Skip to contentSkip to conversation

cifs: remove the retry in cifs_poxis_lock_set

PATCH Superseded yangerkun 2 participants 4 messages Jun 24, 2020 – Jul 1, 2020

Subsystems (3)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

yangerkun

author
#0 lore raw

PATCH 1/1 cifs: remove the retry in cifs_poxis_lock_set

The caller of cifs_posix_lock_set will do retry(like fcntl_setlk64->do_lock_file_wait) if we will wait for any file_lock. So the retry in cifs_poxis_lock_set seems duplicated, remove it to make a cleanup.

---
 fs/cifs/file.c | 8 --------
 1 file changed, 8 deletions(-)
fs/cifs/file.c +0 -8
index 9b0f8f33f832..2c9c24b1805d 100644
@@ -1162,7 +1162,6 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
 	if ((flock->fl_flags & FL_POSIX) == 0)
 		return rc;
 
-try_again:
 	cifs_down_write(&cinode->lock_sem);
 	if (!cinode->can_cache_brlcks) {
 		up_write(&cinode->lock_sem);
@@ -1171,13 +1170,6 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
 
 	rc = posix_lock_file(file, flock, NULL);
 	up_write(&cinode->lock_sem);
-	if (rc == FILE_LOCK_DEFERRED) {
-		rc = wait_event_interruptible(flock->fl_wait,
-					list_empty(&flock->fl_blocked_member));
-		if (!rc)
-			goto try_again;
-		locks_delete_block(flock);
-	}
 	return rc;
 }
 
Signature · 1 line
2.25.4
author
#1 lore raw

Ping...

Quoted 37 lines · yangerkun

The caller of cifs_posix_lock_set will do retry(like fcntl_setlk64->do_lock_file_wait) if we will wait for any file_lock. So the retry in cifs_poxis_lock_set seems duplicated, remove it to make a cleanup.

---
  fs/cifs/file.c | 8 --------
  1 file changed, 8 deletions(-)
fs/cifs/file.c +0 -8
index 9b0f8f33f832..2c9c24b1805d 100644
@@ -1162,7 +1162,6 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
  	if ((flock->fl_flags & FL_POSIX) == 0)
  		return rc;
 
-try_again:
  	cifs_down_write(&cinode->lock_sem);
  	if (!cinode->can_cache_brlcks) {
  		up_write(&cinode->lock_sem);
@@ -1171,13 +1170,6 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
 
  	rc = posix_lock_file(file, flock, NULL);
  	up_write(&cinode->lock_sem);
-	if (rc == FILE_LOCK_DEFERRED) {
-		rc = wait_event_interruptible(flock->fl_wait,
-					list_empty(&flock->fl_blocked_member));
-		if (!rc)
-			goto try_again;
-		locks_delete_block(flock);
-	}
  	return rc;
  }
 
#2 lore raw
Quoted 7 lines · yangerkun

Ping...

Quoted 4 lines · yangerkun

The caller of cifs_posix_lock_set will do retry(like fcntl_setlk64->do_lock_file_wait) if we will wait for any file_lock. So the retry in cifs_poxis_lock_set seems duplicated, remove it to make a cleanup.

If cifs_posix_try_lock() returns FILE_LOCK_DEFERRED (which it might after your patch), then cifs_setlk() will check the return value:

		if (!rc || rc < 0)
			return rc;

These tests will fail (as FILE_LOCK_DEFERRED is 1) and so it will continue on as though the lock was granted.

So I think your patch is wrong.
However I think your goal is correct. cifs shouldn't be waiting. No other filesystem waits when it gets FILE_LOCK_DEFERRED.

So maybe try to fix up your patch.

Thanks,
NeilBrown

Quoted 33 lines
Quoted 33 lines
---
  fs/cifs/file.c | 8 --------
  1 file changed, 8 deletions(-)
fs/cifs/file.c +0 -8
index 9b0f8f33f832..2c9c24b1805d 100644
@@ -1162,7 +1162,6 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
  	if ((flock->fl_flags & FL_POSIX) == 0)
  		return rc;
 
-try_again:
  	cifs_down_write(&cinode->lock_sem);
  	if (!cinode->can_cache_brlcks) {
  		up_write(&cinode->lock_sem);
@@ -1171,13 +1170,6 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
 
  	rc = posix_lock_file(file, flock, NULL);
  	up_write(&cinode->lock_sem);
-	if (rc == FILE_LOCK_DEFERRED) {
-		rc = wait_event_interruptible(flock->fl_wait,
-					list_empty(&flock->fl_blocked_member));
-		if (!rc)
-			goto try_again;
-		locks_delete_block(flock);
-	}
  	return rc;
  }
 
author
#3 lore raw
Quoted 16 lines · NeilBrown
Quoted 7 lines · yangerkun

Ping...

Quoted 4 lines · yangerkun

The caller of cifs_posix_lock_set will do retry(like fcntl_setlk64->do_lock_file_wait) if we will wait for any file_lock. So the retry in cifs_poxis_lock_set seems duplicated, remove it to make a cleanup.

If cifs_posix_try_lock() returns FILE_LOCK_DEFERRED (which it might after your patch), then cifs_setlk() will check the return value:

		if (!rc || rc < 0)
			return rc;
Quoted 8 lines

These tests will fail (as FILE_LOCK_DEFERRED is 1) and so it will continue on as though the lock was granted.

So I think your patch is wrong.
However I think your goal is correct. cifs shouldn't be waiting. No other filesystem waits when it gets FILE_LOCK_DEFERRED.

So maybe try to fix up your patch.

Yes, we should check FILE_LOCK_DEFERRED in cifs_setlk after this patch. Also we may change 'int rc = 1;' exists in cifs_posix_lock_set since FILE_LOCK_DEFERRED equals to 1.

I will send a v2 and thanks a lot for your review!

Thanks,
Kun.

Quoted 37 lines

Thanks,
NeilBrown

Quoted 33 lines
Quoted 33 lines
---
   fs/cifs/file.c | 8 --------
   1 file changed, 8 deletions(-)
fs/cifs/file.c +0 -8
index 9b0f8f33f832..2c9c24b1805d 100644
@@ -1162,7 +1162,6 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
   	if ((flock->fl_flags & FL_POSIX) == 0)
   		return rc;
 
-try_again:
   	cifs_down_write(&cinode->lock_sem);
   	if (!cinode->can_cache_brlcks) {
   		up_write(&cinode->lock_sem);
@@ -1171,13 +1170,6 @@ cifs_posix_lock_set(struct file *file, struct file_lock *flock)
 
   	rc = posix_lock_file(file, flock, NULL);
   	up_write(&cinode->lock_sem);
-	if (rc == FILE_LOCK_DEFERRED) {
-		rc = wait_event_interruptible(flock->fl_wait,
-					list_empty(&flock->fl_blocked_member));
-		if (!rc)
-			goto try_again;
-		locks_delete_block(flock);
-	}
   	return rc;
   }