Re: [PATCH 1/2] dt-bindings: watchdog: Add arm,smc-wdt watchdog arm,smc-wdt compatible

From: Evan Benn <evanbenn@chromium.org>
To: Xingyu Chen <xingyu.chen@amlogic.com>
Cc: Julius Werner <jwerner@chromium.org>,
	Guenter Roeck <linux@roeck-us.net>, Rob Herring <robh@kernel.org>,
	LKML <linux-kernel@vger.kernel.org>,
	devicetree@vger.kernel.org,
	"David S. Miller" <davem@davemloft.net>,
	Jonathan Cameron <Jonathan.Cameron@huawei.com>,
	Mauro Carvalho Chehab <mchehab+samsung@kernel.org>,
	Wim Van Sebroeck <wim@linux-watchdog.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Mark Rutland <mark.rutland@arm.com>,
	linux-watchdog@vger.kernel.org,
	Jianxin Pan <jianxin.pan@amlogic.com>,
	Yonghui Yu <yonghui.yu@amlogic.com>
Subject: Re: [PATCH 1/2] dt-bindings: watchdog: Add arm,smc-wdt watchdog arm,smc-wdt compatible
Date: Mon, 24 Feb 2020 12:10:37 +1100	[thread overview]
Message-ID: <CAKz_xw3z7uXtK1SQKYLu_oP3GmqKR7RXvWj2b1cfyZO-Q-ZxQw@mail.gmail.com> (raw)
In-Reply-To: <1326f594-3cfd-c03d-4f2c-50eeb75724b2@amlogic.com>

Hello,

I think the intention is that this driver talks to a 'standard' arm
smc firmware watchdog call:

https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/3405

Each device could re-implement that ATF driver to talk to the specific hardware,
and could perhaps use a custom SMCWD_FUNC_ID, defined in the dts.

The goal was to provide an ATF patch and linux driver patch that would
be generic. But the above ATF patch
is only for mt8173. Right now it just specifies an interface. It has
less functionality than your meson driver Xingyu.
If it is not suitable, that is fine.

The above ATF patch is deployed on oak, elm, and hana mt8173
chromebook devices, this driver is intended to support those devices.

Evan

On Sat, Feb 22, 2020 at 3:01 PM Xingyu Chen <xingyu.chen@amlogic.com> wrote:
>
> Hi, Julius
>
> On 2020/2/22 3:41, Julius Werner wrote:
> >> Because the ATF does not define standard wdt index, each vendor defines
> >> its own index.
> >> So I don't think that the current driver[0] can fully cover my usecases.
> > I think the best way to solve this would be to put the SMC function ID
> > as another field into the device tree, so that multiple vendors could
> > share the same driver even if their firmware interface uses a
> > different SMC. But they still have to implement the same API for that
> > SMC, of course, not sure if the Meson driver is suitable for that (but
> > if it is then I think merging those drivers would be a good idea).
> The SMC function ID may be solved by the DTS, but the wdt indexs(Eg:
> SMCWD_INFO) are also different
> for each vendor. The imx_sc_wdt.c is also use the SMC to operate the
> WDT, but the wdt indexs(Eg: IMX_SIP_TIMER_START_WDOG)
> are different from ours. IMO, If the ATF can implement a common hal
> interface and index for watchdog, then writing a
> common smc wdt driver will be easier to compatible with all vendors.
>
> Best Regards
> >
> > .