diff mbox series

[v3] arm64: run softirqs on the per-CPU IRQ stack

Message ID 20220815124739.15948-1-zhengqi.arch@bytedance.com (mailing list archive)
State New, archived
Headers show
Series [v3] arm64: run softirqs on the per-CPU IRQ stack | expand

Commit Message

Qi Zheng Aug. 15, 2022, 12:47 p.m. UTC
Currently arm64 supports per-CPU IRQ stack, but softirqs
are still handled in the task context.

Since any call to local_bh_enable() at any level in the task's
call stack may trigger a softirq processing run, which could
potentially cause a task stack overflow if the combined stack
footprints exceed the stack's size, let's run these softirqs
on the IRQ stack as well.

Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Will Deacon <will@kernel.org>
---
v2: https://lore.kernel.org/lkml/20220802065325.39740-1-zhengqi.arch@bytedance.com/
v1: https://lore.kernel.org/lkml/20220708094950.41944-1-zhengqi.arch@bytedance.com/
RFC: https://lore.kernel.org/lkml/20220707110511.52129-1-zhengqi.arch@bytedance.com/

Changelog in v2 -> v3:
 - rebase onto the v6.0-rc1

Changelog in v1 -> v2:
 - temporarily discard [PATCH v1 2/2] to allow this patch to be merged first
 - rebase onto the v5.19
 - collect Reviewed-by and Acked-by

Changelog in RFC -> v1:
 - fix conflicts with commit f2c5092190f2 ("arch/*: Disable softirq stacks on PREEMPT_RT.")

 arch/arm64/Kconfig      |  1 +
 arch/arm64/kernel/irq.c | 13 +++++++++++++
 2 files changed, 14 insertions(+)

Comments

Qi Zheng Aug. 26, 2022, 4:16 a.m. UTC | #1
On 2022/8/15 20:47, Qi Zheng wrote:
> Currently arm64 supports per-CPU IRQ stack, but softirqs
> are still handled in the task context.
> 
> Since any call to local_bh_enable() at any level in the task's
> call stack may trigger a softirq processing run, which could
> potentially cause a task stack overflow if the combined stack
> footprints exceed the stack's size, let's run these softirqs
> on the IRQ stack as well.
> 
> Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
> Acked-by: Will Deacon <will@kernel.org>
> ---
> v2: https://lore.kernel.org/lkml/20220802065325.39740-1-zhengqi.arch@bytedance.com/
> v1: https://lore.kernel.org/lkml/20220708094950.41944-1-zhengqi.arch@bytedance.com/
> RFC: https://lore.kernel.org/lkml/20220707110511.52129-1-zhengqi.arch@bytedance.com/
> 
> Changelog in v2 -> v3:
>   - rebase onto the v6.0-rc1

Gentle ping.

Thanks,
Qi

> 
> Changelog in v1 -> v2:
>   - temporarily discard [PATCH v1 2/2] to allow this patch to be merged first
>   - rebase onto the v5.19
>   - collect Reviewed-by and Acked-by
> 
> Changelog in RFC -> v1:
>   - fix conflicts with commit f2c5092190f2 ("arch/*: Disable softirq stacks on PREEMPT_RT.")
> 
>   arch/arm64/Kconfig      |  1 +
>   arch/arm64/kernel/irq.c | 13 +++++++++++++
>   2 files changed, 14 insertions(+)
> 
> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> index 571cc234d0b3..ee92f5887cf6 100644
> --- a/arch/arm64/Kconfig
> +++ b/arch/arm64/Kconfig
> @@ -230,6 +230,7 @@ config ARM64
>   	select HAVE_ARCH_USERFAULTFD_MINOR if USERFAULTFD
>   	select TRACE_IRQFLAGS_SUPPORT
>   	select TRACE_IRQFLAGS_NMI_SUPPORT
> +	select HAVE_SOFTIRQ_ON_OWN_STACK
>   	help
>   	  ARM 64-bit (AArch64) Linux support.
>   
> diff --git a/arch/arm64/kernel/irq.c b/arch/arm64/kernel/irq.c
> index bda49430c9ea..c36ad20a52f3 100644
> --- a/arch/arm64/kernel/irq.c
> +++ b/arch/arm64/kernel/irq.c
> @@ -22,6 +22,7 @@
>   #include <linux/vmalloc.h>
>   #include <asm/daifflags.h>
>   #include <asm/vmap_stack.h>
> +#include <asm/exception.h>
>   
>   /* Only access this in an NMI enter/exit */
>   DEFINE_PER_CPU(struct nmi_ctx, nmi_contexts);
> @@ -71,6 +72,18 @@ static void init_irq_stacks(void)
>   }
>   #endif
>   
> +#ifndef CONFIG_PREEMPT_RT
> +static void ____do_softirq(struct pt_regs *regs)
> +{
> +	__do_softirq();
> +}
> +
> +void do_softirq_own_stack(void)
> +{
> +	call_on_irq_stack(NULL, ____do_softirq);
> +}
> +#endif
> +
>   static void default_handle_irq(struct pt_regs *regs)
>   {
>   	panic("IRQ taken without a root IRQ handler\n");
Qi Zheng Sept. 7, 2022, 7:04 a.m. UTC | #2
On 2022/8/26 12:16, Qi Zheng wrote:
> 
> 
> On 2022/8/15 20:47, Qi Zheng wrote:
>> Currently arm64 supports per-CPU IRQ stack, but softirqs
>> are still handled in the task context.
>>
>> Since any call to local_bh_enable() at any level in the task's
>> call stack may trigger a softirq processing run, which could
>> potentially cause a task stack overflow if the combined stack
>> footprints exceed the stack's size, let's run these softirqs
>> on the IRQ stack as well.
>>
>> Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
>> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>> Acked-by: Will Deacon <will@kernel.org>
>> ---
>> v2: 
>> https://lore.kernel.org/lkml/20220802065325.39740-1-zhengqi.arch@bytedance.com/
>> v1: 
>> https://lore.kernel.org/lkml/20220708094950.41944-1-zhengqi.arch@bytedance.com/
>> RFC: 
>> https://lore.kernel.org/lkml/20220707110511.52129-1-zhengqi.arch@bytedance.com/
>>
>> Changelog in v2 -> v3:
>>   - rebase onto the v6.0-rc1

Hi Will,

Are we good to merge this patch? Or if there is anything else I need to
do, please let me know. :)

Looking forward to your reply.

Thanks,
Qi

> 
> Gentle ping.
> 
> Thanks,
> Qi
> 
>>
>> Changelog in v1 -> v2:
>>   - temporarily discard [PATCH v1 2/2] to allow this patch to be 
>> merged first
>>   - rebase onto the v5.19
>>   - collect Reviewed-by and Acked-by
>>
>> Changelog in RFC -> v1:
>>   - fix conflicts with commit f2c5092190f2 ("arch/*: Disable softirq 
>> stacks on PREEMPT_RT.")
>>
>>   arch/arm64/Kconfig      |  1 +
>>   arch/arm64/kernel/irq.c | 13 +++++++++++++
>>   2 files changed, 14 insertions(+)
>>
>> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
>> index 571cc234d0b3..ee92f5887cf6 100644
>> --- a/arch/arm64/Kconfig
>> +++ b/arch/arm64/Kconfig
>> @@ -230,6 +230,7 @@ config ARM64
>>       select HAVE_ARCH_USERFAULTFD_MINOR if USERFAULTFD
>>       select TRACE_IRQFLAGS_SUPPORT
>>       select TRACE_IRQFLAGS_NMI_SUPPORT
>> +    select HAVE_SOFTIRQ_ON_OWN_STACK
>>       help
>>         ARM 64-bit (AArch64) Linux support.
>> diff --git a/arch/arm64/kernel/irq.c b/arch/arm64/kernel/irq.c
>> index bda49430c9ea..c36ad20a52f3 100644
>> --- a/arch/arm64/kernel/irq.c
>> +++ b/arch/arm64/kernel/irq.c
>> @@ -22,6 +22,7 @@
>>   #include <linux/vmalloc.h>
>>   #include <asm/daifflags.h>
>>   #include <asm/vmap_stack.h>
>> +#include <asm/exception.h>
>>   /* Only access this in an NMI enter/exit */
>>   DEFINE_PER_CPU(struct nmi_ctx, nmi_contexts);
>> @@ -71,6 +72,18 @@ static void init_irq_stacks(void)
>>   }
>>   #endif
>> +#ifndef CONFIG_PREEMPT_RT
>> +static void ____do_softirq(struct pt_regs *regs)
>> +{
>> +    __do_softirq();
>> +}
>> +
>> +void do_softirq_own_stack(void)
>> +{
>> +    call_on_irq_stack(NULL, ____do_softirq);
>> +}
>> +#endif
>> +
>>   static void default_handle_irq(struct pt_regs *regs)
>>   {
>>       panic("IRQ taken without a root IRQ handler\n");
>
Will Deacon Sept. 7, 2022, 1:34 p.m. UTC | #3
On Wed, Sep 07, 2022 at 03:04:48PM +0800, Qi Zheng wrote:
> 
> 
> On 2022/8/26 12:16, Qi Zheng wrote:
> > 
> > 
> > On 2022/8/15 20:47, Qi Zheng wrote:
> > > Currently arm64 supports per-CPU IRQ stack, but softirqs
> > > are still handled in the task context.
> > > 
> > > Since any call to local_bh_enable() at any level in the task's
> > > call stack may trigger a softirq processing run, which could
> > > potentially cause a task stack overflow if the combined stack
> > > footprints exceed the stack's size, let's run these softirqs
> > > on the IRQ stack as well.
> > > 
> > > Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
> > > Reviewed-by: Arnd Bergmann <arnd@arndb.de>
> > > Acked-by: Will Deacon <will@kernel.org>
> > > ---
> > > v2: https://lore.kernel.org/lkml/20220802065325.39740-1-zhengqi.arch@bytedance.com/
> > > v1: https://lore.kernel.org/lkml/20220708094950.41944-1-zhengqi.arch@bytedance.com/
> > > RFC: https://lore.kernel.org/lkml/20220707110511.52129-1-zhengqi.arch@bytedance.com/
> > > 
> > > Changelog in v2 -> v3:
> > >   - rebase onto the v6.0-rc1
> 
> Hi Will,
> 
> Are we good to merge this patch? Or if there is anything else I need to
> do, please let me know. :)

I'm expecting Catalin to pick this one up for 6.1.

Will
Qi Zheng Sept. 7, 2022, 2:50 p.m. UTC | #4
On 2022/9/7 21:34, Will Deacon wrote:
> On Wed, Sep 07, 2022 at 03:04:48PM +0800, Qi Zheng wrote:
>>
>>
>> On 2022/8/26 12:16, Qi Zheng wrote:
>>>
>>>
>>> On 2022/8/15 20:47, Qi Zheng wrote:
>>>> Currently arm64 supports per-CPU IRQ stack, but softirqs
>>>> are still handled in the task context.
>>>>
>>>> Since any call to local_bh_enable() at any level in the task's
>>>> call stack may trigger a softirq processing run, which could
>>>> potentially cause a task stack overflow if the combined stack
>>>> footprints exceed the stack's size, let's run these softirqs
>>>> on the IRQ stack as well.
>>>>
>>>> Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
>>>> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>>>> Acked-by: Will Deacon <will@kernel.org>
>>>> ---
>>>> v2: https://lore.kernel.org/lkml/20220802065325.39740-1-zhengqi.arch@bytedance.com/
>>>> v1: https://lore.kernel.org/lkml/20220708094950.41944-1-zhengqi.arch@bytedance.com/
>>>> RFC: https://lore.kernel.org/lkml/20220707110511.52129-1-zhengqi.arch@bytedance.com/
>>>>
>>>> Changelog in v2 -> v3:
>>>>    - rebase onto the v6.0-rc1
>>
>> Hi Will,
>>
>> Are we good to merge this patch? Or if there is anything else I need to
>> do, please let me know. :)
> 
> I'm expecting Catalin to pick this one up for 6.1.

Oh, I see. Looking forward to this.

Thanks a lot.
Qi

> 
> Will
Catalin Marinas Sept. 8, 2022, 6:34 p.m. UTC | #5
On Mon, 15 Aug 2022 20:47:39 +0800, Qi Zheng wrote:
> Currently arm64 supports per-CPU IRQ stack, but softirqs
> are still handled in the task context.
> 
> Since any call to local_bh_enable() at any level in the task's
> call stack may trigger a softirq processing run, which could
> potentially cause a task stack overflow if the combined stack
> footprints exceed the stack's size, let's run these softirqs
> on the IRQ stack as well.
> 
> [...]

Applied to arm64 (for-next/misc), thanks!

[1/1] arm64: run softirqs on the per-CPU IRQ stack
      https://git.kernel.org/arm64/c/2d2f3bb897a3
diff mbox series

Patch

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 571cc234d0b3..ee92f5887cf6 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -230,6 +230,7 @@  config ARM64
 	select HAVE_ARCH_USERFAULTFD_MINOR if USERFAULTFD
 	select TRACE_IRQFLAGS_SUPPORT
 	select TRACE_IRQFLAGS_NMI_SUPPORT
+	select HAVE_SOFTIRQ_ON_OWN_STACK
 	help
 	  ARM 64-bit (AArch64) Linux support.
 
diff --git a/arch/arm64/kernel/irq.c b/arch/arm64/kernel/irq.c
index bda49430c9ea..c36ad20a52f3 100644
--- a/arch/arm64/kernel/irq.c
+++ b/arch/arm64/kernel/irq.c
@@ -22,6 +22,7 @@ 
 #include <linux/vmalloc.h>
 #include <asm/daifflags.h>
 #include <asm/vmap_stack.h>
+#include <asm/exception.h>
 
 /* Only access this in an NMI enter/exit */
 DEFINE_PER_CPU(struct nmi_ctx, nmi_contexts);
@@ -71,6 +72,18 @@  static void init_irq_stacks(void)
 }
 #endif
 
+#ifndef CONFIG_PREEMPT_RT
+static void ____do_softirq(struct pt_regs *regs)
+{
+	__do_softirq();
+}
+
+void do_softirq_own_stack(void)
+{
+	call_on_irq_stack(NULL, ____do_softirq);
+}
+#endif
+
 static void default_handle_irq(struct pt_regs *regs)
 {
 	panic("IRQ taken without a root IRQ handler\n");