- 20 Feb, 2024 40 commits
-
-
Mike Snitzer authored
Addresses various sparse warnings like: warning: symbol 'SYMBOL' was not declared. Should it be static? Signed-off-by:
Mike Snitzer <snitzer@kernel.org> Signed-off-by:
Susan LeGendre-McGhee <slegendr@redhat.com> Signed-off-by:
Matthew Sakai <msakai@redhat.com>
-
Mike Snitzer authored
Signed-off-by:
Ken Raeburn <raeburn@redhat.com> Signed-off-by:
-
Mike Snitzer authored
Signed-off-by:
-
Mike Snitzer authored
Add braces around multi-line while loops and if statements. Also remove excess newlines. Signed-off-by:
Chung Chung <cchung@redhat.com> Signed-off-by:
-
Mike Snitzer authored
Remove extra blank line, mark function inline, add missing braces, and fix a typo in a comment. Signed-off-by:
-
Mike Snitzer authored
Comment typo, whitespace issues, mark function inline. Signed-off-by:
-
Mike Snitzer authored
Eliminate use of "trim" in favor of "discard" since it reflects the top-level Linux discard primative rather than the ATA specific ditto. Signed-off-by:
-
Mike Snitzer authored
Signed-off-by:
-
Mike Snitzer authored
Signed-off-by:
-
Mike Snitzer authored
No need to increment each UDS_ error code manually (relative to UDS_ERROR_CODE_BASE). Also, remove unused PRP_BLOCK_START and PRP_BLOCK_END. Lastly, UDS_SUCCESS and VDO_SUCCESS are used interchangeably; so best to explicitly set VDO_SUCCESS equal to UDS_SUCCESS. Signed-off-by:
-
Mike Snitzer authored
'completion' is more informative name for a 'struct vdo_completion' than 'parent'. Signed-off-by:
-
Mike Snitzer authored
The vdo_page_cache's 'vdo' is the same as the block_map's vdo instance, so use that to save 2 extra dereferences. Signed-off-by:
-
Mike Snitzer authored
The block_map is passed to initialize_block_map_zone, but the block_map's vdo member is already initialized with the same vdo instance, so just use it. Signed-off-by:
-
Mike Snitzer authored
Signed-off-by:
-
Mike Snitzer authored
Rename 'pages' to 'num_pages' in distribute_page_over_waitq(). Update assert message in validate_completed_page() to model others. Tweak line-wrapping on a comment that was needlessly long. Signed-off-by:
-
Mike Snitzer authored
Requires moving dm-vdo-target.c into drivers/md/dm-vdo/ This change adds a proper drivers/md/dm-vdo/Makefile and eliminates the abnormal use of patsubst in drivers/md/Makefile -- which was the cause of at least one build failure that was reported by the upstream build bot. Also, split out VDO's drivers/md/dm-vdo/Kconfig and include it from drivers/md/Kconfig Signed-off-by:
-
Mike Snitzer authored
dm_kcopyd_client_create() returns an ERR_PTR so its return must be checked with IS_ERR(). Signed-off-by:
-
Bruce Johnston authored
Reviewed-by:
Bruce Johnston <bjohnsto@redhat.com> Signed-off-by:
-
Bruce Johnston authored
Reviewed-by:
-
Bruce Johnston authored
Use get_unaligned_le64() on the hash lock's record name to serve as the key to use with the int hash-map. Switching to using int hash-map removes the only consumer of pointer hash-map, as such it is removed. Reviewed-by:
-
Mike Snitzer authored
Rename vdo_waitq_dequeue_next_waiter to vdo_waitq_dequeue_waiter. The "next" aspect of returned waiter is implied. "next" also isn't informative ("oldest" would be). Removing "next_" adds symmetry to vdo_waitq_enqueue_waiter(). Also fix whitespace and comments from previous waitq commit. Reviewed-by: -
Mike Snitzer authored
Rather than incrementally dequeue from the zone->flush_waiters vdo_wait_queue, simply re-initialize it. Reviewed-by:
-
Mike Snitzer authored
Remove temporary 'matched_waiters' waitq and just enqueue matched waiters directly to the caller provided 'matched_waitq'. Reviewed-by:
-
Mike Snitzer authored
Reviewed-by:
-
Mike Snitzer authored
Rename various interfaces and structs associated with vdo's wait-queue, e.g.: s/wait_queue/vdo_wait_queue/, s/waiter/vdo_waiter/, etc. Now all function names start with "vdo_waitq_" or "vdo_waiter_". Reviewed-by:
-
Mike Snitzer authored
Rename process_vio_io() to vdo_submit_vio(), and process_data_vio_io() to submit_data_vio(). Reviewed-by:
-
Mike Snitzer authored
Rename submit_data_vio_io() to vdo_submit_data_vio(). Reviewed-by:
-
Mike Snitzer authored
Rename submit_flush_vio() to vdo_submit_flush_vio(). Reviewed-by:
-
Mike Snitzer authored
Rename submit_metadata_vio() to vdo_submit_metadata_vio(). Reviewed-by:
-
Mike Snitzer authored
Just open-code access to bio's sector. Reviewed-by:
-
Matthew Sakai authored
Signed-off-by:
-
Matthew Sakai authored
dm-vdo targets are not supported for 32-bit configurations. A vdo target typically requires 1 to 1.5 GB of memory at any given time, which is likely a large fraction of the addressable memory of a 32-bit system. At the same time, the amount of addressable storage attached to a 32-bit system may not be large enough for deduplication to provide much benefit. Because of these concerns, 32-bit platforms are deemed unlikely to benefit from using a vdo target, so dm-vdo is targeted only at 64-bit platforms. Co-developed-by:
J. corwin Coburn <corwin@hurlbutnet.net> Signed-off-by:
John Wiele <jwiele@redhat.com> Signed-off-by:
-
Matthew Sakai authored
This adds the dm-vdo target. The dm-vdo target provides inline deduplication, compression, and zero-block elimination, allowing applications to consume less actual storage than a normal target. By layering it with other device mapper targets, it can add these features to any storage stack. It can also provide a common deduplication pool for groups of targets. The vdo target does not protect against data corruption, relying instead on integrity protection of the storage below it. Co-developed-by:
Michael Sclafani <dm-devel@lists.linux.dev> Signed-off-by:
Sweet Tea Dorminy <sweettea-kernel@dorminy.me> Signed-off-by:
-
Matthew Sakai authored
Add support for dumping detailed vdo state to the kernel log via a dmsetup message. The dump code is not thread-safe and is generally intended for use only when the vdo is hung. Co-developed-by:
-
Matthew Sakai authored
Add data and methods setting run time parameters via sysfs, and to make state and statistics information available through sysfs. Co-developed-by:
-
Matthew Sakai authored
Add data and methods to report statisics. Co-developed-by:
-
Matthew Sakai authored
Add data and methods for marshalling and unmarshalling the persistent metadata. Co-developed-by:
-
Matthew Sakai authored
Add the data and methods that manage the dm-vdo target itself. This includes the overall state of the target and its threads, the state of the logical volumes, startup, shutdown, and statistics. Co-developed-by:
-
Matthew Sakai authored
When a vdo is restarted after a crash, it will automatically attempt to recover from its journals. If a vdo encounters an unrecoverable error, it will enter read-only mode. This mode indicates that some previously acknowledged data may have been lost. The vdo may be instructed to rebuild as best it can in order to return to a writable state. Although some data may be lost, this process will ensure that the vdo's own metadata is self-consistent. Co-developed-by:
-
Matthew Sakai authored
The recovery journal is used to amortize updates across the block map and slab depot. Each write request causes an entry to be made in the journal. Entries are either "data remappings" or "block map remappings." For a data remapping, the journal records the logical address affected and its old and new physical mappings. For a block map remapping, the journal records the block map page number and the physical block allocated for it (block map pages are never reclaimed, so the old mapping is always 0). Each journal entry and the data write it represents must be stable on disk before the other metadata structures may be updated to reflect the operation. Co-developed-by:
-
