dlan17/linux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge tag 'pm-5.12-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Browse Source 
Pull more power management updates from Rafael Wysocki:
 "These are fixes and cleanups on top of the power management material
  for 5.12-rc1 merged previously.

  Specifics:

   - Address cpufreq regression introduced in 5.11 that causes CPU
     frequency reporting to be distorted on systems with CPPC that use
     acpi-cpufreq as the scaling driver (Rafael Wysocki).

   - Fix regression introduced during the 5.10 development cycle related
     to CPU hotplug and policy recreation in the qcom-cpufreq-hw driver
     (Shawn Guo).

   - Fix recent regression in the operating performance points (OPP)
     framework that may cause frequency updates to be skipped by mistake
     in some cases (Jonathan Marek).

   - Simplify schedutil governor code and remove a misleading comment
     from it (Yue Hu).

   - Fix kerneldoc comment typo in the cpufreq core (Yue Hu)"

* tag 'pm-5.12-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  cpufreq: Fix typo in kerneldoc comment
  cpufreq: schedutil: Remove update_lock comment from struct sugov_policy definition
  cpufreq: schedutil: Remove needless sg_policy parameter from ignore_dl_rate_limit()
  cpufreq: ACPI: Set cpuinfo.max_freq directly if max boost is known
  cpufreq: qcom-hw: drop devm_xxx() calls from init/exit hooks
  opp: Don't skip freq update for different frequency
This commit is contained in:
Linus Torvalds
2021-02-23 14:59:46 -08:00
parent e0fbd25bb3 08c2a406b9
commit 005d3bd9e3
7 changed files with 69 additions and 67 deletions
Download Patch File Download Diff File Expand all files Collapse all files
kernel/sched/cpufreq_schedutil.c
+6 -8
View File
@@ -26,7 +26,7 @@ struct sugov_policy {
struct sugov_tunables *tunables;
struct list_head tunables_hook;
raw_spinlock_t update_lock; /* For shared policies */
raw_spinlock_t update_lock;
u64 last_freq_update_time;
s64 freq_update_delay_ns;
unsigned int next_freq;
@@ -320,23 +320,21 @@ static inline bool sugov_cpu_is_busy(struct sugov_cpu *sg_cpu) { return false; }
* Make sugov_should_update_freq() ignore the rate limit when DL
* has increased the utilization.
*/
static inline void ignore_dl_rate_limit(struct sugov_cpu *sg_cpu, struct sugov_policy *sg_policy)
static inline void ignore_dl_rate_limit(struct sugov_cpu *sg_cpu)
{
if (cpu_bw_dl(cpu_rq(sg_cpu->cpu)) > sg_cpu->bw_dl)
sg_policy->limits_changed = true;
sg_cpu->sg_policy->limits_changed = true;
}
static inline bool sugov_update_single_common(struct sugov_cpu *sg_cpu,
u64 time, unsigned int flags)
{
struct sugov_policy *sg_policy = sg_cpu->sg_policy;
sugov_iowait_boost(sg_cpu, time, flags);
sg_cpu->last_update = time;
ignore_dl_rate_limit(sg_cpu, sg_policy);
ignore_dl_rate_limit(sg_cpu);
if (!sugov_should_update_freq(sg_policy, time))
if (!sugov_should_update_freq(sg_cpu->sg_policy, time))
return false;
sugov_get_util(sg_cpu);
@@ -451,7 +449,7 @@ sugov_update_shared(struct update_util_data *hook, u64 time, unsigned int flags)
sugov_iowait_boost(sg_cpu, time, flags);
sg_cpu->last_update = time;
ignore_dl_rate_limit(sg_cpu, sg_policy);
ignore_dl_rate_limit(sg_cpu);
if (sugov_should_update_freq(sg_policy, time)) {
next_f = sugov_next_freq_shared(sg_cpu, time);