Merge branch 'pci/resource-mmap' into next

* pci/resource-mmap:
  ia64: Use generic pci_mmap_resource_range()
  ia64: Remove redundant checks for WC in pci_mmap_page_range()
  ia64: Remove redundant valid_mmap_phys_addr_range() from pci_mmap_page_range()
  PCI: Add I/O BAR support to generic pci_mmap_resource_range()
  x86/PCI: Use generic pci_mmap_resource_range()
  unicore32/PCI: Use generic pci_mmap_resource_range()
  sh/PCI: Use generic pci_mmap_resource_range()
  parisc: Use generic pci_mmap_resource_range()
  mn10300/PCI: Use generic pci_mmap_resource_range()
  MIPS: PCI: Use generic pci_mmap_resource_range()
  cris/PCI: Use generic pci_mmap_resource_range()
  ARM/PCI: Use generic pci_mmap_resource_range()
  PCI: Add pci_mmap_resource_range() and use it for ARM64
  PCI: Add BAR index argument to pci_mmap_page_range()
  PCI: Use BAR index in sysfs attr->private instead of resource pointer
  PCI: Add arch_can_pci_mmap_io() on architectures which can mmap() I/O space
  PCI: Move multiple declarations of pci_mmap_page_range() to <linux/pci.h>
  PCI: Add arch_can_pci_mmap_wc() macro
  xtensa/PCI: Do not mmap PCI BARs to userspace as write-through
  PCI: Only allow WC mmap on prefetchable resources
  PCI: Fix another sanity check bug in /proc/pci mmap
  PCI: Fix pci_mmap_fits() for HAVE_PCI_RESOURCE_TO_USER platforms

Documentation/filesystems/sysfs-pci.txt

@@ -113,9 +113,18 @@ Supporting PCI access on new platforms
 --------------------------------------
 
 In order to support PCI resource mapping as described above, Linux platform
-code must define HAVE_PCI_MMAP and provide a pci_mmap_page_range function.
-Platforms are free to only support subsets of the mmap functionality, but
-useful return codes should be provided.
+code should ideally define ARCH_GENERIC_PCI_MMAP_RESOURCE and use the generic
+implementation of that functionality. To support the historical interface of
+mmap() through files in /proc/bus/pci, platforms may also set HAVE_PCI_MMAP.
+
+Alternatively, platforms which set HAVE_PCI_MMAP may provide their own
+implementation of pci_mmap_page_range() instead of defining
+ARCH_GENERIC_PCI_MMAP_RESOURCE.
+
+Platforms which support write-combining maps of PCI resources must define
+arch_can_pci_mmap_wc() which shall evaluate to non-zero at runtime when
+write-combining is permitted. Platforms which support maps of I/O resources
+define arch_can_pci_mmap_io() similarly.
 
 Legacy resources are protected by the HAVE_PCI_LEGACY define.  Platforms
 wishing to support legacy functionality should define it and provide

arch/arm/include/asm/pci.h

@@ -29,8 +29,7 @@ static inline int pci_proc_domain(struct pci_bus *bus)
 #define PCI_DMA_BUS_IS_PHYS     (1)
 
 #define HAVE_PCI_MMAP
-extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-                               enum pci_mmap_state mmap_state, int write_combine);
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE
 
 static inline int pci_get_legacy_ide_irq(struct pci_dev *dev, int channel)
 {

arch/arm/kernel/bios32.c

@@ -597,25 +597,6 @@ resource_size_t pcibios_align_resource(void *data, const struct resource *res,
 	return start;
 }
 
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine)
-{
-	if (mmap_state == pci_mmap_io)
-		return -EINVAL;
-
-	/*
-	 * Mark this as IO
-	 */
-	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
-
-	if (remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
-			     vma->vm_end - vma->vm_start,
-			     vma->vm_page_prot))
-		return -EAGAIN;
-
-	return 0;
-}
-
 void __init pci_map_io_early(unsigned long pfn)
 {
 	struct map_desc pci_io_desc = {

arch/arm64/include/asm/pci.h

@@ -22,6 +22,8 @@
  */
 #define PCI_DMA_BUS_IS_PHYS	(0)
 
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE	1
+
 extern int isa_dma_bridge_buggy;
 
 #ifdef CONFIG_PCI

arch/cris/arch-v32/drivers/pci/bios.c

@@ -14,28 +14,6 @@ void pcibios_set_master(struct pci_dev *dev)
 	pci_write_config_byte(dev, PCI_LATENCY_TIMER, lat);
 }
 
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine)
-{
-	unsigned long prot;
-
-	/* Leave vm_pgoff as-is, the PCI space address is the physical
-	 * address on this platform.
-	 */
-	prot = pgprot_val(vma->vm_page_prot);
-	vma->vm_page_prot = __pgprot(prot);
-
-	/* Write-combine setting is ignored, it is changed via the mtrr
-	 * interfaces on this platform.
-	 */
-	if (remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
-			     vma->vm_end - vma->vm_start,
-			     vma->vm_page_prot))
-		return -EAGAIN;
-
-	return 0;
-}
-
 resource_size_t
 pcibios_align_resource(void *data, const struct resource *res,
 		       resource_size_t size, resource_size_t align)

arch/cris/include/asm/pci.h

@@ -42,9 +42,7 @@ struct pci_dev;
 #define PCI_DMA_BUS_IS_PHYS	(1)
 
 #define HAVE_PCI_MMAP
-extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			       enum pci_mmap_state mmap_state, int write_combine);
-
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE
 
 #endif /* __KERNEL__ */
 

arch/ia64/include/asm/pci.h

@@ -51,8 +51,9 @@ extern unsigned long ia64_max_iommu_merge_mask;
 #define PCI_DMA_BUS_IS_PHYS	(ia64_max_iommu_merge_mask == ~0UL)
 
 #define HAVE_PCI_MMAP
-extern int pci_mmap_page_range (struct pci_dev *dev, struct vm_area_struct *vma,
-				enum pci_mmap_state mmap_state, int write_combine);
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE
+#define arch_can_pci_mmap_wc()	1
+
 #define HAVE_PCI_LEGACY
 extern int pci_mmap_legacy_page_range(struct pci_bus *bus,
 				      struct vm_area_struct *vma,

arch/ia64/pci/pci.c

@@ -418,52 +418,6 @@ pcibios_align_resource (void *data, const struct resource *res,
 	return res->start;
 }
 
-int
-pci_mmap_page_range (struct pci_dev *dev, struct vm_area_struct *vma,
-		     enum pci_mmap_state mmap_state, int write_combine)
-{
-	unsigned long size = vma->vm_end - vma->vm_start;
-	pgprot_t prot;
-
-	/*
-	 * I/O space cannot be accessed via normal processor loads and
-	 * stores on this platform.
-	 */
-	if (mmap_state == pci_mmap_io)
-		/*
-		 * XXX we could relax this for I/O spaces for which ACPI
-		 * indicates that the space is 1-to-1 mapped.  But at the
-		 * moment, we don't support multiple PCI address spaces and
-		 * the legacy I/O space is not 1-to-1 mapped, so this is moot.
-		 */
-		return -EINVAL;
-
-	if (!valid_mmap_phys_addr_range(vma->vm_pgoff, size))
-		return -EINVAL;
-
-	prot = phys_mem_access_prot(NULL, vma->vm_pgoff, size,
-				    vma->vm_page_prot);
-
-	/*
-	 * If the user requested WC, the kernel uses UC or WC for this region,
-	 * and the chipset supports WC, we can use WC. Otherwise, we have to
-	 * use the same attribute the kernel uses.
-	 */
-	if (write_combine &&
-	    ((pgprot_val(prot) & _PAGE_MA_MASK) == _PAGE_MA_UC ||
-	     (pgprot_val(prot) & _PAGE_MA_MASK) == _PAGE_MA_WC) &&
-	    efi_range_is_wc(vma->vm_start, vma->vm_end - vma->vm_start))
-		vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
-	else
-		vma->vm_page_prot = prot;
-
-	if (remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
-			     vma->vm_end - vma->vm_start, vma->vm_page_prot))
-		return -EAGAIN;
-
-	return 0;
-}
-
 /**
  * ia64_pci_get_legacy_mem - generic legacy mem routine
  * @bus: bus to get legacy memory base address for

arch/microblaze/include/asm/pci.h

@@ -46,12 +46,10 @@ extern int pci_domain_nr(struct pci_bus *bus);
 extern int pci_proc_domain(struct pci_bus *bus);
 
 struct vm_area_struct;
-/* Map a range of PCI memory or I/O space for a device into user space */
-int pci_mmap_page_range(struct pci_dev *pdev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine);
 
 /* Tell drivers/pci/proc.c that we have pci_mmap_page_range() */
-#define HAVE_PCI_MMAP	1
+#define HAVE_PCI_MMAP		1
+#define arch_can_pci_mmap_io()	1
 
 extern int pci_legacy_read(struct pci_bus *bus, loff_t port, u32 *val,
 			   size_t count);

arch/microblaze/pci/pci-common.c

@@ -278,7 +278,7 @@ pgprot_t pci_phys_mem_access_prot(struct file *file,
  *
  * Returns a negative error code on failure, zero on success.
  */
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
+int pci_mmap_page_range(struct pci_dev *dev, int bar, struct vm_area_struct *vma,
 			enum pci_mmap_state mmap_state, int write_combine)
 {
 	resource_size_t offset =

arch/mips/include/asm/pci.h

@@ -110,10 +110,7 @@ extern unsigned long PCIBIOS_MIN_MEM;
 extern void pcibios_set_master(struct pci_dev *dev);
 
 #define HAVE_PCI_MMAP
-
-extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-	enum pci_mmap_state mmap_state, int write_combine);
-
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE
 #define HAVE_ARCH_PCI_RESOURCE_TO_USER
 
 /*

arch/mips/pci/pci.c

@@ -57,27 +57,3 @@ void pci_resource_to_user(const struct pci_dev *dev, int bar,
 	*start = fixup_bigphys_addr(rsrc->start, size);
 	*end = rsrc->start + size;
 }
-
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine)
-{
-	unsigned long prot;
-
-	/*
-	 * I/O space can be accessed via normal processor loads and stores on
-	 * this platform but for now we elect not to do this and portable
-	 * drivers should not do this anyway.
-	 */
-	if (mmap_state == pci_mmap_io)
-		return -EINVAL;
-
-	/*
-	 * Ignore write-combine; for now only return uncached mappings.
-	 */
-	prot = pgprot_val(vma->vm_page_prot);
-	prot = (prot & ~_CACHE_MASK) | _CACHE_UNCACHED;
-	vma->vm_page_prot = __pgprot(prot);
-
-	return remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
-		vma->vm_end - vma->vm_start, vma->vm_page_prot);
-}

arch/mn10300/include/asm/pci.h

@@ -74,9 +74,7 @@ static inline int pci_controller_num(struct pci_dev *dev)
 }
 
 #define HAVE_PCI_MMAP
-extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			       enum pci_mmap_state mmap_state,
-			       int write_combine);
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE
 
 #endif /* __KERNEL__ */
 

arch/mn10300/unit-asb2305/pci-asb2305.c

@@ -210,26 +210,3 @@ void __init pcibios_resource_survey(void)
 	pcibios_allocate_resources(0);
 	pcibios_allocate_resources(1);
 }
-
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine)
-{
-	unsigned long prot;
-
-	/* Leave vm_pgoff as-is, the PCI space address is the physical
-	 * address on this platform.
-	 */
-	vma->vm_flags |= VM_LOCKED;
-
-	prot = pgprot_val(vma->vm_page_prot);
-	prot &= ~_PAGE_CACHE;
-	vma->vm_page_prot = __pgprot(prot);
-
-	/* Write-combine setting is ignored */
-	if (io_remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
-			       vma->vm_end - vma->vm_start,
-			       vma->vm_page_prot))
-		return -EAGAIN;
-
-	return 0;
-}

arch/parisc/include/asm/pci.h

@@ -200,8 +200,6 @@ static inline int pci_get_legacy_ide_irq(struct pci_dev *dev, int channel)
 }
 
 #define HAVE_PCI_MMAP
-
-extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-	enum pci_mmap_state mmap_state, int write_combine);
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE
 
 #endif /* __ASM_PARISC_PCI_H */

arch/parisc/kernel/pci.c

@@ -227,34 +227,6 @@ resource_size_t pcibios_align_resource(void *data, const struct resource *res,
 	return start;
 }
 
-
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine)
-{
-	unsigned long prot;
-
-	/*
-	 * I/O space can be accessed via normal processor loads and stores on
-	 * this platform but for now we elect not to do this and portable
-	 * drivers should not do this anyway.
-	 */
-	if (mmap_state == pci_mmap_io)
-		return -EINVAL;
-
-	if (write_combine)
-		return -EINVAL;
-
-	/*
-	 * Ignore write-combine; for now only return uncached mappings.
-	 */
-	prot = pgprot_val(vma->vm_page_prot);
-	prot |= _PAGE_NO_CACHE;
-	vma->vm_page_prot = __pgprot(prot);
-
-	return remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
-		vma->vm_end - vma->vm_start, vma->vm_page_prot);
-}
-
 /*
  * A driver is enabling the device.  We make sure that all the appropriate
  * bits are set to allow the device to operate as the driver is expecting.

arch/powerpc/include/asm/pci.h

@@ -77,12 +77,11 @@ extern int pci_domain_nr(struct pci_bus *bus);
 extern int pci_proc_domain(struct pci_bus *bus);
 
 struct vm_area_struct;
-/* Map a range of PCI memory or I/O space for a device into user space */
-int pci_mmap_page_range(struct pci_dev *pdev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine);
 
-/* Tell drivers/pci/proc.c that we have pci_mmap_page_range() */
-#define HAVE_PCI_MMAP	1
+/* Tell drivers/pci/proc.c that we have pci_mmap_page_range() and it does WC */
+#define HAVE_PCI_MMAP		1
+#define arch_can_pci_mmap_io()	1
+#define arch_can_pci_mmap_wc()	1
 
 extern int pci_legacy_read(struct pci_bus *bus, loff_t port, u32 *val,
 			   size_t count);

arch/powerpc/kernel/pci-common.c

@@ -521,7 +521,8 @@ pgprot_t pci_phys_mem_access_prot(struct file *file,
  *
  * Returns a negative error code on failure, zero on success.
  */
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
+int pci_mmap_page_range(struct pci_dev *dev, int bar,
+			struct vm_area_struct *vma,
 			enum pci_mmap_state mmap_state, int write_combine)
 {
 	resource_size_t offset =

arch/sh/drivers/pci/pci.c

@@ -269,27 +269,6 @@ void __ref pcibios_report_status(unsigned int status_mask, int warn)
 	}
 }
 
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine)
-{
-	/*
-	 * I/O space can be accessed via normal processor loads and stores on
-	 * this platform but for now we elect not to do this and portable
-	 * drivers should not do this anyway.
-	 */
-	if (mmap_state == pci_mmap_io)
-		return -EINVAL;
-
-	/*
-	 * Ignore write-combine; for now only return uncached mappings.
-	 */
-	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
-
-	return remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
-			       vma->vm_end - vma->vm_start,
-			       vma->vm_page_prot);
-}
-
 #ifndef CONFIG_GENERIC_IOMAP
 
 void __iomem *__pci_ioport_map(struct pci_dev *dev,

arch/sh/include/asm/pci.h

@@ -66,8 +66,8 @@ extern unsigned long PCIBIOS_MIN_IO, PCIBIOS_MIN_MEM;
 struct pci_dev;
 
 #define HAVE_PCI_MMAP
-extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-	enum pci_mmap_state mmap_state, int write_combine);
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE
+
 extern void pcibios_set_master(struct pci_dev *dev);
 
 /* Dynamic DMA mapping stuff.

arch/sparc/include/asm/pci_64.h

@@ -42,13 +42,10 @@ static inline int pci_proc_domain(struct pci_bus *bus)
 /* Platform support for /proc/bus/pci/X/Y mmap()s. */
 
 #define HAVE_PCI_MMAP
+#define arch_can_pci_mmap_io()	1
 #define HAVE_ARCH_PCI_GET_UNMAPPED_AREA
 #define get_pci_unmapped_area get_fb_unmapped_area
 
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state,
-			int write_combine);
-
 static inline int pci_get_legacy_ide_irq(struct pci_dev *dev, int channel)
 {
 	return PCI_IRQ_NONE;

arch/sparc/kernel/pci.c

@@ -862,9 +862,9 @@ static void __pci_mmap_set_pgprot(struct pci_dev *dev, struct vm_area_struct *vm
  *
  * Returns a negative error code on failure, zero on success.
  */
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state,
-			int write_combine)
+int pci_mmap_page_range(struct pci_dev *dev, int bar,
+			struct vm_area_struct *vma,
+			enum pci_mmap_state mmap_state, int write_combine)
 {
 	int ret;
 

arch/unicore32/include/asm/pci.h

@@ -17,8 +17,7 @@
 #include <mach/hardware.h> /* for PCIBIOS_MIN_* */
 
 #define HAVE_PCI_MMAP
-extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-	enum pci_mmap_state mmap_state, int write_combine);
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE
 
 #endif /* __KERNEL__ */
 #endif

arch/unicore32/kernel/pci.c

@@ -356,26 +356,3 @@ int pcibios_enable_device(struct pci_dev *dev, int mask)
 	}
 	return 0;
 }
-
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine)
-{
-	unsigned long phys;
-
-	if (mmap_state == pci_mmap_io)
-		return -EINVAL;
-
-	phys = vma->vm_pgoff;
-
-	/*
-	 * Mark this as IO
-	 */
-	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
-
-	if (remap_pfn_range(vma, vma->vm_start, phys,
-			     vma->vm_end - vma->vm_start,
-			     vma->vm_page_prot))
-		return -EAGAIN;
-
-	return 0;
-}

arch/x86/include/asm/pci.h

@@ -7,6 +7,7 @@
 #include <linux/string.h>
 #include <linux/scatterlist.h>
 #include <asm/io.h>
+#include <asm/pat.h>
 #include <asm/x86_init.h>
 
 #ifdef __KERNEL__
@@ -102,10 +103,8 @@ int pcibios_set_irq_routing(struct pci_dev *dev, int pin, int irq);
 
 
 #define HAVE_PCI_MMAP
-extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			       enum pci_mmap_state mmap_state,
-			       int write_combine);
-
+#define arch_can_pci_mmap_wc()	pat_enabled()
+#define ARCH_GENERIC_PCI_MMAP_RESOURCE
 
 #ifdef CONFIG_PCI
 extern void early_quirks(void);

arch/x86/pci/i386.c

@@ -406,50 +406,3 @@ void __init pcibios_resource_survey(void)
 	 */
 	ioapic_insert_resources();
 }
-
-static const struct vm_operations_struct pci_mmap_ops = {
-	.access = generic_access_phys,
-};
-
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine)
-{
-	unsigned long prot;
-
-	/* I/O space cannot be accessed via normal processor loads and
-	 * stores on this platform.
-	 */
-	if (mmap_state == pci_mmap_io)
-		return -EINVAL;
-
-	prot = pgprot_val(vma->vm_page_prot);
-
-	/*
- 	 * Return error if pat is not enabled and write_combine is requested.
- 	 * Caller can followup with UC MINUS request and add a WC mtrr if there
- 	 * is a free mtrr slot.
- 	 */
-	if (!pat_enabled() && write_combine)
-		return -EINVAL;
-
-	if (pat_enabled() && write_combine)
-		prot |= cachemode2protval(_PAGE_CACHE_MODE_WC);
-	else if (pat_enabled() || boot_cpu_data.x86 > 3)
-		/*
-		 * ioremap() and ioremap_nocache() defaults to UC MINUS for now.
-		 * To avoid attribute conflicts, request UC MINUS here
-		 * as well.
-		 */
-		prot |= cachemode2protval(_PAGE_CACHE_MODE_UC_MINUS);
-
-	vma->vm_page_prot = __pgprot(prot);
-
-	if (io_remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
-			       vma->vm_end - vma->vm_start,
-			       vma->vm_page_prot))
-		return -EAGAIN;
-
-	vma->vm_ops = &pci_mmap_ops;
-
-	return 0;
-}

arch/xtensa/include/asm/pci.h

@@ -46,12 +46,9 @@ struct pci_dev;
 
 #define PCI_DMA_BUS_IS_PHYS	(1)
 
-/* Map a range of PCI memory or I/O space for a device into user space */
-int pci_mmap_page_range(struct pci_dev *pdev, struct vm_area_struct *vma,
-			enum pci_mmap_state mmap_state, int write_combine);
-
 /* Tell drivers/pci/proc.c that we have pci_mmap_page_range() */
-#define HAVE_PCI_MMAP	1
+#define HAVE_PCI_MMAP		1
+#define arch_can_pci_mmap_io()	1
 
 #endif /* __KERNEL__ */
 

arch/xtensa/kernel/pci.c

@@ -333,25 +333,6 @@ __pci_mmap_make_offset(struct pci_dev *dev, struct vm_area_struct *vma,
 	return -EINVAL;
 }
 
-/*
- * Set vm_page_prot of VMA, as appropriate for this architecture, for a pci
- * device mapping.
- */
-static __inline__ void
-__pci_mmap_set_pgprot(struct pci_dev *dev, struct vm_area_struct *vma,
-		      enum pci_mmap_state mmap_state, int write_combine)
-{
-	int prot = pgprot_val(vma->vm_page_prot);
-
-	/* Set to write-through */
-	prot = (prot & _PAGE_CA_MASK) | _PAGE_CA_WT;
-#if 0
-	if (!write_combine)
-		prot |= _PAGE_WRITETHRU;
-#endif
-	vma->vm_page_prot = __pgprot(prot);
-}
-
 /*
  * Perform the actual remap of the pages for a PCI device mapping, as
  * appropriate for this architecture.  The region in the process to map
@@ -362,7 +343,8 @@ __pci_mmap_set_pgprot(struct pci_dev *dev, struct vm_area_struct *vma,
  *
  * Returns a negative error code on failure, zero on success.
  */
-int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
+int pci_mmap_page_range(struct pci_dev *dev, int bar,
+			struct vm_area_struct *vma,
 			enum pci_mmap_state mmap_state,
 			int write_combine)
 {
@@ -372,7 +354,7 @@ int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
 	if (ret < 0)
 		return ret;
 
-	__pci_mmap_set_pgprot(dev, vma, mmap_state, write_combine);
+	vma->vm_page_prot = pgprot_device(vma->vm_page_prot);
 
 	ret = io_remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
 			         vma->vm_end - vma->vm_start,vma->vm_page_prot);

drivers/pci/Makefile

@@ -4,7 +4,7 @@
 
 obj-y		+= access.o bus.o probe.o host-bridge.o remove.o pci.o \
 			pci-driver.o search.o pci-sysfs.o rom.o setup-res.o \
-			irq.o vpd.o setup-bus.o vc.o
+			irq.o vpd.o setup-bus.o vc.o mmap.o
 obj-$(CONFIG_PROC_FS) += proc.o
 obj-$(CONFIG_SYSFS) += slot.o
 

drivers/pci/mmap.c

+/*
+ * mmap.c — generic PCI resource mmap helper
+ *
+ * Copyright © 2017 Amazon.com, Inc. or its affiliates.
+ *
+ * Author: David Woodhouse <dwmw2@infradead.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/pci.h>
+
+#ifdef ARCH_GENERIC_PCI_MMAP_RESOURCE
+
+/*
+ * Modern setup: generic pci_mmap_resource_range(), and implement the legacy
+ * pci_mmap_page_range() (if needed) as a wrapper round it.
+ */
+
+#ifdef HAVE_PCI_MMAP
+int pci_mmap_page_range(struct pci_dev *pdev, int bar,
+			struct vm_area_struct *vma,
+			enum pci_mmap_state mmap_state, int write_combine)
+{
+	resource_size_t start, end;
+
+	pci_resource_to_user(pdev, bar, &pdev->resource[bar], &start, &end);
+
+	/* Adjust vm_pgoff to be the offset within the resource */
+	vma->vm_pgoff -= start >> PAGE_SHIFT;
+	return pci_mmap_resource_range(pdev, bar, vma, mmap_state,
+				       write_combine);
+}
+#endif
+
+static const struct vm_operations_struct pci_phys_vm_ops = {
+#ifdef CONFIG_HAVE_IOREMAP_PROT
+	.access = generic_access_phys,
+#endif
+};
+
+int pci_mmap_resource_range(struct pci_dev *pdev, int bar,
+			    struct vm_area_struct *vma,
+			    enum pci_mmap_state mmap_state, int write_combine)
+{
+	unsigned long size;
+	int ret;
+
+	size = ((pci_resource_len(pdev, bar) - 1) >> PAGE_SHIFT) + 1;
+	if (vma->vm_pgoff + vma_pages(vma) > size)
+		return -EINVAL;
+
+	if (write_combine)
+		vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
+	else
+		vma->vm_page_prot = pgprot_device(vma->vm_page_prot);
+
+	if (mmap_state == pci_mmap_io) {
+		ret = pci_iobar_pfn(pdev, bar, vma);
+		if (ret)
+			return ret;
+	} else
+		vma->vm_pgoff += (pci_resource_start(pdev, bar) >> PAGE_SHIFT);
+
+	vma->vm_ops = &pci_phys_vm_ops;
+
+	return io_remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
+				  vma->vm_end - vma->vm_start,
+				  vma->vm_page_prot);
+}
+
+#elif defined(HAVE_PCI_MMAP) /* && !ARCH_GENERIC_PCI_MMAP_RESOURCE */
+
+/*
+ * Legacy setup: Impement pci_mmap_resource_range() as a wrapper around
+ * the architecture's pci_mmap_page_range(), converting to "user visible"
+ * addresses as necessary.
+ */
+
+int pci_mmap_resource_range(struct pci_dev *pdev, int bar,
+			    struct vm_area_struct *vma,
+			    enum pci_mmap_state mmap_state, int write_combine)
+{
+	resource_size_t start, end;
+
+	/*
+	 * pci_mmap_page_range() expects the same kind of entry as coming
+	 * from /proc/bus/pci/ which is a "user visible" value. If this is
+	 * different from the resource itself, arch will do necessary fixup.
+	 */
+	pci_resource_to_user(pdev, bar, &pdev->resource[bar], &start, &end);
+	vma->vm_pgoff += start >> PAGE_SHIFT;
+	return pci_mmap_page_range(pdev, bar, vma, mmap_state, write_combine);
+}
+#endif

drivers/pci/pci-sysfs.c

@@ -980,20 +980,24 @@ void pci_remove_legacy_files(struct pci_bus *b)
 }
 #endif /* HAVE_PCI_LEGACY */
 
-#ifdef HAVE_PCI_MMAP
+#if defined(HAVE_PCI_MMAP) || defined(ARCH_GENERIC_PCI_MMAP_RESOURCE)
 
 int pci_mmap_fits(struct pci_dev *pdev, int resno, struct vm_area_struct *vma,
 		  enum pci_mmap_api mmap_api)
 {
-	unsigned long nr, start, size, pci_start;
+	unsigned long nr, start, size;
+	resource_size_t pci_start = 0, pci_end;
 
 	if (pci_resource_len(pdev, resno) == 0)
 		return 0;
 	nr = vma_pages(vma);
 	start = vma->vm_pgoff;
 	size = ((pci_resource_len(pdev, resno) - 1) >> PAGE_SHIFT) + 1;
-	pci_start = (mmap_api == PCI_MMAP_PROCFS) ?
-			pci_resource_start(pdev, resno) >> PAGE_SHIFT : 0;
+	if (mmap_api == PCI_MMAP_PROCFS) {
+		pci_resource_to_user(pdev, resno, &pdev->resource[resno],
+				     &pci_start, &pci_end);
+		pci_start >>= PAGE_SHIFT;
+	}
 	if (start >= pci_start && start < pci_start + size &&
 			start + nr <= pci_start + size)
 		return 1;
@@ -1013,37 +1017,24 @@ static int pci_mmap_resource(struct kobject *kobj, struct bin_attribute *attr,
 			     struct vm_area_struct *vma, int write_combine)
 {
 	struct pci_dev *pdev = to_pci_dev(kobj_to_dev(kobj));
-	struct resource *res = attr->private;
+	int bar = (unsigned long)attr->private;
 	enum pci_mmap_state mmap_type;
-	resource_size_t start, end;
-	int i;
-
-	for (i = 0; i < PCI_ROM_RESOURCE; i++)
-		if (res == &pdev->resource[i])
-			break;
-	if (i >= PCI_ROM_RESOURCE)
-		return -ENODEV;
+	struct resource *res = &pdev->resource[bar];
 
 	if (res->flags & IORESOURCE_MEM && iomem_is_exclusive(res->start))
 		return -EINVAL;
 
-	if (!pci_mmap_fits(pdev, i, vma, PCI_MMAP_SYSFS)) {
+	if (!pci_mmap_fits(pdev, bar, vma, PCI_MMAP_SYSFS)) {
 		WARN(1, "process \"%s\" tried to map 0x%08lx bytes at page 0x%08lx on %s BAR %d (start 0x%16Lx, size 0x%16Lx)\n",
 			current->comm, vma->vm_end-vma->vm_start, vma->vm_pgoff,
-			pci_name(pdev), i,
-			(u64)pci_resource_start(pdev, i),
-			(u64)pci_resource_len(pdev, i));
+			pci_name(pdev), bar,
+			(u64)pci_resource_start(pdev, bar),
+			(u64)pci_resource_len(pdev, bar));
 		return -EINVAL;
 	}
-
-	/* pci_mmap_page_range() expects the same kind of entry as coming
-	 * from /proc/bus/pci/ which is a "user visible" value. If this is
-	 * different from the resource itself, arch will do necessary fixup.
-	 */
-	pci_resource_to_user(pdev, i, res, &start, &end);
-	vma->vm_pgoff += start >> PAGE_SHIFT;
 	mmap_type = res->flags & IORESOURCE_MEM ? pci_mmap_mem : pci_mmap_io;
-	return pci_mmap_page_range(pdev, vma, mmap_type, write_combine);
+
+	return pci_mmap_resource_range(pdev, bar, vma, mmap_type, write_combine);
 }
 
 static int pci_mmap_resource_uc(struct file *filp, struct kobject *kobj,
@@ -1065,22 +1056,18 @@ static ssize_t pci_resource_io(struct file *filp, struct kobject *kobj,
 			       loff_t off, size_t count, bool write)
 {
 	struct pci_dev *pdev = to_pci_dev(kobj_to_dev(kobj));
-	struct resource *res = attr->private;
+	int bar = (unsigned long)attr->private;
+	struct resource *res;
 	unsigned long port = off;
-	int i;
 
-	for (i = 0; i < PCI_ROM_RESOURCE; i++)
-		if (res == &pdev->resource[i])
-			break;
-	if (i >= PCI_ROM_RESOURCE)
-		return -ENODEV;
+	res = &pdev->resource[bar];
 
-	port += pci_resource_start(pdev, i);
+	port += pci_resource_start(pdev, bar);
 
-	if (port > pci_resource_end(pdev, i))
+	if (port > pci_resource_end(pdev, bar))
 		return 0;
 
-	if (port + count - 1 > pci_resource_end(pdev, i))
+	if (port + count - 1 > pci_resource_end(pdev, bar))
 		return -EINVAL;
 
 	switch (count) {
@@ -1170,16 +1157,19 @@ static int pci_create_attr(struct pci_dev *pdev, int num, int write_combine)
 	} else {
 		pdev->res_attr[num] = res_attr;
 		sprintf(res_attr_name, "resource%d", num);
-		res_attr->mmap = pci_mmap_resource_uc;
-	}
-	if (pci_resource_flags(pdev, num) & IORESOURCE_IO) {
-		res_attr->read = pci_read_resource_io;
-		res_attr->write = pci_write_resource_io;
+		if (pci_resource_flags(pdev, num) & IORESOURCE_IO) {
+			res_attr->read = pci_read_resource_io;
+			res_attr->write = pci_write_resource_io;
+			if (arch_can_pci_mmap_io())
+				res_attr->mmap = pci_mmap_resource_uc;
+		} else {
+			res_attr->mmap = pci_mmap_resource_uc;
+		}
 	}
 	res_attr->attr.name = res_attr_name;
 	res_attr->attr.mode = S_IRUSR | S_IWUSR;
 	res_attr->size = pci_resource_len(pdev, num);
-	res_attr->private = &pdev->resource[num];
+	res_attr->private = (void *)(unsigned long)num;
 	retval = sysfs_create_bin_file(&pdev->dev.kobj, res_attr);
 	if (retval)
 		kfree(res_attr);
@@ -1207,9 +1197,9 @@ static int pci_create_resource_files(struct pci_dev *pdev)
 
 		retval = pci_create_attr(pdev, i, 0);
 		/* for prefetchable resources, create a WC mappable file */
-		if (!retval && pdev->resource[i].flags & IORESOURCE_PREFETCH)
+		if (!retval && arch_can_pci_mmap_wc() &&
+		    pdev->resource[i].flags & IORESOURCE_PREFETCH)
 			retval = pci_create_attr(pdev, i, 1);
-
 		if (retval) {
 			pci_remove_resource_files(pdev);
 			return retval;

drivers/pci/pci.h

@@ -21,14 +21,14 @@ void pci_create_firmware_label_files(struct pci_dev *pdev);
 void pci_remove_firmware_label_files(struct pci_dev *pdev);
 #endif
 void pci_cleanup_rom(struct pci_dev *dev);
-#ifdef HAVE_PCI_MMAP
+
 enum pci_mmap_api {
 	PCI_MMAP_SYSFS,	/* mmap on /sys/bus/pci/devices/<BDF>/resource<N> */
 	PCI_MMAP_PROCFS	/* mmap on /proc/bus/pci/<BDF> */
 };
 int pci_mmap_fits(struct pci_dev *pdev, int resno, struct vm_area_struct *vmai,
 		  enum pci_mmap_api mmap_api);
-#endif
+
 int pci_probe_reset_function(struct pci_dev *dev);
 
 /**

drivers/pci/proc.c

@@ -202,6 +202,8 @@ static long proc_bus_pci_ioctl(struct file *file, unsigned int cmd,
 
 #ifdef HAVE_PCI_MMAP
 	case PCIIOC_MMAP_IS_IO:
+		if (!arch_can_pci_mmap_io())
+			return -EINVAL;
 		fpriv->mmap_state = pci_mmap_io;
 		break;
 
@@ -210,14 +212,15 @@ static long proc_bus_pci_ioctl(struct file *file, unsigned int cmd,
 		break;
 
 	case PCIIOC_WRITE_COMBINE:
-		if (arg)
-			fpriv->write_combine = 1;
-		else
-			fpriv->write_combine = 0;
-		break;
-
+		if (arch_can_pci_mmap_wc()) {
+			if (arg)
+				fpriv->write_combine = 1;
+			else
+				fpriv->write_combine = 0;
+			break;
+		}
+		/* If arch decided it can't, fall through... */
 #endif /* HAVE_PCI_MMAP */
-
 	default:
 		ret = -EINVAL;
 		break;
@@ -231,25 +234,35 @@ static int proc_bus_pci_mmap(struct file *file, struct vm_area_struct *vma)
 {
 	struct pci_dev *dev = PDE_DATA(file_inode(file));
 	struct pci_filp_private *fpriv = file->private_data;
-	int i, ret, write_combine;
+	int i, ret, write_combine = 0, res_bit = IORESOURCE_MEM;
 
 	if (!capable(CAP_SYS_RAWIO))
 		return -EPERM;
 
+	if (fpriv->mmap_state == pci_mmap_io) {
+		if (!arch_can_pci_mmap_io())
+			return -EINVAL;
+		res_bit = IORESOURCE_IO;
+	}
+
 	/* Make sure the caller is mapping a real resource for this device */
 	for (i = 0; i < PCI_ROM_RESOURCE; i++) {
-		if (pci_mmap_fits(dev, i, vma,  PCI_MMAP_PROCFS))
+		if (dev->resource[i].flags & res_bit &&
+		    pci_mmap_fits(dev, i, vma,  PCI_MMAP_PROCFS))
 			break;
 	}
 
 	if (i >= PCI_ROM_RESOURCE)
 		return -ENODEV;
 
-	if (fpriv->mmap_state == pci_mmap_mem)
-		write_combine = fpriv->write_combine;
-	else
-		write_combine = 0;
-	ret = pci_mmap_page_range(dev, vma,
+	if (fpriv->mmap_state == pci_mmap_mem &&
+	    fpriv->write_combine) {
+		if (dev->resource[i].flags & IORESOURCE_PREFETCH)
+			write_combine = 1;
+		else
+			return -EINVAL;
+	}
+	ret = pci_mmap_page_range(dev, i, vma,
 				  fpriv->mmap_state, write_combine);
 	if (ret < 0)
 		return ret;

include/linux/pci.h

@@ -1617,6 +1617,36 @@ static inline int pci_get_new_domain_nr(void) { return -ENOSYS; }
 
 #include <asm/pci.h>
 
+/* These two functions provide almost identical functionality. Depennding
+ * on the architecture, one will be implemented as a wrapper around the
+ * other (in drivers/pci/mmap.c).
+ *
+ * pci_mmap_resource_range() maps a specific BAR, and vm->vm_pgoff
+ * is expected to be an offset within that region.
+ *
+ * pci_mmap_page_range() is the legacy architecture-specific interface,
+ * which accepts a "user visible" resource address converted by
+ * pci_resource_to_user(), as used in the legacy mmap() interface in
+ * /proc/bus/pci/.
+ */
+int pci_mmap_resource_range(struct pci_dev *dev, int bar,
+			    struct vm_area_struct *vma,
+			    enum pci_mmap_state mmap_state, int write_combine);
+int pci_mmap_page_range(struct pci_dev *pdev, int bar,
+			struct vm_area_struct *vma,
+			enum pci_mmap_state mmap_state, int write_combine);
+
+#ifndef arch_can_pci_mmap_wc
+#define arch_can_pci_mmap_wc()		0
+#endif
+
+#ifndef arch_can_pci_mmap_io
+#define arch_can_pci_mmap_io()		0
+#define pci_iobar_pfn(pdev, bar, vma) (-EINVAL)
+#else
+int pci_iobar_pfn(struct pci_dev *pdev, int bar, struct vm_area_struct *vma);
+#endif
+
 #ifndef pci_root_bus_fwnode
 #define pci_root_bus_fwnode(bus)	NULL
 #endif