Eric Dumazet 47d3a07528 net/mlx4_en: fix overflow in mlx4_en_init_timestamp()
The cited commit makes a great job of finding optimal shift/multiplier
values assuming a 10 seconds wrap around, but forgot to change the
overflow_period computation.

It overflows in cyclecounter_cyc2ns(), and the final result is 804 ms,
which is silly.

Lets simply use 5 seconds, no need to recompute this, given how it is
supposed to work.

Later, we will use a timer instead of a work queue, since the new RX
allocation schem will no longer need mlx4_en_recover_from_oom() and the
service_task firing every 250 ms.

Fixes: 31c128b66e5b ("net/mlx4_en: Choose time-stamping shift value according to HW frequency")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Tariq Toukan <tariqt@mellanox.com>
Cc: Eugenia Emantayev <eugenia@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-02-26 15:39:43 -05:00
..
2017-02-23 11:53:22 -08:00
2017-02-23 15:57:04 -08:00
2017-02-22 11:44:32 -08:00
2017-02-22 16:41:30 -08:00
2017-02-23 15:57:04 -08:00
2017-02-23 15:46:25 -08:00
2017-02-22 10:30:38 -08:00
2017-02-20 14:23:40 +01:00
2017-02-23 15:57:04 -08:00
2017-02-20 16:42:43 -08:00
2017-02-21 16:58:32 -08:00
2017-02-23 15:57:04 -08:00
2017-02-21 11:51:42 -08:00
2017-02-23 08:50:22 -08:00
2017-02-21 12:04:54 -08:00
2017-02-21 10:57:33 -08:00
2017-02-22 19:23:14 -08:00
2017-02-23 11:53:22 -08:00
2017-02-23 11:53:22 -08:00
2017-02-23 08:18:01 -08:00
2017-02-20 17:23:57 -08:00
2017-02-23 15:57:04 -08:00
2017-02-23 15:57:04 -08:00
2017-02-20 17:55:15 -08:00
2017-02-20 14:22:50 +01:00
2017-02-14 10:20:35 -08:00
2017-02-22 11:15:59 -08:00
2017-02-23 11:26:09 -08:00
2017-02-11 20:59:41 -05:00
2017-02-23 11:53:22 -08:00