Re: [PATCH v2 2/2] drivers/watchdog: ASPEED reference dev tree properties for config

[Christopher Bostic <cbostic@linux.vnet.ibm.com>]

On 6/28/17 11:06 AM, Guenter Roeck wrote:
> On Wed, Jun 28, 2017 at 10:55:08AM -0500, Christopher Bostic wrote:
>>
>> On 6/28/17 10:08 AM, Guenter Roeck wrote:
>>> On Wed, Jun 28, 2017 at 09:59:22AM -0500, Christopher Bostic wrote:
>>>> On 6/28/17 9:54 AM, Guenter Roeck wrote:
>>>>> On Wed, Jun 28, 2017 at 09:29:50AM -0500, Christopher Bostic wrote:
>>>>>> On 6/28/17 6:31 AM, Guenter Roeck wrote:
>>>>>>> On 06/27/2017 02:17 PM, Christopher Bostic wrote:
>>>>>>>> Reference the system device tree when configuring the watchdog
>>>>>>>> engines.  Set external signal mode on timeout if specified.
>>>>>>>> Set system reset on timeout if specified.
>>>>>>>>
>>>>>>>> Signed-off-by: Christopher Bostic <cbostic@linux.vnet.ibm.com>
>>>>>>>> ---
>>>>>>>> v2 - Change of_get_property() to of_property_read_bool()
>>>>>>>>     - Remove redundant check for NULL struct device_node pointer
>>>>>>>>     - Optional property names now start with prefix 'aspeed,'
>>>>>>>> ---
>>>>>>>>   drivers/watchdog/aspeed_wdt.c | 13 +++++++++++--
>>>>>>>>   1 file changed, 11 insertions(+), 2 deletions(-)
>>>>>>>>
>>>>>>>> diff --git a/drivers/watchdog/aspeed_wdt.c
>>>>>>>> b/drivers/watchdog/aspeed_wdt.c
>>>>>>>> index 1c65258..71ce5f5 100644
>>>>>>>> --- a/drivers/watchdog/aspeed_wdt.c
>>>>>>>> +++ b/drivers/watchdog/aspeed_wdt.c
>>>>>>>> @@ -140,6 +140,7 @@ static int aspeed_wdt_probe(struct platform_device
>>>>>>>> *pdev)
>>>>>>>>   {
>>>>>>>>       struct aspeed_wdt *wdt;
>>>>>>>>       struct resource *res;
>>>>>>>> +    struct device_node *np;
>>>>>>>>       int ret;
>>>>>>>>         wdt = devm_kzalloc(&pdev->dev, sizeof(*wdt), GFP_KERNEL);
>>>>>>>> @@ -170,8 +171,16 @@ static int aspeed_wdt_probe(struct platform_device
>>>>>>>> *pdev)
>>>>>>>>        * the SOC and not the full chip
>>>>>>>>        */
>>>>>>>>       wdt->ctrl = WDT_CTRL_RESET_MODE_SOC |
>>>>>>>> -        WDT_CTRL_1MHZ_CLK |
>>>>>>>> -        WDT_CTRL_RESET_SYSTEM;
>>>>>>>> +        WDT_CTRL_1MHZ_CLK;
>>>>>>>> +
>>>>>>>> +    np = pdev->dev.of_node;
>>>>>>>> +    if (of_property_read_bool(np, "aspeed,sys-reset"))
>>>>>>>> +        wdt->ctrl |= WDT_CTRL_RESET_SYSTEM;
>>>>>>>> +
>>>>>>> For backward compatibility, this should default to WDT_CTRL_RESET_SYSTEM
>>>>>>> if no optional property is provided.
>>>>>>>
>>>>>> I had the logic inverted for this property in a previous patch.  The
>>>>>> property was 'no-system-reset' so that when not present the default was to
>>>>>> set WDT_CTRL_RESET_SYSTEM.   As it is in this patch, the only way to
>>>>>> indicate that no system reset is to be done is to not specify the property.
>>>>>> No system reset is desired under circumstances when another wdt engine is to
>>>>>> be responsible for this.   Given the issue with backward compatibility
>>>>>> that's not a solution.  Given this, would creating a property
>>>>>> 'no-system-reset' be acceptable?
>>>>>>
>>>>> Sorry, I fail to see the problem. There are half a dozen properties. What is the
>>>>> problem with having a default if no property is specified ? Your default is "do
>>>>> nothing", which does not really make any sense to me. If the user wants the
>>>>> watchdog to do nothing, the simple means to accomplish that would be to not
>>>>> instantiate it.
>>>>>
>>>>> Sure, that means specifying "system-reset" is redundant, but I don't see a
>>>>> problem with that either. But I do see a problem with loading a watchdog driver
>>>>> that doesn't do anything.
>>>>>
>>>>> If there is really some use case where it makes sense to load a watchdog driver
>>>>> and have it do nothing, please explain and provide a respective devicetree
>>>>> property. "aspeed,do-nothing" (or whatever similar) doesn't sound good, but
>>>>> at least makes it obvious that the driver isn't doing anything besides
>>>>> creating a false sense of "the system is watchdog protected".
>>>> If system-reset is not wanted there are other properties that might still be
>>>> needed, for example
>>>> ARM reset only.   We'd still want to instantiate it.
>>>>
>>> But then you would presumably specify that property, so there would be one.
>>> The question here is what to do if _no_ property is provided.
>> I agree that the default should be to perform a system reset  when no
>> property is provided.  I will make that change.   Additionally for backwards
>> compatibility the SOC reset should be enabled by default when no property is
>> provided.   The other parameters are not configured in the watchdog control
>> register if no property is provided.
>>
>> There is still a need to to configure for no system reset.  For example,
>> optional parameter 'aspeed,no-sys-reset' that can be specified if system
>> reset is to be disabled.  In addition the default 'SOC reset enabled'
>> behavior could be disabled with an optional parameter 'aspeed,no-soc-reset'.
>>
>> Why would there even be a need configure for no system reset?  The external
>> mode 'wdt ext: External signal enable after timeout' can be sent off chip
>> even if system reset is disabled.  The watchdog is doing something in this
>> case and would require wdt instantiation.
>>
> Wouldn't that be "aspeed,external-signal" ?
Yes that optional parameter would need to be specified in that 
situation, as well as 'aspeed,no-system-reset'.

Chris
>
> Guenter
>
>> Hope this is more clear, sorry for the confusion.
>>
>> Thanks,
>> -Chris AST2500 Software Programming Guide
>>> Sorry, I don't get your point.
>>>
>>> Guenter
>>>
>>>> Thanks,
>>>> Chris
>>>>> Thanks,
>>>>> Guenter
>>>>>
>>>>>
>>>>>> Thanks,
>>>>>> Chris
>>>>>>
>>>>>>>> +    if (of_property_read_bool(np, "aspeed,external-signal"))
>>>>>>>> +        wdt->ctrl |= WDT_CTRL_WDT_EXT;
>>>>>>>> +
>>>>>>>> +    writel(wdt->ctrl, wdt->base + WDT_CTRL);
>>>>>>>>         if (readl(wdt->base + WDT_CTRL) & WDT_CTRL_ENABLE)  {
>>>>>>>>           aspeed_wdt_start(&wdt->wdd);
>>>>>>>>
>>>> --
>>>> To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in
>>>> the body of a message to majordomo@vger.kernel.org
>>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html