- 08 Nov, 2012 1 commit
-
-
Lars Ellenberg authored
This commit breaks the API again. Move per-volume former syncer options into disk_conf. Move per-connection former syncer options into net_conf. Renamed the remainign sync_conf to res_opts Syncer settings have been changeable at runtime, so we need to prepare for these settings to be runtime-changeable in their new home as well. Introduce new configuration operations, and share the netlink attribute between "attach" (create new disk) and "disk-opts" (change options). Same for "connect" and "net-opts". Some fields cannot be changed at runtime, however. Introduce a new flag GENLA_F_INVARIANT to be able to trigger on that in the generated validation and assignment functions. Signed-off-by:
Philipp Reisner <philipp.reisner@linbit.com> Signed-off-by:
Lars Ellenberg <lars.ellenberg@linbit.com>
-
- 03 Nov, 2012 25 commits
-
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Lars Ellenberg authored
If drbd_adm_attach failed early, it left the CONFIG_PENDING bit on, blocking any further conn_reconfig_start on that connection. Signed-off-by:
-
Philipp Reisner authored
New function got_conn_RqSReply() Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
That is necessary in case a connection does not have a volume 0 Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
In the context of drbd-8.4 it no longer makes sense to dissalow that. Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Took the chance and converted tconn_process_done_ee() to use idr_for_each_entry() Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Lars Ellenberg authored
This greatly simplifies deconfiguration of whole resources. Signed-off-by:
-
Lars Ellenberg authored
somehow a "goto abort" was introduced with commit drbd: Extracted is_valid_transition() out of sanitize_state() which left drbd_req_state still holding the spin lock. Signed-off-by:
-
Lars Ellenberg authored
We have resources resp. connections, volumes, and minor numbers. A config request may specifies all three of them. If it turns out that the minor belongs to a different connection, or a different volume number in the same connection, that configuration request is invalid. Signed-off-by:
-
Lars Ellenberg authored
Follow O_CREAT semantics when creating connection or minor device/volume objects. If we need O_CREAT|O_EXCL semantics some time down the road, we can add NLM_F_EXCL to the netlink message flags. Signed-off-by:
-
Lars Ellenberg authored
Even if the connection is still established. We should be able to reduce a volume from a replication group, without taking the whole group offline. Signed-off-by:
-
Lars Ellenberg authored
Get rid of a temporary variable and, funny bitand assignment. Just short circuit, returning false, once we encounter the first still configured volume. FIXME verify call sites for need of rcu_read_lock or stronger. Signed-off-by:
-
Lars Ellenberg authored
We want to see existing connection objects, even if they do not currently have volumes attached. Change the .dumpit variant of drbd_adm_get_status to iterate not over minor devices, but over connections + volumes. Signed-off-by:
-
Lars Ellenberg authored
Signed-off-by:
-
Lars Ellenberg authored
s/idr_get_new/idr_get_new_above/ Signed-off-by:
-
Lars Ellenberg authored
Signed-off-by:
-
- 14 Oct, 2011 14 commits
-
-
Lars Ellenberg authored
This adds the new API header and helper files. Signed-off-by:
-
Lars Ellenberg authored
This function was used to broadcast the (leading part of the) bio payload in case we see a data integrity error. It could be received from userland with the drbdsetup events subcommand, to have a peek into the payload that caused the checksum mismatch, and guess from there what may have caused the mismatch, mainly to guess wether it was modification of in-flight data, or data corruption by broken hardware or software bugs. Meanwhile we support bios that are larger than the maximum payload a netlink datagram can carry. And we have means to reliably detect modification of in-flight data by calculating, and comparing, the checksum before and after sendmsg. There is no need to carry this around anymore. Signed-off-by:
-
Lars Ellenberg authored
drbd: fix drbd_delete_device: remove vnr from volumes; idr_remove(); synchronize_rcu(); before cleanup Still missing: rcu_readlock() on the various call sites that access/iterate over those idrs. We don't need a specific write lock, as we only modify from configuration context, which is already strictly serialized. Signed-off-by:
-
Lars Ellenberg authored
Don't rely on availability of bios from the global fs_bio_set, we should use our own bio_set for meta data IO. Signed-off-by:
-
Lars Ellenberg authored
Signed-off-by:
-
Lars Ellenberg authored
Signed-off-by:
-
Lars Ellenberg authored
This commit got it wrong: drbd: Make the peer_seq updating code more obvious Make it more clear that update_peer_seq() is supposed to wake up the seq_wait queue whenever the sequence number changes. We don't need to wake up everytime we receive a sequence number that is _different_ from our currently stored "newest" sequence number, but only if we receive a sequence number _newer_ than what we already have, when we actually change mdev->peer_seq. Signed-off-by: -
Lars Ellenberg authored
Old default behaviour was "pass-on", which is not useful in production at all. Signed-off-by:
-
Lars Ellenberg authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Up to now it only operated on minor numbers. Now it can work also on named connections. Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
