- 19 Jun, 2023 40 commits
-
-
Filipe Manana authored
There's really no need to BUG_ON() if we find a symlink with an extent that is not inline or is compressed. We can just make send fail with an error (-EUCLEAN) and log an informative error message, so just do that instead of BUG_ON(). Reviewed-by:
Qu Wenruo <wqu@suse.com> Signed-off-by:
Filipe Manana <fdmanana@suse.com> Reviewed-by:
David Sterba <dsterba@suse.com> Signed-off-by:
-
Filipe Manana authored
When dropping inode items from a log tree at drop_inode_items(), we this BUG_ON() on the result of btrfs_search_slot() because we don't expect an exact match since having a key with an offset of (u64)-1 is unexpected. That is generally true, but for dir index keys for example, we can get a key with such an offset value, even though it's very unlikely and it would take ages to increase the sequence counter for a dir index up to (u64)-1. We can deal with an exact match, we just have to delete the key at that slot, so there is really no need to BUG_ON(), error out or trigger any warning. So remove the BUG_ON(). Reviewed-by:
Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by:
-
Filipe Manana authored
There's no need to BUG_ON() at split_item() if the leaf does not have enough free space for the new item, we can just return -ENOSPC since the caller can deal with errors from split_item(). Also, as this is a very unlikely condition to happen, because the caller has invoked setup_leaf_for_split() before calling split_item(), surround the condition with a WARN_ON() which makes it easier to notice this unexpected condition and tags the if branch with 'unlikely' as well. I've actually once hit this BUG_ON() with some incorrect code changes I had, which was very inconvenient as it required rebooting the VM. Reviewed-by:
-
Filipe Manana authored
At btrfs_del_ptr(), instead of doing a BUG_ON() in case we fail to record tree mod log operations, do a transaction abort and return the error to the callers. There's really no need for the BUG_ON() as we can release all resources in the context of all callers, and we have to abort because other future tree searches that use the tree mod log (btrfs_search_old_slot()) may get inconsistent results if other operations modify the tree after that failure and before the tree mod log based search. This implies btrfs_del_ptr() return an int instead of void, and making all callers check for returned errors. Signed-off-by:
-
Filipe Manana authored
At insert_ptr(), instead of doing a BUG_ON() in case we fail to record tree mod log operations, do a transaction abort and return the error to the callers. There's really no need for the BUG_ON() as we can release all resources in the context of all callers, and we have to abort because other future tree searches that use the tree mod log (btrfs_search_old_slot()) may get inconsistent results if other operations modify the tree after that failure and before the tree mod log based search. This implies making insert_ptr() return an int instead of void, and making all callers check for returned errors. Signed-off-by:
-
Filipe Manana authored
At insert_new_root(), instead of doing a BUG_ON() in case we fail to record the tree mod log operation, just return the error to the callers after releasing the allocated tree block. At this point we haven't made any changes to the b+tree, so we haven't left it in an inconsistent state and therefore have no need to abort the transaction. All we need to do is to unlock and free the extent buffer we just allocated with the purpose of making it the new root. Reviewed-by:
-
Filipe Manana authored
At push_nodes_for_insert(), instead of doing a BUG_ON() in case we fail to record tree mod log operations, do a transaction abort and return the error to the caller. There's really no need for the BUG_ON() as we can release all resources in this context, and we have to abort because other future tree searches that use the tree mod log (btrfs_search_old_slot()) may get inconsistent results if other operations modify the tree after that failure and before the tree mod log based search. Reviewed-by:
-
Filipe Manana authored
At update_ref_for_cow() we are calling btrfs_handle_fs_error() if we find that the extent buffer has an unexpected ref count of zero, however we can simply use btrfs_abort_transaction(), which achieves the same purposes: to turn the fs to error state, abort the current transaction and turn the fs to RO mode as well. Besides that, btrfs_abort_transaction() also prints a stack trace which makes it more useful. Also, as this is a very unexpected situation, indicating a serious corruption/inconsistency, tag the if branch as 'unlikely', set the error code to -EUCLEAN instead of -EROFS, and log an explicit message. Reviewed-by:
-
Filipe Manana authored
At balance_level() we are calling btrfs_handle_fs_error() when the middle child only has 1 item and the left child is missing, however we can simply use btrfs_abort_transaction(), which achieves the same purposes: to turn the fs to error state, abort the current transaction and turn the fs to RO mode. Besides that, btrfs_abort_transaction() also prints a stack trace which makes it more useful. Also, as this is a highly unexpected case and it's about a b+tree inconsistency, change the error code from -EROFS to -EUCLEAN, tag the if branch as 'unlikely' and log an explicit error message. Reviewed-by:
-
Filipe Manana authored
At balance_level(), when trying to promote a child node to a root node, if we fail to read the child we call btrfs_handle_fs_error(), which turns the fs to RO mode and sets it to error state as well, causing any ongoing transaction to abort. This however is not necessary because at that point we have not made any change yet at balance_level(), so any error reading the child node does not leaves us in any inconsistent state. Therefore we can just return the error to the caller. Reviewed-by:
-
Filipe Manana authored
At balance_level() we have this 'enospc' label where we jump to in case we get an error at several places. However that error is certainly not -ENOSPC in call cases, it can be -EIO or -ENOMEM when reading a child extent buffer for example, or -ENOMEM when trying to record tree mod log operations. So to make this less confusing, rename the label to 'out'. Reviewed-by:
Anand Jain <anand.jain@oracle.com> Signed-off-by:
-
Filipe Manana authored
At balance_level(), instead of doing a BUG_ON() in case we fail to record tree mod log operations, do a transaction abort and return the error to the callers. There's really no need for the BUG_ON() as we can release all resources in this context, and we have to abort because other future tree searches that use the tree mod log (btrfs_search_old_slot()) may get inconsistent results if other operations modify the tree after that failure and before the tree mod log based search. CC: stable@vger.kernel.org # 5.4+ Reviewed-by:
-
Filipe Manana authored
At __btrfs_cow_block(), instead of doing a BUG_ON() in case we fail to record a tree mod log root insertion operation, do a transaction abort instead. There's really no need for the BUG_ON(), we can properly release all resources in this context and turn the filesystem to RO mode and in an error state instead. CC: stable@vger.kernel.org # 5.4+ Reviewed-by:
-
Filipe Manana authored
When logging tree mod log operations we start by checking, in a lockless manner, if we need to log - if we don't, we just return and do nothing, otherwise we will allocate one or more tree mod log operations and then check again if we need to log. This second check will take the tree mod log lock in write mode if we need to log, otherwise it will do nothing and we just free the allocated memory and return success. We can improve on this by not returning an error in case the memory allocations fail, unless the second check tells us that we actually need to log. That is, if we fail to allocate memory and the second check tells use that we don't need to log, we can just return success and avoid returning -ENOMEM to the caller. Currently tree mod log failures are dealt with either a BUG_ON() or a transaction abort, as tree mod log operations are logged in code paths that modify a b+tree. So just avoid failing with -ENOMEM if we fail to allocate a tree mod log operation unless we actually need to log the operations, that is, if tree_mod_dont_log() returns true. Signed-off-by:
-
Filipe Manana authored
At split_node(), if we fail to log the tree mod log copy operation, we return without unlocking the split extent buffer we just allocated and without decrementing the reference we own on it. Fix this by unlocking it and decrementing the ref count before returning. Fixes: 5de865ee ("Btrfs: fix tree mod logging") CC: stable@vger.kernel.org # 5.4+ Reviewed-by:
-
Filipe Manana authored
When COWing an extent buffer that is not the root node, we need to log in the tree mod log that we replaced a pointer in the parent node, otherwise a tree mod log user doing a search on the b+tree can return incorrect results (that miss something). We are doing the call to btrfs_tree_mod_log_insert_key() but we totally ignore its return value. So fix this by adding the missing error handling, resulting in a transaction abort and freeing the COWed extent buffer. Fixes: f230475e ("Btrfs: put all block modifications into the tree mod log") CC: stable@vger.kernel.org # 5.4+ Reviewed-by:
-
Christoph Hellwig authored
Since commit a2ad63da ("VFS: add FMODE_CAN_ODIRECT file flag") file systems can just set the FMODE_CAN_ODIRECT flag at open time instead of wiring up a dummy direct_IO method to indicate support for direct I/O. Do that for btrfs so that noop_direct_IO can eventually be removed. Signed-off-by:
Christoph Hellwig <hch@lst.de> Reviewed-by:
-
Filipe Manana authored
Currently we are only documenting two uses of the bg_list member of a block group, but there two more: 1) To track deleted block groups for discard purposes, introduced in commit e33e17ee ("btrfs: add missing discards when unpinning extents with -o discard"); 2) To track block groups for automatic reclaim, introduced more recently by commit 18bb8bbf ("btrfs: zoned: automatically reclaim zones") So document those two other use cases. Reviewed-by:
-
Naohiro Aota authored
The reclaim process can temporarily fail. For example, if the space is getting tight, it fails to make the block group read-only. If there are no further writes on that block group, the block group will never get back to the reclaim list, and the BG never gets reclaimed. In a certain workload, we can leave many such block groups never reclaimed. So, let's get it back to the list and give it a chance to be reclaimed. Fixes: 18bb8bbf ("btrfs: zoned: automatically reclaim zones") CC: stable@vger.kernel.org # 5.15+ Signed-off-by:
Naohiro Aota <naohiro.aota@wdc.com> Signed-off-by:
-
Naohiro Aota authored
When a filesystem is read-only, we cannot reclaim a block group as it cannot rewrite the data. Just bail out in that case. Note that it can drop block groups in this case. As we did sb_start_write(), read-only filesystem means we got a fatal error and forced read-only. There is no chance to reclaim them again. Fixes: 18bb8bbf ("btrfs: zoned: automatically reclaim zones") CC: stable@vger.kernel.org # 5.15+ Reviewed-by:
-
Naohiro Aota authored
An unused block group is easy to remove to free up space and should be reclaimed fast. Such block group can often already be a target of the reclaim process. As we check list_empty(&bg->bg_list), we keep it in the reclaim list. That block group is never reclaimed until the file system is filled e.g. up to 75%. Instead, we can move unused block group to the unused list and delete it fast. Fixes: 18bb8bbf ("btrfs: zoned: automatically reclaim zones") CC: stable@vger.kernel.org # 5.15+ Reviewed-by:
-
Naohiro Aota authored
The reclaiming process only starts after the filesystem volumes are allocated to a certain level (75% by default). Thus, the list of reclaiming target block groups can build up so huge at the time the reclaim process kicks in. On a test run, there were over 1000 BGs in the reclaim list. As the reclaim involves rewriting the data, it takes really long time to reclaim the BGs. While the reclaim is running, btrfs_delete_unused_bgs() won't proceed because the reclaim side is holding fs_info->reclaim_bgs_lock. As a result, we will have a large number of unused BGs kept in the unused list. On my test run, I got 1057 unused BGs. Since deleting a block group is relatively easy and fast work, we can call btrfs_delete_unused_bgs() while it reclaims BGs, to avoid building up unused BGs. Fixes: 18bb8bbf ("btrfs: zoned: automatically reclaim zones") CC: stable@vger.kernel.org # 5.15+ Reviewed-by:
-
Christoph Hellwig authored
Use the btrfs_finish_ordered_extent helper to complete compressed writes using the bbio->ordered pointer instead of requiring an rbtree lookup in the otherwise equivalent btrfs_mark_ordered_io_finished called from btrfs_writepage_endio_finish_ordered. Reviewed-by:
Josef Bacik <josef@toxicpanda.com> Signed-off-by:
-
Christoph Hellwig authored
Use the btrfs_finish_ordered_extent helper to complete compressed writes using the bbio->ordered pointer instead of requiring an rbtree lookup in the otherwise equivalent btrfs_mark_ordered_io_finished called from btrfs_writepage_endio_finish_ordered. Reviewed-by:
-
Christoph Hellwig authored
Use the btrfs_finish_ordered_extent helper to complete compressed writes using the bbio->ordered pointer instead of requiring an rbtree lookup in the otherwise equivalent btrfs_mark_ordered_io_finished called from btrfs_writepage_endio_finish_ordered. Reviewed-by:
-
Christoph Hellwig authored
This prepares for switching to more efficient ordered_extent processing and already removes the forth and back conversion from len to end back to len. Reviewed-by:
-
Christoph Hellwig authored
Add a helper to complete an ordered_extent without first doing a lookup. The tracepoint cannot use the ordered_extent class as we also want to print the range. Reviewed-by:
-
Christoph Hellwig authored
Factor out a helper to queue up an ordered_extent completion in a work queue. This new helper will later be used complete an ordered_extent without first doing a lookup. Reviewed-by:
-
Christoph Hellwig authored
Factor out a helper from btrfs_mark_ordered_io_finished that does the actual per-ordered_extent work to check if we want to schedule an I/O completion. This new helper will later be used complete an ordered_extent without first doing a lookup. Reviewed-by:
-
Christoph Hellwig authored
Use the ordered_extent pointer in the btrfs_bio instead of looking it up manually. Reviewed-by:
-
Christoph Hellwig authored
Add a pointer to the ordered_extent to the existing union in struct btrfs_bio, so all code dealing with data write bios can just use a pointer dereference to retrieve the ordered_extent instead of doing multiple rbtree lookups per I/O. The reference to this ordered_extent is dropped at end I/O time, which implies that an extra one must be acquired when the bio is split. This also requires moving the btrfs_extract_ordered_extent call into btrfs_split_bio so that the invariant of always having a valid ordered_extent reference for the btrfs_bio is kept. Reviewed-by:
-
Christoph Hellwig authored
btrfs_dio_submit_io is the only place that uses btrfs_bio_end_io to end a bio that hasn't been submitted using btrfs_submit_bio yet, and this invariant will become a problem with upcoming changes to the btrfs bio layer. Just open code the assignment of bi_status and the call to btrfs_dio_end_io. Reviewed-by:
-
Christoph Hellwig authored
Add a helper to check for that a btrfs_bio has a valid inode, and that it is a data inode to key off all the special handling for data path checksumming. Note that this uses is_data_inode instead of REQ_META as REQ_META is only set directly before submission in submit_one_bio and we'll also want to use this helper for error handling where REQ_META isn't set yet. Reviewed-by:
-
Christoph Hellwig authored
All callers are gone now. Reviewed-by:
-
Christoph Hellwig authored
btrfs_submit_compressed_write always operates on a single ordered_extent. Make that explicit by using btrfs_alloc_ordered_extent in the callers and passing the ordered_extent to btrfs_submit_compressed_write. This will help with storing and ordered_extent pointer in the btrfs_bio in subsequent patches. Reviewed-by:
-
Christoph Hellwig authored
Both callers of btrfs_reloc_clone_csums allocate the ordered_extent that btrfs_reloc_clone_csums operates on. Switch them to use btrfs_alloc_ordered_extent instead of btrfs_add_ordered_extent and pass the ordered_extent to btrfs_reloc_clone_csums instead of doing an extra lookup. Reviewed-by:
-
Christoph Hellwig authored
Refactor run_delalloc_nocow a little bit so that there is only a single call to btrfs_add_ordered_extent instead of two. Reviewed-by:
-
Christoph Hellwig authored
Currently buffered writeback bios are allowed to span multiple ordered_extents, although that basically never actually happens since commit 4a445b7b ("btrfs: don't merge pages into bio if their page offset is not contiguous"). Supporting bios than span ordered_extents complicates the file checksumming code, and prevents us from adding an ordered_extent pointer to the btrfs_bio structure. Use the existing code to limit a bio to single ordered_extent for zoned device writes for all writes. This allows to remove the REQ_BTRFS_ONE_ORDERED flags, and the handling of multiple ordered_extents in btrfs_csum_one_bio. Reviewed-by:
-
Christoph Hellwig authored
If a bio gets split, it needs to have a proper file_offset for checksum validation and repair to work properly. Based on feedback from Josef, commit 852eee62 ("btrfs: allow btrfs_submit_bio to split bios") skipped this adjustment for ONE_ORDERED bios. But if we actually ever need to split a ONE_ORDERED read bio, this will lead to a wrong file offset in the repair code. Right now the only user of the file_offset is logging of an error message so this is mostly harmless, but the wrong offset might be more problematic for additional users in the future. Fixes: 852eee62 ("btrfs: allow btrfs_submit_bio to split bios") Reviewed-by:
-
David Sterba authored
The block group tree was not present among the lockdep classes. We could get potentially lockdep warnings but so far none has been seen, also because block-group-tree is a relatively new feature. CC: stable@vger.kernel.org # 6.1+ Signed-off-by:
-
