| Message ID | 20241011023724.614584-9-mailingradian@gmail.com (mailing list archive) |
|---|---|
| State | Not Applicable |
| Headers | show |
| Series | Add SDM670 camera subsystem | expand |
Hello Richard, On 10/11/24 05:37, Richard Acayan wrote: > As found in the Pixel 3a, the Snapdragon 670 has a camera subsystem with > 3 CSIDs and 3 VFEs (including 1 VFE lite). Add this camera subsystem to > the bindings. > > Adapted from SC8280XP camera subsystem. > > Signed-off-by: Richard Acayan <mailingradian@gmail.com> > --- > .../bindings/media/qcom,sdm670-camss.yaml | 318 ++++++++++++++++++ > 1 file changed, 318 insertions(+) > create mode 100644 Documentation/devicetree/bindings/media/qcom,sdm670-camss.yaml > > diff --git a/Documentation/devicetree/bindings/media/qcom,sdm670-camss.yaml b/Documentation/devicetree/bindings/media/qcom,sdm670-camss.yaml > new file mode 100644 > index 000000000000..670502532d28 > --- /dev/null > +++ b/Documentation/devicetree/bindings/media/qcom,sdm670-camss.yaml > @@ -0,0 +1,318 @@ > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) > +%YAML 1.2 > +--- > +$id: http://devicetree.org/schemas/media/qcom,sdm670-camss.yaml# > +$schema: http://devicetree.org/meta-schemas/core.yaml# > + > +title: Qualcomm SDM670 Camera Subsystem (CAMSS) > + > +maintainers: > + - Richard Acayan <mailingradian@gmail.com> > + > +description: > + The CAMSS IP is a CSI decoder and ISP present on Qualcomm platforms. > + > +properties: > + compatible: > + const: qcom,sdm670-camss > + > + reg: > + maxItems: 9 > + > + reg-names: > + items: > + - const: csid0 > + - const: csid1 > + - const: csid2 > + - const: csiphy0 > + - const: csiphy1 > + - const: csiphy2 > + - const: vfe0 > + - const: vfe1 > + - const: vfe_lite > + > + clocks: > + maxItems: 22 > + > + clock-names: > + items: > + - const: gcc_camera_ahb > + - const: gcc_camera_axi > + - const: soc_ahb > + - const: camnoc_axi > + - const: cpas_ahb > + - const: csi0 > + - const: csi1 > + - const: csi2 > + - const: csiphy0 > + - const: csiphy0_timer > + - const: csiphy1 > + - const: csiphy1_timer > + - const: csiphy2 > + - const: csiphy2_timer > + - const: vfe0_axi > + - const: vfe0 > + - const: vfe0_cphy_rx > + - const: vfe1_axi > + - const: vfe1 > + - const: vfe1_cphy_rx > + - const: vfe_lite > + - const: vfe_lite_cphy_rx > + > + interrupts: > + maxItems: 9 > + > + interrupt-names: > + items: > + - const: csid0 > + - const: csid1 > + - const: csid2 > + - const: csiphy0 > + - const: csiphy1 > + - const: csiphy2 > + - const: vfe0 > + - const: vfe1 > + - const: vfe_lite > + > + iommus: > + maxItems: 4 > + > + power-domains: > + items: > + - description: IFE0 GDSC - Image Front End, Global Distributed Switch Controller. > + - description: IFE1 GDSC - Image Front End, Global Distributed Switch Controller. > + - description: Titan Top GDSC - Titan ISP Block, Global Distributed Switch Controller. > + > + power-domain-names: > + items: > + - const: ife0 > + - const: ife1 > + - const: top > + > + ports: > + $ref: /schemas/graph.yaml#/properties/ports > + > + description: > + CSI input ports. > + > + properties: > + port@0: > + $ref: /schemas/graph.yaml#/$defs/port-base > + unevaluatedProperties: false > + description: > + Input port for receiving CSI data from CSIPHY0. > + > + properties: > + endpoint: > + $ref: video-interfaces.yaml# > + unevaluatedProperties: false > + > + properties: > + clock-lanes: > + maxItems: 1 > + > + data-lanes: > + minItems: 1 > + maxItems: 4 > + > + required: > + - clock-lanes > + - data-lanes > + > + port@1: > + $ref: /schemas/graph.yaml#/$defs/port-base > + unevaluatedProperties: false > + description: > + Input port for receiving CSI data from CSIPHY1. > + > + properties: > + endpoint: > + $ref: video-interfaces.yaml# > + unevaluatedProperties: false > + > + properties: > + clock-lanes: > + maxItems: 1 > + > + data-lanes: > + minItems: 1 > + maxItems: 4 > + > + required: > + - clock-lanes > + - data-lanes > + > + port@2: > + $ref: /schemas/graph.yaml#/$defs/port-base > + unevaluatedProperties: false > + description: > + Input port for receiving CSI data from CSIPHY2. > + > + properties: > + endpoint: > + $ref: video-interfaces.yaml# > + unevaluatedProperties: false > + > + properties: > + clock-lanes: > + maxItems: 1 > + > + data-lanes: > + minItems: 1 > + maxItems: 4 > + > + required: > + - clock-lanes > + - data-lanes > + > + vdda-phy-supply: > + description: > + Phandle to a regulator supply to PHY core block. > + > + vdda-pll-supply: > + description: > + Phandle to 1.8V regulator supply to PHY refclk pll block. > + > +required: > + - reg > + - reg-names > + - clock-names > + - clocks > + - compatible > + - interrupts > + - interrupt-names > + - iommus > + - power-domains > + - power-domain-names > + - vdda-phy-supply > + - vdda-pll-supply > + > +additionalProperties: false > + > +examples: > + - | > + #include <dt-bindings/clock/qcom,camcc-sdm845.h> > + #include <dt-bindings/clock/qcom,gcc-sdm845.h> > + #include <dt-bindings/interrupt-controller/arm-gic.h> > + > + soc { > + #address-cells = <2>; > + #size-cells = <2>; > + > + camss@ac65000 { > + compatible = "qcom,sdm670-camss"; > + > + reg = <0 0x0acb3000 0 0x1000>, This is immediately wrong, unit address shall be the same as the address of the first value of reg property. I still object to the sorting order of reg values dictated by reg-names property. There are a few recently added CAMSS device tree binding descriptions, where reg values are sorted by address values without a connection to another property values, and I believe this is the correct way to go. Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe from now on it should be assumed that all subsequently added CAMSS IP descriptions to follow the same established policy. I vote for it. > + <0 0x0acba000 0 0x1000>, > + <0 0x0acc8000 0 0x1000>, > + <0 0x0ac65000 0 0x1000>, > + <0 0x0ac66000 0 0x1000>, > + <0 0x0ac67000 0 0x1000>, > + <0 0x0acaf000 0 0x4000>, > + <0 0x0acb6000 0 0x4000>, > + <0 0x0acc4000 0 0x4000>; > + reg-names = "csid0", > + "csid1", > + "csid2", > + "csiphy0", > + "csiphy1", > + "csiphy2", > + "vfe0", > + "vfe1", > + "vfe_lite"; > + > + interrupts = <GIC_SPI 464 IRQ_TYPE_EDGE_RISING>, > + <GIC_SPI 466 IRQ_TYPE_EDGE_RISING>, > + <GIC_SPI 468 IRQ_TYPE_EDGE_RISING>, > + <GIC_SPI 477 IRQ_TYPE_EDGE_RISING>, > + <GIC_SPI 478 IRQ_TYPE_EDGE_RISING>, > + <GIC_SPI 479 IRQ_TYPE_EDGE_RISING>, > + <GIC_SPI 465 IRQ_TYPE_EDGE_RISING>, > + <GIC_SPI 467 IRQ_TYPE_EDGE_RISING>, > + <GIC_SPI 469 IRQ_TYPE_EDGE_RISING>; > + interrupt-names = "csid0", > + "csid1", > + "csid2", > + "csiphy0", > + "csiphy1", > + "csiphy2", > + "vfe0", > + "vfe1", > + "vfe_lite"; > + > + clocks = <&gcc GCC_CAMERA_AHB_CLK>, > + <&gcc GCC_CAMERA_AXI_CLK>, > + <&camcc CAM_CC_SOC_AHB_CLK>, > + <&camcc CAM_CC_CAMNOC_AXI_CLK>, > + <&camcc CAM_CC_CPAS_AHB_CLK>, > + <&camcc CAM_CC_IFE_0_CSID_CLK>, > + <&camcc CAM_CC_IFE_1_CSID_CLK>, > + <&camcc CAM_CC_IFE_LITE_CSID_CLK>, > + <&camcc CAM_CC_CSIPHY0_CLK>, > + <&camcc CAM_CC_CSI0PHYTIMER_CLK>, > + <&camcc CAM_CC_CSIPHY1_CLK>, > + <&camcc CAM_CC_CSI1PHYTIMER_CLK>, > + <&camcc CAM_CC_CSIPHY2_CLK>, > + <&camcc CAM_CC_CSI2PHYTIMER_CLK>, > + <&camcc CAM_CC_IFE_0_AXI_CLK>, > + <&camcc CAM_CC_IFE_0_CLK>, > + <&camcc CAM_CC_IFE_0_CPHY_RX_CLK>, > + <&camcc CAM_CC_IFE_1_AXI_CLK>, > + <&camcc CAM_CC_IFE_1_CLK>, > + <&camcc CAM_CC_IFE_1_CPHY_RX_CLK>, > + <&camcc CAM_CC_IFE_LITE_CLK>, > + <&camcc CAM_CC_IFE_LITE_CPHY_RX_CLK>; > + clock-names = "gcc_camera_ahb", > + "gcc_camera_axi", > + "soc_ahb", > + "camnoc_axi", > + "cpas_ahb", > + "csi0", > + "csi1", > + "csi2", > + "csiphy0", > + "csiphy0_timer", > + "csiphy1", > + "csiphy1_timer", > + "csiphy2", > + "csiphy2_timer", > + "vfe0_axi", > + "vfe0", > + "vfe0_cphy_rx", > + "vfe1_axi", > + "vfe1", > + "vfe1_cphy_rx", > + "vfe_lite", > + "vfe_lite_cphy_rx"; > + > + iommus = <&apps_smmu 0x808 0x0>, > + <&apps_smmu 0x810 0x8>, > + <&apps_smmu 0xc08 0x0>, > + <&apps_smmu 0xc10 0x8>; > + > + power-domains = <&camcc IFE_0_GDSC>, > + <&camcc IFE_1_GDSC>, > + <&camcc TITAN_TOP_GDSC>; > + power-domain-names = "ife0", > + "ife1", > + "top"; > + > + vdda-phy-supply = <&vreg_l1a_1p225>; > + vdda-pll-supply = <&vreg_l8a_1p8>; > + > + ports { > + #address-cells = <1>; > + #size-cells = <0>; > + > + port@0 { > + reg = <0>; > + > + csiphy_ep0: endpoint { > + clock-lanes = <7>; > + data-lanes = <0 1 2 3>; > + remote-endpoint = <&front_sensor_ep>; > + }; > + }; > + }; > + }; > + }; -- Best wishes, Vladimir
On 11/10/2024 08:14, Vladimir Zapolskiy wrote: > > Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and > qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe > from now on > it should be assumed that all subsequently added CAMSS IP descriptions > to follow > the same established policy. My preference is sort by address not sort by name => we sort the device nodes themselves by address so it seems more consistent to sort by address inside of the devices too. Which means sorting reg by address and irq too. --- bod
On Fri, Oct 11, 2024 at 10:14:49AM +0300, Vladimir Zapolskiy wrote: > > + soc { > > + #address-cells = <2>; > > + #size-cells = <2>; > > + > > + camss@ac65000 { > > + compatible = "qcom,sdm670-camss"; > > + > > + reg = <0 0x0acb3000 0 0x1000>, > > This is immediately wrong, unit address shall be the same as the address of the > first value of reg property. > > I still object to the sorting order of reg values dictated by reg-names property. > > There are a few recently added CAMSS device tree binding descriptions, where > reg values are sorted by address values without a connection to another property > values, and I believe this is the correct way to go. > > Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and > qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe from now on > it should be assumed that all subsequently added CAMSS IP descriptions to follow > the same established policy. Heh, sc8280xp introduced entirely different sorting also in interrupt-names. Just look at interrupts of sm8250 and sc8280xp. Luckily clocks are still keeping style. Can you start keeping consistency? All bindings from the same family of devices, especially if they share something, should have similar order in lists. How do you imagine writing drivers and request items by order (not by name) if the order is different in each flavor? And such change to randomness in style - sometimes reg sorted that way, sometimes other, interrupts sometimes like this or like that - should not come from Linaro. Best regards, Krzysztof
On Fri, Oct 11, 2024 at 09:31:06AM +0100, Bryan O'Donoghue wrote: > On 11/10/2024 08:14, Vladimir Zapolskiy wrote: > > > > Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and > > qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe > > from now on > > it should be assumed that all subsequently added CAMSS IP descriptions > > to follow > > the same established policy. > > My preference is sort by address not sort by name => we sort the device > nodes themselves by address so it seems more consistent to sort by address > inside of the devices too. Strictly speaking, the values of addresses are unknown to the binding, so you can't sort by address. However, if something is truly a single block, then the offsets are probably fixed in order by offset makes sense. But when a block is changed, any rule on sorting may go out the window since we add new regions on the end. This one in particular I have to wonder why csiphy is not a separate node. > > Which means sorting reg by address and irq too. IRQs make little sense to sort IMO. Rob
On 11/10/2024 15:41, Rob Herring wrote: > But when a block is changed, any rule on sorting may go out > the window since we add new regions on the end. Ah I see, I didn't TBH know that. > This one in particular I have to wonder why csiphy is not a separate > node. Hmm, to be honest with you Rob, even though I realise I'm digging myself into a hole of yet more work, yes - we should probably structure camss along the lines of mdss which has separate nodes for DSI phys. -> mdss: display-subsystem@ae00000{} We have 4 SoCs "in flight" at the moment. sdm670 and sc7280 don't require any real driver update to facilitate. sm8550 and x1e80100 do require new VFE, CSID etc. We've been debating how to model the regulators for the CSIPHYs too which are on rails supplied by PMICs external to the SoC. I'm congniscient of the fact 670, 7280 and to an extent sm8550 have been on the list for quite some time, so I'd rather not push 670 and 7280 to have to wait for a whole new way of doing CSIPHYs - especially because these are old hardware with little to no driver change required to support. OTOH x1e80100 and sm8550 already need development work so we can do the CSIPHY transition there. --- bod
Hi Krzysztof, On 10/11/24 17:29, Krzysztof Kozlowski wrote: > On Fri, Oct 11, 2024 at 10:14:49AM +0300, Vladimir Zapolskiy wrote: >>> + soc { >>> + #address-cells = <2>; >>> + #size-cells = <2>; >>> + >>> + camss@ac65000 { >>> + compatible = "qcom,sdm670-camss"; >>> + >>> + reg = <0 0x0acb3000 0 0x1000>, >> >> This is immediately wrong, unit address shall be the same as the address of the >> first value of reg property. >> >> I still object to the sorting order of reg values dictated by reg-names property. >> >> There are a few recently added CAMSS device tree binding descriptions, where >> reg values are sorted by address values without a connection to another property >> values, and I believe this is the correct way to go. >> >> Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and >> qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe from now on >> it should be assumed that all subsequently added CAMSS IP descriptions to follow >> the same established policy. > > Heh, sc8280xp introduced entirely different sorting also in interrupt-names. > > Just look at interrupts of sm8250 and sc8280xp. Luckily clocks are still > keeping style. > > Can you start keeping consistency? All bindings from the same family of > devices, especially if they share something, should have similar order > in lists. > > How do you imagine writing drivers and request items by order (not by > name) if the order is different in each flavor? I don't see a problem here, and I don't remember any reports about this kind of problem while adding CAMSS support in the driver to new platforms. While the problem of improper CAMSS unit address appears again and again, the focus shall be on removing a chance to make a commin mistake here. As I've already said above, device tree bindings of CAMSS in two most recently added platforms sm8250 and sc8280xp follow the numerical order of addresses from reg value. This becomes the policy. Sorting lists of interrupts or clocks by numerical values makes no sense, thus the argument of *-names sorting becomes valid here. For clarity, reg property is very special, also a snippet of its value goes as a unit address. -- Best wishes, Vladimir
Hi Rob. On 10/11/24 17:41, Rob Herring wrote: > On Fri, Oct 11, 2024 at 09:31:06AM +0100, Bryan O'Donoghue wrote: >> On 11/10/2024 08:14, Vladimir Zapolskiy wrote: >>> >>> Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and >>> qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe >>> from now on >>> it should be assumed that all subsequently added CAMSS IP descriptions >>> to follow >>> the same established policy. >> >> My preference is sort by address not sort by name => we sort the device >> nodes themselves by address so it seems more consistent to sort by address >> inside of the devices too. > > Strictly speaking, the values of addresses are unknown to the binding, > so you can't sort by address. However, if something is truly a single > block, then the offsets are probably fixed in order by offset makes > sense. But when a block is changed, any rule on sorting may go out > the window since we add new regions on the end. Exactly, and this is an argument why the sorting is a subject to a device driver policy, kind of any sorting order is equally bad. Sorting 'reg' values by addresses helps to avoid a notorious problem with unit addresses. > This one in particular I have to wonder why csiphy is not a separate > node. There were dicussions about it in the past, and kind of enforced outcome of the discussions is to keep all CAMSS IP components together under one huge plain device tree node. I personally dislike this approach, but obedience is the way to get things merged. >> >> Which means sorting reg by address and irq too. > > IRQs make little sense to sort IMO. For all non-reg properties with a present *-names property the sorting order should be done by *-names property. Only 'reg' is very special. -- Best wishes, Vladimir
On 30/10/2024 15:06, Vladimir Zapolskiy wrote: > Hi Krzysztof, > > On 10/11/24 17:29, Krzysztof Kozlowski wrote: >> On Fri, Oct 11, 2024 at 10:14:49AM +0300, Vladimir Zapolskiy wrote: >>>> + soc { >>>> + #address-cells = <2>; >>>> + #size-cells = <2>; >>>> + >>>> + camss@ac65000 { >>>> + compatible = "qcom,sdm670-camss"; >>>> + >>>> + reg = <0 0x0acb3000 0 0x1000>, >>> >>> This is immediately wrong, unit address shall be the same as the address of the >>> first value of reg property. >>> >>> I still object to the sorting order of reg values dictated by reg-names property. >>> >>> There are a few recently added CAMSS device tree binding descriptions, where >>> reg values are sorted by address values without a connection to another property >>> values, and I believe this is the correct way to go. >>> >>> Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and >>> qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe from now on >>> it should be assumed that all subsequently added CAMSS IP descriptions to follow >>> the same established policy. >> >> Heh, sc8280xp introduced entirely different sorting also in interrupt-names. >> >> Just look at interrupts of sm8250 and sc8280xp. Luckily clocks are still >> keeping style. >> >> Can you start keeping consistency? All bindings from the same family of >> devices, especially if they share something, should have similar order >> in lists. >> >> How do you imagine writing drivers and request items by order (not by >> name) if the order is different in each flavor? > > I don't see a problem here, and I don't remember any reports about this > kind of problem while adding CAMSS support in the driver to new platforms. And I see problem, would create enormous probe code to handle different variants for clock[0] and then clock[1], etc. > > While the problem of improper CAMSS unit address appears again and again, > the focus shall be on removing a chance to make a commin mistake here. This is not a problem. Tools already point it out. Order of reg-names also does not affect that, you can put fake unit address regardless of the order of reg-names items. > > As I've already said above, device tree bindings of CAMSS in two most > recently added platforms sm8250 and sc8280xp follow the numerical order > of addresses from reg value. This becomes the policy. > > Sorting lists of interrupts or clocks by numerical values makes no sense, > thus the argument of *-names sorting becomes valid here. For clarity, reg There is no such argument, no such coding style. > property is very special, also a snippet of its value goes as a unit > address. And order of items does not matter for above "specialness of reg". Best regards, Krzysztof
On Wed, Oct 30, 2024 at 9:20 AM Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org> wrote: > > Hi Rob. > > On 10/11/24 17:41, Rob Herring wrote: > > On Fri, Oct 11, 2024 at 09:31:06AM +0100, Bryan O'Donoghue wrote: > >> On 11/10/2024 08:14, Vladimir Zapolskiy wrote: > >>> > >>> Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and > >>> qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe > >>> from now on > >>> it should be assumed that all subsequently added CAMSS IP descriptions > >>> to follow > >>> the same established policy. > >> > >> My preference is sort by address not sort by name => we sort the device > >> nodes themselves by address so it seems more consistent to sort by address > >> inside of the devices too. > > > > Strictly speaking, the values of addresses are unknown to the binding, > > so you can't sort by address. However, if something is truly a single > > block, then the offsets are probably fixed in order by offset makes > > sense. But when a block is changed, any rule on sorting may go out > > the window since we add new regions on the end. > > Exactly, and this is an argument why the sorting is a subject to a device > driver policy, kind of any sorting order is equally bad. Sorting 'reg' > values by addresses helps to avoid a notorious problem with unit addresses. What notorious problem? > > > This one in particular I have to wonder why csiphy is not a separate > > node. > > There were dicussions about it in the past, and kind of enforced outcome of > the discussions is to keep all CAMSS IP components together under one huge > plain device tree node. I personally dislike this approach, but obedience > is the way to get things merged. Who are you saying would be in the way to get things merged? DT maintainers? I feel certain I would have pushed for separate blocks, but I'll defer to people that know the h/w. I can't learn the details of everyone's h/w. If they get it wrong, it's their problem not mine. > >> Which means sorting reg by address and irq too. > > > > IRQs make little sense to sort IMO. > > For all non-reg properties with a present *-names property the sorting > order should be done by *-names property. Only 'reg' is very special. No. If you had 'main' and 'error', I'd put 'main' first. If they are somewhat equal (e.g. rx, tx), then sure, sort them however you like (assuming no existing binding). The only real rules here are how new entries should be added (on the end). Otherwise, there is no policy. Rob
On 10/30/24 23:06, Rob Herring wrote: > On Wed, Oct 30, 2024 at 9:20 AM Vladimir Zapolskiy > <vladimir.zapolskiy@linaro.org> wrote: >> >> Hi Rob. >> >> On 10/11/24 17:41, Rob Herring wrote: >>> On Fri, Oct 11, 2024 at 09:31:06AM +0100, Bryan O'Donoghue wrote: >>>> On 11/10/2024 08:14, Vladimir Zapolskiy wrote: >>>>> >>>>> Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and >>>>> qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe >>>>> from now on >>>>> it should be assumed that all subsequently added CAMSS IP descriptions >>>>> to follow >>>>> the same established policy. >>>> >>>> My preference is sort by address not sort by name => we sort the device >>>> nodes themselves by address so it seems more consistent to sort by address >>>> inside of the devices too. >>> >>> Strictly speaking, the values of addresses are unknown to the binding, >>> so you can't sort by address. However, if something is truly a single >>> block, then the offsets are probably fixed in order by offset makes >>> sense. But when a block is changed, any rule on sorting may go out >>> the window since we add new regions on the end. >> >> Exactly, and this is an argument why the sorting is a subject to a device >> driver policy, kind of any sorting order is equally bad. Sorting 'reg' >> values by addresses helps to avoid a notorious problem with unit addresses. > > What notorious problem? > Here the problem I reference to is the problem of an incorrespondence between device tree node unit address and the address of the first value of 'reg' values. Having a sorting by addresses allows to grasp IO ranges easily, and setting device tree node unit addresses to some almost arbitrary chosen value from the middle of IP's IO range is suspicious and confusing in my opinion. >> >>> This one in particular I have to wonder why csiphy is not a separate >>> node. >> >> There were dicussions about it in the past, and kind of enforced outcome of >> the discussions is to keep all CAMSS IP components together under one huge >> plain device tree node. I personally dislike this approach, but obedience >> is the way to get things merged. > > Who are you saying would be in the way to get things merged? DT > maintainers? I feel certain I would have pushed for separate blocks, > but I'll defer to people that know the h/w. I can't learn the details > of everyone's h/w. If they get it wrong, it's their problem not mine. I had this discussion with Qualcomm/CAMSS maintainers long time ago, it may be restarted, if there is a necessity. >>>> Which means sorting reg by address and irq too. >>> >>> IRQs make little sense to sort IMO. >> >> For all non-reg properties with a present *-names property the sorting >> order should be done by *-names property. Only 'reg' is very special. > > No. If you had 'main' and 'error', I'd put 'main' first. If they are > somewhat equal (e.g. rx, tx), then sure, sort them however you like > (assuming no existing binding). The only real rules here are how new > entries should be added (on the end). Otherwise, there is no policy. > Here in the proposed terms the start of an IO region is 'main', while some value in the middle of it (the first one in alphabetical sorting) is too secondary to dictate the device tree node unit address, I believe. -- Best wishes, Vladimir
On 11/10/2024 15:29, Krzysztof Kozlowski wrote: > How do you imagine writing drivers and request items by order (not by > name) if the order is different in each flavor? I don't think I'd be much in favour of relying on declaration order in the dts, favouring names to find resources instead, tbh. The 8250 has regs that sort by address and name in the same order. For 8280xp we preferred sort by address and you're right the interrupt sorting isn't consistent. However the latest applied dts for CAMSS is sort by address/irq not sort by reg-name irq-name. Unless its a NAK from yourself and Rob, that would certainly be my preference for any _new_ additions subsequent. --- bod
On 31/10/2024 16:42, Bryan O'Donoghue wrote: > On 11/10/2024 15:29, Krzysztof Kozlowski wrote: >> How do you imagine writing drivers and request items by order (not by >> name) if the order is different in each flavor? > > I don't think I'd be much in favour of relying on declaration order in > the dts, favouring names to find resources instead, tbh. > > The 8250 has regs that sort by address and name in the same order. For > 8280xp we preferred sort by address and you're right the interrupt > sorting isn't consistent. > > However the latest applied dts for CAMSS is sort by address/irq not sort > by reg-name irq-name. > > Unless its a NAK from yourself and Rob, that would certainly be my > preference for any _new_ additions subsequent. It's not a NAK as long you keep the same order in new bindings, which I think it is not possible. I repeat myself: there is no rule/style that list should be ordered by values, but there is a rule that all devices from the same family should have the same order of items in the list. I don't think it is achievable with your approach - sorting by value. Best regards, Krzysztof
On 01/11/2024 09:17, Krzysztof Kozlowski wrote: > On 31/10/2024 16:42, Bryan O'Donoghue wrote: >> On 11/10/2024 15:29, Krzysztof Kozlowski wrote: >>> How do you imagine writing drivers and request items by order (not by >>> name) if the order is different in each flavor? >> >> I don't think I'd be much in favour of relying on declaration order in >> the dts, favouring names to find resources instead, tbh. >> >> The 8250 has regs that sort by address and name in the same order. For >> 8280xp we preferred sort by address and you're right the interrupt >> sorting isn't consistent. >> >> However the latest applied dts for CAMSS is sort by address/irq not sort >> by reg-name irq-name. >> >> Unless its a NAK from yourself and Rob, that would certainly be my >> preference for any _new_ additions subsequent. > > It's not a NAK as long you keep the same order in new bindings, which I > think it is not possible. I repeat myself: there is no rule/style that > list should be ordered by values, but there is a rule that all devices > from the same family should have the same order of items in the list. I > don't think it is achievable with your approach - sorting by value. Grand. I'm happy enough to sort by IP alpha TBH. --- bod
On 30/10/2024 23:13, Vladimir Zapolskiy wrote: > On 10/30/24 23:06, Rob Herring wrote: >> On Wed, Oct 30, 2024 at 9:20 AM Vladimir Zapolskiy >> <vladimir.zapolskiy@linaro.org> wrote: >>> >>> Hi Rob. >>> >>> On 10/11/24 17:41, Rob Herring wrote: >>>> On Fri, Oct 11, 2024 at 09:31:06AM +0100, Bryan O'Donoghue wrote: >>>>> On 11/10/2024 08:14, Vladimir Zapolskiy wrote: >>>>>> >>>>>> Two most recently added CAMSS IP descriptions (qcom,sm8250-camss.yaml and >>>>>> qcom,sc8280xp-camss.yaml) do implement sorting by reg values, I believe >>>>>> from now on >>>>>> it should be assumed that all subsequently added CAMSS IP descriptions >>>>>> to follow >>>>>> the same established policy. >>>>> >>>>> My preference is sort by address not sort by name => we sort the device >>>>> nodes themselves by address so it seems more consistent to sort by address >>>>> inside of the devices too. >>>> >>>> Strictly speaking, the values of addresses are unknown to the binding, >>>> so you can't sort by address. However, if something is truly a single >>>> block, then the offsets are probably fixed in order by offset makes >>>> sense. But when a block is changed, any rule on sorting may go out >>>> the window since we add new regions on the end. >>> >>> Exactly, and this is an argument why the sorting is a subject to a device >>> driver policy, kind of any sorting order is equally bad. Sorting 'reg' >>> values by addresses helps to avoid a notorious problem with unit addresses. >> >> What notorious problem? >> > > Here the problem I reference to is the problem of an incorrespondence between > device tree node unit address and the address of the first value of 'reg' > values. Sorting by unit address does not solve it. Sorting by anything else does not cause it. You repeat this statement multiple times already. Sorry, it's not true. You can sort by address and still put wrong unit address. Tools address it already, not sorting. > > Having a sorting by addresses allows to grasp IO ranges easily, and setting > device tree node unit addresses to some almost arbitrary chosen value from > the middle of IP's IO range is suspicious and confusing in my opinion. It's not arbitrary. It's the main, control address space often. At least it should be. Your sorting rule enforces putting irrelevant address as unit address, imagine: isp@1000 { reg = <0x1000 0x4>, <0x2000, 0x1000>; reg-names = "reset-control-block-one-unimportant-register", "main-control-block-with-99%-of-registers"; Sorry, this is just wrong coding style. Best regards, Krzysztof
On 01/11/2024 10:36, Bryan O'Donoghue wrote: > On 01/11/2024 09:17, Krzysztof Kozlowski wrote: >> On 31/10/2024 16:42, Bryan O'Donoghue wrote: >>> On 11/10/2024 15:29, Krzysztof Kozlowski wrote: >>>> How do you imagine writing drivers and request items by order (not by >>>> name) if the order is different in each flavor? >>> >>> I don't think I'd be much in favour of relying on declaration order in >>> the dts, favouring names to find resources instead, tbh. >>> >>> The 8250 has regs that sort by address and name in the same order. For >>> 8280xp we preferred sort by address and you're right the interrupt >>> sorting isn't consistent. >>> >>> However the latest applied dts for CAMSS is sort by address/irq not sort >>> by reg-name irq-name. >>> >>> Unless its a NAK from yourself and Rob, that would certainly be my >>> preference for any _new_ additions subsequent. >> >> It's not a NAK as long you keep the same order in new bindings, which I >> think it is not possible. I repeat myself: there is no rule/style that >> list should be ordered by values, but there is a rule that all devices >> from the same family should have the same order of items in the list. I >> don't think it is achievable with your approach - sorting by value. > > Grand. > > I'm happy enough to sort by IP alpha TBH. I actually missed that Rob responded here clarifying his point of view. Two of DT maintainers expressed their dislike for such coding style of sorting by value. Regardless whether it is helping or not, it is not even possible to implement. Binding does not know the addresses and one could add a binding without DTS and drivers for some other user or future implementation. It's not even possible to implement that coding style. Best regards, Krzysztof
diff --git a/Documentation/devicetree/bindings/media/qcom,sdm670-camss.yaml b/Documentation/devicetree/bindings/media/qcom,sdm670-camss.yaml new file mode 100644 index 000000000000..670502532d28 --- /dev/null +++ b/Documentation/devicetree/bindings/media/qcom,sdm670-camss.yaml @@ -0,0 +1,318 @@ +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +%YAML 1.2 +--- +$id: http://devicetree.org/schemas/media/qcom,sdm670-camss.yaml# +$schema: http://devicetree.org/meta-schemas/core.yaml# + +title: Qualcomm SDM670 Camera Subsystem (CAMSS) + +maintainers: + - Richard Acayan <mailingradian@gmail.com> + +description: + The CAMSS IP is a CSI decoder and ISP present on Qualcomm platforms. + +properties: + compatible: + const: qcom,sdm670-camss + + reg: + maxItems: 9 + + reg-names: + items: + - const: csid0 + - const: csid1 + - const: csid2 + - const: csiphy0 + - const: csiphy1 + - const: csiphy2 + - const: vfe0 + - const: vfe1 + - const: vfe_lite + + clocks: + maxItems: 22 + + clock-names: + items: + - const: gcc_camera_ahb + - const: gcc_camera_axi + - const: soc_ahb + - const: camnoc_axi + - const: cpas_ahb + - const: csi0 + - const: csi1 + - const: csi2 + - const: csiphy0 + - const: csiphy0_timer + - const: csiphy1 + - const: csiphy1_timer + - const: csiphy2 + - const: csiphy2_timer + - const: vfe0_axi + - const: vfe0 + - const: vfe0_cphy_rx + - const: vfe1_axi + - const: vfe1 + - const: vfe1_cphy_rx + - const: vfe_lite + - const: vfe_lite_cphy_rx + + interrupts: + maxItems: 9 + + interrupt-names: + items: + - const: csid0 + - const: csid1 + - const: csid2 + - const: csiphy0 + - const: csiphy1 + - const: csiphy2 + - const: vfe0 + - const: vfe1 + - const: vfe_lite + + iommus: + maxItems: 4 + + power-domains: + items: + - description: IFE0 GDSC - Image Front End, Global Distributed Switch Controller. + - description: IFE1 GDSC - Image Front End, Global Distributed Switch Controller. + - description: Titan Top GDSC - Titan ISP Block, Global Distributed Switch Controller. + + power-domain-names: + items: + - const: ife0 + - const: ife1 + - const: top + + ports: + $ref: /schemas/graph.yaml#/properties/ports + + description: + CSI input ports. + + properties: + port@0: + $ref: /schemas/graph.yaml#/$defs/port-base + unevaluatedProperties: false + description: + Input port for receiving CSI data from CSIPHY0. + + properties: + endpoint: + $ref: video-interfaces.yaml# + unevaluatedProperties: false + + properties: + clock-lanes: + maxItems: 1 + + data-lanes: + minItems: 1 + maxItems: 4 + + required: + - clock-lanes + - data-lanes + + port@1: + $ref: /schemas/graph.yaml#/$defs/port-base + unevaluatedProperties: false + description: + Input port for receiving CSI data from CSIPHY1. + + properties: + endpoint: + $ref: video-interfaces.yaml# + unevaluatedProperties: false + + properties: + clock-lanes: + maxItems: 1 + + data-lanes: + minItems: 1 + maxItems: 4 + + required: + - clock-lanes + - data-lanes + + port@2: + $ref: /schemas/graph.yaml#/$defs/port-base + unevaluatedProperties: false + description: + Input port for receiving CSI data from CSIPHY2. + + properties: + endpoint: + $ref: video-interfaces.yaml# + unevaluatedProperties: false + + properties: + clock-lanes: + maxItems: 1 + + data-lanes: + minItems: 1 + maxItems: 4 + + required: + - clock-lanes + - data-lanes + + vdda-phy-supply: + description: + Phandle to a regulator supply to PHY core block. + + vdda-pll-supply: + description: + Phandle to 1.8V regulator supply to PHY refclk pll block. + +required: + - reg + - reg-names + - clock-names + - clocks + - compatible + - interrupts + - interrupt-names + - iommus + - power-domains + - power-domain-names + - vdda-phy-supply + - vdda-pll-supply + +additionalProperties: false + +examples: + - | + #include <dt-bindings/clock/qcom,camcc-sdm845.h> + #include <dt-bindings/clock/qcom,gcc-sdm845.h> + #include <dt-bindings/interrupt-controller/arm-gic.h> + + soc { + #address-cells = <2>; + #size-cells = <2>; + + camss@ac65000 { + compatible = "qcom,sdm670-camss"; + + reg = <0 0x0acb3000 0 0x1000>, + <0 0x0acba000 0 0x1000>, + <0 0x0acc8000 0 0x1000>, + <0 0x0ac65000 0 0x1000>, + <0 0x0ac66000 0 0x1000>, + <0 0x0ac67000 0 0x1000>, + <0 0x0acaf000 0 0x4000>, + <0 0x0acb6000 0 0x4000>, + <0 0x0acc4000 0 0x4000>; + reg-names = "csid0", + "csid1", + "csid2", + "csiphy0", + "csiphy1", + "csiphy2", + "vfe0", + "vfe1", + "vfe_lite"; + + interrupts = <GIC_SPI 464 IRQ_TYPE_EDGE_RISING>, + <GIC_SPI 466 IRQ_TYPE_EDGE_RISING>, + <GIC_SPI 468 IRQ_TYPE_EDGE_RISING>, + <GIC_SPI 477 IRQ_TYPE_EDGE_RISING>, + <GIC_SPI 478 IRQ_TYPE_EDGE_RISING>, + <GIC_SPI 479 IRQ_TYPE_EDGE_RISING>, + <GIC_SPI 465 IRQ_TYPE_EDGE_RISING>, + <GIC_SPI 467 IRQ_TYPE_EDGE_RISING>, + <GIC_SPI 469 IRQ_TYPE_EDGE_RISING>; + interrupt-names = "csid0", + "csid1", + "csid2", + "csiphy0", + "csiphy1", + "csiphy2", + "vfe0", + "vfe1", + "vfe_lite"; + + clocks = <&gcc GCC_CAMERA_AHB_CLK>, + <&gcc GCC_CAMERA_AXI_CLK>, + <&camcc CAM_CC_SOC_AHB_CLK>, + <&camcc CAM_CC_CAMNOC_AXI_CLK>, + <&camcc CAM_CC_CPAS_AHB_CLK>, + <&camcc CAM_CC_IFE_0_CSID_CLK>, + <&camcc CAM_CC_IFE_1_CSID_CLK>, + <&camcc CAM_CC_IFE_LITE_CSID_CLK>, + <&camcc CAM_CC_CSIPHY0_CLK>, + <&camcc CAM_CC_CSI0PHYTIMER_CLK>, + <&camcc CAM_CC_CSIPHY1_CLK>, + <&camcc CAM_CC_CSI1PHYTIMER_CLK>, + <&camcc CAM_CC_CSIPHY2_CLK>, + <&camcc CAM_CC_CSI2PHYTIMER_CLK>, + <&camcc CAM_CC_IFE_0_AXI_CLK>, + <&camcc CAM_CC_IFE_0_CLK>, + <&camcc CAM_CC_IFE_0_CPHY_RX_CLK>, + <&camcc CAM_CC_IFE_1_AXI_CLK>, + <&camcc CAM_CC_IFE_1_CLK>, + <&camcc CAM_CC_IFE_1_CPHY_RX_CLK>, + <&camcc CAM_CC_IFE_LITE_CLK>, + <&camcc CAM_CC_IFE_LITE_CPHY_RX_CLK>; + clock-names = "gcc_camera_ahb", + "gcc_camera_axi", + "soc_ahb", + "camnoc_axi", + "cpas_ahb", + "csi0", + "csi1", + "csi2", + "csiphy0", + "csiphy0_timer", + "csiphy1", + "csiphy1_timer", + "csiphy2", + "csiphy2_timer", + "vfe0_axi", + "vfe0", + "vfe0_cphy_rx", + "vfe1_axi", + "vfe1", + "vfe1_cphy_rx", + "vfe_lite", + "vfe_lite_cphy_rx"; + + iommus = <&apps_smmu 0x808 0x0>, + <&apps_smmu 0x810 0x8>, + <&apps_smmu 0xc08 0x0>, + <&apps_smmu 0xc10 0x8>; + + power-domains = <&camcc IFE_0_GDSC>, + <&camcc IFE_1_GDSC>, + <&camcc TITAN_TOP_GDSC>; + power-domain-names = "ife0", + "ife1", + "top"; + + vdda-phy-supply = <&vreg_l1a_1p225>; + vdda-pll-supply = <&vreg_l8a_1p8>; + + ports { + #address-cells = <1>; + #size-cells = <0>; + + port@0 { + reg = <0>; + + csiphy_ep0: endpoint { + clock-lanes = <7>; + data-lanes = <0 1 2 3>; + remote-endpoint = <&front_sensor_ep>; + }; + }; + }; + }; + };
As found in the Pixel 3a, the Snapdragon 670 has a camera subsystem with 3 CSIDs and 3 VFEs (including 1 VFE lite). Add this camera subsystem to the bindings. Adapted from SC8280XP camera subsystem. Signed-off-by: Richard Acayan <mailingradian@gmail.com> --- .../bindings/media/qcom,sdm670-camss.yaml | 318 ++++++++++++++++++ 1 file changed, 318 insertions(+) create mode 100644 Documentation/devicetree/bindings/media/qcom,sdm670-camss.yaml