linux-kernel - Re: [PATCH v3 2/3] media: hantro: Support color conversion via post-processing

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20191114104600.5c6c3e26@collabora.com>
Date:   Thu, 14 Nov 2019 10:46:00 +0100
From:   Boris Brezillon <boris.brezillon@...labora.com>
To:     Ezequiel Garcia <ezequiel@...labora.com>
Cc:     linux-media@...r.kernel.org, kernel@...labora.com,
        Tomasz Figa <tfiga@...omium.org>,
        linux-rockchip@...ts.infradead.org,
        Heiko Stuebner <heiko@...ech.de>,
        Jonas Karlman <jonas@...boo.se>,
        Philipp Zabel <p.zabel@...gutronix.de>,
        Chris Healy <cphealy@...il.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 2/3] media: hantro: Support color conversion via
 post-processing

Hi Ezequiel,

On Wed, 13 Nov 2019 14:56:02 -0300
Ezequiel Garcia <ezequiel@...labora.com> wrote:

> +
> +int hantro_postproc_alloc(struct hantro_ctx *ctx)
> +{
> +	struct hantro_dev *vpu = ctx->dev;
> +	unsigned int i, buf_size;
> +
> +	buf_size = ctx->dst_fmt.plane_fmt[0].sizeimage;
> +
> +	for (i = 0; i < VB2_MAX_FRAME; ++i) {

Don't we know at that point how big the queue is (vq->num_buffers)?
Sounds a bit expensive to always allocate VB2_MAX_FRAME aux buffers.

> +		struct hantro_aux_buf *priv = &ctx->postproc.dec_q[i];
> +
> +		/*
> +		 * The buffers on this queue are meant as intermediate
> +		 * buffers for the decoder, so no mapping is needed.
> +		 */
> +		priv->attrs = DMA_ATTR_NO_KERNEL_MAPPING;
> +		priv->cpu = dma_alloc_attrs(vpu->dev, buf_size, &priv->dma,
> +					    GFP_KERNEL, priv->attrs);
> +		if (!priv->cpu)
> +			return -ENOMEM;
> +		priv->size = buf_size;
> +	}
> +	return 0;
> +}

Other than that, the post-proc extension looks pretty good. Thought it
would be much more invasive than that.

Reviewed-by: Boris Brezillon <boris.brezillon@...labora.com>