*
* 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
*/
* Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2010, 2014, Intel Corporation.
+ * Copyright (c) 2010, 2016, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
#define PTLRPC_MD_OPTIONS 0
/**
- * Max # of bulk operations in one request.
+ * log2 max # of bulk operations in one request: 2=4MB/RPC, 5=32MB/RPC, ...
* In order for the client and server to properly negotiate the maximum
* possible transfer size, PTLRPC_BULK_OPS_COUNT must be a power-of-two
* value. The client is free to limit the actual RPC size for any bulk
- * transfer via cl_max_pages_per_rpc to some non-power-of-two value. */
-#define PTLRPC_BULK_OPS_BITS 2
+ * transfer via cl_max_pages_per_rpc to some non-power-of-two value.
+ * NOTE: This is limited to 16 (=64GB RPCs) by IOOBJ_MAX_BRW_BITS. */
+#define PTLRPC_BULK_OPS_BITS 4
+#if PTLRPC_BULK_OPS_BITS > 16
+#error "More than 65536 BRW RPCs not allowed by IOOBJ_MAX_BRW_BITS."
+#endif
#define PTLRPC_BULK_OPS_COUNT (1U << PTLRPC_BULK_OPS_BITS)
/**
* PTLRPC_BULK_OPS_MASK is for the convenience of the client only, and
*/
#define PTLRPC_MAX_BRW_BITS (LNET_MTU_BITS + PTLRPC_BULK_OPS_BITS)
#define PTLRPC_MAX_BRW_SIZE (1 << PTLRPC_MAX_BRW_BITS)
-#define PTLRPC_MAX_BRW_PAGES (PTLRPC_MAX_BRW_SIZE >> PAGE_CACHE_SHIFT)
+#define PTLRPC_MAX_BRW_PAGES (PTLRPC_MAX_BRW_SIZE >> PAGE_SHIFT)
#define ONE_MB_BRW_SIZE (1 << LNET_MTU_BITS)
#define MD_MAX_BRW_SIZE (1 << LNET_MTU_BITS)
-#define MD_MAX_BRW_PAGES (MD_MAX_BRW_SIZE >> PAGE_CACHE_SHIFT)
+#define MD_MAX_BRW_PAGES (MD_MAX_BRW_SIZE >> PAGE_SHIFT)
#define DT_MAX_BRW_SIZE PTLRPC_MAX_BRW_SIZE
-#define DT_MAX_BRW_PAGES (DT_MAX_BRW_SIZE >> PAGE_CACHE_SHIFT)
+#define DT_MAX_BRW_PAGES (DT_MAX_BRW_SIZE >> PAGE_SHIFT)
#define OFD_MAX_BRW_SIZE (1 << LNET_MTU_BITS)
/* When PAGE_SIZE is a constant, we can check our arithmetic here with cpp! */
#if ((PTLRPC_MAX_BRW_PAGES & (PTLRPC_MAX_BRW_PAGES - 1)) != 0)
# error "PTLRPC_MAX_BRW_PAGES isn't a power of two"
#endif
-#if (PTLRPC_MAX_BRW_SIZE != (PTLRPC_MAX_BRW_PAGES * PAGE_CACHE_SIZE))
-# error "PTLRPC_MAX_BRW_SIZE isn't PTLRPC_MAX_BRW_PAGES * PAGE_CACHE_SIZE"
+#if (PTLRPC_MAX_BRW_SIZE != (PTLRPC_MAX_BRW_PAGES * PAGE_SIZE))
+# error "PTLRPC_MAX_BRW_SIZE isn't PTLRPC_MAX_BRW_PAGES * PAGE_SIZE"
#endif
#if (PTLRPC_MAX_BRW_SIZE > LNET_MTU * PTLRPC_BULK_OPS_COUNT)
# error "PTLRPC_MAX_BRW_SIZE too big"
*/
/* depress threads factor for VM with small memory size */
#define OSS_THR_FACTOR min_t(int, 8, \
- NUM_CACHEPAGES >> (28 - PAGE_CACHE_SHIFT))
+ NUM_CACHEPAGES >> (28 - PAGE_SHIFT))
#define OSS_NTHRS_INIT (PTLRPC_NTHRS_INIT + 1)
#define OSS_NTHRS_BASE 64
-#define OSS_NTHRS_MAX 512
/* threads for handling "create" request */
#define OSS_CR_THR_FACTOR 1
set_producer_func set_producer;
/** opaq argument passed to the producer callback */
void *set_producer_arg;
+ unsigned int set_allow_intr:1;
};
/**
unsigned long rs_committed:1;/* the transaction was committed
and the rs was dispatched
by ptlrpc_commit_replies */
+ atomic_t rs_refcount; /* number of users */
+ /** Number of locks awaiting client ACK */
+ int rs_nlocks;
+
/** Size of the state */
int rs_size;
/** opcode */
struct obd_export *rs_export;
struct ptlrpc_service_part *rs_svcpt;
/** Lnet metadata handle for the reply */
- lnet_handle_md_t rs_md_h;
- atomic_t rs_refcount;
+ lnet_handle_md_t rs_md_h;
/** Context for the sevice thread */
- struct ptlrpc_svc_ctx *rs_svc_ctx;
+ struct ptlrpc_svc_ctx *rs_svc_ctx;
/** Reply buffer (actually sent to the client), encoded if needed */
- struct lustre_msg *rs_repbuf; /* wrapper */
- /** Size of the reply buffer */
- int rs_repbuf_len; /* wrapper buf length */
- /** Size of the reply message */
- int rs_repdata_len; /* wrapper msg length */
- /**
- * Actual reply message. Its content is encrupted (if needed) to
- * produce reply buffer for actual sending. In simple case
- * of no network encryption we jus set \a rs_repbuf to \a rs_msg
- */
- struct lustre_msg *rs_msg; /* reply message */
+ struct lustre_msg *rs_repbuf; /* wrapper */
+ /** Size of the reply buffer */
+ int rs_repbuf_len; /* wrapper buf length */
+ /** Size of the reply message */
+ int rs_repdata_len; /* wrapper msg length */
+ /**
+ * Actual reply message. Its content is encrupted (if needed) to
+ * produce reply buffer for actual sending. In simple case
+ * of no network encryption we jus set \a rs_repbuf to \a rs_msg
+ */
+ struct lustre_msg *rs_msg; /* reply message */
- /** Number of locks awaiting client ACK */
- int rs_nlocks;
/** Handles of locks awaiting client reply ACK */
struct lustre_handle rs_locks[RS_MAX_LOCKS];
/** Lock modes of locks in \a rs_locks */
/** RPC stages */
enum rq_phase {
- RQ_PHASE_NEW = 0xebc0de00,
- RQ_PHASE_RPC = 0xebc0de01,
- RQ_PHASE_BULK = 0xebc0de02,
- RQ_PHASE_INTERPRET = 0xebc0de03,
- RQ_PHASE_COMPLETE = 0xebc0de04,
- RQ_PHASE_UNREGISTERING = 0xebc0de05,
- RQ_PHASE_UNDEFINED = 0xebc0de06
+ RQ_PHASE_NEW = 0xebc0de00,
+ RQ_PHASE_RPC = 0xebc0de01,
+ RQ_PHASE_BULK = 0xebc0de02,
+ RQ_PHASE_INTERPRET = 0xebc0de03,
+ RQ_PHASE_COMPLETE = 0xebc0de04,
+ RQ_PHASE_UNREG_RPC = 0xebc0de05,
+ RQ_PHASE_UNREG_BULK = 0xebc0de06,
+ RQ_PHASE_UNDEFINED = 0xebc0de07
};
/** Type of request interpreter call-back */
time_t cr_reply_deadline;
/** when req bulk unlink must finish. */
time_t cr_bulk_deadline;
+ /** when req unlink must finish. */
+ time_t cr_req_deadline;
/** Portal to which this request would be sent */
short cr_req_ptl;
/** Portal where to wait for reply and where reply would be sent */
#define rq_real_sent rq_cli.cr_sent_out
#define rq_reply_deadline rq_cli.cr_reply_deadline
#define rq_bulk_deadline rq_cli.cr_bulk_deadline
+#define rq_req_deadline rq_cli.cr_req_deadline
#define rq_nr_resend rq_cli.cr_resend_nr
#define rq_request_portal rq_cli.cr_req_ptl
#define rq_reply_portal rq_cli.cr_rep_ptl
struct ptlrpc_svc_ctx *sr_svc_ctx;
/** (server side), pointed directly into req buffer */
struct ptlrpc_user_desc *sr_user_desc;
- /** separated reply state */
+ /** separated reply state, may be vmalloc'd */
struct ptlrpc_reply_state *sr_reply_state;
/** server-side hp handlers */
struct ptlrpc_hpreq_ops *sr_ops;
* status */
rq_allow_replay:1,
/* bulk request, sent to server, but uncommitted */
- rq_unstable:1;
+ rq_unstable:1,
+ rq_allow_intr:1;
/** @} */
/** server-side flags @{ */
/** bulk match bits */
__u64 rq_mbits;
/**
- * List item to for replay list. Not yet commited requests get linked
+ * List item to for replay list. Not yet committed requests get linked
* there.
* Also see \a rq_replay comment above.
* It's also link chain on obd_export::exp_req_replay_queue
rq_bulk_write:1, /* request bulk write */
/* server authentication flags */
rq_auth_gss:1, /* authenticated by gss */
- rq_auth_remote:1, /* authed as remote user */
rq_auth_usr_root:1, /* authed as root */
rq_auth_usr_mdt:1, /* authed as mdt */
rq_auth_usr_ost:1, /* authed as ost */
/** various buffer pointers */
- struct lustre_msg *rq_reqbuf; /**< req wrapper */
- char *rq_repbuf; /**< rep buffer */
- struct lustre_msg *rq_repdata; /**< rep wrapper msg */
+ struct lustre_msg *rq_reqbuf; /**< req wrapper, vmalloc*/
+ char *rq_repbuf; /**< rep buffer, vmalloc */
+ struct lustre_msg *rq_repdata; /**< rep wrapper msg */
/** only in priv mode */
struct lustre_msg *rq_clrbuf;
int rq_reqbuf_len; /* req wrapper buf len */
static inline const char *
ptlrpc_phase2str(enum rq_phase phase)
{
- switch (phase) {
- case RQ_PHASE_NEW:
- return "New";
- case RQ_PHASE_RPC:
- return "Rpc";
- case RQ_PHASE_BULK:
- return "Bulk";
- case RQ_PHASE_INTERPRET:
- return "Interpret";
- case RQ_PHASE_COMPLETE:
- return "Complete";
- case RQ_PHASE_UNREGISTERING:
- return "Unregistering";
- default:
- return "?Phase?";
- }
+ switch (phase) {
+ case RQ_PHASE_NEW:
+ return "New";
+ case RQ_PHASE_RPC:
+ return "Rpc";
+ case RQ_PHASE_BULK:
+ return "Bulk";
+ case RQ_PHASE_INTERPRET:
+ return "Interpret";
+ case RQ_PHASE_COMPLETE:
+ return "Complete";
+ case RQ_PHASE_UNREG_RPC:
+ return "UnregRPC";
+ case RQ_PHASE_UNREG_BULK:
+ return "UnregBULK";
+ default:
+ return "?Phase?";
+ }
}
/**
#define FLAG(field, str) (field ? str : "")
/** Convert bit flags into a string */
-#define DEBUG_REQ_FLAGS(req) \
- ptlrpc_rqphase2str(req), \
- FLAG(req->rq_intr, "I"), FLAG(req->rq_replied, "R"), \
- FLAG(req->rq_err, "E"), \
- FLAG(req->rq_timedout, "X") /* eXpired */, FLAG(req->rq_resend, "S"), \
- FLAG(req->rq_restart, "T"), FLAG(req->rq_replay, "P"), \
- FLAG(req->rq_no_resend, "N"), \
- FLAG(req->rq_waiting, "W"), \
- FLAG(req->rq_wait_ctx, "C"), FLAG(req->rq_hp, "H"), \
- FLAG(req->rq_committed, "M")
-
-#define REQ_FLAGS_FMT "%s:%s%s%s%s%s%s%s%s%s%s%s%s"
+#define DEBUG_REQ_FLAGS(req) \
+ ptlrpc_rqphase2str(req), \
+ FLAG(req->rq_intr, "I"), FLAG(req->rq_replied, "R"), \
+ FLAG(req->rq_err, "E"), FLAG(req->rq_net_err, "e"), \
+ FLAG(req->rq_timedout, "X") /* eXpired */, FLAG(req->rq_resend, "S"), \
+ FLAG(req->rq_restart, "T"), FLAG(req->rq_replay, "P"), \
+ FLAG(req->rq_no_resend, "N"), \
+ FLAG(req->rq_waiting, "W"), \
+ FLAG(req->rq_wait_ctx, "C"), FLAG(req->rq_hp, "H"), \
+ FLAG(req->rq_committed, "M")
+
+#define REQ_FLAGS_FMT "%s:%s%s%s%s%s%s%s%s%s%s%s%s%s"
void _debug_req(struct ptlrpc_request *req,
struct libcfs_debug_msg_data *data, const char *fmt, ...)
#define DEBUG_REQ(level, req, fmt, args...) \
do { \
if ((level) & (D_ERROR | D_WARNING)) { \
- static cfs_debug_limit_state_t cdls; \
+ static struct cfs_debug_limit_state cdls; \
LIBCFS_DEBUG_MSG_DATA_DECL(msgdata, level, &cdls); \
debug_req(&msgdata, level, &cdls, req, "@@@ "fmt" ", ## args);\
} else { \
* encrypt iov, size is either 0 or bd_iov_count.
*/
lnet_kiov_t *bd_enc_vec;
- lnet_kiov_t bd_vec[0];
+ lnet_kiov_t *bd_vec;
} bd_kiov;
struct {
struct kvec *bd_enc_kvec;
- struct kvec bd_kvec[0];
+ struct kvec *bd_kvec;
} bd_kvec;
} bd_u;
* Definition of server service thread structure
*/
struct ptlrpc_thread {
- /**
- * List of active threads in svc->srv_threads
- */
+ /**
+ * List of active threads in svc->srv_threads
+ */
struct list_head t_link;
- /**
- * thread-private data (preallocated memory)
- */
- void *t_data;
- __u32 t_flags;
- /**
- * service thread index, from ptlrpc_start_threads
- */
- unsigned int t_id;
- /**
- * service thread pid
- */
+ /**
+ * thread-private data (preallocated vmalloc'd memory)
+ */
+ void *t_data;
+ __u32 t_flags;
+ /**
+ * service thread index, from ptlrpc_start_threads
+ */
+ unsigned int t_id;
+ /**
+ * service thread pid
+ */
pid_t t_pid;
- /**
- * put watchdog in the structure per thread b=14840
- */
- struct lc_watchdog *t_watchdog;
- /**
- * the svc this thread belonged to b=18582
- */
+ /**
+ * put watchdog in the structure per thread b=14840
+ */
+ struct lc_watchdog *t_watchdog;
+ /**
+ * the svc this thread belonged to b=18582
+ */
struct ptlrpc_service_part *t_svcpt;
wait_queue_head_t t_ctl_waitq;
struct lu_env *t_env;
*/
char pc_name[16];
/**
- * Environment for request interpreters to run in.
- */
- struct lu_env pc_env;
- /**
* CPT the thread is bound on.
*/
int pc_cpt;
static inline int ptlrpc_client_bulk_active(struct ptlrpc_request *req)
{
struct ptlrpc_bulk_desc *desc;
- int rc;
+ int rc;
- LASSERT(req != NULL);
+ LASSERT(req != NULL);
desc = req->rq_bulk;
- if (OBD_FAIL_CHECK(OBD_FAIL_PTLRPC_LONG_BULK_UNLINK) &&
- req->rq_bulk_deadline > cfs_time_current_sec())
- return 1;
+ if (req->rq_bulk_deadline > cfs_time_current_sec())
+ return 1;
- if (!desc)
- return 0;
+ if (!desc)
+ return 0;
spin_lock(&desc->bd_lock);
rc = desc->bd_md_count;
void ptlrpc_init_client(int req_portal, int rep_portal, char *name,
struct ptlrpc_client *);
void ptlrpc_cleanup_client(struct obd_import *imp);
-struct ptlrpc_connection *ptlrpc_uuid_to_connection(struct obd_uuid *uuid);
+struct ptlrpc_connection *ptlrpc_uuid_to_connection(struct obd_uuid *uuid,
+ lnet_nid_t nid4refnet);
int ptlrpc_queue_wait(struct ptlrpc_request *req);
int ptlrpc_replay_req(struct ptlrpc_request *req);
int ptlrpc_request_bufs_pack(struct ptlrpc_request *request,
__u32 version, int opcode, char **bufs,
struct ptlrpc_cli_ctx *ctx);
-struct ptlrpc_request *ptlrpc_prep_req(struct obd_import *imp, __u32 version,
- int opcode, int count, __u32 *lengths,
- char **bufs);
-struct ptlrpc_request *ptlrpc_prep_req_pool(struct obd_import *imp,
- __u32 version, int opcode,
- int count, __u32 *lengths, char **bufs,
- struct ptlrpc_request_pool *pool);
void ptlrpc_req_finished(struct ptlrpc_request *request);
void ptlrpc_req_finished_with_imp_lock(struct ptlrpc_request *request);
struct ptlrpc_request *ptlrpc_request_addref(struct ptlrpc_request *req);
int i;
for (i = 0; i < desc->bd_iov_count ; i++)
- page_cache_release(BD_GET_KIOV(desc, i).kiov_page);
+ put_page(BD_GET_KIOV(desc, i).kiov_page);
}
static inline void ptlrpc_release_bulk_noop(struct ptlrpc_bulk_desc *desc)
if (req->rq_phase == new_phase)
return;
- if (new_phase == RQ_PHASE_UNREGISTERING) {
+ if (new_phase == RQ_PHASE_UNREG_RPC ||
+ new_phase == RQ_PHASE_UNREG_BULK) {
+ /* No embedded unregistering phases */
+ if (req->rq_phase == RQ_PHASE_UNREG_RPC ||
+ req->rq_phase == RQ_PHASE_UNREG_BULK)
+ return;
+
req->rq_next_phase = req->rq_phase;
if (req->rq_import)
atomic_inc(&req->rq_import->imp_unregistering);
}
- if (req->rq_phase == RQ_PHASE_UNREGISTERING) {
+ if (req->rq_phase == RQ_PHASE_UNREG_RPC ||
+ req->rq_phase == RQ_PHASE_UNREG_BULK) {
if (req->rq_import)
atomic_dec(&req->rq_import->imp_unregistering);
}
static inline int
ptlrpc_client_early(struct ptlrpc_request *req)
{
- if (OBD_FAIL_CHECK(OBD_FAIL_PTLRPC_LONG_REPL_UNLINK) &&
- req->rq_reply_deadline > cfs_time_current_sec())
- return 0;
return req->rq_early;
}
static inline int
ptlrpc_client_replied(struct ptlrpc_request *req)
{
- if (OBD_FAIL_CHECK(OBD_FAIL_PTLRPC_LONG_REPL_UNLINK) &&
- req->rq_reply_deadline > cfs_time_current_sec())
- return 0;
- return req->rq_replied;
+ if (req->rq_reply_deadline > cfs_time_current_sec())
+ return 0;
+ return req->rq_replied;
}
/** Returns true if request \a req is in process of receiving server reply */
static inline int
ptlrpc_client_recv(struct ptlrpc_request *req)
{
- if (OBD_FAIL_CHECK(OBD_FAIL_PTLRPC_LONG_REPL_UNLINK) &&
- req->rq_reply_deadline > cfs_time_current_sec())
- return 1;
- return req->rq_receiving_reply;
+ if (req->rq_reply_deadline > cfs_time_current_sec())
+ return 1;
+ return req->rq_receiving_reply;
}
static inline int
int rc;
spin_lock(&req->rq_lock);
- if (OBD_FAIL_CHECK(OBD_FAIL_PTLRPC_LONG_REPL_UNLINK) &&
- req->rq_reply_deadline > cfs_time_current_sec()) {
+ if (req->rq_reply_deadline > cfs_time_current_sec()) {
+ spin_unlock(&req->rq_lock);
+ return 1;
+ }
+ if (req->rq_req_deadline > cfs_time_current_sec()) {
spin_unlock(&req->rq_lock);
return 1;
}
+
rc = !req->rq_req_unlinked || !req->rq_reply_unlinked ||
req->rq_receiving_reply;
spin_unlock(&req->rq_lock);
git://git.whamcloud.com / fs / lustre-release.git / blobdiff