- 23 Mar, 2020 40 commits
-
-
Josef Bacik authored
Since we're not only checking for metadata reservations but also if we need to throttle our delayed ref generation, reorder reserve_metadata_space() above the select_one_root() call in relocate_tree_block(). The reason we want this is because select_reloc_root() will mess with the backref cache, and if we're going to bail we want to be able to cleanly remove this node from the backref cache and come back along to regenerate it. Move it up so this is the first thing we do to make restarting cleaner. Signed-off-by:
Josef Bacik <josef@toxicpanda.com> Reviewed-by:
David Sterba <dsterba@suse.com> Signed-off-by:
-
Josef Bacik authored
Here we are just searching down to the bytenr we're building the backref tree for, and all of it's paths to the roots. These bytenrs are not guaranteed to be anywhere near each other, so readahead just generates extra latency. Reviewed-by:
Qu Wenruo <wqu@suse.com> Signed-off-by:
-
Josef Bacik authored
Readahead will generate a lot of extra reads for adjacent nodes, but when running delayed refs we have no idea if the next ref is going to be adjacent or not, so this potentially just generates a lot of extra IO. To make matters worse each ref is truly just looking for one item, it doesn't generally search forward, so we simply don't need it here. Reviewed-by:
-
Nikolay Borisov authored
With BTRFS_SUBVOL_CREATE_ASYNC support remove it's no longer required to pass the async_transid parameter so remove it and any code using it. Signed-off-by:
Nikolay Borisov <nborisov@suse.com> Reviewed-by:
-
Nikolay Borisov authored
btrfs_ioctl_snap_create_transid no longer takes a transid argument, so remove it and rename the function to __btrfs_ioctl_snap_create to reflect it's an internal, worker function. Signed-off-by:
-
Nikolay Borisov authored
This functionality was deprecated in kernel 5.4. Since no one has complained of the impending removal it's time we did so. Signed-off-by:
-
Josef Bacik authored
Now that we have proper root ref counting everywhere we can kill the subvol_srcu. * removal of fs_info::subvol_srcu reduces size of fs_info by 1176 bytes * the refcount_t used for the references checks for accidental 0->1 in cases where the root lifetime would not be properly protected * there's a leak detector for roots to catch unfreed roots at umount time * SRCU served us well over the years but is was not a proper synchronization mechanism for some cases Signed-off-by:
-
Josef Bacik authored
The radix root is primarily protected by the fs_roots_radix_lock, so use that to lookup and get a ref on all of our fs roots in btrfs_cleanup_fs_roots. The tree reference is taken in the protected section as before. Signed-off-by:
-
Josef Bacik authored
Now that all the users of roots take references for them we can drop the extra root ref we've been taking. Before we had roots at 2 refs for the life of the file system, one for the radix tree, and one simply for existing. Now that we have proper ref accounting in all places that use roots we can drop this extra ref simply for existing as we no longer need it. Signed-off-by:
-
Josef Bacik authored
At the point we add a root to the dead roots list we have no open inodes for that root, so we need to hold a ref on that root to keep it from disappearing. Signed-off-by:
-
Josef Bacik authored
If we make sure all the inodes have refs on their root we don't have to worry about the root disappearing while we have open inodes. Signed-off-by:
-
Josef Bacik authored
There are a few different ways to free roots, either you allocated them yourself and you just do free_extent_buffer(root->node); free_extent_buffer(root->commit_node); btrfs_put_root(root); Which is the pattern for log roots. Or for snapshots/subvolumes that are being dropped you simply call btrfs_free_fs_root() which does all the cleanup for you. Unify this all into btrfs_put_root(), so that we don't free up things associated with the root until the last reference is dropped. This makes the root freeing code much more significant. The only caveat is at close_ctree() time we have to free the extent buffers for all of our main roots (extent_root, chunk_root, etc) because we have to drop the btree_inode and we'll run into issues if we hold onto those nodes until ->kill_sb() time. This will be addressed in the future when we kill the btree_inode. Signed-off-by:
-
Josef Bacik authored
We are going to make root life be controlled soley by refcounting, and inodes will be one of the things that hold a ref on the root. This means we need to handle dropping the ino_cache_inode outside of the root freeing logic, so move it into btrfs_drop_and_free_fs_root() so it is cleaned up properly on unmount. Signed-off-by:
-
Josef Bacik authored
I'm going to make the entire destruction of btrfs_root's controlled by their refcount, so it will be helpful to notice if we're leaking their eb's on umount. Signed-off-by:
-
Josef Bacik authored
This was pretty subtle, we default to reloc roots having 0 root refs, so if we crash in the middle of the relocation they can just be deleted. If we successfully complete the relocation operations we'll set our root refs to 1 in prepare_to_merge() and then go on to merge_reloc_roots(). At prepare_to_merge() time if any of the reloc roots have a 0 reference still, we will remove that reloc root from our reloc root rb tree, and then clean it up later. However this only happens if we successfully start a transaction. If we've aborted previously we will skip this step completely, and only have reloc roots with a reference count of 0, but were never properly removed from the reloc control's rb tree. This isn't a problem per-se, our references are held by the list the reloc roots are on, and by the original root the reloc root belongs to. If we end up in this situation all the reloc roots will be added to the dirty_reloc_list, and then properly dropped at that point. The reloc control will be free'd and the rb tree is no longer used. There were two options when fixing this, one was to remove the BUG_ON(), the other was to make prepare_to_merge() handle the case where we couldn't start a trans handle. IMO this is the cleaner solution. I started with handling the error in prepare_to_merge(), but it turned out super ugly. And in the end this BUG_ON() simply doesn't matter, the cleanup was happening properly, we were just panicing because this BUG_ON() only matters in the success case. So I've opted to just remove it and add a comment where it was. Reviewed-by:
-
Josef Bacik authored
We previously were relying on root->reloc_root to be cleaned up by the drop snapshot, or the error handling. However if btrfs_drop_snapshot() failed it wouldn't drop the ref for the root. Also we sort of depend on the right thing to happen with moving reloc roots between lists and the fs root they belong to, which makes it hard to figure out who owns the reference. Fix this by explicitly holding a reference on the reloc root for roo->reloc_root. This means that we hold two references on reloc roots, one for whichever reloc_roots list it's attached to, and the root->reloc_root we're on. This makes it easier to reason out who owns a reference on the root, and when it needs to be dropped. Reviewed-by:
-
Josef Bacik authored
The DEAD_RELOC_TREE flag is in place in order to avoid a use after free in init_reloc_root, tracking the presence of reloc_root. However adding the explicit tree references in previous patches makes the use after free impossible because at this point we no longer have a reloc_control set on the fs_info and thus cannot enter the function. So move this to be coupled with clearing the root->reloc_root so we're consistent with all other operations of the reloc root. Signed-off-by:
-
Josef Bacik authored
If we have an error while processing the reloc roots we could leak roots that were added to rc->reloc_roots before we hit the error. We could have also not removed the reloc tree mapping from our rb_tree, so clean up any remaining nodes in the reloc root rb_tree. Signed-off-by:
-
Josef Bacik authored
We previously were checking if the root had a dead root before accessing root->reloc_root in order to avoid a use-after-free type bug. However this scenario happens after we've unset the reloc control, so we would have been saved if we'd simply checked for fs_info->reloc_control. At this point during relocation we no longer need to be creating new reloc roots, so simply move this check above the reloc_root checks to avoid any future races and confusion. Reviewed-by:
-
Josef Bacik authored
If we do merge_reloc_roots() we could insert a few roots onto the dirty subvol roots list, where we hold a ref on them. If we fail to start the transaction we need to run clean_dirty_subvols() in order to cleanup the refs. CC: stable@vger.kernel.org # 5.4+ Signed-off-by:
-
Josef Bacik authored
If we fail to load an fs root, or fail to start a transaction we can bail without unsetting the reloc control, which leads to problems later when we free the reloc control but still have it attached to the file system. In the normal path we'll end up calling unset_reloc_control() twice, but all it does is set fs_info->reloc_control = NULL, and we can only have one balance at a time so it's not racey. CC: stable@vger.kernel.org # 5.4+ Reviewed-by:
-
Josef Bacik authored
If we have an error while building the backref tree in relocation we'll process all the pending edges and then free the node. However if we integrated some edges into the cache we'll lose our link to those edges by simply freeing this node, which means we'll leak memory and references to any roots that we've found. Instead we need to use remove_backref_node(), which walks through all of the edges that are still linked to this node and free's them up and drops any root references we may be holding. CC: stable@vger.kernel.org # 4.9+ Reviewed-by:
-
Qu Wenruo authored
In relocation, we need to locate all parent tree leaves referring to one data extent, thus we have a complex mechanism to iterate throught extent tree and subvolume trees to locate the related leaves. However this is already done in backref.c, we have btrfs_find_all_leafs(), which can return a ulist containing all leaves referring to that data extent. Use btrfs_find_all_leafs() to replace find_data_references(). There is a special handling for v1 space cache data extents, where we need to delete the v1 space cache data extents, to avoid those data extents to hang the data relocation. In this patch, the special handling is done by re-iterating the root tree leaf. Although it's a little less efficient than the old handling, considering we can reuse a lot of code, it should be acceptable. Signed-off-by:
-
Josef Bacik authored
While debugging I noticed I wasn't getting ref verify errors before everything blew up. Turns out it's because we don't warn when we try to add a normal ref via btrfs_inc_ref() if the block entry exists but has 0 references. This is incorrect, we should never be doing anything other than adding a new extent once a block entry drops to 0 references. Signed-off-by:
-
Filipe Manana authored
Commit 0c713cba ("Btrfs: fix race between ranged fsync and writeback of adjacent ranges") fixed a bug where we could end up with file extent items in a log tree that represent file ranges that overlap due to a race between the hole detection of a ranged full fsync and writeback for a different file range. The problem was solved by forcing any ranged full fsync to become a non-ranged full fsync - setting the range start to 0 and the end offset to LLONG_MAX. This was a simple solution because the code that detected and marked holes was very complex, it used to be done at copy_items() and implied several searches on the fs/subvolume tree. The drawback of that solution was that we started to flush delalloc for the entire file and wait for all the ordered extents to complete for ranged full fsyncs (including ordered extents covering ranges completely outside the given range). Fortunatelly ranged full fsyncs are not the most common case (hopefully for most workloads). However a later fix for detecting and marking holes was made by commit 0e56315c ("Btrfs: fix missing hole after hole punching and fsync when using NO_HOLES") and it simplified a lot the detection of holes, and now copy_items() no longer does it and we do it in a much more simple way at btrfs_log_holes(). This makes it now possible to simply make the code that detects holes to operate only on the initial range and no longer need to operate on the whole file, while also avoiding the need to flush delalloc for the entire file and wait for ordered extents that cover ranges that don't overlap the given range. Another special care is that we must skip file extent items that fall entirely outside the fsync range when copying inode items from the fs/subvolume tree into the log tree - this is to avoid races with ordered extent completion for extents falling outside the fsync range, which could cause us to end up with file extent items in the log tree that have overlapping ranges - for example if the fsync range is [1Mb, 2Mb], when we copy inode items we could copy an extent item for the range [0, 512K], then release the search path and before moving to the next leaf, an ordered extent for a range of [256Kb, 512Kb] completes - this would cause us to copy the new extent item for range [256Kb, 512Kb] into the log tree after we have copied one for the range [0, 512Kb] - the extents overlap, resulting in a corruption. So this change just does these steps: 1) When the NO_HOLES feature is enabled it leaves the initial range intact - no longer sets it to [0, LLONG_MAX] when the full sync bit is set in the inode. If NO_HOLES is not enabled, always set the range to a full, just like before this change, to avoid missing file extent items representing holes after replaying the log (for both full and fast fsyncs); 2) Make the hole detection code to operate only on the fsync range; 3) Make the code that copies items from the fs/subvolume tree to skip copying file extent items that cover a range completely outside the range of the fsync. Reviewed-by:
Filipe Manana <fdmanana@suse.com> Signed-off-by:
-
Filipe Manana authored
The function btrfs_log_inode() is quite large and so is its loop which iterates the inode items from the fs/subvolume tree and copies them into a log tree. Because this is a large loop inside a very large function and because an upcoming patch in this series needs to add some more logic inside that loop, move the loop into a helper function to make it a bit more manageable. Reviewed-by:
-
Filipe Manana authored
Getting the end offset for a file extent item requires a bit of code since the extent can be either inline or regular/prealloc. There are some places all over the code base that open code this logic and in another patch later in this series it will be needed again. Therefore encapsulate this logic in a helper function and use it. Reviewed-by:
-
Filipe Manana authored
When doing a fast fsync for a range that starts at an offset greater than zero, we can end up with a log that when replayed causes the respective inode miss a file extent item representing a hole if we are not using the NO_HOLES feature. This is because for fast fsyncs we don't log any extents that cover a range different from the one requested in the fsync. Example scenario to trigger it: $ mkfs.btrfs -O ^no-holes -f /dev/sdd $ mount /dev/sdd /mnt # Create a file with a single 256K and fsync it to clear to full sync # bit in the inode - we want the msync below to trigger a fast fsync. $ xfs_io -f -c "pwrite -S 0xab 0 256K" -c "fsync" /mnt/foo # Force a transaction commit and wipe out the log tree. $ sync # Dirty 768K of data, increasing the file size to 1Mb, and flush only # the range from 256K to 512K without updating the log tree # (sync_file_range() does not trigger fsync, it only starts writeback # and waits for it to finish). $ xfs_io -c "pwrite -S 0xcd 256K 768K" /mnt/foo $ xfs_io -c "sync_range -abw 256K 256K" /mnt/foo # Now dirty the range from 768K to 1M again and sync that range. $ xfs_io -c "mmap -w 768K 256K" \ -c "mwrite -S 0xef 768K 256K" \ -c "msync -s 768K 256K" \ -c "munmap" \ /mnt/foo <power fail> # Mount to replay the log. $ mount /dev/sdd /mnt $ umount /mnt $ btrfs check /dev/sdd Opening filesystem to check... Checking filesystem on /dev/sdd UUID: 482fb574-b288-478e-a190-a9c44a78fca6 [1/7] checking root items [2/7] checking extents [3/7] checking free space cache [4/7] checking fs roots root 5 inode 257 errors 100, file extent discount Found file extent holes: start: 262144, len: 524288 ERROR: errors found in fs roots found 720896 bytes used, error(s) found total csum bytes: 512 total tree bytes: 131072 total fs tree bytes: 32768 total extent tree bytes: 16384 btree space waste bytes: 123514 file data blocks allocated: 589824 referenced 589824 Fix this issue by setting the range to full (0 to LLONG_MAX) when the NO_HOLES feature is not enabled. This results in extra work being done but it gives the guarantee we don't end up with missing holes after replaying the log. CC: stable@vger.kernel.org # 4.19+ Reviewed-by: -
Nikolay Borisov authored
Instead of iterating all pending tickets on the normal/priority list to sum their total size the cost can be amortized across ticket addition/ removal. This turns O(n) + O(m) (where n is the size of the normal list and m of the priority list) into O(1). This will mostly have effect in workloads that experience heavy flushing. Signed-off-by:
-
Roman Gushchin authored
Currently btrfs doesn't provide a migratepage callback for data pages. It means that fallback_migrate_page() is used to migrate btrfs pages. fallback_migrate_page() cannot move dirty pages, instead it tries to flush them (in sync mode) or just fails (in async mode). In the sync mode pages which are scheduled to be processed by btrfs_writepage_fixup_worker() can't be effectively flushed by the migration code, because there is no established way to wait for the completion of the delayed work. It all leads to page migration failures. To fix it the patch implements a btrs-specific migratepage callback, which is similar to iomap_migrate_page() used by some other fs, except it does take care of the PagePrivate2 flag which is used for data ordering purposes. Reviewed-by:
Chris Mason <clm@fb.com> Signed-off-by:
Roman Gushchin <guro@fb.com> Signed-off-by:
-
Nikolay Borisov authored
It's no longer used following 30d40577 ("btrfs: reloc: Also queue orphan reloc tree for cleanup to avoid BUG_ON()"), so just remove it. Signed-off-by:
-
Nikolay Borisov authored
Currently the non-prefixed version is a simple wrapper used to hide the 4th argument of the prefixed version. This doesn't bring much value in practice and only makes the code harder to follow by adding another level of indirection. Rectify this by removing the __ prefix and have only one public function to release bytes from a block reservation. No semantic changes. Signed-off-by:
-
Qu Wenruo authored
When relocating data block groups with tons of small extents, or large metadata block groups, there can be over 200,000 extents. We will iterate all extents of such block group in relocate_block_group(), where iteration itself can be kinda time-consuming. So when user want to cancel the balance, the extent iteration loop can be another target. This patch will add the cancelling check in the extent iteration loop of relocate_block_group() to make balance cancelling faster. Reviewed-by:
Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by:
-
Qu Wenruo authored
When relocating a data extents with large large data extents, we spend most of our time in relocate_file_extent_cluster() at stage "moving data extents": 1) | btrfs_relocate_block_group [btrfs]() { 1) | relocate_file_extent_cluster [btrfs]() { 1) $ 6586769 us | } 1) + 18.260 us | relocate_file_extent_cluster [btrfs](); 1) + 15.770 us | relocate_file_extent_cluster [btrfs](); 1) $ 8916340 us | } 1) | btrfs_relocate_block_group [btrfs]() { 1) | relocate_file_extent_cluster [btrfs]() { 1) $ 11611586 us | } 1) + 16.930 us | relocate_file_extent_cluster [btrfs](); 1) + 15.870 us | relocate_file_extent_cluster [btrfs](); 1) $ 14986130 us | } To make data relocation cancelling quicker, add extra balance cancelling check after each page read in relocate_file_extent_cluster(). Cleanup and error handling uses the same mechanism as if the whole process finished Reviewed-by: -
Qu Wenruo authored
Introduce a new error injection point, should_cancel_balance(). It's just a wrapper of atomic_read(&fs_info->balance_cancel_req), but allows us to override the return value. Currently there are only one locations using this function: - btrfs_balance() It checks cancel before each block group. There are other locations checking fs_info->balance_cancel_req, but they are not used as an indicator to exit, so there is no need to use the wrapper. But there will be more locations coming, and some locations can cause kernel panic if not handled properly. So introduce this error injection to provide better test interface. Reviewed-by:
-
Filipe Manana authored
There are a few cases where we don't allow cloning an inline extent into the destination inode, returning -EOPNOTSUPP to user space. This was done to prevent several types of file corruption and because it's not very straightforward to deal with these cases, as they can't rely on simply copying the inline extent between leaves. Such cases require copying the inline extent's data into the respective page of the destination inode. Not supporting these cases makes it harder and more cumbersome to write applications/libraries that work on any filesystem with reflink support, since all these cases for which btrfs fails with -EOPNOTSUPP work just fine on xfs for example. These unsupported cases are also not documented anywhere and explaining which exact cases fail require a bit of too technical understanding of btrfs's internal (inline extents and when and where can they exist in a file), so it's not really user friendly. Also some test cases from fstests that use fsx, such as generic/522 for example, can sporadically fail because they trigger one of these cases, and fsx expects all operations to succeed. This change adds supports for cloning all these cases by copying the inline extent's data into the respective page of the destination inode. With this change test case btrfs/112 from fstests fails because it expects some clone operations to fail, so it will be updated. Also a new test case that exercises all these previously unsupported cases will be added to fstests. Signed-off-by:
-
Filipe Manana authored
We can not reflink parts of an inline extent, we must always reflink the whole inline extent. We know that inline extents always start at file offset 0 and that can never represent an amount of data larger then the filesystem's sector size (both compressed and uncompressed). We also have had the constraints that reflink operations must have a start offset that is aligned to the sector size and an end offset that is also aligned or it ends the inode's i_size, so there's no way for user space to be able to do a reflink operation that will refer to only a part of an inline extent. Initially there was a bug in the inlining code that could allow compressed inline extents that encoded more than 1 page, but that was fixed in 2008 by commit 70b99e69 ("Btrfs: Compression corner fixes") since that was problematic. So remove all the extent cloning code that deals with the possibility of cloning only partial inline extents. Reviewed-by:
-
Filipe Manana authored
The reflink code is quite large and has been living in ioctl.c since ever. It has grown over the years after many bug fixes and improvements, and since I'm planning on making some further improvements on it, it's time to get it better organized by moving into its own file, reflink.c (similar to what xfs does for example). This change only moves the code out of ioctl.c into the new file, it doesn't do any other change. Reviewed-by:
-
Gustavo A. R. Silva authored
The current codebase makes use of the zero-length array language extension to the C90 standard, but the preferred mechanism to declare variable-length types such as these ones is a flexible array member[1][2], introduced in C99: struct foo { int stuff; struct boo array[]; }; By making use of the mechanism above, we will get a compiler warning in case the flexible array does not occur last in the structure, which will help us prevent some kind of undefined behavior bugs from being inadvertently introduced[3] to the codebase from now on. Also, notice that, dynamic memory allocations won't be affected by this change: "Flexible array members have incomplete type, and so the sizeof operator may not be applied. As a quirk of the original implementation of zero-length arrays, sizeof evaluates to zero." [1] This issue was found with the help of Coccinelle. [1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html [2] https://github.com/KSPP/linux/issues/21 [3] commit 76497732 ("cxgb3/l2t: Fix undefined behaviour") Signed-off-by:Gustavo A. R. Silva <gustavo@embeddedor.com> Reviewed-by:
-
Gustavo A. R. Silva authored
The current codebase makes use of the zero-length array language extension to the C90 standard, but the preferred mechanism to declare variable-length types such as these ones is a flexible array member[1][2], introduced in C99: struct foo { int stuff; struct boo array[]; }; By making use of the mechanism above, we will get a compiler warning in case the flexible array does not occur last in the structure, which will help us prevent some kind of undefined behavior bugs from being inadvertently introduced[3] to the codebase from now on. Also, notice that, dynamic memory allocations won't be affected by this change: "Flexible array members have incomplete type, and so the sizeof operator may not be applied. As a quirk of the original implementation of zero-length arrays, sizeof evaluates to zero." [1] This issue was found with the help of Coccinelle. [1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html [2] https://github.com/KSPP/linux/issues/21 [3] commit 76497732 ("cxgb3/l2t: Fix undefined behaviour") Signed-off-by:
-
