Now we have a Error** passed into the return path thread stack, which is
even clearer than an int retval. Change ram_dirty_bitmap_reload() and the
callers to use a bool instead to replace errnos.
Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231017202633.296756-5-peterx@redhat.com>
Normally the postcopy recover phase should only exist for a super short
period, that's the duration when QEMU is trying to recover from an
interrupted postcopy migration, during which handshake will be carried out
for continuing the procedure with state changes from PAUSED -> RECOVER ->
POSTCOPY_ACTIVE again.
Here RECOVER phase should be super small, that happens right after the
admin specified a new but working network link for QEMU to reconnect to
dest QEMU.
However there can still be case where the channel is broken in this small
RECOVER window.
If it happens, with current code there's no way the src QEMU can got kicked
out of RECOVER stage. No way either to retry the recover in another channel
when established.
This patch allows the RECOVER phase to fail itself too - we're mostly
ready, just some small things missing, e.g. properly kick the main
migration thread out when sleeping on rp_sem when we found that we're at
RECOVER stage. When this happens, it fails the RECOVER itself, and
rollback to PAUSED stage. Then the user can retry another round of
recovery.
To make it even stronger, teach QMP command migrate-pause to explicitly
kick src/dst QEMU out when needed, so even if for some reason the migration
thread didn't got kicked out already by a failing rethrn-path thread, the
admin can also kick it out.
This will be an super, super corner case, but still try to cover that.
One can try to test this with two proxy channels for migration:
(a) socat unix-listen:/tmp/src.sock,reuseaddr,fork tcp:localhost:10000
(b) socat tcp-listen:10000,reuseaddr,fork unix:/tmp/dst.sock
So the migration channel will be:
(a) (b)
src -> /tmp/src.sock -> tcp:10000 -> /tmp/dst.sock -> dst
Then to make QEMU hang at RECOVER stage, one can do below:
(1) stop the postcopy using QMP command postcopy-pause
(2) kill the 2nd proxy (b)
(3) try to recover the postcopy using /tmp/src.sock on src
(4) src QEMU will go into RECOVER stage but won't be able to continue
from there, because the channel is actually broken at (b)
Before this patch, step (4) will make src QEMU stuck in RECOVER stage,
without a way to kick the QEMU out or continue the postcopy again. After
this patch, (4) will quickly fail qemu and bounce back to PAUSED stage.
Admin can also kick QEMU from (4) into PAUSED when needed using
migrate-pause when needed.
After bouncing back to PAUSED stage, one can recover again.
Reported-by: Xiaohui Li <xiaohli@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2111332
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231017202633.296756-3-peterx@redhat.com>
rp_state.error was a boolean used to show error happened in return path
thread. That's not only duplicating error reporting (migrate_set_error),
but also not good enough in that we only do error_report() and set it to
true, we never can keep a history of the exact error and show it in
query-migrate.
To make this better, a few things done:
- Use error_setg() rather than error_report() across the whole lifecycle
of return path thread, keeping the error in an Error*.
- With above, no need to have mark_source_rp_bad(), remove it, alongside
with rp_state.error itself.
- Use migrate_set_error() to apply that captured error to the global
migration object when error occured in this thread.
- Do the same when detected qemufile error in source return path
We need to re-export qemu_file_get_error_obj() to do the last one.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231017202633.296756-2-peterx@redhat.com>
This patch is inspired by Joao Martin's patch here:
https://lore.kernel.org/r/20230926161841.98464-1-joao.m.martins@oracle.com
Add tracepoints for major downtime checkpoints on both src and dst. They
share the same tracepoint with a string showing its stage.
Besides the checkpoints in the previous patch, this patch also added
destination checkpoints.
On src, we have these checkpoints added:
- src-downtime-start: right before vm stops on src
- src-vm-stopped: after vm is fully stopped
- src-iterable-saved: after all iterables saved (END sections)
- src-non-iterable-saved: after all non-iterable saved (FULL sections)
- src-downtime-stop: migration fully completed
On dst, we have these checkpoints added:
- dst-precopy-loadvm-completes: after loadvm all done for precopy
- dst-precopy-bh-*: record BH steps to resume VM for precopy
- dst-postcopy-bh-*: record BH steps to resume VM for postcopy
On dst side, we don't have a good way to trace total time consumed by
iterable or non-iterable for now. We can mark it by 1st time receiving a
FULL / END section, but rather than that let's just rely on the other
tracepoints added for vmstates to back up the information.
With this patch, one can enable "vmstate_downtime*" tracepoints and it'll
enable all tracepoints for downtime measurements necessary.
Drop loadvm_postcopy_handle_run_bh() tracepoint alongside, because they
service the same purpose, which was only for postcopy. We then have
unified prefix for all downtime relevant tracepoints.
Co-developed-by: Joao Martins <joao.m.martins@oracle.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231030163346.765724-6-peterx@redhat.com>
We have a bunch of savevm_section* tracepoints, they're good to analyze
migration stream, but not always suitable if someone would like to analyze
the migration downtime. Two major problems:
- savevm_section* tracepoints are dumping all sections, we only care
about the sections that contribute to the downtime
- They don't have an identifier to show the type of sections, so no way
to filter downtime information either easily.
We can add type into the tracepoints, but instead of doing so, this patch
kept them untouched, instead of adding a bunch of downtime specific
tracepoints, so one can enable "vmstate_downtime*" tracepoints and get a
full picture of how the downtime is distributed across iterative and
non-iterative vmstate save/load.
Note that here both save() and load() need to be traced, because both of
them may contribute to the downtime. The contribution is not a simple "add
them together", though: consider when the src is doing a save() of device1
while the dest can be load()ing for device2, so they can happen
concurrently.
Tracking both sides make sense because device load() and save() can be
imbalanced, one device can save() super fast, but load() super slow, vice
versa. We can't figure that out without tracing both.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231030163346.765724-4-peterx@redhat.com>
Postcopy calculates its downtime separately. It always sets
MigrationState.downtime properly, but not MigrationState.downtime_start.
Make postcopy do the same as other modes on properly recording the
timestamp when the VM is going to be stopped. Drop the temporary variable
in postcopy_start() along the way.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231030163346.765724-2-peterx@redhat.com>
Before finally register one SaveStateEntry, we detect for duplicated
entries. This could be helpful to notify us asap instead of get
silent migration failures which could be hard to diagnose.
For example, this patch will generate a message like this (if without
previous fixes on x2apic) as long as we wants to boot a VM instance
with "-smp 200,maxcpus=288,sockets=2,cores=72,threads=2" and QEMU will
bail out even before VM starts:
savevm_state_handler_insert: Detected duplicate SaveStateEntry: id=apic, instance_id=0x0
Suggested-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231020090731.28701-10-quintela@redhat.com>
It's possible that some errors can be overwritten with success retval later
on, and then ignored. Always capture all errors and report.
Reported by Coverity 1522861, but actually I spot one more in the same
function.
Fixes: CID 1522861
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231017203855.298260-1-peterx@redhat.com>
Migration bandwidth is a very important value to live migration. It's
because it's one of the major factors that we'll make decision on when to
switchover to destination in a precopy process.
This value is currently estimated by QEMU during the whole live migration
process by monitoring how fast we were sending the data. This can be the
most accurate bandwidth if in the ideal world, where we're always feeding
unlimited data to the migration channel, and then it'll be limited to the
bandwidth that is available.
However in reality it may be very different, e.g., over a 10Gbps network we
can see query-migrate showing migration bandwidth of only a few tens of
MB/s just because there are plenty of other things the migration thread
might be doing. For example, the migration thread can be busy scanning
zero pages, or it can be fetching dirty bitmap from other external dirty
sources (like vhost or KVM). It means we may not be pushing data as much
as possible to migration channel, so the bandwidth estimated from "how many
data we sent in the channel" can be dramatically inaccurate sometimes.
With that, the decision to switchover will be affected, by assuming that we
may not be able to switchover at all with such a low bandwidth, but in
reality we can.
The migration may not even converge at all with the downtime specified,
with that wrong estimation of bandwidth, keeping iterations forever with a
low estimation of bandwidth.
The issue is QEMU itself may not be able to avoid those uncertainties on
measuing the real "available migration bandwidth". At least not something
I can think of so far.
One way to fix this is when the user is fully aware of the available
bandwidth, then we can allow the user to help providing an accurate value.
For example, if the user has a dedicated channel of 10Gbps for migration
for this specific VM, the user can specify this bandwidth so QEMU can
always do the calculation based on this fact, trusting the user as long as
specified. It may not be the exact bandwidth when switching over (in which
case qemu will push migration data as fast as possible), but much better
than QEMU trying to wildly guess, especially when very wrong.
A new parameter "avail-switchover-bandwidth" is introduced just for this.
So when the user specified this parameter, instead of trusting the
estimated value from QEMU itself (based on the QEMUFile send speed), it
trusts the user more by using this value to decide when to switchover,
assuming that we'll have such bandwidth available then.
Note that specifying this value will not throttle the bandwidth for
switchover yet, so QEMU will always use the full bandwidth possible for
sending switchover data, assuming that should always be the most important
way to use the network at that time.
This can resolve issues like "unconvergence migration" which is caused by
hilarious low "migration bandwidth" detected for whatever reason.
Reported-by: Zhiyi Guo <zhguo@redhat.com>
Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231010221922.40638-1-peterx@redhat.com>
Instead of only relying on the count of rp_sem, make the counter be part of
RAMState so it can be used in both threads to synchronize on the process.
rp_sem will be further reused in follow up patches, as a way to kick the
main thread, e.g., on recovery failures.
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231004220240.167175-7-peterx@redhat.com>
There're a lot of cases where we only have an errno set in last_error but
without a detailed error description. When this happens, try to generate
an error contains the errno as a descriptive error.
This will be helpful in cases where one relies on the Error*. E.g.,
migration state only caches Error* in MigrationState.error. With this,
we'll display correct error messages in e.g. query-migrate when the error
was only set by qemu_file_set_error().
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231004220240.167175-6-peterx@redhat.com>
Introduce a helper to detect whether MigrationState.error is set for
whatever reason.
This is preparation work for any thread (e.g. source return path thread) to
setup errors in an unified way to MigrationState, rather than relying on
its own way to set errors (mark_source_rp_bad()).
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231004220240.167175-3-peterx@redhat.com>
Display it as long as being set, irrelevant of FAILED status. E.g., it may
also be applicable to PAUSED stage of postcopy, to provide hint on what has
gone wrong.
The error_mutex seems to be overlooked when referencing the error, add it
to be very safe.
This will change QAPI behavior by showing up error message outside !FAILED
status, but it's intended and doesn't expect to break anyone.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2018404
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231004220240.167175-2-peterx@redhat.com>
There's a bug on dest that if a double fault triggered on dest qemu (a
network issue during postcopy-recover), we won't set PAUSED correctly
because we assumed we always came from ACTIVE.
Fix that by always overwriting the state to PAUSE.
We could also check for these two states, but maybe it's an overkill. We
did the same on the src QEMU to unconditionally switch to PAUSE anyway.
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231004220240.167175-10-peterx@redhat.com>
For both save/load we actually share the logic on deciding whether a field
should exist. Merge the checks into a helper and use it for both save and
load. When doing so, add documentations and reformat the code to make it
much easier to read.
The real benefit here (besides code cleanups) is we add a trace-point for
this; this is a known spot where we can easily break migration
compatibilities between binaries, and this trace point will be critical for
us to identify such issues.
For example, this will be handy when debugging things like:
https://gitlab.com/qemu-project/qemu/-/issues/932
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230906204722.514474-1-peterx@redhat.com>
This patch fixes the warning of shadowed local variable:
../hw/i386/intel_iommu.c: In function ‘vtd_address_space_unmap’:
../hw/i386/intel_iommu.c:3773:18: warning: declaration of ‘size’ shadows a previous local [-Wshadow=compatible-local]
3773 | uint64_t size = mask + 1;
| ^~~~
../hw/i386/intel_iommu.c:3747:12: note: shadowed declaration is here
3747 | hwaddr size, remain;
| ^~~~
Cc: Jason Wang <jasowang@redhat.com>
Cc: Eric Auger <eric.auger@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-ID: <20230922160410.138786-1-peterx@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
We hit intermit CI issue on failing at migration-test over the unit test
preempt/plain:
qemu-system-x86_64: Unable to read from socket: Connection reset by peer
Memory content inconsistency at 5b43000 first_byte = bd last_byte = bc current = 4f hit_edge = 1
**
ERROR:../tests/qtest/migration-test.c:300:check_guests_ram: assertion failed: (bad == 0)
(test program exited with status code -6)
Fabiano debugged into it and found that the preempt thread can quit even
without receiving all the pages, which can cause guest not receiving all
the pages and corrupt the guest memory.
To make sure preempt thread finished receiving all the pages, we can rely
on the page_requested_count being zero because preempt channel will only
receive requested page faults. Note, not all the faulted pages are required
to be sent via the preempt channel/thread; imagine the case when a
requested page is just queued into the background main channel for
migration, the src qemu will just still send it via the background channel.
Here instead of spinning over reading the count, we add a condvar so the
main thread can wait on it if that unusual case happened, without burning
the cpu for no good reason, even if the duration is short; so even if we
spin in this rare case is probably fine. It's just better to not do so.
The condvar is only used when that special case is triggered. Some memory
ordering trick is needed to guarantee it from happening (against the
preempt thread status field), so the main thread will always get a kick
when that triggers correctly.
Closes: https://gitlab.com/qemu-project/qemu/-/issues/1886
Debugged-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230918172822.19052-2-farosas@suse.de>
We have postcopy recovery but not reflected in the document, do an update
for that.
Add a very small section on postcopy preempt.
Touch up the pagemap section, dropping the unsent map because it's already
been dropped in the source code in commit 1e7cf8c323 ("migration/postcopy:
unsentmap is not necessary for postcopy").
Touch up the postcopy section to remove "network connection" failures as
downside, because now it's not fatal and can be recovered. Suggested by
Laszlo.
Acked-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230706115611.371048-1-peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Recognize this field for VMS_ARRAY typed vmsd fields, then we can do proper
size matching with previous patch.
Note that this is compatible with old -dump-vmstate output, because when
"num" is not there we'll still use the old "size" only.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
For VMS_ARRAY typed vmsd fields, also dump the number of entries in the
array in -vmstate-dump.
Without such information, vmstate static checker can report false negatives
of incompatible vmsd on VMS_ARRAY typed fields, when the src/dst do not
have the same type of array defined. It's because in the checker we only
check against size of fields within a VMSD field.
One example: e1000e used to have a field defined as a boolean array with 5
entries, then removed it and replaced it with UNUSED (in 31e3f318c8):
- VMSTATE_BOOL_ARRAY(core.eitr_intr_pending, E1000EState,
- E1000E_MSIX_VEC_NUM),
+ VMSTATE_UNUSED(E1000E_MSIX_VEC_NUM),
It's a legal replacement but vmstate static checker is not happy with it,
because it checks only against the "size" field between the two
fields (here one is BOOL_ARRAY, the other is UNUSED):
For BOOL_ARRAY:
{
"field": "core.eitr_intr_pending",
"version_id": 0,
"field_exists": false,
"size": 1
},
For UNUSED:
{
"field": "unused",
"version_id": 0,
"field_exists": false,
"size": 5
},
It's not the script to blame because there's just not enough information
dumped to show the total size of the entry for an array. Add it.
Note that this will not break old vmstate checker because the field will
just be ignored.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Instead of print it to STDERR, bring the error upwards so that it can be
reported via QMP responses.
E.g.:
{ "execute": "migrate-set-capabilities" ,
"arguments": { "capabilities":
[ { "capability": "postcopy-ram", "state": true } ] } }
{ "error":
{ "class": "GenericError",
"desc": "Postcopy is not supported: Host backend files need to be TMPFS
or HUGETLBFS only" } }
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Postcopy requires the memory support userfaultfd to work. Right now we
check it but it's a bit too late (when switching to postcopy migration).
Do that early right at enabling of postcopy.
Note that this is still only a best effort because ramblocks can be
dynamically created. We can add check in hostmem creations and fail if
postcopy enabled, but maybe that's too aggressive.
Still, we have chance to fail the most obvious where we know there's an
existing unsupported ramblock.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
The migration object may want to check against different types of memory
when initialized. Delay the creation to be after late backends.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
This new helper fetches file system type for a fd. Only Linux is
implemented so far. Currently only tmpfs and hugetlbfs are defined,
but it can grow as needed.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
In 8.0 devel window we reworked preempt channel creation, so that there'll
be no race condition when the migration channel and preempt channel got
established in the wrong order in commit 5655aab079.
However no one noticed that the change will also be not compatible with
older qemus, majorly 7.1/7.2 versions where preempt mode started to be
supported.
Leverage the same pre-7.2 flag introduced in the previous patch to recover
the behavior hopefully before 8.0 releases, so we don't break migration
when we migrate from 8.0 to older qemu binaries.
Fixes: 5655aab079 ("migration: Postpone postcopy preempt channel to be after main")
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
postcopy_qemufile_src object should be owned by one thread, either the main
thread (e.g. when at the beginning, or at the end of migration), or by the
return path thread (when during a preempt enabled postcopy migration). If
that's not the case the access to the object might be racy.
postcopy_preempt_shutdown_file() can be potentially racy, because it's
called at the end phase of migration on the main thread, however during
which the return path thread hasn't yet been recycled; the recycle happens
in await_return_path_close_on_source() which is after this point.
It means, logically it's posslbe the main thread and the return path thread
are both operating on the same qemufile. While I don't think qemufile is
thread safe at all.
postcopy_preempt_shutdown_file() used to be needed because that's where we
send EOS to dest so that dest can safely shutdown the preempt thread.
To avoid the possible race, remove this only place that a race can happen.
Instead we figure out another way to safely close the preempt thread on
dest.
The core idea during postcopy on deciding "when to stop" is that dest will
send a postcopy SHUT message to src, telling src that all data is there.
Hence to shut the dest preempt thread maybe better to do it directly on
dest node.
This patch proposed such a way that we change postcopy_prio_thread_created
into PreemptThreadStatus, so that we kick the preempt thread on dest qemu
by a sequence of:
mis->preempt_thread_status = PREEMPT_THREAD_QUIT;
qemu_file_shutdown(mis->postcopy_qemufile_dst);
While here shutdown() is probably so far the easiest way to kick preempt
thread from a blocked qemu_get_be64(). Then it reads preempt_thread_status
to make sure it's not a network failure but a willingness to quit the
thread.
We could have avoided that extra status but just rely on migration status.
The problem is postcopy_ram_incoming_cleanup() is just called early enough
so we're still during POSTCOPY_ACTIVE no matter what.. So just make it
simple to have the status introduced.
One flag x-preempt-pre-7-2 is added to keep old pre-7.2 behaviors of
postcopy preempt.
Fixes: 9358982744 ("migration: Send requested page directly in rp-return thread")
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
TLS iochannel will inherit io_shutdown() from the master ioc, however we
missed to do that on the server side.
This will e.g. allow qemu_file_shutdown() to work on dest QEMU too for
migration.
Acked-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
It's possible that we want to reap a dirty ring on a vcpu that is during
creation, because the vcpu is put onto list (CPU_FOREACH visible) before
initialization of the structures. In this case:
qemu_init_vcpu
x86_cpu_realizefn
cpu_exec_realizefn
cpu_list_add <---- can be probed by CPU_FOREACH
qemu_init_vcpu
cpus_accel->create_vcpu_thread(cpu);
kvm_init_vcpu
map kvm_dirty_gfns <--- kvm_dirty_gfns valid
Don't try to reap dirty ring on vcpus during creation or it'll crash.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=2124756
Reported-by: Xiaohui Li <xiaohli@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <1d14deb6684bcb7de1c9633c5bd21113988cc698.1676563222.git.huangy81@chinatelecom.cn>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
QEMU main thread will wait until dest preempt channel established during
processing the LISTEN command (within the whole postcopy PACKAGED data), by
waiting on the semaphore postcopy_qemufile_dst_done.
That's racy, because it's possible that the dest QEMU main thread hasn't
yet accept()ed the new connection when processing the LISTEN event. The
sem_wait() will yield the main thread without being able to run anything
else including the accept() of the new socket, which can cause deadlock
within the main thread.
To avoid the race, move the "wait channel" from main thread to the preempt
thread right at the start.
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Fixes: 5655aab079 ("migration: Postpone postcopy preempt channel to be after main")
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
We don't send UNMAP notification upon domain or global invalidation
which will lead the notifier can't work correctly. One example is to
use vhost remote IOTLB without enabling device IOTLB.
Fixing this by sending UNMAP notification.
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20230223065924.42503-6-jasowang@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Postcopy with preempt-mode enabled needs two channels to communicate. The
order of channel establishment is not guaranteed. It can happen that the
dest QEMU got the preempt channel connection request before the main
channel is established, then the migration may make no progress even during
precopy due to the wrong order.
To fix it, create the preempt channel only if we know the main channel is
established.
For a general postcopy migration, we delay it until postcopy_start(),
that's where we already went through some part of precopy on the main
channel. To make sure dest QEMU has already established the channel, we
wait until we got the first PONG received. That's something we do at the
start of precopy when postcopy enabled so it's guaranteed to happen sooner
or later.
For a postcopy recovery, we delay it to qemu_savevm_state_resume_prepare()
where we'll have round trips of data on bitmap synchronizations, which
means the main channel must have been established.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
This is mostly useless, but useful for us to know whether the main channel
is correctly established without changing the migration protocol.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Since we just dropped the only case where postcopy_preempt_setup() can
return an error, it doesn't need a retval anymore because it never fails.
Move the preempt check to the caller, preparing it to be used elsewhere to
do nothing but as simple as kicking the async connection.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
The whole idea of multi-channel checks was not properly done, IMHO.
Currently we check multi-channel in a lot of places, but actually that's
not needed because we only need to check it right after we get the URI and
that should be it.
If the URI check succeeded, we should never need to check it again because
we must have it. If it check fails, we should fail immediately on either
the qmp_migrate or qmp_migrate_incoming, instead of failingg it later after
the connection established.
Neither should we fail any set capabiliities like what we used to do here:
5ad15e8614 ("migration: allow enabling mutilfd for specific protocol only", 2021-10-19)
Because logically the URI will only be set later after the capability is
set, so it doesn't make a lot of sense to check the URI type when setting
the capability, because we're checking the cap with an old URI passed in,
and that may not even be the URI we're going to use later.
This patch mostly reverted all such checks for before, dropping the
variable migrate_allow_multi_channels and helpers. Instead, add a common
helper to check URI for multi-channels for either qmp_migrate and
qmp_migrate_incoming and that should do all the proper checks. The failure
will only trigger with the "migrate" or "migrate_incoming" command, or when
user specified "-incoming xxx" where "xxx" is not "defer".
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Teach QEMU to use /dev/userfaultfd when it existed and fallback to the
system call if either it's not there or doesn't have enough permission.
Firstly, as long as the app has permission to access /dev/userfaultfd, it
always have the ability to trap kernel faults which QEMU mostly wants.
Meanwhile, in some context (e.g. containers) the userfaultfd syscall can be
forbidden, so it can be the major way to use postcopy in a restricted
environment with strict seccomp setup.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
The downtime should be displayed during postcopy phase because the
switchover phase is done. OTOH it's weird to show "expected downtime"
which can confuse what does that mean if the switchover has already
happened anyway.
This is a slight ABI change on QMP, but I assume it shouldn't affect
anyone.
Reviewed-by: Leonardo Bras <leobras@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Commit d9e474ea56 overlooked the case where the target psize is even larger
than the host psize. One example is Alpha has 8K page size and migration
will start to crash the source QEMU when running Alpha migration on x86.
Fix it by detecting that case and set host start/end just to cover the
single page to be migrated.
This will slightly optimize the common case where host psize equals to
guest psize so we don't even need to do the roundups, but that's trivial.
Cc: qemu-stable@nongnu.org
Reported-by: Thomas Huth <thuth@redhat.com>
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1456
Fixes: d9e474ea56 ("migration: Teach PSS about host page")
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
It seems not super clear on when iova_tree is used, and why. Add a rich
comment above iova_tree to track why we needed the iova_tree, and when we
need it.
Also comment for the map/unmap messages, on how they're used and
implications (e.g. unmap can be larger than the mapped ranges).
Suggested-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20230109193727.1360190-1-peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Now with rs->pss we can already cache channels in pss->pss_channels. That
pss_channel contains more infromation than rs->f because it's per-channel.
So rs->f could be replaced by rss->pss[RAM_CHANNEL_PRECOPY].pss_channel,
while rs->f itself is a bit vague now.
Note that vanilla postcopy still send pages via pss[RAM_CHANNEL_PRECOPY],
that's slightly confusing but it reflects the reality.
Then, after the replacement we can safely drop rs->f.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
With the new code to send pages in rp-return thread, there's little help to
keep lots of the old code on maintaining the preempt state in migration
thread, because the new way should always be faster..
Then if we'll always send pages in the rp-return thread anyway, we don't
need those logic to maintain preempt state anymore because now we serialize
things using the mutex directly instead of using those fields.
It's very unfortunate to have those code for a short period, but that's
still one intermediate step that we noticed the next bottleneck on the
migration thread. Now what we can do best is to drop unnecessary code as
long as the new code is stable to reduce the burden. It's actually a good
thing because the new "sending page in rp-return thread" model is (IMHO)
even cleaner and with better performance.
Remove the old code that was responsible for maintaining preempt states, at
the meantime also remove x-postcopy-preempt-break-huge parameter because
with concurrent sender threads we don't really need to break-huge anymore.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
With all the facilities ready, send the requested page directly in the
rp-return thread rather than queuing it in the request queue, if and only
if postcopy preempt is enabled. It can achieve so because it uses separate
channel for sending urgent pages. The only shared data is bitmap and it's
protected by the bitmap_mutex.
Note that since we're moving the ownership of the urgent channel from the
migration thread to rp thread it also means the rp thread is responsible
for managing the qemufile, e.g. properly close it when pausing migration
happens. For this, let migration_release_from_dst_file to cover shutdown
of the urgent channel too, renaming it as migration_release_dst_files() to
better show what it does.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Since we use PageSearchStatus to represent a channel, it makes perfect
sense to keep last_sent_block (aka, leverage RAM_SAVE_FLAG_CONTINUE) to be
per-channel rather than global because each channel can be sending
different pages on ramblocks.
Hence move it from RAMState into PageSearchStatus.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
We used to allocate PSS structure on the stack for precopy when sending
pages. Make it static, so as to describe per-channel ram migration status.
Here we declared RAM_CHANNEL_MAX instances, preparing for postcopy to use
it, even though this patch has not yet to start using the 2nd instance.
This should not have any functional change per se, but it already starts to
export PSS information via the RAMState, so that e.g. one PSS channel can
start to reference the other PSS channel.
Always protect PSS access using the same RAMState.bitmap_mutex. We already
do so, so no code change needed, just some comment update. Maybe we should
consider renaming bitmap_mutex some day as it's going to be a more commonly
and big mutex we use for ram states, but just leave it for later.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Introduce pss_channel for PageSearchStatus, define it as "the migration
channel to be used to transfer this host page".
We used to have rs->f, which is a mirror to MigrationState.to_dst_file.
After postcopy preempt initial version, rs->f can be dynamically changed
depending on which channel we want to use.
But that later work still doesn't grant full concurrency of sending pages
in e.g. different threads, because rs->f can either be the PRECOPY channel
or POSTCOPY channel. This needs to be per-thread too.
PageSearchStatus is actually a good piece of struct which we can leverage
if we want to have multiple threads sending pages. Sending a single guest
page may not make sense, so we make the granule to be "host page", and in
the PSS structure we allow specify a QEMUFile* to migrate a specific host
page. Then we open the possibility to specify different channels in
different threads with different PSS structures.
The PSS prefix can be slightly misleading here because e.g. for the
upcoming usage of postcopy channel/thread it's not "searching" (or,
scanning) at all but sending the explicit page that was requested. However
since PSS existed for some years keep it as-is until someone complains.
This patch mostly (simply) replace rs->f with pss->pss_channel only. No
functional change intended for this patch yet. But it does prepare to
finally drop rs->f, and make ram_save_guest_page() thread safe.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Migration code has a lot to do with host pages. Teaching PSS core about
the idea of host page helps a lot and makes the code clean. Meanwhile,
this prepares for the future changes that can leverage the new PSS helpers
that this patch introduces to send host page in another thread.
Three more fields are introduced for this:
(1) host_page_sending: this is set to true when QEMU is sending a host
page, false otherwise.
(2) host_page_{start|end}: these point to the start/end of host page
we're sending, and it's only valid when host_page_sending==true.
For example, when we look up the next dirty page on the ramblock, with
host_page_sending==true, we'll not try to look for anything beyond the
current host page boundary. This can be slightly efficient than current
code because currently we'll set pss->page to next dirty bit (which can be
over current host page boundary) and reset it to host page boundary if we
found it goes beyond that.
With above, we can easily make migration_bitmap_find_dirty() self contained
by updating pss->page properly. rs* parameter is removed because it's not
even used in old code.
When sending a host page, we should use the pss helpers like this:
- pss_host_page_prepare(pss): called before sending host page
- pss_within_range(pss): whether we're still working on the cur host page?
- pss_host_page_finish(pss): called after sending a host page
Then we can use ram_save_target_page() to save one small page.
Currently ram_save_host_page() is still the only user. If there'll be
another function to send host page (e.g. in return path thread) in the
future, it should follow the same style.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
To prepare for thread-safety on page accountings, at least below counters
need to be accessed only atomically, they are:
ram_counters.transferred
ram_counters.duplicate
ram_counters.normal
ram_counters.postcopy_bytes
There are a lot of other counters but they won't be accessed outside
migration thread, then they're still safe to be accessed without atomic
ops.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Don't take the bitmap mutex when sending pages, or when being throttled by
migration_rate_limit() (which is a bit tricky to call it here in ram code,
but seems still helpful).
It prepares for the possibility of concurrently sending pages in >1 threads
using the function ram_save_host_page() because all threads may need the
bitmap_mutex to operate on bitmaps, so that either sendmsg() or any kind of
qemu_sem_wait() blocking for one thread will not block the other from
progressing.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Removing referencing to RAMState.f in compress_page_with_multi_thread() and
flush_compressed_data().
Compression code by default isn't compatible with having >1 channels (or it
won't currently know which channel to flush the compressed data), so to
make it simple we always flush on the default to_dst_file port until
someone wants to add >1 ports support, as rs->f right now can really
change (after postcopy preempt is introduced).
There should be no functional change at all after patch applied, since as
long as rs->f referenced in compression code, it must be to_dst_file.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
The major change is to replace "!save_page_use_compression()" with
"xbzrle_enabled" to make it clear.
Reasonings:
(1) When compression enabled, "!save_page_use_compression()" is exactly the
same as checking "xbzrle_enabled".
(2) When compression disabled, "!save_page_use_compression()" always return
true. We used to try calling the xbzrle code, but after this change we
won't, and we shouldn't need to.
Since at it, drop the xbzrle_enabled check in xbzrle_cache_zero_page()
because with this change it's not needed anymore.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Any call to ram_find_and_save_block() needs to take the bitmap mutex. We
used to not take it for most of ram_save_complete() because we thought
we're the only one left using the bitmap, but it's not true after the
preempt full patchset applied, since the return path can be taking it too.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Multifd thread model does not work for compression, explicitly disable it.
Note that previuosly even we can enable both of them, nothing will go
wrong, because the compression code has higher priority so multifd feature
will just be ignored. Now we'll fail even earlier at config time so the
user should be aware of the consequence better.
Note that there can be a slight chance of breaking existing users, but
let's assume they're not majority and not serious users, or they should
have found that multifd is not working already.
With that, we can safely drop the check in ram_save_target_page() for using
multifd, because when multifd=on then compression=off, then the removed
check on save_page_use_compression() will also always return false too.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Since we already have bitmap_mutex to protect either the dirty bitmap or
the clear log bitmap, we don't need atomic operations to set/clear/test on
the clear log bitmap. Switching all ops from atomic to non-atomic
versions, meanwhile touch up the comments to show which lock is in charge.
Introduced non-atomic version of bitmap_test_and_clear_atomic(), mostly the
same as the atomic version but simplified a few places, e.g. dropped the
"old_bits" variable, and also the explicit memory barriers.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
The preempt mode requires the capability to assign channel for each of the
page, while the compression logic will currently assign pages to different
compress thread/local-channel so potentially they're incompatible.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
In qemu_file_shutdown(), there's a possible race if with current order of
operation. There're two major things to do:
(1) Do real shutdown() (e.g. shutdown() syscall on socket)
(2) Update qemufile's last_error
We must do (2) before (1) otherwise there can be a race condition like:
page receiver other thread
------------- ------------
qemu_get_buffer()
do shutdown()
returns 0 (buffer all zero)
(meanwhile we didn't check this retcode)
try to detect IO error
last_error==NULL, IO okay
install ALL-ZERO page
set last_error
--> guest crash!
To fix this, we can also check retval of qemu_get_buffer(), but not all
APIs can be properly checked and ultimately we still need to go back to
qemu_file_get_error(). E.g. qemu_get_byte() doesn't return error.
Maybe some day a rework of qemufile API is really needed, but for now keep
using qemu_file_get_error() and fix it by not allowing that race condition
to happen. Here shutdown() is indeed special because the last_error was
emulated. For real -EIO errors it'll always be set when e.g. sendmsg()
error triggers so we won't miss those ones, only shutdown() is a bit tricky
here.
Cc: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
When starting ram saving procedure (especially at the completion phase),
always set last_seen_block to non-NULL to make sure we can always correctly
detect the case where "we've migrated all the dirty pages".
Then we'll guarantee both last_seen_block and pss.block will be valid
always before the loop starts.
See the comment in the code for some details.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
It's true that when vcpus<=255 we don't require the length of 32bit APIC
IDs. However here since we already have EIM=ON it means the hypervisor
will declare the VM as x2apic supported (e.g. VT-d ECAP register will have
EIM bit 4 set), so the guest should assume the APIC IDs are 32bits width
even if vcpus<=255. In short, commit 77250171bd breaks any simple cmdline
that wants to boot a VM with >=9 but <=255 vcpus with:
-device intel-iommu,intremap=on
For anyone who does not want to enable x2apic, we can use eim=off in the
intel-iommu parameters to skip enabling KVM x2apic.
This partly reverts commit 77250171bd, while
keeping the valid bit on checking split irqchip, but revert the other change.
One thing to mention is that this patch may break migration compatibility
of such VM, however that's probably the best thing we can do, because the
old behavior was simply wrong and not working for >8 vcpus. For <=8 vcpus,
there could be a light guest ABI change (by enabling KVM x2apic after this
patch), but logically it shouldn't affect the migration from working.
Also, this is not the 1st commit to change x2apic behavior. Igor provided
a full history of how this evolved for the past few years:
https://lore.kernel.org/qemu-devel/20220922154617.57d1a1fb@redhat.com/
Relevant commits for reference:
fb506e701e ("intel_iommu: reject broken EIM", 2016-10-17)
c1bb5418e3 ("target/i386: Support up to 32768 CPUs without IRQ remapping", 2020-12-10)
77250171bd ("intel_iommu: Fix irqchip / X2APIC configuration checks", 2022-05-16)
dc89f32d92 ("target/i386: Fix sanity check on max APIC ID / X2APIC enablement", 2022-05-16)
We may want to have this for stable too (mostly for 7.1.0 only). Adding a
fixes tag.
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Claudio Fontana <cfontana@suse.de>
Cc: Igor Mammedov <imammedo@redhat.com>
Fixes: 77250171bd ("intel_iommu: Fix irqchip / X2APIC configuration checks")
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220926153206.10881-1-peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Thomas reported that auto-converge test will timeout on MacOS CI gatings.
Use the migrate_ensure_converge() helper too in the auto-converge as when
Daniel reworked the other test cases.
Since both max_bandwidth / downtime_limit will not be used for converge
calculations, make it simple by removing the remaining check, then we can
completely remove both variables altogether, since migrate_ensure_converge
is used the remaining check won't make much sense anyway.
Reported-by: Thomas Huth <thuth@redhat.com>
Suggested-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220728133516.92061-2-peterx@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
We just added TLS tests for precopy but not postcopy. Add the
corresponding test for vanilla postcopy.
Rename the vanilla postcopy to "postcopy/plain" because all postcopy tests
will only use unix sockets as channel.
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185525.27692-1-peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
dgilbert: Manual merge
With preemption mode on, when we see a postcopy request that was requesting
for exactly the page that we have preempted before (so we've partially sent
the page already via PRECOPY channel and it got preempted by another
postcopy request), currently we drop the request so that after all the
other postcopy requests are serviced then we'll go back to precopy stream
and start to handle that.
We dropped the request because we can't send it via postcopy channel since
the precopy channel already contains partial of the data, and we can only
send a huge page via one channel as a whole. We can't split a huge page
into two channels.
That's a very corner case and that works, but there's a change on the order
of postcopy requests that we handle since we're postponing this (unlucky)
postcopy request to be later than the other queued postcopy requests. The
problem is there's a possibility that when the guest was very busy, the
postcopy queue can be always non-empty, it means this dropped request will
never be handled until the end of postcopy migration. So, there's a chance
that there's one dest QEMU vcpu thread waiting for a page fault for an
extremely long time just because it's unluckily accessing the specific page
that was preempted before.
The worst case time it needs can be as long as the whole postcopy migration
procedure. It's extremely unlikely to happen, but when it happens it's not
good.
The root cause of this problem is because we treat pss->postcopy_requested
variable as with two meanings bound together, as the variable shows:
1. Whether this page request is urgent, and,
2. Which channel we should use for this page request.
With the old code, when we set postcopy_requested it means either both (1)
and (2) are true, or both (1) and (2) are false. We can never have (1)
and (2) to have different values.
However it doesn't necessarily need to be like that. It's very legal that
there's one request that has (1) very high urgency, but (2) we'd like to
use the precopy channel. Just like the corner case we were discussing
above.
To differenciate the two meanings better, introduce a new field called
postcopy_target_channel, showing which channel we should use for this page
request, so as to cover the old meaning (2) only. Then we leave the
postcopy_requested variable to stand only for meaning (1), which is the
urgency of this page request.
With this change, we can easily boost priority of a preempted precopy page
as long as we know that page is also requested as a postcopy page. So with
the new approach in get_queued_page() instead of dropping that request, we
send it right away with the precopy channel so we get back the ordering of
the page faults just like how they're requested on dest.
Reported-by: Manish Mishra <manish.mishra@nutanix.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Manish Mishra <manish.mishra@nutanix.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185520.27583-1-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
This patch is based on the async preempt channel creation. It continues
wiring up the new channel with TLS handshake to destionation when enabled.
Note that only the src QEMU needs such operation; the dest QEMU does not
need any change for TLS support due to the fact that all channels are
established synchronously there, so all the TLS magic is already properly
handled by migration_tls_channel_process_incoming().
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185518.27529-1-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
It's useful for specifying tls credentials all in the cmdline (along with
the -object tls-creds-*), especially for debugging purpose.
The trick here is we must remember to not free these fields again in the
finalize() function of migration object, otherwise it'll cause double-free.
The thing is when destroying an object, we'll first destroy the properties
that bound to the object, then the object itself. To be explicit, when
destroy the object in object_finalize() we have such sequence of
operations:
object_property_del_all(obj);
object_deinit(obj, ti);
So after this change the two fields are properly released already even
before reaching the finalize() function but in object_property_del_all(),
hence we don't need to free them anymore in finalize() or it's double-free.
This also fixes a trivial memory leak for tls-authz as we forgot to free it
before this patch.
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185515.27475-1-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Add migrate_channel_requires_tls() to detect whether the specific channel
requires TLS, leveraging the recently introduced migrate_use_tls(). No
functional change intended.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185513.27421-1-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Add a property field that can conditionally disable the "break sending huge
page" behavior in postcopy preemption. By default it's enabled.
It should only be used for debugging purposes, and we should never remove
the "x-" prefix.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Manish Mishra <manish.mishra@nutanix.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185511.27366-1-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
This patch allows the postcopy preempt channel to be created
asynchronously. The benefit is that when the connection is slow, we won't
take the BQL (and potentially block all things like QMP) for a long time
without releasing.
A function postcopy_preempt_wait_channel() is introduced, allowing the
migration thread to be able to wait on the channel creation. The channel
is always created by the main thread, in which we'll kick a new semaphore
to tell the migration thread that the channel has created.
We'll need to wait for the new channel in two places: (1) when there's a
new postcopy migration that is starting, or (2) when there's a postcopy
migration to resume.
For the start of migration, we don't need to wait for this channel until
when we want to start postcopy, aka, postcopy_start(). We'll fail the
migration if we found that the channel creation failed (which should
probably not happen at all in 99% of the cases, because the main channel is
using the same network topology).
For a postcopy recovery, we'll need to wait in postcopy_pause(). In that
case if the channel creation failed, we can't fail the migration or we'll
crash the VM, instead we keep in PAUSED state, waiting for yet another
recovery.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Manish Mishra <manish.mishra@nutanix.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185509.27311-1-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
To allow postcopy recovery, the ram fast load (preempt-only) dest QEMU thread
needs similar handling on fault tolerance. When ram_load_postcopy() fails,
instead of stopping the thread it halts with a semaphore, preparing to be
kicked again when recovery is detected.
A mutex is introduced to make sure there's no concurrent operation upon the
socket. To make it simple, the fast ram load thread will take the mutex during
its whole procedure, and only release it if it's paused. The fast-path socket
will be properly released by the main loading thread safely when there's
network failures during postcopy with that mutex held.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185506.27257-1-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
This patch enables postcopy-preempt feature.
It contains two major changes to the migration logic:
(1) Postcopy requests are now sent via a different socket from precopy
background migration stream, so as to be isolated from very high page
request delays.
(2) For huge page enabled hosts: when there's postcopy requests, they can now
intercept a partial sending of huge host pages on src QEMU.
After this patch, we'll live migrate a VM with two channels for postcopy: (1)
PRECOPY channel, which is the default channel that transfers background pages;
and (2) POSTCOPY channel, which only transfers requested pages.
There's no strict rule of which channel to use, e.g., if a requested page is
already being transferred on precopy channel, then we will keep using the same
precopy channel to transfer the page even if it's explicitly requested. In 99%
of the cases we'll prioritize the channels so we send requested page via the
postcopy channel as long as possible.
On the source QEMU, when we found a postcopy request, we'll interrupt the
PRECOPY channel sending process and quickly switch to the POSTCOPY channel.
After we serviced all the high priority postcopy pages, we'll switch back to
PRECOPY channel so that we'll continue to send the interrupted huge page again.
There's no new thread introduced on src QEMU.
On the destination QEMU, one new thread is introduced to receive page data from
the postcopy specific socket (done in the preparation patch).
This patch has a side effect: after sending postcopy pages, previously we'll
assume the guest will access follow up pages so we'll keep sending from there.
Now it's changed. Instead of going on with a postcopy requested page, we'll go
back and continue sending the precopy huge page (which can be intercepted by a
postcopy request so the huge page can be sent partially before).
Whether that's a problem is debatable, because "assuming the guest will
continue to access the next page" may not really suite when huge pages are
used, especially if the huge page is large (e.g. 1GB pages). So that locality
hint is much meaningless if huge pages are used.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185504.27203-1-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Create a new socket for postcopy to be prepared to send postcopy requested
pages via this specific channel, so as to not get blocked by precopy pages.
A new thread is also created on dest qemu to receive data from this new channel
based on the ram_load_postcopy() routine.
The ram_load_postcopy(POSTCOPY) branch and the thread has not started to
function, and that'll be done in follow up patches.
Cleanup the new sockets on both src/dst QEMUs, meanwhile look after the new
thread too to make sure it'll be recycled properly.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185502.27149-1-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
dgilbert: With Peter's fix to quieten compiler warning on
start_migration
Firstly, postcopy already preempts precopy due to the fact that we do
unqueue_page() first before looking into dirty bits.
However that's not enough, e.g., when there're host huge page enabled, when
sending a precopy huge page, a postcopy request needs to wait until the whole
huge page that is sending to finish. That could introduce quite some delay,
the bigger the huge page is the larger delay it'll bring.
This patch adds a new capability to allow postcopy requests to preempt existing
precopy page during sending a huge page, so that postcopy requests can be
serviced even faster.
Meanwhile to send it even faster, bypass the precopy stream by providing a
standalone postcopy socket for sending requested pages.
Since the new behavior will not be compatible with the old behavior, this will
not be the default, it's enabled only when the new capability is set on both
src/dst QEMUs.
This patch only adds the capability itself, the logic will be added in follow
up patches.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220707185342.26794-2-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Previously migration didn't have an easy way to cleanup the listening
transport, migrate recovery only allows to execute once. That's done with a
trick flag in postcopy_recover_triggered.
Now the facility is already there.
Drop postcopy_recover_triggered and instead allows a new migrate-recover to
release the previous listener transport.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-8-peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
We used to use postcopy_try_recover() to replace migration_incoming_setup() to
setup incoming channels. That's fine for the old world, but in the new world
there can be more than one channels that need setup. Better move the channel
setup out of it so that postcopy_try_recover() only handles the last phase of
switching to the recovery phase.
To do that in migration_fd_process_incoming(), move the postcopy_try_recover()
call to be after migration_incoming_setup(), which will setup the channels.
While in migration_ioc_process_incoming(), postpone the recover() routine right
before we'll jump into migration_incoming_process().
A side benefit is we don't need to pass in QEMUFile* to postcopy_try_recover()
anymore. Remove it.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-7-peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
This variable, along with its helpers, is used to detect whether multiple
channel will be supported for migration. In follow up patches, there'll be
other capability that requires multi-channels. Hence move it outside multifd
specific code and make it public. Meanwhile rename it from "multifd" to
"multi_channels" to show its real meaning.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-5-peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
This boolean flag shows whether the current page during migration is triggered
by postcopy or not. Then in ram_save_host_page() and deeper stack we'll be
able to have a reference on the priority of this page.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-4-peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
We used to release it right after migrate_fd_connect(). That's not good
enough when there're more than one socket pair required, because it'll be
needed to establish TLS connection for the rest channels.
One example is multifd, where we copied over the hostname for each channel
but that's actually not needed.
Keeping the hostname until the cleanup phase of migration.
Cc: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-2-peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
dgilbert: Fixup checkpatch error; don't need to check for NULL
around g_free
test_migrate_start() will release the MigrateStart structure that passed
in, however that's not super clear to the caller because after the call
returned the pointer can still be referenced by the callers. It can easily
be a source of use-after-free.
Let's pass in a double pointer of that, then we can safely clear the
pointer for the caller after the struct is released.
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220301083925.33483-26-peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
dgilbert: Fixup apply since I didn't take 24/25
Add a helper to cleanup the transport listener.
When do it, we should also null-ify the cleanup hook and the data, then it's
even safe to call it multiple times.
Move the socket_address_list cleanup altogether, because that's a mirror of the
listener channels and only for the purpose of query-migrate. Hence when
someone wants to cleanup the listener transport, it should also want to cleanup
the socket list too, always.
No functional change intended.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220301083925.33483-15-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
We used to have quite a few places making sure -EIO happened and that's the
only way to trigger postcopy recovery. That's based on the assumption that
we'll only return -EIO for channel issues.
It'll work in 99.99% cases but logically that won't cover some corner cases.
One example is e.g. ram_block_from_stream() could fail with an interrupted
network, then -EINVAL will be returned instead of -EIO.
I remembered Dave Gilbert pointed that out before, but somehow this is
overlooked. Neither did I encounter anything outside the -EIO error.
However we'd better touch that up before it triggers a rare VM data loss during
live migrating.
To cover as much those cases as possible, remove the -EIO restriction on
triggering the postcopy recovery, because even if it's not a channel failure,
we can't do anything better than halting QEMU anyway - the corpse of the
process may even be used by a good hand to dig out useful memory regions, or
the admin could simply kill the process later on.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220301083925.33483-11-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Static variable is very unfriendly to threading of ram_block_from_stream().
Move it into MigrationIncomingState.
Make the incoming state pointer to be passed over to ram_block_from_stream() on
both caller sites.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220301083925.33483-8-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Postcopy create threads. A common manner is we init a sem and use it to sync
with the thread. Namely, we have fault_thread_sem and listen_thread_sem and
they're only used for this.
Make it a shared infrastructure so it's easier to create yet another thread.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220301083925.33483-7-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
In ram_load_postcopy() we'll try to detect non-same-page case and dump error.
This error is very helpful for debugging. Adding ramblock & offset into the
error log too.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220301083925.33483-6-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
dgilbert: Fix up long line
Peter Xu