linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Huacai Chen <chenhuacai@gmail.com>
To: Marc Zyngier <maz@kernel.org>
Cc: Huacai Chen <chenhuacai@loongson.cn>,
	Thomas Gleixner <tglx@linutronix.de>,
	LKML <linux-kernel@vger.kernel.org>,
	Xuefeng Li <lixuefeng@loongson.cn>,
	Jiaxun Yang <jiaxun.yang@flygoat.com>
Subject: Re: [PATCH V3 08/10] irqchip: Add LoongArch CPU interrupt controller support
Date: Sun, 29 Aug 2021 17:37:48 +0800	[thread overview]
Message-ID: <CAAhV-H41rridOo_3Eq5t9LPz-mefketAKNdhgbguFtZ0Cqz5Ng@mail.gmail.com> (raw)
In-Reply-To: <87v93pddzu.wl-maz@kernel.org>

Hi, Marc,

On Sat, Aug 28, 2021 at 7:07 PM Marc Zyngier <maz@kernel.org> wrote:
>
> Huacai,
>
> On Sat, 28 Aug 2021 11:07:16 +0100,
> Huacai Chen <chenhuacai@gmail.com> wrote:
> >
> > Hi, Marc,
> >
> > On Wed, Aug 25, 2021 at 4:40 PM Marc Zyngier <maz@kernel.org> wrote:
> > >
> > > On Wed, 25 Aug 2021 07:11:50 +0100,
> > > Huacai Chen <chenhuacai@loongson.cn> wrote:
> > > >
> > > > We are preparing to add new Loongson (based on LoongArch, not MIPS)
> > >
> > > You keep saying "not MIPS", and yet all I see is a blind copy of the
> > > MIPS code.
> > >
> > > > support. This patch add LoongArch CPU interrupt controller support.
> > > >
> > > > Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
> > > > ---
> > > >  drivers/irqchip/Kconfig             | 10 ++++
> > > >  drivers/irqchip/Makefile            |  1 +
> > > >  drivers/irqchip/irq-loongarch-cpu.c | 76 +++++++++++++++++++++++++++++
> > > >  3 files changed, 87 insertions(+)
> > > >  create mode 100644 drivers/irqchip/irq-loongarch-cpu.c
> > > >
> > > > diff --git a/drivers/irqchip/Kconfig b/drivers/irqchip/Kconfig
> > > > index 084bc4c2eebd..443c3a7a0cc1 100644
> > > > --- a/drivers/irqchip/Kconfig
> > > > +++ b/drivers/irqchip/Kconfig
> > > > @@ -528,6 +528,16 @@ config EXYNOS_IRQ_COMBINER
> > > >         Say yes here to add support for the IRQ combiner devices embedded
> > > >         in Samsung Exynos chips.
> > > >
> > > > +config IRQ_LOONGARCH_CPU
> > > > +     bool
> > > > +     select GENERIC_IRQ_CHIP
> > > > +     select IRQ_DOMAIN
> > > > +     select GENERIC_IRQ_EFFECTIVE_AFF_MASK
> > > > +     help
> > > > +       Support for the LoongArch CPU Interrupt Controller. For details of
> > > > +       irq chip hierarchy on LoongArch platforms please read the document
> > > > +       Documentation/loongarch/irq-chip-model.rst.
> > > > +
> > > >  config LOONGSON_LIOINTC
> > > >       bool "Loongson Local I/O Interrupt Controller"
> > > >       depends on MACH_LOONGSON64
> > > > diff --git a/drivers/irqchip/Makefile b/drivers/irqchip/Makefile
> > > > index f88cbf36a9d2..4e34eebe180b 100644
> > > > --- a/drivers/irqchip/Makefile
> > > > +++ b/drivers/irqchip/Makefile
> > > > @@ -105,6 +105,7 @@ obj-$(CONFIG_LS1X_IRQ)                    += irq-ls1x.o
> > > >  obj-$(CONFIG_TI_SCI_INTR_IRQCHIP)    += irq-ti-sci-intr.o
> > > >  obj-$(CONFIG_TI_SCI_INTA_IRQCHIP)    += irq-ti-sci-inta.o
> > > >  obj-$(CONFIG_TI_PRUSS_INTC)          += irq-pruss-intc.o
> > > > +obj-$(CONFIG_IRQ_LOONGARCH_CPU)              += irq-loongarch-cpu.o
> > > >  obj-$(CONFIG_LOONGSON_LIOINTC)               += irq-loongson-liointc.o
> > > >  obj-$(CONFIG_LOONGSON_HTPIC)         += irq-loongson-htpic.o
> > > >  obj-$(CONFIG_LOONGSON_HTVEC)         += irq-loongson-htvec.o
> > > > diff --git a/drivers/irqchip/irq-loongarch-cpu.c b/drivers/irqchip/irq-loongarch-cpu.c
> > > > new file mode 100644
> > > > index 000000000000..8e9e8d39cb22
> > > > --- /dev/null
> > > > +++ b/drivers/irqchip/irq-loongarch-cpu.c
> > > > @@ -0,0 +1,76 @@
> > > > +// SPDX-License-Identifier: GPL-2.0
> > > > +/*
> > > > + * Copyright (C) 2020-2021 Loongson Technology Corporation Limited
> > > > + */
> > > > +
> > > > +#include <linux/init.h>
> > > > +#include <linux/kernel.h>
> > > > +#include <linux/interrupt.h>
> > > > +#include <linux/irq.h>
> > > > +#include <linux/irqchip.h>
> > > > +#include <linux/irqdomain.h>
> > > > +
> > > > +#include <asm/loongarch.h>
> > > > +#include <asm/setup.h>
> > > > +
> > > > +static struct irq_domain *irq_domain;
> > > > +
> > > > +static inline void enable_loongarch_irq(struct irq_data *d)
> > >
> > > Why 'inline' given that it is used as a function pointer?
> > >
> > > > +{
> > > > +     set_csr_ecfg(ECFGF(d->hwirq));
> > > > +}
> > > > +
> > > > +#define eoi_loongarch_irq enable_loongarch_irq
> > >
> > > NAK. EOI and enable cannot be the same operation.
> > >
> > > > +
> > > > +static inline void disable_loongarch_irq(struct irq_data *d)
> > > > +{
> > > > +     clear_csr_ecfg(ECFGF(d->hwirq));
> > > > +}
> > > > +
> > > > +#define ack_loongarch_irq disable_loongarch_irq
> > >
> > > Same thing. Either you have different operations, or this only
> > > supports mask/unmask.
> > >
> > > > +
> > > > +static struct irq_chip loongarch_cpu_irq_controller = {
> > > > +     .name           = "LoongArch",
> > > > +     .irq_ack        = ack_loongarch_irq,
> > > > +     .irq_eoi        = eoi_loongarch_irq,
> > > > +     .irq_enable     = enable_loongarch_irq,
> > > > +     .irq_disable    = disable_loongarch_irq,
> > > > +};
> > > > +
> > > > +asmlinkage void default_handle_irq(int irq)
> > > > +{
> > > > +     do_IRQ(irq_linear_revmap(irq_domain, irq));
> > >
> > > This looks both wrong and short sighted:
> > >
> > > - irq_linear_revmap() is now another name for irq_find_mapping().
> > >   Which means it uses a RCU read critical section. If, as I expect,
> > >   this is just a blind copy of the MIPS code, do_IRQ() will not do
> > >   anything with respect to irq_enter()/irq_exit(), which will result
> > >   in something pretty bad on the exit from idle path. Lockdep will
> > >   probably shout at you pretty loudly.
> > >
> > > - A single root interrupt controller is, in my modest experience,
> > >   something that rarely happen. You will eventually have a variety of
> > >   them, and you will have to join the other arches such as arm, arm64,
> > >   riscv and csky that use CONFIG_GENERIC_IRQ_MULTI_HANDLER instead of
> > >   following the existing MIPS model.
> > I try to use CONFIG_GENERIC_IRQ_MULTI_HANDLER and
> > set_handle_irq()/handle_arch_irq() as arm64, riscv and csky do. But I
> > found a problem:
> > The main handler (e.g., handle_arch_irq()) take only one argument
> > (i.e., struct pt_regs *regs) and polling all interrupts, but we want
> > to use vectored interrupts which take a "irq" argument (as
> > default_handle_irq() does) which can directly handle it.
>
> Are you saying that there is no way for the interrupt controller
> driver to figure out the hwirq number on its own?  That would seem
> pretty odd (even the MIPS GIC has that). Worse case, you can provide
> an arch-specific helper that exposes the current hwirq based on the
> vector that triggered.
We can get the hwirq number by reading CSR.ESTAT register, but in this
way "vectored interrupts" is meaningless.

Huacai
>
> do_IRQ() is a terrible abstraction, and only outlines that your arch
> code is badly structured. What does the arch code have to do with a
> Linux irq number? It shouldn't care at all, because as a value it has
> no significance to the arch code at all. You just go back there
> because the management of your interrupt context is upside down, and
> it really shouldn't matter *what interrupt fired*.
>
> > This seems that if I want to use vectored interrupts, then I will fall
> > to the MIPS model.
>
> Not happening, I'm afraid.
>
>         M.
>
> --
> Without deviation from the norm, progress is not possible.

  reply	other threads:[~2021-08-29  9:38 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-25  6:11 [PATCH V3 00/10] irqchip: Add LoongArch-related irqchip drivers Huacai Chen
2021-08-25  6:11 ` [PATCH V3 01/10] irqchip: Adjust Kconfig for Loongson Huacai Chen
2021-08-25  6:11 ` [PATCH V3 02/10] irqchip/loongson-pch-pic: Add ACPI init support Huacai Chen
2021-08-25  6:11 ` [PATCH V3 03/10] irqchip/loongson-pch-pic: Add suspend/resume support Huacai Chen
2021-08-25  6:11 ` [PATCH V3 04/10] irqchip/loongson-pch-msi: Add ACPI init support Huacai Chen
2021-08-25  6:11 ` [PATCH V3 05/10] irqchip/loongson-htvec: " Huacai Chen
2021-08-25  6:11 ` [PATCH V3 06/10] irqchip/loongson-htvec: Add suspend/resume support Huacai Chen
2021-08-25  6:11 ` [PATCH V3 07/10] irqchip/loongson-liointc: Add ACPI init support Huacai Chen
2021-08-25  6:11 ` [PATCH V3 08/10] irqchip: Add LoongArch CPU interrupt controller support Huacai Chen
2021-08-25  8:40   ` Marc Zyngier
2021-08-27  8:12     ` Huacai Chen
2021-08-28 10:07     ` Huacai Chen
2021-08-28 11:07       ` Marc Zyngier
2021-08-29  9:37         ` Huacai Chen [this message]
2021-08-29 10:10           ` Marc Zyngier
2021-08-29 10:34             ` Huacai Chen
2021-08-29 11:00               ` Marc Zyngier
2021-08-30  3:19                 ` Huacai Chen
2021-08-25  6:11 ` [PATCH V3 09/10] irqchip: Add Loongson Extended I/O " Huacai Chen
2021-08-25  6:11 ` [PATCH V3 10/10] irqchip: Add Loongson PCH LPC " Huacai Chen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAAhV-H41rridOo_3Eq5t9LPz-mefketAKNdhgbguFtZ0Cqz5Ng@mail.gmail.com \
    --to=chenhuacai@gmail.com \
    --cc=chenhuacai@loongson.cn \
    --cc=jiaxun.yang@flygoat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lixuefeng@loongson.cn \
    --cc=maz@kernel.org \
    --cc=tglx@linutronix.de \
    --subject='Re: [PATCH V3 08/10] irqchip: Add LoongArch CPU interrupt controller support' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).