1 HXCOMM See docs
/devel
/docs
.rst
for the format of
this file
.
3 HXCOMM Use
DEFHEADING() to define headings
in both help text and rST
.
4 HXCOMM Text between SRST and ERST is copied to the rST version and
5 HXCOMM discarded from C version
.
6 HXCOMM
DEF(option
, HAS_ARG
/0, opt_enum
, opt_help
, arch_mask
) is used to
7 HXCOMM construct option structures
, enums and help message
for specified
9 HXCOMM HXCOMM can be used
for comments
, discarded from both rST and C
.
11 DEFHEADING(Standard options
:)
13 DEF("help", 0, QEMU_OPTION_h
,
14 "-h or -help display this help and exit\n", QEMU_ARCH_ALL
)
20 DEF("version", 0, QEMU_OPTION_version
,
21 "-version display version information and exit\n", QEMU_ARCH_ALL
)
24 Display version information and exit
27 DEF("machine", HAS_ARG
, QEMU_OPTION_machine
, \
28 "-machine [type=]name[,prop[=value][,...]]\n"
29 " selects emulated machine ('-machine help' for list)\n"
30 " property accel=accel1[:accel2[:...]] selects accelerator\n"
31 " supported accelerators are kvm, xen, hvf, nvmm, whpx or tcg (default: tcg)\n"
32 " vmport=on|off|auto controls emulation of vmport (default: auto)\n"
33 " dump-guest-core=on|off include guest memory in a core dump (default=on)\n"
34 " mem-merge=on|off controls memory merge support (default: on)\n"
35 " aes-key-wrap=on|off controls support for AES key wrapping (default=on)\n"
36 " dea-key-wrap=on|off controls support for DEA key wrapping (default=on)\n"
37 " suppress-vmdesc=on|off disables self-describing migration (default=off)\n"
38 " nvdimm=on|off controls NVDIMM support (default=off)\n"
39 " memory-encryption=@var{} memory encryption object to use (default=none)\n"
40 " hmat=on|off controls ACPI HMAT support (default=off)\n"
41 " memory-backend='backend-id' specifies explicitly provided backend for main RAM (default=none)\n"
42 " cxl-fmw.0.targets.0=firsttarget,cxl-fmw.0.targets.1=secondtarget,cxl-fmw.0.size=size[,cxl-fmw.0.interleave-granularity=granularity]\n",
45 ``
-machine
[type
=]name
[,prop
=value
[,...]]``
46 Select the emulated machine by name
. Use ``
-machine help`` to list
49 For architectures which aim to support live migration compatibility
50 across releases
, each release will introduce a
new versioned machine
51 type
. For example
, the
2.8.0 release introduced machine types
52 "pc-i440fx-2.8" and
"pc-q35-2.8" for the x86\_64
/i686 architectures
.
54 To allow live migration of guests from QEMU version
2.8.0, to QEMU
55 version
2.9.0, the
2.9.0 version must support the
"pc-i440fx-2.8"
56 and
"pc-q35-2.8" machines too
. To allow users live migrating VMs to
57 skip multiple intermediate releases when upgrading
, new releases of
58 QEMU will support machine types from many previous versions
.
60 Supported machine properties are
:
62 ``accel
=accels1
[:accels2
[:...]]``
63 This is used to enable an accelerator
. Depending on the target
64 architecture
, kvm
, xen
, hvf
, nvmm
, whpx or tcg can be available
.
65 By
default, tcg is used
. If there is more than one accelerator
66 specified
, the next one is used
if the previous one fails to
69 ``vmport
=on|off|auto``
70 Enables emulation of VMWare IO port
, for vmmouse etc
. auto says
71 to select the value based on accel
. For accel
=xen the
default is
72 off otherwise the
default is on
.
74 ``dump
-guest
-core
=on|off``
75 Include guest memory
in a core dump
. The
default is on
.
78 Enables or disables memory merge support
. This feature
, when
79 supported by the host
, de
-duplicates identical memory pages
80 among VMs
instances (enabled by
default).
82 ``aes
-key
-wrap
=on|off``
83 Enables or disables AES key wrapping support on s390
-ccw hosts
.
84 This feature controls whether AES wrapping keys will be created
85 to allow execution of AES cryptographic functions
. The
default
88 ``dea
-key
-wrap
=on|off``
89 Enables or disables DEA key wrapping support on s390
-ccw hosts
.
90 This feature controls whether DEA wrapping keys will be created
91 to allow execution of DEA cryptographic functions
. The
default
95 Enables or disables NVDIMM support
. The
default is off
.
97 ``memory
-encryption
=``
98 Memory encryption object to use
. The
default is none
.
101 Enables or disables ACPI Heterogeneous Memory Attribute Table
102 (HMAT
) support
. The
default is off
.
104 ``memory
-backend
='id'``
105 An alternative to legacy ``
-mem
-path`` and ``mem
-prealloc`` options
.
106 Allows to use a memory backend as main RAM
.
111 -object memory
-backend
-file
,id
=pc
.ram
,size
=512M
,mem
-path
=/hugetlbfs
,prealloc
=on
,share
=on
112 -machine memory
-backend
=pc
.ram
115 Migration compatibility note
:
117 * as backend id one shall use value of
'default-ram-id', advertised by
118 machine
type (available via ``query
-machines`` QMP command
), if migration
119 to
/from old
QEMU (<5.0) is expected
.
120 * for machine types
4.0 and older
, user shall
121 use ``x
-use
-canonical
-path
-for-ramblock
-id
=off`` backend option
122 if migration to
/from old
QEMU (<5.0) is expected
.
127 -object memory
-backend
-ram
,id
=pc
.ram
,size
=512M
,x
-use
-canonical
-path
-for-ramblock
-id
=off
128 -machine memory
-backend
=pc
.ram
131 ``cxl
-fmw
.0.targets
.0=firsttarget
,cxl
-fmw
.0.targets
.1=secondtarget
,cxl
-fmw
.0.size
=size
[,cxl
-fmw
.0.interleave
-granularity
=granularity
]``
132 Define a CXL Fixed Memory
Window (CFMW
).
134 Described
in the CXL
2.0 ECN
: CEDT CFMWS
& QTG _DSM
.
136 They are regions of Host Physical
Addresses (HPA
) on a system which
137 may be interleaved across one or more CXL host bridges
. The system
138 software will assign particular devices into these windows and
139 configure the downstream Host
-managed Device
Memory (HDM
) decoders
140 in root ports
, switch ports and devices appropriately to meet the
141 interleave requirements before enabling the memory devices
.
143 ``targets
.X
=target`` provides the mapping to CXL host bridges
144 which may be identified by the id provided
in the
-device entry
.
145 Multiple entries are needed to specify all the targets when
146 the fixed memory window represents interleaved memory
. X is the
149 ``size
=size`` sets the size of the CFMW
. This must be a multiple of
150 256MiB
. The region will be aligned to
256MiB but the location is
151 platform and configuration dependent
.
153 ``interleave
-granularity
=granularity`` sets the granularity of
154 interleave
. Default
256 (bytes
). Only
256, 512, 1k
, 2k
,
155 4k
, 8k and
16k granularities supported
.
161 -machine cxl
-fmw
.0.targets
.0=cxl
.0,cxl
-fmw
.0.targets
.1=cxl
.1,cxl
-fmw
.0.size
=128G
,cxl
-fmw
.0.interleave
-granularity
=512
164 DEF("M", HAS_ARG
, QEMU_OPTION_M
,
165 " sgx-epc.0.memdev=memid,sgx-epc.0.node=numaid\n",
169 ``sgx
-epc
.0.memdev
=@
var{memid
},sgx
-epc
.0.node
=@
var{numaid
}``
170 Define an SGX EPC section
.
173 DEF("cpu", HAS_ARG
, QEMU_OPTION_cpu
,
174 "-cpu cpu select CPU ('-cpu help' for list)\n", QEMU_ARCH_ALL
)
177 Select CPU
model (``
-cpu help``
for list and additional feature
181 DEF("accel", HAS_ARG
, QEMU_OPTION_accel
,
182 "-accel [accel=]accelerator[,prop[=value][,...]]\n"
183 " select accelerator (kvm, xen, hvf, nvmm, whpx or tcg; use 'help' for a list)\n"
184 " igd-passthru=on|off (enable Xen integrated Intel graphics passthrough, default=off)\n"
185 " kernel-irqchip=on|off|split controls accelerated irqchip support (default=on)\n"
186 " kvm-shadow-mem=size of KVM shadow MMU in bytes\n"
187 " one-insn-per-tb=on|off (one guest instruction per TCG translation block)\n"
188 " split-wx=on|off (enable TCG split w^x mapping)\n"
189 " tb-size=n (TCG translation block cache size)\n"
190 " dirty-ring-size=n (KVM dirty ring GFN count, default 0)\n"
191 " eager-split-size=n (KVM Eager Page Split chunk size, default 0, disabled. ARM only)\n"
192 " notify-vmexit=run|internal-error|disable,notify-window=n (enable notify VM exit and set notify window, x86 only)\n"
193 " thread=single|multi (enable multi-threaded TCG)\n"
194 " device=path (KVM device path, default /dev/kvm)\n", QEMU_ARCH_ALL
)
196 ``
-accel name
[,prop
=value
[,...]]``
197 This is used to enable an accelerator
. Depending on the target
198 architecture
, kvm
, xen
, hvf
, nvmm
, whpx or tcg can be available
. By
199 default, tcg is used
. If there is more than one accelerator
200 specified
, the next one is used
if the previous one fails to
203 ``igd
-passthru
=on|off``
204 When Xen is
in use
, this option controls whether Intel
205 integrated graphics devices can be passed through to the guest
208 ``kernel
-irqchip
=on|off|split``
209 Controls KVM
in-kernel irqchip support
. The
default is full
210 acceleration of the interrupt controllers
. On x86
, split irqchip
211 reduces the kernel attack surface
, at a performance cost
for
212 non
-MSI interrupts
. Disabling the
in-kernel irqchip completely
213 is not recommended except
for debugging purposes
.
215 ``kvm
-shadow
-mem
=size``
216 Defines the size of the KVM shadow MMU
.
218 ``one
-insn
-per
-tb
=on|off``
219 Makes the TCG accelerator put only one guest instruction into
220 each translation block
. This slows down emulation a lot
, but
221 can be useful
in some situations
, such as when trying to analyse
222 the logs produced by the ``
-d`` option
.
225 Controls the use of split w^x mapping
for the TCG code generation
226 buffer
. Some operating systems require
this to be enabled
, and
in
227 such a
case this will
default on
. On other operating systems
, this
228 will
default off
, but one may enable
this for testing or debugging
.
231 Controls the
size (in MiB
) of the TCG translation block cache
.
233 ``thread
=single|multi``
234 Controls number of TCG threads
. When the TCG is multi
-threaded
235 there will be one thread per vCPU therefore taking advantage of
236 additional host cores
. The
default is to enable multi
-threading
237 where both the back
-end and front
-ends support it and no
238 incompatible TCG features have been
enabled (e
.g
.
241 ``dirty
-ring
-size
=n``
242 When the KVM accelerator is used
, it controls the size of the per
-vCPU
243 dirty page ring
buffer (number of entries
for each vCPU
). It should
244 be a value that is power of two
, and it should be
1024 or
bigger (but
245 still less than the maximum value that the kernel supports
). 4096
246 could be a good initial value
if you have no idea which is the best
.
247 Set
this value to
0 to disable the feature
. By
default, this feature
248 is
disabled (dirty
-ring
-size
=0). When enabled
, KVM will instead
249 record dirty pages
in a bitmap
.
251 ``eager
-split
-size
=n``
252 KVM
implements dirty page logging at the PAGE_SIZE granularity and
253 enabling dirty
-logging on a huge
-page requires breaking it into
254 PAGE_SIZE pages
in the first place
. KVM on ARM does
this splitting
255 lazily by
default. There are performance benefits
in doing huge
-page
256 split eagerly
, especially
in situations where TLBI costs associated
257 with
break-before
-make sequences are considerable and also
if guest
258 workloads are read intensive
. The size
here specifies how many pages
259 to
break at a time and needs to be a valid block size which is
260 1GB
/2MB
/4KB
, 32MB
/16KB and
512MB
/64KB
for 4KB
/16KB
/64KB PAGE_SIZE
261 respectively
. Be wary of specifying a higher size as it will have an
262 impact on the memory
. By
default, this feature is disabled
263 (eager
-split
-size
=0).
265 ``notify
-vmexit
=run|internal
-error|disable
,notify
-window
=n``
266 Enables or disables notify VM exit support on x86 host and specify
267 the corresponding notify window to trigger the VM exit
if enabled
.
268 ``run`` option enables the feature
. It does nothing and
continue
269 if the exit happens
. ``internal
-error`` option enables the feature
.
270 It raises a internal error
. ``disable`` option doesn
't enable the feature.
271 This feature can mitigate the CPU stuck issue due to event windows don't
272 open up
for a specified of
time (i
.e
. notify
-window
).
273 Default
: notify
-vmexit
=run
,notify
-window
=0.
276 Sets the path to the KVM device node
. Defaults to ``
/dev
/kvm``
. This
277 option can be used to pass the KVM device to use via a file descriptor
278 by setting the value to ``
/dev
/fdset
/NN``
.
282 DEF("smp", HAS_ARG
, QEMU_OPTION_smp
,
283 "-smp [[cpus=]n][,maxcpus=maxcpus][,drawers=drawers][,books=books][,sockets=sockets]\n"
284 " [,dies=dies][,clusters=clusters][,cores=cores][,threads=threads]\n"
285 " set the number of initial CPUs to 'n' [default=1]\n"
286 " maxcpus= maximum number of total CPUs, including\n"
287 " offline CPUs for hotplug, etc\n"
288 " drawers= number of drawers on the machine board\n"
289 " books= number of books in one drawer\n"
290 " sockets= number of sockets in one book\n"
291 " dies= number of dies in one socket\n"
292 " clusters= number of clusters in one die\n"
293 " cores= number of cores in one cluster\n"
294 " threads= number of threads in one core\n"
295 "Note: Different machines may have different subsets of the CPU topology\n"
296 " parameters supported, so the actual meaning of the supported parameters\n"
297 " will vary accordingly. For example, for a machine type that supports a\n"
298 " three-level CPU hierarchy of sockets/cores/threads, the parameters will\n"
299 " sequentially mean as below:\n"
300 " sockets means the number of sockets on the machine board\n"
301 " cores means the number of cores in one socket\n"
302 " threads means the number of threads in one core\n"
303 " For a particular machine type board, an expected CPU topology hierarchy\n"
304 " can be defined through the supported sub-option. Unsupported parameters\n"
305 " can also be provided in addition to the sub-option, but their values\n"
306 " must be set as 1 in the purpose of correct parsing.\n",
309 ``
-smp
[[cpus
=]n
][,maxcpus
=maxcpus
][,sockets
=sockets
][,dies
=dies
][,clusters
=clusters
][,cores
=cores
][,threads
=threads
]``
310 Simulate a SMP system with
'\ ``n``\ ' CPUs initially present on
311 the machine type board
. On boards supporting CPU hotplug
, the optional
312 '\ ``maxcpus``\ ' parameter can be set to enable further CPUs to be
313 added at runtime
. When both parameters are omitted
, the maximum number
314 of CPUs will be calculated from the provided topology members and the
315 initial CPU count will match the maximum number
. When only one of them
316 is given then the omitted one will be set to its counterpart
's value.
317 Both parameters may be specified, but the maximum number of CPUs must
318 be equal to or greater than the initial CPU count. Product of the
319 CPU topology hierarchy must be equal to the maximum number of CPUs.
320 Both parameters are subject to an upper limit that is determined by
321 the specific machine type chosen.
323 To control reporting of CPU topology information, values of the topology
324 parameters can be specified. Machines may only support a subset of the
325 parameters and different machines may have different subsets supported
326 which vary depending on capacity of the corresponding CPU targets. So
327 for a particular machine type board, an expected topology hierarchy can
328 be defined through the supported sub-option. Unsupported parameters can
329 also be provided in addition to the sub-option, but their values must be
330 set as 1 in the purpose of correct parsing.
332 Either the initial CPU count, or at least one of the topology parameters
333 must be specified. The specified parameters must be greater than zero,
334 explicit configuration like "cpus=0" is not allowed. Values for any
335 omitted parameters will be computed from those which are given.
337 For example, the following sub-option defines a CPU topology hierarchy
338 (2 sockets totally on the machine, 2 cores per socket, 2 threads per
339 core) for a machine that only supports sockets/cores/threads.
340 Some members of the option can be omitted but their values will be
341 automatically computed:
345 -smp 8,sockets=2,cores=2,threads=2,maxcpus=8
347 The following sub-option defines a CPU topology hierarchy (2 sockets
348 totally on the machine, 2 dies per socket, 2 cores per die, 2 threads
349 per core) for PC machines which support sockets/dies/cores/threads.
350 Some members of the option can be omitted but their values will be
351 automatically computed:
355 -smp 16,sockets=2,dies=2,cores=2,threads=2,maxcpus=16
357 The following sub-option defines a CPU topology hierarchy (2 sockets
358 totally on the machine, 2 clusters per socket, 2 cores per cluster,
359 2 threads per core) for ARM virt machines which support sockets/clusters
360 /cores/threads. Some members of the option can be omitted but their values
361 will be automatically computed:
365 -smp 16,sockets=2,clusters=2,cores=2,threads=2,maxcpus=16
367 Historically preference was given to the coarsest topology parameters
368 when computing missing values (ie sockets preferred over cores, which
369 were preferred over threads), however, this behaviour is considered
370 liable to change. Prior to 6.2 the preference was sockets over cores
371 over threads. Since 6.2 the preference is cores over sockets over threads.
373 For example, the following option defines a machine board with 2 sockets
374 of 1 core before 6.2 and 1 socket of 2 cores after 6.2:
380 Note: The cluster topology will only be generated in ACPI and exposed
381 to guest if it's explicitly specified
in -smp
.
384 DEF("numa", HAS_ARG
, QEMU_OPTION_numa
,
385 "-numa node[,mem=size][,cpus=firstcpu[-lastcpu]][,nodeid=node][,initiator=node]\n"
386 "-numa node[,memdev=id][,cpus=firstcpu[-lastcpu]][,nodeid=node][,initiator=node]\n"
387 "-numa dist,src=source,dst=destination,val=distance\n"
388 "-numa cpu,node-id=node[,socket-id=x][,core-id=y][,thread-id=z]\n"
389 "-numa hmat-lb,initiator=node,target=node,hierarchy=memory|first-level|second-level|third-level,data-type=access-latency|read-latency|write-latency[,latency=lat][,bandwidth=bw]\n"
390 "-numa hmat-cache,node-id=node,size=size,level=level[,associativity=none|direct|complex][,policy=none|write-back|write-through][,line=size]\n",
393 ``
-numa node
[,mem
=size
][,cpus
=firstcpu
[-lastcpu
]][,nodeid
=node
][,initiator
=initiator
]``
395 ``
-numa node
[,memdev
=id
][,cpus
=firstcpu
[-lastcpu
]][,nodeid
=node
][,initiator
=initiator
]``
397 ``
-numa dist
,src
=source
,dst
=destination
,val
=distance``
399 ``
-numa cpu
,node
-id
=node
[,socket
-id
=x
][,core
-id
=y
][,thread
-id
=z
]``
401 ``
-numa hmat
-lb
,initiator
=node
,target
=node
,hierarchy
=hierarchy
,data
-type
=type
[,latency
=lat
][,bandwidth
=bw
]``
403 ``
-numa hmat
-cache
,node
-id
=node
,size
=size
,level
=level
[,associativity
=str
][,policy
=str
][,line
=size
]``
404 Define a NUMA node and assign RAM and VCPUs to it
. Set the NUMA
405 distance from a source node to a destination node
. Set the ACPI
406 Heterogeneous Memory Attributes
for the given nodes
.
408 Legacy VCPU assignment uses
'\ ``cpus``\ ' option where firstcpu and
409 lastcpu are CPU indexes
. Each
'\ ``cpus``\ ' option represent a
410 contiguous range of CPU
indexes (or a single VCPU
if lastcpu is
411 omitted
). A non
-contiguous set of VCPUs can be represented by
412 providing multiple
'\ ``cpus``\ ' options
. If
'\ ``cpus``\ ' is
413 omitted on all nodes
, VCPUs are automatically split between them
.
415 For example
, the following option assigns VCPUs
0, 1, 2 and
5 to a
420 -numa node
,cpus
=0-2,cpus
=5
422 '\ ``cpu``\ ' option is a
new alternative to
'\ ``cpus``\ ' option
423 which uses
'\ ``socket-id|core-id|thread-id``\ ' properties to
424 assign CPU objects to a node
using topology layout properties of
425 CPU
. The set of properties is machine specific
, and depends on used
426 machine type
/'\ ``smp``\ ' options
. It could be queried with
427 '\ ``hotpluggable-cpus``\ ' monitor command
. '\ ``node-id``\ '
428 property specifies node to which CPU object will be assigned
, it
's
429 required for node to be declared with '\ ``node``\
' option before
430 it's used with
'\ ``cpu``\ ' option
.
437 -smp
1,sockets
=2,maxcpus
=2 \
438 -numa node
,nodeid
=0 -numa node
,nodeid
=1 \
439 -numa cpu
,node
-id
=0,socket
-id
=0 -numa cpu
,node
-id
=1,socket
-id
=1
441 '\ ``memdev``\ ' option assigns RAM from a given memory backend
442 device to a node
. It is recommended to use
'\ ``memdev``\ ' option
443 over legacy
'\ ``mem``\ ' option
. This is because
'\ ``memdev``\ '
444 option provides better performance and more control over the
445 backend
's RAM (e.g. '\ ``prealloc``\
' parameter of
446 '\ ``
-memory
-backend
-ram``\
' allows memory preallocation).
448 For compatibility reasons, legacy '\ ``mem``\
' option is
449 supported in 5.0 and older machine types. Note that '\ ``mem``\
'
450 and '\ ``memdev``\
' are mutually exclusive. If one node uses
451 '\ ``memdev``\
', the rest nodes have to use '\ ``memdev``\
'
452 option, and vice versa.
454 Users must specify memory for all NUMA nodes by '\ ``memdev``\
'
455 (or legacy '\ ``mem``\
' if available). In QEMU 5.2, the support
456 for '\ ``
-numa node``\
' without memory specified was removed.
458 '\ ``initiator``\
' is an additional option that points to an
459 initiator NUMA node that has best performance (the lowest latency or
460 largest bandwidth) to this NUMA node. Note that this option can be
461 set only when the machine property 'hmat
' is set to 'on
'.
463 Following example creates a machine with 2 NUMA nodes, node 0 has
464 CPU. node 1 has only memory, and its initiator is node 0. Note that
465 because node 0 has CPU, by default the initiator of node 0 is itself
471 -m 2G,slots=2,maxmem=4G \
472 -object memory-backend-ram,size=1G,id=m0 \
473 -object memory-backend-ram,size=1G,id=m1 \
474 -numa node,nodeid=0,memdev=m0 \
475 -numa node,nodeid=1,memdev=m1,initiator=0 \
476 -smp 2,sockets=2,maxcpus=2 \
477 -numa cpu,node-id=0,socket-id=0 \
478 -numa cpu,node-id=0,socket-id=1
480 source and destination are NUMA node IDs. distance is the NUMA
481 distance from source to destination. The distance from a node to
482 itself is always 10. If any pair of nodes is given a distance, then
483 all pairs must be given distances. Although, when distances are only
484 given in one direction for each pair of nodes, then the distances in
485 the opposite directions are assumed to be the same. If, however, an
486 asymmetrical pair of distances is given for even one node pair, then
487 all node pairs must be provided distance values for both directions,
488 even when they are symmetrical. When a node is unreachable from
489 another node, set the pair's distance to
255.
491 Note that the
-``numa`` option doesn
't allocate any of the specified
492 resources, it just assigns existing resources to NUMA nodes. This
493 means that one still has to use the ``-m``, ``-smp`` options to
494 allocate RAM and VCPUs respectively.
496 Use '\ ``hmat
-lb``\
' to set System Locality Latency and Bandwidth
497 Information between initiator and target NUMA nodes in ACPI
498 Heterogeneous Attribute Memory Table (HMAT). Initiator NUMA node can
499 create memory requests, usually it has one or more processors.
500 Target NUMA node contains addressable memory.
502 In '\ ``hmat
-lb``\
' option, node are NUMA node IDs. hierarchy is
503 the memory hierarchy of the target NUMA node: if hierarchy is
504 'memory
', the structure represents the memory performance; if
505 hierarchy is 'first
-level\|second
-level\|third
-level
', this
506 structure represents aggregated performance of memory side caches
507 for each domain. type of 'data
-type
' is type of data represented by
508 this structure instance: if 'hierarchy
' is 'memory
', 'data
-type
' is
509 'access\|read\|write
' latency or 'access\|read\|write
' bandwidth of
510 the target memory; if 'hierarchy
' is
511 'first
-level\|second
-level\|third
-level
', 'data
-type
' is
512 'access\|read\|write
' hit latency or 'access\|read\|write
' hit
513 bandwidth of the target memory side cache.
515 lat is latency value in nanoseconds. bw is bandwidth value, the
516 possible value and units are NUM[M\|G\|T], mean that the bandwidth
517 value are NUM byte per second (or MB/s, GB/s or TB/s depending on
518 used suffix). Note that if latency or bandwidth value is 0, means
519 the corresponding latency or bandwidth information is not provided.
521 In '\ ``hmat
-cache``\
' option, node-id is the NUMA-id of the memory
522 belongs. size is the size of memory side cache in bytes. level is
523 the cache level described in this structure, note that the cache
524 level 0 should not be used with '\ ``hmat
-cache``\
' option.
525 associativity is the cache associativity, the possible value is
526 'none
/direct(direct
-mapped
)/complex(complex cache indexing
)'. policy
527 is the write policy. line is the cache Line size in bytes.
529 For example, the following options describe 2 NUMA nodes. Node 0 has
530 2 cpus and a ram, node 1 has only a ram. The processors in node 0
531 access memory in node 0 with access-latency 5 nanoseconds,
532 access-bandwidth is 200 MB/s; The processors in NUMA node 0 access
533 memory in NUMA node 1 with access-latency 10 nanoseconds,
534 access-bandwidth is 100 MB/s. And for memory side cache information,
535 NUMA node 0 and 1 both have 1 level memory cache, size is 10KB,
536 policy is write-back, the cache Line size is 8 bytes:
542 -object memory-backend-ram,size=1G,id=m0 \
543 -object memory-backend-ram,size=1G,id=m1 \
544 -smp 2,sockets=2,maxcpus=2 \
545 -numa node,nodeid=0,memdev=m0 \
546 -numa node,nodeid=1,memdev=m1,initiator=0 \
547 -numa cpu,node-id=0,socket-id=0 \
548 -numa cpu,node-id=0,socket-id=1 \
549 -numa hmat-lb,initiator=0,target=0,hierarchy=memory,data-type=access-latency,latency=5 \
550 -numa hmat-lb,initiator=0,target=0,hierarchy=memory,data-type=access-bandwidth,bandwidth=200M \
551 -numa hmat-lb,initiator=0,target=1,hierarchy=memory,data-type=access-latency,latency=10 \
552 -numa hmat-lb,initiator=0,target=1,hierarchy=memory,data-type=access-bandwidth,bandwidth=100M \
553 -numa hmat-cache,node-id=0,size=10K,level=1,associativity=direct,policy=write-back,line=8 \
554 -numa hmat-cache,node-id=1,size=10K,level=1,associativity=direct,policy=write-back,line=8
557 DEF("add-fd", HAS_ARG, QEMU_OPTION_add_fd,
558 "-add-fd fd=fd,set=set[,opaque=opaque]\n"
559 " Add 'fd
' to fd 'set
'\n", QEMU_ARCH_ALL)
561 ``-add-fd fd=fd,set=set[,opaque=opaque]``
562 Add a file descriptor to an fd set. Valid options are:
565 This option defines the file descriptor of which a duplicate is
566 added to fd set. The file descriptor cannot be stdin, stdout, or
570 This option defines the ID of the fd set to add the file
574 This option defines a free-form string that can be used to
577 You can open an image using pre-opened file descriptors from an fd
583 -add-fd fd=3,set=2,opaque="rdwr:/path/to/file" \\
584 -add-fd fd=4,set=2,opaque="rdonly:/path/to/file" \\
585 -drive file=/dev/fdset/2,index=0,media=disk
588 DEF("set", HAS_ARG, QEMU_OPTION_set,
589 "-set group.id.arg=value\n"
590 " set <arg> parameter for item <id> of type <group>\n"
591 " i.e. -set drive.$id.file=/path/to/image\n", QEMU_ARCH_ALL)
593 ``-set group.id.arg=value``
594 Set parameter arg for item id of type group
597 DEF("global", HAS_ARG, QEMU_OPTION_global,
598 "-global driver.property=value\n"
599 "-global driver=driver,property=property,value=value\n"
600 " set a global default for a driver property\n",
603 ``-global driver.prop=value``
605 ``-global driver=driver,property=property,value=value``
606 Set default value of driver's property prop to value
, e
.g
.:
610 |qemu_system_x86|
-global ide
-hd
.physical_block_size
=4096 disk
-image
.img
612 In particular
, you can use
this to set driver properties
for devices
613 which are created automatically by the machine model
. To create a
614 device which is not created automatically and set properties on it
,
617 -global driver
.prop
=value is shorthand
for -global
618 driver
=driver
,property
=prop
,value
=value
. The longhand syntax works
619 even when driver contains a dot
.
622 DEF("boot", HAS_ARG
, QEMU_OPTION_boot
,
623 "-boot [order=drives][,once=drives][,menu=on|off]\n"
624 " [,splash=sp_name][,splash-time=sp_time][,reboot-timeout=rb_time][,strict=on|off]\n"
625 " 'drives': floppy (a), hard disk (c), CD-ROM (d), network (n)\n"
626 " 'sp_name': the file's name that would be passed to bios as logo picture, if menu=on\n"
627 " 'sp_time': the period that splash picture last if menu=on, unit is ms\n"
628 " 'rb_timeout': the timeout before guest reboot when boot failed, unit is ms\n",
631 ``
-boot
[order
=drives
][,once
=drives
][,menu
=on|off
][,splash
=sp_name
][,splash
-time
=sp_time
][,reboot
-timeout
=rb_timeout
][,strict
=on|off
]``
632 Specify boot order drives as a string of drive letters
. Valid drive
633 letters depend on the target architecture
. The x86 PC uses
: a
, b
634 (floppy
1 and
2), c (first hard disk
), d (first CD
-ROM
), n
-p
635 (Etherboot from network adapter
1-4), hard disk boot is the
default.
636 To apply a particular boot order only on the first startup
, specify
637 it via ``once``
. Note that the ``order`` or ``once`` parameter
638 should not be used together with the ``bootindex`` property of
639 devices
, since the firmware implementations normally
do not support
640 both at the same time
.
642 Interactive boot menus
/prompts can be enabled via ``menu
=on`` as far
643 as firmware
/BIOS supports them
. The
default is non
-interactive boot
.
645 A splash picture could be passed to bios
, enabling user to show it
646 as logo
, when option splash
=sp\_name is given and menu
=on
, If
647 firmware
/BIOS supports them
. Currently Seabios
for X86 system
648 support it
. limitation
: The splash file could be a jpeg file or a
649 BMP file
in 24 BPP
format(true color
). The resolution should be
650 supported by the SVGA mode
, so the recommended is
320x240
, 640x480
,
653 A timeout could be passed to bios
, guest will pause
for rb\_timeout
654 ms when boot failed
, then reboot
. If rb\_timeout is
'-1', guest will
655 not reboot
, qemu passes
'-1' to bios by
default. Currently Seabios
656 for X86 system support it
.
658 Do strict boot via ``strict
=on`` as far as firmware
/BIOS supports
659 it
. This only effects when boot priority is changed by bootindex
660 options
. The
default is non
-strict boot
.
664 #
try to boot from network first
, then from hard disk
665 |qemu_system_x86|
-boot order
=nc
666 # boot from CD
-ROM first
, switch back to
default order after reboot
667 |qemu_system_x86|
-boot once
=d
668 # boot with a splash picture
for 5 seconds
.
669 |qemu_system_x86|
-boot menu
=on
,splash
=/root
/boot
.bmp
,splash
-time
=5000
671 Note
: The legacy format
'-boot drives' is still supported but its
672 use is discouraged as it may be removed from future versions
.
675 DEF("m", HAS_ARG
, QEMU_OPTION_m
,
676 "-m [size=]megs[,slots=n,maxmem=size]\n"
677 " configure guest RAM\n"
678 " size: initial amount of guest memory\n"
679 " slots: number of hotplug slots (default: none)\n"
680 " maxmem: maximum amount of guest memory (default: none)\n"
681 " Note: Some architectures might enforce a specific granularity\n",
684 ``
-m
[size
=]megs
[,slots
=n
,maxmem
=size
]``
685 Sets guest startup RAM size to megs megabytes
. Default is
128 MiB
.
686 Optionally
, a suffix of
"M" or
"G" can be used to signify a value
in
687 megabytes or gigabytes respectively
. Optional pair slots
, maxmem
688 could be used to set amount of hotpluggable memory slots and maximum
689 amount of memory
. Note that maxmem must be aligned to the page size
.
691 For example
, the following command
-line sets the guest startup RAM
692 size to
1GB
, creates
3 slots to hotplug additional memory and sets
693 the maximum memory the guest can reach to
4GB
:
697 |qemu_system|
-m
1G
,slots
=3,maxmem
=4G
699 If slots and maxmem are not specified
, memory hotplug won
't be
700 enabled and the guest startup RAM will never increase.
703 DEF("mem-path", HAS_ARG, QEMU_OPTION_mempath,
704 "-mem-path FILE provide backing storage for guest RAM\n", QEMU_ARCH_ALL)
707 Allocate guest RAM from a temporarily created file in path.
710 DEF("mem-prealloc", 0, QEMU_OPTION_mem_prealloc,
711 "-mem-prealloc preallocate guest memory (use with -mem-path)\n",
715 Preallocate memory when using -mem-path.
718 DEF("k", HAS_ARG, QEMU_OPTION_k,
719 "-k language use keyboard layout (for example 'fr
' for French)\n",
723 Use keyboard layout language (for example ``fr`` for French). This
724 option is only needed where it is not easy to get raw PC keycodes
725 (e.g. on Macs, with some X11 servers or with a VNC or curses
726 display). You don't normally need to use it on PC
/Linux or
729 The available layouts are
:
733 ar de
-ch es fo fr
-ca hu ja mk no pt
-br sv
734 da en
-gb et fr fr
-ch is lt nl pl ru th
735 de en
-us fi fr
-be hr it lv nl
-be pt sl tr
737 The
default is ``en
-us``
.
741 DEF("audio", HAS_ARG
, QEMU_OPTION_audio
,
742 "-audio [driver=]driver[,prop[=value][,...]]\n"
743 " specifies default audio backend when `audiodev` is not\n"
744 " used to create a machine or sound device;"
745 " options are the same as for -audiodev\n"
746 "-audio [driver=]driver,model=value[,prop[=value][,...]]\n"
747 " specifies the audio backend and device to use;\n"
748 " apart from 'model', options are the same as for -audiodev.\n"
749 " use '-audio model=help' to show possible devices.\n",
752 ``
-audio
[driver
=]driver
[,model
=value
][,prop
[=value
][,...]]``
753 If the ``model`` option is specified
, ``
-audio`` is a shortcut
754 for configuring both the guest audio hardware and the host audio
755 backend
in one go
. The guest hardware model can be set with
756 ``model
=modelname``
. Use ``model
=help`` to list the available
759 The following two example
do exactly the same
, to show how ``
-audio``
760 can be used to shorten the command line length
:
764 |qemu_system|
-audiodev pa
,id
=pa
-device sb16
,audiodev
=pa
765 |qemu_system|
-audio pa
,model
=sb16
767 If the ``model`` option is not specified
, ``
-audio`` is used to
768 configure a
default audio backend that will be used whenever the
769 ``audiodev`` property is not set on a device or machine
. In
770 particular
, ``
-audio none`` ensures that no audio is produced even
771 for machines that have embedded sound hardware
.
773 In both cases
, the driver option is the same as with the corresponding
774 ``
-audiodev`` option below
. Use ``driver
=help`` to list the available
779 DEF("audiodev", HAS_ARG
, QEMU_OPTION_audiodev
,
780 "-audiodev [driver=]driver,id=id[,prop[=value][,...]]\n"
781 " specifies the audio backend to use\n"
782 " Use ``-audiodev help`` to list the available drivers\n"
783 " id= identifier of the backend\n"
784 " timer-period= timer period in microseconds\n"
785 " in|out.mixing-engine= use mixing engine to mix streams inside QEMU\n"
786 " in|out.fixed-settings= use fixed settings for host audio\n"
787 " in|out.frequency= frequency to use with fixed settings\n"
788 " in|out.channels= number of channels to use with fixed settings\n"
789 " in|out.format= sample format to use with fixed settings\n"
790 " valid values: s8, s16, s32, u8, u16, u32, f32\n"
791 " in|out.voices= number of voices to use\n"
792 " in|out.buffer-length= length of buffer in microseconds\n"
793 "-audiodev none,id=id,[,prop[=value][,...]]\n"
794 " dummy driver that discards all output\n"
795 #ifdef CONFIG_AUDIO_ALSA
796 "-audiodev alsa,id=id[,prop[=value][,...]]\n"
797 " in|out.dev= name of the audio device to use\n"
798 " in|out.period-length= length of period in microseconds\n"
799 " in|out.try-poll= attempt to use poll mode\n"
800 " threshold= threshold (in microseconds) when playback starts\n"
802 #ifdef CONFIG_AUDIO_COREAUDIO
803 "-audiodev coreaudio,id=id[,prop[=value][,...]]\n"
804 " in|out.buffer-count= number of buffers\n"
806 #ifdef CONFIG_AUDIO_DSOUND
807 "-audiodev dsound,id=id[,prop[=value][,...]]\n"
808 " latency= add extra latency to playback in microseconds\n"
810 #ifdef CONFIG_AUDIO_OSS
811 "-audiodev oss,id=id[,prop[=value][,...]]\n"
812 " in|out.dev= path of the audio device to use\n"
813 " in|out.buffer-count= number of buffers\n"
814 " in|out.try-poll= attempt to use poll mode\n"
815 " try-mmap= try using memory mapped access\n"
816 " exclusive= open device in exclusive mode\n"
817 " dsp-policy= set timing policy (0..10), -1 to use fragment mode\n"
819 #ifdef CONFIG_AUDIO_PA
820 "-audiodev pa,id=id[,prop[=value][,...]]\n"
821 " server= PulseAudio server address\n"
822 " in|out.name= source/sink device name\n"
823 " in|out.latency= desired latency in microseconds\n"
825 #ifdef CONFIG_AUDIO_PIPEWIRE
826 "-audiodev pipewire,id=id[,prop[=value][,...]]\n"
827 " in|out.name= source/sink device name\n"
828 " in|out.stream-name= name of pipewire stream\n"
829 " in|out.latency= desired latency in microseconds\n"
831 #ifdef CONFIG_AUDIO_SDL
832 "-audiodev sdl,id=id[,prop[=value][,...]]\n"
833 " in|out.buffer-count= number of buffers\n"
835 #ifdef CONFIG_AUDIO_SNDIO
836 "-audiodev sndio,id=id[,prop[=value][,...]]\n"
839 "-audiodev spice,id=id[,prop[=value][,...]]\n"
841 #ifdef CONFIG_DBUS_DISPLAY
842 "-audiodev dbus,id=id[,prop[=value][,...]]\n"
844 "-audiodev wav,id=id[,prop[=value][,...]]\n"
845 " path= path of wav file to record\n",
848 ``
-audiodev
[driver
=]driver
,id
=id
[,prop
[=value
][,...]]``
849 Adds a
new audio backend driver identified by id
. There are global
850 and driver specific properties
. Some values can be set differently
851 for input and output
, they
're marked with ``in|out.``. You can set
852 the input's property with ``
in.prop`` and the output
's property with
853 ``out.prop``. For example:
857 -audiodev alsa,id=example,in.frequency=44110,out.frequency=8000
858 -audiodev alsa,id=example,out.channels=1 # leaves in.channels unspecified
860 NOTE: parameter validation is known to be incomplete, in many cases
861 specifying an invalid option causes QEMU to print an error message
862 and continue emulation without sound.
864 Valid global options are:
867 Identifies the audio backend.
869 ``timer-period=period``
870 Sets the timer period used by the audio subsystem in
871 microseconds. Default is 10000 (10 ms).
873 ``in|out.mixing-engine=on|off``
874 Use QEMU's mixing engine to mix all streams inside QEMU and
875 convert audio formats when not supported by the backend
. When
876 off
, fixed
-settings must be off too
. Note that disabling
this
877 option means that the selected backend must support multiple
878 streams and the audio formats used by the virtual cards
,
879 otherwise you
'll get no sound. It's not recommended to disable
880 this option unless you want to use
5.1 or
7.1 audio
, as mixing
881 engine only supports mono and stereo audio
. Default is on
.
883 ``
in|out
.fixed
-settings
=on|off``
884 Use fixed settings
for host audio
. When off
, it will change
885 based on how the guest opens the sound card
. In
this case you
886 must not specify frequency
, channels or format
. Default is on
.
888 ``
in|out
.frequency
=frequency``
889 Specify the frequency to use when
using fixed
-settings
. Default
892 ``
in|out
.channels
=channels``
893 Specify the number of channels to use when
using fixed
-settings
.
894 Default is
2 (stereo
).
896 ``
in|out
.format
=format``
897 Specify the sample format to use when
using fixed
-settings
.
898 Valid values are
: ``s8``
, ``s16``
, ``s32``
, ``u8``
, ``u16``
,
899 ``u32``
, ``f32``
. Default is ``s16``
.
901 ``
in|out
.voices
=voices``
902 Specify the number of voices to use
. Default is
1.
904 ``
in|out
.buffer
-length
=usecs``
905 Sets the size of the buffer
in microseconds
.
907 ``
-audiodev none
,id
=id
[,prop
[=value
][,...]]``
908 Creates a dummy backend that discards all outputs
. This backend has
909 no backend specific properties
.
911 ``
-audiodev alsa
,id
=id
[,prop
[=value
][,...]]``
912 Creates backend
using the ALSA
. This backend is only available on
915 ALSA specific options are
:
917 ``
in|out
.dev
=device``
918 Specify the ALSA device to use
for input and
/or output
. Default
921 ``
in|out
.period
-length
=usecs``
922 Sets the period length
in microseconds
.
924 ``
in|out
.try-poll
=on|off``
925 Attempt to use poll mode with the device
. Default is on
.
927 ``threshold
=threshold``
928 Threshold (in microseconds
) when playback starts
. Default is
0.
930 ``
-audiodev coreaudio
,id
=id
[,prop
[=value
][,...]]``
931 Creates a backend
using Apple
's Core Audio. This backend is only
932 available on Mac OS and only supports playback.
934 Core Audio specific options are:
936 ``in|out.buffer-count=count``
937 Sets the count of the buffers.
939 ``-audiodev dsound,id=id[,prop[=value][,...]]``
940 Creates a backend using Microsoft's DirectSound
. This backend is
941 only available on Windows and only supports playback
.
943 DirectSound specific options are
:
946 Add extra usecs microseconds latency to playback
. Default is
949 ``
-audiodev oss
,id
=id
[,prop
[=value
][,...]]``
950 Creates a backend
using OSS
. This backend is available on most
953 OSS specific options are
:
955 ``
in|out
.dev
=device``
956 Specify the file name of the OSS device to use
. Default is
959 ``
in|out
.buffer
-count
=count``
960 Sets the count of the buffers
.
962 ``
in|out
.try-poll
=on|of``
963 Attempt to use poll mode with the device
. Default is on
.
966 Try
using memory mapped device access
. Default is off
.
969 Open the device
in exclusive
mode (vmix won
't work in this
970 case). Default is off.
972 ``dsp-policy=policy``
973 Sets the timing policy (between 0 and 10, where smaller number
974 means smaller latency but higher CPU usage). Use -1 to use
975 buffer sizes specified by ``buffer`` and ``buffer-count``. This
976 option is ignored if you do not have OSS 4. Default is 5.
978 ``-audiodev pa,id=id[,prop[=value][,...]]``
979 Creates a backend using PulseAudio. This backend is available on
982 PulseAudio specific options are:
985 Sets the PulseAudio server to connect to.
988 Use the specified source/sink for recording/playback.
990 ``in|out.latency=usecs``
991 Desired latency in microseconds. The PulseAudio server will try
992 to honor this value but actual latencies may be lower or higher.
994 ``-audiodev pipewire,id=id[,prop[=value][,...]]``
995 Creates a backend using PipeWire. This backend is available on
998 PipeWire specific options are:
1000 ``in|out.latency=usecs``
1001 Desired latency in microseconds.
1003 ``in|out.name=sink``
1004 Use the specified source/sink for recording/playback.
1006 ``in|out.stream-name``
1007 Specify the name of pipewire stream.
1009 ``-audiodev sdl,id=id[,prop[=value][,...]]``
1010 Creates a backend using SDL. This backend is available on most
1011 systems, but you should use your platform's native backend
if
1014 SDL specific options are
:
1016 ``
in|out
.buffer
-count
=count``
1017 Sets the count of the buffers
.
1019 ``
-audiodev sndio
,id
=id
[,prop
[=value
][,...]]``
1020 Creates a backend
using SNDIO
. This backend is available on
1021 OpenBSD and most other Unix
-like systems
.
1023 Sndio specific options are
:
1025 ``
in|out
.dev
=device``
1026 Specify the sndio device to use
for input and
/or output
. Default
1029 ``
in|out
.latency
=usecs``
1030 Sets the desired period length
in microseconds
.
1032 ``
-audiodev spice
,id
=id
[,prop
[=value
][,...]]``
1033 Creates a backend that sends audio through SPICE
. This backend
1034 requires ``
-spice`` and automatically selected
in that
case, so
1035 usually you can ignore
this option
. This backend has no backend
1036 specific properties
.
1038 ``
-audiodev wav
,id
=id
[,prop
[=value
][,...]]``
1039 Creates a backend that writes audio to a WAV file
.
1041 Backend specific options are
:
1044 Write recorded audio into the specified file
. Default is
1048 DEF("device", HAS_ARG
, QEMU_OPTION_device
,
1049 "-device driver[,prop[=value][,...]]\n"
1050 " add device (based on driver)\n"
1051 " prop=value,... sets driver properties\n"
1052 " use '-device help' to print all possible drivers\n"
1053 " use '-device driver,help' to print all possible properties\n",
1056 ``
-device driver
[,prop
[=value
][,...]]``
1057 Add device driver
. prop
=value sets driver properties
. Valid
1058 properties depend on the driver
. To get help on possible drivers and
1059 properties
, use ``
-device help`` and ``
-device driver
,help``
.
1063 ``
-device ipmi
-bmc
-sim
,id
=id
[,prop
[=value
][,...]]``
1064 Add an IPMI BMC
. This is a simulation of a hardware management
1065 interface processor that normally sits on a system
. It provides a
1066 watchdog and the ability to reset and power control the system
. You
1067 need to connect
this to an IPMI
interface to make it useful
1069 The IPMI slave address to use
for the BMC
. The
default is
0x20. This
1070 address is the BMC
's address on the I2C network of management
1071 controllers. If you don't know what
this means
, it is safe to ignore
1075 The BMC id
for interfaces to use
this device
.
1078 Define slave address to use
for the BMC
. The
default is
0x20.
1081 file containing raw Sensor Data
Records (SDR
) data
. The
default
1085 size of a Field Replaceable
Unit (FRU
) area
. The
default is
1088 ``frudatafile
=file``
1089 file containing raw Field Replaceable
Unit (FRU
) inventory data
.
1090 The
default is none
.
1093 value
for the GUID
for the BMC
, in standard UUID format
. If
this
1094 is set
, get
"Get GUID" command to the BMC will
return it
.
1095 Otherwise
"Get GUID" will
return an error
.
1097 ``
-device ipmi
-bmc
-extern,id
=id
,chardev
=id
[,slave_addr
=val
]``
1098 Add a connection to an external IPMI BMC simulator
. Instead of
1099 locally emulating the BMC like the above item
, instead connect to an
1100 external entity that provides the IPMI services
.
1102 A connection is made to an external BMC simulator
. If you
do this,
1103 it is strongly recommended that you use the
"reconnect=" chardev
1104 option to reconnect to the simulator
if the connection is lost
. Note
1105 that
if this is not used carefully
, it can be a security issue
, as
1106 the
interface has the ability to send resets
, NMIs
, and power off
1107 the VM
. It
's best if QEMU makes a connection to an external
1108 simulator running on a secure port on localhost, so neither the
1109 simulator nor QEMU is exposed to any outside network.
1111 See the "lanserv/README.vm" file in the OpenIPMI library for more
1112 details on the external interface.
1114 ``-device isa-ipmi-kcs,bmc=id[,ioport=val][,irq=val]``
1115 Add a KCS IPMI interface on the ISA bus. This also adds a
1116 corresponding ACPI and SMBIOS entries, if appropriate.
1119 The BMC to connect to, one of ipmi-bmc-sim or ipmi-bmc-extern
1123 Define the I/O address of the interface. The default is 0xca0
1127 Define the interrupt to use. The default is 5. To disable
1128 interrupts, set this to 0.
1130 ``-device isa-ipmi-bt,bmc=id[,ioport=val][,irq=val]``
1131 Like the KCS interface, but defines a BT interface. The default port
1132 is 0xe4 and the default interrupt is 5.
1134 ``-device pci-ipmi-kcs,bmc=id``
1135 Add a KCS IPMI interface on the PCI bus.
1138 The BMC to connect to, one of ipmi-bmc-sim or ipmi-bmc-extern above.
1140 ``-device pci-ipmi-bt,bmc=id``
1141 Like the KCS interface, but defines a BT interface on the PCI bus.
1143 ``-device intel-iommu[,option=...]``
1144 This is only supported by ``-machine q35``, which will enable Intel VT-d
1145 emulation within the guest. It supports below options:
1147 ``intremap=on|off`` (default: auto)
1148 This enables interrupt remapping feature. It's required to enable
1149 complete x2apic
. Currently it only supports kvm kernel
-irqchip modes
1150 ``off`` or ``split``
, while full kernel
-irqchip is not yet supported
.
1151 The
default value is
"auto", which will be decided by the mode of
1154 ``caching
-mode
=on|off``
(default: off
)
1155 This enables caching mode
for the VT
-d emulated device
. When
1156 caching
-mode is enabled
, each guest DMA buffer mapping will generate an
1157 IOTLB invalidation from the guest IOMMU driver to the vIOMMU device
in
1158 a synchronous way
. It is required
for ``
-device vfio
-pci`` to work
1159 with the VT
-d device
, because host assigned devices requires to setup
1160 the DMA mapping on the host before guest DMA starts
.
1162 ``device
-iotlb
=on|off``
(default: off
)
1163 This enables device
-iotlb capability
for the emulated VT
-d device
. So
1164 far virtio
/vhost should be the only real user
for this parameter
,
1165 paired with ats
=on configured
for the device
.
1167 ``aw
-bits
=39|
48``
(default: 39)
1168 This decides the address width of IOVA address space
. The address
1169 space has
39 bits width
for 3-level IOMMU page tables
, and
48 bits
for
1170 4-level IOMMU page tables
.
1172 Please also refer to the wiki page
for general scenarios of VT
-d
1173 emulation
in QEMU
: https
://wiki.qemu.org/Features/VT-d.
1175 ``
-device virtio
-iommu
-pci
[,option
=...]``
1176 This is only supported by ``
-machine q35``
(x86_64
) and ``
-machine virt``
(ARM
).
1177 It supports below options
:
1179 ``granule
=val``
(possible values are
4k
, 8k
, 16k
, 64k and host
; default: host
)
1180 This decides the
default granule to be be exposed by the
1181 virtio
-iommu
. If host
, the granule matches the host page size
.
1183 ``aw
-bits
=val``
(val between
32 and
64, default depends on machine
)
1184 This decides the address width of the IOVA address space
.
1188 DEF("name", HAS_ARG
, QEMU_OPTION_name
,
1189 "-name string1[,process=string2][,debug-threads=on|off]\n"
1190 " set the name of the guest\n"
1191 " string1 sets the window title and string2 the process name\n"
1192 " When debug-threads is enabled, individual threads are given a separate name\n"
1193 " NOTE: The thread names are for debugging and not a stable API.\n",
1197 Sets the name of the guest
. This name will be displayed
in the SDL
1198 window caption
. The name will also be used
for the VNC server
. Also
1199 optionally set the top visible process name
in Linux
. Naming of
1200 individual threads can also be enabled on Linux to aid debugging
.
1203 DEF("uuid", HAS_ARG
, QEMU_OPTION_uuid
,
1204 "-uuid %08x-%04x-%04x-%04x-%012x\n"
1205 " specify machine UUID\n", QEMU_ARCH_ALL
)
1213 DEFHEADING(Block device options
:)
1216 The QEMU block device handling options have a long history and
1217 have gone through several iterations as the feature set and complexity
1218 of the block layer have grown
. Many online guides to QEMU often
1219 reference older and deprecated options
, which can lead to confusion
.
1221 The most explicit way to describe disks is to use a combination of
1222 ``
-device`` to specify the hardware device and ``
-blockdev`` to
1223 describe the backend
. The device defines what the guest sees and the
1224 backend describes how QEMU handles the data
. It is the only guaranteed
1225 stable
interface for describing block devices and as such is
1226 recommended
for management tools and scripting
.
1228 The ``
-drive`` option combines the device and backend into a single
1229 command line option which is a more human friendly
. There is however no
1230 interface stability guarantee although some older board models still
1231 need updating to work with the modern blockdev forms
.
1233 Older options like ``
-hda`` are essentially macros which expand into
1234 ``
-drive`` options
for various drive interfaces
. The original forms
1235 bake
in a lot of assumptions from the days when QEMU was emulating a
1236 legacy PC
, they are not recommended
for modern configurations
.
1240 DEF("fda", HAS_ARG
, QEMU_OPTION_fda
,
1241 "-fda/-fdb file use 'file' as floppy disk 0/1 image\n", QEMU_ARCH_ALL
)
1242 DEF("fdb", HAS_ARG
, QEMU_OPTION_fdb
, "", QEMU_ARCH_ALL
)
1247 Use file as floppy disk
0/1 image (see the
:ref
:`disk images` chapter
in
1248 the System Emulation Users Guide
).
1251 DEF("hda", HAS_ARG
, QEMU_OPTION_hda
,
1252 "-hda/-hdb file use 'file' as hard disk 0/1 image\n", QEMU_ARCH_ALL
)
1253 DEF("hdb", HAS_ARG
, QEMU_OPTION_hdb
, "", QEMU_ARCH_ALL
)
1254 DEF("hdc", HAS_ARG
, QEMU_OPTION_hdc
,
1255 "-hdc/-hdd file use 'file' as hard disk 2/3 image\n", QEMU_ARCH_ALL
)
1256 DEF("hdd", HAS_ARG
, QEMU_OPTION_hdd
, "", QEMU_ARCH_ALL
)
1265 Use file as hard disk
0, 1, 2 or
3 image on the
default bus of the
1266 emulated
machine (this is
for example the IDE bus on most x86 machines
,
1267 but it can also be SCSI
, virtio or something
else on other target
1268 architectures
). See also the
:ref
:`disk images` chapter
in the System
1269 Emulation Users Guide
.
1272 DEF("cdrom", HAS_ARG
, QEMU_OPTION_cdrom
,
1273 "-cdrom file use 'file' as CD-ROM image\n",
1277 Use file as CD
-ROM image on the
default bus of the emulated machine
1278 (which is IDE1 master on x86
, so you cannot use ``
-hdc`` and ``
-cdrom``
1279 at the same time there
). On systems that support it
, you can use the
1280 host CD
-ROM by
using ``
/dev
/cdrom`` as filename
.
1283 DEF("blockdev", HAS_ARG
, QEMU_OPTION_blockdev
,
1284 "-blockdev [driver=]driver[,node-name=N][,discard=ignore|unmap]\n"
1285 " [,cache.direct=on|off][,cache.no-flush=on|off]\n"
1286 " [,read-only=on|off][,auto-read-only=on|off]\n"
1287 " [,force-share=on|off][,detect-zeroes=on|off|unmap]\n"
1288 " [,driver specific parameters...]\n"
1289 " configure a block backend\n", QEMU_ARCH_ALL
)
1291 ``
-blockdev option
[,option
[,option
[,...]]]``
1292 Define a
new block driver node
. Some of the options apply to all
1293 block drivers
, other options are only accepted
for a specific block
1294 driver
. See below
for a list of generic options and options
for the
1295 most common block drivers
.
1297 Options that expect a reference to another
node (e
.g
. ``file``
) can
1298 be given
in two ways
. Either you specify the node name of an already
1299 existing
node (file
=node
-name
), or you define a
new node
inline,
1300 adding options
for the referenced node after a dot
1301 (file
.filename
=path
,file
.aio
=native
).
1303 A block driver node created with ``
-blockdev`` can be used
for a
1304 guest device by specifying its node name
for the ``drive`` property
1305 in a ``
-device`` argument that defines a block device
.
1307 ``Valid options
for any block driver node
:``
1309 Specifies the block driver to use
for the given node
.
1312 This defines the name of the block driver node by which it
1313 will be referenced later
. The name must be unique
, i
.e
. it
1314 must not match the name of a different block driver node
, or
1315 (if you use ``
-drive`` as well
) the ID of a drive
.
1317 If no node name is specified
, it is automatically generated
.
1318 The generated node name is not intended to be predictable
1319 and changes between QEMU invocations
. For the top level
, an
1320 explicit node name must be specified
.
1323 Open the node read
-only
. Guest write attempts will fail
.
1325 Note that some block drivers support only read
-only access
,
1326 either generally or
in certain configurations
. In
this case,
1327 the
default value ``read
-only
=off`` does not work and the
1328 option must be specified explicitly
.
1331 If ``auto
-read
-only
=on`` is set
, QEMU may fall back to
1332 read
-only usage even when ``read
-only
=off`` is requested
, or
1333 even
switch between modes as needed
, e
.g
. depending on
1334 whether the image file is writable or whether a writing user
1335 is attached to the node
.
1338 Override the image locking system of QEMU by forcing the
1339 node to utilize weaker shared access
for permissions where
1340 it would normally request exclusive access
. When there is
1341 the potential
for multiple instances to have the same file
1342 open (whether
this invocation of QEMU is the first or the
1343 second instance
), both instances must permit shared access
1344 for the second instance to succeed at opening the file
.
1346 Enabling ``force
-share
=on`` requires ``read
-only
=on``
.
1349 The host page cache can be avoided with ``cache
.direct
=on``
.
1350 This will attempt to
do disk IO directly to the guest
's
1351 memory. QEMU may still perform an internal copy of the data.
1354 In case you don't care about data integrity over host
1355 failures
, you can use ``cache
.no
-flush
=on``
. This option
1356 tells QEMU that it
never needs to write any data to the disk
1357 but can instead keep things
in cache
. If anything goes
1358 wrong
, like your host losing power
, the disk storage getting
1359 disconnected accidentally
, etc
. your image will most
1360 probably be rendered unusable
.
1363 discard is one of
"ignore" (or
"off") or
"unmap" (or
"on")
1364 and controls whether ``discard``
(also known as ``trim`` or
1365 ``unmap``
) requests are ignored or passed to the filesystem
.
1366 Some machine types may not support discard requests
.
1368 ``detect
-zeroes
=detect
-zeroes``
1369 detect
-zeroes is
"off", "on" or
"unmap" and enables the
1370 automatic conversion of plain zero writes by the OS to
1371 driver specific optimized zero write commands
. You may even
1372 choose
"unmap" if discard is set to
"unmap" to allow a zero
1373 write to be converted to an ``unmap`` operation
.
1375 ``Driver
-specific options
for file``
1376 This is the protocol
-level block driver
for accessing regular
1380 The path to the image file
in the local filesystem
1383 Specifies the AIO
backend (threads
/native
/io_uring
,
1387 Specifies whether the image file is protected with Linux OFD
1388 / POSIX locks
. The
default is to use the Linux Open File
1389 Descriptor API
if available
, otherwise no lock is applied
.
1390 (auto
/on
/off
, default: auto
)
1396 -blockdev driver
=file
,node
-name
=disk
,filename
=disk
.img
1398 ``Driver
-specific options
for raw``
1399 This is the image format block driver
for raw images
. It is
1400 usually stacked on top of a protocol level block driver such as
1404 Reference to or definition of the data source block driver
1405 node (e
.g
. a ``file`` driver node
)
1411 -blockdev driver
=file
,node
-name
=disk_file
,filename
=disk
.img
1412 -blockdev driver
=raw
,node
-name
=disk
,file
=disk_file
1418 -blockdev driver
=raw
,node
-name
=disk
,file
.driver
=file
,file
.filename
=disk
.img
1420 ``Driver
-specific options
for qcow2``
1421 This is the image format block driver
for qcow2 images
. It is
1422 usually stacked on top of a protocol level block driver such as
1426 Reference to or definition of the data source block driver
1427 node (e
.g
. a ``file`` driver node
)
1430 Reference to or definition of the backing file block device
1431 (default is taken from the image file
). It is allowed to
1432 pass ``
null``
here in order to disable the
default backing
1436 Whether to enable the lazy refcounts
feature (on
/off
;
1437 default is taken from the image file
)
1440 The maximum total size of the L2 table and refcount block
1441 caches
in bytes (default: the sum of l2
-cache
-size and
1442 refcount
-cache
-size
)
1445 The maximum size of the L2 table cache
in bytes (default: if
1446 cache
-size is not specified
- 32M on Linux platforms
, and
8M
1447 on non
-Linux platforms
; otherwise
, as large as possible
1448 within the cache
-size
, while permitting the requested or the
1449 minimal refcount cache size
)
1451 ``refcount
-cache
-size``
1452 The maximum size of the refcount block cache
in bytes
1453 (default: 4 times the cluster size
; or
if cache
-size is
1454 specified
, the part of it which is not used
for the L2
1457 ``cache
-clean
-interval``
1458 Clean unused entries
in the L2 and refcount caches
. The
1459 interval is
in seconds
. The
default value is
600 on
1460 supporting platforms
, and
0 on other platforms
. Setting it
1461 to
0 disables
this feature
.
1463 ``pass
-discard
-request``
1464 Whether discard requests to the qcow2 device should be
1465 forwarded to the data
source (on
/off
; default: on
if
1466 discard
=unmap is specified
, off otherwise
)
1468 ``pass
-discard
-snapshot``
1469 Whether discard requests
for the data source should be
1470 issued when a snapshot
operation (e
.g
. deleting a snapshot
)
1471 frees clusters
in the qcow2
file (on
/off
; default: on
)
1473 ``pass
-discard
-other``
1474 Whether discard requests
for the data source should be
1475 issued on other occasions where a cluster gets freed
1476 (on
/off
; default: off
)
1478 ``discard
-no
-unref``
1479 When enabled
, data clusters will remain preallocated when they are
1480 no longer used
, e
.g
. because they are discarded or converted to
1481 zero clusters
. As usual
, whether the old data is discarded or kept
1482 on the protocol
level (i
.e
. in the image file
) depends on the
1483 setting of the pass
-discard
-request option
. Keeping the clusters
1484 preallocated prevents qcow2 fragmentation that would otherwise be
1485 caused by freeing and re
-allocating them later
. Besides potential
1486 performance degradation
, such fragmentation can lead to increased
1487 allocation of clusters past the end of the image file
,
1488 resulting
in image files whose file length can grow much larger
1489 than their guest disk size would suggest
.
1490 If image file length is of
concern (e
.g
. when storing qcow2
1491 images directly on block devices
), you should consider enabling
1495 Which overlap checks to perform
for writes to the image
1496 (none
/constant
/cached
/all
; default: cached
). For details or
1497 finer granularity control refer to the QAPI documentation of
1504 -blockdev driver
=file
,node
-name
=my_file
,filename
=/tmp
/disk
.qcow2
1505 -blockdev driver
=qcow2
,node
-name
=hda
,file
=my_file
,overlap
-check
=none
,cache
-size
=16777216
1511 -blockdev driver
=qcow2
,node
-name
=disk
,file
.driver
=http
,file
.filename
=http
://example.com/image.qcow2
1513 ``Driver
-specific options
for other drivers``
1514 Please refer to the QAPI documentation of the ``blockdev
-add``
1518 DEF("drive", HAS_ARG
, QEMU_OPTION_drive
,
1519 "-drive [file=file][,if=type][,bus=n][,unit=m][,media=d][,index=i]\n"
1520 " [,cache=writethrough|writeback|none|directsync|unsafe][,format=f]\n"
1521 " [,snapshot=on|off][,rerror=ignore|stop|report]\n"
1522 " [,werror=ignore|stop|report|enospc][,id=name]\n"
1523 " [,aio=threads|native|io_uring]\n"
1524 " [,readonly=on|off][,copy-on-read=on|off]\n"
1525 " [,discard=ignore|unmap][,detect-zeroes=on|off|unmap]\n"
1526 " [[,bps=b]|[[,bps_rd=r][,bps_wr=w]]]\n"
1527 " [[,iops=i]|[[,iops_rd=r][,iops_wr=w]]]\n"
1528 " [[,bps_max=bm]|[[,bps_rd_max=rm][,bps_wr_max=wm]]]\n"
1529 " [[,iops_max=im]|[[,iops_rd_max=irm][,iops_wr_max=iwm]]]\n"
1530 " [[,iops_size=is]]\n"
1532 " use 'file' as a drive image\n", QEMU_ARCH_ALL
)
1534 ``
-drive option
[,option
[,option
[,...]]]``
1535 Define a
new drive
. This includes creating a block driver
node (the
1536 backend
) as well as a guest device
, and is mostly a shortcut
for
1537 defining the corresponding ``
-blockdev`` and ``
-device`` options
.
1539 ``
-drive`` accepts all options that are accepted by ``
-blockdev``
.
1540 In addition
, it knows the following options
:
1543 This option defines which disk
image (see the
:ref
:`disk images`
1544 chapter
in the System Emulation Users Guide
) to use with
this drive
.
1545 If the filename contains comma
, you must double
it (for instance
,
1546 "file=my,,file" to use file
"my,file").
1548 Special files such as iSCSI devices can be specified
using
1549 protocol specific URLs
. See the section
for "Device URL Syntax"
1550 for more information
.
1553 This option defines on which type on
interface the drive is
1554 connected
. Available types are
: ide
, scsi
, sd
, mtd
, floppy
,
1555 pflash
, virtio
, none
.
1557 ``bus
=bus
,unit
=unit``
1558 These options define where is connected the drive by defining
1559 the bus number and the unit id
.
1562 This option defines where the drive is connected by
using an
1563 index
in the list of available connectors of a given
interface
1567 This option defines the type of the media
: disk or cdrom
.
1569 ``snapshot
=snapshot``
1570 snapshot is
"on" or
"off" and controls snapshot mode
for the
1571 given
drive (see ``
-snapshot``
).
1574 cache is
"none", "writeback", "unsafe", "directsync" or
1575 "writethrough" and controls how the host cache is used to access
1576 block data
. This is a shortcut that sets the ``cache
.direct``
1577 and ``cache
.no
-flush``
options (as
in ``
-blockdev``
), and
1578 additionally ``cache
.writeback``
, which provides a
default for
1579 the ``write
-cache`` option of block guest
devices (as
in
1580 ``
-device``
). The modes correspond to the following settings
:
1582 ============= =============== ============ ==============
1583 \ cache
.writeback cache
.direct cache
.no
-flush
1584 ============= =============== ============ ==============
1585 writeback on off off
1587 writethrough off off off
1588 directsync off on off
1590 ============= =============== ============ ==============
1592 The
default mode is ``cache
=writeback``
.
1595 aio is
"threads", "native", or
"io_uring" and selects between pthread
1596 based disk I
/O
, native Linux AIO
, or Linux io_uring API
.
1599 Specify which disk format will be used rather than detecting the
1600 format
. Can be used to specify format
=raw to avoid interpreting
1601 an untrusted format header
.
1603 ``werror
=action
,rerror
=action``
1604 Specify which action to take on write and read errors
. Valid
1605 actions are
: "ignore" (ignore the error and
try to
continue),
1606 "stop" (pause QEMU
), "report" (report the error to the guest
),
1607 "enospc" (pause QEMU only
if the host disk is full
; report the
1608 error to the guest otherwise
). The
default setting is
1609 ``werror
=enospc`` and ``rerror
=report``
.
1611 ``copy
-on
-read
=copy
-on
-read``
1612 copy
-on
-read is
"on" or
"off" and enables whether to copy read
1613 backing file sectors into the image file
.
1615 ``bps
=b
,bps_rd
=r
,bps_wr
=w``
1616 Specify bandwidth throttling limits
in bytes per second
, either
1617 for all request types or
for reads or writes only
. Small values
1618 can lead to timeouts or hangs inside the guest
. A safe minimum
1619 for disks is
2 MB
/s
.
1621 ``bps_max
=bm
,bps_rd_max
=rm
,bps_wr_max
=wm``
1622 Specify bursts
in bytes per second
, either
for all request types
1623 or
for reads or writes only
. Bursts allow the guest I
/O to spike
1624 above the limit temporarily
.
1626 ``iops
=i
,iops_rd
=r
,iops_wr
=w``
1627 Specify request rate limits
in requests per second
, either
for
1628 all request types or
for reads or writes only
.
1630 ``iops_max
=bm
,iops_rd_max
=rm
,iops_wr_max
=wm``
1631 Specify bursts
in requests per second
, either
for all request
1632 types or
for reads or writes only
. Bursts allow the guest I
/O to
1633 spike above the limit temporarily
.
1636 Let every is bytes of a request count as a
new request
for iops
1637 throttling purposes
. Use
this option to prevent guests from
1638 circumventing iops limits by sending fewer but larger requests
.
1641 Join a throttling quota group with given name g
. All drives that
1642 are members of the same group are accounted
for together
. Use
1643 this option to prevent guests from circumventing throttling
1644 limits by
using many small disks instead of a single larger
1647 By
default, the ``cache
.writeback
=on`` mode is used
. It will report
1648 data writes as completed as soon as the data is present
in the host
1649 page cache
. This is safe as long as your guest OS makes sure to
1650 correctly flush disk caches where needed
. If your guest OS does not
1651 handle volatile disk write caches correctly and your host crashes or
1652 loses power
, then the guest may experience data corruption
.
1654 For such guests
, you should consider
using ``cache
.writeback
=off``
.
1655 This means that the host page cache will be used to read and write
1656 data
, but write notification will be sent to the guest only after
1657 QEMU has made sure to flush each write to the disk
. Be aware that
1658 this has a major impact on performance
.
1660 When
using the ``
-snapshot`` option
, unsafe caching is always used
.
1662 Copy
-on
-read avoids accessing the same backing file sectors
1663 repeatedly and is useful when the backing file is over a slow
1664 network
. By
default copy
-on
-read is off
.
1666 Instead of ``
-cdrom`` you can use
:
1670 |qemu_system|
-drive file
=file
,index
=2,media
=cdrom
1672 Instead of ``
-hda``
, ``
-hdb``
, ``
-hdc``
, ``
-hdd``
, you can use
:
1676 |qemu_system|
-drive file
=file
,index
=0,media
=disk
1677 |qemu_system|
-drive file
=file
,index
=1,media
=disk
1678 |qemu_system|
-drive file
=file
,index
=2,media
=disk
1679 |qemu_system|
-drive file
=file
,index
=3,media
=disk
1681 You can open an image
using pre
-opened file descriptors from an fd
1687 -add
-fd fd
=3,set
=2,opaque
="rdwr:/path/to/file" \\
1688 -add
-fd fd
=4,set
=2,opaque
="rdonly:/path/to/file" \\
1689 -drive file
=/dev
/fdset
/2,index
=0,media
=disk
1691 You can connect a CDROM to the slave of ide0
:
1695 |qemu_system_x86|
-drive file
=file
,if=ide
,index
=1,media
=cdrom
1697 If you don
't specify the "file=" argument, you define an empty
1702 |qemu_system_x86| -drive if=ide,index=1,media=cdrom
1704 Instead of ``-fda``, ``-fdb``, you can use:
1708 |qemu_system_x86| -drive file=file,index=0,if=floppy
1709 |qemu_system_x86| -drive file=file,index=1,if=floppy
1711 By default, interface is "ide" and index is automatically
1716 |qemu_system_x86| -drive file=a -drive file=b
1718 is interpreted like:
1722 |qemu_system_x86| -hda a -hdb b
1725 DEF("mtdblock", HAS_ARG, QEMU_OPTION_mtdblock,
1726 "-mtdblock file use 'file
' as on-board Flash memory image\n",
1730 Use file as on-board Flash memory image.
1733 DEF("sd", HAS_ARG, QEMU_OPTION_sd,
1734 "-sd file use 'file
' as SecureDigital card image\n", QEMU_ARCH_ALL)
1737 Use file as SecureDigital card image.
1740 DEF("snapshot", 0, QEMU_OPTION_snapshot,
1741 "-snapshot write to temporary files instead of disk image files\n",
1745 Write to temporary files instead of disk image files. In this case,
1746 the raw disk image you use is not written back. You can however
1747 force the write back by pressing C-a s (see the :ref:`disk images`
1748 chapter in the System Emulation Users Guide).
1751 snapshot is incompatible with ``-blockdev`` (instead use qemu-img
1752 to manually create snapshot images to attach to your blockdev).
1753 If you have mixed ``-blockdev`` and ``-drive`` declarations you
1754 can use the 'snapshot
' property on your drive declarations
1755 instead of this global option.
1759 DEF("fsdev", HAS_ARG, QEMU_OPTION_fsdev,
1760 "-fsdev local,id=id,path=path,security_model=mapped-xattr|mapped-file|passthrough|none\n"
1761 " [,writeout=immediate][,readonly=on][,fmode=fmode][,dmode=dmode]\n"
1762 " [[,throttling.bps-total=b]|[[,throttling.bps-read=r][,throttling.bps-write=w]]]\n"
1763 " [[,throttling.iops-total=i]|[[,throttling.iops-read=r][,throttling.iops-write=w]]]\n"
1764 " [[,throttling.bps-total-max=bm]|[[,throttling.bps-read-max=rm][,throttling.bps-write-max=wm]]]\n"
1765 " [[,throttling.iops-total-max=im]|[[,throttling.iops-read-max=irm][,throttling.iops-write-max=iwm]]]\n"
1766 " [[,throttling.iops-size=is]]\n"
1767 "-fsdev proxy,id=id,socket=socket[,writeout=immediate][,readonly=on]\n"
1768 "-fsdev proxy,id=id,sock_fd=sock_fd[,writeout=immediate][,readonly=on]\n"
1769 "-fsdev synth,id=id\n",
1773 ``-fsdev local,id=id,path=path,security_model=security_model [,writeout=writeout][,readonly=on][,fmode=fmode][,dmode=dmode] [,throttling.option=value[,throttling.option=value[,...]]]``
1775 ``-fsdev proxy,id=id,socket=socket[,writeout=writeout][,readonly=on]``
1777 ``-fsdev proxy,id=id,sock_fd=sock_fd[,writeout=writeout][,readonly=on]``
1779 ``-fsdev synth,id=id[,readonly=on]``
1780 Define a new file system device. Valid options are:
1783 Accesses to the filesystem are done by QEMU.
1786 Accesses to the filesystem are done by virtfs-proxy-helper(1). This
1787 option is deprecated (since QEMU 8.1) and will be removed in a future
1788 version of QEMU. Use ``local`` instead.
1791 Synthetic filesystem, only used by QTests.
1794 Specifies identifier for this device.
1797 Specifies the export path for the file system device. Files
1798 under this path will be available to the 9p client on the guest.
1800 ``security_model=security_model``
1801 Specifies the security model to be used for this export path.
1802 Supported security models are "passthrough", "mapped-xattr",
1803 "mapped-file" and "none". In "passthrough" security model, files
1804 are stored using the same credentials as they are created on the
1805 guest. This requires QEMU to run as root. In "mapped-xattr"
1806 security model, some of the file attributes like uid, gid, mode
1807 bits and link target are stored as file attributes. For
1808 "mapped-file" these attributes are stored in the hidden
1809 .virtfs\_metadata directory. Directories exported by this
1810 security model cannot interact with other unix tools. "none"
1811 security model is same as passthrough except the sever won't
1812 report failures
if it fails to set file attributes like
1813 ownership
. Security model is mandatory only
for local fsdriver
.
1814 Other
fsdrivers (like proxy
) don
't take security model as a
1817 ``writeout=writeout``
1818 This is an optional argument. The only supported value is
1819 "immediate". This means that host page cache will be used to
1820 read and write data but write notification will be sent to the
1821 guest only when the data has been reported as written by the
1825 Enables exporting 9p share as a readonly mount for guests. By
1826 default read-write access is given.
1829 Enables proxy filesystem driver to use passed socket file for
1830 communicating with virtfs-proxy-helper(1).
1833 Enables proxy filesystem driver to use passed socket descriptor
1834 for communicating with virtfs-proxy-helper(1). Usually a helper
1835 like libvirt will create socketpair and pass one of the fds as
1839 Specifies the default mode for newly created files on the host.
1840 Works only with security models "mapped-xattr" and
1844 Specifies the default mode for newly created directories on the
1845 host. Works only with security models "mapped-xattr" and
1848 ``throttling.bps-total=b,throttling.bps-read=r,throttling.bps-write=w``
1849 Specify bandwidth throttling limits in bytes per second, either
1850 for all request types or for reads or writes only.
1852 ``throttling.bps-total-max=bm,bps-read-max=rm,bps-write-max=wm``
1853 Specify bursts in bytes per second, either for all request types
1854 or for reads or writes only. Bursts allow the guest I/O to spike
1855 above the limit temporarily.
1857 ``throttling.iops-total=i,throttling.iops-read=r, throttling.iops-write=w``
1858 Specify request rate limits in requests per second, either for
1859 all request types or for reads or writes only.
1861 ``throttling.iops-total-max=im,throttling.iops-read-max=irm, throttling.iops-write-max=iwm``
1862 Specify bursts in requests per second, either for all request
1863 types or for reads or writes only. Bursts allow the guest I/O to
1864 spike above the limit temporarily.
1866 ``throttling.iops-size=is``
1867 Let every is bytes of a request count as a new request for iops
1868 throttling purposes.
1870 -fsdev option is used along with -device driver "virtio-9p-...".
1872 ``-device virtio-9p-type,fsdev=id,mount_tag=mount_tag``
1873 Options for virtio-9p-... driver are:
1876 Specifies the variant to be used. Supported values are "pci",
1877 "ccw" or "device", depending on the machine type.
1880 Specifies the id value specified along with -fsdev option.
1882 ``mount_tag=mount_tag``
1883 Specifies the tag name to be used by the guest to mount this
1887 DEF("virtfs", HAS_ARG, QEMU_OPTION_virtfs,
1888 "-virtfs local,path=path,mount_tag=tag,security_model=mapped-xattr|mapped-file|passthrough|none\n"
1889 " [,id=id][,writeout=immediate][,readonly=on][,fmode=fmode][,dmode=dmode][,multidevs=remap|forbid|warn]\n"
1890 "-virtfs proxy,mount_tag=tag,socket=socket[,id=id][,writeout=immediate][,readonly=on]\n"
1891 "-virtfs proxy,mount_tag=tag,sock_fd=sock_fd[,id=id][,writeout=immediate][,readonly=on]\n"
1892 "-virtfs synth,mount_tag=tag[,id=id][,readonly=on]\n",
1896 ``-virtfs local,path=path,mount_tag=mount_tag ,security_model=security_model[,writeout=writeout][,readonly=on] [,fmode=fmode][,dmode=dmode][,multidevs=multidevs]``
1898 ``-virtfs proxy,socket=socket,mount_tag=mount_tag [,writeout=writeout][,readonly=on]``
1900 ``-virtfs proxy,sock_fd=sock_fd,mount_tag=mount_tag [,writeout=writeout][,readonly=on]``
1902 ``-virtfs synth,mount_tag=mount_tag``
1903 Define a new virtual filesystem device and expose it to the guest using
1904 a virtio-9p-device (a.k.a. 9pfs), which essentially means that a certain
1905 directory on host is made directly accessible by guest as a pass-through
1906 file system by using the 9P network protocol for communication between
1907 host and guests, if desired even accessible, shared by several guests
1910 Note that ``-virtfs`` is actually just a convenience shortcut for its
1911 generalized form ``-fsdev -device virtio-9p-pci``.
1913 The general form of pass-through file system options are:
1916 Accesses to the filesystem are done by QEMU.
1919 Accesses to the filesystem are done by virtfs-proxy-helper(1).
1920 This option is deprecated (since QEMU 8.1) and will be removed in a
1921 future version of QEMU. Use ``local`` instead.
1924 Synthetic filesystem, only used by QTests.
1927 Specifies identifier for the filesystem device
1930 Specifies the export path for the file system device. Files
1931 under this path will be available to the 9p client on the guest.
1933 ``security_model=security_model``
1934 Specifies the security model to be used for this export path.
1935 Supported security models are "passthrough", "mapped-xattr",
1936 "mapped-file" and "none". In "passthrough" security model, files
1937 are stored using the same credentials as they are created on the
1938 guest. This requires QEMU to run as root. In "mapped-xattr"
1939 security model, some of the file attributes like uid, gid, mode
1940 bits and link target are stored as file attributes. For
1941 "mapped-file" these attributes are stored in the hidden
1942 .virtfs\_metadata directory. Directories exported by this
1943 security model cannot interact with other unix tools. "none"
1944 security model is same as passthrough except the sever won't
1945 report failures
if it fails to set file attributes like
1946 ownership
. Security model is mandatory only
for local fsdriver
.
1947 Other
fsdrivers (like proxy
) don
't take security model as a
1950 ``writeout=writeout``
1951 This is an optional argument. The only supported value is
1952 "immediate". This means that host page cache will be used to
1953 read and write data but write notification will be sent to the
1954 guest only when the data has been reported as written by the
1958 Enables exporting 9p share as a readonly mount for guests. By
1959 default read-write access is given.
1962 Enables proxy filesystem driver to use passed socket file for
1963 communicating with virtfs-proxy-helper(1). Usually a helper like
1964 libvirt will create socketpair and pass one of the fds as
1968 Enables proxy filesystem driver to use passed 'sock\_fd
' as the
1969 socket descriptor for interfacing with virtfs-proxy-helper(1).
1972 Specifies the default mode for newly created files on the host.
1973 Works only with security models "mapped-xattr" and
1977 Specifies the default mode for newly created directories on the
1978 host. Works only with security models "mapped-xattr" and
1981 ``mount_tag=mount_tag``
1982 Specifies the tag name to be used by the guest to mount this
1985 ``multidevs=multidevs``
1986 Specifies how to deal with multiple devices being shared with a
1987 9p export. Supported behaviours are either "remap", "forbid" or
1988 "warn". The latter is the default behaviour on which virtfs 9p
1989 expects only one device to be shared with the same export, and
1990 if more than one device is shared and accessed via the same 9p
1991 export then only a warning message is logged (once) by qemu on
1992 host side. In order to avoid file ID collisions on guest you
1993 should either create a separate virtfs export for each device to
1994 be shared with guests (recommended way) or you might use "remap"
1995 instead which allows you to share multiple devices with only one
1996 export instead, which is achieved by remapping the original
1997 inode numbers from host to guest in a way that would prevent
1998 such collisions. Remapping inodes in such use cases is required
1999 because the original device IDs from host are never passed and
2000 exposed on guest. Instead all files of an export shared with
2001 virtfs always share the same device id on guest. So two files
2002 with identical inode numbers but from actually different devices
2003 on host would otherwise cause a file ID collision and hence
2004 potential misbehaviours on guest. "forbid" on the other hand
2005 assumes like "warn" that only one device is shared by the same
2006 export, however it will not only log a warning message but also
2007 deny access to additional devices on guest. Note though that
2008 "forbid" does currently not block all possible file access
2009 operations (e.g. readdir() would still return entries from other
2013 DEF("iscsi", HAS_ARG, QEMU_OPTION_iscsi,
2014 "-iscsi [user=user][,password=password][,password-secret=secret-id]\n"
2015 " [,header-digest=CRC32C|CR32C-NONE|NONE-CRC32C|NONE]\n"
2016 " [,initiator-name=initiator-iqn][,id=target-iqn]\n"
2017 " [,timeout=timeout]\n"
2018 " iSCSI session parameters\n", QEMU_ARCH_ALL)
2022 Configure iSCSI session parameters.
2027 DEFHEADING(USB convenience options:)
2029 DEF("usb", 0, QEMU_OPTION_usb,
2030 "-usb enable on-board USB host controller (if not enabled by default)\n",
2034 Enable USB emulation on machine types with an on-board USB host
2035 controller (if not enabled by default). Note that on-board USB host
2036 controllers may not support USB 3.0. In this case
2037 ``-device qemu-xhci`` can be used instead on machines with PCI.
2040 DEF("usbdevice", HAS_ARG, QEMU_OPTION_usbdevice,
2041 "-usbdevice name add the host or guest USB device 'name
'\n",
2044 ``-usbdevice devname``
2045 Add the USB device devname, and enable an on-board USB controller
2046 if possible and necessary (just like it can be done via
2047 ``-machine usb=on``). Note that this option is mainly intended for
2048 the user's convenience only
. More fine
-grained control can be
2049 achieved by selecting a USB host
controller (if necessary
) and the
2050 desired USB device via the ``
-device`` option instead
. For example
,
2051 instead of
using ``
-usbdevice mouse`` it is possible to use
2052 ``
-device qemu
-xhci
-device usb
-mouse`` to connect the USB mouse
2053 to a USB
3.0 controller
instead (at least on machines that support
2054 PCI and
do not have an USB controller enabled by
default yet
).
2055 For more details
, see the chapter about
2056 :ref
:`Connecting USB devices`
in the System Emulation Users Guide
.
2057 Possible devices
for devname are
:
2060 Braille device
. This will use BrlAPI to display the braille
2061 output on a real or fake
device (i
.e
. it also creates a
2062 corresponding ``braille`` chardev automatically beside the
2063 ``usb
-braille`` USB device
).
2066 Standard USB keyboard
. Will
override the PS
/2 keyboard (if present
).
2069 Virtual Mouse
. This will
override the PS
/2 mouse emulation when
2073 Pointer device that uses absolute
coordinates (like a
2074 touchscreen
). This means QEMU is able to report the mouse
2075 position without having to grab the mouse
. Also overrides the
2076 PS
/2 mouse emulation when activated
.
2079 Wacom PenPartner USB tablet
.
2086 DEFHEADING(Display options
:)
2088 DEF("display", HAS_ARG
, QEMU_OPTION_display
,
2089 #
if defined(CONFIG_SPICE
)
2090 "-display spice-app[,gl=on|off]\n"
2092 #
if defined(CONFIG_SDL
)
2093 "-display sdl[,gl=on|core|es|off][,grab-mod=<mod>][,show-cursor=on|off]\n"
2094 " [,window-close=on|off]\n"
2096 #
if defined(CONFIG_GTK
)
2097 "-display gtk[,full-screen=on|off][,gl=on|off][,grab-on-hover=on|off]\n"
2098 " [,show-tabs=on|off][,show-cursor=on|off][,window-close=on|off]\n"
2099 " [,show-menubar=on|off][,zoom-to-fit=on|off]\n"
2101 #
if defined(CONFIG_VNC
)
2102 "-display vnc=<display>[,<optargs>]\n"
2104 #
if defined(CONFIG_CURSES
)
2105 "-display curses[,charset=<encoding>]\n"
2107 #
if defined(CONFIG_COCOA
)
2108 "-display cocoa[,full-grab=on|off][,swap-opt-cmd=on|off]\n"
2109 " [,show-cursor=on|off][,left-command-key=on|off]\n"
2110 " [,full-screen=on|off][,zoom-to-fit=on|off]\n"
2112 #
if defined(CONFIG_OPENGL
)
2113 "-display egl-headless[,rendernode=<file>]\n"
2115 #
if defined(CONFIG_DBUS_DISPLAY
)
2116 "-display dbus[,addr=<dbusaddr>]\n"
2117 " [,gl=on|core|es|off][,rendernode=<file>]\n"
2120 " select display backend type\n"
2121 " The default display is equivalent to\n "
2122 #
if defined(CONFIG_GTK
)
2123 "\"-display gtk\"\n"
2124 #elif
defined(CONFIG_SDL
)
2125 "\"-display sdl\"\n"
2126 #elif
defined(CONFIG_COCOA
)
2127 "\"-display cocoa\"\n"
2128 #elif
defined(CONFIG_VNC
)
2129 "\"-vnc localhost:0,to=99,id=default\"\n"
2131 "\"-display none\"\n"
2136 Select type of display to use
. Use ``
-display help`` to list the available
2137 display types
. Valid values
for type are
2139 ``spice
-app
[,gl
=on|off
]``
2140 Start QEMU as a Spice server and launch the
default Spice client
2141 application
. The Spice server will redirect the serial consoles
2142 and QEMU monitors
. (Since
4.0)
2145 Export the display over D
-Bus interfaces
. (Since
7.0)
2147 The connection is registered with the
"org.qemu" name (and queued when
2150 ``addr
=<dbusaddr
>``
: D
-Bus bus address to connect to
.
2152 ``p2p
=yes|no``
: Use peer
-to
-peer connection
, accepted via QMP ``add_client``
.
2154 ``gl
=on|off|core|es``
: Use OpenGL
for rendering (the D
-Bus
interface
2155 will share framebuffers with DMABUF file descriptors
).
2158 Display video output via
SDL (usually
in a separate graphics
2159 window
; see the SDL documentation
for other possibilities
).
2160 Valid parameters are
:
2162 ``grab
-mod
=<mods
>``
: Used to select the modifier keys
for toggling
2163 the mouse grabbing
in conjunction with the
"g" key
. ``
<mods
>`` can be
2164 either ``lshift
-lctrl
-lalt`` or ``rctrl``
.
2166 ``gl
=on|off|core|es``
: Use OpenGL
for displaying
2168 ``show
-cursor
=on|off``
: Force showing the mouse cursor
2170 ``window
-close
=on|off``
: Allow to quit qemu with window close button
2173 Display video output
in a GTK window
. This
interface provides
2174 drop
-down menus and other UI elements to configure and control
2175 the VM during runtime
. Valid parameters are
:
2177 ``full
-screen
=on|off``
: Start
in fullscreen mode
2179 ``gl
=on|off``
: Use OpenGL
for displaying
2181 ``grab
-on
-hover
=on|off``
: Grab keyboard input on mouse hover
2183 ``show
-tabs
=on|off``
: Display the tab bar
for switching between the
2184 various graphical
interfaces (e
.g
. VGA and
2185 virtual console character devices
) by
default.
2187 ``show
-cursor
=on|off``
: Force showing the mouse cursor
2189 ``window
-close
=on|off``
: Allow to quit qemu with window close button
2191 ``show
-menubar
=on|off``
: Display the main window menubar
, defaults to
"on"
2193 ``zoom
-to
-fit
=on|off``
: Expand video output to the window size
,
2196 ``curses
[,charset
=<encoding
>]``
2197 Display video output via curses
. For graphics device models
2198 which support a text mode
, QEMU can display
this output
using a
2199 curses
/ncurses
interface. Nothing is displayed when the graphics
2200 device is
in graphical mode or
if the graphics device does not
2201 support a text mode
. Generally only the VGA device models
2202 support text mode
. The font charset used by the guest can be
2203 specified with the ``charset`` option
, for example
2204 ``charset
=CP850``
for IBM CP850 encoding
. The
default is
2208 Display video output
in a Cocoa window
. Mac only
. This
interface
2209 provides drop
-down menus and other UI elements to configure and
2210 control the VM during runtime
. Valid parameters are
:
2212 ``full
-grab
=on|off``
: Capture all key presses
, including system combos
.
2213 This requires accessibility permissions
, since it
2214 performs a global grab on key events
.
2216 https
://support.apple.com/en-in/guide/mac-help/mh32356/mac
2218 ``swap
-opt
-cmd
=on|off``
: Swap the Option and Command keys so that their
2219 key codes match their position on non
-Mac
2220 keyboards and you can use Meta
/Super and Alt
2221 where you expect them
. (default: off
)
2223 ``show
-cursor
=on|off``
: Force showing the mouse cursor
2225 ``left
-command
-key
=on|off``
: Disable forwarding left command key to host
2227 ``full
-screen
=on|off``
: Start
in fullscreen mode
2229 ``zoom
-to
-fit
=on|off``
: Expand video output to the window size
,
2232 ``egl
-headless
[,rendernode
=<file
>]``
2233 Offload all OpenGL operations to a local DRI device
. For any
2234 graphical display
, this display needs to be paired with either
2235 VNC or SPICE displays
.
2238 Start a VNC server on display
<display
>
2241 Do not display video output
. The guest will still see an
2242 emulated graphics card
, but its output will not be displayed to
2243 the QEMU user
. This option differs from the
-nographic option
in
2244 that it only affects what is done with video output
; -nographic
2245 also changes the destination of the serial and parallel port
2249 DEF("nographic", 0, QEMU_OPTION_nographic
,
2250 "-nographic disable graphical output and redirect serial I/Os to console\n",
2254 Normally
, if QEMU is compiled with graphical window support
, it
2255 displays output such as guest graphics
, guest console
, and the QEMU
2256 monitor
in a window
. With
this option
, you can totally disable
2257 graphical output so that QEMU is a simple command line application
.
2258 The emulated serial port is redirected on the console and muxed with
2259 the
monitor (unless redirected elsewhere explicitly
). Therefore
, you
2260 can still use QEMU to debug a Linux kernel with a serial console
.
2261 Use C
-a h
for help on switching between the console and monitor
.
2265 DEF("spice", HAS_ARG
, QEMU_OPTION_spice
,
2266 "-spice [port=port][,tls-port=secured-port][,x509-dir=<dir>]\n"
2267 " [,x509-key-file=<file>][,x509-key-password=<file>]\n"
2268 " [,x509-cert-file=<file>][,x509-cacert-file=<file>]\n"
2269 " [,x509-dh-key-file=<file>][,addr=addr]\n"
2270 " [,ipv4=on|off][,ipv6=on|off][,unix=on|off]\n"
2271 " [,tls-ciphers=<list>]\n"
2272 " [,tls-channel=[main|display|cursor|inputs|record|playback]]\n"
2273 " [,plaintext-channel=[main|display|cursor|inputs|record|playback]]\n"
2274 " [,sasl=on|off][,disable-ticketing=on|off]\n"
2275 " [,password-secret=<secret-id>]\n"
2276 " [,image-compression=[auto_glz|auto_lz|quic|glz|lz|off]]\n"
2277 " [,jpeg-wan-compression=[auto|never|always]]\n"
2278 " [,zlib-glz-wan-compression=[auto|never|always]]\n"
2279 " [,streaming-video=[off|all|filter]][,disable-copy-paste=on|off]\n"
2280 " [,disable-agent-file-xfer=on|off][,agent-mouse=[on|off]]\n"
2281 " [,playback-compression=[on|off]][,seamless-migration=[on|off]]\n"
2282 " [,gl=[on|off]][,rendernode=<file>]\n"
2284 " at least one of {port, tls-port} is mandatory\n",
2288 ``
-spice option
[,option
[,...]]``
2289 Enable the spice remote desktop protocol
. Valid options are
2292 Set the TCP port spice is listening on
for plaintext channels
.
2295 Set the IP address spice is listening on
. Default is any
2298 ``ipv4
=on|off``
; \ ``ipv6
=on|off``
; \ ``unix
=on|off``
2299 Force
using the specified IP version
.
2301 ``password
-secret
=<secret
-id
>``
2302 Set the ID of the ``secret`` object containing the password
2303 you need to authenticate
.
2306 Require that the client use SASL to authenticate with the spice
.
2307 The exact choice of authentication method used is controlled
2308 from the system
/ user
's SASL configuration file for the 'qemu
'
2309 service. This is typically found in /etc/sasl2/qemu.conf. If
2310 running QEMU as an unprivileged user, an environment variable
2311 SASL\_CONF\_PATH can be used to make it search alternate
2312 locations for the service config. While some SASL auth methods
2313 can also provide data encryption (eg GSSAPI), it is recommended
2314 that SASL always be combined with the 'tls
' and 'x509
' settings
2315 to enable use of SSL and server certificates. This ensures a
2316 data encryption preventing compromise of authentication
2319 ``disable-ticketing=on|off``
2320 Allow client connects without authentication.
2322 ``disable-copy-paste=on|off``
2323 Disable copy paste between the client and the guest.
2325 ``disable-agent-file-xfer=on|off``
2326 Disable spice-vdagent based file-xfer between the client and the
2330 Set the TCP port spice is listening on for encrypted channels.
2333 Set the x509 file directory. Expects same filenames as -vnc
2336 ``x509-key-file=<file>``; \ ``x509-key-password=<file>``; \ ``x509-cert-file=<file>``; \ ``x509-cacert-file=<file>``; \ ``x509-dh-key-file=<file>``
2337 The x509 file names can also be configured individually.
2339 ``tls-ciphers=<list>``
2340 Specify which ciphers to use.
2342 ``tls-channel=[main|display|cursor|inputs|record|playback]``; \ ``plaintext-channel=[main|display|cursor|inputs|record|playback]``
2343 Force specific channel to be used with or without TLS
2344 encryption. The options can be specified multiple times to
2345 configure multiple channels. The special name "default" can be
2346 used to set the default mode. For channels which are not
2347 explicitly forced into one mode the spice client is allowed to
2348 pick tls/plaintext as he pleases.
2350 ``image-compression=[auto_glz|auto_lz|quic|glz|lz|off]``
2351 Configure image compression (lossless). Default is auto\_glz.
2353 ``jpeg-wan-compression=[auto|never|always]``; \ ``zlib-glz-wan-compression=[auto|never|always]``
2354 Configure wan image compression (lossy for slow links). Default
2357 ``streaming-video=[off|all|filter]``
2358 Configure video stream detection. Default is off.
2360 ``agent-mouse=[on|off]``
2361 Enable/disable passing mouse events via vdagent. Default is on.
2363 ``playback-compression=[on|off]``
2364 Enable/disable audio stream compression (using celt 0.5.1).
2367 ``seamless-migration=[on|off]``
2368 Enable/disable spice seamless migration. Default is off.
2371 Enable/disable OpenGL context. Default is off.
2373 ``rendernode=<file>``
2374 DRM render node for OpenGL rendering. If not specified, it will
2375 pick the first available. (Since 2.9)
2378 DEF("portrait", 0, QEMU_OPTION_portrait,
2379 "-portrait rotate graphical output 90 deg left (only PXA LCD)\n",
2383 Rotate graphical output 90 deg left (only PXA LCD).
2386 DEF("rotate", HAS_ARG, QEMU_OPTION_rotate,
2387 "-rotate <deg> rotate graphical output some deg left (only PXA LCD)\n",
2391 Rotate graphical output some deg left (only PXA LCD).
2394 DEF("vga", HAS_ARG, QEMU_OPTION_vga,
2395 "-vga [std|cirrus|vmware|qxl|xenfb|tcx|cg3|virtio|none]\n"
2396 " select video card type\n", QEMU_ARCH_ALL)
2399 Select type of VGA card to emulate. Valid values for type are
2402 Cirrus Logic GD5446 Video card. All Windows versions starting
2403 from Windows 95 should recognize and use this graphic card. For
2404 optimal performances, use 16 bit color depth in the guest and
2405 the host OS. (This card was the default before QEMU 2.2)
2408 Standard VGA card with Bochs VBE extensions. If your guest OS
2409 supports the VESA 2.0 VBE extensions (e.g. Windows XP) and if
2410 you want to use high resolution modes (>= 1280x1024x16) then you
2411 should use this option. (This card is the default since QEMU
2415 VMWare SVGA-II compatible adapter. Use it if you have
2416 sufficiently recent XFree86/XOrg server or Windows guest with a
2417 driver for this card.
2420 QXL paravirtual graphic card. It is VGA compatible (including
2421 VESA 2.0 VBE support). Works best with qxl guest drivers
2422 installed though. Recommended choice when using the spice
2426 (sun4m only) Sun TCX framebuffer. This is the default
2427 framebuffer for sun4m machines and offers both 8-bit and 24-bit
2428 colour depths at a fixed resolution of 1024x768.
2431 (sun4m only) Sun cgthree framebuffer. This is a simple 8-bit
2432 framebuffer for sun4m machines available in both 1024x768
2433 (OpenBIOS) and 1152x900 (OBP) resolutions aimed at people
2434 wishing to run older Solaris versions.
2443 DEF("full-screen", 0, QEMU_OPTION_full_screen,
2444 "-full-screen start in full screen\n", QEMU_ARCH_ALL)
2447 Start in full screen.
2450 DEF("g", HAS_ARG, QEMU_OPTION_g ,
2451 "-g WxH[xDEPTH] Set the initial graphical resolution and depth\n",
2452 QEMU_ARCH_PPC | QEMU_ARCH_SPARC | QEMU_ARCH_M68K)
2454 ``-g`` *width*\ ``x``\ *height*\ ``[x``\ *depth*\ ``]``
2455 Set the initial graphical resolution and depth (PPC, SPARC only).
2457 For PPC the default is 800x600x32.
2459 For SPARC with the TCX graphics device, the default is 1024x768x8
2460 with the option of 1024x768x24. For cgthree, the default is
2461 1024x768x8 with the option of 1152x900x8 for people who wish to use
2466 DEF("vnc", HAS_ARG, QEMU_OPTION_vnc ,
2467 "-vnc <display> shorthand for -display vnc=<display>\n", QEMU_ARCH_ALL)
2470 ``-vnc display[,option[,option[,...]]]``
2471 Normally, if QEMU is compiled with graphical window support, it
2472 displays output such as guest graphics, guest console, and the QEMU
2473 monitor in a window. With this option, you can have QEMU listen on
2474 VNC display display and redirect the VGA display over the VNC
2475 session. It is very useful to enable the usb tablet device when
2476 using this option (option ``-device usb-tablet``). When using the
2477 VNC display, you must use the ``-k`` parameter to set the keyboard
2478 layout if you are not using en-us. Valid syntax for the display is
2481 With this option, QEMU will try next available VNC displays,
2482 until the number L, if the originally defined "-vnc display" is
2483 not available, e.g. port 5900+display is already used by another
2484 application. By default, to=0.
2487 TCP connections will only be allowed from host on display d. By
2488 convention the TCP port is 5900+d. Optionally, host can be
2489 omitted in which case the server will accept connections from
2493 Connections will be allowed over UNIX domain sockets where path
2494 is the location of a unix socket to listen for connections on.
2497 VNC is initialized but not started. The monitor ``change``
2498 command can be used to later start the VNC server.
2500 Following the display value there may be one or more option flags
2501 separated by commas. Valid options are
2504 Connect to a listening VNC client via a "reverse" connection.
2505 The client is specified by the display. For reverse network
2506 connections (host:d,``reverse``), the d argument is a TCP port
2507 number, not a display number.
2509 ``websocket=on|off``
2510 Opens an additional TCP listening port dedicated to VNC
2511 Websocket connections. If a bare websocket option is given, the
2512 Websocket port is 5700+display. An alternative port can be
2513 specified with the syntax ``websocket``\ =port.
2515 If host is specified connections will only be allowed from this
2516 host. It is possible to control the websocket listen address
2517 independently, using the syntax ``websocket``\ =host:port.
2519 Websocket could be allowed over UNIX domain socket, using the syntax
2520 ``websocket``\ =unix:path, where path is the location of a unix socket
2521 to listen for connections on.
2523 If no TLS credentials are provided, the websocket connection
2524 runs in unencrypted mode. If TLS credentials are provided, the
2525 websocket connection requires encrypted client connections.
2528 Require that password based authentication is used for client
2531 The password must be set separately using the ``set_password``
2532 command in the :ref:`QEMU monitor`. The
2533 syntax to change your password is:
2534 ``set_password <protocol> <password>`` where <protocol> could be
2535 either "vnc" or "spice".
2537 If you would like to change <protocol> password expiration, you
2538 should use ``expire_password <protocol> <expiration-time>``
2539 where expiration time could be one of the following options:
2540 now, never, +seconds or UNIX time of expiration, e.g. +60 to
2541 make password expire in 60 seconds, or 1335196800 to make
2542 password expire on "Mon Apr 23 12:00:00 EDT 2012" (UNIX time for
2543 this date and time).
2545 You can also use keywords "now" or "never" for the expiration
2546 time to allow <protocol> password to expire immediately or never
2549 ``password-secret=<secret-id>``
2550 Require that password based authentication is used for client
2551 connections, using the password provided by the ``secret``
2552 object identified by ``secret-id``.
2555 Provides the ID of a set of TLS credentials to use to secure the
2556 VNC server. They will apply to both the normal VNC server socket
2557 and the websocket socket (if enabled). Setting TLS credentials
2558 will cause the VNC server socket to enable the VeNCrypt auth
2559 mechanism. The credentials should have been previously created
2560 using the ``-object tls-creds`` argument.
2563 Provides the ID of the QAuthZ authorization object against which
2564 the client's x509 distinguished name will validated
. This object
2565 is only resolved at time of use
, so can be deleted and recreated
2566 on the fly
while the VNC server is active
. If missing
, it will
2567 default to denying access
.
2570 Require that the client use SASL to authenticate with the VNC
2571 server
. The exact choice of authentication method used is
2572 controlled from the system
/ user
's SASL configuration file for
2573 the 'qemu
' service. This is typically found in
2574 /etc/sasl2/qemu.conf. If running QEMU as an unprivileged user,
2575 an environment variable SASL\_CONF\_PATH can be used to make it
2576 search alternate locations for the service config. While some
2577 SASL auth methods can also provide data encryption (eg GSSAPI),
2578 it is recommended that SASL always be combined with the 'tls
'
2579 and 'x509
' settings to enable use of SSL and server
2580 certificates. This ensures a data encryption preventing
2581 compromise of authentication credentials. See the
2582 :ref:`VNC security` section in the System Emulation Users Guide
2583 for details on using SASL authentication.
2586 Provides the ID of the QAuthZ authorization object against which
2587 the client's SASL username will validated
. This object is only
2588 resolved at time of use
, so can be deleted and recreated on the
2589 fly
while the VNC server is active
. If missing
, it will
default
2593 Legacy method
for enabling authorization of clients against the
2594 x509 distinguished name and SASL username
. It results
in the
2595 creation of two ``authz
-list`` objects with IDs of
2596 ``vnc
.username`` and ``vnc
.x509dname``
. The rules
for these
2597 objects must be configured with the HMP ACL commands
.
2599 This option is deprecated and should no longer be used
. The
new
2600 ``sasl
-authz`` and ``tls
-authz`` options are a replacement
.
2603 Enable lossy compression
methods (gradient
, JPEG
, ...). If
this
2604 option is set
, VNC client may receive lossy framebuffer updates
2605 depending on its encoding settings
. Enabling
this option can
2606 save a lot of bandwidth at the expense of quality
.
2608 ``non
-adaptive
=on|off``
2609 Disable adaptive encodings
. Adaptive encodings are enabled by
2610 default. An adaptive encoding will
try to detect frequently
2611 updated screen regions
, and send updates
in these regions
using
2612 a lossy
encoding (like JPEG
). This can be really helpful to save
2613 bandwidth when playing videos
. Disabling adaptive encodings
2614 restores the original
static behavior of encodings like Tight
.
2616 ``share
=[allow
-exclusive|force
-shared|ignore
]``
2617 Set display sharing policy
. 'allow-exclusive' allows clients to
2618 ask
for exclusive access
. As suggested by the rfb spec
this is
2619 implemented by dropping other connections
. Connecting multiple
2620 clients
in parallel requires all clients asking
for a shared
2621 session (vncviewer
: -shared
switch). This is the
default.
2622 'force-shared' disables exclusive client access
. Useful
for
2623 shared desktop sessions
, where you don
't want someone forgetting
2624 specify -shared disconnect everybody else. 'ignore
' completely
2625 ignores the shared flag and allows everybody connect
2626 unconditionally. Doesn't conform to the rfb spec but is
2627 traditional QEMU behavior
.
2630 Set keyboard delay
, for key down and key up events
, in
2631 milliseconds
. Default is
10. Keyboards are low
-bandwidth
2632 devices
, so
this slowdown can help the device and guest to keep
2633 up and not lose events
in case events are arriving
in bulk
.
2634 Possible causes
for the latter are flaky network connections
, or
2635 scripts
for automated testing
.
2637 ``audiodev
=audiodev``
2638 Use the specified audiodev when the VNC client requests audio
2639 transmission
. When not
using an
-audiodev argument
, this option
2640 must be omitted
, otherwise is must be present and specify a
2643 ``power
-control
=on|off``
2644 Permit the remote client to issue shutdown
, reboot or reset power
2648 ARCHHEADING(, QEMU_ARCH_I386
)
2650 ARCHHEADING(i386 target only
:, QEMU_ARCH_I386
)
2652 DEF("win2k-hack", 0, QEMU_OPTION_win2k_hack
,
2653 "-win2k-hack use it when installing Windows 2000 to avoid a disk full bug\n",
2657 Use it when installing Windows
2000 to avoid a disk full bug
. After
2658 Windows
2000 is installed
, you no longer need
this option (this
2659 option slows down the IDE transfers
). Synonym of ``
-global
2660 ide
-device
.win2k
-install
-hack
=on``
.
2663 DEF("no-fd-bootchk", 0, QEMU_OPTION_no_fd_bootchk
,
2664 "-no-fd-bootchk disable boot signature checking for floppy disks\n",
2668 Disable boot signature checking
for floppy disks
in BIOS
. May be
2669 needed to boot from old floppy disks
. Synonym of ``
-m fd
-bootchk
=off``
.
2672 DEF("acpitable", HAS_ARG
, QEMU_OPTION_acpitable
,
2673 "-acpitable [sig=str][,rev=n][,oem_id=str][,oem_table_id=str][,oem_rev=n][,asl_compiler_id=str][,asl_compiler_rev=n][,{data|file}=file1[:file2]...]\n"
2674 " ACPI table description\n", QEMU_ARCH_I386
)
2676 ``
-acpitable
[sig
=str
][,rev
=n
][,oem_id
=str
][,oem_table_id
=str
][,oem_rev
=n
] [,asl_compiler_id
=str
][,asl_compiler_rev
=n
][,data
=file1
[:file2
]...]``
2677 Add ACPI table with specified header fields and context from
2678 specified files
. For file
=, take whole ACPI table from the specified
2679 files
, including all ACPI
headers (possible overridden by other
2680 options
). For data
=, only data portion of the table is used
, all
2681 header information is specified
in the command line
. If a SLIC table
2682 is supplied to QEMU
, then the SLIC
's oem\_id and oem\_table\_id
2683 fields will override the same in the RSDT and the FADT (a.k.a.
2684 FACP), in order to ensure the field matches required by the
2685 Microsoft SLIC spec and the ACPI spec.
2688 DEF("smbios", HAS_ARG, QEMU_OPTION_smbios,
2689 "-smbios file=binary\n"
2690 " load SMBIOS entry from binary file\n"
2691 "-smbios type=0[,vendor=str][,version=str][,date=str][,release=%d.%d]\n"
2693 " specify SMBIOS type 0 fields\n"
2694 "-smbios type=1[,manufacturer=str][,product=str][,version=str][,serial=str]\n"
2695 " [,uuid=uuid][,sku=str][,family=str]\n"
2696 " specify SMBIOS type 1 fields\n"
2697 "-smbios type=2[,manufacturer=str][,product=str][,version=str][,serial=str]\n"
2698 " [,asset=str][,location=str]\n"
2699 " specify SMBIOS type 2 fields\n"
2700 "-smbios type=3[,manufacturer=str][,version=str][,serial=str][,asset=str]\n"
2702 " specify SMBIOS type 3 fields\n"
2703 "-smbios type=4[,sock_pfx=str][,manufacturer=str][,version=str][,serial=str]\n"
2704 " [,asset=str][,part=str][,max-speed=%d][,current-speed=%d]\n"
2705 " [,processor-family=%d,processor-id=%d]\n"
2706 " specify SMBIOS type 4 fields\n"
2707 "-smbios type=8[,external_reference=str][,internal_reference=str][,connector_type=%d][,port_type=%d]\n"
2708 " specify SMBIOS type 8 fields\n"
2709 "-smbios type=11[,value=str][,path=filename]\n"
2710 " specify SMBIOS type 11 fields\n"
2711 "-smbios type=17[,loc_pfx=str][,bank=str][,manufacturer=str][,serial=str]\n"
2712 " [,asset=str][,part=str][,speed=%d]\n"
2713 " specify SMBIOS type 17 fields\n"
2714 "-smbios type=41[,designation=str][,kind=str][,instance=%d][,pcidev=str]\n"
2715 " specify SMBIOS type 41 fields\n",
2716 QEMU_ARCH_I386 | QEMU_ARCH_ARM | QEMU_ARCH_LOONGARCH | QEMU_ARCH_RISCV)
2718 ``-smbios file=binary``
2719 Load SMBIOS entry from binary file.
2721 ``-smbios type=0[,vendor=str][,version=str][,date=str][,release=%d.%d][,uefi=on|off]``
2722 Specify SMBIOS type 0 fields
2724 ``-smbios type=1[,manufacturer=str][,product=str][,version=str][,serial=str][,uuid=uuid][,sku=str][,family=str]``
2725 Specify SMBIOS type 1 fields
2727 ``-smbios type=2[,manufacturer=str][,product=str][,version=str][,serial=str][,asset=str][,location=str]``
2728 Specify SMBIOS type 2 fields
2730 ``-smbios type=3[,manufacturer=str][,version=str][,serial=str][,asset=str][,sku=str]``
2731 Specify SMBIOS type 3 fields
2733 ``-smbios type=4[,sock_pfx=str][,manufacturer=str][,version=str][,serial=str][,asset=str][,part=str][,processor-family=%d][,processor-id=%d]``
2734 Specify SMBIOS type 4 fields
2736 ``-smbios type=9[,slot_designation=str][,slot_type=%d][,slot_data_bus_width=%d][,current_usage=%d][,slot_length=%d][,slot_id=%d][,slot_characteristics1=%d][,slot_characteristics12=%d][,pci_device=str]``
2737 Specify SMBIOS type 9 fields
2739 ``-smbios type=11[,value=str][,path=filename]``
2740 Specify SMBIOS type 11 fields
2742 This argument can be repeated multiple times, and values are added in the order they are parsed.
2743 Applications intending to use OEM strings data are encouraged to use their application name as
2744 a prefix for the value string. This facilitates passing information for multiple applications
2747 The ``value=str`` syntax provides the string data inline, while the ``path=filename`` syntax
2748 loads data from a file on disk. Note that the file is not permitted to contain any NUL bytes.
2750 Both the ``value`` and ``path`` options can be repeated multiple times and will be added to
2751 the SMBIOS table in the order in which they appear.
2753 Note that on the x86 architecture, the total size of all SMBIOS tables is limited to 65535
2754 bytes. Thus the OEM strings data is not suitable for passing large amounts of data into the
2755 guest. Instead it should be used as a indicator to inform the guest where to locate the real
2756 data set, for example, by specifying the serial ID of a block device.
2758 An example passing three strings is
2762 -smbios type=11,value=cloud-init:ds=nocloud-net;s=http://10.10.0.1:8000/,\\
2763 value=anaconda:method=http://dl.fedoraproject.org/pub/fedora/linux/releases/25/x86_64/os,\\
2764 path=/some/file/with/oemstringsdata.txt
2766 In the guest OS this is visible with the ``dmidecode`` command
2771 Handle 0x0E00, DMI type 11, 5 bytes
2773 String 1: cloud-init:ds=nocloud-net;s=http://10.10.0.1:8000/
2774 String 2: anaconda:method=http://dl.fedoraproject.org/pub/fedora/linux/releases/25/x86_64/os
2775 String 3: myapp:some extra data
2778 ``-smbios type=17[,loc_pfx=str][,bank=str][,manufacturer=str][,serial=str][,asset=str][,part=str][,speed=%d]``
2779 Specify SMBIOS type 17 fields
2781 ``-smbios type=41[,designation=str][,kind=str][,instance=%d][,pcidev=str]``
2782 Specify SMBIOS type 41 fields
2784 This argument can be repeated multiple times. Its main use is to allow network interfaces be created
2785 as ``enoX`` on Linux, with X being the instance number, instead of the name depending on the interface
2786 position on the PCI bus.
2788 Here is an example of use:
2792 -netdev user,id=internet \\
2793 -device virtio-net-pci,mac=50:54:00:00:00:42,netdev=internet,id=internet-dev \\
2794 -smbios type=41,designation='Onboard LAN
',instance=1,kind=ethernet,pcidev=internet-dev
2796 In the guest OS, the device should then appear as ``eno1``:
2801 lo UNKNOWN 00:00:00:00:00:00 <LOOPBACK,UP,LOWER_UP>
2802 eno1 UP 50:54:00:00:00:42 <BROADCAST,MULTICAST,UP,LOWER_UP>
2804 Currently, the PCI device has to be attached to the root bus.
2810 DEFHEADING(Network options:)
2812 DEF("netdev", HAS_ARG, QEMU_OPTION_netdev,
2814 "-netdev user,id=str[,ipv4=on|off][,net=addr[/mask]][,host=addr]\n"
2815 " [,ipv6=on|off][,ipv6-net=addr[/int]][,ipv6-host=addr]\n"
2816 " [,restrict=on|off][,hostname=host][,dhcpstart=addr]\n"
2817 " [,dns=addr][,ipv6-dns=addr][,dnssearch=domain][,domainname=domain]\n"
2818 " [,tftp=dir][,tftp-server-name=name][,bootfile=f][,hostfwd=rule][,guestfwd=rule]"
2820 "[,smb=dir[,smbserver=addr]]\n"
2822 " configure a user mode network backend with ID 'str
',\n"
2823 " its DHCP server and optional services\n"
2826 "-netdev tap,id=str,ifname=name\n"
2827 " configure a host TAP network backend with ID 'str
'\n"
2829 "-netdev tap,id=str[,fd=h][,fds=x:y:...:z][,ifname=name][,script=file][,downscript=dfile]\n"
2830 " [,br=bridge][,helper=helper][,sndbuf=nbytes][,vnet_hdr=on|off][,vhost=on|off]\n"
2831 " [,vhostfd=h][,vhostfds=x:y:...:z][,vhostforce=on|off][,queues=n]\n"
2833 " configure a host TAP network backend with ID 'str
'\n"
2834 " connected to a bridge (default=" DEFAULT_BRIDGE_INTERFACE ")\n"
2835 " use network scripts 'file
' (default=" DEFAULT_NETWORK_SCRIPT ")\n"
2836 " to configure it and 'dfile
' (default=" DEFAULT_NETWORK_DOWN_SCRIPT ")\n"
2837 " to deconfigure it\n"
2838 " use '[down
]script
=no
' to disable script execution\n"
2839 " use network helper 'helper
' (default=" DEFAULT_BRIDGE_HELPER ") to\n"
2841 " use 'fd
=h
' to connect to an already opened TAP interface\n"
2842 " use 'fds
=x
:y
:...:z
' to connect to already opened multiqueue capable TAP interfaces\n"
2843 " use 'sndbuf
=nbytes
' to limit the size of the send buffer (the\n"
2844 " default is disabled 'sndbuf
=0' to enable flow control set 'sndbuf
=1048576')\n"
2845 " use vnet_hdr=off to avoid enabling the IFF_VNET_HDR tap flag\n"
2846 " use vnet_hdr=on to make the lack of IFF_VNET_HDR support an error condition\n"
2847 " use vhost=on to enable experimental in kernel accelerator\n"
2848 " (only has effect for virtio guests which use MSIX)\n"
2849 " use vhostforce=on to force vhost on for non-MSIX virtio guests\n"
2850 " use 'vhostfd
=h
' to connect to an already opened vhost net device\n"
2851 " use 'vhostfds
=x
:y
:...:z to connect to multiple already opened vhost net devices
\n"
2852 " use
'queues=n' to specify the number of queues to be created
for multiqueue TAP
\n"
2853 " use
'poll-us=n' to specify the maximum number of microseconds that could be
\n"
2854 " spent on busy polling
for vhost net
\n"
2855 "-netdev bridge
,id
=str
[,br
=bridge
][,helper
=helper
]\n"
2856 " configure a host TAP network backend with ID
'str' that is
\n"
2857 " connected to a
bridge (default=" DEFAULT_BRIDGE_INTERFACE ")\n"
2858 " using the program
'helper (default=" DEFAULT_BRIDGE_HELPER ")\n"
2861 "-netdev l2tpv3,id=str,src=srcaddr,dst=dstaddr[,srcport=srcport][,dstport=dstport]\n"
2862 " [,rxsession=rxsession],txsession=txsession[,ipv6=on|off][,udp=on|off]\n"
2863 " [,cookie64=on|off][,counter][,pincounter][,txcookie=txcookie]\n"
2864 " [,rxcookie=rxcookie][,offset=offset]\n"
2865 " configure a network backend with ID 'str
' connected to\n"
2866 " an Ethernet over L2TPv3 pseudowire.\n"
2867 " Linux kernel 3.3+ as well as most routers can talk\n"
2868 " L2TPv3. This transport allows connecting a VM to a VM,\n"
2869 " VM to a router and even VM to Host. It is a nearly-universal\n"
2870 " standard (RFC3931). Note - this implementation uses static\n"
2871 " pre-configured tunnels (same as the Linux kernel).\n"
2872 " use 'src
=' to specify source address\n"
2873 " use 'dst
=' to specify destination address\n"
2874 " use 'udp
=on
' to specify udp encapsulation\n"
2875 " use 'srcport
=' to specify source udp port\n"
2876 " use 'dstport
=' to specify destination udp port\n"
2877 " use 'ipv6
=on
' to force v6\n"
2878 " L2TPv3 uses cookies to prevent misconfiguration as\n"
2879 " well as a weak security measure\n"
2880 " use 'rxcookie
=0x012345678' to specify a rxcookie\n"
2881 " use 'txcookie
=0x012345678' to specify a txcookie\n"
2882 " use 'cookie64
=on
' to set cookie size to 64 bit, otherwise 32\n"
2883 " use 'counter
=off
' to force a 'cut
-down
' L2TPv3 with no counter\n"
2884 " use 'pincounter
=on
' to work around broken counter handling in peer\n"
2885 " use 'offset
=X
' to add an extra offset between header and data\n"
2887 "-netdev socket,id=str[,fd=h][,listen=[host]:port][,connect=host:port]\n"
2888 " configure a network backend to connect to another network\n"
2889 " using a socket connection\n"
2890 "-netdev socket,id=str[,fd=h][,mcast=maddr:port[,localaddr=addr]]\n"
2891 " configure a network backend to connect to a multicast maddr and port\n"
2892 " use 'localaddr
=addr
' to specify the host address to send packets from\n"
2893 "-netdev socket,id=str[,fd=h][,udp=host:port][,localaddr=host:port]\n"
2894 " configure a network backend to connect to another network\n"
2895 " using an UDP tunnel\n"
2896 "-netdev stream,id=str[,server=on|off],addr.type=inet,addr.host=host,addr.port=port[,to=maxport][,numeric=on|off][,keep-alive=on|off][,mptcp=on|off][,addr.ipv4=on|off][,addr.ipv6=on|off][,reconnect=seconds]\n"
2897 "-netdev stream,id=str[,server=on|off],addr.type=unix,addr.path=path[,abstract=on|off][,tight=on|off][,reconnect=seconds]\n"
2898 "-netdev stream,id=str[,server=on|off],addr.type=fd,addr.str=file-descriptor[,reconnect=seconds]\n"
2899 " configure a network backend to connect to another network\n"
2900 " using a socket connection in stream mode.\n"
2901 "-netdev dgram,id=str,remote.type=inet,remote.host=maddr,remote.port=port[,local.type=inet,local.host=addr]\n"
2902 "-netdev dgram,id=str,remote.type=inet,remote.host=maddr,remote.port=port[,local.type=fd,local.str=file-descriptor]\n"
2903 " configure a network backend to connect to a multicast maddr and port\n"
2904 " use ``local.host=addr`` to specify the host address to send packets from\n"
2905 "-netdev dgram,id=str,local.type=inet,local.host=addr,local.port=port[,remote.type=inet,remote.host=addr,remote.port=port]\n"
2906 "-netdev dgram,id=str,local.type=unix,local.path=path[,remote.type=unix,remote.path=path]\n"
2907 "-netdev dgram,id=str,local.type=fd,local.str=file-descriptor\n"
2908 " configure a network backend to connect to another network\n"
2909 " using an UDP tunnel\n"
2911 "-netdev vde,id=str[,sock=socketpath][,port=n][,group=groupname][,mode=octalmode]\n"
2912 " configure a network backend to connect to port 'n
' of a vde switch\n"
2913 " running on host and listening for incoming connections on 'socketpath
'.\n"
2914 " Use group 'groupname
' and mode 'octalmode
' to change default\n"
2915 " ownership and permissions for communication port.\n"
2917 #ifdef CONFIG_NETMAP
2918 "-netdev netmap,id=str,ifname=name[,devname=nmname]\n"
2919 " attach to the existing netmap-enabled network interface 'name
', or to a\n"
2920 " VALE port (created on the fly) called 'name
' ('nmname
' is name of the \n"
2921 " netmap device, defaults to '/dev
/netmap
')\n"
2923 #ifdef CONFIG_AF_XDP
2924 "-netdev af-xdp,id=str,ifname=name[,mode=native|skb][,force-copy=on|off]\n"
2925 " [,queues=n][,start-queue=m][,inhibit=on|off][,sock-fds=x:y:...:z]\n"
2926 " attach to the existing network interface 'name
' with AF_XDP socket\n"
2927 " use 'mode
=MODE
' to specify an XDP program attach mode\n"
2928 " use 'force
-copy
=on|off
' to force XDP copy mode even if device supports zero-copy (default: off)\n"
2929 " use 'inhibit
=on|off
' to inhibit loading of a default XDP program (default: off)\n"
2930 " with inhibit=on,\n"
2931 " use 'sock
-fds
' to provide file descriptors for already open AF_XDP sockets\n"
2932 " added to a socket map in XDP program. One socket per queue.\n"
2933 " use 'queues
=n
' to specify how many queues of a multiqueue interface should be used\n"
2934 " use 'start
-queue
=m
' to specify the first queue that should be used\n"
2937 "-netdev vhost-user,id=str,chardev=dev[,vhostforce=on|off]\n"
2938 " configure a vhost-user network, backed by a chardev 'dev
'\n"
2941 "-netdev vhost-vdpa,id=str[,vhostdev=/path/to/dev][,vhostfd=h]\n"
2942 " configure a vhost-vdpa network,Establish a vhost-vdpa netdev\n"
2943 " use 'vhostdev
=/path
/to
/dev
' to open a vhost vdpa device\n"
2944 " use 'vhostfd
=h
' to connect to an already opened vhost vdpa device\n"
2947 "-netdev vmnet-host,id=str[,isolated=on|off][,net-uuid=uuid]\n"
2948 " [,start-address=addr,end-address=addr,subnet-mask=mask]\n"
2949 " configure a vmnet network backend in host mode with ID 'str
',\n"
2950 " isolate this interface from others with 'isolated
',\n"
2951 " configure the address range and choose a subnet mask,\n"
2952 " specify network UUID 'uuid
' to disable DHCP and interact with\n"
2953 " vmnet-host interfaces within this isolated network\n"
2954 "-netdev vmnet-shared,id=str[,isolated=on|off][,nat66-prefix=addr]\n"
2955 " [,start-address=addr,end-address=addr,subnet-mask=mask]\n"
2956 " configure a vmnet network backend in shared mode with ID 'str
',\n"
2957 " configure the address range and choose a subnet mask,\n"
2958 " set IPv6 ULA prefix (of length 64) to use for internal network,\n"
2959 " isolate this interface from others with 'isolated
'\n"
2960 "-netdev vmnet-bridged,id=str,ifname=name[,isolated=on|off]\n"
2961 " configure a vmnet network backend in bridged mode with ID 'str
',\n"
2962 " use 'ifname
=name
' to select a physical network interface to be bridged,\n"
2963 " isolate this interface from others with 'isolated
'\n"
2965 "-netdev hubport,id=str,hubid=n[,netdev=nd]\n"
2966 " configure a hub port on the hub with ID 'n
'\n", QEMU_ARCH_ALL)
2967 DEF("nic", HAS_ARG, QEMU_OPTION_nic,
2978 #ifdef CONFIG_NETMAP
2981 #ifdef CONFIG_AF_XDP
2988 "vmnet-host|vmnet-shared|vmnet-bridged|"
2990 "socket][,option][,...][mac=macaddr]\n"
2991 " initialize an on-board / default host NIC (using MAC address\n"
2992 " macaddr) and connect it to the given host network backend\n"
2993 "-nic none use it alone to have zero network devices (the default is to\n"
2994 " provided a 'user
' network connection)\n",
2996 DEF("net", HAS_ARG, QEMU_OPTION_net,
2997 "-net nic[,macaddr=mac][,model=type][,name=str][,addr=str][,vectors=v]\n"
2998 " configure or create an on-board (or machine default) NIC and\n"
2999 " connect it to hub 0 (please use -nic unless you need a hub)\n"
3009 #ifdef CONFIG_NETMAP
3012 #ifdef CONFIG_AF_XDP
3016 "vmnet-host|vmnet-shared|vmnet-bridged|"
3018 "socket][,option][,option][,...]\n"
3019 " old way to initialize a host network interface\n"
3020 " (use the -netdev option if possible instead)\n", QEMU_ARCH_ALL)
3022 ``-nic [tap|bridge|user|l2tpv3|vde|netmap|af-xdp|vhost-user|socket][,...][,mac=macaddr][,model=mn]``
3023 This option is a shortcut for configuring both the on-board
3024 (default) guest NIC hardware and the host network backend in one go.
3025 The host backend options are the same as with the corresponding
3026 ``-netdev`` options below. The guest NIC model can be set with
3027 ``model=modelname``. Use ``model=help`` to list the available device
3028 types. The hardware MAC address can be set with ``mac=macaddr``.
3030 The following two example do exactly the same, to show how ``-nic``
3031 can be used to shorten the command line length:
3035 |qemu_system| -netdev user,id=n1,ipv6=off -device e1000,netdev=n1,mac=52:54:98:76:54:32
3036 |qemu_system| -nic user,ipv6=off,model=e1000,mac=52:54:98:76:54:32
3039 Indicate that no network devices should be configured. It is used to
3040 override the default configuration (default NIC with "user" host
3041 network backend) which is activated if no other networking options
3044 ``-netdev user,id=id[,option][,option][,...]``
3045 Configure user mode host network backend which requires no
3046 administrator privilege to run. Valid options are:
3049 Assign symbolic name for use in monitor commands.
3051 ``ipv4=on|off and ipv6=on|off``
3052 Specify that either IPv4 or IPv6 must be enabled. If neither is
3053 specified both protocols are enabled.
3056 Set IP network address the guest will see. Optionally specify
3057 the netmask, either in the form a.b.c.d or as number of valid
3058 top-most bits. Default is 10.0.2.0/24.
3061 Specify the guest-visible address of the host. Default is the
3062 2nd IP in the guest network, i.e. x.x.x.2.
3064 ``ipv6-net=addr[/int]``
3065 Set IPv6 network address the guest will see (default is
3066 fec0::/64). The network prefix is given in the usual hexadecimal
3067 IPv6 address notation. The prefix size is optional, and is given
3068 as the number of valid top-most bits (default is 64).
3071 Specify the guest-visible IPv6 address of the host. Default is
3072 the 2nd IPv6 in the guest network, i.e. xxxx::2.
3075 If this option is enabled, the guest will be isolated, i.e. it
3076 will not be able to contact the host and no guest IP packets
3077 will be routed over the host to the outside. This option does
3078 not affect any explicitly set forwarding rules.
3081 Specifies the client hostname reported by the built-in DHCP
3085 Specify the first of the 16 IPs the built-in DHCP server can
3086 assign. Default is the 15th to 31st IP in the guest network,
3087 i.e. x.x.x.15 to x.x.x.31.
3090 Specify the guest-visible address of the virtual nameserver. The
3091 address must be different from the host address. Default is the
3092 3rd IP in the guest network, i.e. x.x.x.3.
3095 Specify the guest-visible address of the IPv6 virtual
3096 nameserver. The address must be different from the host address.
3097 Default is the 3rd IP in the guest network, i.e. xxxx::3.
3099 ``dnssearch=domain``
3100 Provides an entry for the domain-search list sent by the
3101 built-in DHCP server. More than one domain suffix can be
3102 transmitted by specifying this option multiple times. If
3103 supported, this will cause the guest to automatically try to
3104 append the given domain suffix(es) in case a domain name can not
3111 |qemu_system| -nic user,dnssearch=mgmt.example.org,dnssearch=example.org
3113 ``domainname=domain``
3114 Specifies the client domain name reported by the built-in DHCP
3118 When using the user mode network stack, activate a built-in TFTP
3119 server. The files in dir will be exposed as the root of a TFTP
3120 server. The TFTP client on the guest must be configured in
3121 binary mode (use the command ``bin`` of the Unix TFTP client).
3122 The built-in TFTP server is read-only; it does not implement any
3123 command for writing files. QEMU will not write to this directory.
3125 ``tftp-server-name=name``
3126 In BOOTP reply, broadcast name as the "TFTP server name"
3127 (RFC2132 option 66). This can be used to advise the guest to
3128 load boot files or configurations from a different server than
3132 When using the user mode network stack, broadcast file as the
3133 BOOTP filename. In conjunction with ``tftp``, this can be used
3134 to network boot a guest from a local directory.
3136 Example (using pxelinux):
3140 |qemu_system| -hda linux.img -boot n -device e1000,netdev=n1 \\
3141 -netdev user,id=n1,tftp=/path/to/tftp/files,bootfile=/pxelinux.0
3143 ``smb=dir[,smbserver=addr]``
3144 When using the user mode network stack, activate a built-in SMB
3145 server so that Windows OSes can access to the host files in
3146 ``dir`` transparently. The IP address of the SMB server can be
3147 set to addr. By default the 4th IP in the guest network is used,
3150 In the guest Windows OS, the line:
3156 must be added in the file ``C:\WINDOWS\LMHOSTS`` (for windows
3157 9x/Me) or ``C:\WINNT\SYSTEM32\DRIVERS\ETC\LMHOSTS`` (Windows
3160 Then ``dir`` can be accessed in ``\\smbserver\qemu``.
3162 Note that a SAMBA server must be installed on the host OS.
3164 ``hostfwd=[tcp|udp]:[hostaddr]:hostport-[guestaddr]:guestport``
3165 Redirect incoming TCP or UDP connections to the host port
3166 hostport to the guest IP address guestaddr on guest port
3167 guestport. If guestaddr is not specified, its value is x.x.x.15
3168 (default first address given by the built-in DHCP server). By
3169 specifying hostaddr, the rule can be bound to a specific host
3170 interface. If no connection type is set, TCP is used. This
3171 option can be given multiple times.
3173 For example, to redirect host X11 connection from screen 1 to
3174 guest screen 0, use the following:
3179 |qemu_system| -nic user,hostfwd=tcp:127.0.0.1:6001-:6000
3180 # this host xterm should open in the guest X11 server
3183 To redirect telnet connections from host port 5555 to telnet
3184 port on the guest, use the following:
3189 |qemu_system| -nic user,hostfwd=tcp::5555-:23
3190 telnet localhost 5555
3192 Then when you use on the host ``telnet localhost 5555``, you
3193 connect to the guest telnet server.
3195 ``guestfwd=[tcp]:server:port-dev``; \ ``guestfwd=[tcp]:server:port-cmd:command``
3196 Forward guest TCP connections to the IP address server on port
3197 port to the character device dev or to a program executed by
3198 cmd:command which gets spawned for each connection. This option
3199 can be given multiple times.
3201 You can either use a chardev directly and have that one used
3202 throughout QEMU's lifetime
, like
in the following example
:
3206 # open
10.10.1.1:4321 on bootup
, connect
10.0.2.100:1234 to it whenever
3207 # the guest accesses it
3208 |qemu_system|
-nic user
,guestfwd
=tcp
:10.0.2.100:1234-tcp
:10.10.1.1:4321
3210 Or you can execute a command on every TCP connection established
3211 by the guest
, so that QEMU behaves similar to an inetd process
3212 for that virtual server
:
3216 # call
"netcat 10.10.1.1 4321" on every TCP connection to
10.0.2.100:1234
3217 # and connect the TCP stream to its stdin
/stdout
3218 |qemu_system|
-nic
'user,id=n1,guestfwd=tcp:10.0.2.100:1234-cmd:netcat 10.10.1.1 4321'
3220 ``
-netdev tap
,id
=id
[,fd
=h
][,ifname
=name
][,script
=file
][,downscript
=dfile
][,br
=bridge
][,helper
=helper
]``
3221 Configure a host TAP network backend with ID id
.
3223 Use the network script file to configure it and the network script
3224 dfile to deconfigure it
. If name is not provided
, the OS
3225 automatically provides one
. The
default network configure script is
3226 ``
/etc
/qemu
-ifup`` and the
default network deconfigure script is
3227 ``
/etc
/qemu
-ifdown``
. Use ``script
=no`` or ``downscript
=no`` to
3228 disable script execution
.
3230 If running QEMU as an unprivileged user
, use the network helper
3231 to configure the TAP
interface and attach it to the bridge
.
3232 The
default network helper executable is
3233 ``
/path
/to
/qemu
-bridge
-helper`` and the
default bridge device is
3236 ``fd``\
=h can be used to specify the handle of an already opened
3243 #launch a QEMU instance with the
default network script
3244 |qemu_system| linux
.img
-nic tap
3248 #launch a QEMU instance with two NICs
, each one connected
3250 |qemu_system| linux
.img
\\
3251 -netdev tap
,id
=nd0
,ifname
=tap0
-device e1000
,netdev
=nd0
\\
3252 -netdev tap
,id
=nd1
,ifname
=tap1
-device rtl8139
,netdev
=nd1
3256 #launch a QEMU instance with the
default network helper to
3257 #connect a TAP device to bridge br0
3258 |qemu_system| linux
.img
-device virtio
-net
-pci
,netdev
=n1
\\
3259 -netdev tap
,id
=n1
,"helper=/path/to/qemu-bridge-helper"
3261 ``
-netdev bridge
,id
=id
[,br
=bridge
][,helper
=helper
]``
3262 Connect a host TAP network
interface to a host bridge device
.
3264 Use the network helper helper to configure the TAP
interface and
3265 attach it to the bridge
. The
default network helper executable is
3266 ``
/path
/to
/qemu
-bridge
-helper`` and the
default bridge device is
3273 #launch a QEMU instance with the
default network helper to
3274 #connect a TAP device to bridge br0
3275 |qemu_system| linux
.img
-netdev bridge
,id
=n1
-device virtio
-net
,netdev
=n1
3279 #launch a QEMU instance with the
default network helper to
3280 #connect a TAP device to bridge qemubr0
3281 |qemu_system| linux
.img
-netdev bridge
,br
=qemubr0
,id
=n1
-device virtio
-net
,netdev
=n1
3283 ``
-netdev socket
,id
=id
[,fd
=h
][,listen
=[host
]:port
][,connect
=host
:port
]``
3284 This host network backend can be used to connect the guest
's network
3285 to another QEMU virtual machine using a TCP socket connection. If
3286 ``listen`` is specified, QEMU waits for incoming connections on port
3287 (host is optional). ``connect`` is used to connect to another QEMU
3288 instance using the ``listen`` option. ``fd``\ =h specifies an
3289 already opened TCP socket.
3295 # launch a first QEMU instance
3296 |qemu_system| linux.img \\
3297 -device e1000,netdev=n1,mac=52:54:00:12:34:56 \\
3298 -netdev socket,id=n1,listen=:1234
3299 # connect the network of this instance to the network of the first instance
3300 |qemu_system| linux.img \\
3301 -device e1000,netdev=n2,mac=52:54:00:12:34:57 \\
3302 -netdev socket,id=n2,connect=127.0.0.1:1234
3304 ``-netdev socket,id=id[,fd=h][,mcast=maddr:port[,localaddr=addr]]``
3305 Configure a socket host network backend to share the guest's network
3306 traffic with another QEMU virtual machines
using a UDP multicast
3307 socket
, effectively making a bus
for every QEMU with same multicast
3308 address maddr and port
. NOTES
:
3310 1. Several QEMU can be running on different hosts and share same bus
3311 (assuming correct multicast setup
for these hosts
).
3313 2. mcast support is compatible with User Mode
Linux (argument
3314 ``ethN
=mcast``
), see http
://user-mode-linux.sf.net.
3316 3. Use ``fd
=h`` to specify an already opened UDP multicast socket
.
3322 # launch one QEMU instance
3323 |qemu_system| linux
.img
\\
3324 -device e1000
,netdev
=n1
,mac
=52:54:00:12:34:56 \\
3325 -netdev socket
,id
=n1
,mcast
=230.0.0.1:1234
3326 # launch another QEMU instance on same
"bus"
3327 |qemu_system| linux
.img
\\
3328 -device e1000
,netdev
=n2
,mac
=52:54:00:12:34:57 \\
3329 -netdev socket
,id
=n2
,mcast
=230.0.0.1:1234
3330 # launch yet another QEMU instance on same
"bus"
3331 |qemu_system| linux
.img
\\
3332 -device e1000
,netdev
=n3
,mac
=52:54:00:12:34:58 \\
3333 -netdev socket
,id
=n3
,mcast
=230.0.0.1:1234
3335 Example (User Mode Linux compat
.):
3339 # launch QEMU
instance (note mcast address selected is UML
's default)
3340 |qemu_system| linux.img \\
3341 -device e1000,netdev=n1,mac=52:54:00:12:34:56 \\
3342 -netdev socket,id=n1,mcast=239.192.168.1:1102
3344 /path/to/linux ubd0=/path/to/root_fs eth0=mcast
3346 Example (send packets from host's
1.2.3.4):
3350 |qemu_system| linux
.img
\\
3351 -device e1000
,netdev
=n1
,mac
=52:54:00:12:34:56 \\
3352 -netdev socket
,id
=n1
,mcast
=239.192.168.1:1102,localaddr
=1.2.3.4
3354 ``
-netdev l2tpv3
,id
=id
,src
=srcaddr
,dst
=dstaddr
[,srcport
=srcport
][,dstport
=dstport
],txsession
=txsession
[,rxsession
=rxsession
][,ipv6
=on|off
][,udp
=on|off
][,cookie64
][,counter
][,pincounter
][,txcookie
=txcookie
][,rxcookie
=rxcookie
][,offset
=offset
]``
3355 Configure a L2TPv3 pseudowire host network backend
. L2TPv3 (RFC3931
)
3356 is a popular protocol to transport
Ethernet (and other Layer
2) data
3357 frames between two systems
. It is present
in routers
, firewalls and
3358 the Linux
kernel (from version
3.3 onwards
).
3360 This transport allows a VM to communicate to another VM
, router or
3364 source
address (mandatory
)
3367 destination
address (mandatory
)
3370 select udp
encapsulation (default is ip
).
3376 destination udp port
.
3379 force v6
, otherwise defaults to v4
.
3381 ``rxcookie
=rxcookie``
; \ ``txcookie
=txcookie``
3382 Cookies are a weak form of security
in the l2tpv3 specification
.
3383 Their
function is mostly to prevent misconfiguration
. By
default
3387 Set cookie size to
64 bit instead of the
default 32
3390 Force a
'cut-down' L2TPv3 with no counter as
in
3391 draft
-mkonstan
-l2tpext
-keyed
-ipv6
-tunnel
-00
3394 Work around broken counter handling
in peer
. This may also help
3395 on networks which have packet reorder
.
3398 Add an extra offset between header and data
3400 For example
, to attach a VM running on host
4.3.2.1 via L2TPv3 to
3401 the bridge br
-lan on the remote Linux host
1.2.3.4:
3405 # Setup tunnel on linux host
using raw ip as encapsulation
3407 ip l2tp add tunnel remote
4.3.2.1 local
1.2.3.4 tunnel_id
1 peer_tunnel_id
1 \\
3408 encap udp udp_sport
16384 udp_dport
16384
3409 ip l2tp add session tunnel_id
1 name vmtunnel0 session_id
\\
3410 0xFFFFFFFF peer_session_id
0xFFFFFFFF
3411 ifconfig vmtunnel0 mtu
1500
3412 ifconfig vmtunnel0 up
3413 brctl addif br
-lan vmtunnel0
3417 # launch QEMU instance
- if your network has reorder or is very lossy add
,pincounter
3419 |qemu_system| linux
.img
-device e1000
,netdev
=n1
\\
3420 -netdev l2tpv3
,id
=n1
,src
=4.2.3.1,dst
=1.2.3.4,udp
,srcport
=16384,dstport
=16384,rxsession
=0xffffffff,txsession
=0xffffffff,counter
3422 ``
-netdev vde
,id
=id
[,sock
=socketpath
][,port
=n
][,group
=groupname
][,mode
=octalmode
]``
3423 Configure VDE backend to connect to PORT n of a vde
switch running
3424 on host and listening
for incoming connections on socketpath
. Use
3425 GROUP groupname and MODE octalmode to change
default ownership and
3426 permissions
for communication port
. This option is only available
if
3427 QEMU has been compiled with vde support enabled
.
3434 vde_switch
-F
-sock
/tmp
/myswitch
3435 # launch QEMU instance
3436 |qemu_system| linux
.img
-nic vde
,sock
=/tmp
/myswitch
3438 ``
-netdev af
-xdp
,id
=str
,ifname
=name
[,mode
=native|skb
][,force
-copy
=on|off
][,queues
=n
][,start
-queue
=m
][,inhibit
=on|off
][,sock
-fds
=x
:y
:...:z
]``
3439 Configure AF_XDP backend to connect to a network
interface 'name'
3440 using AF_XDP socket
. A specific program attach mode
for a
default
3441 XDP program can be forced with
'mode', defaults to best
-effort
,
3442 where the likely most performant mode will be
in use
. Number of queues
3443 'n' should generally match the number or queues
in the
interface,
3444 defaults to
1. Traffic arriving on non
-configured device queues will
3445 not be delivered to the network backend
.
3449 # set number of queues to
4
3450 ethtool
-L eth0 combined
4
3451 # launch QEMU instance
3452 |qemu_system| linux
.img
-device virtio
-net
-pci
,netdev
=n1
\\
3453 -netdev af
-xdp
,id
=n1
,ifname
=eth0
,queues
=4
3455 'start-queue' option can be specified
if a particular range of queues
3456 [m
, m
+ n
] should be
in use
. For example
, this is may be necessary
in
3457 order to use certain NICs
in native mode
. Kernel allows the driver to
3458 create a separate set of XDP queues on top of regular ones
, and only
3459 these queues can be used
for AF_XDP sockets
. NICs that work
this way
3460 may also require an additional traffic redirection with ethtool to these
3465 # set number of queues to
1
3466 ethtool
-L eth0 combined
1
3467 # redirect all the traffic to the second
queue (id
: 1)
3468 # note
: drivers may require non
-empty key
/mask pair
.
3469 ethtool
-N eth0 flow
-type ether
\\
3470 dst
00:00:00:00:00:00 m FF
:FF
:FF
:FF
:FF
:FE action
1
3471 ethtool
-N eth0 flow
-type ether
\\
3472 dst
00:00:00:00:00:01 m FF
:FF
:FF
:FF
:FF
:FE action
1
3473 # launch QEMU instance
3474 |qemu_system| linux
.img
-device virtio
-net
-pci
,netdev
=n1
\\
3475 -netdev af
-xdp
,id
=n1
,ifname
=eth0
,queues
=1,start
-queue
=1
3477 XDP program can also be loaded externally
. In
this case 'inhibit' option
3478 should be set to
'on' and
'sock-fds' provided with file descriptors
for
3479 already open but not bound XDP sockets already added to a socket map
for
3480 corresponding queues
. One socket per queue
.
3484 |qemu_system| linux
.img
-device virtio
-net
-pci
,netdev
=n1
\\
3485 -netdev af
-xdp
,id
=n1
,ifname
=eth0
,queues
=3,inhibit
=on
,sock
-fds
=15:16:17
3487 ``
-netdev vhost
-user
,chardev
=id
[,vhostforce
=on|off
][,queues
=n
]``
3488 Establish a vhost
-user netdev
, backed by a chardev id
. The chardev
3489 should be a unix domain socket backed one
. The vhost
-user uses a
3490 specifically defined protocol to pass vhost ioctl replacement
3491 messages to an application on the other end of the socket
. On
3492 non
-MSIX guests
, the feature can be forced with vhostforce
. Use
3493 'queues=n' to specify the number of queues to be created
for
3494 multiqueue vhost
-user
.
3500 qemu
-m
512 -object memory
-backend
-file
,id
=mem
,size
=512M
,mem
-path
=/hugetlbfs
,share
=on \
3501 -numa node
,memdev
=mem \
3502 -chardev socket
,id
=chr0
,path
=/path
/to
/socket \
3503 -netdev type
=vhost
-user
,id
=net0
,chardev
=chr0 \
3504 -device virtio
-net
-pci
,netdev
=net0
3506 ``
-netdev vhost
-vdpa
[,vhostdev
=/path
/to
/dev
][,vhostfd
=h
]``
3507 Establish a vhost
-vdpa netdev
.
3509 vDPA device is a device that uses a datapath which complies with
3510 the virtio specifications with a vendor specific control path
.
3511 vDPA devices can be both physically located on the hardware or
3512 emulated by software
.
3514 ``
-netdev hubport
,id
=id
,hubid
=hubid
[,netdev
=nd
]``
3515 Create a hub port on the emulated hub with ID hubid
.
3517 The hubport netdev lets you connect a NIC to a QEMU emulated hub
3518 instead of a single netdev
. Alternatively
, you can also connect the
3519 hubport to another netdev with ID nd by
using the ``netdev
=nd``
3522 ``
-net nic
[,netdev
=nd
][,macaddr
=mac
][,model
=type
] [,name
=name
][,addr
=addr
][,vectors
=v
]``
3523 Legacy option to configure or create an on
-board (or machine
3524 default) Network Interface
Card(NIC
) and connect it either to the
3525 emulated hub with ID
0 (i
.e
. the
default hub
), or to the netdev nd
.
3526 If model is omitted
, then the
default NIC model associated with the
3527 machine type is used
. Note that the
default NIC model may change
in
3528 future QEMU releases
, so it is highly recommended to always specify
3529 a model
. Optionally
, the MAC address can be changed to mac
, the
3530 device address set to
addr (PCI cards only
), and a name can be
3531 assigned
for use
in monitor commands
. Optionally
, for PCI cards
, you
3532 can specify the number v of MSI
-X vectors that the card should have
;
3533 this option currently only affects virtio cards
; set v
= 0 to
3534 disable MSI
-X
. If no ``
-net`` option is specified
, a single NIC is
3535 created
. QEMU can emulate several different models of network card
.
3536 Use ``
-net nic
,model
=help``
for a list of available devices
for your
3539 ``
-net user|tap|bridge|socket|l2tpv3|vde
[,...][,name
=name
]``
3540 Configure a host network
backend (with the options corresponding to
3541 the same ``
-netdev`` option
) and connect it to the emulated hub
0
3542 (the
default hub
). Use name to specify the name of the hub port
.
3547 DEFHEADING(Character device options
:)
3549 DEF("chardev", HAS_ARG
, QEMU_OPTION_chardev
,
3551 "-chardev null,id=id[,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3552 "-chardev socket,id=id[,host=host],port=port[,to=to][,ipv4=on|off][,ipv6=on|off][,nodelay=on|off]\n"
3553 " [,server=on|off][,wait=on|off][,telnet=on|off][,websocket=on|off][,reconnect=seconds][,mux=on|off]\n"
3554 " [,logfile=PATH][,logappend=on|off][,tls-creds=ID][,tls-authz=ID] (tcp)\n"
3555 "-chardev socket,id=id,path=path[,server=on|off][,wait=on|off][,telnet=on|off][,websocket=on|off][,reconnect=seconds]\n"
3556 " [,mux=on|off][,logfile=PATH][,logappend=on|off][,abstract=on|off][,tight=on|off] (unix)\n"
3557 "-chardev udp,id=id[,host=host],port=port[,localaddr=localaddr]\n"
3558 " [,localport=localport][,ipv4=on|off][,ipv6=on|off][,mux=on|off]\n"
3559 " [,logfile=PATH][,logappend=on|off]\n"
3560 "-chardev msmouse,id=id[,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3561 "-chardev vc,id=id[[,width=width][,height=height]][[,cols=cols][,rows=rows]]\n"
3562 " [,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3563 "-chardev ringbuf,id=id[,size=size][,logfile=PATH][,logappend=on|off]\n"
3564 "-chardev file,id=id,path=path[,input-path=input-file][,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3565 "-chardev pipe,id=id,path=path[,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3567 "-chardev console,id=id[,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3568 "-chardev serial,id=id,path=path[,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3570 "-chardev pty,id=id[,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3571 "-chardev stdio,id=id[,mux=on|off][,signal=on|off][,logfile=PATH][,logappend=on|off]\n"
3573 #ifdef CONFIG_BRLAPI
3574 "-chardev braille,id=id[,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3576 #
if defined(__linux__
) ||
defined(__sun__
) ||
defined(__FreeBSD__
) \
3577 ||
defined(__NetBSD__
) ||
defined(__OpenBSD__
) ||
defined(__DragonFly__
)
3578 "-chardev serial,id=id,path=path[,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3580 #
if defined(__linux__
) ||
defined(__FreeBSD__
) ||
defined(__DragonFly__
)
3581 "-chardev parallel,id=id,path=path[,mux=on|off][,logfile=PATH][,logappend=on|off]\n"
3583 #
if defined(CONFIG_SPICE
)
3584 "-chardev spicevmc,id=id,name=name[,debug=debug][,logfile=PATH][,logappend=on|off]\n"
3585 "-chardev spiceport,id=id,name=name[,debug=debug][,logfile=PATH][,logappend=on|off]\n"
3591 The general form of a character device option is
:
3593 ``
-chardev backend
,id
=id
[,mux
=on|off
][,options
]``
3594 Backend is one of
: ``
null``
, ``socket``
, ``udp``
, ``msmouse``
,
3595 ``vc``
, ``ringbuf``
, ``file``
, ``pipe``
, ``console``
, ``serial``
,
3596 ``pty``
, ``stdio``
, ``braille``
, ``parallel``
,
3597 ``spicevmc``
, ``spiceport``
. The specific backend will determine the
3600 Use ``
-chardev help`` to print all available chardev backend types
.
3602 All devices must have an id
, which can be any string up to
127
3603 characters long
. It is used to uniquely identify
this device
in
3604 other command line directives
.
3606 A character device may be used
in multiplexing mode by multiple
3607 front
-ends
. Specify ``mux
=on`` to enable
this mode
. A multiplexer is
3608 a
"1:N" device
, and
here the
"1" end is your specified chardev
3609 backend
, and the
"N" end is the various parts of QEMU that can talk
3610 to a chardev
. If you create a chardev with ``id
=myid`` and
3611 ``mux
=on``
, QEMU will create a multiplexer with your specified ID
,
3612 and you can then configure multiple front ends to use that chardev
3613 ID
for their input
/output
. Up to four different front ends can be
3614 connected to a single multiplexed chardev
. (Without multiplexing
3615 enabled
, a chardev can only be used by a single front end
.) For
3616 instance you could use
this to allow a single stdio chardev to be
3617 used by two serial ports and the QEMU monitor
:
3621 -chardev stdio
,mux
=on
,id
=char0 \
3622 -mon chardev
=char0
,mode
=readline \
3623 -serial chardev
:char0 \
3624 -serial chardev
:char0
3626 You can have more than one multiplexer
in a system configuration
;
3627 for instance you could have a TCP port multiplexed between UART
0
3628 and UART
1, and stdio multiplexed between the QEMU monitor and a
3633 -chardev stdio
,mux
=on
,id
=char0 \
3634 -mon chardev
=char0
,mode
=readline \
3635 -parallel chardev
:char0 \
3636 -chardev tcp
,...,mux
=on
,id
=char1 \
3637 -serial chardev
:char1 \
3638 -serial chardev
:char1
3640 When you
're using a multiplexed character device, some escape
3641 sequences are interpreted in the input. See the chapter about
3642 :ref:`keys in the character backend multiplexer` in the
3643 System Emulation Users Guide for more details.
3645 Note that some other command line options may implicitly create
3646 multiplexed character backends; for instance ``-serial mon:stdio``
3647 creates a multiplexed stdio backend connected to the serial port and
3648 the QEMU monitor, and ``-nographic`` also multiplexes the console
3649 and the monitor to stdio.
3651 There is currently no support for multiplexing in the other
3652 direction (where a single QEMU front end takes input and output from
3655 Every backend supports the ``logfile`` option, which supplies the
3656 path to a file to record all data transmitted via the backend. The
3657 ``logappend`` option controls whether the log file will be truncated
3658 or appended to when opened.
3660 The available backends are:
3662 ``-chardev null,id=id``
3663 A void device. This device will not emit any data, and will drop any
3664 data it receives. The null backend does not take any options.
3666 ``-chardev socket,id=id[,TCP options or unix options][,server=on|off][,wait=on|off][,telnet=on|off][,websocket=on|off][,reconnect=seconds][,tls-creds=id][,tls-authz=id]``
3667 Create a two-way stream socket, which can be either a TCP or a unix
3668 socket. A unix socket will be created if ``path`` is specified.
3669 Behaviour is undefined if TCP options are specified for a unix
3672 ``server=on|off`` specifies that the socket shall be a listening socket.
3674 ``wait=on|off`` specifies that QEMU should not block waiting for a client
3675 to connect to a listening socket.
3677 ``telnet=on|off`` specifies that traffic on the socket should interpret
3678 telnet escape sequences.
3680 ``websocket=on|off`` specifies that the socket uses WebSocket protocol for
3683 ``reconnect`` sets the timeout for reconnecting on non-server
3684 sockets when the remote end goes away. qemu will delay this many
3685 seconds and then attempt to reconnect. Zero disables reconnecting,
3688 ``tls-creds`` requests enablement of the TLS protocol for
3689 encryption, and specifies the id of the TLS credentials to use for
3690 the handshake. The credentials must be previously created with the
3691 ``-object tls-creds`` argument.
3693 ``tls-auth`` provides the ID of the QAuthZ authorization object
3694 against which the client's x509 distinguished name will be
3695 validated
. This object is only resolved at time of use
, so can be
3696 deleted and recreated on the fly
while the chardev server is active
.
3697 If missing
, it will
default to denying access
.
3699 TCP and unix socket options are given below
:
3701 ``TCP options
: port
=port
[,host
=host
][,to
=to
][,ipv4
=on|off
][,ipv6
=on|off
][,nodelay
=on|off
]``
3702 ``host``
for a listening socket specifies the local address to
3703 be bound
. For a connecting socket species the remote host to
3704 connect to
. ``host`` is optional
for listening sockets
. If not
3705 specified it defaults to ``
0.0.0.0``
.
3707 ``port``
for a listening socket specifies the local port to be
3708 bound
. For a connecting socket specifies the port on the remote
3709 host to connect to
. ``port`` can be given as either a port
3710 number or a service name
. ``port`` is required
.
3712 ``to`` is only relevant to listening sockets
. If it is
3713 specified
, and ``port`` cannot be bound
, QEMU will attempt to
3714 bind to subsequent ports up to and including ``to`` until it
3715 succeeds
. ``to`` must be specified as a port number
.
3717 ``ipv4
=on|off`` and ``ipv6
=on|off`` specify that either IPv4
3718 or IPv6 must be used
. If neither is specified the socket may
3719 use either protocol
.
3721 ``nodelay
=on|off`` disables the Nagle algorithm
.
3723 ``unix options
: path
=path
[,abstract
=on|off
][,tight
=on|off
]``
3724 ``path`` specifies the local path of the unix socket
. ``path``
3726 ``abstract
=on|off`` specifies the use of the abstract socket namespace
,
3727 rather than the filesystem
. Optional
, defaults to
false.
3728 ``tight
=on|off`` sets the socket length of abstract sockets to their minimum
,
3729 rather than the full sun_path length
. Optional
, defaults to
true.
3731 ``
-chardev udp
,id
=id
[,host
=host
],port
=port
[,localaddr
=localaddr
][,localport
=localport
][,ipv4
=on|off
][,ipv6
=on|off
]``
3732 Sends all traffic from the guest to a remote host over UDP
.
3734 ``host`` specifies the remote host to connect to
. If not specified
3735 it defaults to ``localhost``
.
3737 ``port`` specifies the port on the remote host to connect to
.
3738 ``port`` is required
.
3740 ``localaddr`` specifies the local address to bind to
. If not
3741 specified it defaults to ``
0.0.0.0``
.
3743 ``localport`` specifies the local port to bind to
. If not specified
3744 any available local port will be used
.
3746 ``ipv4
=on|off`` and ``ipv6
=on|off`` specify that either IPv4 or IPv6 must be used
.
3747 If neither is specified the device may use either protocol
.
3749 ``
-chardev msmouse
,id
=id``
3750 Forward QEMU
's emulated msmouse events to the guest. ``msmouse``
3751 does not take any options.
3753 ``-chardev vc,id=id[[,width=width][,height=height]][[,cols=cols][,rows=rows]]``
3754 Connect to a QEMU text console. ``vc`` may optionally be given a
3757 ``width`` and ``height`` specify the width and height respectively
3758 of the console, in pixels.
3760 ``cols`` and ``rows`` specify that the console be sized to fit a
3761 text console with the given dimensions.
3763 ``-chardev ringbuf,id=id[,size=size]``
3764 Create a ring buffer with fixed size ``size``. size must be a power
3765 of two and defaults to ``64K``.
3767 ``-chardev file,id=id,path=path[,input-path=input-path]``
3768 Log all traffic received from the guest to a file.
3770 ``path`` specifies the path of the file to be opened. This file will
3771 be created if it does not already exist, and overwritten if it does.
3772 ``path`` is required.
3774 If ``input-path`` is specified, this is the path of a second file
3775 which will be used for input. If ``input-path`` is not specified,
3776 no input will be available from the chardev.
3778 Note that ``input-path`` is not supported on Windows hosts.
3780 ``-chardev pipe,id=id,path=path``
3781 Create a two-way connection to the guest. The behaviour differs
3782 slightly between Windows hosts and other hosts:
3784 On Windows, a single duplex pipe will be created at
3787 On other hosts, 2 pipes will be created called ``path.in`` and
3788 ``path.out``. Data written to ``path.in`` will be received by the
3789 guest. Data written by the guest can be read from ``path.out``. QEMU
3790 will not create these fifos, and requires them to be present.
3792 ``path`` forms part of the pipe path as described above. ``path`` is
3795 ``-chardev console,id=id``
3796 Send traffic from the guest to QEMU's standard output
. ``console``
3797 does not take any options
.
3799 ``console`` is only available on Windows hosts
.
3801 ``
-chardev serial
,id
=id
,path
=path``
3802 Send traffic from the guest to a serial device on the host
.
3804 On Unix hosts serial will actually accept any tty device
, not only
3807 ``path`` specifies the name of the serial device to open
.
3809 ``
-chardev pty
,id
=id``
3810 Create a
new pseudo
-terminal on the host and connect to it
. ``pty``
3811 does not take any options
.
3813 ``pty`` is not available on Windows hosts
.
3815 ``
-chardev stdio
,id
=id
[,signal
=on|off
]``
3816 Connect to standard input and standard output of the QEMU process
.
3818 ``signal`` controls
if signals are enabled on the terminal
, that
3819 includes exiting QEMU with the key sequence Control
-c
. This option
3820 is enabled by
default, use ``signal
=off`` to disable it
.
3822 ``
-chardev braille
,id
=id``
3823 Connect to a local BrlAPI server
. ``braille`` does not take any
3826 ``
-chardev parallel
,id
=id
,path
=path``
3828 ``parallel`` is only available on Linux
, FreeBSD and DragonFlyBSD
3831 Connect to a local parallel port
.
3833 ``path`` specifies the path to the parallel port device
. ``path`` is
3836 ``
-chardev spicevmc
,id
=id
,debug
=debug
,name
=name``
3837 ``spicevmc`` is only available when spice support is built
in.
3839 ``debug`` debug level
for spicevmc
3841 ``name`` name of spice channel to connect to
3843 Connect to a spice virtual machine channel
, such as vdiport
.
3845 ``
-chardev spiceport
,id
=id
,debug
=debug
,name
=name``
3846 ``spiceport`` is only available when spice support is built
in.
3848 ``debug`` debug level
for spicevmc
3850 ``name`` name of spice port to connect to
3852 Connect to a spice port
, allowing a Spice client to handle the
3853 traffic identified by a
name (preferably a fqdn
).
3859 DEFHEADING(TPM device options
:)
3861 DEF("tpmdev", HAS_ARG
, QEMU_OPTION_tpmdev
, \
3862 "-tpmdev passthrough,id=id[,path=path][,cancel-path=path]\n"
3863 " use path to provide path to a character device; default is /dev/tpm0\n"
3864 " use cancel-path to provide path to TPM's cancel sysfs entry; if\n"
3865 " not provided it will be searched for in /sys/class/misc/tpm?/device\n"
3866 "-tpmdev emulator,id=id,chardev=dev\n"
3867 " configure the TPM device using chardev backend\n",
3870 The general form of a TPM device option is
:
3872 ``
-tpmdev backend
,id
=id
[,options
]``
3873 The specific backend type will determine the applicable options
. The
3874 ``
-tpmdev`` option creates the TPM backend and requires a
3875 ``
-device`` option that specifies the TPM frontend
interface model
.
3877 Use ``
-tpmdev help`` to print all available TPM backend types
.
3879 The available backends are
:
3881 ``
-tpmdev passthrough
,id
=id
,path
=path
,cancel
-path
=cancel
-path``
3882 (Linux
-host only
) Enable access to the host
's TPM using the
3885 ``path`` specifies the path to the host's TPM device
, i
.e
., on a
3886 Linux host
this would be ``
/dev
/tpm0``
. ``path`` is optional and by
3887 default ``
/dev
/tpm0`` is used
.
3889 ``cancel
-path`` specifies the path to the host TPM device
's sysfs
3890 entry allowing for cancellation of an ongoing TPM command.
3891 ``cancel-path`` is optional and by default QEMU will search for the
3894 Some notes about using the host's TPM with the passthrough driver
:
3896 The TPM device accessed by the passthrough driver must not be used
3897 by any other application on the host
.
3899 Since the host
's firmware (BIOS/UEFI) has already initialized the
3900 TPM, the VM's
firmware (BIOS
/UEFI
) will not be able to initialize
3901 the TPM again and may therefore not show a TPM
-specific menu that
3902 would otherwise allow the user to configure the TPM
, e
.g
., allow the
3903 user to enable
/disable or activate
/deactivate the TPM
. Further
, if
3904 TPM ownership is released from within a VM then the host
's TPM will
3905 get disabled and deactivated. To enable and activate the TPM again
3906 afterwards, the host has to be rebooted and the user is required to
3907 enter the firmware's menu to enable and activate the TPM
. If the TPM
3908 is left disabled and
/or deactivated most TPM commands will fail
.
3910 To create a passthrough TPM use the following two options
:
3914 -tpmdev passthrough
,id
=tpm0
-device tpm
-tis
,tpmdev
=tpm0
3916 Note that the ``
-tpmdev`` id is ``tpm0`` and is referenced by
3917 ``tpmdev
=tpm0``
in the device option
.
3919 ``
-tpmdev emulator
,id
=id
,chardev
=dev``
3920 (Linux
-host only
) Enable access to a TPM emulator
using Unix domain
3921 socket based chardev backend
.
3923 ``chardev`` specifies the unique ID of a character device backend
3924 that provides connection to the software TPM server
.
3926 To create a TPM emulator backend device with chardev socket backend
:
3930 -chardev socket
,id
=chrtpm
,path
=/tmp
/swtpm
-sock
-tpmdev emulator
,id
=tpm0
,chardev
=chrtpm
-device tpm
-tis
,tpmdev
=tpm0
3937 DEFHEADING(Boot Image or Kernel specific
:)
3939 There are broadly
4 ways you can boot a system with QEMU
.
3941 - specify a firmware and let it control finding a kernel
3942 - specify a firmware and pass a hint to the kernel to boot
3943 - direct kernel image boot
3944 - manually load files into the guest
's address space
3946 The third method is useful for quickly testing kernels but as there is
3947 no firmware to pass configuration information to the kernel the
3948 hardware must either be probeable, the kernel built for the exact
3949 configuration or passed some configuration data (e.g. a DTB blob)
3950 which tells the kernel what drivers it needs. This exact details are
3951 often hardware specific.
3953 The final method is the most generic way of loading images into the
3954 guest address space and used mostly for ``bare metal`` type
3955 development where the reset vectors of the processor are taken into
3962 For x86 machines and some other architectures ``-bios`` will generally
3963 do the right thing with whatever it is given. For other machines the
3964 more strict ``-pflash`` option needs an image that is sized for the
3965 flash device for the given machine type.
3967 Please see the :ref:`system-targets-ref` section of the manual for
3968 more detailed documentation.
3972 DEF("bios", HAS_ARG, QEMU_OPTION_bios, \
3973 "-bios file set the filename for the BIOS\n", QEMU_ARCH_ALL)
3976 Set the filename for the BIOS.
3979 DEF("pflash", HAS_ARG, QEMU_OPTION_pflash,
3980 "-pflash file use 'file
' as a parallel flash image\n", QEMU_ARCH_ALL)
3983 Use file as a parallel flash image.
3988 The kernel options were designed to work with Linux kernels although
3989 other things (like hypervisors) can be packaged up as a kernel
3990 executable image. The exact format of a executable image is usually
3991 architecture specific.
3993 The way in which the kernel is started (what address it is loaded at,
3994 what if any information is passed to it via CPU registers, the state
3995 of the hardware when it is started, and so on) is also architecture
3996 specific. Typically it follows the specification laid down by the
3997 Linux kernel for how kernels for that architecture must be started.
4001 DEF("kernel", HAS_ARG, QEMU_OPTION_kernel, \
4002 "-kernel bzImage use 'bzImage
' as kernel image\n", QEMU_ARCH_ALL)
4005 Use bzImage as kernel image. The kernel can be either a Linux kernel
4006 or in multiboot format.
4009 DEF("append", HAS_ARG, QEMU_OPTION_append, \
4010 "-append cmdline use 'cmdline
' as kernel command line\n", QEMU_ARCH_ALL)
4013 Use cmdline as kernel command line
4016 DEF("initrd", HAS_ARG, QEMU_OPTION_initrd, \
4017 "-initrd file use 'file
' as initial ram disk\n", QEMU_ARCH_ALL)
4021 Use file as initial ram disk.
4023 ``-initrd "file1 arg=foo,file2"``
4024 This syntax is only available with multiboot.
4026 Use file1 and file2 as modules and pass ``arg=foo`` as parameter to the
4027 first module. Commas can be provided in module parameters by doubling
4028 them on the command line to escape them:
4030 ``-initrd "bzImage earlyprintk=xen,,keep root=/dev/xvda1,initrd.img"``
4031 Multiboot only. Use bzImage as the first module with
4032 "``earlyprintk=xen,keep root=/dev/xvda1``" as its command line,
4033 and initrd.img as the second module.
4037 DEF("dtb", HAS_ARG, QEMU_OPTION_dtb, \
4038 "-dtb file use 'file
' as device tree image\n", QEMU_ARCH_ALL)
4041 Use file as a device tree binary (dtb) image and pass it to the
4047 Finally you can also manually load images directly into the address
4048 space of the guest. This is most useful for developers who already
4049 know the layout of their guest and take care to ensure something sane
4050 will happen when the reset vector executes.
4052 The generic loader can be invoked by using the loader device:
4054 ``-device loader,addr=<addr>,data=<data>,data-len=<data-len>[,data-be=<data-be>][,cpu-num=<cpu-num>]``
4056 there is also the guest loader which operates in a similar way but
4057 tweaks the DTB so a hypervisor loaded via ``-kernel`` can find where
4060 ``-device guest-loader,addr=<addr>[,kernel=<path>,[bootargs=<arguments>]][,initrd=<path>]``
4066 DEFHEADING(Debug/Expert options:)
4068 DEF("compat", HAS_ARG, QEMU_OPTION_compat,
4069 "-compat [deprecated-input=accept|reject|crash][,deprecated-output=accept|hide]\n"
4070 " Policy for handling deprecated management interfaces\n"
4071 "-compat [unstable-input=accept|reject|crash][,unstable-output=accept|hide]\n"
4072 " Policy for handling unstable management interfaces\n",
4075 ``-compat [deprecated-input=@var{input-policy}][,deprecated-output=@var{output-policy}]``
4076 Set policy for handling deprecated management interfaces (experimental):
4078 ``deprecated-input=accept`` (default)
4079 Accept deprecated commands and arguments
4080 ``deprecated-input=reject``
4081 Reject deprecated commands and arguments
4082 ``deprecated-input=crash``
4083 Crash on deprecated commands and arguments
4084 ``deprecated-output=accept`` (default)
4085 Emit deprecated command results and events
4086 ``deprecated-output=hide``
4087 Suppress deprecated command results and events
4089 Limitation: covers only syntactic aspects of QMP.
4091 ``-compat [unstable-input=@var{input-policy}][,unstable-output=@var{output-policy}]``
4092 Set policy for handling unstable management interfaces (experimental):
4094 ``unstable-input=accept`` (default)
4095 Accept unstable commands and arguments
4096 ``unstable-input=reject``
4097 Reject unstable commands and arguments
4098 ``unstable-input=crash``
4099 Crash on unstable commands and arguments
4100 ``unstable-output=accept`` (default)
4101 Emit unstable command results and events
4102 ``unstable-output=hide``
4103 Suppress unstable command results and events
4105 Limitation: covers only syntactic aspects of QMP.
4108 DEF("fw_cfg", HAS_ARG, QEMU_OPTION_fwcfg,
4109 "-fw_cfg [name=]<name>,file=<file>\n"
4110 " add named fw_cfg entry with contents from file\n"
4111 "-fw_cfg [name=]<name>,string=<str>\n"
4112 " add named fw_cfg entry with contents from string\n",
4115 ``-fw_cfg [name=]name,file=file``
4116 Add named fw\_cfg entry with contents from file file.
4117 If the filename contains comma, you must double it (for instance,
4118 "file=my,,file" to use file "my,file").
4120 ``-fw_cfg [name=]name,string=str``
4121 Add named fw\_cfg entry with contents from string str.
4122 If the string contains comma, you must double it (for instance,
4123 "string=my,,string" to use file "my,string").
4125 The terminating NUL character of the contents of str will not be
4126 included as part of the fw\_cfg item data. To insert contents with
4127 embedded NUL characters, you have to use the file parameter.
4129 The fw\_cfg entries are passed by QEMU through to the guest.
4135 -fw_cfg name=opt/com.mycompany/blob,file=./my_blob.bin
4137 creates an fw\_cfg entry named opt/com.mycompany/blob with contents
4138 from ./my\_blob.bin.
4141 DEF("serial", HAS_ARG, QEMU_OPTION_serial, \
4142 "-serial dev redirect the serial port to char device 'dev
'\n",
4146 Redirect the virtual serial port to host character device dev. The
4147 default device is ``vc`` in graphical mode and ``stdio`` in non
4150 This option can be used several times to simulate multiple serial
4153 You can use ``-serial none`` to suppress the creation of default
4156 Available character devices are:
4159 Virtual console. Optionally, a width and height can be given in
4166 It is also possible to specify width or height in characters:
4173 [Linux only] Pseudo TTY (a new PTY is automatically allocated)
4176 No device is allocated. Note that for machine types which
4177 emulate systems where a serial device is always present in
4178 real hardware, this may be equivalent to the ``null`` option,
4179 in that the serial device is still present but all output
4180 is discarded. For boards where the number of serial ports is
4181 truly variable, this suppresses the creation of the device.
4184 A guest will see the UART or serial device as present in the
4185 machine, but all output is discarded, and there is no input.
4186 Conceptually equivalent to redirecting the output to ``/dev/null``.
4189 Use a named character device defined with the ``-chardev``
4193 [Linux only] Use host tty, e.g. ``/dev/ttyS0``. The host serial
4194 port parameters are set according to the emulated ones.
4197 [Linux only, parallel port only] Use host parallel port N.
4198 Currently SPP and EPP parallel port features can be used.
4201 Write output to filename. No character can be read.
4204 [Unix only] standard input/output
4210 [Windows only] Use host serial port n
4212 ``udp:[remote_host]:remote_port[@[src_ip]:src_port]``
4213 This implements UDP Net Console. When remote\_host or src\_ip
4214 are not specified they default to ``0.0.0.0``. When not using a
4215 specified src\_port a random port is automatically chosen.
4217 If you just want a simple readonly console you can use
4218 ``netcat`` or ``nc``, by starting QEMU with:
4219 ``-serial udp::4555`` and nc as: ``nc -u -l -p 4555``. Any time
4220 QEMU writes something to that port it will appear in the
4223 If you plan to send characters back via netconsole or you want
4224 to stop and start QEMU a lot of times, you should have QEMU use
4225 the same source port each time by using something like ``-serial
4226 udp::4555@:4556`` to QEMU. Another approach is to use a patched
4227 version of netcat which can listen to a TCP port and send and
4228 receive characters via udp. If you have a patched version of
4229 netcat which activates telnet remote echo and single char
4230 transfer, then you can use the following options to set up a
4231 netcat redirector to allow telnet on port 5555 to access the
4235 -serial udp::4555@:4556
4238 -u -P 4555 -L 0.0.0.0:4556 -t -p 5555 -I -T
4243 ``tcp:[host]:port[,server=on|off][,wait=on|off][,nodelay=on|off][,reconnect=seconds]``
4244 The TCP Net Console has two modes of operation. It can send the
4245 serial I/O to a location or wait for a connection from a
4246 location. By default the TCP Net Console is sent to host at the
4247 port. If you use the ``server=on`` option QEMU will wait for a client
4248 socket application to connect to the port before continuing,
4249 unless the ``wait=on|off`` option was specified. The ``nodelay=on|off``
4250 option disables the Nagle buffering algorithm. The ``reconnect=on``
4251 option only applies if ``server=no`` is set, if the connection goes
4252 down it will attempt to reconnect at the given interval. If host
4253 is omitted, 0.0.0.0 is assumed. Only one TCP connection at a
4254 time is accepted. You can use ``telnet=on`` to connect to the
4255 corresponding character device.
4257 ``Example to send tcp console to 192.168.0.2 port 4444``
4258 -serial tcp:192.168.0.2:4444
4260 ``Example to listen and wait on port 4444 for connection``
4261 -serial tcp::4444,server=on
4263 ``Example to not wait and listen on ip 192.168.0.100 port 4444``
4264 -serial tcp:192.168.0.100:4444,server=on,wait=off
4266 ``telnet:host:port[,server=on|off][,wait=on|off][,nodelay=on|off]``
4267 The telnet protocol is used instead of raw tcp sockets. The
4268 options work the same as if you had specified ``-serial tcp``.
4269 The difference is that the port acts like a telnet server or
4270 client using telnet option negotiation. This will also allow you
4271 to send the MAGIC\_SYSRQ sequence if you use a telnet that
4272 supports sending the break sequence. Typically in unix telnet
4273 you do it with Control-] and then type "send break" followed by
4274 pressing the enter key.
4276 ``websocket:host:port,server=on[,wait=on|off][,nodelay=on|off]``
4277 The WebSocket protocol is used instead of raw tcp socket. The
4278 port acts as a WebSocket server. Client mode is not supported.
4280 ``unix:path[,server=on|off][,wait=on|off][,reconnect=seconds]``
4281 A unix domain socket is used instead of a tcp socket. The option
4282 works the same as if you had specified ``-serial tcp`` except
4283 the unix domain socket path is used for connections.
4286 This is a special option to allow the monitor to be multiplexed
4287 onto another serial port. The monitor is accessed with key
4288 sequence of Control-a and then pressing c. dev\_string should be
4289 any one of the serial devices specified above. An example to
4290 multiplex the monitor onto a telnet server listening on port
4293 ``-serial mon:telnet::4444,server=on,wait=off``
4295 When the monitor is multiplexed to stdio in this way, Ctrl+C
4296 will not terminate QEMU any more but will be passed to the guest
4300 Braille device. This will use BrlAPI to display the braille
4301 output on a real or fake device.
4304 Three button serial mouse. Configure the guest to use Microsoft
4308 DEF("parallel", HAS_ARG, QEMU_OPTION_parallel, \
4309 "-parallel dev redirect the parallel port to char device 'dev
'\n",
4313 Redirect the virtual parallel port to host device dev (same devices
4314 as the serial port). On Linux hosts, ``/dev/parportN`` can be used
4315 to use hardware devices connected on the corresponding host parallel
4318 This option can be used several times to simulate up to 3 parallel
4321 Use ``-parallel none`` to disable all parallel ports.
4324 DEF("monitor", HAS_ARG, QEMU_OPTION_monitor, \
4325 "-monitor dev redirect the monitor to char device 'dev
'\n",
4329 Redirect the monitor to host device dev (same devices as the serial
4330 port). The default device is ``vc`` in graphical mode and ``stdio``
4331 in non graphical mode. Use ``-monitor none`` to disable the default
4334 DEF("qmp", HAS_ARG, QEMU_OPTION_qmp, \
4335 "-qmp dev like -monitor but opens in 'control
' mode\n",
4339 Like ``-monitor`` but opens in 'control
' mode. For example, to make
4340 QMP available on localhost port 4444::
4342 -qmp tcp:localhost:4444,server=on,wait=off
4344 Not all options are configurable via this syntax; for maximum
4345 flexibility use the ``-mon`` option and an accompanying ``-chardev``.
4348 DEF("qmp-pretty", HAS_ARG, QEMU_OPTION_qmp_pretty, \
4349 "-qmp-pretty dev like -qmp but uses pretty JSON formatting\n",
4353 Like ``-qmp`` but uses pretty JSON formatting.
4356 DEF("mon", HAS_ARG, QEMU_OPTION_mon, \
4357 "-mon [chardev=]name[,mode=readline|control][,pretty[=on|off]]\n", QEMU_ARCH_ALL)
4359 ``-mon [chardev=]name[,mode=readline|control][,pretty[=on|off]]``
4360 Set up a monitor connected to the chardev ``name``.
4361 QEMU supports two monitors: the Human Monitor Protocol
4362 (HMP; for human interaction), and the QEMU Monitor Protocol
4363 (QMP; a JSON RPC-style protocol).
4364 The default is HMP; ``mode=control`` selects QMP instead.
4365 ``pretty`` is only valid when ``mode=control``,
4366 turning on JSON pretty printing to ease
4367 human reading and debugging.
4371 -chardev socket,id=mon1,host=localhost,port=4444,server=on,wait=off \
4372 -mon chardev=mon1,mode=control,pretty=on
4374 enables the QMP monitor on localhost port 4444 with pretty-printing.
4377 DEF("debugcon", HAS_ARG, QEMU_OPTION_debugcon, \
4378 "-debugcon dev redirect the debug console to char device 'dev
'\n",
4382 Redirect the debug console to host device dev (same devices as the
4383 serial port). The debug console is an I/O port which is typically
4384 port 0xe9; writing to that I/O port sends output to this device. The
4385 default device is ``vc`` in graphical mode and ``stdio`` in non
4389 DEF("pidfile", HAS_ARG, QEMU_OPTION_pidfile, \
4390 "-pidfile file write PID to 'file
'\n", QEMU_ARCH_ALL)
4393 Store the QEMU process PID in file. It is useful if you launch QEMU
4397 DEF("preconfig", 0, QEMU_OPTION_preconfig, \
4398 "--preconfig pause QEMU before machine is initialized (experimental)\n",
4402 Pause QEMU for interactive configuration before the machine is
4403 created, which allows querying and configuring properties that will
4404 affect machine initialization. Use QMP command 'x
-exit
-preconfig
' to
4405 exit the preconfig state and move to the next state (i.e. run guest
4406 if -S isn't used or pause the second time
if -S is used
). This
4407 option is experimental
.
4410 DEF("S", 0, QEMU_OPTION_S
, \
4411 "-S freeze CPU at startup (use 'c' to start execution)\n",
4415 Do not start CPU at
startup (you must type
'c' in the monitor
).
4418 DEF("overcommit", HAS_ARG
, QEMU_OPTION_overcommit
,
4419 "-overcommit [mem-lock=on|off][cpu-pm=on|off]\n"
4420 " run qemu with overcommit hints\n"
4421 " mem-lock=on|off controls memory lock support (default: off)\n"
4422 " cpu-pm=on|off controls cpu power management (default: off)\n",
4425 ``
-overcommit mem
-lock
=on|off``
4427 ``
-overcommit cpu
-pm
=on|off``
4428 Run qemu with hints about host resource overcommit
. The
default is
4429 to assume that host overcommits all resources
.
4431 Locking qemu and guest memory can be enabled via ``mem
-lock
=on``
4432 (disabled by
default). This works when host memory is not
4433 overcommitted and reduces the worst
-case latency
for guest
.
4435 Guest ability to manage power state of host
cpus (increasing latency
4436 for other processes on the same host cpu
, but decreasing latency
for
4437 guest
) can be enabled via ``cpu
-pm
=on``
(disabled by
default). This
4438 works best when host CPU is not overcommitted
. When used
, host
4439 estimates of CPU cycle and power utilization will be incorrect
, not
4440 taking into account guest idle time
.
4443 DEF("gdb", HAS_ARG
, QEMU_OPTION_gdb
, \
4444 "-gdb dev accept gdb connection on 'dev'. (QEMU defaults to starting\n"
4445 " the guest without waiting for gdb to connect; use -S too\n"
4446 " if you want it to not start execution.)\n",
4450 Accept a gdb connection on device
dev (see the
:ref
:`GDB usage` chapter
4451 in the System Emulation Users Guide
). Note that
this option does not pause QEMU
4452 execution
-- if you want QEMU to not start the guest until you
4453 connect with gdb and issue a ``
continue`` command
, you will need to
4454 also pass the ``
-S`` option to QEMU
.
4456 The most usual configuration is to listen on a local TCP socket
::
4460 but you can specify other backends
; UDP
, pseudo TTY
, or even stdio
4461 are all reasonable use cases
. For example
, a stdio connection
4462 allows you to start QEMU from within gdb and establish the
4463 connection via a pipe
:
4467 (gdb
) target remote | exec |qemu_system|
-gdb stdio
...
4470 DEF("s", 0, QEMU_OPTION_s
, \
4471 "-s shorthand for -gdb tcp::" DEFAULT_GDBSTUB_PORT
"\n",
4475 Shorthand
for -gdb tcp
::1234, i
.e
. open a gdbserver on TCP port
1234
4476 (see the
:ref
:`GDB usage` chapter
in the System Emulation Users Guide
).
4479 DEF("d", HAS_ARG
, QEMU_OPTION_d
, \
4480 "-d item1,... enable logging of specified items (use '-d help' for a list of log items)\n",
4484 Enable logging of specified items
. Use
'-d help' for a list of log
4488 DEF("D", HAS_ARG
, QEMU_OPTION_D
, \
4489 "-D logfile output log to logfile (default stderr)\n",
4493 Output log
in logfile instead of to stderr
4496 DEF("dfilter", HAS_ARG
, QEMU_OPTION_DFILTER
, \
4497 "-dfilter range,.. filter debug output to range of addresses (useful for -d cpu,exec,etc..)\n",
4500 ``
-dfilter range1
[,...]``
4501 Filter debug output to that relevant to a range of target addresses
.
4502 The filter spec can be either start
+size
, start
-size or start
..end
4503 where start end and size are the addresses and sizes required
. For
4508 -dfilter
0x8000..0x8fff
,0xffffffc000080000+0x200,0xffffffc000060000-0x1000
4510 Will dump output
for any code
in the
0x1000 sized block starting at
4511 0x8000 and the
0x200 sized block starting at
0xffffffc000080000 and
4512 another
0x1000 sized block starting at
0xffffffc00005f000.
4515 DEF("seed", HAS_ARG
, QEMU_OPTION_seed
, \
4516 "-seed number seed the pseudo-random number generator\n",
4520 Force the guest to use a deterministic pseudo
-random number
4521 generator
, seeded with number
. This does not affect crypto routines
4525 DEF("L", HAS_ARG
, QEMU_OPTION_L
, \
4526 "-L path set the directory for the BIOS, VGA BIOS and keymaps\n",
4530 Set the directory
for the BIOS
, VGA BIOS and keymaps
.
4532 To list all the data directories
, use ``
-L help``
.
4535 DEF("enable-kvm", 0, QEMU_OPTION_enable_kvm
, \
4536 "-enable-kvm enable KVM full virtualization support\n",
4537 QEMU_ARCH_ARM | QEMU_ARCH_I386 | QEMU_ARCH_MIPS | QEMU_ARCH_PPC |
4538 QEMU_ARCH_RISCV | QEMU_ARCH_S390X
)
4541 Enable KVM full virtualization support
. This option is only
4542 available
if KVM support is enabled when compiling
.
4545 DEF("xen-domid", HAS_ARG
, QEMU_OPTION_xen_domid
,
4546 "-xen-domid id specify xen guest domain id\n",
4547 QEMU_ARCH_ARM | QEMU_ARCH_I386
)
4548 DEF("xen-attach", 0, QEMU_OPTION_xen_attach
,
4549 "-xen-attach attach to existing xen domain\n"
4550 " libxl will use this when starting QEMU\n",
4551 QEMU_ARCH_ARM | QEMU_ARCH_I386
)
4552 DEF("xen-domid-restrict", 0, QEMU_OPTION_xen_domid_restrict
,
4553 "-xen-domid-restrict restrict set of available xen operations\n"
4554 " to specified domain id. (Does not affect\n"
4555 " xenpv machine type).\n",
4556 QEMU_ARCH_ARM | QEMU_ARCH_I386
)
4559 Specify xen guest domain
id (XEN only
).
4562 Attach to existing xen domain
. libxl will use
this when starting
4563 QEMU (XEN only
). Restrict set of available xen operations to
4564 specified domain
id (XEN only
).
4567 DEF("no-reboot", 0, QEMU_OPTION_no_reboot
, \
4568 "-no-reboot exit instead of rebooting\n", QEMU_ARCH_ALL
)
4571 Exit instead of rebooting
.
4574 DEF("no-shutdown", 0, QEMU_OPTION_no_shutdown
, \
4575 "-no-shutdown stop before shutdown\n", QEMU_ARCH_ALL
)
4578 Don
't exit QEMU on guest shutdown, but instead only stop the
4579 emulation. This allows for instance switching to monitor to commit
4580 changes to the disk image.
4583 DEF("action", HAS_ARG, QEMU_OPTION_action,
4584 "-action reboot=reset|shutdown\n"
4585 " action when guest reboots [default=reset]\n"
4586 "-action shutdown=poweroff|pause\n"
4587 " action when guest shuts down [default=poweroff]\n"
4588 "-action panic=pause|shutdown|exit-failure|none\n"
4589 " action when guest panics [default=shutdown]\n"
4590 "-action watchdog=reset|shutdown|poweroff|inject-nmi|pause|debug|none\n"
4591 " action when watchdog fires [default=reset]\n",
4594 ``-action event=action``
4595 The action parameter serves to modify QEMU's
default behavior when
4596 certain guest events occur
. It provides a generic method
for specifying the
4597 same behaviors that are modified by the ``
-no
-reboot`` and ``
-no
-shutdown``
4602 ``
-action panic
=none``
4603 ``
-action reboot
=shutdown
,shutdown
=pause``
4604 ``
-device i6300esb
-action watchdog
=pause``
4608 DEF("loadvm", HAS_ARG
, QEMU_OPTION_loadvm
, \
4609 "-loadvm [tag|id]\n" \
4610 " start right away with a saved state (loadvm in monitor)\n",
4614 Start right away with a saved
state (``loadvm``
in monitor
)
4618 DEF("daemonize", 0, QEMU_OPTION_daemonize
, \
4619 "-daemonize daemonize QEMU after initializing\n", QEMU_ARCH_ALL
)
4623 Daemonize the QEMU process after initialization
. QEMU will not
4624 detach from standard IO until it is ready to receive connections on
4625 any of its devices
. This option is a useful way
for external
4626 programs to launch QEMU without having to cope with initialization
4630 DEF("option-rom", HAS_ARG
, QEMU_OPTION_option_rom
, \
4631 "-option-rom rom load a file, rom, into the option ROM space\n",
4634 ``
-option
-rom file``
4635 Load the contents of file as an option ROM
. This option is useful to
4636 load things like EtherBoot
.
4639 DEF("rtc", HAS_ARG
, QEMU_OPTION_rtc
, \
4640 "-rtc [base=utc|localtime|<datetime>][,clock=host|rt|vm][,driftfix=none|slew]\n" \
4641 " set the RTC base and clock, enable drift fix for clock ticks (x86 only)\n",
4645 ``
-rtc
[base
=utc|localtime|datetime
][,clock
=host|rt|vm
][,driftfix
=none|slew
]``
4646 Specify ``base`` as ``utc`` or ``localtime`` to let the RTC start at
4647 the current UTC or local time
, respectively
. ``localtime`` is
4648 required
for correct date
in MS
-DOS or Windows
. To start at a
4649 specific point
in time
, provide datetime
in the format
4650 ``
2006-06-17T16
:01:21`` or ``
2006-06-17``
. The
default base is UTC
.
4652 By
default the RTC is driven by the host system time
. This allows
4653 using of the RTC as accurate reference clock inside the guest
,
4654 specifically
if the host time is smoothly following an accurate
4655 external reference clock
, e
.g
. via NTP
. If you want to isolate the
4656 guest time from the host
, you can set ``clock`` to ``rt`` instead
,
4657 which provides a host monotonic clock
if host support it
. To even
4658 prevent the RTC from progressing during suspension
, you can set
4659 ``clock`` to ``vm``
(virtual clock
). '\ ``clock=vm``\ ' is
4660 recommended especially
in icount mode
in order to preserve
4661 determinism
; however
, note that
in icount mode the speed of the
4662 virtual clock is variable and can
in general differ from the host
4665 Enable ``driftfix``
(i386 targets only
) if you experience time drift
4666 problems
, specifically with Windows
' ACPI HAL. This option will try
4667 to figure out how many timer interrupts were not processed by the
4668 Windows guest and will re-inject them.
4671 DEF("icount", HAS_ARG, QEMU_OPTION_icount, \
4672 "-icount [shift=N|auto][,align=on|off][,sleep=on|off][,rr=record|replay,rrfile=<filename>[,rrsnapshot=<snapshot>]]\n" \
4673 " enable virtual instruction counter with 2^N clock ticks per\n" \
4674 " instruction, enable aligning the host and virtual clocks\n" \
4675 " or disable real time cpu sleeping, and optionally enable\n" \
4676 " record-and-replay mode\n", QEMU_ARCH_ALL)
4678 ``-icount [shift=N|auto][,align=on|off][,sleep=on|off][,rr=record|replay,rrfile=filename[,rrsnapshot=snapshot]]``
4679 Enable virtual instruction counter. The virtual cpu will execute one
4680 instruction every 2^N ns of virtual time. If ``auto`` is specified
4681 then the virtual cpu speed will be automatically adjusted to keep
4682 virtual time within a few seconds of real time.
4684 Note that while this option can give deterministic behavior, it does
4685 not provide cycle accurate emulation. Modern CPUs contain
4686 superscalar out of order cores with complex cache hierarchies. The
4687 number of instructions executed often has little or no correlation
4688 with actual performance.
4690 When the virtual cpu is sleeping, the virtual time will advance at
4691 default speed unless ``sleep=on`` is specified. With
4692 ``sleep=on``, the virtual time will jump to the next timer
4693 deadline instantly whenever the virtual cpu goes to sleep mode and
4694 will not advance if no timer is enabled. This behavior gives
4695 deterministic execution times from the guest point of view.
4696 The default if icount is enabled is ``sleep=off``.
4697 ``sleep=on`` cannot be used together with either ``shift=auto``
4700 ``align=on`` will activate the delay algorithm which will try to
4701 synchronise the host clock and the virtual clock. The goal is to
4702 have a guest running at the real frequency imposed by the shift
4703 option. Whenever the guest clock is behind the host clock and if
4704 ``align=on`` is specified then we print a message to the user to
4705 inform about the delay. Currently this option does not work when
4706 ``shift`` is ``auto``. Note: The sync algorithm will work for those
4707 shift values for which the guest clock runs ahead of the host clock.
4708 Typically this happens when the shift value is high (how high
4709 depends on the host machine). The default if icount is enabled
4712 When the ``rr`` option is specified deterministic record/replay is
4713 enabled. The ``rrfile=`` option must also be provided to
4714 specify the path to the replay log. In record mode data is written
4715 to this file, and in replay mode it is read back.
4716 If the ``rrsnapshot`` option is given then it specifies a VM snapshot
4717 name. In record mode, a new VM snapshot with the given name is created
4718 at the start of execution recording. In replay mode this option
4719 specifies the snapshot name used to load the initial VM state.
4722 DEF("watchdog-action", HAS_ARG, QEMU_OPTION_watchdog_action, \
4723 "-watchdog-action reset|shutdown|poweroff|inject-nmi|pause|debug|none\n" \
4724 " action when watchdog fires [default=reset]\n",
4727 ``-watchdog-action action``
4728 The action controls what QEMU will do when the watchdog timer
4729 expires. The default is ``reset`` (forcefully reset the guest).
4730 Other possible actions are: ``shutdown`` (attempt to gracefully
4731 shutdown the guest), ``poweroff`` (forcefully poweroff the guest),
4732 ``inject-nmi`` (inject a NMI into the guest), ``pause`` (pause the
4733 guest), ``debug`` (print a debug message and continue), or ``none``
4736 Note that the ``shutdown`` action requires that the guest responds
4737 to ACPI signals, which it may not be able to do in the sort of
4738 situations where the watchdog would have expired, and thus
4739 ``-watchdog-action shutdown`` is not recommended for production use.
4743 ``-device i6300esb -watchdog-action pause``
4747 DEF("echr", HAS_ARG, QEMU_OPTION_echr, \
4748 "-echr chr set terminal escape character instead of ctrl-a\n",
4751 ``-echr numeric_ascii_value``
4752 Change the escape character used for switching to the monitor when
4753 using monitor and serial sharing. The default is ``0x01`` when using
4754 the ``-nographic`` option. ``0x01`` is equal to pressing
4755 ``Control-a``. You can select a different character from the ascii
4756 control keys where 1 through 26 map to Control-a through Control-z.
4757 For instance you could use the either of the following to change the
4758 escape character to Control-t.
4760 ``-echr 0x14``; \ ``-echr 20``
4764 DEF("incoming", HAS_ARG, QEMU_OPTION_incoming, \
4765 "-incoming tcp:[host]:port[,to=maxport][,ipv4=on|off][,ipv6=on|off]\n" \
4766 "-incoming rdma:host:port[,ipv4=on|off][,ipv6=on|off]\n" \
4767 "-incoming unix:socketpath\n" \
4768 " prepare for incoming migration, listen on\n" \
4769 " specified protocol and socket address\n" \
4770 "-incoming fd:fd\n" \
4771 "-incoming file:filename[,offset=offset]\n" \
4772 "-incoming exec:cmdline\n" \
4773 " accept incoming migration on given file descriptor\n" \
4774 " or from given external command\n" \
4775 "-incoming defer\n" \
4776 " wait for the URI to be specified via migrate_incoming\n",
4779 ``-incoming tcp:[host]:port[,to=maxport][,ipv4=on|off][,ipv6=on|off]``
4781 ``-incoming rdma:host:port[,ipv4=on|off][,ipv6=on|off]``
4782 Prepare for incoming migration, listen on a given tcp port.
4784 ``-incoming unix:socketpath``
4785 Prepare for incoming migration, listen on a given unix socket.
4788 Accept incoming migration from a given file descriptor.
4790 ``-incoming file:filename[,offset=offset]``
4791 Accept incoming migration from a given file starting at offset.
4792 offset allows the common size suffixes, or a 0x prefix, but not both.
4794 ``-incoming exec:cmdline``
4795 Accept incoming migration as an output from specified external
4799 Wait for the URI to be specified via migrate\_incoming. The monitor
4800 can be used to change settings (such as migration parameters) prior
4801 to issuing the migrate\_incoming to allow the migration to begin.
4804 DEF("only-migratable", 0, QEMU_OPTION_only_migratable, \
4805 "-only-migratable allow only migratable devices\n", QEMU_ARCH_ALL)
4807 ``-only-migratable``
4808 Only allow migratable devices. Devices will not be allowed to enter
4809 an unmigratable state.
4812 DEF("nodefaults", 0, QEMU_OPTION_nodefaults, \
4813 "-nodefaults don't create
default devices
\n", QEMU_ARCH_ALL)
4816 Don't create default devices. Normally, QEMU sets the default
4817 devices like serial port, parallel port, virtual console, monitor
4818 device, VGA adapter, floppy and CD-ROM drive and others. The
4819 ``-nodefaults`` option will disable all those default devices.
4823 DEF("runas
", HAS_ARG, QEMU_OPTION_runas, \
4824 "-runas user change to user id user just before starting the VM
\n" \
4825 " user can be numeric uid
:gid instead
\n",
4830 Immediately before starting guest execution, drop root privileges,
4831 switching to the specified user. This option is deprecated, use
4832 ``-run-with user=...`` instead.
4835 DEF("prom
-env
", HAS_ARG, QEMU_OPTION_prom_env,
4836 "-prom
-env variable
=value
\n"
4837 " set OpenBIOS nvram variables
\n",
4838 QEMU_ARCH_PPC | QEMU_ARCH_SPARC)
4840 ``-prom-env variable=value``
4841 Set OpenBIOS nvram variable to given value (PPC, SPARC only).
4845 qemu-system-sparc -prom-env 'auto-boot?=false' \
4846 -prom-env 'boot-device=sd(0,2,0):d' -prom-env 'boot-args=linux single'
4850 qemu-system-ppc -prom-env 'auto-boot?=false' \
4851 -prom-env 'boot-device=hd:2,\yaboot' \
4852 -prom-env 'boot-args=conf=hd:2,\yaboot.conf'
4854 DEF("semihosting
", 0, QEMU_OPTION_semihosting,
4855 "-semihosting semihosting mode
\n",
4856 QEMU_ARCH_ARM | QEMU_ARCH_M68K | QEMU_ARCH_XTENSA |
4857 QEMU_ARCH_MIPS | QEMU_ARCH_RISCV)
4860 Enable :ref:`Semihosting` mode (ARM, M68K, Xtensa, MIPS, RISC-V only).
4863 Note that this allows guest direct access to the host filesystem, so
4864 should only be used with a trusted guest OS.
4866 See the -semihosting-config option documentation for further
4867 information about the facilities this enables.
4869 DEF("semihosting
-config
", HAS_ARG, QEMU_OPTION_semihosting_config,
4870 "-semihosting
-config
[enable
=on|off
][,target
=native|gdb|auto
][,chardev
=id
][,userspace
=on|off
][,arg
=str
[,...]]\n" \
4871 " semihosting configuration
\n",
4872 QEMU_ARCH_ARM | QEMU_ARCH_M68K | QEMU_ARCH_XTENSA |
4873 QEMU_ARCH_MIPS | QEMU_ARCH_RISCV)
4875 ``-semihosting-config [enable=on|off][,target=native|gdb|auto][,chardev=id][,userspace=on|off][,arg=str[,...]]``
4876 Enable and configure :ref:`Semihosting` (ARM, M68K, Xtensa, MIPS, RISC-V
4880 Note that this allows guest direct access to the host filesystem, so
4881 should only be used with a trusted guest OS.
4883 ``target=native|gdb|auto``
4884 Defines where the semihosting calls will be addressed, to QEMU
4885 (``native``) or to GDB (``gdb``). The default is ``auto``, which
4886 means ``gdb`` during debug sessions and ``native`` otherwise.
4889 Send the output to a chardev backend output for native or auto
4890 output when not in gdb
4892 ``userspace=on|off``
4893 Allows code running in guest userspace to access the semihosting
4894 interface. The default is that only privileged guest code can
4895 make semihosting calls. Note that setting ``userspace=on`` should
4896 only be used if all guest code is trusted (for example, in
4897 bare-metal test case code).
4899 ``arg=str1,arg=str2,...``
4900 Allows the user to pass input arguments, and can be used
4901 multiple times to build up a list. The old-style
4902 ``-kernel``/``-append`` method of passing a command line is
4903 still supported for backward compatibility. If both the
4904 ``--semihosting-config arg`` and the ``-kernel``/``-append`` are
4905 specified, the former is passed to semihosting as it always
4908 DEF("old
-param
", 0, QEMU_OPTION_old_param,
4909 "-old
-param old param mode
\n", QEMU_ARCH_ARM)
4912 Old param mode (ARM only).
4915 DEF("sandbox
", HAS_ARG, QEMU_OPTION_sandbox, \
4916 "-sandbox on
[,obsolete
=allow|deny
][,elevateprivileges
=allow|deny|children
]\n" \
4917 " [,spawn
=allow|deny
][,resourcecontrol
=allow|deny
]\n" \
4918 " Enable seccomp mode
2 system call
filter (default 'off').\n" \
4919 " use
'obsolete' to allow obsolete system calls that are provided
\n" \
4920 " by the kernel
, but typically no longer used by modern
\n" \
4921 " C library implementations
.\n" \
4922 " use
'elevateprivileges' to allow or deny the QEMU process ability
\n" \
4923 " to elevate privileges
using set
*uid|gid system calls
.\n" \
4924 " The value
'children' will deny set
*uid|gid system calls
for\n" \
4925 " main QEMU process but will allow forks and execves to run unprivileged
\n" \
4926 " use
'spawn' to avoid QEMU to spawn
new threads or processes by
\n" \
4927 " blocking
*fork and execve
\n" \
4928 " use
'resourcecontrol' to disable process affinity and schedular priority
\n",
4931 ``-sandbox arg[,obsolete=string][,elevateprivileges=string][,spawn=string][,resourcecontrol=string]``
4932 Enable Seccomp mode 2 system call filter. 'on' will enable syscall
4933 filtering and 'off' will disable it. The default is 'off'.
4936 Enable Obsolete system calls
4938 ``elevateprivileges=string``
4939 Disable set\*uid\|gid system calls
4942 Disable \*fork and execve
4944 ``resourcecontrol=string``
4945 Disable process affinity and schedular priority
4948 DEF("readconfig
", HAS_ARG, QEMU_OPTION_readconfig,
4949 "-readconfig
<file
>\n"
4950 " read config file
\n", QEMU_ARCH_ALL)
4952 ``-readconfig file``
4953 Read device configuration from file. This approach is useful when
4954 you want to spawn QEMU process with many command line options but
4955 you don't want to exceed the command line character limit.
4958 DEF("no
-user
-config
", 0, QEMU_OPTION_nouserconfig,
4960 " do not load
default user
-provided config files at startup
\n",
4964 The ``-no-user-config`` option makes QEMU not load any of the
4965 user-provided config files on sysconfdir.
4968 DEF("trace", HAS_ARG, QEMU_OPTION_trace,
4969 "-trace [[enable
=]<pattern
>][,events
=<file
>][,file
=<file
>]\n"
4970 " specify tracing options
\n",
4973 ``-trace [[enable=]pattern][,events=file][,file=file]``
4974 .. include:: ../qemu-option-trace.rst.inc
4977 DEF("plugin
", HAS_ARG, QEMU_OPTION_plugin,
4978 "-plugin
[file
=]<file
>[,<argname
>=<argvalue
>]\n"
4982 ``-plugin file=file[,argname=argvalue]``
4986 Load the given plugin from a shared library file.
4988 ``argname=argvalue``
4989 Argument passed to the plugin. (Can be given multiple times.)
4993 DEF("qtest
", HAS_ARG, QEMU_OPTION_qtest, "", QEMU_ARCH_ALL)
4994 DEF("qtest
-log
", HAS_ARG, QEMU_OPTION_qtest_log, "", QEMU_ARCH_ALL)
4997 DEF("run
-with
", HAS_ARG, QEMU_OPTION_run_with,
4998 "-run
-with
[async
-teardown
=on|off
][,chroot
=dir
][user
=username|uid
:gid
]\n"
4999 " Set miscellaneous QEMU process lifecycle options
:\n"
5000 " async
-teardown
=on enables asynchronous
teardown (Linux only
)\n"
5001 " chroot
=dir chroot to dir just before starting the VM
\n"
5002 " user
=username
switch to the specified user before starting the VM
\n"
5003 " user
=uid
:gid ditto
, but use specified user
-ID and group
-ID instead
\n",
5006 ``-run-with [async-teardown=on|off][,chroot=dir][user=username|uid:gid]``
5007 Set QEMU process lifecycle options.
5009 ``async-teardown=on`` enables asynchronous teardown. A new process called
5010 "cleanup
/<QEMU_PID
>" will be created at startup sharing the address
5011 space with the main QEMU process, using clone. It will wait for the
5012 main QEMU process to terminate completely, and then exit. This allows
5013 QEMU to terminate very quickly even if the guest was huge, leaving the
5014 teardown of the address space to the cleanup process. Since the cleanup
5015 process shares the same cgroups as the main QEMU process, accounting is
5016 performed correctly. This only works if the cleanup process is not
5017 forcefully killed with SIGKILL before the main QEMU process has
5018 terminated completely.
5020 ``chroot=dir`` can be used for doing a chroot to the specified directory
5021 immediately before starting the guest execution. This is especially useful
5022 in combination with -runas.
5024 ``user=username`` or ``user=uid:gid`` can be used to drop root privileges
5025 by switching to the specified user (via username) or user and group
5026 (via uid:gid) immediately before starting guest execution.
5030 DEF("msg
", HAS_ARG, QEMU_OPTION_msg,
5031 "-msg
[timestamp
[=on|off
]][,guest
-name
=[on|off
]]\n"
5032 " control error message format
\n"
5033 " timestamp
=on enables
timestamps (default: off
)\n"
5034 " guest
-name
=on enables guest name prefix but only
if\n"
5035 " -name guest option is
set (default: off
)\n",
5038 ``-msg [timestamp[=on|off]][,guest-name[=on|off]]``
5039 Control error message format.
5041 ``timestamp=on|off``
5042 Prefix messages with a timestamp. Default is off.
5044 ``guest-name=on|off``
5045 Prefix messages with guest name but only if -name guest option is set
5046 otherwise the option is ignored. Default is off.
5049 DEF("dump
-vmstate
", HAS_ARG, QEMU_OPTION_dump_vmstate,
5050 "-dump
-vmstate
<file
>\n"
5051 " Output vmstate information
in JSON format to file
.\n"
5052 " Use the scripts
/vmstate
-static-checker
.py file to
\n"
5053 " check
for possible regressions
in migration code
\n"
5054 " by comparing two such vmstate dumps
.\n",
5057 ``-dump-vmstate file``
5058 Dump json-encoded vmstate information for current machine type to
5062 DEF("enable
-sync
-profile
", 0, QEMU_OPTION_enable_sync_profile,
5063 "-enable
-sync
-profile
\n"
5064 " enable synchronization profiling
\n",
5067 ``-enable-sync-profile``
5068 Enable synchronization profiling.
5071 #if defined(CONFIG_TCG) && defined(CONFIG_LINUX)
5072 DEF("perfmap
", 0, QEMU_OPTION_perfmap,
5073 "-perfmap generate a
/tmp
/perf
-$
{pid
}.map file
for perf
\n",
5077 Generate a map file for Linux perf tools that will allow basic profiling
5078 information to be broken down into basic blocks.
5081 DEF("jitdump
", 0, QEMU_OPTION_jitdump,
5082 "-jitdump generate a jit
-$
{pid
}.dump file
for perf
\n",
5086 Generate a dump file for Linux perf tools that maps basic blocks to symbol
5087 names, line numbers and JITted code.
5093 DEFHEADING(Generic object creation:)
5095 DEF("object
", HAS_ARG, QEMU_OPTION_object,
5096 "-object TYPENAME
[,PROP1
=VALUE1
,...]\n"
5097 " create a
new object of type TYPENAME setting properties
\n"
5098 " in the order they are specified
. Note that the
'id'\n"
5099 " property must be set
. These objects are placed
in the
\n"
5100 " '/objects' path
.\n",
5103 ``-object typename[,prop1=value1,...]``
5104 Create a new object of type typename setting properties in the order
5105 they are specified. Note that the 'id' property must be set. These
5106 objects are placed in the '/objects' path.
5108 ``-object memory-backend-file,id=id,size=size,mem-path=dir,share=on|off,discard-data=on|off,merge=on|off,dump=on|off,prealloc=on|off,host-nodes=host-nodes,policy=default|preferred|bind|interleave,align=align,offset=offset,readonly=on|off,rom=on|off|auto``
5109 Creates a memory file backend object, which can be used to back
5110 the guest RAM with huge pages.
5112 The ``id`` parameter is a unique ID that will be used to
5113 reference this memory region in other parameters, e.g. ``-numa``,
5114 ``-device nvdimm``, etc.
5116 The ``size`` option provides the size of the memory region, and
5117 accepts common suffixes, e.g. ``500M``.
5119 The ``mem-path`` provides the path to either a shared memory or
5120 huge page filesystem mount.
5122 The ``share`` boolean option determines whether the memory
5123 region is marked as private to QEMU, or shared. The latter
5124 allows a co-operating external process to access the QEMU memory
5127 Setting share=on might affect the ability to configure NUMA
5128 bindings for the memory backend under some circumstances, see
5129 Documentation/vm/numa\_memory\_policy.txt on the Linux kernel
5130 source tree for additional details.
5132 Setting the ``discard-data`` boolean option to on indicates that
5133 file contents can be destroyed when QEMU exits, to avoid
5134 unnecessarily flushing data to the backing file. Note that
5135 ``discard-data`` is only an optimization, and QEMU might not
5136 discard file contents if it aborts unexpectedly or is terminated
5139 The ``merge`` boolean option enables memory merge, also known as
5140 MADV\_MERGEABLE, so that Kernel Samepage Merging will consider
5141 the pages for memory deduplication.
5143 Setting the ``dump`` boolean option to off excludes the memory
5144 from core dumps. This feature is also known as MADV\_DONTDUMP.
5146 The ``prealloc`` boolean option enables memory preallocation.
5148 The ``host-nodes`` option binds the memory range to a list of
5151 The ``policy`` option sets the NUMA policy to one of the
5158 prefer the given host node list for allocation
5161 restrict memory allocation to the given host node list
5164 interleave memory allocations across the given host node
5167 The ``align`` option specifies the base address alignment when
5168 QEMU mmap(2) ``mem-path``, and accepts common suffixes, eg
5169 ``2M``. Some backend store specified by ``mem-path`` requires an
5170 alignment different than the default one used by QEMU, eg the
5171 device DAX /dev/dax0.0 requires 2M alignment rather than 4K. In
5172 such cases, users can specify the required alignment via this
5175 The ``offset`` option specifies the offset into the target file
5176 that the region starts at. You can use this parameter to back
5177 multiple regions with a single file.
5179 The ``pmem`` option specifies whether the backing file specified
5180 by ``mem-path`` is in host persistent memory that can be
5181 accessed using the SNIA NVM programming model (e.g. Intel
5182 NVDIMM). If ``pmem`` is set to 'on', QEMU will take necessary
5183 operations to guarantee the persistence of its own writes to
5184 ``mem-path`` (e.g. in vNVDIMM label emulation and live
5185 migration). Also, we will map the backend-file with MAP\_SYNC
5186 flag, which ensures the file metadata is in sync for
5187 ``mem-path`` in case of host crash or a power failure. MAP\_SYNC
5188 requires support from both the host kernel (since Linux kernel
5189 4.15) and the filesystem of ``mem-path`` mounted with DAX
5192 The ``readonly`` option specifies whether the backing file is opened
5193 read-only or read-write (default).
5195 The ``rom`` option specifies whether to create Read Only Memory
5196 (ROM) that cannot be modified by the VM. Any write attempts to such
5197 ROM will be denied. Most use cases want proper RAM instead of ROM.
5198 However, selected use cases, like R/O NVDIMMs, can benefit from
5199 ROM. If set to ``on``, create ROM; if set to ``off``, create
5200 writable RAM; if set to ``auto`` (default), the value of the
5201 ``readonly`` option is used. This option is primarily helpful when
5202 we want to have writable RAM in configurations that would
5203 traditionally create ROM before the ``rom`` option was introduced:
5204 VM templating, where we want to open a file readonly
5205 (``readonly=on``) and mark the memory to be private for QEMU
5206 (``share=off``). For this use case, we need writable RAM instead
5207 of ROM, and want to also set ``rom=off``.
5209 ``-object memory-backend-ram,id=id,merge=on|off,dump=on|off,share=on|off,prealloc=on|off,size=size,host-nodes=host-nodes,policy=default|preferred|bind|interleave``
5210 Creates a memory backend object, which can be used to back the
5211 guest RAM. Memory backend objects offer more control than the
5212 ``-m`` option that is traditionally used to define guest RAM.
5213 Please refer to ``memory-backend-file`` for a description of the
5216 ``-object memory-backend-memfd,id=id,merge=on|off,dump=on|off,share=on|off,prealloc=on|off,size=size,host-nodes=host-nodes,policy=default|preferred|bind|interleave,seal=on|off,hugetlb=on|off,hugetlbsize=size``
5217 Creates an anonymous memory file backend object, which allows
5218 QEMU to share the memory with an external process (e.g. when
5219 using vhost-user). The memory is allocated with memfd and
5220 optional sealing. (Linux only)
5222 The ``seal`` option creates a sealed-file, that will block
5223 further resizing the memory ('on' by default).
5225 The ``hugetlb`` option specify the file to be created resides in
5226 the hugetlbfs filesystem (since Linux 4.14). Used in conjunction
5227 with the ``hugetlb`` option, the ``hugetlbsize`` option specify
5228 the hugetlb page size on systems that support multiple hugetlb
5229 page sizes (it must be a power of 2 value supported by the
5232 In some versions of Linux, the ``hugetlb`` option is
5233 incompatible with the ``seal`` option (requires at least Linux
5236 Please refer to ``memory-backend-file`` for a description of the
5239 The ``share`` boolean option is on by default with memfd.
5241 ``-object iommufd,id=id[,fd=fd]``
5242 Creates an iommufd backend which allows control of DMA mapping
5243 through the ``/dev/iommu`` device.
5245 The ``id`` parameter is a unique ID which frontends (such as
5246 vfio-pci of vdpa) will use to connect with the iommufd backend.
5248 The ``fd`` parameter is an optional pre-opened file descriptor
5249 resulting from ``/dev/iommu`` opening. Usually the iommufd is shared
5250 across all subsystems, bringing the benefit of centralized
5253 ``-object rng-builtin,id=id``
5254 Creates a random number generator backend which obtains entropy
5255 from QEMU builtin functions. The ``id`` parameter is a unique ID
5256 that will be used to reference this entropy backend from the
5257 ``virtio-rng`` device. By default, the ``virtio-rng`` device
5258 uses this RNG backend.
5260 ``-object rng-random,id=id,filename=/dev/random``
5261 Creates a random number generator backend which obtains entropy
5262 from a device on the host. The ``id`` parameter is a unique ID
5263 that will be used to reference this entropy backend from the
5264 ``virtio-rng`` device. The ``filename`` parameter specifies
5265 which file to obtain entropy from and if omitted defaults to
5268 ``-object rng-egd,id=id,chardev=chardevid``
5269 Creates a random number generator backend which obtains entropy
5270 from an external daemon running on the host. The ``id``
5271 parameter is a unique ID that will be used to reference this
5272 entropy backend from the ``virtio-rng`` device. The ``chardev``
5273 parameter is the unique ID of a character device backend that
5274 provides the connection to the RNG daemon.
5276 ``-object tls-creds-anon,id=id,endpoint=endpoint,dir=/path/to/cred/dir,verify-peer=on|off``
5277 Creates a TLS anonymous credentials object, which can be used to
5278 provide TLS support on network backends. The ``id`` parameter is
5279 a unique ID which network backends will use to access the
5280 credentials. The ``endpoint`` is either ``server`` or ``client``
5281 depending on whether the QEMU network backend that uses the
5282 credentials will be acting as a client or as a server. If
5283 ``verify-peer`` is enabled (the default) then once the handshake
5284 is completed, the peer credentials will be verified, though this
5285 is a no-op for anonymous credentials.
5287 The dir parameter tells QEMU where to find the credential files.
5288 For server endpoints, this directory may contain a file
5289 dh-params.pem providing diffie-hellman parameters to use for the
5290 TLS server. If the file is missing, QEMU will generate a set of
5291 DH parameters at startup. This is a computationally expensive
5292 operation that consumes random pool entropy, so it is
5293 recommended that a persistent set of parameters be generated
5296 ``-object tls-creds-psk,id=id,endpoint=endpoint,dir=/path/to/keys/dir[,username=username]``
5297 Creates a TLS Pre-Shared Keys (PSK) credentials object, which
5298 can be used to provide TLS support on network backends. The
5299 ``id`` parameter is a unique ID which network backends will use
5300 to access the credentials. The ``endpoint`` is either ``server``
5301 or ``client`` depending on whether the QEMU network backend that
5302 uses the credentials will be acting as a client or as a server.
5303 For clients only, ``username`` is the username which will be
5304 sent to the server. If omitted it defaults to "qemu
".
5306 The dir parameter tells QEMU where to find the keys file. It is
5307 called "dir
/keys
.psk
" and contains "username
:key
" pairs. This
5308 file can most easily be created using the GnuTLS ``psktool``
5311 For server endpoints, dir may also contain a file dh-params.pem
5312 providing diffie-hellman parameters to use for the TLS server.
5313 If the file is missing, QEMU will generate a set of DH
5314 parameters at startup. This is a computationally expensive
5315 operation that consumes random pool entropy, so it is
5316 recommended that a persistent set of parameters be generated up
5319 ``-object tls-creds-x509,id=id,endpoint=endpoint,dir=/path/to/cred/dir,priority=priority,verify-peer=on|off,passwordid=id``
5320 Creates a TLS anonymous credentials object, which can be used to
5321 provide TLS support on network backends. The ``id`` parameter is
5322 a unique ID which network backends will use to access the
5323 credentials. The ``endpoint`` is either ``server`` or ``client``
5324 depending on whether the QEMU network backend that uses the
5325 credentials will be acting as a client or as a server. If
5326 ``verify-peer`` is enabled (the default) then once the handshake
5327 is completed, the peer credentials will be verified. With x509
5328 certificates, this implies that the clients must be provided
5329 with valid client certificates too.
5331 The dir parameter tells QEMU where to find the credential files.
5332 For server endpoints, this directory may contain a file
5333 dh-params.pem providing diffie-hellman parameters to use for the
5334 TLS server. If the file is missing, QEMU will generate a set of
5335 DH parameters at startup. This is a computationally expensive
5336 operation that consumes random pool entropy, so it is
5337 recommended that a persistent set of parameters be generated
5340 For x509 certificate credentials the directory will contain
5341 further files providing the x509 certificates. The certificates
5342 must be stored in PEM format, in filenames ca-cert.pem,
5343 ca-crl.pem (optional), server-cert.pem (only servers),
5344 server-key.pem (only servers), client-cert.pem (only clients),
5345 and client-key.pem (only clients).
5347 For the server-key.pem and client-key.pem files which contain
5348 sensitive private keys, it is possible to use an encrypted
5349 version by providing the passwordid parameter. This provides the
5350 ID of a previously created ``secret`` object containing the
5351 password for decryption.
5353 The priority parameter allows to override the global default
5354 priority used by gnutls. This can be useful if the system
5355 administrator needs to use a weaker set of crypto priorities for
5356 QEMU without potentially forcing the weakness onto all
5357 applications. Or conversely if one wants wants a stronger
5358 default for QEMU than for all other applications, they can do
5359 this through this parameter. Its format is a gnutls priority
5360 string as described at
5361 https://gnutls.org/manual/html_node/Priority-Strings.html.
5363 ``-object tls-cipher-suites,id=id,priority=priority``
5364 Creates a TLS cipher suites object, which can be used to control
5365 the TLS cipher/protocol algorithms that applications are permitted
5368 The ``id`` parameter is a unique ID which frontends will use to
5369 access the ordered list of permitted TLS cipher suites from the
5372 The ``priority`` parameter allows to override the global default
5373 priority used by gnutls. This can be useful if the system
5374 administrator needs to use a weaker set of crypto priorities for
5375 QEMU without potentially forcing the weakness onto all
5376 applications. Or conversely if one wants wants a stronger
5377 default for QEMU than for all other applications, they can do
5378 this through this parameter. Its format is a gnutls priority
5379 string as described at
5380 https://gnutls.org/manual/html_node/Priority-Strings.html.
5382 An example of use of this object is to control UEFI HTTPS Boot.
5383 The tls-cipher-suites object exposes the ordered list of permitted
5384 TLS cipher suites from the host side to the guest firmware, via
5385 fw_cfg. The list is represented as an array of IANA_TLS_CIPHER
5386 objects. The firmware uses the IANA_TLS_CIPHER array for configuring
5389 In the following example, the priority at which the host-side policy
5390 is retrieved is given by the ``priority`` property.
5391 Given that QEMU uses GNUTLS, ``priority=@SYSTEM`` may be used to
5392 refer to /etc/crypto-policies/back-ends/gnutls.config.
5397 -object tls-cipher-suites,id=mysuite0,priority=@SYSTEM \\
5398 -fw_cfg name=etc/edk2/https/ciphers,gen_id=mysuite0
5400 ``-object filter-buffer,id=id,netdev=netdevid,interval=t[,queue=all|rx|tx][,status=on|off][,position=head|tail|id=<id>][,insert=behind|before]``
5401 Interval t can't be 0, this filter batches the packet delivery:
5402 all packets arriving in a given interval on netdev netdevid are
5403 delayed until the end of the interval. Interval is in
5404 microseconds. ``status`` is optional that indicate whether the
5405 netfilter is on (enabled) or off (disabled), the default status
5406 for netfilter will be 'on'.
5408 queue all\|rx\|tx is an option that can be applied to any
5411 ``all``: the filter is attached both to the receive and the
5412 transmit queue of the netdev (default).
5414 ``rx``: the filter is attached to the receive queue of the
5415 netdev, where it will receive packets sent to the netdev.
5417 ``tx``: the filter is attached to the transmit queue of the
5418 netdev, where it will receive packets sent by the netdev.
5420 position head\|tail\|id=<id> is an option to specify where the
5421 filter should be inserted in the filter list. It can be applied
5424 ``head``: the filter is inserted at the head of the filter list,
5425 before any existing filters.
5427 ``tail``: the filter is inserted at the tail of the filter list,
5428 behind any existing filters (default).
5430 ``id=<id>``: the filter is inserted before or behind the filter
5431 specified by <id>, see the insert option below.
5433 insert behind\|before is an option to specify where to insert
5434 the new filter relative to the one specified with
5435 position=id=<id>. It can be applied to any netfilter.
5437 ``before``: insert before the specified filter.
5439 ``behind``: insert behind the specified filter (default).
5441 ``-object filter-mirror,id=id,netdev=netdevid,outdev=chardevid,queue=all|rx|tx[,vnet_hdr_support][,position=head|tail|id=<id>][,insert=behind|before]``
5442 filter-mirror on netdev netdevid,mirror net packet to
5443 chardevchardevid, if it has the vnet\_hdr\_support flag,
5444 filter-mirror will mirror packet with vnet\_hdr\_len.
5446 ``-object filter-redirector,id=id,netdev=netdevid,indev=chardevid,outdev=chardevid,queue=all|rx|tx[,vnet_hdr_support][,position=head|tail|id=<id>][,insert=behind|before]``
5447 filter-redirector on netdev netdevid,redirect filter's net
5448 packet to chardev chardevid,and redirect indev's packet to
5449 filter.if it has the vnet\_hdr\_support flag, filter-redirector
5450 will redirect packet with vnet\_hdr\_len. Create a
5451 filter-redirector we need to differ outdev id from indev id, id
5452 can not be the same. we can just use indev or outdev, but at
5453 least one of indev or outdev need to be specified.
5455 ``-object filter-rewriter,id=id,netdev=netdevid,queue=all|rx|tx,[vnet_hdr_support][,position=head|tail|id=<id>][,insert=behind|before]``
5456 Filter-rewriter is a part of COLO project.It will rewrite tcp
5457 packet to secondary from primary to keep secondary tcp
5458 connection,and rewrite tcp packet to primary from secondary make
5459 tcp packet can be handled by client.if it has the
5460 vnet\_hdr\_support flag, we can parse packet with vnet header.
5462 usage: colo secondary: -object
5463 filter-redirector,id=f1,netdev=hn0,queue=tx,indev=red0 -object
5464 filter-redirector,id=f2,netdev=hn0,queue=rx,outdev=red1 -object
5465 filter-rewriter,id=rew0,netdev=hn0,queue=all
5467 ``-object filter-dump,id=id,netdev=dev[,file=filename][,maxlen=len][,position=head|tail|id=<id>][,insert=behind|before]``
5468 Dump the network traffic on netdev dev to the file specified by
5469 filename. At most len bytes (64k by default) per packet are
5470 stored. The file format is libpcap, so it can be analyzed with
5471 tools such as tcpdump or Wireshark.
5473 ``-object colo-compare,id=id,primary_in=chardevid,secondary_in=chardevid,outdev=chardevid,iothread=id[,vnet_hdr_support][,notify_dev=id][,compare_timeout=@var{ms}][,expired_scan_cycle=@var{ms}][,max_queue_size=@var{size}]``
5474 Colo-compare gets packet from primary\_in chardevid and
5475 secondary\_in, then compare whether the payload of primary packet
5476 and secondary packet are the same. If same, it will output
5477 primary packet to out\_dev, else it will notify COLO-framework to do
5478 checkpoint and send primary packet to out\_dev. In order to
5479 improve efficiency, we need to put the task of comparison in
5480 another iothread. If it has the vnet\_hdr\_support flag,
5481 colo compare will send/recv packet with vnet\_hdr\_len.
5482 The compare\_timeout=@var{ms} determines the maximum time of the
5483 colo-compare hold the packet. The expired\_scan\_cycle=@var{ms}
5484 is to set the period of scanning expired primary node network packets.
5485 The max\_queue\_size=@var{size} is to set the max compare queue
5486 size depend on user environment.
5487 If user want to use Xen COLO, need to add the notify\_dev to
5488 notify Xen colo-frame to do checkpoint.
5490 COLO-compare must be used with the help of filter-mirror,
5491 filter-redirector and filter-rewriter.
5498 -netdev tap,id=hn0,vhost=off
5499 -device e1000,id=e0,netdev=hn0,mac=52:a4:00:12:78:66
5500 -chardev socket,id=mirror0,host=3.3.3.3,port=9003,server=on,wait=off
5501 -chardev socket,id=compare1,host=3.3.3.3,port=9004,server=on,wait=off
5502 -chardev socket,id=compare0,host=3.3.3.3,port=9001,server=on,wait=off
5503 -chardev socket,id=compare0-0,host=3.3.3.3,port=9001
5504 -chardev socket,id=compare_out,host=3.3.3.3,port=9005,server=on,wait=off
5505 -chardev socket,id=compare_out0,host=3.3.3.3,port=9005
5506 -object iothread,id=iothread1
5507 -object filter-mirror,id=m0,netdev=hn0,queue=tx,outdev=mirror0
5508 -object filter-redirector,netdev=hn0,id=redire0,queue=rx,indev=compare_out
5509 -object filter-redirector,netdev=hn0,id=redire1,queue=rx,outdev=compare0
5510 -object colo-compare,id=comp0,primary_in=compare0-0,secondary_in=compare1,outdev=compare_out0,iothread=iothread1
5513 -netdev tap,id=hn0,vhost=off
5514 -device e1000,netdev=hn0,mac=52:a4:00:12:78:66
5515 -chardev socket,id=red0,host=3.3.3.3,port=9003
5516 -chardev socket,id=red1,host=3.3.3.3,port=9004
5517 -object filter-redirector,id=f1,netdev=hn0,queue=tx,indev=red0
5518 -object filter-redirector,id=f2,netdev=hn0,queue=rx,outdev=red1
5524 -netdev tap,id=hn0,vhost=off
5525 -device e1000,id=e0,netdev=hn0,mac=52:a4:00:12:78:66
5526 -chardev socket,id=mirror0,host=3.3.3.3,port=9003,server=on,wait=off
5527 -chardev socket,id=compare1,host=3.3.3.3,port=9004,server=on,wait=off
5528 -chardev socket,id=compare0,host=3.3.3.3,port=9001,server=on,wait=off
5529 -chardev socket,id=compare0-0,host=3.3.3.3,port=9001
5530 -chardev socket,id=compare_out,host=3.3.3.3,port=9005,server=on,wait=off
5531 -chardev socket,id=compare_out0,host=3.3.3.3,port=9005
5532 -chardev socket,id=notify_way,host=3.3.3.3,port=9009,server=on,wait=off
5533 -object filter-mirror,id=m0,netdev=hn0,queue=tx,outdev=mirror0
5534 -object filter-redirector,netdev=hn0,id=redire0,queue=rx,indev=compare_out
5535 -object filter-redirector,netdev=hn0,id=redire1,queue=rx,outdev=compare0
5536 -object iothread,id=iothread1
5537 -object colo-compare,id=comp0,primary_in=compare0-0,secondary_in=compare1,outdev=compare_out0,notify_dev=notify_way,iothread=iothread1
5540 -netdev tap,id=hn0,vhost=off
5541 -device e1000,netdev=hn0,mac=52:a4:00:12:78:66
5542 -chardev socket,id=red0,host=3.3.3.3,port=9003
5543 -chardev socket,id=red1,host=3.3.3.3,port=9004
5544 -object filter-redirector,id=f1,netdev=hn0,queue=tx,indev=red0
5545 -object filter-redirector,id=f2,netdev=hn0,queue=rx,outdev=red1
5547 If you want to know the detail of above command line, you can
5548 read the colo-compare git log.
5550 ``-object cryptodev-backend-builtin,id=id[,queues=queues]``
5551 Creates a cryptodev backend which executes crypto operations from
5552 the QEMU cipher APIs. The id parameter is a unique ID that will
5553 be used to reference this cryptodev backend from the
5554 ``virtio-crypto`` device. The queues parameter is optional,
5555 which specify the queue number of cryptodev backend, the default
5562 -object cryptodev-backend-builtin,id=cryptodev0 \\
5563 -device virtio-crypto-pci,id=crypto0,cryptodev=cryptodev0 \\
5566 ``-object cryptodev-vhost-user,id=id,chardev=chardevid[,queues=queues]``
5567 Creates a vhost-user cryptodev backend, backed by a chardev
5568 chardevid. The id parameter is a unique ID that will be used to
5569 reference this cryptodev backend from the ``virtio-crypto``
5570 device. The chardev should be a unix domain socket backed one.
5571 The vhost-user uses a specifically defined protocol to pass
5572 vhost ioctl replacement messages to an application on the other
5573 end of the socket. The queues parameter is optional, which
5574 specify the queue number of cryptodev backend for multiqueue
5575 vhost-user, the default of queues is 1.
5581 -chardev socket,id=chardev0,path=/path/to/socket \\
5582 -object cryptodev-vhost-user,id=cryptodev0,chardev=chardev0 \\
5583 -device virtio-crypto-pci,id=crypto0,cryptodev=cryptodev0 \\
5586 ``-object secret,id=id,data=string,format=raw|base64[,keyid=secretid,iv=string]``
5588 ``-object secret,id=id,file=filename,format=raw|base64[,keyid=secretid,iv=string]``
5589 Defines a secret to store a password, encryption key, or some
5590 other sensitive data. The sensitive data can either be passed
5591 directly via the data parameter, or indirectly via the file
5592 parameter. Using the data parameter is insecure unless the
5593 sensitive data is encrypted.
5595 The sensitive data can be provided in raw format (the default),
5596 or base64. When encoded as JSON, the raw format only supports
5597 valid UTF-8 characters, so base64 is recommended for sending
5598 binary data. QEMU will convert from which ever format is
5599 provided to the format it needs internally. eg, an RBD password
5600 can be provided in raw format, even though it will be base64
5601 encoded when passed onto the RBD sever.
5603 For added protection, it is possible to encrypt the data
5604 associated with a secret using the AES-256-CBC cipher. Use of
5605 encryption is indicated by providing the keyid and iv
5606 parameters. The keyid parameter provides the ID of a previously
5607 defined secret that contains the AES-256 decryption key. This
5608 key should be 32-bytes long and be base64 encoded. The iv
5609 parameter provides the random initialization vector used for
5610 encryption of this particular secret and should be a base64
5611 encrypted string of the 16-byte IV.
5613 The simplest (insecure) usage is to provide the secret inline
5617 # |qemu_system| -object secret,id=sec0,data=letmein,format=raw
5619 The simplest secure usage is to provide the secret via a file
5621 # printf "letmein
" > mypasswd.txt # QEMU\_SYSTEM\_MACRO -object
5622 secret,id=sec0,file=mypasswd.txt,format=raw
5624 For greater security, AES-256-CBC should be used. To illustrate
5625 usage, consider the openssl command line tool which can encrypt
5626 the data. Note that when encrypting, the plaintext must be
5627 padded to the cipher block size (32 bytes) using the standard
5628 PKCS#5/6 compatible padding algorithm.
5630 First a master key needs to be created in base64 encoding:
5634 # openssl rand -base64 32 > key.b64
5635 # KEY=$(base64 -d key.b64 | hexdump -v -e '/1 "%02X
"')
5637 Each secret to be encrypted needs to have a random
5638 initialization vector generated. These do not need to be kept
5643 # openssl rand -base64 16 > iv.b64
5644 # IV=$(base64 -d iv.b64 | hexdump -v -e '/1 "%02X
"')
5646 The secret to be defined can now be encrypted, in this case
5647 we're telling openssl to base64 encode the result, but it could
5648 be left as raw bytes if desired.
5652 # SECRET=$(printf "letmein
" |
5653 openssl enc -aes-256-cbc -a -K $KEY -iv $IV)
5655 When launching QEMU, create a master secret pointing to
5656 ``key.b64`` and specify that to be used to decrypt the user
5657 password. Pass the contents of ``iv.b64`` to the second secret
5662 -object secret,id=secmaster0,format=base64,file=key.b64 \\
5663 -object secret,id=sec0,keyid=secmaster0,format=base64,\\
5664 data=$SECRET,iv=$(<iv.b64)
5666 ``-object sev-guest,id=id,cbitpos=cbitpos,reduced-phys-bits=val,[sev-device=string,policy=policy,handle=handle,dh-cert-file=file,session-file=file,kernel-hashes=on|off]``
5667 Create a Secure Encrypted Virtualization (SEV) guest object,
5668 which can be used to provide the guest memory encryption support
5671 When memory encryption is enabled, one of the physical address
5672 bit (aka the C-bit) is utilized to mark if a memory page is
5673 protected. The ``cbitpos`` is used to provide the C-bit
5674 position. The C-bit position is Host family dependent hence user
5675 must provide this value. On EPYC, the value should be 47.
5677 When memory encryption is enabled, we loose certain bits in
5678 physical address space. The ``reduced-phys-bits`` is used to
5679 provide the number of bits we loose in physical address space.
5680 Similar to C-bit, the value is Host family dependent. On EPYC,
5681 a guest will lose a maximum of 1 bit, so the value should be 1.
5683 The ``sev-device`` provides the device file to use for
5684 communicating with the SEV firmware running inside AMD Secure
5685 Processor. The default device is '/dev/sev'. If hardware
5686 supports memory encryption then /dev/sev devices are created by
5689 The ``policy`` provides the guest policy to be enforced by the
5690 SEV firmware and restrict what configuration and operational
5691 commands can be performed on this guest by the hypervisor. The
5692 policy should be provided by the guest owner and is bound to the
5693 guest and cannot be changed throughout the lifetime of the
5694 guest. The default is 0.
5696 If guest ``policy`` allows sharing the key with another SEV
5697 guest then ``handle`` can be use to provide handle of the guest
5698 from which to share the key.
5700 The ``dh-cert-file`` and ``session-file`` provides the guest
5701 owner's Public Diffie-Hillman key defined in SEV spec. The PDH
5702 and session parameters are used for establishing a cryptographic
5703 session with the guest owner to negotiate keys used for
5704 attestation. The file must be encoded in base64.
5706 The ``kernel-hashes`` adds the hashes of given kernel/initrd/
5707 cmdline to a designated guest firmware page for measured Linux
5708 boot with -kernel. The default is off. (Since 6.2)
5710 e.g to launch a SEV guest
5714 # |qemu_system_x86| \\
5716 -object sev-guest,id=sev0,cbitpos=47,reduced-phys-bits=1 \\
5717 -machine ...,memory-encryption=sev0 \\
5720 ``-object authz-simple,id=id,identity=string``
5721 Create an authorization object that will control access to
5724 The ``identity`` parameter is identifies the user and its format
5725 depends on the network service that authorization object is
5726 associated with. For authorizing based on TLS x509 certificates,
5727 the identity must be the x509 distinguished name. Note that care
5728 must be taken to escape any commas in the distinguished name.
5730 An example authorization object to validate a x509 distinguished
5731 name would look like:
5737 -object 'authz-simple,id=auth0,identity=CN=laptop.example.com,,O=Example Org,,L=London,,ST=London,,C=GB' \\
5740 Note the use of quotes due to the x509 distinguished name
5741 containing whitespace, and escaping of ','.
5743 ``-object authz-listfile,id=id,filename=path,refresh=on|off``
5744 Create an authorization object that will control access to
5747 The ``filename`` parameter is the fully qualified path to a file
5748 containing the access control list rules in JSON format.
5750 An example set of rules that match against SASL usernames might
5757 { "match
": "fred
", "policy
": "allow
", "format
": "exact
" },
5758 { "match
": "bob
", "policy
": "allow
", "format
": "exact
" },
5759 { "match
": "danb
", "policy
": "deny
", "format
": "glob
" },
5760 { "match
": "dan
*", "policy
": "allow
", "format
": "exact
" },
5765 When checking access the object will iterate over all the rules
5766 and the first rule to match will have its ``policy`` value
5767 returned as the result. If no rules match, then the default
5768 ``policy`` value is returned.
5770 The rules can either be an exact string match, or they can use
5771 the simple UNIX glob pattern matching to allow wildcards to be
5774 If ``refresh`` is set to true the file will be monitored and
5775 automatically reloaded whenever its content changes.
5777 As with the ``authz-simple`` object, the format of the identity
5778 strings being matched depends on the network service, but is
5779 usually a TLS x509 distinguished name, or a SASL username.
5781 An example authorization object to validate a SASL username
5788 -object authz-simple,id=auth0,filename=/etc/qemu/vnc-sasl.acl,refresh=on \\
5791 ``-object authz-pam,id=id,service=string``
5792 Create an authorization object that will control access to
5795 The ``service`` parameter provides the name of a PAM service to
5796 use for authorization. It requires that a file
5797 ``/etc/pam.d/service`` exist to provide the configuration for
5798 the ``account`` subsystem.
5800 An example authorization object to validate a TLS x509
5801 distinguished name would look like:
5807 -object authz-pam,id=auth0,service=qemu-vnc \\
5810 There would then be a corresponding config file for PAM at
5811 ``/etc/pam.d/qemu-vnc`` that contains:
5815 account requisite pam_listfile.so item=user sense=allow \
5816 file=/etc/qemu/vnc.allow
5818 Finally the ``/etc/qemu/vnc.allow`` file would contain the list
5819 of x509 distinguished names that are permitted access
5823 CN=laptop.example.com,O=Example Home,L=London,ST=London,C=GB
5825 ``-object iothread,id=id,poll-max-ns=poll-max-ns,poll-grow=poll-grow,poll-shrink=poll-shrink,aio-max-batch=aio-max-batch``
5826 Creates a dedicated event loop thread that devices can be
5827 assigned to. This is known as an IOThread. By default device
5828 emulation happens in vCPU threads or the main event loop thread.
5829 This can become a scalability bottleneck. IOThreads allow device
5830 emulation and I/O to run on other host CPUs.
5832 The ``id`` parameter is a unique ID that will be used to
5833 reference this IOThread from ``-device ...,iothread=id``.
5834 Multiple devices can be assigned to an IOThread. Note that not
5835 all devices support an ``iothread`` parameter.
5837 The ``query-iothreads`` QMP command lists IOThreads and reports
5838 their thread IDs so that the user can configure host CPU
5841 IOThreads use an adaptive polling algorithm to reduce event loop
5842 latency. Instead of entering a blocking system call to monitor
5843 file descriptors and then pay the cost of being woken up when an
5844 event occurs, the polling algorithm spins waiting for events for
5845 a short time. The algorithm's default parameters are suitable
5846 for many cases but can be adjusted based on knowledge of the
5847 workload and/or host device latency.
5849 The ``poll-max-ns`` parameter is the maximum number of
5850 nanoseconds to busy wait for events. Polling can be disabled by
5851 setting this value to 0.
5853 The ``poll-grow`` parameter is the multiplier used to increase
5854 the polling time when the algorithm detects it is missing events
5855 due to not polling long enough.
5857 The ``poll-shrink`` parameter is the divisor used to decrease
5858 the polling time when the algorithm detects it is spending too
5859 long polling without encountering events.
5861 The ``aio-max-batch`` parameter is the maximum number of requests
5862 in a batch for the AIO engine, 0 means that the engine will use
5865 The IOThread parameters can be modified at run-time using the
5866 ``qom-set`` command (where ``iothread1`` is the IOThread's
5871 (qemu) qom-set /objects/iothread1 poll-max-ns 100000
5875 HXCOMM This is the last statement. Insert new options before this line!