- 23 Mar, 2020 40 commits
-
-
Qu Wenruo authored
This function is only used in read_fs_root(), which is just a wrapper of btrfs_get_fs_root(). For all the mentioned essential roots except log root tree, btrfs_get_fs_root() has its own quick path to grab them from fs_info directly, thus no need for key.offset modification. For subvolume trees, btrfs_get_fs_root() with key.offset == -1 is completely fine. For log trees and log root tree, it's impossible to hit them, as for relocation all backrefs are fetched from commit root, which never records log tree blocks. Log tree blocks either get freed in regular transaction commit, or replayed at mount time. At runtime we should never hit an backref for log tree in extent tree. Reviewed-by:
Josef Bacik <josef@toxicpanda.com> Signed-off-by:
Qu Wenruo <wqu@suse.com> Reviewed-by:
David Sterba <dsterba@suse.com> Signed-off-by:
-
Nikolay Borisov authored
This commit flips the switch to start tracking/processing pinned extents on a per-transaction basis. It mostly replaces all references from btrfs_fs_info::(pinned_extents|freed_extents[]) to btrfs_transaction::pinned_extents. Two notable modifications that warrant explicit mention are changing clean_pinned_extents to get a reference to the previously running transaction. The other one is removal of call to btrfs_destroy_pinned_extent since transactions are going to be cleaned in btrfs_cleanup_one_transaction. Reviewed-by:
Nikolay Borisov <nborisov@suse.com> Reviewed-by:
-
Nikolay Borisov authored
Next patch is going to refactor how pinned extents are tracked which will necessitate changing this code. To ease that work and contain the changes factor the code now in preparation, this will also help review. Reviewed-by:
-
Nikolay Borisov authored
In preparation to making pinned extents per-transaction ensure that log such extents are always excluded from caching. To achieve this in addition to marking them via btrfs_pin_extent_for_log_replay they also need to be marked with btrfs_add_excluded_extent to prevent log tree extent buffer being loaded by the free space caching thread. That's required since log tree blocks are not recorded in the extent tree, hence they always look free. Reviewed-by:
-
Nikolay Borisov authored
Preparation for refactoring pinned extents tracking. Reviewed-by:
-
Nikolay Borisov authored
All callers have a reference to a transaction handle so pass it to pin_down_extent. This is the final step before switching pinned extent tracking to a per-transaction basis. Reviewed-by:
-
Nikolay Borisov authored
Preparation for refactoring pinned extents tracking. Reviewed-by:
-
Nikolay Borisov authored
btrfs_pin_reserved_extent is now only called with a valid transaction so exploit the fact to take a transaction. This is preparation for tracking pinned extents on a per-transaction basis. Reviewed-by:
-
Nikolay Borisov authored
Calling btrfs_pin_reserved_extent makes sense only with a valid transaction since pinned extents are processed from transaction commit in btrfs_finish_extent_commit. In case of error it's sufficient to adjust the reserved counter to account for log tree extents allocated in the last transaction. This commit moves btrfs_pin_reserved_extent to be called only with valid transaction handle and otherwise uses the newly introduced unaccount_log_buffer to adjust "reserved". If this is not done if a failure occurs before transaction is committed WARN_ON are going to be triggered on unmount. This was especially pronounced with generic/475 test. Reviewed-by:
-
Nikolay Borisov authored
This function correctly adjusts the reserved bytes occupied by a log tree extent buffer. It will be used instead of calling btrfs_pin_reserved_extent. Reviewed-by:
-
Nikolay Borisov authored
Preparation for switching pinned extent tracking to a per-transaction basis. Reviewed-by:
-
Nikolay Borisov authored
Having btrfs_destroy_delayed_refs call btrfs_pin_extent is problematic for making pinned extents tracking per-transaction since btrfs_trans_handle cannot be passed to btrfs_pin_extent in this context. Additionally delayed refs heads pinned in btrfs_destroy_delayed_refs are going to be handled very closely, in btrfs_destroy_pinned_extent. To enable btrfs_pin_extent to take btrfs_trans_handle simply open code it in btrfs_destroy_delayed_refs and call btrfs_error_unpin_extent_range on the range. This enables us to do less work in btrfs_destroy_pinned_extent and leaves btrfs_pin_extent being called in contexts which have a valid btrfs_trans_handle. Reviewed-by:
-
Anand Jain authored
The devinfo attribute handlers were added in 668e48af ("btrfs: sysfs, add devid/dev_state kobject and device attributes") and the name should contain _devinfo_, there's one that does not conform, so unify it with the rest. Signed-off-by:
Anand Jain <anand.jain@oracle.com> Reviewed-by:
-
Anand Jain authored
Since commit 668e48af ("btrfs: sysfs, add devid/dev_state kobject and device attributes"), the functions btrfs_sysfs_add_device_link() and btrfs_sysfs_rm_device_link() do more than just adding and removing the device link as its name indicated. Rename them to be more specific that's about the directory with the attirbutes Signed-off-by:
-
Anand Jain authored
We have one simple function btrfs_sysfs_remove_fsid() to undo btrfs_sysfs_add_fsid(), which also does proper checks before releasing objects. One difference, if btrfs_sysfs_remove_fsid is used that now we also call kobject_del() which was missing before. This was tested (with kobject debug turned on) and no change in behaviour was found. This is a cleanup patch. Signed-off-by:
-
David Sterba authored
The tree pointer can be safely read from the inode, use it and drop the redundant argument. Reviewed-by:
Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by:
-
David Sterba authored
The tree pointer can be safely read from the inode, use it and drop the redundant argument. Reviewed-by:
-
David Sterba authored
The tree pointer can be safely read from the inode, use it and drop the redundant argument. Reviewed-by:
-
David Sterba authored
The tree pointer can be safely read from the page's inode, use it and drop the redundant argument. Reviewed-by:
-
David Sterba authored
The tree pointer can be safely read from the inode so we can drop the redundant argument from btrfs_lock_and_flush_ordered_range. Reviewed-by:
-
David Sterba authored
Add assertions to all helpers that get tree as argument and verify that it's the same that can be obtained from the inode or from its pages. In followup patches the redundant arguments and assertions will be removed one by one. Reviewed-by:
-
David Sterba authored
Now that we're sure the tree from argument is same as the one we can get from the page's inode io_tree, drop the redundant argument. Reviewed-by:
-
David Sterba authored
All functions that set up extent_page_data::tree set it to the inode io_tree. That's passed down the callstack that accesses either the same inode or its pages. In the end submit_extent_page can pull the tree out of the page and we don't have to store it in the structure. Reviewed-by:
-
David Sterba authored
The status of aborted transaction can change between calls and it needs to be accessed by READ_ONCE. Add a helper that also wraps the unlikely hint. Reviewed-by:
-
David Sterba authored
The helpers are related to locking so move them there, update comments. Reviewed-by:
-
Josef Bacik authored
We are now using these for all roots, rename them to btrfs_put_root() and btrfs_grab_root(); Reviewed-by:
-
Josef Bacik authored
Now that we're going to start relying on getting ref counting right for roots, add a list to track allocated roots and print out any roots that aren't freed up at free_fs_info time. Hide this behind CONFIG_BTRFS_DEBUG because this will just be used for developers to verify they aren't breaking things. Reviewed-by:
-
Josef Bacik authored
In adding things like eb leak checking and root leak checking there were a lot of weird corner cases that come from the fact that 1) We do not init the fs_info until we get to open_ctree time in the normal case and 2) The test infrastructure half-init's the fs_info for things that it needs. This makes it really annoying to make changes because you have to add init in two different places, have special cases for testing fs_info's that may not have certain things initialized, and cases for fs_info's that didn't make it to open_ctree and thus are not fully set up. Fix this by extracting out the non-allocating init of the fs info into it's own public function and use that to make sure we're all getting consistent views of an allocated fs_info. Signed-off-by: -
Josef Bacik authored
open_ctree mixes initialization of fs stuff and fs_info stuff, which makes it confusing when doing things like adding the root leak detection. Make a separate function that inits all the static structures inside of the fs_info needed for the fs to operate, and then call that before we start setting up the fs_info to be mounted. Reviewed-by:
-
Josef Bacik authored
Things like the percpu_counters, the mapping_tree, and the csum hash can all be freed at btrfs_free_fs_info time, since the helpers all check if the structure has been initialized already. This significantly cleans up the error cases in open_ctree. Signed-off-by:
-
Josef Bacik authored
Now that all callers of btrfs_get_fs_root are subsequently calling btrfs_grab_fs_root and handling dropping the ref when they are done appropriately, go ahead and push btrfs_grab_fs_root up into btrfs_get_fs_root. Signed-off-by:
-
Josef Bacik authored
If we are going to track leaked roots we need to free them all the same way, so don't kfree() roots directly, use btrfs_put_fs_root. Reviewed-by:
-
Josef Bacik authored
We lookup the fs_root and put it in our fs_info directly, we should hold a ref on this root for the lifetime of the fs_info. Signed-off-by:
-
Josef Bacik authored
We're going to start freeing roots and doing other complicated things in free_fs_info, so we need to move it to disk-io.c and export it in order to use things lik btrfs_put_fs_root(). Reviewed-by:
-
Josef Bacik authored
We lookup the uuid of arbitrary subvolumes, hold a ref on the root while we're doing this. Signed-off-by:
-
Josef Bacik authored
We replay the log into arbitrary fs roots, hold a ref on the root while we're doing this. Signed-off-by:
-
Josef Bacik authored
We create the snapshot and then use it for a bunch of things, we need to hold a ref on it while we're messing with it. Signed-off-by:
-
Josef Bacik authored
We lookup the name of a subvol which means we'll cross into different roots. Hold a ref while we're doing the look ups in the fs_root we're searching. Signed-off-by:
-
Josef Bacik authored
We lookup all the clone roots and the parent root for send, so we need to hold refs on all of these roots while we're processing them. Signed-off-by:
-
Josef Bacik authored
We look up the root for the bytenr that is failing, so we need to hold a ref on the root for that operation. Signed-off-by:
-
