Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
maestro-core
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Container registry
Model registry
Analyze
Contributor analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
GitLab community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
maestro
maestro-core
Commits
6b320574
Commit
6b320574
authored
Jul 6, 2021
by
Christopher Haine
Browse files
Options
Downloads
Patches
Plain Diff
fix timely rdma completion
parent
1d4ab5ad
Branches
Branches containing commit
No related tags found
No related merge requests found
Pipeline
#72343
failed
Jul 6, 2021
Stage: build-and-test
Changes
2
Pipelines
1
Show whitespace changes
Inline
Side-by-side
Showing
2 changed files
transport/rdma.c
+11
-1
11 additions, 1 deletion
transport/rdma.c
transport/transport.c
+18
-6
18 additions, 6 deletions
transport/transport.c
with
29 additions
and
7 deletions
transport/rdma.c
+
11
−
1
View file @
6b320574
...
...
@@ -421,12 +421,22 @@ mstro_transport_rdma_cb(mstro_event ev, void* closure)
// mark readiness for (the transport thread and then) the local pool
// there is only one waiter (the transport thread) for one CDO transfer completion
regentry
->
done
=
1
;
/*
regentry->done = 1;
err = pthread_cond_signal(&(regentry->cond));
if (err) {
ERR("Couldn't signal (errno: %d)\n", err);
args->status = MSTRO_FAIL;
}
*/
/* indicate that CDO is filled */
mstro_cdo
cdo
;
mstro_cdo__find_cdo
(
&
(
args
->
cdoid
),
&
cdo
);
mstro_status
s
=
mstro_cdo__mark_transfer_complete
(
cdo
);
if
(
s
!=
MSTRO_OK
)
{
ERR
(
"Failed to indicate completion of incoming CDO: %d (%s)
\n
"
,
s
,
mstro_status_description
(
s
));
}
DEBUG
(
"RDMA transport for CDO `%s` successful
\n
"
,
args
->
name
);
}
...
...
This diff is collapsed.
Click to expand it.
transport/transport.c
+
18
−
6
View file @
6b320574
...
...
@@ -526,7 +526,18 @@ mstro_transport_execute(
ERR
(
"RDMA transfer completion message does not map to a documented transfer on destination
\n
"
);
return
MSTRO_FAIL
;
}
/*
pthread_mutex_lock(&(regentry->lock));
while (regentry->done == 0) {
int err = pthread_cond_wait(&(regentry->cond), &(regentry->lock));
if (err) {
ERR("Couldn't cond wait (errno: %d)\n", err);
return MSTRO_FAIL;
}
}
pthread_mutex_unlock(&(regentry->lock));
*/
}
else
{
/* indicate that CDO is filled */
s
=
mstro_cdo__mark_transfer_complete
(
cdo
);
...
...
@@ -534,6 +545,7 @@ mstro_transport_execute(
ERR
(
"Failed to indicate completion of incoming CDO: %d (%s)
\n
"
,
s
,
mstro_status_description
(
s
));
}
}
}
else
{
ERR
(
"Unexpected CDO state %s: %d -- unsolicted incoming transfer?
\n
"
,
cdo
->
name
,
cdo
->
state
);
...
...
This diff is collapsed.
Click to expand it.
Christopher Haine
@haine2
mentioned in issue
#44 (closed)
·
Jul 6, 2021
mentioned in issue
#44 (closed)
mentioned in issue #44
Toggle commit list
Ali Mohammed
@mohammed1
mentioned in issue
#36 (closed)
·
Jul 12, 2021
mentioned in issue
#36 (closed)
mentioned in issue #36
Toggle commit list
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment