Commit Graph

438 Commits

Author SHA1 Message Date
Chris Mason
6324fbf334 Btrfs: Dynamic chunk and block group allocation
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:01 -04:00
Chris Mason
0b86a832a1 Btrfs: Add support for multiple devices per filesystem
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
7f93bf8d27 Match the extent tree code to btrfs-progs for multi-device merging
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
eb20978f31 Btrfs: Use KM_USERN instead of KM_IRQ during data summing
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
55d4741489 Btrfs: Disable sysfs files on older kernels
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
2e1a992e31 Btrfs: Make sure bio pages are adjacent during bulk csumming
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
6885f308b5 Btrfs: Misc 2.6.25 updates
Remove the btrfs read_inode method, and use save_mount_options

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Greg KH
e3fe4e7120 btrfs: fixes for kobject changes in mainline
Here's a patch against the unstable tree that gets the code to build
against Linus's current tree (2.6.24-git12).  This is needed as the
kobject/kset api has changed there.

I tried to make the smallest changes needed, and it builds and loads
successfully, but I don't have a btrfs volume anywhere (yet) to try to
see if things still work properly :)

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
6e92f5e651 Btrfs: While doing checksums on bios, cache the extent_buffer mapping
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
065631f6dc Btrfs: checksum file data at bio submission time instead of during writepage
When we checkum file data during writepage, the checksumming is done one
page at a time, making it difficult to do bulk metadata modifications
to insert checksums for large ranges of the file at once.

This patch changes btrfs to checksum on a per-bio basis instead.  The
bios are checksummed before they are handed off to the block layer, so
each bio is contiguous and only has pages from the same inode.

Checksumming on a bio basis allows us to insert and modify the file
checksum items in large groups.  It also allows the checksumming to
be done more easily by async worker threads.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
0740c82bc7 Btrfs: Properly cast before shifting
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
d99cb30a11 Btrfs: Take the extent lock before dropping the delalloc bits
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Yan Zheng
5e591a0703 Btrfs: Fix looping on readdir of the subvol roots
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
0762704b19 Btrfs: Properly clear dirty and delalloc extent bits while preparing the file for write
Yan Zheng noticed that we don't clear the extent state tree dirty and delalloc
bits when we clear the dirty bits on the page during file write.

This leads to csum errors later on.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
952fccac50 Btrfs: Remove extent back refs in batches, and avoid duplicate searches
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
d7fc640e6f Btrfs: Allocator improvements
Reduce CPU time searching for free blocks by optimizing find_first_extent_bit

Fix find_free_extent to make better use of the last_alloc hint.  Before it
was often finding blocks just before the hint.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
9afbb0b752 Btrfs: Disable tree defrag in SSD mode
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
David Miller
df68b8a7ad Btrfs: unaligned access fixes
Btrfs set/get macros lose type information needed to avoid
unaligned accesses on sparc64.
ere is a patch for the kernel bits which fixes most of the
unaligned accesses on sparc64.

btrfs_name_hash is modified to return the hash value instead
of getting a return location via a (potentially unaligned)
pointer.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Yan
39b5637f6f Btrfs: Fix "no csum found for inode" issue.
A few codes were not properly updated for changes of extent map.  This
may be the causes of "no csum found for inode" issue.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
9069218d44 Btrfs: Fix i_blocks accounting
Now that delayed allocation accounting works, i_blocks accounting is changed
to only modify i_blocks when extents inserted or removed.

The fillattr call is changed to include the delayed allocation byte count
in the i_blocks result.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
a86c12c73d Btrfs: Create larger bios for btree blocks
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
961d023213 Btrfs: Don't case unsigned long to int in bio submission
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
5d196fc15d Btrfs: Use 2MB as the empty_size for clustered allocations
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Yan
7261009cea btrfs_drop_extents: handle BTRFS_INODE_REF_KEY types
It's possible "key.type == BTRFS_INODE_REF_KEY" and "key.offset >= end".

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
068fe39fa1 Btrfs: Add checks for last byte in disk to allocator grouping
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Yan
dacb473ad8 Btrfs: Fix hole creation in file_write
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
47b0c4f8c7 Btrfs: Update magic
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
f594706643 Btrfs: Add debugging for block group update failure
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Yan
c2e639f02c Btrfs: Fix typo in extent_io.c
---

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
80b6794d11 Btrfs: Lower stack usage in transaction.c
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
ae9d12853b Btrfs: Fix delalloc account on state deletion
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Yan
2f375ab9c5 Call btrfs_cow_block while lowering tree level.
When freeing root block of a tree,  btrfs_free_extent' parameter
'ref_generation' is from root block itseft.  When freeing non-root
block,  'ref_generation' is from its parent. so when converting a
non-root block to root block, we must guarantee its generation is
equal to its parent's generation.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
60cde612c8 Btrfs: Use last_alloc optimizations for metadata, even without -o ssd
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
21a4989d26 Btrfs: Hash in the offset and owner for file extent backref keys
This makes searches for backrefs and backref insertion much more efficient
when there are many backrefs for a single extent

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
47e4bb988c Btrfs: Insert extent record and the first backref in a single balance
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
80ea96b1f3 Btrfs: Add a lookup cache to the extent state tree
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
4529ba495c Btrfs: Add data block hints to SSD mode too
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
b0c68f8bed Btrfs: Enable delalloc accounting
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Yan
0181e58f91 btrfs_drop_extent fixe for inline items > 8K
When truncating a inline extent, btrfs_drop_extents doesn't properly
handle the case "key.offset > inline_limit". This bug can only happen
when max line size is larger than 8K.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
1b0f7c29e2 Fix hole start calculation in btrfs_settar
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
f392a938f3 Properly align the hole size in btrfs_setattr
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Yan
b1632b10c0 Btrfs: Align extent length to sectorsize in
---

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
5a01a2e3a9 Btrfs: Copy correct tree when inserting into slot 0
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
b3236e68bf Btrfs: Leave on the tree defragger in mount -o ssd, it still helps there
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
6f568d35a0 Btrfs: mount -o max_inline=size to control the maximum inline extent size
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
291d673e6a Btrfs: Do delalloc accounting via hooks in the extent_state code
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
9c58309d6c Btrfs: Add inode item and backref in one insert, reducing cpu usage
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
85e21bac16 Btrfs: During deletes and truncate, remove many items at once from the tree
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:04:00 -04:00
Chris Mason
70dec8079d Btrfs: extent_io and extent_state optimizations
The end_bio routines are changed to take a pointer to the extent state
struct, and the state tree is walked in order to set/clear appropriate
bits as IO completes.  This greatly reduces the number of rbtree searches
done by the end_bio handlers, and reduces lock contention.

The extent_io releasepage function is changed to avoid expensive searches
for locked state.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:03:59 -04:00
Chris Mason
aadfeb6e39 Btrfs: Add some extra debugging around file data checksum failures
Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-09-25 11:03:59 -04:00