[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YXJ7YzZoJ1ZkLl4X@FVFF77S0Q05N>
Date: Fri, 22 Oct 2021 09:52:02 +0100
From: Mark Rutland <mark.rutland@....com>
To: Guo Ren <guoren@...nel.org>
Cc: Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Albert Ou <aou@...s.berkeley.edu>,
Catalin Marinas <catalin.marinas@....com>,
Vincent Chen <deanbo422@...il.com>,
Greentime Hu <green.hu@...il.com>,
Jonas Bonn <jonas@...thpole.se>, kernelfans@...il.com,
Linux ARM <linux-arm-kernel@...ts.infradead.org>,
Russell King <linux@...linux.org.uk>,
Marc Zyngier <maz@...nel.org>, Nick Hu <nickhu@...estech.com>,
Palmer Dabbelt <palmer@...belt.com>,
Paul McKenney <paulmck@...nel.org>,
Paul Walmsley <paul.walmsley@...ive.com>,
Peter Zijlstra <peterz@...radead.org>,
Stafford Horne <shorne@...il.com>,
Stefan Kristiansson <stefan.kristiansson@...nalahti.fi>,
Thomas Gleixner <tglx@...utronix.de>,
Linus Torvalds <torvalds@...ux-foundation.org>,
Thomas Bogendoerfer <tsbogend@...ha.franken.de>,
vgupta@...nel.org, Will Deacon <will@...nel.org>
Subject: Re: [PATCH 05/15] irq: add generic_handle_arch_irq()
On Fri, Oct 22, 2021 at 10:33:53AM +0800, Guo Ren wrote:
> On Fri, Oct 22, 2021 at 2:03 AM Mark Rutland <mark.rutland@....com> wrote:
> >
> > Several architectures select GENERIC_IRQ_MULTI_HANDLER and branch to
> > handle_arch_irq() without performing any entry accounting.
> >
> > Add a generic wrapper to handle the commoon irqentry work when invoking
Ah, I'd typo'd 'common' there; I'll go fix that...
> > handle_arch_irq(). Where an architecture needs to perform some entry
> > accounting itself, it will need to invoke handle_arch_irq() itself.
> >
> > In subsequent patches it will become the responsibilty of the entry code
> > to set the irq regs when entering an IRQ (rather than deferring this to
> > an irqchip handler), so generic_handle_arch_irq() is made to set the irq
> > regs now. This can be redundant in some cases, but is never harmful as
> > saving/restoring the old regs nests safely.
> Shall we remove old_regs save/restore in handle_domain_irq? It's duplicated.
We do, in patch 15 when handle_domain_irq() is removed entirely. :)
Removing it immediately in this patch would require more ifdeffery and/or
another copy of handle_domain_irq(), and since the duplication doesn't cause a
functional problem, I think it's better to live with the temporary duplication
for the next few patches than to try to optimize the intermediate steps at the
cost of legibility.
Thanks,
Mark.
> > Signed-off-by: Mark Rutland <mark.rutland@....com>
> > Cc: Marc Zyngier <maz@...nel.org>
> > Cc: Thomas Gleixner <tglx@...utronix.de>
> > ---
> > kernel/irq/handle.c | 18 ++++++++++++++++++
> > 1 file changed, 18 insertions(+)
> >
> > diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c
> > index 221d80c31e94..27182003b879 100644
> > --- a/kernel/irq/handle.c
> > +++ b/kernel/irq/handle.c
> > @@ -14,6 +14,8 @@
> > #include <linux/interrupt.h>
> > #include <linux/kernel_stat.h>
> >
> > +#include <asm/irq_regs.h>
> > +
> > #include <trace/events/irq.h>
> >
> > #include "internals.h"
> > @@ -226,4 +228,20 @@ int __init set_handle_irq(void (*handle_irq)(struct pt_regs *))
> > handle_arch_irq = handle_irq;
> > return 0;
> > }
> > +
> > +/**
> > + * generic_handle_arch_irq - root irq handler for architectures which do no
> > + * entry accounting themselves
> > + * @regs: Register file coming from the low-level handling code
> > + */
> > +asmlinkage void noinstr generic_handle_arch_irq(struct pt_regs *regs)
> > +{
> > + struct pt_regs *old_regs;
> > +
> > + irq_enter();
> > + old_regs = set_irq_regs(regs);
> > + handle_arch_irq(regs);
> > + set_irq_regs(old_regs);
> > + irq_exit();
> > +}
> > #endif
> > --
> > 2.11.0
> >
>
>
> --
> Best Regards
> Guo Ren
>
> ML: https://lore.kernel.org/linux-csky/
Powered by blists - more mailing lists