- 08 Nov, 2012 40 commits
-
-
Andreas Gruenbacher authored
The 8 byte header finally becomes too small. With the protocol 100 header we have 16 bit for the volume number, proper 32 bit for the data length, and 32 bit for further extensions in the future. Previous versions of drbd are using version 80 headers for all packets short enough for protocol 80. They support both header versions in worker context, but only version 80 headers in asynchronous context. For backwards compatibility, continue to use version 80 headers for short packets before protocol version 100. From protocol version 100 on, use the same header version for all packets. Signed-off-by:
Philipp Reisner <philipp.reisner@linbit.com> Signed-off-by:
Lars Ellenberg <lars.ellenberg@linbit.com>
-
Andreas Gruenbacher authored
Prepare the introduction of the protocol 100 headers. The actual protocol header is removed for the packet declarations. I.e. allow us to use the packets with different headers. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Centralize sock->mutex locking and unlocking in [drbd|conn]_prepare_command() and [drbd|conn]_send_comman(). Therefore all *_send_* functions are touched to use these primitives instead of drbd_get_data_sock()/drbd_put_data_sock() and former helper functions. That change makes the *_send_* functions more standardized. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher 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
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
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Lars Ellenberg authored
Recent commit drbd: get rid of bio_split, allow bios of "arbitrary" size had a reference count leak: it only deactivated the first of several activity log extents for intervals crossing extent boundaries. This commit generalizes on bios spanning multiple activity log extents in drbd_al_begin_io, and adds the necessary loop around lc_put in drbd_al_complete_io as well. Signed-off-by:
-
Lars Ellenberg authored
Where "arbitrary" size is currently 1 MiB, which is the BIO_MAX_SIZE for architectures with 4k PAGE_CACHE_SIZE (most). Signed-off-by:
-
Lars Ellenberg authored
Signed-off-by:
-
Lars Ellenberg authored
We want to avoid bio_split for bios crossing activity log boundaries. So we may need to activate two activity log extents "atomically". drbd_al_begin_io() needs to know more than just the start sector. Signed-off-by:
-
Lars Ellenberg authored
So we can initialize a clean on disk activity log area, without the module complaining with loud assert messages because of checksum or magic value mismatches. Signed-off-by:
-
Andreas Gruenbacher authored
Packets of type P_HAND_SHAKE define which protocol versions and features a node supports. For clarity, call those packets P_CONNECTION_FEATURES instead. (This does not determine the features that a specific drbd device supports, such as drbd protocol A, B, C.) Signed-off-by:
-
Andreas Gruenbacher authored
The first packets exchanged when a connection is established are referred to as P_HAND_SHAKE_S and P_HAND_SHAKE_M in the code, followed by P_HAND_SHAKE packets. To avoid confusion between these two unrelated things, call the initial packets P_INITIAL_DATA and P_INITIAL_META. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher 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
During a disconnect the oc variable in _conn_request_state() could become outdated. Determin the common old state after sleeping. While at it, I implemented that for all parts of the state Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Andreas Gruenbacher authored
The receive handlers do not all handle unknown volume numbers the same way. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
These messages can only trigger in case there is a pretty obvious internal programming error. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
