tcmur_device: fix racy between reopening devices and reporting events #695
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
pkalever
reviewed
Mar 20, 2023
idryomov
reviewed
Mar 20, 2023
idryomov
reviewed
Mar 22, 2023
tcmur_device.c
Outdated
| @@ -46,26 +46,15 @@ int __tcmu_reopen_dev(struct tcmu_device *dev, int retries) | |||
| int ret, attempt = 0; | |||
| bool needs_close = false; | |||
|
|
|||
| assert(rdev->flags & TCMUR_DEV_FLAG_IN_RECOVERY); | |||
There was a problem hiding this comment.
Shouldn't this be done with rdev->rdev_lock held (i.e. on line 50)?
There was a problem hiding this comment.
This is not that critical IMO. But I will fix it.
target.c
Outdated
| @@ -336,11 +336,14 @@ int tcmu_add_dev_to_recovery_list(struct tcmu_device *dev) | |||
| goto done; | |||
| } | |||
| list_add(&tpg_recovery_list, &tpg->recovery_entry); | |||
| rdev->flags |= TCMUR_DEV_FLAG_IN_RECOVERY; | |||
There was a problem hiding this comment.
I think this should be done after adding rdev to tpg->devs list:
add_to_list:
list_add(&tpg->devs, &rdev->recovery_entry);
rdev->flags |= TCMUR_DEV_FLAG_IN_RECOVERY;
Otherwise TCMUR_DEV_FLAG_IN_RECOVERY wouldn't get set in the "another device already kicked off recovery" case (or, if it somehow would, it's really not obvious).
There was a problem hiding this comment.
Correct. I will fix it.
target.c
Outdated
| @@ -336,11 +336,14 @@ int tcmu_add_dev_to_recovery_list(struct tcmu_device *dev) | |||
| goto done; | |||
| } | |||
| list_add(&tpg_recovery_list, &tpg->recovery_entry); | |||
| rdev->flags |= TCMUR_DEV_FLAG_IN_RECOVERY; | |||
| rdev->conn_lost_cnt++; | |||
There was a problem hiding this comment.
The increment should stay in __tcmu_notify_conn_lost IMO. The point of my suggestion was to keep related things close and not mix them:
- set
TCMUR_DEV_FLAG_IN_RECOVERYin the function that addsrdevto the respective list (usingrdev->recovery_entry) - increment
rdev->conn_lost_cntin the function that handles a lost connection
target.c
Outdated
|
|
||
| add_to_list: | ||
| list_add(&tpg->devs, &rdev->recovery_entry); | ||
| done: | ||
| tcmu_release_alua_grps(&alua_list); | ||
| pthread_mutex_unlock(&tpg_recovery_lock); | ||
|
|
To make sure all the in-flight IOs have been finished before flushing the event_work. Or just after we flush the event_work a new timedout IO callback could come and then it will fire a new event work, which may access the device while we are closing the device later. Signed-off-by: Xiubo Li <xiubli@redhat.com>
idryomov
approved these changes
Mar 22, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
To make sure all the in-flight IOs have been finished before flushing the event_work. Or just after we flush the event_work a new timedout IO callback could come and then it will fire a new event work, which may access the device while we are closing the device later.
Signed-off-by: Xiubo Li xiubli@redhat.com