[PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[Kuninori Morimoto]

mt9t112 camera needs 100 milliseconds for initializing
Special thanks to Phil

Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
Reported-by: Phil Edworthy <Phil.Edworthy@renesas.com>
---
 drivers/media/video/mt9t112.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/media/video/mt9t112.c b/drivers/media/video/mt9t112.c
index 7438f8d..e581d8a 100644
--- a/drivers/media/video/mt9t112.c
+++ b/drivers/media/video/mt9t112.c
@@ -885,7 +885,7 @@ static int mt9t112_s_stream(struct v4l2_subdev *sd, int enable)
 		/* Invert PCLK (Data sampled on falling edge of pixclk) */
 		mt9t112_reg_write(ret, client, 0x3C20, param);
 
-		mdelay(5);
+		mdelay(100);
 
 		priv->flags |= INIT_DONE;
 	}
-- 
1.6.3.3

Re: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[Mauro Carvalho Chehab]

Em 02-02-2010 02:54, Kuninori Morimoto escreveu:
> mt9t112 camera needs 100 milliseconds for initializing
> Special thanks to Phil
> 
> Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
> Reported-by: Phil Edworthy <Phil.Edworthy@renesas.com>
> ---
>  drivers/media/video/mt9t112.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/media/video/mt9t112.c b/drivers/media/video/mt9t112.c
> index 7438f8d..e581d8a 100644
> --- a/drivers/media/video/mt9t112.c
> +++ b/drivers/media/video/mt9t112.c
> @@ -885,7 +885,7 @@ static int mt9t112_s_stream(struct v4l2_subdev *sd, int enable)
>  		/* Invert PCLK (Data sampled on falling edge of pixclk) */
>  		mt9t112_reg_write(ret, client, 0x3C20, param);
>  
> -		mdelay(5);
> +		mdelay(100);
>  
>  		priv->flags |= INIT_DONE;
>  	}

Hi Guennadi,

What's the status of this patch?

It applies ok for me, and I couldn't find any reference at the
ML why it was not applied yet.

Thanks,
Mauro

Re: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[Guennadi Liakhovetski]

On Sun, 18 Sep 2011, Mauro Carvalho Chehab wrote:

> Em 02-02-2010 02:54, Kuninori Morimoto escreveu:
> > mt9t112 camera needs 100 milliseconds for initializing
> > Special thanks to Phil
> > 
> > Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
> > Reported-by: Phil Edworthy <Phil.Edworthy@renesas.com>
> > ---
> >  drivers/media/video/mt9t112.c |    2 +-
> >  1 files changed, 1 insertions(+), 1 deletions(-)
> > 
> > diff --git a/drivers/media/video/mt9t112.c b/drivers/media/video/mt9t112.c
> > index 7438f8d..e581d8a 100644
> > --- a/drivers/media/video/mt9t112.c
> > +++ b/drivers/media/video/mt9t112.c
> > @@ -885,7 +885,7 @@ static int mt9t112_s_stream(struct v4l2_subdev *sd, int enable)
> >  		/* Invert PCLK (Data sampled on falling edge of pixclk) */
> >  		mt9t112_reg_write(ret, client, 0x3C20, param);
> >  
> > -		mdelay(5);
> > +		mdelay(100);
> >  
> >  		priv->flags |= INIT_DONE;
> >  	}
> 
> Hi Guennadi,
> 
> What's the status of this patch?
> 
> It applies ok for me, and I couldn't find any reference at the
> ML why it was not applied yet.

Hm, yeah... Looks like also this patch:

> Subject: [PATCH 3/3] soc-camera: mt9t112: The flag which control camera-init is removed
> 
> mt9t112 should always be initialized when camera start.
> Because current driver doesn't run this operation,
> it will be un-stable if user side player run open/close several times.
> Special thanks to Namiki-san
> 
> Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
> Reported-by: Takashi Namiki <Takashi.Namiki@renesas.com>

has not been applied nor discussed on the list... For patches that old I 
would tend to say: if the author / submitter didn't re-submit, then, 
probably, patches aren't relevant anymore... Although it is quite 
possible, that I failed to process them back then. Morimoto-san, do you 
have any information on these patches? Have these problems been solved 
somehow, so that the patches have become obsolete, or are the problems, 
that they address, still there?

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[Kuninori Morimoto]

Hi Guennadi, all

> > > mt9t112 camera needs 100 milliseconds for initializing
> > > Special thanks to Phil
> > > 
> > > Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
> > > Reported-by: Phil Edworthy <Phil.Edworthy@renesas.com>
> > > ---
> > >  drivers/media/video/mt9t112.c |    2 +-
> > >  1 files changed, 1 insertions(+), 1 deletions(-)
> > > 
> > > diff --git a/drivers/media/video/mt9t112.c b/drivers/media/video/mt9t112.c
> > > index 7438f8d..e581d8a 100644
> > > --- a/drivers/media/video/mt9t112.c
> > > +++ b/drivers/media/video/mt9t112.c
> > > @@ -885,7 +885,7 @@ static int mt9t112_s_stream(struct v4l2_subdev *sd, int enable)
> > >  		/* Invert PCLK (Data sampled on falling edge of pixclk) */
> > >  		mt9t112_reg_write(ret, client, 0x3C20, param);
> > >  
> > > -		mdelay(5);
> > > +		mdelay(100);
> > >  
> > >  		priv->flags |= INIT_DONE;
> > >  	}
> > 
> > Hi Guennadi,
> > 
> > What's the status of this patch?
> > 
> > It applies ok for me, and I couldn't find any reference at the
> > ML why it was not applied yet.
> 
> Hm, yeah... Looks like also this patch:
> 
> > Subject: [PATCH 3/3] soc-camera: mt9t112: The flag which control camera-init is removed
> > 
> > mt9t112 should always be initialized when camera start.
> > Because current driver doesn't run this operation,
> > it will be un-stable if user side player run open/close several times.
> > Special thanks to Namiki-san
> > 
> > Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
> > Reported-by: Takashi Namiki <Takashi.Namiki@renesas.com>
> 
> has not been applied nor discussed on the list... For patches that old I 
> would tend to say: if the author / submitter didn't re-submit, then, 
> probably, patches aren't relevant anymore... Although it is quite 
> possible, that I failed to process them back then. Morimoto-san, do you 
> have any information on these patches? Have these problems been solved 
> somehow, so that the patches have become obsolete, or are the problems, 
> that they address, still there?

This patch is needed for mt9t112 camera initialize.
I thought that it was already applied.

Best regards
---
Kuninori Morimoto

Re: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[Guennadi Liakhovetski]

Morimoto-san

Thanks for your reply.

On Mon, 19 Sep 2011, Kuninori Morimoto wrote:

> Hi Guennadi, all
> 
> > > > mt9t112 camera needs 100 milliseconds for initializing
> > > > Special thanks to Phil
> > > > 
> > > > Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
> > > > Reported-by: Phil Edworthy <Phil.Edworthy@renesas.com>
> > > > ---
> > > >  drivers/media/video/mt9t112.c |    2 +-
> > > >  1 files changed, 1 insertions(+), 1 deletions(-)
> > > > 
> > > > diff --git a/drivers/media/video/mt9t112.c b/drivers/media/video/mt9t112.c
> > > > index 7438f8d..e581d8a 100644
> > > > --- a/drivers/media/video/mt9t112.c
> > > > +++ b/drivers/media/video/mt9t112.c
> > > > @@ -885,7 +885,7 @@ static int mt9t112_s_stream(struct v4l2_subdev *sd, int enable)
> > > >  		/* Invert PCLK (Data sampled on falling edge of pixclk) */
> > > >  		mt9t112_reg_write(ret, client, 0x3C20, param);
> > > >  
> > > > -		mdelay(5);
> > > > +		mdelay(100);
> > > >  
> > > >  		priv->flags |= INIT_DONE;
> > > >  	}
> > > 
> > > Hi Guennadi,
> > > 
> > > What's the status of this patch?
> > > 
> > > It applies ok for me, and I couldn't find any reference at the
> > > ML why it was not applied yet.
> > 
> > Hm, yeah... Looks like also this patch:
> > 
> > > Subject: [PATCH 3/3] soc-camera: mt9t112: The flag which control camera-init is removed
> > > 
> > > mt9t112 should always be initialized when camera start.
> > > Because current driver doesn't run this operation,
> > > it will be un-stable if user side player run open/close several times.
> > > Special thanks to Namiki-san
> > > 
> > > Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
> > > Reported-by: Takashi Namiki <Takashi.Namiki@renesas.com>
> > 
> > has not been applied nor discussed on the list... For patches that old I 
> > would tend to say: if the author / submitter didn't re-submit, then, 
> > probably, patches aren't relevant anymore... Although it is quite 
> > possible, that I failed to process them back then. Morimoto-san, do you 
> > have any information on these patches? Have these problems been solved 
> > somehow, so that the patches have become obsolete, or are the problems, 
> > that they address, still there?
> 
> This patch is needed for mt9t112 camera initialize.
> I thought that it was already applied.

Which patch do you mean? Patch 2/3, or 3/3, or both are needed?

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[Guennadi Liakhovetski]

Morimoto-san

There was a question at the bottom of this email, which you might have 
overseen:-) Could you give me an idea, which patche(es) exactly you meant?

Thanks
Guennadi

On Tue, 20 Sep 2011, Guennadi Liakhovetski wrote:

> Morimoto-san
> 
> Thanks for your reply.
> 
> On Mon, 19 Sep 2011, Kuninori Morimoto wrote:
> 
> > Hi Guennadi, all
> > 
> > > > > mt9t112 camera needs 100 milliseconds for initializing
> > > > > Special thanks to Phil
> > > > > 
> > > > > Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
> > > > > Reported-by: Phil Edworthy <Phil.Edworthy@renesas.com>
> > > > > ---
> > > > >  drivers/media/video/mt9t112.c |    2 +-
> > > > >  1 files changed, 1 insertions(+), 1 deletions(-)
> > > > > 
> > > > > diff --git a/drivers/media/video/mt9t112.c b/drivers/media/video/mt9t112.c
> > > > > index 7438f8d..e581d8a 100644
> > > > > --- a/drivers/media/video/mt9t112.c
> > > > > +++ b/drivers/media/video/mt9t112.c
> > > > > @@ -885,7 +885,7 @@ static int mt9t112_s_stream(struct v4l2_subdev *sd, int enable)
> > > > >  		/* Invert PCLK (Data sampled on falling edge of pixclk) */
> > > > >  		mt9t112_reg_write(ret, client, 0x3C20, param);
> > > > >  
> > > > > -		mdelay(5);
> > > > > +		mdelay(100);
> > > > >  
> > > > >  		priv->flags |= INIT_DONE;
> > > > >  	}
> > > > 
> > > > Hi Guennadi,
> > > > 
> > > > What's the status of this patch?
> > > > 
> > > > It applies ok for me, and I couldn't find any reference at the
> > > > ML why it was not applied yet.
> > > 
> > > Hm, yeah... Looks like also this patch:
> > > 
> > > > Subject: [PATCH 3/3] soc-camera: mt9t112: The flag which control camera-init is removed
> > > > 
> > > > mt9t112 should always be initialized when camera start.
> > > > Because current driver doesn't run this operation,
> > > > it will be un-stable if user side player run open/close several times.
> > > > Special thanks to Namiki-san
> > > > 
> > > > Signed-off-by: Kuninori Morimoto <morimoto.kuninori@renesas.com>
> > > > Reported-by: Takashi Namiki <Takashi.Namiki@renesas.com>
> > > 
> > > has not been applied nor discussed on the list... For patches that old I 
> > > would tend to say: if the author / submitter didn't re-submit, then, 
> > > probably, patches aren't relevant anymore... Although it is quite 
> > > possible, that I failed to process them back then. Morimoto-san, do you 
> > > have any information on these patches? Have these problems been solved 
> > > somehow, so that the patches have become obsolete, or are the problems, 
> > > that they address, still there?
> > 
> > This patch is needed for mt9t112 camera initialize.
> > I thought that it was already applied.
> 
> Which patch do you mean? Patch 2/3, or 3/3, or both are needed?
> 
> Thanks
> Guennadi
> ---
> Guennadi Liakhovetski, Ph.D.
> Freelance Open-Source Software Developer
> http://www.open-technology.de/
> 

---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[Kuninori Morimoto]

Hi Guennadi

> There was a question at the bottom of this email, which you might have 
> overseen:-) Could you give me an idea, which patche(es) exactly you meant?

sorry for my super late response.
I losted this email.


> > > > > Subject: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize
(snip)
> > > > > Subject: [PATCH 3/3] soc-camera: mt9t112: The flag which control camera-init is removed

> > > This patch is needed for mt9t112 camera initialize.
> > > I thought that it was already applied.
> > 
> > Which patch do you mean? Patch 2/3, or 3/3, or both are needed?

Both are needed.
These are bug fix patches

Best regards
---
Kuninori Morimoto

Re: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[Guennadi Liakhovetski]

Hi all

On Wed, 12 Oct 2011, Kuninori Morimoto wrote:

> Hi Guennadi
> 
> > There was a question at the bottom of this email, which you might have 
> > overseen:-) Could you give me an idea, which patche(es) exactly you meant?
> 
> sorry for my super late response.
> I losted this email.
> 
> 
> > > > > > Subject: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize
> (snip)
> > > > > > Subject: [PATCH 3/3] soc-camera: mt9t112: The flag which control camera-init is removed
> 
> > > > This patch is needed for mt9t112 camera initialize.
> > > > I thought that it was already applied.
> > > 
> > > Which patch do you mean? Patch 2/3, or 3/3, or both are needed?
> 
> Both are needed.
> These are bug fix patches

I tried to capture several frames beginning with the very first one (as 
much as performance allowed), and I do see several black or wrongly 
coloured framed in the beginning, but none of those patches, including the 
proposed 300ms at the end of .s_stream() fixes the problem reliably. So, 
either this problems, that these patches fix, are specific to the Solution 
Engine board (is it the one, where the problems have been observed?), or 
one needs a different testing method. If they are SE-specific - I don't 
think, getting those fixes in the driver is very important, because 
mt9t112 data for SE is not in the mainline. If I was testing wrongly, 
please, tell me how exactly to reproduce those problems and see, how one 
or another patch fixes them.

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[Kuninori Morimoto]

Hi Guennadi, and Phil

Thank you for your check/test

> > Both are needed.
> > These are bug fix patches
> 
> I tried to capture several frames beginning with the very first one (as 
> much as performance allowed), and I do see several black or wrongly 
> coloured framed in the beginning, but none of those patches, including the 
> proposed 300ms at the end of .s_stream() fixes the problem reliably. So, 
> either this problems, that these patches fix, are specific to the Solution 
> Engine board (is it the one, where the problems have been observed?), or 
> one needs a different testing method. If they are SE-specific - I don't 
> think, getting those fixes in the driver is very important, because 
> mt9t112 data for SE is not in the mainline. If I was testing wrongly, 
> please, tell me how exactly to reproduce those problems and see, how one 
> or another patch fixes them.

I guess mt9t112 camera is used in SE (with local circuit ?)
and Ecovec.
But I forgot detail of this issue (I have no mt9t112 for now).

I think Phil is the person who wanted this patch.

But if this issue happen only SE, I can agree to Guennadi

Best regards
---
Kuninori Morimoto

Re: [PATCH 2/3] soc-camera: mt9t112: modify delay time after initialize

[phil.edworthy]

Hi Guennadi, Morimoto-san,

> > > Both are needed.
> > > These are bug fix patches
> > 
> > I tried to capture several frames beginning with the very first one 
(as 
> > much as performance allowed), and I do see several black or wrongly 
> > coloured framed in the beginning, but none of those patches, including 
the 
> > proposed 300ms at the end of .s_stream() fixes the problem reliably. 
So, 
> > either this problems, that these patches fix, are specific to the 
Solution 
> > Engine board (is it the one, where the problems have been observed?), 
or 
> > one needs a different testing method. If they are SE-specific - I 
don't 
> > think, getting those fixes in the driver is very important, because 
> > mt9t112 data for SE is not in the mainline. If I was testing wrongly, 
> > please, tell me how exactly to reproduce those problems and see, how 
one 
> > or another patch fixes them.
> 
> I guess mt9t112 camera is used in SE (with local circuit ?)
> and Ecovec.
> But I forgot detail of this issue (I have no mt9t112 for now).
> 
> I think Phil is the person who wanted this patch.

There are capture issues on the Ecovec board with this camera. iirc, these 
patches made the situation better but still didn't completely fix all 
issues. Morimoto-san has made comments previously that the mt9t112 is a 
little difficult to setup and we don't have the relevant information from 
the manufacturer.

Thanks
Phil