| 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: <64b78b9f-1e9d-6178-ec13-7f354716928e@suse.com>
Date: Fri, 25 Sep 2020 13:01:06 +0200
From: Matthias Brugger <mbrugger@...e.com>
To: Weiyi Lu <weiyi.lu@...iatek.com>,
Enric Balletbo i Serra <enric.balletbo@...labora.com>
Cc: linux-kernel@...r.kernel.org,
Collabora Kernel ML <kernel@...labora.com>,
fparent@...libre.com, matthias.bgg@...il.com,
drinkcat@...omium.org, hsinyi@...omium.org,
linux-arm-kernel@...ts.infradead.org,
linux-mediatek@...ts.infradead.org
Subject: Re: [PATCH 06/12] soc: mediatek: pm-domains: Add SMI block as bus
protection block
On 25/09/2020 12:45, Weiyi Lu wrote:
> On Thu, 2020-09-10 at 19:28 +0200, Enric Balletbo i Serra wrote:
>> From: Matthias Brugger <mbrugger@...e.com>
>>
>> Apart from the infracfg block, the SMI block is used to enable the bus
>> protection for some power domains. Add support for this block.
>>
>> Signed-off-by: Matthias Brugger <mbrugger@...e.com>
>> Signed-off-by: Enric Balletbo i Serra <enric.balletbo@...labora.com>
>> ---
>>
>> drivers/soc/mediatek/mtk-pm-domains.c | 64 ++++++++++++++++++++-------
>> include/linux/soc/mediatek/infracfg.h | 6 +++
>> 2 files changed, 53 insertions(+), 17 deletions(-)
>>
>> diff --git a/drivers/soc/mediatek/mtk-pm-domains.c b/drivers/soc/mediatek/mtk-pm-domains.c
>> index f609c2d454fa..3aa430a60602 100644
>> --- a/drivers/soc/mediatek/mtk-pm-domains.c
>> +++ b/drivers/soc/mediatek/mtk-pm-domains.c
>> @@ -56,8 +56,25 @@
>>
>> #define SPM_MAX_BUS_PROT_DATA 3
>>
>> +#define _BUS_PROT(_mask, _set, _clr, _sta, _update) { \
>> + .bus_prot_mask = (_mask), \
>> + .bus_prot_set = _set, \
>> + .bus_prot_clr = _clr, \
>> + .bus_prot_sta = _sta, \
>> + .bus_prot_reg_update = _update, \
>> + }
>> +
>> +#define BUS_PROT_WR(_mask, _set, _clr, _sta) \
>> + _BUS_PROT(_mask, _set, _clr, _sta, false)
>> +
>> +#define BUS_PROT_UPDATE(_mask, _set, _clr, _sta) \
>> + _BUS_PROT(_mask, _set, _clr, _sta, true)
>> +
>> struct scpsys_bus_prot_data {
>> u32 bus_prot_mask;
>> + u32 bus_prot_set;
>> + u32 bus_prot_clr;
>> + u32 bus_prot_sta;
>> bool bus_prot_reg_update;
>> };
>>
>> @@ -69,6 +86,7 @@ struct scpsys_bus_prot_data {
>> * @sram_pdn_ack_bits: The mask for sram power control acked bits.
>> * @caps: The flag for active wake-up action.
>> * @bp_infracfg: bus protection for infracfg subsystem
>> + * @bp_smi: bus protection for smi subsystem
>> */
>> struct scpsys_domain_data {
>> u32 sta_mask;
>> @@ -77,6 +95,7 @@ struct scpsys_domain_data {
>> u32 sram_pdn_ack_bits;
>> u8 caps;
>> const struct scpsys_bus_prot_data bp_infracfg[SPM_MAX_BUS_PROT_DATA];
>> + const struct scpsys_bus_prot_data bp_smi[SPM_MAX_BUS_PROT_DATA];
>> };
>>
>> struct scpsys_domain {
>> @@ -86,6 +105,7 @@ struct scpsys_domain {
>> int num_clks;
>> struct clk_bulk_data *clks;
>> struct regmap *infracfg;
>> + struct regmap *smi;
>> };
>>
>> struct scpsys_soc_data {
>> @@ -173,9 +193,9 @@ static int _scpsys_bus_protect_enable(const struct scpsys_bus_prot_data *bpd, st
>> if (bpd[i].bus_prot_reg_update)
>> regmap_update_bits(regmap, INFRA_TOPAXI_PROTECTEN, mask, mask);
>> else
>> - regmap_write(regmap, INFRA_TOPAXI_PROTECTEN_SET, mask);
>> + regmap_write(regmap, bpd[i].bus_prot_set, mask);
>>
>
> Could it be?
>
> if (bpd[i].bus_prot_set)
> regmap_write(regmap, bpd[i].bus_prot_set, mask);
> else
> regmap_update_bits(regmap, INFRA_TOPAXI_PROTECTEN, mask, mask);
>
Nice catch. That's a bug here:
regmap_update_bits(regmap, INFRA_TOPAXI_PROTECTEN, mask, mask)
should be
regmap_update_bits(regmap, bpd[i].bus_prot_set, mask, mask)
Regarding your question. No we can't do that as we use bus_prot_set variable for
write and update and decide on bus_prot_reg_update which method to use.
>> - ret = regmap_read_poll_timeout(regmap, INFRA_TOPAXI_PROTECTSTA1,
>> + ret = regmap_read_poll_timeout(regmap, bpd[i].bus_prot_sta,
>> val, (val & mask) == mask,
>> MTK_POLL_DELAY_US, MTK_POLL_TIMEOUT);
>> if (ret)
>> @@ -191,7 +211,11 @@ static int scpsys_bus_protect_enable(struct scpsys_domain *pd)
>> int ret;
>>
>> ret = _scpsys_bus_protect_enable(bpd, pd->infracfg);
>> - return ret;
>> + if (ret)
>> + return ret;
>> +
>> + bpd = pd->data->bp_smi;
>> + return _scpsys_bus_protect_enable(bpd, pd->smi);
>> }
>>
>> static int _scpsys_bus_protect_disable(const struct scpsys_bus_prot_data *bpd,
>> @@ -206,11 +230,11 @@ static int _scpsys_bus_protect_disable(const struct scpsys_bus_prot_data *bpd,
>> return 0;
>>
>> if (bpd[i].bus_prot_reg_update)
>
> ditto.
>
>> - regmap_update_bits(regmap, INFRA_TOPAXI_PROTECTEN, mask, 0);
>> + regmap_update_bits(regmap, bpd[i].bus_prot_set, mask, 0);
As you can see here we use bpd[i].bus_prot_set
You could argue that we could check the bus_prot_clr to decide if we need an
update or not, but I prefer an explicit variable to flag this. Implicite
reasoning depending on values for bus_prot_clr will make the code harder to read
and won't save a lot of memory.
>> else
>> - regmap_write(regmap, INFRA_TOPAXI_PROTECTEN_CLR, mask);
>> + regmap_write(regmap, bpd[i].bus_prot_clr, mask);
>>
>> - ret = regmap_read_poll_timeout(regmap, INFRA_TOPAXI_PROTECTSTA1,
>> + ret = regmap_read_poll_timeout(regmap, bpd[i].bus_prot_sta,
>> val, !(val & mask),
>> MTK_POLL_DELAY_US, MTK_POLL_TIMEOUT);
>> if (ret)
>> @@ -226,7 +250,11 @@ static int scpsys_bus_protect_disable(struct scpsys_domain *pd)
>> int ret;
>>
>> ret = _scpsys_bus_protect_disable(bpd, pd->infracfg);
>> - return ret;
>> + if (ret)
>> + return ret;
>> +
>> + bpd = pd->data->bp_smi;
>> + return _scpsys_bus_protect_disable(bpd, pd->smi);
>
> It should have a reverse order compared to the enable control. But I'd
> like to make it more flexible to any sequence, like INFRA->SMI->INFRA
Correct, we whould disable first SMI and then infracfg. Good catch!
Right now we don't have any hardware supported that does
infracfg->SMI->infracfg. When that comes up we will need a good solution for the
driver :)
Thanks for your review!
Matthias
>
>> }
>>
>> static int scpsys_power_on(struct generic_pm_domain *genpd)
>> @@ -360,6 +388,10 @@ static int scpsys_add_one_domain(struct scpsys *scpsys, struct device_node *node
>> if (IS_ERR(pd->infracfg))
>> pd->infracfg = NULL;
>>
>> + pd->smi = syscon_regmap_lookup_by_phandle(node, "mediatek,smi");
>> + if (IS_ERR(pd->smi))
>> + pd->smi = NULL;
>> +
>> pd->num_clks = of_clk_get_parent_count(node);
>> if (pd->num_clks > 0) {
>> pd->clks = devm_kcalloc(scpsys->dev, pd->num_clks, sizeof(*pd->clks), GFP_KERNEL);
>> @@ -532,10 +564,9 @@ static const struct scpsys_domain_data scpsys_domain_data_mt8173[] = {
>> .ctl_offs = SPM_DIS_PWR_CON,
>> .sram_pdn_bits = GENMASK(11, 8),
>> .sram_pdn_ack_bits = GENMASK(12, 12),
>> - .bp_infracfg[0] = {
>> - .bus_prot_reg_update = true,
>> - .bus_prot_mask = MT8173_TOP_AXI_PROT_EN_MM_M0 |
>> - MT8173_TOP_AXI_PROT_EN_MM_M1,
>> + .bp_infracfg = {
>> + BUS_PROT_UPDATE_MT8173(MT8173_TOP_AXI_PROT_EN_MM_M0 |
>> + MT8173_TOP_AXI_PROT_EN_MM_M1),
>> },
>> },
>> [MT8173_POWER_DOMAIN_VENC_LT] = {
>> @@ -574,12 +605,11 @@ static const struct scpsys_domain_data scpsys_domain_data_mt8173[] = {
>> .ctl_offs = SPM_MFG_PWR_CON,
>> .sram_pdn_bits = GENMASK(13, 8),
>> .sram_pdn_ack_bits = GENMASK(21, 16),
>> - .bp_infracfg[0] = {
>> - .bus_prot_reg_update = true,
>> - .bus_prot_mask = MT8173_TOP_AXI_PROT_EN_MFG_S |
>> - MT8173_TOP_AXI_PROT_EN_MFG_M0 |
>> - MT8173_TOP_AXI_PROT_EN_MFG_M1 |
>> - MT8173_TOP_AXI_PROT_EN_MFG_SNOOP_OUT,
>> + .bp_infracfg = {
>> + BUS_PROT_UPDATE_MT8173(MT8173_TOP_AXI_PROT_EN_MFG_S |
>> + MT8173_TOP_AXI_PROT_EN_MFG_M0 |
>> + MT8173_TOP_AXI_PROT_EN_MFG_M1 |
>> + MT8173_TOP_AXI_PROT_EN_MFG_SNOOP_OUT),
>> },
>> },
>> };
>> diff --git a/include/linux/soc/mediatek/infracfg.h b/include/linux/soc/mediatek/infracfg.h
>> index f967d02cc2ff..3f18cddffb44 100644
>> --- a/include/linux/soc/mediatek/infracfg.h
>> +++ b/include/linux/soc/mediatek/infracfg.h
>> @@ -37,6 +37,12 @@
>> #define INFRA_TOPAXI_PROTECTEN_SET 0x0260
>> #define INFRA_TOPAXI_PROTECTEN_CLR 0x0264
>>
>> +#define BUS_PROT_UPDATE_MT8173(_mask) \
>> + BUS_PROT_UPDATE(_mask, \
>> + INFRA_TOPAXI_PROTECTEN, \
>> + INFRA_TOPAXI_PROTECTEN_CLR, \
>> + INFRA_TOPAXI_PROTECTSTA1)
>> +
>> int mtk_infracfg_set_bus_protection(struct regmap *infracfg, u32 mask,
>> bool reg_update);
>> int mtk_infracfg_clear_bus_protection(struct regmap *infracfg, u32 mask,
>
Powered by blists - more mailing lists