* [Qemu-devel] [PATCH v2 01/40] blockjob: Update block-job-pause/resume documentation
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 14:20 ` Eric Blake
2018-05-18 17:12 ` John Snow
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 02/40] blockjob: Improve BlockJobInfo.offset/len documentation Kevin Wolf
` (41 subsequent siblings)
42 siblings, 2 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
Commit 0ec4dfb8d changed block-job_pause/resume so that they return an
error if they don't do anything because the job is already
paused/running. It forgot to update the documentation, so do that now.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
qapi/block-core.json | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/qapi/block-core.json b/qapi/block-core.json
index 55728cb823..d32ec95666 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -2338,8 +2338,7 @@
#
# This command returns immediately after marking the active background block
# operation for pausing. It is an error to call this command if no
-# operation is in progress. Pausing an already paused job has no cumulative
-# effect; a single block-job-resume command will resume the job.
+# operation is in progress or if the job is already paused.
#
# The operation will pause as soon as possible. No event is emitted when
# the operation is actually paused. Cancelling a paused job automatically
@@ -2363,7 +2362,7 @@
#
# This command returns immediately after resuming a paused background block
# operation. It is an error to call this command if no operation is in
-# progress. Resuming an already running job is not an error.
+# progress or if the job is not paused.
#
# This command also clears the error status of the job.
#
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 01/40] blockjob: Update block-job-pause/resume documentation
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 01/40] blockjob: Update block-job-pause/resume documentation Kevin Wolf
@ 2018-05-18 14:20 ` Eric Blake
2018-05-18 17:12 ` John Snow
1 sibling, 0 replies; 86+ messages in thread
From: Eric Blake @ 2018-05-18 14:20 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:20 AM, Kevin Wolf wrote:
> Commit 0ec4dfb8d changed block-job_pause/resume so that they return an
> error if they don't do anything because the job is already
> paused/running. It forgot to update the documentation, so do that now.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
> qapi/block-core.json | 5 ++---
> 1 file changed, 2 insertions(+), 3 deletions(-)
>
Reviewed-by: Eric Blake <eblake@redhat.com>
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 01/40] blockjob: Update block-job-pause/resume documentation
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 01/40] blockjob: Update block-job-pause/resume documentation Kevin Wolf
2018-05-18 14:20 ` Eric Blake
@ 2018-05-18 17:12 ` John Snow
1 sibling, 0 replies; 86+ messages in thread
From: John Snow @ 2018-05-18 17:12 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: qemu-devel, jcody, armbru, mreitz
On 05/18/2018 09:20 AM, Kevin Wolf wrote:
> Commit 0ec4dfb8d changed block-job_pause/resume so that they return an
> error if they don't do anything because the job is already
> paused/running. It forgot to update the documentation, so do that now.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
D:
Reviewed-by: John Snow <jsnow@redhat.com>
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 02/40] blockjob: Improve BlockJobInfo.offset/len documentation
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 01/40] blockjob: Update block-job-pause/resume documentation Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 14:25 ` Eric Blake
2018-05-18 17:47 ` John Snow
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 03/40] job: Create Job, JobDriver and job_create() Kevin Wolf
` (40 subsequent siblings)
42 siblings, 2 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
Clarify that len is just an estimation of the end value of offset, and
that offset increases monotonically while len can change arbitrarily.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
qapi/block-core.json | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/qapi/block-core.json b/qapi/block-core.json
index d32ec95666..0e29abf099 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -1148,7 +1148,12 @@
# @device: The job identifier. Originally the device name but other
# values are allowed since QEMU 2.7
#
-# @len: the maximum progress value
+# @len: Estimated @offset value at the completion of the job. This value can
+# arbitrarily change while the job is running, in both directions.
+#
+# @offset: Progress made until now. The unit is arbitrary and the value can
+# only meaningfully be used for the ratio of @offset to @len. The
+# value is monotonically increasing.
#
# @busy: false if the job is known to be in a quiescent state, with
# no pending I/O. Since 1.3.
@@ -1156,8 +1161,6 @@
# @paused: whether the job is paused or, if @busy is true, will
# pause itself as soon as possible. Since 1.3.
#
-# @offset: the current progress value
-#
# @speed: the rate limit, bytes per second
#
# @io-status: the status of the job (since 1.3)
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 02/40] blockjob: Improve BlockJobInfo.offset/len documentation
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 02/40] blockjob: Improve BlockJobInfo.offset/len documentation Kevin Wolf
@ 2018-05-18 14:25 ` Eric Blake
2018-05-18 17:47 ` John Snow
1 sibling, 0 replies; 86+ messages in thread
From: Eric Blake @ 2018-05-18 14:25 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:20 AM, Kevin Wolf wrote:
> Clarify that len is just an estimation of the end value of offset, and
> that offset increases monotonically while len can change arbitrarily.
That's tighter than what libvirt promises (and in fact, there are cases
where libvirt synthesizes an offset/len of 1/1 to indicate completion
when the information is no longer available from qemu, which means the
offset has gone backwards to libvirt clients).
But it matches the qemu implementation, and I'm okay if we want to stick
to those semantics of monotonically increasing offset (the more
important semantics of being an estimate of time remaining is possible
whether or not you have a guarantee of a monotonic increase on one of
the two numbers).
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
> qapi/block-core.json | 9 ++++++---
> 1 file changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/qapi/block-core.json b/qapi/block-core.json
> index d32ec95666..0e29abf099 100644
> --- a/qapi/block-core.json
> +++ b/qapi/block-core.json
> @@ -1148,7 +1148,12 @@
> # @device: The job identifier. Originally the device name but other
> # values are allowed since QEMU 2.7
> #
> -# @len: the maximum progress value
> +# @len: Estimated @offset value at the completion of the job. This value can
> +# arbitrarily change while the job is running, in both directions.
> +#
> +# @offset: Progress made until now. The unit is arbitrary and the value can
> +# only meaningfully be used for the ratio of @offset to @len. The
> +# value is monotonically increasing.
So I'm okay whether or not you drop that last sentence.
You're also rearranging the docs to occur in the order that the struct
declares things (good change, and trivial; not sure if you want to call
it out in the commit message).
Reviewed-by: Eric Blake <eblake@redhat.com>
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 02/40] blockjob: Improve BlockJobInfo.offset/len documentation
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 02/40] blockjob: Improve BlockJobInfo.offset/len documentation Kevin Wolf
2018-05-18 14:25 ` Eric Blake
@ 2018-05-18 17:47 ` John Snow
1 sibling, 0 replies; 86+ messages in thread
From: John Snow @ 2018-05-18 17:47 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: qemu-devel, jcody, armbru, mreitz
On 05/18/2018 09:20 AM, Kevin Wolf wrote:
> Clarify that len is just an estimation of the end value of offset, and
> that offset increases monotonically while len can change arbitrarily.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
> qapi/block-core.json | 9 ++++++---
> 1 file changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/qapi/block-core.json b/qapi/block-core.json
> index d32ec95666..0e29abf099 100644
> --- a/qapi/block-core.json
> +++ b/qapi/block-core.json
> @@ -1148,7 +1148,12 @@
> # @device: The job identifier. Originally the device name but other
> # values are allowed since QEMU 2.7
> #
> -# @len: the maximum progress value
> +# @len: Estimated @offset value at the completion of the job. This value can
> +# arbitrarily change while the job is running, in both directions.
> +#
> +# @offset: Progress made until now. The unit is arbitrary and the value can
> +# only meaningfully be used for the ratio of @offset to @len. The
> +# value is monotonically increasing.
> #
> # @busy: false if the job is known to be in a quiescent state, with
> # no pending I/O. Since 1.3.
> @@ -1156,8 +1161,6 @@
> # @paused: whether the job is paused or, if @busy is true, will
> # pause itself as soon as possible. Since 1.3.
> #
> -# @offset: the current progress value
> -#
> # @speed: the rate limit, bytes per second
> #
> # @io-status: the status of the job (since 1.3)
>
It matches current actual behavior, so it's probably a good update. It
feels like a change in behavior, but it's rather just codifying the
existing reality.
OK.
Reviewed-by: John Snow <jsnow@redhat.com>
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 03/40] job: Create Job, JobDriver and job_create()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 01/40] blockjob: Update block-job-pause/resume documentation Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 02/40] blockjob: Improve BlockJobInfo.offset/len documentation Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 04/40] job: Rename BlockJobType into JobType Kevin Wolf
` (39 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This is the first step towards creating an infrastructure for generic
background jobs that aren't tied to a block device. For now, Job only
stores its ID and JobDriver, the rest stays in BlockJob.
The following patches will move over more parts of BlockJob to Job if
they are meaningful outside the context of a block job.
BlockJob.driver is now redundant, but this patch leaves it around to
avoid unnecessary churn. The next patches will get rid of almost all of
its uses anyway so that it can be removed later with much less churn.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
include/block/blockjob.h | 9 +++----
include/block/blockjob_int.h | 4 +--
include/qemu/job.h | 60 ++++++++++++++++++++++++++++++++++++++++++++
block/backup.c | 4 ++-
block/commit.c | 4 ++-
block/mirror.c | 10 +++++---
block/stream.c | 4 ++-
blockjob.c | 46 ++++++++++++++++-----------------
job.c | 48 +++++++++++++++++++++++++++++++++++
tests/test-bdrv-drain.c | 4 ++-
tests/test-blockjob-txn.c | 4 ++-
tests/test-blockjob.c | 12 ++++++---
MAINTAINERS | 2 ++
Makefile.objs | 2 +-
14 files changed, 169 insertions(+), 44 deletions(-)
create mode 100644 include/qemu/job.h
create mode 100644 job.c
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 0f56f723de..640e649034 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -26,6 +26,7 @@
#ifndef BLOCKJOB_H
#define BLOCKJOB_H
+#include "qemu/job.h"
#include "block/block.h"
#include "qemu/ratelimit.h"
@@ -40,6 +41,9 @@ typedef struct BlockJobTxn BlockJobTxn;
* Long-running operation on a BlockDriverState.
*/
typedef struct BlockJob {
+ /** Data belonging to the generic Job infrastructure */
+ Job job;
+
/** The job type, including the job vtable. */
const BlockJobDriver *driver;
@@ -47,11 +51,6 @@ typedef struct BlockJob {
BlockBackend *blk;
/**
- * The ID of the block job. May be NULL for internal jobs.
- */
- char *id;
-
- /**
* The coroutine that executes the job. If not NULL, it is
* reentered when busy is false and the job is cancelled.
*/
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 62ec964d09..e8eca44747 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -35,8 +35,8 @@
* A class type for block job driver.
*/
struct BlockJobDriver {
- /** Derived BlockJob struct size */
- size_t instance_size;
+ /** Generic JobDriver callbacks and settings */
+ JobDriver job_driver;
/** String describing the operation, part of query-block-jobs QMP API */
BlockJobType job_type;
diff --git a/include/qemu/job.h b/include/qemu/job.h
new file mode 100644
index 0000000000..b4b49f19e1
--- /dev/null
+++ b/include/qemu/job.h
@@ -0,0 +1,60 @@
+/*
+ * Declarations for background jobs
+ *
+ * Copyright (c) 2011 IBM Corp.
+ * Copyright (c) 2012, 2018 Red Hat, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef JOB_H
+#define JOB_H
+
+typedef struct JobDriver JobDriver;
+
+/**
+ * Long-running operation.
+ */
+typedef struct Job {
+ /** The ID of the job. May be NULL for internal jobs. */
+ char *id;
+
+ /** The type of this job. */
+ const JobDriver *driver;
+} Job;
+
+/**
+ * Callbacks and other information about a Job driver.
+ */
+struct JobDriver {
+ /** Derived Job struct size */
+ size_t instance_size;
+};
+
+
+/**
+ * Create a new long-running job and return it.
+ *
+ * @job_id: The id of the newly-created job, or %NULL for internal jobs
+ * @driver: The class object for the newly-created job.
+ * @errp: Error object.
+ */
+void *job_create(const char *job_id, const JobDriver *driver, Error **errp);
+
+#endif
diff --git a/block/backup.c b/block/backup.c
index e14d99560d..9e672bbd5e 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -523,7 +523,9 @@ static void coroutine_fn backup_run(void *opaque)
}
static const BlockJobDriver backup_job_driver = {
- .instance_size = sizeof(BackupBlockJob),
+ .job_driver = {
+ .instance_size = sizeof(BackupBlockJob),
+ },
.job_type = BLOCK_JOB_TYPE_BACKUP,
.start = backup_run,
.commit = backup_commit,
diff --git a/block/commit.c b/block/commit.c
index ba5df6aa0a..18cbb2f9c4 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -215,7 +215,9 @@ out:
}
static const BlockJobDriver commit_job_driver = {
- .instance_size = sizeof(CommitBlockJob),
+ .job_driver = {
+ .instance_size = sizeof(CommitBlockJob),
+ },
.job_type = BLOCK_JOB_TYPE_COMMIT,
.start = commit_run,
};
diff --git a/block/mirror.c b/block/mirror.c
index a4197bb975..aa1d6b742e 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -913,7 +913,7 @@ static void mirror_complete(BlockJob *job, Error **errp)
if (!s->synced) {
error_setg(errp, "The active block job '%s' cannot be completed",
- job->id);
+ job->job.id);
return;
}
@@ -986,7 +986,9 @@ static void mirror_drain(BlockJob *job)
}
static const BlockJobDriver mirror_job_driver = {
- .instance_size = sizeof(MirrorBlockJob),
+ .job_driver = {
+ .instance_size = sizeof(MirrorBlockJob),
+ },
.job_type = BLOCK_JOB_TYPE_MIRROR,
.start = mirror_run,
.complete = mirror_complete,
@@ -996,7 +998,9 @@ static const BlockJobDriver mirror_job_driver = {
};
static const BlockJobDriver commit_active_job_driver = {
- .instance_size = sizeof(MirrorBlockJob),
+ .job_driver = {
+ .instance_size = sizeof(MirrorBlockJob),
+ },
.job_type = BLOCK_JOB_TYPE_COMMIT,
.start = mirror_run,
.complete = mirror_complete,
diff --git a/block/stream.c b/block/stream.c
index df9660d2fc..f88fc75141 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -209,7 +209,9 @@ out:
}
static const BlockJobDriver stream_job_driver = {
- .instance_size = sizeof(StreamBlockJob),
+ .job_driver = {
+ .instance_size = sizeof(StreamBlockJob),
+ },
.job_type = BLOCK_JOB_TYPE_STREAM,
.start = stream_run,
};
diff --git a/blockjob.c b/blockjob.c
index 112672a68b..1464856eb5 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -34,7 +34,6 @@
#include "qapi/qapi-events-block-core.h"
#include "qapi/qmp/qerror.h"
#include "qemu/coroutine.h"
-#include "qemu/id.h"
#include "qemu/timer.h"
/* Right now, this mutex is only needed to synchronize accesses to job->busy
@@ -92,7 +91,8 @@ static int block_job_apply_verb(BlockJob *job, BlockJobVerb bv, Error **errp)
return 0;
}
error_setg(errp, "Job '%s' in state '%s' cannot accept command verb '%s'",
- job->id, BlockJobStatus_str(job->status), BlockJobVerb_str(bv));
+ job->job.id, BlockJobStatus_str(job->status),
+ BlockJobVerb_str(bv));
return -EPERM;
}
@@ -159,7 +159,7 @@ BlockJob *block_job_get(const char *id)
BlockJob *job;
QLIST_FOREACH(job, &block_jobs, job_list) {
- if (job->id && !strcmp(id, job->id)) {
+ if (job->job.id && !strcmp(id, job->job.id)) {
return job;
}
}
@@ -261,7 +261,7 @@ void block_job_unref(BlockJob *job)
block_job_detach_aio_context, job);
blk_unref(job->blk);
error_free(job->blocker);
- g_free(job->id);
+ g_free(job->job.id);
assert(!timer_pending(&job->sleep_timer));
g_free(job);
}
@@ -311,7 +311,7 @@ static char *child_job_get_parent_desc(BdrvChild *c)
BlockJob *job = c->opaque;
return g_strdup_printf("%s job '%s'",
BlockJobType_str(job->driver->job_type),
- job->id);
+ job->job.id);
}
static void child_job_drained_begin(BdrvChild *c)
@@ -365,7 +365,7 @@ int block_job_add_bdrv(BlockJob *job, const char *name, BlockDriverState *bs,
bool block_job_is_internal(BlockJob *job)
{
- return (job->id == NULL);
+ return (job->job.id == NULL);
}
static bool block_job_started(BlockJob *job)
@@ -705,13 +705,13 @@ int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n)
void block_job_complete(BlockJob *job, Error **errp)
{
/* Should not be reachable via external interface for internal jobs */
- assert(job->id);
+ assert(job->job.id);
if (block_job_apply_verb(job, BLOCK_JOB_VERB_COMPLETE, errp)) {
return;
}
if (job->pause_count || job->cancelled || !job->driver->complete) {
error_setg(errp, "The active block job '%s' cannot be completed",
- job->id);
+ job->job.id);
return;
}
@@ -720,7 +720,7 @@ void block_job_complete(BlockJob *job, Error **errp)
void block_job_finalize(BlockJob *job, Error **errp)
{
- assert(job && job->id);
+ assert(job && job->job.id);
if (block_job_apply_verb(job, BLOCK_JOB_VERB_FINALIZE, errp)) {
return;
}
@@ -731,7 +731,7 @@ void block_job_dismiss(BlockJob **jobptr, Error **errp)
{
BlockJob *job = *jobptr;
/* similarly to _complete, this is QMP-interface only. */
- assert(job->id);
+ assert(job->job.id);
if (block_job_apply_verb(job, BLOCK_JOB_VERB_DISMISS, errp)) {
return;
}
@@ -848,7 +848,7 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
}
info = g_new0(BlockJobInfo, 1);
info->type = g_strdup(BlockJobType_str(job->driver->job_type));
- info->device = g_strdup(job->id);
+ info->device = g_strdup(job->job.id);
info->len = job->len;
info->busy = atomic_read(&job->busy);
info->paused = job->pause_count > 0;
@@ -879,7 +879,7 @@ static void block_job_event_cancelled(BlockJob *job)
}
qapi_event_send_block_job_cancelled(job->driver->job_type,
- job->id,
+ job->job.id,
job->len,
job->offset,
job->speed,
@@ -893,7 +893,7 @@ static void block_job_event_completed(BlockJob *job, const char *msg)
}
qapi_event_send_block_job_completed(job->driver->job_type,
- job->id,
+ job->job.id,
job->len,
job->offset,
job->speed,
@@ -907,7 +907,7 @@ static int block_job_event_pending(BlockJob *job)
block_job_state_transition(job, BLOCK_JOB_STATUS_PENDING);
if (!job->auto_finalize && !block_job_is_internal(job)) {
qapi_event_send_block_job_pending(job->driver->job_type,
- job->id,
+ job->job.id,
&error_abort);
}
return 0;
@@ -945,12 +945,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
error_setg(errp, "Cannot specify job ID for internal block job");
return NULL;
}
-
- if (!id_wellformed(job_id)) {
- error_setg(errp, "Invalid job ID '%s'", job_id);
- return NULL;
- }
-
if (block_job_get(job_id)) {
error_setg(errp, "Job ID '%s' already in use", job_id);
return NULL;
@@ -964,9 +958,13 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
return NULL;
}
- job = g_malloc0(driver->instance_size);
+ job = job_create(job_id, &driver->job_driver, errp);
+ if (job == NULL) {
+ blk_unref(blk);
+ return NULL;
+ }
+
job->driver = driver;
- job->id = g_strdup(job_id);
job->blk = blk;
job->cb = cb;
job->opaque = opaque;
@@ -1187,7 +1185,7 @@ void block_job_event_ready(BlockJob *job)
}
qapi_event_send_block_job_ready(job->driver->job_type,
- job->id,
+ job->job.id,
job->len,
job->offset,
job->speed, &error_abort);
@@ -1217,7 +1215,7 @@ BlockErrorAction block_job_error_action(BlockJob *job, BlockdevOnError on_err,
abort();
}
if (!block_job_is_internal(job)) {
- qapi_event_send_block_job_error(job->id,
+ qapi_event_send_block_job_error(job->job.id,
is_read ? IO_OPERATION_TYPE_READ :
IO_OPERATION_TYPE_WRITE,
action, &error_abort);
diff --git a/job.c b/job.c
new file mode 100644
index 0000000000..87fd48468c
--- /dev/null
+++ b/job.c
@@ -0,0 +1,48 @@
+/*
+ * Background jobs (long-running operations)
+ *
+ * Copyright (c) 2011 IBM Corp.
+ * Copyright (c) 2012, 2018 Red Hat, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "qemu/osdep.h"
+#include "qemu-common.h"
+#include "qapi/error.h"
+#include "qemu/job.h"
+#include "qemu/id.h"
+
+void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
+{
+ Job *job;
+
+ if (job_id) {
+ if (!id_wellformed(job_id)) {
+ error_setg(errp, "Invalid job ID '%s'", job_id);
+ return NULL;
+ }
+ }
+
+ job = g_malloc0(driver->instance_size);
+ job->driver = driver;
+ job->id = g_strdup(job_id);
+
+ return job;
+}
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index 7673de1062..fe9f412b39 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -520,7 +520,9 @@ static void test_job_complete(BlockJob *job, Error **errp)
}
BlockJobDriver test_job_driver = {
- .instance_size = sizeof(TestBlockJob),
+ .job_driver = {
+ .instance_size = sizeof(TestBlockJob),
+ },
.start = test_job_start,
.complete = test_job_complete,
};
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 5789893dda..48b12d1744 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -74,7 +74,9 @@ static void test_block_job_cb(void *opaque, int ret)
}
static const BlockJobDriver test_block_job_driver = {
- .instance_size = sizeof(TestBlockJob),
+ .job_driver = {
+ .instance_size = sizeof(TestBlockJob),
+ },
.start = test_block_job_run,
};
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 8946bfd37b..b82026180a 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -17,7 +17,9 @@
#include "sysemu/block-backend.h"
static const BlockJobDriver test_block_job_driver = {
- .instance_size = sizeof(BlockJob),
+ .job_driver = {
+ .instance_size = sizeof(BlockJob),
+ },
};
static void block_job_cb(void *opaque, int ret)
@@ -38,9 +40,9 @@ static BlockJob *mk_job(BlockBackend *blk, const char *id,
g_assert_null(errp);
g_assert_nonnull(job);
if (id) {
- g_assert_cmpstr(job->id, ==, id);
+ g_assert_cmpstr(job->job.id, ==, id);
} else {
- g_assert_cmpstr(job->id, ==, blk_name(blk));
+ g_assert_cmpstr(job->job.id, ==, blk_name(blk));
}
} else {
g_assert_nonnull(errp);
@@ -192,7 +194,9 @@ static void coroutine_fn cancel_job_start(void *opaque)
}
static const BlockJobDriver test_cancel_driver = {
- .instance_size = sizeof(CancelJob),
+ .job_driver = {
+ .instance_size = sizeof(CancelJob),
+ },
.start = cancel_job_start,
.complete = cancel_job_complete,
};
diff --git a/MAINTAINERS b/MAINTAINERS
index e187b1f18f..9fba3307d9 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1369,6 +1369,8 @@ L: qemu-block@nongnu.org
S: Supported
F: blockjob.c
F: include/block/blockjob.h
+F: job.c
+F: include/block/job.h
F: block/backup.c
F: block/commit.c
F: block/stream.c
diff --git a/Makefile.objs b/Makefile.objs
index c6c9b8fc21..92b73fc272 100644
--- a/Makefile.objs
+++ b/Makefile.objs
@@ -63,7 +63,7 @@ chardev-obj-y = chardev/
# block-obj-y is code used by both qemu system emulation and qemu-img
block-obj-y += nbd/
-block-obj-y += block.o blockjob.o
+block-obj-y += block.o blockjob.o job.o
block-obj-y += block/ scsi/
block-obj-y += qemu-io-cmds.o
block-obj-$(CONFIG_REPLICATION) += replication.o
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 04/40] job: Rename BlockJobType into JobType
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (2 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 03/40] job: Create Job, JobDriver and job_create() Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 05/40] job: Add JobDriver.job_type Kevin Wolf
` (38 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
QAPI types aren't externally visible, so we can rename them without
causing problems. Before we add a job type to Job, rename the enum
so it can be used for more than just block jobs.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
qapi/block-core.json | 14 +++++++-------
include/block/blockjob_int.h | 2 +-
block/backup.c | 2 +-
block/commit.c | 2 +-
block/mirror.c | 4 ++--
block/stream.c | 2 +-
blockjob.c | 6 +++---
7 files changed, 16 insertions(+), 16 deletions(-)
diff --git a/qapi/block-core.json b/qapi/block-core.json
index 0e29abf099..63c6011411 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -1050,9 +1050,9 @@
'data': ['top', 'full', 'none', 'incremental'] }
##
-# @BlockJobType:
+# @JobType:
#
-# Type of a block job.
+# Type of a background job.
#
# @commit: block commit job type, see "block-commit"
#
@@ -1064,7 +1064,7 @@
#
# Since: 1.7
##
-{ 'enum': 'BlockJobType',
+{ 'enum': 'JobType',
'data': ['commit', 'stream', 'mirror', 'backup'] }
##
@@ -4499,7 +4499,7 @@
#
##
{ 'event': 'BLOCK_JOB_COMPLETED',
- 'data': { 'type' : 'BlockJobType',
+ 'data': { 'type' : 'JobType',
'device': 'str',
'len' : 'int',
'offset': 'int',
@@ -4535,7 +4535,7 @@
#
##
{ 'event': 'BLOCK_JOB_CANCELLED',
- 'data': { 'type' : 'BlockJobType',
+ 'data': { 'type' : 'JobType',
'device': 'str',
'len' : 'int',
'offset': 'int',
@@ -4600,7 +4600,7 @@
#
##
{ 'event': 'BLOCK_JOB_READY',
- 'data': { 'type' : 'BlockJobType',
+ 'data': { 'type' : 'JobType',
'device': 'str',
'len' : 'int',
'offset': 'int',
@@ -4627,7 +4627,7 @@
#
##
{ 'event': 'BLOCK_JOB_PENDING',
- 'data': { 'type' : 'BlockJobType',
+ 'data': { 'type' : 'JobType',
'id' : 'str' } }
##
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index e8eca44747..8e7e0a2f57 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -39,7 +39,7 @@ struct BlockJobDriver {
JobDriver job_driver;
/** String describing the operation, part of query-block-jobs QMP API */
- BlockJobType job_type;
+ JobType job_type;
/** Mandatory: Entrypoint for the Coroutine. */
CoroutineEntry *start;
diff --git a/block/backup.c b/block/backup.c
index 9e672bbd5e..c49ea92dca 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -526,7 +526,7 @@ static const BlockJobDriver backup_job_driver = {
.job_driver = {
.instance_size = sizeof(BackupBlockJob),
},
- .job_type = BLOCK_JOB_TYPE_BACKUP,
+ .job_type = JOB_TYPE_BACKUP,
.start = backup_run,
.commit = backup_commit,
.abort = backup_abort,
diff --git a/block/commit.c b/block/commit.c
index 18cbb2f9c4..afa2b2bacf 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -218,7 +218,7 @@ static const BlockJobDriver commit_job_driver = {
.job_driver = {
.instance_size = sizeof(CommitBlockJob),
},
- .job_type = BLOCK_JOB_TYPE_COMMIT,
+ .job_type = JOB_TYPE_COMMIT,
.start = commit_run,
};
diff --git a/block/mirror.c b/block/mirror.c
index aa1d6b742e..ed72656a23 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -989,7 +989,7 @@ static const BlockJobDriver mirror_job_driver = {
.job_driver = {
.instance_size = sizeof(MirrorBlockJob),
},
- .job_type = BLOCK_JOB_TYPE_MIRROR,
+ .job_type = JOB_TYPE_MIRROR,
.start = mirror_run,
.complete = mirror_complete,
.pause = mirror_pause,
@@ -1001,7 +1001,7 @@ static const BlockJobDriver commit_active_job_driver = {
.job_driver = {
.instance_size = sizeof(MirrorBlockJob),
},
- .job_type = BLOCK_JOB_TYPE_COMMIT,
+ .job_type = JOB_TYPE_COMMIT,
.start = mirror_run,
.complete = mirror_complete,
.pause = mirror_pause,
diff --git a/block/stream.c b/block/stream.c
index f88fc75141..048bceb5d0 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -212,7 +212,7 @@ static const BlockJobDriver stream_job_driver = {
.job_driver = {
.instance_size = sizeof(StreamBlockJob),
},
- .job_type = BLOCK_JOB_TYPE_STREAM,
+ .job_type = JOB_TYPE_STREAM,
.start = stream_run,
};
diff --git a/blockjob.c b/blockjob.c
index 1464856eb5..2a3844721e 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -310,7 +310,7 @@ static char *child_job_get_parent_desc(BdrvChild *c)
{
BlockJob *job = c->opaque;
return g_strdup_printf("%s job '%s'",
- BlockJobType_str(job->driver->job_type),
+ JobType_str(job->driver->job_type),
job->job.id);
}
@@ -847,7 +847,7 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
return NULL;
}
info = g_new0(BlockJobInfo, 1);
- info->type = g_strdup(BlockJobType_str(job->driver->job_type));
+ info->type = g_strdup(JobType_str(job->driver->job_type));
info->device = g_strdup(job->job.id);
info->len = job->len;
info->busy = atomic_read(&job->busy);
@@ -980,7 +980,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
block_job_sleep_timer_cb, job);
error_setg(&job->blocker, "block device is in use by block job: %s",
- BlockJobType_str(driver->job_type));
+ JobType_str(driver->job_type));
block_job_add_bdrv(job, "main node", bs, 0, BLK_PERM_ALL, &error_abort);
bs->job = job;
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 05/40] job: Add JobDriver.job_type
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (3 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 04/40] job: Rename BlockJobType into JobType Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 06/40] job: Add job_delete() Kevin Wolf
` (37 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves the job_type field from BlockJobDriver to JobDriver.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
include/block/blockjob_int.h | 3 ---
include/qemu/job.h | 11 +++++++++++
block/backup.c | 2 +-
block/commit.c | 2 +-
block/mirror.c | 4 ++--
block/stream.c | 2 +-
blockjob.c | 16 +++++++---------
job.c | 10 ++++++++++
8 files changed, 33 insertions(+), 17 deletions(-)
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 8e7e0a2f57..1e62d6dd30 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -38,9 +38,6 @@ struct BlockJobDriver {
/** Generic JobDriver callbacks and settings */
JobDriver job_driver;
- /** String describing the operation, part of query-block-jobs QMP API */
- JobType job_type;
-
/** Mandatory: Entrypoint for the Coroutine. */
CoroutineEntry *start;
diff --git a/include/qemu/job.h b/include/qemu/job.h
index b4b49f19e1..279ce688fd 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -26,6 +26,8 @@
#ifndef JOB_H
#define JOB_H
+#include "qapi/qapi-types-block-core.h"
+
typedef struct JobDriver JobDriver;
/**
@@ -45,6 +47,9 @@ typedef struct Job {
struct JobDriver {
/** Derived Job struct size */
size_t instance_size;
+
+ /** Enum describing the operation */
+ JobType job_type;
};
@@ -57,4 +62,10 @@ struct JobDriver {
*/
void *job_create(const char *job_id, const JobDriver *driver, Error **errp);
+/** Returns the JobType of a given Job. */
+JobType job_type(const Job *job);
+
+/** Returns the enum string for the JobType of a given Job. */
+const char *job_type_str(const Job *job);
+
#endif
diff --git a/block/backup.c b/block/backup.c
index c49ea92dca..baf8d432da 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -525,8 +525,8 @@ static void coroutine_fn backup_run(void *opaque)
static const BlockJobDriver backup_job_driver = {
.job_driver = {
.instance_size = sizeof(BackupBlockJob),
+ .job_type = JOB_TYPE_BACKUP,
},
- .job_type = JOB_TYPE_BACKUP,
.start = backup_run,
.commit = backup_commit,
.abort = backup_abort,
diff --git a/block/commit.c b/block/commit.c
index afa2b2bacf..32d29c890e 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -217,8 +217,8 @@ out:
static const BlockJobDriver commit_job_driver = {
.job_driver = {
.instance_size = sizeof(CommitBlockJob),
+ .job_type = JOB_TYPE_COMMIT,
},
- .job_type = JOB_TYPE_COMMIT,
.start = commit_run,
};
diff --git a/block/mirror.c b/block/mirror.c
index ed72656a23..35fcc1f0b7 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -988,8 +988,8 @@ static void mirror_drain(BlockJob *job)
static const BlockJobDriver mirror_job_driver = {
.job_driver = {
.instance_size = sizeof(MirrorBlockJob),
+ .job_type = JOB_TYPE_MIRROR,
},
- .job_type = JOB_TYPE_MIRROR,
.start = mirror_run,
.complete = mirror_complete,
.pause = mirror_pause,
@@ -1000,8 +1000,8 @@ static const BlockJobDriver mirror_job_driver = {
static const BlockJobDriver commit_active_job_driver = {
.job_driver = {
.instance_size = sizeof(MirrorBlockJob),
+ .job_type = JOB_TYPE_COMMIT,
},
- .job_type = JOB_TYPE_COMMIT,
.start = mirror_run,
.complete = mirror_complete,
.pause = mirror_pause,
diff --git a/block/stream.c b/block/stream.c
index 048bceb5d0..cb723f190a 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -211,8 +211,8 @@ out:
static const BlockJobDriver stream_job_driver = {
.job_driver = {
.instance_size = sizeof(StreamBlockJob),
+ .job_type = JOB_TYPE_STREAM,
},
- .job_type = JOB_TYPE_STREAM,
.start = stream_run,
};
diff --git a/blockjob.c b/blockjob.c
index 2a3844721e..ea71ec0129 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -309,9 +309,7 @@ static void block_job_detach_aio_context(void *opaque)
static char *child_job_get_parent_desc(BdrvChild *c)
{
BlockJob *job = c->opaque;
- return g_strdup_printf("%s job '%s'",
- JobType_str(job->driver->job_type),
- job->job.id);
+ return g_strdup_printf("%s job '%s'", job_type_str(&job->job), job->job.id);
}
static void child_job_drained_begin(BdrvChild *c)
@@ -847,7 +845,7 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
return NULL;
}
info = g_new0(BlockJobInfo, 1);
- info->type = g_strdup(JobType_str(job->driver->job_type));
+ info->type = g_strdup(job_type_str(&job->job));
info->device = g_strdup(job->job.id);
info->len = job->len;
info->busy = atomic_read(&job->busy);
@@ -878,7 +876,7 @@ static void block_job_event_cancelled(BlockJob *job)
return;
}
- qapi_event_send_block_job_cancelled(job->driver->job_type,
+ qapi_event_send_block_job_cancelled(job_type(&job->job),
job->job.id,
job->len,
job->offset,
@@ -892,7 +890,7 @@ static void block_job_event_completed(BlockJob *job, const char *msg)
return;
}
- qapi_event_send_block_job_completed(job->driver->job_type,
+ qapi_event_send_block_job_completed(job_type(&job->job),
job->job.id,
job->len,
job->offset,
@@ -906,7 +904,7 @@ static int block_job_event_pending(BlockJob *job)
{
block_job_state_transition(job, BLOCK_JOB_STATUS_PENDING);
if (!job->auto_finalize && !block_job_is_internal(job)) {
- qapi_event_send_block_job_pending(job->driver->job_type,
+ qapi_event_send_block_job_pending(job_type(&job->job),
job->job.id,
&error_abort);
}
@@ -980,7 +978,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
block_job_sleep_timer_cb, job);
error_setg(&job->blocker, "block device is in use by block job: %s",
- JobType_str(driver->job_type));
+ job_type_str(&job->job));
block_job_add_bdrv(job, "main node", bs, 0, BLK_PERM_ALL, &error_abort);
bs->job = job;
@@ -1184,7 +1182,7 @@ void block_job_event_ready(BlockJob *job)
return;
}
- qapi_event_send_block_job_ready(job->driver->job_type,
+ qapi_event_send_block_job_ready(job_type(&job->job),
job->job.id,
job->len,
job->offset,
diff --git a/job.c b/job.c
index 87fd48468c..83724a43de 100644
--- a/job.c
+++ b/job.c
@@ -29,6 +29,16 @@
#include "qemu/job.h"
#include "qemu/id.h"
+JobType job_type(const Job *job)
+{
+ return job->driver->job_type;
+}
+
+const char *job_type_str(const Job *job)
+{
+ return JobType_str(job_type(job));
+}
+
void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
{
Job *job;
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 06/40] job: Add job_delete()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (4 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 05/40] job: Add JobDriver.job_type Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 07/40] job: Maintain a list of all jobs Kevin Wolf
` (36 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves freeing the Job object and its fields from block_job_unref()
to job_delete().
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
include/qemu/job.h | 3 +++
blockjob.c | 3 +--
job.c | 6 ++++++
3 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 279ce688fd..43dc2e4a7d 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -62,6 +62,9 @@ struct JobDriver {
*/
void *job_create(const char *job_id, const JobDriver *driver, Error **errp);
+/** Frees the @job object. */
+void job_delete(Job *job);
+
/** Returns the JobType of a given Job. */
JobType job_type(const Job *job);
diff --git a/blockjob.c b/blockjob.c
index ea71ec0129..430a67ba63 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -261,9 +261,8 @@ void block_job_unref(BlockJob *job)
block_job_detach_aio_context, job);
blk_unref(job->blk);
error_free(job->blocker);
- g_free(job->job.id);
assert(!timer_pending(&job->sleep_timer));
- g_free(job);
+ job_delete(&job->job);
}
}
diff --git a/job.c b/job.c
index 83724a43de..cfdd008c52 100644
--- a/job.c
+++ b/job.c
@@ -56,3 +56,9 @@ void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
return job;
}
+
+void job_delete(Job *job)
+{
+ g_free(job->id);
+ g_free(job);
+}
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 07/40] job: Maintain a list of all jobs
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (5 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 06/40] job: Add job_delete() Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 08/40] job: Move state transitions to Job Kevin Wolf
` (35 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves the job list from BlockJob to Job. Now we can check for
duplicate IDs in job_create().
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
include/block/blockjob.h | 3 ---
include/qemu/job.h | 19 +++++++++++++++++++
blockjob.c | 46 ++++++++++++++++++++++++----------------------
job.c | 31 +++++++++++++++++++++++++++++++
4 files changed, 74 insertions(+), 25 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 640e649034..10bd9f7059 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -105,9 +105,6 @@ typedef struct BlockJob {
*/
bool deferred_to_main_loop;
- /** Element of the list of block jobs */
- QLIST_ENTRY(BlockJob) job_list;
-
/** Status that is published by the query-block-jobs QMP API */
BlockDeviceIoStatus iostatus;
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 43dc2e4a7d..bae2b0920c 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -27,6 +27,7 @@
#define JOB_H
#include "qapi/qapi-types-block-core.h"
+#include "qemu/queue.h"
typedef struct JobDriver JobDriver;
@@ -39,6 +40,9 @@ typedef struct Job {
/** The type of this job. */
const JobDriver *driver;
+
+ /** Element of the list of jobs */
+ QLIST_ENTRY(Job) job_list;
} Job;
/**
@@ -71,4 +75,19 @@ JobType job_type(const Job *job);
/** Returns the enum string for the JobType of a given Job. */
const char *job_type_str(const Job *job);
+/**
+ * Get the next element from the list of block jobs after @job, or the
+ * first one if @job is %NULL.
+ *
+ * Returns the requested job, or %NULL if there are no more jobs left.
+ */
+Job *job_next(Job *job);
+
+/**
+ * Get the job identified by @id (which must not be %NULL).
+ *
+ * Returns the requested job, or %NULL if it doesn't exist.
+ */
+Job *job_get(const char *id);
+
#endif
diff --git a/blockjob.c b/blockjob.c
index 430a67ba63..c69b2e7cf5 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -129,8 +129,6 @@ struct BlockJobTxn {
int refcnt;
};
-static QLIST_HEAD(, BlockJob) block_jobs = QLIST_HEAD_INITIALIZER(block_jobs);
-
/*
* The block job API is composed of two categories of functions.
*
@@ -146,25 +144,34 @@ static QLIST_HEAD(, BlockJob) block_jobs = QLIST_HEAD_INITIALIZER(block_jobs);
* blockjob_int.h.
*/
-BlockJob *block_job_next(BlockJob *job)
+static bool is_block_job(Job *job)
{
- if (!job) {
- return QLIST_FIRST(&block_jobs);
- }
- return QLIST_NEXT(job, job_list);
+ return job_type(job) == JOB_TYPE_BACKUP ||
+ job_type(job) == JOB_TYPE_COMMIT ||
+ job_type(job) == JOB_TYPE_MIRROR ||
+ job_type(job) == JOB_TYPE_STREAM;
+}
+
+BlockJob *block_job_next(BlockJob *bjob)
+{
+ Job *job = bjob ? &bjob->job : NULL;
+
+ do {
+ job = job_next(job);
+ } while (job && !is_block_job(job));
+
+ return job ? container_of(job, BlockJob, job) : NULL;
}
BlockJob *block_job_get(const char *id)
{
- BlockJob *job;
+ Job *job = job_get(id);
- QLIST_FOREACH(job, &block_jobs, job_list) {
- if (job->job.id && !strcmp(id, job->job.id)) {
- return job;
- }
+ if (job && is_block_job(job)) {
+ return container_of(job, BlockJob, job);
+ } else {
+ return NULL;
}
-
- return NULL;
}
BlockJobTxn *block_job_txn_new(void)
@@ -253,7 +260,6 @@ void block_job_unref(BlockJob *job)
assert(job->status == BLOCK_JOB_STATUS_NULL);
assert(!job->txn);
BlockDriverState *bs = blk_bs(job->blk);
- QLIST_REMOVE(job, job_list);
bs->job = NULL;
block_job_remove_all_bdrv(job);
blk_remove_aio_context_notifier(job->blk,
@@ -812,7 +818,7 @@ void block_job_cancel_sync_all(void)
BlockJob *job;
AioContext *aio_context;
- while ((job = QLIST_FIRST(&block_jobs))) {
+ while ((job = block_job_next(NULL))) {
aio_context = blk_get_aio_context(job->blk);
aio_context_acquire(aio_context);
block_job_cancel_sync(job);
@@ -942,10 +948,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
error_setg(errp, "Cannot specify job ID for internal block job");
return NULL;
}
- if (block_job_get(job_id)) {
- error_setg(errp, "Job ID '%s' already in use", job_id);
- return NULL;
- }
}
blk = blk_new(perm, shared_perm);
@@ -961,6 +963,8 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
return NULL;
}
+ assert(is_block_job(&job->job));
+
job->driver = driver;
job->blk = blk;
job->cb = cb;
@@ -983,8 +987,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
bdrv_op_unblock(bs, BLOCK_OP_TYPE_DATAPLANE, job->blocker);
- QLIST_INSERT_HEAD(&block_jobs, job, job_list);
-
blk_add_aio_context_notifier(blk, block_job_attached_aio_context,
block_job_detach_aio_context, job);
diff --git a/job.c b/job.c
index cfdd008c52..e57303c6bb 100644
--- a/job.c
+++ b/job.c
@@ -29,6 +29,8 @@
#include "qemu/job.h"
#include "qemu/id.h"
+static QLIST_HEAD(, Job) jobs = QLIST_HEAD_INITIALIZER(jobs);
+
JobType job_type(const Job *job)
{
return job->driver->job_type;
@@ -39,6 +41,27 @@ const char *job_type_str(const Job *job)
return JobType_str(job_type(job));
}
+Job *job_next(Job *job)
+{
+ if (!job) {
+ return QLIST_FIRST(&jobs);
+ }
+ return QLIST_NEXT(job, job_list);
+}
+
+Job *job_get(const char *id)
+{
+ Job *job;
+
+ QLIST_FOREACH(job, &jobs, job_list) {
+ if (job->id && !strcmp(id, job->id)) {
+ return job;
+ }
+ }
+
+ return NULL;
+}
+
void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
{
Job *job;
@@ -48,17 +71,25 @@ void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
error_setg(errp, "Invalid job ID '%s'", job_id);
return NULL;
}
+ if (job_get(job_id)) {
+ error_setg(errp, "Job ID '%s' already in use", job_id);
+ return NULL;
+ }
}
job = g_malloc0(driver->instance_size);
job->driver = driver;
job->id = g_strdup(job_id);
+ QLIST_INSERT_HEAD(&jobs, job, job_list);
+
return job;
}
void job_delete(Job *job)
{
+ QLIST_REMOVE(job, job_list);
+
g_free(job->id);
g_free(job);
}
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 08/40] job: Move state transitions to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (6 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 07/40] job: Maintain a list of all jobs Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 14:36 ` Eric Blake
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 09/40] job: Add reference counting Kevin Wolf
` (34 subsequent siblings)
42 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves BlockJob.status and the closely related functions
(block_)job_state_transition() and (block_)job_apply_verb to Job. The
two QAPI enums are renamed to JobStatus and JobVerb.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
qapi/block-core.json | 16 ++++----
include/block/blockjob.h | 3 --
include/qemu/job.h | 13 ++++++
blockjob.c | 102 +++++++++++------------------------------------
job.c | 56 ++++++++++++++++++++++++++
tests/test-blockjob.c | 39 +++++++++---------
block/trace-events | 2 -
trace-events | 4 ++
8 files changed, 123 insertions(+), 112 deletions(-)
diff --git a/qapi/block-core.json b/qapi/block-core.json
index 63c6011411..bb964b4319 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -1068,9 +1068,9 @@
'data': ['commit', 'stream', 'mirror', 'backup'] }
##
-# @BlockJobVerb:
+# @JobVerb:
#
-# Represents command verbs that can be applied to a blockjob.
+# Represents command verbs that can be applied to a job.
#
# @cancel: see @block-job-cancel
#
@@ -1088,14 +1088,14 @@
#
# Since: 2.12
##
-{ 'enum': 'BlockJobVerb',
+{ 'enum': 'JobVerb',
'data': ['cancel', 'pause', 'resume', 'set-speed', 'complete', 'dismiss',
'finalize' ] }
##
-# @BlockJobStatus:
+# @JobStatus:
#
-# Indicates the present state of a given blockjob in its lifetime.
+# Indicates the present state of a given job in its lifetime.
#
# @undefined: Erroneous, default state. Should not ever be visible.
#
@@ -1134,7 +1134,7 @@
#
# Since: 2.12
##
-{ 'enum': 'BlockJobStatus',
+{ 'enum': 'JobStatus',
'data': ['undefined', 'created', 'running', 'paused', 'ready', 'standby',
'waiting', 'pending', 'aborting', 'concluded', 'null' ] }
@@ -1184,7 +1184,7 @@
'data': {'type': 'str', 'device': 'str', 'len': 'int',
'offset': 'int', 'busy': 'bool', 'paused': 'bool', 'speed': 'int',
'io-status': 'BlockDeviceIoStatus', 'ready': 'bool',
- 'status': 'BlockJobStatus',
+ 'status': 'JobStatus',
'auto-finalize': 'bool', 'auto-dismiss': 'bool',
'*error': 'str' } }
@@ -2416,7 +2416,7 @@
# QEMU 2.12+ job lifetime management semantics.
#
# This command will refuse to operate on any job that has not yet reached
-# its terminal state, BLOCK_JOB_STATUS_CONCLUDED. For jobs that make use of
+# its terminal state, JOB_STATUS_CONCLUDED. For jobs that make use of the
# BLOCK_JOB_READY event, block-job-cancel or block-job-complete will still need
# to be used as appropriate.
#
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 10bd9f7059..01cdee6d15 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -147,9 +147,6 @@ typedef struct BlockJob {
*/
QEMUTimer sleep_timer;
- /** Current state; See @BlockJobStatus for details. */
- BlockJobStatus status;
-
/** True if this job should automatically finalize itself */
bool auto_finalize;
diff --git a/include/qemu/job.h b/include/qemu/job.h
index bae2b0920c..0b78778b9e 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -41,6 +41,9 @@ typedef struct Job {
/** The type of this job. */
const JobDriver *driver;
+ /** Current state; See @JobStatus for details. */
+ JobStatus status;
+
/** Element of the list of jobs */
QLIST_ENTRY(Job) job_list;
} Job;
@@ -90,4 +93,14 @@ Job *job_next(Job *job);
*/
Job *job_get(const char *id);
+/**
+ * Check whether the verb @verb can be applied to @job in its current state.
+ * Returns 0 if the verb can be applied; otherwise errp is set and -EPERM
+ * returned.
+ */
+int job_apply_verb(Job *job, JobVerb verb, Error **errp);
+
+/* TODO To be removed from the public interface */
+void job_state_transition(Job *job, JobStatus s1);
+
#endif
diff --git a/blockjob.c b/blockjob.c
index c69b2e7cf5..0fba01edd4 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -41,61 +41,6 @@
* block_job_enter. */
static QemuMutex block_job_mutex;
-/* BlockJob State Transition Table */
-bool BlockJobSTT[BLOCK_JOB_STATUS__MAX][BLOCK_JOB_STATUS__MAX] = {
- /* U, C, R, P, Y, S, W, D, X, E, N */
- /* U: */ [BLOCK_JOB_STATUS_UNDEFINED] = {0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0},
- /* C: */ [BLOCK_JOB_STATUS_CREATED] = {0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1},
- /* R: */ [BLOCK_JOB_STATUS_RUNNING] = {0, 0, 0, 1, 1, 0, 1, 0, 1, 0, 0},
- /* P: */ [BLOCK_JOB_STATUS_PAUSED] = {0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0},
- /* Y: */ [BLOCK_JOB_STATUS_READY] = {0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0},
- /* S: */ [BLOCK_JOB_STATUS_STANDBY] = {0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0},
- /* W: */ [BLOCK_JOB_STATUS_WAITING] = {0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0},
- /* D: */ [BLOCK_JOB_STATUS_PENDING] = {0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0},
- /* X: */ [BLOCK_JOB_STATUS_ABORTING] = {0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0},
- /* E: */ [BLOCK_JOB_STATUS_CONCLUDED] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1},
- /* N: */ [BLOCK_JOB_STATUS_NULL] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
-};
-
-bool BlockJobVerbTable[BLOCK_JOB_VERB__MAX][BLOCK_JOB_STATUS__MAX] = {
- /* U, C, R, P, Y, S, W, D, X, E, N */
- [BLOCK_JOB_VERB_CANCEL] = {0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0},
- [BLOCK_JOB_VERB_PAUSE] = {0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0},
- [BLOCK_JOB_VERB_RESUME] = {0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0},
- [BLOCK_JOB_VERB_SET_SPEED] = {0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0},
- [BLOCK_JOB_VERB_COMPLETE] = {0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0},
- [BLOCK_JOB_VERB_FINALIZE] = {0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0},
- [BLOCK_JOB_VERB_DISMISS] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0},
-};
-
-static void block_job_state_transition(BlockJob *job, BlockJobStatus s1)
-{
- BlockJobStatus s0 = job->status;
- assert(s1 >= 0 && s1 <= BLOCK_JOB_STATUS__MAX);
- trace_block_job_state_transition(job, job->ret, BlockJobSTT[s0][s1] ?
- "allowed" : "disallowed",
- BlockJobStatus_str(s0),
- BlockJobStatus_str(s1));
- assert(BlockJobSTT[s0][s1]);
- job->status = s1;
-}
-
-static int block_job_apply_verb(BlockJob *job, BlockJobVerb bv, Error **errp)
-{
- assert(bv >= 0 && bv <= BLOCK_JOB_VERB__MAX);
- trace_block_job_apply_verb(job, BlockJobStatus_str(job->status),
- BlockJobVerb_str(bv),
- BlockJobVerbTable[bv][job->status] ?
- "allowed" : "prohibited");
- if (BlockJobVerbTable[bv][job->status]) {
- return 0;
- }
- error_setg(errp, "Job '%s' in state '%s' cannot accept command verb '%s'",
- job->job.id, BlockJobStatus_str(job->status),
- BlockJobVerb_str(bv));
- return -EPERM;
-}
-
static void block_job_lock(void)
{
qemu_mutex_lock(&block_job_mutex);
@@ -257,7 +202,7 @@ static void block_job_detach_aio_context(void *opaque);
void block_job_unref(BlockJob *job)
{
if (--job->refcnt == 0) {
- assert(job->status == BLOCK_JOB_STATUS_NULL);
+ assert(job->job.status == JOB_STATUS_NULL);
assert(!job->txn);
BlockDriverState *bs = blk_bs(job->blk);
bs->job = NULL;
@@ -409,7 +354,7 @@ void block_job_start(BlockJob *job)
job->pause_count--;
job->busy = true;
job->paused = false;
- block_job_state_transition(job, BLOCK_JOB_STATUS_RUNNING);
+ job_state_transition(&job->job, JOB_STATUS_RUNNING);
bdrv_coroutine_enter(blk_bs(job->blk), job->co);
}
@@ -421,7 +366,7 @@ static void block_job_decommission(BlockJob *job)
job->paused = false;
job->deferred_to_main_loop = true;
block_job_txn_del_job(job);
- block_job_state_transition(job, BLOCK_JOB_STATUS_NULL);
+ job_state_transition(&job->job, JOB_STATUS_NULL);
block_job_unref(job);
}
@@ -432,7 +377,7 @@ static void block_job_do_dismiss(BlockJob *job)
static void block_job_conclude(BlockJob *job)
{
- block_job_state_transition(job, BLOCK_JOB_STATUS_CONCLUDED);
+ job_state_transition(&job->job, JOB_STATUS_CONCLUDED);
if (job->auto_dismiss || !block_job_started(job)) {
block_job_do_dismiss(job);
}
@@ -444,7 +389,7 @@ static void block_job_update_rc(BlockJob *job)
job->ret = -ECANCELED;
}
if (job->ret) {
- block_job_state_transition(job, BLOCK_JOB_STATUS_ABORTING);
+ job_state_transition(&job->job, JOB_STATUS_ABORTING);
}
}
@@ -652,7 +597,7 @@ static void block_job_completed_txn_success(BlockJob *job)
BlockJobTxn *txn = job->txn;
BlockJob *other_job;
- block_job_state_transition(job, BLOCK_JOB_STATUS_WAITING);
+ job_state_transition(&job->job, JOB_STATUS_WAITING);
/*
* Successful completion, see if there are other running jobs in this
@@ -677,7 +622,7 @@ void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp)
{
int64_t old_speed = job->speed;
- if (block_job_apply_verb(job, BLOCK_JOB_VERB_SET_SPEED, errp)) {
+ if (job_apply_verb(&job->job, JOB_VERB_SET_SPEED, errp)) {
return;
}
if (speed < 0) {
@@ -709,7 +654,7 @@ void block_job_complete(BlockJob *job, Error **errp)
{
/* Should not be reachable via external interface for internal jobs */
assert(job->job.id);
- if (block_job_apply_verb(job, BLOCK_JOB_VERB_COMPLETE, errp)) {
+ if (job_apply_verb(&job->job, JOB_VERB_COMPLETE, errp)) {
return;
}
if (job->pause_count || job->cancelled || !job->driver->complete) {
@@ -724,7 +669,7 @@ void block_job_complete(BlockJob *job, Error **errp)
void block_job_finalize(BlockJob *job, Error **errp)
{
assert(job && job->job.id);
- if (block_job_apply_verb(job, BLOCK_JOB_VERB_FINALIZE, errp)) {
+ if (job_apply_verb(&job->job, JOB_VERB_FINALIZE, errp)) {
return;
}
block_job_do_finalize(job);
@@ -735,7 +680,7 @@ void block_job_dismiss(BlockJob **jobptr, Error **errp)
BlockJob *job = *jobptr;
/* similarly to _complete, this is QMP-interface only. */
assert(job->job.id);
- if (block_job_apply_verb(job, BLOCK_JOB_VERB_DISMISS, errp)) {
+ if (job_apply_verb(&job->job, JOB_VERB_DISMISS, errp)) {
return;
}
@@ -745,7 +690,7 @@ void block_job_dismiss(BlockJob **jobptr, Error **errp)
void block_job_user_pause(BlockJob *job, Error **errp)
{
- if (block_job_apply_verb(job, BLOCK_JOB_VERB_PAUSE, errp)) {
+ if (job_apply_verb(&job->job, JOB_VERB_PAUSE, errp)) {
return;
}
if (job->user_paused) {
@@ -768,7 +713,7 @@ void block_job_user_resume(BlockJob *job, Error **errp)
error_setg(errp, "Can't resume a job that was not paused");
return;
}
- if (block_job_apply_verb(job, BLOCK_JOB_VERB_RESUME, errp)) {
+ if (job_apply_verb(&job->job, JOB_VERB_RESUME, errp)) {
return;
}
block_job_iostatus_reset(job);
@@ -778,7 +723,7 @@ void block_job_user_resume(BlockJob *job, Error **errp)
void block_job_cancel(BlockJob *job, bool force)
{
- if (job->status == BLOCK_JOB_STATUS_CONCLUDED) {
+ if (job->job.status == JOB_STATUS_CONCLUDED) {
block_job_do_dismiss(job);
return;
}
@@ -794,7 +739,7 @@ void block_job_cancel(BlockJob *job, bool force)
void block_job_user_cancel(BlockJob *job, bool force, Error **errp)
{
- if (block_job_apply_verb(job, BLOCK_JOB_VERB_CANCEL, errp)) {
+ if (job_apply_verb(&job->job, JOB_VERB_CANCEL, errp)) {
return;
}
block_job_cancel(job, force);
@@ -859,7 +804,7 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
info->speed = job->speed;
info->io_status = job->iostatus;
info->ready = job->ready;
- info->status = job->status;
+ info->status = job->job.status;
info->auto_finalize = job->auto_finalize;
info->auto_dismiss = job->auto_dismiss;
info->has_error = job->ret != 0;
@@ -907,7 +852,7 @@ static void block_job_event_completed(BlockJob *job, const char *msg)
static int block_job_event_pending(BlockJob *job)
{
- block_job_state_transition(job, BLOCK_JOB_STATUS_PENDING);
+ job_state_transition(&job->job, JOB_STATUS_PENDING);
if (!job->auto_finalize && !block_job_is_internal(job)) {
qapi_event_send_block_job_pending(job_type(&job->job),
job->job.id,
@@ -975,7 +920,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
job->refcnt = 1;
job->auto_finalize = !(flags & BLOCK_JOB_MANUAL_FINALIZE);
job->auto_dismiss = !(flags & BLOCK_JOB_MANUAL_DISMISS);
- block_job_state_transition(job, BLOCK_JOB_STATUS_CREATED);
aio_timer_init(qemu_get_aio_context(), &job->sleep_timer,
QEMU_CLOCK_REALTIME, SCALE_NS,
block_job_sleep_timer_cb, job);
@@ -1017,7 +961,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
void block_job_early_fail(BlockJob *job)
{
- assert(job->status == BLOCK_JOB_STATUS_CREATED);
+ assert(job->job.status == JOB_STATUS_CREATED);
block_job_decommission(job);
}
@@ -1077,14 +1021,14 @@ void coroutine_fn block_job_pause_point(BlockJob *job)
}
if (block_job_should_pause(job) && !block_job_is_cancelled(job)) {
- BlockJobStatus status = job->status;
- block_job_state_transition(job, status == BLOCK_JOB_STATUS_READY ? \
- BLOCK_JOB_STATUS_STANDBY : \
- BLOCK_JOB_STATUS_PAUSED);
+ JobStatus status = job->job.status;
+ job_state_transition(&job->job, status == JOB_STATUS_READY
+ ? JOB_STATUS_STANDBY
+ : JOB_STATUS_PAUSED);
job->paused = true;
block_job_do_yield(job, -1);
job->paused = false;
- block_job_state_transition(job, status);
+ job_state_transition(&job->job, status);
}
if (job->driver->resume) {
@@ -1176,7 +1120,7 @@ void block_job_iostatus_reset(BlockJob *job)
void block_job_event_ready(BlockJob *job)
{
- block_job_state_transition(job, BLOCK_JOB_STATUS_READY);
+ job_state_transition(&job->job, JOB_STATUS_READY);
job->ready = true;
if (block_job_is_internal(job)) {
diff --git a/job.c b/job.c
index e57303c6bb..4960b30838 100644
--- a/job.c
+++ b/job.c
@@ -28,9 +28,63 @@
#include "qapi/error.h"
#include "qemu/job.h"
#include "qemu/id.h"
+#include "trace-root.h"
static QLIST_HEAD(, Job) jobs = QLIST_HEAD_INITIALIZER(jobs);
+/* Job State Transition Table */
+bool JobSTT[JOB_STATUS__MAX][JOB_STATUS__MAX] = {
+ /* U, C, R, P, Y, S, W, D, X, E, N */
+ /* U: */ [JOB_STATUS_UNDEFINED] = {0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0},
+ /* C: */ [JOB_STATUS_CREATED] = {0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1},
+ /* R: */ [JOB_STATUS_RUNNING] = {0, 0, 0, 1, 1, 0, 1, 0, 1, 0, 0},
+ /* P: */ [JOB_STATUS_PAUSED] = {0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0},
+ /* Y: */ [JOB_STATUS_READY] = {0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0},
+ /* S: */ [JOB_STATUS_STANDBY] = {0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0},
+ /* W: */ [JOB_STATUS_WAITING] = {0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0},
+ /* D: */ [JOB_STATUS_PENDING] = {0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0},
+ /* X: */ [JOB_STATUS_ABORTING] = {0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0},
+ /* E: */ [JOB_STATUS_CONCLUDED] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1},
+ /* N: */ [JOB_STATUS_NULL] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
+};
+
+bool JobVerbTable[JOB_VERB__MAX][JOB_STATUS__MAX] = {
+ /* U, C, R, P, Y, S, W, D, X, E, N */
+ [JOB_VERB_CANCEL] = {0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0},
+ [JOB_VERB_PAUSE] = {0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0},
+ [JOB_VERB_RESUME] = {0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0},
+ [JOB_VERB_SET_SPEED] = {0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0},
+ [JOB_VERB_COMPLETE] = {0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0},
+ [JOB_VERB_FINALIZE] = {0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0},
+ [JOB_VERB_DISMISS] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0},
+};
+
+/* TODO Make static once the whole state machine is in job.c */
+void job_state_transition(Job *job, JobStatus s1)
+{
+ JobStatus s0 = job->status;
+ assert(s1 >= 0 && s1 <= JOB_STATUS__MAX);
+ trace_job_state_transition(job, /* TODO re-enable: job->ret */ 0,
+ JobSTT[s0][s1] ? "allowed" : "disallowed",
+ JobStatus_str(s0), JobStatus_str(s1));
+ assert(JobSTT[s0][s1]);
+ job->status = s1;
+}
+
+int job_apply_verb(Job *job, JobVerb verb, Error **errp)
+{
+ assert(verb >= 0 && verb <= JOB_VERB__MAX);
+ trace_job_apply_verb(job, JobStatus_str(job->status), JobVerb_str(verb),
+ JobVerbTable[verb][job->status] ?
+ "allowed" : "prohibited");
+ if (JobVerbTable[verb][job->status]) {
+ return 0;
+ }
+ error_setg(errp, "Job '%s' in state '%s' cannot accept command verb '%s'",
+ job->id, JobStatus_str(job->status), JobVerb_str(verb));
+ return -EPERM;
+}
+
JobType job_type(const Job *job)
{
return job->driver->job_type;
@@ -81,6 +135,8 @@ void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
job->driver = driver;
job->id = g_strdup(job_id);
+ job_state_transition(job, JOB_STATUS_CREATED);
+
QLIST_INSERT_HEAD(&jobs, job, job_list);
return job;
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index b82026180a..6ccd585dee 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -211,7 +211,7 @@ static CancelJob *create_common(BlockJob **pjob)
job = mk_job(blk, "Steve", &test_cancel_driver, true,
BLOCK_JOB_MANUAL_FINALIZE | BLOCK_JOB_MANUAL_DISMISS);
block_job_ref(job);
- assert(job->status == BLOCK_JOB_STATUS_CREATED);
+ assert(job->job.status == JOB_STATUS_CREATED);
s = container_of(job, CancelJob, common);
s->blk = blk;
@@ -223,15 +223,14 @@ static void cancel_common(CancelJob *s)
{
BlockJob *job = &s->common;
BlockBackend *blk = s->blk;
- BlockJobStatus sts = job->status;
+ JobStatus sts = job->job.status;
block_job_cancel_sync(job);
- if ((sts != BLOCK_JOB_STATUS_CREATED) &&
- (sts != BLOCK_JOB_STATUS_CONCLUDED)) {
+ if (sts != JOB_STATUS_CREATED && sts != JOB_STATUS_CONCLUDED) {
BlockJob *dummy = job;
block_job_dismiss(&dummy, &error_abort);
}
- assert(job->status == BLOCK_JOB_STATUS_NULL);
+ assert(job->job.status == JOB_STATUS_NULL);
block_job_unref(job);
destroy_blk(blk);
}
@@ -253,7 +252,7 @@ static void test_cancel_running(void)
s = create_common(&job);
block_job_start(job);
- assert(job->status == BLOCK_JOB_STATUS_RUNNING);
+ assert(job->job.status == JOB_STATUS_RUNNING);
cancel_common(s);
}
@@ -266,11 +265,11 @@ static void test_cancel_paused(void)
s = create_common(&job);
block_job_start(job);
- assert(job->status == BLOCK_JOB_STATUS_RUNNING);
+ assert(job->job.status == JOB_STATUS_RUNNING);
block_job_user_pause(job, &error_abort);
block_job_enter(job);
- assert(job->status == BLOCK_JOB_STATUS_PAUSED);
+ assert(job->job.status == JOB_STATUS_PAUSED);
cancel_common(s);
}
@@ -283,11 +282,11 @@ static void test_cancel_ready(void)
s = create_common(&job);
block_job_start(job);
- assert(job->status == BLOCK_JOB_STATUS_RUNNING);
+ assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
block_job_enter(job);
- assert(job->status == BLOCK_JOB_STATUS_READY);
+ assert(job->job.status == JOB_STATUS_READY);
cancel_common(s);
}
@@ -300,15 +299,15 @@ static void test_cancel_standby(void)
s = create_common(&job);
block_job_start(job);
- assert(job->status == BLOCK_JOB_STATUS_RUNNING);
+ assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
block_job_enter(job);
- assert(job->status == BLOCK_JOB_STATUS_READY);
+ assert(job->job.status == JOB_STATUS_READY);
block_job_user_pause(job, &error_abort);
block_job_enter(job);
- assert(job->status == BLOCK_JOB_STATUS_STANDBY);
+ assert(job->job.status == JOB_STATUS_STANDBY);
cancel_common(s);
}
@@ -321,18 +320,18 @@ static void test_cancel_pending(void)
s = create_common(&job);
block_job_start(job);
- assert(job->status == BLOCK_JOB_STATUS_RUNNING);
+ assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
block_job_enter(job);
- assert(job->status == BLOCK_JOB_STATUS_READY);
+ assert(job->job.status == JOB_STATUS_READY);
block_job_complete(job, &error_abort);
block_job_enter(job);
while (!s->completed) {
aio_poll(qemu_get_aio_context(), true);
}
- assert(job->status == BLOCK_JOB_STATUS_PENDING);
+ assert(job->job.status == JOB_STATUS_PENDING);
cancel_common(s);
}
@@ -345,21 +344,21 @@ static void test_cancel_concluded(void)
s = create_common(&job);
block_job_start(job);
- assert(job->status == BLOCK_JOB_STATUS_RUNNING);
+ assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
block_job_enter(job);
- assert(job->status == BLOCK_JOB_STATUS_READY);
+ assert(job->job.status == JOB_STATUS_READY);
block_job_complete(job, &error_abort);
block_job_enter(job);
while (!s->completed) {
aio_poll(qemu_get_aio_context(), true);
}
- assert(job->status == BLOCK_JOB_STATUS_PENDING);
+ assert(job->job.status == JOB_STATUS_PENDING);
block_job_finalize(job, &error_abort);
- assert(job->status == BLOCK_JOB_STATUS_CONCLUDED);
+ assert(job->job.status == JOB_STATUS_CONCLUDED);
cancel_common(s);
}
diff --git a/block/trace-events b/block/trace-events
index f8c50b4063..93b927908a 100644
--- a/block/trace-events
+++ b/block/trace-events
@@ -6,8 +6,6 @@ bdrv_lock_medium(void *bs, bool locked) "bs %p locked %d"
# blockjob.c
block_job_completed(void *job, int ret, int jret) "job %p ret %d corrected ret %d"
-block_job_state_transition(void *job, int ret, const char *legal, const char *s0, const char *s1) "job %p (ret: %d) attempting %s transition (%s-->%s)"
-block_job_apply_verb(void *job, const char *state, const char *verb, const char *legal) "job %p in state %s; applying verb %s (%s)"
# block/block-backend.c
blk_co_preadv(void *blk, void *bs, int64_t offset, unsigned int bytes, int flags) "blk %p bs %p offset %"PRId64" bytes %u flags 0x%x"
diff --git a/trace-events b/trace-events
index ed71f44649..2507e1327d 100644
--- a/trace-events
+++ b/trace-events
@@ -104,6 +104,10 @@ gdbstub_err_invalid_rle(void) "got invalid RLE sequence"
gdbstub_err_checksum_invalid(uint8_t ch) "got invalid command checksum digit: 0x%02x"
gdbstub_err_checksum_incorrect(uint8_t expected, uint8_t got) "got command packet with incorrect checksum, expected=0x%02x, received=0x%02x"
+# job.c
+job_state_transition(void *job, int ret, const char *legal, const char *s0, const char *s1) "job %p (ret: %d) attempting %s transition (%s-->%s)"
+job_apply_verb(void *job, const char *state, const char *verb, const char *legal) "job %p in state %s; applying verb %s (%s)"
+
### Guest events, keep at bottom
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 08/40] job: Move state transitions to Job
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 08/40] job: Move state transitions to Job Kevin Wolf
@ 2018-05-18 14:36 ` Eric Blake
0 siblings, 0 replies; 86+ messages in thread
From: Eric Blake @ 2018-05-18 14:36 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:20 AM, Kevin Wolf wrote:
> This moves BlockJob.status and the closely related functions
> (block_)job_state_transition() and (block_)job_apply_verb to Job. The
> two QAPI enums are renamed to JobStatus and JobVerb.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> Reviewed-by: Max Reitz <mreitz@redhat.com>
> Reviewed-by: John Snow <jsnow@redhat.com>
> ---
> @@ -1077,14 +1021,14 @@ void coroutine_fn block_job_pause_point(BlockJob *job)
> }
>
> if (block_job_should_pause(job) && !block_job_is_cancelled(job)) {
> - BlockJobStatus status = job->status;
> - block_job_state_transition(job, status == BLOCK_JOB_STATUS_READY ? \
> - BLOCK_JOB_STATUS_STANDBY : \
> - BLOCK_JOB_STATUS_PAUSED);
> + JobStatus status = job->job.status;
> + job_state_transition(&job->job, status == JOB_STATUS_READY
> + ? JOB_STATUS_STANDBY
> + : JOB_STATUS_PAUSED);
Nice that you are getting rid of the pointless \. Here, you split the
?: operator by wrapping the ? onto the start of the next line...
> +++ b/job.c
> +int job_apply_verb(Job *job, JobVerb verb, Error **errp)
> +{
> + assert(verb >= 0 && verb <= JOB_VERB__MAX);
> + trace_job_apply_verb(job, JobStatus_str(job->status), JobVerb_str(verb),
> + JobVerbTable[verb][job->status] ?
> + "allowed" : "prohibited");
while here, you put it at the end of the previous line. We have a
slight preference for one style over the other (there are also false
positives in both of these greps):
$ git grep '?$' | wc -c
27251
$ git grep '^ *?' | wc -c
9619
But I don't care which style you use, other than pointing out that a
consistent style within the patch doesn't hurt. :)
So, whether or not you make a whitespace-only tweak to one of those two
spots,
Reviewed-by: Eric Blake <eblake@redhat.com>
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 09/40] job: Add reference counting
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (7 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 08/40] job: Move state transitions to Job Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 10/40] job: Move cancelled to Job Kevin Wolf
` (33 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves reference counting from BlockJob to Job.
In order to keep calling the BlockJob cleanup code when the job is
deleted via job_unref(), introduce a new JobDriver.free callback. Every
block job must use block_job_free() for this callback, this is asserted
in block_job_create().
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
include/block/blockjob.h | 21 -------------------
include/block/blockjob_int.h | 7 +++++++
include/qemu/job.h | 19 ++++++++++++++++--
block/backup.c | 1 +
block/commit.c | 1 +
block/mirror.c | 2 ++
block/stream.c | 1 +
blockjob.c | 48 +++++++++++++++++++-------------------------
job.c | 22 ++++++++++++++++----
qemu-img.c | 4 ++--
tests/test-bdrv-drain.c | 1 +
tests/test-blockjob-txn.c | 1 +
tests/test-blockjob.c | 6 ++++--
13 files changed, 76 insertions(+), 58 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 01cdee6d15..087e7820f5 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -132,9 +132,6 @@ typedef struct BlockJob {
/** The opaque value that is passed to the completion function. */
void *opaque;
- /** Reference count of the block job */
- int refcnt;
-
/** True when job has reported completion by calling block_job_completed. */
bool completed;
@@ -400,24 +397,6 @@ void block_job_iostatus_reset(BlockJob *job);
BlockJobTxn *block_job_txn_new(void);
/**
- * block_job_ref:
- *
- * Add a reference to BlockJob refcnt, it will be decreased with
- * block_job_unref, and then be freed if it comes to be the last
- * reference.
- */
-void block_job_ref(BlockJob *job);
-
-/**
- * block_job_unref:
- *
- * Release a reference that was previously acquired with block_job_ref
- * or block_job_create. If it's the last reference to the object, it will be
- * freed.
- */
-void block_job_unref(BlockJob *job);
-
-/**
* block_job_txn_unref:
*
* Release a reference that was previously acquired with block_job_txn_add_job
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 1e62d6dd30..6f0fe3c48d 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -144,6 +144,13 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
BlockCompletionFunc *cb, void *opaque, Error **errp);
/**
+ * block_job_free:
+ * Callback to be used for JobDriver.free in all block jobs. Frees block job
+ * specific resources in @job.
+ */
+void block_job_free(Job *job);
+
+/**
* block_job_sleep_ns:
* @job: The job that calls the function.
* @ns: How many nanoseconds to stop for.
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 0b78778b9e..0751e2afab 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -41,6 +41,9 @@ typedef struct Job {
/** The type of this job. */
const JobDriver *driver;
+ /** Reference count of the block job */
+ int refcnt;
+
/** Current state; See @JobStatus for details. */
JobStatus status;
@@ -57,6 +60,9 @@ struct JobDriver {
/** Enum describing the operation */
JobType job_type;
+
+ /** Called when the job is freed */
+ void (*free)(Job *job);
};
@@ -69,8 +75,17 @@ struct JobDriver {
*/
void *job_create(const char *job_id, const JobDriver *driver, Error **errp);
-/** Frees the @job object. */
-void job_delete(Job *job);
+/**
+ * Add a reference to Job refcnt, it will be decreased with job_unref, and then
+ * be freed if it comes to be the last reference.
+ */
+void job_ref(Job *job);
+
+/**
+ * Release a reference that was previously acquired with job_ref() or
+ * job_create(). If it's the last reference to the object, it will be freed.
+ */
+void job_unref(Job *job);
/** Returns the JobType of a given Job. */
JobType job_type(const Job *job);
diff --git a/block/backup.c b/block/backup.c
index baf8d432da..cfdb89d977 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -526,6 +526,7 @@ static const BlockJobDriver backup_job_driver = {
.job_driver = {
.instance_size = sizeof(BackupBlockJob),
.job_type = JOB_TYPE_BACKUP,
+ .free = block_job_free,
},
.start = backup_run,
.commit = backup_commit,
diff --git a/block/commit.c b/block/commit.c
index 32d29c890e..925c96abe7 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -218,6 +218,7 @@ static const BlockJobDriver commit_job_driver = {
.job_driver = {
.instance_size = sizeof(CommitBlockJob),
.job_type = JOB_TYPE_COMMIT,
+ .free = block_job_free,
},
.start = commit_run,
};
diff --git a/block/mirror.c b/block/mirror.c
index 35fcc1f0b7..0df4f709fd 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -989,6 +989,7 @@ static const BlockJobDriver mirror_job_driver = {
.job_driver = {
.instance_size = sizeof(MirrorBlockJob),
.job_type = JOB_TYPE_MIRROR,
+ .free = block_job_free,
},
.start = mirror_run,
.complete = mirror_complete,
@@ -1001,6 +1002,7 @@ static const BlockJobDriver commit_active_job_driver = {
.job_driver = {
.instance_size = sizeof(MirrorBlockJob),
.job_type = JOB_TYPE_COMMIT,
+ .free = block_job_free,
},
.start = mirror_run,
.complete = mirror_complete,
diff --git a/block/stream.c b/block/stream.c
index cb723f190a..7273d2213c 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -212,6 +212,7 @@ static const BlockJobDriver stream_job_driver = {
.job_driver = {
.instance_size = sizeof(StreamBlockJob),
.job_type = JOB_TYPE_STREAM,
+ .free = block_job_free,
},
.start = stream_run,
};
diff --git a/blockjob.c b/blockjob.c
index 0fba01edd4..0bf0a26af0 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -190,31 +190,25 @@ static void block_job_resume(BlockJob *job)
block_job_enter_cond(job, block_job_timer_not_pending);
}
-void block_job_ref(BlockJob *job)
-{
- ++job->refcnt;
-}
-
static void block_job_attached_aio_context(AioContext *new_context,
void *opaque);
static void block_job_detach_aio_context(void *opaque);
-void block_job_unref(BlockJob *job)
+void block_job_free(Job *job)
{
- if (--job->refcnt == 0) {
- assert(job->job.status == JOB_STATUS_NULL);
- assert(!job->txn);
- BlockDriverState *bs = blk_bs(job->blk);
- bs->job = NULL;
- block_job_remove_all_bdrv(job);
- blk_remove_aio_context_notifier(job->blk,
- block_job_attached_aio_context,
- block_job_detach_aio_context, job);
- blk_unref(job->blk);
- error_free(job->blocker);
- assert(!timer_pending(&job->sleep_timer));
- job_delete(&job->job);
- }
+ BlockJob *bjob = container_of(job, BlockJob, job);
+ BlockDriverState *bs = blk_bs(bjob->blk);
+
+ assert(!bjob->txn);
+
+ bs->job = NULL;
+ block_job_remove_all_bdrv(bjob);
+ blk_remove_aio_context_notifier(bjob->blk,
+ block_job_attached_aio_context,
+ block_job_detach_aio_context, bjob);
+ blk_unref(bjob->blk);
+ error_free(bjob->blocker);
+ assert(!timer_pending(&bjob->sleep_timer));
}
static void block_job_attached_aio_context(AioContext *new_context,
@@ -245,7 +239,7 @@ static void block_job_detach_aio_context(void *opaque)
BlockJob *job = opaque;
/* In case the job terminates during aio_poll()... */
- block_job_ref(job);
+ job_ref(&job->job);
block_job_pause(job);
@@ -253,7 +247,7 @@ static void block_job_detach_aio_context(void *opaque)
block_job_drain(job);
}
- block_job_unref(job);
+ job_unref(&job->job);
}
static char *child_job_get_parent_desc(BdrvChild *c)
@@ -367,7 +361,7 @@ static void block_job_decommission(BlockJob *job)
job->deferred_to_main_loop = true;
block_job_txn_del_job(job);
job_state_transition(&job->job, JOB_STATUS_NULL);
- block_job_unref(job);
+ job_unref(&job->job);
}
static void block_job_do_dismiss(BlockJob *job)
@@ -506,14 +500,14 @@ static int block_job_finish_sync(BlockJob *job,
assert(blk_bs(job->blk)->job == job);
- block_job_ref(job);
+ job_ref(&job->job);
if (finish) {
finish(job, &local_err);
}
if (local_err) {
error_propagate(errp, local_err);
- block_job_unref(job);
+ job_unref(&job->job);
return -EBUSY;
}
/* block_job_drain calls block_job_enter, and it should be enough to
@@ -526,7 +520,7 @@ static int block_job_finish_sync(BlockJob *job,
aio_poll(qemu_get_aio_context(), true);
}
ret = (job->cancelled && job->ret == 0) ? -ECANCELED : job->ret;
- block_job_unref(job);
+ job_unref(&job->job);
return ret;
}
@@ -909,6 +903,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
}
assert(is_block_job(&job->job));
+ assert(job->job.driver->free == &block_job_free);
job->driver = driver;
job->blk = blk;
@@ -917,7 +912,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
job->busy = false;
job->paused = true;
job->pause_count = 1;
- job->refcnt = 1;
job->auto_finalize = !(flags & BLOCK_JOB_MANUAL_FINALIZE);
job->auto_dismiss = !(flags & BLOCK_JOB_MANUAL_DISMISS);
aio_timer_init(qemu_get_aio_context(), &job->sleep_timer,
diff --git a/job.c b/job.c
index 4960b30838..e588fbe53d 100644
--- a/job.c
+++ b/job.c
@@ -134,6 +134,7 @@ void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
job = g_malloc0(driver->instance_size);
job->driver = driver;
job->id = g_strdup(job_id);
+ job->refcnt = 1;
job_state_transition(job, JOB_STATUS_CREATED);
@@ -142,10 +143,23 @@ void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
return job;
}
-void job_delete(Job *job)
+void job_ref(Job *job)
{
- QLIST_REMOVE(job, job_list);
+ ++job->refcnt;
+}
+
+void job_unref(Job *job)
+{
+ if (--job->refcnt == 0) {
+ assert(job->status == JOB_STATUS_NULL);
- g_free(job->id);
- g_free(job);
+ if (job->driver->free) {
+ job->driver->free(job);
+ }
+
+ QLIST_REMOVE(job, job_list);
+
+ g_free(job->id);
+ g_free(job);
+ }
}
diff --git a/qemu-img.c b/qemu-img.c
index 60e45ec103..27bbcc5100 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -862,7 +862,7 @@ static void run_block_job(BlockJob *job, Error **errp)
int ret = 0;
aio_context_acquire(aio_context);
- block_job_ref(job);
+ job_ref(&job->job);
do {
aio_poll(aio_context, true);
qemu_progress_print(job->len ?
@@ -874,7 +874,7 @@ static void run_block_job(BlockJob *job, Error **errp)
} else {
ret = job->ret;
}
- block_job_unref(job);
+ job_unref(&job->job);
aio_context_release(aio_context);
/* publish completion progress only when success */
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index fe9f412b39..f9e37d479c 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -522,6 +522,7 @@ static void test_job_complete(BlockJob *job, Error **errp)
BlockJobDriver test_job_driver = {
.job_driver = {
.instance_size = sizeof(TestBlockJob),
+ .free = block_job_free,
},
.start = test_job_start,
.complete = test_job_complete,
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 48b12d1744..b49b28ca27 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -76,6 +76,7 @@ static void test_block_job_cb(void *opaque, int ret)
static const BlockJobDriver test_block_job_driver = {
.job_driver = {
.instance_size = sizeof(TestBlockJob),
+ .free = block_job_free,
},
.start = test_block_job_run,
};
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 6ccd585dee..e24fc3f140 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -19,6 +19,7 @@
static const BlockJobDriver test_block_job_driver = {
.job_driver = {
.instance_size = sizeof(BlockJob),
+ .free = block_job_free,
},
};
@@ -196,6 +197,7 @@ static void coroutine_fn cancel_job_start(void *opaque)
static const BlockJobDriver test_cancel_driver = {
.job_driver = {
.instance_size = sizeof(CancelJob),
+ .free = block_job_free,
},
.start = cancel_job_start,
.complete = cancel_job_complete,
@@ -210,7 +212,7 @@ static CancelJob *create_common(BlockJob **pjob)
blk = create_blk(NULL);
job = mk_job(blk, "Steve", &test_cancel_driver, true,
BLOCK_JOB_MANUAL_FINALIZE | BLOCK_JOB_MANUAL_DISMISS);
- block_job_ref(job);
+ job_ref(&job->job);
assert(job->job.status == JOB_STATUS_CREATED);
s = container_of(job, CancelJob, common);
s->blk = blk;
@@ -231,7 +233,7 @@ static void cancel_common(CancelJob *s)
block_job_dismiss(&dummy, &error_abort);
}
assert(job->job.status == JOB_STATUS_NULL);
- block_job_unref(job);
+ job_unref(&job->job);
destroy_blk(blk);
}
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 10/40] job: Move cancelled to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (8 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 09/40] job: Add reference counting Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 11/40] job: Add Job.aio_context Kevin Wolf
` (32 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
We cannot yet move the whole logic around job cancelling to Job because
it depends on quite a few other things that are still only in BlockJob,
but we can move the cancelled field at least.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
include/block/blockjob.h | 8 --------
include/block/blockjob_int.h | 8 --------
include/qemu/job.h | 11 +++++++++++
block/backup.c | 6 +++---
block/commit.c | 4 ++--
block/mirror.c | 20 ++++++++++----------
block/stream.c | 4 ++--
blockjob.c | 28 +++++++++++++---------------
job.c | 5 +++++
tests/test-blockjob-txn.c | 6 +++---
tests/test-blockjob.c | 2 +-
11 files changed, 50 insertions(+), 52 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 087e7820f5..1e708f468a 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -57,14 +57,6 @@ typedef struct BlockJob {
Coroutine *co;
/**
- * Set to true if the job should cancel itself. The flag must
- * always be tested just before toggling the busy flag from false
- * to true. After a job has been cancelled, it should only yield
- * if #aio_poll will ("sooner or later") reenter the coroutine.
- */
- bool cancelled;
-
- /**
* Set to true if the job should abort immediately without waiting
* for data to be in sync.
*/
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 6f0fe3c48d..d64f30e6b0 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -196,14 +196,6 @@ void block_job_early_fail(BlockJob *job);
void block_job_completed(BlockJob *job, int ret);
/**
- * block_job_is_cancelled:
- * @job: The job being queried.
- *
- * Returns whether the job is scheduled for cancellation.
- */
-bool block_job_is_cancelled(BlockJob *job);
-
-/**
* block_job_pause_point:
* @job: The job that is ready to pause.
*
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 0751e2afab..5dfbec5d69 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -47,6 +47,14 @@ typedef struct Job {
/** Current state; See @JobStatus for details. */
JobStatus status;
+ /**
+ * Set to true if the job should cancel itself. The flag must
+ * always be tested just before toggling the busy flag from false
+ * to true. After a job has been cancelled, it should only yield
+ * if #aio_poll will ("sooner or later") reenter the coroutine.
+ */
+ bool cancelled;
+
/** Element of the list of jobs */
QLIST_ENTRY(Job) job_list;
} Job;
@@ -93,6 +101,9 @@ JobType job_type(const Job *job);
/** Returns the enum string for the JobType of a given Job. */
const char *job_type_str(const Job *job);
+/** Returns whether the job is scheduled for cancellation. */
+bool job_is_cancelled(Job *job);
+
/**
* Get the next element from the list of block jobs after @job, or the
* first one if @job is %NULL.
diff --git a/block/backup.c b/block/backup.c
index cfdb89d977..ef0aa0e24e 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -329,7 +329,7 @@ static bool coroutine_fn yield_and_check(BackupBlockJob *job)
{
uint64_t delay_ns;
- if (block_job_is_cancelled(&job->common)) {
+ if (job_is_cancelled(&job->common.job)) {
return true;
}
@@ -339,7 +339,7 @@ static bool coroutine_fn yield_and_check(BackupBlockJob *job)
job->bytes_read = 0;
block_job_sleep_ns(&job->common, delay_ns);
- if (block_job_is_cancelled(&job->common)) {
+ if (job_is_cancelled(&job->common.job)) {
return true;
}
@@ -441,7 +441,7 @@ static void coroutine_fn backup_run(void *opaque)
if (job->sync_mode == MIRROR_SYNC_MODE_NONE) {
/* All bits are set in copy_bitmap to allow any cluster to be copied.
* This does not actually require them to be copied. */
- while (!block_job_is_cancelled(&job->common)) {
+ while (!job_is_cancelled(&job->common.job)) {
/* Yield until the job is cancelled. We just let our before_write
* notify callback service CoW requests. */
block_job_yield(&job->common);
diff --git a/block/commit.c b/block/commit.c
index 925c96abe7..85baea8f92 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -90,7 +90,7 @@ static void commit_complete(BlockJob *job, void *opaque)
* the normal backing chain can be restored. */
blk_unref(s->base);
- if (!block_job_is_cancelled(&s->common) && ret == 0) {
+ if (!job_is_cancelled(&s->common.job) && ret == 0) {
/* success */
ret = bdrv_drop_intermediate(s->commit_top_bs, base,
s->backing_file_str);
@@ -172,7 +172,7 @@ static void coroutine_fn commit_run(void *opaque)
* with no pending I/O here so that bdrv_drain_all() returns.
*/
block_job_sleep_ns(&s->common, delay_ns);
- if (block_job_is_cancelled(&s->common)) {
+ if (job_is_cancelled(&s->common.job)) {
break;
}
/* Copy if allocated above the base */
diff --git a/block/mirror.c b/block/mirror.c
index 0df4f709fd..424072ed00 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -622,7 +622,7 @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
mirror_throttle(s);
- if (block_job_is_cancelled(&s->common)) {
+ if (job_is_cancelled(&s->common.job)) {
s->initial_zeroing_ongoing = false;
return 0;
}
@@ -650,7 +650,7 @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s)
mirror_throttle(s);
- if (block_job_is_cancelled(&s->common)) {
+ if (job_is_cancelled(&s->common.job)) {
return 0;
}
@@ -695,7 +695,7 @@ static void coroutine_fn mirror_run(void *opaque)
checking for a NULL string */
int ret = 0;
- if (block_job_is_cancelled(&s->common)) {
+ if (job_is_cancelled(&s->common.job)) {
goto immediate_exit;
}
@@ -729,10 +729,10 @@ static void coroutine_fn mirror_run(void *opaque)
/* Report BLOCK_JOB_READY and wait for complete. */
block_job_event_ready(&s->common);
s->synced = true;
- while (!block_job_is_cancelled(&s->common) && !s->should_complete) {
+ while (!job_is_cancelled(&s->common.job) && !s->should_complete) {
block_job_yield(&s->common);
}
- s->common.cancelled = false;
+ s->common.job.cancelled = false;
goto immediate_exit;
}
@@ -768,7 +768,7 @@ static void coroutine_fn mirror_run(void *opaque)
s->last_pause_ns = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
if (!s->is_none_mode) {
ret = mirror_dirty_init(s);
- if (ret < 0 || block_job_is_cancelled(&s->common)) {
+ if (ret < 0 || job_is_cancelled(&s->common.job)) {
goto immediate_exit;
}
}
@@ -828,7 +828,7 @@ static void coroutine_fn mirror_run(void *opaque)
}
should_complete = s->should_complete ||
- block_job_is_cancelled(&s->common);
+ job_is_cancelled(&s->common.job);
cnt = bdrv_get_dirty_count(s->dirty_bitmap);
}
@@ -856,7 +856,7 @@ static void coroutine_fn mirror_run(void *opaque)
* completion.
*/
assert(QLIST_EMPTY(&bs->tracked_requests));
- s->common.cancelled = false;
+ s->common.job.cancelled = false;
need_drain = false;
break;
}
@@ -869,7 +869,7 @@ static void coroutine_fn mirror_run(void *opaque)
}
trace_mirror_before_sleep(s, cnt, s->synced, delay_ns);
block_job_sleep_ns(&s->common, delay_ns);
- if (block_job_is_cancelled(&s->common) &&
+ if (job_is_cancelled(&s->common.job) &&
(!s->synced || s->common.force))
{
break;
@@ -884,7 +884,7 @@ immediate_exit:
* the target is a copy of the source.
*/
assert(ret < 0 || ((s->common.force || !s->synced) &&
- block_job_is_cancelled(&s->common)));
+ job_is_cancelled(&s->common.job)));
assert(need_drain);
mirror_wait_for_all_io(s);
}
diff --git a/block/stream.c b/block/stream.c
index 7273d2213c..22c71ae100 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -66,7 +66,7 @@ static void stream_complete(BlockJob *job, void *opaque)
BlockDriverState *base = s->base;
Error *local_err = NULL;
- if (!block_job_is_cancelled(&s->common) && bs->backing &&
+ if (!job_is_cancelled(&s->common.job) && bs->backing &&
data->ret == 0) {
const char *base_id = NULL, *base_fmt = NULL;
if (base) {
@@ -141,7 +141,7 @@ static void coroutine_fn stream_run(void *opaque)
* with no pending I/O here so that bdrv_drain_all() returns.
*/
block_job_sleep_ns(&s->common, delay_ns);
- if (block_job_is_cancelled(&s->common)) {
+ if (job_is_cancelled(&s->common.job)) {
break;
}
diff --git a/blockjob.c b/blockjob.c
index 0bf0a26af0..f4f9956678 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -379,7 +379,7 @@ static void block_job_conclude(BlockJob *job)
static void block_job_update_rc(BlockJob *job)
{
- if (!job->ret && block_job_is_cancelled(job)) {
+ if (!job->ret && job_is_cancelled(&job->job)) {
job->ret = -ECANCELED;
}
if (job->ret) {
@@ -438,7 +438,7 @@ static int block_job_finalize_single(BlockJob *job)
/* Emit events only if we actually started */
if (block_job_started(job)) {
- if (block_job_is_cancelled(job)) {
+ if (job_is_cancelled(&job->job)) {
block_job_event_cancelled(job);
} else {
const char *msg = NULL;
@@ -464,7 +464,7 @@ static void block_job_cancel_async(BlockJob *job, bool force)
job->user_paused = false;
job->pause_count--;
}
- job->cancelled = true;
+ job->job.cancelled = true;
/* To prevent 'force == false' overriding a previous 'force == true' */
job->force |= force;
}
@@ -519,7 +519,8 @@ static int block_job_finish_sync(BlockJob *job,
while (!job->completed) {
aio_poll(qemu_get_aio_context(), true);
}
- ret = (job->cancelled && job->ret == 0) ? -ECANCELED : job->ret;
+ ret = (job_is_cancelled(&job->job) && job->ret == 0)
+ ? -ECANCELED : job->ret;
job_unref(&job->job);
return ret;
}
@@ -557,7 +558,7 @@ static void block_job_completed_txn_abort(BlockJob *job)
other_job = QLIST_FIRST(&txn->jobs);
ctx = blk_get_aio_context(other_job->blk);
if (!other_job->completed) {
- assert(other_job->cancelled);
+ assert(job_is_cancelled(&other_job->job));
block_job_finish_sync(other_job, NULL, NULL);
}
block_job_finalize_single(other_job);
@@ -651,7 +652,9 @@ void block_job_complete(BlockJob *job, Error **errp)
if (job_apply_verb(&job->job, JOB_VERB_COMPLETE, errp)) {
return;
}
- if (job->pause_count || job->cancelled || !job->driver->complete) {
+ if (job->pause_count || job_is_cancelled(&job->job) ||
+ !job->driver->complete)
+ {
error_setg(errp, "The active block job '%s' cannot be completed",
job->job.id);
return;
@@ -1006,7 +1009,7 @@ void coroutine_fn block_job_pause_point(BlockJob *job)
if (!block_job_should_pause(job)) {
return;
}
- if (block_job_is_cancelled(job)) {
+ if (job_is_cancelled(&job->job)) {
return;
}
@@ -1014,7 +1017,7 @@ void coroutine_fn block_job_pause_point(BlockJob *job)
job->driver->pause(job);
}
- if (block_job_should_pause(job) && !block_job_is_cancelled(job)) {
+ if (block_job_should_pause(job) && !job_is_cancelled(&job->job)) {
JobStatus status = job->job.status;
job_state_transition(&job->job, status == JOB_STATUS_READY
? JOB_STATUS_STANDBY
@@ -1066,17 +1069,12 @@ void block_job_enter(BlockJob *job)
block_job_enter_cond(job, NULL);
}
-bool block_job_is_cancelled(BlockJob *job)
-{
- return job->cancelled;
-}
-
void block_job_sleep_ns(BlockJob *job, int64_t ns)
{
assert(job->busy);
/* Check cancellation *before* setting busy = false, too! */
- if (block_job_is_cancelled(job)) {
+ if (job_is_cancelled(&job->job)) {
return;
}
@@ -1092,7 +1090,7 @@ void block_job_yield(BlockJob *job)
assert(job->busy);
/* Check cancellation *before* setting busy = false, too! */
- if (block_job_is_cancelled(job)) {
+ if (job_is_cancelled(&job->job)) {
return;
}
diff --git a/job.c b/job.c
index e588fbe53d..92729cffa4 100644
--- a/job.c
+++ b/job.c
@@ -95,6 +95,11 @@ const char *job_type_str(const Job *job)
return JobType_str(job_type(job));
}
+bool job_is_cancelled(Job *job)
+{
+ return job->cancelled;
+}
+
Job *job_next(Job *job)
{
if (!job) {
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index b49b28ca27..26b4bbb230 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -29,7 +29,7 @@ static void test_block_job_complete(BlockJob *job, void *opaque)
BlockDriverState *bs = blk_bs(job->blk);
int rc = (intptr_t)opaque;
- if (block_job_is_cancelled(job)) {
+ if (job_is_cancelled(&job->job)) {
rc = -ECANCELED;
}
@@ -49,7 +49,7 @@ static void coroutine_fn test_block_job_run(void *opaque)
block_job_yield(job);
}
- if (block_job_is_cancelled(job)) {
+ if (job_is_cancelled(&job->job)) {
break;
}
}
@@ -66,7 +66,7 @@ typedef struct {
static void test_block_job_cb(void *opaque, int ret)
{
TestBlockJobCBData *data = opaque;
- if (!ret && block_job_is_cancelled(&data->job->common)) {
+ if (!ret && job_is_cancelled(&data->job->common.job)) {
ret = -ECANCELED;
}
*data->result = ret;
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index e24fc3f140..fa31481537 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -179,7 +179,7 @@ static void coroutine_fn cancel_job_start(void *opaque)
CancelJob *s = opaque;
while (!s->should_complete) {
- if (block_job_is_cancelled(&s->common)) {
+ if (job_is_cancelled(&s->common.job)) {
goto defer;
}
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 11/40] job: Add Job.aio_context
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (9 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 10/40] job: Move cancelled to Job Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 12/40] job: Move defer_to_main_loop to Job Kevin Wolf
` (31 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
When block jobs need an AioContext, they just take it from their main
block node. Generic jobs don't have a main block node, so we need to
assign them an AioContext explicitly.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
include/qemu/job.h | 7 ++++++-
blockjob.c | 5 ++++-
job.c | 4 +++-
3 files changed, 13 insertions(+), 3 deletions(-)
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 5dfbec5d69..01e083f97a 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -47,6 +47,9 @@ typedef struct Job {
/** Current state; See @JobStatus for details. */
JobStatus status;
+ /** AioContext to run the job coroutine in */
+ AioContext *aio_context;
+
/**
* Set to true if the job should cancel itself. The flag must
* always be tested just before toggling the busy flag from false
@@ -79,9 +82,11 @@ struct JobDriver {
*
* @job_id: The id of the newly-created job, or %NULL for internal jobs
* @driver: The class object for the newly-created job.
+ * @ctx: The AioContext to run the job coroutine in.
* @errp: Error object.
*/
-void *job_create(const char *job_id, const JobDriver *driver, Error **errp);
+void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
+ Error **errp);
/**
* Add a reference to Job refcnt, it will be decreased with job_unref, and then
diff --git a/blockjob.c b/blockjob.c
index f4f9956678..0a0b1c41dd 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -216,6 +216,7 @@ static void block_job_attached_aio_context(AioContext *new_context,
{
BlockJob *job = opaque;
+ job->job.aio_context = new_context;
if (job->driver->attached_aio_context) {
job->driver->attached_aio_context(job, new_context);
}
@@ -247,6 +248,7 @@ static void block_job_detach_aio_context(void *opaque)
block_job_drain(job);
}
+ job->job.aio_context = NULL;
job_unref(&job->job);
}
@@ -899,7 +901,8 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
return NULL;
}
- job = job_create(job_id, &driver->job_driver, errp);
+ job = job_create(job_id, &driver->job_driver, blk_get_aio_context(blk),
+ errp);
if (job == NULL) {
blk_unref(blk);
return NULL;
diff --git a/job.c b/job.c
index 92729cffa4..7ff1f66273 100644
--- a/job.c
+++ b/job.c
@@ -121,7 +121,8 @@ Job *job_get(const char *id)
return NULL;
}
-void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
+void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
+ Error **errp)
{
Job *job;
@@ -140,6 +141,7 @@ void *job_create(const char *job_id, const JobDriver *driver, Error **errp)
job->driver = driver;
job->id = g_strdup(job_id);
job->refcnt = 1;
+ job->aio_context = ctx;
job_state_transition(job, JOB_STATUS_CREATED);
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 12/40] job: Move defer_to_main_loop to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (10 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 11/40] job: Add Job.aio_context Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 17:56 ` John Snow
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 13/40] job: Move coroutine and related code " Kevin Wolf
` (30 subsequent siblings)
42 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
Move the defer_to_main_loop functionality from BlockJob to Job.
The code can be simplified because we can use job->aio_context in
job_defer_to_main_loop_bh() now, instead of having to access the
BlockDriverState.
Probably taking the data->aio_context lock in addition was already
unnecessary in the old code because we didn't actually make use of
anything protected by the old AioContext except getting the new
AioContext, in case it changed between scheduling the BH and running it.
But it's certainly unnecessary now that the BDS isn't accessed at all
any more.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 5 ----
include/block/blockjob_int.h | 19 ---------------
include/qemu/job.h | 20 ++++++++++++++++
block/backup.c | 7 +++---
block/commit.c | 11 +++++----
block/mirror.c | 15 ++++++------
block/stream.c | 14 +++++------
blockjob.c | 57 ++++----------------------------------------
job.c | 32 +++++++++++++++++++++++++
tests/test-bdrv-drain.c | 7 +++---
tests/test-blockjob-txn.c | 13 +++++-----
tests/test-blockjob.c | 7 +++---
12 files changed, 97 insertions(+), 110 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 1e708f468a..2a9e865e31 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -92,11 +92,6 @@ typedef struct BlockJob {
*/
bool ready;
- /**
- * Set to true when the job has deferred work to the main loop.
- */
- bool deferred_to_main_loop;
-
/** Status that is published by the query-block-jobs QMP API */
BlockDeviceIoStatus iostatus;
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index d64f30e6b0..0c2f8de381 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -233,23 +233,4 @@ void block_job_event_ready(BlockJob *job);
BlockErrorAction block_job_error_action(BlockJob *job, BlockdevOnError on_err,
int is_read, int error);
-typedef void BlockJobDeferToMainLoopFn(BlockJob *job, void *opaque);
-
-/**
- * block_job_defer_to_main_loop:
- * @job: The job
- * @fn: The function to run in the main loop
- * @opaque: The opaque value that is passed to @fn
- *
- * This function must be called by the main job coroutine just before it
- * returns. @fn is executed in the main loop with the BlockDriverState
- * AioContext acquired. Block jobs must call bdrv_unref(), bdrv_close(), and
- * anything that uses bdrv_drain_all() in the main loop.
- *
- * The @job AioContext is held while @fn executes.
- */
-void block_job_defer_to_main_loop(BlockJob *job,
- BlockJobDeferToMainLoopFn *fn,
- void *opaque);
-
#endif
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 01e083f97a..933e0ab328 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -58,6 +58,9 @@ typedef struct Job {
*/
bool cancelled;
+ /** Set to true when the job has deferred work to the main loop. */
+ bool deferred_to_main_loop;
+
/** Element of the list of jobs */
QLIST_ENTRY(Job) job_list;
} Job;
@@ -131,6 +134,23 @@ Job *job_get(const char *id);
*/
int job_apply_verb(Job *job, JobVerb verb, Error **errp);
+typedef void JobDeferToMainLoopFn(Job *job, void *opaque);
+
+/**
+ * @job: The job
+ * @fn: The function to run in the main loop
+ * @opaque: The opaque value that is passed to @fn
+ *
+ * This function must be called by the main job coroutine just before it
+ * returns. @fn is executed in the main loop with the job AioContext acquired.
+ *
+ * Block jobs must call bdrv_unref(), bdrv_close(), and anything that uses
+ * bdrv_drain_all() in the main loop.
+ *
+ * The @job AioContext is held while @fn executes.
+ */
+void job_defer_to_main_loop(Job *job, JobDeferToMainLoopFn *fn, void *opaque);
+
/* TODO To be removed from the public interface */
void job_state_transition(Job *job, JobStatus s1);
diff --git a/block/backup.c b/block/backup.c
index ef0aa0e24e..22dd368c90 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -317,11 +317,12 @@ typedef struct {
int ret;
} BackupCompleteData;
-static void backup_complete(BlockJob *job, void *opaque)
+static void backup_complete(Job *job, void *opaque)
{
+ BlockJob *bjob = container_of(job, BlockJob, job);
BackupCompleteData *data = opaque;
- block_job_completed(job, data->ret);
+ block_job_completed(bjob, data->ret);
g_free(data);
}
@@ -519,7 +520,7 @@ static void coroutine_fn backup_run(void *opaque)
data = g_malloc(sizeof(*data));
data->ret = ret;
- block_job_defer_to_main_loop(&job->common, backup_complete, data);
+ job_defer_to_main_loop(&job->common.job, backup_complete, data);
}
static const BlockJobDriver backup_job_driver = {
diff --git a/block/commit.c b/block/commit.c
index 85baea8f92..d326766e4d 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -72,9 +72,10 @@ typedef struct {
int ret;
} CommitCompleteData;
-static void commit_complete(BlockJob *job, void *opaque)
+static void commit_complete(Job *job, void *opaque)
{
- CommitBlockJob *s = container_of(job, CommitBlockJob, common);
+ CommitBlockJob *s = container_of(job, CommitBlockJob, common.job);
+ BlockJob *bjob = &s->common;
CommitCompleteData *data = opaque;
BlockDriverState *top = blk_bs(s->top);
BlockDriverState *base = blk_bs(s->base);
@@ -90,7 +91,7 @@ static void commit_complete(BlockJob *job, void *opaque)
* the normal backing chain can be restored. */
blk_unref(s->base);
- if (!job_is_cancelled(&s->common.job) && ret == 0) {
+ if (!job_is_cancelled(job) && ret == 0) {
/* success */
ret = bdrv_drop_intermediate(s->commit_top_bs, base,
s->backing_file_str);
@@ -114,7 +115,7 @@ static void commit_complete(BlockJob *job, void *opaque)
* block_job_finish_sync()), block_job_completed() won't free it and
* therefore the blockers on the intermediate nodes remain. This would
* cause bdrv_set_backing_hd() to fail. */
- block_job_remove_all_bdrv(job);
+ block_job_remove_all_bdrv(bjob);
block_job_completed(&s->common, ret);
g_free(data);
@@ -211,7 +212,7 @@ out:
data = g_malloc(sizeof(*data));
data->ret = ret;
- block_job_defer_to_main_loop(&s->common, commit_complete, data);
+ job_defer_to_main_loop(&s->common.job, commit_complete, data);
}
static const BlockJobDriver commit_job_driver = {
diff --git a/block/mirror.c b/block/mirror.c
index 424072ed00..90d4ac9cb6 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -484,9 +484,10 @@ typedef struct {
int ret;
} MirrorExitData;
-static void mirror_exit(BlockJob *job, void *opaque)
+static void mirror_exit(Job *job, void *opaque)
{
- MirrorBlockJob *s = container_of(job, MirrorBlockJob, common);
+ MirrorBlockJob *s = container_of(job, MirrorBlockJob, common.job);
+ BlockJob *bjob = &s->common;
MirrorExitData *data = opaque;
AioContext *replace_aio_context = NULL;
BlockDriverState *src = s->source;
@@ -568,7 +569,7 @@ static void mirror_exit(BlockJob *job, void *opaque)
* the blockers on the intermediate nodes so that the resulting state is
* valid. Also give up permissions on mirror_top_bs->backing, which might
* block the removal. */
- block_job_remove_all_bdrv(job);
+ block_job_remove_all_bdrv(bjob);
bdrv_child_try_set_perm(mirror_top_bs->backing, 0, BLK_PERM_ALL,
&error_abort);
bdrv_replace_node(mirror_top_bs, backing_bs(mirror_top_bs), &error_abort);
@@ -576,9 +577,9 @@ static void mirror_exit(BlockJob *job, void *opaque)
/* We just changed the BDS the job BB refers to (with either or both of the
* bdrv_replace_node() calls), so switch the BB back so the cleanup does
* the right thing. We don't need any permissions any more now. */
- blk_remove_bs(job->blk);
- blk_set_perm(job->blk, 0, BLK_PERM_ALL, &error_abort);
- blk_insert_bs(job->blk, mirror_top_bs, &error_abort);
+ blk_remove_bs(bjob->blk);
+ blk_set_perm(bjob->blk, 0, BLK_PERM_ALL, &error_abort);
+ blk_insert_bs(bjob->blk, mirror_top_bs, &error_abort);
block_job_completed(&s->common, data->ret);
@@ -901,7 +902,7 @@ immediate_exit:
if (need_drain) {
bdrv_drained_begin(bs);
}
- block_job_defer_to_main_loop(&s->common, mirror_exit, data);
+ job_defer_to_main_loop(&s->common.job, mirror_exit, data);
}
static void mirror_complete(BlockJob *job, Error **errp)
diff --git a/block/stream.c b/block/stream.c
index 22c71ae100..0bba81678c 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -58,16 +58,16 @@ typedef struct {
int ret;
} StreamCompleteData;
-static void stream_complete(BlockJob *job, void *opaque)
+static void stream_complete(Job *job, void *opaque)
{
- StreamBlockJob *s = container_of(job, StreamBlockJob, common);
+ StreamBlockJob *s = container_of(job, StreamBlockJob, common.job);
+ BlockJob *bjob = &s->common;
StreamCompleteData *data = opaque;
- BlockDriverState *bs = blk_bs(job->blk);
+ BlockDriverState *bs = blk_bs(bjob->blk);
BlockDriverState *base = s->base;
Error *local_err = NULL;
- if (!job_is_cancelled(&s->common.job) && bs->backing &&
- data->ret == 0) {
+ if (!job_is_cancelled(job) && bs->backing && data->ret == 0) {
const char *base_id = NULL, *base_fmt = NULL;
if (base) {
base_id = s->backing_file_str;
@@ -88,7 +88,7 @@ out:
/* Reopen the image back in read-only mode if necessary */
if (s->bs_flags != bdrv_get_flags(bs)) {
/* Give up write permissions before making it read-only */
- blk_set_perm(job->blk, 0, BLK_PERM_ALL, &error_abort);
+ blk_set_perm(bjob->blk, 0, BLK_PERM_ALL, &error_abort);
bdrv_reopen(bs, s->bs_flags, NULL);
}
@@ -205,7 +205,7 @@ out:
/* Modify backing chain and close BDSes in main loop */
data = g_malloc(sizeof(*data));
data->ret = ret;
- block_job_defer_to_main_loop(&s->common, stream_complete, data);
+ job_defer_to_main_loop(&s->common.job, stream_complete, data);
}
static const BlockJobDriver stream_job_driver = {
diff --git a/blockjob.c b/blockjob.c
index 0a0b1c41dd..3ede511da0 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -360,7 +360,7 @@ static void block_job_decommission(BlockJob *job)
job->completed = true;
job->busy = false;
job->paused = false;
- job->deferred_to_main_loop = true;
+ job->job.deferred_to_main_loop = true;
block_job_txn_del_job(job);
job_state_transition(&job->job, JOB_STATUS_NULL);
job_unref(&job->job);
@@ -515,7 +515,7 @@ static int block_job_finish_sync(BlockJob *job,
/* block_job_drain calls block_job_enter, and it should be enough to
* induce progress until the job completes or moves to the main thread.
*/
- while (!job->deferred_to_main_loop && !job->completed) {
+ while (!job->job.deferred_to_main_loop && !job->completed) {
block_job_drain(job);
}
while (!job->completed) {
@@ -729,7 +729,7 @@ void block_job_cancel(BlockJob *job, bool force)
block_job_cancel_async(job, force);
if (!block_job_started(job)) {
block_job_completed(job, -ECANCELED);
- } else if (job->deferred_to_main_loop) {
+ } else if (job->job.deferred_to_main_loop) {
block_job_completed_txn_abort(job);
} else {
block_job_enter(job);
@@ -1045,7 +1045,7 @@ static void block_job_enter_cond(BlockJob *job, bool(*fn)(BlockJob *job))
if (!block_job_started(job)) {
return;
}
- if (job->deferred_to_main_loop) {
+ if (job->job.deferred_to_main_loop) {
return;
}
@@ -1060,7 +1060,7 @@ static void block_job_enter_cond(BlockJob *job, bool(*fn)(BlockJob *job))
return;
}
- assert(!job->deferred_to_main_loop);
+ assert(!job->job.deferred_to_main_loop);
timer_del(&job->sleep_timer);
job->busy = true;
block_job_unlock();
@@ -1166,50 +1166,3 @@ BlockErrorAction block_job_error_action(BlockJob *job, BlockdevOnError on_err,
}
return action;
}
-
-typedef struct {
- BlockJob *job;
- AioContext *aio_context;
- BlockJobDeferToMainLoopFn *fn;
- void *opaque;
-} BlockJobDeferToMainLoopData;
-
-static void block_job_defer_to_main_loop_bh(void *opaque)
-{
- BlockJobDeferToMainLoopData *data = opaque;
- AioContext *aio_context;
-
- /* Prevent race with block_job_defer_to_main_loop() */
- aio_context_acquire(data->aio_context);
-
- /* Fetch BDS AioContext again, in case it has changed */
- aio_context = blk_get_aio_context(data->job->blk);
- if (aio_context != data->aio_context) {
- aio_context_acquire(aio_context);
- }
-
- data->fn(data->job, data->opaque);
-
- if (aio_context != data->aio_context) {
- aio_context_release(aio_context);
- }
-
- aio_context_release(data->aio_context);
-
- g_free(data);
-}
-
-void block_job_defer_to_main_loop(BlockJob *job,
- BlockJobDeferToMainLoopFn *fn,
- void *opaque)
-{
- BlockJobDeferToMainLoopData *data = g_malloc(sizeof(*data));
- data->job = job;
- data->aio_context = blk_get_aio_context(job->blk);
- data->fn = fn;
- data->opaque = opaque;
- job->deferred_to_main_loop = true;
-
- aio_bh_schedule_oneshot(qemu_get_aio_context(),
- block_job_defer_to_main_loop_bh, data);
-}
diff --git a/job.c b/job.c
index 7ff1f66273..2e0cc2e7a8 100644
--- a/job.c
+++ b/job.c
@@ -28,6 +28,7 @@
#include "qapi/error.h"
#include "qemu/job.h"
#include "qemu/id.h"
+#include "qemu/main-loop.h"
#include "trace-root.h"
static QLIST_HEAD(, Job) jobs = QLIST_HEAD_INITIALIZER(jobs);
@@ -170,3 +171,34 @@ void job_unref(Job *job)
g_free(job);
}
}
+
+typedef struct {
+ Job *job;
+ JobDeferToMainLoopFn *fn;
+ void *opaque;
+} JobDeferToMainLoopData;
+
+static void job_defer_to_main_loop_bh(void *opaque)
+{
+ JobDeferToMainLoopData *data = opaque;
+ Job *job = data->job;
+ AioContext *aio_context = job->aio_context;
+
+ aio_context_acquire(aio_context);
+ data->fn(data->job, data->opaque);
+ aio_context_release(aio_context);
+
+ g_free(data);
+}
+
+void job_defer_to_main_loop(Job *job, JobDeferToMainLoopFn *fn, void *opaque)
+{
+ JobDeferToMainLoopData *data = g_malloc(sizeof(*data));
+ data->job = job;
+ data->fn = fn;
+ data->opaque = opaque;
+ job->deferred_to_main_loop = true;
+
+ aio_bh_schedule_oneshot(qemu_get_aio_context(),
+ job_defer_to_main_loop_bh, data);
+}
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index f9e37d479c..4f8cba8377 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -496,9 +496,10 @@ typedef struct TestBlockJob {
bool should_complete;
} TestBlockJob;
-static void test_job_completed(BlockJob *job, void *opaque)
+static void test_job_completed(Job *job, void *opaque)
{
- block_job_completed(job, 0);
+ BlockJob *bjob = container_of(job, BlockJob, job);
+ block_job_completed(bjob, 0);
}
static void coroutine_fn test_job_start(void *opaque)
@@ -510,7 +511,7 @@ static void coroutine_fn test_job_start(void *opaque)
block_job_sleep_ns(&s->common, 100000);
}
- block_job_defer_to_main_loop(&s->common, test_job_completed, NULL);
+ job_defer_to_main_loop(&s->common.job, test_job_completed, NULL);
}
static void test_job_complete(BlockJob *job, Error **errp)
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 26b4bbb230..c03f9662d8 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -24,16 +24,17 @@ typedef struct {
int *result;
} TestBlockJob;
-static void test_block_job_complete(BlockJob *job, void *opaque)
+static void test_block_job_complete(Job *job, void *opaque)
{
- BlockDriverState *bs = blk_bs(job->blk);
+ BlockJob *bjob = container_of(job, BlockJob, job);
+ BlockDriverState *bs = blk_bs(bjob->blk);
int rc = (intptr_t)opaque;
- if (job_is_cancelled(&job->job)) {
+ if (job_is_cancelled(job)) {
rc = -ECANCELED;
}
- block_job_completed(job, rc);
+ block_job_completed(bjob, rc);
bdrv_unref(bs);
}
@@ -54,8 +55,8 @@ static void coroutine_fn test_block_job_run(void *opaque)
}
}
- block_job_defer_to_main_loop(job, test_block_job_complete,
- (void *)(intptr_t)s->rc);
+ job_defer_to_main_loop(&job->job, test_block_job_complete,
+ (void *)(intptr_t)s->rc);
}
typedef struct {
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index fa31481537..5f43bd72a4 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -161,11 +161,12 @@ typedef struct CancelJob {
bool completed;
} CancelJob;
-static void cancel_job_completed(BlockJob *job, void *opaque)
+static void cancel_job_completed(Job *job, void *opaque)
{
+ BlockJob *bjob = container_of(job, BlockJob, job);
CancelJob *s = opaque;
s->completed = true;
- block_job_completed(job, 0);
+ block_job_completed(bjob, 0);
}
static void cancel_job_complete(BlockJob *job, Error **errp)
@@ -191,7 +192,7 @@ static void coroutine_fn cancel_job_start(void *opaque)
}
defer:
- block_job_defer_to_main_loop(&s->common, cancel_job_completed, s);
+ job_defer_to_main_loop(&s->common.job, cancel_job_completed, s);
}
static const BlockJobDriver test_cancel_driver = {
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 12/40] job: Move defer_to_main_loop to Job
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 12/40] job: Move defer_to_main_loop to Job Kevin Wolf
@ 2018-05-18 17:56 ` John Snow
0 siblings, 0 replies; 86+ messages in thread
From: John Snow @ 2018-05-18 17:56 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: qemu-devel, jcody, armbru, mreitz
On 05/18/2018 09:20 AM, Kevin Wolf wrote:
> Move the defer_to_main_loop functionality from BlockJob to Job.
>
> The code can be simplified because we can use job->aio_context in
> job_defer_to_main_loop_bh() now, instead of having to access the
> BlockDriverState.
>
> Probably taking the data->aio_context lock in addition was already
> unnecessary in the old code because we didn't actually make use of
> anything protected by the old AioContext except getting the new
> AioContext, in case it changed between scheduling the BH and running it.
> But it's certainly unnecessary now that the BDS isn't accessed at all
> any more.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 13/40] job: Move coroutine and related code to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (11 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 12/40] job: Move defer_to_main_loop to Job Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 18:43 ` John Snow
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 14/40] job: Add job_sleep_ns() Kevin Wolf
` (29 subsequent siblings)
42 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This commit moves some core functions for dealing with the job coroutine
from BlockJob to Job. This includes primarily entering the coroutine
(both for the first and reentering) and yielding explicitly and at pause
points.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
include/block/blockjob.h | 40 --------
include/block/blockjob_int.h | 26 -----
include/qemu/job.h | 76 +++++++++++++++
block/backup.c | 2 +-
block/commit.c | 4 +-
block/mirror.c | 22 ++---
block/replication.c | 2 +-
block/stream.c | 4 +-
blockdev.c | 8 +-
blockjob.c | 219 ++++++++-----------------------------------
job.c | 137 +++++++++++++++++++++++++++
tests/test-bdrv-drain.c | 38 ++++----
tests/test-blockjob-txn.c | 12 +--
tests/test-blockjob.c | 14 +--
14 files changed, 305 insertions(+), 299 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 2a9e865e31..b60d919fbf 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -51,43 +51,18 @@ typedef struct BlockJob {
BlockBackend *blk;
/**
- * The coroutine that executes the job. If not NULL, it is
- * reentered when busy is false and the job is cancelled.
- */
- Coroutine *co;
-
- /**
* Set to true if the job should abort immediately without waiting
* for data to be in sync.
*/
bool force;
/**
- * Counter for pause request. If non-zero, the block job is either paused,
- * or if busy == true will pause itself as soon as possible.
- */
- int pause_count;
-
- /**
* Set to true if the job is paused by user. Can be unpaused with the
* block-job-resume QMP command.
*/
bool user_paused;
/**
- * Set to false by the job while the coroutine has yielded and may be
- * re-entered by block_job_enter(). There may still be I/O or event loop
- * activity pending. Accessed under block_job_mutex (in blockjob.c).
- */
- bool busy;
-
- /**
- * Set to true by the job while it is in a quiescent state, where
- * no I/O or event loop activity is pending.
- */
- bool paused;
-
- /**
* Set to true when the job is ready to be completed.
*/
bool ready;
@@ -125,12 +100,6 @@ typedef struct BlockJob {
/** ret code passed to block_job_completed. */
int ret;
- /**
- * Timer that is used by @block_job_sleep_ns. Accessed under
- * block_job_mutex (in blockjob.c).
- */
- QEMUTimer sleep_timer;
-
/** True if this job should automatically finalize itself */
bool auto_finalize;
@@ -208,15 +177,6 @@ void block_job_remove_all_bdrv(BlockJob *job);
void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp);
/**
- * block_job_start:
- * @job: A job that has not yet been started.
- *
- * Begins execution of a block job.
- * Takes ownership of one reference to the job object.
- */
-void block_job_start(BlockJob *job);
-
-/**
* block_job_cancel:
* @job: The job to be canceled.
* @force: Quit a job without waiting for data to be in sync.
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 0c2f8de381..0a614a89b8 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -38,9 +38,6 @@ struct BlockJobDriver {
/** Generic JobDriver callbacks and settings */
JobDriver job_driver;
- /** Mandatory: Entrypoint for the Coroutine. */
- CoroutineEntry *start;
-
/**
* Optional callback for job types whose completion must be triggered
* manually.
@@ -85,20 +82,6 @@ struct BlockJobDriver {
*/
void (*clean)(BlockJob *job);
- /**
- * If the callback is not NULL, it will be invoked when the job transitions
- * into the paused state. Paused jobs must not perform any asynchronous
- * I/O or event loop activity. This callback is used to quiesce jobs.
- */
- void coroutine_fn (*pause)(BlockJob *job);
-
- /**
- * If the callback is not NULL, it will be invoked when the job transitions
- * out of the paused state. Any asynchronous I/O or event loop activity
- * should be restarted from this callback.
- */
- void coroutine_fn (*resume)(BlockJob *job);
-
/*
* If the callback is not NULL, it will be invoked before the job is
* resumed in a new AioContext. This is the place to move any resources
@@ -196,15 +179,6 @@ void block_job_early_fail(BlockJob *job);
void block_job_completed(BlockJob *job, int ret);
/**
- * block_job_pause_point:
- * @job: The job that is ready to pause.
- *
- * Pause now if block_job_pause() has been called. Block jobs that perform
- * lots of I/O must call this between requests so that the job can be paused.
- */
-void coroutine_fn block_job_pause_point(BlockJob *job);
-
-/**
* block_job_enter:
* @job: The job to enter.
*
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 933e0ab328..9dcff12283 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -28,6 +28,7 @@
#include "qapi/qapi-types-block-core.h"
#include "qemu/queue.h"
+#include "qemu/coroutine.h"
typedef struct JobDriver JobDriver;
@@ -51,6 +52,37 @@ typedef struct Job {
AioContext *aio_context;
/**
+ * The coroutine that executes the job. If not NULL, it is reentered when
+ * busy is false and the job is cancelled.
+ */
+ Coroutine *co;
+
+ /**
+ * Timer that is used by @block_job_sleep_ns. Accessed under job_mutex (in
+ * job.c).
+ */
+ QEMUTimer sleep_timer;
+
+ /**
+ * Counter for pause request. If non-zero, the block job is either paused,
+ * or if busy == true will pause itself as soon as possible.
+ */
+ int pause_count;
+
+ /**
+ * Set to false by the job while the coroutine has yielded and may be
+ * re-entered by block_job_enter(). There may still be I/O or event loop
+ * activity pending. Accessed under block_job_mutex (in blockjob.c).
+ */
+ bool busy;
+
+ /**
+ * Set to true by the job while it is in a quiescent state, where
+ * no I/O or event loop activity is pending.
+ */
+ bool paused;
+
+ /**
* Set to true if the job should cancel itself. The flag must
* always be tested just before toggling the busy flag from false
* to true. After a job has been cancelled, it should only yield
@@ -75,6 +107,23 @@ struct JobDriver {
/** Enum describing the operation */
JobType job_type;
+ /** Mandatory: Entrypoint for the Coroutine. */
+ CoroutineEntry *start;
+
+ /**
+ * If the callback is not NULL, it will be invoked when the job transitions
+ * into the paused state. Paused jobs must not perform any asynchronous
+ * I/O or event loop activity. This callback is used to quiesce jobs.
+ */
+ void coroutine_fn (*pause)(Job *job);
+
+ /**
+ * If the callback is not NULL, it will be invoked when the job transitions
+ * out of the paused state. Any asynchronous I/O or event loop activity
+ * should be restarted from this callback.
+ */
+ void coroutine_fn (*resume)(Job *job);
+
/** Called when the job is freed */
void (*free)(Job *job);
};
@@ -103,6 +152,30 @@ void job_ref(Job *job);
*/
void job_unref(Job *job);
+/**
+ * Conditionally enter the job coroutine if the job is ready to run, not
+ * already busy and fn() returns true. fn() is called while under the job_lock
+ * critical section.
+ */
+void job_enter_cond(Job *job, bool(*fn)(Job *job));
+
+/**
+ * @job: A job that has not yet been started.
+ *
+ * Begins execution of a job.
+ * Takes ownership of one reference to the job object.
+ */
+void job_start(Job *job);
+
+/**
+ * @job: The job that is ready to pause.
+ *
+ * Pause now if job_pause() has been called. Jobs that perform lots of I/O
+ * must call this between requests so that the job can be paused.
+ */
+void coroutine_fn job_pause_point(Job *job);
+
+
/** Returns the JobType of a given Job. */
JobType job_type(const Job *job);
@@ -153,5 +226,8 @@ void job_defer_to_main_loop(Job *job, JobDeferToMainLoopFn *fn, void *opaque);
/* TODO To be removed from the public interface */
void job_state_transition(Job *job, JobStatus s1);
+void coroutine_fn job_do_yield(Job *job, uint64_t ns);
+bool job_should_pause(Job *job);
+bool job_started(Job *job);
#endif
diff --git a/block/backup.c b/block/backup.c
index 22dd368c90..7d9aad9749 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -528,8 +528,8 @@ static const BlockJobDriver backup_job_driver = {
.instance_size = sizeof(BackupBlockJob),
.job_type = JOB_TYPE_BACKUP,
.free = block_job_free,
+ .start = backup_run,
},
- .start = backup_run,
.commit = backup_commit,
.abort = backup_abort,
.clean = backup_clean,
diff --git a/block/commit.c b/block/commit.c
index d326766e4d..2fbc31077a 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -220,8 +220,8 @@ static const BlockJobDriver commit_job_driver = {
.instance_size = sizeof(CommitBlockJob),
.job_type = JOB_TYPE_COMMIT,
.free = block_job_free,
+ .start = commit_run,
},
- .start = commit_run,
};
static int coroutine_fn bdrv_commit_top_preadv(BlockDriverState *bs,
@@ -371,7 +371,7 @@ void commit_start(const char *job_id, BlockDriverState *bs,
s->on_error = on_error;
trace_commit_start(bs, base, top, s);
- block_job_start(&s->common);
+ job_start(&s->common.job);
return;
fail:
diff --git a/block/mirror.c b/block/mirror.c
index 90d4ac9cb6..95fc8072b0 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -126,7 +126,7 @@ static void mirror_iteration_done(MirrorOp *op, int ret)
g_free(op);
if (s->waiting_for_io) {
- qemu_coroutine_enter(s->common.co);
+ qemu_coroutine_enter(s->common.job.co);
}
}
@@ -345,7 +345,7 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s)
mirror_wait_for_io(s);
}
- block_job_pause_point(&s->common);
+ job_pause_point(&s->common.job);
/* Find the number of consective dirty chunks following the first dirty
* one, and wait for in flight requests in them. */
@@ -597,7 +597,7 @@ static void mirror_throttle(MirrorBlockJob *s)
s->last_pause_ns = now;
block_job_sleep_ns(&s->common, 0);
} else {
- block_job_pause_point(&s->common);
+ job_pause_point(&s->common.job);
}
}
@@ -786,7 +786,7 @@ static void coroutine_fn mirror_run(void *opaque)
goto immediate_exit;
}
- block_job_pause_point(&s->common);
+ job_pause_point(&s->common.job);
cnt = bdrv_get_dirty_count(s->dirty_bitmap);
/* cnt is the number of dirty bytes remaining and s->bytes_in_flight is
@@ -957,9 +957,9 @@ static void mirror_complete(BlockJob *job, Error **errp)
block_job_enter(&s->common);
}
-static void mirror_pause(BlockJob *job)
+static void mirror_pause(Job *job)
{
- MirrorBlockJob *s = container_of(job, MirrorBlockJob, common);
+ MirrorBlockJob *s = container_of(job, MirrorBlockJob, common.job);
mirror_wait_for_all_io(s);
}
@@ -991,10 +991,10 @@ static const BlockJobDriver mirror_job_driver = {
.instance_size = sizeof(MirrorBlockJob),
.job_type = JOB_TYPE_MIRROR,
.free = block_job_free,
+ .start = mirror_run,
+ .pause = mirror_pause,
},
- .start = mirror_run,
.complete = mirror_complete,
- .pause = mirror_pause,
.attached_aio_context = mirror_attached_aio_context,
.drain = mirror_drain,
};
@@ -1004,10 +1004,10 @@ static const BlockJobDriver commit_active_job_driver = {
.instance_size = sizeof(MirrorBlockJob),
.job_type = JOB_TYPE_COMMIT,
.free = block_job_free,
+ .start = mirror_run,
+ .pause = mirror_pause,
},
- .start = mirror_run,
.complete = mirror_complete,
- .pause = mirror_pause,
.attached_aio_context = mirror_attached_aio_context,
.drain = mirror_drain,
};
@@ -1244,7 +1244,7 @@ static void mirror_start_job(const char *job_id, BlockDriverState *bs,
}
trace_mirror_start(bs, s, opaque);
- block_job_start(&s->common);
+ job_start(&s->common.job);
return;
fail:
diff --git a/block/replication.c b/block/replication.c
index 48148b884a..9ed6e0fb04 100644
--- a/block/replication.c
+++ b/block/replication.c
@@ -576,7 +576,7 @@ static void replication_start(ReplicationState *rs, ReplicationMode mode,
aio_context_release(aio_context);
return;
}
- block_job_start(job);
+ job_start(&job->job);
break;
default:
aio_context_release(aio_context);
diff --git a/block/stream.c b/block/stream.c
index 0bba81678c..6d8b7b6eee 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -213,8 +213,8 @@ static const BlockJobDriver stream_job_driver = {
.instance_size = sizeof(StreamBlockJob),
.job_type = JOB_TYPE_STREAM,
.free = block_job_free,
+ .start = stream_run,
},
- .start = stream_run,
};
void stream_start(const char *job_id, BlockDriverState *bs,
@@ -262,7 +262,7 @@ void stream_start(const char *job_id, BlockDriverState *bs,
s->on_error = on_error;
trace_stream_start(bs, base, s);
- block_job_start(&s->common);
+ job_start(&s->common.job);
return;
fail:
diff --git a/blockdev.c b/blockdev.c
index 3808b1fc00..c551fdf39a 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -1910,7 +1910,7 @@ static void drive_backup_commit(BlkActionState *common)
aio_context_acquire(aio_context);
assert(state->job);
- block_job_start(state->job);
+ job_start(&state->job->job);
aio_context_release(aio_context);
}
@@ -2008,7 +2008,7 @@ static void blockdev_backup_commit(BlkActionState *common)
aio_context_acquire(aio_context);
assert(state->job);
- block_job_start(state->job);
+ job_start(&state->job->job);
aio_context_release(aio_context);
}
@@ -3425,7 +3425,7 @@ void qmp_drive_backup(DriveBackup *arg, Error **errp)
BlockJob *job;
job = do_drive_backup(arg, NULL, errp);
if (job) {
- block_job_start(job);
+ job_start(&job->job);
}
}
@@ -3513,7 +3513,7 @@ void qmp_blockdev_backup(BlockdevBackup *arg, Error **errp)
BlockJob *job;
job = do_blockdev_backup(arg, NULL, errp);
if (job) {
- block_job_start(job);
+ job_start(&job->job);
}
}
diff --git a/blockjob.c b/blockjob.c
index 3ede511da0..313b1ff7ce 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -36,30 +36,9 @@
#include "qemu/coroutine.h"
#include "qemu/timer.h"
-/* Right now, this mutex is only needed to synchronize accesses to job->busy
- * and job->sleep_timer, such as concurrent calls to block_job_do_yield and
- * block_job_enter. */
-static QemuMutex block_job_mutex;
-
-static void block_job_lock(void)
-{
- qemu_mutex_lock(&block_job_mutex);
-}
-
-static void block_job_unlock(void)
-{
- qemu_mutex_unlock(&block_job_mutex);
-}
-
-static void __attribute__((__constructor__)) block_job_init(void)
-{
- qemu_mutex_init(&block_job_mutex);
-}
-
static void block_job_event_cancelled(BlockJob *job);
static void block_job_event_completed(BlockJob *job, const char *msg);
static int block_job_event_pending(BlockJob *job);
-static void block_job_enter_cond(BlockJob *job, bool(*fn)(BlockJob *job));
/* Transactional group of block jobs */
struct BlockJobTxn {
@@ -161,33 +140,27 @@ static void block_job_txn_del_job(BlockJob *job)
}
}
-/* Assumes the block_job_mutex is held */
-static bool block_job_timer_pending(BlockJob *job)
-{
- return timer_pending(&job->sleep_timer);
-}
-
-/* Assumes the block_job_mutex is held */
-static bool block_job_timer_not_pending(BlockJob *job)
+/* Assumes the job_mutex is held */
+static bool job_timer_not_pending(Job *job)
{
- return !block_job_timer_pending(job);
+ return !timer_pending(&job->sleep_timer);
}
static void block_job_pause(BlockJob *job)
{
- job->pause_count++;
+ job->job.pause_count++;
}
static void block_job_resume(BlockJob *job)
{
- assert(job->pause_count > 0);
- job->pause_count--;
- if (job->pause_count) {
+ assert(job->job.pause_count > 0);
+ job->job.pause_count--;
+ if (job->job.pause_count) {
return;
}
/* kick only if no timer is pending */
- block_job_enter_cond(job, block_job_timer_not_pending);
+ job_enter_cond(&job->job, job_timer_not_pending);
}
static void block_job_attached_aio_context(AioContext *new_context,
@@ -208,7 +181,7 @@ void block_job_free(Job *job)
block_job_detach_aio_context, bjob);
blk_unref(bjob->blk);
error_free(bjob->blocker);
- assert(!timer_pending(&bjob->sleep_timer));
+ assert(!timer_pending(&bjob->job.sleep_timer));
}
static void block_job_attached_aio_context(AioContext *new_context,
@@ -226,7 +199,7 @@ static void block_job_attached_aio_context(AioContext *new_context,
static void block_job_drain(BlockJob *job)
{
- /* If job is !job->busy this kicks it into the next pause point. */
+ /* If job is !job->job.busy this kicks it into the next pause point. */
block_job_enter(job);
blk_drain(job->blk);
@@ -244,7 +217,7 @@ static void block_job_detach_aio_context(void *opaque)
block_job_pause(job);
- while (!job->paused && !job->completed) {
+ while (!job->job.paused && !job->completed) {
block_job_drain(job);
}
@@ -312,29 +285,11 @@ bool block_job_is_internal(BlockJob *job)
return (job->job.id == NULL);
}
-static bool block_job_started(BlockJob *job)
-{
- return job->co;
-}
-
const BlockJobDriver *block_job_driver(BlockJob *job)
{
return job->driver;
}
-/**
- * All jobs must allow a pause point before entering their job proper. This
- * ensures that jobs can be paused prior to being started, then resumed later.
- */
-static void coroutine_fn block_job_co_entry(void *opaque)
-{
- BlockJob *job = opaque;
-
- assert(job && job->driver && job->driver->start);
- block_job_pause_point(job);
- job->driver->start(job);
-}
-
static void block_job_sleep_timer_cb(void *opaque)
{
BlockJob *job = opaque;
@@ -342,24 +297,12 @@ static void block_job_sleep_timer_cb(void *opaque)
block_job_enter(job);
}
-void block_job_start(BlockJob *job)
-{
- assert(job && !block_job_started(job) && job->paused &&
- job->driver && job->driver->start);
- job->co = qemu_coroutine_create(block_job_co_entry, job);
- job->pause_count--;
- job->busy = true;
- job->paused = false;
- job_state_transition(&job->job, JOB_STATUS_RUNNING);
- bdrv_coroutine_enter(blk_bs(job->blk), job->co);
-}
-
static void block_job_decommission(BlockJob *job)
{
assert(job);
job->completed = true;
- job->busy = false;
- job->paused = false;
+ job->job.busy = false;
+ job->job.paused = false;
job->job.deferred_to_main_loop = true;
block_job_txn_del_job(job);
job_state_transition(&job->job, JOB_STATUS_NULL);
@@ -374,7 +317,7 @@ static void block_job_do_dismiss(BlockJob *job)
static void block_job_conclude(BlockJob *job)
{
job_state_transition(&job->job, JOB_STATUS_CONCLUDED);
- if (job->auto_dismiss || !block_job_started(job)) {
+ if (job->auto_dismiss || !job_started(&job->job)) {
block_job_do_dismiss(job);
}
}
@@ -439,7 +382,7 @@ static int block_job_finalize_single(BlockJob *job)
}
/* Emit events only if we actually started */
- if (block_job_started(job)) {
+ if (job_started(&job->job)) {
if (job_is_cancelled(&job->job)) {
block_job_event_cancelled(job);
} else {
@@ -464,7 +407,7 @@ static void block_job_cancel_async(BlockJob *job, bool force)
if (job->user_paused) {
/* Do not call block_job_enter here, the caller will handle it. */
job->user_paused = false;
- job->pause_count--;
+ job->job.pause_count--;
}
job->job.cancelled = true;
/* To prevent 'force == false' overriding a previous 'force == true' */
@@ -615,6 +558,12 @@ static void block_job_completed_txn_success(BlockJob *job)
}
}
+/* Assumes the job_mutex is held */
+static bool job_timer_pending(Job *job)
+{
+ return timer_pending(&job->sleep_timer);
+}
+
void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp)
{
int64_t old_speed = job->speed;
@@ -635,7 +584,7 @@ void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp)
}
/* kick only if a timer is pending */
- block_job_enter_cond(job, block_job_timer_pending);
+ job_enter_cond(&job->job, job_timer_pending);
}
int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n)
@@ -654,7 +603,7 @@ void block_job_complete(BlockJob *job, Error **errp)
if (job_apply_verb(&job->job, JOB_VERB_COMPLETE, errp)) {
return;
}
- if (job->pause_count || job_is_cancelled(&job->job) ||
+ if (job->job.pause_count || job_is_cancelled(&job->job) ||
!job->driver->complete)
{
error_setg(errp, "The active block job '%s' cannot be completed",
@@ -708,7 +657,7 @@ bool block_job_user_paused(BlockJob *job)
void block_job_user_resume(BlockJob *job, Error **errp)
{
assert(job);
- if (!job->user_paused || job->pause_count <= 0) {
+ if (!job->user_paused || job->job.pause_count <= 0) {
error_setg(errp, "Can't resume a job that was not paused");
return;
}
@@ -727,7 +676,7 @@ void block_job_cancel(BlockJob *job, bool force)
return;
}
block_job_cancel_async(job, force);
- if (!block_job_started(job)) {
+ if (!job_started(&job->job)) {
block_job_completed(job, -ECANCELED);
} else if (job->job.deferred_to_main_loop) {
block_job_completed_txn_abort(job);
@@ -797,8 +746,8 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
info->type = g_strdup(job_type_str(&job->job));
info->device = g_strdup(job->job.id);
info->len = job->len;
- info->busy = atomic_read(&job->busy);
- info->paused = job->pause_count > 0;
+ info->busy = atomic_read(&job->job.busy);
+ info->paused = job->job.pause_count > 0;
info->offset = job->offset;
info->speed = job->speed;
info->io_status = job->iostatus;
@@ -915,12 +864,9 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
job->blk = blk;
job->cb = cb;
job->opaque = opaque;
- job->busy = false;
- job->paused = true;
- job->pause_count = 1;
job->auto_finalize = !(flags & BLOCK_JOB_MANUAL_FINALIZE);
job->auto_dismiss = !(flags & BLOCK_JOB_MANUAL_DISMISS);
- aio_timer_init(qemu_get_aio_context(), &job->sleep_timer,
+ aio_timer_init(qemu_get_aio_context(), &job->job.sleep_timer,
QEMU_CLOCK_REALTIME, SCALE_NS,
block_job_sleep_timer_cb, job);
@@ -980,128 +926,41 @@ void block_job_completed(BlockJob *job, int ret)
}
}
-static bool block_job_should_pause(BlockJob *job)
-{
- return job->pause_count > 0;
-}
-
-/* Yield, and schedule a timer to reenter the coroutine after @ns nanoseconds.
- * Reentering the job coroutine with block_job_enter() before the timer has
- * expired is allowed and cancels the timer.
- *
- * If @ns is (uint64_t) -1, no timer is scheduled and block_job_enter() must be
- * called explicitly. */
-static void block_job_do_yield(BlockJob *job, uint64_t ns)
-{
- block_job_lock();
- if (ns != -1) {
- timer_mod(&job->sleep_timer, ns);
- }
- job->busy = false;
- block_job_unlock();
- qemu_coroutine_yield();
-
- /* Set by block_job_enter before re-entering the coroutine. */
- assert(job->busy);
-}
-
-void coroutine_fn block_job_pause_point(BlockJob *job)
-{
- assert(job && block_job_started(job));
-
- if (!block_job_should_pause(job)) {
- return;
- }
- if (job_is_cancelled(&job->job)) {
- return;
- }
-
- if (job->driver->pause) {
- job->driver->pause(job);
- }
-
- if (block_job_should_pause(job) && !job_is_cancelled(&job->job)) {
- JobStatus status = job->job.status;
- job_state_transition(&job->job, status == JOB_STATUS_READY
- ? JOB_STATUS_STANDBY
- : JOB_STATUS_PAUSED);
- job->paused = true;
- block_job_do_yield(job, -1);
- job->paused = false;
- job_state_transition(&job->job, status);
- }
-
- if (job->driver->resume) {
- job->driver->resume(job);
- }
-}
-
-/*
- * Conditionally enter a block_job pending a call to fn() while
- * under the block_job_lock critical section.
- */
-static void block_job_enter_cond(BlockJob *job, bool(*fn)(BlockJob *job))
-{
- if (!block_job_started(job)) {
- return;
- }
- if (job->job.deferred_to_main_loop) {
- return;
- }
-
- block_job_lock();
- if (job->busy) {
- block_job_unlock();
- return;
- }
-
- if (fn && !fn(job)) {
- block_job_unlock();
- return;
- }
-
- assert(!job->job.deferred_to_main_loop);
- timer_del(&job->sleep_timer);
- job->busy = true;
- block_job_unlock();
- aio_co_wake(job->co);
-}
-
void block_job_enter(BlockJob *job)
{
- block_job_enter_cond(job, NULL);
+ job_enter_cond(&job->job, NULL);
}
void block_job_sleep_ns(BlockJob *job, int64_t ns)
{
- assert(job->busy);
+ assert(job->job.busy);
/* Check cancellation *before* setting busy = false, too! */
if (job_is_cancelled(&job->job)) {
return;
}
- if (!block_job_should_pause(job)) {
- block_job_do_yield(job, qemu_clock_get_ns(QEMU_CLOCK_REALTIME) + ns);
+ if (!job_should_pause(&job->job)) {
+ job_do_yield(&job->job, qemu_clock_get_ns(QEMU_CLOCK_REALTIME) + ns);
}
- block_job_pause_point(job);
+ job_pause_point(&job->job);
}
void block_job_yield(BlockJob *job)
{
- assert(job->busy);
+ assert(job->job.busy);
/* Check cancellation *before* setting busy = false, too! */
if (job_is_cancelled(&job->job)) {
return;
}
- if (!block_job_should_pause(job)) {
- block_job_do_yield(job, -1);
+ if (!job_should_pause(&job->job)) {
+ job_do_yield(&job->job, -1);
}
- block_job_pause_point(job);
+ job_pause_point(&job->job);
}
void block_job_iostatus_reset(BlockJob *job)
@@ -1109,7 +968,7 @@ void block_job_iostatus_reset(BlockJob *job)
if (job->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
return;
}
- assert(job->user_paused && job->pause_count > 0);
+ assert(job->user_paused && job->job.pause_count > 0);
job->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
}
diff --git a/job.c b/job.c
index 2e0cc2e7a8..08b9273adf 100644
--- a/job.c
+++ b/job.c
@@ -60,6 +60,26 @@ bool JobVerbTable[JOB_VERB__MAX][JOB_STATUS__MAX] = {
[JOB_VERB_DISMISS] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0},
};
+/* Right now, this mutex is only needed to synchronize accesses to job->busy
+ * and job->sleep_timer, such as concurrent calls to job_do_yield and
+ * job_enter. */
+static QemuMutex job_mutex;
+
+static void job_lock(void)
+{
+ qemu_mutex_lock(&job_mutex);
+}
+
+static void job_unlock(void)
+{
+ qemu_mutex_unlock(&job_mutex);
+}
+
+static void __attribute__((__constructor__)) job_init(void)
+{
+ qemu_mutex_init(&job_mutex);
+}
+
/* TODO Make static once the whole state machine is in job.c */
void job_state_transition(Job *job, JobStatus s1)
{
@@ -101,6 +121,16 @@ bool job_is_cancelled(Job *job)
return job->cancelled;
}
+bool job_started(Job *job)
+{
+ return job->co;
+}
+
+bool job_should_pause(Job *job)
+{
+ return job->pause_count > 0;
+}
+
Job *job_next(Job *job)
{
if (!job) {
@@ -143,6 +173,9 @@ void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
job->id = g_strdup(job_id);
job->refcnt = 1;
job->aio_context = ctx;
+ job->busy = false;
+ job->paused = true;
+ job->pause_count = 1;
job_state_transition(job, JOB_STATUS_CREATED);
@@ -172,6 +205,110 @@ void job_unref(Job *job)
}
}
+void job_enter_cond(Job *job, bool(*fn)(Job *job))
+{
+ if (!job_started(job)) {
+ return;
+ }
+ if (job->deferred_to_main_loop) {
+ return;
+ }
+
+ job_lock();
+ if (job->busy) {
+ job_unlock();
+ return;
+ }
+
+ if (fn && !fn(job)) {
+ job_unlock();
+ return;
+ }
+
+ assert(!job->deferred_to_main_loop);
+ timer_del(&job->sleep_timer);
+ job->busy = true;
+ job_unlock();
+ aio_co_wake(job->co);
+}
+
+/* Yield, and schedule a timer to reenter the coroutine after @ns nanoseconds.
+ * Reentering the job coroutine with block_job_enter() before the timer has
+ * expired is allowed and cancels the timer.
+ *
+ * If @ns is (uint64_t) -1, no timer is scheduled and block_job_enter() must be
+ * called explicitly. */
+void coroutine_fn job_do_yield(Job *job, uint64_t ns)
+{
+ job_lock();
+ if (ns != -1) {
+ timer_mod(&job->sleep_timer, ns);
+ }
+ job->busy = false;
+ job_unlock();
+ qemu_coroutine_yield();
+
+ /* Set by job_enter_cond() before re-entering the coroutine. */
+ assert(job->busy);
+}
+
+void coroutine_fn job_pause_point(Job *job)
+{
+ assert(job && job_started(job));
+
+ if (!job_should_pause(job)) {
+ return;
+ }
+ if (job_is_cancelled(job)) {
+ return;
+ }
+
+ if (job->driver->pause) {
+ job->driver->pause(job);
+ }
+
+ if (job_should_pause(job) && !job_is_cancelled(job)) {
+ JobStatus status = job->status;
+ job_state_transition(job, status == JOB_STATUS_READY
+ ? JOB_STATUS_STANDBY
+ : JOB_STATUS_PAUSED);
+ job->paused = true;
+ job_do_yield(job, -1);
+ job->paused = false;
+ job_state_transition(job, status);
+ }
+
+ if (job->driver->resume) {
+ job->driver->resume(job);
+ }
+}
+
+/**
+ * All jobs must allow a pause point before entering their job proper. This
+ * ensures that jobs can be paused prior to being started, then resumed later.
+ */
+static void coroutine_fn job_co_entry(void *opaque)
+{
+ Job *job = opaque;
+
+ assert(job && job->driver && job->driver->start);
+ job_pause_point(job);
+ job->driver->start(job);
+}
+
+
+void job_start(Job *job)
+{
+ assert(job && !job_started(job) && job->paused &&
+ job->driver && job->driver->start);
+ job->co = qemu_coroutine_create(job_co_entry, job);
+ job->pause_count--;
+ job->busy = true;
+ job->paused = false;
+ job_state_transition(job, JOB_STATUS_RUNNING);
+ aio_co_enter(job->aio_context, job->co);
+}
+
typedef struct {
Job *job;
JobDeferToMainLoopFn *fn;
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index 4f8cba8377..c9f2f9b183 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -524,8 +524,8 @@ BlockJobDriver test_job_driver = {
.job_driver = {
.instance_size = sizeof(TestBlockJob),
.free = block_job_free,
+ .start = test_job_start,
},
- .start = test_job_start,
.complete = test_job_complete,
};
@@ -549,47 +549,47 @@ static void test_blockjob_common(enum drain_type drain_type)
job = block_job_create("job0", &test_job_driver, NULL, src, 0, BLK_PERM_ALL,
0, 0, NULL, NULL, &error_abort);
block_job_add_bdrv(job, "target", target, 0, BLK_PERM_ALL, &error_abort);
- block_job_start(job);
+ job_start(&job->job);
- g_assert_cmpint(job->pause_count, ==, 0);
- g_assert_false(job->paused);
- g_assert_false(job->busy); /* We're in block_job_sleep_ns() */
+ g_assert_cmpint(job->job.pause_count, ==, 0);
+ g_assert_false(job->job.paused);
+ g_assert_false(job->job.busy); /* We're in block_job_sleep_ns() */
do_drain_begin(drain_type, src);
if (drain_type == BDRV_DRAIN_ALL) {
/* bdrv_drain_all() drains both src and target */
- g_assert_cmpint(job->pause_count, ==, 2);
+ g_assert_cmpint(job->job.pause_count, ==, 2);
} else {
- g_assert_cmpint(job->pause_count, ==, 1);
+ g_assert_cmpint(job->job.pause_count, ==, 1);
}
/* XXX We don't wait until the job is actually paused. Is this okay? */
- /* g_assert_true(job->paused); */
- g_assert_false(job->busy); /* The job is paused */
+ /* g_assert_true(job->job.paused); */
+ g_assert_false(job->job.busy); /* The job is paused */
do_drain_end(drain_type, src);
- g_assert_cmpint(job->pause_count, ==, 0);
- g_assert_false(job->paused);
- g_assert_false(job->busy); /* We're in block_job_sleep_ns() */
+ g_assert_cmpint(job->job.pause_count, ==, 0);
+ g_assert_false(job->job.paused);
+ g_assert_false(job->job.busy); /* We're in block_job_sleep_ns() */
do_drain_begin(drain_type, target);
if (drain_type == BDRV_DRAIN_ALL) {
/* bdrv_drain_all() drains both src and target */
- g_assert_cmpint(job->pause_count, ==, 2);
+ g_assert_cmpint(job->job.pause_count, ==, 2);
} else {
- g_assert_cmpint(job->pause_count, ==, 1);
+ g_assert_cmpint(job->job.pause_count, ==, 1);
}
/* XXX We don't wait until the job is actually paused. Is this okay? */
- /* g_assert_true(job->paused); */
- g_assert_false(job->busy); /* The job is paused */
+ /* g_assert_true(job->job.paused); */
+ g_assert_false(job->job.busy); /* The job is paused */
do_drain_end(drain_type, target);
- g_assert_cmpint(job->pause_count, ==, 0);
- g_assert_false(job->paused);
- g_assert_false(job->busy); /* We're in block_job_sleep_ns() */
+ g_assert_cmpint(job->job.pause_count, ==, 0);
+ g_assert_false(job->job.paused);
+ g_assert_false(job->job.busy); /* We're in block_job_sleep_ns() */
ret = block_job_complete_sync(job, &error_abort);
g_assert_cmpint(ret, ==, 0);
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index c03f9662d8..323e154a00 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -78,8 +78,8 @@ static const BlockJobDriver test_block_job_driver = {
.job_driver = {
.instance_size = sizeof(TestBlockJob),
.free = block_job_free,
+ .start = test_block_job_run,
},
- .start = test_block_job_run,
};
/* Create a block job that completes with a given return code after a given
@@ -125,7 +125,7 @@ static void test_single_job(int expected)
txn = block_job_txn_new();
job = test_block_job_start(1, true, expected, &result, txn);
- block_job_start(job);
+ job_start(&job->job);
if (expected == -ECANCELED) {
block_job_cancel(job, false);
@@ -165,8 +165,8 @@ static void test_pair_jobs(int expected1, int expected2)
txn = block_job_txn_new();
job1 = test_block_job_start(1, true, expected1, &result1, txn);
job2 = test_block_job_start(2, true, expected2, &result2, txn);
- block_job_start(job1);
- block_job_start(job2);
+ job_start(&job1->job);
+ job_start(&job2->job);
/* Release our reference now to trigger as many nice
* use-after-free bugs as possible.
@@ -227,8 +227,8 @@ static void test_pair_jobs_fail_cancel_race(void)
txn = block_job_txn_new();
job1 = test_block_job_start(1, true, -ECANCELED, &result1, txn);
job2 = test_block_job_start(2, false, 0, &result2, txn);
- block_job_start(job1);
- block_job_start(job2);
+ job_start(&job1->job);
+ job_start(&job2->job);
block_job_cancel(job1, false);
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 5f43bd72a4..1d18325feb 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -199,8 +199,8 @@ static const BlockJobDriver test_cancel_driver = {
.job_driver = {
.instance_size = sizeof(CancelJob),
.free = block_job_free,
+ .start = cancel_job_start,
},
- .start = cancel_job_start,
.complete = cancel_job_complete,
};
@@ -254,7 +254,7 @@ static void test_cancel_running(void)
s = create_common(&job);
- block_job_start(job);
+ job_start(&job->job);
assert(job->job.status == JOB_STATUS_RUNNING);
cancel_common(s);
@@ -267,7 +267,7 @@ static void test_cancel_paused(void)
s = create_common(&job);
- block_job_start(job);
+ job_start(&job->job);
assert(job->job.status == JOB_STATUS_RUNNING);
block_job_user_pause(job, &error_abort);
@@ -284,7 +284,7 @@ static void test_cancel_ready(void)
s = create_common(&job);
- block_job_start(job);
+ job_start(&job->job);
assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
@@ -301,7 +301,7 @@ static void test_cancel_standby(void)
s = create_common(&job);
- block_job_start(job);
+ job_start(&job->job);
assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
@@ -322,7 +322,7 @@ static void test_cancel_pending(void)
s = create_common(&job);
- block_job_start(job);
+ job_start(&job->job);
assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
@@ -346,7 +346,7 @@ static void test_cancel_concluded(void)
s = create_common(&job);
- block_job_start(job);
+ job_start(&job->job);
assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 13/40] job: Move coroutine and related code to Job
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 13/40] job: Move coroutine and related code " Kevin Wolf
@ 2018-05-18 18:43 ` John Snow
0 siblings, 0 replies; 86+ messages in thread
From: John Snow @ 2018-05-18 18:43 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: qemu-devel, jcody, armbru, mreitz
On 05/18/2018 09:20 AM, Kevin Wolf wrote:
> This commit moves some core functions for dealing with the job coroutine
> from BlockJob to Job. This includes primarily entering the coroutine
> (both for the first and reentering) and yielding explicitly and at pause
> points.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The _cond functions get shuffled around a bit, you address Max's
comments and otherwise it's the same as in V1.
Reviewed-by: John Snow <jsnow@redhat.com>
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 14/40] job: Add job_sleep_ns()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (12 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 13/40] job: Move coroutine and related code " Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 15/40] job: Move pause/resume functions to Job Kevin Wolf
` (28 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
There is nothing block layer specific about block_job_sleep_ns(), so
move the function to Job.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob_int.h | 11 -----------
include/qemu/job.h | 19 ++++++++++++++++++-
block/backup.c | 2 +-
block/commit.c | 2 +-
block/mirror.c | 4 ++--
block/stream.c | 2 +-
blockjob.c | 27 ---------------------------
job.c | 32 ++++++++++++++++++++++++++++++++
tests/test-bdrv-drain.c | 8 ++++----
tests/test-blockjob-txn.c | 2 +-
tests/test-blockjob.c | 2 +-
11 files changed, 61 insertions(+), 50 deletions(-)
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 0a614a89b8..8937f5b163 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -134,17 +134,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
void block_job_free(Job *job);
/**
- * block_job_sleep_ns:
- * @job: The job that calls the function.
- * @ns: How many nanoseconds to stop for.
- *
- * Put the job to sleep (assuming that it wasn't canceled) for @ns
- * %QEMU_CLOCK_REALTIME nanoseconds. Canceling the job will immediately
- * interrupt the wait.
- */
-void block_job_sleep_ns(BlockJob *job, int64_t ns);
-
-/**
* block_job_yield:
* @job: The job that calls the function.
*
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 9dcff12283..509408f747 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -58,7 +58,7 @@ typedef struct Job {
Coroutine *co;
/**
- * Timer that is used by @block_job_sleep_ns. Accessed under job_mutex (in
+ * Timer that is used by @job_sleep_ns. Accessed under job_mutex (in
* job.c).
*/
QEMUTimer sleep_timer;
@@ -168,6 +168,13 @@ void job_enter_cond(Job *job, bool(*fn)(Job *job));
void job_start(Job *job);
/**
+ * @job: The job to enter.
+ *
+ * Continue the specified job by entering the coroutine.
+ */
+void job_enter(Job *job);
+
+/**
* @job: The job that is ready to pause.
*
* Pause now if job_pause() has been called. Jobs that perform lots of I/O
@@ -175,6 +182,16 @@ void job_start(Job *job);
*/
void coroutine_fn job_pause_point(Job *job);
+/**
+ * @job: The job that calls the function.
+ * @ns: How many nanoseconds to stop for.
+ *
+ * Put the job to sleep (assuming that it wasn't canceled) for @ns
+ * %QEMU_CLOCK_REALTIME nanoseconds. Canceling the job will immediately
+ * interrupt the wait.
+ */
+void coroutine_fn job_sleep_ns(Job *job, int64_t ns);
+
/** Returns the JobType of a given Job. */
JobType job_type(const Job *job);
diff --git a/block/backup.c b/block/backup.c
index 7d9aad9749..f3a4f7c898 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -338,7 +338,7 @@ static bool coroutine_fn yield_and_check(BackupBlockJob *job)
* return. Without a yield, the VM would not reboot. */
delay_ns = block_job_ratelimit_get_delay(&job->common, job->bytes_read);
job->bytes_read = 0;
- block_job_sleep_ns(&job->common, delay_ns);
+ job_sleep_ns(&job->common.job, delay_ns);
if (job_is_cancelled(&job->common.job)) {
return true;
diff --git a/block/commit.c b/block/commit.c
index 2fbc31077a..1c6cb6c298 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -172,7 +172,7 @@ static void coroutine_fn commit_run(void *opaque)
/* Note that even when no rate limit is applied we need to yield
* with no pending I/O here so that bdrv_drain_all() returns.
*/
- block_job_sleep_ns(&s->common, delay_ns);
+ job_sleep_ns(&s->common.job, delay_ns);
if (job_is_cancelled(&s->common.job)) {
break;
}
diff --git a/block/mirror.c b/block/mirror.c
index 95fc8072b0..5d8f75c677 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -595,7 +595,7 @@ static void mirror_throttle(MirrorBlockJob *s)
if (now - s->last_pause_ns > BLOCK_JOB_SLICE_TIME) {
s->last_pause_ns = now;
- block_job_sleep_ns(&s->common, 0);
+ job_sleep_ns(&s->common.job, 0);
} else {
job_pause_point(&s->common.job);
}
@@ -869,7 +869,7 @@ static void coroutine_fn mirror_run(void *opaque)
cnt == 0 ? BLOCK_JOB_SLICE_TIME : 0);
}
trace_mirror_before_sleep(s, cnt, s->synced, delay_ns);
- block_job_sleep_ns(&s->common, delay_ns);
+ job_sleep_ns(&s->common.job, delay_ns);
if (job_is_cancelled(&s->common.job) &&
(!s->synced || s->common.force))
{
diff --git a/block/stream.c b/block/stream.c
index 6d8b7b6eee..1faab02086 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -140,7 +140,7 @@ static void coroutine_fn stream_run(void *opaque)
/* Note that even when no rate limit is applied we need to yield
* with no pending I/O here so that bdrv_drain_all() returns.
*/
- block_job_sleep_ns(&s->common, delay_ns);
+ job_sleep_ns(&s->common.job, delay_ns);
if (job_is_cancelled(&s->common.job)) {
break;
}
diff --git a/blockjob.c b/blockjob.c
index 313b1ff7ce..4dc360c794 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -181,7 +181,6 @@ void block_job_free(Job *job)
block_job_detach_aio_context, bjob);
blk_unref(bjob->blk);
error_free(bjob->blocker);
- assert(!timer_pending(&bjob->job.sleep_timer));
}
static void block_job_attached_aio_context(AioContext *new_context,
@@ -290,13 +289,6 @@ const BlockJobDriver *block_job_driver(BlockJob *job)
return job->driver;
}
-static void block_job_sleep_timer_cb(void *opaque)
-{
- BlockJob *job = opaque;
-
- block_job_enter(job);
-}
-
static void block_job_decommission(BlockJob *job)
{
assert(job);
@@ -866,9 +858,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
job->opaque = opaque;
job->auto_finalize = !(flags & BLOCK_JOB_MANUAL_FINALIZE);
job->auto_dismiss = !(flags & BLOCK_JOB_MANUAL_DISMISS);
- aio_timer_init(qemu_get_aio_context(), &job->job.sleep_timer,
- QEMU_CLOCK_REALTIME, SCALE_NS,
- block_job_sleep_timer_cb, job);
error_setg(&job->blocker, "block device is in use by block job: %s",
job_type_str(&job->job));
@@ -931,22 +920,6 @@ void block_job_enter(BlockJob *job)
job_enter_cond(&job->job, NULL);
}
-void block_job_sleep_ns(BlockJob *job, int64_t ns)
-{
- assert(job->job.busy);
-
- /* Check cancellation *before* setting busy = false, too! */
- if (job_is_cancelled(&job->job)) {
- return;
- }
-
- if (!job_should_pause(&job->job)) {
- job_do_yield(&job->job, qemu_clock_get_ns(QEMU_CLOCK_REALTIME) + ns);
- }
-
- job_pause_point(&job->job);
-}
-
void block_job_yield(BlockJob *job)
{
assert(job->job.busy);
diff --git a/job.c b/job.c
index 08b9273adf..bff69bdaa5 100644
--- a/job.c
+++ b/job.c
@@ -152,6 +152,13 @@ Job *job_get(const char *id)
return NULL;
}
+static void job_sleep_timer_cb(void *opaque)
+{
+ Job *job = opaque;
+
+ job_enter(job);
+}
+
void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
Error **errp)
{
@@ -178,6 +185,9 @@ void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
job->pause_count = 1;
job_state_transition(job, JOB_STATUS_CREATED);
+ aio_timer_init(qemu_get_aio_context(), &job->sleep_timer,
+ QEMU_CLOCK_REALTIME, SCALE_NS,
+ job_sleep_timer_cb, job);
QLIST_INSERT_HEAD(&jobs, job, job_list);
@@ -193,6 +203,7 @@ void job_unref(Job *job)
{
if (--job->refcnt == 0) {
assert(job->status == JOB_STATUS_NULL);
+ assert(!timer_pending(&job->sleep_timer));
if (job->driver->free) {
job->driver->free(job);
@@ -232,6 +243,11 @@ void job_enter_cond(Job *job, bool(*fn)(Job *job))
aio_co_wake(job->co);
}
+void job_enter(Job *job)
+{
+ job_enter_cond(job, NULL);
+}
+
/* Yield, and schedule a timer to reenter the coroutine after @ns nanoseconds.
* Reentering the job coroutine with block_job_enter() before the timer has
* expired is allowed and cancels the timer.
@@ -283,6 +299,22 @@ void coroutine_fn job_pause_point(Job *job)
}
}
+void coroutine_fn job_sleep_ns(Job *job, int64_t ns)
+{
+ assert(job->busy);
+
+ /* Check cancellation *before* setting busy = false, too! */
+ if (job_is_cancelled(job)) {
+ return;
+ }
+
+ if (!job_should_pause(job)) {
+ job_do_yield(job, qemu_clock_get_ns(QEMU_CLOCK_REALTIME) + ns);
+ }
+
+ job_pause_point(job);
+}
+
/**
* All jobs must allow a pause point before entering their job proper. This
* ensures that jobs can be paused prior to being started, then resumed later.
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index c9f2f9b183..50232f5eaf 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -508,7 +508,7 @@ static void coroutine_fn test_job_start(void *opaque)
block_job_event_ready(&s->common);
while (!s->should_complete) {
- block_job_sleep_ns(&s->common, 100000);
+ job_sleep_ns(&s->common.job, 100000);
}
job_defer_to_main_loop(&s->common.job, test_job_completed, NULL);
@@ -553,7 +553,7 @@ static void test_blockjob_common(enum drain_type drain_type)
g_assert_cmpint(job->job.pause_count, ==, 0);
g_assert_false(job->job.paused);
- g_assert_false(job->job.busy); /* We're in block_job_sleep_ns() */
+ g_assert_false(job->job.busy); /* We're in job_sleep_ns() */
do_drain_begin(drain_type, src);
@@ -571,7 +571,7 @@ static void test_blockjob_common(enum drain_type drain_type)
g_assert_cmpint(job->job.pause_count, ==, 0);
g_assert_false(job->job.paused);
- g_assert_false(job->job.busy); /* We're in block_job_sleep_ns() */
+ g_assert_false(job->job.busy); /* We're in job_sleep_ns() */
do_drain_begin(drain_type, target);
@@ -589,7 +589,7 @@ static void test_blockjob_common(enum drain_type drain_type)
g_assert_cmpint(job->job.pause_count, ==, 0);
g_assert_false(job->job.paused);
- g_assert_false(job->job.busy); /* We're in block_job_sleep_ns() */
+ g_assert_false(job->job.busy); /* We're in job_sleep_ns() */
ret = block_job_complete_sync(job, &error_abort);
g_assert_cmpint(ret, ==, 0);
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 323e154a00..0e6162bc71 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -45,7 +45,7 @@ static void coroutine_fn test_block_job_run(void *opaque)
while (s->iterations--) {
if (s->use_timer) {
- block_job_sleep_ns(job, 0);
+ job_sleep_ns(&job->job, 0);
} else {
block_job_yield(job);
}
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 1d18325feb..b329bd5274 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -188,7 +188,7 @@ static void coroutine_fn cancel_job_start(void *opaque)
block_job_event_ready(&s->common);
}
- block_job_sleep_ns(&s->common, 100000);
+ job_sleep_ns(&s->common.job, 100000);
}
defer:
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 15/40] job: Move pause/resume functions to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (13 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 14/40] job: Add job_sleep_ns() Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 16/40] job: Replace BlockJob.completed with job_is_completed() Kevin Wolf
` (27 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
While we already moved the state related to job pausing to Job, the
functions to do were still BlockJob only. This commit moves them over to
Job.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
include/block/blockjob.h | 32 -----------------
include/block/blockjob_int.h | 7 ++++
include/qemu/job.h | 37 ++++++++++++++++++++
block/backup.c | 1 +
block/commit.c | 1 +
block/mirror.c | 2 ++
block/stream.c | 1 +
blockdev.c | 6 ++--
blockjob.c | 81 +++++++++-----------------------------------
job.c | 59 ++++++++++++++++++++++++++++++++
tests/test-bdrv-drain.c | 1 +
tests/test-blockjob-txn.c | 1 +
tests/test-blockjob.c | 6 ++--
13 files changed, 133 insertions(+), 102 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index b60d919fbf..556a8f6375 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -57,12 +57,6 @@ typedef struct BlockJob {
bool force;
/**
- * Set to true if the job is paused by user. Can be unpaused with the
- * block-job-resume QMP command.
- */
- bool user_paused;
-
- /**
* Set to true when the job is ready to be completed.
*/
bool ready;
@@ -248,32 +242,6 @@ void block_job_progress_set_remaining(BlockJob *job, uint64_t remaining);
BlockJobInfo *block_job_query(BlockJob *job, Error **errp);
/**
- * block_job_user_pause:
- * @job: The job to be paused.
- *
- * Asynchronously pause the specified job.
- * Do not allow a resume until a matching call to block_job_user_resume.
- */
-void block_job_user_pause(BlockJob *job, Error **errp);
-
-/**
- * block_job_paused:
- * @job: The job to query.
- *
- * Returns true if the job is user-paused.
- */
-bool block_job_user_paused(BlockJob *job);
-
-/**
- * block_job_user_resume:
- * @job: The job to be resumed.
- *
- * Resume the specified job.
- * Must be paired with a preceding block_job_user_pause.
- */
-void block_job_user_resume(BlockJob *job, Error **errp);
-
-/**
* block_job_user_cancel:
* @job: The job to be cancelled.
* @force: Quit a job without waiting for data to be in sync.
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 8937f5b163..7e705ae0e9 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -134,6 +134,13 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
void block_job_free(Job *job);
/**
+ * block_job_user_resume:
+ * Callback to be used for JobDriver.user_resume in all block jobs. Resets the
+ * iostatus when the user resumes @job.
+ */
+void block_job_user_resume(Job *job);
+
+/**
* block_job_yield:
* @job: The job that calls the function.
*
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 509408f747..bc6398568f 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -83,6 +83,12 @@ typedef struct Job {
bool paused;
/**
+ * Set to true if the job is paused by user. Can be unpaused with the
+ * block-job-resume QMP command.
+ */
+ bool user_paused;
+
+ /**
* Set to true if the job should cancel itself. The flag must
* always be tested just before toggling the busy flag from false
* to true. After a job has been cancelled, it should only yield
@@ -124,6 +130,12 @@ struct JobDriver {
*/
void coroutine_fn (*resume)(Job *job);
+ /**
+ * Called when the job is resumed by the user (i.e. user_paused becomes
+ * false). .user_resume is called before .resume.
+ */
+ void (*user_resume)(Job *job);
+
/** Called when the job is freed */
void (*free)(Job *job);
};
@@ -203,6 +215,31 @@ const char *job_type_str(const Job *job);
bool job_is_cancelled(Job *job);
/**
+ * Request @job to pause at the next pause point. Must be paired with
+ * job_resume(). If the job is supposed to be resumed by user action, call
+ * job_user_pause() instead.
+ */
+void job_pause(Job *job);
+
+/** Resumes a @job paused with job_pause. */
+void job_resume(Job *job);
+
+/**
+ * Asynchronously pause the specified @job.
+ * Do not allow a resume until a matching call to job_user_resume.
+ */
+void job_user_pause(Job *job, Error **errp);
+
+/** Returns true if the job is user-paused. */
+bool job_user_paused(Job *job);
+
+/**
+ * Resume the specified @job.
+ * Must be paired with a preceding job_user_pause.
+ */
+void job_user_resume(Job *job, Error **errp);
+
+/**
* Get the next element from the list of block jobs after @job, or the
* first one if @job is %NULL.
*
diff --git a/block/backup.c b/block/backup.c
index f3a4f7c898..4d011d5a5c 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -528,6 +528,7 @@ static const BlockJobDriver backup_job_driver = {
.instance_size = sizeof(BackupBlockJob),
.job_type = JOB_TYPE_BACKUP,
.free = block_job_free,
+ .user_resume = block_job_user_resume,
.start = backup_run,
},
.commit = backup_commit,
diff --git a/block/commit.c b/block/commit.c
index 1c6cb6c298..c4a98e5804 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -220,6 +220,7 @@ static const BlockJobDriver commit_job_driver = {
.instance_size = sizeof(CommitBlockJob),
.job_type = JOB_TYPE_COMMIT,
.free = block_job_free,
+ .user_resume = block_job_user_resume,
.start = commit_run,
},
};
diff --git a/block/mirror.c b/block/mirror.c
index 5d8f75c677..9a7226f38c 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -991,6 +991,7 @@ static const BlockJobDriver mirror_job_driver = {
.instance_size = sizeof(MirrorBlockJob),
.job_type = JOB_TYPE_MIRROR,
.free = block_job_free,
+ .user_resume = block_job_user_resume,
.start = mirror_run,
.pause = mirror_pause,
},
@@ -1004,6 +1005,7 @@ static const BlockJobDriver commit_active_job_driver = {
.instance_size = sizeof(MirrorBlockJob),
.job_type = JOB_TYPE_COMMIT,
.free = block_job_free,
+ .user_resume = block_job_user_resume,
.start = mirror_run,
.pause = mirror_pause,
},
diff --git a/block/stream.c b/block/stream.c
index 1faab02086..e81b488a22 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -214,6 +214,7 @@ static const BlockJobDriver stream_job_driver = {
.job_type = JOB_TYPE_STREAM,
.free = block_job_free,
.start = stream_run,
+ .user_resume = block_job_user_resume,
},
};
diff --git a/blockdev.c b/blockdev.c
index c551fdf39a..3533c0dd6a 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -3844,7 +3844,7 @@ void qmp_block_job_cancel(const char *device,
force = false;
}
- if (block_job_user_paused(job) && !force) {
+ if (job_user_paused(&job->job) && !force) {
error_setg(errp, "The block job for device '%s' is currently paused",
device);
goto out;
@@ -3866,7 +3866,7 @@ void qmp_block_job_pause(const char *device, Error **errp)
}
trace_qmp_block_job_pause(job);
- block_job_user_pause(job, errp);
+ job_user_pause(&job->job, errp);
aio_context_release(aio_context);
}
@@ -3880,7 +3880,7 @@ void qmp_block_job_resume(const char *device, Error **errp)
}
trace_qmp_block_job_resume(job);
- block_job_user_resume(job, errp);
+ job_user_resume(&job->job, errp);
aio_context_release(aio_context);
}
diff --git a/blockjob.c b/blockjob.c
index 4dc360c794..6334a54f50 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -140,29 +140,6 @@ static void block_job_txn_del_job(BlockJob *job)
}
}
-/* Assumes the job_mutex is held */
-static bool job_timer_not_pending(Job *job)
-{
- return !timer_pending(&job->sleep_timer);
-}
-
-static void block_job_pause(BlockJob *job)
-{
- job->job.pause_count++;
-}
-
-static void block_job_resume(BlockJob *job)
-{
- assert(job->job.pause_count > 0);
- job->job.pause_count--;
- if (job->job.pause_count) {
- return;
- }
-
- /* kick only if no timer is pending */
- job_enter_cond(&job->job, job_timer_not_pending);
-}
-
static void block_job_attached_aio_context(AioContext *new_context,
void *opaque);
static void block_job_detach_aio_context(void *opaque);
@@ -193,7 +170,7 @@ static void block_job_attached_aio_context(AioContext *new_context,
job->driver->attached_aio_context(job, new_context);
}
- block_job_resume(job);
+ job_resume(&job->job);
}
static void block_job_drain(BlockJob *job)
@@ -214,7 +191,7 @@ static void block_job_detach_aio_context(void *opaque)
/* In case the job terminates during aio_poll()... */
job_ref(&job->job);
- block_job_pause(job);
+ job_pause(&job->job);
while (!job->job.paused && !job->completed) {
block_job_drain(job);
@@ -233,13 +210,13 @@ static char *child_job_get_parent_desc(BdrvChild *c)
static void child_job_drained_begin(BdrvChild *c)
{
BlockJob *job = c->opaque;
- block_job_pause(job);
+ job_pause(&job->job);
}
static void child_job_drained_end(BdrvChild *c)
{
BlockJob *job = c->opaque;
- block_job_resume(job);
+ job_resume(&job->job);
}
static const BdrvChildRole child_job = {
@@ -396,9 +373,9 @@ static void block_job_cancel_async(BlockJob *job, bool force)
if (job->iostatus != BLOCK_DEVICE_IO_STATUS_OK) {
block_job_iostatus_reset(job);
}
- if (job->user_paused) {
+ if (job->job.user_paused) {
/* Do not call block_job_enter here, the caller will handle it. */
- job->user_paused = false;
+ job->job.user_paused = false;
job->job.pause_count--;
}
job->job.cancelled = true;
@@ -628,39 +605,6 @@ void block_job_dismiss(BlockJob **jobptr, Error **errp)
*jobptr = NULL;
}
-void block_job_user_pause(BlockJob *job, Error **errp)
-{
- if (job_apply_verb(&job->job, JOB_VERB_PAUSE, errp)) {
- return;
- }
- if (job->user_paused) {
- error_setg(errp, "Job is already paused");
- return;
- }
- job->user_paused = true;
- block_job_pause(job);
-}
-
-bool block_job_user_paused(BlockJob *job)
-{
- return job->user_paused;
-}
-
-void block_job_user_resume(BlockJob *job, Error **errp)
-{
- assert(job);
- if (!job->user_paused || job->job.pause_count <= 0) {
- error_setg(errp, "Can't resume a job that was not paused");
- return;
- }
- if (job_apply_verb(&job->job, JOB_VERB_RESUME, errp)) {
- return;
- }
- block_job_iostatus_reset(job);
- job->user_paused = false;
- block_job_resume(job);
-}
-
void block_job_cancel(BlockJob *job, bool force)
{
if (job->job.status == JOB_STATUS_CONCLUDED) {
@@ -851,6 +795,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
assert(is_block_job(&job->job));
assert(job->job.driver->free == &block_job_free);
+ assert(job->job.driver->user_resume == &block_job_user_resume);
job->driver = driver;
job->blk = blk;
@@ -941,10 +886,16 @@ void block_job_iostatus_reset(BlockJob *job)
if (job->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
return;
}
- assert(job->user_paused && job->job.pause_count > 0);
+ assert(job->job.user_paused && job->job.pause_count > 0);
job->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
}
+void block_job_user_resume(Job *job)
+{
+ BlockJob *bjob = container_of(job, BlockJob, job);
+ block_job_iostatus_reset(bjob);
+}
+
void block_job_event_ready(BlockJob *job)
{
job_state_transition(&job->job, JOB_STATUS_READY);
@@ -991,9 +942,9 @@ BlockErrorAction block_job_error_action(BlockJob *job, BlockdevOnError on_err,
action, &error_abort);
}
if (action == BLOCK_ERROR_ACTION_STOP) {
- block_job_pause(job);
+ job_pause(&job->job);
/* make the pause user visible, which will be resumed from QMP. */
- job->user_paused = true;
+ job->job.user_paused = true;
block_job_iostatus_set_err(job, error);
}
return action;
diff --git a/job.c b/job.c
index bff69bdaa5..21e3219495 100644
--- a/job.c
+++ b/job.c
@@ -341,6 +341,65 @@ void job_start(Job *job)
aio_co_enter(job->aio_context, job->co);
}
+/* Assumes the block_job_mutex is held */
+static bool job_timer_not_pending(Job *job)
+{
+ return !timer_pending(&job->sleep_timer);
+}
+
+void job_pause(Job *job)
+{
+ job->pause_count++;
+}
+
+void job_resume(Job *job)
+{
+ assert(job->pause_count > 0);
+ job->pause_count--;
+ if (job->pause_count) {
+ return;
+ }
+
+ /* kick only if no timer is pending */
+ job_enter_cond(job, job_timer_not_pending);
+}
+
+void job_user_pause(Job *job, Error **errp)
+{
+ if (job_apply_verb(job, JOB_VERB_PAUSE, errp)) {
+ return;
+ }
+ if (job->user_paused) {
+ error_setg(errp, "Job is already paused");
+ return;
+ }
+ job->user_paused = true;
+ job_pause(job);
+}
+
+bool job_user_paused(Job *job)
+{
+ return job->user_paused;
+}
+
+void job_user_resume(Job *job, Error **errp)
+{
+ assert(job);
+ if (!job->user_paused || job->pause_count <= 0) {
+ error_setg(errp, "Can't resume a job that was not paused");
+ return;
+ }
+ if (job_apply_verb(job, JOB_VERB_RESUME, errp)) {
+ return;
+ }
+ if (job->driver->user_resume) {
+ job->driver->user_resume(job);
+ }
+ job->user_paused = false;
+ job_resume(job);
+}
+
+
typedef struct {
Job *job;
JobDeferToMainLoopFn *fn;
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index 50232f5eaf..c993512f66 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -524,6 +524,7 @@ BlockJobDriver test_job_driver = {
.job_driver = {
.instance_size = sizeof(TestBlockJob),
.free = block_job_free,
+ .user_resume = block_job_user_resume,
.start = test_job_start,
},
.complete = test_job_complete,
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 0e6162bc71..93d1ff0859 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -78,6 +78,7 @@ static const BlockJobDriver test_block_job_driver = {
.job_driver = {
.instance_size = sizeof(TestBlockJob),
.free = block_job_free,
+ .user_resume = block_job_user_resume,
.start = test_block_job_run,
},
};
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index b329bd5274..ceb59600ed 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -20,6 +20,7 @@ static const BlockJobDriver test_block_job_driver = {
.job_driver = {
.instance_size = sizeof(BlockJob),
.free = block_job_free,
+ .user_resume = block_job_user_resume,
},
};
@@ -199,6 +200,7 @@ static const BlockJobDriver test_cancel_driver = {
.job_driver = {
.instance_size = sizeof(CancelJob),
.free = block_job_free,
+ .user_resume = block_job_user_resume,
.start = cancel_job_start,
},
.complete = cancel_job_complete,
@@ -270,7 +272,7 @@ static void test_cancel_paused(void)
job_start(&job->job);
assert(job->job.status == JOB_STATUS_RUNNING);
- block_job_user_pause(job, &error_abort);
+ job_user_pause(&job->job, &error_abort);
block_job_enter(job);
assert(job->job.status == JOB_STATUS_PAUSED);
@@ -308,7 +310,7 @@ static void test_cancel_standby(void)
block_job_enter(job);
assert(job->job.status == JOB_STATUS_READY);
- block_job_user_pause(job, &error_abort);
+ job_user_pause(&job->job, &error_abort);
block_job_enter(job);
assert(job->job.status == JOB_STATUS_STANDBY);
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 16/40] job: Replace BlockJob.completed with job_is_completed()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (14 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 15/40] job: Move pause/resume functions to Job Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 17/40] job: Move BlockJobCreateFlags to Job Kevin Wolf
` (26 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
Since we introduced an explicit status to block job, BlockJob.completed
is redundant because it can be derived from the status. Remove the field
from BlockJob and add a function to derive it from the status at the Job
level.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
---
include/block/blockjob.h | 3 ---
include/qemu/job.h | 3 +++
blockjob.c | 16 +++++++---------
job.c | 22 ++++++++++++++++++++++
qemu-img.c | 4 ++--
5 files changed, 34 insertions(+), 14 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 556a8f6375..3e94e18850 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -88,9 +88,6 @@ typedef struct BlockJob {
/** The opaque value that is passed to the completion function. */
void *opaque;
- /** True when job has reported completion by calling block_job_completed. */
- bool completed;
-
/** ret code passed to block_job_completed. */
int ret;
diff --git a/include/qemu/job.h b/include/qemu/job.h
index bc6398568f..858f3beea4 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -214,6 +214,9 @@ const char *job_type_str(const Job *job);
/** Returns whether the job is scheduled for cancellation. */
bool job_is_cancelled(Job *job);
+/** Returns whether the job is in a completed state. */
+bool job_is_completed(Job *job);
+
/**
* Request @job to pause at the next pause point. Must be paired with
* job_resume(). If the job is supposed to be resumed by user action, call
diff --git a/blockjob.c b/blockjob.c
index 6334a54f50..a1d1f488e9 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -193,7 +193,7 @@ static void block_job_detach_aio_context(void *opaque)
job_pause(&job->job);
- while (!job->job.paused && !job->completed) {
+ while (!job->job.paused && !job_is_completed(&job->job)) {
block_job_drain(job);
}
@@ -269,7 +269,6 @@ const BlockJobDriver *block_job_driver(BlockJob *job)
static void block_job_decommission(BlockJob *job)
{
assert(job);
- job->completed = true;
job->job.busy = false;
job->job.paused = false;
job->job.deferred_to_main_loop = true;
@@ -334,7 +333,7 @@ static void block_job_clean(BlockJob *job)
static int block_job_finalize_single(BlockJob *job)
{
- assert(job->completed);
+ assert(job_is_completed(&job->job));
/* Ensure abort is called for late-transactional failures */
block_job_update_rc(job);
@@ -427,10 +426,10 @@ static int block_job_finish_sync(BlockJob *job,
/* block_job_drain calls block_job_enter, and it should be enough to
* induce progress until the job completes or moves to the main thread.
*/
- while (!job->job.deferred_to_main_loop && !job->completed) {
+ while (!job->job.deferred_to_main_loop && !job_is_completed(&job->job)) {
block_job_drain(job);
}
- while (!job->completed) {
+ while (!job_is_completed(&job->job)) {
aio_poll(qemu_get_aio_context(), true);
}
ret = (job_is_cancelled(&job->job) && job->ret == 0)
@@ -471,7 +470,7 @@ static void block_job_completed_txn_abort(BlockJob *job)
while (!QLIST_EMPTY(&txn->jobs)) {
other_job = QLIST_FIRST(&txn->jobs);
ctx = blk_get_aio_context(other_job->blk);
- if (!other_job->completed) {
+ if (!job_is_completed(&other_job->job)) {
assert(job_is_cancelled(&other_job->job));
block_job_finish_sync(other_job, NULL, NULL);
}
@@ -513,7 +512,7 @@ static void block_job_completed_txn_success(BlockJob *job)
* txn.
*/
QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
- if (!other_job->completed) {
+ if (!job_is_completed(&other_job->job)) {
return;
}
assert(other_job->ret == 0);
@@ -847,9 +846,8 @@ void block_job_early_fail(BlockJob *job)
void block_job_completed(BlockJob *job, int ret)
{
- assert(job && job->txn && !job->completed);
+ assert(job && job->txn && !job_is_completed(&job->job));
assert(blk_bs(job->blk)->job == job);
- job->completed = true;
job->ret = ret;
block_job_update_rc(job);
trace_block_job_completed(job, ret, job->ret);
diff --git a/job.c b/job.c
index 21e3219495..1e432af0d6 100644
--- a/job.c
+++ b/job.c
@@ -121,6 +121,28 @@ bool job_is_cancelled(Job *job)
return job->cancelled;
}
+bool job_is_completed(Job *job)
+{
+ switch (job->status) {
+ case JOB_STATUS_UNDEFINED:
+ case JOB_STATUS_CREATED:
+ case JOB_STATUS_RUNNING:
+ case JOB_STATUS_PAUSED:
+ case JOB_STATUS_READY:
+ case JOB_STATUS_STANDBY:
+ return false;
+ case JOB_STATUS_WAITING:
+ case JOB_STATUS_PENDING:
+ case JOB_STATUS_ABORTING:
+ case JOB_STATUS_CONCLUDED:
+ case JOB_STATUS_NULL:
+ return true;
+ default:
+ g_assert_not_reached();
+ }
+ return false;
+}
+
bool job_started(Job *job)
{
return job->co;
diff --git a/qemu-img.c b/qemu-img.c
index 27bbcc5100..e6a689c794 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -867,9 +867,9 @@ static void run_block_job(BlockJob *job, Error **errp)
aio_poll(aio_context, true);
qemu_progress_print(job->len ?
((float)job->offset / job->len * 100.f) : 0.0f, 0);
- } while (!job->ready && !job->completed);
+ } while (!job->ready && !job_is_completed(&job->job));
- if (!job->completed) {
+ if (!job_is_completed(&job->job)) {
ret = block_job_complete_sync(job, errp);
} else {
ret = job->ret;
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 17/40] job: Move BlockJobCreateFlags to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (15 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 16/40] job: Replace BlockJob.completed with job_is_completed() Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-23 22:24 ` John Snow
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 18/40] blockjob: Split block_job_event_pending() Kevin Wolf
` (25 subsequent siblings)
42 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This renames the BlockJobCreateFlags constants, moves a few JOB_INTERNAL
checks to job_create() and the auto_{finalize,dismiss} fields from
BlockJob to Job.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 17 -----------------
include/block/blockjob_int.h | 3 +--
include/qemu/job.h | 20 +++++++++++++++++++-
block/commit.c | 2 +-
block/mirror.c | 2 +-
block/replication.c | 4 ++--
block/stream.c | 2 +-
blockdev.c | 14 +++++++-------
blockjob.c | 27 +++++++--------------------
job.c | 11 ++++++++++-
qemu-img.c | 2 +-
tests/test-blockjob-txn.c | 2 +-
tests/test-blockjob.c | 4 ++--
13 files changed, 53 insertions(+), 57 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 3e94e18850..f9aaaaa835 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -91,27 +91,10 @@ typedef struct BlockJob {
/** ret code passed to block_job_completed. */
int ret;
- /** True if this job should automatically finalize itself */
- bool auto_finalize;
-
- /** True if this job should automatically dismiss itself */
- bool auto_dismiss;
-
BlockJobTxn *txn;
QLIST_ENTRY(BlockJob) txn_list;
} BlockJob;
-typedef enum BlockJobCreateFlags {
- /* Default behavior */
- BLOCK_JOB_DEFAULT = 0x00,
- /* BlockJob is not QMP-created and should not send QMP events */
- BLOCK_JOB_INTERNAL = 0x01,
- /* BlockJob requires manual finalize step */
- BLOCK_JOB_MANUAL_FINALIZE = 0x02,
- /* BlockJob requires manual dismiss step */
- BLOCK_JOB_MANUAL_DISMISS = 0x04,
-} BlockJobCreateFlags;
-
/**
* block_job_next:
* @job: A block job, or %NULL.
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 7e705ae0e9..88639f7efc 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -106,8 +106,7 @@ struct BlockJobDriver {
* @bs: The block
* @perm, @shared_perm: Permissions to request for @bs
* @speed: The maximum speed, in bytes per second, or 0 for unlimited.
- * @flags: Creation flags for the Block Job.
- * See @BlockJobCreateFlags
+ * @flags: Creation flags for the Block Job. See @JobCreateFlags.
* @cb: Completion function for the job.
* @opaque: Opaque pointer value passed to @cb.
* @errp: Error object.
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 858f3beea4..9783e4049b 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -99,6 +99,12 @@ typedef struct Job {
/** Set to true when the job has deferred work to the main loop. */
bool deferred_to_main_loop;
+ /** True if this job should automatically finalize itself */
+ bool auto_finalize;
+
+ /** True if this job should automatically dismiss itself */
+ bool auto_dismiss;
+
/** Element of the list of jobs */
QLIST_ENTRY(Job) job_list;
} Job;
@@ -140,6 +146,17 @@ struct JobDriver {
void (*free)(Job *job);
};
+typedef enum JobCreateFlags {
+ /* Default behavior */
+ JOB_DEFAULT = 0x00,
+ /* Job is not QMP-created and should not send QMP events */
+ JOB_INTERNAL = 0x01,
+ /* Job requires manual finalize step */
+ JOB_MANUAL_FINALIZE = 0x02,
+ /* Job requires manual dismiss step */
+ JOB_MANUAL_DISMISS = 0x04,
+} JobCreateFlags;
+
/**
* Create a new long-running job and return it.
@@ -147,10 +164,11 @@ struct JobDriver {
* @job_id: The id of the newly-created job, or %NULL for internal jobs
* @driver: The class object for the newly-created job.
* @ctx: The AioContext to run the job coroutine in.
+ * @flags: Creation flags for the job. See @JobCreateFlags.
* @errp: Error object.
*/
void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
- Error **errp);
+ int flags, Error **errp);
/**
* Add a reference to Job refcnt, it will be decreased with job_unref, and then
diff --git a/block/commit.c b/block/commit.c
index c4a98e5804..7a6ae59d42 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -282,7 +282,7 @@ void commit_start(const char *job_id, BlockDriverState *bs,
}
s = block_job_create(job_id, &commit_job_driver, NULL, bs, 0, BLK_PERM_ALL,
- speed, BLOCK_JOB_DEFAULT, NULL, NULL, errp);
+ speed, JOB_DEFAULT, NULL, NULL, errp);
if (!s) {
return;
}
diff --git a/block/mirror.c b/block/mirror.c
index 9a7226f38c..5091e72554 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -1284,7 +1284,7 @@ void mirror_start(const char *job_id, BlockDriverState *bs,
}
is_none_mode = mode == MIRROR_SYNC_MODE_NONE;
base = mode == MIRROR_SYNC_MODE_TOP ? backing_bs(bs) : NULL;
- mirror_start_job(job_id, bs, BLOCK_JOB_DEFAULT, target, replaces,
+ mirror_start_job(job_id, bs, JOB_DEFAULT, target, replaces,
speed, granularity, buf_size, backing_mode,
on_source_error, on_target_error, unmap, NULL, NULL,
&mirror_job_driver, is_none_mode, base, false,
diff --git a/block/replication.c b/block/replication.c
index 9ed6e0fb04..ff2e2028af 100644
--- a/block/replication.c
+++ b/block/replication.c
@@ -568,7 +568,7 @@ static void replication_start(ReplicationState *rs, ReplicationMode mode,
job = backup_job_create(NULL, s->secondary_disk->bs, s->hidden_disk->bs,
0, MIRROR_SYNC_MODE_NONE, NULL, false,
BLOCKDEV_ON_ERROR_REPORT,
- BLOCKDEV_ON_ERROR_REPORT, BLOCK_JOB_INTERNAL,
+ BLOCKDEV_ON_ERROR_REPORT, JOB_INTERNAL,
backup_job_completed, bs, NULL, &local_err);
if (local_err) {
error_propagate(errp, local_err);
@@ -693,7 +693,7 @@ static void replication_stop(ReplicationState *rs, bool failover, Error **errp)
s->stage = BLOCK_REPLICATION_FAILOVER;
commit_active_start(NULL, s->active_disk->bs, s->secondary_disk->bs,
- BLOCK_JOB_INTERNAL, 0, BLOCKDEV_ON_ERROR_REPORT,
+ JOB_INTERNAL, 0, BLOCKDEV_ON_ERROR_REPORT,
NULL, replication_done, bs, true, errp);
break;
default:
diff --git a/block/stream.c b/block/stream.c
index e81b488a22..eee02538ed 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -242,7 +242,7 @@ void stream_start(const char *job_id, BlockDriverState *bs,
BLK_PERM_GRAPH_MOD,
BLK_PERM_CONSISTENT_READ | BLK_PERM_WRITE_UNCHANGED |
BLK_PERM_WRITE,
- speed, BLOCK_JOB_DEFAULT, NULL, NULL, errp);
+ speed, JOB_DEFAULT, NULL, NULL, errp);
if (!s) {
goto fail;
}
diff --git a/blockdev.c b/blockdev.c
index 3533c0dd6a..278b92ce03 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -3244,7 +3244,7 @@ void qmp_block_commit(bool has_job_id, const char *job_id, const char *device,
goto out;
}
commit_active_start(has_job_id ? job_id : NULL, bs, base_bs,
- BLOCK_JOB_DEFAULT, speed, on_error,
+ JOB_DEFAULT, speed, on_error,
filter_node_name, NULL, NULL, false, &local_err);
} else {
BlockDriverState *overlay_bs = bdrv_find_overlay(bs, top_bs);
@@ -3275,7 +3275,7 @@ static BlockJob *do_drive_backup(DriveBackup *backup, BlockJobTxn *txn,
AioContext *aio_context;
QDict *options = NULL;
Error *local_err = NULL;
- int flags, job_flags = BLOCK_JOB_DEFAULT;
+ int flags, job_flags = JOB_DEFAULT;
int64_t size;
bool set_backing_hd = false;
@@ -3398,10 +3398,10 @@ static BlockJob *do_drive_backup(DriveBackup *backup, BlockJobTxn *txn,
}
}
if (!backup->auto_finalize) {
- job_flags |= BLOCK_JOB_MANUAL_FINALIZE;
+ job_flags |= JOB_MANUAL_FINALIZE;
}
if (!backup->auto_dismiss) {
- job_flags |= BLOCK_JOB_MANUAL_DISMISS;
+ job_flags |= JOB_MANUAL_DISMISS;
}
job = backup_job_create(backup->job_id, bs, target_bs, backup->speed,
@@ -3442,7 +3442,7 @@ BlockJob *do_blockdev_backup(BlockdevBackup *backup, BlockJobTxn *txn,
Error *local_err = NULL;
AioContext *aio_context;
BlockJob *job = NULL;
- int job_flags = BLOCK_JOB_DEFAULT;
+ int job_flags = JOB_DEFAULT;
if (!backup->has_speed) {
backup->speed = 0;
@@ -3491,10 +3491,10 @@ BlockJob *do_blockdev_backup(BlockdevBackup *backup, BlockJobTxn *txn,
}
}
if (!backup->auto_finalize) {
- job_flags |= BLOCK_JOB_MANUAL_FINALIZE;
+ job_flags |= JOB_MANUAL_FINALIZE;
}
if (!backup->auto_dismiss) {
- job_flags |= BLOCK_JOB_MANUAL_DISMISS;
+ job_flags |= JOB_MANUAL_DISMISS;
}
job = backup_job_create(backup->job_id, bs, target_bs, backup->speed,
backup->sync, NULL, backup->compress,
diff --git a/blockjob.c b/blockjob.c
index a1d1f488e9..d9d8ff7f5c 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -285,7 +285,7 @@ static void block_job_do_dismiss(BlockJob *job)
static void block_job_conclude(BlockJob *job)
{
job_state_transition(&job->job, JOB_STATUS_CONCLUDED);
- if (job->auto_dismiss || !job_started(&job->job)) {
+ if (job->job.auto_dismiss || !job_started(&job->job)) {
block_job_do_dismiss(job);
}
}
@@ -483,7 +483,7 @@ static void block_job_completed_txn_abort(BlockJob *job)
static int block_job_needs_finalize(BlockJob *job)
{
- return !job->auto_finalize;
+ return !job->job.auto_finalize;
}
static void block_job_do_finalize(BlockJob *job)
@@ -688,8 +688,8 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
info->io_status = job->iostatus;
info->ready = job->ready;
info->status = job->job.status;
- info->auto_finalize = job->auto_finalize;
- info->auto_dismiss = job->auto_dismiss;
+ info->auto_finalize = job->job.auto_finalize;
+ info->auto_dismiss = job->job.auto_dismiss;
info->has_error = job->ret != 0;
info->error = job->ret ? g_strdup(strerror(-job->ret)) : NULL;
return info;
@@ -736,7 +736,7 @@ static void block_job_event_completed(BlockJob *job, const char *msg)
static int block_job_event_pending(BlockJob *job)
{
job_state_transition(&job->job, JOB_STATUS_PENDING);
- if (!job->auto_finalize && !block_job_is_internal(job)) {
+ if (!job->job.auto_finalize && !block_job_is_internal(job)) {
qapi_event_send_block_job_pending(job_type(&job->job),
job->job.id,
&error_abort);
@@ -763,19 +763,8 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
return NULL;
}
- if (job_id == NULL && !(flags & BLOCK_JOB_INTERNAL)) {
+ if (job_id == NULL && !(flags & JOB_INTERNAL)) {
job_id = bdrv_get_device_name(bs);
- if (!*job_id) {
- error_setg(errp, "An explicit job ID is required for this node");
- return NULL;
- }
- }
-
- if (job_id) {
- if (flags & BLOCK_JOB_INTERNAL) {
- error_setg(errp, "Cannot specify job ID for internal block job");
- return NULL;
- }
}
blk = blk_new(perm, shared_perm);
@@ -786,7 +775,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
}
job = job_create(job_id, &driver->job_driver, blk_get_aio_context(blk),
- errp);
+ flags, errp);
if (job == NULL) {
blk_unref(blk);
return NULL;
@@ -800,8 +789,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
job->blk = blk;
job->cb = cb;
job->opaque = opaque;
- job->auto_finalize = !(flags & BLOCK_JOB_MANUAL_FINALIZE);
- job->auto_dismiss = !(flags & BLOCK_JOB_MANUAL_DISMISS);
error_setg(&job->blocker, "block device is in use by block job: %s",
job_type_str(&job->job));
diff --git a/job.c b/job.c
index 1e432af0d6..743d38d0d9 100644
--- a/job.c
+++ b/job.c
@@ -182,11 +182,15 @@ static void job_sleep_timer_cb(void *opaque)
}
void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
- Error **errp)
+ int flags, Error **errp)
{
Job *job;
if (job_id) {
+ if (flags & JOB_INTERNAL) {
+ error_setg(errp, "Cannot specify job ID for internal job");
+ return NULL;
+ }
if (!id_wellformed(job_id)) {
error_setg(errp, "Invalid job ID '%s'", job_id);
return NULL;
@@ -195,6 +199,9 @@ void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
error_setg(errp, "Job ID '%s' already in use", job_id);
return NULL;
}
+ } else if (!(flags & JOB_INTERNAL)) {
+ error_setg(errp, "An explicit job ID is required");
+ return NULL;
}
job = g_malloc0(driver->instance_size);
@@ -205,6 +212,8 @@ void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
job->busy = false;
job->paused = true;
job->pause_count = 1;
+ job->auto_finalize = !(flags & JOB_MANUAL_FINALIZE);
+ job->auto_dismiss = !(flags & JOB_MANUAL_DISMISS);
job_state_transition(job, JOB_STATUS_CREATED);
aio_timer_init(qemu_get_aio_context(), &job->sleep_timer,
diff --git a/qemu-img.c b/qemu-img.c
index e6a689c794..12b5818641 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -1015,7 +1015,7 @@ static int img_commit(int argc, char **argv)
aio_context = bdrv_get_aio_context(bs);
aio_context_acquire(aio_context);
- commit_active_start("commit", bs, base_bs, BLOCK_JOB_DEFAULT, 0,
+ commit_active_start("commit", bs, base_bs, JOB_DEFAULT, 0,
BLOCKDEV_ON_ERROR_REPORT, NULL, common_block_job_cb,
&cbi, false, &local_err);
aio_context_release(aio_context);
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 93d1ff0859..60e9fa2298 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -107,7 +107,7 @@ static BlockJob *test_block_job_start(unsigned int iterations,
snprintf(job_id, sizeof(job_id), "job%u", counter++);
s = block_job_create(job_id, &test_block_job_driver, txn, bs,
- 0, BLK_PERM_ALL, 0, BLOCK_JOB_DEFAULT,
+ 0, BLK_PERM_ALL, 0, JOB_DEFAULT,
test_block_job_cb, data, &error_abort);
s->iterations = iterations;
s->use_timer = use_timer;
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index ceb59600ed..8bb0aa8f85 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -59,7 +59,7 @@ static BlockJob *do_test_id(BlockBackend *blk, const char *id,
bool should_succeed)
{
return mk_job(blk, id, &test_block_job_driver,
- should_succeed, BLOCK_JOB_DEFAULT);
+ should_succeed, JOB_DEFAULT);
}
/* This creates a BlockBackend (optionally with a name) with a
@@ -214,7 +214,7 @@ static CancelJob *create_common(BlockJob **pjob)
blk = create_blk(NULL);
job = mk_job(blk, "Steve", &test_cancel_driver, true,
- BLOCK_JOB_MANUAL_FINALIZE | BLOCK_JOB_MANUAL_DISMISS);
+ JOB_MANUAL_FINALIZE | JOB_MANUAL_DISMISS);
job_ref(&job->job);
assert(job->job.status == JOB_STATUS_CREATED);
s = container_of(job, CancelJob, common);
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 17/40] job: Move BlockJobCreateFlags to Job
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 17/40] job: Move BlockJobCreateFlags to Job Kevin Wolf
@ 2018-05-23 22:24 ` John Snow
2018-05-24 8:17 ` Kevin Wolf
0 siblings, 1 reply; 86+ messages in thread
From: John Snow @ 2018-05-23 22:24 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: qemu-devel, jcody, armbru, mreitz
On 05/18/2018 09:20 AM, Kevin Wolf wrote:
> + job->auto_finalize = !(flags & JOB_MANUAL_FINALIZE);
> + job->auto_dismiss = !(flags & JOB_MANUAL_DISMISS);
Job API might be a good chance to say "No, this is the default behavior
for this API."
I don't know how possible this is, but could we remove these behavior
flags for jobs (but keep them for block jobs), and then any legacy block
job creation interfaces we have can enable/disable them as the user
requested,
and the block job layer itself has hooks that persuade the core job
layer to automatically transition without user input, if appropriate.
(Unless that happens later?)
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 17/40] job: Move BlockJobCreateFlags to Job
2018-05-23 22:24 ` John Snow
@ 2018-05-24 8:17 ` Kevin Wolf
2018-05-24 17:46 ` John Snow
0 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-24 8:17 UTC (permalink / raw)
To: John Snow; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
Am 24.05.2018 um 00:24 hat John Snow geschrieben:
>
>
> On 05/18/2018 09:20 AM, Kevin Wolf wrote:
> > + job->auto_finalize = !(flags & JOB_MANUAL_FINALIZE);
> > + job->auto_dismiss = !(flags & JOB_MANUAL_DISMISS);
>
> Job API might be a good chance to say "No, this is the default behavior
> for this API."
>
> I don't know how possible this is, but could we remove these behavior
> flags for jobs (but keep them for block jobs), and then any legacy block
> job creation interfaces we have can enable/disable them as the user
> requested,
>
> and the block job layer itself has hooks that persuade the core job
> layer to automatically transition without user input, if appropriate.
>
> (Unless that happens later?)
That's really a question for job-create, which we don't have yet. I'm
not sure whether job-create would expose those options, and if so, what
the defaults would be.
I think auto-dismiss=false is a good default because the reason for
having a manual job-dismiss is shared by all jobs (you want to be able
to query the job result after it finished).
auto-finalize is different. We introduced it to avoid surprise graph
changes, but this implies that every job changes the block graph when it
completes. This is obviously not true for non-block jobs. They might
still commonly change something else that shouldn't be done as a
surprise, but I am working right now on at least a single job type that
doesn't do anything like that on completion (blockdev-create).
For now, I made it unconditionally auto-finalize=true and
auto-dismiss=false without offering the options in QMP, but that's open
for discussion.
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 17/40] job: Move BlockJobCreateFlags to Job
2018-05-24 8:17 ` Kevin Wolf
@ 2018-05-24 17:46 ` John Snow
0 siblings, 0 replies; 86+ messages in thread
From: John Snow @ 2018-05-24 17:46 UTC (permalink / raw)
To: Kevin Wolf; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
On 05/24/2018 04:17 AM, Kevin Wolf wrote:
> Am 24.05.2018 um 00:24 hat John Snow geschrieben:
>>
>>
>> On 05/18/2018 09:20 AM, Kevin Wolf wrote:
>>> + job->auto_finalize = !(flags & JOB_MANUAL_FINALIZE);
>>> + job->auto_dismiss = !(flags & JOB_MANUAL_DISMISS);
>>
>> Job API might be a good chance to say "No, this is the default behavior
>> for this API."
>>
>> I don't know how possible this is, but could we remove these behavior
>> flags for jobs (but keep them for block jobs), and then any legacy block
>> job creation interfaces we have can enable/disable them as the user
>> requested,
>>
>> and the block job layer itself has hooks that persuade the core job
>> layer to automatically transition without user input, if appropriate.
>>
>> (Unless that happens later?)
>
> That's really a question for job-create, which we don't have yet. I'm
> not sure whether job-create would expose those options, and if so, what
> the defaults would be.
>
> I think auto-dismiss=false is a good default because the reason for
> having a manual job-dismiss is shared by all jobs (you want to be able
> to query the job result after it finished).
>
Sounds good to me.
> auto-finalize is different. We introduced it to avoid surprise graph
> changes, but this implies that every job changes the block graph when it
> completes. This is obviously not true for non-block jobs. They might
> still commonly change something else that shouldn't be done as a
> surprise, but I am working right now on at least a single job type that
> doesn't do anything like that on completion (blockdev-create).
>
I'll have to stay tuned.
> For now, I made it unconditionally auto-finalize=true and
> auto-dismiss=false without offering the options in QMP, but that's open
> for discussion.
Sure. I was wondering if we didn't want to change the internal API's
defaults around, but I guess that can happen whenever based on what the
QMP looks like.
>
> Kevin
>
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 18/40] blockjob: Split block_job_event_pending()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (16 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 17/40] job: Move BlockJobCreateFlags to Job Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 19/40] job: Add job_event_*() Kevin Wolf
` (24 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
block_job_event_pending() doesn't only send a QMP event, but it also
transitions to the PENDING state. Split the function so that we get one
part only sending the event (like other block_job_event_* functions) and
another part that does the state transition.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
blockjob.c | 27 ++++++++++++++++++---------
1 file changed, 18 insertions(+), 9 deletions(-)
diff --git a/blockjob.c b/blockjob.c
index d9d8ff7f5c..b4334fb1bf 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -38,7 +38,7 @@
static void block_job_event_cancelled(BlockJob *job);
static void block_job_event_completed(BlockJob *job, const char *msg);
-static int block_job_event_pending(BlockJob *job);
+static void block_job_event_pending(BlockJob *job);
/* Transactional group of block jobs */
struct BlockJobTxn {
@@ -500,6 +500,15 @@ static void block_job_do_finalize(BlockJob *job)
}
}
+static int block_job_transition_to_pending(BlockJob *job)
+{
+ job_state_transition(&job->job, JOB_STATUS_PENDING);
+ if (!job->job.auto_finalize) {
+ block_job_event_pending(job);
+ }
+ return 0;
+}
+
static void block_job_completed_txn_success(BlockJob *job)
{
BlockJobTxn *txn = job->txn;
@@ -518,7 +527,7 @@ static void block_job_completed_txn_success(BlockJob *job)
assert(other_job->ret == 0);
}
- block_job_txn_apply(txn, block_job_event_pending, false);
+ block_job_txn_apply(txn, block_job_transition_to_pending, false);
/* If no jobs need manual finalization, automatically do so */
if (block_job_txn_apply(txn, block_job_needs_finalize, false) == 0) {
@@ -733,15 +742,15 @@ static void block_job_event_completed(BlockJob *job, const char *msg)
&error_abort);
}
-static int block_job_event_pending(BlockJob *job)
+static void block_job_event_pending(BlockJob *job)
{
- job_state_transition(&job->job, JOB_STATUS_PENDING);
- if (!job->job.auto_finalize && !block_job_is_internal(job)) {
- qapi_event_send_block_job_pending(job_type(&job->job),
- job->job.id,
- &error_abort);
+ if (block_job_is_internal(job)) {
+ return;
}
- return 0;
+
+ qapi_event_send_block_job_pending(job_type(&job->job),
+ job->job.id,
+ &error_abort);
}
/*
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 19/40] job: Add job_event_*()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (17 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 18/40] blockjob: Split block_job_event_pending() Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 20/40] job: Move single job finalisation to Job Kevin Wolf
` (23 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
Go through the Job layer in order to send QMP events. For the moment,
these functions only call a notifier in the BlockJob layer that sends
the existing commands.
This uses notifiers rather than JobDriver callbacks because internal
users of jobs won't receive QMP events, but might still be interested
in getting notified for the events.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 9 +++++++++
include/qemu/job.h | 18 ++++++++++++++++++
blockjob.c | 41 +++++++++++++++++++++++++++--------------
job.c | 19 +++++++++++++++++++
4 files changed, 73 insertions(+), 14 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index f9aaaaa835..aef06295f6 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -82,6 +82,15 @@ typedef struct BlockJob {
/** Block other operations when block job is running */
Error *blocker;
+ /** Called when a cancelled job is finalised. */
+ Notifier finalize_cancelled_notifier;
+
+ /** Called when a successfully completed job is finalised. */
+ Notifier finalize_completed_notifier;
+
+ /** Called when the job transitions to PENDING */
+ Notifier pending_notifier;
+
/** BlockDriverStates that are involved in this block job */
GSList *nodes;
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 9783e4049b..14d93778f3 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -105,6 +105,15 @@ typedef struct Job {
/** True if this job should automatically dismiss itself */
bool auto_dismiss;
+ /** Notifiers called when a cancelled job is finalised */
+ NotifierList on_finalize_cancelled;
+
+ /** Notifiers called when a successfully completed job is finalised */
+ NotifierList on_finalize_completed;
+
+ /** Notifiers called when the job transitions to PENDING */
+ NotifierList on_pending;
+
/** Element of the list of jobs */
QLIST_ENTRY(Job) job_list;
} Job;
@@ -182,6 +191,15 @@ void job_ref(Job *job);
*/
void job_unref(Job *job);
+/** To be called when a cancelled job is finalised. */
+void job_event_cancelled(Job *job);
+
+/** To be called when a successfully completed job is finalised. */
+void job_event_completed(Job *job);
+
+/** To be called when the job transitions to PENDING */
+void job_event_pending(Job *job);
+
/**
* Conditionally enter the job coroutine if the job is ready to run, not
* already busy and fn() returns true. fn() is called while under the job_lock
diff --git a/blockjob.c b/blockjob.c
index b4334fb1bf..05d7921b3f 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -36,10 +36,6 @@
#include "qemu/coroutine.h"
#include "qemu/timer.h"
-static void block_job_event_cancelled(BlockJob *job);
-static void block_job_event_completed(BlockJob *job, const char *msg);
-static void block_job_event_pending(BlockJob *job);
-
/* Transactional group of block jobs */
struct BlockJobTxn {
@@ -352,13 +348,9 @@ static int block_job_finalize_single(BlockJob *job)
/* Emit events only if we actually started */
if (job_started(&job->job)) {
if (job_is_cancelled(&job->job)) {
- block_job_event_cancelled(job);
+ job_event_cancelled(&job->job);
} else {
- const char *msg = NULL;
- if (job->ret < 0) {
- msg = strerror(-job->ret);
- }
- block_job_event_completed(job, msg);
+ job_event_completed(&job->job);
}
}
@@ -504,7 +496,7 @@ static int block_job_transition_to_pending(BlockJob *job)
{
job_state_transition(&job->job, JOB_STATUS_PENDING);
if (!job->job.auto_finalize) {
- block_job_event_pending(job);
+ job_event_pending(&job->job);
}
return 0;
}
@@ -712,8 +704,10 @@ static void block_job_iostatus_set_err(BlockJob *job, int error)
}
}
-static void block_job_event_cancelled(BlockJob *job)
+static void block_job_event_cancelled(Notifier *n, void *opaque)
{
+ BlockJob *job = opaque;
+
if (block_job_is_internal(job)) {
return;
}
@@ -726,12 +720,19 @@ static void block_job_event_cancelled(BlockJob *job)
&error_abort);
}
-static void block_job_event_completed(BlockJob *job, const char *msg)
+static void block_job_event_completed(Notifier *n, void *opaque)
{
+ BlockJob *job = opaque;
+ const char *msg = NULL;
+
if (block_job_is_internal(job)) {
return;
}
+ if (job->ret < 0) {
+ msg = strerror(-job->ret);
+ }
+
qapi_event_send_block_job_completed(job_type(&job->job),
job->job.id,
job->len,
@@ -742,8 +743,10 @@ static void block_job_event_completed(BlockJob *job, const char *msg)
&error_abort);
}
-static void block_job_event_pending(BlockJob *job)
+static void block_job_event_pending(Notifier *n, void *opaque)
{
+ BlockJob *job = opaque;
+
if (block_job_is_internal(job)) {
return;
}
@@ -799,6 +802,16 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
job->cb = cb;
job->opaque = opaque;
+ job->finalize_cancelled_notifier.notify = block_job_event_cancelled;
+ job->finalize_completed_notifier.notify = block_job_event_completed;
+ job->pending_notifier.notify = block_job_event_pending;
+
+ notifier_list_add(&job->job.on_finalize_cancelled,
+ &job->finalize_cancelled_notifier);
+ notifier_list_add(&job->job.on_finalize_completed,
+ &job->finalize_completed_notifier);
+ notifier_list_add(&job->job.on_pending, &job->pending_notifier);
+
error_setg(&job->blocker, "block device is in use by block job: %s",
job_type_str(&job->job));
block_job_add_bdrv(job, "main node", bs, 0, BLK_PERM_ALL, &error_abort);
diff --git a/job.c b/job.c
index 743d38d0d9..748354b479 100644
--- a/job.c
+++ b/job.c
@@ -215,6 +215,10 @@ void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
job->auto_finalize = !(flags & JOB_MANUAL_FINALIZE);
job->auto_dismiss = !(flags & JOB_MANUAL_DISMISS);
+ notifier_list_init(&job->on_finalize_cancelled);
+ notifier_list_init(&job->on_finalize_completed);
+ notifier_list_init(&job->on_pending);
+
job_state_transition(job, JOB_STATUS_CREATED);
aio_timer_init(qemu_get_aio_context(), &job->sleep_timer,
QEMU_CLOCK_REALTIME, SCALE_NS,
@@ -247,6 +251,21 @@ void job_unref(Job *job)
}
}
+void job_event_cancelled(Job *job)
+{
+ notifier_list_notify(&job->on_finalize_cancelled, job);
+}
+
+void job_event_completed(Job *job)
+{
+ notifier_list_notify(&job->on_finalize_completed, job);
+}
+
+void job_event_pending(Job *job)
+{
+ notifier_list_notify(&job->on_pending, job);
+}
+
void job_enter_cond(Job *job, bool(*fn)(Job *job))
{
if (!job_started(job)) {
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 20/40] job: Move single job finalisation to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (18 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 19/40] job: Add job_event_*() Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 18:00 ` Eric Blake
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 21/40] job: Convert block_job_cancel_async() " Kevin Wolf
` (22 subsequent siblings)
42 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves the finalisation of a single job from BlockJob to Job.
Some part of this code depends on job transactions, and job transactions
call this code, we introduce some temporary calls from Job functions to
BlockJob ones. This will be fixed once transactions move to Job, too.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 9 ---
include/block/blockjob_int.h | 36 -----------
include/qemu/job.h | 53 +++++++++++++++-
block/backup.c | 22 +++----
block/commit.c | 2 +-
block/mirror.c | 2 +-
blockjob.c | 142 ++++++++-----------------------------------
job.c | 100 +++++++++++++++++++++++++++++-
qemu-img.c | 2 +-
tests/test-blockjob.c | 10 +--
10 files changed, 194 insertions(+), 184 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index aef06295f6..3f405d1fa7 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -76,9 +76,6 @@ typedef struct BlockJob {
/** Rate limiting data structure for implementing @speed. */
RateLimit limit;
- /** The completion function that will be called when the job completes. */
- BlockCompletionFunc *cb;
-
/** Block other operations when block job is running */
Error *blocker;
@@ -94,12 +91,6 @@ typedef struct BlockJob {
/** BlockDriverStates that are involved in this block job */
GSList *nodes;
- /** The opaque value that is passed to the completion function. */
- void *opaque;
-
- /** ret code passed to block_job_completed. */
- int ret;
-
BlockJobTxn *txn;
QLIST_ENTRY(BlockJob) txn_list;
} BlockJob;
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 88639f7efc..bf2b762808 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -54,34 +54,6 @@ struct BlockJobDriver {
*/
int (*prepare)(BlockJob *job);
- /**
- * If the callback is not NULL, it will be invoked when all the jobs
- * belonging to the same transaction complete; or upon this job's
- * completion if it is not in a transaction. Skipped if NULL.
- *
- * All jobs will complete with a call to either .commit() or .abort() but
- * never both.
- */
- void (*commit)(BlockJob *job);
-
- /**
- * If the callback is not NULL, it will be invoked when any job in the
- * same transaction fails; or upon this job's failure (due to error or
- * cancellation) if it is not in a transaction. Skipped if NULL.
- *
- * All jobs will complete with a call to either .commit() or .abort() but
- * never both.
- */
- void (*abort)(BlockJob *job);
-
- /**
- * If the callback is not NULL, it will be invoked after a call to either
- * .commit() or .abort(). Regardless of which callback is invoked after
- * completion, .clean() will always be called, even if the job does not
- * belong to a transaction group.
- */
- void (*clean)(BlockJob *job);
-
/*
* If the callback is not NULL, it will be invoked before the job is
* resumed in a new AioContext. This is the place to move any resources
@@ -156,14 +128,6 @@ void block_job_yield(BlockJob *job);
int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n);
/**
- * block_job_early_fail:
- * @bs: The block device.
- *
- * The block job could not be started, free it.
- */
-void block_job_early_fail(BlockJob *job);
-
-/**
* block_job_completed:
* @job: The job being completed.
* @ret: The status code.
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 14d93778f3..3e817beee9 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -29,6 +29,7 @@
#include "qapi/qapi-types-block-core.h"
#include "qemu/queue.h"
#include "qemu/coroutine.h"
+#include "block/aio.h"
typedef struct JobDriver JobDriver;
@@ -105,6 +106,15 @@ typedef struct Job {
/** True if this job should automatically dismiss itself */
bool auto_dismiss;
+ /** ret code passed to block_job_completed. */
+ int ret;
+
+ /** The completion function that will be called when the job completes. */
+ BlockCompletionFunc *cb;
+
+ /** The opaque value that is passed to the completion function. */
+ void *opaque;
+
/** Notifiers called when a cancelled job is finalised */
NotifierList on_finalize_cancelled;
@@ -151,6 +161,35 @@ struct JobDriver {
*/
void (*user_resume)(Job *job);
+ /**
+ * If the callback is not NULL, it will be invoked when all the jobs
+ * belonging to the same transaction complete; or upon this job's
+ * completion if it is not in a transaction. Skipped if NULL.
+ *
+ * All jobs will complete with a call to either .commit() or .abort() but
+ * never both.
+ */
+ void (*commit)(Job *job);
+
+ /**
+ * If the callback is not NULL, it will be invoked when any job in the
+ * same transaction fails; or upon this job's failure (due to error or
+ * cancellation) if it is not in a transaction. Skipped if NULL.
+ *
+ * All jobs will complete with a call to either .commit() or .abort() but
+ * never both.
+ */
+ void (*abort)(Job *job);
+
+ /**
+ * If the callback is not NULL, it will be invoked after a call to either
+ * .commit() or .abort(). Regardless of which callback is invoked after
+ * completion, .clean() will always be called, even if the job does not
+ * belong to a transaction group.
+ */
+ void (*clean)(Job *job);
+
+
/** Called when the job is freed */
void (*free)(Job *job);
};
@@ -174,10 +213,12 @@ typedef enum JobCreateFlags {
* @driver: The class object for the newly-created job.
* @ctx: The AioContext to run the job coroutine in.
* @flags: Creation flags for the job. See @JobCreateFlags.
+ * @cb: Completion function for the job.
+ * @opaque: Opaque pointer value passed to @cb.
* @errp: Error object.
*/
void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
- int flags, Error **errp);
+ int flags, BlockCompletionFunc *cb, void *opaque, Error **errp);
/**
* Add a reference to Job refcnt, it will be decreased with job_unref, and then
@@ -300,6 +341,10 @@ Job *job_get(const char *id);
*/
int job_apply_verb(Job *job, JobVerb verb, Error **errp);
+/** The @job could not be started, free it. */
+void job_early_fail(Job *job);
+
+
typedef void JobDeferToMainLoopFn(Job *job, void *opaque);
/**
@@ -322,5 +367,11 @@ void job_state_transition(Job *job, JobStatus s1);
void coroutine_fn job_do_yield(Job *job, uint64_t ns);
bool job_should_pause(Job *job);
bool job_started(Job *job);
+void job_do_dismiss(Job *job);
+int job_finalize_single(Job *job);
+void job_update_rc(Job *job);
+
+typedef struct BlockJob BlockJob;
+void block_job_txn_del_job(BlockJob *job);
#endif
diff --git a/block/backup.c b/block/backup.c
index 4d011d5a5c..bd31282924 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -207,25 +207,25 @@ static void backup_cleanup_sync_bitmap(BackupBlockJob *job, int ret)
}
}
-static void backup_commit(BlockJob *job)
+static void backup_commit(Job *job)
{
- BackupBlockJob *s = container_of(job, BackupBlockJob, common);
+ BackupBlockJob *s = container_of(job, BackupBlockJob, common.job);
if (s->sync_bitmap) {
backup_cleanup_sync_bitmap(s, 0);
}
}
-static void backup_abort(BlockJob *job)
+static void backup_abort(Job *job)
{
- BackupBlockJob *s = container_of(job, BackupBlockJob, common);
+ BackupBlockJob *s = container_of(job, BackupBlockJob, common.job);
if (s->sync_bitmap) {
backup_cleanup_sync_bitmap(s, -1);
}
}
-static void backup_clean(BlockJob *job)
+static void backup_clean(Job *job)
{
- BackupBlockJob *s = container_of(job, BackupBlockJob, common);
+ BackupBlockJob *s = container_of(job, BackupBlockJob, common.job);
assert(s->target);
blk_unref(s->target);
s->target = NULL;
@@ -530,10 +530,10 @@ static const BlockJobDriver backup_job_driver = {
.free = block_job_free,
.user_resume = block_job_user_resume,
.start = backup_run,
+ .commit = backup_commit,
+ .abort = backup_abort,
+ .clean = backup_clean,
},
- .commit = backup_commit,
- .abort = backup_abort,
- .clean = backup_clean,
.attached_aio_context = backup_attached_aio_context,
.drain = backup_drain,
};
@@ -678,8 +678,8 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
bdrv_reclaim_dirty_bitmap(bs, sync_bitmap, NULL);
}
if (job) {
- backup_clean(&job->common);
- block_job_early_fail(&job->common);
+ backup_clean(&job->common.job);
+ job_early_fail(&job->common.job);
}
return NULL;
diff --git a/block/commit.c b/block/commit.c
index 7a6ae59d42..e53b2d7d55 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -385,7 +385,7 @@ fail:
if (commit_top_bs) {
bdrv_replace_node(commit_top_bs, top, &error_abort);
}
- block_job_early_fail(&s->common);
+ job_early_fail(&s->common.job);
}
diff --git a/block/mirror.c b/block/mirror.c
index 5091e72554..e9a90ea730 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -1257,7 +1257,7 @@ fail:
g_free(s->replaces);
blk_unref(s->target);
- block_job_early_fail(&s->common);
+ job_early_fail(&s->common.job);
}
bdrv_child_try_set_perm(mirror_top_bs->backing, 0, BLK_PERM_ALL,
diff --git a/blockjob.c b/blockjob.c
index 05d7921b3f..34c57da304 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -127,7 +127,7 @@ void block_job_txn_add_job(BlockJobTxn *txn, BlockJob *job)
block_job_txn_ref(txn);
}
-static void block_job_txn_del_job(BlockJob *job)
+void block_job_txn_del_job(BlockJob *job)
{
if (job->txn) {
QLIST_REMOVE(job, txn_list);
@@ -262,101 +262,12 @@ const BlockJobDriver *block_job_driver(BlockJob *job)
return job->driver;
}
-static void block_job_decommission(BlockJob *job)
-{
- assert(job);
- job->job.busy = false;
- job->job.paused = false;
- job->job.deferred_to_main_loop = true;
- block_job_txn_del_job(job);
- job_state_transition(&job->job, JOB_STATUS_NULL);
- job_unref(&job->job);
-}
-
-static void block_job_do_dismiss(BlockJob *job)
-{
- block_job_decommission(job);
-}
-
-static void block_job_conclude(BlockJob *job)
-{
- job_state_transition(&job->job, JOB_STATUS_CONCLUDED);
- if (job->job.auto_dismiss || !job_started(&job->job)) {
- block_job_do_dismiss(job);
- }
-}
-
-static void block_job_update_rc(BlockJob *job)
-{
- if (!job->ret && job_is_cancelled(&job->job)) {
- job->ret = -ECANCELED;
- }
- if (job->ret) {
- job_state_transition(&job->job, JOB_STATUS_ABORTING);
- }
-}
-
static int block_job_prepare(BlockJob *job)
{
- if (job->ret == 0 && job->driver->prepare) {
- job->ret = job->driver->prepare(job);
- }
- return job->ret;
-}
-
-static void block_job_commit(BlockJob *job)
-{
- assert(!job->ret);
- if (job->driver->commit) {
- job->driver->commit(job);
- }
-}
-
-static void block_job_abort(BlockJob *job)
-{
- assert(job->ret);
- if (job->driver->abort) {
- job->driver->abort(job);
- }
-}
-
-static void block_job_clean(BlockJob *job)
-{
- if (job->driver->clean) {
- job->driver->clean(job);
+ if (job->job.ret == 0 && job->driver->prepare) {
+ job->job.ret = job->driver->prepare(job);
}
-}
-
-static int block_job_finalize_single(BlockJob *job)
-{
- assert(job_is_completed(&job->job));
-
- /* Ensure abort is called for late-transactional failures */
- block_job_update_rc(job);
-
- if (!job->ret) {
- block_job_commit(job);
- } else {
- block_job_abort(job);
- }
- block_job_clean(job);
-
- if (job->cb) {
- job->cb(job->opaque, job->ret);
- }
-
- /* Emit events only if we actually started */
- if (job_started(&job->job)) {
- if (job_is_cancelled(&job->job)) {
- job_event_cancelled(&job->job);
- } else {
- job_event_completed(&job->job);
- }
- }
-
- block_job_txn_del_job(job);
- block_job_conclude(job);
- return 0;
+ return job->job.ret;
}
static void block_job_cancel_async(BlockJob *job, bool force)
@@ -424,8 +335,8 @@ static int block_job_finish_sync(BlockJob *job,
while (!job_is_completed(&job->job)) {
aio_poll(qemu_get_aio_context(), true);
}
- ret = (job_is_cancelled(&job->job) && job->ret == 0)
- ? -ECANCELED : job->ret;
+ ret = (job_is_cancelled(&job->job) && job->job.ret == 0)
+ ? -ECANCELED : job->job.ret;
job_unref(&job->job);
return ret;
}
@@ -466,7 +377,7 @@ static void block_job_completed_txn_abort(BlockJob *job)
assert(job_is_cancelled(&other_job->job));
block_job_finish_sync(other_job, NULL, NULL);
}
- block_job_finalize_single(other_job);
+ job_finalize_single(&other_job->job);
aio_context_release(ctx);
}
@@ -478,6 +389,11 @@ static int block_job_needs_finalize(BlockJob *job)
return !job->job.auto_finalize;
}
+static int block_job_finalize_single(BlockJob *job)
+{
+ return job_finalize_single(&job->job);
+}
+
static void block_job_do_finalize(BlockJob *job)
{
int rc;
@@ -516,7 +432,7 @@ static void block_job_completed_txn_success(BlockJob *job)
if (!job_is_completed(&other_job->job)) {
return;
}
- assert(other_job->ret == 0);
+ assert(other_job->job.ret == 0);
}
block_job_txn_apply(txn, block_job_transition_to_pending, false);
@@ -601,14 +517,14 @@ void block_job_dismiss(BlockJob **jobptr, Error **errp)
return;
}
- block_job_do_dismiss(job);
+ job_do_dismiss(&job->job);
*jobptr = NULL;
}
void block_job_cancel(BlockJob *job, bool force)
{
if (job->job.status == JOB_STATUS_CONCLUDED) {
- block_job_do_dismiss(job);
+ job_do_dismiss(&job->job);
return;
}
block_job_cancel_async(job, force);
@@ -691,8 +607,8 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
info->status = job->job.status;
info->auto_finalize = job->job.auto_finalize;
info->auto_dismiss = job->job.auto_dismiss;
- info->has_error = job->ret != 0;
- info->error = job->ret ? g_strdup(strerror(-job->ret)) : NULL;
+ info->has_error = job->job.ret != 0;
+ info->error = job->job.ret ? g_strdup(strerror(-job->job.ret)) : NULL;
return info;
}
@@ -729,8 +645,8 @@ static void block_job_event_completed(Notifier *n, void *opaque)
return;
}
- if (job->ret < 0) {
- msg = strerror(-job->ret);
+ if (job->job.ret < 0) {
+ msg = strerror(-job->job.ret);
}
qapi_event_send_block_job_completed(job_type(&job->job),
@@ -787,7 +703,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
}
job = job_create(job_id, &driver->job_driver, blk_get_aio_context(blk),
- flags, errp);
+ flags, cb, opaque, errp);
if (job == NULL) {
blk_unref(blk);
return NULL;
@@ -799,8 +715,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
job->driver = driver;
job->blk = blk;
- job->cb = cb;
- job->opaque = opaque;
job->finalize_cancelled_notifier.notify = block_job_event_cancelled;
job->finalize_completed_notifier.notify = block_job_event_completed;
@@ -828,7 +742,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
block_job_set_speed(job, speed, &local_err);
if (local_err) {
- block_job_early_fail(job);
+ job_early_fail(&job->job);
error_propagate(errp, local_err);
return NULL;
}
@@ -847,20 +761,14 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
return job;
}
-void block_job_early_fail(BlockJob *job)
-{
- assert(job->job.status == JOB_STATUS_CREATED);
- block_job_decommission(job);
-}
-
void block_job_completed(BlockJob *job, int ret)
{
assert(job && job->txn && !job_is_completed(&job->job));
assert(blk_bs(job->blk)->job == job);
- job->ret = ret;
- block_job_update_rc(job);
- trace_block_job_completed(job, ret, job->ret);
- if (job->ret) {
+ job->job.ret = ret;
+ job_update_rc(&job->job);
+ trace_block_job_completed(job, ret, job->job.ret);
+ if (job->job.ret) {
block_job_completed_txn_abort(job);
} else {
block_job_completed_txn_success(job);
diff --git a/job.c b/job.c
index 748354b479..f9bd8baff2 100644
--- a/job.c
+++ b/job.c
@@ -85,7 +85,7 @@ void job_state_transition(Job *job, JobStatus s1)
{
JobStatus s0 = job->status;
assert(s1 >= 0 && s1 <= JOB_STATUS__MAX);
- trace_job_state_transition(job, /* TODO re-enable: job->ret */ 0,
+ trace_job_state_transition(job, job->ret,
JobSTT[s0][s1] ? "allowed" : "disallowed",
JobStatus_str(s0), JobStatus_str(s1));
assert(JobSTT[s0][s1]);
@@ -182,7 +182,7 @@ static void job_sleep_timer_cb(void *opaque)
}
void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
- int flags, Error **errp)
+ int flags, BlockCompletionFunc *cb, void *opaque, Error **errp)
{
Job *job;
@@ -214,6 +214,8 @@ void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
job->pause_count = 1;
job->auto_finalize = !(flags & JOB_MANUAL_FINALIZE);
job->auto_dismiss = !(flags & JOB_MANUAL_DISMISS);
+ job->cb = cb;
+ job->opaque = opaque;
notifier_list_init(&job->on_finalize_cancelled);
notifier_list_init(&job->on_finalize_completed);
@@ -449,6 +451,100 @@ void job_user_resume(Job *job, Error **errp)
job_resume(job);
}
+void job_do_dismiss(Job *job)
+{
+ assert(job);
+ job->busy = false;
+ job->paused = false;
+ job->deferred_to_main_loop = true;
+
+ /* TODO Don't assume it's a BlockJob */
+ block_job_txn_del_job((BlockJob*) job);
+
+ job_state_transition(job, JOB_STATUS_NULL);
+ job_unref(job);
+}
+
+void job_early_fail(Job *job)
+{
+ assert(job->status == JOB_STATUS_CREATED);
+ job_do_dismiss(job);
+}
+
+static void job_conclude(Job *job)
+{
+ job_state_transition(job, JOB_STATUS_CONCLUDED);
+ if (job->auto_dismiss || !job_started(job)) {
+ job_do_dismiss(job);
+ }
+}
+
+void job_update_rc(Job *job)
+{
+ if (!job->ret && job_is_cancelled(job)) {
+ job->ret = -ECANCELED;
+ }
+ if (job->ret) {
+ job_state_transition(job, JOB_STATUS_ABORTING);
+ }
+}
+
+static void job_commit(Job *job)
+{
+ assert(!job->ret);
+ if (job->driver->commit) {
+ job->driver->commit(job);
+ }
+}
+
+static void job_abort(Job *job)
+{
+ assert(job->ret);
+ if (job->driver->abort) {
+ job->driver->abort(job);
+ }
+}
+
+static void job_clean(Job *job)
+{
+ if (job->driver->clean) {
+ job->driver->clean(job);
+ }
+}
+
+int job_finalize_single(Job *job)
+{
+ assert(job_is_completed(job));
+
+ /* Ensure abort is called for late-transactional failures */
+ job_update_rc(job);
+
+ if (!job->ret) {
+ job_commit(job);
+ } else {
+ job_abort(job);
+ }
+ job_clean(job);
+
+ if (job->cb) {
+ job->cb(job->opaque, job->ret);
+ }
+
+ /* Emit events only if we actually started */
+ if (job_started(job)) {
+ if (job_is_cancelled(job)) {
+ job_event_cancelled(job);
+ } else {
+ job_event_completed(job);
+ }
+ }
+
+ /* TODO Don't assume it's a BlockJob */
+ block_job_txn_del_job((BlockJob*) job);
+ job_conclude(job);
+ return 0;
+}
+
typedef struct {
Job *job;
diff --git a/qemu-img.c b/qemu-img.c
index 12b5818641..81f338273f 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -872,7 +872,7 @@ static void run_block_job(BlockJob *job, Error **errp)
if (!job_is_completed(&job->job)) {
ret = block_job_complete_sync(job, errp);
} else {
- ret = job->ret;
+ ret = job->job.ret;
}
job_unref(&job->job);
aio_context_release(aio_context);
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 8bb0aa8f85..1fe6803fe0 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -128,11 +128,11 @@ static void test_job_ids(void)
job[1] = do_test_id(blk[1], "id0", false);
/* But once job[0] finishes we can reuse its ID */
- block_job_early_fail(job[0]);
+ job_early_fail(&job[0]->job);
job[1] = do_test_id(blk[1], "id0", true);
/* No job ID specified, defaults to the backend name ('drive1') */
- block_job_early_fail(job[1]);
+ job_early_fail(&job[1]->job);
job[1] = do_test_id(blk[1], NULL, true);
/* Duplicate job ID */
@@ -145,9 +145,9 @@ static void test_job_ids(void)
/* This one is valid */
job[2] = do_test_id(blk[2], "id_2", true);
- block_job_early_fail(job[0]);
- block_job_early_fail(job[1]);
- block_job_early_fail(job[2]);
+ job_early_fail(&job[0]->job);
+ job_early_fail(&job[1]->job);
+ job_early_fail(&job[2]->job);
destroy_blk(blk[0]);
destroy_blk(blk[1]);
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 20/40] job: Move single job finalisation to Job
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 20/40] job: Move single job finalisation to Job Kevin Wolf
@ 2018-05-18 18:00 ` Eric Blake
0 siblings, 0 replies; 86+ messages in thread
From: Eric Blake @ 2018-05-18 18:00 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:20 AM, Kevin Wolf wrote:
> This moves the finalisation of a single job from BlockJob to Job.
>
> Some part of this code depends on job transactions, and job transactions
> call this code, we introduce some temporary calls from Job functions to
> BlockJob ones. This will be fixed once transactions move to Job, too.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> Reviewed-by: Max Reitz <mreitz@redhat.com>
> ---
> +++ b/job.c
> @@ -449,6 +451,100 @@ void job_user_resume(Job *job, Error **errp)
> job_resume(job);
> }
>
> +void job_do_dismiss(Job *job)
> +{
> + assert(job);
> + job->busy = false;
> + job->paused = false;
> + job->deferred_to_main_loop = true;
> +
> + /* TODO Don't assume it's a BlockJob */
> + block_job_txn_del_job((BlockJob*) job);
checkpatch flagged this for style, but it goes away later in the series.
But more than just style, this hard-codes the assumption that BlockJob
has a Job member at offset 0. So would it be better to use
container_of() instead of a bare cast, even though it is transient?
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (19 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 20/40] job: Move single job finalisation to Job Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-23 23:18 ` John Snow
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 22/40] job: Add job_drain() Kevin Wolf
` (21 subsequent siblings)
42 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
block_job_cancel_async() did two things that were still block job
specific:
* Setting job->force. This field makes sense on the Job level, so we can
just move it. While at it, rename it to job->force_cancel to make its
purpose more obvious.
* Resetting the I/O status. This can't be moved because generic Jobs
don't have an I/O status. What the function really implements is a
user resume, except without entering the coroutine. Consequently, it
makes sense to call the .user_resume driver callback here which
already resets the I/O status.
The old block_job_cancel_async() has two separate if statements that
check job->iostatus != BLOCK_DEVICE_IO_STATUS_OK and job->user_paused.
However, the former condition always implies the latter (as is
asserted in block_job_iostatus_reset()), so changing the explicit call
of block_job_iostatus_reset() on the former condition with the
.user_resume callback on the latter condition is equivalent and
doesn't need to access any BlockJob specific state.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 6 ------
include/qemu/job.h | 6 ++++++
block/mirror.c | 4 ++--
blockjob.c | 25 +++++++++++++------------
4 files changed, 21 insertions(+), 20 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 3f405d1fa7..d975efea20 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -51,12 +51,6 @@ typedef struct BlockJob {
BlockBackend *blk;
/**
- * Set to true if the job should abort immediately without waiting
- * for data to be in sync.
- */
- bool force;
-
- /**
* Set to true when the job is ready to be completed.
*/
bool ready;
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 3e817beee9..2648c74281 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -97,6 +97,12 @@ typedef struct Job {
*/
bool cancelled;
+ /**
+ * Set to true if the job should abort immediately without waiting
+ * for data to be in sync.
+ */
+ bool force_cancel;
+
/** Set to true when the job has deferred work to the main loop. */
bool deferred_to_main_loop;
diff --git a/block/mirror.c b/block/mirror.c
index e9a90ea730..c3951d1ca2 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -871,7 +871,7 @@ static void coroutine_fn mirror_run(void *opaque)
trace_mirror_before_sleep(s, cnt, s->synced, delay_ns);
job_sleep_ns(&s->common.job, delay_ns);
if (job_is_cancelled(&s->common.job) &&
- (!s->synced || s->common.force))
+ (!s->synced || s->common.job.force_cancel))
{
break;
}
@@ -884,7 +884,7 @@ immediate_exit:
* or it was cancelled prematurely so that we do not guarantee that
* the target is a copy of the source.
*/
- assert(ret < 0 || ((s->common.force || !s->synced) &&
+ assert(ret < 0 || ((s->common.job.force_cancel || !s->synced) &&
job_is_cancelled(&s->common.job)));
assert(need_drain);
mirror_wait_for_all_io(s);
diff --git a/blockjob.c b/blockjob.c
index 34c57da304..4cac3670ad 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -270,19 +270,20 @@ static int block_job_prepare(BlockJob *job)
return job->job.ret;
}
-static void block_job_cancel_async(BlockJob *job, bool force)
+static void job_cancel_async(Job *job, bool force)
{
- if (job->iostatus != BLOCK_DEVICE_IO_STATUS_OK) {
- block_job_iostatus_reset(job);
- }
- if (job->job.user_paused) {
- /* Do not call block_job_enter here, the caller will handle it. */
- job->job.user_paused = false;
- job->job.pause_count--;
+ if (job->user_paused) {
+ /* Do not call job_enter here, the caller will handle it. */
+ job->user_paused = false;
+ if (job->driver->user_resume) {
+ job->driver->user_resume(job);
+ }
+ assert(job->pause_count > 0);
+ job->pause_count--;
}
- job->job.cancelled = true;
+ job->cancelled = true;
/* To prevent 'force == false' overriding a previous 'force == true' */
- job->force |= force;
+ job->force_cancel |= force;
}
static int block_job_txn_apply(BlockJobTxn *txn, int fn(BlockJob *), bool lock)
@@ -367,7 +368,7 @@ static void block_job_completed_txn_abort(BlockJob *job)
* on the caller, so leave it. */
QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
if (other_job != job) {
- block_job_cancel_async(other_job, false);
+ job_cancel_async(&other_job->job, false);
}
}
while (!QLIST_EMPTY(&txn->jobs)) {
@@ -527,7 +528,7 @@ void block_job_cancel(BlockJob *job, bool force)
job_do_dismiss(&job->job);
return;
}
- block_job_cancel_async(job, force);
+ job_cancel_async(&job->job, force);
if (!job_started(&job->job)) {
block_job_completed(job, -ECANCELED);
} else if (job->job.deferred_to_main_loop) {
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 21/40] job: Convert block_job_cancel_async() " Kevin Wolf
@ 2018-05-23 23:18 ` John Snow
2018-05-24 8:24 ` Kevin Wolf
0 siblings, 1 reply; 86+ messages in thread
From: John Snow @ 2018-05-23 23:18 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: qemu-devel, jcody, armbru, mreitz
On 05/18/2018 09:20 AM, Kevin Wolf wrote:
> block_job_cancel_async() did two things that were still block job
> specific:
>
> * Setting job->force. This field makes sense on the Job level, so we can
> just move it. While at it, rename it to job->force_cancel to make its
> purpose more obvious.
>
> * Resetting the I/O status. This can't be moved because generic Jobs
> don't have an I/O status. What the function really implements is a
> user resume, except without entering the coroutine. Consequently, it
You know, I had noticed this before but because this is called from
contexts which are NOT user-initiated, I didn't want to share the callback.
... I guess it's fine because the job is about to not exist anymore, so
it probably won't spook anything.
> makes sense to call the .user_resume driver callback here which
> already resets the I/O status.
>
> The old block_job_cancel_async() has two separate if statements that
> check job->iostatus != BLOCK_DEVICE_IO_STATUS_OK and job->user_paused.
> However, the former condition always implies the latter (as is
> asserted in block_job_iostatus_reset()), so changing the explicit call
> of block_job_iostatus_reset() on the former condition with the
> .user_resume callback on the latter condition is equivalent and
> doesn't need to access any BlockJob specific state.
Oh, cool! This is nice.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> Reviewed-by: Max Reitz <mreitz@redhat.com>
> ---
> include/block/blockjob.h | 6 ------
> include/qemu/job.h | 6 ++++++
> block/mirror.c | 4 ++--
> blockjob.c | 25 +++++++++++++------------
> 4 files changed, 21 insertions(+), 20 deletions(-)
>
> diff --git a/include/block/blockjob.h b/include/block/blockjob.h
> index 3f405d1fa7..d975efea20 100644
> --- a/include/block/blockjob.h
> +++ b/include/block/blockjob.h
> @@ -51,12 +51,6 @@ typedef struct BlockJob {
> BlockBackend *blk;
>
> /**
> - * Set to true if the job should abort immediately without waiting
> - * for data to be in sync.
> - */
> - bool force;
> -
> - /**
> * Set to true when the job is ready to be completed.
> */
> bool ready;
> diff --git a/include/qemu/job.h b/include/qemu/job.h
> index 3e817beee9..2648c74281 100644
> --- a/include/qemu/job.h
> +++ b/include/qemu/job.h
> @@ -97,6 +97,12 @@ typedef struct Job {
> */
> bool cancelled;
>
> + /**
> + * Set to true if the job should abort immediately without waiting
> + * for data to be in sync.
> + */
> + bool force_cancel;
> +
Does this comment need an update now, though?
Actually, in terms of "new jobs" API, it'd be really nice if cancel
*always meant cancel*.
I think "cancel" should never be used to mean "successful completion,
but different from the one we'd get if we used job_complete."
i.e., either we need a job setting:
job-set completion-mode=[pivot|no-pivot]
or optional parameters to pass to job-complete:
job-complete mode=[understood-by-job-type]
or some other mechanism that accomplishes the same type of behavior. It
would be nice if it did not have to be determined at job creation time
but instead could be determined later.
> /** Set to true when the job has deferred work to the main loop. */
> bool deferred_to_main_loop;
>
> diff --git a/block/mirror.c b/block/mirror.c
> index e9a90ea730..c3951d1ca2 100644
> --- a/block/mirror.c
> +++ b/block/mirror.c
> @@ -871,7 +871,7 @@ static void coroutine_fn mirror_run(void *opaque)
> trace_mirror_before_sleep(s, cnt, s->synced, delay_ns);
> job_sleep_ns(&s->common.job, delay_ns);
> if (job_is_cancelled(&s->common.job) &&
> - (!s->synced || s->common.force))
> + (!s->synced || s->common.job.force_cancel))
> {
> break;
> }
> @@ -884,7 +884,7 @@ immediate_exit:
> * or it was cancelled prematurely so that we do not guarantee that
> * the target is a copy of the source.
> */
> - assert(ret < 0 || ((s->common.force || !s->synced) &&
> + assert(ret < 0 || ((s->common.job.force_cancel || !s->synced) &&
> job_is_cancelled(&s->common.job)));
> assert(need_drain);
> mirror_wait_for_all_io(s);
> diff --git a/blockjob.c b/blockjob.c
> index 34c57da304..4cac3670ad 100644
> --- a/blockjob.c
> +++ b/blockjob.c
> @@ -270,19 +270,20 @@ static int block_job_prepare(BlockJob *job)
> return job->job.ret;
> }
>
> -static void block_job_cancel_async(BlockJob *job, bool force)
> +static void job_cancel_async(Job *job, bool force)
> {
> - if (job->iostatus != BLOCK_DEVICE_IO_STATUS_OK) {
> - block_job_iostatus_reset(job);
> - }
> - if (job->job.user_paused) {
> - /* Do not call block_job_enter here, the caller will handle it. */
> - job->job.user_paused = false;
> - job->job.pause_count--;
> + if (job->user_paused) {
> + /* Do not call job_enter here, the caller will handle it. */
> + job->user_paused = false;
> + if (job->driver->user_resume) {
> + job->driver->user_resume(job);
> + }
> + assert(job->pause_count > 0);
> + job->pause_count--;
> }
> - job->job.cancelled = true;
> + job->cancelled = true;
> /* To prevent 'force == false' overriding a previous 'force == true' */
> - job->force |= force;
> + job->force_cancel |= force;
> }
>
> static int block_job_txn_apply(BlockJobTxn *txn, int fn(BlockJob *), bool lock)
> @@ -367,7 +368,7 @@ static void block_job_completed_txn_abort(BlockJob *job)
> * on the caller, so leave it. */
> QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
> if (other_job != job) {
> - block_job_cancel_async(other_job, false);
> + job_cancel_async(&other_job->job, false);
> }
> }
> while (!QLIST_EMPTY(&txn->jobs)) {
> @@ -527,7 +528,7 @@ void block_job_cancel(BlockJob *job, bool force)
> job_do_dismiss(&job->job);
> return;
> }
> - block_job_cancel_async(job, force);
> + job_cancel_async(&job->job, force);
> if (!job_started(&job->job)) {
> block_job_completed(job, -ECANCELED);
> } else if (job->job.deferred_to_main_loop) {
>
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-23 23:18 ` John Snow
@ 2018-05-24 8:24 ` Kevin Wolf
2018-05-24 17:42 ` John Snow
0 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-24 8:24 UTC (permalink / raw)
To: John Snow; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
Am 24.05.2018 um 01:18 hat John Snow geschrieben:
> > diff --git a/include/qemu/job.h b/include/qemu/job.h
> > index 3e817beee9..2648c74281 100644
> > --- a/include/qemu/job.h
> > +++ b/include/qemu/job.h
> > @@ -97,6 +97,12 @@ typedef struct Job {
> > */
> > bool cancelled;
> >
> > + /**
> > + * Set to true if the job should abort immediately without waiting
> > + * for data to be in sync.
> > + */
> > + bool force_cancel;
> > +
>
> Does this comment need an update now, though?
>
> Actually, in terms of "new jobs" API, it'd be really nice if cancel
> *always meant cancel*.
>
> I think "cancel" should never be used to mean "successful completion,
> but different from the one we'd get if we used job_complete."
>
> i.e., either we need a job setting:
>
> job-set completion-mode=[pivot|no-pivot]
>
> or optional parameters to pass to job-complete:
>
> job-complete mode=[understood-by-job-type]
>
> or some other mechanism that accomplishes the same type of behavior. It
> would be nice if it did not have to be determined at job creation time
> but instead could be determined later.
I agree. We already made sure that job-cancel really means cancel on the
QAPI level, so we're free to do that. We just need to keep supporting
block-job-cancel with the old semantics, so what I have is the easy
conversion. We can change the internal implementation when we actually
implement the selection of a completion mode.
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-24 8:24 ` Kevin Wolf
@ 2018-05-24 17:42 ` John Snow
2018-05-25 8:00 ` Kevin Wolf
0 siblings, 1 reply; 86+ messages in thread
From: John Snow @ 2018-05-24 17:42 UTC (permalink / raw)
To: Kevin Wolf; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
On 05/24/2018 04:24 AM, Kevin Wolf wrote:
> Am 24.05.2018 um 01:18 hat John Snow geschrieben:
>>> diff --git a/include/qemu/job.h b/include/qemu/job.h
>>> index 3e817beee9..2648c74281 100644
>>> --- a/include/qemu/job.h
>>> +++ b/include/qemu/job.h
>>> @@ -97,6 +97,12 @@ typedef struct Job {
>>> */
>>> bool cancelled;
>>>
>>> + /**
>>> + * Set to true if the job should abort immediately without waiting
>>> + * for data to be in sync.
>>> + */
>>> + bool force_cancel;
>>> +
>>
>> Does this comment need an update now, though?
>>
>> Actually, in terms of "new jobs" API, it'd be really nice if cancel
>> *always meant cancel*.
>>
>> I think "cancel" should never be used to mean "successful completion,
>> but different from the one we'd get if we used job_complete."
>>
>> i.e., either we need a job setting:
>>
>> job-set completion-mode=[pivot|no-pivot]
>>
>> or optional parameters to pass to job-complete:
>>
>> job-complete mode=[understood-by-job-type]
>>
>> or some other mechanism that accomplishes the same type of behavior. It
>> would be nice if it did not have to be determined at job creation time
>> but instead could be determined later.
>
> I agree. We already made sure that job-cancel really means cancel on the
> QAPI level, so we're free to do that. We just need to keep supporting
> block-job-cancel with the old semantics, so what I have is the easy
> conversion. We can change the internal implementation when we actually
> implement the selection of a completion mode.
>
> Kevin
>
We need this before 3.0 though, yeah? unless we make job-cancel
x-job-cancel or some other warning that the way it works might change, yeah?
Or do I misunderstand our leeway to change this at a later point in time?
(can job-cancel apply to block jobs created with the legacy
infrastructure? My reading was "yes.")
--js
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-24 17:42 ` John Snow
@ 2018-05-25 8:00 ` Kevin Wolf
2018-05-25 17:43 ` John Snow
2018-05-29 11:59 ` [Qemu-devel] [Qemu-block] " Kashyap Chamarthy
0 siblings, 2 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-25 8:00 UTC (permalink / raw)
To: John Snow; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
Am 24.05.2018 um 19:42 hat John Snow geschrieben:
>
>
> On 05/24/2018 04:24 AM, Kevin Wolf wrote:
> > Am 24.05.2018 um 01:18 hat John Snow geschrieben:
> >>> diff --git a/include/qemu/job.h b/include/qemu/job.h
> >>> index 3e817beee9..2648c74281 100644
> >>> --- a/include/qemu/job.h
> >>> +++ b/include/qemu/job.h
> >>> @@ -97,6 +97,12 @@ typedef struct Job {
> >>> */
> >>> bool cancelled;
> >>>
> >>> + /**
> >>> + * Set to true if the job should abort immediately without waiting
> >>> + * for data to be in sync.
> >>> + */
> >>> + bool force_cancel;
> >>> +
> >>
> >> Does this comment need an update now, though?
> >>
> >> Actually, in terms of "new jobs" API, it'd be really nice if cancel
> >> *always meant cancel*.
> >>
> >> I think "cancel" should never be used to mean "successful completion,
> >> but different from the one we'd get if we used job_complete."
> >>
> >> i.e., either we need a job setting:
> >>
> >> job-set completion-mode=[pivot|no-pivot]
> >>
> >> or optional parameters to pass to job-complete:
> >>
> >> job-complete mode=[understood-by-job-type]
> >>
> >> or some other mechanism that accomplishes the same type of behavior. It
> >> would be nice if it did not have to be determined at job creation time
> >> but instead could be determined later.
> >
> > I agree. We already made sure that job-cancel really means cancel on the
> > QAPI level, so we're free to do that. We just need to keep supporting
> > block-job-cancel with the old semantics, so what I have is the easy
> > conversion. We can change the internal implementation when we actually
> > implement the selection of a completion mode.
> >
> > Kevin
> >
>
> We need this before 3.0 though, yeah? unless we make job-cancel
> x-job-cancel or some other warning that the way it works might change, yeah?
>
> Or do I misunderstand our leeway to change this at a later point in time?
>
> (can job-cancel apply to block jobs created with the legacy
> infrastructure? My reading was "yes.")
It can, and it already has its final semantics, so nothing has to change
before 3.0. job-cancel is equivalent to block-job-cancel with fixed
force=true. If you want the complete-by-cancel behaviour of mirror, you
have to use block-job-cancel for now, because job-cancel doesn't provide
that functionality.
So what we can change later is adding a way to initiate this secondary
completion mode with a job-* command (probably with a new option for
job-complete). But we wouldn't change the semantics of exisiting
commands.
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-25 8:00 ` Kevin Wolf
@ 2018-05-25 17:43 ` John Snow
2018-05-29 11:59 ` [Qemu-devel] [Qemu-block] " Kashyap Chamarthy
1 sibling, 0 replies; 86+ messages in thread
From: John Snow @ 2018-05-25 17:43 UTC (permalink / raw)
To: Kevin Wolf; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
On 05/25/2018 04:00 AM, Kevin Wolf wrote:
> Am 24.05.2018 um 19:42 hat John Snow geschrieben:
>>
>>
>> On 05/24/2018 04:24 AM, Kevin Wolf wrote:
>>> Am 24.05.2018 um 01:18 hat John Snow geschrieben:
>>>>> diff --git a/include/qemu/job.h b/include/qemu/job.h
>>>>> index 3e817beee9..2648c74281 100644
>>>>> --- a/include/qemu/job.h
>>>>> +++ b/include/qemu/job.h
>>>>> @@ -97,6 +97,12 @@ typedef struct Job {
>>>>> */
>>>>> bool cancelled;
>>>>>
>>>>> + /**
>>>>> + * Set to true if the job should abort immediately without waiting
>>>>> + * for data to be in sync.
>>>>> + */
>>>>> + bool force_cancel;
>>>>> +
>>>>
>>>> Does this comment need an update now, though?
>>>>
>>>> Actually, in terms of "new jobs" API, it'd be really nice if cancel
>>>> *always meant cancel*.
>>>>
>>>> I think "cancel" should never be used to mean "successful completion,
>>>> but different from the one we'd get if we used job_complete."
>>>>
>>>> i.e., either we need a job setting:
>>>>
>>>> job-set completion-mode=[pivot|no-pivot]
>>>>
>>>> or optional parameters to pass to job-complete:
>>>>
>>>> job-complete mode=[understood-by-job-type]
>>>>
>>>> or some other mechanism that accomplishes the same type of behavior. It
>>>> would be nice if it did not have to be determined at job creation time
>>>> but instead could be determined later.
>>>
>>> I agree. We already made sure that job-cancel really means cancel on the
>>> QAPI level, so we're free to do that. We just need to keep supporting
>>> block-job-cancel with the old semantics, so what I have is the easy
>>> conversion. We can change the internal implementation when we actually
>>> implement the selection of a completion mode.
>>>
>>> Kevin
>>>
>>
>> We need this before 3.0 though, yeah? unless we make job-cancel
>> x-job-cancel or some other warning that the way it works might change, yeah?
>>
>> Or do I misunderstand our leeway to change this at a later point in time?
>>
>> (can job-cancel apply to block jobs created with the legacy
>> infrastructure? My reading was "yes.")
>
> It can, and it already has its final semantics, so nothing has to change
> before 3.0. job-cancel is equivalent to block-job-cancel with fixed
> force=true. If you want the complete-by-cancel behaviour of mirror, you
> have to use block-job-cancel for now, because job-cancel doesn't provide
> that functionality.
>
> So what we can change later is adding a way to initiate this secondary
> completion mode with a job-* command (probably with a new option for
> job-complete). But we wouldn't change the semantics of exisiting
> commands.
>
> Kevin
>
Oho, I _was_ missing something. I hadn't realized cancel takes on the
newer semantics. Thanks.
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [Qemu-block] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-25 8:00 ` Kevin Wolf
2018-05-25 17:43 ` John Snow
@ 2018-05-29 11:59 ` Kashyap Chamarthy
2018-05-29 12:30 ` Max Reitz
1 sibling, 1 reply; 86+ messages in thread
From: Kashyap Chamarthy @ 2018-05-29 11:59 UTC (permalink / raw)
To: Kevin Wolf; +Cc: John Snow, mreitz, qemu-devel, qemu-block, armbru
On Fri, May 25, 2018 at 10:00:35AM +0200, Kevin Wolf wrote:
> Am 24.05.2018 um 19:42 hat John Snow geschrieben:
[...]
(Randomly chiming in for a small clarification.)
> > >> or some other mechanism that accomplishes the same type of behavior. It
> > >> would be nice if it did not have to be determined at job creation time
> > >> but instead could be determined later.
> > >
> > > I agree. We already made sure that job-cancel really means cancel on the
> > > QAPI level, so we're free to do that. We just need to keep supporting
> > > block-job-cancel with the old semantics, so what I have is the easy
> > > conversion. We can change the internal implementation when we actually
> > > implement the selection of a completion mode.
> > >
> > > Kevin
> > >
> >
> > We need this before 3.0 though, yeah? unless we make job-cancel
> > x-job-cancel or some other warning that the way it works might change, yeah?
> >
> > Or do I misunderstand our leeway to change this at a later point in time?
> >
> > (can job-cancel apply to block jobs created with the legacy
> > infrastructure? My reading was "yes.")
>
> It can, and it already has its final semantics, so nothing has to change
> before 3.0. job-cancel is equivalent to block-job-cancel with fixed
> force=true. If you want the complete-by-cancel behaviour of mirror, you
> have to use block-job-cancel for now, because job-cancel doesn't provide
> that functionality.
I think by "complete-by-cancel" you are referring to this[*] magic
behaviour, mentioned in the last sentence here:
When you cancel an in-progress ‘mirror’ job before the source and
target are synchronized, block-job-cancel will emit the event
BLOCK_JOB_CANCELLED. However, note that if you cancel a ‘mirror’ job
after it has indicated (via the event BLOCK_JOB_READY) that the
source and target have reached synchronization, then the event
emitted by block-job-cancel changes to BLOCK_JOB_COMPLETED.
[*] https://git.qemu.org/?p=qemu.git;a=blob;f=docs/interop/live-block-operations.rst#l515
> So what we can change later is adding a way to initiate this secondary
> completion mode with a job-* command (probably with a new option for
> job-complete). But we wouldn't change the semantics of exisiting
> commands.
Ah, so if I'm understanding it correctly, the existing subtle magic of
"complete-by-cancel" will be rectified by separting the two distinct
operations: 'job-cancel' and 'job-complete'.
--
/kashyap
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [Qemu-block] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-29 11:59 ` [Qemu-devel] [Qemu-block] " Kashyap Chamarthy
@ 2018-05-29 12:30 ` Max Reitz
2018-05-29 13:10 ` Kashyap Chamarthy
2018-05-30 20:33 ` John Snow
0 siblings, 2 replies; 86+ messages in thread
From: Max Reitz @ 2018-05-29 12:30 UTC (permalink / raw)
To: Kashyap Chamarthy, Kevin Wolf; +Cc: John Snow, qemu-devel, qemu-block, armbru
[-- Attachment #1: Type: text/plain, Size: 3314 bytes --]
On 2018-05-29 13:59, Kashyap Chamarthy wrote:
> On Fri, May 25, 2018 at 10:00:35AM +0200, Kevin Wolf wrote:
>> Am 24.05.2018 um 19:42 hat John Snow geschrieben:
>
> [...]
>
> (Randomly chiming in for a small clarification.)
>
>>>>> or some other mechanism that accomplishes the same type of behavior. It
>>>>> would be nice if it did not have to be determined at job creation time
>>>>> but instead could be determined later.
>>>>
>>>> I agree. We already made sure that job-cancel really means cancel on the
>>>> QAPI level, so we're free to do that. We just need to keep supporting
>>>> block-job-cancel with the old semantics, so what I have is the easy
>>>> conversion. We can change the internal implementation when we actually
>>>> implement the selection of a completion mode.
>>>>
>>>> Kevin
>>>>
>>>
>>> We need this before 3.0 though, yeah? unless we make job-cancel
>>> x-job-cancel or some other warning that the way it works might change, yeah?
>>>
>>> Or do I misunderstand our leeway to change this at a later point in time?
>>>
>>> (can job-cancel apply to block jobs created with the legacy
>>> infrastructure? My reading was "yes.")
>>
>> It can, and it already has its final semantics, so nothing has to change
>> before 3.0. job-cancel is equivalent to block-job-cancel with fixed
>> force=true. If you want the complete-by-cancel behaviour of mirror, you
>> have to use block-job-cancel for now, because job-cancel doesn't provide
>> that functionality.
>
> I think by "complete-by-cancel" you are referring to this[*] magic
> behaviour, mentioned in the last sentence here:
>
> When you cancel an in-progress ‘mirror’ job before the source and
> target are synchronized, block-job-cancel will emit the event
> BLOCK_JOB_CANCELLED. However, note that if you cancel a ‘mirror’ job
> after it has indicated (via the event BLOCK_JOB_READY) that the
> source and target have reached synchronization, then the event
> emitted by block-job-cancel changes to BLOCK_JOB_COMPLETED.
>
>
> [*] https://git.qemu.org/?p=qemu.git;a=blob;f=docs/interop/live-block-operations.rst#l515
>
>> So what we can change later is adding a way to initiate this secondary
>> completion mode with a job-* command (probably with a new option for
>> job-complete). But we wouldn't change the semantics of exisiting
>> commands.
>
> Ah, so if I'm understanding it correctly, the existing subtle magic of
> "complete-by-cancel" will be rectified by separting the two distinct
> operations: 'job-cancel' and 'job-complete'.
Not really, because we already had those two operations for block jobs.
The thing is that block-job-complete (and thus job-complete) will pivot
to the target disk, but block-job-cancel doesn't.
The special behavior is that you can use block-job-cancel after
BLOCK_JOB_READY to complete the job, but not pivot to it. I don't think
we have a real plan on how to represent that with the generic job
commands, we just know that we don't want to use job-cancel.
(Maybe we can add a flag to job-complete (which to me does not sound
like a good idea), or you could set flags on jobs while they are
running, so you can set a do-not-pivot flag on the mirror job before you
complete it.)
Max
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [Qemu-block] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-29 12:30 ` Max Reitz
@ 2018-05-29 13:10 ` Kashyap Chamarthy
2018-05-29 13:22 ` Kashyap Chamarthy
2018-05-30 20:33 ` John Snow
1 sibling, 1 reply; 86+ messages in thread
From: Kashyap Chamarthy @ 2018-05-29 13:10 UTC (permalink / raw)
To: Max Reitz; +Cc: Kevin Wolf, John Snow, qemu-devel, qemu-block, armbru
On Tue, May 29, 2018 at 02:30:47PM +0200, Max Reitz wrote:
> On 2018-05-29 13:59, Kashyap Chamarthy wrote:
> > On Fri, May 25, 2018 at 10:00:35AM +0200, Kevin Wolf wrote:
[...]
> >> It can, and it already has its final semantics, so nothing has to change
> >> before 3.0. job-cancel is equivalent to block-job-cancel with fixed
> >> force=true. If you want the complete-by-cancel behaviour of mirror, you
> >> have to use block-job-cancel for now, because job-cancel doesn't provide
> >> that functionality.
> >
> > I think by "complete-by-cancel" you are referring to this[*] magic
> > behaviour, mentioned in the last sentence here:
> >
> > When you cancel an in-progress ‘mirror’ job before the source and
> > target are synchronized, block-job-cancel will emit the event
> > BLOCK_JOB_CANCELLED. However, note that if you cancel a ‘mirror’ job
> > after it has indicated (via the event BLOCK_JOB_READY) that the
> > source and target have reached synchronization, then the event
> > emitted by block-job-cancel changes to BLOCK_JOB_COMPLETED.
> >
> >
> > [*] https://git.qemu.org/?p=qemu.git;a=blob;f=docs/interop/live-block-operations.rst#l515
> >
> >> So what we can change later is adding a way to initiate this secondary
> >> completion mode with a job-* command (probably with a new option for
> >> job-complete). But we wouldn't change the semantics of exisiting
> >> commands.
> >
> > Ah, so if I'm understanding it correctly, the existing subtle magic of
> > "complete-by-cancel" will be rectified by separting the two distinct
> > operations: 'job-cancel' and 'job-complete'.
>
> Not really, because we already had those two operations for block jobs.
> The thing is that block-job-complete (and thus job-complete) will pivot
> to the target disk, but block-job-cancel doesn't.
Indeed; from the doc linked earlier: "Issuing the command
``block-job-complete`` (after it emits the event
``BLOCK_JOB_COMPLETED``) will adjust the guest device (i.e. live QEMU)
to point to the target image, [E], causing all the new writes from this
point on to happen there."
> The special behavior is that you can use block-job-cancel after
> BLOCK_JOB_READY to complete the job, but not pivot to it. I don't think
> we have a real plan on how to represent that with the generic job
> commands, we just know that we don't want to use job-cancel.
Ah, thanks for clarifying. Yes, what you say makes sense — not using
'job-cancel' to represent completion.
> (Maybe we can add a flag to job-complete (which to me does not sound
> like a good idea), or you could set flags on jobs while they are
> running, so you can set a do-not-pivot flag on the mirror job before you
> complete it.)
Yeah, spelling that out, 'do-not-pivot' or something along those lines,
as a flag makes it clearer. "Implicit is better than explicit".
--
/kashyap
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [Qemu-block] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-29 13:10 ` Kashyap Chamarthy
@ 2018-05-29 13:22 ` Kashyap Chamarthy
0 siblings, 0 replies; 86+ messages in thread
From: Kashyap Chamarthy @ 2018-05-29 13:22 UTC (permalink / raw)
To: Max Reitz; +Cc: Kevin Wolf, John Snow, qemu-devel, qemu-block, armbru
On Tue, May 29, 2018 at 03:10:58PM +0200, Kashyap Chamarthy wrote:
> On Tue, May 29, 2018 at 02:30:47PM +0200, Max Reitz wrote:
[...]
> > The special behavior is that you can use block-job-cancel after
> > BLOCK_JOB_READY to complete the job, but not pivot to it. I don't think
> > we have a real plan on how to represent that with the generic job
> > commands, we just know that we don't want to use job-cancel.
>
> Ah, thanks for clarifying. Yes, what you say makes sense — not using
> 'job-cancel' to represent completion.
>
> > (Maybe we can add a flag to job-complete (which to me does not sound
> > like a good idea), or you could set flags on jobs while they are
> > running, so you can set a do-not-pivot flag on the mirror job before you
> > complete it.)
>
> Yeah, spelling that out, 'do-not-pivot' or something along those lines,
> as a flag makes it clearer. "Implicit is better than explicit".
Oops, I got the quote reverse :P. Correct: "Explicit is better than
implicit". But you've auto-corrected in your head already...
--
/kashyap
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [Qemu-block] [PATCH v2 21/40] job: Convert block_job_cancel_async() to Job
2018-05-29 12:30 ` Max Reitz
2018-05-29 13:10 ` Kashyap Chamarthy
@ 2018-05-30 20:33 ` John Snow
1 sibling, 0 replies; 86+ messages in thread
From: John Snow @ 2018-05-30 20:33 UTC (permalink / raw)
To: Max Reitz, Kashyap Chamarthy, Kevin Wolf; +Cc: qemu-devel, qemu-block, armbru
On 05/29/2018 08:30 AM, Max Reitz wrote:
> On 2018-05-29 13:59, Kashyap Chamarthy wrote:
>> On Fri, May 25, 2018 at 10:00:35AM +0200, Kevin Wolf wrote:
>>> Am 24.05.2018 um 19:42 hat John Snow geschrieben:
>>
>> [...]
>>
>> (Randomly chiming in for a small clarification.)
>>
>>>>>> or some other mechanism that accomplishes the same type of behavior. It
>>>>>> would be nice if it did not have to be determined at job creation time
>>>>>> but instead could be determined later.
>>>>>
>>>>> I agree. We already made sure that job-cancel really means cancel on the
>>>>> QAPI level, so we're free to do that. We just need to keep supporting
>>>>> block-job-cancel with the old semantics, so what I have is the easy
>>>>> conversion. We can change the internal implementation when we actually
>>>>> implement the selection of a completion mode.
>>>>>
>>>>> Kevin
>>>>>
>>>>
>>>> We need this before 3.0 though, yeah? unless we make job-cancel
>>>> x-job-cancel or some other warning that the way it works might change, yeah?
>>>>
>>>> Or do I misunderstand our leeway to change this at a later point in time?
>>>>
>>>> (can job-cancel apply to block jobs created with the legacy
>>>> infrastructure? My reading was "yes.")
>>>
>>> It can, and it already has its final semantics, so nothing has to change
>>> before 3.0. job-cancel is equivalent to block-job-cancel with fixed
>>> force=true. If you want the complete-by-cancel behaviour of mirror, you
>>> have to use block-job-cancel for now, because job-cancel doesn't provide
>>> that functionality.
>>
>> I think by "complete-by-cancel" you are referring to this[*] magic
>> behaviour, mentioned in the last sentence here:
>>
>> When you cancel an in-progress ‘mirror’ job before the source and
>> target are synchronized, block-job-cancel will emit the event
>> BLOCK_JOB_CANCELLED. However, note that if you cancel a ‘mirror’ job
>> after it has indicated (via the event BLOCK_JOB_READY) that the
>> source and target have reached synchronization, then the event
>> emitted by block-job-cancel changes to BLOCK_JOB_COMPLETED.
>>
>>
>> [*] https://git.qemu.org/?p=qemu.git;a=blob;f=docs/interop/live-block-operations.rst#l515
>>
>>> So what we can change later is adding a way to initiate this secondary
>>> completion mode with a job-* command (probably with a new option for
>>> job-complete). But we wouldn't change the semantics of exisiting
>>> commands.
>>
>> Ah, so if I'm understanding it correctly, the existing subtle magic of
>> "complete-by-cancel" will be rectified by separting the two distinct
>> operations: 'job-cancel' and 'job-complete'.
>
> Not really, because we already had those two operations for block jobs.
> The thing is that block-job-complete (and thus job-complete) will pivot
> to the target disk, but block-job-cancel doesn't.
>
> The special behavior is that you can use block-job-cancel after
> BLOCK_JOB_READY to complete the job, but not pivot to it. I don't think
> we have a real plan on how to represent that with the generic job
> commands, we just know that we don't want to use job-cancel.
>
> (Maybe we can add a flag to job-complete (which to me does not sound
> like a good idea), or you could set flags on jobs while they are
> running, so you can set a do-not-pivot flag on the mirror job before you
> complete it.)
>
> Max
>
Adding the completion-mode as a "setting" to the job has always been my
favorite way. Whether you set this setting at creation time or later
during runtime is not important.
When we issue job-complete, the job knows based on its own settings the
way it should do so, if there are multiple possibilities.
I only mentioned it as a flag to complete ... err, just in case it was
easier that way somehow. *shrug*.
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 22/40] job: Add job_drain()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (20 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 21/40] job: Convert block_job_cancel_async() " Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 23/40] job: Move .complete callback to Job Kevin Wolf
` (20 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
block_job_drain() contains a blk_drain() call which cannot be moved to
Job, so add a new JobDriver callback JobDriver.drain which has a common
implementation for all BlockJobs. In addition to this we keep the
existing BlockJobDriver.drain callback that is called by the common
drain implementation for all block jobs.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob_int.h | 12 ++++++++++++
include/qemu/job.h | 13 +++++++++++++
block/backup.c | 1 +
block/commit.c | 1 +
block/mirror.c | 2 ++
block/stream.c | 1 +
blockjob.c | 20 ++++++++++----------
job.c | 11 +++++++++++
tests/test-bdrv-drain.c | 1 +
tests/test-blockjob-txn.c | 1 +
tests/test-blockjob.c | 2 ++
11 files changed, 55 insertions(+), 10 deletions(-)
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index bf2b762808..38fe22d7e0 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -65,6 +65,10 @@ struct BlockJobDriver {
* If the callback is not NULL, it will be invoked when the job has to be
* synchronously cancelled or completed; it should drain BlockDriverStates
* as required to ensure progress.
+ *
+ * Block jobs must use the default implementation for job_driver.drain,
+ * which will in turn call this callback after doing generic block job
+ * stuff.
*/
void (*drain)(BlockJob *job);
};
@@ -112,6 +116,14 @@ void block_job_free(Job *job);
void block_job_user_resume(Job *job);
/**
+ * block_job_drain:
+ * Callback to be used for JobDriver.drain in all block jobs. Drains the main
+ * block node associated with the block jobs and calls BlockJobDriver.drain for
+ * job-specific actions.
+ */
+void block_job_drain(Job *job);
+
+/**
* block_job_yield:
* @job: The job that calls the function.
*
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 2648c74281..aebc1959e6 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -167,6 +167,13 @@ struct JobDriver {
*/
void (*user_resume)(Job *job);
+ /*
+ * If the callback is not NULL, it will be invoked when the job has to be
+ * synchronously cancelled or completed; it should drain any activities
+ * as required to ensure progress.
+ */
+ void (*drain)(Job *job);
+
/**
* If the callback is not NULL, it will be invoked when all the jobs
* belonging to the same transaction complete; or upon this job's
@@ -325,6 +332,12 @@ bool job_user_paused(Job *job);
*/
void job_user_resume(Job *job, Error **errp);
+/*
+ * Drain any activities as required to ensure progress. This can be called in a
+ * loop to synchronously complete a job.
+ */
+void job_drain(Job *job);
+
/**
* Get the next element from the list of block jobs after @job, or the
* first one if @job is %NULL.
diff --git a/block/backup.c b/block/backup.c
index bd31282924..ca7d990b21 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -529,6 +529,7 @@ static const BlockJobDriver backup_job_driver = {
.job_type = JOB_TYPE_BACKUP,
.free = block_job_free,
.user_resume = block_job_user_resume,
+ .drain = block_job_drain,
.start = backup_run,
.commit = backup_commit,
.abort = backup_abort,
diff --git a/block/commit.c b/block/commit.c
index e53b2d7d55..02a8af9127 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -221,6 +221,7 @@ static const BlockJobDriver commit_job_driver = {
.job_type = JOB_TYPE_COMMIT,
.free = block_job_free,
.user_resume = block_job_user_resume,
+ .drain = block_job_drain,
.start = commit_run,
},
};
diff --git a/block/mirror.c b/block/mirror.c
index c3951d1ca2..a579bd8cd1 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -992,6 +992,7 @@ static const BlockJobDriver mirror_job_driver = {
.job_type = JOB_TYPE_MIRROR,
.free = block_job_free,
.user_resume = block_job_user_resume,
+ .drain = block_job_drain,
.start = mirror_run,
.pause = mirror_pause,
},
@@ -1006,6 +1007,7 @@ static const BlockJobDriver commit_active_job_driver = {
.job_type = JOB_TYPE_COMMIT,
.free = block_job_free,
.user_resume = block_job_user_resume,
+ .drain = block_job_drain,
.start = mirror_run,
.pause = mirror_pause,
},
diff --git a/block/stream.c b/block/stream.c
index eee02538ed..b996278ab3 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -215,6 +215,7 @@ static const BlockJobDriver stream_job_driver = {
.free = block_job_free,
.start = stream_run,
.user_resume = block_job_user_resume,
+ .drain = block_job_drain,
},
};
diff --git a/blockjob.c b/blockjob.c
index 4cac3670ad..63e166927a 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -169,14 +169,13 @@ static void block_job_attached_aio_context(AioContext *new_context,
job_resume(&job->job);
}
-static void block_job_drain(BlockJob *job)
+void block_job_drain(Job *job)
{
- /* If job is !job->job.busy this kicks it into the next pause point. */
- block_job_enter(job);
+ BlockJob *bjob = container_of(job, BlockJob, job);
- blk_drain(job->blk);
- if (job->driver->drain) {
- job->driver->drain(job);
+ blk_drain(bjob->blk);
+ if (bjob->driver->drain) {
+ bjob->driver->drain(bjob);
}
}
@@ -190,7 +189,7 @@ static void block_job_detach_aio_context(void *opaque)
job_pause(&job->job);
while (!job->job.paused && !job_is_completed(&job->job)) {
- block_job_drain(job);
+ job_drain(&job->job);
}
job->job.aio_context = NULL;
@@ -327,11 +326,11 @@ static int block_job_finish_sync(BlockJob *job,
job_unref(&job->job);
return -EBUSY;
}
- /* block_job_drain calls block_job_enter, and it should be enough to
- * induce progress until the job completes or moves to the main thread.
+ /* job_drain calls job_enter, and it should be enough to induce progress
+ * until the job completes or moves to the main thread.
*/
while (!job->job.deferred_to_main_loop && !job_is_completed(&job->job)) {
- block_job_drain(job);
+ job_drain(&job->job);
}
while (!job_is_completed(&job->job)) {
aio_poll(qemu_get_aio_context(), true);
@@ -713,6 +712,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
assert(is_block_job(&job->job));
assert(job->job.driver->free == &block_job_free);
assert(job->job.driver->user_resume == &block_job_user_resume);
+ assert(job->job.driver->drain == &block_job_drain);
job->driver = driver;
job->blk = blk;
diff --git a/job.c b/job.c
index f9bd8baff2..a2cf411a07 100644
--- a/job.c
+++ b/job.c
@@ -367,6 +367,17 @@ void coroutine_fn job_sleep_ns(Job *job, int64_t ns)
job_pause_point(job);
}
+void job_drain(Job *job)
+{
+ /* If job is !busy this kicks it into the next pause point. */
+ job_enter(job);
+
+ if (job->driver->drain) {
+ job->driver->drain(job);
+ }
+}
+
+
/**
* All jobs must allow a pause point before entering their job proper. This
* ensures that jobs can be paused prior to being started, then resumed later.
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index c993512f66..58ea5663f0 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -525,6 +525,7 @@ BlockJobDriver test_job_driver = {
.instance_size = sizeof(TestBlockJob),
.free = block_job_free,
.user_resume = block_job_user_resume,
+ .drain = block_job_drain,
.start = test_job_start,
},
.complete = test_job_complete,
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 60e9fa2298..1572f8d96f 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -79,6 +79,7 @@ static const BlockJobDriver test_block_job_driver = {
.instance_size = sizeof(TestBlockJob),
.free = block_job_free,
.user_resume = block_job_user_resume,
+ .drain = block_job_drain,
.start = test_block_job_run,
},
};
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 1fe6803fe0..592a13613d 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -21,6 +21,7 @@ static const BlockJobDriver test_block_job_driver = {
.instance_size = sizeof(BlockJob),
.free = block_job_free,
.user_resume = block_job_user_resume,
+ .drain = block_job_drain,
},
};
@@ -201,6 +202,7 @@ static const BlockJobDriver test_cancel_driver = {
.instance_size = sizeof(CancelJob),
.free = block_job_free,
.user_resume = block_job_user_resume,
+ .drain = block_job_drain,
.start = cancel_job_start,
},
.complete = cancel_job_complete,
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 23/40] job: Move .complete callback to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (21 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 22/40] job: Add job_drain() Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 24/40] job: Move job_finish_sync() " Kevin Wolf
` (19 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves the .complete callback that tells a READY job to complete
from BlockJobDriver to JobDriver. The wrapper function job_complete()
doesn't require anything block job specific any more and can be moved
to Job.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 10 ----------
include/block/blockjob_int.h | 6 ------
include/qemu/job.h | 8 ++++++++
block/mirror.c | 10 +++++-----
blockdev.c | 2 +-
blockjob.c | 23 +++++------------------
job.c | 16 ++++++++++++++++
tests/test-bdrv-drain.c | 6 +++---
tests/test-blockjob.c | 10 +++++-----
9 files changed, 43 insertions(+), 48 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index d975efea20..85ce18a381 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -154,16 +154,6 @@ void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp);
void block_job_cancel(BlockJob *job, bool force);
/**
- * block_job_complete:
- * @job: The job to be completed.
- * @errp: Error object.
- *
- * Asynchronously complete the specified job.
- */
-void block_job_complete(BlockJob *job, Error **errp);
-
-
-/**
* block_job_finalize:
* @job: The job to fully commit and finish.
* @errp: Error object.
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 38fe22d7e0..b8ca7bb0c9 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -39,12 +39,6 @@ struct BlockJobDriver {
JobDriver job_driver;
/**
- * Optional callback for job types whose completion must be triggered
- * manually.
- */
- void (*complete)(BlockJob *job, Error **errp);
-
- /**
* If the callback is not NULL, prepare will be invoked when all the jobs
* belonging to the same transaction complete; or upon this job's completion
* if it is not in a transaction.
diff --git a/include/qemu/job.h b/include/qemu/job.h
index aebc1959e6..8f7f71a9b1 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -167,6 +167,12 @@ struct JobDriver {
*/
void (*user_resume)(Job *job);
+ /**
+ * Optional callback for job types whose completion must be triggered
+ * manually.
+ */
+ void (*complete)(Job *job, Error **errp);
+
/*
* If the callback is not NULL, it will be invoked when the job has to be
* synchronously cancelled or completed; it should drain any activities
@@ -363,6 +369,8 @@ int job_apply_verb(Job *job, JobVerb verb, Error **errp);
/** The @job could not be started, free it. */
void job_early_fail(Job *job);
+/** Asynchronously complete the specified @job. */
+void job_complete(Job *job, Error **errp);;
typedef void JobDeferToMainLoopFn(Job *job, void *opaque);
diff --git a/block/mirror.c b/block/mirror.c
index a579bd8cd1..656237af5c 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -905,16 +905,16 @@ immediate_exit:
job_defer_to_main_loop(&s->common.job, mirror_exit, data);
}
-static void mirror_complete(BlockJob *job, Error **errp)
+static void mirror_complete(Job *job, Error **errp)
{
- MirrorBlockJob *s = container_of(job, MirrorBlockJob, common);
+ MirrorBlockJob *s = container_of(job, MirrorBlockJob, common.job);
BlockDriverState *target;
target = blk_bs(s->target);
if (!s->synced) {
error_setg(errp, "The active block job '%s' cannot be completed",
- job->job.id);
+ job->id);
return;
}
@@ -995,8 +995,8 @@ static const BlockJobDriver mirror_job_driver = {
.drain = block_job_drain,
.start = mirror_run,
.pause = mirror_pause,
+ .complete = mirror_complete,
},
- .complete = mirror_complete,
.attached_aio_context = mirror_attached_aio_context,
.drain = mirror_drain,
};
@@ -1010,8 +1010,8 @@ static const BlockJobDriver commit_active_job_driver = {
.drain = block_job_drain,
.start = mirror_run,
.pause = mirror_pause,
+ .complete = mirror_complete,
},
- .complete = mirror_complete,
.attached_aio_context = mirror_attached_aio_context,
.drain = mirror_drain,
};
diff --git a/blockdev.c b/blockdev.c
index 278b92ce03..0967f6ab66 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -3894,7 +3894,7 @@ void qmp_block_job_complete(const char *device, Error **errp)
}
trace_qmp_block_job_complete(job);
- block_job_complete(job, errp);
+ job_complete(&job->job, errp);
aio_context_release(aio_context);
}
diff --git a/blockjob.c b/blockjob.c
index 63e166927a..0ca7672941 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -481,24 +481,6 @@ int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n)
return ratelimit_calculate_delay(&job->limit, n);
}
-void block_job_complete(BlockJob *job, Error **errp)
-{
- /* Should not be reachable via external interface for internal jobs */
- assert(job->job.id);
- if (job_apply_verb(&job->job, JOB_VERB_COMPLETE, errp)) {
- return;
- }
- if (job->job.pause_count || job_is_cancelled(&job->job) ||
- !job->driver->complete)
- {
- error_setg(errp, "The active block job '%s' cannot be completed",
- job->job.id);
- return;
- }
-
- job->driver->complete(job, errp);
-}
-
void block_job_finalize(BlockJob *job, Error **errp)
{
assert(job && job->job.id);
@@ -571,6 +553,11 @@ void block_job_cancel_sync_all(void)
}
}
+static void block_job_complete(BlockJob *job, Error **errp)
+{
+ job_complete(&job->job, errp);
+}
+
int block_job_complete_sync(BlockJob *job, Error **errp)
{
return block_job_finish_sync(job, &block_job_complete, errp);
diff --git a/job.c b/job.c
index a2cf411a07..8a455798e2 100644
--- a/job.c
+++ b/job.c
@@ -556,6 +556,22 @@ int job_finalize_single(Job *job)
return 0;
}
+void job_complete(Job *job, Error **errp)
+{
+ /* Should not be reachable via external interface for internal jobs */
+ assert(job->id);
+ if (job_apply_verb(job, JOB_VERB_COMPLETE, errp)) {
+ return;
+ }
+ if (job->pause_count || job_is_cancelled(job) || !job->driver->complete) {
+ error_setg(errp, "The active block job '%s' cannot be completed",
+ job->id);
+ return;
+ }
+
+ job->driver->complete(job, errp);
+}
+
typedef struct {
Job *job;
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index 58ea5663f0..b428aaca68 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -514,9 +514,9 @@ static void coroutine_fn test_job_start(void *opaque)
job_defer_to_main_loop(&s->common.job, test_job_completed, NULL);
}
-static void test_job_complete(BlockJob *job, Error **errp)
+static void test_job_complete(Job *job, Error **errp)
{
- TestBlockJob *s = container_of(job, TestBlockJob, common);
+ TestBlockJob *s = container_of(job, TestBlockJob, common.job);
s->should_complete = true;
}
@@ -527,8 +527,8 @@ BlockJobDriver test_job_driver = {
.user_resume = block_job_user_resume,
.drain = block_job_drain,
.start = test_job_start,
+ .complete = test_job_complete,
},
- .complete = test_job_complete,
};
static void test_blockjob_common(enum drain_type drain_type)
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 592a13613d..e44c608327 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -171,9 +171,9 @@ static void cancel_job_completed(Job *job, void *opaque)
block_job_completed(bjob, 0);
}
-static void cancel_job_complete(BlockJob *job, Error **errp)
+static void cancel_job_complete(Job *job, Error **errp)
{
- CancelJob *s = container_of(job, CancelJob, common);
+ CancelJob *s = container_of(job, CancelJob, common.job);
s->should_complete = true;
}
@@ -204,8 +204,8 @@ static const BlockJobDriver test_cancel_driver = {
.user_resume = block_job_user_resume,
.drain = block_job_drain,
.start = cancel_job_start,
+ .complete = cancel_job_complete,
},
- .complete = cancel_job_complete,
};
static CancelJob *create_common(BlockJob **pjob)
@@ -333,7 +333,7 @@ static void test_cancel_pending(void)
block_job_enter(job);
assert(job->job.status == JOB_STATUS_READY);
- block_job_complete(job, &error_abort);
+ job_complete(&job->job, &error_abort);
block_job_enter(job);
while (!s->completed) {
aio_poll(qemu_get_aio_context(), true);
@@ -357,7 +357,7 @@ static void test_cancel_concluded(void)
block_job_enter(job);
assert(job->job.status == JOB_STATUS_READY);
- block_job_complete(job, &error_abort);
+ job_complete(&job->job, &error_abort);
block_job_enter(job);
while (!s->completed) {
aio_poll(qemu_get_aio_context(), true);
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 24/40] job: Move job_finish_sync() to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (22 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 23/40] job: Move .complete callback to Job Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 25/40] job: Switch transactions to JobTxn Kevin Wolf
` (18 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
block_job_finish_sync() doesn't contain anything block job specific any
more, so it can be moved to Job.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/qemu/job.h | 9 +++++++++
block/commit.c | 6 +++---
blockjob.c | 55 +++++++++---------------------------------------------
job.c | 28 +++++++++++++++++++++++++++
4 files changed, 49 insertions(+), 49 deletions(-)
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 8f7f71a9b1..17e2ceca87 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -389,6 +389,15 @@ typedef void JobDeferToMainLoopFn(Job *job, void *opaque);
*/
void job_defer_to_main_loop(Job *job, JobDeferToMainLoopFn *fn, void *opaque);
+/**
+ * Synchronously finishes the given @job. If @finish is given, it is called to
+ * trigger completion or cancellation of the job.
+ *
+ * Returns 0 if the job is successfully completed, -ECANCELED if the job was
+ * cancelled before completing, and -errno in other error cases.
+ */
+int job_finish_sync(Job *job, void (*finish)(Job *, Error **errp), Error **errp);
+
/* TODO To be removed from the public interface */
void job_state_transition(Job *job, JobStatus s1);
void coroutine_fn job_do_yield(Job *job, uint64_t ns);
diff --git a/block/commit.c b/block/commit.c
index 02a8af9127..40d97a35a5 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -112,9 +112,9 @@ static void commit_complete(Job *job, void *opaque)
blk_unref(s->top);
/* If there is more than one reference to the job (e.g. if called from
- * block_job_finish_sync()), block_job_completed() won't free it and
- * therefore the blockers on the intermediate nodes remain. This would
- * cause bdrv_set_backing_hd() to fail. */
+ * job_finish_sync()), block_job_completed() won't free it and therefore
+ * the blockers on the intermediate nodes remain. This would cause
+ * bdrv_set_backing_hd() to fail. */
block_job_remove_all_bdrv(bjob);
block_job_completed(&s->common, ret);
diff --git a/blockjob.c b/blockjob.c
index 0ca7672941..1ed3e9c88d 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -307,40 +307,6 @@ static int block_job_txn_apply(BlockJobTxn *txn, int fn(BlockJob *), bool lock)
return rc;
}
-static int block_job_finish_sync(BlockJob *job,
- void (*finish)(BlockJob *, Error **errp),
- Error **errp)
-{
- Error *local_err = NULL;
- int ret;
-
- assert(blk_bs(job->blk)->job == job);
-
- job_ref(&job->job);
-
- if (finish) {
- finish(job, &local_err);
- }
- if (local_err) {
- error_propagate(errp, local_err);
- job_unref(&job->job);
- return -EBUSY;
- }
- /* job_drain calls job_enter, and it should be enough to induce progress
- * until the job completes or moves to the main thread.
- */
- while (!job->job.deferred_to_main_loop && !job_is_completed(&job->job)) {
- job_drain(&job->job);
- }
- while (!job_is_completed(&job->job)) {
- aio_poll(qemu_get_aio_context(), true);
- }
- ret = (job_is_cancelled(&job->job) && job->job.ret == 0)
- ? -ECANCELED : job->job.ret;
- job_unref(&job->job);
- return ret;
-}
-
static void block_job_completed_txn_abort(BlockJob *job)
{
AioContext *ctx;
@@ -375,7 +341,7 @@ static void block_job_completed_txn_abort(BlockJob *job)
ctx = blk_get_aio_context(other_job->blk);
if (!job_is_completed(&other_job->job)) {
assert(job_is_cancelled(&other_job->job));
- block_job_finish_sync(other_job, NULL, NULL);
+ job_finish_sync(&other_job->job, NULL, NULL);
}
job_finalize_single(&other_job->job);
aio_context_release(ctx);
@@ -528,16 +494,18 @@ void block_job_user_cancel(BlockJob *job, bool force, Error **errp)
}
/* A wrapper around block_job_cancel() taking an Error ** parameter so it may be
- * used with block_job_finish_sync() without the need for (rather nasty)
- * function pointer casts there. */
-static void block_job_cancel_err(BlockJob *job, Error **errp)
+ * used with job_finish_sync() without the need for (rather nasty) function
+ * pointer casts there. */
+static void block_job_cancel_err(Job *job, Error **errp)
{
- block_job_cancel(job, false);
+ BlockJob *bjob = container_of(job, BlockJob, job);
+ assert(is_block_job(job));
+ block_job_cancel(bjob, false);
}
int block_job_cancel_sync(BlockJob *job)
{
- return block_job_finish_sync(job, &block_job_cancel_err, NULL);
+ return job_finish_sync(&job->job, &block_job_cancel_err, NULL);
}
void block_job_cancel_sync_all(void)
@@ -553,14 +521,9 @@ void block_job_cancel_sync_all(void)
}
}
-static void block_job_complete(BlockJob *job, Error **errp)
-{
- job_complete(&job->job, errp);
-}
-
int block_job_complete_sync(BlockJob *job, Error **errp)
{
- return block_job_finish_sync(job, &block_job_complete, errp);
+ return job_finish_sync(&job->job, job_complete, errp);
}
void block_job_progress_update(BlockJob *job, uint64_t done)
diff --git a/job.c b/job.c
index 8a455798e2..7c9a3a7b47 100644
--- a/job.c
+++ b/job.c
@@ -603,3 +603,31 @@ void job_defer_to_main_loop(Job *job, JobDeferToMainLoopFn *fn, void *opaque)
aio_bh_schedule_oneshot(qemu_get_aio_context(),
job_defer_to_main_loop_bh, data);
}
+
+int job_finish_sync(Job *job, void (*finish)(Job *, Error **errp), Error **errp)
+{
+ Error *local_err = NULL;
+ int ret;
+
+ job_ref(job);
+
+ if (finish) {
+ finish(job, &local_err);
+ }
+ if (local_err) {
+ error_propagate(errp, local_err);
+ job_unref(job);
+ return -EBUSY;
+ }
+ /* job_drain calls job_enter, and it should be enough to induce progress
+ * until the job completes or moves to the main thread. */
+ while (!job->deferred_to_main_loop && !job_is_completed(job)) {
+ job_drain(job);
+ }
+ while (!job_is_completed(job)) {
+ aio_poll(qemu_get_aio_context(), true);
+ }
+ ret = (job_is_cancelled(job) && job->ret == 0) ? -ECANCELED : job->ret;
+ job_unref(job);
+ return ret;
+}
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 25/40] job: Switch transactions to JobTxn
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (23 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 24/40] job: Move job_finish_sync() " Kevin Wolf
@ 2018-05-18 13:20 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 26/40] job: Move transactions to Job Kevin Wolf
` (17 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:20 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This doesn't actually move any transaction code to Job yet, but it
renames the type for transactions from BlockJobTxn to JobTxn and makes
them contain Jobs rather than BlockJobs
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/block_int.h | 2 +-
include/block/blockjob.h | 11 ++++----
include/block/blockjob_int.h | 2 +-
include/qemu/job.h | 3 +++
block/backup.c | 2 +-
blockdev.c | 14 +++++------
blockjob.c | 60 +++++++++++++++++++++++---------------------
tests/test-blockjob-txn.c | 8 +++---
8 files changed, 54 insertions(+), 48 deletions(-)
diff --git a/include/block/block_int.h b/include/block/block_int.h
index 76b589da57..6c0927bce3 100644
--- a/include/block/block_int.h
+++ b/include/block/block_int.h
@@ -1029,7 +1029,7 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
BlockdevOnError on_target_error,
int creation_flags,
BlockCompletionFunc *cb, void *opaque,
- BlockJobTxn *txn, Error **errp);
+ JobTxn *txn, Error **errp);
void hmp_drive_add_node(Monitor *mon, const char *optstr);
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 85ce18a381..44df025bd0 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -33,7 +33,7 @@
#define BLOCK_JOB_SLICE_TIME 100000000ULL /* ns */
typedef struct BlockJobDriver BlockJobDriver;
-typedef struct BlockJobTxn BlockJobTxn;
+typedef struct JobTxn JobTxn;
/**
* BlockJob:
@@ -85,8 +85,7 @@ typedef struct BlockJob {
/** BlockDriverStates that are involved in this block job */
GSList *nodes;
- BlockJobTxn *txn;
- QLIST_ENTRY(BlockJob) txn_list;
+ JobTxn *txn;
} BlockJob;
/**
@@ -273,7 +272,7 @@ void block_job_iostatus_reset(BlockJob *job);
* group. Jobs wait for each other before completing. Cancelling one job
* cancels all jobs in the transaction.
*/
-BlockJobTxn *block_job_txn_new(void);
+JobTxn *block_job_txn_new(void);
/**
* block_job_txn_unref:
@@ -282,7 +281,7 @@ BlockJobTxn *block_job_txn_new(void);
* or block_job_txn_new. If it's the last reference to the object, it will be
* freed.
*/
-void block_job_txn_unref(BlockJobTxn *txn);
+void block_job_txn_unref(JobTxn *txn);
/**
* block_job_txn_add_job:
@@ -293,7 +292,7 @@ void block_job_txn_unref(BlockJobTxn *txn);
* The caller must call either block_job_txn_unref() or block_job_completed()
* to release the reference that is automatically grabbed here.
*/
-void block_job_txn_add_job(BlockJobTxn *txn, BlockJob *job);
+void block_job_txn_add_job(JobTxn *txn, BlockJob *job);
/**
* block_job_is_internal:
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index b8ca7bb0c9..ce66a9b51c 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -91,7 +91,7 @@ struct BlockJobDriver {
* called from a wrapper that is specific to the job type.
*/
void *block_job_create(const char *job_id, const BlockJobDriver *driver,
- BlockJobTxn *txn, BlockDriverState *bs, uint64_t perm,
+ JobTxn *txn, BlockDriverState *bs, uint64_t perm,
uint64_t shared_perm, int64_t speed, int flags,
BlockCompletionFunc *cb, void *opaque, Error **errp);
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 17e2ceca87..d4aa7fa981 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -132,6 +132,9 @@ typedef struct Job {
/** Element of the list of jobs */
QLIST_ENTRY(Job) job_list;
+
+ /** Element of the list of jobs in a job transaction */
+ QLIST_ENTRY(Job) txn_list;
} Job;
/**
diff --git a/block/backup.c b/block/backup.c
index ca7d990b21..6172f90c90 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -547,7 +547,7 @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs,
BlockdevOnError on_target_error,
int creation_flags,
BlockCompletionFunc *cb, void *opaque,
- BlockJobTxn *txn, Error **errp)
+ JobTxn *txn, Error **errp)
{
int64_t len;
BlockDriverInfo bdi;
diff --git a/blockdev.c b/blockdev.c
index 0967f6ab66..817c3848c0 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -1446,7 +1446,7 @@ typedef struct BlkActionOps {
struct BlkActionState {
TransactionAction *action;
const BlkActionOps *ops;
- BlockJobTxn *block_job_txn;
+ JobTxn *block_job_txn;
TransactionProperties *txn_props;
QSIMPLEQ_ENTRY(BlkActionState) entry;
};
@@ -1864,7 +1864,7 @@ typedef struct DriveBackupState {
BlockJob *job;
} DriveBackupState;
-static BlockJob *do_drive_backup(DriveBackup *backup, BlockJobTxn *txn,
+static BlockJob *do_drive_backup(DriveBackup *backup, JobTxn *txn,
Error **errp);
static void drive_backup_prepare(BlkActionState *common, Error **errp)
@@ -1954,7 +1954,7 @@ typedef struct BlockdevBackupState {
BlockJob *job;
} BlockdevBackupState;
-static BlockJob *do_blockdev_backup(BlockdevBackup *backup, BlockJobTxn *txn,
+static BlockJob *do_blockdev_backup(BlockdevBackup *backup, JobTxn *txn,
Error **errp);
static void blockdev_backup_prepare(BlkActionState *common, Error **errp)
@@ -2243,7 +2243,7 @@ void qmp_transaction(TransactionActionList *dev_list,
Error **errp)
{
TransactionActionList *dev_entry = dev_list;
- BlockJobTxn *block_job_txn = NULL;
+ JobTxn *block_job_txn = NULL;
BlkActionState *state, *next;
Error *local_err = NULL;
@@ -2251,7 +2251,7 @@ void qmp_transaction(TransactionActionList *dev_list,
QSIMPLEQ_INIT(&snap_bdrv_states);
/* Does this transaction get canceled as a group on failure?
- * If not, we don't really need to make a BlockJobTxn.
+ * If not, we don't really need to make a JobTxn.
*/
props = get_transaction_properties(props);
if (props->completion_mode != ACTION_COMPLETION_MODE_INDIVIDUAL) {
@@ -3264,7 +3264,7 @@ out:
aio_context_release(aio_context);
}
-static BlockJob *do_drive_backup(DriveBackup *backup, BlockJobTxn *txn,
+static BlockJob *do_drive_backup(DriveBackup *backup, JobTxn *txn,
Error **errp)
{
BlockDriverState *bs;
@@ -3434,7 +3434,7 @@ BlockDeviceInfoList *qmp_query_named_block_nodes(Error **errp)
return bdrv_named_nodes_list(errp);
}
-BlockJob *do_blockdev_backup(BlockdevBackup *backup, BlockJobTxn *txn,
+BlockJob *do_blockdev_backup(BlockdevBackup *backup, JobTxn *txn,
Error **errp)
{
BlockDriverState *bs;
diff --git a/blockjob.c b/blockjob.c
index 1ed3e9c88d..bd35c4fa7a 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -37,13 +37,13 @@
#include "qemu/timer.h"
/* Transactional group of block jobs */
-struct BlockJobTxn {
+struct JobTxn {
/* Is this txn being cancelled? */
bool aborting;
/* List of jobs */
- QLIST_HEAD(, BlockJob) jobs;
+ QLIST_HEAD(, Job) jobs;
/* Reference count */
int refcnt;
@@ -94,27 +94,27 @@ BlockJob *block_job_get(const char *id)
}
}
-BlockJobTxn *block_job_txn_new(void)
+JobTxn *block_job_txn_new(void)
{
- BlockJobTxn *txn = g_new0(BlockJobTxn, 1);
+ JobTxn *txn = g_new0(JobTxn, 1);
QLIST_INIT(&txn->jobs);
txn->refcnt = 1;
return txn;
}
-static void block_job_txn_ref(BlockJobTxn *txn)
+static void block_job_txn_ref(JobTxn *txn)
{
txn->refcnt++;
}
-void block_job_txn_unref(BlockJobTxn *txn)
+void block_job_txn_unref(JobTxn *txn)
{
if (txn && --txn->refcnt == 0) {
g_free(txn);
}
}
-void block_job_txn_add_job(BlockJobTxn *txn, BlockJob *job)
+void block_job_txn_add_job(JobTxn *txn, BlockJob *job)
{
if (!txn) {
return;
@@ -123,14 +123,14 @@ void block_job_txn_add_job(BlockJobTxn *txn, BlockJob *job)
assert(!job->txn);
job->txn = txn;
- QLIST_INSERT_HEAD(&txn->jobs, job, txn_list);
+ QLIST_INSERT_HEAD(&txn->jobs, &job->job, txn_list);
block_job_txn_ref(txn);
}
void block_job_txn_del_job(BlockJob *job)
{
if (job->txn) {
- QLIST_REMOVE(job, txn_list);
+ QLIST_REMOVE(&job->job, txn_list);
block_job_txn_unref(job->txn);
job->txn = NULL;
}
@@ -285,18 +285,22 @@ static void job_cancel_async(Job *job, bool force)
job->force_cancel |= force;
}
-static int block_job_txn_apply(BlockJobTxn *txn, int fn(BlockJob *), bool lock)
+static int block_job_txn_apply(JobTxn *txn, int fn(BlockJob *), bool lock)
{
AioContext *ctx;
- BlockJob *job, *next;
+ Job *job, *next;
+ BlockJob *bjob;
int rc = 0;
QLIST_FOREACH_SAFE(job, &txn->jobs, txn_list, next) {
+ assert(is_block_job(job));
+ bjob = container_of(job, BlockJob, job);
+
if (lock) {
- ctx = blk_get_aio_context(job->blk);
+ ctx = job->aio_context;
aio_context_acquire(ctx);
}
- rc = fn(job);
+ rc = fn(bjob);
if (lock) {
aio_context_release(ctx);
}
@@ -310,8 +314,8 @@ static int block_job_txn_apply(BlockJobTxn *txn, int fn(BlockJob *), bool lock)
static void block_job_completed_txn_abort(BlockJob *job)
{
AioContext *ctx;
- BlockJobTxn *txn = job->txn;
- BlockJob *other_job;
+ JobTxn *txn = job->txn;
+ Job *other_job;
if (txn->aborting) {
/*
@@ -324,7 +328,7 @@ static void block_job_completed_txn_abort(BlockJob *job)
/* We are the first failed job. Cancel other jobs. */
QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
- ctx = blk_get_aio_context(other_job->blk);
+ ctx = other_job->aio_context;
aio_context_acquire(ctx);
}
@@ -332,18 +336,18 @@ static void block_job_completed_txn_abort(BlockJob *job)
* them; this job, however, may or may not be cancelled, depending
* on the caller, so leave it. */
QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
- if (other_job != job) {
- job_cancel_async(&other_job->job, false);
+ if (other_job != &job->job) {
+ job_cancel_async(other_job, false);
}
}
while (!QLIST_EMPTY(&txn->jobs)) {
other_job = QLIST_FIRST(&txn->jobs);
- ctx = blk_get_aio_context(other_job->blk);
- if (!job_is_completed(&other_job->job)) {
- assert(job_is_cancelled(&other_job->job));
- job_finish_sync(&other_job->job, NULL, NULL);
+ ctx = other_job->aio_context;
+ if (!job_is_completed(other_job)) {
+ assert(job_is_cancelled(other_job));
+ job_finish_sync(other_job, NULL, NULL);
}
- job_finalize_single(&other_job->job);
+ job_finalize_single(other_job);
aio_context_release(ctx);
}
@@ -385,8 +389,8 @@ static int block_job_transition_to_pending(BlockJob *job)
static void block_job_completed_txn_success(BlockJob *job)
{
- BlockJobTxn *txn = job->txn;
- BlockJob *other_job;
+ JobTxn *txn = job->txn;
+ Job *other_job;
job_state_transition(&job->job, JOB_STATUS_WAITING);
@@ -395,10 +399,10 @@ static void block_job_completed_txn_success(BlockJob *job)
* txn.
*/
QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
- if (!job_is_completed(&other_job->job)) {
+ if (!job_is_completed(other_job)) {
return;
}
- assert(other_job->job.ret == 0);
+ assert(other_job->ret == 0);
}
block_job_txn_apply(txn, block_job_transition_to_pending, false);
@@ -628,7 +632,7 @@ static void block_job_event_pending(Notifier *n, void *opaque)
*/
void *block_job_create(const char *job_id, const BlockJobDriver *driver,
- BlockJobTxn *txn, BlockDriverState *bs, uint64_t perm,
+ JobTxn *txn, BlockDriverState *bs, uint64_t perm,
uint64_t shared_perm, int64_t speed, int flags,
BlockCompletionFunc *cb, void *opaque, Error **errp)
{
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 1572f8d96f..ec5d592b68 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -93,7 +93,7 @@ static const BlockJobDriver test_block_job_driver = {
*/
static BlockJob *test_block_job_start(unsigned int iterations,
bool use_timer,
- int rc, int *result, BlockJobTxn *txn)
+ int rc, int *result, JobTxn *txn)
{
BlockDriverState *bs;
TestBlockJob *s;
@@ -122,7 +122,7 @@ static BlockJob *test_block_job_start(unsigned int iterations,
static void test_single_job(int expected)
{
BlockJob *job;
- BlockJobTxn *txn;
+ JobTxn *txn;
int result = -EINPROGRESS;
txn = block_job_txn_new();
@@ -160,7 +160,7 @@ static void test_pair_jobs(int expected1, int expected2)
{
BlockJob *job1;
BlockJob *job2;
- BlockJobTxn *txn;
+ JobTxn *txn;
int result1 = -EINPROGRESS;
int result2 = -EINPROGRESS;
@@ -222,7 +222,7 @@ static void test_pair_jobs_fail_cancel_race(void)
{
BlockJob *job1;
BlockJob *job2;
- BlockJobTxn *txn;
+ JobTxn *txn;
int result1 = -EINPROGRESS;
int result2 = -EINPROGRESS;
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 26/40] job: Move transactions to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (24 preceding siblings ...)
2018-05-18 13:20 ` [Qemu-devel] [PATCH v2 25/40] job: Switch transactions to JobTxn Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 27/40] job: Move completion and cancellation " Kevin Wolf
` (16 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves the logic that implements job transactions from BlockJob to
Job.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 54 ----------
include/block/blockjob_int.h | 10 --
include/qemu/job.h | 71 +++++++++++--
blockdev.c | 6 +-
blockjob.c | 238 +------------------------------------------
job.c | 234 ++++++++++++++++++++++++++++++++++++++++--
tests/test-blockjob-txn.c | 12 +--
tests/test-blockjob.c | 2 +-
8 files changed, 303 insertions(+), 324 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 44df025bd0..09e6bb42bf 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -33,7 +33,6 @@
#define BLOCK_JOB_SLICE_TIME 100000000ULL /* ns */
typedef struct BlockJobDriver BlockJobDriver;
-typedef struct JobTxn JobTxn;
/**
* BlockJob:
@@ -84,8 +83,6 @@ typedef struct BlockJob {
/** BlockDriverStates that are involved in this block job */
GSList *nodes;
-
- JobTxn *txn;
} BlockJob;
/**
@@ -153,22 +150,6 @@ void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp);
void block_job_cancel(BlockJob *job, bool force);
/**
- * block_job_finalize:
- * @job: The job to fully commit and finish.
- * @errp: Error object.
- *
- * For jobs that have finished their work and are pending
- * awaiting explicit acknowledgement to commit their work,
- * This will commit that work.
- *
- * FIXME: Make the below statement universally true:
- * For jobs that support the manual workflow mode, all graph
- * changes that occur as a result will occur after this command
- * and before a successful reply.
- */
-void block_job_finalize(BlockJob *job, Error **errp);
-
-/**
* block_job_dismiss:
* @job: The job to be dismissed.
* @errp: Error object.
@@ -260,41 +241,6 @@ int block_job_complete_sync(BlockJob *job, Error **errp);
void block_job_iostatus_reset(BlockJob *job);
/**
- * block_job_txn_new:
- *
- * Allocate and return a new block job transaction. Jobs can be added to the
- * transaction using block_job_txn_add_job().
- *
- * The transaction is automatically freed when the last job completes or is
- * cancelled.
- *
- * All jobs in the transaction either complete successfully or fail/cancel as a
- * group. Jobs wait for each other before completing. Cancelling one job
- * cancels all jobs in the transaction.
- */
-JobTxn *block_job_txn_new(void);
-
-/**
- * block_job_txn_unref:
- *
- * Release a reference that was previously acquired with block_job_txn_add_job
- * or block_job_txn_new. If it's the last reference to the object, it will be
- * freed.
- */
-void block_job_txn_unref(JobTxn *txn);
-
-/**
- * block_job_txn_add_job:
- * @txn: The transaction (may be NULL)
- * @job: Job to add to the transaction
- *
- * Add @job to the transaction. The @job must not already be in a transaction.
- * The caller must call either block_job_txn_unref() or block_job_completed()
- * to release the reference that is automatically grabbed here.
- */
-void block_job_txn_add_job(JobTxn *txn, BlockJob *job);
-
-/**
* block_job_is_internal:
* @job: The job to determine if it is user-visible or not.
*
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index ce66a9b51c..29a28020ac 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -38,16 +38,6 @@ struct BlockJobDriver {
/** Generic JobDriver callbacks and settings */
JobDriver job_driver;
- /**
- * If the callback is not NULL, prepare will be invoked when all the jobs
- * belonging to the same transaction complete; or upon this job's completion
- * if it is not in a transaction.
- *
- * This callback will not be invoked if the job has already failed.
- * If it fails, abort and then clean will be called.
- */
- int (*prepare)(BlockJob *job);
-
/*
* If the callback is not NULL, it will be invoked before the job is
* resumed in a new AioContext. This is the place to move any resources
diff --git a/include/qemu/job.h b/include/qemu/job.h
index d4aa7fa981..39d74abdec 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -32,6 +32,8 @@
#include "block/aio.h"
typedef struct JobDriver JobDriver;
+typedef struct JobTxn JobTxn;
+
/**
* Long-running operation.
@@ -133,6 +135,9 @@ typedef struct Job {
/** Element of the list of jobs */
QLIST_ENTRY(Job) job_list;
+ /** Transaction this job is part of */
+ JobTxn *txn;
+
/** Element of the list of jobs in a job transaction */
QLIST_ENTRY(Job) txn_list;
} Job;
@@ -184,6 +189,16 @@ struct JobDriver {
void (*drain)(Job *job);
/**
+ * If the callback is not NULL, prepare will be invoked when all the jobs
+ * belonging to the same transaction complete; or upon this job's completion
+ * if it is not in a transaction.
+ *
+ * This callback will not be invoked if the job has already failed.
+ * If it fails, abort and then clean will be called.
+ */
+ int (*prepare)(Job *job);
+
+ /**
* If the callback is not NULL, it will be invoked when all the jobs
* belonging to the same transaction complete; or upon this job's
* completion if it is not in a transaction. Skipped if NULL.
@@ -227,20 +242,52 @@ typedef enum JobCreateFlags {
JOB_MANUAL_DISMISS = 0x04,
} JobCreateFlags;
+/**
+ * Allocate and return a new job transaction. Jobs can be added to the
+ * transaction using job_txn_add_job().
+ *
+ * The transaction is automatically freed when the last job completes or is
+ * cancelled.
+ *
+ * All jobs in the transaction either complete successfully or fail/cancel as a
+ * group. Jobs wait for each other before completing. Cancelling one job
+ * cancels all jobs in the transaction.
+ */
+JobTxn *job_txn_new(void);
+
+/**
+ * Release a reference that was previously acquired with job_txn_add_job or
+ * job_txn_new. If it's the last reference to the object, it will be freed.
+ */
+void job_txn_unref(JobTxn *txn);
+
+/**
+ * @txn: The transaction (may be NULL)
+ * @job: Job to add to the transaction
+ *
+ * Add @job to the transaction. The @job must not already be in a transaction.
+ * The caller must call either job_txn_unref() or block_job_completed() to
+ * release the reference that is automatically grabbed here.
+ *
+ * If @txn is NULL, the function does nothing.
+ */
+void job_txn_add_job(JobTxn *txn, Job *job);
/**
* Create a new long-running job and return it.
*
* @job_id: The id of the newly-created job, or %NULL for internal jobs
* @driver: The class object for the newly-created job.
+ * @txn: The transaction this job belongs to, if any. %NULL otherwise.
* @ctx: The AioContext to run the job coroutine in.
* @flags: Creation flags for the job. See @JobCreateFlags.
* @cb: Completion function for the job.
* @opaque: Opaque pointer value passed to @cb.
* @errp: Error object.
*/
-void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
- int flags, BlockCompletionFunc *cb, void *opaque, Error **errp);
+void *job_create(const char *job_id, const JobDriver *driver, JobTxn *txn,
+ AioContext *ctx, int flags, BlockCompletionFunc *cb,
+ void *opaque, Error **errp);
/**
* Add a reference to Job refcnt, it will be decreased with job_unref, and then
@@ -260,9 +307,6 @@ void job_event_cancelled(Job *job);
/** To be called when a successfully completed job is finalised. */
void job_event_completed(Job *job);
-/** To be called when the job transitions to PENDING */
-void job_event_pending(Job *job);
-
/**
* Conditionally enter the job coroutine if the job is ready to run, not
* already busy and fn() returns true. fn() is called while under the job_lock
@@ -375,6 +419,16 @@ void job_early_fail(Job *job);
/** Asynchronously complete the specified @job. */
void job_complete(Job *job, Error **errp);;
+/**
+ * For a @job that has finished its work and is pending awaiting explicit
+ * acknowledgement to commit its work, this will commit that work.
+ *
+ * FIXME: Make the below statement universally true:
+ * For jobs that support the manual workflow mode, all graph changes that occur
+ * as a result will occur after this command and before a successful reply.
+ */
+void job_finalize(Job *job, Error **errp);
+
typedef void JobDeferToMainLoopFn(Job *job, void *opaque);
/**
@@ -407,10 +461,9 @@ void coroutine_fn job_do_yield(Job *job, uint64_t ns);
bool job_should_pause(Job *job);
bool job_started(Job *job);
void job_do_dismiss(Job *job);
-int job_finalize_single(Job *job);
void job_update_rc(Job *job);
-
-typedef struct BlockJob BlockJob;
-void block_job_txn_del_job(BlockJob *job);
+void job_cancel_async(Job *job, bool force);
+void job_completed_txn_abort(Job *job);
+void job_completed_txn_success(Job *job);
#endif
diff --git a/blockdev.c b/blockdev.c
index 817c3848c0..87a23bed6f 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -2255,7 +2255,7 @@ void qmp_transaction(TransactionActionList *dev_list,
*/
props = get_transaction_properties(props);
if (props->completion_mode != ACTION_COMPLETION_MODE_INDIVIDUAL) {
- block_job_txn = block_job_txn_new();
+ block_job_txn = job_txn_new();
}
/* drain all i/o before any operations */
@@ -2314,7 +2314,7 @@ exit:
if (!has_props) {
qapi_free_TransactionProperties(props);
}
- block_job_txn_unref(block_job_txn);
+ job_txn_unref(block_job_txn);
}
void qmp_eject(bool has_device, const char *device,
@@ -3908,7 +3908,7 @@ void qmp_block_job_finalize(const char *id, Error **errp)
}
trace_qmp_block_job_finalize(job);
- block_job_finalize(job, errp);
+ job_finalize(&job->job, errp);
aio_context_release(aio_context);
}
diff --git a/blockjob.c b/blockjob.c
index bd35c4fa7a..14b21c86b7 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -36,19 +36,6 @@
#include "qemu/coroutine.h"
#include "qemu/timer.h"
-/* Transactional group of block jobs */
-struct JobTxn {
-
- /* Is this txn being cancelled? */
- bool aborting;
-
- /* List of jobs */
- QLIST_HEAD(, Job) jobs;
-
- /* Reference count */
- int refcnt;
-};
-
/*
* The block job API is composed of two categories of functions.
*
@@ -94,48 +81,6 @@ BlockJob *block_job_get(const char *id)
}
}
-JobTxn *block_job_txn_new(void)
-{
- JobTxn *txn = g_new0(JobTxn, 1);
- QLIST_INIT(&txn->jobs);
- txn->refcnt = 1;
- return txn;
-}
-
-static void block_job_txn_ref(JobTxn *txn)
-{
- txn->refcnt++;
-}
-
-void block_job_txn_unref(JobTxn *txn)
-{
- if (txn && --txn->refcnt == 0) {
- g_free(txn);
- }
-}
-
-void block_job_txn_add_job(JobTxn *txn, BlockJob *job)
-{
- if (!txn) {
- return;
- }
-
- assert(!job->txn);
- job->txn = txn;
-
- QLIST_INSERT_HEAD(&txn->jobs, &job->job, txn_list);
- block_job_txn_ref(txn);
-}
-
-void block_job_txn_del_job(BlockJob *job)
-{
- if (job->txn) {
- QLIST_REMOVE(&job->job, txn_list);
- block_job_txn_unref(job->txn);
- job->txn = NULL;
- }
-}
-
static void block_job_attached_aio_context(AioContext *new_context,
void *opaque);
static void block_job_detach_aio_context(void *opaque);
@@ -145,8 +90,6 @@ void block_job_free(Job *job)
BlockJob *bjob = container_of(job, BlockJob, job);
BlockDriverState *bs = blk_bs(bjob->blk);
- assert(!bjob->txn);
-
bs->job = NULL;
block_job_remove_all_bdrv(bjob);
blk_remove_aio_context_notifier(bjob->blk,
@@ -261,158 +204,6 @@ const BlockJobDriver *block_job_driver(BlockJob *job)
return job->driver;
}
-static int block_job_prepare(BlockJob *job)
-{
- if (job->job.ret == 0 && job->driver->prepare) {
- job->job.ret = job->driver->prepare(job);
- }
- return job->job.ret;
-}
-
-static void job_cancel_async(Job *job, bool force)
-{
- if (job->user_paused) {
- /* Do not call job_enter here, the caller will handle it. */
- job->user_paused = false;
- if (job->driver->user_resume) {
- job->driver->user_resume(job);
- }
- assert(job->pause_count > 0);
- job->pause_count--;
- }
- job->cancelled = true;
- /* To prevent 'force == false' overriding a previous 'force == true' */
- job->force_cancel |= force;
-}
-
-static int block_job_txn_apply(JobTxn *txn, int fn(BlockJob *), bool lock)
-{
- AioContext *ctx;
- Job *job, *next;
- BlockJob *bjob;
- int rc = 0;
-
- QLIST_FOREACH_SAFE(job, &txn->jobs, txn_list, next) {
- assert(is_block_job(job));
- bjob = container_of(job, BlockJob, job);
-
- if (lock) {
- ctx = job->aio_context;
- aio_context_acquire(ctx);
- }
- rc = fn(bjob);
- if (lock) {
- aio_context_release(ctx);
- }
- if (rc) {
- break;
- }
- }
- return rc;
-}
-
-static void block_job_completed_txn_abort(BlockJob *job)
-{
- AioContext *ctx;
- JobTxn *txn = job->txn;
- Job *other_job;
-
- if (txn->aborting) {
- /*
- * We are cancelled by another job, which will handle everything.
- */
- return;
- }
- txn->aborting = true;
- block_job_txn_ref(txn);
-
- /* We are the first failed job. Cancel other jobs. */
- QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
- ctx = other_job->aio_context;
- aio_context_acquire(ctx);
- }
-
- /* Other jobs are effectively cancelled by us, set the status for
- * them; this job, however, may or may not be cancelled, depending
- * on the caller, so leave it. */
- QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
- if (other_job != &job->job) {
- job_cancel_async(other_job, false);
- }
- }
- while (!QLIST_EMPTY(&txn->jobs)) {
- other_job = QLIST_FIRST(&txn->jobs);
- ctx = other_job->aio_context;
- if (!job_is_completed(other_job)) {
- assert(job_is_cancelled(other_job));
- job_finish_sync(other_job, NULL, NULL);
- }
- job_finalize_single(other_job);
- aio_context_release(ctx);
- }
-
- block_job_txn_unref(txn);
-}
-
-static int block_job_needs_finalize(BlockJob *job)
-{
- return !job->job.auto_finalize;
-}
-
-static int block_job_finalize_single(BlockJob *job)
-{
- return job_finalize_single(&job->job);
-}
-
-static void block_job_do_finalize(BlockJob *job)
-{
- int rc;
- assert(job && job->txn);
-
- /* prepare the transaction to complete */
- rc = block_job_txn_apply(job->txn, block_job_prepare, true);
- if (rc) {
- block_job_completed_txn_abort(job);
- } else {
- block_job_txn_apply(job->txn, block_job_finalize_single, true);
- }
-}
-
-static int block_job_transition_to_pending(BlockJob *job)
-{
- job_state_transition(&job->job, JOB_STATUS_PENDING);
- if (!job->job.auto_finalize) {
- job_event_pending(&job->job);
- }
- return 0;
-}
-
-static void block_job_completed_txn_success(BlockJob *job)
-{
- JobTxn *txn = job->txn;
- Job *other_job;
-
- job_state_transition(&job->job, JOB_STATUS_WAITING);
-
- /*
- * Successful completion, see if there are other running jobs in this
- * txn.
- */
- QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
- if (!job_is_completed(other_job)) {
- return;
- }
- assert(other_job->ret == 0);
- }
-
- block_job_txn_apply(txn, block_job_transition_to_pending, false);
-
- /* If no jobs need manual finalization, automatically do so */
- if (block_job_txn_apply(txn, block_job_needs_finalize, false) == 0) {
- block_job_do_finalize(job);
- }
-}
-
/* Assumes the job_mutex is held */
static bool job_timer_pending(Job *job)
{
@@ -451,15 +242,6 @@ int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n)
return ratelimit_calculate_delay(&job->limit, n);
}
-void block_job_finalize(BlockJob *job, Error **errp)
-{
- assert(job && job->job.id);
- if (job_apply_verb(&job->job, JOB_VERB_FINALIZE, errp)) {
- return;
- }
- block_job_do_finalize(job);
-}
-
void block_job_dismiss(BlockJob **jobptr, Error **errp)
{
BlockJob *job = *jobptr;
@@ -483,7 +265,7 @@ void block_job_cancel(BlockJob *job, bool force)
if (!job_started(&job->job)) {
block_job_completed(job, -ECANCELED);
} else if (job->job.deferred_to_main_loop) {
- block_job_completed_txn_abort(job);
+ job_completed_txn_abort(&job->job);
} else {
block_job_enter(job);
}
@@ -656,7 +438,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
return NULL;
}
- job = job_create(job_id, &driver->job_driver, blk_get_aio_context(blk),
+ job = job_create(job_id, &driver->job_driver, txn, blk_get_aio_context(blk),
flags, cb, opaque, errp);
if (job == NULL) {
blk_unref(blk);
@@ -703,30 +485,20 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
}
}
- /* Single jobs are modeled as single-job transactions for sake of
- * consolidating the job management logic */
- if (!txn) {
- txn = block_job_txn_new();
- block_job_txn_add_job(txn, job);
- block_job_txn_unref(txn);
- } else {
- block_job_txn_add_job(txn, job);
- }
-
return job;
}
void block_job_completed(BlockJob *job, int ret)
{
- assert(job && job->txn && !job_is_completed(&job->job));
+ assert(job && job->job.txn && !job_is_completed(&job->job));
assert(blk_bs(job->blk)->job == job);
job->job.ret = ret;
job_update_rc(&job->job);
trace_block_job_completed(job, ret, job->job.ret);
if (job->job.ret) {
- block_job_completed_txn_abort(job);
+ job_completed_txn_abort(&job->job);
} else {
- block_job_completed_txn_success(job);
+ job_completed_txn_success(&job->job);
}
}
diff --git a/job.c b/job.c
index 7c9a3a7b47..ee0bf42c5d 100644
--- a/job.c
+++ b/job.c
@@ -60,6 +60,19 @@ bool JobVerbTable[JOB_VERB__MAX][JOB_STATUS__MAX] = {
[JOB_VERB_DISMISS] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0},
};
+/* Transactional group of jobs */
+struct JobTxn {
+
+ /* Is this txn being cancelled? */
+ bool aborting;
+
+ /* List of jobs */
+ QLIST_HEAD(, Job) jobs;
+
+ /* Reference count */
+ int refcnt;
+};
+
/* Right now, this mutex is only needed to synchronize accesses to job->busy
* and job->sleep_timer, such as concurrent calls to job_do_yield and
* job_enter. */
@@ -80,6 +93,71 @@ static void __attribute__((__constructor__)) job_init(void)
qemu_mutex_init(&job_mutex);
}
+JobTxn *job_txn_new(void)
+{
+ JobTxn *txn = g_new0(JobTxn, 1);
+ QLIST_INIT(&txn->jobs);
+ txn->refcnt = 1;
+ return txn;
+}
+
+static void job_txn_ref(JobTxn *txn)
+{
+ txn->refcnt++;
+}
+
+void job_txn_unref(JobTxn *txn)
+{
+ if (txn && --txn->refcnt == 0) {
+ g_free(txn);
+ }
+}
+
+void job_txn_add_job(JobTxn *txn, Job *job)
+{
+ if (!txn) {
+ return;
+ }
+
+ assert(!job->txn);
+ job->txn = txn;
+
+ QLIST_INSERT_HEAD(&txn->jobs, job, txn_list);
+ job_txn_ref(txn);
+}
+
+static void job_txn_del_job(Job *job)
+{
+ if (job->txn) {
+ QLIST_REMOVE(job, txn_list);
+ job_txn_unref(job->txn);
+ job->txn = NULL;
+ }
+}
+
+static int job_txn_apply(JobTxn *txn, int fn(Job *), bool lock)
+{
+ AioContext *ctx;
+ Job *job, *next;
+ int rc = 0;
+
+ QLIST_FOREACH_SAFE(job, &txn->jobs, txn_list, next) {
+ if (lock) {
+ ctx = job->aio_context;
+ aio_context_acquire(ctx);
+ }
+ rc = fn(job);
+ if (lock) {
+ aio_context_release(ctx);
+ }
+ if (rc) {
+ break;
+ }
+ }
+ return rc;
+}
+
+
/* TODO Make static once the whole state machine is in job.c */
void job_state_transition(Job *job, JobStatus s1)
{
@@ -181,8 +259,9 @@ static void job_sleep_timer_cb(void *opaque)
job_enter(job);
}
-void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
- int flags, BlockCompletionFunc *cb, void *opaque, Error **errp)
+void *job_create(const char *job_id, const JobDriver *driver, JobTxn *txn,
+ AioContext *ctx, int flags, BlockCompletionFunc *cb,
+ void *opaque, Error **errp)
{
Job *job;
@@ -228,6 +307,16 @@ void *job_create(const char *job_id, const JobDriver *driver, AioContext *ctx,
QLIST_INSERT_HEAD(&jobs, job, job_list);
+ /* Single jobs are modeled as single-job transactions for sake of
+ * consolidating the job management logic */
+ if (!txn) {
+ txn = job_txn_new();
+ job_txn_add_job(txn, job);
+ job_txn_unref(txn);
+ } else {
+ job_txn_add_job(txn, job);
+ }
+
return job;
}
@@ -241,6 +330,7 @@ void job_unref(Job *job)
if (--job->refcnt == 0) {
assert(job->status == JOB_STATUS_NULL);
assert(!timer_pending(&job->sleep_timer));
+ assert(!job->txn);
if (job->driver->free) {
job->driver->free(job);
@@ -263,7 +353,7 @@ void job_event_completed(Job *job)
notifier_list_notify(&job->on_finalize_completed, job);
}
-void job_event_pending(Job *job)
+static void job_event_pending(Job *job)
{
notifier_list_notify(&job->on_pending, job);
}
@@ -469,8 +559,7 @@ void job_do_dismiss(Job *job)
job->paused = false;
job->deferred_to_main_loop = true;
- /* TODO Don't assume it's a BlockJob */
- block_job_txn_del_job((BlockJob*) job);
+ job_txn_del_job(job);
job_state_transition(job, JOB_STATUS_NULL);
job_unref(job);
@@ -523,7 +612,7 @@ static void job_clean(Job *job)
}
}
-int job_finalize_single(Job *job)
+static int job_finalize_single(Job *job)
{
assert(job_is_completed(job));
@@ -550,12 +639,141 @@ int job_finalize_single(Job *job)
}
}
- /* TODO Don't assume it's a BlockJob */
- block_job_txn_del_job((BlockJob*) job);
+ job_txn_del_job(job);
job_conclude(job);
return 0;
}
+void job_cancel_async(Job *job, bool force)
+{
+ if (job->user_paused) {
+ /* Do not call job_enter here, the caller will handle it. */
+ job->user_paused = false;
+ if (job->driver->user_resume) {
+ job->driver->user_resume(job);
+ }
+ assert(job->pause_count > 0);
+ job->pause_count--;
+ }
+ job->cancelled = true;
+ /* To prevent 'force == false' overriding a previous 'force == true' */
+ job->force_cancel |= force;
+}
+
+void job_completed_txn_abort(Job *job)
+{
+ AioContext *ctx;
+ JobTxn *txn = job->txn;
+ Job *other_job;
+
+ if (txn->aborting) {
+ /*
+ * We are cancelled by another job, which will handle everything.
+ */
+ return;
+ }
+ txn->aborting = true;
+ job_txn_ref(txn);
+
+ /* We are the first failed job. Cancel other jobs. */
+ QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
+ ctx = other_job->aio_context;
+ aio_context_acquire(ctx);
+ }
+
+ /* Other jobs are effectively cancelled by us, set the status for
+ * them; this job, however, may or may not be cancelled, depending
+ * on the caller, so leave it. */
+ QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
+ if (other_job != job) {
+ job_cancel_async(other_job, false);
+ }
+ }
+ while (!QLIST_EMPTY(&txn->jobs)) {
+ other_job = QLIST_FIRST(&txn->jobs);
+ ctx = other_job->aio_context;
+ if (!job_is_completed(other_job)) {
+ assert(job_is_cancelled(other_job));
+ job_finish_sync(other_job, NULL, NULL);
+ }
+ job_finalize_single(other_job);
+ aio_context_release(ctx);
+ }
+
+ job_txn_unref(txn);
+}
+
+static int job_prepare(Job *job)
+{
+ if (job->ret == 0 && job->driver->prepare) {
+ job->ret = job->driver->prepare(job);
+ }
+ return job->ret;
+}
+
+static int job_needs_finalize(Job *job)
+{
+ return !job->auto_finalize;
+}
+
+static void job_do_finalize(Job *job)
+{
+ int rc;
+ assert(job && job->txn);
+
+ /* prepare the transaction to complete */
+ rc = job_txn_apply(job->txn, job_prepare, true);
+ if (rc) {
+ job_completed_txn_abort(job);
+ } else {
+ job_txn_apply(job->txn, job_finalize_single, true);
+ }
+}
+
+void job_finalize(Job *job, Error **errp)
+{
+ assert(job && job->id);
+ if (job_apply_verb(job, JOB_VERB_FINALIZE, errp)) {
+ return;
+ }
+ job_do_finalize(job);
+}
+
+static int job_transition_to_pending(Job *job)
+{
+ job_state_transition(job, JOB_STATUS_PENDING);
+ if (!job->auto_finalize) {
+ job_event_pending(job);
+ }
+ return 0;
+}
+
+void job_completed_txn_success(Job *job)
+{
+ JobTxn *txn = job->txn;
+ Job *other_job;
+
+ job_state_transition(job, JOB_STATUS_WAITING);
+
+ /*
+ * Successful completion, see if there are other running jobs in this
+ * txn.
+ */
+ QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
+ if (!job_is_completed(other_job)) {
+ return;
+ }
+ assert(other_job->ret == 0);
+ }
+
+ job_txn_apply(txn, job_transition_to_pending, false);
+
+ /* If no jobs need manual finalization, automatically do so */
+ if (job_txn_apply(txn, job_needs_finalize, false) == 0) {
+ job_do_finalize(job);
+ }
+}
+
void job_complete(Job *job, Error **errp)
{
/* Should not be reachable via external interface for internal jobs */
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index ec5d592b68..6ee31d59ad 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -125,7 +125,7 @@ static void test_single_job(int expected)
JobTxn *txn;
int result = -EINPROGRESS;
- txn = block_job_txn_new();
+ txn = job_txn_new();
job = test_block_job_start(1, true, expected, &result, txn);
job_start(&job->job);
@@ -138,7 +138,7 @@ static void test_single_job(int expected)
}
g_assert_cmpint(result, ==, expected);
- block_job_txn_unref(txn);
+ job_txn_unref(txn);
}
static void test_single_job_success(void)
@@ -164,7 +164,7 @@ static void test_pair_jobs(int expected1, int expected2)
int result1 = -EINPROGRESS;
int result2 = -EINPROGRESS;
- txn = block_job_txn_new();
+ txn = job_txn_new();
job1 = test_block_job_start(1, true, expected1, &result1, txn);
job2 = test_block_job_start(2, true, expected2, &result2, txn);
job_start(&job1->job);
@@ -173,7 +173,7 @@ static void test_pair_jobs(int expected1, int expected2)
/* Release our reference now to trigger as many nice
* use-after-free bugs as possible.
*/
- block_job_txn_unref(txn);
+ job_txn_unref(txn);
if (expected1 == -ECANCELED) {
block_job_cancel(job1, false);
@@ -226,7 +226,7 @@ static void test_pair_jobs_fail_cancel_race(void)
int result1 = -EINPROGRESS;
int result2 = -EINPROGRESS;
- txn = block_job_txn_new();
+ txn = job_txn_new();
job1 = test_block_job_start(1, true, -ECANCELED, &result1, txn);
job2 = test_block_job_start(2, false, 0, &result2, txn);
job_start(&job1->job);
@@ -247,7 +247,7 @@ static void test_pair_jobs_fail_cancel_race(void)
g_assert_cmpint(result1, ==, -ECANCELED);
g_assert_cmpint(result2, ==, -ECANCELED);
- block_job_txn_unref(txn);
+ job_txn_unref(txn);
}
int main(int argc, char **argv)
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index e44c608327..1e052c2e9c 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -364,7 +364,7 @@ static void test_cancel_concluded(void)
}
assert(job->job.status == JOB_STATUS_PENDING);
- block_job_finalize(job, &error_abort);
+ job_finalize(&job->job, &error_abort);
assert(job->job.status == JOB_STATUS_CONCLUDED);
cancel_common(s);
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 27/40] job: Move completion and cancellation to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (25 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 26/40] job: Move transactions to Job Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 28/40] block: Cancel job in bdrv_close_all() callers Kevin Wolf
` (15 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves the top-level job completion and cancellation functions from
BlockJob to Job.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
include/block/blockjob.h | 55 -----------------------------
include/block/blockjob_int.h | 18 ----------
include/qemu/job.h | 68 +++++++++++++++++++++++++++++------
block.c | 4 ++-
block/backup.c | 3 +-
block/commit.c | 6 ++--
block/mirror.c | 6 ++--
block/replication.c | 4 +--
block/stream.c | 2 +-
blockdev.c | 8 ++---
blockjob.c | 76 ---------------------------------------
job.c | 84 +++++++++++++++++++++++++++++++++++++++-----
qemu-img.c | 2 +-
tests/test-bdrv-drain.c | 5 ++-
tests/test-blockjob-txn.c | 14 ++++----
tests/test-blockjob.c | 21 ++++++-----
block/trace-events | 3 --
trace-events | 1 +
18 files changed, 171 insertions(+), 209 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 09e6bb42bf..e9ed7b8b78 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -141,15 +141,6 @@ void block_job_remove_all_bdrv(BlockJob *job);
void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp);
/**
- * block_job_cancel:
- * @job: The job to be canceled.
- * @force: Quit a job without waiting for data to be in sync.
- *
- * Asynchronously cancel the specified job.
- */
-void block_job_cancel(BlockJob *job, bool force);
-
-/**
* block_job_dismiss:
* @job: The job to be dismissed.
* @errp: Error object.
@@ -186,52 +177,6 @@ void block_job_progress_set_remaining(BlockJob *job, uint64_t remaining);
BlockJobInfo *block_job_query(BlockJob *job, Error **errp);
/**
- * block_job_user_cancel:
- * @job: The job to be cancelled.
- * @force: Quit a job without waiting for data to be in sync.
- *
- * Cancels the specified job, but may refuse to do so if the
- * operation isn't currently meaningful.
- */
-void block_job_user_cancel(BlockJob *job, bool force, Error **errp);
-
-/**
- * block_job_cancel_sync:
- * @job: The job to be canceled.
- *
- * Synchronously cancel the job. The completion callback is called
- * before the function returns. The job may actually complete
- * instead of canceling itself; the circumstances under which this
- * happens depend on the kind of job that is active.
- *
- * Returns the return value from the job if the job actually completed
- * during the call, or -ECANCELED if it was canceled.
- */
-int block_job_cancel_sync(BlockJob *job);
-
-/**
- * block_job_cancel_sync_all:
- *
- * Synchronously cancels all jobs using block_job_cancel_sync().
- */
-void block_job_cancel_sync_all(void);
-
-/**
- * block_job_complete_sync:
- * @job: The job to be completed.
- * @errp: Error object which may be set by block_job_complete(); this is not
- * necessarily set on every error, the job return value has to be
- * checked as well.
- *
- * Synchronously complete the job. The completion callback is called before the
- * function returns, unless it is NULL (which is permissible when using this
- * function).
- *
- * Returns the return value from the job.
- */
-int block_job_complete_sync(BlockJob *job, Error **errp);
-
-/**
* block_job_iostatus_reset:
* @job: The job whose I/O status should be reset.
*
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 29a28020ac..7df07b20cf 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -124,24 +124,6 @@ void block_job_yield(BlockJob *job);
int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n);
/**
- * block_job_completed:
- * @job: The job being completed.
- * @ret: The status code.
- *
- * Call the completion function that was registered at creation time, and
- * free @job.
- */
-void block_job_completed(BlockJob *job, int ret);
-
-/**
- * block_job_enter:
- * @job: The job to enter.
- *
- * Continue the specified job by entering the coroutine.
- */
-void block_job_enter(BlockJob *job);
-
-/**
* block_job_event_ready:
* @job: The job which is now ready to be completed.
*
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 39d74abdec..bbe1b0cd1a 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -74,8 +74,8 @@ typedef struct Job {
/**
* Set to false by the job while the coroutine has yielded and may be
- * re-entered by block_job_enter(). There may still be I/O or event loop
- * activity pending. Accessed under block_job_mutex (in blockjob.c).
+ * re-entered by job_enter(). There may still be I/O or event loop activity
+ * pending. Accessed under block_job_mutex (in blockjob.c).
*/
bool busy;
@@ -114,7 +114,7 @@ typedef struct Job {
/** True if this job should automatically dismiss itself */
bool auto_dismiss;
- /** ret code passed to block_job_completed. */
+ /** ret code passed to job_completed. */
int ret;
/** The completion function that will be called when the job completes. */
@@ -266,8 +266,8 @@ void job_txn_unref(JobTxn *txn);
* @job: Job to add to the transaction
*
* Add @job to the transaction. The @job must not already be in a transaction.
- * The caller must call either job_txn_unref() or block_job_completed() to
- * release the reference that is automatically grabbed here.
+ * The caller must call either job_txn_unref() or job_completed() to release
+ * the reference that is automatically grabbed here.
*
* If @txn is NULL, the function does nothing.
*/
@@ -416,8 +416,59 @@ int job_apply_verb(Job *job, JobVerb verb, Error **errp);
/** The @job could not be started, free it. */
void job_early_fail(Job *job);
+/**
+ * @job: The job being completed.
+ * @ret: The status code.
+ *
+ * Marks @job as completed. If @ret is non-zero, the job transaction it is part
+ * of is aborted. If @ret is zero, the job moves into the WAITING state. If it
+ * is the last job to complete in its transaction, all jobs in the transaction
+ * move from WAITING to PENDING.
+ */
+void job_completed(Job *job, int ret);
+
/** Asynchronously complete the specified @job. */
-void job_complete(Job *job, Error **errp);;
+void job_complete(Job *job, Error **errp);
+
+/**
+ * Asynchronously cancel the specified @job. If @force is true, the job should
+ * be cancelled immediately without waiting for a consistent state.
+ */
+void job_cancel(Job *job, bool force);
+
+/**
+ * Cancels the specified job like job_cancel(), but may refuse to do so if the
+ * operation isn't meaningful in the current state of the job.
+ */
+void job_user_cancel(Job *job, bool force, Error **errp);
+
+/**
+ * Synchronously cancel the @job. The completion callback is called
+ * before the function returns. The job may actually complete
+ * instead of canceling itself; the circumstances under which this
+ * happens depend on the kind of job that is active.
+ *
+ * Returns the return value from the job if the job actually completed
+ * during the call, or -ECANCELED if it was canceled.
+ */
+int job_cancel_sync(Job *job);
+
+/** Synchronously cancels all jobs using job_cancel_sync(). */
+void job_cancel_sync_all(void);
+
+/**
+ * @job: The job to be completed.
+ * @errp: Error object which may be set by job_complete(); this is not
+ * necessarily set on every error, the job return value has to be
+ * checked as well.
+ *
+ * Synchronously complete the job. The completion callback is called before the
+ * function returns, unless it is NULL (which is permissible when using this
+ * function).
+ *
+ * Returns the return value from the job.
+ */
+int job_complete_sync(Job *job, Error **errp);
/**
* For a @job that has finished its work and is pending awaiting explicit
@@ -459,11 +510,6 @@ int job_finish_sync(Job *job, void (*finish)(Job *, Error **errp), Error **errp)
void job_state_transition(Job *job, JobStatus s1);
void coroutine_fn job_do_yield(Job *job, uint64_t ns);
bool job_should_pause(Job *job);
-bool job_started(Job *job);
void job_do_dismiss(Job *job);
-void job_update_rc(Job *job);
-void job_cancel_async(Job *job, bool force);
-void job_completed_txn_abort(Job *job);
-void job_completed_txn_success(Job *job);
#endif
diff --git a/block.c b/block.c
index 676e57f562..d50b08c29b 100644
--- a/block.c
+++ b/block.c
@@ -3362,7 +3362,9 @@ static void bdrv_close(BlockDriverState *bs)
void bdrv_close_all(void)
{
- block_job_cancel_sync_all();
+ /* TODO We do want to cancel all jobs instead of just block jobs on
+ * shutdown, but bdrv_close_all() isn't the right place any more. */
+ job_cancel_sync_all();
nbd_export_close_all();
/* Drop references from requests still in flight, such as canceled block
diff --git a/block/backup.c b/block/backup.c
index 6172f90c90..b13f91d8a7 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -319,10 +319,9 @@ typedef struct {
static void backup_complete(Job *job, void *opaque)
{
- BlockJob *bjob = container_of(job, BlockJob, job);
BackupCompleteData *data = opaque;
- block_job_completed(bjob, data->ret);
+ job_completed(job, data->ret);
g_free(data);
}
diff --git a/block/commit.c b/block/commit.c
index 40d97a35a5..b0a847e678 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -112,12 +112,12 @@ static void commit_complete(Job *job, void *opaque)
blk_unref(s->top);
/* If there is more than one reference to the job (e.g. if called from
- * job_finish_sync()), block_job_completed() won't free it and therefore
- * the blockers on the intermediate nodes remain. This would cause
+ * job_finish_sync()), job_completed() won't free it and therefore the
+ * blockers on the intermediate nodes remain. This would cause
* bdrv_set_backing_hd() to fail. */
block_job_remove_all_bdrv(bjob);
- block_job_completed(&s->common, ret);
+ job_completed(job, ret);
g_free(data);
/* If bdrv_drop_intermediate() didn't already do that, remove the commit
diff --git a/block/mirror.c b/block/mirror.c
index 656237af5c..c63cf7cbc9 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -498,7 +498,7 @@ static void mirror_exit(Job *job, void *opaque)
bdrv_release_dirty_bitmap(src, s->dirty_bitmap);
/* Make sure that the source BDS doesn't go away before we called
- * block_job_completed(). */
+ * job_completed(). */
bdrv_ref(src);
bdrv_ref(mirror_top_bs);
bdrv_ref(target_bs);
@@ -581,7 +581,7 @@ static void mirror_exit(Job *job, void *opaque)
blk_set_perm(bjob->blk, 0, BLK_PERM_ALL, &error_abort);
blk_insert_bs(bjob->blk, mirror_top_bs, &error_abort);
- block_job_completed(&s->common, data->ret);
+ job_completed(job, data->ret);
g_free(data);
bdrv_drained_end(src);
@@ -954,7 +954,7 @@ static void mirror_complete(Job *job, Error **errp)
}
s->should_complete = true;
- block_job_enter(&s->common);
+ job_enter(job);
}
static void mirror_pause(Job *job)
diff --git a/block/replication.c b/block/replication.c
index ff2e2028af..826db7b304 100644
--- a/block/replication.c
+++ b/block/replication.c
@@ -145,7 +145,7 @@ static void replication_close(BlockDriverState *bs)
replication_stop(s->rs, false, NULL);
}
if (s->stage == BLOCK_REPLICATION_FAILOVER) {
- block_job_cancel_sync(s->active_disk->bs->job);
+ job_cancel_sync(&s->active_disk->bs->job->job);
}
if (s->mode == REPLICATION_MODE_SECONDARY) {
@@ -681,7 +681,7 @@ static void replication_stop(ReplicationState *rs, bool failover, Error **errp)
* disk, secondary disk in backup_job_completed().
*/
if (s->secondary_disk->bs->job) {
- block_job_cancel_sync(s->secondary_disk->bs->job);
+ job_cancel_sync(&s->secondary_disk->bs->job->job);
}
if (!failover) {
diff --git a/block/stream.c b/block/stream.c
index b996278ab3..8546c412cd 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -93,7 +93,7 @@ out:
}
g_free(s->backing_file_str);
- block_job_completed(&s->common, data->ret);
+ job_completed(job, data->ret);
g_free(data);
}
diff --git a/blockdev.c b/blockdev.c
index 87a23bed6f..31319a6d5a 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -150,7 +150,7 @@ void blockdev_mark_auto_del(BlockBackend *blk)
aio_context_acquire(aio_context);
if (bs->job) {
- block_job_cancel(bs->job, false);
+ job_cancel(&bs->job->job, false);
}
aio_context_release(aio_context);
@@ -1925,7 +1925,7 @@ static void drive_backup_abort(BlkActionState *common)
aio_context = bdrv_get_aio_context(state->bs);
aio_context_acquire(aio_context);
- block_job_cancel_sync(state->job);
+ job_cancel_sync(&state->job->job);
aio_context_release(aio_context);
}
@@ -2023,7 +2023,7 @@ static void blockdev_backup_abort(BlkActionState *common)
aio_context = bdrv_get_aio_context(state->bs);
aio_context_acquire(aio_context);
- block_job_cancel_sync(state->job);
+ job_cancel_sync(&state->job->job);
aio_context_release(aio_context);
}
@@ -3851,7 +3851,7 @@ void qmp_block_job_cancel(const char *device,
}
trace_qmp_block_job_cancel(job);
- block_job_user_cancel(job, force, errp);
+ job_user_cancel(&job->job, force, errp);
out:
aio_context_release(aio_context);
}
diff --git a/blockjob.c b/blockjob.c
index 14b21c86b7..438baa1778 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -255,63 +255,6 @@ void block_job_dismiss(BlockJob **jobptr, Error **errp)
*jobptr = NULL;
}
-void block_job_cancel(BlockJob *job, bool force)
-{
- if (job->job.status == JOB_STATUS_CONCLUDED) {
- job_do_dismiss(&job->job);
- return;
- }
- job_cancel_async(&job->job, force);
- if (!job_started(&job->job)) {
- block_job_completed(job, -ECANCELED);
- } else if (job->job.deferred_to_main_loop) {
- job_completed_txn_abort(&job->job);
- } else {
- block_job_enter(job);
- }
-}
-
-void block_job_user_cancel(BlockJob *job, bool force, Error **errp)
-{
- if (job_apply_verb(&job->job, JOB_VERB_CANCEL, errp)) {
- return;
- }
- block_job_cancel(job, force);
-}
-
-/* A wrapper around block_job_cancel() taking an Error ** parameter so it may be
- * used with job_finish_sync() without the need for (rather nasty) function
- * pointer casts there. */
-static void block_job_cancel_err(Job *job, Error **errp)
-{
- BlockJob *bjob = container_of(job, BlockJob, job);
- assert(is_block_job(job));
- block_job_cancel(bjob, false);
-}
-
-int block_job_cancel_sync(BlockJob *job)
-{
- return job_finish_sync(&job->job, &block_job_cancel_err, NULL);
-}
-
-void block_job_cancel_sync_all(void)
-{
- BlockJob *job;
- AioContext *aio_context;
-
- while ((job = block_job_next(NULL))) {
- aio_context = blk_get_aio_context(job->blk);
- aio_context_acquire(aio_context);
- block_job_cancel_sync(job);
- aio_context_release(aio_context);
- }
-}
-
-int block_job_complete_sync(BlockJob *job, Error **errp)
-{
- return job_finish_sync(&job->job, job_complete, errp);
-}
-
void block_job_progress_update(BlockJob *job, uint64_t done)
{
job->offset += done;
@@ -488,25 +431,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
return job;
}
-void block_job_completed(BlockJob *job, int ret)
-{
- assert(job && job->job.txn && !job_is_completed(&job->job));
- assert(blk_bs(job->blk)->job == job);
- job->job.ret = ret;
- job_update_rc(&job->job);
- trace_block_job_completed(job, ret, job->job.ret);
- if (job->job.ret) {
- job_completed_txn_abort(&job->job);
- } else {
- job_completed_txn_success(&job->job);
- }
-}
-
-void block_job_enter(BlockJob *job)
-{
- job_enter_cond(&job->job, NULL);
-}
-
void block_job_yield(BlockJob *job)
{
assert(job->job.busy);
diff --git a/job.c b/job.c
index ee0bf42c5d..26f70bcab7 100644
--- a/job.c
+++ b/job.c
@@ -221,7 +221,7 @@ bool job_is_completed(Job *job)
return false;
}
-bool job_started(Job *job)
+static bool job_started(Job *job)
{
return job->co;
}
@@ -391,10 +391,10 @@ void job_enter(Job *job)
}
/* Yield, and schedule a timer to reenter the coroutine after @ns nanoseconds.
- * Reentering the job coroutine with block_job_enter() before the timer has
- * expired is allowed and cancels the timer.
+ * Reentering the job coroutine with job_enter() before the timer has expired
+ * is allowed and cancels the timer.
*
- * If @ns is (uint64_t) -1, no timer is scheduled and block_job_enter() must be
+ * If @ns is (uint64_t) -1, no timer is scheduled and job_enter() must be
* called explicitly. */
void coroutine_fn job_do_yield(Job *job, uint64_t ns)
{
@@ -579,7 +579,7 @@ static void job_conclude(Job *job)
}
}
-void job_update_rc(Job *job)
+static void job_update_rc(Job *job)
{
if (!job->ret && job_is_cancelled(job)) {
job->ret = -ECANCELED;
@@ -644,7 +644,7 @@ static int job_finalize_single(Job *job)
return 0;
}
-void job_cancel_async(Job *job, bool force)
+static void job_cancel_async(Job *job, bool force)
{
if (job->user_paused) {
/* Do not call job_enter here, the caller will handle it. */
@@ -660,7 +660,7 @@ void job_cancel_async(Job *job, bool force)
job->force_cancel |= force;
}
-void job_completed_txn_abort(Job *job)
+static void job_completed_txn_abort(Job *job)
{
AioContext *ctx;
JobTxn *txn = job->txn;
@@ -748,7 +748,7 @@ static int job_transition_to_pending(Job *job)
return 0;
}
-void job_completed_txn_success(Job *job)
+static void job_completed_txn_success(Job *job)
{
JobTxn *txn = job->txn;
Job *other_job;
@@ -774,6 +774,74 @@ void job_completed_txn_success(Job *job)
}
}
+void job_completed(Job *job, int ret)
+{
+ assert(job && job->txn && !job_is_completed(job));
+ job->ret = ret;
+ job_update_rc(job);
+ trace_job_completed(job, ret, job->ret);
+ if (job->ret) {
+ job_completed_txn_abort(job);
+ } else {
+ job_completed_txn_success(job);
+ }
+}
+
+void job_cancel(Job *job, bool force)
+{
+ if (job->status == JOB_STATUS_CONCLUDED) {
+ job_do_dismiss(job);
+ return;
+ }
+ job_cancel_async(job, force);
+ if (!job_started(job)) {
+ job_completed(job, -ECANCELED);
+ } else if (job->deferred_to_main_loop) {
+ job_completed_txn_abort(job);
+ } else {
+ job_enter(job);
+ }
+}
+
+void job_user_cancel(Job *job, bool force, Error **errp)
+{
+ if (job_apply_verb(job, JOB_VERB_CANCEL, errp)) {
+ return;
+ }
+ job_cancel(job, force);
+}
+
+/* A wrapper around job_cancel() taking an Error ** parameter so it may be
+ * used with job_finish_sync() without the need for (rather nasty) function
+ * pointer casts there. */
+static void job_cancel_err(Job *job, Error **errp)
+{
+ job_cancel(job, false);
+}
+
+int job_cancel_sync(Job *job)
+{
+ return job_finish_sync(job, &job_cancel_err, NULL);
+}
+
+void job_cancel_sync_all(void)
+{
+ Job *job;
+ AioContext *aio_context;
+
+ while ((job = job_next(NULL))) {
+ aio_context = job->aio_context;
+ aio_context_acquire(aio_context);
+ job_cancel_sync(job);
+ aio_context_release(aio_context);
+ }
+}
+
+int job_complete_sync(Job *job, Error **errp)
+{
+ return job_finish_sync(job, job_complete, errp);
+}
+
void job_complete(Job *job, Error **errp)
{
/* Should not be reachable via external interface for internal jobs */
diff --git a/qemu-img.c b/qemu-img.c
index 81f338273f..b2bcfc3df1 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -870,7 +870,7 @@ static void run_block_job(BlockJob *job, Error **errp)
} while (!job->ready && !job_is_completed(&job->job));
if (!job_is_completed(&job->job)) {
- ret = block_job_complete_sync(job, errp);
+ ret = job_complete_sync(&job->job, errp);
} else {
ret = job->job.ret;
}
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index b428aaca68..3600ffd3fb 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -498,8 +498,7 @@ typedef struct TestBlockJob {
static void test_job_completed(Job *job, void *opaque)
{
- BlockJob *bjob = container_of(job, BlockJob, job);
- block_job_completed(bjob, 0);
+ job_completed(job, 0);
}
static void coroutine_fn test_job_start(void *opaque)
@@ -593,7 +592,7 @@ static void test_blockjob_common(enum drain_type drain_type)
g_assert_false(job->job.paused);
g_assert_false(job->job.busy); /* We're in job_sleep_ns() */
- ret = block_job_complete_sync(job, &error_abort);
+ ret = job_complete_sync(&job->job, &error_abort);
g_assert_cmpint(ret, ==, 0);
blk_unref(blk_src);
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 6ee31d59ad..34ee179574 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -34,7 +34,7 @@ static void test_block_job_complete(Job *job, void *opaque)
rc = -ECANCELED;
}
- block_job_completed(bjob, rc);
+ job_completed(job, rc);
bdrv_unref(bs);
}
@@ -130,7 +130,7 @@ static void test_single_job(int expected)
job_start(&job->job);
if (expected == -ECANCELED) {
- block_job_cancel(job, false);
+ job_cancel(&job->job, false);
}
while (result == -EINPROGRESS) {
@@ -176,10 +176,10 @@ static void test_pair_jobs(int expected1, int expected2)
job_txn_unref(txn);
if (expected1 == -ECANCELED) {
- block_job_cancel(job1, false);
+ job_cancel(&job1->job, false);
}
if (expected2 == -ECANCELED) {
- block_job_cancel(job2, false);
+ job_cancel(&job2->job, false);
}
while (result1 == -EINPROGRESS || result2 == -EINPROGRESS) {
@@ -232,13 +232,13 @@ static void test_pair_jobs_fail_cancel_race(void)
job_start(&job1->job);
job_start(&job2->job);
- block_job_cancel(job1, false);
+ job_cancel(&job1->job, false);
/* Now make job2 finish before the main loop kicks jobs. This simulates
* the race between a pending kick and another job completing.
*/
- block_job_enter(job2);
- block_job_enter(job2);
+ job_enter(&job2->job);
+ job_enter(&job2->job);
while (result1 == -EINPROGRESS || result2 == -EINPROGRESS) {
aio_poll(qemu_get_aio_context(), true);
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 1e052c2e9c..46a78739aa 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -165,10 +165,9 @@ typedef struct CancelJob {
static void cancel_job_completed(Job *job, void *opaque)
{
- BlockJob *bjob = container_of(job, BlockJob, job);
CancelJob *s = opaque;
s->completed = true;
- block_job_completed(bjob, 0);
+ job_completed(job, 0);
}
static void cancel_job_complete(Job *job, Error **errp)
@@ -232,7 +231,7 @@ static void cancel_common(CancelJob *s)
BlockBackend *blk = s->blk;
JobStatus sts = job->job.status;
- block_job_cancel_sync(job);
+ job_cancel_sync(&job->job);
if (sts != JOB_STATUS_CREATED && sts != JOB_STATUS_CONCLUDED) {
BlockJob *dummy = job;
block_job_dismiss(&dummy, &error_abort);
@@ -275,7 +274,7 @@ static void test_cancel_paused(void)
assert(job->job.status == JOB_STATUS_RUNNING);
job_user_pause(&job->job, &error_abort);
- block_job_enter(job);
+ job_enter(&job->job);
assert(job->job.status == JOB_STATUS_PAUSED);
cancel_common(s);
@@ -292,7 +291,7 @@ static void test_cancel_ready(void)
assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
- block_job_enter(job);
+ job_enter(&job->job);
assert(job->job.status == JOB_STATUS_READY);
cancel_common(s);
@@ -309,11 +308,11 @@ static void test_cancel_standby(void)
assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
- block_job_enter(job);
+ job_enter(&job->job);
assert(job->job.status == JOB_STATUS_READY);
job_user_pause(&job->job, &error_abort);
- block_job_enter(job);
+ job_enter(&job->job);
assert(job->job.status == JOB_STATUS_STANDBY);
cancel_common(s);
@@ -330,11 +329,11 @@ static void test_cancel_pending(void)
assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
- block_job_enter(job);
+ job_enter(&job->job);
assert(job->job.status == JOB_STATUS_READY);
job_complete(&job->job, &error_abort);
- block_job_enter(job);
+ job_enter(&job->job);
while (!s->completed) {
aio_poll(qemu_get_aio_context(), true);
}
@@ -354,11 +353,11 @@ static void test_cancel_concluded(void)
assert(job->job.status == JOB_STATUS_RUNNING);
s->should_converge = true;
- block_job_enter(job);
+ job_enter(&job->job);
assert(job->job.status == JOB_STATUS_READY);
job_complete(&job->job, &error_abort);
- block_job_enter(job);
+ job_enter(&job->job);
while (!s->completed) {
aio_poll(qemu_get_aio_context(), true);
}
diff --git a/block/trace-events b/block/trace-events
index 93b927908a..2d59b53fd3 100644
--- a/block/trace-events
+++ b/block/trace-events
@@ -4,9 +4,6 @@
bdrv_open_common(void *bs, const char *filename, int flags, const char *format_name) "bs %p filename \"%s\" flags 0x%x format_name \"%s\""
bdrv_lock_medium(void *bs, bool locked) "bs %p locked %d"
-# blockjob.c
-block_job_completed(void *job, int ret, int jret) "job %p ret %d corrected ret %d"
-
# block/block-backend.c
blk_co_preadv(void *blk, void *bs, int64_t offset, unsigned int bytes, int flags) "blk %p bs %p offset %"PRId64" bytes %u flags 0x%x"
blk_co_pwritev(void *blk, void *bs, int64_t offset, unsigned int bytes, int flags) "blk %p bs %p offset %"PRId64" bytes %u flags 0x%x"
diff --git a/trace-events b/trace-events
index 2507e1327d..ef7579a285 100644
--- a/trace-events
+++ b/trace-events
@@ -107,6 +107,7 @@ gdbstub_err_checksum_incorrect(uint8_t expected, uint8_t got) "got command packe
# job.c
job_state_transition(void *job, int ret, const char *legal, const char *s0, const char *s1) "job %p (ret: %d) attempting %s transition (%s-->%s)"
job_apply_verb(void *job, const char *state, const char *verb, const char *legal) "job %p in state %s; applying verb %s (%s)"
+job_completed(void *job, int ret, int jret) "job %p ret %d corrected ret %d"
### Guest events, keep at bottom
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 28/40] block: Cancel job in bdrv_close_all() callers
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (26 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 27/40] job: Move completion and cancellation " Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 29/40] job: Add job_yield() Kevin Wolf
` (14 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
Now that we cancel all jobs and not only block jobs on shutdown, doing
that in bdrv_close_all() isn't really appropriate any more. Move the
job_cancel_sync_all() call to the callers, and only assert that there
are no job running in bdrv_close_all().
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
block.c | 4 +---
qemu-nbd.c | 8 +++++++-
vl.c | 1 +
3 files changed, 9 insertions(+), 4 deletions(-)
diff --git a/block.c b/block.c
index d50b08c29b..501b64c819 100644
--- a/block.c
+++ b/block.c
@@ -3362,9 +3362,7 @@ static void bdrv_close(BlockDriverState *bs)
void bdrv_close_all(void)
{
- /* TODO We do want to cancel all jobs instead of just block jobs on
- * shutdown, but bdrv_close_all() isn't the right place any more. */
- job_cancel_sync_all();
+ assert(job_next(NULL) == NULL);
nbd_export_close_all();
/* Drop references from requests still in flight, such as canceled block
diff --git a/qemu-nbd.c b/qemu-nbd.c
index 0af0560ad1..51b9d38c72 100644
--- a/qemu-nbd.c
+++ b/qemu-nbd.c
@@ -482,6 +482,12 @@ static const char *socket_activation_validate_opts(const char *device,
return NULL;
}
+static void qemu_nbd_shutdown(void)
+{
+ job_cancel_sync_all();
+ bdrv_close_all();
+}
+
int main(int argc, char **argv)
{
BlockBackend *blk;
@@ -928,7 +934,7 @@ int main(int argc, char **argv)
exit(EXIT_FAILURE);
}
bdrv_init();
- atexit(bdrv_close_all);
+ atexit(qemu_nbd_shutdown);
srcpath = argv[optind];
if (imageOpts) {
diff --git a/vl.c b/vl.c
index 3b39bbd7a8..d8c53ab9fd 100644
--- a/vl.c
+++ b/vl.c
@@ -4679,6 +4679,7 @@ int main(int argc, char **argv, char **envp)
/* No more vcpu or device emulation activity beyond this point */
vm_shutdown();
+ job_cancel_sync_all();
bdrv_close_all();
res_free();
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 29/40] job: Add job_yield()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (27 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 28/40] block: Cancel job in bdrv_close_all() callers Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 30/40] job: Add job_dismiss() Kevin Wolf
` (13 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves block_job_yield() to the Job layer.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob_int.h | 8 --------
include/qemu/job.h | 9 +++++++--
block/backup.c | 2 +-
block/mirror.c | 2 +-
blockjob.c | 16 ----------------
job.c | 20 ++++++++++++++++++--
tests/test-blockjob-txn.c | 2 +-
7 files changed, 28 insertions(+), 31 deletions(-)
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 7df07b20cf..806ac64d87 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -108,14 +108,6 @@ void block_job_user_resume(Job *job);
void block_job_drain(Job *job);
/**
- * block_job_yield:
- * @job: The job that calls the function.
- *
- * Yield the block job coroutine.
- */
-void block_job_yield(BlockJob *job);
-
-/**
* block_job_ratelimit_get_delay:
*
* Calculate and return delay for the next request in ns. See the documentation
diff --git a/include/qemu/job.h b/include/qemu/job.h
index bbe1b0cd1a..94900ec008 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -339,6 +339,13 @@ void coroutine_fn job_pause_point(Job *job);
/**
* @job: The job that calls the function.
+ *
+ * Yield the job coroutine.
+ */
+void job_yield(Job *job);
+
+/**
+ * @job: The job that calls the function.
* @ns: How many nanoseconds to stop for.
*
* Put the job to sleep (assuming that it wasn't canceled) for @ns
@@ -508,8 +515,6 @@ int job_finish_sync(Job *job, void (*finish)(Job *, Error **errp), Error **errp)
/* TODO To be removed from the public interface */
void job_state_transition(Job *job, JobStatus s1);
-void coroutine_fn job_do_yield(Job *job, uint64_t ns);
-bool job_should_pause(Job *job);
void job_do_dismiss(Job *job);
#endif
diff --git a/block/backup.c b/block/backup.c
index b13f91d8a7..6f4f3df229 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -444,7 +444,7 @@ static void coroutine_fn backup_run(void *opaque)
while (!job_is_cancelled(&job->common.job)) {
/* Yield until the job is cancelled. We just let our before_write
* notify callback service CoW requests. */
- block_job_yield(&job->common);
+ job_yield(&job->common.job);
}
} else if (job->sync_mode == MIRROR_SYNC_MODE_INCREMENTAL) {
ret = backup_run_incremental(job);
diff --git a/block/mirror.c b/block/mirror.c
index c63cf7cbc9..687f955c22 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -731,7 +731,7 @@ static void coroutine_fn mirror_run(void *opaque)
block_job_event_ready(&s->common);
s->synced = true;
while (!job_is_cancelled(&s->common.job) && !s->should_complete) {
- block_job_yield(&s->common);
+ job_yield(&s->common.job);
}
s->common.job.cancelled = false;
goto immediate_exit;
diff --git a/blockjob.c b/blockjob.c
index 438baa1778..f146fe0cbd 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -431,22 +431,6 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
return job;
}
-void block_job_yield(BlockJob *job)
-{
- assert(job->job.busy);
-
- /* Check cancellation *before* setting busy = false, too! */
- if (job_is_cancelled(&job->job)) {
- return;
- }
-
- if (!job_should_pause(&job->job)) {
- job_do_yield(&job->job, -1);
- }
-
- job_pause_point(&job->job);
-}
-
void block_job_iostatus_reset(BlockJob *job)
{
if (job->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
diff --git a/job.c b/job.c
index 26f70bcab7..d7cdcd2456 100644
--- a/job.c
+++ b/job.c
@@ -226,7 +226,7 @@ static bool job_started(Job *job)
return job->co;
}
-bool job_should_pause(Job *job)
+static bool job_should_pause(Job *job)
{
return job->pause_count > 0;
}
@@ -396,7 +396,7 @@ void job_enter(Job *job)
*
* If @ns is (uint64_t) -1, no timer is scheduled and job_enter() must be
* called explicitly. */
-void coroutine_fn job_do_yield(Job *job, uint64_t ns)
+static void coroutine_fn job_do_yield(Job *job, uint64_t ns)
{
job_lock();
if (ns != -1) {
@@ -441,6 +441,22 @@ void coroutine_fn job_pause_point(Job *job)
}
}
+void job_yield(Job *job)
+{
+ assert(job->busy);
+
+ /* Check cancellation *before* setting busy = false, too! */
+ if (job_is_cancelled(job)) {
+ return;
+ }
+
+ if (!job_should_pause(job)) {
+ job_do_yield(job, -1);
+ }
+
+ job_pause_point(job);
+}
+
void coroutine_fn job_sleep_ns(Job *job, int64_t ns)
{
assert(job->busy);
diff --git a/tests/test-blockjob-txn.c b/tests/test-blockjob-txn.c
index 34ee179574..fce836639a 100644
--- a/tests/test-blockjob-txn.c
+++ b/tests/test-blockjob-txn.c
@@ -47,7 +47,7 @@ static void coroutine_fn test_block_job_run(void *opaque)
if (s->use_timer) {
job_sleep_ns(&job->job, 0);
} else {
- block_job_yield(job);
+ job_yield(&job->job);
}
if (job_is_cancelled(&job->job)) {
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 30/40] job: Add job_dismiss()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (28 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 29/40] job: Add job_yield() Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 31/40] job: Add job_is_ready() Kevin Wolf
` (12 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This moves block_job_dismiss() to the Job layer.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 9 ---------
include/qemu/job.h | 7 ++++++-
blockdev.c | 10 ++++++----
blockjob.c | 13 -------------
job.c | 15 ++++++++++++++-
tests/test-blockjob.c | 4 ++--
6 files changed, 28 insertions(+), 30 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index e9ed7b8b78..5a81afc6c3 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -141,15 +141,6 @@ void block_job_remove_all_bdrv(BlockJob *job);
void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp);
/**
- * block_job_dismiss:
- * @job: The job to be dismissed.
- * @errp: Error object.
- *
- * Remove a concluded job from the query list.
- */
-void block_job_dismiss(BlockJob **job, Error **errp);
-
-/**
* block_job_progress_update:
* @job: The job that has made progress
* @done: How much progress the job made
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 94900ec008..1e8050c6fb 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -487,6 +487,12 @@ int job_complete_sync(Job *job, Error **errp);
*/
void job_finalize(Job *job, Error **errp);
+/**
+ * Remove the concluded @job from the query list and resets the passed pointer
+ * to %NULL. Returns an error if the job is not actually concluded.
+ */
+void job_dismiss(Job **job, Error **errp);
+
typedef void JobDeferToMainLoopFn(Job *job, void *opaque);
/**
@@ -515,6 +521,5 @@ int job_finish_sync(Job *job, void (*finish)(Job *, Error **errp), Error **errp)
/* TODO To be removed from the public interface */
void job_state_transition(Job *job, JobStatus s1);
-void job_do_dismiss(Job *job);
#endif
diff --git a/blockdev.c b/blockdev.c
index 31319a6d5a..8de95be8f4 100644
--- a/blockdev.c
+++ b/blockdev.c
@@ -3915,14 +3915,16 @@ void qmp_block_job_finalize(const char *id, Error **errp)
void qmp_block_job_dismiss(const char *id, Error **errp)
{
AioContext *aio_context;
- BlockJob *job = find_block_job(id, &aio_context, errp);
+ BlockJob *bjob = find_block_job(id, &aio_context, errp);
+ Job *job;
- if (!job) {
+ if (!bjob) {
return;
}
- trace_qmp_block_job_dismiss(job);
- block_job_dismiss(&job, errp);
+ trace_qmp_block_job_dismiss(bjob);
+ job = &bjob->job;
+ job_dismiss(&job, errp);
aio_context_release(aio_context);
}
diff --git a/blockjob.c b/blockjob.c
index f146fe0cbd..3ca009bea5 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -242,19 +242,6 @@ int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n)
return ratelimit_calculate_delay(&job->limit, n);
}
-void block_job_dismiss(BlockJob **jobptr, Error **errp)
-{
- BlockJob *job = *jobptr;
- /* similarly to _complete, this is QMP-interface only. */
- assert(job->job.id);
- if (job_apply_verb(&job->job, JOB_VERB_DISMISS, errp)) {
- return;
- }
-
- job_do_dismiss(&job->job);
- *jobptr = NULL;
-}
-
void block_job_progress_update(BlockJob *job, uint64_t done)
{
job->offset += done;
diff --git a/job.c b/job.c
index d7cdcd2456..af31de4669 100644
--- a/job.c
+++ b/job.c
@@ -568,7 +568,7 @@ void job_user_resume(Job *job, Error **errp)
job_resume(job);
}
-void job_do_dismiss(Job *job)
+static void job_do_dismiss(Job *job)
{
assert(job);
job->busy = false;
@@ -581,6 +581,19 @@ void job_do_dismiss(Job *job)
job_unref(job);
}
+void job_dismiss(Job **jobptr, Error **errp)
+{
+ Job *job = *jobptr;
+ /* similarly to _complete, this is QMP-interface only. */
+ assert(job->id);
+ if (job_apply_verb(job, JOB_VERB_DISMISS, errp)) {
+ return;
+ }
+
+ job_do_dismiss(job);
+ *jobptr = NULL;
+}
+
void job_early_fail(Job *job)
{
assert(job->status == JOB_STATUS_CREATED);
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 46a78739aa..7131cabb16 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -233,8 +233,8 @@ static void cancel_common(CancelJob *s)
job_cancel_sync(&job->job);
if (sts != JOB_STATUS_CREATED && sts != JOB_STATUS_CONCLUDED) {
- BlockJob *dummy = job;
- block_job_dismiss(&dummy, &error_abort);
+ Job *dummy = &job->job;
+ job_dismiss(&dummy, &error_abort);
}
assert(job->job.status == JOB_STATUS_NULL);
job_unref(&job->job);
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 31/40] job: Add job_is_ready()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (29 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 30/40] job: Add job_dismiss() Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-23 23:42 ` John Snow
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 32/40] job: Add job_transition_to_ready() Kevin Wolf
` (11 subsequent siblings)
42 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
Instead of having a 'bool ready' in BlockJob, add a function that
derives its value from the job status.
At the same time, this fixes the behaviour to match what the QAPI
documentation promises for query-block-job: 'true if the job may be
completed'. When the ready flag was introduced in commit ef6dbf1e46e,
the flag never had to be reset to match the description because after
being ready, the jobs would immediately complete and disappear.
Job transactions and manual job finalisation were introduced only later.
With these changes, jobs may stay around even after having completed
(and they are not ready to be completed a second time), however their
patches forgot to reset the ready flag.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 5 -----
include/qemu/job.h | 3 +++
blockjob.c | 3 +--
job.c | 22 ++++++++++++++++++++++
qemu-img.c | 2 +-
tests/test-blockjob.c | 2 +-
6 files changed, 28 insertions(+), 9 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 5a81afc6c3..8e1e1ee0de 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -49,11 +49,6 @@ typedef struct BlockJob {
/** The block device on which the job is operating. */
BlockBackend *blk;
- /**
- * Set to true when the job is ready to be completed.
- */
- bool ready;
-
/** Status that is published by the query-block-jobs QMP API */
BlockDeviceIoStatus iostatus;
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 1e8050c6fb..487f9d9a32 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -367,6 +367,9 @@ bool job_is_cancelled(Job *job);
/** Returns whether the job is in a completed state. */
bool job_is_completed(Job *job);
+/** Returns whether the job is ready to be completed. */
+bool job_is_ready(Job *job);
+
/**
* Request @job to pause at the next pause point. Must be paired with
* job_resume(). If the job is supposed to be resumed by user action, call
diff --git a/blockjob.c b/blockjob.c
index 3ca009bea5..38f18e9ba3 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -269,7 +269,7 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
info->offset = job->offset;
info->speed = job->speed;
info->io_status = job->iostatus;
- info->ready = job->ready;
+ info->ready = job_is_ready(&job->job),
info->status = job->job.status;
info->auto_finalize = job->job.auto_finalize;
info->auto_dismiss = job->job.auto_dismiss;
@@ -436,7 +436,6 @@ void block_job_user_resume(Job *job)
void block_job_event_ready(BlockJob *job)
{
job_state_transition(&job->job, JOB_STATUS_READY);
- job->ready = true;
if (block_job_is_internal(job)) {
return;
diff --git a/job.c b/job.c
index af31de4669..66ee26f2a0 100644
--- a/job.c
+++ b/job.c
@@ -199,6 +199,28 @@ bool job_is_cancelled(Job *job)
return job->cancelled;
}
+bool job_is_ready(Job *job)
+{
+ switch (job->status) {
+ case JOB_STATUS_UNDEFINED:
+ case JOB_STATUS_CREATED:
+ case JOB_STATUS_RUNNING:
+ case JOB_STATUS_PAUSED:
+ case JOB_STATUS_WAITING:
+ case JOB_STATUS_PENDING:
+ case JOB_STATUS_ABORTING:
+ case JOB_STATUS_CONCLUDED:
+ case JOB_STATUS_NULL:
+ return false;
+ case JOB_STATUS_READY:
+ case JOB_STATUS_STANDBY:
+ return true;
+ default:
+ g_assert_not_reached();
+ }
+ return false;
+}
+
bool job_is_completed(Job *job)
{
switch (job->status) {
diff --git a/qemu-img.c b/qemu-img.c
index b2bcfc3df1..eecb7d3228 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -867,7 +867,7 @@ static void run_block_job(BlockJob *job, Error **errp)
aio_poll(aio_context, true);
qemu_progress_print(job->len ?
((float)job->offset / job->len * 100.f) : 0.0f, 0);
- } while (!job->ready && !job_is_completed(&job->job));
+ } while (!job_is_ready(&job->job) && !job_is_completed(&job->job));
if (!job_is_completed(&job->job)) {
ret = job_complete_sync(&job->job, errp);
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 7131cabb16..8180d03a5f 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -185,7 +185,7 @@ static void coroutine_fn cancel_job_start(void *opaque)
goto defer;
}
- if (!s->common.ready && s->should_converge) {
+ if (!job_is_ready(&s->common.job) && s->should_converge) {
block_job_event_ready(&s->common);
}
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 31/40] job: Add job_is_ready()
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 31/40] job: Add job_is_ready() Kevin Wolf
@ 2018-05-23 23:42 ` John Snow
2018-05-24 8:30 ` Kevin Wolf
0 siblings, 1 reply; 86+ messages in thread
From: John Snow @ 2018-05-23 23:42 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: qemu-devel, jcody, armbru, mreitz
On 05/18/2018 09:21 AM, Kevin Wolf wrote:
> Instead of having a 'bool ready' in BlockJob, add a function that
> derives its value from the job status.
>
> At the same time, this fixes the behaviour to match what the QAPI
> documentation promises for query-block-job: 'true if the job may be
> completed'. When the ready flag was introduced in commit ef6dbf1e46e,
> the flag never had to be reset to match the description because after
> being ready, the jobs would immediately complete and disappear.
>
> Job transactions and manual job finalisation were introduced only later.
> With these changes, jobs may stay around even after having completed
> (and they are not ready to be completed a second time), however their
> patches forgot to reset the ready flag.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> Reviewed-by: Max Reitz <mreitz@redhat.com>
> ---
> include/block/blockjob.h | 5 -----
> include/qemu/job.h | 3 +++
> blockjob.c | 3 +--
> job.c | 22 ++++++++++++++++++++++
> qemu-img.c | 2 +-
> tests/test-blockjob.c | 2 +-
> 6 files changed, 28 insertions(+), 9 deletions(-)
>
> diff --git a/include/block/blockjob.h b/include/block/blockjob.h
> index 5a81afc6c3..8e1e1ee0de 100644
> --- a/include/block/blockjob.h
> +++ b/include/block/blockjob.h
> @@ -49,11 +49,6 @@ typedef struct BlockJob {
> /** The block device on which the job is operating. */
> BlockBackend *blk;
>
> - /**
> - * Set to true when the job is ready to be completed.
> - */
> - bool ready;
> -
> /** Status that is published by the query-block-jobs QMP API */
> BlockDeviceIoStatus iostatus;
>
> diff --git a/include/qemu/job.h b/include/qemu/job.h
> index 1e8050c6fb..487f9d9a32 100644
> --- a/include/qemu/job.h
> +++ b/include/qemu/job.h
> @@ -367,6 +367,9 @@ bool job_is_cancelled(Job *job);
> /** Returns whether the job is in a completed state. */
> bool job_is_completed(Job *job);
>
> +/** Returns whether the job is ready to be completed. */
> +bool job_is_ready(Job *job);
> +
> /**
> * Request @job to pause at the next pause point. Must be paired with
> * job_resume(). If the job is supposed to be resumed by user action, call
> diff --git a/blockjob.c b/blockjob.c
> index 3ca009bea5..38f18e9ba3 100644
> --- a/blockjob.c
> +++ b/blockjob.c
> @@ -269,7 +269,7 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
> info->offset = job->offset;
> info->speed = job->speed;
> info->io_status = job->iostatus;
> - info->ready = job->ready;
> + info->ready = job_is_ready(&job->job),
> info->status = job->job.status;
> info->auto_finalize = job->job.auto_finalize;
> info->auto_dismiss = job->job.auto_dismiss;
> @@ -436,7 +436,6 @@ void block_job_user_resume(Job *job)
> void block_job_event_ready(BlockJob *job)
> {
> job_state_transition(&job->job, JOB_STATUS_READY);
> - job->ready = true;
>
> if (block_job_is_internal(job)) {
> return;
> diff --git a/job.c b/job.c
> index af31de4669..66ee26f2a0 100644
> --- a/job.c
> +++ b/job.c
> @@ -199,6 +199,28 @@ bool job_is_cancelled(Job *job)
> return job->cancelled;
> }
>
> +bool job_is_ready(Job *job)
> +{
> + switch (job->status) {
> + case JOB_STATUS_UNDEFINED:
> + case JOB_STATUS_CREATED:
> + case JOB_STATUS_RUNNING:
> + case JOB_STATUS_PAUSED:
> + case JOB_STATUS_WAITING:
> + case JOB_STATUS_PENDING:
> + case JOB_STATUS_ABORTING:
> + case JOB_STATUS_CONCLUDED:
> + case JOB_STATUS_NULL:
> + return false;
> + case JOB_STATUS_READY:
> + case JOB_STATUS_STANDBY:
> + return true;
> + default:
> + g_assert_not_reached();
> + }
> + return false;
> +}
> +
What's the benefit to a switch statement with a default clause here,
over the shorter:
if (job->status == READY || job->status == STANDBY) {
return true;
}
return false;
(Yes, I realize you already merged this code, but I'm still curious and
I need to read the series anyway to see what's changed...)
> bool job_is_completed(Job *job)
> {
> switch (job->status) {
> diff --git a/qemu-img.c b/qemu-img.c
> index b2bcfc3df1..eecb7d3228 100644
> --- a/qemu-img.c
> +++ b/qemu-img.c
> @@ -867,7 +867,7 @@ static void run_block_job(BlockJob *job, Error **errp)
> aio_poll(aio_context, true);
> qemu_progress_print(job->len ?
> ((float)job->offset / job->len * 100.f) : 0.0f, 0);
> - } while (!job->ready && !job_is_completed(&job->job));
> + } while (!job_is_ready(&job->job) && !job_is_completed(&job->job));
>
Semantic change in this variable, but due to the way this function is
written it doesn't matter.
> if (!job_is_completed(&job->job)) {
> ret = job_complete_sync(&job->job, errp);
> diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
> index 7131cabb16..8180d03a5f 100644
> --- a/tests/test-blockjob.c
> +++ b/tests/test-blockjob.c
> @@ -185,7 +185,7 @@ static void coroutine_fn cancel_job_start(void *opaque)
> goto defer;
> }
>
> - if (!s->common.ready && s->should_converge) {
> + if (!job_is_ready(&s->common.job) && s->should_converge) {
> block_job_event_ready(&s->common);
> }
>
>
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 31/40] job: Add job_is_ready()
2018-05-23 23:42 ` John Snow
@ 2018-05-24 8:30 ` Kevin Wolf
2018-05-24 17:25 ` John Snow
0 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-24 8:30 UTC (permalink / raw)
To: John Snow; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
Am 24.05.2018 um 01:42 hat John Snow geschrieben:
> On 05/18/2018 09:21 AM, Kevin Wolf wrote:
> > Instead of having a 'bool ready' in BlockJob, add a function that
> > derives its value from the job status.
> >
> > At the same time, this fixes the behaviour to match what the QAPI
> > documentation promises for query-block-job: 'true if the job may be
> > completed'. When the ready flag was introduced in commit ef6dbf1e46e,
> > the flag never had to be reset to match the description because after
> > being ready, the jobs would immediately complete and disappear.
> >
> > Job transactions and manual job finalisation were introduced only later.
> > With these changes, jobs may stay around even after having completed
> > (and they are not ready to be completed a second time), however their
> > patches forgot to reset the ready flag.
> >
> > Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> > Reviewed-by: Max Reitz <mreitz@redhat.com>
> > ---
> > include/block/blockjob.h | 5 -----
> > include/qemu/job.h | 3 +++
> > blockjob.c | 3 +--
> > job.c | 22 ++++++++++++++++++++++
> > qemu-img.c | 2 +-
> > tests/test-blockjob.c | 2 +-
> > 6 files changed, 28 insertions(+), 9 deletions(-)
> >
> > diff --git a/include/block/blockjob.h b/include/block/blockjob.h
> > index 5a81afc6c3..8e1e1ee0de 100644
> > --- a/include/block/blockjob.h
> > +++ b/include/block/blockjob.h
> > @@ -49,11 +49,6 @@ typedef struct BlockJob {
> > /** The block device on which the job is operating. */
> > BlockBackend *blk;
> >
> > - /**
> > - * Set to true when the job is ready to be completed.
> > - */
> > - bool ready;
> > -
> > /** Status that is published by the query-block-jobs QMP API */
> > BlockDeviceIoStatus iostatus;
> >
> > diff --git a/include/qemu/job.h b/include/qemu/job.h
> > index 1e8050c6fb..487f9d9a32 100644
> > --- a/include/qemu/job.h
> > +++ b/include/qemu/job.h
> > @@ -367,6 +367,9 @@ bool job_is_cancelled(Job *job);
> > /** Returns whether the job is in a completed state. */
> > bool job_is_completed(Job *job);
> >
> > +/** Returns whether the job is ready to be completed. */
> > +bool job_is_ready(Job *job);
> > +
> > /**
> > * Request @job to pause at the next pause point. Must be paired with
> > * job_resume(). If the job is supposed to be resumed by user action, call
> > diff --git a/blockjob.c b/blockjob.c
> > index 3ca009bea5..38f18e9ba3 100644
> > --- a/blockjob.c
> > +++ b/blockjob.c
> > @@ -269,7 +269,7 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
> > info->offset = job->offset;
> > info->speed = job->speed;
> > info->io_status = job->iostatus;
> > - info->ready = job->ready;
> > + info->ready = job_is_ready(&job->job),
> > info->status = job->job.status;
> > info->auto_finalize = job->job.auto_finalize;
> > info->auto_dismiss = job->job.auto_dismiss;
> > @@ -436,7 +436,6 @@ void block_job_user_resume(Job *job)
> > void block_job_event_ready(BlockJob *job)
> > {
> > job_state_transition(&job->job, JOB_STATUS_READY);
> > - job->ready = true;
> >
> > if (block_job_is_internal(job)) {
> > return;
> > diff --git a/job.c b/job.c
> > index af31de4669..66ee26f2a0 100644
> > --- a/job.c
> > +++ b/job.c
> > @@ -199,6 +199,28 @@ bool job_is_cancelled(Job *job)
> > return job->cancelled;
> > }
> >
> > +bool job_is_ready(Job *job)
> > +{
> > + switch (job->status) {
> > + case JOB_STATUS_UNDEFINED:
> > + case JOB_STATUS_CREATED:
> > + case JOB_STATUS_RUNNING:
> > + case JOB_STATUS_PAUSED:
> > + case JOB_STATUS_WAITING:
> > + case JOB_STATUS_PENDING:
> > + case JOB_STATUS_ABORTING:
> > + case JOB_STATUS_CONCLUDED:
> > + case JOB_STATUS_NULL:
> > + return false;
> > + case JOB_STATUS_READY:
> > + case JOB_STATUS_STANDBY:
> > + return true;
> > + default:
> > + g_assert_not_reached();
> > + }
> > + return false;
> > +}
> > +
>
> What's the benefit to a switch statement with a default clause here,
> over the shorter:
>
> if (job->status == READY || job->status == STANDBY) {
> return true;
> }
> return false;
>
> (Yes, I realize you already merged this code, but I'm still curious and
> I need to read the series anyway to see what's changed...)
That it's easy to copy and paste from job_is_completed()? :-P
I guess you could argue that the switch ensures that we don't forget to
explicitly handle every state if we ever add a new one, but the real
reason is more like, job_is_completed() was already there and I didn't
see a reason to do something different here.
(And actually, I seem to remember that I initially had everything after
READY return true as well, but then I noticed that you can reach those
states through ABORTED as well, which should not result in
job_is_ready() returning true.)
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 31/40] job: Add job_is_ready()
2018-05-24 8:30 ` Kevin Wolf
@ 2018-05-24 17:25 ` John Snow
2018-05-25 8:06 ` Kevin Wolf
0 siblings, 1 reply; 86+ messages in thread
From: John Snow @ 2018-05-24 17:25 UTC (permalink / raw)
To: Kevin Wolf; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
On 05/24/2018 04:30 AM, Kevin Wolf wrote:
> Am 24.05.2018 um 01:42 hat John Snow geschrieben:
>> On 05/18/2018 09:21 AM, Kevin Wolf wrote:
>>> Instead of having a 'bool ready' in BlockJob, add a function that
>>> derives its value from the job status.
>>>
>>> At the same time, this fixes the behaviour to match what the QAPI
>>> documentation promises for query-block-job: 'true if the job may be
>>> completed'. When the ready flag was introduced in commit ef6dbf1e46e,
>>> the flag never had to be reset to match the description because after
>>> being ready, the jobs would immediately complete and disappear.
>>>
>>> Job transactions and manual job finalisation were introduced only later.
>>> With these changes, jobs may stay around even after having completed
>>> (and they are not ready to be completed a second time), however their
>>> patches forgot to reset the ready flag.
>>>
>>> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
>>> Reviewed-by: Max Reitz <mreitz@redhat.com>
>>> ---
>>> include/block/blockjob.h | 5 -----
>>> include/qemu/job.h | 3 +++
>>> blockjob.c | 3 +--
>>> job.c | 22 ++++++++++++++++++++++
>>> qemu-img.c | 2 +-
>>> tests/test-blockjob.c | 2 +-
>>> 6 files changed, 28 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/include/block/blockjob.h b/include/block/blockjob.h
>>> index 5a81afc6c3..8e1e1ee0de 100644
>>> --- a/include/block/blockjob.h
>>> +++ b/include/block/blockjob.h
>>> @@ -49,11 +49,6 @@ typedef struct BlockJob {
>>> /** The block device on which the job is operating. */
>>> BlockBackend *blk;
>>>
>>> - /**
>>> - * Set to true when the job is ready to be completed.
>>> - */
>>> - bool ready;
>>> -
>>> /** Status that is published by the query-block-jobs QMP API */
>>> BlockDeviceIoStatus iostatus;
>>>
>>> diff --git a/include/qemu/job.h b/include/qemu/job.h
>>> index 1e8050c6fb..487f9d9a32 100644
>>> --- a/include/qemu/job.h
>>> +++ b/include/qemu/job.h
>>> @@ -367,6 +367,9 @@ bool job_is_cancelled(Job *job);
>>> /** Returns whether the job is in a completed state. */
>>> bool job_is_completed(Job *job);
>>>
>>> +/** Returns whether the job is ready to be completed. */
>>> +bool job_is_ready(Job *job);
>>> +
>>> /**
>>> * Request @job to pause at the next pause point. Must be paired with
>>> * job_resume(). If the job is supposed to be resumed by user action, call
>>> diff --git a/blockjob.c b/blockjob.c
>>> index 3ca009bea5..38f18e9ba3 100644
>>> --- a/blockjob.c
>>> +++ b/blockjob.c
>>> @@ -269,7 +269,7 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
>>> info->offset = job->offset;
>>> info->speed = job->speed;
>>> info->io_status = job->iostatus;
>>> - info->ready = job->ready;
>>> + info->ready = job_is_ready(&job->job),
>>> info->status = job->job.status;
>>> info->auto_finalize = job->job.auto_finalize;
>>> info->auto_dismiss = job->job.auto_dismiss;
>>> @@ -436,7 +436,6 @@ void block_job_user_resume(Job *job)
>>> void block_job_event_ready(BlockJob *job)
>>> {
>>> job_state_transition(&job->job, JOB_STATUS_READY);
>>> - job->ready = true;
>>>
>>> if (block_job_is_internal(job)) {
>>> return;
>>> diff --git a/job.c b/job.c
>>> index af31de4669..66ee26f2a0 100644
>>> --- a/job.c
>>> +++ b/job.c
>>> @@ -199,6 +199,28 @@ bool job_is_cancelled(Job *job)
>>> return job->cancelled;
>>> }
>>>
>>> +bool job_is_ready(Job *job)
>>> +{
>>> + switch (job->status) {
>>> + case JOB_STATUS_UNDEFINED:
>>> + case JOB_STATUS_CREATED:
>>> + case JOB_STATUS_RUNNING:
>>> + case JOB_STATUS_PAUSED:
>>> + case JOB_STATUS_WAITING:
>>> + case JOB_STATUS_PENDING:
>>> + case JOB_STATUS_ABORTING:
>>> + case JOB_STATUS_CONCLUDED:
>>> + case JOB_STATUS_NULL:
>>> + return false;
>>> + case JOB_STATUS_READY:
>>> + case JOB_STATUS_STANDBY:
>>> + return true;
>>> + default:
>>> + g_assert_not_reached();
>>> + }
>>> + return false;
>>> +}
>>> +
>>
>> What's the benefit to a switch statement with a default clause here,
>> over the shorter:
>>
>> if (job->status == READY || job->status == STANDBY) {
>> return true;
>> }
>> return false;
>>
>> (Yes, I realize you already merged this code, but I'm still curious and
>> I need to read the series anyway to see what's changed...)
>
> That it's easy to copy and paste from job_is_completed()? :-P
>
Haha! Sure!
> I guess you could argue that the switch ensures that we don't forget to
> explicitly handle every state if we ever add a new one, but the real
> reason is more like, job_is_completed() was already there and I didn't
> see a reason to do something different here.
>
I think the "default" case removes that benefit somewhat; it's nicer
when the compiler yelps at you for forgetting. The cases that might
cause an assertion could be harder to hit.
No need to change anything though, I was really just curious about your
style preferences.
(I often like to write switch/case statements without a default clause
hoping that the compiler will catch when I don't check all possible
enumerations, but this doesn't work well with QAPI values because of the
__MAX enumeration we like to add at the end. You can fix it by adding a
case MY_ENUM__MAX: g_assert_not_reached();, but I'm often afraid this
looks too hacky and silly to check in. I wish C had slightly nicer
enumerations as first class citizens a lot. I also wish I had a yacht
sometimes.)
> (And actually, I seem to remember that I initially had everything after
> READY return true as well, but then I noticed that you can reach those
> states through ABORTED as well, which should not result in
> job_is_ready() returning true.)
>
Makes much more sense.
Thanks!
> Kevin
>
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 31/40] job: Add job_is_ready()
2018-05-24 17:25 ` John Snow
@ 2018-05-25 8:06 ` Kevin Wolf
0 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-25 8:06 UTC (permalink / raw)
To: John Snow; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
Am 24.05.2018 um 19:25 hat John Snow geschrieben:
> >>> diff --git a/job.c b/job.c
> >>> index af31de4669..66ee26f2a0 100644
> >>> --- a/job.c
> >>> +++ b/job.c
> >>> @@ -199,6 +199,28 @@ bool job_is_cancelled(Job *job)
> >>> return job->cancelled;
> >>> }
> >>>
> >>> +bool job_is_ready(Job *job)
> >>> +{
> >>> + switch (job->status) {
> >>> + case JOB_STATUS_UNDEFINED:
> >>> + case JOB_STATUS_CREATED:
> >>> + case JOB_STATUS_RUNNING:
> >>> + case JOB_STATUS_PAUSED:
> >>> + case JOB_STATUS_WAITING:
> >>> + case JOB_STATUS_PENDING:
> >>> + case JOB_STATUS_ABORTING:
> >>> + case JOB_STATUS_CONCLUDED:
> >>> + case JOB_STATUS_NULL:
> >>> + return false;
> >>> + case JOB_STATUS_READY:
> >>> + case JOB_STATUS_STANDBY:
> >>> + return true;
> >>> + default:
> >>> + g_assert_not_reached();
> >>> + }
> >>> + return false;
> >>> +}
> >>> +
> >>
> >> What's the benefit to a switch statement with a default clause here,
> >> over the shorter:
> >>
> >> if (job->status == READY || job->status == STANDBY) {
> >> return true;
> >> }
> >> return false;
> >>
> >> (Yes, I realize you already merged this code, but I'm still curious and
> >> I need to read the series anyway to see what's changed...)
> >
> > That it's easy to copy and paste from job_is_completed()? :-P
> >
>
> Haha! Sure!
>
> > I guess you could argue that the switch ensures that we don't forget to
> > explicitly handle every state if we ever add a new one, but the real
> > reason is more like, job_is_completed() was already there and I didn't
> > see a reason to do something different here.
> >
>
> I think the "default" case removes that benefit somewhat; it's nicer
> when the compiler yelps at you for forgetting. The cases that might
> cause an assertion could be harder to hit.
Good point. I'm not sure which compilers would actually warn here, while
asserting makes sure to crash some tests, but in this context I think we
could actually have both: All enum values that are handled return inside
the switch statement. So if we move the g_assert_not_reached() just
after the switch block, we should get both the compiler warning and the
crash.
Would you like to send a patch?
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 32/40] job: Add job_transition_to_ready()
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (30 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 31/40] job: Add job_is_ready() Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 33/40] job: Move progress fields to Job Kevin Wolf
` (10 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
The transition to the READY state was still performed in the BlockJob
layer, in the same function that sent the BLOCK_JOB_READY QMP event.
This patch brings the state transition to the Job layer and implements
the QMP event using a notifier called from the Job layer, like we
already do for other events related to state transitions.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 3 +++
include/block/blockjob_int.h | 8 --------
include/qemu/job.h | 9 ++++++---
block/mirror.c | 6 +++---
blockjob.c | 33 ++++++++++++++++++---------------
job.c | 16 +++++++++++++---
tests/test-bdrv-drain.c | 2 +-
tests/test-blockjob.c | 2 +-
8 files changed, 45 insertions(+), 34 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 8e1e1ee0de..4fca45f6a1 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -76,6 +76,9 @@ typedef struct BlockJob {
/** Called when the job transitions to PENDING */
Notifier pending_notifier;
+ /** Called when the job transitions to READY */
+ Notifier ready_notifier;
+
/** BlockDriverStates that are involved in this block job */
GSList *nodes;
} BlockJob;
diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h
index 806ac64d87..5cd50c6639 100644
--- a/include/block/blockjob_int.h
+++ b/include/block/blockjob_int.h
@@ -116,14 +116,6 @@ void block_job_drain(Job *job);
int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n);
/**
- * block_job_event_ready:
- * @job: The job which is now ready to be completed.
- *
- * Send a BLOCK_JOB_READY event for the specified job.
- */
-void block_job_event_ready(BlockJob *job);
-
-/**
* block_job_error_action:
* @job: The job to signal an error for.
* @on_err: The error action setting.
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 487f9d9a32..bfc2bc5611 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -132,6 +132,9 @@ typedef struct Job {
/** Notifiers called when the job transitions to PENDING */
NotifierList on_pending;
+ /** Notifiers called when the job transitions to READY */
+ NotifierList on_ready;
+
/** Element of the list of jobs */
QLIST_ENTRY(Job) job_list;
@@ -426,6 +429,9 @@ int job_apply_verb(Job *job, JobVerb verb, Error **errp);
/** The @job could not be started, free it. */
void job_early_fail(Job *job);
+/** Moves the @job from RUNNING to READY */
+void job_transition_to_ready(Job *job);
+
/**
* @job: The job being completed.
* @ret: The status code.
@@ -522,7 +528,4 @@ void job_defer_to_main_loop(Job *job, JobDeferToMainLoopFn *fn, void *opaque);
*/
int job_finish_sync(Job *job, void (*finish)(Job *, Error **errp), Error **errp);
-/* TODO To be removed from the public interface */
-void job_state_transition(Job *job, JobStatus s1);
-
#endif
diff --git a/block/mirror.c b/block/mirror.c
index 687f955c22..bdc1b5b7b9 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -727,8 +727,8 @@ static void coroutine_fn mirror_run(void *opaque)
}
if (s->bdev_length == 0) {
- /* Report BLOCK_JOB_READY and wait for complete. */
- block_job_event_ready(&s->common);
+ /* Transition to the READY state and wait for complete. */
+ job_transition_to_ready(&s->common.job);
s->synced = true;
while (!job_is_cancelled(&s->common.job) && !s->should_complete) {
job_yield(&s->common.job);
@@ -824,7 +824,7 @@ static void coroutine_fn mirror_run(void *opaque)
* report completion. This way, block-job-cancel will leave
* the target in a consistent state.
*/
- block_job_event_ready(&s->common);
+ job_transition_to_ready(&s->common.job);
s->synced = true;
}
diff --git a/blockjob.c b/blockjob.c
index 38f18e9ba3..da11b3b763 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -338,6 +338,22 @@ static void block_job_event_pending(Notifier *n, void *opaque)
&error_abort);
}
+static void block_job_event_ready(Notifier *n, void *opaque)
+{
+ BlockJob *job = opaque;
+
+ if (block_job_is_internal(job)) {
+ return;
+ }
+
+ qapi_event_send_block_job_ready(job_type(&job->job),
+ job->job.id,
+ job->len,
+ job->offset,
+ job->speed, &error_abort);
+}
+
+
/*
* API for block job drivers and the block layer. These functions are
* declared in blockjob_int.h.
@@ -386,12 +402,14 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
job->finalize_cancelled_notifier.notify = block_job_event_cancelled;
job->finalize_completed_notifier.notify = block_job_event_completed;
job->pending_notifier.notify = block_job_event_pending;
+ job->ready_notifier.notify = block_job_event_ready;
notifier_list_add(&job->job.on_finalize_cancelled,
&job->finalize_cancelled_notifier);
notifier_list_add(&job->job.on_finalize_completed,
&job->finalize_completed_notifier);
notifier_list_add(&job->job.on_pending, &job->pending_notifier);
+ notifier_list_add(&job->job.on_ready, &job->ready_notifier);
error_setg(&job->blocker, "block device is in use by block job: %s",
job_type_str(&job->job));
@@ -433,21 +451,6 @@ void block_job_user_resume(Job *job)
block_job_iostatus_reset(bjob);
}
-void block_job_event_ready(BlockJob *job)
-{
- job_state_transition(&job->job, JOB_STATUS_READY);
-
- if (block_job_is_internal(job)) {
- return;
- }
-
- qapi_event_send_block_job_ready(job_type(&job->job),
- job->job.id,
- job->len,
- job->offset,
- job->speed, &error_abort);
-}
-
BlockErrorAction block_job_error_action(BlockJob *job, BlockdevOnError on_err,
int is_read, int error)
{
diff --git a/job.c b/job.c
index 66ee26f2a0..34304f2883 100644
--- a/job.c
+++ b/job.c
@@ -157,9 +157,7 @@ static int job_txn_apply(JobTxn *txn, int fn(Job *), bool lock)
return rc;
}
-
-/* TODO Make static once the whole state machine is in job.c */
-void job_state_transition(Job *job, JobStatus s1)
+static void job_state_transition(Job *job, JobStatus s1)
{
JobStatus s0 = job->status;
assert(s1 >= 0 && s1 <= JOB_STATUS__MAX);
@@ -321,6 +319,7 @@ void *job_create(const char *job_id, const JobDriver *driver, JobTxn *txn,
notifier_list_init(&job->on_finalize_cancelled);
notifier_list_init(&job->on_finalize_completed);
notifier_list_init(&job->on_pending);
+ notifier_list_init(&job->on_ready);
job_state_transition(job, JOB_STATUS_CREATED);
aio_timer_init(qemu_get_aio_context(), &job->sleep_timer,
@@ -380,6 +379,11 @@ static void job_event_pending(Job *job)
notifier_list_notify(&job->on_pending, job);
}
+static void job_event_ready(Job *job)
+{
+ notifier_list_notify(&job->on_ready, job);
+}
+
void job_enter_cond(Job *job, bool(*fn)(Job *job))
{
if (!job_started(job)) {
@@ -799,6 +803,12 @@ static int job_transition_to_pending(Job *job)
return 0;
}
+void job_transition_to_ready(Job *job)
+{
+ job_state_transition(job, JOB_STATUS_READY);
+ job_event_ready(job);
+}
+
static void job_completed_txn_success(Job *job)
{
JobTxn *txn = job->txn;
diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c
index 3600ffd3fb..2cba63b881 100644
--- a/tests/test-bdrv-drain.c
+++ b/tests/test-bdrv-drain.c
@@ -505,7 +505,7 @@ static void coroutine_fn test_job_start(void *opaque)
{
TestBlockJob *s = opaque;
- block_job_event_ready(&s->common);
+ job_transition_to_ready(&s->common.job);
while (!s->should_complete) {
job_sleep_ns(&s->common.job, 100000);
}
diff --git a/tests/test-blockjob.c b/tests/test-blockjob.c
index 8180d03a5f..e408d52351 100644
--- a/tests/test-blockjob.c
+++ b/tests/test-blockjob.c
@@ -186,7 +186,7 @@ static void coroutine_fn cancel_job_start(void *opaque)
}
if (!job_is_ready(&s->common.job) && s->should_converge) {
- block_job_event_ready(&s->common);
+ job_transition_to_ready(&s->common.job);
}
job_sleep_ns(&s->common.job, 100000);
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 33/40] job: Move progress fields to Job
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (31 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 32/40] job: Add job_transition_to_ready() Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 34/40] job: Introduce qapi/job.json Kevin Wolf
` (9 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
BlockJob has fields .offset and .len, which are actually misnomers today
because they are no longer tied to block device sizes, but just progress
counters. As such they make a lot of sense in generic Jobs.
This patch moves the fields to Job and renames them to .progress_current
and .progress_total to describe their function better.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
include/block/blockjob.h | 25 -------------------------
include/qemu/job.h | 28 ++++++++++++++++++++++++++++
block/backup.c | 8 ++++----
block/commit.c | 4 ++--
block/mirror.c | 4 ++--
block/stream.c | 4 ++--
blockjob.c | 26 ++++++++------------------
job.c | 10 ++++++++++
qemu-img.c | 8 ++++++--
9 files changed, 62 insertions(+), 55 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 4fca45f6a1..3021d11126 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -52,12 +52,6 @@ typedef struct BlockJob {
/** Status that is published by the query-block-jobs QMP API */
BlockDeviceIoStatus iostatus;
- /** Offset that is published by the query-block-jobs QMP API */
- int64_t offset;
-
- /** Length that is published by the query-block-jobs QMP API */
- int64_t len;
-
/** Speed that was set with @block_job_set_speed. */
int64_t speed;
@@ -139,25 +133,6 @@ void block_job_remove_all_bdrv(BlockJob *job);
void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp);
/**
- * block_job_progress_update:
- * @job: The job that has made progress
- * @done: How much progress the job made
- *
- * Updates the progress counter of the job.
- */
-void block_job_progress_update(BlockJob *job, uint64_t done);
-
-/**
- * block_job_progress_set_remaining:
- * @job: The job whose expected progress end value is set
- * @remaining: Expected end value of the progress counter of the job
- *
- * Sets the expected end value of the progress counter of a job so that a
- * completion percentage can be calculated when the progress is updated.
- */
-void block_job_progress_set_remaining(BlockJob *job, uint64_t remaining);
-
-/**
* block_job_query:
* @job: The job to get information about.
*
diff --git a/include/qemu/job.h b/include/qemu/job.h
index bfc2bc5611..92d1d249fe 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -114,6 +114,16 @@ typedef struct Job {
/** True if this job should automatically dismiss itself */
bool auto_dismiss;
+ /**
+ * Current progress. The unit is arbitrary as long as the ratio between
+ * progress_current and progress_total represents the estimated percentage
+ * of work already done.
+ */
+ int64_t progress_current;
+
+ /** Estimated progress_current value at the completion of the job */
+ int64_t progress_total;
+
/** ret code passed to job_completed. */
int ret;
@@ -304,6 +314,24 @@ void job_ref(Job *job);
*/
void job_unref(Job *job);
+/**
+ * @job: The job that has made progress
+ * @done: How much progress the job made since the last call
+ *
+ * Updates the progress counter of the job.
+ */
+void job_progress_update(Job *job, uint64_t done);
+
+/**
+ * @job: The job whose expected progress end value is set
+ * @remaining: Missing progress (on top of the current progress counter value)
+ * until the new expected end value is reached
+ *
+ * Sets the expected end value of the progress counter of a job so that a
+ * completion percentage can be calculated when the progress is updated.
+ */
+void job_progress_set_remaining(Job *job, uint64_t remaining);
+
/** To be called when a cancelled job is finalised. */
void job_event_cancelled(Job *job);
diff --git a/block/backup.c b/block/backup.c
index 6f4f3df229..4e228e959b 100644
--- a/block/backup.c
+++ b/block/backup.c
@@ -160,7 +160,7 @@ static int coroutine_fn backup_do_cow(BackupBlockJob *job,
* offset field is an opaque progress value, it is not a disk offset.
*/
job->bytes_read += n;
- block_job_progress_update(&job->common, n);
+ job_progress_update(&job->common.job, n);
}
out:
@@ -406,8 +406,8 @@ static void backup_incremental_init_copy_bitmap(BackupBlockJob *job)
bdrv_set_dirty_iter(dbi, next_cluster * job->cluster_size);
}
- /* TODO block_job_progress_set_remaining() would make more sense */
- block_job_progress_update(&job->common,
+ /* TODO job_progress_set_remaining() would make more sense */
+ job_progress_update(&job->common.job,
job->len - hbitmap_count(job->copy_bitmap) * job->cluster_size);
bdrv_dirty_iter_free(dbi);
@@ -425,7 +425,7 @@ static void coroutine_fn backup_run(void *opaque)
qemu_co_rwlock_init(&job->flush_rwlock);
nb_clusters = DIV_ROUND_UP(job->len, job->cluster_size);
- block_job_progress_set_remaining(&job->common, job->len);
+ job_progress_set_remaining(&job->common.job, job->len);
job->copy_bitmap = hbitmap_alloc(nb_clusters, 0);
if (job->sync_mode == MIRROR_SYNC_MODE_INCREMENTAL) {
diff --git a/block/commit.c b/block/commit.c
index b0a847e678..620666161b 100644
--- a/block/commit.c
+++ b/block/commit.c
@@ -150,7 +150,7 @@ static void coroutine_fn commit_run(void *opaque)
if (len < 0) {
goto out;
}
- block_job_progress_set_remaining(&s->common, len);
+ job_progress_set_remaining(&s->common.job, len);
ret = base_len = blk_getlength(s->base);
if (base_len < 0) {
@@ -196,7 +196,7 @@ static void coroutine_fn commit_run(void *opaque)
}
}
/* Publish progress */
- block_job_progress_update(&s->common, n);
+ job_progress_update(&s->common.job, n);
if (copy) {
delay_ns = block_job_ratelimit_get_delay(&s->common, n);
diff --git a/block/mirror.c b/block/mirror.c
index bdc1b5b7b9..dcb66ec3be 100644
--- a/block/mirror.c
+++ b/block/mirror.c
@@ -119,7 +119,7 @@ static void mirror_iteration_done(MirrorOp *op, int ret)
bitmap_set(s->cow_bitmap, chunk_num, nb_chunks);
}
if (!s->initial_zeroing_ongoing) {
- block_job_progress_update(&s->common, op->bytes);
+ job_progress_update(&s->common.job, op->bytes);
}
}
qemu_iovec_destroy(&op->qiov);
@@ -792,7 +792,7 @@ static void coroutine_fn mirror_run(void *opaque)
/* cnt is the number of dirty bytes remaining and s->bytes_in_flight is
* the number of bytes currently being processed; together those are
* the current remaining operation length */
- block_job_progress_set_remaining(&s->common, s->bytes_in_flight + cnt);
+ job_progress_set_remaining(&s->common.job, s->bytes_in_flight + cnt);
/* Note that even when no rate limit is applied we need to yield
* periodically with no pending I/O so that bdrv_drain_all() returns.
diff --git a/block/stream.c b/block/stream.c
index 8546c412cd..a5d6e0cf8a 100644
--- a/block/stream.c
+++ b/block/stream.c
@@ -121,7 +121,7 @@ static void coroutine_fn stream_run(void *opaque)
ret = len;
goto out;
}
- block_job_progress_set_remaining(&s->common, len);
+ job_progress_set_remaining(&s->common.job, len);
buf = qemu_blockalign(bs, STREAM_BUFFER_SIZE);
@@ -184,7 +184,7 @@ static void coroutine_fn stream_run(void *opaque)
ret = 0;
/* Publish progress */
- block_job_progress_update(&s->common, n);
+ job_progress_update(&s->common.job, n);
if (copy) {
delay_ns = block_job_ratelimit_get_delay(&s->common, n);
} else {
diff --git a/blockjob.c b/blockjob.c
index da11b3b763..5c8ff6f846 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -242,16 +242,6 @@ int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n)
return ratelimit_calculate_delay(&job->limit, n);
}
-void block_job_progress_update(BlockJob *job, uint64_t done)
-{
- job->offset += done;
-}
-
-void block_job_progress_set_remaining(BlockJob *job, uint64_t remaining)
-{
- job->len = job->offset + remaining;
-}
-
BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
{
BlockJobInfo *info;
@@ -263,10 +253,10 @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
info = g_new0(BlockJobInfo, 1);
info->type = g_strdup(job_type_str(&job->job));
info->device = g_strdup(job->job.id);
- info->len = job->len;
info->busy = atomic_read(&job->job.busy);
info->paused = job->job.pause_count > 0;
- info->offset = job->offset;
+ info->offset = job->job.progress_current;
+ info->len = job->job.progress_total;
info->speed = job->speed;
info->io_status = job->iostatus;
info->ready = job_is_ready(&job->job),
@@ -296,8 +286,8 @@ static void block_job_event_cancelled(Notifier *n, void *opaque)
qapi_event_send_block_job_cancelled(job_type(&job->job),
job->job.id,
- job->len,
- job->offset,
+ job->job.progress_total,
+ job->job.progress_current,
job->speed,
&error_abort);
}
@@ -317,8 +307,8 @@ static void block_job_event_completed(Notifier *n, void *opaque)
qapi_event_send_block_job_completed(job_type(&job->job),
job->job.id,
- job->len,
- job->offset,
+ job->job.progress_total,
+ job->job.progress_current,
job->speed,
!!msg,
msg,
@@ -348,8 +338,8 @@ static void block_job_event_ready(Notifier *n, void *opaque)
qapi_event_send_block_job_ready(job_type(&job->job),
job->job.id,
- job->len,
- job->offset,
+ job->job.progress_total,
+ job->job.progress_current,
job->speed, &error_abort);
}
diff --git a/job.c b/job.c
index 34304f2883..f7ab94352a 100644
--- a/job.c
+++ b/job.c
@@ -364,6 +364,16 @@ void job_unref(Job *job)
}
}
+void job_progress_update(Job *job, uint64_t done)
+{
+ job->progress_current += done;
+}
+
+void job_progress_set_remaining(Job *job, uint64_t remaining)
+{
+ job->progress_total = job->progress_current + remaining;
+}
+
void job_event_cancelled(Job *job)
{
notifier_list_notify(&job->on_finalize_cancelled, job);
diff --git a/qemu-img.c b/qemu-img.c
index eecb7d3228..e67b7a5844 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -864,9 +864,13 @@ static void run_block_job(BlockJob *job, Error **errp)
aio_context_acquire(aio_context);
job_ref(&job->job);
do {
+ float progress = 0.0f;
aio_poll(aio_context, true);
- qemu_progress_print(job->len ?
- ((float)job->offset / job->len * 100.f) : 0.0f, 0);
+ if (job->job.progress_total) {
+ progress = (float)job->job.progress_current /
+ job->job.progress_total * 100.f;
+ }
+ qemu_progress_print(progress, 0);
} while (!job_is_ready(&job->job) && !job_is_completed(&job->job));
if (!job_is_completed(&job->job)) {
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 34/40] job: Introduce qapi/job.json
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (32 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 33/40] job: Move progress fields to Job Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 15:59 ` Eric Blake
2018-05-31 21:21 ` Eric Blake
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event Kevin Wolf
` (8 subsequent siblings)
42 siblings, 2 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This adds a separate schema file for all job-related definitions that
aren't tied to the block layer.
For a start, move the enums JobType, JobStatus and JobVerb.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
qapi/block-core.json | 90 +-----------------------------------------------
qapi/job.json | 94 +++++++++++++++++++++++++++++++++++++++++++++++++++
qapi/qapi-schema.json | 1 +
Makefile | 9 +++++
Makefile.objs | 4 +++
5 files changed, 109 insertions(+), 89 deletions(-)
create mode 100644 qapi/job.json
diff --git a/qapi/block-core.json b/qapi/block-core.json
index bb964b4319..7dfa77a05c 100644
--- a/qapi/block-core.json
+++ b/qapi/block-core.json
@@ -6,6 +6,7 @@
{ 'include': 'common.json' }
{ 'include': 'crypto.json' }
+{ 'include': 'job.json' }
{ 'include': 'sockets.json' }
##
@@ -1050,95 +1051,6 @@
'data': ['top', 'full', 'none', 'incremental'] }
##
-# @JobType:
-#
-# Type of a background job.
-#
-# @commit: block commit job type, see "block-commit"
-#
-# @stream: block stream job type, see "block-stream"
-#
-# @mirror: drive mirror job type, see "drive-mirror"
-#
-# @backup: drive backup job type, see "drive-backup"
-#
-# Since: 1.7
-##
-{ 'enum': 'JobType',
- 'data': ['commit', 'stream', 'mirror', 'backup'] }
-
-##
-# @JobVerb:
-#
-# Represents command verbs that can be applied to a job.
-#
-# @cancel: see @block-job-cancel
-#
-# @pause: see @block-job-pause
-#
-# @resume: see @block-job-resume
-#
-# @set-speed: see @block-job-set-speed
-#
-# @complete: see @block-job-complete
-#
-# @dismiss: see @block-job-dismiss
-#
-# @finalize: see @block-job-finalize
-#
-# Since: 2.12
-##
-{ 'enum': 'JobVerb',
- 'data': ['cancel', 'pause', 'resume', 'set-speed', 'complete', 'dismiss',
- 'finalize' ] }
-
-##
-# @JobStatus:
-#
-# Indicates the present state of a given job in its lifetime.
-#
-# @undefined: Erroneous, default state. Should not ever be visible.
-#
-# @created: The job has been created, but not yet started.
-#
-# @running: The job is currently running.
-#
-# @paused: The job is running, but paused. The pause may be requested by
-# either the QMP user or by internal processes.
-#
-# @ready: The job is running, but is ready for the user to signal completion.
-# This is used for long-running jobs like mirror that are designed to
-# run indefinitely.
-#
-# @standby: The job is ready, but paused. This is nearly identical to @paused.
-# The job may return to @ready or otherwise be canceled.
-#
-# @waiting: The job is waiting for other jobs in the transaction to converge
-# to the waiting state. This status will likely not be visible for
-# the last job in a transaction.
-#
-# @pending: The job has finished its work, but has finalization steps that it
-# needs to make prior to completing. These changes may require
-# manual intervention by the management process if manual was set
-# to true. These changes may still fail.
-#
-# @aborting: The job is in the process of being aborted, and will finish with
-# an error. The job will afterwards report that it is @concluded.
-# This status may not be visible to the management process.
-#
-# @concluded: The job has finished all work. If manual was set to true, the job
-# will remain in the query list until it is dismissed.
-#
-# @null: The job is in the process of being dismantled. This state should not
-# ever be visible externally.
-#
-# Since: 2.12
-##
-{ 'enum': 'JobStatus',
- 'data': ['undefined', 'created', 'running', 'paused', 'ready', 'standby',
- 'waiting', 'pending', 'aborting', 'concluded', 'null' ] }
-
-##
# @BlockJobInfo:
#
# Information about a long-running block device operation.
diff --git a/qapi/job.json b/qapi/job.json
new file mode 100644
index 0000000000..a472c0cb29
--- /dev/null
+++ b/qapi/job.json
@@ -0,0 +1,94 @@
+# -*- Mode: Python -*-
+
+##
+# == Background jobs
+##
+
+##
+# @JobType:
+#
+# Type of a background job.
+#
+# @commit: block commit job type, see "block-commit"
+#
+# @stream: block stream job type, see "block-stream"
+#
+# @mirror: drive mirror job type, see "drive-mirror"
+#
+# @backup: drive backup job type, see "drive-backup"
+#
+# Since: 1.7
+##
+{ 'enum': 'JobType',
+ 'data': ['commit', 'stream', 'mirror', 'backup'] }
+
+##
+# @JobStatus:
+#
+# Indicates the present state of a given job in its lifetime.
+#
+# @undefined: Erroneous, default state. Should not ever be visible.
+#
+# @created: The job has been created, but not yet started.
+#
+# @running: The job is currently running.
+#
+# @paused: The job is running, but paused. The pause may be requested by
+# either the QMP user or by internal processes.
+#
+# @ready: The job is running, but is ready for the user to signal completion.
+# This is used for long-running jobs like mirror that are designed to
+# run indefinitely.
+#
+# @standby: The job is ready, but paused. This is nearly identical to @paused.
+# The job may return to @ready or otherwise be canceled.
+#
+# @waiting: The job is waiting for other jobs in the transaction to converge
+# to the waiting state. This status will likely not be visible for
+# the last job in a transaction.
+#
+# @pending: The job has finished its work, but has finalization steps that it
+# needs to make prior to completing. These changes may require
+# manual intervention by the management process if manual was set
+# to true. These changes may still fail.
+#
+# @aborting: The job is in the process of being aborted, and will finish with
+# an error. The job will afterwards report that it is @concluded.
+# This status may not be visible to the management process.
+#
+# @concluded: The job has finished all work. If manual was set to true, the job
+# will remain in the query list until it is dismissed.
+#
+# @null: The job is in the process of being dismantled. This state should not
+# ever be visible externally.
+#
+# Since: 2.12
+##
+{ 'enum': 'JobStatus',
+ 'data': ['undefined', 'created', 'running', 'paused', 'ready', 'standby',
+ 'waiting', 'pending', 'aborting', 'concluded', 'null' ] }
+
+##
+# @JobVerb:
+#
+# Represents command verbs that can be applied to a job.
+#
+# @cancel: see @block-job-cancel
+#
+# @pause: see @block-job-pause
+#
+# @resume: see @block-job-resume
+#
+# @set-speed: see @block-job-set-speed
+#
+# @complete: see @block-job-complete
+#
+# @dismiss: see @block-job-dismiss
+#
+# @finalize: see @block-job-finalize
+#
+# Since: 2.12
+##
+{ 'enum': 'JobVerb',
+ 'data': ['cancel', 'pause', 'resume', 'set-speed', 'complete', 'dismiss',
+ 'finalize' ] }
diff --git a/qapi/qapi-schema.json b/qapi/qapi-schema.json
index 25bce78352..65b6dc2f6f 100644
--- a/qapi/qapi-schema.json
+++ b/qapi/qapi-schema.json
@@ -84,6 +84,7 @@
{ 'include': 'crypto.json' }
{ 'include': 'block.json' }
{ 'include': 'char.json' }
+{ 'include': 'job.json' }
{ 'include': 'net.json' }
{ 'include': 'rocker.json' }
{ 'include': 'tpm.json' }
diff --git a/Makefile b/Makefile
index 35554b5bef..6d588d1f71 100644
--- a/Makefile
+++ b/Makefile
@@ -98,6 +98,7 @@ GENERATED_FILES += qapi/qapi-types-char.h qapi/qapi-types-char.c
GENERATED_FILES += qapi/qapi-types-common.h qapi/qapi-types-common.c
GENERATED_FILES += qapi/qapi-types-crypto.h qapi/qapi-types-crypto.c
GENERATED_FILES += qapi/qapi-types-introspect.h qapi/qapi-types-introspect.c
+GENERATED_FILES += qapi/qapi-types-job.h qapi/qapi-types-job.c
GENERATED_FILES += qapi/qapi-types-migration.h qapi/qapi-types-migration.c
GENERATED_FILES += qapi/qapi-types-misc.h qapi/qapi-types-misc.c
GENERATED_FILES += qapi/qapi-types-net.h qapi/qapi-types-net.c
@@ -116,6 +117,7 @@ GENERATED_FILES += qapi/qapi-visit-char.h qapi/qapi-visit-char.c
GENERATED_FILES += qapi/qapi-visit-common.h qapi/qapi-visit-common.c
GENERATED_FILES += qapi/qapi-visit-crypto.h qapi/qapi-visit-crypto.c
GENERATED_FILES += qapi/qapi-visit-introspect.h qapi/qapi-visit-introspect.c
+GENERATED_FILES += qapi/qapi-visit-job.h qapi/qapi-visit-job.c
GENERATED_FILES += qapi/qapi-visit-migration.h qapi/qapi-visit-migration.c
GENERATED_FILES += qapi/qapi-visit-misc.h qapi/qapi-visit-misc.c
GENERATED_FILES += qapi/qapi-visit-net.h qapi/qapi-visit-net.c
@@ -133,6 +135,7 @@ GENERATED_FILES += qapi/qapi-commands-char.h qapi/qapi-commands-char.c
GENERATED_FILES += qapi/qapi-commands-common.h qapi/qapi-commands-common.c
GENERATED_FILES += qapi/qapi-commands-crypto.h qapi/qapi-commands-crypto.c
GENERATED_FILES += qapi/qapi-commands-introspect.h qapi/qapi-commands-introspect.c
+GENERATED_FILES += qapi/qapi-commands-job.h qapi/qapi-commands-job.c
GENERATED_FILES += qapi/qapi-commands-migration.h qapi/qapi-commands-migration.c
GENERATED_FILES += qapi/qapi-commands-misc.h qapi/qapi-commands-misc.c
GENERATED_FILES += qapi/qapi-commands-net.h qapi/qapi-commands-net.c
@@ -150,6 +153,7 @@ GENERATED_FILES += qapi/qapi-events-char.h qapi/qapi-events-char.c
GENERATED_FILES += qapi/qapi-events-common.h qapi/qapi-events-common.c
GENERATED_FILES += qapi/qapi-events-crypto.h qapi/qapi-events-crypto.c
GENERATED_FILES += qapi/qapi-events-introspect.h qapi/qapi-events-introspect.c
+GENERATED_FILES += qapi/qapi-events-job.h qapi/qapi-events-job.c
GENERATED_FILES += qapi/qapi-events-migration.h qapi/qapi-events-migration.c
GENERATED_FILES += qapi/qapi-events-misc.h qapi/qapi-events-misc.c
GENERATED_FILES += qapi/qapi-events-net.h qapi/qapi-events-net.c
@@ -582,6 +586,7 @@ qapi-modules = $(SRC_PATH)/qapi/qapi-schema.json $(SRC_PATH)/qapi/common.json \
$(SRC_PATH)/qapi/char.json \
$(SRC_PATH)/qapi/crypto.json \
$(SRC_PATH)/qapi/introspect.json \
+ $(SRC_PATH)/qapi/job.json \
$(SRC_PATH)/qapi/migration.json \
$(SRC_PATH)/qapi/misc.json \
$(SRC_PATH)/qapi/net.json \
@@ -601,6 +606,7 @@ qapi/qapi-types-char.c qapi/qapi-types-char.h \
qapi/qapi-types-common.c qapi/qapi-types-common.h \
qapi/qapi-types-crypto.c qapi/qapi-types-crypto.h \
qapi/qapi-types-introspect.c qapi/qapi-types-introspect.h \
+qapi/qapi-types-job.c qapi/qapi-types-job.h \
qapi/qapi-types-migration.c qapi/qapi-types-migration.h \
qapi/qapi-types-misc.c qapi/qapi-types-misc.h \
qapi/qapi-types-net.c qapi/qapi-types-net.h \
@@ -619,6 +625,7 @@ qapi/qapi-visit-char.c qapi/qapi-visit-char.h \
qapi/qapi-visit-common.c qapi/qapi-visit-common.h \
qapi/qapi-visit-crypto.c qapi/qapi-visit-crypto.h \
qapi/qapi-visit-introspect.c qapi/qapi-visit-introspect.h \
+qapi/qapi-visit-job.c qapi/qapi-visit-job.h \
qapi/qapi-visit-migration.c qapi/qapi-visit-migration.h \
qapi/qapi-visit-misc.c qapi/qapi-visit-misc.h \
qapi/qapi-visit-net.c qapi/qapi-visit-net.h \
@@ -636,6 +643,7 @@ qapi/qapi-commands-char.c qapi/qapi-commands-char.h \
qapi/qapi-commands-common.c qapi/qapi-commands-common.h \
qapi/qapi-commands-crypto.c qapi/qapi-commands-crypto.h \
qapi/qapi-commands-introspect.c qapi/qapi-commands-introspect.h \
+qapi/qapi-commands-job.c qapi/qapi-commands-job.h \
qapi/qapi-commands-migration.c qapi/qapi-commands-migration.h \
qapi/qapi-commands-misc.c qapi/qapi-commands-misc.h \
qapi/qapi-commands-net.c qapi/qapi-commands-net.h \
@@ -653,6 +661,7 @@ qapi/qapi-events-char.c qapi/qapi-events-char.h \
qapi/qapi-events-common.c qapi/qapi-events-common.h \
qapi/qapi-events-crypto.c qapi/qapi-events-crypto.h \
qapi/qapi-events-introspect.c qapi/qapi-events-introspect.h \
+qapi/qapi-events-job.c qapi/qapi-events-job.h \
qapi/qapi-events-migration.c qapi/qapi-events-migration.h \
qapi/qapi-events-misc.c qapi/qapi-events-misc.h \
qapi/qapi-events-net.c qapi/qapi-events-net.h \
diff --git a/Makefile.objs b/Makefile.objs
index 92b73fc272..3df8d58e49 100644
--- a/Makefile.objs
+++ b/Makefile.objs
@@ -10,6 +10,7 @@ util-obj-y += qapi/qapi-types-char.o
util-obj-y += qapi/qapi-types-common.o
util-obj-y += qapi/qapi-types-crypto.o
util-obj-y += qapi/qapi-types-introspect.o
+util-obj-y += qapi/qapi-types-job.o
util-obj-y += qapi/qapi-types-migration.o
util-obj-y += qapi/qapi-types-misc.o
util-obj-y += qapi/qapi-types-net.o
@@ -28,6 +29,7 @@ util-obj-y += qapi/qapi-visit-char.o
util-obj-y += qapi/qapi-visit-common.o
util-obj-y += qapi/qapi-visit-crypto.o
util-obj-y += qapi/qapi-visit-introspect.o
+util-obj-y += qapi/qapi-visit-job.o
util-obj-y += qapi/qapi-visit-migration.o
util-obj-y += qapi/qapi-visit-misc.o
util-obj-y += qapi/qapi-visit-net.o
@@ -45,6 +47,7 @@ util-obj-y += qapi/qapi-events-char.o
util-obj-y += qapi/qapi-events-common.o
util-obj-y += qapi/qapi-events-crypto.o
util-obj-y += qapi/qapi-events-introspect.o
+util-obj-y += qapi/qapi-events-job.o
util-obj-y += qapi/qapi-events-migration.o
util-obj-y += qapi/qapi-events-misc.o
util-obj-y += qapi/qapi-events-net.o
@@ -140,6 +143,7 @@ common-obj-y += qapi/qapi-commands-char.o
common-obj-y += qapi/qapi-commands-common.o
common-obj-y += qapi/qapi-commands-crypto.o
common-obj-y += qapi/qapi-commands-introspect.o
+common-obj-y += qapi/qapi-commands-job.o
common-obj-y += qapi/qapi-commands-migration.o
common-obj-y += qapi/qapi-commands-misc.o
common-obj-y += qapi/qapi-commands-net.o
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 34/40] job: Introduce qapi/job.json
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 34/40] job: Introduce qapi/job.json Kevin Wolf
@ 2018-05-18 15:59 ` Eric Blake
2018-05-31 21:21 ` Eric Blake
1 sibling, 0 replies; 86+ messages in thread
From: Eric Blake @ 2018-05-18 15:59 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:21 AM, Kevin Wolf wrote:
> This adds a separate schema file for all job-related definitions that
> aren't tied to the block layer.
>
> For a start, move the enums JobType, JobStatus and JobVerb.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
> qapi/block-core.json | 90 +-----------------------------------------------
> qapi/job.json | 94 +++++++++++++++++++++++++++++++++++++++++++++++++++
> qapi/qapi-schema.json | 1 +
> Makefile | 9 +++++
> Makefile.objs | 4 +++
> 5 files changed, 109 insertions(+), 89 deletions(-)
> create mode 100644 qapi/job.json
Do you want to add a MAINTAINERS entry under a block section, so that
the block list is cc'd on changes to the new file, in addition to the
fact that it is already covered by normal QAPI reviewers? (If we do
that, block-core.json could also use such additional coverage.)
> +++ b/Makefile
> @@ -98,6 +98,7 @@ GENERATED_FILES += qapi/qapi-types-char.h qapi/qapi-types-char.c
> GENERATED_FILES += qapi/qapi-types-common.h qapi/qapi-types-common.c
> GENERATED_FILES += qapi/qapi-types-crypto.h qapi/qapi-types-crypto.c
> GENERATED_FILES += qapi/qapi-types-introspect.h qapi/qapi-types-introspect.c
> +GENERATED_FILES += qapi/qapi-types-job.h qapi/qapi-types-job.c
> GENERATED_FILES += qapi/qapi-types-migration.h qapi/qapi-types-migration.c
> GENERATED_FILES += qapi/qapi-types-misc.h qapi/qapi-types-misc.c
> GENERATED_FILES += qapi/qapi-types-net.h qapi/qapi-types-net.c
I've threatened to post a refactoring patch to make it so that adding a
new schema file doesn't require so many edits in the Makefile. Maybe I
should make good on that threat :)
But in the meantime, this looks good.
Reviewed-by: Eric Blake <eblake@redhat.com>
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 34/40] job: Introduce qapi/job.json
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 34/40] job: Introduce qapi/job.json Kevin Wolf
2018-05-18 15:59 ` Eric Blake
@ 2018-05-31 21:21 ` Eric Blake
1 sibling, 0 replies; 86+ messages in thread
From: Eric Blake @ 2018-05-31 21:21 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:21 AM, Kevin Wolf wrote:
> This adds a separate schema file for all job-related definitions that
> aren't tied to the block layer.
>
> For a start, move the enums JobType, JobStatus and JobVerb.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
> qapi/block-core.json | 90 +-----------------------------------------------
> qapi/job.json | 94 +++++++++++++++++++++++++++++++++++++++++++++++++++
> qapi/qapi-schema.json | 1 +
> Makefile | 9 +++++
> Makefile.objs | 4 +++
> 5 files changed, 109 insertions(+), 89 deletions(-)
> create mode 100644 qapi/job.json
This commit now results in a dirty 'git status' for in-tree builds:
qapi/qapi-commands-job.c
qapi/qapi-commands-job.h
qapi/qapi-events-job.c
qapi/qapi-events-job.h
qapi/qapi-types-job.c
qapi/qapi-types-job.h
qapi/qapi-visit-job.c
qapi/qapi-visit-job.h
I'll propose a cleanup patch.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (33 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 34/40] job: Introduce qapi/job.json Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 17:55 ` Eric Blake
2018-05-24 0:02 ` John Snow
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 36/40] job: Add lifecycle QMP commands Kevin Wolf
` (7 subsequent siblings)
42 siblings, 2 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This adds a QMP event that is emitted whenever a job transitions from
one status to another.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
qapi/job.json | 14 ++++
job.c | 10 +++
tests/qemu-iotests/030 | 17 +++-
tests/qemu-iotests/040 | 2 +
tests/qemu-iotests/041 | 17 +++-
tests/qemu-iotests/094.out | 7 ++
tests/qemu-iotests/095 | 2 +-
tests/qemu-iotests/095.out | 6 ++
tests/qemu-iotests/109 | 2 +-
tests/qemu-iotests/109.out | 178 ++++++++++++++++++++++++++++++++++++------
tests/qemu-iotests/124 | 8 ++
tests/qemu-iotests/127.out | 7 ++
tests/qemu-iotests/141 | 13 ++-
tests/qemu-iotests/141.out | 29 +++++++
tests/qemu-iotests/144 | 2 +-
tests/qemu-iotests/144.out | 7 ++
tests/qemu-iotests/156 | 2 +-
tests/qemu-iotests/156.out | 7 ++
tests/qemu-iotests/185 | 12 +--
tests/qemu-iotests/185.out | 10 +++
tests/qemu-iotests/191 | 4 +-
tests/qemu-iotests/191.out | 132 +++++++++++++++++++++++++++++++
tests/qemu-iotests/iotests.py | 5 ++
23 files changed, 449 insertions(+), 44 deletions(-)
diff --git a/qapi/job.json b/qapi/job.json
index a472c0cb29..9fbdd0ccd9 100644
--- a/qapi/job.json
+++ b/qapi/job.json
@@ -92,3 +92,17 @@
{ 'enum': 'JobVerb',
'data': ['cancel', 'pause', 'resume', 'set-speed', 'complete', 'dismiss',
'finalize' ] }
+
+##
+# @JOB_STATUS_CHANGE:
+#
+# Emitted when a job transitions to a different status.
+#
+# @id: The job identifier
+# @status: The new job status
+#
+# Since: 2.13
+##
+{ 'event': 'JOB_STATUS_CHANGE',
+ 'data': { 'id': 'str',
+ 'status': 'JobStatus' } }
diff --git a/job.c b/job.c
index f7ab94352a..8e106e5d15 100644
--- a/job.c
+++ b/job.c
@@ -30,6 +30,7 @@
#include "qemu/id.h"
#include "qemu/main-loop.h"
#include "trace-root.h"
+#include "qapi/qapi-events-job.h"
static QLIST_HEAD(, Job) jobs = QLIST_HEAD_INITIALIZER(jobs);
@@ -157,6 +158,11 @@ static int job_txn_apply(JobTxn *txn, int fn(Job *), bool lock)
return rc;
}
+static bool job_is_internal(Job *job)
+{
+ return (job->id == NULL);
+}
+
static void job_state_transition(Job *job, JobStatus s1)
{
JobStatus s0 = job->status;
@@ -166,6 +172,10 @@ static void job_state_transition(Job *job, JobStatus s1)
JobStatus_str(s0), JobStatus_str(s1));
assert(JobSTT[s0][s1]);
job->status = s1;
+
+ if (!job_is_internal(job) && s1 != s0) {
+ qapi_event_send_job_status_change(job->id, job->status, &error_abort);
+ }
}
int job_apply_verb(Job *job, JobVerb verb, Error **errp)
diff --git a/tests/qemu-iotests/030 b/tests/qemu-iotests/030
index 640a6dfd10..1dbc2ddc49 100755
--- a/tests/qemu-iotests/030
+++ b/tests/qemu-iotests/030
@@ -304,8 +304,7 @@ class TestParallelOps(iotests.QMPTestCase):
result = self.vm.qmp('block-stream', device='node5', base=self.imgs[3], job_id='stream-node6')
self.assert_qmp(result, 'error/class', 'GenericError')
- event = self.vm.get_qmp_event(wait=True)
- self.assertEqual(event['event'], 'BLOCK_JOB_READY')
+ event = self.vm.event_wait(name='BLOCK_JOB_READY')
self.assert_qmp(event, 'data/device', 'commit-drive0')
self.assert_qmp(event, 'data/type', 'commit')
self.assert_qmp_absent(event, 'data/error')
@@ -565,6 +564,8 @@ class TestEIO(TestErrors):
self.assert_qmp(event, 'data/offset', self.STREAM_BUFFER_SIZE)
self.assert_qmp(event, 'data/len', self.image_len)
completed = True
+ elif event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
self.assert_no_active_block_jobs()
self.vm.shutdown()
@@ -596,6 +597,8 @@ class TestEIO(TestErrors):
self.assert_qmp(event, 'data/offset', self.image_len)
self.assert_qmp(event, 'data/len', self.image_len)
completed = True
+ elif event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
self.assert_no_active_block_jobs()
self.vm.shutdown()
@@ -637,6 +640,8 @@ class TestEIO(TestErrors):
self.assert_qmp(event, 'data/offset', self.image_len)
self.assert_qmp(event, 'data/len', self.image_len)
completed = True
+ elif event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
self.assert_no_active_block_jobs()
self.vm.shutdown()
@@ -663,6 +668,8 @@ class TestEIO(TestErrors):
self.assert_qmp(event, 'data/offset', self.STREAM_BUFFER_SIZE)
self.assert_qmp(event, 'data/len', self.image_len)
completed = True
+ elif event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
self.assert_no_active_block_jobs()
self.vm.shutdown()
@@ -722,6 +729,8 @@ class TestENOSPC(TestErrors):
self.assert_qmp(event, 'data/offset', self.image_len)
self.assert_qmp(event, 'data/len', self.image_len)
completed = True
+ elif event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
self.assert_no_active_block_jobs()
self.vm.shutdown()
@@ -751,7 +760,9 @@ class TestStreamStop(iotests.QMPTestCase):
time.sleep(0.1)
events = self.vm.get_qmp_events(wait=False)
- self.assertEqual(events, [], 'unexpected QMP event: %s' % events)
+ for e in events:
+ self.assert_qmp(e, 'event', 'JOB_STATUS_CHANGE')
+ self.assert_qmp(e, 'data/id', 'drive0')
self.cancel_and_wait(resume=True)
diff --git a/tests/qemu-iotests/040 b/tests/qemu-iotests/040
index 90b5b4f2ad..1beb5e6dab 100755
--- a/tests/qemu-iotests/040
+++ b/tests/qemu-iotests/040
@@ -162,6 +162,8 @@ class TestSingleDrive(ImageCommitTestCase):
elif event['event'] == 'BLOCK_JOB_CANCELLED':
self.assert_qmp(event, 'data/device', 'drive0')
cancelled = True
+ elif event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
else:
self.fail("Unexpected event %s" % (event['event']))
diff --git a/tests/qemu-iotests/041 b/tests/qemu-iotests/041
index a860a31e9a..e94587950c 100755
--- a/tests/qemu-iotests/041
+++ b/tests/qemu-iotests/041
@@ -445,6 +445,8 @@ new_state = "1"
self.assert_qmp(event, 'data/device', 'drive0')
self.assert_qmp(event, 'data/error', 'Input/output error')
completed = True
+ elif event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
self.assert_no_active_block_jobs()
self.vm.shutdown()
@@ -457,6 +459,10 @@ new_state = "1"
self.assert_qmp(result, 'return', {})
event = self.vm.get_qmp_event(wait=True)
+ while event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
+ event = self.vm.get_qmp_event(wait=True)
+
self.assertEquals(event['event'], 'BLOCK_JOB_ERROR')
self.assert_qmp(event, 'data/device', 'drive0')
self.assert_qmp(event, 'data/operation', 'read')
@@ -478,6 +484,10 @@ new_state = "1"
self.assert_qmp(result, 'return', {})
event = self.vm.get_qmp_event(wait=True)
+ while event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
+ event = self.vm.get_qmp_event(wait=True)
+
self.assertEquals(event['event'], 'BLOCK_JOB_ERROR')
self.assert_qmp(event, 'data/device', 'drive0')
self.assert_qmp(event, 'data/operation', 'read')
@@ -608,7 +618,7 @@ new_state = "1"
on_target_error='ignore')
self.assert_qmp(result, 'return', {})
- event = self.vm.get_qmp_event(wait=True)
+ event = self.vm.event_wait(name='BLOCK_JOB_ERROR')
self.assertEquals(event['event'], 'BLOCK_JOB_ERROR')
self.assert_qmp(event, 'data/device', 'drive0')
self.assert_qmp(event, 'data/operation', 'write')
@@ -784,7 +794,12 @@ class TestGranularity(iotests.QMPTestCase):
sync='full', target=target_img,
mode='absolute-paths', granularity=8192)
self.assert_qmp(result, 'return', {})
+
event = self.vm.get_qmp_event(wait=60.0)
+ while event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', 'drive0')
+ event = self.vm.get_qmp_event(wait=60.0)
+
# Failures will manifest as COMPLETED/ERROR.
self.assert_qmp(event, 'event', 'BLOCK_JOB_READY')
self.complete_and_wait(drive='drive0', wait_ready=False)
diff --git a/tests/qemu-iotests/094.out b/tests/qemu-iotests/094.out
index f52baffe70..665b630b08 100644
--- a/tests/qemu-iotests/094.out
+++ b/tests/qemu-iotests/094.out
@@ -2,10 +2,17 @@ QA output created by 094
Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
Formatting 'TEST_DIR/source.IMGFMT', fmt=IMGFMT size=67108864
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 67108864, "offset": 67108864, "speed": 0, "type": "mirror"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 67108864, "offset": 67108864, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
*** done
diff --git a/tests/qemu-iotests/095 b/tests/qemu-iotests/095
index 030adb22e1..72ecc22e1b 100755
--- a/tests/qemu-iotests/095
+++ b/tests/qemu-iotests/095
@@ -72,7 +72,7 @@ _send_qemu_cmd $h "{ 'execute': 'qmp_capabilities' }" "return"
_send_qemu_cmd $h "{ 'execute': 'block-commit',
'arguments': { 'device': 'test',
- 'top': '"${TEST_IMG}.snp1"' } }" "BLOCK_JOB_COMPLETED"
+ 'top': '"${TEST_IMG}.snp1"' } }" '"status": "null"'
_cleanup_qemu
diff --git a/tests/qemu-iotests/095.out b/tests/qemu-iotests/095.out
index 73875cab40..8c093dfff3 100644
--- a/tests/qemu-iotests/095.out
+++ b/tests/qemu-iotests/095.out
@@ -11,8 +11,14 @@ virtual size: 5.0M (5242880 bytes)
=== Running QEMU Live Commit Test ===
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "test"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "test"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "test"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "test"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "test", "len": 104857600, "offset": 104857600, "speed": 0, "type": "commit"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "test"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "test"}}
=== Base image info after commit and resize ===
image: TEST_DIR/t.IMGFMT.base
diff --git a/tests/qemu-iotests/109 b/tests/qemu-iotests/109
index d70b574d88..acbd079136 100755
--- a/tests/qemu-iotests/109
+++ b/tests/qemu-iotests/109
@@ -64,7 +64,7 @@ function run_qemu()
_send_qemu_cmd $QEMU_HANDLE '' "$qmp_event"
if test "$qmp_event" = BLOCK_JOB_ERROR; then
- _send_qemu_cmd $QEMU_HANDLE '' "BLOCK_JOB_COMPLETED"
+ _send_qemu_cmd $QEMU_HANDLE '' '"status": "null"'
fi
_send_qemu_cmd $QEMU_HANDLE '{"execute":"query-block-jobs"}' "return"
_send_qemu_cmd $QEMU_HANDLE '{"execute":"quit"}' "return"
diff --git a/tests/qemu-iotests/109.out b/tests/qemu-iotests/109.out
index 8a9b93672b..ad0ee6fb48 100644
--- a/tests/qemu-iotests/109.out
+++ b/tests/qemu-iotests/109.out
@@ -6,23 +6,35 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
Formatting 'TEST_DIR/t.raw.src', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": 0, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 1024, "offset": 1024, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 1024, "offset": 1024, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 1024, "offset": 1024, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -32,23 +44,35 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
Formatting 'TEST_DIR/t.raw.src', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": 512, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 197120, "offset": 197120, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 197120, "offset": 197120, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 197120, "offset": 197120, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -58,23 +82,35 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
Formatting 'TEST_DIR/t.raw.src', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": 262144, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 327680, "offset": 327680, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 327680, "offset": 327680, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 327680, "offset": 327680, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -84,23 +120,35 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
Formatting 'TEST_DIR/t.raw.src', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": 0, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 1024, "offset": 1024, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 1024, "offset": 1024, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 1024, "offset": 1024, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -110,23 +158,35 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
Formatting 'TEST_DIR/t.raw.src', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": 0, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 65536, "offset": 65536, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 65536, "offset": 65536, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 65536, "offset": 65536, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -136,23 +196,35 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
Formatting 'TEST_DIR/t.raw.src', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": 0, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 2560, "offset": 2560, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 2560, "offset": 2560, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 2560, "offset": 2560, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -161,23 +233,35 @@ Images are identical.
Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": OFFSET, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 2560, "offset": 2560, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 2560, "offset": 2560, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 2560, "offset": 2560, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -186,23 +270,35 @@ Images are identical.
Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": OFFSET, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 31457280, "offset": 31457280, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 31457280, "offset": 31457280, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 31457280, "offset": 31457280, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -211,23 +307,35 @@ Images are identical.
Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": OFFSET, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 327680, "offset": 327680, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 327680, "offset": 327680, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 327680, "offset": 327680, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -236,23 +344,35 @@ Images are identical.
Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_ERROR", "data": {"device": "src", "operation": "write", "action": "report"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": LEN, "offset": OFFSET, "speed": 0, "type": "mirror", "error": "Operation not permitted"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
{"return": []}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 2048, "offset": 2048, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 2048, "offset": 2048, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 2048, "offset": 2048, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
@@ -261,23 +381,37 @@ Images are identical.
Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
{"return": {}}
WARNING: Image format was not specified for 'TEST_DIR/t.raw' and probing guessed raw.
-Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
-Specify the 'raw' format explicitly to remove the restrictions.
+ Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
+ Specify the 'raw' format explicitly to remove the restrictions.
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 512, "offset": 512, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 512, "offset": 512, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 512, "offset": 512, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "src"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "src", "len": 512, "offset": 512, "speed": 0, "type": "mirror"}}
{"return": [{"auto-finalize": true, "io-status": "ok", "device": "src", "auto-dismiss": true, "busy": false, "len": 512, "offset": 512, "status": "ready", "paused": false, "speed": 0, "ready": true, "type": "mirror"}]}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "src"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "src", "len": 512, "offset": 512, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "src"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "src"}}
Warning: Image size mismatch!
Images are identical.
*** done
diff --git a/tests/qemu-iotests/124 b/tests/qemu-iotests/124
index 8e76e62f93..3ea4ac53f5 100755
--- a/tests/qemu-iotests/124
+++ b/tests/qemu-iotests/124
@@ -151,10 +151,17 @@ class TestIncrementalBackupBase(iotests.QMPTestCase):
return self.wait_qmp_backup(kwargs['device'], error)
+ def ignore_job_status_change_events(self):
+ while True:
+ e = self.vm.event_wait(name="JOB_STATUS_CHANGE")
+ if e['data']['status'] == 'null':
+ break
+
def wait_qmp_backup(self, device, error='Input/output error'):
event = self.vm.event_wait(name="BLOCK_JOB_COMPLETED",
match={'data': {'device': device}})
self.assertNotEqual(event, None)
+ self.ignore_job_status_change_events()
try:
failure = self.dictpath(event, 'data/error')
@@ -172,6 +179,7 @@ class TestIncrementalBackupBase(iotests.QMPTestCase):
event = self.vm.event_wait(name='BLOCK_JOB_CANCELLED',
match={'data': {'device': device}})
self.assertNotEqual(event, None)
+ self.ignore_job_status_change_events()
def create_anchor_backup(self, drive=None):
diff --git a/tests/qemu-iotests/127.out b/tests/qemu-iotests/127.out
index 543d075005..83b522d4c2 100644
--- a/tests/qemu-iotests/127.out
+++ b/tests/qemu-iotests/127.out
@@ -5,10 +5,17 @@ Formatting 'TEST_DIR/t.IMGFMT.overlay1', fmt=IMGFMT size=65536 backing_file=TEST
wrote 42/42 bytes at offset 0
42 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "mirror"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "mirror"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "mirror", "len": 65536, "offset": 65536, "speed": 0, "type": "mirror"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "mirror"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "mirror", "len": 65536, "offset": 65536, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "mirror"}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
*** done
diff --git a/tests/qemu-iotests/141 b/tests/qemu-iotests/141
index 2f9d7b9bc2..4246d387a1 100755
--- a/tests/qemu-iotests/141
+++ b/tests/qemu-iotests/141
@@ -107,7 +107,7 @@ test_blockjob \
'format': '$IMGFMT',
'sync': 'none'}}" \
'return' \
- 'BLOCK_JOB_CANCELLED'
+ '"status": "null"'
echo
echo '=== Testing drive-mirror ==='
@@ -124,7 +124,7 @@ test_blockjob \
'format': '$IMGFMT',
'sync': 'none'}}" \
'BLOCK_JOB_READY' \
- 'BLOCK_JOB_COMPLETED'
+ '"status": "null"'
echo
echo '=== Testing active block-commit ==='
@@ -138,7 +138,7 @@ test_blockjob \
"{'execute': 'block-commit',
'arguments': {'job-id': 'job0', 'device': 'drv0'}}" \
'BLOCK_JOB_READY' \
- 'BLOCK_JOB_COMPLETED'
+ '"status": "null"'
echo
echo '=== Testing non-active block-commit ==='
@@ -157,7 +157,7 @@ test_blockjob \
'top': '$TEST_DIR/m.$IMGFMT',
'speed': 1}}" \
'return' \
- 'BLOCK_JOB_CANCELLED'
+ '"status": "null"'
echo
echo '=== Testing block-stream ==='
@@ -170,8 +170,7 @@ echo
$QEMU_IO -c 'write 0 1M' "$TEST_DIR/b.$IMGFMT" | _filter_qemu_io
# With some data to stream (and @speed set to 1), block-stream will not complete
-# until we send the block-job-cancel command. Therefore, no event other than
-# BLOCK_JOB_CANCELLED will be emitted.
+# until we send the block-job-cancel command.
test_blockjob \
"{'execute': 'block-stream',
@@ -179,7 +178,7 @@ test_blockjob \
'device': 'drv0',
'speed': 1}}" \
'return' \
- 'BLOCK_JOB_CANCELLED'
+ '"status": "null"'
_cleanup_qemu
diff --git a/tests/qemu-iotests/141.out b/tests/qemu-iotests/141.out
index 82e763b68d..f252c86875 100644
--- a/tests/qemu-iotests/141.out
+++ b/tests/qemu-iotests/141.out
@@ -8,31 +8,50 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/m.
{"return": {}}
Formatting 'TEST_DIR/o.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.IMGFMT backing_fmt=IMGFMT
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
{"return": {}}
{"error": {"class": "GenericError", "desc": "Node drv0 is in use"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "job0"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_CANCELLED", "data": {"device": "job0", "len": 1048576, "offset": 0, "speed": 0, "type": "backup"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "job0"}}
{"return": {}}
=== Testing drive-mirror ===
{"return": {}}
Formatting 'TEST_DIR/o.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.IMGFMT backing_fmt=IMGFMT
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "job0"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "mirror"}}
{"return": {}}
{"error": {"class": "GenericError", "desc": "Node 'drv0' is busy: node is used as backing hd of 'NODE_NAME'"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "job0"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "job0"}}
{"return": {}}
=== Testing active block-commit ===
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "job0"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}}
{"return": {}}
{"error": {"class": "GenericError", "desc": "Node 'drv0' is busy: node is used as backing hd of 'NODE_NAME'"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "job0"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "job0"}}
{"return": {}}
=== Testing non-active block-commit ===
@@ -40,10 +59,15 @@ Formatting 'TEST_DIR/o.IMGFMT', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.
wrote 1048576/1048576 bytes at offset 0
1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
{"return": {}}
{"error": {"class": "GenericError", "desc": "Node drv0 is in use"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "job0"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_CANCELLED", "data": {"device": "job0", "len": 1048576, "offset": 524288, "speed": 1, "type": "commit"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "job0"}}
{"return": {}}
=== Testing block-stream ===
@@ -51,9 +75,14 @@ wrote 1048576/1048576 bytes at offset 0
wrote 1048576/1048576 bytes at offset 0
1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "job0"}}
{"return": {}}
{"error": {"class": "GenericError", "desc": "Node drv0 is in use"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "job0"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_CANCELLED", "data": {"device": "job0", "len": 1048576, "offset": 524288, "speed": 1, "type": "stream"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "job0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "job0"}}
{"return": {}}
*** done
diff --git a/tests/qemu-iotests/144 b/tests/qemu-iotests/144
index 00de3c33cf..4b915718cd 100755
--- a/tests/qemu-iotests/144
+++ b/tests/qemu-iotests/144
@@ -93,7 +93,7 @@ _send_qemu_cmd $h "{ 'execute': 'block-job-complete',
'arguments': {
'device': 'virtio0'
}
- }" "COMPLETED"
+ }" '"status": "null"'
echo
echo === Performing Live Snapshot 2 ===
diff --git a/tests/qemu-iotests/144.out b/tests/qemu-iotests/144.out
index 014b2817ee..55299201e4 100644
--- a/tests/qemu-iotests/144.out
+++ b/tests/qemu-iotests/144.out
@@ -12,10 +12,17 @@ Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/
=== Performing block-commit on active layer ===
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "virtio0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "virtio0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "virtio0"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "virtio0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}}
{"return": {}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "virtio0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "virtio0"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "virtio0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "virtio0"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "virtio0"}}
=== Performing Live Snapshot 2 ===
diff --git a/tests/qemu-iotests/156 b/tests/qemu-iotests/156
index e75dc4d743..0a9a09802e 100755
--- a/tests/qemu-iotests/156
+++ b/tests/qemu-iotests/156
@@ -119,7 +119,7 @@ _send_qemu_cmd $QEMU_HANDLE \
_send_qemu_cmd $QEMU_HANDLE \
'' \
- 'BLOCK_JOB_COMPLETED'
+ '"status": "null"'
# Remove the source images
rm -f "$TEST_IMG{,.backing,.overlay}"
diff --git a/tests/qemu-iotests/156.out b/tests/qemu-iotests/156.out
index f96a564c1d..34c057b626 100644
--- a/tests/qemu-iotests/156.out
+++ b/tests/qemu-iotests/156.out
@@ -12,13 +12,20 @@ wrote 131072/131072 bytes at offset 131072
128 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": ""}
Formatting 'TEST_DIR/t.IMGFMT.target.overlay', fmt=IMGFMT size=1048576 backing_file=TEST_DIR/t.IMGFMT.target
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "source"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "source"}}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "source"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "source", "len": 131072, "offset": 131072, "speed": 0, "type": "mirror"}}
wrote 65536/65536 bytes at offset 196608
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
{"return": ""}
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "waiting", "id": "source"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "pending", "id": "source"}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_COMPLETED", "data": {"device": "source", "len": 196608, "offset": 196608, "speed": 0, "type": "mirror"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "source"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "source"}}
read 65536/65536 bytes at offset 0
64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
diff --git a/tests/qemu-iotests/185 b/tests/qemu-iotests/185
index 9a2d317414..567ba67e48 100755
--- a/tests/qemu-iotests/185
+++ b/tests/qemu-iotests/185
@@ -118,8 +118,10 @@ _send_qemu_cmd $h \
# If we don't sleep here 'quit' command races with disk I/O
sleep 0.5
+# Ignore the JOB_STATUS_CHANGE events while shutting down the VM. Depending on
+# the timing, jobs may or may not transition through a paused state.
_send_qemu_cmd $h "{ 'execute': 'quit' }" "return"
-wait=1 _cleanup_qemu
+wait=1 _cleanup_qemu | grep -v 'JOB_STATUS_CHANGE'
echo
echo === Start active commit job and exit qemu ===
@@ -141,7 +143,7 @@ _send_qemu_cmd $h \
sleep 0.5
_send_qemu_cmd $h "{ 'execute': 'quit' }" "return"
-wait=1 _cleanup_qemu
+wait=1 _cleanup_qemu | grep -v 'JOB_STATUS_CHANGE'
echo
echo === Start mirror job and exit qemu ===
@@ -166,7 +168,7 @@ _send_qemu_cmd $h \
sleep 0.5
_send_qemu_cmd $h "{ 'execute': 'quit' }" "return"
-wait=1 _cleanup_qemu
+wait=1 _cleanup_qemu | grep -v 'JOB_STATUS_CHANGE'
echo
echo === Start backup job and exit qemu ===
@@ -190,7 +192,7 @@ _send_qemu_cmd $h \
sleep 0.5
_send_qemu_cmd $h "{ 'execute': 'quit' }" "return"
-wait=1 _cleanup_qemu
+wait=1 _cleanup_qemu | grep -v 'JOB_STATUS_CHANGE'
echo
echo === Start streaming job and exit qemu ===
@@ -211,7 +213,7 @@ _send_qemu_cmd $h \
sleep 0.5
_send_qemu_cmd $h "{ 'execute': 'quit' }" "return"
-wait=1 _cleanup_qemu
+wait=1 _cleanup_qemu | grep -v 'JOB_STATUS_CHANGE'
_check_test_img
diff --git a/tests/qemu-iotests/185.out b/tests/qemu-iotests/185.out
index 57eaf8d699..4e0ca0dffa 100644
--- a/tests/qemu-iotests/185.out
+++ b/tests/qemu-iotests/185.out
@@ -17,6 +17,8 @@ Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.q
=== Start commit job and exit qemu ===
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "disk"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "disk"}}
{"return": {}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
@@ -25,6 +27,8 @@ Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.q
=== Start active commit job and exit qemu ===
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "disk"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "disk"}}
{"return": {}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
@@ -34,6 +38,8 @@ Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 backing_file=TEST_DIR/t.q
{"return": {}}
Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "disk"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "disk"}}
{"return": {}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
@@ -43,6 +49,8 @@ Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 l
{"return": {}}
Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 lazy_refcounts=off refcount_bits=16
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "disk"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "disk"}}
{"return": {}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
@@ -51,6 +59,8 @@ Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 size=67108864 cluster_size=65536 l
=== Start streaming job and exit qemu ===
{"return": {}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "created", "id": "disk"}}
+{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "running", "id": "disk"}}
{"return": {}}
{"return": {}}
{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}}
diff --git a/tests/qemu-iotests/191 b/tests/qemu-iotests/191
index dfad6555e4..b3629ff6e8 100755
--- a/tests/qemu-iotests/191
+++ b/tests/qemu-iotests/191
@@ -83,7 +83,7 @@ _send_qemu_cmd $h \
'device': 'top',
'base':'$TEST_IMG.base',
'top': '$TEST_IMG.mid' } }" \
- "BLOCK_JOB_COMPLETED"
+ '"status": "null"'
_send_qemu_cmd $h "" "^}"
echo
@@ -131,7 +131,7 @@ _send_qemu_cmd $h \
'device': 'top',
'base':'$TEST_IMG.base',
'top': '$TEST_IMG.mid' } }" \
- "BLOCK_JOB_COMPLETED"
+ '"status": "null"'
_send_qemu_cmd $h "" "^}"
echo
diff --git a/tests/qemu-iotests/191.out b/tests/qemu-iotests/191.out
index 190c5f049a..31a0c7d4c4 100644
--- a/tests/qemu-iotests/191.out
+++ b/tests/qemu-iotests/191.out
@@ -16,6 +16,28 @@ wrote 65536/65536 bytes at offset 1048576
=== Perform commit job ===
{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "created",
+ "id": "commit0"
+ }
+}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "running",
+ "id": "commit0"
+ }
+}
+{
"return": {
}
}
@@ -24,6 +46,28 @@ wrote 65536/65536 bytes at offset 1048576
"seconds": TIMESTAMP,
"microseconds": TIMESTAMP
},
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "waiting",
+ "id": "commit0"
+ }
+}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "pending",
+ "id": "commit0"
+ }
+}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
"event": "BLOCK_JOB_COMPLETED",
"data": {
"device": "commit0",
@@ -33,6 +77,28 @@ wrote 65536/65536 bytes at offset 1048576
"type": "commit"
}
}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "concluded",
+ "id": "commit0"
+ }
+}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "null",
+ "id": "commit0"
+ }
+}
=== Check that both top and top2 point to base now ===
@@ -356,6 +422,28 @@ wrote 65536/65536 bytes at offset 1048576
=== Perform commit job ===
{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "created",
+ "id": "commit0"
+ }
+}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "running",
+ "id": "commit0"
+ }
+}
+{
"return": {
}
}
@@ -364,6 +452,28 @@ wrote 65536/65536 bytes at offset 1048576
"seconds": TIMESTAMP,
"microseconds": TIMESTAMP
},
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "waiting",
+ "id": "commit0"
+ }
+}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "pending",
+ "id": "commit0"
+ }
+}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
"event": "BLOCK_JOB_COMPLETED",
"data": {
"device": "commit0",
@@ -373,6 +483,28 @@ wrote 65536/65536 bytes at offset 1048576
"type": "commit"
}
}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "concluded",
+ "id": "commit0"
+ }
+}
+{
+ "timestamp": {
+ "seconds": TIMESTAMP,
+ "microseconds": TIMESTAMP
+ },
+ "event": "JOB_STATUS_CHANGE",
+ "data": {
+ "status": "null",
+ "id": "commit0"
+ }
+}
=== Check that both top and top2 point to base now ===
diff --git a/tests/qemu-iotests/iotests.py b/tests/qemu-iotests/iotests.py
index b25d48a91b..ed08e06e1f 100644
--- a/tests/qemu-iotests/iotests.py
+++ b/tests/qemu-iotests/iotests.py
@@ -464,6 +464,9 @@ class QMPTestCase(unittest.TestCase):
self.assert_qmp(event, 'data/device', drive)
result = event
cancelled = True
+ elif event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', drive)
+
self.assert_no_active_block_jobs()
return result
@@ -479,6 +482,8 @@ class QMPTestCase(unittest.TestCase):
self.assert_qmp(event, 'data/offset', event['data']['len'])
self.assert_no_active_block_jobs()
return event
+ elif event['event'] == 'JOB_STATUS_CHANGE':
+ self.assert_qmp(event, 'data/id', drive)
def wait_ready(self, drive='drive0'):
'''Wait until a block job BLOCK_JOB_READY event'''
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event Kevin Wolf
@ 2018-05-18 17:55 ` Eric Blake
2018-05-24 0:02 ` John Snow
1 sibling, 0 replies; 86+ messages in thread
From: Eric Blake @ 2018-05-18 17:55 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:21 AM, Kevin Wolf wrote:
> This adds a QMP event that is emitted whenever a job transitions from
> one status to another.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
> qapi/job.json | 14 ++++
> job.c | 10 +++
> @@ -157,6 +158,11 @@ static int job_txn_apply(JobTxn *txn, int fn(Job *), bool lock)
> return rc;
> }
>
> +static bool job_is_internal(Job *job)
> +{
> + return (job->id == NULL);
The () are not necessary. In fact, you could use !!job->id for less
typing, or even rely on the compiler to auto-convert a pointer into bool
by just 'return job->id' (although the latter feels a bit too terse to me).
Style is minor, so
Reviewed-by: Eric Blake <eblake@redhat.com>
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event Kevin Wolf
2018-05-18 17:55 ` Eric Blake
@ 2018-05-24 0:02 ` John Snow
2018-05-24 8:36 ` Kevin Wolf
1 sibling, 1 reply; 86+ messages in thread
From: John Snow @ 2018-05-24 0:02 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: qemu-devel, jcody, armbru, mreitz
On 05/18/2018 09:21 AM, Kevin Wolf wrote:
> This adds a QMP event that is emitted whenever a job transitions from
> one status to another.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
That's a lot of events, and a lot are redundant to what we already
transmitted under block jobs; it also has the effect of making internal
state changes explicit behavior that we're responsible for maintaining
for external clients.
Is that what we want here?
(I mean, the answer is probably "Yes" because you're here writing the
patch, but I was hoping to find your motivation.)
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event
2018-05-24 0:02 ` John Snow
@ 2018-05-24 8:36 ` Kevin Wolf
2018-05-24 17:36 ` John Snow
0 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-24 8:36 UTC (permalink / raw)
To: John Snow; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
Am 24.05.2018 um 02:02 hat John Snow geschrieben:
>
>
> On 05/18/2018 09:21 AM, Kevin Wolf wrote:
> > This adds a QMP event that is emitted whenever a job transitions from
> > one status to another.
> >
> > Signed-off-by: Kevin Wolf <kwolf@redhat.com>
>
> That's a lot of events, and a lot are redundant to what we already
> transmitted under block jobs; it also has the effect of making internal
> state changes explicit behavior that we're responsible for maintaining
> for external clients.
>
> Is that what we want here?
>
> (I mean, the answer is probably "Yes" because you're here writing the
> patch, but I was hoping to find your motivation.)
The state change is visible in query-jobs anyway, so allowing the client
to get events rather than polling query-jobs doesn't change the amount
of information that is exposed.
You're right that some of the events are redundant with existing block
job events, but that unavoidable because we can't send BLOCK_JOB_*
events for non-block jobs. And not sending JOB_* for block jobs would be
inconsistent.
The question that I was asking myself was just whether I'd literally
duplicate the existing events, just with s/BLOCK_JOB_/JOB_/, or whether
a single event with a status field can do. I think the latter is more
elegant (also because it can be implemented in a single place), even if
it is emitted a bit more often than strictly necessary.
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event
2018-05-24 8:36 ` Kevin Wolf
@ 2018-05-24 17:36 ` John Snow
2018-05-24 18:22 ` Eric Blake
0 siblings, 1 reply; 86+ messages in thread
From: John Snow @ 2018-05-24 17:36 UTC (permalink / raw)
To: Kevin Wolf; +Cc: qemu-block, qemu-devel, jcody, armbru, mreitz
On 05/24/2018 04:36 AM, Kevin Wolf wrote:
> Am 24.05.2018 um 02:02 hat John Snow geschrieben:
>>
>>
>> On 05/18/2018 09:21 AM, Kevin Wolf wrote:
>>> This adds a QMP event that is emitted whenever a job transitions from
>>> one status to another.
>>>
>>> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
>>
>> That's a lot of events, and a lot are redundant to what we already
>> transmitted under block jobs; it also has the effect of making internal
>> state changes explicit behavior that we're responsible for maintaining
>> for external clients.
>>
>> Is that what we want here?
>>
>> (I mean, the answer is probably "Yes" because you're here writing the
>> patch, but I was hoping to find your motivation.)
>
> The state change is visible in query-jobs anyway, so allowing the client
> to get events rather than polling query-jobs doesn't change the amount
> of information that is exposed.
>
With differing amounts of guarantees, though; for querying you have to
so happen to catch it and in practice you may never observe many of the
status transitions. It's not something that an external client could
rely on.
With events, we're guaranteeing you will be able to listen to all
possible state transitions. People may begin depending on them in ways
we hadn't anticipated just yet.
...Well, as long as the client is connected to the stream, which we
suggest may not always be possible with the presence of the concluded state.
> You're right that some of the events are redundant with existing block
> job events, but that unavoidable because we can't send BLOCK_JOB_*
> events for non-block jobs. And not sending JOB_* for block jobs would be
> inconsistent.
>
> The question that I was asking myself was just whether I'd literally
> duplicate the existing events, just with s/BLOCK_JOB_/JOB_/, or whether
> a single event with a status field can do. I think the latter is more
> elegant (also because it can be implemented in a single place), even if
> it is emitted a bit more often than strictly necessary.
>
Code-wise I agree that this is more elegant; just wondering if the
implications of the additional API guarantees were considered. This
means we have to be a bit more careful about how we restructure the
state machine in the future, I think.
It also makes the "NULL" state visible, which I didn't really intend...
It's probably fine, just a little quirky.
> Kevin
>
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event
2018-05-24 17:36 ` John Snow
@ 2018-05-24 18:22 ` Eric Blake
2018-05-24 18:32 ` John Snow
0 siblings, 1 reply; 86+ messages in thread
From: Eric Blake @ 2018-05-24 18:22 UTC (permalink / raw)
To: John Snow, Kevin Wolf; +Cc: jcody, mreitz, qemu-devel, qemu-block, armbru
On 05/24/2018 12:36 PM, John Snow wrote:
>>> On 05/18/2018 09:21 AM, Kevin Wolf wrote:
>>>> This adds a QMP event that is emitted whenever a job transitions from
>>>> one status to another.
>>>>
>>>> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
>>
>> The question that I was asking myself was just whether I'd literally
>> duplicate the existing events, just with s/BLOCK_JOB_/JOB_/, or whether
>> a single event with a status field can do. I think the latter is more
>> elegant (also because it can be implemented in a single place), even if
>> it is emitted a bit more often than strictly necessary.
>>
>
> Code-wise I agree that this is more elegant; just wondering if the
> implications of the additional API guarantees were considered. This
> means we have to be a bit more careful about how we restructure the
> state machine in the future, I think.
>
> It also makes the "NULL" state visible, which I didn't really intend...
> It's probably fine, just a little quirky.
Is it worth a special case to avoid emitting the event on transition to
the NULL state?
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event
2018-05-24 18:22 ` Eric Blake
@ 2018-05-24 18:32 ` John Snow
0 siblings, 0 replies; 86+ messages in thread
From: John Snow @ 2018-05-24 18:32 UTC (permalink / raw)
To: Eric Blake, Kevin Wolf; +Cc: jcody, mreitz, qemu-devel, qemu-block, armbru
On 05/24/2018 02:22 PM, Eric Blake wrote:
> On 05/24/2018 12:36 PM, John Snow wrote:
>
>>>> On 05/18/2018 09:21 AM, Kevin Wolf wrote:
>>>>> This adds a QMP event that is emitted whenever a job transitions from
>>>>> one status to another.
>>>>>
>>>>> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
>
>>>
>>> The question that I was asking myself was just whether I'd literally
>>> duplicate the existing events, just with s/BLOCK_JOB_/JOB_/, or whether
>>> a single event with a status field can do. I think the latter is more
>>> elegant (also because it can be implemented in a single place), even if
>>> it is emitted a bit more often than strictly necessary.
>>>
>>
>> Code-wise I agree that this is more elegant; just wondering if the
>> implications of the additional API guarantees were considered. This
>> means we have to be a bit more careful about how we restructure the
>> state machine in the future, I think.
>>
>> It also makes the "NULL" state visible, which I didn't really intend...
>> It's probably fine, just a little quirky.
>
> Is it worth a special case to avoid emitting the event on transition to
> the NULL state?
>
I would have done it, but it also doesn't necessarily hurt anything to
have it in here either.
Maybe it provides a benefit: I suppose it would definitely indicate to a
client that -- regardless of how they configured the job (with
auto-dismiss or not) that it serves as final record that the job has
*definitely absolutely gone* and can no longer be addressed.
It's just a strange state name for that purpose; I simply didn't
*intend* to expose it.
...OTOH, for early failures it seems like an obviously spurious message
that we don't need or want.
Well, glad I can give you precisely conflicting feelings on it and
arrive at no opinion. Good job everyone, see you tomorrow.
--js
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 36/40] job: Add lifecycle QMP commands
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (34 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 35/40] job: Add JOB_STATUS_CHANGE QMP event Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 18:12 ` Eric Blake
2018-05-23 23:56 ` John Snow
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 37/40] job: Add query-jobs QMP command Kevin Wolf
` (6 subsequent siblings)
42 siblings, 2 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This adds QMP commands that control the transition between states of the
job lifecycle.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
qapi/job.json | 99 +++++++++++++++++++++++++++++++++++++++++++
job-qmp.c | 134 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
MAINTAINERS | 1 +
Makefile.objs | 1 +
trace-events | 9 ++++
5 files changed, 244 insertions(+)
create mode 100644 job-qmp.c
diff --git a/qapi/job.json b/qapi/job.json
index 9fbdd0ccd9..b84dc6c820 100644
--- a/qapi/job.json
+++ b/qapi/job.json
@@ -106,3 +106,102 @@
{ 'event': 'JOB_STATUS_CHANGE',
'data': { 'id': 'str',
'status': 'JobStatus' } }
+
+##
+# @job-pause:
+#
+# Pause an active job.
+#
+# This command returns immediately after marking the active job for pausing.
+# Pausing an already paused job is an error.
+#
+# The job will pause as soon as possible, which means transitioning into the
+# PAUSED state if it was RUNNING, or into STANDBY if it was READY. The
+# corresponding JOB_STATUS_CHANGE event will be emitted.
+#
+# Cancelling a paused job automatically resumes it.
+#
+# @id: The job identifier.
+#
+# Since: 2.13
+##
+{ 'command': 'job-pause', 'data': { 'id': 'str' } }
+
+##
+# @job-resume:
+#
+# Resume a paused job.
+#
+# This command returns immediately after resuming a paused job. Resuming an
+# already running job is an error.
+#
+# @id : The job identifier.
+#
+# Since: 2.13
+##
+{ 'command': 'job-resume', 'data': { 'id': 'str' } }
+
+##
+# @job-cancel:
+#
+# Instruct an active background job to cancel at the next opportunity.
+# This command returns immediately after marking the active job for
+# cancellation.
+#
+# The job will cancel as soon as possible and then emit a JOB_STATUS_CHANGE
+# event. Usually, the status will change to ABORTING, but it is possible that
+# a job successfully completes (e.g. because it was almost done and there was
+# no opportunity to cancel earlier than completing the job) and transitions to
+# PENDING instead.
+#
+# @id: The job identifier.
+#
+# Since: 2.13
+##
+{ 'command': 'job-cancel', 'data': { 'id': 'str' } }
+
+
+##
+# @job-complete:
+#
+# Manually trigger completion of an active job in the READY state.
+#
+# @id: The job identifier.
+#
+# Since: 2.13
+##
+{ 'command': 'job-complete', 'data': { 'id': 'str' } }
+
+##
+# @job-dismiss:
+#
+# Deletes a job that is in the CONCLUDED state. This command only needs to be
+# run explicitly for jobs that don't have automatic dismiss enabled.
+#
+# This command will refuse to operate on any job that has not yet reached its
+# terminal state, JOB_STATUS_CONCLUDED. For jobs that make use of JOB_READY
+# event, job-cancel or job-complete will still need to be used as appropriate.
+#
+# @id: The job identifier.
+#
+# Since: 2.13
+##
+{ 'command': 'job-dismiss', 'data': { 'id': 'str' } }
+
+##
+# @job-finalize:
+#
+# Instructs all jobs in a transaction (or a single job if it is not part of any
+# transaction) to finalize any graph changes and do any necessary cleanup. This
+# command requires that all involved jobs are in the PENDING state.
+#
+# For jobs in a transaction, instructing one job to finalize will force
+# ALL jobs in the transaction to finalize, so it is only necessary to instruct
+# a single member job to finalize.
+#
+# @id: The identifier of any job in the transaction, or of a job that is not
+# part of any transaction.
+#
+# Since: 2.13
+##
+{ 'command': 'job-finalize', 'data': { 'id': 'str' } }
diff --git a/job-qmp.c b/job-qmp.c
new file mode 100644
index 0000000000..b2e18cfd9c
--- /dev/null
+++ b/job-qmp.c
@@ -0,0 +1,134 @@
+/*
+ * QMP interface for background jobs
+ *
+ * Copyright (c) 2011 IBM Corp.
+ * Copyright (c) 2012, 2018 Red Hat, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "qemu/osdep.h"
+#include "qemu-common.h"
+#include "qemu/job.h"
+#include "qapi/qapi-commands-job.h"
+#include "qapi/error.h"
+#include "trace-root.h"
+
+/* Get a job using its ID and acquire its AioContext */
+static Job *find_job(const char *id, AioContext **aio_context, Error **errp)
+{
+ Job *job;
+
+ *aio_context = NULL;
+
+ job = job_get(id);
+ if (!job) {
+ error_setg(errp, "Job not found");
+ return NULL;
+ }
+
+ *aio_context = job->aio_context;
+ aio_context_acquire(*aio_context);
+
+ return job;
+}
+
+void qmp_job_cancel(const char *id, Error **errp)
+{
+ AioContext *aio_context;
+ Job *job = find_job(id, &aio_context, errp);
+
+ if (!job) {
+ return;
+ }
+
+ trace_qmp_job_cancel(job);
+ job_user_cancel(job, true, errp);
+ aio_context_release(aio_context);
+}
+
+void qmp_job_pause(const char *id, Error **errp)
+{
+ AioContext *aio_context;
+ Job *job = find_job(id, &aio_context, errp);
+
+ if (!job) {
+ return;
+ }
+
+ trace_qmp_job_pause(job);
+ job_user_pause(job, errp);
+ aio_context_release(aio_context);
+}
+
+void qmp_job_resume(const char *id, Error **errp)
+{
+ AioContext *aio_context;
+ Job *job = find_job(id, &aio_context, errp);
+
+ if (!job) {
+ return;
+ }
+
+ trace_qmp_job_resume(job);
+ job_user_resume(job, errp);
+ aio_context_release(aio_context);
+}
+
+void qmp_job_complete(const char *id, Error **errp)
+{
+ AioContext *aio_context;
+ Job *job = find_job(id, &aio_context, errp);
+
+ if (!job) {
+ return;
+ }
+
+ trace_qmp_job_complete(job);
+ job_complete(job, errp);
+ aio_context_release(aio_context);
+}
+
+void qmp_job_finalize(const char *id, Error **errp)
+{
+ AioContext *aio_context;
+ Job *job = find_job(id, &aio_context, errp);
+
+ if (!job) {
+ return;
+ }
+
+ trace_qmp_job_finalize(job);
+ job_finalize(job, errp);
+ aio_context_release(aio_context);
+}
+
+void qmp_job_dismiss(const char *id, Error **errp)
+{
+ AioContext *aio_context;
+ Job *job = find_job(id, &aio_context, errp);
+
+ if (!job) {
+ return;
+ }
+
+ trace_qmp_job_dismiss(job);
+ job_dismiss(&job, errp);
+ aio_context_release(aio_context);
+}
diff --git a/MAINTAINERS b/MAINTAINERS
index 9fba3307d9..4215295f05 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1370,6 +1370,7 @@ S: Supported
F: blockjob.c
F: include/block/blockjob.h
F: job.c
+F: job-qmp.c
F: include/block/job.h
F: block/backup.c
F: block/commit.c
diff --git a/Makefile.objs b/Makefile.objs
index 3df8d58e49..c6c3554203 100644
--- a/Makefile.objs
+++ b/Makefile.objs
@@ -97,6 +97,7 @@ io-obj-y = io/
ifeq ($(CONFIG_SOFTMMU),y)
common-obj-y = blockdev.o blockdev-nbd.o block/
common-obj-y += bootdevice.o iothread.o
+common-obj-y += job-qmp.o
common-obj-y += net/
common-obj-y += qdev-monitor.o device-hotplug.o
common-obj-$(CONFIG_WIN32) += os-win32.o
diff --git a/trace-events b/trace-events
index ef7579a285..c445f54773 100644
--- a/trace-events
+++ b/trace-events
@@ -109,6 +109,15 @@ job_state_transition(void *job, int ret, const char *legal, const char *s0, con
job_apply_verb(void *job, const char *state, const char *verb, const char *legal) "job %p in state %s; applying verb %s (%s)"
job_completed(void *job, int ret, int jret) "job %p ret %d corrected ret %d"
+# job-qmp.c
+qmp_job_cancel(void *job) "job %p"
+qmp_job_pause(void *job) "job %p"
+qmp_job_resume(void *job) "job %p"
+qmp_job_complete(void *job) "job %p"
+qmp_job_finalize(void *job) "job %p"
+qmp_job_dismiss(void *job) "job %p"
+
+
### Guest events, keep at bottom
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 36/40] job: Add lifecycle QMP commands
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 36/40] job: Add lifecycle QMP commands Kevin Wolf
@ 2018-05-18 18:12 ` Eric Blake
2018-05-22 10:40 ` Kevin Wolf
2018-05-23 23:56 ` John Snow
1 sibling, 1 reply; 86+ messages in thread
From: Eric Blake @ 2018-05-18 18:12 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:21 AM, Kevin Wolf wrote:
> This adds QMP commands that control the transition between states of the
> job lifecycle.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
> qapi/job.json | 99 +++++++++++++++++++++++++++++++++++++++++++
> job-qmp.c | 134 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> MAINTAINERS | 1 +
> Makefile.objs | 1 +
> trace-events | 9 ++++
> 5 files changed, 244 insertions(+)
> create mode 100644 job-qmp.c
>
> +##
> +# @job-dismiss:
> +#
> +# Deletes a job that is in the CONCLUDED state. This command only needs to be
> +# run explicitly for jobs that don't have automatic dismiss enabled.
Did we decide whether it is valid to expect a job with automatic dismiss
enabled (old-style block jobs) to use the new job control commands? Or
would it be reasonable to require that 'job-dismiss' is an error on jobs
with auto-dismiss enabled (as in, if you're going to use new style jobs,
you are guaranteed to also have auto-dismiss false, because we don't
expose a way to change that flag in new-style jobs; and if you use old
style jobs, all management of the job should be done through the old
interfaces).
> +# This command will refuse to operate on any job that has not yet reached its
> +# terminal state, JOB_STATUS_CONCLUDED. For jobs that make use of JOB_READY
> +# event, job-cancel or job-complete will still need to be used as appropriate.
> +#
> +# @id: The job identifier.
> +#
> +# Since: 2.13
> +##
> +{ 'command': 'job-dismiss', 'data': { 'id': 'str' } }
> +
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 36/40] job: Add lifecycle QMP commands
2018-05-18 18:12 ` Eric Blake
@ 2018-05-22 10:40 ` Kevin Wolf
0 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-22 10:40 UTC (permalink / raw)
To: Eric Blake; +Cc: qemu-block, mreitz, jsnow, jcody, armbru, qemu-devel
Am 18.05.2018 um 20:12 hat Eric Blake geschrieben:
> On 05/18/2018 08:21 AM, Kevin Wolf wrote:
> > This adds QMP commands that control the transition between states of the
> > job lifecycle.
> >
> > Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> > ---
> > qapi/job.json | 99 +++++++++++++++++++++++++++++++++++++++++++
> > job-qmp.c | 134 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> > MAINTAINERS | 1 +
> > Makefile.objs | 1 +
> > trace-events | 9 ++++
> > 5 files changed, 244 insertions(+)
> > create mode 100644 job-qmp.c
> >
> > +##
> > +# @job-dismiss:
> > +#
> > +# Deletes a job that is in the CONCLUDED state. This command only needs to be
> > +# run explicitly for jobs that don't have automatic dismiss enabled.
>
> Did we decide whether it is valid to expect a job with automatic dismiss
> enabled (old-style block jobs) to use the new job control commands?
"job control commands" is a pretty broad term, so I'd say yes, it's
expected that you can use job-* commands on any job. For example,
job-pause/resume make perfect sense on a job with auto-dismiss=true.
> Or would it be reasonable to require that 'job-dismiss' is an error on
> jobs with auto-dismiss enabled
This, too, but only because it implicitly follows from the condition
specified above: You'll never catch a job in the CONCLUDED state when
you have auto-dismiss=true because it will immediately move on to NULL.
> (as in, if you're going to use new style jobs, you are guaranteed to
> also have auto-dismiss false, because we don't expose a way to change
> that flag in new-style jobs; and if you use old style jobs, all
> management of the job should be done through the old interfaces).
I'm not completely convinced that auto-dismiss=false is the only
"correct" setting. That might be the case for libvirt, but possibly not
for simple ad-hoc scripts with lower requirements.
We don't have any new jobs yet, so whether we expose the auto-* flags
there is a decision yet to be made.
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 36/40] job: Add lifecycle QMP commands
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 36/40] job: Add lifecycle QMP commands Kevin Wolf
2018-05-18 18:12 ` Eric Blake
@ 2018-05-23 23:56 ` John Snow
1 sibling, 0 replies; 86+ messages in thread
From: John Snow @ 2018-05-23 23:56 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: qemu-devel, jcody, armbru, mreitz
On 05/18/2018 09:21 AM, Kevin Wolf wrote:
> +{ 'command': 'job-complete', 'data': { 'id': 'str' } }
Do we have to name it this? I've always disliked how "complete" is used
as both a verb and an adjective in our code (hence why I used such
bizarre phrasings like "concluded" in my jobs patches) ...
I don't necessarily have a better suggestion, but now's the chance to
ask the question.
I guess you don't have a better suggestion either, otherwise you'd have
changed it.
--js
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 37/40] job: Add query-jobs QMP command
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (35 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 36/40] job: Add lifecycle QMP commands Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 18:14 ` Eric Blake
2018-05-18 18:22 ` Eric Blake
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 38/40] blockjob: Remove BlockJob.driver Kevin Wolf
` (5 subsequent siblings)
42 siblings, 2 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This adds a minimal query-jobs implementation that shouldn't pose many
design questions. It can later be extended to expose more information,
and especially job-specific information.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
qapi/job.json | 45 +++++++++++++++++++++++++++++++++++++++++++++
include/qemu/job.h | 3 +++
job-qmp.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
job.c | 2 +-
4 files changed, 103 insertions(+), 1 deletion(-)
diff --git a/qapi/job.json b/qapi/job.json
index b84dc6c820..932d5928e9 100644
--- a/qapi/job.json
+++ b/qapi/job.json
@@ -205,3 +205,48 @@
# Since: 2.13
##
{ 'command': 'job-finalize', 'data': { 'id': 'str' } }
+
+##
+# @JobInfo:
+#
+# Information about a job.
+#
+# @id: The job identifier
+#
+# @type: The kind of job that is being performed
+#
+# @status: Current job state/status
+#
+# @current-progress: Progress made until now. The unit is arbitrary and the
+# value can only meaningfully be used for the ratio of
+# @offset to @len. The value is monotonically increasing.
+#
+# @total-progress: Estimated @offset value at the completion of the job.
+# This value can arbitrarily change while the job is
+# running, in both directions.
+#
+# @error: If this field is present, the job failed; if it is
+# still missing in the CONCLUDED state, this indicates
+# successful completion.
+#
+# The value is a human-readable error message to describe
+# the reason for the job failure. It should not be parsed
+# by applications.
+#
+# Since: 2.13
+##
+{ 'struct': 'JobInfo',
+ 'data': { 'id': 'str', 'type': 'JobType', 'status': 'JobStatus',
+ 'current-progress': 'int', 'total-progress': 'int',
+ '*error': 'str' } }
+
+##
+# @query-jobs:
+#
+# Return information about jobs.
+#
+# Returns: a list with a @JobInfo for each active job
+#
+# Since: 2.13
+##
+{ 'command': 'query-jobs', 'returns': ['JobInfo'] }
diff --git a/include/qemu/job.h b/include/qemu/job.h
index 92d1d249fe..8c8badf75e 100644
--- a/include/qemu/job.h
+++ b/include/qemu/job.h
@@ -392,6 +392,9 @@ JobType job_type(const Job *job);
/** Returns the enum string for the JobType of a given Job. */
const char *job_type_str(const Job *job);
+/** Returns true if the job should not be visible to the management layer. */
+bool job_is_internal(Job *job);
+
/** Returns whether the job is scheduled for cancellation. */
bool job_is_cancelled(Job *job);
diff --git a/job-qmp.c b/job-qmp.c
index b2e18cfd9c..7f38f63336 100644
--- a/job-qmp.c
+++ b/job-qmp.c
@@ -132,3 +132,57 @@ void qmp_job_dismiss(const char *id, Error **errp)
job_dismiss(&job, errp);
aio_context_release(aio_context);
}
+
+static JobInfo *job_query_single(Job *job, Error **errp)
+{
+ JobInfo *info;
+ const char *errmsg = NULL;
+
+ assert(!job_is_internal(job));
+
+ if (job->ret < 0) {
+ errmsg = strerror(-job->ret);
+ }
+
+ info = g_new(JobInfo, 1);
+ *info = (JobInfo) {
+ .id = g_strdup(job->id),
+ .type = job_type(job),
+ .status = job->status,
+ .current_progress = job->progress_current,
+ .total_progress = job->progress_total,
+ .has_error = !!errmsg,
+ .error = g_strdup(errmsg),
+ };
+
+ return info;
+}
+
+JobInfoList *qmp_query_jobs(Error **errp)
+{
+ JobInfoList *head = NULL, **p_next = &head;
+ Job *job;
+
+ for (job = job_next(NULL); job; job = job_next(job)) {
+ JobInfoList *elem;
+ AioContext *aio_context;
+
+ if (job_is_internal(job)) {
+ continue;
+ }
+ elem = g_new0(JobInfoList, 1);
+ aio_context = job->aio_context;
+ aio_context_acquire(aio_context);
+ elem->value = job_query_single(job, errp);
+ aio_context_release(aio_context);
+ if (!elem->value) {
+ g_free(elem);
+ qapi_free_JobInfoList(head);
+ return NULL;
+ }
+ *p_next = elem;
+ p_next = &elem->next;
+ }
+
+ return head;
+}
diff --git a/job.c b/job.c
index 8e106e5d15..37f32266ba 100644
--- a/job.c
+++ b/job.c
@@ -158,7 +158,7 @@ static int job_txn_apply(JobTxn *txn, int fn(Job *), bool lock)
return rc;
}
-static bool job_is_internal(Job *job)
+bool job_is_internal(Job *job)
{
return (job->id == NULL);
}
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 37/40] job: Add query-jobs QMP command
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 37/40] job: Add query-jobs QMP command Kevin Wolf
@ 2018-05-18 18:14 ` Eric Blake
2018-05-18 18:22 ` Eric Blake
1 sibling, 0 replies; 86+ messages in thread
From: Eric Blake @ 2018-05-18 18:14 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:21 AM, Kevin Wolf wrote:
> This adds a minimal query-jobs implementation that shouldn't pose many
> design questions. It can later be extended to expose more information,
> and especially job-specific information.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
> qapi/job.json | 45 +++++++++++++++++++++++++++++++++++++++++++++
> include/qemu/job.h | 3 +++
> job-qmp.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
> job.c | 2 +-
> 4 files changed, 103 insertions(+), 1 deletion(-)
> +#
> +# @current-progress: Progress made until now. The unit is arbitrary and the
> +# value can only meaningfully be used for the ratio of
> +# @offset to @len. The value is monotonically increasing.
s/@offset to @len/@current-progress to @total-progress/
> +#
> +# @total-progress: Estimated @offset value at the completion of the job.
s/@offset/@current-progress/
> +# This value can arbitrarily change while the job is
> +# running, in both directions.
> +#
> +# @error: If this field is present, the job failed; if it is
> +# still missing in the CONCLUDED state, this indicates
> +# successful completion.
> +#
> +# The value is a human-readable error message to describe
> +# the reason for the job failure. It should not be parsed
> +# by applications.
> +#
> +# Since: 2.13
> +##
> +{ 'struct': 'JobInfo',
> + 'data': { 'id': 'str', 'type': 'JobType', 'status': 'JobStatus',
> + 'current-progress': 'int', 'total-progress': 'int',
> + '*error': 'str' } }
> +
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 37/40] job: Add query-jobs QMP command
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 37/40] job: Add query-jobs QMP command Kevin Wolf
2018-05-18 18:14 ` Eric Blake
@ 2018-05-18 18:22 ` Eric Blake
2018-05-22 10:44 ` Kevin Wolf
1 sibling, 1 reply; 86+ messages in thread
From: Eric Blake @ 2018-05-18 18:22 UTC (permalink / raw)
To: Kevin Wolf, qemu-block; +Cc: mreitz, jsnow, jcody, armbru, qemu-devel
On 05/18/2018 08:21 AM, Kevin Wolf wrote:
> This adds a minimal query-jobs implementation that shouldn't pose many
> design questions. It can later be extended to expose more information,
> and especially job-specific information.
>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
> qapi/job.json | 45 +++++++++++++++++++++++++++++++++++++++++++++
> include/qemu/job.h | 3 +++
> job-qmp.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
> job.c | 2 +-
> 4 files changed, 103 insertions(+), 1 deletion(-)
>
> +##
> +# @JobInfo:
> +#
> +# Information about a job.
> +#
> +# @id: The job identifier
> +##
> +{ 'struct': 'JobInfo',
> + 'data': { 'id': 'str', 'type': 'JobType', 'status': 'JobStatus',
> + 'current-progress': 'int', 'total-progress': 'int',
> + '*error': 'str' } }
Is it worth exposing whether a job is auto-finalize and auto-complete?
Goes back to the issue of whether clients of the new job API would ever
want/need to rely on the auto- features; while clients of the old
blockjob API that get the auto- features by default will never be
calling the new query-jobs command.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 37/40] job: Add query-jobs QMP command
2018-05-18 18:22 ` Eric Blake
@ 2018-05-22 10:44 ` Kevin Wolf
0 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-22 10:44 UTC (permalink / raw)
To: Eric Blake; +Cc: qemu-block, mreitz, jsnow, jcody, armbru, qemu-devel
Am 18.05.2018 um 20:22 hat Eric Blake geschrieben:
> On 05/18/2018 08:21 AM, Kevin Wolf wrote:
> > This adds a minimal query-jobs implementation that shouldn't pose many
> > design questions. It can later be extended to expose more information,
> > and especially job-specific information.
> >
> > Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> > ---
> > qapi/job.json | 45 +++++++++++++++++++++++++++++++++++++++++++++
> > include/qemu/job.h | 3 +++
> > job-qmp.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
> > job.c | 2 +-
> > 4 files changed, 103 insertions(+), 1 deletion(-)
> >
>
> > +##
> > +# @JobInfo:
> > +#
> > +# Information about a job.
> > +#
> > +# @id: The job identifier
>
> > +##
> > +{ 'struct': 'JobInfo',
> > + 'data': { 'id': 'str', 'type': 'JobType', 'status': 'JobStatus',
> > + 'current-progress': 'int', 'total-progress': 'int',
> > + '*error': 'str' } }
>
> Is it worth exposing whether a job is auto-finalize and auto-complete? Goes
> back to the issue of whether clients of the new job API would ever want/need
> to rely on the auto- features; while clients of the old blockjob API that
> get the auto- features by default will never be calling the new query-jobs
> command.
There are probably more fields that could be exposed. For most of them,
it's not obvious whether we want to expose them, so I went for the
minimal useful information here. We can always add more information to a
query command; but we can't take information away later if it turns out
that exposing it was a bad idea.
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread
* [Qemu-devel] [PATCH v2 38/40] blockjob: Remove BlockJob.driver
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (36 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 37/40] job: Add query-jobs QMP command Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 39/40] iotests: Move qmp_to_opts() to VM Kevin Wolf
` (4 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
BlockJob.driver is redundant with Job.driver and only used in very few
places any more. Remove it.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
include/block/blockjob.h | 3 ---
blockjob.c | 17 ++++++++++-------
2 files changed, 10 insertions(+), 10 deletions(-)
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 3021d11126..32c00b7dc0 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -43,9 +43,6 @@ typedef struct BlockJob {
/** Data belonging to the generic Job infrastructure */
Job job;
- /** The job type, including the job vtable. */
- const BlockJobDriver *driver;
-
/** The block device on which the job is operating. */
BlockBackend *blk;
diff --git a/blockjob.c b/blockjob.c
index 5c8ff6f846..0306533a2e 100644
--- a/blockjob.c
+++ b/blockjob.c
@@ -103,10 +103,12 @@ static void block_job_attached_aio_context(AioContext *new_context,
void *opaque)
{
BlockJob *job = opaque;
+ const JobDriver *drv = job->job.driver;
+ BlockJobDriver *bjdrv = container_of(drv, BlockJobDriver, job_driver);
job->job.aio_context = new_context;
- if (job->driver->attached_aio_context) {
- job->driver->attached_aio_context(job, new_context);
+ if (bjdrv->attached_aio_context) {
+ bjdrv->attached_aio_context(job, new_context);
}
job_resume(&job->job);
@@ -115,10 +117,12 @@ static void block_job_attached_aio_context(AioContext *new_context,
void block_job_drain(Job *job)
{
BlockJob *bjob = container_of(job, BlockJob, job);
+ const JobDriver *drv = job->driver;
+ BlockJobDriver *bjdrv = container_of(drv, BlockJobDriver, job_driver);
blk_drain(bjob->blk);
- if (bjob->driver->drain) {
- bjob->driver->drain(bjob);
+ if (bjdrv->drain) {
+ bjdrv->drain(bjob);
}
}
@@ -201,7 +205,7 @@ bool block_job_is_internal(BlockJob *job)
const BlockJobDriver *block_job_driver(BlockJob *job)
{
- return job->driver;
+ return container_of(job->job.driver, BlockJobDriver, job_driver);
}
/* Assumes the job_mutex is held */
@@ -386,8 +390,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
assert(job->job.driver->user_resume == &block_job_user_resume);
assert(job->job.driver->drain == &block_job_drain);
- job->driver = driver;
- job->blk = blk;
+ job->blk = blk;
job->finalize_cancelled_notifier.notify = block_job_event_cancelled;
job->finalize_completed_notifier.notify = block_job_event_completed;
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 39/40] iotests: Move qmp_to_opts() to VM
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (37 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 38/40] blockjob: Remove BlockJob.driver Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 40/40] qemu-iotests: Test job-* with block jobs Kevin Wolf
` (3 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
qmp_to_opts() used to be a method of QMPTestCase, but recently we
started to add more Python test cases that don't make use of
QMPTestCase. In order to make the method usable there, move it to VM.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
---
tests/qemu-iotests/041 | 6 +++---
tests/qemu-iotests/155 | 2 +-
tests/qemu-iotests/iotests.py | 45 ++++++++++++++++++++++---------------------
3 files changed, 27 insertions(+), 26 deletions(-)
diff --git a/tests/qemu-iotests/041 b/tests/qemu-iotests/041
index e94587950c..c20ac7da87 100755
--- a/tests/qemu-iotests/041
+++ b/tests/qemu-iotests/041
@@ -1030,9 +1030,9 @@ class TestOrphanedSource(iotests.QMPTestCase):
'read-only': 'on' }
self.vm = iotests.VM()
- self.vm.add_blockdev(self.qmp_to_opts(blk0))
- self.vm.add_blockdev(self.qmp_to_opts(blk1))
- self.vm.add_blockdev(self.qmp_to_opts(blk2))
+ self.vm.add_blockdev(self.vm.qmp_to_opts(blk0))
+ self.vm.add_blockdev(self.vm.qmp_to_opts(blk1))
+ self.vm.add_blockdev(self.vm.qmp_to_opts(blk2))
self.vm.launch()
def tearDown(self):
diff --git a/tests/qemu-iotests/155 b/tests/qemu-iotests/155
index 42dae04c83..63a5b5e2c0 100755
--- a/tests/qemu-iotests/155
+++ b/tests/qemu-iotests/155
@@ -63,7 +63,7 @@ class BaseClass(iotests.QMPTestCase):
'driver': iotests.imgfmt,
'file': {'driver': 'file',
'filename': source_img}}
- self.vm.add_blockdev(self.qmp_to_opts(blockdev))
+ self.vm.add_blockdev(self.vm.qmp_to_opts(blockdev))
self.vm.add_device('virtio-blk,id=qdev0,drive=source')
self.vm.launch()
diff --git a/tests/qemu-iotests/iotests.py b/tests/qemu-iotests/iotests.py
index ed08e06e1f..28159d837a 100644
--- a/tests/qemu-iotests/iotests.py
+++ b/tests/qemu-iotests/iotests.py
@@ -363,6 +363,27 @@ class VM(qtest.QEMUQtestMachine):
return self.qmp('human-monitor-command',
command_line='qemu-io %s "%s"' % (drive, cmd))
+ def flatten_qmp_object(self, obj, output=None, basestr=''):
+ if output is None:
+ output = dict()
+ if isinstance(obj, list):
+ for i in range(len(obj)):
+ self.flatten_qmp_object(obj[i], output, basestr + str(i) + '.')
+ elif isinstance(obj, dict):
+ for key in obj:
+ self.flatten_qmp_object(obj[key], output, basestr + key + '.')
+ else:
+ output[basestr[:-1]] = obj # Strip trailing '.'
+ return output
+
+ def qmp_to_opts(self, obj):
+ obj = self.flatten_qmp_object(obj)
+ output_list = list()
+ for key in obj:
+ output_list += [key + '=' + obj[key]]
+ return ','.join(output_list)
+
+
index_re = re.compile(r'([^\[]+)\[([^\]]+)\]')
@@ -390,26 +411,6 @@ class QMPTestCase(unittest.TestCase):
self.fail('invalid index "%s" in path "%s" in "%s"' % (idx, path, str(d)))
return d
- def flatten_qmp_object(self, obj, output=None, basestr=''):
- if output is None:
- output = dict()
- if isinstance(obj, list):
- for i in range(len(obj)):
- self.flatten_qmp_object(obj[i], output, basestr + str(i) + '.')
- elif isinstance(obj, dict):
- for key in obj:
- self.flatten_qmp_object(obj[key], output, basestr + key + '.')
- else:
- output[basestr[:-1]] = obj # Strip trailing '.'
- return output
-
- def qmp_to_opts(self, obj):
- obj = self.flatten_qmp_object(obj)
- output_list = list()
- for key in obj:
- output_list += [key + '=' + obj[key]]
- return ','.join(output_list)
-
def assert_qmp_absent(self, d, path):
try:
result = self.dictpath(d, path)
@@ -444,8 +445,8 @@ class QMPTestCase(unittest.TestCase):
'''Asserts that the given filename is a json: filename and that its
content is equal to the given reference object'''
self.assertEqual(json_filename[:5], 'json:')
- self.assertEqual(self.flatten_qmp_object(json.loads(json_filename[5:])),
- self.flatten_qmp_object(reference))
+ self.assertEqual(self.vm.flatten_qmp_object(json.loads(json_filename[5:])),
+ self.vm.flatten_qmp_object(reference))
def cancel_and_wait(self, drive='drive0', force=False, resume=False):
'''Cancel a block job and wait for it to finish, returning the event'''
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* [Qemu-devel] [PATCH v2 40/40] qemu-iotests: Test job-* with block jobs
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (38 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 39/40] iotests: Move qmp_to_opts() to VM Kevin Wolf
@ 2018-05-18 13:21 ` Kevin Wolf
2018-05-18 14:05 ` [Qemu-devel] [PATCH v2 00/40] Generic background jobs no-reply
` (2 subsequent siblings)
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-18 13:21 UTC (permalink / raw)
To: qemu-block; +Cc: kwolf, mreitz, jsnow, eblake, jcody, armbru, qemu-devel
This adds a test case that tests the new job-* QMP commands with
mirror and backup block jobs.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
tests/qemu-iotests/219 | 209 +++++++++++++++++++++++++++++
tests/qemu-iotests/219.out | 327 +++++++++++++++++++++++++++++++++++++++++++++
tests/qemu-iotests/group | 1 +
3 files changed, 537 insertions(+)
create mode 100755 tests/qemu-iotests/219
create mode 100644 tests/qemu-iotests/219.out
diff --git a/tests/qemu-iotests/219 b/tests/qemu-iotests/219
new file mode 100755
index 0000000000..898a26eef0
--- /dev/null
+++ b/tests/qemu-iotests/219
@@ -0,0 +1,209 @@
+#!/usr/bin/env python
+#
+# Copyright (C) 2018 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: Kevin Wolf <kwolf@redhat.com>
+#
+# Check using the job-* QMP commands with block jobs
+
+import iotests
+
+iotests.verify_image_format(supported_fmts=['qcow2'])
+
+def pause_wait(vm, job_id):
+ with iotests.Timeout(3, "Timeout waiting for job to pause"):
+ while True:
+ result = vm.qmp('query-jobs')
+ for job in result['return']:
+ if job['id'] == job_id and job['status'] in ['paused', 'standby']:
+ return job
+
+# Test that block-job-pause/resume and job-pause/resume can be mixed
+def test_pause_resume(vm):
+ for pause_cmd, pause_arg in [('block-job-pause', 'device'),
+ ('job-pause', 'id')]:
+ for resume_cmd, resume_arg in [('block-job-resume', 'device'),
+ ('job-resume', 'id')]:
+ iotests.log('=== Testing %s/%s ===' % (pause_cmd, resume_cmd))
+
+ iotests.log(vm.qmp(pause_cmd, **{pause_arg: 'job0'}))
+ pause_wait(vm, 'job0')
+ iotests.log(iotests.filter_qmp_event(vm.event_wait('JOB_STATUS_CHANGE')))
+ iotests.log(vm.qmp('query-jobs'))
+
+ iotests.log(vm.qmp(resume_cmd, **{resume_arg: 'job0'}))
+ iotests.log(iotests.filter_qmp_event(vm.event_wait('JOB_STATUS_CHANGE')))
+ iotests.log(vm.qmp('query-jobs'))
+
+def test_job_lifecycle(vm, job, job_args, has_ready=False):
+ iotests.log('')
+ iotests.log('')
+ iotests.log('Starting block job: %s (auto-finalize: %s; auto-dismiss: %s)' %
+ (job,
+ job_args.get('auto-finalize', True),
+ job_args.get('auto-dismiss', True)))
+ iotests.log(vm.qmp(job, job_id='job0', **job_args))
+
+ # Depending on the storage, the first request may or may not have completed
+ # yet, so filter out the progress. Later query-job calls don't need the
+ # filtering because the progress is made deterministic by the block job
+ # speed
+ result = vm.qmp('query-jobs')
+ for j in result['return']:
+ del j['current-progress']
+ iotests.log(result)
+
+ # undefined -> created -> running
+ iotests.log(iotests.filter_qmp_event(vm.event_wait('JOB_STATUS_CHANGE')))
+ iotests.log(iotests.filter_qmp_event(vm.event_wait('JOB_STATUS_CHANGE')))
+
+ # RUNNING state:
+ # pause/resume should work, complete/finalize/dismiss should error out
+ iotests.log('')
+ iotests.log('Pause/resume in RUNNING')
+ test_pause_resume(vm)
+
+ iotests.log(vm.qmp('job-complete', id='job0'))
+ iotests.log(vm.qmp('job-finalize', id='job0'))
+ iotests.log(vm.qmp('job-dismiss', id='job0'))
+
+ iotests.log(vm.qmp('block-job-complete', device='job0'))
+ iotests.log(vm.qmp('block-job-finalize', id='job0'))
+ iotests.log(vm.qmp('block-job-dismiss', id='job0'))
+
+ # Let the job complete (or transition to READY if it supports that)
+ iotests.log(vm.qmp('block-job-set-speed', device='job0', speed=0))
+ if has_ready:
+ iotests.log('')
+ iotests.log('Waiting for READY state...')
+ vm.event_wait('BLOCK_JOB_READY')
+ iotests.log(iotests.filter_qmp_event(vm.event_wait('JOB_STATUS_CHANGE')))
+ iotests.log(vm.qmp('query-jobs'))
+
+ # READY state:
+ # pause/resume/complete should work, finalize/dismiss should error out
+ iotests.log('')
+ iotests.log('Pause/resume in READY')
+ test_pause_resume(vm)
+
+ iotests.log(vm.qmp('job-finalize', id='job0'))
+ iotests.log(vm.qmp('job-dismiss', id='job0'))
+
+ iotests.log(vm.qmp('block-job-finalize', id='job0'))
+ iotests.log(vm.qmp('block-job-dismiss', id='job0'))
+
+ # Transition to WAITING
+ iotests.log(vm.qmp('job-complete', id='job0'))
+
+ # Move to WAITING and PENDING state
+ iotests.log('')
+ iotests.log('Waiting for PENDING state...')
+ iotests.log(iotests.filter_qmp_event(vm.event_wait('JOB_STATUS_CHANGE')))
+ iotests.log(iotests.filter_qmp_event(vm.event_wait('JOB_STATUS_CHANGE')))
+
+ if not job_args.get('auto-finalize', True):
+ # PENDING state:
+ # finalize should work, pause/complete/dismiss should error out
+ iotests.log(vm.qmp('query-jobs'))
+
+ iotests.log(vm.qmp('job-pause', id='job0'))
+ iotests.log(vm.qmp('job-complete', id='job0'))
+ iotests.log(vm.qmp('job-dismiss', id='job0'))
+
+ iotests.log(vm.qmp('block-job-pause', device='job0'))
+ iotests.log(vm.qmp('block-job-complete', device='job0'))
+ iotests.log(vm.qmp('block-job-dismiss', id='job0'))
+
+ # Transition to CONCLUDED
+ iotests.log(vm.qmp('job-finalize', id='job0'))
+
+
+ # Move to CONCLUDED state
+ iotests.log(iotests.filter_qmp_event(vm.event_wait('JOB_STATUS_CHANGE')))
+
+ if not job_args.get('auto-dismiss', True):
+ # CONCLUDED state:
+ # dismiss should work, pause/complete/finalize should error out
+ iotests.log(vm.qmp('query-jobs'))
+
+ iotests.log(vm.qmp('job-pause', id='job0'))
+ iotests.log(vm.qmp('job-complete', id='job0'))
+ iotests.log(vm.qmp('job-finalize', id='job0'))
+
+ iotests.log(vm.qmp('block-job-pause', device='job0'))
+ iotests.log(vm.qmp('block-job-complete', device='job0'))
+ iotests.log(vm.qmp('block-job-finalize', id='job0'))
+
+ # Transition to NULL
+ iotests.log(vm.qmp('job-dismiss', id='job0'))
+
+ # Move to NULL state
+ iotests.log(iotests.filter_qmp_event(vm.event_wait('JOB_STATUS_CHANGE')))
+ iotests.log(vm.qmp('query-jobs'))
+
+
+with iotests.FilePath('disk.img') as disk_path, \
+ iotests.FilePath('copy.img') as copy_path, \
+ iotests.VM() as vm:
+
+ img_size = '4M'
+ iotests.qemu_img_create('-f', iotests.imgfmt, disk_path, img_size)
+ iotests.qemu_io('-c', 'write 0 %s' % (img_size),
+ '-f', iotests.imgfmt, disk_path)
+
+ iotests.log('Launching VM...')
+ vm.add_blockdev(vm.qmp_to_opts({
+ 'driver': iotests.imgfmt,
+ 'node-name': 'drive0-node',
+ 'file': {
+ 'driver': 'file',
+ 'filename': disk_path,
+ },
+ }))
+ vm.launch()
+
+ # In order to keep things deterministic (especially progress in query-job,
+ # but related to this also automatic state transitions like job
+ # completion), but still get pause points often enough to avoid making this
+ # test very slow, it's important to have the right ratio between speed and
+ # buf_size.
+ #
+ # For backup, buf_size is hard-coded to the source image cluster size (64k),
+ # so we'll pick the same for mirror. The slice time, i.e. the granularity
+ # of the rate limiting is 100ms. With a speed of 256k per second, we can
+ # get four pause points per second. This gives us 250ms per iteration,
+ # which should be enough to stay deterministic.
+
+ test_job_lifecycle(vm, 'drive-mirror', has_ready=True, job_args={
+ 'device': 'drive0-node',
+ 'target': copy_path,
+ 'sync': 'full',
+ 'speed': 262144,
+ 'buf_size': 65536,
+ })
+
+ for auto_finalize in [True, False]:
+ for auto_dismiss in [True, False]:
+ test_job_lifecycle(vm, 'drive-backup', job_args={
+ 'device': 'drive0-node',
+ 'target': copy_path,
+ 'sync': 'full',
+ 'speed': 262144,
+ 'auto-finalize': auto_finalize,
+ 'auto-dismiss': auto_dismiss,
+ })
+
+ vm.shutdown()
diff --git a/tests/qemu-iotests/219.out b/tests/qemu-iotests/219.out
new file mode 100644
index 0000000000..346801b655
--- /dev/null
+++ b/tests/qemu-iotests/219.out
@@ -0,0 +1,327 @@
+Launching VM...
+
+
+Starting block job: drive-mirror (auto-finalize: True; auto-dismiss: True)
+{u'return': {}}
+{u'return': [{u'status': u'running', u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'created', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+
+Pause/resume in RUNNING
+=== Testing block-job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 65536, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+=== Testing block-job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+=== Testing job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+=== Testing job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 327680, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'return': {}}
+
+Waiting for READY state...
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'ready', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'ready', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+
+Pause/resume in READY
+=== Testing block-job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'standby', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'standby', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'ready', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'ready', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+=== Testing block-job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'standby', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'standby', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'ready', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'ready', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+=== Testing job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'standby', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'standby', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'ready', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'ready', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+=== Testing job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'standby', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'standby', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'ready', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'ready', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'mirror'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'ready' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'ready' cannot accept command verb 'dismiss'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'ready' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'ready' cannot accept command verb 'dismiss'"}}
+{u'return': {}}
+
+Waiting for PENDING state...
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'waiting', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'pending', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'concluded', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'null', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': []}
+
+
+Starting block job: drive-backup (auto-finalize: True; auto-dismiss: True)
+{u'return': {}}
+{u'return': [{u'status': u'running', u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'created', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+
+Pause/resume in RUNNING
+=== Testing block-job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 65536, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing block-job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 327680, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'return': {}}
+
+Waiting for PENDING state...
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'waiting', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'pending', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'concluded', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'null', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': []}
+
+
+Starting block job: drive-backup (auto-finalize: True; auto-dismiss: False)
+{u'return': {}}
+{u'return': [{u'status': u'running', u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'created', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+
+Pause/resume in RUNNING
+=== Testing block-job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 65536, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing block-job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 327680, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'return': {}}
+
+Waiting for PENDING state...
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'waiting', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'pending', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'concluded', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'concluded', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'pause'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'pause'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'finalize'"}}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'null', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': []}
+
+
+Starting block job: drive-backup (auto-finalize: False; auto-dismiss: True)
+{u'return': {}}
+{u'return': [{u'status': u'running', u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'created', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+
+Pause/resume in RUNNING
+=== Testing block-job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 65536, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing block-job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 327680, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'return': {}}
+
+Waiting for PENDING state...
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'waiting', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'pending', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'pending', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'pause'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'dismiss'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'pause'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'dismiss'"}}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'concluded', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'null', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': []}
+
+
+Starting block job: drive-backup (auto-finalize: False; auto-dismiss: False)
+{u'return': {}}
+{u'return': [{u'status': u'running', u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'created', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+
+Pause/resume in RUNNING
+=== Testing block-job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 65536, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing block-job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 131072, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing job-pause/block-job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 196608, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+=== Testing job-pause/job-resume ===
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'paused', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'paused', u'current-progress': 262144, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'running', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'running', u'current-progress': 327680, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'running' cannot accept command verb 'dismiss'"}}
+{u'return': {}}
+
+Waiting for PENDING state...
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'waiting', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'pending', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'pending', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'pause'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'dismiss'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'pause'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'pending' cannot accept command verb 'dismiss'"}}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'concluded', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': [{u'status': u'concluded', u'current-progress': 4194304, u'total-progress': 4194304, u'id': u'job0', u'type': u'backup'}]}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'pause'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'finalize'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'pause'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'complete'"}}
+{u'error': {u'class': u'GenericError', u'desc': u"Job 'job0' in state 'concluded' cannot accept command verb 'finalize'"}}
+{u'return': {}}
+{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'null', u'id': u'job0'}, u'event': u'JOB_STATUS_CHANGE'}
+{u'return': []}
diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
index cc8cd8cc8e..6f1e07dc6e 100644
--- a/tests/qemu-iotests/group
+++ b/tests/qemu-iotests/group
@@ -216,3 +216,4 @@
215 rw auto quick
216 rw auto quick
218 rw auto quick
+219 rw auto
--
2.13.6
^ permalink raw reply related [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 00/40] Generic background jobs
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (39 preceding siblings ...)
2018-05-18 13:21 ` [Qemu-devel] [PATCH v2 40/40] qemu-iotests: Test job-* with block jobs Kevin Wolf
@ 2018-05-18 14:05 ` no-reply
2018-05-18 18:41 ` Dr. David Alan Gilbert
2018-05-23 12:31 ` Kevin Wolf
42 siblings, 0 replies; 86+ messages in thread
From: no-reply @ 2018-05-18 14:05 UTC (permalink / raw)
To: kwolf; +Cc: famz, qemu-block, qemu-devel, jsnow, jcody, armbru, mreitz
Hi,
This series seems to have some coding style problems. See output below for
more information:
Type: series
Message-id: 20180518132114.4070-1-kwolf@redhat.com
Subject: [Qemu-devel] [PATCH v2 00/40] Generic background jobs
=== TEST SCRIPT BEGIN ===
#!/bin/bash
BASE=base
n=1
total=$(git log --oneline $BASE.. | wc -l)
failed=0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
commits="$(git log --format=%H --reverse $BASE..)"
for c in $commits; do
echo "Checking PATCH $n/$total: $(git log -n 1 --format=%s $c)..."
if ! git show $c --format=email | ./scripts/checkpatch.pl --mailback -; then
failed=1
echo
fi
n=$((n+1))
done
exit $failed
=== TEST SCRIPT END ===
Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
From https://github.com/patchew-project/qemu
t [tag update] patchew/20180517092513.735-1-kraxel@redhat.com -> patchew/20180517092513.735-1-kraxel@redhat.com
* [new tag] patchew/20180518132114.4070-1-kwolf@redhat.com -> patchew/20180518132114.4070-1-kwolf@redhat.com
Switched to a new branch 'test'
be01aa198a qemu-iotests: Test job-* with block jobs
62233b758c iotests: Move qmp_to_opts() to VM
3f946bc442 blockjob: Remove BlockJob.driver
cae6373cbd job: Add query-jobs QMP command
9bf0565e0c job: Add lifecycle QMP commands
f7363a69b4 job: Add JOB_STATUS_CHANGE QMP event
5a63efe4e4 job: Introduce qapi/job.json
2a5e7375fa job: Move progress fields to Job
2ae08c1d06 job: Add job_transition_to_ready()
9d76be11bd job: Add job_is_ready()
9b51cc2a7e job: Add job_dismiss()
555a61a69f job: Add job_yield()
798978bcb1 block: Cancel job in bdrv_close_all() callers
4dcb71e4e5 job: Move completion and cancellation to Job
ffad8fad4b job: Move transactions to Job
0edc41ca48 job: Switch transactions to JobTxn
a1073023f2 job: Move job_finish_sync() to Job
6ec521dbdb job: Move .complete callback to Job
9660a731ab job: Add job_drain()
1958d0ac38 job: Convert block_job_cancel_async() to Job
dc1a8d70b8 job: Move single job finalisation to Job
4e7a12008f job: Add job_event_*()
d973eb6546 blockjob: Split block_job_event_pending()
7d447f8ab3 job: Move BlockJobCreateFlags to Job
ffa1c9ad30 job: Replace BlockJob.completed with job_is_completed()
13e426855c job: Move pause/resume functions to Job
0c52ff1b24 job: Add job_sleep_ns()
5ebf3ba66a job: Move coroutine and related code to Job
4e7b469718 job: Move defer_to_main_loop to Job
896fe05128 job: Add Job.aio_context
d73a94fde9 job: Move cancelled to Job
e386ae83f9 job: Add reference counting
3915da0ada job: Move state transitions to Job
e2ea58c7b0 job: Maintain a list of all jobs
717f2e60bf job: Add job_delete()
f31f336bca job: Add JobDriver.job_type
048bb87d0a job: Rename BlockJobType into JobType
b2707abc22 job: Create Job, JobDriver and job_create()
f7c7867229 blockjob: Improve BlockJobInfo.offset/len documentation
7611d0c75f blockjob: Update block-job-pause/resume documentation
=== OUTPUT BEGIN ===
Checking PATCH 1/40: blockjob: Update block-job-pause/resume documentation...
Checking PATCH 2/40: blockjob: Improve BlockJobInfo.offset/len documentation...
Checking PATCH 3/40: job: Create Job, JobDriver and job_create()...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#353:
new file mode 100644
total: 0 errors, 1 warnings, 424 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 4/40: job: Rename BlockJobType into JobType...
Checking PATCH 5/40: job: Add JobDriver.job_type...
Checking PATCH 6/40: job: Add job_delete()...
Checking PATCH 7/40: job: Maintain a list of all jobs...
Checking PATCH 8/40: job: Move state transitions to Job...
ERROR: space prohibited before open square bracket '['
#342: FILE: job.c:38:
+ /* U: */ [JOB_STATUS_UNDEFINED] = {0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0},
ERROR: space prohibited before open square bracket '['
#343: FILE: job.c:39:
+ /* C: */ [JOB_STATUS_CREATED] = {0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1},
ERROR: space prohibited before open square bracket '['
#344: FILE: job.c:40:
+ /* R: */ [JOB_STATUS_RUNNING] = {0, 0, 0, 1, 1, 0, 1, 0, 1, 0, 0},
ERROR: space prohibited before open square bracket '['
#345: FILE: job.c:41:
+ /* P: */ [JOB_STATUS_PAUSED] = {0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0},
ERROR: space prohibited before open square bracket '['
#346: FILE: job.c:42:
+ /* Y: */ [JOB_STATUS_READY] = {0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0},
ERROR: space prohibited before open square bracket '['
#347: FILE: job.c:43:
+ /* S: */ [JOB_STATUS_STANDBY] = {0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0},
ERROR: space prohibited before open square bracket '['
#348: FILE: job.c:44:
+ /* W: */ [JOB_STATUS_WAITING] = {0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0},
ERROR: space prohibited before open square bracket '['
#349: FILE: job.c:45:
+ /* D: */ [JOB_STATUS_PENDING] = {0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0},
ERROR: space prohibited before open square bracket '['
#350: FILE: job.c:46:
+ /* X: */ [JOB_STATUS_ABORTING] = {0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0},
ERROR: space prohibited before open square bracket '['
#351: FILE: job.c:47:
+ /* E: */ [JOB_STATUS_CONCLUDED] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1},
ERROR: space prohibited before open square bracket '['
#352: FILE: job.c:48:
+ /* N: */ [JOB_STATUS_NULL] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
total: 11 errors, 0 warnings, 528 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 9/40: job: Add reference counting...
Checking PATCH 10/40: job: Move cancelled to Job...
Checking PATCH 11/40: job: Add Job.aio_context...
Checking PATCH 12/40: job: Move defer_to_main_loop to Job...
Checking PATCH 13/40: job: Move coroutine and related code to Job...
Checking PATCH 14/40: job: Add job_sleep_ns()...
Checking PATCH 15/40: job: Move pause/resume functions to Job...
Checking PATCH 16/40: job: Replace BlockJob.completed with job_is_completed()...
Checking PATCH 17/40: job: Move BlockJobCreateFlags to Job...
Checking PATCH 18/40: blockjob: Split block_job_event_pending()...
Checking PATCH 19/40: job: Add job_event_*()...
Checking PATCH 20/40: job: Move single job finalisation to Job...
WARNING: line over 80 characters
#508: FILE: include/qemu/job.h:221:
+ int flags, BlockCompletionFunc *cb, void *opaque, Error **errp);
ERROR: "(foo*)" should be "(foo *)"
#578: FILE: job.c:462:
+ block_job_txn_del_job((BlockJob*) job);
ERROR: "(foo*)" should be "(foo *)"
#659: FILE: job.c:543:
+ block_job_txn_del_job((BlockJob*) job);
total: 2 errors, 1 warnings, 620 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 21/40: job: Convert block_job_cancel_async() to Job...
Checking PATCH 22/40: job: Add job_drain()...
Checking PATCH 23/40: job: Move .complete callback to Job...
Checking PATCH 24/40: job: Move job_finish_sync() to Job...
WARNING: line over 80 characters
#139: FILE: include/qemu/job.h:399:
+int job_finish_sync(Job *job, void (*finish)(Job *, Error **errp), Error **errp);
total: 0 errors, 1 warnings, 144 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 25/40: job: Switch transactions to JobTxn...
Checking PATCH 26/40: job: Move transactions to Job...
Checking PATCH 27/40: job: Move completion and cancellation to Job...
Checking PATCH 28/40: block: Cancel job in bdrv_close_all() callers...
Checking PATCH 29/40: job: Add job_yield()...
Checking PATCH 30/40: job: Add job_dismiss()...
Checking PATCH 31/40: job: Add job_is_ready()...
Checking PATCH 32/40: job: Add job_transition_to_ready()...
Checking PATCH 33/40: job: Move progress fields to Job...
Checking PATCH 34/40: job: Introduce qapi/job.json...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#235:
new file mode 100644
total: 0 errors, 1 warnings, 294 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 35/40: job: Add JOB_STATUS_CHANGE QMP event...
Checking PATCH 36/40: job: Add lifecycle QMP commands...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#37:
new file mode 100644
total: 0 errors, 1 warnings, 265 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 37/40: job: Add query-jobs QMP command...
Checking PATCH 38/40: blockjob: Remove BlockJob.driver...
ERROR: initializer for struct BlockJobDriver should normally be const
#21: FILE: blockjob.c:107:
+ BlockJobDriver *bjdrv = container_of(drv, BlockJobDriver, job_driver);
ERROR: initializer for struct BlockJobDriver should normally be const
#36: FILE: blockjob.c:121:
+ BlockJobDriver *bjdrv = container_of(drv, BlockJobDriver, job_driver);
total: 2 errors, 0 warnings, 54 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 39/40: iotests: Move qmp_to_opts() to VM...
WARNING: line over 80 characters
#109: FILE: tests/qemu-iotests/iotests.py:448:
+ self.assertEqual(self.vm.flatten_qmp_object(json.loads(json_filename[5:])),
total: 0 errors, 1 warnings, 83 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
Checking PATCH 40/40: qemu-iotests: Test job-* with block jobs...
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#13:
new file mode 100755
total: 0 errors, 1 warnings, 540 lines checked
Your patch has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
=== OUTPUT END ===
Test command exited with code: 1
---
Email generated automatically by Patchew [http://patchew.org/].
Please send your feedback to patchew-devel@redhat.com
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 00/40] Generic background jobs
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (40 preceding siblings ...)
2018-05-18 14:05 ` [Qemu-devel] [PATCH v2 00/40] Generic background jobs no-reply
@ 2018-05-18 18:41 ` Dr. David Alan Gilbert
2018-05-22 11:01 ` Kevin Wolf
2018-05-23 12:31 ` Kevin Wolf
42 siblings, 1 reply; 86+ messages in thread
From: Dr. David Alan Gilbert @ 2018-05-18 18:41 UTC (permalink / raw)
To: Kevin Wolf, marcandre.lureau
Cc: qemu-block, qemu-devel, jsnow, jcody, armbru, mreitz
* Kevin Wolf (kwolf@redhat.com) wrote:
> Before we can make x-blockdev-create a background job, we need to
> generalise the job infrastructure so that it can be used without any
> associated block node.
Is there any relationship between what this does, and what
Marc-André's 'monitor: add asynchronous command type' tries to do?
(See 20180326150916.9602-1-marcandre.lureau@redhat.com 26th March)
Dave
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
^ permalink raw reply [flat|nested] 86+ messages in thread* Re: [Qemu-devel] [PATCH v2 00/40] Generic background jobs
2018-05-18 18:41 ` Dr. David Alan Gilbert
@ 2018-05-22 11:01 ` Kevin Wolf
2018-05-22 17:15 ` Marc-André Lureau
0 siblings, 1 reply; 86+ messages in thread
From: Kevin Wolf @ 2018-05-22 11:01 UTC (permalink / raw)
To: Dr. David Alan Gilbert
Cc: marcandre.lureau, qemu-block, qemu-devel, jsnow, jcody, armbru, mreitz
Am 18.05.2018 um 20:41 hat Dr. David Alan Gilbert geschrieben:
> * Kevin Wolf (kwolf@redhat.com) wrote:
> > Before we can make x-blockdev-create a background job, we need to
> > generalise the job infrastructure so that it can be used without any
> > associated block node.
>
> Is there any relationship between what this does, and what
> Marc-André's 'monitor: add asynchronous command type' tries to do?
> (See 20180326150916.9602-1-marcandre.lureau@redhat.com 26th March)
Depends on who you ask. :-)
In a way, yes. Marc-André's async commands are for long-running
operations and jobs are for long-running operations. However, they are
for a different kind of "long-running".
Basically, Marc-André's work is for commands that are generally quick,
but perform some blocking operation. Usually commands that are currently
implemented synchronously, but always bothered us because they block the
VM for a while. They take maybe a few seconds at most, but you really
don't want them to block the guest even for short time.
The important part here is you don't want to modify the operations while
they're running, they just send their return value when they are done.
(In the latest version, Marc-André even made sure not to modify the wire
protocol, so IIUC the commands aren't really async any more in the sense
that you can have multiple commands running, but they are just
non-blocking in the sense that the guest can keep running while they are
doing their work.)
In comparison, jobs are typically really long running, like several
minutes (like copying a whole disk image). Some of them can even run
indefinitely (like mirroring) until you explicitly tell them to stop.
You want to continue using the monitor while they are running, and to be
able to manage them at runtime (pause/resume/set speed/cancel/...).
So I think both address different problems.
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 00/40] Generic background jobs
2018-05-22 11:01 ` Kevin Wolf
@ 2018-05-22 17:15 ` Marc-André Lureau
2018-05-29 17:16 ` Dr. David Alan Gilbert
0 siblings, 1 reply; 86+ messages in thread
From: Marc-André Lureau @ 2018-05-22 17:15 UTC (permalink / raw)
To: Kevin Wolf
Cc: Dr. David Alan Gilbert, qemu-block, qemu-devel, John Snow,
Jeff Cody, Armbruster, Markus, Max Reitz
Hi
On Tue, May 22, 2018 at 1:01 PM, Kevin Wolf <kwolf@redhat.com> wrote:
> Am 18.05.2018 um 20:41 hat Dr. David Alan Gilbert geschrieben:
>> * Kevin Wolf (kwolf@redhat.com) wrote:
>> > Before we can make x-blockdev-create a background job, we need to
>> > generalise the job infrastructure so that it can be used without any
>> > associated block node.
>>
>> Is there any relationship between what this does, and what
>> Marc-André's 'monitor: add asynchronous command type' tries to do?
>> (See 20180326150916.9602-1-marcandre.lureau@redhat.com 26th March)
>
> Depends on who you ask. :-)
>
> In a way, yes. Marc-André's async commands are for long-running
> operations and jobs are for long-running operations. However, they are
> for a different kind of "long-running".
>
> Basically, Marc-André's work is for commands that are generally quick,
> but perform some blocking operation. Usually commands that are currently
> implemented synchronously, but always bothered us because they block the
> VM for a while. They take maybe a few seconds at most, but you really
> don't want them to block the guest even for short time.
>
> The important part here is you don't want to modify the operations while
> they're running, they just send their return value when they are done.
> (In the latest version, Marc-André even made sure not to modify the wire
> protocol, so IIUC the commands aren't really async any more in the sense
> that you can have multiple commands running, but they are just
> non-blocking in the sense that the guest can keep running while they are
> doing their work.)
>
> In comparison, jobs are typically really long running, like several
> minutes (like copying a whole disk image). Some of them can even run
> indefinitely (like mirroring) until you explicitly tell them to stop.
> You want to continue using the monitor while they are running, and to be
> able to manage them at runtime (pause/resume/set speed/cancel/...).
>
> So I think both address different problems.
I agree with Kevin that both address different needs and complement
each other. Right now, QMP commands are handled synchronous, so they
block the main thread (or the "OOB" thread). I needed a simple way to
handle them asynchronously, *without* breaking or changing an existing
QMP command API. Job, on the other hand, provides a lot of facilities
that are unnecessary for most short living commands, while adding
complexity, at the protocol level (for instance, by using an extra
"job-id" space and not being tied to the qmp session). Depending on
the facilities Job provide, and the guarantee that must hold when
using it, it may make sense or help to transform an existing command
to an asynchronous version, but I doubt many of the existing commands
will need it. However, if there is a need to introduce job-like
facilities to QMP async (such as running concurrent commands, listing
running commands, being able to cancel them, being notified on
progression etc), then I think we should be careful to use Job. For
now, this is unneeded, as QMP async is internal to qemu, the client
can't run anything to manage an ongoing async command, so I think
both are improvements and will coexist harmoniously.
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 00/40] Generic background jobs
2018-05-22 17:15 ` Marc-André Lureau
@ 2018-05-29 17:16 ` Dr. David Alan Gilbert
0 siblings, 0 replies; 86+ messages in thread
From: Dr. David Alan Gilbert @ 2018-05-29 17:16 UTC (permalink / raw)
To: Marc-André Lureau
Cc: Kevin Wolf, qemu-block, qemu-devel, John Snow, Jeff Cody,
Armbruster, Markus, Max Reitz
* Marc-André Lureau (marcandre.lureau@redhat.com) wrote:
> Hi
>
> On Tue, May 22, 2018 at 1:01 PM, Kevin Wolf <kwolf@redhat.com> wrote:
> > Am 18.05.2018 um 20:41 hat Dr. David Alan Gilbert geschrieben:
> >> * Kevin Wolf (kwolf@redhat.com) wrote:
> >> > Before we can make x-blockdev-create a background job, we need to
> >> > generalise the job infrastructure so that it can be used without any
> >> > associated block node.
> >>
> >> Is there any relationship between what this does, and what
> >> Marc-André's 'monitor: add asynchronous command type' tries to do?
> >> (See 20180326150916.9602-1-marcandre.lureau@redhat.com 26th March)
> >
> > Depends on who you ask. :-)
> >
> > In a way, yes. Marc-André's async commands are for long-running
> > operations and jobs are for long-running operations. However, they are
> > for a different kind of "long-running".
> >
> > Basically, Marc-André's work is for commands that are generally quick,
> > but perform some blocking operation. Usually commands that are currently
> > implemented synchronously, but always bothered us because they block the
> > VM for a while. They take maybe a few seconds at most, but you really
> > don't want them to block the guest even for short time.
> >
> > The important part here is you don't want to modify the operations while
> > they're running, they just send their return value when they are done.
> > (In the latest version, Marc-André even made sure not to modify the wire
> > protocol, so IIUC the commands aren't really async any more in the sense
> > that you can have multiple commands running, but they are just
> > non-blocking in the sense that the guest can keep running while they are
> > doing their work.)
> >
> > In comparison, jobs are typically really long running, like several
> > minutes (like copying a whole disk image). Some of them can even run
> > indefinitely (like mirroring) until you explicitly tell them to stop.
> > You want to continue using the monitor while they are running, and to be
> > able to manage them at runtime (pause/resume/set speed/cancel/...).
> >
> > So I think both address different problems.
>
> I agree with Kevin that both address different needs and complement
> each other. Right now, QMP commands are handled synchronous, so they
> block the main thread (or the "OOB" thread). I needed a simple way to
> handle them asynchronously, *without* breaking or changing an existing
> QMP command API. Job, on the other hand, provides a lot of facilities
> that are unnecessary for most short living commands, while adding
> complexity, at the protocol level (for instance, by using an extra
> "job-id" space and not being tied to the qmp session). Depending on
> the facilities Job provide, and the guarantee that must hold when
> using it, it may make sense or help to transform an existing command
> to an asynchronous version, but I doubt many of the existing commands
> will need it. However, if there is a need to introduce job-like
> facilities to QMP async (such as running concurrent commands, listing
> running commands, being able to cancel them, being notified on
> progression etc), then I think we should be careful to use Job. For
> now, this is unneeded, as QMP async is internal to qemu, the client
> can't run anything to manage an ongoing async command, so I think
> both are improvements and will coexist harmoniously.
OK; it seems we are carving stuff up into a whole bunch of different
categories based on their longevity and blockableness.
(Normal, OOB, Async, Job)
I do worry a bit that pretty much anything that's not 'instant' means
it's waiting for something somewhere, and because you're waiting
for something that means you've got to be able to fail or cancel
it in case that thing never happens.
Dave
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
^ permalink raw reply [flat|nested] 86+ messages in thread
* Re: [Qemu-devel] [PATCH v2 00/40] Generic background jobs
2018-05-18 13:20 [Qemu-devel] [PATCH v2 00/40] Generic background jobs Kevin Wolf
` (41 preceding siblings ...)
2018-05-18 18:41 ` Dr. David Alan Gilbert
@ 2018-05-23 12:31 ` Kevin Wolf
42 siblings, 0 replies; 86+ messages in thread
From: Kevin Wolf @ 2018-05-23 12:31 UTC (permalink / raw)
To: qemu-block; +Cc: mreitz, jsnow, eblake, jcody, armbru, qemu-devel
Am 18.05.2018 um 15:20 hat Kevin Wolf geschrieben:
> Before we can make x-blockdev-create a background job, we need to
> generalise the job infrastructure so that it can be used without any
> associated block node.
>
> This series extracts a Job object from the block job infrastructure,
> which should contain everything related to jobs that doesn't require the
> block layer to be involved.
>
> Job creation with a centralised job-create command (if we even want
> this) as well as the actual conversion of x-blockdev-create to Job is
> left for another series.
Applied to the block branch.
Kevin
^ permalink raw reply [flat|nested] 86+ messages in thread