On 02/11/17 14:32, Arnd Bergmann wrote:
This driver is the only one using the deprecated timeval_to_ns()
helper. Changing it from do_gettimeofday() to ktime_get() makes
the code more efficient, more robust against concurrent
settimeofday(), more accurate and lets us get rid of that helper
in the future.

Signed-off-by: Arnd Bergmann <a...@arndb.de>
---
  drivers/staging/greybus/loopback.c | 42 ++++++++++++++++----------------------
  1 file changed, 18 insertions(+), 24 deletions(-)

diff --git a/drivers/staging/greybus/loopback.c 
b/drivers/staging/greybus/loopback.c
index 85046fb16aad..3d92638c424b 100644
--- a/drivers/staging/greybus/loopback.c
+++ b/drivers/staging/greybus/loopback.c
@@ -58,7 +58,7 @@ static struct gb_loopback_device gb_dev;
  struct gb_loopback_async_operation {
        struct gb_loopback *gb;
        struct gb_operation *operation;
-       struct timeval ts;
+       ktime_t ts;
        struct timer_list timer;
        struct list_head entry;
        struct work_struct work;
@@ -81,7 +81,7 @@ struct gb_loopback {
        atomic_t outstanding_operations;
/* Per connection stats */
-       struct timeval ts;
+       ktime_t ts;
        struct gb_loopback_stats latency;
        struct gb_loopback_stats throughput;
        struct gb_loopback_stats requests_per_second;
@@ -375,14 +375,9 @@ static u64 __gb_loopback_calc_latency(u64 t1, u64 t2)
                return NSEC_PER_DAY - t2 + t1;
  }
-static u64 gb_loopback_calc_latency(struct timeval *ts, struct timeval *te)
+static u64 gb_loopback_calc_latency(ktime_t ts, ktime_t te)
  {
-       u64 t1, t2;
-
-       t1 = timeval_to_ns(ts);
-       t2 = timeval_to_ns(te);
-
-       return __gb_loopback_calc_latency(t1, t2);
+       return __gb_loopback_calc_latency(ktime_to_ns(ts), ktime_to_ns(te));
  }
static int gb_loopback_operation_sync(struct gb_loopback *gb, int type,
@@ -390,10 +385,10 @@ static int gb_loopback_operation_sync(struct gb_loopback 
*gb, int type,
                                      void *response, int response_size)
  {
        struct gb_operation *operation;
-       struct timeval ts, te;
+       ktime_t ts, te;
        int ret;
- do_gettimeofday(&ts);
+       ts = ktime_get();
        operation = gb_operation_create(gb->connection, type, request_size,
                                        response_size, GFP_KERNEL);
        if (!operation)
@@ -421,10 +416,10 @@ static int gb_loopback_operation_sync(struct gb_loopback 
*gb, int type,
                }
        }
- do_gettimeofday(&te);
+       te = ktime_get();
/* Calculate the total time the message took */
-       gb->elapsed_nsecs = gb_loopback_calc_latency(&ts, &te);
+       gb->elapsed_nsecs = gb_loopback_calc_latency(ts, te);
out_put_operation:
        gb_operation_put(operation);
@@ -492,10 +487,10 @@ static void gb_loopback_async_operation_callback(struct 
gb_operation *operation)
  {
        struct gb_loopback_async_operation *op_async;
        struct gb_loopback *gb;
-       struct timeval te;
+       ktime_t te;
        bool err = false;
- do_gettimeofday(&te);
+       te = ktime_get();
        op_async = gb_loopback_operation_find(operation->id);
        if (!op_async)
                return;
@@ -512,8 +507,7 @@ static void gb_loopback_async_operation_callback(struct 
gb_operation *operation)
        }
if (!err)
-               gb->elapsed_nsecs = gb_loopback_calc_latency(&op_async->ts,
-                                                            &te);
+               gb->elapsed_nsecs = gb_loopback_calc_latency(op_async->ts, te);
if (op_async->pending) {
                if (err)
@@ -608,7 +602,7 @@ static int gb_loopback_async_operation(struct gb_loopback 
*gb, int type,
        list_add_tail(&op_async->entry, &gb_dev.list_op_async);
        spin_unlock_irqrestore(&gb_dev.lock, flags);
- do_gettimeofday(&op_async->ts);
+       op_async->ts = ktime_get();
        op_async->pending = true;
        atomic_inc(&gb->outstanding_operations);
        mutex_lock(&gb->mutex);
@@ -842,7 +836,7 @@ static void gb_loopback_reset_stats(struct gb_loopback *gb)
        /* Should be initialized at least once per transaction set */
        gb->apbridge_latency_ts = 0;
        gb->gbphy_latency_ts = 0;
-       memset(&gb->ts, 0, sizeof(struct timeval));
+       gb->ts = ktime_set(0, 0);
  }
static void gb_loopback_update_stats(struct gb_loopback_stats *stats, u32 val)
@@ -925,15 +919,15 @@ static void gb_loopback_calculate_stats(struct 
gb_loopback *gb, bool error)
  {
        u64 nlat;
        u32 lat;
-       struct timeval te;
+       ktime_t te;
if (!error) {
                gb->requests_completed++;
                gb_loopback_calculate_latency_stats(gb);
        }
- do_gettimeofday(&te);
-       nlat = gb_loopback_calc_latency(&gb->ts, &te);
+       te = ktime_get();
+       nlat = gb_loopback_calc_latency(gb->ts, te);
        if (nlat >= NSEC_PER_SEC || gb->iteration_count == gb->iteration_max) {
                lat = gb_loopback_nsec_to_usec_latency(nlat);
@@ -1017,8 +1011,8 @@ static int gb_loopback_fn(void *data)
                size = gb->size;
                us_wait = gb->us_wait;
                type = gb->type;
-               if (gb->ts.tv_usec == 0 && gb->ts.tv_sec == 0)
-                       do_gettimeofday(&gb->ts);
+               if (ktime_to_ns(gb->ts) == 0)
+                       gb->ts = ktime_get();
                mutex_unlock(&gb->mutex);
/* Else operations to perform */


Fine with this change too.

Acked-by: Bryan O'Donoghue <pure.lo...@nexus-software.ie>

Reply via email to