| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <4836344.GXAFRqVoOG@kreacher>
Date: Thu, 03 Aug 2023 23:07:41 +0200
From: "Rafael J. Wysocki" <rjw@...ysocki.net>
To: Linux PM <linux-pm@...r.kernel.org>,
Peter Zijlstra <peterz@...radead.org>,
Anna-Maria Behnsen <anna-maria@...utronix.de>
Cc: LKML <linux-kernel@...r.kernel.org>,
Frederic Weisbecker <frederic@...nel.org>,
Kajetan Puchalski <kajetan.puchalski@....com>
Subject: [RFT][PATCH v2 1/3] cpuidle: teo: Do not call tick_nohz_get_sleep_length() upfront
From: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
Because the cost of calling tick_nohz_get_sleep_length() may increase
in the future, reorder the code in teo_select() so it first uses the
statistics to pick up a candidate idle state and applies the utilization
heuristic to it and only then calls tick_nohz_get_sleep_length() to
obtain the sleep length value and refine the selection if necessary.
This change by itself does not cause tick_nohz_get_sleep_length() to
be called less often, but it prepares the code for subsequent changes
that will do so.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
---
v1 -> v2:
* Fix up the "2 states and utilized CPU" special case handling, so it
allows the tick to be stopped in all of the cases when the target
residency of state 1 is above TICK_NSEC.
* Fix up a stale comment.
---
drivers/cpuidle/governors/teo.c | 105 ++++++++++++++++------------------------
1 file changed, 44 insertions(+), 61 deletions(-)
Index: linux-pm/drivers/cpuidle/governors/teo.c
===================================================================
--- linux-pm.orig/drivers/cpuidle/governors/teo.c
+++ linux-pm/drivers/cpuidle/governors/teo.c
@@ -306,15 +306,10 @@ static void teo_update(struct cpuidle_dr
cpu_data->total += PULSE;
}
-static bool teo_time_ok(u64 interval_ns)
+static bool teo_state_ok(int i, struct cpuidle_driver *drv)
{
- return !tick_nohz_tick_stopped() || interval_ns >= TICK_NSEC;
-}
-
-static s64 teo_middle_of_bin(int idx, struct cpuidle_driver *drv)
-{
- return (drv->states[idx].target_residency_ns +
- drv->states[idx+1].target_residency_ns) / 2;
+ return !tick_nohz_tick_stopped() ||
+ drv->states[i].target_residency_ns >= TICK_NSEC;
}
/**
@@ -354,6 +349,7 @@ static int teo_select(struct cpuidle_dri
{
struct teo_cpu *cpu_data = per_cpu_ptr(&teo_cpus, dev->cpu);
s64 latency_req = cpuidle_governor_latency_req(dev->cpu);
+ ktime_t delta_tick = TICK_NSEC / 2;
unsigned int idx_intercept_sum = 0;
unsigned int intercept_sum = 0;
unsigned int idx_recent_sum = 0;
@@ -363,7 +359,6 @@ static int teo_select(struct cpuidle_dri
int constraint_idx = 0;
int idx0 = 0, idx = -1;
bool alt_intercepts, alt_recent;
- ktime_t delta_tick;
bool cpu_utilized;
s64 duration_ns;
int i;
@@ -374,9 +369,11 @@ static int teo_select(struct cpuidle_dri
}
cpu_data->time_span_ns = local_clock();
-
- duration_ns = tick_nohz_get_sleep_length(&delta_tick);
- cpu_data->sleep_length_ns = duration_ns;
+ /*
+ * Set the expected sleep length to infinity in case of an early
+ * return.
+ */
+ cpu_data->sleep_length_ns = KTIME_MAX;
/* Check if there is any choice in the first place. */
if (drv->state_count < 2) {
@@ -384,11 +381,8 @@ static int teo_select(struct cpuidle_dri
goto out_tick;
}
- if (!dev->states_usage[0].disable) {
+ if (!dev->states_usage[0].disable)
idx = 0;
- if (drv->states[1].target_residency_ns > duration_ns)
- goto out_tick;
- }
cpu_utilized = teo_cpu_is_utilized(dev->cpu, cpu_data);
/*
@@ -397,8 +391,6 @@ static int teo_select(struct cpuidle_dri
* the shallowest non-polling state and exit.
*/
if (drv->state_count < 3 && cpu_utilized) {
- /* The CPU is utilized, so assume a short idle duration. */
- duration_ns = teo_middle_of_bin(0, drv);
/*
* If state 0 is enabled and it is not a polling one, select it
* right away unless the scheduler tick has been stopped, in
@@ -408,22 +400,17 @@ static int teo_select(struct cpuidle_dri
* anyway.
*/
if ((!idx && !(drv->states[0].flags & CPUIDLE_FLAG_POLLING) &&
- teo_time_ok(duration_ns)) || dev->states_usage[1].disable) {
+ teo_state_ok(0, drv)) || dev->states_usage[1].disable) {
idx = 0;
goto out_tick;
}
/* Assume that state 1 is not a polling one and use it. */
idx = 1;
+ duration_ns = drv->states[1].target_residency_ns;
goto end;
}
- /*
- * Find the deepest idle state whose target residency does not exceed
- * the current sleep length and the deepest idle state not deeper than
- * the former whose exit latency does not exceed the current latency
- * constraint. Compute the sums of metrics for early wakeup pattern
- * detection.
- */
+ /* Compute the sums of metrics for early wakeup pattern detection. */
for (i = 1; i < drv->state_count; i++) {
struct teo_bin *prev_bin = &cpu_data->state_bins[i-1];
struct cpuidle_state *s = &drv->states[i];
@@ -439,19 +426,15 @@ static int teo_select(struct cpuidle_dri
if (dev->states_usage[i].disable)
continue;
- if (idx < 0) {
- idx = i; /* first enabled state */
- idx0 = i;
- }
-
- if (s->target_residency_ns > duration_ns)
- break;
+ if (idx < 0)
+ idx0 = i; /* first enabled state */
idx = i;
if (s->exit_latency_ns <= latency_req)
constraint_idx = i;
+ /* Save the sums for the current state. */
idx_intercept_sum = intercept_sum;
idx_hit_sum = hit_sum;
idx_recent_sum = recent_sum;
@@ -465,7 +448,7 @@ static int teo_select(struct cpuidle_dri
if (idx == idx0) {
/*
- * This is the first enabled idle state, so use it, but do not
+ * Only one idle state is enabled, so use it, but do not
* allow the tick to be stopped it is shallow enough.
*/
duration_ns = drv->states[idx].target_residency_ns;
@@ -479,13 +462,11 @@ static int teo_select(struct cpuidle_dri
* all of the deeper states, or the sum of the numbers of recent
* intercepts over all of the states shallower than the candidate one
* is greater than a half of the number of recent events taken into
- * account, the CPU is likely to wake up early, so find an alternative
- * idle state to select.
+ * account, a shallower idle state is likely to be a better choice.
*/
alt_intercepts = 2 * idx_intercept_sum > cpu_data->total - idx_hit_sum;
alt_recent = idx_recent_sum > NR_RECENT / 2;
if (alt_recent || alt_intercepts) {
- s64 first_suitable_span_ns = duration_ns;
int first_suitable_idx = idx;
/*
@@ -494,44 +475,39 @@ static int teo_select(struct cpuidle_dri
* cases (both with respect to intercepts overall and with
* respect to the recent intercepts only) in the past.
*
- * Take the possible latency constraint and duration limitation
- * present if the tick has been stopped already into account.
+ * Take the possible duration limitation present if the tick
+ * has been stopped already into account.
*/
intercept_sum = 0;
recent_sum = 0;
for (i = idx - 1; i >= 0; i--) {
struct teo_bin *bin = &cpu_data->state_bins[i];
- s64 span_ns;
intercept_sum += bin->intercepts;
recent_sum += bin->recent;
- span_ns = teo_middle_of_bin(i, drv);
-
if ((!alt_recent || 2 * recent_sum > idx_recent_sum) &&
(!alt_intercepts ||
2 * intercept_sum > idx_intercept_sum)) {
- if (teo_time_ok(span_ns) &&
- !dev->states_usage[i].disable) {
+ /*
+ * Use the current state unless it is too
+ * shallow or disabled, in which case take the
+ * first enabled state that is deep enough.
+ */
+ if (teo_state_ok(i, drv) &&
+ !dev->states_usage[i].disable)
idx = i;
- duration_ns = span_ns;
- } else {
- /*
- * The current state is too shallow or
- * disabled, so take the first enabled
- * deeper state with suitable time span.
- */
+ else
idx = first_suitable_idx;
- duration_ns = first_suitable_span_ns;
- }
+
break;
}
if (dev->states_usage[i].disable)
continue;
- if (!teo_time_ok(span_ns)) {
+ if (!teo_state_ok(i, drv)) {
/*
* The current state is too shallow, but if an
* alternative candidate state has been found,
@@ -543,7 +519,6 @@ static int teo_select(struct cpuidle_dri
break;
}
- first_suitable_span_ns = span_ns;
first_suitable_idx = i;
}
}
@@ -562,14 +537,22 @@ static int teo_select(struct cpuidle_dri
* not sufficiently large.
*/
if (cpu_utilized) {
- s64 span_ns;
+ i = teo_find_shallower_state(drv, dev, idx, KTIME_MAX, true);
+ if (teo_state_ok(i, drv))
+ idx = i;
+ }
- i = teo_find_shallower_state(drv, dev, idx, duration_ns, true);
- span_ns = teo_middle_of_bin(i, drv);
- if (teo_time_ok(span_ns)) {
+ duration_ns = tick_nohz_get_sleep_length(&delta_tick);
+ cpu_data->sleep_length_ns = duration_ns;
+
+ /*
+ * If the closest expected timer is before the terget residency of the
+ * candidate state, a shallower one needs to be found.
+ */
+ if (drv->states[idx].target_residency_ns > duration_ns) {
+ i = teo_find_shallower_state(drv, dev, idx, duration_ns, false);
+ if (teo_state_ok(i, drv))
idx = i;
- duration_ns = span_ns;
- }
}
end:
Powered by blists - more mailing lists