- 17 Apr, 2023 40 commits
-
-
Filipe Manana authored
At btrfs_update_global_block_rsv(), we are assuming an unlink operation uses 5 metadata units, but that's not true anymore, it uses 6 since the commit bca4ad7c ("btrfs: reserve correct number of items for unlink and rmdir"). So update the code and comments to consider 6 units. Reviewed-by:
Josef Bacik <josef@toxicpanda.com> Signed-off-by:
Filipe Manana <fdmanana@suse.com> Reviewed-by:
David Sterba <dsterba@suse.com> Signed-off-by:
-
Filipe Manana authored
When evicting an inode, we are incorrectly calculating the amount of space required for a single delayed reference in case the free space tree is enabled. We have to multiply by 2 the result of btrfs_calc_insert_metadata_size(). We should be calculating according to the size update and space release of the delayed block reserve logic at btrfs_update_delayed_refs_rsv() and btrfs_delayed_refs_rsv_release(). Fix this by using the btrfs_calc_delayed_ref_bytes() helper at evict_refill_and_join() instead of btrfs_calc_insert_metadata_size(). Reviewed-by:
-
Filipe Manana authored
Instead of duplicating the logic for calculating how much space is required for a given number of delayed references, add an inline helper to encapsulate that logic and use it everywhere we are calculating the space required. Reviewed-by:
-
Filipe Manana authored
Now that btrfs_calc_insert_metadata_size() can take a const fs_info argument, make the fs_info argument of calc_reclaim_items_nr() and of calc_delayed_refs_nr() const as well. Reviewed-by:
-
Filipe Manana authored
The fs_info argument of the helpers btrfs_calc_insert_metadata_size() and btrfs_calc_metadata_size() is not modified so it can be const. This will also allow a new helper function in one of the next patches to have its fs_info argument as const. Reviewed-by:
-
Filipe Manana authored
When flushing a limited number of delayed references (FLUSH_DELAYED_REFS_NR state), we are assuming each delayed reference is holding a number of bytes matching the needed space for inserting for a single metadata item (the result of btrfs_calc_insert_metadata_size()). That is not correct when using the free space tree, as in that case we have to multiply that value by 2 since we need to touch the free space tree as well. This is the same computation as we do at btrfs_update_delayed_refs_rsv() and at btrfs_delayed_refs_rsv_release(). So correct the computation for the amount of delayed references we need to flush in case we have the free space tree. This does not fix a functional issue, instead it makes the flush code flush less delayed references, only the minimum necessary to satisfy a ticket. Reviewed-by:
-
Filipe Manana authored
When refilling the delayed block reserve we are incorrectly computing the amount of bytes for a single delayed reference if the free space tree is being used. In that case we should double the calculated amount. Everywhere else we compute the correct amount, like when updating the delayed block reserve, at btrfs_update_delayed_refs_rsv(), or when releasing space from the delayed block reserve, at btrfs_delayed_refs_rsv_release(). So fix btrfs_delayed_refs_rsv_refill() to multiply the amount of bytes for a single delayed reference by two in case the free space tree is used. Reviewed-by:
-
Filipe Manana authored
During inode eviction, if we are truncating a deleted inode, we don't add delayed items for our inode, so there's no need to throttle on delayed items on each iteration of the loop that truncates inode items from its subvolume tree. But we dirty extent buffers from its subvolume tree, so we only need to throttle on btree inode dirty pages. So use btrfs_btree_balance_dirty_nodelay() in the loop that truncates inode items. Reviewed-by:
-
Filipe Manana authored
We have this logic encapsulated in btrfs_should_throttle_delayed_refs() where we try to estimate if running the current amount of delayed references we have will take more than half a second, and if so, the caller btrfs_should_throttle_delayed_refs() should do something to prevent more and more delayed refs from being accumulated. This logic was added in commit 0a2b2a84 ("Btrfs: throttle delayed refs better") and then further refined in commit a79b7d4b ("Btrfs: async delayed refs"). The idea back then was that the caller of btrfs_should_throttle_delayed_refs() would release its transaction handle (by calling btrfs_end_transaction()) when that function returned true, then btrfs_end_transaction() would trigger an async job to run delayed references in a workqueue, and later start/join a transaction again and do more work. However we don't run delayed references asynchronously anymore, that was removed in commit db2462a6 ("btrfs: don't run delayed refs in the end transaction logic"). That makes the logic that tries to estimate how long we will take to run our current delayed references, at btrfs_should_throttle_delayed_refs(), pointless as we don't take any action to run delayed references anymore. We do have other type of throttling, which consists of checking the size and reserved space of the delayed and global block reserves, as well as if fluhsing delayed references for the current transaction was already started, etc - this is all done by btrfs_should_end_transaction(), and the only user of btrfs_should_throttle_delayed_refs() does periodically call btrfs_should_end_transaction(). So remove btrfs_should_throttle_delayed_refs() and the infrastructure that keeps track of the average time used for running delayed references, as well as adapting btrfs_truncate_inode_items() to call btrfs_check_space_for_delayed_refs() instead. Reviewed-by:
-
Filipe Manana authored
At btrfs_block_rsv_refill(), there's no point in initializing the 'num_bytes' variable to 0 and then, after taking the block reserve's spinlock, initializing it to the value of the 'min_reserved' parameter. So just get rid of the 'num_bytes' local variable and rename the 'min_reserved' parameter to 'num_bytes'. Reviewed-by:
-
Filipe Manana authored
At btrfs_truncate_inode_items(), in the while loop when we decide that we are going to delete an item, it's pointless to check that 'pending_del_nr' is non-zero in an else clause because the corresponding if statement is checking if 'pending_del_nr' has a value of zero. So just remove that condition from the else clause. Reviewed-by:
-
Filipe Manana authored
When reserving metadata space for delalloc (and direct IO too), at btrfs_delalloc_reserve_metadata(), there's no need to count the number of extents while holding the inode's spinlock, since that does not require access to any field of the inode. This section of code can be called concurrently, when we have direct IO writes against different file ranges that don't increase the inode's i_size, so it's beneficial to shorten the critical section by counting the number of extents before taking the inode's spinlock. Reviewed-by:
-
Filipe Manana authored
The only caller of btrfs_make_block_group() always passes 0 as the value for the bytes_used argument, so remove it. Reviewed-by:
-
Filipe Manana authored
The function should_end_transaction() is very short and only has one caller, which is btrfs_should_end_transaction(). So move the code from should_end_transaction() into btrfs_should_end_transaction(). Reviewed-by:
-
Filipe Manana authored
Currently btrfs_should_throttle_delayed_refs() returns 1 or 2 in case the delayed refs should be throttled, however the only caller (inode eviction and truncation path) does not care about those two different conditions, it treats the return value as a boolean. This allows us to remove one of the conditions in btrfs_should_throttle_delayed_refs() and change its return value from 'int' to 'bool'. So just do that. Reviewed-by:
-
Filipe Manana authored
At space-info.c:__reserve_bytes(), instead of initializing 'ret' to 0 when it's declared and then shortly after set it to -ENOSPC under the space info's spinlock, initialize it to -ENOSPC when declaring it. Reviewed-by:
-
Filipe Manana authored
When reserving space, at space-info.c:__reserve_bytes(), we assert that either the current task is not holding a transacion handle, or, if it is, that the flush method is not BTRFS_RESERVE_FLUSH_ALL. This is because that flush method can trigger transaction commits, and therefore could lead to a deadlock. However there are other 2 flush methods that can trigger transaction commits: 1) BTRFS_RESERVE_FLUSH_ALL_STEAL 2) BTRFS_RESERVE_FLUSH_EVICT So update the assertion to check the flush method is also not one those two methods if the current task is holding a transaction handle. Reviewed-by:
-
Filipe Manana authored
The BTRFS_RESERVE_FLUSH_EVICT flush method can also commit transactions, see the definition of the evict_flush_states const array at space-info.c, but the documentation for it at space-info.h does not mention it. So update the documentation. Reviewed-by:
-
Filipe Manana authored
The block reserve passed to btrfs_block_rsv_check() is never NULL, so remove the check. In case it can ever become NULL in the future, then we'll get a pretty obvious and clear NULL pointer dereference crash and stack trace. Reviewed-by:
-
Filipe Manana authored
At btrfs_delayed_refs_rsv_refill(), we are passing a value of 0 to the 'update_size' argument of btrfs_block_rsv_add_bytes(), which is defined as a boolean. Functionally this is fine because a 0 is, implicitly, converted to a boolean false value. However it's easier to read an explicit 'false' value, so just pass 'false' instead of 0. Reviewed-by:
-
Filipe Manana authored
The last argument of btrfs_block_rsv_migrate() is a boolean, but we are passing an integer, with a value of 1, to it at evict_refill_and_join(). While this is not a bug, due to type conversion, it's a lot more clear to simply pass the boolean true value instead. So just do that. Reviewed-by:
-
Filipe Manana authored
It's not used anywhere at the moment, but it was used in earlier version of a patch that removed its use in the second version. So just remove btrfs_lru_cache_is_full(). Reviewed-by:
Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by:
-
Christoph Hellwig authored
btrfs_add_compressed_bio_pages is needlessly complicated. Instead of iterating over the logic disk offset just to add pages to the bio use a simple offset starting at 0, which also removes most of the claiming. Additionally __bio_add_pages already takes care of the assert that the bio is always properly sized, and btrfs_submit_bio called right after asserts that the bio size is non-zero. Signed-off-by:
Christoph Hellwig <hch@lst.de> Reviewed-by:
-
Christoph Hellwig authored
Adding pages to a bio has nothing to do with the sector. Move the assignment to the two callers in preparation for cleaning up btrfs_add_compressed_bio_pages. Signed-off-by:
-
Naohiro Aota authored
Currently, /sys/fs/btrfs/<UUID>/bg_reclaim_threshold is limited to 0 (disable) or [50 .. 100]%, so we need to fill 50% of a device to start the auto reclaim process. It is cumbersome to do so when we want to shake out possible race issues of normal write vs reclaim. Relax the threshold check under the BTRFS_DEBUG option. Reviewed-by:
Naohiro Aota <naohiro.aota@wdc.com> Reviewed-by:
-
Christoph Hellwig authored
btrfs_split_bio expects a btrfs_bio as argument and always allocates one. Type both the orig_bio argument and the return value as struct btrfs_bio to improve type safety. Reviewed-by:
Anand Jain <anand.jain@oracle.com> Reviewed-by:
Qu Wenruo <wqu@suse.com> Signed-off-by:
-
Christoph Hellwig authored
Return the containing struct btrfs_bio instead of the less type safe struct bio from btrfs_bio_alloc. Reviewed-by:
-
Christoph Hellwig authored
The bio in struct btrfs_bio_ctrl must be a btrfs_bio, so store a pointer to the btrfs_bio for better type checking. Reviewed-by:
-
Christoph Hellwig authored
struct btrfs_bio now has an always valid inode pointer that can be used to find the inode in submit_one_bio, so use that and initialize all variables for which it is possible at declaration time. Reviewed-by:
-
Christoph Hellwig authored
The original bio must be a btrfs_bio, so store a pointer to the btrfs_bio for better type checking. Reviewed-by:
-
Christoph Hellwig authored
btrfs_submit_compressed_read expects the bio passed to it to be embedded into a btrfs_bio structure. Pass the btrfs_bio directly to increase type safety and make the code self-documenting. Reviewed-by:
-
Christoph Hellwig authored
btrfs_submit_bio expects the bio passed to it to be embedded into a btrfs_bio structure. Pass the btrfs_bio directly to increase type safety and make the code self-documenting. Reviewed-by:
-
Christoph Hellwig authored
All algorithms have to fill the remainder of the orig_bio with zeroes, so do it in common code. Reviewed-by:
-
Christoph Hellwig authored
btrfs_encoded_read_regular_fill_pages has a pretty odd control flow. Unwind it so that there is a single loop over the pages array. Reviewed-by:
-
Christoph Hellwig authored
The inode and file_offset members in struct btrfs_encoded_read_private are unused, so remove them. Last used in commit 7959bd44 ("btrfs: remove the start argument to check_data_csum and export") and commit 7609afac ("btrfs: handle checksum validation and repair at the storage layer"). Reviewed-by:
-
David Sterba authored
The DREW lock uses percpu variable to track lock counters and for that it needs to allocate the structure. In btrfs_read_tree_root() or btrfs_init_fs_root() this may add another error case or requires the NOFS scope protection. One way is to preallocate the structure as was suggested in https://lore.kernel.org/linux-btrfs/20221214021125.28289-1-robbieko@synology.com/ We may avoid the allocation altogether if we don't use the percpu variables but an atomic for the writer counter. This should not make any difference, the DREW lock is used for truncate and NOCOW writes along with other IO operations. The percpu counter for writers has been there since the original commit 8257b2dc "Btrfs: introduce btrfs_{start, end}_nocow_write() for each subvolume". The reason could be to avoid hammering the same cacheline from all the readers but then the writers do that anyway. Signed-off-by:
-
Anand Jain authored
If no discard mount option is specified, including the NODISCARD option, we make the async discard the default option then we don't have to call the clear_opt again to clear the NODISCARD flag. Though this makes no difference, that the call is redundant has been pointed out several times so we better remove it. Signed-off-by:
-
Anand Jain authored
BTRFS_FEATURE_INCOMPAT_SUPP is defined twice, once under CONFIG_BTRFS_DEBUG and once without it, resulting in repetitive code. The reason for this is to add experimental features under CONFIG_BTRFS_DEBUG. To avoid repetitive code, add a common list BTRFS_FEATURE_INCOMPAT_SUPP_STABLE, and append experimental features only under CONFIG_BTRFS_DEBUG. Signed-off-by:
-
Qu Wenruo authored
We don't need to pass the roots as arguments, reading them from the rb-tree is cheap. Thus there is really not much need to pre-fetch it and pass it all the way from scrub_stripe(). And we already have more than enough arguments in scrub_simple_mirror() and scrub_simple_stripe(), it's better to remove them and only grab those roots in scrub_simple_mirror(). Signed-off-by:
-
Qu Wenruo authored
The variable @path is no longer passed into any call sites after commit 18d30ab9 ("btrfs: scrub: use scrub_simple_mirror() to handle RAID56 data stripe scrub"), thus we can remove the variable completely. Signed-off-by:
-
