linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Boris Brezillon <boris.brezillon@collabora.com>
To: Mauro Carvalho Chehab <mchehab@kernel.org>,
	Hans Verkuil <hans.verkuil@cisco.com>,
	Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
	Sakari Ailus <sakari.ailus@iki.fi>,
	linux-media@vger.kernel.org
Cc: Philipp Zabel <p.zabel@pengutronix.de>,
	Ezequiel Garcia <ezequiel@collabora.com>,
	Nicolas Dufresne <nicolas.dufresne@collabora.com>,
	Tomasz Figa <tfiga@chromium.org>, Jonas Karlman <jonas@kwiboo.se>,
	Francois Buergisser <fbuergisser@google.com>
Subject: Re: [PATCH 3/3] media: hantro: h264: Fix the frame_num wraparound case
Date: Mon, 9 Sep 2019 09:31:30 +0200	[thread overview]
Message-ID: <20190909093130.64bcfa02@collabora.com> (raw)
In-Reply-To: <20190909072815.23981-3-boris.brezillon@collabora.com>

On Mon,  9 Sep 2019 09:28:15 +0200
Boris Brezillon <boris.brezillon@collabora.com> wrote:

> Step '8.2.4.1 Decoding process for picture numbers' was missing in the
> reflist creation logic, leading to invalid P reflists when a
> ->frame_num wraparound happens.  
> 
> Fixes: a9471e25629b ("media: hantro: Add core bits to support H264 decoding")
> Reported-by: Francois Buergisser <fbuergisser@google.com>
> Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
> ---

For those who would like to test this patch, here is a video [1] showing
artifacts without this change.

[1]http://crosvideo.commondatastorage.googleapis.com/1080.mp4

>  drivers/staging/media/hantro/hantro_h264.c | 23 +++++++++++++++++++++-
>  1 file changed, 22 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/staging/media/hantro/hantro_h264.c b/drivers/staging/media/hantro/hantro_h264.c
> index 881d73977ff6..e3f757a7de34 100644
> --- a/drivers/staging/media/hantro/hantro_h264.c
> +++ b/drivers/staging/media/hantro/hantro_h264.c
> @@ -271,6 +271,7 @@ struct hantro_h264_reflist_builder {
>  	const struct v4l2_h264_dpb_entry *dpb;
>  	s32 pocs[HANTRO_H264_DPB_SIZE];
>  	u8 unordered_reflist[HANTRO_H264_DPB_SIZE];
> +	int frame_nums[HANTRO_H264_DPB_SIZE];
>  	s32 curpoc;
>  	u8 num_valid;
>  };
> @@ -294,13 +295,20 @@ static void
>  init_reflist_builder(struct hantro_ctx *ctx,
>  		     struct hantro_h264_reflist_builder *b)
>  {
> +	const struct v4l2_ctrl_h264_slice_params *slice_params;
>  	const struct v4l2_ctrl_h264_decode_params *dec_param;
> +	const struct v4l2_ctrl_h264_sps *sps;
>  	struct vb2_v4l2_buffer *buf = hantro_get_dst_buf(ctx);
>  	const struct v4l2_h264_dpb_entry *dpb = ctx->h264_dec.dpb;
>  	struct vb2_queue *cap_q = &ctx->fh.m2m_ctx->cap_q_ctx.q;
> +	int cur_frame_num, max_frame_num;
>  	unsigned int i;
>  
>  	dec_param = ctx->h264_dec.ctrls.decode;
> +	slice_params = &ctx->h264_dec.ctrls.slices[0];
> +	sps = ctx->h264_dec.ctrls.sps;
> +	max_frame_num = 1 << (sps->log2_max_frame_num_minus4 + 4);
> +	cur_frame_num = slice_params->frame_num;
>  
>  	memset(b, 0, sizeof(*b));
>  	b->dpb = dpb;
> @@ -318,6 +326,18 @@ init_reflist_builder(struct hantro_ctx *ctx,
>  			continue;
>  
>  		buf = to_vb2_v4l2_buffer(vb2_get_buffer(cap_q, buf_idx));
> +
> +		/*
> +		 * Handle frame_num wraparound as described in section
> +		 * '8.2.4.1 Decoding process for picture numbers' of the spec.
> +		 * TODO: This logic will have to be adjusted when we start
> +		 * supporting interlaced content.
> +		 */
> +		if (dpb[i].frame_num > cur_frame_num)
> +			b->frame_nums[i] = (int)dpb[i].frame_num - max_frame_num;
> +		else
> +			b->frame_nums[i] = dpb[i].frame_num;
> +
>  		b->pocs[i] = get_poc(buf->field, dpb[i].top_field_order_cnt,
>  				     dpb[i].bottom_field_order_cnt);
>  		b->unordered_reflist[b->num_valid] = i;
> @@ -353,7 +373,8 @@ static int p_ref_list_cmp(const void *ptra, const void *ptrb, const void *data)
>  	 * ascending order.
>  	 */
>  	if (!(a->flags & V4L2_H264_DPB_ENTRY_FLAG_LONG_TERM))
> -		return HANTRO_CMP(b->frame_num, a->frame_num);
> +		return HANTRO_CMP(builder->frame_nums[idxb],
> +				  builder->frame_nums[idxa]);
>  
>  	return HANTRO_CMP(a->pic_num, b->pic_num);
>  }


  reply	other threads:[~2019-09-09  7:31 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-09  7:28 [PATCH 1/3] media: hantro: h264: Fix a comment in b1_ref_list_cmp() Boris Brezillon
2019-09-09  7:28 ` [PATCH 2/3] media: hantro: h264: Rename POC_CMP() into HANTRO_CMP() Boris Brezillon
2019-09-18  5:03   ` Francois Buergisser
2019-09-18  6:19     ` Tomasz Figa
2019-09-09  7:28 ` [PATCH 3/3] media: hantro: h264: Fix the frame_num wraparound case Boris Brezillon
2019-09-09  7:31   ` Boris Brezillon [this message]
2019-09-09  8:51   ` Philipp Zabel
2019-09-18  5:04   ` Francois Buergisser
2019-09-18  6:29     ` Tomasz Figa
2019-09-18  5:03 ` [PATCH 1/3] media: hantro: h264: Fix a comment in b1_ref_list_cmp() Francois Buergisser
2019-09-18  6:18   ` Tomasz Figa

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=20190909093130.64bcfa02@collabora.com \
    --to=boris.brezillon@collabora.com \
    --cc=ezequiel@collabora.com \
    --cc=fbuergisser@google.com \
    --cc=hans.verkuil@cisco.com \
    --cc=jonas@kwiboo.se \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@kernel.org \
    --cc=nicolas.dufresne@collabora.com \
    --cc=p.zabel@pengutronix.de \
    --cc=sakari.ailus@iki.fi \
    --cc=tfiga@chromium.org \
    /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).