lockd: fix hang on shutdown when there are active locks
Subsystems (2)
Jeff Layton
maintainercover lockd: fix hang on shutdown when there are active locks
We had a report that shutting down nfsd would hang when there were active NFSv3 locks. The first patch fixes that. While testing that I hit a crash in nlm_close_files. The second patch fixes that one.
Jeff Layton (2):
lockd: set fl_owner when unlocking files
lockd: fix nlm_close_files
fs/lockd/svcsubs.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
Signature · 1 line
2.36.1
Jeff Layton
maintainerPATCH 1/2 lockd: set fl_owner when unlocking files
Unlocking a POSIX inode with vfs_lock_file only works if the owner matches. Ensure we set it in the request.
--- fs/lockd/svcsubs.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
fs/lockd/svcsubs.c +3 -2
@@ -176,7 +176,7 @@ nlm_delete_file(struct nlm_file *file) } } -static int nlm_unlock_files (struct nlm_file *file) +static int nlm_unlock_files (struct nlm_file *file,fl_owner_t owner) {struct file_lock lock;
@@ -184,6 +184,7 @@ static int nlm_unlock_files(struct nlm_file *file) lock.fl_type =F_UNLCK ; lock.fl_start =0 ; lock.fl_end =OFFSET_MAX ; + lock.fl_owner = owner;if (file->f_file[O_RDONLY ] &&vfs_lock_file (file->f_file[O_RDONLY ],F_SETLK , &lock,NULL )) goto out_err;
@@ -225,7 +226,7 @@ nlm_traverse_locks(struct nlm_host *host, struct nlm_file *file,if (match (lockhost, host)) {spin_unlock (&flctx->flc_lock); -if (nlm_unlock_files (file)) +if (nlm_unlock_files (file, fl->fl_owner))return 1 ; goto again; }
Signature · 1 line
2.36.1
Jeff Layton
maintainerRe: [PATCH 1/2] lockd: set fl_owner when unlocking files
Quoted 3 lines · Jeff Layton
Unlocking a POSIX inode with
vfs_lock_fileonly works if the owner matches. Ensure we set it in the request.
Oof, that description makes no sense. How about:
"Unlocking a POSIX lock on an inode with vfs_lock_file..."
Quoted 37 lines
Cc J. Bruce Fields <[email protected]>Fixes 7f024fcd5c97 ("Keep read and write fds with each nlm_file")Signed-off-by Jeff Layton <[email protected]>--- fs/lockd/svcsubs.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)fs/lockd/svcsubs.c +3
-2@@ -176,7 +176,7 @@ nlm_delete_file(struct nlm_file *file) } } -static int nlm_unlock_files (struct nlm_file *file) +static int nlm_unlock_files (struct nlm_file *file,fl_owner_t owner) {struct file_lock lock;@@ -184,6 +184,7 @@ static int nlm_unlock_files(struct nlm_file *file) lock.fl_type =F_UNLCK ; lock.fl_start =0 ; lock.fl_end =OFFSET_MAX ; + lock.fl_owner = owner;if (file->f_file[O_RDONLY ] &&vfs_lock_file (file->f_file[O_RDONLY ],F_SETLK , &lock,NULL )) goto out_err;@@ -225,7 +226,7 @@ nlm_traverse_locks(struct nlm_host *host, struct nlm_file *file,if (match (lockhost, host)) {spin_unlock (&flctx->flc_lock); -if (nlm_unlock_files (file)) +if (nlm_unlock_files (file, fl->fl_owner))return 1 ; goto again; }
Signature · 1 line
Jeff Layton <[email protected]>
Jeff Layton
maintainerPATCH 2/2 lockd: fix nlm_close_files
In reply to Jeff LaytonThis loop condition tries a bit too hard to be clever. Just test for the two indices we care about explicitly.
--- fs/lockd/svcsubs.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)
fs/lockd/svcsubs.c +4 -5
@@ -283,11 +283,10 @@ nlm_file_inuse(struct nlm_file *file)static void nlm_close_files (struct nlm_file *file) { -struct file *f; - -for (f = file->f_file[0 ]; f <= file->f_file[1 ]; f++) -if (f) - nlmsvc_ops->fclose (f); +if (file->f_file[O_RDONLY ]) + nlmsvc_ops->fclose (file->f_file[O_RDONLY ]); +if (file->f_file[O_WRONLY ]) + nlmsvc_ops->fclose (file->f_file[O_WRONLY ]); }/*
Signature · 1 line
2.36.1
Quoted 8 lines · Jeff Layton
We had a report that shutting down nfsd would hang when there were active NFSv3 locks. The first patch fixes that. While testing that I hit a crash in
nlm_close_files. The second patch fixes that one.
Quoted 6 lines
Jeff Layton (2):
lockd: setfl_ownerwhen unlocking files
lockd: fixnlm_close_files
fs/lockd/svcsubs.c| 14 +++++++-------1 file changed, 7 insertions(+), 7 deletions(-)
Grabbed both for the next 5.18-rc PR.
--
Chuck Lever