- 28 Sep, 2011 2 commits
-
-
Philipp Reisner authored
Now drbd communication with protocol 100 actually works. Replaced the remaining p_header80 with p_header where we no longer know which header it is. In the places where p_header80 is still in use, it is on purpose, because we know that it is an old style header there. Signed-off-by:
Philipp Reisner <philipp.reisner@linbit.com> Signed-off-by:
Lars Ellenberg <lars.ellenberg@linbit.com>
-
Philipp Reisner authored
The new header layout will only be used if the peer supports it of course. For the first packet and the handshake packet the old (h80) layout is used for compatibility reasons. Signed-off-by:
-
- 29 Aug, 2011 23 commits
-
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
recv_bm_rle_bits() should not make any assumptions abou the layout of the packet header Signed-off-by:
-
Philipp Reisner authored
Signed-off-by:
-
Philipp Reisner authored
sed -i \ -e 's/mdev->cram_hmac_tfm/mdev->tconn->cram_hmac_tfm/g' \ -e 's/mdev->integrity_w_tfm/mdev->tconn->integrity_w_tfm/g' \ -e 's/mdev->integrity_r_tfm/mdev->tconn->integrity_r_tfm/g' \ -e 's/mdev->int_dig_out/mdev->tconn->int_dig_out/g' \ -e 's/mdev->int_dig_in/mdev->tconn->int_dig_in/g' \ -e 's/mdev->int_dig_vv/mdev->tconn->int_dig_vv/g' \ *.[ch] Signed-off-by: -
Philipp Reisner authored
sed -i \ -e 's/mdev->req_lock/mdev->tconn->req_lock/g' \ -e 's/mdev->unused_spare_tle/mdev->tconn->unused_spare_tle/g' \ -e 's/mdev->newest_tle/mdev->tconn->newest_tle/g' \ -e 's/mdev->oldest_tle/mdev->tconn->oldest_tle/g' \ -e 's/mdev->out_of_sequence_requests/mdev->tconn->out_of_sequence_requests/g' \ *.[ch] Signed-off-by: -
Philipp Reisner authored
sed -i \ -e 's/mdev->agreed_pro_version/mdev->tconn->agreed_pro_version/g' \ -e 's/mdev->last_received/mdev->tconn->last_received/g' \ -e 's/mdev->ko_count/mdev->tconn->ko_count/g' \ *.[ch] Signed-off-by: -
Philipp Reisner authored
Patch mostly: sed -i -e 's/mdev->receiver/mdev->tconn->receiver/g' \ -e 's/mdev->worker/mdev->tconn->worker/g' \ -e 's/mdev->asender/mdev->tconn->asender/g' \ *.[ch] Signed-off-by: -
Philipp Reisner authored
Patch mostly: sed -i -e 's/mdev->data/mdev->tconn->data/g' \ -e 's/mdev->meta/mdev->tconn->meta/g' \ *.[ch] Signed-off-by: -
Philipp Reisner authored
Patch partly generated by: sed -i -e 's/get_net_conf(mdev)/get_net_conf(mdev->tconn)/g' \ -e 's/put_net_conf(mdev)/put_net_conf(mdev->tconn)/g' \ -e 's/get_net_conf(odev)/get_net_conf(odev->tconn)/g' \ -e 's/put_net_conf(odev)/put_net_conf(odev->tconn)/g' \ *.[ch] Signed-off-by: -
Philipp Reisner authored
Besides moving the struct member, everything else is generated by: sed -i -e 's/mdev->net_conf/mdev->tconn->net_conf/g' \ -e 's/odev->net_conf/odev->tconn->net_conf/g' \ *.[ch] Signed-off-by: -
Philipp Reisner authored
Starting to dissolve the network connection from the actual block devices. Signed-off-by:
-
Philipp Reisner authored
Inspired by the list_for_each_entry() macro
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
The open_cnt check is an open-coded D_ASSERT() check. In case the data.work queue is not empty, it does not really help to know which drbd_work elements remained on that list: they will be freed immediately afterwards, anyway. Signed-off-by:
-
Andreas Gruenbacher authored
Remove the file name and line number from the syslog messages generated: we have no duplicate function names, and no function contains the same assertion more than once. 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:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
We do not do collision detection for read requests, but we still need to look up the request objects when we receive a package over the network. Using the same data structure for read and write requests results in simpler code once the tl_hash and app_reads_hash tables are removed. Signed-off-by:
-
- 25 Aug, 2011 15 commits
-
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Factor out duplicate code in got_NegAck(). Signed-off-by:
-
Andreas Gruenbacher authored
Get rid of the ar_id_to_req() and ack_id_to_req() wrappers. Signed-off-by:
-
Andreas Gruenbacher authored
Unify the ar_id_to_req() and ack_id_to_req() functions: make both fail if the consistency check fails. Move the request lookup code now duplicated in both functions into its own function. Signed-off-by:
-
Andreas Gruenbacher authored
Move _ar_id_to_req() to drbd_receiver.c and mark it non-inline. Remove the leading underscores from _ar_id_to_req() and _ack_id_to_req(). Mark ar_hash_slot() inline. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
This is the only place where this function is used. Make it static. Signed-off-by:
-
Andreas Gruenbacher authored
The ID_VACANT definition has become entirely irrelevant by now. The is_syncer_block_id() macro does not improve the code, so eliminated it. Signed-off-by:
-
Andreas Gruenbacher authored
Converting the constants happens at compile time. Signed-off-by:
-
Andreas Gruenbacher authored
Converting the constants happens at compile time. Signed-off-by:
-
Andreas Gruenbacher authored
DRBD_MAGIC has nothing to do with block ids and the funny values computed were not actually used, anyway. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
