A couple lock manager calls used by the lease code look unnecessary to me; the following patches remove them. I'm planning on submitting them with the nfsd patches for 2.6.38, barring any objections. --b. --

When we converted to sharing struct filess between nfs4 opens I went too far and also used the same mechanism for delegations. But keeping a reference to the struct file ensures it will outlast the lease, and allows us to remove the lease with the same file as we added it. Signed-off-by: J. Bruce Fields <bfields@redhat.com> --- fs/nfsd/nfs4state.c | 10 +++++----- fs/nfsd/state.h | 1 + 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/fs/nfsd/nfs4state.c ...

Signed-off-by: J. Bruce Fields <bfields@redhat.com> --- fs/locks.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/fs/locks.c b/fs/locks.c index 5cb6506..feaac63 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -1399,7 +1399,7 @@ int generic_setlease(struct file *filp, long arg, struct file_lock **flp) for (before = &inode->i_flock; ((fl = *before) != NULL) && IS_LEASE(fl); before = &fl->fl_next) { - if (fl->fl_file == lease->fl_file) + if (fl->fl_file == ...

Add an ioctl to dump filesystem's metadata in memory in vfs. Userspace collects such info and uses it to do metadata readahead. Filesystem can hook to super_operations.metadata_incore to get metadata in specific approach. Next patch will give an example how to implement .metadata_incore in btrfs. Signed-off-by: Shaohua Li <shaohua.li@intel.com> --- fs/compat_ioctl.c | 1 fs/ioctl.c | 75 +++++++++++++++++++++++++++++++++++++++++++++++++++++ include/linux/fs.h | 15 ...

Add metadata readahead ioctl in vfs. Filesystem can hook to super_operations.metadata_readahead to handle filesystem specific task. Next patch will give an example how btrfs implements it. Signed-off-by: Shaohua Li <shaohua.li@intel.com> --- fs/compat_ioctl.c | 1 + fs/ioctl.c | 21 +++++++++++++++++++++ include/linux/fs.h | 8 ++++++++ 3 files changed, 30 insertions(+) Index: linux/fs/ioctl.c =================================================================== --- ...

do validation for extent_buffer if it's skipped before With metadata readahead, we slightly change the behavior. Before it, we allocate an extent_buffer (so set page->private), do metadata read and btree_readpage_end_io_hook() will do validation. After it, we directly do metadata readahead, and since in this case page hasn't ->private, btree_readpage_end_io_hook() will not do validation. This patch fixes this. It addes a new flag to indicate if a buffer is validated. If not and even the buffer ...

Sorry, it was the commit 93bb41f4f8b89ac8 ("fs: Do not dispatch FITRIM through separate super_operation"). --

The ->trim_fs has been removed meanwhile, so remove it from the documentation as well. Signed-off-by: Christoph Hellwig <hch@lst.de> Reported-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> Index: linux-2.6/Documentation/filesystems/Locking =================================================================== --- linux-2.6.orig/Documentation/filesystems/Locking 2011-01-04 07:08:41.296003372 +0100 +++ linux-2.6/Documentation/filesystems/Locking 2011-01-04 07:08:45.561004280 +0100 @@ -115,7 ...

And actually I think it is much better to have sync_inode, which means we'll be able to get rid of commit_metadata (which should be an inode operation anyway, not an export operation which really should deal with exporting filesystems to a non-vfs namespace, not nfsd hacks). commit_metadata would just be sync_inode with a null range or no data sync flag set. --

As explained in the previous mail it's not just not writing data, it's conceptually quite different from fsync. --

Did you notice the comment above the function? ;-) You should really add the new ioctls to compat_sys_ioctl, not to the COMPATIBLE_IOCTL() list, in order to make the behavior consistent between 32 and 64 bit user space. The main difference is that all ioctl commands that are hardcoded in the functions get called before trying to call the file system specific This should be _IOW, not _IOR. Otherwise looks good. Arnd --

Hmm I could not reproduce this with the latest state (ef9bf3b7144bee6ce) of branch 'origin/for-next' from git.infradead.org/users/eparis/notify.git What i did was: - register with fanotify - set mark for OPEN_PERM event - read an event - cause a segfault before response is returned to fanotify The process terminates and the core file is created as expected. I dont think that this will work. A fanotify registration is not tracked by its pid, but by its group which may be valid ...

Whoops, thanks for the reminder. Looks like we never added fl_mylease? That leaves the fl_release_private patch, updated as follows. --b. commit 3d801116bb23a1f446627ce1976950c7a126541e Author: J. Bruce Fields <bfields@redhat.com> Date: Sat Oct 30 17:41:26 2010 -0400 nfsd4: eliminate lease delete callback nfsd controls the lifetime of the lease, not the lock code, so there's no need for this callback on lease destruction. Signed-off-by: J. Bruce Fields ...