pwm: mxs: set pwm_chip can_sleep flag

The .config() calls clk_get_rate() which might sleep, so we need to set pwm_chip can_sleep flag. Otherwise, we see the following warning when using pwm driven heartbeat led.

WARNING: CPU: 0 PID: 0 at kernel/locking/mutex.c:856 mutex_trylock+0x184/0x1a4()
DEBUG_LOCKS_WARN_ON(in_interrupt())
Modules linked in:
CPU: 0 PID: 0 Comm: swapper Not tainted 3.14.0-rc5 #18
[<c0015420>] (unwind_backtrace) from [<c0012cb0>] (show_stack+0x10/0x14)
[<c0012cb0>] (show_stack) from [<c001daf8>] (warn_slowpath_common+0x6c/0x8c)
[<c001daf8>] (warn_slowpath_common) from [<c001dbac>] (warn_slowpath_fmt+0x30/0x40)
[<c001dbac>] (warn_slowpath_fmt) from [<c045df74>] (mutex_trylock+0x184/0x1a4)
[<c045df74>] (mutex_trylock) from [<c0360950>] (clk_prepare_lock+0xc/0xec)
[<c0360950>] (clk_prepare_lock) from [<c0362020>] (clk_get_rate+0xc/0x68)
[<c0362020>] (clk_get_rate) from [<c028d07c>] (mxs_pwm_config+0x20/0x198)
[<c028d07c>] (mxs_pwm_config) from [<c028bde8>] (pwm_config+0x60/0x70)
[<c028bde8>] (pwm_config) from [<c034b61c>] (__led_pwm_set+0x1c/0x3c)
[<c034b61c>] (__led_pwm_set) from [<c034bc3c>] (led_heartbeat_function+0x70/0x110)
[<c034bc3c>] (led_heartbeat_function) from [<c00292f0>] (call_timer_fn+0x7c/0x164)
[<c00292f0>] (call_timer_fn) from [<c00295c8>] (run_timer_softirq+0x1f0/0x260)
[<c00295c8>] (run_timer_softirq) from [<c002255c>] (__do_softirq+0xc4/0x2f0)
[<c002255c>] (__do_softirq) from [<c0022890>] (irq_exit+0xa4/0x10c)
[<c0022890>] (irq_exit) from [<c0010240>] (handle_IRQ+0x34/0x84)
[<c0010240>] (handle_IRQ) from [<c0013524>] (__irq_svc+0x44/0x54)
[<c0013524>] (__irq_svc) from [<c00107f8>] (arch_cpu_idle+0x40/0x48)
[<c00107f8>] (arch_cpu_idle) from [<c005deb8>] (cpu_startup_entry+0x70/0x198)
[<c005deb8>] (cpu_startup_entry) from [<c060aac8>] (start_kernel+0x2a8/0x2f8)

Reported-by Stefan Wahren <[email protected]>

Signed-off-by Shawn Guo <[email protected]>

---
 drivers/pwm/pwm-mxs.c | 1 +
 1 file changed, 1 insertion(+)

drivers/pwm/pwm-mxs.c +0 -0

index 9475bc7..4f1bb4e 100644

Hi Shawn,

Am 08.04.2014 13:29, schrieb Shawn Guo:

Quoted 39 lines

The .config() calls clk_get_rate() which might sleep, so we need to set pwm_chip can_sleep flag. Otherwise, we see the following warning when using pwm driven heartbeat led.

WARNING: CPU: 0 PID: 0 at kernel/locking/mutex.c:856 mutex_trylock+0x184/0x1a4()
DEBUG_LOCKS_WARN_ON(in_interrupt())
Modules linked in:
CPU: 0 PID: 0 Comm: swapper Not tainted 3.14.0-rc5 #18
[<c0015420>] (unwind_backtrace) from [<c0012cb0>] (show_stack+0x10/0x14)
[<c0012cb0>] (show_stack) from [<c001daf8>] (warn_slowpath_common+0x6c/0x8c)
[<c001daf8>] (warn_slowpath_common) from [<c001dbac>] (warn_slowpath_fmt+0x30/0x40)
[<c001dbac>] (warn_slowpath_fmt) from [<c045df74>] (mutex_trylock+0x184/0x1a4)
[<c045df74>] (mutex_trylock) from [<c0360950>] (clk_prepare_lock+0xc/0xec)
[<c0360950>] (clk_prepare_lock) from [<c0362020>] (clk_get_rate+0xc/0x68)
[<c0362020>] (clk_get_rate) from [<c028d07c>] (mxs_pwm_config+0x20/0x198)
[<c028d07c>] (mxs_pwm_config) from [<c028bde8>] (pwm_config+0x60/0x70)
[<c028bde8>] (pwm_config) from [<c034b61c>] (__led_pwm_set+0x1c/0x3c)
[<c034b61c>] (__led_pwm_set) from [<c034bc3c>] (led_heartbeat_function+0x70/0x110)
[<c034bc3c>] (led_heartbeat_function) from [<c00292f0>] (call_timer_fn+0x7c/0x164)
[<c00292f0>] (call_timer_fn) from [<c00295c8>] (run_timer_softirq+0x1f0/0x260)
[<c00295c8>] (run_timer_softirq) from [<c002255c>] (__do_softirq+0xc4/0x2f0)
[<c002255c>] (__do_softirq) from [<c0022890>] (irq_exit+0xa4/0x10c)
[<c0022890>] (irq_exit) from [<c0010240>] (handle_IRQ+0x34/0x84)
[<c0010240>] (handle_IRQ) from [<c0013524>] (__irq_svc+0x44/0x54)
[<c0013524>] (__irq_svc) from [<c00107f8>] (arch_cpu_idle+0x40/0x48)
[<c00107f8>] (arch_cpu_idle) from [<c005deb8>] (cpu_startup_entry+0x70/0x198)
[<c005deb8>] (cpu_startup_entry) from [<c060aac8>] (start_kernel+0x2a8/0x2f8)

Reported-by Stefan Wahren <[email protected]>

Signed-off-by Shawn Guo <[email protected]>

---
 drivers/pwm/pwm-mxs.c | 1 +
 1 file changed, 1 insertion(+)

drivers/pwm/pwm-mxs.c +0 -0

index 9475bc7..4f1bb4e 100644

drivers/pwm/pwm-mxs.c static int mxs_pwm_probe(struct platform_device *pdev) · line 153

Quoted 7 lines

drivers/pwm/pwm-mxs.c +1 -0

@@ -147,6 +147,7 @@ static int mxs_pwm_probe(struct platform_device *pdev)
 	mxs->chip.dev = &pdev->dev;
 	mxs->chip.ops = &mxs_pwm_ops;
 	mxs->chip.base = -1;
+	mxs->chip.can_sleep = true;
 	ret = of_property_read_u32(np, "fsl,pwm-number", &mxs->chip.npwm);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed to get pwm number: %d\n", ret);

thanks for the patch. I've tested it with our i.MX28 board and the warning above never came. So it works.
Unfortunately the led still don't behave as expected. If i set the led trigger to heartbeat the led goes on and stays in this state.

May be this has something to do with the following discussion

http://comments.gmane.org/gmane.linux.leds/208

Quoted 54 lines

Hi Shawn,

Am 08.04.2014 13:29, schrieb Shawn Guo:

Quoted 39 lines

The .config() calls clk_get_rate() which might sleep, so we need to set pwm_chip can_sleep flag. Otherwise, we see the following warning when using pwm driven heartbeat led.

WARNING: CPU: 0 PID: 0 at kernel/locking/mutex.c:856 mutex_trylock+0x184/0x1a4()
DEBUG_LOCKS_WARN_ON(in_interrupt())
Modules linked in:
CPU: 0 PID: 0 Comm: swapper Not tainted 3.14.0-rc5 #18
[<c0015420>] (unwind_backtrace) from [<c0012cb0>] (show_stack+0x10/0x14)
[<c0012cb0>] (show_stack) from [<c001daf8>] (warn_slowpath_common+0x6c/0x8c)
[<c001daf8>] (warn_slowpath_common) from [<c001dbac>] (warn_slowpath_fmt+0x30/0x40)
[<c001dbac>] (warn_slowpath_fmt) from [<c045df74>] (mutex_trylock+0x184/0x1a4)
[<c045df74>] (mutex_trylock) from [<c0360950>] (clk_prepare_lock+0xc/0xec)
[<c0360950>] (clk_prepare_lock) from [<c0362020>] (clk_get_rate+0xc/0x68)
[<c0362020>] (clk_get_rate) from [<c028d07c>] (mxs_pwm_config+0x20/0x198)
[<c028d07c>] (mxs_pwm_config) from [<c028bde8>] (pwm_config+0x60/0x70)
[<c028bde8>] (pwm_config) from [<c034b61c>] (__led_pwm_set+0x1c/0x3c)
[<c034b61c>] (__led_pwm_set) from [<c034bc3c>] (led_heartbeat_function+0x70/0x110)
[<c034bc3c>] (led_heartbeat_function) from [<c00292f0>] (call_timer_fn+0x7c/0x164)
[<c00292f0>] (call_timer_fn) from [<c00295c8>] (run_timer_softirq+0x1f0/0x260)
[<c00295c8>] (run_timer_softirq) from [<c002255c>] (__do_softirq+0xc4/0x2f0)
[<c002255c>] (__do_softirq) from [<c0022890>] (irq_exit+0xa4/0x10c)
[<c0022890>] (irq_exit) from [<c0010240>] (handle_IRQ+0x34/0x84)
[<c0010240>] (handle_IRQ) from [<c0013524>] (__irq_svc+0x44/0x54)
[<c0013524>] (__irq_svc) from [<c00107f8>] (arch_cpu_idle+0x40/0x48)
[<c00107f8>] (arch_cpu_idle) from [<c005deb8>] (cpu_startup_entry+0x70/0x198)
[<c005deb8>] (cpu_startup_entry) from [<c060aac8>] (start_kernel+0x2a8/0x2f8)

Reported-by Stefan Wahren <[email protected]>

Signed-off-by Shawn Guo <[email protected]>

---
 drivers/pwm/pwm-mxs.c | 1 +
 1 file changed, 1 insertion(+)

drivers/pwm/pwm-mxs.c +0 -0

index 9475bc7..4f1bb4e 100644

drivers/pwm/pwm-mxs.c static int mxs_pwm_probe(struct platform_device *pdev) · line 153

Quoted 7 lines

drivers/pwm/pwm-mxs.c +1 -0

@@ -147,6 +147,7 @@ static int mxs_pwm_probe(struct platform_device *pdev)
 	mxs->chip.dev = &pdev->dev;
 	mxs->chip.ops = &mxs_pwm_ops;
 	mxs->chip.base = -1;
+	mxs->chip.can_sleep = true;
 	ret = of_property_read_u32(np, "fsl,pwm-number", &mxs->chip.npwm);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed to get pwm number: %d\n", ret);

thanks for the patch. I've tested it with our i.MX28 board and the warning above never came. So it works.
Unfortunately the led still don't behave as expected. If i set the led trigger to heartbeat the led goes on and stays in this state.

Quoted 4 lines

May be this has something to do with the following discussion

http://comments.gmane.org/gmane.linux.leds/208

This may be but the solution is not correct we want to keep disabling the pwm when brightness is 0. Can you try the latest series from Russell:

http://thread.gmane.org/gmane.linux.leds/579 and set active_low but keep the pwm polarity normal.

Regards,

i applied these 5 patches, but it doesn't fix the problem.

Here are my results for the pwm driver led on i.MX28:

version | trigger: none | trigger: heartbeat
-----------------------------------------------------------------------------
3.14 | OK | NOT OK (WARNING, LED
permanent on)
+ shawn's patch | OK | NOT OK (LED
permanent on)
+ russell's patches | OK | NOT OK (LED
permanent on)
+ active-low | NOT OK (inverse logic) | NOT OK (LED
permanent on)

Russell's patches work as expected, but have not influence on the heartbeat.

Quoted 15 lines · Stefan Wahren

i applied these 5 patches, but it doesn't fix the problem.

Here are my results for the pwm driver led on i.MX28:

version | trigger: none | trigger: heartbeat
-----------------------------------------------------------------------------
3.14 | OK | NOT OK (WARNING, LED
permanent on)
+ shawn's patch | OK | NOT OK (LED
permanent on)
+ russell's patches | OK | NOT OK (LED
permanent on)
+ active-low | NOT OK (inverse logic) | NOT OK (LED
permanent on)

Quoted 2 lines

Russell's patches work as expected, but have not influence on the heartbeat.

Yeah, I got it wrong or explained badly.

So, if without Russell's patches, your pwm is not inversed, then you would have to set active_low and inverse the pwm.
If it was inversed, then you'll have to set active_low and let the pwm with the normal polarity. I know this is a bit contrived but this was exactly my point why we shouldn't call it active_low.

Can you try to switch the polarity when setting active_low ?

I'm not home, else I'll try on my i.mx28 boards...

Quoted 1 line · Alexandre Belloni

I'm not home, else I'll try on my i.mx28 boards...

I just tried heartbeat here, on iMX6, and it works fine. That's with this in the DT:

Quoted 5 lines · Russell King - ARM Linux

Quoted 1 line · Alexandre Belloni

I'm not home, else I'll try on my i.mx28 boards...

I just tried heartbeat here, on iMX6, and it works fine. That's with this in the DT:

FYI. i.MX28 is known as 'mxs' sub-architecture and uses pwm-mxs driver rather than pwm-imx which is the one for i.MX6.

Am 08.04.2014 13:29, schrieb Shawn Guo:

Quoted 39 lines

The .config() calls clk_get_rate() which might sleep, so we need to set pwm_chip can_sleep flag. Otherwise, we see the following warning when using pwm driven heartbeat led.

WARNING: CPU: 0 PID: 0 at kernel/locking/mutex.c:856 mutex_trylock+0x184/0x1a4()
DEBUG_LOCKS_WARN_ON(in_interrupt())
Modules linked in:
CPU: 0 PID: 0 Comm: swapper Not tainted 3.14.0-rc5 #18
[<c0015420>] (unwind_backtrace) from [<c0012cb0>] (show_stack+0x10/0x14)
[<c0012cb0>] (show_stack) from [<c001daf8>] (warn_slowpath_common+0x6c/0x8c)
[<c001daf8>] (warn_slowpath_common) from [<c001dbac>] (warn_slowpath_fmt+0x30/0x40)
[<c001dbac>] (warn_slowpath_fmt) from [<c045df74>] (mutex_trylock+0x184/0x1a4)
[<c045df74>] (mutex_trylock) from [<c0360950>] (clk_prepare_lock+0xc/0xec)
[<c0360950>] (clk_prepare_lock) from [<c0362020>] (clk_get_rate+0xc/0x68)
[<c0362020>] (clk_get_rate) from [<c028d07c>] (mxs_pwm_config+0x20/0x198)
[<c028d07c>] (mxs_pwm_config) from [<c028bde8>] (pwm_config+0x60/0x70)
[<c028bde8>] (pwm_config) from [<c034b61c>] (__led_pwm_set+0x1c/0x3c)
[<c034b61c>] (__led_pwm_set) from [<c034bc3c>] (led_heartbeat_function+0x70/0x110)
[<c034bc3c>] (led_heartbeat_function) from [<c00292f0>] (call_timer_fn+0x7c/0x164)
[<c00292f0>] (call_timer_fn) from [<c00295c8>] (run_timer_softirq+0x1f0/0x260)
[<c00295c8>] (run_timer_softirq) from [<c002255c>] (__do_softirq+0xc4/0x2f0)
[<c002255c>] (__do_softirq) from [<c0022890>] (irq_exit+0xa4/0x10c)
[<c0022890>] (irq_exit) from [<c0010240>] (handle_IRQ+0x34/0x84)
[<c0010240>] (handle_IRQ) from [<c0013524>] (__irq_svc+0x44/0x54)
[<c0013524>] (__irq_svc) from [<c00107f8>] (arch_cpu_idle+0x40/0x48)
[<c00107f8>] (arch_cpu_idle) from [<c005deb8>] (cpu_startup_entry+0x70/0x198)
[<c005deb8>] (cpu_startup_entry) from [<c060aac8>] (start_kernel+0x2a8/0x2f8)

Reported-by Stefan Wahren <[email protected]>

Signed-off-by Shawn Guo <[email protected]>

---
 drivers/pwm/pwm-mxs.c | 1 +
 1 file changed, 1 insertion(+)

drivers/pwm/pwm-mxs.c +0 -0

index 9475bc7..4f1bb4e 100644

drivers/pwm/pwm-mxs.c static int mxs_pwm_probe(struct platform_device *pdev) · line 153

Quoted 7 lines

drivers/pwm/pwm-mxs.c +1 -0

@@ -147,6 +147,7 @@ static int mxs_pwm_probe(struct platform_device *pdev)
 	mxs->chip.dev = &pdev->dev;
 	mxs->chip.ops = &mxs_pwm_ops;
 	mxs->chip.base = -1;
+	mxs->chip.can_sleep = true;
 	ret = of_property_read_u32(np, "fsl,pwm-number", &mxs->chip.npwm);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed to get pwm number: %d\n", ret);

thanks for the patch. I've tested it with our i.MX28 board and the warning above never came. So it works.
Unfortunately the led still don't behave as expected. If i set the led trigger to heartbeat the led goes on and stays in this state.

May be this has something to do with the following discussion

http://comments.gmane.org/gmane.linux.leds/208

Hello,

Quoted 53 lines · Stefan Wahren

Am 08.04.2014 13:29, schrieb Shawn Guo:

Quoted 39 lines

The .config() calls clk_get_rate() which might sleep, so we need to set pwm_chip can_sleep flag. Otherwise, we see the following warning when using pwm driven heartbeat led.

WARNING: CPU: 0 PID: 0 at kernel/locking/mutex.c:856 mutex_trylock+0x184/0x1a4()
DEBUG_LOCKS_WARN_ON(in_interrupt())
Modules linked in:
CPU: 0 PID: 0 Comm: swapper Not tainted 3.14.0-rc5 #18
[<c0015420>] (unwind_backtrace) from [<c0012cb0>] (show_stack+0x10/0x14)
[<c0012cb0>] (show_stack) from [<c001daf8>] (warn_slowpath_common+0x6c/0x8c)
[<c001daf8>] (warn_slowpath_common) from [<c001dbac>] (warn_slowpath_fmt+0x30/0x40)
[<c001dbac>] (warn_slowpath_fmt) from [<c045df74>] (mutex_trylock+0x184/0x1a4)
[<c045df74>] (mutex_trylock) from [<c0360950>] (clk_prepare_lock+0xc/0xec)
[<c0360950>] (clk_prepare_lock) from [<c0362020>] (clk_get_rate+0xc/0x68)
[<c0362020>] (clk_get_rate) from [<c028d07c>] (mxs_pwm_config+0x20/0x198)
[<c028d07c>] (mxs_pwm_config) from [<c028bde8>] (pwm_config+0x60/0x70)
[<c028bde8>] (pwm_config) from [<c034b61c>] (__led_pwm_set+0x1c/0x3c)
[<c034b61c>] (__led_pwm_set) from [<c034bc3c>] (led_heartbeat_function+0x70/0x110)
[<c034bc3c>] (led_heartbeat_function) from [<c00292f0>] (call_timer_fn+0x7c/0x164)
[<c00292f0>] (call_timer_fn) from [<c00295c8>] (run_timer_softirq+0x1f0/0x260)
[<c00295c8>] (run_timer_softirq) from [<c002255c>] (__do_softirq+0xc4/0x2f0)
[<c002255c>] (__do_softirq) from [<c0022890>] (irq_exit+0xa4/0x10c)
[<c0022890>] (irq_exit) from [<c0010240>] (handle_IRQ+0x34/0x84)
[<c0010240>] (handle_IRQ) from [<c0013524>] (__irq_svc+0x44/0x54)
[<c0013524>] (__irq_svc) from [<c00107f8>] (arch_cpu_idle+0x40/0x48)
[<c00107f8>] (arch_cpu_idle) from [<c005deb8>] (cpu_startup_entry+0x70/0x198)
[<c005deb8>] (cpu_startup_entry) from [<c060aac8>] (start_kernel+0x2a8/0x2f8)

Reported-by Stefan Wahren <[email protected]>

Signed-off-by Shawn Guo <[email protected]>

---
 drivers/pwm/pwm-mxs.c | 1 +
 1 file changed, 1 insertion(+)

drivers/pwm/pwm-mxs.c +0 -0

index 9475bc7..4f1bb4e 100644

drivers/pwm/pwm-mxs.c static int mxs_pwm_probe(struct platform_device *pdev) · line 153

Quoted 7 lines

drivers/pwm/pwm-mxs.c +1 -0

@@ -147,6 +147,7 @@ static int mxs_pwm_probe(struct platform_device *pdev)
 	mxs->chip.dev = &pdev->dev;
 	mxs->chip.ops = &mxs_pwm_ops;
 	mxs->chip.base = -1;
+	mxs->chip.can_sleep = true;
 	ret = of_property_read_u32(np, "fsl,pwm-number", &mxs->chip.npwm);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed to get pwm number: %d\n", ret);

thanks for the patch. I've tested it with our i.MX28 board and the warning above never came. So it works.
Unfortunately the led still don't behave as expected. If i set the led trigger to heartbeat the led goes on and stays in this state.

Quoted 3 lines

May be this has something to do with the following discussion

http://comments.gmane.org/gmane.linux.leds/208

I guess that's your problem, yes. Does patch 2 of my series (i.e.
http://thread.gmane.org/gmane.linux.ports.arm.kernel/282593/focus=282596)
help you?

Quoted 5 lines · Uwe Kleine-K?nig

Quoted 4 lines · Stefan Wahren

Am 08.04.2014 13:29, schrieb Shawn Guo:
Unfortunately the led still don't behave as expected. If i set the led trigger to heartbeat the led goes on and stays in this state.

Quoted 7 lines

Quoted 3 lines

May be this has something to do with the following discussion

http://comments.gmane.org/gmane.linux.leds/208

I guess that's your problem, yes. Does patch 2 of my series (i.e.
http://thread.gmane.org/gmane.linux.ports.arm.kernel/282593/focus=282596)
help you?

While I had exactly the same issue back in september and wrote the exact same patch that got rejected. I would agree with Thierry here taht we need to keep the pwm_disable() there as it potentially allows to save power.

However, it seems to not work quite well with PWMs from Freescale (see the thread from Russell). Or I would say with PWM with an undefined state when disabled and I believe we are soon to find more.

We could either:
- add a flag like can_sleep that would allow driver to know that the pwm always has to be enable to get sane results.

- or introduce functions like prepare/unprepare to be called from probe and remove in the leds-pwm driver and that will enable/disable the
channel in pwm-mxs. pwm_enable/pwm_disable not doing anything.

Quoted 9 lines

While I had exactly the same issue back in september and wrote the exact same patch that got rejected. I would agree with Thierry here taht we need to keep the pwm_disable() there as it potentially allows to save power.

However, it seems to not work quite well with PWMs from Freescale (see the thread from Russell). Or I would say with PWM with an undefined state when disabled and I believe we are soon to find more.

Quoted 7 lines

We could either:
- add a flag like can_sleep that would allow driver to know that the pwm always has to be enable to get sane results.

- or introduce functions like prepare/unprepare to be called from probe and remove in the leds-pwm driver and that will enable/disable the
channel in pwm-mxs. pwm_enable/pwm_disable not doing anything.

There are a few more options. I repeat your's to get a single list:

a) Somehow tell the API users that pwm_disable might not result in a flat zero after set_duty(0).
a') Anchor in the API that users must not expect anything from the pwm
pin after pwm_disable.
b) Make pwm_enable/pwm_disable noops on i.MX28
c) Make set_duty only return when the new value reached the pin.
d) Make pwm_disable wait to yield the expected result
d') Make pwm_disable wait if the last programmed duty cycle is 0 or
full period.
e) Introduce a new callback that does the waiting, e.g. a .commit callback
e') Introduce a new callback pwm_config_sync that does wait
appropriatly, keep pwm_config as is.
f) Make the i.MX28 driver switch the pin the gpio mode with the expected output level in pwm_disable.

a') is a special case of a); d') is a special case of d), ditto for e') and e).

a) and e) would result in changes in the pwm API. For c) I'd like to make the API docs more explicit, too. (i.e. demand that the new configuration is already active after pwm_config returns)

Both e) and a) have the drawback that the API becomes more complicated and users will get it wrong. a) and b) are bad from a power management POV. Maybe c) is what most users expect, but d) and d') are more effective as they result in less waiting and still are good enough. f) is more complicated to implement, also it depends on a gpio being available in hardware. So if we choose f) we still might need a fallback implementation from the other options.

My series implemented a'), but Thierry didn't like it. I think my favorite is e').

Quoted 49 lines · Uwe Kleine-K?nig

Quoted 9 lines

While I had exactly the same issue back in september and wrote the exact same patch that got rejected. I would agree with Thierry here taht we need to keep the pwm_disable() there as it potentially allows to save power.

However, it seems to not work quite well with PWMs from Freescale (see the thread from Russell). Or I would say with PWM with an undefined state when disabled and I believe we are soon to find more.

Quoted 7 lines

We could either:
- add a flag like can_sleep that would allow driver to know that the pwm always has to be enable to get sane results.

- or introduce functions like prepare/unprepare to be called from probe and remove in the leds-pwm driver and that will enable/disable the
channel in pwm-mxs. pwm_enable/pwm_disable not doing anything.

There are a few more options. I repeat your's to get a single list:

a) Somehow tell the API users that pwm_disable might not result in a flat zero after set_duty(0).
a') Anchor in the API that users must not expect anything from the pwm
pin after pwm_disable.
b) Make pwm_enable/pwm_disable noops on i.MX28
c) Make set_duty only return when the new value reached the pin.
d) Make pwm_disable wait to yield the expected result
d') Make pwm_disable wait if the last programmed duty cycle is 0 or
full period.
e) Introduce a new callback that does the waiting, e.g. a .commit callback
e') Introduce a new callback pwm_config_sync that does wait
appropriatly, keep pwm_config as is.
f) Make the i.MX28 driver switch the pin the gpio mode with the expected output level in pwm_disable.

a') is a special case of a); d') is a special case of d), ditto for e') and e).

a) and e) would result in changes in the pwm API. For c) I'd like to make the API docs more explicit, too. (i.e. demand that the new configuration is already active after pwm_config returns)

Both e) and a) have the drawback that the API becomes more complicated and users will get it wrong. a) and b) are bad from a power management POV. Maybe c) is what most users expect, but d) and d') are more effective as they result in less waiting and still are good enough. f) is more complicated to implement, also it depends on a gpio being available in hardware. So if we choose f) we still might need a fallback implementation from the other options.

Quoted 3 lines

My series implemented a'), but Thierry didn't like it. I think my favorite is e').

I think I would actually prefer d'. Doing the synchronisation in d' is abstracting it for the user.

Quoted 39 lines · Alexandre Belloni

Quoted 37 lines · Uwe Kleine-K?nig

Quoted 9 lines

While I had exactly the same issue back in september and wrote the exact same patch that got rejected. I would agree with Thierry here taht we need to keep the pwm_disable() there as it potentially allows to save power.

However, it seems to not work quite well with PWMs from Freescale (see the thread from Russell). Or I would say with PWM with an undefined state when disabled and I believe we are soon to find more.

Quoted 7 lines

We could either:
- add a flag like can_sleep that would allow driver to know that the pwm always has to be enable to get sane results.

- or introduce functions like prepare/unprepare to be called from probe and remove in the leds-pwm driver and that will enable/disable the
channel in pwm-mxs. pwm_enable/pwm_disable not doing anything.

There are a few more options. I repeat your's to get a single list:

a) Somehow tell the API users that pwm_disable might not result in a flat zero after set_duty(0).
a') Anchor in the API that users must not expect anything from the pwm
pin after pwm_disable.
b) Make pwm_enable/pwm_disable noops on i.MX28
c) Make set_duty only return when the new value reached the pin.
d) Make pwm_disable wait to yield the expected result
d') Make pwm_disable wait if the last programmed duty cycle is 0 or
full period.
e) Introduce a new callback that does the waiting, e.g. a .commit callback
e') Introduce a new callback pwm_config_sync that does wait
appropriatly, keep pwm_config as is.
f) Make the i.MX28 driver switch the pin the gpio mode with the expected output level in pwm_disable.
[...]
My series implemented a'), but Thierry didn't like it. I think my favorite is e').

Quoted 2 lines

I think I would actually prefer d'. Doing the synchronisation in d' is abstracting it for the user.

My preference would be simply d). Mostly because I don't like special cases and especially because I don't see an advantage in special casing 0 and full period duty cycles. That doesn't mean of course that drivers can't special case if they really want or have to.

That said I've also been thinking about adding support for e), which would allow atomically changing the duty cycle, period and polarity of a channel. This might become necessary at some point.

Hello,

Quoted 5 lines · Thierry Reding

My preference would be simply d). Mostly because I don't like special cases and especially because I don't see an advantage in special casing 0 and full period duty cycles. That doesn't mean of course that drivers can't special case if they really want or have to.

Well d' is just an optimisation, because if you have a duty cycle between 0 and full (exclusive) and no other means of syncronisation you cannot influence where pwm_disable stops the output, at low or high.

Do we have cases where pwm_disable makes the pin high-z? Or something else which violates the assumption "pwm_config(pwm, 0, period); pwm_disable(pwm); makes the pin 0"?

Quoted 3 lines

That said I've also been thinking about adding support for e), which would allow atomically changing the duty cycle, period and polarity of a channel. This might become necessary at some point.

On i.MX28 you cannot change atomically, only program atomically. The current period will end regularily anyhow as programmed before. But maybe that's what you mean?!

Best regards
Uwe

The .config() calls clk_get_rate() which might sleep, so we need to set pwm_chip can_sleep flag. Otherwise, we see the following warning when using pwm driven heartbeat led.

WARNING: CPU: 0 PID: 0 at kernel/locking/mutex.c:856 mutex_trylock+0x184/0x1a4()
DEBUG_LOCKS_WARN_ON(in_interrupt())
Modules linked in:
CPU: 0 PID: 0 Comm: swapper Not tainted 3.14.0-rc5 #18
[<c0015420>] (unwind_backtrace) from [<c0012cb0>] (show_stack+0x10/0x14)
[<c0012cb0>] (show_stack) from [<c001daf8>] (warn_slowpath_common+0x6c/0x8c)
[<c001daf8>] (warn_slowpath_common) from [<c001dbac>] (warn_slowpath_fmt+0x30/0x40)
[<c001dbac>] (warn_slowpath_fmt) from [<c045df74>] (mutex_trylock+0x184/0x1a4)
[<c045df74>] (mutex_trylock) from [<c0360950>] (clk_prepare_lock+0xc/0xec)
[<c0360950>] (clk_prepare_lock) from [<c0362020>] (clk_get_rate+0xc/0x68)
[<c0362020>] (clk_get_rate) from [<c028d07c>] (mxs_pwm_config+0x20/0x198)
[<c028d07c>] (mxs_pwm_config) from [<c028bde8>] (pwm_config+0x60/0x70)
[<c028bde8>] (pwm_config) from [<c034b61c>] (__led_pwm_set+0x1c/0x3c)
[<c034b61c>] (__led_pwm_set) from [<c034bc3c>] (led_heartbeat_function+0x70/0x110)
[<c034bc3c>] (led_heartbeat_function) from [<c00292f0>] (call_timer_fn+0x7c/0x164)
[<c00292f0>] (call_timer_fn) from [<c00295c8>] (run_timer_softirq+0x1f0/0x260)
[<c00295c8>] (run_timer_softirq) from [<c002255c>] (__do_softirq+0xc4/0x2f0)
[<c002255c>] (__do_softirq) from [<c0022890>] (irq_exit+0xa4/0x10c)
[<c0022890>] (irq_exit) from [<c0010240>] (handle_IRQ+0x34/0x84)
[<c0010240>] (handle_IRQ) from [<c0013524>] (__irq_svc+0x44/0x54)
[<c0013524>] (__irq_svc) from [<c00107f8>] (arch_cpu_idle+0x40/0x48)
[<c00107f8>] (arch_cpu_idle) from [<c005deb8>] (cpu_startup_entry+0x70/0x198)
[<c005deb8>] (cpu_startup_entry) from [<c060aac8>] (start_kernel+0x2a8/0x2f8)

Reported-by Stefan Wahren <[email protected]>

Signed-off-by Shawn Guo <[email protected]>

---
 drivers/pwm/pwm-mxs.c | 1 +
 1 file changed, 1 insertion(+)

drivers/pwm/pwm-mxs.c +1 -0

index 9475bc7..4f1bb4e 100644
@@ -147,6 +147,7 @@ static int mxs_pwm_probe(struct platform_device *pdev)
 	mxs->chip.dev = &pdev->dev;
 	mxs->chip.ops = &mxs_pwm_ops;
 	mxs->chip.base = -1;
+	mxs->chip.can_sleep = true;
 	ret = of_property_read_u32(np, "fsl,pwm-number", &mxs->chip.npwm);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed to get pwm number: %d\n", ret);

Signature · 1 line

1.8.3.2

The .config() calls clk_get_rate() which might sleep, so we need to set pwm_chip can_sleep flag. Otherwise, we see the following warning when using pwm driven heartbeat led.

WARNING: CPU: 0 PID: 0 at kernel/locking/mutex.c:856 mutex_trylock+0x184/0x1a4()
DEBUG_LOCKS_WARN_ON(in_interrupt())
Modules linked in:
CPU: 0 PID: 0 Comm: swapper Not tainted 3.14.0-rc5 #18
[<c0015420>] (unwind_backtrace) from [<c0012cb0>] (show_stack+0x10/0x14)
[<c0012cb0>] (show_stack) from [<c001daf8>] (warn_slowpath_common+0x6c/0x8c)
[<c001daf8>] (warn_slowpath_common) from [<c001dbac>] (warn_slowpath_fmt+0x30/0x40)
[<c001dbac>] (warn_slowpath_fmt) from [<c045df74>] (mutex_trylock+0x184/0x1a4)
[<c045df74>] (mutex_trylock) from [<c0360950>] (clk_prepare_lock+0xc/0xec)
[<c0360950>] (clk_prepare_lock) from [<c0362020>] (clk_get_rate+0xc/0x68)
[<c0362020>] (clk_get_rate) from [<c028d07c>] (mxs_pwm_config+0x20/0x198)
[<c028d07c>] (mxs_pwm_config) from [<c028bde8>] (pwm_config+0x60/0x70)
[<c028bde8>] (pwm_config) from [<c034b61c>] (__led_pwm_set+0x1c/0x3c)
[<c034b61c>] (__led_pwm_set) from [<c034bc3c>] (led_heartbeat_function+0x70/0x110)
[<c034bc3c>] (led_heartbeat_function) from [<c00292f0>] (call_timer_fn+0x7c/0x164)
[<c00292f0>] (call_timer_fn) from [<c00295c8>] (run_timer_softirq+0x1f0/0x260)
[<c00295c8>] (run_timer_softirq) from [<c002255c>] (__do_softirq+0xc4/0x2f0)
[<c002255c>] (__do_softirq) from [<c0022890>] (irq_exit+0xa4/0x10c)
[<c0022890>] (irq_exit) from [<c0010240>] (handle_IRQ+0x34/0x84)
[<c0010240>] (handle_IRQ) from [<c0013524>] (__irq_svc+0x44/0x54)
[<c0013524>] (__irq_svc) from [<c00107f8>] (arch_cpu_idle+0x40/0x48)
[<c00107f8>] (arch_cpu_idle) from [<c005deb8>] (cpu_startup_entry+0x70/0x198)
[<c005deb8>] (cpu_startup_entry) from [<c060aac8>] (start_kernel+0x2a8/0x2f8)

Reported-by Stefan Wahren <[email protected]>

Signed-off-by Shawn Guo <[email protected]>

---
 drivers/pwm/pwm-mxs.c | 1 +
 1 file changed, 1 insertion(+)

drivers/pwm/pwm-mxs.c +1 -0

index 9475bc7..4f1bb4e 100644
@@ -147,6 +147,7 @@ static int mxs_pwm_probe(struct platform_device *pdev)
 	mxs->chip.dev = &pdev->dev;
 	mxs->chip.ops = &mxs_pwm_ops;
 	mxs->chip.base = -1;
+	mxs->chip.can_sleep = true;
 	ret = of_property_read_u32(np, "fsl,pwm-number", &mxs->chip.npwm);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed to get pwm number: %d\n", ret);