- 01 Sep, 2013 40 commits
-
-
Josef Bacik authored
I added a patch where we started taking the ordered operations mutex when we waited on ordered extents. We need this because we splice the list and process it, so if a flusher came in during this scenario it would think the list was empty and we'd usually get an early ENOSPC. The problem with this is that this lock is used in transaction committing. So we end up with something like this Transaction commit -> wait on writers Delalloc flusher -> run_ordered_operations (holds mutex) ->wait for filemap-flush to do its thing flush task -> cow_file_range ->wait on btrfs_join_transaction because we're commiting some other task -> commit_transaction because we notice trans->transaction->flush is set -> run_ordered_operations (hang on mutex) We need to disentangle the ordered operations flushing from the delalloc flushing, since they are separate things. This solves the deadlock issue I was seeing. Thanks, Signed-off-by:
Josef Bacik <jbacik@fusionio.com> Signed-off-by:
Chris Mason <chris.mason@fusionio.com>
-
Josef Bacik authored
There are several places where we BUG_ON() if we fail to remove the orphan items and such, which is not ok, so remove those and either abort or just carry on. This also fixes a problem where if we couldn't start a transaction we wouldn't actually remove the orphan item reserve for the inode. Thanks, Signed-off-by:
-
Josef Bacik authored
Eric pointed out that btrfs will happily allow you to delete the default subvol. This is a problem obviously since the next time you go to mount the file system it will freak out because it can't find the root. Fix this by adding a check to see if our default subvol points to the subvol we are trying to delete, and if it does not allowing it to happen. Thanks, Signed-off-by:
-
Josef Bacik authored
We have logic to see if we've already created a parent directory by check to see if an inode inside of that directory has a lower inode number than the one we are currently processing. The logic is that if there is a lower inode number then we would have had to made sure the directory was created at that previous point. The problem is that subvols inode numbers count from the lowest objectid in the root tree, which may be less than our current progress. So just skip if our dir item key is a root item. This fixes the original test and the xfstest version I made that added an extra subvol create. Thanks, Reported-by:
Emil Karlson <jekarlson@gmail.com> Signed-off-by:
-
Mark Fasheh authored
This patch adds an ioctl, BTRFS_IOC_FILE_EXTENT_SAME which will try to de-duplicate a list of extents across a range of files. Internally, the ioctl re-uses code from the clone ioctl. This avoids rewriting a large chunk of extent handling code. Userspace passes in an array of file, offset pairs along with a length argument. The ioctl will then (for each dedupe) do a byte-by-byte comparison of the user data before deduping the extent. Status and number of bytes deduped are returned for each operation. Signed-off-by:
Mark Fasheh <mfasheh@suse.de> Reviewed-by:
Zach Brown <zab@redhat.com> Signed-off-by:
-
Mark Fasheh authored
We want this for btrfs_extent_same. Basically readpage and friends do their own extent locking but for the purposes of dedupe, we want to have both files locked down across a set of readpage operations (so that we can compare data). Introduce this variant and a flag which can be set for extent_read_full_page() to indicate that we are already locked. Partial credit for this patch goes to Gabriel de Perthuis <g2p.code@gmail.com> as I have included a fix from him to the original patch which avoids a deadlock on compressed extents. Signed-off-by:
-
Mark Fasheh authored
There's some 250+ lines here that are easily encapsulated into their own function. I don't change how anything works here, just create and document the new btrfs_clone() function from btrfs_ioctl_clone() code. Signed-off-by:
-
Mark Fasheh authored
The range locking in btrfs_ioctl_clone is trivially broken out into it's own function. This reduces the complexity of btrfs_ioctl_clone() by a small bit and makes that locking code available to future functions in fs/btrfs/ioctl.c Signed-off-by:
-
Wang Shilong authored
Signed-off-by:
Wang Shilong <wangsl.fnst@cn.fujitsu.com> Reviewed-by:
Miao Xie <miaox@cn.fujitsu.com> Signed-off-by:
-
Filipe David Borba Manana authored
In extent-tree.c:do_chunk_alloc(), early on we returned 0 (success) when the target space was full and when chunk allocation is needed. However, later on in that same function we return ENOSPC if btrfs_alloc_chunk() fails (and chunk allocation was needed) and set the space's full flag. This was inconsistent, as -ENOSPC should be returned if the space is full and a chunk allocation needs to performed. If the space is full but no chunk allocation is needed, just return 0 (success). Signed-off-by:
Filipe David Borba Manana <fdmanana@gmail.com> Signed-off-by:
-
Filipe David Borba Manana authored
tree-log.c was ignoring the return value from btrfs_run_delayed_items() in several places. Signed-off-by:
-
Filipe David Borba Manana authored
In tree-log.c:replay_one_name(), if memory allocation for the name fails, ensure we iput the dir inode we got before before we return. Signed-off-by:
-
Liu Bo authored
The rule originally comes from nocow writing, but snapshot-aware defrag is a different case, the extent has been writen and we're not going to change the extent but add a reference on the data. So we're able to allow such compressed extents to be merged into one bigger extent if they're pointing to the same data. Reviewed-by:
Liu Bo <bo.li.liu@oracle.com> Signed-off-by:
-
David Sterba authored
I'ts hardcoded to 30 seconds which is fine for most users. Higher values defer data being synced to permanent storage with obvious consequences when the system crashes. The upper bound is not forced, but a warning is printed if it's more than 300 seconds (5 minutes). Signed-off-by:
David Sterba <dsterba@suse.cz> Signed-off-by:
-
Josef Bacik authored
There is no reason we can't just set the path to blocking and then do normal GFP_NOFS allocations for these extent buffers. Thanks, Signed-off-by:
-
Josef Bacik authored
We can get ENOMEM trying to allocate dummy bufs for the rewind operation of the tree mod log. Instead of BUG_ON()'ing in this case pass up ENOMEM. I looked back through the callers and I'm pretty sure I got everybody who did BUG_ON(ret) in this path. Thanks, Signed-off-by:
-
Josef Bacik authored
When doing a send with a parent subvol we will check to see if the file we are acting on is being overwritten and move it if we think it may be needed further down the line during the send. We check this by checking its directory and making sure it existed in the parent and making sure the file existed in the parent. The problem with this check is that if we create a directory and a file in that directory, and then snapshot, and then remove and re-create that same directory and file with different inode numbers and then try to snapshot and send with the original parent we will try and save the original file inside of that directory. This is a problem because during the receive we move the directory out of the way because it is a completely new inode, which makes us unable to find the old file inside of the directory when we try to move that out of the way for the overwrite. We fix this by checking the parent directory of the inode we think we are overwriting. If the parent directory generation in the send root != the parent directory generation in the parent root then we know it is a completely new directory and we need not bother with moving the file out of the way because it would have been completely destroyed. This fixes bz 60673. Thanks, Signed-off-by:
-
Josef Bacik authored
Alex Lyakas reported a bug where wait_block_group_cache_progress() would wait forever if a drive failed. This is because we just bail out if there is an error while trying to cache a block group, we don't update anybody who may be waiting. So this introduces a new enum for the cache state in case of error and makes everybody bail out if we have an error. Alex tested and verified this patch fixed his problem. This fixes bz 59431. Thanks, Signed-off-by:
-
Filipe David Borba Manana authored
Signed-off-by:
-
Dave Jones authored
If we bail out when the stripe alloc fails, we need to undo the earlier allocation of raid_map. Signed-off-by:
Dave Jones <davej@redhat.com> Signed-off-by:
-
Filipe David Borba Manana authored
Signed-off-by:
Jan Schmidt <list.btrfs@jan-o-sch.net> Signed-off-by:
-
Filipe David Borba Manana authored
After reading all device items from the chunk tree, don't exit the loop and then navigate down the tree again to find the chunk items. Instead just read all device items and chunk items with a single tree search. This is possible because all device items are found before any chunk item in the chunks tree. Signed-off-by:
-
Josef Bacik authored
There is no reason for this sort of jackassery. Thanks, Signed-off-by:
-
Josef Bacik authored
Previously we only added blocks to the list to have their backrefs checked if the level of the block is right above the one we are searching for. This is because we want to make sure we don't add the entire path up to the root to the lists to make sure we process things one at a time. This assumes that if any blocks in the path to the root are going to be not checked (shared in other words) then they will be in the level right above the current block on up. This isn't quite right though since we can have blocks higher up the list that are shared because they are attached to a reloc root. But we won't add this block to be checked and then later on we will BUG_ON(!upper->checked). So instead keep track of wether or not we've queued a block to be checked in this current search, and if we haven't go ahead and queue it to be checked. This patch fixed the panic I was seeing where we BUG_ON(!upper->checked). Thanks, Signed-off-by:
-
Josef Bacik authored
If our item isn't big enough to have an actual inline item when we have skinny metadata enabled just return 1 in find_inline_backref so we can move on to the next item. This probably wasn't causing a problem since we check the values of ptr and end properly, but just in case this will keep us from doing extra work. Thanks, Signed-off-by:
-
Josef Bacik authored
First of all we no longer set EXTENT_DIRTY when we dirty an extent so this patch removes the clearing of EXTENT_DIRTY since it confuses me. This patch also adds clearing EXTENT_DEFRAG and also doing EXTENT_DO_ACCOUNTING when we have errors. This is because if we are clearing delalloc without adding an ordered extent then we need to make sure the enospc handling stuff is accounted for. Also if this range was DEFRAG we need to make sure that bit is cleared so we dont leak it. Thanks, Signed-off-by:
-
Josef Bacik authored
This patch removes the io_tree argument for extent_clear_unlock_delalloc since we always use &BTRFS_I(inode)->io_tree, and it separates out the extent tree operations from the page operations. This way we just pass in the extent bits we want to clear and then pass in the operations we want done to the pages. This is because I'm going to fix what extent bits we clear in some cases and rather than add a bunch of new flags we'll just use the actual extent bits we want to clear. Thanks, Signed-off-by:
-
Anand Jain authored
Signed-off-by:
Anand Jain <anand.jain@oracle.com> Signed-off-by:
-
Miao Xie authored
When we read several pages at once, we needn't get the extent map object every time we deal with a page, and we can cache the extent map object. So, we can reduce the search time of the extent map, and besides that, we also can reduce the lock contention of the extent map tree. Signed-off-by:
-
Miao Xie authored
In the past, we cached the checksum value in the extent state object, so we had to split the extent state object by the block size, or we had no space to keep this checksum value. But it increased the lock contention of the extent state tree. Now we removed this limit by caching the checksum into the bio object, so it is unnecessary to do the extent state operations by the block size, we can do it in batches, in this way, we can reduce the lock contention of the extent state tree. Signed-off-by:
-
Miao Xie authored
Before applying this patch, we set the uptodate flag and unlock the extent by the page size, it is unnecessary, we can do it in batches, it can reduce the lock contention of the extent state tree. Signed-off-by:
-
Miao Xie authored
Before applying this patch, we cached the csum value into the extent state tree when reading some data from the disk, this operation increased the lock contention of the state tree. Now, we just store the csum value into the bio structure or other unshared structure, so we can reduce the lock contention. Signed-off-by:
-
Miao Xie authored
This patch add some branch prediction hints into the end IO function of the read page, it reduced the percentage of the branch misses from 5.5% to 4.9%. Signed-off-by:
-
Miao Xie authored
Signed-off-by:
-
Wang Shilong authored
Some options are missing in btrfs_show_options(), this patch adds them. Signed-off-by:
-
Wang Shilong authored
Although for most time, int is enough for subvolid, we should ensure safety in theory. Signed-off-by:
-
Wang Shilong authored
I just notice the following commands succeed: mount <dev> <mnt> -o thread_pool=-1 This is ridiculous, only positive thread_pool makes sense,this patch adds sanity checks for them, and also catches the error of ENOMEM if allocating memory fails. Signed-off-by:
-
Miao Xie authored
Signed-off-by:
-
Dan Carpenter authored
The alloc_rbio() frees "raid_map" and "bbio" on error, so there is a potential double free bug in raid56_parity_write(). The raid56_parity_write() and raid56_parity_recover() functions should still free "raid_map" and "bbio" on error if other errors occur though, so I have added some more calls to kfree(). Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by:
-
Josef Bacik authored
We can end up with inodes on the auto defrag list that exist on roots that are going to be deleted. This is extra work we don't need to do, so just bail if our root has 0 root refs. Thanks, Signed-off-by:
-
