[PATCH v2] arm64: run softirqs on the per-CPU IRQ stack

[PATCH v2] arm64: run softirqs on the per-CPU IRQ stack

[Qi Zheng]

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>
---
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 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 1652a9800ebe..90f1ab403724 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -226,6 +226,7 @@ config ARM64
 	select THREAD_INFO_IN_TASK
 	select HAVE_ARCH_USERFAULTFD_MINOR if USERFAULTFD
 	select TRACE_IRQFLAGS_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");
-- 
2.20.1

Re: [PATCH v2] arm64: run softirqs on the per-CPU IRQ stack

[Qi Zheng]

On 2022/8/2 14:53, 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>

Hi Will,

Are we good to merge it into 6.0-rc1?

Thanks,
Qi

> ---
> 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 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 1652a9800ebe..90f1ab403724 100644
> --- a/arch/arm64/Kconfig
> +++ b/arch/arm64/Kconfig
> @@ -226,6 +226,7 @@ config ARM64
>   	select THREAD_INFO_IN_TASK
>   	select HAVE_ARCH_USERFAULTFD_MINOR if USERFAULTFD
>   	select TRACE_IRQFLAGS_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");

-- 
Thanks,
Qi

Re: [PATCH v2] arm64: run softirqs on the per-CPU IRQ stack

[Arnd Bergmann]

On Fri, Aug 5, 2022 at 9:10 AM Qi Zheng <zhengqi.arch@bytedance.com> wrote:
> On 2022/8/2 14:53, 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>
>
> Are we good to merge it into 6.0-rc1?

I think you misunderstood the timing that Will proposed in

https://lore.kernel.org/linux-arm-kernel/d540aaff-ec6a-3f25-dd79-b27b4ad81b36@bytedance.com/

You should send your patch after 6.0-rc1 has been released, to be merged
into the linux-next tree and sent as part of the 6.1 merge window.

The two ways patches get merged are:

- bugfixes can get merged into maintainer trees at any time and sent
  upstream regardless of the merge window, these never need to wait

- non-bugfix patches can get merged into maintainer trees based on an
   -rc release and then sent upstream during the following merge window.

        Arnd

Re: [PATCH v2] arm64: run softirqs on the per-CPU IRQ stack

[Qi Zheng]

On 2022/8/5 15:50, Arnd Bergmann wrote:
> On Fri, Aug 5, 2022 at 9:10 AM Qi Zheng <zhengqi.arch@bytedance.com> wrote:
>> On 2022/8/2 14:53, 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>
>>
>> Are we good to merge it into 6.0-rc1?
> 
> I think you misunderstood the timing that Will proposed in
> 
> https://lore.kernel.org/linux-arm-kernel/d540aaff-ec6a-3f25-dd79-b27b4ad81b36@bytedance.com/
> 
> You should send your patch after 6.0-rc1 has been released, to be merged
> into the linux-next tree and sent as part of the 6.1 merge window.

Oh, got it. I will repost it after 6.0-rc1.

> 
> The two ways patches get merged are:
> 
> - bugfixes can get merged into maintainer trees at any time and sent
>    upstream regardless of the merge window, these never need to wait
> 
> - non-bugfix patches can get merged into maintainer trees based on an
>     -rc release and then sent upstream during the following merge window.

Thanks for such detailed information. :)

> 
>          Arnd

Thanks,
Qi

-- 
Thanks,
Qi

Re: [PATCH v2] arm64: run softirqs on the per-CPU IRQ stack

[kernel test robot]

Hi Qi,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on soc/for-next]
[also build test WARNING on v5.19]
[cannot apply to arm64/for-next/core linus/master next-20220805]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Qi-Zheng/arm64-run-softirqs-on-the-per-CPU-IRQ-stack/20220802-145547
base:   https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: arm64-randconfig-r032-20220801 (https://download.01.org/0day-ci/archive/20220807/202208072125.pTMPctXF-lkp@intel.com/config)
compiler: clang version 16.0.0 (https://github.com/llvm/llvm-project 52cd00cabf479aa7eb6dbb063b7ba41ea57bce9e)
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # install arm64 cross compiling tool for clang build
        # apt-get install binutils-aarch64-linux-gnu
        # https://github.com/intel-lab-lkp/linux/commit/34ac9375ce5e75127084159f724fcb2208c022a3
        git remote add linux-review https://github.com/intel-lab-lkp/linux
        git fetch --no-tags linux-review Qi-Zheng/arm64-run-softirqs-on-the-per-CPU-IRQ-stack/20220802-145547
        git checkout 34ac9375ce5e75127084159f724fcb2208c022a3
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=arm64 SHELL=/bin/bash arch/arm64/kernel/

If you fix the issue, kindly add following tag where applicable
Reported-by: kernel test robot <lkp@intel.com>

All warnings (new ones prefixed by >>):

>> arch/arm64/kernel/irq.c:81:6: warning: no previous prototype for function 'do_softirq_own_stack' [-Wmissing-prototypes]
   void do_softirq_own_stack(void)
        ^
   arch/arm64/kernel/irq.c:81:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
   void do_softirq_own_stack(void)
   ^
   static 
   arch/arm64/kernel/irq.c:120:13: warning: no previous prototype for function 'init_IRQ' [-Wmissing-prototypes]
   void __init init_IRQ(void)
               ^
   arch/arm64/kernel/irq.c:120:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
   void __init init_IRQ(void)
   ^
   static 
   2 warnings generated.


vim +/do_softirq_own_stack +81 arch/arm64/kernel/irq.c

    80	
  > 81	void do_softirq_own_stack(void)
    82	{
    83		call_on_irq_stack(NULL, ____do_softirq);
    84	}
    85	#endif
    86	

-- 
0-DAY CI Kernel Test Service
https://01.org/lkp

Re: [PATCH v2] arm64: run softirqs on the per-CPU IRQ stack

[kernel test robot]

Hi Qi,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on soc/for-next]
[also build test WARNING on v5.19]
[cannot apply to arm64/for-next/core linus/master next-20220805]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Qi-Zheng/arm64-run-softirqs-on-the-per-CPU-IRQ-stack/20220802-145547
base:   https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: arm64-randconfig-r024-20220801 (https://download.01.org/0day-ci/archive/20220807/202208072123.sWU13g72-lkp@intel.com/config)
compiler: aarch64-linux-gcc (GCC) 12.1.0
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # https://github.com/intel-lab-lkp/linux/commit/34ac9375ce5e75127084159f724fcb2208c022a3
        git remote add linux-review https://github.com/intel-lab-lkp/linux
        git fetch --no-tags linux-review Qi-Zheng/arm64-run-softirqs-on-the-per-CPU-IRQ-stack/20220802-145547
        git checkout 34ac9375ce5e75127084159f724fcb2208c022a3
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=arm64 SHELL=/bin/bash arch/arm64/kernel/

If you fix the issue, kindly add following tag where applicable
Reported-by: kernel test robot <lkp@intel.com>

All warnings (new ones prefixed by >>):

>> arch/arm64/kernel/irq.c:81:6: warning: no previous prototype for 'do_softirq_own_stack' [-Wmissing-prototypes]
      81 | void do_softirq_own_stack(void)
         |      ^~~~~~~~~~~~~~~~~~~~
   arch/arm64/kernel/irq.c:120:13: warning: no previous prototype for 'init_IRQ' [-Wmissing-prototypes]
     120 | void __init init_IRQ(void)
         |             ^~~~~~~~


vim +/do_softirq_own_stack +81 arch/arm64/kernel/irq.c

    80	
  > 81	void do_softirq_own_stack(void)
    82	{
    83		call_on_irq_stack(NULL, ____do_softirq);
    84	}
    85	#endif
    86	

-- 
0-DAY CI Kernel Test Service
https://01.org/lkp