[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-Id: <20180605090827.u535p5du5em3kgil@aksadiga.ibm>
Date: Tue, 5 Jun 2018 14:38:27 +0530
From: Akshay Adiga <akshay.adiga@...ux.vnet.ibm.com>
To: Abhishek <huntbag@...ux.vnet.ibm.com>
Cc: Benjamin Herrenschmidt <benh@...nel.crashing.org>,
stewart@...ux.vnet.ibm.com, linux-pm@...r.kernel.org,
daniel.lezcano@...aro.org, rjw@...ysocki.net,
linux-kernel@...r.kernel.org, paulus@...ba.org,
linuxppc-dev@...ts.ozlabs.org
Subject: Re: [PATCH v2] cpuidle/powernv : Add Description for cpuidle state
On Tue, Jun 05, 2018 at 02:24:39PM +0530, Abhishek wrote:
>
>
> On 06/04/2018 05:15 PM, Akshay Adiga wrote:
> > On Mon, Jun 04, 2018 at 07:04:14PM +1000, Benjamin Herrenschmidt wrote:
> > > Is this a new property ? I'm not fan of adding yet another of those
> > > silly arrays.
> > >
> > > I would say this is the right time now to switch over to a node per
> > > state instead, as we discussed with Vaidy.
>
> It is not a new property. Name was being used for description as description
> was not present in device tree. A skiboot patch adding description to device
> tree have been posted. This patch reads those description instead of copying
> name itself into description. And we fall back to reading name into
> description to not break the comaptibility with older firmware.
>From a cpuidle point of view this is a exisiting property, but for
powernv there was no device-tree property describing this.
Abhishek has added the following skiboot patch for adding description
for each idle state in device-tree .
https://patchwork.ozlabs.org/patch/924879/
I agree this can go into new device tree format which each idle state
as a node. Probably i will roll this patch into mine in the next
version.
>
> Thanks
> Abhishek
>
> > I posted the node based device tree here :
> > skiboot patch : https://patchwork.ozlabs.org/patch/923120/
> > kernel patch : https://lkml.org/lkml/2018/5/30/1146
> >
> > Do you have any inputs for this design ?
> >
> > > Additionally, while doing that, we can provide the versioning mechanism
> > > I proposed so we can deal with state specific issues and erratas.
> > >
> > > Cheers,
> > > Ben.
> > >
>
Powered by blists - more mailing lists