| Message ID | 20190107160512.9480-1-tony@atomide.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | ARM: dts: omap3-gta04: Fix graph_port warning | expand |
Hi Tony, > Am 07.01.2019 um 17:05 schrieb Tony Lindgren <tony@atomide.com>: > > We're currently getting a warning with make dtbs: > > arch/arm/boot/dts/omap3-gta04.dtsi:720.7-727.4: Warning (graph_port): > /ocp@68000000/dss@48050000/encoder@48050c0 0/port: graph node unit > address error, expected "0" yes, dtc has become more critical nowadays... To me it looks like redundant information, but it is how it is. > > Cc: H. Nikolaus Schaller <hns@goldelico.com> > Signed-off-by: Tony Lindgren <tony@atomide.com> > --- > arch/arm/boot/dts/omap3-gta04.dtsi | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/arm/boot/dts/omap3-gta04.dtsi b/arch/arm/boot/dts/omap3-gta04.dtsi > --- a/arch/arm/boot/dts/omap3-gta04.dtsi > +++ b/arch/arm/boot/dts/omap3-gta04.dtsi > @@ -717,7 +717,7 @@ > #address-cells = <1>; > #size-cells = <0>; > > - port { > + port@0 { > reg = <0>; > venc_out: endpoint { > remote-endpoint = <&opa_in>; Looks good to me. BR and thanks, NIkolaus
On Mon, Jan 7, 2019 at 10:21 AM H. Nikolaus Schaller <hns@goldelico.com> wrote: > > Hi Tony, > > > Am 07.01.2019 um 17:05 schrieb Tony Lindgren <tony@atomide.com>: > > > > We're currently getting a warning with make dtbs: > > > > arch/arm/boot/dts/omap3-gta04.dtsi:720.7-727.4: Warning (graph_port): > > /ocp@68000000/dss@48050000/encoder@48050c0 0/port: graph node unit > > address error, expected "0" > > yes, dtc has become more critical nowadays... > > To me it looks like redundant information, but it is how it is. If there's only one port, then it is redundant and you should drop reg instead. There should be a warning for that too. That one is probably off by default as that's more subjective and noisy. Rob
* Rob Herring <robh@kernel.org> [190107 18:32]: > On Mon, Jan 7, 2019 at 10:21 AM H. Nikolaus Schaller <hns@goldelico.com> wrote: > > > > Hi Tony, > > > > > Am 07.01.2019 um 17:05 schrieb Tony Lindgren <tony@atomide.com>: > > > > > > We're currently getting a warning with make dtbs: > > > > > > arch/arm/boot/dts/omap3-gta04.dtsi:720.7-727.4: Warning (graph_port): > > > /ocp@68000000/dss@48050000/encoder@48050c0 0/port: graph node unit > > > address error, expected "0" > > > > yes, dtc has become more critical nowadays... > > > > To me it looks like redundant information, but it is how it is. > > If there's only one port, then it is redundant and you should drop reg > instead. There should be a warning for that too. That one is probably > off by default as that's more subjective and noisy. OK dropping this patch. Nikolaus, care to post a better version since you can test it? Regards, Tony
> Am 07.01.2019 um 19:34 schrieb Tony Lindgren <tony@atomide.com>: > > * Rob Herring <robh@kernel.org> [190107 18:32]: >> On Mon, Jan 7, 2019 at 10:21 AM H. Nikolaus Schaller <hns@goldelico.com> wrote: >>> >>> Hi Tony, >>> >>>> Am 07.01.2019 um 17:05 schrieb Tony Lindgren <tony@atomide.com>: >>>> >>>> We're currently getting a warning with make dtbs: >>>> >>>> arch/arm/boot/dts/omap3-gta04.dtsi:720.7-727.4: Warning (graph_port): >>>> /ocp@68000000/dss@48050000/encoder@48050c0 0/port: graph node unit >>>> address error, expected "0" >>> >>> yes, dtc has become more critical nowadays... >>> >>> To me it looks like redundant information, but it is how it is. >> >> If there's only one port, then it is redundant and you should drop reg >> instead. If I remember correctly it was required by older omapdrm code (which did scan for a reg property) but that may have changed. Therefore I've added Tomi to this discussion. >> There should be a warning for that too. That one is probably >> off by default as that's more subjective and noisy. > > OK dropping this patch. Nikolaus, care to post a better > version since you can test it? > > Regards, > > Tony
On 07/01/19 21:34, H. Nikolaus Schaller wrote: >>> If there's only one port, then it is redundant and you should drop reg >>> instead. > > If I remember correctly it was required by older omapdrm code (which > did scan for a reg property) but that may have changed. Therefore I've > added Tomi to this discussion. I don't think so. The code looks for 'reg', but if it's not there, then it's considered 0. Tomi
* Tomi Valkeinen <tomi.valkeinen@ti.com> [190108 07:57]: > On 07/01/19 21:34, H. Nikolaus Schaller wrote: > > >>> If there's only one port, then it is redundant and you should drop reg > >>> instead. > > > > If I remember correctly it was required by older omapdrm code (which > > did scan for a reg property) but that may have changed. Therefore I've > > added Tomi to this discussion. > > I don't think so. The code looks for 'reg', but if it's not there, then > it's considered 0. FYI, the binding doc for the graph bindings is at Documentation/devicetree/bindings/graph.txt if there is any questions on how things should work. Regards, Tony
Ho Tony, Tomi, > Am 08.01.2019 um 16:56 schrieb Tony Lindgren <tony@atomide.com>: > > * Tomi Valkeinen <tomi.valkeinen@ti.com> [190108 07:57]: >> On 07/01/19 21:34, H. Nikolaus Schaller wrote: >> >>>>> If there's only one port, then it is redundant and you should drop reg >>>>> instead. >>> >>> If I remember correctly it was required by older omapdrm code (which >>> did scan for a reg property) but that may have changed. Therefore I've >>> added Tomi to this discussion. >> >> I don't think so. The code looks for 'reg', but if it's not there, then >> it's considered 0. > > FYI, the binding doc for the graph bindings is at > Documentation/devicetree/bindings/graph.txt if there is > any questions on how things should work. I interpret it that "port@number" is only needed if there are multiple ports and there should not be any "@" or "reg" for the single port case. Although it seems not to be explicitly forbidden. The &dss port doesn't have it either. So let's remove the reg = <0>; It also seems that #address-cells = <1>; #size-cells = <0>; is not required in this case and can also be removed. I've tested and it works with v5.0-rc1. There is an edited patch [1] but IMHO you as the original author should post it. BR, Nikolaus [1]: http://git.goldelico.com/?p=letux-kernel.git;a=commit;h=e46369390f62ce7176e91dfa443185e4b097ef4b
* H. Nikolaus Schaller <hns@goldelico.com> [190109 16:47]: > Ho Tony, Tomi, > > > Am 08.01.2019 um 16:56 schrieb Tony Lindgren <tony@atomide.com>: > > > > * Tomi Valkeinen <tomi.valkeinen@ti.com> [190108 07:57]: > >> On 07/01/19 21:34, H. Nikolaus Schaller wrote: > >> > >>>>> If there's only one port, then it is redundant and you should drop reg > >>>>> instead. > >>> > >>> If I remember correctly it was required by older omapdrm code (which > >>> did scan for a reg property) but that may have changed. Therefore I've > >>> added Tomi to this discussion. > >> > >> I don't think so. The code looks for 'reg', but if it's not there, then > >> it's considered 0. > > > > FYI, the binding doc for the graph bindings is at > > Documentation/devicetree/bindings/graph.txt if there is > > any questions on how things should work. > > I interpret it that "port@number" is only needed if there are > multiple ports and there should not be any "@" or "reg" for the > single port case. Although it seems not to be explicitly > forbidden. > > The &dss port doesn't have it either. > > So let's remove the reg = <0>; > > It also seems that > > #address-cells = <1>; > #size-cells = <0>; > > is not required in this case and can also be removed. > > I've tested and it works with v5.0-rc1. > > There is an edited patch [1] but IMHO you as the original > author should post it. Looks good to me. But since you already modified it, all you need to do is add your Signed-off-by after mine and mail it out :) That is if you want to keep me as the original author I guess for the description. Otherwise we have to go through one more round of emails where I post it you ack it and so on as I can't test it. Regards, Tony > [1]: http://git.goldelico.com/?p=letux-kernel.git;a=commit;h=e46369390f62ce7176e91dfa443185e4b097ef4b
Hi Tony, > Am 09.01.2019 um 18:01 schrieb Tony Lindgren <tony@atomide.com>: > > * H. Nikolaus Schaller <hns@goldelico.com> [190109 16:47]: >> Ho Tony, Tomi, >> >>> Am 08.01.2019 um 16:56 schrieb Tony Lindgren <tony@atomide.com>: >>> >>> * Tomi Valkeinen <tomi.valkeinen@ti.com> [190108 07:57]: >>>> On 07/01/19 21:34, H. Nikolaus Schaller wrote: >>>> >>>>>>> If there's only one port, then it is redundant and you should drop reg >>>>>>> instead. >>>>> >>>>> If I remember correctly it was required by older omapdrm code (which >>>>> did scan for a reg property) but that may have changed. Therefore I've >>>>> added Tomi to this discussion. >>>> >>>> I don't think so. The code looks for 'reg', but if it's not there, then >>>> it's considered 0. >>> >>> FYI, the binding doc for the graph bindings is at >>> Documentation/devicetree/bindings/graph.txt if there is >>> any questions on how things should work. >> >> I interpret it that "port@number" is only needed if there are >> multiple ports and there should not be any "@" or "reg" for the >> single port case. Although it seems not to be explicitly >> forbidden. >> >> The &dss port doesn't have it either. >> >> So let's remove the reg = <0>; >> >> It also seems that >> >> #address-cells = <1>; >> #size-cells = <0>; >> >> is not required in this case and can also be removed. >> >> I've tested and it works with v5.0-rc1. >> >> There is an edited patch [1] but IMHO you as the original >> author should post it. > > Looks good to me. But since you already modified it, all > you need to do is add your Signed-off-by after mine and > mail it out :) Well, the main job is mailing it out (collecting the destination addresses)... > That is if you want to keep me as the > original author I guess for the description. > > Otherwise we have to go through one more round of emails > where I post it you ack it and so on as I can't test it. Anyways, I send it asap. BR, Nikolaus > > Regards, > > Tony > >> [1]: http://git.goldelico.com/?p=letux-kernel.git;a=commit;h=e46369390f62ce7176e91dfa443185e4b097ef4b
diff --git a/arch/arm/boot/dts/omap3-gta04.dtsi b/arch/arm/boot/dts/omap3-gta04.dtsi --- a/arch/arm/boot/dts/omap3-gta04.dtsi +++ b/arch/arm/boot/dts/omap3-gta04.dtsi @@ -717,7 +717,7 @@ #address-cells = <1>; #size-cells = <0>; - port { + port@0 { reg = <0>; venc_out: endpoint { remote-endpoint = <&opa_in>;
We're currently getting a warning with make dtbs: arch/arm/boot/dts/omap3-gta04.dtsi:720.7-727.4: Warning (graph_port): /ocp@68000000/dss@48050000/encoder@48050c0 0/port: graph node unit address error, expected "0" Cc: H. Nikolaus Schaller <hns@goldelico.com> Signed-off-by: Tony Lindgren <tony@atomide.com> --- arch/arm/boot/dts/omap3-gta04.dtsi | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)