- 22 Jan, 2018 40 commits
-
-
Timofey Titovets authored
We need to call extent_range_clear_dirty_for_io() on compression range to prevent application from changing page content, while pages compressing. extent_range_clear_dirty_for_io() runs on each loop iteration, "(end - start)" can be much (up to 1024 times) bigger then compression range (BTRFS_MAX_UNCOMPRESSED). The start pointer is advanced each time we manage to compress part of the range. The end pointer does not change so we could redirty the remaining parts repeatedly. Fix that behaviour by call extent_range_clear_dirty_for_io() only once, the first time it happens. This is the safest but probably not the best behaviour. Previous iterations of the patch tried to redirty only the range that we were not able to compress. This has been refused by David for safety reasons, the writeout callchain is complex and there could be some path that relies on redirtying the entire unwritten range. Signed-off-by:
Timofey Titovets <nefelim4ag@gmail.com> Reviewed-by:
David Sterba <dsterba@suse.com> [ enhance changelog, the history and safety concerns, add comment ] Signed-off-by:
-
Timofey Titovets authored
Slowest part of heuristic for now is kernel heap sort() It's can take up to 55% of runtime on sorting bucket items. As sorting will always call on most data sets to get correctly byte_core_set_size, the only way to speed up heuristic, is to speed up sort on bucket. Add a general radix_sort function. Radix sort require 2 buffers, one full size of input array and one for store counters (jump addresses). That increase usage per heuristic workspace +1KiB 8KiB + 1KiB -> 8KiB + 2KiB That is LSD Radix, i use 4 bit as a base for calculating, to make counters array acceptable small (16 elements * 8 byte). That Radix sort implementation have several points to adjust, I added him to make radix sort general usable in kernel, like heap sort, if needed. Performance tested in userspace copy of heuristic code, throughput: - average <-> random data: ~3500 MiB/s - heap sort - average <-> random data: ~6000 MiB/s - radix sort Signed-off-by: -
Anand Jain authored
Currently device state is being managed by each individual int variable such as struct btrfs_device::is_tgtdev_for_dev_replace. Instead of that declare btrfs_device::dev_state BTRFS_DEV_STATE_FLUSH_SENT and use the bit operations. Signed-off-by:
Anand Jain <anand.jain@oracle.com> Signed-off-by:
-
Anand Jain authored
Currently device state is being managed by each individual int variable such as struct btrfs_device::is_tgtdev_for_dev_replace. Instead of that declare btrfs_device::dev_state BTRFS_DEV_STATE_MISSING and use the bit operations. Signed-off-by:
-
Anand Jain authored
Currently device state is being managed by each individual int variable such as struct btrfs_device::missing. Instead of that declare btrfs_device::dev_state BTRFS_DEV_STATE_MISSING and use the bit operations. Signed-off-by:
-
Anand Jain authored
Currently device state is being managed by each individual int variable such as struct btrfs_device::in_fs_metadata. Instead of that declare device state BTRFS_DEV_STATE_IN_FS_METADATA and use the bit operations. Signed-off-by:
Nikolay Borisov <nborisov@suse.com> [ whitespace adjustments ] Signed-off-by:
-
Anand Jain authored
Currently device state is being managed by each individual int variable such as struct btrfs_device::writeable. Instead of that declare device state BTRFS_DEV_STATE_WRITEABLE and use the bit operations. Signed-off-by:
-
Anand Jain authored
This patch creates a helper function to get either the rcu device path or missing. Signed-off-by:
-
Anand Jain authored
We can query the bdev directly when needed at btrfs_discard_extent() so drop btrfs_device::can_discard. Signed-off-by:
-
Colin Ian King authored
The function update_share_count is local to the source and does not need to be in global scope, so make it static. Cleans up sparse warning: fs/btrfs/backref.c:219:6: warning: symbol 'update_share_count' was not declared. Should it be static? Signed-off-by:
Colin Ian King <colin.king@canonical.com> Reviewed-by:
-
Nikolay Borisov authored
Commit 9036c102 ("Btrfs: update hole handling v2") added the FLAG_VACANCY to denote holes, however there was already a consistent way of flagging extents which represent hole - ->block_start = EXTENT_MAP_HOLE. And also the only place where this flag is checked is in the fiemap code, but the block_start value is also checked and every other place in the filesystem detects holes by using block_start value's. So remove the extra flag. This survived a full xfstest run. Signed-off-by:
-
Qu Wenruo authored
This function is no longer used outside of extent-tree.c. Make it static. Signed-off-by:
Qu Wenruo <wqu@suse.com> Reviewed-by:
-
David Sterba authored
* ZSTD_inBuffer in_buf * ZSTD_outBuffer out_buf are used in all functions to pass the compression parameters and the local variables consume some space. We can move them to the workspace and reduce the stack consumption: zstd.c:zstd_decompress -24 (136 -> 112) zstd.c:zstd_decompress_bio -24 (144 -> 120) zstd.c:zstd_compress_pages -24 (264 -> 240) Signed-off-by:
Nick Terrell <terrelln@fb.com> Signed-off-by:
-
David Sterba authored
There are now 20 bytes of holes, we can reduce that to 4 by minor changes. Moving 'aborted' to the status and flags is also more logical, similar for num_dirty_bgs. The size goes from 432 to 416. Reviewed-by:
Liu Bo <bo.li.liu@oracle.com> Signed-off-by:
-
David Sterba authored
The u64 is an overkill here, we could not possibly create that many blockgroups in one transaction. Signed-off-by: -
David Sterba authored
Recent updates to the structure left some holes, reorder the types so the packing is tight. The size goes from 112 to 104 on 64bit. Signed-off-by: -
David Sterba authored
The use_count is a reference counter, we can use the refcount_t type, though we don't use the atomicity. This is not a performance critical code and we could catch the underflows. The type is changed from long, but the number of references will fit an int. Signed-off-by: -
David Sterba authored
Last user was removed in a monster commit a22285a6 ("Btrfs: Integrate metadata reservation with start_transaction") in 2010. Signed-off-by:
-
David Sterba authored
The semantics of adding_csums matches bool, 'short' was most likely used to save space in a698d075 ("Btrfs: add a type field for the transaction handle"). Signed-off-by:
-
Edmund Nadolski authored
Due to new_inline logic, the create == 0 is always true at this point in the code, so the create != 0 branch can be removed. Signed-off-by:
Edmund Nadolski <enadolski@suse.com> Reviewed-by:
-
Edmund Nadolski authored
Replace hardcoded numeric argument values for inode_only with the constants defined for that use. Signed-off-by:
-
David Sterba authored
The maximum size of a checksum buffer is known, BTRFS_CSUM_SIZE, and we don't have to allocate it dynamically. This code path is not used at all as we have only the crc32c and use an on-stack buffer already. Signed-off-by: -
Liu Bo authored
Setting plug can merge adjacent IOs before dispatching IOs to the disk driver. Without plug, it'd not be a problem for single disk usecases, but for multiple disks using raid profile, a large IO can be split to several IOs of stripe length, and plug can be helpful to bring them together for each disk so that we can save several disk access. Moreover, fsync issues synchronous writes, so plug can really take effect. Signed-off-by:
-
Anand Jain authored
No functional changes, create btrfs_open_one_device() from __btrfs_open_devices(). This is a preparatory work to add dynamic device scan. Signed-off-by:
-
Anand Jain authored
No functional changes. This helps to move the entire section into a new function. Signed-off-by:
-
Anand Jain authored
This is in preparation to move a section of code in __btrfs_open_devices() into a new function so that it can be reused. As we set seeding if any of the device is having SB flag BTRFS_SUPER_FLAG_SEEDING, so do it in the device list loop itself. No functional changes. Signed-off-by:
-
Geert Uytterhoeven authored
With gcc-4.1.2: fs/btrfs/ref-verify.c: In function ‘btrfs_build_ref_tree’: fs/btrfs/ref-verify.c:1017: warning: ‘root’ is used uninitialized in this function The variable is indeed passed uninitialized, but it is never used by the callee. However, not all versions of gcc are smart enough to notice. Hence remove the unused parameter from walk_up_tree() to silence the compiler warning. Signed-off-by:Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by:
-
David Sterba authored
All callers pass btree_get_extent, which needs to be exported. Signed-off-by: -
David Sterba authored
All callers pass btrfs_get_extent. Signed-off-by: -
David Sterba authored
All callers pass btrfs_get_extent. Signed-off-by: -
David Sterba authored
There's only one caller that passes btrfs_get_extent. Signed-off-by: -
David Sterba authored
All callers pass btrfs_get_extent_fiemap and get_extent_skip_holes itself is used only as a fiemap helper. Signed-off-by: -
David Sterba authored
All callers pass btrfs_get_extent_fiemap and we don't expect anything else in the context of extent_fiemap. Signed-off-by: -
David Sterba authored
Previous patches cleaned up all places where extent_page_data::get_extent was set and it was btrfs_get_extent all the time, so we can simply call that instead. This also reduces size of extent_page_data by 8 bytes which has positive effect on stack consumption on various functions on the write out path. Signed-off-by: -
David Sterba authored
There's only one caller. Signed-off-by: -
David Sterba authored
There's only one caller. Signed-off-by: -
David Sterba authored
There's only one caller. Signed-off-by: -
Qu Wenruo authored
Since tree-checker has verified leaf when reading from disk, we don't need the existing verify_dir_item() or btrfs_is_name_len_valid() checks. Signed-off-by:
-
Qu Wenruo authored
Add checker for dir item, for key types DIR_ITEM, DIR_INDEX and XATTR_ITEM. This checker does comprehensive checks for: 1) dir_item header and its data size Against item boundary and maximum name/xattr length. This part is mostly the same as old verify_dir_item(). 2) dir_type Against maximum file types, and against key type. Since XATTR key should only have FT_XATTR dir item, and normal dir item type should not have XATTR key. The check between key->type and dir_type is newly introduced by this patch. 3) name hash For XATTR and DIR_ITEM key, key->offset is name hash (crc32c). Check the hash of the name against the key to ensure it's correct. The name hash check is only found in btrfs-progs before this patch. Signed-off-by:
Su Yue <suy.fnst@cn.fujitsu.com> Reviewed-by:
-
David Sterba authored
This callback is called directly from VFS, no locks are held at the allocation time. Signed-off-by:
-
