Add ethernet support for MediaTek SoCs from mt8188 family.
As mt8188 and mt8195 have same ethernet design, so private data
"mt8195_gmac_variant" can be reused for mt8188.
Signed-off-by: Jianguo Zhang <[email protected]>
---
drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
index d42e1afb6521..f45be440b6d0 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
@@ -720,6 +720,8 @@ static const struct of_device_id mediatek_dwmac_match[] = {
.data = &mt2712_gmac_variant },
{ .compatible = "mediatek,mt8195-gmac",
.data = &mt8195_gmac_variant },
+ { .compatible = "mediatek,mt8188-gmac",
+ .data = &mt8195_gmac_variant },
{ }
};
--
2.25.1
On 19/09/2022 10:04, Jianguo Zhang wrote:
> Add ethernet support for MediaTek SoCs from mt8188 family.
> As mt8188 and mt8195 have same ethernet design, so private data
> "mt8195_gmac_variant" can be reused for mt8188.
>
> Signed-off-by: Jianguo Zhang <[email protected]>
> ---
> drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> index d42e1afb6521..f45be440b6d0 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> @@ -720,6 +720,8 @@ static const struct of_device_id mediatek_dwmac_match[] = {
> .data = &mt2712_gmac_variant },
> { .compatible = "mediatek,mt8195-gmac",
> .data = &mt8195_gmac_variant },
> + { .compatible = "mediatek,mt8188-gmac",
> + .data = &mt8195_gmac_variant },
It's the same. No need for new entry.
Best regards,
Krzysztof
Dear Krzysztof,
Thanks for your comments.
On Mon, 2022-09-19 at 10:19 +0200, Krzysztof Kozlowski wrote:
> On 19/09/2022 10:04, Jianguo Zhang wrote:
> > Add ethernet support for MediaTek SoCs from mt8188 family.
> > As mt8188 and mt8195 have same ethernet design, so private data
> > "mt8195_gmac_variant" can be reused for mt8188.
> >
> > Signed-off-by: Jianguo Zhang <[email protected]>
> > ---
> > drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> > b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> > index d42e1afb6521..f45be440b6d0 100644
> > --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> > +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> > @@ -720,6 +720,8 @@ static const struct of_device_id
> > mediatek_dwmac_match[] = {
> > .data = &mt2712_gmac_variant },
> > { .compatible = "mediatek,mt8195-gmac",
> > .data = &mt8195_gmac_variant },
> > + { .compatible = "mediatek,mt8188-gmac",
> > + .data = &mt8195_gmac_variant },
>
> It's the same. No need for new entry.
>
mt8188 and mt8195 are different SoCs and we need to distinguish mt8188
from mt8195, so I think a new entry is needed for mt8188 with the
specific "compatiable".
On the other hand, mt8188 and mt8195 have same ethernet design, so the
private data "mt8195_gmac_variant" can be resued to reduce redundant
info in driver.
>
> Best regards,
> Krzysztof
On 19/09/2022 11:56, Jianguo Zhang wrote:
>> No, this does not justify new entry. You need specific compatible,
>> but
>> not new entry.
>>
>>> On the other hand, mt8188 and mt8195 have same ethernet design, so
>>> the
>>> private data "mt8195_gmac_variant" can be resued to reduce
>>> redundant
>>> info in driver.
>>
>> And you do not need new entry in the driver.
> Do you mean that I can use "mediatek,mt8195-gmac" as compatible for
> ethernet in mt8188 DTS file?
Yes, as a fallback. Example schema describes such case.
Best regards,
Krzysztof
On 19/09/2022 10:37, Jianguo Zhang wrote:
> Dear Krzysztof,
>
> Thanks for your comments.
>
>
> On Mon, 2022-09-19 at 10:19 +0200, Krzysztof Kozlowski wrote:
>> On 19/09/2022 10:04, Jianguo Zhang wrote:
>>> Add ethernet support for MediaTek SoCs from mt8188 family.
>>> As mt8188 and mt8195 have same ethernet design, so private data
>>> "mt8195_gmac_variant" can be reused for mt8188.
>>>
>>> Signed-off-by: Jianguo Zhang <[email protected]>
>>> ---
>>> drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c | 2 ++
>>> 1 file changed, 2 insertions(+)
>>>
>>> diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
>>> b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
>>> index d42e1afb6521..f45be440b6d0 100644
>>> --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
>>> +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
>>> @@ -720,6 +720,8 @@ static const struct of_device_id
>>> mediatek_dwmac_match[] = {
>>> .data = &mt2712_gmac_variant },
>>> { .compatible = "mediatek,mt8195-gmac",
>>> .data = &mt8195_gmac_variant },
>>> + { .compatible = "mediatek,mt8188-gmac",
>>> + .data = &mt8195_gmac_variant },
>>
>> It's the same. No need for new entry.
>>
> mt8188 and mt8195 are different SoCs and we need to distinguish mt8188
> from mt8195, so I think a new entry is needed for mt8188 with the
> specific "compatiable".
No, this does not justify new entry. You need specific compatible, but
not new entry.
> On the other hand, mt8188 and mt8195 have same ethernet design, so the
> private data "mt8195_gmac_variant" can be resued to reduce redundant
> info in driver.
And you do not need new entry in the driver.
Best regards,
Krzysztof
Dear Krzysztof,
On Mon, 2022-09-19 at 11:27 +0200, Krzysztof Kozlowski wrote:
> On 19/09/2022 10:37, Jianguo Zhang wrote:
> > Dear Krzysztof,
> >
> > Thanks for your comments.
> >
> >
> > On Mon, 2022-09-19 at 10:19 +0200, Krzysztof Kozlowski wrote:
> > > On 19/09/2022 10:04, Jianguo Zhang wrote:
> > > > Add ethernet support for MediaTek SoCs from mt8188 family.
> > > > As mt8188 and mt8195 have same ethernet design, so private data
> > > > "mt8195_gmac_variant" can be reused for mt8188.
> > > >
> > > > Signed-off-by: Jianguo Zhang <[email protected]>
> > > > ---
> > > > drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c | 2 ++
> > > > 1 file changed, 2 insertions(+)
> > > >
> > > > diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-
> > > > mediatek.c
> > > > b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> > > > index d42e1afb6521..f45be440b6d0 100644
> > > > --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> > > > +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-mediatek.c
> > > > @@ -720,6 +720,8 @@ static const struct of_device_id
> > > > mediatek_dwmac_match[] = {
> > > > .data = &mt2712_gmac_variant },
> > > > { .compatible = "mediatek,mt8195-gmac",
> > > > .data = &mt8195_gmac_variant },
> > > > + { .compatible = "mediatek,mt8188-gmac",
> > > > + .data = &mt8195_gmac_variant },
> > >
> > > It's the same. No need for new entry.
> > >
> >
> > mt8188 and mt8195 are different SoCs and we need to distinguish
> > mt8188
> > from mt8195, so I think a new entry is needed for mt8188 with the
> > specific "compatiable".
>
> No, this does not justify new entry. You need specific compatible,
> but
> not new entry.
>
> > On the other hand, mt8188 and mt8195 have same ethernet design, so
> > the
> > private data "mt8195_gmac_variant" can be resued to reduce
> > redundant
> > info in driver.
>
> And you do not need new entry in the driver.
Do you mean that I can use "mediatek,mt8195-gmac" as compatible for
ethernet in mt8188 DTS file?
>
> Best regards,
> Krzysztof
BRS
Jianguo
Dear Krzysztof,
Thanks for your comment.
On Mon, 2022-09-19 at 12:06 +0200, Krzysztof Kozlowski wrote:
> On 19/09/2022 11:56, Jianguo Zhang wrote:
> > > No, this does not justify new entry. You need specific
> > > compatible,
> > > but
> > > not new entry.
> > >
> > > > On the other hand, mt8188 and mt8195 have same ethernet design,
> > > > so
> > > > the
> > > > private data "mt8195_gmac_variant" can be resued to reduce
> > > > redundant
> > > > info in driver.
> > >
> > > And you do not need new entry in the driver.
> >
> > Do you mean that I can use "mediatek,mt8195-gmac" as compatible for
> > ethernet in mt8188 DTS file?
>
> Yes, as a fallback. Example schema describes such case.
>
OK, we will drop the changes in patch #1 and add MT8188 related info in
binding document in next version patches.
> Best regards,
> Krzysztof
>
BRS
Jianguo