cluster/afr: Make local->child_up immutable

Afr transaction performs lock, pre-op, op, post-op and unlock steps in that
order. The child_up[] is overloaded with the information of where all
the first two steps succeeded. This works perfectly fine for
Transaction, but the locking/unlocking part of the code is re-used by
data self-heal. In that each loop_frame does lock, rchecksum,
read-from-source and write-to-sinks, unlock steps.

Rchecksum fop assumes that the fop needs to happen on one source + all
sinks and sets the call_count to that number. But if the lock step fails
on any of the sinks it will mark the child_up of that child to 0, which
will result in call_count mismatch and the frame will hang thinking that
some more cbks need to come. When this happens loop_frame will never go
to unlock step leading to hangs on that file.

Change-Id: I3dd0449cc6193a980bacf637d935881f4b22210a
BUG: 3597
Reviewed-on: http://review.gluster.com/474
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amar@gluster.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>

1 files changed, 24 insertions, 40 deletions

diff --git a/xlators/cluster/afr/src/afr-dir-write.c b/xlators/cluster/afr/src/afr-dir-write.c
index 9a17c20309b..2929ad741d8 100644
--- a/xlators/cluster/afr/src/afr-dir-write.c
+++ b/xlators/cluster/afr/src/afr-dir-write.c
@@ -212,16 +212,14 @@ afr_create_wind (call_frame_t *frame, xlator_t *this)
 {
         afr_local_t *local = NULL;
         afr_private_t *priv = NULL;
-        afr_internal_lock_t *int_lock = NULL;
         int call_count = -1;
         int i = 0;
 
         local = frame->local;
         priv = this->private;
-        int_lock = &local->internal_lock;
 
-        call_count = afr_locked_children_count (int_lock->entry_locked_nodes,
-                                                priv->child_count);
+        call_count = afr_pre_op_done_children_count (local->transaction.pre_op,
+                                                     priv->child_count);
 
         if (call_count == 0) {
                 local->transaction.resume (frame, this);
@@ -231,7 +229,7 @@ afr_create_wind (call_frame_t *frame, xlator_t *this)
         local->call_count = call_count;
 
         for (i = 0; i < priv->child_count; i++) {
-                if (local->child_up[i] && int_lock->entry_locked_nodes[i]) {
+                if (local->transaction.pre_op[i]) {
                         STACK_WIND_COOKIE (frame, afr_create_wind_cbk,
                                            (void *) (long) i,
                                            priv->children[i],
@@ -445,16 +443,14 @@ afr_mknod_wind (call_frame_t *frame, xlator_t *this)
 {
         afr_local_t *local = NULL;
         afr_private_t *priv = NULL;
-        afr_internal_lock_t *int_lock = NULL;
         int call_count = -1;
         int i = 0;
 
         local = frame->local;
         priv  = this->private;
-        int_lock = &local->internal_lock;
 
-        call_count = afr_locked_children_count (int_lock->entry_locked_nodes,
-                                                priv->child_count);
+        call_count = afr_pre_op_done_children_count (local->transaction.pre_op,
+                                                     priv->child_count);
 
         if (call_count == 0) {
                 local->transaction.resume (frame, this);
@@ -464,7 +460,7 @@ afr_mknod_wind (call_frame_t *frame, xlator_t *this)
         local->call_count = call_count;
 
         for (i = 0; i < priv->child_count; i++) {
-                if (local->child_up[i] && int_lock->entry_locked_nodes[i]) {
+                if (local->transaction.pre_op[i]) {
                         STACK_WIND_COOKIE (frame, afr_mknod_wind_cbk, (void *) (long) i,
                                            priv->children[i],
                                            priv->children[i]->fops->mknod,
@@ -673,16 +669,14 @@ afr_mkdir_wind (call_frame_t *frame, xlator_t *this)
 {
         afr_local_t *local = NULL;
         afr_private_t *priv = NULL;
-        afr_internal_lock_t *int_lock = NULL;
         int call_count = -1;
         int i = 0;
 
         local = frame->local;
         priv  = this->private;
-        int_lock = &local->internal_lock;
 
-        call_count = afr_locked_children_count (int_lock->entry_locked_nodes,
-                                                priv->child_count);
+        call_count = afr_pre_op_done_children_count (local->transaction.pre_op,
+                                                     priv->child_count);
 
         if (call_count == 0) {
                 local->transaction.resume (frame, this);
@@ -692,7 +686,7 @@ afr_mkdir_wind (call_frame_t *frame, xlator_t *this)
         local->call_count = call_count;
 
         for (i = 0; i < priv->child_count; i++) {
-                if (local->child_up[i] && int_lock->entry_locked_nodes[i]) {
+                if (local->transaction.pre_op[i]) {
                         STACK_WIND_COOKIE (frame, afr_mkdir_wind_cbk,
                                            (void *) (long) i,
                                            priv->children[i],
@@ -903,16 +897,14 @@ afr_link_wind (call_frame_t *frame, xlator_t *this)
 {
         afr_local_t *local = NULL;
         afr_private_t *priv = NULL;
-        afr_internal_lock_t *int_lock = NULL;
         int call_count = -1;
         int i = 0;
 
         local = frame->local;
         priv  = this->private;
-        int_lock = &local->internal_lock;
 
-        call_count = afr_locked_children_count (int_lock->entry_locked_nodes,
-                                                priv->child_count);
+        call_count = afr_pre_op_done_children_count (local->transaction.pre_op,
+                                                     priv->child_count);
 
         if (call_count == 0) {
                 local->transaction.resume (frame, this);
@@ -922,7 +914,7 @@ afr_link_wind (call_frame_t *frame, xlator_t *this)
         local->call_count = call_count;
 
         for (i = 0; i < priv->child_count; i++) {
-                if (local->child_up[i] && int_lock->entry_locked_nodes[i]) {
+                if (local->transaction.pre_op[i]) {
                         STACK_WIND_COOKIE (frame, afr_link_wind_cbk, (void *) (long) i,
                                            priv->children[i],
                                            priv->children[i]->fops->link,
@@ -1129,16 +1121,14 @@ afr_symlink_wind (call_frame_t *frame, xlator_t *this)
 {
         afr_local_t *local = NULL;
         afr_private_t *priv = NULL;
-        afr_internal_lock_t *int_lock = NULL;
         int call_count = -1;
         int i = 0;
 
         local = frame->local;
         priv = this->private;
-        int_lock = &local->internal_lock;
 
-        call_count = afr_locked_children_count (int_lock->entry_locked_nodes,
-                                                priv->child_count);
+        call_count = afr_pre_op_done_children_count (local->transaction.pre_op,
+                                                     priv->child_count);
 
         if (call_count == 0) {
                 local->transaction.resume (frame, this);
@@ -1148,7 +1138,7 @@ afr_symlink_wind (call_frame_t *frame, xlator_t *this)
         local->call_count = call_count;
 
         for (i = 0; i < priv->child_count; i++) {
-                if (local->child_up[i] && int_lock->entry_locked_nodes[i]) {
+                if (local->transaction.pre_op[i]) {
                         STACK_WIND_COOKIE (frame, afr_symlink_wind_cbk,
                                            (void *) (long) i,
                                            priv->children[i],
@@ -1353,16 +1343,14 @@ afr_rename_wind (call_frame_t *frame, xlator_t *this)
 {
         afr_local_t *local = NULL;
         afr_private_t *priv = NULL;
-        afr_internal_lock_t *int_lock = NULL;
         int call_count = -1;
         int i = 0;
 
         local = frame->local;
         priv = this->private;
-        int_lock = &local->internal_lock;
 
-        call_count = afr_locked_children_count (int_lock->entry_locked_nodes,
-                                                priv->child_count);
+        call_count = afr_pre_op_done_children_count (local->transaction.pre_op,
+                                                     priv->child_count);
 
         if (call_count == 0) {
                 local->transaction.resume (frame, this);
@@ -1372,7 +1360,7 @@ afr_rename_wind (call_frame_t *frame, xlator_t *this)
         local->call_count = call_count;
 
         for (i = 0; i < priv->child_count; i++) {
-                if (local->child_up[i] && int_lock->entry_locked_nodes[i]) {
+                if (local->transaction.pre_op[i]) {
                         STACK_WIND_COOKIE (frame, afr_rename_wind_cbk,
                                            (void *) (long) i,
                                            priv->children[i],
@@ -1559,16 +1547,14 @@ afr_unlink_wind (call_frame_t *frame, xlator_t *this)
 {
         afr_local_t *local = NULL;
         afr_private_t *priv = NULL;
-        afr_internal_lock_t *int_lock = NULL;
         int call_count = -1;
         int i = 0;
 
         local = frame->local;
         priv  = this->private;
-        int_lock = &local->internal_lock;
 
-        call_count = afr_locked_children_count (int_lock->entry_locked_nodes,
-                                                priv->child_count);
+        call_count = afr_pre_op_done_children_count (local->transaction.pre_op,
+                                                     priv->child_count);
 
         if (call_count == 0) {
                 local->transaction.resume (frame, this);
@@ -1578,7 +1564,7 @@ afr_unlink_wind (call_frame_t *frame, xlator_t *this)
         local->call_count = call_count;
 
         for (i = 0; i < priv->child_count; i++) {
-                if (local->child_up[i] && int_lock->entry_locked_nodes[i]) {
+                if (local->transaction.pre_op[i]) {
                         STACK_WIND_COOKIE (frame, afr_unlink_wind_cbk,
                                            (void *) (long) i,
                                            priv->children[i],
@@ -1758,16 +1744,14 @@ afr_rmdir_wind (call_frame_t *frame, xlator_t *this)
 {
         afr_local_t *local = NULL;
         afr_private_t *priv = NULL;
-        afr_internal_lock_t *int_lock = NULL;
         int call_count = -1;
         int i = 0;
 
         local = frame->local;
         priv  = this->private;
-        int_lock = &local->internal_lock;
 
-        call_count = afr_locked_children_count (int_lock->entry_locked_nodes,
-                                                priv->child_count);
+        call_count = afr_pre_op_done_children_count (local->transaction.pre_op,
+                                                     priv->child_count);
 
         if (call_count == 0) {
                 local->transaction.resume (frame, this);
@@ -1777,7 +1761,7 @@ afr_rmdir_wind (call_frame_t *frame, xlator_t *this)
         local->call_count = call_count;
 
         for (i = 0; i < priv->child_count; i++) {
-                if (local->child_up[i] && int_lock->entry_locked_nodes[i]) {
+                if (local->transaction.pre_op[i]) {
                         STACK_WIND_COOKIE (frame, afr_rmdir_wind_cbk,
                                            (void *) (long) i,
                                            priv->children[i],