| Message ID | 1c7f83fe32ca8699c745629a3ad0581c7f75b1ee.1508337840.git.bcodding@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Self-NACK. I think the client assumes that it always sends operations matching the openmode, so NFS4ERR_OPENMODE returned from the server can be assumed to be the result of a revoked or expired state. In order to pass the server's results back up to fcntl(), we'd need a way to determine if the LOCK's state hasn't been revoked. I don't think its worth the work or time to fix up error handling for this one case, so probably the best thing to do is accept that commit e12937279c8b was just plain wrong and revert it. Ben On 18 Oct 2017, at 10:44, Benjamin Coddington wrote: > Commit e12937279c8b "NFS: Move the flock open mode check into > nfs_flock()" > changed NFSv3 behavior for flock() such that the open mode must match > the > lock type, however this requirement shouldn't be enforced for flock(). > > Since FL_POSIX locks are already checked in fs/locks.c, remove this > check > in NFS altogether. For NFSv4, allow the server to determine whether > or not > the mode should match the lock type for flock(). > > Suggested-by: Trond Myklebust <trond.myklebust@primarydata.com> > Signed-off-by: Benjamin Coddington <bcodding@redhat.com> > --- > fs/nfs/file.c | 16 +--------------- > 1 file changed, 1 insertion(+), 15 deletions(-) > > diff --git a/fs/nfs/file.c b/fs/nfs/file.c > index af330c31f627..ca7e737810e4 100644 > --- a/fs/nfs/file.c > +++ b/fs/nfs/file.c > @@ -822,22 +822,8 @@ int nfs_flock(struct file *filp, int cmd, struct > file_lock *fl) > if (NFS_SERVER(inode)->flags & NFS_MOUNT_LOCAL_FLOCK) > is_local = 1; > > - /* > - * VFS doesn't require the open mode to match a flock() lock's type. > - * NFS, however, may simulate flock() locking with posix locking > which > - * requires the open mode to match the lock type. > - */ > - switch (fl->fl_type) { > - case F_UNLCK: > + if (fl->fl_type == F_UNLCK) > return do_unlk(filp, cmd, fl, is_local); > - case F_RDLCK: > - if (!(filp->f_mode & FMODE_READ)) > - return -EBADF; > - break; > - case F_WRLCK: > - if (!(filp->f_mode & FMODE_WRITE)) > - return -EBADF; > - } > > return do_setlk(filp, cmd, fl, is_local); > } > -- > 2.9.3 > > -- > To unsubscribe from this list: send the line "unsubscribe linux-nfs" > in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/fs/nfs/file.c b/fs/nfs/file.c index af330c31f627..ca7e737810e4 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -822,22 +822,8 @@ int nfs_flock(struct file *filp, int cmd, struct file_lock *fl) if (NFS_SERVER(inode)->flags & NFS_MOUNT_LOCAL_FLOCK) is_local = 1; - /* - * VFS doesn't require the open mode to match a flock() lock's type. - * NFS, however, may simulate flock() locking with posix locking which - * requires the open mode to match the lock type. - */ - switch (fl->fl_type) { - case F_UNLCK: + if (fl->fl_type == F_UNLCK) return do_unlk(filp, cmd, fl, is_local); - case F_RDLCK: - if (!(filp->f_mode & FMODE_READ)) - return -EBADF; - break; - case F_WRLCK: - if (!(filp->f_mode & FMODE_WRITE)) - return -EBADF; - } return do_setlk(filp, cmd, fl, is_local); }
Commit e12937279c8b "NFS: Move the flock open mode check into nfs_flock()" changed NFSv3 behavior for flock() such that the open mode must match the lock type, however this requirement shouldn't be enforced for flock(). Since FL_POSIX locks are already checked in fs/locks.c, remove this check in NFS altogether. For NFSv4, allow the server to determine whether or not the mode should match the lock type for flock(). Suggested-by: Trond Myklebust <trond.myklebust@primarydata.com> Signed-off-by: Benjamin Coddington <bcodding@redhat.com> --- fs/nfs/file.c | 16 +--------------- 1 file changed, 1 insertion(+), 15 deletions(-)