asm-generic/io.h: Add a non-posted variant of ioremap()

ARM64 currently defaults to posted MMIO (nGnRE), but some devices
require the use of non-posted MMIO (nGnRnE). Introduce a new ioremap()
variant to handle this case. ioremap_np() returns NULL on arches that
do not implement this variant.

sparc64 is the only architecture that needs to be touched directly,
because it includes neither of the generic io.h or iomap.h headers.

This adds the IORESOURCE_MEM_NONPOSTED flag, which maps to this
variant and marks a given resource as requiring non-posted mappings.
This is implemented in the resource system because it is a SoC-level
requirement, so existing drivers do not need special-case code to pick
this ioremap variant.

Then this is implemented in devres by introducing devm_ioremap_np(),
and making devm_ioremap_resource() automatically select this variant
when the resource has the IORESOURCE_MEM_NONPOSTED flag set.
Acked-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Hector Martin <marcan@marcan.st>

Documentation/driver-api/driver-model/devres.rst

@@ -309,6 +309,7 @@ IOMAP
   devm_ioremap()
   devm_ioremap_uc()
   devm_ioremap_wc()
+  devm_ioremap_np()
   devm_ioremap_resource() : checks resource, requests memory region, ioremaps
   devm_ioremap_resource_wc()
   devm_platform_ioremap_resource() : calls devm_ioremap_resource() for platform device

arch/sparc/include/asm/io_64.h

@@ -409,6 +409,10 @@ static inline void __iomem *ioremap(unsigned long offset, unsigned long size)
 #define ioremap_uc(X,Y)			ioremap((X),(Y))
 #define ioremap_wc(X,Y)			ioremap((X),(Y))
 #define ioremap_wt(X,Y)			ioremap((X),(Y))
+static inline void __iomem *ioremap_np(unsigned long offset, unsigned long size)
+{
+	return NULL;
+}
 
 static inline void iounmap(volatile void __iomem *addr)
 {

include/asm-generic/io.h

@@ -942,7 +942,9 @@ static inline void *phys_to_virt(unsigned long address)
  *
  * ioremap_wc() and ioremap_wt() can provide more relaxed caching attributes
  * for specific drivers if the architecture choses to implement them.  If they
- * are not implemented we fall back to plain ioremap.
+ * are not implemented we fall back to plain ioremap. Conversely, ioremap_np()
+ * can provide stricter non-posted write semantics if the architecture
+ * implements them.
  */
 #ifndef CONFIG_MMU
 #ifndef ioremap
@@ -993,6 +995,24 @@ static inline void __iomem *ioremap_uc(phys_addr_t offset, size_t size)
 {
 	return NULL;
 }
+
+/*
+ * ioremap_np needs an explicit architecture implementation, as it
+ * requests stronger semantics than regular ioremap(). Portable drivers
+ * should instead use one of the higher-level abstractions, like
+ * devm_ioremap_resource(), to choose the correct variant for any given
+ * device and bus. Portable drivers with a good reason to want non-posted
+ * write semantics should always provide an ioremap() fallback in case
+ * ioremap_np() is not available.
+ */
+#ifndef ioremap_np
+#define ioremap_np ioremap_np
+static inline void __iomem *ioremap_np(phys_addr_t offset, size_t size)
+{
+	return NULL;
+}
+#endif
+
 #endif
 
 #ifdef CONFIG_HAS_IOPORT_MAP

include/asm-generic/iomap.h

@@ -101,6 +101,15 @@ extern void ioport_unmap(void __iomem *);
 #define ioremap_wt ioremap
 #endif
 
+#ifndef ARCH_HAS_IOREMAP_NP
+/* See the comment in asm-generic/io.h about ioremap_np(). */
+#define ioremap_np ioremap_np
+static inline void __iomem *ioremap_np(phys_addr_t offset, size_t size)
+{
+	return NULL;
+}
+#endif
+
 #ifdef CONFIG_PCI
 /* Destroy a virtual mapping cookie for a PCI BAR (memory or IO) */
 struct pci_dev;

include/linux/io.h

@@ -68,6 +68,8 @@ void __iomem *devm_ioremap_uc(struct device *dev, resource_size_t offset,
 				   resource_size_t size);
 void __iomem *devm_ioremap_wc(struct device *dev, resource_size_t offset,
 				   resource_size_t size);
+void __iomem *devm_ioremap_np(struct device *dev, resource_size_t offset,
+				   resource_size_t size);
 void devm_iounmap(struct device *dev, void __iomem *addr);
 int check_signature(const volatile void __iomem *io_addr,
 			const unsigned char *signature, int length);

include/linux/ioport.h

@@ -108,6 +108,7 @@ struct resource {
 #define IORESOURCE_MEM_32BIT		(3<<3)
 #define IORESOURCE_MEM_SHADOWABLE	(1<<5)	/* dup: IORESOURCE_SHADOWABLE */
 #define IORESOURCE_MEM_EXPANSIONROM	(1<<6)
+#define IORESOURCE_MEM_NONPOSTED	(1<<7)
 
 /* PnP I/O specific bits (IORESOURCE_BITS) */
 #define IORESOURCE_IO_16BIT_ADDR	(1<<0)

lib/devres.c

@@ -10,6 +10,7 @@ enum devm_ioremap_type {
 	DEVM_IOREMAP = 0,
 	DEVM_IOREMAP_UC,
 	DEVM_IOREMAP_WC,
+	DEVM_IOREMAP_NP,
 };
 
 void devm_ioremap_release(struct device *dev, void *res)
@@ -42,6 +43,9 @@ static void __iomem *__devm_ioremap(struct device *dev, resource_size_t offset,
 	case DEVM_IOREMAP_WC:
 		addr = ioremap_wc(offset, size);
 		break;
+	case DEVM_IOREMAP_NP:
+		addr = ioremap_np(offset, size);
+		break;
 	}
 
 	if (addr) {
@@ -98,6 +102,21 @@ void __iomem *devm_ioremap_wc(struct device *dev, resource_size_t offset,
 }
 EXPORT_SYMBOL(devm_ioremap_wc);
 
+/**
+ * devm_ioremap_np - Managed ioremap_np()
+ * @dev: Generic device to remap IO address for
+ * @offset: Resource address to map
+ * @size: Size of map
+ *
+ * Managed ioremap_np().  Map is automatically unmapped on driver detach.
+ */
+void __iomem *devm_ioremap_np(struct device *dev, resource_size_t offset,
+			      resource_size_t size)
+{
+	return __devm_ioremap(dev, offset, size, DEVM_IOREMAP_NP);
+}
+EXPORT_SYMBOL(devm_ioremap_np);
+
 /**
  * devm_iounmap - Managed iounmap()
  * @dev: Generic device to unmap for
@@ -128,6 +147,9 @@ __devm_ioremap_resource(struct device *dev, const struct resource *res,
 		return IOMEM_ERR_PTR(-EINVAL);
 	}
 
+	if (type == DEVM_IOREMAP && res->flags & IORESOURCE_MEM_NONPOSTED)
+		type = DEVM_IOREMAP_NP;
+
 	size = resource_size(res);
 
 	if (res->name)