I would like to see some performance measurements for this patch on a system with fast storage and multiple 10 GbE links.
If not, at least a good analysis of the expected performance impact the patch will have on major architectures.
Tonight I will think about whether the 2038 thing even matters or whether we just need a comment explaining why it's safe.
On May 11, 2015 11:38 AM, Arnd Bergmann arnd@arndb.de wrote:
On Monday 11 May 2015 08:05:05 Tina Ruchandani wrote:
'struct frame' uses two variables to store the sent timestamp - 'struct timeval' and jiffies. jiffies is used to avoid discrepancies caused by updates to system time. 'struct timeval' uses 32-bit representation for seconds which will overflow in year 2038. This patch does the following:
- Replace the use of 'struct timeval' and jiffies with ktime_t, which
is a 64-bit timestamp and is year 2038 safe.
- ktime_t provides both long range (like jiffies) and high resolution
(like timeval). Using ktime_get (monotonic time) instead of wall-clock time prevents any discprepancies caused by updates to system time.
Signed-off-by: Tina Ruchandani ruchandani.tina@gmail.com
Very nice!
@@ -499,32 +497,15 @@ resend(struct aoedev *d, struct frame *f) static int tsince_hr(struct frame *f) {
- struct timeval now;
- ktime_t now;
int n;
- do_gettimeofday(&now);
- n = now.tv_usec - f->sent.tv_usec;
- n += (now.tv_sec - f->sent.tv_sec) * USEC_PER_SEC;
- now = ktime_get();
- n = ktime_to_us(ktime_sub(now, f->sent));
I would cut four extra lines by writing this as
return ktime_us_delta(ktime_get(), f->sent));
but the effect is exactly the same.
With that change, please add
Reviewed-by: Arnd Bergmann arnd@arndb.de
Arnd