SUNRPC: Replace jiffies-based metrics with ktime-based metrics

Currently RPC performance metrics that tabulate elapsed time use
jiffies time values.  This is problematic on systems that use slow
jiffies (for instance 100HZ systems built for paravirtualized
environments).  It is also a problem for computing precise latency
statistics for advanced network transports, such as InfiniBand,
that can have round-trip latencies significanly faster than a single
clock tick.

For the RPC client, adopt the high resolution time stamp mechanism
already used by the network layer and blktrace: ktime.

We use ktime format time stamps for all internal computations, and
convert to milliseconds for presentation.  As a result, we need only
addition operations in the performance critical paths; multiply/divide
is required only for presentation.

We could report RTT metrics in microseconds.  In fact the mountstats
format is versioned to accomodate exactly this kind of interface
improvement.

For now, however, we'll stay with millisecond precision for
presentation to maintain backwards compatibility with the handful of
currently deployed user space tools.  At a later point, we'll move to
an API such as BDI_STATS where a finer timestamp precision can be
reported.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

1 files changed, 9 insertions, 18 deletions

diff --git a/net/sunrpc/stats.c b/net/sunrpc/stats.c
index 5785d2037f45..aacd95f0dce5 100644
--- a/net/sunrpc/stats.c
+++ b/net/sunrpc/stats.c
@@ -144,7 +144,7 @@ void rpc_count_iostats(struct rpc_task *task)
         struct rpc_rqst *req = task->tk_rqstp;
         struct rpc_iostats *stats;
         struct rpc_iostats *op_metrics;
-        long rtt, execute, queue;
+        ktime_t delta;
 
         if (!task->tk_client || !task->tk_client->cl_metrics || !req)
                 return;
@@ -159,20 +159,13 @@ void rpc_count_iostats(struct rpc_task *task)
         op_metrics->om_bytes_sent += task->tk_bytes_sent;
         op_metrics->om_bytes_recv += req->rq_reply_bytes_recvd;
 
-        queue = (long)req->rq_xtime - task->tk_start;
-        if (queue < 0)
-                queue = -queue;
-        op_metrics->om_queue += queue;
+        delta = ktime_sub(req->rq_xtime, task->tk_start);
+        op_metrics->om_queue = ktime_add(op_metrics->om_queue, delta);
 
-        rtt = task->tk_rtt;
-        if (rtt < 0)
-                rtt = -rtt;
-        op_metrics->om_rtt += rtt;
+        op_metrics->om_rtt = ktime_add(op_metrics->om_rtt, task->tk_rtt);
 
-        execute = (long)jiffies - task->tk_start;
-        if (execute < 0)
-                execute = -execute;
-        op_metrics->om_execute += execute;
+        delta = ktime_sub(ktime_get(), task->tk_start);
+        op_metrics->om_execute = ktime_add(op_metrics->om_execute, delta);
 }
 
 static void _print_name(struct seq_file *seq, unsigned int op,
@@ -186,8 +179,6 @@ static void _print_name(struct seq_file *seq, unsigned int op,
                 seq_printf(seq, "\t%12u: ", op);
 }
 
-#define MILLISECS_PER_JIFFY     (1000 / HZ)
-
 void rpc_print_iostats(struct seq_file *seq, struct rpc_clnt *clnt)
 {
         struct rpc_iostats *stats = clnt->cl_metrics;
@@ -214,9 +205,9 @@ void rpc_print_iostats(struct seq_file *seq, struct rpc_clnt *clnt)
                                 metrics->om_timeouts,
                                 metrics->om_bytes_sent,
                                 metrics->om_bytes_recv,
-                                metrics->om_queue * MILLISECS_PER_JIFFY,
-                                metrics->om_rtt * MILLISECS_PER_JIFFY,
-                                metrics->om_execute * MILLISECS_PER_JIFFY);
+                                ktime_to_ms(metrics->om_queue),
+                                ktime_to_ms(metrics->om_rtt),
+                                ktime_to_ms(metrics->om_execute));
         }
 }
 EXPORT_SYMBOL_GPL(rpc_print_iostats);