[RESEND PATCH v4] iommu/mediatek: check 4GB mode by reading infracfg

Miles Chen miles.chen at mediatek.com
Sat Aug 29 04:51:20 EDT 2020


On Thu, 2020-08-27 at 20:27 +0100, Robin Murphy wrote:
> On 2020-08-27 06:31, Yong Wu wrote:
> > On Wed, 2020-08-26 at 16:56 +0800, Miles Chen wrote:
> >> In previous discussion [1] and [2], we found that it is risky to
> >> use max_pfn or totalram_pages to tell if 4GB mode is enabled.
> >>
> >> Check 4GB mode by reading infracfg register, remove the usage
> >> of the un-exported symbol max_pfn.
> >>
> >> This is a step towards building mtk_iommu as a kernel module.
> >>
> >> [1] https://lore.kernel.org/lkml/20200603161132.2441-1-miles.chen@mediatek.com/
> >> [2] https://lore.kernel.org/lkml/20200604080120.2628-1-miles.chen@mediatek.com/
> >> [3] https://lore.kernel.org/lkml/20200715205120.GA778876@bogus/
> >>
> >> Cc: Mike Rapoport <rppt at linux.ibm.com>
> >> Cc: David Hildenbrand <david at redhat.com>
> >> Cc: Yong Wu <yong.wu at mediatek.com>
> >> Cc: Yingjoe Chen <yingjoe.chen at mediatek.com>
> >> Cc: Christoph Hellwig <hch at lst.de>
> >> Cc: Rob Herring <robh at kernel.org>
> >> Cc: Matthias Brugger <matthias.bgg at gmail.com>
> >> Signed-off-by: Miles Chen <miles.chen at mediatek.com>
> >>
> >> ---
> >>
> >> Change since v3
> >> - use lore.kernel.org links
> >> - move "change since..." after "---"
> >>
> >> Change since v2:
> >> - determine compatible string by m4u_plat
> >> - rebase to next-20200720
> >> - add "---"
> >>
> >> Change since v1:
> >> - remove the phandle usage, search for infracfg instead [3]
> >> - use infracfg instead of infracfg_regmap
> >> - move infracfg definitaions to linux/soc/mediatek/infracfg.h
> >> - update enable_4GB only when has_4gb_mode
> >> ---
> >>   drivers/iommu/mtk_iommu.c             | 34 +++++++++++++++++++++++----
> >>   include/linux/soc/mediatek/infracfg.h |  3 +++
> >>   2 files changed, 32 insertions(+), 5 deletions(-)
> >>
> >> diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c
> >> index 785b228d39a6..adc350150492 100644
> >> --- a/drivers/iommu/mtk_iommu.c
> >> +++ b/drivers/iommu/mtk_iommu.c
> >> @@ -3,7 +3,6 @@
> >>    * Copyright (c) 2015-2016 MediaTek Inc.
> >>    * Author: Yong Wu <yong.wu at mediatek.com>
> >>    */
> >> -#include <linux/memblock.h>
> >>   #include <linux/bug.h>
> >>   #include <linux/clk.h>
> >>   #include <linux/component.h>
> >> @@ -15,13 +14,16 @@
> >>   #include <linux/iommu.h>
> >>   #include <linux/iopoll.h>
> >>   #include <linux/list.h>
> >> +#include <linux/mfd/syscon.h>
> >>   #include <linux/of_address.h>
> >>   #include <linux/of_iommu.h>
> >>   #include <linux/of_irq.h>
> >>   #include <linux/of_platform.h>
> >>   #include <linux/platform_device.h>
> >> +#include <linux/regmap.h>
> >>   #include <linux/slab.h>
> >>   #include <linux/spinlock.h>
> >> +#include <linux/soc/mediatek/infracfg.h>
> >>   #include <asm/barrier.h>
> >>   #include <soc/mediatek/smi.h>
> >>   
> >> @@ -640,8 +642,11 @@ static int mtk_iommu_probe(struct platform_device *pdev)
> >>   	struct resource         *res;
> >>   	resource_size_t		ioaddr;
> >>   	struct component_match  *match = NULL;
> >> +	struct regmap		*infracfg;
> >>   	void                    *protect;
> >>   	int                     i, larb_nr, ret;
> >> +	u32			val;
> >> +	char                    *p;
> >>   
> >>   	data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
> >>   	if (!data)
> >> @@ -655,10 +660,29 @@ static int mtk_iommu_probe(struct platform_device *pdev)
> >>   		return -ENOMEM;
> >>   	data->protect_base = ALIGN(virt_to_phys(protect), MTK_PROTECT_PA_ALIGN);
> >>   
> >> -	/* Whether the current dram is over 4GB */
> >> -	data->enable_4GB = !!(max_pfn > (BIT_ULL(32) >> PAGE_SHIFT));
> >> -	if (!MTK_IOMMU_HAS_FLAG(data->plat_data, HAS_4GB_MODE))
> >> -		data->enable_4GB = false;
> >> +	data->enable_4GB = false;
> 
> Nit: this isn't really necessary, since the structure is kzalloc()ed.

Thanks for the comment, I will remove this this in v5.
> 
> >> +	if (MTK_IOMMU_HAS_FLAG(data->plat_data, HAS_4GB_MODE)) {
> >> +		switch (data->plat_data->m4u_plat) {
> >> +		case M4U_MT2712:
> >> +			p = "mediatek,mt2712-infracfg";
> >> +			break;
> >> +		case M4U_MT8173:
> >> +			p = "mediatek,mt8173-infracfg";
> >> +			break;
> >> +		default:
> >> +			p = NULL;
> >> +		}
> >> +
> > 
> > This can be simplified:
> > 
> >          if (data->plat_data->m4u_plat == M4U_MT2712)
> > 		p = "mediatek,mt2712-infracfg";
> > 	else if(data->plat_data->m4u_plat == M4U_MT8173)
> > 		p = "mediatek,mt8173-infracfg";
> > 	else
> > 		return -EINVAL;
> 
> Right, at this point the HAS_4GB_MODE flag is entirely redundant and 
> should be removed. FWIW I still think your suggestion of putting the 
> infracfg names into plat_data would be even better and cleaner - there's 
> plenty of precedent for that sort of thing (see "git grep '\.clk_name'" 
> for example).
> 
> Robin.

hmm, I have no strong opinion about this. We have discussed this
approach before and built this patch based on the discussion.
https://lore.kernel.org/patchwork/patch/1276801/



Miles
> 
> > 
> > Then,
> > Reviewed-by: Yong Wu <yong.wu at mediatek.com>
> > 	
> > 
> >> +		infracfg = syscon_regmap_lookup_by_compatible(p);
> >> +
> >> +		if (IS_ERR(infracfg))
> >> +			return PTR_ERR(infracfg);
> >> +
> >> +		ret = regmap_read(infracfg, REG_INFRA_MISC, &val);
> >> +		if (ret)
> >> +			return ret;
> >> +		data->enable_4GB = !!(val & F_DDR_4GB_SUPPORT_EN);
> >> +	}
> >>   
> >>   	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> >>   	data->base = devm_ioremap_resource(dev, res);
> >> diff --git a/include/linux/soc/mediatek/infracfg.h b/include/linux/soc/mediatek/infracfg.h
> >> index fd25f0148566..233463d789c6 100644
> >> --- a/include/linux/soc/mediatek/infracfg.h
> >> +++ b/include/linux/soc/mediatek/infracfg.h
> >> @@ -32,6 +32,9 @@
> >>   #define MT7622_TOP_AXI_PROT_EN_WB		(BIT(2) | BIT(6) | \
> >>   						 BIT(7) | BIT(8))
> >>   
> >> +#define REG_INFRA_MISC				0xf00
> >> +#define F_DDR_4GB_SUPPORT_EN			BIT(13)
> >> +
> >>   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,
> > 
> > _______________________________________________
> > linux-arm-kernel mailing list
> > linux-arm-kernel at lists.infradead.org
> > http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
> > 



More information about the Linux-mediatek mailing list