* Re: [PATCH 2/2] pkt_sched: Change PSCHED_SHIFT from 10 to 6
2009-06-09 8:05 [PATCH 2/2] pkt_sched: Change PSCHED_SHIFT from 10 to 6 Jarek Poplawski
@ 2009-06-09 12:21 ` David Miller
2009-06-15 9:15 ` [PATCH 3/2] pkt_sched: Rename PSCHED_US2NS and PSCHED_NS2US Jarek Poplawski
1 sibling, 0 replies; 4+ messages in thread
From: David Miller @ 2009-06-09 12:21 UTC (permalink / raw)
To: jarkao2; +Cc: kaber, vexwek, shemminger, netdev, devik, dada1, hazard, slavon
From: Jarek Poplawski <jarkao2@gmail.com>
Date: Tue, 9 Jun 2009 08:05:13 +0000
> Change PSCHED_SHIFT from 10 to 6 to increase schedulers time
> resolution. This will increase 16x a number of (internal) ticks per
> nanosecond, and is needed to improve accuracy of schedulers based on
> rate tables, like HTB, TBF or CBQ, with rates above 100Mbit. It is
> assumed this change is safe for 32bit accounting of time diffs up
> to 2 minutes, which should be enough for common use (extremely low
> rate values may overflow, so get inaccurate instead). To make full
> use of this change an updated iproute2 will be needed. (But using
> older iproute2 should be safe too.)
>
> This change breaks ticks - microseconds similarity, so some minor code
> fixes might be needed. It is also planned to change naming adequately
> eg. to PSCHED_TICKS2NS() etc. in the near future.
>
> Reported-by: Antonio Almeida <vexwek@gmail.com>
> Tested-by: Antonio Almeida <vexwek@gmail.com>
> Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Applied.
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH 3/2] pkt_sched: Rename PSCHED_US2NS and PSCHED_NS2US
2009-06-09 8:05 [PATCH 2/2] pkt_sched: Change PSCHED_SHIFT from 10 to 6 Jarek Poplawski
2009-06-09 12:21 ` David Miller
@ 2009-06-15 9:15 ` Jarek Poplawski
2009-06-15 9:32 ` David Miller
1 sibling, 1 reply; 4+ messages in thread
From: Jarek Poplawski @ 2009-06-15 9:15 UTC (permalink / raw)
To: David Miller
Cc: Patrick McHardy, Antonio Almeida, Stephen Hemminger, netdev,
Martin Devera, Eric Dumazet, Vladimir Ivashchenko,
Badalian Vyacheslav
On Tue, Jun 09, 2009 at 08:05:13AM +0000, Jarek Poplawski wrote:
...
> This change breaks ticks - microseconds similarity, so some minor code
> fixes might be needed. It is also planned to change naming adequately
> eg. to PSCHED_TICKS2NS() etc. in the near future.
-------------------->
pkt_sched: Rename PSCHED_US2NS and PSCHED_NS2US
Let's use TICKS instead of US, so PSCHED_TICKS2NS and PSCHED_NS2TICKS
(like in PSCHED_TICKS_PER_SEC already) to avoid misleading.
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
---
include/net/pkt_sched.h | 8 ++++----
net/sched/sch_api.c | 4 ++--
net/sched/sch_cbq.c | 4 ++--
3 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/include/net/pkt_sched.h b/include/net/pkt_sched.h
index 120935b..82a3191 100644
--- a/include/net/pkt_sched.h
+++ b/include/net/pkt_sched.h
@@ -43,15 +43,15 @@ typedef long psched_tdiff_t;
/* Avoid doing 64 bit divide */
#define PSCHED_SHIFT 6
-#define PSCHED_US2NS(x) ((s64)(x) << PSCHED_SHIFT)
-#define PSCHED_NS2US(x) ((x) >> PSCHED_SHIFT)
+#define PSCHED_TICKS2NS(x) ((s64)(x) << PSCHED_SHIFT)
+#define PSCHED_NS2TICKS(x) ((x) >> PSCHED_SHIFT)
-#define PSCHED_TICKS_PER_SEC PSCHED_NS2US(NSEC_PER_SEC)
+#define PSCHED_TICKS_PER_SEC PSCHED_NS2TICKS(NSEC_PER_SEC)
#define PSCHED_PASTPERFECT 0
static inline psched_time_t psched_get_time(void)
{
- return PSCHED_NS2US(ktime_to_ns(ktime_get()));
+ return PSCHED_NS2TICKS(ktime_to_ns(ktime_get()));
}
static inline psched_tdiff_t
diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c
index 3200979..24d17ce 100644
--- a/net/sched/sch_api.c
+++ b/net/sched/sch_api.c
@@ -484,7 +484,7 @@ void qdisc_watchdog_schedule(struct qdisc_watchdog *wd, psched_time_t expires)
wd->qdisc->flags |= TCQ_F_THROTTLED;
time = ktime_set(0, 0);
- time = ktime_add_ns(time, PSCHED_US2NS(expires));
+ time = ktime_add_ns(time, PSCHED_TICKS2NS(expires));
hrtimer_start(&wd->timer, time, HRTIMER_MODE_ABS);
}
EXPORT_SYMBOL(qdisc_watchdog_schedule);
@@ -1680,7 +1680,7 @@ static int psched_show(struct seq_file *seq, void *v)
hrtimer_get_res(CLOCK_MONOTONIC, &ts);
seq_printf(seq, "%08x %08x %08x %08x\n",
- (u32)NSEC_PER_USEC, (u32)PSCHED_US2NS(1),
+ (u32)NSEC_PER_USEC, (u32)PSCHED_TICKS2NS(1),
1000000,
(u32)NSEC_PER_SEC/(u32)ktime_to_ns(timespec_to_ktime(ts)));
diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c
index d728d81..23a1676 100644
--- a/net/sched/sch_cbq.c
+++ b/net/sched/sch_cbq.c
@@ -509,7 +509,7 @@ static void cbq_ovl_delay(struct cbq_class *cl)
q->pmask |= (1<<TC_CBQ_MAXPRIO);
expires = ktime_set(0, 0);
- expires = ktime_add_ns(expires, PSCHED_US2NS(sched));
+ expires = ktime_add_ns(expires, PSCHED_TICKS2NS(sched));
if (hrtimer_try_to_cancel(&q->delay_timer) &&
ktime_to_ns(ktime_sub(
hrtimer_get_expires(&q->delay_timer),
@@ -620,7 +620,7 @@ static enum hrtimer_restart cbq_undelay(struct hrtimer *timer)
ktime_t time;
time = ktime_set(0, 0);
- time = ktime_add_ns(time, PSCHED_US2NS(now + delay));
+ time = ktime_add_ns(time, PSCHED_TICKS2NS(now + delay));
hrtimer_start(&q->delay_timer, time, HRTIMER_MODE_ABS);
}
^ permalink raw reply related [flat|nested] 4+ messages in thread