MIME-Version: 1.0
In-Reply-To: <CAMz4kuLQyD5=X7PEOX98_drt66yc=Fq+KxMFHDW6tp_yZsxtsQ@mail.gmail.com>
References: <1b80566bd849d68b0fc8de54ecbbc7b4efbb1077.1512708743.git.baolin.wang@spreadtrum.com>
 <c85f55a434efada592c50f46c0d0532fdb5340b0.1512708743.git.baolin.wang@spreadtrum.com>
 <05f19500-f8a3-7afc-a8d9-bcf0aff3e6bc@linaro.org> <CAMz4ku+10jTXe4bykByO_iLM33ACTu9jSnrX=i66MD3F79w9dQ@mail.gmail.com>
 <3c1da7bb-a380-072f-957d-221a6f177217@linaro.org> <CAMz4kuLQyD5=X7PEOX98_drt66yc=Fq+KxMFHDW6tp_yZsxtsQ@mail.gmail.com>
From: Arnd Bergmann <arnd@arndb.de>
Date: Tue, 12 Dec 2017 10:42:19 +0100
Message-ID: <CAK8P3a2okz==g-NvfZhYXettQfw+SmQ6aacZ8AoaKhrbJ36peQ@mail.gmail.com>
Subject: Re: [PATCH v3 2/2] clocksource: sprd: Add timer driver for Spreadtrum
 SC9860 platform
To: Baolin Wang <baolin.wang@linaro.org>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>,
        Baolin Wang <baolin.wang@spreadtrum.com>,
        Thomas Gleixner <tglx@linutronix.de>, Rob Herring <robh+dt@kernel.org>,
        Mark Rutland <mark.rutland@arm.com>, DTML <devicetree@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>, Mark Brown <broonie@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1543
Lines: 39

On Tue, Dec 12, 2017 at 10:26 AM, Baolin Wang <baolin.wang@linaro.org> wrote:
> On 12 December 2017 at 17:16, Daniel Lezcano <daniel.lezcano@linaro.org> wrote:
>> On 08/12/2017 09:20, Baolin Wang wrote:
>>>>> +static irqreturn_t sprd_timer_interrupt(int irq, void *dev_id)
>>>>> +{
>>>>> +     struct clock_event_device *ce = (struct clock_event_device *)dev_id;
>>>>> +     struct timer_of *to = to_timer_of(ce);
>>>>> +
>>>>> +     sprd_timer_clear_interrupt(timer_of_base(to));
>>>>> +
>>>>> +     if (clockevent_state_oneshot(ce))
>>>>> +             sprd_timer_disable(timer_of_base(to));
>>>>> +
>>>>> +     ce->event_handler(ce);
>>>>> +     return IRQ_HANDLED;
>>>>> +}
>>>>> +
>>>>> +static struct timer_of to = {
>>>>> +     .flags = TIMER_OF_IRQ | TIMER_OF_BASE,
>>>>
>>>> Why not the TIMER_OF_CLOCK ?
>>>
>>> The timer's clock is fixed to 32.768K and no need to divide the
>>> frequency, so our clock tree does not supply the timer's clock now.
>>
>> The driver is fine. However, I would like to unify the clk usage in the
>> timer driver, so if you refer to a clock so the TIMER_OF_CLOCK can be
>> used, that will nice.
>
> I understand your concern, but I've asked our clock driver owner in
> Spreadtrum, we have no related registers to describe the topology of
> the RTC fixed 32.768K, so the clock driver can not add timer's clock
> node.

I think you can easily put a separate DT node in the tree with
compatible="fixed-clock". This will be available during early boot,
and interpreted by setting TIMER_OF_CLOCK.

       Arnd