summaryrefslogtreecommitdiff
path: root/net/rds
diff options
context:
space:
mode:
authorSantosh Shilimkar <santosh.shilimkar@oracle.com>2016-07-05 08:35:15 +0300
committerSantosh Shilimkar <santosh.shilimkar@oracle.com>2017-01-03 01:02:59 +0300
commit3289025aedc018f8fd9d0e37fb9efa0c6d531ffa (patch)
tree251ddb4c6606d91809d98d40a22e18382025664e /net/rds
parentf9fb69adb6c7acca60977a4db5a5f95b8e66c041 (diff)
downloadlinux-3289025aedc018f8fd9d0e37fb9efa0c6d531ffa.tar.xz
RDS: add receive message trace used by application
Socket option to tap receive path latency in various stages in nano seconds. It can be enabled on selective sockets using using SO_RDS_MSG_RXPATH_LATENCY socket option. RDS will return the data to application with RDS_CMSG_RXPATH_LATENCY in defined format. Scope is left to add more trace points for future without need of change in the interface. Reviewed-by: Sowmini Varadhan <sowmini.varadhan@oracle.com> Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Diffstat (limited to 'net/rds')
-rw-r--r--net/rds/af_rds.c28
-rw-r--r--net/rds/ib_recv.c4
-rw-r--r--net/rds/rds.h10
-rw-r--r--net/rds/recv.c32
-rw-r--r--net/rds/tcp_recv.c5
5 files changed, 76 insertions, 3 deletions
diff --git a/net/rds/af_rds.c b/net/rds/af_rds.c
index 2ac1e6194be3..fd8217404162 100644
--- a/net/rds/af_rds.c
+++ b/net/rds/af_rds.c
@@ -298,6 +298,30 @@ static int rds_enable_recvtstamp(struct sock *sk, char __user *optval,
return 0;
}
+static int rds_recv_track_latency(struct rds_sock *rs, char __user *optval,
+ int optlen)
+{
+ struct rds_rx_trace_so trace;
+ int i;
+
+ if (optlen != sizeof(struct rds_rx_trace_so))
+ return -EFAULT;
+
+ if (copy_from_user(&trace, optval, sizeof(trace)))
+ return -EFAULT;
+
+ rs->rs_rx_traces = trace.rx_traces;
+ for (i = 0; i < rs->rs_rx_traces; i++) {
+ if (trace.rx_trace_pos[i] > RDS_MSG_RX_DGRAM_TRACE_MAX) {
+ rs->rs_rx_traces = 0;
+ return -EFAULT;
+ }
+ rs->rs_rx_trace[i] = trace.rx_trace_pos[i];
+ }
+
+ return 0;
+}
+
static int rds_setsockopt(struct socket *sock, int level, int optname,
char __user *optval, unsigned int optlen)
{
@@ -338,6 +362,9 @@ static int rds_setsockopt(struct socket *sock, int level, int optname,
ret = rds_enable_recvtstamp(sock->sk, optval, optlen);
release_sock(sock->sk);
break;
+ case SO_RDS_MSG_RXPATH_LATENCY:
+ ret = rds_recv_track_latency(rs, optval, optlen);
+ break;
default:
ret = -ENOPROTOOPT;
}
@@ -484,6 +511,7 @@ static int __rds_create(struct socket *sock, struct sock *sk, int protocol)
INIT_LIST_HEAD(&rs->rs_cong_list);
spin_lock_init(&rs->rs_rdma_lock);
rs->rs_rdma_keys = RB_ROOT;
+ rs->rs_rx_traces = 0;
spin_lock_bh(&rds_sock_lock);
list_add_tail(&rs->rs_item, &rds_sock_list);
diff --git a/net/rds/ib_recv.c b/net/rds/ib_recv.c
index 4b0f12679219..e10624aa6959 100644
--- a/net/rds/ib_recv.c
+++ b/net/rds/ib_recv.c
@@ -911,8 +911,12 @@ static void rds_ib_process_recv(struct rds_connection *conn,
ic->i_ibinc = ibinc;
hdr = &ibinc->ii_inc.i_hdr;
+ ibinc->ii_inc.i_rx_lat_trace[RDS_MSG_RX_HDR] =
+ local_clock();
memcpy(hdr, ihdr, sizeof(*hdr));
ic->i_recv_data_rem = be32_to_cpu(hdr->h_len);
+ ibinc->ii_inc.i_rx_lat_trace[RDS_MSG_RX_START] =
+ local_clock();
rdsdebug("ic %p ibinc %p rem %u flag 0x%x\n", ic, ibinc,
ic->i_recv_data_rem, hdr->h_flags);
diff --git a/net/rds/rds.h b/net/rds/rds.h
index f713194e4620..07fff73dd4f3 100644
--- a/net/rds/rds.h
+++ b/net/rds/rds.h
@@ -253,6 +253,11 @@ struct rds_ext_header_rdma_dest {
#define RDS_EXTHDR_GEN_NUM 6
#define __RDS_EXTHDR_MAX 16 /* for now */
+#define RDS_RX_MAX_TRACES (RDS_MSG_RX_DGRAM_TRACE_MAX + 1)
+#define RDS_MSG_RX_HDR 0
+#define RDS_MSG_RX_START 1
+#define RDS_MSG_RX_END 2
+#define RDS_MSG_RX_CMSG 3
struct rds_incoming {
atomic_t i_refcount;
@@ -265,6 +270,7 @@ struct rds_incoming {
rds_rdma_cookie_t i_rdma_cookie;
struct timeval i_rx_tstamp;
+ u64 i_rx_lat_trace[RDS_RX_MAX_TRACES];
};
struct rds_mr {
@@ -575,6 +581,10 @@ struct rds_sock {
unsigned char rs_recverr,
rs_cong_monitor;
u32 rs_hash_initval;
+
+ /* Socket receive path trace points*/
+ u8 rs_rx_traces;
+ u8 rs_rx_trace[RDS_MSG_RX_DGRAM_TRACE_MAX];
};
static inline struct rds_sock *rds_sk_to_rs(const struct sock *sk)
diff --git a/net/rds/recv.c b/net/rds/recv.c
index ba19eeeae85a..8b7e7b7f2c2d 100644
--- a/net/rds/recv.c
+++ b/net/rds/recv.c
@@ -43,6 +43,8 @@
void rds_inc_init(struct rds_incoming *inc, struct rds_connection *conn,
__be32 saddr)
{
+ int i;
+
atomic_set(&inc->i_refcount, 1);
INIT_LIST_HEAD(&inc->i_item);
inc->i_conn = conn;
@@ -50,6 +52,9 @@ void rds_inc_init(struct rds_incoming *inc, struct rds_connection *conn,
inc->i_rdma_cookie = 0;
inc->i_rx_tstamp.tv_sec = 0;
inc->i_rx_tstamp.tv_usec = 0;
+
+ for (i = 0; i < RDS_RX_MAX_TRACES; i++)
+ inc->i_rx_lat_trace[i] = 0;
}
EXPORT_SYMBOL_GPL(rds_inc_init);
@@ -373,6 +378,7 @@ void rds_recv_incoming(struct rds_connection *conn, __be32 saddr, __be32 daddr,
if (sock_flag(sk, SOCK_RCVTSTAMP))
do_gettimeofday(&inc->i_rx_tstamp);
rds_inc_addref(inc);
+ inc->i_rx_lat_trace[RDS_MSG_RX_END] = local_clock();
list_add_tail(&inc->i_item, &rs->rs_recv_queue);
__rds_wake_sk_sleep(sk);
} else {
@@ -534,7 +540,7 @@ static int rds_cmsg_recv(struct rds_incoming *inc, struct msghdr *msg,
ret = put_cmsg(msg, SOL_RDS, RDS_CMSG_RDMA_DEST,
sizeof(inc->i_rdma_cookie), &inc->i_rdma_cookie);
if (ret)
- return ret;
+ goto out;
}
if ((inc->i_rx_tstamp.tv_sec != 0) &&
@@ -543,10 +549,30 @@ static int rds_cmsg_recv(struct rds_incoming *inc, struct msghdr *msg,
sizeof(struct timeval),
&inc->i_rx_tstamp);
if (ret)
- return ret;
+ goto out;
}
- return 0;
+ if (rs->rs_rx_traces) {
+ struct rds_cmsg_rx_trace t;
+ int i, j;
+
+ inc->i_rx_lat_trace[RDS_MSG_RX_CMSG] = local_clock();
+ t.rx_traces = rs->rs_rx_traces;
+ for (i = 0; i < rs->rs_rx_traces; i++) {
+ j = rs->rs_rx_trace[i];
+ t.rx_trace_pos[i] = j;
+ t.rx_trace[i] = inc->i_rx_lat_trace[j + 1] -
+ inc->i_rx_lat_trace[j];
+ }
+
+ ret = put_cmsg(msg, SOL_RDS, RDS_CMSG_RXPATH_LATENCY,
+ sizeof(t), &t);
+ if (ret)
+ goto out;
+ }
+
+out:
+ return ret;
}
int rds_recvmsg(struct socket *sock, struct msghdr *msg, size_t size,
diff --git a/net/rds/tcp_recv.c b/net/rds/tcp_recv.c
index ad4892e97f91..e006ef8e6d40 100644
--- a/net/rds/tcp_recv.c
+++ b/net/rds/tcp_recv.c
@@ -180,6 +180,9 @@ static int rds_tcp_data_recv(read_descriptor_t *desc, struct sk_buff *skb,
rdsdebug("alloced tinc %p\n", tinc);
rds_inc_path_init(&tinc->ti_inc, cp,
cp->cp_conn->c_faddr);
+ tinc->ti_inc.i_rx_lat_trace[RDS_MSG_RX_HDR] =
+ local_clock();
+
/*
* XXX * we might be able to use the __ variants when
* we've already serialized at a higher level.
@@ -204,6 +207,8 @@ static int rds_tcp_data_recv(read_descriptor_t *desc, struct sk_buff *skb,
/* could be 0 for a 0 len message */
tc->t_tinc_data_rem =
be32_to_cpu(tinc->ti_inc.i_hdr.h_len);
+ tinc->ti_inc.i_rx_lat_trace[RDS_MSG_RX_START] =
+ local_clock();
}
}