| Message ID | 20210601024408.24485-2-jianjun.wang@mediatek.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Delegated to: | Lorenzo Pieralisi |
| Headers | show |
| Series | PCI: mediatek-gen3: Add controller support for MT8195 | expand |
Hi, On Tue, Jun 1, 2021 at 10:50 AM Jianjun Wang <jianjun.wang@mediatek.com> wrote: > > MT8195 is an ARM platform SoC which has the same PCIe IP with MT8192. Based on what I'm seeing internally, there seems to be some inconsistency across the MediaTek platform on whether new compatible strings should be introduced for "fully compatible" IP blocks. If this hardware block in MT8195 is "the same" as the one in MT8192, do we really need the new compatible string? Are there any concerns? Thanks ChenYu > Signed-off-by: Jianjun Wang <jianjun.wang@mediatek.com> > --- > Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml | 4 +++- > 1 file changed, 3 insertions(+), 1 deletion(-) > > diff --git a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > index e7b1f9892da4..d5e4a3e63d97 100644 > --- a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > +++ b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > @@ -48,7 +48,9 @@ allOf: > > properties: > compatible: > - const: mediatek,mt8192-pcie > + oneOf: > + - const: mediatek,mt8192-pcie > + - const: mediatek,mt8195-pcie > > reg: > maxItems: 1 > -- > 2.18.0 > _______________________________________________ > Linux-mediatek mailing list > Linux-mediatek@lists.infradead.org > http://lists.infradead.org/mailman/listinfo/linux-mediatek
On Tue, 2021-06-01 at 11:53 +0800, Chen-Yu Tsai wrote: > Hi, > > On Tue, Jun 1, 2021 at 10:50 AM Jianjun Wang <jianjun.wang@mediatek.com> wrote: > > > > MT8195 is an ARM platform SoC which has the same PCIe IP with MT8192. > > Based on what I'm seeing internally, there seems to be some inconsistency > across the MediaTek platform on whether new compatible strings should be > introduced for "fully compatible" IP blocks. > > If this hardware block in MT8195 is "the same" as the one in MT8192, do we > really need the new compatible string? Are there any concerns? Hi Chen-Yu, It's ok to reuse the compatible string with MT8192, but I think this will be easier to find which platforms this driver is compatible with, especially when we have more and more platforms in the future. Thanks. > > > Thanks > ChenYu > > > > Signed-off-by: Jianjun Wang <jianjun.wang@mediatek.com> > > --- > > Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml | 4 +++- > > 1 file changed, 3 insertions(+), 1 deletion(-) > > > > diff --git a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > > index e7b1f9892da4..d5e4a3e63d97 100644 > > --- a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > > +++ b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > > @@ -48,7 +48,9 @@ allOf: > > > > properties: > > compatible: > > - const: mediatek,mt8192-pcie > > + oneOf: > > + - const: mediatek,mt8192-pcie > > + - const: mediatek,mt8195-pcie > > > > reg: > > maxItems: 1 > > -- > > 2.18.0 > > _______________________________________________ > > Linux-mediatek mailing list > > Linux-mediatek@lists.infradead.org > > http://lists.infradead.org/mailman/listinfo/linux-mediatek
Hi, On Tue, Jun 1, 2021 at 1:50 PM Jianjun Wang <jianjun.wang@mediatek.com> wrote: > > On Tue, 2021-06-01 at 11:53 +0800, Chen-Yu Tsai wrote: > > Hi, > > > > On Tue, Jun 1, 2021 at 10:50 AM Jianjun Wang <jianjun.wang@mediatek.com> wrote: > > > > > > MT8195 is an ARM platform SoC which has the same PCIe IP with MT8192. > > > > Based on what I'm seeing internally, there seems to be some inconsistency > > across the MediaTek platform on whether new compatible strings should be > > introduced for "fully compatible" IP blocks. > > > > If this hardware block in MT8195 is "the same" as the one in MT8192, do we > > really need the new compatible string? Are there any concerns? > > Hi Chen-Yu, > > It's ok to reuse the compatible string with MT8192, but I think this > will be easier to find which platforms this driver is compatible with, > especially when we have more and more platforms in the future. If it's just for informational purposes, then having the MT8192 compatible as a fallback would work, and we wouldn't need to make changes to the driver. This works better especially if we have to support multiple operating systems that use device tree. So we would want "mediatek,mt8195-pcie", "mediatek,mt8192-pcie" and "mediatek,mt8192-pcie" be the valid options. Personally I'm not seeing enough value to justify adding the compatible string just for informational purposes though. One could easily discern which hardware is used by looking at the device tree. Regards ChenYu > Thanks. > > > > > > Thanks > > ChenYu > > > > > > > Signed-off-by: Jianjun Wang <jianjun.wang@mediatek.com> > > > --- > > > Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml | 4 +++- > > > 1 file changed, 3 insertions(+), 1 deletion(-) > > > > > > diff --git a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > > > index e7b1f9892da4..d5e4a3e63d97 100644 > > > --- a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > > > +++ b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > > > @@ -48,7 +48,9 @@ allOf: > > > > > > properties: > > > compatible: > > > - const: mediatek,mt8192-pcie > > > + oneOf: > > > + - const: mediatek,mt8192-pcie > > > + - const: mediatek,mt8195-pcie > > > > > > reg: > > > maxItems: 1 > > > -- > > > 2.18.0 > > > _______________________________________________ > > > Linux-mediatek mailing list > > > Linux-mediatek@lists.infradead.org > > > http://lists.infradead.org/mailman/listinfo/linux-mediatek >
On 01/06/2021 08:07, Chen-Yu Tsai wrote: > Hi, > > On Tue, Jun 1, 2021 at 1:50 PM Jianjun Wang <jianjun.wang@mediatek.com> wrote: >> >> On Tue, 2021-06-01 at 11:53 +0800, Chen-Yu Tsai wrote: >>> Hi, >>> >>> On Tue, Jun 1, 2021 at 10:50 AM Jianjun Wang <jianjun.wang@mediatek.com> wrote: >>>> >>>> MT8195 is an ARM platform SoC which has the same PCIe IP with MT8192. >>> >>> Based on what I'm seeing internally, there seems to be some inconsistency >>> across the MediaTek platform on whether new compatible strings should be >>> introduced for "fully compatible" IP blocks. >>> >>> If this hardware block in MT8195 is "the same" as the one in MT8192, do we >>> really need the new compatible string? Are there any concerns? >> >> Hi Chen-Yu, >> >> It's ok to reuse the compatible string with MT8192, but I think this >> will be easier to find which platforms this driver is compatible with, >> especially when we have more and more platforms in the future. > > If it's just for informational purposes, then having the MT8192 compatible > as a fallback would work, and we wouldn't need to make changes to the driver. > This works better especially if we have to support multiple operating systems > that use device tree. > > So we would want > > "mediatek,mt8195-pcie", "mediatek,mt8192-pcie" > > and > > "mediatek,mt8192-pcie" > > be the valid options. > > Personally I'm not seeing enough value to justify adding the compatible string > just for informational purposes though. One could easily discern which hardware > is used by looking at the device tree. > I agree, if no differences between the two chips are known, adding a binding withe new compatible and a fallback is a good thing. If we later on realize that mt8195 PCI block has differences, we can add the matching to the driver. Regards, Matthias > > Regards > ChenYu > > >> Thanks. >>> >>> >>> Thanks >>> ChenYu >>> >>> >>>> Signed-off-by: Jianjun Wang <jianjun.wang@mediatek.com> >>>> --- >>>> Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml | 4 +++- >>>> 1 file changed, 3 insertions(+), 1 deletion(-) >>>> >>>> diff --git a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml >>>> index e7b1f9892da4..d5e4a3e63d97 100644 >>>> --- a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml >>>> +++ b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml >>>> @@ -48,7 +48,9 @@ allOf: >>>> >>>> properties: >>>> compatible: >>>> - const: mediatek,mt8192-pcie >>>> + oneOf: >>>> + - const: mediatek,mt8192-pcie >>>> + - const: mediatek,mt8195-pcie >>>> >>>> reg: >>>> maxItems: 1 >>>> -- >>>> 2.18.0 >>>> _______________________________________________ >>>> Linux-mediatek mailing list >>>> Linux-mediatek@lists.infradead.org >>>> http://lists.infradead.org/mailman/listinfo/linux-mediatek >>
On Wed, Jun 02, 2021 at 01:33:07PM +0200, Matthias Brugger wrote: > > > On 01/06/2021 08:07, Chen-Yu Tsai wrote: > > Hi, > > > > On Tue, Jun 1, 2021 at 1:50 PM Jianjun Wang <jianjun.wang@mediatek.com> wrote: > >> > >> On Tue, 2021-06-01 at 11:53 +0800, Chen-Yu Tsai wrote: > >>> Hi, > >>> > >>> On Tue, Jun 1, 2021 at 10:50 AM Jianjun Wang <jianjun.wang@mediatek.com> wrote: > >>>> > >>>> MT8195 is an ARM platform SoC which has the same PCIe IP with MT8192. > >>> > >>> Based on what I'm seeing internally, there seems to be some inconsistency > >>> across the MediaTek platform on whether new compatible strings should be > >>> introduced for "fully compatible" IP blocks. > >>> > >>> If this hardware block in MT8195 is "the same" as the one in MT8192, do we > >>> really need the new compatible string? Are there any concerns? > >> > >> Hi Chen-Yu, > >> > >> It's ok to reuse the compatible string with MT8192, but I think this > >> will be easier to find which platforms this driver is compatible with, > >> especially when we have more and more platforms in the future. > > > > If it's just for informational purposes, then having the MT8192 compatible > > as a fallback would work, and we wouldn't need to make changes to the driver. > > This works better especially if we have to support multiple operating systems > > that use device tree. > > > > So we would want > > > > "mediatek,mt8195-pcie", "mediatek,mt8192-pcie" > > > > and > > > > "mediatek,mt8192-pcie" > > > > be the valid options. > > > > Personally I'm not seeing enough value to justify adding the compatible string > > just for informational purposes though. One could easily discern which hardware > > is used by looking at the device tree. > > > > I agree, if no differences between the two chips are known, adding a > binding withe new compatible and a fallback is a good thing. If we > later on realize that mt8195 PCI block has differences, we can add the > matching to the driver. So this series can be dropped, right ? Thanks, Lorenzo > Regards, > Matthias > > > > > Regards > > ChenYu > > > > > >> Thanks. > >>> > >>> > >>> Thanks > >>> ChenYu > >>> > >>> > >>>> Signed-off-by: Jianjun Wang <jianjun.wang@mediatek.com> > >>>> --- > >>>> Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml | 4 +++- > >>>> 1 file changed, 3 insertions(+), 1 deletion(-) > >>>> > >>>> diff --git a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > >>>> index e7b1f9892da4..d5e4a3e63d97 100644 > >>>> --- a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > >>>> +++ b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > >>>> @@ -48,7 +48,9 @@ allOf: > >>>> > >>>> properties: > >>>> compatible: > >>>> - const: mediatek,mt8192-pcie > >>>> + oneOf: > >>>> + - const: mediatek,mt8192-pcie > >>>> + - const: mediatek,mt8195-pcie > >>>> > >>>> reg: > >>>> maxItems: 1 > >>>> -- > >>>> 2.18.0 > >>>> _______________________________________________ > >>>> Linux-mediatek mailing list > >>>> Linux-mediatek@lists.infradead.org > >>>> http://lists.infradead.org/mailman/listinfo/linux-mediatek > >>
On Tue, 2021-06-22 at 12:05 +0100, Lorenzo Pieralisi wrote: > On Wed, Jun 02, 2021 at 01:33:07PM +0200, Matthias Brugger wrote: > > > > > > On 01/06/2021 08:07, Chen-Yu Tsai wrote: > > > Hi, > > > > > > On Tue, Jun 1, 2021 at 1:50 PM Jianjun Wang <jianjun.wang@mediatek.com> wrote: > > >> > > >> On Tue, 2021-06-01 at 11:53 +0800, Chen-Yu Tsai wrote: > > >>> Hi, > > >>> > > >>> On Tue, Jun 1, 2021 at 10:50 AM Jianjun Wang <jianjun.wang@mediatek.com> wrote: > > >>>> > > >>>> MT8195 is an ARM platform SoC which has the same PCIe IP with MT8192. > > >>> > > >>> Based on what I'm seeing internally, there seems to be some inconsistency > > >>> across the MediaTek platform on whether new compatible strings should be > > >>> introduced for "fully compatible" IP blocks. > > >>> > > >>> If this hardware block in MT8195 is "the same" as the one in MT8192, do we > > >>> really need the new compatible string? Are there any concerns? > > >> > > >> Hi Chen-Yu, > > >> > > >> It's ok to reuse the compatible string with MT8192, but I think this > > >> will be easier to find which platforms this driver is compatible with, > > >> especially when we have more and more platforms in the future. > > > > > > If it's just for informational purposes, then having the MT8192 compatible > > > as a fallback would work, and we wouldn't need to make changes to the driver. > > > This works better especially if we have to support multiple operating systems > > > that use device tree. > > > > > > So we would want > > > > > > "mediatek,mt8195-pcie", "mediatek,mt8192-pcie" > > > > > > and > > > > > > "mediatek,mt8192-pcie" > > > > > > be the valid options. > > > > > > Personally I'm not seeing enough value to justify adding the compatible string > > > just for informational purposes though. One could easily discern which hardware > > > is used by looking at the device tree. > > > > > > > I agree, if no differences between the two chips are known, adding a > > binding withe new compatible and a fallback is a good thing. If we > > later on realize that mt8195 PCI block has differences, we can add the > > matching to the driver. > > So this series can be dropped, right ? Yes, we will send dt-bindings with dts changes in another series, thanks. > > Thanks, > Lorenzo > > > Regards, > > Matthias > > > > > > > > Regards > > > ChenYu > > > > > > > > >> Thanks. > > >>> > > >>> > > >>> Thanks > > >>> ChenYu > > >>> > > >>> > > >>>> Signed-off-by: Jianjun Wang <jianjun.wang@mediatek.com> > > >>>> --- > > >>>> Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml | 4 +++- > > >>>> 1 file changed, 3 insertions(+), 1 deletion(-) > > >>>> > > >>>> diff --git a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > > >>>> index e7b1f9892da4..d5e4a3e63d97 100644 > > >>>> --- a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > > >>>> +++ b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml > > >>>> @@ -48,7 +48,9 @@ allOf: > > >>>> > > >>>> properties: > > >>>> compatible: > > >>>> - const: mediatek,mt8192-pcie > > >>>> + oneOf: > > >>>> + - const: mediatek,mt8192-pcie > > >>>> + - const: mediatek,mt8195-pcie > > >>>> > > >>>> reg: > > >>>> maxItems: 1 > > >>>> -- > > >>>> 2.18.0 > > >>>> _______________________________________________ > > >>>> Linux-mediatek mailing list > > >>>> Linux-mediatek@lists.infradead.org > > >>>> http://lists.infradead.org/mailman/listinfo/linux-mediatek > > >>
diff --git a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml index e7b1f9892da4..d5e4a3e63d97 100644 --- a/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml +++ b/Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml @@ -48,7 +48,9 @@ allOf: properties: compatible: - const: mediatek,mt8192-pcie + oneOf: + - const: mediatek,mt8192-pcie + - const: mediatek,mt8195-pcie reg: maxItems: 1
MT8195 is an ARM platform SoC which has the same PCIe IP with MT8192. Signed-off-by: Jianjun Wang <jianjun.wang@mediatek.com> --- Documentation/devicetree/bindings/pci/mediatek-pcie-gen3.yaml | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)