summaryrefslogtreecommitdiffstats
path: root/xlators/mgmt/glusterd/src/glusterd-syncop.c
diff options
context:
space:
mode:
authorKrishnan Parthasarathi <kparthas@redhat.com>2013-04-02 07:56:25 +0530
committerAnand Avati <avati@redhat.com>2013-04-12 13:47:46 -0700
commitf34343d3751cd73e8eabe6d5544fb1f58b316595 (patch)
tree869aa908771b3708f4ad9a7a7ec57a623239b9a5 /xlators/mgmt/glusterd/src/glusterd-syncop.c
parent732cd267c924554a638519cff0df146b2688d6e8 (diff)
glusterd: big lock - a coarse-grained locking to prevent races
There are primarily three lists that are part of glusterd process, that are concurrently accessed. Namely, priv->volumes, priv->peers and volinfo->bricks_list. Big-lock approach ----------------- WHAT IS IT? Big lock is a coarse-grained lock which protects all three lists, mentioned above, from racy access. HOW DOES IT WORK? At any given point in time, glusterd's thread(s) are in execution _iff_ there is a preceding, inbound network event. Of course, the sigwaiter thread and timer thread are exceptions. A network event is an external trigger to glusterd, via the epoll thread, in the form of POLLIN and POLLERR. As long as we take the big-lock at all such entry points and yield it when we are done, we are guaranteed that all the network events, accessing the global lists, are serialised. This amounts to holding the big lock at - all the handlers of all the actors in glusterd. (POLLIN) - all the cbks in glusterd. (POLLIN) - rpc_notify (DISCONNECT event), if we access/modify one of the three lists. (POLLERR) In the case of synctask'ized volume operations, we must remember that, if we held the big lock for the entire duration of the handler, we may block other non-synctask rpc actors from executing. For eg, volume-start would block in PMAP SIGNIN, if done incorrectly. To prevent this, we need to yield the big lock, when we yield the synctask, and reacquire on waking up of the synctask. Change-Id: Ib929f9905b55fb6c3fc27fefb497a26dba058e4f BUG: 948686 Signed-off-by: Krishnan Parthasarathi <kparthas@redhat.com> Reviewed-on: http://review.gluster.org/4784 Reviewed-by: Jeff Darcy <jdarcy@redhat.com> Tested-by: Gluster Build System <jenkins@build.gluster.com>
Diffstat (limited to 'xlators/mgmt/glusterd/src/glusterd-syncop.c')
-rw-r--r--xlators/mgmt/glusterd/src/glusterd-syncop.c24
1 files changed, 20 insertions, 4 deletions
diff --git a/xlators/mgmt/glusterd/src/glusterd-syncop.c b/xlators/mgmt/glusterd/src/glusterd-syncop.c
index 5bdf5917d..d36459632 100644
--- a/xlators/mgmt/glusterd/src/glusterd-syncop.c
+++ b/xlators/mgmt/glusterd/src/glusterd-syncop.c
@@ -18,6 +18,16 @@
#include "glusterd-op-sm.h"
#include "glusterd-utils.h"
+static inline void
+gd_synctask_barrier_wait (struct syncargs *args, int count)
+{
+ glusterd_conf_t *conf = THIS->private;
+
+ synclock_unlock (&conf->big_lock);
+ synctask_barrier_wait (args, count);
+ synclock_lock (&conf->big_lock);
+}
+
static void
gd_collate_errors (struct syncargs *args, int op_ret, int op_errno,
char *op_errstr)
@@ -664,7 +674,7 @@ gd_lock_op_phase (struct list_head *peers, glusterd_op_t op, dict_t *op_ctx,
gd_syncop_mgmt_lock (peerinfo->rpc, &args, MY_UUID, peer_uuid);
peer_cnt++;
}
- synctask_barrier_wait((&args), peer_cnt);
+ gd_synctask_barrier_wait((&args), peer_cnt);
ret = args.op_ret;
if (ret) {
gf_asprintf (op_errstr, "Another transaction could be "
@@ -739,7 +749,7 @@ stage_done:
op, req_dict, op_ctx);
peer_cnt++;
}
- synctask_barrier_wait((&args), peer_cnt);
+ gd_synctask_barrier_wait((&args), peer_cnt);
ret = args.op_ret;
if (dict_get_str (op_ctx, "errstr", &errstr) == 0)
*op_errstr = gf_strdup (errstr);
@@ -814,7 +824,7 @@ commit_done:
op, req_dict, op_ctx);
peer_cnt++;
}
- synctask_barrier_wait((&args), peer_cnt);
+ gd_synctask_barrier_wait((&args), peer_cnt);
ret = args.op_ret;
if (dict_get_str (op_ctx, "errstr", &errstr) == 0)
*op_errstr = gf_strdup (errstr);
@@ -851,7 +861,7 @@ gd_unlock_op_phase (struct list_head *peers, glusterd_op_t op, int op_ret,
list_del_init (&peerinfo->op_peers_list);
peer_cnt++;
}
- synctask_barrier_wait((&args), peer_cnt);
+ gd_synctask_barrier_wait((&args), peer_cnt);
ret = args.op_ret;
if (ret) {
gf_log (this->name, GF_LOG_ERROR, "Failed to unlock "
@@ -887,8 +897,10 @@ gd_brick_op_phase (glusterd_op_t op, dict_t *op_ctx, dict_t *req_dict, char **op
int ret = -1;
rpc_clnt_t *rpc = NULL;
dict_t *rsp_dict = NULL;
+ glusterd_conf_t *conf = NULL;
this = THIS;
+ conf = this->private;
rsp_dict = dict_new ();
if (!rsp_dict) {
ret = -1;
@@ -928,8 +940,12 @@ gd_brick_op_phase (glusterd_op_t op, dict_t *op_ctx, dict_t *req_dict, char **op
"due to rpc failure.");
goto out;
}
+ /*This is to ensure that the brick_op_cbk is able to take
+ * the big lock*/
+ synclock_unlock (&conf->big_lock);
ret = gd_syncop_mgmt_brick_op (rpc, pending_node, op, req_dict,
op_ctx, op_errstr);
+ synclock_lock (&conf->big_lock);
if (ret)
goto out;