drbd: New disk option al-updates

By disabling al-updates one might increase performace. The price for
that is that in case a crashed primary (that had al-updates disabled)
is reintegraded, it will receive a full-resync instead of a bitmap
based resync.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>

5 files changed, 30 insertions, 4 deletions

diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c
index 9eae28944312..83d48d210b69 100644
--- a/drivers/block/drbd/drbd_actlog.c
+++ b/drivers/block/drbd/drbd_actlog.c
@@ -276,8 +276,16 @@ void drbd_al_begin_io(struct drbd_conf *mdev, struct drbd_interval *i)
                 /* Double check: it may have been committed by someone else,
                  * while we have been waiting for the lock. */
                 if (mdev->act_log->pending_changes) {
-                        al_write_transaction(mdev);
-                        mdev->al_writ_cnt++;
+                        bool write_al_updates;
+
+                        rcu_read_lock();
+                        write_al_updates = rcu_dereference(mdev->ldev->disk_conf)->al_updates;
+                        rcu_read_unlock();
+
+                        if (write_al_updates) {
+                                al_write_transaction(mdev);
+                                mdev->al_writ_cnt++;
+                        }
 
                         spin_lock_irq(&mdev->al_lock);
                         /* FIXME
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index dc5bd6bbb280..c5d4fac1a111 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -1230,6 +1230,11 @@ int drbd_adm_disk_opts(struct sk_buff *skb, struct genl_info *info)
 
         mutex_unlock(&mdev->tconn->conf_update);
 
+        if (new_disk_conf->al_updates)
+                mdev->ldev->md.flags &= MDF_AL_DISABLED;
+        else
+                mdev->ldev->md.flags |= MDF_AL_DISABLED;
+
         drbd_bump_write_ordering(mdev->tconn, WO_bdev_flush);
 
         drbd_md_sync(mdev);
@@ -1545,7 +1550,9 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
         } else if (dd == grew)
                 set_bit(RESYNC_AFTER_NEG, &mdev->flags);
 
-        if (drbd_md_test_flag(mdev->ldev, MDF_FULL_SYNC)) {
+        if (drbd_md_test_flag(mdev->ldev, MDF_FULL_SYNC) ||
+            (test_bit(CRASHED_PRIMARY, &mdev->flags) &&
+             drbd_md_test_flag(mdev->ldev, MDF_AL_DISABLED))) {
                 dev_info(DEV, "Assuming that all blocks are out of sync "
                      "(aka FullSync)\n");
                 if (drbd_bitmap_io(mdev, &drbd_bmio_set_n_write,
@@ -1588,13 +1595,19 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
         if (ns.disk == D_CONSISTENT &&
             (ns.pdsk == D_OUTDATED || rcu_dereference(mdev->ldev->disk_conf)->fencing == FP_DONT_CARE))
                 ns.disk = D_UP_TO_DATE;
-        rcu_read_unlock();
 
         /* All tests on MDF_PRIMARY_IND, MDF_CONNECTED_IND,
            MDF_CONSISTENT and MDF_WAS_UP_TO_DATE must happen before
            this point, because drbd_request_state() modifies these
            flags. */
 
+        if (rcu_dereference(mdev->ldev->disk_conf)->al_updates)
+                mdev->ldev->md.flags &= MDF_AL_DISABLED;
+        else
+                mdev->ldev->md.flags |= MDF_AL_DISABLED;
+
+        rcu_read_unlock();
+
         /* In case we are C_CONNECTED postpone any decision on the new disk
            state after the negotiation phase. */
         if (mdev->state.conn == C_CONNECTED) {
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index 1e86156c10f7..36ae7dd28d90 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -338,6 +338,7 @@ extern const char *drbd_set_st_err_str(enum drbd_state_rv);
 #define MDF_PEER_OUT_DATED      (1 << 5)
 #define MDF_CRASHED_PRIMARY     (1 << 6)
 #define MDF_AL_CLEAN            (1 << 7)
+#define MDF_AL_DISABLED         (1 << 8)
 
 enum drbd_uuid_index {
         UI_CURRENT,
diff --git a/include/linux/drbd_genl.h b/include/linux/drbd_genl.h
index 826008f297fe..92ec4b50a885 100644
--- a/include/linux/drbd_genl.h
+++ b/include/linux/drbd_genl.h
@@ -130,6 +130,8 @@ GENL_struct(DRBD_NLA_DISK_CONF, 3, disk_conf,
         __flg_field_def(19, DRBD_GENLA_F_MANDATORY,     md_flushes, DRBD_MD_FLUSHES_DEF)
         __u32_field_def(20,     DRBD_GENLA_F_MANDATORY, disk_timeout, DRBD_DISK_TIMEOUT_DEF)
         __u32_field_def(21,     0 /* OPTIONAL */,       read_balancing, DRBD_READ_BALANCING_DEF)
+        /* 9: __u32_field_def(22,       DRBD_GENLA_F_MANDATORY, unplug_watermark, DRBD_UNPLUG_WATERMARK_DEF) */
+        __flg_field_def(23,     0 /* OPTIONAL */,       al_updates, DRBD_AL_UPDATES_DEF)
 )
 
 GENL_struct(DRBD_NLA_RESOURCE_OPTS, 4, res_opts,
@@ -168,6 +170,7 @@ GENL_struct(DRBD_NLA_NET_CONF, 5, net_conf,
         __flg_field_def(27, DRBD_GENLA_F_MANDATORY,     always_asbp, DRBD_ALWAYS_ASBP_DEF)
         __flg_field(28, DRBD_GENLA_F_MANDATORY | DRBD_F_INVARIANT,      tentative)
         __flg_field_def(29,     DRBD_GENLA_F_MANDATORY, use_rle, DRBD_USE_RLE_DEF)
+        /* 9: __u32_field_def(30,       DRBD_GENLA_F_MANDATORY, fencing_policy, DRBD_FENCING_DEF) */
 )
 
 GENL_struct(DRBD_NLA_SET_ROLE_PARMS, 6, set_role_parms,
diff --git a/include/linux/drbd_limits.h b/include/linux/drbd_limits.h
index 17ef66a5c114..1fa19c5f5e64 100644
--- a/include/linux/drbd_limits.h
+++ b/include/linux/drbd_limits.h
@@ -210,6 +210,7 @@
 #define DRBD_DISK_DRAIN_DEF     1
 #define DRBD_MD_FLUSHES_DEF     1
 #define DRBD_TCP_CORK_DEF       1
+#define DRBD_AL_UPDATES_DEF     1
 
 #define DRBD_ALLOW_TWO_PRIMARIES_DEF    0
 #define DRBD_ALWAYS_ASBP_DEF    0