add several new RPC statistics counters to help sysadmins more easily
debug client-side NFS performance and reliability problems. make the RPC
counters longs instead of ints to take advantage of 64-bit counters on
platforms that can support them.
it adds some new counter values in /proc/net/rpc/nfs, so it should be
applied before the feature freeze deadline. please apply this to your
tree.
diff -drN -U3 00-stock/include/linux/sunrpc/stats.h 01-counters/include/linux/sunrpc/stats.h
--- 00-stock/include/linux/sunrpc/stats.h Tue Oct 15 23:28:20 2002
+++ 01-counters/include/linux/sunrpc/stats.h Sat Oct 19 15:48:23 2002
@@ -1,7 +1,7 @@
/*
* linux/include/linux/sunrpc/stats.h
*
- * Client statistics collection for SUN RPC
+ * Statistics collection for SUN RPC
*
* Copyright (C) 1996 Olaf Kirch <[email protected]>
*/
@@ -11,33 +11,40 @@
#include <linux/config.h>
#include <linux/proc_fs.h>
+#include <linux/cache.h>
struct rpc_stat {
struct rpc_program * program;
- unsigned int netcnt,
+ unsigned long netcnt,
netudpcnt,
nettcpcnt,
nettcpconn,
netreconn;
- unsigned int rpccnt,
+ unsigned long rpccnt,
rpcretrans,
rpcauthrefresh,
- rpcgarbage;
-};
+ rpcgarbage,
+ rpcnospace,
+ rpcbadxids,
+ rpcbadverfs,
+ rpccantconn,
+ rpcnomem,
+ rpccantsend;
+} ____cacheline_aligned;
struct svc_stat {
struct svc_program * program;
- unsigned int netcnt,
+ unsigned long netcnt,
netudpcnt,
nettcpcnt,
nettcpconn;
- unsigned int rpccnt,
+ unsigned long rpccnt,
rpcbadfmt,
rpcbadauth,
rpcbadclnt;
-};
+} ____cacheline_aligned;
void rpc_proc_init(void);
void rpc_proc_exit(void);
diff -drN -U3 00-stock/include/linux/sunrpc/xprt.h 01-counters/include/linux/sunrpc/xprt.h
--- 00-stock/include/linux/sunrpc/xprt.h Tue Oct 15 23:28:32 2002
+++ 01-counters/include/linux/sunrpc/xprt.h Sat Oct 19 15:54:19 2002
@@ -146,6 +146,7 @@
unsigned long sockstate; /* Socket state */
unsigned char shutdown : 1, /* being shut down */
nocong : 1, /* no congestion control */
+ neverconn : 1, /* no connection attempt yet */
resvport : 1, /* use a reserved port */
stream : 1; /* TCP */
diff -drN -U3 00-stock/net/sunrpc/clnt.c 01-counters/net/sunrpc/clnt.c
--- 00-stock/net/sunrpc/clnt.c Tue Oct 15 23:27:53 2002
+++ 01-counters/net/sunrpc/clnt.c Sat Oct 19 15:48:23 2002
@@ -453,6 +453,8 @@
xprt_release(task);
}
+ task->tk_client->cl_stats->rpcnomem++;
+
switch (status) {
case -EAGAIN: /* woken up; retry */
task->tk_action = call_reserve;
@@ -494,6 +496,7 @@
if (RPC_IS_ASYNC(task) || !(task->tk_client->cl_intr && signalled())) {
xprt_release(task);
task->tk_action = call_reserve;
+ task->tk_client->cl_stats->rpcnomem++;
rpc_delay(task, HZ>>4);
return;
}
@@ -669,6 +672,7 @@
task->tk_action = call_bind;
break;
case -EAGAIN:
+ clnt->cl_stats->rpcnospace++;
task->tk_action = call_transmit;
break;
case -EIO:
@@ -757,14 +761,16 @@
}
/* Verify the RPC header */
- if (!(p = call_verify(task)))
+ if (!(p = call_verify(task))) {
+ clnt->cl_stats->rpcbadverfs++;
return;
+ }
/*
* The following is an NFS-specific hack to cater for setuid
* processes whose uid is mapped to nobody on the server.
*/
- if (task->tk_client->cl_droppriv &&
+ if (clnt->cl_droppriv &&
(ntohl(*p) == NFSERR_ACCES || ntohl(*p) == NFSERR_PERM)) {
if (RPC_IS_SETUID(task) && task->tk_suid_retry) {
dprintk("RPC: %4d retry squashed uid\n", task->tk_pid);
diff -drN -U3 00-stock/net/sunrpc/stats.c 01-counters/net/sunrpc/stats.c
--- 00-stock/net/sunrpc/stats.c Tue Oct 15 23:29:07 2002
+++ 01-counters/net/sunrpc/stats.c Sat Oct 19 15:48:23 2002
@@ -40,16 +40,23 @@
int len, i, j;
len = sprintf(buffer,
- "net %d %d %d %d\n",
+ "net %lu %lu %lu %lu\n",
statp->netcnt,
statp->netudpcnt,
statp->nettcpcnt,
statp->nettcpconn);
len += sprintf(buffer + len,
- "rpc %d %d %d\n",
+ "rpc %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu\n",
statp->rpccnt,
statp->rpcretrans,
- statp->rpcauthrefresh);
+ statp->rpcauthrefresh,
+ statp->rpcgarbage,
+ statp->rpcnospace,
+ statp->rpcbadxids,
+ statp->rpcbadverfs,
+ statp->rpccantconn,
+ statp->rpcnomem,
+ statp->rpccantsend);
for (i = 0; i < prog->nrvers; i++) {
if (!(vers = prog->version[i]))
@@ -88,13 +95,13 @@
int len, i, j;
len = sprintf(buffer,
- "net %d %d %d %d\n",
+ "net %lu %lu %lu %lu\n",
statp->netcnt,
statp->netudpcnt,
statp->nettcpcnt,
statp->nettcpconn);
len += sprintf(buffer + len,
- "rpc %d %d %d %d %d\n",
+ "rpc %lu %lu %lu %lu %lu\n",
statp->rpccnt,
statp->rpcbadfmt+statp->rpcbadauth+statp->rpcbadclnt,
statp->rpcbadfmt,
diff -drN -U3 00-stock/net/sunrpc/xprt.c 01-counters/net/sunrpc/xprt.c
--- 00-stock/net/sunrpc/xprt.c Tue Oct 15 23:28:29 2002
+++ 01-counters/net/sunrpc/xprt.c Sat Oct 19 16:21:26 2002
@@ -213,6 +213,7 @@
static inline int
xprt_sendmsg(struct rpc_xprt *xprt, struct rpc_rqst *req)
{
+ struct rpc_stat *stats = req->rq_task->tk_client->cl_stats;
struct socket *sock = xprt->sock;
struct msghdr msg;
struct xdr_buf *xdr = &req->rq_snd_buf;
@@ -224,6 +225,12 @@
if (!sock)
return -ENOTCONN;
+ stats->netcnt++;
+ if (xprt->stream)
+ stats->nettcpcnt++;
+ else
+ stats->netudpcnt++;
+
xprt_pktdump("packet data:",
req->rq_svec->iov_base,
req->rq_svec->iov_len);
@@ -258,16 +265,21 @@
/* When the server has died, an ICMP port unreachable message
* prompts ECONNREFUSED.
*/
+ stats->rpccantsend++;
+ break;
case -EAGAIN:
break;
case -ENOTCONN:
case -EPIPE:
/* connection broken */
+ stats->rpccantsend++;
if (xprt->stream)
result = -ENOTCONN;
break;
default:
+ stats->rpccantsend++;
printk(KERN_NOTICE "RPC: sendmsg returned error %d\n", -result);
+ break;
}
return result;
}
@@ -416,6 +428,7 @@
void
xprt_connect(struct rpc_task *task)
{
+ struct rpc_stat *stats = task->tk_client->cl_stats;
struct rpc_xprt *xprt = task->tk_xprt;
struct socket *sock = xprt->sock;
struct sock *inet;
@@ -487,6 +500,11 @@
if (inet->state != TCP_ESTABLISHED) {
xprt_close(xprt);
task->tk_status = -EAGAIN;
+ stats->nettcpconn++;
+ if (!xprt->neverconn) {
+ stats->netreconn++;
+ xprt->neverconn = 0;
+ }
goto out_write;
}
@@ -497,12 +515,14 @@
case -EPIPE:
xprt_close(xprt);
task->tk_status = -ENOTCONN;
+ stats->rpccantconn++;
goto out_write;
default:
/* Report myriad other possible returns. If this file
* system is soft mounted, just error out, like Solaris. */
xprt_close(xprt);
+ stats->rpccantconn++;
if (task->tk_client->cl_softrtry) {
printk(KERN_WARNING
"RPC: error %d connecting to server %s, exiting\n",
@@ -528,12 +548,18 @@
static void
xprt_conn_status(struct rpc_task *task)
{
+ struct rpc_stat *stats = task->tk_client->cl_stats;
struct rpc_xprt *xprt = task->tk_xprt;
switch (task->tk_status) {
case 0:
dprintk("RPC: %4d xprt_conn_status: connection established\n",
task->tk_pid);
+ stats->nettcpconn++;
+ if (!xprt->neverconn) {
+ stats->netreconn++;
+ xprt->neverconn = 0;
+ }
goto out;
case -ETIMEDOUT:
dprintk("RPC: %4d xprt_conn_status: timed out\n",
@@ -548,6 +574,8 @@
rpc_delay(task, RPC_REESTABLISH_TIMEOUT);
break;
}
+ stats->rpccantconn++;
+
/* if soft mounted, cause this RPC to fail */
if (task->tk_client->cl_softrtry)
task->tk_status = -EIO;
@@ -1374,6 +1402,7 @@
if (xprt->stream) {
xprt->cwnd = RPC_MAXCWND;
xprt->nocong = 1;
+ xprt->neverconn = 1;
} else
xprt->cwnd = RPC_INITCWND;
spin_lock_init(&xprt->sock_lock);