*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
#include <lustre_intent.h>
#include <libcfs/libcfs.h>
#include <obd_class.h>
-#include <lustre/lustre_idl.h>
+#include <uapi/linux/lustre/lustre_idl.h>
#include <lustre_lib.h>
#include <lustre_dlm.h>
#include <lustre_export.h>
struct ptlrpc_request;
struct obd_device;
-/**
- * Serializes in-flight MDT-modifying RPC requests to preserve idempotency.
- *
- * This mutex is used to implement execute-once semantics on the MDT.
- * The MDT stores the last transaction ID and result for every client in
- * its last_rcvd file. If the client doesn't get a reply, it can safely
- * resend the request and the MDT will reconstruct the reply being aware
- * that the request has already been executed. Without this lock,
- * execution status of concurrent in-flight requests would be
- * overwritten.
- *
- * This design limits the extent to which we can keep a full pipeline of
- * in-flight requests from a single client. This limitation could be
- * overcome by allowing multiple slots per client in the last_rcvd file.
- */
-struct mdc_rpc_lock {
- /** Lock protecting in-flight RPC concurrency. */
- struct mutex rpcl_mutex;
- /** Intent associated with currently executing request. */
- struct lookup_intent *rpcl_it;
- /** Used for MDS/RPC load testing purposes. */
- int rpcl_fakes;
-};
-
-#define MDC_FAKE_RPCL_IT ((void *)0x2c0012bfUL)
-
-static inline void mdc_init_rpc_lock(struct mdc_rpc_lock *lck)
-{
- mutex_init(&lck->rpcl_mutex);
- lck->rpcl_it = NULL;
-}
-
-static inline void mdc_get_rpc_lock(struct mdc_rpc_lock *lck,
- struct lookup_intent *it)
+static inline void mdc_get_mod_rpc_slot(struct ptlrpc_request *req,
+ struct lookup_intent *it)
{
- ENTRY;
-
- if (it != NULL && (it->it_op == IT_GETATTR || it->it_op == IT_LOOKUP ||
- it->it_op == IT_LAYOUT || it->it_op == IT_READDIR))
- return;
-
- /* This would normally block until the existing request finishes.
- * If fail_loc is set it will block until the regular request is
- * done, then set rpcl_it to MDC_FAKE_RPCL_IT. Once that is set
- * it will only be cleared when all fake requests are finished.
- * Only when all fake requests are finished can normal requests
- * be sent, to ensure they are recoverable again. */
- again:
- mutex_lock(&lck->rpcl_mutex);
-
- if (CFS_FAIL_CHECK_QUIET(OBD_FAIL_MDC_RPCS_SEM)) {
- lck->rpcl_it = MDC_FAKE_RPCL_IT;
- lck->rpcl_fakes++;
- mutex_unlock(&lck->rpcl_mutex);
- return;
- }
+ struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
+ __u32 opc;
+ __u16 tag;
- /* This will only happen when the CFS_FAIL_CHECK() was
- * just turned off but there are still requests in progress.
- * Wait until they finish. It doesn't need to be efficient
- * in this extremely rare case, just have low overhead in
- * the common case when it isn't true. */
- while (unlikely(lck->rpcl_it == MDC_FAKE_RPCL_IT)) {
- mutex_unlock(&lck->rpcl_mutex);
- schedule_timeout(cfs_time_seconds(1) / 4);
- goto again;
- }
-
- LASSERT(lck->rpcl_it == NULL);
- lck->rpcl_it = it;
+ opc = lustre_msg_get_opc(req->rq_reqmsg);
+ tag = obd_get_mod_rpc_slot(cli, opc, it);
+ lustre_msg_set_tag(req->rq_reqmsg, tag);
}
-static inline void mdc_put_rpc_lock(struct mdc_rpc_lock *lck,
- struct lookup_intent *it)
+static inline void mdc_put_mod_rpc_slot(struct ptlrpc_request *req,
+ struct lookup_intent *it)
{
- if (it != NULL && (it->it_op == IT_GETATTR || it->it_op == IT_LOOKUP ||
- it->it_op == IT_LAYOUT || it->it_op == IT_READDIR))
- goto out;
-
- if (lck->rpcl_it == MDC_FAKE_RPCL_IT) { /* OBD_FAIL_MDC_RPCS_SEM */
- mutex_lock(&lck->rpcl_mutex);
+ struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
+ __u32 opc;
+ __u16 tag;
- LASSERTF(lck->rpcl_fakes > 0, "%d\n", lck->rpcl_fakes);
- lck->rpcl_fakes--;
-
- if (lck->rpcl_fakes == 0)
- lck->rpcl_it = NULL;
-
- } else {
- LASSERTF(it == lck->rpcl_it, "%p != %p\n", it, lck->rpcl_it);
- lck->rpcl_it = NULL;
- }
-
- mutex_unlock(&lck->rpcl_mutex);
- out:
- EXIT;
+ opc = lustre_msg_get_opc(req->rq_reqmsg);
+ tag = lustre_msg_get_tag(req->rq_reqmsg);
+ obd_put_mod_rpc_slot(cli, opc, it, tag);
}
-/* Update the maximum observed easize and cookiesize. The default easize
- * and cookiesize is initialized to the minimum value but allowed to grow
- * up to a single page in size if required to handle the common case.
+
+/**
+ * Update the maximum possible easize.
+ *
+ * This value is learned from ptlrpc replies sent by the MDT. The
+ * default easize is initialized to the minimum value but allowed to
+ * grow up to a single page in size if required to handle the common
+ * case.
+ *
+ * \see client_obd::cl_default_mds_easize
+ *
+ * \param[in] exp export for MDC device
+ * \param[in] body body of ptlrpc reply from MDT
+ *
*/
static inline void mdc_update_max_ea_from_body(struct obd_export *exp,
struct mdt_body *body)
{
if (body->mbo_valid & OBD_MD_FLMODEASIZE) {
struct client_obd *cli = &exp->exp_obd->u.cli;
+ __u32 def_easize;
- if (cli->cl_max_mds_easize < body->mbo_max_mdsize) {
+ if (cli->cl_max_mds_easize < body->mbo_max_mdsize)
cli->cl_max_mds_easize = body->mbo_max_mdsize;
- cli->cl_default_mds_easize =
- min_t(__u32, body->mbo_max_mdsize,
- PAGE_CACHE_SIZE);
- }
- if (cli->cl_max_mds_cookiesize < body->mbo_max_cookiesize) {
- cli->cl_max_mds_cookiesize = body->mbo_max_cookiesize;
- cli->cl_default_mds_cookiesize =
- min_t(__u32, body->mbo_max_cookiesize,
- PAGE_CACHE_SIZE);
- }
+
+ def_easize = min_t(__u32, body->mbo_max_mdsize,
+ OBD_MAX_DEFAULT_EA_SIZE);
+ cli->cl_default_mds_easize = def_easize;
}
}