state->ios_left isn't decremented for requests that don't need a file,
so it might be larger than number of SQEs left. That in some
circumstances makes us to grab more files that is needed so imposing
extra put.
Deaccount one ios_left for each request.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Keep ->needs_file_no_error check out of io_file_get(), and let callers
handle it. It makes it more straightforward. Also, as the only error it
can hand back -EBADF, make it return a file or NULL.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
ctx->nr_user_files == 0 IFF ctx->file_data == NULL and there fixed files
are not used. Hence, verifying fds only against ctx->nr_user_files is
enough. Remove the other check from hot path.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Move work.files grabbing into io_prep_async_work() to all other work
resources initialisation. We don't need to keep it separately now, as
->ring_fd/file are gone. It also allows to not grab it when a request
is not going to io-wq.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There is no real reason left for preparing io-wq work context for linked
requests in advance, remove it as this might become a bottleneck in some
cases.
Reported-by: Roman Gershman <romger@amazon.com>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If compressed inode has inconsistent fields on i_compress_algorithm,
i_compr_blocks and i_log_cluster_size, we missed to set SBI_NEED_FSCK
to notice fsck to repair the inode, fix it.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
There are no bugs here that I've spotted, it's just easier to use the
normal API and there are no performance advantages to using the more
verbose advanced API.
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The xas_store() wasn't paired with an xas_nomem() loop, so if it couldn't
allocate memory using GFP_NOWAIT, it would leak the reference to the file
descriptor. Also the node pointed to by the xas could be freed between
the call to xas_load() under the rcu_read_lock() and the acquisition of
the xa_lock.
It's easier to just use the normal xa_load/xa_store interface here.
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
[axboe: fix missing assign after alloc, cur_uring -> tctx rename]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We have to drop the lock during each iteration, so there's no advantage
to using the advanced API. Convert this to a standard xa_for_each() loop.
Reported-by: syzbot+27c12725d8ff0bfe1a13@syzkaller.appspotmail.com
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
FUSE servers can indicate crossmount points by setting FUSE_ATTR_SUBMOUNT
in fuse_attr.flags. The inode will then be marked as S_AUTOMOUNT, and the
.d_automount implementation creates a new submount at that location, so
that the submount gets a distinct st_dev value.
Note that all submounts get a distinct superblock and a distinct st_dev
value, so for virtio-fs, even if the same filesystem is mounted more than
once on the host, none of its mount points will have the same st_dev. We
need distinct superblocks because the superblock points to the root node,
but the different host mounts may show different trees (e.g. due to
submounts in some of them, but not in others).
Right now, this behavior is only enabled when fuse_conn.auto_submounts is
set, which is the case only for virtio-fs.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
If a container sets a net namespace specific uniquifier, then use that
in the setclientid/exchangeid process.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
When the user sets a uniquifier, then ensure we copy the string
so that calls to strlen() etc are atomic with calls to snprintf().
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
syzbot reported:
general protection fault, probably for non-canonical address 0xdffffc0000000001: 0000 [#1] PREEMPT SMP KASAN
KASAN: null-ptr-deref in range [0x0000000000000008-0x000000000000000f]
CPU: 0 PID: 6860 Comm: syz-executor835 Not tainted 5.9.0-rc8-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:utf8_casefold+0x43/0x1b0 fs/unicode/utf8-core.c:107
[...]
Call Trace:
f2fs_init_casefolded_name fs/f2fs/dir.c:85 [inline]
__f2fs_setup_filename fs/f2fs/dir.c:118 [inline]
f2fs_prepare_lookup+0x3bf/0x640 fs/f2fs/dir.c:163
f2fs_lookup+0x10d/0x920 fs/f2fs/namei.c:494
__lookup_hash+0x115/0x240 fs/namei.c:1445
filename_create+0x14b/0x630 fs/namei.c:3467
user_path_create fs/namei.c:3524 [inline]
do_mkdirat+0x56/0x310 fs/namei.c:3664
do_syscall_64+0x31/0x70 arch/x86/entry/common.c:46
entry_SYSCALL_64_after_hwframe+0x44/0xa9
[...]
The problem is that an inode has F2FS_CASEFOLD_FL set, but the
filesystem doesn't have the casefold feature flag set, and therefore
super_block::s_encoding is NULL.
Fix this by making sanity_check_inode() reject inodes that have
F2FS_CASEFOLD_FL when the filesystem doesn't have the casefold feature.
Reported-by: syzbot+05139c4039d0679e19ff@syzkaller.appspotmail.com
Fixes: 2c2eb7a300 ("f2fs: Support case-insensitive file name lookups")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@collabora.com>
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
In 32bit system, 64-bits key breaks memory alignment.
This fixes the commit "f2fs: support 64-bits key in f2fs rb-tree node entry".
Reported-by: Nicolas Chauvet <kwizart@gmail.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Colin reports that there's unreachable code, since we only ever break
if ret == 0. This is correct, and is due to a reversed logic condition
in when to break or not.
Break out of the loop if we don't process any task work, in that case
we do want to return -EINTR.
Fixes: af9c1a44f8 ("io_uring: process task work in io_uring_register()")
Reported-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Small conflict around locking in rxrpc_process_event() -
channel_lock moved to bundle in next, while state lock
needs _bh() from net.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The afs filesystem has a lock[*] that it uses to serialise I/O operations
going to the server (vnode->io_lock), as the server will only perform one
modification operation at a time on any given file or directory. This
prevents the the filesystem from filling up all the call slots to a server
with calls that aren't going to be executed in parallel anyway, thereby
allowing operations on other files to obtain slots.
[*] Note that is probably redundant for directories at least since
i_rwsem is used to serialise directory modifications and
lookup/reading vs modification. The server does allow parallel
non-modification ops, however.
When a file truncation op completes, we truncate the in-memory copy of the
file to match - but we do it whilst still holding the io_lock, the idea
being to prevent races with other operations.
However, if writeback starts in a worker thread simultaneously with
truncation (whilst notify_change() is called with i_rwsem locked, writeback
pays it no heed), it may manage to set PG_writeback bits on the pages that
will get truncated before afs_setattr_success() manages to call
truncate_pagecache(). Truncate will then wait for those pages - whilst
still inside io_lock:
# cat /proc/8837/stack
[<0>] wait_on_page_bit_common+0x184/0x1e7
[<0>] truncate_inode_pages_range+0x37f/0x3eb
[<0>] truncate_pagecache+0x3c/0x53
[<0>] afs_setattr_success+0x4d/0x6e
[<0>] afs_wait_for_operation+0xd8/0x169
[<0>] afs_do_sync_operation+0x16/0x1f
[<0>] afs_setattr+0x1fb/0x25d
[<0>] notify_change+0x2cf/0x3c4
[<0>] do_truncate+0x7f/0xb2
[<0>] do_sys_ftruncate+0xd1/0x104
[<0>] do_syscall_64+0x2d/0x3a
[<0>] entry_SYSCALL_64_after_hwframe+0x44/0xa9
The writeback operation, however, stalls indefinitely because it needs to
get the io_lock to proceed:
# cat /proc/5940/stack
[<0>] afs_get_io_locks+0x58/0x1ae
[<0>] afs_begin_vnode_operation+0xc7/0xd1
[<0>] afs_store_data+0x1b2/0x2a3
[<0>] afs_write_back_from_locked_page+0x418/0x57c
[<0>] afs_writepages_region+0x196/0x224
[<0>] afs_writepages+0x74/0x156
[<0>] do_writepages+0x2d/0x56
[<0>] __writeback_single_inode+0x84/0x207
[<0>] writeback_sb_inodes+0x238/0x3cf
[<0>] __writeback_inodes_wb+0x68/0x9f
[<0>] wb_writeback+0x145/0x26c
[<0>] wb_do_writeback+0x16a/0x194
[<0>] wb_workfn+0x74/0x177
[<0>] process_one_work+0x174/0x264
[<0>] worker_thread+0x117/0x1b9
[<0>] kthread+0xec/0xf1
[<0>] ret_from_fork+0x1f/0x30
and thus deadlock has occurred.
Note that whilst afs_setattr() calls filemap_write_and_wait(), the fact
that the caller is holding i_rwsem doesn't preclude more pages being
dirtied through an mmap'd region.
Fix this by:
(1) Use the vnode validate_lock to mediate access between afs_setattr()
and afs_writepages():
(a) Exclusively lock validate_lock in afs_setattr() around the whole
RPC operation.
(b) If WB_SYNC_ALL isn't set on entry to afs_writepages(), trying to
shared-lock validate_lock and returning immediately if we couldn't
get it.
(c) If WB_SYNC_ALL is set, wait for the lock.
The validate_lock is also used to validate a file and to zap its cache
if the file was altered by a third party, so it's probably a good fit
for this.
(2) Move the truncation outside of the io_lock in setattr, using the same
hook as is used for local directory editing.
This requires the old i_size to be retained in the operation record as
we commit the revised status to the inode members inside the io_lock
still, but we still need to know if we reduced the file size.
Fixes: d2ddc776a4 ("afs: Overhaul volume and server record caching and fileserver rotation")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Prior to commit 9fe55eea7e ("Fix race when checking i_size on direct
i/o read"), an unaligned direct read past end of file would trigger EOF,
since generic_file_aio_read detected this read-at-EOF condition and
skipped the direct IO read entirely, returning 0. After that change, the
read now reaches dio_generic, which detects the misalignment and returns
EINVAL.
This consolidates the generic direct-io to follow the same behavior of
filesystems. Apparently, this fix will only affect ocfs2 since other
filesystems do this verification before calling do_blockdev_direct_IO,
with the exception of f2fs, which has the same bug, but is fixed in the
next patch.
it can be verified by a read loop on a file that does a partial read
before EOF (On file that doesn't end at an aligned address). The
following code fails on an unaligned file on filesystems without
prior validation without this patch, but not on btrfs, ext4, and xfs.
while (done < total) {
ssize_t delta = pread(fd, buf + done, total - done, off + done);
if (!delta)
break;
...
}
Fix this regression by moving the misalignment check to after the EOF
check added by commit 74cedf9b6c ("direct-io: Fix negative return from
dio read beyond eof").
Based on a patch by Jamie Liu.
Link: https://lore.kernel.org/r/20201008062620.2928326-4-krisman@collabora.com
Reported-by: Jamie Liu <jamieliu@google.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.com>
Signed-off-by: Jan Kara <jack@suse.cz>
If a DIO read starts past EOF, the kernel won't attempt it, so we don't
need to flush dirty pages before failing the syscall.
Link: https://lore.kernel.org/r/20201008062620.2928326-3-krisman@collabora.com
Suggested-by: Jan Kara <jack@suse.cz>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.com>
Signed-off-by: Jan Kara <jack@suse.cz>
In preparation to resort DIO checks, reduce code duplication of error
handling in do_blockdev_direct_IO.
Link: https://lore.kernel.org/r/20201008062620.2928326-2-krisman@collabora.com
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Identical to how we handle the ctx reference counts, increase by the
batch we're expecting to submit, and handle any slow path residual,
if any. The request alloc-and-issue path is very hot, and this makes
a noticeable difference by avoiding an two atomic incs for each
individual request.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Decode multiple hole and data segments sent by the server, placing
everything directly where they need to go in the xdr pages.
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
We keep things simple for now by only decoding a single hole or data
segment returned by the server, even if they returned more to us.
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
This patch adds client support for decoding a single NFS4_CONTENT_DATA
segment returned by the server. This is the simplest implementation
possible, since it does not account for any hole segments in the reply.
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
The formatting is strange in xfs_trans_mod_dquot, so do a reindent.
Signed-off-by: Kaixu Xia <kaixuxia@tencent.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
If we pass in XFS_QMOPT_{U,G,P}QUOTA flags and different uid/gid/prid
than them currently associated with the inode, the arguments
O_{u,g,p}dqpp shouldn't be NULL, so add the ASSERT for them.
Signed-off-by: Kaixu Xia <kaixuxia@tencent.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Refactor xfs_getfsmap to improve its performance: instead of indirectly
calling a function that copies one record to userspace at a time, create
a shadow buffer in the kernel and copy the whole array once at the end.
On the author's computer, this reduces the runtime on his /home by ~20%.
This also eliminates a deadlock when running GETFSMAP against the
realtime device. The current code locks the rtbitmap to create
fsmappings and copies them into userspace, having not released the
rtbitmap lock. If the userspace buffer is an mmap of a sparse file that
itself resides on the realtime device, the write page fault will recurse
into the fs for allocation, which will deadlock on the rtbitmap lock.
Fixes: 4c934c7dd6 ("xfs: report realtime space information via the rtbitmap")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
If userspace asked fsmap to count the number of entries, we cannot
return more than UINT_MAX entries because fmh_entries is u32.
Therefore, stop counting if we hit this limit or else we will waste time
to return truncated results.
Fixes: e89c041338 ("xfs: implement the GETFSMAP ioctl")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Now that we have the ability to ask the log how far the tail needs to be
pushed to maintain its free space targets, augment the decision to relog
an intent item so that we only do it if the log has hit the 75% full
threshold. There's no point in relogging an intent into the same
checkpoint, and there's no need to relog if there's plenty of free space
in the log.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Separate the computation of the log push threshold and the push logic in
xlog_grant_push_ail. This enables higher level code to determine (for
example) that it is holding on to a logged intent item and the log is so
busy that it is more than 75% full. In that case, it would be desirable
to move the log item towards the head to release the tail, which we will
cover in the next patch.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
There's a subtle design flaw in the deferred log item code that can lead
to pinning the log tail. Taking up the defer ops chain examples from
the previous commit, we can get trapped in sequences like this:
Caller hands us a transaction t0 with D0-D3 attached. The defer ops
chain will look like the following if the transaction rolls succeed:
t1: D0(t0), D1(t0), D2(t0), D3(t0)
t2: d4(t1), d5(t1), D1(t0), D2(t0), D3(t0)
t3: d5(t1), D1(t0), D2(t0), D3(t0)
...
t9: d9(t7), D3(t0)
t10: D3(t0)
t11: d10(t10), d11(t10)
t12: d11(t10)
In transaction 9, we finish d9 and try to roll to t10 while holding onto
an intent item for D3 that we logged in t0.
The previous commit changed the order in which we place new defer ops in
the defer ops processing chain to reduce the maximum chain length. Now
make xfs_defer_finish_noroll capable of relogging the entire chain
periodically so that we can always move the log tail forward. Most
chains will never get relogged, except for operations that generate very
long chains (large extents containing many blocks with different sharing
levels) or are on filesystems with small logs and a lot of ongoing
metadata updates.
Callers are now required to ensure that the transaction reservation is
large enough to handle logging done items and new intent items for the
maximum possible chain length. Most callers are careful to keep the
chain lengths low, so the overhead should be minimal.
The decision to relog an intent item is made based on whether the intent
was logged in a previous checkpoint, since there's no point in relogging
an intent into the same checkpoint.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
The defer ops code has been finishing items in the wrong order -- if a
top level defer op creates items A and B, and finishing item A creates
more defer ops A1 and A2, we'll put the new items on the end of the
chain and process them in the order A B A1 A2. This is kind of weird,
since it's convenient for programmers to be able to think of A and B as
an ordered sequence where all the sub-tasks for A must finish before we
move on to B, e.g. A A1 A2 D.
Right now, our log intent items are not so complex that this matters,
but this will become important for the atomic extent swapping patchset.
In order to maintain correct reference counting of extents, we have to
unmap and remap extents in that order, and we want to complete that work
before moving on to the next range that the user wants to swap. This
patch fixes defer ops to satsify that requirement.
The primary symptom of the incorrect order was noticed in an early
performance analysis of the atomic extent swap code. An astonishingly
large number of deferred work items accumulated when userspace requested
an atomic update of two very fragmented files. The cause of this was
traced to the same ordering bug in the inner loop of
xfs_defer_finish_noroll.
If the ->finish_item method of a deferred operation queues new deferred
operations, those new deferred ops are appended to the tail of the
pending work list. To illustrate, say that a caller creates a
transaction t0 with four deferred operations D0-D3. The first thing
defer ops does is roll the transaction to t1, leaving us with:
t1: D0(t0), D1(t0), D2(t0), D3(t0)
Let's say that finishing each of D0-D3 will create two new deferred ops.
After finish D0 and roll, we'll have the following chain:
t2: D1(t0), D2(t0), D3(t0), d4(t1), d5(t1)
d4 and d5 were logged to t1. Notice that while we're about to start
work on D1, we haven't actually completed all the work implied by D0
being finished. So far we've been careful (or lucky) to structure the
dfops callers such that D1 doesn't depend on d4 or d5 being finished,
but this is a potential logic bomb.
There's a second problem lurking. Let's see what happens as we finish
D1-D3:
t3: D2(t0), D3(t0), d4(t1), d5(t1), d6(t2), d7(t2)
t4: D3(t0), d4(t1), d5(t1), d6(t2), d7(t2), d8(t3), d9(t3)
t5: d4(t1), d5(t1), d6(t2), d7(t2), d8(t3), d9(t3), d10(t4), d11(t4)
Let's say that d4-d11 are simple work items that don't queue any other
operations, which means that we can complete each d4 and roll to t6:
t6: d5(t1), d6(t2), d7(t2), d8(t3), d9(t3), d10(t4), d11(t4)
t7: d6(t2), d7(t2), d8(t3), d9(t3), d10(t4), d11(t4)
...
t11: d10(t4), d11(t4)
t12: d11(t4)
<done>
When we try to roll to transaction #12, we're holding defer op d11,
which we logged way back in t4. This means that the tail of the log is
pinned at t4. If the log is very small or there are a lot of other
threads updating metadata, this means that we might have wrapped the log
and cannot get roll to t11 because there isn't enough space left before
we'd run into t4.
Let's shift back to the original failure. I mentioned before that I
discovered this flaw while developing the atomic file update code. In
that scenario, we have a defer op (D0) that finds a range of file blocks
to remap, creates a handful of new defer ops to do that, and then asks
to be continued with however much work remains.
So, D0 is the original swapext deferred op. The first thing defer ops
does is rolls to t1:
t1: D0(t0)
We try to finish D0, logging d1 and d2 in the process, but can't get all
the work done. We log a done item and a new intent item for the work
that D0 still has to do, and roll to t2:
t2: D0'(t1), d1(t1), d2(t1)
We roll and try to finish D0', but still can't get all the work done, so
we log a done item and a new intent item for it, requeue D0 a second
time, and roll to t3:
t3: D0''(t2), d1(t1), d2(t1), d3(t2), d4(t2)
If it takes 48 more rolls to complete D0, then we'll finally dispense
with D0 in t50:
t50: D<fifty primes>(t49), d1(t1), ..., d102(t50)
We then try to roll again to get a chain like this:
t51: d1(t1), d2(t1), ..., d101(t50), d102(t50)
...
t152: d102(t50)
<done>
Notice that in rolling to transaction #51, we're holding on to a log
intent item for d1 that was logged in transaction #1. This means that
the tail of the log is pinned at t1. If the log is very small or there
are a lot of other threads updating metadata, this means that we might
have wrapped the log and cannot roll to t51 because there isn't enough
space left before we'd run into t1. This is of course problem #2 again.
But notice the third problem with this scenario: we have 102 defer ops
tied to this transaction! Each of these items are backed by pinned
kernel memory, which means that we risk OOM if the chains get too long.
Yikes. Problem #1 is a subtle logic bomb that could hit someone in the
future; problem #2 applies (rarely) to the current upstream, and problem
#3 applies to work under development.
This is not how incremental deferred operations were supposed to work.
The dfops design of logging in the same transaction an intent-done item
and a new intent item for the work remaining was to make it so that we
only have to juggle enough deferred work items to finish that one small
piece of work. Deferred log item recovery will find that first
unfinished work item and restart it, no matter how many other intent
items might follow it in the log. Therefore, it's ok to put the new
intents at the start of the dfops chain.
For the first example, the chains look like this:
t2: d4(t1), d5(t1), D1(t0), D2(t0), D3(t0)
t3: d5(t1), D1(t0), D2(t0), D3(t0)
...
t9: d9(t7), D3(t0)
t10: D3(t0)
t11: d10(t10), d11(t10)
t12: d11(t10)
For the second example, the chains look like this:
t1: D0(t0)
t2: d1(t1), d2(t1), D0'(t1)
t3: d2(t1), D0'(t1)
t4: D0'(t1)
t5: d1(t4), d2(t4), D0''(t4)
...
t148: D0<50 primes>(t147)
t149: d101(t148), d102(t148)
t150: d102(t148)
<done>
This actually sucks more for pinning the log tail (we try to roll to t10
while holding an intent item that was logged in t1) but we've solved
problem #1. We've also reduced the maximum chain length from:
sum(all the new items) + nr_original_items
to:
max(new items that each original item creates) + nr_original_items
This solves problem #3 by sharply reducing the number of defer ops that
can be attached to a transaction at any given time. The change makes
the problem of log tail pinning worse, but is improvement we need to
solve problem #2. Actually solving #2, however, is left to the next
patch.
Note that a subsequent analysis of some hard-to-trigger reflink and COW
livelocks on extremely fragmented filesystems (or systems running a lot
of IO threads) showed the same symptoms -- uncomfortably large numbers
of incore deferred work items and occasional stalls in the transaction
grant code while waiting for log reservations. I think this patch and
the next one will also solve these problems.
As originally written, the code used list_splice_tail_init instead of
list_splice_init, so change that, and leave a short comment explaining
our actions.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
In xfs_bui_item_recover, there exists a use-after-free bug with regards
to the inode that is involved in the bmap replay operation. If the
mapping operation does not complete, we call xfs_bmap_unmap_extent to
create a deferred op to finish the unmapping work, and we retain a
pointer to the incore inode.
Unfortunately, the very next thing we do is commit the transaction and
drop the inode. If reclaim tears down the inode before we try to finish
the defer ops, we dereference garbage and blow up. Therefore, create a
way to join inodes to the defer ops freezer so that we can maintain the
xfs_inode reference until we're done with the inode.
Note: This imposes the requirement that there be enough memory to keep
every incore inode in memory throughout recovery.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
In most places in XFS, we have a specific order in which we gather
resources: grab the inode, allocate a transaction, then lock the inode.
xfs_bui_item_recover doesn't do it in that order, so fix it to be more
consistent. This also makes the error bailout code a bit less weird.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Brian Foster <bfoster@redhat.com>
The bmap intent item checking code in xfs_bui_item_recover is spread all
over the function. We should check the recovered log item at the top
before we allocate any resources or do anything else, so do that.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
When xfs_defer_capture extracts the deferred ops and transaction state
from a transaction, it should record the transaction reservation type
from the old transaction so that when we continue the dfops chain, we
still use the same reservation parameters.
Doing this means that the log item recovery functions get to determine
the transaction reservation instead of abusing tr_itruncate in yet
another part of xfs.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
When xfs_defer_capture extracts the deferred ops and transaction state
from a transaction, it should record the remaining block reservations so
that when we continue the dfops chain, we can reserve the same number of
blocks to use. We capture the reservations for both data and realtime
volumes.
This adds the requirement that every log intent item recovery function
must be careful to reserve enough blocks to handle both itself and all
defer ops that it can queue. On the other hand, this enables us to do
away with the handwaving block estimation nonsense that was going on in
xlog_finish_defer_ops.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Brian Foster <bfoster@redhat.com>
When we replay unfinished intent items that have been recovered from the
log, it's possible that the replay will cause the creation of more
deferred work items. As outlined in commit 509955823c ("xfs: log
recovery should replay deferred ops in order"), later work items have an
implicit ordering dependency on earlier work items. Therefore, recovery
must replay the items (both recovered and created) in the same order
that they would have been during normal operation.
For log recovery, we enforce this ordering by using an empty transaction
to collect deferred ops that get created in the process of recovering a
log intent item to prevent them from being committed before the rest of
the recovered intent items. After we finish committing all the
recovered log items, we allocate a transaction with an enormous block
reservation, splice our huge list of created deferred ops into that
transaction, and commit it, thereby finishing all those ops.
This is /really/ hokey -- it's the one place in XFS where we allow
nested transactions; the splicing of the defer ops list is is inelegant
and has to be done twice per recovery function; and the broken way we
handle inode pointers and block reservations cause subtle use-after-free
and allocator problems that will be fixed by this patch and the two
patches after it.
Therefore, replace the hokey empty transaction with a structure designed
to capture each chain of deferred ops that are created as part of
recovering a single unfinished log intent. Finally, refactor the loop
that replays those chains to do so using one transaction per chain.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
The ->iop_recover method of a log intent item removes the recovered
intent item from the AIL by logging an intent done item and committing
the transaction, so it's superfluous to have this flag check. Nothing
else uses it, so get rid of the flag entirely.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Remove this one-line helper since the assert is trivially true in one
call site and the rest obscures a bitmask operation.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Commit 8d875f95da ("btrfs: disable strict file flushes for
renames and truncates") eliminated the notion of ordered operations and
instead BTRFS_INODE_ORDERED_DATA_CLOSE only remained as a flag
indicating that a file's content should be synced to disk in case a
file is truncated and any writes happen to it concurrently. In fact
this intendend behavior was broken until it was fixed in
f6dc45c7a9 ("Btrfs: fix filemap_flush call in btrfs_file_release").
All things considered let's give the flag a more descriptive name. Also
slightly reword comments.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This patch fixes the following sparse errors in
fs/btrfs/super.c in function btrfs_show_devname()
fs/btrfs/super.c: error: incompatible types in comparison expression (different address spaces):
fs/btrfs/super.c: struct rcu_string [noderef] <asn:4> *
fs/btrfs/super.c: struct rcu_string *
The error was because of the following line in function btrfs_show_devname():
if (first_dev)
seq_escape(m, rcu_str_deref(first_dev->name), " \t\n\\");
Annotating the btrfs_device::name member with __rcu fixes the sparse
error.
Acked-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Madhuparna Bhowmik <madhuparnabhowmik04@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Many things can happen after the device is scanned and before the device
is mounted. One such thing is losing the BTRFS_MAGIC on the device.
If it happens we still won't free that device from the memory and cause
the userland confusion.
For example: As the BTRFS_IOC_DEV_INFO still carries the device path
which does not have the BTRFS_MAGIC, 'btrfs fi show' still lists
device which does not belong to the filesystem anymore:
$ mkfs.btrfs -fq -draid1 -mraid1 /dev/sda /dev/sdb
$ wipefs -a /dev/sdb
# /dev/sdb does not contain magic signature
$ mount -o degraded /dev/sda /btrfs
$ btrfs fi show -m
Label: none uuid: 470ec6fb-646b-4464-b3cb-df1b26c527bd
Total devices 2 FS bytes used 128.00KiB
devid 1 size 3.00GiB used 571.19MiB path /dev/sda
devid 2 size 3.00GiB used 571.19MiB path /dev/sdb
We need to distinguish the missing signature and invalid superblock, so
add a specific error code ENODATA for that. This also fixes failure of
fstest btrfs/198.
CC: stable@vger.kernel.org # 4.19+
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In fstest btrfs/064 a transaction abort in __btrfs_cow_block could lead
to a system lockup. It gets stuck trying to write back inodes, and the
write back thread was trying to lock an extent buffer:
$ cat /proc/2143497/stack
[<0>] __btrfs_tree_lock+0x108/0x250
[<0>] lock_extent_buffer_for_io+0x35e/0x3a0
[<0>] btree_write_cache_pages+0x15a/0x3b0
[<0>] do_writepages+0x28/0xb0
[<0>] __writeback_single_inode+0x54/0x5c0
[<0>] writeback_sb_inodes+0x1e8/0x510
[<0>] wb_writeback+0xcc/0x440
[<0>] wb_workfn+0xd7/0x650
[<0>] process_one_work+0x236/0x560
[<0>] worker_thread+0x55/0x3c0
[<0>] kthread+0x13a/0x150
[<0>] ret_from_fork+0x1f/0x30
This is because we got an error while COWing a block, specifically here
if (test_bit(BTRFS_ROOT_SHAREABLE, &root->state)) {
ret = btrfs_reloc_cow_block(trans, root, buf, cow);
if (ret) {
btrfs_abort_transaction(trans, ret);
return ret;
}
}
[16402.241552] BTRFS: Transaction aborted (error -2)
[16402.242362] WARNING: CPU: 1 PID: 2563188 at fs/btrfs/ctree.c:1074 __btrfs_cow_block+0x376/0x540
[16402.249469] CPU: 1 PID: 2563188 Comm: fsstress Not tainted 5.9.0-rc6+ #8
[16402.249936] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-2.fc32 04/01/2014
[16402.250525] RIP: 0010:__btrfs_cow_block+0x376/0x540
[16402.252417] RSP: 0018:ffff9cca40e578b0 EFLAGS: 00010282
[16402.252787] RAX: 0000000000000025 RBX: 0000000000000002 RCX: ffff9132bbd19388
[16402.253278] RDX: 00000000ffffffd8 RSI: 0000000000000027 RDI: ffff9132bbd19380
[16402.254063] RBP: ffff9132b41a49c0 R08: 0000000000000000 R09: 0000000000000000
[16402.254887] R10: 0000000000000000 R11: ffff91324758b080 R12: ffff91326ef17ce0
[16402.255694] R13: ffff91325fc0f000 R14: ffff91326ef176b0 R15: ffff9132815e2000
[16402.256321] FS: 00007f542c6d7b80(0000) GS:ffff9132bbd00000(0000) knlGS:0000000000000000
[16402.256973] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[16402.257374] CR2: 00007f127b83f250 CR3: 0000000133480002 CR4: 0000000000370ee0
[16402.257867] Call Trace:
[16402.258072] btrfs_cow_block+0x109/0x230
[16402.258356] btrfs_search_slot+0x530/0x9d0
[16402.258655] btrfs_lookup_file_extent+0x37/0x40
[16402.259155] __btrfs_drop_extents+0x13c/0xd60
[16402.259628] ? btrfs_block_rsv_migrate+0x4f/0xb0
[16402.259949] btrfs_replace_file_extents+0x190/0x820
[16402.260873] btrfs_clone+0x9ae/0xc00
[16402.261139] btrfs_extent_same_range+0x66/0x90
[16402.261771] btrfs_remap_file_range+0x353/0x3b1
[16402.262333] vfs_dedupe_file_range_one.part.0+0xd5/0x140
[16402.262821] vfs_dedupe_file_range+0x189/0x220
[16402.263150] do_vfs_ioctl+0x552/0x700
[16402.263662] __x64_sys_ioctl+0x62/0xb0
[16402.264023] do_syscall_64+0x33/0x40
[16402.264364] entry_SYSCALL_64_after_hwframe+0x44/0xa9
[16402.264862] RIP: 0033:0x7f542c7d15cb
[16402.266901] RSP: 002b:00007ffd35944ea8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[16402.267627] RAX: ffffffffffffffda RBX: 00000000009d1968 RCX: 00007f542c7d15cb
[16402.268298] RDX: 00000000009d2490 RSI: 00000000c0189436 RDI: 0000000000000003
[16402.268958] RBP: 00000000009d2520 R08: 0000000000000036 R09: 00000000009d2e64
[16402.269726] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000002
[16402.270659] R13: 000000000001f000 R14: 00000000009d1970 R15: 00000000009d2e80
[16402.271498] irq event stamp: 0
[16402.271846] hardirqs last enabled at (0): [<0000000000000000>] 0x0
[16402.272497] hardirqs last disabled at (0): [<ffffffff910dbf59>] copy_process+0x6b9/0x1ba0
[16402.273343] softirqs last enabled at (0): [<ffffffff910dbf59>] copy_process+0x6b9/0x1ba0
[16402.273905] softirqs last disabled at (0): [<0000000000000000>] 0x0
[16402.274338] ---[ end trace 737874a5a41a8236 ]---
[16402.274669] BTRFS: error (device dm-9) in __btrfs_cow_block:1074: errno=-2 No such entry
[16402.276179] BTRFS info (device dm-9): forced readonly
[16402.277046] BTRFS: error (device dm-9) in btrfs_replace_file_extents:2723: errno=-2 No such entry
[16402.278744] BTRFS: error (device dm-9) in __btrfs_cow_block:1074: errno=-2 No such entry
[16402.279968] BTRFS: error (device dm-9) in __btrfs_cow_block:1074: errno=-2 No such entry
[16402.280582] BTRFS info (device dm-9): balance: ended with status: -30
The problem here is that as soon as we allocate the new block it is
locked and marked dirty in the btree inode. This means that we could
attempt to writeback this block and need to lock the extent buffer.
However we're not unlocking it here and thus we deadlock.
Fix this by unlocking the cow block if we have any errors inside of
__btrfs_cow_block, and also free it so we do not leak it.
CC: stable@vger.kernel.org # 4.4+
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Since we now perform direct reads using i_rwsem, we can remove this
inode flag used to co-ordinate unlocked reads.
The truncate call takes i_rwsem. This means it is correctly synchronized
with concurrent direct reads.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Johannes Thumshirn <jth@kernel.org>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
