S390's init_idle_preempt_count(p, cpu) doesn't actually let us initialize the
preempt_count of the requested CPU's idle task: it unconditionally writes
to the current CPU's. This clearly conflicts with idle_threads_init(),
which intends to initialize *all* the idle tasks, including their
preempt_count (or their CPU's, if the arch uses a per-CPU preempt_count).
Unfortunately, it seems the way s390 does things doesn't let us initialize
every possible CPU's preempt_count early on, as the pages where this
resides are only allocated when a CPU is brought up and are freed when it
is brought down.
Let the arch-specific code set a CPU's preempt_count when its lowcore is
allocated, and turn init_idle_preempt_count() into an empty stub.
Fixes: f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
Reported-by: Guenter Roeck <[email protected]>
Signed-off-by: Valentin Schneider <[email protected]>
---
arch/s390/include/asm/preempt.h | 16 ++++------------
arch/s390/kernel/setup.c | 1 +
arch/s390/kernel/smp.c | 1 +
3 files changed, 6 insertions(+), 12 deletions(-)
diff --git a/arch/s390/include/asm/preempt.h b/arch/s390/include/asm/preempt.h
index 23ff51be7e29..d9d5350cc3ec 100644
--- a/arch/s390/include/asm/preempt.h
+++ b/arch/s390/include/asm/preempt.h
@@ -29,12 +29,6 @@ static inline void preempt_count_set(int pc)
old, new) != old);
}
-#define init_task_preempt_count(p) do { } while (0)
-
-#define init_idle_preempt_count(p, cpu) do { \
- S390_lowcore.preempt_count = PREEMPT_DISABLED; \
-} while (0)
-
static inline void set_preempt_need_resched(void)
{
__atomic_and(~PREEMPT_NEED_RESCHED, &S390_lowcore.preempt_count);
@@ -88,12 +82,6 @@ static inline void preempt_count_set(int pc)
S390_lowcore.preempt_count = pc;
}
-#define init_task_preempt_count(p) do { } while (0)
-
-#define init_idle_preempt_count(p, cpu) do { \
- S390_lowcore.preempt_count = PREEMPT_DISABLED; \
-} while (0)
-
static inline void set_preempt_need_resched(void)
{
}
@@ -130,6 +118,10 @@ static inline bool should_resched(int preempt_offset)
#endif /* CONFIG_HAVE_MARCH_Z196_FEATURES */
+#define init_task_preempt_count(p) do { } while (0)
+/* Deferred to CPU bringup time */
+#define init_idle_preempt_count(p, cpu) do { } while (0)
+
#ifdef CONFIG_PREEMPTION
extern void preempt_schedule(void);
#define __preempt_schedule() preempt_schedule()
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index 5aab59ad5688..382d73da134c 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -466,6 +466,7 @@ static void __init setup_lowcore_dat_off(void)
lc->br_r1_trampoline = 0x07f1; /* br %r1 */
lc->return_lpswe = gen_lpswe(__LC_RETURN_PSW);
lc->return_mcck_lpswe = gen_lpswe(__LC_RETURN_MCCK_PSW);
+ lc->preempt_count = PREEMPT_DISABLED;
set_prefix((u32)(unsigned long) lc);
lowcore_ptr[0] = lc;
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 111909aeb8d2..1fb483e06a64 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -219,6 +219,7 @@ static int pcpu_alloc_lowcore(struct pcpu *pcpu, int cpu)
lc->br_r1_trampoline = 0x07f1; /* br %r1 */
lc->return_lpswe = gen_lpswe(__LC_RETURN_PSW);
lc->return_mcck_lpswe = gen_lpswe(__LC_RETURN_MCCK_PSW);
+ lc->preempt_count = PREEMPT_DISABLED;
if (nmi_alloc_per_cpu(lc))
goto out_stack;
lowcore_ptr[cpu] = lc;
--
2.25.1
On Wed, Jul 07, 2021 at 05:33:38PM +0100, Valentin Schneider wrote:
> S390's init_idle_preempt_count(p, cpu) doesn't actually let us initialize the
> preempt_count of the requested CPU's idle task: it unconditionally writes
> to the current CPU's. This clearly conflicts with idle_threads_init(),
> which intends to initialize *all* the idle tasks, including their
> preempt_count (or their CPU's, if the arch uses a per-CPU preempt_count).
>
> Unfortunately, it seems the way s390 does things doesn't let us initialize
> every possible CPU's preempt_count early on, as the pages where this
> resides are only allocated when a CPU is brought up and are freed when it
> is brought down.
>
> Let the arch-specific code set a CPU's preempt_count when its lowcore is
> allocated, and turn init_idle_preempt_count() into an empty stub.
>
> Fixes: f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
> Reported-by: Guenter Roeck <[email protected]>
> Signed-off-by: Valentin Schneider <[email protected]>
Tested-by: Guenter Roeck <[email protected]>
Guenter
On Wed, Jul 07, 2021 at 05:33:38PM +0100, Valentin Schneider wrote:
> S390's init_idle_preempt_count(p, cpu) doesn't actually let us initialize the
> preempt_count of the requested CPU's idle task: it unconditionally writes
> to the current CPU's. This clearly conflicts with idle_threads_init(),
> which intends to initialize *all* the idle tasks, including their
> preempt_count (or their CPU's, if the arch uses a per-CPU preempt_count).
>
> Unfortunately, it seems the way s390 does things doesn't let us initialize
> every possible CPU's preempt_count early on, as the pages where this
> resides are only allocated when a CPU is brought up and are freed when it
> is brought down.
>
> Let the arch-specific code set a CPU's preempt_count when its lowcore is
> allocated, and turn init_idle_preempt_count() into an empty stub.
>
> Fixes: f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
> Reported-by: Guenter Roeck <[email protected]>
> Signed-off-by: Valentin Schneider <[email protected]>
> ---
> arch/s390/include/asm/preempt.h | 16 ++++------------
> arch/s390/kernel/setup.c | 1 +
> arch/s390/kernel/smp.c | 1 +
> 3 files changed, 6 insertions(+), 12 deletions(-)
Reviewed-by: Heiko Carstens <[email protected]>
Vasily, can you pick this one up, please?
On Thu, Jul 08, 2021 at 04:17:10PM +0200, Heiko Carstens wrote:
> On Wed, Jul 07, 2021 at 05:33:38PM +0100, Valentin Schneider wrote:
> > S390's init_idle_preempt_count(p, cpu) doesn't actually let us initialize the
> > preempt_count of the requested CPU's idle task: it unconditionally writes
> > to the current CPU's. This clearly conflicts with idle_threads_init(),
> > which intends to initialize *all* the idle tasks, including their
> > preempt_count (or their CPU's, if the arch uses a per-CPU preempt_count).
> >
> > Unfortunately, it seems the way s390 does things doesn't let us initialize
> > every possible CPU's preempt_count early on, as the pages where this
> > resides are only allocated when a CPU is brought up and are freed when it
> > is brought down.
> >
> > Let the arch-specific code set a CPU's preempt_count when its lowcore is
> > allocated, and turn init_idle_preempt_count() into an empty stub.
> >
> > Fixes: f1a0a376ca0c ("sched/core: Initialize the idle task with preemption disabled")
> > Reported-by: Guenter Roeck <[email protected]>
> > Signed-off-by: Valentin Schneider <[email protected]>
> > ---
> > arch/s390/include/asm/preempt.h | 16 ++++------------
> > arch/s390/kernel/setup.c | 1 +
> > arch/s390/kernel/smp.c | 1 +
> > 3 files changed, 6 insertions(+), 12 deletions(-)
>
> Reviewed-by: Heiko Carstens <[email protected]>
>
> Vasily, can you pick this one up, please?
Will pick it up right away for rc1, thanks!