* Re: [PATCH 1/3] dt-bindings: watchdog: ti,rti-wdt: Add support for WDIOF_CARDRESET
[not found] ` <20230704-revisit-punctured-f7342a00926b@spud>
@ 2023-07-05 6:18 ` Li, Hua Qian
0 siblings, 0 replies; only message in thread
From: Li, Hua Qian @ 2023-07-05 6:18 UTC (permalink / raw)
To: conor
Cc: linux-arm-kernel, linux, wim, krzysztof.kozlowski, kristo,
devicetree, Su, Bao Cheng, huaqianlee, nm, vigneshr, conor+dt,
krzysztof.kozlowski+dt, robh+dt, linux-kernel, linux-watchdog
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@siemens.com wrote:
> > > > > > From: Li Hua Qian <huaqian.li@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
^ permalink raw reply [flat|nested] only message in thread