| 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
| ||
|
Message-ID: <CAGXv+5EG00P4EzNjm=7nRNwYwEF9aXorhELsmbxtrYM20SW0KQ@mail.gmail.com>
Date: Tue, 20 Jul 2021 16:51:45 +0800
From: Chen-Yu Tsai <wenst@...omium.org>
To: Chun-Jie Chen <chun-jie.chen@...iatek.com>
Cc: Enric Balletbo i Serra <enric.balletbo@...labora.com>,
Matthias Brugger <matthias.bgg@...il.com>,
Nicolas Boichat <drinkcat@...omium.org>,
Rob Herring <robh+dt@...nel.org>,
linux-arm-kernel@...ts.infradead.org,
LKML <linux-kernel@...r.kernel.org>, devicetree@...r.kernel.org,
linux-mediatek@...ts.infradead.org,
srv_heupstream <srv_heupstream@...iatek.com>,
Project_Global_Chrome_Upstream_Group
<Project_Global_Chrome_Upstream_Group@...iatek.com>
Subject: Re: [v3 3/5] soc: mediatek: pm-domains: Add support for mt8195
Hi,
On Mon, Jul 5, 2021 at 1:42 PM Chun-Jie Chen <chun-jie.chen@...iatek.com> wrote:
>
> Add domain control data including bus protection data size
> change due to more protection steps in mt8195.
>
> Signed-off-by: Chun-Jie Chen <chun-jie.chen@...iatek.com>
> ---
> drivers/soc/mediatek/mt8195-pm-domains.h | 738 +++++++++++++++++++++++
> drivers/soc/mediatek/mtk-pm-domains.c | 5 +
> drivers/soc/mediatek/mtk-pm-domains.h | 2 +-
> include/linux/soc/mediatek/infracfg.h | 103 ++++
> 4 files changed, 847 insertions(+), 1 deletion(-)
> create mode 100644 drivers/soc/mediatek/mt8195-pm-domains.h
>
> diff --git a/drivers/soc/mediatek/mt8195-pm-domains.h b/drivers/soc/mediatek/mt8195-pm-domains.h
> new file mode 100644
> index 000000000000..54bb7af8e9a3
> --- /dev/null
> +++ b/drivers/soc/mediatek/mt8195-pm-domains.h
> @@ -0,0 +1,738 @@
> +/* SPDX-License-Identifier: GPL-2.0-only */
> +/*
> + * Copyright (c) 2021 MediaTek Inc.
> + * Author: Chun-Jie Chen <chun-jie.chen@...iatek.com>
> + */
> +
> +#ifndef __SOC_MEDIATEK_MT8195_PM_DOMAINS_H
> +#define __SOC_MEDIATEK_MT8195_PM_DOMAINS_H
> +
> +#include "mtk-pm-domains.h"
> +#include <dt-bindings/power/mt8195-power.h>
> +
> +/*
> + * MT8195 power domain support
> + */
> +
> +static const struct scpsys_domain_data scpsys_domain_data_mt8195[] = {
The SCPSYS block is not documented in the datasheets available. However
I did look at all the register and bit offsets and confirmed nothing
overlapped.
> diff --git a/drivers/soc/mediatek/mtk-pm-domains.c b/drivers/soc/mediatek/mtk-pm-domains.c
> index 2689f02d7a41..12552c9996ac 100644
> --- a/drivers/soc/mediatek/mtk-pm-domains.c
> +++ b/drivers/soc/mediatek/mtk-pm-domains.c
> @@ -20,6 +20,7 @@
> #include "mt8173-pm-domains.h"
> #include "mt8183-pm-domains.h"
> #include "mt8192-pm-domains.h"
> +#include "mt8195-pm-domains.h"
>
> #define MTK_POLL_DELAY_US 10
> #define MTK_POLL_TIMEOUT USEC_PER_SEC
> @@ -576,6 +577,10 @@ static const struct of_device_id scpsys_of_match[] = {
> .compatible = "mediatek,mt8192-power-controller",
> .data = &mt8192_scpsys_data,
> },
> + {
> + .compatible = "mediatek,mt8195-power-controller",
> + .data = &mt8195_scpsys_data,
> + },
> { }
> };
>
> diff --git a/drivers/soc/mediatek/mtk-pm-domains.h b/drivers/soc/mediatek/mtk-pm-domains.h
> index 8b86ed22ca56..caaa38100093 100644
> --- a/drivers/soc/mediatek/mtk-pm-domains.h
> +++ b/drivers/soc/mediatek/mtk-pm-domains.h
> @@ -37,7 +37,7 @@
> #define PWR_STATUS_AUDIO BIT(24)
> #define PWR_STATUS_USB BIT(25)
>
> -#define SPM_MAX_BUS_PROT_DATA 5
> +#define SPM_MAX_BUS_PROT_DATA 6
>
> #define _BUS_PROT(_mask, _set, _clr, _sta, _update, _ignore) { \
> .bus_prot_mask = (_mask), \
> diff --git a/include/linux/soc/mediatek/infracfg.h b/include/linux/soc/mediatek/infracfg.h
> index 4615a228da51..3e90fb9b926a 100644
> --- a/include/linux/soc/mediatek/infracfg.h
> +++ b/include/linux/soc/mediatek/infracfg.h
> @@ -2,6 +2,109 @@
> #ifndef __SOC_MEDIATEK_INFRACFG_H
> #define __SOC_MEDIATEK_INFRACFG_H
>
> +#define MT8195_TOP_AXI_PROT_EN_STA1 0x228
> +#define MT8195_TOP_AXI_PROT_EN_1_STA1 0x258
> +#define MT8195_TOP_AXI_PROT_EN_SET 0x2a0
> +#define MT8195_TOP_AXI_PROT_EN_CLR 0x2a4
> +#define MT8195_TOP_AXI_PROT_EN_1_SET 0x2a8
> +#define MT8195_TOP_AXI_PROT_EN_1_CLR 0x2ac
> +#define MT8195_TOP_AXI_PROT_EN_MM_SET 0x2d4
> +#define MT8195_TOP_AXI_PROT_EN_MM_CLR 0x2d8
> +#define MT8195_TOP_AXI_PROT_EN_MM_STA1 0x2ec
> +#define MT8195_TOP_AXI_PROT_EN_2_SET 0x714
> +#define MT8195_TOP_AXI_PROT_EN_2_CLR 0x718
> +#define MT8195_TOP_AXI_PROT_EN_2_STA1 0x724
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_SET 0xb84
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_CLR 0xb88
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_STA1 0xb90
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_1_SET 0xba4
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_1_CLR 0xba8
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_1_STA1 0xbb0
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_2_SET 0xbb8
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_2_CLR 0xbbc
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_2_STA1 0xbc4
> +#define MT8195_TOP_AXI_PROT_EN_SUB_INFRA_VDNR_SET 0xbcc
> +#define MT8195_TOP_AXI_PROT_EN_SUB_INFRA_VDNR_CLR 0xbd0
> +#define MT8195_TOP_AXI_PROT_EN_SUB_INFRA_VDNR_STA1 0xbd8
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_SET 0xdcc
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_CLR 0xdd0
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_STA1 0xdd8
These all look correct.
> +#define MT8195_TOP_AXI_PROT_EN_NNA0 BIT(1)
> +#define MT8195_TOP_AXI_PROT_EN_NNA1 BIT(2)
> +#define MT8195_TOP_AXI_PROT_EN_NNA GENMASK(2, 1)
> +#define MT8195_TOP_AXI_PROT_EN_VDOSYS0 BIT(6)
> +#define MT8195_TOP_AXI_PROT_EN_VPPSYS0 BIT(10)
> +#define MT8195_TOP_AXI_PROT_EN_MFG1 BIT(11)
> +#define MT8195_TOP_AXI_PROT_EN_MFG1_2ND GENMASK(22, 21)
> +#define MT8195_TOP_AXI_PROT_EN_VPPSYS0_2ND BIT(23)
> +#define MT8195_TOP_AXI_PROT_EN_1_MFG1 GENMASK(20, 19)
> +#define MT8195_TOP_AXI_PROT_EN_1_CAM BIT(22)
> +#define MT8195_TOP_AXI_PROT_EN_2_CAM BIT(0)
> +#define MT8195_TOP_AXI_PROT_EN_2_MFG1_2ND GENMASK(6, 5)
> +#define MT8195_TOP_AXI_PROT_EN_2_MFG1 BIT(7)
> +#define MT8195_TOP_AXI_PROT_EN_2_AUDIO_ASRC (BIT(8) | BIT(17))
> +#define MT8195_TOP_AXI_PROT_EN_2_AUDIO (BIT(9) | BIT(11))
> +#define MT8195_TOP_AXI_PROT_EN_2_ADSP (BIT(12) | GENMASK(16, 14))
> +#define MT8195_TOP_AXI_PROT_EN_2_NNA0_2ND BIT(19)
> +#define MT8195_TOP_AXI_PROT_EN_2_NNA1_2ND BIT(20)
> +#define MT8195_TOP_AXI_PROT_EN_2_NNA_2ND GENMASK(20, 19)
> +#define MT8195_TOP_AXI_PROT_EN_2_NNA0 BIT(21)
> +#define MT8195_TOP_AXI_PROT_EN_2_NNA1 BIT(22)
> +#define MT8195_TOP_AXI_PROT_EN_2_NNA GENMASK(22, 21)
> +#define MT8195_TOP_AXI_PROT_EN_MM_CAM (BIT(0) | BIT(2) | BIT(4))
> +#define MT8195_TOP_AXI_PROT_EN_MM_IPE BIT(1)
> +#define MT8195_TOP_AXI_PROT_EN_MM_IMG (BIT(1) | BIT(3))
> +#define MT8195_TOP_AXI_PROT_EN_MM_VPPSYS0 (GENMASK(2, 0) | GENMASK(8, 6) | \
> + GENMASK(12, 10) | GENMASK(21, 19) | \
> + BIT(31))
> +#define MT8195_TOP_AXI_PROT_EN_MM_VDOSYS0 (GENMASK(5, 3) | BIT(9) | \
> + GENMASK(14, 13) | GENMASK(21, 17) | \
> + BIT(30))
> +#define MT8195_TOP_AXI_PROT_EN_MM_VPPSYS1 GENMASK(8, 5)
> +#define MT8195_TOP_AXI_PROT_EN_MM_VENC (BIT(9) | BIT(11))
> +#define MT8195_TOP_AXI_PROT_EN_MM_VENC_CORE1 (BIT(10) | BIT(12))
> +#define MT8195_TOP_AXI_PROT_EN_MM_VDEC0 BIT(13)
> +#define MT8195_TOP_AXI_PROT_EN_MM_VDEC1 BIT(14)
> +#define MT8195_TOP_AXI_PROT_EN_MM_VDOSYS1_2ND BIT(22)
> +#define MT8195_TOP_AXI_PROT_EN_MM_VPPSYS1_2ND BIT(23)
> +#define MT8195_TOP_AXI_PROT_EN_MM_CAM_2ND BIT(24)
> +#define MT8195_TOP_AXI_PROT_EN_MM_IMG_2ND BIT(25)
> +#define MT8195_TOP_AXI_PROT_EN_MM_VENC_2ND BIT(26)
> +#define MT8195_TOP_AXI_PROT_EN_MM_WPESYS BIT(27)
> +#define MT8195_TOP_AXI_PROT_EN_MM_VDEC0_2ND BIT(28)
> +#define MT8195_TOP_AXI_PROT_EN_MM_VDEC1_2ND BIT(29)
> +#define MT8195_TOP_AXI_PROT_EN_MM_VDOSYS0_2ND GENMASK(29, 22)
> +#define MT8195_TOP_AXI_PROT_EN_MM_VDOSYS1 GENMASK(31, 30)
There's significant overlap within this block. This means when the base
VDOSYS0 power domain is on, all the overlapped protection bits get turned
off. I'm not sure that's correct.
Same goes for IMG, which overlaps with IPE.
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VPPSYS0_2ND (GENMASK(7, 0) | GENMASK(18, 11))
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VENC BIT(2)
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VENC_CORE1 (BIT(3) | BIT(15))
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_CAM (BIT(5) | BIT(17))
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VPPSYS1 (GENMASK(7, 6) | BIT(18))
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VPPSYS0 (GENMASK(9, 8) | GENMASK(22, 21) | BIT(24))
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VDOSYS1 BIT(10)
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VDEC2_2ND BIT(12)
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VDEC0_2ND BIT(13)
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_WPESYS_2ND BIT(14)
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_IMG BIT(16)
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_IPE BIT(16)
And here, IMG and IPE are the same.
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VDEC2 BIT(21)
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VDEC0 BIT(22)
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_VDOSYS0 BIT(23)
> +#define MT8195_TOP_AXI_PROT_EN_MM_2_WPESYS GENMASK(24, 23)
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_1_EPD_TX BIT(1)
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_1_DP_TX BIT(2)
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_PCIE_MAC_P0 (BIT(11) | BIT(28))
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_PCIE_MAC_P1 (BIT(12) | BIT(29))
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_1_PCIE_MAC_P0 BIT(13)
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_1_PCIE_MAC_P1 BIT(14)
> +#define MT8195_TOP_AXI_PROT_EN_SUB_INFRA_VDNR_MFG1 (BIT(17) | BIT(19))
> +#define MT8195_TOP_AXI_PROT_EN_SUB_INFRA_VDNR_VPPSYS0 BIT(20)
> +#define MT8195_TOP_AXI_PROT_EN_SUB_INFRA_VDNR_VDOSYS0 BIT(21)
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_2_NNA0 BIT(25)
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_2_NNA1 BIT(26)
> +#define MT8195_TOP_AXI_PROT_EN_VDNR_2_NNA GENMASK(26, 25)
> +
All these MT8195_TOP_AXI_PROT_EN_* bit offsets aren't documented. Besides
the huge overlap above, it seems NNA also includes NNA0 and NNA1.
Regards
ChenYu
> #define MT8192_TOP_AXI_PROT_EN_STA1 0x228
> #define MT8192_TOP_AXI_PROT_EN_1_STA1 0x258
> #define MT8192_TOP_AXI_PROT_EN_SET 0x2a0
> --
> 2.18.0
> _______________________________________________
> Linux-mediatek mailing list
> Linux-mediatek@...ts.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-mediatek
Powered by blists - more mailing lists