2020-10-29 20:05:43 +03:00
|
|
|
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
|
|
|
|
From: Stefan Reiter <s.reiter@proxmox.com>
|
|
|
|
Date: Thu, 22 Oct 2020 17:34:18 +0200
|
|
|
|
Subject: [PATCH] PVE: Migrate dirty bitmap state via savevm
|
|
|
|
|
|
|
|
QEMU provides 'savevm' registrations as a mechanism for arbitrary state
|
|
|
|
to be migrated along with a VM. Use this to send a serialized version of
|
|
|
|
dirty bitmap state data from proxmox-backup-qemu, and restore it on the
|
|
|
|
target node.
|
|
|
|
|
|
|
|
Also add a flag to query-proxmox-support so qemu-server can determine if
|
|
|
|
safe migration is possible and makes sense.
|
|
|
|
|
|
|
|
Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
|
2022-01-13 12:34:33 +03:00
|
|
|
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
|
update submodule and patches to QEMU 8.0.0
Many changes were necessary this time around:
* QAPI was changed to avoid redundant has_* variables, see commit
44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C")
for details. This affected many QMP commands added by Proxmox too.
* Pending querying for migration got split into two functions, one to
estimate, one for exact value, see commit c8df4a7aef ("migration:
Split save_live_pending() into state_pending_*") for details. Relevant
for savevm-async and PBS dirty bitmap.
* Some block (driver) functions got converted to coroutines, so the
Proxmox block drivers needed to be adapted.
* Alloc track auto-detaching during PBS live restore got broken by
AioContext-related changes resulting in a deadlock. The current, hacky
method was replaced by a simpler one. Stefan apparently ran into a
problem with that when he wrote the driver, but there were
improvements in the stream job code since then and I didn't manage to
reproduce the issue. It's a separate patch "alloc-track: fix deadlock
during drop" for now, you can find the details there.
* Async snapshot-related changes:
- The pending querying got adapted to the above-mentioned split and
a patch is added to optimize it/make it more similar to what
upstream code does.
- Added initialization of the compression counters (for
future-proofing).
- It's necessary the hold the BQL (big QEMU lock = iothread mutex)
during the setup phase, because block layer functions are used there
and not doing so leads to racy, hard-to-debug crashes or hangs. It's
necessary to change some upstream code too for this, a version of
the patch "migration: for snapshots, hold the BQL during setup
callbacks" is intended to be upstreamed.
- Need to take the bdrv graph read lock before flushing.
* hmp_info_balloon was moved to a different file.
* Needed to include a new headers from time to time to still get the
correct functions.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
|
|
|
[FE: split up state_pending for 8.0]
|
|
|
|
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
|
2020-10-29 20:05:43 +03:00
|
|
|
---
|
2020-11-10 19:16:15 +03:00
|
|
|
include/migration/misc.h | 3 ++
|
2021-02-11 19:11:11 +03:00
|
|
|
migration/meson.build | 2 +
|
2021-05-27 13:43:32 +03:00
|
|
|
migration/migration.c | 1 +
|
update submodule and patches to QEMU 8.0.0
Many changes were necessary this time around:
* QAPI was changed to avoid redundant has_* variables, see commit
44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C")
for details. This affected many QMP commands added by Proxmox too.
* Pending querying for migration got split into two functions, one to
estimate, one for exact value, see commit c8df4a7aef ("migration:
Split save_live_pending() into state_pending_*") for details. Relevant
for savevm-async and PBS dirty bitmap.
* Some block (driver) functions got converted to coroutines, so the
Proxmox block drivers needed to be adapted.
* Alloc track auto-detaching during PBS live restore got broken by
AioContext-related changes resulting in a deadlock. The current, hacky
method was replaced by a simpler one. Stefan apparently ran into a
problem with that when he wrote the driver, but there were
improvements in the stream job code since then and I didn't manage to
reproduce the issue. It's a separate patch "alloc-track: fix deadlock
during drop" for now, you can find the details there.
* Async snapshot-related changes:
- The pending querying got adapted to the above-mentioned split and
a patch is added to optimize it/make it more similar to what
upstream code does.
- Added initialization of the compression counters (for
future-proofing).
- It's necessary the hold the BQL (big QEMU lock = iothread mutex)
during the setup phase, because block layer functions are used there
and not doing so leads to racy, hard-to-debug crashes or hangs. It's
necessary to change some upstream code too for this, a version of
the patch "migration: for snapshots, hold the BQL during setup
callbacks" is intended to be upstreamed.
- Need to take the bdrv graph read lock before flushing.
* hmp_info_balloon was moved to a different file.
* Needed to include a new headers from time to time to still get the
correct functions.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
|
|
|
migration/pbs-state.c | 104 +++++++++++++++++++++++++++++++++++++++
|
2020-11-10 19:16:15 +03:00
|
|
|
pve-backup.c | 1 +
|
|
|
|
qapi/block-core.json | 6 +++
|
update submodule and patches to QEMU 8.0.0
Many changes were necessary this time around:
* QAPI was changed to avoid redundant has_* variables, see commit
44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C")
for details. This affected many QMP commands added by Proxmox too.
* Pending querying for migration got split into two functions, one to
estimate, one for exact value, see commit c8df4a7aef ("migration:
Split save_live_pending() into state_pending_*") for details. Relevant
for savevm-async and PBS dirty bitmap.
* Some block (driver) functions got converted to coroutines, so the
Proxmox block drivers needed to be adapted.
* Alloc track auto-detaching during PBS live restore got broken by
AioContext-related changes resulting in a deadlock. The current, hacky
method was replaced by a simpler one. Stefan apparently ran into a
problem with that when he wrote the driver, but there were
improvements in the stream job code since then and I didn't manage to
reproduce the issue. It's a separate patch "alloc-track: fix deadlock
during drop" for now, you can find the details there.
* Async snapshot-related changes:
- The pending querying got adapted to the above-mentioned split and
a patch is added to optimize it/make it more similar to what
upstream code does.
- Added initialization of the compression counters (for
future-proofing).
- It's necessary the hold the BQL (big QEMU lock = iothread mutex)
during the setup phase, because block layer functions are used there
and not doing so leads to racy, hard-to-debug crashes or hangs. It's
necessary to change some upstream code too for this, a version of
the patch "migration: for snapshots, hold the BQL during setup
callbacks" is intended to be upstreamed.
- Need to take the bdrv graph read lock before flushing.
* hmp_info_balloon was moved to a different file.
* Needed to include a new headers from time to time to still get the
correct functions.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
|
|
|
6 files changed, 117 insertions(+)
|
2020-10-29 20:05:43 +03:00
|
|
|
create mode 100644 migration/pbs-state.c
|
|
|
|
|
|
|
|
diff --git a/include/migration/misc.h b/include/migration/misc.h
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
index 1bc8902e6d..43eb0e46d6 100644
|
2020-10-29 20:05:43 +03:00
|
|
|
--- a/include/migration/misc.h
|
|
|
|
+++ b/include/migration/misc.h
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
@@ -80,4 +80,7 @@ bool migration_in_bg_snapshot(void);
|
2020-10-29 20:05:43 +03:00
|
|
|
/* migration/block-dirty-bitmap.c */
|
|
|
|
void dirty_bitmap_mig_init(void);
|
|
|
|
|
|
|
|
+/* migration/pbs-state.c */
|
|
|
|
+void pbs_state_mig_init(void);
|
|
|
|
+
|
|
|
|
#endif
|
2021-02-11 19:11:11 +03:00
|
|
|
diff --git a/migration/meson.build b/migration/meson.build
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
index 8f9d122187..2b9d2cc881 100644
|
2021-02-11 19:11:11 +03:00
|
|
|
--- a/migration/meson.build
|
|
|
|
+++ b/migration/meson.build
|
2023-10-17 15:10:09 +03:00
|
|
|
@@ -7,7 +7,9 @@ migration_files = files(
|
update submodule and patches to 7.1.0
Notable changes:
* The only big change is the switch to using a custom QIOChannel for
savevm-async, because the previously used QEMUFileOps was dropped.
Changes to the current implementation:
* Switch to vector based methods as required for an IO channel. For
short reads the passed-in IO vector is stuffed with zeroes at the
end, just to be sure.
* For reading: The documentation in include/io/channel.h states that
at least one byte should be read, so also error out when whe are
at the very end instead of returning 0.
* For reading: Fix off-by-one error when request goes beyond end.
The wrong code piece was:
if ((pos + size) > maxlen) {
size = maxlen - pos - 1;
}
Previously, the last byte would not be read. It's actually
possible to get a snapshot .raw file that has content all the way
up the final 512 byte (= BDRV_SECTOR_SIZE) boundary without any
trailing zero bytes (I wrote a script to do it).
Luckily, it didn't cause a real issue, because qemu_loadvm_state()
is not interested in the final (i.e. QEMU_VM_VMDESCRIPTION)
section. The buffer for reading it is simply freed up afterwards
and the function will assume that it read the whole section, even
if that's not the case.
* For writing: Make use of the generated blk_pwritev() wrapper
instead of manually wrapping the coroutine to simplify and save a
few lines.
* Adapt to changed interfaces for blk_{pread,pwrite}:
* a9262f551e ("block: Change blk_{pread,pwrite}() param order")
* 3b35d4542c ("block: Add a 'flags' param to blk_pread()")
* bf5b16fa40 ("block: Make blk_{pread,pwrite}() return 0 on success")
Those changes especially affected the qemu-img dd patches, because
the context also changed, but also some of our block drivers used
the functions.
* Drop qemu-common.h include: it got renamed after essentially
everything was moved to other headers. The only remaining user I
could find for things dropped from the header between 7.0 and 7.1
was qemu_get_vm_name() in the iscsi-initiatorname patch, but it
already includes the header to which the function was moved.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2022-10-14 15:07:13 +03:00
|
|
|
'vmstate.c',
|
2021-02-11 19:11:11 +03:00
|
|
|
'qemu-file.c',
|
2021-05-27 13:43:32 +03:00
|
|
|
'yank_functions.c',
|
2021-02-11 19:11:11 +03:00
|
|
|
+ 'pbs-state.c',
|
|
|
|
)
|
2023-10-17 15:10:09 +03:00
|
|
|
+system_ss.add(libproxmox_backup_qemu)
|
2020-10-29 20:05:43 +03:00
|
|
|
|
2023-10-17 15:10:09 +03:00
|
|
|
system_ss.add(files(
|
2021-02-11 19:11:11 +03:00
|
|
|
'block-dirty-bitmap.c',
|
2021-05-27 13:43:32 +03:00
|
|
|
diff --git a/migration/migration.c b/migration/migration.c
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
index 982ab85f04..b84924442d 100644
|
2021-05-27 13:43:32 +03:00
|
|
|
--- a/migration/migration.c
|
|
|
|
+++ b/migration/migration.c
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
@@ -202,6 +202,7 @@ void migration_object_init(void)
|
2021-05-27 13:43:32 +03:00
|
|
|
blk_mig_init();
|
|
|
|
ram_mig_init();
|
|
|
|
dirty_bitmap_mig_init();
|
|
|
|
+ pbs_state_mig_init();
|
|
|
|
}
|
|
|
|
|
2022-02-11 12:24:33 +03:00
|
|
|
void migration_cancel(const Error *error)
|
2020-10-29 20:05:43 +03:00
|
|
|
diff --git a/migration/pbs-state.c b/migration/pbs-state.c
|
|
|
|
new file mode 100644
|
update submodule and patches to QEMU 8.0.0
Many changes were necessary this time around:
* QAPI was changed to avoid redundant has_* variables, see commit
44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C")
for details. This affected many QMP commands added by Proxmox too.
* Pending querying for migration got split into two functions, one to
estimate, one for exact value, see commit c8df4a7aef ("migration:
Split save_live_pending() into state_pending_*") for details. Relevant
for savevm-async and PBS dirty bitmap.
* Some block (driver) functions got converted to coroutines, so the
Proxmox block drivers needed to be adapted.
* Alloc track auto-detaching during PBS live restore got broken by
AioContext-related changes resulting in a deadlock. The current, hacky
method was replaced by a simpler one. Stefan apparently ran into a
problem with that when he wrote the driver, but there were
improvements in the stream job code since then and I didn't manage to
reproduce the issue. It's a separate patch "alloc-track: fix deadlock
during drop" for now, you can find the details there.
* Async snapshot-related changes:
- The pending querying got adapted to the above-mentioned split and
a patch is added to optimize it/make it more similar to what
upstream code does.
- Added initialization of the compression counters (for
future-proofing).
- It's necessary the hold the BQL (big QEMU lock = iothread mutex)
during the setup phase, because block layer functions are used there
and not doing so leads to racy, hard-to-debug crashes or hangs. It's
necessary to change some upstream code too for this, a version of
the patch "migration: for snapshots, hold the BQL during setup
callbacks" is intended to be upstreamed.
- Need to take the bdrv graph read lock before flushing.
* hmp_info_balloon was moved to a different file.
* Needed to include a new headers from time to time to still get the
correct functions.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
|
|
|
index 0000000000..887e998b9e
|
2020-10-29 20:05:43 +03:00
|
|
|
--- /dev/null
|
|
|
|
+++ b/migration/pbs-state.c
|
update submodule and patches to QEMU 8.0.0
Many changes were necessary this time around:
* QAPI was changed to avoid redundant has_* variables, see commit
44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C")
for details. This affected many QMP commands added by Proxmox too.
* Pending querying for migration got split into two functions, one to
estimate, one for exact value, see commit c8df4a7aef ("migration:
Split save_live_pending() into state_pending_*") for details. Relevant
for savevm-async and PBS dirty bitmap.
* Some block (driver) functions got converted to coroutines, so the
Proxmox block drivers needed to be adapted.
* Alloc track auto-detaching during PBS live restore got broken by
AioContext-related changes resulting in a deadlock. The current, hacky
method was replaced by a simpler one. Stefan apparently ran into a
problem with that when he wrote the driver, but there were
improvements in the stream job code since then and I didn't manage to
reproduce the issue. It's a separate patch "alloc-track: fix deadlock
during drop" for now, you can find the details there.
* Async snapshot-related changes:
- The pending querying got adapted to the above-mentioned split and
a patch is added to optimize it/make it more similar to what
upstream code does.
- Added initialization of the compression counters (for
future-proofing).
- It's necessary the hold the BQL (big QEMU lock = iothread mutex)
during the setup phase, because block layer functions are used there
and not doing so leads to racy, hard-to-debug crashes or hangs. It's
necessary to change some upstream code too for this, a version of
the patch "migration: for snapshots, hold the BQL during setup
callbacks" is intended to be upstreamed.
- Need to take the bdrv graph read lock before flushing.
* hmp_info_balloon was moved to a different file.
* Needed to include a new headers from time to time to still get the
correct functions.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
|
|
|
@@ -0,0 +1,104 @@
|
2020-10-29 20:05:43 +03:00
|
|
|
+/*
|
|
|
|
+ * PBS (dirty-bitmap) state migration
|
|
|
|
+ */
|
|
|
|
+
|
|
|
|
+#include "qemu/osdep.h"
|
|
|
|
+#include "migration/misc.h"
|
|
|
|
+#include "qemu-file.h"
|
|
|
|
+#include "migration/vmstate.h"
|
|
|
|
+#include "migration/register.h"
|
|
|
|
+#include "proxmox-backup-qemu.h"
|
|
|
|
+
|
fix dirty-bitmap state migration freeze
The idea in general is to migrate all the state, which is small for
us, in a single step once. But, QEMU only calls save state if we
return active true.
Hardcoding is-active to return true, like done initially, makes the
migration freeze, as QEMU thinks this is never done, and only stops
calling us and finishes after a few seconds.
So, add a state with an "active" boolean, set to true when
initializing a migration, and set it to false when the state was
saved.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2020-11-05 20:43:01 +03:00
|
|
|
+typedef struct PBSState {
|
|
|
|
+ bool active;
|
|
|
|
+} PBSState;
|
|
|
|
+
|
2020-11-10 19:16:15 +03:00
|
|
|
+/* state is accessed via this static variable directly, 'opaque' is NULL */
|
fix dirty-bitmap state migration freeze
The idea in general is to migrate all the state, which is small for
us, in a single step once. But, QEMU only calls save state if we
return active true.
Hardcoding is-active to return true, like done initially, makes the
migration freeze, as QEMU thinks this is never done, and only stops
calling us and finishes after a few seconds.
So, add a state with an "active" boolean, set to true when
initializing a migration, and set it to false when the state was
saved.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2020-11-05 20:43:01 +03:00
|
|
|
+static PBSState pbs_state;
|
|
|
|
+
|
update submodule and patches to QEMU 8.0.0
Many changes were necessary this time around:
* QAPI was changed to avoid redundant has_* variables, see commit
44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C")
for details. This affected many QMP commands added by Proxmox too.
* Pending querying for migration got split into two functions, one to
estimate, one for exact value, see commit c8df4a7aef ("migration:
Split save_live_pending() into state_pending_*") for details. Relevant
for savevm-async and PBS dirty bitmap.
* Some block (driver) functions got converted to coroutines, so the
Proxmox block drivers needed to be adapted.
* Alloc track auto-detaching during PBS live restore got broken by
AioContext-related changes resulting in a deadlock. The current, hacky
method was replaced by a simpler one. Stefan apparently ran into a
problem with that when he wrote the driver, but there were
improvements in the stream job code since then and I didn't manage to
reproduce the issue. It's a separate patch "alloc-track: fix deadlock
during drop" for now, you can find the details there.
* Async snapshot-related changes:
- The pending querying got adapted to the above-mentioned split and
a patch is added to optimize it/make it more similar to what
upstream code does.
- Added initialization of the compression counters (for
future-proofing).
- It's necessary the hold the BQL (big QEMU lock = iothread mutex)
during the setup phase, because block layer functions are used there
and not doing so leads to racy, hard-to-debug crashes or hangs. It's
necessary to change some upstream code too for this, a version of
the patch "migration: for snapshots, hold the BQL during setup
callbacks" is intended to be upstreamed.
- Need to take the bdrv graph read lock before flushing.
* hmp_info_balloon was moved to a different file.
* Needed to include a new headers from time to time to still get the
correct functions.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
|
|
|
+static void pbs_state_pending(void *opaque, uint64_t *must_precopy,
|
|
|
|
+ uint64_t *can_postcopy)
|
2020-10-29 20:05:43 +03:00
|
|
|
+{
|
|
|
|
+ /* we send everything in save_setup, so nothing is ever pending */
|
|
|
|
+}
|
|
|
|
+
|
|
|
|
+/* receive PBS state via f and deserialize, called on target */
|
|
|
|
+static int pbs_state_load(QEMUFile *f, void *opaque, int version_id)
|
|
|
|
+{
|
|
|
|
+ /* safe cast, we cannot migrate to target with less bits than source */
|
|
|
|
+ size_t buf_size = (size_t)qemu_get_be64(f);
|
|
|
|
+
|
|
|
|
+ uint8_t *buf = (uint8_t *)malloc(buf_size);
|
|
|
|
+ size_t read = qemu_get_buffer(f, buf, buf_size);
|
|
|
|
+
|
|
|
|
+ if (read < buf_size) {
|
|
|
|
+ fprintf(stderr, "error receiving PBS state: not enough data\n");
|
|
|
|
+ return -EIO;
|
|
|
|
+ }
|
|
|
|
+
|
|
|
|
+ proxmox_import_state(buf, buf_size);
|
|
|
|
+
|
|
|
|
+ free(buf);
|
|
|
|
+ return 0;
|
|
|
|
+}
|
|
|
|
+
|
|
|
|
+/* serialize PBS state and send to target via f, called on source */
|
|
|
|
+static int pbs_state_save_setup(QEMUFile *f, void *opaque)
|
|
|
|
+{
|
|
|
|
+ size_t buf_size;
|
|
|
|
+ uint8_t *buf = proxmox_export_state(&buf_size);
|
|
|
|
+
|
|
|
|
+ /* LV encoding */
|
|
|
|
+ qemu_put_be64(f, buf_size);
|
|
|
|
+ qemu_put_buffer(f, buf, buf_size);
|
|
|
|
+
|
|
|
|
+ proxmox_free_state_buf(buf);
|
fix dirty-bitmap state migration freeze
The idea in general is to migrate all the state, which is small for
us, in a single step once. But, QEMU only calls save state if we
return active true.
Hardcoding is-active to return true, like done initially, makes the
migration freeze, as QEMU thinks this is never done, and only stops
calling us and finishes after a few seconds.
So, add a state with an "active" boolean, set to true when
initializing a migration, and set it to false when the state was
saved.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2020-11-05 20:43:01 +03:00
|
|
|
+ pbs_state.active = false;
|
2020-10-29 20:05:43 +03:00
|
|
|
+ return 0;
|
|
|
|
+}
|
|
|
|
+
|
|
|
|
+static bool pbs_state_is_active(void *opaque)
|
|
|
|
+{
|
2020-11-10 19:16:15 +03:00
|
|
|
+ /* we need to return active exactly once, else .save_setup is never called,
|
|
|
|
+ * but if we'd just return true the migration doesn't make progress since
|
|
|
|
+ * it'd be waiting for us */
|
fix dirty-bitmap state migration freeze
The idea in general is to migrate all the state, which is small for
us, in a single step once. But, QEMU only calls save state if we
return active true.
Hardcoding is-active to return true, like done initially, makes the
migration freeze, as QEMU thinks this is never done, and only stops
calling us and finishes after a few seconds.
So, add a state with an "active" boolean, set to true when
initializing a migration, and set it to false when the state was
saved.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2020-11-05 20:43:01 +03:00
|
|
|
+ return pbs_state.active;
|
2020-10-29 20:05:43 +03:00
|
|
|
+}
|
|
|
|
+
|
|
|
|
+static bool pbs_state_is_active_iterate(void *opaque)
|
|
|
|
+{
|
|
|
|
+ /* we don't iterate, everything is sent in save_setup */
|
fix dirty-bitmap state migration freeze
The idea in general is to migrate all the state, which is small for
us, in a single step once. But, QEMU only calls save state if we
return active true.
Hardcoding is-active to return true, like done initially, makes the
migration freeze, as QEMU thinks this is never done, and only stops
calling us and finishes after a few seconds.
So, add a state with an "active" boolean, set to true when
initializing a migration, and set it to false when the state was
saved.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2020-11-05 20:43:01 +03:00
|
|
|
+ return pbs_state_is_active(opaque);
|
2020-10-29 20:05:43 +03:00
|
|
|
+}
|
|
|
|
+
|
|
|
|
+static bool pbs_state_has_postcopy(void *opaque)
|
|
|
|
+{
|
|
|
|
+ /* PBS state can't change during a migration (since that's blocking any
|
|
|
|
+ * potential backups), so we can copy everything before the VM is stopped */
|
|
|
|
+ return false;
|
|
|
|
+}
|
|
|
|
+
|
2020-11-10 19:16:15 +03:00
|
|
|
+static void pbs_state_save_cleanup(void *opaque)
|
|
|
|
+{
|
|
|
|
+ /* reset active after migration succeeds or fails */
|
|
|
|
+ pbs_state.active = false;
|
|
|
|
+}
|
|
|
|
+
|
2020-10-29 20:05:43 +03:00
|
|
|
+static SaveVMHandlers savevm_pbs_state_handlers = {
|
|
|
|
+ .save_setup = pbs_state_save_setup,
|
|
|
|
+ .has_postcopy = pbs_state_has_postcopy,
|
update submodule and patches to QEMU 8.0.0
Many changes were necessary this time around:
* QAPI was changed to avoid redundant has_* variables, see commit
44ea9d9be3 ("qapi: Start to elide redundant has_FOO in generated C")
for details. This affected many QMP commands added by Proxmox too.
* Pending querying for migration got split into two functions, one to
estimate, one for exact value, see commit c8df4a7aef ("migration:
Split save_live_pending() into state_pending_*") for details. Relevant
for savevm-async and PBS dirty bitmap.
* Some block (driver) functions got converted to coroutines, so the
Proxmox block drivers needed to be adapted.
* Alloc track auto-detaching during PBS live restore got broken by
AioContext-related changes resulting in a deadlock. The current, hacky
method was replaced by a simpler one. Stefan apparently ran into a
problem with that when he wrote the driver, but there were
improvements in the stream job code since then and I didn't manage to
reproduce the issue. It's a separate patch "alloc-track: fix deadlock
during drop" for now, you can find the details there.
* Async snapshot-related changes:
- The pending querying got adapted to the above-mentioned split and
a patch is added to optimize it/make it more similar to what
upstream code does.
- Added initialization of the compression counters (for
future-proofing).
- It's necessary the hold the BQL (big QEMU lock = iothread mutex)
during the setup phase, because block layer functions are used there
and not doing so leads to racy, hard-to-debug crashes or hangs. It's
necessary to change some upstream code too for this, a version of
the patch "migration: for snapshots, hold the BQL during setup
callbacks" is intended to be upstreamed.
- Need to take the bdrv graph read lock before flushing.
* hmp_info_balloon was moved to a different file.
* Needed to include a new headers from time to time to still get the
correct functions.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:53 +03:00
|
|
|
+ .state_pending_exact = pbs_state_pending,
|
|
|
|
+ .state_pending_estimate = pbs_state_pending,
|
2020-10-29 20:05:43 +03:00
|
|
|
+ .is_active_iterate = pbs_state_is_active_iterate,
|
|
|
|
+ .load_state = pbs_state_load,
|
|
|
|
+ .is_active = pbs_state_is_active,
|
2020-11-10 19:16:15 +03:00
|
|
|
+ .save_cleanup = pbs_state_save_cleanup,
|
2020-10-29 20:05:43 +03:00
|
|
|
+};
|
|
|
|
+
|
|
|
|
+void pbs_state_mig_init(void)
|
|
|
|
+{
|
fix dirty-bitmap state migration freeze
The idea in general is to migrate all the state, which is small for
us, in a single step once. But, QEMU only calls save state if we
return active true.
Hardcoding is-active to return true, like done initially, makes the
migration freeze, as QEMU thinks this is never done, and only stops
calling us and finishes after a few seconds.
So, add a state with an "active" boolean, set to true when
initializing a migration, and set it to false when the state was
saved.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2020-11-05 20:43:01 +03:00
|
|
|
+ pbs_state.active = true;
|
2020-10-29 20:05:43 +03:00
|
|
|
+ register_savevm_live("pbs-state", 0, 1,
|
|
|
|
+ &savevm_pbs_state_handlers,
|
2020-11-10 19:16:15 +03:00
|
|
|
+ NULL);
|
2020-10-29 20:05:43 +03:00
|
|
|
+}
|
|
|
|
diff --git a/pve-backup.c b/pve-backup.c
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
index 903afcd7e9..777db7938e 100644
|
2020-10-29 20:05:43 +03:00
|
|
|
--- a/pve-backup.c
|
|
|
|
+++ b/pve-backup.c
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
@@ -1096,6 +1096,7 @@ ProxmoxSupportStatus *qmp_query_proxmox_support(Error **errp)
|
2020-11-24 18:41:20 +03:00
|
|
|
ret->pbs_library_version = g_strdup(proxmox_backup_qemu_version());
|
2020-10-29 20:05:43 +03:00
|
|
|
ret->pbs_dirty_bitmap = true;
|
2021-03-16 19:30:22 +03:00
|
|
|
ret->pbs_dirty_bitmap_savevm = true;
|
2020-10-29 20:05:43 +03:00
|
|
|
+ ret->pbs_dirty_bitmap_migration = true;
|
2021-03-16 19:30:22 +03:00
|
|
|
ret->query_bitmap_info = true;
|
squash related patches
where there is no good reason to keep them separate. It's a pain
during rebase if there are multiple patches changing the same code
over and over again. This was especially bad for the backup-related
patches. If the history of patches really is needed, it can be
extracted via git. Additionally, compilation with partial application
of patches was broken since a long time, because one of the master key
changes became part of an earlier patch during a past rebase.
If only the same files were changed by a subsequent patch and the
changes felt to belong together (obvious for later bug fixes, but also
done for features e.g. adding master key support for PBS), the patches
were squashed together.
The PBS namespace support patch was split into the individual parts
it changes, i.e. PBS block driver, pbs-restore binary and QMP backup
infrastructure, and squashed into the respective patches.
No code change is intended, git diff in the submodule should not show
any difference between applying all patches before this commit and
applying all patches after this commit.
The query-proxmox-support QMP function has been left as part of the
"PVE-Backup: Proxmox backup patches for QEMU" patch, because it's
currently only used there. If it ever is used elsewhere too, it can
be split out from there.
The recent alloc-track and BQL-related savevm-async changes have been
left separate for now, because it's not 100% clear they are the best
approach yet. This depends on what upstream decides about the BQL
stuff and whether and what kind of issues with the changes pop up.
The qemu-img dd snapshot patch has been re-ordered to after the other
qemu-img dd patches.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:56 +03:00
|
|
|
ret->pbs_masterkey = true;
|
|
|
|
ret->backup_max_workers = true;
|
2020-10-29 20:05:43 +03:00
|
|
|
diff --git a/qapi/block-core.json b/qapi/block-core.json
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
index a4050268ca..7b977459fa 100644
|
2020-10-29 20:05:43 +03:00
|
|
|
--- a/qapi/block-core.json
|
|
|
|
+++ b/qapi/block-core.json
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
@@ -993,6 +993,11 @@
|
2021-03-16 19:30:22 +03:00
|
|
|
# @pbs-dirty-bitmap-savevm: True if 'dirty-bitmaps' migration capability can
|
|
|
|
# safely be set for savevm-async.
|
2020-10-29 20:05:43 +03:00
|
|
|
#
|
|
|
|
+# @pbs-dirty-bitmap-migration: True if safe migration of dirty-bitmaps including
|
|
|
|
+# PBS state is supported. Enabling 'dirty-bitmaps'
|
|
|
|
+# migration cap if this is false/unset may lead
|
|
|
|
+# to crashes on migration!
|
|
|
|
+#
|
squash related patches
where there is no good reason to keep them separate. It's a pain
during rebase if there are multiple patches changing the same code
over and over again. This was especially bad for the backup-related
patches. If the history of patches really is needed, it can be
extracted via git. Additionally, compilation with partial application
of patches was broken since a long time, because one of the master key
changes became part of an earlier patch during a past rebase.
If only the same files were changed by a subsequent patch and the
changes felt to belong together (obvious for later bug fixes, but also
done for features e.g. adding master key support for PBS), the patches
were squashed together.
The PBS namespace support patch was split into the individual parts
it changes, i.e. PBS block driver, pbs-restore binary and QMP backup
infrastructure, and squashed into the respective patches.
No code change is intended, git diff in the submodule should not show
any difference between applying all patches before this commit and
applying all patches after this commit.
The query-proxmox-support QMP function has been left as part of the
"PVE-Backup: Proxmox backup patches for QEMU" patch, because it's
currently only used there. If it ever is used elsewhere too, it can
be split out from there.
The recent alloc-track and BQL-related savevm-async changes have been
left separate for now, because it's not 100% clear they are the best
approach yet. This depends on what upstream decides about the BQL
stuff and whether and what kind of issues with the changes pop up.
The qemu-img dd snapshot patch has been re-ordered to after the other
qemu-img dd patches.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:56 +03:00
|
|
|
# @pbs-masterkey: True if the QMP backup call supports the 'master_keyfile'
|
|
|
|
# parameter.
|
2020-11-24 18:41:20 +03:00
|
|
|
#
|
update submodule and patches to QEMU 8.2.2
This version includes both the AioContext lock and the block graph
lock, so there might be some deadlocks lurking. It's not possible to
disable the block graph lock like was done in QEMU 8.1, because there
are no changes like the function bdrv_schedule_unref() that require
it. QEMU 9.0 will finally get rid of the AioContext locking.
During live-restore with a VirtIO SCSI drive with iothread there is a
known racy deadlock related to the AioContext lock. Not new [1], but
not sure if more likely now. Should be fixed in QEMU 9.0.
The block graph lock comes with annotations that can be checked by
clang's TSA. This required changes to the block drivers, i.e.
alloc-track, pbs, zeroinit as well as taking the appropriate locks
in pve-backup, savevm-async, vma-reader.
Local variable shadowing is prohibited via a compiler flag now,
required slight adaptation in vma.c.
Major changes only affect alloc-track:
* It is not possible to call a generated co-wrapper like
bdrv_get_info() while holding the block graph lock exclusively [0],
which does happen during initialization of alloc-track when the
backing hd is set and the refresh_limits driver callback is invoked.
The bdrv_get_info() call to get the cluster size is moved to
directly after opening the file child in track_open().
The important thing is that at least the request alignment for the
write target is used, because then the RMW cycle in bdrv_pwritev
will gather enough data from the backing file. Partial cluster
allocations in the target are not a fundamental issue, because the
driver returns its allocation status based on the bitmap, so any
other data that maps to the same cluster will still be copied later
by a stream job (or during writes to that cluster).
* Replacing the node cannot be done in the
track_co_change_backing_file() callback, because it is a coroutine
and cannot hold the block graph lock exclusively. So it is moved to
the stream job itself with the auto-remove option not having an
effect anymore (qemu-server would always set it anyways).
In the future, there could either be a special option for the stream
job, or maybe the upcoming blockdev-replace QMP command can be used.
Replacing the backing child is actually already done in the stream
job, so no need to do it in the track_co_change_backing_file()
callback. It also cannot be called from a coroutine. Looking at the
implementation in the qcow2 driver, it doesn't seem to be intended
to change the backing child itself, just update driver-internal
state.
Other changes:
* alloc-track: Error out early when used without auto-remove. Since
replacing the node now happens in the stream job, where the option
cannot be read from (it's internal to the driver), it will always be
treated as 'on'. Makes sure to have users beside qemu-server notice
the change (should they even exist). The option can be fully dropped
in the future while adding a version guard in qemu-server.
* alloc-track: Avoid seemingly superfluous child permission update.
Doesn't seem necessary nowadays (maybe after commit "alloc-track:
fix deadlock during drop" where the dropping is not rescheduled and
delayed anymore or some upstream change). Replacing the block node
will already update the permissions of the new node (which was the
file child before). Should there really be some issue, instead of
having a drop state, this could also be just based off the fact
whether there is still a backing child.
Dumping the cumulative (shared) permissions for the BDS with a debug
print yields the same values after this patch and with QEMU 8.1,
namely 3 and 5.
* PBS block driver: compile unconditionally. Proxmox VE always needs
it and something in the build process changed to make it not enabled
by default. Probably would need to move the build option to meson
otherwise.
* backup: job unreferencing during cleanup needs to happen outside of
coroutine, so it was moved to before invoking the clean
* mirror: Cherry-pick stable fix to avoid potential deadlock.
* savevm-async: migrate_init now can fail, so propagate potential
error.
* savevm-async: compression counters are not accessible outside
migration/ram-compress now, so drop code that prophylactically set
it to zero.
[0]: https://lore.kernel.org/qemu-devel/220be383-3b0d-4938-b584-69ad214e5d5d@proxmox.com/
[1]: https://lore.kernel.org/qemu-devel/e13b488e-bf13-44f2-acca-e724d14f43fd@proxmox.com/
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2024-04-25 18:21:28 +03:00
|
|
|
@@ -1003,6 +1008,7 @@
|
2020-11-24 18:41:20 +03:00
|
|
|
'data': { 'pbs-dirty-bitmap': 'bool',
|
|
|
|
'query-bitmap-info': 'bool',
|
2021-03-16 19:30:22 +03:00
|
|
|
'pbs-dirty-bitmap-savevm': 'bool',
|
2020-11-24 18:41:20 +03:00
|
|
|
+ 'pbs-dirty-bitmap-migration': 'bool',
|
squash related patches
where there is no good reason to keep them separate. It's a pain
during rebase if there are multiple patches changing the same code
over and over again. This was especially bad for the backup-related
patches. If the history of patches really is needed, it can be
extracted via git. Additionally, compilation with partial application
of patches was broken since a long time, because one of the master key
changes became part of an earlier patch during a past rebase.
If only the same files were changed by a subsequent patch and the
changes felt to belong together (obvious for later bug fixes, but also
done for features e.g. adding master key support for PBS), the patches
were squashed together.
The PBS namespace support patch was split into the individual parts
it changes, i.e. PBS block driver, pbs-restore binary and QMP backup
infrastructure, and squashed into the respective patches.
No code change is intended, git diff in the submodule should not show
any difference between applying all patches before this commit and
applying all patches after this commit.
The query-proxmox-support QMP function has been left as part of the
"PVE-Backup: Proxmox backup patches for QEMU" patch, because it's
currently only used there. If it ever is used elsewhere too, it can
be split out from there.
The recent alloc-track and BQL-related savevm-async changes have been
left separate for now, because it's not 100% clear they are the best
approach yet. This depends on what upstream decides about the BQL
stuff and whether and what kind of issues with the changes pop up.
The qemu-img dd snapshot patch has been re-ordered to after the other
qemu-img dd patches.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
2023-05-15 16:39:56 +03:00
|
|
|
'pbs-masterkey': 'bool',
|
|
|
|
'pbs-library-version': 'str',
|
|
|
|
'backup-max-workers': 'bool' } }
|