Skip to contentSkip to conversation

PROBLEM: BUG appearing when trying to allocate interrupt on Exynos MCT after CPU hotplug

DISCUSSION Marcin Jabrzyk 4 participants 9 messages Oct 23, 2014 – Feb 2, 2015

Subsystems (2)
Arm Samsung Exynos Clocksource, clockevent drivers
lore
Share & tools
lore

Marcin Jabrzyk

author
#0 lore raw

[1.] One line summary of the problem: "BUG: sleeping function called from invalid context at mm/slub.c:1250" after CPU hotplug [2.] Full description of the problem/report:

This was tested on Exynos 3250 board with
https://lkml.org/lkml/2014/9/24/441 applied. Board is booting to /bin/sh. After executing:

mount -t sysfs sys /sys && echo 0 > /sys/devices/system/cpu/cpu1/online && echo 1 > /sys/devices/system/cpu/cpu1/online

I'm getting:

[    7.226405] IRQ258 no longer affine to CPU1
[    7.226629] CPU1: shutdown
[    7.230037] CPU1: Software reset
[    7.231822] CPU1: Booted secondary processor
[    7.231843] BUG: sleeping function called from invalid context at 
mm/slub.c:1250
[    7.231850] in_atomic(): 1, irqs_disabled(): 128, pid: 0, name: swapper/1
[    7.231861] Preemption disabled at:[<  (null)>]   (null)
[    7.231864]
[    7.231876] CPU: 1 PID: 0 Comm: swapper/1 Not tainted 3.17.0-dirty #45
[    7.231914] [<c0013c04>] (unwind_backtrace) from [<c0010eac>] 
(show_stack+0x10/0x14)
[    7.231931] [<c0010eac>] (show_stack) from [<c03ffd0c>] 
(dump_stack+0x70/0xbc)
[    7.231950] [<c03ffd0c>] (dump_stack) from [<c00b9a20>] 
(kmem_cache_alloc+0xe8/0x184)
[    7.231968] [<c00b9a20>] (kmem_cache_alloc) from [<c0059710>] 
(request_threaded_irq+0x64/0x128)
[    7.231985] [<c0059710>] (request_threaded_irq) from [<c030ecc8>] 
(exynos4_local_timer_setup+0xc0/0x13c)
[    7.232000] [<c030ecc8>] (exynos4_local_timer_setup) from 
[<c030ede4>] (exynos4_mct_cpu_notify+0x30/0xa8)
[    7.232016] [<c030ede4>] (exynos4_mct_cpu_notify) from [<c0038540>] 
(notifier_call_chain+0x44/0x84)
[    7.232034] [<c0038540>] (notifier_call_chain) from [<c0021144>] 
(__cpu_notify+0x28/0x44)
[    7.232049] [<c0021144>] (__cpu_notify) from [<c0012af0>] 
(secondary_start_kernel+0xe8/0x138)
[    7.232062] [<c0012af0>] (secondary_start_kernel) from [<400086a4>]

(0x400086a4)

The problem is that request_irq is calling allocation with GFP_KERNEL flag in atomic block.
This bug should be easy observable on any board with "samsung,exynos4210-mct" compatible MCT block.

[4.1.] Kernel version (from /proc/version):
3.17.0
[4.2.] Kernel .config file:
exynos_defconfig + DEBUG_ATOMIC_SLEEP and DEBUG_PREEMPT

[7.] A small shell script or example program which triggers the
problem (if possible)
mount -t sysfs sys /sys && echo 0 > /sys/devices/system/cpu/cpu1/online && echo 1 > /sys/devices/system/cpu/cpu1/online
[8.] Environment
/bin/sh

When SoC have MCT_INT_SPI interrupt it is being allocated after hotplugging of the CPU, secondary_start_kernel() is sending CPU boot notifications which are send when preemption and interrupts are disabled. Exynos_mct notification handler tries to set up and allocate IRQ for SPI type interrupt for started CPU and then BUG appears. There might be similar problem on qcom-timer I think just after looking on the code.

Best regards,
--
Marcin Jabrzyk
Samsung R&D Institute Poland
Samsung Electronics

Acked-by Kukjin Kim Acked-by Marc Zyngier
#1 lore raw
Quoted 2 lines · Marcin Jabrzyk

[1.] One line summary of the problem: "BUG: sleeping function called from invalid context at mm/slub.c:1250" after CPU hotplug

I'm really not surprised.

Quoted 7 lines

When SoC have MCT_INT_SPI interrupt it is being allocated after hotplugging of the CPU, secondary_start_kernel() is sending CPU boot notifications which are send when preemption and interrupts are disabled. Exynos_mct notification handler tries to set up and allocate IRQ for SPI type interrupt for started CPU and then BUG appears.
There might be similar problem on qcom-timer I think just after looking on the code.

The CPU notifier is called via notify_cpu_starting(), which is called with interrupts disabled, and a reason code of CPU_STARTING. Interrupts at this point /must/ remain disabled.

The Exynos code then goes on to call exynos4_local_timer_setup() which tries to reverse the free_irq() in exynos4_local_timer_stop() by calling request_irq(). Calling request_irq() with interrupts off has never been permissible.

So, this code is wrong today, and it was also wrong when it was written. It /couldn't/ have been tested. It looks like this commit added this buggy code:

commit ee98d27df6827b5ba4bd99cb7d5cb1239b6a1a31

Author: Stephen Boyd <[email protected]>
Date:   Fri Feb 15 16:40:51 2013 -0800

ARM: EXYNOS4: Divorce mct from local timer API

Separate the mct local timers from the local timer API. This will allow us to remove ARM local timer support in the near future and gets us closer to moving this driver to drivers/clocksource.

Acked-by Kukjin Kim <[email protected]>
Acked-by Marc Zyngier <[email protected]>
Cc Thomas Abraham <[email protected]>
Signed-off-by Stephen Boyd <[email protected]>

A good question would be: why doesn't this happen at boot time when CPU1 is first brought up? The conditions here are no different from hotplugging CPU1 back in. Do you see a similar warning on boot too?

Signature · 2 lines
FTTC broadband for 0.8mile line: currently at 9.5Mbps down 400kbps up
according to speedtest.net.
Acked-by Mark Rutland Reviewed-by Tomasz Figa
#2 lore raw
Quoted 12 lines · Russell King - ARM Linux
Quoted 2 lines · Marcin Jabrzyk

[1.] One line summary of the problem: "BUG: sleeping function called from invalid context at mm/slub.c:1250" after CPU hotplug

I'm really not surprised.

Quoted 7 lines

When SoC have MCT_INT_SPI interrupt it is being allocated after hotplugging of the CPU, secondary_start_kernel() is sending CPU boot notifications which are send when preemption and interrupts are disabled. Exynos_mct notification handler tries to set up and allocate IRQ for SPI type interrupt for started CPU and then BUG appears.
There might be similar problem on qcom-timer I think just after looking on the code.

There's no problem for qcom-timer because there are only PPIs on SMP platforms.

Quoted 19 lines

The CPU notifier is called via notify_cpu_starting(), which is called with interrupts disabled, and a reason code of CPU_STARTING. Interrupts at this point /must/ remain disabled.

The Exynos code then goes on to call exynos4_local_timer_setup() which tries to reverse the free_irq() in exynos4_local_timer_stop() by calling request_irq(). Calling request_irq() with interrupts off has never been permissible.

So, this code is wrong today, and it was also wrong when it was written. It /couldn't/ have been tested. It looks like this commit added this buggy code:

commit ee98d27df6827b5ba4bd99cb7d5cb1239b6a1a31

Author: Stephen Boyd <[email protected]>
Date:   Fri Feb 15 16:40:51 2013 -0800

ARM: EXYNOS4: Divorce mct from local timer API

Quoted 8 lines

Separate the mct local timers from the local timer API. This will allow us to remove ARM local timer support in the near future and gets us closer to moving this driver to drivers/clocksource.

Acked-by Kukjin Kim <[email protected]>
Acked-by Marc Zyngier <[email protected]>
Cc Thomas Abraham <[email protected]>
Signed-off-by Stephen Boyd <[email protected]>

I'm not so sure. It looks like in that patch I didn't change anything with respect to when things are called. In fact, it looks like we were calling setup_irq() there, but another patch around the same time changed that to request_irq()

commit 7114cd749a12ff9fd64a2f6f04919760f45ab183

Author: Chander Kashyap <[email protected]>
Date:   Wed Jun 19 00:29:35 2013 +0900

clocksource: exynos_mct: use (request/free)_irq calls for local timer registration

Replace the (setup/remove)_irq calls for local timer registration with (request/free)_irq calls. This generalizes the local timer registration API. Suggested by Mark Rutland.

Signed-off-by Chander Kashyap <[email protected]>
Acked-by Mark Rutland <[email protected]>
Reviewed-by Tomasz Figa <[email protected]>
Signed-off-by Kukjin Kim <[email protected]>

I don't believe setup_irq() allocates anything so we should probably go back to using that over request_irq() or explore requesting the irqs once and then enabling/disabling instead.

Quoted 4 lines

A good question would be: why doesn't this happen at boot time when CPU1 is first brought up? The conditions here are no different from hotplugging CPU1 back in. Do you see a similar warning on boot too?

Probably because such checks are completely avoided until the system state is switched to SYSTEM_RUNNING (see the first if statement in __might_sleep()). It would be nice if we could remove that.

Signature · 2 lines
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
author
#3 lore raw
Quoted 14 lines · Stephen Boyd
Quoted 12 lines · Russell King - ARM Linux
Quoted 2 lines · Marcin Jabrzyk

[1.] One line summary of the problem: "BUG: sleeping function called from invalid context at mm/slub.c:1250" after CPU hotplug

I'm really not surprised.

Quoted 7 lines

When SoC have MCT_INT_SPI interrupt it is being allocated after hotplugging of the CPU, secondary_start_kernel() is sending CPU boot notifications which are send when preemption and interrupts are disabled. Exynos_mct notification handler tries to set up and allocate IRQ for SPI type interrupt for started CPU and then BUG appears.
There might be similar problem on qcom-timer I think just after looking on the code.

Quoted 3 lines

There's no problem for qcom-timer because there are only PPIs on SMP platforms.

Ok, so it's only a problem on Exynos platform for now.

Quoted 43 lines
Quoted 19 lines

The CPU notifier is called via notify_cpu_starting(), which is called with interrupts disabled, and a reason code of CPU_STARTING. Interrupts at this point /must/ remain disabled.

The Exynos code then goes on to call exynos4_local_timer_setup() which tries to reverse the free_irq() in exynos4_local_timer_stop() by calling request_irq(). Calling request_irq() with interrupts off has never been permissible.

So, this code is wrong today, and it was also wrong when it was written. It /couldn't/ have been tested. It looks like this commit added this buggy code:

commit ee98d27df6827b5ba4bd99cb7d5cb1239b6a1a31

Author: Stephen Boyd <[email protected]>
Date:   Fri Feb 15 16:40:51 2013 -0800

ARM: EXYNOS4: Divorce mct from local timer API

Quoted 8 lines

Separate the mct local timers from the local timer API. This will allow us to remove ARM local timer support in the near future and gets us closer to moving this driver to drivers/clocksource.

Acked-by Kukjin Kim <[email protected]>
Acked-by Marc Zyngier <[email protected]>
Cc Thomas Abraham <[email protected]>
Signed-off-by Stephen Boyd <[email protected]>

I'm not so sure. It looks like in that patch I didn't change anything with respect to when things are called. In fact, it looks like we were calling setup_irq() there, but another patch around the same time changed that to request_irq()

commit 7114cd749a12ff9fd64a2f6f04919760f45ab183

Author: Chander Kashyap <[email protected]>
Date:   Wed Jun 19 00:29:35 2013 +0900

clocksource: exynos_mct: use (request/free)_irq calls for local timer registration

Replace the (setup/remove)_irq calls for local timer registration with (request/free)_irq calls. This generalizes the local timer registration API. Suggested by Mark Rutland.

Quoted 9 lines
Signed-off-by Chander Kashyap <[email protected]>
Acked-by Mark Rutland <[email protected]>
Reviewed-by Tomasz Figa <[email protected]>
Signed-off-by Kukjin Kim <[email protected]>

I don't believe setup_irq() allocates anything so we should probably go back to using that over request_irq() or explore requesting the irqs once and then enabling/disabling instead.

So what would be a better way to handle this? Going back to setup_irq or trying to enable/disable irqs on CPU hotplug? As this touched low level things and it's rare case for setting/enabling irqs just after CPU is coming back to life again.

Quoted 4 lines
Quoted 4 lines

A good question would be: why doesn't this happen at boot time when CPU1 is first brought up? The conditions here are no different from hotplugging CPU1 back in. Do you see a similar warning on boot too?

No the boot looks clean and there is not any sign of that problem.

Quoted 5 lines

Probably because such checks are completely avoided until the system state is switched to SYSTEM_RUNNING (see the first if statement in __might_sleep()). It would be nice if we could remove that.

That's most probably the reason of no warnings on boot process.

Best regards,
--
Marcin Jabrzyk
Samsung R&D Institute Poland
Samsung Electronics

#4 lore raw
Quoted 64 lines · Marcin Jabrzyk
Quoted 51 lines · Stephen Boyd
Quoted 23 lines · Russell King - ARM Linux

The CPU notifier is called via notify_cpu_starting(), which is called with interrupts disabled, and a reason code of CPU_STARTING. Interrupts
at this point /must/ remain disabled.

The Exynos code then goes on to call exynos4_local_timer_setup() which tries to reverse the free_irq() in exynos4_local_timer_stop() by calling
request_irq(). Calling request_irq() with interrupts off has never been
permissible.

So, this code is wrong today, and it was also wrong when it was written.
It /couldn't/ have been tested. It looks like this commit added this buggy code:

commit ee98d27df6827b5ba4bd99cb7d5cb1239b6a1a31

Author: Stephen Boyd <[email protected]>
Date:   Fri Feb 15 16:40:51 2013 -0800

ARM: EXYNOS4: Divorce mct from local timer API

Quoted 8 lines

Separate the mct local timers from the local timer API. This will allow us to remove ARM local timer support in the near future and gets us closer to moving this driver to drivers/clocksource.

Acked-by Kukjin Kim <[email protected]>
Acked-by Marc Zyngier <[email protected]>
Cc Thomas Abraham <[email protected]>
Signed-off-by Stephen Boyd <[email protected]>

I'm not so sure. It looks like in that patch I didn't change anything with respect to when things are called. In fact, it looks like we were calling setup_irq() there, but another patch around the same time changed that to request_irq()

commit 7114cd749a12ff9fd64a2f6f04919760f45ab183

Author: Chander Kashyap <[email protected]>
Date:   Wed Jun 19 00:29:35 2013 +0900

clocksource: exynos_mct: use (request/free)_irq calls for local
timer registration

Replace the (setup/remove)_irq calls for local timer
registration with
(request/free)_irq calls. This generalizes the local timer
registration API.
Suggested by Mark Rutland.

Quoted 9 lines
Signed-off-by Chander Kashyap <[email protected]>
Acked-by Mark Rutland <[email protected]>
Reviewed-by Tomasz Figa <[email protected]>
Signed-off-by Kukjin Kim <[email protected]>

I don't believe setup_irq() allocates anything so we should probably go back to using that over request_irq() or explore requesting the irqs once and then enabling/disabling instead.

Quoted 5 lines

So what would be a better way to handle this? Going back to setup_irq or trying to enable/disable irqs on CPU hotplug? As this touched low level things and it's rare case for setting/enabling irqs just after CPU is coming back to life again.

The safest thing is setup_irq(), but do you care to try this patch? Doing the enable/disable is not as robust because request_irq() returns with the irq enabled and then we have to disable the irq to make things symmetric. This whole driver doesn't look like it's prepared for such a situation where the interrupt triggers before the clockevent is registered so this doesn't look like a problem in practice. Doing the disable right after request is typically bad though, and may not pass review.

----8<-----

From: Stephen Boyd <[email protected]>
Subject: [PATCH] clocksource: exynos_mct: Avoid scheduling while atomic

If we call request_irq() during the CPU_STARTING notifier we'll try to allocate an irq descriptor with GFP_KERNEL while we're running with irqs disabled. Just request the irqs at boot time and enable/disable them when a CPU comes up or goes down to avoid such problems.

Signed-off-by Stephen Boyd <[email protected]>
---
 drivers/clocksource/exynos_mct.c | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)
drivers/clocksource/exynos_mct.c +13 -10
index 9403061a2acc..1800053b4644 100644
@@ -467,13 +467,7 @@ static int exynos4_local_timer_setup(struct clock_event_device *evt)
 
 	if (mct_int_type == MCT_INT_SPI) {
 		evt->irq = mct_irqs[MCT_L0_IRQ + cpu];
-		if (request_irq(evt->irq, exynos4_mct_tick_isr,
-				IRQF_TIMER | IRQF_NOBALANCING,
-				evt->name, mevt)) {
-			pr_err("exynos-mct: cannot register IRQ %d\n",
-				evt->irq);
-			return -EIO;
-		}
+		enable_irq(evt->irq);
 		irq_force_affinity(mct_irqs[MCT_L0_IRQ + cpu], cpumask_of(cpu));
 	} else {
 		enable_percpu_irq(mct_irqs[MCT_L0_IRQ], 0);

@@ -488,7 +482,7 @@ static void exynos4_local_timer_stop(struct clock_event_device *evt)
 {
 	evt->set_mode(CLOCK_EVT_MODE_UNUSED, evt);
 	if (mct_int_type == MCT_INT_SPI)
-		free_irq(evt->irq, this_cpu_ptr(&percpu_mct_tick));
+		disable_irq(evt->irq);
 	else
 		disable_percpu_irq(mct_irqs[MCT_L0_IRQ]);
 }

@@ -522,8 +516,9 @@ static struct notifier_block exynos4_mct_cpu_nb = {
 
 static void __init exynos4_timer_resources(struct device_node *np, void __iomem *base)
 {
-	int err;
+	int err, cpu;
 	struct mct_clock_event_device *mevt = this_cpu_ptr(&percpu_mct_tick);
+	struct mct_clock_event_device *evt;
 	struct clk *mct_clk, *tick_clk;
 
 	tick_clk = np ? of_clk_get_by_name(np, "fin_pll") :

@@ -549,7 +544,15 @@ static void __init exynos4_timer_resources(struct device_node *np, void __iomem
 		WARN(err, "MCT: can't request IRQ %d (%d)\n",
 		     mct_irqs[MCT_L0_IRQ], err);
 	} else {
-		irq_set_affinity(mct_irqs[MCT_L0_IRQ], cpumask_of(0));
+		for_each_present_cpu(cpu) {
+			evt = per_cpu_ptr(&percpu_mct_tick, cpu);
+			if (request_irq(mct_irqs[MCT_L0_IRQ + cpu],
+					exynos4_mct_tick_isr,
+					IRQF_TIMER | IRQF_NOBALANCING,
+					"MCT", evt))
+				pr_err("exynos-mct: cannot register IRQ\n");
+			disable_irq(mct_irqs[MCT_L0_IRQ + cpu]);
+		}
 	}
 
 	err = register_cpu_notifier(&exynos4_mct_cpu_nb);
Signature · 2 lines
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
author
#5 lore raw

So I've tried this patch, it resolves one problem but introduces also new ones. As expected the BUG warning is not showing after applying this patch but there are some interesting side effects.
I was looking on /proc/interrupts output. IRQ for CPU0 have "MCT" name and IRQ for CPU1 has unexpectedly no name at all.
After making hotplug cycle of CPU1 I've observed that IRQs attached originally for that CPU are generating on really low count and not in order with IRQ for CPU0.
What's more the interrupt for CPU1 is showing to me as being counted for both CPUs, so it's probably not being attached to CPU1.

Best regards,
Marcin Jabrzyk

Quoted 143 lines · Stephen Boyd
Quoted 64 lines · Marcin Jabrzyk
Quoted 51 lines · Stephen Boyd
Quoted 23 lines · Russell King - ARM Linux

The CPU notifier is called via notify_cpu_starting(), which is called with interrupts disabled, and a reason code of CPU_STARTING. Interrupts
at this point /must/ remain disabled.

The Exynos code then goes on to call exynos4_local_timer_setup() which tries to reverse the free_irq() in exynos4_local_timer_stop() by calling
request_irq(). Calling request_irq() with interrupts off has never been
permissible.

So, this code is wrong today, and it was also wrong when it was written.
It /couldn't/ have been tested. It looks like this commit added this buggy code:

commit ee98d27df6827b5ba4bd99cb7d5cb1239b6a1a31

Author: Stephen Boyd <[email protected]>
Date:   Fri Feb 15 16:40:51 2013 -0800

ARM: EXYNOS4: Divorce mct from local timer API

Quoted 8 lines

Separate the mct local timers from the local timer API. This will allow us to remove ARM local timer support in the near future and gets us closer to moving this driver to drivers/clocksource.

Acked-by Kukjin Kim <[email protected]>
Acked-by Marc Zyngier <[email protected]>
Cc Thomas Abraham <[email protected]>
Signed-off-by Stephen Boyd <[email protected]>

I'm not so sure. It looks like in that patch I didn't change anything with respect to when things are called. In fact, it looks like we were calling setup_irq() there, but another patch around the same time changed that to request_irq()

commit 7114cd749a12ff9fd64a2f6f04919760f45ab183

Author: Chander Kashyap <[email protected]>
Date:   Wed Jun 19 00:29:35 2013 +0900

clocksource: exynos_mct: use (request/free)_irq calls for local
timer registration

Replace the (setup/remove)_irq calls for local timer
registration with
(request/free)_irq calls. This generalizes the local timer
registration API.
Suggested by Mark Rutland.

Quoted 9 lines
Signed-off-by Chander Kashyap <[email protected]>
Acked-by Mark Rutland <[email protected]>
Reviewed-by Tomasz Figa <[email protected]>
Signed-off-by Kukjin Kim <[email protected]>

I don't believe setup_irq() allocates anything so we should probably go back to using that over request_irq() or explore requesting the irqs once and then enabling/disabling instead.

Quoted 5 lines

So what would be a better way to handle this? Going back to setup_irq or trying to enable/disable irqs on CPU hotplug? As this touched low level things and it's rare case for setting/enabling irqs just after CPU is coming back to life again.

The safest thing is setup_irq(), but do you care to try this patch? Doing the enable/disable is not as robust because request_irq() returns with the irq enabled and then we have to disable the irq to make things symmetric. This whole driver doesn't look like it's prepared for such a situation where the interrupt triggers before the clockevent is registered so this doesn't look like a problem in practice. Doing the disable right after request is typically bad though, and may not pass review.

----8<-----

From: Stephen Boyd <[email protected]>
Subject: [PATCH] clocksource: exynos_mct: Avoid scheduling while atomic

If we call request_irq() during the CPU_STARTING notifier we'll try to allocate an irq descriptor with GFP_KERNEL while we're running with irqs disabled. Just request the irqs at boot time and enable/disable them when a CPU comes up or goes down to avoid such problems.

Signed-off-by Stephen Boyd <[email protected]>
---
  drivers/clocksource/exynos_mct.c | 23 +++++++++++++----------
  1 file changed, 13 insertions(+), 10 deletions(-)
drivers/clocksource/exynos_mct.c +6 -10
index 9403061a2acc..1800053b4644 100644
@@ -467,13 +467,7 @@ static int exynos4_local_timer_setup(struct clock_event_device *evt)
 
  	if (mct_int_type == MCT_INT_SPI) {
  		evt->irq = mct_irqs[MCT_L0_IRQ + cpu];
-		if (request_irq(evt->irq, exynos4_mct_tick_isr,
-				IRQF_TIMER | IRQF_NOBALANCING,
-				evt->name, mevt)) {
-			pr_err("exynos-mct: cannot register IRQ %d\n",
-				evt->irq);
-			return -EIO;
-		}
+		enable_irq(evt->irq);
  		irq_force_affinity(mct_irqs[MCT_L0_IRQ + cpu], cpumask_of(cpu));
  	} else {
  		enable_percpu_irq(mct_irqs[MCT_L0_IRQ], 0);

@@ -488,7 +482,7 @@ static void exynos4_local_timer_stop(struct clock_event_device *evt)
  {
  	evt->set_mode(CLOCK_EVT_MODE_UNUSED, evt);
  	if (mct_int_type == MCT_INT_SPI)
-		free_irq(evt->irq, this_cpu_ptr(&percpu_mct_tick));
+		disable_irq(evt->irq);
  	else
  		disable_percpu_irq(mct_irqs[MCT_L0_IRQ]);
  }

@@ -522,8 +516,9 @@ static struct notifier_block exynos4_mct_cpu_nb = {
 
  static void __init exynos4_timer_resources(struct device_node *np, void __iomem *base)
  {
-	int err;
+	int err, cpu;
  	struct mct_clock_event_device *mevt = this_cpu_ptr(&percpu_mct_tick);
+	struct mct_clock_event_device *evt;
  	struct clk *mct_clk, *tick_clk;
 
  	tick_clk = np ? of_clk_get_by_name(np, "fin_pll") :

@@ -549,7 +544,15 @@ static void __init exynos4_timer_resources(struct device_node *np, void __iomem
  		WARN(err, "MCT: can't request IRQ %d (%d)\n",
  		     mct_irqs[MCT_L0_IRQ], err);
  	} else {
-		irq_set_affinity(mct_irqs[MCT_L0_IRQ], cpumask_of(0));
+		for_each_present_cpu(cpu) {
+			evt = per_cpu_ptr(&percpu_mct_tick, cpu);

drivers/clocksource/exynos_mct.c static void __init exynos4_timer_resources(struct device_node *np, void __iomem · line 558

Quoted 10 lines
drivers/clocksource/exynos_mct.c +7 -0
+			if (request_irq(mct_irqs[MCT_L0_IRQ + cpu],
+					exynos4_mct_tick_isr,
+					IRQF_TIMER | IRQF_NOBALANCING,
+					"MCT", evt))
+				pr_err("exynos-mct: cannot register IRQ\n");
+			disable_irq(mct_irqs[MCT_L0_IRQ + cpu]);
+		}
  	}
 
  	err = register_cpu_notifier(&exynos4_mct_cpu_nb);

--
Marcin Jabrzyk
Samsung R&D Institute Poland
Samsung Electronics

#6 lore raw

Kept meaning to get back to this thread. Have you resolved it?

Quoted 3 lines · Marcin Jabrzyk

So I've tried this patch, it resolves one problem but introduces also new ones. As expected the BUG warning is not showing after applying this patch but there are some interesting side effects.

Well that's half good news.

Quoted 2 lines

I was looking on /proc/interrupts output. IRQ for CPU0 have "MCT" name and IRQ for CPU1 has unexpectedly no name at all.

This is pretty confusing. I don't see how the patch could cause this to happen.

Quoted 6 lines

After making hotplug cycle of CPU1 I've observed that IRQs attached originally for that CPU are generating on really low count and not in order with IRQ for CPU0.
What's more the interrupt for CPU1 is showing to me as being counted for both CPUs, so it's probably not being attached to CPU1.

yeah. Can you give the output of /proc/timer_list in addition to /proc/interrupts? It may give some hints on what's going on. It may also be interesting to see if irq_force_affinity() is failing. Please check the return value and print an error

drivers/clocksource/exynos_mct.c +4 -1
index 1800053b4644..3c4538e26731 100644
@@ -450,6 +450,7 @@ static int exynos4_local_timer_setup(struct clock_event_device *evt)
 {
 	struct mct_clock_event_device *mevt;
 	unsigned int cpu = smp_processor_id();
+	int ret;
 
 	mevt = container_of(evt, struct mct_clock_event_device, evt);
 

@@ -468,7 +469,9 @@ static int exynos4_local_timer_setup(struct clock_event_device *evt)
 	if (mct_int_type == MCT_INT_SPI) {
 		evt->irq = mct_irqs[MCT_L0_IRQ + cpu];
 		enable_irq(evt->irq);
-		irq_force_affinity(mct_irqs[MCT_L0_IRQ + cpu], cpumask_of(cpu));
+		ret = irq_force_affinity(mct_irqs[MCT_L0_IRQ + cpu], cpumask_of(cpu));
+		if (ret)
+			pr_err("force failed %d\n", ret);
 	} else {
 		enable_percpu_irq(mct_irqs[MCT_L0_IRQ], 0);
 	}
Signature · 2 lines
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
#7 lore raw
Quoted 22 lines · Stephen Boyd

Kept meaning to get back to this thread. Have you resolved it?

Quoted 3 lines · Marcin Jabrzyk

So I've tried this patch, it resolves one problem but introduces also new ones. As expected the BUG warning is not showing after applying this patch but there are some interesting side effects.

Well that's half good news.

Quoted 2 lines

I was looking on /proc/interrupts output. IRQ for CPU0 have "MCT" name and IRQ for CPU1 has unexpectedly no name at all.

This is pretty confusing. I don't see how the patch could cause this to happen.

Quoted 6 lines

After making hotplug cycle of CPU1 I've observed that IRQs attached originally for that CPU are generating on really low count and not in order with IRQ for CPU0.
What's more the interrupt for CPU1 is showing to me as being counted for both CPUs, so it's probably not being attached to CPU1.

Quoted 4 lines

yeah. Can you give the output of /proc/timer_list in addition to /proc/interrupts? It may give some hints on what's going on. It may also be interesting to see if irq_force_affinity() is failing. Please check the return value and print an error

Hi Stephen, Marcin,

can you have a look if the patch [1] fixes this issue ?

-- Daniel

[1] https://lkml.org/lkml/2015/1/30/423

Quoted 24 lines
drivers/clocksource/exynos_mct.c +4 -1
index 1800053b4644..3c4538e26731 100644
@@ -450,6 +450,7 @@ static int exynos4_local_timer_setup(struct clock_event_device *evt)
  {
  	struct mct_clock_event_device *mevt;
  	unsigned int cpu = smp_processor_id();
+	int ret;
 
  	mevt = container_of(evt, struct mct_clock_event_device, evt);
 

@@ -468,7 +469,9 @@ static int exynos4_local_timer_setup(struct clock_event_device *evt)
  	if (mct_int_type == MCT_INT_SPI) {
  		evt->irq = mct_irqs[MCT_L0_IRQ + cpu];
  		enable_irq(evt->irq);
-		irq_force_affinity(mct_irqs[MCT_L0_IRQ + cpu], cpumask_of(cpu));
+		ret = irq_force_affinity(mct_irqs[MCT_L0_IRQ + cpu], cpumask_of(cpu));
+		if (ret)
+			pr_err("force failed %d\n", ret);
  	} else {
  		enable_percpu_irq(mct_irqs[MCT_L0_IRQ], 0);
  	}
Signature · 5 lines
  <http://www.linaro.org/> Linaro.org ? Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog
author
#8 lore raw
Quoted 28 lines · Daniel Lezcano
Quoted 22 lines · Stephen Boyd

Kept meaning to get back to this thread. Have you resolved it?

Quoted 3 lines · Marcin Jabrzyk

So I've tried this patch, it resolves one problem but introduces also new ones. As expected the BUG warning is not showing after applying this patch but there are some interesting side effects.

Well that's half good news.

Quoted 2 lines

I was looking on /proc/interrupts output. IRQ for CPU0 have "MCT" name and IRQ for CPU1 has unexpectedly no name at all.

This is pretty confusing. I don't see how the patch could cause this to happen.

Quoted 6 lines

After making hotplug cycle of CPU1 I've observed that IRQs attached originally for that CPU are generating on really low count and not in order with IRQ for CPU0.
What's more the interrupt for CPU1 is showing to me as being counted for both CPUs, so it's probably not being attached to CPU1.

Quoted 4 lines

yeah. Can you give the output of /proc/timer_list in addition to /proc/interrupts? It may give some hints on what's going on. It may also be interesting to see if irq_force_affinity() is failing. Please check the return value and print an error

Quoted 9 lines

Hi Stephen, Marcin,

can you have a look if the patch [1] fixes this issue ?

-- Daniel

[1] https://lkml.org/lkml/2015/1/30/423

Hi Daniel,

I've checked this patch on the board that have problems and it fixes this issue completely. Everything looks fine after power cycle of the CPU.

Best regards,
Marcin Jabrzyk

Quoted 30 lines
Quoted 28 lines
drivers/clocksource/exynos_mct.c +4 -1
index 1800053b4644..3c4538e26731 100644
@@ -450,6 +450,7 @@ static int exynos4_local_timer_setup(struct
 clock_event_device *evt)
  {
      struct mct_clock_event_device *mevt;
      unsigned int cpu = smp_processor_id();
+    int ret;
 
      mevt = container_of(evt, struct mct_clock_event_device, evt);

@@ -468,7 +469,9 @@ static int exynos4_local_timer_setup(struct
 clock_event_device *evt)
      if (mct_int_type == MCT_INT_SPI) {
          evt->irq = mct_irqs[MCT_L0_IRQ + cpu];
          enable_irq(evt->irq);
-        irq_force_affinity(mct_irqs[MCT_L0_IRQ + cpu], cpumask_of(cpu));
+        ret = irq_force_affinity(mct_irqs[MCT_L0_IRQ + cpu],
 cpumask_of(cpu));
+        if (ret)
+            pr_err("force failed %d\n", ret);
      } else {

          enable_percpu_irq(mct_irqs[MCT_L0_IRQ], 0);
      }
Signature · 2 lines
Samsung R&D Institute Poland
Samsung Electronics