automatically reboot after 5 seconds.
= <boolean>
Disable SMP support. No secondary processors will be booted. Defaults to booting secondary processors.
This option is ignored in pv-shim mode.
= <integer>
= on | off | fake=<integer> | noacpi
Default:
on
= <boolean>
Default:
false
Flag to enable or disable partial emulation of system/coprocessor registers. Only effective if CONFIG_PARTIAL_EMULATION is enabled.
WARNING: Enabling this option might result in unwanted/non-spec compliant behavior.
= List of [ serr=<bool>, perr=<bool> ]
Default: Signaling left as set by firmware.
Override the firmware settings, and explicitly enable or disable the signalling of PCI System and Parity errors.
=[<seg>:]<bus>:<device>,<stride>
Mark a group of PCI devices as using phantom functions without actually advertising so, so the IOMMU can create translation contexts for them.
All numbers specified must be hexadecimal ones.
This option can be specified more than once (up to 8 times at present).
= <boolean>
Default:
false
Flag to enable or disable support for PCI passthrough
= <boolean> | xpti=<bool>
Default:
xpti
Can be modified at runtime (change takes effect only for domains created afterwards)
If available, control usage of the PCID feature of the processor for
64-bit pv-domains. PCID can be used either for no domain at all
(false
), for all of them (true
), only for
those subject to XPTI (xpti
) or for those not subject to
XPTI (no-xpti
). The feature is used only in case INVPCID is
supported and not disabled via invpcid=false
.
= <integer>
= <integer>
= ( <integer> | List of ( C1 | C1E | C2 | ... )
This is a mask of C-states which are to be used preferably. This option is applicable only on hardware were certain C-states are exclusive of one another.
= <boolean>
Default:
true
outside of shim mode,false
in shim mode
Certain devices accessible by I/O ports may be accessible also through “alias” ports (originally a result of incomplete address decoding). When such devices are solely under Xen’s control, Xen disallows even Dom0 access to the “primary” ports. When alias probing is active and aliases are detected, “alias” ports would then be treated similar to the “primary” ones.
= List of ( cmt:<boolean> | rmid_max:<integer> | cat:<boolean> | cos_max:<integer> | cdp:<boolean> )
Default:
psr=cmt:0,rmid_max:255,cat:0,cos_max:255,cdp:0
Platform Shared Resource(PSR) Services. Intel Haswell and later server platforms offer information about the sharing of resources.
To use the PSR monitoring service for a certain domain, a Resource Monitoring ID(RMID) is used to bind the domain to corresponding shared resource. RMID is a hardware-provided layer of abstraction between software and logical processors.
To use the PSR cache allocation service for a certain domain, a capacity bitmasks(CBM) is used to bind the domain to corresponding shared resource. CBM represents cache capacity and indicates the degree of overlap and isolation between domains. In hypervisor a Class of Service(COS) ID is allocated for each unique CBM.
The following resources are available:
Cache Monitoring Technology (Haswell and later). Information regarding the L3 cache occupancy.
cmt
instructs Xen to enable/disable Cache Monitoring
Technology.rmid_max
indicates the max value for rmid.Memory Bandwidth Monitoring (Broadwell and later). Information regarding the total/local memory bandwidth. Follow the same options with Cache Monitoring Technology.
Cache Allocation Technology (Broadwell and later). Information regarding the cache allocation.
cat
instructs Xen to enable/disable Cache Allocation
Technology.cos_max
indicates the max value for COS ID.Code and Data Prioritization Technology (Broadwell and later). Information regarding the code cache and the data cache allocation. CDP is based on CAT.
cdp
instructs Xen to enable/disable Code and Data
Prioritization. Note that cos_max
of CDP is a little
different from cos_max
of CAT. With CDP, one COS will
corespond two CBMs other than one with CAT, due to the sum of CBMs is
fixed, that means actual cos_max
in use will automatically
reduce to half when CDP is enabled.= List of [ 32=<bool> ]
Applicability: x86
Controls for aspects of PV guest support.
The 32
boolean controls whether 32bit PV guests can
be created. It defaults to true
, and is ignored when
CONFIG_PV32
is compiled out.
32bit PV guests are incompatible with CET Shadow Stacks. If Xen is
using shadow stacks, this option will be overridden to
false
. Backwards compatibility can be maintained with the
pv-shim
mechanism.
= <boolean>
Default:
true
Only available if Xen is compiled with
CONFIG_PV_LINEAR_PT
support enabled.
Allow PV guests to have pagetable entries pointing to other pagetables of the same level (i.e., allowing L2 PTEs to point to other L2 pages). This technique is often called “linear pagetables”, and is sometimes used to allow operating systems a simple way to consistently map the current process’s pagetables into its own virtual address space.
Linux and MiniOS don’t use this technique. NetBSD and Novell Netware do; there may be other custom operating systems which do. If you’re certain you don’t plan on having PV guests which use this feature, turning it off can reduce the attack surface.
= List of [ <bool>, dom0=<bool>, domu=<bool> ]
Default:
false
on believed-unaffected hardware, or in pv-shim mode.domu
on believed-affected hardware.
Mitigations for L1TF / XSA-273 / CVE-2018-3620 for PV guests.
For backwards compatibility, we may not alter an architecturally-legitimate pagetable entry a PV guest chooses to write. We can however force such a guest into shadow mode so that Xen controls the PTEs which are reachable by the CPU pagewalk.
Shadowing is performed at the point where a PV guest first tries to write an L1TF-vulnerable PTE. Therefore, a PV guest kernel which has been updated with its own L1TF mitigations will not trigger shadow mode if it is well behaved.
If CONFIG_SHADOW_PAGING
is not compiled in, this
mitigation instead crashes the guest when an L1TF-vulnerable PTE is
written, which still allows updated, well-behaved PV guests to run,
despite Shadow being compiled out.
In the pv-shim case, Shadow is expected to be compiled out, and a malicious guest kernel can only leak data from the shim Xen, rather than the host Xen.
= <boolean>
Default:
false
This option is intended for use by a toolstack, when choosing to run a PV guest compatibly inside an HVM container.
In this mode, the kernel and initrd passed as modules to the hypervisor are constructed into a plain unprivileged PV domain.
= <integer>
Default:
10
How frequently a CPU which has gone idle, but with pending RCU callbacks, should be woken up to check if the grace period has completed, and the callbacks are safe to be executed. Expressed in milliseconds; maximum is 100, and it can’t be 0.
= t[riple] | k[bd] | a[cpi] | p[ci] | P[ower] | e[fi] | n[o] [, [w]arm | [c]old]
Default: system dependent
Specify the host reboot method.
warm
instructs Xen to not set the cold reboot flag.
cold
instructs Xen to set the cold reboot flag.
no
instructs Xen to not automatically reboot after
panics or crashes.
triple
instructs Xen to reboot the host by causing a
triple fault.
kbd
instructs Xen to reboot the host via the keyboard
controller.
acpi
instructs Xen to reboot the host using RESET_REG in
the ACPI FADT (this is default mode if available).
pci
instructs Xen to reboot the host using PCI reset
register (port CF9).
Power
instructs Xen to power-cycle the host using PCI
reset register (port CF9).
efi
instructs Xen to reboot using the EFI reboot
call.
xen
instructs Xen to reboot using Xen’s SCHEDOP
hypercall (this is the default when running nested Xen)
= start<-end>=[s1]bdf1[,[s1]bdf2[,...]];start<-end>=[s2]bdf1[,[s2]bdf2[,...]]
Define RMRR units that are missing from ACPI table along with device they belong to and use them for 1:1 mapping. End addresses can be omitted and one page will be mapped. The ranges are inclusive when start and end are specified. If segment of the first device is not specified, segment zero will be used. If other segments are not specified, first device segment will be used. If a segment is specified for other than the first device and it does not match the one specified for the first one, an error will be reported.
‘start’ and ‘end’ values are page numbers (not full physical addresses), in hexadecimal format (can optionally be preceded by “0x”).
Usage example: If device 0:0:1d.0 requires one page (0xd5d45) to be reserved, and device 0:0:1a.0 requires three pages (0xd5d46 thru 0xd5d48) to be reserved, one usage would be:
rmrr=d5d45=0:0:1d.0;0xd5d46-0xd5d48=0:0:1a.0
Note: grub2 requires to escape or use quotations if special characters are used, namely ‘;’, refer to the grub2 documentation if multiple ranges are specified.
= <boolean>
Default:
true
Map the HPET page as read only in Dom0. If disabled the page will be mapped with read and write permissions.
= credit | credit2 | arinc653 | rtds | null
Default:
sched=credit2
Choose the default scheduler. Note the default scheduler is
selectable via Kconfig and depends on enabled schedulers. Check
CONFIG_SCHED_DEFAULT
to see which scheduler is the
default.
= <integer>
Default:
16
Defines how many CPUs will be put, at most, in each Credit2 runqueue.
Runqueues are still arranged according to the host topology (and following what indicated by the ‘credit2_runqueue’ parameter). But we also have a cap to the number of CPUs that share each runqueues.
A value that is a submultiple of the number of online CPUs is recommended, as that would likely produce a perfectly balanced runqueue configuration.
= <integer>
= <integer>
Set the timeslice of the credit1 scheduler, in milliseconds. The default is 30ms. Reasonable values may include 10, 5, or even 1 for very latency-sensitive workloads.
= cpu | core | socket
Default:
sched-gran=cpu
Set the scheduling granularity. In case the granularity is larger
than 1 (e.g. core
on a SMT-enabled system, or
socket
) multiple vcpus are assigned statically to a
“scheduling unit” which will then be subject to scheduling. This
assignment of vcpus to scheduling units is fixed.
cpu
: Vcpus will be scheduled individually on single cpus
(e.g. a hyperthread using x86/Intel terminology)
core
: As many vcpus as there are cpus on a physical core
are scheduled together on a physical core.
socket
: As many vcpus as there are cpus on a physical
sockets are scheduled together on a physical socket.
Note: a value other than cpu
will result in rejecting a
runtime modification attempt of the “smt” setting.
Note: for AMD x86 processors before Fam17 the terminology in the official data sheets is different: a cpu is named “core” and multiple “cores” are running in the same “compute unit”. As from Fam17 on AMD is using the same names as Intel (“thread” and “core”) the topology levels are named “cpu”, “core” and “socket” even on older AMD processors.
= <integer>
In order to limit the rate of context switching, set the minimum amount of time that a vcpu can be scheduled for before preempting it, in microseconds. The default is 1000us (1ms). Setting this to 0 disables it altogether.
= <boolean>
Normally Xen will try to maximize performance and cache utilization by spreading out vcpus across as many different divisions as possible (i.e, numa nodes, sockets, cores threads, &c). This often maximizes throughput, but also maximizes energy usage, since it reduces the depth to which a processor can sleep.
This option inverts the logic, so that the scheduler in effect tries to keep the vcpus on the smallest amount of silicon possible; i.e., first fill up sibling threads, then sibling cores, then sibling sockets, &c. This will reduce performance somewhat, particularly on systems with hyperthreading enabled, but should reduce power by enabling more sockets and cores to go into deeper sleep states.
= <boolean>
Default:
false
Scrub domains’ freed pages. This is a safety net against a (buggy) domain accidentally leaking secrets by releasing pages without proper sanitization.
= <size>
Default:
16kB
Set the serial transmit buffer size.
= diverse | panic
Default:
diverse
This parameter is provided to administrators to determine how the hypervisor handles SErrors.
diverse
: The hypervisor will distinguish guest SErrors
from hypervisor SErrors:
panic
: All SErrors will cause the whole system to
crash. This option should only be used if you trust all your guests
and/or they don’t have a gadget (e.g. device) to generate SErrors in
normal run.
= List of ( min:<size> | max:<size> | <size> )
Set the amount of memory that xen-shim uses. Only has effect if pv-shim mode is enabled. Note that this value accounts for the memory used by the shim itself plus the free memory slack given to the shim for runtime allocations.
min:<size>
specifies the minimum amount of
memory. Ignored if greater than max.max:<size>
specifies the maximum amount of
memory.<size>
specifies the exact amount of memory.
Overrides both min and max.By default, the amount of free memory slack given to the shim for runtime usage is 1MB.
= <boolean> | hvm
Default:
true
unless running in pv-shim mode on AMD or Hygon hardware
Flag to enable Supervisor Mode Access Prevention Use
smap=hvm
to allow SMAP use by HVM guests only.
In PV shim mode on AMD or Hygon hardware due to significant performance impact in some cases and generally lower security risk the option defaults to false.
= <boolean> | hvm
Default:
true
unless running in pv-shim mode on AMD or Hygon hardware
Flag to enable Supervisor Mode Execution Protection Use
smep=hvm
to allow SMEP use by HVM guests only.
In PV shim mode on AMD or Hygon hardware due to significant performance impact in some cases and generally lower security risk the option defaults to false.
= <boolean>
Default: true
Control bring up of multiple hyper-threads per CPU core.
= <boolean> | cap | <integer>
A true boolean value enables legacy behavior (1s timeout), while
cap
enforces the maximum theoretically necessary timeout of
670ms. Any number is being interpreted as a custom timeout in
milliseconds. Zero or boolean false disable the quirk workaround, which
is also the default.
= List of [ ssbd=force-disable|runtime|force-enable ]
Controls for speculative execution sidechannel mitigations.
The option ssbd=
is used to control the state of
Speculative Store Bypass Disable (SSBD) mitigation.
ssbd=force-disable
will keep the mitigation permanently
off. The guest will not be able to control the state of the
mitigation.ssbd=runtime
will always turn on the mitigation when
running in the hypervisor context. The guest will be to turn on/off the
mitigation for itself by using the firmware interface
ARCH_WORKAROUND_2
.ssbd=force-enable
will keep the mitigation permanently
on. The guest will not be able to control the state of the
mitigation.By default SSBD will be mitigated at runtime (i.e
ssbd=runtime
).
= List of [ <bool>, xen=<bool>, {pv,hvm}=<bool>, {msr-sc,rsb,verw,{ibpb,bhb}-entry}=<bool>|{pv,hvm}=<bool>, bti-thunk=retpoline|lfence|jmp,bhb-seq=short|tsx|long, {ibrs,ibpb,ssbd,psfd, eager-fpu,l1d-flush,branch-harden,srb-lock, unpriv-mmio,gds-mit,div-scrub,lock-harden, bhi-dis-s}=<bool> ]
Controls for speculative execution sidechannel mitigations. By default, Xen will pick the most appropriate mitigations based on compiled in support, loaded microcode, and hardware details, and will virtualise appropriate mitigations for guests to use.
WARNING: Any use of this option may interfere with heuristics. Use with extreme care.
An overall boolean value, spec-ctrl=no
, can be specified
to turn off all mitigations, including pieces of infrastructure used to
virtualise certain mitigation features for guests. This also includes
settings which xpti
, smt
,
pv-l1tf
, tsx
control, unless the respective
option(s) have been specified earlier on the command line.
Alternatively, a slightly more restricted
spec-ctrl=no-xen
can be used to turn off all of Xen’s
mitigations, while leaving the virtualisation support in place for
guests to use.
Use of a positive boolean value for either of these options is invalid.
The pv=
, hvm=
, msr-sc=
,
rsb=
, verw=
, ibpb-entry=
and
bhb-entry=
options offer fine grained control over the
primitives by Xen. These impact Xen’s ability to protect itself, and/or
Xen’s ability to virtualise support for guests to use.
pv=
and hvm=
offer control over all
suboptions for PV and HVM guests respectively.
Each other option can be used either as a plain boolean
(e.g. spec-ctrl=rsb
to control both the PV and HVM
sub-options), or with pv=
or hvm=
subsuboptions (e.g. spec-ctrl=rsb=no-hvm
to disable HVM RSB
only).
msr-sc=
offers control over Xen’s support for
manipulating MSR_SPEC_CTRL
on entry and exit. These blocks
are necessary to virtualise support for guests and if disabled, guests
will be unable to use IBRS/STIBP/SSBD/etc.
rsb=
offers control over whether to overwrite the
Return Stack Buffer / Return Address Stack on entry to Xen and on
idle.
verw=
offers control over whether to use VERW for
its scrubbing side effects at appropriate privilege transitions. The
exact side effects are microarchitecture and microcode specific.
Note: md-clear=
is accepted as a deprecated alias. For
compatibility with development versions of XSA-297, mds=
is
also accepted on Xen 4.12 and earlier as an alias. Consult vendor
documentation in preference to here.
ibpb-entry=
offers control over whether IBPB
(Indirect Branch Prediction Barrier) is used on entry to Xen. This is
used by default on hardware vulnerable to Branch Type Confusion, and
hardware vulnerable to Speculative Return Stack Overflow if appropriate
microcode has been loaded, but for performance reasons dom0 is
unprotected by default. If it is necessary to protect dom0 too, boot
with spec-ctrl=ibpb-entry
.
bhb-entry=
offers control over whether BHB-clearing
(Branch History Buffer) sequences are used on entry to Xen. This is used
by default on hardware vulnerable to Branch History Injection, when the
BHI_DIS_S control is not available (see bhi-dis-s
). The
choice of scrubbing sequence can be selected using the
bhb-seq=
option. If it is necessary to protect dom0 too,
boot with spec-ctrl=bhb-entry
.
If Xen was compiled with CONFIG_INDIRECT_THUNK
support,
bti-thunk=
can be used to select which of the thunks gets
patched into the __x86_indirect_thunk_%reg
locations. The
default thunk is retpoline
(generally preferred), with the
alternatives being jmp
(a jmp *%reg
gadget,
minimal overhead), and lfence
(an
lfence; jmp *%reg
gadget).
On all hardware, bhb-seq=
can be used to select which of
the BHB-clearing sequences gets used. This interacts with the
bhb-entry=
and bhi-dis-s=
options in order to
mitigate Branch History Injection on affected hardware. The default
sequence is short
, with tsx
as an alternative
available capable hardware, and long
that can be opted in
to.
On hardware supporting IBRS (Indirect Branch Restricted Speculation),
the ibrs=
option can be used to force or prevent Xen using
the feature itself. If Xen is not using IBRS itself, functionality is
still set up so IBRS can be virtualised for guests.
On hardware supporting STIBP (Single Thread Indirect Branch
Predictors), the stibp=
option can be used to force or
prevent Xen using the feature itself. By default, Xen will use STIBP
when IBRS is in use (IBRS implies STIBP), and when hardware hints
recommend using it as a blanket setting.
On hardware supporting SSBD (Speculative Store Bypass Disable), the
ssbd=
option can be used to force or prevent Xen using the
feature itself. The feature is virtualised for guests, independently of
Xen’s choice of setting. On AMD hardware, disabling Xen SSBD usage on
the command line (ssbd=0
which is the default value) can
lead to Xen running with the guest SSBD selection depending on hardware
support, on the same hardware setting ssbd=1
will result in
SSBD always being enabled, regardless of guest choice.
On hardware supporting PSFD (Predictive Store Forwarding Disable),
the psfd=
option can be used to force or prevent Xen using
the feature itself. By default, Xen will not use PSFD. PSFD is implied
by SSBD, and SSBD is off by default.
On hardware supporting BHI_DIS_S (Branch History Injection Disable
Supervisor), the bhi-dis-s=
option can be used to force or
prevent Xen using the feature itself. By default Xen will use BHI_DIS_S
on hardware susceptible to Branch History Injection.
On hardware supporting IBPB (Indirect Branch Prediction Barrier), the
ibpb=
option can be used to force (the default) or prevent
Xen from issuing branch prediction barriers on vcpu context
switches.
On all hardware, the eager-fpu=
option can be used to
force or prevent Xen from using fully eager FPU context switches. This
is currently implemented as a global control. By default, Xen will
choose to use fully eager context switches on hardware believed to
speculate past #NM exceptions.
On hardware supporting L1D_FLUSH, the l1d-flush=
option
can be used to force or prevent Xen from issuing an L1 data cache flush
on each VMEntry. Irrespective of Xen’s setting, the feature is
virtualised for HVM guests to use. By default, Xen will enable this
mitigation on hardware believed to be vulnerable to L1TF.
If Xen is compiled with
CONFIG_SPECULATIVE_HARDEN_BRANCH
, the
branch-harden=
boolean can be used to force or prevent Xen
from using speculation barriers to protect selected conditional
branches. By default, Xen will enable this mitigation.
On hardware supporting SRBDS_CTRL, the srb-lock=
option
can be used to force or prevent Xen from protect the Special Register
Buffer from leaking stale data. By default, Xen will enable this
mitigation, except on parts where MDS is fixed and TAA is
fixed/mitigated and there are no unprivileged MMIO mappings (in which
case, there is believed to be no way for an attacker to obtain stale
data).
The unpriv-mmio=
boolean indicates whether the system
has (or will have) less than fully privileged domains granted access to
MMIO devices. By default, this option is disabled. If enabled, Xen will
use the FB_CLEAR
and/or SRBDS_CTRL
functionality available in the Intel May 2022 microcode release to
mitigate cross-domain leakage of data via the MMIO Stale Data
vulnerabilities.
On all hardware, the gds-mit=
option can be used to
force or prevent Xen from mitigating the GDS (Gather Data Sampling)
vulnerability. By default, Xen will mitigate GDS on hardware believed to
be vulnerable. On hardware supporting GDS_CTRL (requires the August 2023
microcode), and where firmware has elected not to lock the
configuration, Xen will use GDS_CTRL to mitigate GDS with. Otherwise,
Xen will mitigate by disabling AVX, which blocks the use of the AVX2
Gather instructions.
On all hardware, the div-scrub=
option can be used to
force or prevent Xen from mitigating the DIV-leakage vulnerability. By
default, Xen will mitigate DIV-leakage on hardware believed to be
vulnerable.
If Xen is compiled with CONFIG_SPECULATIVE_HARDEN_LOCK
,
the lock-harden=
boolean can be used to force or prevent
Xen from using speculation barriers to protect lock critical regions.
This mitigation won’t be engaged by default, and needs to be explicitly
enabled on the command line.
= <boolean>
Default:
false
Flag to force synchronous console output. Useful for debugging, but not suitable for production environments due to incurred overhead.
= 0x<phys_addr>
Specify the physical address of the trusted boot shared page.
= <integer>
Specify the per-cpu trace buffer size in pages.
= <boolean>
Default:
true
Flag to enable TSC deadline as the APIC timer mode.
= <integer>
Specify a mask for Xen event tracing. This allows Xen tracing to be enabled at boot. Refer to the xentrace(8) documentation for a list of valid event mask values. In order to enable tracing, a buffer size (in pages) must also be specified via the tbuf_size parameter.
= <boolean>
= <integer>
= unstable | skewed | stable:socket
= <bool>
Applicability: x86 with CONFIG_INTEL active
Default: false on parts vulnerable to TAA, true otherwise
Controls for the use of Transactional Synchronization eXtensions.
Several microcode updates are relevant:
March 2019, fixing the TSX memory ordering errata on all TSX-enabled CPUs to date. Introduced MSR_TSX_FORCE_ABORT on SKL/SKX/KBL/WHL/CFL parts. The errata workaround uses Performance Counter 3, so the user can select between working TSX and working perfcounters.
November 2019, fixing the TSX Async Abort speculative vulnerability. Introduced MSR_TSX_CTRL on all TSX-enabled MDS_NO parts to date, CLX/WHL-R/CFL-R, with the controls becoming architectural moving forward and formally retiring HLE from the architecture. The user can disable TSX to mitigate TAA, and elect to hide the HLE/RTM CPUID bits. Also causes VERW to once-again flush the microarchiectural buffers in case a TAA mitigation is wanted along with TSX being enabled.
June 2021, removing the workaround for March 2019 on client CPUs and formally de-featured TSX on SKL/KBL/WHL/CFL (Note: SKX still retains the March 2019 fix). Introduced the ability to hide the HLE/RTM CPUID bits. PCR3 works fine, and TSX is disabled by default, but the user can re-enable TSX at their own risk, accepting that the memory order erratum is unfixed.
February 2022, removing the VERW flushing workaround from November 2019 on client CPUs and formally de-featuring TSX on WHL-R/CFL-R (Note: CLX still retains the VERW flushing workaround). TSX defaults to disabled, and is locked off when SGX is enabled in the BIOS. When SGX is not enabled, TSX can be re-enabled at the users own risk, as it reintroduces the TSX Async Abort speculative vulnerability.
On systems with the ability to configure TSX, this boolean offers system wide control of whether TSX is enabled or disabled.
When TSX is disabled, transactions unconditionally abort. This is compatible with the TSX spec, which requires software to have a non-transactional path as a fallback. The RTM and HLE CPUID bits are hidden from VMs by default, but can be re-enabled if required. This allows VMs which previously saw RTM/HLE to be migrated in, although any TSX-enabled software will run with reduced performance.
When TSX is locked off by firmware, tsx=
is ignored
and treated as false
.
An explicit tsx=
choice is honoured, even if it is
true
and would result in a vulnerable system.
When no explicit tsx=
choice is given, parts
vulnerable to TAA will be mitigated by disabling TSX, as this is the
lowest overhead option.
When no explicit tsx=
option is given, parts
susceptible to the memory ordering errata default to true
to enable working TSX. Alternatively, selecting tsx=0
will
disable TSX and restore PCR3 to a working state.
SKX and SKL/KBL/WHL/CFL on pre-June 2021 microcode default to
true
. Alternatively, selecting tsx=0
will
disable TSX and restore PCR3 to a working state.
SKL/KBL/WHL/CFL on the June 2021 microcode or later default to
false
. Alternatively, selecting tsx=1
will
re-enable TSX at the users own risk.
= List of [ <integer> | scan=<bool>, nmi=<bool> ]
Applicability: x86
Default: `nmi`
Controls for CPU microcode loading. For early loading, this parameter can specify how and where to find the microcode update blob. For late loading, this parameter specifies if the update happens within a NMI handler.
‘integer’ specifies the CPU microcode update blob module index. When
positive, this specifies the n-th module (in the GrUB entry, zero based)
to be used for updating CPU micrcode. When negative, counting starts at
the end of the modules in the GrUB entry (so with the blob commonly
being last, one could specify ucode=-1
). Note that the
value of zero is not valid here (entry zero, i.e. the first module, is
always the Dom0 kernel image). Note further that use of this option has
an unspecified effect when used with xen.efi (there the concept of
modules doesn’t exist, and the blob gets specified via the
ucode=<filename>
config file/section entry; see EFI configuration file description).
‘scan’ instructs the hypervisor to scan the multiboot images for an
cpio image that contains microcode. Depending on the platform the blob
with the microcode in the cpio name space must be: - on Intel:
kernel/x86/microcode/GenuineIntel.bin - on AMD :
kernel/x86/microcode/AuthenticAMD.bin When using xen.efi, the
ucode=<filename>
config file setting takes precedence
over scan
.
‘nmi’ determines late loading is performed in NMI handler or just in
stop_machine context. In NMI handler, even NMIs are blocked, which is
considered safer. The default value is true
.
= <boolean>
= <integer>
Default:
0
Specify a delay, in microseconds, between migrations of a VCPU between PCPUs when using the credit1 scheduler. This prevents rapid fluttering of a VCPU between CPUs, and reduces the implicit overheads such as cache-warming. 1ms (1000) has been measured as a good value.
= <integer>
Default:
0
This allows to override the amount of video RAM, in MiB, determined to be present.
= ( ask | current | text-80x<rows> | gfx-<width>x<height>x<depth> | mode-<mode> )[,keep]
ask
causes Xen to display a menu of available modes and
request the user to choose one of them.
current
causes Xen to use the graphics adapter in its
current state, without further setup.
text-80x<rows>
instructs Xen to set up text mode.
Valid values for <rows>
are
25, 28, 30, 34, 43, 50, 80
gfx-<width>x<height>x<depth>
instructs
Xen to set up graphics mode with the specified width, height and
depth.
mode-<mode>
instructs Xen to use a specific mode,
as shown with the ask
option. (N.B menu modes are displayed
in hex, so <mode>
should be a hexadecimal number)
The optional keep
parameter causes Xen to continue using
the vga console even after dom0 has been started. The default behaviour
is to relinquish control to dom0.
= <integer>
Default:
2047
Specify the maximum number of retries before an enlightened Windows guest will notify Xen that it has failed to acquire a spinlock.
= [<major>],[<minor>],[<build>]
Default:
6,0,0x1772
= <integer>
Default:
0
Specify the value of the VM Notify window used to detect locked VMs. Set to -1 to disable the feature. Value is in units of crystal clock cycles.
Note the hardware might add a threshold to the provided value in order to make it safe, and hence using 0 is fine.
= <boolean>
Default:
true
Use Virtual Processor ID support if available. This prevents the need for TLB flushes on VM entry and exit, increasing performance.
= List of [ <bool>, bts, ipc, arch, rtm-abort=<bool> ]
Applicability: x86. Default: false
Controls for Performance Monitoring Unit virtualisation.
Performance monitoring facilities tend to be very hardware specific, and provide access to a wealth of low level processor information.
An overall boolean can be used to enable or disable vPMU support. vPMU is disabled by default.
When enabled, guests have full access to all performance counter
settings, including model specific functionality. This is a superset of
the functionality offered by ipc
and/or arch
,
but a subset of the functionality offered by bts
.
Xen’s watchdog functionality is implemented using performance counters. As a result, use of the watchdog option will override and disable vPMU.
The bts
option enables performance monitoring, and
permits additional access to the Branch Trace Store controls. BTS is an
Intel feature where the processor can write data into a buffer whenever
a branch occurs. However, as this feature isn’t virtualised, a
misconfiguration by the guest can lock the entire system up.
The ipc
option allows access to the most minimal set
of counters possible: instructions, cycles, and reference cycles. These
can be used to calculate instructions per cycle (IPC).
The arch
option allows access to the pre-defined
architectural events.
The rtm-abort
boolean has been superseded. Use
tsx=0
instead.
Warning: As the virtualisation is not 100% safe, don’t use the vpmu flag on production systems (see https://xenbits.xen.org/xsa/advisory-163.html)!
= trap | native
Default:
trap
WFI is the ARM instruction to “wait for interrupt”. WFE is similar
and means “wait for event”. This option, which is ARM specific, changes
the way guest WFI and WFE are implemented in Xen. By default, Xen traps
both instructions. In the case of WFI, Xen blocks the guest vcpu; in the
case of WFE, Xen yield the guest vcpu. When setting vwfi to
native
, Xen doesn’t trap either instruction, running them
in guest context. Setting vwfi to native
reduces irq
latency significantly. It can also lead to suboptimal scheduling
decisions, but only when the system is oversubscribed (i.e., in total
there are more vCPUs than pCPUs).
= force | <boolean>
Default:
false
Run an NMI watchdog on each processor. If a processor is stuck for
longer than the watchdog_timeout, a panic occurs. When
force
is specified, in addition to running an NMI watchdog
on each processor, unknown NMIs will still be processed.
= <integer>
Default:
5
Set the NMI watchdog timeout in seconds. Specifying 0
will turn off the watchdog.
= <boolean>
Default:
true
Permit use of x2apic setup for SMP environments.
= physical | mixed
Default:
physical
if FADT mandates physical mode, otherwise set at build time by CONFIG_X2APIC_{PHYSICAL,MIXED}.
In the case that x2apic is in use, this option switches between modes to address APICs in the system as interrupt destinations.
= <boolean>
Default:
true
if FADT mandates physical mode or if interrupt remapping is not available,false
otherwise.
In the case that x2apic is in use, this option switches between physical and clustered mode. The default, given no hint from the FADT, is cluster mode.
WARNING: x2apic_phys
is deprecated and
superseded by x2apic-mode
. The latter takes precedence if
both are set.
= List of [ <integer> | <integer>-<integer> ]
Default:
0: the lowermost color
Specify Xen LLC color configuration. This options is available only
when CONFIG_LLC_COLORING
is enabled. Two colors are most
likely needed on platforms where private caches are physically indexed,
e.g. the L1 instruction cache of the Arm Cortex-A57.
= <size>
Default:
0
(1/32 of RAM)
Amount of RAM to set aside for the Xenheap. Must be an integer multiple of 32.
By default will use 1/32 of the RAM up to a maximum of 1GB and with a minimum of 32M, subject to a suitably aligned and sized contiguous region of memory being available.
= List of [ default | <boolean> | dom0=<bool> | domu=<bool> ]
Default:
false
on hardware known not to be vulnerable to Meltdown (e.g. AMD) Default:true
everywhere else
Override default selection of whether to isolate 64-bit PV guest page tables.
true
activates page table isolation even on hardware not
vulnerable by Meltdown for all domains.
false
deactivates page table isolation on all systems
for all domains.
default
sets the default behaviour.
With dom0
and domu
it is possible to
control page table isolation for dom0 or guest domains only.
= <boolean>
Default:
true
Permit use of the xsave/xrstor
instructions.
= dummy | flask | silo
Default: selectable via Kconfig. Depends on enabled XSM modules.
Specify which XSM module should be enabled. This option is only
available if the hypervisor was compiled with CONFIG_XSM
enabled.
dummy
: this is the default choice. Basic restriction
for common deployment (the dummy module) will be applied. It’s also used
when XSM is compiled out.flask
: this is the policy based access control. To
choose this, the separated option in kconfig must also be enabled.silo
: this will deny any unmediated communication
channels between unprivileged VMs. To choose this, the separated option
in kconfig must also be enabled.