The offset field in struct btrfs_extent_ref records the position
inside file that file extent is referenced by. In the new back
reference system, tree leaves holding reference to file extent
are recorded explicitly. We can quickly scan these tree leaves, so the
offset field is not required.
This patch also makes the back reference system check the objectid
when extents are being deleted
Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
This patch makes the back reference system to explicit record the
location of parent node for all types of extents. The location of
parent node is placed into the offset field of backref key. Every
time a tree block is balanced, the back references for the affected
lower level extents are updated.
This patch improves converter's allocator and fixes a bug in data relocation
function. The new allocator caches free blocks as Btrfs's default allocator.
In testing here, the user CPU time reduced to half of the original when
checksum and small file packing was disabled. This patch also enlarges the
size of block groups created by the converter.
balance level starts by trying to empty the middle block, and then
pushes from the right to the middle. This might empty the right block
and leave a small number of pointers in the middle.
The main changes in this patch are adding chunk handing and data relocation
ability. In the last step of conversion, the converter relocates data in system
chunk and move chunk tree into system chunk. In the rollback process, the
converter remove chunk tree from system chunk and copy data back.
Regards
YZ
---
Block headers now store the chunk tree uuid
Chunk items records the device uuid for each stripes
Device extent items record better back refs to the chunk tree
Block groups record better back refs to the chunk tree
The chunk tree format has also changed. The objectid of BTRFS_CHUNK_ITEM_KEY
used to be the logical offset of the chunk. Now it is a chunk tree id,
with the logical offset being stored in the offset field of the key.
This allows a single chunk tree to record multiple logical address spaces,
upping the number of bytes indexed by a chunk tree from 2^64 to
2^128.
When a block is freed, it can be immediately reused if it is from
the current transaction. But, an extra check is required to make sure
the block had not been written yet. If it were reused after being written,
the transid in the block header might match the transid of the
next time the block was allocated.
The parent node records the transaction ID of the block it is pointing to,
and this is used as part of validating the block on reads. So, there
can only be one version of a block per transaction.
The mkfs code bootstraps the filesystem on a single device. Once
the raid block groups are setup, it needs to recow all of the blocks so
that each tree is properly allocated.
We get lots of warnings of the flavor:
utils.c:441: warning: format '%Lu' expects type 'long long unsigned int' but argument 2 has type 'u64'
And thanks to -Werror, the build fails. Clean up these printfs
by properly casting the arg to the format specified.
Signed-off-by: Alex Chiang <achiang@hp.com>
This is done by doing a two-step conversion (rather than a one-step).
First, the variable goes from type * to void *, and then to
implicitly to void **.
(Not sure if this is "good practice", but it shuts up the compiler,
so it seems the compiler takes into account that we are actually punning
it this way.)
This adds support for keeping track of the number of blocks used by
root_item's. This makes it so that mkfs lays down the "default" subvol with
the correct block accounting in place. Thank you,
