Message-ID: <1c2ac5b1-4ba4-b4e9-8775-b1597d9aae99@roeck-us.net>
Date:   Mon, 14 Oct 2019 06:49:46 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Xingyu Chen <xingyu.chen@...ogic.com>,
        Wim Van Sebroeck <wim@...ux-watchdog.org>,
        Kevin Hilman <khilman@...libre.com>,
        Neil Armstrong <narmstrong@...libre.com>
Cc:     Rob Herring <robh+dt@...nel.org>,
        Jerome Brunet <jbrunet@...libre.com>,
        Qianggui Song <qianggui.song@...ogic.com>,
        Jianxin Pan <jianxin.pan@...ogic.com>,
        Jian Hu <jian.hu@...ogic.com>, linux-watchdog@...r.kernel.org,
        linux-amlogic@...ts.infradead.org,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/4] watchdog: add meson secure watchdog driver

On 10/14/19 4:42 AM, Xingyu Chen wrote:
> Hi, Guenter
> Thanks for your review.
> 
> On 2019/10/12 22:29, Guenter Roeck wrote:
>> On 10/12/19 3:05 AM, Xingyu Chen wrote:
>>> The watchdog controller on the Meson-A/C series SoCs is moved to secure
>>> world, watchdog operation needs to be done in secure EL3 mode via ATF,
>>> Non-secure world can call SMC instruction to trap to AFT for watchdog
>>> operation.
>>>
>>> Signed-off-by: Xingyu Chen <xingyu.chen@...ogic.com>
>>> ---
>>>   drivers/watchdog/Kconfig         |  16 +++
>>>   drivers/watchdog/Makefile        |   1 +
>>>   drivers/watchdog/meson_sec_wdt.c | 205 +++++++++++++++++++++++++++++++++++++++
>>>   3 files changed, 222 insertions(+)
>>>   create mode 100644 drivers/watchdog/meson_sec_wdt.c
>>>
>>> diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
>>> index 58e7c10..e6b0707 100644
>>> --- a/drivers/watchdog/Kconfig
>>> +++ b/drivers/watchdog/Kconfig
>>> @@ -826,6 +826,22 @@ config MESON_GXBB_WATCHDOG
>>>         To compile this driver as a module, choose M here: the
>>>         module will be called meson_gxbb_wdt.
>>> +config MESON_SEC_WATCHDOG
>>> +    tristate "Amlogic Meson Secure watchdog support"
>>> +    depends on ARCH_MESON || COMPILE_TEST
>>
>> Did you try COMPILE_TEST (eg allmodconfig) on, say x86_64 ?
>> AFAICS the meson sm calls are only available if MESON_SM is
>> enabled, and that depends on both ARCH_MESON and ARM64_4K_PAGES.
>> This dependency is not expressed here, and neither is enabled
>> with COMPILE_TEST.Sorry, I have't done this kind of test before for this patchset. There
> is a kernel build error which related to the current driver when i try to use allmodconfig with x86. I will fix it by adding "depends on MESON_SM" in next version.
>>
>>> +    select WATCHDOG_CORE
>>> +    help
>>> +      The watchdog controller on the Meson-A/C series SoCs is moved to
>>> +      secure world, watchdog operation needs to be done in secure EL3
>>> +      mode via ATF, non-secure world can call SMC instruction to trap
>>> +      to ATF for the watchdog operation.
>>> +
>>> +      Say Y here if watchdog controller on Meson SoCs is located in
>>> +      secure world.
>>> +
>>> +      To compile this driver as a module, choose M here: the
>>> +      module will be called meson_sec_wdt.
>>> +
>>>   config MESON_WATCHDOG
>>>       tristate "Amlogic Meson SoCs watchdog support"
>>>       depends on ARCH_MESON || COMPILE_TEST
>>> diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
>>> index 2ee352b..5e6b73d 100644
>>> --- a/drivers/watchdog/Makefile
>>> +++ b/drivers/watchdog/Makefile
>>> @@ -78,6 +78,7 @@ obj-$(CONFIG_QCOM_WDT) += qcom-wdt.o
>>>   obj-$(CONFIG_BCM_KONA_WDT) += bcm_kona_wdt.o
>>>   obj-$(CONFIG_TEGRA_WATCHDOG) += tegra_wdt.o
>>>   obj-$(CONFIG_MESON_GXBB_WATCHDOG) += meson_gxbb_wdt.o
>>> +obj-$(CONFIG_MESON_SEC_WATCHDOG) += meson_sec_wdt.o
>>>   obj-$(CONFIG_MESON_WATCHDOG) += meson_wdt.o
>>>   obj-$(CONFIG_MEDIATEK_WATCHDOG) += mtk_wdt.o
>>>   obj-$(CONFIG_DIGICOLOR_WATCHDOG) += digicolor_wdt.o
>>> diff --git a/drivers/watchdog/meson_sec_wdt.c b/drivers/watchdog/meson_sec_wdt.c
>>> new file mode 100644
>>> index 00000000..2b5357c
>>> --- /dev/null
>>> +++ b/drivers/watchdog/meson_sec_wdt.c
>>> @@ -0,0 +1,205 @@
>>> +// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
>>> +/*
>>> + * Copyright (c) 2019 Amlogic, Inc. All rights reserved.
>>> + * Author: Xingyu Chen <xingyu.chen@...ogic.com>
>>> + *
>>> + */
>>> +#include <linux/err.h>
>>> +#include <linux/module.h>
>>> +#include <linux/of.h>
>>> +#include <linux/platform_device.h>
>>> +#include <linux/types.h>
>>> +#include <linux/watchdog.h>
>>> +#include <linux/firmware/meson/meson_sm.h>
>>> +
>>> +#define MESON_SIP_WDT_DISABLE        0x1
>>> +#define MESON_SIP_WDT_ENABLE        0x2
>>> +#define MESON_SIP_WDT_PING        0x3
>>> +#define MESON_SIP_WDT_INIT        0x4
>>> +#define MESON_SIP_WDT_RESETNOW        0x5
>>> +#define MESON_SIP_WDT_SETTIMEOUT    0x6
>>> +#define MESON_SIP_WDT_GETTIMELEFT    0x7
>>> +
>>> +#define DEFAULT_TIMEOUT            30 /* seconds */
>>> +
>>> +/*
>>> + * Watchdog timer tick is set to 1ms in secfw side, and tick count is
>>> + * stored in the bit[16-31] of WATCHDOG_CNT register, so the maximum
>>> + * timeout value is 0xffff ms.
>>> + */
>>> +#define MAX_TIMEOUT_MS            0xFFFF
>>> +
>>> +struct meson_sec_wdt {
>>> +    struct watchdog_device wdt_dev;
>>> +    struct meson_sm_firmware *fw;
>>> +};
>>> +
>>> +static int meson_sec_wdt_start(struct watchdog_device *wdt_dev)
>>> +{
>>> +    int ret;
>>> +    struct meson_sec_wdt *data = watchdog_get_drvdata(wdt_dev);
>>> +
>>> +    ret = meson_sm_call(data->fw, SM_WATCHDOG_OPS, NULL,
>>> +                MESON_SIP_WDT_ENABLE, 0, 0, 0, 0); > +    if (ret)
>>> +        return ret;
>>> +
>>> +    return 0;
>>
>> This is equivalent to
>>      return ret;
>> or even
>>      return meson_sm_call(...);
>> I will fix it in next version.
>>> +}
>>> +
>>> +static int meson_sec_wdt_stop(struct watchdog_device *wdt_dev)
>>> +{
>>> +    int ret;
>>> +    struct meson_sec_wdt *data = watchdog_get_drvdata(wdt_dev);
>>> +
>>> +    ret = meson_sm_call(data->fw, SM_WATCHDOG_OPS, NULL,
>>> +                MESON_SIP_WDT_DISABLE, 0, 0, 0, 0);
>>> +    if (ret)
>>> +        return ret;
>>> +
>>> +    return 0;
>>
>> Same as above.
> I will fix it in next version.
>>
>>> +}
>>> +
>>> +static int meson_sec_wdt_ping(struct watchdog_device *wdt_dev)
>>> +{
>>> +    struct meson_sec_wdt *data = watchdog_get_drvdata(wdt_dev);
>>> +
>>> +    meson_sm_call(data->fw, SM_WATCHDOG_OPS, NULL,
>>> +              MESON_SIP_WDT_PING, 0, 0, 0, 0);
>>> +
>>> +    return 0;
>>
>> Why ignore errors ?
> I will fix it as above.
>>
>>> +}
>>> +
>>> +static int meson_sec_wdt_set_timeout(struct watchdog_device *wdt_dev,
>>> +                     unsigned int timeout)
>>> +{
>>> +    int ret;
>>> +    struct meson_sec_wdt *data = watchdog_get_drvdata(wdt_dev);
>>> +
>>> +    wdt_dev->timeout = timeout;
>>> +    meson_sec_wdt_ping(wdt_dev);
>>> +
>>
>> Unconditionally ? Also, the core does that after setting the timeoutIt is used to avoid watchdog timer expire once the timeout is updated, 
> and that the watchdog_ping from core not been invoked. If it is considered useless, i will remove it.
> 

As mentioned, the core sends a ping immediately after the update.

>>
>>> +    ret = meson_sm_call(data->fw, SM_WATCHDOG_OPS, NULL,
>>> +                MESON_SIP_WDT_SETTIMEOUT,
>>> +                wdt_dev->timeout, 0, 0, 0);
>>> +    if (ret)
>>> +        return ret;
>>> +
>>> +    return 0;
>>
>> same as above.
> I will fix it in next version.
>>
>>> +}
>>> +
>>> +static unsigned int meson_sec_wdt_get_timeleft(struct watchdog_device *wdt_dev)
>>> +{
>>> +    int ret;
>>> +    int timeleft;
>>> +    struct meson_sec_wdt *data = watchdog_get_drvdata(wdt_dev);
>>> +
>>> +    ret = meson_sm_call(data->fw, SM_WATCHDOG_OPS, &timeleft,
>>> +                MESON_SIP_WDT_GETTIMELEFT, 0, 0, 0, 0);
>>> +
>>> +    if (ret)
>>> +        return 0;
>>
>> Really ? Why ? 0 is most definitely incorrect here.
> I just want to return a invalid value after I failed to get valid timeleft. Could you tell me what is corrent value ?

Why not return the error ?

>>
>>> +
>>> +    return timeleft;
>>> +}
>>> +
>>> +static const struct watchdog_ops meson_sec_wdt_ops = {
>>> +    .start = meson_sec_wdt_start,
>>> +    .stop = meson_sec_wdt_stop,
>>> +    .ping = meson_sec_wdt_ping,
>>> +    .set_timeout = meson_sec_wdt_set_timeout,
>>> +    .get_timeleft = meson_sec_wdt_get_timeleft,
>>> +};
>>> +
>>> +static const struct watchdog_info meson_sec_wdt_info = {
>>> +    .identity = "Meson Secure Watchdog Timer",
>>> +    .options = WDIOF_SETTIMEOUT | WDIOF_KEEPALIVEPING | WDIOF_MAGICCLOSE,
>>> +};
>>> +
>>> +static int __maybe_unused meson_sec_wdt_resume(struct device *dev)
>>> +{
>>> +    struct meson_sec_wdt *data = dev_get_drvdata(dev);
>>> +
>>> +    if (watchdog_active(&data->wdt_dev))
>>> +        meson_sec_wdt_start(&data->wdt_dev);
>>
>> No error return ?
> I will fix it in next version
>>
>>> +
>>> +    return 0;
>>> +}
>>> +
>>> +static int __maybe_unused meson_sec_wdt_suspend(struct device *dev)
>>> +{
>>> +    struct meson_sec_wdt *data = dev_get_drvdata(dev);
>>> +
>>> +    if (watchdog_active(&data->wdt_dev))
>>> +        meson_sec_wdt_stop(&data->wdt_dev);
>>
>> No error return ?
> same as above
>>
>>> +
>>> +    return 0;
>>> +}
>>> +
>>> +static const struct dev_pm_ops meson_sec_wdt_pm_ops = {
>>> +    SET_SYSTEM_SLEEP_PM_OPS(meson_sec_wdt_suspend, meson_sec_wdt_resume)
>>> +};
>>> +
>>> +static const struct of_device_id meson_sec_wdt_dt_ids[] = {
>>> +     { .compatible = "amlogic,meson-sec-wdt", },
>>> +     { /* sentinel */ },
>>> +};
>>> +MODULE_DEVICE_TABLE(of, meson_sec_wdt_dt_ids);
>>> +
>>> +static int meson_sec_wdt_probe(struct platform_device *pdev)
>>> +{
>>> +    struct device *dev = &pdev->dev;
>>> +    struct meson_sec_wdt *data;
>>> +    struct device_node *sm_np;
>>> +    int ret;
>>> +
>>> +    data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
>>> +    if (!data)
>>> +        return -ENOMEM;
>>> +
>>> +    sm_np = of_parse_phandle(pdev->dev.of_node, "secure-monitor", 0);
>>> +    if (!sm_np) {
>>> +        dev_err(&pdev->dev, "no secure-monitor node\n");
>>> +        return -ENODEV;
>>
>> ENODEV is wrong here.
> I found some drivers use the ENODEV in this case, Eg:
> - ./watchdog/ts4800_wdt.c -122
> - ./nvmem/meson-efuse.c   -55
> ...
> 
> Could you tell me which error code is more suitable in this case ?
> 

Yes, I understand, with Millions of LOC to choose from, you will
find examples for everything somewhere in the Linux kernel.
That doesn't mean it is a good idea.

The property is listed as mandatory, so EINVAL would probably be
most appropriate, unless DT handling proposes a different error code
in such situations.

>>
>>> +    }
>>> +
>>> +    data->fw = meson_sm_get(sm_np);
>>> +    of_node_put(sm_np);
>>> +    if (!data->fw)
>>> +        return -EPROBE_DEFER;
>>
>> How do you know ?
> The fw pointer is static variable which is initialized to NULL. if it is equal to NULL, and indicate the meson_sm driver is not registered, then wdt driver use EPROBE_DEFER to defer probe.
>>
>>> +
>>> +    platform_set_drvdata(pdev, data);
>>> +
>>> +    data->wdt_dev.parent = dev;
>>> +    data->wdt_dev.info = &meson_sec_wdt_info;
>>> +    data->wdt_dev.ops = &meson_sec_wdt_ops;
>>> +    data->wdt_dev.max_hw_heartbeat_ms = MAX_TIMEOUT_MS;
>>> +    data->wdt_dev.min_timeout = 1;
>>> +    data->wdt_dev.timeout = DEFAULT_TIMEOUT;
>>> +    watchdog_set_drvdata(&data->wdt_dev, data);
>>> +
>> No watchdog_init_timeout() ? Any special reason for not supporting
>> to set the timeout with a devicetree property ?
> I want to keep consistency with previous meson watchdog driver.
> as follows:
> - meson_gxbb_wdt.c
> - meson_wdt.c
> 

Yes, again, it is possible to find examples for pretty much everything
in the linux kernel. But, sure, as you wish.

Guenter

Powered by blists - more mailing lists