Skip to contentSkip to conversation

i2c: designware: Clean up MODEL_* definitions

PATCH In review Andy Shevchenko 3 participants 12 messages Jan 14 – Jan 22

Subsystems (3)
lore
Share & tools
lore

Andy Shevchenko

author reviewer
#0 lore raw

cover i2c: designware: Clean up MODEL_* definitions

The current way of recognizing models is not scalable and in most of the cases just a replacement of well established device_is_compatible() approach.

For now take the low-hanging fruits and partially clean up those MODEL_* definitions.

Andy Shevchenko (2):
i2c: designware: Remove not-going-to-be-supported code for Baikal SoC
i2c: designware: Use device_is_compatible() instead of custom approach

 drivers/i2c/busses/Kconfig                  |  1 -
 drivers/i2c/busses/i2c-designware-common.c  |  6 +-
 drivers/i2c/busses/i2c-designware-core.h    |  2 -
 drivers/i2c/busses/i2c-designware-platdrv.c | 70 +--------------------
 4 files changed, 2 insertions(+), 77 deletions(-)
Signature · 1 line
2.50.1
author reviewer
#1 lore raw

PATCH 1/2 i2c: designware: Remove not-going-to-be-supported code for Baikal SoC

As noticed in the discussion [1] the Baikal SoC and platforms are not going to be finalized, hence remove stale code.

---
 drivers/i2c/busses/Kconfig                  |  1 -
 drivers/i2c/busses/i2c-designware-core.h    |  1 -
 drivers/i2c/busses/i2c-designware-platdrv.c | 68 ---------------------
 3 files changed, 70 deletions(-)
drivers/i2c/busses/Kconfig +0 -1
index 860812e224a0..e11d50750e63 100644
@@ -580,7 +580,6 @@ if I2C_DESIGNWARE_CORE
 config I2C_DESIGNWARE_PLATFORM
 	tristate "Synopsys DesignWare Platform driver"
 	depends on (ACPI && COMMON_CLK) || !ACPI
-	select MFD_SYSCON if MIPS_BAIKAL_T1
 	default I2C_DESIGNWARE_CORE
 	help
 	  If you say yes to this option, support will be included for the
drivers/i2c/busses/i2c-designware-core.h +0 -1
index 2a7decc24931..cf0364079b55 100644
@@ -314,7 +314,6 @@ struct dw_i2c_dev {
 #define ACCESS_POLLING				BIT(3)
 
 #define MODEL_MSCC_OCELOT			BIT(8)
-#define MODEL_BAIKAL_BT1			BIT(9)
 #define MODEL_AMD_NAVI_GPU			BIT(10)
 #define MODEL_WANGXUN_SP			BIT(11)
 #define MODEL_MASK				GENMASK(11, 8)
drivers/i2c/busses/i2c-designware-platdrv.c +0 -68

86 lines not on this page · raw message

Signature · 1 line
2.50.1
maintainer
#3 lore raw

Re: [PATCH v1 1/2] i2c: designware: Remove not-going-to-be-supported code for Baikal SoC

In reply to Andy Shevchenko
Quoted 5 lines · Andy Shevchenko

As noticed in the discussion [1] the Baikal SoC and platforms are not going to be finalized, hence remove stale code.

Until you guys sort things out with patch 2, I'm applying patch 1.

Thanks,
Andi

author reviewer
#4 lore raw

PATCH 2/2 i2c: designware: Use device_is_compatible() instead of custom approach

In reply to Andy Shevchenko

We use MODEL_MSCC_OCELOT effectively is a flag for comparing against "compatible" property. Use device_is_compatible() directly to make it clear.

---
 drivers/i2c/busses/i2c-designware-common.c  | 6 +-----
 drivers/i2c/busses/i2c-designware-core.h    | 1 -
 drivers/i2c/busses/i2c-designware-platdrv.c | 2 +-
 3 files changed, 2 insertions(+), 7 deletions(-)
drivers/i2c/busses/i2c-designware-common.c +1 -5
index 5b1e8f74c4ac..c766d9821975 100644
@@ -238,14 +238,10 @@ static void i2c_dw_of_configure(struct device *device)
 	struct platform_device *pdev = to_platform_device(device);
 	struct dw_i2c_dev *dev = dev_get_drvdata(device);
 
-	switch (dev->flags & MODEL_MASK) {
-	case MODEL_MSCC_OCELOT:
+	if (device_is_compatible(dev->dev, "mscc,ocelot-i2c")) {
 		dev->ext = devm_platform_ioremap_resource(pdev, 1);
 		if (!IS_ERR(dev->ext))
 			dev->set_sda_hold_time = mscc_twi_set_sda_hold_time;
-		break;
-	default:
-		break;
 	}
 }
 
drivers/i2c/busses/i2c-designware-core.h +0 -1
index cf0364079b55..10055f0e0ec3 100644
@@ -313,7 +313,6 @@ struct dw_i2c_dev {
 #define ARBITRATION_SEMAPHORE			BIT(2)
 #define ACCESS_POLLING				BIT(3)
 
-#define MODEL_MSCC_OCELOT			BIT(8)
 #define MODEL_AMD_NAVI_GPU			BIT(10)
 #define MODEL_WANGXUN_SP			BIT(11)
 #define MODEL_MASK				GENMASK(11, 8)
drivers/i2c/busses/i2c-designware-platdrv.c +1 -1
index 2e532f16691b..4e6fe3b55322 100644
@@ -267,7 +267,7 @@ static void dw_i2c_plat_remove(struct platform_device *pdev)
 }
 
 static const struct of_device_id dw_i2c_of_match[] = {
-	{ .compatible = "mscc,ocelot-i2c", .data = (void *)MODEL_MSCC_OCELOT },
+	{ .compatible = "mscc,ocelot-i2c" },
 	{ .compatible = "snps,designware-i2c" },
 	{}
 };
Signature · 1 line
2.50.1
maintainer
#5 lore raw

Re: [PATCH v1 2/2] i2c: designware: Use device_is_compatible() instead of custom approach

Quoted 1 line · Andy Shevchenko

We use MODEL_MSCC_OCELOT effectively is a flag for comparing against

as a flag?

Quoted 2 lines

"compatible" property. Use device_is_compatible() directly to make it clear.

Okay but if something else ever needs this same quirk then we would need to add new entry here and also to the IDs list.

With the flag you can have the IDs in a single place not all over the driver.

Quoted 55 lines
---
 drivers/i2c/busses/i2c-designware-common.c  | 6 +-----
 drivers/i2c/busses/i2c-designware-core.h    | 1 -
 drivers/i2c/busses/i2c-designware-platdrv.c | 2 +-
 3 files changed, 2 insertions(+), 7 deletions(-)
drivers/i2c/busses/i2c-designware-common.c +1 -5
index 5b1e8f74c4ac..c766d9821975 100644
@@ -238,14 +238,10 @@ static void i2c_dw_of_configure(struct device *device)
 	struct platform_device *pdev = to_platform_device(device);
 	struct dw_i2c_dev *dev = dev_get_drvdata(device);
 
-	switch (dev->flags & MODEL_MASK) {
-	case MODEL_MSCC_OCELOT:
+	if (device_is_compatible(dev->dev, "mscc,ocelot-i2c")) {
 		dev->ext = devm_platform_ioremap_resource(pdev, 1);
 		if (!IS_ERR(dev->ext))
 			dev->set_sda_hold_time = mscc_twi_set_sda_hold_time;
-		break;
-	default:
-		break;
 	}
 }
 
drivers/i2c/busses/i2c-designware-core.h +0 -1
index cf0364079b55..10055f0e0ec3 100644
@@ -313,7 +313,6 @@ struct dw_i2c_dev {
 #define ARBITRATION_SEMAPHORE			BIT(2)
 #define ACCESS_POLLING				BIT(3)
 
-#define MODEL_MSCC_OCELOT			BIT(8)
 #define MODEL_AMD_NAVI_GPU			BIT(10)
 #define MODEL_WANGXUN_SP			BIT(11)
 #define MODEL_MASK				GENMASK(11, 8)
drivers/i2c/busses/i2c-designware-platdrv.c +1 -1
index 2e532f16691b..4e6fe3b55322 100644
@@ -267,7 +267,7 @@ static void dw_i2c_plat_remove(struct platform_device *pdev)
 }
 
 static const struct of_device_id dw_i2c_of_match[] = {
-	{ .compatible = "mscc,ocelot-i2c", .data = (void *)MODEL_MSCC_OCELOT },
+	{ .compatible = "mscc,ocelot-i2c" },
 	{ .compatible = "snps,designware-i2c" },
 	{}
 };
Signature · 1 line
2.50.1
author reviewer
#6 lore raw

Re: [PATCH v1 2/2] i2c: designware: Use device_is_compatible() instead of custom approach

Quoted 4 lines · Mika Westerberg
Quoted 1 line · Andy Shevchenko

We use MODEL_MSCC_OCELOT effectively is a flag for comparing against

as a flag?

Yes, sorry for the typo.

Quoted 5 lines
Quoted 2 lines

"compatible" property. Use device_is_compatible() directly to make it clear.

Okay but if something else ever needs this same quirk then we would need to add new entry here and also to the IDs list.

Yes, that's how DT works and there are, of course, examples all over the kernel, first that comes to my mind: drivers/mmc/host/sdhci-pltfm.c.

Quoted 2 lines

With the flag you can have the IDs in a single place not all over the driver.

Makes the reality harder to read. If I know that the same quirk is used by different platform (in terms of compatible string) I will see it immediately from the code. Flag is meaningless.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
maintainer
#7 lore raw

Re: [PATCH v1 2/2] i2c: designware: Use device_is_compatible() instead of custom approach

Quoted 17 lines · Andy Shevchenko
Quoted 4 lines · Mika Westerberg
Quoted 1 line · Andy Shevchenko

We use MODEL_MSCC_OCELOT effectively is a flag for comparing against

as a flag?

Yes, sorry for the typo.

Quoted 5 lines
Quoted 2 lines

"compatible" property. Use device_is_compatible() directly to make it clear.

Okay but if something else ever needs this same quirk then we would need to add new entry here and also to the IDs list.

Yes, that's how DT works and there are, of course, examples all over the kernel, first that comes to my mind: drivers/mmc/host/sdhci-pltfm.c.

Quoted 6 lines
Quoted 2 lines

With the flag you can have the IDs in a single place not all over the driver.

Makes the reality harder to read. If I know that the same quirk is used by different platform (in terms of compatible string) I will see it immediately from the code. Flag is meaningless.

Well with the flag you get the help from the compiler if you typo it but with the string comparison you are on your own. Therefore I prefer the flag and as I said it also avoid duplicating the compatible string.

author reviewer
#8 lore raw

Re: [PATCH v1 2/2] i2c: designware: Use device_is_compatible() instead of custom approach

Quoted 21 lines · Mika Westerberg
Quoted 20 lines · Andy Shevchenko
Quoted 4 lines · Mika Westerberg
Quoted 1 line · Andy Shevchenko

We use MODEL_MSCC_OCELOT effectively is a flag for comparing against

as a flag?

Yes, sorry for the typo.

Quoted 5 lines
Quoted 2 lines

"compatible" property. Use device_is_compatible() directly to make it clear.

Okay but if something else ever needs this same quirk then we would need to add new entry here and also to the IDs list.

Yes, that's how DT works and there are, of course, examples all over the kernel, first that comes to my mind: drivers/mmc/host/sdhci-pltfm.c.

Quoted 2 lines

With the flag you can have the IDs in a single place not all over the driver.

Quoted 6 lines
Quoted 3 lines

Makes the reality harder to read. If I know that the same quirk is used by different platform (in terms of compatible string) I will see it immediately from the code. Flag is meaningless.

Well with the flag you get the help from the compiler if you typo it but with the string comparison you are on your own.

It only affects the development cycle, isn't it obvious that any quirk should be tested by the author, otherwise it won't work.

Quoted 2 lines

Therefore I prefer the flag and as I said it also avoid duplicating the compatible string.

Duplicating compatible string is not a problem, linker nowadays combines the same strings into one. In case you are worrying about the space, it's not a problem.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
maintainer
#9 lore raw

Re: [PATCH v1 2/2] i2c: designware: Use device_is_compatible() instead of custom approach

Quoted 26 lines · Andy Shevchenko
Quoted 25 lines · Mika Westerberg
Quoted 23 lines · Andy Shevchenko
Quoted 4 lines · Mika Westerberg
Quoted 1 line · Andy Shevchenko

We use MODEL_MSCC_OCELOT effectively is a flag for comparing against

as a flag?

Yes, sorry for the typo.

Quoted 5 lines
Quoted 2 lines

"compatible" property. Use device_is_compatible() directly to make it clear.

Okay but if something else ever needs this same quirk then we would need to add new entry here and also to the IDs list.

Yes, that's how DT works and there are, of course, examples all over the kernel, first that comes to my mind: drivers/mmc/host/sdhci-pltfm.c.

Quoted 2 lines

With the flag you can have the IDs in a single place not all over the driver.

Makes the reality harder to read. If I know that the same quirk is used by different platform (in terms of compatible string) I will see it immediately from the code. Flag is meaningless.

Quoted 5 lines
Quoted 2 lines

Well with the flag you get the help from the compiler if you typo it but with the string comparison you are on your own.

It only affects the development cycle, isn't it obvious that any quirk should be tested by the author, otherwise it won't work.

Of course but then are these "cleanups" that probably will not be tested that well if at all. If it is possible to take advantage of the compiler then I prefer that.

Quoted 6 lines
Quoted 2 lines

Therefore I prefer the flag and as I said it also avoid duplicating the compatible string.

Duplicating compatible string is not a problem, linker nowadays combines the same strings into one. In case you are worrying about the space, it's not a problem.

I don't care about the space but just duplication.

maintainer
#10 lore raw

Re: [PATCH v1 2/2] i2c: designware: Use device_is_compatible() instead of custom approach

In reply to Andy Shevchenko

Hi,

Quoted 5 lines · Andy Shevchenko

We use MODEL_MSCC_OCELOT effectively is a flag for comparing against "compatible" property. Use device_is_compatible() directly to make it clear.

I'm coming into the discussion and I personally like Andy's approach. As a third party, I believe this patch has reached the consensus and I'm taking it in i2c/i2c-host.

Thanks Andy,
Andi

author reviewer
#11 lore raw

Re: [PATCH v1 2/2] i2c: designware: Use device_is_compatible() instead of custom approach

Quoted 10 lines · Andi Shyti
Quoted 5 lines · Andy Shevchenko

We use MODEL_MSCC_OCELOT effectively is a flag for comparing against "compatible" property. Use device_is_compatible() directly to make it clear.

I'm coming into the discussion and I personally like Andy's approach. As a third party, I believe this patch has reached the consensus and I'm taking it in i2c/i2c-host.

Thank you!

Signature · 2 lines
With Best Regards,
Andy Shevchenko