From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benoit Cousson Date: Mon, 02 Jun 2014 07:11:01 +0000 Subject: Re: [PATCH 3/7] ARM: shmobile: Disable DVS support for r8a7790 Message-Id: <538C2385.40702@baylibre.com> List-Id: References: <5386E66D.3020502@bp.renesas.com> In-Reply-To: <5386E66D.3020502@bp.renesas.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable To: linux-sh@vger.kernel.org Hi, On 02/06/2014 05:14, Gaku Inami wrote: > Hi Geert > > Thank you for your feedback. > >> The role of the DT is to describe the hardware. >> Not using DVS because the effect of power consumption is low or non-exis= tent >> is a software policy. >> >> So perhaps it should be kept? > > I understand what you mean. I will keep the setting of VDD into DT, > and consider to disable DVS in a different way. > Indeed, the line "cpu0-supply =3D <&vdd_dvfs>;" does just provide the name = of the power rail that supply the ca15 cluster. The postfix _dvfs is a little bit misleading in this case, I do agree. If you do no want voltage scaling, it is just a matter of keeping the=20 same voltage in the OPP table. This is what I have done in the series I=20 sent a while back. That being said, what is the exact voltage range supported by the CPU? AFAIR, with 28nm process some SoCs are able to scale down to 0.9v. Thanks, Benoit --=20 Beno=C3=AEt Cousson BayLibre Embedded Linux Technology Lab www.baylibre.com