1QEMU(1) QEMU QEMU(1)
2
6 qemu - QEMU User Documentation
7
9 qemu-system-x86_64 [options] [disk_image]
10
12 The QEMU PC System emulator simulates the following peripherals:
13 • i440FX host PCI bridge and PIIX3 PCI to ISA bridge
15 • Cirrus CLGD 5446 PCI VGA card or dummy VGA card with Bochs VESA ex‐
17 tensions (hardware level, including all non standard modes).
18 • PS/2 mouse and keyboard
20 • 2 PCI IDE interfaces with hard disk and CD-ROM support
22 • Floppy disk
24 • PCI and ISA network adapters
26 • Serial ports
28 • IPMI BMC, either and internal or external one
30 • Creative SoundBlaster 16 sound card
32 • ENSONIQ AudioPCI ES1370 sound card
34 • Intel 82801AA AC97 Audio compatible sound card
36 • Intel HD Audio Controller and HDA codec
38 • Adlib (OPL2) - Yamaha YM3812 compatible chip
40 • Gravis Ultrasound GF1 sound card
42 • CS4231A compatible sound card
44 • PC speaker
46 • PCI UHCI, OHCI, EHCI or XHCI USB controller and a virtual USB-1.1
48 hub.
49 SMP is supported with up to 255 CPUs.
51 QEMU uses the PC BIOS from the Seabios project and the Plex86/Bochs
53 LGPL VGA BIOS.
54 QEMU uses YM3812 emulation by Tatsuyuki Satoh.
56 QEMU uses GUS emulation (GUSEMU32 http://www.deinmeister.de/gusemu/) by
58 Tibor "TS" Schütz.
59 Note that, by default, GUS shares IRQ(7) with parallel ports and so
61 QEMU must be told to not have parallel ports to have working GUS.
62 qemu-system-x86_64 dos.img -device gus -parallel none
64 Alternatively:
66 qemu-system-x86_64 dos.img -device gus,irq=5
68 Or some other unclaimed IRQ.
70 CS4231A is the chip used in Windows Sound System and GUSMAX products
72 The PC speaker audio device can be configured using the pcspk-audiodev
74 machine property, i.e.
75 qemu-system-x86_64 some.img -audiodev <backend>,id=<name> -machine pcspk-audiodev=<name>
77
79 disk_image is a raw hard disk image for IDE hard disk 0. Some targets
80 do not need a disk image.
81 Standard options
83 -h Display help and exit
84 -version
86 Display version information and exit
87 -machine [type=]name[,prop=value[,...]]
89 Select the emulated machine by name. Use -machine help to list
90 available machines.
91 For architectures which aim to support live migration compati‐
93 bility across releases, each release will introduce a new ver‐
94 sioned machine type. For example, the 2.8.0 release introduced
95 machine types "pc-i440fx-2.8" and "pc-q35-2.8" for the
96 x86_64/i686 architectures.
97 To allow live migration of guests from QEMU version 2.8.0, to
99 QEMU version 2.9.0, the 2.9.0 version must support the
100 "pc-i440fx-2.8" and "pc-q35-2.8" machines too. To allow users
101 live migrating VMs to skip multiple intermediate releases when
102 upgrading, new releases of QEMU will support machine types from
103 many previous versions.
104 Supported machine properties are:
106 accel=accels1[:accels2[:...]]
108 This is used to enable an accelerator. Depending on the
109 target architecture, kvm, xen, hax, hvf, nvmm, whpx or
110 tcg can be available. By default, tcg is used. If there
111 is more than one accelerator specified, the next one is
112 used if the previous one fails to initialize.
113 vmport=on|off|auto
115 Enables emulation of VMWare IO port, for vmmouse etc.
116 auto says to select the value based on accel. For ac‐
117 cel=xen the default is off otherwise the default is on.
118 dump-guest-core=on|off
120 Include guest memory in a core dump. The default is on.
121 mem-merge=on|off
123 Enables or disables memory merge support. This feature,
124 when supported by the host, de-duplicates identical mem‐
125 ory pages among VMs instances (enabled by default).
126 aes-key-wrap=on|off
128 Enables or disables AES key wrapping support on s390-ccw
129 hosts. This feature controls whether AES wrapping keys
130 will be created to allow execution of AES cryptographic
131 functions. The default is on.
132 dea-key-wrap=on|off
134 Enables or disables DEA key wrapping support on s390-ccw
135 hosts. This feature controls whether DEA wrapping keys
136 will be created to allow execution of DEA cryptographic
137 functions. The default is on.
138 nvdimm=on|off
140 Enables or disables NVDIMM support. The default is off.
141 memory-encryption=
143 Memory encryption object to use. The default is none.
144 hmat=on|off
146 Enables or disables ACPI Heterogeneous Memory Attribute
147 Table (HMAT) support. The default is off.
148 memory-backend='id'
150 An alternative to legacy -mem-path and mem-prealloc op‐
151 tions. Allows to use a memory backend as main RAM.
152 For example:
154 -object memory-backend-file,id=pc.ram,size=512M,mem-path=/hugetlbfs,prealloc=on,share=on
156 -machine memory-backend=pc.ram
157 -m 512M
158 Migration compatibility note:
160 • as backend id one shall use value of 'default-ram-id',
162 advertised by machine type (available via query-ma‐
163 chines QMP command), if migration to/from old QEMU
164 (<5.0) is expected.
165 • for machine types 4.0 and older, user shall use
167 x-use-canonical-path-for-ramblock-id=off backend option
168 if migration to/from old QEMU (<5.0) is expected.
169 For example:
171 -object memory-backend-ram,id=pc.ram,size=512M,x-use-canonical-path-for-ramblock-id=off
173 -machine memory-backend=pc.ram
174 -m 512M
175 sgx-epc.0.memdev=@var{memid},sgx-epc.0.node=@var{numaid}
177 Define an SGX EPC section.
178 -cpu model
180 Select CPU model (-cpu help for list and additional feature se‐
181 lection)
182 -accel name[,prop=value[,...]]
184 This is used to enable an accelerator. Depending on the target
185 architecture, kvm, xen, hax, hvf, nvmm, whpx or tcg can be
186 available. By default, tcg is used. If there is more than one
187 accelerator specified, the next one is used if the previous one
188 fails to initialize.
189 igd-passthru=on|off
191 When Xen is in use, this option controls whether Intel
192 integrated graphics devices can be passed through to the
193 guest (default=off)
194 kernel-irqchip=on|off|split
196 Controls KVM in-kernel irqchip support. The default is
197 full acceleration of the interrupt controllers. On x86,
198 split irqchip reduces the kernel attack surface, at a
199 performance cost for non-MSI interrupts. Disabling the
200 in-kernel irqchip completely is not recommended except
201 for debugging purposes.
202 kvm-shadow-mem=size
204 Defines the size of the KVM shadow MMU.
205 split-wx=on|off
207 Controls the use of split w^x mapping for the TCG code
208 generation buffer. Some operating systems require this to
209 be enabled, and in such a case this will default on. On
210 other operating systems, this will default off, but one
211 may enable this for testing or debugging.
212 tb-size=n
214 Controls the size (in MiB) of the TCG translation block
215 cache.
216 thread=single|multi
218 Controls number of TCG threads. When the TCG is
219 multi-threaded there will be one thread per vCPU there‐
220 fore taking advantage of additional host cores. The de‐
221 fault is to enable multi-threading where both the
222 back-end and front-ends support it and no incompatible
223 TCG features have been enabled (e.g. icount/replay).
224 dirty-ring-size=n
226 When the KVM accelerator is used, it controls the size of
227 the per-vCPU dirty page ring buffer (number of entries
228 for each vCPU). It should be a value that is power of
229 two, and it should be 1024 or bigger (but still less than
230 the maximum value that the kernel supports). 4096 could
231 be a good initial value if you have no idea which is the
232 best. Set this value to 0 to disable the feature. By
233 default, this feature is disabled (dirty-ring-size=0).
234 When enabled, KVM will instead record dirty pages in a
235 bitmap.
236 -smp [[cpus=]n][,maxcpus=maxcpus][,sockets=sockets][,dies=dies][,clus‐
238 ters=clusters][,cores=cores][,threads=threads]
239 Simulate a SMP system with 'n' CPUs initially present on the ma‐
240 chine type board. On boards supporting CPU hotplug, the optional
241 'maxcpus' parameter can be set to enable further CPUs to be
242 added at runtime. When both parameters are omitted, the maximum
243 number of CPUs will be calculated from the provided topology
244 members and the initial CPU count will match the maximum number.
245 When only one of them is given then the omitted one will be set
246 to its counterpart's value. Both parameters may be specified,
247 but the maximum number of CPUs must be equal to or greater than
248 the initial CPU count. Product of the CPU topology hierarchy
249 must be equal to the maximum number of CPUs. Both parameters
250 are subject to an upper limit that is determined by the specific
251 machine type chosen.
252 To control reporting of CPU topology information, values of the
254 topology parameters can be specified. Machines may only support
255 a subset of the parameters and different machines may have dif‐
256 ferent subsets supported which vary depending on capacity of the
257 corresponding CPU targets. So for a particular machine type
258 board, an expected topology hierarchy can be defined through the
259 supported sub-option. Unsupported parameters can also be pro‐
260 vided in addition to the sub-option, but their values must be
261 set as 1 in the purpose of correct parsing.
262 Either the initial CPU count, or at least one of the topology
264 parameters must be specified. The specified parameters must be
265 greater than zero, explicit configuration like "cpus=0" is not
266 allowed. Values for any omitted parameters will be computed from
267 those which are given.
268 For example, the following sub-option defines a CPU topology hi‐
270 erarchy (2 sockets totally on the machine, 2 cores per socket, 2
271 threads per core) for a machine that only supports sock‐
272 ets/cores/threads. Some members of the option can be omitted
273 but their values will be automatically computed:
274 -smp 8,sockets=2,cores=2,threads=2,maxcpus=8
276 The following sub-option defines a CPU topology hierarchy (2
278 sockets totally on the machine, 2 dies per socket, 2 cores per
279 die, 2 threads per core) for PC machines which support sock‐
280 ets/dies/cores/threads. Some members of the option can be omit‐
281 ted but their values will be automatically computed:
282 -smp 16,sockets=2,dies=2,cores=2,threads=2,maxcpus=16
284 The following sub-option defines a CPU topology hierarchy (2
286 sockets totally on the machine, 2 clusters per socket, 2 cores
287 per cluster, 2 threads per core) for ARM virt machines which
288 support sockets/clusters /cores/threads. Some members of the op‐
289 tion can be omitted but their values will be automatically com‐
290 puted:
291 -smp 16,sockets=2,clusters=2,cores=2,threads=2,maxcpus=16
293 Historically preference was given to the coarsest topology pa‐
295 rameters when computing missing values (ie sockets preferred
296 over cores, which were preferred over threads), however, this
297 behaviour is considered liable to change. Prior to 6.2 the pref‐
298 erence was sockets over cores over threads. Since 6.2 the pref‐
299 erence is cores over sockets over threads.
300 For example, the following option defines a machine board with 2
302 sockets of 1 core before 6.2 and 1 socket of 2 cores after 6.2:
303 -smp 2
305 -numa node[,mem=size][,cpus=firstcpu[-lastcpu]][,nodeid=node][,initia‐
307 tor=initiator]
308 -numa node[,memdev=id][,cpus=firstcpu[-lastcpu]][,nodeid=node][,initia‐
311 tor=initiator]
312 -numa dist,src=source,dst=destination,val=distance
315 -numa cpu,node-id=node[,socket-id=x][,core-id=y][,thread-id=z]
318 -numa hmat-lb,initiator=node,target=node,hierarchy=hierar‐
321 chy,data-type=tpye[,latency=lat][,bandwidth=bw]
322 -numa hmat-cache,node-id=node,size=size,level=level[,associativ‐
325 ity=str][,policy=str][,line=size]
326 Define a NUMA node and assign RAM and VCPUs to it. Set the NUMA
327 distance from a source node to a destination node. Set the ACPI
328 Heterogeneous Memory Attributes for the given nodes.
329 Legacy VCPU assignment uses 'cpus' option where firstcpu and
331 lastcpu are CPU indexes. Each 'cpus' option represent a contigu‐
332 ous range of CPU indexes (or a single VCPU if lastcpu is omit‐
333 ted). A non-contiguous set of VCPUs can be represented by pro‐
334 viding multiple 'cpus' options. If 'cpus' is omitted on all
335 nodes, VCPUs are automatically split between them.
336 For example, the following option assigns VCPUs 0, 1, 2 and 5 to
338 a NUMA node:
339 -numa node,cpus=0-2,cpus=5
341 'cpu' option is a new alternative to 'cpus' option which uses
343 'socket-id|core-id|thread-id' properties to assign CPU objects
344 to a node using topology layout properties of CPU. The set of
345 properties is machine specific, and depends on used machine
346 type/'smp' options. It could be queried with 'hotpluggable-cpus'
347 monitor command. 'node-id' property specifies node to which CPU
348 object will be assigned, it's required for node to be declared
349 with 'node' option before it's used with 'cpu' option.
350 For example:
352 -M pc \
354 -smp 1,sockets=2,maxcpus=2 \
355 -numa node,nodeid=0 -numa node,nodeid=1 \
356 -numa cpu,node-id=0,socket-id=0 -numa cpu,node-id=1,socket-id=1
357 Legacy 'mem' assigns a given RAM amount to a node (not supported
359 for 5.1 and newer machine types). 'memdev' assigns RAM from a
360 given memory backend device to a node. If 'mem' and 'memdev' are
361 omitted in all nodes, RAM is split equally between them.
362 'mem' and 'memdev' are mutually exclusive. Furthermore, if one
364 node uses 'memdev', all of them have to use it.
365 'initiator' is an additional option that points to an initiator
367 NUMA node that has best performance (the lowest latency or
368 largest bandwidth) to this NUMA node. Note that this option can
369 be set only when the machine property 'hmat' is set to 'on'.
370 Following example creates a machine with 2 NUMA nodes, node 0
372 has CPU. node 1 has only memory, and its initiator is node 0.
373 Note that because node 0 has CPU, by default the initiator of
374 node 0 is itself and must be itself.
375 -machine hmat=on \
377 -m 2G,slots=2,maxmem=4G \
378 -object memory-backend-ram,size=1G,id=m0 \
379 -object memory-backend-ram,size=1G,id=m1 \
380 -numa node,nodeid=0,memdev=m0 \
381 -numa node,nodeid=1,memdev=m1,initiator=0 \
382 -smp 2,sockets=2,maxcpus=2 \
383 -numa cpu,node-id=0,socket-id=0 \
384 -numa cpu,node-id=0,socket-id=1
385 source and destination are NUMA node IDs. distance is the NUMA
387 distance from source to destination. The distance from a node to
388 itself is always 10. If any pair of nodes is given a distance,
389 then all pairs must be given distances. Although, when distances
390 are only given in one direction for each pair of nodes, then the
391 distances in the opposite directions are assumed to be the same.
392 If, however, an asymmetrical pair of distances is given for even
393 one node pair, then all node pairs must be provided distance
394 values for both directions, even when they are symmetrical. When
395 a node is unreachable from another node, set the pair's distance
396 to 255.
397 Note that the -numa option doesn't allocate any of the specified
399 resources, it just assigns existing resources to NUMA nodes.
400 This means that one still has to use the -m, -smp options to al‐
401 locate RAM and VCPUs respectively.
402 Use 'hmat-lb' to set System Locality Latency and Bandwidth In‐
404 formation between initiator and target NUMA nodes in ACPI Het‐
405 erogeneous Attribute Memory Table (HMAT). Initiator NUMA node
406 can create memory requests, usually it has one or more proces‐
407 sors. Target NUMA node contains addressable memory.
408 In 'hmat-lb' option, node are NUMA node IDs. hierarchy is the
410 memory hierarchy of the target NUMA node: if hierarchy is 'mem‐
411 ory', the structure represents the memory performance; if hier‐
412 archy is 'first-level|second-level|third-level', this structure
413 represents aggregated performance of memory side caches for each
414 domain. type of 'data-type' is type of data represented by this
415 structure instance: if 'hierarchy' is 'memory', 'data-type' is
416 'access|read|write' latency or 'access|read|write' bandwidth of
417 the target memory; if 'hierarchy' is 'first-level|sec‐
418 ond-level|third-level', 'data-type' is 'access|read|write' hit
419 latency or 'access|read|write' hit bandwidth of the target mem‐
420 ory side cache.
421 lat is latency value in nanoseconds. bw is bandwidth value, the
423 possible value and units are NUM[M|G|T], mean that the bandwidth
424 value are NUM byte per second (or MB/s, GB/s or TB/s depending
425 on used suffix). Note that if latency or bandwidth value is 0,
426 means the corresponding latency or bandwidth information is not
427 provided.
428 In 'hmat-cache' option, node-id is the NUMA-id of the memory be‐
430 longs. size is the size of memory side cache in bytes. level is
431 the cache level described in this structure, note that the cache
432 level 0 should not be used with 'hmat-cache' option. associa‐
433 tivity is the cache associativity, the possible value is
434 'none/direct(direct-mapped)/complex(complex cache indexing)'.
435 policy is the write policy. line is the cache Line size in
436 bytes.
437 For example, the following options describe 2 NUMA nodes. Node 0
439 has 2 cpus and a ram, node 1 has only a ram. The processors in
440 node 0 access memory in node 0 with access-latency 5 nanosec‐
441 onds, access-bandwidth is 200 MB/s; The processors in NUMA node
442 0 access memory in NUMA node 1 with access-latency 10 nanosec‐
443 onds, access-bandwidth is 100 MB/s. And for memory side cache
444 information, NUMA node 0 and 1 both have 1 level memory cache,
445 size is 10KB, policy is write-back, the cache Line size is 8
446 bytes:
447 -machine hmat=on \
449 -m 2G \
450 -object memory-backend-ram,size=1G,id=m0 \
451 -object memory-backend-ram,size=1G,id=m1 \
452 -smp 2,sockets=2,maxcpus=2 \
453 -numa node,nodeid=0,memdev=m0 \
454 -numa node,nodeid=1,memdev=m1,initiator=0 \
455 -numa cpu,node-id=0,socket-id=0 \
456 -numa cpu,node-id=0,socket-id=1 \
457 -numa hmat-lb,initiator=0,target=0,hierarchy=memory,data-type=access-latency,latency=5 \
458 -numa hmat-lb,initiator=0,target=0,hierarchy=memory,data-type=access-bandwidth,bandwidth=200M \
459 -numa hmat-lb,initiator=0,target=1,hierarchy=memory,data-type=access-latency,latency=10 \
460 -numa hmat-lb,initiator=0,target=1,hierarchy=memory,data-type=access-bandwidth,bandwidth=100M \
461 -numa hmat-cache,node-id=0,size=10K,level=1,associativity=direct,policy=write-back,line=8 \
462 -numa hmat-cache,node-id=1,size=10K,level=1,associativity=direct,policy=write-back,line=8
463 -add-fd fd=fd,set=set[,opaque=opaque]
465 Add a file descriptor to an fd set. Valid options are:
466 fd=fd This option defines the file descriptor of which a dupli‐
468 cate is added to fd set. The file descriptor cannot be
469 stdin, stdout, or stderr.
470 set=set
472 This option defines the ID of the fd set to add the file
473 descriptor to.
474 opaque=opaque
476 This option defines a free-form string that can be used
477 to describe fd.
478 You can open an image using pre-opened file descriptors from an
480 fd set:
481 qemu-system-x86_64 \
483 -add-fd fd=3,set=2,opaque="rdwr:/path/to/file" \
484 -add-fd fd=4,set=2,opaque="rdonly:/path/to/file" \
485 -drive file=/dev/fdset/2,index=0,media=disk
486 -set group.id.arg=value
488 Set parameter arg for item id of type group
489 -global driver.prop=value
491 -global driver=driver,property=property,value=value
494 Set default value of driver's property prop to value, e.g.:
495 qemu-system-x86_64 -global ide-hd.physical_block_size=4096 disk-image.img
497 In particular, you can use this to set driver properties for de‐
499 vices which are created automatically by the machine model. To
500 create a device which is not created automatically and set prop‐
501 erties on it, use -device.
502 -global driver.prop=value is shorthand for -global
504 driver=driver,property=prop,value=value. The longhand syntax
505 works even when driver contains a dot.
506 -boot [or‐
508 der=drives][,once=drives][,menu=on|off][,splash=sp_name][,splash-time=sp_time][,re‐
509 boot-timeout=rb_timeout][,strict=on|off]
510 Specify boot order drives as a string of drive letters. Valid
511 drive letters depend on the target architecture. The x86 PC
512 uses: a, b (floppy 1 and 2), c (first hard disk), d (first
513 CD-ROM), n-p (Etherboot from network adapter 1-4), hard disk
514 boot is the default. To apply a particular boot order only on
515 the first startup, specify it via once. Note that the order or
516 once parameter should not be used together with the bootindex
517 property of devices, since the firmware implementations normally
518 do not support both at the same time.
519 Interactive boot menus/prompts can be enabled via menu=on as far
521 as firmware/BIOS supports them. The default is non-interactive
522 boot.
523 A splash picture could be passed to bios, enabling user to show
525 it as logo, when option splash=sp_name is given and menu=on, If
526 firmware/BIOS supports them. Currently Seabios for X86 system
527 support it. limitation: The splash file could be a jpeg file or
528 a BMP file in 24 BPP format(true color). The resolution should
529 be supported by the SVGA mode, so the recommended is 320x240,
530 640x480, 800x640.
531 A timeout could be passed to bios, guest will pause for rb_time‐
533 out ms when boot failed, then reboot. If rb_timeout is '-1',
534 guest will not reboot, qemu passes '-1' to bios by default. Cur‐
535 rently Seabios for X86 system support it.
536 Do strict boot via strict=on as far as firmware/BIOS supports
538 it. This only effects when boot priority is changed by bootindex
539 options. The default is non-strict boot.
540 # try to boot from network first, then from hard disk
542 qemu-system-x86_64 -boot order=nc
543 # boot from CD-ROM first, switch back to default order after reboot
544 qemu-system-x86_64 -boot once=d
545 # boot with a splash picture for 5 seconds.
546 qemu-system-x86_64 -boot menu=on,splash=/root/boot.bmp,splash-time=5000
547 Note: The legacy format '-boot drives' is still supported but
549 its use is discouraged as it may be removed from future ver‐
550 sions.
551 -m [size=]megs[,slots=n,maxmem=size]
553 Sets guest startup RAM size to megs megabytes. Default is 128
554 MiB. Optionally, a suffix of "M" or "G" can be used to signify
555 a value in megabytes or gigabytes respectively. Optional pair
556 slots, maxmem could be used to set amount of hotpluggable memory
557 slots and maximum amount of memory. Note that maxmem must be
558 aligned to the page size.
559 For example, the following command-line sets the guest startup
561 RAM size to 1GB, creates 3 slots to hotplug additional memory
562 and sets the maximum memory the guest can reach to 4GB:
563 qemu-system-x86_64 -m 1G,slots=3,maxmem=4G
565 If slots and maxmem are not specified, memory hotplug won't be
567 enabled and the guest startup RAM will never increase.
568 -mem-path path
570 Allocate guest RAM from a temporarily created file in path.
571 -mem-prealloc
573 Preallocate memory when using -mem-path.
574 -k language
576 Use keyboard layout language (for example fr for French). This
577 option is only needed where it is not easy to get raw PC key‐
578 codes (e.g. on Macs, with some X11 servers or with a VNC or
579 curses display). You don't normally need to use it on PC/Linux
580 or PC/Windows hosts.
581 The available layouts are:
583 ar de-ch es fo fr-ca hu ja mk no pt-br sv
585 da en-gb et fr fr-ch is lt nl pl ru th
586 de en-us fi fr-be hr it lv nl-be pt sl tr
587 The default is en-us.
589 -audio-help
591 Will show the -audiodev equivalent of the currently specified
592 (deprecated) environment variables.
593 -audiodev [driver=]driver,id=id[,prop[=value][,...]]
595 Adds a new audio backend driver identified by id. There are
596 global and driver specific properties. Some values can be set
597 differently for input and output, they're marked with in|out..
598 You can set the input's property with in.prop and the output's
599 property with out.prop. For example:
600 -audiodev alsa,id=example,in.frequency=44110,out.frequency=8000
602 -audiodev alsa,id=example,out.channels=1 # leaves in.channels unspecified
603 NOTE: parameter validation is known to be incomplete, in many
605 cases specifying an invalid option causes QEMU to print an error
606 message and continue emulation without sound.
607 Valid global options are:
609 id=identifier
611 Identifies the audio backend.
612 timer-period=period
614 Sets the timer period used by the audio subsystem in mi‐
615 croseconds. Default is 10000 (10 ms).
616 in|out.mixing-engine=on|off
618 Use QEMU's mixing engine to mix all streams inside QEMU
619 and convert audio formats when not supported by the back‐
620 end. When off, fixed-settings must be off too. Note that
621 disabling this option means that the selected backend
622 must support multiple streams and the audio formats used
623 by the virtual cards, otherwise you'll get no sound. It's
624 not recommended to disable this option unless you want to
625 use 5.1 or 7.1 audio, as mixing engine only supports mono
626 and stereo audio. Default is on.
627 in|out.fixed-settings=on|off
629 Use fixed settings for host audio. When off, it will
630 change based on how the guest opens the sound card. In
631 this case you must not specify frequency, channels or
632 format. Default is on.
633 in|out.frequency=frequency
635 Specify the frequency to use when using fixed-settings.
636 Default is 44100Hz.
637 in|out.channels=channels
639 Specify the number of channels to use when using
640 fixed-settings. Default is 2 (stereo).
641 in|out.format=format
643 Specify the sample format to use when using fixed-set‐
644 tings. Valid values are: s8, s16, s32, u8, u16, u32,
645 f32. Default is s16.
646 in|out.voices=voices
648 Specify the number of voices to use. Default is 1.
649 in|out.buffer-length=usecs
651 Sets the size of the buffer in microseconds.
652 -audiodev none,id=id[,prop[=value][,...]]
654 Creates a dummy backend that discards all outputs. This backend
655 has no backend specific properties.
656 -audiodev alsa,id=id[,prop[=value][,...]]
658 Creates backend using the ALSA. This backend is only available
659 on Linux.
660 ALSA specific options are:
662 in|out.dev=device
664 Specify the ALSA device to use for input and/or output.
665 Default is default.
666 in|out.period-length=usecs
668 Sets the period length in microseconds.
669 in|out.try-poll=on|off
671 Attempt to use poll mode with the device. Default is on.
672 threshold=threshold
674 Threshold (in microseconds) when playback starts. Default
675 is 0.
676 -audiodev coreaudio,id=id[,prop[=value][,...]]
678 Creates a backend using Apple's Core Audio. This backend is only
679 available on Mac OS and only supports playback.
680 Core Audio specific options are:
682 in|out.buffer-count=count
684 Sets the count of the buffers.
685 -audiodev dsound,id=id[,prop[=value][,...]]
687 Creates a backend using Microsoft's DirectSound. This backend is
688 only available on Windows and only supports playback.
689 DirectSound specific options are:
691 latency=usecs
693 Add extra usecs microseconds latency to playback. Default
694 is 10000 (10 ms).
695 -audiodev oss,id=id[,prop[=value][,...]]
697 Creates a backend using OSS. This backend is available on most
698 Unix-like systems.
699 OSS specific options are:
701 in|out.dev=device
703 Specify the file name of the OSS device to use. Default
704 is /dev/dsp.
705 in|out.buffer-count=count
707 Sets the count of the buffers.
708 in|out.try-poll=on|of
710 Attempt to use poll mode with the device. Default is on.
711 try-mmap=on|off
713 Try using memory mapped device access. Default is off.
714 exclusive=on|off
716 Open the device in exclusive mode (vmix won't work in
717 this case). Default is off.
718 dsp-policy=policy
720 Sets the timing policy (between 0 and 10, where smaller
721 number means smaller latency but higher CPU usage). Use
722 -1 to use buffer sizes specified by buffer and buf‐
723 fer-count. This option is ignored if you do not have OSS
724 4. Default is 5.
725 -audiodev pa,id=id[,prop[=value][,...]]
727 Creates a backend using PulseAudio. This backend is available on
728 most systems.
729 PulseAudio specific options are:
731 server=server
733 Sets the PulseAudio server to connect to.
734 in|out.name=sink
736 Use the specified source/sink for recording/playback.
737 in|out.latency=usecs
739 Desired latency in microseconds. The PulseAudio server
740 will try to honor this value but actual latencies may be
741 lower or higher.
742 -audiodev sdl,id=id[,prop[=value][,...]]
744 Creates a backend using SDL. This backend is available on most
745 systems, but you should use your platform's native backend if
746 possible.
747 SDL specific options are:
749 in|out.buffer-count=count
751 Sets the count of the buffers.
752 -audiodev spice,id=id[,prop[=value][,...]]
754 Creates a backend that sends audio through SPICE. This backend
755 requires -spice and automatically selected in that case, so usu‐
756 ally you can ignore this option. This backend has no backend
757 specific properties.
758 -audiodev wav,id=id[,prop[=value][,...]]
760 Creates a backend that writes audio to a WAV file.
761 Backend specific options are:
763 path=path
765 Write recorded audio into the specified file. Default is
766 qemu.wav.
767 -soundhw card1[,card2,...] or -soundhw all
769 Enable audio and selected sound hardware. Use 'help' to print
770 all available sound hardware. For example:
771 qemu-system-x86_64 -soundhw sb16,adlib disk.img
773 qemu-system-x86_64 -soundhw es1370 disk.img
774 qemu-system-x86_64 -soundhw ac97 disk.img
775 qemu-system-x86_64 -soundhw hda disk.img
776 qemu-system-x86_64 -soundhw all disk.img
777 qemu-system-x86_64 -soundhw help
778 Note that Linux's i810_audio OSS kernel (for AC97) module might
780 require manually specifying clocking.
781 modprobe i810_audio clocking=48000
783 -device driver[,prop[=value][,...]]
785 Add device driver. prop=value sets driver properties. Valid
786 properties depend on the driver. To get help on possible drivers
787 and properties, use -device help and -device driver,help.
788 Some drivers are:
790 -device ipmi-bmc-sim,id=id[,prop[=value][,...]]
792 Add an IPMI BMC. This is a simulation of a hardware management
793 interface processor that normally sits on a system. It provides
794 a watchdog and the ability to reset and power control the sys‐
795 tem. You need to connect this to an IPMI interface to make it
796 useful
797 The IPMI slave address to use for the BMC. The default is 0x20.
799 This address is the BMC's address on the I2C network of manage‐
800 ment controllers. If you don't know what this means, it is safe
801 to ignore it.
802 id=id The BMC id for interfaces to use this device.
804 slave_addr=val
806 Define slave address to use for the BMC. The default is
807 0x20.
808 sdrfile=file
810 file containing raw Sensor Data Records (SDR) data. The
811 default is none.
812 fruareasize=val
814 size of a Field Replaceable Unit (FRU) area. The default
815 is 1024.
816 frudatafile=file
818 file containing raw Field Replaceable Unit (FRU) inven‐
819 tory data. The default is none.
820 guid=uuid
822 value for the GUID for the BMC, in standard UUID format.
823 If this is set, get "Get GUID" command to the BMC will
824 return it. Otherwise "Get GUID" will return an error.
825 -device ipmi-bmc-extern,id=id,chardev=id[,slave_addr=val]
827 Add a connection to an external IPMI BMC simulator. Instead of
828 locally emulating the BMC like the above item, instead connect
829 to an external entity that provides the IPMI services.
830 A connection is made to an external BMC simulator. If you do
832 this, it is strongly recommended that you use the "reconnect="
833 chardev option to reconnect to the simulator if the connection
834 is lost. Note that if this is not used carefully, it can be a
835 security issue, as the interface has the ability to send resets,
836 NMIs, and power off the VM. It's best if QEMU makes a connection
837 to an external simulator running on a secure port on localhost,
838 so neither the simulator nor QEMU is exposed to any outside net‐
839 work.
840 See the "lanserv/README.vm" file in the OpenIPMI library for
842 more details on the external interface.
843 -device isa-ipmi-kcs,bmc=id[,ioport=val][,irq=val]
845 Add a KCS IPMI interafce on the ISA bus. This also adds a corre‐
846 sponding ACPI and SMBIOS entries, if appropriate.
847 bmc=id The BMC to connect to, one of ipmi-bmc-sim or
849 ipmi-bmc-extern above.
850 ioport=val
852 Define the I/O address of the interface. The default is
853 0xca0 for KCS.
854 irq=val
856 Define the interrupt to use. The default is 5. To disable
857 interrupts, set this to 0.
858 -device isa-ipmi-bt,bmc=id[,ioport=val][,irq=val]
860 Like the KCS interface, but defines a BT interface. The default
861 port is 0xe4 and the default interrupt is 5.
862 -device pci-ipmi-kcs,bmc=id
864 Add a KCS IPMI interafce on the PCI bus.
865 bmc=id The BMC to connect to, one of ipmi-bmc-sim or
867 ipmi-bmc-extern above.
868 -device pci-ipmi-bt,bmc=id
870 Like the KCS interface, but defines a BT interface on the PCI
871 bus.
872 -device intel-iommu[,option=...]
874 This is only supported by -machine q35, which will enable Intel
875 VT-d emulation within the guest. It supports below options:
876 intremap=on|off (default: auto)
878 This enables interrupt remapping feature. It's required
879 to enable complete x2apic. Currently it only supports
880 kvm kernel-irqchip modes off or split, while full ker‐
881 nel-irqchip is not yet supported. The default value is
882 "auto", which will be decided by the mode of ker‐
883 nel-irqchip.
884 caching-mode=on|off (default: off)
886 This enables caching mode for the VT-d emulated device.
887 When caching-mode is enabled, each guest DMA buffer map‐
888 ping will generate an IOTLB invalidation from the guest
889 IOMMU driver to the vIOMMU device in a synchronous way.
890 It is required for -device vfio-pci to work with the VT-d
891 device, because host assigned devices requires to setup
892 the DMA mapping on the host before guest DMA starts.
893 device-iotlb=on|off (default: off)
895 This enables device-iotlb capability for the emulated
896 VT-d device. So far virtio/vhost should be the only real
897 user for this parameter, paired with ats=on configured
898 for the device.
899 aw-bits=39|48 (default: 39)
901 This decides the address width of IOVA address space.
902 The address space has 39 bits width for 3-level IOMMU
903 page tables, and 48 bits for 4-level IOMMU page tables.
904 Please also refer to the wiki page for general scenarios of VT-d
906 emulation in QEMU: https://wiki.qemu.org/Features/VT-d.
907 -name name
909 Sets the name of the guest. This name will be displayed in the
910 SDL window caption. The name will also be used for the VNC
911 server. Also optionally set the top visible process name in
912 Linux. Naming of individual threads can also be enabled on Linux
913 to aid debugging.
914 -uuid uuid
916 Set system UUID.
917 Block device options
919 -fda file
920 -fdb file
923 Use file as floppy disk 0/1 image (see the Disk Images chapter
924 in the System Emulation Users Guide).
925 -hda file
927 -hdb file
930 -hdc file
933 -hdd file
936 Use file as hard disk 0, 1, 2 or 3 image (see the Disk Images
937 chapter in the System Emulation Users Guide).
938 -cdrom file
940 Use file as CD-ROM image (you cannot use -hdc and -cdrom at the
941 same time). You can use the host CD-ROM by using /dev/cdrom as
942 filename.
943 -blockdev option[,option[,option[,...]]]
945 Define a new block driver node. Some of the options apply to all
946 block drivers, other options are only accepted for a specific
947 block driver. See below for a list of generic options and op‐
948 tions for the most common block drivers.
949 Options that expect a reference to another node (e.g. file) can
951 be given in two ways. Either you specify the node name of an al‐
952 ready existing node (file=node-name), or you define a new node
953 inline, adding options for the referenced node after a dot
954 (file.filename=path,file.aio=native).
955 A block driver node created with -blockdev can be used for a
957 guest device by specifying its node name for the drive property
958 in a -device argument that defines a block device.
959 Valid options for any block driver node:
961 driver Specifies the block driver to use for the given
963 node.
964 node-name
966 This defines the name of the block driver node by
967 which it will be referenced later. The name must
968 be unique, i.e. it must not match the name of a
969 different block driver node, or (if you use -drive
970 as well) the ID of a drive.
971 If no node name is specified, it is automatically
973 generated. The generated node name is not in‐
974 tended to be predictable and changes between QEMU
975 invocations. For the top level, an explicit node
976 name must be specified.
977 read-only
979 Open the node read-only. Guest write attempts will
980 fail.
981 Note that some block drivers support only
983 read-only access, either generally or in certain
984 configurations. In this case, the default value
985 read-only=off does not work and the option must be
986 specified explicitly.
987 auto-read-only
989 If auto-read-only=on is set, QEMU may fall back to
990 read-only usage even when read-only=off is re‐
991 quested, or even switch between modes as needed,
992 e.g. depending on whether the image file is
993 writable or whether a writing user is attached to
994 the node.
995 force-share
997 Override the image locking system of QEMU by forc‐
998 ing the node to utilize weaker shared access for
999 permissions where it would normally request exclu‐
1000 sive access. When there is the potential for mul‐
1001 tiple instances to have the same file open
1002 (whether this invocation of QEMU is the first or
1003 the second instance), both instances must permit
1004 shared access for the second instance to succeed
1005 at opening the file.
1006 Enabling force-share=on requires read-only=on.
1008 cache.direct
1010 The host page cache can be avoided with cache.di‐
1011 rect=on. This will attempt to do disk IO directly
1012 to the guest's memory. QEMU may still perform an
1013 internal copy of the data.
1014 cache.no-flush
1016 In case you don't care about data integrity over
1017 host failures, you can use cache.no-flush=on. This
1018 option tells QEMU that it never needs to write any
1019 data to the disk but can instead keep things in
1020 cache. If anything goes wrong, like your host los‐
1021 ing power, the disk storage getting disconnected
1022 accidentally, etc. your image will most probably
1023 be rendered unusable.
1024 discard=discard
1026 discard is one of "ignore" (or "off") or "unmap"
1027 (or "on") and controls whether discard (also known
1028 as trim or unmap) requests are ignored or passed
1029 to the filesystem. Some machine types may not
1030 support discard requests.
1031 detect-zeroes=detect-zeroes
1033 detect-zeroes is "off", "on" or "unmap" and en‐
1034 ables the automatic conversion of plain zero
1035 writes by the OS to driver specific optimized zero
1036 write commands. You may even choose "unmap" if
1037 discard is set to "unmap" to allow a zero write to
1038 be converted to an unmap operation.
1039 Driver-specific options for file
1041 This is the protocol-level block driver for accessing
1042 regular files.
1043 filename
1045 The path to the image file in the local filesystem
1046 aio Specifies the AIO backend (threads/native/io_ur‐
1048 ing, default: threads)
1049 locking
1051 Specifies whether the image file is protected with
1052 Linux OFD / POSIX locks. The default is to use the
1053 Linux Open File Descriptor API if available, oth‐
1054 erwise no lock is applied. (auto/on/off, default:
1055 auto)
1056 Example:
1058 -blockdev driver=file,node-name=disk,filename=disk.img
1060 Driver-specific options for raw
1062 This is the image format block driver for raw images. It
1063 is usually stacked on top of a protocol level block
1064 driver such as file.
1065 file Reference to or definition of the data source
1067 block driver node (e.g. a file driver node)
1068 Example 1:
1070 -blockdev driver=file,node-name=disk_file,filename=disk.img
1072 -blockdev driver=raw,node-name=disk,file=disk_file
1073 Example 2:
1075 -blockdev driver=raw,node-name=disk,file.driver=file,file.filename=disk.img
1077 Driver-specific options for qcow2
1079 This is the image format block driver for qcow2 images.
1080 It is usually stacked on top of a protocol level block
1081 driver such as file.
1082 file Reference to or definition of the data source
1084 block driver node (e.g. a file driver node)
1085 backing
1087 Reference to or definition of the backing file
1088 block device (default is taken from the image
1089 file). It is allowed to pass null here in order to
1090 disable the default backing file.
1091 lazy-refcounts
1093 Whether to enable the lazy refcounts feature
1094 (on/off; default is taken from the image file)
1095 cache-size
1097 The maximum total size of the L2 table and ref‐
1098 count block caches in bytes (default: the sum of
1099 l2-cache-size and refcount-cache-size)
1100 l2-cache-size
1102 The maximum size of the L2 table cache in bytes
1103 (default: if cache-size is not specified - 32M on
1104 Linux platforms, and 8M on non-Linux platforms;
1105 otherwise, as large as possible within the
1106 cache-size, while permitting the requested or the
1107 minimal refcount cache size)
1108 refcount-cache-size
1110 The maximum size of the refcount block cache in
1111 bytes (default: 4 times the cluster size; or if
1112 cache-size is specified, the part of it which is
1113 not used for the L2 cache)
1114 cache-clean-interval
1116 Clean unused entries in the L2 and refcount
1117 caches. The interval is in seconds. The default
1118 value is 600 on supporting platforms, and 0 on
1119 other platforms. Setting it to 0 disables this
1120 feature.
1121 pass-discard-request
1123 Whether discard requests to the qcow2 device
1124 should be forwarded to the data source (on/off;
1125 default: on if discard=unmap is specified, off
1126 otherwise)
1127 pass-discard-snapshot
1129 Whether discard requests for the data source
1130 should be issued when a snapshot operation (e.g.
1131 deleting a snapshot) frees clusters in the qcow2
1132 file (on/off; default: on)
1133 pass-discard-other
1135 Whether discard requests for the data source
1136 should be issued on other occasions where a clus‐
1137 ter gets freed (on/off; default: off)
1138 overlap-check
1140 Which overlap checks to perform for writes to the
1141 image (none/constant/cached/all; default: cached).
1142 For details or finer granularity control refer to
1143 the QAPI documentation of blockdev-add.
1144 Example 1:
1146 -blockdev driver=file,node-name=my_file,filename=/tmp/disk.qcow2
1148 -blockdev driver=qcow2,node-name=hda,file=my_file,overlap-check=none,cache-size=16777216
1149 Example 2:
1151 -blockdev driver=qcow2,node-name=disk,file.driver=http,file.filename=http://example.com/image.qcow2
1153 Driver-specific options for other drivers
1155 Please refer to the QAPI documentation of the block‐
1156 dev-add QMP command.
1157 -drive option[,option[,option[,...]]]
1159 Define a new drive. This includes creating a block driver node
1160 (the backend) as well as a guest device, and is mostly a short‐
1161 cut for defining the corresponding -blockdev and -device op‐
1162 tions.
1163 -drive accepts all options that are accepted by -blockdev. In
1165 addition, it knows the following options:
1166 file=file
1168 This option defines which disk image (see the Disk Images
1169 chapter in the System Emulation Users Guide) to use with
1170 this drive. If the filename contains comma, you must
1171 double it (for instance, "file=my,,file" to use file
1172 "my,file").
1173 Special files such as iSCSI devices can be specified us‐
1175 ing protocol specific URLs. See the section for "Device
1176 URL Syntax" for more information.
1177 if=interface
1179 This option defines on which type on interface the drive
1180 is connected. Available types are: ide, scsi, sd, mtd,
1181 floppy, pflash, virtio, none.
1182 bus=bus,unit=unit
1184 These options define where is connected the drive by
1185 defining the bus number and the unit id.
1186 index=index
1188 This option defines where the drive is connected by using
1189 an index in the list of available connectors of a given
1190 interface type.
1191 media=media
1193 This option defines the type of the media: disk or cdrom.
1194 snapshot=snapshot
1196 snapshot is "on" or "off" and controls snapshot mode for
1197 the given drive (see -snapshot).
1198 cache=cache
1200 cache is "none", "writeback", "unsafe", "directsync" or
1201 "writethrough" and controls how the host cache is used to
1202 access block data. This is a shortcut that sets the
1203 cache.direct and cache.no-flush options (as in -block‐
1204 dev), and additionally cache.writeback, which provides a
1205 default for the write-cache option of block guest devices
1206 (as in -device). The modes correspond to the following
1207 settings:
1208 ┌─────────────┬─────────────────┬──────────────┬────────────────┐
1210 │ │ cache.writeback │ cache.direct │ cache.no-flush │
1211 ├─────────────┼─────────────────┼──────────────┼────────────────┤
1212 │writeback │ on │ off │ off │
1213 ├─────────────┼─────────────────┼──────────────┼────────────────┤
1214 │none │ on │ on │ off │
1215 ├─────────────┼─────────────────┼──────────────┼────────────────┤
1216 │writethrough │ off │ off │ off │
1217 ├─────────────┼─────────────────┼──────────────┼────────────────┤
1218 │directsync │ off │ on │ off │
1219 ├─────────────┼─────────────────┼──────────────┼────────────────┤
1220 │unsafe │ on │ off │ on │
1221 └─────────────┴─────────────────┴──────────────┴────────────────┘
1222 The default mode is cache=writeback.
1224 aio=aio
1226 aio is "threads", "native", or "io_uring" and selects be‐
1227 tween pthread based disk I/O, native Linux AIO, or Linux
1228 io_uring API.
1229 format=format
1231 Specify which disk format will be used rather than de‐
1232 tecting the format. Can be used to specify format=raw to
1233 avoid interpreting an untrusted format header.
1234 werror=action,rerror=action
1236 Specify which action to take on write and read errors.
1237 Valid actions are: "ignore" (ignore the error and try to
1238 continue), "stop" (pause QEMU), "report" (report the er‐
1239 ror to the guest), "enospc" (pause QEMU only if the host
1240 disk is full; report the error to the guest otherwise).
1241 The default setting is werror=enospc and rerror=report.
1242 copy-on-read=copy-on-read
1244 copy-on-read is "on" or "off" and enables whether to copy
1245 read backing file sectors into the image file.
1246 bps=b,bps_rd=r,bps_wr=w
1248 Specify bandwidth throttling limits in bytes per second,
1249 either for all request types or for reads or writes only.
1250 Small values can lead to timeouts or hangs inside the
1251 guest. A safe minimum for disks is 2 MB/s.
1252 bps_max=bm,bps_rd_max=rm,bps_wr_max=wm
1254 Specify bursts in bytes per second, either for all re‐
1255 quest types or for reads or writes only. Bursts allow the
1256 guest I/O to spike above the limit temporarily.
1257 iops=i,iops_rd=r,iops_wr=w
1259 Specify request rate limits in requests per second, ei‐
1260 ther for all request types or for reads or writes only.
1261 iops_max=bm,iops_rd_max=rm,iops_wr_max=wm
1263 Specify bursts in requests per second, either for all re‐
1264 quest types or for reads or writes only. Bursts allow the
1265 guest I/O to spike above the limit temporarily.
1266 iops_size=is
1268 Let every is bytes of a request count as a new request
1269 for iops throttling purposes. Use this option to prevent
1270 guests from circumventing iops limits by sending fewer
1271 but larger requests.
1272 group=g
1274 Join a throttling quota group with given name g. All
1275 drives that are members of the same group are accounted
1276 for together. Use this option to prevent guests from cir‐
1277 cumventing throttling limits by using many small disks
1278 instead of a single larger disk.
1279 By default, the cache.writeback=on mode is used. It will report
1281 data writes as completed as soon as the data is present in the
1282 host page cache. This is safe as long as your guest OS makes
1283 sure to correctly flush disk caches where needed. If your guest
1284 OS does not handle volatile disk write caches correctly and your
1285 host crashes or loses power, then the guest may experience data
1286 corruption.
1287 For such guests, you should consider using cache.writeback=off.
1289 This means that the host page cache will be used to read and
1290 write data, but write notification will be sent to the guest
1291 only after QEMU has made sure to flush each write to the disk.
1292 Be aware that this has a major impact on performance.
1293 When using the -snapshot option, unsafe caching is always used.
1295 Copy-on-read avoids accessing the same backing file sectors re‐
1297 peatedly and is useful when the backing file is over a slow net‐
1298 work. By default copy-on-read is off.
1299 Instead of -cdrom you can use:
1301 qemu-system-x86_64 -drive file=file,index=2,media=cdrom
1303 Instead of -hda, -hdb, -hdc, -hdd, you can use:
1305 qemu-system-x86_64 -drive file=file,index=0,media=disk
1307 qemu-system-x86_64 -drive file=file,index=1,media=disk
1308 qemu-system-x86_64 -drive file=file,index=2,media=disk
1309 qemu-system-x86_64 -drive file=file,index=3,media=disk
1310 You can open an image using pre-opened file descriptors from an
1312 fd set:
1313 qemu-system-x86_64 \
1315 -add-fd fd=3,set=2,opaque="rdwr:/path/to/file" \
1316 -add-fd fd=4,set=2,opaque="rdonly:/path/to/file" \
1317 -drive file=/dev/fdset/2,index=0,media=disk
1318 You can connect a CDROM to the slave of ide0:
1320 qemu-system-x86_64 -drive file=file,if=ide,index=1,media=cdrom
1322 If you don't specify the "file=" argument, you define an empty
1324 drive:
1325 qemu-system-x86_64 -drive if=ide,index=1,media=cdrom
1327 Instead of -fda, -fdb, you can use:
1329 qemu-system-x86_64 -drive file=file,index=0,if=floppy
1331 qemu-system-x86_64 -drive file=file,index=1,if=floppy
1332 By default, interface is "ide" and index is automatically incre‐
1334 mented:
1335 qemu-system-x86_64 -drive file=a -drive file=b"
1337 is interpreted like:
1339 qemu-system-x86_64 -hda a -hdb b
1341 -mtdblock file
1343 Use file as on-board Flash memory image.
1344 -sd file
1346 Use file as SecureDigital card image.
1347 -pflash file
1349 Use file as a parallel flash image.
1350 -snapshot
1352 Write to temporary files instead of disk image files. In this
1353 case, the raw disk image you use is not written back. You can
1354 however force the write back by pressing C-a s (see the Disk Im‐
1355 ages chapter in the System Emulation Users Guide).
1356 -fsdev local,id=id,path=path,security_model=security_model [,write‐
1358 out=writeout][,readonly=on][,fmode=fmode][,dmode=dmode] [,throt‐
1359 tling.option=value[,throttling.option=value[,...]]]
1360 -fsdev proxy,id=id,socket=socket[,writeout=writeout][,readonly=on]
1363 -fsdev proxy,id=id,sock_fd=sock_fd[,writeout=writeout][,readonly=on]
1366 -fsdev synth,id=id[,readonly=on]
1369 Define a new file system device. Valid options are:
1370 local Accesses to the filesystem are done by QEMU.
1372 proxy Accesses to the filesystem are done by
1374 virtfs-proxy-helper(1).
1375 synth Synthetic filesystem, only used by QTests.
1377 id=id Specifies identifier for this device.
1379 path=path
1381 Specifies the export path for the file system device.
1382 Files under this path will be available to the 9p client
1383 on the guest.
1384 security_model=security_model
1386 Specifies the security model to be used for this export
1387 path. Supported security models are "passthrough",
1388 "mapped-xattr", "mapped-file" and "none". In
1389 "passthrough" security model, files are stored using the
1390 same credentials as they are created on the guest. This
1391 requires QEMU to run as root. In "mapped-xattr" security
1392 model, some of the file attributes like uid, gid, mode
1393 bits and link target are stored as file attributes. For
1394 "mapped-file" these attributes are stored in the hidden
1395 .virtfs_metadata directory. Directories exported by this
1396 security model cannot interact with other unix tools.
1397 "none" security model is same as passthrough except the
1398 sever won't report failures if it fails to set file at‐
1399 tributes like ownership. Security model is mandatory only
1400 for local fsdriver. Other fsdrivers (like proxy) don't
1401 take security model as a parameter.
1402 writeout=writeout
1404 This is an optional argument. The only supported value is
1405 "immediate". This means that host page cache will be used
1406 to read and write data but write notification will be
1407 sent to the guest only when the data has been reported as
1408 written by the storage subsystem.
1409 readonly=on
1411 Enables exporting 9p share as a readonly mount for
1412 guests. By default read-write access is given.
1413 socket=socket
1415 Enables proxy filesystem driver to use passed socket file
1416 for communicating with virtfs-proxy-helper(1).
1417 sock_fd=sock_fd
1419 Enables proxy filesystem driver to use passed socket de‐
1420 scriptor for communicating with virtfs-proxy-helper(1).
1421 Usually a helper like libvirt will create socketpair and
1422 pass one of the fds as sock_fd.
1423 fmode=fmode
1425 Specifies the default mode for newly created files on the
1426 host. Works only with security models "mapped-xattr" and
1427 "mapped-file".
1428 dmode=dmode
1430 Specifies the default mode for newly created directories
1431 on the host. Works only with security models
1432 "mapped-xattr" and "mapped-file".
1433 throttling.bps-total=b,throttling.bps-read=r,throt‐
1435 tling.bps-write=w
1436 Specify bandwidth throttling limits in bytes per second,
1437 either for all request types or for reads or writes only.
1438 throttling.bps-total-max=bm,bps-read-max=rm,bps-write-max=wm
1440 Specify bursts in bytes per second, either for all re‐
1441 quest types or for reads or writes only. Bursts allow the
1442 guest I/O to spike above the limit temporarily.
1443 throttling.iops-total=i,throttling.iops-read=r, throt‐
1445 tling.iops-write=w
1446 Specify request rate limits in requests per second, ei‐
1447 ther for all request types or for reads or writes only.
1448 throttling.iops-total-max=im,throttling.iops-read-max=irm,
1450 throttling.iops-write-max=iwm
1451 Specify bursts in requests per second, either for all re‐
1452 quest types or for reads or writes only. Bursts allow the
1453 guest I/O to spike above the limit temporarily.
1454 throttling.iops-size=is
1456 Let every is bytes of a request count as a new request
1457 for iops throttling purposes.
1458 -fsdev option is used along with -device driver "virtio-9p-...".
1460 -device virtio-9p-type,fsdev=id,mount_tag=mount_tag
1462 Options for virtio-9p-... driver are:
1463 type Specifies the variant to be used. Supported values are
1465 "pci", "ccw" or "device", depending on the machine type.
1466 fsdev=id
1468 Specifies the id value specified along with -fsdev op‐
1469 tion.
1470 mount_tag=mount_tag
1472 Specifies the tag name to be used by the guest to mount
1473 this export point.
1474 -virtfs local,path=path,mount_tag=mount_tag ,security_model=secu‐
1476 rity_model[,writeout=writeout][,readonly=on]
1477 [,fmode=fmode][,dmode=dmode][,multidevs=multidevs]
1478 -virtfs proxy,socket=socket,mount_tag=mount_tag [,writeout=write‐
1481 out][,readonly=on]
1482 -virtfs proxy,sock_fd=sock_fd,mount_tag=mount_tag [,writeout=write‐
1485 out][,readonly=on]
1486 -virtfs synth,mount_tag=mount_tag
1489 Define a new virtual filesystem device and expose it to the
1490 guest using a virtio-9p-device (a.k.a. 9pfs), which essentially
1491 means that a certain directory on host is made directly accessi‐
1492 ble by guest as a pass-through file system by using the 9P net‐
1493 work protocol for communication between host and guests, if de‐
1494 sired even accessible, shared by several guests simultaniously.
1495 Note that -virtfs is actually just a convenience shortcut for
1497 its generalized form -fsdev -device virtio-9p-pci.
1498 The general form of pass-through file system options are:
1500 local Accesses to the filesystem are done by QEMU.
1502 proxy Accesses to the filesystem are done by
1504 virtfs-proxy-helper(1).
1505 synth Synthetic filesystem, only used by QTests.
1507 id=id Specifies identifier for the filesystem device
1509 path=path
1511 Specifies the export path for the file system device.
1512 Files under this path will be available to the 9p client
1513 on the guest.
1514 security_model=security_model
1516 Specifies the security model to be used for this export
1517 path. Supported security models are "passthrough",
1518 "mapped-xattr", "mapped-file" and "none". In
1519 "passthrough" security model, files are stored using the
1520 same credentials as they are created on the guest. This
1521 requires QEMU to run as root. In "mapped-xattr" security
1522 model, some of the file attributes like uid, gid, mode
1523 bits and link target are stored as file attributes. For
1524 "mapped-file" these attributes are stored in the hidden
1525 .virtfs_metadata directory. Directories exported by this
1526 security model cannot interact with other unix tools.
1527 "none" security model is same as passthrough except the
1528 sever won't report failures if it fails to set file at‐
1529 tributes like ownership. Security model is mandatory only
1530 for local fsdriver. Other fsdrivers (like proxy) don't
1531 take security model as a parameter.
1532 writeout=writeout
1534 This is an optional argument. The only supported value is
1535 "immediate". This means that host page cache will be used
1536 to read and write data but write notification will be
1537 sent to the guest only when the data has been reported as
1538 written by the storage subsystem.
1539 readonly=on
1541 Enables exporting 9p share as a readonly mount for
1542 guests. By default read-write access is given.
1543 socket=socket
1545 Enables proxy filesystem driver to use passed socket file
1546 for communicating with virtfs-proxy-helper(1). Usually a
1547 helper like libvirt will create socketpair and pass one
1548 of the fds as sock_fd.
1549 sock_fd
1551 Enables proxy filesystem driver to use passed 'sock_fd'
1552 as the socket descriptor for interfacing with
1553 virtfs-proxy-helper(1).
1554 fmode=fmode
1556 Specifies the default mode for newly created files on the
1557 host. Works only with security models "mapped-xattr" and
1558 "mapped-file".
1559 dmode=dmode
1561 Specifies the default mode for newly created directories
1562 on the host. Works only with security models
1563 "mapped-xattr" and "mapped-file".
1564 mount_tag=mount_tag
1566 Specifies the tag name to be used by the guest to mount
1567 this export point.
1568 multidevs=multidevs
1570 Specifies how to deal with multiple devices being shared
1571 with a 9p export. Supported behaviours are either
1572 "remap", "forbid" or "warn". The latter is the default
1573 behaviour on which virtfs 9p expects only one device to
1574 be shared with the same export, and if more than one de‐
1575 vice is shared and accessed via the same 9p export then
1576 only a warning message is logged (once) by qemu on host
1577 side. In order to avoid file ID collisions on guest you
1578 should either create a separate virtfs export for each
1579 device to be shared with guests (recommended way) or you
1580 might use "remap" instead which allows you to share mul‐
1581 tiple devices with only one export instead, which is
1582 achieved by remapping the original inode numbers from
1583 host to guest in a way that would prevent such colli‐
1584 sions. Remapping inodes in such use cases is required be‐
1585 cause the original device IDs from host are never passed
1586 and exposed on guest. Instead all files of an export
1587 shared with virtfs always share the same device id on
1588 guest. So two files with identical inode numbers but from
1589 actually different devices on host would otherwise cause
1590 a file ID collision and hence potential misbehaviours on
1591 guest. "forbid" on the other hand assumes like "warn"
1592 that only one device is shared by the same export, how‐
1593 ever it will not only log a warning message but also deny
1594 access to additional devices on guest. Note though that
1595 "forbid" does currently not block all possible file ac‐
1596 cess operations (e.g. readdir() would still return en‐
1597 tries from other devices).
1598 -iscsi Configure iSCSI session parameters.
1600 USB convenience options
1602 -usb Enable USB emulation on machine types with an on-board USB host
1603 controller (if not enabled by default). Note that on-board USB
1604 host controllers may not support USB 3.0. In this case -device
1605 qemu-xhci can be used instead on machines with PCI.
1606 -usbdevice devname
1608 Add the USB device devname, and enable an on-board USB con‐
1609 troller if possible and necessary (just like it can be done via
1610 -machine usb=on). Note that this option is mainly intended for
1611 the user's convenience only. More fine-grained control can be
1612 achieved by selecting a USB host controller (if necessary) and
1613 the desired USB device via the -device option instead. For exam‐
1614 ple, instead of using -usbdevice mouse it is possible to use
1615 -device qemu-xhci -device usb-mouse to connect the USB mouse to
1616 a USB 3.0 controller instead (at least on machines that support
1617 PCI and do not have an USB controller enabled by default yet).
1618 For more details, see the chapter about Connecting USB devices
1619 in the System Emulation Users Guide. Possible devices for dev‐
1620 name are:
1621 braille
1623 Braille device. This will use BrlAPI to display the
1624 braille output on a real or fake device (i.e. it also
1625 creates a corresponding braille chardev automatically be‐
1626 side the usb-braille USB device).
1627 keyboard
1629 Standard USB keyboard. Will override the PS/2 keyboard
1630 (if present).
1631 mouse Virtual Mouse. This will override the PS/2 mouse emula‐
1633 tion when activated.
1634 tablet Pointer device that uses absolute coordinates (like a
1636 touchscreen). This means QEMU is able to report the mouse
1637 position without having to grab the mouse. Also overrides
1638 the PS/2 mouse emulation when activated.
1639 wacom-tablet
1641 Wacom PenPartner USB tablet.
1642 Display options
1644 -display type
1645 Select type of display to use. This option is a replacement for
1646 the old style -sdl/-curses/... options. Use -display help to
1647 list the available display types. Valid values for type are
1648 spice-app[,gl=on|off]
1650 Start QEMU as a Spice server and launch the default Spice
1651 client application. The Spice server will redirect the
1652 serial consoles and QEMU monitors. (Since 4.0)
1653 dbus Export the display over D-Bus interfaces. (Since 7.0)
1655 The connection is registered with the "org.qemu" name
1657 (and queued when already owned).
1658 addr=<dbusaddr> : D-Bus bus address to connect to.
1660 p2p=yes|no : Use peer-to-peer connection, accepted via
1662 QMP add_client.
1663 gl=on|off|core|es : Use OpenGL for rendering (the D-Bus
1665 interface will share framebuffers with DMABUF file de‐
1666 scriptors).
1667 sdl Display video output via SDL (usually in a separate
1669 graphics window; see the SDL documentation for other pos‐
1670 sibilities). Valid parameters are:
1671 grab-mod=<mods> : Used to select the modifier keys for
1673 toggling the mouse grabbing in conjunction with the "g"
1674 key. <mods> can be either lshift-lctrl-lalt or rctrl.
1675 alt_grab=on|off : Use Control+Alt+Shift-g to toggle mouse
1677 grabbing. This parameter is deprecated - use grab-mod
1678 instead.
1679 ctrl_grab=on|off : Use Right-Control-g to toggle mouse
1681 grabbing. This parameter is deprecated - use grab-mod
1682 instead.
1683 gl=on|off|core|es : Use OpenGL for displaying
1685 show-cursor=on|off : Force showing the mouse cursor
1687 window-close=on|off : Allow to quit qemu with window
1689 close button
1690 gtk Display video output in a GTK window. This interface pro‐
1692 vides drop-down menus and other UI elements to configure
1693 and control the VM during runtime. Valid parameters are:
1694 full-screen=on|off : Start in fullscreen mode
1696 gl=on|off : Use OpenGL for displaying
1698 grab-on-hover=on|off : Grab keyboard input on mouse hover
1700 show-cursor=on|off : Force showing the mouse cursor
1702 window-close=on|off : Allow to quit qemu with window
1704 close button
1705 curses[,charset=<encoding>]
1707 Display video output via curses. For graphics device mod‐
1708 els which support a text mode, QEMU can display this out‐
1709 put using a curses/ncurses interface. Nothing is dis‐
1710 played when the graphics device is in graphical mode or
1711 if the graphics device does not support a text mode. Gen‐
1712 erally only the VGA device models support text mode. The
1713 font charset used by the guest can be specified with the
1714 charset option, for example charset=CP850 for IBM CP850
1715 encoding. The default is CP437.
1716 cocoa Display video output in a Cocoa window. Mac only. This
1718 interface provides drop-down menus and other UI elements
1719 to configure and control the VM during runtime. Valid pa‐
1720 rameters are:
1721 show-cursor=on|off : Force showing the mouse cursor
1723 left-command-key=on|off : Disable forwarding left command
1725 key to host
1726 egl-headless[,rendernode=<file>]
1728 Offload all OpenGL operations to a local DRI device. For
1729 any graphical display, this display needs to be paired
1730 with either VNC or SPICE displays.
1731 vnc=<display>
1733 Start a VNC server on display <display>
1734 none Do not display video output. The guest will still see an
1736 emulated graphics card, but its output will not be dis‐
1737 played to the QEMU user. This option differs from the
1738 -nographic option in that it only affects what is done
1739 with video output; -nographic also changes the destina‐
1740 tion of the serial and parallel port data.
1741 -nographic
1743 Normally, if QEMU is compiled with graphical window support, it
1744 displays output such as guest graphics, guest console, and the
1745 QEMU monitor in a window. With this option, you can totally dis‐
1746 able graphical output so that QEMU is a simple command line ap‐
1747 plication. The emulated serial port is redirected on the con‐
1748 sole and muxed with the monitor (unless redirected elsewhere ex‐
1749 plicitly). Therefore, you can still use QEMU to debug a Linux
1750 kernel with a serial console. Use C-a h for help on switching
1751 between the console and monitor.
1752 -curses
1754 Normally, if QEMU is compiled with graphical window support, it
1755 displays output such as guest graphics, guest console, and the
1756 QEMU monitor in a window. With this option, QEMU can display the
1757 VGA output when in text mode using a curses/ncurses interface.
1758 Nothing is displayed in graphical mode.
1759 -alt-grab
1761 Use Ctrl-Alt-Shift to grab mouse (instead of Ctrl-Alt). Note
1762 that this also affects the special keys (for fullscreen, moni‐
1763 tor-mode switching, etc). This option is deprecated - please use
1764 -display sdl,grab-mod=lshift-lctrl-lalt instead.
1765 -ctrl-grab
1767 Use Right-Ctrl to grab mouse (instead of Ctrl-Alt). Note that
1768 this also affects the special keys (for fullscreen, monitor-mode
1769 switching, etc). This option is deprecated - please use -display
1770 sdl,grab-mod=rctrl instead.
1771 -sdl Enable SDL.
1773 -spice option[,option[,...]]
1775 Enable the spice remote desktop protocol. Valid options are
1776 port=<nr>
1778 Set the TCP port spice is listening on for plaintext
1779 channels.
1780 addr=<addr>
1782 Set the IP address spice is listening on. Default is any
1783 address.
1784 ipv4=on|off; ipv6=on|off; unix=on|off
1786 Force using the specified IP version.
1787 password=<string>
1789 Set the password you need to authenticate.
1790 This option is deprecated and insecure because it leaves
1792 the password visible in the process listing. Use pass‐
1793 word-secret instead.
1794 password-secret=<secret-id>
1796 Set the ID of the secret object containing the password
1797 you need to authenticate.
1798 sasl=on|off
1800 Require that the client use SASL to authenticate with the
1801 spice. The exact choice of authentication method used is
1802 controlled from the system / user's SASL configuration
1803 file for the 'qemu' service. This is typically found in
1804 /etc/sasl2/qemu.conf. If running QEMU as an unprivileged
1805 user, an environment variable SASL_CONF_PATH can be used
1806 to make it search alternate locations for the service
1807 config. While some SASL auth methods can also provide
1808 data encryption (eg GSSAPI), it is recommended that SASL
1809 always be combined with the 'tls' and 'x509' settings to
1810 enable use of SSL and server certificates. This ensures a
1811 data encryption preventing compromise of authentication
1812 credentials.
1813 disable-ticketing=on|off
1815 Allow client connects without authentication.
1816 disable-copy-paste=on|off
1818 Disable copy paste between the client and the guest.
1819 disable-agent-file-xfer=on|off
1821 Disable spice-vdagent based file-xfer between the client
1822 and the guest.
1823 tls-port=<nr>
1825 Set the TCP port spice is listening on for encrypted
1826 channels.
1827 x509-dir=<dir>
1829 Set the x509 file directory. Expects same filenames as
1830 -vnc $display,x509=$dir
1831 x509-key-file=<file>; x509-key-password=<file>;
1833 x509-cert-file=<file>; x509-cacert-file=<file>;
1834 x509-dh-key-file=<file>
1835 The x509 file names can also be configured individually.
1836 tls-ciphers=<list>
1838 Specify which ciphers to use.
1839 tls-channel=[main|display|cursor|inputs|record|playback]; plain‐
1841 text-channel=[main|display|cursor|inputs|record|playback]
1842 Force specific channel to be used with or without TLS en‐
1843 cryption. The options can be specified multiple times to
1844 configure multiple channels. The special name "default"
1845 can be used to set the default mode. For channels which
1846 are not explicitly forced into one mode the spice client
1847 is allowed to pick tls/plaintext as he pleases.
1848 image-compression=[auto_glz|auto_lz|quic|glz|lz|off]
1850 Configure image compression (lossless). Default is
1851 auto_glz.
1852 jpeg-wan-compression=[auto|never|always]; zlib-glz-wan-compres‐
1854 sion=[auto|never|always]
1855 Configure wan image compression (lossy for slow links).
1856 Default is auto.
1857 streaming-video=[off|all|filter]
1859 Configure video stream detection. Default is off.
1860 agent-mouse=[on|off]
1862 Enable/disable passing mouse events via vdagent. Default
1863 is on.
1864 playback-compression=[on|off]
1866 Enable/disable audio stream compression (using celt
1867 0.5.1). Default is on.
1868 seamless-migration=[on|off]
1870 Enable/disable spice seamless migration. Default is off.
1871 gl=[on|off]
1873 Enable/disable OpenGL context. Default is off.
1874 rendernode=<file>
1876 DRM render node for OpenGL rendering. If not specified,
1877 it will pick the first available. (Since 2.9)
1878 -portrait
1880 Rotate graphical output 90 deg left (only PXA LCD).
1881 -rotate deg
1883 Rotate graphical output some deg left (only PXA LCD).
1884 -vga type
1886 Select type of VGA card to emulate. Valid values for type are
1887 cirrus Cirrus Logic GD5446 Video card. All Windows versions
1889 starting from Windows 95 should recognize and use this
1890 graphic card. For optimal performances, use 16 bit color
1891 depth in the guest and the host OS. (This card was the
1892 default before QEMU 2.2)
1893 std Standard VGA card with Bochs VBE extensions. If your
1895 guest OS supports the VESA 2.0 VBE extensions (e.g. Win‐
1896 dows XP) and if you want to use high resolution modes (>=
1897 1280x1024x16) then you should use this option. (This card
1898 is the default since QEMU 2.2)
1899 vmware VMWare SVGA-II compatible adapter. Use it if you have
1901 sufficiently recent XFree86/XOrg server or Windows guest
1902 with a driver for this card.
1903 qxl QXL paravirtual graphic card. It is VGA compatible (in‐
1905 cluding VESA 2.0 VBE support). Works best with qxl guest
1906 drivers installed though. Recommended choice when using
1907 the spice protocol.
1908 tcx (sun4m only) Sun TCX framebuffer. This is the default
1910 framebuffer for sun4m machines and offers both 8-bit and
1911 24-bit colour depths at a fixed resolution of 1024x768.
1912 cg3 (sun4m only) Sun cgthree framebuffer. This is a simple
1914 8-bit framebuffer for sun4m machines available in both
1915 1024x768 (OpenBIOS) and 1152x900 (OBP) resolutions aimed
1916 at people wishing to run older Solaris versions.
1917 virtio Virtio VGA card.
1919 none Disable VGA card.
1921 -full-screen
1923 Start in full screen.
1924 -g widthxheight[xdepth]
1926 Set the initial graphical resolution and depth (PPC, SPARC
1927 only).
1928 For PPC the default is 800x600x32.
1930 For SPARC with the TCX graphics device, the default is
1932 1024x768x8 with the option of 1024x768x24. For cgthree, the de‐
1933 fault is 1024x768x8 with the option of 1152x900x8 for people who
1934 wish to use OBP.
1935 -vnc display[,option[,option[,...]]]
1937 Normally, if QEMU is compiled with graphical window support, it
1938 displays output such as guest graphics, guest console, and the
1939 QEMU monitor in a window. With this option, you can have QEMU
1940 listen on VNC display display and redirect the VGA display over
1941 the VNC session. It is very useful to enable the usb tablet de‐
1942 vice when using this option (option -device usb-tablet). When
1943 using the VNC display, you must use the -k parameter to set the
1944 keyboard layout if you are not using en-us. Valid syntax for the
1945 display is
1946 to=L With this option, QEMU will try next available VNC dis‐
1948 plays, until the number L, if the origianlly defined
1949 "-vnc display" is not available, e.g. port 5900+display
1950 is already used by another application. By default, to=0.
1951 host:d TCP connections will only be allowed from host on display
1953 d. By convention the TCP port is 5900+d. Optionally, host
1954 can be omitted in which case the server will accept con‐
1955 nections from any host.
1956 unix:path
1958 Connections will be allowed over UNIX domain sockets
1959 where path is the location of a unix socket to listen for
1960 connections on.
1961 none VNC is initialized but not started. The monitor change
1963 command can be used to later start the VNC server.
1964 Following the display value there may be one or more option
1966 flags separated by commas. Valid options are
1967 reverse=on|off
1969 Connect to a listening VNC client via a "reverse" connec‐
1970 tion. The client is specified by the display. For re‐
1971 verse network connections (host:d,``reverse``), the d ar‐
1972 gument is a TCP port number, not a display number.
1973 websocket=on|off
1975 Opens an additional TCP listening port dedicated to VNC
1976 Websocket connections. If a bare websocket option is
1977 given, the Websocket port is 5700+display. An alternative
1978 port can be specified with the syntax websocket=port.
1979 If host is specified connections will only be allowed
1981 from this host. It is possible to control the websocket
1982 listen address independently, using the syntax web‐
1983 socket=host:port.
1984 If no TLS credentials are provided, the websocket connec‐
1986 tion runs in unencrypted mode. If TLS credentials are
1987 provided, the websocket connection requires encrypted
1988 client connections.
1989 password=on|off
1991 Require that password based authentication is used for
1992 client connections.
1993 The password must be set separately using the set_pass‐
1995 word command in the QEMU Monitor. The syntax to change
1996 your password is: set_password <protocol> <password>
1997 where <protocol> could be either "vnc" or "spice".
1998 If you would like to change <protocol> password expira‐
2000 tion, you should use expire_password <protocol> <expira‐
2001 tion-time> where expiration time could be one of the fol‐
2002 lowing options: now, never, +seconds or UNIX time of ex‐
2003 piration, e.g. +60 to make password expire in 60 seconds,
2004 or 1335196800 to make password expire on "Mon Apr 23
2005 12:00:00 EDT 2012" (UNIX time for this date and time).
2006 You can also use keywords "now" or "never" for the expi‐
2008 ration time to allow <protocol> password to expire imme‐
2009 diately or never expire.
2010 password-secret=<secret-id>
2012 Require that password based authentication is used for
2013 client connections, using the password provided by the
2014 secret object identified by secret-id.
2015 tls-creds=ID
2017 Provides the ID of a set of TLS credentials to use to se‐
2018 cure the VNC server. They will apply to both the normal
2019 VNC server socket and the websocket socket (if enabled).
2020 Setting TLS credentials will cause the VNC server socket
2021 to enable the VeNCrypt auth mechanism. The credentials
2022 should have been previously created using the -object
2023 tls-creds argument.
2024 tls-authz=ID
2026 Provides the ID of the QAuthZ authorization object
2027 against which the client's x509 distinguished name will
2028 validated. This object is only resolved at time of use,
2029 so can be deleted and recreated on the fly while the VNC
2030 server is active. If missing, it will default to denying
2031 access.
2032 sasl=on|off
2034 Require that the client use SASL to authenticate with the
2035 VNC server. The exact choice of authentication method
2036 used is controlled from the system / user's SASL configu‐
2037 ration file for the 'qemu' service. This is typically
2038 found in /etc/sasl2/qemu.conf. If running QEMU as an un‐
2039 privileged user, an environment variable SASL_CONF_PATH
2040 can be used to make it search alternate locations for the
2041 service config. While some SASL auth methods can also
2042 provide data encryption (eg GSSAPI), it is recommended
2043 that SASL always be combined with the 'tls' and 'x509'
2044 settings to enable use of SSL and server certificates.
2045 This ensures a data encryption preventing compromise of
2046 authentication credentials. See the VNC security section
2047 in the System Emulation Users Guide for details on using
2048 SASL authentication.
2049 sasl-authz=ID
2051 Provides the ID of the QAuthZ authorization object
2052 against which the client's SASL username will validated.
2053 This object is only resolved at time of use, so can be
2054 deleted and recreated on the fly while the VNC server is
2055 active. If missing, it will default to denying access.
2056 acl=on|off
2058 Legacy method for enabling authorization of clients
2059 against the x509 distinguished name and SASL username. It
2060 results in the creation of two authz-list objects with
2061 IDs of vnc.username and vnc.x509dname. The rules for
2062 these objects must be configured with the HMP ACL com‐
2063 mands.
2064 This option is deprecated and should no longer be used.
2066 The new sasl-authz and tls-authz options are a replace‐
2067 ment.
2068 lossy=on|off
2070 Enable lossy compression methods (gradient, JPEG, ...).
2071 If this option is set, VNC client may receive lossy
2072 framebuffer updates depending on its encoding settings.
2073 Enabling this option can save a lot of bandwidth at the
2074 expense of quality.
2075 non-adaptive=on|off
2077 Disable adaptive encodings. Adaptive encodings are en‐
2078 abled by default. An adaptive encoding will try to detect
2079 frequently updated screen regions, and send updates in
2080 these regions using a lossy encoding (like JPEG). This
2081 can be really helpful to save bandwidth when playing
2082 videos. Disabling adaptive encodings restores the origi‐
2083 nal static behavior of encodings like Tight.
2084 share=[allow-exclusive|force-shared|ignore]
2086 Set display sharing policy. 'allow-exclusive' allows
2087 clients to ask for exclusive access. As suggested by the
2088 rfb spec this is implemented by dropping other connec‐
2089 tions. Connecting multiple clients in parallel requires
2090 all clients asking for a shared session (vncviewer:
2091 -shared switch). This is the default. 'force-shared'
2092 disables exclusive client access. Useful for shared desk‐
2093 top sessions, where you don't want someone forgetting
2094 specify -shared disconnect everybody else. 'ignore' com‐
2095 pletely ignores the shared flag and allows everybody con‐
2096 nect unconditionally. Doesn't conform to the rfb spec but
2097 is traditional QEMU behavior.
2098 key-delay-ms
2100 Set keyboard delay, for key down and key up events, in
2101 milliseconds. Default is 10. Keyboards are low-bandwidth
2102 devices, so this slowdown can help the device and guest
2103 to keep up and not lose events in case events are arriv‐
2104 ing in bulk. Possible causes for the latter are flaky
2105 network connections, or scripts for automated testing.
2106 audiodev=audiodev
2108 Use the specified audiodev when the VNC client requests
2109 audio transmission. When not using an -audiodev argument,
2110 this option must be omitted, otherwise is must be present
2111 and specify a valid audiodev.
2112 power-control=on|off
2114 Permit the remote client to issue shutdown, reboot or re‐
2115 set power control requests.
2116 i386 target only
2118 -win2k-hack
2119 Use it when installing Windows 2000 to avoid a disk full bug.
2120 After Windows 2000 is installed, you no longer need this option
2121 (this option slows down the IDE transfers).
2122 -no-fd-bootchk
2124 Disable boot signature checking for floppy disks in BIOS. May be
2125 needed to boot from old floppy disks.
2126 -no-acpi
2128 Disable ACPI (Advanced Configuration and Power Interface) sup‐
2129 port. Use it if your guest OS complains about ACPI problems (PC
2130 target machine only).
2131 -no-hpet
2133 Disable HPET support.
2134 -acpitable [sig=str][,rev=n][,oem_id=str][,oem_ta‐
2136 ble_id=str][,oem_rev=n] [,asl_compiler_id=str][,asl_com‐
2137 piler_rev=n][,data=file1[:file2]...]
2138 Add ACPI table with specified header fields and context from
2139 specified files. For file=, take whole ACPI table from the spec‐
2140 ified files, including all ACPI headers (possible overridden by
2141 other options). For data=, only data portion of the table is
2142 used, all header information is specified in the command line.
2143 If a SLIC table is supplied to QEMU, then the SLIC's oem_id and
2144 oem_table_id fields will override the same in the RSDT and the
2145 FADT (a.k.a. FACP), in order to ensure the field matches re‐
2146 quired by the Microsoft SLIC spec and the ACPI spec.
2147 -smbios file=binary
2149 Load SMBIOS entry from binary file.
2150 -smbios type=0[,vendor=str][,version=str][,date=str][,re‐
2152 lease=%d.%d][,uefi=on|off]
2153 Specify SMBIOS type 0 fields
2154 -smbios type=1[,manufacturer=str][,product=str][,version=str][,se‐
2156 rial=str][,uuid=uuid][,sku=str][,family=str]
2157 Specify SMBIOS type 1 fields
2158 -smbios type=2[,manufacturer=str][,product=str][,version=str][,se‐
2160 rial=str][,asset=str][,location=str]
2161 Specify SMBIOS type 2 fields
2162 -smbios type=3[,manufacturer=str][,version=str][,serial=str][,as‐
2164 set=str][,sku=str]
2165 Specify SMBIOS type 3 fields
2166 -smbios type=4[,sock_pfx=str][,manufacturer=str][,version=str][,se‐
2168 rial=str][,asset=str][,part=str][,processor-id=%d]
2169 Specify SMBIOS type 4 fields
2170 -smbios type=11[,value=str][,path=filename]
2172 Specify SMBIOS type 11 fields
2173 This argument can be repeated multiple times, and values are
2175 added in the order they are parsed. Applications intending to
2176 use OEM strings data are encouraged to use their application
2177 name as a prefix for the value string. This facilitates passing
2178 information for multiple applications concurrently.
2179 The value=str syntax provides the string data inline, while the
2181 path=filename syntax loads data from a file on disk. Note that
2182 the file is not permitted to contain any NUL bytes.
2183 Both the value and path options can be repeated multiple times
2185 and will be added to the SMBIOS table in the order in which they
2186 appear.
2187 Note that on the x86 architecture, the total size of all SMBIOS
2189 tables is limited to 65535 bytes. Thus the OEM strings data is
2190 not suitable for passing large amounts of data into the guest.
2191 Instead it should be used as a indicator to inform the guest
2192 where to locate the real data set, for example, by specifying
2193 the serial ID of a block device.
2194 An example passing three strings is
2196 -smbios type=11,value=cloud-init:ds=nocloud-net;s=http://10.10.0.1:8000/,\
2198 value=anaconda:method=http://dl.fedoraproject.org/pub/fedora/linux/releases/25/x86_64/os,\
2199 path=/some/file/with/oemstringsdata.txt
2200 In the guest OS this is visible with the dmidecode command
2202 $ dmidecode -t 11
2204 Handle 0x0E00, DMI type 11, 5 bytes
2205 OEM Strings
2206 String 1: cloud-init:ds=nocloud-net;s=http://10.10.0.1:8000/
2207 String 2: anaconda:method=http://dl.fedoraproject.org/pub/fedora/linux/releases/25/x86_64/os
2208 String 3: myapp:some extra data
2209 -smbios type=17[,loc_pfx=str][,bank=str][,manufacturer=str][,se‐
2211 rial=str][,asset=str][,part=str][,speed=%d]
2212 Specify SMBIOS type 17 fields
2213 -smbios type=41[,designation=str][,kind=str][,instance=%d][,pcidev=str]
2215 Specify SMBIOS type 41 fields
2216 This argument can be repeated multiple times. Its main use is
2218 to allow network interfaces be created as enoX on Linux, with X
2219 being the instance number, instead of the name depending on the
2220 interface position on the PCI bus.
2221 Here is an example of use:
2223 -netdev user,id=internet \
2225 -device virtio-net-pci,mac=50:54:00:00:00:42,netdev=internet,id=internet-dev \
2226 -smbios type=41,designation='Onboard LAN',instance=1,kind=ethernet,pcidev=internet-dev
2227 In the guest OS, the device should then appear as eno1:
2229 ..parsed-literal:
2231 $ ip -brief l
2233 lo UNKNOWN 00:00:00:00:00:00 <LOOPBACK,UP,LOWER_UP>
2234 eno1 UP 50:54:00:00:00:42 <BROADCAST,MULTICAST,UP,LOWER_UP>
2235 Currently, the PCI device has to be attached to the root bus.
2237 Network options
2239 -nic
2240 [tap|bridge|user|l2tpv3|vde|netmap|vhost-user|socket][,...][,mac=macaddr][,model=mn]
2241 This option is a shortcut for configuring both the on-board (de‐
2242 fault) guest NIC hardware and the host network backend in one
2243 go. The host backend options are the same as with the corre‐
2244 sponding -netdev options below. The guest NIC model can be set
2245 with model=modelname. Use model=help to list the available de‐
2246 vice types. The hardware MAC address can be set with
2247 mac=macaddr.
2248 The following two example do exactly the same, to show how -nic
2250 can be used to shorten the command line length:
2251 qemu-system-x86_64 -netdev user,id=n1,ipv6=off -device e1000,netdev=n1,mac=52:54:98:76:54:32
2253 qemu-system-x86_64 -nic user,ipv6=off,model=e1000,mac=52:54:98:76:54:32
2254 -nic none
2256 Indicate that no network devices should be configured. It is
2257 used to override the default configuration (default NIC with
2258 "user" host network backend) which is activated if no other net‐
2259 working options are provided.
2260 -netdev user,id=id[,option][,option][,...]
2262 Configure user mode host network backend which requires no ad‐
2263 ministrator privilege to run. Valid options are:
2264 id=id Assign symbolic name for use in monitor commands.
2266 ipv4=on|off and ipv6=on|off
2268 Specify that either IPv4 or IPv6 must be enabled. If nei‐
2269 ther is specified both protocols are enabled.
2270 net=addr[/mask]
2272 Set IP network address the guest will see. Optionally
2273 specify the netmask, either in the form a.b.c.d or as
2274 number of valid top-most bits. Default is 10.0.2.0/24.
2275 host=addr
2277 Specify the guest-visible address of the host. Default is
2278 the 2nd IP in the guest network, i.e. x.x.x.2.
2279 ipv6-net=addr[/int]
2281 Set IPv6 network address the guest will see (default is
2282 fec0::/64). The network prefix is given in the usual
2283 hexadecimal IPv6 address notation. The prefix size is op‐
2284 tional, and is given as the number of valid top-most bits
2285 (default is 64).
2286 ipv6-host=addr
2288 Specify the guest-visible IPv6 address of the host. De‐
2289 fault is the 2nd IPv6 in the guest network, i.e. xxxx::2.
2290 restrict=on|off
2292 If this option is enabled, the guest will be isolated,
2293 i.e. it will not be able to contact the host and no guest
2294 IP packets will be routed over the host to the outside.
2295 This option does not affect any explicitly set forwarding
2296 rules.
2297 hostname=name
2299 Specifies the client hostname reported by the built-in
2300 DHCP server.
2301 dhcpstart=addr
2303 Specify the first of the 16 IPs the built-in DHCP server
2304 can assign. Default is the 15th to 31st IP in the guest
2305 network, i.e. x.x.x.15 to x.x.x.31.
2306 dns=addr
2308 Specify the guest-visible address of the virtual name‐
2309 server. The address must be different from the host ad‐
2310 dress. Default is the 3rd IP in the guest network, i.e.
2311 x.x.x.3.
2312 ipv6-dns=addr
2314 Specify the guest-visible address of the IPv6 virtual
2315 nameserver. The address must be different from the host
2316 address. Default is the 3rd IP in the guest network,
2317 i.e. xxxx::3.
2318 dnssearch=domain
2320 Provides an entry for the domain-search list sent by the
2321 built-in DHCP server. More than one domain suffix can be
2322 transmitted by specifying this option multiple times. If
2323 supported, this will cause the guest to automatically try
2324 to append the given domain suffix(es) in case a domain
2325 name can not be resolved.
2326 Example:
2328 qemu-system-x86_64 -nic user,dnssearch=mgmt.example.org,dnssearch=example.org
2330 domainname=domain
2332 Specifies the client domain name reported by the built-in
2333 DHCP server.
2334 tftp=dir
2336 When using the user mode network stack, activate a
2337 built-in TFTP server. The files in dir will be exposed as
2338 the root of a TFTP server. The TFTP client on the guest
2339 must be configured in binary mode (use the command bin of
2340 the Unix TFTP client).
2341 tftp-server-name=name
2343 In BOOTP reply, broadcast name as the "TFTP server name"
2344 (RFC2132 option 66). This can be used to advise the guest
2345 to load boot files or configurations from a different
2346 server than the host address.
2347 bootfile=file
2349 When using the user mode network stack, broadcast file as
2350 the BOOTP filename. In conjunction with tftp, this can be
2351 used to network boot a guest from a local directory.
2352 Example (using pxelinux):
2354 qemu-system-x86_64 -hda linux.img -boot n -device e1000,netdev=n1 \
2356 -netdev user,id=n1,tftp=/path/to/tftp/files,bootfile=/pxelinux.0
2357 smb=dir[,smbserver=addr]
2359 When using the user mode network stack, activate a
2360 built-in SMB server so that Windows OSes can access to
2361 the host files in dir transparently. The IP address of
2362 the SMB server can be set to addr. By default the 4th IP
2363 in the guest network is used, i.e. x.x.x.4.
2364 In the guest Windows OS, the line:
2366 10.0.2.4 smbserver
2368 must be added in the file C:\WINDOWS\LMHOSTS (for windows
2370 9x/Me) or C:\WINNT\SYSTEM32\DRIVERS\ETC\LMHOSTS (Windows
2371 NT/2000).
2372 Then dir can be accessed in \\smbserver\qemu.
2374 Note that a SAMBA server must be installed on the host
2376 OS.
2377 hostfwd=[tcp|udp]:[hostaddr]:hostport-[guestaddr]:guestport
2379 Redirect incoming TCP or UDP connections to the host port
2380 hostport to the guest IP address guestaddr on guest port
2381 guestport. If guestaddr is not specified, its value is
2382 x.x.x.15 (default first address given by the built-in
2383 DHCP server). By specifying hostaddr, the rule can be
2384 bound to a specific host interface. If no connection type
2385 is set, TCP is used. This option can be given multiple
2386 times.
2387 For example, to redirect host X11 connection from screen
2389 1 to guest screen 0, use the following:
2390 # on the host
2392 qemu-system-x86_64 -nic user,hostfwd=tcp:127.0.0.1:6001-:6000
2393 # this host xterm should open in the guest X11 server
2394 xterm -display :1
2395 To redirect telnet connections from host port 5555 to
2397 telnet port on the guest, use the following:
2398 # on the host
2400 qemu-system-x86_64 -nic user,hostfwd=tcp::5555-:23
2401 telnet localhost 5555
2402 Then when you use on the host telnet localhost 5555, you
2404 connect to the guest telnet server.
2405 guestfwd=[tcp]:server:port-dev; guest‐
2407 fwd=[tcp]:server:port-cmd:command
2408 Forward guest TCP connections to the IP address server on
2409 port port to the character device dev or to a program ex‐
2410 ecuted by cmd:command which gets spawned for each connec‐
2411 tion. This option can be given multiple times.
2412 You can either use a chardev directly and have that one
2414 used throughout QEMU's lifetime, like in the following
2415 example:
2416 # open 10.10.1.1:4321 on bootup, connect 10.0.2.100:1234 to it whenever
2418 # the guest accesses it
2419 qemu-system-x86_64 -nic user,guestfwd=tcp:10.0.2.100:1234-tcp:10.10.1.1:4321
2420 Or you can execute a command on every TCP connection es‐
2422 tablished by the guest, so that QEMU behaves similar to
2423 an inetd process for that virtual server:
2424 # call "netcat 10.10.1.1 4321" on every TCP connection to 10.0.2.100:1234
2426 # and connect the TCP stream to its stdin/stdout
2427 qemu-system-x86_64 -nic 'user,id=n1,guestfwd=tcp:10.0.2.100:1234-cmd:netcat 10.10.1.1 4321'
2428 -netdev tap,id=id[,fd=h][,ifname=name][,script=file][,down‐
2430 script=dfile][,br=bridge][,helper=helper]
2431 Configure a host TAP network backend with ID id.
2432 Use the network script file to configure it and the network
2434 script dfile to deconfigure it. If name is not provided, the OS
2435 automatically provides one. The default network configure script
2436 is /etc/qemu-ifup and the default network deconfigure script is
2437 /etc/qemu-ifdown. Use script=no or downscript=no to disable
2438 script execution.
2439 If running QEMU as an unprivileged user, use the network helper
2441 to configure the TAP interface and attach it to the bridge. The
2442 default network helper executable is /path/to/qemu-bridge-helper
2443 and the default bridge device is br0.
2444 fd=h can be used to specify the handle of an already opened host
2446 TAP interface.
2447 Examples:
2449 #launch a QEMU instance with the default network script
2451 qemu-system-x86_64 linux.img -nic tap
2452 #launch a QEMU instance with two NICs, each one connected
2454 #to a TAP device
2455 qemu-system-x86_64 linux.img \
2456 -netdev tap,id=nd0,ifname=tap0 -device e1000,netdev=nd0 \
2457 -netdev tap,id=nd1,ifname=tap1 -device rtl8139,netdev=nd1
2458 #launch a QEMU instance with the default network helper to
2460 #connect a TAP device to bridge br0
2461 qemu-system-x86_64 linux.img -device virtio-net-pci,netdev=n1 \
2462 -netdev tap,id=n1,"helper=/path/to/qemu-bridge-helper"
2463 -netdev bridge,id=id[,br=bridge][,helper=helper]
2465 Connect a host TAP network interface to a host bridge device.
2466 Use the network helper helper to configure the TAP interface and
2468 attach it to the bridge. The default network helper executable
2469 is /path/to/qemu-bridge-helper and the default bridge device is
2470 br0.
2471 Examples:
2473 #launch a QEMU instance with the default network helper to
2475 #connect a TAP device to bridge br0
2476 qemu-system-x86_64 linux.img -netdev bridge,id=n1 -device virtio-net,netdev=n1
2477 #launch a QEMU instance with the default network helper to
2479 #connect a TAP device to bridge qemubr0
2480 qemu-system-x86_64 linux.img -netdev bridge,br=qemubr0,id=n1 -device virtio-net,netdev=n1
2481 -netdev socket,id=id[,fd=h][,listen=[host]:port][,connect=host:port]
2483 This host network backend can be used to connect the guest's
2484 network to another QEMU virtual machine using a TCP socket con‐
2485 nection. If listen is specified, QEMU waits for incoming connec‐
2486 tions on port (host is optional). connect is used to connect to
2487 another QEMU instance using the listen option. fd=h specifies an
2488 already opened TCP socket.
2489 Example:
2491 # launch a first QEMU instance
2493 qemu-system-x86_64 linux.img \
2494 -device e1000,netdev=n1,mac=52:54:00:12:34:56 \
2495 -netdev socket,id=n1,listen=:1234
2496 # connect the network of this instance to the network of the first instance
2497 qemu-system-x86_64 linux.img \
2498 -device e1000,netdev=n2,mac=52:54:00:12:34:57 \
2499 -netdev socket,id=n2,connect=127.0.0.1:1234
2500 -netdev socket,id=id[,fd=h][,mcast=maddr:port[,localaddr=addr]]
2502 Configure a socket host network backend to share the guest's
2503 network traffic with another QEMU virtual machines using a UDP
2504 multicast socket, effectively making a bus for every QEMU with
2505 same multicast address maddr and port. NOTES:
2506 1. Several QEMU can be running on different hosts and share same
2508 bus (assuming correct multicast setup for these hosts).
2509 2. mcast support is compatible with User Mode Linux (argument
2511 ethN=mcast), see http://user-mode-linux.sf.net.
2512 3. Use fd=h to specify an already opened UDP multicast socket.
2514 Example:
2516 # launch one QEMU instance
2518 qemu-system-x86_64 linux.img \
2519 -device e1000,netdev=n1,mac=52:54:00:12:34:56 \
2520 -netdev socket,id=n1,mcast=230.0.0.1:1234
2521 # launch another QEMU instance on same "bus"
2522 qemu-system-x86_64 linux.img \
2523 -device e1000,netdev=n2,mac=52:54:00:12:34:57 \
2524 -netdev socket,id=n2,mcast=230.0.0.1:1234
2525 # launch yet another QEMU instance on same "bus"
2526 qemu-system-x86_64 linux.img \
2527 -device e1000,netdev=n3,mac=52:54:00:12:34:58 \
2528 -netdev socket,id=n3,mcast=230.0.0.1:1234
2529 Example (User Mode Linux compat.):
2531 # launch QEMU instance (note mcast address selected is UML's default)
2533 qemu-system-x86_64 linux.img \
2534 -device e1000,netdev=n1,mac=52:54:00:12:34:56 \
2535 -netdev socket,id=n1,mcast=239.192.168.1:1102
2536 # launch UML
2537 /path/to/linux ubd0=/path/to/root_fs eth0=mcast
2538 Example (send packets from host's 1.2.3.4):
2540 qemu-system-x86_64 linux.img \
2542 -device e1000,netdev=n1,mac=52:54:00:12:34:56 \
2543 -netdev socket,id=n1,mcast=239.192.168.1:1102,localaddr=1.2.3.4
2544 -netdev l2tpv3,id=id,src=srcaddr,dst=dstaddr[,srcport=srcport][,dst‐
2546 port=dstport],txsession=txsession[,rxsession=rxses‐
2547 sion][,ipv6=on|off][,udp=on|off][,cookie64][,counter][,pincounter][,tx‐
2548 cookie=txcookie][,rxcookie=rxcookie][,offset=offset]
2549 Configure a L2TPv3 pseudowire host network backend. L2TPv3
2550 (RFC3931) is a popular protocol to transport Ethernet (and other
2551 Layer 2) data frames between two systems. It is present in
2552 routers, firewalls and the Linux kernel (from version 3.3 on‐
2553 wards).
2554 This transport allows a VM to communicate to another VM, router
2556 or firewall directly.
2557 src=srcaddr
2559 source address (mandatory)
2560 dst=dstaddr
2562 destination address (mandatory)
2563 udp select udp encapsulation (default is ip).
2565 srcport=srcport
2567 source udp port.
2568 dstport=dstport
2570 destination udp port.
2571 ipv6 force v6, otherwise defaults to v4.
2573 rxcookie=rxcookie; txcookie=txcookie
2575 Cookies are a weak form of security in the l2tpv3 speci‐
2576 fication. Their function is mostly to prevent misconfig‐
2577 uration. By default they are 32 bit.
2578 cookie64
2580 Set cookie size to 64 bit instead of the default 32
2581 counter=off
2583 Force a 'cut-down' L2TPv3 with no counter as in
2584 draft-mkonstan-l2tpext-keyed-ipv6-tunnel-00
2585 pincounter=on
2587 Work around broken counter handling in peer. This may
2588 also help on networks which have packet reorder.
2589 offset=offset
2591 Add an extra offset between header and data
2592 For example, to attach a VM running on host 4.3.2.1 via L2TPv3
2594 to the bridge br-lan on the remote Linux host 1.2.3.4:
2595 # Setup tunnel on linux host using raw ip as encapsulation
2597 # on 1.2.3.4
2598 ip l2tp add tunnel remote 4.3.2.1 local 1.2.3.4 tunnel_id 1 peer_tunnel_id 1 \
2599 encap udp udp_sport 16384 udp_dport 16384
2600 ip l2tp add session tunnel_id 1 name vmtunnel0 session_id \
2601 0xFFFFFFFF peer_session_id 0xFFFFFFFF
2602 ifconfig vmtunnel0 mtu 1500
2603 ifconfig vmtunnel0 up
2604 brctl addif br-lan vmtunnel0
2605 # on 4.3.2.1
2608 # launch QEMU instance - if your network has reorder or is very lossy add ,pincounter
2609 qemu-system-x86_64 linux.img -device e1000,netdev=n1 \
2611 -netdev l2tpv3,id=n1,src=4.2.3.1,dst=1.2.3.4,udp,srcport=16384,dstport=16384,rxsession=0xffffffff,txsession=0xffffffff,counter
2612 -netdev vde,id=id[,sock=socketpath][,port=n][,group=group‐
2614 name][,mode=octalmode]
2615 Configure VDE backend to connect to PORT n of a vde switch run‐
2616 ning on host and listening for incoming connections on socket‐
2617 path. Use GROUP groupname and MODE octalmode to change default
2618 ownership and permissions for communication port. This option is
2619 only available if QEMU has been compiled with vde support en‐
2620 abled.
2621 Example:
2623 # launch vde switch
2625 vde_switch -F -sock /tmp/myswitch
2626 # launch QEMU instance
2627 qemu-system-x86_64 linux.img -nic vde,sock=/tmp/myswitch
2628 -netdev vhost-user,chardev=id[,vhostforce=on|off][,queues=n]
2630 Establish a vhost-user netdev, backed by a chardev id. The
2631 chardev should be a unix domain socket backed one. The
2632 vhost-user uses a specifically defined protocol to pass vhost
2633 ioctl replacement messages to an application on the other end of
2634 the socket. On non-MSIX guests, the feature can be forced with
2635 vhostforce. Use 'queues=n' to specify the number of queues to be
2636 created for multiqueue vhost-user.
2637 Example:
2639 qemu -m 512 -object memory-backend-file,id=mem,size=512M,mem-path=/hugetlbfs,share=on \
2641 -numa node,memdev=mem \
2642 -chardev socket,id=chr0,path=/path/to/socket \
2643 -netdev type=vhost-user,id=net0,chardev=chr0 \
2644 -device virtio-net-pci,netdev=net0
2645 -netdev vhost-vdpa,vhostdev=/path/to/dev
2647 Establish a vhost-vdpa netdev.
2648 vDPA device is a device that uses a datapath which complies with
2650 the virtio specifications with a vendor specific control path.
2651 vDPA devices can be both physically located on the hardware or
2652 emulated by software.
2653 -netdev hubport,id=id,hubid=hubid[,netdev=nd]
2655 Create a hub port on the emulated hub with ID hubid.
2656 The hubport netdev lets you connect a NIC to a QEMU emulated hub
2658 instead of a single netdev. Alternatively, you can also connect
2659 the hubport to another netdev with ID nd by using the netdev=nd
2660 option.
2661 -net nic[,netdev=nd][,macaddr=mac][,model=type]
2663 [,name=name][,addr=addr][,vectors=v]
2664 Legacy option to configure or create an on-board (or machine de‐
2665 fault) Network Interface Card(NIC) and connect it either to the
2666 emulated hub with ID 0 (i.e. the default hub), or to the netdev
2667 nd. If model is omitted, then the default NIC model associated
2668 with the machine type is used. Note that the default NIC model
2669 may change in future QEMU releases, so it is highly recommended
2670 to always specify a model. Optionally, the MAC address can be
2671 changed to mac, the device address set to addr (PCI cards only),
2672 and a name can be assigned for use in monitor commands. Option‐
2673 ally, for PCI cards, you can specify the number v of MSI-X vec‐
2674 tors that the card should have; this option currently only af‐
2675 fects virtio cards; set v = 0 to disable MSI-X. If no -net op‐
2676 tion is specified, a single NIC is created. QEMU can emulate
2677 several different models of network card. Use -net
2678 nic,model=help for a list of available devices for your target.
2679 -net user|tap|bridge|socket|l2tpv3|vde[,...][,name=name]
2681 Configure a host network backend (with the options corresponding
2682 to the same -netdev option) and connect it to the emulated hub 0
2683 (the default hub). Use name to specify the name of the hub port.
2684 Character device options
2686 The general form of a character device option is:
2687 -chardev backend,id=id[,mux=on|off][,options]
2689 Backend is one of: null, socket, udp, msmouse, vc, ringbuf,
2690 file, pipe, console, serial, pty, stdio, braille, tty, parallel,
2691 parport, spicevmc, spiceport. The specific backend will deter‐
2692 mine the applicable options.
2693 Use -chardev help to print all available chardev backend types.
2695 All devices must have an id, which can be any string up to 127
2697 characters long. It is used to uniquely identify this device in
2698 other command line directives.
2699 A character device may be used in multiplexing mode by multiple
2701 front-ends. Specify mux=on to enable this mode. A multiplexer is
2702 a "1:N" device, and here the "1" end is your specified chardev
2703 backend, and the "N" end is the various parts of QEMU that can
2704 talk to a chardev. If you create a chardev with id=myid and
2705 mux=on, QEMU will create a multiplexer with your specified ID,
2706 and you can then configure multiple front ends to use that
2707 chardev ID for their input/output. Up to four different front
2708 ends can be connected to a single multiplexed chardev. (Without
2709 multiplexing enabled, a chardev can only be used by a single
2710 front end.) For instance you could use this to allow a single
2711 stdio chardev to be used by two serial ports and the QEMU moni‐
2712 tor:
2713 -chardev stdio,mux=on,id=char0 \
2715 -mon chardev=char0,mode=readline \
2716 -serial chardev:char0 \
2717 -serial chardev:char0
2718 You can have more than one multiplexer in a system configura‐
2720 tion; for instance you could have a TCP port multiplexed between
2721 UART 0 and UART 1, and stdio multiplexed between the QEMU moni‐
2722 tor and a parallel port:
2723 -chardev stdio,mux=on,id=char0 \
2725 -mon chardev=char0,mode=readline \
2726 -parallel chardev:char0 \
2727 -chardev tcp,...,mux=on,id=char1 \
2728 -serial chardev:char1 \
2729 -serial chardev:char1
2730 When you're using a multiplexed character device, some escape
2732 sequences are interpreted in the input. See the chapter about
2733 Keys in the character backend multiplexer in the System Emula‐
2734 tion Users Guide for more details.
2735 Note that some other command line options may implicitly create
2737 multiplexed character backends; for instance -serial mon:stdio
2738 creates a multiplexed stdio backend connected to the serial port
2739 and the QEMU monitor, and -nographic also multiplexes the con‐
2740 sole and the monitor to stdio.
2741 There is currently no support for multiplexing in the other di‐
2743 rection (where a single QEMU front end takes input and output
2744 from multiple chardevs).
2745 Every backend supports the logfile option, which supplies the
2747 path to a file to record all data transmitted via the backend.
2748 The logappend option controls whether the log file will be trun‐
2749 cated or appended to when opened.
2750 The available backends are:
2752 -chardev null,id=id
2754 A void device. This device will not emit any data, and will drop
2755 any data it receives. The null backend does not take any op‐
2756 tions.
2757 -chardev socket,id=id[,TCP options or unix op‐
2759 tions][,server=on|off][,wait=on|off][,telnet=on|off][,web‐
2760 socket=on|off][,reconnect=seconds][,tls-creds=id][,tls-authz=id]
2761 Create a two-way stream socket, which can be either a TCP or a
2762 unix socket. A unix socket will be created if path is specified.
2763 Behaviour is undefined if TCP options are specified for a unix
2764 socket.
2765 server=on|off specifies that the socket shall be a listening
2767 socket.
2768 wait=on|off specifies that QEMU should not block waiting for a
2770 client to connect to a listening socket.
2771 telnet=on|off specifies that traffic on the socket should inter‐
2773 pret telnet escape sequences.
2774 websocket=on|off specifies that the socket uses WebSocket proto‐
2776 col for communication.
2777 reconnect sets the timeout for reconnecting on non-server sock‐
2779 ets when the remote end goes away. qemu will delay this many
2780 seconds and then attempt to reconnect. Zero disables reconnect‐
2781 ing, and is the default.
2782 tls-creds requests enablement of the TLS protocol for encryp‐
2784 tion, and specifies the id of the TLS credentials to use for the
2785 handshake. The credentials must be previously created with the
2786 -object tls-creds argument.
2787 tls-auth provides the ID of the QAuthZ authorization object
2789 against which the client's x509 distinguished name will be vali‐
2790 dated. This object is only resolved at time of use, so can be
2791 deleted and recreated on the fly while the chardev server is ac‐
2792 tive. If missing, it will default to denying access.
2793 TCP and unix socket options are given below:
2795 TCP options:
2797 port=port[,host=host][,to=to][,ipv4=on|off][,ipv6=on|off][,node‐
2798 lay=on|off]
2799 host for a listening socket specifies the local address
2800 to be bound. For a connecting socket species the remote
2801 host to connect to. host is optional for listening sock‐
2802 ets. If not specified it defaults to 0.0.0.0.
2803 port for a listening socket specifies the local port to
2805 be bound. For a connecting socket specifies the port on
2806 the remote host to connect to. port can be given as ei‐
2807 ther a port number or a service name. port is required.
2808 to is only relevant to listening sockets. If it is speci‐
2810 fied, and port cannot be bound, QEMU will attempt to bind
2811 to subsequent ports up to and including to until it suc‐
2812 ceeds. to must be specified as a port number.
2813 ipv4=on|off and ipv6=on|off specify that either IPv4 or
2815 IPv6 must be used. If neither is specified the socket may
2816 use either protocol.
2817 nodelay=on|off disables the Nagle algorithm.
2819 unix options: path=path[,abstract=on|off][,tight=on|off]
2821 path specifies the local path of the unix socket. path is
2822 required. abstract=on|off specifies the use of the ab‐
2823 stract socket namespace, rather than the filesystem. Op‐
2824 tional, defaults to false. tight=on|off sets the socket
2825 length of abstract sockets to their minimum, rather than
2826 the full sun_path length. Optional, defaults to true.
2827 -chardev udp,id=id[,host=host],port=port[,localaddr=localaddr][,local‐
2829 port=localport][,ipv4=on|off][,ipv6=on|off]
2830 Sends all traffic from the guest to a remote host over UDP.
2831 host specifies the remote host to connect to. If not specified
2833 it defaults to localhost.
2834 port specifies the port on the remote host to connect to. port
2836 is required.
2837 localaddr specifies the local address to bind to. If not speci‐
2839 fied it defaults to 0.0.0.0.
2840 localport specifies the local port to bind to. If not specified
2842 any available local port will be used.
2843 ipv4=on|off and ipv6=on|off specify that either IPv4 or IPv6
2845 must be used. If neither is specified the device may use either
2846 protocol.
2847 -chardev msmouse,id=id
2849 Forward QEMU's emulated msmouse events to the guest. msmouse
2850 does not take any options.
2851 -chardev
2853 vc,id=id[[,width=width][,height=height]][[,cols=cols][,rows=rows]]
2854 Connect to a QEMU text console. vc may optionally be given a
2855 specific size.
2856 width and height specify the width and height respectively of
2858 the console, in pixels.
2859 cols and rows specify that the console be sized to fit a text
2861 console with the given dimensions.
2862 -chardev ringbuf,id=id[,size=size]
2864 Create a ring buffer with fixed size size. size must be a power
2865 of two and defaults to 64K.
2866 -chardev file,id=id,path=path
2868 Log all traffic received from the guest to a file.
2869 path specifies the path of the file to be opened. This file will
2871 be created if it does not already exist, and overwritten if it
2872 does. path is required.
2873 -chardev pipe,id=id,path=path
2875 Create a two-way connection to the guest. The behaviour differs
2876 slightly between Windows hosts and other hosts:
2877 On Windows, a single duplex pipe will be created at
2879 \\.pipe\path.
2880 On other hosts, 2 pipes will be created called path.in and
2882 path.out. Data written to path.in will be received by the guest.
2883 Data written by the guest can be read from path.out. QEMU will
2884 not create these fifos, and requires them to be present.
2885 path forms part of the pipe path as described above. path is re‐
2887 quired.
2888 -chardev console,id=id
2890 Send traffic from the guest to QEMU's standard output. console
2891 does not take any options.
2892 console is only available on Windows hosts.
2894 -chardev serial,id=id,path=path
2896 Send traffic from the guest to a serial device on the host.
2897 On Unix hosts serial will actually accept any tty device, not
2899 only serial lines.
2900 path specifies the name of the serial device to open.
2902 -chardev pty,id=id
2904 Create a new pseudo-terminal on the host and connect to it. pty
2905 does not take any options.
2906 pty is not available on Windows hosts.
2908 -chardev stdio,id=id[,signal=on|off]
2910 Connect to standard input and standard output of the QEMU
2911 process.
2912 signal controls if signals are enabled on the terminal, that in‐
2914 cludes exiting QEMU with the key sequence Control-c. This option
2915 is enabled by default, use signal=off to disable it.
2916 -chardev braille,id=id
2918 Connect to a local BrlAPI server. braille does not take any op‐
2919 tions.
2920 -chardev tty,id=id,path=path
2922 tty is only available on Linux, Sun, FreeBSD, NetBSD, OpenBSD
2923 and DragonFlyBSD hosts. It is an alias for serial.
2924 path specifies the path to the tty. path is required.
2926 -chardev parallel,id=id,path=path
2928 -chardev parport,id=id,path=path
2931 parallel is only available on Linux, FreeBSD and DragonFlyBSD
2932 hosts.
2933 Connect to a local parallel port.
2935 path specifies the path to the parallel port device. path is re‐
2937 quired.
2938 -chardev spicevmc,id=id,debug=debug,name=name
2940 spicevmc is only available when spice support is built in.
2941 debug debug level for spicevmc
2943 name name of spice channel to connect to
2945 Connect to a spice virtual machine channel, such as vdiport.
2947 -chardev spiceport,id=id,debug=debug,name=name
2949 spiceport is only available when spice support is built in.
2950 debug debug level for spicevmc
2952 name name of spice port to connect to
2954 Connect to a spice port, allowing a Spice client to handle the
2956 traffic identified by a name (preferably a fqdn).
2957 TPM device options
2959 The general form of a TPM device option is:
2960 -tpmdev backend,id=id[,options]
2962 The specific backend type will determine the applicable options.
2963 The -tpmdev option creates the TPM backend and requires a -de‐
2964 vice option that specifies the TPM frontend interface model.
2965 Use -tpmdev help to print all available TPM backend types.
2967 The available backends are:
2969 -tpmdev passthrough,id=id,path=path,cancel-path=cancel-path
2971 (Linux-host only) Enable access to the host's TPM using the
2972 passthrough driver.
2973 path specifies the path to the host's TPM device, i.e., on a
2975 Linux host this would be /dev/tpm0. path is optional and by de‐
2976 fault /dev/tpm0 is used.
2977 cancel-path specifies the path to the host TPM device's sysfs
2979 entry allowing for cancellation of an ongoing TPM command. can‐
2980 cel-path is optional and by default QEMU will search for the
2981 sysfs entry to use.
2982 Some notes about using the host's TPM with the passthrough
2984 driver:
2985 The TPM device accessed by the passthrough driver must not be
2987 used by any other application on the host.
2988 Since the host's firmware (BIOS/UEFI) has already initialized
2990 the TPM, the VM's firmware (BIOS/UEFI) will not be able to ini‐
2991 tialize the TPM again and may therefore not show a TPM-specific
2992 menu that would otherwise allow the user to configure the TPM,
2993 e.g., allow the user to enable/disable or activate/deactivate
2994 the TPM. Further, if TPM ownership is released from within a VM
2995 then the host's TPM will get disabled and deactivated. To enable
2996 and activate the TPM again afterwards, the host has to be re‐
2997 booted and the user is required to enter the firmware's menu to
2998 enable and activate the TPM. If the TPM is left disabled and/or
2999 deactivated most TPM commands will fail.
3000 To create a passthrough TPM use the following two options:
3002 -tpmdev passthrough,id=tpm0 -device tpm-tis,tpmdev=tpm0
3004 Note that the -tpmdev id is tpm0 and is referenced by tp‐
3006 mdev=tpm0 in the device option.
3007 -tpmdev emulator,id=id,chardev=dev
3009 (Linux-host only) Enable access to a TPM emulator using Unix do‐
3010 main socket based chardev backend.
3011 chardev specifies the unique ID of a character device backend
3013 that provides connection to the software TPM server.
3014 To create a TPM emulator backend device with chardev socket
3016 backend:
3017 -chardev socket,id=chrtpm,path=/tmp/swtpm-sock -tpmdev emulator,id=tpm0,chardev=chrtpm -device tpm-tis,tpmdev=tpm0
3019 Linux/Multiboot boot specific
3021 When using these options, you can use a given Linux or Multiboot kernel
3022 without installing it in the disk image. It can be useful for easier
3023 testing of various kernels.
3024 -kernel bzImage
3026 Use bzImage as kernel image. The kernel can be either a Linux
3027 kernel or in multiboot format.
3028 -append cmdline
3030 Use cmdline as kernel command line
3031 -initrd file
3033 Use file as initial ram disk.
3034 -initrd "file1 arg=foo,file2"
3036 This syntax is only available with multiboot.
3037 Use file1 and file2 as modules and pass arg=foo as parameter to
3039 the first module.
3040 -dtb file
3042 Use file as a device tree binary (dtb) image and pass it to the
3043 kernel on boot.
3044 Debug/Expert options
3046 -compat [deprecated-input=@var{input-policy}][,deprecated-out‐
3047 put=@var{output-policy}]
3048 Set policy for handling deprecated management interfaces (exper‐
3049 imental):
3050 deprecated-input=accept (default)
3052 Accept deprecated commands and arguments
3053 deprecated-input=reject
3055 Reject deprecated commands and arguments
3056 deprecated-input=crash
3058 Crash on deprecated commands and arguments
3059 deprecated-output=accept (default)
3061 Emit deprecated command results and events
3062 deprecated-output=hide
3064 Suppress deprecated command results and events
3065 Limitation: covers only syntactic aspects of QMP.
3067 -compat [unstable-input=@var{input-policy}][,unstable-output=@var{out‐
3069 put-policy}]
3070 Set policy for handling unstable management interfaces (experi‐
3071 mental):
3072 unstable-input=accept (default)
3074 Accept unstable commands and arguments
3075 unstable-input=reject
3077 Reject unstable commands and arguments
3078 unstable-input=crash
3080 Crash on unstable commands and arguments
3081 unstable-output=accept (default)
3083 Emit unstable command results and events
3084 unstable-output=hide
3086 Suppress unstable command results and events
3087 Limitation: covers only syntactic aspects of QMP.
3089 -fw_cfg [name=]name,file=file
3091 Add named fw_cfg entry with contents from file file.
3092 -fw_cfg [name=]name,string=str
3094 Add named fw_cfg entry with contents from string str.
3095 The terminating NUL character of the contents of str will not be
3097 included as part of the fw_cfg item data. To insert contents
3098 with embedded NUL characters, you have to use the file parame‐
3099 ter.
3100 The fw_cfg entries are passed by QEMU through to the guest.
3102 Example:
3104 -fw_cfg name=opt/com.mycompany/blob,file=./my_blob.bin
3106 creates an fw_cfg entry named opt/com.mycompany/blob with con‐
3108 tents from ./my_blob.bin.
3109 -serial dev
3111 Redirect the virtual serial port to host character device dev.
3112 The default device is vc in graphical mode and stdio in non
3113 graphical mode.
3114 This option can be used several times to simulate up to 4 serial
3116 ports.
3117 Use -serial none to disable all serial ports.
3119 Available character devices are:
3121 vc[:WxH]
3123 Virtual console. Optionally, a width and height can be
3124 given in pixel with
3125 vc:800x600
3127 It is also possible to specify width or height in charac‐
3129 ters:
3130 vc:80Cx24C
3132 pty [Linux only] Pseudo TTY (a new PTY is automatically allo‐
3134 cated)
3135 none No device is allocated.
3137 null void device
3139 chardev:id
3141 Use a named character device defined with the -chardev
3142 option.
3143 /dev/XXX
3145 [Linux only] Use host tty, e.g. /dev/ttyS0. The host se‐
3146 rial port parameters are set according to the emulated
3147 ones.
3148 /dev/parportN
3150 [Linux only, parallel port only] Use host parallel port
3151 N. Currently SPP and EPP parallel port features can be
3152 used.
3153 file:filename
3155 Write output to filename. No character can be read.
3156 stdio [Unix only] standard input/output
3158 pipe:filename
3160 name pipe filename
3161 COMn [Windows only] Use host serial port n
3163 udp:[remote_host]:remote_port[@[src_ip]:src_port]
3165 This implements UDP Net Console. When remote_host or
3166 src_ip are not specified they default to 0.0.0.0. When
3167 not using a specified src_port a random port is automati‐
3168 cally chosen.
3169 If you just want a simple readonly console you can use
3171 netcat or nc, by starting QEMU with: -serial udp::4555
3172 and nc as: nc -u -l -p 4555. Any time QEMU writes some‐
3173 thing to that port it will appear in the netconsole ses‐
3174 sion.
3175 If you plan to send characters back via netconsole or you
3177 want to stop and start QEMU a lot of times, you should
3178 have QEMU use the same source port each time by using
3179 something like -serial udp::4555@:4556 to QEMU. Another
3180 approach is to use a patched version of netcat which can
3181 listen to a TCP port and send and receive characters via
3182 udp. If you have a patched version of netcat which acti‐
3183 vates telnet remote echo and single char transfer, then
3184 you can use the following options to set up a netcat
3185 redirector to allow telnet on port 5555 to access the
3186 QEMU port.
3187 QEMU Options:
3189 -serial udp::4555@:4556
3190 netcat options:
3192 -u -P 4555 -L 0.0.0.0:4556 -t -p 5555 -I -T
3193 telnet options:
3195 localhost 5555
3196 tcp:[host]:port[,server=on|off][,wait=on|off][,node‐
3198 lay=on|off][,reconnect=seconds]
3199 The TCP Net Console has two modes of operation. It can
3200 send the serial I/O to a location or wait for a connec‐
3201 tion from a location. By default the TCP Net Console is
3202 sent to host at the port. If you use the server=on option
3203 QEMU will wait for a client socket application to connect
3204 to the port before continuing, unless the wait=on|off op‐
3205 tion was specified. The nodelay=on|off option disables
3206 the Nagle buffering algorithm. The reconnect=on option
3207 only applies if server=no is set, if the connection goes
3208 down it will attempt to reconnect at the given interval.
3209 If host is omitted, 0.0.0.0 is assumed. Only one TCP con‐
3210 nection at a time is accepted. You can use telnet=on to
3211 connect to the corresponding character device.
3212 Example to send tcp console to 192.168.0.2 port 4444
3214 -serial tcp:192.168.0.2:4444
3215 Example to listen and wait on port 4444 for connection
3217 -serial tcp::4444,server=on
3218 Example to not wait and listen on ip 192.168.0.100 port
3220 4444
3221 -serial tcp:192.168.0.100:4444,server=on,wait=off
3222 telnet:host:port[,server=on|off][,wait=on|off][,nodelay=on|off]
3224 The telnet protocol is used instead of raw tcp sockets.
3225 The options work the same as if you had specified -serial
3226 tcp. The difference is that the port acts like a telnet
3227 server or client using telnet option negotiation. This
3228 will also allow you to send the MAGIC_SYSRQ sequence if
3229 you use a telnet that supports sending the break se‐
3230 quence. Typically in unix telnet you do it with Control-]
3231 and then type "send break" followed by pressing the enter
3232 key.
3233 websocket:host:port,server=on[,wait=on|off][,nodelay=on|off]
3235 The WebSocket protocol is used instead of raw tcp socket.
3236 The port acts as a WebSocket server. Client mode is not
3237 supported.
3238 unix:path[,server=on|off][,wait=on|off][,reconnect=seconds]
3240 A unix domain socket is used instead of a tcp socket. The
3241 option works the same as if you had specified -serial tcp
3242 except the unix domain socket path is used for connec‐
3243 tions.
3244 mon:dev_string
3246 This is a special option to allow the monitor to be mul‐
3247 tiplexed onto another serial port. The monitor is ac‐
3248 cessed with key sequence of Control-a and then pressing
3249 c. dev_string should be any one of the serial devices
3250 specified above. An example to multiplex the monitor onto
3251 a telnet server listening on port 4444 would be:
3252 -serial mon:telnet::4444,server=on,wait=off
3254 When the monitor is multiplexed to stdio in this way,
3256 Ctrl+C will not terminate QEMU any more but will be
3257 passed to the guest instead.
3258 braille
3260 Braille device. This will use BrlAPI to display the
3261 braille output on a real or fake device.
3262 msmouse
3264 Three button serial mouse. Configure the guest to use Mi‐
3265 crosoft protocol.
3266 -parallel dev
3268 Redirect the virtual parallel port to host device dev (same de‐
3269 vices as the serial port). On Linux hosts, /dev/parportN can be
3270 used to use hardware devices connected on the corresponding host
3271 parallel port.
3272 This option can be used several times to simulate up to 3 paral‐
3274 lel ports.
3275 Use -parallel none to disable all parallel ports.
3277 -monitor dev
3279 Redirect the monitor to host device dev (same devices as the se‐
3280 rial port). The default device is vc in graphical mode and stdio
3281 in non graphical mode. Use -monitor none to disable the default
3282 monitor.
3283 -qmp dev
3285 Like -monitor but opens in 'control' mode.
3286 -qmp-pretty dev
3288 Like -qmp but uses pretty JSON formatting.
3289 -mon [chardev=]name[,mode=readline|control][,pretty[=on|off]]
3291 Setup monitor on chardev name. mode=control configures a QMP
3292 monitor (a JSON RPC-style protocol) and it is not the same as
3293 HMP, the human monitor that has a "(qemu)" prompt. pretty is
3294 only valid when mode=control, turning on JSON pretty printing to
3295 ease human reading and debugging.
3296 -debugcon dev
3298 Redirect the debug console to host device dev (same devices as
3299 the serial port). The debug console is an I/O port which is typ‐
3300 ically port 0xe9; writing to that I/O port sends output to this
3301 device. The default device is vc in graphical mode and stdio in
3302 non graphical mode.
3303 -pidfile file
3305 Store the QEMU process PID in file. It is useful if you launch
3306 QEMU from a script.
3307 -singlestep
3309 Run the emulation in single step mode.
3310 --preconfig
3312 Pause QEMU for interactive configuration before the machine is
3313 created, which allows querying and configuring properties that
3314 will affect machine initialization. Use QMP command 'x-exit-pre‐
3315 config' to exit the preconfig state and move to the next state
3316 (i.e. run guest if -S isn't used or pause the second time if -S
3317 is used). This option is experimental.
3318 -S Do not start CPU at startup (you must type 'c' in the monitor).
3320 -overcommit mem-lock=on|off
3322 -overcommit cpu-pm=on|off
3325 Run qemu with hints about host resource overcommit. The default
3326 is to assume that host overcommits all resources.
3327 Locking qemu and guest memory can be enabled via mem-lock=on
3329 (disabled by default). This works when host memory is not over‐
3330 committed and reduces the worst-case latency for guest.
3331 Guest ability to manage power state of host cpus (increasing la‐
3333 tency for other processes on the same host cpu, but decreasing
3334 latency for guest) can be enabled via cpu-pm=on (disabled by de‐
3335 fault). This works best when host CPU is not overcommitted. When
3336 used, host estimates of CPU cycle and power utilization will be
3337 incorrect, not taking into account guest idle time.
3338 -gdb dev
3340 Accept a gdb connection on device dev (see the GDB usage chapter
3341 in the System Emulation Users Guide). Note that this option does
3342 not pause QEMU execution -- if you want QEMU to not start the
3343 guest until you connect with gdb and issue a continue command,
3344 you will need to also pass the -S option to QEMU.
3345 The most usual configuration is to listen on a local TCP socket:
3347 -gdb tcp::3117
3349 but you can specify other backends; UDP, pseudo TTY, or even
3351 stdio are all reasonable use cases. For example, a stdio connec‐
3352 tion allows you to start QEMU from within gdb and establish the
3353 connection via a pipe:
3354 (gdb) target remote | exec qemu-system-x86_64 -gdb stdio ...
3356 -s Shorthand for -gdb tcp::1234, i.e. open a gdbserver on TCP port
3358 1234 (see the GDB usage chapter in the System Emulation Users
3359 Guide).
3360 -d item1[,...]
3362 Enable logging of specified items. Use '-d help' for a list of
3363 log items.
3364 -D logfile
3366 Output log in logfile instead of to stderr
3367 -dfilter range1[,...]
3369 Filter debug output to that relevant to a range of target ad‐
3370 dresses. The filter spec can be either start+size, start-size
3371 or start..end where start end and size are the addresses and
3372 sizes required. For example:
3373 -dfilter 0x8000..0x8fff,0xffffffc000080000+0x200,0xffffffc000060000-0x1000
3375 Will dump output for any code in the 0x1000 sized block starting
3377 at 0x8000 and the 0x200 sized block starting at
3378 0xffffffc000080000 and another 0x1000 sized block starting at
3379 0xffffffc00005f000.
3380 -seed number
3382 Force the guest to use a deterministic pseudo-random number gen‐
3383 erator, seeded with number. This does not affect crypto routines
3384 within the host.
3385 -L path
3387 Set the directory for the BIOS, VGA BIOS and keymaps.
3388 To list all the data directories, use -L help.
3390 -bios file
3392 Set the filename for the BIOS.
3393 -enable-kvm
3395 Enable KVM full virtualization support. This option is only
3396 available if KVM support is enabled when compiling.
3397 -xen-domid id
3399 Specify xen guest domain id (XEN only).
3400 -xen-attach
3402 Attach to existing xen domain. libxl will use this when starting
3403 QEMU (XEN only). Restrict set of available xen operations to
3404 specified domain id (XEN only).
3405 -no-reboot
3407 Exit instead of rebooting.
3408 -no-shutdown
3410 Don't exit QEMU on guest shutdown, but instead only stop the em‐
3411 ulation. This allows for instance switching to monitor to commit
3412 changes to the disk image.
3413 -action event=action
3415 The action parameter serves to modify QEMU's default behavior
3416 when certain guest events occur. It provides a generic method
3417 for specifying the same behaviors that are modified by the
3418 -no-reboot and -no-shutdown parameters.
3419 Examples:
3421 -action panic=none -action reboot=shutdown,shutdown=pause
3423 -watchdog i6300esb -action watchdog=pause
3424 -loadvm file
3426 Start right away with a saved state (loadvm in monitor)
3427 -daemonize
3429 Daemonize the QEMU process after initialization. QEMU will not
3430 detach from standard IO until it is ready to receive connections
3431 on any of its devices. This option is a useful way for external
3432 programs to launch QEMU without having to cope with initializa‐
3433 tion race conditions.
3434 -option-rom file
3436 Load the contents of file as an option ROM. This option is use‐
3437 ful to load things like EtherBoot.
3438 -rtc [base=utc|localtime|datetime][,clock=host|rt|vm][,drift‐
3440 fix=none|slew]
3441 Specify base as utc or localtime to let the RTC start at the
3442 current UTC or local time, respectively. localtime is required
3443 for correct date in MS-DOS or Windows. To start at a specific
3444 point in time, provide datetime in the format
3445 2006-06-17T16:01:21 or 2006-06-17. The default base is UTC.
3446 By default the RTC is driven by the host system time. This al‐
3448 lows using of the RTC as accurate reference clock inside the
3449 guest, specifically if the host time is smoothly following an
3450 accurate external reference clock, e.g. via NTP. If you want to
3451 isolate the guest time from the host, you can set clock to rt
3452 instead, which provides a host monotonic clock if host support
3453 it. To even prevent the RTC from progressing during suspension,
3454 you can set clock to vm (virtual clock). 'clock=vm' is recom‐
3455 mended especially in icount mode in order to preserve determin‐
3456 ism; however, note that in icount mode the speed of the virtual
3457 clock is variable and can in general differ from the host clock.
3458 Enable driftfix (i386 targets only) if you experience time drift
3460 problems, specifically with Windows' ACPI HAL. This option will
3461 try to figure out how many timer interrupts were not processed
3462 by the Windows guest and will re-inject them.
3463 -icount [shift=N|auto][,align=on|off][,sleep=on|off][,rr=record|re‐
3465 play,rrfile=filename[,rrsnapshot=snapshot]]
3466 Enable virtual instruction counter. The virtual cpu will execute
3467 one instruction every 2^N ns of virtual time. If auto is speci‐
3468 fied then the virtual cpu speed will be automatically adjusted
3469 to keep virtual time within a few seconds of real time.
3470 Note that while this option can give deterministic behavior, it
3472 does not provide cycle accurate emulation. Modern CPUs contain
3473 superscalar out of order cores with complex cache hierarchies.
3474 The number of instructions executed often has little or no cor‐
3475 relation with actual performance.
3476 When the virtual cpu is sleeping, the virtual time will advance
3478 at default speed unless sleep=on is specified. With sleep=on,
3479 the virtual time will jump to the next timer deadline instantly
3480 whenever the virtual cpu goes to sleep mode and will not advance
3481 if no timer is enabled. This behavior gives deterministic execu‐
3482 tion times from the guest point of view. The default if icount
3483 is enabled is sleep=off. sleep=on cannot be used together with
3484 either shift=auto or align=on.
3485 align=on will activate the delay algorithm which will try to
3487 synchronise the host clock and the virtual clock. The goal is to
3488 have a guest running at the real frequency imposed by the shift
3489 option. Whenever the guest clock is behind the host clock and if
3490 align=on is specified then we print a message to the user to in‐
3491 form about the delay. Currently this option does not work when
3492 shift is auto. Note: The sync algorithm will work for those
3493 shift values for which the guest clock runs ahead of the host
3494 clock. Typically this happens when the shift value is high (how
3495 high depends on the host machine). The default if icount is en‐
3496 abled is align=off.
3497 When the rr option is specified deterministic record/replay is
3499 enabled. The rrfile= option must also be provided to specify the
3500 path to the replay log. In record mode data is written to this
3501 file, and in replay mode it is read back. If the rrsnapshot op‐
3502 tion is given then it specifies a VM snapshot name. In record
3503 mode, a new VM snapshot with the given name is created at the
3504 start of execution recording. In replay mode this option speci‐
3505 fies the snapshot name used to load the initial VM state.
3506 -watchdog model
3508 Create a virtual hardware watchdog device. Once enabled (by a
3509 guest action), the watchdog must be periodically polled by an
3510 agent inside the guest or else the guest will be restarted.
3511 Choose a model for which your guest has drivers.
3512 The model is the model of hardware watchdog to emulate. Use
3514 -watchdog help to list available hardware models. Only one
3515 watchdog can be enabled for a guest.
3516 The following models may be available:
3518 ib700 iBASE 700 is a very simple ISA watchdog with a single
3520 timer.
3521 i6300esb
3523 Intel 6300ESB I/O controller hub is a much more feature‐
3524 ful PCI-based dual-timer watchdog.
3525 diag288
3527 A virtual watchdog for s390x backed by the diagnose 288
3528 hypercall (currently KVM only).
3529 -watchdog-action action
3531 The action controls what QEMU will do when the watchdog timer
3532 expires. The default is reset (forcefully reset the guest).
3533 Other possible actions are: shutdown (attempt to gracefully
3534 shutdown the guest), poweroff (forcefully poweroff the guest),
3535 inject-nmi (inject a NMI into the guest), pause (pause the
3536 guest), debug (print a debug message and continue), or none (do
3537 nothing).
3538 Note that the shutdown action requires that the guest responds
3540 to ACPI signals, which it may not be able to do in the sort of
3541 situations where the watchdog would have expired, and thus
3542 -watchdog-action shutdown is not recommended for production use.
3543 Examples:
3545 -watchdog i6300esb -watchdog-action pause; -watchdog ib700
3547 -echr numeric_ascii_value
3549 Change the escape character used for switching to the monitor
3550 when using monitor and serial sharing. The default is 0x01 when
3551 using the -nographic option. 0x01 is equal to pressing Con‐
3552 trol-a. You can select a different character from the ascii con‐
3553 trol keys where 1 through 26 map to Control-a through Control-z.
3554 For instance you could use the either of the following to change
3555 the escape character to Control-t.
3556 -echr 0x14; -echr 20
3558 -incoming tcp:[host]:port[,to=maxport][,ipv4=on|off][,ipv6=on|off]
3560 -incoming rdma:host:port[,ipv4=on|off][,ipv6=on|off]
3563 Prepare for incoming migration, listen on a given tcp port.
3564 -incoming unix:socketpath
3566 Prepare for incoming migration, listen on a given unix socket.
3567 -incoming fd:fd
3569 Accept incoming migration from a given filedescriptor.
3570 -incoming exec:cmdline
3572 Accept incoming migration as an output from specified external
3573 command.
3574 -incoming defer
3576 Wait for the URI to be specified via migrate_incoming. The moni‐
3577 tor can be used to change settings (such as migration parame‐
3578 ters) prior to issuing the migrate_incoming to allow the migra‐
3579 tion to begin.
3580 -only-migratable
3582 Only allow migratable devices. Devices will not be allowed to
3583 enter an unmigratable state.
3584 -nodefaults
3586 Don't create default devices. Normally, QEMU sets the default
3587 devices like serial port, parallel port, virtual console, moni‐
3588 tor device, VGA adapter, floppy and CD-ROM drive and others. The
3589 -nodefaults option will disable all those default devices.
3590 -chroot dir
3592 Immediately before starting guest execution, chroot to the spec‐
3593 ified directory. Especially useful in combination with -runas.
3594 -runas user
3596 Immediately before starting guest execution, drop root privi‐
3597 leges, switching to the specified user.
3598 -prom-env variable=value
3600 Set OpenBIOS nvram variable to given value (PPC, SPARC only).
3601 qemu-system-sparc -prom-env 'auto-boot?=false' \
3603 -prom-env 'boot-device=sd(0,2,0):d' -prom-env 'boot-args=linux single'
3604 qemu-system-ppc -prom-env 'auto-boot?=false' \
3606 -prom-env 'boot-device=hd:2,\yaboot' \
3607 -prom-env 'boot-args=conf=hd:2,\yaboot.conf'
3608 -semihosting
3610 Enable semihosting mode (ARM, M68K, Xtensa, MIPS, Nios II,
3611 RISC-V only).
3612 Note that this allows guest direct access to the host filesys‐
3614 tem, so should only be used with a trusted guest OS.
3615 See the -semihosting-config option documentation for further in‐
3617 formation about the facilities this enables.
3618 -semihosting-config [enable=on|off][,target=na‐
3620 tive|gdb|auto][,chardev=id][,arg=str[,...]]
3621 Enable and configure semihosting (ARM, M68K, Xtensa, MIPS, Nios
3622 II, RISC-V only).
3623 Note that this allows guest direct access to the host filesys‐
3625 tem, so should only be used with a trusted guest OS.
3626 On Arm this implements the standard semihosting API, version
3628 2.0.
3629 On M68K this implements the "ColdFire GDB" interface used by
3631 libgloss.
3632 Xtensa semihosting provides basic file IO calls, such as
3634 open/read/write/seek/select. Tensilica baremetal libc for ISS
3635 and linux platform "sim" use this interface.
3636 On RISC-V this implements the standard semihosting API, version
3638 0.2.
3639 target=native|gdb|auto
3641 Defines where the semihosting calls will be addressed, to
3642 QEMU (native) or to GDB (gdb). The default is auto, which
3643 means gdb during debug sessions and native otherwise.
3644 chardev=str1
3646 Send the output to a chardev backend output for native or
3647 auto output when not in gdb
3648 arg=str1,arg=str2,...
3650 Allows the user to pass input arguments, and can be used
3651 multiple times to build up a list. The old-style -ker‐
3652 nel/-append method of passing a command line is still
3653 supported for backward compatibility. If both the --semi‐
3654 hosting-config arg and the -kernel/-append are specified,
3655 the former is passed to semihosting as it always takes
3656 precedence.
3657 -old-param
3659 Old param mode (ARM only).
3660 -sandbox arg[,obsolete=string][,elevateprivi‐
3662 leges=string][,spawn=string][,resourcecontrol=string]
3663 Enable Seccomp mode 2 system call filter. 'on' will enable
3664 syscall filtering and 'off' will disable it. The default is
3665 'off'.
3666 obsolete=string
3668 Enable Obsolete system calls
3669 elevateprivileges=string
3671 Disable set*uid|gid system calls
3672 spawn=string
3674 Disable *fork and execve
3675 resourcecontrol=string
3677 Disable process affinity and schedular priority
3678 -readconfig file
3680 Read device configuration from file. This approach is useful
3681 when you want to spawn QEMU process with many command line op‐
3682 tions but you don't want to exceed the command line character
3683 limit.
3684 -no-user-config
3686 The -no-user-config option makes QEMU not load any of the
3687 user-provided config files on sysconfdir.
3688 -trace [[enable=]pattern][,events=file][,file=file]
3690 Specify tracing options.
3691 [enable=]PATTERN
3693 Immediately enable events matching PATTERN (either event name
3694 or a globbing pattern). This option is only available if
3695 QEMU has been compiled with the simple, log or ftrace tracing
3696 backend. To specify multiple events or patterns, specify the
3697 -trace option multiple times.
3698 Use -trace help to print a list of names of trace points.
3700 events=FILE
3702 Immediately enable events listed in FILE. The file must con‐
3703 tain one event name (as listed in the trace-events-all file)
3704 per line; globbing patterns are accepted too. This option is
3705 only available if QEMU has been compiled with the simple, log
3706 or ftrace tracing backend.
3707 file=FILE
3709 Log output traces to FILE. This option is only available if
3710 QEMU has been compiled with the simple tracing backend.
3711 -plugin file=file[,argname=argvalue]
3713 Load a plugin.
3714 file=file
3716 Load the given plugin from a shared library file.
3717 argname=argvalue
3719 Argument passed to the plugin. (Can be given multiple
3720 times.)
3721 -enable-fips
3723 Enable FIPS 140-2 compliance mode.
3724 -msg [timestamp[=on|off]][,guest-name[=on|off]]
3726 Control error message format.
3727 timestamp=on|off
3729 Prefix messages with a timestamp. Default is off.
3730 guest-name=on|off
3732 Prefix messages with guest name but only if -name guest
3733 option is set otherwise the option is ignored. Default is
3734 off.
3735 -dump-vmstate file
3737 Dump json-encoded vmstate information for current machine type
3738 to file in file
3739 -enable-sync-profile
3741 Enable synchronization profiling.
3742 Generic object creation
3744 -object typename[,prop1=value1,...]
3745 Create a new object of type typename setting properties in the
3746 order they are specified. Note that the 'id' property must be
3747 set. These objects are placed in the '/objects' path.
3748 -object memory-back‐
3750 end-file,id=id,size=size,mem-path=dir,share=on|off,dis‐
3751 card-data=on|off,merge=on|off,dump=on|off,preal‐
3752 loc=on|off,host-nodes=host-nodes,policy=default|pre‐
3753 ferred|bind|interleave,align=align,readonly=on|off
3754 Creates a memory file backend object, which can be used
3755 to back the guest RAM with huge pages.
3756 The id parameter is a unique ID that will be used to ref‐
3758 erence this memory region in other parameters, e.g.
3759 -numa, -device nvdimm, etc.
3760 The size option provides the size of the memory region,
3762 and accepts common suffixes, e.g. 500M.
3763 The mem-path provides the path to either a shared memory
3765 or huge page filesystem mount.
3766 The share boolean option determines whether the memory
3768 region is marked as private to QEMU, or shared. The lat‐
3769 ter allows a co-operating external process to access the
3770 QEMU memory region.
3771 The share is also required for pvrdma devices due to lim‐
3773 itations in the RDMA API provided by Linux.
3774 Setting share=on might affect the ability to configure
3776 NUMA bindings for the memory backend under some circum‐
3777 stances, see Documentation/vm/numa_memory_policy.txt on
3778 the Linux kernel source tree for additional details.
3779 Setting the discard-data boolean option to on indicates
3781 that file contents can be destroyed when QEMU exits, to
3782 avoid unnecessarily flushing data to the backing file.
3783 Note that discard-data is only an optimization, and QEMU
3784 might not discard file contents if it aborts unexpectedly
3785 or is terminated using SIGKILL.
3786 The merge boolean option enables memory merge, also known
3788 as MADV_MERGEABLE, so that Kernel Samepage Merging will
3789 consider the pages for memory deduplication.
3790 Setting the dump boolean option to off excludes the mem‐
3792 ory from core dumps. This feature is also known as
3793 MADV_DONTDUMP.
3794 The prealloc boolean option enables memory preallocation.
3796 The host-nodes option binds the memory range to a list of
3798 NUMA host nodes.
3799 The policy option sets the NUMA policy to one of the fol‐
3801 lowing values:
3802 default
3804 default host policy
3805 preferred
3807 prefer the given host node list for allocation
3808 bind restrict memory allocation to the given host node
3810 list
3811 interleave
3813 interleave memory allocations across the given
3814 host node list
3815 The align option specifies the base address alignment
3817 when QEMU mmap(2) mem-path, and accepts common suffixes,
3818 eg 2M. Some backend store specified by mem-path requires
3819 an alignment different than the default one used by QEMU,
3820 eg the device DAX /dev/dax0.0 requires 2M alignment
3821 rather than 4K. In such cases, users can specify the re‐
3822 quired alignment via this option.
3823 The pmem option specifies whether the backing file speci‐
3825 fied by mem-path is in host persistent memory that can be
3826 accessed using the SNIA NVM programming model (e.g. Intel
3827 NVDIMM). If pmem is set to 'on', QEMU will take necessary
3828 operations to guarantee the persistence of its own writes
3829 to mem-path (e.g. in vNVDIMM label emulation and live mi‐
3830 gration). Also, we will map the backend-file with
3831 MAP_SYNC flag, which ensures the file metadata is in sync
3832 for mem-path in case of host crash or a power failure.
3833 MAP_SYNC requires support from both the host kernel
3834 (since Linux kernel 4.15) and the filesystem of mem-path
3835 mounted with DAX option.
3836 The readonly option specifies whether the backing file is
3838 opened read-only or read-write (default).
3839 -object memory-back‐
3841 end-ram,id=id,merge=on|off,dump=on|off,share=on|off,preal‐
3842 loc=on|off,size=size,host-nodes=host-nodes,policy=default|pre‐
3843 ferred|bind|interleave
3844 Creates a memory backend object, which can be used to
3845 back the guest RAM. Memory backend objects offer more
3846 control than the -m option that is traditionally used to
3847 define guest RAM. Please refer to memory-backend-file
3848 for a description of the options.
3849 -object memory-back‐
3851 end-memfd,id=id,merge=on|off,dump=on|off,share=on|off,preal‐
3852 loc=on|off,size=size,host-nodes=host-nodes,policy=default|pre‐
3853 ferred|bind|interleave,seal=on|off,hugetlb=on|off,hugetlb‐
3854 size=size
3855 Creates an anonymous memory file backend object, which
3856 allows QEMU to share the memory with an external process
3857 (e.g. when using vhost-user). The memory is allocated
3858 with memfd and optional sealing. (Linux only)
3859 The seal option creates a sealed-file, that will block
3861 further resizing the memory ('on' by default).
3862 The hugetlb option specify the file to be created resides
3864 in the hugetlbfs filesystem (since Linux 4.14). Used in
3865 conjunction with the hugetlb option, the hugetlbsize op‐
3866 tion specify the hugetlb page size on systems that sup‐
3867 port multiple hugetlb page sizes (it must be a power of 2
3868 value supported by the system).
3869 In some versions of Linux, the hugetlb option is incom‐
3871 patible with the seal option (requires at least Linux
3872 4.16).
3873 Please refer to memory-backend-file for a description of
3875 the other options.
3876 The share boolean option is on by default with memfd.
3878 -object rng-builtin,id=id
3880 Creates a random number generator backend which obtains
3881 entropy from QEMU builtin functions. The id parameter is
3882 a unique ID that will be used to reference this entropy
3883 backend from the virtio-rng device. By default, the vir‐
3884 tio-rng device uses this RNG backend.
3885 -object rng-random,id=id,filename=/dev/random
3887 Creates a random number generator backend which obtains
3888 entropy from a device on the host. The id parameter is a
3889 unique ID that will be used to reference this entropy
3890 backend from the virtio-rng device. The filename parame‐
3891 ter specifies which file to obtain entropy from and if
3892 omitted defaults to /dev/urandom.
3893 -object rng-egd,id=id,chardev=chardevid
3895 Creates a random number generator backend which obtains
3896 entropy from an external daemon running on the host. The
3897 id parameter is a unique ID that will be used to refer‐
3898 ence this entropy backend from the virtio-rng device. The
3899 chardev parameter is the unique ID of a character device
3900 backend that provides the connection to the RNG daemon.
3901 -object tls-creds-anon,id=id,endpoint=end‐
3903 point,dir=/path/to/cred/dir,verify-peer=on|off
3904 Creates a TLS anonymous credentials object, which can be
3905 used to provide TLS support on network backends. The id
3906 parameter is a unique ID which network backends will use
3907 to access the credentials. The endpoint is either server
3908 or client depending on whether the QEMU network backend
3909 that uses the credentials will be acting as a client or
3910 as a server. If verify-peer is enabled (the default) then
3911 once the handshake is completed, the peer credentials
3912 will be verified, though this is a no-op for anonymous
3913 credentials.
3914 The dir parameter tells QEMU where to find the credential
3916 files. For server endpoints, this directory may contain
3917 a file dh-params.pem providing diffie-hellman parameters
3918 to use for the TLS server. If the file is missing, QEMU
3919 will generate a set of DH parameters at startup. This is
3920 a computationally expensive operation that consumes ran‐
3921 dom pool entropy, so it is recommended that a persistent
3922 set of parameters be generated upfront and saved.
3923 -object tls-creds-psk,id=id,endpoint=end‐
3925 point,dir=/path/to/keys/dir[,username=username]
3926 Creates a TLS Pre-Shared Keys (PSK) credentials object,
3927 which can be used to provide TLS support on network back‐
3928 ends. The id parameter is a unique ID which network back‐
3929 ends will use to access the credentials. The endpoint is
3930 either server or client depending on whether the QEMU
3931 network backend that uses the credentials will be acting
3932 as a client or as a server. For clients only, username
3933 is the username which will be sent to the server. If
3934 omitted it defaults to "qemu".
3935 The dir parameter tells QEMU where to find the keys file.
3937 It is called "dir/keys.psk" and contains "username:key"
3938 pairs. This file can most easily be created using the
3939 GnuTLS psktool program.
3940 For server endpoints, dir may also contain a file
3942 dh-params.pem providing diffie-hellman parameters to use
3943 for the TLS server. If the file is missing, QEMU will
3944 generate a set of DH parameters at startup. This is a
3945 computationally expensive operation that consumes random
3946 pool entropy, so it is recommended that a persistent set
3947 of parameters be generated up front and saved.
3948 -object tls-creds-x509,id=id,endpoint=end‐
3950 point,dir=/path/to/cred/dir,priority=priority,ver‐
3951 ify-peer=on|off,passwordid=id
3952 Creates a TLS anonymous credentials object, which can be
3953 used to provide TLS support on network backends. The id
3954 parameter is a unique ID which network backends will use
3955 to access the credentials. The endpoint is either server
3956 or client depending on whether the QEMU network backend
3957 that uses the credentials will be acting as a client or
3958 as a server. If verify-peer is enabled (the default) then
3959 once the handshake is completed, the peer credentials
3960 will be verified. With x509 certificates, this implies
3961 that the clients must be provided with valid client cer‐
3962 tificates too.
3963 The dir parameter tells QEMU where to find the credential
3965 files. For server endpoints, this directory may contain
3966 a file dh-params.pem providing diffie-hellman parameters
3967 to use for the TLS server. If the file is missing, QEMU
3968 will generate a set of DH parameters at startup. This is
3969 a computationally expensive operation that consumes ran‐
3970 dom pool entropy, so it is recommended that a persistent
3971 set of parameters be generated upfront and saved.
3972 For x509 certificate credentials the directory will con‐
3974 tain further files providing the x509 certificates. The
3975 certificates must be stored in PEM format, in filenames
3976 ca-cert.pem, ca-crl.pem (optional), server-cert.pem (only
3977 servers), server-key.pem (only servers), client-cert.pem
3978 (only clients), and client-key.pem (only clients).
3979 For the server-key.pem and client-key.pem files which
3981 contain sensitive private keys, it is possible to use an
3982 encrypted version by providing the passwordid parameter.
3983 This provides the ID of a previously created secret ob‐
3984 ject containing the password for decryption.
3985 The priority parameter allows to override the global de‐
3987 fault priority used by gnutls. This can be useful if the
3988 system administrator needs to use a weaker set of crypto
3989 priorities for QEMU without potentially forcing the weak‐
3990 ness onto all applications. Or conversely if one wants
3991 wants a stronger default for QEMU than for all other ap‐
3992 plications, they can do this through this parameter. Its
3993 format is a gnutls priority string as described at
3994 https://gnutls.org/manual/html_node/Priority-Strings.html.
3995 -object tls-cipher-suites,id=id,priority=priority
3997 Creates a TLS cipher suites object, which can be used to
3998 control the TLS cipher/protocol algorithms that applica‐
3999 tions are permitted to use.
4000 The id parameter is a unique ID which frontends will use
4002 to access the ordered list of permitted TLS cipher suites
4003 from the host.
4004 The priority parameter allows to override the global de‐
4006 fault priority used by gnutls. This can be useful if the
4007 system administrator needs to use a weaker set of crypto
4008 priorities for QEMU without potentially forcing the weak‐
4009 ness onto all applications. Or conversely if one wants
4010 wants a stronger default for QEMU than for all other ap‐
4011 plications, they can do this through this parameter. Its
4012 format is a gnutls priority string as described at
4013 https://gnutls.org/manual/html_node/Priority-Strings.html.
4014 An example of use of this object is to control UEFI HTTPS
4016 Boot. The tls-cipher-suites object exposes the ordered
4017 list of permitted TLS cipher suites from the host side to
4018 the guest firmware, via fw_cfg. The list is represented
4019 as an array of IANA_TLS_CIPHER objects. The firmware uses
4020 the IANA_TLS_CIPHER array for configuring guest-side TLS.
4021 In the following example, the priority at which the
4023 host-side policy is retrieved is given by the priority
4024 property. Given that QEMU uses GNUTLS, priority=@SYSTEM
4025 may be used to refer to /etc/crypto-poli‐
4026 cies/back-ends/gnutls.config.
4027 # qemu-system-x86_64 \
4029 -object tls-cipher-suites,id=mysuite0,priority=@SYSTEM \
4030 -fw_cfg name=etc/edk2/https/ciphers,gen_id=mysuite0
4031 -object filter-buffer,id=id,netdev=netdevid,inter‐
4033 val=t[,queue=all|rx|tx][,status=on|off][,posi‐
4034 tion=head|tail|id=<id>][,insert=behind|before]
4035 Interval t can't be 0, this filter batches the packet de‐
4036 livery: all packets arriving in a given interval on net‐
4037 dev netdevid are delayed until the end of the interval.
4038 Interval is in microseconds. status is optional that in‐
4039 dicate whether the netfilter is on (enabled) or off (dis‐
4040 abled), the default status for netfilter will be 'on'.
4041 queue all|rx|tx is an option that can be applied to any
4043 netfilter.
4044 all: the filter is attached both to the receive and the
4046 transmit queue of the netdev (default).
4047 rx: the filter is attached to the receive queue of the
4049 netdev, where it will receive packets sent to the netdev.
4050 tx: the filter is attached to the transmit queue of the
4052 netdev, where it will receive packets sent by the netdev.
4053 position head|tail|id=<id> is an option to specify where
4055 the filter should be inserted in the filter list. It can
4056 be applied to any netfilter.
4057 head: the filter is inserted at the head of the filter
4059 list, before any existing filters.
4060 tail: the filter is inserted at the tail of the filter
4062 list, behind any existing filters (default).
4063 id=<id>: the filter is inserted before or behind the fil‐
4065 ter specified by <id>, see the insert option below.
4066 insert behind|before is an option to specify where to in‐
4068 sert the new filter relative to the one specified with
4069 position=id=<id>. It can be applied to any netfilter.
4070 before: insert before the specified filter.
4072 behind: insert behind the specified filter (default).
4074 -object filter-mirror,id=id,netdev=netdevid,outdev=charde‐
4076 vid,queue=all|rx|tx[,vnet_hdr_support][,posi‐
4077 tion=head|tail|id=<id>][,insert=behind|before]
4078 filter-mirror on netdev netdevid,mirror net packet to
4079 chardevchardevid, if it has the vnet_hdr_support flag,
4080 filter-mirror will mirror packet with vnet_hdr_len.
4081 -object filter-redirector,id=id,netdev=netdevid,indev=charde‐
4083 vid,outdev=chardevid,queue=all|rx|tx[,vnet_hdr_support][,posi‐
4084 tion=head|tail|id=<id>][,insert=behind|before]
4085 filter-redirector on netdev netdevid,redirect filter's
4086 net packet to chardev chardevid,and redirect indev's
4087 packet to filter.if it has the vnet_hdr_support flag,
4088 filter-redirector will redirect packet with vnet_hdr_len.
4089 Create a filter-redirector we need to differ outdev id
4090 from indev id, id can not be the same. we can just use
4091 indev or outdev, but at least one of indev or outdev need
4092 to be specified.
4093 -object filter-rewriter,id=id,netdev=netde‐
4095 vid,queue=all|rx|tx,[vnet_hdr_support][,posi‐
4096 tion=head|tail|id=<id>][,insert=behind|before]
4097 Filter-rewriter is a part of COLO project.It will rewrite
4098 tcp packet to secondary from primary to keep secondary
4099 tcp connection,and rewrite tcp packet to primary from
4100 secondary make tcp packet can be handled by client.if it
4101 has the vnet_hdr_support flag, we can parse packet with
4102 vnet header.
4103 usage: colo secondary: -object filter-redirec‐
4105 tor,id=f1,netdev=hn0,queue=tx,indev=red0 -object fil‐
4106 ter-redirector,id=f2,netdev=hn0,queue=rx,outdev=red1 -ob‐
4107 ject filter-rewriter,id=rew0,netdev=hn0,queue=all
4108 -object filter-dump,id=id,netdev=dev[,file=file‐
4110 name][,maxlen=len][,position=head|tail|id=<id>][,insert=be‐
4111 hind|before]
4112 Dump the network traffic on netdev dev to the file speci‐
4113 fied by filename. At most len bytes (64k by default) per
4114 packet are stored. The file format is libpcap, so it can
4115 be analyzed with tools such as tcpdump or Wireshark.
4116 -object colo-compare,id=id,primary_in=chardevid,sec‐
4118 ondary_in=chardevid,outdev=chardevid,iothread=id[,vnet_hdr_sup‐
4119 port][,notify_dev=id][,compare_timeout=@var{ms}][,ex‐
4120 pired_scan_cycle=@var{ms}][,max_queue_size=@var{size}]
4121 Colo-compare gets packet from primary_in chardevid and
4122 secondary_in, then compare whether the payload of primary
4123 packet and secondary packet are the same. If same, it
4124 will output primary packet to out_dev, else it will no‐
4125 tify COLO-framework to do checkpoint and send primary
4126 packet to out_dev. In order to improve efficiency, we
4127 need to put the task of comparison in another iothread.
4128 If it has the vnet_hdr_support flag, colo compare will
4129 send/recv packet with vnet_hdr_len. The
4130 compare_timeout=@var{ms} determines the maximum time of
4131 the colo-compare hold the packet. The
4132 expired_scan_cycle=@var{ms} is to set the period of scan‐
4133 ning expired primary node network packets. The
4134 max_queue_size=@var{size} is to set the max compare queue
4135 size depend on user environment. If user want to use Xen
4136 COLO, need to add the notify_dev to notify Xen colo-frame
4137 to do checkpoint.
4138 COLO-compare must be used with the help of filter-mirror,
4140 filter-redirector and filter-rewriter.
4141 KVM COLO
4143 primary:
4145 -netdev tap,id=hn0,vhost=off,script=/etc/qemu-ifup,downscript=/etc/qemu-ifdown
4146 -device e1000,id=e0,netdev=hn0,mac=52:a4:00:12:78:66
4147 -chardev socket,id=mirror0,host=3.3.3.3,port=9003,server=on,wait=off
4148 -chardev socket,id=compare1,host=3.3.3.3,port=9004,server=on,wait=off
4149 -chardev socket,id=compare0,host=3.3.3.3,port=9001,server=on,wait=off
4150 -chardev socket,id=compare0-0,host=3.3.3.3,port=9001
4151 -chardev socket,id=compare_out,host=3.3.3.3,port=9005,server=on,wait=off
4152 -chardev socket,id=compare_out0,host=3.3.3.3,port=9005
4153 -object iothread,id=iothread1
4154 -object filter-mirror,id=m0,netdev=hn0,queue=tx,outdev=mirror0
4155 -object filter-redirector,netdev=hn0,id=redire0,queue=rx,indev=compare_out
4156 -object filter-redirector,netdev=hn0,id=redire1,queue=rx,outdev=compare0
4157 -object colo-compare,id=comp0,primary_in=compare0-0,secondary_in=compare1,outdev=compare_out0,iothread=iothread1
4158 secondary:
4160 -netdev tap,id=hn0,vhost=off,script=/etc/qemu-ifup,down script=/etc/qemu-ifdown
4161 -device e1000,netdev=hn0,mac=52:a4:00:12:78:66
4162 -chardev socket,id=red0,host=3.3.3.3,port=9003
4163 -chardev socket,id=red1,host=3.3.3.3,port=9004
4164 -object filter-redirector,id=f1,netdev=hn0,queue=tx,indev=red0
4165 -object filter-redirector,id=f2,netdev=hn0,queue=rx,outdev=red1
4166 Xen COLO
4169 primary:
4171 -netdev tap,id=hn0,vhost=off,script=/etc/qemu-ifup,downscript=/etc/qemu-ifdown
4172 -device e1000,id=e0,netdev=hn0,mac=52:a4:00:12:78:66
4173 -chardev socket,id=mirror0,host=3.3.3.3,port=9003,server=on,wait=off
4174 -chardev socket,id=compare1,host=3.3.3.3,port=9004,server=on,wait=off
4175 -chardev socket,id=compare0,host=3.3.3.3,port=9001,server=on,wait=off
4176 -chardev socket,id=compare0-0,host=3.3.3.3,port=9001
4177 -chardev socket,id=compare_out,host=3.3.3.3,port=9005,server=on,wait=off
4178 -chardev socket,id=compare_out0,host=3.3.3.3,port=9005
4179 -chardev socket,id=notify_way,host=3.3.3.3,port=9009,server=on,wait=off
4180 -object filter-mirror,id=m0,netdev=hn0,queue=tx,outdev=mirror0
4181 -object filter-redirector,netdev=hn0,id=redire0,queue=rx,indev=compare_out
4182 -object filter-redirector,netdev=hn0,id=redire1,queue=rx,outdev=compare0
4183 -object iothread,id=iothread1
4184 -object colo-compare,id=comp0,primary_in=compare0-0,secondary_in=compare1,outdev=compare_out0,notify_dev=nofity_way,iothread=iothread1
4185 secondary:
4187 -netdev tap,id=hn0,vhost=off,script=/etc/qemu-ifup,down script=/etc/qemu-ifdown
4188 -device e1000,netdev=hn0,mac=52:a4:00:12:78:66
4189 -chardev socket,id=red0,host=3.3.3.3,port=9003
4190 -chardev socket,id=red1,host=3.3.3.3,port=9004
4191 -object filter-redirector,id=f1,netdev=hn0,queue=tx,indev=red0
4192 -object filter-redirector,id=f2,netdev=hn0,queue=rx,outdev=red1
4193 If you want to know the detail of above command line, you
4195 can read the colo-compare git log.
4196 -object cryptodev-backend-builtin,id=id[,queues=queues]
4198 Creates a cryptodev backend which executes crypto
4199 opreation from the QEMU cipher APIS. The id parameter is
4200 a unique ID that will be used to reference this cryptodev
4201 backend from the virtio-crypto device. The queues parame‐
4202 ter is optional, which specify the queue number of cryp‐
4203 todev backend, the default of queues is 1.
4204 # qemu-system-x86_64 \
4206 [...] \
4207 -object cryptodev-backend-builtin,id=cryptodev0 \
4208 -device virtio-crypto-pci,id=crypto0,cryptodev=cryptodev0 \
4209 [...]
4210 -object cryptodev-vhost-user,id=id,chardev=charde‐
4212 vid[,queues=queues]
4213 Creates a vhost-user cryptodev backend, backed by a
4214 chardev chardevid. The id parameter is a unique ID that
4215 will be used to reference this cryptodev backend from the
4216 virtio-crypto device. The chardev should be a unix domain
4217 socket backed one. The vhost-user uses a specifically
4218 defined protocol to pass vhost ioctl replacement messages
4219 to an application on the other end of the socket. The
4220 queues parameter is optional, which specify the queue
4221 number of cryptodev backend for multiqueue vhost-user,
4222 the default of queues is 1.
4223 # qemu-system-x86_64 \
4225 [...] \
4226 -chardev socket,id=chardev0,path=/path/to/socket \
4227 -object cryptodev-vhost-user,id=cryptodev0,chardev=chardev0 \
4228 -device virtio-crypto-pci,id=crypto0,cryptodev=cryptodev0 \
4229 [...]
4230 -object secret,id=id,data=string,format=raw|base64[,keyid=se‐
4232 cretid,iv=string]
4233 -object secret,id=id,file=filename,format=raw|base64[,keyid=se‐
4236 cretid,iv=string]
4237 Defines a secret to store a password, encryption key, or
4238 some other sensitive data. The sensitive data can either
4239 be passed directly via the data parameter, or indirectly
4240 via the file parameter. Using the data parameter is inse‐
4241 cure unless the sensitive data is encrypted.
4242 The sensitive data can be provided in raw format (the de‐
4244 fault), or base64. When encoded as JSON, the raw format
4245 only supports valid UTF-8 characters, so base64 is recom‐
4246 mended for sending binary data. QEMU will convert from
4247 which ever format is provided to the format it needs in‐
4248 ternally. eg, an RBD password can be provided in raw for‐
4249 mat, even though it will be base64 encoded when passed
4250 onto the RBD sever.
4251 For added protection, it is possible to encrypt the data
4253 associated with a secret using the AES-256-CBC cipher.
4254 Use of encryption is indicated by providing the keyid and
4255 iv parameters. The keyid parameter provides the ID of a
4256 previously defined secret that contains the AES-256 de‐
4257 cryption key. This key should be 32-bytes long and be
4258 base64 encoded. The iv parameter provides the random ini‐
4259 tialization vector used for encryption of this particular
4260 secret and should be a base64 encrypted string of the
4261 16-byte IV.
4262 The simplest (insecure) usage is to provide the secret
4264 inline
4265 # qemu-system-x86_64 -object secret,id=sec0,data=letmein,format=raw
4267 The simplest secure usage is to provide the secret via a
4269 file
4270 # printf "letmein" > mypasswd.txt # QEMU_SYSTEM_MACRO
4272 -object secret,id=sec0,file=mypasswd.txt,format=raw
4273 For greater security, AES-256-CBC should be used. To il‐
4275 lustrate usage, consider the openssl command line tool
4276 which can encrypt the data. Note that when encrypting,
4277 the plaintext must be padded to the cipher block size (32
4278 bytes) using the standard PKCS#5/6 compatible padding al‐
4279 gorithm.
4280 First a master key needs to be created in base64 encod‐
4282 ing:
4283 # openssl rand -base64 32 > key.b64
4285 # KEY=$(base64 -d key.b64 | hexdump -v -e '/1 "%02X"')
4286 Each secret to be encrypted needs to have a random ini‐
4288 tialization vector generated. These do not need to be
4289 kept secret
4290 # openssl rand -base64 16 > iv.b64
4292 # IV=$(base64 -d iv.b64 | hexdump -v -e '/1 "%02X"')
4293 The secret to be defined can now be encrypted, in this
4295 case we're telling openssl to base64 encode the result,
4296 but it could be left as raw bytes if desired.
4297 # SECRET=$(printf "letmein" |
4299 openssl enc -aes-256-cbc -a -K $KEY -iv $IV)
4300 When launching QEMU, create a master secret pointing to
4302 key.b64 and specify that to be used to decrypt the user
4303 password. Pass the contents of iv.b64 to the second se‐
4304 cret
4305 # qemu-system-x86_64 \
4307 -object secret,id=secmaster0,format=base64,file=key.b64 \
4308 -object secret,id=sec0,keyid=secmaster0,format=base64,\
4309 data=$SECRET,iv=$(<iv.b64)
4310 -object sev-guest,id=id,cbitpos=cbitpos,re‐
4312 duced-phys-bits=val,[sev-device=string,policy=policy,handle=han‐
4313 dle,dh-cert-file=file,session-file=file,kernel-hashes=on|off]
4314 Create a Secure Encrypted Virtualization (SEV) guest ob‐
4315 ject, which can be used to provide the guest memory en‐
4316 cryption support on AMD processors.
4317 When memory encryption is enabled, one of the physical
4319 address bit (aka the C-bit) is utilized to mark if a mem‐
4320 ory page is protected. The cbitpos is used to provide the
4321 C-bit position. The C-bit position is Host family depen‐
4322 dent hence user must provide this value. On EPYC, the
4323 value should be 47.
4324 When memory encryption is enabled, we loose certain bits
4326 in physical address space. The reduced-phys-bits is used
4327 to provide the number of bits we loose in physical ad‐
4328 dress space. Similar to C-bit, the value is Host family
4329 dependent. On EPYC, the value should be 5.
4330 The sev-device provides the device file to use for commu‐
4332 nicating with the SEV firmware running inside AMD Secure
4333 Processor. The default device is '/dev/sev'. If hardware
4334 supports memory encryption then /dev/sev devices are cre‐
4335 ated by CCP driver.
4336 The policy provides the guest policy to be enforced by
4338 the SEV firmware and restrict what configuration and op‐
4339 erational commands can be performed on this guest by the
4340 hypervisor. The policy should be provided by the guest
4341 owner and is bound to the guest and cannot be changed
4342 throughout the lifetime of the guest. The default is 0.
4343 If guest policy allows sharing the key with another SEV
4345 guest then handle can be use to provide handle of the
4346 guest from which to share the key.
4347 The dh-cert-file and session-file provides the guest
4349 owner's Public Diffie-Hillman key defined in SEV spec.
4350 The PDH and session parameters are used for establishing
4351 a cryptographic session with the guest owner to negotiate
4352 keys used for attestation. The file must be encoded in
4353 base64.
4354 The kernel-hashes adds the hashes of given kernel/initrd/
4356 cmdline to a designated guest firmware page for measured
4357 Linux boot with -kernel. The default is off. (Since 6.2)
4358 e.g to launch a SEV guest
4360 # qemu-system-x86_64 \
4362 ...... \
4363 -object sev-guest,id=sev0,cbitpos=47,reduced-phys-bits=5 \
4364 -machine ...,memory-encryption=sev0 \
4365 .....
4366 -object authz-simple,id=id,identity=string
4368 Create an authorization object that will control access
4369 to network services.
4370 The identity parameter is identifies the user and its
4372 format depends on the network service that authorization
4373 object is associated with. For authorizing based on TLS
4374 x509 certificates, the identity must be the x509 distin‐
4375 guished name. Note that care must be taken to escape any
4376 commas in the distinguished name.
4377 An example authorization object to validate a x509 dis‐
4379 tinguished name would look like:
4380 # qemu-system-x86_64 \
4382 ... \
4383 -object 'authz-simple,id=auth0,identity=CN=laptop.example.com,,O=Example Org,,L=London,,ST=London,,C=GB' \
4384 ...
4385 Note the use of quotes due to the x509 distinguished name
4387 containing whitespace, and escaping of ','.
4388 -object authz-listfile,id=id,filename=path,refresh=on|off
4390 Create an authorization object that will control access
4391 to network services.
4392 The filename parameter is the fully qualified path to a
4394 file containing the access control list rules in JSON
4395 format.
4396 An example set of rules that match against SASL usernames
4398 might look like:
4399 {
4401 "rules": [
4402 { "match": "fred", "policy": "allow", "format": "exact" },
4403 { "match": "bob", "policy": "allow", "format": "exact" },
4404 { "match": "danb", "policy": "deny", "format": "glob" },
4405 { "match": "dan*", "policy": "allow", "format": "exact" },
4406 ],
4407 "policy": "deny"
4408 }
4409 When checking access the object will iterate over all the
4411 rules and the first rule to match will have its policy
4412 value returned as the result. If no rules match, then the
4413 default policy value is returned.
4414 The rules can either be an exact string match, or they
4416 can use the simple UNIX glob pattern matching to allow
4417 wildcards to be used.
4418 If refresh is set to true the file will be monitored and
4420 automatically reloaded whenever its content changes.
4421 As with the authz-simple object, the format of the iden‐
4423 tity strings being matched depends on the network ser‐
4424 vice, but is usually a TLS x509 distinguished name, or a
4425 SASL username.
4426 An example authorization object to validate a SASL user‐
4428 name would look like:
4429 # qemu-system-x86_64 \
4431 ... \
4432 -object authz-simple,id=auth0,filename=/etc/qemu/vnc-sasl.acl,refresh=on \
4433 ...
4434 -object authz-pam,id=id,service=string
4436 Create an authorization object that will control access
4437 to network services.
4438 The service parameter provides the name of a PAM service
4440 to use for authorization. It requires that a file
4441 /etc/pam.d/service exist to provide the configuration for
4442 the account subsystem.
4443 An example authorization object to validate a TLS x509
4445 distinguished name would look like:
4446 # qemu-system-x86_64 \
4448 ... \
4449 -object authz-pam,id=auth0,service=qemu-vnc \
4450 ...
4451 There would then be a corresponding config file for PAM
4453 at /etc/pam.d/qemu-vnc that contains:
4454 account requisite pam_listfile.so item=user sense=allow \
4456 file=/etc/qemu/vnc.allow
4457 Finally the /etc/qemu/vnc.allow file would contain the
4459 list of x509 distingished names that are permitted access
4460 CN=laptop.example.com,O=Example Home,L=London,ST=London,C=GB
4462 -object io‐
4464 thread,id=id,poll-max-ns=poll-max-ns,poll-grow=poll-grow,poll-shrink=poll-shrink,aio-max-batch=aio-max-batch
4465 Creates a dedicated event loop thread that devices can be
4466 assigned to. This is known as an IOThread. By default de‐
4467 vice emulation happens in vCPU threads or the main event
4468 loop thread. This can become a scalability bottleneck.
4469 IOThreads allow device emulation and I/O to run on other
4470 host CPUs.
4471 The id parameter is a unique ID that will be used to ref‐
4473 erence this IOThread from -device ...,iothread=id. Mul‐
4474 tiple devices can be assigned to an IOThread. Note that
4475 not all devices support an iothread parameter.
4476 The query-iothreads QMP command lists IOThreads and re‐
4478 ports their thread IDs so that the user can configure
4479 host CPU pinning/affinity.
4480 IOThreads use an adaptive polling algorithm to reduce
4482 event loop latency. Instead of entering a blocking system
4483 call to monitor file descriptors and then pay the cost of
4484 being woken up when an event occurs, the polling algo‐
4485 rithm spins waiting for events for a short time. The al‐
4486 gorithm's default parameters are suitable for many cases
4487 but can be adjusted based on knowledge of the workload
4488 and/or host device latency.
4489 The poll-max-ns parameter is the maximum number of
4491 nanoseconds to busy wait for events. Polling can be dis‐
4492 abled by setting this value to 0.
4493 The poll-grow parameter is the multiplier used to in‐
4495 crease the polling time when the algorithm detects it is
4496 missing events due to not polling long enough.
4497 The poll-shrink parameter is the divisor used to decrease
4499 the polling time when the algorithm detects it is spend‐
4500 ing too long polling without encountering events.
4501 The aio-max-batch parameter is the maximum number of re‐
4503 quests in a batch for the AIO engine, 0 means that the
4504 engine will use its default.
4505 The IOThread parameters can be modified at run-time using
4507 the qom-set command (where iothread1 is the IOThread's
4508 id):
4509 (qemu) qom-set /objects/iothread1 poll-max-ns 100000
4511 During the graphical emulation, you can use special key combinations to
4513 change modes. The default key mappings are shown below, but if you use
4514 -alt-grab then the modifier is Ctrl-Alt-Shift (instead of Ctrl-Alt) and
4515 if you use -ctrl-grab then the modifier is the right Ctrl key (instead
4516 of Ctrl-Alt):
4517 Ctrl-Alt-f
4519 Toggle full screen
4520 Ctrl-Alt-+
4522 Enlarge the screen
4523 Ctrl-Alt--
4525 Shrink the screen
4526 Ctrl-Alt-u
4528 Restore the screen's un-scaled dimensions
4529 Ctrl-Alt-n
4531 Switch to virtual console 'n'. Standard console mappings are:
4532 1 Target system display
4534 2 Monitor
4536 3 Serial port
4538 Ctrl-Alt
4540 Toggle mouse and keyboard grab.
4541 In the virtual consoles, you can use Ctrl-Up, Ctrl-Down, Ctrl-PageUp
4543 and Ctrl-PageDown to move in the back log.
4544 During emulation, if you are using a character backend multiplexer
4546 (which is the default if you are using -nographic) then several com‐
4547 mands are available via an escape sequence. These key sequences all
4548 start with an escape character, which is Ctrl-a by default, but can be
4549 changed with -echr. The list below assumes you're using the default.
4550 Ctrl-a h
4552 Print this help
4553 Ctrl-a x
4555 Exit emulator
4556 Ctrl-a s
4558 Save disk data back to file (if -snapshot)
4559 Ctrl-a t
4561 Toggle console timestamps
4562 Ctrl-a b
4564 Send break (magic sysrq in Linux)
4565 Ctrl-a c
4567 Rotate between the frontends connected to the multiplexer (usu‐
4568 ally this switches between the monitor and the console)
4569 Ctrl-a Ctrl-a
4571 Send the escape character to the frontend
4572
4574 In addition to using normal file images for the emulated storage de‐
4575 vices, QEMU can also use networked resources such as iSCSI devices.
4576 These are specified using a special URL syntax.
4577 iSCSI iSCSI support allows QEMU to access iSCSI resources directly and
4579 use as images for the guest storage. Both disk and cdrom images
4580 are supported.
4581 Syntax for specifying iSCSI LUNs is "iscsi://<tar‐
4583 get-ip>[:<port>]/<target-iqn>/<lun>"
4584 By default qemu will use the iSCSI initiator-name
4586 'iqn.2008-11.org.linux-kvm[:<name>]' but this can also be set
4587 from the command line or a configuration file.
4588 Since version QEMU 2.4 it is possible to specify a iSCSI request
4590 timeout to detect stalled requests and force a reestablishment
4591 of the session. The timeout is specified in seconds. The default
4592 is 0 which means no timeout. Libiscsi 1.15.0 or greater is re‐
4593 quired for this feature.
4594 Example (without authentication):
4596 qemu-system-x86_64 -iscsi initiator-name=iqn.2001-04.com.example:my-initiator \
4598 -cdrom iscsi://192.0.2.1/iqn.2001-04.com.example/2 \
4599 -drive file=iscsi://192.0.2.1/iqn.2001-04.com.example/1
4600 Example (CHAP username/password via URL):
4602 qemu-system-x86_64 -drive file=iscsi://user%password@192.0.2.1/iqn.2001-04.com.example/1
4604 Example (CHAP username/password via environment variables):
4606 LIBISCSI_CHAP_USERNAME="user" \
4608 LIBISCSI_CHAP_PASSWORD="password" \
4609 qemu-system-x86_64 -drive file=iscsi://192.0.2.1/iqn.2001-04.com.example/1
4610 NBD QEMU supports NBD (Network Block Devices) both using TCP proto‐
4612 col as well as Unix Domain Sockets. With TCP, the default port
4613 is 10809.
4614 Syntax for specifying a NBD device using TCP, in preferred URI
4616 form: "nbd://<server-ip>[:<port>]/[<export>]"
4617 Syntax for specifying a NBD device using Unix Domain Sockets;
4619 remember that '?' is a shell glob character and may need quot‐
4620 ing: "nbd+unix:///[<export>]?socket=<domain-socket>"
4621 Older syntax that is also recognized:
4623 "nbd:<server-ip>:<port>[:exportname=<export>]"
4624 Syntax for specifying a NBD device using Unix Domain Sockets
4626 "nbd:unix:<domain-socket>[:exportname=<export>]"
4627 Example for TCP
4629 qemu-system-x86_64 --drive file=nbd:192.0.2.1:30000
4631 Example for Unix Domain Sockets
4633 qemu-system-x86_64 --drive file=nbd:unix:/tmp/nbd-socket
4635 SSH QEMU supports SSH (Secure Shell) access to remote disks.
4637 Examples:
4639 qemu-system-x86_64 -drive file=ssh://user@host/path/to/disk.img
4641 qemu-system-x86_64 -drive file.driver=ssh,file.user=user,file.host=host,file.port=22,file.path=/path/to/disk.img
4642 Currently authentication must be done using ssh-agent. Other au‐
4644 thentication methods may be supported in future.
4645 GlusterFS
4647 GlusterFS is a user space distributed file system. QEMU supports
4648 the use of GlusterFS volumes for hosting VM disk images using
4649 TCP, Unix Domain Sockets and RDMA transport protocols.
4650 Syntax for specifying a VM disk image on GlusterFS volume is
4652 URI:
4654 gluster[+type]://[host[:port]]/volume/path[?socket=...][,debug=N][,logfile=...]
4655 JSON:
4657 'json:{"driver":"qcow2","file":{"driver":"gluster","volume":"testvol","path":"a.img","debug":N,"logfile":"...",
4658 "server":[{"type":"tcp","host":"...","port":"..."},
4659 {"type":"unix","socket":"..."}]}}'
4660 Example
4662 URI:
4664 qemu-system-x86_64 --drive file=gluster://192.0.2.1/testvol/a.img,
4665 file.debug=9,file.logfile=/var/log/qemu-gluster.log
4666 JSON:
4668 qemu-system-x86_64 'json:{"driver":"qcow2",
4669 "file":{"driver":"gluster",
4670 "volume":"testvol","path":"a.img",
4671 "debug":9,"logfile":"/var/log/qemu-gluster.log",
4672 "server":[{"type":"tcp","host":"1.2.3.4","port":24007},
4673 {"type":"unix","socket":"/var/run/glusterd.socket"}]}}'
4674 qemu-system-x86_64 -drive driver=qcow2,file.driver=gluster,file.volume=testvol,file.path=/path/a.img,
4675 file.debug=9,file.logfile=/var/log/qemu-gluster.log,
4676 file.server.0.type=tcp,file.server.0.host=1.2.3.4,file.server.0.port=24007,
4677 file.server.1.type=unix,file.server.1.socket=/var/run/glusterd.socket
4678 See also http://www.gluster.org.
4680 HTTP/HTTPS/FTP/FTPS
4682 QEMU supports read-only access to files accessed over http(s)
4683 and ftp(s).
4684 Syntax using a single filename:
4686 <protocol>://[<username>[:<password>]@]<host>/<path>
4688 where:
4690 protocol
4692 'http', 'https', 'ftp', or 'ftps'.
4693 username
4695 Optional username for authentication to the remote
4696 server.
4697 password
4699 Optional password for authentication to the remote
4700 server.
4701 host Address of the remote server.
4703 path Path on the remote server, including any query string.
4705 The following options are also supported:
4707 url The full URL when passing options to the driver explic‐
4709 itly.
4710 readahead
4712 The amount of data to read ahead with each range request
4713 to the remote server. This value may optionally have the
4714 suffix 'T', 'G', 'M', 'K', 'k' or 'b'. If it does not
4715 have a suffix, it will be assumed to be in bytes. The
4716 value must be a multiple of 512 bytes. It defaults to
4717 256k.
4718 sslverify
4720 Whether to verify the remote server's certificate when
4721 connecting over SSL. It can have the value 'on' or 'off'.
4722 It defaults to 'on'.
4723 cookie Send this cookie (it can also be a list of cookies sepa‐
4725 rated by ';') with each outgoing request. Only supported
4726 when using protocols such as HTTP which support cookies,
4727 otherwise ignored.
4728 timeout
4730 Set the timeout in seconds of the CURL connection. This
4731 timeout is the time that CURL waits for a response from
4732 the remote server to get the size of the image to be
4733 downloaded. If not set, the default timeout of 5 seconds
4734 is used.
4735 Note that when passing options to qemu explicitly, driver is the
4737 value of <protocol>.
4738 Example: boot from a remote Fedora 20 live ISO image
4740 qemu-system-x86_64 --drive media=cdrom,file=https://archives.fedoraproject.org/pub/archive/fedora/linux/releases/20/Live/x86_64/Fedora-Live-Desktop-x86_64-20-1.iso,readonly
4742 qemu-system-x86_64 --drive media=cdrom,file.driver=http,file.url=http://archives.fedoraproject.org/pub/fedora/linux/releases/20/Live/x86_64/Fedora-Live-Desktop-x86_64-20-1.iso,readonly
4744 Example: boot from a remote Fedora 20 cloud image using a local
4746 overlay for writes, copy-on-read, and a readahead of 64k
4747 qemu-img create -f qcow2 -o backing_file='json:{"file.driver":"http",, "file.url":"http://archives.fedoraproject.org/pub/archive/fedora/linux/releases/20/Images/x86_64/Fedora-x86_64-20-20131211.1-sda.qcow2",, "file.readahead":"64k"}' /tmp/Fedora-x86_64-20-20131211.1-sda.qcow2
4749 qemu-system-x86_64 -drive file=/tmp/Fedora-x86_64-20-20131211.1-sda.qcow2,copy-on-read=on
4751 Example: boot from an image stored on a VMware vSphere server
4753 with a self-signed certificate using a local overlay for writes,
4754 a readahead of 64k and a timeout of 10 seconds.
4755 qemu-img create -f qcow2 -o backing_file='json:{"file.driver":"https",, "file.url":"https://user:password@vsphere.example.com/folder/test/test-flat.vmdk?dcPath=Datacenter&dsName=datastore1",, "file.sslverify":"off",, "file.readahead":"64k",, "file.timeout":10}' /tmp/test.qcow2
4757 qemu-system-x86_64 -drive file=/tmp/test.qcow2
4759
4761 The HTML documentation of QEMU for more precise information and Linux
4762 user mode emulator invocation.
4763
4765 Fabrice Bellard
4766
4768 2023, The QEMU Project Developers
47697.0.0 Jan 19, 2023 QEMU(1)