From mboxrd@z Thu Jan 1 00:00:00 1970 From: robh@kernel.org (Rob Herring) Date: Mon, 15 Oct 2018 09:34:20 -0500 Subject: [PATCH 6/9] PM / OPP: dt-bindings: Add opp-interconnect-bw In-Reply-To: <20180827151112.25211-7-jcrouse@codeaurora.org> References: <20180827151112.25211-1-jcrouse@codeaurora.org> <20180827151112.25211-7-jcrouse@codeaurora.org> Message-ID: <20181015143420.GA16932@bogus> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Mon, Aug 27, 2018 at 09:11:09AM -0600, Jordan Crouse wrote: > Add the "opp-interconnect-bw" property to specify the > average and peak bandwidth for an interconnect path for > a specific operating power point. A separate bandwidth > pair can be specified for each of the interconnects > defined for the device by appending the interconnect > name to the property. > > Signed-off-by: Jordan Crouse > --- > Documentation/devicetree/bindings/opp/opp.txt | 36 +++++++++++++++++++ > 1 file changed, 36 insertions(+) > > diff --git a/Documentation/devicetree/bindings/opp/opp.txt b/Documentation/devicetree/bindings/opp/opp.txt > index c396c4c0af92..d714c084f36d 100644 > --- a/Documentation/devicetree/bindings/opp/opp.txt > +++ b/Documentation/devicetree/bindings/opp/opp.txt > @@ -170,6 +170,11 @@ Optional properties: > functioning of the current device at the current OPP (where this property is > present). > > +- opp-interconnect-bw-: This is an array of pairs specifying the average > + and peak bandwidth in bytes per second for the interconnect path known by > + 'name'. This should match the name(s) specified by interconnect-names in the > + device definition. > + I don't think this is good design with the name defined in one node and then used in the OPP table. First, '*-names' is typically the a name local to that node/device. If you had 2 instances of a device with a shared OPP table for the 2 instances, then you are going to have to make the names unique. Second, how exactly would having multiple b/w entries work? A given OPP frequency supports the sum of the b/w entries? What if some devices for b/w entries aren't currently active? This also seems like a mixture of using OPP table for setting GPU's bandwidth/freq and then the interconnect binding to set the interconnect's bandwidth. Perhaps we need a more unified approach. Not sure what that would look like though. Rob