linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 0/2] v4l: vsp1: crop and a single input issues
@ 2014-11-26  6:19 Takanari Hayama
  2014-11-26  6:19 ` [PATCH 1/2] v4l: vsp1: Reset VSP1 RPF source address Takanari Hayama
  2014-11-26  6:19 ` [PATCH 2/2] v4l: vsp1: Always enable virtual RPF when BRU is in use Takanari Hayama
  0 siblings, 2 replies; 7+ messages in thread
From: Takanari Hayama @ 2014-11-26  6:19 UTC (permalink / raw)
  To: linux-media; +Cc: linux-sh

Hi,

We've found a couple of issues in VSP1 driver. Each issue is described
in each patch. They can be applied separetely.

Takanari Hayama (2):
  v4l: vsp1: Reset VSP1 RPF source address
  v4l: vsp1: Always enable virtual RPF when BRU is in use

 drivers/media/platform/vsp1/vsp1_rpf.c  | 15 +++++++++++++++
 drivers/media/platform/vsp1/vsp1_rwpf.h |  2 ++
 drivers/media/platform/vsp1/vsp1_wpf.c  | 11 ++++++-----
 3 files changed, 23 insertions(+), 5 deletions(-)

Best Regards,
Takanari Hayama
-- 
1.8.0


^ permalink raw reply	[flat|nested] 7+ messages in thread

* [PATCH 1/2] v4l: vsp1: Reset VSP1 RPF source address
  2014-11-26  6:19 [PATCH 0/2] v4l: vsp1: crop and a single input issues Takanari Hayama
@ 2014-11-26  6:19 ` Takanari Hayama
  2014-11-26  8:59   ` Geert Uytterhoeven
  2014-11-26  6:19 ` [PATCH 2/2] v4l: vsp1: Always enable virtual RPF when BRU is in use Takanari Hayama
  1 sibling, 1 reply; 7+ messages in thread
From: Takanari Hayama @ 2014-11-26  6:19 UTC (permalink / raw)
  To: linux-media; +Cc: linux-sh

Source address of VSP1 RPF needs to be reset whenever crop offsets are
recalculated.

This correctly reflects a crop setting even VIDIOC_QBUF is called
before VIDIOIC_STREAMON is called.

Signed-off-by: Takanari Hayama <taki@igel.co.jp>
---
 drivers/media/platform/vsp1/vsp1_rpf.c  | 15 +++++++++++++++
 drivers/media/platform/vsp1/vsp1_rwpf.h |  2 ++
 2 files changed, 17 insertions(+)

diff --git a/drivers/media/platform/vsp1/vsp1_rpf.c b/drivers/media/platform/vsp1/vsp1_rpf.c
index d14d26b..79c0db8 100644
--- a/drivers/media/platform/vsp1/vsp1_rpf.c
+++ b/drivers/media/platform/vsp1/vsp1_rpf.c
@@ -106,11 +106,22 @@ static int rpf_s_stream(struct v4l2_subdev *subdev, int enable)
 			+ crop->left * fmtinfo->bpp[0] / 8;
 	pstride = format->plane_fmt[0].bytesperline
 		<< VI6_RPF_SRCM_PSTRIDE_Y_SHIFT;
+
+	vsp1_rpf_write(rpf, VI6_RPF_SRCM_ADDR_Y,
+		       rpf->buf_addr[0] + rpf->offsets[0]);
+
 	if (format->num_planes > 1) {
 		rpf->offsets[1] = crop->top * format->plane_fmt[1].bytesperline
 				+ crop->left * fmtinfo->bpp[1] / 8;
 		pstride |= format->plane_fmt[1].bytesperline
 			<< VI6_RPF_SRCM_PSTRIDE_C_SHIFT;
+
+		vsp1_rpf_write(rpf, VI6_RPF_SRCM_ADDR_C0,
+			       rpf->buf_addr[1] + rpf->offsets[1]);
+
+		if (format->num_planes > 2)
+			vsp1_rpf_write(rpf, VI6_RPF_SRCM_ADDR_C1,
+				       rpf->buf_addr[2] + rpf->offsets[1]);
 	}
 
 	vsp1_rpf_write(rpf, VI6_RPF_SRCM_PSTRIDE, pstride);
@@ -179,6 +190,10 @@ static void rpf_vdev_queue(struct vsp1_video *video,
 			   struct vsp1_video_buffer *buf)
 {
 	struct vsp1_rwpf *rpf = container_of(video, struct vsp1_rwpf, video);
+	int i;
+
+	for (i = 0; i < 3; i++)
+		rpf->buf_addr[i] = buf->addr[i];
 
 	vsp1_rpf_write(rpf, VI6_RPF_SRCM_ADDR_Y,
 		       buf->addr[0] + rpf->offsets[0]);
diff --git a/drivers/media/platform/vsp1/vsp1_rwpf.h b/drivers/media/platform/vsp1/vsp1_rwpf.h
index 28dd9e7..1f98fe3 100644
--- a/drivers/media/platform/vsp1/vsp1_rwpf.h
+++ b/drivers/media/platform/vsp1/vsp1_rwpf.h
@@ -39,6 +39,8 @@ struct vsp1_rwpf {
 	struct v4l2_rect crop;
 
 	unsigned int offsets[2];
+
+	unsigned int buf_addr[3];
 };
 
 static inline struct vsp1_rwpf *to_rwpf(struct v4l2_subdev *subdev)
-- 
1.8.0


^ permalink raw reply related	[flat|nested] 7+ messages in thread

* [PATCH 2/2] v4l: vsp1: Always enable virtual RPF when BRU is in use
  2014-11-26  6:19 [PATCH 0/2] v4l: vsp1: crop and a single input issues Takanari Hayama
  2014-11-26  6:19 ` [PATCH 1/2] v4l: vsp1: Reset VSP1 RPF source address Takanari Hayama
@ 2014-11-26  6:19 ` Takanari Hayama
  2014-11-26 12:55   ` Sergei Shtylyov
  1 sibling, 1 reply; 7+ messages in thread
From: Takanari Hayama @ 2014-11-26  6:19 UTC (permalink / raw)
  To: linux-media; +Cc: linux-sh

Regardless of a number of inputs, we should always enable virtual RPF
when BRU is used. This allows the case when there's only one input to
BRU, and a size of the input is smaller than a size of an output of BRU.

Signed-off-by: Takanari Hayama <taki@igel.co.jp>
---
 drivers/media/platform/vsp1/vsp1_wpf.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/media/platform/vsp1/vsp1_wpf.c b/drivers/media/platform/vsp1/vsp1_wpf.c
index 6e05776..29ea28b 100644
--- a/drivers/media/platform/vsp1/vsp1_wpf.c
+++ b/drivers/media/platform/vsp1/vsp1_wpf.c
@@ -92,19 +92,20 @@ static int wpf_s_stream(struct v4l2_subdev *subdev, int enable)
 		return 0;
 	}
 
-	/* Sources. If the pipeline has a single input configure it as the
-	 * master layer. Otherwise configure all inputs as sub-layers and
-	 * select the virtual RPF as the master layer.
+	/* Sources. If the pipeline has a single input and BRU is not used,
+	 * configure it as the master layer. Otherwise configure all
+	 * inputs as sub-layers and select the virtual RPF as the master
+	 * layer.
 	 */
 	for (i = 0; i < pipe->num_inputs; ++i) {
 		struct vsp1_rwpf *input = pipe->inputs[i];
 
-		srcrpf |= pipe->num_inputs == 1
+		srcrpf |= ((!pipe->bru) && (pipe->num_inputs == 1))
 			? VI6_WPF_SRCRPF_RPF_ACT_MST(input->entity.index)
 			: VI6_WPF_SRCRPF_RPF_ACT_SUB(input->entity.index);
 	}
 
-	if (pipe->num_inputs > 1)
+	if ((pipe->bru) || (pipe->num_inputs > 1))
 		srcrpf |= VI6_WPF_SRCRPF_VIRACT_MST;
 
 	vsp1_wpf_write(wpf, VI6_WPF_SRCRPF, srcrpf);
-- 
1.8.0


^ permalink raw reply related	[flat|nested] 7+ messages in thread

* Re: [PATCH 1/2] v4l: vsp1: Reset VSP1 RPF source address
  2014-11-26  6:19 ` [PATCH 1/2] v4l: vsp1: Reset VSP1 RPF source address Takanari Hayama
@ 2014-11-26  8:59   ` Geert Uytterhoeven
  2014-11-27  0:57     ` Takanari Hayama
  0 siblings, 1 reply; 7+ messages in thread
From: Geert Uytterhoeven @ 2014-11-26  8:59 UTC (permalink / raw)
  To: Takanari Hayama; +Cc: Linux Media Mailing List, Linux-sh list

Hi Hayama-san,

On Wed, Nov 26, 2014 at 7:19 AM, Takanari Hayama <taki@igel.co.jp> wrote:
> @@ -179,6 +190,10 @@ static void rpf_vdev_queue(struct vsp1_video *video,
>                            struct vsp1_video_buffer *buf)
>  {
>         struct vsp1_rwpf *rpf = container_of(video, struct vsp1_rwpf, video);
> +       int i;
> +
> +       for (i = 0; i < 3; i++)
> +               rpf->buf_addr[i] = buf->addr[i];

vsp1_video_buffer.addr is "dma_addr_t addr[3];"...

BTW, you can use memcpy() instead of an explicit loop.

>
>         vsp1_rpf_write(rpf, VI6_RPF_SRCM_ADDR_Y,
>                        buf->addr[0] + rpf->offsets[0]);
> diff --git a/drivers/media/platform/vsp1/vsp1_rwpf.h b/drivers/media/platform/vsp1/vsp1_rwpf.h
> index 28dd9e7..1f98fe3 100644
> --- a/drivers/media/platform/vsp1/vsp1_rwpf.h
> +++ b/drivers/media/platform/vsp1/vsp1_rwpf.h
> @@ -39,6 +39,8 @@ struct vsp1_rwpf {
>         struct v4l2_rect crop;
>
>         unsigned int offsets[2];
> +
> +       unsigned int buf_addr[3];

... hence the above should use dma_addr_t, too.

If CONFIG_ARM_LPAE is enabled, CONFIG_ARCH_DMA_ADDR_T_64BIT
will be enabled, too, and dma_addr_t will be u64.

>  };

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH 2/2] v4l: vsp1: Always enable virtual RPF when BRU is in use
  2014-11-26  6:19 ` [PATCH 2/2] v4l: vsp1: Always enable virtual RPF when BRU is in use Takanari Hayama
@ 2014-11-26 12:55   ` Sergei Shtylyov
  2014-11-27  1:01     ` Takanari Hayama
  0 siblings, 1 reply; 7+ messages in thread
From: Sergei Shtylyov @ 2014-11-26 12:55 UTC (permalink / raw)
  To: Takanari Hayama, linux-media; +Cc: linux-sh

Hello.

On 11/26/2014 9:19 AM, Takanari Hayama wrote:

> Regardless of a number of inputs, we should always enable virtual RPF
> when BRU is used. This allows the case when there's only one input to
> BRU, and a size of the input is smaller than a size of an output of BRU.

> Signed-off-by: Takanari Hayama <taki@igel.co.jp>
> ---
>   drivers/media/platform/vsp1/vsp1_wpf.c | 11 ++++++-----
>   1 file changed, 6 insertions(+), 5 deletions(-)

> diff --git a/drivers/media/platform/vsp1/vsp1_wpf.c b/drivers/media/platform/vsp1/vsp1_wpf.c
> index 6e05776..29ea28b 100644
> --- a/drivers/media/platform/vsp1/vsp1_wpf.c
> +++ b/drivers/media/platform/vsp1/vsp1_wpf.c
> @@ -92,19 +92,20 @@ static int wpf_s_stream(struct v4l2_subdev *subdev, int enable)
>   		return 0;
>   	}
>
> -	/* Sources. If the pipeline has a single input configure it as the
> -	 * master layer. Otherwise configure all inputs as sub-layers and
> -	 * select the virtual RPF as the master layer.
> +	/* Sources. If the pipeline has a single input and BRU is not used,
> +	 * configure it as the master layer. Otherwise configure all
> +	 * inputs as sub-layers and select the virtual RPF as the master
> +	 * layer.
>   	 */
>   	for (i = 0; i < pipe->num_inputs; ++i) {
>   		struct vsp1_rwpf *input = pipe->inputs[i];
>
> -		srcrpf |= pipe->num_inputs == 1
> +		srcrpf |= ((!pipe->bru) && (pipe->num_inputs == 1))

    Inner parens not needed, especially in the first case.

>   			? VI6_WPF_SRCRPF_RPF_ACT_MST(input->entity.index)
>   			: VI6_WPF_SRCRPF_RPF_ACT_SUB(input->entity.index);
>   	}
>
> -	if (pipe->num_inputs > 1)
> +	if ((pipe->bru) || (pipe->num_inputs > 1))

    Likewise.

[...]

WBR, Sergei


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH 1/2] v4l: vsp1: Reset VSP1 RPF source address
  2014-11-26  8:59   ` Geert Uytterhoeven
@ 2014-11-27  0:57     ` Takanari Hayama
  0 siblings, 0 replies; 7+ messages in thread
From: Takanari Hayama @ 2014-11-27  0:57 UTC (permalink / raw)
  To: Geert Uytterhoeven; +Cc: Linux Media Mailing List, Linux-sh list

Hi Geert,

On 11/26/14, 5:59 PM, Geert Uytterhoeven wrote:
> Hi Hayama-san,
> 
> On Wed, Nov 26, 2014 at 7:19 AM, Takanari Hayama <taki@igel.co.jp> wrote:
>> @@ -179,6 +190,10 @@ static void rpf_vdev_queue(struct vsp1_video *video,
>>                            struct vsp1_video_buffer *buf)
>>  {
>>         struct vsp1_rwpf *rpf = container_of(video, struct vsp1_rwpf, video);
>> +       int i;
>> +
>> +       for (i = 0; i < 3; i++)
>> +               rpf->buf_addr[i] = buf->addr[i];
> 
> vsp1_video_buffer.addr is "dma_addr_t addr[3];"...

Oops. Thank you for pointing that out.

> BTW, you can use memcpy() instead of an explicit loop.

I thought about it too. However, it might not be that straight forward.

VSP1 accepts only 32-bit address. If we enable LPAE, the address should
be converted and mapped to 32-bit address space via IPMMU. So, once
IPMMU is supported, we should do address mapping.

So, I guess we should leave this loop as is here, so that we can add
some address conversion in the future.

>>
>>         vsp1_rpf_write(rpf, VI6_RPF_SRCM_ADDR_Y,
>>                        buf->addr[0] + rpf->offsets[0]);
>> diff --git a/drivers/media/platform/vsp1/vsp1_rwpf.h b/drivers/media/platform/vsp1/vsp1_rwpf.h
>> index 28dd9e7..1f98fe3 100644
>> --- a/drivers/media/platform/vsp1/vsp1_rwpf.h
>> +++ b/drivers/media/platform/vsp1/vsp1_rwpf.h
>> @@ -39,6 +39,8 @@ struct vsp1_rwpf {
>>         struct v4l2_rect crop;
>>
>>         unsigned int offsets[2];
>> +
>> +       unsigned int buf_addr[3];
> 
> ... hence the above should use dma_addr_t, too.
> 
> If CONFIG_ARM_LPAE is enabled, CONFIG_ARCH_DMA_ADDR_T_64BIT
> will be enabled, too, and dma_addr_t will be u64.

Thanks. Although we cannot support LPAE for VSP1 without IPMMU, I'll
change it to dma_addr_t anyway.

> 
>>  };

Cheers,
Takanari Hayama, Ph.D. (taki@igel.co.jp)
IGEL Co.,Ltd.
http://www.igel.co.jp/

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH 2/2] v4l: vsp1: Always enable virtual RPF when BRU is in use
  2014-11-26 12:55   ` Sergei Shtylyov
@ 2014-11-27  1:01     ` Takanari Hayama
  0 siblings, 0 replies; 7+ messages in thread
From: Takanari Hayama @ 2014-11-27  1:01 UTC (permalink / raw)
  To: Sergei Shtylyov, linux-media; +Cc: linux-sh

Hi Sergei,

On 11/26/14, 9:55 PM, Sergei Shtylyov wrote:
> Hello.
> 
> On 11/26/2014 9:19 AM, Takanari Hayama wrote:
> 
>> Regardless of a number of inputs, we should always enable virtual RPF
>> when BRU is used. This allows the case when there's only one input to
>> BRU, and a size of the input is smaller than a size of an output of BRU.
> 
>> Signed-off-by: Takanari Hayama <taki@igel.co.jp>
>> ---
>>   drivers/media/platform/vsp1/vsp1_wpf.c | 11 ++++++-----
>>   1 file changed, 6 insertions(+), 5 deletions(-)
> 
>> diff --git a/drivers/media/platform/vsp1/vsp1_wpf.c
>> b/drivers/media/platform/vsp1/vsp1_wpf.c
>> index 6e05776..29ea28b 100644
>> --- a/drivers/media/platform/vsp1/vsp1_wpf.c
>> +++ b/drivers/media/platform/vsp1/vsp1_wpf.c
>> @@ -92,19 +92,20 @@ static int wpf_s_stream(struct v4l2_subdev
>> *subdev, int enable)
>>           return 0;
>>       }
>>
>> -    /* Sources. If the pipeline has a single input configure it as the
>> -     * master layer. Otherwise configure all inputs as sub-layers and
>> -     * select the virtual RPF as the master layer.
>> +    /* Sources. If the pipeline has a single input and BRU is not used,
>> +     * configure it as the master layer. Otherwise configure all
>> +     * inputs as sub-layers and select the virtual RPF as the master
>> +     * layer.
>>        */
>>       for (i = 0; i < pipe->num_inputs; ++i) {
>>           struct vsp1_rwpf *input = pipe->inputs[i];
>>
>> -        srcrpf |= pipe->num_inputs == 1
>> +        srcrpf |= ((!pipe->bru) && (pipe->num_inputs == 1))
> 
>    Inner parens not needed, especially in the first case.

Ok. Will fix it.

>>               ? VI6_WPF_SRCRPF_RPF_ACT_MST(input->entity.index)
>>               : VI6_WPF_SRCRPF_RPF_ACT_SUB(input->entity.index);
>>       }
>>
>> -    if (pipe->num_inputs > 1)
>> +    if ((pipe->bru) || (pipe->num_inputs > 1))
> 
>    Likewise.

Here too. Thanks!

> [...]
> 
> WBR, Sergei

Cheers,
Takanari Hayama, Ph.D. (taki@igel.co.jp)
IGEL Co.,Ltd.
http://www.igel.co.jp/

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2014-11-27  1:01 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-11-26  6:19 [PATCH 0/2] v4l: vsp1: crop and a single input issues Takanari Hayama
2014-11-26  6:19 ` [PATCH 1/2] v4l: vsp1: Reset VSP1 RPF source address Takanari Hayama
2014-11-26  8:59   ` Geert Uytterhoeven
2014-11-27  0:57     ` Takanari Hayama
2014-11-26  6:19 ` [PATCH 2/2] v4l: vsp1: Always enable virtual RPF when BRU is in use Takanari Hayama
2014-11-26 12:55   ` Sergei Shtylyov
2014-11-27  1:01     ` Takanari Hayama

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).