[PATCH v5 0/7] Convert the intel iommu driver to the dma-iommu api

[PATCH v5 0/7] Convert the intel iommu driver to the dma-iommu api

[Lu Baolu]

Hi Will,

The previous post of this series could be found here.

https://lore.kernel.org/linux-iommu/20200927063437.13988-1-baolu.lu@linux.intel.com/

Changes in this version:
- Add Tested-by: Logan Gunthorpe <logang@deltatee.com>
- Rebase to v5.10-rc3

Please review and consider it for iommu/next.

Best regards,
baolu

Lu Baolu (3):
  iommu: Add quirk for Intel graphic devices in map_sg
  iommu/vt-d: Update domain geometry in iommu_ops.at(de)tach_dev
  iommu/vt-d: Cleanup after converting to dma-iommu ops

Tom Murphy (4):
  iommu: Handle freelists when using deferred flushing in iommu drivers
  iommu: Add iommu_dma_free_cpu_cached_iovas()
  iommu: Allow the dma-iommu api to use bounce buffers
  iommu/vt-d: Convert intel iommu driver to the iommu ops

 .../admin-guide/kernel-parameters.txt         |   5 -
 drivers/iommu/dma-iommu.c                     | 228 ++++-
 drivers/iommu/intel/Kconfig                   |   1 +
 drivers/iommu/intel/iommu.c                   | 905 +++---------------
 include/linux/dma-iommu.h                     |   8 +
 include/linux/iommu.h                         |   1 +
 6 files changed, 336 insertions(+), 812 deletions(-)

-- 
2.25.1

[PATCH v5 1/7] iommu: Handle freelists when using deferred flushing in iommu drivers

[Lu Baolu]

From: Tom Murphy <murphyt7@tcd.ie>

Allow the iommu_unmap_fast to return newly freed page table pages and
pass the freelist to queue_iova in the dma-iommu ops path.

This is useful for iommu drivers (in this case the intel iommu driver)
which need to wait for the ioTLB to be flushed before newly
free/unmapped page table pages can be freed. This way we can still batch
ioTLB free operations and handle the freelists.

Signed-off-by: Tom Murphy <murphyt7@tcd.ie>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Logan Gunthorpe <logang@deltatee.com>
---
 drivers/iommu/dma-iommu.c   | 29 +++++++++++++------
 drivers/iommu/intel/iommu.c | 55 ++++++++++++++++++++++++-------------
 include/linux/iommu.h       |  1 +
 3 files changed, 58 insertions(+), 27 deletions(-)

diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index 0cbcd3fc3e7e..9c827a4d2207 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -49,6 +49,18 @@ struct iommu_dma_cookie {
 	struct iommu_domain		*fq_domain;
 };
 
+static void iommu_dma_entry_dtor(unsigned long data)
+{
+	struct page *freelist = (struct page *)data;
+
+	while (freelist) {
+		unsigned long p = (unsigned long)page_address(freelist);
+
+		freelist = freelist->freelist;
+		free_page(p);
+	}
+}
+
 static inline size_t cookie_msi_granule(struct iommu_dma_cookie *cookie)
 {
 	if (cookie->type == IOMMU_DMA_IOVA_COOKIE)
@@ -343,7 +355,7 @@ static int iommu_dma_init_domain(struct iommu_domain *domain, dma_addr_t base,
 	if (!cookie->fq_domain && !iommu_domain_get_attr(domain,
 			DOMAIN_ATTR_DMA_USE_FLUSH_QUEUE, &attr) && attr) {
 		if (init_iova_flush_queue(iovad, iommu_dma_flush_iotlb_all,
-					NULL))
+					  iommu_dma_entry_dtor))
 			pr_warn("iova flush queue initialization failed\n");
 		else
 			cookie->fq_domain = domain;
@@ -440,7 +452,7 @@ static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain,
 }
 
 static void iommu_dma_free_iova(struct iommu_dma_cookie *cookie,
-		dma_addr_t iova, size_t size)
+		dma_addr_t iova, size_t size, struct page *freelist)
 {
 	struct iova_domain *iovad = &cookie->iovad;
 
@@ -449,7 +461,8 @@ static void iommu_dma_free_iova(struct iommu_dma_cookie *cookie,
 		cookie->msi_iova -= size;
 	else if (cookie->fq_domain)	/* non-strict mode */
 		queue_iova(iovad, iova_pfn(iovad, iova),
-				size >> iova_shift(iovad), 0);
+				size >> iova_shift(iovad),
+				(unsigned long)freelist);
 	else
 		free_iova_fast(iovad, iova_pfn(iovad, iova),
 				size >> iova_shift(iovad));
@@ -474,7 +487,7 @@ static void __iommu_dma_unmap(struct device *dev, dma_addr_t dma_addr,
 
 	if (!cookie->fq_domain)
 		iommu_iotlb_sync(domain, &iotlb_gather);
-	iommu_dma_free_iova(cookie, dma_addr, size);
+	iommu_dma_free_iova(cookie, dma_addr, size, iotlb_gather.freelist);
 }
 
 static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
@@ -496,7 +509,7 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
 		return DMA_MAPPING_ERROR;
 
 	if (iommu_map_atomic(domain, iova, phys - iova_off, size, prot)) {
-		iommu_dma_free_iova(cookie, iova, size);
+		iommu_dma_free_iova(cookie, iova, size, NULL);
 		return DMA_MAPPING_ERROR;
 	}
 	return iova + iova_off;
@@ -649,7 +662,7 @@ static void *iommu_dma_alloc_remap(struct device *dev, size_t size,
 out_free_sg:
 	sg_free_table(&sgt);
 out_free_iova:
-	iommu_dma_free_iova(cookie, iova, size);
+	iommu_dma_free_iova(cookie, iova, size, NULL);
 out_free_pages:
 	__iommu_dma_free_pages(pages, count);
 	return NULL;
@@ -900,7 +913,7 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg,
 	return __finalise_sg(dev, sg, nents, iova);
 
 out_free_iova:
-	iommu_dma_free_iova(cookie, iova, iova_len);
+	iommu_dma_free_iova(cookie, iova, iova_len, NULL);
 out_restore_sg:
 	__invalidate_sg(sg, nents);
 	return 0;
@@ -1228,7 +1241,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev,
 	return msi_page;
 
 out_free_iova:
-	iommu_dma_free_iova(cookie, iova, size);
+	iommu_dma_free_iova(cookie, iova, size, NULL);
 out_free_page:
 	kfree(msi_page);
 	return NULL;
diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
index af3abd285214..77fba7f8336a 100644
--- a/drivers/iommu/intel/iommu.c
+++ b/drivers/iommu/intel/iommu.c
@@ -1243,17 +1243,17 @@ static struct page *dma_pte_clear_level(struct dmar_domain *domain, int level,
    pages can only be freed after the IOTLB flush has been done. */
 static struct page *domain_unmap(struct dmar_domain *domain,
 				 unsigned long start_pfn,
-				 unsigned long last_pfn)
+				 unsigned long last_pfn,
+				 struct page *freelist)
 {
-	struct page *freelist;
-
 	BUG_ON(!domain_pfn_supported(domain, start_pfn));
 	BUG_ON(!domain_pfn_supported(domain, last_pfn));
 	BUG_ON(start_pfn > last_pfn);
 
 	/* we don't need lock here; nobody else touches the iova range */
 	freelist = dma_pte_clear_level(domain, agaw_to_level(domain->agaw),
-				       domain->pgd, 0, start_pfn, last_pfn, NULL);
+				       domain->pgd, 0, start_pfn, last_pfn,
+				       freelist);
 
 	/* free pgd */
 	if (start_pfn == 0 && last_pfn == DOMAIN_MAX_PFN(domain->gaw)) {
@@ -2011,7 +2011,8 @@ static void domain_exit(struct dmar_domain *domain)
 	if (domain->pgd) {
 		struct page *freelist;
 
-		freelist = domain_unmap(domain, 0, DOMAIN_MAX_PFN(domain->gaw));
+		freelist = domain_unmap(domain, 0,
+					DOMAIN_MAX_PFN(domain->gaw), NULL);
 		dma_free_pagelist(freelist);
 	}
 
@@ -3570,7 +3571,7 @@ static void intel_unmap(struct device *dev, dma_addr_t dev_addr, size_t size)
 	if (dev_is_pci(dev))
 		pdev = to_pci_dev(dev);
 
-	freelist = domain_unmap(domain, start_pfn, last_pfn);
+	freelist = domain_unmap(domain, start_pfn, last_pfn, NULL);
 	if (intel_iommu_strict || (pdev && pdev->untrusted) ||
 			!has_iova_flush_queue(&domain->iovad)) {
 		iommu_flush_iotlb_psi(iommu, domain, start_pfn,
@@ -4637,7 +4638,8 @@ static int intel_iommu_memory_notifier(struct notifier_block *nb,
 			struct page *freelist;
 
 			freelist = domain_unmap(si_domain,
-						start_vpfn, last_vpfn);
+						start_vpfn, last_vpfn,
+						NULL);
 
 			rcu_read_lock();
 			for_each_active_iommu(iommu, drhd)
@@ -5610,10 +5612,8 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain,
 				struct iommu_iotlb_gather *gather)
 {
 	struct dmar_domain *dmar_domain = to_dmar_domain(domain);
-	struct page *freelist = NULL;
 	unsigned long start_pfn, last_pfn;
-	unsigned int npages;
-	int iommu_id, level = 0;
+	int level = 0;
 
 	/* Cope with horrid API which requires us to unmap more than the
 	   size argument if it happens to be a large-page mapping. */
@@ -5625,22 +5625,38 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain,
 	start_pfn = iova >> VTD_PAGE_SHIFT;
 	last_pfn = (iova + size - 1) >> VTD_PAGE_SHIFT;
 
-	freelist = domain_unmap(dmar_domain, start_pfn, last_pfn);
-
-	npages = last_pfn - start_pfn + 1;
-
-	for_each_domain_iommu(iommu_id, dmar_domain)
-		iommu_flush_iotlb_psi(g_iommus[iommu_id], dmar_domain,
-				      start_pfn, npages, !freelist, 0);
-
-	dma_free_pagelist(freelist);
+	gather->freelist = domain_unmap(dmar_domain, start_pfn,
+					last_pfn, gather->freelist);
 
 	if (dmar_domain->max_addr == iova + size)
 		dmar_domain->max_addr = iova;
 
+	iommu_iotlb_gather_add_page(domain, gather, iova, size);
+
 	return size;
 }
 
+static void intel_iommu_tlb_sync(struct iommu_domain *domain,
+				 struct iommu_iotlb_gather *gather)
+{
+	struct dmar_domain *dmar_domain = to_dmar_domain(domain);
+	unsigned long iova_pfn = IOVA_PFN(gather->start);
+	size_t size = gather->end - gather->start;
+	unsigned long start_pfn, last_pfn;
+	unsigned long nrpages;
+	int iommu_id;
+
+	nrpages = aligned_nrpages(gather->start, size);
+	start_pfn = mm_to_dma_pfn(iova_pfn);
+	last_pfn = start_pfn + nrpages - 1;
+
+	for_each_domain_iommu(iommu_id, dmar_domain)
+		iommu_flush_iotlb_psi(g_iommus[iommu_id], dmar_domain,
+				      start_pfn, nrpages, !gather->freelist, 0);
+
+	dma_free_pagelist(gather->freelist);
+}
+
 static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain,
 					    dma_addr_t iova)
 {
@@ -6100,6 +6116,7 @@ const struct iommu_ops intel_iommu_ops = {
 	.aux_get_pasid		= intel_iommu_aux_get_pasid,
 	.map			= intel_iommu_map,
 	.unmap			= intel_iommu_unmap,
+	.iotlb_sync		= intel_iommu_tlb_sync,
 	.iova_to_phys		= intel_iommu_iova_to_phys,
 	.probe_device		= intel_iommu_probe_device,
 	.probe_finalize		= intel_iommu_probe_finalize,
diff --git a/include/linux/iommu.h b/include/linux/iommu.h
index b95a6f8db6ff..e56bae327e35 100644
--- a/include/linux/iommu.h
+++ b/include/linux/iommu.h
@@ -180,6 +180,7 @@ struct iommu_iotlb_gather {
 	unsigned long		start;
 	unsigned long		end;
 	size_t			pgsize;
+	struct page		*freelist;
 };
 
 /**
-- 
2.25.1

[PATCH v5 2/7] iommu: Add iommu_dma_free_cpu_cached_iovas()

[Lu Baolu]

From: Tom Murphy <murphyt7@tcd.ie>

Add a iommu_dma_free_cpu_cached_iovas function to allow drivers which
use the dma-iommu ops to free cached cpu iovas.

Signed-off-by: Tom Murphy <murphyt7@tcd.ie>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Logan Gunthorpe <logang@deltatee.com>
---
 drivers/iommu/dma-iommu.c | 9 +++++++++
 include/linux/dma-iommu.h | 8 ++++++++
 2 files changed, 17 insertions(+)

diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index 9c827a4d2207..de521e22bafb 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -49,6 +49,15 @@ struct iommu_dma_cookie {
 	struct iommu_domain		*fq_domain;
 };
 
+void iommu_dma_free_cpu_cached_iovas(unsigned int cpu,
+		struct iommu_domain *domain)
+{
+	struct iommu_dma_cookie *cookie = domain->iova_cookie;
+	struct iova_domain *iovad = &cookie->iovad;
+
+	free_cpu_cached_iovas(cpu, iovad);
+}
+
 static void iommu_dma_entry_dtor(unsigned long data)
 {
 	struct page *freelist = (struct page *)data;
diff --git a/include/linux/dma-iommu.h b/include/linux/dma-iommu.h
index 2112f21f73d8..706b68d1359b 100644
--- a/include/linux/dma-iommu.h
+++ b/include/linux/dma-iommu.h
@@ -37,6 +37,9 @@ void iommu_dma_compose_msi_msg(struct msi_desc *desc,
 
 void iommu_dma_get_resv_regions(struct device *dev, struct list_head *list);
 
+void iommu_dma_free_cpu_cached_iovas(unsigned int cpu,
+		struct iommu_domain *domain);
+
 #else /* CONFIG_IOMMU_DMA */
 
 struct iommu_domain;
@@ -78,5 +81,10 @@ static inline void iommu_dma_get_resv_regions(struct device *dev, struct list_he
 {
 }
 
+static inline void iommu_dma_free_cpu_cached_iovas(unsigned int cpu,
+		struct iommu_domain *domain)
+{
+}
+
 #endif	/* CONFIG_IOMMU_DMA */
 #endif	/* __DMA_IOMMU_H */
-- 
2.25.1

[PATCH v5 3/7] iommu: Allow the dma-iommu api to use bounce buffers

[Lu Baolu]

From: Tom Murphy <murphyt7@tcd.ie>

Allow the dma-iommu api to use bounce buffers for untrusted devices.
This is a copy of the intel bounce buffer code.

Signed-off-by: Tom Murphy <murphyt7@tcd.ie>
Co-developed-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Logan Gunthorpe <logang@deltatee.com>
---
 drivers/iommu/dma-iommu.c | 163 +++++++++++++++++++++++++++++++++++---
 1 file changed, 150 insertions(+), 13 deletions(-)

diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index de521e22bafb..10dafbc3d9e0 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -20,9 +20,11 @@
 #include <linux/mm.h>
 #include <linux/mutex.h>
 #include <linux/pci.h>
+#include <linux/swiotlb.h>
 #include <linux/scatterlist.h>
 #include <linux/vmalloc.h>
 #include <linux/crash_dump.h>
+#include <linux/dma-direct.h>
 
 struct iommu_dma_msi_page {
 	struct list_head	list;
@@ -499,6 +501,31 @@ static void __iommu_dma_unmap(struct device *dev, dma_addr_t dma_addr,
 	iommu_dma_free_iova(cookie, dma_addr, size, iotlb_gather.freelist);
 }
 
+static void __iommu_dma_unmap_swiotlb(struct device *dev, dma_addr_t dma_addr,
+		size_t size, enum dma_data_direction dir,
+		unsigned long attrs)
+{
+	struct iommu_domain *domain = iommu_get_dma_domain(dev);
+	struct iommu_dma_cookie *cookie = domain->iova_cookie;
+	struct iova_domain *iovad = &cookie->iovad;
+	phys_addr_t phys;
+
+	phys = iommu_iova_to_phys(domain, dma_addr);
+	if (WARN_ON(!phys))
+		return;
+
+	__iommu_dma_unmap(dev, dma_addr, size);
+
+	if (unlikely(is_swiotlb_buffer(phys)))
+		swiotlb_tbl_unmap_single(dev, phys, size,
+				iova_align(iovad, size), dir, attrs);
+}
+
+static bool dev_is_untrusted(struct device *dev)
+{
+	return dev_is_pci(dev) && to_pci_dev(dev)->untrusted;
+}
+
 static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
 		size_t size, int prot, u64 dma_mask)
 {
@@ -524,6 +551,55 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
 	return iova + iova_off;
 }
 
+static dma_addr_t __iommu_dma_map_swiotlb(struct device *dev, phys_addr_t phys,
+		size_t org_size, dma_addr_t dma_mask, bool coherent,
+		enum dma_data_direction dir, unsigned long attrs)
+{
+	int prot = dma_info_to_prot(dir, coherent, attrs);
+	struct iommu_domain *domain = iommu_get_dma_domain(dev);
+	struct iommu_dma_cookie *cookie = domain->iova_cookie;
+	struct iova_domain *iovad = &cookie->iovad;
+	size_t aligned_size = org_size;
+	void *padding_start;
+	size_t padding_size;
+	dma_addr_t iova;
+
+	/*
+	 * If both the physical buffer start address and size are
+	 * page aligned, we don't need to use a bounce page.
+	 */
+	if (IS_ENABLED(CONFIG_SWIOTLB) && dev_is_untrusted(dev) &&
+	    iova_offset(iovad, phys | org_size)) {
+		aligned_size = iova_align(iovad, org_size);
+		phys = swiotlb_tbl_map_single(dev,
+				phys_to_dma(dev, io_tlb_start),
+				phys, org_size, aligned_size, dir, attrs);
+
+		if (phys == DMA_MAPPING_ERROR)
+			return DMA_MAPPING_ERROR;
+
+		/* Cleanup the padding area. */
+		padding_start = phys_to_virt(phys);
+		padding_size = aligned_size;
+
+		if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC) &&
+		    (dir == DMA_TO_DEVICE ||
+		     dir == DMA_BIDIRECTIONAL)) {
+			padding_start += org_size;
+			padding_size -= org_size;
+		}
+
+		memset(padding_start, 0, padding_size);
+	}
+
+	iova = __iommu_dma_map(dev, phys, aligned_size, prot, dma_mask);
+	if ((iova == DMA_MAPPING_ERROR) && is_swiotlb_buffer(phys))
+		swiotlb_tbl_unmap_single(dev, phys, org_size,
+				aligned_size, dir, attrs);
+
+	return iova;
+}
+
 static void __iommu_dma_free_pages(struct page **pages, int count)
 {
 	while (count--)
@@ -697,11 +773,15 @@ static void iommu_dma_sync_single_for_cpu(struct device *dev,
 {
 	phys_addr_t phys;
 
-	if (dev_is_dma_coherent(dev))
+	if (dev_is_dma_coherent(dev) && !dev_is_untrusted(dev))
 		return;
 
 	phys = iommu_iova_to_phys(iommu_get_dma_domain(dev), dma_handle);
-	arch_sync_dma_for_cpu(phys, size, dir);
+	if (!dev_is_dma_coherent(dev))
+		arch_sync_dma_for_cpu(phys, size, dir);
+
+	if (is_swiotlb_buffer(phys))
+		swiotlb_tbl_sync_single(dev, phys, size, dir, SYNC_FOR_CPU);
 }
 
 static void iommu_dma_sync_single_for_device(struct device *dev,
@@ -709,11 +789,15 @@ static void iommu_dma_sync_single_for_device(struct device *dev,
 {
 	phys_addr_t phys;
 
-	if (dev_is_dma_coherent(dev))
+	if (dev_is_dma_coherent(dev) && !dev_is_untrusted(dev))
 		return;
 
 	phys = iommu_iova_to_phys(iommu_get_dma_domain(dev), dma_handle);
-	arch_sync_dma_for_device(phys, size, dir);
+	if (is_swiotlb_buffer(phys))
+		swiotlb_tbl_sync_single(dev, phys, size, dir, SYNC_FOR_DEVICE);
+
+	if (!dev_is_dma_coherent(dev))
+		arch_sync_dma_for_device(phys, size, dir);
 }
 
 static void iommu_dma_sync_sg_for_cpu(struct device *dev,
@@ -723,11 +807,17 @@ static void iommu_dma_sync_sg_for_cpu(struct device *dev,
 	struct scatterlist *sg;
 	int i;
 
-	if (dev_is_dma_coherent(dev))
+	if (dev_is_dma_coherent(dev) && !dev_is_untrusted(dev))
 		return;
 
-	for_each_sg(sgl, sg, nelems, i)
-		arch_sync_dma_for_cpu(sg_phys(sg), sg->length, dir);
+	for_each_sg(sgl, sg, nelems, i) {
+		if (!dev_is_dma_coherent(dev))
+			arch_sync_dma_for_cpu(sg_phys(sg), sg->length, dir);
+
+		if (is_swiotlb_buffer(sg_phys(sg)))
+			swiotlb_tbl_sync_single(dev, sg_phys(sg), sg->length,
+						dir, SYNC_FOR_CPU);
+	}
 }
 
 static void iommu_dma_sync_sg_for_device(struct device *dev,
@@ -737,11 +827,17 @@ static void iommu_dma_sync_sg_for_device(struct device *dev,
 	struct scatterlist *sg;
 	int i;
 
-	if (dev_is_dma_coherent(dev))
+	if (dev_is_dma_coherent(dev) && !dev_is_untrusted(dev))
 		return;
 
-	for_each_sg(sgl, sg, nelems, i)
-		arch_sync_dma_for_device(sg_phys(sg), sg->length, dir);
+	for_each_sg(sgl, sg, nelems, i) {
+		if (is_swiotlb_buffer(sg_phys(sg)))
+			swiotlb_tbl_sync_single(dev, sg_phys(sg), sg->length,
+						dir, SYNC_FOR_DEVICE);
+
+		if (!dev_is_dma_coherent(dev))
+			arch_sync_dma_for_device(sg_phys(sg), sg->length, dir);
+	}
 }
 
 static dma_addr_t iommu_dma_map_page(struct device *dev, struct page *page,
@@ -750,10 +846,10 @@ static dma_addr_t iommu_dma_map_page(struct device *dev, struct page *page,
 {
 	phys_addr_t phys = page_to_phys(page) + offset;
 	bool coherent = dev_is_dma_coherent(dev);
-	int prot = dma_info_to_prot(dir, coherent, attrs);
 	dma_addr_t dma_handle;
 
-	dma_handle = __iommu_dma_map(dev, phys, size, prot, dma_get_mask(dev));
+	dma_handle = __iommu_dma_map_swiotlb(dev, phys, size, dma_get_mask(dev),
+			coherent, dir, attrs);
 	if (!coherent && !(attrs & DMA_ATTR_SKIP_CPU_SYNC) &&
 	    dma_handle != DMA_MAPPING_ERROR)
 		arch_sync_dma_for_device(phys, size, dir);
@@ -765,7 +861,7 @@ static void iommu_dma_unmap_page(struct device *dev, dma_addr_t dma_handle,
 {
 	if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC))
 		iommu_dma_sync_single_for_cpu(dev, dma_handle, size, dir);
-	__iommu_dma_unmap(dev, dma_handle, size);
+	__iommu_dma_unmap_swiotlb(dev, dma_handle, size, dir, attrs);
 }
 
 /*
@@ -843,6 +939,39 @@ static void __invalidate_sg(struct scatterlist *sg, int nents)
 	}
 }
 
+static void iommu_dma_unmap_sg_swiotlb(struct device *dev, struct scatterlist *sg,
+		int nents, enum dma_data_direction dir, unsigned long attrs)
+{
+	struct scatterlist *s;
+	int i;
+
+	for_each_sg(sg, s, nents, i)
+		__iommu_dma_unmap_swiotlb(dev, sg_dma_address(s),
+				sg_dma_len(s), dir, attrs);
+}
+
+static int iommu_dma_map_sg_swiotlb(struct device *dev, struct scatterlist *sg,
+		int nents, enum dma_data_direction dir, unsigned long attrs)
+{
+	struct scatterlist *s;
+	int i;
+
+	for_each_sg(sg, s, nents, i) {
+		sg_dma_address(s) = __iommu_dma_map_swiotlb(dev, sg_phys(s),
+				s->length, dma_get_mask(dev),
+				dev_is_dma_coherent(dev), dir, attrs);
+		if (sg_dma_address(s) == DMA_MAPPING_ERROR)
+			goto out_unmap;
+		sg_dma_len(s) = s->length;
+	}
+
+	return nents;
+
+out_unmap:
+	iommu_dma_unmap_sg_swiotlb(dev, sg, i, dir, attrs | DMA_ATTR_SKIP_CPU_SYNC);
+	return 0;
+}
+
 /*
  * The DMA API client is passing in a scatterlist which could describe
  * any old buffer layout, but the IOMMU API requires everything to be
@@ -869,6 +998,9 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg,
 	if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC))
 		iommu_dma_sync_sg_for_device(dev, sg, nents, dir);
 
+	if (dev_is_untrusted(dev))
+		return iommu_dma_map_sg_swiotlb(dev, sg, nents, dir, attrs);
+
 	/*
 	 * Work out how much IOVA space we need, and align the segments to
 	 * IOVA granules for the IOMMU driver to handle. With some clever
@@ -938,6 +1070,11 @@ static void iommu_dma_unmap_sg(struct device *dev, struct scatterlist *sg,
 	if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC))
 		iommu_dma_sync_sg_for_cpu(dev, sg, nents, dir);
 
+	if (dev_is_untrusted(dev)) {
+		iommu_dma_unmap_sg_swiotlb(dev, sg, nents, dir, attrs);
+		return;
+	}
+
 	/*
 	 * The scatterlist segments are mapped into a single
 	 * contiguous IOVA allocation, so this is incredibly easy.
-- 
2.25.1

[PATCH v5 4/7] iommu: Add quirk for Intel graphic devices in map_sg

[Lu Baolu]

Combining the sg segments exposes a bug in the Intel i915 driver which
causes visual artifacts and the screen to freeze. This is most likely
because of how the i915 handles the returned list. It probably doesn't
respect the returned value specifying the number of elements in the list
and instead depends on the previous behaviour of the Intel iommu driver
which would return the same number of elements in the output list as in
the input list.

Signed-off-by: Tom Murphy <murphyt7@tcd.ie>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Logan Gunthorpe <logang@deltatee.com>
---
 drivers/iommu/dma-iommu.c | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index 10dafbc3d9e0..5591d6593583 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -879,6 +879,33 @@ static int __finalise_sg(struct device *dev, struct scatterlist *sg, int nents,
 	unsigned int cur_len = 0, max_len = dma_get_max_seg_size(dev);
 	int i, count = 0;
 
+	/*
+	 * The Intel graphic driver is used to assume that the returned
+	 * sg list is not combound. This blocks the efforts of converting
+	 * Intel IOMMU driver to dma-iommu api's. Add this quirk to make the
+	 * device driver work and should be removed once it's fixed in i915
+	 * driver.
+	 */
+	if (IS_ENABLED(CONFIG_DRM_I915) && dev_is_pci(dev) &&
+	    to_pci_dev(dev)->vendor == PCI_VENDOR_ID_INTEL &&
+	    (to_pci_dev(dev)->class >> 16) == PCI_BASE_CLASS_DISPLAY) {
+		for_each_sg(sg, s, nents, i) {
+			unsigned int s_iova_off = sg_dma_address(s);
+			unsigned int s_length = sg_dma_len(s);
+			unsigned int s_iova_len = s->length;
+
+			s->offset += s_iova_off;
+			s->length = s_length;
+			sg_dma_address(s) = dma_addr + s_iova_off;
+			sg_dma_len(s) = s_length;
+			dma_addr += s_iova_len;
+
+			pr_info_once("sg combining disabled due to i915 driver\n");
+		}
+
+		return nents;
+	}
+
 	for_each_sg(sg, s, nents, i) {
 		/* Restore this segment's original unaligned fields first */
 		unsigned int s_iova_off = sg_dma_address(s);
-- 
2.25.1

[PATCH v5 5/7] iommu/vt-d: Update domain geometry in iommu_ops.at(de)tach_dev

[Lu Baolu]

The iommu-dma constrains IOVA allocation based on the domain geometry
that the driver reports. Update domain geometry everytime a domain is
attached to or detached from a device.

Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Logan Gunthorpe <logang@deltatee.com>
---
 drivers/iommu/intel/iommu.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
index 77fba7f8336a..80a7e4e1781b 100644
--- a/drivers/iommu/intel/iommu.c
+++ b/drivers/iommu/intel/iommu.c
@@ -67,8 +67,8 @@
 #define MAX_AGAW_WIDTH 64
 #define MAX_AGAW_PFN_WIDTH	(MAX_AGAW_WIDTH - VTD_PAGE_SHIFT)
 
-#define __DOMAIN_MAX_PFN(gaw)  ((((uint64_t)1) << (gaw-VTD_PAGE_SHIFT)) - 1)
-#define __DOMAIN_MAX_ADDR(gaw) ((((uint64_t)1) << gaw) - 1)
+#define __DOMAIN_MAX_PFN(gaw)  ((((uint64_t)1) << ((gaw) - VTD_PAGE_SHIFT)) - 1)
+#define __DOMAIN_MAX_ADDR(gaw) ((((uint64_t)1) << (gaw)) - 1)
 
 /* We limit DOMAIN_MAX_PFN to fit in an unsigned long, and DOMAIN_MAX_ADDR
    to match. That way, we can use 'unsigned long' for PFNs with impunity. */
@@ -739,6 +739,18 @@ static void domain_update_iommu_cap(struct dmar_domain *domain)
 	 */
 	if (domain->nid == NUMA_NO_NODE)
 		domain->nid = domain_update_device_node(domain);
+
+	/*
+	 * First-level translation restricts the input-address to a
+	 * canonical address (i.e., address bits 63:N have the same
+	 * value as address bit [N-1], where N is 48-bits with 4-level
+	 * paging and 57-bits with 5-level paging). Hence, skip bit
+	 * [N-1].
+	 */
+	if (domain_use_first_level(domain))
+		domain->domain.geometry.aperture_end = __DOMAIN_MAX_ADDR(domain->gaw - 1);
+	else
+		domain->domain.geometry.aperture_end = __DOMAIN_MAX_ADDR(domain->gaw);
 }
 
 struct context_entry *iommu_context_addr(struct intel_iommu *iommu, u8 bus,
-- 
2.25.1

[PATCH v5 6/7] iommu/vt-d: Convert intel iommu driver to the iommu ops

[Lu Baolu]

From: Tom Murphy <murphyt7@tcd.ie>

Convert the intel iommu driver to the dma-iommu api. Remove the iova
handling and reserve region code from the intel iommu driver.

Signed-off-by: Tom Murphy <murphyt7@tcd.ie>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Logan Gunthorpe <logang@deltatee.com>
---
 drivers/iommu/intel/Kconfig |   1 +
 drivers/iommu/intel/iommu.c | 746 ++----------------------------------
 2 files changed, 43 insertions(+), 704 deletions(-)

diff --git a/drivers/iommu/intel/Kconfig b/drivers/iommu/intel/Kconfig
index 5337ee1584b0..28a3d1596c76 100644
--- a/drivers/iommu/intel/Kconfig
+++ b/drivers/iommu/intel/Kconfig
@@ -13,6 +13,7 @@ config INTEL_IOMMU
 	select DMAR_TABLE
 	select SWIOTLB
 	select IOASID
+	select IOMMU_DMA
 	help
 	  DMA remapping (DMAR) devices support enables independent address
 	  translations for Direct Memory Access (DMA) from devices.
diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
index 80a7e4e1781b..7b36b48b9887 100644
--- a/drivers/iommu/intel/iommu.c
+++ b/drivers/iommu/intel/iommu.c
@@ -31,6 +31,7 @@
 #include <linux/io.h>
 #include <linux/iova.h>
 #include <linux/iommu.h>
+#include <linux/dma-iommu.h>
 #include <linux/intel-iommu.h>
 #include <linux/syscore_ops.h>
 #include <linux/tboot.h>
@@ -41,7 +42,6 @@
 #include <linux/dma-direct.h>
 #include <linux/crash_dump.h>
 #include <linux/numa.h>
-#include <linux/swiotlb.h>
 #include <asm/irq_remapping.h>
 #include <asm/cacheflush.h>
 #include <asm/iommu.h>
@@ -382,9 +382,6 @@ struct device_domain_info *get_domain_info(struct device *dev)
 DEFINE_SPINLOCK(device_domain_lock);
 static LIST_HEAD(device_domain_list);
 
-#define device_needs_bounce(d) (!intel_no_bounce && dev_is_pci(d) &&	\
-				to_pci_dev(d)->untrusted)
-
 /*
  * Iterate over elements in device_domain_list and call the specified
  * callback @fn against each element.
@@ -1289,13 +1286,6 @@ static void dma_free_pagelist(struct page *freelist)
 	}
 }
 
-static void iova_entry_free(unsigned long data)
-{
-	struct page *freelist = (struct page *)data;
-
-	dma_free_pagelist(freelist);
-}
-
 /* iommu handling */
 static int iommu_alloc_root_entry(struct intel_iommu *iommu)
 {
@@ -1660,19 +1650,17 @@ static inline void __mapping_notify_one(struct intel_iommu *iommu,
 		iommu_flush_write_buffer(iommu);
 }
 
-static void iommu_flush_iova(struct iova_domain *iovad)
+static void intel_flush_iotlb_all(struct iommu_domain *domain)
 {
-	struct dmar_domain *domain;
+	struct dmar_domain *dmar_domain = to_dmar_domain(domain);
 	int idx;
 
-	domain = container_of(iovad, struct dmar_domain, iovad);
-
-	for_each_domain_iommu(idx, domain) {
+	for_each_domain_iommu(idx, dmar_domain) {
 		struct intel_iommu *iommu = g_iommus[idx];
-		u16 did = domain->iommu_did[iommu->seq_id];
+		u16 did = dmar_domain->iommu_did[iommu->seq_id];
 
-		if (domain_use_first_level(domain))
-			domain_flush_piotlb(iommu, domain, 0, -1, 0);
+		if (domain_use_first_level(dmar_domain))
+			domain_flush_piotlb(iommu, dmar_domain, 0, -1, 0);
 		else
 			iommu->flush.flush_iotlb(iommu, did, 0, 0,
 						 DMA_TLB_DSI_FLUSH);
@@ -1954,48 +1942,6 @@ static int domain_detach_iommu(struct dmar_domain *domain,
 	return count;
 }
 
-static struct iova_domain reserved_iova_list;
-static struct lock_class_key reserved_rbtree_key;
-
-static int dmar_init_reserved_ranges(void)
-{
-	struct pci_dev *pdev = NULL;
-	struct iova *iova;
-	int i;
-
-	init_iova_domain(&reserved_iova_list, VTD_PAGE_SIZE, IOVA_START_PFN);
-
-	lockdep_set_class(&reserved_iova_list.iova_rbtree_lock,
-		&reserved_rbtree_key);
-
-	/* IOAPIC ranges shouldn't be accessed by DMA */
-	iova = reserve_iova(&reserved_iova_list, IOVA_PFN(IOAPIC_RANGE_START),
-		IOVA_PFN(IOAPIC_RANGE_END));
-	if (!iova) {
-		pr_err("Reserve IOAPIC range failed\n");
-		return -ENODEV;
-	}
-
-	/* Reserve all PCI MMIO to avoid peer-to-peer access */
-	for_each_pci_dev(pdev) {
-		struct resource *r;
-
-		for (i = 0; i < PCI_NUM_RESOURCES; i++) {
-			r = &pdev->resource[i];
-			if (!r->flags || !(r->flags & IORESOURCE_MEM))
-				continue;
-			iova = reserve_iova(&reserved_iova_list,
-					    IOVA_PFN(r->start),
-					    IOVA_PFN(r->end));
-			if (!iova) {
-				pci_err(pdev, "Reserve iova for %pR failed\n", r);
-				return -ENODEV;
-			}
-		}
-	}
-	return 0;
-}
-
 static inline int guestwidth_to_adjustwidth(int gaw)
 {
 	int agaw;
@@ -2018,7 +1964,7 @@ static void domain_exit(struct dmar_domain *domain)
 
 	/* destroy iovas */
 	if (domain->domain.type == IOMMU_DOMAIN_DMA)
-		put_iova_domain(&domain->iovad);
+		iommu_put_dma_cookie(&domain->domain);
 
 	if (domain->pgd) {
 		struct page *freelist;
@@ -2552,16 +2498,6 @@ struct dmar_domain *find_domain(struct device *dev)
 	return NULL;
 }
 
-static void do_deferred_attach(struct device *dev)
-{
-	struct iommu_domain *domain;
-
-	dev_iommu_priv_set(dev, NULL);
-	domain = iommu_get_domain_for_dev(dev);
-	if (domain)
-		intel_iommu_attach_device(domain, dev);
-}
-
 static inline struct device_domain_info *
 dmar_search_domain_by_dev_info(int segment, int bus, int devfn)
 {
@@ -3434,595 +3370,6 @@ static int __init init_dmars(void)
 	return ret;
 }
 
-/* This takes a number of _MM_ pages, not VTD pages */
-static unsigned long intel_alloc_iova(struct device *dev,
-				     struct dmar_domain *domain,
-				     unsigned long nrpages, uint64_t dma_mask)
-{
-	unsigned long iova_pfn;
-
-	/*
-	 * Restrict dma_mask to the width that the iommu can handle.
-	 * First-level translation restricts the input-address to a
-	 * canonical address (i.e., address bits 63:N have the same
-	 * value as address bit [N-1], where N is 48-bits with 4-level
-	 * paging and 57-bits with 5-level paging). Hence, skip bit
-	 * [N-1].
-	 */
-	if (domain_use_first_level(domain))
-		dma_mask = min_t(uint64_t, DOMAIN_MAX_ADDR(domain->gaw - 1),
-				 dma_mask);
-	else
-		dma_mask = min_t(uint64_t, DOMAIN_MAX_ADDR(domain->gaw),
-				 dma_mask);
-
-	/* Ensure we reserve the whole size-aligned region */
-	nrpages = __roundup_pow_of_two(nrpages);
-
-	if (!dmar_forcedac && dma_mask > DMA_BIT_MASK(32)) {
-		/*
-		 * First try to allocate an io virtual address in
-		 * DMA_BIT_MASK(32) and if that fails then try allocating
-		 * from higher range
-		 */
-		iova_pfn = alloc_iova_fast(&domain->iovad, nrpages,
-					   IOVA_PFN(DMA_BIT_MASK(32)), false);
-		if (iova_pfn)
-			return iova_pfn;
-	}
-	iova_pfn = alloc_iova_fast(&domain->iovad, nrpages,
-				   IOVA_PFN(dma_mask), true);
-	if (unlikely(!iova_pfn)) {
-		dev_err_once(dev, "Allocating %ld-page iova failed\n",
-			     nrpages);
-		return 0;
-	}
-
-	return iova_pfn;
-}
-
-static dma_addr_t __intel_map_single(struct device *dev, phys_addr_t paddr,
-				     size_t size, int dir, u64 dma_mask)
-{
-	struct dmar_domain *domain;
-	phys_addr_t start_paddr;
-	unsigned long iova_pfn;
-	int prot = 0;
-	int ret;
-	struct intel_iommu *iommu;
-	unsigned long paddr_pfn = paddr >> PAGE_SHIFT;
-
-	BUG_ON(dir == DMA_NONE);
-
-	if (unlikely(attach_deferred(dev)))
-		do_deferred_attach(dev);
-
-	domain = find_domain(dev);
-	if (!domain)
-		return DMA_MAPPING_ERROR;
-
-	iommu = domain_get_iommu(domain);
-	size = aligned_nrpages(paddr, size);
-
-	iova_pfn = intel_alloc_iova(dev, domain, dma_to_mm_pfn(size), dma_mask);
-	if (!iova_pfn)
-		goto error;
-
-	/*
-	 * Check if DMAR supports zero-length reads on write only
-	 * mappings..
-	 */
-	if (dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL || \
-			!cap_zlr(iommu->cap))
-		prot |= DMA_PTE_READ;
-	if (dir == DMA_FROM_DEVICE || dir == DMA_BIDIRECTIONAL)
-		prot |= DMA_PTE_WRITE;
-	/*
-	 * paddr - (paddr + size) might be partial page, we should map the whole
-	 * page.  Note: if two part of one page are separately mapped, we
-	 * might have two guest_addr mapping to the same host paddr, but this
-	 * is not a big problem
-	 */
-	ret = domain_pfn_mapping(domain, mm_to_dma_pfn(iova_pfn),
-				 mm_to_dma_pfn(paddr_pfn), size, prot);
-	if (ret)
-		goto error;
-
-	start_paddr = (phys_addr_t)iova_pfn << PAGE_SHIFT;
-	start_paddr += paddr & ~PAGE_MASK;
-
-	trace_map_single(dev, start_paddr, paddr, size << VTD_PAGE_SHIFT);
-
-	return start_paddr;
-
-error:
-	if (iova_pfn)
-		free_iova_fast(&domain->iovad, iova_pfn, dma_to_mm_pfn(size));
-	dev_err(dev, "Device request: %zx@%llx dir %d --- failed\n",
-		size, (unsigned long long)paddr, dir);
-	return DMA_MAPPING_ERROR;
-}
-
-static dma_addr_t intel_map_page(struct device *dev, struct page *page,
-				 unsigned long offset, size_t size,
-				 enum dma_data_direction dir,
-				 unsigned long attrs)
-{
-	return __intel_map_single(dev, page_to_phys(page) + offset,
-				  size, dir, *dev->dma_mask);
-}
-
-static dma_addr_t intel_map_resource(struct device *dev, phys_addr_t phys_addr,
-				     size_t size, enum dma_data_direction dir,
-				     unsigned long attrs)
-{
-	return __intel_map_single(dev, phys_addr, size, dir, *dev->dma_mask);
-}
-
-static void intel_unmap(struct device *dev, dma_addr_t dev_addr, size_t size)
-{
-	struct dmar_domain *domain;
-	unsigned long start_pfn, last_pfn;
-	unsigned long nrpages;
-	unsigned long iova_pfn;
-	struct intel_iommu *iommu;
-	struct page *freelist;
-	struct pci_dev *pdev = NULL;
-
-	domain = find_domain(dev);
-	BUG_ON(!domain);
-
-	iommu = domain_get_iommu(domain);
-
-	iova_pfn = IOVA_PFN(dev_addr);
-
-	nrpages = aligned_nrpages(dev_addr, size);
-	start_pfn = mm_to_dma_pfn(iova_pfn);
-	last_pfn = start_pfn + nrpages - 1;
-
-	if (dev_is_pci(dev))
-		pdev = to_pci_dev(dev);
-
-	freelist = domain_unmap(domain, start_pfn, last_pfn, NULL);
-	if (intel_iommu_strict || (pdev && pdev->untrusted) ||
-			!has_iova_flush_queue(&domain->iovad)) {
-		iommu_flush_iotlb_psi(iommu, domain, start_pfn,
-				      nrpages, !freelist, 0);
-		/* free iova */
-		free_iova_fast(&domain->iovad, iova_pfn, dma_to_mm_pfn(nrpages));
-		dma_free_pagelist(freelist);
-	} else {
-		queue_iova(&domain->iovad, iova_pfn, nrpages,
-			   (unsigned long)freelist);
-		/*
-		 * queue up the release of the unmap to save the 1/6th of the
-		 * cpu used up by the iotlb flush operation...
-		 */
-	}
-
-	trace_unmap_single(dev, dev_addr, size);
-}
-
-static void intel_unmap_page(struct device *dev, dma_addr_t dev_addr,
-			     size_t size, enum dma_data_direction dir,
-			     unsigned long attrs)
-{
-	intel_unmap(dev, dev_addr, size);
-}
-
-static void intel_unmap_resource(struct device *dev, dma_addr_t dev_addr,
-		size_t size, enum dma_data_direction dir, unsigned long attrs)
-{
-	intel_unmap(dev, dev_addr, size);
-}
-
-static void *intel_alloc_coherent(struct device *dev, size_t size,
-				  dma_addr_t *dma_handle, gfp_t flags,
-				  unsigned long attrs)
-{
-	struct page *page = NULL;
-	int order;
-
-	if (unlikely(attach_deferred(dev)))
-		do_deferred_attach(dev);
-
-	size = PAGE_ALIGN(size);
-	order = get_order(size);
-
-	if (gfpflags_allow_blocking(flags)) {
-		unsigned int count = size >> PAGE_SHIFT;
-
-		page = dma_alloc_from_contiguous(dev, count, order,
-						 flags & __GFP_NOWARN);
-	}
-
-	if (!page)
-		page = alloc_pages(flags, order);
-	if (!page)
-		return NULL;
-	memset(page_address(page), 0, size);
-
-	*dma_handle = __intel_map_single(dev, page_to_phys(page), size,
-					 DMA_BIDIRECTIONAL,
-					 dev->coherent_dma_mask);
-	if (*dma_handle != DMA_MAPPING_ERROR)
-		return page_address(page);
-	if (!dma_release_from_contiguous(dev, page, size >> PAGE_SHIFT))
-		__free_pages(page, order);
-
-	return NULL;
-}
-
-static void intel_free_coherent(struct device *dev, size_t size, void *vaddr,
-				dma_addr_t dma_handle, unsigned long attrs)
-{
-	int order;
-	struct page *page = virt_to_page(vaddr);
-
-	size = PAGE_ALIGN(size);
-	order = get_order(size);
-
-	intel_unmap(dev, dma_handle, size);
-	if (!dma_release_from_contiguous(dev, page, size >> PAGE_SHIFT))
-		__free_pages(page, order);
-}
-
-static void intel_unmap_sg(struct device *dev, struct scatterlist *sglist,
-			   int nelems, enum dma_data_direction dir,
-			   unsigned long attrs)
-{
-	dma_addr_t startaddr = sg_dma_address(sglist) & PAGE_MASK;
-	unsigned long nrpages = 0;
-	struct scatterlist *sg;
-	int i;
-
-	for_each_sg(sglist, sg, nelems, i) {
-		nrpages += aligned_nrpages(sg_dma_address(sg), sg_dma_len(sg));
-	}
-
-	intel_unmap(dev, startaddr, nrpages << VTD_PAGE_SHIFT);
-
-	trace_unmap_sg(dev, startaddr, nrpages << VTD_PAGE_SHIFT);
-}
-
-static int intel_map_sg(struct device *dev, struct scatterlist *sglist, int nelems,
-			enum dma_data_direction dir, unsigned long attrs)
-{
-	int i;
-	struct dmar_domain *domain;
-	size_t size = 0;
-	int prot = 0;
-	unsigned long iova_pfn;
-	int ret;
-	struct scatterlist *sg;
-	unsigned long start_vpfn;
-	struct intel_iommu *iommu;
-
-	BUG_ON(dir == DMA_NONE);
-
-	if (unlikely(attach_deferred(dev)))
-		do_deferred_attach(dev);
-
-	domain = find_domain(dev);
-	if (!domain)
-		return 0;
-
-	iommu = domain_get_iommu(domain);
-
-	for_each_sg(sglist, sg, nelems, i)
-		size += aligned_nrpages(sg->offset, sg->length);
-
-	iova_pfn = intel_alloc_iova(dev, domain, dma_to_mm_pfn(size),
-				*dev->dma_mask);
-	if (!iova_pfn) {
-		sglist->dma_length = 0;
-		return 0;
-	}
-
-	/*
-	 * Check if DMAR supports zero-length reads on write only
-	 * mappings..
-	 */
-	if (dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL || \
-			!cap_zlr(iommu->cap))
-		prot |= DMA_PTE_READ;
-	if (dir == DMA_FROM_DEVICE || dir == DMA_BIDIRECTIONAL)
-		prot |= DMA_PTE_WRITE;
-
-	start_vpfn = mm_to_dma_pfn(iova_pfn);
-
-	ret = domain_sg_mapping(domain, start_vpfn, sglist, size, prot);
-	if (unlikely(ret)) {
-		dma_pte_free_pagetable(domain, start_vpfn,
-				       start_vpfn + size - 1,
-				       agaw_to_level(domain->agaw) + 1);
-		free_iova_fast(&domain->iovad, iova_pfn, dma_to_mm_pfn(size));
-		return 0;
-	}
-
-	for_each_sg(sglist, sg, nelems, i)
-		trace_map_sg(dev, i + 1, nelems, sg);
-
-	return nelems;
-}
-
-static u64 intel_get_required_mask(struct device *dev)
-{
-	return DMA_BIT_MASK(32);
-}
-
-static const struct dma_map_ops intel_dma_ops = {
-	.alloc = intel_alloc_coherent,
-	.free = intel_free_coherent,
-	.map_sg = intel_map_sg,
-	.unmap_sg = intel_unmap_sg,
-	.map_page = intel_map_page,
-	.unmap_page = intel_unmap_page,
-	.map_resource = intel_map_resource,
-	.unmap_resource = intel_unmap_resource,
-	.dma_supported = dma_direct_supported,
-	.mmap = dma_common_mmap,
-	.get_sgtable = dma_common_get_sgtable,
-	.alloc_pages = dma_common_alloc_pages,
-	.free_pages = dma_common_free_pages,
-	.get_required_mask = intel_get_required_mask,
-};
-
-static void
-bounce_sync_single(struct device *dev, dma_addr_t addr, size_t size,
-		   enum dma_data_direction dir, enum dma_sync_target target)
-{
-	struct dmar_domain *domain;
-	phys_addr_t tlb_addr;
-
-	domain = find_domain(dev);
-	if (WARN_ON(!domain))
-		return;
-
-	tlb_addr = intel_iommu_iova_to_phys(&domain->domain, addr);
-	if (is_swiotlb_buffer(tlb_addr))
-		swiotlb_tbl_sync_single(dev, tlb_addr, size, dir, target);
-}
-
-static dma_addr_t
-bounce_map_single(struct device *dev, phys_addr_t paddr, size_t size,
-		  enum dma_data_direction dir, unsigned long attrs,
-		  u64 dma_mask)
-{
-	size_t aligned_size = ALIGN(size, VTD_PAGE_SIZE);
-	struct dmar_domain *domain;
-	struct intel_iommu *iommu;
-	unsigned long iova_pfn;
-	unsigned long nrpages;
-	phys_addr_t tlb_addr;
-	int prot = 0;
-	int ret;
-
-	if (unlikely(attach_deferred(dev)))
-		do_deferred_attach(dev);
-
-	domain = find_domain(dev);
-
-	if (WARN_ON(dir == DMA_NONE || !domain))
-		return DMA_MAPPING_ERROR;
-
-	iommu = domain_get_iommu(domain);
-	if (WARN_ON(!iommu))
-		return DMA_MAPPING_ERROR;
-
-	nrpages = aligned_nrpages(0, size);
-	iova_pfn = intel_alloc_iova(dev, domain,
-				    dma_to_mm_pfn(nrpages), dma_mask);
-	if (!iova_pfn)
-		return DMA_MAPPING_ERROR;
-
-	/*
-	 * Check if DMAR supports zero-length reads on write only
-	 * mappings..
-	 */
-	if (dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL ||
-			!cap_zlr(iommu->cap))
-		prot |= DMA_PTE_READ;
-	if (dir == DMA_FROM_DEVICE || dir == DMA_BIDIRECTIONAL)
-		prot |= DMA_PTE_WRITE;
-
-	/*
-	 * If both the physical buffer start address and size are
-	 * page aligned, we don't need to use a bounce page.
-	 */
-	if (!IS_ALIGNED(paddr | size, VTD_PAGE_SIZE)) {
-		tlb_addr = swiotlb_tbl_map_single(dev,
-				phys_to_dma_unencrypted(dev, io_tlb_start),
-				paddr, size, aligned_size, dir, attrs);
-		if (tlb_addr == DMA_MAPPING_ERROR) {
-			goto swiotlb_error;
-		} else {
-			/* Cleanup the padding area. */
-			void *padding_start = phys_to_virt(tlb_addr);
-			size_t padding_size = aligned_size;
-
-			if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC) &&
-			    (dir == DMA_TO_DEVICE ||
-			     dir == DMA_BIDIRECTIONAL)) {
-				padding_start += size;
-				padding_size -= size;
-			}
-
-			memset(padding_start, 0, padding_size);
-		}
-	} else {
-		tlb_addr = paddr;
-	}
-
-	ret = domain_pfn_mapping(domain, mm_to_dma_pfn(iova_pfn),
-				 tlb_addr >> VTD_PAGE_SHIFT, nrpages, prot);
-	if (ret)
-		goto mapping_error;
-
-	trace_bounce_map_single(dev, iova_pfn << PAGE_SHIFT, paddr, size);
-
-	return (phys_addr_t)iova_pfn << PAGE_SHIFT;
-
-mapping_error:
-	if (is_swiotlb_buffer(tlb_addr))
-		swiotlb_tbl_unmap_single(dev, tlb_addr, size,
-					 aligned_size, dir, attrs);
-swiotlb_error:
-	free_iova_fast(&domain->iovad, iova_pfn, dma_to_mm_pfn(nrpages));
-	dev_err(dev, "Device bounce map: %zx@%llx dir %d --- failed\n",
-		size, (unsigned long long)paddr, dir);
-
-	return DMA_MAPPING_ERROR;
-}
-
-static void
-bounce_unmap_single(struct device *dev, dma_addr_t dev_addr, size_t size,
-		    enum dma_data_direction dir, unsigned long attrs)
-{
-	size_t aligned_size = ALIGN(size, VTD_PAGE_SIZE);
-	struct dmar_domain *domain;
-	phys_addr_t tlb_addr;
-
-	domain = find_domain(dev);
-	if (WARN_ON(!domain))
-		return;
-
-	tlb_addr = intel_iommu_iova_to_phys(&domain->domain, dev_addr);
-	if (WARN_ON(!tlb_addr))
-		return;
-
-	intel_unmap(dev, dev_addr, size);
-	if (is_swiotlb_buffer(tlb_addr))
-		swiotlb_tbl_unmap_single(dev, tlb_addr, size,
-					 aligned_size, dir, attrs);
-
-	trace_bounce_unmap_single(dev, dev_addr, size);
-}
-
-static dma_addr_t
-bounce_map_page(struct device *dev, struct page *page, unsigned long offset,
-		size_t size, enum dma_data_direction dir, unsigned long attrs)
-{
-	return bounce_map_single(dev, page_to_phys(page) + offset,
-				 size, dir, attrs, *dev->dma_mask);
-}
-
-static dma_addr_t
-bounce_map_resource(struct device *dev, phys_addr_t phys_addr, size_t size,
-		    enum dma_data_direction dir, unsigned long attrs)
-{
-	return bounce_map_single(dev, phys_addr, size,
-				 dir, attrs, *dev->dma_mask);
-}
-
-static void
-bounce_unmap_page(struct device *dev, dma_addr_t dev_addr, size_t size,
-		  enum dma_data_direction dir, unsigned long attrs)
-{
-	bounce_unmap_single(dev, dev_addr, size, dir, attrs);
-}
-
-static void
-bounce_unmap_resource(struct device *dev, dma_addr_t dev_addr, size_t size,
-		      enum dma_data_direction dir, unsigned long attrs)
-{
-	bounce_unmap_single(dev, dev_addr, size, dir, attrs);
-}
-
-static void
-bounce_unmap_sg(struct device *dev, struct scatterlist *sglist, int nelems,
-		enum dma_data_direction dir, unsigned long attrs)
-{
-	struct scatterlist *sg;
-	int i;
-
-	for_each_sg(sglist, sg, nelems, i)
-		bounce_unmap_page(dev, sg->dma_address,
-				  sg_dma_len(sg), dir, attrs);
-}
-
-static int
-bounce_map_sg(struct device *dev, struct scatterlist *sglist, int nelems,
-	      enum dma_data_direction dir, unsigned long attrs)
-{
-	int i;
-	struct scatterlist *sg;
-
-	for_each_sg(sglist, sg, nelems, i) {
-		sg->dma_address = bounce_map_page(dev, sg_page(sg),
-						  sg->offset, sg->length,
-						  dir, attrs);
-		if (sg->dma_address == DMA_MAPPING_ERROR)
-			goto out_unmap;
-		sg_dma_len(sg) = sg->length;
-	}
-
-	for_each_sg(sglist, sg, nelems, i)
-		trace_bounce_map_sg(dev, i + 1, nelems, sg);
-
-	return nelems;
-
-out_unmap:
-	bounce_unmap_sg(dev, sglist, i, dir, attrs | DMA_ATTR_SKIP_CPU_SYNC);
-	return 0;
-}
-
-static void
-bounce_sync_single_for_cpu(struct device *dev, dma_addr_t addr,
-			   size_t size, enum dma_data_direction dir)
-{
-	bounce_sync_single(dev, addr, size, dir, SYNC_FOR_CPU);
-}
-
-static void
-bounce_sync_single_for_device(struct device *dev, dma_addr_t addr,
-			      size_t size, enum dma_data_direction dir)
-{
-	bounce_sync_single(dev, addr, size, dir, SYNC_FOR_DEVICE);
-}
-
-static void
-bounce_sync_sg_for_cpu(struct device *dev, struct scatterlist *sglist,
-		       int nelems, enum dma_data_direction dir)
-{
-	struct scatterlist *sg;
-	int i;
-
-	for_each_sg(sglist, sg, nelems, i)
-		bounce_sync_single(dev, sg_dma_address(sg),
-				   sg_dma_len(sg), dir, SYNC_FOR_CPU);
-}
-
-static void
-bounce_sync_sg_for_device(struct device *dev, struct scatterlist *sglist,
-			  int nelems, enum dma_data_direction dir)
-{
-	struct scatterlist *sg;
-	int i;
-
-	for_each_sg(sglist, sg, nelems, i)
-		bounce_sync_single(dev, sg_dma_address(sg),
-				   sg_dma_len(sg), dir, SYNC_FOR_DEVICE);
-}
-
-static const struct dma_map_ops bounce_dma_ops = {
-	.alloc			= intel_alloc_coherent,
-	.free			= intel_free_coherent,
-	.map_sg			= bounce_map_sg,
-	.unmap_sg		= bounce_unmap_sg,
-	.map_page		= bounce_map_page,
-	.unmap_page		= bounce_unmap_page,
-	.sync_single_for_cpu	= bounce_sync_single_for_cpu,
-	.sync_single_for_device	= bounce_sync_single_for_device,
-	.sync_sg_for_cpu	= bounce_sync_sg_for_cpu,
-	.sync_sg_for_device	= bounce_sync_sg_for_device,
-	.map_resource		= bounce_map_resource,
-	.unmap_resource		= bounce_unmap_resource,
-	.alloc_pages		= dma_common_alloc_pages,
-	.free_pages		= dma_common_free_pages,
-	.dma_supported		= dma_direct_supported,
-};
-
 static inline int iommu_domain_cache_init(void)
 {
 	int ret = 0;
@@ -4690,7 +4037,7 @@ static void free_all_cpu_cached_iovas(unsigned int cpu)
 			if (!domain || domain->domain.type != IOMMU_DOMAIN_DMA)
 				continue;
 
-			free_cpu_cached_iovas(cpu, &domain->iovad);
+			iommu_dma_free_cpu_cached_iovas(cpu, &domain->domain);
 		}
 	}
 }
@@ -4963,12 +4310,6 @@ int __init intel_iommu_init(void)
 	if (list_empty(&dmar_atsr_units))
 		pr_info("No ATSR found\n");
 
-	if (dmar_init_reserved_ranges()) {
-		if (force_on)
-			panic("tboot: Failed to reserve iommu ranges\n");
-		goto out_free_reserved_range;
-	}
-
 	if (dmar_map_gfx)
 		intel_iommu_gfx_mapped = 1;
 
@@ -4979,7 +4320,7 @@ int __init intel_iommu_init(void)
 		if (force_on)
 			panic("tboot: Failed to initialize DMARs\n");
 		pr_err("Initialization failed\n");
-		goto out_free_reserved_range;
+		goto out_free_dmar;
 	}
 	up_write(&dmar_global_lock);
 
@@ -5020,8 +4361,6 @@ int __init intel_iommu_init(void)
 
 	return 0;
 
-out_free_reserved_range:
-	put_iova_domain(&reserved_iova_list);
 out_free_dmar:
 	intel_iommu_free_dmars();
 	up_write(&dmar_global_lock);
@@ -5119,17 +4458,6 @@ static int md_domain_init(struct dmar_domain *domain, int guest_width)
 	return 0;
 }
 
-static void intel_init_iova_domain(struct dmar_domain *dmar_domain)
-{
-	init_iova_domain(&dmar_domain->iovad, VTD_PAGE_SIZE, IOVA_START_PFN);
-	copy_reserved_iova(&reserved_iova_list, &dmar_domain->iovad);
-
-	if (!intel_iommu_strict &&
-	    init_iova_flush_queue(&dmar_domain->iovad,
-				  iommu_flush_iova, iova_entry_free))
-		pr_info("iova flush queue initialization failed\n");
-}
-
 static struct iommu_domain *intel_iommu_domain_alloc(unsigned type)
 {
 	struct dmar_domain *dmar_domain;
@@ -5149,8 +4477,9 @@ static struct iommu_domain *intel_iommu_domain_alloc(unsigned type)
 			return NULL;
 		}
 
-		if (type == IOMMU_DOMAIN_DMA)
-			intel_init_iova_domain(dmar_domain);
+		if (type == IOMMU_DOMAIN_DMA &&
+		    iommu_get_dma_cookie(&dmar_domain->domain))
+			return NULL;
 
 		domain = &dmar_domain->domain;
 		domain->geometry.aperture_start = 0;
@@ -5779,13 +5108,13 @@ static void intel_iommu_release_device(struct device *dev)
 
 static void intel_iommu_probe_finalize(struct device *dev)
 {
-	struct iommu_domain *domain;
+	dma_addr_t base = IOVA_START_PFN << VTD_PAGE_SHIFT;
+	struct iommu_domain *domain = iommu_get_domain_for_dev(dev);
+	struct dmar_domain *dmar_domain = to_dmar_domain(domain);
 
-	domain = iommu_get_domain_for_dev(dev);
-	if (device_needs_bounce(dev))
-		set_dma_ops(dev, &bounce_dma_ops);
-	else if (domain && domain->type == IOMMU_DOMAIN_DMA)
-		set_dma_ops(dev, &intel_dma_ops);
+	if (domain && domain->type == IOMMU_DOMAIN_DMA)
+		iommu_setup_dma_ops(dev, base,
+				    __DOMAIN_MAX_ADDR(dmar_domain->gaw) - base);
 	else
 		set_dma_ops(dev, NULL);
 }
@@ -5898,19 +5227,6 @@ int intel_iommu_enable_pasid(struct intel_iommu *iommu, struct device *dev)
 	return ret;
 }
 
-static void intel_iommu_apply_resv_region(struct device *dev,
-					  struct iommu_domain *domain,
-					  struct iommu_resv_region *region)
-{
-	struct dmar_domain *dmar_domain = to_dmar_domain(domain);
-	unsigned long start, end;
-
-	start = IOVA_PFN(region->start);
-	end   = IOVA_PFN(region->start + region->length - 1);
-
-	WARN_ON_ONCE(!reserve_iova(&dmar_domain->iovad, start, end));
-}
-
 static struct iommu_group *intel_iommu_device_group(struct device *dev)
 {
 	if (dev_is_pci(dev))
@@ -6099,6 +5415,27 @@ intel_iommu_domain_set_attr(struct iommu_domain *domain,
 	return ret;
 }
 
+static int
+intel_iommu_domain_get_attr(struct iommu_domain *domain,
+			    enum iommu_attr attr, void *data)
+{
+	switch (domain->type) {
+	case IOMMU_DOMAIN_UNMANAGED:
+		return -ENODEV;
+	case IOMMU_DOMAIN_DMA:
+		switch (attr) {
+		case DOMAIN_ATTR_DMA_USE_FLUSH_QUEUE:
+			*(int *)data = !intel_iommu_strict;
+			return 0;
+		default:
+			return -ENODEV;
+		}
+		break;
+	default:
+		return -EINVAL;
+	}
+}
+
 /*
  * Check that the device does not live on an external facing PCI port that is
  * marked as untrusted. Such devices should not be able to apply quirks and
@@ -6120,6 +5457,7 @@ const struct iommu_ops intel_iommu_ops = {
 	.capable		= intel_iommu_capable,
 	.domain_alloc		= intel_iommu_domain_alloc,
 	.domain_free		= intel_iommu_domain_free,
+	.domain_get_attr        = intel_iommu_domain_get_attr,
 	.domain_set_attr	= intel_iommu_domain_set_attr,
 	.attach_dev		= intel_iommu_attach_device,
 	.detach_dev		= intel_iommu_detach_device,
@@ -6128,6 +5466,7 @@ const struct iommu_ops intel_iommu_ops = {
 	.aux_get_pasid		= intel_iommu_aux_get_pasid,
 	.map			= intel_iommu_map,
 	.unmap			= intel_iommu_unmap,
+	.flush_iotlb_all        = intel_flush_iotlb_all,
 	.iotlb_sync		= intel_iommu_tlb_sync,
 	.iova_to_phys		= intel_iommu_iova_to_phys,
 	.probe_device		= intel_iommu_probe_device,
@@ -6135,7 +5474,6 @@ const struct iommu_ops intel_iommu_ops = {
 	.release_device		= intel_iommu_release_device,
 	.get_resv_regions	= intel_iommu_get_resv_regions,
 	.put_resv_regions	= generic_iommu_put_resv_regions,
-	.apply_resv_region	= intel_iommu_apply_resv_region,
 	.device_group		= intel_iommu_device_group,
 	.dev_has_feat		= intel_iommu_dev_has_feat,
 	.dev_feat_enabled	= intel_iommu_dev_feat_enabled,
-- 
2.25.1

[PATCH v5 7/7] iommu/vt-d: Cleanup after converting to dma-iommu ops

[Lu Baolu]

Some cleanups after converting the driver to use dma-iommu ops.
- Remove nobounce option;
- Cleanup and simplify the path in domain mapping.

Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Tested-by: Logan Gunthorpe <logang@deltatee.com>
---
 .../admin-guide/kernel-parameters.txt         |  5 --
 drivers/iommu/intel/iommu.c                   | 90 ++++++-------------
 2 files changed, 28 insertions(+), 67 deletions(-)

diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
index 526d65d8573a..76b2a2063fd0 100644
--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -1883,11 +1883,6 @@
 			Note that using this option lowers the security
 			provided by tboot because it makes the system
 			vulnerable to DMA attacks.
-		nobounce [Default off]
-			Disable bounce buffer for untrusted devices such as
-			the Thunderbolt devices. This will treat the untrusted
-			devices as the trusted ones, hence might expose security
-			risks of DMA attacks.
 
 	intel_idle.max_cstate=	[KNL,HW,ACPI,X86]
 			0	disables intel_idle and fall back on acpi_idle.
diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
index 7b36b48b9887..e81fdbf79b96 100644
--- a/drivers/iommu/intel/iommu.c
+++ b/drivers/iommu/intel/iommu.c
@@ -355,7 +355,6 @@ static int dmar_forcedac;
 static int intel_iommu_strict;
 static int intel_iommu_superpage = 1;
 static int iommu_identity_mapping;
-static int intel_no_bounce;
 static int iommu_skip_te_disable;
 
 #define IDENTMAP_GFX		2
@@ -457,9 +456,6 @@ static int __init intel_iommu_setup(char *str)
 		} else if (!strncmp(str, "tboot_noforce", 13)) {
 			pr_info("Intel-IOMMU: not forcing on after tboot. This could expose security risk for tboot\n");
 			intel_iommu_tboot_noforce = 1;
-		} else if (!strncmp(str, "nobounce", 8)) {
-			pr_info("Intel-IOMMU: No bounce buffer. This could expose security risks of DMA attacks\n");
-			intel_no_bounce = 1;
 		}
 
 		str += strcspn(str, ",");
@@ -2277,15 +2273,14 @@ static inline int hardware_largepage_caps(struct dmar_domain *domain,
 	return level;
 }
 
-static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
-			    struct scatterlist *sg, unsigned long phys_pfn,
-			    unsigned long nr_pages, int prot)
+static int
+__domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
+		 unsigned long phys_pfn, unsigned long nr_pages, int prot)
 {
 	struct dma_pte *first_pte = NULL, *pte = NULL;
-	phys_addr_t pteval;
-	unsigned long sg_res = 0;
 	unsigned int largepage_lvl = 0;
 	unsigned long lvl_pages = 0;
+	phys_addr_t pteval;
 	u64 attr;
 
 	BUG_ON(!domain_pfn_supported(domain, iov_pfn + nr_pages - 1));
@@ -2297,26 +2292,14 @@ static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
 	if (domain_use_first_level(domain))
 		attr |= DMA_FL_PTE_PRESENT | DMA_FL_PTE_XD | DMA_FL_PTE_US;
 
-	if (!sg) {
-		sg_res = nr_pages;
-		pteval = ((phys_addr_t)phys_pfn << VTD_PAGE_SHIFT) | attr;
-	}
+	pteval = ((phys_addr_t)phys_pfn << VTD_PAGE_SHIFT) | attr;
 
 	while (nr_pages > 0) {
 		uint64_t tmp;
 
-		if (!sg_res) {
-			unsigned int pgoff = sg->offset & ~PAGE_MASK;
-
-			sg_res = aligned_nrpages(sg->offset, sg->length);
-			sg->dma_address = ((dma_addr_t)iov_pfn << VTD_PAGE_SHIFT) + pgoff;
-			sg->dma_length = sg->length;
-			pteval = (sg_phys(sg) - pgoff) | attr;
-			phys_pfn = pteval >> VTD_PAGE_SHIFT;
-		}
-
 		if (!pte) {
-			largepage_lvl = hardware_largepage_caps(domain, iov_pfn, phys_pfn, sg_res);
+			largepage_lvl = hardware_largepage_caps(domain, iov_pfn,
+					phys_pfn, nr_pages);
 
 			first_pte = pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl);
 			if (!pte)
@@ -2328,7 +2311,7 @@ static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
 				pteval |= DMA_PTE_LARGE_PAGE;
 				lvl_pages = lvl_to_nr_pages(largepage_lvl);
 
-				nr_superpages = sg_res / lvl_pages;
+				nr_superpages = nr_pages / lvl_pages;
 				end_pfn = iov_pfn + nr_superpages * lvl_pages - 1;
 
 				/*
@@ -2362,48 +2345,45 @@ static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
 		lvl_pages = lvl_to_nr_pages(largepage_lvl);
 
 		BUG_ON(nr_pages < lvl_pages);
-		BUG_ON(sg_res < lvl_pages);
 
 		nr_pages -= lvl_pages;
 		iov_pfn += lvl_pages;
 		phys_pfn += lvl_pages;
 		pteval += lvl_pages * VTD_PAGE_SIZE;
-		sg_res -= lvl_pages;
 
 		/* If the next PTE would be the first in a new page, then we
-		   need to flush the cache on the entries we've just written.
-		   And then we'll need to recalculate 'pte', so clear it and
-		   let it get set again in the if (!pte) block above.
-
-		   If we're done (!nr_pages) we need to flush the cache too.
-
-		   Also if we've been setting superpages, we may need to
-		   recalculate 'pte' and switch back to smaller pages for the
-		   end of the mapping, if the trailing size is not enough to
-		   use another superpage (i.e. sg_res < lvl_pages). */
+		 * need to flush the cache on the entries we've just written.
+		 * And then we'll need to recalculate 'pte', so clear it and
+		 * let it get set again in the if (!pte) block above.
+		 *
+		 * If we're done (!nr_pages) we need to flush the cache too.
+		 *
+		 * Also if we've been setting superpages, we may need to
+		 * recalculate 'pte' and switch back to smaller pages for the
+		 * end of the mapping, if the trailing size is not enough to
+		 * use another superpage (i.e. nr_pages < lvl_pages).
+		 */
 		pte++;
 		if (!nr_pages || first_pte_in_page(pte) ||
-		    (largepage_lvl > 1 && sg_res < lvl_pages)) {
+		    (largepage_lvl > 1 && nr_pages < lvl_pages)) {
 			domain_flush_cache(domain, first_pte,
 					   (void *)pte - (void *)first_pte);
 			pte = NULL;
 		}
-
-		if (!sg_res && nr_pages)
-			sg = sg_next(sg);
 	}
+
 	return 0;
 }
 
-static int domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
-			  struct scatterlist *sg, unsigned long phys_pfn,
-			  unsigned long nr_pages, int prot)
+static int
+domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
+	       unsigned long phys_pfn, unsigned long nr_pages, int prot)
 {
 	int iommu_id, ret;
 	struct intel_iommu *iommu;
 
 	/* Do the real mapping first */
-	ret = __domain_mapping(domain, iov_pfn, sg, phys_pfn, nr_pages, prot);
+	ret = __domain_mapping(domain, iov_pfn, phys_pfn, nr_pages, prot);
 	if (ret)
 		return ret;
 
@@ -2415,20 +2395,6 @@ static int domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
 	return 0;
 }
 
-static inline int domain_sg_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
-				    struct scatterlist *sg, unsigned long nr_pages,
-				    int prot)
-{
-	return domain_mapping(domain, iov_pfn, sg, 0, nr_pages, prot);
-}
-
-static inline int domain_pfn_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
-				     unsigned long phys_pfn, unsigned long nr_pages,
-				     int prot)
-{
-	return domain_mapping(domain, iov_pfn, NULL, phys_pfn, nr_pages, prot);
-}
-
 static void domain_context_clear_one(struct intel_iommu *iommu, u8 bus, u8 devfn)
 {
 	unsigned long flags;
@@ -2688,7 +2654,7 @@ static int iommu_domain_identity_map(struct dmar_domain *domain,
 	 */
 	dma_pte_clear_range(domain, first_vpfn, last_vpfn);
 
-	return __domain_mapping(domain, first_vpfn, NULL,
+	return __domain_mapping(domain, first_vpfn,
 				first_vpfn, last_vpfn - first_vpfn + 1,
 				DMA_PTE_READ|DMA_PTE_WRITE);
 }
@@ -4943,8 +4909,8 @@ static int intel_iommu_map(struct iommu_domain *domain,
 	/* Round up size to next multiple of PAGE_SIZE, if it and
 	   the low bits of hpa would take us onto the next page */
 	size = aligned_nrpages(hpa, size);
-	ret = domain_pfn_mapping(dmar_domain, iova >> VTD_PAGE_SHIFT,
-				 hpa >> VTD_PAGE_SHIFT, size, prot);
+	ret = domain_mapping(dmar_domain, iova >> VTD_PAGE_SHIFT,
+			     hpa >> VTD_PAGE_SHIFT, size, prot);
 	return ret;
 }
 
-- 
2.25.1

Re: [PATCH v5 0/7] Convert the intel iommu driver to the dma-iommu api

[Chris Wilson]

Quoting Lu Baolu (2020-11-20 10:17:12)
> Lu Baolu (3):
>   iommu: Add quirk for Intel graphic devices in map_sg
>   iommu/vt-d: Update domain geometry in iommu_ops.at(de)tach_dev
>   iommu/vt-d: Cleanup after converting to dma-iommu ops
> 
> Tom Murphy (4):
>   iommu: Handle freelists when using deferred flushing in iommu drivers
>   iommu: Add iommu_dma_free_cpu_cached_iovas()
>   iommu: Allow the dma-iommu api to use bounce buffers
>   iommu/vt-d: Convert intel iommu driver to the iommu ops

Something that may be of interest is that we encounter problems with
using intel-iommu across a PCI remove event. All HW generations fail
with faults like:

DMAR: DRHD: handling fault status reg 3
DMAR: [DMA Write] Request device [00:02.0] PASID ffffffff fault addr 4b822000 [fault reason 02] Present bit in context entry is clear

i.e. they all report missing present bit after re-adding the device to the
iommu group. Forcing an identity map (or disabling iommu) works fine.

I applied this series just on the off-chance it changed the symptoms; it
does not. If you have any ideas on how to chase down this fault, that
would be very useful. We have a few other DMAR faults visible on many
platforms, all "[fault reason 07] Next page table ptr is invalid" that
are again not affected by this series, that we also need to resolve.
-Chris

Re: [PATCH v5 0/7] Convert the intel iommu driver to the dma-iommu api

[Lu Baolu]

Hi Chris,

On 2020/11/20 20:24, Chris Wilson wrote:
> Quoting Lu Baolu (2020-11-20 10:17:12)
>> Lu Baolu (3):
>>    iommu: Add quirk for Intel graphic devices in map_sg
>>    iommu/vt-d: Update domain geometry in iommu_ops.at(de)tach_dev
>>    iommu/vt-d: Cleanup after converting to dma-iommu ops
>>
>> Tom Murphy (4):
>>    iommu: Handle freelists when using deferred flushing in iommu drivers
>>    iommu: Add iommu_dma_free_cpu_cached_iovas()
>>    iommu: Allow the dma-iommu api to use bounce buffers
>>    iommu/vt-d: Convert intel iommu driver to the iommu ops
> 
> Something that may be of interest is that we encounter problems with
> using intel-iommu across a PCI remove event. All HW generations fail
> with faults like:
> 
> DMAR: DRHD: handling fault status reg 3
> DMAR: [DMA Write] Request device [00:02.0] PASID ffffffff fault addr 4b822000 [fault reason 02] Present bit in context entry is clear
> 
> i.e. they all report missing present bit after re-adding the device to the
> iommu group. Forcing an identity map (or disabling iommu) works fine.
> 
> I applied this series just on the off-chance it changed the symptoms; it
> does not. If you have any ideas on how to chase down this fault, that
> would be very useful. We have a few other DMAR faults visible on many
> platforms, all "[fault reason 07] Next page table ptr is invalid" that
> are again not affected by this series, that we also need to resolve.

This series only replaces the previous vt-d self-made dma api's with the
generic one (which has already been used by some other vendor iommu
drivers). It shouldn't solve the problem you're having.

How about reporting them in https://bugzilla.kernel.org/? I have
interests to help to figure out possible fixes.

> -Chris
> 

Best regards,
baolu

Re: [PATCH v5 3/7] iommu: Allow the dma-iommu api to use bounce buffers

[Christoph Hellwig]

> +	/*
> +	 * If both the physical buffer start address and size are
> +	 * page aligned, we don't need to use a bounce page.
> +	 */
> +	if (IS_ENABLED(CONFIG_SWIOTLB) && dev_is_untrusted(dev) &&
> +	    iova_offset(iovad, phys | org_size)) {
> +		aligned_size = iova_align(iovad, org_size);
> +		phys = swiotlb_tbl_map_single(dev,
> +				phys_to_dma(dev, io_tlb_start),
> +				phys, org_size, aligned_size, dir, attrs);

swiotlb_tbl_map_single takes one less argument in 5.10-rc now.

Re: [PATCH v5 3/7] iommu: Allow the dma-iommu api to use bounce buffers

[Lu Baolu]

Hi Christoph,

On 2020/11/23 18:08, Christoph Hellwig wrote:
>> +	/*
>> +	 * If both the physical buffer start address and size are
>> +	 * page aligned, we don't need to use a bounce page.
>> +	 */
>> +	if (IS_ENABLED(CONFIG_SWIOTLB) && dev_is_untrusted(dev) &&
>> +	    iova_offset(iovad, phys | org_size)) {
>> +		aligned_size = iova_align(iovad, org_size);
>> +		phys = swiotlb_tbl_map_single(dev,
>> +				phys_to_dma(dev, io_tlb_start),
>> +				phys, org_size, aligned_size, dir, attrs);
> 
> swiotlb_tbl_map_single takes one less argument in 5.10-rc now.
> 

Yes. But Will's iommu/next branch is based on 5.10-rc3. I synced with
him, we agreed to keep it 5.10-rc3 and resolve this conflict when
merging it.

Best regards,
baolu

Re: [PATCH v5 3/7] iommu: Allow the dma-iommu api to use bounce buffers

[Will Deacon]

On Mon, Nov 23, 2020 at 07:40:57PM +0800, Lu Baolu wrote:
> On 2020/11/23 18:08, Christoph Hellwig wrote:
> > > +	/*
> > > +	 * If both the physical buffer start address and size are
> > > +	 * page aligned, we don't need to use a bounce page.
> > > +	 */
> > > +	if (IS_ENABLED(CONFIG_SWIOTLB) && dev_is_untrusted(dev) &&
> > > +	    iova_offset(iovad, phys | org_size)) {
> > > +		aligned_size = iova_align(iovad, org_size);
> > > +		phys = swiotlb_tbl_map_single(dev,
> > > +				phys_to_dma(dev, io_tlb_start),
> > > +				phys, org_size, aligned_size, dir, attrs);
> > 
> > swiotlb_tbl_map_single takes one less argument in 5.10-rc now.
> > 
> 
> Yes. But Will's iommu/next branch is based on 5.10-rc3. I synced with
> him, we agreed to keep it 5.10-rc3 and resolve this conflict when
> merging it.

That's right, although I failed to appreciate the conflict was due to a
change in function prototype rather than just a context collision. So
I've updated the vt-d branch to contain the stuff fron Konrad:

https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=for-next/iommu/vt-d

Sorry for messing you around!

Will

Re: [PATCH v5 3/7] iommu: Allow the dma-iommu api to use bounce buffers

[Lu Baolu]

Hi Will,

On 2020/11/23 19:47, Will Deacon wrote:
> On Mon, Nov 23, 2020 at 07:40:57PM +0800, Lu Baolu wrote:
>> On 2020/11/23 18:08, Christoph Hellwig wrote:
>>>> +	/*
>>>> +	 * If both the physical buffer start address and size are
>>>> +	 * page aligned, we don't need to use a bounce page.
>>>> +	 */
>>>> +	if (IS_ENABLED(CONFIG_SWIOTLB) && dev_is_untrusted(dev) &&
>>>> +	    iova_offset(iovad, phys | org_size)) {
>>>> +		aligned_size = iova_align(iovad, org_size);
>>>> +		phys = swiotlb_tbl_map_single(dev,
>>>> +				phys_to_dma(dev, io_tlb_start),
>>>> +				phys, org_size, aligned_size, dir, attrs);
>>>
>>> swiotlb_tbl_map_single takes one less argument in 5.10-rc now.
>>>
>>
>> Yes. But Will's iommu/next branch is based on 5.10-rc3. I synced with
>> him, we agreed to keep it 5.10-rc3 and resolve this conflict when
>> merging it.
> 
> That's right, although I failed to appreciate the conflict was due to a
> change in function prototype rather than just a context collision. So
> I've updated the vt-d branch to contain the stuff fron Konrad:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=for-next/iommu/vt-d
> 
> Sorry for messing you around!

It's okay. I will re-base the patch series later.

Best regards,
baolu