regulator: tps6586x: add version detection and voltage tables
Subsystems (4)
v4 cover regulator: tps6586x: add version detection and voltage tables
This is the 4rd version of this patchset which adds version detection for the tps6586x mfd family. Just discovered a missing null reference check which did not triggered in my tests, sorry about that.
This is required because some regulator versions use different voltage tables. The regulator driver now uses the right voltage table according to the version.
While at it, there are a few other incorrect regulator entries in the device tree, change them accordingly (this patch is new and grew out of a patch with the same name which Lucas Stach posted in July).
Tested on Colibri T20 V1.1 and V1.2 which makes use of diffent version of that regulator.
Changes since v3:
- Check if version specific table is necessary
Changes since v2:
- Avoid moving devm_kzalloc
- Removed reg_ from reg_version
- Moved walk through version dependent tables to find_regulator_info, removed the inline definition. This reduces .o size and encapsulates the logic of finding the right regulator into one function. It comes with a slight code duplication, the table search now appears twice.
Changes since v1:
- Merged DT and driver change (preserves bisectability)
- Use different regulator tables for different regulator version
* This also makes the TPS6586X_ANY version needless
- Replaced enum with defines
- Move version message in a sperate function
- Include regulator bugfix, use correct supply voltage
- Checked all voltages, fix SM0/SM1 (Core/CPU) and LDO6 voltage (VDAC/VI) in another commit
- Rebased on top of v3.13-rc2
Stefan Agner (3):
mfd: tps6586x: add version detection
regulator: tps6586x: add and use correct voltage table
ARM: tegra: correct Colibri T20 regulator settings
arch/arm/boot/dts/tegra20-colibri-512.dtsi | 34 +++++------ drivers/mfd/tps6586x.c | 43 ++++++++++++-- drivers/regulator/tps6586x-regulator.c | 93 ++++++++++++++++++++++++------ include/linux/mfd/tps6586x.h | 7 +++ 4 files changed, 136 insertions(+), 41 deletions(-)
Signature · 1 line
1.8.4.2
PATCH v4 1/3 mfd: tps6586x: add version detection
Use the VERSIONCRC to determine the exact device version. According to the datasheet this register can be used as device identifier. The identification is needed since some tps6586x regulators use a different voltage table.
---
Since the struct tps6586x is inside the c file, I could not easily move the tps6586x_get_version function as inline to the header file.
Changes since v2:
- Avoid moving devm_kzalloc
--- drivers/mfd/tps6586x.c | 43 ++++++++++++++++++++++++++++++++++++++----- include/linux/mfd/tps6586x.h | 7 +++++++ 2 files changed, 45 insertions(+), 5 deletions(-)
drivers/mfd/tps6586x.c +38 -5
@@ -124,6 +124,7 @@ struct tps6586x {struct device *dev;struct i2c_client *client;struct regmap *regmap; +int version;int irq;struct irq_chip irq_chip;
@@ -208,6 +209,14 @@ int tps6586x_irq_get_virq(struct device *dev, int irq) }EXPORT_SYMBOL_GPL (tps6586x_irq_get_virq); +int tps6586x_get_version (struct device *dev) +{ +struct tps6586x *tps6586x =dev_get_drvdata (dev); + +return tps6586x->version; +} +EXPORT_SYMBOL_GPL (tps6586x_get_version); +static int __remove_subdev (struct device *dev,void *unused) {platform_device_unregister (to_platform_device (dev));
@@ -472,6 +481,31 @@ static void tps6586x_power_off(void)tps6586x_set_bits (tps6586x_dev,TPS6586X_SUPPLYENE ,SLEEP_MODE_BIT ); } +static void tps6586x_print_version (struct i2c_client *client,int version) +{ +const char *name; + +switch (version) { +case TPS658621A : + name ="TPS658621A" ; +break ; +case TPS658621CD : + name ="TPS658621C/D" ; +break ; +case TPS658623 : + name ="TPS658623" ; +break ; +case TPS658643 : + name ="TPS658643" ; +break ; +default : + name ="TPS6586X" ; +break ; + } + +dev_info (&client->dev,"Found %s, VERSIONCRC is %02x\n" , name, version); +} +static int tps6586x_i2c_probe (struct i2c_client *client,const struct i2c_device_id *id) {
@@ -493,13 +527,12 @@ static int tps6586x_i2c_probe(struct i2c_client *client,return -EIO ; } -dev_info (&client->dev,"VERSIONCRC is %02x\n" , ret); - tps6586x =devm_kzalloc (&client->dev,sizeof (*tps6586x),GFP_KERNEL ); -if (tps6586x ==NULL ) { -dev_err (&client->dev,"memory for tps6586x alloc failed\n" ); +if (!tps6586x)return -ENOMEM ; - } + + tps6586x->version = ret; +tps6586x_print_version (client, tps6586x->version); tps6586x->client = client; tps6586x->dev = &client->dev;
include/linux/mfd/tps6586x.h +7 -0
@@ -13,6 +13,12 @@#define TPS6586X_SLEW_RATE_SET 0x08#define TPS6586X_SLEW_RATE_MASK 0x07 +/* VERSION CRC */ +#define TPS658621A 0x15 +#define TPS658621CD 0x2c +#define TPS658623 0x1b +#define TPS658643 0x03 +enum {TPS6586X_ID_SYS ,TPS6586X_ID_SM_0 ,
@@ -97,5 +103,6 @@ extern int tps6586x_clr_bits(struct device *dev, int reg, uint8_t bit_mask);extern int tps6586x_update (struct device *dev,int reg,uint8_t val,uint8_t mask);extern int tps6586x_irq_get_virq (struct device *dev,int irq); +extern int tps6586x_get_version (struct device *dev);#endif /*__LINUX_MFD_TPS6586X_H */
Signature · 1 line
1.8.4.2
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
drivers/mfd/tps6586x.c int tps6586x_irq_get_virq(struct device *dev, int irq) · line 218
Quoted 7 lines
drivers/mfd/tps6586x.c +7
-0+int tps6586x_get_version (struct device *dev) +{ +struct tps6586x *tps6586x =dev_get_drvdata (dev); + +return tps6586x->version; +} +EXPORT_SYMBOL_GPL (tps6586x_get_version);
I thought Mark suggested that this routine was converted to a 'static inline' and moved into the header? Did you not think this was a good idea?
Signature · 4 lines
Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org ? Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
Am 2013-12-04 09:10, schrieb Lee Jones:
Quoted 11 lines
drivers/mfd/tps6586x.c int tps6586x_irq_get_virq(struct device *dev, int irq) · line 218
Quoted 7 lines
drivers/mfd/tps6586x.c +7
-0+int tps6586x_get_version (struct device *dev) +{ +struct tps6586x *tps6586x =dev_get_drvdata (dev); + +return tps6586x->version; +} +EXPORT_SYMBOL_GPL (tps6586x_get_version);I thought Mark suggested that this routine was converted to a 'static inline' and moved into the header? Did you not think this was a good idea?
As I pointed out in the comment above, the struct tps6586x is in the C file, so I would need to move that too. That's why I did not made that change in the end. What do you think, should I still move (and move the struct too?)
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
Quoted 12 lines · Stefan Agner
Am 2013-12-04 09:10, schrieb Lee Jones:
Quoted 11 lines
drivers/mfd/tps6586x.c int tps6586x_irq_get_virq(struct device *dev, int irq) · line 218
Quoted 7 lines
drivers/mfd/tps6586x.c +7
-0+int tps6586x_get_version (struct device *dev) +{ +struct tps6586x *tps6586x =dev_get_drvdata (dev); + +return tps6586x->version; +} +EXPORT_SYMBOL_GPL (tps6586x_get_version);I thought Mark suggested that this routine was converted to a 'static inline' and moved into the header? Did you not think this was a good idea?
Quoted 4 lines
As I pointed out in the comment above, the struct tps6586x is in the C file, so I would need to move that too. That's why I did not made that change in the end. What do you think, should I still move (and move the struct too?)
Why would the struct have to be moved if the function is inline?
Signature · 4 lines
Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org ? Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
Am 2013-12-04 11:07, schrieb Lee Jones:
Quoted 14 lines · Stefan Agner
Quoted 13 lines
Am 2013-12-04 09:10, schrieb Lee Jones:
Quoted 11 lines
drivers/mfd/tps6586x.c int tps6586x_irq_get_virq(struct device *dev, int irq) · line 218
Quoted 7 lines
drivers/mfd/tps6586x.c +7
-0+int tps6586x_get_version (struct device *dev) +{ +struct tps6586x *tps6586x =dev_get_drvdata (dev); + +return tps6586x->version; +} +EXPORT_SYMBOL_GPL (tps6586x_get_version);I thought Mark suggested that this routine was converted to a 'static inline' and moved into the header? Did you not think this was a good idea?
Quoted 6 lines
Quoted 4 lines
As I pointed out in the comment above, the struct tps6586x is in the C file, so I would need to move that too. That's why I did not made that change in the end. What do you think, should I still move (and move the struct too?)
Why would the struct have to be moved if the function is inline?
True, the inline I could have done without moving the struct and the function. Would you like me to create another revision doing this?
But moving the function needs moving of the struct tps6586x declaration...
[Sorry, this time with answer all]
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
Quoted 22 lines · Stefan Agner
Am 2013-12-04 11:07, schrieb Lee Jones:
Quoted 14 lines · Stefan Agner
Quoted 13 lines
Am 2013-12-04 09:10, schrieb Lee Jones:
Quoted 11 lines
drivers/mfd/tps6586x.c int tps6586x_irq_get_virq(struct device *dev, int irq) · line 218
Quoted 7 lines
drivers/mfd/tps6586x.c +7
-0+int tps6586x_get_version (struct device *dev) +{ +struct tps6586x *tps6586x =dev_get_drvdata (dev); + +return tps6586x->version; +} +EXPORT_SYMBOL_GPL (tps6586x_get_version);I thought Mark suggested that this routine was converted to a 'static inline' and moved into the header? Did you not think this was a good idea?
Quoted 6 lines
Quoted 4 lines
As I pointed out in the comment above, the struct tps6586x is in the C file, so I would need to move that too. That's why I did not made that change in the end. What do you think, should I still move (and move the struct too?)
Why would the struct have to be moved if the function is inline?
Quoted 7 lines
True, the inline I could have done without moving the struct and the function. Would you like me to create another revision doing this?
But moving the function needs moving of the struct tps6586x declaration...
[Sorry, this time with answer all]
Do you know what, it's really not that important.
Patch applied.
Signature · 4 lines
Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org ? Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
Mark Brown
maintainerRe: [PATCH v4 1/3] mfd: tps6586x: add version detection
In reply to Lee JonesQuoted 8 lines · Lee Jones
Quoted 4 lines · Stefan Agner
As I pointed out in the comment above, the struct tps6586x is in the C file, so I would need to move that too. That's why I did not made that change in the end. What do you think, should I still move (and move the struct too?)
Why would the struct have to be moved if the function is inline?
If the function is in the header and trying to use a struct that's only
defined in the C file then it's not going to build - keeping the struct
in the C file only does seem like a worthwhile thing for encapsulation.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131204/971a8602/attachment.sig>
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
Quoted 10 lines
Quoted 6 lines
Quoted 4 lines
As I pointed out in the comment above, the struct tps6586x is in the C file, so I would need to move that too. That's why I did not made that change in the end. What do you think, should I still move (and move the struct too?)
Why would the struct have to be moved if the function is inline?
If the function is in the header and trying to use a struct that's only defined in the C file then it's not going to build - keeping the struct in the C file only does seem like a worthwhile thing for encapsulation.
Yes, I just carried out my own testing and found that out. Prior to this I was under the impression that inline functions were directly transposed into the location of the call, were perhaps it could make use of any resources declared within that context. It appears that impression was not correct.
Every day continues to be a school day. :)
Signature · 4 lines
Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org ? Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
In reply to Stefan AgnerQuoted 11 lines · Stefan Agner
Use the VERSIONCRC to determine the exact device version. According to the datasheet this register can be used as device identifier. The identification is needed since some tps6586x regulators use a different voltage table.
drivers/mfd/tps6586x.c +0
-0@@ -493,13 +527,12 @@ static int tps6586x_i2c_probe(struct i2c_client *client,return -EIO ; }
drivers/mfd/tps6586x.c static int tps6586x_i2c_probe(struct i2c_client *client, · line 534
Quoted 10 lines
drivers/mfd/tps6586x.c +3
-5-dev_info (&client->dev,"VERSIONCRC is %02x\n" , ret); - tps6586x =devm_kzalloc (&client->dev,sizeof (*tps6586x),GFP_KERNEL ); -if (tps6586x ==NULL ) { -dev_err (&client->dev,"memory for tps6586x alloc failed\n" ); +if (!tps6586x)return -ENOMEM ; - } + + tps6586x->version = ret;
I have to say, I dislike this version of the patch. Separating the reading of the version register from the assignment to tps6586x->version doesn't make any sense, especially given that the version value is stored in a variable named "ret"; that name isn't remotely related to what's stored there. What if someone comes along later and adds more code that assigns to ret between where it's repurposed for the version value and where it's assigned to tps6586x->version? It'd be extremely difficult for a patch reviewer to spot that given the limited context in a diff, and quite non-obvious to the person changing the code too..
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
Am 2013-12-05 18:06, schrieb Stephen Warren:
<snip>
Quoted 16 lines
Quoted 4 lines
drivers/mfd/tps6586x.c +0
-0@@ -493,13 +527,12 @@ static int tps6586x_i2c_probe(struct i2c_client *client,return -EIO ; }drivers/mfd/tps6586x.c static int tps6586x_i2c_probe(struct i2c_client *client, · line 534
Quoted 10 lines
drivers/mfd/tps6586x.c +3
-5-dev_info (&client->dev,"VERSIONCRC is %02x\n" , ret); - tps6586x =devm_kzalloc (&client->dev,sizeof (*tps6586x),GFP_KERNEL ); -if (tps6586x ==NULL ) { -dev_err (&client->dev,"memory for tps6586x alloc failed\n" ); +if (!tps6586x)return -ENOMEM ; - } + + tps6586x->version = ret;I have to say, I dislike this version of the patch. Separating the
Quoted 8 lines
reading of the version register from the assignment to
tps6586x->versiondoesn't make any sense, especially given that the version value is stored in a variable named "ret"; that name isn't remotely related to what's stored there. What if someone comes along later and adds more code that assigns to ret between where it's repurposed for the version value and where it's assigned totps6586x->version? It'd be extremely difficult for a patch reviewer to spot that given the limited context in a diff, and quite non-obvious to the person changing the code too..
The value comes from the return value of i2c_smbus_read_byte_data. If the value is below zero its an EIO error.
I could add a variable "version", but for me it felt strange because we check if version is below zero. This feels like its a wrong version rather than a transmit error. So I would prefer ret over version. But I agree, when one just reads the patch, its not obvious what exactly happens.
In v2, I moved the i2c_smbus_read_byte_data function call after the allocation, so it was more obvious for the reader. But then, as Thierry Reding pointed out, not moving it is an optimization: In case reading fails, we don't allocate memory first.
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
Quoted 27 lines · Stefan Agner
Am 2013-12-05 18:06, schrieb Stephen Warren:
<snip>Quoted 16 lines
Quoted 4 lines
drivers/mfd/tps6586x.c +0
-0@@ -493,13 +527,12 @@ static int tps6586x_i2c_probe(struct i2c_client *client,return -EIO ; }drivers/mfd/tps6586x.c static int tps6586x_i2c_probe(struct i2c_client *client, · line 534
Quoted 10 lines
drivers/mfd/tps6586x.c +3
-5-dev_info (&client->dev,"VERSIONCRC is %02x\n" , ret); - tps6586x =devm_kzalloc (&client->dev,sizeof (*tps6586x),GFP_KERNEL ); -if (tps6586x ==NULL ) { -dev_err (&client->dev,"memory for tps6586x alloc failed\n" ); +if (!tps6586x)return -ENOMEM ; - } + + tps6586x->version = ret;I have to say, I dislike this version of the patch. Separating the
Quoted 8 lines
reading of the version register from the assignment to
tps6586x->versiondoesn't make any sense, especially given that the version value is stored in a variable named "ret"; that name isn't remotely related to what's stored there. What if someone comes along later and adds more code that assigns to ret between where it's repurposed for the version value and where it's assigned totps6586x->version? It'd be extremely difficult for a patch reviewer to spot that given the limited context in a diff, and quite non-obvious to the person changing the code too..
Quoted 8 lines
The value comes from the return value of
i2c_smbus_read_byte_data. If the value is below zero its an EIO error.I could add a variable "version", but for me it felt strange because we check if version is below zero. This feels like its a wrong version rather than a transmit error. So I would prefer ret over version. But I agree, when one just reads the patch, its not obvious what exactly happens.
In my opinion, using a variable named "version" here would be preferable. Testing that against <0 is just the way the I2C API works, so the same argument could be applied to any I2C access.
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
Am 2013-12-05 18:40, schrieb Stephen Warren:
Quoted 37 lines · Stefan Agner
Quoted 27 lines
Am 2013-12-05 18:06, schrieb Stephen Warren:
<snip>Quoted 16 lines
Quoted 4 lines
drivers/mfd/tps6586x.c +0
-0@@ -493,13 +527,12 @@ static int tps6586x_i2c_probe(struct i2c_client *client,return -EIO ; }drivers/mfd/tps6586x.c static int tps6586x_i2c_probe(struct i2c_client *client, · line 534
Quoted 10 lines
drivers/mfd/tps6586x.c +3
-5-dev_info (&client->dev,"VERSIONCRC is %02x\n" , ret); - tps6586x =devm_kzalloc (&client->dev,sizeof (*tps6586x),GFP_KERNEL ); -if (tps6586x ==NULL ) { -dev_err (&client->dev,"memory for tps6586x alloc failed\n" ); +if (!tps6586x)return -ENOMEM ; - } + + tps6586x->version = ret;I have to say, I dislike this version of the patch. Separating the
Quoted 8 lines
reading of the version register from the assignment to
tps6586x->versiondoesn't make any sense, especially given that the version value is stored in a variable named "ret"; that name isn't remotely related to what's stored there. What if someone comes along later and adds more code that assigns to ret between where it's repurposed for the version value and where it's assigned totps6586x->version? It'd be extremely difficult for a patch reviewer to spot that given the limited context in a diff, and quite non-obvious to the person changing the code too..Quoted 8 lines
The value comes from the return value of
i2c_smbus_read_byte_data. If the value is below zero its an EIO error.I could add a variable "version", but for me it felt strange because we check if version is below zero. This feels like its a wrong version rather than a transmit error. So I would prefer ret over version. But I agree, when one just reads the patch, its not obvious what exactly happens.
Quoted 3 lines
In my opinion, using a variable named "version" here would be preferable. Testing that against <0 is just the way the I2C API works, so the same argument could be applied to any I2C access.
Hm, I try the empiric way:
$ grep -r -e i2c_smbus_read_byte_data | grep "ret =" | wc -l 139 $ grep -r -e i2c_smbus_read_byte_data | grep "version =" | wc -l 3
Ok, thats not fair at all, version is usage specific whilst ret is not.
$ grep -r -e i2c_smbus_read_byte_data | grep " = " | wc -l 703
On the other hand is the additional variable. But I think the compiler will optimize that anyway, so this might not be an argument at all :-)
I see your point... Should I create another patch revision? Lee, is the patch already merged?
Re: [PATCH v4 1/3] mfd: tps6586x: add version detection
Quoted 35 lines
Quoted 35 lines
Quoted 34 lines
<snip>
Quoted 16 lines
Quoted 4 lines
drivers/mfd/tps6586x.c +0
-0@@ -493,13 +527,12 @@ static int tps6586x_i2c_probe(struct i2c_client *client,return -EIO ; }drivers/mfd/tps6586x.c static int tps6586x_i2c_probe(struct i2c_client *client, · line 534
Quoted 10 lines
drivers/mfd/tps6586x.c +3
-5-dev_info (&client->dev,"VERSIONCRC is %02x\n" , ret); - tps6586x =devm_kzalloc (&client->dev,sizeof (*tps6586x),GFP_KERNEL ); -if (tps6586x ==NULL ) { -dev_err (&client->dev,"memory for tps6586x alloc failed\n" ); +if (!tps6586x)return -ENOMEM ; - } + + tps6586x->version = ret;I have to say, I dislike this version of the patch. Separating the
Quoted 8 lines
reading of the version register from the assignment to
tps6586x->versiondoesn't make any sense, especially given that the version value is stored in a variable named "ret"; that name isn't remotely related to what's stored there. What if someone comes along later and adds more code that assigns to ret between where it's repurposed for the version value and where it's assigned totps6586x->version? It'd be extremely difficult for a patch reviewer to spot that given the limited context in a diff, and quite non-obvious to the person changing the code too..The value comes from the return value of
i2c_smbus_read_byte_data. If the value is below zero its an EIO error.I could add a variable "version", but for me it felt strange because we check if version is below zero. This feels like its a wrong version rather than a transmit error. So I would prefer ret over version. But I agree, when one just reads the patch, its not obvious what exactly happens.
Quoted 3 lines
Quoted 3 lines
In my opinion, using a variable named "version" here would be preferable. Testing that against <0 is just the way the I2C API works, so the same argument could be applied to any I2C access.
So, FWIW I agree with Stephen and have done from the start. Please see my original comment from the first submission:
Quoted 10 lines
Quoted 1 line
ret = i2c_smbus_read_byte_data(client, TPS6586X_VERSIONCRC);If you're going to do this, please change 'ret' to 'version'.
Hm, I try the empiric way:
$ grep -r -e i2c_smbus_read_byte_data | grep "ret =" | wc -l 139 $ grep -r -e i2c_smbus_read_byte_data | grep "version =" | wc -l 3
Quoted 4 lines
Ok, thats not fair at all, version is usage specific whilst ret is not.
$ grep -r -e i2c_smbus_read_byte_data | grep " = " | wc -l 703
I not really that worried about what everyone else does. I'm more concerned with doing what we deem to be the correct thing here.
Quoted 5 lines
On the other hand is the additional variable. But I think the compiler will optimize that anyway, so this might not be an argument at all :-)
I see your point... Should I create another patch revision? Lee, is the patch already merged?
It isn't. Please submit another version as Stephen requests.
Signature · 4 lines
Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org ? Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
PATCH v4 2/3 regulator: tps6586x: add and use correct voltage table
In reply to Stefan AgnerDepending on the regulator version, the voltage table might be different. Use version specific regulator tables in order to select correct voltage table. For the following regulator versions different voltage tables are now used:
* TPS658623: Use correct voltage table for SM2
* TPS658643: New voltage table for SM2
Both versions are in use on the Colibri T20 module. Make use of the correct tables by requesting the correct SM2 voltage of 1.8V.
This change is not backward compatible since an old driver is not able to correctly set that value. The value 1.8V is out of range for the old driver and will refuse to probe the device. The regulator starts with default settings and the driver shows appropriate error messages.
On Colibri T20, the old value used to work with TPS658623 since the driver applied a wrong voltage table too. However, the TPS658643 used on V1.2 devices uses yet another voltage table and those broke that pseudo-compatibility. The regulator driver now has the correct voltage table for both regulator versions and those the correct voltage can be used in the device tree.
---
Changes since v3:
- Check if version specific table is necessary
Changes since v2:
- Removed reg_ from reg_version
- Moved walk through version dependent tables to find_regulator_info, removed the inline definition. This reduces .o size and encapsulates the logic of finding the right regulator into one function. It comes with a slight code duplication, the table search now appears twice.
--- arch/arm/boot/dts/tegra20-colibri-512.dtsi | 4 +- drivers/regulator/tps6586x-regulator.c | 93 ++++++++++++++++++++++++------ 2 files changed, 76 insertions(+), 21 deletions(-)
arch/arm/boot/dts/tegra20-colibri-512.dtsi +2 -2
@@ -268,8 +268,8 @@ reg = <3>; regulator-compatible = "sm2"; regulator-name = "vdd_sm2,vin_ldo*"; - regulator-min-microvolt = <3700000>; - regulator-max-microvolt = <3700000>; + regulator-min-microvolt = <1800000>; + regulator-max-microvolt = <1800000>; regulator-always-on; };
drivers/regulator/tps6586x-regulator.c +74 -19
148 lines not on this page · raw message
Signature · 1 line
1.8.4.2
Re: [PATCH v4 2/3] regulator: tps6586x: add and use correct voltage table
On Tue, Dec 03, 2013 at 11:18:47PM +0100, Stefan Agner wrote:
[...]
Quoted 1 line
drivers/regulator/tps6586x-regulator.c +0
-0
[...]
drivers/regulator/tps6586x-regulator.c static int tps6586x_regulator_set_slew_rate(struct platform_device *pdev, · line 311
Quoted 8 lines
drivers/regulator/tps6586x-regulator.c +8
-0+/* Search version specific table first */ +if (table) { +for (i =0 ; i < num; i++) { + ri = &table[i]; +if (ri->desc.id == id) +return ri; + } + }
Ah... there you go. Looks good:
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131204/417edcbe/attachment-0001.sig>
Re: [PATCH v4 2/3] regulator: tps6586x: add and use correct voltage table
In reply to Stefan AgnerQuoted 4 lines · Stefan Agner
Depending on the regulator version, the voltage table might be different. Use version specific regulator tables in order to select correct voltage table. For the following regulator versions different voltage tables are now used:
I won't apply since I think that patch 3 needs to go along with this,
unless people decide the regulator tree is the easiest way to merge.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131204/4e21d4ac/attachment-0001.sig>
Re: [PATCH v4 2/3] regulator: tps6586x: add and use correct voltage table
Am 2013-12-04 13:14, schrieb Mark Brown:
Quoted 10 lines · Stefan Agner
Quoted 4 lines
Depending on the regulator version, the voltage table might be different. Use version specific regulator tables in order to select correct voltage table. For the following regulator versions different voltage tables are now used:
Acked-by Mark Brown <[email protected]>I won't apply since I think that patch 3 needs to go along with this, unless people decide the regulator tree is the easiest way to merge.
I merged the dependent DT changes into that patch, so from my side, patch 3 don't needs to go along with this...
Mark Brown
maintainerRe: [PATCH v4 2/3] regulator: tps6586x: add and use correct voltage table
Quoted 9 lines · Stefan Agner
Am 2013-12-04 13:14, schrieb Mark Brown:
Quoted 4 lines
Acked-by Mark Brown <[email protected]>I won't apply since I think that patch 3 needs to go along with this, unless people decide the regulator tree is the easiest way to merge.
I merged the dependent DT changes into that patch, so from my side, patch 3 don't needs to go along with this...
OK, though now I look I see it also depends on patch 1 itself so at
least those two will need to go together.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20131204/813a4258/attachment-0001.sig>
Re: [PATCH v4 2/3] regulator: tps6586x: add and use correct voltage table
Quoted 10 lines
Quoted 7 lines
Quoted 4 lines
Acked-by Mark Brown <[email protected]>I won't apply since I think that patch 3 needs to go along with this, unless people decide the regulator tree is the easiest way to merge.
I merged the dependent DT changes into that patch, so from my side, patch 3 don't needs to go along with this...
OK, though now I look I see it also depends on patch 1 itself so at least those two will need to go together.
Well as I've already applied patch 1, I can take this too, no problem.
Signature · 4 lines
Lee Jones Linaro STMicroelectronics Landing Team Lead Linaro.org ? Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
Re: [PATCH v4 2/3] regulator: tps6586x: add and use correct voltage table
In reply to Stefan AgnerQuoted 17 lines · Stefan Agner
Depending on the regulator version, the voltage table might be different. Use version specific regulator tables in order to select correct voltage table. For the following regulator versions different voltage tables are now used:
* TPS658623: Use correct voltage table for SM2 * TPS658643: New voltage table for SM2
Both versions are in use on the Colibri T20 module. Make use of the correct tables by requesting the correct SM2 voltage of 1.8V.
This change is not backward compatible since an old driver is not able to correctly set that value. The value 1.8V is out of range for the old driver and will refuse to probe the device. The regulator starts with default settings and the driver shows appropriate error messages.
Quoted 6 lines
On Colibri T20, the old value used to work with TPS658623 since the driver applied a wrong voltage table too. However, the TPS658643 used on V1.2 devices uses yet another voltage table and those broke that pseudo-compatibility. The regulator driver now has the correct voltage table for both regulator versions and those the correct voltage can be used in the device tree.
(yes, taking this through the MFD tree with patch 1/3 makes sense)
PATCH v4 3/3 ARM: tegra: correct Colibri T20 regulator settings
In reply to Stefan AgnerSet the parent of the regulators LDO2 to LDO9 according to the schematic. Set the base voltage to 3.3V, there is only 3.3V on the module itself.
Set the Core and CPU voltage to the specified voltages of 1.2V and 1.0V respectivly.
LDO6 should deliver 2.85V. The attached peripherals were not in use so far.
--- arch/arm/boot/dts/tegra20-colibri-512.dtsi | 30 +++++++++++++++--------------- 1 file changed, 15 insertions(+), 15 deletions(-)
arch/arm/boot/dts/tegra20-colibri-512.dtsi +15 -15
@@ -225,15 +225,15 @@ #gpio-cells = <2>; gpio-controller; - sys-supply = <&vdd_5v0_reg>; + sys-supply = <&vdd_3v3_reg>; vin-sm0-supply = <&sys_reg>; vin-sm1-supply = <&sys_reg>; vin-sm2-supply = <&sys_reg>; vinldo01-supply = <&sm2_reg>; - vinldo23-supply = <&sm2_reg>; - vinldo4-supply = <&sm2_reg>; - vinldo678-supply = <&sm2_reg>; - vinldo9-supply = <&sm2_reg>; + vinldo23-supply = <&vdd_3v3_reg>; + vinldo4-supply = <&vdd_3v3_reg>; + vinldo678-supply = <&vdd_3v3_reg>; + vinldo9-supply = <&vdd_3v3_reg>; regulators { #address-cells = <1>;
@@ -250,8 +250,8 @@ reg = <1>; regulator-compatible = "sm0"; regulator-name = "vdd_sm0,vdd_core"; - regulator-min-microvolt = <1275000>; - regulator-max-microvolt = <1275000>; + regulator-min-microvolt = <1200000>; + regulator-max-microvolt = <1200000>; regulator-always-on; };
@@ -259,8 +259,8 @@ reg = <2>; regulator-compatible = "sm1"; regulator-name = "vdd_sm1,vdd_cpu"; - regulator-min-microvolt = <1100000>; - regulator-max-microvolt = <1100000>; + regulator-min-microvolt = <1000000>; + regulator-max-microvolt = <1000000>; regulator-always-on; };
@@ -316,8 +316,8 @@ reg = <10>; regulator-compatible = "ldo6"; regulator-name = "vdd_ldo6,avdd_vdac,vddio_vi,vddio_cam"; - regulator-min-microvolt = <1800000>; - regulator-max-microvolt = <1800000>; + regulator-min-microvolt = <2850000>; + regulator-max-microvolt = <2850000>; }; hdmi_vdd_reg: regulator at 11 {
@@ -504,12 +504,12 @@ #address-cells = <1>; #size-cells = <0>; - vdd_5v0_reg: regulator at 100 { + vdd_3v3_reg: regulator at 100 { compatible = "regulator-fixed"; reg = <100>; - regulator-name = "vdd_5v0"; - regulator-min-microvolt = <5000000>; - regulator-max-microvolt = <5000000>; + regulator-name = "vdd_3v3"; + regulator-min-microvolt = <3300000>; + regulator-max-microvolt = <3300000>; regulator-always-on; };
Signature · 1 line
1.8.4.2
Re: [PATCH v4 3/3] ARM: tegra: correct Colibri T20 regulator settings
Am Dienstag, den 03.12.2013, 23:18 +0100 schrieb Stefan Agner:
Quoted 11 lines
Set the parent of the regulators LDO2 to LDO9 according to the schematic. Set the base voltage to 3.3V, there is only 3.3V on the module itself.
Set the Core and CPU voltage to the specified voltages of 1.2V and 1.0V respectivly.
LDO6 should deliver 2.85V. The attached peripherals were not in use so far.
Signed-off-by Stefan Agner <[email protected]>
Quoted 79 lines
--- arch/arm/boot/dts/tegra20-colibri-512.dtsi | 30 +++++++++++++++--------------- 1 file changed, 15 insertions(+), 15 deletions(-)arch/arm/boot/dts/tegra20-colibri-512.dtsi +15
-15@@ -225,15 +225,15 @@ #gpio-cells = <2>; gpio-controller; - sys-supply = <&vdd_5v0_reg>; + sys-supply = <&vdd_3v3_reg>; vin-sm0-supply = <&sys_reg>; vin-sm1-supply = <&sys_reg>; vin-sm2-supply = <&sys_reg>; vinldo01-supply = <&sm2_reg>; - vinldo23-supply = <&sm2_reg>; - vinldo4-supply = <&sm2_reg>; - vinldo678-supply = <&sm2_reg>; - vinldo9-supply = <&sm2_reg>; + vinldo23-supply = <&vdd_3v3_reg>; + vinldo4-supply = <&vdd_3v3_reg>; + vinldo678-supply = <&vdd_3v3_reg>; + vinldo9-supply = <&vdd_3v3_reg>; regulators { #address-cells = <1>;@@ -250,8 +250,8 @@ reg = <1>; regulator-compatible = "sm0"; regulator-name = "vdd_sm0,vdd_core"; - regulator-min-microvolt = <1275000>; - regulator-max-microvolt = <1275000>; + regulator-min-microvolt = <1200000>; + regulator-max-microvolt = <1200000>; regulator-always-on; };@@ -259,8 +259,8 @@ reg = <2>; regulator-compatible = "sm1"; regulator-name = "vdd_sm1,vdd_cpu"; - regulator-min-microvolt = <1100000>; - regulator-max-microvolt = <1100000>; + regulator-min-microvolt = <1000000>; + regulator-max-microvolt = <1000000>; regulator-always-on; };@@ -316,8 +316,8 @@ reg = <10>; regulator-compatible = "ldo6"; regulator-name = "vdd_ldo6,avdd_vdac,vddio_vi,vddio_cam"; - regulator-min-microvolt = <1800000>; - regulator-max-microvolt = <1800000>; + regulator-min-microvolt = <2850000>; + regulator-max-microvolt = <2850000>; }; hdmi_vdd_reg: regulator at 11 {@@ -504,12 +504,12 @@ #address-cells = <1>; #size-cells = <0>; - vdd_5v0_reg: regulator at 100 { + vdd_3v3_reg: regulator at 100 { compatible = "regulator-fixed"; reg = <100>; - regulator-name = "vdd_5v0"; - regulator-min-microvolt = <5000000>; - regulator-max-microvolt = <5000000>; + regulator-name = "vdd_3v3"; + regulator-min-microvolt = <3300000>; + regulator-max-microvolt = <3300000>; regulator-always-on; };
Signature · 4 lines
Pengutronix e.K. | Lucas Stach | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-5076 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
Re: [PATCH v4 3/3] ARM: tegra: correct Colibri T20 regulator settings
In reply to Stefan AgnerQuoted 9 lines · Stefan Agner
Set the parent of the regulators LDO2 to LDO9 according to the schematic. Set the base voltage to 3.3V, there is only 3.3V on the module itself.
Set the Core and CPU voltage to the specified voltages of 1.2V and 1.0V respectivly.
LDO6 should deliver 2.85V. The attached peripherals were not in use so far.
Does this depend on patch 1/3 or 2/3, or can I apply it to the Tegra tree which doesn't include those patches?
If this does depend on patch 1/3 or 2/3... this patch feels a bit big to apply to the MFD tree. Perhaps patch 1/3 and 2/3 could be applied alone to a topic branch in the MFD tree so I can merge that into the Tegra tree in order to apply this?
Re: [PATCH v4 3/3] ARM: tegra: correct Colibri T20 regulator settings
Am 2013-12-05 18:12, schrieb Stephen Warren:
Quoted 5 lines · Stefan Agner
Quoted 4 lines
Set the parent of the regulators LDO2 to LDO9 according to the schematic. Set the base voltage to 3.3V, there is only 3.3V on the module itself.
Quoted 8 lines
Quoted 5 lines
Set the Core and CPU voltage to the specified voltages of 1.2V and 1.0V respectivly.
LDO6 should deliver 2.85V. The attached peripherals were not in use so far.
Does this depend on patch 1/3 or 2/3, or can I apply it to the Tegra tree which doesn't include those patches?
This patch is independent from 1/3 and 2/3, so you can apply it to the Tegra tree. The relevant DT change for 2/3 is included in 2/3...