Mitch kept hitting a panic because he was getting ENOSPC. One of my previous
patches makes it so we are much better at not allocating new metadata chunks.
Unfortunately coupled with the overcommit patch this works us into a bit of a
problem if we are removing a bunch of space and end up chewing up all of our
space with pinned extents. We can allocate chunks fine and overflow is ok, but
the only way to reclaim this space is to commit the transaction. So if we go to
overcommit, first check and see how much pinned space we have. If we have more
than 80% of the free space chewed up with pinned extents, just commit the
transaction, this will free up enough space for our reservation and we won't
have this problem anymore. With this patch Mitch's test doesn't blow up
anymore. Thanks,
Reported-and-tested-by: Mitch Harder <mitch.harder@sabayonlinux.org>
Signed-off-by: Josef Bacik <josef@redhat.com>
Currently btrfs_block_rsv_check does 2 things, it will either refill a block
reserve like in the truncate or refill case, or it will check to see if there is
enough space in the global reserve and possibly refill it. However because of
overcommit we could be well overcommitting ourselves just to try and refill the
global reserve, when really we should just be committing the transaction. So
breack this out into btrfs_block_rsv_refill and btrfs_block_rsv_check. Refill
will try to reserve more metadata if it can and btrfs_block_rsv_check will not,
it will only tell you if the factor of the total space is still reserved.
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
In __unlink_start_trans() if we don't have enough room for a reservation we will
check to see if the unlink will free up space. If it does that's great, but we
will still could add an orphan item, so we need to reserve enough space to add
the orphan item. Do this and migrate the space the global reserve so it all
works out right. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
We started setting trans->block_rsv = NULL to allow the delayed refs flushing
stuff to use the right block_rsv and then just made
btrfs_trans_release_metadata() unconditionally use the trans block rsv. The
problem with this is we need to reserve some space in the transaction and then
migrate it to the global block rsv, so we need to be able to free that out
properly. So instead just move btrfs_trans_release_metadata() before the
delayed ref flushing and use trans->block_rsv for the freeing. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Currently we only allow a maximum of 2 megabytes of pages to be flushed at a
time. This was ok before, but now we have overcommit which will screw us in a
heartbeat if we are quickly filling the disk. So instead pick either 2
megabytes or the number of pages we need to reclaim to be safe again, which ever
is larger. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
The only way we actually reclaim delalloc space is waiting for the IO to
completely finish. Usually we kick off a bunch of IO and wait for a little bit
and hope we can make our reservation, and usually this works out pretty well.
With overcommit however we can get seriously underwater if we're filling up the
disk quickly, so we need to be able to force the delalloc shrinker to wait for
the ordered IO to finish to give us a better chance of actually reclaiming
enough space to get our reservation. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Before the only reason to commit the transaction to recover space in
reserve_metadata_bytes() was if there were enough pinned_bytes to satisfy our
reservation. But now we have the delayed inode stuff which will hold it's
reservations until we commit the transaction. So say we max out our reservation
by creating a bunch of files but don't have any pinned bytes we will ENOSPC out
early even though we could commit the transaction and get that space back. So
now just unconditionally commit the transaction since currently there is no way
to know how much metadata space is being reserved by delayed inode stuff.
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Recently I changed the xattr stuff to unconditionally set the xattr first in
case the xattr didn't exist yet. This has introduced a regression when setting
an xattr that already exists with a large value. If we find the key we are
looking for split_leaf will assume that we're extending that item. The problem
is the size we pass down to btrfs_search_slot includes the size of the item
already, so if we have the largest xattr we can possibly have plus the size of
the xattr item plus the xattr item that btrfs_search_slot we'd overflow the
leaf. Thankfully this is not what we're doing, but split_leaf doesn't know this
so it just returns EOVERFLOW. So in the xattr code we need to check and see if
we got back EOVERFLOW and treat it like EEXIST since that's really what
happened. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Our unlink reservations were a bit much, we were reserving 10 and I only count 8
possible items we're touching, so comment what we're reserving for and fix the
count value. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
I noticed recently that my overcommit patch was causing one of my enospc tests
to fail 25% of the time with early ENOSPC. This is because my overcommit patch
was letting us go way over board, but it wasn't waiting long enough to let the
delalloc shrinker do it's job. The problem is we just start writeback and wait
a little bit hoping we flush enough, but we only free up delalloc space by
having the writes complete all the way. We do this by waiting for ordered
extents, which we do but only if we already free'd enough for the reservation,
which isn't right, we should flush ordered extents if we didn't reclaim enough
in case that will push us over the edge. With this patch I've not seen a
failure in this enospc test after running it in a loop for an hour. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Yeah yeah I know this is how we used to do it and then I changed it, but damnit
I'm changing it back. The fact is that writing out checksums will modify
metadata, which could cause us to dirty a block group we've already written out,
so we have to truncate it and all of it's checksums and re-write it which will
write new checksums which could dirty a blockg roup that has already been
written and you see where I'm going with this? This can cause unmount or really
anything that depends on a transaction to commit to take it's sweet damned time
to happen. So go back to the way it was, only this time we're specifically
setting NODATACOW because we can't go through the COW pathway anyway and we're
doing our own built-in cow'ing by truncating the free space cache. The other
new thing is once we truncate the old cache and preallocate the new space, we
don't need to do that song and dance at all for the rest of the transaction, we
can just overwrite the existing space with the new cache if the block group
changes for whatever reason, and the NODATACOW will let us do this fine. So
keep track of which transaction we last cleared our cache in and if we cleared
it in this transaction just say we're all setup and carry on. This survives
xfstests and stress.sh.
The inode cache will continue to use the normal csum infrastructure since it
only gets written once and there will be no more modifications to the fs tree in
a transaction commit.
Signed-off-by: Josef Bacik <josef@redhat.com>
My overcommit stuff can be a little racy when we're filling up the disk with
fs_mark and we overcommit into things that quickly get used up for data. So use
num_bytes to see if we have enough available space so we're less likely to
overcommit ourselves out of the ability to make reservations. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
We need to check the return value of filemap_write_and_wait in the space cache
writeout code. Also don't set the inode's generation until we're sure nothing
else is going to fail. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
In writing and reading the space cache we have one big loop that keeps track of
which page we are on and then a bunch of sizeable loops underneath this big loop
to try and read/write out properly. Especially in the write case this makes
things hugely complicated and hard to follow, and makes our error checking and
recovery equally as complex. So add a io_ctl struct with a bunch of helpers to
keep track of the pages we have, where we are, if we have enough space etc.
This unifies how we deal with the pages we're writing and keeps all the messy
tracking internal. This allows us to kill the big loops in both the read and
write case and makes reviewing and chaning the write and read paths much
simpler. I've run xfstests and stress.sh on this code and it survives. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
I noticed a slight bug where we will not bother writing out the block group
cache's space cache if it's space tree is empty. Since it could have a cluster
or pinned extents that need to be written out this is just not a valid test.
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Some users have requested this and I've found I needed a way to disable cache
loading without actually clearing the cache, so introduce the no_space_cache
option. Before we check the super blocks cache generation field and if it was
populated we always turned space caching on. Now we check this and set the
space cache option on, and then parse the mount options so that if we want it
off it get's turned off. Then we check the mount option all the places we do
the caching work instead of checking the super's cache generation. This makes
things more consistent and lets us turn space caching off. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Xfstests 79 was failing because we were inheriting the S_APPEND flag when we
weren't supposed to. There isn't any specific documentation on this so I'm
taking the test as the standard of how things work, and having S_APPEND set on a
directory doesn't mean that S_APPEND gets inherited by its children according to
this test. So only inherit btrfs specific things. This will let us set
compress/nocompress on specific directories and everything in the directories
will inherit this flag, same with nodatacow. With this patch test 79 passes.
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
One of the things that kills us is the fact that our ENOSPC reservations are
horribly over the top in most normal cases. There isn't too much that can be
done about this because when we are completely full we really need them to work
like this so we don't under reserve. However if there is plenty of unallocated
chunks on the disk we can use that to gauge how much we can overcommit. So this
patch adds chunk free space accounting so we always know how much unallocated
space we have. Then if we fail to make a reservation within our allocated
space, check to see if we can overcommit. In the normal flushing case (like
with delalloc metadata reservations) we'll take the free space and divide it by
2 if our metadata profile is setup for DUP or any of those, and then divide it
by 8 to make sure we don't overcommit too much. Then if we're in a non-flushing
case (we really need this reservation now!) we only limit ourselves to half of
the free space. This makes this fio test
[torrent]
filename=torrent-test
rw=randwrite
size=4g
ioengine=sync
directory=/mnt/btrfs-test
go from taking around 45 minutes to 10 seconds on my freshly formatted 3 TiB
file system. This doesn't seem to break my other enospc tests, but could really
use some more testing as this is a super scary change. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
I noticed while running xfstests 83 that if we didn't have enough space to
delete our inode the orphan cleanup would just loop. This is because it keeps
finding the same orphan item and keeps trying to kill it but can't because we
don't get an error back from iput for deleting the inode. So keep track of the
last guy we tried to kill, if it's the same as the one we're trying to kill
currently we know we are having problems and can just error out. I don't have a
way to test this so look hard and make sure it's right. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Xfstests 83 really stresses our ENOSPC since it uses a 100mb fs which ends up
with the mixed block group stuff. Because of this we can run into a situation
where we don't have enough space to delete inodes, or even worse we can't free
the inodes when we next mount the fs which causes the orphan code to lose its
mind. So if we fail to make our reservation, steal from the global reserve.
The global reserve will end up taking up the entire rest of the free space on
the fs in this worst case so there really is no other option. With this patch
test 83 doesn't freak out. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
While looking for a performance regression a user was complaining about, I
noticed that we had a regression with the varmail test of filebench. This was
introduced by
0d10ee2e6d
which keeps us from calling writepages in writepage. This is a correct change,
however it happens to help the varmail test because we write out in larger
chunks. This is largly to do with how we write out dirty pages for each
transaction. If you run filebench with
load varmail
set $dir=/mnt/btrfs-test
run 60
prior to this patch you would get ~1420 ops/second, but with the patch you get
~1200 ops/second. This is a 16% decrease. So since we know the range of dirty
pages we want to write out, don't write out in one page chunks, write out in
ranges. So to do this we call filemap_fdatawrite_range() on the range of bytes.
Then we convert the DIRTY extents to NEED_WAIT extents. When we then call
btrfs_wait_marked_extents() we only have to filemap_fdatawait_range() on that
range and clear the NEED_WAIT extents. This doesn't get us back to our original
speeds, but I've been seeing ~1380 ops/second, which is a <5% regression as
opposed to a >15% regression. That is acceptable given that the original commit
greatly reduces our latency to begin with. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
If I have a range where I know a certain bit is and I want to set it to another
bit the only option I have is to call set and then clear bit, which will result
in 2 tree searches. This is inefficient, so introduce convert_extent_bit which
will go through and set the bit I want and clear the old bit I don't want.
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
There is a bug that may lead to early ENOSPC in our reservation code. We've
been checking against num_bytes which may be above and beyond what we want to
actually reserve, which could give us a false ENOSPC. Fix this by making sure
the unused space is above how much we want to reserve and not how much we're
trying to flush. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
In fixing how we deal with bad inodes, we had a regression in the orphan cleanup
code, since it expects to get a bad inode back. So fix it to deal with getting
-ESTALE back by deleting the orphan item manually and moving on. Thanks,
Reported-by: Simon Kirby <sim@hostway.ca>
Signed-off-by: Josef Bacik <josef@redhat.com>
Johannes pointed out we were allocating only kernel pages for doing writes,
which is kind of a big deal if you are on 32bit and have more than a gig of ram.
So fix our allocations to use the mapping's gfp but still clear __GFP_FS so we
don't re-enter. Thanks,
Reported-by: Johannes Weiner <jweiner@redhat.com>
Signed-off-by: Josef Bacik <josef@redhat.com>
I kept getting warnings from evict because we were calling
btrfs_start_transaction() with a transaction already started when doing a
balance. This is because we remove a block group which requires a transaction,
and the put the last reference on the cache inode. Instead of doing this we
need to delay the iput so it is done not within a transaction having started.
This gets rid of our warnings. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Checksums are charged in 2 different ways. The first case is when we're writing
to the disk, we account for the new checksums with the delalloc block rsv. In
order for this to work we check if we're allocating a block for the csum root
and if trans->block_rsv == the delalloc block rsv. But when we're deleting the
csums because of cow, this is charged to the global block rsv, and is done when
we run the delayed refs. So we need to make sure that trans->block_rsv == NULL
when running the delayed refs. So set it to NULL and reset it in
should_end_transaction, and set it to NULL in commit_transaction. This got rid
of the ridiculous amount of warnings I was seeing when trying to do a balance.
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
The only thing that we need to have a trans handle for is in
reserve_metadata_bytes and thats to know how much flushing we can do. So
instead of passing it around, just check current->journal_info for a
trans_handle so we know if we can commit a transaction to try and free up space
or not. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Since the durable block rsv stuff has been killed there is no need to get the
block_rsv in btrfs_free_tree_block anymore.
Signed-off-by: Josef Bacik <josef@redhat.com>
The alloc warnings everybody has been seeing is because we have been reserving
space for csums, but we weren't actually using that space. So make
get_block_rsv() return the trans->block_rsv if we're modifying the csum root.
Also set the trans->block_rsv to NULL so that if we modify the csum root when
running delayed ref's that comes out of the global reserve like it's supposed
to. With this patch I'm not seeing those alloc warnings anymore. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Since free space inodes now use normal checksumming we need to make sure to
account for their metadata use. So reserve metadata space, and then if we fail
to write out the metadata we can just release it, otherwise it will be freed up
when the io completes. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
In moving some enospc stuff around I noticed that when we unmount we are often
evicting the free space cache inodes before we do our last commit. This isn't
bad, but it makes us constantly have to re-read the inodes back. So instead
don't evict the cache until after we do our last commit, this will make things a
little less crappy and makes a future enospc change work properly. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
While debugging a different issue I noticed that we were always reserving space
when we tried to use our truncate block rsv's. This is because they didn't have
a ->size value, so use_block_rsv just assumes there is nothing reserved and it
does a reserve_metadata_bytes. This is because btrfs_check_block_rsv() doesn't
actually add to the size of the block rsv. That seems to be the right thing to
do so set ->size to the minimum truncate size we need, since we will always only
refill to that size anyway, and this way everything works out correctly.
Signed-off-by: Josef Bacik <josef@redhat.com>
If we have to emergency reserve space we need to not increase the block_rsv
size, otherwise we'll leak space. Take for instance delalloc, say we reserve
4k, and we use that 4k, and then we have to emergency allocate another 4k, we
bump the size up to 8k, however we've only accounted for 4k in reservations in
all of our supporting logic, so we'll go to free the 4k and end up having a size
of 4k, which will cause us to later not free as much space. I saw this doing
testing where I wasn't reserving enough space for something but was still
leaking space, very frustrating. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
When changing back to using a spin_lock to protect the extent counters I decided
that since we would only be dropping our original extent, it was ok to just drop
the extent and return. However since somebody else could have come in and done
a reservation, we need to do the normal song and dance to clear the reservation
out properly. So calculate how much space we need to free, and then subtract
what we just attempted to reserve. If it's more then we know we need to drop
those bytes from the delalloc block rsv. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
We are setting ins_len to 1 even tho we are just modifying an item that should
be there already. This may cause the search stuff to split nodes on the way
down needelessly. Set this to 0 since we aren't inserting anything. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
If you run xfstest 224 it you will get lots of messages about not being able to
delete inodes and that they will be cleaned up next mount. This is because
btrfs_block_rsv_check was not calling reserve_metadata_bytes with the ability to
flush, so if there was not enough space, it simply failed. But in truncate and
evict case we could easily flush space to try and get enough space to do our
work, so make btrfs_block_rsv_check take a flush argument to pass down to
reserve_metadata_bytes. Now xfstests 224 runs fine without all those
complaints. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
With btrfs_truncate_inode_items we always return if we have to go to another
leaf, which makes us do our reservation again. This means we will only ever
modify one leaf at a time, so we only need 1 items worth of slack space. Also,
since we are deleting we will not be creating nodes as we go down, if anything
we'll be free'ing them as we merge them together, so make a different
calculation for truncate which will only have the worst case useage of COW'ing
the entire path down to the leaf. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Lukas found a problem where if he tries to fallocate over the same region twice
and the first fallocate took up all the space we would fail with ENOSPC. This
is because we reserve the total space we want to use for fallocate, regardless
of wether or not we will have to actually preallocate. So instead move the
check into the loop where we actually have to do the preallocate. Thanks,
Tested-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: Josef Bacik <josef@redhat.com>
Currently we're starting and stopping a transaction for no real reason, so kill
that and just reserve enough space as if we can truncate all in one transaction.
Also use btrfs_block_rsv_check() for our reserve to minimize the amount of space
we may have to allocate for our slack space. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
We will try and reserve metadata bytes in btrfs_block_rsv_check and if we cannot
because we have a transaction open it will return EAGAIN, so we do not need to
try and commit the transaction again.
Signed-off-by: Josef Bacik <josef@redhat.com>
The priority and refill_used flags are not used anymore, and neither is the
usage counter, so just remove them from btrfs_block_rsv.
Signed-off-by: Josef Bacik <josef@redhat.com>
A user reported getting spammed when moving to 3.0 by this message. Since we
switched to the normal checksumming infrastructure all old free space caches
will be wrong and need to be regenerated so people are likely to see this
message a lot, so ratelimit it so it doesn't fill up their logs and freak them
out. Thanks,
Reported-by: Andrew Lutomirski <luto@mit.edu>
Signed-off-by: Josef Bacik <josef@redhat.com>
I converted btrfs_truncate to do sane reservations for truncate, but didn't
convert btrfs_evict_inode. Basically we need to save the orphan_rsv for
deleting the orphan item, and do normal reservations for our truncate. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
This patch kills off the calculation for the amount of space needed for the
orphan operations during a snapshot. The thing is we only do snapshots on
commit, so any space that is in the block_rsv->freed[] isn't going to be in the
new snapshot anyway, so there isn't any reason to require that space to be
reserved for the snapshot to occur. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
We have not been reserving enough space for checksums. We were just reserving
bytes for the checksum items themselves, we were not taking into account having
to cow the tree and such. This patch adds a csum_bytes counter to the inode for
keeping track of the number of bytes outstanding we have for checksums. Then we
calculate how many leaves would be required for the checksums we are given and
use that to reserve space. This adds a significant amount of bytes to our
reservations, but we will handle this later. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
We always look for delalloc bytes in our io_tree so we can fill in delalloc.
This is fine in most cases, but if we're writing out the btree_inode this is
just a superfluous tree search on the io_tree, and if we have a lot of metadata
dirty this could be an expensive check. So instead check to see if our io_tree
has a ->fill_delalloc op, and if not don't even bother doing the lookup.
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
We have been using bytes_reserved for metadata reservations, which is wrong
since we use that to keep track of outstanding reservations from the allocator.
This resulted in us doing a lot of silly things to make sure we don't allocate a
bunch of metadata chunks since we never had a real view of how much space was
actually in use by metadata.
This passes Arne's enospc test and xfstests as well as my own enospc tests.
Hopefully this will get us moving in the right direction. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
We've only been able to mount with subvol=<whatever> where whatever was a subvol
within whatever root we had as the default. This allows us to mount -o
subvol=path/to/subvol/you/want relative from the normal fs_tree root. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Currently what we do is just wrong. We either
1) Alloc a new "root" dentry with sb->s_root as it's parent which is just wrong
as we could walk into this subvol later on via another path and hilarity could
ensue. Also we don't check the return value of d_splice_alias which isn't good
either.
or
2) Do a d_find_alias() which we could have lost our dentry from cache at this
point and found nothing.
So use d_obtain_alias(). In the case that we already have the inode/dentry in
cache we will get the correct dentry. If not we will get a disconnected dentry
tree so if we walk into it later on everything will be connected up properly.
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Moving things around to give us better packing in the btrfs_inode. This reduces
the size of our inode by 8 bytes. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
The btrfs file defrag code will loop through the extents and
force COW on them. But there is a concurrent truncate in the middle of
the defrag, it might end up defragging the same range over and over
again.
The problem is that writepage won't go through and do anything on pages
past i_size, so the cow won't happen, so the file will appear to still
be fragmented. defrag will end up hitting the same extents again and
again.
In the worst case, the truncate can actually live lock with the defrag
because the defrag keeps creating new ordered extents which the truncate
code keeps waiting on.
The fix here is to make defrag check for i_size inside the main loop,
instead of just once before the looping starts.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Follow those steps:
# mount -o autodefrag /dev/sda7 /mnt
# dd if=/dev/urandom of=/mnt/tmp bs=200K count=1
# sync
# dd if=/dev/urandom of=/mnt/tmp bs=8K count=1 conv=notrunc
and then it'll go into a loop: writeback -> defrag -> writeback ...
It's because writeback writes [8K, 200K] and then writes [0, 8K].
I tried to make writeback know if the pages are dirtied by defrag,
but the patch was a bit intrusive. Here I simply set writeback_index
when we defrag a file.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Scrub uses a simple tree-enumeration to bring the relevant portions
of the extent- and csum-tree into the page cache before starting the
scrub-I/O. This is now replaced by using the new readahead-API.
During readahead the scrub is being accounted as paused, so it won't
hold off transaction commits.
This change raises the average disk bandwith utilisation on my test
volume from 70% to 90%. On another volume, the time for a test run
went down from 89s to 43s.
Changes v5:
- reada1/2 are now of type struct reada_control *
Signed-off-by: Arne Jansen <sensille@gmx.net>
This adds the hooks needed for readahead. In the readpage_end_io_hook,
the extent state is checked for the EXTENT_READAHEAD flag. Only in this
case the readahead hook is called, to keep the impact on non-ra as low
as possible.
Additionally, a hook for a failed IO is added, otherwise readahead would
wait indefinitely for the extent to finish.
Changes for v2:
- eliminate race condition
Signed-off-by: Arne Jansen <sensille@gmx.net>
This is the implementation for the generic read ahead framework.
To trigger a readahead, btrfs_reada_add must be called. It will start
a read ahead for the given range [start, end) on tree root. The returned
handle can either be used to wait on the readahead to finish
(btrfs_reada_wait), or to send it to the background (btrfs_reada_detach).
The read ahead works as follows:
On btrfs_reada_add, the root of the tree is inserted into a radix_tree.
reada_start_machine will then search for extents to prefetch and trigger
some reads. When a read finishes for a node, all contained node/leaf
pointers that lie in the given range will also be enqueued. The reads will
be triggered in sequential order, thus giving a big win over a naive
enumeration. It will also make use of multi-device layouts. Each disk
will have its on read pointer and all disks will by utilized in parallel.
Also will no two disks read both sides of a mirror simultaneously, as this
would waste seeking capacity. Instead both disks will read different parts
of the filesystem.
Any number of readaheads can be started in parallel. The read order will be
determined globally, i.e. 2 parallel readaheads will normally finish faster
than the 2 started one after another.
Changes v2:
- protect root->node by transaction instead of node_lock
- fix missed branches:
The readahead had a too simple check to determine if a branch from
a node should be checked or not. It now also records the upper bound
of each node to see if the requested RA range lies within.
- use KERN_CONT to debug output, to avoid line breaks
- defer reada_start_machine to worker to avoid deadlock
Changes v3:
- protect root->node by rcu
Changes v5:
- changed EIO-semantics of reada_tree_block_flagged
- remove spin_lock from reada_control and make elems an atomic_t
- remove unused read_total from reada_control
- kill reada_key_cmp, use btrfs_comp_cpu_keys instead
- use kref-style release functions where possible
- return struct reada_control * instead of void * from btrfs_reada_add
Signed-off-by: Arne Jansen <sensille@gmx.net>
Add state information for readahead to btrfs_fs_info and btrfs_device
Changes v2:
- don't wait in radix_trees
- add own set of workers for readahead
Reviewed-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Arne Jansen <sensille@gmx.net>
Add a READAHEAD extent buffer flag.
Add a function to trigger a read with this flag set.
Changes v2:
- use extent buffer flags instead of extent state flags
Changes v5:
- adapt to changed read_extent_buffer_pages interface
- don't return eb from reada_tree_block_flagged if it has CORRUPT flag set
Signed-off-by: Arne Jansen <sensille@gmx.net>
read_extent_buffer_pages currently has two modes, either trigger a read
without waiting for anything, or wait for the I/O to finish. The former
also bails when it's unable to lock the page. This patch now adds an
additional parameter to allow it to block on page lock, but don't wait
for completion.
Changes v5:
- merge the 2 wait parameters into one and define WAIT_NONE, WAIT_COMPLETE and
WAIT_PAGE_LOCK
Change v6:
- fix bug introduced in v5
Signed-off-by: Arne Jansen <sensille@gmx.net>
A user reported a problem where ceph was getting into 100% cpu usage while doing
some writing. It turns out it's because we were doing a short write on a not
uptodate page, which means we'd fall back at one page at a time and fault the
page in. The problem is our position is on the page boundary, so our fault in
logic wasn't actually reading the page, so we'd just spin forever or until the
page got read in by somebody else. This will force a readpage if we end up
doing a short copy. Alexandre could reproduce this easily with ceph and reports
it fixes his problem. I also wrote a reproducer that no longer hangs my box
with this patch. Thanks,
Reported-and-tested-by: Alexandre Oliva <aoliva@redhat.com>
Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This ties nodatasum fixup in scrub together with raid repair patches. While
both series are working fine alone, scrub will report uncorrectable errors
if they occur in a nodatasum extent *and* the page is in the page cache.
Previously, we would have triggered readpage to find good data and do the
repair. However, readpage wouldn't read anything in the case where the page
is up to date in the cache. So, we simply take that good data we have and
call repair_io_failure directly (unless the page in the cache is dirty).
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
The raid-retry code in inode.c can be generalized so that it works for
metadata as well. Thus, this patch moves it to extent_io.c and makes the
raid-retry code a raid-repair code.
Repair works that way: Whenever a read error occurs and we have more
mirrors to try, note the failed mirror, and retry another. If we find a
good one, check if we did note a failure earlier and if so, do not allow
the read to complete until after the bad sector was written with the good
data we just fetched. As we have the extent locked while reading, no one
can change the data in between.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
The error correction code wants to make sure that only the bad mirror is
rewritten. Thus, we need to know which mirror is the bad one. I did not
find a more apropriate field than bi_bdev. But I think using this is fine,
because it is modified by the block layer, anyway, and should not be read
after the bio returned.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
The block layer modifies bio->bi_bdev and bio->bi_sector while working on
the bio, they do _not_ come back unmodified in the completion callback.
To call add_page, we need at least some bi_bdev set, which is why the code
was working, previously. With this patch, we use the latest_bdev from
fsinfo instead of the leftover in the bio. This gives us the possibility to
use the bi_bdev field for another purpose.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
btrfs_bio is a bio abstraction able to split and not complete after the last
bio has returned (like the old btrfs_multi_bio). Additionally, btrfs_bio
tracks the mirror_num used to read data which can be used for error
correction purposes.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
these ioctls make use of the new functions initially added for scrub. they
return all inodes belonging to a logical address (BTRFS_IOC_LOGICAL_INO) and
all paths belonging to an inode (BTRFS_IOC_INO_PATHS).
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
This removes a FIXME comment and introduces the first part of nodatasum
fixup: It gets the corresponding inode for a logical address and triggers a
regular readpage for the corrupted sector.
Once we have on-the-fly error correction our error will be automatically
corrected. The correction code is expected to clear the newly introduced
EXTENT_DAMAGED flag, making scrub report that error as "corrected" instead
of "uncorrectable" eventually.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
Currently, extent_read_full_page always assumes we are trying to read mirror
0, which generally is the best we can do. To add flexibility, pass it as a
parameter. This will be needed by scrub fixup code.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
Fix the mirror_num determination in scrub_stripe. The rest of the scrub code
did not use mirror_num for anything important and that error went unnoticed.
The nodatasum fixup patch of this set depends on a correct mirror_num.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
While scrubbing, we may encounter various errors. Previously, a logical
address was printed to the log only. Now, all paths belonging to that
address are resolved and printed separately. That should work for hardlinks
as well as reflinks.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
In normal operation, scrub is reading data sequentially in large portions.
In case of an i/o error, we try to find the corrupted area(s) by issuing
page sized read requests. With this commit we increment the
unverified_errors counter if all of the small size requests succeed.
Userland patches carrying such conspicous events to the administrator should
already be around.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
These helper functions iterate back references and call a function for each
backref. There is also a function to resolve an inode to a path in the
file system.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
Fix a crash/BUG_ON in the clone ioctl due to insufficient reservation. We
need to reserve space for:
- adjusting the old extent (possibly splitting it)
- adding the new extent
- updating the inode
Signed-off-by: Sage Weil <sage@newdream.net>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
We can race with readdir and the RCU path walking stuff. This is because we
clear the need lookup flag before actually instantiating the inode. This will
lead the RCU path walk stuff to find a dentry it thinks is valid without a
d_inode attached. So instead unhash the dentry when we first start the lookup,
and then clear the flag after we've instantiated the dentry so we're garunteed
to either try the slow lookup, or have the d_inode set properly.
Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
The recent reworking of btrfs' lseek lead to incorrect
values being returned. This adds checks for seeking
beyond EOF in SEEK_HOLE and makes sure the error
values come back correct.
Andi Kleen also sent in similar patches.
Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Reported-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
The dst file will have the same inode flags with dst file after
file clone, and I think it's unexpected.
For example, the dst file will suddenly become immutable after
getting some share of data with src file, if the src is immutable.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
To reproduce the bug:
# mount /dev/sda7 /mnt
# dd if=/dev/zero of=/mnt/src bs=4K count=1
# umount /mnt
# mount -o nodatasum /dev/sda7 /mnt
# dd if=/dev/zero of=/mnt/dst bs=4K count=1
# clone_range -s 4K -l 4K /mnt/src /mnt/dst
# echo 3 > /proc/sys/vm/drop_caches
# cat /mnt/dst
# dmesg
...
btrfs no csum found for inode 258 start 0
btrfs csum failed ino 258 off 0 csum 2566472073 private 0
It's because part of the file is checksummed and the other part is not,
and then btrfs will complain checksum is not found when we read the file.
Disallow file clone if src and dst file have different checksum flag,
so we ensure a file is completely checksummed or unchecksummed.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
It's a bug in commit f81c9cdc56
(Btrfs: truncate pages from clone ioctl target range)
We should pass the dest range to the truncate function, but not the
src range.
Also move the function before locking extent state.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Since the d_off in the first dirent for "." (that originates from
the 4th argument "offset" of filldir() for the 2nd dirent for "..")
is wrongly assigned in btrfs_real_readdir(), telldir returns same
offset for different locations.
| # mkfs.btrfs /dev/sdb1
| # mount /dev/sdb1 fs0
| # cd fs0
| # touch file0 file1
| # ../test
| telldir: 0
| readdir: d_off = 2, d_name = "."
| telldir: 2
| readdir: d_off = 2, d_name = ".."
| telldir: 2
| readdir: d_off = 3, d_name = "file0"
| telldir: 3
| readdir: d_off = 2147483647, d_name = "file1"
| telldir: 2147483647
To fix this problem, pass filp->f_pos (which is loff_t) instead.
| # ../test
| telldir: 0
| readdir: d_off = 1, d_name = "."
| telldir: 1
| readdir: d_off = 2, d_name = ".."
| telldir: 2
| readdir: d_off = 3, d_name = "file0"
:
At the moment the "offset" for "." is unused because there is no
preceding dirent, however it is better to pass filp->f_pos to follow
grammatical usage.
Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
* 'for-linus' of git://github.com/chrismason/linux:
Btrfs: add dummy extent if dst offset excceeds file end in
Btrfs: calc file extent num_bytes correctly in file clone
btrfs: xattr: fix attribute removal
Btrfs: fix wrong nbytes information of the inode
Btrfs: fix the file extent gap when doing direct IO
Btrfs: fix unclosed transaction handle in btrfs_cont_expand
Btrfs: fix misuse of trans block rsv
Btrfs: reset to appropriate block rsv after orphan operations
Btrfs: skip locking if searching the commit root in csum lookup
btrfs: fix warning in iput for bad-inode
Btrfs: fix an oops when deleting snapshots
You can see there's no file extent with range [0, 4096]. Check this by
btrfsck:
# btrfsck /dev/sda7
root 5 inode 258 errors 100
...
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
An attribute is not removed by 'setfattr -x attr file' and remains
visible in attr list. This makes xfstests/062 pass again.
Signed-off-by: David Sterba <dsterba@suse.cz>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
If we write some data into the data hole of the file(no preallocation for this
hole), Btrfs will allocate some disk space, and update nbytes of the inode, but
the other element--disk_i_size needn't be updated. At this condition, we must
update inode metadata though disk_i_size is not changed(btrfs_ordered_update_i_size()
return 1).
# mkfs.btrfs /dev/sdb1
# mount /dev/sdb1 /mnt
# touch /mnt/a
# truncate -s 856002 /mnt/a
# dd if=/dev/zero of=/mnt/a bs=4K count=1 conv=nocreat,notrunc
# umount /mnt
# btrfsck /dev/sdb1
root 5 inode 257 errors 400
found 32768 bytes used err is 1
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
When we write some data to the place that is beyond the end of the file
in direct I/O mode, a data hole will be created. And Btrfs should insert
a file extent item that point to this hole into the fs tree. But unfortunately
Btrfs forgets doing it.
The following is a simple way to reproduce it:
# mkfs.btrfs /dev/sdc2
# mount /dev/sdc2 /test4
# touch /test4/a
# dd if=/dev/zero of=/test4/a seek=8 count=1 bs=4K oflag=direct conv=nocreat,notrunc
# umount /test4
# btrfsck /dev/sdc2
root 5 inode 257 errors 100
Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Tested-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
The function - btrfs_cont_expand() forgot to close the transaction handle before
it jump out the while loop. Fix it.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
At the beginning of create_pending_snapshot, trans->block_rsv is set
to pending->block_rsv and is used for snapshot things, however, when
it is done, we do not recover it as will.
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
While truncating free space cache, we forget to change trans->block_rsv
back to the original one, but leave it with the orphan_block_rsv, and
then with option inode_cache enable, it leads to countless warnings of
btrfs_alloc_free_block and btrfs_orphan_commit_root:
WARNING: at fs/btrfs/extent-tree.c:5711 btrfs_alloc_free_block+0x180/0x350 [btrfs]()
...
WARNING: at fs/btrfs/inode.c:2193 btrfs_orphan_commit_root+0xb0/0xc0 [btrfs]()
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
It's not enough to just search the commit root, since we could be cow'ing the
very block we need to search through, which would mean that its locked and we'll
still deadlock. So use path->skip_locking as well. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
iput() shouldn't be called for inodes in I_NEW state.
We need to mark inode as constructed first.
WARNING: at fs/inode.c:1309 iput+0x20b/0x210()
Call Trace:
[<ffffffff8103e7ba>] warn_slowpath_common+0x7a/0xb0
[<ffffffff8103e805>] warn_slowpath_null+0x15/0x20
[<ffffffff810eaf0b>] iput+0x20b/0x210
[<ffffffff811b96fb>] btrfs_iget+0x1eb/0x4a0
[<ffffffff811c3ad6>] btrfs_run_defrag_inodes+0x136/0x210
[<ffffffff811ad55f>] cleaner_kthread+0x17f/0x1a0
[<ffffffff81035b7d>] ? sub_preempt_count+0x9d/0xd0
[<ffffffff811ad3e0>] ? transaction_kthread+0x280/0x280
[<ffffffff8105af86>] kthread+0x96/0xa0
[<ffffffff814336d4>] kernel_thread_helper+0x4/0x10
[<ffffffff8105aef0>] ? kthread_worker_fn+0x190/0x190
[<ffffffff814336d0>] ? gs_change+0xb/0xb
Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>
CC: Konstantin Khlebnikov <khlebnikov@openvz.org>
Tested-by: David Sterba <dsterba@suse.cz>
CC: Josef Bacik <josef@redhat.com>
CC: Chris Mason <chris.mason@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
We can reproduce this oops via the following steps:
$ mkfs.btrfs /dev/sdb7
$ mount /dev/sdb7 /mnt/btrfs
$ for ((i=0; i<3; i++)); do btrfs sub snap /mnt/btrfs /mnt/btrfs/s_$i; done
$ rm -fr /mnt/btrfs/*
$ rm -fr /mnt/btrfs/*
then we'll get
------------[ cut here ]------------
kernel BUG at fs/btrfs/inode.c:2264!
[...]
Call Trace:
[<ffffffffa05578c7>] btrfs_rmdir+0xf7/0x1b0 [btrfs]
[<ffffffff81150b95>] vfs_rmdir+0xa5/0xf0
[<ffffffff81153cc3>] do_rmdir+0x123/0x140
[<ffffffff81145ac7>] ? fput+0x197/0x260
[<ffffffff810aecff>] ? audit_syscall_entry+0x1bf/0x1f0
[<ffffffff81153d0d>] sys_unlinkat+0x2d/0x40
[<ffffffff8147896b>] system_call_fastpath+0x16/0x1b
RIP [<ffffffffa054f7b9>] btrfs_orphan_add+0x179/0x1a0 [btrfs]
When it comes to btrfs_lookup_dentry, we may set a snapshot's inode->i_ino
to BTRFS_EMPTY_SUBVOL_DIR_OBJECTID instead of BTRFS_FIRST_FREE_OBJECTID,
while the snapshot's location.objectid remains unchanged.
However, btrfs_ino() does not take this into account, and returns a wrong ino,
and causes the oops.
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
