linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: Evan Benn <evanbenn@gmail.com>
To: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Evan Benn <evanbenn@chromium.org>,
	Matthias Brugger <matthias.bgg@gmail.com>,
	 Stanley Chu <stanley.chu@mediatek.com>,
	Alexey Klimov <alexey.klimov@linaro.org>,
	 Julia Lawall <Julia.Lawall@lip6.fr>,
	Yingjoe Chen <yingjoe.chen@mediatek.com>,
	 Thomas Gleixner <tglx@linutronix.de>,
	linux-mediatek@lists.infradead.org,
	 linux-arm-kernel@lists.infradead.org,
	Viresh Kumar <viresh.kumar@linaro.org>,
	 LKML <linux-kernel@vger.kernel.org>,
	Fabien Parent <fparent@baylibre.com>
Subject: Re: [PATCH v2] drivers/clocksource/mediatek: Ack and disable interrupts on shutdown
Date: Sat, 27 Mar 2021 12:31:46 +1100	[thread overview]
Message-ID: <CAP8nV8rv6bKARZ31fVqcjVgMe+5UyVG8UAyNsm1hDCP2BbRSgw@mail.gmail.com> (raw)
In-Reply-To: <049946ac-5263-21ee-9651-7295f0bcf387@linaro.org>

Hi Daniel,

That is a good point, and I did try that at first and it works fine. I
uploaded this version because the suspend/resume callbacks were
undocumented and mostly not used by other clocksource drivers. I
thought a smaller diff might be preferable.
I also thought it would be better to shut off the interrupt as soon as
it is not needed, avoiding any other potential bugs, instead of just
fixing the one we know about with suspend. I'm not sure how the other
driver / timer disable flows are intended to work for example
(shutdown, detach, etc).

That said I am happy to upload that version if people think it is better.

https://elixir.bootlin.com/linux/latest/source/include/linux/clockchips.h#L120


On Thu, 25 Mar 2021 at 19:10, Daniel Lezcano <daniel.lezcano@linaro.org> wrote:
>
> On 25/03/2021 02:35, Evan Benn wrote:
> > set_state_shutdown is called during system suspend after interrupts have
> > been disabled. If the timer has fired in the meantime, there will be
> > a pending IRQ. So we ack that now and disable the timer. Without this
> > ARM trusted firmware will abort the suspend due to the pending
> > interrupt.
> >
> > Now always disable the IRQ in state transitions, and re-enable in
> > set_periodic and next_event.
>
> Why not put add the suspend/resume callbacks and put there the specific
> code and let the irq untouched in the normal flow ?
>
> > Signed-off-by: Evan Benn <evanbenn@chromium.org>
> > ---
> >
> > Changes in v2:
> > Remove the patch that splits the drivers into 2 files.
> >
> >  drivers/clocksource/timer-mediatek.c | 49 +++++++++++++++++-----------
> >  1 file changed, 30 insertions(+), 19 deletions(-)
> >
> > diff --git a/drivers/clocksource/timer-mediatek.c b/drivers/clocksource/timer-mediatek.c
> > index 9318edcd8963..fba2f9494d90 100644
> > --- a/drivers/clocksource/timer-mediatek.c
> > +++ b/drivers/clocksource/timer-mediatek.c
> > @@ -132,13 +132,33 @@ static u64 notrace mtk_gpt_read_sched_clock(void)
> >       return readl_relaxed(gpt_sched_reg);
> >  }
> >
> > +static void mtk_gpt_disable_ack_interrupts(struct timer_of *to, u8 timer)
> > +{
> > +     u32 val;
> > +
> > +     /* Disable interrupts */
> > +     val = readl(timer_of_base(to) + GPT_IRQ_EN_REG);
> > +     writel(val & ~GPT_IRQ_ENABLE(timer), timer_of_base(to) +
> > +            GPT_IRQ_EN_REG);
> > +
> > +     /* Ack interrupts */
> > +     writel(GPT_IRQ_ACK(timer), timer_of_base(to) + GPT_IRQ_ACK_REG);
> > +}
> > +
> >  static void mtk_gpt_clkevt_time_stop(struct timer_of *to, u8 timer)
> >  {
> >       u32 val;
> >
> > +     /* Disable timer */
> >       val = readl(timer_of_base(to) + GPT_CTRL_REG(timer));
> >       writel(val & ~GPT_CTRL_ENABLE, timer_of_base(to) +
> >              GPT_CTRL_REG(timer));
> > +
> > +     /* This may be called with interrupts disabled,
> > +      * so we need to ack any interrupt that is pending
> > +      * Or for example ATF will prevent a suspend from completing.
> > +      */
> > +     mtk_gpt_disable_ack_interrupts(to, timer);
> >  }
> >
> >  static void mtk_gpt_clkevt_time_setup(struct timer_of *to,
> > @@ -152,8 +172,10 @@ static void mtk_gpt_clkevt_time_start(struct timer_of *to,
> >  {
> >       u32 val;
> >
> > -     /* Acknowledge interrupt */
> > -     writel(GPT_IRQ_ACK(timer), timer_of_base(to) + GPT_IRQ_ACK_REG);
> > +     /* Enable interrupts */
> > +     val = readl(timer_of_base(to) + GPT_IRQ_EN_REG);
> > +     writel(val | GPT_IRQ_ENABLE(timer),
> > +            timer_of_base(to) + GPT_IRQ_EN_REG);
> >
> >       val = readl(timer_of_base(to) + GPT_CTRL_REG(timer));
> >
> > @@ -226,21 +248,6 @@ __init mtk_gpt_setup(struct timer_of *to, u8 timer, u8 option)
> >              timer_of_base(to) + GPT_CTRL_REG(timer));
> >  }
> >
> > -static void mtk_gpt_enable_irq(struct timer_of *to, u8 timer)
> > -{
> > -     u32 val;
> > -
> > -     /* Disable all interrupts */
> > -     writel(0x0, timer_of_base(to) + GPT_IRQ_EN_REG);
> > -
> > -     /* Acknowledge all spurious pending interrupts */
> > -     writel(0x3f, timer_of_base(to) + GPT_IRQ_ACK_REG);
> > -
> > -     val = readl(timer_of_base(to) + GPT_IRQ_EN_REG);
> > -     writel(val | GPT_IRQ_ENABLE(timer),
> > -            timer_of_base(to) + GPT_IRQ_EN_REG);
> > -}
> > -
> >  static struct timer_of to = {
> >       .flags = TIMER_OF_IRQ | TIMER_OF_BASE | TIMER_OF_CLOCK,
> >
> > @@ -292,6 +299,12 @@ static int __init mtk_gpt_init(struct device_node *node)
> >       if (ret)
> >               return ret;
> >
> > +     /* In case the firmware left the interrupts enabled
> > +      * disable and ack those now
> > +      */
> > +     mtk_gpt_disable_ack_interrupts(&to, TIMER_CLK_SRC);
> > +     mtk_gpt_disable_ack_interrupts(&to, TIMER_CLK_EVT);
> > +
> >       /* Configure clock source */
> >       mtk_gpt_setup(&to, TIMER_CLK_SRC, GPT_CTRL_OP_FREERUN);
> >       clocksource_mmio_init(timer_of_base(&to) + GPT_CNT_REG(TIMER_CLK_SRC),
> > @@ -305,8 +318,6 @@ static int __init mtk_gpt_init(struct device_node *node)
> >       clockevents_config_and_register(&to.clkevt, timer_of_rate(&to),
> >                                       TIMER_SYNC_TICKS, 0xffffffff);
> >
> > -     mtk_gpt_enable_irq(&to, TIMER_CLK_EVT);
> > -
> >       return 0;
> >  }
> >  TIMER_OF_DECLARE(mtk_mt6577, "mediatek,mt6577-timer", mtk_gpt_init);
> >
>
>
> --
> <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
>
> Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
> <http://twitter.com/#!/linaroorg> Twitter |
> <http://www.linaro.org/linaro-blog/> Blog

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

  reply	other threads:[~2021-03-27  1:34 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-25  1:35 [PATCH v2] drivers/clocksource/mediatek: Ack and disable interrupts on shutdown Evan Benn
2021-03-25  8:07 ` Daniel Lezcano
2021-03-27  1:31   ` Evan Benn [this message]
2021-04-02 14:15     ` Daniel Lezcano
2021-04-06  1:01       ` Evan Benn

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=CAP8nV8rv6bKARZ31fVqcjVgMe+5UyVG8UAyNsm1hDCP2BbRSgw@mail.gmail.com \
    --to=evanbenn@gmail.com \
    --cc=Julia.Lawall@lip6.fr \
    --cc=alexey.klimov@linaro.org \
    --cc=daniel.lezcano@linaro.org \
    --cc=evanbenn@chromium.org \
    --cc=fparent@baylibre.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mediatek@lists.infradead.org \
    --cc=matthias.bgg@gmail.com \
    --cc=stanley.chu@mediatek.com \
    --cc=tglx@linutronix.de \
    --cc=viresh.kumar@linaro.org \
    --cc=yingjoe.chen@mediatek.com \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).