add support for the global coherent pool to the dma core

add support for the global coherent pool to the dma core

[Christoph Hellwig]

Hi all,

this series adds support for using the global coherent (aka uncached)
pool to the generic dma-direct code and then switches arm-nommu and
hexagon over to it, together with a bunch of cleanups.

Diffstat:
 arch/arm/Kconfig                |    5 -
 arch/arm/mm/dma-mapping-nommu.c |  173 +---------------------------------------
 arch/hexagon/Kconfig            |    1 
 arch/hexagon/kernel/dma.c       |   57 ++-----------
 include/linux/dma-map-ops.h     |   18 ++--
 kernel/dma/Kconfig              |    4 
 kernel/dma/coherent.c           |  159 +++++++++++++++++-------------------
 kernel/dma/direct.c             |   15 +++
 8 files changed, 124 insertions(+), 308 deletions(-)

[PATCH 1/7] dma-direct: add support for dma_coherent_default_memory

[Christoph Hellwig]

Add an option to allocate uncached memory for dma_alloc_coherent from
the global dma_coherent_default_memory.  This will allow to move
arm-nommu (and eventually other platforms) to use generic code for
allocating uncached memory from a pre-populated pool.

Note that this is a different pool from the one that platforms that
can remap at runtime use for GFP_ATOMIC allocations for now, although
there might be opportunities to eventually end up with a common codebase
for the two use cases.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Dillon Min <dillon.minfei@gmail.com>
---
 kernel/dma/Kconfig  |  4 ++++
 kernel/dma/direct.c | 15 +++++++++++++++
 2 files changed, 19 insertions(+)

diff --git a/kernel/dma/Kconfig b/kernel/dma/Kconfig
index 77b405508743..725cfd51762b 100644
--- a/kernel/dma/Kconfig
+++ b/kernel/dma/Kconfig
@@ -93,6 +93,10 @@ config DMA_COHERENT_POOL
 	select GENERIC_ALLOCATOR
 	bool
 
+config DMA_GLOBAL_POOL
+	select DMA_DECLARE_COHERENT
+	bool
+
 config DMA_REMAP
 	bool
 	depends on MMU
diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c
index f737e3347059..d1d0258ed6d0 100644
--- a/kernel/dma/direct.c
+++ b/kernel/dma/direct.c
@@ -156,9 +156,14 @@ void *dma_direct_alloc(struct device *dev, size_t size,
 
 	if (!IS_ENABLED(CONFIG_ARCH_HAS_DMA_SET_UNCACHED) &&
 	    !IS_ENABLED(CONFIG_DMA_DIRECT_REMAP) &&
+	    !IS_ENABLED(CONFIG_DMA_GLOBAL_POOL) &&
 	    !dev_is_dma_coherent(dev))
 		return arch_dma_alloc(dev, size, dma_handle, gfp, attrs);
 
+	if (IS_ENABLED(CONFIG_DMA_GLOBAL_POOL) &&
+	    !dev_is_dma_coherent(dev))
+		return dma_alloc_from_global_coherent(dev, size, dma_handle);
+
 	/*
 	 * Remapping or decrypting memory may block. If either is required and
 	 * we can't block, allocate the memory from the atomic pools.
@@ -255,11 +260,19 @@ void dma_direct_free(struct device *dev, size_t size,
 
 	if (!IS_ENABLED(CONFIG_ARCH_HAS_DMA_SET_UNCACHED) &&
 	    !IS_ENABLED(CONFIG_DMA_DIRECT_REMAP) &&
+	    !IS_ENABLED(CONFIG_DMA_GLOBAL_POOL) &&
 	    !dev_is_dma_coherent(dev)) {
 		arch_dma_free(dev, size, cpu_addr, dma_addr, attrs);
 		return;
 	}
 
+	if (IS_ENABLED(CONFIG_DMA_GLOBAL_POOL) &&
+	    !dev_is_dma_coherent(dev)) {
+		if (!dma_release_from_global_coherent(page_order, cpu_addr))
+			WARN_ON_ONCE(1);
+		return;
+	}
+
 	/* If cpu_addr is not from an atomic pool, dma_free_from_pool() fails */
 	if (IS_ENABLED(CONFIG_DMA_COHERENT_POOL) &&
 	    dma_free_from_pool(dev, cpu_addr, PAGE_ALIGN(size)))
@@ -462,6 +475,8 @@ int dma_direct_mmap(struct device *dev, struct vm_area_struct *vma,
 
 	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
 		return ret;
+	if (dma_mmap_from_global_coherent(vma, cpu_addr, size, &ret))
+		return ret;
 
 	if (vma->vm_pgoff >= count || user_count > count - vma->vm_pgoff)
 		return -ENXIO;
-- 
2.30.2

[PATCH 2/7] ARM/nommu: use the generic dma-direct code for non-coherent devices

[Christoph Hellwig]

Select the right options to just use the generic dma-direct code
instead of reimplementing it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Dillon Min <dillon.minfei@gmail.com>
---
 arch/arm/Kconfig                |   5 +-
 arch/arm/mm/dma-mapping-nommu.c | 173 ++------------------------------
 2 files changed, 9 insertions(+), 169 deletions(-)

diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index 3ea1c417339f..fcb192bfd35a 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -18,8 +18,8 @@ config ARM
 	select ARCH_HAS_SET_MEMORY
 	select ARCH_HAS_STRICT_KERNEL_RWX if MMU && !XIP_KERNEL
 	select ARCH_HAS_STRICT_MODULE_RWX if MMU
-	select ARCH_HAS_SYNC_DMA_FOR_DEVICE if SWIOTLB
-	select ARCH_HAS_SYNC_DMA_FOR_CPU if SWIOTLB
+	select ARCH_HAS_SYNC_DMA_FOR_DEVICE if SWIOTLB || !MMU
+	select ARCH_HAS_SYNC_DMA_FOR_CPU if SWIOTLB || !MMU
 	select ARCH_HAS_TEARDOWN_DMA_OPS if MMU
 	select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST
 	select ARCH_HAVE_CUSTOM_GPIO_H
@@ -44,6 +44,7 @@ config ARM
 	select CPU_PM if SUSPEND || CPU_IDLE
 	select DCACHE_WORD_ACCESS if HAVE_EFFICIENT_UNALIGNED_ACCESS
 	select DMA_DECLARE_COHERENT
+	select DMA_GLOBAL_POOL if !MMU
 	select DMA_OPS
 	select DMA_REMAP if MMU
 	select EDAC_SUPPORT
diff --git a/arch/arm/mm/dma-mapping-nommu.c b/arch/arm/mm/dma-mapping-nommu.c
index 6bfd2b884505..cfd9c933d2f0 100644
--- a/arch/arm/mm/dma-mapping-nommu.c
+++ b/arch/arm/mm/dma-mapping-nommu.c
@@ -5,12 +5,7 @@
  *  Copyright (C) 2000-2004 Russell King
  */
 
-#include <linux/export.h>
-#include <linux/mm.h>
-#include <linux/dma-direct.h>
 #include <linux/dma-map-ops.h>
-#include <linux/scatterlist.h>
-
 #include <asm/cachetype.h>
 #include <asm/cacheflush.h>
 #include <asm/outercache.h>
@@ -18,65 +13,8 @@
 
 #include "dma.h"
 
-/*
- *  The generic direct mapping code is used if
- *   - MMU/MPU is off
- *   - cpu is v7m w/o cache support
- *   - device is coherent
- *  otherwise arm_nommu_dma_ops is used.
- *
- *  arm_nommu_dma_ops rely on consistent DMA memory (please, refer to
- *  [1] on how to declare such memory).
- *
- *  [1] Documentation/devicetree/bindings/reserved-memory/reserved-memory.txt
- */
-
-static void *arm_nommu_dma_alloc(struct device *dev, size_t size,
-				 dma_addr_t *dma_handle, gfp_t gfp,
-				 unsigned long attrs)
-
-{
-	void *ret = dma_alloc_from_global_coherent(dev, size, dma_handle);
-
-	/*
-	 * dma_alloc_from_global_coherent() may fail because:
-	 *
-	 * - no consistent DMA region has been defined, so we can't
-	 *   continue.
-	 * - there is no space left in consistent DMA region, so we
-	 *   only can fallback to generic allocator if we are
-	 *   advertised that consistency is not required.
-	 */
-
-	WARN_ON_ONCE(ret == NULL);
-	return ret;
-}
-
-static void arm_nommu_dma_free(struct device *dev, size_t size,
-			       void *cpu_addr, dma_addr_t dma_addr,
-			       unsigned long attrs)
-{
-	int ret = dma_release_from_global_coherent(get_order(size), cpu_addr);
-
-	WARN_ON_ONCE(ret == 0);
-}
-
-static int arm_nommu_dma_mmap(struct device *dev, struct vm_area_struct *vma,
-			      void *cpu_addr, dma_addr_t dma_addr, size_t size,
-			      unsigned long attrs)
-{
-	int ret;
-
-	if (dma_mmap_from_global_coherent(vma, cpu_addr, size, &ret))
-		return ret;
-	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
-		return ret;
-	return -ENXIO;
-}
-
-
-static void __dma_page_cpu_to_dev(phys_addr_t paddr, size_t size,
-				  enum dma_data_direction dir)
+void arch_sync_dma_for_device(phys_addr_t paddr, size_t size,
+		enum dma_data_direction dir)
 {
 	dmac_map_area(__va(paddr), size, dir);
 
@@ -86,8 +24,8 @@ static void __dma_page_cpu_to_dev(phys_addr_t paddr, size_t size,
 		outer_clean_range(paddr, paddr + size);
 }
 
-static void __dma_page_dev_to_cpu(phys_addr_t paddr, size_t size,
-				  enum dma_data_direction dir)
+void arch_sync_dma_for_cpu(phys_addr_t paddr, size_t size,
+		enum dma_data_direction dir)
 {
 	if (dir != DMA_TO_DEVICE) {
 		outer_inv_range(paddr, paddr + size);
@@ -95,102 +33,6 @@ static void __dma_page_dev_to_cpu(phys_addr_t paddr, size_t size,
 	}
 }
 
-static dma_addr_t arm_nommu_dma_map_page(struct device *dev, struct page *page,
-					 unsigned long offset, size_t size,
-					 enum dma_data_direction dir,
-					 unsigned long attrs)
-{
-	dma_addr_t handle = page_to_phys(page) + offset;
-
-	__dma_page_cpu_to_dev(handle, size, dir);
-
-	return handle;
-}
-
-static void arm_nommu_dma_unmap_page(struct device *dev, dma_addr_t handle,
-				     size_t size, enum dma_data_direction dir,
-				     unsigned long attrs)
-{
-	__dma_page_dev_to_cpu(handle, size, dir);
-}
-
-
-static int arm_nommu_dma_map_sg(struct device *dev, struct scatterlist *sgl,
-				int nents, enum dma_data_direction dir,
-				unsigned long attrs)
-{
-	int i;
-	struct scatterlist *sg;
-
-	for_each_sg(sgl, sg, nents, i) {
-		sg_dma_address(sg) = sg_phys(sg);
-		sg_dma_len(sg) = sg->length;
-		__dma_page_cpu_to_dev(sg_dma_address(sg), sg_dma_len(sg), dir);
-	}
-
-	return nents;
-}
-
-static void arm_nommu_dma_unmap_sg(struct device *dev, struct scatterlist *sgl,
-				   int nents, enum dma_data_direction dir,
-				   unsigned long attrs)
-{
-	struct scatterlist *sg;
-	int i;
-
-	for_each_sg(sgl, sg, nents, i)
-		__dma_page_dev_to_cpu(sg_dma_address(sg), sg_dma_len(sg), dir);
-}
-
-static void arm_nommu_dma_sync_single_for_device(struct device *dev,
-		dma_addr_t handle, size_t size, enum dma_data_direction dir)
-{
-	__dma_page_cpu_to_dev(handle, size, dir);
-}
-
-static void arm_nommu_dma_sync_single_for_cpu(struct device *dev,
-		dma_addr_t handle, size_t size, enum dma_data_direction dir)
-{
-	__dma_page_cpu_to_dev(handle, size, dir);
-}
-
-static void arm_nommu_dma_sync_sg_for_device(struct device *dev, struct scatterlist *sgl,
-					     int nents, enum dma_data_direction dir)
-{
-	struct scatterlist *sg;
-	int i;
-
-	for_each_sg(sgl, sg, nents, i)
-		__dma_page_cpu_to_dev(sg_dma_address(sg), sg_dma_len(sg), dir);
-}
-
-static void arm_nommu_dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sgl,
-					  int nents, enum dma_data_direction dir)
-{
-	struct scatterlist *sg;
-	int i;
-
-	for_each_sg(sgl, sg, nents, i)
-		__dma_page_dev_to_cpu(sg_dma_address(sg), sg_dma_len(sg), dir);
-}
-
-const struct dma_map_ops arm_nommu_dma_ops = {
-	.alloc			= arm_nommu_dma_alloc,
-	.free			= arm_nommu_dma_free,
-	.alloc_pages		= dma_direct_alloc_pages,
-	.free_pages		= dma_direct_free_pages,
-	.mmap			= arm_nommu_dma_mmap,
-	.map_page		= arm_nommu_dma_map_page,
-	.unmap_page		= arm_nommu_dma_unmap_page,
-	.map_sg			= arm_nommu_dma_map_sg,
-	.unmap_sg		= arm_nommu_dma_unmap_sg,
-	.sync_single_for_device	= arm_nommu_dma_sync_single_for_device,
-	.sync_single_for_cpu	= arm_nommu_dma_sync_single_for_cpu,
-	.sync_sg_for_device	= arm_nommu_dma_sync_sg_for_device,
-	.sync_sg_for_cpu	= arm_nommu_dma_sync_sg_for_cpu,
-};
-EXPORT_SYMBOL(arm_nommu_dma_ops);
-
 void arch_setup_dma_ops(struct device *dev, u64 dma_base, u64 size,
 			const struct iommu_ops *iommu, bool coherent)
 {
@@ -201,14 +43,11 @@ void arch_setup_dma_ops(struct device *dev, u64 dma_base, u64 size,
 		 * enough to check if MPU is in use or not since in absense of
 		 * MPU system memory map is used.
 		 */
-		dev->archdata.dma_coherent = (cacheid) ? coherent : true;
+		dev->dma_coherent = cacheid ? coherent : true;
 	} else {
 		/*
 		 * Assume coherent DMA in case MMU/MPU has not been set up.
 		 */
-		dev->archdata.dma_coherent = (get_cr() & CR_M) ? coherent : true;
+		dev->dma_coherent = (get_cr() & CR_M) ? coherent : true;
 	}
-
-	if (!dev->archdata.dma_coherent)
-		set_dma_ops(dev, &arm_nommu_dma_ops);
 }
-- 
2.30.2

[PATCH 3/7] dma-mapping: allow using the global coherent pool for !ARM

[Christoph Hellwig]

Switch an ifdef so that the global coherent pool is initialized for
any architecture that selects the DMA_GLOBAL_POOL symbol insted of
hardcoding ARM.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Dillon Min <dillon.minfei@gmail.com>
---
 kernel/dma/coherent.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/kernel/dma/coherent.c b/kernel/dma/coherent.c
index 794e76b03b34..67b126afac5a 100644
--- a/kernel/dma/coherent.c
+++ b/kernel/dma/coherent.c
@@ -361,7 +361,9 @@ static int __init rmem_dma_setup(struct reserved_mem *rmem)
 		pr_err("Reserved memory: regions without no-map are not yet supported\n");
 		return -EINVAL;
 	}
+#endif
 
+#ifdef CONFIG_DMA_GLOBAL_POOL
 	if (of_get_flat_dt_prop(node, "linux,dma-default", NULL)) {
 		WARN(dma_reserved_default_memory,
 		     "Reserved memory: region for default DMA coherent area is redefined\n");
-- 
2.30.2

[PATCH 4/7] dma-mapping: simplify dma_init_coherent_memory

[Christoph Hellwig]

Return the allocated dma_coherent_mem structure, set the
use_dma_pfn_offset and print the failure warning inside of
dma_init_coherent_memory instead of leaving that to the callers.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Dillon Min <dillon.minfei@gmail.com>
---
 kernel/dma/coherent.c | 78 ++++++++++++++++++-------------------------
 1 file changed, 33 insertions(+), 45 deletions(-)

diff --git a/kernel/dma/coherent.c b/kernel/dma/coherent.c
index 67b126afac5a..ab397ebfd5ad 100644
--- a/kernel/dma/coherent.c
+++ b/kernel/dma/coherent.c
@@ -37,51 +37,44 @@ static inline dma_addr_t dma_get_device_base(struct device *dev,
 	return mem->device_base;
 }
 
-static int dma_init_coherent_memory(phys_addr_t phys_addr,
-		dma_addr_t device_addr, size_t size,
-		struct dma_coherent_mem **mem)
+static struct dma_coherent_mem *dma_init_coherent_memory(phys_addr_t phys_addr,
+		dma_addr_t device_addr, size_t size, bool use_dma_pfn_offset)
 {
-	struct dma_coherent_mem *dma_mem = NULL;
-	void *mem_base = NULL;
+	struct dma_coherent_mem *dma_mem;
 	int pages = size >> PAGE_SHIFT;
 	int bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
-	int ret;
+	void *mem_base;
 
-	if (!size) {
-		ret = -EINVAL;
-		goto out;
-	}
+	if (!size)
+		return ERR_PTR(-EINVAL);
 
 	mem_base = memremap(phys_addr, size, MEMREMAP_WC);
-	if (!mem_base) {
-		ret = -EINVAL;
-		goto out;
-	}
+	if (!mem_base)
+		return ERR_PTR(-EINVAL);
+
 	dma_mem = kzalloc(sizeof(struct dma_coherent_mem), GFP_KERNEL);
-	if (!dma_mem) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!dma_mem)
+		goto out_unmap_membase;
 	dma_mem->bitmap = kzalloc(bitmap_size, GFP_KERNEL);
-	if (!dma_mem->bitmap) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!dma_mem->bitmap)
+		goto out_free_dma_mem;
 
 	dma_mem->virt_base = mem_base;
 	dma_mem->device_base = device_addr;
 	dma_mem->pfn_base = PFN_DOWN(phys_addr);
 	dma_mem->size = pages;
+	dma_mem->use_dev_dma_pfn_offset = use_dma_pfn_offset;
 	spin_lock_init(&dma_mem->spinlock);
 
-	*mem = dma_mem;
-	return 0;
+	return dma_mem;
 
-out:
+out_free_dma_mem:
 	kfree(dma_mem);
-	if (mem_base)
-		memunmap(mem_base);
-	return ret;
+out_unmap_membase:
+	memunmap(mem_base);
+	pr_err("Reserved memory: failed to init DMA memory pool at %pa, size %zd MiB\n",
+		&phys_addr, size / SZ_1M);
+	return ERR_PTR(-ENOMEM);
 }
 
 static void dma_release_coherent_memory(struct dma_coherent_mem *mem)
@@ -130,9 +123,9 @@ int dma_declare_coherent_memory(struct device *dev, phys_addr_t phys_addr,
 	struct dma_coherent_mem *mem;
 	int ret;
 
-	ret = dma_init_coherent_memory(phys_addr, device_addr, size, &mem);
-	if (ret)
-		return ret;
+	mem = dma_init_coherent_memory(phys_addr, device_addr, size, false);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem);
 
 	ret = dma_assign_coherent_memory(dev, mem);
 	if (ret)
@@ -319,21 +312,16 @@ static struct reserved_mem *dma_reserved_default_memory __initdata;
 
 static int rmem_dma_device_init(struct reserved_mem *rmem, struct device *dev)
 {
-	struct dma_coherent_mem *mem = rmem->priv;
-	int ret;
-
-	if (!mem) {
-		ret = dma_init_coherent_memory(rmem->base, rmem->base,
-					       rmem->size, &mem);
-		if (ret) {
-			pr_err("Reserved memory: failed to init DMA memory pool at %pa, size %ld MiB\n",
-				&rmem->base, (unsigned long)rmem->size / SZ_1M);
-			return ret;
-		}
+	if (!rmem->priv) {
+		struct dma_coherent_mem *mem;
+
+		mem = dma_init_coherent_memory(rmem->base, rmem->base,
+					       rmem->size, true);
+		if (IS_ERR(mem))
+			return PTR_ERR(mem);
+		rmem->priv = mem;
 	}
-	mem->use_dev_dma_pfn_offset = true;
-	rmem->priv = mem;
-	dma_assign_coherent_memory(dev, mem);
+	dma_assign_coherent_memory(dev, rmem->priv);
 	return 0;
 }
 
-- 
2.30.2

[PATCH 5/7] dma-mapping: add a dma_init_global_coherent helper

[Christoph Hellwig]

Add a new helper to initialize the global coherent pool.  This both
cleans up the existing initialization which indirects through the
reserved_mem_ops that are normally only used for struct device, and
also allows using the global pool for non-devicetree architectures.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Dillon Min <dillon.minfei@gmail.com>
---
 include/linux/dma-map-ops.h |  2 +-
 kernel/dma/coherent.c       | 32 ++++++++++++++------------------
 2 files changed, 15 insertions(+), 19 deletions(-)

diff --git a/include/linux/dma-map-ops.h b/include/linux/dma-map-ops.h
index 0d53a96a3d64..7e49bb86a0c0 100644
--- a/include/linux/dma-map-ops.h
+++ b/include/linux/dma-map-ops.h
@@ -176,7 +176,7 @@ void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
 int dma_release_from_global_coherent(int order, void *vaddr);
 int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *cpu_addr,
 		size_t size, int *ret);
-
+int dma_init_global_coherent(phys_addr_t phys_addr, size_t size);
 #else
 static inline int dma_declare_coherent_memory(struct device *dev,
 		phys_addr_t phys_addr, dma_addr_t device_addr, size_t size)
diff --git a/kernel/dma/coherent.c b/kernel/dma/coherent.c
index ab397ebfd5ad..160d4e246ecb 100644
--- a/kernel/dma/coherent.c
+++ b/kernel/dma/coherent.c
@@ -300,6 +300,18 @@ int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *vaddr,
 					vaddr, size, ret);
 }
 
+int dma_init_global_coherent(phys_addr_t phys_addr, size_t size)
+{
+	struct dma_coherent_mem *mem;
+
+	mem = dma_init_coherent_memory(phys_addr, phys_addr, size, true);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem);
+	dma_coherent_default_memory = mem;
+	pr_info("DMA: default coherent area is set\n");
+	return 0;
+}
+
 /*
  * Support for reserved memory regions defined in device tree
  */
@@ -367,26 +379,10 @@ static int __init rmem_dma_setup(struct reserved_mem *rmem)
 
 static int __init dma_init_reserved_memory(void)
 {
-	const struct reserved_mem_ops *ops;
-	int ret;
-
 	if (!dma_reserved_default_memory)
 		return -ENOMEM;
-
-	ops = dma_reserved_default_memory->ops;
-
-	/*
-	 * We rely on rmem_dma_device_init() does not propagate error of
-	 * dma_assign_coherent_memory() for "NULL" device.
-	 */
-	ret = ops->device_init(dma_reserved_default_memory, NULL);
-
-	if (!ret) {
-		dma_coherent_default_memory = dma_reserved_default_memory->priv;
-		pr_info("DMA: default coherent area is set\n");
-	}
-
-	return ret;
+	return dma_init_global_coherent(dma_reserved_default_memory->base,
+					dma_reserved_default_memory->size);
 }
 
 core_initcall(dma_init_reserved_memory);
-- 
2.30.2

[PATCH 6/7] dma-mapping: make the global coherent pool conditional

[Christoph Hellwig]

Only build the code to support the global coherent pool if support for
it is enabled.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Dillon Min <dillon.minfei@gmail.com>
---
 include/linux/dma-map-ops.h | 18 +++++++-------
 kernel/dma/coherent.c       | 47 ++++++++++++++++++++-----------------
 2 files changed, 35 insertions(+), 30 deletions(-)

diff --git a/include/linux/dma-map-ops.h b/include/linux/dma-map-ops.h
index 7e49bb86a0c0..9b79aa4dd300 100644
--- a/include/linux/dma-map-ops.h
+++ b/include/linux/dma-map-ops.h
@@ -170,13 +170,6 @@ int dma_alloc_from_dev_coherent(struct device *dev, ssize_t size,
 int dma_release_from_dev_coherent(struct device *dev, int order, void *vaddr);
 int dma_mmap_from_dev_coherent(struct device *dev, struct vm_area_struct *vma,
 		void *cpu_addr, size_t size, int *ret);
-
-void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
-		dma_addr_t *dma_handle);
-int dma_release_from_global_coherent(int order, void *vaddr);
-int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *cpu_addr,
-		size_t size, int *ret);
-int dma_init_global_coherent(phys_addr_t phys_addr, size_t size);
 #else
 static inline int dma_declare_coherent_memory(struct device *dev,
 		phys_addr_t phys_addr, dma_addr_t device_addr, size_t size)
@@ -186,7 +179,16 @@ static inline int dma_declare_coherent_memory(struct device *dev,
 #define dma_alloc_from_dev_coherent(dev, size, handle, ret) (0)
 #define dma_release_from_dev_coherent(dev, order, vaddr) (0)
 #define dma_mmap_from_dev_coherent(dev, vma, vaddr, order, ret) (0)
+#endif /* CONFIG_DMA_DECLARE_COHERENT */
 
+#ifdef CONFIG_DMA_GLOBAL_POOL
+void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
+		dma_addr_t *dma_handle);
+int dma_release_from_global_coherent(int order, void *vaddr);
+int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *cpu_addr,
+		size_t size, int *ret);
+int dma_init_global_coherent(phys_addr_t phys_addr, size_t size);
+#else
 static inline void *dma_alloc_from_global_coherent(struct device *dev,
 		ssize_t size, dma_addr_t *dma_handle)
 {
@@ -201,7 +203,7 @@ static inline int dma_mmap_from_global_coherent(struct vm_area_struct *vma,
 {
 	return 0;
 }
-#endif /* CONFIG_DMA_DECLARE_COHERENT */
+#endif /* CONFIG_DMA_GLOBAL_POOL */
 
 /*
  * This is the actual return value from the ->alloc_noncontiguous method.
diff --git a/kernel/dma/coherent.c b/kernel/dma/coherent.c
index 160d4e246ecb..c05408902a68 100644
--- a/kernel/dma/coherent.c
+++ b/kernel/dma/coherent.c
@@ -20,8 +20,6 @@ struct dma_coherent_mem {
 	bool		use_dev_dma_pfn_offset;
 };
 
-static struct dma_coherent_mem *dma_coherent_default_memory __ro_after_init;
-
 static inline struct dma_coherent_mem *dev_get_coherent_memory(struct device *dev)
 {
 	if (dev && dev->dma_mem)
@@ -191,16 +189,6 @@ int dma_alloc_from_dev_coherent(struct device *dev, ssize_t size,
 	return 1;
 }
 
-void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
-				     dma_addr_t *dma_handle)
-{
-	if (!dma_coherent_default_memory)
-		return NULL;
-
-	return __dma_alloc_from_coherent(dev, dma_coherent_default_memory, size,
-					 dma_handle);
-}
-
 static int __dma_release_from_coherent(struct dma_coherent_mem *mem,
 				       int order, void *vaddr)
 {
@@ -236,15 +224,6 @@ int dma_release_from_dev_coherent(struct device *dev, int order, void *vaddr)
 	return __dma_release_from_coherent(mem, order, vaddr);
 }
 
-int dma_release_from_global_coherent(int order, void *vaddr)
-{
-	if (!dma_coherent_default_memory)
-		return 0;
-
-	return __dma_release_from_coherent(dma_coherent_default_memory, order,
-			vaddr);
-}
-
 static int __dma_mmap_from_coherent(struct dma_coherent_mem *mem,
 		struct vm_area_struct *vma, void *vaddr, size_t size, int *ret)
 {
@@ -290,6 +269,28 @@ int dma_mmap_from_dev_coherent(struct device *dev, struct vm_area_struct *vma,
 	return __dma_mmap_from_coherent(mem, vma, vaddr, size, ret);
 }
 
+#ifdef CONFIG_DMA_GLOBAL_POOL
+static struct dma_coherent_mem *dma_coherent_default_memory __ro_after_init;
+
+void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
+				     dma_addr_t *dma_handle)
+{
+	if (!dma_coherent_default_memory)
+		return NULL;
+
+	return __dma_alloc_from_coherent(dev, dma_coherent_default_memory, size,
+					 dma_handle);
+}
+
+int dma_release_from_global_coherent(int order, void *vaddr)
+{
+	if (!dma_coherent_default_memory)
+		return 0;
+
+	return __dma_release_from_coherent(dma_coherent_default_memory, order,
+			vaddr);
+}
+
 int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *vaddr,
 				   size_t size, int *ret)
 {
@@ -311,6 +312,7 @@ int dma_init_global_coherent(phys_addr_t phys_addr, size_t size)
 	pr_info("DMA: default coherent area is set\n");
 	return 0;
 }
+#endif /* CONFIG_DMA_GLOBAL_POOL */
 
 /*
  * Support for reserved memory regions defined in device tree
@@ -377,6 +379,7 @@ static int __init rmem_dma_setup(struct reserved_mem *rmem)
 	return 0;
 }
 
+#ifdef CONFIG_DMA_GLOBAL_POOL
 static int __init dma_init_reserved_memory(void)
 {
 	if (!dma_reserved_default_memory)
@@ -384,8 +387,8 @@ static int __init dma_init_reserved_memory(void)
 	return dma_init_global_coherent(dma_reserved_default_memory->base,
 					dma_reserved_default_memory->size);
 }
-
 core_initcall(dma_init_reserved_memory);
+#endif /* CONFIG_DMA_GLOBAL_POOL */
 
 RESERVEDMEM_OF_DECLARE(dma, "shared-dma-pool", rmem_dma_setup);
 #endif
-- 
2.30.2

Re: [PATCH 6/7] dma-mapping: make the global coherent pool conditional

[Atish Patra]

On Sun, Jul 11, 2021 at 11:25 PM Christoph Hellwig <hch@lst.de> wrote:
>
> Only build the code to support the global coherent pool if support for
> it is enabled.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> Tested-by: Dillon Min <dillon.minfei@gmail.com>
> ---
>  include/linux/dma-map-ops.h | 18 +++++++-------
>  kernel/dma/coherent.c       | 47 ++++++++++++++++++++-----------------
>  2 files changed, 35 insertions(+), 30 deletions(-)
>
> diff --git a/include/linux/dma-map-ops.h b/include/linux/dma-map-ops.h
> index 7e49bb86a0c0..9b79aa4dd300 100644
> --- a/include/linux/dma-map-ops.h
> +++ b/include/linux/dma-map-ops.h
> @@ -170,13 +170,6 @@ int dma_alloc_from_dev_coherent(struct device *dev, ssize_t size,
>  int dma_release_from_dev_coherent(struct device *dev, int order, void *vaddr);
>  int dma_mmap_from_dev_coherent(struct device *dev, struct vm_area_struct *vma,
>                 void *cpu_addr, size_t size, int *ret);
> -
> -void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
> -               dma_addr_t *dma_handle);
> -int dma_release_from_global_coherent(int order, void *vaddr);
> -int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *cpu_addr,
> -               size_t size, int *ret);
> -int dma_init_global_coherent(phys_addr_t phys_addr, size_t size);
>  #else
>  static inline int dma_declare_coherent_memory(struct device *dev,
>                 phys_addr_t phys_addr, dma_addr_t device_addr, size_t size)
> @@ -186,7 +179,16 @@ static inline int dma_declare_coherent_memory(struct device *dev,
>  #define dma_alloc_from_dev_coherent(dev, size, handle, ret) (0)
>  #define dma_release_from_dev_coherent(dev, order, vaddr) (0)
>  #define dma_mmap_from_dev_coherent(dev, vma, vaddr, order, ret) (0)
> +#endif /* CONFIG_DMA_DECLARE_COHERENT */
>
> +#ifdef CONFIG_DMA_GLOBAL_POOL
> +void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
> +               dma_addr_t *dma_handle);
> +int dma_release_from_global_coherent(int order, void *vaddr);
> +int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *cpu_addr,
> +               size_t size, int *ret);
> +int dma_init_global_coherent(phys_addr_t phys_addr, size_t size);
> +#else
>  static inline void *dma_alloc_from_global_coherent(struct device *dev,
>                 ssize_t size, dma_addr_t *dma_handle)
>  {
> @@ -201,7 +203,7 @@ static inline int dma_mmap_from_global_coherent(struct vm_area_struct *vma,
>  {
>         return 0;
>  }
> -#endif /* CONFIG_DMA_DECLARE_COHERENT */
> +#endif /* CONFIG_DMA_GLOBAL_POOL */
>
>  /*
>   * This is the actual return value from the ->alloc_noncontiguous method.
> diff --git a/kernel/dma/coherent.c b/kernel/dma/coherent.c
> index 160d4e246ecb..c05408902a68 100644
> --- a/kernel/dma/coherent.c
> +++ b/kernel/dma/coherent.c
> @@ -20,8 +20,6 @@ struct dma_coherent_mem {
>         bool            use_dev_dma_pfn_offset;
>  };
>
> -static struct dma_coherent_mem *dma_coherent_default_memory __ro_after_init;
> -
>  static inline struct dma_coherent_mem *dev_get_coherent_memory(struct device *dev)
>  {
>         if (dev && dev->dma_mem)
> @@ -191,16 +189,6 @@ int dma_alloc_from_dev_coherent(struct device *dev, ssize_t size,
>         return 1;
>  }
>
> -void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
> -                                    dma_addr_t *dma_handle)
> -{
> -       if (!dma_coherent_default_memory)
> -               return NULL;
> -
> -       return __dma_alloc_from_coherent(dev, dma_coherent_default_memory, size,
> -                                        dma_handle);
> -}
> -
>  static int __dma_release_from_coherent(struct dma_coherent_mem *mem,
>                                        int order, void *vaddr)
>  {
> @@ -236,15 +224,6 @@ int dma_release_from_dev_coherent(struct device *dev, int order, void *vaddr)
>         return __dma_release_from_coherent(mem, order, vaddr);
>  }
>
> -int dma_release_from_global_coherent(int order, void *vaddr)
> -{
> -       if (!dma_coherent_default_memory)
> -               return 0;
> -
> -       return __dma_release_from_coherent(dma_coherent_default_memory, order,
> -                       vaddr);
> -}
> -
>  static int __dma_mmap_from_coherent(struct dma_coherent_mem *mem,
>                 struct vm_area_struct *vma, void *vaddr, size_t size, int *ret)
>  {
> @@ -290,6 +269,28 @@ int dma_mmap_from_dev_coherent(struct device *dev, struct vm_area_struct *vma,
>         return __dma_mmap_from_coherent(mem, vma, vaddr, size, ret);
>  }
>
> +#ifdef CONFIG_DMA_GLOBAL_POOL
> +static struct dma_coherent_mem *dma_coherent_default_memory __ro_after_init;
> +
> +void *dma_alloc_from_global_coherent(struct device *dev, ssize_t size,
> +                                    dma_addr_t *dma_handle)
> +{
> +       if (!dma_coherent_default_memory)
> +               return NULL;
> +
> +       return __dma_alloc_from_coherent(dev, dma_coherent_default_memory, size,
> +                                        dma_handle);
> +}
> +
> +int dma_release_from_global_coherent(int order, void *vaddr)
> +{
> +       if (!dma_coherent_default_memory)
> +               return 0;
> +
> +       return __dma_release_from_coherent(dma_coherent_default_memory, order,
> +                       vaddr);
> +}
> +
>  int dma_mmap_from_global_coherent(struct vm_area_struct *vma, void *vaddr,
>                                    size_t size, int *ret)
>  {
> @@ -311,6 +312,7 @@ int dma_init_global_coherent(phys_addr_t phys_addr, size_t size)
>         pr_info("DMA: default coherent area is set\n");
>         return 0;
>  }
> +#endif /* CONFIG_DMA_GLOBAL_POOL */
>
>  /*
>   * Support for reserved memory regions defined in device tree
> @@ -377,6 +379,7 @@ static int __init rmem_dma_setup(struct reserved_mem *rmem)
>         return 0;
>  }
>
> +#ifdef CONFIG_DMA_GLOBAL_POOL
>  static int __init dma_init_reserved_memory(void)
>  {
>         if (!dma_reserved_default_memory)
> @@ -384,8 +387,8 @@ static int __init dma_init_reserved_memory(void)
>         return dma_init_global_coherent(dma_reserved_default_memory->base,
>                                         dma_reserved_default_memory->size);

dma_reserved_default_memory also needs to be moved inside
CONFIG_DMA_GLOBAL_POOL because
all the usages are already part of that config.


>  }
> -
>  core_initcall(dma_init_reserved_memory);
> +#endif /* CONFIG_DMA_GLOBAL_POOL */
>
>  RESERVEDMEM_OF_DECLARE(dma, "shared-dma-pool", rmem_dma_setup);
>  #endif
> --
> 2.30.2
>
> _______________________________________________
> iommu mailing list
> iommu@lists.linux-foundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/iommu



--
Regards,
Atish

[PATCH 7/7] hexagon: use the generic global coherent pool

[Christoph Hellwig]

Switch hexagon to use the generic code for dma_alloc_coherent from
a global pre-filled pool.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 arch/hexagon/Kconfig      |  1 +
 arch/hexagon/kernel/dma.c | 57 ++++++++-------------------------------
 2 files changed, 12 insertions(+), 46 deletions(-)

diff --git a/arch/hexagon/Kconfig b/arch/hexagon/Kconfig
index e5a852080730..aab1a40eb653 100644
--- a/arch/hexagon/Kconfig
+++ b/arch/hexagon/Kconfig
@@ -7,6 +7,7 @@ config HEXAGON
 	select ARCH_32BIT_OFF_T
 	select ARCH_HAS_SYNC_DMA_FOR_DEVICE
 	select ARCH_NO_PREEMPT
+	select DMA_GLOBAL_POOL
 	# Other pending projects/to-do items.
 	# select HAVE_REGS_AND_STACK_ACCESS_API
 	# select HAVE_HW_BREAKPOINT if PERF_EVENTS
diff --git a/arch/hexagon/kernel/dma.c b/arch/hexagon/kernel/dma.c
index 00b9a81075dd..882680e81a30 100644
--- a/arch/hexagon/kernel/dma.c
+++ b/arch/hexagon/kernel/dma.c
@@ -7,54 +7,8 @@
 
 #include <linux/dma-map-ops.h>
 #include <linux/memblock.h>
-#include <linux/genalloc.h>
-#include <linux/module.h>
 #include <asm/page.h>
 
-static struct gen_pool *coherent_pool;
-
-
-/* Allocates from a pool of uncached memory that was reserved at boot time */
-
-void *arch_dma_alloc(struct device *dev, size_t size, dma_addr_t *dma_addr,
-		gfp_t flag, unsigned long attrs)
-{
-	void *ret;
-
-	/*
-	 * Our max_low_pfn should have been backed off by 16MB in
-	 * mm/init.c to create DMA coherent space.  Use that as the VA
-	 * for the pool.
-	 */
-
-	if (coherent_pool == NULL) {
-		coherent_pool = gen_pool_create(PAGE_SHIFT, -1);
-
-		if (coherent_pool == NULL)
-			panic("Can't create %s() memory pool!", __func__);
-		else
-			gen_pool_add(coherent_pool,
-				(unsigned long)pfn_to_virt(max_low_pfn),
-				hexagon_coherent_pool_size, -1);
-	}
-
-	ret = (void *) gen_pool_alloc(coherent_pool, size);
-
-	if (ret) {
-		memset(ret, 0, size);
-		*dma_addr = (dma_addr_t) virt_to_phys(ret);
-	} else
-		*dma_addr = ~0;
-
-	return ret;
-}
-
-void arch_dma_free(struct device *dev, size_t size, void *vaddr,
-		dma_addr_t dma_addr, unsigned long attrs)
-{
-	gen_pool_free(coherent_pool, (unsigned long) vaddr, size);
-}
-
 void arch_sync_dma_for_device(phys_addr_t paddr, size_t size,
 		enum dma_data_direction dir)
 {
@@ -77,3 +31,14 @@ void arch_sync_dma_for_device(phys_addr_t paddr, size_t size,
 		BUG();
 	}
 }
+
+/*
+ * Our max_low_pfn should have been backed off by 16MB in mm/init.c to create
+ * DMA coherent space.  Use that for the pool.
+ */
+static int __init hexagon_dma_init(void)
+{
+	return dma_init_global_coherent(PFN_PHYS(max_low_pfn),
+					hexagon_coherent_pool_size);
+}
+core_initcall(hexagon_dma_init);
-- 
2.30.2

RE: [PATCH 7/7] hexagon: use the generic global coherent pool

[Brian Cain]

> -----Original Message-----
> From: Christoph Hellwig <hch@lst.de>
...
> Switch hexagon to use the generic code for dma_alloc_coherent from
> a global pre-filled pool.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Reviewed-by: Brian Cain <bcain@codeaurora.org>

Re: [PATCH 7/7] hexagon: use the generic global coherent pool

['Christoph Hellwig']

Thanks,

I've pulled the whole series into the dma-mapping for-next tree.

RE: add support for the global coherent pool to the dma core

[Brian Cain]

> -----Original Message-----
> From: Christoph Hellwig <hch@lst.de>
> Sent: Monday, July 12, 2021 1:17 AM
> To: iommu@lists.linux-foundation.org; Russell King
<linux@armlinux.org.uk>;
> Brian Cain <bcain@codeaurora.org>
> Cc: Dillon Min <dillon.minfei@gmail.com>; Vladimir Murzin
> <vladimir.murzin@arm.com>; linux-arm-kernel@lists.infradead.org; linux-
> hexagon@vger.kernel.org; linux-kernel@vger.kernel.org
> Subject: add support for the global coherent pool to the dma core
> 
> Hi all,
> 
> this series adds support for using the global coherent (aka uncached)
> pool to the generic dma-direct code and then switches arm-nommu and
> hexagon over to it, together with a bunch of cleanups.

Christoph,

Thanks for sending this -- I will take a look and give some feedback soon.

> Diffstat:
>  arch/arm/Kconfig                |    5 -
>  arch/arm/mm/dma-mapping-nommu.c |  173
+--------------------------------------
> -
>  arch/hexagon/Kconfig            |    1
>  arch/hexagon/kernel/dma.c       |   57 ++-----------
>  include/linux/dma-map-ops.h     |   18 ++--
>  kernel/dma/Kconfig              |    4
>  kernel/dma/coherent.c           |  159
+++++++++++++++++-------------------
>  kernel/dma/direct.c             |   15 +++
>  8 files changed, 124 insertions(+), 308 deletions(-)

Re: add support for the global coherent pool to the dma core

['Christoph Hellwig']

On Wed, Jul 28, 2021 at 10:20:34AM -0500, Brian Cain wrote:
> > this series adds support for using the global coherent (aka uncached)
> > pool to the generic dma-direct code and then switches arm-nommu and
> > hexagon over to it, together with a bunch of cleanups.
> 
> Christoph,
> 
> Thanks for sending this -- I will take a look and give some feedback soon.

Any chance you'll get to this in the next days?  I'd love to merge this
series for the current merge window with a little time in linux-next.

Re: add support for the global coherent pool to the dma core

[Christoph Hellwig]

Any more comments?

Otherwise I'm going to pull this into the dma-mapping tree next week.