Skip to contentSkip to conversation

sched: Provide USF for the portable equipment.

PATCH In review Dongdong Yang 5 participants 8 messages Jul 30, 2020 – Aug 3, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 sched: Provide USF for the portable equipment.

From: Dongdong Yang <[email protected]>

This patch provides USF(User Sensitive Feedback factor) auxiliary cpufreq governor to support high level layer sysfs inodes setting for utils adjustment purpose from the identified scenario on portable equipment. Because the power consumption and UI response are more cared for by portable equipment users. And the "screen off" status stands for no request from the user, however, the kernel is still expected to notify the user in time on modem, network or powerkey events occur. USF provides "sched_usf_non_ux_r" sysfs inode to cut down the utils from user space tasks according to high level scenario. In addition, it usually hints more cpufreq demand that the preemptive counts of the tasks on the cpu burst and over the user expecting completed time such as the ratio sysctl_sched_latency to sysctl_sched_min_granularity on "screen on" status, which more likely with more UI. The sysfs inodes "sched_usf_up_l0_r" and "sched_usf_down_r" have been provided to adjust the utils according to high level identified scenario to alloc the cpufreq in time.

Dongdong Yang (1):
sched: Provide USF for portable equipment.

 drivers/staging/Kconfig          |   2 +
 drivers/staging/Makefile         |   1 +
 drivers/staging/fbsched/Kconfig  |  10 ++
 drivers/staging/fbsched/Makefile |   2 +
 drivers/staging/fbsched/usf.c    | 351 +++++++++++++++++++++++++++++++++++++++
 kernel/sched/cpufreq_schedutil.c |  11 +-
 6 files changed, 376 insertions(+), 1 deletion(-)
 create mode 100644 drivers/staging/fbsched/Kconfig
 create mode 100644 drivers/staging/fbsched/Makefile
 create mode 100644 drivers/staging/fbsched/usf.c
Signature · 1 line
2.7.4
#1 lore raw

PATCH 1/1 sched: Provide USF for the portable equipment.

From: Dongdong Yang <[email protected]>

The power consumption and UI response are more cared for by the portable equipment users. USF(User Sensitive Feedback factor) auxiliary cpufreq governor is providing more utils adjustment settings to a high level by scenario identification.

From the view of portable equipment, screen off status usually stands for no request from the user, however, the kernel is still expected to notify the user in time on modem, network or powerkey events occur. In some scenarios, such as listening to music, low power processors, such as DSP, take more actions and CPU load requirements cut down.  It would bring more power consumption benefit if high level have interfaces to adjust utils according to the current scenario and load.

In addition, the portable equipment user usually heavy interact with devices by touch, and other peripherals. The boost preemptive counts are marking the load requirement urgent, vice versa. If such feedback factor could be set to high level according to the scenario, it would contribute to the power consumption and UI response.

If no USF sysfs inode is set, and no screen on or off event, adjust_task_pred_demand shall not be invoked. Once sched_usf_up_l0_r/down_r/non_ux_r be set, adjust_task_pred_demand_impl shall be called back to update settings according to high level scenario identification.

We can get about 17% mean power consumption save at listening to music with speaker on "screen off" scenario, as below statistical data from 7766 XiaoMi devices for two weeks with sched_usf_non_ux_r be set:

        day1         day2         day3         day4
count   7766.000000  7766.000000  7766.000000  7766.000000
mean    88.035525    85.500282    83.829305    86.054997
std     111.049980   108.258834   107.562583   108.558240
min     0.099000     0.037000     0.067000     0.045000
25%     34.765500    34.021750    34.101500    34.423000
50%     54.950000    55.286500    54.189500    54.248500
75%     95.954000    93.942000    91.738000    94.0592500
80%     114.675000   107.430000   106.378000   108.673000
85%     137.851000   129.511000   127.156500   131.750750
90%     179.669000   170.208500   164.027000   172.348000
95%     272.395000   257.845500   247.750500   263.275750
98%     399.034500   412.170400   391.484000   402.835600
        day5         day6        day7         day8
count   7766.000000  7766.00000  7766.000000  7766.000000
mean    82.532677    79.21923    77.611380    81.075081
std     104.870079   101.34819   103.140037   97.506221
min     0.051000     0.02900     0.007000     0.068000
25%     32.873000    33.44400    31.965500    33.863500
50%     52.180500    51.56550    50.806500    53.080000
75%     90.905750    86.82625    83.859250    89.973000
80%     105.455000   99.64700    97.271000    104.225000
85%     128.300000   118.47825   116.570250   126.648250
90%     166.647500   149.18000   150.649500   161.087000
95%     247.208500   224.36050   226.380000   245.291250
98%     393.002000   347.92060   369.791800   378.778600
        day9         day10        day11        day12
count   7766.000000  7766.000000  7766.000000  7766.000000
mean    79.989170    83.859417    78.032930    77.060542
std     104.226122   108.893043   102.561715   99.844276
min     0.118000     0.017000     0.028000     0.039000
25%     32.056250    33.454500    31.176250    30.897750
50%     51.506000    54.056000    48.969500    49.069000
75%     88.513500    92.953500    83.506750    84.096000
80%     102.876000   107.845000   97.717000    98.073000
85%     124.363000   128.288000   118.366500   116.869250
90%     160.557000   167.084000   154.342500   148.187500
95%     231.149000   242.925750   236.759000   228.131250
98%     367.206600   388.619100   385.269100   376.541600
        day13        day14
count   7766.000000  7766.000000
mean    75.528036    73.702878
std     90.750594    86.796016
min     0.066000     0.054000
25%     31.170500    31.608500
50%     48.758500    49.215000
75%     84.522750    83.053000
80%     97.879000    94.875000
85%     116.680250   113.573750
90%     149.083500   144.089500
95%     226.177750   211.488750
98%     347.011100   331.317100
---
 drivers/staging/Kconfig          |   2 +
 drivers/staging/Makefile         |   1 +
 drivers/staging/fbsched/Kconfig  |  10 ++
 drivers/staging/fbsched/Makefile |   2 +
 drivers/staging/fbsched/usf.c    | 351 +++++++++++++++++++++++++++++++++++++++
 kernel/sched/cpufreq_schedutil.c |  11 +-
 6 files changed, 376 insertions(+), 1 deletion(-)
 create mode 100644 drivers/staging/fbsched/Kconfig
 create mode 100644 drivers/staging/fbsched/Makefile
 create mode 100644 drivers/staging/fbsched/usf.c
drivers/staging/Kconfig +2 -0
index 4ec5528..05b231e 100644
@@ -120,4 +120,6 @@ source "drivers/staging/qlge/Kconfig"
 
 source "drivers/staging/wfx/Kconfig"
 
+source "drivers/staging/fbsched/Kconfig"
+
 endif # STAGING
drivers/staging/Makefile +1 -0
index 4d34198..e016ec6 100644
@@ -50,3 +50,4 @@ obj-$(CONFIG_FIELDBUS_DEV)     += fieldbus/
 obj-$(CONFIG_KPC2000)		+= kpc2000/
 obj-$(CONFIG_QLGE)		+= qlge/
 obj-$(CONFIG_WFX)		+= wfx/
+obj-$(CONFIG_SCHED_USF)		+= fbsched/
drivers/staging/fbsched/Kconfig new file +10 -0
new file mode 100644
index 0000000..954b6af
@@ -0,0 +1,10 @@
+config SCHED_USF
+	bool "User Sensitive Factors for Scheduler"
+	depends on CPU_FREQ_GOV_SCHEDUTIL && FB
+	help
+	  Select this option to enable the adjustment on the cpufreq with
+	  the user sensitive factors on schedule. It is special for portable
+	  equipment which more power care and quick response requirement on
+	  screen on.
+
+	  If unsure, say N.
drivers/staging/fbsched/Makefile new file +2 -0
new file mode 100644
index 0000000..f56aa6c
@@ -0,0 +1,2 @@
+LINUXINCLUDE += -include $(srctree)/kernel/sched/sched.h
+obj-$(CONFIG_SCHED_USF) += usf.o
drivers/staging/fbsched/usf.c new file +351 -0

351 lines not on this page · raw message

kernel/sched/cpufreq_schedutil.c +10 -1
index 7fbaee2..7bc3429 100644
@@ -289,12 +289,21 @@ unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs,
 	return min(max, util);
 }
 
+#ifdef CONFIG_SCHED_USF
+void (*adjust_task_pred_demand)(int cpuid, unsigned long *util,
+	struct rq *rq) = NULL;
+EXPORT_SYMBOL(adjust_task_pred_demand);
+#endif
+
 static unsigned long sugov_get_util(struct sugov_cpu *sg_cpu)
 {
 	struct rq *rq = cpu_rq(sg_cpu->cpu);
 	unsigned long util = cpu_util_cfs(rq);
 	unsigned long max = arch_scale_cpu_capacity(sg_cpu->cpu);
-
+#ifdef CONFIG_SCHED_USF
+	if (adjust_task_pred_demand)
+		adjust_task_pred_demand(sg_cpu->cpu, &util, rq);
+#endif
 	sg_cpu->max = max;
 	sg_cpu->bw_dl = cpu_bw_dl(rq);
 
Signature · 1 line
2.7.4
#2 lore raw

On Thu, 30 Jul 2020 21:35:43 +0800 Dongdong Yang <[email protected]> wrote:

I'll let others decide the value of this, but I have some comments.

Quoted 6 lines

Why all the signed-off-bys? All of you worked on it?

drivers/staging/fbsched/usf.c line 111

Quoted 11 lines
drivers/staging/fbsched/usf.c new file +11 -0
+	if (evdata->data && val == FB_EVENT_BLANK) {
+		blank = *(int *)(evdata->data);
+
+		switch (blank) {
+		case FB_BLANK_POWERDOWN:
+			usf_vdev.is_screen_on = 0;
+			if (usf_vdev.sysctl_sched_usf_non_ux != 0)
+				adjust_task_pred_demand =
+				    &adjust_task_pred_demand_impl;
+			else
+				adjust_task_pred_demand = NULL;

So sysctl can enable and disable this?

Quoted 17 lines
drivers/staging/fbsched/usf.c new file +3 -0
+
+			break;
+
kernel/sched/cpufreq_schedutil.c +5 -0
index 7fbaee2..7bc3429 100644
@@ -289,12 +289,21 @@ unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs,
 	return min(max, util);
 }
 
+#ifdef CONFIG_SCHED_USF
+void (*adjust_task_pred_demand)(int cpuid, unsigned long *util,
+	struct rq *rq) = NULL;
+EXPORT_SYMBOL(adjust_task_pred_demand);
+#endif

kernel/sched/cpufreq_schedutil.c unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs, · line 305

Quoted 10 lines
kernel/sched/cpufreq_schedutil.c +4 -1
+
 static unsigned long sugov_get_util(struct sugov_cpu *sg_cpu)
 {
 	struct rq *rq = cpu_rq(sg_cpu->cpu);
 	unsigned long util = cpu_util_cfs(rq);
 	unsigned long max = arch_scale_cpu_capacity(sg_cpu->cpu);
-
+#ifdef CONFIG_SCHED_USF
+	if (adjust_task_pred_demand)
+		adjust_task_pred_demand(sg_cpu->cpu, &util, rq);

The above is racy. Nothing stops adjust_task_pred_demand from being non-null at the if condition, then becoming NULL before it is called.

Instead I would have the following:

DEFINE_STATIC_KEY_FALSE(adjust_task_pred_set);

#ifdef CONFIG_SCHED_USF
void adjust_task_pred_demand(int cpuid, unsigned long *util,

				struct rq *rq);

#else
static inline void adjust_task_pred_demand(int cpuid,
unsigned long *util, struct rq *rq)
{ }
#endif

	if (static_key_unlikely(adjust_task_pred_set))
		adjust_task_pred_demand(sg_cpu->cpu, &util, rq);

And hopefully the compiler will just remove all of it if it's not enabled.

Then you set the static branch when you want it to be called, and do not use a racy function pointer.

-- Steve

Quoted 4 lines
kernel/sched/cpufreq_schedutil.c +1 -0
+#endif
 	sg_cpu->max = max;
 	sg_cpu->bw_dl = cpu_bw_dl(rq);
 
maintainer
#3 lore raw
In reply to Dongdong Yang
Quoted 34 lines · Dongdong Yang

From: Dongdong Yang <[email protected]>

The power consumption and UI response are more cared for by the portable equipment users. USF(User Sensitive Feedback factor) auxiliary cpufreq governor is providing more utils adjustment settings to a high level by scenario identification.

Quoted 1 line

From the view of portable equipment, screen off status

usually stands for no request from the user, however, the kernel is still expected to notify the user in time on modem, network or powerkey events occur. In some scenarios, such as listening to music, low power processors, such as DSP, take more actions and CPU load requirements cut down.  It would bring more power consumption benefit if high level have interfaces to adjust utils according to the current scenario and load.

In addition, the portable equipment user usually heavy interact with devices by touch, and other peripherals. The boost preemptive counts are marking the load requirement urgent, vice versa. If such feedback factor could be set to high level according to the scenario, it would contribute to the power consumption and UI response.

If no USF sysfs inode is set, and no screen on or off event, adjust_task_pred_demand shall not be invoked. Once sched_usf_up_l0_r/down_r/non_ux_r be set, adjust_task_pred_demand_impl shall be called back to update settings according to high level scenario identification.

Quoted 5 lines

We can get about 17% mean power consumption save at listening to music with speaker on "screen off" scenario, as below statistical data from 7766 XiaoMi devices for two weeks with sched_usf_non_ux_r be set:

Nit, you can wrap your changelog text at 72 columns to make it easier to read.

Quoted 62 lines
        day1         day2         day3         day4
count   7766.000000  7766.000000  7766.000000  7766.000000
mean    88.035525    85.500282    83.829305    86.054997
std     111.049980   108.258834   107.562583   108.558240
min     0.099000     0.037000     0.067000     0.045000
25%     34.765500    34.021750    34.101500    34.423000
50%     54.950000    55.286500    54.189500    54.248500
75%     95.954000    93.942000    91.738000    94.0592500
80%     114.675000   107.430000   106.378000   108.673000
85%     137.851000   129.511000   127.156500   131.750750
90%     179.669000   170.208500   164.027000   172.348000
95%     272.395000   257.845500   247.750500   263.275750
98%     399.034500   412.170400   391.484000   402.835600
        day5         day6        day7         day8
count   7766.000000  7766.00000  7766.000000  7766.000000
mean    82.532677    79.21923    77.611380    81.075081
std     104.870079   101.34819   103.140037   97.506221
min     0.051000     0.02900     0.007000     0.068000
25%     32.873000    33.44400    31.965500    33.863500
50%     52.180500    51.56550    50.806500    53.080000
75%     90.905750    86.82625    83.859250    89.973000
80%     105.455000   99.64700    97.271000    104.225000
85%     128.300000   118.47825   116.570250   126.648250
90%     166.647500   149.18000   150.649500   161.087000
95%     247.208500   224.36050   226.380000   245.291250
98%     393.002000   347.92060   369.791800   378.778600
        day9         day10        day11        day12
count   7766.000000  7766.000000  7766.000000  7766.000000
mean    79.989170    83.859417    78.032930    77.060542
std     104.226122   108.893043   102.561715   99.844276
min     0.118000     0.017000     0.028000     0.039000
25%     32.056250    33.454500    31.176250    30.897750
50%     51.506000    54.056000    48.969500    49.069000
75%     88.513500    92.953500    83.506750    84.096000
80%     102.876000   107.845000   97.717000    98.073000
85%     124.363000   128.288000   118.366500   116.869250
90%     160.557000   167.084000   154.342500   148.187500
95%     231.149000   242.925750   236.759000   228.131250
98%     367.206600   388.619100   385.269100   376.541600
        day13        day14
count   7766.000000  7766.000000
mean    75.528036    73.702878
std     90.750594    86.796016
min     0.066000     0.054000
25%     31.170500    31.608500
50%     48.758500    49.215000
75%     84.522750    83.053000
80%     97.879000    94.875000
85%     116.680250   113.573750
90%     149.083500   144.089500
95%     226.177750   211.488750
98%     347.011100   331.317100
Quoted 6 lines
---
 drivers/staging/Kconfig          |   2 +
 drivers/staging/Makefile         |   1 +
 drivers/staging/fbsched/Kconfig  |  10 ++
 drivers/staging/fbsched/Makefile |   2 +
 drivers/staging/fbsched/usf.c    | 351 +++++++++++++++++++++++++++++++++++++++

Why the different names, "fbsched" and "usf"? what does "fbsched" mean?

Quoted 1 line
 kernel/sched/cpufreq_schedutil.c |  11 +-

Why are you touching code outside of drivers/staging/ at all? That's usually a good sign that this should not be a staging driver as they should all be self-contained so nothing else in the kernel is messed with.

kernel/sched/cpufreq_schedutil.c unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs, · line 296

Quoted 11 lines
kernel/sched/cpufreq_schedutil.c +5 -0
@@ -289,12 +289,21 @@ unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs,
 	return min(max, util);
 }
 
+#ifdef CONFIG_SCHED_USF
+void (*adjust_task_pred_demand)(int cpuid, unsigned long *util,
+	struct rq *rq) = NULL;
+EXPORT_SYMBOL(adjust_task_pred_demand);
+#endif

No #ifdef in .c code. And why not EXPORT_SYMBOL_GPL?

kernel/sched/cpufreq_schedutil.c unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs, · line 306

Quoted 11 lines
kernel/sched/cpufreq_schedutil.c +5 -1
+
 static unsigned long sugov_get_util(struct sugov_cpu *sg_cpu)
 {
 	struct rq *rq = cpu_rq(sg_cpu->cpu);
 	unsigned long util = cpu_util_cfs(rq);
 	unsigned long max = arch_scale_cpu_capacity(sg_cpu->cpu);
-
+#ifdef CONFIG_SCHED_USF
+	if (adjust_task_pred_demand)
+		adjust_task_pred_demand(sg_cpu->cpu, &util, rq);
+#endif

Again, no #ifdef in .c code should be ever done, especially for something as simple as this. Otherwise the code is totally unmaintainable over time.

thanks,

greg k-h

maintainer NAK
#4 lore raw
In reply to Dongdong Yang
Quoted 16 lines · Dongdong Yang
kernel/sched/cpufreq_schedutil.c +6 -0
index 7fbaee2..7bc3429 100644
@@ -289,12 +289,21 @@ unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs,
 	return min(max, util);
 }
 
+#ifdef CONFIG_SCHED_USF
+void (*adjust_task_pred_demand)(int cpuid, unsigned long *util,
+	struct rq *rq) = NULL;
+EXPORT_SYMBOL(adjust_task_pred_demand);
+#endif
+
 static unsigned long sugov_get_util(struct sugov_cpu *sg_cpu)
 {

kernel/sched/cpufreq_schedutil.c unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs, · line 308

Quoted 10 lines
kernel/sched/cpufreq_schedutil.c +4 -1
 	struct rq *rq = cpu_rq(sg_cpu->cpu);
 	unsigned long util = cpu_util_cfs(rq);
 	unsigned long max = arch_scale_cpu_capacity(sg_cpu->cpu);
-
+#ifdef CONFIG_SCHED_USF
+	if (adjust_task_pred_demand)
+		adjust_task_pred_demand(sg_cpu->cpu, &util, rq);
+#endif
 	sg_cpu->max = max;
 	sg_cpu->bw_dl = cpu_bw_dl(rq);

NAK

#5 lore raw
Quoted 21 lines · On Fri, 31 Jul 2020 20:15:38 +0200 [email protected] wrote:
Quoted 20 lines · Dongdong Yang
kernel/sched/cpufreq_schedutil.c +6 -1
index 7fbaee2..7bc3429 100644
@@ -289,12 +289,21 @@ unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs,
 	return min(max, util);
 }
 
+#ifdef CONFIG_SCHED_USF
+void (*adjust_task_pred_demand)(int cpuid, unsigned long *util,
+	struct rq *rq) = NULL;
+EXPORT_SYMBOL(adjust_task_pred_demand);
+#endif
+
 static unsigned long sugov_get_util(struct sugov_cpu *sg_cpu)
 {
 	struct rq *rq = cpu_rq(sg_cpu->cpu);
 	unsigned long util = cpu_util_cfs(rq);
 	unsigned long max = arch_scale_cpu_capacity(sg_cpu->cpu);
-
Quoted 8 lines

kernel/sched/cpufreq_schedutil.c unsigned long schedutil_cpu_util(int cpu, unsigned long util_cfs, · line 308

Quoted 6 lines
kernel/sched/cpufreq_schedutil.c +4 -0
+#ifdef CONFIG_SCHED_USF
+	if (adjust_task_pred_demand)
+		adjust_task_pred_demand(sg_cpu->cpu, &util, rq);
+#endif
 	sg_cpu->max = max;
 	sg_cpu->bw_dl = cpu_bw_dl(rq);  

NAK

Peter,

It's more informative if you include rational with a NAK.

-- Steve

#6 lore raw
In reply to Dongdong Yang
Quoted 197 lines · Dongdong Yang

From: Dongdong Yang <[email protected]>

The power consumption and UI response are more cared for by the portable equipment users. USF(User Sensitive Feedback factor) auxiliary cpufreq governor is providing more utils adjustment settings to a high level by scenario identification.

From the view of portable equipment, screen off status usually stands for no request from the user, however, the kernel is still expected to notify the user in time on modem, network or powerkey events occur. In some scenarios, such as listening to music, low power processors, such as DSP, take more actions and CPU load requirements cut down.  It would bring more power consumption benefit if high level have interfaces to adjust utils according to the current scenario and load.

In addition, the portable equipment user usually heavy interact with devices by touch, and other peripherals. The boost preemptive counts are marking the load requirement urgent, vice versa. If such feedback factor could be set to high level according to the scenario, it would contribute to the power consumption and UI response.

If no USF sysfs inode is set, and no screen on or off event, adjust_task_pred_demand shall not be invoked. Once sched_usf_up_l0_r/down_r/non_ux_r be set, adjust_task_pred_demand_impl shall be called back to update settings according to high level scenario identification.

We can get about 17% mean power consumption save at listening to music with speaker on "screen off" scenario, as below statistical data from 7766 XiaoMi devices for two weeks with sched_usf_non_ux_r be set:

        day1         day2         day3         day4
count   7766.000000  7766.000000  7766.000000  7766.000000
mean    88.035525    85.500282    83.829305    86.054997
std     111.049980   108.258834   107.562583   108.558240
min     0.099000     0.037000     0.067000     0.045000
25%     34.765500    34.021750    34.101500    34.423000
50%     54.950000    55.286500    54.189500    54.248500
75%     95.954000    93.942000    91.738000    94.0592500
80%     114.675000   107.430000   106.378000   108.673000
85%     137.851000   129.511000   127.156500   131.750750
90%     179.669000   170.208500   164.027000   172.348000
95%     272.395000   257.845500   247.750500   263.275750
98%     399.034500   412.170400   391.484000   402.835600
        day5         day6        day7         day8
count   7766.000000  7766.00000  7766.000000  7766.000000
mean    82.532677    79.21923    77.611380    81.075081
std     104.870079   101.34819   103.140037   97.506221
min     0.051000     0.02900     0.007000     0.068000
25%     32.873000    33.44400    31.965500    33.863500
50%     52.180500    51.56550    50.806500    53.080000
75%     90.905750    86.82625    83.859250    89.973000
80%     105.455000   99.64700    97.271000    104.225000
85%     128.300000   118.47825   116.570250   126.648250
90%     166.647500   149.18000   150.649500   161.087000
95%     247.208500   224.36050   226.380000   245.291250
98%     393.002000   347.92060   369.791800   378.778600
        day9         day10        day11        day12
count   7766.000000  7766.000000  7766.000000  7766.000000
mean    79.989170    83.859417    78.032930    77.060542
std     104.226122   108.893043   102.561715   99.844276
min     0.118000     0.017000     0.028000     0.039000
25%     32.056250    33.454500    31.176250    30.897750
50%     51.506000    54.056000    48.969500    49.069000
75%     88.513500    92.953500    83.506750    84.096000
80%     102.876000   107.845000   97.717000    98.073000
85%     124.363000   128.288000   118.366500   116.869250
90%     160.557000   167.084000   154.342500   148.187500
95%     231.149000   242.925750   236.759000   228.131250
98%     367.206600   388.619100   385.269100   376.541600
        day13        day14
count   7766.000000  7766.000000
mean    75.528036    73.702878
std     90.750594    86.796016
min     0.066000     0.054000
25%     31.170500    31.608500
50%     48.758500    49.215000
75%     84.522750    83.053000
80%     97.879000    94.875000
85%     116.680250   113.573750
90%     149.083500   144.089500
95%     226.177750   211.488750
98%     347.011100   331.317100
---
 drivers/staging/Kconfig          |   2 +
 drivers/staging/Makefile         |   1 +
 drivers/staging/fbsched/Kconfig  |  10 ++
 drivers/staging/fbsched/Makefile |   2 +
 drivers/staging/fbsched/usf.c    | 351 +++++++++++++++++++++++++++++++++++++++
 kernel/sched/cpufreq_schedutil.c |  11 +-
 6 files changed, 376 insertions(+), 1 deletion(-)
 create mode 100644 drivers/staging/fbsched/Kconfig
 create mode 100644 drivers/staging/fbsched/Makefile
 create mode 100644 drivers/staging/fbsched/usf.c
drivers/staging/Kconfig +2 -0
index 4ec5528..05b231e 100644
@@ -120,4 +120,6 @@ source "drivers/staging/qlge/Kconfig"
 
 source "drivers/staging/wfx/Kconfig"
 
+source "drivers/staging/fbsched/Kconfig"
+
 endif # STAGING
drivers/staging/Makefile +1 -0
index 4d34198..e016ec6 100644
@@ -50,3 +50,4 @@ obj-$(CONFIG_FIELDBUS_DEV)     += fieldbus/
 obj-$(CONFIG_KPC2000)		+= kpc2000/
 obj-$(CONFIG_QLGE)		+= qlge/
 obj-$(CONFIG_WFX)		+= wfx/
+obj-$(CONFIG_SCHED_USF)		+= fbsched/
drivers/staging/fbsched/Kconfig new file +10 -0
new file mode 100644
index 0000000..954b6af
@@ -0,0 +1,10 @@
+config SCHED_USF
+	bool "User Sensitive Factors for Scheduler"
+	depends on CPU_FREQ_GOV_SCHEDUTIL && FB
+	help
+	  Select this option to enable the adjustment on the cpufreq with
+	  the user sensitive factors on schedule. It is special for portable
+	  equipment which more power care and quick response requirement on
+	  screen on.
+
+	  If unsure, say N.
drivers/staging/fbsched/Makefile new file +2 -0
new file mode 100644
index 0000000..f56aa6c
@@ -0,0 +1,2 @@
+LINUXINCLUDE += -include $(srctree)/kernel/sched/sched.h
+obj-$(CONFIG_SCHED_USF) += usf.o
drivers/staging/fbsched/usf.c new file +34 -0
new file mode 100644
index 0000000..8582992
@@ -0,0 +1,351 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2020 XiaoMi Inc.
+ * Author: Yang Dongdong <[email protected]>
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+ * See http://www.gnu.org/licenses/gpl-2.0.html for more details.
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/platform_device.h>
+#include <linux/kthread.h>
+#include <linux/cpu.h>
+#include <linux/sysfs.h>
+#include <linux/kthread.h>
+#include <linux/kobject.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/kallsyms.h>
+#include <linux/debugfs.h>
+#include <linux/fb.h>
+#include <linux/notifier.h>
+
+#define BOOST_MIN_V -100
+#define BOOST_MAX_V 100
+#define LEVEL_TOP 3
+

drivers/staging/fbsched/usf.c line 44

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+#define USF_TAG	"[usf_sched]"
+
+extern void (*adjust_task_pred_demand)(int cpuid,
+				       unsigned long *util,
+				       struct rq *rq);
+DEFINE_PER_CPU(unsigned long[PID_MAX_LIMIT], task_hist_nivcsw);
+
+static struct {
+	bool is_sched_usf_enabled;
+	int enable_debug;

Make this a bool. Use true false.

drivers/staging/fbsched/usf.c line 45

Quoted 1 line
drivers/staging/fbsched/usf.c new file +1 -0
+	int is_screen_on;

Bool

Quoted 73 lines
drivers/staging/fbsched/usf.c new file +73 -0
+	struct kobject *kobj;
+	struct dentry *debugfs_entry;
+	int sysctl_sched_usf_up_l0;
+	int sysctl_sched_usf_down;
+	int sysctl_sched_usf_non_ux;
+	int usf_up_l0;
+	int usf_down;
+	int usf_non_ux;
+} usf_vdev;
+
+static void adjust_task_pred_demand_impl(int cpuid,
+					 unsigned long *util,
+					 struct rq *rq)
+{
+	/* sysctl_sched_latency/sysctl_sched_min_granularity */
+	u32 bl_sw_num = 3;
+
+	if (!usf_vdev.is_sched_usf_enabled || !rq || !rq->curr)
+		return;
+
+	if (usf_vdev.is_screen_on) {
+		if (rq->curr->nivcsw >
+		    (per_cpu(task_hist_nivcsw, cpuid)[rq->curr->pid]
+		     + bl_sw_num + 1)) {
+			(*util) += (*util) >> usf_vdev.usf_up_l0;
+		} else if (rq->curr->nivcsw <
+			   (per_cpu(task_hist_nivcsw, cpuid)[rq->curr->pid]
+			    + bl_sw_num - 1) && (rq->nr_running < bl_sw_num)) {
+			(*util) >>= usf_vdev.usf_down;
+		}
+		per_cpu(task_hist_nivcsw, cpuid)[rq->curr->pid] =
+		    rq->curr->nivcsw;
+	} else if (rq->curr->mm) {
+		(*util) >>= usf_vdev.usf_non_ux;
+	}
+
+	if (unlikely(usf_vdev.enable_debug))
+		trace_printk
+		    ("%s: cpu_id=%d non_ux=%d usf_up=%d usf_down=%d util=%lu\n",
+		     USF_TAG, cpuid, usf_vdev.usf_non_ux,
+		     usf_vdev.usf_up_l0, usf_vdev.usf_down, *util);
+}
+
+static int usf_lcd_notifier(struct notifier_block *nb,
+			    unsigned long val, void *data)
+{
+	struct fb_event *evdata = data;
+	unsigned int blank;
+
+	if (!evdata)
+		return 0;
+
+	if (val != FB_EVENT_BLANK)
+		return 0;
+
+	if (evdata->data && val == FB_EVENT_BLANK) {
+		blank = *(int *)(evdata->data);
+
+		switch (blank) {
+		case FB_BLANK_POWERDOWN:
+			usf_vdev.is_screen_on = 0;
+			if (usf_vdev.sysctl_sched_usf_non_ux != 0)
+				adjust_task_pred_demand =
+				    &adjust_task_pred_demand_impl;
+			else
+				adjust_task_pred_demand = NULL;
+
+			break;
+
+		case FB_BLANK_UNBLANK:
+			usf_vdev.is_screen_on = 1;
+			if (usf_vdev.sysctl_sched_usf_up_l0 != 0 ||
+			    usf_vdev.sysctl_sched_usf_down != 0)

drivers/staging/fbsched/usf.c line 128

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+				adjust_task_pred_demand =
+				    &adjust_task_pred_demand_impl;
+			else
+				adjust_task_pred_demand = NULL;
+			break;
+		default:
+			break;
+		}
+
+		usf_vdev.is_sched_usf_enabled = 1;

s/1/true/

Quoted 12 lines
drivers/staging/fbsched/usf.c new file +12 -0
+		if (usf_vdev.enable_debug)
+			trace_printk("%s : usf_vdev.is_screen_on:%d\n",
+				     __func__, usf_vdev.is_screen_on);
+	}
+	return NOTIFY_OK;
+}
+
+static struct notifier_block usf_lcd_nb = {
+	.notifier_call = usf_lcd_notifier,
+	.priority = INT_MAX,
+};
+

drivers/staging/fbsched/usf.c line 150

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+static ssize_t store_sched_usf_up_l0_r(struct kobject *kobj,
+				       struct kobj_attribute *attr,
+				       const char *buf, size_t count)
+{
+	int val = 0;
+	int ret;
+
+	ret = kstrtoint(buf, 0, &val);
+	if (ret) {
+		pr_err(USF_TAG "set state fail ret=%d\n", ret);

Don't print an error. Just return ret;

drivers/staging/fbsched/usf.c line 160

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+		return ret;
+	}
+
+	if ((val >= 0) && (val <= BOOST_MAX_V)) {
+		usf_vdev.sysctl_sched_usf_up_l0 = val;
+		usf_vdev.usf_up_l0 = LEVEL_TOP -
+				DIV_ROUND_UP(val, BOOST_MAX_V / 2);
+	} else {
+		pr_err(USF_TAG "%d should fall into [%d %d]",
+		       val, 0, BOOST_MAX_V);

Flip this test around and return an error code on invalide input:

	if (val < 0 || val > BOOST_MAX_V)
		return -EINVAL;

Print an error here if you want.

drivers/staging/fbsched/usf.c line 166

Quoted 6 lines
drivers/staging/fbsched/usf.c new file +6 -0
+	}
+	if ((usf_vdev.sysctl_sched_usf_up_l0 == 0) &&
+	    (usf_vdev.sysctl_sched_usf_down == 0))
+		adjust_task_pred_demand = NULL;
+	else
+		adjust_task_pred_demand = &adjust_task_pred_demand_impl;

There needs to be some locking around adjust_task_pred_demand.

Quoted 13 lines
drivers/staging/fbsched/usf.c new file +13 -0
+
+	return count;
+}
+
+static ssize_t store_sched_usf_down_r(struct kobject *kobj,
+				      struct kobj_attribute *attr,
+				      const char *buf, size_t count)
+{
+	int val = 0;
+	int ret;
+
+	ret = kstrtoint(buf, 0, &val);
+	if (ret) {

drivers/staging/fbsched/usf.c line 189

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+		pr_err(USF_TAG "set state fail ret=%d\n", ret);
+		return ret;
+	}
+
+	if ((val >= BOOST_MIN_V) && (val <= 0)) {
+		usf_vdev.sysctl_sched_usf_down = val;
+		usf_vdev.usf_down = DIV_ROUND_UP(-val, -BOOST_MIN_V / 2);
+	} else {
+		pr_err(USF_TAG "%d should fall into [%d %d]",
+		       val, BOOST_MIN_V, 0);

All the same comments as above.

Quoted 26 lines
drivers/staging/fbsched/usf.c new file +26 -0
+	}i
+	if ((usf_vdev.sysctl_sched_usf_up_l0 == 0) &&
+	    (usf_vdev.sysctl_sched_usf_down == 0))
+		adjust_task_pred_demand = NULL;
+	else
+		adjust_task_pred_demand = &adjust_task_pred_demand_impl;
+
+	return count;
+}
+
+static ssize_t store_sched_usf_non_ux_r(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					const char *buf, size_t count)
+{
+	int val = 0;
+	int ret;
+
+	ret = kstrtoint(buf, 0, &val);
+	if (ret) {
+		pr_err(USF_TAG "set state fail ret=%d\n", ret);
+		return ret;
+	}
+
+	if ((val >= BOOST_MIN_V) && (val <= 0)) {
+		usf_vdev.sysctl_sched_usf_non_ux = val;
+		usf_vdev.usf_non_ux = DIV_ROUND_UP(-val, -BOOST_MIN_V / 2);

drivers/staging/fbsched/usf.c line 225

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+	} else {
+		pr_err(USF_TAG "%d should fall into [%d %d]",
+		       val, BOOST_MIN_V, 0);
+	}
+	if (usf_vdev.sysctl_sched_usf_non_ux == 0)
+		adjust_task_pred_demand = NULL;
+	else
+		adjust_task_pred_demand = &adjust_task_pred_demand_impl;
+
+	return count;

Same.

Quoted 6 lines
drivers/staging/fbsched/usf.c new file +6 -0
+}
+
+#define usf_attr_rw(_name)						\
+static struct kobj_attribute _name =					\
+__ATTR(_name, 0664, show_##_name, store_##_name)
+

drivers/staging/fbsched/usf.c line 241

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+#define usf_show_node(_name, _value)					\
+static ssize_t show_##_name						\
+(struct kobject *kobj, struct kobj_attribute *attr,  char *buf)		\
+{									\
+	unsigned int len = 0;						\
+	unsigned int max_len = 8;					\
+	len +=								\
+	    snprintf(buf + len, max_len - len, "%d",			\
+		     usf_vdev.sysctl_##_value);				\
+	return len;							\

Just do this:

	return sprintf(buf, "%d", usf_vdev.sysctl_##_value);
Quoted 30 lines
drivers/staging/fbsched/usf.c new file +30 -0
+}
+
+usf_show_node(sched_usf_up_l0_r, sched_usf_up_l0);
+usf_show_node(sched_usf_down_r, sched_usf_down);
+usf_show_node(sched_usf_non_ux_r, sched_usf_non_ux);
+
+usf_attr_rw(sched_usf_up_l0_r);
+usf_attr_rw(sched_usf_down_r);
+usf_attr_rw(sched_usf_non_ux_r);
+
+static struct attribute *sched_attrs[] = {
+	&sched_usf_up_l0_r.attr,
+	&sched_usf_down_r.attr,
+	&sched_usf_non_ux_r.attr,
+	NULL,
+};
+
+static struct attribute_group sched_attr_group = {
+	.attrs = sched_attrs,
+};
+
+static int usf_dbg_get(void *data, u64 *val)
+{
+	*val = (u64)usf_vdev.enable_debug;
+
+	return 0;
+}
+
+static int usf_dbg_set(void *data, u64 val)
+{

drivers/staging/fbsched/usf.c line 281

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+	usf_vdev.enable_debug = !!val;
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(usf_dbg_fops, usf_dbg_get,
+			usf_dbg_set, "%llu\n");
+
+static int __init intera_monitor_init(void)
+{
+	int res = -1;

Don't initialize this. It just silences uninitialized variable warnings so it will lead to accidental bugs.

Quoted 11 lines
drivers/staging/fbsched/usf.c new file +11 -0
+	struct attribute_group *attr_group;
+
+	res = fb_register_client(&usf_lcd_nb);
+	if (res < 0) {
+		pr_err("Failed to register usf_lcd_nb!\n");
+		return res;
+	}
+
+	/*
+	 * create a sched_usf in cpu_subsys:
+	 * /sys/devices/system/cpu/sched_usf/...

drivers/staging/fbsched/usf.c line 302

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+	 */
+	attr_group = &sched_attr_group;
+	usf_vdev.kobj = kobject_create_and_add("sched_usf",
+					       &cpu_subsys.dev_root->kobj);
+
+	if (!usf_vdev.kobj) {
+		pr_err("Failed to create sched_usf!\n");
+		res = -ENOMEM;
+		goto out;
+	} else {

Delete the else. Pull the code in one indent level.

drivers/staging/fbsched/usf.c line 307

Quoted 5 lines
drivers/staging/fbsched/usf.c new file +5 -0
+		res = sysfs_create_group(usf_vdev.kobj, attr_group);
+		if (res) {
+			kobject_put(usf_vdev.kobj);
+			goto out;
+		} else {

Delete this else statement.

drivers/staging/fbsched/usf.c line 317

Quoted 10 lines
drivers/staging/fbsched/usf.c new file +10 -0
+			kobject_uevent(usf_vdev.kobj, KOBJ_ADD);
+		}
+	}
+
+	usf_vdev.enable_debug = 0;
+	usf_vdev.debugfs_entry = debugfs_create_file("usf_dbg",
+						     0660, NULL, NULL,
+						     &usf_dbg_fops);
+	if (!usf_vdev.debugfs_entry)
+		pr_err("Failed to create usf_dbg!\n");

debugfs never returns NULL. Generally debugfs_create_file() returns are not supposed to be error checked. Just delete the check.

drivers/staging/fbsched/usf.c line 323

Quoted 6 lines
drivers/staging/fbsched/usf.c new file +6 -0
+
+	usf_vdev.is_sched_usf_enabled = 0;
+	usf_vdev.sysctl_sched_usf_up_l0 = 0;
+	usf_vdev.sysctl_sched_usf_down = 0;
+	usf_vdev.sysctl_sched_usf_non_ux = 0;
+	adjust_task_pred_demand = NULL;

No need to set all these to zero/NULL. It's already zero.

drivers/staging/fbsched/usf.c line 335

Quoted 12 lines
drivers/staging/fbsched/usf.c new file +12 -0
+
+	return 0;
+out:
+	fb_unregister_client(&usf_lcd_nb);
+	return res;
+}
+
+module_init(intera_monitor_init);
+
+static void __exit intera_monitor_exit(void)
+{
+	if (usf_vdev.kobj)

Delet this impossible condition.

drivers/staging/fbsched/usf.c line 337

Quoted 2 lines
drivers/staging/fbsched/usf.c new file +2 -0
+		kobject_put(usf_vdev.kobj);
+	if (usf_vdev.debugfs_entry)

Delete this unnecessary condition.

drivers/staging/fbsched/usf.c line 344

Quoted 7 lines
drivers/staging/fbsched/usf.c new file +7 -0
+		debugfs_remove(usf_vdev.debugfs_entry);
+	fb_unregister_client(&usf_lcd_nb);
+	usf_vdev.is_sched_usf_enabled = 0;
+	usf_vdev.sysctl_sched_usf_up_l0 = 0;
+	usf_vdev.sysctl_sched_usf_down = 0;
+	usf_vdev.sysctl_sched_usf_non_ux = 0;
+	adjust_task_pred_demand = NULL;

No need to set all these to zero/NULL.

regards,
dan carpenter

maintainer
#7 lore raw
In reply to Dongdong Yang
Quoted 25 lines · Dongdong Yang

From: Dongdong Yang <[email protected]>

This patch provides USF(User Sensitive Feedback factor) auxiliary cpufreq governor to support high level layer sysfs inodes setting for utils adjustment purpose from the identified scenario on portable equipment. Because the power consumption and UI response are more cared for by portable equipment users. And the "screen off" status stands for no request from the user, however, the kernel is still expected to notify the user in time on modem, network or powerkey events occur. USF provides "sched_usf_non_ux_r" sysfs inode to cut down the utils from user space tasks according to high level scenario. In addition, it usually hints more cpufreq demand that the preemptive counts of the tasks on the cpu burst and over the user expecting completed time such as the ratio sysctl_sched_latency to sysctl_sched_min_granularity on "screen on" status, which more likely with more UI. The sysfs inodes "sched_usf_up_l0_r" and "sched_usf_down_r" have been provided to adjust the utils according to high level identified scenario to alloc the cpufreq in time.

Dongdong Yang (1):
sched: Provide USF for portable equipment.

Quoted 10 lines
 drivers/staging/Kconfig          |   2 +
 drivers/staging/Makefile         |   1 +
 drivers/staging/fbsched/Kconfig  |  10 ++
 drivers/staging/fbsched/Makefile |   2 +
 drivers/staging/fbsched/usf.c    | 351 +++++++++++++++++++++++++++++++++++++++
 kernel/sched/cpufreq_schedutil.c |  11 +-
 6 files changed, 376 insertions(+), 1 deletion(-)
 create mode 100644 drivers/staging/fbsched/Kconfig
 create mode 100644 drivers/staging/fbsched/Makefile
 create mode 100644 drivers/staging/fbsched/usf.c

For new staging drivers/code, we need a TODO file that lists what remains to be done on the code to get it out of staging/

I don't see any good reason why this has to go to staging now, what prevents it from being merged to the "real" part of the kernel today?

thanks,

greg k-h