- Both callers(truncate and punch_hole) already aligned left end point
so we no longer need split logic here.
- Remove dead duplicated code.
- Call ext4_ext_dirty only after we have updated eh_entries, otherwise
we'll loose entries update. Regression caused by d583fb87a3
266'th testcase in xfstests (http://patchwork.ozlabs.org/patch/120872)
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
* 'next' of git://selinuxproject.org/~jmorris/linux-security: (95 commits)
TOMOYO: Fix incomplete read after seek.
Smack: allow to access /smack/access as normal user
TOMOYO: Fix unused kernel config option.
Smack: fix: invalid length set for the result of /smack/access
Smack: compilation fix
Smack: fix for /smack/access output, use string instead of byte
Smack: domain transition protections (v3)
Smack: Provide information for UDS getsockopt(SO_PEERCRED)
Smack: Clean up comments
Smack: Repair processing of fcntl
Smack: Rule list lookup performance
Smack: check permissions from user space (v2)
TOMOYO: Fix quota and garbage collector.
TOMOYO: Remove redundant tasklist_lock.
TOMOYO: Fix domain transition failure warning.
TOMOYO: Remove tomoyo_policy_memory_lock spinlock.
TOMOYO: Simplify garbage collector.
TOMOYO: Fix make namespacecheck warnings.
target: check hex2bin result
encrypted-keys: check hex2bin result
...
Now that we support raid5 Enable it at mount. Raid6 will come next
raid4 is not demanded for so it will probably not be enabled.
(Until some one wants it)
NOTE: That mkfs.exofs had support for raid5/6 since long time
ago. (Making an empty raidX FS is just as easy as raid0 ;-} )
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
The ore need suplied a r4w_get_page/r4w_put_page API
from Filesystem so it can get cache pages to read-into when
writing parial stripes.
Also I commented out and NULLed the .writepage (singular)
vector. Because it gives terrible write pattern to raid
and is apparently not needed. Even in OOM conditions the
system copes (even better) with out it.
TODO: How to specify to write_cache_pages() to start
or include a certain page?
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
This is finally the RAID5 Write support.
The bigger part of this patch is not the XOR engine itself, But the
read4write logic, which is a complete mini prepare_for_striping
reading engine that can read scattered pages of a stripe into cache
so it can be used for XOR calculation. That is, if the write was not
stripe aligned.
The main algorithm behind the XOR engine is the 2 dimensional array:
struct __stripe_pages_2d.
A drawing might save 1000 words
---
__stripe_pages_2d
|
n = pages_in_stripe_unit;
w = group_width - parity;
| pages array presented to the XOR lib
| |
V |
__1_page_stripe[0].pages --> [c0][c1]..[cw][c_par] <---|
| |
__1_page_stripe[1].pages --> [c0][c1]..[cw][c_par] <---
|
... | ...
|
__1_page_stripe[n].pages --> [c0][c1]..[cw][c_par]
^
|
data added columns first then row
---
The pages are put on this array columns first. .i.e:
p0-of-c0, p1-of-c0, ... pn-of-c0, p0-of-c1, ...
So we are doing a corner turn of the pages.
Note that pages will zigzag down and left. but are put sequentially
in growing order. So when the time comes to XOR the stripe, only the
beginning and end of the array need be checked. We scan the array
and any NULL spot will be field by pages-to-be-read.
The FS that wants to support RAID5 needs to supply an
operations-vector that searches a given page in cache, and specifies
if the page is uptodate or need reading. All these pages to be read
are put on a slave ore_io_state and synchronously read. All the pages
of a stripe are read in one IO, using the scatter gather mechanism.
In write we constrain our IO to only be incomplete on a single
stripe. Meaning either the complete IO is within a single stripe so
we might have pages to read from both beginning or end of the
strip. Or we have some reading to do at beginning but end at strip
boundary. The left over pages are pushed to the next IO by the API
already established by previous work, where an IO offset/length
combination presented to the ORE might get the length truncated and
the user must re-submit the leftover pages. (Both exofs and NFS
support this)
But any ORE user should make it's best effort to align it's IO
before hand and avoid complications. A cached ore_layout->stripe_size
member can be used for that calculation. (NOTE: that ORE demands
that stripe_size may not be bigger then 32bit)
What else? Well read it and tell me.
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
This patch introduces the first stage of RAID5 support
mainly the skip-over-raid-units when reading. For
writes it inserts BLANK units, into where XOR blocks
should be calculated and written to.
It introduces the new "general raid maths", and the main
additional parameters and components needed for raid5.
Since at this stage it could corrupt future version that
actually do support raid5. The enablement of raid5
mounting and setting of parity-count > 0 is disabled. So
the raid5 code will never be used. Mounting of raid5 is
only enabled later once the basic XOR write is also in.
But if the patch "enable RAID5" is applied this code has
been tested to be able to properly read raid5 volumes
and is according to standard.
Also it has been tested that the new maths still properly
supports RAID0 and grouping code just as before.
(BTW: I have found more bugs in the pnfs-obj RAID math
fixed here)
The ore.c file is getting too big, so new ore_raid.[hc]
files are added that will include the special raid stuff
that are not used in striping and mirrors. In future write
support these will get bigger.
When adding the ore_raid.c to Kbuild file I was forced to
rename ore.ko to libore.ko. Is it possible to keep source
file, say ore.c and module file ore.ko the same even if there
are multiple files inside ore.ko?
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
fs/exofs directory has multiple targets now, of which the
ore.ko will be needed by the pnfs-objects-layout-driver
(fs/nfs/objlayout).
As suggested by: Michal Marek <mmarek@suse.cz> convert
inclusion of exofs/ from obj-$(CONFIG_EXOFS_FS) => obj-$(y).
So ORE can be selected also from fs/nfs/Kconfig
CC: Michal Marek <mmarek@suse.cz>
CC: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
that reduces a traffic and increases a performance.
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Acked-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
that reduces a traffic and increases a performance.
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Acked-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
to handle all lock requests on the client in an exclusive oplock case.
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Acked-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
If we have an oplock and negotiate mandatory locking style we handle
all brlock requests on the client.
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Acked-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Instead of saying all integer argument option should be listed in the beginning
move integer parsing to each option type.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
* remove lot of update to different data structure
* add a seperate callback for zero copy request.
* above makes non zero copy code path simpler
* remove conditionalizing TREAD/TREADDIR/TWRITE in the zero copy path
* Fix the dotu p9_check_errors with zero copy. Add sufficient doc around
* Add support for both in and output buffers in zero copy callback
* pin and unpin pages in the same context
* use helpers instead of defining page offset and rest of page ourself
* Fix mem leak in p9_check_errors
* Remove 'E' and 'F' in p9pdu_vwritef
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
bio originally has the functionality to set the complete cpu, but
it is broken.
Chirstoph said that "This code is unused, and from the all the
discussions lately pretty obviously broken. The only thing keeping
it serves is creating more confusion and possibly more bugs."
And Jens replied with "We can kill bio_set_completion_cpu(). I'm fine
with leaving cpu control to the request based drivers, they are the
only ones that can toggle the setting anyway".
So this patch tries to remove all the work of controling complete cpu
from a bio.
Cc: Shaohua Li <shaohua.li@intel.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The WARN_ON under some circumstances heavily polute log and slow down
the machine. This is just a safety, as the warning should be fixed by
another patch, nevertheless, it still pops up during testing.
Signed-off-by: David Sterba <dsterba@suse.cz>
There's a missing test whether the path passed to subvol=path option
during mount is a real subvolume, allowing any directory located in
default subovlume to be passed and accepted for mount.
(current btrfs progs prevent this early)
$ btrfs subvol snapshot . p1-snap
ERROR: '.' is not a subvolume
(with "is subvolume?" test bypassed)
$ btrfs subvol snapshot . p1-snap
Create a snapshot of '.' in './p1-snap'
$ btrfs subvol list -p .
ID 258 parent 5 top level 5 path subvol
ID 259 parent 5 top level 5 path subvol1
ID 260 parent 5 top level 5 path default-subvol1
ID 262 parent 5 top level 5 path p1/p1-snapshot
ID 263 parent 259 top level 5 path subvol1/subvol1-snap
The problem I see is that this makes a false impression of snapshotting the
given subvolume but in fact snapshots the default one: a user expects outcome
like ID 263 but in fact gets ID 262 .
This patch makes mount fail with EINVAL with a message in syslog.
Signed-off-by: David Sterba <dsterba@suse.cz>
According to rfc5661 18.50, implement DESTROY_CLIENTID operation.
Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
RFC5661 says:
The client may set one or both of
OPEN4_SHARE_ACCESS_WANT_SIGNAL_DELEG_WHEN_RESRC_AVAIL and
OPEN4_SHARE_ACCESS_WANT_PUSH_DELEG_WHEN_UNCONTENDED.
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
We should call cifs_all_info_to_fattr in rc == 0 case only.
Cc: <stable@kernel.org>
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Currently code make an impression what grow procedure is very complicated
and some mythical paths, blocks are involved. But in fact grow in depth
it relatively simple procedure:
1) Just create new meta block and copy root data to that block.
2) Convert root from extent to index if old depth == 0
3) Update root block pointer
This patch does:
- Reorganize code to make it more self explanatory
- Do not pass path parameter to new_meta_block() in order to
provoke allocation from inode's group because top-level block
should site closer to it's inode, but not to leaf data block.
[ This happens anyway, due to logic in mballoc; we should drop
the path parameter from new_meta_block() entirely. -- tytso ]
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In cifs_readv_receive we don't update rdata->result to error value
after kmap'ing a page. We should kunmap the page in the no error
case only.
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Move the recently added readahead of the indirect pointer
tree during deallocation into its own function in order
that we can use it elsewhere in the future. Also this
fixes the resetting of the "first" variable in the
original patch.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
The two variables being initialised in gfs2_inplace_reserve
to track the file & line number of the caller are never
used, so we might as well remove them.
If something does go wrong, then a stack trace is probably
more useful anyway.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Some items picked up through automated code analysis. A few bits
of unreachable code and two unchecked return values.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
GFS2's fallocate code currently goes through the page cache. Since it's only
writing to the end of the file or to holes in it, it doesn't need to, and it
was causing issues on low memory environments. This patch pulls in some of
Steve's block allocation work, and uses it to simply allocate the blocks for
the file, and zero them out at allocation time. It provides a slight
performance increase, and it dramatically simplifies the code.
Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
This patch improves the performance of delete/unlink
operations in a GFS2 file system where the files are large
by adding a layer of metadata read-ahead for indirect blocks.
Mileage will vary, but on my system, deleting an 8.6G file
dropped from 22 seconds to about 4.5 seconds.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Bob reported:
I found an off-by-one problem with how I coded this section:
It should be:
+ else if (blk >= cur->rd_data0 + cur->rd_data)
In fact, cur->rd_data0 + cur->rd_data is the start of the next
rgrp (the next ri_addr), so without the "=" check it can land on
the wrong rgrp.
In all normal cases, this won't be a problem: you're searching
for a block _within_ the rgrp, which will pass the test properly.
Where it gets into trouble is if you search the rgrps for the
block exactly equal to ri_addr. I don't think anything in the
kernel does this, but I found a place in gfs2-utils gfs2_edit
where it does. So I definitely need to fix it in libgfs2. I'd
like to suggest we fix it in the kernel as well for the sake of
keeping the functions similar.
So this patch fixes the above mentioned off by one error as well
as removing the unused parent pointer.
Reported-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
This patch brings gfs2's ->page_mkwrite uptodate with respect to the
expectations set by the VM. Also added is a check to wait if the fs
is frozen, before we attempt to get a glock. This will only work on
the node which initiates the freeze, but thats ok since the transaction
lock will still provide the expected barrier on other nodes.
The major change here is that we return a locked page now, except when
we don't return a page at all (error cases). This removes the race
which required rechecking the page after it was returned.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Cc: Nick Piggin <npiggin@kernel.dk>
Unfortunately, it is not enough to just ignore locked buffers during
the AIL flush from fsync. We need to be able to ignore all buffers
which are locked, dirty or pinned at this stage as they might have
been added subsequent to the log flush earlier in the fsync function.
In addition, this means that we no longer need to rely on i_mutex to
keep out writes during fsync, so we can, as a side-effect, remove
that protection too.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Tested-By: Abhijith Das <adas@redhat.com>
Each block which is deallocated, requires a call to gfs2_rlist_add()
and each of those calls was calling gfs2_blk2rgrpd() in order to
figure out which rgrp the block belonged in. This can be speeded up
by making use of the rgrp cached in the inode. We also reset this
cached rgrp in case the block has changed rgrp. This should provide
a big reduction in gfs2_blk2rgrpd() calls during deallocation.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
The recursive_scan() function only ever takes a single "bc"
argument, so we might as well just call do_strip() directly
from resource_scan() rather than pass it in as an argument.
Also the "data" argument is always a struct strip_mine, so
we can pass that in, rather than using a void pointer.
This also moves do_strip() ahead of recursive_scan() so that
we don't need to add a prototype.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Given that a resource group has been locked, there is no reason why
we should not be able to allocate as many blocks as are free. The
al_requested parameter should really be considered as a minimum
number of blocks to be available. Should this limit be overshot,
there are other mechanisms which will prevent over allocation.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
This means that after the initial allocation for any inode, the
last used resource group is cached in the inode for future use.
This drastically reduces the number of lookups of resource
groups in the common case, and this the contention on that
data structure.
The allocation algorithm is the same as previously, except that we
always check to see if the goal block is within the cached rgrp
first before going to the rbtree to look one up.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Since we have ruled out supporting online filesystem shrink,
it is possible to make the resource group list append only
during the life of a super block. This gives several benefits:
Firstly, we only need to read new rindex elements as they are added
rather than needing to reread the whole rindex file each time one
element is added.
Secondly, the rindex glock can be held for much shorter periods of
time, and is completely removed from the fast path for allocations.
The lock is taken in shared mode only when updating the resource
groups when the first allocation occurs, and after a grow has
taken place.
Thirdly, this results in a reduction in code size, and everything
gets a lot simpler to understand in this area.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Here is an update of Bob's original rbtree patch which, in addition, also
resolves the rather strange ref counting that was being done relating to
the bitmap blocks.
Originally we had a dual system for journaling resource groups. The metadata
blocks were journaled and also the rgrp itself was added to a list. The reason
for adding the rgrp to the list in the journal was so that the "repolish
clones" code could be run to update the free space, and potentially send any
discard requests when the log was flushed. This was done by comparing the
"cloned" bitmap with what had been written back on disk during the transaction
commit.
Due to this, there was a requirement to hang on to the rgrps' bitmap buffers
until the journal had been flushed. For that reason, there was a rather
complicated set up in the ->go_lock ->go_unlock functions for rgrps involving
both a mutex and a spinlock (the ->sd_rindex_spin) to maintain a reference
count on the buffers.
However, the journal maintains a reference count on the buffers anyway, since
they are being journaled as metadata buffers. So by moving the code which deals
with the post-journal accounting for bitmap blocks to the metadata journaling
code, we can entirely dispense with the rather strange buffer ref counting
scheme and also the requirement to journal the rgrps.
The net result of all this is that the ->sd_rindex_spin is left to do exactly
one job, and that is to look after the rbtree or rgrps.
This patch is designed to be a stepping stone towards using RCU for the rbtree
of resource groups, however the reduction in the number of uses of the
->sd_rindex_spin is likely to have benefits for multi-threaded workloads,
anyway.
The patch retains ->go_lock and ->go_unlock for rgrps, however these maybe also
be removed in future in favour of calling the functions directly where required
in the code. That will allow locking of resource groups without needing to
actually read them in - something that could be useful in speeding up statfs.
In the mean time though it is valid to dereference ->bi_bh only when the rgrp
is locked. This is basically the same rule as before, modulo the references not
being valid until the following journal flush.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Cc: Benjamin Marzinski <bmarzins@redhat.com>
We need to take the inode's glock whenever the inode's size
is referenced, otherwise it might not be uptodate. Even
though generic_file_llseek_unlocked() doesn't implement
SEEK_DATA, SEEK_HOLE directly, it does reference the inode's
size in those cases, so we need to add them to the list
of origins which need the glock.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
If we pass through knowledge of whether the creation is intended to be
exclusive or not, then we can deal with that in gfs2_create_inode
and remove one set of locking. Also this removes the loop in
gfs2_create and simplifies the code a bit.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
The aim of this patch is to use the newly enhanced ->dirty_inode()
super block operation to deal with atime updates, rather than
piggy backing that code into ->write_inode() as is currently
done.
The net result is a simplification of the code in various places
and a reduction of the number of gfs2_dinode_out() calls since
this is now implied by ->dirty_inode().
Some of the mark_inode_dirty() calls have been moved under glocks
in order to take advantage of then being able to avoid locking in
->dirty_inode() when we already have suitable locks.
One consequence is that generic_write_end() now correctly deals
with file size updates, so that we do not need a separate check
for that afterwards. This also, indirectly, means that fdatasync
should work correctly on GFS2 - the current code always syncs the
metadata whether it needs to or not.
Has survived testing with postmark (with and without atime) and
also fsx.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Journaled data requires that a complete flush of all dirty data for
the file is done, in order that the ail flush which comes after
will succeed.
Also the recently enhanced bug trap can trigger falsely in case
an ail flush from fsync races with a page read. This updates the
bug trap such that it will ignore buffers which are locked and
only trigger on dirty and/or pinned buffers when the ail flush
is run from fsync. The original bug trap is retained when ail
flush is run from ->go_sync()
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
If we have got far enough through the inode allocation code
path that an inode has already been allocated, then we must
call iput to dispose of it, if an error occurs during a
later part of the process. This will always be the final iput
since there will be no other references to the inode.
Unlike when the inode has been unlinked, its block state will
be GFS2_BLKST_INODE rather than GFS2_BLKST_UNLINKED so we need
to skip the test in ->evict_inode() for this one case in order
to ensure that it will be deallocated correctly. This patch adds
a new flag in order to ensure that this will happen correctly.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
We do not need to start a transaction unless the atime
check has proved positive. Also if we are going to flush
the complete ail list anyway, we might as well skip the
writeback for this specific inode's metadata, since that
will be done as part of the ail writeback process in an
order offering potentially more efficient I/O.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
The assert was being tested under the wrong lock, a
legacy of the original code. Also, if it does trigger,
the resulting information was not always a lot of help.
This moves the patch under the correct lock and also
prints out more useful information in tacking down the
source of the problem.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Now that the data writing is part of fsync proper, we can split
the waiting part out and do it later on. This reduces the
number of waits that we do during fsync on average.
There is also no need to take the i_mutex unless we are flushing
metadata to disk, so we can move that to within the metadata
flushing code.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Since there is now only a single caller to gfs2_dir_read_data()
and it has a number of constant arguments, we can factor
those out. Also some tests relating to the inode size were
being done twice.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Quota file is fs's metadata, so it is reasonable to permit use
root resevation if necessary. This patch fix 265'th xfstest failure
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
As soon as the nfs_client gets created, its cl_rpcclient is set to
ERR_PTR(-EINVAL). The rpc client structure is allocated later. Check
if the client is ready before using the cl_rpcclient pointer.
Signed-off-by: Malahal Naineni <malahal@us.ibm.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
If ext4_jbd2_file_inode() in mpage_da_map_and_submit() fails due to
journal abort, this function returns to caller without unlocking the
page. It leads to the deadlock, and the patch fixes this issue by
calling mpage_da_submit_io().
Signed-off-by: Kazuya Mio <k-mio@sx.jp.nec.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If ext4_jbd2_file_inode() in ext4_ordered_write_end() fails for some
reasons, this function returns to caller without unlocking the page.
It leads to the deadlock, and the patch fixes this issue.
Signed-off-by: Akira Fujita <a-fujita@rs.jp.nec.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Fix a bug introduced by 20b45077. We have to return EINVAL on mount
failure, but doing that too early in the sequence leaves all of the
devices opened exclusively. This also fixes an issue where under some
scenarios only a second mount -o degraded <devices> command would
succeed.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Initialize fs_info->bdev_holder a bit earlier to be able to pass a
correct holder id to blkdev_get() when opening seed devices with O_EXCL.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
If lookup_extent_backref fails, path->nodes[0] reasonably could be
null along with other callers of btrfs_print_leaf, so ensure we have a
valid extent buffer before dereferencing.
Signed-off-by: Daniel J Blueman <daniel.blueman@gmail.com>
The task may fail to get free space though it is enough when multi-task
space allocation and caching space happen at the same time.
Task1 Caching Thread Task2
------------------------------------------------------------------------
find_free_extent
The space has not
be cached, and start
caching thread. And
wait for it.
cache space, if
the space is > 2MB
wake up Task1
find_free_extent
get all the space that
is cached.
try to allocate space,
but there is no space
now.
trigger BUG_ON()
The message is following:
btrfs allocation failed flags 1, wanted 4096
space_info has 1040187392 free, is not full
space_info total=1082130432, used=4096, pinned=41938944, reserved=0, may_use=40828928, readonly=0
block group 12582912 has 8388608 bytes, 0 used 8388608 pinned 0 reserved
block group has cluster?: no
0 blocks of free space at or bigger than bytes is
block group 1103101952 has 1073741824 bytes, 4096 used 33550336 pinned 0 reserved
block group has cluster?: no
0 blocks of free space at or bigger than bytes is
------------[ cut here ]------------
kernel BUG at fs/btrfs/inode.c:835!
[<ffffffffa031261b>] __extent_writepage+0x1bf/0x5ce [btrfs]
[<ffffffff810cbcb8>] ? __set_page_dirty_nobuffers+0xfe/0x108
[<ffffffffa02f8ada>] ? wait_current_trans+0x23/0xec [btrfs]
[<ffffffff810c3fbf>] ? find_get_pages_tag+0x73/0xe2
[<ffffffffa0312d12>] extent_write_cache_pages.clone.0+0x176/0x29a [btrfs]
[<ffffffffa0312e74>] extent_writepages+0x3e/0x53 [btrfs]
[<ffffffff8110ad2c>] ? do_sync_write+0xc6/0x103
[<ffffffffa0302d6e>] ? btrfs_submit_direct+0x414/0x414 [btrfs]
[<ffffffff811380fa>] ? fsnotify+0x236/0x266
[<ffffffffa02fc930>] btrfs_writepages+0x22/0x24 [btrfs]
[<ffffffff810cc215>] do_writepages+0x1c/0x25
[<ffffffff810c4958>] __filemap_fdatawrite_range+0x4e/0x50
[<ffffffff810c4982>] filemap_write_and_wait_range+0x28/0x51
[<ffffffffa0306b2e>] btrfs_sync_file+0x7d/0x198 [btrfs]
[<ffffffff8110aa26>] ? fsnotify_modify+0x5d/0x65
[<ffffffff8112d150>] vfs_fsync_range+0x18/0x21
[<ffffffff8112d170>] vfs_fsync+0x17/0x19
[<ffffffff8112d316>] do_fsync+0x29/0x3e
[<ffffffff8112d348>] sys_fsync+0xb/0xf
[<ffffffff81468352>] system_call_fastpath+0x16/0x1b
[SNIP]
RIP [<ffffffffa02fe08c>] cow_file_range+0x1c4/0x32b [btrfs]
We fix this bug by trying to allocate the space again if there are block groups
in caching.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
In btrfs_get_acl(), when the second __btrfs_getxattr() call fails,
acl is not correctly set.
Therefore, a wrong value might return to the caller.
Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Free space items are located in tree of tree roots, not in the extent
tree. It didn't pop up because lookup_free_space_inode() grabs the
inode all the time instead of actually searching the tree.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
To reproduce the bug:
# mount -o nodatacow /dev/sda7 /mnt/
# dd if=/dev/zero of=/mnt/tmp bs=4K count=1
1+0 records in
1+0 records out
4096 bytes (4.1 kB) copied, 0.000136115 s, 30.1 MB/s
# dd if=/dev/zero of=/mnt/tmp bs=4K count=1 conv=notrunc oflag=direct
dd: writing `/mnt/tmp': Input/output error
1+0 records in
0+0 records out
btrfs_ordered_update_i_size() may return 1, but btrfs_endio_direct_write()
mistakenly takes it as an error.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
It's not a big deal if we fail to allocate the array, and instead of
panic we can just give up compressing.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
We should retirn EINVAL if the start is beyond the end of the file
system in the btrfs_ioctl_fitrim(). Fix that by adding the appropriate
check for it.
Also in the btrfs_trim_fs() it is possible that len+start might overflow
if big values are passed. Fix it by decrementing the len so that start+len
is equal to the file system size in the worst case.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
We won't defrag an extent, if it's bigger than the threshold we
specified and there's no small extent before it, but actually
the code doesn't work this way.
There are three bugs:
- When should_defrag_range() decides we should keep on defragmenting
an extent, last_len is not incremented. (old bug)
- The length that passes to should_defrag_range() is not the length
we're going to defrag. (new bug)
- We always defrag 256K bytes data, and a big extent can be part of
this range. (new bug)
For a file with 4 extents:
| 4K | 4K | 256K | 256K |
The result of defrag with (the default) 256K extent thresh should be:
| 264K | 256K |
but with those bugs, we'll get:
| 520K |
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
It's off-by-one, and thus we may skip the last page while defragmenting.
An example case:
# create /mnt/file with 2 4K file extents
# btrfs fi defrag /mnt/file
# sync
# filefrag /mnt/file
/mnt/file: 2 extents found
So it's not defragmented.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Don't use inode->i_size directly, since we're not holding i_mutex.
This also fixes another bug, that i_size can change after it's checked
against 0 and then (i_size - 1) can be negative.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Offset field in data extent backref can underflow if clone range ioctl
is used. We can reliably detect the underflow because max file size is
limited to 2^63 and max data extent size is limited by block group size.
Signed-off-by: Zheng Yan <zheng.z.yan@intel.com>
Add support to print nostrictsync and noperm mount options in
/proc/mounts for shares mounted with these options.
(cleanup merge conflict in Sachin's original patch)
Suggested-by: Sachin Prabhu <sprabhu@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
sysfs is a core piece of ifrastructure that many people use and
few people have all of the rules in their head on how to use
it correctly. Add warnings for people using tagged directories
improperly to that any misuses can be caught and diagnosed quickly.
A single inexpensive test in sysfs_find_dirent is almost sufficient
to catch all possible misuses. An additional warning is needed
in sysfs_add_dirent so that we actually fail when attempting to
add an untagged dirent in a tagged directory.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now that /sys/class/net/bonding_masters is implemented as a tagged sysfs
file we can remove support for untagged files in tagged directories.
This change removes any ambiguity of what a NULL namespace value
means. A NULL namespace parameter after this patch means
that we are talking about an untagged sysfs dirent.
This makes the sysfs code much less prone to mistakes when during
maintenance.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Looking up files in sysfs is hard to understand and analyize because we
currently allow placing untagged files in tagged directories. In the
implementation of that we have two subtly different meanings of NULL.
NULL meaning there is no tag on a directory entry and NULL meaning
we don't care which namespace the lookup is performed for. This
multiple uses of NULL have resulted in subtle bugs (since fixed)
in the code.
Currently it is only the bonding driver that needs to have an untagged
file in a tagged directory.
To untagle this mess I am adding support for tagged files to sysfs.
Modifying the bonding driver to implement bonding_masters as a tagged
file. Registering bonding_masters once for each network namespace.
Then I am removing support for untagged entries in tagged sysfs
directories.
Resulting in code that is much easier to reason about.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Don't rely on the PageError flag to tell us if one of the partial reads of
the page failed. Instead, replace that with a dedicated flag in the
struct nfs_page.
Then clean out redundant uses of the PageError flag: the VM no longer
checks it for reads.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
CIFS currently uses wait_event_killable to put tasks to sleep while
they await replies from the server. That function though does not
allow the freezer to run. In many cases, the network interface may
be going down anyway, in which case the reply will never come. The
client then ends up blocking the computer from suspending.
Fix this by adding a new wait_event_freezable variant --
wait_event_freezekillable. The idea is to combine the behavior of
wait_event_killable and wait_event_freezable -- put the task to
sleep and only allow it to be awoken by fatal signals, but also
allow the freezer to do its job.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Tune bdi.ra_pages to be a multiple of the rsize. This prevents the VFS
from asking for pages that require small reads to satisfy.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Currently we cap the rsize at a value that fits in CIFSMaxBufSize. That's
not needed any longer for readpages. Allow the use of larger values for
readpages. cifs_iovec_read and cifs_read however are still limited to the
CIFSMaxBufSize. Make sure they don't exceed that.
The patch also changes the rsize defaults. The default when unix
extensions are enabled is set to 1M for parity with the wsize, and there
is a hard cap of ~16M.
When unix extensions are not enabled, the default is set to 60k. According
to MS-CIFS, Windows servers can only send a max of 60k at a time, so
this is more efficient than requesting a larger size. If the user wishes
however, the max can be extended up to 128k - the length of the READ_RSP
header.
Really old servers however require a special hack to ensure that we don't
request too large a read.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Now that we have code in place to do asynchronous reads, convert
cifs_readpages to use it. The new cifs_readpages walks the page_list
that gets passed in, locks and adds the pages to the pagecache and
sets up cifs_readdata to handle the reads.
The rest is handled by the cifs_async_readv infrastructure.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
...which will allow cifs to do an asynchronous read call to the server.
The caller will allocate and set up cifs_readdata for each READ_AND_X
call that should be issued on the wire. The pages passed in are added
to the pagecache, but not placed on the LRU list yet (as we need the
page->lru to keep the pages on the list in the readdata).
When cifsd identifies the mid, it will see that there is a special
receive handler for the call, and use that to receive the rest of the
frame. cifs_readv_receive will then marshal up a kvec array with
kmapped pages from the pagecache, which eliminates one copy of the
data. Once the data is received, the pages are added to the LRU list,
set uptodate, and unlocked.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
There is no pad, and it simplifies the code to remove the "Data" field.
None of the existing code relies on these fields, or on the READ_RSP
being a particular length.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
In order to handle larger SMBs for readpages and other calls, we want
to be able to read into a preallocated set of buffers. Rather than
changing all of the existing code to preallocate buffers however, we
instead add a receive callback function to the MID.
cifsd will call this function once the mid_q_entry has been identified
in order to receive the rest of the SMB. If the mid can't be identified
or the receive pointer is unset, then the standard 3rd phase receive
function will be called.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Move the entire 3rd phase of the receive codepath into a separate
function in preparation for the addition of a pluggable receive
function.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
In order to receive directly into a preallocated buffer, we need to ID
the mid earlier, before the bulk of the response is read. Call the mid
finding routine as soon as we're able to read the mid.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
We have several functions that need to access these pointers. Currently
that's done with a lot of double pointer passing. Instead, move them
into the TCP_Server_Info and simplify the handling.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Change find_cifs_mid to only return NULL if a mid could not be found.
If we got part of a multi-part T2 response, then coalesce it and still
return the mid. The caller can determine the T2 receive status from
the flags in the mid.
With this change, there is no need to pass a pointer to "length" as
well so just pass by value. If a mid is found, then we can just mark
it as malformed. If one isn't found, then the value of "length" won't
change anyway.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Begin breaking up find_cifs_mid into smaller pieces. The parts that
coalesce T2 responses don't really need to be done under the
GlobalMid_lock anyway. Create a new function that just finds the
mid on the list, and then later takes it off the list if the entire
response has been received.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Have the demultiplex thread receive just enough to get to the MID, and
then find it before receiving the rest. Later, we'll use this to swap
in a preallocated receive buffer for some calls.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Having to continually allocate a new kvec array is expensive. Allocate
one that's big enough, and only reallocate it as needed.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Eventually we'll want to allow cifsd to read data directly into the
pagecache. In order to do that we'll need a routine that can take a
kvec array and pass that directly to kernel_recvmsg.
Unfortunately though, the kernel's recvmsg routines modify the kvec
array that gets passed in, so we need to use a copy of the kvec array
and refresh that copy on each pass through the loop.
Reviewed-and-Tested-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
I noticed we had a little bit of latency when writing out the space cache
inodes. It's because we flush it before we write anything in case we have dirty
pages already there. This doesn't matter though since we're just going to
overwrite the space, and there really shouldn't be any dirty pages anyway. This
makes some of my tests run a little bit faster. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
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>
0c12eaffdf "nfsd: don't break lease on
CLAIM_DELEGATE_CUR" was a temporary workaround for a problem fixed
properly in the vfs layer by 778fc546f7
"locks: fix tracking of inprogress lease breaks", so we can revert that
change (but keeping some minor cleanup from that commit).
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
..to allow forcing of either compression scheme. This will override
compiled-in defaults. jffs2_compress is reworked a bit, as the lzo/zlib
override shares lots of code w/ the PRIORITY mode.
v2: update show_options accordingly.
Signed-off-by: Andres Salomon <dilinger@queued.net>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@intel.com>
Currently jffs2 has compile-time constants (and .config options)
controlling whether or not the various compression/decompression
drivers are built in and enabled. This is fine for embedded
systems, but it clashes with distribution kernels. Distro kernels
tend to turn on everything; this causes OpenFirmware to fall
over, as it understands ZLIB-compressed inodes. Booting a kernel
that has LZO compression enabled, writing to the boot partition,
and then rebooting causes OFW to fail to read the kernel from
the filesystem. This is because LZO compression has priority
when writing new data to jffs2, if LZO is enabled.
This patch adds mount option parsing, and a single supported
option ("compr=none"). This adds the flexibility of being
able to specify which compressor overrides on a per-superblock
basis. For now, we can simply disable compression;
additional flexibility coming soon.
v2: kill some printks, and implement show_options as suggested
by Artem Bityutskiy.
Signed-off-by: Andres Salomon <dilinger@queued.net>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@intel.com>
The following command sequence triggers an oops.
# mount /dev/sdb1 /mnt
# echo 1 > /sys/class/scsi_device/0\:0\:1\:0/device/delete
# umount /mnt
general protection fault: 0000 [#1] PREEMPT SMP
CPU 2
Modules linked in:
Pid: 791, comm: umount Not tainted 3.1.0-rc3-work+ #8 Bochs Bochs
RIP: 0010:[<ffffffff810d0879>] [<ffffffff810d0879>] __lock_acquire+0x389/0x1d60
...
Call Trace:
[<ffffffff810d2845>] lock_acquire+0x95/0x140
[<ffffffff81aed87b>] _raw_spin_lock+0x3b/0x50
[<ffffffff811573bc>] bdi_lock_two+0x5c/0x70
[<ffffffff811c2f6c>] bdev_inode_switch_bdi+0x4c/0xf0
[<ffffffff811c3fcb>] __blkdev_put+0x11b/0x1d0
[<ffffffff811c4010>] __blkdev_put+0x160/0x1d0
[<ffffffff811c40df>] blkdev_put+0x5f/0x190
[<ffffffff8118f18d>] kill_block_super+0x4d/0x80
[<ffffffff8118f4a5>] deactivate_locked_super+0x45/0x70
[<ffffffff8119003a>] deactivate_super+0x4a/0x70
[<ffffffff811ac4ad>] mntput_no_expire+0xed/0x130
[<ffffffff811acf2e>] sys_umount+0x7e/0x3a0
[<ffffffff81aeeeab>] system_call_fastpath+0x16/0x1b
This is because bdev holds on to disk but disk doesn't pin the
associated queue. If a SCSI device is removed while the device is
still open, the sdev puts the base reference to the queue on release.
When the bdev is finally released, the associated queue is already
gone along with the bdi and bdev_inode_switch_bdi() ends up
dereferencing already freed bdi.
Even if it were not for this bug, disk not holding onto the associated
queue is very unusual and error-prone.
Fix it by making add_disk() take an extra reference to its queue and
put it on disk_release() and ensuring that disk and its fops owner are
put in that order after all accesses to the disk and queue are
complete.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: stable@kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Both LOOKUP and OPEN operations may return NFS4ERR_BADNAME if we send a
an invalid name as a filename argument. As far as the application is
concerned, it just has to know that the file doesn't exist, and so
ENOENT would be the appropriate reply. We should only return EINVAL
if the filename is being used to _create_ a new object on the
remote filesystem.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
I intended to do this as part of fixing part of the conflict with
the merge with Linus' tree, but evidently it didn't get included in
the commit.
Signed-off-by: Alex Elder <aelder@sgi.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
commit ae50c0b5 "pnfs: client stats" added additional information to
the output of /proc/self/mountstats. The new functions introduced are
only used in this file and should be marked static.
If CONFIG_NFS_V4_1 is not defined, empty stub functions are used. If
CONFIG_NFS_V4 is not defined these stub functions are not used at all.
Adding static for the functions results in compile warnings:
fs/nfs/super.c:743: warning: 'show_sessions' defined but not used
fs/nfs/super.c:756: warning: 'show_pnfs' defined but not used
Fix this by adding a #ifdef CONFIG_NFS_V4 guard around the two
show_ functions.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
bl_add_page_to_bio returns error pointer. bio should be reset to
NULL in failure cases as the out path always calls bl_submit_bio.
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Cc: stable@kernel.org [3.0]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
For pnfs pagelist read failure, we need to pg_recoalesce and resend IO to
mds.
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Cc: stable@kernel.org [3.0]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
For pnfs pagelist write failure, we need to pg_recoalesce and resend IO to
mds.
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Cc: stable@kernel.org [3.0]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
file layout and block layout both use it to set mark layout io failure
bit. So make it generic.
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Cc: stable@kernel.org [3.0]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The same function is used by idmap, gss and blocklayout code. Make it
generic.
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Cc: stable@kernel.org [3.0]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Make the status field explicitly 32 bits. "...it's unlikely that the kernel
and userspace would differ on the size of an int here, but it might be a
good idea to go ahead and make that explicitly 32 bits in case we end up
dealing with more exotic arches at some point in the future."
Suggested-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Cc: stable@kernel.org [3.0]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Always return PTR_ERR, not NULL, from nfs4_blk_get_deviceinfo and
nfs4_blk_decode_device.
Check for IS_ERR, not NULL, in bl_set_layoutdriver when calling
nfs4_blk_get_deviceinfo.
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Cc: stable@kernel.org [3.0]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
nfs_find_and_lock_request will take a reference to the nfs_page and
will then put it if the req is already locked. It's possible though
that the reference will be the last one. That put then can kick off
a whole series of reference puts:
nfs_page
nfs_open_context
dentry
inode
If the inode ends up being deleted, then the VFS will call
truncate_inode_pages. That function will try to take the page lock, but
it was already locked when migrate_page was called. The code
deadlocks.
Fix this by simply refusing the migration request if PagePrivate is
already set, indicating that the page is already associated with an
active read or write request.
We've had a customer test a backported version of this patch and
the preliminary results seem good.
Cc: stable@kernel.org
Cc: Andrea Arcangeli <aarcange@redhat.com>
Reported-by: Harshula Jayasuriya <harshula@redhat.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The result from ipv6_addr_scope() always not be a single SCOPE,
so we can't use equal to compare the result with IPV6_ADDR_SCOPE_LINKLOCAL
at nfs_sockaddr_match_ipaddr6.
This patch fixs the problem, and lets checking address before scope_id.
Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
commit 420e3646 allowed the kernel to reduce the number of unnecessary
commit calls by skipping the commit when there are a large number of
outstanding pages.
However, the current test in nfs_commit_unstable_pages does not handle
the edge condition properly. When ncommit == 0, then that means that the
kernel doesn't need to do anything more for the inode. The current test
though in the WB_SYNC_NONE case will return true, and the inode will end
up being marked dirty. Once that happens the inode will never be clean
until there's a WB_SYNC_ALL flush.
Fix this by immediately returning from nfs_commit_unstable_pages when
ncommit == 0.
Mike noticed this problem initially in RHEL5 (2.6.18-based kernel) which
has a backported version of 420e3646. The inode cache there was growing
very large. The inode cache was unable to be shrunk since the inodes
were all marked dirty. Calling sync() would essentially "fix" the
problem -- the WB_SYNC_ALL flush would result in the inodes all being
marked clean.
What I'm not clear on is how big a problem this is in mainline kernels
as the writeback code there is very different. Either way, it seems
incorrect to re-mark the inode dirty in this case.
Reported-by: Mike McLean <mikem@redhat.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Cc: stable@kernel.org [2.6.34+]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
This reverts commit b80c3cb628.
The reverted commit was rendered obsolete by a VFS fix: commit
5547e8aac6 (writeback: Update dirty flags in
two steps). We now no longer need to worry about writeback_single_inode()
missing our marking the inode for COMMIT in 'do_writepages()' call.
Reverting this patch, fixes a performance regression in which the inode
would continuously get queued to the dirty list, causing the writeback
code to unnecessarily try to send a COMMIT.
Signed-off-by: Trond Myklebust <Trond.Myklebust>
Tested-by: Simon Kirby <sim@hostway.ca>
Cc: stable@kernel.org [2.6.35+]
Automounting directories are now invalidated by .d_revalidate()
so to be d_instantiate()d again with the right DCACHE_NEED_AUTOMOUNT
flag
Reviewed-by: Jeff Layton <jlayton@samba.org>
Signed-off-by: Gerlando Falauto <gerlando.falauto@keymile.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Smatch complains that the cast to "int" in min_t() changes very large
values of current_read_size into negative values and so min_t()
could return the wrong value. I removed the const as well, as that
doesn't do anything here.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Steve French <smfrench@gmail.com>
The third parameter to ext4_free_blocks is a struct buffer_head *. This
parameter should be NULL not 0.
This quiets the sparse noise:
warning: Using plain integer as NULL pointer
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The function declarations in ext4.h are already marked extern, so it's
not necessary to do so in the .c files.
This quiets the sparse noise:
warning: function 'ext4_flush_completed_IO' with external linkage has definition
warning: function 'ext4_init_inode_table' with external linkage has definition
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Add block plug for ext4 .writepages. Though ext4 .writepages
already handles request merge very well, block plug is still
helpful to reduce block lock contention.
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
As part of startup, the MMP initialization code does this:
mmp->mmp_seq = seq = cpu_to_le32(mmp_new_seq());
Next, mmp->mmp_seq is written out to disk, a delay happens, and then
the MMP block is read back in and the sequence value is tested:
if (seq != le32_to_cpu(mmp->mmp_seq)) {
/* fail the mount */
On a LE system such as x86, the *le32* functions do nothing and this
works. Unfortunately, on a BE system such as ppc64, this comparison
becomes:
if (cpu_to_le32(new_seq) != le32_to_cpu(cpu_to_le32(new_seq)) {
/* fail the mount */
Except for a few palindromic sequence numbers, this test always causes
the mount to fail, which makes MMP filesystems generally unmountable
on ppc64. The attached patch fixes this situation.
Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Current logic would print an error message only once, and then
'failed_writes' would stay at 1. Rework the loop to increment
'failed_writes' and print the error message every
s_mmp_update_interval * 60 seconds, as intended according to the
comment.
Signed-off-by: Nikitas Angelinas <nikitas_angelinas@xyratex.com>
Signed-off-by: Andrew Perepechko <andrew_perepechko@xyratex.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Acked-by: Andreas Dilger <adilger@dilger.ca>
sysname holds "Linux" by default, i.e. what appears when doing a "uname
-s"; nodename should be used to print the machine's hostname, i.e. what
is returned when doing a "uname -n" or "hostname", and what
gethostname(2)/sethostname(2) manipulate, in order to notify the
administrator of the node which is contending to mount the filesystem.
Acked-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Nikitas Angelinas <nikitas_angelinas@xyratex.com>
Signed-off-by: Andrew Perepechko <andrew_perepechko@xyratex.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If we create the object and then return failure to the client, we're
left with an unexpected file in the filesystem.
I'm trying to eliminate such cases but not 100% sure I have so an
assertion might be helpful for now.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
As with the nfs4_file, we'd prefer to find out about any failure before
creating a new file rather than after.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Move idr preallocation out of stateid initialization, into stateid
allocation, so that we no longer have to handle any errors from the
former.
This is a little subtle due to the way the idr code manages these
preallocated items--document that in comments.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Creating a new file is an irrevocable step--once it's visible in the
filesystem, other processes may have seen it and done something with it,
and unlinking it wouldn't simply undo the effects of the create.
Therefore, in the case where OPEN creates a new file, we shouldn't do
the create until we know that the rest of the OPEN processing will
succeed.
For example, we should preallocate a struct file in case we need it
until waiting to allocate it till process_open2(), which is already too
late.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
If process_open1() creates a new open owner, but the open later fails,
the current code will leave the open owner around. It won't be on the
close_lru list, and the client isn't expected to send a CLOSE, so it
will hang around as long as the client does.
Similarly, if process_open1() removes an existing open owner from the
close lru, anticipating that an open owner that previously had no
associated stateid's now will, but the open subsequently fails, then
we'll again be left with the same leak.
Fix both problems.
Reported-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
There doesn't seem to be any harm to renewing the client a bit earlier,
when it is looked up. That saves us from having to sprinkle
renew_client calls over quite so many places.
Also remove a redundant comment and do a little cleanup.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Resolved conflicts:
fs/xfs/xfs_trans_priv.h:
- deleted struct xfs_ail field xa_flags
- kept field xa_log_flush in struct xfs_ail
fs/xfs/xfs_trans_ail.c:
- in xfsaild_push(), in XFS_ITEM_PUSHBUF case, replaced
"flush_log = 1" with "ailp->xa_log_flush++"
Signed-off-by: Alex Elder <aelder@sgi.com>
Add a sanity check to make sure ix hasn't gone beyond the valid bounds
of the extent block.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Now build security descriptor to change either owner or group at the
server. Initially security descriptor was built to change only
(D)ACL, that functionality has been extended.
When either an Owner or a Group of a file object at the server is changed,
rest of security descriptor remains same (DACL etc.).
To set security descriptor, it is necessary to open that file
with permission bits of either WRITE_DAC if DACL is being modified or
WRITE_OWNER (Take Ownership) if Owner or Group is being changed.
It is the server that decides whether a set security descriptor with
either owner or group change succeeds or not.
Signed-off-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Steve French <smfrench@gmail.com>
This should be a bitwise negate here. It silences a Sparse warning:
fs/nfsd/nfs4xdr.c:693:16: warning: dubious: x & !y
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Current ore_check_io API receives a residual
pointer, to report partial IO. But it is actually
not used, because in a multiple devices IO there
is never a linearity in the IO failure.
On the other hand if every failing device is reported
through a received callback measures can be taken to
handle only failed devices. One at a time.
This will also be needed by the objects-layout-driver
for it's error reporting facility.
Exofs is not currently using the new information and
keeps the old behaviour of failing the complete IO in
case of an error. (No partial completion)
TODO: Use an ore_check_io callback to set_page_error only
the failing pages. And re-dirty write pages.
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
All users of the ore will need to check if current code
supports the given layout. For example RAID5/6 is not
currently supported.
So move all the checks from exofs/super.c to a new
ore_verify_layout() to be used by ore users.
Note that any new layout should be passed through the
ore_verify_layout() because the ore engine will prepare
and verify some internal members of ore_layout, and
assumes it's called.
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Users like the objlayout-driver would like to only pass
a partial device table that covers the IO in question.
For example exofs divides the file into raid-group-sized
chunks and only serves group_width number of devices at
a time.
The partiality is communicated by setting
ore_componets->first_dev and the array covers all logical
devices from oc->first_dev upto (oc->first_dev + oc->numdevs)
The ore_comp_dev() API receives a logical device index
and returns the actual present device in the table.
An out-of-range dev_index will BUG.
Logical device index is the theoretical device index as if
all the devices of a file are present. .i.e:
total_devs = group_width * mirror_p1 * group_count
0 <= dev_index < total_devs
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Memory conditions and max_bio constraints might cause us to
not comply to the full length of the requested IO. Instead of
failing the complete IO we can issue a shorter read/write and
report how much was actually executed in the ios->length
member.
All users must check ios->length at IO_done or upon return of
ore_read/write and re-issue the reminder of the bytes. Because
other wise there is no error returned like before.
This is part of the effort to support the pnfs-obj layout driver.
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
If at read/write_done the actual IO was shorter then requested,
reported in returned ios->length. It is not an error. The reminder
of the pages should just be unlocked but not marked uptodate or
end_page_writeback. They will be re issued later by the VFS.
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Move the check and preparation of the ios->kern_buff case to
later inside _write_mirror().
Since read was never used with ios->kern_buff its support is removed
instead of fixed.
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Now that each ore_io_state covers only a single raid group.
A single striping_info math is needed. Embed one inside
ore_io_state to cache the calculation results and eliminate
an extra call.
Also the outer _prepare_for_striping is removed since it does nothing.
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Usually a single IO is confined to one group of devices
(group_width) and at the boundary of a raid group it can
spill into a second group. Current code would allocate a
full device_table size array at each io_state so it can
comply to requests that span two groups. Needless to say
that is very wasteful, specially when device_table count
can get very large (hundreds even thousands), while a
group_width is usually 8 or 10.
* Change ore API to trim on IO that spans two raid groups.
The user passes offset+length to ore_get_rw_state, the
ore might trim on that length if spanning a group boundary.
The user must check ios->length or ios->nrpages to see
how much IO will be preformed. It is the responsibility
of the user to re-issue the reminder of the IO.
* Modify exofs To copy spilled pages on to the next IO.
This means one last kick is needed after all coalescing
of pages is done.
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
* 'for-linus' of git://oss.sgi.com/xfs/xfs:
xfs: revert to using a kthread for AIL pushing
xfs: force the log if we encounter pinned buffers in .iop_pushbuf
xfs: do not update xa_last_pushed_lsn for locked items
that let us do local lock checks before requesting to the server.
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Steve French <smfrench@gmail.com>
Rename it for better clarity as to what it does and have the caller pass
in just the single type byte. Turn the if statement into a switch and
optimize it by placing the most common message type at the top. Move the
header length check back into cifs_demultiplex_thread in preparation
for adding a new receive phase and normalize the cFYI messages.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Split cifs_lock into several functions and let CIFSSMBLock get pid
as an argument.
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Steve French <smfrench@gmail.com>
..the length field has only 17 bits.
Cc: <stable@kernel.org>
Acked-by: Jeff Layton <jlayton@samba.org>
Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
Signed-off-by: Steve French <smfrench@gmail.com>
Move the iovec handling entirely into read_from_socket. That simplifies
the code and gets rid of the special handling for header reads. With
this we can also get rid of the "goto incomplete_rcv" label in the main
demultiplex thread function since we can now treat header and non-header
receives the same way.
Also, make it return an int (since we'll never receive enough to worry
about the sign bit anyway), and simply make it return the amount of bytes
read or a negative error code.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Add data structures and functions necessary to map a uid and gid to SID.
These functions are very similar to the ones used to map a SID to uid and gid.
This time, instead of storing sid to id mapping sorted on a sid value,
id to sid is stored, sorted on an id.
A cifs upcall sends an id (uid or gid) and expects a SID structure
in return, if mapping was done successfully.
A failed id to sid mapping to EINVAL.
This patchset aims to enable chown and chgrp commands when
cifsacl mount option is specified, especially to Windows SMB servers.
Currently we can't do that. So now along with chmod command,
chown and chgrp work.
Winbind is used to map id to a SID. chown and chgrp use an upcall
to provide an id to winbind and upcall returns with corrosponding
SID if any exists. That SID is used to build security descriptor.
The DACL part of a security descriptor is not changed by either
chown or chgrp functionality.
cifs client maintains a separate caches for uid to SID and
gid to SID mapping. This is similar to the one used earlier
to map SID to id (as part of ID mapping code).
I tested it by mounting shares from a Windows (2003) server by
authenticating as two users, one at a time, as Administrator and
as a ordinary user.
And then attempting to change owner of a file on the share.
Depending on the permissions/privileges at the server for that file,
chown request fails to either open a file (to change the ownership)
or to set security descriptor.
So it all depends on privileges on the file at the server and what
user you are authenticated as at the server, cifs client is just a
conduit.
I compared the security descriptor during chown command to that
what smbcacls sends when it is used with -M OWNNER: option
and they are similar.
This patchset aim to enable chown and chgrp commands when
cifsacl mount option is specified, especially to Windows SMB servers.
Currently we can't do that. So now along with chmod command,
chown and chgrp work.
I tested it by mounting shares from a Windows (2003) server by
authenticating as two users, one at a time, as Administrator and
as a ordinary user.
And then attempting to change owner of a file on the share.
Depending on the permissions/privileges at the server for that file,
chown request fails to either open a file (to change the ownership)
or to set security descriptor.
So it all depends on privileges on the file at the server and what
user you are authenticated as at the server, cifs client is just a
conduit.
Signed-off-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Suresh had a typo in his recent patch adding information on
the new oplock_endabled parm. Should be documented as in
directory /sys/module/cifs/parameters not /proc/module/cifs/parameters
Signed-off-by: Steve French <smfrench@gmail.com>
Add mount options backupuid and backugid.
It allows an authenticated user to access files with the intent to back them
up including their ACLs, who may not have access permission but has
"Backup files and directories user right" on them (by virtue of being part
of the built-in group Backup Operators.
When mount options backupuid is specified, cifs client restricts the
use of backup intents to the user whose effective user id is specified
along with the mount option.
When mount options backupgid is specified, cifs client restricts the
use of backup intents to the users whose effective user id belongs to the
group id specified along with the mount option.
If an authenticated user is not part of the built-in group Backup Operators
at the server, access to such files is denied, even if allowed by the client.
Signed-off-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
The plan is to deprecate this interface by kernel version 3.4.
Changes since v1
- add a '\n' to the printk.
Reported-by: Alexander Swen <alex@swen.nu>
Cc: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <smfrench@gmail.com>
Reported-by: Alexander Swen <alex@swen.nu>
Cc: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <smfrench@gmail.com>
Thus spake Jeff Layton:
"Making that a module parm would allow you to set that parameter at boot
time without needing to add special startup scripts. IMO, all of the
procfile "switches" under /proc/fs/cifs should be module parms
instead."
This patch doesn't alter the default behavior (Oplocks are enabled by
default).
To disable oplocks when loading the module, use
modprobe cifs enable_oplocks=0
(any of '0' or 'n' or 'N' conventions can be used).
To disable oplocks at runtime using the new interface, use
echo 0 > /sys/module/cifs/parameters/enable_oplocks
The older /proc/fs/cifs/OplockEnabled interface will be deprecated
after two releases. A subsequent patch will add an warning message
about this deprecation.
Changes since v2:
- make enable_oplocks a 'bool'
Changes since v1:
- eliminate the use of extra variable by renaming the old one to
enable_oplocks and make it an 'int' type.
Reported-by: Alexander Swen <alex@swen.nu>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: Steve French <smfrench@gmail.com>
If the server stops sending data while in the middle of sending a
response then we still want to reconnect it if it doesn't come back.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
If msg_controllen is 0, then the socket layer should never touch these
fields. Thus, there's no need to continually reset them. Also, there's
no need to keep this field on the stack for the demultiplex thread, just
make it a local variable in read_from_socket.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
We have two versions of signature generating code. A vectorized and
non-vectorized version. Eliminate a large chunk of cut-and-paste
code by turning the non-vectorized version into a wrapper around the
vectorized one.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
The variable names in this function are so ambiguous that it's very
difficult to know what it's doing. Rename them to make it a bit more
clear.
Also, remove a redundant length check. cifsd checks to make sure that
the rfclen isn't larger than the maximum frame size when it does the
receive.
Finally, change checkSMB to return a real error code (-EIO) when
it finds an error. That will help simplify some coming changes in the
callers.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
server->maxBuf is the maximum SMB size (including header) that the
server can handle. CIFSMaxBufSize is the maximum amount of data (sans
header) that the client can handle. Currently maxBuf is being capped at
CIFSMaxBufSize + the max headers size, and the two values are used
somewhat interchangeably in the code.
This makes little sense as these two values are not related at all.
Separate them and make sure the code uses the right values in the right
places.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
It should be 'CONFIG_CIFS_NFSD_EXPORT'. No-one noticed because that
symbol depends on BROKEN.
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: Steve French <smfrench@gmail.com>
...as that's more efficient when we know that the lengths are equal.
Reported-by: David Howells <dhowells@redhat.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>
Currently pstore write interface employs record id as return
value, but it is not enough because it can't tell caller if
the write operation is successful. Pass the record id back via
an argument pointer and return zero for success, non-zero for
failure.
Signed-off-by: Chen Gong <gong.chen@linux.intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
The result from ipv6_addr_scope() is a set of flags, not a single value,
so we can't just compare the result with IPV6_ADDR_SCOPE_LINKLOCAL.
This patch fixs the problem, and checks for unequal addresses before
scope_id.
Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
When we call xfs_flush_buftarg (generally from sync or umount) it already
is too late to flush the data workqueues, as I/O completion is signalled
for them and we are thus already done with the data we would flush here.
There are places where flushing them might be useful, but the current
sync interface doesn't give us that opportunity.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
The calling convention that returns a pointer to a static buffer is
fairly nasty, so just opencode it in the only caller that is left.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
Use xfs_ioerror_alert instead of opencoding a very similar error
message.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
Instead of passing the block number and mount structure explicitly
get them off the bp and fix make the argument order more natural.
Also move it to xfs_buf.c and stop printing the device name given
that we already get the fs name as part of xfs_alert, and we know
what device is operates on because of the caller that gets printed,
finally rename it to xfs_buf_ioerror_alert and pass __func__ as
argument where it makes sense.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
Change _xfs_buf_initialize to allocate the buffer directly and rename it to
xfs_buf_alloc now that is the only buffer allocation routine. Also remove
the xfs_buf_deallocate wrapper around the kmem_zone_free calls for buffers.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
For each call to xfs_buf_stale we call xfs_buf_delwri_dequeue either
directly before or after it, or are guaranteed by the surrounding
conditionals that we are never called on delwri buffers. Simply
this situation by moving the call to xfs_buf_delwri_dequeue into
xfs_buf_stale.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
The code is unused and under a config option that doesn't exist, remove it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
The code to flush buffers in the umount code is a bit iffy: we first
flush all delwri buffers out, but then might be able to queue up a
new one when logging the sb counts. On a normal shutdown that one
would get flushed out when doing the synchronous superblock write in
xfs_unmountfs_writesb, but we skip that one if the filesystem has
been shut down.
Fix this by moving the delwri list flushing until just before unmounting
the log, and while we're at it also remove the superflous delwri list
and buffer lru flusing for the rt and log device that can never have
cached or delwri buffers.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reported-by: Amit Sahrawat <amit.sahrawat83@gmail.com>
Tested-by: Amit Sahrawat <amit.sahrawat83@gmail.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
Directories are only updated transactionally, which means fsync only
needs to flush the log the inode is currently dirty, but not bother
with checking for dirty data, non-transactional updates, and most
importanly doesn't have to flush disk caches except as part of a
transaction commit.
While the first two optimizations can't easily be measured, the
latter actually makes a difference when doing lots of fsync that do
not actually have to commit the inode, e.g. because an earlier fsync
already pushed the log far enough.
The new xfs_dir_fsync is identical to xfs_nfs_commit_metadata except
for the prototype, but I'm not sure creating a common helper for the
two is worth it given how simple the functions are.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
The AIL push code will issue a log force on ever single push loop
that it exits and has encountered pinned items. It doesn't rescan
these pinned items until it revisits the AIL from the start. Hence
we only need to force the log once per walk from the start of the
AIL to the target LSN.
This results in numbers like this:
xs_push_ail_flush..... 1456
xs_log_force......... 1485
For an 8-way 50M inode create workload - almost all the log forces
are coming from the AIL pushing code.
Reduce the number of log forces by only forcing the log if the
previous walk found pinned buffers. This reduces the numbers to:
xs_push_ail_flush..... 665
xs_log_force......... 682
For the same test.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
Stats show that for an 8-way unlink @ ~80,000 unlinks/s we are doing
~1 million cache hit lookups to ~3000 buffer creates. That's almost
3 orders of magnitude more cahce hits than misses, so optimising for
cache hits is quite important. In the cache hit case, we do not need
to allocate a new buffer in case of a cache miss, so we are
effectively hitting the allocator for no good reason for vast the
majority of calls to _xfs_buf_find. 8-way create workloads are
showing similar cache hit/miss ratios.
The result is profiles that look like this:
samples pcnt function DSO
_______ _____ _______________________________ _________________
1036.00 10.0% _xfs_buf_find [kernel.kallsyms]
582.00 5.6% kmem_cache_alloc [kernel.kallsyms]
519.00 5.0% __memcpy [kernel.kallsyms]
468.00 4.5% __ticket_spin_lock [kernel.kallsyms]
388.00 3.7% kmem_cache_free [kernel.kallsyms]
331.00 3.2% xfs_log_commit_cil [kernel.kallsyms]
Further, there is a fair bit of work involved in initialising a new
buffer once a cache miss has occurred and we currently do that under
the rbtree spinlock. That increases spinlock hold time on what are
heavily used trees.
To fix this, remove the initialisation of the buffer from
_xfs_buf_find() and only allocate the new buffer once we've had a
cache miss. Initialise the buffer immediately after allocating it in
xfs_buf_get, too, so that is it ready for insert if we get another
cache miss after allocation. This minimises lock hold time and
avoids unnecessary allocator churn. The resulting profiles look
like:
samples pcnt function DSO
_______ _____ ___________________________ _________________
8111.00 9.1% _xfs_buf_find [kernel.kallsyms]
4380.00 4.9% __memcpy [kernel.kallsyms]
4341.00 4.8% __ticket_spin_lock [kernel.kallsyms]
3401.00 3.8% kmem_cache_alloc [kernel.kallsyms]
2856.00 3.2% xfs_log_commit_cil [kernel.kallsyms]
2625.00 2.9% __kmalloc [kernel.kallsyms]
2380.00 2.7% kfree [kernel.kallsyms]
2016.00 2.3% kmem_cache_free [kernel.kallsyms]
Showing a significant reduction in time spent doing allocation and
freeing from slabs (kmem_cache_alloc and kmem_cache_free).
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
Dmitry Monakhov