Powerpc currently resets a CPU's idle task preempt_count to 0 before said
task starts executing the secondary startup routine (and becomes an idle
task proper).
This conflicts with commit
f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
which initializes all of the idle tasks' preempt_count to PREEMPT_DISABLED
during smp_init(). Note that this was superfluous before said commit, as
back then the hotplug machinery would invoke init_idle() via
idle_thread_get(), which would have already reset the CPU's idle task's
preempt_count to PREEMPT_ENABLED.
Get rid of this preempt_count write.
Cc: Guenter Roeck <[email protected]>
Fixes: f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
Reported-by: Bharata B Rao <[email protected]>
Signed-off-by: Valentin Schneider <[email protected]>
---
arch/powerpc/platforms/cell/smp.c | 3 ---
arch/powerpc/platforms/pseries/smp.c | 5 +----
2 files changed, 1 insertion(+), 7 deletions(-)
diff --git a/arch/powerpc/platforms/cell/smp.c b/arch/powerpc/platforms/cell/smp.c
index c855a0aeb49c..d7ab868aab54 100644
--- a/arch/powerpc/platforms/cell/smp.c
+++ b/arch/powerpc/platforms/cell/smp.c
@@ -78,9 +78,6 @@ static inline int smp_startup_cpu(unsigned int lcpu)
pcpu = get_hard_smp_processor_id(lcpu);
- /* Fixup atomic count: it exited inside IRQ handler. */
- task_thread_info(paca_ptrs[lcpu]->__current)->preempt_count = 0;
-
/*
* If the RTAS start-cpu token does not exist then presume the
* cpu is already spinning.
diff --git a/arch/powerpc/platforms/pseries/smp.c b/arch/powerpc/platforms/pseries/smp.c
index 096629f54576..7ebf3382816a 100644
--- a/arch/powerpc/platforms/pseries/smp.c
+++ b/arch/powerpc/platforms/pseries/smp.c
@@ -105,10 +105,7 @@ static inline int smp_startup_cpu(unsigned int lcpu)
return 1;
}
- /* Fixup atomic count: it exited inside IRQ handler. */
- task_thread_info(paca_ptrs[lcpu]->__current)->preempt_count = 0;
-
- /*
+ /*
* If the RTAS start-cpu token does not exist then presume the
* cpu is already spinning.
*/
--
2.25.1
On Wed, Jul 07, 2021 at 07:38:31PM +0100, Valentin Schneider wrote:
> Powerpc currently resets a CPU's idle task preempt_count to 0 before said
> task starts executing the secondary startup routine (and becomes an idle
> task proper).
>
> This conflicts with commit
>
> f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
>
> which initializes all of the idle tasks' preempt_count to PREEMPT_DISABLED
> during smp_init(). Note that this was superfluous before said commit, as
> back then the hotplug machinery would invoke init_idle() via
> idle_thread_get(), which would have already reset the CPU's idle task's
> preempt_count to PREEMPT_ENABLED.
>
> Get rid of this preempt_count write.
>
> Cc: Guenter Roeck <[email protected]>
> Fixes: f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
> Reported-by: Bharata B Rao <[email protected]>
> Signed-off-by: Valentin Schneider <[email protected]>
Tested-by: Guenter Roeck <[email protected]>
> ---
> arch/powerpc/platforms/cell/smp.c | 3 ---
> arch/powerpc/platforms/pseries/smp.c | 5 +----
> 2 files changed, 1 insertion(+), 7 deletions(-)
>
> diff --git a/arch/powerpc/platforms/cell/smp.c b/arch/powerpc/platforms/cell/smp.c
> index c855a0aeb49c..d7ab868aab54 100644
> --- a/arch/powerpc/platforms/cell/smp.c
> +++ b/arch/powerpc/platforms/cell/smp.c
> @@ -78,9 +78,6 @@ static inline int smp_startup_cpu(unsigned int lcpu)
>
> pcpu = get_hard_smp_processor_id(lcpu);
>
> - /* Fixup atomic count: it exited inside IRQ handler. */
> - task_thread_info(paca_ptrs[lcpu]->__current)->preempt_count = 0;
> -
> /*
> * If the RTAS start-cpu token does not exist then presume the
> * cpu is already spinning.
> diff --git a/arch/powerpc/platforms/pseries/smp.c b/arch/powerpc/platforms/pseries/smp.c
> index 096629f54576..7ebf3382816a 100644
> --- a/arch/powerpc/platforms/pseries/smp.c
> +++ b/arch/powerpc/platforms/pseries/smp.c
> @@ -105,10 +105,7 @@ static inline int smp_startup_cpu(unsigned int lcpu)
> return 1;
> }
>
> - /* Fixup atomic count: it exited inside IRQ handler. */
> - task_thread_info(paca_ptrs[lcpu]->__current)->preempt_count = 0;
> -
> - /*
> + /*
> * If the RTAS start-cpu token does not exist then presume the
> * cpu is already spinning.
> */
> --
> 2.25.1
>
On Wed, Jul 07, 2021 at 07:38:31PM +0100, Valentin Schneider wrote:
> Powerpc currently resets a CPU's idle task preempt_count to 0 before said
> task starts executing the secondary startup routine (and becomes an idle
> task proper).
>
> This conflicts with commit
>
> f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
>
> which initializes all of the idle tasks' preempt_count to PREEMPT_DISABLED
> during smp_init(). Note that this was superfluous before said commit, as
> back then the hotplug machinery would invoke init_idle() via
> idle_thread_get(), which would have already reset the CPU's idle task's
> preempt_count to PREEMPT_ENABLED.
>
> Get rid of this preempt_count write.
>
> Cc: Guenter Roeck <[email protected]>
> Fixes: f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
> Reported-by: Bharata B Rao <[email protected]>
> Signed-off-by: Valentin Schneider <[email protected]>
> ---
> arch/powerpc/platforms/cell/smp.c | 3 ---
> arch/powerpc/platforms/pseries/smp.c | 5 +----
> 2 files changed, 1 insertion(+), 7 deletions(-)
The messages like "BUG: scheduling while atomic: swapper/1/0/0x00000000"
for each secondary CPU are no longer seen after this patch on powerpc.
Tested-by: Bharata B Rao <[email protected]>
On Wed, 7 Jul 2021 19:38:31 +0100, Valentin Schneider wrote:
> Powerpc currently resets a CPU's idle task preempt_count to 0 before said
> task starts executing the secondary startup routine (and becomes an idle
> task proper).
>
> This conflicts with commit
>
> f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
>
> [...]
Applied to powerpc/fixes.
[1/1] powerpc: preempt: Don't touch the idle task's preempt_count during hotplug
https://git.kernel.org/powerpc/c/2c669ef6979c370f98d4b876e54f19613c81e075
cheers