* [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP
@ 2020-09-15 11:35 Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 1/9] block: push error reporting into bdrv_all_*_snapshot functions Daniel P. Berrangé
` (8 more replies)
0 siblings, 9 replies; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
v1: https://lists.gnu.org/archive/html/qemu-devel/2020-07/msg00866.html
v2: https://lists.gnu.org/archive/html/qemu-devel/2020-07/msg07523.html
v3: https://lists.gnu.org/archive/html/qemu-devel/2020-08/msg07076.html
This series aims to provide a better designed replacement for the
savevm/loadvm/delvm HMP commands, which despite their flaws continue
to be actively used in the QMP world via the HMP command passthrough
facility.
The main problems addressed are:
- The logic to pick which disk to store the vmstate in is not
satsifactory.
The first block driver state cannot be assumed to be the root disk
image, it might be OVMF varstore and we don't want to store vmstate
in there.
- The logic to decide which disks must be snapshotted is hardwired
to all disks which are writable
Again with OVMF there might be a writable varstore, but this can be
raw rather than qcow2 format, and thus unable to be snapshotted.
While users might wish to snapshot their varstore, in some/many/most
cases it is entirely uneccessary. Users are blocked from snapshotting
their VM though due to this varstore.
- The commands are synchronous blocking execution and returning
errors immediately.
This is partially addressed by integrating with the job framework.
This forces the client to use the async commands to determine
the completion status or error message from the operations.
In the block code I've only dealt with node names for block devices, as
IIUC, this is all that libvirt should need in the -blockdev world it now
lives in. IOW, I've made not attempt to cope with people wanting to use
these QMP commands in combination with -drive args, as libvirt will
never use -drive with a QEMU new enough to have these new commands.
The main limitations of this current impl
- The snapshot process runs serialized in the main thread. ie QEMU
guest execution is blocked for the duration. The job framework
lets us fix this in future without changing the QMP semantics
exposed to the apps.
- Most vmstate loading errors just go to stderr, as they are not
using Error **errp reporting. Thus the job framework just
reports a fairly generic message
"Error -22 while loading VM state"
Again this can be fixed later without changing the QMP semantics
exposed to apps.
I've done some minimal work in libvirt to start to make use of the new
commands to validate their functionality, but this isn't finished yet.
My ultimate goal is to make the GNOME Boxes maintainer happy again by
having internal snapshots work with OVMF:
https://gitlab.gnome.org/GNOME/gnome-boxes/-/commit/c486da262f6566326fbcb5e=
f45c5f64048f16a6e
Changed in v4:
- Make the device lists mandatory, dropping all support for
QEMU's built-in heuristics to select devices.
- Improve some error reporting and I/O test coverage
Changed in v3:
- Schedule a bottom half to escape from coroutine context in
the jobs. This is needed because the locking in the snapshot
code goes horribly wrong when run from a background coroutine
instead of the main event thread.
- Re-factor way we iterate over devices, so that we correctly
report non-existant devices passed by the user over QMP.
- Add QAPI docs notes about limitations wrt vmstate error
reporting (it all goes to stderr not an Error **errp)
so QMP only gets a fairly generic error message currently.
- Add I/O test to validate many usage scenarios / errors
- Add I/O test helpers to handle QMP events with a deterministic
ordering
- Ensure 'delete-snapshot' reports an error if requesting
delete from devices that don't support snapshot, instead of
silently succeeding with no erro.
Changed in v2:
- Use new command names "snapshot-{load,save,delete}" to make it
clear that these are different from the "savevm|loadvm|delvm"
as they use the Job framework
- Use an include list for block devs, not an exclude list
Daniel P. Berrang=C3=A9 (9):
block: push error reporting into bdrv_all_*_snapshot functions
migration: stop returning errno from load_snapshot()
block: add ability to specify list of blockdevs during snapshot
block: allow specifying name of block device for vmstate storage
migration: control whether snapshots are ovewritten
migration: wire up support for snapshot device selection
migration: introduce a delete_snapshot wrapper
iotests: add support for capturing and matching QMP events
migration: introduce snapshot-{save,load,delete} QMP commands
block/monitor/block-hmp-cmds.c | 7 +-
block/snapshot.c | 248 ++++++++++++++------
include/block/snapshot.h | 23 +-
include/migration/snapshot.h | 14 +-
migration/savevm.c | 260 +++++++++++++++++----
monitor/hmp-cmds.c | 11 +-
qapi/job.json | 9 +-
qapi/migration.json | 120 ++++++++++
replay/replay-snapshot.c | 4 +-
softmmu/vl.c | 2 +-
tests/qemu-iotests/267.out | 10 +-
tests/qemu-iotests/310 | 338 +++++++++++++++++++++++++++
tests/qemu-iotests/310.out | 412 +++++++++++++++++++++++++++++++++
tests/qemu-iotests/common.qemu | 107 ++++++++-
tests/qemu-iotests/group | 1 +
15 files changed, 1425 insertions(+), 141 deletions(-)
create mode 100755 tests/qemu-iotests/310
create mode 100644 tests/qemu-iotests/310.out
--=20
2.26.2
^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH v4 1/9] block: push error reporting into bdrv_all_*_snapshot functions
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
@ 2020-09-15 11:35 ` Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 2/9] migration: stop returning errno from load_snapshot() Daniel P. Berrangé
` (7 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
The bdrv_all_*_snapshot functions return a BlockDriverState pointer
for the invalid backend, which the callers then use to report an
error message. In some cases multiple callers are reporting the
same error message, but with slightly different text. In the future
there will be more error scenarios for some of these methods, which
will benefit from fine grained error message reporting. So it is
helpful to push error reporting down a level.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
block/monitor/block-hmp-cmds.c | 7 ++--
block/snapshot.c | 77 +++++++++++++++++-----------------
include/block/snapshot.h | 14 +++----
migration/savevm.c | 37 +++++-----------
monitor/hmp-cmds.c | 7 +---
tests/qemu-iotests/267.out | 10 ++---
6 files changed, 65 insertions(+), 87 deletions(-)
diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c
index 4d3db5ed3c..bec794253d 100644
--- a/block/monitor/block-hmp-cmds.c
+++ b/block/monitor/block-hmp-cmds.c
@@ -898,10 +898,11 @@ void hmp_info_snapshots(Monitor *mon, const QDict *qdict)
ImageEntry *image_entry, *next_ie;
SnapshotEntry *snapshot_entry;
+ Error *err = NULL;
- bs = bdrv_all_find_vmstate_bs();
+ bs = bdrv_all_find_vmstate_bs(&err);
if (!bs) {
- monitor_printf(mon, "No available block device supports snapshots\n");
+ error_report_err(err);
return;
}
aio_context = bdrv_get_aio_context(bs);
@@ -951,7 +952,7 @@ void hmp_info_snapshots(Monitor *mon, const QDict *qdict)
total = 0;
for (i = 0; i < nb_sns; i++) {
SnapshotEntry *next_sn;
- if (bdrv_all_find_snapshot(sn_tab[i].name, &bs1) == 0) {
+ if (bdrv_all_find_snapshot(sn_tab[i].name, NULL) == 0) {
global_snapshots[total] = i;
total++;
QTAILQ_FOREACH(image_entry, &image_list, next) {
diff --git a/block/snapshot.c b/block/snapshot.c
index a2bf3a54eb..50e35bb9fa 100644
--- a/block/snapshot.c
+++ b/block/snapshot.c
@@ -462,14 +462,14 @@ static bool bdrv_all_snapshots_includes_bs(BlockDriverState *bs)
* These functions will properly handle dataplane (take aio_context_acquire
* when appropriate for appropriate block drivers) */
-bool bdrv_all_can_snapshot(BlockDriverState **first_bad_bs)
+bool bdrv_all_can_snapshot(Error **errp)
{
- bool ok = true;
BlockDriverState *bs;
BdrvNextIterator it;
for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
AioContext *ctx = bdrv_get_aio_context(bs);
+ bool ok;
aio_context_acquire(ctx);
if (bdrv_all_snapshots_includes_bs(bs)) {
@@ -477,26 +477,25 @@ bool bdrv_all_can_snapshot(BlockDriverState **first_bad_bs)
}
aio_context_release(ctx);
if (!ok) {
+ error_setg(errp, "Device '%s' is writable but does not support "
+ "snapshots", bdrv_get_device_or_node_name(bs));
bdrv_next_cleanup(&it);
- goto fail;
+ return false;
}
}
-fail:
- *first_bad_bs = bs;
- return ok;
+ return true;
}
-int bdrv_all_delete_snapshot(const char *name, BlockDriverState **first_bad_bs,
- Error **errp)
+int bdrv_all_delete_snapshot(const char *name, Error **errp)
{
- int ret = 0;
BlockDriverState *bs;
BdrvNextIterator it;
QEMUSnapshotInfo sn1, *snapshot = &sn1;
for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
AioContext *ctx = bdrv_get_aio_context(bs);
+ int ret;
aio_context_acquire(ctx);
if (bdrv_all_snapshots_includes_bs(bs) &&
@@ -507,26 +506,25 @@ int bdrv_all_delete_snapshot(const char *name, BlockDriverState **first_bad_bs,
}
aio_context_release(ctx);
if (ret < 0) {
+ error_prepend(errp, "Could not delete snapshot '%s' on '%s': ",
+ name, bdrv_get_device_or_node_name(bs));
bdrv_next_cleanup(&it);
- goto fail;
+ return -1;
}
}
-fail:
- *first_bad_bs = bs;
- return ret;
+ return 0;
}
-int bdrv_all_goto_snapshot(const char *name, BlockDriverState **first_bad_bs,
- Error **errp)
+int bdrv_all_goto_snapshot(const char *name, Error **errp)
{
- int ret = 0;
BlockDriverState *bs;
BdrvNextIterator it;
for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
AioContext *ctx = bdrv_get_aio_context(bs);
+ int ret;
aio_context_acquire(ctx);
if (bdrv_all_snapshots_includes_bs(bs)) {
@@ -534,75 +532,75 @@ int bdrv_all_goto_snapshot(const char *name, BlockDriverState **first_bad_bs,
}
aio_context_release(ctx);
if (ret < 0) {
+ error_prepend(errp, "Could not load snapshot '%s' on '%s': ",
+ name, bdrv_get_device_or_node_name(bs));
bdrv_next_cleanup(&it);
- goto fail;
+ return -1;
}
}
-fail:
- *first_bad_bs = bs;
- return ret;
+ return 0;
}
-int bdrv_all_find_snapshot(const char *name, BlockDriverState **first_bad_bs)
+int bdrv_all_find_snapshot(const char *name, Error **errp)
{
QEMUSnapshotInfo sn;
- int err = 0;
BlockDriverState *bs;
BdrvNextIterator it;
for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
AioContext *ctx = bdrv_get_aio_context(bs);
+ int ret;
aio_context_acquire(ctx);
if (bdrv_all_snapshots_includes_bs(bs)) {
- err = bdrv_snapshot_find(bs, &sn, name);
+ ret = bdrv_snapshot_find(bs, &sn, name);
}
aio_context_release(ctx);
- if (err < 0) {
+ if (ret < 0) {
+ error_setg(errp, "Could not find snapshot '%s' on '%s'",
+ name, bdrv_get_device_or_node_name(bs));
bdrv_next_cleanup(&it);
- goto fail;
+ return -1;
}
}
-fail:
- *first_bad_bs = bs;
- return err;
+ return 0;
}
int bdrv_all_create_snapshot(QEMUSnapshotInfo *sn,
BlockDriverState *vm_state_bs,
uint64_t vm_state_size,
- BlockDriverState **first_bad_bs)
+ Error **errp)
{
- int err = 0;
BlockDriverState *bs;
BdrvNextIterator it;
for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
AioContext *ctx = bdrv_get_aio_context(bs);
+ int ret;
aio_context_acquire(ctx);
if (bs == vm_state_bs) {
sn->vm_state_size = vm_state_size;
- err = bdrv_snapshot_create(bs, sn);
+ ret = bdrv_snapshot_create(bs, sn);
} else if (bdrv_all_snapshots_includes_bs(bs)) {
sn->vm_state_size = 0;
- err = bdrv_snapshot_create(bs, sn);
+ ret = bdrv_snapshot_create(bs, sn);
}
aio_context_release(ctx);
- if (err < 0) {
+ if (ret < 0) {
+ error_setg(errp, "Could not create snapshot '%s' on '%s'",
+ sn->name, bdrv_get_device_or_node_name(bs));
bdrv_next_cleanup(&it);
- goto fail;
+ return -1;
}
}
-fail:
- *first_bad_bs = bs;
- return err;
+ return 0;
}
-BlockDriverState *bdrv_all_find_vmstate_bs(void)
+BlockDriverState *bdrv_all_find_vmstate_bs(Error **errp)
{
BlockDriverState *bs;
BdrvNextIterator it;
@@ -620,5 +618,8 @@ BlockDriverState *bdrv_all_find_vmstate_bs(void)
break;
}
}
+ if (!bs) {
+ error_setg(errp, "No block device supports snapshots");
+ }
return bs;
}
diff --git a/include/block/snapshot.h b/include/block/snapshot.h
index 2bfcd57578..ba1528eee0 100644
--- a/include/block/snapshot.h
+++ b/include/block/snapshot.h
@@ -76,17 +76,15 @@ int bdrv_snapshot_load_tmp_by_id_or_name(BlockDriverState *bs,
* These functions will properly handle dataplane (take aio_context_acquire
* when appropriate for appropriate block drivers */
-bool bdrv_all_can_snapshot(BlockDriverState **first_bad_bs);
-int bdrv_all_delete_snapshot(const char *name, BlockDriverState **first_bsd_bs,
- Error **errp);
-int bdrv_all_goto_snapshot(const char *name, BlockDriverState **first_bad_bs,
- Error **errp);
-int bdrv_all_find_snapshot(const char *name, BlockDriverState **first_bad_bs);
+bool bdrv_all_can_snapshot(Error **errp);
+int bdrv_all_delete_snapshot(const char *name, Error **errp);
+int bdrv_all_goto_snapshot(const char *name, Error **errp);
+int bdrv_all_find_snapshot(const char *name, Error **errp);
int bdrv_all_create_snapshot(QEMUSnapshotInfo *sn,
BlockDriverState *vm_state_bs,
uint64_t vm_state_size,
- BlockDriverState **first_bad_bs);
+ Error **errp);
-BlockDriverState *bdrv_all_find_vmstate_bs(void);
+BlockDriverState *bdrv_all_find_vmstate_bs(Error **errp);
#endif
diff --git a/migration/savevm.c b/migration/savevm.c
index 304d98ff78..3826c437cc 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -2660,7 +2660,7 @@ int qemu_load_device_state(QEMUFile *f)
int save_snapshot(const char *name, Error **errp)
{
- BlockDriverState *bs, *bs1;
+ BlockDriverState *bs;
QEMUSnapshotInfo sn1, *sn = &sn1, old_sn1, *old_sn = &old_sn1;
int ret = -1, ret2;
QEMUFile *f;
@@ -2680,25 +2680,19 @@ int save_snapshot(const char *name, Error **errp)
return ret;
}
- if (!bdrv_all_can_snapshot(&bs)) {
- error_setg(errp, "Device '%s' is writable but does not support "
- "snapshots", bdrv_get_device_or_node_name(bs));
+ if (!bdrv_all_can_snapshot(errp)) {
return ret;
}
/* Delete old snapshots of the same name */
if (name) {
- ret = bdrv_all_delete_snapshot(name, &bs1, errp);
- if (ret < 0) {
- error_prepend(errp, "Error while deleting snapshot on device "
- "'%s': ", bdrv_get_device_or_node_name(bs1));
+ if (bdrv_all_delete_snapshot(name, errp) < 0) {
return ret;
}
}
- bs = bdrv_all_find_vmstate_bs();
+ bs = bdrv_all_find_vmstate_bs(errp);
if (bs == NULL) {
- error_setg(errp, "No block device can accept snapshots");
return ret;
}
aio_context = bdrv_get_aio_context(bs);
@@ -2763,10 +2757,8 @@ int save_snapshot(const char *name, Error **errp)
aio_context_release(aio_context);
aio_context = NULL;
- ret = bdrv_all_create_snapshot(sn, bs, vm_state_size, &bs);
+ ret = bdrv_all_create_snapshot(sn, bs, vm_state_size, errp);
if (ret < 0) {
- error_setg(errp, "Error while creating snapshot on '%s'",
- bdrv_get_device_or_node_name(bs));
goto the_end;
}
@@ -2868,7 +2860,7 @@ void qmp_xen_load_devices_state(const char *filename, Error **errp)
int load_snapshot(const char *name, Error **errp)
{
- BlockDriverState *bs, *bs_vm_state;
+ BlockDriverState *bs_vm_state;
QEMUSnapshotInfo sn;
QEMUFile *f;
int ret;
@@ -2881,23 +2873,16 @@ int load_snapshot(const char *name, Error **errp)
return -EINVAL;
}
- if (!bdrv_all_can_snapshot(&bs)) {
- error_setg(errp,
- "Device '%s' is writable but does not support snapshots",
- bdrv_get_device_or_node_name(bs));
+ if (!bdrv_all_can_snapshot(errp)) {
return -ENOTSUP;
}
- ret = bdrv_all_find_snapshot(name, &bs);
+ ret = bdrv_all_find_snapshot(name, errp);
if (ret < 0) {
- error_setg(errp,
- "Device '%s' does not have the requested snapshot '%s'",
- bdrv_get_device_or_node_name(bs), name);
return ret;
}
- bs_vm_state = bdrv_all_find_vmstate_bs();
+ bs_vm_state = bdrv_all_find_vmstate_bs(errp);
if (!bs_vm_state) {
- error_setg(errp, "No block device supports snapshots");
return -ENOTSUP;
}
aio_context = bdrv_get_aio_context(bs_vm_state);
@@ -2917,10 +2902,8 @@ int load_snapshot(const char *name, Error **errp)
/* Flush all IO requests so they don't interfere with the new state. */
bdrv_drain_all_begin();
- ret = bdrv_all_goto_snapshot(name, &bs, errp);
+ ret = bdrv_all_goto_snapshot(name, errp);
if (ret < 0) {
- error_prepend(errp, "Could not load snapshot '%s' on '%s': ",
- name, bdrv_get_device_or_node_name(bs));
goto err_drain;
}
diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
index 7711726fd2..9bb50b9abf 100644
--- a/monitor/hmp-cmds.c
+++ b/monitor/hmp-cmds.c
@@ -1137,15 +1137,10 @@ void hmp_savevm(Monitor *mon, const QDict *qdict)
void hmp_delvm(Monitor *mon, const QDict *qdict)
{
- BlockDriverState *bs;
Error *err = NULL;
const char *name = qdict_get_str(qdict, "name");
- if (bdrv_all_delete_snapshot(name, &bs, &err) < 0) {
- error_prepend(&err,
- "deleting snapshot on device '%s': ",
- bdrv_get_device_name(bs));
- }
+ bdrv_all_delete_snapshot(name, &err);
hmp_handle_error(mon, err);
}
diff --git a/tests/qemu-iotests/267.out b/tests/qemu-iotests/267.out
index 215902b3ad..c65cce893a 100644
--- a/tests/qemu-iotests/267.out
+++ b/tests/qemu-iotests/267.out
@@ -6,9 +6,9 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=134217728
Testing:
QEMU X.Y.Z monitor - type 'help' for more information
(qemu) savevm snap0
-Error: No block device can accept snapshots
+Error: No block device supports snapshots
(qemu) info snapshots
-No available block device supports snapshots
+No block device supports snapshots
(qemu) loadvm snap0
Error: No block device supports snapshots
(qemu) quit
@@ -22,7 +22,7 @@ QEMU X.Y.Z monitor - type 'help' for more information
(qemu) savevm snap0
Error: Device 'none0' is writable but does not support snapshots
(qemu) info snapshots
-No available block device supports snapshots
+No block device supports snapshots
(qemu) loadvm snap0
Error: Device 'none0' is writable but does not support snapshots
(qemu) quit
@@ -58,7 +58,7 @@ QEMU X.Y.Z monitor - type 'help' for more information
(qemu) savevm snap0
Error: Device 'virtio0' is writable but does not support snapshots
(qemu) info snapshots
-No available block device supports snapshots
+No block device supports snapshots
(qemu) loadvm snap0
Error: Device 'virtio0' is writable but does not support snapshots
(qemu) quit
@@ -83,7 +83,7 @@ QEMU X.Y.Z monitor - type 'help' for more information
(qemu) savevm snap0
Error: Device 'file' is writable but does not support snapshots
(qemu) info snapshots
-No available block device supports snapshots
+No block device supports snapshots
(qemu) loadvm snap0
Error: Device 'file' is writable but does not support snapshots
(qemu) quit
--
2.26.2
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v4 2/9] migration: stop returning errno from load_snapshot()
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 1/9] block: push error reporting into bdrv_all_*_snapshot functions Daniel P. Berrangé
@ 2020-09-15 11:35 ` Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 3/9] block: add ability to specify list of blockdevs during snapshot Daniel P. Berrangé
` (6 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
None of the callers care about the errno value since there is a full
Error object populated. This gives consistency with save_snapshot()
which already just returns -1.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
migration/savevm.c | 17 ++++++++---------
1 file changed, 8 insertions(+), 9 deletions(-)
diff --git a/migration/savevm.c b/migration/savevm.c
index 3826c437cc..711137bcbe 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -2870,20 +2870,20 @@ int load_snapshot(const char *name, Error **errp)
if (!replay_can_snapshot()) {
error_setg(errp, "Record/replay does not allow loading snapshot "
"right now. Try once more later.");
- return -EINVAL;
+ return -1;
}
if (!bdrv_all_can_snapshot(errp)) {
- return -ENOTSUP;
+ return -1;
}
ret = bdrv_all_find_snapshot(name, errp);
if (ret < 0) {
- return ret;
+ return -1;
}
bs_vm_state = bdrv_all_find_vmstate_bs(errp);
if (!bs_vm_state) {
- return -ENOTSUP;
+ return -1;
}
aio_context = bdrv_get_aio_context(bs_vm_state);
@@ -2892,11 +2892,11 @@ int load_snapshot(const char *name, Error **errp)
ret = bdrv_snapshot_find(bs_vm_state, &sn, name);
aio_context_release(aio_context);
if (ret < 0) {
- return ret;
+ return -1;
} else if (sn.vm_state_size == 0) {
error_setg(errp, "This is a disk-only snapshot. Revert to it "
" offline using qemu-img");
- return -EINVAL;
+ return -1;
}
/* Flush all IO requests so they don't interfere with the new state. */
@@ -2911,7 +2911,6 @@ int load_snapshot(const char *name, Error **errp)
f = qemu_fopen_bdrv(bs_vm_state, 0);
if (!f) {
error_setg(errp, "Could not open VM state file");
- ret = -EINVAL;
goto err_drain;
}
@@ -2927,14 +2926,14 @@ int load_snapshot(const char *name, Error **errp)
if (ret < 0) {
error_setg(errp, "Error %d while loading VM state", ret);
- return ret;
+ return -1;
}
return 0;
err_drain:
bdrv_drain_all_end();
- return ret;
+ return -1;
}
void vmstate_register_ram(MemoryRegion *mr, DeviceState *dev)
--
2.26.2
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v4 3/9] block: add ability to specify list of blockdevs during snapshot
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 1/9] block: push error reporting into bdrv_all_*_snapshot functions Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 2/9] migration: stop returning errno from load_snapshot() Daniel P. Berrangé
@ 2020-09-15 11:35 ` Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 4/9] block: allow specifying name of block device for vmstate storage Daniel P. Berrangé
` (5 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
When running snapshot operations, there are various rules for which
blockdevs are included/excluded. While this provides reasonable default
behaviour, there are scenarios that are not well handled by the default
logic. Some of the conditions do not have a single correct answer.
Thus there needs to be a way for the mgmt app to provide an explicit
list of blockdevs to perform snapshots across. This can be achieved
by passing a list of node names that should be used.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
block/monitor/block-hmp-cmds.c | 4 +-
block/snapshot.c | 180 ++++++++++++++++++++++++---------
include/block/snapshot.h | 22 ++--
migration/savevm.c | 16 +--
monitor/hmp-cmds.c | 2 +-
5 files changed, 160 insertions(+), 64 deletions(-)
diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c
index bec794253d..30279bd15a 100644
--- a/block/monitor/block-hmp-cmds.c
+++ b/block/monitor/block-hmp-cmds.c
@@ -900,7 +900,7 @@ void hmp_info_snapshots(Monitor *mon, const QDict *qdict)
SnapshotEntry *snapshot_entry;
Error *err = NULL;
- bs = bdrv_all_find_vmstate_bs(&err);
+ bs = bdrv_all_find_vmstate_bs(false, NULL, &err);
if (!bs) {
error_report_err(err);
return;
@@ -952,7 +952,7 @@ void hmp_info_snapshots(Monitor *mon, const QDict *qdict)
total = 0;
for (i = 0; i < nb_sns; i++) {
SnapshotEntry *next_sn;
- if (bdrv_all_find_snapshot(sn_tab[i].name, NULL) == 0) {
+ if (bdrv_all_find_snapshot(sn_tab[i].name, false, NULL, NULL) == 0) {
global_snapshots[total] = i;
total++;
QTAILQ_FOREACH(image_entry, &image_list, next) {
diff --git a/block/snapshot.c b/block/snapshot.c
index 50e35bb9fa..155b8aad88 100644
--- a/block/snapshot.c
+++ b/block/snapshot.c
@@ -447,6 +447,41 @@ int bdrv_snapshot_load_tmp_by_id_or_name(BlockDriverState *bs,
return ret;
}
+
+static int bdrv_all_get_snapshot_devices(bool has_devices, strList *devices,
+ GList **all_bdrvs,
+ Error **errp)
+{
+ g_autoptr(GList) bdrvs = NULL;
+
+ if (has_devices) {
+ if (!devices) {
+ error_setg(errp, "At least one device is required for snapshot");
+ return -1;
+ }
+
+ while (devices) {
+ BlockDriverState *bs = bdrv_find_node(devices->value);
+ if (!bs) {
+ error_setg(errp, "No block device node '%s'", devices->value);
+ return -1;
+ }
+ bdrvs = g_list_append(bdrvs, bs);
+ devices = devices->next;
+ }
+ } else {
+ BlockDriverState *bs;
+ BdrvNextIterator it;
+ for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
+ bdrvs = g_list_append(bdrvs, bs);
+ }
+ }
+
+ *all_bdrvs = g_steal_pointer(&bdrvs);
+ return 0;
+}
+
+
static bool bdrv_all_snapshots_includes_bs(BlockDriverState *bs)
{
if (!bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) {
@@ -462,43 +497,59 @@ static bool bdrv_all_snapshots_includes_bs(BlockDriverState *bs)
* These functions will properly handle dataplane (take aio_context_acquire
* when appropriate for appropriate block drivers) */
-bool bdrv_all_can_snapshot(Error **errp)
+bool bdrv_all_can_snapshot(bool has_devices, strList *devices,
+ Error **errp)
{
- BlockDriverState *bs;
- BdrvNextIterator it;
+ g_autoptr(GList) bdrvs = NULL;
+ GList *iterbdrvs;
- for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
+ if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) {
+ return false;
+ }
+
+ iterbdrvs = bdrvs;
+ while (iterbdrvs) {
+ BlockDriverState *bs = iterbdrvs->data;
AioContext *ctx = bdrv_get_aio_context(bs);
bool ok;
aio_context_acquire(ctx);
- if (bdrv_all_snapshots_includes_bs(bs)) {
+ if (devices || bdrv_all_snapshots_includes_bs(bs)) {
ok = bdrv_can_snapshot(bs);
}
aio_context_release(ctx);
if (!ok) {
error_setg(errp, "Device '%s' is writable but does not support "
"snapshots", bdrv_get_device_or_node_name(bs));
- bdrv_next_cleanup(&it);
return false;
}
+
+ iterbdrvs = iterbdrvs->next;
}
return true;
}
-int bdrv_all_delete_snapshot(const char *name, Error **errp)
+int bdrv_all_delete_snapshot(const char *name,
+ bool has_devices, strList *devices,
+ Error **errp)
{
- BlockDriverState *bs;
- BdrvNextIterator it;
- QEMUSnapshotInfo sn1, *snapshot = &sn1;
+ g_autoptr(GList) bdrvs = NULL;
+ GList *iterbdrvs;
+
+ if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) {
+ return -1;
+ }
- for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
+ iterbdrvs = bdrvs;
+ while (iterbdrvs) {
+ BlockDriverState *bs = iterbdrvs->data;
AioContext *ctx = bdrv_get_aio_context(bs);
- int ret;
+ QEMUSnapshotInfo sn1, *snapshot = &sn1;
+ int ret = 0;
aio_context_acquire(ctx);
- if (bdrv_all_snapshots_includes_bs(bs) &&
+ if ((devices || bdrv_all_snapshots_includes_bs(bs)) &&
bdrv_snapshot_find(bs, snapshot, name) >= 0)
{
ret = bdrv_snapshot_delete(bs, snapshot->id_str,
@@ -508,61 +559,80 @@ int bdrv_all_delete_snapshot(const char *name, Error **errp)
if (ret < 0) {
error_prepend(errp, "Could not delete snapshot '%s' on '%s': ",
name, bdrv_get_device_or_node_name(bs));
- bdrv_next_cleanup(&it);
return -1;
}
+
+ iterbdrvs = iterbdrvs->next;
}
return 0;
}
-int bdrv_all_goto_snapshot(const char *name, Error **errp)
+int bdrv_all_goto_snapshot(const char *name,
+ bool has_devices, strList *devices,
+ Error **errp)
{
- BlockDriverState *bs;
- BdrvNextIterator it;
+ g_autoptr(GList) bdrvs = NULL;
+ GList *iterbdrvs;
- for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
+ if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) {
+ return -1;
+ }
+
+ iterbdrvs = bdrvs;
+ while (iterbdrvs) {
+ BlockDriverState *bs = iterbdrvs->data;
AioContext *ctx = bdrv_get_aio_context(bs);
- int ret;
+ int ret = 0;
aio_context_acquire(ctx);
- if (bdrv_all_snapshots_includes_bs(bs)) {
+ if (devices || bdrv_all_snapshots_includes_bs(bs)) {
ret = bdrv_snapshot_goto(bs, name, errp);
}
aio_context_release(ctx);
if (ret < 0) {
error_prepend(errp, "Could not load snapshot '%s' on '%s': ",
name, bdrv_get_device_or_node_name(bs));
- bdrv_next_cleanup(&it);
return -1;
}
+
+ iterbdrvs = iterbdrvs->next;
}
return 0;
}
-int bdrv_all_find_snapshot(const char *name, Error **errp)
+int bdrv_all_find_snapshot(const char *name,
+ bool has_devices, strList *devices,
+ Error **errp)
{
- QEMUSnapshotInfo sn;
- BlockDriverState *bs;
- BdrvNextIterator it;
+ g_autoptr(GList) bdrvs = NULL;
+ GList *iterbdrvs;
+
+ if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) {
+ return -1;
+ }
- for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
+ iterbdrvs = bdrvs;
+ while (iterbdrvs) {
+ BlockDriverState *bs = iterbdrvs->data;
AioContext *ctx = bdrv_get_aio_context(bs);
- int ret;
+ QEMUSnapshotInfo sn;
+ int ret = 0;
aio_context_acquire(ctx);
- if (bdrv_all_snapshots_includes_bs(bs)) {
+ if (devices || bdrv_all_snapshots_includes_bs(bs)) {
ret = bdrv_snapshot_find(bs, &sn, name);
}
aio_context_release(ctx);
if (ret < 0) {
error_setg(errp, "Could not find snapshot '%s' on '%s'",
name, bdrv_get_device_or_node_name(bs));
- bdrv_next_cleanup(&it);
return -1;
}
+
+ iterbdrvs = iterbdrvs->next;
}
return 0;
@@ -571,20 +641,27 @@ int bdrv_all_find_snapshot(const char *name, Error **errp)
int bdrv_all_create_snapshot(QEMUSnapshotInfo *sn,
BlockDriverState *vm_state_bs,
uint64_t vm_state_size,
+ bool has_devices, strList *devices,
Error **errp)
{
- BlockDriverState *bs;
- BdrvNextIterator it;
+ g_autoptr(GList) bdrvs = NULL;
+ GList *iterbdrvs;
+
+ if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) {
+ return -1;
+ }
- for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
+ iterbdrvs = bdrvs;
+ while (iterbdrvs) {
+ BlockDriverState *bs = iterbdrvs->data;
AioContext *ctx = bdrv_get_aio_context(bs);
- int ret;
+ int ret = 0;
aio_context_acquire(ctx);
if (bs == vm_state_bs) {
sn->vm_state_size = vm_state_size;
ret = bdrv_snapshot_create(bs, sn);
- } else if (bdrv_all_snapshots_includes_bs(bs)) {
+ } else if (devices || bdrv_all_snapshots_includes_bs(bs)) {
sn->vm_state_size = 0;
ret = bdrv_snapshot_create(bs, sn);
}
@@ -592,34 +669,43 @@ int bdrv_all_create_snapshot(QEMUSnapshotInfo *sn,
if (ret < 0) {
error_setg(errp, "Could not create snapshot '%s' on '%s'",
sn->name, bdrv_get_device_or_node_name(bs));
- bdrv_next_cleanup(&it);
return -1;
}
+
+ iterbdrvs = iterbdrvs->next;
}
return 0;
}
-BlockDriverState *bdrv_all_find_vmstate_bs(Error **errp)
+BlockDriverState *bdrv_all_find_vmstate_bs(bool has_devices, strList *devices,
+ Error **errp)
{
- BlockDriverState *bs;
- BdrvNextIterator it;
+ g_autoptr(GList) bdrvs = NULL;
+ GList *iterbdrvs;
- for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
+ if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) {
+ return NULL;
+ }
+
+ iterbdrvs = bdrvs;
+ while (iterbdrvs) {
+ BlockDriverState *bs = iterbdrvs->data;
AioContext *ctx = bdrv_get_aio_context(bs);
- bool found;
+ bool found = false;
aio_context_acquire(ctx);
- found = bdrv_all_snapshots_includes_bs(bs) && bdrv_can_snapshot(bs);
+ found = (devices || bdrv_all_snapshots_includes_bs(bs)) &&
+ bdrv_can_snapshot(bs);
aio_context_release(ctx);
if (found) {
- bdrv_next_cleanup(&it);
- break;
+ return bs;
}
+
+ iterbdrvs = iterbdrvs->next;
}
- if (!bs) {
- error_setg(errp, "No block device supports snapshots");
- }
- return bs;
+
+ error_setg(errp, "No block device supports snapshots");
+ return NULL;
}
diff --git a/include/block/snapshot.h b/include/block/snapshot.h
index ba1528eee0..bac3409669 100644
--- a/include/block/snapshot.h
+++ b/include/block/snapshot.h
@@ -25,7 +25,7 @@
#ifndef SNAPSHOT_H
#define SNAPSHOT_H
-
+#include "qapi/qapi-builtin-types.h"
#define SNAPSHOT_OPT_BASE "snapshot."
#define SNAPSHOT_OPT_ID "snapshot.id"
@@ -76,15 +76,25 @@ int bdrv_snapshot_load_tmp_by_id_or_name(BlockDriverState *bs,
* These functions will properly handle dataplane (take aio_context_acquire
* when appropriate for appropriate block drivers */
-bool bdrv_all_can_snapshot(Error **errp);
-int bdrv_all_delete_snapshot(const char *name, Error **errp);
-int bdrv_all_goto_snapshot(const char *name, Error **errp);
-int bdrv_all_find_snapshot(const char *name, Error **errp);
+bool bdrv_all_can_snapshot(bool has_devices, strList *devices,
+ Error **errp);
+int bdrv_all_delete_snapshot(const char *name,
+ bool has_devices, strList *devices,
+ Error **errp);
+int bdrv_all_goto_snapshot(const char *name,
+ bool has_devices, strList *devices,
+ Error **errp);
+int bdrv_all_find_snapshot(const char *name,
+ bool has_devices, strList *devices,
+ Error **errp);
int bdrv_all_create_snapshot(QEMUSnapshotInfo *sn,
BlockDriverState *vm_state_bs,
uint64_t vm_state_size,
+ bool has_devices,
+ strList *devices,
Error **errp);
-BlockDriverState *bdrv_all_find_vmstate_bs(Error **errp);
+BlockDriverState *bdrv_all_find_vmstate_bs(bool has_devices, strList *devices,
+ Error **errp);
#endif
diff --git a/migration/savevm.c b/migration/savevm.c
index 711137bcbe..0b3454f2e8 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -2680,18 +2680,18 @@ int save_snapshot(const char *name, Error **errp)
return ret;
}
- if (!bdrv_all_can_snapshot(errp)) {
+ if (!bdrv_all_can_snapshot(false, NULL, errp)) {
return ret;
}
/* Delete old snapshots of the same name */
if (name) {
- if (bdrv_all_delete_snapshot(name, errp) < 0) {
+ if (bdrv_all_delete_snapshot(name, false, NULL, errp) < 0) {
return ret;
}
}
- bs = bdrv_all_find_vmstate_bs(errp);
+ bs = bdrv_all_find_vmstate_bs(false ,NULL, errp);
if (bs == NULL) {
return ret;
}
@@ -2757,7 +2757,7 @@ int save_snapshot(const char *name, Error **errp)
aio_context_release(aio_context);
aio_context = NULL;
- ret = bdrv_all_create_snapshot(sn, bs, vm_state_size, errp);
+ ret = bdrv_all_create_snapshot(sn, bs, vm_state_size, false, NULL, errp);
if (ret < 0) {
goto the_end;
}
@@ -2873,15 +2873,15 @@ int load_snapshot(const char *name, Error **errp)
return -1;
}
- if (!bdrv_all_can_snapshot(errp)) {
+ if (!bdrv_all_can_snapshot(false, NULL, errp)) {
return -1;
}
- ret = bdrv_all_find_snapshot(name, errp);
+ ret = bdrv_all_find_snapshot(name, false, NULL, errp);
if (ret < 0) {
return -1;
}
- bs_vm_state = bdrv_all_find_vmstate_bs(errp);
+ bs_vm_state = bdrv_all_find_vmstate_bs(false, NULL, errp);
if (!bs_vm_state) {
return -1;
}
@@ -2902,7 +2902,7 @@ int load_snapshot(const char *name, Error **errp)
/* Flush all IO requests so they don't interfere with the new state. */
bdrv_drain_all_begin();
- ret = bdrv_all_goto_snapshot(name, errp);
+ ret = bdrv_all_goto_snapshot(name, false, NULL, errp);
if (ret < 0) {
goto err_drain;
}
diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
index 9bb50b9abf..98c98ae182 100644
--- a/monitor/hmp-cmds.c
+++ b/monitor/hmp-cmds.c
@@ -1140,7 +1140,7 @@ void hmp_delvm(Monitor *mon, const QDict *qdict)
Error *err = NULL;
const char *name = qdict_get_str(qdict, "name");
- bdrv_all_delete_snapshot(name, &err);
+ bdrv_all_delete_snapshot(name, false, NULL, &err);
hmp_handle_error(mon, err);
}
--
2.26.2
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v4 4/9] block: allow specifying name of block device for vmstate storage
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
` (2 preceding siblings ...)
2020-09-15 11:35 ` [PATCH v4 3/9] block: add ability to specify list of blockdevs during snapshot Daniel P. Berrangé
@ 2020-09-15 11:35 ` Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 5/9] migration: control whether snapshots are ovewritten Daniel P. Berrangé
` (4 subsequent siblings)
8 siblings, 0 replies; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
Currently the vmstate will be stored in the first block device that
supports snapshots. Historically this would have usually been the
root device, but with UEFI it might be the variable store. There
needs to be a way to override the choice of block device to store
the state in.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
block/monitor/block-hmp-cmds.c | 2 +-
block/snapshot.c | 23 ++++++++++++++++++++---
include/block/snapshot.h | 3 ++-
migration/savevm.c | 4 ++--
4 files changed, 25 insertions(+), 7 deletions(-)
diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c
index 30279bd15a..4a1fc1d6b0 100644
--- a/block/monitor/block-hmp-cmds.c
+++ b/block/monitor/block-hmp-cmds.c
@@ -900,7 +900,7 @@ void hmp_info_snapshots(Monitor *mon, const QDict *qdict)
SnapshotEntry *snapshot_entry;
Error *err = NULL;
- bs = bdrv_all_find_vmstate_bs(false, NULL, &err);
+ bs = bdrv_all_find_vmstate_bs(NULL, false, NULL, &err);
if (!bs) {
error_report_err(err);
return;
diff --git a/block/snapshot.c b/block/snapshot.c
index 155b8aad88..d2caaa88b5 100644
--- a/block/snapshot.c
+++ b/block/snapshot.c
@@ -678,7 +678,9 @@ int bdrv_all_create_snapshot(QEMUSnapshotInfo *sn,
return 0;
}
-BlockDriverState *bdrv_all_find_vmstate_bs(bool has_devices, strList *devices,
+
+BlockDriverState *bdrv_all_find_vmstate_bs(const char *vmstate_bs,
+ bool has_devices, strList *devices,
Error **errp)
{
g_autoptr(GList) bdrvs = NULL;
@@ -699,13 +701,28 @@ BlockDriverState *bdrv_all_find_vmstate_bs(bool has_devices, strList *devices,
bdrv_can_snapshot(bs);
aio_context_release(ctx);
- if (found) {
+ if (vmstate_bs) {
+ if (g_str_equal(vmstate_bs,
+ bdrv_get_node_name(bs))) {
+ if (found) {
+ return bs;
+ } else {
+ error_setg(errp, "vmstate block device '%s' does not support snapshots",
+ vmstate_bs);
+ return NULL;
+ }
+ }
+ } else if (found) {
return bs;
}
iterbdrvs = iterbdrvs->next;
}
- error_setg(errp, "No block device supports snapshots");
+ if (vmstate_bs) {
+ error_setg(errp, "vmstate block device '%s' does not exist", vmstate_bs);
+ } else {
+ error_setg(errp, "no block device can store vmstate for snapshot");
+ }
return NULL;
}
diff --git a/include/block/snapshot.h b/include/block/snapshot.h
index bac3409669..4d25f17728 100644
--- a/include/block/snapshot.h
+++ b/include/block/snapshot.h
@@ -94,7 +94,8 @@ int bdrv_all_create_snapshot(QEMUSnapshotInfo *sn,
strList *devices,
Error **errp);
-BlockDriverState *bdrv_all_find_vmstate_bs(bool has_devices, strList *devices,
+BlockDriverState *bdrv_all_find_vmstate_bs(const char *vmstate_bs,
+ bool has_devices, strList *devices,
Error **errp);
#endif
diff --git a/migration/savevm.c b/migration/savevm.c
index 0b3454f2e8..76972d69b0 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -2691,7 +2691,7 @@ int save_snapshot(const char *name, Error **errp)
}
}
- bs = bdrv_all_find_vmstate_bs(false ,NULL, errp);
+ bs = bdrv_all_find_vmstate_bs(NULL, false, NULL, errp);
if (bs == NULL) {
return ret;
}
@@ -2881,7 +2881,7 @@ int load_snapshot(const char *name, Error **errp)
return -1;
}
- bs_vm_state = bdrv_all_find_vmstate_bs(false, NULL, errp);
+ bs_vm_state = bdrv_all_find_vmstate_bs(NULL, false, NULL, errp);
if (!bs_vm_state) {
return -1;
}
--
2.26.2
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v4 5/9] migration: control whether snapshots are ovewritten
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
` (3 preceding siblings ...)
2020-09-15 11:35 ` [PATCH v4 4/9] block: allow specifying name of block device for vmstate storage Daniel P. Berrangé
@ 2020-09-15 11:35 ` Daniel P. Berrangé
2020-09-16 7:47 ` Markus Armbruster
2020-09-15 11:35 ` [PATCH v4 6/9] migration: wire up support for snapshot device selection Daniel P. Berrangé
` (3 subsequent siblings)
8 siblings, 1 reply; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
The traditional HMP "savevm" command will overwrite an existing snapshot
if it already exists with the requested name. This new flag allows this
to be controlled allowing for safer behaviour with a future QMP command.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
include/migration/snapshot.h | 2 +-
migration/savevm.c | 4 ++--
monitor/hmp-cmds.c | 2 +-
replay/replay-snapshot.c | 2 +-
4 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/include/migration/snapshot.h b/include/migration/snapshot.h
index c85b6ec75b..d7db1174ef 100644
--- a/include/migration/snapshot.h
+++ b/include/migration/snapshot.h
@@ -15,7 +15,7 @@
#ifndef QEMU_MIGRATION_SNAPSHOT_H
#define QEMU_MIGRATION_SNAPSHOT_H
-int save_snapshot(const char *name, Error **errp);
+int save_snapshot(const char *name, bool overwrite, Error **errp);
int load_snapshot(const char *name, Error **errp);
#endif
diff --git a/migration/savevm.c b/migration/savevm.c
index 76972d69b0..2025e3e579 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -2658,7 +2658,7 @@ int qemu_load_device_state(QEMUFile *f)
return 0;
}
-int save_snapshot(const char *name, Error **errp)
+int save_snapshot(const char *name, bool overwrite, Error **errp)
{
BlockDriverState *bs;
QEMUSnapshotInfo sn1, *sn = &sn1, old_sn1, *old_sn = &old_sn1;
@@ -2685,7 +2685,7 @@ int save_snapshot(const char *name, Error **errp)
}
/* Delete old snapshots of the same name */
- if (name) {
+ if (name && overwrite) {
if (bdrv_all_delete_snapshot(name, false, NULL, errp) < 0) {
return ret;
}
diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
index 98c98ae182..c1b8533d0c 100644
--- a/monitor/hmp-cmds.c
+++ b/monitor/hmp-cmds.c
@@ -1131,7 +1131,7 @@ void hmp_savevm(Monitor *mon, const QDict *qdict)
{
Error *err = NULL;
- save_snapshot(qdict_get_try_str(qdict, "name"), &err);
+ save_snapshot(qdict_get_try_str(qdict, "name"), true, &err);
hmp_handle_error(mon, err);
}
diff --git a/replay/replay-snapshot.c b/replay/replay-snapshot.c
index e26fa4c892..8e7ff97d11 100644
--- a/replay/replay-snapshot.c
+++ b/replay/replay-snapshot.c
@@ -77,7 +77,7 @@ void replay_vmstate_init(void)
if (replay_snapshot) {
if (replay_mode == REPLAY_MODE_RECORD) {
- if (save_snapshot(replay_snapshot, &err) != 0) {
+ if (save_snapshot(replay_snapshot, true, &err) != 0) {
error_report_err(err);
error_report("Could not create snapshot for icount record");
exit(1);
--
2.26.2
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v4 6/9] migration: wire up support for snapshot device selection
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
` (4 preceding siblings ...)
2020-09-15 11:35 ` [PATCH v4 5/9] migration: control whether snapshots are ovewritten Daniel P. Berrangé
@ 2020-09-15 11:35 ` Daniel P. Berrangé
2020-09-21 17:41 ` Dr. David Alan Gilbert
2020-09-15 11:35 ` [PATCH v4 7/9] migration: introduce a delete_snapshot wrapper Daniel P. Berrangé
` (2 subsequent siblings)
8 siblings, 1 reply; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
Modify load_snapshot/save_snapshot to accept the device list and vmstate
node name parameters previously added to the block layer.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
include/migration/snapshot.h | 12 ++++++++++--
migration/savevm.c | 24 ++++++++++++++----------
monitor/hmp-cmds.c | 4 ++--
replay/replay-snapshot.c | 4 ++--
softmmu/vl.c | 2 +-
5 files changed, 29 insertions(+), 17 deletions(-)
diff --git a/include/migration/snapshot.h b/include/migration/snapshot.h
index d7db1174ef..b2c72e0a1b 100644
--- a/include/migration/snapshot.h
+++ b/include/migration/snapshot.h
@@ -15,7 +15,15 @@
#ifndef QEMU_MIGRATION_SNAPSHOT_H
#define QEMU_MIGRATION_SNAPSHOT_H
-int save_snapshot(const char *name, bool overwrite, Error **errp);
-int load_snapshot(const char *name, Error **errp);
+#include "qapi/qapi-builtin-types.h"
+
+int save_snapshot(const char *name, bool overwrite,
+ const char *vmstate,
+ bool has_devices, strList *devices,
+ Error **errp);
+int load_snapshot(const char *name,
+ const char *vmstate,
+ bool has_devices, strList *devices,
+ Error **errp);
#endif
diff --git a/migration/savevm.c b/migration/savevm.c
index 2025e3e579..b3d2ce7045 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -43,6 +43,8 @@
#include "qapi/error.h"
#include "qapi/qapi-commands-migration.h"
#include "qapi/qapi-commands-misc.h"
+#include "qapi/clone-visitor.h"
+#include "qapi/qapi-builtin-visit.h"
#include "qapi/qmp/qerror.h"
#include "qemu/error-report.h"
#include "sysemu/cpus.h"
@@ -2658,7 +2660,8 @@ int qemu_load_device_state(QEMUFile *f)
return 0;
}
-int save_snapshot(const char *name, bool overwrite, Error **errp)
+int save_snapshot(const char *name, bool overwrite, const char *vmstate,
+ bool has_devices, strList *devices, Error **errp)
{
BlockDriverState *bs;
QEMUSnapshotInfo sn1, *sn = &sn1, old_sn1, *old_sn = &old_sn1;
@@ -2680,18 +2683,18 @@ int save_snapshot(const char *name, bool overwrite, Error **errp)
return ret;
}
- if (!bdrv_all_can_snapshot(false, NULL, errp)) {
+ if (!bdrv_all_can_snapshot(has_devices, devices, errp)) {
return ret;
}
/* Delete old snapshots of the same name */
if (name && overwrite) {
- if (bdrv_all_delete_snapshot(name, false, NULL, errp) < 0) {
+ if (bdrv_all_delete_snapshot(name, has_devices, devices, errp) < 0) {
return ret;
}
}
- bs = bdrv_all_find_vmstate_bs(NULL, false, NULL, errp);
+ bs = bdrv_all_find_vmstate_bs(vmstate, has_devices, devices, errp);
if (bs == NULL) {
return ret;
}
@@ -2757,7 +2760,7 @@ int save_snapshot(const char *name, bool overwrite, Error **errp)
aio_context_release(aio_context);
aio_context = NULL;
- ret = bdrv_all_create_snapshot(sn, bs, vm_state_size, false, NULL, errp);
+ ret = bdrv_all_create_snapshot(sn, bs, vm_state_size, has_devices, devices, errp);
if (ret < 0) {
goto the_end;
}
@@ -2858,7 +2861,8 @@ void qmp_xen_load_devices_state(const char *filename, Error **errp)
migration_incoming_state_destroy();
}
-int load_snapshot(const char *name, Error **errp)
+int load_snapshot(const char *name, const char *vmstate,
+ bool has_devices, strList *devices, Error **errp)
{
BlockDriverState *bs_vm_state;
QEMUSnapshotInfo sn;
@@ -2873,15 +2877,15 @@ int load_snapshot(const char *name, Error **errp)
return -1;
}
- if (!bdrv_all_can_snapshot(false, NULL, errp)) {
+ if (!bdrv_all_can_snapshot(has_devices, devices, errp)) {
return -1;
}
- ret = bdrv_all_find_snapshot(name, false, NULL, errp);
+ ret = bdrv_all_find_snapshot(name, has_devices, devices, errp);
if (ret < 0) {
return -1;
}
- bs_vm_state = bdrv_all_find_vmstate_bs(NULL, false, NULL, errp);
+ bs_vm_state = bdrv_all_find_vmstate_bs(vmstate, has_devices, devices, errp);
if (!bs_vm_state) {
return -1;
}
@@ -2902,7 +2906,7 @@ int load_snapshot(const char *name, Error **errp)
/* Flush all IO requests so they don't interfere with the new state. */
bdrv_drain_all_begin();
- ret = bdrv_all_goto_snapshot(name, false, NULL, errp);
+ ret = bdrv_all_goto_snapshot(name, has_devices, devices, errp);
if (ret < 0) {
goto err_drain;
}
diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
index c1b8533d0c..1191aac3ae 100644
--- a/monitor/hmp-cmds.c
+++ b/monitor/hmp-cmds.c
@@ -1121,7 +1121,7 @@ void hmp_loadvm(Monitor *mon, const QDict *qdict)
vm_stop(RUN_STATE_RESTORE_VM);
- if (load_snapshot(name, &err) == 0 && saved_vm_running) {
+ if (load_snapshot(name, NULL, false, NULL, &err) == 0 && saved_vm_running) {
vm_start();
}
hmp_handle_error(mon, err);
@@ -1131,7 +1131,7 @@ void hmp_savevm(Monitor *mon, const QDict *qdict)
{
Error *err = NULL;
- save_snapshot(qdict_get_try_str(qdict, "name"), true, &err);
+ save_snapshot(qdict_get_try_str(qdict, "name"), true, NULL, false, NULL, &err);
hmp_handle_error(mon, err);
}
diff --git a/replay/replay-snapshot.c b/replay/replay-snapshot.c
index 8e7ff97d11..8ed09177b5 100644
--- a/replay/replay-snapshot.c
+++ b/replay/replay-snapshot.c
@@ -77,13 +77,13 @@ void replay_vmstate_init(void)
if (replay_snapshot) {
if (replay_mode == REPLAY_MODE_RECORD) {
- if (save_snapshot(replay_snapshot, true, &err) != 0) {
+ if (save_snapshot(replay_snapshot, true, NULL, false, NULL, &err) != 0) {
error_report_err(err);
error_report("Could not create snapshot for icount record");
exit(1);
}
} else if (replay_mode == REPLAY_MODE_PLAY) {
- if (load_snapshot(replay_snapshot, &err) != 0) {
+ if (load_snapshot(replay_snapshot, NULL, false, NULL, &err) != 0) {
error_report_err(err);
error_report("Could not load snapshot for icount replay");
exit(1);
diff --git a/softmmu/vl.c b/softmmu/vl.c
index f7b103467c..820d380229 100644
--- a/softmmu/vl.c
+++ b/softmmu/vl.c
@@ -4459,7 +4459,7 @@ void qemu_init(int argc, char **argv, char **envp)
register_global_state();
if (loadvm) {
Error *local_err = NULL;
- if (load_snapshot(loadvm, &local_err) < 0) {
+ if (load_snapshot(loadvm, NULL, false, NULL, &local_err) < 0) {
error_report_err(local_err);
autostart = 0;
exit(1);
--
2.26.2
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v4 7/9] migration: introduce a delete_snapshot wrapper
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
` (5 preceding siblings ...)
2020-09-15 11:35 ` [PATCH v4 6/9] migration: wire up support for snapshot device selection Daniel P. Berrangé
@ 2020-09-15 11:35 ` Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 8/9] iotests: add support for capturing and matching QMP events Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands Daniel P. Berrangé
8 siblings, 0 replies; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
Make snapshot deletion consistent with the snapshot save
and load commands by using a wrapper around the blockdev
layer. The main difference is that we get upfront validation
of the passed in device list (if any).
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
include/migration/snapshot.h | 4 +++-
migration/savevm.c | 13 +++++++++++++
monitor/hmp-cmds.c | 2 +-
3 files changed, 17 insertions(+), 2 deletions(-)
diff --git a/include/migration/snapshot.h b/include/migration/snapshot.h
index b2c72e0a1b..f46f28745d 100644
--- a/include/migration/snapshot.h
+++ b/include/migration/snapshot.h
@@ -25,5 +25,7 @@ int load_snapshot(const char *name,
const char *vmstate,
bool has_devices, strList *devices,
Error **errp);
-
+int delete_snapshot(const char *name,
+ bool has_devices, strList *devices,
+ Error **errp);
#endif
diff --git a/migration/savevm.c b/migration/savevm.c
index b3d2ce7045..56f85be250 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -2940,6 +2940,19 @@ err_drain:
return -1;
}
+int delete_snapshot(const char *name, bool has_devices, strList *devices, Error **errp)
+{
+ if (!bdrv_all_can_snapshot(has_devices, devices, errp)) {
+ return -1;
+ }
+
+ if (bdrv_all_delete_snapshot(name, has_devices, devices, errp) < 0) {
+ return -1;
+ }
+
+ return 0;
+}
+
void vmstate_register_ram(MemoryRegion *mr, DeviceState *dev)
{
qemu_ram_set_idstr(mr->ram_block,
diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
index 1191aac3ae..278dca59b0 100644
--- a/monitor/hmp-cmds.c
+++ b/monitor/hmp-cmds.c
@@ -1140,7 +1140,7 @@ void hmp_delvm(Monitor *mon, const QDict *qdict)
Error *err = NULL;
const char *name = qdict_get_str(qdict, "name");
- bdrv_all_delete_snapshot(name, false, NULL, &err);
+ delete_snapshot(name, false, NULL, &err);
hmp_handle_error(mon, err);
}
--
2.26.2
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v4 8/9] iotests: add support for capturing and matching QMP events
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
` (6 preceding siblings ...)
2020-09-15 11:35 ` [PATCH v4 7/9] migration: introduce a delete_snapshot wrapper Daniel P. Berrangé
@ 2020-09-15 11:35 ` Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands Daniel P. Berrangé
8 siblings, 0 replies; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
When using the _launch_qemu and _send_qemu_cmd functions from
common.qemu, any QMP events get mixed in with the output from
the commands and responses.
This makes it difficult to write a test case as the ordering
of events in the output is not stable.
This introduces a variable 'capture_events' which can be set
to a list of event names. Any events listed in this variable
will not be printed, instead collected in the $QEMU_EVENTS
environment variable.
A new '_wait_event' function can be invoked to collect events
at a fixed point in time. The function will first pull events
cached in $QEMU_EVENTS variable, and if none are found, will
then read more from QMP.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
tests/qemu-iotests/common.qemu | 107 ++++++++++++++++++++++++++++++++-
1 file changed, 106 insertions(+), 1 deletion(-)
diff --git a/tests/qemu-iotests/common.qemu b/tests/qemu-iotests/common.qemu
index de680cf1c7..87d7a54001 100644
--- a/tests/qemu-iotests/common.qemu
+++ b/tests/qemu-iotests/common.qemu
@@ -53,6 +53,15 @@ _in_fd=4
# If $mismatch_only is set, only non-matching responses will
# be echoed.
#
+# If $capture_events is non-empty, then any QMP event names it lists
+# will not be echoed out, but instead collected in the $QEMU_EVENTS
+# variable. The _wait_event function can later be used to received
+# the cached events.
+#
+# If $only_capture_events is set to anything but an empty string,
+# when an error will be raised if a QMP message is seen which is
+# not an event listed in $capture_events.
+#
# If $success_or_failure is set, the meaning of the arguments is
# changed as follows:
# $2: A string to search for in the response; if found, this indicates
@@ -78,6 +87,32 @@ _timed_wait_for()
QEMU_STATUS[$h]=0
while IFS= read -t ${QEMU_COMM_TIMEOUT} resp <&${QEMU_OUT[$h]}
do
+ if [ -n "$capture_events" ]; then
+ capture=0
+ local evname
+ for evname in $capture_events
+ do
+ grep -q "\"event\": \"${evname}\"" < <(echo "${resp}")
+ if [ $? -eq 0 ]; then
+ capture=1
+ fi
+ done
+ if [ $capture = 1 ];
+ then
+ ev=$(echo "${resp}" | tr -d '\r' | tr % .)
+ QEMU_EVENTS="${QEMU_EVENTS:+${QEMU_EVENTS}%}${ev}"
+ if [ -n "$only_capture_events" ]; then
+ return
+ else
+ continue
+ fi
+ fi
+ fi
+ if [ -n "$only_capture_events" ]; then
+ echo "Only expected $capture_events but got ${resp}"
+ exit 1
+ fi
+
if [ -z "${silent}" ] && [ -z "${mismatch_only}" ]; then
echo "${resp}" | _filter_testdir | _filter_qemu \
| _filter_qemu_io | _filter_qmp | _filter_hmp
@@ -177,12 +212,82 @@ _send_qemu_cmd()
let count--;
done
if [ ${QEMU_STATUS[$h]} -ne 0 ] && [ -z "${qemu_error_no_exit}" ]; then
- echo "Timeout waiting for ${1} on handle ${h}"
+ echo "Timeout waiting for command ${1} response on handle ${h}"
exit 1 #Timeout means the test failed
fi
}
+# Check event cache for a named QMP event
+#
+# Input parameters:
+# $1: Name of the QMP event to check for
+#
+# Checks if the named QMP event that was previously captured
+# into $QEMU_EVENTS. When matched, the QMP event will be echoed
+# and the $matched variable set to 1.
+#
+# _wait_event is more suitable for test usage in most cases
+_check_cached_events()
+{
+ local evname=${1}
+
+ local match="\"event\": \"$evname\""
+
+ matched=0
+ if [ -n "$QEMU_EVENTS" ]; then
+ CURRENT_QEMU_EVENTS=$QEMU_EVENTS
+ QEMU_EVENTS=
+ old_IFS=$IFS
+ IFS="%"
+ for ev in $CURRENT_QEMU_EVENTS
+ do
+ grep -q "$match" < <(echo "${ev}")
+ if [ $? -eq 0 -a $matched = 0 ]; then
+ echo "${ev}" | _filter_testdir | _filter_qemu \
+ | _filter_qemu_io | _filter_qmp | _filter_hmp
+ matched=1
+ else
+ QEMU_EVENTS="${QEMU_EVENTS:+${QEMU_EVENTS}%}${ev}"
+ fi
+ done
+ IFS=$old_IFS
+ fi
+}
+
+# Wait for a named QMP event
+#
+# Input parameters:
+# $1: QEMU handle to use
+# $2: Name of the QMP event to wait for
+#
+# Checks if the named QMP event that was previously captured
+# into $QEMU_EVENTS. If none are present, then waits for the
+# event to arrive on the QMP channel. When matched, the QMP
+# event will be echoed
+_wait_event()
+{
+ local h=${1}
+ local evname=${2}
+
+ while true
+ do
+ _check_cached_events $evname
+
+ if [ $matched = 1 ];
+ then
+ return
+ fi
+
+ only_capture_events=1 qemu_error_no_exit=1 _timed_wait_for ${h}
+
+ if [ ${QEMU_STATUS[$h]} -ne 0 ] ; then
+ echo "Timeout waiting for event ${evname} on handle ${h}"
+ exit 1 #Timeout means the test failed
+ fi
+ done
+}
+
# Launch a QEMU process.
#
# Input parameters:
--
2.26.2
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
` (7 preceding siblings ...)
2020-09-15 11:35 ` [PATCH v4 8/9] iotests: add support for capturing and matching QMP events Daniel P. Berrangé
@ 2020-09-15 11:35 ` Daniel P. Berrangé
2020-09-16 8:17 ` Markus Armbruster
8 siblings, 1 reply; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-15 11:35 UTC (permalink / raw)
To: qemu-devel
Cc: Kevin Wolf, Peter Krempa, Daniel P. Berrangé,
Denis V. Lunev, qemu-block, Juan Quintela,
Dr. David Alan Gilbert, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
savevm, loadvm and delvm are some of the few HMP commands that have never
been converted to use QMP. The reasons for the lack of conversion are
that they blocked execution of the event thread, and the semantics
around choice of disks were ill-defined.
Despite this downside, however, libvirt and applications using libvirt
have used these commands for as long as QMP has existed, via the
"human-monitor-command" passthrough command. IOW, while it is clearly
desirable to be able to fix the problems, they are not a blocker to
all real world usage.
Meanwhile there is a need for other features which involve adding new
parameters to the commands. This is possible with HMP passthrough, but
it provides no reliable way for apps to introspect features, so using
QAPI modelling is highly desirable.
This patch thus introduces new snapshot-{load,save,delete} commands to
QMP that are intended to replace the old HMP counterparts. The new
commands are given different names, because they will be using the new
QEMU job framework and thus will have diverging behaviour from the HMP
originals. It would thus be misleading to keep the same name.
While this design uses the generic job framework, the current impl is
still blocking. The intention that the blocking problem is fixed later.
None the less applications using these new commands should assume that
they are asynchronous and thus wait for the job status change event to
indicate completion.
In addition to using the job framework, the new commands require the
caller to be explicit about all the block device nodes used in the
snapshot operations, with no built-in default heuristics in use.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
---
migration/savevm.c | 183 ++++++++++++++++
qapi/job.json | 9 +-
qapi/migration.json | 120 +++++++++++
tests/qemu-iotests/310 | 338 ++++++++++++++++++++++++++++++
tests/qemu-iotests/310.out | 412 +++++++++++++++++++++++++++++++++++++
tests/qemu-iotests/group | 1 +
6 files changed, 1062 insertions(+), 1 deletion(-)
create mode 100755 tests/qemu-iotests/310
create mode 100644 tests/qemu-iotests/310.out
diff --git a/migration/savevm.c b/migration/savevm.c
index 56f85be250..85b50953ad 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -2980,3 +2980,186 @@ bool vmstate_check_only_migratable(const VMStateDescription *vmsd)
return !(vmsd && vmsd->unmigratable);
}
+
+typedef struct SnapshotJob {
+ Job common;
+ char *tag;
+ char *vmstate;
+ strList *devices;
+ Coroutine *co;
+ Error **errp;
+ int ret;
+} SnapshotJob;
+
+static void qmp_snapshot_job_free(SnapshotJob *s)
+{
+ g_free(s->tag);
+ g_free(s->vmstate);
+ qapi_free_strList(s->devices);
+}
+
+
+static void snapshot_load_job_bh(void *opaque)
+{
+ Job *job = opaque;
+ SnapshotJob *s = container_of(job, SnapshotJob, common);
+ int saved_vm_running;
+
+ job_progress_set_remaining(&s->common, 1);
+
+ saved_vm_running = runstate_is_running();
+ vm_stop(RUN_STATE_RESTORE_VM);
+
+ s->ret = load_snapshot(s->tag, s->vmstate, true, s->devices, s->errp);
+ if (s->ret == 0 && saved_vm_running) {
+ vm_start();
+ }
+
+ job_progress_update(&s->common, 1);
+
+ qmp_snapshot_job_free(s);
+ aio_co_wake(s->co);
+}
+
+static void snapshot_save_job_bh(void *opaque)
+{
+ Job *job = opaque;
+ SnapshotJob *s = container_of(job, SnapshotJob, common);
+
+ job_progress_set_remaining(&s->common, 1);
+ s->ret = save_snapshot(s->tag, false, s->vmstate, true, s->devices, s->errp);
+ job_progress_update(&s->common, 1);
+
+ qmp_snapshot_job_free(s);
+ aio_co_wake(s->co);
+}
+
+static void snapshot_delete_job_bh(void *opaque)
+{
+ Job *job = opaque;
+ SnapshotJob *s = container_of(job, SnapshotJob, common);
+
+ job_progress_set_remaining(&s->common, 1);
+ s->ret = delete_snapshot(s->tag, true, s->devices, s->errp);
+ job_progress_update(&s->common, 1);
+
+ qmp_snapshot_job_free(s);
+ aio_co_wake(s->co);
+}
+
+static int coroutine_fn snapshot_save_job_run(Job *job, Error **errp)
+{
+ SnapshotJob *s = container_of(job, SnapshotJob, common);
+ s->errp = errp;
+ s->co = qemu_coroutine_self();
+ aio_bh_schedule_oneshot(qemu_get_aio_context(),
+ snapshot_save_job_bh, job);
+ qemu_coroutine_yield();
+ return s->ret;
+}
+
+static int coroutine_fn snapshot_load_job_run(Job *job, Error **errp)
+{
+ SnapshotJob *s = container_of(job, SnapshotJob, common);
+ s->errp = errp;
+ s->co = qemu_coroutine_self();
+ aio_bh_schedule_oneshot(qemu_get_aio_context(),
+ snapshot_load_job_bh, job);
+ qemu_coroutine_yield();
+ return s->ret;
+}
+
+static int coroutine_fn snapshot_delete_job_run(Job *job, Error **errp)
+{
+ SnapshotJob *s = container_of(job, SnapshotJob, common);
+ s->errp = errp;
+ s->co = qemu_coroutine_self();
+ aio_bh_schedule_oneshot(qemu_get_aio_context(),
+ snapshot_delete_job_bh, job);
+ qemu_coroutine_yield();
+ return s->ret;
+}
+
+
+static const JobDriver snapshot_load_job_driver = {
+ .instance_size = sizeof(SnapshotJob),
+ .job_type = JOB_TYPE_SNAPSHOT_LOAD,
+ .run = snapshot_load_job_run,
+};
+
+static const JobDriver snapshot_save_job_driver = {
+ .instance_size = sizeof(SnapshotJob),
+ .job_type = JOB_TYPE_SNAPSHOT_SAVE,
+ .run = snapshot_save_job_run,
+};
+
+static const JobDriver snapshot_delete_job_driver = {
+ .instance_size = sizeof(SnapshotJob),
+ .job_type = JOB_TYPE_SNAPSHOT_DELETE,
+ .run = snapshot_delete_job_run,
+};
+
+
+void qmp_snapshot_save(const char *job_id,
+ const char *tag,
+ const char *vmstate,
+ strList *devices,
+ Error **errp)
+{
+ SnapshotJob *s;
+
+ s = job_create(job_id, &snapshot_save_job_driver, NULL,
+ qemu_get_aio_context(), JOB_MANUAL_DISMISS,
+ NULL, NULL, errp);
+ if (!s) {
+ return;
+ }
+
+ s->tag = g_strdup(tag);
+ s->vmstate = g_strdup(vmstate);
+ s->devices = QAPI_CLONE(strList, devices);
+
+ job_start(&s->common);
+}
+
+void qmp_snapshot_load(const char *job_id,
+ const char *tag,
+ const char *vmstate,
+ strList *devices,
+ Error **errp)
+{
+ SnapshotJob *s;
+
+ s = job_create(job_id, &snapshot_load_job_driver, NULL,
+ qemu_get_aio_context(), JOB_MANUAL_DISMISS,
+ NULL, NULL, errp);
+ if (!s) {
+ return;
+ }
+
+ s->tag = g_strdup(tag);
+ s->vmstate = g_strdup(vmstate);
+ s->devices = QAPI_CLONE(strList, devices);
+
+ job_start(&s->common);
+}
+
+void qmp_snapshot_delete(const char *job_id,
+ const char *tag,
+ strList *devices,
+ Error **errp)
+{
+ SnapshotJob *s;
+
+ s = job_create(job_id, &snapshot_delete_job_driver, NULL,
+ qemu_get_aio_context(), JOB_MANUAL_DISMISS,
+ NULL, NULL, errp);
+ if (!s) {
+ return;
+ }
+
+ s->tag = g_strdup(tag);
+ s->devices = QAPI_CLONE(strList, devices);
+
+ job_start(&s->common);
+}
diff --git a/qapi/job.json b/qapi/job.json
index 280c2f76f1..b2cbb4fead 100644
--- a/qapi/job.json
+++ b/qapi/job.json
@@ -22,10 +22,17 @@
#
# @amend: image options amend job type, see "x-blockdev-amend" (since 5.1)
#
+# @snapshot-load: snapshot load job type, see "snapshot-load" (since 5.2)
+#
+# @snapshot-save: snapshot save job type, see "snapshot-save" (since 5.2)
+#
+# @snapshot-delete: snapshot delete job type, see "snapshot-delete" (since 5.2)
+#
# Since: 1.7
##
{ 'enum': 'JobType',
- 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend'] }
+ 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend',
+ 'snapshot-load', 'snapshot-save', 'snapshot-delete'] }
##
# @JobStatus:
diff --git a/qapi/migration.json b/qapi/migration.json
index 675f70bb67..b584c0be31 100644
--- a/qapi/migration.json
+++ b/qapi/migration.json
@@ -1720,3 +1720,123 @@
##
{ 'event': 'UNPLUG_PRIMARY',
'data': { 'device-id': 'str' } }
+
+##
+# @snapshot-save:
+#
+# Save a VM snapshot
+#
+# @job-id: identifier for the newly created job
+# @tag: name of the snapshot to create
+# @devices: list of block device node names to save a snapshot to
+# @vmstate: block device node name to save vmstate to
+#
+# Applications should not assume that the snapshot save is complete
+# when this command returns. The job commands / events must be used
+# to determine completion and to fetch details of any errors that arise.
+#
+# Note that the VM CPUs will be paused during the time it takes to
+# save the snapshot
+#
+# It is strongly recommended that @devices contain all writable
+# block device nodes if a consistent snapshot is required.
+#
+# If @tag already exists, an error will be reported
+#
+# Returns: nothing
+#
+# Example:
+#
+# -> { "execute": "snapshot-save",
+# "data": {
+# "job-id": "snapsave0",
+# "tag": "my-snap",
+# "vmstate": "disk0",
+# "devices": ["disk0", "disk1"]
+# }
+# }
+# <- { "return": { } }
+#
+# Since: 5.2
+##
+{ 'command': 'snapshot-save',
+ 'data': { 'job-id': 'str',
+ 'tag': 'str',
+ 'vmstate': 'str',
+ 'devices': ['str'] } }
+
+##
+# @snapshot-load:
+#
+# Load a VM snapshot
+#
+# @job-id: identifier for the newly created job
+# @tag: name of the snapshot to load.
+# @devices: list of block device node names to load a snapshot from
+# @vmstate: block device node name to load vmstate from
+#
+# Applications should not assume that the snapshot save is complete
+# when this command returns. The job commands / events must be used
+# to determine completion and to fetch details of any errors that arise.
+#
+# Note that the VM CPUs will be paused during the time it takes to
+# save the snapshot
+#
+# It is strongly recommended that @devices contain all writable
+# block device nodes that can have changed since the original
+# @snapshot-save command execution.
+#
+# Returns: nothing
+#
+# Example:
+#
+# -> { "execute": "snapshot-load",
+# "data": {
+# "job-id": "snapload0",
+# "tag": "my-snap",
+# "vmstate": "disk0",
+# "devices": ["disk0", "disk1"]
+# }
+# }
+# <- { "return": { } }
+#
+# Since: 5.2
+##
+{ 'command': 'snapshot-load',
+ 'data': { 'job-id': 'str',
+ 'tag': 'str',
+ 'vmstate': 'str',
+ 'devices': ['str'] } }
+
+##
+# @snapshot-delete:
+#
+# Delete a VM snapshot
+#
+# @job-id: identifier for the newly created job
+# @tag: name of the snapshot to delete.
+# @devices: list of block device node names to delete a snapshot from
+#
+# Applications should not assume that the snapshot save is complete
+# when this command returns. The job commands / events must be used
+# to determine completion and to fetch details of any errors that arise.
+#
+# Returns: nothing
+#
+# Example:
+#
+# -> { "execute": "snapshot-delete",
+# "data": {
+# "job-id": "snapdelete0",
+# "tag": "my-snap",
+# "devices": ["disk0", "disk1"]
+# }
+# }
+# <- { "return": { } }
+#
+# Since: 5.2
+##
+{ 'command': 'snapshot-delete',
+ 'data': { 'job-id': 'str',
+ 'tag': 'str',
+ 'devices': ['str'] } }
diff --git a/tests/qemu-iotests/310 b/tests/qemu-iotests/310
new file mode 100755
index 0000000000..622451c79f
--- /dev/null
+++ b/tests/qemu-iotests/310
@@ -0,0 +1,338 @@
+#!/usr/bin/env bash
+#
+# Test which nodes are involved in internal snapshots
+#
+# Copyright (C) 2020 Red Hat, Inc.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+#
+
+# creator
+owner=berrange@redhat.com
+
+seq=`basename $0`
+echo "QA output created by $seq"
+
+status=1 # failure is the default!
+
+_cleanup()
+{
+ _cleanup_qemu
+ _cleanup_test_img
+ TEST_IMG="$TEST_IMG.alt1" _cleanup_test_img
+ TEST_IMG="$TEST_IMG.alt2" _cleanup_test_img
+ rm -f "$SOCK_DIR/nbd"
+}
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+. ./common.qemu
+
+_supported_fmt qcow2
+_supported_proto file
+_supported_os Linux
+_require_drivers copy-on-read
+
+# Internal snapshots are (currently) impossible with refcount_bits=1,
+# and generally impossible with external data files
+_unsupported_imgopts 'refcount_bits=1[^0-9]' data_file
+
+_require_devices virtio-blk
+
+
+size=128M
+
+if [ -n "$BACKING_FILE" ]; then
+ _make_test_img -b "$BACKING_FILE" -F $IMGFMT $size
+else
+ _make_test_img $size
+fi
+TEST_IMG="$TEST_IMG.alt1" _make_test_img $size
+IMGOPTS= IMGFMT=raw TEST_IMG="$TEST_IMG.alt2" _make_test_img $size
+
+export capture_events="JOB_STATUS_CHANGE STOP RESUME"
+
+wait_job()
+{
+ local job=$1
+ shift
+
+ # All jobs start with two events...
+ #
+ # created
+ _wait_event $QEMU_HANDLE "JOB_STATUS_CHANGE"
+ # running
+ _wait_event $QEMU_HANDLE "JOB_STATUS_CHANGE"
+
+ # Next events vary depending on job type and
+ # whether it succeeds or not.
+ for evname in $@
+ do
+ _wait_event $QEMU_HANDLE $evname
+ done
+
+ # All jobs finish off with two more events...
+ # concluded
+ _wait_event $QEMU_HANDLE "JOB_STATUS_CHANGE"
+ _send_qemu_cmd $QEMU_HANDLE "{\"execute\": \"query-jobs\"}" "return"
+ _send_qemu_cmd $QEMU_HANDLE "{\"execute\": \"job-dismiss\", \"arguments\": {\"id\": \"$job\"}}" "return"
+ # null
+ _wait_event $QEMU_HANDLE "JOB_STATUS_CHANGE"
+}
+
+run_save()
+{
+ local job=$1
+ local vmstate=$2
+ local devices=$3
+ local fail=$4
+
+ _send_qemu_cmd $QEMU_HANDLE "{\"execute\": \"snapshot-save\",
+ \"arguments\": {
+ \"job-id\": \"$job\",
+ \"tag\": \"snap0\",
+ \"vmstate\": \"$vmstate\",
+ \"devices\": $devices}}" "return"
+
+ if [ $fail = 0 ]; then
+ # job status: waiting, pending
+ wait_job $job "STOP" "RESUME" "JOB_STATUS_CHANGE" "JOB_STATUS_CHANGE"
+ else
+ # job status: aborting
+ wait_job $job "JOB_STATUS_CHANGE"
+ fi
+}
+
+run_load()
+{
+ local job=$1
+ local vmstate=$2
+ local devices=$3
+ local fail=$4
+
+ _send_qemu_cmd $QEMU_HANDLE "{\"execute\": \"snapshot-load\",
+ \"arguments\": {
+ \"job-id\": \"$job\",
+ \"tag\": \"snap0\",
+ \"vmstate\": \"$vmstate\",
+ \"devices\": $devices}}" "return"
+ if [ $fail = 0 ]; then
+ # job status: waiting, pending
+ wait_job $job "STOP" "RESUME" "JOB_STATUS_CHANGE" "JOB_STATUS_CHANGE"
+ else
+ # job status: aborting
+ wait_job $job "STOP" "JOB_STATUS_CHANGE"
+ fi
+}
+
+run_delete()
+{
+ local job=$1
+ local devices=$2
+ local fail=$3
+
+ _send_qemu_cmd $QEMU_HANDLE "{\"execute\": \"snapshot-delete\",
+ \"arguments\": {
+ \"job-id\": \"$job\",
+ \"tag\": \"snap0\",
+ \"devices\": $devices}}" "return"
+ if [ $fail = 0 ]; then
+ # job status: waiting, pending
+ wait_job $job "JOB_STATUS_CHANGE" "JOB_STATUS_CHANGE"
+ else
+ # job status: aborting
+ wait_job $job "JOB_STATUS_CHANGE"
+ fi
+}
+
+start_qemu()
+{
+ keep_stderr=y
+ _launch_qemu -nodefaults -nographic "$@"
+
+ _send_qemu_cmd $QEMU_HANDLE '{"execute": "qmp_capabilities"}' 'return'
+}
+
+stop_qemu()
+{
+ _send_qemu_cmd $QEMU_HANDLE '{"execute": "quit"}' 'return'
+
+ wait=1 _cleanup_qemu
+}
+
+run_test()
+{
+ local job=$1
+ local vmstate=$2
+ local devices=$3
+ local savefail=$4
+ local loadfail=$5
+ local delfail=$6
+ shift
+ shift
+ shift
+ shift
+ shift
+ shift
+
+ start_qemu $@
+ run_save "save-$job" "$vmstate" "$devices" "$savefail"
+ run_load "load-$job" "$vmstate" "$devices" "$loadfail"
+ run_delete "delete-$job" "$devices" "$delfail"
+ stop_qemu
+}
+
+
+echo
+echo "===== Snapshot single qcow2 image ====="
+echo
+
+run_test "simple" "diskfmt0" "[\"diskfmt0\"]" 0 0 0 \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}"
+
+
+echo
+echo "===== Snapshot no image ====="
+echo
+
+# When snapshotting we need to pass at least one writable disk
+# otherwise there's no work to do
+
+run_test "no-image" "diskfmt0" "[]" 1 1 1 \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}"
+
+
+echo
+echo "===== Snapshot missing image ====="
+echo
+
+# The block node names we pass need to actually exist
+
+run_test "missing-image" "diskfmt1729" "[\"diskfmt1729\"]" 1 1 1 \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}"
+
+
+echo
+echo "===== Snapshot vmstate not in devices list ====="
+echo
+
+# The node name referred to for vmstate must be one of the nodes
+# being included in the snapshot, otherwise the vmstate that is
+# captured is liable to be overwritten making subsequent load
+# impossible
+
+run_test "excluded-vmstate" "diskfmt0" "[\"diskfmt1\"]" 1 1 0 \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}" \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG.alt1','node-name':'disk1'}" \
+ -blockdev "{'driver':'qcow2','file':'disk1','node-name':'diskfmt1'}"
+
+
+echo
+echo "===== Snapshot protocol instead of format ====="
+echo
+
+# The snapshot has to be done against the qcow2 format layer
+# not the underlying file protocol layer
+
+run_test "proto-not-fmt" "disk0" "[\"disk0\"]" 1 1 1 \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}"
+
+
+echo
+echo "===== Snapshot dual qcow2 image ====="
+echo
+
+# We can snapshot multiple qcow2 disks at the same time
+
+run_test "dual-image" "diskfmt0" "[\"diskfmt0\", \"diskfmt1\"]" 0 0 0 \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}" \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG.alt1','node-name':'disk1'}" \
+ -blockdev "{'driver':'qcow2','file':'disk1','node-name':'diskfmt1'}"
+
+
+echo
+echo "===== Snapshot error with raw image ====="
+echo
+
+# If we're snapshotting multiple disks, all must be capable
+# of supporting snapshots. A raw disk in the list must cause
+# an error.
+
+run_test "raw-fmt" "diskfmt0" "[\"diskfmt0\", \"diskfmt1\", \"diskfmt2\"]" 1 1 1 \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}" \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG.alt1','node-name':'disk1'}" \
+ -blockdev "{'driver':'qcow2','file':'disk1','node-name':'diskfmt1'}" \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG.alt2','node-name':'disk2'}" \
+ -blockdev "{'driver':'raw','file':'disk2','node-name':'diskfmt2'}"
+
+
+echo
+echo "===== Snapshot with raw image excluded ====="
+echo
+
+# If we're snapshotting multiple disks, all must be capable
+# of supporting snapshots. A writable raw disk can be excluded
+# from the snapshot, though it means its data won't be restored
+# by later snapshot load operation.
+
+run_test "skip-raw" "diskfmt0" "[\"diskfmt0\", \"diskfmt1\"]" 0 0 0 \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}" \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG.alt1','node-name':'disk1'}" \
+ -blockdev "{'driver':'qcow2','file':'disk1','node-name':'diskfmt1'}" \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG.alt2','node-name':'disk2'}" \
+ -blockdev "{'driver':'raw','file':'disk2','node-name':'diskfmt2'}"
+
+
+echo
+echo "===== Snapshot bad error reporting to stderr ====="
+echo
+
+# This demonstrates that we're not capturing vmstate loading failures
+# into QMP errors, they're ending up in stderr instead. vmstate needs
+# to report errors via Error object but that is a major piece of work
+# for the future. This test case's expected output log will need
+# adjusting when that is done.
+
+start_qemu \
+ -device virtio-rng \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}"
+
+run_save "save-err-stderr" "diskfmt0" "[\"diskfmt0\"]" 0
+stop_qemu
+
+# leave off virtio-rng to provoke vmstate failure
+start_qemu \
+ -blockdev "{'driver':'file','filename':'$TEST_IMG','node-name':'disk0'}" \
+ -blockdev "{'driver':'qcow2','file':'disk0','node-name':'diskfmt0'}"
+
+run_load "load-err-stderr" "diskfmt0" "[\"diskfmt0\"]" 1
+run_delete "delete-err-stderr" "[\"diskfmt0\"]" 0
+
+stop_qemu
+
+# success, all done
+echo "*** done"
+rm -f $seq.full
+status=0
diff --git a/tests/qemu-iotests/310.out b/tests/qemu-iotests/310.out
new file mode 100644
index 0000000000..b2001b2551
--- /dev/null
+++ b/tests/qemu-iotests/310.out
@@ -0,0 +1,412 @@
+QA output created by 310
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=134217728
+Formatting 'TEST_DIR/t.IMGFMT.alt1', fmt=IMGFMT size=134217728
+Formatting 'TEST_DIR/t.qcow2.alt2', fmt=IMGFMT size=134217728
+
+===== Snapshot single qcow2 image =====
+
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-save", "arguments": {"job-id": "save-simple", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "save-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "save-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "RESUME"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "save-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "save-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "save-simple"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-save", "id": "save-simple"}]}
+{"execute": "job-dismiss", "arguments": {"id": "save-simple"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "save-simple"}}
+{"execute": "snapshot-load", "arguments": {"job-id": "load-simple", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "load-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "load-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "RESUME"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "load-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "load-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "load-simple"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-load", "id": "load-simple"}]}
+{"execute": "job-dismiss", "arguments": {"id": "load-simple"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "load-simple"}}
+{"execute": "snapshot-delete", "arguments": {"job-id": "delete-simple", "tag": "snap0", "devices": ["diskfmt0"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "delete-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "delete-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "delete-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "delete-simple"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "delete-simple"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-delete", "id": "delete-simple"}]}
+{"execute": "job-dismiss", "arguments": {"id": "delete-simple"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "delete-simple"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+
+===== Snapshot no image =====
+
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-save", "arguments": {"job-id": "save-no-image", "tag": "snap0", "vmstate": "diskfmt0", "devices": []}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "save-no-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "save-no-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "save-no-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "save-no-image"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-save", "id": "save-no-image", "error": "At least one device is required for snapshot"}]}
+{"execute": "job-dismiss", "arguments": {"id": "save-no-image"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "save-no-image"}}
+{"execute": "snapshot-load", "arguments": {"job-id": "load-no-image", "tag": "snap0", "vmstate": "diskfmt0", "devices": []}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "load-no-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "load-no-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "load-no-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "load-no-image"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-load", "id": "load-no-image", "error": "At least one device is required for snapshot"}]}
+{"execute": "job-dismiss", "arguments": {"id": "load-no-image"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "load-no-image"}}
+{"execute": "snapshot-delete", "arguments": {"job-id": "delete-no-image", "tag": "snap0", "devices": []}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "delete-no-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "delete-no-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "delete-no-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "delete-no-image"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-delete", "id": "delete-no-image", "error": "At least one device is required for snapshot"}]}
+{"execute": "job-dismiss", "arguments": {"id": "delete-no-image"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "delete-no-image"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+
+===== Snapshot missing image =====
+
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-save", "arguments": {"job-id": "save-missing-image", "tag": "snap0", "vmstate": "diskfmt1729", "devices": ["diskfmt1729"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "save-missing-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "save-missing-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "save-missing-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "save-missing-image"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-save", "id": "save-missing-image", "error": "No block device node 'diskfmt1729'"}]}
+{"execute": "job-dismiss", "arguments": {"id": "save-missing-image"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "save-missing-image"}}
+{"execute": "snapshot-load", "arguments": {"job-id": "load-missing-image", "tag": "snap0", "vmstate": "diskfmt1729", "devices": ["diskfmt1729"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "load-missing-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "load-missing-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "load-missing-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "load-missing-image"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-load", "id": "load-missing-image", "error": "No block device node 'diskfmt1729'"}]}
+{"execute": "job-dismiss", "arguments": {"id": "load-missing-image"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "load-missing-image"}}
+{"execute": "snapshot-delete", "arguments": {"job-id": "delete-missing-image", "tag": "snap0", "devices": ["diskfmt1729"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "delete-missing-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "delete-missing-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "delete-missing-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "delete-missing-image"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-delete", "id": "delete-missing-image", "error": "No block device node 'diskfmt1729'"}]}
+{"execute": "job-dismiss", "arguments": {"id": "delete-missing-image"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "delete-missing-image"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+
+===== Snapshot vmstate not in devices list =====
+
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-save", "arguments": { "job-id": "save-excluded-vmstate", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt1"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "save-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "save-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "save-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "save-excluded-vmstate"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-save", "id": "save-excluded-vmstate", "error": "vmstate block device 'diskfmt0' does not exist"}]}
+{"execute": "job-dismiss", "arguments": {"id": "save-excluded-vmstate"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "save-excluded-vmstate"}}
+{"execute": "snapshot-load", "arguments": { "job-id": "load-excluded-vmstate", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt1"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "load-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "load-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "load-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "load-excluded-vmstate"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-load", "id": "load-excluded-vmstate", "error": "Could not find snapshot 'snap0' on 'diskfmt1'"}]}
+{"execute": "job-dismiss", "arguments": {"id": "load-excluded-vmstate"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "load-excluded-vmstate"}}
+{"execute": "snapshot-delete", "arguments": { "job-id": "delete-excluded-vmstate", "tag": "snap0", "devices": ["diskfmt1"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "delete-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "delete-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "delete-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "delete-excluded-vmstate"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "delete-excluded-vmstate"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-delete", "id": "delete-excluded-vmstate"}]}
+{"execute": "job-dismiss", "arguments": {"id": "delete-excluded-vmstate"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "delete-excluded-vmstate"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+
+===== Snapshot protocol instead of format =====
+
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-save", "arguments": {"job-id": "save-proto-not-fmt", "tag": "snap0", "vmstate": "disk0", "devices": ["disk0"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "save-proto-not-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "save-proto-not-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "save-proto-not-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "save-proto-not-fmt"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-save", "id": "save-proto-not-fmt", "error": "Device 'disk0' is writable but does not support snapshots"}]}
+{"execute": "job-dismiss", "arguments": {"id": "save-proto-not-fmt"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "save-proto-not-fmt"}}
+{"execute": "snapshot-load", "arguments": {"job-id": "load-proto-not-fmt", "tag": "snap0", "vmstate": "disk0", "devices": ["disk0"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "load-proto-not-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "load-proto-not-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "load-proto-not-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "load-proto-not-fmt"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-load", "id": "load-proto-not-fmt", "error": "Device 'disk0' is writable but does not support snapshots"}]}
+{"execute": "job-dismiss", "arguments": {"id": "load-proto-not-fmt"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "load-proto-not-fmt"}}
+{"execute": "snapshot-delete", "arguments": {"job-id": "delete-proto-not-fmt", "tag": "snap0", "devices": ["disk0"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "delete-proto-not-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "delete-proto-not-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "delete-proto-not-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "delete-proto-not-fmt"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-delete", "id": "delete-proto-not-fmt", "error": "Device 'disk0' is writable but does not support snapshots"}]}
+{"execute": "job-dismiss", "arguments": {"id": "delete-proto-not-fmt"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "delete-proto-not-fmt"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+
+===== Snapshot dual qcow2 image =====
+
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-save", "arguments": {"job-id": "save-dual-image", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0", "diskfmt1"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "save-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "save-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "RESUME"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "save-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "save-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "save-dual-image"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-save", "id": "save-dual-image"}]}
+{"execute": "job-dismiss", "arguments": {"id": "save-dual-image"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "save-dual-image"}}
+{"execute": "snapshot-load", "arguments": {"job-id": "load-dual-image", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0", "diskfmt1"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "load-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "load-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "RESUME"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "load-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "load-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "load-dual-image"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-load", "id": "load-dual-image"}]}
+{"execute": "job-dismiss", "arguments": {"id": "load-dual-image"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "load-dual-image"}}
+{"execute": "snapshot-delete", "arguments": {"job-id": "delete-dual-image", "tag": "snap0", "devices": ["diskfmt0", "diskfmt1"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "delete-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "delete-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "delete-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "delete-dual-image"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "delete-dual-image"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-delete", "id": "delete-dual-image"}]}
+{"execute": "job-dismiss", "arguments": {"id": "delete-dual-image"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "delete-dual-image"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+
+===== Snapshot error with raw image =====
+
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-save", "arguments": {"job-id": "save-raw-fmt", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0", "diskfmt1", "diskfmt2"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "save-raw-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "save-raw-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "save-raw-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "save-raw-fmt"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-save", "id": "save-raw-fmt", "error": "Device 'diskfmt2' is writable but does not support snapshots"}]}
+{"execute": "job-dismiss", "arguments": {"id": "save-raw-fmt"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "save-raw-fmt"}}
+{"execute": "snapshot-load", "arguments": {"job-id": "load-raw-fmt", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0", "diskfmt1", "diskfmt2"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "load-raw-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "load-raw-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "load-raw-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "load-raw-fmt"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-load", "id": "load-raw-fmt", "error": "Device 'diskfmt2' is writable but does not support snapshots"}]}
+{"execute": "job-dismiss", "arguments": {"id": "load-raw-fmt"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "load-raw-fmt"}}
+{"execute": "snapshot-delete", "arguments": {"job-id": "delete-raw-fmt", "tag": "snap0", "devices": ["diskfmt0", "diskfmt1", "diskfmt2"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "delete-raw-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "delete-raw-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "delete-raw-fmt"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "delete-raw-fmt"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-delete", "id": "delete-raw-fmt", "error": "Device 'diskfmt2' is writable but does not support snapshots"}]}
+{"execute": "job-dismiss", "arguments": {"id": "delete-raw-fmt"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "delete-raw-fmt"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+
+===== Snapshot with raw image excluded =====
+
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-save", "arguments": {"job-id": "save-skip-raw", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0", "diskfmt1"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "save-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "save-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "RESUME"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "save-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "save-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "save-skip-raw"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-save", "id": "save-skip-raw"}]}
+{"execute": "job-dismiss", "arguments": {"id": "save-skip-raw"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "save-skip-raw"}}
+{"execute": "snapshot-load", "arguments": {"job-id": "load-skip-raw", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0", "diskfmt1"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "load-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "load-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "RESUME"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "load-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "load-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "load-skip-raw"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-load", "id": "load-skip-raw"}]}
+{"execute": "job-dismiss", "arguments": {"id": "load-skip-raw"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "load-skip-raw"}}
+{"execute": "snapshot-delete", "arguments": {"job-id": "delete-skip-raw", "tag": "snap0", "devices": ["diskfmt0", "diskfmt1"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "delete-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "delete-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "delete-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "delete-skip-raw"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "delete-skip-raw"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-delete", "id": "delete-skip-raw"}]}
+{"execute": "job-dismiss", "arguments": {"id": "delete-skip-raw"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "delete-skip-raw"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+
+===== Snapshot bad error reporting to stderr =====
+
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-save", "arguments": {"job-id": "save-err-stderr", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "save-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "save-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "RESUME"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "save-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "save-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "save-err-stderr"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-save", "id": "save-err-stderr"}]}
+{"execute": "job-dismiss", "arguments": {"id": "save-err-stderr"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "save-err-stderr"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+{"execute": "qmp_capabilities"}
+{"return": {}}
+{"execute": "snapshot-load", "arguments": {"job-id": "load-err-stderr", "tag": "snap0", "vmstate": "diskfmt0", "devices": ["diskfmt0"]}}
+qemu-system-x86_64: Unknown savevm section or instance '0000:00:02.0/virtio-rng' 0. Make sure that your current VM setup matches your saved VM setup, including any hotplugged devices
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "load-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "load-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "STOP"}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "load-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "load-err-stderr"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-load", "id": "load-err-stderr", "error": "Error -22 while loading VM state"}]}
+{"execute": "job-dismiss", "arguments": {"id": "load-err-stderr"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "load-err-stderr"}}
+{"execute": "snapshot-delete", "arguments": {"job-id": "delete-err-stderr", "tag": "snap0", "devices": ["diskfmt0"]}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "delete-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "delete-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "delete-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "delete-err-stderr"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "delete-err-stderr"}}
+{"execute": "query-jobs"}
+{"return": [{"current-progress": 1, "status": "concluded", "total-progress": 1, "type": "snapshot-delete", "id": "delete-err-stderr"}]}
+{"execute": "job-dismiss", "arguments": {"id": "delete-err-stderr"}}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "delete-err-stderr"}}
+{"execute": "quit"}
+{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}}
+*** done
diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
index 5cad015231..30a04b38c8 100644
--- a/tests/qemu-iotests/group
+++ b/tests/qemu-iotests/group
@@ -291,6 +291,7 @@
277 rw quick
279 rw backing quick
280 rw migration quick
+310 rw quick
281 rw quick
282 rw img quick
283 auto quick
--
2.26.2
^ permalink raw reply related [flat|nested] 18+ messages in thread
* Re: [PATCH v4 5/9] migration: control whether snapshots are ovewritten
2020-09-15 11:35 ` [PATCH v4 5/9] migration: control whether snapshots are ovewritten Daniel P. Berrangé
@ 2020-09-16 7:47 ` Markus Armbruster
2020-09-16 8:25 ` Daniel P. Berrangé
0 siblings, 1 reply; 18+ messages in thread
From: Markus Armbruster @ 2020-09-16 7:47 UTC (permalink / raw)
To: Daniel P. Berrangé
Cc: Kevin Wolf, Peter Krempa, Denis V. Lunev, qemu-block,
Juan Quintela, qemu-devel, Dr. David Alan Gilbert,
Pavel Dovgalyuk, Paolo Bonzini, Max Reitz, John Snow
Daniel P. Berrangé <berrange@redhat.com> writes:
> The traditional HMP "savevm" command will overwrite an existing snapshot
> if it already exists with the requested name. This new flag allows this
> to be controlled allowing for safer behaviour with a future QMP command.
>
> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
> ---
> include/migration/snapshot.h | 2 +-
> migration/savevm.c | 4 ++--
> monitor/hmp-cmds.c | 2 +-
> replay/replay-snapshot.c | 2 +-
> 4 files changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/include/migration/snapshot.h b/include/migration/snapshot.h
> index c85b6ec75b..d7db1174ef 100644
> --- a/include/migration/snapshot.h
> +++ b/include/migration/snapshot.h
> @@ -15,7 +15,7 @@
> #ifndef QEMU_MIGRATION_SNAPSHOT_H
> #define QEMU_MIGRATION_SNAPSHOT_H
>
> -int save_snapshot(const char *name, Error **errp);
> +int save_snapshot(const char *name, bool overwrite, Error **errp);
> int load_snapshot(const char *name, Error **errp);
>
> #endif
> diff --git a/migration/savevm.c b/migration/savevm.c
> index 76972d69b0..2025e3e579 100644
> --- a/migration/savevm.c
> +++ b/migration/savevm.c
> @@ -2658,7 +2658,7 @@ int qemu_load_device_state(QEMUFile *f)
> return 0;
> }
>
> -int save_snapshot(const char *name, Error **errp)
> +int save_snapshot(const char *name, bool overwrite, Error **errp)
> {
> BlockDriverState *bs;
> QEMUSnapshotInfo sn1, *sn = &sn1, old_sn1, *old_sn = &old_sn1;
> @@ -2685,7 +2685,7 @@ int save_snapshot(const char *name, Error **errp)
> }
>
> /* Delete old snapshots of the same name */
> - if (name) {
> + if (name && overwrite) {
> if (bdrv_all_delete_snapshot(name, false, NULL, errp) < 0) {
> return ret;
> }
Are you sure this is sane?
To see what happens, I set a breakpoint on this function, set overwrite
to false. I got a *second* snapshot with the same ID.
> diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
> index 98c98ae182..c1b8533d0c 100644
> --- a/monitor/hmp-cmds.c
> +++ b/monitor/hmp-cmds.c
> @@ -1131,7 +1131,7 @@ void hmp_savevm(Monitor *mon, const QDict *qdict)
> {
> Error *err = NULL;
>
> - save_snapshot(qdict_get_try_str(qdict, "name"), &err);
> + save_snapshot(qdict_get_try_str(qdict, "name"), true, &err);
> hmp_handle_error(mon, err);
> }
>
> diff --git a/replay/replay-snapshot.c b/replay/replay-snapshot.c
> index e26fa4c892..8e7ff97d11 100644
> --- a/replay/replay-snapshot.c
> +++ b/replay/replay-snapshot.c
> @@ -77,7 +77,7 @@ void replay_vmstate_init(void)
>
> if (replay_snapshot) {
> if (replay_mode == REPLAY_MODE_RECORD) {
> - if (save_snapshot(replay_snapshot, &err) != 0) {
> + if (save_snapshot(replay_snapshot, true, &err) != 0) {
> error_report_err(err);
> error_report("Could not create snapshot for icount record");
> exit(1);
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands
2020-09-15 11:35 ` [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands Daniel P. Berrangé
@ 2020-09-16 8:17 ` Markus Armbruster
2020-09-16 8:27 ` Daniel P. Berrangé
2020-09-21 18:16 ` Eric Blake
0 siblings, 2 replies; 18+ messages in thread
From: Markus Armbruster @ 2020-09-16 8:17 UTC (permalink / raw)
To: Daniel P. Berrangé
Cc: Kevin Wolf, Peter Krempa, Denis V. Lunev, qemu-block,
Juan Quintela, Markus Armbruster, qemu-devel,
Dr. David Alan Gilbert, Pavel Dovgalyuk, Paolo Bonzini,
Max Reitz, John Snow
Daniel P. Berrangé <berrange@redhat.com> writes:
> savevm, loadvm and delvm are some of the few HMP commands that have never
> been converted to use QMP. The reasons for the lack of conversion are
> that they blocked execution of the event thread, and the semantics
> around choice of disks were ill-defined.
>
> Despite this downside, however, libvirt and applications using libvirt
> have used these commands for as long as QMP has existed, via the
> "human-monitor-command" passthrough command. IOW, while it is clearly
> desirable to be able to fix the problems, they are not a blocker to
> all real world usage.
>
> Meanwhile there is a need for other features which involve adding new
> parameters to the commands. This is possible with HMP passthrough, but
> it provides no reliable way for apps to introspect features, so using
> QAPI modelling is highly desirable.
>
> This patch thus introduces new snapshot-{load,save,delete} commands to
> QMP that are intended to replace the old HMP counterparts. The new
> commands are given different names, because they will be using the new
> QEMU job framework and thus will have diverging behaviour from the HMP
> originals. It would thus be misleading to keep the same name.
>
> While this design uses the generic job framework, the current impl is
> still blocking. The intention that the blocking problem is fixed later.
> None the less applications using these new commands should assume that
> they are asynchronous and thus wait for the job status change event to
> indicate completion.
>
> In addition to using the job framework, the new commands require the
> caller to be explicit about all the block device nodes used in the
> snapshot operations, with no built-in default heuristics in use.
>
> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
[...]
> diff --git a/qapi/job.json b/qapi/job.json
> index 280c2f76f1..b2cbb4fead 100644
> --- a/qapi/job.json
> +++ b/qapi/job.json
> @@ -22,10 +22,17 @@
> #
> # @amend: image options amend job type, see "x-blockdev-amend" (since 5.1)
> #
> +# @snapshot-load: snapshot load job type, see "snapshot-load" (since 5.2)
> +#
> +# @snapshot-save: snapshot save job type, see "snapshot-save" (since 5.2)
> +#
> +# @snapshot-delete: snapshot delete job type, see "snapshot-delete" (since 5.2)
> +#
> # Since: 1.7
> ##
> { 'enum': 'JobType',
> - 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend'] }
> + 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend',
> + 'snapshot-load', 'snapshot-save', 'snapshot-delete'] }
>
> ##
> # @JobStatus:
> diff --git a/qapi/migration.json b/qapi/migration.json
> index 675f70bb67..b584c0be31 100644
> --- a/qapi/migration.json
> +++ b/qapi/migration.json
> @@ -1720,3 +1720,123 @@
> ##
> { 'event': 'UNPLUG_PRIMARY',
> 'data': { 'device-id': 'str' } }
> +
> +##
> +# @snapshot-save:
> +#
> +# Save a VM snapshot
> +#
> +# @job-id: identifier for the newly created job
> +# @tag: name of the snapshot to create
> +# @devices: list of block device node names to save a snapshot to
Looks like you dropped the idea to also accept drive IDs. Is that for
good, or would you like to add it later?
> +# @vmstate: block device node name to save vmstate to
> +#
> +# Applications should not assume that the snapshot save is complete
> +# when this command returns. The job commands / events must be used
> +# to determine completion and to fetch details of any errors that arise.
> +#
> +# Note that the VM CPUs will be paused during the time it takes to
> +# save the snapshot
End the sentence with a period, please.
> +#
> +# It is strongly recommended that @devices contain all writable
> +# block device nodes if a consistent snapshot is required.
If it doesn't, the snapshot is partial, and a consistent restore from a
partial snapshot is generally impossible. The comment is okay as is.
> +#
> +# If @tag already exists, an error will be reported
> +#
> +# Returns: nothing
> +#
> +# Example:
> +#
> +# -> { "execute": "snapshot-save",
> +# "data": {
> +# "job-id": "snapsave0",
> +# "tag": "my-snap",
> +# "vmstate": "disk0",
> +# "devices": ["disk0", "disk1"]
> +# }
> +# }
> +# <- { "return": { } }
> +#
> +# Since: 5.2
> +##
> +{ 'command': 'snapshot-save',
> + 'data': { 'job-id': 'str',
> + 'tag': 'str',
> + 'vmstate': 'str',
> + 'devices': ['str'] } }
> +
> +##
> +# @snapshot-load:
> +#
> +# Load a VM snapshot
> +#
> +# @job-id: identifier for the newly created job
> +# @tag: name of the snapshot to load.
> +# @devices: list of block device node names to load a snapshot from
> +# @vmstate: block device node name to load vmstate from
> +#
> +# Applications should not assume that the snapshot save is complete
> +# when this command returns. The job commands / events must be used
> +# to determine completion and to fetch details of any errors that arise.
> +#
> +# Note that the VM CPUs will be paused during the time it takes to
> +# save the snapshot
> +#
> +# It is strongly recommended that @devices contain all writable
> +# block device nodes that can have changed since the original
> +# @snapshot-save command execution.
> +#
> +# Returns: nothing
> +#
> +# Example:
> +#
> +# -> { "execute": "snapshot-load",
> +# "data": {
> +# "job-id": "snapload0",
> +# "tag": "my-snap",
> +# "vmstate": "disk0",
> +# "devices": ["disk0", "disk1"]
> +# }
> +# }
> +# <- { "return": { } }
> +#
> +# Since: 5.2
> +##
> +{ 'command': 'snapshot-load',
> + 'data': { 'job-id': 'str',
> + 'tag': 'str',
> + 'vmstate': 'str',
> + 'devices': ['str'] } }
> +
> +##
> +# @snapshot-delete:
> +#
> +# Delete a VM snapshot
> +#
> +# @job-id: identifier for the newly created job
> +# @tag: name of the snapshot to delete.
> +# @devices: list of block device node names to delete a snapshot from
> +#
> +# Applications should not assume that the snapshot save is complete
> +# when this command returns. The job commands / events must be used
> +# to determine completion and to fetch details of any errors that arise.
> +#
> +# Returns: nothing
> +#
> +# Example:
> +#
> +# -> { "execute": "snapshot-delete",
> +# "data": {
> +# "job-id": "snapdelete0",
> +# "tag": "my-snap",
> +# "devices": ["disk0", "disk1"]
> +# }
> +# }
> +# <- { "return": { } }
> +#
> +# Since: 5.2
> +##
> +{ 'command': 'snapshot-delete',
> + 'data': { 'job-id': 'str',
> + 'tag': 'str',
> + 'devices': ['str'] } }
[...]
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v4 5/9] migration: control whether snapshots are ovewritten
2020-09-16 7:47 ` Markus Armbruster
@ 2020-09-16 8:25 ` Daniel P. Berrangé
0 siblings, 0 replies; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-16 8:25 UTC (permalink / raw)
To: Markus Armbruster
Cc: Kevin Wolf, Peter Krempa, Denis V. Lunev, qemu-block,
Juan Quintela, qemu-devel, Dr. David Alan Gilbert,
Pavel Dovgalyuk, Paolo Bonzini, Max Reitz, John Snow
On Wed, Sep 16, 2020 at 09:47:32AM +0200, Markus Armbruster wrote:
> Daniel P. Berrangé <berrange@redhat.com> writes:
>
> > The traditional HMP "savevm" command will overwrite an existing snapshot
> > if it already exists with the requested name. This new flag allows this
> > to be controlled allowing for safer behaviour with a future QMP command.
> >
> > Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
> > ---
> > include/migration/snapshot.h | 2 +-
> > migration/savevm.c | 4 ++--
> > monitor/hmp-cmds.c | 2 +-
> > replay/replay-snapshot.c | 2 +-
> > 4 files changed, 5 insertions(+), 5 deletions(-)
> >
> > diff --git a/include/migration/snapshot.h b/include/migration/snapshot.h
> > index c85b6ec75b..d7db1174ef 100644
> > --- a/include/migration/snapshot.h
> > +++ b/include/migration/snapshot.h
> > @@ -15,7 +15,7 @@
> > #ifndef QEMU_MIGRATION_SNAPSHOT_H
> > #define QEMU_MIGRATION_SNAPSHOT_H
> >
> > -int save_snapshot(const char *name, Error **errp);
> > +int save_snapshot(const char *name, bool overwrite, Error **errp);
> > int load_snapshot(const char *name, Error **errp);
> >
> > #endif
> > diff --git a/migration/savevm.c b/migration/savevm.c
> > index 76972d69b0..2025e3e579 100644
> > --- a/migration/savevm.c
> > +++ b/migration/savevm.c
> > @@ -2658,7 +2658,7 @@ int qemu_load_device_state(QEMUFile *f)
> > return 0;
> > }
> >
> > -int save_snapshot(const char *name, Error **errp)
> > +int save_snapshot(const char *name, bool overwrite, Error **errp)
> > {
> > BlockDriverState *bs;
> > QEMUSnapshotInfo sn1, *sn = &sn1, old_sn1, *old_sn = &old_sn1;
> > @@ -2685,7 +2685,7 @@ int save_snapshot(const char *name, Error **errp)
> > }
> >
> > /* Delete old snapshots of the same name */
> > - if (name) {
> > + if (name && overwrite) {
> > if (bdrv_all_delete_snapshot(name, false, NULL, errp) < 0) {
> > return ret;
> > }
>
> Are you sure this is sane?
>
> To see what happens, I set a breakpoint on this function, set overwrite
> to false. I got a *second* snapshot with the same ID.
Sigh. No, it doesn't do what I was meaning it to, and I forgot to add
a test case for this scenario in the last patch.
Regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands
2020-09-16 8:17 ` Markus Armbruster
@ 2020-09-16 8:27 ` Daniel P. Berrangé
2020-09-16 11:44 ` Markus Armbruster
2020-09-21 18:16 ` Eric Blake
1 sibling, 1 reply; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-09-16 8:27 UTC (permalink / raw)
To: Markus Armbruster
Cc: Kevin Wolf, Peter Krempa, Denis V. Lunev, qemu-block,
Juan Quintela, qemu-devel, Dr. David Alan Gilbert,
Pavel Dovgalyuk, Paolo Bonzini, Max Reitz, John Snow
On Wed, Sep 16, 2020 at 10:17:52AM +0200, Markus Armbruster wrote:
> Daniel P. Berrangé <berrange@redhat.com> writes:
>
> > savevm, loadvm and delvm are some of the few HMP commands that have never
> > been converted to use QMP. The reasons for the lack of conversion are
> > that they blocked execution of the event thread, and the semantics
> > around choice of disks were ill-defined.
> >
> > Despite this downside, however, libvirt and applications using libvirt
> > have used these commands for as long as QMP has existed, via the
> > "human-monitor-command" passthrough command. IOW, while it is clearly
> > desirable to be able to fix the problems, they are not a blocker to
> > all real world usage.
> >
> > Meanwhile there is a need for other features which involve adding new
> > parameters to the commands. This is possible with HMP passthrough, but
> > it provides no reliable way for apps to introspect features, so using
> > QAPI modelling is highly desirable.
> >
> > This patch thus introduces new snapshot-{load,save,delete} commands to
> > QMP that are intended to replace the old HMP counterparts. The new
> > commands are given different names, because they will be using the new
> > QEMU job framework and thus will have diverging behaviour from the HMP
> > originals. It would thus be misleading to keep the same name.
> >
> > While this design uses the generic job framework, the current impl is
> > still blocking. The intention that the blocking problem is fixed later.
> > None the less applications using these new commands should assume that
> > they are asynchronous and thus wait for the job status change event to
> > indicate completion.
> >
> > In addition to using the job framework, the new commands require the
> > caller to be explicit about all the block device nodes used in the
> > snapshot operations, with no built-in default heuristics in use.
> >
> > Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
> [...]
> > diff --git a/qapi/job.json b/qapi/job.json
> > index 280c2f76f1..b2cbb4fead 100644
> > --- a/qapi/job.json
> > +++ b/qapi/job.json
> > @@ -22,10 +22,17 @@
> > #
> > # @amend: image options amend job type, see "x-blockdev-amend" (since 5.1)
> > #
> > +# @snapshot-load: snapshot load job type, see "snapshot-load" (since 5.2)
> > +#
> > +# @snapshot-save: snapshot save job type, see "snapshot-save" (since 5.2)
> > +#
> > +# @snapshot-delete: snapshot delete job type, see "snapshot-delete" (since 5.2)
> > +#
> > # Since: 1.7
> > ##
> > { 'enum': 'JobType',
> > - 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend'] }
> > + 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend',
> > + 'snapshot-load', 'snapshot-save', 'snapshot-delete'] }
> >
> > ##
> > # @JobStatus:
> > diff --git a/qapi/migration.json b/qapi/migration.json
> > index 675f70bb67..b584c0be31 100644
> > --- a/qapi/migration.json
> > +++ b/qapi/migration.json
> > @@ -1720,3 +1720,123 @@
> > ##
> > { 'event': 'UNPLUG_PRIMARY',
> > 'data': { 'device-id': 'str' } }
> > +
> > +##
> > +# @snapshot-save:
> > +#
> > +# Save a VM snapshot
> > +#
> > +# @job-id: identifier for the newly created job
> > +# @tag: name of the snapshot to create
> > +# @devices: list of block device node names to save a snapshot to
>
> Looks like you dropped the idea to also accept drive IDs. Is that for
> good, or would you like to add it later?
I'm still kind of on the fence, but if general opinion is that we should
accept drive IDs, I'll add it.
I wonder what the other blockdev-* APIs accept - some consistency between
APIs is desirable.
Regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands
2020-09-16 8:27 ` Daniel P. Berrangé
@ 2020-09-16 11:44 ` Markus Armbruster
0 siblings, 0 replies; 18+ messages in thread
From: Markus Armbruster @ 2020-09-16 11:44 UTC (permalink / raw)
To: Daniel P. Berrangé
Cc: Kevin Wolf, Peter Krempa, Denis V. Lunev, qemu-block,
Juan Quintela, qemu-devel, Dr. David Alan Gilbert,
Pavel Dovgalyuk, Paolo Bonzini, Max Reitz, John Snow
Daniel P. Berrangé <berrange@redhat.com> writes:
> On Wed, Sep 16, 2020 at 10:17:52AM +0200, Markus Armbruster wrote:
>> Daniel P. Berrangé <berrange@redhat.com> writes:
>>
>> > savevm, loadvm and delvm are some of the few HMP commands that have never
>> > been converted to use QMP. The reasons for the lack of conversion are
>> > that they blocked execution of the event thread, and the semantics
>> > around choice of disks were ill-defined.
>> >
>> > Despite this downside, however, libvirt and applications using libvirt
>> > have used these commands for as long as QMP has existed, via the
>> > "human-monitor-command" passthrough command. IOW, while it is clearly
>> > desirable to be able to fix the problems, they are not a blocker to
>> > all real world usage.
>> >
>> > Meanwhile there is a need for other features which involve adding new
>> > parameters to the commands. This is possible with HMP passthrough, but
>> > it provides no reliable way for apps to introspect features, so using
>> > QAPI modelling is highly desirable.
>> >
>> > This patch thus introduces new snapshot-{load,save,delete} commands to
>> > QMP that are intended to replace the old HMP counterparts. The new
>> > commands are given different names, because they will be using the new
>> > QEMU job framework and thus will have diverging behaviour from the HMP
>> > originals. It would thus be misleading to keep the same name.
>> >
>> > While this design uses the generic job framework, the current impl is
>> > still blocking. The intention that the blocking problem is fixed later.
>> > None the less applications using these new commands should assume that
>> > they are asynchronous and thus wait for the job status change event to
>> > indicate completion.
>> >
>> > In addition to using the job framework, the new commands require the
>> > caller to be explicit about all the block device nodes used in the
>> > snapshot operations, with no built-in default heuristics in use.
>> >
>> > Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
>> [...]
>> > diff --git a/qapi/job.json b/qapi/job.json
>> > index 280c2f76f1..b2cbb4fead 100644
>> > --- a/qapi/job.json
>> > +++ b/qapi/job.json
>> > @@ -22,10 +22,17 @@
>> > #
>> > # @amend: image options amend job type, see "x-blockdev-amend" (since 5.1)
>> > #
>> > +# @snapshot-load: snapshot load job type, see "snapshot-load" (since 5.2)
>> > +#
>> > +# @snapshot-save: snapshot save job type, see "snapshot-save" (since 5.2)
>> > +#
>> > +# @snapshot-delete: snapshot delete job type, see "snapshot-delete" (since 5.2)
>> > +#
>> > # Since: 1.7
>> > ##
>> > { 'enum': 'JobType',
>> > - 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend'] }
>> > + 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend',
>> > + 'snapshot-load', 'snapshot-save', 'snapshot-delete'] }
>> >
>> > ##
>> > # @JobStatus:
>> > diff --git a/qapi/migration.json b/qapi/migration.json
>> > index 675f70bb67..b584c0be31 100644
>> > --- a/qapi/migration.json
>> > +++ b/qapi/migration.json
>> > @@ -1720,3 +1720,123 @@
>> > ##
>> > { 'event': 'UNPLUG_PRIMARY',
>> > 'data': { 'device-id': 'str' } }
>> > +
>> > +##
>> > +# @snapshot-save:
>> > +#
>> > +# Save a VM snapshot
>> > +#
>> > +# @job-id: identifier for the newly created job
>> > +# @tag: name of the snapshot to create
>> > +# @devices: list of block device node names to save a snapshot to
>>
>> Looks like you dropped the idea to also accept drive IDs. Is that for
>> good, or would you like to add it later?
>
> I'm still kind of on the fence, but if general opinion is that we should
> accept drive IDs, I'll add it.
I'm fine with accepting only node names. But unless we're fairly
certain node names will do, we should try to pick an interface that can
be extended to drive IDs painlessly.
> I wonder what the other blockdev-* APIs accept - some consistency between
> APIs is desirable.
The common pattern appears to be
# Either @device or @node-name must be set but not both.
#
# @device: the name of the device to get the image resized
#
# @node-name: graph node name to get the image resized (Since 2.0)
#
[...]
'*device': 'str',
'*node-name': 'str',
For snapshot-save & friends, I can see two reasonably consistent ways:
1. Have two optional lists, must specify exactly one of them.
2. Change the list element from 'str' to a struct with the two optional
members, must specify exactly one.
The second way lets you mix drive IDs and node-names freely. Do we want
to?
If yes, we can still use a variation of the first way: accept *both*
lists.
Permitting mixing makes it possible to specify the same device twice.
Could be silently accepted, or made a hard error. Matter of taste.
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v4 6/9] migration: wire up support for snapshot device selection
2020-09-15 11:35 ` [PATCH v4 6/9] migration: wire up support for snapshot device selection Daniel P. Berrangé
@ 2020-09-21 17:41 ` Dr. David Alan Gilbert
0 siblings, 0 replies; 18+ messages in thread
From: Dr. David Alan Gilbert @ 2020-09-21 17:41 UTC (permalink / raw)
To: Daniel P. Berrangé
Cc: Kevin Wolf, Peter Krempa, Denis V. Lunev, qemu-block,
Juan Quintela, qemu-devel, Markus Armbruster, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow
* Daniel P. Berrangé (berrange@redhat.com) wrote:
> Modify load_snapshot/save_snapshot to accept the device list and vmstate
> node name parameters previously added to the block layer.
>
> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
> ---
> include/migration/snapshot.h | 12 ++++++++++--
> migration/savevm.c | 24 ++++++++++++++----------
> monitor/hmp-cmds.c | 4 ++--
> replay/replay-snapshot.c | 4 ++--
> softmmu/vl.c | 2 +-
> 5 files changed, 29 insertions(+), 17 deletions(-)
>
> diff --git a/include/migration/snapshot.h b/include/migration/snapshot.h
> index d7db1174ef..b2c72e0a1b 100644
> --- a/include/migration/snapshot.h
> +++ b/include/migration/snapshot.h
> @@ -15,7 +15,15 @@
> #ifndef QEMU_MIGRATION_SNAPSHOT_H
> #define QEMU_MIGRATION_SNAPSHOT_H
>
> -int save_snapshot(const char *name, bool overwrite, Error **errp);
> -int load_snapshot(const char *name, Error **errp);
> +#include "qapi/qapi-builtin-types.h"
> +
> +int save_snapshot(const char *name, bool overwrite,
> + const char *vmstate,
> + bool has_devices, strList *devices,
> + Error **errp);
> +int load_snapshot(const char *name,
> + const char *vmstate,
> + bool has_devices, strList *devices,
> + Error **errp);
>
> #endif
> diff --git a/migration/savevm.c b/migration/savevm.c
> index 2025e3e579..b3d2ce7045 100644
> --- a/migration/savevm.c
> +++ b/migration/savevm.c
> @@ -43,6 +43,8 @@
> #include "qapi/error.h"
> #include "qapi/qapi-commands-migration.h"
> #include "qapi/qapi-commands-misc.h"
> +#include "qapi/clone-visitor.h"
> +#include "qapi/qapi-builtin-visit.h"
> #include "qapi/qmp/qerror.h"
> #include "qemu/error-report.h"
> #include "sysemu/cpus.h"
> @@ -2658,7 +2660,8 @@ int qemu_load_device_state(QEMUFile *f)
> return 0;
> }
>
> -int save_snapshot(const char *name, bool overwrite, Error **errp)
> +int save_snapshot(const char *name, bool overwrite, const char *vmstate,
> + bool has_devices, strList *devices, Error **errp)
> {
> BlockDriverState *bs;
> QEMUSnapshotInfo sn1, *sn = &sn1, old_sn1, *old_sn = &old_sn1;
> @@ -2680,18 +2683,18 @@ int save_snapshot(const char *name, bool overwrite, Error **errp)
> return ret;
> }
>
> - if (!bdrv_all_can_snapshot(false, NULL, errp)) {
> + if (!bdrv_all_can_snapshot(has_devices, devices, errp)) {
> return ret;
> }
>
> /* Delete old snapshots of the same name */
> if (name && overwrite) {
> - if (bdrv_all_delete_snapshot(name, false, NULL, errp) < 0) {
> + if (bdrv_all_delete_snapshot(name, has_devices, devices, errp) < 0) {
> return ret;
> }
> }
>
> - bs = bdrv_all_find_vmstate_bs(NULL, false, NULL, errp);
> + bs = bdrv_all_find_vmstate_bs(vmstate, has_devices, devices, errp);
> if (bs == NULL) {
> return ret;
> }
> @@ -2757,7 +2760,7 @@ int save_snapshot(const char *name, bool overwrite, Error **errp)
> aio_context_release(aio_context);
> aio_context = NULL;
>
> - ret = bdrv_all_create_snapshot(sn, bs, vm_state_size, false, NULL, errp);
> + ret = bdrv_all_create_snapshot(sn, bs, vm_state_size, has_devices, devices, errp);
> if (ret < 0) {
> goto the_end;
> }
> @@ -2858,7 +2861,8 @@ void qmp_xen_load_devices_state(const char *filename, Error **errp)
> migration_incoming_state_destroy();
> }
>
> -int load_snapshot(const char *name, Error **errp)
> +int load_snapshot(const char *name, const char *vmstate,
> + bool has_devices, strList *devices, Error **errp)
> {
> BlockDriverState *bs_vm_state;
> QEMUSnapshotInfo sn;
> @@ -2873,15 +2877,15 @@ int load_snapshot(const char *name, Error **errp)
> return -1;
> }
>
> - if (!bdrv_all_can_snapshot(false, NULL, errp)) {
> + if (!bdrv_all_can_snapshot(has_devices, devices, errp)) {
> return -1;
> }
> - ret = bdrv_all_find_snapshot(name, false, NULL, errp);
> + ret = bdrv_all_find_snapshot(name, has_devices, devices, errp);
> if (ret < 0) {
> return -1;
> }
>
> - bs_vm_state = bdrv_all_find_vmstate_bs(NULL, false, NULL, errp);
> + bs_vm_state = bdrv_all_find_vmstate_bs(vmstate, has_devices, devices, errp);
> if (!bs_vm_state) {
> return -1;
> }
> @@ -2902,7 +2906,7 @@ int load_snapshot(const char *name, Error **errp)
> /* Flush all IO requests so they don't interfere with the new state. */
> bdrv_drain_all_begin();
>
> - ret = bdrv_all_goto_snapshot(name, false, NULL, errp);
> + ret = bdrv_all_goto_snapshot(name, has_devices, devices, errp);
> if (ret < 0) {
> goto err_drain;
> }
> diff --git a/monitor/hmp-cmds.c b/monitor/hmp-cmds.c
> index c1b8533d0c..1191aac3ae 100644
> --- a/monitor/hmp-cmds.c
> +++ b/monitor/hmp-cmds.c
> @@ -1121,7 +1121,7 @@ void hmp_loadvm(Monitor *mon, const QDict *qdict)
>
> vm_stop(RUN_STATE_RESTORE_VM);
>
> - if (load_snapshot(name, &err) == 0 && saved_vm_running) {
> + if (load_snapshot(name, NULL, false, NULL, &err) == 0 && saved_vm_running) {
> vm_start();
> }
> hmp_handle_error(mon, err);
> @@ -1131,7 +1131,7 @@ void hmp_savevm(Monitor *mon, const QDict *qdict)
> {
> Error *err = NULL;
>
> - save_snapshot(qdict_get_try_str(qdict, "name"), true, &err);
> + save_snapshot(qdict_get_try_str(qdict, "name"), true, NULL, false, NULL, &err);
> hmp_handle_error(mon, err);
> }
>
> diff --git a/replay/replay-snapshot.c b/replay/replay-snapshot.c
> index 8e7ff97d11..8ed09177b5 100644
> --- a/replay/replay-snapshot.c
> +++ b/replay/replay-snapshot.c
> @@ -77,13 +77,13 @@ void replay_vmstate_init(void)
>
> if (replay_snapshot) {
> if (replay_mode == REPLAY_MODE_RECORD) {
> - if (save_snapshot(replay_snapshot, true, &err) != 0) {
> + if (save_snapshot(replay_snapshot, true, NULL, false, NULL, &err) != 0) {
> error_report_err(err);
> error_report("Could not create snapshot for icount record");
> exit(1);
> }
> } else if (replay_mode == REPLAY_MODE_PLAY) {
> - if (load_snapshot(replay_snapshot, &err) != 0) {
> + if (load_snapshot(replay_snapshot, NULL, false, NULL, &err) != 0) {
> error_report_err(err);
> error_report("Could not load snapshot for icount replay");
> exit(1);
> diff --git a/softmmu/vl.c b/softmmu/vl.c
> index f7b103467c..820d380229 100644
> --- a/softmmu/vl.c
> +++ b/softmmu/vl.c
> @@ -4459,7 +4459,7 @@ void qemu_init(int argc, char **argv, char **envp)
> register_global_state();
> if (loadvm) {
> Error *local_err = NULL;
> - if (load_snapshot(loadvm, &local_err) < 0) {
> + if (load_snapshot(loadvm, NULL, false, NULL, &local_err) < 0) {
> error_report_err(local_err);
> autostart = 0;
> exit(1);
> --
> 2.26.2
>
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands
2020-09-16 8:17 ` Markus Armbruster
2020-09-16 8:27 ` Daniel P. Berrangé
@ 2020-09-21 18:16 ` Eric Blake
2020-10-02 15:56 ` Daniel P. Berrangé
1 sibling, 1 reply; 18+ messages in thread
From: Eric Blake @ 2020-09-21 18:16 UTC (permalink / raw)
To: Markus Armbruster, Daniel P. Berrangé
Cc: Kevin Wolf, Peter Krempa, Denis V. Lunev, qemu-block,
Juan Quintela, qemu-devel, Dr. David Alan Gilbert,
Pavel Dovgalyuk, Paolo Bonzini, Max Reitz, John Snow
On 9/16/20 3:17 AM, Markus Armbruster wrote:
> Daniel P. Berrangé <berrange@redhat.com> writes:
>
>> savevm, loadvm and delvm are some of the few HMP commands that have never
>> been converted to use QMP. The reasons for the lack of conversion are
>> that they blocked execution of the event thread, and the semantics
>> around choice of disks were ill-defined.
>>
>> Despite this downside, however, libvirt and applications using libvirt
>> have used these commands for as long as QMP has existed, via the
>> "human-monitor-command" passthrough command. IOW, while it is clearly
>> desirable to be able to fix the problems, they are not a blocker to
>> all real world usage.
>>
>> Meanwhile there is a need for other features which involve adding new
>> parameters to the commands. This is possible with HMP passthrough, but
>> it provides no reliable way for apps to introspect features, so using
>> QAPI modelling is highly desirable.
>>
>> This patch thus introduces new snapshot-{load,save,delete} commands to
>> QMP that are intended to replace the old HMP counterparts. The new
>> commands are given different names, because they will be using the new
>> QEMU job framework and thus will have diverging behaviour from the HMP
>> originals. It would thus be misleading to keep the same name.
>>
>> While this design uses the generic job framework, the current impl is
>> still blocking. The intention that the blocking problem is fixed later.
>> None the less applications using these new commands should assume that
>> they are asynchronous and thus wait for the job status change event to
>> indicate completion.
>>
>> In addition to using the job framework, the new commands require the
>> caller to be explicit about all the block device nodes used in the
>> snapshot operations, with no built-in default heuristics in use.
>>
>> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
> [...]
>> diff --git a/qapi/job.json b/qapi/job.json
>> index 280c2f76f1..b2cbb4fead 100644
>> --- a/qapi/job.json
>> +++ b/qapi/job.json
>> @@ -22,10 +22,17 @@
>> #
>> # @amend: image options amend job type, see "x-blockdev-amend" (since 5.1)
>> #
>> +# @snapshot-load: snapshot load job type, see "snapshot-load" (since 5.2)
>> +#
>> +# @snapshot-save: snapshot save job type, see "snapshot-save" (since 5.2)
>> +#
>> +# @snapshot-delete: snapshot delete job type, see "snapshot-delete" (since 5.2)
>> +#
>> # Since: 1.7
>> ##
>> { 'enum': 'JobType',
>> - 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend'] }
>> + 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend',
>> + 'snapshot-load', 'snapshot-save', 'snapshot-delete'] }
>>
>> ##
>> # @JobStatus:
>> diff --git a/qapi/migration.json b/qapi/migration.json
>> index 675f70bb67..b584c0be31 100644
>> --- a/qapi/migration.json
>> +++ b/qapi/migration.json
>> @@ -1720,3 +1720,123 @@
>> ##
>> { 'event': 'UNPLUG_PRIMARY',
>> 'data': { 'device-id': 'str' } }
>> +
>> +##
>> +# @snapshot-save:
>> +#
>> +# Save a VM snapshot
>> +#
>> +# @job-id: identifier for the newly created job
>> +# @tag: name of the snapshot to create
>> +# @devices: list of block device node names to save a snapshot to
>
> Looks like you dropped the idea to also accept drive IDs. Is that for
> good, or would you like to add it later?
Is it necessary? Several of our newer block interfaces have required
node names, rather than permitting alternation. If we rewrite the
existing HMP commands to operate on top of the new QMP command, it is
still possible for HMP to support drive names even when QMP does not. I
don't think the complexity of worrying about drive names is worth it;
after all, the QMP command is new enough that the only libvirt that will
use it is also a libvirt that knows how to use -blockdev, and thus node
names are sufficient.
Yes, we can add drive ids later if I turn out to be wrong, but for now,
I'm hoping their exclusion is intentional.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3226
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands
2020-09-21 18:16 ` Eric Blake
@ 2020-10-02 15:56 ` Daniel P. Berrangé
0 siblings, 0 replies; 18+ messages in thread
From: Daniel P. Berrangé @ 2020-10-02 15:56 UTC (permalink / raw)
To: Eric Blake
Cc: Kevin Wolf, Peter Krempa, Denis V. Lunev, qemu-block,
Juan Quintela, Markus Armbruster, qemu-devel, Pavel Dovgalyuk,
Paolo Bonzini, Max Reitz, John Snow, Dr. David Alan Gilbert
On Mon, Sep 21, 2020 at 01:16:16PM -0500, Eric Blake wrote:
> On 9/16/20 3:17 AM, Markus Armbruster wrote:
> > Daniel P. Berrangé <berrange@redhat.com> writes:
> >
> > > savevm, loadvm and delvm are some of the few HMP commands that have never
> > > been converted to use QMP. The reasons for the lack of conversion are
> > > that they blocked execution of the event thread, and the semantics
> > > around choice of disks were ill-defined.
> > >
> > > Despite this downside, however, libvirt and applications using libvirt
> > > have used these commands for as long as QMP has existed, via the
> > > "human-monitor-command" passthrough command. IOW, while it is clearly
> > > desirable to be able to fix the problems, they are not a blocker to
> > > all real world usage.
> > >
> > > Meanwhile there is a need for other features which involve adding new
> > > parameters to the commands. This is possible with HMP passthrough, but
> > > it provides no reliable way for apps to introspect features, so using
> > > QAPI modelling is highly desirable.
> > >
> > > This patch thus introduces new snapshot-{load,save,delete} commands to
> > > QMP that are intended to replace the old HMP counterparts. The new
> > > commands are given different names, because they will be using the new
> > > QEMU job framework and thus will have diverging behaviour from the HMP
> > > originals. It would thus be misleading to keep the same name.
> > >
> > > While this design uses the generic job framework, the current impl is
> > > still blocking. The intention that the blocking problem is fixed later.
> > > None the less applications using these new commands should assume that
> > > they are asynchronous and thus wait for the job status change event to
> > > indicate completion.
> > >
> > > In addition to using the job framework, the new commands require the
> > > caller to be explicit about all the block device nodes used in the
> > > snapshot operations, with no built-in default heuristics in use.
> > >
> > > Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
> > [...]
> > > diff --git a/qapi/job.json b/qapi/job.json
> > > index 280c2f76f1..b2cbb4fead 100644
> > > --- a/qapi/job.json
> > > +++ b/qapi/job.json
> > > @@ -22,10 +22,17 @@
> > > #
> > > # @amend: image options amend job type, see "x-blockdev-amend" (since 5.1)
> > > #
> > > +# @snapshot-load: snapshot load job type, see "snapshot-load" (since 5.2)
> > > +#
> > > +# @snapshot-save: snapshot save job type, see "snapshot-save" (since 5.2)
> > > +#
> > > +# @snapshot-delete: snapshot delete job type, see "snapshot-delete" (since 5.2)
> > > +#
> > > # Since: 1.7
> > > ##
> > > { 'enum': 'JobType',
> > > - 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend'] }
> > > + 'data': ['commit', 'stream', 'mirror', 'backup', 'create', 'amend',
> > > + 'snapshot-load', 'snapshot-save', 'snapshot-delete'] }
> > > ##
> > > # @JobStatus:
> > > diff --git a/qapi/migration.json b/qapi/migration.json
> > > index 675f70bb67..b584c0be31 100644
> > > --- a/qapi/migration.json
> > > +++ b/qapi/migration.json
> > > @@ -1720,3 +1720,123 @@
> > > ##
> > > { 'event': 'UNPLUG_PRIMARY',
> > > 'data': { 'device-id': 'str' } }
> > > +
> > > +##
> > > +# @snapshot-save:
> > > +#
> > > +# Save a VM snapshot
> > > +#
> > > +# @job-id: identifier for the newly created job
> > > +# @tag: name of the snapshot to create
> > > +# @devices: list of block device node names to save a snapshot to
> >
> > Looks like you dropped the idea to also accept drive IDs. Is that for
> > good, or would you like to add it later?
>
> Is it necessary? Several of our newer block interfaces have required node
> names, rather than permitting alternation. If we rewrite the existing HMP
> commands to operate on top of the new QMP command, it is still possible for
> HMP to support drive names even when QMP does not. I don't think the
> complexity of worrying about drive names is worth it; after all, the QMP
> command is new enough that the only libvirt that will use it is also a
> libvirt that knows how to use -blockdev, and thus node names are sufficient.
>
> Yes, we can add drive ids later if I turn out to be wrong, but for now, I'm
> hoping their exclusion is intentional.
I didn't realize we have precedent for new commands only accepting
node names. Given that, I'm going to stick with this design and
only support node names.
Regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
^ permalink raw reply [flat|nested] 18+ messages in thread
end of thread, other threads:[~2020-10-02 15:59 UTC | newest]
Thread overview: 18+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-09-15 11:35 [PATCH v4 0/9] migration: bring improved savevm/loadvm/delvm to QMP Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 1/9] block: push error reporting into bdrv_all_*_snapshot functions Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 2/9] migration: stop returning errno from load_snapshot() Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 3/9] block: add ability to specify list of blockdevs during snapshot Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 4/9] block: allow specifying name of block device for vmstate storage Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 5/9] migration: control whether snapshots are ovewritten Daniel P. Berrangé
2020-09-16 7:47 ` Markus Armbruster
2020-09-16 8:25 ` Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 6/9] migration: wire up support for snapshot device selection Daniel P. Berrangé
2020-09-21 17:41 ` Dr. David Alan Gilbert
2020-09-15 11:35 ` [PATCH v4 7/9] migration: introduce a delete_snapshot wrapper Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 8/9] iotests: add support for capturing and matching QMP events Daniel P. Berrangé
2020-09-15 11:35 ` [PATCH v4 9/9] migration: introduce snapshot-{save, load, delete} QMP commands Daniel P. Berrangé
2020-09-16 8:17 ` Markus Armbruster
2020-09-16 8:27 ` Daniel P. Berrangé
2020-09-16 11:44 ` Markus Armbruster
2020-09-21 18:16 ` Eric Blake
2020-10-02 15:56 ` Daniel P. Berrangé
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).