[PATCH -next V10 09/10] riscv: stack: Support HAVE_SOFTIRQ_ON_OWN_STACK

[Guo Ren]

On Fri, Dec 9, 2022 at 3:50 PM Björn Töpel <bjorn at kernel.org> wrote:
>
> Guo Ren <guoren at kernel.org> writes:
>
> > On Thu, Dec 8, 2022 at 6:12 PM Björn Töpel <bjorn at kernel.org> wrote:
> >>
> >> guoren at kernel.org writes:
> >>
> >> > From: Guo Ren <guoren at linux.alibaba.com>
> >> >
> >> > Add the HAVE_SOFTIRQ_ON_OWN_STACK feature for the IRQ_STACKS config. The
> >> > irq and softirq use the same independent irq_stack of percpu by time
> >> > division multiplexing.
> >> >
> >> > Tested-by: Jisheng Zhang <jszhang at kernel.org>
> >> > Signed-off-by: Guo Ren <guoren at linux.alibaba.com>
> >> > Signed-off-by: Guo Ren <guoren at kernel.org>
> >> > ---
> >> >  arch/riscv/Kconfig      |  7 ++++---
> >> >  arch/riscv/kernel/irq.c | 33 +++++++++++++++++++++++++++++++++
> >> >  2 files changed, 37 insertions(+), 3 deletions(-)
> >> >
> >> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> >> > index 0a9d4bdc0338..bd4c4ae4cdc9 100644
> >> > --- a/arch/riscv/Kconfig
> >> > +++ b/arch/riscv/Kconfig
> >> > @@ -447,12 +447,13 @@ config FPU
> >> >         If you don't know what to do here, say Y.
> >> >
> >> >  config IRQ_STACKS
> >> > -     bool "Independent irq stacks" if EXPERT
> >> > +     bool "Independent irq & softirq stacks" if EXPERT
> >> >       default y
> >> >       select HAVE_IRQ_EXIT_ON_IRQ_STACK
> >> > +     select HAVE_SOFTIRQ_ON_OWN_STACK
> >>
> >> HAVE_IRQ_EXIT_ON_IRQ_STACK is used by softirq.c Shouldn't that be
> >> selected introduced in this patch, instead of the previous one?
> > This patch depends on the previous one. And the previous one could
> > work separately.
>
> Let me try to be more clear: IRQ_STACKS should be introduced in the
> previous patch, which adds per-cpu stacks to hardirq. This patch adds
> per-cpu stacks for softirq, and the softirq related selects:
>
>  select HAVE_IRQ_EXIT_ON_IRQ_STACK
>  select HAVE_SOFTIRQ_ON_OWN_STACK
>
> Hence, the "HAVE_IRQ_EXIT_ON_IRQ_STACK" select should be part of *this*
> patch, not the previous.
>
> Or am I missing something?
You are right, HAVE_IRQ_EXIT_ON_IRQ_STACK is belong to SOFTIRQ:
static inline void invoke_softirq(void)
{
...
        if (!force_irqthreads() || !__this_cpu_read(ksoftirqd)) {
#ifdef CONFIG_HAVE_IRQ_EXIT_ON_IRQ_STACK
...
                __do_softirq();
#else
...
                do_softirq_own_stack();
#endif
...
}

I would fix that in the next version.

>
>
> Björn



-- 
Best Regards
 Guo Ren