| 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: <1501604425.17690.5.camel@mtkswgap22>
Date: Wed, 2 Aug 2017 00:20:25 +0800
From: Sean Wang <sean.wang@...iatek.com>
To: Matthias Brugger <matthias.bgg@...il.com>
CC: <robh+dt@...nel.org>, <rjw@...ysocki.net>, <khilman@...libre.com>,
<ulf.hansson@...aro.org>, <devicetree@...r.kernel.org>,
<linux-mediatek@...ts.infradead.org>, <linux-pm@...r.kernel.org>,
<linux-arm-kernel@...ts.infradead.org>,
<linux-kernel@...r.kernel.org>,
"Chen Zhong" <chen.zhong@...iatek.com>
Subject: Re: [PATCH v2 3/3] soc: mediatek: add SCPSYS power domain driver
for MediaTek MT7622 SoC
On Tue, 2017-08-01 at 16:58 +0200, Matthias Brugger wrote:
>
> On 08/01/2017 12:06 PM, sean.wang@...iatek.com wrote:
> > From: Chen Zhong <chen.zhong@...iatek.com>
> >
> > Add SCPSYS power domain driver for MT7622 SoC having four power domains
> > which are respectively ETHSYS for Ethernet including embedded switch,
> > WBSYS for WIFI and Bluetooth, HIF0SYS for PCI-E and SATA, and HIF1SYS for
> > USB. Those functions could be selectively powered gated when the
> > corresponding function is no longer to use in order to reach more minimal
> > power dissipation.
> >
> > Signed-off-by: Chen Zhong <chen.zhong@...iatek.com>
> > Signed-off-by: Sean Wang <sean.wang@...iatek.com>
> > ---
> > drivers/soc/mediatek/mtk-scpsys.c | 81 +++++++++++++++++++++++++++++++++++
> > include/linux/soc/mediatek/infracfg.h | 8 +++-
> > 2 files changed, 88 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/soc/mediatek/mtk-scpsys.c b/drivers/soc/mediatek/mtk-scpsys.c
> > index ceb2cc4..513aa73 100644
> > --- a/drivers/soc/mediatek/mtk-scpsys.c
> > +++ b/drivers/soc/mediatek/mtk-scpsys.c
> > @@ -22,6 +22,7 @@
> >
> > #include <dt-bindings/power/mt2701-power.h>
> > #include <dt-bindings/power/mt6797-power.h>
> > +#include <dt-bindings/power/mt7622-power.h>
> > #include <dt-bindings/power/mt8173-power.h>
> >
> > #define SPM_VDE_PWR_CON 0x0210
> > @@ -39,6 +40,11 @@
> > #define SPM_MFG_2D_PWR_CON 0x02c0
> > #define SPM_MFG_ASYNC_PWR_CON 0x02c4
> > #define SPM_USB_PWR_CON 0x02cc
> > +#define SPM_ETHSYS_PWR_CON 0x02e0 /* MT7622 */
> > +#define SPM_HIF0_PWR_CON 0x02e4 /* MT7622 */
> > +#define SPM_HIF1_PWR_CON 0x02e8 /* MT7622 */
> > +#define SPM_WB_PWR_CON 0x02ec /* MT7622 */
> > +
> >
> > #define SPM_PWR_STATUS 0x060c
> > #define SPM_PWR_STATUS_2ND 0x0610
> > @@ -64,6 +70,10 @@
> > #define PWR_STATUS_MFG_ASYNC BIT(23)
> > #define PWR_STATUS_AUDIO BIT(24)
> > #define PWR_STATUS_USB BIT(25)
> > +#define PWR_STATUS_ETHSYS BIT(24) /* MT7622 */
> > +#define PWR_STATUS_HIF0 BIT(25) /* MT7622 */
> > +#define PWR_STATUS_HIF1 BIT(26) /* MT7622 */
> > +#define PWR_STATUS_WB BIT(27) /* MT7622 */
> >
> > enum clk_id {
> > CLK_NONE,
> > @@ -73,6 +83,7 @@ enum clk_id {
> > CLK_VENC_LT,
> > CLK_ETHIF,
> > CLK_VDEC,
> > + CLK_HIFSEL,
> > CLK_MAX,
> > };
> >
> > @@ -84,6 +95,7 @@ static const char * const clk_names[] = {
> > "venc_lt",
> > "ethif",
> > "vdec",
> > + "hif_sel",
> > NULL,
> > };
> >
> > @@ -696,6 +708,72 @@ static int __init scpsys_probe_mt6797(struct platform_device *pdev)
> > }
> >
> > /*
> > + * MT7622 power domain support
> > + */
> > +static const struct scp_domain_data scp_domain_data_mt7622[] = {
> > + [MT7622_POWER_DOMAIN_ETHSYS] = {
> > + .name = "ethsys",
> > + .sta_mask = PWR_STATUS_ETHSYS,
> > + .ctl_offs = SPM_ETHSYS_PWR_CON,
> > + .sram_pdn_bits = GENMASK(11, 8),
> > + .sram_pdn_ack_bits = GENMASK(15, 12),
> > + .clk_id = {CLK_NONE},
> > + .bus_prot_mask = MT7622_TOP_AXI_PROT_EN_ETHSYS,
> > + .active_wakeup = true,
> > + },
> > + [MT7622_POWER_DOMAIN_HIF0] = {
> > + .name = "hif0",
> > + .sta_mask = PWR_STATUS_HIF0,
> > + .ctl_offs = SPM_HIF0_PWR_CON,
> > + .sram_pdn_bits = GENMASK(11, 8),
> > + .sram_pdn_ack_bits = GENMASK(15, 12),
> > + .clk_id = {CLK_HIFSEL},
> > + .bus_prot_mask = MT7622_TOP_AXI_PROT_EN_HIF0,
> > + .active_wakeup = true,
> > + },
> > + [MT7622_POWER_DOMAIN_HIF1] = {
> > + .name = "hif1",
> > + .sta_mask = PWR_STATUS_HIF1,
> > + .ctl_offs = SPM_HIF1_PWR_CON,
> > + .sram_pdn_bits = GENMASK(11, 8),
> > + .sram_pdn_ack_bits = GENMASK(15, 12),
> > + .clk_id = {CLK_HIFSEL},
> > + .bus_prot_mask = MT7622_TOP_AXI_PROT_EN_HIF1,
> > + .active_wakeup = true,
> > + },
> > + [MT7622_POWER_DOMAIN_WB] = {
> > + .name = "wb",
> > + .sta_mask = PWR_STATUS_WB,
> > + .ctl_offs = SPM_WB_PWR_CON,
> > + .sram_pdn_bits = 0,
> > + .sram_pdn_ack_bits = 0,
> > + .clk_id = {CLK_NONE},
> > + .bus_prot_mask = MT7622_TOP_AXI_PROT_EN_WB,
> > + .active_wakeup = true,
> > + },
> > +};
> > +
> > +#define NUM_DOMAINS_MT7622 ARRAY_SIZE(scp_domain_data_mt7622)
> > +
> > +static int __init scpsys_probe_mt7622(struct platform_device *pdev)
> > +{
> > + struct scp *scp;
> > + struct scp_ctrl_reg scp_reg;
> > +
> > + scp_reg.pwr_sta_offs = SPM_PWR_STATUS;
> > + scp_reg.pwr_sta2nd_offs = SPM_PWR_STATUS_2ND;
> > +
> > + scp = init_scp(pdev, scp_domain_data_mt7622, NUM_DOMAINS_MT7622,
> > + &scp_reg);
> > + if (IS_ERR(scp))
> > + return PTR_ERR(scp);
> > +
> > + mtk_register_power_domains(pdev, scp, NUM_DOMAINS_MT7622);
> > +
> > + return 0;
> > +}
>
> All the scpsys_probe_mtXXXX are following the same pattern, which get's really
> obvious when comparing this function to scpsys_probe_mt2701.
> Instead of adding the scpsys_probe_mtXXXX callback to the of_device_id data
> pointer, couldn't we pass a struct with a pointer to the scp_domain_data_mtXXXX
> array and a possible callback to register the subdomains?
>
> This would reduce code duplication for every new SoC, which starts to bloat the
> driver.
>
> Can you implement this as a separate patch of this series? Or do you prefer me
> to do so and you just rebase your series on top of that?
>
> Regards,
> Matthias
>
Hi, Matthias
Okay, I would add some common code as a separate patch in the next
version to reduce code duplication. This is indeed required if more
SoC support are added to the driver. Thanks for your suggestion.
Sean
> > +
> > +/*
> > * MT8173 power domain support
> > */
> >
> > @@ -835,6 +913,9 @@ static const struct of_device_id of_scpsys_match_tbl[] = {
> > .compatible = "mediatek,mt6797-scpsys",
> > .data = scpsys_probe_mt6797,
> > }, {
> > + .compatible = "mediatek,mt7622-scpsys",
> > + .data = scpsys_probe_mt7622,
> > + }, {
> > .compatible = "mediatek,mt8173-scpsys",
> > .data = scpsys_probe_mt8173,
> > }, {
> > diff --git a/include/linux/soc/mediatek/infracfg.h b/include/linux/soc/mediatek/infracfg.h
> > index a5714e9..c1e5062 100644
> > --- a/include/linux/soc/mediatek/infracfg.h
> > +++ b/include/linux/soc/mediatek/infracfg.h
> > @@ -20,7 +20,13 @@
> > #define MT8173_TOP_AXI_PROT_EN_MFG_M1 BIT(22)
> > #define MT8173_TOP_AXI_PROT_EN_MFG_SNOOP_OUT BIT(23)
> >
> > +#define MT7622_TOP_AXI_PROT_EN_ETHSYS (BIT(3) | BIT(17))
> > +#define MT7622_TOP_AXI_PROT_EN_HIF0 (BIT(24) | BIT(25))
> > +#define MT7622_TOP_AXI_PROT_EN_HIF1 (BIT(26) | BIT(27) | \
> > + BIT(28))
> > +#define MT7622_TOP_AXI_PROT_EN_WB (BIT(2) | BIT(6) | \
> > + BIT(7) | BIT(8))
> > +
> > int mtk_infracfg_set_bus_protection(struct regmap *infracfg, u32 mask);
> > int mtk_infracfg_clear_bus_protection(struct regmap *infracfg, u32 mask);
> > -
> > #endif /* __SOC_MEDIATEK_INFRACFG_H */
> >
Powered by blists - more mailing lists