afr_frame_return() should be called within the previous lock (without taking it again of course) #728
Comments
gluster-ant
pushed a commit
that referenced
this issue
Oct 7, 2019
If you are already under lock, just decrement the call count directly instead of removing the lock, re-taking the lock and decrementing. Implements #728 updates: bz#1193929 Signed-off-by: Yaniv Kaul <ykaul@redhat.com> Change-Id: I3fa20b4651fbdb826655c5a03baeed46e99b5487
amarts
pushed a commit
to amarts/glusterfs
that referenced
this issue
Oct 9, 2019
If you are already under lock, just decrement the call count directly instead of removing the lock, re-taking the lock and decrementing. Implements gluster/glusterfs#728 updates: bz#1193929 Signed-off-by: Yaniv Kaul <ykaul@redhat.com> Change-Id: I3fa20b4651fbdb826655c5a03baeed46e99b5487
|
@amarts haven't given up on this for DHT ( https://review.gluster.org/#/c/glusterfs/+/23458/.) and will try to get it in to 8.0 |
amarts
pushed a commit
to amarts/glusterfs
that referenced
this issue
Mar 24, 2020
If you are already under lock, just decrement the call count directly instead of removing the lock, re-taking the lock and decrementing. Implements gluster/glusterfs#728 updates: bz#1193929 Signed-off-by: Yaniv Kaul <ykaul@redhat.com> Change-Id: I3fa20b4651fbdb826655c5a03baeed46e99b5487
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
In several locations, we see the following code:
Which essentially frees the frame lock that afr_frame_return() would take immediately after, just to decrement the call count. Instead, just perform
--local->call_count;inside that lock.Functions that I've seen this:
afr_flush_cbk
afr_fsyncdir_cbk
afr_statfs_cbk
afr_opendir_cbk
__afr_dir_write_cbk
__afr_inode_write_cbk
afr_open_cbk
The text was updated successfully, but these errors were encountered: