next: suspicious RCU usage message since commit 'rcu: Remove superfluous versions of rcu_read_lock_sched_held()'

Hi,

I see the following log message when running a qemu test for 'beagle' with omap2plus_defconfig.

===============================
[ INFO: suspicious RCU usage. ]
4.6.0-rc4-next-20160422 #1 Not tainted
-------------------------------
include/trace/events/power.h:328 suspicious rcu_dereference_check() usage!

other info that might help us debug this:

RCU used illegally from idle CPU!
rcu_scheduler_active = 1, debug_locks = 0
RCU used illegally from extended quiescent state!
no locks held by swapper/0/0.

stack backtrace:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc4-next-20160422 #1
Hardware name: Generic OMAP3-GP (Flattened Device Tree)
[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
[<c010b64c>] (show_stack) from [<c047acbc>] (dump_stack+0xa8/0xe0)
[<c047acbc>] (dump_stack) from [<c012bc10>] (pwrdm_set_next_pwrst+0xf8/0x1cc)
[<c012bc10>] (pwrdm_set_next_pwrst) from [<c01269fc>] (omap3_enter_idle_bm+0x1b8/0x1e8)
[<c01269fc>] (omap3_enter_idle_bm) from [<c05fa0b8>] (cpuidle_enter_state+0x84/0x408)
[<c05fa0b8>] (cpuidle_enter_state) from [<c0182c1c>] (cpu_startup_entry+0x1c8/0x3f0)
[<c0182c1c>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

bisect points to commit 'rcu: Remove superfluous versions of rcu_read_lock_sched_held()'. Bisect log is attached.

Quoted 29 lines · Guenter Roeck

Hi,

I see the following log message when running a qemu test for 'beagle' with omap2plus_defconfig.

===============================
[ INFO: suspicious RCU usage. ]
4.6.0-rc4-next-20160422 #1 Not tainted
-------------------------------
include/trace/events/power.h:328 suspicious rcu_dereference_check() usage!

other info that might help us debug this:

RCU used illegally from idle CPU!
rcu_scheduler_active = 1, debug_locks = 0
RCU used illegally from extended quiescent state!
no locks held by swapper/0/0.

stack backtrace:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc4-next-20160422 #1
Hardware name: Generic OMAP3-GP (Flattened Device Tree)
[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
[<c010b64c>] (show_stack) from [<c047acbc>] (dump_stack+0xa8/0xe0)
[<c047acbc>] (dump_stack) from [<c012bc10>] (pwrdm_set_next_pwrst+0xf8/0x1cc)
[<c012bc10>] (pwrdm_set_next_pwrst) from [<c01269fc>] (omap3_enter_idle_bm+0x1b8/0x1e8)
[<c01269fc>] (omap3_enter_idle_bm) from [<c05fa0b8>] (cpuidle_enter_state+0x84/0x408)
[<c05fa0b8>] (cpuidle_enter_state) from [<c0182c1c>] (cpu_startup_entry+0x1c8/0x3f0)
[<c0182c1c>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

Quoted 2 lines

bisect points to commit 'rcu: Remove superfluous versions of rcu_read_lock_sched_held()'. Bisect log is attached.

I believe that the real fix is not a revert of that commit, but rather that some of the tracing statements need an "_rcuidle" suffix.

Something like the following (untested, probably does not build) patch.

Thanx, Paul

------------------------------------------------------------------------

commit ca91304178e1cf53ee391236a0ac3969cc814e5f

Author: Paul E. McKenney <[email protected]>
Date:   Sun Apr 24 14:30:16 2016 -0700

arm: Use _rcuidle tracepoint to allow use from idle

Signed-off-by Paul E. McKenney <[email protected]>

arch/arm/mach-omap2/powerdomain.c +0 -0

index 78af6d8cf2e2..12b66b5bcc55 100644

Hi Paul,

Quoted 56 lines · Paul E. McKenney

Quoted 29 lines · Guenter Roeck

Hi,

I see the following log message when running a qemu test for 'beagle' with omap2plus_defconfig.

===============================
[ INFO: suspicious RCU usage. ]
4.6.0-rc4-next-20160422 #1 Not tainted
-------------------------------
include/trace/events/power.h:328 suspicious rcu_dereference_check() usage!

other info that might help us debug this:

RCU used illegally from idle CPU!
rcu_scheduler_active = 1, debug_locks = 0
RCU used illegally from extended quiescent state!
no locks held by swapper/0/0.

stack backtrace:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc4-next-20160422 #1
Hardware name: Generic OMAP3-GP (Flattened Device Tree)
[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
[<c010b64c>] (show_stack) from [<c047acbc>] (dump_stack+0xa8/0xe0)
[<c047acbc>] (dump_stack) from [<c012bc10>] (pwrdm_set_next_pwrst+0xf8/0x1cc)
[<c012bc10>] (pwrdm_set_next_pwrst) from [<c01269fc>] (omap3_enter_idle_bm+0x1b8/0x1e8)
[<c01269fc>] (omap3_enter_idle_bm) from [<c05fa0b8>] (cpuidle_enter_state+0x84/0x408)
[<c05fa0b8>] (cpuidle_enter_state) from [<c0182c1c>] (cpu_startup_entry+0x1c8/0x3f0)
[<c0182c1c>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

Quoted 2 lines

bisect points to commit 'rcu: Remove superfluous versions of rcu_read_lock_sched_held()'. Bisect log is attached.

I believe that the real fix is not a revert of that commit, but rather that some of the tracing statements need an "_rcuidle" suffix.

Something like the following (untested, probably does not build) patch.

Thanx, Paul

------------------------------------------------------------------------

commit ca91304178e1cf53ee391236a0ac3969cc814e5f

Author: Paul E. McKenney <[email protected]>
Date:   Sun Apr 24 14:30:16 2016 -0700

arm: Use _rcuidle tracepoint to allow use from idle

Signed-off-by Paul E. McKenney <[email protected]>

arch/arm/mach-omap2/powerdomain.c +0 -0

index 78af6d8cf2e2..12b66b5bcc55 100644

arch/arm/mach-omap2/powerdomain.c int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst) · line 530

Quoted 10 lines

arch/arm/mach-omap2/powerdomain.c +2 -2

@@ -523,8 +523,8 @@ int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst)
 
 	if (arch_pwrdm && arch_pwrdm->pwrdm_set_next_pwrst) {
 		/* Trace the pwrdm desired target state */
-		trace_power_domain_target(pwrdm->name, pwrst,
-					  smp_processor_id());
+		trace_power_domain_target_rcuidle(pwrdm->name, pwrst,
+						  smp_processor_id());
 		/* Program the pwrdm desired target state */
 		ret = arch_pwrdm->pwrdm_set_next_pwrst(pwrdm, pwrst);
 	}

It does build. After applying it, I get a different traceback.

[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14) [<c010b64c>] (show_stack) from [<c047ac3c>] (dump_stack+0xa8/0xe0) [<c047ac3c>] (dump_stack) from [<c012c340>] (_pwrdm_state_switch+0x188/0x32c) [<c012c340>] (_pwrdm_state_switch) from [<c012c4f0>] (_pwrdm_post_transition_cb+0xc/0x14) [<c012c4f0>] (_pwrdm_post_transition_cb) from [<c012ba74>] (pwrdm_for_each+0x30/0x5c) [<c012ba74>] (pwrdm_for_each) from [<c012c72c>] (pwrdm_post_transition+0x24/0x30) [<c012c72c>] (pwrdm_post_transition) from [<c012548c>] (omap_sram_idle+0xfc/0x240) [<c012548c>] (omap_sram_idle) from [<c0126934>] (omap3_enter_idle_bm+0xf0/0x1e8) [<c0126934>] (omap3_enter_idle_bm) from [<c05fa038>] (cpuidle_enter_state+0x84/0x408) [<c05fa038>] (cpuidle_enter_state) from [<c0182b90>] (cpu_startup_entry+0x1c8/0x3f0) [<c0182b90>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Quoted 84 lines · Guenter Roeck

Hi Paul,

Quoted 56 lines · Paul E. McKenney

Quoted 29 lines · Guenter Roeck

Hi,

I see the following log message when running a qemu test for 'beagle' with omap2plus_defconfig.

===============================
[ INFO: suspicious RCU usage. ]
4.6.0-rc4-next-20160422 #1 Not tainted
-------------------------------
include/trace/events/power.h:328 suspicious rcu_dereference_check() usage!

other info that might help us debug this:

RCU used illegally from idle CPU!
rcu_scheduler_active = 1, debug_locks = 0
RCU used illegally from extended quiescent state!
no locks held by swapper/0/0.

stack backtrace:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc4-next-20160422 #1
Hardware name: Generic OMAP3-GP (Flattened Device Tree)
[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
[<c010b64c>] (show_stack) from [<c047acbc>] (dump_stack+0xa8/0xe0)
[<c047acbc>] (dump_stack) from [<c012bc10>] (pwrdm_set_next_pwrst+0xf8/0x1cc)
[<c012bc10>] (pwrdm_set_next_pwrst) from [<c01269fc>] (omap3_enter_idle_bm+0x1b8/0x1e8)
[<c01269fc>] (omap3_enter_idle_bm) from [<c05fa0b8>] (cpuidle_enter_state+0x84/0x408)
[<c05fa0b8>] (cpuidle_enter_state) from [<c0182c1c>] (cpu_startup_entry+0x1c8/0x3f0)
[<c0182c1c>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

Quoted 2 lines

bisect points to commit 'rcu: Remove superfluous versions of rcu_read_lock_sched_held()'. Bisect log is attached.

I believe that the real fix is not a revert of that commit, but rather that some of the tracing statements need an "_rcuidle" suffix.

Something like the following (untested, probably does not build) patch.

Thanx, Paul

------------------------------------------------------------------------

commit ca91304178e1cf53ee391236a0ac3969cc814e5f

Author: Paul E. McKenney <[email protected]>
Date:   Sun Apr 24 14:30:16 2016 -0700

arm: Use _rcuidle tracepoint to allow use from idle

Signed-off-by Paul E. McKenney <[email protected]>

arch/arm/mach-omap2/powerdomain.c +0 -0

index 78af6d8cf2e2..12b66b5bcc55 100644

arch/arm/mach-omap2/powerdomain.c int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst) · line 530

Quoted 10 lines

arch/arm/mach-omap2/powerdomain.c +2 -2

@@ -523,8 +523,8 @@ int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst)
 
  	if (arch_pwrdm && arch_pwrdm->pwrdm_set_next_pwrst) {
  		/* Trace the pwrdm desired target state */
-		trace_power_domain_target(pwrdm->name, pwrst,
-					  smp_processor_id());
+		trace_power_domain_target_rcuidle(pwrdm->name, pwrst,
+						  smp_processor_id());
  		/* Program the pwrdm desired target state */
  		ret = arch_pwrdm->pwrdm_set_next_pwrst(pwrdm, pwrst);
  	}

It does build. After applying it, I get a different traceback.

[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14) [<c010b64c>] (show_stack) from [<c047ac3c>] (dump_stack+0xa8/0xe0) [<c047ac3c>] (dump_stack) from [<c012c340>] (_pwrdm_state_switch+0x188/0x32c) [<c012c340>] (_pwrdm_state_switch) from [<c012c4f0>] (_pwrdm_post_transition_cb+0xc/0x14) [<c012c4f0>] (_pwrdm_post_transition_cb) from [<c012ba74>] (pwrdm_for_each+0x30/0x5c) [<c012ba74>] (pwrdm_for_each) from [<c012c72c>] (pwrdm_post_transition+0x24/0x30) [<c012c72c>] (pwrdm_post_transition) from [<c012548c>] (omap_sram_idle+0xfc/0x240) [<c012548c>] (omap_sram_idle) from [<c0126934>] (omap3_enter_idle_bm+0xf0/0x1e8) [<c0126934>] (omap3_enter_idle_bm) from [<c05fa038>] (cpuidle_enter_state+0x84/0x408) [<c05fa038>] (cpuidle_enter_state) from [<c0182b90>] (cpu_startup_entry+0x1c8/0x3f0) [<c0182b90>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

Quoted 2 lines

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Very good!

(And yes, you normally find these one at a time...)

Quoted 85 lines · Paul E. McKenney

Quoted 84 lines · Guenter Roeck

Hi Paul,

Quoted 56 lines · Paul E. McKenney

Quoted 29 lines · Guenter Roeck

Hi,

I see the following log message when running a qemu test for 'beagle' with omap2plus_defconfig.

===============================
[ INFO: suspicious RCU usage. ]
4.6.0-rc4-next-20160422 #1 Not tainted
-------------------------------
include/trace/events/power.h:328 suspicious rcu_dereference_check() usage!

other info that might help us debug this:

RCU used illegally from idle CPU!
rcu_scheduler_active = 1, debug_locks = 0
RCU used illegally from extended quiescent state!
no locks held by swapper/0/0.

stack backtrace:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc4-next-20160422 #1
Hardware name: Generic OMAP3-GP (Flattened Device Tree)
[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
[<c010b64c>] (show_stack) from [<c047acbc>] (dump_stack+0xa8/0xe0)
[<c047acbc>] (dump_stack) from [<c012bc10>] (pwrdm_set_next_pwrst+0xf8/0x1cc)
[<c012bc10>] (pwrdm_set_next_pwrst) from [<c01269fc>] (omap3_enter_idle_bm+0x1b8/0x1e8)
[<c01269fc>] (omap3_enter_idle_bm) from [<c05fa0b8>] (cpuidle_enter_state+0x84/0x408)
[<c05fa0b8>] (cpuidle_enter_state) from [<c0182c1c>] (cpu_startup_entry+0x1c8/0x3f0)
[<c0182c1c>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

Quoted 2 lines

bisect points to commit 'rcu: Remove superfluous versions of rcu_read_lock_sched_held()'. Bisect log is attached.

I believe that the real fix is not a revert of that commit, but rather that some of the tracing statements need an "_rcuidle" suffix.

Something like the following (untested, probably does not build) patch.

Thanx, Paul

------------------------------------------------------------------------

commit ca91304178e1cf53ee391236a0ac3969cc814e5f

Author: Paul E. McKenney <[email protected]>
Date:   Sun Apr 24 14:30:16 2016 -0700

arm: Use _rcuidle tracepoint to allow use from idle

Signed-off-by Paul E. McKenney <[email protected]>

arch/arm/mach-omap2/powerdomain.c +0 -0

index 78af6d8cf2e2..12b66b5bcc55 100644

arch/arm/mach-omap2/powerdomain.c int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst) · line 530

Quoted 10 lines

arch/arm/mach-omap2/powerdomain.c +2 -2

@@ -523,8 +523,8 @@ int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst)
 
 	if (arch_pwrdm && arch_pwrdm->pwrdm_set_next_pwrst) {
 		/* Trace the pwrdm desired target state */
-		trace_power_domain_target(pwrdm->name, pwrst,
-					  smp_processor_id());
+		trace_power_domain_target_rcuidle(pwrdm->name, pwrst,
+						  smp_processor_id());
 		/* Program the pwrdm desired target state */
 		ret = arch_pwrdm->pwrdm_set_next_pwrst(pwrdm, pwrst);
 	}

It does build. After applying it, I get a different traceback.

[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14) [<c010b64c>] (show_stack) from [<c047ac3c>] (dump_stack+0xa8/0xe0) [<c047ac3c>] (dump_stack) from [<c012c340>] (_pwrdm_state_switch+0x188/0x32c) [<c012c340>] (_pwrdm_state_switch) from [<c012c4f0>] (_pwrdm_post_transition_cb+0xc/0x14) [<c012c4f0>] (_pwrdm_post_transition_cb) from [<c012ba74>] (pwrdm_for_each+0x30/0x5c) [<c012ba74>] (pwrdm_for_each) from [<c012c72c>] (pwrdm_post_transition+0x24/0x30) [<c012c72c>] (pwrdm_post_transition) from [<c012548c>] (omap_sram_idle+0xfc/0x240) [<c012548c>] (omap_sram_idle) from [<c0126934>] (omap3_enter_idle_bm+0xf0/0x1e8) [<c0126934>] (omap3_enter_idle_bm) from [<c05fa038>] (cpuidle_enter_state+0x84/0x408) [<c05fa038>] (cpuidle_enter_state) from [<c0182b90>] (cpu_startup_entry+0x1c8/0x3f0) [<c0182b90>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

Quoted 7 lines

Quoted 2 lines

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Very good!

(And yes, you normally find these one at a time...)

Are you going to submit a formal patch ?

Quoted 89 lines · Guenter Roeck

Quoted 88 lines · Paul E. McKenney

Quoted 86 lines · Guenter Roeck

Hi Paul,

Quoted 56 lines · Paul E. McKenney

Quoted 29 lines · Guenter Roeck

Hi,

I see the following log message when running a qemu test for 'beagle' with omap2plus_defconfig.

===============================
[ INFO: suspicious RCU usage. ]
4.6.0-rc4-next-20160422 #1 Not tainted
-------------------------------
include/trace/events/power.h:328 suspicious rcu_dereference_check() usage!

other info that might help us debug this:

RCU used illegally from idle CPU!
rcu_scheduler_active = 1, debug_locks = 0
RCU used illegally from extended quiescent state!
no locks held by swapper/0/0.

stack backtrace:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc4-next-20160422 #1
Hardware name: Generic OMAP3-GP (Flattened Device Tree)
[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
[<c010b64c>] (show_stack) from [<c047acbc>] (dump_stack+0xa8/0xe0)
[<c047acbc>] (dump_stack) from [<c012bc10>] (pwrdm_set_next_pwrst+0xf8/0x1cc)
[<c012bc10>] (pwrdm_set_next_pwrst) from [<c01269fc>] (omap3_enter_idle_bm+0x1b8/0x1e8)
[<c01269fc>] (omap3_enter_idle_bm) from [<c05fa0b8>] (cpuidle_enter_state+0x84/0x408)
[<c05fa0b8>] (cpuidle_enter_state) from [<c0182c1c>] (cpu_startup_entry+0x1c8/0x3f0)
[<c0182c1c>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

Quoted 2 lines

bisect points to commit 'rcu: Remove superfluous versions of rcu_read_lock_sched_held()'. Bisect log is attached.

I believe that the real fix is not a revert of that commit, but rather that some of the tracing statements need an "_rcuidle" suffix.

Something like the following (untested, probably does not build) patch.

Thanx, Paul

------------------------------------------------------------------------

commit ca91304178e1cf53ee391236a0ac3969cc814e5f

Author: Paul E. McKenney <[email protected]>
Date:   Sun Apr 24 14:30:16 2016 -0700

arm: Use _rcuidle tracepoint to allow use from idle

Signed-off-by Paul E. McKenney <[email protected]>

arch/arm/mach-omap2/powerdomain.c +0 -0

index 78af6d8cf2e2..12b66b5bcc55 100644

arch/arm/mach-omap2/powerdomain.c int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst) · line 530

Quoted 10 lines

arch/arm/mach-omap2/powerdomain.c +2 -2

@@ -523,8 +523,8 @@ int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst)
 
  	if (arch_pwrdm && arch_pwrdm->pwrdm_set_next_pwrst) {
  		/* Trace the pwrdm desired target state */
-		trace_power_domain_target(pwrdm->name, pwrst,
-					  smp_processor_id());
+		trace_power_domain_target_rcuidle(pwrdm->name, pwrst,
+						  smp_processor_id());
  		/* Program the pwrdm desired target state */
  		ret = arch_pwrdm->pwrdm_set_next_pwrst(pwrdm, pwrst);
  	}

It does build. After applying it, I get a different traceback.

[<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14) [<c010b64c>] (show_stack) from [<c047ac3c>] (dump_stack+0xa8/0xe0) [<c047ac3c>] (dump_stack) from [<c012c340>] (_pwrdm_state_switch+0x188/0x32c) [<c012c340>] (_pwrdm_state_switch) from [<c012c4f0>] (_pwrdm_post_transition_cb+0xc/0x14) [<c012c4f0>] (_pwrdm_post_transition_cb) from [<c012ba74>] (pwrdm_for_each+0x30/0x5c) [<c012ba74>] (pwrdm_for_each) from [<c012c72c>] (pwrdm_post_transition+0x24/0x30) [<c012c72c>] (pwrdm_post_transition) from [<c012548c>] (omap_sram_idle+0xfc/0x240) [<c012548c>] (omap_sram_idle) from [<c0126934>] (omap3_enter_idle_bm+0xf0/0x1e8) [<c0126934>] (omap3_enter_idle_bm) from [<c05fa038>] (cpuidle_enter_state+0x84/0x408) [<c05fa038>] (cpuidle_enter_state) from [<c0182b90>] (cpu_startup_entry+0x1c8/0x3f0) [<c0182b90>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Quoted 5 lines

Quoted 4 lines

Very good!

(And yes, you normally find these one at a time...)

Are you going to submit a formal patch ?

I can, but please feel free to send mine along with yours, if you wish.

Quoted 3 lines · Paul E. McKenney

Quoted 2 lines · Guenter Roeck

Quoted 1 line · Paul E. McKenney

On Sun, Apr 24, 2016 at 04:56:38PM -0700, Guenter Roeck wrote:

Quoted 5 lines

Quoted 5 lines

Quoted 5 lines

Quoted 2 lines

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Very good!

Quoted 6 lines

Quoted 3 lines

Quoted 2 lines

(And yes, you normally find these one at a time...)

Are you going to submit a formal patch ?

I can, but please feel free to send mine along with yours, if you wish.

I think it would be best if you send a single patch which fixes both calls.

Quoted 4 lines · Guenter Roeck

Quoted 3 lines · Paul E. McKenney

Quoted 2 lines · Guenter Roeck

Quoted 1 line · Paul E. McKenney

On Sun, Apr 24, 2016 at 04:56:38PM -0700, Guenter Roeck wrote:

[ . . . ]

Quoted 9 lines

Quoted 9 lines

Quoted 8 lines

Quoted 7 lines

Quoted 2 lines

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Very good!

(And yes, you normally find these one at a time...)

Are you going to submit a formal patch ?

Quoted 3 lines

Quoted 2 lines

I can, but please feel free to send mine along with yours, if you wish.

I think it would be best if you send a single patch which fixes both calls.

Like this one?

If so, could you please run it to make sure that it actually fixes the problem? And if it does, would you be willing to give me a Tested-by?

Thanx, Paul

------------------------------------------------------------------------

commit e3c9e1a46984b8c3e4204f138aced7630c4b85c2

Author: Paul E. McKenney <[email protected]>
Date:   Sun Apr 24 14:30:16 2016 -0700

    arm: Use _rcuidle tracepoint to allow use from idle
    
    Testing on ARM encountered the following pair of lockdep-RCU splats:
    
    ------------------------------------------------------------------------
    
    ===============================
    [ INFO: suspicious RCU usage. ]
    4.6.0-rc4-next-20160422 #1 Not tainted
    -------------------------------
    include/trace/events/power.h:328 suspicious rcu_dereference_check() usage!
    
    other info that might help us debug this:
    
    RCU used illegally from idle CPU!
    rcu_scheduler_active = 1, debug_locks = 0
    RCU used illegally from extended quiescent state!
    no locks held by swapper/0/0.
    
    stack backtrace:
    CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc4-next-20160422 #1
    Hardware name: Generic OMAP3-GP (Flattened Device Tree)
    [<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
    [<c010b64c>] (show_stack) from [<c047acbc>] (dump_stack+0xa8/0xe0)
    [<c047acbc>] (dump_stack) from [<c012bc10>] (pwrdm_set_next_pwrst+0xf8/0x1cc)
    [<c012bc10>] (pwrdm_set_next_pwrst) from [<c01269fc>] (omap3_enter_idle_bm+0x1b8/0x1e8)
    [<c01269fc>] (omap3_enter_idle_bm) from [<c05fa0b8>] (cpuidle_enter_state+0x84/0x408)
    [<c05fa0b8>] (cpuidle_enter_state) from [<c0182c1c>] (cpu_startup_entry+0x1c8/0x3f0)
    [<c0182c1c>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)
    
    ------------------------------------------------------------------------
    
    [<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
    [<c010b64c>] (show_stack) from [<c047ac3c>] (dump_stack+0xa8/0xe0)
    [<c047ac3c>] (dump_stack) from [<c012c340>] (_pwrdm_state_switch+0x188/0x32c)
    [<c012c340>] (_pwrdm_state_switch) from [<c012c4f0>] (_pwrdm_post_transition_cb+0xc/0x14)
    [<c012c4f0>] (_pwrdm_post_transition_cb) from [<c012ba74>] (pwrdm_for_each+0x30/0x5c)
    [<c012ba74>] (pwrdm_for_each) from [<c012c72c>] (pwrdm_post_transition+0x24/0x30)
    [<c012c72c>] (pwrdm_post_transition) from [<c012548c>] (omap_sram_idle+0xfc/0x240)
    [<c012548c>] (omap_sram_idle) from [<c0126934>] (omap3_enter_idle_bm+0xf0/0x1e8)
    [<c0126934>] (omap3_enter_idle_bm) from [<c05fa038>] (cpuidle_enter_state+0x84/0x408)
    [<c05fa038>] (cpuidle_enter_state) from [<c0182b90>] (cpu_startup_entry+0x1c8/0x3f0)
    [<c0182b90>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

    ------------------------------------------------------------------------
    
    These are caused by event tracing from the idle loop.  This commit
    therefore adds the _rcuidle suffix to make RCU aware of this implicit
    use of RCU by event tracing, thus preventing both splats.
    
    Reported-by: Guenter Roeck <[email protected]>

Signed-off-by Paul E. McKenney <[email protected]>

Cc Tony Lindgren <[email protected]>

Cc Russell King <[email protected]>

Cc Steven Rostedt <[email protected]>

Cc <[email protected]>

Cc <[email protected]>

arch/arm/mach-omap2/powerdomain.c +5 -4

index 78af6d8cf2e2..daf2753de7aa 100644
@@ -186,8 +186,9 @@ static int _pwrdm_state_switch(struct powerdomain *pwrdm, int flag)
 			trace_state = (PWRDM_TRACE_STATES_FLAG |
 				       ((next & OMAP_POWERSTATE_MASK) << 8) |
 				       ((prev & OMAP_POWERSTATE_MASK) << 0));
-			trace_power_domain_target(pwrdm->name, trace_state,
-						  smp_processor_id());
+			trace_power_domain_target_rcuidle(pwrdm->name,
+							  trace_state,
+							  smp_processor_id());
 		}
 		break;
 	default:

@@ -523,8 +524,8 @@ int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst)
 
 	if (arch_pwrdm && arch_pwrdm->pwrdm_set_next_pwrst) {
 		/* Trace the pwrdm desired target state */
-		trace_power_domain_target(pwrdm->name, pwrst,
-					  smp_processor_id());
+		trace_power_domain_target_rcuidle(pwrdm->name, pwrst,
+						  smp_processor_id());
 		/* Program the pwrdm desired target state */
 		ret = arch_pwrdm->pwrdm_set_next_pwrst(pwrdm, pwrst);
 	}

Quoted 19 lines · Paul E. McKenney

Quoted 4 lines · Guenter Roeck

Quoted 3 lines · Paul E. McKenney

Quoted 2 lines · Guenter Roeck

Quoted 1 line · Paul E. McKenney

On Sun, Apr 24, 2016 at 04:56:38PM -0700, Guenter Roeck wrote:

[ . . . ]

Quoted 11 lines

Quoted 11 lines

Quoted 8 lines

Quoted 7 lines

Quoted 2 lines

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Very good!

(And yes, you normally find these one at a time...)

Are you going to submit a formal patch ?

I can, but please feel free to send mine along with yours, if you wish.

Quoted 7 lines

Quoted 1 line

I think it would be best if you send a single patch which fixes both calls.

Like this one?

If so, could you please run it to make sure that it actually fixes the problem? And if it does, would you be willing to give me a Tested-by?

It does. Tested-by: inline below.

Thanks,
Guenter

Quoted 52 lines

Thanx, Paul

------------------------------------------------------------------------

commit e3c9e1a46984b8c3e4204f138aced7630c4b85c2

Author: Paul E. McKenney <[email protected]>
Date:   Sun Apr 24 14:30:16 2016 -0700

    arm: Use _rcuidle tracepoint to allow use from idle
    
    Testing on ARM encountered the following pair of lockdep-RCU splats:
    
    ------------------------------------------------------------------------
    
    ===============================
    [ INFO: suspicious RCU usage. ]
    4.6.0-rc4-next-20160422 #1 Not tainted
    -------------------------------
    include/trace/events/power.h:328 suspicious rcu_dereference_check() usage!
    
    other info that might help us debug this:
    
    RCU used illegally from idle CPU!
    rcu_scheduler_active = 1, debug_locks = 0
    RCU used illegally from extended quiescent state!
    no locks held by swapper/0/0.
    
    stack backtrace:
    CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc4-next-20160422 #1
    Hardware name: Generic OMAP3-GP (Flattened Device Tree)
    [<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
    [<c010b64c>] (show_stack) from [<c047acbc>] (dump_stack+0xa8/0xe0)
    [<c047acbc>] (dump_stack) from [<c012bc10>] (pwrdm_set_next_pwrst+0xf8/0x1cc)
    [<c012bc10>] (pwrdm_set_next_pwrst) from [<c01269fc>] (omap3_enter_idle_bm+0x1b8/0x1e8)
    [<c01269fc>] (omap3_enter_idle_bm) from [<c05fa0b8>] (cpuidle_enter_state+0x84/0x408)
    [<c05fa0b8>] (cpuidle_enter_state) from [<c0182c1c>] (cpu_startup_entry+0x1c8/0x3f0)
    [<c0182c1c>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)
    
    ------------------------------------------------------------------------
    
    [<c010f55c>] (unwind_backtrace) from [<c010b64c>] (show_stack+0x10/0x14)
    [<c010b64c>] (show_stack) from [<c047ac3c>] (dump_stack+0xa8/0xe0)
    [<c047ac3c>] (dump_stack) from [<c012c340>] (_pwrdm_state_switch+0x188/0x32c)
    [<c012c340>] (_pwrdm_state_switch) from [<c012c4f0>] (_pwrdm_post_transition_cb+0xc/0x14)
    [<c012c4f0>] (_pwrdm_post_transition_cb) from [<c012ba74>] (pwrdm_for_each+0x30/0x5c)
    [<c012ba74>] (pwrdm_for_each) from [<c012c72c>] (pwrdm_post_transition+0x24/0x30)
    [<c012c72c>] (pwrdm_post_transition) from [<c012548c>] (omap_sram_idle+0xfc/0x240)
    [<c012548c>] (omap_sram_idle) from [<c0126934>] (omap3_enter_idle_bm+0xf0/0x1e8)
    [<c0126934>] (omap3_enter_idle_bm) from [<c05fa038>] (cpuidle_enter_state+0x84/0x408)
    [<c05fa038>] (cpuidle_enter_state) from [<c0182b90>] (cpu_startup_entry+0x1c8/0x3f0)
    [<c0182b90>] (cpu_startup_entry) from [<c0b00c20>] (start_kernel+0x354/0x3cc)

Quoted 7 lines

    ------------------------------------------------------------------------
    
    These are caused by event tracing from the idle loop.  This commit
    therefore adds the _rcuidle suffix to make RCU aware of this implicit
    use of RCU by event tracing, thus preventing both splats.
    
    Reported-by: Guenter Roeck <[email protected]>

Tested-by Guenter Roeck <[email protected]>

Quoted 35 lines

Signed-off-by Paul E. McKenney <[email protected]>

Cc Tony Lindgren <[email protected]>

Cc Russell King <[email protected]>

Cc Steven Rostedt <[email protected]>

Cc <[email protected]>

Cc <[email protected]>

arch/arm/mach-omap2/powerdomain.c +5 -4

index 78af6d8cf2e2..daf2753de7aa 100644
@@ -186,8 +186,9 @@ static int _pwrdm_state_switch(struct powerdomain *pwrdm, int flag)
 			trace_state = (PWRDM_TRACE_STATES_FLAG |
 				       ((next & OMAP_POWERSTATE_MASK) << 8) |
 				       ((prev & OMAP_POWERSTATE_MASK) << 0));
-			trace_power_domain_target(pwrdm->name, trace_state,
-						  smp_processor_id());
+			trace_power_domain_target_rcuidle(pwrdm->name,
+							  trace_state,
+							  smp_processor_id());
 		}
 		break;
 	default:

@@ -523,8 +524,8 @@ int pwrdm_set_next_pwrst(struct powerdomain *pwrdm, u8 pwrst)
 
 	if (arch_pwrdm && arch_pwrdm->pwrdm_set_next_pwrst) {
 		/* Trace the pwrdm desired target state */
-		trace_power_domain_target(pwrdm->name, pwrst,
-					  smp_processor_id());
+		trace_power_domain_target_rcuidle(pwrdm->name, pwrst,
+						  smp_processor_id());
 		/* Program the pwrdm desired target state */
 		ret = arch_pwrdm->pwrdm_set_next_pwrst(pwrdm, pwrst);
 	}

Quoted 20 lines · Guenter Roeck

Quoted 19 lines · Paul E. McKenney

Quoted 4 lines · Guenter Roeck

Quoted 3 lines · Paul E. McKenney

Quoted 2 lines · Guenter Roeck

Quoted 1 line · Paul E. McKenney

On Sun, Apr 24, 2016 at 04:56:38PM -0700, Guenter Roeck wrote:

[ . . . ]

Quoted 11 lines

Quoted 11 lines

Quoted 8 lines

Quoted 7 lines

Quoted 2 lines

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Very good!

(And yes, you normally find these one at a time...)

Are you going to submit a formal patch ?

I can, but please feel free to send mine along with yours, if you wish.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

I think it would be best if you send a single patch which fixes both calls.

Like this one?

If so, could you please run it to make sure that it actually fixes the problem? And if it does, would you be willing to give me a Tested-by?

It does. Tested-by: inline below.

Got it, thank you!

If the ARM guys are willing to take this, it might hit the next merge window, or perhaps they will take it as an exception. If I push it up my usual route, it will be a bit later.

I just now sent it out, so hopefully they will grab it. ;-)

Quoted 20 lines · Guenter Roeck

Quoted 19 lines · Paul E. McKenney

Quoted 4 lines · Guenter Roeck

Quoted 3 lines · Paul E. McKenney

Quoted 2 lines · Guenter Roeck

Quoted 1 line · Paul E. McKenney

On Sun, Apr 24, 2016 at 04:56:38PM -0700, Guenter Roeck wrote:

[ . . . ]

Quoted 11 lines

Quoted 11 lines

Quoted 8 lines

Quoted 7 lines

Quoted 2 lines

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Very good!

(And yes, you normally find these one at a time...)

Are you going to submit a formal patch ?

I can, but please feel free to send mine along with yours, if you wish.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

I think it would be best if you send a single patch which fixes both calls.

Like this one?

If so, could you please run it to make sure that it actually fixes the problem? And if it does, would you be willing to give me a Tested-by?

It does. Tested-by: inline below.

Got it, thank you!

If the ARM guys are willing to take this, it might hit the next merge window, or perhaps they will take it as an exception. If I push it up my usual route, it will be a bit later.

I just now sent it out, so hopefully they will grab it. ;-)

Quoted 30 lines · Paul E. McKenney

Quoted 20 lines · Guenter Roeck

Quoted 19 lines · Paul E. McKenney

Quoted 4 lines · Guenter Roeck

Quoted 3 lines · Paul E. McKenney

Quoted 2 lines · Guenter Roeck

Quoted 1 line · Paul E. McKenney

On Sun, Apr 24, 2016 at 04:56:38PM -0700, Guenter Roeck wrote:

[ . . . ]

Quoted 11 lines

Quoted 11 lines

Quoted 8 lines

Quoted 7 lines

Quoted 2 lines

After making the same change in _pwrdm_state_switch(), the traceback is gone from my tests (beagle, beagle-xm, and overo-tobi).

Very good!

(And yes, you normally find these one at a time...)

Are you going to submit a formal patch ?

I can, but please feel free to send mine along with yours, if you wish.

Quoted 8 lines

Quoted 7 lines

Quoted 1 line

I think it would be best if you send a single patch which fixes both calls.

Like this one?

If so, could you please run it to make sure that it actually fixes the problem? And if it does, would you be willing to give me a Tested-by?

It does. Tested-by: inline below.

Quoted 8 lines

Got it, thank you!

If the ARM guys are willing to take this, it might hit the next merge window, or perhaps they will take it as an exception. If I push it up my usual route, it will be a bit later.

I just now sent it out, so hopefully they will grab it. ;-)

The problem is still seen in next-20160513, so it looks like the patch was not accepted.

I recently learned that arm has a special way of submitting patches. See
http://www.arm.linux.org.uk/developer/patches/ for details. If I understand correctly, you'll have to send the patch to [email protected], and it has to be formatted correctly (eg no "[PATCH]" in the subject line, and some other information added). I never tried it myself, so I don't really know how exactly it works.