[PATCH 1/3] dt-bindings: watchdog: ti,rti-wdt: Add support for WDIOF_CARDRESET

Li, Hua Qian HuaQian.Li at siemens.com
Tue Jul 4 23:18:58 PDT 2023


On Tue, 2023-07-04 at 17:48 +0100, Conor Dooley wrote:
> On Tue, Jul 04, 2023 at 06:33:01AM +0000, Li, Hua Qian wrote:
> > On Tue, 2023-07-04 at 07:46 +0200, Krzysztof Kozlowski wrote:
> > > On 04/07/2023 05:33, Li, Hua Qian wrote:
> > > > On Mon, 2023-07-03 at 17:35 +0200, Krzysztof Kozlowski wrote:
> > > > > On 03/07/2023 05:01, huaqian.li at siemens.com wrote:
> > > > > > From: Li Hua Qian <huaqian.li at siemens.com>
> 
> > > > > 
> > > > > > +
> > > > > > +  memory-size:
> > > > > > +    maxItems: 1
> > > > > > +    description:
> > > > > > +      Contains the watchdog reserved memory size.
> > > > > 
> > > > > There is no such property in reserved-memory consumer
> > > > > bindings...
> > > > > I
> > > > > mean, what size? Missing type/ref anyway.
> > > > > 
> > > > > Best regards,
> > > > > Krzysztof
> > > > 
> > > > Actually, I wasn't a big fan of the current binding and
> > > > properties
> > > > definition. I wanted to describe a region with specific size to
> > > > reserve the reset reason in specific memory. If possible,
> > > > please
> > > > give
> > > > me some advice, thanks!
> > > 
> > > I don't understand what do you need here. There is a size already
> > > -
> > > just
> > > open your DTS...
> > > 
> > > Best regards,
> > > Krzysztof
> > > 
> > Yes, you are right, writing them here looks strange. I was treating
> > them as part of hardware, then felt it is better to define in dts. 
> > 
> > If I define them in driver directly, assume the following, do you
> > think
> > it is a good idea?
> > 
> > // file: drivers/watchdog/rti_wdt.c 
> > #define PON_REASON_ADDR 0xA2000000
> > #define PON_REASON_SIZE 0x1000
> 
> I don't understand this, why can you not just use a reg entry?
> 
I personally thought it was not good to write like this before because
this region is a customized reserved region not the device's resources.

Anyway, I currently think it is the best way, and I plan to refactor it
like this.

Any other suggestions please let me know, thanks!

Best Regards,
Hua qian


More information about the linux-arm-kernel mailing list