Skip to contentSkip to conversation

Obtaining memory information for kexec/kdump

RFC Hariprasad Nellitheertha 2 participants 13 messages Mar 24, 2005 – Mar 28, 2005

Subsystems (18)
lore
Share & tools
lore
#0 lore raw

Hi,

The topic of creating a common interface across architectures for obtaining system RAM information has been discussed on lkml and fastboot for a while now. Kexec needs information about the entire physical RAM present in the system while kdump needs information on the memory that the kernel has booted with.

The /proc/iomem interface is insufficient because its behavior varies across architectures.

- on i386, /proc/iomem reflects the memory that the system has booted with. But, it does not reflect memory above 4GB.
- on x86_64, /proc/iomem reflects the entire physical memory present irrespective of how much memory the kernel has booted with.
- on ppc64, /proc/iomem does not reflect system RAM at all.

The patches that follow provide two views in the proc file system so we have a common mechanism of extracting this information.

- A map of the entire physical memory present, irrespective of whether the system has been booted with mem= or memmap= options. This view comes up as /proc/physmem.

- A map of the memory currently used by the kernel. This map can be accessed as /proc/activemem. This view will vary from the previous one depending upon the values provided by mem= or memmap= options.

Since the patches made use of the "resource" struct to store and retrieve this information, some modifications were needed to convert the "start" and "end" fields of struct resource to u64. The first patch of the series makes this conversion. The rest of the patches provide the code for the two maps.

The patches are for the i386 and x86_64 architectures and against 2.6.12-rc1-mm1.

Kindly review and comment on the patches.

Thanks and Regards, Hari

#1 lore raw

Re: [RFC][PATCH 1/7] Converting resource struct fields to 64 bit

Regards, Hari

---

This patch converts the start and end fields of the resource structure to 64 bit. The patch makes the necesary changes in several files to handle this migration.

---

 linux-2.6.12-rc1-hari/arch/arm/kernel/bios32.c                    |    6 -
 linux-2.6.12-rc1-hari/arch/i386/kernel/efi.c                      |    2 
 linux-2.6.12-rc1-hari/arch/i386/pci/i386.c                        |    2 
 linux-2.6.12-rc1-hari/arch/ppc/kernel/pci.c                       |   26 ++---
 linux-2.6.12-rc1-hari/drivers/atm/ambassador.c                    |    2 
 linux-2.6.12-rc1-hari/drivers/atm/firestream.c                    |    4 
 linux-2.6.12-rc1-hari/drivers/block/sx8.c                         |    4 
 linux-2.6.12-rc1-hari/drivers/char/applicom.c                     |    7 -
 linux-2.6.12-rc1-hari/drivers/ide/pci/aec62xx.c                   |    3 
 linux-2.6.12-rc1-hari/drivers/ide/pci/cmd64x.c                    |    3 
 linux-2.6.12-rc1-hari/drivers/ide/pci/hpt34x.c                    |    4 
 linux-2.6.12-rc1-hari/drivers/ide/pci/pdc202xx_new.c              |    4 
 linux-2.6.12-rc1-hari/drivers/ide/pci/pdc202xx_old.c              |    4 
 linux-2.6.12-rc1-hari/drivers/ieee1394/ohci1394.c                 |   10 +-
 linux-2.6.12-rc1-hari/drivers/input/joystick/iforce/iforce-main.c |    4 
 linux-2.6.12-rc1-hari/drivers/input/serio/ct82c710.c              |    6 -
 linux-2.6.12-rc1-hari/drivers/isdn/hisax/telespci.c               |    4 
 linux-2.6.12-rc1-hari/drivers/media/video/bttv-driver.c           |    8 -
 linux-2.6.12-rc1-hari/drivers/media/video/cx88/cx88-core.c        |    4 
 linux-2.6.12-rc1-hari/drivers/media/video/cx88/cx88-mpeg.c        |    4 
 linux-2.6.12-rc1-hari/drivers/media/video/cx88/cx88-video.c       |    4 
 linux-2.6.12-rc1-hari/drivers/media/video/saa7134/saa7134-core.c  |    8 -
 linux-2.6.12-rc1-hari/drivers/message/i2o/iop.c                   |   14 +--
 linux-2.6.12-rc1-hari/drivers/mtd/devices/pmc551.c                |    8 -
 linux-2.6.12-rc1-hari/drivers/mtd/maps/amd76xrom.c                |    5 -
 linux-2.6.12-rc1-hari/drivers/mtd/maps/ichxrom.c                  |    5 -
 linux-2.6.12-rc1-hari/drivers/mtd/maps/scx200_docflash.c          |    4 
 linux-2.6.12-rc1-hari/drivers/net/8139cp.c                        |    8 -
 linux-2.6.12-rc1-hari/drivers/net/8139too.c                       |    4 
 linux-2.6.12-rc1-hari/drivers/net/e100.c                          |    4 
 linux-2.6.12-rc1-hari/drivers/net/skge.c                          |    4 
 linux-2.6.12-rc1-hari/drivers/net/tulip/de2104x.c                 |    8 -
 linux-2.6.12-rc1-hari/drivers/net/tulip/tulip_core.c              |    6 -
 linux-2.6.12-rc1-hari/drivers/net/typhoon.c                       |    5 -
 linux-2.6.12-rc1-hari/drivers/net/wan/dscc4.c                     |   12 +-
 linux-2.6.12-rc1-hari/drivers/net/wan/pc300_drv.c                 |    4 
 linux-2.6.12-rc1-hari/drivers/pci/hotplug/cpcihp_zt5550.c         |    9 +-
 linux-2.6.12-rc1-hari/drivers/pci/hotplug/cpqphp_core.c           |   10 +-
 linux-2.6.12-rc1-hari/drivers/pci/hotplug/pciehp_hpc.c            |    5 -
 linux-2.6.12-rc1-hari/drivers/pci/hotplug/shpchp_hpc.c            |   12 +-
 linux-2.6.12-rc1-hari/drivers/pci/pci-sysfs.c                     |    6 -
 linux-2.6.12-rc1-hari/drivers/pci/pci.c                           |    5 -
 linux-2.6.12-rc1-hari/drivers/pci/proc.c                          |   18 +---
 linux-2.6.12-rc1-hari/drivers/pci/setup-bus.c                     |   11 +-
 linux-2.6.12-rc1-hari/drivers/pci/setup-res.c                     |   18 ++--
 linux-2.6.12-rc1-hari/drivers/pcmcia/i82365.c                     |    5 -
 linux-2.6.12-rc1-hari/drivers/pcmcia/pd6729.c                     |    3 
 linux-2.6.12-rc1-hari/drivers/pcmcia/tcic.c                       |    5 -
 linux-2.6.12-rc1-hari/drivers/pnp/manager.c                       |   12 +-
 linux-2.6.12-rc1-hari/drivers/pnp/resource.c                      |    8 -
 linux-2.6.12-rc1-hari/drivers/scsi/sata_via.c                     |    6 -
 linux-2.6.12-rc1-hari/drivers/serial/8250_pci.c                   |    4 
 linux-2.6.12-rc1-hari/include/linux/ioport.h                      |   21 ++--
 linux-2.6.12-rc1-hari/kernel/resource.c                           |   45 ++++------
 linux-2.6.12-rc1-hari/sound/drivers/mpu401/mpu401.c               |    4 
 linux-2.6.12-rc1-hari/sound/isa/es18xx.c                          |    2 
 linux-2.6.12-rc1-hari/sound/isa/gus/interwave.c                   |    8 -
 linux-2.6.12-rc1-hari/sound/isa/sb/sb16.c                         |    2 
 linux-2.6.12-rc1-hari/sound/oss/forte.c                           |    5 -
 linux-2.6.12-rc1-hari/sound/pci/bt87x.c                           |    5 -
 linux-2.6.12-rc1-hari/sound/pci/sonicvibes.c                      |    4 
 61 files changed, 233 insertions(+), 214 deletions(-)
arch/arm/kernel/bios32.c renamed from arch/arm/kernel/bios32.c~resource-64bit +3 -3
diff -puN arch/arm/kernel/bios32.c~resource-64bit arch/arm/kernel/bios32.c
@@ -304,7 +304,7 @@ static inline int pdev_bad_for_parity(st
 static void __devinit
 pdev_fixup_device_resources(struct pci_sys_data *root, struct pci_dev *dev)
 {
-	unsigned long offset;
+	u64 offset;
 	int i;
 
 	for (i = 0; i < PCI_NUM_RESOURCES; i++) {
@@ -619,9 +619,9 @@ char * __init pcibios_setup(char *str)
  * which might be mirrored at 0x0100-0x03ff..
  */
 void pcibios_align_resource(void *data, struct resource *res,
-			    unsigned long size, unsigned long align)
+			    u64 size, u64 align)
 {
-	unsigned long start = res->start;
+	u64 start = res->start;
 
 	if (res->flags & IORESOURCE_IO && start & 0x300)
 		start = (start + 0x3ff) & ~0x3ff;
arch/i386/kernel/efi.c renamed from arch/i386/kernel/efi.c~resource-64bit +1 -1
diff -puN arch/i386/kernel/efi.c~resource-64bit arch/i386/kernel/efi.c
@@ -590,7 +590,7 @@ efi_initialize_iomem_resources(struct re
 		res->end = res->start + ((md->num_pages << EFI_PAGE_SHIFT) - 1);
 		res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
 		if (request_resource(&iomem_resource, res) < 0)
-			printk(KERN_ERR PFX "Failed to allocate res %s : 0x%lx-0x%lx\n",
+			printk(KERN_ERR PFX "Failed to allocate res %s : 0x%llx-0x%llx\n",
 				res->name, res->start, res->end);
 		/*
 		 * We don't know which region contains kernel data so we try
arch/i386/pci/i386.c renamed from arch/i386/pci/i386.c~resource-64bit +1 -1
diff -puN arch/i386/pci/i386.c~resource-64bit arch/i386/pci/i386.c
@@ -51,7 +51,7 @@ pcibios_align_resource(void *data, struc
 		       unsigned long size, unsigned long align)
 {
 	if (res->flags & IORESOURCE_IO) {
-		unsigned long start = res->start;
+		u64 start = res->start;
 
 		if (start & 0x300) {
 			start = (start + 0x3ff) & ~0x3ff;
arch/ppc/kernel/pci.c renamed from arch/ppc/kernel/pci.c~resource-64bit +13 -13
diff -puN arch/ppc/kernel/pci.c~resource-64bit arch/ppc/kernel/pci.c
@@ -114,7 +114,7 @@ pcibios_fixup_resources(struct pci_dev *
 		if (!res->flags)
 			continue;
 		if (res->end == 0xffffffff) {
-			DBG("PCI:%s Resource %d [%08lx-%08lx] is unassigned\n",
+			DBG("PCI:%s Resource %d [%08Lx-%08Lx] is unassigned\n",
 			    pci_name(dev), i, res->start, res->end);
 			res->end -= res->start;
 			res->start = 0;
@@ -173,17 +173,17 @@ EXPORT_SYMBOL(pcibios_resource_to_bus);
  * but we want to try to avoid allocating at 0x2900-0x2bff
  * which might have be mirrored at 0x0100-0x03ff..
  */
-void pcibios_align_resource(void *data, struct resource *res, unsigned long size,
-		       unsigned long align)
+void pcibios_align_resource(void *data, struct resource *res, u64 size,
+		       u64 align)
 {
 	struct pci_dev *dev = data;
 
 	if (res->flags & IORESOURCE_IO) {
-		unsigned long start = res->start;
+		u64 start = res->start;
 
 		if (size > 0x100) {
 			printk(KERN_ERR "PCI: I/O Region %s/%d too large"
-			       " (%ld bytes)\n", pci_name(dev),
+			       " (%Ld bytes)\n", pci_name(dev),
 			       dev->resource - res, size);
 		}
 
@@ -255,7 +255,7 @@ pcibios_allocate_bus_resources(struct li
 				}
 			}
 
-			DBG("PCI: bridge rsrc %lx..%lx (%lx), parent %p\n",
+			DBG("PCI: bridge rsrc %Lx..%Lx (%lx), parent %p\n",
 			    res->start, res->end, res->flags, pr);
 			if (pr) {
 				if (request_resource(pr, res) == 0)
@@ -306,7 +306,7 @@ reparent_resources(struct resource *pare
 	*pp = NULL;
 	for (p = res->child; p != NULL; p = p->sibling) {
 		p->parent = res;
-		DBG(KERN_INFO "PCI: reparented %s [%lx..%lx] under %s\n",
+		DBG(KERN_INFO "PCI: reparented %s [%Lx..%Lx] under %s\n",
 		    p->name, p->start, p->end, res->name);
 	}
 	return 0;
@@ -362,12 +362,12 @@ pci_relocate_bridge_resource(struct pci_
 		try = conflict->start - 1;
 	}
 	if (request_resource(pr, res)) {
-		DBG(KERN_ERR "PCI: huh? couldn't move to %lx..%lx\n",
+		DBG(KERN_ERR "PCI: huh? couldn't move to %Lx..%Lx\n",
 		    res->start, res->end);
 		return -1;		/* "can't happen" */
 	}
 	update_bridge_base(bus, i);
-	printk(KERN_INFO "PCI: bridge %d resource %d moved to %lx..%lx\n",
+	printk(KERN_INFO "PCI: bridge %d resource %d moved to %Lx..%Lx\n",
 	       bus->number, i, res->start, res->end);
 	return 0;
 }
@@ -479,14 +479,14 @@ static inline void alloc_resource(struct
 {
 	struct resource *pr, *r = &dev->resource[idx];
 
-	DBG("PCI:%s: Resource %d: %08lx-%08lx (f=%lx)\n",
+	DBG("PCI:%s: Resource %d: %016Lx-%016Lx (f=%lx)\n",
 	    pci_name(dev), idx, r->start, r->end, r->flags);
 	pr = pci_find_parent_resource(dev, r);
 	if (!pr || request_resource(pr, r) < 0) {
 		printk(KERN_ERR "PCI: Cannot allocate resource region %d"
 		       " of device %s\n", idx, pci_name(dev));
 		if (pr)
-			DBG("PCI:  parent is %p: %08lx-%08lx (f=%lx)\n",
+			DBG("PCI:  parent is %p: %016Lx-%016Lx (f=%lx)\n",
 			    pr, pr->start, pr->end, pr->flags);
 		/* We'll assign a new address later */
 		r->flags |= IORESOURCE_UNSET;
@@ -1061,7 +1061,7 @@ do_update_p2p_io_resource(struct pci_bus
 	DBG("Remapping Bus %d, bridge: %s\n", bus->number, pci_name(bridge));
 	res.start -= ((unsigned long) hose->io_base_virt - isa_io_base);
 	res.end -= ((unsigned long) hose->io_base_virt - isa_io_base);
-	DBG("  IO window: %08lx-%08lx\n", res.start, res.end);
+	DBG("  IO window: %016Lx-%016Lx\n", res.start, res.end);
 
 	/* Set up the top and bottom of the PCI I/O segment for this bus. */
 	pci_read_config_dword(bridge, PCI_IO_BASE, &l);
@@ -1210,7 +1210,7 @@ do_fixup_p2p_level(struct pci_bus *bus)
 				if ((r->flags & IORESOURCE_IO) == 0)
 					continue;
 				DBG("Trying to allocate from %08lx, size %08lx from parent"
-				    " res %d: %08lx -> %08lx\n",
+				    " res %d: %016Lx -> %016Lx\n",
 					res->start, res->end, i, r->start, r->end);
 			
 				if (allocate_resource(r, res, res->end + 1, res->start, max,
drivers/atm/ambassador.c renamed from drivers/atm/ambassador.c~resource-64bit +1 -1
diff -puN drivers/atm/ambassador.c~resource-64bit drivers/atm/ambassador.c
@@ -2255,7 +2255,7 @@ static int __devinit amb_probe(struct pc
 	}
 
 	PRINTD (DBG_INFO, "found Madge ATM adapter (amb) at"
-		" IO %lx, IRQ %u, MEM %p", pci_resource_start(pci_dev, 1),
+		" IO %llx, IRQ %u, MEM %p", (unsigned long long)pci_resource_start(pci_dev, 1),
 		irq, bus_to_virt(pci_resource_start(pci_dev, 0)));
 
 	// check IO region
drivers/atm/firestream.c renamed from drivers/atm/firestream.c~resource-64bit +2 -2
diff -puN drivers/atm/firestream.c~resource-64bit drivers/atm/firestream.c
@@ -1656,9 +1656,9 @@ static int __devinit fs_init (struct fs_
 	func_enter ();
 	pci_dev = dev->pci_dev;
 
-	printk (KERN_INFO "found a FireStream %d card, base %08lx, irq%d.\n", 
+	printk (KERN_INFO "found a FireStream %d card, base %08llx, irq%d.\n",
 		IS_FS50(dev)?50:155,
-		pci_resource_start(pci_dev, 0), dev->pci_dev->irq);
+		(unsigned long long)pci_resource_start(pci_dev, 0), dev->pci_dev->irq);
 
 	if (fs_debug & FS_DEBUG_INIT)
 		my_hd ((unsigned char *) dev, sizeof (*dev));
drivers/block/sx8.c renamed from drivers/block/sx8.c~resource-64bit +2 -2
diff -puN drivers/block/sx8.c~resource-64bit drivers/block/sx8.c
@@ -1689,9 +1689,9 @@ static int carm_init_one (struct pci_dev
 	DPRINTK("waiting for probe_sem\n");
 	down(&host->probe_sem);
 
-	printk(KERN_INFO "%s: pci %s, ports %d, io %lx, irq %u, major %d\n",
+	printk(KERN_INFO "%s: pci %s, ports %d, io %llx, irq %u, major %d\n",
 	       host->name, pci_name(pdev), (int) CARM_MAX_PORTS,
-	       pci_resource_start(pdev, 0), pdev->irq, host->major);
+	       (unsigned long long)pci_resource_start(pdev, 0), pdev->irq, host->major);
 
 	carm_host_id++;
 	pci_set_drvdata(pdev, host);
drivers/char/applicom.c renamed from drivers/char/applicom.c~resource-64bit +4 -3
diff -puN drivers/char/applicom.c~resource-64bit drivers/char/applicom.c
@@ -215,13 +215,14 @@ int __init applicom_init(void)
 		RamIO = ioremap(dev->resource[0].start, LEN_RAM_IO);
 
 		if (!RamIO) {
-			printk(KERN_INFO "ac.o: Failed to ioremap PCI memory space at 0x%lx\n", dev->resource[0].start);
+			printk(KERN_INFO "ac.o: Failed to ioremap PCI memory space at 0x%llx\n",
+				(unsigned long long)dev->resource[0].start);
 			pci_disable_device(dev);
 			return -EIO;
 		}
 
-		printk(KERN_INFO "Applicom %s found at mem 0x%lx, irq %d\n",
-		       applicom_pci_devnames[dev->device-1], dev->resource[0].start, 
+		printk(KERN_INFO "Applicom %s found at mem 0x%llx, irq %d\n",
+		       applicom_pci_devnames[dev->device-1], (unsigned long long)dev->resource[0].start,
 		       dev->irq);
 
 		boardno = ac_register_board(dev->resource[0].start, RamIO,0);
drivers/ide/pci/aec62xx.c renamed from drivers/ide/pci/aec62xx.c~resource-64bit +2 -1
diff -puN drivers/ide/pci/aec62xx.c~resource-64bit drivers/ide/pci/aec62xx.c
@@ -301,7 +301,8 @@ static unsigned int __devinit init_chips
 
 	if (dev->resource[PCI_ROM_RESOURCE].start) {
 		pci_write_config_dword(dev, PCI_ROM_ADDRESS, dev->resource[PCI_ROM_RESOURCE].start | PCI_ROM_ADDRESS_ENABLE);
-		printk(KERN_INFO "%s: ROM enabled at 0x%08lx\n", name, dev->resource[PCI_ROM_RESOURCE].start);
+		printk(KERN_INFO "%s: ROM enabled at 0x%08llx\n", name,
+			(unsigned long long)dev->resource[PCI_ROM_RESOURCE].start);
 	}
 
 	if (bus_speed <= 33)
drivers/ide/pci/cmd64x.c renamed from drivers/ide/pci/cmd64x.c~resource-64bit +2 -1
diff -puN drivers/ide/pci/cmd64x.c~resource-64bit drivers/ide/pci/cmd64x.c
@@ -609,7 +609,8 @@ static unsigned int __devinit init_chips
 #ifdef __i386__
 	if (dev->resource[PCI_ROM_RESOURCE].start) {
 		pci_write_config_byte(dev, PCI_ROM_ADDRESS, dev->resource[PCI_ROM_RESOURCE].start | PCI_ROM_ADDRESS_ENABLE);
-		printk(KERN_INFO "%s: ROM enabled at 0x%08lx\n", name, dev->resource[PCI_ROM_RESOURCE].start);
+		printk(KERN_INFO "%s: ROM enabled at 0x%08llx\n", name,
+		(unsigned long long)dev->resource[PCI_ROM_RESOURCE].start);
 	}
 #endif
 
drivers/ide/pci/hpt34x.c renamed from drivers/ide/pci/hpt34x.c~resource-64bit +2 -2
diff -puN drivers/ide/pci/hpt34x.c~resource-64bit drivers/ide/pci/hpt34x.c
@@ -175,8 +175,8 @@ static unsigned int __devinit init_chips
 		if (pci_resource_start(dev, PCI_ROM_RESOURCE)) {
 			pci_write_config_byte(dev, PCI_ROM_ADDRESS,
 				dev->resource[PCI_ROM_RESOURCE].start | PCI_ROM_ADDRESS_ENABLE);
-			printk(KERN_INFO "HPT345: ROM enabled at 0x%08lx\n",
-				dev->resource[PCI_ROM_RESOURCE].start);
+			printk(KERN_INFO "HPT345: ROM enabled at 0x%08llx\n",
+				(unsigned long long)dev->resource[PCI_ROM_RESOURCE].start);
 		}
 		pci_write_config_byte(dev, PCI_LATENCY_TIMER, 0xF0);
 	} else {
drivers/ide/pci/pdc202xx_new.c renamed from drivers/ide/pci/pdc202xx_new.c~resource-64bit +2 -2
diff -puN drivers/ide/pci/pdc202xx_new.c~resource-64bit drivers/ide/pci/pdc202xx_new.c
@@ -313,8 +313,8 @@ static unsigned int __devinit init_chips
 	if (dev->resource[PCI_ROM_RESOURCE].start) {
 		pci_write_config_dword(dev, PCI_ROM_ADDRESS,
 			dev->resource[PCI_ROM_RESOURCE].start | PCI_ROM_ADDRESS_ENABLE);
-		printk(KERN_INFO "%s: ROM enabled at 0x%08lx\n",
-			name, dev->resource[PCI_ROM_RESOURCE].start);
+		printk(KERN_INFO "%s: ROM enabled at 0x%08llx\n",
+			name, (unsigned long long)dev->resource[PCI_ROM_RESOURCE].start);
 	}
 
 #ifdef CONFIG_PPC_PMAC
drivers/ide/pci/pdc202xx_old.c renamed from drivers/ide/pci/pdc202xx_old.c~resource-64bit +2 -2
diff -puN drivers/ide/pci/pdc202xx_old.c~resource-64bit drivers/ide/pci/pdc202xx_old.c
@@ -580,8 +580,8 @@ static unsigned int __devinit init_chips
 	if (dev->resource[PCI_ROM_RESOURCE].start) {
 		pci_write_config_dword(dev, PCI_ROM_ADDRESS,
 			dev->resource[PCI_ROM_RESOURCE].start | PCI_ROM_ADDRESS_ENABLE);
-		printk(KERN_INFO "%s: ROM enabled at 0x%08lx\n",
-			name, dev->resource[PCI_ROM_RESOURCE].start);
+		printk(KERN_INFO "%s: ROM enabled at 0x%08llx\n",
+			name, (unsigned long long)dev->resource[PCI_ROM_RESOURCE].start);
 	}
 
 	/*
drivers/ieee1394/ohci1394.c renamed from drivers/ieee1394/ohci1394.c~resource-64bit +5 -5
diff -puN drivers/ieee1394/ohci1394.c~resource-64bit drivers/ieee1394/ohci1394.c
@@ -584,11 +584,11 @@ static void ohci_initialize(struct ti_oh
 	sprintf (irq_buf, "%s", __irq_itoa(ohci->dev->irq));
 #endif
 	PRINT(KERN_INFO, "OHCI-1394 %d.%d (PCI): IRQ=[%s]  "
-	      "MMIO=[%lx-%lx]  Max Packet=[%d]",
+	      "MMIO=[%llx-%llx]  Max Packet=[%d]",
 	      ((((buf) >> 16) & 0xf) + (((buf) >> 20) & 0xf) * 10),
 	      ((((buf) >> 4) & 0xf) + ((buf) & 0xf) * 10), irq_buf,
-	      pci_resource_start(ohci->dev, 0),
-	      pci_resource_start(ohci->dev, 0) + OHCI1394_REGISTER_SIZE - 1,
+	      (unsigned long long)pci_resource_start(ohci->dev, 0),
+	      (unsigned long long)pci_resource_start(ohci->dev, 0) + OHCI1394_REGISTER_SIZE - 1,
 	      ohci->max_packet_size);
 
 	/* Check all of our ports to make sure that if anything is
@@ -3251,8 +3251,8 @@ static int __devinit ohci1394_pci_probe(
 	 * clearly says it's 2kb, so this shouldn't be a problem. */
 	ohci_base = pci_resource_start(dev, 0);
 	if (pci_resource_len(dev, 0) != OHCI1394_REGISTER_SIZE)
-		PRINT(KERN_WARNING, "Unexpected PCI resource length of %lx!",
-		      pci_resource_len(dev, 0));
+		PRINT(KERN_WARNING, "Unexpected PCI resource length of %llx!",
+		      (unsigned long long)pci_resource_len(dev, 0));
 
 	/* Seems PCMCIA handles this internally. Not sure why. Seems
 	 * pretty bogus to force a driver to special case this.  */
drivers/input/joystick/iforce/iforce-main.c renamed from drivers/input/joystick/iforce/iforce-main.c~resource-64bit +2 -2
diff -puN drivers/input/joystick/iforce/iforce-main.c~resource-64bit drivers/input/joystick/iforce/iforce-main.c
@@ -526,9 +526,9 @@ int iforce_init_device(struct iforce *if
 
 	printk(KERN_DEBUG "iforce->dev.open = %p\n", iforce->dev.open);
 
-	printk(KERN_INFO "input: %s [%d effects, %ld bytes memory]\n",
+	printk(KERN_INFO "input: %s [%d effects, %lld bytes memory]\n",
 		iforce->dev.name, iforce->dev.ff_effects_max,
-		iforce->device_memory.end);
+		(unsigned long long)iforce->device_memory.end);
 
 	return 0;
 }
drivers/input/serio/ct82c710.c renamed from drivers/input/serio/ct82c710.c~resource-64bit +3 -3
diff -puN drivers/input/serio/ct82c710.c~resource-64bit drivers/input/serio/ct82c710.c
@@ -187,7 +187,7 @@ static struct serio * __init ct82c710_al
 		serio->write = ct82c710_write;
 		serio->dev.parent = &ct82c710_device->dev;
 		strlcpy(serio->name, "C&T 82c710 mouse port", sizeof(serio->name));
-		snprintf(serio->phys, sizeof(serio->phys), "isa%04lx/serio0", CT82C710_DATA);
+		snprintf(serio->phys, sizeof(serio->phys), "isa%04llx/serio0", (unsigned long long)CT82C710_DATA);
 	}
 
 	return serio;
@@ -209,8 +209,8 @@ static int __init ct82c710_init(void)
 
 	serio_register_port(ct82c710_port);
 
-	printk(KERN_INFO "serio: C&T 82c710 mouse port at %#lx irq %d\n",
-		CT82C710_DATA, CT82C710_IRQ);
+	printk(KERN_INFO "serio: C&T 82c710 mouse port at %#llx irq %d\n",
+		(unsigned long long)CT82C710_DATA, CT82C710_IRQ);
 
 	return 0;
 }
drivers/isdn/hisax/telespci.c renamed from drivers/isdn/hisax/telespci.c~resource-64bit +2 -2
diff -puN drivers/isdn/hisax/telespci.c~resource-64bit drivers/isdn/hisax/telespci.c
@@ -311,8 +311,8 @@ setup_telespci(struct IsdnCard *card)
 		}
 		cs->hw.teles0.membase = ioremap(pci_resource_start(dev_tel, 0),
 			PAGE_SIZE);
-		printk(KERN_INFO "Found: Zoran, base-address: 0x%lx, irq: 0x%x\n",
-			pci_resource_start(dev_tel, 0), dev_tel->irq);
+		printk(KERN_INFO "Found: Zoran, base-address: 0x%llx, irq: 0x%x\n",
+			(unsigned long long)pci_resource_start(dev_tel, 0), dev_tel->irq);
 	} else {
 		printk(KERN_WARNING "TelesPCI: No PCI card found\n");
 		return(0);
drivers/media/video/bttv-driver.c renamed from drivers/media/video/bttv-driver.c~resource-64bit +4 -4
diff -puN drivers/media/video/bttv-driver.c~resource-64bit drivers/media/video/bttv-driver.c
@@ -3821,8 +3821,8 @@ static int __devinit bttv_probe(struct p
 	if (!request_mem_region(pci_resource_start(dev,0),
 				pci_resource_len(dev,0),
 				btv->c.name)) {
-                printk(KERN_WARNING "bttv%d: can't request iomem (0x%lx).\n",
-		       btv->c.nr, pci_resource_start(dev,0));
+                printk(KERN_WARNING "bttv%d: can't request iomem (0x%llx).\n",
+		       btv->c.nr, (unsigned long long)pci_resource_start(dev,0));
 		return -EBUSY;
 	}
         pci_set_master(dev);
@@ -3838,8 +3838,8 @@ static int __devinit bttv_probe(struct p
         pci_read_config_byte(dev, PCI_LATENCY_TIMER, &lat);
         printk(KERN_INFO "bttv%d: Bt%d (rev %d) at %s, ",
                bttv_num,btv->id, btv->revision, pci_name(dev));
-        printk("irq: %d, latency: %d, mmio: 0x%lx\n",
-	       btv->c.pci->irq, lat, pci_resource_start(dev,0));
+        printk("irq: %d, latency: %d, mmio: 0x%llx\n",
+	       btv->c.pci->irq, lat, (unsigned long long)pci_resource_start(dev,0));
 	schedule();
 
 	btv->bt848_mmio=ioremap(pci_resource_start(dev,0), 0x1000);
drivers/media/video/cx88/cx88-core.c renamed from drivers/media/video/cx88/cx88-core.c~resource-64bit +2 -2
diff -puN drivers/media/video/cx88/cx88-core.c~resource-64bit drivers/media/video/cx88/cx88-core.c
@@ -1096,8 +1096,8 @@ static int get_ressources(struct cx88_co
 			       pci_resource_len(pci,0),
 			       core->name))
 		return 0;
-	printk(KERN_ERR "%s: can't get MMIO memory @ 0x%lx\n",
-	       core->name,pci_resource_start(pci,0));
+	printk(KERN_ERR "%s: can't get MMIO memory @ 0x%llx\n",
+	       core->name, (unsigned long long)pci_resource_start(pci,0));
 	return -EBUSY;
 }
 
drivers/media/video/cx88/cx88-mpeg.c renamed from drivers/media/video/cx88/cx88-mpeg.c~resource-64bit +2 -2
diff -puN drivers/media/video/cx88/cx88-mpeg.c~resource-64bit drivers/media/video/cx88/cx88-mpeg.c
@@ -344,9 +344,9 @@ int cx8802_init_common(struct cx8802_dev
 	pci_read_config_byte(dev->pci, PCI_CLASS_REVISION, &dev->pci_rev);
         pci_read_config_byte(dev->pci, PCI_LATENCY_TIMER,  &dev->pci_lat);
         printk(KERN_INFO "%s/2: found at %s, rev: %d, irq: %d, "
-	       "latency: %d, mmio: 0x%lx\n", dev->core->name,
+	       "latency: %d, mmio: 0x%llx\n", dev->core->name,
 	       pci_name(dev->pci), dev->pci_rev, dev->pci->irq,
-	       dev->pci_lat,pci_resource_start(dev->pci,0));
+	       dev->pci_lat, (unsigned long long)pci_resource_start(dev->pci,0));
 
 	/* initialize driver struct */
         init_MUTEX(&dev->lock);
drivers/media/video/cx88/cx88-video.c renamed from drivers/media/video/cx88/cx88-video.c~resource-64bit +2 -2
diff -puN drivers/media/video/cx88/cx88-video.c~resource-64bit drivers/media/video/cx88/cx88-video.c
@@ -2016,9 +2016,9 @@ static int __devinit cx8800_initdev(stru
 	pci_read_config_byte(pci_dev, PCI_CLASS_REVISION, &dev->pci_rev);
         pci_read_config_byte(pci_dev, PCI_LATENCY_TIMER,  &dev->pci_lat);
         printk(KERN_INFO "%s/0: found at %s, rev: %d, irq: %d, "
-	       "latency: %d, mmio: 0x%lx\n", core->name,
+	       "latency: %d, mmio: 0x%llx\n", core->name,
 	       pci_name(pci_dev), dev->pci_rev, pci_dev->irq,
-	       dev->pci_lat,pci_resource_start(pci_dev,0));
+	       dev->pci_lat, (unsigned long long)pci_resource_start(pci_dev,0));
 
 	pci_set_master(pci_dev);
 	if (!pci_dma_supported(pci_dev,0xffffffff)) {
drivers/media/video/saa7134/saa7134-core.c renamed from drivers/media/video/saa7134/saa7134-core.c~resource-64bit +4 -4
diff -puN drivers/media/video/saa7134/saa7134-core.c~resource-64bit drivers/media/video/saa7134/saa7134-core.c
@@ -906,9 +906,9 @@ static int __devinit saa7134_initdev(str
 	pci_read_config_byte(pci_dev, PCI_CLASS_REVISION, &dev->pci_rev);
         pci_read_config_byte(pci_dev, PCI_LATENCY_TIMER,  &dev->pci_lat);
         printk(KERN_INFO "%s: found at %s, rev: %d, irq: %d, "
-	       "latency: %d, mmio: 0x%lx\n", dev->name,
+	       "latency: %d, mmio: 0x%llx\n", dev->name,
 	       pci_name(pci_dev), dev->pci_rev, pci_dev->irq,
-	       dev->pci_lat,pci_resource_start(pci_dev,0));
+	       dev->pci_lat, (unsigned long long)pci_resource_start(pci_dev,0));
 	pci_set_master(pci_dev);
 	if (!pci_dma_supported(pci_dev,0xffffffff)) {
 		printk("%s: Oops: no 32bit PCI DMA ???\n",dev->name);
@@ -940,8 +940,8 @@ static int __devinit saa7134_initdev(str
 				pci_resource_len(pci_dev,0),
 				dev->name)) {
 		err = -EBUSY;
-		printk(KERN_ERR "%s: can't get MMIO memory @ 0x%lx\n",
-		       dev->name,pci_resource_start(pci_dev,0));
+		printk(KERN_ERR "%s: can't get MMIO memory @ 0x%llx\n",
+		       dev->name, (unsigned long long)pci_resource_start(pci_dev,0));
 		goto fail1;
 	}
 	dev->lmmio = ioremap(pci_resource_start(pci_dev,0), 0x1000);
drivers/message/i2o/iop.c renamed from drivers/message/i2o/iop.c~resource-64bit +8 -6
diff -puN drivers/message/i2o/iop.c~resource-64bit drivers/message/i2o/iop.c
@@ -767,9 +767,10 @@ static int i2o_iop_systab_set(struct i2o
 			c->mem_alloc = 1;
 			sb->current_mem_size = 1 + res->end - res->start;
 			sb->current_mem_base = res->start;
-			printk(KERN_INFO "%s: allocated %ld bytes of PCI memory"
-			       " at 0x%08lX.\n", c->name,
-			       1 + res->end - res->start, res->start);
+			printk(KERN_INFO "%s: allocated %lld bytes of PCI memory"
+			       " at 0x%08llX.\n", c->name,
+				(unsigned long long)(1 + res->end - res->start),
+				(unsigned long long)(res->start));
 		}
 	}
 
@@ -790,9 +791,10 @@ static int i2o_iop_systab_set(struct i2o
 			c->io_alloc = 1;
 			sb->current_io_size = 1 + res->end - res->start;
 			sb->current_mem_base = res->start;
-			printk(KERN_INFO "%s: allocated %ld bytes of PCI I/O at"
-			       " 0x%08lX.\n", c->name,
-			       1 + res->end - res->start, res->start);
+			printk(KERN_INFO "%s: allocated %lld bytes of PCI I/O at"
+			       " 0x%08llX.\n", c->name,
+				(unsigned long long)1 + res->end - res->start,
+				(unsigned long long)(res->start));
 		}
 	}
 
drivers/mtd/devices/pmc551.c renamed from drivers/mtd/devices/pmc551.c~resource-64bit +4 -4
diff -puN drivers/mtd/devices/pmc551.c~resource-64bit drivers/mtd/devices/pmc551.c
@@ -552,11 +552,11 @@ static u32 fixup_pmc551 (struct pci_dev 
         /*
          * Some screen fun
          */
-        printk(KERN_DEBUG "pmc551: %d%c (0x%x) of %sprefetchable memory at 0x%lx\n",
+        printk(KERN_DEBUG "pmc551: %d%c (0x%x) of %sprefetchable memory at 0x%llx\n",
 	       (size<1024)?size:(size<1048576)?size>>10:size>>20,
                (size<1024)?'B':(size<1048576)?'K':'M',
 	       size, ((dcmd&(0x1<<3)) == 0)?"non-":"",
-               (dev->resource[0].start)&PCI_BASE_ADDRESS_MEM_MASK );
+               (unsigned long long)(dev->resource[0].start)&PCI_BASE_ADDRESS_MEM_MASK );
 
         /*
          * Check to see the state of the memory
@@ -686,8 +686,8 @@ static int __init init_pmc551(void)
                         break;
                 }
 
-                printk(KERN_NOTICE "pmc551: Found PCI V370PDC at 0x%lX\n",
-				    PCI_Device->resource[0].start);
+                printk(KERN_NOTICE "pmc551: Found PCI V370PDC at 0x%llX\n",
+				    (unsigned long long)PCI_Device->resource[0].start);
 
                 /*
                  * The PMC551 device acts VERY weird if you don't init it
drivers/mtd/maps/amd76xrom.c renamed from drivers/mtd/maps/amd76xrom.c~resource-64bit +3 -2
diff -puN drivers/mtd/maps/amd76xrom.c~resource-64bit drivers/mtd/maps/amd76xrom.c
@@ -123,9 +123,10 @@ static int __devinit amd76xrom_init_one 
 		window->rsrc.parent = NULL;
 		printk(KERN_ERR MOD_NAME
 			" %s(): Unable to register resource"
-			" 0x%.08lx-0x%.08lx - kernel bug?\n",
+			" 0x%.08llx-0x%.08llx - kernel bug?\n",
 			__func__,
-			window->rsrc.start, window->rsrc.end);
+			(unsigned long long)window->rsrc.start,
+			(unsigned long long)window->rsrc.end);
 	}
 
 #if 0
drivers/mtd/maps/ichxrom.c renamed from drivers/mtd/maps/ichxrom.c~resource-64bit +3 -2
diff -puN drivers/mtd/maps/ichxrom.c~resource-64bit drivers/mtd/maps/ichxrom.c
@@ -177,9 +177,10 @@ static int __devinit ichxrom_init_one (s
 		window->rsrc.parent = NULL;
 		printk(KERN_DEBUG MOD_NAME
 			": %s(): Unable to register resource"
-			" 0x%.08lx-0x%.08lx - kernel bug?\n",
+			" 0x%.08llx-0x%.08llx - kernel bug?\n",
 			__func__,
-			window->rsrc.start, window->rsrc.end);
+			(unsigned long long)window->rsrc.start,
+			(unsigned long long)window->rsrc.end);
 	}
 
 	/* Map the firmware hub into my address space. */
drivers/mtd/maps/scx200_docflash.c renamed from drivers/mtd/maps/scx200_docflash.c~resource-64bit +2 -2
diff -puN drivers/mtd/maps/scx200_docflash.c~resource-64bit drivers/mtd/maps/scx200_docflash.c
@@ -164,8 +164,8 @@ static int __init init_scx200_docflash(v
 		outl(pmr, scx200_cb_base + SCx200_PMR);
 	}
 	
-       	printk(KERN_INFO NAME ": DOCCS mapped at 0x%lx-0x%lx, width %d\n", 
-	       docmem.start, docmem.end, width);
+       	printk(KERN_INFO NAME ": DOCCS mapped at 0x%llx-0x%llx, width %d\n",
+	       (unsigned long long)docmem.start, (unsigned long long)docmem.end, width);
 
 	scx200_docflash_map.size = size;
 	if (width == 8)
drivers/net/8139cp.c renamed from drivers/net/8139cp.c~resource-64bit +4 -4
diff -puN drivers/net/8139cp.c~resource-64bit drivers/net/8139cp.c
@@ -1725,8 +1725,8 @@ static int cp_init_one (struct pci_dev *
 	}
 	if (pci_resource_len(pdev, 1) < CP_REGS_SIZE) {
 		rc = -EIO;
-		printk(KERN_ERR PFX "MMIO resource (%lx) too small on pci dev %s\n",
-		       pci_resource_len(pdev, 1), pci_name(pdev));
+		printk(KERN_ERR PFX "MMIO resource (%llx) too small on pci dev %s\n",
+		       (unsigned long long)pci_resource_len(pdev, 1), pci_name(pdev));
 		goto err_out_res;
 	}
 
@@ -1758,8 +1758,8 @@ static int cp_init_one (struct pci_dev *
 	regs = ioremap(pciaddr, CP_REGS_SIZE);
 	if (!regs) {
 		rc = -EIO;
-		printk(KERN_ERR PFX "Cannot map PCI MMIO (%lx@%lx) on pci dev %s\n",
-		       pci_resource_len(pdev, 1), pciaddr, pci_name(pdev));
+		printk(KERN_ERR PFX "Cannot map PCI MMIO (%llx@%lx) on pci dev %s\n",
+		       (unsigned long long)pci_resource_len(pdev, 1), pciaddr, pci_name(pdev));
 		goto err_out_res;
 	}
 	dev->base_addr = (unsigned long) regs;
drivers/net/8139too.c renamed from drivers/net/8139too.c~resource-64bit +2 -2
diff -puN drivers/net/8139too.c~resource-64bit drivers/net/8139too.c
@@ -1337,9 +1337,9 @@ static int rtl8139_open (struct net_devi
 	netif_start_queue (dev);
 
 	if (netif_msg_ifup(tp))
-		printk(KERN_DEBUG "%s: rtl8139_open() ioaddr %#lx IRQ %d"
+		printk(KERN_DEBUG "%s: rtl8139_open() ioaddr %#llx IRQ %d"
 			" GP Pins %2.2x %s-duplex.\n",
-			dev->name, pci_resource_start (tp->pci_dev, 1),
+			dev->name, (unsigned long long)pci_resource_start (tp->pci_dev, 1),
 			dev->irq, RTL_R8 (MediaStatus),
 			tp->mii.full_duplex ? "full" : "half");
 
drivers/net/e100.c renamed from drivers/net/e100.c~resource-64bit +2 -2
diff -puN drivers/net/e100.c~resource-64bit drivers/net/e100.c
@@ -2316,9 +2316,9 @@ static int __devinit e100_probe(struct p
 		goto err_out_free;
 	}
 
-	DPRINTK(PROBE, INFO, "addr 0x%lx, irq %d, "
+	DPRINTK(PROBE, INFO, "addr 0x%llx, irq %d, "
 		"MAC addr %02X:%02X:%02X:%02X:%02X:%02X\n",
-		pci_resource_start(pdev, 0), pdev->irq,
+		(unsigned long long)pci_resource_start(pdev, 0), pdev->irq,
 		netdev->dev_addr[0], netdev->dev_addr[1], netdev->dev_addr[2],
 		netdev->dev_addr[3], netdev->dev_addr[4], netdev->dev_addr[5]);
 
drivers/net/skge.c renamed from drivers/net/skge.c~resource-64bit +2 -2
diff -puN drivers/net/skge.c~resource-64bit drivers/net/skge.c
@@ -3229,8 +3229,8 @@ static int __devinit skge_probe(struct p
 	if (err)
 		goto err_out_free_irq;
 
-	printk(KERN_INFO PFX "addr 0x%lx irq %d chip %s rev %d\n",
-	       pci_resource_start(pdev, 0), pdev->irq,
+	printk(KERN_INFO PFX "addr 0x%llx irq %d chip %s rev %d\n",
+	       (unsigned long long)pci_resource_start(pdev, 0), pdev->irq,
 	       skge_board_name(hw), chip_rev(hw));
 
 	if ((dev = skge_devinit(hw, 0)) == NULL)
drivers/net/tulip/de2104x.c renamed from drivers/net/tulip/de2104x.c~resource-64bit +4 -4
diff -puN drivers/net/tulip/de2104x.c~resource-64bit drivers/net/tulip/de2104x.c
@@ -2007,8 +2007,8 @@ static int __init de_init_one (struct pc
 	}
 	if (pci_resource_len(pdev, 1) < DE_REGS_SIZE) {
 		rc = -EIO;
-		printk(KERN_ERR PFX "MMIO resource (%lx) too small on pci dev %s\n",
-		       pci_resource_len(pdev, 1), pci_name(pdev));
+		printk(KERN_ERR PFX "MMIO resource (%llx) too small on pci dev %s\n",
+		       (unsigned long long)pci_resource_len(pdev, 1), pci_name(pdev));
 		goto err_out_res;
 	}
 
@@ -2016,8 +2016,8 @@ static int __init de_init_one (struct pc
 	regs = ioremap_nocache(pciaddr, DE_REGS_SIZE);
 	if (!regs) {
 		rc = -EIO;
-		printk(KERN_ERR PFX "Cannot map PCI MMIO (%lx@%lx) on pci dev %s\n",
-		       pci_resource_len(pdev, 1), pciaddr, pci_name(pdev));
+		printk(KERN_ERR PFX "Cannot map PCI MMIO (%llx@%lx) on pci dev %s\n",
+		       (unsigned long long)pci_resource_len(pdev, 1), pciaddr, pci_name(pdev));
 		goto err_out_res;
 	}
 	dev->base_addr = (unsigned long) regs;
drivers/net/tulip/tulip_core.c renamed from drivers/net/tulip/tulip_core.c~resource-64bit +3 -3
diff -puN drivers/net/tulip/tulip_core.c~resource-64bit drivers/net/tulip/tulip_core.c
@@ -1359,10 +1359,10 @@ static int __devinit tulip_init_one (str
 	SET_MODULE_OWNER(dev);
 	SET_NETDEV_DEV(dev, &pdev->dev);
 	if (pci_resource_len (pdev, 0) < tulip_tbl[chip_idx].io_size) {
-		printk (KERN_ERR PFX "%s: I/O region (0x%lx@0x%lx) too small, "
+		printk (KERN_ERR PFX "%s: I/O region (0x%llx@0x%llx) too small, "
 			"aborting\n", pci_name(pdev),
-			pci_resource_len (pdev, 0),
-			pci_resource_start (pdev, 0));
+			(unsigned long long)pci_resource_len (pdev, 0),
+			(unsigned long long)pci_resource_start (pdev, 0));
 		goto err_out_free_netdev;
 	}
 
drivers/net/typhoon.c renamed from drivers/net/typhoon.c~resource-64bit +3 -2
diff -puN drivers/net/typhoon.c~resource-64bit drivers/net/typhoon.c
@@ -2569,9 +2569,10 @@ typhoon_init_one(struct pci_dev *pdev, c
 
 	pci_set_drvdata(pdev, dev);
 
-	printk(KERN_INFO "%s: %s at %s 0x%lx, ",
+	printk(KERN_INFO "%s: %s at %s 0x%llx, ",
 	       dev->name, typhoon_card_info[card_id].name,
-	       use_mmio ? "MMIO" : "IO", pci_resource_start(pdev, use_mmio));
+	       use_mmio ? "MMIO" : "IO",
+		(unsigned long long)pci_resource_start(pdev, use_mmio));
 	for(i = 0; i < 5; i++)
 		printk("%2.2x:", dev->dev_addr[i]);
 	printk("%2.2x\n", dev->dev_addr[i]);
drivers/net/wan/dscc4.c renamed from drivers/net/wan/dscc4.c~resource-64bit +6 -6
diff -puN drivers/net/wan/dscc4.c~resource-64bit drivers/net/wan/dscc4.c
@@ -731,15 +731,15 @@ static int __devinit dscc4_init_one(stru
 	ioaddr = ioremap(pci_resource_start(pdev, 0),
 					pci_resource_len(pdev, 0));
 	if (!ioaddr) {
-		printk(KERN_ERR "%s: cannot remap MMIO region %lx @ %lx\n",
-			DRV_NAME, pci_resource_len(pdev, 0),
-			pci_resource_start(pdev, 0));
+		printk(KERN_ERR "%s: cannot remap MMIO region %llx @ %llx\n",
+			DRV_NAME, (unsigned long long)pci_resource_len(pdev, 0),
+			(unsigned long long)pci_resource_start(pdev, 0));
 		rc = -EIO;
 		goto err_free_mmio_regions_2;
 	}
-	printk(KERN_DEBUG "Siemens DSCC4, MMIO at %#lx (regs), %#lx (lbi), IRQ %d\n",
-	        pci_resource_start(pdev, 0),
-	        pci_resource_start(pdev, 1), pdev->irq);
+	printk(KERN_DEBUG "Siemens DSCC4, MMIO at %#llx (regs), %#llx (lbi), IRQ %d\n",
+	        (unsigned long long)pci_resource_start(pdev, 0),
+	        (unsigned long long)pci_resource_start(pdev, 1), pdev->irq);
 
 	/* Cf errata DS5 p.2 */
 	pci_write_config_byte(pdev, PCI_LATENCY_TIMER, 0xf8);
drivers/net/wan/pc300_drv.c renamed from drivers/net/wan/pc300_drv.c~resource-64bit +2 -2
diff -puN drivers/net/wan/pc300_drv.c~resource-64bit drivers/net/wan/pc300_drv.c
@@ -3441,9 +3441,9 @@ cpc_init_one(struct pci_dev *pdev, const
 
 	card = (pc300_t *) kmalloc(sizeof(pc300_t), GFP_KERNEL);
 	if (card == NULL) {
-		printk("PC300 found at RAM 0x%08lx, "
+		printk("PC300 found at RAM 0x%08llx, "
 		       "but could not allocate card structure.\n",
-		       pci_resource_start(pdev, 3));
+		       (unsigned long long)pci_resource_start(pdev, 3));
 		return -ENOMEM;
 	}
 	memset(card, 0, sizeof(pc300_t));
drivers/pci/hotplug/cpcihp_zt5550.c renamed from drivers/pci/hotplug/cpcihp_zt5550.c~resource-64bit +5 -4
diff -puN drivers/pci/hotplug/cpcihp_zt5550.c~resource-64bit drivers/pci/hotplug/cpcihp_zt5550.c
@@ -85,8 +85,8 @@ static int zt5550_hc_config(struct pci_d
 	}
 	hc_dev = pdev;
 	dbg("hc_dev = %p", hc_dev);
-	dbg("pci resource start %lx", pci_resource_start(hc_dev, 1));
-	dbg("pci resource len %lx", pci_resource_len(hc_dev, 1));
+	dbg("pci resource start %llx", (unsigned long long)pci_resource_start(hc_dev, 1));
+	dbg("pci resource len %llx", (unsigned long long)pci_resource_len(hc_dev, 1));
 
 	if(!request_mem_region(pci_resource_start(hc_dev, 1),
 				pci_resource_len(hc_dev, 1), MY_NAME)) {
@@ -97,8 +97,9 @@ static int zt5550_hc_config(struct pci_d
 	hc_registers =
 	    ioremap(pci_resource_start(hc_dev, 1), pci_resource_len(hc_dev, 1));
 	if(!hc_registers) {
-		err("cannot remap MMIO region %lx @ %lx",
-		    pci_resource_len(hc_dev, 1), pci_resource_start(hc_dev, 1));
+		err("cannot remap MMIO region %llx @ %llx",
+		    (unsigned long long)pci_resource_len(hc_dev, 1),
+			(unsigned long long)pci_resource_start(hc_dev, 1));
 		release_mem_region(pci_resource_start(hc_dev, 1),
 				   pci_resource_len(hc_dev, 1));
 		return -ENODEV;
drivers/pci/hotplug/cpqphp_core.c renamed from drivers/pci/hotplug/cpqphp_core.c~resource-64bit +5 -5
diff -puN drivers/pci/hotplug/cpqphp_core.c~resource-64bit drivers/pci/hotplug/cpqphp_core.c
@@ -1066,8 +1066,8 @@ static int cpqhpc_probe(struct pci_dev *
 	}
 	
 	dbg("pdev = %p\n", pdev);
-	dbg("pci resource start %lx\n", pci_resource_start(pdev, 0));
-	dbg("pci resource len %lx\n", pci_resource_len(pdev, 0));
+	dbg("pci resource start %llx\n", (unsigned long long)pci_resource_start(pdev, 0));
+	dbg("pci resource len %llx\n", (unsigned long long)pci_resource_len(pdev, 0));
 
 	if (!request_mem_region(pci_resource_start(pdev, 0),
 				pci_resource_len(pdev, 0), MY_NAME)) {
@@ -1079,9 +1079,9 @@ static int cpqhpc_probe(struct pci_dev *
 	ctrl->hpc_reg = ioremap(pci_resource_start(pdev, 0),
 					pci_resource_len(pdev, 0));
 	if (!ctrl->hpc_reg) {
-		err("cannot remap MMIO region %lx @ %lx\n",
-				pci_resource_len(pdev, 0),
-				pci_resource_start(pdev, 0));
+		err("cannot remap MMIO region %llx @ %llx\n",
+				(unsigned long long)pci_resource_len(pdev, 0),
+				(unsigned long long)pci_resource_start(pdev, 0));
 		rc = -ENODEV;
 		goto err_free_mem_region;
 	}
drivers/pci/hotplug/pciehp_hpc.c renamed from drivers/pci/hotplug/pciehp_hpc.c~resource-64bit +3 -2
diff -puN drivers/pci/hotplug/pciehp_hpc.c~resource-64bit drivers/pci/hotplug/pciehp_hpc.c
@@ -1343,8 +1343,9 @@ int pcie_init(struct controller * ctrl,
 		PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn), dev->irq);
 	for ( rc = 0; rc < DEVICE_COUNT_RESOURCE; rc++)
 		if (pci_resource_len(pdev, rc) > 0)
-			dbg("pci resource[%d] start=0x%lx(len=0x%lx)\n", rc,
-				pci_resource_start(pdev, rc), pci_resource_len(pdev, rc));
+			dbg("pci resource[%d] start=0x%llx(len=0x%llx)\n", rc,
+				(unsigned long long)pci_resource_start(pdev, rc),
+				(unsigned long long)pci_resource_len(pdev, rc));
 
 	info("HPC vendor_id %x device_id %x ss_vid %x ss_did %x\n", pdev->vendor, pdev->device, 
 		pdev->subsystem_vendor, pdev->subsystem_device);
drivers/pci/hotplug/shpchp_hpc.c renamed from drivers/pci/hotplug/shpchp_hpc.c~resource-64bit +7 -5
diff -puN drivers/pci/hotplug/shpchp_hpc.c~resource-64bit drivers/pci/hotplug/shpchp_hpc.c
@@ -1482,8 +1482,9 @@ int shpc_init(struct controller * ctrl,
 		PCI_FUNC(pdev->devfn), pdev->irq);
 	for ( rc = 0; rc < DEVICE_COUNT_RESOURCE; rc++)
 		if (pci_resource_len(pdev, rc) > 0)
-			dbg("pci resource[%d] start=0x%lx(len=0x%lx), shpc_base_offset %x\n", rc,
-				pci_resource_start(pdev, rc), pci_resource_len(pdev, rc), shpc_base_offset);
+			dbg("pci resource[%d] start=0x%llx(len=0x%llx), shpc_base_offset %x\n", rc,
+				(unsigned long long)pci_resource_start(pdev, rc),
+				(unsigned long long)pci_resource_len(pdev, rc), shpc_base_offset);
 
 	info("HPC vendor_id %x device_id %x ss_vid %x ss_did %x\n", pdev->vendor, pdev->device, pdev->subsystem_vendor, 
 		pdev->subsystem_device);
@@ -1498,13 +1499,14 @@ int shpc_init(struct controller * ctrl,
 
 	php_ctlr->creg = ioremap(pci_resource_start(pdev, 0) + shpc_base_offset, pci_resource_len(pdev, 0));
 	if (!php_ctlr->creg) {
-		err("%s: cannot remap MMIO region %lx @ %lx\n", __FUNCTION__, pci_resource_len(pdev, 0), 
-			pci_resource_start(pdev, 0) + shpc_base_offset);
+		err("%s: cannot remap MMIO region %llx @ %llx\n", __FUNCTION__,
+			(unsigned long long)pci_resource_len(pdev, 0),
+			(unsigned long long)pci_resource_start(pdev, 0) + shpc_base_offset);
 		release_mem_region(pci_resource_start(pdev, 0) + shpc_base_offset, pci_resource_len(pdev, 0));
 		goto abort_free_ctlr;
 	}
 	dbg("%s: php_ctlr->creg %p\n", __FUNCTION__, php_ctlr->creg);
-	dbg("%s: physical addr %p\n", __FUNCTION__, (void*)pci_resource_start(pdev, 0));
+	dbg("%s: physical addr %llx\n", __FUNCTION__, (unsigned long long)pci_resource_start(pdev, 0));
 
 	init_MUTEX(&ctrl->crit_sect);
 	/* Setup wait queue */
drivers/pci/pci.c renamed from drivers/pci/pci.c~resource-64bit +3 -2
diff -puN drivers/pci/pci.c~resource-64bit drivers/pci/pci.c
@@ -623,10 +623,11 @@ int pci_request_region(struct pci_dev *p
 	return 0;
 
 err_out:
-	printk (KERN_WARNING "PCI: Unable to reserve %s region #%d:%lx@%lx for device %s\n",
+	printk (KERN_WARNING "PCI: Unable to reserve %s region #%d:%llx@%llx for device %s\n",
 		pci_resource_flags(pdev, bar) & IORESOURCE_IO ? "I/O" : "mem",
 		bar + 1, /* PCI BAR # */
-		pci_resource_len(pdev, bar), pci_resource_start(pdev, bar),
+		(unsigned long long)pci_resource_len(pdev, bar),
+		(unsigned long long)pci_resource_start(pdev, bar),
 		pci_name(pdev));
 	return -EBUSY;
 }
drivers/pci/pci-sysfs.c renamed from drivers/pci/pci-sysfs.c~resource-64bit +3 -3
diff -puN drivers/pci/pci-sysfs.c~resource-64bit drivers/pci/pci-sysfs.c
@@ -65,9 +65,9 @@ resource_show(struct device * dev, char 
 		max = DEVICE_COUNT_RESOURCE;
 
 	for (i = 0; i < max; i++) {
-		str += sprintf(str,"0x%016lx 0x%016lx 0x%016lx\n",
-			       pci_resource_start(pci_dev,i),
-			       pci_resource_end(pci_dev,i),
+		str += sprintf(str,"0x%016llx 0x%016llx 0x%016lx\n",
+			       (unsigned long long)pci_resource_start(pci_dev,i),
+			       (unsigned long long)pci_resource_end(pci_dev,i),
 			       pci_resource_flags(pci_dev,i));
 	}
 	return (str - buf);
drivers/pci/proc.c renamed from drivers/pci/proc.c~resource-64bit +6 -12
diff -puN drivers/pci/proc.c~resource-64bit drivers/pci/proc.c
@@ -301,12 +301,6 @@ static struct file_operations proc_bus_p
 #endif /* HAVE_PCI_MMAP */
 };
 
-#if BITS_PER_LONG == 32
-#define LONG_FORMAT "\t%08lx"
-#else
-#define LONG_FORMAT "\t%16lx"
-#endif
-
 /* iterator */
 static void *pci_seq_start(struct seq_file *m, loff_t *pos)
 {
@@ -358,13 +352,13 @@ static int show_device(struct seq_file *
 			dev->irq);
 	/* Here should be 7 and not PCI_NUM_RESOURCES as we need to preserve compatibility */
 	for(i=0; i<7; i++)
-		seq_printf(m, LONG_FORMAT,
-			dev->resource[i].start |
-			(dev->resource[i].flags & PCI_REGION_FLAG_MASK));
+		seq_printf(m, "\t%16llx",
+			(unsigned long long)(dev->resource[i].start |
+			(dev->resource[i].flags & PCI_REGION_FLAG_MASK)));
 	for(i=0; i<7; i++)
-		seq_printf(m, LONG_FORMAT,
-			dev->resource[i].start < dev->resource[i].end ?
-			dev->resource[i].end - dev->resource[i].start + 1 : 0);
+		seq_printf(m, "\t%16llx",
+			(unsigned long long)(dev->resource[i].start < dev->resource[i].end ?
+			dev->resource[i].end - dev->resource[i].start + 1 : 0));
 	seq_putc(m, '\t');
 	if (drv)
 		seq_printf(m, "%s", drv->name);
drivers/pci/setup-bus.c renamed from drivers/pci/setup-bus.c~resource-64bit +6 -5
diff -puN drivers/pci/setup-bus.c~resource-64bit drivers/pci/setup-bus.c
@@ -332,8 +332,8 @@ static int __devinit
 pbus_size_mem(struct pci_bus *bus, unsigned long mask, unsigned long type)
 {
 	struct pci_dev *dev;
-	unsigned long min_align, align, size;
-	unsigned long aligns[12];	/* Alignments from 1Mb to 2Gb */
+	u64 min_align, align, size;
+	u64 aligns[12];	/* Alignments from 1Mb to 2Gb */
 	int order, max_order;
 	struct resource *b_res = find_free_bus_resource(bus, type);
 
@@ -359,8 +359,9 @@ pbus_size_mem(struct pci_bus *bus, unsig
 			order = __ffs(align) - 20;
 			if (order > 11) {
 				printk(KERN_WARNING "PCI: region %s/%d "
-				       "too large: %lx-%lx\n",
-				       pci_name(dev), i, r->start, r->end);
+				       "too large: %llx-%llx\n",
+				       pci_name(dev), i, (unsigned long long)r->start,
+				       (unsigned long long)r->end);
 				r->flags = 0;
 				continue;
 			}
@@ -379,7 +380,7 @@ pbus_size_mem(struct pci_bus *bus, unsig
 	align = 0;
 	min_align = 0;
 	for (order = 0; order <= max_order; order++) {
-		unsigned long align1 = 1UL << (order + 20);
+		u64 align1 = 1UL << (order + 20);
 
 		if (!align)
 			min_align = align1;
drivers/pci/setup-res.c renamed from drivers/pci/setup-res.c~resource-64bit +10 -8
diff -puN drivers/pci/setup-res.c~resource-64bit drivers/pci/setup-res.c
@@ -108,10 +108,11 @@ pci_claim_resource(struct pci_dev *dev, 
 		err = insert_resource(root, res);
 
 	if (err) {
-		printk(KERN_ERR "PCI: %s region %d of %s %s [%lx:%lx]\n",
+		printk(KERN_ERR "PCI: %s region %d of %s %s [%llx:%llx]\n",
 		       root ? "Address space collision on" :
 			      "No parent found for",
-		       resource, dtype, pci_name(dev), res->start, res->end);
+		       resource, dtype, pci_name(dev), (unsigned long long)res->start,
+			(unsigned long long)res->end);
 	}
 
 	return err;
@@ -121,7 +122,7 @@ int pci_assign_resource(struct pci_dev *
 {
 	struct pci_bus *bus = dev->bus;
 	struct resource *res = dev->resource + resno;
-	unsigned long size, min, align;
+	u64 size, min, align;
 	int ret;
 
 	size = res->end - res->start + 1;
@@ -148,9 +149,9 @@ int pci_assign_resource(struct pci_dev *
 	}
 
 	if (ret) {
-		printk(KERN_ERR "PCI: Failed to allocate %s resource #%d:%lx@%lx for %s\n",
+		printk(KERN_ERR "PCI: Failed to allocate %s resource #%d:%llx@%llx for %s\n",
 		       res->flags & IORESOURCE_IO ? "I/O" : "mem",
-		       resno, size, res->start, pci_name(dev));
+		       resno, (unsigned long long)size, (unsigned long long)res->start, pci_name(dev));
 	} else if (resno < PCI_BRIDGE_RESOURCES) {
 		pci_update_resource(dev, res, resno);
 	}
@@ -167,7 +168,7 @@ pdev_sort_resources(struct pci_dev *dev,
 	for (i = 0; i < PCI_NUM_RESOURCES; i++) {
 		struct resource *r;
 		struct resource_list *list, *tmp;
-		unsigned long r_align;
+		u64 r_align;
 
 		r = &dev->resource[i];
 		r_align = r->end - r->start;
@@ -176,8 +177,9 @@ pdev_sort_resources(struct pci_dev *dev,
 			continue;
 		if (!r_align) {
 			printk(KERN_WARNING "PCI: Ignore bogus resource %d "
-					    "[%lx:%lx] of %s\n",
-					    i, r->start, r->end, pci_name(dev));
+					    "[%llx:%llx] of %s\n",
+					    i, (unsigned long long)r->start,
+						(unsigned long long)r->end, pci_name(dev));
 			continue;
 		}
 		r_align = (i < PCI_BRIDGE_RESOURCES) ? r_align + 1 : r->start;
drivers/pcmcia/i82365.c renamed from drivers/pcmcia/i82365.c~resource-64bit +3 -2
diff -puN drivers/pcmcia/i82365.c~resource-64bit drivers/pcmcia/i82365.c
@@ -1165,9 +1165,10 @@ static int i365_set_mem_map(u_short sock
     u_short base, i;
     u_char map;
     
-    debug(1, "SetMemMap(%d, %d, %#2.2x, %d ns, %#lx-%#lx, "
+    debug(1, "SetMemMap(%d, %d, %#2.2x, %d ns, %#llx-%#llx, "
 	  "%#x)\n", sock, mem->map, mem->flags, mem->speed,
-	  mem->res->start, mem->res->end, mem->card_start);
+	  (unsigned long long)mem->res->start,
+		(unsigned long long)mem->res->end, mem->card_start);
 
     map = mem->map;
     if ((map > 4) || (mem->card_start > 0x3ffffff) ||
drivers/pcmcia/pd6729.c renamed from drivers/pcmcia/pd6729.c~resource-64bit +2 -1
diff -puN drivers/pcmcia/pd6729.c~resource-64bit drivers/pcmcia/pd6729.c
@@ -715,7 +715,8 @@ static int __devinit pd6729_pci_probe(st
 		goto err_out_free_mem;
 
 	printk(KERN_INFO "pd6729: Cirrus PD6729 PCI to PCMCIA Bridge "
-		"at 0x%lx on irq %d\n",	pci_resource_start(dev, 0), dev->irq);
+		"at 0x%llx on irq %d\n",
+		(unsigned long long)pci_resource_start(dev, 0), dev->irq);
  	/*
 	 * Since we have no memory BARs some firmware may not
 	 * have had PCI_COMMAND_MEMORY enabled, yet the device needs it.
drivers/pcmcia/tcic.c renamed from drivers/pcmcia/tcic.c~resource-64bit +3 -2
diff -puN drivers/pcmcia/tcic.c~resource-64bit drivers/pcmcia/tcic.c
@@ -833,8 +833,9 @@ static int tcic_set_mem_map(struct pcmci
     u_long base, len, mmap;
 
     debug(1, "SetMemMap(%d, %d, %#2.2x, %d ns, "
-	  "%#lx-%#lx, %#x)\n", psock, mem->map, mem->flags,
-	  mem->speed, mem->res->start, mem->res->end, mem->card_start);
+	  "%#llx-%#llx, %#x)\n", psock, mem->map, mem->flags,
+	  mem->speed, (unsigned long long)mem->res->start,
+		(unsigned long long)mem->res->end, mem->card_start);
     if ((mem->map > 3) || (mem->card_start > 0x3ffffff) ||
 	(mem->res->start > 0xffffff) || (mem->res->end > 0xffffff) ||
 	(mem->res->start > mem->res->end) || (mem->speed > 1000))
drivers/pnp/manager.c renamed from drivers/pnp/manager.c~resource-64bit +8 -4
diff -puN drivers/pnp/manager.c~resource-64bit drivers/pnp/manager.c
@@ -25,7 +25,8 @@ DECLARE_MUTEX(pnp_res_mutex);
 
 static int pnp_assign_port(struct pnp_dev *dev, struct pnp_port *rule, int idx)
 {
-	unsigned long *start, *end, *flags;
+	u64 *start, *end;
+	unsigned long *flags;
 
 	if (!dev || !rule)
 		return -EINVAL;
@@ -68,7 +69,8 @@ static int pnp_assign_port(struct pnp_de
 
 static int pnp_assign_mem(struct pnp_dev *dev, struct pnp_mem *rule, int idx)
 {
-	unsigned long *start, *end, *flags;
+	u64 *start, *end;
+	unsigned long *flags;
 
 	if (!dev || !rule)
 		return -EINVAL;
@@ -121,7 +123,8 @@ static int pnp_assign_mem(struct pnp_dev
 
 static int pnp_assign_irq(struct pnp_dev * dev, struct pnp_irq *rule, int idx)
 {
-	unsigned long *start, *end, *flags;
+	u64 *start, *end;
+	unsigned long *flags;
 	int i;
 
 	/* IRQ priority: this table is good for i386 */
@@ -173,7 +176,8 @@ static int pnp_assign_irq(struct pnp_dev
 
 static int pnp_assign_dma(struct pnp_dev *dev, struct pnp_dma *rule, int idx)
 {
-	unsigned long *start, *end, *flags;
+	u64 *start, *end;
+	unsigned long *flags;
 	int i;
 
 	/* DMA priority: this table is good for i386 */
drivers/pnp/resource.c renamed from drivers/pnp/resource.c~resource-64bit +4 -4
diff -puN drivers/pnp/resource.c~resource-64bit drivers/pnp/resource.c
@@ -242,7 +242,7 @@ int pnp_check_port(struct pnp_dev * dev,
 {
 	int tmp;
 	struct pnp_dev *tdev;
-	unsigned long *port, *end, *tport, *tend;
+	u64 *port, *end, *tport, *tend;
 	port = &dev->res.port_resource[idx].start;
 	end = &dev->res.port_resource[idx].end;
 
@@ -298,7 +298,7 @@ int pnp_check_mem(struct pnp_dev * dev, 
 {
 	int tmp;
 	struct pnp_dev *tdev;
-	unsigned long *addr, *end, *taddr, *tend;
+	u64 *addr, *end, *taddr, *tend;
 	addr = &dev->res.mem_resource[idx].start;
 	end = &dev->res.mem_resource[idx].end;
 
@@ -359,7 +359,7 @@ int pnp_check_irq(struct pnp_dev * dev, 
 {
 	int tmp;
 	struct pnp_dev *tdev;
-	unsigned long * irq = &dev->res.irq_resource[idx].start;
+	u64 * irq = &dev->res.irq_resource[idx].start;
 
 	/* if the resource doesn't exist, don't complain about it */
 	if (cannot_compare(dev->res.irq_resource[idx].flags))
@@ -424,7 +424,7 @@ int pnp_check_dma(struct pnp_dev * dev, 
 #ifndef CONFIG_IA64
 	int tmp;
 	struct pnp_dev *tdev;
-	unsigned long * dma = &dev->res.dma_resource[idx].start;
+	u64 * dma = &dev->res.dma_resource[idx].start;
 
 	/* if the resource doesn't exist, don't complain about it */
 	if (cannot_compare(dev->res.dma_resource[idx].flags))
drivers/scsi/sata_via.c renamed from drivers/scsi/sata_via.c~resource-64bit +3 -3
diff -puN drivers/scsi/sata_via.c~resource-64bit drivers/scsi/sata_via.c
@@ -327,10 +327,10 @@ static int svia_init_one (struct pci_dev
 	for (i = 0; i < ARRAY_SIZE(svia_bar_sizes); i++)
 		if ((pci_resource_start(pdev, i) == 0) ||
 		    (pci_resource_len(pdev, i) < bar_sizes[i])) {
-			printk(KERN_ERR DRV_NAME "(%s): invalid PCI BAR %u (sz 0x%lx, val 0x%lx)\n",
+			printk(KERN_ERR DRV_NAME "(%s): invalid PCI BAR %u (sz 0x%llx, val 0x%llx)\n",
 			       pci_name(pdev), i,
-			       pci_resource_start(pdev, i),
-			       pci_resource_len(pdev, i));
+			       (unsigned long long)pci_resource_start(pdev, i),
+			       (unsigned long long)pci_resource_len(pdev, i));
 			rc = -ENODEV;
 			goto err_out_regions;
 		}
drivers/serial/8250_pci.c renamed from drivers/serial/8250_pci.c~resource-64bit +2 -2
diff -puN drivers/serial/8250_pci.c~resource-64bit drivers/serial/8250_pci.c
@@ -589,8 +589,8 @@ pci_default_setup(struct pci_dev *dev, s
 	else
 		offset += idx * board->uart_offset;
 
-	maxnr = (pci_resource_len(dev, bar) - board->first_offset) /
-		(8 << board->reg_shift);
+	maxnr = (pci_resource_len(dev, bar) - board->first_offset) >>
+		(board->reg_shift + 3);
 
 	if (board->flags & FL_REGION_SZ_CAP && idx >= maxnr)
 		return 1;
include/linux/ioport.h renamed from include/linux/ioport.h~resource-64bit +11 -10
diff -puN include/linux/ioport.h~resource-64bit include/linux/ioport.h
@@ -9,13 +9,14 @@
 #define _LINUX_IOPORT_H
 
 #include <linux/compiler.h>
+#include <linux/types.h>
 /*
  * Resources are tree-like, allowing
  * nesting etc..
  */
 struct resource {
 	const char *name;
-	unsigned long start, end;
+	u64 start, end;
 	unsigned long flags;
 	struct resource *parent, *sibling, *child;
 };
@@ -96,31 +97,31 @@ extern struct resource * ____request_res
 extern int release_resource(struct resource *new);
 extern int insert_resource(struct resource *parent, struct resource *new);
 extern int allocate_resource(struct resource *root, struct resource *new,
-			     unsigned long size,
-			     unsigned long min, unsigned long max,
-			     unsigned long align,
+			     u64 size,
+			     u64 min, u64 max,
+			     u64 align,
 			     void (*alignf)(void *, struct resource *,
 					    unsigned long, unsigned long),
 			     void *alignf_data);
-int adjust_resource(struct resource *res, unsigned long start,
-		    unsigned long size);
+int adjust_resource(struct resource *res, u64 start,
+		    u64 size);
 
 /* Convenience shorthand with allocation */
 #define request_region(start,n,name)	__request_region(&ioport_resource, (start), (n), (name))
 #define request_mem_region(start,n,name) __request_region(&iomem_resource, (start), (n), (name))
 #define rename_region(region, newname) do { (region)->name = (newname); } while (0)
 
-extern struct resource * __request_region(struct resource *, unsigned long start, unsigned long n, const char *name);
+extern struct resource * __request_region(struct resource *, u64 start, u64 n, const char *name);
 
 /* Compatibility cruft */
 #define release_region(start,n)	__release_region(&ioport_resource, (start), (n))
 #define check_mem_region(start,n)	__check_region(&iomem_resource, (start), (n))
 #define release_mem_region(start,n)	__release_region(&iomem_resource, (start), (n))
 
-extern int __check_region(struct resource *, unsigned long, unsigned long);
-extern void __release_region(struct resource *, unsigned long, unsigned long);
+extern int __check_region(struct resource *, u64, u64);
+extern void __release_region(struct resource *, u64, u64);
 
-static inline int __deprecated check_region(unsigned long s, unsigned long n)
+static inline int __deprecated check_region(u64 s, u64 n)
 {
 	return __check_region(&ioport_resource, s, n);
 }
kernel/resource.c renamed from kernel/resource.c~resource-64bit +22 -23

122 lines not on this page · raw message

sound/drivers/mpu401/mpu401.c renamed from sound/drivers/mpu401/mpu401.c~resource-64bit +2 -2
diff -puN sound/drivers/mpu401/mpu401.c~resource-64bit sound/drivers/mpu401/mpu401.c
@@ -126,8 +126,8 @@ static int __init snd_mpu401_pnp(int dev
 		return -ENODEV;
 	}
 	if (pnp_port_len(device, 0) < IO_EXTENT) {
-		snd_printk(KERN_ERR "PnP port length is %ld, expected %d\n",
-			   pnp_port_len(device, 0), IO_EXTENT);
+		snd_printk(KERN_ERR "PnP port length is %lld, expected %d\n",
+			   (unsigned long long)pnp_port_len(device, 0), IO_EXTENT);
 		return -ENODEV;
 	}
 	port[dev] = pnp_port_start(device, 0);
sound/isa/es18xx.c renamed from sound/isa/es18xx.c~resource-64bit +1 -1
diff -puN sound/isa/es18xx.c~resource-64bit sound/isa/es18xx.c
@@ -1938,7 +1938,7 @@ static int __devinit snd_audiodrive_pnp(
 		snd_printk(KERN_ERR PFX "PnP control configure failure (out of resources?)\n");
 		return -EAGAIN;
 	}
-	snd_printdd("pnp: port=0x%lx\n", pnp_port_start(acard->devc, 0));
+	snd_printdd("pnp: port=0x%llx\n", (unsigned long long)pnp_port_start(acard->devc, 0));
 	/* PnP initialization */
 	pdev = acard->dev;
 	pnp_init_resource_table(cfg);
sound/isa/gus/interwave.c renamed from sound/isa/gus/interwave.c~resource-64bit +4 -4
diff -puN sound/isa/gus/interwave.c~resource-64bit sound/isa/gus/interwave.c
@@ -616,10 +616,10 @@ static int __devinit snd_interwave_pnp(i
 	if (dma2[dev] >= 0)
 		dma2[dev] = pnp_dma(pdev, 1);
 	irq[dev] = pnp_irq(pdev, 0);
-	snd_printdd("isapnp IW: sb port=0x%lx, gf1 port=0x%lx, codec port=0x%lx\n",
-				pnp_port_start(pdev, 0),
-				pnp_port_start(pdev, 1),
-				pnp_port_start(pdev, 2));
+	snd_printdd("isapnp IW: sb port=0x%llx, gf1 port=0x%llx, codec port=0x%llx\n",
+				(unsigned long long)pnp_port_start(pdev, 0),
+				(unsigned long long)pnp_port_start(pdev, 1),
+				(unsigned long long)pnp_port_start(pdev, 2));
 	snd_printdd("isapnp IW: dma1=%i, dma2=%i, irq=%i\n", dma1[dev], dma2[dev], irq[dev]);
 #ifdef SNDRV_STB
 	/* Tone Control initialization */
sound/isa/sb/sb16.c renamed from sound/isa/sb/sb16.c~resource-64bit +1 -1
diff -puN sound/isa/sb/sb16.c~resource-64bit sound/isa/sb/sb16.c
@@ -322,7 +322,7 @@ static int __devinit snd_card_sb16_pnp(i
 			goto __wt_error; 
 		} 
 		awe_port[dev] = pnp_port_start(pdev, 0);
-		snd_printdd("pnp SB16: wavetable port=0x%lx\n", pnp_port_start(pdev, 0));
+		snd_printdd("pnp SB16: wavetable port=0x%llx\n", (unsigned long long)pnp_port_start(pdev, 0));
 	} else {
 __wt_error:
 		if (pdev) {
sound/oss/forte.c renamed from sound/oss/forte.c~resource-64bit +3 -2
diff -puN sound/oss/forte.c~resource-64bit sound/oss/forte.c
@@ -2034,8 +2034,9 @@ forte_probe (struct pci_dev *pci_dev, co
 	
 	pci_set_drvdata (pci_dev, chip);
 
-	printk (KERN_INFO PFX "FM801 chip found at 0x%04lX-0x%04lX IRQ %u\n", 
-		chip->iobase, pci_resource_end (pci_dev, 0), chip->irq);
+	printk (KERN_INFO PFX "FM801 chip found at 0x%04lX-0x%04llX IRQ %u\n",
+		chip->iobase, (unsigned long long)pci_resource_end (pci_dev, 0),
+		chip->irq);
 
 	/* Power it up */
 	if ((ret = forte_chip_init (chip)) == 0)
sound/pci/bt87x.c renamed from sound/pci/bt87x.c~resource-64bit +3 -2
diff -puN sound/pci/bt87x.c~resource-64bit sound/pci/bt87x.c
@@ -876,8 +876,9 @@ static int __devinit snd_bt87x_probe(str
 
 	strcpy(card->driver, "Bt87x");
 	sprintf(card->shortname, "Brooktree Bt%x", pci->device);
-	sprintf(card->longname, "%s at %#lx, irq %i",
-		card->shortname, pci_resource_start(pci, 0), chip->irq);
+	sprintf(card->longname, "%s at %#llx, irq %i",
+		card->shortname, (unsigned long long)pci_resource_start(pci, 0),
+		chip->irq);
 	strcpy(card->mixername, "Bt87x");
 
 	err = snd_card_register(card);
sound/pci/sonicvibes.c renamed from sound/pci/sonicvibes.c~resource-64bit +2 -2
diff -puN sound/pci/sonicvibes.c~resource-64bit sound/pci/sonicvibes.c
@@ -1462,10 +1462,10 @@ static int __devinit snd_sonic_probe(str
 
 	strcpy(card->driver, "SonicVibes");
 	strcpy(card->shortname, "S3 SonicVibes");
-	sprintf(card->longname, "%s rev %i at 0x%lx, irq %i",
+	sprintf(card->longname, "%s rev %i at 0x%llx, irq %i",
 		card->shortname,
 		sonic->revision,
-		pci_resource_start(pci, 1),
+		(unsigned long long)pci_resource_start(pci, 1),
 		sonic->irq);
 
 	if ((err = snd_sonicvibes_pcm(sonic, 0, NULL)) < 0) {

_

#2 lore raw

Re: [RFC][PATCH 2/7] Common code for the physmem map

---
This patch provides the arch independent code to create the physmem view in the proc file system. /proc/physmem reflects the entire RAM resources that present in the system. Even if the system has been booted with mem= or memmap= options, this view will reflect the complete physical memory map.

---

---

 linux-2.6.12-rc1-hari/kernel/resource.c |   29 +++++++++++++++++++++++++++++
 1 files changed, 29 insertions(+)
kernel/resource.c renamed from kernel/resource.c~physmem-common +29 -0
diff -puN kernel/resource.c~physmem-common kernel/resource.c
@@ -39,6 +39,15 @@ struct resource iomem_resource = {
 
 EXPORT_SYMBOL(iomem_resource);
 
+struct resource physmem_resource = {
+	.name	= "Phys mem",
+	.start	= 0ULL,
+	.end	= ~0ULL,
+	.flags	= IORESOURCE_MEM,
+};
+
+EXPORT_SYMBOL(physmem_resource);
+
 static DEFINE_RWLOCK(resource_lock);
 
 #ifdef CONFIG_PROC_FS
@@ -118,6 +127,16 @@ static int iomem_open(struct inode *inod
 	return res;
 }
 
+static int physmem_open(struct inode *inode, struct file *file)
+{
+	int res = seq_open(file, &resource_op);
+	if (!res) {
+		struct seq_file *m = file->private_data;
+		m->private = &physmem_resource;
+	}
+	return res;
+}
+
 static struct file_operations proc_ioports_operations = {
 	.open		= ioports_open,
 	.read		= seq_read,
@@ -132,6 +151,13 @@ static struct file_operations proc_iomem
 	.release	= seq_release,
 };
 
+static struct file_operations proc_physmem_operations = {
+	.open		= physmem_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= seq_release,
+};
+
 static int __init ioresources_init(void)
 {
 	struct proc_dir_entry *entry;
@@ -142,6 +168,9 @@ static int __init ioresources_init(void)
 	entry = create_proc_entry("iomem", 0, NULL);
 	if (entry)
 		entry->proc_fops = &proc_iomem_operations;
+	entry = create_proc_entry("physmem", 0, NULL);
+	if (entry)
+		entry->proc_fops = &proc_physmem_operations;
 	return 0;
 }
 __initcall(ioresources_init);

_

#3 lore raw

Re: [RFC][PATCH 3/7] i386 code for the physmem map

Regards, Hari

---

This patch contains the i386 specific code to generate the /proc/physmem view.

---

 linux-2.6.12-rc1-hari/arch/i386/kernel/efi.c   |  134 +++++++++++++++----------
 linux-2.6.12-rc1-hari/arch/i386/kernel/setup.c |   82 +++++++++++++--
 linux-2.6.12-rc1-hari/include/linux/efi.h      |    2 
 linux-2.6.12-rc1-hari/include/linux/ioport.h   |    1 
 4 files changed, 156 insertions(+), 63 deletions(-)
arch/i386/kernel/efi.c renamed from arch/i386/kernel/efi.c~physmem-i386 +83 -51
diff -puN arch/i386/kernel/efi.c~physmem-i386 arch/i386/kernel/efi.c
@@ -49,6 +49,7 @@ struct efi efi;
 EXPORT_SYMBOL(efi);
 static struct efi efi_phys __initdata;
 struct efi_memory_map memmap __initdata;
+struct efi_memory_map memmapcopy __initdata;
 
 /*
  * We require an early boot_ioremap mapping mechanism initially
@@ -527,6 +528,67 @@ void __init efi_enter_virtual_mode(void)
 					efi.systab->runtime->reset_system;
 }
 
+static struct resource * __init alloc_efi_resource(efi_memory_desc_t *md,
+						   struct resource *resource,
+						   int i)
+{
+	struct resource *res;
+
+	res = alloc_bootmem_low(sizeof(struct resource));
+	switch (md->type) {
+	case EFI_RESERVED_TYPE:
+		res->name = "Reserved Memory";
+		break;
+	case EFI_LOADER_CODE:
+		res->name = "Loader Code";
+		break;
+	case EFI_LOADER_DATA:
+		res->name = "Loader Data";
+		break;
+	case EFI_BOOT_SERVICES_DATA:
+		res->name = "BootServices Data";
+		break;
+	case EFI_BOOT_SERVICES_CODE:
+		res->name = "BootServices Code";
+		break;
+	case EFI_RUNTIME_SERVICES_CODE:
+		res->name = "Runtime Service Code";
+		break;
+	case EFI_RUNTIME_SERVICES_DATA:
+		res->name = "Runtime Service Data";
+		break;
+	case EFI_CONVENTIONAL_MEMORY:
+		res->name = "Conventional Memory";
+		break;
+	case EFI_UNUSABLE_MEMORY:
+		res->name = "Unusable Memory";
+		break;
+	case EFI_ACPI_RECLAIM_MEMORY:
+		res->name = "ACPI Reclaim";
+		break;
+	case EFI_ACPI_MEMORY_NVS:
+		res->name = "ACPI NVS";
+		break;
+	case EFI_MEMORY_MAPPED_IO:
+		res->name = "Memory Mapped IO";
+		break;
+	case EFI_MEMORY_MAPPED_IO_PORT_SPACE:
+		res->name = "Memory Mapped IO Port Space";
+		break;
+	default:
+		res->name = "Reserved";
+		break;
+	}
+
+	res->start = md->phys_addr;
+	res->end = res->start + ((md->num_pages << EFI_PAGE_SHIFT) - 1);
+	res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
+	if (request_resource(resource, res) < 0)
+		printk(KERN_ERR PFX "Failed to allocate res %s : 0x%llx-0x%llx\n",
+			res->name, res->start, res->end);
+	return res;
+}
+
 void __init
 efi_initialize_iomem_resources(struct resource *code_resource,
 			       struct resource *data_resource)
@@ -541,57 +603,7 @@ efi_initialize_iomem_resources(struct re
 		if ((md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT)) >
 		    0x100000000ULL)
 			continue;
-		res = alloc_bootmem_low(sizeof(struct resource));
-		switch (md->type) {
-		case EFI_RESERVED_TYPE:
-			res->name = "Reserved Memory";
-			break;
-		case EFI_LOADER_CODE:
-			res->name = "Loader Code";
-			break;
-		case EFI_LOADER_DATA:
-			res->name = "Loader Data";
-			break;
-		case EFI_BOOT_SERVICES_DATA:
-			res->name = "BootServices Data";
-			break;
-		case EFI_BOOT_SERVICES_CODE:
-			res->name = "BootServices Code";
-			break;
-		case EFI_RUNTIME_SERVICES_CODE:
-			res->name = "Runtime Service Code";
-			break;
-		case EFI_RUNTIME_SERVICES_DATA:
-			res->name = "Runtime Service Data";
-			break;
-		case EFI_CONVENTIONAL_MEMORY:
-			res->name = "Conventional Memory";
-			break;
-		case EFI_UNUSABLE_MEMORY:
-			res->name = "Unusable Memory";
-			break;
-		case EFI_ACPI_RECLAIM_MEMORY:
-			res->name = "ACPI Reclaim";
-			break;
-		case EFI_ACPI_MEMORY_NVS:
-			res->name = "ACPI NVS";
-			break;
-		case EFI_MEMORY_MAPPED_IO:
-			res->name = "Memory Mapped IO";
-			break;
-		case EFI_MEMORY_MAPPED_IO_PORT_SPACE:
-			res->name = "Memory Mapped IO Port Space";
-			break;
-		default:
-			res->name = "Reserved";
-			break;
-		}
-		res->start = md->phys_addr;
-		res->end = res->start + ((md->num_pages << EFI_PAGE_SHIFT) - 1);
-		res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
-		if (request_resource(&iomem_resource, res) < 0)
-			printk(KERN_ERR PFX "Failed to allocate res %s : 0x%llx-0x%llx\n",
-				res->name, res->start, res->end);
+		res = alloc_efi_resource(md, &iomem_resource, i);
 		/*
 		 * We don't know which region contains kernel data so we try
 		 * it repeatedly and let the resource manager test it.
@@ -606,6 +618,26 @@ efi_initialize_iomem_resources(struct re
 	}
 }
 
+void __init efi_initialize_physmem_resources(void)
+{
+	struct resource *res;
+	efi_memory_desc_t *md;
+	int i;
+
+	for (i = 0; i < memmapcopy.nr_map; i++) {
+		md = &memmapcopy.map[i];
+		res = alloc_efi_resource(md, &physmem_resource, i);
+	}
+}
+
+/*
+ * Make a copy of memmap for creating the physmem resources later.
+ */
+void __init preserve_memmap(void)
+{
+	memcpy(&memmapcopy, &memmap, sizeof(memmap));
+}
+
 /*
  * Convenience functions to obtain memory types and attributes
  */
arch/i386/kernel/setup.c renamed from arch/i386/kernel/setup.c~physmem-i386 +70 -12

141 lines not on this page · raw message

include/linux/efi.h renamed from include/linux/efi.h~physmem-i386 +2 -0
diff -puN include/linux/efi.h~physmem-i386 include/linux/efi.h
@@ -301,6 +301,8 @@ extern u64 efi_mem_attributes (unsigned 
 extern int __init efi_uart_console_only (void);
 extern void efi_initialize_iomem_resources(struct resource *code_resource,
 					struct resource *data_resource);
+extern void efi_initialize_physmem_resources(void);
+extern void preserve_memmap(void);
 extern unsigned long __init efi_get_time(void);
 extern int __init efi_set_rtc_mmss(unsigned long nowtime);
 extern struct efi_memory_map memmap;
include/linux/ioport.h renamed from include/linux/ioport.h~physmem-i386 +1 -0
diff -puN include/linux/ioport.h~physmem-i386 include/linux/ioport.h
@@ -91,6 +91,7 @@ struct resource_list {
 /* PC/ISA/whatever - the normal PC address spaces: IO and memory */
 extern struct resource ioport_resource;
 extern struct resource iomem_resource;
+extern struct resource physmem_resource;
 
 extern int request_resource(struct resource *root, struct resource *new);
 extern struct resource * ____request_resource(struct resource *root, struct resource *new);

_

#4 lore raw

Re: [RFC][PATCH 4/7] x86_64 code for the physmem map

Regards, Hari

---

This patch contains the x86_64 specific code to generate the /proc/physmem view.
---

---

 linux-2.6.12-rc1-hari/arch/x86_64/kernel/e820.c |   55 +++++++++++++++++-------
 1 files changed, 39 insertions(+), 16 deletions(-)
arch/x86_64/kernel/e820.c renamed from arch/x86_64/kernel/e820.c~physmem-x8664 +39 -16
diff -puN arch/x86_64/kernel/e820.c~physmem-x8664 arch/x86_64/kernel/e820.c
@@ -178,25 +178,30 @@ unsigned long __init e820_end_of_ram(voi
 	return end_pfn;	
 }
 
-/* 
- * Mark e820 reserved areas as busy for the resource manager.
- */
-void __init e820_reserve_resources(void)
+static struct resource * __init alloc_e820_resource(struct resource *resource, int i)
+{
+	struct resource *res;
+
+	res = alloc_bootmem_low(sizeof(struct resource));
+	switch (e820.map[i].type) {
+	case E820_RAM:	res->name = "System RAM"; break;
+	case E820_ACPI:	res->name = "ACPI Tables"; break;
+	case E820_NVS:	res->name = "ACPI Non-volatile Storage"; break;
+	default:	res->name = "reserved";
+	}
+	res->start = e820.map[i].addr;
+	res->end = res->start + e820.map[i].size - 1;
+	res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
+	request_resource(resource, res);
+	return res;
+}
+
+static void __init e820_reserve_iomem_resources(void)
 {
 	int i;
+	struct resource *res;
 	for (i = 0; i < e820.nr_map; i++) {
-		struct resource *res;
-		res = alloc_bootmem_low(sizeof(struct resource));
-		switch (e820.map[i].type) {
-		case E820_RAM:	res->name = "System RAM"; break;
-		case E820_ACPI:	res->name = "ACPI Tables"; break;
-		case E820_NVS:	res->name = "ACPI Non-volatile Storage"; break;
-		default:	res->name = "reserved";
-		}
-		res->start = e820.map[i].addr;
-		res->end = res->start + e820.map[i].size - 1;
-		res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
-		request_resource(&iomem_resource, res);
+		res = alloc_e820_resource(&iomem_resource, i);
 		if (e820.map[i].type == E820_RAM) {
 			/*
 			 *  We don't know which RAM region contains kernel data,
@@ -212,6 +217,24 @@ void __init e820_reserve_resources(void)
 	}
 }
 
+static void __init e820_reserve_physmem_resources(void)
+{
+	int i;
+	struct resource *res;
+	for (i = 0; i < e820.nr_map; i++) {
+		res = alloc_e820_resource(&physmem_resource, i);
+	}
+}
+
+/*
+ * Mark e820 reserved areas as busy for the resource manager.
+ */
+void __init e820_reserve_resources(void)
+{
+	e820_reserve_iomem_resources();
+	e820_reserve_physmem_resources();
+}
+
 /* 
  * Add a memory region to the kernel e820 map.
  */ 

_

#5 lore raw

Re: [RFC][PATCH 5/7] Common code for the activemem map

Regards, Hari

---

This patch provides the arch independent code to create the activemem view in the proc file system. /proc/activemem reflects the RAM resources that are in use by the kernel. If the system has been booted with mem= or memmap= options, then this view will reflect the truncated map.
---

---

 linux-2.6.12-rc1-hari/kernel/resource.c |   29 +++++++++++++++++++++++++++++
 1 files changed, 29 insertions(+)
kernel/resource.c renamed from kernel/resource.c~activemem-common +29 -0
diff -puN kernel/resource.c~activemem-common kernel/resource.c
@@ -48,6 +48,15 @@ struct resource physmem_resource = {
 
 EXPORT_SYMBOL(physmem_resource);
 
+struct resource activemem_resource = {
+	.name	= "Active mem",
+	.start	= 0ULL,
+	.end	= ~0ULL,
+	.flags	= IORESOURCE_MEM,
+};
+
+EXPORT_SYMBOL(activemem_resource);
+
 static DEFINE_RWLOCK(resource_lock);
 
 #ifdef CONFIG_PROC_FS
@@ -137,6 +146,16 @@ static int physmem_open(struct inode *in
 	return res;
 }
 
+static int activemem_open(struct inode *inode, struct file *file)
+{
+	int res = seq_open(file, &resource_op);
+	if (!res) {
+		struct seq_file *m = file->private_data;
+		m->private = &activemem_resource;
+	}
+	return res;
+}
+
 static struct file_operations proc_ioports_operations = {
 	.open		= ioports_open,
 	.read		= seq_read,
@@ -158,6 +177,13 @@ static struct file_operations proc_physm
 	.release	= seq_release,
 };
 
+static struct file_operations proc_activemem_operations = {
+	.open		= activemem_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= seq_release,
+};
+
 static int __init ioresources_init(void)
 {
 	struct proc_dir_entry *entry;
@@ -171,6 +197,9 @@ static int __init ioresources_init(void)
 	entry = create_proc_entry("physmem", 0, NULL);
 	if (entry)
 		entry->proc_fops = &proc_physmem_operations;
+	entry = create_proc_entry("activemem", 0, NULL);
+	if (entry)
+		entry->proc_fops = &proc_activemem_operations;
 	return 0;
 }
 __initcall(ioresources_init);

_

#6 lore raw

Re: [RFC][PATCH 6/7] i386 code for the activemem map

Regards, Hari

---

This patch contains the i386 specific code to generate the /proc/activemem view.
---

---

 linux-2.6.12-rc1-hari/arch/i386/kernel/efi.c   |   31 +++++++++++++++
 linux-2.6.12-rc1-hari/arch/i386/kernel/setup.c |   50 +++++++++++++++++++++++++
 linux-2.6.12-rc1-hari/include/linux/efi.h      |    2 +
 linux-2.6.12-rc1-hari/include/linux/ioport.h   |    1 
 4 files changed, 84 insertions(+)
arch/i386/kernel/efi.c renamed from arch/i386/kernel/efi.c~activemem-i386 +31 -0
diff -puN arch/i386/kernel/efi.c~activemem-i386 arch/i386/kernel/efi.c
@@ -630,6 +630,37 @@ void __init efi_initialize_physmem_resou
 	}
 }
 
+void __init efi_initialize_activemem_resources(struct resource *code_resource,
+					       struct resource *data_resource)
+{
+	struct resource *res, *code_resource_copy, *data_resource_copy;
+	efi_memory_desc_t *md;
+	int i;
+#ifdef CONFIG_KEXEC
+	struct resource *crashk_res_copy;
+
+	crashk_res_copy = alloc_bootmem_low(sizeof(struct resource));
+	memcpy(crashk_res_copy, &crashk_res, sizeof(struct resource));
+#endif
+
+	code_resource_copy = alloc_bootmem_low(2*sizeof(struct resource));
+	data_resource_copy = code_resource_copy + sizeof(struct resource);
+	memcpy(code_resource_copy, code_resource, sizeof(struct resource));
+	memcpy(data_resource_copy, data_resource, sizeof(struct resource));
+
+	for (i = 0; i < memmap.nr_map; i++) {
+		md = &memmap.map[i];
+		res = alloc_efi_resource(md, &activemem_resource, i);
+		if (md->type == EFI_CONVENTIONAL_MEMORY) {
+			request_resource(res, code_resource_copy);
+			request_resource(res, data_resource_copy);
+#ifdef CONFIG_KEXEC
+			request_resource(res, crashk_res_copy);
+#endif
+		}
+	}
+}
+
 /*
  * Make a copy of memmap for creating the physmem resources later.
  */
arch/i386/kernel/setup.c renamed from arch/i386/kernel/setup.c~activemem-i386 +50 -0

68 lines not on this page · raw message

include/linux/efi.h renamed from include/linux/efi.h~activemem-i386 +2 -0
diff -puN include/linux/efi.h~activemem-i386 include/linux/efi.h
@@ -302,6 +302,8 @@ extern int __init efi_uart_console_only 
 extern void efi_initialize_iomem_resources(struct resource *code_resource,
 					struct resource *data_resource);
 extern void efi_initialize_physmem_resources(void);
+extern void efi_initialize_activemem_resources(struct resource *code_resource,
+					struct resource *data_resource);
 extern void preserve_memmap(void);
 extern unsigned long __init efi_get_time(void);
 extern int __init efi_set_rtc_mmss(unsigned long nowtime);
include/linux/ioport.h renamed from include/linux/ioport.h~activemem-i386 +1 -0
diff -puN include/linux/ioport.h~activemem-i386 include/linux/ioport.h
@@ -92,6 +92,7 @@ struct resource_list {
 extern struct resource ioport_resource;
 extern struct resource iomem_resource;
 extern struct resource physmem_resource;
+extern struct resource activemem_resource;
 
 extern int request_resource(struct resource *root, struct resource *new);
 extern struct resource * ____request_resource(struct resource *root, struct resource *new);

_

#7 lore raw

Re: [RFC][PATCH 7/7] x86_64 code for the activemem map

Regards, Hari

---

This patch contains the x86_64 specific code to generate the /proc/activemem view.
---

---

 linux-2.6.12-rc1-hari/arch/x86_64/kernel/e820.c |   37 ++++++++++++++++++++++++
 1 files changed, 37 insertions(+)
arch/x86_64/kernel/e820.c renamed from arch/x86_64/kernel/e820.c~activemem-x8664 +37 -0
diff -puN arch/x86_64/kernel/e820.c~activemem-x8664 arch/x86_64/kernel/e820.c
@@ -226,6 +226,42 @@ static void __init e820_reserve_physmem_
 	}
 }
 
+static void __init e820_reserve_activemem_resources(void)
+{
+	int i;
+	struct resource *res, *code_resource_copy, *data_resource_copy;
+#ifdef CONFIG_KEXEC
+	struct resource *crashk_res_copy;
+
+	crashk_res_copy = alloc_bootmem_low(sizeof(struct resource));
+	memcpy(crashk_res_copy, &crashk_res, sizeof(struct resource));
+#endif
+
+	code_resource_copy = alloc_bootmem_low(2*sizeof(struct resource));
+	data_resource_copy = code_resource_copy + sizeof(struct resource);
+	memcpy(code_resource_copy, &code_resource, sizeof(struct resource));
+	memcpy(data_resource_copy, &data_resource, sizeof(struct resource));
+
+	for (i = 0; i < e820.nr_map; i++) {
+		res = alloc_e820_resource(&activemem_resource, i);
+		if (e820.map[i].type == E820_RAM) {
+			request_resource(res, code_resource_copy);
+			request_resource(res, data_resource_copy);
+#ifdef CONFIG_KEXEC
+			request_resource(res, crashk_res_copy);
+#endif
+		}
+		/* If the system has booted with less memory, reflect
+		 * only those entries
+		 */
+		if ((e820.map[i].addr + e820.map[i].size) > (end_user_pfn << PAGE_SHIFT)) {
+			/* Trim the last entry to reflect the actual pfn allowed */
+			res->end = (end_user_pfn << PAGE_SHIFT) - 1;
+			break;
+		}
+	}
+}
+
 /*
  * Mark e820 reserved areas as busy for the resource manager.
  */
@@ -233,6 +269,7 @@ void __init e820_reserve_resources(void)
 {
 	e820_reserve_iomem_resources();
 	e820_reserve_physmem_resources();
+	e820_reserve_activemem_resources();
 }
 
 /* 

_

#8 lore raw
In reply to Hariprasad Nellitheertha
Quoted 3 lines · Hariprasad Nellitheertha

The topic of creating a common interface across architectures for obtaining system RAM information has been discussed on lkml and fastboot for a while now.

Sorry, I missed this on LKML.

Quoted 4 lines

Kexec needs
information about the entire physical RAM present in the system while kdump needs information on the memory that the kernel has booted with.

I think there's likely a lot of commonality with the needs of memory hotplug systems here. We effectively dump out the physical layout of the system, but in sysfs. We do this mostly because any memory hotplug changes generate hotplug events, just like all other hardware. If you do this in /proc, it's another thing that memory hotplug will have to update.

Also, we already have a concept of active and non-active physical
memory: we call it online and offline. Some tweaks to the information that we export might be all that you need, instead of creating a new interface. I've attached a document I started writing a couple days ago about the sysfs layout and the call paths for hotplug. It's horribly incomplete, but not a bad start.

If you want to see some more details of the layout, please check out this patch set:

http://www.sr71.net/patches/2.6.12/2.6.12-rc1-mhp1/patch-2.6.12-rc1-mhp1.gz

A good example of all of the hotplug stuff enabled for a normal machine is this .config, it boots on my 4-way PIII Xeon.

http://www.sr71.net/patches/2.6.12/2.6.12-rc1-mhp1/configs/config-i386-sparse-hotplug

You're welcome to borrow the machine that I normally boot this config on. Should make booting it relatively foolproof. :)

-- Dave

block_size_bytes: The size of each memory section (in hex)

memoryXXXX: Each directory contains individual information about
that memory section. The number associated with the memory section is arbitrary, and should not be used for any calculations, it is a useless tag.

memoryXXXX/phys_index: multiply this by block_size_bytes and you
can calculate the starting physical address of this memory section. This is so that you can directly correlate the memoryXXXX directory with any probe that you've issued.

probe: The user may write a physical address into this file to
initate a hot-addition of a new memory section.

echo 0x00000000 > /sys/devices/system/memory/probe memory_probe_store()
add_memory()

	  __add_pages()
	  __add_section()
		allocate section memmap
		lock zone
		register_new_memory()
			add_memory_block()
			register_memory()
				create kobject -> generates hotplug event
			sysfs create files()
		unlock zone

/sbin/hotplug sees "ADD" event, calls:

#!/bin/sh

# goes in /etc/hotplug.d/memory/onlinenow.hotplug
if [ "$ACTION" != "add" ]; then
	exit 0;
fi
while ! [ -f $SYSFS/$DEVPATH/state ]; do
	sleep 0.1;
done
echo online > "$SYSFS/$DEVPATH/state"

store_mem_state()

	memory_block_change_state()
		memory_block_action()
		|------>online_pages()
		|		online_page()
		|			__free_page()
		\------>remove_memory()
				__remove_pages()
					capture_page_range()
					unregister_memory_section()
						-> generates unplug event
#9 lore raw
Quoted 2 lines · Dave Hansen

On Thu, 2005-03-24 at 11:19 +0530, Hariprasad Nellitheertha wrote:

...

Quoted 6 lines

I think there's likely a lot of commonality with the needs of memory hotplug systems here. We effectively dump out the physical layout of the system, but in sysfs. We do this mostly because any memory hotplug changes generate hotplug events, just like all other hardware. If you do this in /proc, it's another thing that memory hotplug will have to update.

We put it in /proc primarily because what we wanted was similar in many ways to /proc/iomem and so we (re)use a bit of the code. Also, we were wondering if it is appropriate to put in multiple values in a single file in sysfs.

Quoted 5 lines

Also, we already have a concept of active and non-active physical
memory: we call it online and offline. Some tweaks to the information that we export might be all that you need, instead of creating a new interface.

Looks like. And the tweaks could be handled by the user space kexec-tools.

I've attached a document I started writing a couple days ago

Quoted 7 lines

about the sysfs layout and the call paths for hotplug. It's horribly incomplete, but not a bad start.

If you want to see some more details of the layout, please check out this patch set:

http://www.sr71.net/patches/2.6.12/2.6.12-rc1-mhp1/patch-2.6.12-rc1-mhp1.gz

This does not have the sysfs related code. Is there a separate patch for adding the sysfs entries?

Quoted 9 lines

A good example of all of the hotplug stuff enabled for a normal machine is this .config, it boots on my 4-way PIII Xeon.

http://www.sr71.net/patches/2.6.12/2.6.12-rc1-mhp1/configs/config-i386-sparse-hotplug

You're welcome to borrow the machine that I normally boot this config on. Should make booting it relatively foolproof. :)

Quoted 6 lines

-- Dave

------------------------------------------------------------------------

block_size_bytes: The size of each memory section (in hex)

This value is per memoryXXXX directory, right?

Regards, Hari

#10 lore raw
Quoted 11 lines · Hariprasad Nellitheertha
Quoted 6 lines · Dave Hansen

I think there's likely a lot of commonality with the needs of memory hotplug systems here. We effectively dump out the physical layout of the system, but in sysfs. We do this mostly because any memory hotplug changes generate hotplug events, just like all other hardware. If you do this in /proc, it's another thing that memory hotplug will have to update.

We put it in /proc primarily because what we wanted was similar in many ways to /proc/iomem and so we (re)use a bit of the code.

The code reuse is nice, but the expanded use of /proc is not.

Quoted 2 lines

Also, we were wondering if it is appropriate to put in multiple values in a single file in sysfs.

Why would you need to do that?

Quoted 11 lines

I've attached a document I started writing a couple days ago

Quoted 7 lines

about the sysfs layout and the call paths for hotplug. It's horribly incomplete, but not a bad start.

If you want to see some more details of the layout, please check out this patch set:

http://www.sr71.net/patches/2.6.12/2.6.12-rc1-mhp1/patch-2.6.12-rc1-mhp1.gz

This does not have the sysfs related code. Is there a separate patch for adding the sysfs entries?

Hmmm. I think my rollup script broke. Try this:

http://www.sr71.net/patches/2.6.12/2.6.12-rc1-mhp1/broken-out/L0-sysfs-memory-class.patch

Quoted 3 lines
Quoted 1 line

block_size_bytes: The size of each memory section (in hex)

This value is per memoryXXXX directory, right?

No, it's global. However, we have discussed doing it per-section in the future to collapse some of the contiguous areas into a single directory.

-- Dave

#11 lore raw
Quoted 14 lines · Dave Hansen
Quoted 10 lines · Hariprasad Nellitheertha
Quoted 6 lines · Dave Hansen

I think there's likely a lot of commonality with the needs of memory hotplug systems here. We effectively dump out the physical layout of the system, but in sysfs. We do this mostly because any memory hotplug changes generate hotplug events, just like all other hardware. If you do this in /proc, it's another thing that memory hotplug will have to update.

We put it in /proc primarily because what we wanted was similar in many ways to /proc/iomem and so we (re)use a bit of the code.

Quoted 7 lines

The code reuse is nice, but the expanded use of /proc is not.

Quoted 1 line

Also, we were wondering if it is appropriate to put in multiple values in a single file in sysfs.

Why would you need to do that?

Because we are putting the starting address, end address and the memory type against each entry (just like in /proc/iomem). Of course, we can figure out the ending address knowing the starting address and the section size.

Quoted 9 lines
Quoted 9 lines

I've attached a document I started writing a couple days ago

Quoted 6 lines

about the sysfs layout and the call paths for hotplug. It's horribly incomplete, but not a bad start.

If you want to see some more details of the layout, please check out this patch set:

Quoted 8 lines
Quoted 3 lines
Quoted 1 line

http://www.sr71.net/patches/2.6.12/2.6.12-rc1-mhp1/patch-2.6.12-rc1-mhp1.gz

This does not have the sysfs related code. Is there a separate patch for adding the sysfs entries?

Hmmm. I think my rollup script broke. Try this:

http://www.sr71.net/patches/2.6.12/2.6.12-rc1-mhp1/broken-out/L0-sysfs-memory-class.patch

In addition to this, I also needed to pull-in the J-zone_resize_sem.patch to get it to compile.

Would it be possible to make this a separate patch-set so that it does not depend on memory hotplug.

Quoted 9 lines
Quoted 5 lines
Quoted 3 lines

block_size_bytes: The size of each memory section (in hex)

This value is per memoryXXXX directory, right?

No, it's global. However, we have discussed doing it per-section in the future to collapse some of the contiguous areas into a single directory.

I tested this on my PIII 256M machine. /sys/devices/system/memory showed 4 memory sections each of size 64MB. There are a couple of issues that we noticed. We will not be able to spot those physical memory areas which the OS does not use (such as the region between 640k and 1MB). Also, when I booted the system with the mem=100M option, two entries (memory0 and memory1) turned up. With block_size_bytes being 64M, this turns out equivalent to a system with 128M memory.

If block_size_bytes was per-directory, it would be easier in such situations.

Regards, Hari

#12 lore raw
Quoted 7 lines · Hariprasad Nellitheertha
Quoted 6 lines · Dave Hansen

On Thu, 2005-03-24 at 15:49 +0530, Hariprasad Nellitheertha wrote:
The code reuse is nice, but the expanded use of /proc is not.

Quoted 2 lines

Also, we were wondering if it is appropriate to
put in multiple values in a single file in sysfs.

Quoted 6 lines
Quoted 1 line

Why would you need to do that?

Because we are putting the starting address, end address and the memory type against each entry (just like in /proc/iomem). Of course, we can figure out the ending address knowing the starting address and the section size.

That sounds like what you *want* and not what you need :)

Quoted 7 lines
Quoted 1 line

http://www.sr71.net/patches/2.6.12/2.6.12-rc1-mhp1/broken-out/L0-sysfs-memory-class.patch

In addition to this, I also needed to pull-in the J-zone_resize_sem.patch to get it to compile.

Would it be possible to make this a separate patch-set so that it does not depend on memory hotplug.

Yes, it's quite possible. However, I've already done this for the page-migration patches, and I'm not looking forward to doing it again. If it was as simple as you describe, is there a real reason to break it out?

Quoted 9 lines

I tested this on my PIII 256M machine. /sys/devices/system/memory showed 4 memory sections each of size 64MB. There are a couple of issues that we noticed. We will not be able to spot those physical memory areas which the OS does not use (such as the region between 640k and 1MB). Also, when I booted the system with the mem=100M option, two entries (memory0 and memory1) turned up. With block_size_bytes being 64M, this turns out equivalent to a system with 128M memory.

This turns out to be a minor issue for memory hotplug systems as well, because it means that you can't add back that last 28MB of memory, either.

Quoted 2 lines

If block_size_bytes was per-directory, it would be easier in such situations.

First of all, I think there are lots of solutions to the problem, not just changing the scope of "block_size_bytes". We could also present a value inside of each file that represents which pages in that memory section area actually online and real RAM. That could be generated from (slowly) from hardware information like the e820 table. It could be very slow because the only users would be swsusp and kexec, which aren't performance-critical.

Also, having variably-sized sysfs objects presents some serious obstacles for hotplug memory. A memory remove could involve splitting existing memory areas, and lots of small additions could involve merging memory ares, just like VMAs.

We haven't implemented either of these things yet, because it hasn't been necessary, and we don't want to bloat the code. However, if there's another user, it's a reason to go do it now. Also, it may be a good idea to move block_size_bytes into the memoryXX directory now, just in case we need to change it later.

-- Dave