Skip to contentSkip to conversation

iommu/arm-smmu: Enable split pagetable support

PATCH Superseded Jordan Crouse 3 participants 13 messages Jun 11, 2020 – Jun 17, 2020

Subsystems (8)

v8 of 19 Superseded Latest is v19 →

lore
Share & tools
lore
#0 lore raw

v8 cover iommu/arm-smmu: Enable split pagetable support

Another iteration of the split-pagetable support for arm-smmu and the Adreno GPU SMMU. After email discussions [1] we opted to make a arm-smmu implementation for specifically for the Adreno GPU and use that to enable split pagetable support and later other implementation specific bits that we need.

On the hardware side this is very close to the same code from before [2] only the TTBR1 quirk is turned on by the implementation and not a domain attribute. In drm/msm we use the returned size of the aperture as a clue to let us know which virtual address space we should use for global memory objects.

There are two open items that you should be aware of. First, in the implementation specific code we have to check the compatible string of the device so that we only enable TTBR1 for the GPU (SID 0) and not the GMU (SID 4). I went back and forth trying to decide if I wanted to use the compatbile string or the SID as the filter and settled on the compatible string but I could be talked out of it.

The other open item is that in drm/msm the hardware only uses 49 bits of the address space but arm-smmu expects the address to be sign extended all the way to 64 bits. This isn't a problem normally unless you look at the hardware registers that contain a IOVA and then the upper bits will be zero. I opted to restrict the internal drm/msm IOVA range to only 49 bits and then sign extend right before calling iommu_map / iommu_unmap. This is a bit wonky but I thought that matching the hardware would be less confusing when debugging a hang.

v8: Pass the attached device in the smmu_domain to the implementation specific functions

[1] https://lists.linuxfoundation.org/pipermail/iommu/2020-May/044537.html
[2] https://patchwork.kernel.org/patch/11482591/

Jordan Crouse (7):
iommu/arm-smmu: Pass io-pgtable config to implementation specific
function
iommu/arm-smmu: Add support for split pagetables
dt-bindings: arm-smmu: Add compatible string for Adreno GPU SMMU iommu/arm-smmu: Add a pointer to the attached device to smmu_domain iommu/arm-smmu: Add implementation for the adreno GPU SMMU drm/msm: Set the global virtual address range from the IOMMU domain
arm: dts: qcom: sm845: Set the compatible string for the GPU SMMU

 .../devicetree/bindings/iommu/arm,smmu.yaml   |  4 ++
 arch/arm64/boot/dts/qcom/sdm845.dtsi          |  2 +-
 drivers/gpu/drm/msm/adreno/adreno_gpu.c       | 13 +++++-
 drivers/gpu/drm/msm/msm_iommu.c               |  7 +++
 drivers/iommu/arm-smmu-impl.c                 |  6 ++-
 drivers/iommu/arm-smmu-qcom.c                 | 45 ++++++++++++++++++-
 drivers/iommu/arm-smmu.c                      | 33 +++++++++-----
 drivers/iommu/arm-smmu.h                      | 30 ++++++++++---
 8 files changed, 117 insertions(+), 23 deletions(-)
Signature · 6 lines
2.17.1

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
#1 lore raw

PATCH v8 1/7 iommu/arm-smmu: Pass io-pgtable config to implementation specific function

Construct the io-pgtable config before calling the implementation specific init_context function and pass it so the implementation specific function can get a chance to change it before the io-pgtable is created.

---

 drivers/iommu/arm-smmu-impl.c |  3 ++-
 drivers/iommu/arm-smmu.c      | 11 ++++++-----
 drivers/iommu/arm-smmu.h      |  3 ++-
 3 files changed, 10 insertions(+), 7 deletions(-)
drivers/iommu/arm-smmu-impl.c +2 -1
index c75b9d957b70..a20e426d81ac 100644
@@ -68,7 +68,8 @@ static int cavium_cfg_probe(struct arm_smmu_device *smmu)
 	return 0;
 }
 
-static int cavium_init_context(struct arm_smmu_domain *smmu_domain)
+static int cavium_init_context(struct arm_smmu_domain *smmu_domain,
+		struct io_pgtable_cfg *pgtbl_cfg)
 {
 	struct cavium_smmu *cs = container_of(smmu_domain->smmu,
 					      struct cavium_smmu, smmu);
drivers/iommu/arm-smmu.c +6 -5
index 243bc4cb2705..8a3a6c8c887a 100644
@@ -797,11 +797,6 @@ static int arm_smmu_init_domain_context(struct iommu_domain *domain,
 		cfg->asid = cfg->cbndx;
 
 	smmu_domain->smmu = smmu;
-	if (smmu->impl && smmu->impl->init_context) {
-		ret = smmu->impl->init_context(smmu_domain);
-		if (ret)
-			goto out_unlock;
-	}
 
 	pgtbl_cfg = (struct io_pgtable_cfg) {
 		.pgsize_bitmap	= smmu->pgsize_bitmap,
@@ -812,6 +807,12 @@ static int arm_smmu_init_domain_context(struct iommu_domain *domain,
 		.iommu_dev	= smmu->dev,
 	};
 
+	if (smmu->impl && smmu->impl->init_context) {
+		ret = smmu->impl->init_context(smmu_domain, &pgtbl_cfg);
+		if (ret)
+			goto out_unlock;
+	}
+
 	if (smmu_domain->non_strict)
 		pgtbl_cfg.quirks |= IO_PGTABLE_QUIRK_NON_STRICT;
 
drivers/iommu/arm-smmu.h +2 -1
index d172c024be61..38b041530a4f 100644
@@ -383,7 +383,8 @@ struct arm_smmu_impl {
 			    u64 val);
 	int (*cfg_probe)(struct arm_smmu_device *smmu);
 	int (*reset)(struct arm_smmu_device *smmu);
-	int (*init_context)(struct arm_smmu_domain *smmu_domain);
+	int (*init_context)(struct arm_smmu_domain *smmu_domain,
+			struct io_pgtable_cfg *cfg);
 	void (*tlb_sync)(struct arm_smmu_device *smmu, int page, int sync,
 			 int status);
 	int (*def_domain_type)(struct device *dev);
Signature · 6 lines
2.17.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
#2 lore raw

PATCH v8 2/7 iommu/arm-smmu: Add support for split pagetables

In reply to Jordan Crouse

Enable TTBR1 for a context bank if IO_PGTABLE_QUIRK_ARM_TTBR1 is selected by the io-pgtable configuration.

---

 drivers/iommu/arm-smmu.c | 21 ++++++++++++++++-----
 drivers/iommu/arm-smmu.h | 25 +++++++++++++++++++------
 2 files changed, 35 insertions(+), 11 deletions(-)
drivers/iommu/arm-smmu.c +16 -5
index 8a3a6c8c887a..048de2681670 100644
@@ -555,11 +555,15 @@ static void arm_smmu_init_context_bank(struct arm_smmu_domain *smmu_domain,
 			cb->ttbr[0] = pgtbl_cfg->arm_v7s_cfg.ttbr;
 			cb->ttbr[1] = 0;
 		} else {
-			cb->ttbr[0] = pgtbl_cfg->arm_lpae_s1_cfg.ttbr;
-			cb->ttbr[0] |= FIELD_PREP(ARM_SMMU_TTBRn_ASID,
-						  cfg->asid);
+			cb->ttbr[0] = FIELD_PREP(ARM_SMMU_TTBRn_ASID,
+				cfg->asid);
 			cb->ttbr[1] = FIELD_PREP(ARM_SMMU_TTBRn_ASID,
-						 cfg->asid);
+				cfg->asid);
+
+			if (pgtbl_cfg->quirks & IO_PGTABLE_QUIRK_ARM_TTBR1)
+				cb->ttbr[1] |= pgtbl_cfg->arm_lpae_s1_cfg.ttbr;
+			else
+				cb->ttbr[0] |= pgtbl_cfg->arm_lpae_s1_cfg.ttbr;
 		}
 	} else {
 		cb->ttbr[0] = pgtbl_cfg->arm_lpae_s2_cfg.vttbr;
@@ -824,7 +828,14 @@ static int arm_smmu_init_domain_context(struct iommu_domain *domain,
 
 	/* Update the domain's page sizes to reflect the page table format */
 	domain->pgsize_bitmap = pgtbl_cfg.pgsize_bitmap;
-	domain->geometry.aperture_end = (1UL << ias) - 1;
+
+	if (pgtbl_cfg.quirks & IO_PGTABLE_QUIRK_ARM_TTBR1) {
+		domain->geometry.aperture_start = ~0UL << ias;
+		domain->geometry.aperture_end = ~0UL;
+	} else {
+		domain->geometry.aperture_end = (1UL << ias) - 1;
+	}
+
 	domain->geometry.force_aperture = true;
 
 	/* Initialise the context bank with our page table cfg */
drivers/iommu/arm-smmu.h +19 -6
index 38b041530a4f..5f2de20e883b 100644
@@ -168,10 +168,12 @@ enum arm_smmu_cbar_type {
 #define ARM_SMMU_CB_TCR			0x30
 #define ARM_SMMU_TCR_EAE		BIT(31)
 #define ARM_SMMU_TCR_EPD1		BIT(23)
+#define ARM_SMMU_TCR_A1			BIT(22)
 #define ARM_SMMU_TCR_TG0		GENMASK(15, 14)
 #define ARM_SMMU_TCR_SH0		GENMASK(13, 12)
 #define ARM_SMMU_TCR_ORGN0		GENMASK(11, 10)
 #define ARM_SMMU_TCR_IRGN0		GENMASK(9, 8)
+#define ARM_SMMU_TCR_EPD0		BIT(7)
 #define ARM_SMMU_TCR_T0SZ		GENMASK(5, 0)
 
 #define ARM_SMMU_VTCR_RES1		BIT(31)
@@ -347,12 +349,23 @@ struct arm_smmu_domain {
 
 static inline u32 arm_smmu_lpae_tcr(struct io_pgtable_cfg *cfg)
 {
-	return ARM_SMMU_TCR_EPD1 |
-	       FIELD_PREP(ARM_SMMU_TCR_TG0, cfg->arm_lpae_s1_cfg.tcr.tg) |
-	       FIELD_PREP(ARM_SMMU_TCR_SH0, cfg->arm_lpae_s1_cfg.tcr.sh) |
-	       FIELD_PREP(ARM_SMMU_TCR_ORGN0, cfg->arm_lpae_s1_cfg.tcr.orgn) |
-	       FIELD_PREP(ARM_SMMU_TCR_IRGN0, cfg->arm_lpae_s1_cfg.tcr.irgn) |
-	       FIELD_PREP(ARM_SMMU_TCR_T0SZ, cfg->arm_lpae_s1_cfg.tcr.tsz);
+	u32 tcr = FIELD_PREP(ARM_SMMU_TCR_TG0, cfg->arm_lpae_s1_cfg.tcr.tg) |
+		FIELD_PREP(ARM_SMMU_TCR_SH0, cfg->arm_lpae_s1_cfg.tcr.sh) |
+		FIELD_PREP(ARM_SMMU_TCR_ORGN0, cfg->arm_lpae_s1_cfg.tcr.orgn) |
+		FIELD_PREP(ARM_SMMU_TCR_IRGN0, cfg->arm_lpae_s1_cfg.tcr.irgn) |
+		FIELD_PREP(ARM_SMMU_TCR_T0SZ, cfg->arm_lpae_s1_cfg.tcr.tsz);
+
+       /*
+	* When TTBR1 is selected shift the TCR fields by 16 bits and disable
+	* translation in TTBR0
+	*/
+	if (cfg->quirks & IO_PGTABLE_QUIRK_ARM_TTBR1) {
+		tcr = (tcr << 16) & ~ARM_SMMU_TCR_A1;
+		tcr |= ARM_SMMU_TCR_EPD0;
+	} else
+		tcr |= ARM_SMMU_TCR_EPD1;
+
+	return tcr;
 }
 
 static inline u32 arm_smmu_lpae_tcr2(struct io_pgtable_cfg *cfg)
Signature · 6 lines
2.17.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
#3 lore raw

PATCH v8 3/7 dt-bindings: arm-smmu: Add compatible string for Adreno GPU SMMU

In reply to Jordan Crouse

Every Qcom Adreno GPU has an embedded SMMU for its own use. These devices depend on unique features such as split pagetables, different stall/halt requirements and other settings. Identify them with a compatible string so that they can be identified in the arm-smmu implementation specific code.

---

 Documentation/devicetree/bindings/iommu/arm,smmu.yaml | 4 ++++
 1 file changed, 4 insertions(+)
Documentation/devicetree/bindings/iommu/arm,smmu.yaml +4 -0
index d7ceb4c34423..e52a1b146c97 100644
@@ -38,6 +38,10 @@ properties:
               - qcom,sc7180-smmu-500
               - qcom,sdm845-smmu-500
           - const: arm,mmu-500
+      - description: Qcom Adreno GPUs implementing "arm,smmu-v2"
+        items:
+          - const: qcom,adreno-smmu
+          - const: qcom,smmu-v2
       - items:
           - const: arm,mmu-500
           - const: arm,smmu-v2
Signature · 1 line
2.17.1
maintainer Reviewed-by on v8 3/7
#4 lore raw

Re: [PATCH v8 3/7] dt-bindings: arm-smmu: Add compatible string for Adreno GPU SMMU

Quoted 6 lines · Jordan Crouse

Every Qcom Adreno GPU has an embedded SMMU for its own use. These devices depend on unique features such as split pagetables, different stall/halt requirements and other settings. Identify them with a compatible string so that they can be identified in the arm-smmu implementation specific code.

Quoted 6 lines

---

 Documentation/devicetree/bindings/iommu/arm,smmu.yaml | 4 ++++
 1 file changed, 4 insertions(+)
#5 lore raw

PATCH v8 4/7 iommu/arm-smmu: Add a pointer to the attached device to smmu_domain

In reply to Jordan Crouse

Add a link to the pointer to the struct device that is attached to a domain. This makes it easy to get the pointer if it is needed in the implementation specific code.

---

 drivers/iommu/arm-smmu.c | 1 +
 drivers/iommu/arm-smmu.h | 1 +
 2 files changed, 2 insertions(+)
drivers/iommu/arm-smmu.c +1 -0
index 048de2681670..743d75b9ff3f 100644
@@ -801,6 +801,7 @@ static int arm_smmu_init_domain_context(struct iommu_domain *domain,
 		cfg->asid = cfg->cbndx;
 
 	smmu_domain->smmu = smmu;
+	smmu_domain->dev = dev;
 
 	pgtbl_cfg = (struct io_pgtable_cfg) {
 		.pgsize_bitmap	= smmu->pgsize_bitmap,
drivers/iommu/arm-smmu.h +1 -0
index 5f2de20e883b..d33cfe26b2f5 100644
@@ -345,6 +345,7 @@ struct arm_smmu_domain {
 	struct mutex			init_mutex; /* Protects smmu pointer */
 	spinlock_t			cb_lock; /* Serialises ATS1* ops and TLB syncs */
 	struct iommu_domain		domain;
+	struct device			*dev;	/* Device attached to this domain */
 };
 
 static inline u32 arm_smmu_lpae_tcr(struct io_pgtable_cfg *cfg)
Signature · 6 lines
2.17.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
#6 lore raw

Re: [PATCH v8 4/7] iommu/arm-smmu: Add a pointer to the attached device to smmu_domain

Thank you for the patch! Yet something to improve:

Hi Jordan,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on v5.7 next-20200611]
[cannot apply to iommu/next robh/for-next arm/for-next keystone/next rockchip/for-next arm64/for-next/core shawnguo/for-next soc/for-next] [if your patch is applied to the wrong git tree, please drop us a note to help
improve the system. BTW, we also suggest to use '--base' option to specify the
base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url:    https://github.com/0day-ci/linux/commits/Jordan-Crouse/iommu-arm-smmu-Enable-split-pagetable-support/20200612-094718
base:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git b961f8dc8976c091180839f4483d67b7c2ca2578
config: arm64-randconfig-s031-20200612 (attached as .config)
compiler: aarch64-linux-gcc (GCC) 9.3.0
reproduce:
        # apt-get install sparse
        # sparse version: v0.6.1-250-g42323db3-dirty
        # save the attached .config to linux build tree
        make W=1 C=1 ARCH=arm64 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__'

If you fix the issue, kindly add following tag as appropriate

All errors (new ones prefixed by >>, old ones prefixed by <<):

drivers/iommu/arm-smmu.c: In function 'arm_smmu_init_domain_context':

>> drivers/iommu/arm-smmu.c:804:21: error: 'dev' undeclared (first use in this function); did you mean 'cdev'?
804 |  smmu_domain->dev = dev;
|                     ^~~
|                     cdev

drivers/iommu/arm-smmu.c:804:21: note: each undeclared identifier is reported only once for each function it appears in

vim +804 drivers/iommu/arm-smmu.c

669
670 static int arm_smmu_init_domain_context(struct iommu_domain *domain,
671 struct arm_smmu_device *smmu)
672 {
673 int irq, start, ret = 0;
674 unsigned long ias, oas;
675 struct io_pgtable_ops *pgtbl_ops;
676 struct io_pgtable_cfg pgtbl_cfg;
677 enum io_pgtable_fmt fmt;
678 struct arm_smmu_domain *smmu_domain = to_smmu_domain(domain);
679 struct arm_smmu_cfg *cfg = &smmu_domain->cfg;
680
681 mutex_lock(&smmu_domain->init_mutex);
682 if (smmu_domain->smmu)
683 goto out_unlock;
684
685 if (domain->type == IOMMU_DOMAIN_IDENTITY) {
686 smmu_domain->stage = ARM_SMMU_DOMAIN_BYPASS;
687 smmu_domain->smmu = smmu;
688 goto out_unlock;
689 }
690
691 /*
692 * Mapping the requested stage onto what we support is surprisingly
693 * complicated, mainly because the spec allows S1+S2 SMMUs without
694 * support for nested translation. That means we end up with the
695 * following table:
696 *
697 * Requested Supported Actual

   698		 *     S1               N              S1
   699		 *     S1             S1+S2            S1
   700		 *     S1               S2             S2
   701		 *     S1               S1             S1
   702		 *     N                N              N
   703		 *     N              S1+S2            S2
   704		 *     N                S2             S2
   705		 *     N                S1             S1
   706		 *
   707		 * Note that you can't actually request stage-2 mappings.
   708		 */
   709		if (!(smmu->features & ARM_SMMU_FEAT_TRANS_S1))
   710			smmu_domain->stage = ARM_SMMU_DOMAIN_S2;
   711		if (!(smmu->features & ARM_SMMU_FEAT_TRANS_S2))
   712			smmu_domain->stage = ARM_SMMU_DOMAIN_S1;
   713	
   714		/*
   715		 * Choosing a suitable context format is even more fiddly. Until we
   716		 * grow some way for the caller to express a preference, and/or move
   717		 * the decision into the io-pgtable code where it arguably belongs,
   718		 * just aim for the closest thing to the rest of the system, and hope
   719		 * that the hardware isn't esoteric enough that we can't assume AArch64
   720		 * support to be a superset of AArch32 support...
   721		 */
   722		if (smmu->features & ARM_SMMU_FEAT_FMT_AARCH32_L)
   723			cfg->fmt = ARM_SMMU_CTX_FMT_AARCH32_L;
   724		if (IS_ENABLED(CONFIG_IOMMU_IO_PGTABLE_ARMV7S) &&
   725		    !IS_ENABLED(CONFIG_64BIT) && !IS_ENABLED(CONFIG_ARM_LPAE) &&
   726		    (smmu->features & ARM_SMMU_FEAT_FMT_AARCH32_S) &&
   727		    (smmu_domain->stage == ARM_SMMU_DOMAIN_S1))
   728			cfg->fmt = ARM_SMMU_CTX_FMT_AARCH32_S;
   729		if ((IS_ENABLED(CONFIG_64BIT) || cfg->fmt == ARM_SMMU_CTX_FMT_NONE) &&
   730		    (smmu->features & (ARM_SMMU_FEAT_FMT_AARCH64_64K |
   731				       ARM_SMMU_FEAT_FMT_AARCH64_16K |
   732				       ARM_SMMU_FEAT_FMT_AARCH64_4K)))
   733			cfg->fmt = ARM_SMMU_CTX_FMT_AARCH64;
   734	
   735		if (cfg->fmt == ARM_SMMU_CTX_FMT_NONE) {
   736			ret = -EINVAL;
   737			goto out_unlock;
   738		}
   739	
   740		switch (smmu_domain->stage) {
   741		case ARM_SMMU_DOMAIN_S1:
   742			cfg->cbar = CBAR_TYPE_S1_TRANS_S2_BYPASS;
   743			start = smmu->num_s2_context_banks;
   744			ias = smmu->va_size;
   745			oas = smmu->ipa_size;
   746			if (cfg->fmt == ARM_SMMU_CTX_FMT_AARCH64) {
   747				fmt = ARM_64_LPAE_S1;
   748			} else if (cfg->fmt == ARM_SMMU_CTX_FMT_AARCH32_L) {
   749				fmt = ARM_32_LPAE_S1;
   750				ias = min(ias, 32UL);
   751				oas = min(oas, 40UL);
   752			} else {
   753				fmt = ARM_V7S;
   754				ias = min(ias, 32UL);
   755				oas = min(oas, 32UL);
   756			}
   757			smmu_domain->flush_ops = &arm_smmu_s1_tlb_ops;
   758			break;
   759		case ARM_SMMU_DOMAIN_NESTED:
   760			/*
   761			 * We will likely want to change this if/when KVM gets
   762			 * involved.
   763			 */
   764		case ARM_SMMU_DOMAIN_S2:
   765			cfg->cbar = CBAR_TYPE_S2_TRANS;
   766			start = 0;
   767			ias = smmu->ipa_size;
   768			oas = smmu->pa_size;
   769			if (cfg->fmt == ARM_SMMU_CTX_FMT_AARCH64) {
   770				fmt = ARM_64_LPAE_S2;
   771			} else {
   772				fmt = ARM_32_LPAE_S2;
   773				ias = min(ias, 40UL);
   774				oas = min(oas, 40UL);
   775			}
   776			if (smmu->version == ARM_SMMU_V2)
   777				smmu_domain->flush_ops = &arm_smmu_s2_tlb_ops_v2;
   778			else
   779				smmu_domain->flush_ops = &arm_smmu_s2_tlb_ops_v1;
   780			break;
   781		default:
   782			ret = -EINVAL;
   783			goto out_unlock;
   784		}
   785		ret = __arm_smmu_alloc_bitmap(smmu->context_map, start,
   786					      smmu->num_context_banks);
   787		if (ret < 0)
   788			goto out_unlock;
   789	
   790		cfg->cbndx = ret;
   791		if (smmu->version < ARM_SMMU_V2) {
   792			cfg->irptndx = atomic_inc_return(&smmu->irptndx);
   793			cfg->irptndx %= smmu->num_context_irqs;
   794		} else {
   795			cfg->irptndx = cfg->cbndx;
   796		}
   797	
   798		if (smmu_domain->stage == ARM_SMMU_DOMAIN_S2)
   799			cfg->vmid = cfg->cbndx + 1;
   800		else
   801			cfg->asid = cfg->cbndx;
   802	
   803		smmu_domain->smmu = smmu;
 > 804		smmu_domain->dev = dev;

805
806 pgtbl_cfg = (struct io_pgtable_cfg) {
807 .pgsize_bitmap = smmu->pgsize_bitmap,
808 .ias = ias,
809 .oas = oas,
810 .coherent_walk = smmu->features & ARM_SMMU_FEAT_COHERENT_WALK,
811 .tlb = smmu_domain->flush_ops,
812 .iommu_dev = smmu->dev,
813 };
814
815 if (smmu->impl && smmu->impl->init_context) {
816 ret = smmu->impl->init_context(smmu_domain, &pgtbl_cfg);
817 if (ret)
818 goto out_unlock;
819 }
820
821 if (smmu_domain->non_strict)
822 pgtbl_cfg.quirks |= IO_PGTABLE_QUIRK_NON_STRICT;
823
824 pgtbl_ops = alloc_io_pgtable_ops(fmt, &pgtbl_cfg, smmu_domain);
825 if (!pgtbl_ops) {
826 ret = -ENOMEM;
827 goto out_clear_smmu;
828 }
829
830 /* Update the domain's page sizes to reflect the page table format */
831 domain->pgsize_bitmap = pgtbl_cfg.pgsize_bitmap;
832
833 if (pgtbl_cfg.quirks & IO_PGTABLE_QUIRK_ARM_TTBR1) {
834 domain->geometry.aperture_start = ~0UL << ias;
835 domain->geometry.aperture_end = ~0UL;
836 } else {
837 domain->geometry.aperture_end = (1UL << ias) - 1;
838 }
839
840 domain->geometry.force_aperture = true;
841
842 /* Initialise the context bank with our page table cfg */
843 arm_smmu_init_context_bank(smmu_domain, &pgtbl_cfg);
844 arm_smmu_write_context_bank(smmu, cfg->cbndx);
845
846 /*
847 * Request context fault interrupt. Do this last to avoid the
848 * handler seeing a half-initialised domain state.
849 */
850 irq = smmu->irqs[smmu->num_global_irqs + cfg->irptndx];
851 ret = devm_request_irq(smmu->dev, irq, arm_smmu_context_fault,
852 IRQF_SHARED, "arm-smmu-context-fault", domain);
853 if (ret < 0) {
854 dev_err(smmu->dev, "failed to request context IRQ %d (%u)\n",
855 cfg->irptndx, irq);
856 cfg->irptndx = ARM_SMMU_INVALID_IRPTNDX;
857 }
858
859 mutex_unlock(&smmu_domain->init_mutex);
860
861 /* Publish page table ops for map/unmap */
862 smmu_domain->pgtbl_ops = pgtbl_ops;
863 return 0;
864
865 out_clear_smmu:
866 __arm_smmu_free_bitmap(smmu->context_map, cfg->cbndx);
867 smmu_domain->smmu = NULL;
868 out_unlock:
869 mutex_unlock(&smmu_domain->init_mutex);
870 return ret;
871 }
872

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

#7 lore raw

PATCH v8 5/7 iommu/arm-smmu: Add implementation for the adreno GPU SMMU

In reply to Jordan Crouse

Add a special implementation for the SMMU attached to most Adreno GPU target triggered from the qcom,adreno-gpu-smmu compatible string. When selected the driver will attempt to enable split pagetables.

---

 drivers/iommu/arm-smmu-impl.c |  3 +++
 drivers/iommu/arm-smmu-qcom.c | 45 +++++++++++++++++++++++++++++++++--
 drivers/iommu/arm-smmu.h      |  1 +
 3 files changed, 47 insertions(+), 2 deletions(-)
drivers/iommu/arm-smmu-impl.c +3 -0
index a20e426d81ac..309675cf6699 100644
@@ -176,5 +176,8 @@ struct arm_smmu_device *arm_smmu_impl_init(struct arm_smmu_device *smmu)
 	    of_device_is_compatible(np, "qcom,sc7180-smmu-500"))
 		return qcom_smmu_impl_init(smmu);
 
+	if (of_device_is_compatible(smmu->dev->of_node, "qcom,adreno-smmu"))
+		return qcom_adreno_smmu_impl_init(smmu);
+
 	return smmu;
 }
drivers/iommu/arm-smmu-qcom.c +43 -2

64 lines not on this page · raw message

drivers/iommu/arm-smmu.h +1 -0
index d33cfe26b2f5..c417814f1d98 100644
@@ -466,6 +466,7 @@ static inline void arm_smmu_writeq(struct arm_smmu_device *smmu, int page,
 
 struct arm_smmu_device *arm_smmu_impl_init(struct arm_smmu_device *smmu);
 struct arm_smmu_device *qcom_smmu_impl_init(struct arm_smmu_device *smmu);
+struct arm_smmu_device *qcom_adreno_smmu_impl_init(struct arm_smmu_device *smmu);
 
 int arm_mmu500_reset(struct arm_smmu_device *smmu);
 
Signature · 6 lines
2.17.1

_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
#8 lore raw

Re: [PATCH v8 5/7] iommu/arm-smmu: Add implementation for the adreno GPU SMMU

Thank you for the patch! Yet something to improve:

Hi Jordan,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on next-20200611]
[cannot apply to iommu/next robh/for-next arm/for-next keystone/next rockchip/for-next arm64/for-next/core shawnguo/for-next soc/for-next v5.7] [if your patch is applied to the wrong git tree, please drop us a note to help
improve the system. BTW, we also suggest to use '--base' option to specify the
base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url:    https://github.com/0day-ci/linux/commits/Jordan-Crouse/iommu-arm-smmu-Enable-split-pagetable-support/20200612-094718
base:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git b961f8dc8976c091180839f4483d67b7c2ca2578
config: arm64-randconfig-s031-20200612 (attached as .config)
compiler: aarch64-linux-gcc (GCC) 9.3.0
reproduce:
        # apt-get install sparse
        # sparse version: v0.6.1-250-g42323db3-dirty
        # save the attached .config to linux build tree
        make W=1 C=1 ARCH=arm64 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__'

If you fix the issue, kindly add following tag as appropriate

All error/warnings (new ones prefixed by >>, old ones prefixed by <<):

drivers/iommu/arm-smmu-qcom.c: In function 'qcom_adreno_smmu_is_gpu_device':

>> drivers/iommu/arm-smmu-qcom.c:17:49: error: 'smmu_domain->dev' is a pointer; did you mean to use '->'?
17 |  return of_device_is_compatible(smmu_domain->dev.of_node, "qcom,adreno");
|                                                 ^
|                                                 ->
>> drivers/iommu/arm-smmu-qcom.c:18:1: warning: control reaches end of non-void function [-Wreturn-type]

18 | }
| ^

vim +17 drivers/iommu/arm-smmu-qcom.c

    14	
    15	static bool qcom_adreno_smmu_is_gpu_device(struct arm_smmu_domain *smmu_domain)
    16	{
  > 17		return of_device_is_compatible(smmu_domain->dev.of_node, "qcom,adreno");
  > 18	}
    19	

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

#9 lore raw

Re: [PATCH v8 5/7] iommu/arm-smmu: Add implementation for the adreno GPU SMMU

In reply to Jordan Crouse
Thank you for the patch! Yet something to improve:

Hi Jordan,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on next-20200612]
[cannot apply to iommu/next robh/for-next arm/for-next keystone/next rockchip/for-next arm64/for-next/core shawnguo/for-next soc/for-next v5.7] [if your patch is applied to the wrong git tree, please drop us a note to help
improve the system. BTW, we also suggest to use '--base' option to specify the
base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url:    https://github.com/0day-ci/linux/commits/Jordan-Crouse/iommu-arm-smmu-Enable-split-pagetable-support/20200612-094718
base:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git b961f8dc8976c091180839f4483d67b7c2ca2578
config: arm64-randconfig-r026-20200612 (attached as .config)
compiler: clang version 11.0.0 (https://github.com/llvm/llvm-project 3b43f006294971b8049d4807110032169780e5b8)

reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# install arm64 cross compiling tool for clang build
# apt-get install binutils-aarch64-linux-gnu
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross ARCH=arm64

If you fix the issue, kindly add following tag as appropriate

All errors (new ones prefixed by >>, old ones prefixed by <<):

>> drivers/iommu/arm-smmu-qcom.c:17:49: error: member reference type 'struct device *' is a pointer; did you mean to use '->'?

return of_device_is_compatible(smmu_domain->dev.of_node, "qcom,adreno"); ~~~~~~~~~~~~~~~~^
->
1 error generated.

vim +17 drivers/iommu/arm-smmu-qcom.c

    14	
    15	static bool qcom_adreno_smmu_is_gpu_device(struct arm_smmu_domain *smmu_domain)
    16	{
  > 17		return of_device_is_compatible(smmu_domain->dev.of_node, "qcom,adreno");
    18	}
    19	

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

#10 lore raw

PATCH v8 6/7 drm/msm: Set the global virtual address range from the IOMMU domain

In reply to Jordan Crouse

Use the aperture settings from the IOMMU domain to set up the virtual address range for the GPU. This allows us to transparently deal with IOMMU side features (like split pagetables).

---

 drivers/gpu/drm/msm/adreno/adreno_gpu.c | 13 +++++++++++--
 drivers/gpu/drm/msm/msm_iommu.c         |  7 +++++++
 2 files changed, 18 insertions(+), 2 deletions(-)
drivers/gpu/drm/msm/adreno/adreno_gpu.c +11 -2
index 89673c7ed473..3e717c1ebb7f 100644
@@ -192,9 +192,18 @@ adreno_iommu_create_address_space(struct msm_gpu *gpu,
 	struct iommu_domain *iommu = iommu_domain_alloc(&platform_bus_type);
 	struct msm_mmu *mmu = msm_iommu_new(&pdev->dev, iommu);
 	struct msm_gem_address_space *aspace;
+	u64 start, size;
 
-	aspace = msm_gem_address_space_create(mmu, "gpu", SZ_16M,
-		0xfffffff);
+	/*
+	 * Use the aperture start or SZ_16M, whichever is greater. This will
+	 * ensure that we align with the allocated pagetable range while still
+	 * allowing room in the lower 32 bits for GMEM and whatnot
+	 */
+	start = max_t(u64, SZ_16M, iommu->geometry.aperture_start);
+	size = iommu->geometry.aperture_end - start + 1;
+
+	aspace = msm_gem_address_space_create(mmu, "gpu",
+		start & GENMASK(48, 0), size);
 
 	if (IS_ERR(aspace) && !IS_ERR(mmu))
 		mmu->funcs->destroy(mmu);
drivers/gpu/drm/msm/msm_iommu.c +7 -0
index 3a381a9674c9..bbe129867590 100644
@@ -36,6 +36,10 @@ static int msm_iommu_map(struct msm_mmu *mmu, uint64_t iova,
 	struct msm_iommu *iommu = to_msm_iommu(mmu);
 	size_t ret;
 
+	/* The arm-smmu driver expects the addresses to be sign extended */
+	if (iova & BIT(48))
+		iova |= GENMASK(63, 49);
+
 	ret = iommu_map_sg(iommu->domain, iova, sgt->sgl, sgt->nents, prot);
 	WARN_ON(!ret);
 
@@ -46,6 +50,9 @@ static int msm_iommu_unmap(struct msm_mmu *mmu, uint64_t iova, size_t len)
 {
 	struct msm_iommu *iommu = to_msm_iommu(mmu);
 
+	if (iova & BIT(48))
+		iova |= GENMASK(63, 49);
+
 	iommu_unmap(iommu->domain, iova, len);
 
 	return 0;
Signature · 6 lines
2.17.1

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
#11 lore raw

Re: [PATCH v8 6/7] drm/msm: Set the global virtual address range from the IOMMU domain

Thank you for the patch! Perhaps something to improve:

Hi Jordan,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on linus/master]
[also build test WARNING on next-20200611]
[cannot apply to iommu/next robh/for-next arm/for-next keystone/next rockchip/for-next arm64/for-next/core shawnguo/for-next soc/for-next v5.7] [if your patch is applied to the wrong git tree, please drop us a note to help
improve the system. BTW, we also suggest to use '--base' option to specify the
base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url:    https://github.com/0day-ci/linux/commits/Jordan-Crouse/iommu-arm-smmu-Enable-split-pagetable-support/20200612-094718
base:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git b961f8dc8976c091180839f4483d67b7c2ca2578
config: arm-defconfig (attached as .config)
compiler: arm-linux-gnueabi-gcc (GCC) 9.3.0

reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=arm

If you fix the issue, kindly add following tag as appropriate

All warnings (new ones prefixed by >>, old ones prefixed by <<):

In file included from include/linux/bitops.h:5,
from include/linux/kernel.h:12,
from include/linux/ascii85.h:11,
from drivers/gpu/drm/msm/adreno/adreno_gpu.c:9:
drivers/gpu/drm/msm/adreno/adreno_gpu.c: In function 'adreno_iommu_create_address_space':
include/linux/bits.h:37:11: warning: right shift count is negative [-Wshift-count-negative]
37 | (~UL(0) >> (BITS_PER_LONG - 1 - (h))))
| ^~
include/linux/bits.h:39:31: note: in expansion of macro '__GENMASK'
39 | (GENMASK_INPUT_CHECK(h, l) + __GENMASK(h, l))
| ^~~~~~~~~

>> drivers/gpu/drm/msm/adreno/adreno_gpu.c:206:11: note: in expansion of macro 'GENMASK'

206 | start & GENMASK(48, 0), size);
| ^~~~~~~
--
In file included from include/linux/bits.h:6,
from include/linux/bitops.h:5,
from include/linux/kernel.h:12,
from drivers/gpu/drm/msm/msm_drv.h:11,
from drivers/gpu/drm/msm/msm_iommu.c:7:
drivers/gpu/drm/msm/msm_iommu.c: In function 'msm_iommu_map':

>> include/vdso/bits.h:7:26: warning: left shift count >= width of type [-Wshift-count-overflow]
7 | #define BIT(nr)   (UL(1) << (nr))
|                          ^~
>> drivers/gpu/drm/msm/msm_iommu.c:40:13: note: in expansion of macro 'BIT'

40 | if (iova & BIT(48))
| ^~~
In file included from include/linux/bitops.h:5,
from include/linux/kernel.h:12,
from drivers/gpu/drm/msm/msm_drv.h:11,
from drivers/gpu/drm/msm/msm_iommu.c:7:

>> include/linux/bits.h:36:22: warning: left shift count >= width of type [-Wshift-count-overflow]

36 | (((~UL(0)) - (UL(1) << (l)) + 1) & | ^~
include/linux/bits.h:39:31: note: in expansion of macro '__GENMASK' 39 | (GENMASK_INPUT_CHECK(h, l) + __GENMASK(h, l))
| ^~~~~~~~~

>> drivers/gpu/drm/msm/msm_iommu.c:41:11: note: in expansion of macro 'GENMASK'

41 | iova |= GENMASK(63, 49);
| ^~~~~~~
include/linux/bits.h:37:11: warning: right shift count is negative [-Wshift-count-negative]
37 | (~UL(0) >> (BITS_PER_LONG - 1 - (h))))
| ^~
include/linux/bits.h:39:31: note: in expansion of macro '__GENMASK'
39 | (GENMASK_INPUT_CHECK(h, l) + __GENMASK(h, l))
| ^~~~~~~~~

>> drivers/gpu/drm/msm/msm_iommu.c:41:11: note: in expansion of macro 'GENMASK'

41 | iova |= GENMASK(63, 49);
| ^~~~~~~
In file included from include/linux/bits.h:6,
from include/linux/bitops.h:5,
from include/linux/kernel.h:12,
from drivers/gpu/drm/msm/msm_drv.h:11,
from drivers/gpu/drm/msm/msm_iommu.c:7:
drivers/gpu/drm/msm/msm_iommu.c: In function 'msm_iommu_unmap':

>> include/vdso/bits.h:7:26: warning: left shift count >= width of type [-Wshift-count-overflow]

7 | #define BIT(nr) (UL(1) << (nr))
| ^~
drivers/gpu/drm/msm/msm_iommu.c:53:13: note: in expansion of macro 'BIT' 53 | if (iova & BIT(48))
| ^~~
In file included from include/linux/bitops.h:5,
from include/linux/kernel.h:12,
from drivers/gpu/drm/msm/msm_drv.h:11,
from drivers/gpu/drm/msm/msm_iommu.c:7:

>> include/linux/bits.h:36:22: warning: left shift count >= width of type [-Wshift-count-overflow]

36 | (((~UL(0)) - (UL(1) << (l)) + 1) & | ^~
include/linux/bits.h:39:31: note: in expansion of macro '__GENMASK'
39 | (GENMASK_INPUT_CHECK(h, l) + __GENMASK(h, l))
| ^~~~~~~~~
drivers/gpu/drm/msm/msm_iommu.c:54:11: note: in expansion of macro 'GENMASK'
54 | iova |= GENMASK(63, 49);
| ^~~~~~~
include/linux/bits.h:37:11: warning: right shift count is negative [-Wshift-count-negative]
37 | (~UL(0) >> (BITS_PER_LONG - 1 - (h))))
| ^~
include/linux/bits.h:39:31: note: in expansion of macro '__GENMASK'
39 | (GENMASK_INPUT_CHECK(h, l) + __GENMASK(h, l))
| ^~~~~~~~~
drivers/gpu/drm/msm/msm_iommu.c:54:11: note: in expansion of macro 'GENMASK'
54 | iova |= GENMASK(63, 49);
| ^~~~~~~

vim +/GENMASK +206 drivers/gpu/drm/msm/adreno/adreno_gpu.c

   > 9	#include <linux/ascii85.h>
    10	#include <linux/interconnect.h>
    11	#include <linux/qcom_scm.h>
    12	#include <linux/kernel.h>
    13	#include <linux/of_address.h>
    14	#include <linux/pm_opp.h>
    15	#include <linux/slab.h>
    16	#include <linux/soc/qcom/mdt_loader.h>
    17	#include <soc/qcom/ocmem.h>
    18	#include "adreno_gpu.h"
    19	#include "msm_gem.h"
    20	#include "msm_mmu.h"
    21	
    22	static bool zap_available = true;
    23	
    24	static int zap_shader_load_mdt(struct msm_gpu *gpu, const char *fwname,
    25			u32 pasid)
    26	{
    27		struct device *dev = &gpu->pdev->dev;
    28		const struct firmware *fw;
    29		const char *signed_fwname = NULL;
    30		struct device_node *np, *mem_np;
    31		struct resource r;
    32		phys_addr_t mem_phys;
    33		ssize_t mem_size;
    34		void *mem_region = NULL;
    35		int ret;
    36	
    37		if (!IS_ENABLED(CONFIG_ARCH_QCOM)) {
    38			zap_available = false;
    39			return -EINVAL;
    40		}
    41	
    42		np = of_get_child_by_name(dev->of_node, "zap-shader");
    43		if (!np) {
    44			zap_available = false;
    45			return -ENODEV;
    46		}
    47	
    48		mem_np = of_parse_phandle(np, "memory-region", 0);
    49		of_node_put(np);
    50		if (!mem_np) {
    51			zap_available = false;
    52			return -EINVAL;
    53		}
    54	
    55		ret = of_address_to_resource(mem_np, 0, &r);
    56		of_node_put(mem_np);
    57		if (ret)
    58			return ret;
    59	
    60		mem_phys = r.start;
    61	
    62		/*
    63		 * Check for a firmware-name property.  This is the new scheme
    64		 * to handle firmware that may be signed with device specific
    65		 * keys, allowing us to have a different zap fw path for different
    66		 * devices.
    67		 *
    68		 * If the firmware-name property is found, we bypass the
    69		 * adreno_request_fw() mechanism, because we don't need to handle
    70		 * the /lib/firmware/qcom/... vs /lib/firmware/... case.
    71		 *
    72		 * If the firmware-name property is not found, for backwards
    73		 * compatibility we fall back to the fwname from the gpulist
    74		 * table.
    75		 */
    76		of_property_read_string_index(np, "firmware-name", 0, &signed_fwname);
    77		if (signed_fwname) {
    78			fwname = signed_fwname;
    79			ret = request_firmware_direct(&fw, fwname, gpu->dev->dev);
    80			if (ret)
    81				fw = ERR_PTR(ret);
    82		} else if (fwname) {
    83			/* Request the MDT file from the default location: */
    84			fw = adreno_request_fw(to_adreno_gpu(gpu), fwname);
    85		} else {
    86			/*
    87			 * For new targets, we require the firmware-name property,
    88			 * if a zap-shader is required, rather than falling back
    89			 * to a firmware name specified in gpulist.
    90			 *
    91			 * Because the firmware is signed with a (potentially)
    92			 * device specific key, having the name come from gpulist
    93			 * was a bad idea, and is only provided for backwards
    94			 * compatibility for older targets.
    95			 */
    96			return -ENODEV;
    97		}
    98	
    99		if (IS_ERR(fw)) {
   100			DRM_DEV_ERROR(dev, "Unable to load %s\n", fwname);
   101			return PTR_ERR(fw);
   102		}
   103	
   104		/* Figure out how much memory we need */
   105		mem_size = qcom_mdt_get_size(fw);
   106		if (mem_size < 0) {
   107			ret = mem_size;
   108			goto out;
   109		}
   110	
   111		if (mem_size > resource_size(&r)) {
   112			DRM_DEV_ERROR(dev,
   113				"memory region is too small to load the MDT\n");
   114			ret = -E2BIG;
   115			goto out;
   116		}
   117	
   118		/* Allocate memory for the firmware image */
   119		mem_region = memremap(mem_phys, mem_size,  MEMREMAP_WC);
   120		if (!mem_region) {
   121			ret = -ENOMEM;
   122			goto out;
   123		}
   124	
   125		/*
   126		 * Load the rest of the MDT
   127		 *
   128		 * Note that we could be dealing with two different paths, since
   129		 * with upstream linux-firmware it would be in a qcom/ subdir..
   130		 * adreno_request_fw() handles this, but qcom_mdt_load() does
   131		 * not.  But since we've already gotten through adreno_request_fw()
   132		 * we know which of the two cases it is:
   133		 */
   134		if (signed_fwname || (to_adreno_gpu(gpu)->fwloc == FW_LOCATION_LEGACY)) {
   135			ret = qcom_mdt_load(dev, fw, fwname, pasid,
   136					mem_region, mem_phys, mem_size, NULL);
   137		} else {
   138			char *newname;
   139	
   140			newname = kasprintf(GFP_KERNEL, "qcom/%s", fwname);
   141	
   142			ret = qcom_mdt_load(dev, fw, newname, pasid,
   143					mem_region, mem_phys, mem_size, NULL);
   144			kfree(newname);
   145		}
   146		if (ret)
   147			goto out;
   148	
   149		/* Send the image to the secure world */
   150		ret = qcom_scm_pas_auth_and_reset(pasid);
   151	
   152		/*
   153		 * If the scm call returns -EOPNOTSUPP we assume that this target
   154		 * doesn't need/support the zap shader so quietly fail
   155		 */
   156		if (ret == -EOPNOTSUPP)
   157			zap_available = false;
   158		else if (ret)
   159			DRM_DEV_ERROR(dev, "Unable to authorize the image\n");
   160	
   161	out:
   162		if (mem_region)
   163			memunmap(mem_region);
   164	
   165		release_firmware(fw);
   166	
   167		return ret;
   168	}
   169	
   170	int adreno_zap_shader_load(struct msm_gpu *gpu, u32 pasid)
   171	{
   172		struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
   173		struct platform_device *pdev = gpu->pdev;
   174	
   175		/* Short cut if we determine the zap shader isn't available/needed */
   176		if (!zap_available)
   177			return -ENODEV;
   178	
   179		/* We need SCM to be able to load the firmware */
   180		if (!qcom_scm_is_available()) {
   181			DRM_DEV_ERROR(&pdev->dev, "SCM is not available\n");
   182			return -EPROBE_DEFER;
   183		}
   184	
   185		return zap_shader_load_mdt(gpu, adreno_gpu->info->zapfw, pasid);
   186	}
   187	
   188	struct msm_gem_address_space *
   189	adreno_iommu_create_address_space(struct msm_gpu *gpu,
   190			struct platform_device *pdev)
   191	{
   192		struct iommu_domain *iommu = iommu_domain_alloc(&platform_bus_type);
   193		struct msm_mmu *mmu = msm_iommu_new(&pdev->dev, iommu);
   194		struct msm_gem_address_space *aspace;
   195		u64 start, size;
   196	
   197		/*
   198		 * Use the aperture start or SZ_16M, whichever is greater. This will
   199		 * ensure that we align with the allocated pagetable range while still
   200		 * allowing room in the lower 32 bits for GMEM and whatnot
   201		 */
   202		start = max_t(u64, SZ_16M, iommu->geometry.aperture_start);
   203		size = iommu->geometry.aperture_end - start + 1;
   204	
   205		aspace = msm_gem_address_space_create(mmu, "gpu",
 > 206			start & GENMASK(48, 0), size);

207
208 if (IS_ERR(aspace) && !IS_ERR(mmu))
209 mmu->funcs->destroy(mmu);
210
211 return aspace;
212 }
213

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

#12 lore raw

PATCH v8 7/7 arm: dts: qcom: sm845: Set the compatible string for the GPU SMMU

In reply to Jordan Crouse

Set the qcom,adreno-smmu compatible string for the GPU SMMU to enable split pagetables.

---

 arch/arm64/boot/dts/qcom/sdm845.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm64/boot/dts/qcom/sdm845.dtsi +1 -1
index 8eb5a31346d2..8b15cd74e9ba 100644
@@ -3556,7 +3556,7 @@
 		};
 
 		adreno_smmu: iommu@5040000 {
-			compatible = "qcom,sdm845-smmu-v2", "qcom,smmu-v2";
+			compatible = "qcom,adreno-smmu", "qcom,smmu-v2";
 			reg = <0 0x5040000 0 0x10000>;
 			#iommu-cells = <1>;
 			#global-interrupts = <2>;
Signature · 1 line
2.17.1