[pbs-devel] [PATCH 00/11] Locking and rustdoc improvements

[Stefan Reiter]

Getting our locking into better shape, now with at least 10% more flocks/mutexes
and 14% fewer bugs! *

Order now, and you will even receive an updated documentation providing lots of
detail on how said locking actually works!

* terms and conditions may apply


With these patches applied, all interactions displayed in the table in patch 11
resolve fine in theory, in practice I couldn't trigger any of the known races we
had previously anymore either. Extra thourough review/testing still appreciated,
locking is hard...


proxmox-backup: Stefan Reiter (11):
  prune: respect snapshot flock
  prune: never fail, just warn about failed removals
  backup: use shared flock for base snapshot
  reader: acquire shared flock on open snapshot
  verify: acquire shared snapshot flock and skip on error
  gc: avoid race between phase1 and forget/prune
  datastore: remove load_manifest_json
  datastore: add manifest locking
  datastore: remove individual snapshots before group
  rustdoc: add crate level doc
  rustdoc: overhaul backup rustdoc and add locking table

 src/api2/admin/datastore.rs    |  84 +++++++-------
 src/api2/backup.rs             |   4 +-
 src/api2/backup/environment.rs |   6 +-
 src/api2/reader.rs             |  19 +++-
 src/backup.rs                  | 201 ++++++++++++++++++++-------------
 src/backup/datastore.rs        | 132 ++++++++++++++++++----
 src/backup/dynamic_index.rs    |   5 +-
 src/backup/fixed_index.rs      |   7 +-
 src/backup/verify.rs           |  22 +++-
 src/lib.rs                     |   5 +
 src/tools/fs.rs                |  22 +++-
 11 files changed, 345 insertions(+), 162 deletions(-)

-- 
2.20.1