This is a note to let you know that I've just added the patch titled
cpufreq: intel_pstate: Refine computation of P-state for given frequency
to the 6.6-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
cpufreq-intel_pstate-refine-computation-of-p-state-for-given-frequency.patch
and it can be found in the queue-6.6 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
>From 192cdb1c907fd8df2d764c5bb17496e415e59391 Mon Sep 17 00:00:00 2001
From: "Rafael J. Wysocki" <rafael.j.wysocki@xxxxxxxxx>
Date: Mon, 22 Jan 2024 15:18:11 +0100
Subject: cpufreq: intel_pstate: Refine computation of P-state for given frequency
From: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
commit 192cdb1c907fd8df2d764c5bb17496e415e59391 upstream.
On systems using HWP, if a given frequency is equal to the maximum turbo
frequency or the maximum non-turbo frequency, the HWP performance level
corresponding to it is already known and can be used directly without
any computation.
Accordingly, adjust the code to use the known HWP performance levels in
the cases mentioned above.
This also helps to avoid limiting CPU capacity artificially in some
cases when the BIOS produces the HWP_CAP numbers using a different
E-core-to-P-core performance scaling factor than expected by the kernel.
Fixes: f5c8cf2a4992 ("cpufreq: intel_pstate: hybrid: Use known scaling factor for P-cores")
Cc: 6.1+ <stable@xxxxxxxxxxxxxxx> # 6.1+
Tested-by: Srinivas Pandruvada <srinivas.pandruvada@xxxxxxxxxxxxxxx>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/cpufreq/intel_pstate.c | 55 +++++++++++++++++++++++++----------------
1 file changed, 34 insertions(+), 21 deletions(-)
--- a/drivers/cpufreq/intel_pstate.c
+++ b/drivers/cpufreq/intel_pstate.c
@@ -526,6 +526,30 @@ static int intel_pstate_cppc_get_scaling
}
#endif /* CONFIG_ACPI_CPPC_LIB */
+static int intel_pstate_freq_to_hwp_rel(struct cpudata *cpu, int freq,
+ unsigned int relation)
+{
+ if (freq == cpu->pstate.turbo_freq)
+ return cpu->pstate.turbo_pstate;
+
+ if (freq == cpu->pstate.max_freq)
+ return cpu->pstate.max_pstate;
+
+ switch (relation) {
+ case CPUFREQ_RELATION_H:
+ return freq / cpu->pstate.scaling;
+ case CPUFREQ_RELATION_C:
+ return DIV_ROUND_CLOSEST(freq, cpu->pstate.scaling);
+ }
+
+ return DIV_ROUND_UP(freq, cpu->pstate.scaling);
+}
+
+static int intel_pstate_freq_to_hwp(struct cpudata *cpu, int freq)
+{
+ return intel_pstate_freq_to_hwp_rel(cpu, freq, CPUFREQ_RELATION_L);
+}
+
/**
* intel_pstate_hybrid_hwp_adjust - Calibrate HWP performance levels.
* @cpu: Target CPU.
@@ -543,6 +567,7 @@ static void intel_pstate_hybrid_hwp_adju
int perf_ctl_scaling = cpu->pstate.perf_ctl_scaling;
int perf_ctl_turbo = pstate_funcs.get_turbo(cpu->cpu);
int scaling = cpu->pstate.scaling;
+ int freq;
pr_debug("CPU%d: perf_ctl_max_phys = %d\n", cpu->cpu, perf_ctl_max_phys);
pr_debug("CPU%d: perf_ctl_turbo = %d\n", cpu->cpu, perf_ctl_turbo);
@@ -556,16 +581,16 @@ static void intel_pstate_hybrid_hwp_adju
cpu->pstate.max_freq = rounddown(cpu->pstate.max_pstate * scaling,
perf_ctl_scaling);
- cpu->pstate.max_pstate_physical =
- DIV_ROUND_UP(perf_ctl_max_phys * perf_ctl_scaling,
- scaling);
+ freq = perf_ctl_max_phys * perf_ctl_scaling;
+ cpu->pstate.max_pstate_physical = intel_pstate_freq_to_hwp(cpu, freq);
- cpu->pstate.min_freq = cpu->pstate.min_pstate * perf_ctl_scaling;
+ freq = cpu->pstate.min_pstate * perf_ctl_scaling;
+ cpu->pstate.min_freq = freq;
/*
* Cast the min P-state value retrieved via pstate_funcs.get_min() to
* the effective range of HWP performance levels.
*/
- cpu->pstate.min_pstate = DIV_ROUND_UP(cpu->pstate.min_freq, scaling);
+ cpu->pstate.min_pstate = intel_pstate_freq_to_hwp(cpu, freq);
}
static inline void update_turbo_state(void)
@@ -2528,13 +2553,12 @@ static void intel_pstate_update_perf_lim
* abstract values to represent performance rather than pure ratios.
*/
if (hwp_active && cpu->pstate.scaling != perf_ctl_scaling) {
- int scaling = cpu->pstate.scaling;
int freq;
freq = max_policy_perf * perf_ctl_scaling;
- max_policy_perf = DIV_ROUND_UP(freq, scaling);
+ max_policy_perf = intel_pstate_freq_to_hwp(cpu, freq);
freq = min_policy_perf * perf_ctl_scaling;
- min_policy_perf = DIV_ROUND_UP(freq, scaling);
+ min_policy_perf = intel_pstate_freq_to_hwp(cpu, freq);
}
pr_debug("cpu:%d min_policy_perf:%d max_policy_perf:%d\n",
@@ -2908,18 +2932,7 @@ static int intel_cpufreq_target(struct c
cpufreq_freq_transition_begin(policy, &freqs);
- switch (relation) {
- case CPUFREQ_RELATION_L:
- target_pstate = DIV_ROUND_UP(freqs.new, cpu->pstate.scaling);
- break;
- case CPUFREQ_RELATION_H:
- target_pstate = freqs.new / cpu->pstate.scaling;
- break;
- default:
- target_pstate = DIV_ROUND_CLOSEST(freqs.new, cpu->pstate.scaling);
- break;
- }
-
+ target_pstate = intel_pstate_freq_to_hwp_rel(cpu, freqs.new, relation);
target_pstate = intel_cpufreq_update_pstate(policy, target_pstate, false);
freqs.new = target_pstate * cpu->pstate.scaling;
@@ -2937,7 +2950,7 @@ static unsigned int intel_cpufreq_fast_s
update_turbo_state();
- target_pstate = DIV_ROUND_UP(target_freq, cpu->pstate.scaling);
+ target_pstate = intel_pstate_freq_to_hwp(cpu, target_freq);
target_pstate = intel_cpufreq_update_pstate(policy, target_pstate, true);
Patches currently in stable-queue which might be from rafael.j.wysocki@xxxxxxxxx are
queue-6.6/async-split-async_schedule_node_domain.patch
queue-6.6/cpufreq-intel_pstate-refine-computation-of-p-state-for-given-frequency.patch
queue-6.6/pm-sleep-fix-possible-deadlocks-in-core-system-wide-pm-code.patch
queue-6.6/async-introduce-async_schedule_dev_nocall.patch
queue-6.6/pm-hibernate-enforce-ordering-during-image-compression-decompression.patch
