[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <d77e445b764cab4085390c0dd09a69f9479aec05.camel@mediatek.com>
Date: Fri, 22 Apr 2022 10:52:27 +0800
From: Irui Wang <irui.wang@...iatek.com>
To: AngeloGioacchino Del Regno
<angelogioacchino.delregno@...labora.com>,
<minghsiu.tsai@...iatek.com>
CC: <houlong.wei@...iatek.com>, <andrew-ct.chen@...iatek.com>,
<mchehab@...nel.org>, <matthias.bgg@...il.com>,
<hans.verkuil@...co.com>, <linux-media@...r.kernel.org>,
<linux-arm-kernel@...ts.infradead.org>,
<linux-mediatek@...ts.infradead.org>,
<linux-kernel@...r.kernel.org>, <kernel@...labora.com>,
<acourbot@...omium.org>
Subject: Re: [PATCH] media: platform: mtk-mdp: Fix mdp_ipi_comm structure
alignment
Dear Angelo,
On Thu, 2022-04-21 at 13:03 +0200, AngeloGioacchino Del Regno wrote:
> Il 07/03/22 16:56, AngeloGioacchino Del Regno ha scritto:
> > The mdp_ipi_comm structure defines a command that is either
> > PROCESS (start processing) or DEINIT (destroy instance); we
> > are using this one to send PROCESS or DEINIT commands from Linux
> > to an MDP instance through a VPU write but, while the first wants
> > us to stay 4-bytes aligned, the VPU instead requires an 8-bytes
> > data alignment.
> >
> > Keeping in mind that these commands are executed immediately
> > after sending them (hence not chained with others before the
> > VPU/MDP "actually" start executing), it is fine to simply add
> > a padding of 4 bytes to this structure: this keeps the same
> > performance as before, as we're still stack-allocating it,
> > while avoiding hackery inside of mtk-vpu to ensure alignment
> > bringing a definitely bigger performance impact.
> >
> > Fixes: c8eb2d7e8202 ("[media] media: Add Mediatek MDP Driver")
> > Signed-off-by: AngeloGioacchino Del Regno <
> > angelogioacchino.delregno@...labora.com>
>
> Hello,
> is there any issue with this patch?
>
> Regards,
> Angelo
>
> > ---
> >
> > This patch has been tested on Acer Chromebook R 13 (MT8173 Elm) on
> > Debian Sid.
> >
> > This is an alternative solution to the mtk-vpu approach, found
> > here:
> >
https://lore.kernel.org/all/20210920170408.1561-1-dafna.hirschfeld@collabora.com
> >
> > drivers/media/platform/mtk-mdp/mtk_mdp_ipi.h | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/drivers/media/platform/mtk-mdp/mtk_mdp_ipi.h
> > b/drivers/media/platform/mtk-mdp/mtk_mdp_ipi.h
> > index 2cb8cecb3077..b810c96695c8 100644
> > --- a/drivers/media/platform/mtk-mdp/mtk_mdp_ipi.h
> > +++ b/drivers/media/platform/mtk-mdp/mtk_mdp_ipi.h
> > @@ -40,12 +40,14 @@ struct mdp_ipi_init {
> > * @ipi_id : IPI_MDP
> > * @ap_inst : AP mtk_mdp_vpu address
> > * @vpu_inst_addr : VPU MDP instance address
> > + * @padding : Alignment padding
> > */
> > struct mdp_ipi_comm {
> > uint32_t msg_id;
> > uint32_t ipi_id;
> > uint64_t ap_inst;
> > uint32_t vpu_inst_addr;
> > + uint32_t padding;
> > };
> >
> > /**
The struct definition should be keep align between Kernel and
User(sizeof(kernel) == sizeof(md32)), for backward compatible, the new
"padding" variable can't be accessed by Kernel and md32, just for
alignment padding in kernel side.
Reviewed-by: Irui Wang <irui.wang@...iatek.com>
>
Powered by blists - more mailing lists