Migration

QEMU has code to load/save the state of the guest that it is running. These are two complementary operations. Saving the state just does that, saves the state for each device that the guest is running. Restoring a guest is just the opposite operation: we need to load the state of each device.

For this to work, QEMU has to be launched with the same arguments the two times. I.e. it can only restore the state in one guest that has the same devices that the one it was saved (this last requirement can be relaxed a bit, but for now we can consider that configuration has to be exactly the same).

Once that we are able to save/restore a guest, a new functionality is requested: migration. This means that QEMU is able to start in one machine and being “migrated” to another machine. I.e. being moved to another machine.

Next was the “live migration” functionality. This is important because some guests run with a lot of state (specially RAM), and it can take a while to move all state from one machine to another. Live migration allows the guest to continue running while the state is transferred. Only while the last part of the state is transferred has the guest to be stopped. Typically the time that the guest is unresponsive during live migration is the low hundred of milliseconds (notice that this depends on a lot of things).

Transports

The migration stream is normally just a byte stream that can be passed over any transport.

  • tcp migration: do the migration using tcp sockets

  • unix migration: do the migration using unix sockets

  • exec migration: do the migration using the stdin/stdout through a process.

  • fd migration: do the migration using a file descriptor that is passed to QEMU. QEMU doesn’t care how this file descriptor is opened.

In addition, support is included for migration using RDMA, which transports the page data using RDMA, where the hardware takes care of transporting the pages, and the load on the CPU is much lower. While the internals of RDMA migration are a bit different, this isn’t really visible outside the RAM migration code.

All these migration protocols use the same infrastructure to save/restore state devices. This infrastructure is shared with the savevm/loadvm functionality.

Debugging

The migration stream can be analyzed thanks to scripts/analyze-migration.py.

Example usage:

$ qemu-system-x86_64 -display none -monitor stdio
(qemu) migrate "exec:cat > mig"
(qemu) q
$ ./scripts/analyze-migration.py -f mig
{
  "ram (3)": {
      "section sizes": {
          "pc.ram": "0x0000000008000000",
...

See also analyze-migration.py -h help for more options.

Common infrastructure

The files, sockets or fd’s that carry the migration stream are abstracted by the QEMUFile type (see migration/qemu-file.h). In most cases this is connected to a subtype of QIOChannel (see io/).

Saving the state of one device

For most devices, the state is saved in a single call to the migration infrastructure; these are non-iterative devices. The data for these devices is sent at the end of precopy migration, when the CPUs are paused. There are also iterative devices, which contain a very large amount of data (e.g. RAM or large tables). See the iterative device section below.

General advice for device developers

  • The migration state saved should reflect the device being modelled rather than the way your implementation works. That way if you change the implementation later the migration stream will stay compatible. That model may include internal state that’s not directly visible in a register.

  • When saving a migration stream the device code may walk and check the state of the device. These checks might fail in various ways (e.g. discovering internal state is corrupt or that the guest has done something bad). Consider carefully before asserting/aborting at this point, since the normal response from users is that migration broke their VM since it had apparently been running fine until then. In these error cases, the device should log a message indicating the cause of error, and should consider putting the device into an error state, allowing the rest of the VM to continue execution.

  • The migration might happen at an inconvenient point, e.g. right in the middle of the guest reprogramming the device, during guest reboot or shutdown or while the device is waiting for external IO. It’s strongly preferred that migrations do not fail in this situation, since in the cloud environment migrations might happen automatically to VMs that the administrator doesn’t directly control.

  • If you do need to fail a migration, ensure that sufficient information is logged to identify what went wrong.

  • The destination should treat an incoming migration stream as hostile (which we do to varying degrees in the existing code). Check that offsets into buffers and the like can’t cause overruns. Fail the incoming migration in the case of a corrupted stream like this.

  • Take care with internal device state or behaviour that might become migration version dependent. For example, the order of PCI capabilities is required to stay constant across migration. Another example would be that a special case handled by subsections (see below) might become much more common if a default behaviour is changed.

  • The state of the source should not be changed or destroyed by the outgoing migration. Migrations timing out or being failed by higher levels of management, or failures of the destination host are not unusual, and in that case the VM is restarted on the source. Note that the management layer can validly revert the migration even though the QEMU level of migration has succeeded as long as it does it before starting execution on the destination.

  • Buses and devices should be able to explicitly specify addresses when instantiated, and management tools should use those. For example, when hot adding USB devices it’s important to specify the ports and addresses, since implicit ordering based on the command line order may be different on the destination. This can result in the device state being loaded into the wrong device.

VMState

Most device data can be described using the VMSTATE macros (mostly defined in include/migration/vmstate.h).

An example (from hw/input/pckbd.c)

static const VMStateDescription vmstate_kbd = {
    .name = "pckbd",
    .version_id = 3,
    .minimum_version_id = 3,
    .fields = (VMStateField[]) {
        VMSTATE_UINT8(write_cmd, KBDState),
        VMSTATE_UINT8(status, KBDState),
        VMSTATE_UINT8(mode, KBDState),
        VMSTATE_UINT8(pending, KBDState),
        VMSTATE_END_OF_LIST()
    }
};

We are declaring the state with name “pckbd”. The version_id is 3, and there are 4 uint8_t fields in the KBDState structure. We registered this VMSTATEDescription with one of the following functions. The first one will generate a device instance_id different for each registration. Use the second one if you already have an id that is different for each instance of the device:

vmstate_register_any(NULL, &vmstate_kbd, s);
vmstate_register(NULL, instance_id, &vmstate_kbd, s);

For devices that are qdev based, we can register the device in the class init function:

dc->vmsd = &vmstate_kbd_isa;

The VMState macros take care of ensuring that the device data section is formatted portably (normally big endian) and make some compile time checks against the types of the fields in the structures.

VMState macros can include other VMStateDescriptions to store substructures (see VMSTATE_STRUCT_), arrays (VMSTATE_ARRAY_) and variable length arrays (VMSTATE_VARRAY_). Various other macros exist for special cases.

Note that the format on the wire is still very raw; i.e. a VMSTATE_UINT32 ends up with a 4 byte bigendian representation on the wire; in the future it might be possible to use a more structured format.

Legacy way

This way is going to disappear as soon as all current users are ported to VMSTATE; although converting existing code can be tricky, and thus ‘soon’ is relative.

Each device has to register two functions, one to save the state and another to load the state back.

int register_savevm_live(const char *idstr,
                         int instance_id,
                         int version_id,
                         SaveVMHandlers *ops,
                         void *opaque);

Two functions in the ops structure are the save_state and load_state functions. Notice that load_state receives a version_id parameter to know what state format is receiving. save_state doesn’t have a version_id parameter because it always uses the latest version.

Note that because the VMState macros still save the data in a raw format, in many cases it’s possible to replace legacy code with a carefully constructed VMState description that matches the byte layout of the existing code.

Changing migration data structures

When we migrate a device, we save/load the state as a series of fields. Sometimes, due to bugs or new functionality, we need to change the state to store more/different information. Changing the migration state saved for a device can break migration compatibility unless care is taken to use the appropriate techniques. In general QEMU tries to maintain forward migration compatibility (i.e. migrating from QEMU n->n+1) and there are users who benefit from backward compatibility as well.

Subsections

The most common structure change is adding new data, e.g. when adding a newer form of device, or adding that state that you previously forgot to migrate. This is best solved using a subsection.

A subsection is “like” a device vmstate, but with a particularity, it has a Boolean function that tells if that values are needed to be sent or not. If this functions returns false, the subsection is not sent. Subsections have a unique name, that is looked for on the receiving side.

On the receiving side, if we found a subsection for a device that we don’t understand, we just fail the migration. If we understand all the subsections, then we load the state with success. There’s no check that a subsection is loaded, so a newer QEMU that knows about a subsection can (with care) load a stream from an older QEMU that didn’t send the subsection.

If the new data is only needed in a rare case, then the subsection can be made conditional on that case and the migration will still succeed to older QEMUs in most cases. This is OK for data that’s critical, but in some use cases it’s preferred that the migration should succeed even with the data missing. To support this the subsection can be connected to a device property and from there to a versioned machine type.

The ‘pre_load’ and ‘post_load’ functions on subsections are only called if the subsection is loaded.

One important note is that the outer post_load() function is called “after” loading all subsections, because a newer subsection could change the same value that it uses. A flag, and the combination of outer pre_load and post_load can be used to detect whether a subsection was loaded, and to fall back on default behaviour when the subsection isn’t present.

Example:

static bool ide_drive_pio_state_needed(void *opaque)
{
    IDEState *s = opaque;

    return ((s->status & DRQ_STAT) != 0)
        || (s->bus->error_status & BM_STATUS_PIO_RETRY);
}

const VMStateDescription vmstate_ide_drive_pio_state = {
    .name = "ide_drive/pio_state",
    .version_id = 1,
    .minimum_version_id = 1,
    .pre_save = ide_drive_pio_pre_save,
    .post_load = ide_drive_pio_post_load,
    .needed = ide_drive_pio_state_needed,
    .fields = (VMStateField[]) {
        VMSTATE_INT32(req_nb_sectors, IDEState),
        VMSTATE_VARRAY_INT32(io_buffer, IDEState, io_buffer_total_len, 1,
                             vmstate_info_uint8, uint8_t),
        VMSTATE_INT32(cur_io_buffer_offset, IDEState),
        VMSTATE_INT32(cur_io_buffer_len, IDEState),
        VMSTATE_UINT8(end_transfer_fn_idx, IDEState),
        VMSTATE_INT32(elementary_transfer_size, IDEState),
        VMSTATE_INT32(packet_transfer_size, IDEState),
        VMSTATE_END_OF_LIST()
    }
};

const VMStateDescription vmstate_ide_drive = {
    .name = "ide_drive",
    .version_id = 3,
    .minimum_version_id = 0,
    .post_load = ide_drive_post_load,
    .fields = (VMStateField[]) {
        .... several fields ....
        VMSTATE_END_OF_LIST()
    },
    .subsections = (const VMStateDescription*[]) {
        &vmstate_ide_drive_pio_state,
        NULL
    }
};

Here we have a subsection for the pio state. We only need to save/send this state when we are in the middle of a pio operation (that is what ide_drive_pio_state_needed() checks). If DRQ_STAT is not enabled, the values on that fields are garbage and don’t need to be sent.

Connecting subsections to properties

Using a condition function that checks a ‘property’ to determine whether to send a subsection allows backward migration compatibility when new subsections are added, especially when combined with versioned machine types.

For example:

  1. Add a new property using DEFINE_PROP_BOOL - e.g. support-foo and default it to true.

  2. Add an entry to the hw_compat_ for the previous version that sets the property to false.

  3. Add a static bool support_foo function that tests the property.

  4. Add a subsection with a .needed set to the support_foo function

  5. (potentially) Add an outer pre_load that sets up a default value for ‘foo’ to be used if the subsection isn’t loaded.

Now that subsection will not be generated when using an older machine type and the migration stream will be accepted by older QEMU versions.

Not sending existing elements

Sometimes members of the VMState are no longer needed:

  • removing them will break migration compatibility

  • making them version dependent and bumping the version will break backward migration compatibility.

Adding a dummy field into the migration stream is normally the best way to preserve compatibility.

If the field really does need to be removed then:

  1. Add a new property/compatibility/function in the same way for subsections above.

  2. replace the VMSTATE macro with the _TEST version of the macro, e.g.:

VMSTATE_UINT32(foo, barstruct)

becomes

VMSTATE_UINT32_TEST(foo, barstruct, pre_version_baz)

Sometime in the future when we no longer care about the ancient versions these can be killed off. Note that for backward compatibility it’s important to fill in the structure with data that the destination will understand.

Any difference in the predicates on the source and destination will end up with different fields being enabled and data being loaded into the wrong fields; for this reason conditional fields like this are very fragile.

Versions

Version numbers are intended for major incompatible changes to the migration of a device, and using them breaks backward-migration compatibility; in general most changes can be made by adding Subsections (see above) or _TEST macros (see above) which won’t break compatibility.

Each version is associated with a series of fields saved. The save_state always saves the state as the newer version. But load_state sometimes is able to load state from an older version.

You can see that there are two version fields:

  • version_id: the maximum version_id supported by VMState for that device.

  • minimum_version_id: the minimum version_id that VMState is able to understand for that device.

VMState is able to read versions from minimum_version_id to version_id.

There are _V forms of many VMSTATE_ macros to load fields for version dependent fields, e.g.

VMSTATE_UINT16_V(ip_id, Slirp, 2),

only loads that field for versions 2 and newer.

Saving state will always create a section with the ‘version_id’ value and thus can’t be loaded by any older QEMU.

Massaging functions

Sometimes, it is not enough to be able to save the state directly from one structure, we need to fill the correct values there. One example is when we are using kvm. Before saving the cpu state, we need to ask kvm to copy to QEMU the state that it is using. And the opposite when we are loading the state, we need a way to tell kvm to load the state for the cpu that we have just loaded from the QEMUFile.

The functions to do that are inside a vmstate definition, and are called:

  • int (*pre_load)(void *opaque);

    This function is called before we load the state of one device.

  • int (*post_load)(void *opaque, int version_id);

    This function is called after we load the state of one device.

  • int (*pre_save)(void *opaque);

    This function is called before we save the state of one device.

  • int (*post_save)(void *opaque);

    This function is called after we save the state of one device (even upon failure, unless the call to pre_save returned an error).

Example: You can look at hpet.c, that uses the first three functions to massage the state that is transferred.

The VMSTATE_WITH_TMP macro may be useful when the migration data doesn’t match the stored device data well; it allows an intermediate temporary structure to be populated with migration data and then transferred to the main structure.

If you use memory API functions that update memory layout outside initialization (i.e., in response to a guest action), this is a strong indication that you need to call these functions in a post_load callback. Examples of such memory API functions are:

  • memory_region_add_subregion()

  • memory_region_del_subregion()

  • memory_region_set_readonly()

  • memory_region_set_nonvolatile()

  • memory_region_set_enabled()

  • memory_region_set_address()

  • memory_region_set_alias_offset()

Iterative device migration

Some devices, such as RAM, Block storage or certain platform devices, have large amounts of data that would mean that the CPUs would be paused for too long if they were sent in one section. For these devices an iterative approach is taken.

The iterative devices generally don’t use VMState macros (although it may be possible in some cases) and instead use qemu_put_*/qemu_get_* macros to read/write data to the stream. Specialist versions exist for high bandwidth IO.

An iterative device must provide:

  • A save_setup function that initialises the data structures and transmits a first section containing information on the device. In the case of RAM this transmits a list of RAMBlocks and sizes.

  • A load_setup function that initialises the data structures on the destination.

  • A state_pending_exact function that indicates how much more data we must save. The core migration code will use this to determine when to pause the CPUs and complete the migration.

  • A state_pending_estimate function that indicates how much more data we must save. When the estimated amount is smaller than the threshold, we call state_pending_exact.

  • A save_live_iterate function should send a chunk of data until the point that stream bandwidth limits tell it to stop. Each call generates one section.

  • A save_live_complete_precopy function that must transmit the last section for the device containing any remaining data.

  • A load_state function used to load sections generated by any of the save functions that generate sections.

  • cleanup functions for both save and load that are called at the end of migration.

Note that the contents of the sections for iterative migration tend to be open-coded by the devices; care should be taken in parsing the results and structuring the stream to make them easy to validate.

Device ordering

There are cases in which the ordering of device loading matters; for example in some systems where a device may assert an interrupt during loading, if the interrupt controller is loaded later then it might lose the state.

Some ordering is implicitly provided by the order in which the machine definition creates devices, however this is somewhat fragile.

The MigrationPriority enum provides a means of explicitly enforcing ordering. Numerically higher priorities are loaded earlier. The priority is set by setting the priority field of the top level VMStateDescription for the device.

Stream structure

The stream tries to be word and endian agnostic, allowing migration between hosts of different characteristics running the same VM.

  • Header

    • Magic

    • Version

    • VM configuration section

      • Machine type

      • Target page bits

  • List of sections Each section contains a device, or one iteration of a device save.

    • section type

    • section id

    • ID string (First section of each device)

    • instance id (First section of each device)

    • version id (First section of each device)

    • <device data>

    • Footer mark

  • EOF mark

  • VM Description structure Consisting of a JSON description of the contents for analysis only

The device data in each section consists of the data produced by the code described above. For non-iterative devices they have a single section; iterative devices have an initial and last section and a set of parts in between. Note that there is very little checking by the common code of the integrity of the device data contents, that’s up to the devices themselves. The footer mark provides a little bit of protection for the case where the receiving side reads more or less data than expected.

The ID string is normally unique, having been formed from a bus name and device address, PCI devices and storage devices hung off PCI controllers fit this pattern well. Some devices are fixed single instances (e.g. “pc-ram”). Others (especially either older devices or system devices which for some reason don’t have a bus concept) make use of the instance id for otherwise identically named devices.

Return path

Only a unidirectional stream is required for normal migration, however a return path can be created when bidirectional communication is desired. This is primarily used by postcopy, but is also used to return a success flag to the source at the end of migration.

qemu_file_get_return_path(QEMUFile* fwdpath) gives the QEMUFile* for the return path.

Source side

Forward path - written by migration thread Return path - opened by main thread, read by return-path thread

Destination side

Forward path - read by main thread Return path - opened by main thread, written by main thread AND postcopy thread (protected by rp_mutex)

Dirty limit

The dirty limit, short for dirty page rate upper limit, is a new capability introduced in the 8.1 QEMU release that uses a new algorithm based on the KVM dirty ring to throttle down the guest during live migration.

The algorithm framework is as follows:

------------------------------------------------------------------------------
main   --------------> throttle thread ------------> PREPARE(1) <--------
thread  \                                                |              |
         \                                               |              |
          \                                              V              |
           -\                                        CALCULATE(2)       |
             \                                           |              |
              \                                          |              |
               \                                         V              |
                \                                    SET PENALTY(3) -----
                 -\                                      |
                   \                                     |
                    \                                    V
                     -> virtual CPU thread -------> ACCEPT PENALTY(4)
------------------------------------------------------------------------------

When the qmp command qmp_set_vcpu_dirty_limit is called for the first time, the QEMU main thread starts the throttle thread. The throttle thread, once launched, executes the loop, which consists of three steps:

  • PREPARE (1)

    The entire work of PREPARE (1) is preparation for the second stage, CALCULATE(2), as the name implies. It involves preparing the dirty page rate value and the corresponding upper limit of the VM: The dirty page rate is calculated via the KVM dirty ring mechanism, which tells QEMU how many dirty pages a virtual CPU has had since the last KVM_EXIT_DIRTY_RING_FULL exception; The dirty page rate upper limit is specified by caller, therefore fetch it directly.

  • CALCULATE (2)

    Calculate a suitable sleep period for each virtual CPU, which will be used to determine the penalty for the target virtual CPU. The computation must be done carefully in order to reduce the dirty page rate progressively down to the upper limit without oscillation. To achieve this, two strategies are provided: the first is to add or subtract sleep time based on the ratio of the current dirty page rate to the limit, which is used when the current dirty page rate is far from the limit; the second is to add or subtract a fixed time when the current dirty page rate is close to the limit.

  • SET PENALTY (3)

    Set the sleep time for each virtual CPU that should be penalized based on the results of the calculation supplied by step CALCULATE (2).

After completing the three above stages, the throttle thread loops back to step PREPARE (1) until the dirty limit is reached.

On the other hand, each virtual CPU thread reads the sleep duration and sleeps in the path of the KVM_EXIT_DIRTY_RING_FULL exception handler, that is ACCEPT PENALTY (4). Virtual CPUs tied with writing processes will obviously exit to the path and get penalized, whereas virtual CPUs involved with read processes will not.

In summary, thanks to the KVM dirty ring technology, the dirty limit algorithm will restrict virtual CPUs as needed to keep their dirty page rate inside the limit. This leads to more steady reading performance during live migration and can aid in improving large guest responsiveness.

Postcopy

‘Postcopy’ migration is a way to deal with migrations that refuse to converge (or take too long to converge) its plus side is that there is an upper bound on the amount of migration traffic and time it takes, the down side is that during the postcopy phase, a failure of either side causes the guest to be lost.

In postcopy the destination CPUs are started before all the memory has been transferred, and accesses to pages that are yet to be transferred cause a fault that’s translated by QEMU into a request to the source QEMU.

Postcopy can be combined with precopy (i.e. normal migration) so that if precopy doesn’t finish in a given time the switch is made to postcopy.

Enabling postcopy

To enable postcopy, issue this command on the monitor (both source and destination) prior to the start of migration:

migrate_set_capability postcopy-ram on

The normal commands are then used to start a migration, which is still started in precopy mode. Issuing:

migrate_start_postcopy

will now cause the transition from precopy to postcopy. It can be issued immediately after migration is started or any time later on. Issuing it after the end of a migration is harmless.

Blocktime is a postcopy live migration metric, intended to show how long the vCPU was in state of interruptible sleep due to pagefault. That metric is calculated both for all vCPUs as overlapped value, and separately for each vCPU. These values are calculated on destination side. To enable postcopy blocktime calculation, enter following command on destination monitor:

migrate_set_capability postcopy-blocktime on

Postcopy blocktime can be retrieved by query-migrate qmp command. postcopy-blocktime value of qmp command will show overlapped blocking time for all vCPU, postcopy-vcpu-blocktime will show list of blocking time per vCPU.

Note

During the postcopy phase, the bandwidth limits set using migrate_set_parameter is ignored (to avoid delaying requested pages that the destination is waiting for).

Postcopy device transfer

Loading of device data may cause the device emulation to access guest RAM that may trigger faults that have to be resolved by the source, as such the migration stream has to be able to respond with page data during the device load, and hence the device data has to be read from the stream completely before the device load begins to free the stream up. This is achieved by ‘packaging’ the device data into a blob that’s read in one go.

Source behaviour

Until postcopy is entered the migration stream is identical to normal precopy, except for the addition of a ‘postcopy advise’ command at the beginning, to tell the destination that postcopy might happen. When postcopy starts the source sends the page discard data and then forms the ‘package’ containing:

  • Command: ‘postcopy listen’

  • The device state

    A series of sections, identical to the precopy streams device state stream containing everything except postcopiable devices (i.e. RAM)

  • Command: ‘postcopy run’

The ‘package’ is sent as the data part of a Command: CMD_PACKAGED, and the contents are formatted in the same way as the main migration stream.

During postcopy the source scans the list of dirty pages and sends them to the destination without being requested (in much the same way as precopy), however when a page request is received from the destination, the dirty page scanning restarts from the requested location. This causes requested pages to be sent quickly, and also causes pages directly after the requested page to be sent quickly in the hope that those pages are likely to be used by the destination soon.

Destination behaviour

Initially the destination looks the same as precopy, with a single thread reading the migration stream; the ‘postcopy advise’ and ‘discard’ commands are processed to change the way RAM is managed, but don’t affect the stream processing.

------------------------------------------------------------------------------
                        1      2   3     4 5                      6   7
main -----DISCARD-CMD_PACKAGED ( LISTEN  DEVICE     DEVICE DEVICE RUN )
thread                             |       |
                                   |     (page request)
                                   |        \___
                                   v            \
listen thread:                     --- page -- page -- page -- page -- page --

                                   a   b        c
------------------------------------------------------------------------------
  • On receipt of CMD_PACKAGED (1)

    All the data associated with the package - the ( … ) section in the diagram - is read into memory, and the main thread recurses into qemu_loadvm_state_main to process the contents of the package (2) which contains commands (3,6) and devices (4…)

  • On receipt of ‘postcopy listen’ - 3 -(i.e. the 1st command in the package)

    a new thread (a) is started that takes over servicing the migration stream, while the main thread carries on loading the package. It loads normal background page data (b) but if during a device load a fault happens (5) the returned page (c) is loaded by the listen thread allowing the main threads device load to carry on.

  • The last thing in the CMD_PACKAGED is a ‘RUN’ command (6)

    letting the destination CPUs start running. At the end of the CMD_PACKAGED (7) the main thread returns to normal running behaviour and is no longer used by migration, while the listen thread carries on servicing page data until the end of migration.

Postcopy Recovery

Comparing to precopy, postcopy is special on error handlings. When any error happens (in this case, mostly network errors), QEMU cannot easily fail a migration because VM data resides in both source and destination QEMU instances. On the other hand, when issue happens QEMU on both sides will go into a paused state. It’ll need a recovery phase to continue a paused postcopy migration.

The recovery phase normally contains a few steps:

  • When network issue occurs, both QEMU will go into PAUSED state

  • When the network is recovered (or a new network is provided), the admin can setup the new channel for migration using QMP command ‘migrate-recover’ on destination node, preparing for a resume.

  • On source host, the admin can continue the interrupted postcopy migration using QMP command ‘migrate’ with resume=true flag set.

  • After the connection is re-established, QEMU will continue the postcopy migration on both sides.

During a paused postcopy migration, the VM can logically still continue running, and it will not be impacted from any page access to pages that were already migrated to destination VM before the interruption happens. However, if any of the missing pages got accessed on destination VM, the VM thread will be halted waiting for the page to be migrated, it means it can be halted until the recovery is complete.

The impact of accessing missing pages can be relevant to different configurations of the guest. For example, when with async page fault enabled, logically the guest can proactively schedule out the threads accessing missing pages.

Postcopy states

Postcopy moves through a series of states (see postcopy_state) from ADVISE->DISCARD->LISTEN->RUNNING->END

  • Advise

    Set at the start of migration if postcopy is enabled, even if it hasn’t had the start command; here the destination checks that its OS has the support needed for postcopy, and performs setup to ensure the RAM mappings are suitable for later postcopy. The destination will fail early in migration at this point if the required OS support is not present. (Triggered by reception of POSTCOPY_ADVISE command)

  • Discard

    Entered on receipt of the first ‘discard’ command; prior to the first Discard being performed, hugepages are switched off (using madvise) to ensure that no new huge pages are created during the postcopy phase, and to cause any huge pages that have discards on them to be broken.

  • Listen

    The first command in the package, POSTCOPY_LISTEN, switches the destination state to Listen, and starts a new thread (the ‘listen thread’) which takes over the job of receiving pages off the migration stream, while the main thread carries on processing the blob. With this thread able to process page reception, the destination now ‘sensitises’ the RAM to detect any access to missing pages (on Linux using the ‘userfault’ system).

  • Running

    POSTCOPY_RUN causes the destination to synchronise all state and start the CPUs and IO devices running. The main thread now finishes processing the migration package and now carries on as it would for normal precopy migration (although it can’t do the cleanup it would do as it finishes a normal migration).

  • Paused

    Postcopy can run into a paused state (normally on both sides when happens), where all threads will be temporarily halted mostly due to network errors. When reaching paused state, migration will make sure the qemu binary on both sides maintain the data without corrupting the VM. To continue the migration, the admin needs to fix the migration channel using the QMP command ‘migrate-recover’ on the destination node, then resume the migration using QMP command ‘migrate’ again on source node, with resume=true flag set.

  • End

    The listen thread can now quit, and perform the cleanup of migration state, the migration is now complete.

Source side page map

The ‘migration bitmap’ in postcopy is basically the same as in the precopy, where each of the bit to indicate that page is ‘dirty’ - i.e. needs sending. During the precopy phase this is updated as the CPU dirties pages, however during postcopy the CPUs are stopped and nothing should dirty anything any more. Instead, dirty bits are cleared when the relevant pages are sent during postcopy.

Postcopy with hugepages

Postcopy now works with hugetlbfs backed memory:

  1. The linux kernel on the destination must support userfault on hugepages.

  2. The huge-page configuration on the source and destination VMs must be identical; i.e. RAMBlocks on both sides must use the same page size.

  3. Note that -mem-path /dev/hugepages will fall back to allocating normal RAM if it doesn’t have enough hugepages, triggering (b) to fail. Using -mem-prealloc enforces the allocation using hugepages.

  4. Care should be taken with the size of hugepage used; postcopy with 2MB hugepages works well, however 1GB hugepages are likely to be problematic since it takes ~1 second to transfer a 1GB hugepage across a 10Gbps link, and until the full page is transferred the destination thread is blocked.

Postcopy with shared memory

Postcopy migration with shared memory needs explicit support from the other processes that share memory and from QEMU. There are restrictions on the type of memory that userfault can support shared.

The Linux kernel userfault support works on /dev/shm memory and on hugetlbfs (although the kernel doesn’t provide an equivalent to madvise(MADV_DONTNEED) for hugetlbfs which may be a problem in some configurations).

The vhost-user code in QEMU supports clients that have Postcopy support, and the vhost-user-bridge (in tests/) and the DPDK package have changes to support postcopy.

The client needs to open a userfaultfd and register the areas of memory that it maps with userfault. The client must then pass the userfaultfd back to QEMU together with a mapping table that allows fault addresses in the clients address space to be converted back to RAMBlock/offsets. The client’s userfaultfd is added to the postcopy fault-thread and page requests are made on behalf of the client by QEMU. QEMU performs ‘wake’ operations on the client’s userfaultfd to allow it to continue after a page has arrived.

Note

There are two future improvements that would be nice:
  1. Some way to make QEMU ignorant of the addresses in the clients address space

  2. Avoiding the need for QEMU to perform ufd-wake calls after the pages have arrived

Retro-fitting postcopy to existing clients is possible:
  1. A mechanism is needed for the registration with userfault as above, and the registration needs to be coordinated with the phases of postcopy. In vhost-user extra messages are added to the existing control channel.

  2. Any thread that can block due to guest memory accesses must be identified and the implication understood; for example if the guest memory access is made while holding a lock then all other threads waiting for that lock will also be blocked.

Postcopy Preemption Mode

Postcopy preempt is a new capability introduced in 8.0 QEMU release, it allows urgent pages (those got page fault requested from destination QEMU explicitly) to be sent in a separate preempt channel, rather than queued in the background migration channel. Anyone who cares about latencies of page faults during a postcopy migration should enable this feature. By default, it’s not enabled.

Firmware

Migration migrates the copies of RAM and ROM, and thus when running on the destination it includes the firmware from the source. Even after resetting a VM, the old firmware is used. Only once QEMU has been restarted is the new firmware in use.

  • Changes in firmware size can cause changes in the required RAMBlock size to hold the firmware and thus migration can fail. In practice it’s best to pad firmware images to convenient powers of 2 with plenty of space for growth.

  • Care should be taken with device emulation code so that newer emulation code can work with older firmware to allow forward migration.

  • Care should be taken with newer firmware so that backward migration to older systems with older device emulation code will work.

In some cases it may be best to tie specific firmware versions to specific versioned machine types to cut down on the combinations that will need support. This is also useful when newer versions of firmware outgrow the padding.

Backwards compatibility

How backwards compatibility works

When we do migration, we have two QEMU processes: the source and the target. There are two cases, they are the same version or they are different versions. The easy case is when they are the same version. The difficult one is when they are different versions.

There are two things that are different, but they have very similar names and sometimes get confused:

  • QEMU version

  • machine type version

Let’s start with a practical example, we start with:

  • qemu-system-x86_64 (v5.2), from now on qemu-5.2.

  • qemu-system-x86_64 (v5.1), from now on qemu-5.1.

Related to this are the “latest” machine types defined on each of them:

  • pc-q35-5.2 (newer one in qemu-5.2) from now on pc-5.2

  • pc-q35-5.1 (newer one in qemu-5.1) from now on pc-5.1

First of all, migration is only supposed to work if you use the same machine type in both source and destination. The QEMU hardware configuration needs to be the same also on source and destination. Most aspects of the backend configuration can be changed at will, except for a few cases where the backend features influence frontend device feature exposure. But that is not relevant for this section.

I am going to list the number of combinations that we can have. Let’s start with the trivial ones, QEMU is the same on source and destination:

1 - qemu-5.2 -M pc-5.2 -> migrates to -> qemu-5.2 -M pc-5.2

This is the latest QEMU with the latest machine type. This have to work, and if it doesn’t work it is a bug.

2 - qemu-5.1 -M pc-5.1 -> migrates to -> qemu-5.1 -M pc-5.1

Exactly the same case than the previous one, but for 5.1. Nothing to see here either.

This are the easiest ones, we will not talk more about them in this section.

Now we start with the more interesting cases. Consider the case where we have the same QEMU version in both sides (qemu-5.2) but we are using the latest machine type for that version (pc-5.2) but one of an older QEMU version, in this case pc-5.1.

3 - qemu-5.2 -M pc-5.1 -> migrates to -> qemu-5.2 -M pc-5.1

It needs to use the definition of pc-5.1 and the devices as they were configured on 5.1, but this should be easy in the sense that both sides are the same QEMU and both sides have exactly the same idea of what the pc-5.1 machine is.

4 - qemu-5.1 -M pc-5.2 -> migrates to -> qemu-5.1 -M pc-5.2

This combination is not possible as the qemu-5.1 doesn’t understand pc-5.2 machine type. So nothing to worry here.

Now it comes the interesting ones, when both QEMU processes are different. Notice also that the machine type needs to be pc-5.1, because we have the limitation than qemu-5.1 doesn’t know pc-5.2. So the possible cases are:

5 - qemu-5.2 -M pc-5.1 -> migrates to -> qemu-5.1 -M pc-5.1

This migration is known as newer to older. We need to make sure when we are developing 5.2 we need to take care about not to break migration to qemu-5.1. Notice that we can’t make updates to qemu-5.1 to understand whatever qemu-5.2 decides to change, so it is in qemu-5.2 side to make the relevant changes.

6 - qemu-5.1 -M pc-5.1 -> migrates to -> qemu-5.2 -M pc-5.1

This migration is known as older to newer. We need to make sure than we are able to receive migrations from qemu-5.1. The problem is similar to the previous one.

If qemu-5.1 and qemu-5.2 were the same, there will not be any compatibility problems. But the reason that we create qemu-5.2 is to get new features, devices, defaults, etc.

If we get a device that has a new feature, or change a default value, we have a problem when we try to migrate between different QEMU versions.

So we need a way to tell qemu-5.2 that when we are using machine type pc-5.1, it needs to not use the feature, to be able to migrate to real qemu-5.1.

And the equivalent part when migrating from qemu-5.1 to qemu-5.2. qemu-5.2 has to expect that it is not going to get data for the new feature, because qemu-5.1 doesn’t know about it.

How do we tell QEMU about these device feature changes? In hw/core/machine.c:hw_compat_X_Y arrays.

If we change a default value, we need to put back the old value on that array. And the device, during initialization needs to look at that array to see what value it needs to get for that feature. And what are we going to put in that array, the value of a property.

To create a property for a device, we need to use one of the DEFINE_PROP_*() macros. See include/hw/qdev-properties.h to find the macros that exist. With it, we set the default value for that property, and that is what it is going to get in the latest released version. But if we want a different value for a previous version, we can change that in the hw_compat_X_Y arrays.

hw_compat_X_Y is an array of registers that have the format:

  • name_device

  • name_property

  • value

Let’s see a practical example.

In qemu-5.2 virtio-blk-device got multi queue support. This is a change that is not backward compatible. In qemu-5.1 it has one queue. In qemu-5.2 it has the same number of queues as the number of cpus in the system.

When we are doing migration, if we migrate from a device that has 4 queues to a device that have only one queue, we don’t know where to put the extra information for the other 3 queues, and we fail migration.

Similar problem when we migrate from qemu-5.1 that has only one queue to qemu-5.2, we only sent information for one queue, but destination has 4, and we have 3 queues that are not properly initialized and anything can happen.

So, how can we address this problem. Easy, just convince qemu-5.2 that when it is running pc-5.1, it needs to set the number of queues for virtio-blk-devices to 1.

That way we fix the cases 5 and 6.

5 - qemu-5.2 -M pc-5.1 -> migrates to -> qemu-5.1 -M pc-5.1

qemu-5.2 -M pc-5.1 sets number of queues to be 1. qemu-5.1 -M pc-5.1 expects number of queues to be 1.

correct. migration works.

6 - qemu-5.1 -M pc-5.1 -> migrates to -> qemu-5.2 -M pc-5.1

qemu-5.1 -M pc-5.1 sets number of queues to be 1. qemu-5.2 -M pc-5.1 expects number of queues to be 1.

correct. migration works.

And now the other interesting case, case 3. In this case we have:

3 - qemu-5.2 -M pc-5.1 -> migrates to -> qemu-5.2 -M pc-5.1

Here we have the same QEMU in both sides. So it doesn’t matter a lot if we have set the number of queues to 1 or not, because they are the same.

WRONG!

Think what happens if we do one of this double migrations:

A -> migrates -> B -> migrates -> C

where:

A: qemu-5.1 -M pc-5.1 B: qemu-5.2 -M pc-5.1 C: qemu-5.2 -M pc-5.1

migration A -> B is case 6, so number of queues needs to be 1.

migration B -> C is case 3, so we don’t care. But actually we care because we haven’t started the guest in qemu-5.2, it came migrated from qemu-5.1. So to be in the safe place, we need to always use number of queues 1 when we are using pc-5.1.

Now, how was this done in reality? The following commit shows how it was done:

commit 9445e1e15e66c19e42bea942ba810db28052cd05
Author: Stefan Hajnoczi <stefanha@redhat.com>
Date:   Tue Aug 18 15:33:47 2020 +0100

virtio-blk-pci: default num_queues to -smp N

The relevant parts for migration are:

@@ -1281,7 +1284,8 @@ static Property virtio_blk_properties[] = {
 #endif
     DEFINE_PROP_BIT("request-merging", VirtIOBlock, conf.request_merging, 0,
                     true),
-    DEFINE_PROP_UINT16("num-queues", VirtIOBlock, conf.num_queues, 1),
+    DEFINE_PROP_UINT16("num-queues", VirtIOBlock, conf.num_queues,
+                       VIRTIO_BLK_AUTO_NUM_QUEUES),
     DEFINE_PROP_UINT16("queue-size", VirtIOBlock, conf.queue_size, 256),

It changes the default value of num_queues. But it fishes it for old machine types to have the right value:

@@ -31,6 +31,7 @@
 GlobalProperty hw_compat_5_1[] = {
     ...
+    { "virtio-blk-device", "num-queues", "1"},
     ...
 };

A device with different features on both sides

Let’s assume that we are using the same QEMU binary on both sides, just to make the things easier. But we have a device that has different features on both sides of the migration. That can be because the devices are different, because the kernel driver of both devices have different features, whatever.

How can we get this to work with migration. The way to do that is “theoretically” easy. You have to get the features that the device has in the source of the migration. The features that the device has on the target of the migration, you get the intersection of the features of both sides, and that is the way that you should launch QEMU.

Notice that this is not completely related to QEMU. The most important thing here is that this should be handled by the managing application that launches QEMU. If QEMU is configured correctly, the migration will succeed.

That said, actually doing it is complicated. Almost all devices are bad at being able to be launched with only some features enabled. With one big exception: cpus.

You can read the documentation for QEMU x86 cpu models here:

https://qemu-project.gitlab.io/qemu/system/qemu-cpu-models.html

See when they talk about migration they recommend that one chooses the newest cpu model that is supported for all cpus.

Let’s say that we have:

Host A:

Device X has the feature Y

Host B:

Device X has not the feature Y

If we try to migrate without any care from host A to host B, it will fail because when migration tries to load the feature Y on destination, it will find that the hardware is not there.

Doing this would be the equivalent of doing with cpus:

Host A:

$ qemu-system-x86_64 -cpu host

Host B:

$ qemu-system-x86_64 -cpu host

When both hosts have different cpu features this is guaranteed to fail. Especially if Host B has less features than host A. If host A has less features than host B, sometimes it works. Important word of last sentence is “sometimes”.

So, forgetting about cpu models and continuing with the -cpu host example, let’s see that the differences of the cpus is that Host A and B have the following features:

Features: ‘pcid’ ‘stibp’ ‘taa-no’ Host A: X X Host B: X

And we want to migrate between them, the way configure both QEMU cpu will be:

Host A:

$ qemu-system-x86_64 -cpu host,pcid=off,stibp=off

Host B:

$ qemu-system-x86_64 -cpu host,taa-no=off

And you would be able to migrate between them. It is responsibility of the management application or of the user to make sure that the configuration is correct. QEMU doesn’t know how to look at this kind of features in general.

Notice that we don’t recommend to use -cpu host for migration. It is used in this example because it makes the example simpler.

Other devices have worse control about individual features. If they want to be able to migrate between hosts that show different features, the device needs a way to configure which ones it is going to use.

In this section we have considered that we are using the same QEMU binary in both sides of the migration. If we use different QEMU versions process, then we need to have into account all other differences and the examples become even more complicated.

How to mitigate when we have a backward compatibility error

We broke migration for old machine types continuously during development. But as soon as we find that there is a problem, we fix it. The problem is what happens when we detect after we have done a release that something has gone wrong.

Let see how it worked with one example.

After the release of qemu-8.0 we found a problem when doing migration of the machine type pc-7.2.

  • $ qemu-7.2 -M pc-7.2 -> qemu-7.2 -M pc-7.2

    This migration works

  • $ qemu-8.0 -M pc-7.2 -> qemu-8.0 -M pc-7.2

    This migration works

  • $ qemu-8.0 -M pc-7.2 -> qemu-7.2 -M pc-7.2

    This migration fails

  • $ qemu-7.2 -M pc-7.2 -> qemu-8.0 -M pc-7.2

    This migration fails

So clearly something fails when migration between qemu-7.2 and qemu-8.0 with machine type pc-7.2. The error messages, and git bisect pointed to this commit.

In qemu-8.0 we got this commit:

commit 010746ae1db7f52700cb2e2c46eb94f299cfa0d2
Author: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Date:   Thu Mar 2 13:37:02 2023 +0000

hw/pci/aer: Implement PCI_ERR_UNCOR_MASK register

The relevant bits of the commit for our example are this ones:

--- a/hw/pci/pcie_aer.c
+++ b/hw/pci/pcie_aer.c
@@ -112,6 +112,10 @@ int pcie_aer_init(PCIDevice *dev,

     pci_set_long(dev->w1cmask + offset + PCI_ERR_UNCOR_STATUS,
                  PCI_ERR_UNC_SUPPORTED);
+    pci_set_long(dev->config + offset + PCI_ERR_UNCOR_MASK,
+                 PCI_ERR_UNC_MASK_DEFAULT);
+    pci_set_long(dev->wmask + offset + PCI_ERR_UNCOR_MASK,
+                 PCI_ERR_UNC_SUPPORTED);

     pci_set_long(dev->config + offset + PCI_ERR_UNCOR_SEVER,
                 PCI_ERR_UNC_SEVERITY_DEFAULT);

The patch changes how we configure PCI space for AER. But QEMU fails when the PCI space configuration is different between source and destination.

The following commit shows how this got fixed:

commit 5ed3dabe57dd9f4c007404345e5f5bf0e347317f
Author: Leonardo Bras <leobras@redhat.com>
Date:   Tue May 2 21:27:02 2023 -0300

hw/pci: Disable PCI_ERR_UNCOR_MASK register for machine type < 8.0

[...]

The relevant parts of the fix in QEMU are as follow:

First, we create a new property for the device to be able to configure the old behaviour or the new behaviour:

diff --git a/hw/pci/pci.c b/hw/pci/pci.c
index 8a87ccc8b0..5153ad63d6 100644
--- a/hw/pci/pci.c
+++ b/hw/pci/pci.c
@@ -79,6 +79,8 @@ static Property pci_props[] = {
     DEFINE_PROP_STRING("failover_pair_id", PCIDevice,
                        failover_pair_id),
     DEFINE_PROP_UINT32("acpi-index",  PCIDevice, acpi_index, 0),
+    DEFINE_PROP_BIT("x-pcie-err-unc-mask", PCIDevice, cap_present,
+                    QEMU_PCIE_ERR_UNC_MASK_BITNR, true),
     DEFINE_PROP_END_OF_LIST()
 };

Notice that we enable the feature for new machine types.

Now we see how the fix is done. This is going to depend on what kind of breakage happens, but in this case it is quite simple:

diff --git a/hw/pci/pcie_aer.c b/hw/pci/pcie_aer.c
index 103667c368..374d593ead 100644
--- a/hw/pci/pcie_aer.c
+++ b/hw/pci/pcie_aer.c
@@ -112,10 +112,13 @@ int pcie_aer_init(PCIDevice *dev, uint8_t cap_ver,
uint16_t offset,

     pci_set_long(dev->w1cmask + offset + PCI_ERR_UNCOR_STATUS,
                  PCI_ERR_UNC_SUPPORTED);
-    pci_set_long(dev->config + offset + PCI_ERR_UNCOR_MASK,
-                 PCI_ERR_UNC_MASK_DEFAULT);
-    pci_set_long(dev->wmask + offset + PCI_ERR_UNCOR_MASK,
-                 PCI_ERR_UNC_SUPPORTED);
+
+    if (dev->cap_present & QEMU_PCIE_ERR_UNC_MASK) {
+        pci_set_long(dev->config + offset + PCI_ERR_UNCOR_MASK,
+                     PCI_ERR_UNC_MASK_DEFAULT);
+        pci_set_long(dev->wmask + offset + PCI_ERR_UNCOR_MASK,
+                     PCI_ERR_UNC_SUPPORTED);
+    }

     pci_set_long(dev->config + offset + PCI_ERR_UNCOR_SEVER,
                  PCI_ERR_UNC_SEVERITY_DEFAULT);

I.e. If the property bit is enabled, we configure it as we did for qemu-8.0. If the property bit is not set, we configure it as it was in 7.2.

And now, everything that is missing is disabling the feature for old machine types:

diff --git a/hw/core/machine.c b/hw/core/machine.c
index 47a34841a5..07f763eb2e 100644
--- a/hw/core/machine.c
+++ b/hw/core/machine.c
@@ -48,6 +48,7 @@ GlobalProperty hw_compat_7_2[] = {
     { "e1000e", "migrate-timadj", "off" },
     { "virtio-mem", "x-early-migration", "false" },
     { "migration", "x-preempt-pre-7-2", "true" },
+    { TYPE_PCI_DEVICE, "x-pcie-err-unc-mask", "off" },
 };
 const size_t hw_compat_7_2_len = G_N_ELEMENTS(hw_compat_7_2);

And now, when qemu-8.0.1 is released with this fix, all combinations are going to work as supposed.

  • $ qemu-7.2 -M pc-7.2 -> qemu-7.2 -M pc-7.2 (works)

  • $ qemu-8.0.1 -M pc-7.2 -> qemu-8.0.1 -M pc-7.2 (works)

  • $ qemu-8.0.1 -M pc-7.2 -> qemu-7.2 -M pc-7.2 (works)

  • $ qemu-7.2 -M pc-7.2 -> qemu-8.0.1 -M pc-7.2 (works)

So the normality has been restored and everything is ok, no?

Not really, now our matrix is much bigger. We started with the easy cases, migration from the same version to the same version always works:

  • $ qemu-7.2 -M pc-7.2 -> qemu-7.2 -M pc-7.2

  • $ qemu-8.0 -M pc-7.2 -> qemu-8.0 -M pc-7.2

  • $ qemu-8.0.1 -M pc-7.2 -> qemu-8.0.1 -M pc-7.2

Now the interesting ones. When the QEMU processes versions are different. For the 1st set, their fail and we can do nothing, both versions are released and we can’t change anything.

  • $ qemu-7.2 -M pc-7.2 -> qemu-8.0 -M pc-7.2

  • $ qemu-8.0 -M pc-7.2 -> qemu-7.2 -M pc-7.2

This two are the ones that work. The whole point of making the change in qemu-8.0.1 release was to fix this issue:

  • $ qemu-7.2 -M pc-7.2 -> qemu-8.0.1 -M pc-7.2

  • $ qemu-8.0.1 -M pc-7.2 -> qemu-7.2 -M pc-7.2

But now we found that qemu-8.0 neither can migrate to qemu-7.2 not qemu-8.0.1.

  • $ qemu-8.0 -M pc-7.2 -> qemu-8.0.1 -M pc-7.2

  • $ qemu-8.0.1 -M pc-7.2 -> qemu-8.0 -M pc-7.2

So, if we start a pc-7.2 machine in qemu-8.0 we can’t migrate it to anything except to qemu-8.0.

Can we do better?

Yeap. If we know that we are going to do this migration:

  • $ qemu-8.0 -M pc-7.2 -> qemu-8.0.1 -M pc-7.2

We can launch the appropriate devices with:

--device...,x-pci-e-err-unc-mask=on

And now we can receive a migration from 8.0. And from now on, we can do that migration to new machine types if we remember to enable that property for pc-7.2. Notice that we need to remember, it is not enough to know that the source of the migration is qemu-8.0. Think of this example:

$ qemu-8.0 -M pc-7.2 -> qemu-8.0.1 -M pc-7.2 -> qemu-8.2 -M pc-7.2

In the second migration, the source is not qemu-8.0, but we still have that “problem” and have that property enabled. Notice that we need to continue having this mark/property until we have this machine rebooted. But it is not a normal reboot (that don’t reload QEMU) we need the machine to poweroff/poweron on a fixed QEMU. And from now on we can use the proper real machine.