Re: [git pull] vfs.git; pile 1

From: Linus Torvalds
Date: Sun Jul 22 2012 - 13:34:30 EST

I'm not pulling this until the mess with the NFS tree is sorted out.
Apparently you rebased your (public!) VFS tree, and now half of your
old pre-rebase patches are in the NFS tree.

Rebasing public trees IS NOT A VALID OPERATION! Exactly because of
messes like this.

So no. No way am I pulling a big VFS tree that apparently has random
unknown aliased commits in other trees.


On Sun, Jul 22, 2012 at 3:09 AM, Al Viro <viro@xxxxxxxxxxxxxxxxxx> wrote:
> This one is *big* and changes quite a few things around VFS. What's in there:
> * the first of two really major architecture changes - death to
> open intents. The former is finally there; it was very long in making,
> but with Miklos getting through really hard and messy final push in
> fs/namei.c, we finally have it. Unlike his variant, this one doesn't
> introduce struct opendata; what we have instead is ->atomic_open()
> taking preallocated struct file * and passing everything via its fields.
> Instead of returning struct file *, it returns -E... on error, 0 on success
> and 1 in "deal with it yourself" case (e.g. symlink found on server,
> etc.). See comments before fs/namei.c:atomic_open(). That made a lot of
> goodies finally possible and quite a few are in that pile: ->lookup(),
> ->d_revalidate() and ->create() do not get struct nameidata * anymore;
> ->lookup() and ->d_revalidate() get lookup flags instead, ->create()
> gets "do we want it exclusive" flag. With the introduction of new helper
> (kern_path_locked()) we are rid of all struct nameidata instances outside
> of fs/namei.c; it's still visible in namei.h, but not for long. Come the
> next cycle, declaration will move either to fs/internal.h or to fs/namei.c
> itself. [me, miklos, hch]
> * the second major change: behaviour of final fput(). Now we
> have __fput() done without any locks held by caller *and* not from deep
> in call stack. That obviously lifts a lot of constraints on the locking
> in there. Moreover, it's legal now to call fput() from atomic contexts
> (which has immediately simplified life for aio.c). We also don't need
> anti-recursion logics in __scm_destroy() anymore. There is a price,
> though - the damn thing has become partially asynchronous. For fput()
> from normal process we are guaranteed that pending __fput() will be done
> before the caller returns to userland, exits or gets stopped for ptrace.
> For kernel threads and atomic contexts it's done via schedule_work(),
> so theoretically we might need a way to make sure it's finished; so far
> only one such place had been found, but there might be more. There's
> flush_delayed_fput() (do all pending __fput()) and there's __fput_sync()
> (fput() analog doing __fput() immediately). I hope we won't need them
> often; see warnings in fs/file_table.c for details. [me, based on task_work
> series from Oleg merged last cycle]
> * sync series from Jan
> * large part of "death to sync_supers()" work from Artem; the only
> bits missing here are exofs and ext4 ones. As far as I understand, those
> are going via the exofs and ext4 trees resp.; once they are in, we can
> put ->write_super() to the rest, along with the thread calling it.
> * preparatory bits from unionmount series (from dhowells).
> * assorted cleanups and fixes all over the place, as usual.
> This is not the last pile for this cycle; there's at least jlayton's
> ESTALE work and fsfreeze series (the latter - in dire need of
> fixes, so I'm not sure it'll make the cut this cycle). I'll
> probably throw symlink/hardlink restrictions stuff from Kees into
> the next pile, too. Plus there's a lot of misc patches I hadn't
> thrown into that one - it's large enough as it is...
> Please, pull from the usual place -
> git:// for-linus
> Shortlog:
> Al Viro (61):
> get rid of ->mnt_longterm
> get rid of magic in proc_namespace.c
> qnx6: don't bother with ->i_dentry in inode-freeing callback
> cifs: don't bother with ->i_dentry in ->destroy_inode()
> adfs: don't bother with ->i_dentry in ->destroy_inode()
> vfs: update documentation on ->i_dentry handling
> affs: get rid of open-coded list_for_each_entry()
> affs: unobfuscate affs_fix_dcache()
> ocfs2: use list_for_each_entry in ocfs2_find_local_alias()
> ext4: get rid of open-coded d_find_any_alias()
> vfs: switch i_dentry/d_alias to hlist
> coda: use list_for_each_entry
> namei.c: let follow_link() do put_link() on failure
> ->atomic_open() prototype change - pass int * instead of bool *
> don't modify od->filp at all
> make ->atomic_open() return int
> kill opendata->{mnt,dentry}
> kill struct opendata
> fs/namei.c: get do_last() and friends return int
> make finish_no_open() return int
> switch do_dentry_open() to returning int
> fold __dentry_open() into its sole caller
> do_dentry_open(): take initialization of file->f_path to caller
> switch nfs_lookup_check_intent() away from nameidata
> nfs_lookup_verify_inode() - nd is *always* non-NULL here
> fs/nfs/dir.c: switch to passing nd->flags instead of nd wherever possible
> stop passing nameidata * to ->d_revalidate()
> fs/namei.c: don't pass nameidata to d_revalidate()
> fs/namei.c: don't pass namedata to lookup_dcache()
> stop passing nameidata to ->lookup()
> fs/namei.c: don't pass nameidata to __lookup_hash() and lookup_real()
> don't pass nameidata to ->create()
> don't pass nameidata * to vfs_create()
> get rid of kern_path_parent()
> sysfs: switch to ->s_d_op and ->d_release()
> sysfs: just use d_materialise_unique()
> __d_unalias() should refuse to move mountpoints
> debugfs: make sure that debugfs_create_file() gets used only for regulars
> debugfs: fold debugfs_create_by_name() into the only caller
> debugfs: get rid of useless arguments to debugfs_{mkdir,symlink}
> mark_files_ro(): don't bother with mntget/mntput
> do_dentry_open(): close the race with mark_files_ro() in failure exit
> signal: make sure we don't get stopped with pending task_work
> trimming task_work: kill ->data
> trim task_work: get rid of hlist
> merge task_work and rcu_head, get rid of separate allocation for keyring case
> move exit_task_work() past exit_files()
> deal with task_work callbacks adding more work
> switch fput to task_work_add
> aio: now fput() is OK from interrupt context; get rid of manual delayed __fput()
> get rid of ->scm_work_list
> hold task_lock around checks in keyctl
> unobfuscate follow_up() a bit
> tidy up namei.c a bit
> don't expose I_NEW inodes via dentry->d_inode
> ecryptfs: don't reinvent the wheels, please - use struct completion
> zoran: don't bother with struct file * in zoran_map
> spufs: shift dget/mntget towards dentry_open()
> switch dentry_open() to struct path, make it grab references itself
> btrfs: switch btrfs_ioctl_balance() to mnt_want_write_file()
> ext4: switch EXT4_IOC_RESIZE_FS to mnt_want_write_file()
> Andrew Morton (1):
> notify_change(): check that i_mutex is held
> Artem Bityutskiy (23):
> affs: stop setting bm_flags
> affs: remove useless superblock writeout on unmount
> affs: remove useless superblock writeout on remount
> affs: re-structure superblock locking a bit
> affs: stop using lock_super
> affs: introduce VFS superblock object back-reference
> affs: get rid of affs_sync_super
> hfsplus: make hfsplus_sync_fs static
> hfsplus: amend debugging print
> hfsplus: remove useless check
> hfsplus: get rid of write_super
> hfs: push lock_super down
> hfs: get rid of lock_super
> hfs: remove extra mdb write on unmount
> hfs: simplify a bit checking for R/O
> hfs: introduce VFS superblock object back-reference
> hfs: get rid of hfs_sync_super
> fs/sysv: remove useless write_super call
> fs/sysv: remove another useless write_super call
> fs/sysv: stop using write_super and s_dirt
> fs/ufs: remove extra superblock write on unmount
> fs/ufs: re-arrange the code a bit
> fs/ufs: get rid of write_super
> Christoph Hellwig (2):
> fs: move path_put on failure out of ->follow_link
> fs: add nd_jump_link
> David Howells (6):
> VFS: Fix the banner comment on lookup_open()
> VFS: Make chown() and lchown() call fchownat()
> VFS: Make clone_mnt()/copy_tree()/collect_mounts() return errors
> VFS: Comment mount following code
> VFS: Pass mount flags to sget()
> VFS: Split inode_permission()
> Eric Sandeen (3):
> vfs: allow custom EOF in generic_file_llseek code
> ext4: use core vfs llseek code for dir seeks
> ext3: pass custom EOF to generic_file_llseek_size()
> Jan Kara (8):
> vfs: Move noop_backing_dev_info check from sync into writeback
> quota: Split dquot_quota_sync() to writeback and cache flushing part
> quota: Move quota syncing to ->sync_fs method
> vfs: Reorder operations during sys_sync
> vfs: Create function for iterating over block devices
> vfs: Make sys_sync writeout also block device inodes
> vfs: Remove unnecessary flushing of block devices
> vfs: Avoid unnecessary WB_SYNC_NONE writeback during sys_sync and reorder sync passes
> Julia Lawall (1):
> fs/direct-io.c: adjust suspicious bit operation
> Miklos Szeredi (21):
> vfs: do_last(): inline lookup_slow()
> vfs: do_last(): separate O_CREAT specific code
> vfs: do_last(): common slow lookup
> vfs: add lookup_open()
> vfs: lookup_open(): expand lookup_hash()
> vfs: add i_op->atomic_open()
> nfs: implement i_op->atomic_open()
> nfs: clean up ->create in nfs_rpc_ops
> nfs: don't use nd->
> nfs: don't use intents for checking atomic open
> fuse: implement i_op->atomic_open()
> cifs: implement i_op->atomic_open()
> ceph: remove unused arg from ceph_lookup_open()
> ceph: implement i_op->atomic_open()
> 9p: implement i_op->atomic_open()
> vfs: remove open intents from nameidata
> vfs: do_last(): clean up error handling
> vfs: do_last(): clean up labels
> vfs: do_last(): clean up bool
> vfs: do_last(): clean up retry
> vfs: move O_DIRECT check to common code
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at