From: Oleg Drokin Date: Mon, 22 Jul 2013 16:06:36 +0000 (+0800) Subject: staging/lustre: Make quota namespace refcounting consistent X-Git-Url: http://drtracing.org/?a=commitdiff_plain;h=df9fcbeb7730ec77a8f643ff1e00b99167cbeabe;p=deliverable%2Flinux.git staging/lustre: Make quota namespace refcounting consistent It seems quota namespace is needlessly referenced on connect, but that's not necessary as it could not go away until entire obd goes away. On the other hand this extra reference disturbs other logic depending on empty namespace having zero refcount, so this patch drops such extra referencing. This picks client side change of the original patch. Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2924 Lustre-change: http://review.whamcloud.com/6234 Signed-off-by: Oleg Drokin Reviewed-by: Niu Yawei Reviewed-by: Johann Lombardi Signed-off-by: Peng Tao Signed-off-by: Andreas Dilger Signed-off-by: Greg Kroah-Hartman --- diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h b/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h index 785c1a19da57..bf6f00341a76 100644 --- a/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h +++ b/drivers/staging/lustre/lustre/ldlm/ldlm_internal.h @@ -84,8 +84,7 @@ static inline struct mutex *ldlm_namespace_lock(ldlm_side_t client) &ldlm_srv_namespace_lock : &ldlm_cli_namespace_lock; } -/* ns_bref is the number of resources in this namespace with the notable - * exception of quota namespaces which have their empty refcount at 1 */ +/* ns_bref is the number of resources in this namespace */ static inline int ldlm_ns_empty(struct ldlm_namespace *ns) { return atomic_read(&ns->ns_bref) == 0; diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c index 716283859d72..17eab79ba839 100644 --- a/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c +++ b/drivers/staging/lustre/lustre/ldlm/ldlm_resource.c @@ -1171,11 +1171,7 @@ ldlm_resource_get(struct ldlm_namespace *ns, struct ldlm_resource *parent, /* Let's see if we happened to be the very first resource in this * namespace. If so, and this is a client namespace, we need to move * the namespace into the active namespaces list to be patrolled by - * the ldlm_poold. - * A notable exception, for quota namespaces qsd_lib.c already took a - * namespace reference, so it won't be participating in all of this, - * but I guess that's ok since we have no business cancelling quota - * locks anyway */ + * the ldlm_poold. */ if (ns_is_client(ns) && ns_refcount == 1) { mutex_lock(ldlm_namespace_lock(LDLM_NAMESPACE_CLIENT)); ldlm_namespace_move_to_active_locked(ns, LDLM_NAMESPACE_CLIENT);