Re: DSS2 panel question

Re: DSS2 panel question

[Tomi Valkeinen]

On Sat, 2009-11-21 at 16:16 +0100, ext Grazvydas Ignotas wrote:
> Hello,
> 
> I've been updating Pandora's panel for upstream-going DSS2, which now
> uses device/device_driver model for panels. The thing is that the
> panel we use is SPI controlled, so it needs to be set up as
> spi_driver. From SPI probe function I can do
> omap_dss_register_driver(), but how do I pass SPI data to
> omap_dss_device cleanly? Surely some globals would help, but this is
> not a "clean" way I guess. There used to be panel-n800.c with similar
> requirements, but it's not yet converted, right? Do you have any ideas
> how this can be done?

I don't have any good answer for this. For N900's panel there was a
static global variable, if I remember right. Not very clean, as you
said.

 Tomi

Re: DSS2 panel question

[TAO HU]

Hi,

One idea as below

struct spi_dss_device {
             struct omap_dss_device dss_dev;
             struct spi_dss_dev_priv  * spi_data;
} spi_dss_dev;


spi_dss_probe.c

spi_probe(spi_device * spi)
{
    spi_dss_dev. spi_data = spi;
    omap_dss_register_device((spi_ device *) &spi_dss_dev);

 }

 spi_dss_drv.c
 spi_dss_drv_probe(omap_dss_device * dev)
 {
        my_spi_dss_dev = container_of(dev, struct spi_dss_device, dss_dev);
        spi_data = my_spi_dss_dev;
 }

- Hu Tao

On Mon, Nov 23, 2009 at 7:04 PM, Tomi Valkeinen
<tomi.valkeinen@nokia.com> wrote:
>
> On Sat, 2009-11-21 at 16:16 +0100, ext Grazvydas Ignotas wrote:
> > Hello,
> >
> > I've been updating Pandora's panel for upstream-going DSS2, which now
> > uses device/device_driver model for panels. The thing is that the
> > panel we use is SPI controlled, so it needs to be set up as
> > spi_driver. From SPI probe function I can do
> > omap_dss_register_driver(), but how do I pass SPI data to
> > omap_dss_device cleanly? Surely some globals would help, but this is
> > not a "clean" way I guess. There used to be panel-n800.c with similar
> > requirements, but it's not yet converted, right? Do you have any ideas
> > how this can be done?
>
> I don't have any good answer for this. For N900's panel there was a
> static global variable, if I remember right. Not very clean, as you
> said.
>
>  Tomi
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-omap" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: DSS2 panel question

[Grazvydas Ignotas]

On Mon, Nov 23, 2009 at 7:41 PM, TAO HU <tghk48@motorola.com> wrote:
> Hi,
>
> One idea as below
>
> struct spi_dss_device {
>              struct omap_dss_device dss_dev;
>              struct spi_dss_dev_priv  * spi_data;
> } spi_dss_dev;

Hm, omap_dss_register_device() takes omap_dss_driver, not
omap_dss_device. Anyway that wouldn't work if there were 2 instances
of same panel in one system I guess, as 2 different spi_data instances
would probe, when there is only one omap_dss_driver.

Instead I'm thinking to link omap_dss_device with spi in board file
itself by (ab)using platform_data:

static struct omap_dss_device pandora_lcd_device = {
...
};

struct spi_board_info pandora_spi_board_info = {
  .bus_num = 1,
  ...
  .platform_data = pandora_lcd_device,
};

Then in panel driver:

int spi_probe(struct spi_device *spi)
{
  struct omap_dss_device *dssdev = spi->dev.platform_data;
  dev_set_drvdata(&dssdev->dev, spi);
}

static int lcd_probe(struct omap_dss_device *dssdev)
{
  struct spi_device *spi = dev_get_drvdata(&dssdev->dev);
}

Tomi, does that look reasonable to you?
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: DSS2 panel question

[TAO HU]

Hi, Grazvydas Ignotas

What did you mean "Hm, omap_dss_register_device() takes omap_dss_driver"?

Below is what I saw from the code
 int omap_dss_register_device(struct omap_dss_device *dssdev)

I assume the code I have is for DSS. Does DSS2 change the API?
BTW, what's the location for DSS2 code in linux-omap tree? Seems not
see it anywhere


- Hu Tao


On Wed, Nov 25, 2009 at 6:53 PM, Grazvydas Ignotas <notasas@gmail.com> wrote:
> On Mon, Nov 23, 2009 at 7:41 PM, TAO HU <tghk48@motorola.com> wrote:
>> Hi,
>>
>> One idea as below
>>
>> struct spi_dss_device {
>>              struct omap_dss_device dss_dev;
>>              struct spi_dss_dev_priv  * spi_data;
>> } spi_dss_dev;
>
> Hm, omap_dss_register_device() takes omap_dss_driver, not
> omap_dss_device. Anyway that wouldn't work if there were 2 instances
> of same panel in one system I guess, as 2 different spi_data instances
> would probe, when there is only one omap_dss_driver.
>
> Instead I'm thinking to link omap_dss_device with spi in board file
> itself by (ab)using platform_data:
>
> static struct omap_dss_device pandora_lcd_device = {
> ...
> };
>
> struct spi_board_info pandora_spi_board_info = {
>  .bus_num = 1,
>  ...
>  .platform_data = pandora_lcd_device,
> };
>
> Then in panel driver:
>
> int spi_probe(struct spi_device *spi)
> {
>  struct omap_dss_device *dssdev = spi->dev.platform_data;
>  dev_set_drvdata(&dssdev->dev, spi);
> }
>
> static int lcd_probe(struct omap_dss_device *dssdev)
> {
>  struct spi_device *spi = dev_get_drvdata(&dssdev->dev);
> }
>
> Tomi, does that look reasonable to you?
>
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: DSS2 panel question

[Grazvydas Ignotas]

On Wed, Nov 25, 2009 at 7:44 PM, TAO HU <tghk48@motorola.com> wrote:
> Hi, Grazvydas Ignotas
>
> What did you mean "Hm, omap_dss_register_device() takes omap_dss_driver"?

Whoops, I meant omap_dss_register_driver()

> Below is what I saw from the code
>  int omap_dss_register_device(struct omap_dss_device *dssdev)
>
> I assume the code I have is for DSS. Does DSS2 change the API?

I was referring to version that is being submitted to mainline, and
panel drivers in particular, like:
http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=blob;f=drivers/video/omap2/displays/panel-generic.c;h=eb48d1afd8008437458bb3fe87e351b5ce785c42;hb=HEAD

> BTW, what's the location for DSS2 code in linux-omap tree? Seems not
> see it anywhere

Tony is probably waiting for it to reach mainline and fall from there.

>
> - Hu Tao
>
>
> On Wed, Nov 25, 2009 at 6:53 PM, Grazvydas Ignotas <notasas@gmail.com> wrote:
>> On Mon, Nov 23, 2009 at 7:41 PM, TAO HU <tghk48@motorola.com> wrote:
>>> Hi,
>>>
>>> One idea as below
>>>
>>> struct spi_dss_device {
>>>              struct omap_dss_device dss_dev;
>>>              struct spi_dss_dev_priv  * spi_data;
>>> } spi_dss_dev;
>>
>> Hm, omap_dss_register_device() takes omap_dss_driver, not
>> omap_dss_device. Anyway that wouldn't work if there were 2 instances
>> of same panel in one system I guess, as 2 different spi_data instances
>> would probe, when there is only one omap_dss_driver.
>>
>> Instead I'm thinking to link omap_dss_device with spi in board file
>> itself by (ab)using platform_data:
>>
>> static struct omap_dss_device pandora_lcd_device = {
>> ...
>> };
>>
>> struct spi_board_info pandora_spi_board_info = {
>>  .bus_num = 1,
>>  ...
>>  .platform_data = pandora_lcd_device,
>> };
>>
>> Then in panel driver:
>>
>> int spi_probe(struct spi_device *spi)
>> {
>>  struct omap_dss_device *dssdev = spi->dev.platform_data;
>>  dev_set_drvdata(&dssdev->dev, spi);
>> }
>>
>> static int lcd_probe(struct omap_dss_device *dssdev)
>> {
>>  struct spi_device *spi = dev_get_drvdata(&dssdev->dev);
>> }
>>
>> Tomi, does that look reasonable to you?
>>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: DSS2 panel question

[Tony Lindgren]

* Grazvydas Ignotas <notasas@gmail.com> [091125 12:23]:
> On Wed, Nov 25, 2009 at 7:44 PM, TAO HU <tghk48@motorola.com> wrote:
> > Hi, Grazvydas Ignotas
> >
> > What did you mean "Hm, omap_dss_register_device() takes omap_dss_driver"?
> 
> Whoops, I meant omap_dss_register_driver()
> 
> > Below is what I saw from the code
> >  int omap_dss_register_device(struct omap_dss_device *dssdev)
> >
> > I assume the code I have is for DSS. Does DSS2 change the API?
> 
> I was referring to version that is being submitted to mainline, and
> panel drivers in particular, like:
> http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=blob;f=drivers/video/omap2/displays/panel-generic.c;h=eb48d1afd8008437458bb3fe87e351b5ce785c42;hb=HEAD
> 
> > BTW, what's the location for DSS2 code in linux-omap tree? Seems not
> > see it anywhere
> 
> Tony is probably waiting for it to reach mainline and fall from there.

Pulled Tomi's for-next now into linux-omap master branch for some last
testing before the merge window starts :)

Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: DSS2 panel question

[Tomi Valkeinen]

On Wed, 2009-11-25 at 11:53 +0100, ext Grazvydas Ignotas wrote:
> On Mon, Nov 23, 2009 at 7:41 PM, TAO HU <tghk48@motorola.com> wrote:
> > Hi,
> >
> > One idea as below
> >
> > struct spi_dss_device {
> >              struct omap_dss_device dss_dev;
> >              struct spi_dss_dev_priv  * spi_data;
> > } spi_dss_dev;
> 
> Hm, omap_dss_register_device() takes omap_dss_driver, not
> omap_dss_device. Anyway that wouldn't work if there were 2 instances
> of same panel in one system I guess, as 2 different spi_data instances
> would probe, when there is only one omap_dss_driver.
> 
> Instead I'm thinking to link omap_dss_device with spi in board file
> itself by (ab)using platform_data:
> 
> static struct omap_dss_device pandora_lcd_device = {
> ...
> };
> 
> struct spi_board_info pandora_spi_board_info = {
>   .bus_num = 1,
>   ...
>   .platform_data = pandora_lcd_device,
> };
> 
> Then in panel driver:
> 
> int spi_probe(struct spi_device *spi)
> {
>   struct omap_dss_device *dssdev = spi->dev.platform_data;
>   dev_set_drvdata(&dssdev->dev, spi);
> }
> 
> static int lcd_probe(struct omap_dss_device *dssdev)
> {
>   struct spi_device *spi = dev_get_drvdata(&dssdev->dev);
> }
> 
> Tomi, does that look reasonable to you?

Yes, the linking has to be done inside the board file somehow. I need to
try this at some point with N900's panel driver, but I think it looks
good.

I think you could also link it other way around, store the
spi_board_info inside the dss_device, but I don't immediately see that
this way is better than the other.

 Tomi