[PATCH v5 0/8] Introduce automatic DMA configuration for IOMMU masters

[PATCH v5 0/8] Introduce automatic DMA configuration for IOMMU masters

[Will Deacon]

Hi all,

Here is v5 of the patches I've previously sent here:

  RFCv1: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/283023.html
  RFCv2: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-September/283752.html
  RFCv3: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-September/287031.html
  RFCv4: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/302711.html

Changes since RFCv4 include:

  - Dropped the RFC tag, since has been used by a couple of people now

  - Dropped the DMA segment configuration from of_dma_configure, as there
    appear to be assumptions about 64k segments elsewhere in the kernel

  - Added acks/tested-bys (thanks to everybody who reviewed the series)

  - A few small fixes for issues found by Marek

Arnd: Is this too late for 3.19? We could merge the first 6 patches
with no issues, since there aren't any callers of of_iommu_init
without patch 7 anyway.

Up to you.

Will

--->8

Marek Szyprowski (1):
  iommu: fix initialization without 'add_device' callback

Will Deacon (7):
  iommu: provide early initialisation hook for IOMMU drivers
  dma-mapping: replace set_arch_dma_coherent_ops with arch_setup_dma_ops
  iommu: add new iommu_ops callback for adding an OF device
  iommu: provide helper function to configure an IOMMU for an of master
  dma-mapping: detect and configure IOMMU in of_dma_configure
  arm: call iommu_init before of_platform_populate
  arm: dma-mapping: plumb our iommu mapping ops into arch_setup_dma_ops

 arch/arm/include/asm/dma-mapping.h | 12 +++---
 arch/arm/kernel/setup.c            |  2 +
 arch/arm/mm/dma-mapping.c          | 83 ++++++++++++++++++++++++++++++++++----
 drivers/iommu/Kconfig              |  2 +-
 drivers/iommu/iommu.c              |  2 +-
 drivers/iommu/of_iommu.c           | 50 +++++++++++++++++++++++
 drivers/of/platform.c              | 50 ++++++++++++-----------
 include/asm-generic/vmlinux.lds.h  |  2 +
 include/linux/dma-mapping.h        | 13 +++---
 include/linux/iommu.h              |  8 ++++
 include/linux/of_iommu.h           | 31 ++++++++++++++
 11 files changed, 211 insertions(+), 44 deletions(-)

-- 
2.1.1

[PATCH v5 1/8] iommu: provide early initialisation hook for IOMMU drivers

[Will Deacon]

IOMMU drivers must be initialised before any of their upstream devices,
otherwise the relevant iommu_ops won't be configured for the bus in
question. To solve this, a number of IOMMU drivers use initcalls to
initialise the driver before anything has a chance to be probed.

Whilst this solves the immediate problem, it leaves the job of probing
the IOMMU completely separate from the iommu_ops to configure the IOMMU,
which are called on a per-bus basis and require the driver to figure out
exactly which instance of the IOMMU is being requested. In particular,
the add_device callback simply passes a struct device to the driver,
which then has to parse firmware tables or probe buses to identify the
relevant IOMMU instance.

This patch takes the first step in addressing this problem by adding an
early initialisation pass for IOMMU drivers, giving them the ability to
store some per-instance data in their iommu_ops structure and store that
in their of_node. This can later be used when parsing OF masters to
identify the IOMMU instance in question.

Acked-by: Joerg Roedel <jroedel@suse.de>
Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Tested-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
 drivers/iommu/of_iommu.c          | 17 +++++++++++++++++
 include/asm-generic/vmlinux.lds.h |  2 ++
 include/linux/iommu.h             |  2 ++
 include/linux/of_iommu.h          | 25 +++++++++++++++++++++++++
 4 files changed, 46 insertions(+)

diff --git a/drivers/iommu/of_iommu.c b/drivers/iommu/of_iommu.c
index e550ccb7634e..89b903406968 100644
--- a/drivers/iommu/of_iommu.c
+++ b/drivers/iommu/of_iommu.c
@@ -22,6 +22,9 @@
 #include <linux/of.h>
 #include <linux/of_iommu.h>
 
+static const struct of_device_id __iommu_of_table_sentinel
+	__used __section(__iommu_of_table_end);
+
 /**
  * of_get_dma_window - Parse *dma-window property and returns 0 if found.
  *
@@ -89,3 +92,17 @@ int of_get_dma_window(struct device_node *dn, const char *prefix, int index,
 	return 0;
 }
 EXPORT_SYMBOL_GPL(of_get_dma_window);
+
+void __init of_iommu_init(void)
+{
+	struct device_node *np;
+	const struct of_device_id *match, *matches = &__iommu_of_table;
+
+	for_each_matching_node_and_match(np, matches, &match) {
+		const of_iommu_init_fn init_fn = match->data;
+
+		if (init_fn(np))
+			pr_err("Failed to initialise IOMMU %s\n",
+				of_node_full_name(np));
+	}
+}
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index aa70cbda327c..bee5d683074d 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -164,6 +164,7 @@
 #define CLKSRC_OF_TABLES()	OF_TABLE(CONFIG_CLKSRC_OF, clksrc)
 #define IRQCHIP_OF_MATCH_TABLE() OF_TABLE(CONFIG_IRQCHIP, irqchip)
 #define CLK_OF_TABLES()		OF_TABLE(CONFIG_COMMON_CLK, clk)
+#define IOMMU_OF_TABLES()	OF_TABLE(CONFIG_OF_IOMMU, iommu)
 #define RESERVEDMEM_OF_TABLES()	OF_TABLE(CONFIG_OF_RESERVED_MEM, reservedmem)
 #define CPU_METHOD_OF_TABLES()	OF_TABLE(CONFIG_SMP, cpu_method)
 #define EARLYCON_OF_TABLES()	OF_TABLE(CONFIG_SERIAL_EARLYCON, earlycon)
@@ -497,6 +498,7 @@
 	CLK_OF_TABLES()							\
 	RESERVEDMEM_OF_TABLES()						\
 	CLKSRC_OF_TABLES()						\
+	IOMMU_OF_TABLES()						\
 	CPU_METHOD_OF_TABLES()						\
 	KERNEL_DTB()							\
 	IRQCHIP_OF_MATCH_TABLE()					\
diff --git a/include/linux/iommu.h b/include/linux/iommu.h
index e6a7c9ff72f2..7b83f9f8e11d 100644
--- a/include/linux/iommu.h
+++ b/include/linux/iommu.h
@@ -103,6 +103,7 @@ enum iommu_attr {
  * @domain_get_attr: Query domain attributes
  * @domain_set_attr: Change domain attributes
  * @pgsize_bitmap: bitmap of supported page sizes
+ * @priv: per-instance data private to the iommu driver
  */
 struct iommu_ops {
 	bool (*capable)(enum iommu_cap);
@@ -133,6 +134,7 @@ struct iommu_ops {
 	u32 (*domain_get_windows)(struct iommu_domain *domain);
 
 	unsigned long pgsize_bitmap;
+	void *priv;
 };
 
 #define IOMMU_GROUP_NOTIFY_ADD_DEVICE		1 /* Device added */
diff --git a/include/linux/of_iommu.h b/include/linux/of_iommu.h
index 51a560f34bca..5762cdc8effe 100644
--- a/include/linux/of_iommu.h
+++ b/include/linux/of_iommu.h
@@ -1,12 +1,17 @@
 #ifndef __OF_IOMMU_H
 #define __OF_IOMMU_H
 
+#include <linux/iommu.h>
+#include <linux/of.h>
+
 #ifdef CONFIG_OF_IOMMU
 
 extern int of_get_dma_window(struct device_node *dn, const char *prefix,
 			     int index, unsigned long *busno, dma_addr_t *addr,
 			     size_t *size);
 
+extern void of_iommu_init(void);
+
 #else
 
 static inline int of_get_dma_window(struct device_node *dn, const char *prefix,
@@ -16,6 +21,26 @@ static inline int of_get_dma_window(struct device_node *dn, const char *prefix,
 	return -EINVAL;
 }
 
+static inline void of_iommu_init(void) { }
+
 #endif	/* CONFIG_OF_IOMMU */
 
+static inline void of_iommu_set_ops(struct device_node *np,
+				    const struct iommu_ops *ops)
+{
+	np->data = (struct iommu_ops *)ops;
+}
+
+static inline struct iommu_ops *of_iommu_get_ops(struct device_node *np)
+{
+	return np->data;
+}
+
+extern struct of_device_id __iommu_of_table;
+
+typedef int (*of_iommu_init_fn)(struct device_node *);
+
+#define IOMMU_OF_DECLARE(name, compat, fn) \
+	_OF_DECLARE(iommu, name, compat, fn, of_iommu_init_fn)
+
 #endif /* __OF_IOMMU_H */
-- 
2.1.1

[PATCH v5 2/8] dma-mapping: replace set_arch_dma_coherent_ops with arch_setup_dma_ops

[Will Deacon]

set_arch_dma_coherent_ops is called from of_dma_configure in order to
swizzle the architectural dma-mapping functions over to a cache-coherent
implementation. This is currently implemented only for ARM.

In anticipation of re-using this mechanism for IOMMU-backed dma-mapping
ops too, this patch replaces the function with a broader
arch_setup_dma_ops callback which will be extended in future.

Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Tested-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
 arch/arm/include/asm/dma-mapping.h |  8 ++++----
 drivers/of/platform.c              | 31 +++++++++++++------------------
 include/linux/dma-mapping.h        |  7 ++-----
 3 files changed, 19 insertions(+), 27 deletions(-)

diff --git a/arch/arm/include/asm/dma-mapping.h b/arch/arm/include/asm/dma-mapping.h
index 85738b200023..dc3420e77758 100644
--- a/arch/arm/include/asm/dma-mapping.h
+++ b/arch/arm/include/asm/dma-mapping.h
@@ -121,12 +121,12 @@ static inline unsigned long dma_max_pfn(struct device *dev)
 }
 #define dma_max_pfn(dev) dma_max_pfn(dev)
 
-static inline int set_arch_dma_coherent_ops(struct device *dev)
+static inline void arch_setup_dma_ops(struct device *dev, bool coherent)
 {
-	set_dma_ops(dev, &arm_coherent_dma_ops);
-	return 0;
+	if (coherent)
+		set_dma_ops(dev, &arm_coherent_dma_ops);
 }
-#define set_arch_dma_coherent_ops(dev)	set_arch_dma_coherent_ops(dev)
+#define arch_setup_dma_ops arch_setup_dma_ops
 
 static inline dma_addr_t phys_to_dma(struct device *dev, phys_addr_t paddr)
 {
diff --git a/drivers/of/platform.c b/drivers/of/platform.c
index 3b64d0bf5bba..ff1f4e9afccb 100644
--- a/drivers/of/platform.c
+++ b/drivers/of/platform.c
@@ -164,6 +164,8 @@ static void of_dma_configure(struct device *dev)
 {
 	u64 dma_addr, paddr, size;
 	int ret;
+	bool coherent;
+	unsigned long offset;
 
 	/*
 	 * Set default dma-mask to 32 bit. Drivers are expected to setup
@@ -178,28 +180,21 @@ static void of_dma_configure(struct device *dev)
 	if (!dev->dma_mask)
 		dev->dma_mask = &dev->coherent_dma_mask;
 
-	/*
-	 * if dma-coherent property exist, call arch hook to setup
-	 * dma coherent operations.
-	 */
-	if (of_dma_is_coherent(dev->of_node)) {
-		set_arch_dma_coherent_ops(dev);
-		dev_dbg(dev, "device is dma coherent\n");
-	}
-
-	/*
-	 * if dma-ranges property doesn't exist - just return else
-	 * setup the dma offset
-	 */
 	ret = of_dma_get_range(dev->of_node, &dma_addr, &paddr, &size);
 	if (ret < 0) {
-		dev_dbg(dev, "no dma range information to setup\n");
-		return;
+		dma_addr = offset = 0;
+		size = dev->coherent_dma_mask;
+	} else {
+		offset = PFN_DOWN(paddr - dma_addr);
+		dev_dbg(dev, "dma_pfn_offset(%#08lx)\n", dev->dma_pfn_offset);
 	}
+	dev->dma_pfn_offset = offset;
+
+	coherent = of_dma_is_coherent(dev->of_node);
+	dev_dbg(dev, "device is%sdma coherent\n",
+		coherent ? " " : " not ");
 
-	/* DMA ranges found. Calculate and set dma_pfn_offset */
-	dev->dma_pfn_offset = PFN_DOWN(paddr - dma_addr);
-	dev_dbg(dev, "dma_pfn_offset(%#08lx)\n", dev->dma_pfn_offset);
+	arch_setup_dma_ops(dev, coherent);
 }
 
 /**
diff --git a/include/linux/dma-mapping.h b/include/linux/dma-mapping.h
index d5d388160f42..8a1560f95d4a 100644
--- a/include/linux/dma-mapping.h
+++ b/include/linux/dma-mapping.h
@@ -129,11 +129,8 @@ static inline int dma_coerce_mask_and_coherent(struct device *dev, u64 mask)
 
 extern u64 dma_get_required_mask(struct device *dev);
 
-#ifndef set_arch_dma_coherent_ops
-static inline int set_arch_dma_coherent_ops(struct device *dev)
-{
-	return 0;
-}
+#ifndef arch_setup_dma_ops
+static inline void arch_setup_dma_ops(struct device *dev, bool coherent) { }
 #endif
 
 static inline unsigned int dma_get_max_seg_size(struct device *dev)
-- 
2.1.1

[PATCH v5 3/8] iommu: add new iommu_ops callback for adding an OF device

[Will Deacon]

This patch adds a new function to the iommu_ops structure to allow an
OF device to be added to a specific IOMMU instance using the recently
merged generic devicetree binding for IOMMUs. The callback (of_xlate)
takes a struct device representing the master and an of_phandle_args
representing the IOMMU and the correspondong IDs for the new master.

Acked-by: Joerg Roedel <jroedel@suse.de>
Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Tested-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
 include/linux/iommu.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/include/linux/iommu.h b/include/linux/iommu.h
index 7b83f9f8e11d..415c7613d02c 100644
--- a/include/linux/iommu.h
+++ b/include/linux/iommu.h
@@ -21,6 +21,7 @@
 
 #include <linux/errno.h>
 #include <linux/err.h>
+#include <linux/of.h>
 #include <linux/types.h>
 #include <trace/events/iommu.h>
 
@@ -102,6 +103,7 @@ enum iommu_attr {
  * @remove_device: remove device from iommu grouping
  * @domain_get_attr: Query domain attributes
  * @domain_set_attr: Change domain attributes
+ * @of_xlate: add OF master IDs to iommu grouping
  * @pgsize_bitmap: bitmap of supported page sizes
  * @priv: per-instance data private to the iommu driver
  */
@@ -133,6 +135,10 @@ struct iommu_ops {
 	/* Get the numer of window per domain */
 	u32 (*domain_get_windows)(struct iommu_domain *domain);
 
+#ifdef CONFIG_OF_IOMMU
+	int (*of_xlate)(struct device *dev, struct of_phandle_args *args);
+#endif
+
 	unsigned long pgsize_bitmap;
 	void *priv;
 };
-- 
2.1.1

[PATCH v5 4/8] iommu: provide helper function to configure an IOMMU for an of master

[Will Deacon]

The generic IOMMU device-tree bindings can be used to add arbitrary OF
masters to an IOMMU with a compliant binding.

This patch introduces of_iommu_configure, which does exactly that.

Acked-by: Joerg Roedel <jroedel@suse.de>
Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Tested-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
 drivers/iommu/Kconfig    |  2 +-
 drivers/iommu/of_iommu.c | 33 +++++++++++++++++++++++++++++++++
 include/linux/of_iommu.h |  6 ++++++
 3 files changed, 40 insertions(+), 1 deletion(-)

diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
index dd5112265cc9..6d13f962f156 100644
--- a/drivers/iommu/Kconfig
+++ b/drivers/iommu/Kconfig
@@ -15,7 +15,7 @@ if IOMMU_SUPPORT
 
 config OF_IOMMU
        def_bool y
-       depends on OF
+       depends on OF && IOMMU_API
 
 config FSL_PAMU
 	bool "Freescale IOMMU support"
diff --git a/drivers/iommu/of_iommu.c b/drivers/iommu/of_iommu.c
index 89b903406968..73236d3cc955 100644
--- a/drivers/iommu/of_iommu.c
+++ b/drivers/iommu/of_iommu.c
@@ -18,6 +18,7 @@
  */
 
 #include <linux/export.h>
+#include <linux/iommu.h>
 #include <linux/limits.h>
 #include <linux/of.h>
 #include <linux/of_iommu.h>
@@ -93,6 +94,38 @@ int of_get_dma_window(struct device_node *dn, const char *prefix, int index,
 }
 EXPORT_SYMBOL_GPL(of_get_dma_window);
 
+struct iommu_ops *of_iommu_configure(struct device *dev)
+{
+	struct of_phandle_args iommu_spec;
+	struct device_node *np;
+	struct iommu_ops *ops = NULL;
+	int idx = 0;
+
+	/*
+	 * We don't currently walk up the tree looking for a parent IOMMU.
+	 * See the `Notes:' section of
+	 * Documentation/devicetree/bindings/iommu/iommu.txt
+	 */
+	while (!of_parse_phandle_with_args(dev->of_node, "iommus",
+					   "#iommu-cells", idx,
+					   &iommu_spec)) {
+		np = iommu_spec.np;
+		ops = of_iommu_get_ops(np);
+
+		if (!ops || !ops->of_xlate || ops->of_xlate(dev, &iommu_spec))
+			goto err_put_node;
+
+		of_node_put(np);
+		idx++;
+	}
+
+	return ops;
+
+err_put_node:
+	of_node_put(np);
+	return NULL;
+}
+
 void __init of_iommu_init(void)
 {
 	struct device_node *np;
diff --git a/include/linux/of_iommu.h b/include/linux/of_iommu.h
index 5762cdc8effe..d03abbb11c34 100644
--- a/include/linux/of_iommu.h
+++ b/include/linux/of_iommu.h
@@ -1,6 +1,7 @@
 #ifndef __OF_IOMMU_H
 #define __OF_IOMMU_H
 
+#include <linux/device.h>
 #include <linux/iommu.h>
 #include <linux/of.h>
 
@@ -11,6 +12,7 @@ extern int of_get_dma_window(struct device_node *dn, const char *prefix,
 			     size_t *size);
 
 extern void of_iommu_init(void);
+extern struct iommu_ops *of_iommu_configure(struct device *dev);
 
 #else
 
@@ -22,6 +24,10 @@ static inline int of_get_dma_window(struct device_node *dn, const char *prefix,
 }
 
 static inline void of_iommu_init(void) { }
+static inline struct iommu_ops *of_iommu_configure(struct device *dev)
+{
+	return NULL;
+}
 
 #endif	/* CONFIG_OF_IOMMU */
 
-- 
2.1.1

[PATCH v5 5/8] iommu: fix initialization without 'add_device' callback

[Will Deacon]

From: Marek Szyprowski <m.szyprowski@samsung.com>

IOMMU drivers can be initialized from of_iommu helpers. Such drivers don't
need to provide device_add callbacks to operate properly, so there is no
need to fail initialization if the callback is missing.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
 drivers/iommu/iommu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
index ed8b04867b1f..02f798b7e295 100644
--- a/drivers/iommu/iommu.c
+++ b/drivers/iommu/iommu.c
@@ -737,7 +737,7 @@ static int add_iommu_group(struct device *dev, void *data)
 	const struct iommu_ops *ops = cb->ops;
 
 	if (!ops->add_device)
-		return -ENODEV;
+		return 0;
 
 	WARN_ON(dev->iommu_group);
 
-- 
2.1.1

[PATCH v5 6/8] dma-mapping: detect and configure IOMMU in of_dma_configure

[Will Deacon]

This patch extends of_dma_configure so that it sets up the IOMMU for a
device, as well as the coherent/non-coherent DMA mapping ops.

Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Tested-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
 arch/arm/include/asm/dma-mapping.h |  4 +++-
 drivers/of/platform.c              | 21 ++++++++++++++-------
 include/linux/dma-mapping.h        |  8 +++++++-
 3 files changed, 24 insertions(+), 9 deletions(-)

diff --git a/arch/arm/include/asm/dma-mapping.h b/arch/arm/include/asm/dma-mapping.h
index dc3420e77758..f3c0d953f6a2 100644
--- a/arch/arm/include/asm/dma-mapping.h
+++ b/arch/arm/include/asm/dma-mapping.h
@@ -121,7 +121,9 @@ static inline unsigned long dma_max_pfn(struct device *dev)
 }
 #define dma_max_pfn(dev) dma_max_pfn(dev)
 
-static inline void arch_setup_dma_ops(struct device *dev, bool coherent)
+static inline void arch_setup_dma_ops(struct device *dev, u64 dma_base,
+				      u64 size, struct iommu_ops *iommu,
+				      bool coherent)
 {
 	if (coherent)
 		set_dma_ops(dev, &arm_coherent_dma_ops);
diff --git a/drivers/of/platform.c b/drivers/of/platform.c
index ff1f4e9afccb..b89caf8c7586 100644
--- a/drivers/of/platform.c
+++ b/drivers/of/platform.c
@@ -19,6 +19,7 @@
 #include <linux/slab.h>
 #include <linux/of_address.h>
 #include <linux/of_device.h>
+#include <linux/of_iommu.h>
 #include <linux/of_irq.h>
 #include <linux/of_platform.h>
 #include <linux/platform_device.h>
@@ -166,6 +167,7 @@ static void of_dma_configure(struct device *dev)
 	int ret;
 	bool coherent;
 	unsigned long offset;
+	struct iommu_ops *iommu;
 
 	/*
 	 * Set default dma-mask to 32 bit. Drivers are expected to setup
@@ -194,7 +196,16 @@ static void of_dma_configure(struct device *dev)
 	dev_dbg(dev, "device is%sdma coherent\n",
 		coherent ? " " : " not ");
 
-	arch_setup_dma_ops(dev, coherent);
+	iommu = of_iommu_configure(dev);
+	dev_dbg(dev, "device is%sbehind an iommu\n",
+		iommu ? " " : " not ");
+
+	arch_setup_dma_ops(dev, dma_addr, size, iommu, coherent);
+}
+
+static void of_dma_deconfigure(struct device *dev)
+{
+	arch_teardown_dma_ops(dev);
 }
 
 /**
@@ -223,16 +234,12 @@ static struct platform_device *of_platform_device_create_pdata(
 	if (!dev)
 		goto err_clear_flag;
 
-	of_dma_configure(&dev->dev);
 	dev->dev.bus = &platform_bus_type;
 	dev->dev.platform_data = platform_data;
-
-	/* We do not fill the DMA ops for platform devices by default.
-	 * This is currently the responsibility of the platform code
-	 * to do such, possibly using a device notifier
-	 */
+	of_dma_configure(&dev->dev);
 
 	if (of_device_add(dev) != 0) {
+		of_dma_deconfigure(&dev->dev);
 		platform_device_put(dev);
 		goto err_clear_flag;
 	}
diff --git a/include/linux/dma-mapping.h b/include/linux/dma-mapping.h
index 8a1560f95d4a..c3007cb4bfa6 100644
--- a/include/linux/dma-mapping.h
+++ b/include/linux/dma-mapping.h
@@ -130,7 +130,13 @@ static inline int dma_coerce_mask_and_coherent(struct device *dev, u64 mask)
 extern u64 dma_get_required_mask(struct device *dev);
 
 #ifndef arch_setup_dma_ops
-static inline void arch_setup_dma_ops(struct device *dev, bool coherent) { }
+static inline void arch_setup_dma_ops(struct device *dev, u64 dma_base,
+				      u64 size, struct iommu_ops *iommu,
+				      bool coherent) { }
+#endif
+
+#ifndef arch_teardown_dma_ops
+static inline void arch_teardown_dma_ops(struct device *dev) { }
 #endif
 
 static inline unsigned int dma_get_max_seg_size(struct device *dev)
-- 
2.1.1

[PATCH v5 7/8] arm: call iommu_init before of_platform_populate

[Will Deacon]

We need to ensure that the IOMMUs in the system have a chance to perform
some basic initialisation before we start adding masters to them.

This patch adds a call to of_iommu_init before of_platform_populate.

Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
 arch/arm/kernel/setup.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c
index c03106378b49..1cfc94aa7fa2 100644
--- a/arch/arm/kernel/setup.c
+++ b/arch/arm/kernel/setup.c
@@ -18,6 +18,7 @@
 #include <linux/bootmem.h>
 #include <linux/seq_file.h>
 #include <linux/screen_info.h>
+#include <linux/of_iommu.h>
 #include <linux/of_platform.h>
 #include <linux/init.h>
 #include <linux/kexec.h>
@@ -806,6 +807,7 @@ static int __init customize_machine(void)
 	 * machine from the device tree, if no callback is provided,
 	 * otherwise we would always need an init_machine callback.
 	 */
+	of_iommu_init();
 	if (machine_desc->init_machine)
 		machine_desc->init_machine();
 #ifdef CONFIG_OF
-- 
2.1.1

[PATCH v5 8/8] arm: dma-mapping: plumb our iommu mapping ops into arch_setup_dma_ops

[Will Deacon]

This patch plumbs the existing ARM IOMMU DMA infrastructure (which isn't
actually called outside of a few drivers) into arch_setup_dma_ops, so
that we can use IOMMUs for DMA transfers in a more generic fashion.

Since this significantly complicates the arch_setup_dma_ops function,
it is moved out of line into dma-mapping.c. If CONFIG_ARM_DMA_USE_IOMMU
is not set, the iommu parameter is ignored and the normal ops are used
instead.

Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
 arch/arm/include/asm/dma-mapping.h | 12 +++---
 arch/arm/mm/dma-mapping.c          | 83 ++++++++++++++++++++++++++++++++++----
 2 files changed, 81 insertions(+), 14 deletions(-)

diff --git a/arch/arm/include/asm/dma-mapping.h b/arch/arm/include/asm/dma-mapping.h
index f3c0d953f6a2..9410b7e548fc 100644
--- a/arch/arm/include/asm/dma-mapping.h
+++ b/arch/arm/include/asm/dma-mapping.h
@@ -121,14 +121,12 @@ static inline unsigned long dma_max_pfn(struct device *dev)
 }
 #define dma_max_pfn(dev) dma_max_pfn(dev)
 
-static inline void arch_setup_dma_ops(struct device *dev, u64 dma_base,
-				      u64 size, struct iommu_ops *iommu,
-				      bool coherent)
-{
-	if (coherent)
-		set_dma_ops(dev, &arm_coherent_dma_ops);
-}
 #define arch_setup_dma_ops arch_setup_dma_ops
+extern void arch_setup_dma_ops(struct device *dev, u64 dma_base, u64 size,
+			       struct iommu_ops *iommu, bool coherent);
+
+#define arch_teardown_dma_ops arch_teardown_dma_ops
+extern void arch_teardown_dma_ops(struct device *dev);
 
 static inline dma_addr_t phys_to_dma(struct device *dev, phys_addr_t paddr)
 {
diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c
index e8907117861e..09645f00bd17 100644
--- a/arch/arm/mm/dma-mapping.c
+++ b/arch/arm/mm/dma-mapping.c
@@ -1947,9 +1947,8 @@ EXPORT_SYMBOL_GPL(arm_iommu_release_mapping);
  *	arm_iommu_create_mapping)
  *
  * Attaches specified io address space mapping to the provided device,
- * this replaces the dma operations (dma_map_ops pointer) with the
- * IOMMU aware version. More than one client might be attached to
- * the same io address space mapping.
+ * More than one client might be attached to the same io address space
+ * mapping.
  */
 int arm_iommu_attach_device(struct device *dev,
 			    struct dma_iommu_mapping *mapping)
@@ -1962,7 +1961,6 @@ int arm_iommu_attach_device(struct device *dev,
 
 	kref_get(&mapping->kref);
 	dev->archdata.mapping = mapping;
-	set_dma_ops(dev, &iommu_ops);
 
 	pr_debug("Attached IOMMU controller to %s device.\n", dev_name(dev));
 	return 0;
@@ -1974,7 +1972,6 @@ EXPORT_SYMBOL_GPL(arm_iommu_attach_device);
  * @dev: valid struct device pointer
  *
  * Detaches the provided device from a previously attached map.
- * This voids the dma operations (dma_map_ops pointer)
  */
 void arm_iommu_detach_device(struct device *dev)
 {
@@ -1989,10 +1986,82 @@ void arm_iommu_detach_device(struct device *dev)
 	iommu_detach_device(mapping->domain, dev);
 	kref_put(&mapping->kref, release_iommu_mapping);
 	dev->archdata.mapping = NULL;
-	set_dma_ops(dev, NULL);
 
 	pr_debug("Detached IOMMU controller from %s device.\n", dev_name(dev));
 }
 EXPORT_SYMBOL_GPL(arm_iommu_detach_device);
 
-#endif
+static struct dma_map_ops *arm_get_iommu_dma_map_ops(bool coherent)
+{
+	return coherent ? &iommu_coherent_ops : &iommu_ops;
+}
+
+static bool arm_setup_iommu_dma_ops(struct device *dev, u64 dma_base, u64 size,
+				    struct iommu_ops *iommu)
+{
+	struct dma_iommu_mapping *mapping;
+
+	if (!iommu)
+		return false;
+
+	mapping = arm_iommu_create_mapping(dev->bus, dma_base, size);
+	if (IS_ERR(mapping)) {
+		pr_warn("Failed to create %llu-byte IOMMU mapping for device %s\n",
+				size, dev_name(dev));
+		return false;
+	}
+
+	if (arm_iommu_attach_device(dev, mapping)) {
+		pr_warn("Failed to attached device %s to IOMMU_mapping\n",
+				dev_name(dev));
+		arm_iommu_release_mapping(mapping);
+		return false;
+	}
+
+	return true;
+}
+
+static void arm_teardown_iommu_dma_ops(struct device *dev)
+{
+	struct dma_iommu_mapping *mapping = dev->archdata.mapping;
+
+	arm_iommu_detach_device(dev);
+	arm_iommu_release_mapping(mapping);
+}
+
+#else
+
+static bool arm_setup_iommu_dma_ops(struct device *dev, u64 dma_base, u64 size,
+				    struct iommu_ops *iommu)
+{
+	return false;
+}
+
+static void arm_teardown_iommu_dma_ops(struct device *dev) { }
+
+#define arm_get_iommu_dma_map_ops arm_get_dma_map_ops
+
+#endif	/* CONFIG_ARM_DMA_USE_IOMMU */
+
+static struct dma_map_ops *arm_get_dma_map_ops(bool coherent)
+{
+	return coherent ? &arm_coherent_dma_ops : &arm_dma_ops;
+}
+
+void arch_setup_dma_ops(struct device *dev, u64 dma_base, u64 size,
+			struct iommu_ops *iommu, bool coherent)
+{
+	struct dma_map_ops *dma_ops;
+
+	if (arm_setup_iommu_dma_ops(dev, dma_base, size, iommu))
+		dma_ops = arm_get_iommu_dma_map_ops(coherent);
+	else
+		dma_ops = arm_get_dma_map_ops(coherent);
+
+	set_dma_ops(dev, dma_ops);
+}
+
+void arch_teardown_dma_ops(struct device *dev)
+{
+	arm_teardown_iommu_dma_ops(dev);
+}
-- 
2.1.1

[PATCH v5 0/8] Introduce automatic DMA configuration for IOMMU masters

[Arnd Bergmann]

On Friday 28 November 2014 13:29:38 Will Deacon wrote:
> Hi all,
> 
> Here is v5 of the patches I've previously sent here:
> 
>   RFCv1: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/283023.html
>   RFCv2: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-September/283752.html
>   RFCv3: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-September/287031.html
>   RFCv4: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/302711.html
> 
> Changes since RFCv4 include:
> 
>   - Dropped the RFC tag, since has been used by a couple of people now
> 
>   - Dropped the DMA segment configuration from of_dma_configure, as there
>     appear to be assumptions about 64k segments elsewhere in the kernel
> 
>   - Added acks/tested-bys (thanks to everybody who reviewed the series)
> 
>   - A few small fixes for issues found by Marek
> 
> Arnd: Is this too late for 3.19? We could merge the first 6 patches
> with no issues, since there aren't any callers of of_iommu_init
> without patch 7 anyway.
> 
> Up to you.

I think this looks great overall. My only feedback is the exact same
comment that Joerg already made:

On Friday 28 November 2014 14:03:36 jroedel at suse.de wrote:
> Hmm, I don't like the idea of storing private data in iommu_ops. But
> given that this is already an improvement we can build on later, here is
> my
> 
>         Acked-by: Joerg Roedel <jroedel@suse.de>
> 
> To further improve this we should probably introduce a seperate
> iommu-descriptor data-structure later which then describes a single
> hardware iommu device.

so I second that and add my

Acked-by: Arnd Bergmann <arnd@arndb.de>

Now, who should merge this series? I think someone should put all eight
patches into linux-next now, and if something goes wrong with the last
two, then we skip them for 3.19.

	Arnd

[PATCH v5 0/8] Introduce automatic DMA configuration for IOMMU masters

[Will Deacon]

On Mon, Dec 01, 2014 at 01:52:57PM +0000, Arnd Bergmann wrote:
> On Friday 28 November 2014 13:29:38 Will Deacon wrote:
> > Here is v5 of the patches I've previously sent here:
> > 
> >   RFCv1: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/283023.html
> >   RFCv2: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-September/283752.html
> >   RFCv3: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-September/287031.html
> >   RFCv4: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/302711.html
> > 
> > Changes since RFCv4 include:
> > 
> >   - Dropped the RFC tag, since has been used by a couple of people now
> > 
> >   - Dropped the DMA segment configuration from of_dma_configure, as there
> >     appear to be assumptions about 64k segments elsewhere in the kernel
> > 
> >   - Added acks/tested-bys (thanks to everybody who reviewed the series)
> > 
> >   - A few small fixes for issues found by Marek
> > 
> > Arnd: Is this too late for 3.19? We could merge the first 6 patches
> > with no issues, since there aren't any callers of of_iommu_init
> > without patch 7 anyway.
> > 
> > Up to you.
> 
> I think this looks great overall. My only feedback is the exact same
> comment that Joerg already made:
> 
> On Friday 28 November 2014 14:03:36 jroedel at suse.de wrote:
> > Hmm, I don't like the idea of storing private data in iommu_ops. But
> > given that this is already an improvement we can build on later, here is
> > my
> > 
> >         Acked-by: Joerg Roedel <jroedel@suse.de>
> > 
> > To further improve this we should probably introduce a seperate
> > iommu-descriptor data-structure later which then describes a single
> > hardware iommu device.
> 
> so I second that and add my
> 
> Acked-by: Arnd Bergmann <arnd@arndb.de>

Yup, I'll take a look at that as part of a separate series, since I really
want to have different pgsize_bitmaps anyway.

> Now, who should merge this series? I think someone should put all eight
> patches into linux-next now, and if something goes wrong with the last
> two, then we skip them for 3.19.

I think it makes most sense to go via arm-soc, but we'd need rmk's ack
on the last two patches.

Russell, are you ok with that plan?

Will