Whamcloud - gitweb
LU-10467 lustre: convert most users of LWI_TIMEOUT_INTERVAL()
[fs/lustre-release.git] / lustre / ptlrpc / client.c
index b6c7782..bee8d3f 100644 (file)
@@ -37,6 +37,7 @@
 #include <linux/delay.h>
 #include <linux/random.h>
 
+#include <lnet/lib-lnet.h>
 #include <obd_support.h>
 #include <obd_class.h>
 #include <lustre_lib.h>
@@ -68,6 +69,28 @@ static void ptlrpc_release_bulk_page_pin(struct ptlrpc_bulk_desc *desc)
                put_page(BD_GET_KIOV(desc, i).kiov_page);
 }
 
+static int ptlrpc_prep_bulk_frag_pages(struct ptlrpc_bulk_desc *desc,
+                                      void *frag, int len)
+{
+       unsigned int offset = (uintptr_t)frag & ~PAGE_MASK;
+
+       ENTRY;
+       while (len > 0) {
+               int page_len = min_t(unsigned int, PAGE_SIZE - offset,
+                                    len);
+               uintptr_t vaddr = (uintptr_t) frag;
+
+               ptlrpc_prep_bulk_page_nopin(desc,
+                                           lnet_kvaddr_to_page(vaddr),
+                                           offset, page_len);
+               offset = 0;
+               len -= page_len;
+               frag += page_len;
+       }
+
+       RETURN(desc->bd_nob);
+}
+
 const struct ptlrpc_bulk_frag_ops ptlrpc_bulk_kiov_pin_ops = {
        .add_kiov_frag  = ptlrpc_prep_bulk_page_pin,
        .release_frags  = ptlrpc_release_bulk_page_pin,
@@ -77,6 +100,7 @@ EXPORT_SYMBOL(ptlrpc_bulk_kiov_pin_ops);
 const struct ptlrpc_bulk_frag_ops ptlrpc_bulk_kiov_nopin_ops = {
        .add_kiov_frag  = ptlrpc_prep_bulk_page_nopin,
        .release_frags  = ptlrpc_release_bulk_noop,
+       .add_iov_frag   = ptlrpc_prep_bulk_frag_pages,
 };
 EXPORT_SYMBOL(ptlrpc_bulk_kiov_nopin_ops);
 
@@ -217,7 +241,6 @@ struct ptlrpc_bulk_desc *ptlrpc_prep_bulk_imp(struct ptlrpc_request *req,
        if (!desc)
                RETURN(NULL);
 
-       desc->bd_import_generation = req->rq_import_generation;
        desc->bd_import = class_import_get(imp);
        desc->bd_req = req;
 
@@ -625,7 +648,7 @@ ptlrpc_init_rq_pool(int num_rq, int msgsize,
 {
        struct ptlrpc_request_pool *pool;
 
-       OBD_ALLOC(pool, sizeof(struct ptlrpc_request_pool));
+       OBD_ALLOC_PTR(pool);
        if (!pool)
                return NULL;
 
@@ -738,6 +761,41 @@ static inline void ptlrpc_assign_next_xid(struct ptlrpc_request *req)
 
 static atomic64_t ptlrpc_last_xid;
 
+static void ptlrpc_reassign_next_xid(struct ptlrpc_request *req)
+{
+       spin_lock(&req->rq_import->imp_lock);
+       list_del_init(&req->rq_unreplied_list);
+       ptlrpc_assign_next_xid_nolock(req);
+       spin_unlock(&req->rq_import->imp_lock);
+       DEBUG_REQ(D_RPCTRACE, req, "reassign xid");
+}
+
+void ptlrpc_get_mod_rpc_slot(struct ptlrpc_request *req)
+{
+       struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
+       __u32 opc;
+       __u16 tag;
+
+       opc = lustre_msg_get_opc(req->rq_reqmsg);
+       tag = obd_get_mod_rpc_slot(cli, opc);
+       lustre_msg_set_tag(req->rq_reqmsg, tag);
+       ptlrpc_reassign_next_xid(req);
+}
+EXPORT_SYMBOL(ptlrpc_get_mod_rpc_slot);
+
+void ptlrpc_put_mod_rpc_slot(struct ptlrpc_request *req)
+{
+       __u16 tag = lustre_msg_get_tag(req->rq_reqmsg);
+
+       if (tag != 0) {
+               struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
+               __u32 opc = lustre_msg_get_opc(req->rq_reqmsg);
+
+               obd_put_mod_rpc_slot(cli, opc, tag);
+       }
+}
+EXPORT_SYMBOL(ptlrpc_put_mod_rpc_slot);
+
 int ptlrpc_request_bufs_pack(struct ptlrpc_request *request,
                             __u32 version, int opcode, char **bufs,
                             struct ptlrpc_cli_ctx *ctx)
@@ -895,7 +953,6 @@ ptlrpc_request_alloc_internal(struct obd_import *imp,
                              const struct req_format *format)
 {
        struct ptlrpc_request *request;
-       int connect = 0;
 
        request = __ptlrpc_request_alloc(imp, pool);
        if (!request)
@@ -916,17 +973,17 @@ ptlrpc_request_alloc_internal(struct obd_import *imp,
                if (imp->imp_state == LUSTRE_IMP_IDLE) {
                        imp->imp_generation++;
                        imp->imp_initiated_at = imp->imp_generation;
-                       imp->imp_state =  LUSTRE_IMP_NEW;
-                       connect = 1;
-               }
-               spin_unlock(&imp->imp_lock);
-               if (connect) {
-                       rc = ptlrpc_connect_import(imp);
+                       imp->imp_state = LUSTRE_IMP_NEW;
+
+                       /* connect_import_locked releases imp_lock */
+                       rc = ptlrpc_connect_import_locked(imp);
                        if (rc < 0) {
                                ptlrpc_request_free(request);
                                return NULL;
                        }
                        ptlrpc_pinger_add_import(imp);
+               } else {
+                       spin_unlock(&imp->imp_lock);
                }
        }
 
@@ -2434,7 +2491,7 @@ int ptlrpc_set_wait(const struct lu_env *env, struct ptlrpc_request_set *set)
                       set, timeout);
 
                if ((timeout == 0 && !signal_pending(current)) ||
-                   set->set_allow_intr)
+                   set->set_allow_intr) {
                        /*
                         * No requests are in-flight (ether timed out
                         * or delayed), so we can allow interrupts.
@@ -2445,7 +2502,10 @@ int ptlrpc_set_wait(const struct lu_env *env, struct ptlrpc_request_set *set)
                                        cfs_time_seconds(timeout ? timeout : 1),
                                        ptlrpc_expired_set,
                                        ptlrpc_interrupted_set, set);
-               else
+
+                       rc = l_wait_event(set->set_waitq,
+                                         ptlrpc_check_set(NULL, set), &lwi);
+               } else {
                        /*
                         * At least one request is in flight, so no
                         * interrupts are allowed. Wait until all
@@ -2454,29 +2514,32 @@ int ptlrpc_set_wait(const struct lu_env *env, struct ptlrpc_request_set *set)
                        lwi = LWI_TIMEOUT(cfs_time_seconds(timeout ? timeout : 1),
                                          ptlrpc_expired_set, set);
 
-               rc = l_wait_event(set->set_waitq,
-                                 ptlrpc_check_set(NULL, set), &lwi);
-
-               /*
-                * LU-769 - if we ignored the signal because it was already
-                * pending when we started, we need to handle it now or we risk
-                * it being ignored forever
-                */
-               if (rc == -ETIMEDOUT &&
-                   (!lwi.lwi_allow_intr || set->set_allow_intr) &&
-                   signal_pending(current)) {
-                       sigset_t blocked_sigs =
-                                          cfs_block_sigsinv(LUSTRE_FATAL_SIGS);
+                       rc = l_wait_event(set->set_waitq,
+                                         ptlrpc_check_set(NULL, set), &lwi);
 
                        /*
-                        * In fact we only interrupt for the "fatal" signals
-                        * like SIGINT or SIGKILL. We still ignore less
-                        * important signals since ptlrpc set is not easily
-                        * reentrant from userspace again
+                        * LU-769 - if we ignored the signal because
+                        * it was already pending when we started, we
+                        * need to handle it now or we risk it being
+                        * ignored forever
                         */
-                       if (signal_pending(current))
-                               ptlrpc_interrupted_set(set);
-                       cfs_restore_sigs(blocked_sigs);
+                       if (rc == -ETIMEDOUT &&
+                           signal_pending(current)) {
+                               sigset_t blocked_sigs =
+                                       cfs_block_sigsinv(LUSTRE_FATAL_SIGS);
+
+                               /*
+                                * In fact we only interrupt for the
+                                * "fatal" signals like SIGINT or
+                                * SIGKILL. We still ignore less
+                                * important signals since ptlrpc set
+                                * is not easily reentrant from
+                                * userspace again
+                                */
+                               if (signal_pending(current))
+                                       ptlrpc_interrupted_set(set);
+                               cfs_restore_sigs(blocked_sigs);
+                       }
                }
 
                LASSERT(rc == 0 || rc == -EINTR || rc == -ETIMEDOUT);
@@ -2678,9 +2741,6 @@ EXPORT_SYMBOL(ptlrpc_req_xid);
  */
 static int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async)
 {
-       int rc;
-       struct l_wait_info lwi;
-
        /*
         * Might sleep.
         */
@@ -2721,24 +2781,25 @@ static int ptlrpc_unregister_reply(struct ptlrpc_request *request, int async)
         * unlinked before returning a req to the pool.
         */
        for (;;) {
-               /* The wq argument is ignored by user-space wait_event macros */
                wait_queue_head_t *wq = (request->rq_set) ?
                                        &request->rq_set->set_waitq :
                                        &request->rq_reply_waitq;
+               int seconds = LONG_UNLINK;
                /*
                 * Network access will complete in finite time but the HUGE
                 * timeout lets us CWARN for visibility of sluggish NALs
                 */
-               lwi = LWI_TIMEOUT_INTERVAL(cfs_time_seconds(LONG_UNLINK),
-                                          cfs_time_seconds(1), NULL, NULL);
-               rc = l_wait_event(*wq, !ptlrpc_client_recv_or_unlink(request),
-                                 &lwi);
-               if (rc == 0) {
+               while (seconds > 0 &&
+                      wait_event_idle_timeout(
+                              *wq,
+                              !ptlrpc_client_recv_or_unlink(request),
+                              cfs_time_seconds(1)) == 0)
+                       seconds -= 1;
+               if (seconds > 0) {
                        ptlrpc_rqphase_move(request, request->rq_next_phase);
                        RETURN(1);
                }
 
-               LASSERT(rc == -ETIMEDOUT);
                DEBUG_REQ(D_WARNING, request,
                          "Unexpectedly long timeout receiving_reply=%d req_ulinked=%d reply_unlinked=%d",
                          request->rq_receiving_reply,
@@ -3337,7 +3398,8 @@ void ptlrpc_init_xid(void)
        }
 
        /* Need to always be aligned to a power-of-two for mutli-bulk BRW */
-       CLASSERT((PTLRPC_BULK_OPS_COUNT & (PTLRPC_BULK_OPS_COUNT - 1)) == 0);
+       BUILD_BUG_ON((PTLRPC_BULK_OPS_COUNT & (PTLRPC_BULK_OPS_COUNT - 1)) !=
+                    0);
        xid &= PTLRPC_BULK_OPS_MASK;
        atomic64_set(&ptlrpc_last_xid, xid);
 }