linux-toradex.git/fs/super.c, branch v3.3.5

vfs: Provide function to get superblock and wait for it to thaw

2012-02-14T01:45:38+00:00

In quota code we need to find a superblock corresponding to a device and wait
for superblock to be unfrozen. However this waiting has to happen without
s_umount semaphore because that is required for superblock to thaw. So provide
a function in VFS for this to keep dances with s_umount where they belong.

[AV: implementation switched to saner variant]

Signed-off-by: Jan Kara 
Signed-off-by: Al Viro

wake up s_wait_unfrozen when ->freeze_fs fails

2012-01-17T21:38:47+00:00

dd slept infinitely when fsfeeze failed because of EIO.
To fix this problem, if ->freeze_fs fails, freeze_super() wakes up
the tasks waiting for the filesystem to become unfrozen.

When s_frozen isn't SB_UNFROZEN in __generic_file_aio_write(),
the function sleeps until FITHAW ioctl wakes up s_wait_unfrozen.

However, if ->freeze_fs fails, s_frozen is set to SB_UNFROZEN and then
freeze_super() returns an error number. In this case, FITHAW ioctl returns
EINVAL because s_frozen is already SB_UNFROZEN. There is no way to wake up
s_wait_unfrozen, so __generic_file_aio_write() sleeps infinitely.

Signed-off-by: Kazuya Mio 
Signed-off-by: Al Viro

vfs: prevent remount read-only if pending removes

2012-01-07T04:20:13+00:00

If there are any inodes on the super block that have been unlinked
(i_nlink == 0) but have not yet been deleted then prevent the
remounting the super block read-only.

Reported-by: Toshiyuki Okajima 
Signed-off-by: Miklos Szeredi 
Tested-by: Toshiyuki Okajima 
Signed-off-by: Al Viro

vfs: protect remounting superblock read-only

2012-01-07T04:20:12+00:00

Currently remouting superblock read-only is racy in a major way.

With the per mount read-only infrastructure it is now possible to
prevent most races, which this patch attempts.

Before starting the remount read-only, iterate through all mounts
belonging to the superblock and if none of them have any pending
writes, set sb->s_readonly_remount.  This indicates that remount is in
progress and no further write requests are allowed.  If the remount
succeeds set MS_RDONLY and reset s_readonly_remount.

If the remounting is unsuccessful just reset s_readonly_remount.
This can result in transient EROFS errors, despite the fact the
remount failed.  Unfortunately hodling off writes is difficult as
remount itself may touch the filesystem (e.g. through load_nls())
which would deadlock.

A later patch deals with delayed writes due to nlink going to zero.

Signed-off-by: Miklos Szeredi 
Tested-by: Toshiyuki Okajima 
Signed-off-by: Al Viro

vfs: keep list of mounts for each superblock

2012-01-07T04:20:12+00:00

Keep track of vfsmounts belonging to a superblock.  List is protected
by vfsmount_lock.

Signed-off-by: Miklos Szeredi 
Tested-by: Toshiyuki Okajima 
Signed-off-by: Al Viro

vfs: fix the rest of sget() races

2012-01-04T03:53:10+00:00

unfortunately, just checking MS_BORN after having grabbed ->s_umount in
sget() is not enough; places that pick superblock from a list and
grab s_umount shared need the same check in addition to checking for
->s_root; otherwise three-way race between failing mount, sget() and
such list-walker can leave us with list-walker coming *second*, when
temporary active ref grabbed by sget() (to be dropped when sget()
notices that original mount has failed by checking MS_BORN) has
lead to deactivate_locked_super() from failing ->mount() *not* doing
->kill_sb() and just releasing ->s_umount.  Once sget() gets through
and notices that MS_BORN had never been set it will drop the active
ref and fs will be shut down and kicked out of all lists, but it's
too late for something like sync_supers().

Signed-off-by: Al Viro

vfs: convert fs_supers to hlist

2012-01-04T03:52:39+00:00

Signed-off-by: Al Viro

trim fs/internal.h

2012-01-04T03:52:35+00:00

some stuff in there can actually become static; some belongs to pnode.h
as it's a private interface between namespace.c and pnode.c...

Signed-off-by: Al Viro

vfs: ignore error on forced remount

2011-11-02T11:53:42+00:00

On emergency remount we want to force MS_RDONLY on the super block
even if ->remount_fs() failed for some reason.

Signed-off-by: Miklos Szeredi 
Tested-by: Toshiyuki Okajima 
Signed-off-by: Christoph Hellwig

vmscan: fix shrinker callback bug in fs/super.c

2011-11-01T00:30:49+00:00

The callback must not return -1 when nr_to_scan is zero. Fix the bug in
fs/super.c and add this requirement to the callback specification.

Signed-off-by: Mikulas Patocka 
Cc: Dave Chinner 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds