linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Hans Verkuil <hverkuil@xs4all.nl>
To: Philipp Zabel <p.zabel@pengutronix.de>,
	Hans Verkuil <hans.verkuil@cisco.com>
Cc: Mauro Carvalho Chehab <m.chehab@samsung.com>,
	linux-media@vger.kernel.org, kernel@pengutronix.de
Subject: Re: [PATCH 4/5] [media] vivid: add support for contiguous DMA buffers
Date: Mon, 03 Nov 2014 14:38:25 +0100	[thread overview]
Message-ID: <54578551.7070002@xs4all.nl> (raw)
In-Reply-To: <1413972221-13669-5-git-send-email-p.zabel@pengutronix.de>

Hi Philipp,

I've been playing with this and I cannot make it work. One thing that is missing
in this patch is that the device struct isn't passed to v4l2_device_register.
Without that the vb2 allocation context will actually be a NULL pointer.

But after fixing that and a few other minor things (see this branch of mine:
git://linuxtv.org/hverkuil/media_tree.git vivid) it still won't work because
dma_alloc_coherent fails and that's because the device is not DMA capable.

I'm not sure how to fix this. I'd like to support dma-contig (and dma-sg in the
future), but I think someone with a better understanding of this needs to look
at this.

BTW, I'm testing this on a regular x86_64 architecture.

Regards,

	Hans

On 10/22/2014 12:03 PM, Philipp Zabel wrote:
> To simulate the behaviour of real hardware with such limitations or to
> connect vivid to real hardware with such limitations, add an option to
> let vivid use the dma-contig allocator instead of vmalloc.
> 
> Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
> ---
>  drivers/media/platform/vivid/Kconfig         |  1 +
>  drivers/media/platform/vivid/vivid-core.c    | 30 +++++++++++++++++++++++-----
>  drivers/media/platform/vivid/vivid-core.h    |  1 +
>  drivers/media/platform/vivid/vivid-vid-cap.c |  4 +++-
>  drivers/media/platform/vivid/vivid-vid-out.c |  5 ++++-
>  5 files changed, 34 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/media/platform/vivid/Kconfig b/drivers/media/platform/vivid/Kconfig
> index 3bfda25..f48c998 100644
> --- a/drivers/media/platform/vivid/Kconfig
> +++ b/drivers/media/platform/vivid/Kconfig
> @@ -4,6 +4,7 @@ config VIDEO_VIVID
>  	select FONT_SUPPORT
>  	select FONT_8x16
>  	select VIDEOBUF2_VMALLOC
> +	select VIDEOBUF2_DMA_CONTIG
>  	select FB_CFB_FILLRECT
>  	select FB_CFB_COPYAREA
>  	select FB_CFB_IMAGEBLIT
> diff --git a/drivers/media/platform/vivid/vivid-core.c b/drivers/media/platform/vivid/vivid-core.c
> index c79d60d..4c4fc3d 100644
> --- a/drivers/media/platform/vivid/vivid-core.c
> +++ b/drivers/media/platform/vivid/vivid-core.c
> @@ -29,6 +29,7 @@
>  #include <linux/platform_device.h>
>  #include <linux/videodev2.h>
>  #include <linux/v4l2-dv-timings.h>
> +#include <media/videobuf2-dma-contig.h>
>  #include <media/videobuf2-vmalloc.h>
>  #include <media/v4l2-dv-timings.h>
>  #include <media/v4l2-ioctl.h>
> @@ -107,6 +108,11 @@ MODULE_PARM_DESC(multiplanar, " 0 (default) is alternating single and multiplana
>  			      "\t\t    1 is single planar devices,\n"
>  			      "\t\t    2 is multiplanar devices");
>  
> +static unsigned allocators[VIVID_MAX_DEVS];
> +module_param_array(allocators, uint, NULL, 0444);
> +MODULE_PARM_DESC(allocators, " memory allocator selection, default is 0.\n"
> +			     "\t\t    0=vmalloc, 1=dma-contig");
> +
>  /* Default: video + vbi-cap (raw and sliced) + radio rx + radio tx + sdr + vbi-out + vid-out */
>  static unsigned node_types[VIVID_MAX_DEVS] = { [0 ... (VIVID_MAX_DEVS - 1)] = 0x1d3d };
>  module_param_array(node_types, uint, NULL, 0444);
> @@ -640,6 +646,10 @@ static int __init vivid_create_instance(int inst)
>  {
>  	static const struct v4l2_dv_timings def_dv_timings =
>  					V4L2_DV_BT_CEA_1280X720P60;
> +	static const struct vb2_mem_ops * const vivid_mem_ops[2] = {
> +		&vb2_vmalloc_memops,
> +		&vb2_dma_contig_memops,
> +	};
>  	unsigned in_type_counter[4] = { 0, 0, 0, 0 };
>  	unsigned out_type_counter[4] = { 0, 0, 0, 0 };
>  	int ccs_cap = ccs_cap_mode[inst];
> @@ -650,6 +660,7 @@ static int __init vivid_create_instance(int inst)
>  	struct video_device *vfd;
>  	struct vb2_queue *q;
>  	unsigned node_type = node_types[inst];
> +	unsigned allocator = allocators[inst];
>  	v4l2_std_id tvnorms_cap = 0, tvnorms_out = 0;
>  	int ret;
>  	int i;
> @@ -1001,6 +1012,15 @@ static int __init vivid_create_instance(int inst)
>  	dev->fb_cap.fmt.bytesperline = dev->src_rect.width * tpg_g_twopixelsize(&dev->tpg, 0) / 2;
>  	dev->fb_cap.fmt.sizeimage = dev->src_rect.height * dev->fb_cap.fmt.bytesperline;
>  
> +	/* initialize allocator context */
> +	if (allocator == 1) {
> +		dev->alloc_ctx = vb2_dma_contig_init_ctx(dev->v4l2_dev.dev);
> +		if (IS_ERR(dev->alloc_ctx))
> +			goto unreg_dev;
> +	} else if (allocator >= ARRAY_SIZE(vivid_mem_ops)) {
> +		allocator = 0;
> +	}
> +
>  	/* initialize locks */
>  	spin_lock_init(&dev->slock);
>  	mutex_init(&dev->mutex);
> @@ -1022,7 +1042,7 @@ static int __init vivid_create_instance(int inst)
>  		q->drv_priv = dev;
>  		q->buf_struct_size = sizeof(struct vivid_buffer);
>  		q->ops = &vivid_vid_cap_qops;
> -		q->mem_ops = &vb2_vmalloc_memops;
> +		q->mem_ops = vivid_mem_ops[allocator];
>  		q->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC;
>  		q->min_buffers_needed = 2;
>  
> @@ -1040,7 +1060,7 @@ static int __init vivid_create_instance(int inst)
>  		q->drv_priv = dev;
>  		q->buf_struct_size = sizeof(struct vivid_buffer);
>  		q->ops = &vivid_vid_out_qops;
> -		q->mem_ops = &vb2_vmalloc_memops;
> +		q->mem_ops = vivid_mem_ops[allocator];
>  		q->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC;
>  		q->min_buffers_needed = 2;
>  
> @@ -1058,7 +1078,7 @@ static int __init vivid_create_instance(int inst)
>  		q->drv_priv = dev;
>  		q->buf_struct_size = sizeof(struct vivid_buffer);
>  		q->ops = &vivid_vbi_cap_qops;
> -		q->mem_ops = &vb2_vmalloc_memops;
> +		q->mem_ops = vivid_mem_ops[allocator];
>  		q->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC;
>  		q->min_buffers_needed = 2;
>  
> @@ -1076,7 +1096,7 @@ static int __init vivid_create_instance(int inst)
>  		q->drv_priv = dev;
>  		q->buf_struct_size = sizeof(struct vivid_buffer);
>  		q->ops = &vivid_vbi_out_qops;
> -		q->mem_ops = &vb2_vmalloc_memops;
> +		q->mem_ops = vivid_mem_ops[allocator];
>  		q->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC;
>  		q->min_buffers_needed = 2;
>  
> @@ -1093,7 +1113,7 @@ static int __init vivid_create_instance(int inst)
>  		q->drv_priv = dev;
>  		q->buf_struct_size = sizeof(struct vivid_buffer);
>  		q->ops = &vivid_sdr_cap_qops;
> -		q->mem_ops = &vb2_vmalloc_memops;
> +		q->mem_ops = vivid_mem_ops[allocator];
>  		q->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC;
>  		q->min_buffers_needed = 8;
>  
> diff --git a/drivers/media/platform/vivid/vivid-core.h b/drivers/media/platform/vivid/vivid-core.h
> index 811c286..4fefb0d 100644
> --- a/drivers/media/platform/vivid/vivid-core.h
> +++ b/drivers/media/platform/vivid/vivid-core.h
> @@ -142,6 +142,7 @@ struct vivid_dev {
>  	struct v4l2_ctrl_handler	ctrl_hdl_radio_tx;
>  	struct video_device		sdr_cap_dev;
>  	struct v4l2_ctrl_handler	ctrl_hdl_sdr_cap;
> +	struct vb2_alloc_ctx		*alloc_ctx;
>  	spinlock_t			slock;
>  	struct mutex			mutex;
>  
> diff --git a/drivers/media/platform/vivid/vivid-vid-cap.c b/drivers/media/platform/vivid/vivid-vid-cap.c
> index 331c544..04b5fbf 100644
> --- a/drivers/media/platform/vivid/vivid-vid-cap.c
> +++ b/drivers/media/platform/vivid/vivid-vid-cap.c
> @@ -151,8 +151,10 @@ static int vid_cap_queue_setup(struct vb2_queue *vq, const struct v4l2_format *f
>  
>  	/*
>  	 * videobuf2-vmalloc allocator is context-less so no need to set
> -	 * alloc_ctxs array.
> +	 * alloc_ctxs array. videobuf2-dma-contig needs a context, though.
>  	 */
> +	for (p = 0; p < planes; p++)
> +		alloc_ctxs[p] = dev->alloc_ctx;
>  
>  	if (planes == 2)
>  		dprintk(dev, 1, "%s, count=%d, sizes=%u, %u\n", __func__,
> diff --git a/drivers/media/platform/vivid/vivid-vid-out.c b/drivers/media/platform/vivid/vivid-vid-out.c
> index 69c2dbd..6b8dfd6 100644
> --- a/drivers/media/platform/vivid/vivid-vid-out.c
> +++ b/drivers/media/platform/vivid/vivid-vid-out.c
> @@ -39,6 +39,7 @@ static int vid_out_queue_setup(struct vb2_queue *vq, const struct v4l2_format *f
>  	unsigned planes = dev->fmt_out->planes;
>  	unsigned h = dev->fmt_out_rect.height;
>  	unsigned size = dev->bytesperline_out[0] * h;
> +	unsigned p;
>  
>  	if (dev->field_out == V4L2_FIELD_ALTERNATE) {
>  		/*
> @@ -98,8 +99,10 @@ static int vid_out_queue_setup(struct vb2_queue *vq, const struct v4l2_format *f
>  
>  	/*
>  	 * videobuf2-vmalloc allocator is context-less so no need to set
> -	 * alloc_ctxs array.
> +	 * alloc_ctxs array. videobuf2-dma-contig needs a context, though.
>  	 */
> +	for (p = 0; p < planes; p++)
> +		alloc_ctxs[p] = dev->alloc_ctx;
>  
>  	if (planes == 2)
>  		dprintk(dev, 1, "%s, count=%d, sizes=%u, %u\n", __func__,
> 


  parent reply	other threads:[~2014-11-03 13:38 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-10-22 10:03 [PATCH 0/5] Contiguous DMA buffer support for the Virtual Video Test Driver Philipp Zabel
2014-10-22 10:03 ` [PATCH 1/5] [media] vivid: select CONFIG_FB_CFB_FILLRECT/COPYAREA/IMAGEBLIT Philipp Zabel
2014-10-22 10:07   ` Hans Verkuil
2014-10-22 10:03 ` [PATCH 2/5] [media] vivid: remove unused videobuf2-vmalloc headers Philipp Zabel
2014-10-22 10:08   ` Hans Verkuil
2014-10-22 10:03 ` [PATCH 3/5] [media] vivid: convert to platform device Philipp Zabel
2014-10-22 10:10   ` Hans Verkuil
2014-10-22 10:03 ` [PATCH 4/5] [media] vivid: add support for contiguous DMA buffers Philipp Zabel
2014-10-22 10:14   ` Hans Verkuil
2014-10-22 11:50     ` Philipp Zabel
2014-11-03 13:38   ` Hans Verkuil [this message]
2014-11-04 11:51     ` Philipp Zabel
2014-10-22 10:03 ` [PATCH 5/5] [media] vivid: enable VIDIOC_EXPBUF Philipp Zabel
2014-10-22 10:15   ` Hans Verkuil

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=54578551.7070002@xs4all.nl \
    --to=hverkuil@xs4all.nl \
    --cc=hans.verkuil@cisco.com \
    --cc=kernel@pengutronix.de \
    --cc=linux-media@vger.kernel.org \
    --cc=m.chehab@samsung.com \
    --cc=p.zabel@pengutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).