| Message ID | 20220209153535.818830-8-mark.rutland@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | arm64 / sched/preempt: support PREEMPT_DYNAMIC with static keys | expand |
On Wed, Feb 09, 2022 at 03:35:35PM +0000, Mark Rutland wrote: > This patch enables support for PREEMPT_DYNAMIC on arm64, allowing the > preemption model to be chosen at boot time. > > Specifically, this patch selects HAVE_PREEMPT_DYNAMIC_KEY, so that each > preemption function is an out-of-line call with an early return > depending upon a static key. This leaves almost all the codegen up to > the compiler, and side-steps a number of pain points with static calls > (e.g. interaction with CFI schemes). This should have no worse overhead > than using non-inline static calls, as those use out-of-line trampolines > with early returns. > > For example, the dynamic_cond_resched() wrapper looks as follows when > enabled. When disabled, the first `B` is replaced with a `NOP`, > resulting in an early return. > > | <dynamic_cond_resched>: > | bti c > | b <dynamic_cond_resched+0x10> // or `nop` > | mov w0, #0x0 > | ret > | mrs x0, sp_el0 > | ldr x0, [x0, #8] > | cbnz x0, <dynamic_cond_resched+0x8> > | paciasp > | stp x29, x30, [sp, #-16]! > | mov x29, sp > | bl <preempt_schedule_common> > | mov w0, #0x1 > | ldp x29, x30, [sp], #16 > | autiasp > | ret > > ... compared to the regular form of the function: > > | <__cond_resched>: > | bti c > | mrs x0, sp_el0 > | ldr x1, [x0, #8] > | cbz x1, <__cond_resched+0x18> > | mov w0, #0x0 > | ret > | paciasp > | stp x29, x30, [sp, #-16]! > | mov x29, sp > | bl <preempt_schedule_common> > | mov w0, #0x1 > | ldp x29, x30, [sp], #16 > | autiasp > | ret > > Since arm64 does not yet use the generic entry code, we must define our > own `sk_dynamic_irqentry_exit_cond_resched`, which will be > enabled/disabled by the common code in kernel/sched/core.c. All other > preemption functions and associated static keys are defined there. > > Note that PREEMPT_DYNAMIC is `def bool y`, so this will default to > enabled. > > Signed-off-by: Mark Rutland <mark.rutland@arm.com> > Cc: Ard Biesheuvel <ardb@kernel.org> > Cc: Catalin Marinas <catalin.marinas@arm.com> > Cc: Frederic Weisbecker <frederic@kernel.org> > Cc: Ingo Molnar <mingo@redhat.com> > Cc: Joey Gouly <joey.gouly@arm.com> > Cc: Juri Lelli <juri.lelli@redhat.com> > Cc: Peter Zijlstra <peterz@infradead.org> > Cc: Valentin Schneider <valentin.schneider@arm.com> > Cc: Will Deacon <will@kernel.org> Acked-by: Catalin Marinas <catalin.marinas@arm.com>
On Wed, Feb 09, 2022 at 03:35:35PM +0000, Mark Rutland wrote: > Note that PREEMPT_DYNAMIC is `def bool y`, so this will default to > enabled. It should probably be "def_bool y if HAVE_STATIC_CALL_INLINE"... Thanks.
On Wed, Feb 09, 2022 at 08:57:09PM +0100, Frederic Weisbecker wrote: > On Wed, Feb 09, 2022 at 03:35:35PM +0000, Mark Rutland wrote: > > Note that PREEMPT_DYNAMIC is `def bool y`, so this will default to > > enabled. > > It should probably be "def_bool y if HAVE_STATIC_CALL_INLINE"... Sure; I'm more than happy to fold that into patch 5. Thanks, Mark.
On Thu, Feb 10, 2022 at 09:38:37AM +0000, Mark Rutland wrote: > On Wed, Feb 09, 2022 at 08:57:09PM +0100, Frederic Weisbecker wrote: > > On Wed, Feb 09, 2022 at 03:35:35PM +0000, Mark Rutland wrote: > > > Note that PREEMPT_DYNAMIC is `def bool y`, so this will default to > > > enabled. > > > > It should probably be "def_bool y if HAVE_STATIC_CALL_INLINE"... > > Sure; I'm more than happy to fold that into patch 5. For the moment I've made that: def_bool y if HAVE_PREEMPT_DYNAMIC_CALL ... since that fit more neatly with the other bits I had to add, and didn't change the existing behaviour of 32-bit x86. Please shout if you think that should be HAVE_STATIC_CALL_INLINE specifically! Thanks, Mark.
On Thu, Feb 10, 2022 at 12:00:56PM +0000, Mark Rutland wrote: > On Thu, Feb 10, 2022 at 09:38:37AM +0000, Mark Rutland wrote: > > On Wed, Feb 09, 2022 at 08:57:09PM +0100, Frederic Weisbecker wrote: > > > On Wed, Feb 09, 2022 at 03:35:35PM +0000, Mark Rutland wrote: > > > > Note that PREEMPT_DYNAMIC is `def bool y`, so this will default to > > > > enabled. > > > > > > It should probably be "def_bool y if HAVE_STATIC_CALL_INLINE"... > > > > Sure; I'm more than happy to fold that into patch 5. > > For the moment I've made that: > > def_bool y if HAVE_PREEMPT_DYNAMIC_CALL > > ... since that fit more neatly with the other bits I had to add, and didn't > change the existing behaviour of 32-bit x86. > > Please shout if you think that should be HAVE_STATIC_CALL_INLINE specifically! I seem to remember peterz didn't mind keeping it default y as long as HAVE_STATIC_CALL*. So I guess that's fine. Thanks!
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index f2b5a4abef21..3831d922a81d 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -192,6 +192,7 @@ config ARM64 select HAVE_PERF_EVENTS select HAVE_PERF_REGS select HAVE_PERF_USER_STACK_DUMP + select HAVE_PREEMPT_DYNAMIC_KEY select HAVE_REGS_AND_STACK_ACCESS_API select HAVE_POSIX_CPU_TIMERS_TASK_WORK select HAVE_FUNCTION_ARG_ACCESS_API diff --git a/arch/arm64/include/asm/preempt.h b/arch/arm64/include/asm/preempt.h index e83f0982b99c..0159b625cc7f 100644 --- a/arch/arm64/include/asm/preempt.h +++ b/arch/arm64/include/asm/preempt.h @@ -2,6 +2,7 @@ #ifndef __ASM_PREEMPT_H #define __ASM_PREEMPT_H +#include <linux/jump_label.h> #include <linux/thread_info.h> #define PREEMPT_NEED_RESCHED BIT(32) @@ -80,10 +81,24 @@ static inline bool should_resched(int preempt_offset) } #ifdef CONFIG_PREEMPTION + void preempt_schedule(void); -#define __preempt_schedule() preempt_schedule() void preempt_schedule_notrace(void); -#define __preempt_schedule_notrace() preempt_schedule_notrace() + +#ifdef CONFIG_PREEMPT_DYNAMIC + +DECLARE_STATIC_KEY_TRUE(sk_dynamic_irqentry_exit_cond_resched); +void dynamic_preempt_schedule(void); +#define __preempt_schedule() dynamic_preempt_schedule() +void dynamic_preempt_schedule_notrace(void); +#define __preempt_schedule_notrace() dynamic_preempt_schedule_notrace() + +#else /* CONFIG_PREEMPT_DYNAMIC */ + +#define __preempt_schedule() preempt_schedule() +#define __preempt_schedule_notrace() preempt_schedule_notrace() + +#endif /* CONFIG_PREEMPT_DYNAMIC */ #endif /* CONFIG_PREEMPTION */ #endif /* __ASM_PREEMPT_H */ diff --git a/arch/arm64/kernel/entry-common.c b/arch/arm64/kernel/entry-common.c index 2c639b6b676d..675352ec1368 100644 --- a/arch/arm64/kernel/entry-common.c +++ b/arch/arm64/kernel/entry-common.c @@ -220,9 +220,17 @@ static void noinstr arm64_exit_el1_dbg(struct pt_regs *regs) lockdep_hardirqs_on(CALLER_ADDR0); } +#ifdef CONFIG_PREEMPT_DYNAMIC +DEFINE_STATIC_KEY_TRUE(sk_dynamic_irqentry_exit_cond_resched); +#define need_irq_preemption() \ + (static_branch_unlikely(&sk_dynamic_irqentry_exit_cond_resched)) +#else +#define need_irq_preemption() (IS_ENABLED(CONFIG_PREEMPTION)) +#endif + static void __sched arm64_preempt_schedule_irq(void) { - if (!IS_ENABLED(CONFIG_PREEMPTION)) + if (!need_irq_preemption()) return; /*
This patch enables support for PREEMPT_DYNAMIC on arm64, allowing the preemption model to be chosen at boot time. Specifically, this patch selects HAVE_PREEMPT_DYNAMIC_KEY, so that each preemption function is an out-of-line call with an early return depending upon a static key. This leaves almost all the codegen up to the compiler, and side-steps a number of pain points with static calls (e.g. interaction with CFI schemes). This should have no worse overhead than using non-inline static calls, as those use out-of-line trampolines with early returns. For example, the dynamic_cond_resched() wrapper looks as follows when enabled. When disabled, the first `B` is replaced with a `NOP`, resulting in an early return. | <dynamic_cond_resched>: | bti c | b <dynamic_cond_resched+0x10> // or `nop` | mov w0, #0x0 | ret | mrs x0, sp_el0 | ldr x0, [x0, #8] | cbnz x0, <dynamic_cond_resched+0x8> | paciasp | stp x29, x30, [sp, #-16]! | mov x29, sp | bl <preempt_schedule_common> | mov w0, #0x1 | ldp x29, x30, [sp], #16 | autiasp | ret ... compared to the regular form of the function: | <__cond_resched>: | bti c | mrs x0, sp_el0 | ldr x1, [x0, #8] | cbz x1, <__cond_resched+0x18> | mov w0, #0x0 | ret | paciasp | stp x29, x30, [sp, #-16]! | mov x29, sp | bl <preempt_schedule_common> | mov w0, #0x1 | ldp x29, x30, [sp], #16 | autiasp | ret Since arm64 does not yet use the generic entry code, we must define our own `sk_dynamic_irqentry_exit_cond_resched`, which will be enabled/disabled by the common code in kernel/sched/core.c. All other preemption functions and associated static keys are defined there. Note that PREEMPT_DYNAMIC is `def bool y`, so this will default to enabled. Signed-off-by: Mark Rutland <mark.rutland@arm.com> Cc: Ard Biesheuvel <ardb@kernel.org> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Frederic Weisbecker <frederic@kernel.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Joey Gouly <joey.gouly@arm.com> Cc: Juri Lelli <juri.lelli@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Valentin Schneider <valentin.schneider@arm.com> Cc: Will Deacon <will@kernel.org> --- arch/arm64/Kconfig | 1 + arch/arm64/include/asm/preempt.h | 19 +++++++++++++++++-- arch/arm64/kernel/entry-common.c | 10 +++++++++- 3 files changed, 27 insertions(+), 3 deletions(-)