kernel export thread (nfsd/lockd/ksmbd) uses F_SETLK cmd with the FL_SLEEP flag set to request asynchronous processing of blocking locks. Currently cifs does not support such requests, it ignores F_SETLK cmd and handles the FL_SLEEP flag as an usual blocking lock request. To work around the problem let's detect such request, drop FL_SLEEP and wait_flag flags before cifs_setlk() execution. Dropped FL_SLEEP flag should be restored back because some calling function (nfsd4_lock) require it. https://bugzilla.kernel.org/show_bug.cgi?id=215383 Signed-off-by: Vasily Averin <vvs@xxxxxxxxxxxxx> --- fs/cifs/file.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/fs/cifs/file.c b/fs/cifs/file.c index 9fee3af83a73..6835458ee845 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -1847,6 +1847,7 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *flock) int lock = 0, unlock = 0; bool wait_flag = false; bool posix_lck = false; + bool async = false; struct cifs_sb_info *cifs_sb; struct cifs_tcon *tcon; struct cifsFileInfo *cfile; @@ -1890,8 +1891,17 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *flock) return -EOPNOTSUPP; } + async = (flock->fl_flags & FL_SLEEP) && IS_SETLK(cmd); + if (async) { + flock->fl_flags &= ~FL_SLEEP; + wait_flag = false; + } + rc = cifs_setlk(file, flock, type, wait_flag, posix_lck, lock, unlock, xid); + if (async) + flock->fl_flags |= FL_SLEEP; + free_xid(xid); return rc; } -- 2.25.1
