[PATCH net-next resend 0/6] Support for ASP 2.2 and optimizations

[PATCH net-next resend 0/6] Support for ASP 2.2 and optimizations

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 897 bytes --]

ASP 2.2 adds some power savings during low power modes.

Also make various improvements when entering low power modes and
reduce MDIO traffic by hooking up interrupts.

Justin Chen (6):
  dt-bindings: net: brcm,unimac-mdio: Add asp-v2.2
  dt-bindings: net: brcm,asp-v2.0: Add asp-v2.2
  net: bcmasp: Add support for ASP 2.2
  net: phy: mdio-bcm-unimac: Add asp v2.2 support
  net: bcmasp: Keep buffers through power management
  net: bcmasp: Add support for PHY interrupts

 .../bindings/net/brcm,asp-v2.0.yaml           |   4 +
 .../bindings/net/brcm,unimac-mdio.yaml        |   1 +
 drivers/net/ethernet/broadcom/asp2/bcmasp.c   |  90 +++++++-
 drivers/net/ethernet/broadcom/asp2/bcmasp.h   |  23 ++-
 .../net/ethernet/broadcom/asp2/bcmasp_intf.c  | 192 ++++++++----------
 drivers/net/mdio/mdio-bcm-unimac.c            |   1 +
 6 files changed, 195 insertions(+), 116 deletions(-)

-- 
2.34.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

[PATCH net-next resend 1/6] dt-bindings: net: brcm,unimac-mdio: Add asp-v2.2

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 708 bytes --]

The ASP 2.2 Ethernet controller uses a brcm unimac.

Signed-off-by: Justin Chen <justin.chen@broadcom.com>
---
 Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml b/Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml
index 6684810fcbf0..23dfe0838dca 100644
--- a/Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml
+++ b/Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml
@@ -24,6 +24,7 @@ properties:
       - brcm,genet-mdio-v5
       - brcm,asp-v2.0-mdio
       - brcm,asp-v2.1-mdio
+      - brcm,asp-v2.2-mdio
       - brcm,unimac-mdio
 
   reg:
-- 
2.34.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

Re: [PATCH net-next resend 1/6] dt-bindings: net: brcm,unimac-mdio: Add asp-v2.2

[Krzysztof Kozlowski]

On 23/02/2024 23:24, Justin Chen wrote:
> The ASP 2.2 Ethernet controller uses a brcm unimac.
> 
> Signed-off-by: Justin Chen <justin.chen@broadcom.com>
> ---
>  Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml | 1 +
>  1 file changed, 1 insertion(+)

Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>

Best regards,
Krzysztof

Re: [PATCH net-next resend 1/6] dt-bindings: net: brcm,unimac-mdio: Add asp-v2.2

[Florian Fainelli]

[-- Attachment #1: Type: text/plain, Size: 223 bytes --]

On 2/23/24 14:24, Justin Chen wrote:
> The ASP 2.2 Ethernet controller uses a brcm unimac.
> 
> Signed-off-by: Justin Chen <justin.chen@broadcom.com>

Acked-by: Florian Fainelli <florian.fainelli@broadcom.com>
-- 
Florian


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4221 bytes --]

[PATCH net-next resend 2/6] dt-bindings: net: brcm,asp-v2.0: Add asp-v2.2

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 795 bytes --]

Add support for ASP 2.2.

Signed-off-by: Justin Chen <justin.chen@broadcom.com>
---
 Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
index 75d8138298fb..5a345f03de17 100644
--- a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
+++ b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
@@ -15,6 +15,10 @@ description: Broadcom Ethernet controller first introduced with 72165
 properties:
   compatible:
     oneOf:
+      - items:
+          - enum:
+              - brcm,bcm74165-asp
+          - const: brcm,asp-v2.2
       - items:
           - enum:
               - brcm,bcm74165-asp
-- 
2.34.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

Re: [PATCH net-next resend 2/6] dt-bindings: net: brcm,asp-v2.0: Add asp-v2.2

[Krzysztof Kozlowski]

On 23/02/2024 23:24, Justin Chen wrote:
> Add support for ASP 2.2.
> 
> Signed-off-by: Justin Chen <justin.chen@broadcom.com>
> ---
>  Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
> index 75d8138298fb..5a345f03de17 100644
> --- a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
> +++ b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
> @@ -15,6 +15,10 @@ description: Broadcom Ethernet controller first introduced with 72165
>  properties:
>    compatible:
>      oneOf:
> +      - items:
> +          - enum:
> +              - brcm,bcm74165-asp
> +          - const: brcm,asp-v2.2
>        - items:
>            - enum:
>                - brcm,bcm74165-asp

Hm, this confuses me: why do you have same SoC with three different
versions of the same block?

Best regards,
Krzysztof

Re: [PATCH net-next resend 2/6] dt-bindings: net: brcm,asp-v2.0: Add asp-v2.2

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 1317 bytes --]



On 2/24/24 2:22 AM, Krzysztof Kozlowski wrote:
> On 23/02/2024 23:24, Justin Chen wrote:
>> Add support for ASP 2.2.
>>
>> Signed-off-by: Justin Chen <justin.chen@broadcom.com>
>> ---
>>   Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml | 4 ++++
>>   1 file changed, 4 insertions(+)
>>
>> diff --git a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
>> index 75d8138298fb..5a345f03de17 100644
>> --- a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
>> +++ b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
>> @@ -15,6 +15,10 @@ description: Broadcom Ethernet controller first introduced with 72165
>>   properties:
>>     compatible:
>>       oneOf:
>> +      - items:
>> +          - enum:
>> +              - brcm,bcm74165-asp
>> +          - const: brcm,asp-v2.2
>>         - items:
>>             - enum:
>>                 - brcm,bcm74165-asp
> 
> Hm, this confuses me: why do you have same SoC with three different
> versions of the same block?
> 

bcm72165 -> asp-v2.0
bcm74165 -> asp-v2.1
Are two different SoCs.

The entry I just added is
bcm74165 -> asp-v2.2
This is a SoC minor revision. Maybe it should bcm74165b0-asp instead? 
Not sure what the protocol is.

Thanks,
Justin

> Best regards,
> Krzysztof
> 

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

Re: [PATCH net-next resend 2/6] dt-bindings: net: brcm,asp-v2.0: Add asp-v2.2

[Krzysztof Kozlowski]

On 26/02/2024 20:42, Justin Chen wrote:
> 
> 
> On 2/24/24 2:22 AM, Krzysztof Kozlowski wrote:
>> On 23/02/2024 23:24, Justin Chen wrote:
>>> Add support for ASP 2.2.
>>>
>>> Signed-off-by: Justin Chen <justin.chen@broadcom.com>
>>> ---
>>>   Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml | 4 ++++
>>>   1 file changed, 4 insertions(+)
>>>
>>> diff --git a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
>>> index 75d8138298fb..5a345f03de17 100644
>>> --- a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
>>> +++ b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
>>> @@ -15,6 +15,10 @@ description: Broadcom Ethernet controller first introduced with 72165
>>>   properties:
>>>     compatible:
>>>       oneOf:
>>> +      - items:
>>> +          - enum:
>>> +              - brcm,bcm74165-asp
>>> +          - const: brcm,asp-v2.2
>>>         - items:
>>>             - enum:
>>>                 - brcm,bcm74165-asp
>>
>> Hm, this confuses me: why do you have same SoC with three different
>> versions of the same block?
>>
> 
> bcm72165 -> asp-v2.0
> bcm74165 -> asp-v2.1
> Are two different SoCs.

Ah, right, existing bindings has two SoCs.

> 
> The entry I just added is
> bcm74165 -> asp-v2.2
> This is a SoC minor revision. Maybe it should bcm74165b0-asp instead? 
> Not sure what the protocol is.

So still the confusion - same SoC with different IP blocks. That's
totally opposite of what we expect: same version of IP block used in
multiple SoCs.



Best regards,
Krzysztof

Re: [PATCH net-next resend 2/6] dt-bindings: net: brcm,asp-v2.0: Add asp-v2.2

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 1960 bytes --]



On 2/26/2024 10:55 PM, Krzysztof Kozlowski wrote:
> On 26/02/2024 20:42, Justin Chen wrote:
>>
>>
>> On 2/24/24 2:22 AM, Krzysztof Kozlowski wrote:
>>> On 23/02/2024 23:24, Justin Chen wrote:
>>>> Add support for ASP 2.2.
>>>>
>>>> Signed-off-by: Justin Chen <justin.chen@broadcom.com>
>>>> ---
>>>>    Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml | 4 ++++
>>>>    1 file changed, 4 insertions(+)
>>>>
>>>> diff --git a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
>>>> index 75d8138298fb..5a345f03de17 100644
>>>> --- a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
>>>> +++ b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
>>>> @@ -15,6 +15,10 @@ description: Broadcom Ethernet controller first introduced with 72165
>>>>    properties:
>>>>      compatible:
>>>>        oneOf:
>>>> +      - items:
>>>> +          - enum:
>>>> +              - brcm,bcm74165-asp
>>>> +          - const: brcm,asp-v2.2
>>>>          - items:
>>>>              - enum:
>>>>                  - brcm,bcm74165-asp
>>>
>>> Hm, this confuses me: why do you have same SoC with three different
>>> versions of the same block?
>>>
>>
>> bcm72165 -> asp-v2.0
>> bcm74165 -> asp-v2.1
>> Are two different SoCs.
> 
> Ah, right, existing bindings has two SoCs.
> 
>>
>> The entry I just added is
>> bcm74165 -> asp-v2.2
>> This is a SoC minor revision. Maybe it should bcm74165b0-asp instead?
>> Not sure what the protocol is.
> 
> So still the confusion - same SoC with different IP blocks. That's
> totally opposite of what we expect: same version of IP block used in
> multiple SoCs.
> 
>

Agreed. Unfortunately what we expect is not always what comes to fruition...

Thinking about it again, I prefer bcm74165b0-asp. Otherwise it doesn't 
properly describe the hardware as we do not have one SoC with two 
different IP blocks.

Thanks,
Justin

> 
> Best regards,
> Krzysztof
> 

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

Re: [PATCH net-next resend 2/6] dt-bindings: net: brcm,asp-v2.0: Add asp-v2.2

[Paolo Abeni]

On Mon, 2024-02-26 at 23:07 -0800, Justin Chen wrote:
> On 2/26/2024 10:55 PM, Krzysztof Kozlowski wrote:
> > On 26/02/2024 20:42, Justin Chen wrote:
> > > On 2/24/24 2:22 AM, Krzysztof Kozlowski wrote:
> > > > On 23/02/2024 23:24, Justin Chen wrote:
> > > > > Add support for ASP 2.2.
> > > > > 
> > > > > Signed-off-by: Justin Chen <justin.chen@broadcom.com>
> > > > > ---
> > > > >    Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml | 4 ++++
> > > > >    1 file changed, 4 insertions(+)
> > > > > 
> > > > > diff --git a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
> > > > > index 75d8138298fb..5a345f03de17 100644
> > > > > --- a/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
> > > > > +++ b/Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
> > > > > @@ -15,6 +15,10 @@ description: Broadcom Ethernet controller first introduced with 72165
> > > > >    properties:
> > > > >      compatible:
> > > > >        oneOf:
> > > > > +      - items:
> > > > > +          - enum:
> > > > > +              - brcm,bcm74165-asp
> > > > > +          - const: brcm,asp-v2.2
> > > > >          - items:
> > > > >              - enum:
> > > > >                  - brcm,bcm74165-asp
> > > > 
> > > > Hm, this confuses me: why do you have same SoC with three different
> > > > versions of the same block?
> > > > 
> > > 
> > > bcm72165 -> asp-v2.0
> > > bcm74165 -> asp-v2.1
> > > Are two different SoCs.
> > 
> > Ah, right, existing bindings has two SoCs.
> > 
> > > 
> > > The entry I just added is
> > > bcm74165 -> asp-v2.2
> > > This is a SoC minor revision. Maybe it should bcm74165b0-asp instead?
> > > Not sure what the protocol is.
> > 
> > So still the confusion - same SoC with different IP blocks. That's
> > totally opposite of what we expect: same version of IP block used in
> > multiple SoCs.
> 
> Agreed. Unfortunately what we expect is not always what comes to fruition...
> 
> Thinking about it again, I prefer bcm74165b0-asp. Otherwise it doesn't 
> properly describe the hardware as we do not have one SoC with two 
> different IP blocks.

I read the above as you intend to send a v2 with an update dt-binding.
Please correct me if I'm wrong.

Meanwhile dropping this revision from PW.

Cheers,

Paolo

Re: [PATCH net-next resend 2/6] dt-bindings: net: brcm,asp-v2.0: Add asp-v2.2

[Florian Fainelli]

[-- Attachment #1: Type: text/plain, Size: 196 bytes --]

On 2/23/24 14:24, Justin Chen wrote:
> Add support for ASP 2.2.
> 
> Signed-off-by: Justin Chen <justin.chen@broadcom.com>

Acked-by: Florian Fainelli <florian.fainelli@broadcom.com>
-- 
Florian


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4221 bytes --]

[PATCH net-next resend 3/6] net: bcmasp: Add support for ASP 2.2

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 8557 bytes --]

ASP 2.2 improves power savings during low power modes.

A new register was added to toggle to a slower clock during low
power modes.

EEE was broken for ASP 2.0/2.1. A HW workaround was added for
ASP 2.2 that requires toggling a chicken bit.

Signed-off-by: Justin Chen <justin.chen@broadcom.com>
---
 drivers/net/ethernet/broadcom/asp2/bcmasp.c   | 73 +++++++++++++++++--
 drivers/net/ethernet/broadcom/asp2/bcmasp.h   | 18 ++++-
 .../net/ethernet/broadcom/asp2/bcmasp_intf.c  |  6 ++
 3 files changed, 87 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/asp2/bcmasp.c b/drivers/net/ethernet/broadcom/asp2/bcmasp.c
index 80245c65cc90..100c69f3307a 100644
--- a/drivers/net/ethernet/broadcom/asp2/bcmasp.c
+++ b/drivers/net/ethernet/broadcom/asp2/bcmasp.c
@@ -972,7 +972,26 @@ static void bcmasp_core_init(struct bcmasp_priv *priv)
 		      ASP_INTR2_CLEAR);
 }
 
-static void bcmasp_core_clock_select(struct bcmasp_priv *priv, bool slow)
+static void bcmasp_core_clock_select_many(struct bcmasp_priv *priv, bool slow)
+{
+	u32 reg;
+
+	reg = ctrl2_core_rl(priv, ASP_CTRL2_CORE_CLOCK_SELECT);
+	if (slow)
+		reg &= ~ASP_CTRL2_CORE_CLOCK_SELECT_MAIN;
+	else
+		reg |= ASP_CTRL2_CORE_CLOCK_SELECT_MAIN;
+	ctrl2_core_wl(priv, reg, ASP_CTRL2_CORE_CLOCK_SELECT);
+
+	reg = ctrl2_core_rl(priv, ASP_CTRL2_CPU_CLOCK_SELECT);
+	if (slow)
+		reg &= ~ASP_CTRL2_CPU_CLOCK_SELECT_MAIN;
+	else
+		reg |= ASP_CTRL2_CPU_CLOCK_SELECT_MAIN;
+	ctrl2_core_wl(priv, reg, ASP_CTRL2_CPU_CLOCK_SELECT);
+}
+
+static void bcmasp_core_clock_select_one(struct bcmasp_priv *priv, bool slow)
 {
 	u32 reg;
 
@@ -1166,6 +1185,24 @@ static void bcmasp_wol_irq_destroy_per_intf(struct bcmasp_priv *priv)
 	}
 }
 
+static void bcmasp_eee_fixup(struct bcmasp_intf *intf, bool en)
+{
+	u32 reg, phy_lpi_overwrite;
+
+	reg = rx_edpkt_core_rl(intf->parent, ASP_EDPKT_SPARE_REG);
+	phy_lpi_overwrite = intf->internal_phy ? ASP_EDPKT_SPARE_REG_EPHY_LPI :
+			    ASP_EDPKT_SPARE_REG_GPHY_LPI;
+
+	if (en)
+		reg |= phy_lpi_overwrite;
+	else
+		reg &= ~phy_lpi_overwrite;
+
+	rx_edpkt_core_wl(intf->parent, reg, ASP_EDPKT_SPARE_REG);
+
+	usleep_range(50, 100);
+}
+
 static struct bcmasp_hw_info v20_hw_info = {
 	.rx_ctrl_flush = ASP_RX_CTRL_FLUSH,
 	.umac2fb = UMAC2FB_OFFSET,
@@ -1178,6 +1215,7 @@ static const struct bcmasp_plat_data v20_plat_data = {
 	.init_wol = bcmasp_init_wol_per_intf,
 	.enable_wol = bcmasp_enable_wol_per_intf,
 	.destroy_wol = bcmasp_wol_irq_destroy_per_intf,
+	.core_clock_select = bcmasp_core_clock_select_one,
 	.hw_info = &v20_hw_info,
 };
 
@@ -1194,17 +1232,39 @@ static const struct bcmasp_plat_data v21_plat_data = {
 	.init_wol = bcmasp_init_wol_shared,
 	.enable_wol = bcmasp_enable_wol_shared,
 	.destroy_wol = bcmasp_wol_irq_destroy_shared,
+	.core_clock_select = bcmasp_core_clock_select_one,
 	.hw_info = &v21_hw_info,
 };
 
+static const struct bcmasp_plat_data v22_plat_data = {
+	.init_wol = bcmasp_init_wol_shared,
+	.enable_wol = bcmasp_enable_wol_shared,
+	.destroy_wol = bcmasp_wol_irq_destroy_shared,
+	.core_clock_select = bcmasp_core_clock_select_many,
+	.hw_info = &v21_hw_info,
+	.eee_fixup = bcmasp_eee_fixup,
+};
+
+static void bcmasp_set_pdata(struct bcmasp_priv *priv, const struct bcmasp_plat_data *pdata)
+{
+	priv->init_wol = pdata->init_wol;
+	priv->enable_wol = pdata->enable_wol;
+	priv->destroy_wol = pdata->destroy_wol;
+	priv->core_clock_select = pdata->core_clock_select;
+	priv->eee_fixup = pdata->eee_fixup;
+	priv->hw_info = pdata->hw_info;
+}
+
 static const struct of_device_id bcmasp_of_match[] = {
 	{ .compatible = "brcm,asp-v2.0", .data = &v20_plat_data },
 	{ .compatible = "brcm,asp-v2.1", .data = &v21_plat_data },
+	{ .compatible = "brcm,asp-v2.2", .data = &v22_plat_data },
 	{ /* sentinel */ },
 };
 MODULE_DEVICE_TABLE(of, bcmasp_of_match);
 
 static const struct of_device_id bcmasp_mdio_of_match[] = {
+	{ .compatible = "brcm,asp-v2.2-mdio", },
 	{ .compatible = "brcm,asp-v2.1-mdio", },
 	{ .compatible = "brcm,asp-v2.0-mdio", },
 	{ /* sentinel */ },
@@ -1265,16 +1325,13 @@ static int bcmasp_probe(struct platform_device *pdev)
 	if (!pdata)
 		return dev_err_probe(dev, -EINVAL, "unable to find platform data\n");
 
-	priv->init_wol = pdata->init_wol;
-	priv->enable_wol = pdata->enable_wol;
-	priv->destroy_wol = pdata->destroy_wol;
-	priv->hw_info = pdata->hw_info;
+	bcmasp_set_pdata(priv, pdata);
 
 	/* Enable all clocks to ensure successful probing */
 	bcmasp_core_clock_set(priv, ASP_CTRL_CLOCK_CTRL_ASP_ALL_DISABLE, 0);
 
 	/* Switch to the main clock */
-	bcmasp_core_clock_select(priv, false);
+	priv->core_clock_select(priv, false);
 
 	bcmasp_intr2_mask_set_all(priv);
 	bcmasp_intr2_clear_all(priv);
@@ -1381,7 +1438,7 @@ static int __maybe_unused bcmasp_suspend(struct device *d)
 	 */
 	bcmasp_core_clock_set(priv, 0, ASP_CTRL_CLOCK_CTRL_ASP_TX_DISABLE);
 
-	bcmasp_core_clock_select(priv, true);
+	priv->core_clock_select(priv, true);
 
 	clk_disable_unprepare(priv->clk);
 
@@ -1399,7 +1456,7 @@ static int __maybe_unused bcmasp_resume(struct device *d)
 		return ret;
 
 	/* Switch to the main clock domain */
-	bcmasp_core_clock_select(priv, false);
+	priv->core_clock_select(priv, false);
 
 	/* Re-enable all clocks for re-initialization */
 	bcmasp_core_clock_set(priv, ASP_CTRL_CLOCK_CTRL_ASP_ALL_DISABLE, 0);
diff --git a/drivers/net/ethernet/broadcom/asp2/bcmasp.h b/drivers/net/ethernet/broadcom/asp2/bcmasp.h
index 312bf9b6576e..61598dc070b1 100644
--- a/drivers/net/ethernet/broadcom/asp2/bcmasp.h
+++ b/drivers/net/ethernet/broadcom/asp2/bcmasp.h
@@ -33,6 +33,12 @@
 #define ASP_WAKEUP_INTR2_FILT_1			BIT(3)
 #define ASP_WAKEUP_INTR2_FW			BIT(4)
 
+#define ASP_CTRL2_OFFSET			0x2000
+#define  ASP_CTRL2_CORE_CLOCK_SELECT		0x0
+#define   ASP_CTRL2_CORE_CLOCK_SELECT_MAIN	BIT(0)
+#define  ASP_CTRL2_CPU_CLOCK_SELECT		0x4
+#define   ASP_CTRL2_CPU_CLOCK_SELECT_MAIN	BIT(0)
+
 #define ASP_TX_ANALYTICS_OFFSET			0x4c000
 #define  ASP_TX_ANALYTICS_CTRL			0x0
 
@@ -134,8 +140,11 @@ enum asp_rx_net_filter_block {
 #define ASP_EDPKT_RX_PKT_CNT			0x138
 #define ASP_EDPKT_HDR_EXTR_CNT			0x13c
 #define ASP_EDPKT_HDR_OUT_CNT			0x140
+#define ASP_EDPKT_SPARE_REG			0x174
+#define  ASP_EDPKT_SPARE_REG_EPHY_LPI		BIT(4)
+#define  ASP_EDPKT_SPARE_REG_GPHY_LPI		BIT(3)
 
-#define ASP_CTRL				0x101000
+#define ASP_CTRL_OFFSET				0x101000
 #define  ASP_CTRL_ASP_SW_INIT			0x04
 #define   ASP_CTRL_ASP_SW_INIT_ACPUSS_CORE	BIT(0)
 #define   ASP_CTRL_ASP_SW_INIT_ASP_TX		BIT(1)
@@ -372,6 +381,8 @@ struct bcmasp_plat_data {
 	void (*init_wol)(struct bcmasp_priv *priv);
 	void (*enable_wol)(struct bcmasp_intf *intf, bool en);
 	void (*destroy_wol)(struct bcmasp_priv *priv);
+	void (*core_clock_select)(struct bcmasp_priv *priv, bool slow);
+	void (*eee_fixup)(struct bcmasp_intf *priv, bool en);
 	struct bcmasp_hw_info		*hw_info;
 };
 
@@ -390,6 +401,8 @@ struct bcmasp_priv {
 	void (*init_wol)(struct bcmasp_priv *priv);
 	void (*enable_wol)(struct bcmasp_intf *intf, bool en);
 	void (*destroy_wol)(struct bcmasp_priv *priv);
+	void (*core_clock_select)(struct bcmasp_priv *priv, bool slow);
+	void (*eee_fixup)(struct bcmasp_intf *intf, bool en);
 
 	void __iomem			*base;
 	struct	bcmasp_hw_info		*hw_info;
@@ -530,7 +543,8 @@ BCMASP_CORE_IO_MACRO(rx_analytics, ASP_RX_ANALYTICS_OFFSET);
 BCMASP_CORE_IO_MACRO(rx_ctrl, ASP_RX_CTRL_OFFSET);
 BCMASP_CORE_IO_MACRO(rx_filter, ASP_RX_FILTER_OFFSET);
 BCMASP_CORE_IO_MACRO(rx_edpkt, ASP_EDPKT_OFFSET);
-BCMASP_CORE_IO_MACRO(ctrl, ASP_CTRL);
+BCMASP_CORE_IO_MACRO(ctrl, ASP_CTRL_OFFSET);
+BCMASP_CORE_IO_MACRO(ctrl2, ASP_CTRL2_OFFSET);
 
 struct bcmasp_intf *bcmasp_interface_create(struct bcmasp_priv *priv,
 					    struct device_node *ndev_dn, int i);
diff --git a/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c b/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
index e429876c7291..36e6fae937ea 100644
--- a/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
+++ b/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
@@ -1333,6 +1333,9 @@ static void bcmasp_suspend_to_wol(struct bcmasp_intf *intf)
 				     ASP_WAKEUP_INTR2_MASK_CLEAR);
 	}
 
+	if (intf->eee.eee_enabled && intf->parent->eee_fixup)
+		intf->parent->eee_fixup(intf, true);
+
 	netif_dbg(intf, wol, ndev, "entered WOL mode\n");
 }
 
@@ -1381,6 +1384,9 @@ static void bcmasp_resume_from_wol(struct bcmasp_intf *intf)
 {
 	u32 reg;
 
+	if (intf->eee.eee_enabled && intf->parent->eee_fixup)
+		intf->parent->eee_fixup(intf, false);
+
 	reg = umac_rl(intf, UMC_MPD_CTRL);
 	reg &= ~UMC_MPD_CTRL_MPD_EN;
 	umac_wl(intf, reg, UMC_MPD_CTRL);
-- 
2.34.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

Re: [PATCH net-next resend 3/6] net: bcmasp: Add support for ASP 2.2

[Florian Fainelli]

[-- Attachment #1: Type: text/plain, Size: 425 bytes --]

On 2/23/24 14:24, Justin Chen wrote:
> ASP 2.2 improves power savings during low power modes.
> 
> A new register was added to toggle to a slower clock during low
> power modes.
> 
> EEE was broken for ASP 2.0/2.1. A HW workaround was added for
> ASP 2.2 that requires toggling a chicken bit.
> 
> Signed-off-by: Justin Chen <justin.chen@broadcom.com>

Acked-by: Florian Fainelli <florian.fainelli@broadcom.com>
-- 
Florian


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4221 bytes --]

[PATCH net-next resend 4/6] net: phy: mdio-bcm-unimac: Add asp v2.2 support

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 732 bytes --]

Add mdio compat string for ASP 2.0 ethernet driver.

Signed-off-by: Justin Chen <justin.chen@broadcom.com>
---
 drivers/net/mdio/mdio-bcm-unimac.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/mdio/mdio-bcm-unimac.c b/drivers/net/mdio/mdio-bcm-unimac.c
index 6fe08427fdd4..f40eb50bb978 100644
--- a/drivers/net/mdio/mdio-bcm-unimac.c
+++ b/drivers/net/mdio/mdio-bcm-unimac.c
@@ -334,6 +334,7 @@ static SIMPLE_DEV_PM_OPS(unimac_mdio_pm_ops,
 			 NULL, unimac_mdio_resume);
 
 static const struct of_device_id unimac_mdio_ids[] = {
+	{ .compatible = "brcm,asp-v2.2-mdio", },
 	{ .compatible = "brcm,asp-v2.1-mdio", },
 	{ .compatible = "brcm,asp-v2.0-mdio", },
 	{ .compatible = "brcm,genet-mdio-v5", },
-- 
2.34.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

Re: [PATCH net-next resend 4/6] net: phy: mdio-bcm-unimac: Add asp v2.2 support

[Florian Fainelli]

[-- Attachment #1: Type: text/plain, Size: 223 bytes --]

On 2/23/24 14:24, Justin Chen wrote:
> Add mdio compat string for ASP 2.0 ethernet driver.
> 
> Signed-off-by: Justin Chen <justin.chen@broadcom.com>

Acked-by: Florian Fainelli <florian.fainelli@broadcom.com>
-- 
Florian


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4221 bytes --]

[PATCH net-next resend 5/6] net: bcmasp: Keep buffers through power management

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 9153 bytes --]

There is no advantage of freeing and re-allocating buffers through
suspend and resume. This waste cycles and makes suspend/resume time
longer. We also open ourselves to failed allocations in systems with
heavy memory fragmentation.

Signed-off-by: Justin Chen <justin.chen@broadcom.com>
---
 drivers/net/ethernet/broadcom/asp2/bcmasp.h   |   1 +
 .../net/ethernet/broadcom/asp2/bcmasp_intf.c  | 181 ++++++++----------
 2 files changed, 76 insertions(+), 106 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/asp2/bcmasp.h b/drivers/net/ethernet/broadcom/asp2/bcmasp.h
index 61598dc070b1..127a5340625e 100644
--- a/drivers/net/ethernet/broadcom/asp2/bcmasp.h
+++ b/drivers/net/ethernet/broadcom/asp2/bcmasp.h
@@ -315,6 +315,7 @@ struct bcmasp_intf {
 	struct bcmasp_desc		*rx_edpkt_cpu;
 	dma_addr_t			rx_edpkt_dma_addr;
 	dma_addr_t			rx_edpkt_dma_read;
+	dma_addr_t			rx_edpkt_dma_valid;
 
 	/* RX buffer prefetcher ring*/
 	void				*rx_ring_cpu;
diff --git a/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c b/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
index 36e6fae937ea..0b378a6d43e7 100644
--- a/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
+++ b/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
@@ -674,40 +674,78 @@ static void bcmasp_adj_link(struct net_device *dev)
 		phy_print_status(phydev);
 }
 
-static int bcmasp_init_rx(struct bcmasp_intf *intf)
+static int bcmasp_alloc_buffers(struct bcmasp_intf *intf)
 {
 	struct device *kdev = &intf->parent->pdev->dev;
 	struct page *buffer_pg;
-	dma_addr_t dma;
-	void *p;
-	u32 reg;
-	int ret;
 
+	/* Alloc RX */
 	intf->rx_buf_order = get_order(RING_BUFFER_SIZE);
 	buffer_pg = alloc_pages(GFP_KERNEL, intf->rx_buf_order);
 	if (!buffer_pg)
 		return -ENOMEM;
 
-	dma = dma_map_page(kdev, buffer_pg, 0, RING_BUFFER_SIZE,
-			   DMA_FROM_DEVICE);
-	if (dma_mapping_error(kdev, dma)) {
-		__free_pages(buffer_pg, intf->rx_buf_order);
-		return -ENOMEM;
-	}
 	intf->rx_ring_cpu = page_to_virt(buffer_pg);
-	intf->rx_ring_dma = dma;
-	intf->rx_ring_dma_valid = intf->rx_ring_dma + RING_BUFFER_SIZE - 1;
+	intf->rx_ring_dma = dma_map_page(kdev, buffer_pg, 0, RING_BUFFER_SIZE,
+					 DMA_FROM_DEVICE);
+	if (dma_mapping_error(kdev, intf->rx_ring_dma))
+		goto free_rx_buffer;
+
+	intf->rx_edpkt_cpu = dma_alloc_coherent(kdev, DESC_RING_SIZE,
+						&intf->rx_edpkt_dma_addr, GFP_KERNEL);
+	if (!intf->rx_edpkt_cpu)
+		goto free_rx_buffer_dma;
+
+	/* Alloc TX */
+	intf->tx_spb_cpu = dma_alloc_coherent(kdev, DESC_RING_SIZE,
+					      &intf->tx_spb_dma_addr, GFP_KERNEL);
+	if (!intf->tx_spb_cpu)
+		goto free_rx_edpkt_dma;
 
-	p = dma_alloc_coherent(kdev, DESC_RING_SIZE, &intf->rx_edpkt_dma_addr,
+	intf->tx_cbs = kcalloc(DESC_RING_COUNT, sizeof(struct bcmasp_tx_cb),
 			       GFP_KERNEL);
-	if (!p) {
-		ret = -ENOMEM;
-		goto free_rx_ring;
-	}
-	intf->rx_edpkt_cpu = p;
+	if (!intf->tx_cbs)
+		goto free_tx_spb_dma;
 
-	netif_napi_add(intf->ndev, &intf->rx_napi, bcmasp_rx_poll);
+	return 0;
+
+free_tx_spb_dma:
+	dma_free_coherent(kdev, DESC_RING_SIZE, intf->tx_spb_cpu,
+			  intf->tx_spb_dma_addr);
+free_rx_edpkt_dma:
+	dma_free_coherent(kdev, DESC_RING_SIZE, intf->rx_edpkt_cpu,
+			  intf->rx_edpkt_dma_addr);
+free_rx_buffer_dma:
+	dma_unmap_page(kdev, intf->rx_ring_dma, RING_BUFFER_SIZE,
+		       DMA_FROM_DEVICE);
+free_rx_buffer:
+	__free_pages(buffer_pg, intf->rx_buf_order);
+
+	return -ENOMEM;
+}
+
+static void bcmasp_reclaim_free_buffers(struct bcmasp_intf *intf)
+{
+	struct device *kdev = &intf->parent->pdev->dev;
+
+	/* RX buffers */
+	dma_free_coherent(kdev, DESC_RING_SIZE, intf->rx_edpkt_cpu,
+			  intf->rx_edpkt_dma_addr);
+	dma_unmap_page(kdev, intf->rx_ring_dma, RING_BUFFER_SIZE,
+		       DMA_FROM_DEVICE);
+	__free_pages(virt_to_page(intf->rx_ring_cpu), intf->rx_buf_order);
+
+	/* TX buffers */
+	dma_free_coherent(kdev, DESC_RING_SIZE, intf->tx_spb_cpu,
+			  intf->tx_spb_dma_addr);
+	kfree(intf->tx_cbs);
+}
 
+static void bcmasp_init_rx(struct bcmasp_intf *intf)
+{
+	/* Restart from index 0 */
+	intf->rx_ring_dma_valid = intf->rx_ring_dma + RING_BUFFER_SIZE - 1;
+	intf->rx_edpkt_dma_valid = intf->rx_edpkt_dma_addr + (DESC_RING_SIZE - 1);
 	intf->rx_edpkt_dma_read = intf->rx_edpkt_dma_addr;
 	intf->rx_edpkt_index = 0;
 
@@ -733,64 +771,23 @@ static int bcmasp_init_rx(struct bcmasp_intf *intf)
 	rx_edpkt_dma_wq(intf, intf->rx_edpkt_dma_addr, RX_EDPKT_DMA_WRITE);
 	rx_edpkt_dma_wq(intf, intf->rx_edpkt_dma_addr, RX_EDPKT_DMA_READ);
 	rx_edpkt_dma_wq(intf, intf->rx_edpkt_dma_addr, RX_EDPKT_DMA_BASE);
-	rx_edpkt_dma_wq(intf, intf->rx_edpkt_dma_addr + (DESC_RING_SIZE - 1),
-			RX_EDPKT_DMA_END);
-	rx_edpkt_dma_wq(intf, intf->rx_edpkt_dma_addr + (DESC_RING_SIZE - 1),
-			RX_EDPKT_DMA_VALID);
-
-	reg = UMAC2FB_CFG_DEFAULT_EN |
-	      ((intf->channel + 11) << UMAC2FB_CFG_CHID_SHIFT);
-	reg |= (0xd << UMAC2FB_CFG_OK_SEND_SHIFT);
-	umac2fb_wl(intf, reg, UMAC2FB_CFG);
+	rx_edpkt_dma_wq(intf, intf->rx_edpkt_dma_valid, RX_EDPKT_DMA_END);
+	rx_edpkt_dma_wq(intf, intf->rx_edpkt_dma_valid, RX_EDPKT_DMA_VALID);
 
-	return 0;
-
-free_rx_ring:
-	dma_unmap_page(kdev, intf->rx_ring_dma, RING_BUFFER_SIZE,
-		       DMA_FROM_DEVICE);
-	__free_pages(virt_to_page(intf->rx_ring_cpu), intf->rx_buf_order);
-
-	return ret;
+	umac2fb_wl(intf, UMAC2FB_CFG_DEFAULT_EN | ((intf->channel + 11) <<
+		   UMAC2FB_CFG_CHID_SHIFT) | (0xd << UMAC2FB_CFG_OK_SEND_SHIFT),
+		   UMAC2FB_CFG);
 }
 
-static void bcmasp_reclaim_free_all_rx(struct bcmasp_intf *intf)
-{
-	struct device *kdev = &intf->parent->pdev->dev;
-
-	dma_free_coherent(kdev, DESC_RING_SIZE, intf->rx_edpkt_cpu,
-			  intf->rx_edpkt_dma_addr);
-	dma_unmap_page(kdev, intf->rx_ring_dma, RING_BUFFER_SIZE,
-		       DMA_FROM_DEVICE);
-	__free_pages(virt_to_page(intf->rx_ring_cpu), intf->rx_buf_order);
-}
 
-static int bcmasp_init_tx(struct bcmasp_intf *intf)
+static void bcmasp_init_tx(struct bcmasp_intf *intf)
 {
-	struct device *kdev = &intf->parent->pdev->dev;
-	void *p;
-	int ret;
-
-	p = dma_alloc_coherent(kdev, DESC_RING_SIZE, &intf->tx_spb_dma_addr,
-			       GFP_KERNEL);
-	if (!p)
-		return -ENOMEM;
-
-	intf->tx_spb_cpu = p;
+	/* Restart from index 0 */
 	intf->tx_spb_dma_valid = intf->tx_spb_dma_addr + DESC_RING_SIZE - 1;
 	intf->tx_spb_dma_read = intf->tx_spb_dma_addr;
-
-	intf->tx_cbs = kcalloc(DESC_RING_COUNT, sizeof(struct bcmasp_tx_cb),
-			       GFP_KERNEL);
-	if (!intf->tx_cbs) {
-		ret = -ENOMEM;
-		goto free_tx_spb;
-	}
-
 	intf->tx_spb_index = 0;
 	intf->tx_spb_clean_index = 0;
 
-	netif_napi_add_tx(intf->ndev, &intf->tx_napi, bcmasp_tx_poll);
-
 	/* Make sure channels are disabled */
 	tx_spb_ctrl_wl(intf, 0x0, TX_SPB_CTRL_ENABLE);
 	tx_epkt_core_wl(intf, 0x0, TX_EPKT_C_CFG_MISC);
@@ -806,26 +803,6 @@ static int bcmasp_init_tx(struct bcmasp_intf *intf)
 	tx_spb_dma_wq(intf, intf->tx_spb_dma_addr, TX_SPB_DMA_BASE);
 	tx_spb_dma_wq(intf, intf->tx_spb_dma_valid, TX_SPB_DMA_END);
 	tx_spb_dma_wq(intf, intf->tx_spb_dma_valid, TX_SPB_DMA_VALID);
-
-	return 0;
-
-free_tx_spb:
-	dma_free_coherent(kdev, DESC_RING_SIZE, intf->tx_spb_cpu,
-			  intf->tx_spb_dma_addr);
-
-	return ret;
-}
-
-static void bcmasp_reclaim_free_all_tx(struct bcmasp_intf *intf)
-{
-	struct device *kdev = &intf->parent->pdev->dev;
-
-	/* Free descriptors */
-	dma_free_coherent(kdev, DESC_RING_SIZE, intf->tx_spb_cpu,
-			  intf->tx_spb_dma_addr);
-
-	/* Free cbs */
-	kfree(intf->tx_cbs);
 }
 
 static void bcmasp_ephy_enable_set(struct bcmasp_intf *intf, bool enable)
@@ -915,10 +892,7 @@ static void bcmasp_netif_deinit(struct net_device *dev)
 	bcmasp_enable_rx_irq(intf, 0);
 
 	netif_napi_del(&intf->tx_napi);
-	bcmasp_reclaim_free_all_tx(intf);
-
 	netif_napi_del(&intf->rx_napi);
-	bcmasp_reclaim_free_all_rx(intf);
 }
 
 static int bcmasp_stop(struct net_device *dev)
@@ -932,6 +906,8 @@ static int bcmasp_stop(struct net_device *dev)
 
 	bcmasp_netif_deinit(dev);
 
+	bcmasp_reclaim_free_buffers(intf);
+
 	phy_disconnect(dev->phydev);
 
 	/* Disable internal EPHY or external PHY */
@@ -1073,17 +1049,12 @@ static int bcmasp_netif_init(struct net_device *dev, bool phy_connect)
 	intf->old_link = -1;
 	intf->old_pause = -1;
 
-	ret = bcmasp_init_tx(intf);
-	if (ret)
-		goto err_phy_disconnect;
-
-	/* Turn on asp */
+	bcmasp_init_tx(intf);
+	netif_napi_add_tx(intf->ndev, &intf->tx_napi, bcmasp_tx_poll);
 	bcmasp_enable_tx(intf, 1);
 
-	ret = bcmasp_init_rx(intf);
-	if (ret)
-		goto err_reclaim_tx;
-
+	bcmasp_init_rx(intf);
+	netif_napi_add(intf->ndev, &intf->rx_napi, bcmasp_rx_poll);
 	bcmasp_enable_rx(intf, 1);
 
 	/* Turn on UniMAC TX/RX */
@@ -1097,12 +1068,6 @@ static int bcmasp_netif_init(struct net_device *dev, bool phy_connect)
 
 	return 0;
 
-err_reclaim_tx:
-	netif_napi_del(&intf->tx_napi);
-	bcmasp_reclaim_free_all_tx(intf);
-err_phy_disconnect:
-	if (phydev)
-		phy_disconnect(phydev);
 err_phy_disable:
 	if (intf->internal_phy)
 		bcmasp_ephy_enable_set(intf, false);
@@ -1118,6 +1083,10 @@ static int bcmasp_open(struct net_device *dev)
 
 	netif_dbg(intf, ifup, dev, "bcmasp open\n");
 
+	ret = bcmasp_alloc_buffers(intf);
+	if (ret)
+		return ret;
+
 	ret = clk_prepare_enable(intf->parent->clk);
 	if (ret)
 		return ret;
-- 
2.34.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

Re: [PATCH net-next resend 5/6] net: bcmasp: Keep buffers through power management

[Florian Fainelli]

[-- Attachment #1: Type: text/plain, Size: 409 bytes --]

On 2/23/24 14:24, Justin Chen wrote:
> There is no advantage of freeing and re-allocating buffers through
> suspend and resume. This waste cycles and makes suspend/resume time
> longer. We also open ourselves to failed allocations in systems with
> heavy memory fragmentation.
> 
> Signed-off-by: Justin Chen <justin.chen@broadcom.com>

Acked-by: Florian Fainelli <florian.fainelli@broadcom.com>
-- 
Florian


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4221 bytes --]

[PATCH net-next resend 6/6] net: bcmasp: Add support for PHY interrupts

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 3461 bytes --]

Hook up the phy interrupts for internal phys to reduce mdio traffic
and improve responsiveness of link changes.

Signed-off-by: Justin Chen <justin.chen@broadcom.com>
---
 drivers/net/ethernet/broadcom/asp2/bcmasp.c     | 17 +++++++++++++++++
 drivers/net/ethernet/broadcom/asp2/bcmasp.h     |  4 ++++
 .../net/ethernet/broadcom/asp2/bcmasp_intf.c    |  5 +++++
 3 files changed, 26 insertions(+)

diff --git a/drivers/net/ethernet/broadcom/asp2/bcmasp.c b/drivers/net/ethernet/broadcom/asp2/bcmasp.c
index 100c69f3307a..a806dadc4196 100644
--- a/drivers/net/ethernet/broadcom/asp2/bcmasp.c
+++ b/drivers/net/ethernet/broadcom/asp2/bcmasp.c
@@ -31,6 +31,20 @@ static void _intr2_mask_set(struct bcmasp_priv *priv, u32 mask)
 	priv->irq_mask |= mask;
 }
 
+void bcmasp_enable_phy_irq(struct bcmasp_intf *intf, int en)
+{
+	struct bcmasp_priv *priv = intf->parent;
+
+	/* Only supported with internal phys */
+	if (!intf->internal_phy)
+		return;
+
+	if (en)
+		_intr2_mask_clear(priv, ASP_INTR2_PHY_EVENT(intf->channel));
+	else
+		_intr2_mask_set(priv, ASP_INTR2_PHY_EVENT(intf->channel));
+}
+
 void bcmasp_enable_tx_irq(struct bcmasp_intf *intf, int en)
 {
 	struct bcmasp_priv *priv = intf->parent;
@@ -79,6 +93,9 @@ static void bcmasp_intr2_handling(struct bcmasp_intf *intf, u32 status)
 			__napi_schedule_irqoff(&intf->tx_napi);
 		}
 	}
+
+	if (status & ASP_INTR2_PHY_EVENT(intf->channel))
+		phy_mac_interrupt(intf->ndev->phydev);
 }
 
 static irqreturn_t bcmasp_isr(int irq, void *data)
diff --git a/drivers/net/ethernet/broadcom/asp2/bcmasp.h b/drivers/net/ethernet/broadcom/asp2/bcmasp.h
index 127a5340625e..f93cb3da44b0 100644
--- a/drivers/net/ethernet/broadcom/asp2/bcmasp.h
+++ b/drivers/net/ethernet/broadcom/asp2/bcmasp.h
@@ -19,6 +19,8 @@
 #define ASP_INTR2_TX_DESC(intr)			BIT((intr) + 14)
 #define ASP_INTR2_UMC0_WAKE			BIT(22)
 #define ASP_INTR2_UMC1_WAKE			BIT(28)
+#define ASP_INTR2_PHY_EVENT(intr)		((intr) ? BIT(30) | BIT(31) : \
+						BIT(24) | BIT(25))
 
 #define ASP_WAKEUP_INTR2_OFFSET			0x1200
 #define  ASP_WAKEUP_INTR2_STATUS		0x0
@@ -556,6 +558,8 @@ void bcmasp_enable_tx_irq(struct bcmasp_intf *intf, int en);
 
 void bcmasp_enable_rx_irq(struct bcmasp_intf *intf, int en);
 
+void bcmasp_enable_phy_irq(struct bcmasp_intf *intf, int en);
+
 void bcmasp_flush_rx_port(struct bcmasp_intf *intf);
 
 extern const struct ethtool_ops bcmasp_ethtool_ops;
diff --git a/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c b/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
index 0b378a6d43e7..8fbeb506abb9 100644
--- a/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
+++ b/drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
@@ -382,6 +382,7 @@ static void bcmasp_netif_start(struct net_device *dev)
 
 	bcmasp_enable_rx_irq(intf, 1);
 	bcmasp_enable_tx_irq(intf, 1);
+	bcmasp_enable_phy_irq(intf, 1);
 
 	phy_start(dev->phydev);
 }
@@ -890,6 +891,7 @@ static void bcmasp_netif_deinit(struct net_device *dev)
 	/* Disable interrupts */
 	bcmasp_enable_tx_irq(intf, 0);
 	bcmasp_enable_rx_irq(intf, 0);
+	bcmasp_enable_phy_irq(intf, 0);
 
 	netif_napi_del(&intf->tx_napi);
 	netif_napi_del(&intf->rx_napi);
@@ -1028,6 +1030,9 @@ static int bcmasp_netif_init(struct net_device *dev, bool phy_connect)
 			goto err_phy_disable;
 		}
 
+		if (intf->internal_phy)
+			dev->phydev->irq = PHY_MAC_INTERRUPT;
+
 		/* Indicate that the MAC is responsible for PHY PM */
 		phydev->mac_managed_pm = true;
 	} else if (!intf->wolopts) {
-- 
2.34.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

Re: [PATCH net-next resend 6/6] net: bcmasp: Add support for PHY interrupts

[Florian Fainelli]

[-- Attachment #1: Type: text/plain, Size: 285 bytes --]

On 2/23/24 14:24, Justin Chen wrote:
> Hook up the phy interrupts for internal phys to reduce mdio traffic
> and improve responsiveness of link changes.
> 
> Signed-off-by: Justin Chen <justin.chen@broadcom.com>

Acked-by: Florian Fainelli <florian.fainelli@broadcom.com>
-- 
Florian


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4221 bytes --]

Re: [PATCH net-next resend 0/6] Support for ASP 2.2 and optimizations

[Krzysztof Kozlowski]

On 23/02/2024 23:24, Justin Chen wrote:
> ASP 2.2 adds some power savings during low power modes.
> 
> Also make various improvements when entering low power modes and
> reduce MDIO traffic by hooking up interrupts.

Why are you resending?

Why did you ignore tags/feedback?

Best regards,
Krzysztof

Re: [PATCH net-next resend 0/6] Support for ASP 2.2 and optimizations

[Justin Chen]

[-- Attachment #1: Type: text/plain, Size: 716 bytes --]



On 2/24/24 2:23 AM, Krzysztof Kozlowski wrote:
> On 23/02/2024 23:24, Justin Chen wrote:
>> ASP 2.2 adds some power savings during low power modes.
>>
>> Also make various improvements when entering low power modes and
>> reduce MDIO traffic by hooking up interrupts.
> 
> Why are you resending?
> 
> Why did you ignore tags/feedback?
> 

I think I misunderstood the resend tag. The feedback was I needed to 
rebase onto a newer net-next to resolve a merge conflict. I addressed 
this feedback, but this should have been v2 and not resend. Apologies. 
Totally forgot about Florian's tags. Will send a v2 after we resolve the 
current comments with the device tree.

Thanks,
Justin


> Best regards,
> Krzysztof
> 

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]