Whamcloud - gitweb
LU-6142 ldlm: Fix style issues for ldlm_extent.c 36/49536/6
authorArshad Hussain <arshad.hussain@aeoncomputing.com>
Fri, 23 Dec 2022 12:26:27 +0000 (17:56 +0530)
committerOleg Drokin <green@whamcloud.com>
Fri, 3 Feb 2023 06:48:16 +0000 (06:48 +0000)
This patch fixes issues reported by checkpatch
for file lustre/ldlm/ldlm_extent.c

Test-Parameters: trivial
Signed-off-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Change-Id: I9cecd1f377f33f3d4129cddcd7b59c3a7c003e04
Reviewed-on: https://review.whamcloud.com/c/fs/lustre-release/+/49536
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
lustre/ldlm/ldlm_extent.c

index 4209310..544bfaf 100644 (file)
@@ -74,39 +74,41 @@ static void ldlm_extent_internal_policy_fixup(struct ldlm_lock *req,
        __u64 req_end = req->l_req_extent.end;
        __u64 req_align, mask;
 
-        if (conflicting > 32 && (req_mode == LCK_PW || req_mode == LCK_CW)) {
-                if (req_end < req_start + LDLM_MAX_GROWN_EXTENT)
-                        new_ex->end = min(req_start + LDLM_MAX_GROWN_EXTENT,
-                                          new_ex->end);
-        }
-
-        if (new_ex->start == 0 && new_ex->end == OBD_OBJECT_EOF) {
-                EXIT;
-                return;
-        }
-
-        /* we need to ensure that the lock extent is properly aligned to what
-         * the client requested. Also we need to make sure it's also server
-         * page size aligned otherwise a server page can be covered by two
-         * write locks. */
+       if (conflicting > 32 && (req_mode == LCK_PW || req_mode == LCK_CW)) {
+               if (req_end < req_start + LDLM_MAX_GROWN_EXTENT)
+                       new_ex->end = min(req_start + LDLM_MAX_GROWN_EXTENT,
+                                         new_ex->end);
+       }
+
+       if (new_ex->start == 0 && new_ex->end == OBD_OBJECT_EOF) {
+               EXIT;
+               return;
+       }
+
+       /* we need to ensure that the lock extent is properly aligned to what
+        * the client requested. Also we need to make sure it's also server
+        * page size aligned otherwise a server page can be covered by two
+        * write locks.
+        */
        mask = PAGE_SIZE;
-        req_align = (req_end + 1) | req_start;
-        if (req_align != 0 && (req_align & (mask - 1)) == 0) {
-                while ((req_align & mask) == 0)
-                        mask <<= 1;
-        }
-        mask -= 1;
-        /* We can only shrink the lock, not grow it.
-         * This should never cause lock to be smaller than requested,
-         * since requested lock was already aligned on these boundaries. */
-        new_ex->start = ((new_ex->start - 1) | mask) + 1;
-        new_ex->end = ((new_ex->end + 1) & ~mask) - 1;
-        LASSERTF(new_ex->start <= req_start,
+       req_align = (req_end + 1) | req_start;
+       if (req_align != 0 && (req_align & (mask - 1)) == 0) {
+               while ((req_align & mask) == 0)
+                       mask <<= 1;
+       }
+       mask -= 1;
+       /* We can only shrink the lock, not grow it.
+        * This should never cause lock to be smaller than requested,
+        * since requested lock was already aligned on these boundaries.
+        */
+       new_ex->start = ((new_ex->start - 1) | mask) + 1;
+       new_ex->end = ((new_ex->end + 1) & ~mask) - 1;
+       LASSERTF(new_ex->start <= req_start,
                 "mask %#llx grant start %llu req start %llu\n",
-                 mask, new_ex->start, req_start);
-        LASSERTF(new_ex->end >= req_end,
+                mask, new_ex->start, req_start);
+       LASSERTF(new_ex->end >= req_end,
                 "mask %#llx grant end %llu req end %llu\n",
-                 mask, new_ex->end, req_end);
+                mask, new_ex->end, req_end);
 }
 
 /**
@@ -122,7 +124,7 @@ static void ldlm_extent_internal_policy_fixup(struct ldlm_lock *req,
  * Use interval tree to expand the lock extent for granted lock.
  */
 static void ldlm_extent_internal_policy_granted(struct ldlm_lock *req,
-                                                struct ldlm_extent *new_ex)
+                                               struct ldlm_extent *new_ex)
 {
        struct ldlm_resource *res = req->l_resource;
        enum ldlm_mode req_mode = req->l_req_mode;
@@ -135,44 +137,44 @@ static void ldlm_extent_internal_policy_granted(struct ldlm_lock *req,
        };
        int conflicting = 0;
        int idx;
+
        ENTRY;
 
        lockmode_verify(req_mode);
 
        /* Using interval tree to handle the LDLM extent granted locks. */
-        for (idx = 0; idx < LCK_MODE_NUM; idx++) {
+       for (idx = 0; idx < LCK_MODE_NUM; idx++) {
                struct interval_node_extent ext = {
                        .start  = req_start,
                        .end    = req_end,
                };
 
-                tree = &res->lr_itree[idx];
-                if (lockmode_compat(tree->lit_mode, req_mode))
-                        continue;
-
-                conflicting += tree->lit_size;
-                if (conflicting > 4)
-                        limiter.start = req_start;
-
-                if (interval_is_overlapped(tree->lit_root, &ext))
-                        CDEBUG(D_INFO, 
-                               "req_mode = %d, tree->lit_mode = %d, "
-                               "tree->lit_size = %d\n",
-                               req_mode, tree->lit_mode, tree->lit_size);
-                interval_expand(tree->lit_root, &ext, &limiter);
-                limiter.start = max(limiter.start, ext.start);
-                limiter.end = min(limiter.end, ext.end);
-                if (limiter.start == req_start && limiter.end == req_end)
-                        break;
-        }
-
-        new_ex->start = limiter.start;
-        new_ex->end = limiter.end;
-        LASSERT(new_ex->start <= req_start);
-        LASSERT(new_ex->end >= req_end);
-
-        ldlm_extent_internal_policy_fixup(req, new_ex, conflicting);
-        EXIT;
+               tree = &res->lr_itree[idx];
+               if (lockmode_compat(tree->lit_mode, req_mode))
+                       continue;
+
+               conflicting += tree->lit_size;
+               if (conflicting > 4)
+                       limiter.start = req_start;
+
+               if (interval_is_overlapped(tree->lit_root, &ext))
+                       CDEBUG(D_INFO,
+                              "req_mode = %d, tree->lit_mode = %d, tree->lit_size = %d\n",
+                              req_mode, tree->lit_mode, tree->lit_size);
+               interval_expand(tree->lit_root, &ext, &limiter);
+               limiter.start = max(limiter.start, ext.start);
+               limiter.end = min(limiter.end, ext.end);
+               if (limiter.start == req_start && limiter.end == req_end)
+                       break;
+       }
+
+       new_ex->start = limiter.start;
+       new_ex->end = limiter.end;
+       LASSERT(new_ex->start <= req_start);
+       LASSERT(new_ex->end >= req_end);
+
+       ldlm_extent_internal_policy_fixup(req, new_ex, conflicting);
+       EXIT;
 }
 
 /* The purpose of this function is to return:
@@ -182,7 +184,7 @@ static void ldlm_extent_internal_policy_granted(struct ldlm_lock *req,
  */
 static void
 ldlm_extent_internal_policy_waiting(struct ldlm_lock *req,
-                                    struct ldlm_extent *new_ex)
+                                   struct ldlm_extent *new_ex)
 {
        struct ldlm_resource *res = req->l_resource;
        enum ldlm_mode req_mode = req->l_req_mode;
@@ -190,6 +192,7 @@ ldlm_extent_internal_policy_waiting(struct ldlm_lock *req,
        __u64 req_end = req->l_req_extent.end;
        struct ldlm_lock *lock;
        int conflicting = 0;
+
        ENTRY;
 
        lockmode_verify(req_mode);
@@ -204,75 +207,81 @@ ldlm_extent_internal_policy_waiting(struct ldlm_lock *req,
                        return;
                }
 
-                /* Don't conflict with ourselves */
-                if (req == lock)
-                        continue;
-
-                /* Locks are compatible, overlap doesn't matter */
-                /* Until bug 20 is fixed, try to avoid granting overlapping
-                 * locks on one client (they take a long time to cancel) */
-                if (lockmode_compat(lock->l_req_mode, req_mode) &&
-                    lock->l_export != req->l_export)
-                        continue;
-
-                /* If this is a high-traffic lock, don't grow downwards at all
-                 * or grow upwards too much */
-                ++conflicting;
-                if (conflicting > 4)
-                        new_ex->start = req_start;
-
-                /* If lock doesn't overlap new_ex, skip it. */
-                if (!ldlm_extent_overlap(l_extent, new_ex))
-                        continue;
-
-                /* Locks conflicting in requested extents and we can't satisfy
-                 * both locks, so ignore it.  Either we will ping-pong this
-                 * extent (we would regardless of what extent we granted) or
-                 * lock is unused and it shouldn't limit our extent growth. */
-                if (ldlm_extent_overlap(&lock->l_req_extent,&req->l_req_extent))
-                        continue;
-
-                /* We grow extents downwards only as far as they don't overlap
-                 * with already-granted locks, on the assumption that clients
-                 * will be writing beyond the initial requested end and would
-                 * then need to enqueue a new lock beyond previous request.
-                 * l_req_extent->end strictly < req_start, checked above. */
-                if (l_extent->start < req_start && new_ex->start != req_start) {
-                        if (l_extent->end >= req_start)
-                                new_ex->start = req_start;
-                        else
-                                new_ex->start = min(l_extent->end+1, req_start);
-                }
-
-                /* If we need to cancel this lock anyways because our request
-                 * overlaps the granted lock, we grow up to its requested
-                 * extent start instead of limiting this extent, assuming that
-                 * clients are writing forwards and the lock had over grown
-                 * its extent downwards before we enqueued our request. */
-                if (l_extent->end > req_end) {
-                        if (l_extent->start <= req_end)
-                                new_ex->end = max(lock->l_req_extent.start - 1,
-                                                  req_end);
-                        else
-                                new_ex->end = max(l_extent->start - 1, req_end);
-                }
-        }
-
-        ldlm_extent_internal_policy_fixup(req, new_ex, conflicting);
-        EXIT;
+               /* Don't conflict with ourselves */
+               if (req == lock)
+                       continue;
+
+               /* Locks are compatible, overlap doesn't matter
+                * Until bug 20 is fixed, try to avoid granting overlapping
+                * locks on one client (they take a long time to cancel)
+                */
+               if (lockmode_compat(lock->l_req_mode, req_mode) &&
+                   lock->l_export != req->l_export)
+                       continue;
+
+               /* If this is a high-traffic lock, don't grow downwards at all
+                * or grow upwards too much
+                */
+               ++conflicting;
+               if (conflicting > 4)
+                       new_ex->start = req_start;
+
+               /* If lock doesn't overlap new_ex, skip it. */
+               if (!ldlm_extent_overlap(l_extent, new_ex))
+                       continue;
+
+               /* Locks conflicting in requested extents and we can't satisfy
+                * both locks, so ignore it.  Either we will ping-pong this
+                * extent (we would regardless of what extent we granted) or
+                * lock is unused and it shouldn't limit our extent growth.
+                */
+               if (ldlm_extent_overlap(&lock->l_req_extent,
+                                        &req->l_req_extent))
+                       continue;
+
+               /* We grow extents downwards only as far as they don't overlap
+                * with already-granted locks, on the assumption that clients
+                * will be writing beyond the initial requested end and would
+                * then need to enqueue a new lock beyond previous request.
+                * l_req_extent->end strictly < req_start, checked above.
+                */
+               if (l_extent->start < req_start && new_ex->start != req_start) {
+                       if (l_extent->end >= req_start)
+                               new_ex->start = req_start;
+                       else
+                               new_ex->start = min(l_extent->end+1, req_start);
+               }
+
+               /* If we need to cancel this lock anyways because our request
+                * overlaps the granted lock, we grow up to its requested
+                * extent start instead of limiting this extent, assuming that
+                * clients are writing forwards and the lock had over grown
+                * its extent downwards before we enqueued our request.
+                */
+               if (l_extent->end > req_end) {
+                       if (l_extent->start <= req_end)
+                               new_ex->end = max(lock->l_req_extent.start - 1,
+                                                 req_end);
+                       else
+                               new_ex->end = max(l_extent->start - 1, req_end);
+               }
+       }
+
+       ldlm_extent_internal_policy_fixup(req, new_ex, conflicting);
+       EXIT;
 }
 
 
 /* In order to determine the largest possible extent we can grant, we need
- * to scan all of the queues. */
+ * to scan all of the queues.
+ */
 static void ldlm_extent_policy(struct ldlm_resource *res,
                               struct ldlm_lock *lock, __u64 *flags)
 {
        struct ldlm_extent new_ex = { .start = 0, .end = OBD_OBJECT_EOF };
 
        if (lock->l_export == NULL)
-               /*
-                * this is a local lock taken by server (e.g., as a part of
+               /* this is a local lock taken by server (e.g., as a part of
                 * OST-side locking, or unlink handling). Expansion doesn't
                 * make a lot of sense for local locks, because they are
                 * dropped immediately on operation completion and would only
@@ -287,7 +296,8 @@ static void ldlm_extent_policy(struct ldlm_resource *res,
 
        /* Because reprocess_queue zeroes flags and uses it to return
         * LDLM_FL_LOCK_CHANGED, we must check for the NO_EXPANSION flag
-        * in the lock flags rather than the 'flags' argument */
+        * in the lock flags rather than the 'flags' argument
+        */
        if (likely(!(lock->l_flags & LDLM_FL_NO_EXPANSION))) {
                ldlm_extent_internal_policy_granted(lock, &new_ex);
                ldlm_extent_internal_policy_waiting(lock, &new_ex);
@@ -296,7 +306,8 @@ static void ldlm_extent_policy(struct ldlm_resource *res,
                new_ex.start = lock->l_policy_data.l_extent.start;
                new_ex.end = lock->l_policy_data.l_extent.end;
                /* In case the request is not on correct boundaries, we call
-                * fixup. (normally called in ldlm_extent_internal_policy_*) */
+                * fixup. (normally called in ldlm_extent_internal_policy_*)
+                */
                ldlm_extent_internal_policy_fixup(lock, &new_ex, 0);
        }
 
@@ -320,7 +331,7 @@ static bool ldlm_check_contention(struct ldlm_lock *lock, int contended_locks)
                res->lr_contention_time = now;
 
        return now < res->lr_contention_time +
-                    ldlm_res_to_ns(res)->ns_contention_time;
+               ldlm_res_to_ns(res)->ns_contention_time;
 }
 
 struct ldlm_extent_compat_args {
@@ -341,32 +352,32 @@ static enum interval_iter ldlm_extent_compat_cb(struct interval_node *n,
        struct ldlm_lock *lock, *enq = priv->lock;
        enum ldlm_mode mode = priv->mode;
        int count = 0;
+
        ENTRY;
 
        LASSERT(!list_empty(&node->li_group));
 
        list_for_each_entry(lock, &node->li_group, l_sl_policy) {
-                /* interval tree is for granted lock */
-                LASSERTF(mode == lock->l_granted_mode,
-                         "mode = %s, lock->l_granted_mode = %s\n",
-                         ldlm_lockname[mode],
-                         ldlm_lockname[lock->l_granted_mode]);
-                count++;
-               if (lock->l_blocking_ast &&
-                   lock->l_granted_mode != LCK_GROUP)
-                        ldlm_add_ast_work_item(lock, enq, work_list);
-        }
-
-        /* don't count conflicting glimpse locks */
-        extent = ldlm_interval_extent(node);
-        if (!(mode == LCK_PR &&
-            extent->start == 0 && extent->end == OBD_OBJECT_EOF))
-                *priv->locks += count;
-
-        if (priv->compat)
-                *priv->compat = 0;
-
-        RETURN(INTERVAL_ITER_CONT);
+               /* interval tree is for granted lock */
+               LASSERTF(mode == lock->l_granted_mode,
+                        "mode = %s, lock->l_granted_mode = %s\n",
+                        ldlm_lockname[mode],
+                        ldlm_lockname[lock->l_granted_mode]);
+               count++;
+               if (lock->l_blocking_ast && lock->l_granted_mode != LCK_GROUP)
+                       ldlm_add_ast_work_item(lock, enq, work_list);
+       }
+
+       /* don't count conflicting glimpse locks */
+       extent = ldlm_interval_extent(node);
+       if (!(mode == LCK_PR && extent->start == 0 &&
+           extent->end == OBD_OBJECT_EOF))
+               *priv->locks += count;
+
+       if (priv->compat)
+               *priv->compat = 0;
+
+       RETURN(INTERVAL_ITER_CONT);
 }
 
 /**
@@ -393,60 +404,64 @@ ldlm_extent_compat_queue(struct list_head *queue, struct ldlm_lock *req,
        struct ldlm_lock *lock;
        int check_contention;
        int compat = 1;
+
        ENTRY;
 
-        lockmode_verify(req_mode);
-
-        /* Using interval tree for granted lock */
-        if (queue == &res->lr_granted) {
-                struct ldlm_interval_tree *tree;
-                struct ldlm_extent_compat_args data = {.work_list = work_list,
-                                               .lock = req,
-                                               .locks = contended_locks,
-                                               .compat = &compat };
-                struct interval_node_extent ex = { .start = req_start,
-                                                   .end = req_end };
-                int idx, rc;
-
-                for (idx = 0; idx < LCK_MODE_NUM; idx++) {
-                        tree = &res->lr_itree[idx];
-                        if (tree->lit_root == NULL) /* empty tree, skipped */
-                                continue;
-
-                        data.mode = tree->lit_mode;
-                        if (lockmode_compat(req_mode, tree->lit_mode)) {
-                                struct ldlm_interval *node;
-                                struct ldlm_extent *extent;
-
-                                if (req_mode != LCK_GROUP)
-                                        continue;
-
-                                /* group lock, grant it immediately if
-                                 * compatible */
-                                node = to_ldlm_interval(tree->lit_root);
-                                extent = ldlm_interval_extent(node);
-                                if (req->l_policy_data.l_extent.gid ==
-                                    extent->gid)
-                                        RETURN(2);
-                        }
-
-                        if (tree->lit_mode == LCK_GROUP) {
+       lockmode_verify(req_mode);
+
+       /* Using interval tree for granted lock */
+       if (queue == &res->lr_granted) {
+               struct ldlm_interval_tree *tree;
+               struct ldlm_extent_compat_args data = {
+                       .work_list = work_list,
+                       .lock = req,
+                       .locks = contended_locks,
+                       .compat = &compat };
+               struct interval_node_extent ex = {
+                       .start = req_start,
+                       .end = req_end };
+               int idx, rc;
+
+               for (idx = 0; idx < LCK_MODE_NUM; idx++) {
+                       tree = &res->lr_itree[idx];
+                       if (tree->lit_root == NULL) /* empty tree, skipped */
+                               continue;
+
+                       data.mode = tree->lit_mode;
+                       if (lockmode_compat(req_mode, tree->lit_mode)) {
+                               struct ldlm_interval *node;
+                               struct ldlm_extent *extent;
+
+                               if (req_mode != LCK_GROUP)
+                                       continue;
+
+                               /* group lock,grant it immediately if
+                                * compatible */
+                               node = to_ldlm_interval(tree->lit_root);
+                               extent = ldlm_interval_extent(node);
+                               if (req->l_policy_data.l_extent.gid ==
+                                   extent->gid)
+                                       RETURN(2);
+                       }
+
+                       if (tree->lit_mode == LCK_GROUP) {
                                if (*flags & (LDLM_FL_BLOCK_NOWAIT |
-                                             LDLM_FL_SPECULATIVE)) {
+                                   LDLM_FL_SPECULATIVE)) {
                                        compat = -EAGAIN;
-                                        goto destroylock;
-                                }
+                                       goto destroylock;
+                               }
 
-                                if (!work_list)
-                                        RETURN(0);
+                               if (!work_list)
+                                       RETURN(0);
 
-                                /* if work list is not NULL,add all
-                                   locks in the tree to work list */
-                                compat = 0;
-                                interval_iterate(tree->lit_root,
-                                                 ldlm_extent_compat_cb, &data);
-                                continue;
-                        }
+                               /* if work list is not NULL,add all
+                                * locks in the tree to work list
+                                */
+                               compat = 0;
+                               interval_iterate(tree->lit_root,
+                                                ldlm_extent_compat_cb, &data);
+                               continue;
+                       }
 
                        /* We've found a potentially blocking lock, check
                         * compatibility.  This handles locks other than GROUP
@@ -454,7 +469,8 @@ ldlm_extent_compat_queue(struct list_head *queue, struct ldlm_lock *req,
                         *
                         * Locks with FL_SPECULATIVE are asynchronous requests
                         * which must never wait behind another lock, so they
-                        * fail if any conflicting lock is found. */
+                        * fail if any conflicting lock is found.
+                        */
                        if (!work_list || (*flags & LDLM_FL_SPECULATIVE)) {
                                rc = interval_is_overlapped(tree->lit_root,
                                                            &ex);
@@ -466,166 +482,182 @@ ldlm_extent_compat_queue(struct list_head *queue, struct ldlm_lock *req,
                                                goto destroylock;
                                        }
                                }
-                        } else {
-                                interval_search(tree->lit_root, &ex,
-                                                ldlm_extent_compat_cb, &data);
+                       } else {
+                               interval_search(tree->lit_root, &ex,
+                                               ldlm_extent_compat_cb, &data);
                                if (!list_empty(work_list) && compat)
-                                        compat = 0;
-                        }
-                }
-        } else { /* for waiting queue */
+                                       compat = 0;
+                       }
+               }
+       } else { /* for waiting queue */
                list_for_each_entry(lock, queue, l_res_link) {
-                        check_contention = 1;
+                       check_contention = 1;
 
                        /* We stop walking the queue if we hit ourselves so
                         * we don't take conflicting locks enqueued after us
-                        * into account, or we'd wait forever. */
-                        if (req == lock)
-                                break;
-
-                        /* locks are compatible, overlap doesn't matter */
-                        if (lockmode_compat(lock->l_req_mode, req_mode)) {
-                                if (req_mode == LCK_PR &&
-                                    ((lock->l_policy_data.l_extent.start <=
-                                      req->l_policy_data.l_extent.start) &&
-                                     (lock->l_policy_data.l_extent.end >=
-                                      req->l_policy_data.l_extent.end))) {
+                        * into account, or we'd wait forever.
+                        */
+                       if (req == lock)
+                               break;
+
+                       /* locks are compatible, overlap doesn't matter */
+                       if (lockmode_compat(lock->l_req_mode, req_mode)) {
+                               if (req_mode == LCK_PR &&
+                               ((lock->l_policy_data.l_extent.start <=
+                               req->l_policy_data.l_extent.start) &&
+                               (lock->l_policy_data.l_extent.end >=
+                               req->l_policy_data.l_extent.end))) {
                                        /* If we met a PR lock just like us or
-                                          wider, and nobody down the list
-                                          conflicted with it, that means we
-                                          can skip processing of the rest of
-                                          the list and safely place ourselves
-                                          at the end of the list, or grant
-                                          (dependent if we met an conflicting
-                                          locks before in the list).  In case
-                                          of 1st enqueue only we continue
-                                          traversing if there is something
-                                          conflicting down the list because
-                                          we need to make sure that something
-                                          is marked as AST_SENT as well, in
-                                          cse of empy worklist we would exit
-                                          on first conflict met. */
+                                        * wider, and nobody down the list
+                                        * conflicted with it, that means we
+                                        * can skip processing of the rest of
+                                        * the list and safely place ourselves
+                                        * at the end of the list, or grant
+                                        * (dependent if we met an conflicting
+                                        * locks before in the list).  In case
+                                        * of 1st enqueue only we continue
+                                        * traversing if there is something
+                                        * conflicting down the list because
+                                        * we need to make sure that something
+                                        * is marked as AST_SENT as well, in
+                                        * case of empy worklist we would exit
+                                        * on first conflict met.
+                                        */
+
                                        /* There IS a case where such flag is
-                                          not set for a lock, yet it blocks
-                                          something. Luckily for us this is
-                                          only during destroy, so lock is
-                                          exclusive. So here we are safe */
+                                        * not set for a lock, yet it blocks
+                                        * something. Luckily for us this is
+                                        * only during destroy, so lock is
+                                        * exclusive. So here we are safe
+                                        */
                                        if (!ldlm_is_ast_sent(lock))
                                                RETURN(compat);
-                                }
-
-                                /* non-group locks are compatible, overlap doesn't
-                                   matter */
-                                if (likely(req_mode != LCK_GROUP))
-                                        continue;
-
-                                /* If we are trying to get a GROUP lock and there is
-                                   another one of this kind, we need to compare gid */
-                                if (req->l_policy_data.l_extent.gid ==
-                                    lock->l_policy_data.l_extent.gid) {
-                                        /* If existing lock with matched gid is granted,
-                                           we grant new one too. */
+                               }
+
+                               /* non-group locks are compatible,
+                                * overlap doesn't matter
+                                */
+                               if (likely(req_mode != LCK_GROUP))
+                                       continue;
+
+                               /* If we are trying to get a GROUP lock and
+                                * there is another one of this kind, we need
+                                * to compare gid
+                                */
+                               if (req->l_policy_data.l_extent.gid ==
+                               lock->l_policy_data.l_extent.gid) {
+                                       /* If existing lock with matched gid
+                                        * is granted, we grant new one too.
+                                        */
                                        if (ldlm_is_granted(lock))
                                                RETURN(2);
 
-                                        /* Otherwise we are scanning queue of waiting
-                                         * locks and it means current request would
-                                         * block along with existing lock (that is
-                                         * already blocked.
-                                         * If we are in nonblocking mode - return
-                                         * immediately */
-                                       if (*flags & (LDLM_FL_BLOCK_NOWAIT
-                                                     | LDLM_FL_SPECULATIVE)) {
+                                       /* Otherwise we are scanning queue of
+                                        * waiting locks and it means current
+                                        * request would block along with
+                                        * existing lock (that is already
+                                        * blocked.  If we are in nonblocking
+                                        * mode - return immediately
+                                        */
+                                       if (*flags & (LDLM_FL_BLOCK_NOWAIT |
+                                                     LDLM_FL_SPECULATIVE)) {
                                                compat = -EAGAIN;
-                                                goto destroylock;
-                                        }
-                                        /* If this group lock is compatible with another
-                                         * group lock on the waiting list, they must be
-                                         * together in the list, so they can be granted
-                                         * at the same time.  Otherwise the later lock
-                                         * can get stuck behind another, incompatible,
-                                         * lock. */
-                                        ldlm_resource_insert_lock_after(lock, req);
-                                        /* Because 'lock' is not granted, we can stop
-                                         * processing this queue and return immediately.
-                                         * There is no need to check the rest of the
-                                         * list. */
-                                        RETURN(0);
-                                }
-                        }
+                                               goto destroylock;
+                                       }
+                                       /* If this group lock is compatible with
+                                        * another group lock on the waiting
+                                        * list, they must be together in the
+                                        * list, so they can be granted at the
+                                        * same time.  Otherwise the later lock
+                                        * can get stuck behind another,
+                                        * incompatible, lock.
+                                        */
+                                       ldlm_resource_insert_lock_after(lock,
+                                                                       req);
+                                       /* Because 'lock' is not granted, we can
+                                        * stop processing this queue and return
+                                        * immediately. There is no need to
+                                        * check the rest of the list.
+                                        */
+                                       RETURN(0);
+                               }
+                       }
 
                        if (unlikely(req_mode == LCK_GROUP &&
-                                    !ldlm_is_granted(lock))) {
-                                compat = 0;
-                                if (lock->l_req_mode != LCK_GROUP) {
-                                        /* Ok, we hit non-GROUP lock, there should be no
-                                           more GROUP locks later on, queue in front of
-                                           first non-GROUP lock */
-
-                                       ldlm_resource_insert_lock_before(lock, req);
-                                        break;
-                                }
+                           !ldlm_is_granted(lock))) {
+                               compat = 0;
+                               if (lock->l_req_mode != LCK_GROUP) {
+                                       /* Ok, we hit non-GROUP lock, there
+                                        * should be no more GROUP locks later
+                                        * on, queue in front of first
+                                        * non-GROUP lock
+                                        */
+
+                                       ldlm_resource_insert_lock_before(lock,
+                                                                        req);
+                                       break;
+                               }
                                LASSERT(req->l_policy_data.l_extent.gid !=
                                        lock->l_policy_data.l_extent.gid);
                                continue;
-                        }
+                       }
 
-                        if (unlikely(lock->l_req_mode == LCK_GROUP)) {
+                       if (unlikely(lock->l_req_mode == LCK_GROUP)) {
                                /* If compared lock is GROUP, then requested is
                                 * PR/PW so this is not compatible; extent
-                                * range does not matter */
-                               if (*flags & (LDLM_FL_BLOCK_NOWAIT
-                                             | LDLM_FL_SPECULATIVE)) {
+                                * range does not matter
+                                */
+                               if (*flags & (LDLM_FL_BLOCK_NOWAIT |
+                                   LDLM_FL_SPECULATIVE)) {
                                        compat = -EAGAIN;
-                                        goto destroylock;
-                                }
-                        } else if (lock->l_policy_data.l_extent.end < req_start ||
-                                   lock->l_policy_data.l_extent.start > req_end) {
-                                /* if a non group lock doesn't overlap skip it */
-                                continue;
-                        } else if (lock->l_req_extent.end < req_start ||
-                                   lock->l_req_extent.start > req_end) {
-                                /* false contention, the requests doesn't really overlap */
-                                check_contention = 0;
-                        }
-
-                        if (!work_list)
-                                RETURN(0);
+                                       goto destroylock;
+                               }
+                       } else if (lock->l_policy_data.l_extent.end < req_start ||
+                                  lock->l_policy_data.l_extent.start > req_end) {
+                               /* if non group lock doesn't overlap skip it */
+                               continue;
+                       } else if (lock->l_req_extent.end < req_start ||
+                                  lock->l_req_extent.start > req_end) {
+                               /* false contention, the requests doesn't
+                                * really overlap */
+                               check_contention = 0;
+                       }
+
+                       if (!work_list)
+                               RETURN(0);
 
                        if (*flags & LDLM_FL_SPECULATIVE) {
                                compat = -EAGAIN;
                                goto destroylock;
                        }
 
-                        /* don't count conflicting glimpse locks */
-                        if (lock->l_req_mode == LCK_PR &&
-                            lock->l_policy_data.l_extent.start == 0 &&
-                            lock->l_policy_data.l_extent.end == OBD_OBJECT_EOF)
-                                check_contention = 0;
+                       /* don't count conflicting glimpse locks */
+                       if (lock->l_req_mode == LCK_PR &&
+                           lock->l_policy_data.l_extent.start == 0 &&
+                            lock->l_policy_data.l_extent.end == OBD_OBJECT_EOF)
+                               check_contention = 0;
 
-                        *contended_locks += check_contention;
+                       *contended_locks += check_contention;
 
-                        compat = 0;
+                       compat = 0;
                        if (lock->l_blocking_ast &&
                            lock->l_req_mode != LCK_GROUP)
-                                ldlm_add_ast_work_item(lock, req, work_list);
-                }
-        }
-
-        if (ldlm_check_contention(req, *contended_locks) &&
-            compat == 0 &&
-            (*flags & LDLM_FL_DENY_ON_CONTENTION) &&
-            req->l_req_mode != LCK_GROUP &&
-            req_end - req_start <=
-            ldlm_res_to_ns(req->l_resource)->ns_max_nolock_size)
-                GOTO(destroylock, compat = -EUSERS);
-
-        RETURN(compat);
+                               ldlm_add_ast_work_item(lock, req, work_list);
+               }
+       }
+
+       if (ldlm_check_contention(req, *contended_locks) &&
+           compat == 0 && (*flags & LDLM_FL_DENY_ON_CONTENTION) &&
+            req->l_req_mode != LCK_GROUP && req_end - req_start <=
+             ldlm_res_to_ns(req->l_resource)->ns_max_nolock_size)
+               GOTO(destroylock, compat = -EUSERS);
+
+       RETURN(compat);
 destroylock:
        list_del_init(&req->l_res_link);
-        ldlm_lock_destroy_nolock(req);
-        RETURN(compat);
-}
+       ldlm_lock_destroy_nolock(req);
+       RETURN(compat);
+       }
 
 /**
  * This function refresh eviction timer for cancelled lock.
@@ -690,7 +722,7 @@ void ldlm_resource_prolong(struct ldlm_prolong_args *arg)
        struct ldlm_interval_tree *tree;
        struct ldlm_resource *res;
        struct interval_node_extent ex = { .start = arg->lpa_extent.start,
-                                          .end = arg->lpa_extent.end };
+                                       .end = arg->lpa_extent.end };
        int idx;
 
        ENTRY;
@@ -699,7 +731,7 @@ void ldlm_resource_prolong(struct ldlm_prolong_args *arg)
                                &arg->lpa_resid, LDLM_EXTENT, 0);
        if (IS_ERR(res)) {
                CDEBUG(D_DLMTRACE, "Failed to get resource for resid %llu/%llu\n",
-                      arg->lpa_resid.name[0], arg->lpa_resid.name[1]);
+               arg->lpa_resid.name[0], arg->lpa_resid.name[1]);
                RETURN_EXIT;
        }
 
@@ -720,7 +752,6 @@ void ldlm_resource_prolong(struct ldlm_prolong_args *arg)
                interval_search(tree->lit_root, &ex,
                                ldlm_resource_prolong_cb, arg);
        }
-
        unlock_res(res);
        ldlm_resource_putref(res);
 
@@ -737,8 +768,8 @@ EXPORT_SYMBOL(ldlm_resource_prolong);
  * either queue.
  */
 int ldlm_process_extent_lock(struct ldlm_lock *lock, __u64 *flags,
-                            enum ldlm_process_intention intention,
-                            enum ldlm_error *err, struct list_head *work_list)
+               enum ldlm_process_intention intention,
+               enum ldlm_error *err, struct list_head *work_list)
 {
        struct ldlm_resource *res = lock->l_resource;
        int rc, rc2 = 0;
@@ -759,35 +790,35 @@ int ldlm_process_extent_lock(struct ldlm_lock *lock, __u64 *flags,
                 * can only return -EAGAIN if (flags & BLOCK_NOWAIT |
                 * SPECULATIVE). flags should always be zero here, and if that
                 * ever stops being true, we want to find out. */
-                LASSERT(*flags == 0);
-                rc = ldlm_extent_compat_queue(&res->lr_granted, lock, flags,
-                                             NULL, &contended_locks);
-                if (rc == 1) {
-                        rc = ldlm_extent_compat_queue(&res->lr_waiting, lock,
-                                                     flags, NULL,
-                                                      &contended_locks);
-                }
-                if (rc == 0)
-                        RETURN(LDLM_ITER_STOP);
-
-                ldlm_resource_unlink_lock(lock);
-
-                if (!OBD_FAIL_CHECK(OBD_FAIL_LDLM_CANCEL_EVICT_RACE))
-                        ldlm_extent_policy(res, lock, flags);
+               LASSERT(*flags == 0);
+               rc = ldlm_extent_compat_queue(&res->lr_granted, lock, flags,
+                                       NULL, &contended_locks);
+               if (rc == 1) {
+                       rc = ldlm_extent_compat_queue(&res->lr_waiting, lock,
+                                               flags, NULL,
+                                               &contended_locks);
+               }
+               if (rc == 0)
+                       RETURN(LDLM_ITER_STOP);
+
+               ldlm_resource_unlink_lock(lock);
+
+               if (!OBD_FAIL_CHECK(OBD_FAIL_LDLM_CANCEL_EVICT_RACE))
+                       ldlm_extent_policy(res, lock, flags);
                ldlm_grant_lock(lock, grant_work);
-                RETURN(LDLM_ITER_CONTINUE);
-        }
+               RETURN(LDLM_ITER_CONTINUE);
+       }
 
-        contended_locks = 0;
+       contended_locks = 0;
        rc = ldlm_extent_compat_queue(&res->lr_granted, lock, flags,
-                                     work_list, &contended_locks);
+                               work_list, &contended_locks);
        if (rc < 0)
                GOTO(out, *err = rc);
 
        if (rc != 2) {
                rc2 = ldlm_extent_compat_queue(&res->lr_waiting, lock,
-                                              flags, work_list,
-                                              &contended_locks);
+                                       flags, work_list,
+                                       &contended_locks);
                if (rc2 < 0)
                        GOTO(out, *err = rc = rc2);
        }
@@ -845,7 +876,7 @@ static enum interval_iter ldlm_kms_shift_cb(struct interval_node *n,
         * highest lock (or we share that distinction with another lock), and
         * don't need to update KMS.  Return old_kms and stop looking. */
        if (lock->l_policy_data.l_extent.end == OBD_OBJECT_EOF ||
-           lock->l_policy_data.l_extent.end + 1 >= arg->old_kms) {
+       lock->l_policy_data.l_extent.end + 1 >= arg->old_kms) {
                arg->kms = arg->old_kms;
                arg->complete = true;
                RETURN(INTERVAL_ITER_STOP);
@@ -872,7 +903,8 @@ static enum interval_iter ldlm_kms_shift_cb(struct interval_node *n,
  *
  * Caller must hold lr_lock already.
  *
- * NB: A lock on [x,y] protects a KMS of up to y + 1 bytes! */
+ * NB: A lock on [x,y] protects a KMS of up to y + 1 bytes!
+ */
 __u64 ldlm_extent_shift_kms(struct ldlm_lock *lock, __u64 old_kms)
 {
        struct ldlm_resource *res = lock->l_resource;
@@ -923,6 +955,7 @@ struct kmem_cache *ldlm_interval_slab;
 static struct ldlm_interval *ldlm_interval_alloc(struct ldlm_lock *lock)
 {
        struct ldlm_interval *node;
+
        ENTRY;
 
        LASSERT(lock->l_resource->lr_type == LDLM_EXTENT);
@@ -937,33 +970,33 @@ static struct ldlm_interval *ldlm_interval_alloc(struct ldlm_lock *lock)
 
 void ldlm_interval_free(struct ldlm_interval *node)
 {
-        if (node) {
+       if (node) {
                LASSERT(list_empty(&node->li_group));
-                LASSERT(!interval_is_intree(&node->li_node));
-                OBD_SLAB_FREE(node, ldlm_interval_slab, sizeof(*node));
-        }
+               LASSERT(!interval_is_intree(&node->li_node));
+               OBD_SLAB_FREE(node, ldlm_interval_slab, sizeof(*node));
+       }
 }
 
 /* interval tree, for LDLM_EXTENT. */
 void ldlm_interval_attach(struct ldlm_interval *n,
-                          struct ldlm_lock *l)
+                         struct ldlm_lock *l)
 {
-        LASSERT(l->l_tree_node == NULL);
-        LASSERT(l->l_resource->lr_type == LDLM_EXTENT);
+       LASSERT(l->l_tree_node == NULL);
+       LASSERT(l->l_resource->lr_type == LDLM_EXTENT);
 
        list_add_tail(&l->l_sl_policy, &n->li_group);
-        l->l_tree_node = n;
+       l->l_tree_node = n;
 }
 
 struct ldlm_interval *ldlm_interval_detach(struct ldlm_lock *l)
 {
-        struct ldlm_interval *n = l->l_tree_node;
+       struct ldlm_interval *n = l->l_tree_node;
 
-        if (n == NULL)
-                return NULL;
+       if (n == NULL)
+               return NULL;
 
        LASSERT(!list_empty(&n->li_group));
-        l->l_tree_node = NULL;
+       l->l_tree_node = NULL;
        list_del_init(&l->l_sl_policy);
 
        return list_empty(&n->li_group) ? n : NULL;
@@ -983,6 +1016,7 @@ static inline int ldlm_mode_to_index(enum ldlm_mode mode)
 int ldlm_extent_alloc_lock(struct ldlm_lock *lock)
 {
        lock->l_tree_node = NULL;
+
        if (ldlm_interval_alloc(lock) == NULL)
                return -ENOMEM;
        return 0;
@@ -990,57 +1024,59 @@ int ldlm_extent_alloc_lock(struct ldlm_lock *lock)
 
 /** Add newly granted lock into interval tree for the resource. */
 void ldlm_extent_add_lock(struct ldlm_resource *res,
-                          struct ldlm_lock *lock)
+                         struct ldlm_lock *lock)
 {
-        struct interval_node *found, **root;
-        struct ldlm_interval *node;
-        struct ldlm_extent *extent;
+       struct interval_node *found, **root;
+       struct ldlm_interval *node;
+       struct ldlm_extent *extent;
        int idx, rc;
 
        LASSERT(ldlm_is_granted(lock));
 
-        node = lock->l_tree_node;
-        LASSERT(node != NULL);
-        LASSERT(!interval_is_intree(&node->li_node));
+       node = lock->l_tree_node;
+       LASSERT(node != NULL);
+       LASSERT(!interval_is_intree(&node->li_node));
 
        idx = ldlm_mode_to_index(lock->l_granted_mode);
        LASSERT(lock->l_granted_mode == BIT(idx));
        LASSERT(lock->l_granted_mode == res->lr_itree[idx].lit_mode);
 
-        /* node extent initialize */
-        extent = &lock->l_policy_data.l_extent;
+       /* node extent initialize */
+       extent = &lock->l_policy_data.l_extent;
 
        rc = interval_set(&node->li_node, extent->start, extent->end);
        LASSERT(!rc);
 
-        root = &res->lr_itree[idx].lit_root;
-        found = interval_insert(&node->li_node, root);
-        if (found) { /* The policy group found. */
-                struct ldlm_interval *tmp = ldlm_interval_detach(lock);
-                LASSERT(tmp != NULL);
-                ldlm_interval_free(tmp);
-                ldlm_interval_attach(to_ldlm_interval(found), lock);
-        }
-        res->lr_itree[idx].lit_size++;
+       root = &res->lr_itree[idx].lit_root;
+       found = interval_insert(&node->li_node, root);
+       if (found) { /* The policy group found. */
+               struct ldlm_interval *tmp = ldlm_interval_detach(lock);
 
-        /* even though we use interval tree to manage the extent lock, we also
-         * add the locks into grant list, for debug purpose, .. */
-        ldlm_resource_add_lock(res, &res->lr_granted, lock);
+               LASSERT(tmp != NULL);
+               ldlm_interval_free(tmp);
+               ldlm_interval_attach(to_ldlm_interval(found), lock);
+       }
+       res->lr_itree[idx].lit_size++;
+
+       /* even though we use interval tree to manage the extent lock, we also
+        * add the locks into grant list, for debug purpose, .. */
+       ldlm_resource_add_lock(res, &res->lr_granted, lock);
 
        if (OBD_FAIL_CHECK(OBD_FAIL_LDLM_GRANT_CHECK)) {
                struct ldlm_lock *lck;
 
                list_for_each_entry_reverse(lck, &res->lr_granted,
-                                           l_res_link) {
+                                       l_res_link) {
                        if (lck == lock)
                                continue;
                        if (lockmode_compat(lck->l_granted_mode,
-                                           lock->l_granted_mode))
+                                       lock->l_granted_mode))
                                continue;
                        if (ldlm_extent_overlap(&lck->l_req_extent,
                                                &lock->l_req_extent)) {
-                               CDEBUG(D_ERROR, "granting conflicting lock %p "
-                                               "%p\n", lck, lock);
+                               CDEBUG(D_ERROR,
+                                      "granting conflicting lock %p %p\n",
+                                      lck, lock);
                                ldlm_resource_dump(D_ERROR, res);
                                LBUG();
                        }
@@ -1089,4 +1125,3 @@ void ldlm_extent_policy_local_to_wire(const union ldlm_policy_data *lpolicy,
        wpolicy->l_extent.end = lpolicy->l_extent.end;
        wpolicy->l_extent.gid = lpolicy->l_extent.gid;
 }
-