* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ryusuke/nilfs2:
nilfs2: fix oopses with doubly mounted snapshots
nilfs2: missing a read lock for segment writer in nilfs_attach_checkpoint()
The commit 2ff05b2b (oom: move oom_adj value) moveed the oom_adj value to
the mm_struct. It was a very good first step for sanitize OOM.
However Paul Menage reported the commit makes regression to his job
scheduler. Current OOM logic can kill OOM_DISABLED process.
Why? His program has the code of similar to the following.
...
set_oom_adj(OOM_DISABLE); /* The job scheduler never killed by oom */
...
if (vfork() == 0) {
set_oom_adj(0); /* Invoked child can be killed */
execve("foo-bar-cmd");
}
....
vfork() parent and child are shared the same mm_struct. then above
set_oom_adj(0) doesn't only change oom_adj for vfork() child, it's also
change oom_adj for vfork() parent. Then, vfork() parent (job scheduler)
lost OOM immune and it was killed.
Actually, fork-setting-exec idiom is very frequently used in userland program.
We must not break this assumption.
Then, this patch revert commit 2ff05b2b and related commit.
Reverted commit list
---------------------
- commit 2ff05b2b4e (oom: move oom_adj value from task_struct to mm_struct)
- commit 4d8b9135c3 (oom: avoid unnecessary mm locking and scanning for OOM_DISABLE)
- commit 8123681022 (oom: only oom kill exiting tasks with attached memory)
- commit 933b787b57 (mm: copy over oom_adj value at fork time)
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Nick Piggin <npiggin@suse.de>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The last correction to the tcp_connect_to_sock error exit path,
commit a89d63a159, can free an already
freed socket, due to collision with a previous (incomplete) attempt
to fix the same issue, commit 311f6fc77c.
Signed-off-by: Casey Dahlin <cdahlin@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
will fix kernel oopses like the following:
# mount -t nilfs2 -r -o cp=20 /dev/sdb1 /test1
# mount -t nilfs2 -r -o cp=20 /dev/sdb1 /test2
# umount /test1
# umount /test2
BUG: sleeping function called from invalid context at arch/x86/mm/fault.c:1069
in_atomic(): 0, irqs_disabled(): 1, pid: 3886, name: umount.nilfs2
1 lock held by umount.nilfs2/3886:
#0: (&type->s_umount_key#31){+.+...}, at: [<c10b398a>] deactivate_super+0x52/0x6c
irq event stamp: 1219
hardirqs last enabled at (1219): [<c135c774>] __mutex_unlock_slowpath+0xf8/0x119
hardirqs last disabled at (1218): [<c135c6d5>] __mutex_unlock_slowpath+0x59/0x119
softirqs last enabled at (1214): [<c1033316>] __do_softirq+0x1a5/0x1ad
softirqs last disabled at (1205): [<c1033354>] do_softirq+0x36/0x5a
Pid: 3886, comm: umount.nilfs2 Not tainted 2.6.31-rc6 #55
Call Trace:
[<c1023549>] __might_sleep+0x107/0x10e
[<c13603c0>] do_page_fault+0x246/0x397
[<c136017a>] ? do_page_fault+0x0/0x397
[<c135e753>] error_code+0x6b/0x70
[<c136017a>] ? do_page_fault+0x0/0x397
[<c104f805>] ? __lock_acquire+0x91/0x12fd
[<c1050a62>] ? __lock_acquire+0x12ee/0x12fd
[<c1050a62>] ? __lock_acquire+0x12ee/0x12fd
[<c1050b2b>] lock_acquire+0xba/0xdd
[<d0d17d3f>] ? nilfs_detach_segment_constructor+0x2f/0x2fa [nilfs2]
[<c135d4fe>] down_write+0x2a/0x46
[<d0d17d3f>] ? nilfs_detach_segment_constructor+0x2f/0x2fa [nilfs2]
[<d0d17d3f>] nilfs_detach_segment_constructor+0x2f/0x2fa [nilfs2]
[<c104ea2c>] ? mark_held_locks+0x43/0x5b
[<c104ecb1>] ? trace_hardirqs_on_caller+0x10b/0x133
[<c104ece4>] ? trace_hardirqs_on+0xb/0xd
[<d0d09ac1>] nilfs_put_super+0x2f/0xca [nilfs2]
[<c10b3352>] generic_shutdown_super+0x49/0xb8
[<c10b33de>] kill_block_super+0x1d/0x31
[<c10e6599>] ? vfs_quota_off+0x0/0x12
[<c10b398f>] deactivate_super+0x57/0x6c
[<c10c4bc3>] mntput_no_expire+0x8c/0xb4
[<c10c5094>] sys_umount+0x27f/0x2a4
[<c10c50c6>] sys_oldumount+0xd/0xf
[<c10031a4>] sysenter_do_call+0x12/0x38
...
This turns out to be a bug brought by an -rc1 patch ("nilfs2: simplify
remaining sget() use").
In the patch, a new "put resource" function, nilfs_put_sbinfo()
was introduced to delay freeing nilfs_sb_info struct.
But the nilfs_put_sbinfo() mistakenly used atomic_dec_and_test()
function to check the reference count, and it caused the nilfs_sb_info
was freed when user mounted a snapshot twice.
This bug also suggests there was unseen memory leak in usual mount
/umount operations for nilfs.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
this patch is for the same problem that Benjamin Marzinski fixes at commit
b94a170e96
quotation of the original problem:
---cut here---
When a file is deleted from a gfs2 filesystem on one node, a dcache
entry for it may still exist on other nodes in the cluster. If this
happens, gfs2 will be unable to free this file on disk. Because of this,
it's possible to have a gfs2 filesystem with no files on it and no free
space. With this patch, when a node receives a callback notifying it
that the file is being deleted on another node, it schedules a new
workqueue thread to remove the file's dcache entry.
---end cut---
after applying Benjamin's patch, I think there is still a case in which the disk
inode remains even when "no space" is hit. the case is that when running
d_prune_aliases() against the inode, there are one or more dentries(aliases)
which have reference count number > 0. in this case the dentries won't be pruned.
and even later, the reference count becomes to 0, the dentries can still be
cached in memory. unfortunately, no callback come again, things come back to
the state before the callback runs. thus the on disk inode remains there until
in memoryinode is removed for some other reason(shrinking inode cache or unmount
the volume..).
this patch is to remove those dentries when their reference count becomes to 0 and
the inode is deleted by remote node. for implementation, gfs2_dentry_delete() is
added as dentry_operations.d_delete. the function returns true when the inode is
deleted by remote node. in dput(), gfs2_dentry_delete() is called and since it
returns true, the dentry is unhashed from dcache and then removed. when all dentries
are removed, the in memory inode get removed so that the on disk inode is freed.
Signed-off-by: Wengang Wang <wen.gang.wang@oracle.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
'ns_cno' of structure 'the_nilfs' must be protected from segment
writer, in other words, the caller of nilfs_get_checkpoint should hold
read lock for nilfs->ns_segctor_sem. This patch adds the lock/unlock
operations in nilfs_attach_checkpoint() when calling
nilfs_cpfile_get_checkpoint().
Signed-off-by: Zhang Qiang <zhangqiang.buaa@gmail.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
A user reported that although his root ext4 filesystem was mounting
fine, other filesystems would not mount, with the:
"Filesystem with huge files cannot be mounted RDWR without CONFIG_LBDAF"
error on his 32-bit box built without CONFIG_LBDAF. This is because
the test at mount time for this situation was not being re-checked
on remount, and the normal boot process makes an ro->rw transition,
so this was being missed.
Refactor to make a common helper function to test the filesystem
features against the type of mount request (RO vs. RW) so that we
stay consistent.
Addresses Red-Hat-Bugzilla: #517650
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
While reading through some of the mballoc code it seems that a couple
spots in the size normalization function could be streamlined.
The test for non-overlapping PAs can be or'd for the start & end
conditions, and the tests for adjacent PAs can be else-if'd -
it's essentially independently testing:
if (A + B <= C)
...
if (A > C)
...
These cannot both be true so it seems like the else-if might
be slightly more efficient and/or informative.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext4_mb_update_group_info is only called in one place, and it's
extremely simple. There's no reason to have it in a separate function
in a separate file as far as I can tell, it just obfuscates what's
really going on.
Perhaps it was intended to keep the grp->bb_* manipulation local to
mballoc.c but we're already accessing other grp-> fields in balloc.c
directly so this seems ok.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext4 will happily mount a > 16T filesystem on a 32-bit box, but
this is not safe; writes to the block device will wrap past 16T
and the page cache can't index past 16T (232 index * 4k pages).
Adding another test to the existing "too many sectors" test
should do the trick.
Add a comment, a relevant return value, and fix the reference
to the CONFIG_LBD(AF) option as well.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
During truncate we are sometimes forced to start a new transaction as
the amount of blocks to be journaled is both quite large and hard to
predict. So far we restarted a transaction while holding i_data_sem
and that violates lock ordering because i_data_sem ranks below a
transaction start (and it can lead to a real deadlock with
ext4_get_blocks() mapping blocks in some page while having a
transaction open).
We fix the problem by dropping the i_data_sem before restarting the
transaction and acquire it afterwards. It's slightly subtle that this
works:
1) By the time ext4_truncate() is called, all the page cache for the
truncated part of the file is dropped so get_block() should not be
called on it (we only have to invalidate extent cache after we
reacquire i_data_sem because some extent from not-truncated part could
extend also into the part we are going to truncate).
2) Writes, migrate or defrag hold i_mutex so they are stopped for all
the time of the truncate.
This bug has been found and analyzed by Theodore Tso <tytso@mit.edu>.
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
lockdep annotation for a transaction start has been at the end of
jbd2_journal_start(). But a transaction is also started from
jbd2_journal_restart(). Move the lockdep annotation to start_this_handle()
which covers both cases.
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext4_ext_show_leaf() will display the leaf extents when extent
debugging is enabled.
Printing out the unwritten bit is useful for debugging unwritten
extent, allow us to see the unwritten extents vs written extents,
after the unwritten extents are splitted or converted.
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
When EXT_DEBUG is enabled I received the following compile warning on
PPC64:
CC [M] fs/ext4/inode.o
CC [M] fs/ext4/extents.o
fs/ext4/extents.c: In function ‘ext4_ext_rm_leaf’:
fs/ext4/extents.c:2097: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 2 has type ‘ext4_lblk_t’
fs/ext4/extents.c: In function ‘ext4_ext_get_blocks’:
fs/ext4/extents.c:2789: warning: format ‘%u’ expects type ‘unsigned int’, but argument 4 has type ‘long unsigned int’
fs/ext4/extents.c:2852: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 3 has type ‘ext4_lblk_t’
fs/ext4/extents.c:2953: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 4 has type ‘unsigned int’
CC [M] fs/ext4/migrate.o
The patch fixes compile warning.
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Index: linux-2.6.31-rc4/fs/ext4/extents.c
===================================================================
Currently the group preallocation code tries to find a large (512)
free block from which to do per-cpu group allocation for small files.
The problem with this scheme is that it leaves the filesystem horribly
fragmented. In the worst case, if the filesystem is unmounted and
remounted (after a system shutdown, for example) we forget the fact
that wee were using a particular (now-partially filled) 512 block
extent. So the next time we try to allocate space for a small file,
we will find *another* completely free 512 block chunk to allocate
small files. Given that there are 32,768 blocks in a block group,
after 64 iterations of "mount, write one 4k file in a directory,
unmount", the block group will have 64 files, each separated by 511
blocks, and the block group will no longer have any free 512
completely free chunks of blocks for group preallocation space.
So if we try to allocate blocks for a file that has been closed, such
that we know the final size of the file, and the filesystem is not
busy, avoid using group preallocation.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The mount options string is saved in sb->s_options. This patch removes
the redundant duplicating of the mount options. Also, since we are not
displaying anything special in show options, we replace v9fs_show_options
with generic_show_options for now.
Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Cast the error return value (ENOMEM) in v9fs_get_inode() to its
correct type using ERR_PTR.
Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
If we fail to mount the filesystem, we have to be careful not to dereference
uninitialized structures in ocfs2_kill_sb.
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
Remove a redundant update of inode's i_uid and i_gid
after v9fs_get_inode() since the latter already sets up
a new inode and sets the proper uid and gid values.
Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
->get_sb can fail causing some badness. this patch fixes
* clear sb->fs_s_info in kill_sb.
* deactivate_locked_super() calls kill_sb (v9fs_kill_super) which closes the
destroys the client, clunks all its fids and closes the v9fs session.
Attempting to do it twice will cause an oops.
Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Add the delimiter ',' before the options when they are passed
and check if no option parameters are passed to prevent displaying
NULL in /proc/mounts.
Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Add missing p9stat_free in v9fs_inode_from_fid to avoid
any possible leaks.
Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Fix the comments -- mostly the improper and/or missing descriptions
of function parameters.
Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Add a missing iput when cleaning up if v9fs_get_inode
fails after returning a valid inode.
Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Check if v9fs_fid_add was successful or not based on its
return value.
Signed-off-by: Abhishek Kulkarni <adkulkar@umail.iu.edu>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
The inotify_add_watch man page specifies that inotify_add_watch() will
return a non-negative integer. However, historically the inotify
watches started at 1, not at 0.
Turns out that the inotifywait program provided by the inotify-tools
package doesn't properly handle a 0 watch descriptor. In 7e790dd5 we
changed from starting at 1 to starting at 0. This patch starts at 1,
just like in previous kernels, but also just like in previous kernels
it's possible for it to wrap back to 0. This preserves the kernel
functionality exactly like it was before the patch (neither method broke
the spec)
Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In f44aebcc the tail drop logic of events with no file backing
(q_overflow and in_ignored) was reversed so IN_IGNORED events would
never be tail dropped. This now means that Q_OVERFLOW events are NOT
tail dropped. The fix is to not tail drop IN_IGNORED, but to tail drop
Q_OVERFLOW.
Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
inotify decides if private data it passed to get added to an event was
used by checking list_empty(). But it's possible that the event may
have been dequeued and the private event removed so it would look empty.
The fix is to use the return code from fsnotify_add_notify_event rather
than looking at the list.
Signed-off-by: Eric Paris <eparis@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In ocfs2_do_truncate, we forget to release last_eb_bh which
will cause memleak. So call brelse in the end.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
ocfs2_read_virt_blocks() does BUG when we try to read a block from a file
beyond its end. Since this can happen due to filesystem corruption, it
is not really an appropriate answer. Make ocfs2_read_quota_block() check
the condition and handle it by calling ocfs2_error() and returning EIO.
[ Modified to print ip_blkno in the error - Joel ]
Reported-by: Tristan Ye <tristan.ye@oracle.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
This adds a link from the per-gfs2 sb sysfs directory to
the block device upon which the filesystem is mounted. The
link is called "device", strangely enough :-)
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
A little while back, block allocation was given some improved
error handling which meant that -EIO was returned in the case
of there being a problem in the resource group data. In addition
a message is printed explaning what went wrong and how to fix it.
This extends that error handling so that it also covers inode
allocation too.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
With each uevent, we now always include the journal ID. We
can't call it JID since that is already in use by some of
the individual events relating to recovery, so we use
JOURNALID instead. We don't send the JOURNALID for spectator
mounts, since there isn't one.
Also the ADD event now has both RDONLY and SPECTATOR information
to match that of the ONLINE event.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
We already have an offline uevent (used when a withdraw occurs)
but no online uevent. This adds an online uevent so that userspace
will be able to detect a successful mount by means other than
not receiving a remove event after the add & recovery (change)
uevents.
It has also been added to the remount path as well - we can't use
a change uevent there as older GFS2 userspace acts on change uevents
according to the state that it thinks the fs is in, so we can't
easily add any new ones.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
The locking in xfs_iget_cache_hit currently has numerous problems:
- we clear the reclaim tag without i_flags_lock which protects
modifications to it
- we call inode_init_always which can sleep with pag_ici_lock
held (this is oss.sgi.com BZ #819)
- we acquire and drop i_flags_lock a lot and thus provide no
consistency between the various flags we set/clear under it
This patch fixes all that with a major revamp of the locking in
the function. The new version acquires i_flags_lock early and
only drops it once we need to call into inode_init_always or before
calling xfs_ilock.
This patch fixes a bug seen in the wild where we race modifying the
reclaim tag.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Felix Blyakher <felixb@sgi.com>
Reviewed-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Felix Blyakher <felixb@sgi.com>
The triggered field of struct poll_wqueues introduced in commit
5f820f648c ("poll: allow f_op->poll to
sleep").
It was first set to 1 in pollwake() (now __pollwake() ), tested and
later set to 0 in poll_schedule_timeout(), but not initialized before.
As a result when the process needs to sleep, triggered was likely to be
non-zero even if pollwake() is not called before the first
poll_schedule_timeout(), meaning schedule_hrtimeout_range() would not be
called and an extra loop calling all ->poll() would be done.
This patch initialize triggered to 0 in poll_initwait() so the ->poll()
are not called twice before the process goes to sleep when it needs to.
Signed-off-by: Guillaume Knispel <gknispel@proformatique.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
