lkml.org 
[lkml]   [2022]   [Apr]   [24]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[PATCH 2/3] sched_clock: Use Hz as the unit for clock rate reporting below 4kHz
We have always used kHz as the unit for clock rates reported between 
1MHz (inclusive) and 4MHz (exclusive), e.g.:

riscv_timer_init_dt: Registering clocksource cpuid [0] hartid [1]
clocksource: riscv_clocksource: mask: 0xffffffffffffffff max_cycles: 0x1d854df40, max_idle_ns: 3526361616960 ns
sched_clock: 64 bits at 1000kHz, resolution 1000ns, wraps every 2199023255500ns

This reduces the amount of data lost due to rounding, but hasn't been
replicated for the kHz range when support was added for proper reporting
of sub-kHz clock rates. Take the same approach then for rates between
1kHz (inclusive) and 4kHz (exclusive), also for consistency.

Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Fixes: 2f0778afac79 ("ARM: 7205/2: sched_clock: allow sched_clock to be selected at runtime")
---
Russell,

Please correct me if I am wrong with my guess for the origin of the 4MHz
boundary as there's no mention of it with your commit 112f38a4a316 ("ARM:
sched_clock: provide common infrastructure for sched_clock()"), which is
where this code has come from.

Maciej
---
kernel/time/sched_clock.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

linux-sched-clock-rate-4khz.diff
Index: linux-macro/kernel/time/sched_clock.c
===================================================================
--- linux-macro.orig/kernel/time/sched_clock.c
+++ linux-macro/kernel/time/sched_clock.c
@@ -203,7 +203,7 @@ sched_clock_register(u64 (*read)(void),
r = DIV_ROUND_CLOSEST(r, 1000000);
r_unit = 'M';
} else {
- if (r >= 1000) {
+ if (r >= 4000) {
r = DIV_ROUND_CLOSEST(r, 1000);
r_unit = 'k';
} else {
\
 
 \ /
  Last update: 2022-04-24 13:48    [W:0.091 / U:0.284 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site