Skip to contentSkip to conversation

net: phy: mediatek: permit to compile test GE SOC PHY driver

PATCH Applied Christian Marangi 4 participants 11 messages Apr 10, 2025 – Apr 14, 2025

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/2 net: phy: mediatek: permit to compile test GE SOC PHY driver

When commit 462a3daad679 ("net: phy: mediatek: fix compile-test dependencies") fixed the dependency, it should have also introduced an or on COMPILE_TEST to permit this driver to be compile-tested even if NVMEM_MTK_EFUSE wasn't selected. The driver makes use of NVMEM API that are always compiled (return error) so the driver can actually be compiled even without that config.

Fix and simplify the dependency condition of this kernel config.

Fixes 462a3daad679 ("net: phy: mediatek: fix compile-test dependencies")

---
Changes v2:
- Add Ack and Reviewed-by tag
- Address suggested dependency from Russell

 drivers/net/phy/mediatek/Kconfig | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/net/phy/mediatek/Kconfig +1 -2
index 2a8ac5aed0f8..6a4c2b328c41 100644
@@ -15,8 +15,7 @@ config MEDIATEK_GE_PHY
 
 config MEDIATEK_GE_SOC_PHY
 	tristate "MediaTek SoC Ethernet PHYs"
-	depends on (ARM64 && ARCH_MEDIATEK) || COMPILE_TEST
-	depends on NVMEM_MTK_EFUSE
+	depends on (ARM64 && ARCH_MEDIATEK && NVMEM_MTK_EFUSE) || COMPILE_TEST
 	select MTK_NET_PHYLIB
 	help
 	  Supports MediaTek SoC built-in Gigabit Ethernet PHYs.
Signature · 1 line
2.48.1
#1 lore raw

PATCH v2 2/2 net: phy: mediatek: add Airoha PHY ID to SoC driver

Airoha AN7581 SoC ship with a Switch based on the MT753x Switch embedded in other SoC like the MT7581 and the MT7988. Similar to these they require configuring some pin to enable LED PHYs.

Add support for the PHY ID for the Airoha embedded Switch and define a simple probe function to toggle these pins. Also fill the LED functions and add dedicated function to define LED polarity.

---
Changes v2:
- Add Reviewed-by tag
- Address suggested dependency from Russell

 drivers/net/phy/mediatek/Kconfig      |  4 +-
 drivers/net/phy/mediatek/mtk-ge-soc.c | 62 +++++++++++++++++++++++++++
 2 files changed, 65 insertions(+), 1 deletion(-)
drivers/net/phy/mediatek/Kconfig +3 -1
index 6a4c2b328c41..4308002bb82c 100644
@@ -15,7 +15,9 @@ config MEDIATEK_GE_PHY
 
 config MEDIATEK_GE_SOC_PHY
 	tristate "MediaTek SoC Ethernet PHYs"
-	depends on (ARM64 && ARCH_MEDIATEK && NVMEM_MTK_EFUSE) || COMPILE_TEST
+	depends on ARM64 || COMPILE_TEST
+	depends on ARCH_AIROHA || (ARCH_MEDIATEK && NVMEM_MTK_EFUSE) || \
+		   COMPILE_TEST
 	select MTK_NET_PHYLIB
 	help
 	  Supports MediaTek SoC built-in Gigabit Ethernet PHYs.
drivers/net/phy/mediatek/mtk-ge-soc.c +62 -0

88 lines not on this page · raw message

Signature · 1 line
2.48.1
#2 lore raw

Re: [net-next PATCH v2 2/2] net: phy: mediatek: add Airoha PHY ID to SoC driver

drivers/net/phy/mediatek/Kconfig config MEDIATEK_GE_PHY · line 23

Quoted 10 lines · Christian Marangi

config MEDIATEK_GE_SOC_PHY

drivers/net/phy/mediatek/Kconfig +3 -1
 	tristate "MediaTek SoC Ethernet PHYs"
-	depends on (ARM64 && ARCH_MEDIATEK && NVMEM_MTK_EFUSE) || COMPILE_TEST
+	depends on ARM64 || COMPILE_TEST
+	depends on ARCH_AIROHA || (ARCH_MEDIATEK && NVMEM_MTK_EFUSE) || \
+		   COMPILE_TEST
 	select MTK_NET_PHYLIB
 	help
 	  Supports MediaTek SoC built-in Gigabit Ethernet PHYs.

This now also fails for non-compile-test builds with NVMEM=m, ARCH_MEDIATEK=n, ARCH_AIROHA=y and MEDIATEK_GE_SOC_PHY=y.

Arnd

reviewer
#3 lore raw

Re: [net-next PATCH v2 2/2] net: phy: mediatek: add Airoha PHY ID to SoC driver

In reply to Christian Marangi
Quoted 10 lines · Christian Marangi

Airoha AN7581 SoC ship with a Switch based on the MT753x Switch embedded in other SoC like the MT7581 and the MT7988. Similar to these they require configuring some pin to enable LED PHYs.

Add support for the PHY ID for the Airoha embedded Switch and define a simple probe function to toggle these pins. Also fill the LED functions and add dedicated function to define LED polarity.

...

Quoted 1 line
drivers/net/phy/mediatek/mtk-ge-soc.c +0 -0

...

Quoted 15 lines
drivers/net/phy/mediatek/mtk-ge-soc.c +15 -0
+static int an7581_phy_led_polarity_set(struct phy_device *phydev, int index,
+				       unsigned long modes)
+{
+	u32 mode;
+	u16 val;
+
+	if (index >= MTK_PHY_MAX_LEDS)
+		return -EINVAL;
+
+	for_each_set_bit(mode, &modes, __PHY_LED_MODES_NUM) {
+		switch (mode) {
+		case PHY_LED_ACTIVE_LOW:
+			val = MTK_PHY_LED_ON_POLARITY;
+			break;
+		case PHY_LED_ACTIVE_HIGH:

drivers/net/phy/mediatek/mtk-ge-soc.c static int mt7981_phy_probe(struct phy_device *phydev) · line 1456

Quoted 10 lines
drivers/net/phy/mediatek/mtk-ge-soc.c +10 -0
+			val = 0;
+			break;
+		default:
+			return -EINVAL;
+		}
+	}
+
+	return phy_modify_mmd(phydev, MDIO_MMD_VEND2, index ?
+			      MTK_PHY_LED1_ON_CTRL : MTK_PHY_LED0_ON_CTRL,
+			      MTK_PHY_LED_ON_POLARITY, val);

Hi Christian,

Perhaps this cannot occur in practice, but if the for_each_set_bit loop iterates zero times then val will be used uninitialised here.

Flagged by Smatch.

Quoted 1 line

+}

...

#4 lore raw

Re: [net-next PATCH v2 2/2] net: phy: mediatek: add Airoha PHY ID to SoC driver

Quoted 46 lines · Simon Horman
Quoted 10 lines · Christian Marangi

Airoha AN7581 SoC ship with a Switch based on the MT753x Switch embedded in other SoC like the MT7581 and the MT7988. Similar to these they require configuring some pin to enable LED PHYs.

Add support for the PHY ID for the Airoha embedded Switch and define a simple probe function to toggle these pins. Also fill the LED functions and add dedicated function to define LED polarity.

...

Quoted 1 line
drivers/net/phy/mediatek/mtk-ge-soc.c +0 -0

...

Quoted 15 lines
drivers/net/phy/mediatek/mtk-ge-soc.c +15 -0
+static int an7581_phy_led_polarity_set(struct phy_device *phydev, int index,
+				       unsigned long modes)
+{
+	u32 mode;
+	u16 val;
+
+	if (index >= MTK_PHY_MAX_LEDS)
+		return -EINVAL;
+
+	for_each_set_bit(mode, &modes, __PHY_LED_MODES_NUM) {
+		switch (mode) {
+		case PHY_LED_ACTIVE_LOW:
+			val = MTK_PHY_LED_ON_POLARITY;
+			break;
+		case PHY_LED_ACTIVE_HIGH:

drivers/net/phy/mediatek/mtk-ge-soc.c static int mt7981_phy_probe(struct phy_device *phydev) · line 1456

Quoted 10 lines
drivers/net/phy/mediatek/mtk-ge-soc.c +10 -0
+			val = 0;
+			break;
+		default:
+			return -EINVAL;
+		}
+	}
+
+	return phy_modify_mmd(phydev, MDIO_MMD_VEND2, index ?
+			      MTK_PHY_LED1_ON_CTRL : MTK_PHY_LED0_ON_CTRL,
+			      MTK_PHY_LED_ON_POLARITY, val);

Hi Christian,

Quoted 8 lines

Perhaps this cannot occur in practice, but if the for_each_set_bit loop iterates zero times then val will be used uninitialised here.

Flagged by Smatch.

Quoted 1 line

+}

...

Almost impossible but yes I will post a follow-up patch fixing this!

Signature · 1 line
	Ansuel
#5 lore raw
In reply to Christian Marangi
Quoted 4 lines · Christian Marangi

When commit 462a3daad679 ("net: phy: mediatek: fix compile-test dependencies") fixed the dependency, it should have also introduced an or on COMPILE_TEST to permit this driver to be compile-tested even if NVMEM_MTK_EFUSE wasn't selected

Why does this matter? NVMEM_MTK_EFUSE can be enabled for both allmodconfig and randconfig builds on any architecture, so you get build coverage either way, it's just a little less likely to be enabled in randconfig I guess?

Quoted 6 lines
drivers/net/phy/mediatek/Kconfig +0 -0
index 2a8ac5aed0f8..6a4c2b328c41 100644

drivers/net/phy/mediatek/Kconfig config MEDIATEK_GE_PHY · line 21

Quoted 9 lines
drivers/net/phy/mediatek/Kconfig +1 -2
@@ -15,8 +15,7 @@ config MEDIATEK_GE_PHY
 
 config MEDIATEK_GE_SOC_PHY
 	tristate "MediaTek SoC Ethernet PHYs"
-	depends on (ARM64 && ARCH_MEDIATEK) || COMPILE_TEST
-	depends on NVMEM_MTK_EFUSE
+	depends on (ARM64 && ARCH_MEDIATEK && NVMEM_MTK_EFUSE) || COMPILE_TEST
 	select MTK_NET_PHYLIB
 	help
 	  Supports MediaTek SoC built-in Gigabit Ethernet PHYs.

I would expect this to break the build with CONFIG_NVMEM=m and MEDIATEK_GE_SOC_PHY=y.

The normal thing here would be to have a dependency on CONFIG_NVMEM in place of the NVMEM_MTK_EFUSE dependency, or possible on 'NVMEM || !NVMEM' if you want to make it more likely to be enabled in randconfig builds.

Arnd

#6 lore raw
Quoted 11 lines · Arnd Bergmann
Quoted 4 lines · Christian Marangi

When commit 462a3daad679 ("net: phy: mediatek: fix compile-test dependencies") fixed the dependency, it should have also introduced an or on COMPILE_TEST to permit this driver to be compile-tested even if NVMEM_MTK_EFUSE wasn't selected

Why does this matter? NVMEM_MTK_EFUSE can be enabled for both allmodconfig and randconfig builds on any architecture, so you get build coverage either way, it's just a little less likely to be enabled in randconfig I guess?

If we base stuff on the fact that everything is selected or that a random config by luck selects it, then COMPILE_TEST doesn't make sense at all.

For my personal test, I wanted to test the driver on a simple x86 build without having to depend on ARCH or having to cross compile. Won't happen on real world scenario? Totally. I should be able to compile it? Yes.

Quoted 19 lines
Quoted 6 lines
drivers/net/phy/mediatek/Kconfig +0 -0
index 2a8ac5aed0f8..6a4c2b328c41 100644

drivers/net/phy/mediatek/Kconfig config MEDIATEK_GE_PHY · line 21

Quoted 9 lines
drivers/net/phy/mediatek/Kconfig +1 -2
@@ -15,8 +15,7 @@ config MEDIATEK_GE_PHY
 
 config MEDIATEK_GE_SOC_PHY
 	tristate "MediaTek SoC Ethernet PHYs"
-	depends on (ARM64 && ARCH_MEDIATEK) || COMPILE_TEST
-	depends on NVMEM_MTK_EFUSE
+	depends on (ARM64 && ARCH_MEDIATEK && NVMEM_MTK_EFUSE) || COMPILE_TEST
 	select MTK_NET_PHYLIB
 	help
 	  Supports MediaTek SoC built-in Gigabit Ethernet PHYs.

I would expect this to break the build with CONFIG_NVMEM=m and MEDIATEK_GE_SOC_PHY=y.

Quoted 5 lines

The normal thing here would be to have a dependency on CONFIG_NVMEM in place of the NVMEM_MTK_EFUSE dependency, or possible on 'NVMEM || !NVMEM' if you want to make it more likely to be enabled in randconfig builds.

The big idea of these dependency is that... In MTK the internal PHY of the switch needs calibration or it won't work hence it doesn't make sense to select the PHY as it won't ever work without the NVMEM driver.

But from a compile test view where we only evaluate if the driver have compilation error or other kind of warning, we should not care...

Signature · 1 line
	Ansuel
#7 lore raw
Quoted 47 lines · Christian Marangi
Quoted 11 lines · Arnd Bergmann
Quoted 4 lines · Christian Marangi

When commit 462a3daad679 ("net: phy: mediatek: fix compile-test dependencies") fixed the dependency, it should have also introduced an or on COMPILE_TEST to permit this driver to be compile-tested even if NVMEM_MTK_EFUSE wasn't selected

Why does this matter? NVMEM_MTK_EFUSE can be enabled for both allmodconfig and randconfig builds on any architecture, so you get build coverage either way, it's just a little less likely to be enabled in randconfig I guess?

If we base stuff on the fact that everything is selected or that a random config by luck selects it, then COMPILE_TEST doesn't make sense at all.

For my personal test, I wanted to test the driver on a simple x86 build without having to depend on ARCH or having to cross compile. Won't happen on real world scenario? Totally. I should be able to compile it? Yes.

Quoted 19 lines
Quoted 6 lines
drivers/net/phy/mediatek/Kconfig +0 -0
index 2a8ac5aed0f8..6a4c2b328c41 100644

drivers/net/phy/mediatek/Kconfig config MEDIATEK_GE_PHY · line 21

Quoted 9 lines
drivers/net/phy/mediatek/Kconfig +1 -2
@@ -15,8 +15,7 @@ config MEDIATEK_GE_PHY
 
 config MEDIATEK_GE_SOC_PHY
 	tristate "MediaTek SoC Ethernet PHYs"
-	depends on (ARM64 && ARCH_MEDIATEK) || COMPILE_TEST
-	depends on NVMEM_MTK_EFUSE
+	depends on (ARM64 && ARCH_MEDIATEK && NVMEM_MTK_EFUSE) || COMPILE_TEST
 	select MTK_NET_PHYLIB
 	help
 	  Supports MediaTek SoC built-in Gigabit Ethernet PHYs.

I would expect this to break the build with CONFIG_NVMEM=m and MEDIATEK_GE_SOC_PHY=y.

Quoted 5 lines

The normal thing here would be to have a dependency on CONFIG_NVMEM in place of the NVMEM_MTK_EFUSE dependency, or possible on 'NVMEM || !NVMEM' if you want to make it more likely to be enabled in randconfig builds.

Quoted 7 lines

The big idea of these dependency is that... In MTK the internal PHY of the switch needs calibration or it won't work hence it doesn't make sense to select the PHY as it won't ever work without the NVMEM driver.

But from a compile test view where we only evaluate if the driver have compilation error or other kind of warning, we should not care...

Also 99% I could be wrong but from what I can see in NVMEM kconfig, NVMEM is not tristate but only bool? So NVMEM=m is not a thing?

Signature · 1 line
	Ansuel
Acked-by
#8 lore raw
Quoted 4 lines · Christian Marangi

On Thu, Apr 10, 2025 at 12:40:15PM +0200, Christian Marangi wrote:

Also 99% I could be wrong but from what I can see in NVMEM kconfig, NVMEM is not tristate but only bool? So NVMEM=m is not a thing?

Ah, I missed that. In this case your patches are both fine.

#9 lore raw
In reply to Christian Marangi
Quoted 13 lines · Christian Marangi
Quoted 11 lines · Arnd Bergmann
Quoted 4 lines · Christian Marangi

When commit 462a3daad679 ("net: phy: mediatek: fix compile-test dependencies") fixed the dependency, it should have also introduced an or on COMPILE_TEST to permit this driver to be compile-tested even if NVMEM_MTK_EFUSE wasn't selected

Why does this matter? NVMEM_MTK_EFUSE can be enabled for both allmodconfig and randconfig builds on any architecture, so you get build coverage either way, it's just a little less likely to be enabled in randconfig I guess?

Quoted 8 lines

If we base stuff on the fact that everything is selected or that a random config by luck selects it, then COMPILE_TEST doesn't make sense at all.

For my personal test, I wanted to test the driver on a simple x86 build without having to depend on ARCH or having to cross compile. Won't happen on real world scenario? Totally. I should be able to compile it? Yes.

Being able to compile test the driver yourself is clearly a good idea, I just don't think that requires the dependency to be conditional here.

Quoted 9 lines
Quoted 8 lines

I would expect this to break the build with CONFIG_NVMEM=m and MEDIATEK_GE_SOC_PHY=y.

The normal thing here would be to have a dependency on CONFIG_NVMEM in place of the NVMEM_MTK_EFUSE dependency, or possible on 'NVMEM || !NVMEM' if you want to make it more likely to be enabled in randconfig builds.

Quoted 3 lines

The big idea of these dependency is that... In MTK the internal PHY of the switch needs calibration or it won't work hence it doesn't make sense to select the PHY as it won't ever work without the NVMEM driver.

That's not how dependencies normally work: the driver also fails to work correctly if you are missing the correct pinctrl or led driver, or the syscon, yet you don't list them explicitly because you just need them for the machine to work correctly.

The driver dependencies should just list what you need to build, regardless of whether you are compile-testing or building a driver that is going to be used.

Arnd

#10 lore raw
In reply to Christian Marangi
This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This series was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Quoted 8 lines · On Thu, 10 Apr 2025 12:04:03 +0200 you wrote:

When commit 462a3daad679 ("net: phy: mediatek: fix compile-test dependencies") fixed the dependency, it should have also introduced an or on COMPILE_TEST to permit this driver to be compile-tested even if NVMEM_MTK_EFUSE wasn't selected. The driver makes use of NVMEM API that are always compiled (return error) so the driver can actually be compiled even without that config.

[...]

Here is the summary with links:
- [net-next,v2,1/2] net: phy: mediatek: permit to compile test GE SOC PHY driver

    https://git.kernel.org/netdev/net-next/c/e5566162af8b
  - [net-next,v2,2/2] net: phy: mediatek: add Airoha PHY ID to SoC driver
    https://git.kernel.org/netdev/net-next/c/6a325aed130b

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html