[PATCH v2 0/3] Add Legacy PM OPS usage checks to class, bus, and driver register functions

[PATCH v2 0/3] Add Legacy PM OPS usage checks to class, bus, and driver register functions

[Shuah Khan]

Add Legacy PM OPS usage checks to class, bus, and driver register functions.
If Legacy PM OPS usage is found, print warning message to indicate that the
driver code needs updating to use Dev PM OPS interfaces. This will help serve
as a way to track drivers that still use Legacy PM OPS and fix them.

This patch set adds Legacy PM OPS usage check and warning to bus_register().
__class_register(), and driver_register().

Individual patches in this series are not dependent on each other. The only
reason this is a series is for context and facilitating discussion on the
overall change as opposed individual patches.

Shuah Khan (3):
  drivers/bus: Add Legacy PM OPS usage check and warning to
    bus_register()
  drivers/class: Add Legacy PM OPS usage check and warning to
    __class_register()
  driver: Add Legacy PM OPS usage check and warning to driver_register()

 drivers/base/bus.c    | 3 +++
 drivers/base/class.c  | 4 ++++
 drivers/base/driver.c | 4 ++++
 3 files changed, 11 insertions(+)

-- 
1.8.3.2

[PATCH v2 1/3] drivers/bus: Add Legacy PM OPS usage check and warning to bus_register()

[Shuah Khan]

Add Legacy PM OPS usage checks to bus_register() function. If Legacy PM OPS
usage is found, print warning message to indicate that the driver code needs
updating to use Dev PM OPS interfaces. This will help serve as a way to track
drivers that still use Legacy PM OPS and fix them.

The Legacy PM OPS check looks for suspend(struct device *, pm_message_t) or
resume(struct device *) bus level interfaces.

Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
---
 drivers/base/bus.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/base/bus.c b/drivers/base/bus.c
index 73f6c29..e8753a7 100644
--- a/drivers/base/bus.c
+++ b/drivers/base/bus.c
@@ -916,6 +916,9 @@ int bus_register(struct bus_type *bus)
 		goto bus_groups_fail;
 
 	pr_debug("bus: '%s': registered\n", bus->name);
+	if (bus->suspend || bus->resume)
+		pr_warn("bus '%s' needs updating - use pm pointer.\n",
+			bus->name);
 	return 0;
 
 bus_groups_fail:
-- 
1.8.3.2

Re: [PATCH v2 1/3] drivers/bus: Add Legacy PM OPS usage check and warning to bus_register()

[Greg KH]

On Mon, Jan 06, 2014 at 01:03:21PM -0700, Shuah Khan wrote:
> Add Legacy PM OPS usage checks to bus_register() function. If Legacy PM OPS
> usage is found, print warning message to indicate that the driver code needs
> updating to use Dev PM OPS interfaces. This will help serve as a way to track
> drivers that still use Legacy PM OPS and fix them.
> 
> The Legacy PM OPS check looks for suspend(struct device *, pm_message_t) or
> resume(struct device *) bus level interfaces.
> 
> Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
> ---
>  drivers/base/bus.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/base/bus.c b/drivers/base/bus.c
> index 73f6c29..e8753a7 100644
> --- a/drivers/base/bus.c
> +++ b/drivers/base/bus.c
> @@ -916,6 +916,9 @@ int bus_register(struct bus_type *bus)
>  		goto bus_groups_fail;
>  
>  	pr_debug("bus: '%s': registered\n", bus->name);
> +	if (bus->suspend || bus->resume)
> +		pr_warn("bus '%s' needs updating - use pm pointer.\n",
> +			bus->name);

Why can't we just sweep the tree for all of these now, fix them up, and
then delete these fields and be done with it?

Same for the other ones, putting warnings in the kernel log files
doesn't work at all for getting people to fix up their code (see
examples of scsi log messages being there for _years_ about obsolete
driver interfaces being used.)

thanks,

greg k-h

Re: [PATCH v2 1/3] drivers/bus: Add Legacy PM OPS usage check and warning to bus_register()

[Shuah Khan]

On 01/06/2014 01:17 PM, Greg KH wrote:
> On Mon, Jan 06, 2014 at 01:03:21PM -0700, Shuah Khan wrote:
>> Add Legacy PM OPS usage checks to bus_register() function. If Legacy PM OPS
>> usage is found, print warning message to indicate that the driver code needs
>> updating to use Dev PM OPS interfaces. This will help serve as a way to track
>> drivers that still use Legacy PM OPS and fix them.
>>
>> The Legacy PM OPS check looks for suspend(struct device *, pm_message_t) or
>> resume(struct device *) bus level interfaces.
>>
>> Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
>> ---
>>   drivers/base/bus.c | 3 +++
>>   1 file changed, 3 insertions(+)
>>
>> diff --git a/drivers/base/bus.c b/drivers/base/bus.c
>> index 73f6c29..e8753a7 100644
>> --- a/drivers/base/bus.c
>> +++ b/drivers/base/bus.c
>> @@ -916,6 +916,9 @@ int bus_register(struct bus_type *bus)
>>   		goto bus_groups_fail;
>>
>>   	pr_debug("bus: '%s': registered\n", bus->name);
>> +	if (bus->suspend || bus->resume)
>> +		pr_warn("bus '%s' needs updating - use pm pointer.\n",
>> +			bus->name);
>
> Why can't we just sweep the tree for all of these now, fix them up, and
> then delete these fields and be done with it?
>
> Same for the other ones, putting warnings in the kernel log files
> doesn't work at all for getting people to fix up their code (see
> examples of scsi log messages being there for _years_ about obsolete
> driver interfaces being used.)

This change is not a simple case of replacing legacy with pm ops. I have 
been working on changing drivers that use legacy, in some cases it is an 
easy change, however in some cases it requires adding new interfaces to 
port over from legacy to pm ops. Legacy suspend/resume is more course 
grain and pm ops offer fine grain control over suspend, hibernate etc. 
Also it has been a challenge without the hardware to test.

My thinking is adding warning might get the attention of individual 
driver owners.

-- Shuah

-- 
Shuah Khan
Senior Linux Kernel Developer - Open Source Group
Samsung Research America(Silicon Valley)
shuah.kh@samsung.com | (970) 672-0658

Re: [PATCH v2 1/3] drivers/bus: Add Legacy PM OPS usage check and warning to bus_register()

[Greg KH]

On Mon, Jan 06, 2014 at 01:30:31PM -0700, Shuah Khan wrote:
> On 01/06/2014 01:17 PM, Greg KH wrote:
> >On Mon, Jan 06, 2014 at 01:03:21PM -0700, Shuah Khan wrote:
> >>Add Legacy PM OPS usage checks to bus_register() function. If Legacy PM OPS
> >>usage is found, print warning message to indicate that the driver code needs
> >>updating to use Dev PM OPS interfaces. This will help serve as a way to track
> >>drivers that still use Legacy PM OPS and fix them.
> >>
> >>The Legacy PM OPS check looks for suspend(struct device *, pm_message_t) or
> >>resume(struct device *) bus level interfaces.
> >>
> >>Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
> >>---
> >>  drivers/base/bus.c | 3 +++
> >>  1 file changed, 3 insertions(+)
> >>
> >>diff --git a/drivers/base/bus.c b/drivers/base/bus.c
> >>index 73f6c29..e8753a7 100644
> >>--- a/drivers/base/bus.c
> >>+++ b/drivers/base/bus.c
> >>@@ -916,6 +916,9 @@ int bus_register(struct bus_type *bus)
> >>  		goto bus_groups_fail;
> >>
> >>  	pr_debug("bus: '%s': registered\n", bus->name);
> >>+	if (bus->suspend || bus->resume)
> >>+		pr_warn("bus '%s' needs updating - use pm pointer.\n",
> >>+			bus->name);
> >
> >Why can't we just sweep the tree for all of these now, fix them up, and
> >then delete these fields and be done with it?
> >
> >Same for the other ones, putting warnings in the kernel log files
> >doesn't work at all for getting people to fix up their code (see
> >examples of scsi log messages being there for _years_ about obsolete
> >driver interfaces being used.)
> 
> This change is not a simple case of replacing legacy with pm ops. I have
> been working on changing drivers that use legacy, in some cases it is an
> easy change, however in some cases it requires adding new interfaces to port
> over from legacy to pm ops. Legacy suspend/resume is more course grain and
> pm ops offer fine grain control over suspend, hibernate etc. Also it has
> been a challenge without the hardware to test.
> 
> My thinking is adding warning might get the attention of individual driver
> owners.

If it's not a trivial change, and takes a lot of work by someone who
knows this type of thing (i.e. you), then I wouldn't start making kernel
log warnings about this, it's just going to annoy a bunch of people :)

Again, see the SCSI log warnings for examples of this, people just tune
them out, assuming someone else is going to fix the issue, someday.  And
that someday never comes...

thanks,

greg k-h

Re: [PATCH v2 1/3] drivers/bus: Add Legacy PM OPS usage check and warning to bus_register()

[Shuah Khan]

On 01/06/2014 02:29 PM, Greg KH wrote:
> On Mon, Jan 06, 2014 at 01:30:31PM -0700, Shuah Khan wrote:
>> On 01/06/2014 01:17 PM, Greg KH wrote:
>>> On Mon, Jan 06, 2014 at 01:03:21PM -0700, Shuah Khan wrote:
>>>> Add Legacy PM OPS usage checks to bus_register() function. If Legacy PM OPS
>>>> usage is found, print warning message to indicate that the driver code needs
>>>> updating to use Dev PM OPS interfaces. This will help serve as a way to track
>>>> drivers that still use Legacy PM OPS and fix them.
>>>>
>>>> The Legacy PM OPS check looks for suspend(struct device *, pm_message_t) or
>>>> resume(struct device *) bus level interfaces.
>>>>
>>>> Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
>>>> ---
>>>>   drivers/base/bus.c | 3 +++
>>>>   1 file changed, 3 insertions(+)
>>>>
>>>> diff --git a/drivers/base/bus.c b/drivers/base/bus.c
>>>> index 73f6c29..e8753a7 100644
>>>> --- a/drivers/base/bus.c
>>>> +++ b/drivers/base/bus.c
>>>> @@ -916,6 +916,9 @@ int bus_register(struct bus_type *bus)
>>>>   		goto bus_groups_fail;
>>>>
>>>>   	pr_debug("bus: '%s': registered\n", bus->name);
>>>> +	if (bus->suspend || bus->resume)
>>>> +		pr_warn("bus '%s' needs updating - use pm pointer.\n",
>>>> +			bus->name);
>>>
>>> Why can't we just sweep the tree for all of these now, fix them up, and
>>> then delete these fields and be done with it?
>>>
>>> Same for the other ones, putting warnings in the kernel log files
>>> doesn't work at all for getting people to fix up their code (see
>>> examples of scsi log messages being there for _years_ about obsolete
>>> driver interfaces being used.)
>>
>> This change is not a simple case of replacing legacy with pm ops. I have
>> been working on changing drivers that use legacy, in some cases it is an
>> easy change, however in some cases it requires adding new interfaces to port
>> over from legacy to pm ops. Legacy suspend/resume is more course grain and
>> pm ops offer fine grain control over suspend, hibernate etc. Also it has
>> been a challenge without the hardware to test.
>>
>> My thinking is adding warning might get the attention of individual driver
>> owners.
>
> If it's not a trivial change, and takes a lot of work by someone who
> knows this type of thing (i.e. you), then I wouldn't start making kernel
> log warnings about this, it's just going to annoy a bunch of people :)

ok I will continue fixing the drivers. :)

>
> Again, see the SCSI log warnings for examples of this, people just tune
> them out, assuming someone else is going to fix the issue, someday.  And
> that someday never comes...
>

Yeah. Warnings could just get ignored.

-- Shuah

-- 
Shuah Khan
Senior Linux Kernel Developer - Open Source Group
Samsung Research America(Silicon Valley)
shuah.kh@samsung.com | (970) 672-0658

[PATCH v2 2/3] drivers/class: Add Legacy PM OPS usage check and warning to __class_register()

[Shuah Khan]

Add Legacy PM OPS usage checks to __class_register() function. If Legacy PM OPS
usage is found, print warning message to indicate that the driver code needs
updating to use Dev PM OPS interfaces. This will help serve as a way to track
drivers that still use Legacy PM OPS and fix them.

The Legacy PM OPS check looks for suspend(struct device *, pm_message_t) or
resume(struct device *) class level interfaces.

Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
---
 drivers/base/class.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/base/class.c b/drivers/base/class.c
index f96f704..2f20be5 100644
--- a/drivers/base/class.c
+++ b/drivers/base/class.c
@@ -169,6 +169,10 @@ int __class_register(struct class *cls, struct lock_class_key *key)
 
 	pr_debug("device class '%s': registering\n", cls->name);
 
+	if (cls->suspend || cls->resume)
+		pr_warn("device class '%s' needs updating - use pm pointer.\n",
+			cls->name);
+
 	cp = kzalloc(sizeof(*cp), GFP_KERNEL);
 	if (!cp)
 		return -ENOMEM;
-- 
1.8.3.2

[PATCH v2 3/3] driver: Add Legacy PM OPS usage check and warning to driver_register()

[Shuah Khan]

Add Legacy PM OPS usage checks to driver_register() function. If Legacy PM OPS
usage is found, print warning message to indicate that the driver code needs
updating to use Dev PM OPS interfaces. This will help serve as a way to track
drivers that still use Legacy PM OPS and fix them.

The Legacy PM OPS check looks for suspend(struct device *, pm_message_t) or
resume(struct device *) struct device_driver interfaces.

Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
---
 drivers/base/driver.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/base/driver.c b/drivers/base/driver.c
index 9e29943..7838a8a 100644
--- a/drivers/base/driver.c
+++ b/drivers/base/driver.c
@@ -157,6 +157,10 @@ int driver_register(struct device_driver *drv)
 		printk(KERN_WARNING "Driver '%s' needs updating - please use "
 			"bus_type methods\n", drv->name);
 
+	if (drv->suspend || drv->resume)
+		pr_warn("Driver '%s' needs updating - use pm pointer.\n",
+			drv->name);
+
 	other = driver_find(drv->name, drv->bus);
 	if (other) {
 		printk(KERN_ERR "Error: Driver '%s' is already registered, "
-- 
1.8.3.2