linux-kernel - Re: [PATCH v1] memory: mtk-smi: Add ostd setting for mt8186

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: <9c594eaa-9ddc-4340-ac0d-d911073764ac@collabora.com>
Date: Wed, 21 May 2025 11:26:21 +0200
From: AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>
To: Friday Yang <friday.yang@...iatek.com>, Yong Wu <yong.wu@...iatek.com>,
 Krzysztof Kozlowski <krzk@...nel.org>, Rob Herring <robh@...nel.org>,
 Conor Dooley <conor+dt@...nel.org>, Matthias Brugger
 <matthias.bgg@...il.com>, Philipp Zabel <p.zabel@...gutronix.de>
Cc: linux-mediatek@...ts.infradead.org, linux-kernel@...r.kernel.org,
 devicetree@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
 Project_Global_Chrome_Upstream_Group@...iatek.com
Subject: Re: [PATCH v1] memory: mtk-smi: Add ostd setting for mt8186

Il 21/05/25 11:16, Friday Yang ha scritto:
> Add initial ostd setting for mt8186. All the settings come from DE.
> These settings help adjust Multimedia HW's bandwidth limits to achieve
> a balanced bandwidth requirement. Without this, the VENC HW works
> abnormal while stress testing.
> 
> Fixes: 86a010bfc739 ("memory: mtk-smi: mt8186: Add smi support")
> Signed-off-by: Friday Yang <friday.yang@...iatek.com>

I agree about this commit and you can get my

Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>

...but I still have a question.

This driver is getting lots of those big OSTD arrays, and I can foresee this
getting bigger and bigger with every new SoC getting supported in there.

I'd like to understand how we can improve that, hence, can you please describe
how the OSTD values are calculated and how are they limiting the bandwidth?

I'm thinking that we can do something such that we get this runtime calculated
instead of just holding fixed values, so that we may eventually replace all those
big arrays with just a few values (foreseeing 3-4 values) and performing a big
cleanup (which may bring further improvements in the future).

Cheers,
Angelo

> ---
>   drivers/memory/mtk-smi.c | 33 +++++++++++++++++++++++++++++++++
>   1 file changed, 33 insertions(+)
> 
> diff --git a/drivers/memory/mtk-smi.c b/drivers/memory/mtk-smi.c
> index c086c22511f7..733e22f695ab 100644
> --- a/drivers/memory/mtk-smi.c
> +++ b/drivers/memory/mtk-smi.c
> @@ -320,6 +320,38 @@ static const u8 mtk_smi_larb_mt6893_ostd[][SMI_LARB_PORT_NR_MAX] = {
>   	[20] = {0x9, 0x9, 0x5, 0x5, 0x1, 0x1},
>   };
> 
> +static const u8 mtk_smi_larb_mt8186_ostd[][SMI_LARB_PORT_NR_MAX] = {
> +	[0] = {0x2, 0x1, 0x8, 0x1,},
> +	[1] = {0x1, 0x3, 0x1, 0x1,},
> +	[2] = {0x6, 0x1, 0x4, 0x1,},
> +	[3] = {},
> +	[4] = {0xf, 0x1, 0x5, 0x1, 0x1, 0x1, 0x1, 0x1, 0x1, 0x1, 0x1,
> +	       0x1, 0x1, 0x1,},
> +	[5] = {},
> +	[6] = {},
> +	[7] = {0x1, 0x3, 0x1, 0x1, 0x1, 0x3, 0x2, 0xd, 0x7, 0x5, 0x3,
> +	       0x1, 0x5,},
> +	[8] = {0x1, 0x2, 0x2,},
> +	[9] = {0x9, 0x7, 0xf, 0x8, 0x1, 0x8, 0x9, 0x3, 0x3, 0xb, 0x7, 0x4,
> +	       0x9, 0x1, 0x1, 0x1, 0x1, 0x1, 0x1, 0x1, 0x1, 0x1, 0x1, 0x1,
> +	       0x1, 0x1, 0x1, 0x1, 0x1,},
> +	[10] = {},
> +	[11] = {0x9, 0x7, 0xf, 0x8, 0x1, 0x8, 0x9, 0x3, 0x3, 0xb, 0x7, 0x4,
> +		0x9, 0x1, 0x1, 0x1, 0x1, 0x1, 0x8, 0x7, 0x7, 0x1, 0x6, 0x2,
> +		0xf, 0x8, 0x1, 0x1, 0x1,},
> +	[12] = {},
> +	[13] = {0x1, 0x1, 0x1, 0x1, 0x1, 0x1, 0x6, 0x6, 0x6, 0x1, 0x1, 0x1,},
> +	[14] = {0x1, 0x1, 0x1, 0x1, 0x1, 0x1,},
> +	[15] = {},
> +	[16] = {0x28, 0x14, 0x2, 0xc, 0x18, 0x1, 0x14, 0x1, 0x4, 0x4, 0x4,
> +		0x2, 0x4, 0x2, 0x8, 0x4, 0x4,},
> +	[17] = {0x28, 0x14, 0x2, 0xc, 0x18, 0x1, 0x14, 0x1, 0x4, 0x4, 0x4,
> +		0x2, 0x4, 0x2, 0x8, 0x4, 0x4,},
> +	[18] = {},
> +	[19] = {0x1, 0x1, 0x1, 0x1,},
> +	[20] = {0x2, 0x2, 0x2, 0x2, 0x1, 0x1,},
> +};
> +
>   static const u8 mtk_smi_larb_mt8188_ostd[][SMI_LARB_PORT_NR_MAX] = {
>   	[0] = {0x02, 0x18, 0x22, 0x22, 0x01, 0x02, 0x0a,},
>   	[1] = {0x12, 0x02, 0x14, 0x14, 0x01, 0x18, 0x0a,},
> @@ -491,6 +523,7 @@ static const struct mtk_smi_larb_gen mtk_smi_larb_mt8183 = {
>   static const struct mtk_smi_larb_gen mtk_smi_larb_mt8186 = {
>   	.config_port                = mtk_smi_larb_config_port_gen2_general,
>   	.flags_general	            = MTK_SMI_FLAG_SLEEP_CTL,
> +	.ostd			    = mtk_smi_larb_mt8186_ostd,
>   };
> 
>   static const struct mtk_smi_larb_gen mtk_smi_larb_mt8188 = {
> --
> 2.46.0
>