qa/suites: check dirfrag counter in cephfs workunits

(just the ones which really do do enough work to generate
some fragmentation)

Signed-off-by: John Spray <john.spray@redhat.com>

qa/cephfs/tasks/cfuse_workunit_suites_blogbench.yaml

@@ -1,4 +1,8 @@
 tasks:
+- check-counter:
+    counters:
+      mds:
+        - "mds.dir_split"
 - workunit:
     clients:
       all:

qa/cephfs/tasks/cfuse_workunit_suites_ffsb.yaml

@@ -4,6 +4,10 @@ overrides:
       osd:
         filestore flush min: 0
 tasks:
+- check-counter:
+    counters:
+      mds:
+        - "mds.dir_split"
 - workunit:
     clients:
       all:

qa/cephfs/tasks/libcephfs_interface_tests.yaml

@@ -4,6 +4,10 @@ overrides:
   kclient:
     disabled: true
 tasks:
+- check-counter:
+    counters:
+      mds:
+        - "mds.dir_split"
 - workunit:
     clients:
       client.0:

qa/suites/fs/basic/tasks/cfuse_workunit_kernel_untar_build.yaml

@@ -4,6 +4,10 @@ overrides:
       client:
         fuse_default_permissions: 0
 tasks:
+- check-counter:
+    counters:
+      mds:
+        - "mds.dir_split"
 - workunit:
     clients:
       all:

qa/suites/fs/basic/tasks/cfuse_workunit_misc.yaml

@@ -1,4 +1,8 @@
 tasks:
+- check-counter:
+    counters:
+      mds:
+        - "mds.dir_split"
 - workunit:
     timeout: 6h
     clients:

qa/suites/fs/basic/tasks/cfuse_workunit_norstats.yaml

@@ -1,4 +1,8 @@
 tasks:
+- check-counter:
+    counters:
+      mds:
+        - "mds.dir_split"
 - workunit:
     timeout: 6h
     clients:

qa/suites/fs/basic/tasks/cfuse_workunit_suites_fsx.yaml

@@ -1,4 +1,8 @@
 tasks:
+- check-counter:
+    counters:
+      mds:
+        - "mds.dir_split"
 - workunit:
     clients:
       all: