Skip to contentSkip to conversation

ASoC: rt5677: Convert I2C driver to ->probe_new()

PATCH In review Andy Shevchenko 3 participants 12 messages Aug 24, 2017 – Aug 25, 2017

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ASoC: rt5677: Convert I2C driver to ->probe_new()

There is no platform code that uses i2c module table. Remove it altogether and adjust ->probe() to be ->probe_new().

---
 sound/soc/codecs/rt5677.c | 11 ++---------
 1 file changed, 2 insertions(+), 9 deletions(-)
sound/soc/codecs/rt5677.c +2 -9
index c735590c5a25..6448b7a00203 100644
@@ -5009,11 +5009,6 @@ static const struct regmap_config rt5677_regmap = {
 	.num_ranges = ARRAY_SIZE(rt5677_ranges),
 };
 
-static const struct i2c_device_id rt5677_i2c_id[] = {
-	{ }
-};
-MODULE_DEVICE_TABLE(i2c, rt5677_i2c_id);
-
 static const struct of_device_id rt5677_of_match[] = {
 	{ .compatible = "realtek,rt5677", RT5677 },
 	{ }
@@ -5131,8 +5126,7 @@ static void rt5677_free_irq(struct i2c_client *i2c)
 		regmap_del_irq_chip(i2c->irq, rt5677->irq_data);
 }
 
-static int rt5677_i2c_probe(struct i2c_client *i2c,
-		    const struct i2c_device_id *id)
+static int rt5677_i2c_probe(struct i2c_client *i2c)
 {
 	struct rt5677_priv *rt5677;
 	int ret;
@@ -5279,9 +5273,8 @@ static struct i2c_driver rt5677_i2c_driver = {
 		.of_match_table = rt5677_of_match,
 		.acpi_match_table = ACPI_PTR(rt5677_acpi_match),
 	},
-	.probe = rt5677_i2c_probe,
+	.probe_new = rt5677_i2c_probe,
 	.remove   = rt5677_i2c_remove,
-	.id_table = rt5677_i2c_id,
 };
 module_i2c_driver(rt5677_i2c_driver);
 
Signature · 1 line
2.14.1
maintainer
#1 lore raw
Quoted 2 lines · Andy Shevchenko

There is no platform code that uses i2c module table. Remove it altogether and adjust ->probe() to be ->probe_new().

What is the value in doing this? Removing the ability to instantiate via board files is just inconveniencing people doing that, it's not a goal in itself.

#2 lore raw
Quoted 9 lines · Mark Brown
Quoted 2 lines · Andy Shevchenko

There is no platform code that uses i2c module table. Remove it altogether and adjust ->probe() to be ->probe_new().

What is the value in doing this? Removing the ability to instantiate via board files is just inconveniencing people doing that, it's not a goal in itself.

But the id table is already empty...?

Takashi

maintainer
#3 lore raw
Quoted 4 lines · Takashi Iwai
Quoted 2 lines · Mark Brown

On Thu, Aug 24, 2017 at 04:41:10PM +0300, Andy Shevchenko wrote:

Quoted 8 lines
Quoted 6 lines
Quoted 2 lines

There is no platform code that uses i2c module table. Remove it altogether and adjust ->probe() to be ->probe_new().

What is the value in doing this? Removing the ability to instantiate via board files is just inconveniencing people doing that, it's not a goal in itself.

But the id table is already empty...?

Right, that seems like a bug though.

#4 lore raw
Quoted 9 lines · Mark Brown
Quoted 6 lines · Takashi Iwai
Quoted 4 lines · Mark Brown
Quoted 2 lines · Andy Shevchenko

There is no platform code that uses i2c module table. Remove it altogether and adjust ->probe() to be ->probe_new().

Quoted 7 lines
Quoted 5 lines
Quoted 3 lines

What is the value in doing this? Removing the ability to instantiate via board files is just inconveniencing people doing that, it's not a goal in itself.

But the id table is already empty...?

Right, that seems like a bug though.

Well, that's what the commit ddc9e69b9dc2 does after all: it cleans up the legacy usage, and moves all either OF or ACPI matching.

If we want to take back old i2c ids, better to put the explicit i2c_match_id() check in rt5677_i2c_probe() instead of sticking with the old i2c probe callback.

Takashi

#5 lore raw
Quoted 16 lines · Takashi Iwai
Quoted 14 lines · Mark Brown
Quoted 11 lines · Takashi Iwai
Quoted 9 lines · Mark Brown
Quoted 2 lines · Andy Shevchenko

There is no platform code that uses i2c module table. Remove it altogether and adjust ->probe() to be ->probe_new().

What is the value in doing this?  Removing the ability to instantiate
via board files is just inconveniencing people doing that, it's not a
goal in itself.

But the id table is already empty...?

Quoted 8 lines
Quoted 1 line

Right, that seems like a bug though.

Well, that's what the commit ddc9e69b9dc2 does after all: it cleans up the legacy usage, and moves all either OF or ACPI matching.

If we want to take back old i2c ids, better to put the explicit i2c_match_id() check in rt5677_i2c_probe() instead of sticking with the old i2c probe callback.

Well, there was never user for rt5676, thus, platform code can use driver name to register / create platform device to be matched.

Signature · 6 lines
Andy Shevchenko <[email protected]>
Intel Finland Oy
_______________________________________________
Alsa-devel mailing list
[email protected]
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
maintainer
#7 lore raw
In reply to Takashi Iwai
Quoted 8 lines · Takashi Iwai
Quoted 3 lines · Mark Brown
Quoted 1 line

But the id table is already empty...?

Right, that seems like a bug though.

Well, that's what the commit ddc9e69b9dc2 does after all: it cleans up the legacy usage, and moves all either OF or ACPI matching.

Which is broken since not all the world is ACPI or DT.

Quoted 3 lines

If we want to take back old i2c ids, better to put the explicit i2c_match_id() check in rt5677_i2c_probe() instead of sticking with the old i2c probe callback.

That's what I said, yes.

#8 lore raw
Quoted 9 lines · Mark Brown
Quoted 6 lines · Takashi Iwai
Quoted 2 lines · Mark Brown
Quoted 1 line

But the id table is already empty...?

Right, that seems like a bug though.

Well, that's what the commit ddc9e69b9dc2 does after all: it cleans up
the legacy usage, and moves all either OF or ACPI matching.

Which is broken since not all the world is ACPI or DT.

While I agree on the second part, I don't think it's broken per se, there never was a single user of that, and looking how ASoC board files are organized it is quite unlikely there will be one...

Quoted 5 lines
Quoted 3 lines

If we want to take back old i2c ids, better to put the explicit i2c_match_id() check in rt5677_i2c_probe() instead of sticking with the old i2c probe callback.

That's what I said, yes.

...though we can (re)introduce this dead code back for a potential user in the future.

Signature · 2 lines
Andy Shevchenko <[email protected]>
Intel Finland Oy
maintainer
#6 lore raw
In reply to Andy Shevchenko
Quoted 8 lines · Andy Shevchenko
Quoted 3 lines · Takashi Iwai

If we want to take back old i2c ids, better to put the explicit i2c_match_id() check in rt5677_i2c_probe() instead of sticking with the old i2c probe callback.

Well, there was never user for rt5676, thus, platform code can use driver name to register / create platform device to be matched.

The driver handle both rt5676 and rt5677 using different IDs so it can't just rely on the driver table.

maintainer
#9 lore raw
In reply to Andy Shevchenko
Quoted 7 lines · Andy Shevchenko
Quoted 5 lines · Mark Brown
Quoted 3 lines · Takashi Iwai

Well, that's what the commit ddc9e69b9dc2 does after all: it cleans up
the legacy usage, and moves all either OF or ACPI matching.

Quoted 5 lines
Quoted 1 line

Which is broken since not all the world is ACPI or DT.

While I agree on the second part, I don't think it's broken per se, there never was a single user of that, and looking how ASoC board files are organized it is quite unlikely there will be one...

I have no idea why you say the ASoC machine drivers would prevent someone adding a platform based user, obviously ASoC predates the use of both ACPI and DT quite considerably.

Quoted 4 lines
Quoted 1 line

That's what I said, yes.

...though we can (re)introduce this dead code back for a potential user in the future.

The IDs are back as a result of the merge resolution I did.

#10 lore raw
Quoted 7 lines · Mark Brown
Quoted 6 lines · Andy Shevchenko
Quoted 5 lines · Mark Brown
Quoted 4 lines · Takashi Iwai

Well, that's what the commit ddc9e69b9dc2 does after all: it cleans
up
the legacy usage, and moves all either OF or ACPI matching.

Quoted 8 lines
Quoted 5 lines
Quoted 1 line

Which is broken since not all the world is ACPI or DT.

While I agree on the second part, I don't think it's broken per se, there never was a single user of that, and looking how ASoC board files
are organized it is quite unlikely there will be one...

I have no idea why you say the ASoC machine drivers would prevent someone adding a platform based user,

For example, looking into
sound/soc/mediatek/mt8173/mt8173-rt5650-rt5676.c
I see a direct use of codec driver without any platform driver involvement.

Is it exception from the rule, or a common use?

Quoted 2 lines

obviously ASoC predates the use of
both ACPI and DT quite considerably.

Yep, the current users of driver in question either DT or ACPI, would it be pure platform in the future?

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

That's what I said, yes.

...though we can (re)introduce this dead code back for a potential user
in the future.

The IDs are back as a result of the merge resolution I did.

I'm fine as long as it doesn't include broken 'ACPI instance as I2C ID'. Thanks!

Signature · 2 lines
Andy Shevchenko <[email protected]>
Intel Finland Oy
maintainer
#11 lore raw
Quoted 11 lines · Andy Shevchenko
Quoted 2 lines · Mark Brown

I have no idea why you say the ASoC machine drivers would prevent someone adding a platform based user,

For example, looking into
sound/soc/mediatek/mt8173/mt8173-rt5650-rt5676.c
I see a direct use of codec driver without any platform driver involvement.

Is it exception from the rule, or a common use?

That's an entirely standard driver using normally registered devices?

Quoted 5 lines
Quoted 2 lines

obviously ASoC predates the use of
both ACPI and DT quite considerably.

Yep, the current users of driver in question either DT or ACPI, would it be pure platform in the future?

It could be potentially (well, I2C but registered using a board file). Most architectures still use board files and neither ACPI nor DT has a sensible story for instantiating plugin modules yet.