On Tue, May 03, 2022 at 02:26:15PM +0200, AngeloGioacchino Del Regno wrote: > Il 03/05/22 14:19, Krzysztof Kozlowski ha scritto: > > On 29/04/2022 23:11, Nícolas F. R. A. Prado wrote: > > > The SCP has three memory regions: sram, l1tcm and cfg. Only sram is > > > required, the other two are optional. Fix the dt-binding so that the > > > optional regions can be omitted and passed in any order. > > > > No, cannot be passed in any order. > > > > > > > > Also add the missing minItems to the reg property and update the > > > description. > > > > > > Signed-off-by: Nícolas F. R. A. Prado <nfraprado@xxxxxxxxxxxxx> > > > > > > --- > > > > > > .../devicetree/bindings/remoteproc/mtk,scp.yaml | 12 +++++++++--- > > > 1 file changed, 9 insertions(+), 3 deletions(-) > > > > > > diff --git a/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml b/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml > > > index 823a236242de..ec9ddeb6ca2c 100644 > > > --- a/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml > > > +++ b/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml > > > @@ -24,14 +24,20 @@ properties: > > > reg: > > > description: > > > Should contain the address ranges for memory regions SRAM, CFG, and > > > - L1TCM. > > > + L1TCM. Only SRAM is required, while CFG and L1TCM are optional. > > > + minItems: 1 > > > maxItems: 3 > > > reg-names: > > > + minItems: 1 > > > items: > > > - const: sram > > > - - const: cfg > > > - - const: l1tcm > > > + - enum: > > > + - l1tcm > > > + - cfg > > > + - enum: > > > + - l1tcm > > > + - cfg > > > > This allows them in any combination which is not what we want. If both > > are optional and both can appear, then last should be a const:l1tcm. > > > > Best regards, > > Krzysztof > > Nicolas, I think that you weren't clear about what you're trying to solve with this > commit in the description. > > I remember you had this kind of instance, but I don't really remember if it was > about mtk,scp or (and?) something else.... so.... are you trying to fix issues > with devicetrees declaring > > reg-names = "sram", "l1tcm"; ? > > Was this giving dtbs_check errors? Hi Angelo, yes, some devicetrees (like mt8183) have reg-names = "sram", "cfg"; I'll include mention of this in the commit description for v3. This was also supposed to fix the warning for mt8192, where the order is different: "sram", "l1tcm", "cfg". But since Krzysztof said that we want a fixed order, then the mt8192 DT will need to be updated. (I also just noticed that it's just l1tcm that is optional, I was just being blind, so I'll also fix that in v3) Thanks, Nícolas > > Cheers, > Angelo
