aboutsummaryrefslogtreecommitdiffstats
path: root/arch
diff options
context:
space:
mode:
Diffstat (limited to 'arch')
-rw-r--r--arch/alpha/Kconfig1
-rw-r--r--arch/arm/Kconfig19
-rw-r--r--arch/avr32/Kconfig1
-rw-r--r--arch/blackfin/Kconfig1
-rw-r--r--arch/cris/Kconfig1
-rw-r--r--arch/frv/Kconfig1
-rw-r--r--arch/h8300/Kconfig1
-rw-r--r--arch/ia64/Kconfig1
-rw-r--r--arch/ia64/pci/pci.c25
-rw-r--r--arch/ia64/sn/pci/tioce_provider.c16
-rw-r--r--arch/m32r/Kconfig1
-rw-r--r--arch/m68k/Kconfig1
-rw-r--r--arch/m68knommu/Kconfig1
-rw-r--r--arch/mips/Kconfig1
-rw-r--r--arch/parisc/Kconfig1
-rw-r--r--arch/powerpc/Kconfig1
-rw-r--r--arch/ppc/Kconfig1
-rw-r--r--arch/s390/defconfig87
-rw-r--r--arch/s390/kernel/binfmt_elf32.c11
-rw-r--r--arch/s390/kernel/compat_wrapper.S20
-rw-r--r--arch/s390/kernel/process.c1
-rw-r--r--arch/s390/kernel/setup.c1
-rw-r--r--arch/s390/kernel/syscalls.S3
-rw-r--r--arch/s390/kernel/traps.c3
-rw-r--r--arch/s390/mm/fault.c40
-rw-r--r--arch/s390/mm/init.c3
-rw-r--r--arch/s390/mm/mmap.c65
-rw-r--r--arch/s390/mm/pgtable.c176
-rw-r--r--arch/s390/mm/vmem.c28
-rw-r--r--arch/sh/Kconfig1
-rw-r--r--arch/sparc/Kconfig1
-rw-r--r--arch/sparc/kernel/errtbls.c144
-rw-r--r--arch/sparc/kernel/of_device.c27
-rw-r--r--arch/sparc/kernel/vmlinux.lds.S4
-rw-r--r--arch/sparc64/Kconfig5
-rw-r--r--arch/sparc64/kernel/iommu.c352
-rw-r--r--arch/sparc64/kernel/iommu_common.h33
-rw-r--r--arch/sparc64/kernel/kprobes.c113
-rw-r--r--arch/sparc64/kernel/of_device.c26
-rw-r--r--arch/sparc64/kernel/pci_sun4v.c278
-rw-r--r--arch/v850/Kconfig1
-rw-r--r--arch/x86/Kconfig4
-rw-r--r--arch/x86/Kconfig.debug6
-rw-r--r--arch/x86/Makefile4
-rw-r--r--arch/x86/boot/.gitignore2
-rw-r--r--arch/x86/boot/printf.c24
-rw-r--r--arch/x86/configs/i386_defconfig1
-rw-r--r--arch/x86/configs/x86_64_defconfig1
-rw-r--r--arch/x86/kernel/Makefile2
-rw-r--r--arch/x86/kernel/cpu/mcheck/therm_throt.c2
-rw-r--r--arch/x86/kernel/entry_32.S15
-rw-r--r--arch/x86/kernel/entry_64.S18
-rw-r--r--arch/x86/kernel/geode_32.c5
-rw-r--r--arch/x86/kernel/head_32.S151
-rw-r--r--arch/x86/kernel/mfgpt_32.c123
-rw-r--r--arch/x86/kernel/quirks.c2
-rw-r--r--arch/x86/kernel/setup_32.c4
-rw-r--r--arch/x86/kernel/topology.c2
-rw-r--r--arch/x86/mm/init_32.c74
-rw-r--r--arch/x86/mm/init_64.c8
-rw-r--r--arch/x86/mm/ioremap.c55
-rw-r--r--arch/x86/mm/pageattr.c140
-rw-r--r--arch/x86/pci/common.c25
-rw-r--r--arch/x86/pci/direct.c4
-rw-r--r--arch/x86/pci/fixup.c6
-rw-r--r--arch/x86/pci/legacy.c2
-rw-r--r--arch/x86/pci/mmconfig-shared.c41
-rw-r--r--arch/x86/pci/mmconfig_32.c20
-rw-r--r--arch/x86/pci/mmconfig_64.c18
-rw-r--r--arch/x86/pci/pci.h22
-rw-r--r--arch/x86/pci/visws.c3
-rw-r--r--arch/x86/power/Makefile4
-rw-r--r--arch/x86/power/cpu_32.c (renamed from arch/x86/power/cpu.c)2
-rw-r--r--arch/x86/power/cpu_64.c (renamed from arch/x86/kernel/suspend_64.c)160
-rw-r--r--arch/x86/power/hibernate_32.c (renamed from arch/x86/power/suspend.c)6
-rw-r--r--arch/x86/power/hibernate_64.c169
-rw-r--r--arch/x86/power/hibernate_asm_32.S (renamed from arch/x86/power/swsusp.S)3
-rw-r--r--arch/x86/power/hibernate_asm_64.S (renamed from arch/x86/kernel/suspend_asm_64.S)9
-rw-r--r--arch/x86/xen/mmu.c6
-rw-r--r--arch/x86/xen/time.c10
-rw-r--r--arch/xtensa/Kconfig1
81 files changed, 1624 insertions, 1027 deletions
diff --git a/arch/alpha/Kconfig b/arch/alpha/Kconfig
index 5b7dcd5a0e75..002703b8c0b0 100644
--- a/arch/alpha/Kconfig
+++ b/arch/alpha/Kconfig
@@ -5,6 +5,7 @@
5config ALPHA 5config ALPHA
6 bool 6 bool
7 default y 7 default y
8 select HAVE_IDE
8 select HAVE_OPROFILE 9 select HAVE_OPROFILE
9 help 10 help
10 The Alpha is a 64-bit general-purpose processor designed and 11 The Alpha is a 64-bit general-purpose processor designed and
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index 4127af93c5f3..9619c43783ff 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -222,6 +222,7 @@ config ARCH_CLPS7500
222 select TIMER_ACORN 222 select TIMER_ACORN
223 select ISA 223 select ISA
224 select NO_IOPORT 224 select NO_IOPORT
225 select HAVE_IDE
225 help 226 help
226 Support for the Cirrus Logic PS7500FE system-on-a-chip. 227 Support for the Cirrus Logic PS7500FE system-on-a-chip.
227 228
@@ -234,6 +235,7 @@ config ARCH_CO285
234 bool "Co-EBSA285" 235 bool "Co-EBSA285"
235 select FOOTBRIDGE 236 select FOOTBRIDGE
236 select FOOTBRIDGE_ADDIN 237 select FOOTBRIDGE_ADDIN
238 select HAVE_IDE
237 help 239 help
238 Support for Intel's EBSA285 companion chip. 240 Support for Intel's EBSA285 companion chip.
239 241
@@ -258,6 +260,7 @@ config ARCH_EP93XX
258config ARCH_FOOTBRIDGE 260config ARCH_FOOTBRIDGE
259 bool "FootBridge" 261 bool "FootBridge"
260 select FOOTBRIDGE 262 select FOOTBRIDGE
263 select HAVE_IDE
261 help 264 help
262 Support for systems based on the DC21285 companion chip 265 Support for systems based on the DC21285 companion chip
263 ("FootBridge"), such as the Simtec CATS and the Rebel NetWinder. 266 ("FootBridge"), such as the Simtec CATS and the Rebel NetWinder.
@@ -296,6 +299,7 @@ config ARCH_IOP32X
296 depends on MMU 299 depends on MMU
297 select PLAT_IOP 300 select PLAT_IOP
298 select PCI 301 select PCI
302 select HAVE_IDE
299 help 303 help
300 Support for Intel's 80219 and IOP32X (XScale) family of 304 Support for Intel's 80219 and IOP32X (XScale) family of
301 processors. 305 processors.
@@ -305,12 +309,14 @@ config ARCH_IOP33X
305 depends on MMU 309 depends on MMU
306 select PLAT_IOP 310 select PLAT_IOP
307 select PCI 311 select PCI
312 select HAVE_IDE
308 help 313 help
309 Support for Intel's IOP33X (XScale) family of processors. 314 Support for Intel's IOP33X (XScale) family of processors.
310 315
311config ARCH_IXP23XX 316config ARCH_IXP23XX
312 bool "IXP23XX-based" 317 bool "IXP23XX-based"
313 depends on MMU 318 depends on MMU
319 select HAVE_IDE
314 select PCI 320 select PCI
315 help 321 help
316 Support for Intel's IXP23xx (XScale) family of processors. 322 Support for Intel's IXP23xx (XScale) family of processors.
@@ -328,12 +334,14 @@ config ARCH_IXP4XX
328 select GENERIC_GPIO 334 select GENERIC_GPIO
329 select GENERIC_TIME 335 select GENERIC_TIME
330 select GENERIC_CLOCKEVENTS 336 select GENERIC_CLOCKEVENTS
337 select HAVE_IDE
331 help 338 help
332 Support for Intel's IXP4XX (XScale) family of processors. 339 Support for Intel's IXP4XX (XScale) family of processors.
333 340
334config ARCH_L7200 341config ARCH_L7200
335 bool "LinkUp-L7200" 342 bool "LinkUp-L7200"
336 select FIQ 343 select FIQ
344 select HAVE_IDE
337 help 345 help
338 Say Y here if you intend to run this kernel on a LinkUp Systems 346 Say Y here if you intend to run this kernel on a LinkUp Systems
339 L7200 Software Development Board which uses an ARM720T processor. 347 L7200 Software Development Board which uses an ARM720T processor.
@@ -388,6 +396,7 @@ config ARCH_PXA
388 depends on MMU 396 depends on MMU
389 select ARCH_MTD_XIP 397 select ARCH_MTD_XIP
390 select GENERIC_GPIO 398 select GENERIC_GPIO
399 select HAVE_IDE
391 select HAVE_GPIO_LIB 400 select HAVE_GPIO_LIB
392 select GENERIC_TIME 401 select GENERIC_TIME
393 select GENERIC_CLOCKEVENTS 402 select GENERIC_CLOCKEVENTS
@@ -403,6 +412,7 @@ config ARCH_RPC
403 select ARCH_MAY_HAVE_PC_FDC 412 select ARCH_MAY_HAVE_PC_FDC
404 select ISA_DMA_API 413 select ISA_DMA_API
405 select NO_IOPORT 414 select NO_IOPORT
415 select HAVE_IDE
406 help 416 help
407 On the Acorn Risc-PC, Linux can support the internal IDE disk and 417 On the Acorn Risc-PC, Linux can support the internal IDE disk and
408 CD-ROM interface, serial and parallel port, and the floppy drive. 418 CD-ROM interface, serial and parallel port, and the floppy drive.
@@ -414,12 +424,14 @@ config ARCH_SA1100
414 select ARCH_MTD_XIP 424 select ARCH_MTD_XIP
415 select GENERIC_GPIO 425 select GENERIC_GPIO
416 select GENERIC_TIME 426 select GENERIC_TIME
427 select HAVE_IDE
417 help 428 help
418 Support for StrongARM 11x0 based boards. 429 Support for StrongARM 11x0 based boards.
419 430
420config ARCH_S3C2410 431config ARCH_S3C2410
421 bool "Samsung S3C2410, S3C2412, S3C2413, S3C2440, S3C2442, S3C2443" 432 bool "Samsung S3C2410, S3C2412, S3C2413, S3C2440, S3C2442, S3C2443"
422 select GENERIC_GPIO 433 select GENERIC_GPIO
434 select HAVE_IDE
423 help 435 help
424 Samsung S3C2410X CPU based systems, such as the Simtec Electronics 436 Samsung S3C2410X CPU based systems, such as the Simtec Electronics
425 BAST (<http://www.simtec.co.uk/products/EB110ITX/>), the IPAQ 1940 or 437 BAST (<http://www.simtec.co.uk/products/EB110ITX/>), the IPAQ 1940 or
@@ -427,6 +439,7 @@ config ARCH_S3C2410
427 439
428config ARCH_SHARK 440config ARCH_SHARK
429 bool "Shark" 441 bool "Shark"
442 select HAVE_IDE
430 select ISA 443 select ISA
431 select ISA_DMA 444 select ISA_DMA
432 select PCI 445 select PCI
@@ -436,6 +449,7 @@ config ARCH_SHARK
436 449
437config ARCH_LH7A40X 450config ARCH_LH7A40X
438 bool "Sharp LH7A40X" 451 bool "Sharp LH7A40X"
452 select HAVE_IDE
439 help 453 help
440 Say Y here for systems based on one of the Sharp LH7A40X 454 Say Y here for systems based on one of the Sharp LH7A40X
441 System on a Chip processors. These CPUs include an ARM922T 455 System on a Chip processors. These CPUs include an ARM922T
@@ -1093,12 +1107,7 @@ source "drivers/block/Kconfig"
1093 1107
1094source "drivers/misc/Kconfig" 1108source "drivers/misc/Kconfig"
1095 1109
1096if PCMCIA || ARCH_CLPS7500 || ARCH_IOP32X || ARCH_IOP33X || ARCH_IXP4XX \
1097 || ARCH_L7200 || ARCH_LH7A40X || ARCH_PXA || ARCH_RPC \
1098 || ARCH_S3C2410 || ARCH_SA1100 || ARCH_SHARK || FOOTBRIDGE \
1099 || ARCH_IXP23XX
1100source "drivers/ide/Kconfig" 1110source "drivers/ide/Kconfig"
1101endif
1102 1111
1103source "drivers/scsi/Kconfig" 1112source "drivers/scsi/Kconfig"
1104 1113
diff --git a/arch/avr32/Kconfig b/arch/avr32/Kconfig
index 28e0caf4156c..c75d7089f982 100644
--- a/arch/avr32/Kconfig
+++ b/arch/avr32/Kconfig
@@ -10,6 +10,7 @@ config AVR32
10 # With EMBEDDED=n, we get lots of stuff automatically selected 10 # With EMBEDDED=n, we get lots of stuff automatically selected
11 # that we usually don't need on AVR32. 11 # that we usually don't need on AVR32.
12 select EMBEDDED 12 select EMBEDDED
13 select HAVE_IDE
13 select HAVE_OPROFILE 14 select HAVE_OPROFILE
14 select HAVE_KPROBES 15 select HAVE_KPROBES
15 help 16 help
diff --git a/arch/blackfin/Kconfig b/arch/blackfin/Kconfig
index 368bc7fe167e..589c6aca4803 100644
--- a/arch/blackfin/Kconfig
+++ b/arch/blackfin/Kconfig
@@ -24,6 +24,7 @@ config RWSEM_XCHGADD_ALGORITHM
24config BLACKFIN 24config BLACKFIN
25 bool 25 bool
26 default y 26 default y
27 select HAVE_IDE
27 select HAVE_OPROFILE 28 select HAVE_OPROFILE
28 29
29config ZONE_DMA 30config ZONE_DMA
diff --git a/arch/cris/Kconfig b/arch/cris/Kconfig
index 8456bc8efb7c..9389d38f222f 100644
--- a/arch/cris/Kconfig
+++ b/arch/cris/Kconfig
@@ -54,6 +54,7 @@ config FORCE_MAX_ZONEORDER
54config CRIS 54config CRIS
55 bool 55 bool
56 default y 56 default y
57 select HAVE_IDE
57 58
58config HZ 59config HZ
59 int 60 int
diff --git a/arch/frv/Kconfig b/arch/frv/Kconfig
index 9e561ede0925..a5aac1b07562 100644
--- a/arch/frv/Kconfig
+++ b/arch/frv/Kconfig
@@ -5,6 +5,7 @@
5config FRV 5config FRV
6 bool 6 bool
7 default y 7 default y
8 select HAVE_IDE
8 9
9config ZONE_DMA 10config ZONE_DMA
10 bool 11 bool
diff --git a/arch/h8300/Kconfig b/arch/h8300/Kconfig
index f69e5ea38558..085dc6ec152b 100644
--- a/arch/h8300/Kconfig
+++ b/arch/h8300/Kconfig
@@ -8,6 +8,7 @@ mainmenu "uClinux/h8300 (w/o MMU) Kernel Configuration"
8config H8300 8config H8300
9 bool 9 bool
10 default y 10 default y
11 select HAVE_IDE
11 12
12config MMU 13config MMU
13 bool 14 bool
diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig
index b0de1132dfc0..2d4fcd01bc91 100644
--- a/arch/ia64/Kconfig
+++ b/arch/ia64/Kconfig
@@ -15,6 +15,7 @@ config IA64
15 select ACPI if (!IA64_HP_SIM) 15 select ACPI if (!IA64_HP_SIM)
16 select PM if (!IA64_HP_SIM) 16 select PM if (!IA64_HP_SIM)
17 select ARCH_SUPPORTS_MSI 17 select ARCH_SUPPORTS_MSI
18 select HAVE_IDE
18 select HAVE_OPROFILE 19 select HAVE_OPROFILE
19 select HAVE_KPROBES 20 select HAVE_KPROBES
20 default y 21 default y
diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c
index 488e48a5deea..8fd7e825192b 100644
--- a/arch/ia64/pci/pci.c
+++ b/arch/ia64/pci/pci.c
@@ -43,8 +43,7 @@
43#define PCI_SAL_EXT_ADDRESS(seg, bus, devfn, reg) \ 43#define PCI_SAL_EXT_ADDRESS(seg, bus, devfn, reg) \
44 (((u64) seg << 28) | (bus << 20) | (devfn << 12) | (reg)) 44 (((u64) seg << 28) | (bus << 20) | (devfn << 12) | (reg))
45 45
46static int 46int raw_pci_read(unsigned int seg, unsigned int bus, unsigned int devfn,
47pci_sal_read (unsigned int seg, unsigned int bus, unsigned int devfn,
48 int reg, int len, u32 *value) 47 int reg, int len, u32 *value)
49{ 48{
50 u64 addr, data = 0; 49 u64 addr, data = 0;
@@ -68,8 +67,7 @@ pci_sal_read (unsigned int seg, unsigned int bus, unsigned int devfn,
68 return 0; 67 return 0;
69} 68}
70 69
71static int 70int raw_pci_write(unsigned int seg, unsigned int bus, unsigned int devfn,
72pci_sal_write (unsigned int seg, unsigned int bus, unsigned int devfn,
73 int reg, int len, u32 value) 71 int reg, int len, u32 value)
74{ 72{
75 u64 addr; 73 u64 addr;
@@ -91,24 +89,17 @@ pci_sal_write (unsigned int seg, unsigned int bus, unsigned int devfn,
91 return 0; 89 return 0;
92} 90}
93 91
94static struct pci_raw_ops pci_sal_ops = { 92static int pci_read(struct pci_bus *bus, unsigned int devfn, int where,
95 .read = pci_sal_read, 93 int size, u32 *value)
96 .write = pci_sal_write
97};
98
99struct pci_raw_ops *raw_pci_ops = &pci_sal_ops;
100
101static int
102pci_read (struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *value)
103{ 94{
104 return raw_pci_ops->read(pci_domain_nr(bus), bus->number, 95 return raw_pci_read(pci_domain_nr(bus), bus->number,
105 devfn, where, size, value); 96 devfn, where, size, value);
106} 97}
107 98
108static int 99static int pci_write(struct pci_bus *bus, unsigned int devfn, int where,
109pci_write (struct pci_bus *bus, unsigned int devfn, int where, int size, u32 value) 100 int size, u32 value)
110{ 101{
111 return raw_pci_ops->write(pci_domain_nr(bus), bus->number, 102 return raw_pci_write(pci_domain_nr(bus), bus->number,
112 devfn, where, size, value); 103 devfn, where, size, value);
113} 104}
114 105
diff --git a/arch/ia64/sn/pci/tioce_provider.c b/arch/ia64/sn/pci/tioce_provider.c
index e1a3e19d3d9c..999f14f986e2 100644
--- a/arch/ia64/sn/pci/tioce_provider.c
+++ b/arch/ia64/sn/pci/tioce_provider.c
@@ -752,13 +752,13 @@ tioce_kern_init(struct tioce_common *tioce_common)
752 * Determine the secondary bus number of the port2 logical PPB. 752 * Determine the secondary bus number of the port2 logical PPB.
753 * This is used to decide whether a given pci device resides on 753 * This is used to decide whether a given pci device resides on
754 * port1 or port2. Note: We don't have enough plumbing set up 754 * port1 or port2. Note: We don't have enough plumbing set up
755 * here to use pci_read_config_xxx() so use the raw_pci_ops vector. 755 * here to use pci_read_config_xxx() so use raw_pci_read().
756 */ 756 */
757 757
758 seg = tioce_common->ce_pcibus.bs_persist_segment; 758 seg = tioce_common->ce_pcibus.bs_persist_segment;
759 bus = tioce_common->ce_pcibus.bs_persist_busnum; 759 bus = tioce_common->ce_pcibus.bs_persist_busnum;
760 760
761 raw_pci_ops->read(seg, bus, PCI_DEVFN(2, 0), PCI_SECONDARY_BUS, 1,&tmp); 761 raw_pci_read(seg, bus, PCI_DEVFN(2, 0), PCI_SECONDARY_BUS, 1,&tmp);
762 tioce_kern->ce_port1_secondary = (u8) tmp; 762 tioce_kern->ce_port1_secondary = (u8) tmp;
763 763
764 /* 764 /*
@@ -799,11 +799,11 @@ tioce_kern_init(struct tioce_common *tioce_common)
799 799
800 /* mem base/limit */ 800 /* mem base/limit */
801 801
802 raw_pci_ops->read(seg, bus, PCI_DEVFN(dev, 0), 802 raw_pci_read(seg, bus, PCI_DEVFN(dev, 0),
803 PCI_MEMORY_BASE, 2, &tmp); 803 PCI_MEMORY_BASE, 2, &tmp);
804 base = (u64)tmp << 16; 804 base = (u64)tmp << 16;
805 805
806 raw_pci_ops->read(seg, bus, PCI_DEVFN(dev, 0), 806 raw_pci_read(seg, bus, PCI_DEVFN(dev, 0),
807 PCI_MEMORY_LIMIT, 2, &tmp); 807 PCI_MEMORY_LIMIT, 2, &tmp);
808 limit = (u64)tmp << 16; 808 limit = (u64)tmp << 16;
809 limit |= 0xfffffUL; 809 limit |= 0xfffffUL;
@@ -817,21 +817,21 @@ tioce_kern_init(struct tioce_common *tioce_common)
817 * attributes. 817 * attributes.
818 */ 818 */
819 819
820 raw_pci_ops->read(seg, bus, PCI_DEVFN(dev, 0), 820 raw_pci_read(seg, bus, PCI_DEVFN(dev, 0),
821 PCI_PREF_MEMORY_BASE, 2, &tmp); 821 PCI_PREF_MEMORY_BASE, 2, &tmp);
822 base = ((u64)tmp & PCI_PREF_RANGE_MASK) << 16; 822 base = ((u64)tmp & PCI_PREF_RANGE_MASK) << 16;
823 823
824 raw_pci_ops->read(seg, bus, PCI_DEVFN(dev, 0), 824 raw_pci_read(seg, bus, PCI_DEVFN(dev, 0),
825 PCI_PREF_BASE_UPPER32, 4, &tmp); 825 PCI_PREF_BASE_UPPER32, 4, &tmp);
826 base |= (u64)tmp << 32; 826 base |= (u64)tmp << 32;
827 827
828 raw_pci_ops->read(seg, bus, PCI_DEVFN(dev, 0), 828 raw_pci_read(seg, bus, PCI_DEVFN(dev, 0),
829 PCI_PREF_MEMORY_LIMIT, 2, &tmp); 829 PCI_PREF_MEMORY_LIMIT, 2, &tmp);
830 830
831 limit = ((u64)tmp & PCI_PREF_RANGE_MASK) << 16; 831 limit = ((u64)tmp & PCI_PREF_RANGE_MASK) << 16;
832 limit |= 0xfffffUL; 832 limit |= 0xfffffUL;
833 833
834 raw_pci_ops->read(seg, bus, PCI_DEVFN(dev, 0), 834 raw_pci_read(seg, bus, PCI_DEVFN(dev, 0),
835 PCI_PREF_LIMIT_UPPER32, 4, &tmp); 835 PCI_PREF_LIMIT_UPPER32, 4, &tmp);
836 limit |= (u64)tmp << 32; 836 limit |= (u64)tmp << 32;
837 837
diff --git a/arch/m32r/Kconfig b/arch/m32r/Kconfig
index d4679ab55b96..de153de2ea9f 100644
--- a/arch/m32r/Kconfig
+++ b/arch/m32r/Kconfig
@@ -8,6 +8,7 @@ mainmenu "Linux/M32R Kernel Configuration"
8config M32R 8config M32R
9 bool 9 bool
10 default y 10 default y
11 select HAVE_IDE
11 select HAVE_OPROFILE 12 select HAVE_OPROFILE
12 13
13config SBUS 14config SBUS
diff --git a/arch/m68k/Kconfig b/arch/m68k/Kconfig
index 2b0ed89cd173..65db2261b9ea 100644
--- a/arch/m68k/Kconfig
+++ b/arch/m68k/Kconfig
@@ -5,6 +5,7 @@
5config M68K 5config M68K
6 bool 6 bool
7 default y 7 default y
8 select HAVE_IDE
8 9
9config MMU 10config MMU
10 bool 11 bool
diff --git a/arch/m68knommu/Kconfig b/arch/m68knommu/Kconfig
index 548a7b321633..07eb4c4bab82 100644
--- a/arch/m68knommu/Kconfig
+++ b/arch/m68knommu/Kconfig
@@ -8,6 +8,7 @@ mainmenu "uClinux/68k (w/o MMU) Kernel Configuration"
8config M68K 8config M68K
9 bool 9 bool
10 default y 10 default y
11 select HAVE_IDE
11 12
12config MMU 13config MMU
13 bool 14 bool
diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig
index ec78a5762e9e..ade230d445d9 100644
--- a/arch/mips/Kconfig
+++ b/arch/mips/Kconfig
@@ -1,6 +1,7 @@
1config MIPS 1config MIPS
2 bool 2 bool
3 default y 3 default y
4 select HAVE_IDE
4 select HAVE_OPROFILE 5 select HAVE_OPROFILE
5 # Horrible source of confusion. Die, die, die ... 6 # Horrible source of confusion. Die, die, die ...
6 select EMBEDDED 7 select EMBEDDED
diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig
index d929ac84f25a..bc7a19da6245 100644
--- a/arch/parisc/Kconfig
+++ b/arch/parisc/Kconfig
@@ -7,6 +7,7 @@ mainmenu "Linux/PA-RISC Kernel Configuration"
7 7
8config PARISC 8config PARISC
9 def_bool y 9 def_bool y
10 select HAVE_IDE
10 select HAVE_OPROFILE 11 select HAVE_OPROFILE
11 help 12 help
12 The PA-RISC microprocessor is designed by Hewlett-Packard and used 13 The PA-RISC microprocessor is designed by Hewlett-Packard and used
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index 26b963c33c88..485513c9f1af 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -87,6 +87,7 @@ config ARCH_NO_VIRT_TO_BUS
87config PPC 87config PPC
88 bool 88 bool
89 default y 89 default y
90 select HAVE_IDE
90 select HAVE_OPROFILE 91 select HAVE_OPROFILE
91 select HAVE_KPROBES 92 select HAVE_KPROBES
92 93
diff --git a/arch/ppc/Kconfig b/arch/ppc/Kconfig
index 531156f8919c..abc877faf123 100644
--- a/arch/ppc/Kconfig
+++ b/arch/ppc/Kconfig
@@ -42,6 +42,7 @@ config GENERIC_CALIBRATE_DELAY
42config PPC 42config PPC
43 bool 43 bool
44 default y 44 default y
45 select HAVE_IDE
45 select HAVE_OPROFILE 46 select HAVE_OPROFILE
46 select HAVE_KPROBES 47 select HAVE_KPROBES
47 48
diff --git a/arch/s390/defconfig b/arch/s390/defconfig
index ece7b99da895..39921f3a9685 100644
--- a/arch/s390/defconfig
+++ b/arch/s390/defconfig
@@ -1,12 +1,13 @@
1# 1#
2# Automatically generated make config: don't edit 2# Automatically generated make config: don't edit
3# Linux kernel version: 2.6.23 3# Linux kernel version: 2.6.24
4# Mon Oct 22 12:10:44 2007 4# Sat Feb 9 12:13:01 2008
5# 5#
6CONFIG_MMU=y 6CONFIG_MMU=y
7CONFIG_ZONE_DMA=y 7CONFIG_ZONE_DMA=y
8CONFIG_LOCKDEP_SUPPORT=y 8CONFIG_LOCKDEP_SUPPORT=y
9CONFIG_STACKTRACE_SUPPORT=y 9CONFIG_STACKTRACE_SUPPORT=y
10CONFIG_HAVE_LATENCYTOP_SUPPORT=y
10CONFIG_RWSEM_XCHGADD_ALGORITHM=y 11CONFIG_RWSEM_XCHGADD_ALGORITHM=y
11# CONFIG_ARCH_HAS_ILOG2_U32 is not set 12# CONFIG_ARCH_HAS_ILOG2_U32 is not set
12# CONFIG_ARCH_HAS_ILOG2_U64 is not set 13# CONFIG_ARCH_HAS_ILOG2_U64 is not set
@@ -15,6 +16,7 @@ CONFIG_GENERIC_TIME=y
15CONFIG_GENERIC_BUG=y 16CONFIG_GENERIC_BUG=y
16CONFIG_NO_IOMEM=y 17CONFIG_NO_IOMEM=y
17CONFIG_NO_DMA=y 18CONFIG_NO_DMA=y
19CONFIG_GENERIC_LOCKBREAK=y
18CONFIG_S390=y 20CONFIG_S390=y
19CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config" 21CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
20 22
@@ -32,7 +34,6 @@ CONFIG_SYSVIPC_SYSCTL=y
32CONFIG_POSIX_MQUEUE=y 34CONFIG_POSIX_MQUEUE=y
33# CONFIG_BSD_PROCESS_ACCT is not set 35# CONFIG_BSD_PROCESS_ACCT is not set
34# CONFIG_TASKSTATS is not set 36# CONFIG_TASKSTATS is not set
35# CONFIG_USER_NS is not set
36CONFIG_AUDIT=y 37CONFIG_AUDIT=y
37# CONFIG_AUDITSYSCALL is not set 38# CONFIG_AUDITSYSCALL is not set
38CONFIG_IKCONFIG=y 39CONFIG_IKCONFIG=y
@@ -41,13 +42,19 @@ CONFIG_LOG_BUF_SHIFT=17
41CONFIG_CGROUPS=y 42CONFIG_CGROUPS=y
42# CONFIG_CGROUP_DEBUG is not set 43# CONFIG_CGROUP_DEBUG is not set
43CONFIG_CGROUP_NS=y 44CONFIG_CGROUP_NS=y
44CONFIG_CGROUP_CPUACCT=y
45# CONFIG_CPUSETS is not set 45# CONFIG_CPUSETS is not set
46CONFIG_FAIR_GROUP_SCHED=y 46CONFIG_FAIR_GROUP_SCHED=y
47CONFIG_FAIR_USER_SCHED=y 47CONFIG_FAIR_USER_SCHED=y
48# CONFIG_FAIR_CGROUP_SCHED is not set 48# CONFIG_FAIR_CGROUP_SCHED is not set
49# CONFIG_CGROUP_CPUACCT is not set
50# CONFIG_RESOURCE_COUNTERS is not set
49CONFIG_SYSFS_DEPRECATED=y 51CONFIG_SYSFS_DEPRECATED=y
50# CONFIG_RELAY is not set 52# CONFIG_RELAY is not set
53CONFIG_NAMESPACES=y
54CONFIG_UTS_NS=y
55CONFIG_IPC_NS=y
56# CONFIG_USER_NS is not set
57# CONFIG_PID_NS is not set
51CONFIG_BLK_DEV_INITRD=y 58CONFIG_BLK_DEV_INITRD=y
52CONFIG_INITRAMFS_SOURCE="" 59CONFIG_INITRAMFS_SOURCE=""
53# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set 60# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
@@ -61,17 +68,26 @@ CONFIG_HOTPLUG=y
61CONFIG_PRINTK=y 68CONFIG_PRINTK=y
62CONFIG_BUG=y 69CONFIG_BUG=y
63CONFIG_ELF_CORE=y 70CONFIG_ELF_CORE=y
71# CONFIG_COMPAT_BRK is not set
64CONFIG_BASE_FULL=y 72CONFIG_BASE_FULL=y
65CONFIG_FUTEX=y 73CONFIG_FUTEX=y
66CONFIG_ANON_INODES=y 74CONFIG_ANON_INODES=y
67CONFIG_EPOLL=y 75CONFIG_EPOLL=y
68CONFIG_SIGNALFD=y 76CONFIG_SIGNALFD=y
77CONFIG_TIMERFD=y
69CONFIG_EVENTFD=y 78CONFIG_EVENTFD=y
70CONFIG_SHMEM=y 79CONFIG_SHMEM=y
71CONFIG_VM_EVENT_COUNTERS=y 80CONFIG_VM_EVENT_COUNTERS=y
72CONFIG_SLAB=y 81CONFIG_SLAB=y
73# CONFIG_SLUB is not set 82# CONFIG_SLUB is not set
74# CONFIG_SLOB is not set 83# CONFIG_SLOB is not set
84# CONFIG_PROFILING is not set
85# CONFIG_MARKERS is not set
86CONFIG_HAVE_OPROFILE=y
87CONFIG_KPROBES=y
88CONFIG_HAVE_KPROBES=y
89CONFIG_PROC_PAGE_MONITOR=y
90CONFIG_SLABINFO=y
75CONFIG_RT_MUTEXES=y 91CONFIG_RT_MUTEXES=y
76# CONFIG_TINY_SHMEM is not set 92# CONFIG_TINY_SHMEM is not set
77CONFIG_BASE_SMALL=0 93CONFIG_BASE_SMALL=0
@@ -99,6 +115,8 @@ CONFIG_DEFAULT_DEADLINE=y
99# CONFIG_DEFAULT_CFQ is not set 115# CONFIG_DEFAULT_CFQ is not set
100# CONFIG_DEFAULT_NOOP is not set 116# CONFIG_DEFAULT_NOOP is not set
101CONFIG_DEFAULT_IOSCHED="deadline" 117CONFIG_DEFAULT_IOSCHED="deadline"
118CONFIG_CLASSIC_RCU=y
119# CONFIG_PREEMPT_RCU is not set
102 120
103# 121#
104# Base setup 122# Base setup
@@ -137,7 +155,7 @@ CONFIG_ARCH_POPULATES_NODE_MAP=y
137# CONFIG_PREEMPT_NONE is not set 155# CONFIG_PREEMPT_NONE is not set
138# CONFIG_PREEMPT_VOLUNTARY is not set 156# CONFIG_PREEMPT_VOLUNTARY is not set
139CONFIG_PREEMPT=y 157CONFIG_PREEMPT=y
140CONFIG_PREEMPT_BKL=y 158# CONFIG_RCU_TRACE is not set
141CONFIG_SELECT_MEMORY_MODEL=y 159CONFIG_SELECT_MEMORY_MODEL=y
142CONFIG_FLATMEM_MANUAL=y 160CONFIG_FLATMEM_MANUAL=y
143# CONFIG_DISCONTIGMEM_MANUAL is not set 161# CONFIG_DISCONTIGMEM_MANUAL is not set
@@ -151,7 +169,6 @@ CONFIG_RESOURCES_64BIT=y
151CONFIG_ZONE_DMA_FLAG=1 169CONFIG_ZONE_DMA_FLAG=1
152CONFIG_BOUNCE=y 170CONFIG_BOUNCE=y
153CONFIG_VIRT_TO_BUS=y 171CONFIG_VIRT_TO_BUS=y
154CONFIG_HOLES_IN_ZONE=y
155 172
156# 173#
157# I/O subsystem configuration 174# I/O subsystem configuration
@@ -180,6 +197,7 @@ CONFIG_HZ_100=y
180# CONFIG_HZ_300 is not set 197# CONFIG_HZ_300 is not set
181# CONFIG_HZ_1000 is not set 198# CONFIG_HZ_1000 is not set
182CONFIG_HZ=100 199CONFIG_HZ=100
200# CONFIG_SCHED_HRTICK is not set
183CONFIG_NO_IDLE_HZ=y 201CONFIG_NO_IDLE_HZ=y
184CONFIG_NO_IDLE_HZ_INIT=y 202CONFIG_NO_IDLE_HZ_INIT=y
185CONFIG_S390_HYPFS_FS=y 203CONFIG_S390_HYPFS_FS=y
@@ -201,6 +219,7 @@ CONFIG_XFRM=y
201# CONFIG_XFRM_USER is not set 219# CONFIG_XFRM_USER is not set
202# CONFIG_XFRM_SUB_POLICY is not set 220# CONFIG_XFRM_SUB_POLICY is not set
203# CONFIG_XFRM_MIGRATE is not set 221# CONFIG_XFRM_MIGRATE is not set
222# CONFIG_XFRM_STATISTICS is not set
204CONFIG_NET_KEY=y 223CONFIG_NET_KEY=y
205# CONFIG_NET_KEY_MIGRATE is not set 224# CONFIG_NET_KEY_MIGRATE is not set
206CONFIG_IUCV=m 225CONFIG_IUCV=m
@@ -251,6 +270,7 @@ CONFIG_IPV6_SIT=y
251# CONFIG_NETWORK_SECMARK is not set 270# CONFIG_NETWORK_SECMARK is not set
252CONFIG_NETFILTER=y 271CONFIG_NETFILTER=y
253# CONFIG_NETFILTER_DEBUG is not set 272# CONFIG_NETFILTER_DEBUG is not set
273CONFIG_NETFILTER_ADVANCED=y
254 274
255# 275#
256# Core Netfilter Configuration 276# Core Netfilter Configuration
@@ -258,7 +278,6 @@ CONFIG_NETFILTER=y
258CONFIG_NETFILTER_NETLINK=m 278CONFIG_NETFILTER_NETLINK=m
259CONFIG_NETFILTER_NETLINK_QUEUE=m 279CONFIG_NETFILTER_NETLINK_QUEUE=m
260CONFIG_NETFILTER_NETLINK_LOG=m 280CONFIG_NETFILTER_NETLINK_LOG=m
261CONFIG_NF_CONNTRACK_ENABLED=m
262CONFIG_NF_CONNTRACK=m 281CONFIG_NF_CONNTRACK=m
263# CONFIG_NF_CT_ACCT is not set 282# CONFIG_NF_CT_ACCT is not set
264# CONFIG_NF_CONNTRACK_MARK is not set 283# CONFIG_NF_CONNTRACK_MARK is not set
@@ -286,7 +305,7 @@ CONFIG_NF_CONNTRACK=m
286# CONFIG_IP_NF_ARPTABLES is not set 305# CONFIG_IP_NF_ARPTABLES is not set
287 306
288# 307#
289# IPv6: Netfilter Configuration (EXPERIMENTAL) 308# IPv6: Netfilter Configuration
290# 309#
291# CONFIG_NF_CONNTRACK_IPV6 is not set 310# CONFIG_NF_CONNTRACK_IPV6 is not set
292# CONFIG_IP6_NF_QUEUE is not set 311# CONFIG_IP6_NF_QUEUE is not set
@@ -343,6 +362,7 @@ CONFIG_NET_CLS_U32=m
343CONFIG_CLS_U32_MARK=y 362CONFIG_CLS_U32_MARK=y
344CONFIG_NET_CLS_RSVP=m 363CONFIG_NET_CLS_RSVP=m
345CONFIG_NET_CLS_RSVP6=m 364CONFIG_NET_CLS_RSVP6=m
365CONFIG_NET_CLS_FLOW=m
346# CONFIG_NET_EMATCH is not set 366# CONFIG_NET_EMATCH is not set
347CONFIG_NET_CLS_ACT=y 367CONFIG_NET_CLS_ACT=y
348CONFIG_NET_ACT_POLICE=y 368CONFIG_NET_ACT_POLICE=y
@@ -351,7 +371,6 @@ CONFIG_NET_ACT_POLICE=y
351CONFIG_NET_ACT_NAT=m 371CONFIG_NET_ACT_NAT=m
352# CONFIG_NET_ACT_PEDIT is not set 372# CONFIG_NET_ACT_PEDIT is not set
353# CONFIG_NET_ACT_SIMP is not set 373# CONFIG_NET_ACT_SIMP is not set
354CONFIG_NET_CLS_POLICE=y
355# CONFIG_NET_CLS_IND is not set 374# CONFIG_NET_CLS_IND is not set
356CONFIG_NET_SCH_FIFO=y 375CONFIG_NET_SCH_FIFO=y
357 376
@@ -360,6 +379,15 @@ CONFIG_NET_SCH_FIFO=y
360# 379#
361# CONFIG_NET_PKTGEN is not set 380# CONFIG_NET_PKTGEN is not set
362# CONFIG_NET_TCPPROBE is not set 381# CONFIG_NET_TCPPROBE is not set
382CONFIG_CAN=m
383CONFIG_CAN_RAW=m
384CONFIG_CAN_BCM=m
385
386#
387# CAN Device Drivers
388#
389CONFIG_CAN_VCAN=m
390# CONFIG_CAN_DEBUG_DEVICES is not set
363# CONFIG_AF_RXRPC is not set 391# CONFIG_AF_RXRPC is not set
364# CONFIG_RFKILL is not set 392# CONFIG_RFKILL is not set
365# CONFIG_NET_9P is not set 393# CONFIG_NET_9P is not set
@@ -389,7 +417,7 @@ CONFIG_BLK_DEV_NBD=m
389CONFIG_BLK_DEV_RAM=y 417CONFIG_BLK_DEV_RAM=y
390CONFIG_BLK_DEV_RAM_COUNT=16 418CONFIG_BLK_DEV_RAM_COUNT=16
391CONFIG_BLK_DEV_RAM_SIZE=4096 419CONFIG_BLK_DEV_RAM_SIZE=4096
392CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024 420CONFIG_BLK_DEV_XIP=y
393# CONFIG_CDROM_PKTCDVD is not set 421# CONFIG_CDROM_PKTCDVD is not set
394# CONFIG_ATA_OVER_ETH is not set 422# CONFIG_ATA_OVER_ETH is not set
395 423
@@ -406,6 +434,7 @@ CONFIG_DASD_DIAG=y
406CONFIG_DASD_EER=y 434CONFIG_DASD_EER=y
407CONFIG_MISC_DEVICES=y 435CONFIG_MISC_DEVICES=y
408# CONFIG_EEPROM_93CX6 is not set 436# CONFIG_EEPROM_93CX6 is not set
437# CONFIG_ENCLOSURE_SERVICES is not set
409 438
410# 439#
411# SCSI device support 440# SCSI device support
@@ -487,6 +516,7 @@ CONFIG_NET_ETHERNET=y
487# CONFIG_IBM_NEW_EMAC_TAH is not set 516# CONFIG_IBM_NEW_EMAC_TAH is not set
488# CONFIG_IBM_NEW_EMAC_EMAC4 is not set 517# CONFIG_IBM_NEW_EMAC_EMAC4 is not set
489CONFIG_NETDEV_1000=y 518CONFIG_NETDEV_1000=y
519# CONFIG_E1000E_ENABLED is not set
490CONFIG_NETDEV_10000=y 520CONFIG_NETDEV_10000=y
491# CONFIG_TR is not set 521# CONFIG_TR is not set
492# CONFIG_WAN is not set 522# CONFIG_WAN is not set
@@ -508,7 +538,6 @@ CONFIG_QETH=y
508CONFIG_CCWGROUP=y 538CONFIG_CCWGROUP=y
509# CONFIG_PPP is not set 539# CONFIG_PPP is not set
510# CONFIG_SLIP is not set 540# CONFIG_SLIP is not set
511# CONFIG_SHAPER is not set
512# CONFIG_NETCONSOLE is not set 541# CONFIG_NETCONSOLE is not set
513# CONFIG_NETPOLL is not set 542# CONFIG_NETPOLL is not set
514# CONFIG_NET_POLL_CONTROLLER is not set 543# CONFIG_NET_POLL_CONTROLLER is not set
@@ -558,6 +587,7 @@ CONFIG_S390_TAPE_34XX=m
558CONFIG_MONWRITER=m 587CONFIG_MONWRITER=m
559CONFIG_S390_VMUR=m 588CONFIG_S390_VMUR=m
560# CONFIG_POWER_SUPPLY is not set 589# CONFIG_POWER_SUPPLY is not set
590# CONFIG_THERMAL is not set
561# CONFIG_WATCHDOG is not set 591# CONFIG_WATCHDOG is not set
562 592
563# 593#
@@ -584,12 +614,10 @@ CONFIG_FS_POSIX_ACL=y
584# CONFIG_XFS_FS is not set 614# CONFIG_XFS_FS is not set
585# CONFIG_GFS2_FS is not set 615# CONFIG_GFS2_FS is not set
586# CONFIG_OCFS2_FS is not set 616# CONFIG_OCFS2_FS is not set
587# CONFIG_MINIX_FS is not set 617CONFIG_DNOTIFY=y
588# CONFIG_ROMFS_FS is not set
589CONFIG_INOTIFY=y 618CONFIG_INOTIFY=y
590CONFIG_INOTIFY_USER=y 619CONFIG_INOTIFY_USER=y
591# CONFIG_QUOTA is not set 620# CONFIG_QUOTA is not set
592CONFIG_DNOTIFY=y
593# CONFIG_AUTOFS_FS is not set 621# CONFIG_AUTOFS_FS is not set
594# CONFIG_AUTOFS4_FS is not set 622# CONFIG_AUTOFS4_FS is not set
595# CONFIG_FUSE_FS is not set 623# CONFIG_FUSE_FS is not set
@@ -632,8 +660,10 @@ CONFIG_CONFIGFS_FS=m
632# CONFIG_EFS_FS is not set 660# CONFIG_EFS_FS is not set
633# CONFIG_CRAMFS is not set 661# CONFIG_CRAMFS is not set
634# CONFIG_VXFS_FS is not set 662# CONFIG_VXFS_FS is not set
663# CONFIG_MINIX_FS is not set
635# CONFIG_HPFS_FS is not set 664# CONFIG_HPFS_FS is not set
636# CONFIG_QNX4FS_FS is not set 665# CONFIG_QNX4FS_FS is not set
666# CONFIG_ROMFS_FS is not set
637# CONFIG_SYSV_FS is not set 667# CONFIG_SYSV_FS is not set
638# CONFIG_UFS_FS is not set 668# CONFIG_UFS_FS is not set
639CONFIG_NETWORK_FILESYSTEMS=y 669CONFIG_NETWORK_FILESYSTEMS=y
@@ -686,16 +716,13 @@ CONFIG_MSDOS_PARTITION=y
686# CONFIG_NLS is not set 716# CONFIG_NLS is not set
687CONFIG_DLM=m 717CONFIG_DLM=m
688# CONFIG_DLM_DEBUG is not set 718# CONFIG_DLM_DEBUG is not set
689CONFIG_INSTRUMENTATION=y
690# CONFIG_PROFILING is not set
691CONFIG_KPROBES=y
692# CONFIG_MARKERS is not set
693 719
694# 720#
695# Kernel hacking 721# Kernel hacking
696# 722#
697CONFIG_TRACE_IRQFLAGS_SUPPORT=y 723CONFIG_TRACE_IRQFLAGS_SUPPORT=y
698# CONFIG_PRINTK_TIME is not set 724# CONFIG_PRINTK_TIME is not set
725CONFIG_ENABLE_WARN_DEPRECATED=y
699CONFIG_ENABLE_MUST_CHECK=y 726CONFIG_ENABLE_MUST_CHECK=y
700CONFIG_MAGIC_SYSRQ=y 727CONFIG_MAGIC_SYSRQ=y
701# CONFIG_UNUSED_SYMBOLS is not set 728# CONFIG_UNUSED_SYMBOLS is not set
@@ -721,12 +748,18 @@ CONFIG_DEBUG_BUGVERBOSE=y
721# CONFIG_DEBUG_INFO is not set 748# CONFIG_DEBUG_INFO is not set
722# CONFIG_DEBUG_VM is not set 749# CONFIG_DEBUG_VM is not set
723# CONFIG_DEBUG_LIST is not set 750# CONFIG_DEBUG_LIST is not set
751# CONFIG_DEBUG_SG is not set
724# CONFIG_FRAME_POINTER is not set 752# CONFIG_FRAME_POINTER is not set
725CONFIG_FORCED_INLINING=y 753CONFIG_FORCED_INLINING=y
726# CONFIG_RCU_TORTURE_TEST is not set 754# CONFIG_RCU_TORTURE_TEST is not set
755# CONFIG_KPROBES_SANITY_TEST is not set
756# CONFIG_BACKTRACE_SELF_TEST is not set
727# CONFIG_LKDTM is not set 757# CONFIG_LKDTM is not set
728# CONFIG_FAULT_INJECTION is not set 758# CONFIG_FAULT_INJECTION is not set
759# CONFIG_LATENCYTOP is not set
729CONFIG_SAMPLES=y 760CONFIG_SAMPLES=y
761# CONFIG_SAMPLE_KOBJECT is not set
762# CONFIG_DEBUG_PAGEALLOC is not set
730 763
731# 764#
732# Security options 765# Security options
@@ -738,6 +771,7 @@ CONFIG_CRYPTO=y
738CONFIG_CRYPTO_ALGAPI=y 771CONFIG_CRYPTO_ALGAPI=y
739CONFIG_CRYPTO_AEAD=m 772CONFIG_CRYPTO_AEAD=m
740CONFIG_CRYPTO_BLKCIPHER=y 773CONFIG_CRYPTO_BLKCIPHER=y
774CONFIG_CRYPTO_SEQIV=m
741CONFIG_CRYPTO_HASH=m 775CONFIG_CRYPTO_HASH=m
742CONFIG_CRYPTO_MANAGER=y 776CONFIG_CRYPTO_MANAGER=y
743CONFIG_CRYPTO_HMAC=m 777CONFIG_CRYPTO_HMAC=m
@@ -745,17 +779,20 @@ CONFIG_CRYPTO_HMAC=m
745# CONFIG_CRYPTO_NULL is not set 779# CONFIG_CRYPTO_NULL is not set
746# CONFIG_CRYPTO_MD4 is not set 780# CONFIG_CRYPTO_MD4 is not set
747CONFIG_CRYPTO_MD5=m 781CONFIG_CRYPTO_MD5=m
748# CONFIG_CRYPTO_SHA1 is not set 782CONFIG_CRYPTO_SHA1=m
749# CONFIG_CRYPTO_SHA256 is not set 783# CONFIG_CRYPTO_SHA256 is not set
750# CONFIG_CRYPTO_SHA512 is not set 784# CONFIG_CRYPTO_SHA512 is not set
751# CONFIG_CRYPTO_WP512 is not set 785# CONFIG_CRYPTO_WP512 is not set
752# CONFIG_CRYPTO_TGR192 is not set 786# CONFIG_CRYPTO_TGR192 is not set
753# CONFIG_CRYPTO_GF128MUL is not set 787CONFIG_CRYPTO_GF128MUL=m
754CONFIG_CRYPTO_ECB=m 788CONFIG_CRYPTO_ECB=m
755CONFIG_CRYPTO_CBC=y 789CONFIG_CRYPTO_CBC=y
756CONFIG_CRYPTO_PCBC=m 790CONFIG_CRYPTO_PCBC=m
757# CONFIG_CRYPTO_LRW is not set 791# CONFIG_CRYPTO_LRW is not set
758# CONFIG_CRYPTO_XTS is not set 792# CONFIG_CRYPTO_XTS is not set
793CONFIG_CRYPTO_CTR=m
794CONFIG_CRYPTO_GCM=m
795CONFIG_CRYPTO_CCM=m
759# CONFIG_CRYPTO_CRYPTD is not set 796# CONFIG_CRYPTO_CRYPTD is not set
760# CONFIG_CRYPTO_DES is not set 797# CONFIG_CRYPTO_DES is not set
761CONFIG_CRYPTO_FCRYPT=m 798CONFIG_CRYPTO_FCRYPT=m
@@ -770,20 +807,22 @@ CONFIG_CRYPTO_FCRYPT=m
770# CONFIG_CRYPTO_KHAZAD is not set 807# CONFIG_CRYPTO_KHAZAD is not set
771# CONFIG_CRYPTO_ANUBIS is not set 808# CONFIG_CRYPTO_ANUBIS is not set
772CONFIG_CRYPTO_SEED=m 809CONFIG_CRYPTO_SEED=m
810CONFIG_CRYPTO_SALSA20=m
773# CONFIG_CRYPTO_DEFLATE is not set 811# CONFIG_CRYPTO_DEFLATE is not set
774# CONFIG_CRYPTO_MICHAEL_MIC is not set 812# CONFIG_CRYPTO_MICHAEL_MIC is not set
775# CONFIG_CRYPTO_CRC32C is not set 813# CONFIG_CRYPTO_CRC32C is not set
776CONFIG_CRYPTO_CAMELLIA=m 814CONFIG_CRYPTO_CAMELLIA=m
777# CONFIG_CRYPTO_TEST is not set 815# CONFIG_CRYPTO_TEST is not set
778CONFIG_CRYPTO_AUTHENC=m 816CONFIG_CRYPTO_AUTHENC=m
817CONFIG_CRYPTO_LZO=m
779CONFIG_CRYPTO_HW=y 818CONFIG_CRYPTO_HW=y
819CONFIG_ZCRYPT=m
820# CONFIG_ZCRYPT_MONOLITHIC is not set
780# CONFIG_CRYPTO_SHA1_S390 is not set 821# CONFIG_CRYPTO_SHA1_S390 is not set
781# CONFIG_CRYPTO_SHA256_S390 is not set 822# CONFIG_CRYPTO_SHA256_S390 is not set
782# CONFIG_CRYPTO_DES_S390 is not set 823# CONFIG_CRYPTO_DES_S390 is not set
783# CONFIG_CRYPTO_AES_S390 is not set 824# CONFIG_CRYPTO_AES_S390 is not set
784CONFIG_S390_PRNG=m 825CONFIG_S390_PRNG=m
785CONFIG_ZCRYPT=m
786# CONFIG_ZCRYPT_MONOLITHIC is not set
787 826
788# 827#
789# Library routines 828# Library routines
@@ -794,5 +833,7 @@ CONFIG_BITREVERSE=m
794# CONFIG_CRC_ITU_T is not set 833# CONFIG_CRC_ITU_T is not set
795CONFIG_CRC32=m 834CONFIG_CRC32=m
796CONFIG_CRC7=m 835CONFIG_CRC7=m
797# CONFIG_LIBCRC32C is not set 836CONFIG_LIBCRC32C=m
837CONFIG_LZO_COMPRESS=m
838CONFIG_LZO_DECOMPRESS=m
798CONFIG_PLIST=y 839CONFIG_PLIST=y
diff --git a/arch/s390/kernel/binfmt_elf32.c b/arch/s390/kernel/binfmt_elf32.c
index f1e40ca00d8d..3e1c315b736d 100644
--- a/arch/s390/kernel/binfmt_elf32.c
+++ b/arch/s390/kernel/binfmt_elf32.c
@@ -134,6 +134,7 @@ static inline int dump_task_fpu(struct task_struct *tsk, elf_fpregset_t *fpregs)
134} 134}
135 135
136#include <asm/processor.h> 136#include <asm/processor.h>
137#include <asm/pgalloc.h>
137#include <linux/module.h> 138#include <linux/module.h>
138#include <linux/elfcore.h> 139#include <linux/elfcore.h>
139#include <linux/binfmts.h> 140#include <linux/binfmts.h>
@@ -183,6 +184,16 @@ struct elf_prpsinfo32
183#undef start_thread 184#undef start_thread
184#define start_thread start_thread31 185#define start_thread start_thread31
185 186
187static inline void start_thread31(struct pt_regs *regs, unsigned long new_psw,
188 unsigned long new_stackp)
189{
190 set_fs(USER_DS);
191 regs->psw.mask = psw_user32_bits;
192 regs->psw.addr = new_psw;
193 regs->gprs[15] = new_stackp;
194 crst_table_downgrade(current->mm, 1UL << 31);
195}
196
186MODULE_DESCRIPTION("Binary format loader for compatibility with 32bit Linux for S390 binaries," 197MODULE_DESCRIPTION("Binary format loader for compatibility with 32bit Linux for S390 binaries,"
187 " Copyright 2000 IBM Corporation"); 198 " Copyright 2000 IBM Corporation");
188MODULE_AUTHOR("Gerhard Tonn <ton@de.ibm.com>"); 199MODULE_AUTHOR("Gerhard Tonn <ton@de.ibm.com>");
diff --git a/arch/s390/kernel/compat_wrapper.S b/arch/s390/kernel/compat_wrapper.S
index 062c3d4c0394..743d54f0b8db 100644
--- a/arch/s390/kernel/compat_wrapper.S
+++ b/arch/s390/kernel/compat_wrapper.S
@@ -1712,3 +1712,23 @@ sys_fallocate_wrapper:
1712 sllg %r5,%r6,32 # get high word of 64bit loff_t 1712 sllg %r5,%r6,32 # get high word of 64bit loff_t
1713 l %r5,164(%r15) # get low word of 64bit loff_t 1713 l %r5,164(%r15) # get low word of 64bit loff_t
1714 jg sys_fallocate 1714 jg sys_fallocate
1715
1716 .globl sys_timerfd_create_wrapper
1717sys_timerfd_create_wrapper:
1718 lgfr %r2,%r2 # int
1719 lgfr %r3,%r3 # int
1720 jg sys_timerfd_create
1721
1722 .globl compat_sys_timerfd_settime_wrapper
1723compat_sys_timerfd_settime_wrapper:
1724 lgfr %r2,%r2 # int
1725 lgfr %r3,%r3 # int
1726 llgtr %r4,%r4 # struct compat_itimerspec *
1727 llgtr %r5,%r5 # struct compat_itimerspec *
1728 jg compat_sys_timerfd_settime
1729
1730 .globl compat_sys_timerfd_gettime_wrapper
1731compat_sys_timerfd_gettime_wrapper:
1732 lgfr %r2,%r2 # int
1733 llgtr %r3,%r3 # struct compat_itimerspec *
1734 jg compat_sys_timerfd_gettime
diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c
index 0e7aca039307..a6a4729e0e94 100644
--- a/arch/s390/kernel/process.c
+++ b/arch/s390/kernel/process.c
@@ -29,7 +29,6 @@
29#include <linux/slab.h> 29#include <linux/slab.h>
30#include <linux/vmalloc.h> 30#include <linux/vmalloc.h>
31#include <linux/user.h> 31#include <linux/user.h>
32#include <linux/a.out.h>
33#include <linux/interrupt.h> 32#include <linux/interrupt.h>
34#include <linux/delay.h> 33#include <linux/delay.h>
35#include <linux/reboot.h> 34#include <linux/reboot.h>
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index f9f8779022a0..290e504061a3 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -24,7 +24,6 @@
24#include <linux/ptrace.h> 24#include <linux/ptrace.h>
25#include <linux/slab.h> 25#include <linux/slab.h>
26#include <linux/user.h> 26#include <linux/user.h>
27#include <linux/a.out.h>
28#include <linux/tty.h> 27#include <linux/tty.h>
29#include <linux/ioport.h> 28#include <linux/ioport.h>
30#include <linux/delay.h> 29#include <linux/delay.h>
diff --git a/arch/s390/kernel/syscalls.S b/arch/s390/kernel/syscalls.S
index 25eac7802fc4..c87ec687d4c6 100644
--- a/arch/s390/kernel/syscalls.S
+++ b/arch/s390/kernel/syscalls.S
@@ -327,3 +327,6 @@ SYSCALL(sys_utimensat,sys_utimensat,compat_sys_utimensat_wrapper) /* 315 */
327SYSCALL(sys_signalfd,sys_signalfd,compat_sys_signalfd_wrapper) 327SYSCALL(sys_signalfd,sys_signalfd,compat_sys_signalfd_wrapper)
328NI_SYSCALL /* 317 old sys_timer_fd */ 328NI_SYSCALL /* 317 old sys_timer_fd */
329SYSCALL(sys_eventfd,sys_eventfd,sys_eventfd_wrapper) 329SYSCALL(sys_eventfd,sys_eventfd,sys_eventfd_wrapper)
330SYSCALL(sys_timerfd_create,sys_timerfd_create,sys_timerfd_create_wrapper)
331SYSCALL(sys_timerfd_settime,sys_timerfd_settime,compat_sys_timerfd_settime_wrapper) /* 320 */
332SYSCALL(sys_timerfd_gettime,sys_timerfd_gettime,compat_sys_timerfd_gettime_wrapper)
diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index a4d29025ddbd..60f728aeaf12 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -60,6 +60,7 @@ int sysctl_userprocess_debug = 0;
60extern pgm_check_handler_t do_protection_exception; 60extern pgm_check_handler_t do_protection_exception;
61extern pgm_check_handler_t do_dat_exception; 61extern pgm_check_handler_t do_dat_exception;
62extern pgm_check_handler_t do_monitor_call; 62extern pgm_check_handler_t do_monitor_call;
63extern pgm_check_handler_t do_asce_exception;
63 64
64#define stack_pointer ({ void **sp; asm("la %0,0(15)" : "=&d" (sp)); sp; }) 65#define stack_pointer ({ void **sp; asm("la %0,0(15)" : "=&d" (sp)); sp; })
65 66
@@ -730,7 +731,7 @@ void __init trap_init(void)
730 pgm_check_table[0x12] = &translation_exception; 731 pgm_check_table[0x12] = &translation_exception;
731 pgm_check_table[0x13] = &special_op_exception; 732 pgm_check_table[0x13] = &special_op_exception;
732#ifdef CONFIG_64BIT 733#ifdef CONFIG_64BIT
733 pgm_check_table[0x38] = &do_dat_exception; 734 pgm_check_table[0x38] = &do_asce_exception;
734 pgm_check_table[0x39] = &do_dat_exception; 735 pgm_check_table[0x39] = &do_dat_exception;
735 pgm_check_table[0x3A] = &do_dat_exception; 736 pgm_check_table[0x3A] = &do_dat_exception;
736 pgm_check_table[0x3B] = &do_dat_exception; 737 pgm_check_table[0x3B] = &do_dat_exception;
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index 2456b52ed068..ed13d429a487 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -32,6 +32,7 @@
32#include <asm/system.h> 32#include <asm/system.h>
33#include <asm/pgtable.h> 33#include <asm/pgtable.h>
34#include <asm/s390_ext.h> 34#include <asm/s390_ext.h>
35#include <asm/mmu_context.h>
35 36
36#ifndef CONFIG_64BIT 37#ifndef CONFIG_64BIT
37#define __FAIL_ADDR_MASK 0x7ffff000 38#define __FAIL_ADDR_MASK 0x7ffff000
@@ -444,6 +445,45 @@ void __kprobes do_dat_exception(struct pt_regs *regs, unsigned long error_code)
444 do_exception(regs, error_code & 0xff, 0); 445 do_exception(regs, error_code & 0xff, 0);
445} 446}
446 447
448#ifdef CONFIG_64BIT
449void __kprobes do_asce_exception(struct pt_regs *regs, unsigned long error_code)
450{
451 struct mm_struct *mm;
452 struct vm_area_struct *vma;
453 unsigned long address;
454 int space;
455
456 mm = current->mm;
457 address = S390_lowcore.trans_exc_code & __FAIL_ADDR_MASK;
458 space = check_space(current);
459
460 if (unlikely(space == 0 || in_atomic() || !mm))
461 goto no_context;
462
463 local_irq_enable();
464
465 down_read(&mm->mmap_sem);
466 vma = find_vma(mm, address);
467 up_read(&mm->mmap_sem);
468
469 if (vma) {
470 update_mm(mm, current);
471 return;
472 }
473
474 /* User mode accesses just cause a SIGSEGV */
475 if (regs->psw.mask & PSW_MASK_PSTATE) {
476 current->thread.prot_addr = address;
477 current->thread.trap_no = error_code;
478 do_sigsegv(regs, error_code, SEGV_MAPERR, address);
479 return;
480 }
481
482no_context:
483 do_no_context(regs, error_code, address);
484}
485#endif
486
447#ifdef CONFIG_PFAULT 487#ifdef CONFIG_PFAULT
448/* 488/*
449 * 'pfault' pseudo page faults routines. 489 * 'pfault' pseudo page faults routines.
diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c
index 983ec6ec0e7c..8053245fe259 100644
--- a/arch/s390/mm/init.c
+++ b/arch/s390/mm/init.c
@@ -112,6 +112,7 @@ void __init paging_init(void)
112 init_mm.pgd = swapper_pg_dir; 112 init_mm.pgd = swapper_pg_dir;
113 S390_lowcore.kernel_asce = __pa(init_mm.pgd) & PAGE_MASK; 113 S390_lowcore.kernel_asce = __pa(init_mm.pgd) & PAGE_MASK;
114#ifdef CONFIG_64BIT 114#ifdef CONFIG_64BIT
115 /* A three level page table (4TB) is enough for the kernel space. */
115 S390_lowcore.kernel_asce |= _ASCE_TYPE_REGION3 | _ASCE_TABLE_LENGTH; 116 S390_lowcore.kernel_asce |= _ASCE_TYPE_REGION3 | _ASCE_TABLE_LENGTH;
116 pgd_type = _REGION3_ENTRY_EMPTY; 117 pgd_type = _REGION3_ENTRY_EMPTY;
117#else 118#else
@@ -184,7 +185,7 @@ void kernel_map_pages(struct page *page, int numpages, int enable)
184 pmd = pmd_offset(pud, address); 185 pmd = pmd_offset(pud, address);
185 pte = pte_offset_kernel(pmd, address); 186 pte = pte_offset_kernel(pmd, address);
186 if (!enable) { 187 if (!enable) {
187 ptep_invalidate(address, pte); 188 ptep_invalidate(&init_mm, address, pte);
188 continue; 189 continue;
189 } 190 }
190 *pte = mk_pte_phys(address, __pgprot(_PAGE_TYPE_RW)); 191 *pte = mk_pte_phys(address, __pgprot(_PAGE_TYPE_RW));
diff --git a/arch/s390/mm/mmap.c b/arch/s390/mm/mmap.c
index 356257c171de..5932a824547a 100644
--- a/arch/s390/mm/mmap.c
+++ b/arch/s390/mm/mmap.c
@@ -27,6 +27,7 @@
27#include <linux/personality.h> 27#include <linux/personality.h>
28#include <linux/mm.h> 28#include <linux/mm.h>
29#include <linux/module.h> 29#include <linux/module.h>
30#include <asm/pgalloc.h>
30 31
31/* 32/*
32 * Top of mmap area (just below the process stack). 33 * Top of mmap area (just below the process stack).
@@ -62,6 +63,8 @@ static inline int mmap_is_legacy(void)
62 current->signal->rlim[RLIMIT_STACK].rlim_cur == RLIM_INFINITY; 63 current->signal->rlim[RLIMIT_STACK].rlim_cur == RLIM_INFINITY;
63} 64}
64 65
66#ifndef CONFIG_64BIT
67
65/* 68/*
66 * This function, called very early during the creation of a new 69 * This function, called very early during the creation of a new
67 * process VM image, sets up which VM layout function to use: 70 * process VM image, sets up which VM layout function to use:
@@ -84,3 +87,65 @@ void arch_pick_mmap_layout(struct mm_struct *mm)
84} 87}
85EXPORT_SYMBOL_GPL(arch_pick_mmap_layout); 88EXPORT_SYMBOL_GPL(arch_pick_mmap_layout);
86 89
90#else
91
92static unsigned long
93s390_get_unmapped_area(struct file *filp, unsigned long addr,
94 unsigned long len, unsigned long pgoff, unsigned long flags)
95{
96 struct mm_struct *mm = current->mm;
97 int rc;
98
99 addr = arch_get_unmapped_area(filp, addr, len, pgoff, flags);
100 if (addr & ~PAGE_MASK)
101 return addr;
102 if (unlikely(mm->context.asce_limit < addr + len)) {
103 rc = crst_table_upgrade(mm, addr + len);
104 if (rc)
105 return (unsigned long) rc;
106 }
107 return addr;
108}
109
110static unsigned long
111s390_get_unmapped_area_topdown(struct file *filp, const unsigned long addr0,
112 const unsigned long len, const unsigned long pgoff,
113 const unsigned long flags)
114{
115 struct mm_struct *mm = current->mm;
116 unsigned long addr = addr0;
117 int rc;
118
119 addr = arch_get_unmapped_area_topdown(filp, addr, len, pgoff, flags);
120 if (addr & ~PAGE_MASK)
121 return addr;
122 if (unlikely(mm->context.asce_limit < addr + len)) {
123 rc = crst_table_upgrade(mm, addr + len);
124 if (rc)
125 return (unsigned long) rc;
126 }
127 return addr;
128}
129/*
130 * This function, called very early during the creation of a new
131 * process VM image, sets up which VM layout function to use:
132 */
133void arch_pick_mmap_layout(struct mm_struct *mm)
134{
135 /*
136 * Fall back to the standard layout if the personality
137 * bit is set, or if the expected stack growth is unlimited:
138 */
139 if (mmap_is_legacy()) {
140 mm->mmap_base = TASK_UNMAPPED_BASE;
141 mm->get_unmapped_area = s390_get_unmapped_area;
142 mm->unmap_area = arch_unmap_area;
143 } else {
144 mm->mmap_base = mmap_base();
145 mm->get_unmapped_area = s390_get_unmapped_area_topdown;
146 mm->unmap_area = arch_unmap_area_topdown;
147 }
148}
149EXPORT_SYMBOL_GPL(arch_pick_mmap_layout);
150
151#endif
diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c
index 019f518cd5a0..fd072013f88c 100644
--- a/arch/s390/mm/pgtable.c
+++ b/arch/s390/mm/pgtable.c
@@ -23,11 +23,18 @@
23#include <asm/pgalloc.h> 23#include <asm/pgalloc.h>
24#include <asm/tlb.h> 24#include <asm/tlb.h>
25#include <asm/tlbflush.h> 25#include <asm/tlbflush.h>
26#include <asm/mmu_context.h>
26 27
27#ifndef CONFIG_64BIT 28#ifndef CONFIG_64BIT
28#define ALLOC_ORDER 1 29#define ALLOC_ORDER 1
30#define TABLES_PER_PAGE 4
31#define FRAG_MASK 15UL
32#define SECOND_HALVES 10UL
29#else 33#else
30#define ALLOC_ORDER 2 34#define ALLOC_ORDER 2
35#define TABLES_PER_PAGE 2
36#define FRAG_MASK 3UL
37#define SECOND_HALVES 2UL
31#endif 38#endif
32 39
33unsigned long *crst_table_alloc(struct mm_struct *mm, int noexec) 40unsigned long *crst_table_alloc(struct mm_struct *mm, int noexec)
@@ -45,52 +52,179 @@ unsigned long *crst_table_alloc(struct mm_struct *mm, int noexec)
45 } 52 }
46 page->index = page_to_phys(shadow); 53 page->index = page_to_phys(shadow);
47 } 54 }
55 spin_lock(&mm->page_table_lock);
56 list_add(&page->lru, &mm->context.crst_list);
57 spin_unlock(&mm->page_table_lock);
48 return (unsigned long *) page_to_phys(page); 58 return (unsigned long *) page_to_phys(page);
49} 59}
50 60
51void crst_table_free(unsigned long *table) 61void crst_table_free(struct mm_struct *mm, unsigned long *table)
52{ 62{
53 unsigned long *shadow = get_shadow_table(table); 63 unsigned long *shadow = get_shadow_table(table);
64 struct page *page = virt_to_page(table);
54 65
66 spin_lock(&mm->page_table_lock);
67 list_del(&page->lru);
68 spin_unlock(&mm->page_table_lock);
55 if (shadow) 69 if (shadow)
56 free_pages((unsigned long) shadow, ALLOC_ORDER); 70 free_pages((unsigned long) shadow, ALLOC_ORDER);
57 free_pages((unsigned long) table, ALLOC_ORDER); 71 free_pages((unsigned long) table, ALLOC_ORDER);
58} 72}
59 73
74#ifdef CONFIG_64BIT
75int crst_table_upgrade(struct mm_struct *mm, unsigned long limit)
76{
77 unsigned long *table, *pgd;
78 unsigned long entry;
79
80 BUG_ON(limit > (1UL << 53));
81repeat:
82 table = crst_table_alloc(mm, mm->context.noexec);
83 if (!table)
84 return -ENOMEM;
85 spin_lock(&mm->page_table_lock);
86 if (mm->context.asce_limit < limit) {
87 pgd = (unsigned long *) mm->pgd;
88 if (mm->context.asce_limit <= (1UL << 31)) {
89 entry = _REGION3_ENTRY_EMPTY;
90 mm->context.asce_limit = 1UL << 42;
91 mm->context.asce_bits = _ASCE_TABLE_LENGTH |
92 _ASCE_USER_BITS |
93 _ASCE_TYPE_REGION3;
94 } else {
95 entry = _REGION2_ENTRY_EMPTY;
96 mm->context.asce_limit = 1UL << 53;
97 mm->context.asce_bits = _ASCE_TABLE_LENGTH |
98 _ASCE_USER_BITS |
99 _ASCE_TYPE_REGION2;
100 }
101 crst_table_init(table, entry);
102 pgd_populate(mm, (pgd_t *) table, (pud_t *) pgd);
103 mm->pgd = (pgd_t *) table;
104 table = NULL;
105 }
106 spin_unlock(&mm->page_table_lock);
107 if (table)
108 crst_table_free(mm, table);
109 if (mm->context.asce_limit < limit)
110 goto repeat;
111 update_mm(mm, current);
112 return 0;
113}
114
115void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
116{
117 pgd_t *pgd;
118
119 if (mm->context.asce_limit <= limit)
120 return;
121 __tlb_flush_mm(mm);
122 while (mm->context.asce_limit > limit) {
123 pgd = mm->pgd;
124 switch (pgd_val(*pgd) & _REGION_ENTRY_TYPE_MASK) {
125 case _REGION_ENTRY_TYPE_R2:
126 mm->context.asce_limit = 1UL << 42;
127 mm->context.asce_bits = _ASCE_TABLE_LENGTH |
128 _ASCE_USER_BITS |
129 _ASCE_TYPE_REGION3;
130 break;
131 case _REGION_ENTRY_TYPE_R3:
132 mm->context.asce_limit = 1UL << 31;
133 mm->context.asce_bits = _ASCE_TABLE_LENGTH |
134 _ASCE_USER_BITS |
135 _ASCE_TYPE_SEGMENT;
136 break;
137 default:
138 BUG();
139 }
140 mm->pgd = (pgd_t *) (pgd_val(*pgd) & _REGION_ENTRY_ORIGIN);
141 crst_table_free(mm, (unsigned long *) pgd);
142 }
143 update_mm(mm, current);
144}
145#endif
146
60/* 147/*
61 * page table entry allocation/free routines. 148 * page table entry allocation/free routines.
62 */ 149 */
63unsigned long *page_table_alloc(int noexec) 150unsigned long *page_table_alloc(struct mm_struct *mm)
64{ 151{
65 struct page *page = alloc_page(GFP_KERNEL); 152 struct page *page;
66 unsigned long *table; 153 unsigned long *table;
154 unsigned long bits;
67 155
68 if (!page) 156 bits = mm->context.noexec ? 3UL : 1UL;
69 return NULL; 157 spin_lock(&mm->page_table_lock);
70 page->index = 0; 158 page = NULL;
71 if (noexec) { 159 if (!list_empty(&mm->context.pgtable_list)) {
72 struct page *shadow = alloc_page(GFP_KERNEL); 160 page = list_first_entry(&mm->context.pgtable_list,
73 if (!shadow) { 161 struct page, lru);
74 __free_page(page); 162 if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
163 page = NULL;
164 }
165 if (!page) {
166 spin_unlock(&mm->page_table_lock);
167 page = alloc_page(GFP_KERNEL|__GFP_REPEAT);
168 if (!page)
75 return NULL; 169 return NULL;
76 } 170 pgtable_page_ctor(page);
77 table = (unsigned long *) page_to_phys(shadow); 171 page->flags &= ~FRAG_MASK;
172 table = (unsigned long *) page_to_phys(page);
78 clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE); 173 clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE);
79 page->index = (addr_t) table; 174 spin_lock(&mm->page_table_lock);
175 list_add(&page->lru, &mm->context.pgtable_list);
80 } 176 }
81 pgtable_page_ctor(page);
82 table = (unsigned long *) page_to_phys(page); 177 table = (unsigned long *) page_to_phys(page);
83 clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE); 178 while (page->flags & bits) {
179 table += 256;
180 bits <<= 1;
181 }
182 page->flags |= bits;
183 if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
184 list_move_tail(&page->lru, &mm->context.pgtable_list);
185 spin_unlock(&mm->page_table_lock);
84 return table; 186 return table;
85} 187}
86 188
87void page_table_free(unsigned long *table) 189void page_table_free(struct mm_struct *mm, unsigned long *table)
88{ 190{
89 unsigned long *shadow = get_shadow_pte(table); 191 struct page *page;
192 unsigned long bits;
90 193
91 pgtable_page_dtor(virt_to_page(table)); 194 bits = mm->context.noexec ? 3UL : 1UL;
92 if (shadow) 195 bits <<= (__pa(table) & (PAGE_SIZE - 1)) / 256 / sizeof(unsigned long);
93 free_page((unsigned long) shadow); 196 page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
94 free_page((unsigned long) table); 197 spin_lock(&mm->page_table_lock);
198 page->flags ^= bits;
199 if (page->flags & FRAG_MASK) {
200 /* Page now has some free pgtable fragments. */
201 list_move(&page->lru, &mm->context.pgtable_list);
202 page = NULL;
203 } else
204 /* All fragments of the 4K page have been freed. */
205 list_del(&page->lru);
206 spin_unlock(&mm->page_table_lock);
207 if (page) {
208 pgtable_page_dtor(page);
209 __free_page(page);
210 }
211}
95 212
213void disable_noexec(struct mm_struct *mm, struct task_struct *tsk)
214{
215 struct page *page;
216
217 spin_lock(&mm->page_table_lock);
218 /* Free shadow region and segment tables. */
219 list_for_each_entry(page, &mm->context.crst_list, lru)
220 if (page->index) {
221 free_pages((unsigned long) page->index, ALLOC_ORDER);
222 page->index = 0;
223 }
224 /* "Free" second halves of page tables. */
225 list_for_each_entry(page, &mm->context.pgtable_list, lru)
226 page->flags &= ~SECOND_HALVES;
227 spin_unlock(&mm->page_table_lock);
228 mm->context.noexec = 0;
229 update_mm(mm, tsk);
96} 230}
diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c
index 7c1287ccf788..35d90a4720fd 100644
--- a/arch/s390/mm/vmem.c
+++ b/arch/s390/mm/vmem.c
@@ -69,7 +69,19 @@ static void __ref *vmem_alloc_pages(unsigned int order)
69 return alloc_bootmem_pages((1 << order) * PAGE_SIZE); 69 return alloc_bootmem_pages((1 << order) * PAGE_SIZE);
70} 70}
71 71
72#define vmem_pud_alloc() ({ BUG(); ((pud_t *) NULL); }) 72static inline pud_t *vmem_pud_alloc(void)
73{
74 pud_t *pud = NULL;
75
76#ifdef CONFIG_64BIT
77 pud = vmem_alloc_pages(2);
78 if (!pud)
79 return NULL;
80 pud_val(*pud) = _REGION3_ENTRY_EMPTY;
81 memcpy(pud + 1, pud, (PTRS_PER_PUD - 1)*sizeof(pud_t));
82#endif
83 return pud;
84}
73 85
74static inline pmd_t *vmem_pmd_alloc(void) 86static inline pmd_t *vmem_pmd_alloc(void)
75{ 87{
@@ -84,13 +96,18 @@ static inline pmd_t *vmem_pmd_alloc(void)
84 return pmd; 96 return pmd;
85} 97}
86 98
87static inline pte_t *vmem_pte_alloc(void) 99static pte_t __init_refok *vmem_pte_alloc(void)
88{ 100{
89 pte_t *pte = vmem_alloc_pages(0); 101 pte_t *pte;
90 102
103 if (slab_is_available())
104 pte = (pte_t *) page_table_alloc(&init_mm);
105 else
106 pte = alloc_bootmem(PTRS_PER_PTE * sizeof(pte_t));
91 if (!pte) 107 if (!pte)
92 return NULL; 108 return NULL;
93 clear_table((unsigned long *) pte, _PAGE_TYPE_EMPTY, PAGE_SIZE); 109 clear_table((unsigned long *) pte, _PAGE_TYPE_EMPTY,
110 PTRS_PER_PTE * sizeof(pte_t));
94 return pte; 111 return pte;
95} 112}
96 113
@@ -360,6 +377,9 @@ void __init vmem_map_init(void)
360{ 377{
361 int i; 378 int i;
362 379
380 INIT_LIST_HEAD(&init_mm.context.crst_list);
381 INIT_LIST_HEAD(&init_mm.context.pgtable_list);
382 init_mm.context.noexec = 0;
363 NODE_DATA(0)->node_mem_map = VMEM_MAP; 383 NODE_DATA(0)->node_mem_map = VMEM_MAP;
364 for (i = 0; i < MEMORY_CHUNKS && memory_chunk[i].size > 0; i++) 384 for (i = 0; i < MEMORY_CHUNKS && memory_chunk[i].size > 0; i++)
365 vmem_add_mem(memory_chunk[i].addr, memory_chunk[i].size); 385 vmem_add_mem(memory_chunk[i].addr, memory_chunk[i].size);
diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig
index 6e035d1cf789..d87d4bf88803 100644
--- a/arch/sh/Kconfig
+++ b/arch/sh/Kconfig
@@ -8,6 +8,7 @@ mainmenu "Linux/SuperH Kernel Configuration"
8config SUPERH 8config SUPERH
9 def_bool y 9 def_bool y
10 select EMBEDDED 10 select EMBEDDED
11 select HAVE_IDE
11 select HAVE_OPROFILE 12 select HAVE_OPROFILE
12 help 13 help
13 The SuperH is a RISC processor targeted for use in embedded systems 14 The SuperH is a RISC processor targeted for use in embedded systems
diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig
index 7c674a3503b6..c40343c54920 100644
--- a/arch/sparc/Kconfig
+++ b/arch/sparc/Kconfig
@@ -69,6 +69,7 @@ config NR_CPUS
69config SPARC 69config SPARC
70 bool 70 bool
71 default y 71 default y
72 select HAVE_IDE
72 select HAVE_OPROFILE 73 select HAVE_OPROFILE
73 74
74# Identify this as a Sparc32 build 75# Identify this as a Sparc32 build
diff --git a/arch/sparc/kernel/errtbls.c b/arch/sparc/kernel/errtbls.c
index bb36f6eadfee..ed14df7116e9 100644
--- a/arch/sparc/kernel/errtbls.c
+++ b/arch/sparc/kernel/errtbls.c
@@ -1,21 +1,18 @@
1/* $Id: errtbls.c,v 1.2 1995/11/25 00:57:55 davem Exp $ 1/* errtbls.c: Error number conversion tables.
2 * errtbls.c: Error number conversion tables between various syscall
3 * OS semantics.
4 * 2 *
5 * Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu) 3 * Copyright (C) 1995, 2007 David S. Miller (davem@davemloft.net)
6 * 4 *
7 * Based upon preliminary work which is: 5 * Based upon preliminary work which is:
8 * 6 *
9 * Copyright (C) 1995 Adrian M. Rodriguez (adrian@remus.rutgers.edu) 7 * Copyright (C) 1995 Adrian M. Rodriguez (adrian@remus.rutgers.edu)
10 */ 8 */
11 9
12#include <asm/bsderrno.h> /* NetBSD (bsd4.4) errnos */
13#include <asm/solerrno.h> /* Solaris errnos */ 10#include <asm/solerrno.h> /* Solaris errnos */
14 11
15/* Here are tables which convert between Linux/SunOS error number 12/* Here is the table which converts between Linux error number values
16 * values to the equivalent in other OSs. Note that since the Linux 13 * to the equivalent under Solaris. Note that since the Linux ones
17 * ones have been set up to match exactly those of SunOS, no 14 * have been set up to match exactly those of SunOS, no translation
18 * translation table is needed for that OS. 15 * table is needed for that OS.
19 */ 16 */
20 17
21int solaris_errno[] = { 18int solaris_errno[] = {
@@ -145,132 +142,3 @@ int solaris_errno[] = {
145 SOL_ELIBMAX, 142 SOL_ELIBMAX,
146 SOL_ELIBSCN, 143 SOL_ELIBSCN,
147}; 144};
148
149int netbsd_errno[] = {
150 0,
151 BSD_EPERM,
152 BSD_ENOENT,
153 BSD_ESRCH,
154 BSD_EINTR,
155 BSD_EIO,
156 BSD_ENXIO,
157 BSD_E2BIG,
158 BSD_ENOEXEC,
159 BSD_EBADF,
160 BSD_ECHILD,
161 BSD_EAGAIN,
162 BSD_ENOMEM,
163 BSD_EACCES,
164 BSD_EFAULT,
165 BSD_NOTBLK,
166 BSD_EBUSY,
167 BSD_EEXIST,
168 BSD_EXDEV,
169 BSD_ENODEV,
170 BSD_ENOTDIR,
171 BSD_EISDIR,
172 BSD_EINVAL,
173 BSD_ENFILE,
174 BSD_EMFILE,
175 BSD_ENOTTY,
176 BSD_ETXTBSY,
177 BSD_EFBIG,
178 BSD_ENOSPC,
179 BSD_ESPIPE,
180 BSD_EROFS,
181 BSD_EMLINK,
182 BSD_EPIPE,
183 BSD_EDOM,
184 BSD_ERANGE,
185 BSD_EWOULDBLOCK,
186 BSD_EINPROGRESS,
187 BSD_EALREADY,
188 BSD_ENOTSOCK,
189 BSD_EDESTADDRREQ,
190 BSD_EMSGSIZE,
191 BSD_EPROTOTYPE,
192 BSD_ENOPROTOOPT,
193 BSD_EPROTONOSUPPORT,
194 BSD_ESOCKTNOSUPPORT,
195 BSD_EOPNOTSUPP,
196 BSD_EPFNOSUPPORT,
197 BSD_EAFNOSUPPORT,
198 BSD_EADDRINUSE,
199 BSD_EADDRNOTAVAIL,
200 BSD_ENETDOWN,
201 BSD_ENETUNREACH,
202 BSD_ENETRESET,
203 BSD_ECONNABORTED,
204 BSD_ECONNRESET,
205 BSD_ENOBUFS,
206 BSD_EISCONN,
207 BSD_ENOTONN,
208 BSD_ESHUTDOWN,
209 BSD_ETOOMANYREFS,
210 BSD_ETIMEDOUT,
211 BSD_ECONNREFUSED,
212 BSD_ELOOP,
213 BSD_ENAMETOOLONG,
214 BSD_EHOSTDOWN,
215 BSD_EHOSTUNREACH,
216 BSD_ENOTEMPTY,
217 BSD_EPROCLIM,
218 BSD_EUSERS,
219 BSD_EDQUOT,
220 BSD_ESTALE,
221 BSD_EREMOTE,
222 BSD_ENOSTR,
223 BSD_ETIME,
224 BSD_ENOSR,
225 BSD_ENOMSG,
226 BSD_EBADMSG,
227 BSD_IDRM,
228 BSD_EDEADLK,
229 BSD_ENOLCK,
230 BSD_ENONET,
231 BSD_ERREMOTE,
232 BSD_ENOLINK,
233 BSD_EADV,
234 BSD_ESRMNT,
235 BSD_ECOMM,
236 BSD_EPROTO,
237 BSD_EMULTIHOP,
238 BSD_EINVAL, /* EDOTDOT XXX??? */
239 BSD_REMCHG,
240 BSD_NOSYS,
241 BSD_STRPIPE,
242 BSD_EOVERFLOW,
243 BSD_EBADFD,
244 BSD_ECHRNG,
245 BSD_EL2NSYNC,
246 BSD_EL3HLT,
247 BSD_EL3RST,
248 BSD_NRNG,
249 BSD_EUNATCH,
250 BSD_ENOCSI,
251 BSD_EL2HLT,
252 BSD_EBADE,
253 BSD_EBADR,
254 BSD_EXFULL,
255 BSD_ENOANO,
256 BSD_EBADRQC,
257 BSD_EBADSLT,
258 BSD_EDEADLOCK,
259 BSD_EBFONT,
260 BSD_ELIBEXEC,
261 BSD_ENODATA,
262 BSD_ELIBBAD,
263 BSD_ENOPKG,
264 BSD_ELIBACC,
265 BSD_ENOTUNIQ,
266 BSD_ERESTART,
267 BSD_EUCLEAN,
268 BSD_ENOTNAM,
269 BSD_ENAVAIL,
270 BSD_EISNAM,
271 BSD_EREMOTEIO,
272 BSD_EILSEQ,
273 BSD_ELIBMAX,
274 BSD_ELIBSCN,
275};
276
diff --git a/arch/sparc/kernel/of_device.c b/arch/sparc/kernel/of_device.c
index 3ea000d15e3a..cc4c235c4f59 100644
--- a/arch/sparc/kernel/of_device.c
+++ b/arch/sparc/kernel/of_device.c
@@ -584,30 +584,3 @@ static int __init of_debug(char *str)
584} 584}
585 585
586__setup("of_debug=", of_debug); 586__setup("of_debug=", of_debug);
587
588struct of_device* of_platform_device_create(struct device_node *np,
589 const char *bus_id,
590 struct device *parent,
591 struct bus_type *bus)
592{
593 struct of_device *dev;
594
595 dev = kzalloc(sizeof(*dev), GFP_KERNEL);
596 if (!dev)
597 return NULL;
598
599 dev->dev.parent = parent;
600 dev->dev.bus = bus;
601 dev->dev.release = of_release_dev;
602
603 strlcpy(dev->dev.bus_id, bus_id, BUS_ID_SIZE);
604
605 if (of_device_register(dev) != 0) {
606 kfree(dev);
607 return NULL;
608 }
609
610 return dev;
611}
612
613EXPORT_SYMBOL(of_platform_device_create);
diff --git a/arch/sparc/kernel/vmlinux.lds.S b/arch/sparc/kernel/vmlinux.lds.S
index 216147d6e61f..b1002c607196 100644
--- a/arch/sparc/kernel/vmlinux.lds.S
+++ b/arch/sparc/kernel/vmlinux.lds.S
@@ -89,6 +89,10 @@ SECTIONS
89 .data.cacheline_aligned : { 89 .data.cacheline_aligned : {
90 *(.data.cacheline_aligned) 90 *(.data.cacheline_aligned)
91 } 91 }
92 . = ALIGN(32);
93 .data.read_mostly : {
94 *(.data.read_mostly)
95 }
92 96
93 __bss_start = .; 97 __bss_start = .;
94 .sbss : { 98 .sbss : {
diff --git a/arch/sparc64/Kconfig b/arch/sparc64/Kconfig
index b810f2b7526a..3af378ddb6ae 100644
--- a/arch/sparc64/Kconfig
+++ b/arch/sparc64/Kconfig
@@ -14,6 +14,7 @@ config SPARC
14config SPARC64 14config SPARC64
15 bool 15 bool
16 default y 16 default y
17 select HAVE_IDE
17 help 18 help
18 SPARC is a family of RISC microprocessors designed and marketed by 19 SPARC is a family of RISC microprocessors designed and marketed by
19 Sun Microsystems, incorporated. This port covers the newer 64-bit 20 Sun Microsystems, incorporated. This port covers the newer 64-bit
@@ -40,6 +41,10 @@ config MMU
40 bool 41 bool
41 default y 42 default y
42 43
44config IOMMU_HELPER
45 bool
46 default y
47
43config QUICKLIST 48config QUICKLIST
44 bool 49 bool
45 default y 50 default y
diff --git a/arch/sparc64/kernel/iommu.c b/arch/sparc64/kernel/iommu.c
index 5623a4d59dff..d3276ebcfb47 100644
--- a/arch/sparc64/kernel/iommu.c
+++ b/arch/sparc64/kernel/iommu.c
@@ -1,6 +1,6 @@
1/* iommu.c: Generic sparc64 IOMMU support. 1/* iommu.c: Generic sparc64 IOMMU support.
2 * 2 *
3 * Copyright (C) 1999, 2007 David S. Miller (davem@davemloft.net) 3 * Copyright (C) 1999, 2007, 2008 David S. Miller (davem@davemloft.net)
4 * Copyright (C) 1999, 2000 Jakub Jelinek (jakub@redhat.com) 4 * Copyright (C) 1999, 2000 Jakub Jelinek (jakub@redhat.com)
5 */ 5 */
6 6
@@ -10,6 +10,7 @@
10#include <linux/device.h> 10#include <linux/device.h>
11#include <linux/dma-mapping.h> 11#include <linux/dma-mapping.h>
12#include <linux/errno.h> 12#include <linux/errno.h>
13#include <linux/iommu-helper.h>
13 14
14#ifdef CONFIG_PCI 15#ifdef CONFIG_PCI
15#include <linux/pci.h> 16#include <linux/pci.h>
@@ -41,7 +42,7 @@
41 "i" (ASI_PHYS_BYPASS_EC_E)) 42 "i" (ASI_PHYS_BYPASS_EC_E))
42 43
43/* Must be invoked under the IOMMU lock. */ 44/* Must be invoked under the IOMMU lock. */
44static void __iommu_flushall(struct iommu *iommu) 45static void iommu_flushall(struct iommu *iommu)
45{ 46{
46 if (iommu->iommu_flushinv) { 47 if (iommu->iommu_flushinv) {
47 iommu_write(iommu->iommu_flushinv, ~(u64)0); 48 iommu_write(iommu->iommu_flushinv, ~(u64)0);
@@ -83,54 +84,91 @@ static inline void iopte_make_dummy(struct iommu *iommu, iopte_t *iopte)
83 iopte_val(*iopte) = val; 84 iopte_val(*iopte) = val;
84} 85}
85 86
86/* Based largely upon the ppc64 iommu allocator. */ 87/* Based almost entirely upon the ppc64 iommu allocator. If you use the 'handle'
87static long arena_alloc(struct iommu *iommu, unsigned long npages) 88 * facility it must all be done in one pass while under the iommu lock.
89 *
90 * On sun4u platforms, we only flush the IOMMU once every time we've passed
91 * over the entire page table doing allocations. Therefore we only ever advance
92 * the hint and cannot backtrack it.
93 */
94unsigned long iommu_range_alloc(struct device *dev,
95 struct iommu *iommu,
96 unsigned long npages,
97 unsigned long *handle)
88{ 98{
99 unsigned long n, end, start, limit, boundary_size;
89 struct iommu_arena *arena = &iommu->arena; 100 struct iommu_arena *arena = &iommu->arena;
90 unsigned long n, i, start, end, limit; 101 int pass = 0;
91 int pass; 102
103 /* This allocator was derived from x86_64's bit string search */
104
105 /* Sanity check */
106 if (unlikely(npages == 0)) {
107 if (printk_ratelimit())
108 WARN_ON(1);
109 return DMA_ERROR_CODE;
110 }
111
112 if (handle && *handle)
113 start = *handle;
114 else
115 start = arena->hint;
92 116
93 limit = arena->limit; 117 limit = arena->limit;
94 start = arena->hint;
95 pass = 0;
96 118
97again: 119 /* The case below can happen if we have a small segment appended
98 n = find_next_zero_bit(arena->map, limit, start); 120 * to a large, or when the previous alloc was at the very end of
99 end = n + npages; 121 * the available space. If so, go back to the beginning and flush.
100 if (unlikely(end >= limit)) { 122 */
123 if (start >= limit) {
124 start = 0;
125 if (iommu->flush_all)
126 iommu->flush_all(iommu);
127 }
128
129 again:
130
131 if (dev)
132 boundary_size = ALIGN(dma_get_seg_boundary(dev) + 1,
133 1 << IO_PAGE_SHIFT);
134 else
135 boundary_size = ALIGN(1UL << 32, 1 << IO_PAGE_SHIFT);
136
137 n = iommu_area_alloc(arena->map, limit, start, npages, 0,
138 boundary_size >> IO_PAGE_SHIFT, 0);
139 if (n == -1) {
101 if (likely(pass < 1)) { 140 if (likely(pass < 1)) {
102 limit = start; 141 /* First failure, rescan from the beginning. */
103 start = 0; 142 start = 0;
104 __iommu_flushall(iommu); 143 if (iommu->flush_all)
144 iommu->flush_all(iommu);
105 pass++; 145 pass++;
106 goto again; 146 goto again;
107 } else { 147 } else {
108 /* Scanned the whole thing, give up. */ 148 /* Second failure, give up */
109 return -1; 149 return DMA_ERROR_CODE;
110 }
111 }
112
113 for (i = n; i < end; i++) {
114 if (test_bit(i, arena->map)) {
115 start = i + 1;
116 goto again;
117 } 150 }
118 } 151 }
119 152
120 for (i = n; i < end; i++) 153 end = n + npages;
121 __set_bit(i, arena->map);
122 154
123 arena->hint = end; 155 arena->hint = end;
124 156
157 /* Update handle for SG allocations */
158 if (handle)
159 *handle = end;
160
125 return n; 161 return n;
126} 162}
127 163
128static void arena_free(struct iommu_arena *arena, unsigned long base, unsigned long npages) 164void iommu_range_free(struct iommu *iommu, dma_addr_t dma_addr, unsigned long npages)
129{ 165{
130 unsigned long i; 166 struct iommu_arena *arena = &iommu->arena;
167 unsigned long entry;
131 168
132 for (i = base; i < (base + npages); i++) 169 entry = (dma_addr - iommu->page_table_map_base) >> IO_PAGE_SHIFT;
133 __clear_bit(i, arena->map); 170
171 iommu_area_free(arena->map, entry, npages);
134} 172}
135 173
136int iommu_table_init(struct iommu *iommu, int tsbsize, 174int iommu_table_init(struct iommu *iommu, int tsbsize,
@@ -156,6 +194,9 @@ int iommu_table_init(struct iommu *iommu, int tsbsize,
156 } 194 }
157 iommu->arena.limit = num_tsb_entries; 195 iommu->arena.limit = num_tsb_entries;
158 196
197 if (tlb_type != hypervisor)
198 iommu->flush_all = iommu_flushall;
199
159 /* Allocate and initialize the dummy page which we 200 /* Allocate and initialize the dummy page which we
160 * set inactive IO PTEs to point to. 201 * set inactive IO PTEs to point to.
161 */ 202 */
@@ -192,22 +233,18 @@ out_free_map:
192 return -ENOMEM; 233 return -ENOMEM;
193} 234}
194 235
195static inline iopte_t *alloc_npages(struct iommu *iommu, unsigned long npages) 236static inline iopte_t *alloc_npages(struct device *dev, struct iommu *iommu,
237 unsigned long npages)
196{ 238{
197 long entry; 239 unsigned long entry;
198 240
199 entry = arena_alloc(iommu, npages); 241 entry = iommu_range_alloc(dev, iommu, npages, NULL);
200 if (unlikely(entry < 0)) 242 if (unlikely(entry == DMA_ERROR_CODE))
201 return NULL; 243 return NULL;
202 244
203 return iommu->page_table + entry; 245 return iommu->page_table + entry;
204} 246}
205 247
206static inline void free_npages(struct iommu *iommu, dma_addr_t base, unsigned long npages)
207{
208 arena_free(&iommu->arena, base >> IO_PAGE_SHIFT, npages);
209}
210
211static int iommu_alloc_ctx(struct iommu *iommu) 248static int iommu_alloc_ctx(struct iommu *iommu)
212{ 249{
213 int lowest = iommu->ctx_lowest_free; 250 int lowest = iommu->ctx_lowest_free;
@@ -258,7 +295,7 @@ static void *dma_4u_alloc_coherent(struct device *dev, size_t size,
258 iommu = dev->archdata.iommu; 295 iommu = dev->archdata.iommu;
259 296
260 spin_lock_irqsave(&iommu->lock, flags); 297 spin_lock_irqsave(&iommu->lock, flags);
261 iopte = alloc_npages(iommu, size >> IO_PAGE_SHIFT); 298 iopte = alloc_npages(dev, iommu, size >> IO_PAGE_SHIFT);
262 spin_unlock_irqrestore(&iommu->lock, flags); 299 spin_unlock_irqrestore(&iommu->lock, flags);
263 300
264 if (unlikely(iopte == NULL)) { 301 if (unlikely(iopte == NULL)) {
@@ -296,7 +333,7 @@ static void dma_4u_free_coherent(struct device *dev, size_t size,
296 333
297 spin_lock_irqsave(&iommu->lock, flags); 334 spin_lock_irqsave(&iommu->lock, flags);
298 335
299 free_npages(iommu, dvma - iommu->page_table_map_base, npages); 336 iommu_range_free(iommu, dvma, npages);
300 337
301 spin_unlock_irqrestore(&iommu->lock, flags); 338 spin_unlock_irqrestore(&iommu->lock, flags);
302 339
@@ -327,7 +364,7 @@ static dma_addr_t dma_4u_map_single(struct device *dev, void *ptr, size_t sz,
327 npages >>= IO_PAGE_SHIFT; 364 npages >>= IO_PAGE_SHIFT;
328 365
329 spin_lock_irqsave(&iommu->lock, flags); 366 spin_lock_irqsave(&iommu->lock, flags);
330 base = alloc_npages(iommu, npages); 367 base = alloc_npages(dev, iommu, npages);
331 ctx = 0; 368 ctx = 0;
332 if (iommu->iommu_ctxflush) 369 if (iommu->iommu_ctxflush)
333 ctx = iommu_alloc_ctx(iommu); 370 ctx = iommu_alloc_ctx(iommu);
@@ -465,7 +502,7 @@ static void dma_4u_unmap_single(struct device *dev, dma_addr_t bus_addr,
465 for (i = 0; i < npages; i++) 502 for (i = 0; i < npages; i++)
466 iopte_make_dummy(iommu, base + i); 503 iopte_make_dummy(iommu, base + i);
467 504
468 free_npages(iommu, bus_addr - iommu->page_table_map_base, npages); 505 iommu_range_free(iommu, bus_addr, npages);
469 506
470 iommu_free_ctx(iommu, ctx); 507 iommu_free_ctx(iommu, ctx);
471 508
@@ -475,124 +512,209 @@ static void dma_4u_unmap_single(struct device *dev, dma_addr_t bus_addr,
475static int dma_4u_map_sg(struct device *dev, struct scatterlist *sglist, 512static int dma_4u_map_sg(struct device *dev, struct scatterlist *sglist,
476 int nelems, enum dma_data_direction direction) 513 int nelems, enum dma_data_direction direction)
477{ 514{
478 unsigned long flags, ctx, i, npages, iopte_protection; 515 struct scatterlist *s, *outs, *segstart;
479 struct scatterlist *sg; 516 unsigned long flags, handle, prot, ctx;
517 dma_addr_t dma_next = 0, dma_addr;
518 unsigned int max_seg_size;
519 int outcount, incount, i;
480 struct strbuf *strbuf; 520 struct strbuf *strbuf;
481 struct iommu *iommu; 521 struct iommu *iommu;
482 iopte_t *base; 522
483 u32 dma_base; 523 BUG_ON(direction == DMA_NONE);
484
485 /* Fast path single entry scatterlists. */
486 if (nelems == 1) {
487 sglist->dma_address =
488 dma_4u_map_single(dev, sg_virt(sglist),
489 sglist->length, direction);
490 if (unlikely(sglist->dma_address == DMA_ERROR_CODE))
491 return 0;
492 sglist->dma_length = sglist->length;
493 return 1;
494 }
495 524
496 iommu = dev->archdata.iommu; 525 iommu = dev->archdata.iommu;
497 strbuf = dev->archdata.stc; 526 strbuf = dev->archdata.stc;
498 527 if (nelems == 0 || !iommu)
499 if (unlikely(direction == DMA_NONE)) 528 return 0;
500 goto bad_no_ctx;
501
502 npages = calc_npages(sglist, nelems);
503 529
504 spin_lock_irqsave(&iommu->lock, flags); 530 spin_lock_irqsave(&iommu->lock, flags);
505 531
506 base = alloc_npages(iommu, npages);
507 ctx = 0; 532 ctx = 0;
508 if (iommu->iommu_ctxflush) 533 if (iommu->iommu_ctxflush)
509 ctx = iommu_alloc_ctx(iommu); 534 ctx = iommu_alloc_ctx(iommu);
510 535
511 spin_unlock_irqrestore(&iommu->lock, flags);
512
513 if (base == NULL)
514 goto bad;
515
516 dma_base = iommu->page_table_map_base +
517 ((base - iommu->page_table) << IO_PAGE_SHIFT);
518
519 if (strbuf->strbuf_enabled) 536 if (strbuf->strbuf_enabled)
520 iopte_protection = IOPTE_STREAMING(ctx); 537 prot = IOPTE_STREAMING(ctx);
521 else 538 else
522 iopte_protection = IOPTE_CONSISTENT(ctx); 539 prot = IOPTE_CONSISTENT(ctx);
523 if (direction != DMA_TO_DEVICE) 540 if (direction != DMA_TO_DEVICE)
524 iopte_protection |= IOPTE_WRITE; 541 prot |= IOPTE_WRITE;
525 542
526 for_each_sg(sglist, sg, nelems, i) { 543 outs = s = segstart = &sglist[0];
527 unsigned long paddr = SG_ENT_PHYS_ADDRESS(sg); 544 outcount = 1;
528 unsigned long slen = sg->length; 545 incount = nelems;
529 unsigned long this_npages; 546 handle = 0;
547
548 /* Init first segment length for backout at failure */
549 outs->dma_length = 0;
550
551 max_seg_size = dma_get_max_seg_size(dev);
552 for_each_sg(sglist, s, nelems, i) {
553 unsigned long paddr, npages, entry, slen;
554 iopte_t *base;
555
556 slen = s->length;
557 /* Sanity check */
558 if (slen == 0) {
559 dma_next = 0;
560 continue;
561 }
562 /* Allocate iommu entries for that segment */
563 paddr = (unsigned long) SG_ENT_PHYS_ADDRESS(s);
564 npages = iommu_num_pages(paddr, slen);
565 entry = iommu_range_alloc(dev, iommu, npages, &handle);
566
567 /* Handle failure */
568 if (unlikely(entry == DMA_ERROR_CODE)) {
569 if (printk_ratelimit())
570 printk(KERN_INFO "iommu_alloc failed, iommu %p paddr %lx"
571 " npages %lx\n", iommu, paddr, npages);
572 goto iommu_map_failed;
573 }
530 574
531 this_npages = iommu_num_pages(paddr, slen); 575 base = iommu->page_table + entry;
532 576
533 sg->dma_address = dma_base | (paddr & ~IO_PAGE_MASK); 577 /* Convert entry to a dma_addr_t */
534 sg->dma_length = slen; 578 dma_addr = iommu->page_table_map_base +
579 (entry << IO_PAGE_SHIFT);
580 dma_addr |= (s->offset & ~IO_PAGE_MASK);
535 581
582 /* Insert into HW table */
536 paddr &= IO_PAGE_MASK; 583 paddr &= IO_PAGE_MASK;
537 while (this_npages--) { 584 while (npages--) {
538 iopte_val(*base) = iopte_protection | paddr; 585 iopte_val(*base) = prot | paddr;
539
540 base++; 586 base++;
541 paddr += IO_PAGE_SIZE; 587 paddr += IO_PAGE_SIZE;
542 dma_base += IO_PAGE_SIZE;
543 } 588 }
589
590 /* If we are in an open segment, try merging */
591 if (segstart != s) {
592 /* We cannot merge if:
593 * - allocated dma_addr isn't contiguous to previous allocation
594 */
595 if ((dma_addr != dma_next) ||
596 (outs->dma_length + s->length > max_seg_size)) {
597 /* Can't merge: create a new segment */
598 segstart = s;
599 outcount++;
600 outs = sg_next(outs);
601 } else {
602 outs->dma_length += s->length;
603 }
604 }
605
606 if (segstart == s) {
607 /* This is a new segment, fill entries */
608 outs->dma_address = dma_addr;
609 outs->dma_length = slen;
610 }
611
612 /* Calculate next page pointer for contiguous check */
613 dma_next = dma_addr + slen;
544 } 614 }
545 615
546 return nelems; 616 spin_unlock_irqrestore(&iommu->lock, flags);
617
618 if (outcount < incount) {
619 outs = sg_next(outs);
620 outs->dma_address = DMA_ERROR_CODE;
621 outs->dma_length = 0;
622 }
623
624 return outcount;
625
626iommu_map_failed:
627 for_each_sg(sglist, s, nelems, i) {
628 if (s->dma_length != 0) {
629 unsigned long vaddr, npages, entry, i;
630 iopte_t *base;
631
632 vaddr = s->dma_address & IO_PAGE_MASK;
633 npages = iommu_num_pages(s->dma_address, s->dma_length);
634 iommu_range_free(iommu, vaddr, npages);
635
636 entry = (vaddr - iommu->page_table_map_base)
637 >> IO_PAGE_SHIFT;
638 base = iommu->page_table + entry;
639
640 for (i = 0; i < npages; i++)
641 iopte_make_dummy(iommu, base + i);
642
643 s->dma_address = DMA_ERROR_CODE;
644 s->dma_length = 0;
645 }
646 if (s == outs)
647 break;
648 }
649 spin_unlock_irqrestore(&iommu->lock, flags);
547 650
548bad:
549 iommu_free_ctx(iommu, ctx);
550bad_no_ctx:
551 if (printk_ratelimit())
552 WARN_ON(1);
553 return 0; 651 return 0;
554} 652}
555 653
654/* If contexts are being used, they are the same in all of the mappings
655 * we make for a particular SG.
656 */
657static unsigned long fetch_sg_ctx(struct iommu *iommu, struct scatterlist *sg)
658{
659 unsigned long ctx = 0;
660
661 if (iommu->iommu_ctxflush) {
662 iopte_t *base;
663 u32 bus_addr;
664
665 bus_addr = sg->dma_address & IO_PAGE_MASK;
666 base = iommu->page_table +
667 ((bus_addr - iommu->page_table_map_base) >> IO_PAGE_SHIFT);
668
669 ctx = (iopte_val(*base) & IOPTE_CONTEXT) >> 47UL;
670 }
671 return ctx;
672}
673
556static void dma_4u_unmap_sg(struct device *dev, struct scatterlist *sglist, 674static void dma_4u_unmap_sg(struct device *dev, struct scatterlist *sglist,
557 int nelems, enum dma_data_direction direction) 675 int nelems, enum dma_data_direction direction)
558{ 676{
559 unsigned long flags, ctx, i, npages; 677 unsigned long flags, ctx;
678 struct scatterlist *sg;
560 struct strbuf *strbuf; 679 struct strbuf *strbuf;
561 struct iommu *iommu; 680 struct iommu *iommu;
562 iopte_t *base;
563 u32 bus_addr;
564 681
565 if (unlikely(direction == DMA_NONE)) { 682 BUG_ON(direction == DMA_NONE);
566 if (printk_ratelimit())
567 WARN_ON(1);
568 }
569 683
570 iommu = dev->archdata.iommu; 684 iommu = dev->archdata.iommu;
571 strbuf = dev->archdata.stc; 685 strbuf = dev->archdata.stc;
572 686
573 bus_addr = sglist->dma_address & IO_PAGE_MASK; 687 ctx = fetch_sg_ctx(iommu, sglist);
574 688
575 npages = calc_npages(sglist, nelems); 689 spin_lock_irqsave(&iommu->lock, flags);
576 690
577 base = iommu->page_table + 691 sg = sglist;
578 ((bus_addr - iommu->page_table_map_base) >> IO_PAGE_SHIFT); 692 while (nelems--) {
693 dma_addr_t dma_handle = sg->dma_address;
694 unsigned int len = sg->dma_length;
695 unsigned long npages, entry;
696 iopte_t *base;
697 int i;
579 698
580 spin_lock_irqsave(&iommu->lock, flags); 699 if (!len)
700 break;
701 npages = iommu_num_pages(dma_handle, len);
702 iommu_range_free(iommu, dma_handle, npages);
581 703
582 /* Record the context, if any. */ 704 entry = ((dma_handle - iommu->page_table_map_base)
583 ctx = 0; 705 >> IO_PAGE_SHIFT);
584 if (iommu->iommu_ctxflush) 706 base = iommu->page_table + entry;
585 ctx = (iopte_val(*base) & IOPTE_CONTEXT) >> 47UL;
586 707
587 /* Step 1: Kick data out of streaming buffers if necessary. */ 708 dma_handle &= IO_PAGE_MASK;
588 if (strbuf->strbuf_enabled) 709 if (strbuf->strbuf_enabled)
589 strbuf_flush(strbuf, iommu, bus_addr, ctx, npages, direction); 710 strbuf_flush(strbuf, iommu, dma_handle, ctx,
711 npages, direction);
590 712
591 /* Step 2: Clear out the TSB entries. */ 713 for (i = 0; i < npages; i++)
592 for (i = 0; i < npages; i++) 714 iopte_make_dummy(iommu, base + i);
593 iopte_make_dummy(iommu, base + i);
594 715
595 free_npages(iommu, bus_addr - iommu->page_table_map_base, npages); 716 sg = sg_next(sg);
717 }
596 718
597 iommu_free_ctx(iommu, ctx); 719 iommu_free_ctx(iommu, ctx);
598 720
diff --git a/arch/sparc64/kernel/iommu_common.h b/arch/sparc64/kernel/iommu_common.h
index 4b5cafa2877a..0713bd58499c 100644
--- a/arch/sparc64/kernel/iommu_common.h
+++ b/arch/sparc64/kernel/iommu_common.h
@@ -1,9 +1,11 @@
1/* $Id: iommu_common.h,v 1.5 2001/12/11 09:41:01 davem Exp $ 1/* iommu_common.h: UltraSparc SBUS/PCI common iommu declarations.
2 * iommu_common.h: UltraSparc SBUS/PCI common iommu declarations.
3 * 2 *
4 * Copyright (C) 1999 David S. Miller (davem@redhat.com) 3 * Copyright (C) 1999, 2008 David S. Miller (davem@davemloft.net)
5 */ 4 */
6 5
6#ifndef _IOMMU_COMMON_H
7#define _IOMMU_COMMON_H
8
7#include <linux/kernel.h> 9#include <linux/kernel.h>
8#include <linux/types.h> 10#include <linux/types.h>
9#include <linux/sched.h> 11#include <linux/sched.h>
@@ -56,21 +58,12 @@ static inline unsigned long calc_npages(struct scatterlist *sglist, int nelems)
56 return npages; 58 return npages;
57} 59}
58 60
59/* You are _strongly_ advised to enable the following debugging code 61extern unsigned long iommu_range_alloc(struct device *dev,
60 * any time you make changes to the sg code below, run it for a while 62 struct iommu *iommu,
61 * with filesystems mounted read-only before buying the farm... -DaveM 63 unsigned long npages,
62 */ 64 unsigned long *handle);
63#undef VERIFY_SG 65extern void iommu_range_free(struct iommu *iommu,
64 66 dma_addr_t dma_addr,
65#ifdef VERIFY_SG 67 unsigned long npages);
66extern void verify_sglist(struct scatterlist *sg, int nents, iopte_t *iopte, int npages);
67#endif
68
69/* Two addresses are "virtually contiguous" if and only if:
70 * 1) They are equal, or...
71 * 2) They are both on a page boundary
72 */
73#define VCONTIG(__X, __Y) (((__X) == (__Y)) || \
74 (((__X) | (__Y)) << (64UL - PAGE_SHIFT)) == 0UL)
75 68
76extern unsigned long prepare_sg(struct device *dev, struct scatterlist *sg, int nents); 69#endif /* _IOMMU_COMMON_H */
diff --git a/arch/sparc64/kernel/kprobes.c b/arch/sparc64/kernel/kprobes.c
index d94f901d321e..34fc3ddd5002 100644
--- a/arch/sparc64/kernel/kprobes.c
+++ b/arch/sparc64/kernel/kprobes.c
@@ -480,8 +480,117 @@ int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
480 return 0; 480 return 0;
481} 481}
482 482
483/* architecture specific initialization */ 483/* Called with kretprobe_lock held. The value stored in the return
484int arch_init_kprobes(void) 484 * address register is actually 2 instructions before where the
485 * callee will return to. Sequences usually look something like this
486 *
487 * call some_function <--- return register points here
488 * nop <--- call delay slot
489 * whatever <--- where callee returns to
490 *
491 * To keep trampoline_probe_handler logic simpler, we normalize the
492 * value kept in ri->ret_addr so we don't need to keep adjusting it
493 * back and forth.
494 */
495void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
496 struct pt_regs *regs)
497{
498 ri->ret_addr = (kprobe_opcode_t *)(regs->u_regs[UREG_RETPC] + 8);
499
500 /* Replace the return addr with trampoline addr */
501 regs->u_regs[UREG_RETPC] =
502 ((unsigned long)kretprobe_trampoline) - 8;
503}
504
505/*
506 * Called when the probe at kretprobe trampoline is hit
507 */
508int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
509{
510 struct kretprobe_instance *ri = NULL;
511 struct hlist_head *head, empty_rp;
512 struct hlist_node *node, *tmp;
513 unsigned long flags, orig_ret_address = 0;
514 unsigned long trampoline_address =(unsigned long)&kretprobe_trampoline;
515
516 INIT_HLIST_HEAD(&empty_rp);
517 spin_lock_irqsave(&kretprobe_lock, flags);
518 head = kretprobe_inst_table_head(current);
519
520 /*
521 * It is possible to have multiple instances associated with a given
522 * task either because an multiple functions in the call path
523 * have a return probe installed on them, and/or more then one return
524 * return probe was registered for a target function.
525 *
526 * We can handle this because:
527 * - instances are always inserted at the head of the list
528 * - when multiple return probes are registered for the same
529 * function, the first instance's ret_addr will point to the
530 * real return address, and all the rest will point to
531 * kretprobe_trampoline
532 */
533 hlist_for_each_entry_safe(ri, node, tmp, head, hlist) {
534 if (ri->task != current)
535 /* another task is sharing our hash bucket */
536 continue;
537
538 if (ri->rp && ri->rp->handler)
539 ri->rp->handler(ri, regs);
540
541 orig_ret_address = (unsigned long)ri->ret_addr;
542 recycle_rp_inst(ri, &empty_rp);
543
544 if (orig_ret_address != trampoline_address)
545 /*
546 * This is the real return address. Any other
547 * instances associated with this task are for
548 * other calls deeper on the call stack
549 */
550 break;
551 }
552
553 kretprobe_assert(ri, orig_ret_address, trampoline_address);
554 regs->tpc = orig_ret_address;
555 regs->tnpc = orig_ret_address + 4;
556
557 reset_current_kprobe();
558 spin_unlock_irqrestore(&kretprobe_lock, flags);
559 preempt_enable_no_resched();
560
561 hlist_for_each_entry_safe(ri, node, tmp, &empty_rp, hlist) {
562 hlist_del(&ri->hlist);
563 kfree(ri);
564 }
565 /*
566 * By returning a non-zero value, we are telling
567 * kprobe_handler() that we don't want the post_handler
568 * to run (and have re-enabled preemption)
569 */
570 return 1;
571}
572
573void kretprobe_trampoline_holder(void)
574{
575 asm volatile(".global kretprobe_trampoline\n"
576 "kretprobe_trampoline:\n"
577 "\tnop\n"
578 "\tnop\n");
579}
580static struct kprobe trampoline_p = {
581 .addr = (kprobe_opcode_t *) &kretprobe_trampoline,
582 .pre_handler = trampoline_probe_handler
583};
584
585int __init arch_init_kprobes(void)
485{ 586{
587 return register_kprobe(&trampoline_p);
588}
589
590int __kprobes arch_trampoline_kprobe(struct kprobe *p)
591{
592 if (p->addr == (kprobe_opcode_t *)&kretprobe_trampoline)
593 return 1;
594
486 return 0; 595 return 0;
487} 596}
diff --git a/arch/sparc64/kernel/of_device.c b/arch/sparc64/kernel/of_device.c
index fc5c0cc793b8..0fd9db95b896 100644
--- a/arch/sparc64/kernel/of_device.c
+++ b/arch/sparc64/kernel/of_device.c
@@ -868,29 +868,3 @@ static int __init of_debug(char *str)
868} 868}
869 869
870__setup("of_debug=", of_debug); 870__setup("of_debug=", of_debug);
871
872struct of_device* of_platform_device_create(struct device_node *np,
873 const char *bus_id,
874 struct device *parent,
875 struct bus_type *bus)
876{
877 struct of_device *dev;
878
879 dev = kzalloc(sizeof(*dev), GFP_KERNEL);
880 if (!dev)
881 return NULL;
882
883 dev->dev.parent = parent;
884 dev->dev.bus = bus;
885 dev->dev.release = of_release_dev;
886
887 strlcpy(dev->dev.bus_id, bus_id, BUS_ID_SIZE);
888
889 if (of_device_register(dev) != 0) {
890 kfree(dev);
891 return NULL;
892 }
893
894 return dev;
895}
896EXPORT_SYMBOL(of_platform_device_create);
diff --git a/arch/sparc64/kernel/pci_sun4v.c b/arch/sparc64/kernel/pci_sun4v.c
index 61baf8dc095e..ddca6c6c0b49 100644
--- a/arch/sparc64/kernel/pci_sun4v.c
+++ b/arch/sparc64/kernel/pci_sun4v.c
@@ -1,6 +1,6 @@
1/* pci_sun4v.c: SUN4V specific PCI controller support. 1/* pci_sun4v.c: SUN4V specific PCI controller support.
2 * 2 *
3 * Copyright (C) 2006, 2007 David S. Miller (davem@davemloft.net) 3 * Copyright (C) 2006, 2007, 2008 David S. Miller (davem@davemloft.net)
4 */ 4 */
5 5
6#include <linux/kernel.h> 6#include <linux/kernel.h>
@@ -89,6 +89,17 @@ static long iommu_batch_flush(struct iommu_batch *p)
89 return 0; 89 return 0;
90} 90}
91 91
92static inline void iommu_batch_new_entry(unsigned long entry)
93{
94 struct iommu_batch *p = &__get_cpu_var(iommu_batch);
95
96 if (p->entry + p->npages == entry)
97 return;
98 if (p->entry != ~0UL)
99 iommu_batch_flush(p);
100 p->entry = entry;
101}
102
92/* Interrupts must be disabled. */ 103/* Interrupts must be disabled. */
93static inline long iommu_batch_add(u64 phys_page) 104static inline long iommu_batch_add(u64 phys_page)
94{ 105{
@@ -113,54 +124,6 @@ static inline long iommu_batch_end(void)
113 return iommu_batch_flush(p); 124 return iommu_batch_flush(p);
114} 125}
115 126
116static long arena_alloc(struct iommu_arena *arena, unsigned long npages)
117{
118 unsigned long n, i, start, end, limit;
119 int pass;
120
121 limit = arena->limit;
122 start = arena->hint;
123 pass = 0;
124
125again:
126 n = find_next_zero_bit(arena->map, limit, start);
127 end = n + npages;
128 if (unlikely(end >= limit)) {
129 if (likely(pass < 1)) {
130 limit = start;
131 start = 0;
132 pass++;
133 goto again;
134 } else {
135 /* Scanned the whole thing, give up. */
136 return -1;
137 }
138 }
139
140 for (i = n; i < end; i++) {
141 if (test_bit(i, arena->map)) {
142 start = i + 1;
143 goto again;
144 }
145 }
146
147 for (i = n; i < end; i++)
148 __set_bit(i, arena->map);
149
150 arena->hint = end;
151
152 return n;
153}
154
155static void arena_free(struct iommu_arena *arena, unsigned long base,
156 unsigned long npages)
157{
158 unsigned long i;
159
160 for (i = base; i < (base + npages); i++)
161 __clear_bit(i, arena->map);
162}
163
164static void *dma_4v_alloc_coherent(struct device *dev, size_t size, 127static void *dma_4v_alloc_coherent(struct device *dev, size_t size,
165 dma_addr_t *dma_addrp, gfp_t gfp) 128 dma_addr_t *dma_addrp, gfp_t gfp)
166{ 129{
@@ -185,11 +148,11 @@ static void *dma_4v_alloc_coherent(struct device *dev, size_t size,
185 iommu = dev->archdata.iommu; 148 iommu = dev->archdata.iommu;
186 149
187 spin_lock_irqsave(&iommu->lock, flags); 150 spin_lock_irqsave(&iommu->lock, flags);
188 entry = arena_alloc(&iommu->arena, npages); 151 entry = iommu_range_alloc(dev, iommu, npages, NULL);
189 spin_unlock_irqrestore(&iommu->lock, flags); 152 spin_unlock_irqrestore(&iommu->lock, flags);
190 153
191 if (unlikely(entry < 0L)) 154 if (unlikely(entry == DMA_ERROR_CODE))
192 goto arena_alloc_fail; 155 goto range_alloc_fail;
193 156
194 *dma_addrp = (iommu->page_table_map_base + 157 *dma_addrp = (iommu->page_table_map_base +
195 (entry << IO_PAGE_SHIFT)); 158 (entry << IO_PAGE_SHIFT));
@@ -219,10 +182,10 @@ static void *dma_4v_alloc_coherent(struct device *dev, size_t size,
219iommu_map_fail: 182iommu_map_fail:
220 /* Interrupts are disabled. */ 183 /* Interrupts are disabled. */
221 spin_lock(&iommu->lock); 184 spin_lock(&iommu->lock);
222 arena_free(&iommu->arena, entry, npages); 185 iommu_range_free(iommu, *dma_addrp, npages);
223 spin_unlock_irqrestore(&iommu->lock, flags); 186 spin_unlock_irqrestore(&iommu->lock, flags);
224 187
225arena_alloc_fail: 188range_alloc_fail:
226 free_pages(first_page, order); 189 free_pages(first_page, order);
227 return NULL; 190 return NULL;
228} 191}
@@ -243,7 +206,7 @@ static void dma_4v_free_coherent(struct device *dev, size_t size, void *cpu,
243 206
244 spin_lock_irqsave(&iommu->lock, flags); 207 spin_lock_irqsave(&iommu->lock, flags);
245 208
246 arena_free(&iommu->arena, entry, npages); 209 iommu_range_free(iommu, dvma, npages);
247 210
248 do { 211 do {
249 unsigned long num; 212 unsigned long num;
@@ -281,10 +244,10 @@ static dma_addr_t dma_4v_map_single(struct device *dev, void *ptr, size_t sz,
281 npages >>= IO_PAGE_SHIFT; 244 npages >>= IO_PAGE_SHIFT;
282 245
283 spin_lock_irqsave(&iommu->lock, flags); 246 spin_lock_irqsave(&iommu->lock, flags);
284 entry = arena_alloc(&iommu->arena, npages); 247 entry = iommu_range_alloc(dev, iommu, npages, NULL);
285 spin_unlock_irqrestore(&iommu->lock, flags); 248 spin_unlock_irqrestore(&iommu->lock, flags);
286 249
287 if (unlikely(entry < 0L)) 250 if (unlikely(entry == DMA_ERROR_CODE))
288 goto bad; 251 goto bad;
289 252
290 bus_addr = (iommu->page_table_map_base + 253 bus_addr = (iommu->page_table_map_base +
@@ -319,7 +282,7 @@ bad:
319iommu_map_fail: 282iommu_map_fail:
320 /* Interrupts are disabled. */ 283 /* Interrupts are disabled. */
321 spin_lock(&iommu->lock); 284 spin_lock(&iommu->lock);
322 arena_free(&iommu->arena, entry, npages); 285 iommu_range_free(iommu, bus_addr, npages);
323 spin_unlock_irqrestore(&iommu->lock, flags); 286 spin_unlock_irqrestore(&iommu->lock, flags);
324 287
325 return DMA_ERROR_CODE; 288 return DMA_ERROR_CODE;
@@ -350,9 +313,9 @@ static void dma_4v_unmap_single(struct device *dev, dma_addr_t bus_addr,
350 313
351 spin_lock_irqsave(&iommu->lock, flags); 314 spin_lock_irqsave(&iommu->lock, flags);
352 315
353 entry = (bus_addr - iommu->page_table_map_base) >> IO_PAGE_SHIFT; 316 iommu_range_free(iommu, bus_addr, npages);
354 arena_free(&iommu->arena, entry, npages);
355 317
318 entry = (bus_addr - iommu->page_table_map_base) >> IO_PAGE_SHIFT;
356 do { 319 do {
357 unsigned long num; 320 unsigned long num;
358 321
@@ -368,88 +331,131 @@ static void dma_4v_unmap_single(struct device *dev, dma_addr_t bus_addr,
368static int dma_4v_map_sg(struct device *dev, struct scatterlist *sglist, 331static int dma_4v_map_sg(struct device *dev, struct scatterlist *sglist,
369 int nelems, enum dma_data_direction direction) 332 int nelems, enum dma_data_direction direction)
370{ 333{
371 unsigned long flags, npages, i, prot; 334 struct scatterlist *s, *outs, *segstart;
372 struct scatterlist *sg; 335 unsigned long flags, handle, prot;
336 dma_addr_t dma_next = 0, dma_addr;
337 unsigned int max_seg_size;
338 int outcount, incount, i;
373 struct iommu *iommu; 339 struct iommu *iommu;
374 long entry, err; 340 long err;
375 u32 dma_base; 341
376 342 BUG_ON(direction == DMA_NONE);
377 /* Fast path single entry scatterlists. */
378 if (nelems == 1) {
379 sglist->dma_address =
380 dma_4v_map_single(dev, sg_virt(sglist),
381 sglist->length, direction);
382 if (unlikely(sglist->dma_address == DMA_ERROR_CODE))
383 return 0;
384 sglist->dma_length = sglist->length;
385 return 1;
386 }
387 343
388 iommu = dev->archdata.iommu; 344 iommu = dev->archdata.iommu;
345 if (nelems == 0 || !iommu)
346 return 0;
389 347
390 if (unlikely(direction == DMA_NONE)) 348 prot = HV_PCI_MAP_ATTR_READ;
391 goto bad; 349 if (direction != DMA_TO_DEVICE)
392 350 prot |= HV_PCI_MAP_ATTR_WRITE;
393 npages = calc_npages(sglist, nelems);
394 351
395 spin_lock_irqsave(&iommu->lock, flags); 352 outs = s = segstart = &sglist[0];
396 entry = arena_alloc(&iommu->arena, npages); 353 outcount = 1;
397 spin_unlock_irqrestore(&iommu->lock, flags); 354 incount = nelems;
355 handle = 0;
398 356
399 if (unlikely(entry < 0L)) 357 /* Init first segment length for backout at failure */
400 goto bad; 358 outs->dma_length = 0;
401 359
402 dma_base = iommu->page_table_map_base + 360 spin_lock_irqsave(&iommu->lock, flags);
403 (entry << IO_PAGE_SHIFT);
404 361
405 prot = HV_PCI_MAP_ATTR_READ; 362 iommu_batch_start(dev, prot, ~0UL);
406 if (direction != DMA_TO_DEVICE)
407 prot |= HV_PCI_MAP_ATTR_WRITE;
408 363
409 local_irq_save(flags); 364 max_seg_size = dma_get_max_seg_size(dev);
365 for_each_sg(sglist, s, nelems, i) {
366 unsigned long paddr, npages, entry, slen;
410 367
411 iommu_batch_start(dev, prot, entry); 368 slen = s->length;
369 /* Sanity check */
370 if (slen == 0) {
371 dma_next = 0;
372 continue;
373 }
374 /* Allocate iommu entries for that segment */
375 paddr = (unsigned long) SG_ENT_PHYS_ADDRESS(s);
376 npages = iommu_num_pages(paddr, slen);
377 entry = iommu_range_alloc(dev, iommu, npages, &handle);
412 378
413 for_each_sg(sglist, sg, nelems, i) { 379 /* Handle failure */
414 unsigned long paddr = SG_ENT_PHYS_ADDRESS(sg); 380 if (unlikely(entry == DMA_ERROR_CODE)) {
415 unsigned long slen = sg->length; 381 if (printk_ratelimit())
416 unsigned long this_npages; 382 printk(KERN_INFO "iommu_alloc failed, iommu %p paddr %lx"
383 " npages %lx\n", iommu, paddr, npages);
384 goto iommu_map_failed;
385 }
417 386
418 this_npages = iommu_num_pages(paddr, slen); 387 iommu_batch_new_entry(entry);
419 388
420 sg->dma_address = dma_base | (paddr & ~IO_PAGE_MASK); 389 /* Convert entry to a dma_addr_t */
421 sg->dma_length = slen; 390 dma_addr = iommu->page_table_map_base +
391 (entry << IO_PAGE_SHIFT);
392 dma_addr |= (s->offset & ~IO_PAGE_MASK);
422 393
394 /* Insert into HW table */
423 paddr &= IO_PAGE_MASK; 395 paddr &= IO_PAGE_MASK;
424 while (this_npages--) { 396 while (npages--) {
425 err = iommu_batch_add(paddr); 397 err = iommu_batch_add(paddr);
426 if (unlikely(err < 0L)) { 398 if (unlikely(err < 0L))
427 local_irq_restore(flags);
428 goto iommu_map_failed; 399 goto iommu_map_failed;
400 paddr += IO_PAGE_SIZE;
401 }
402
403 /* If we are in an open segment, try merging */
404 if (segstart != s) {
405 /* We cannot merge if:
406 * - allocated dma_addr isn't contiguous to previous allocation
407 */
408 if ((dma_addr != dma_next) ||
409 (outs->dma_length + s->length > max_seg_size)) {
410 /* Can't merge: create a new segment */
411 segstart = s;
412 outcount++;
413 outs = sg_next(outs);
414 } else {
415 outs->dma_length += s->length;
429 } 416 }
417 }
430 418
431 paddr += IO_PAGE_SIZE; 419 if (segstart == s) {
432 dma_base += IO_PAGE_SIZE; 420 /* This is a new segment, fill entries */
421 outs->dma_address = dma_addr;
422 outs->dma_length = slen;
433 } 423 }
424
425 /* Calculate next page pointer for contiguous check */
426 dma_next = dma_addr + slen;
434 } 427 }
435 428
436 err = iommu_batch_end(); 429 err = iommu_batch_end();
437 430
438 local_irq_restore(flags);
439
440 if (unlikely(err < 0L)) 431 if (unlikely(err < 0L))
441 goto iommu_map_failed; 432 goto iommu_map_failed;
442 433
443 return nelems; 434 spin_unlock_irqrestore(&iommu->lock, flags);
444 435
445bad: 436 if (outcount < incount) {
446 if (printk_ratelimit()) 437 outs = sg_next(outs);
447 WARN_ON(1); 438 outs->dma_address = DMA_ERROR_CODE;
448 return 0; 439 outs->dma_length = 0;
440 }
441
442 return outcount;
449 443
450iommu_map_failed: 444iommu_map_failed:
451 spin_lock_irqsave(&iommu->lock, flags); 445 for_each_sg(sglist, s, nelems, i) {
452 arena_free(&iommu->arena, entry, npages); 446 if (s->dma_length != 0) {
447 unsigned long vaddr, npages;
448
449 vaddr = s->dma_address & IO_PAGE_MASK;
450 npages = iommu_num_pages(s->dma_address, s->dma_length);
451 iommu_range_free(iommu, vaddr, npages);
452 /* XXX demap? XXX */
453 s->dma_address = DMA_ERROR_CODE;
454 s->dma_length = 0;
455 }
456 if (s == outs)
457 break;
458 }
453 spin_unlock_irqrestore(&iommu->lock, flags); 459 spin_unlock_irqrestore(&iommu->lock, flags);
454 460
455 return 0; 461 return 0;
@@ -458,39 +464,43 @@ iommu_map_failed:
458static void dma_4v_unmap_sg(struct device *dev, struct scatterlist *sglist, 464static void dma_4v_unmap_sg(struct device *dev, struct scatterlist *sglist,
459 int nelems, enum dma_data_direction direction) 465 int nelems, enum dma_data_direction direction)
460{ 466{
461 unsigned long flags, npages;
462 struct pci_pbm_info *pbm; 467 struct pci_pbm_info *pbm;
463 u32 devhandle, bus_addr; 468 struct scatterlist *sg;
464 struct iommu *iommu; 469 struct iommu *iommu;
465 long entry; 470 unsigned long flags;
471 u32 devhandle;
466 472
467 if (unlikely(direction == DMA_NONE)) { 473 BUG_ON(direction == DMA_NONE);
468 if (printk_ratelimit())
469 WARN_ON(1);
470 }
471 474
472 iommu = dev->archdata.iommu; 475 iommu = dev->archdata.iommu;
473 pbm = dev->archdata.host_controller; 476 pbm = dev->archdata.host_controller;
474 devhandle = pbm->devhandle; 477 devhandle = pbm->devhandle;
475 478
476 bus_addr = sglist->dma_address & IO_PAGE_MASK;
477
478 npages = calc_npages(sglist, nelems);
479
480 entry = ((bus_addr - iommu->page_table_map_base) >> IO_PAGE_SHIFT);
481
482 spin_lock_irqsave(&iommu->lock, flags); 479 spin_lock_irqsave(&iommu->lock, flags);
483 480
484 arena_free(&iommu->arena, entry, npages); 481 sg = sglist;
485 482 while (nelems--) {
486 do { 483 dma_addr_t dma_handle = sg->dma_address;
487 unsigned long num; 484 unsigned int len = sg->dma_length;
485 unsigned long npages, entry;
486
487 if (!len)
488 break;
489 npages = iommu_num_pages(dma_handle, len);
490 iommu_range_free(iommu, dma_handle, npages);
491
492 entry = ((dma_handle - iommu->page_table_map_base) >> IO_PAGE_SHIFT);
493 while (npages) {
494 unsigned long num;
495
496 num = pci_sun4v_iommu_demap(devhandle, HV_PCI_TSBID(0, entry),
497 npages);
498 entry += num;
499 npages -= num;
500 }
488 501
489 num = pci_sun4v_iommu_demap(devhandle, HV_PCI_TSBID(0, entry), 502 sg = sg_next(sg);
490 npages); 503 }
491 entry += num;
492 npages -= num;
493 } while (npages != 0);
494 504
495 spin_unlock_irqrestore(&iommu->lock, flags); 505 spin_unlock_irqrestore(&iommu->lock, flags);
496} 506}
diff --git a/arch/v850/Kconfig b/arch/v850/Kconfig
index 7b6d3716efca..4379f43505ef 100644
--- a/arch/v850/Kconfig
+++ b/arch/v850/Kconfig
@@ -78,6 +78,7 @@ config MCA
78config V850 78config V850
79 bool 79 bool
80 default y 80 default y
81 select HAVE_IDE
81 82
82menu "Processor type and features" 83menu "Processor type and features"
83 84
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 65a70b777c12..aaed1a3b92d6 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -18,6 +18,7 @@ config X86_64
18### Arch settings 18### Arch settings
19config X86 19config X86
20 def_bool y 20 def_bool y
21 select HAVE_IDE
21 select HAVE_OPROFILE 22 select HAVE_OPROFILE
22 select HAVE_KPROBES 23 select HAVE_KPROBES
23 24
@@ -102,6 +103,9 @@ config ARCH_HAS_ILOG2_U32
102config ARCH_HAS_ILOG2_U64 103config ARCH_HAS_ILOG2_U64
103 def_bool n 104 def_bool n
104 105
106config ARCH_HAS_CPU_IDLE_WAIT
107 def_bool y
108
105config GENERIC_CALIBRATE_DELAY 109config GENERIC_CALIBRATE_DELAY
106 def_bool y 110 def_bool y
107 111
diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug
index fa555148823d..864affc9a7b0 100644
--- a/arch/x86/Kconfig.debug
+++ b/arch/x86/Kconfig.debug
@@ -34,13 +34,9 @@ config DEBUG_STACK_USAGE
34 34
35 This option will slow down process creation somewhat. 35 This option will slow down process creation somewhat.
36 36
37comment "Page alloc debug is incompatible with Software Suspend on i386"
38 depends on DEBUG_KERNEL && HIBERNATION
39 depends on X86_32
40
41config DEBUG_PAGEALLOC 37config DEBUG_PAGEALLOC
42 bool "Debug page memory allocations" 38 bool "Debug page memory allocations"
43 depends on DEBUG_KERNEL && X86_32 39 depends on DEBUG_KERNEL
44 help 40 help
45 Unmap pages from the kernel linear mapping after free_pages(). 41 Unmap pages from the kernel linear mapping after free_pages().
46 This results in a large slowdown, but helps to find certain types 42 This results in a large slowdown, but helps to find certain types
diff --git a/arch/x86/Makefile b/arch/x86/Makefile
index 364865b1b08d..204af43535c5 100644
--- a/arch/x86/Makefile
+++ b/arch/x86/Makefile
@@ -191,8 +191,10 @@ drivers-$(CONFIG_PCI) += arch/x86/pci/
191# must be linked after kernel/ 191# must be linked after kernel/
192drivers-$(CONFIG_OPROFILE) += arch/x86/oprofile/ 192drivers-$(CONFIG_OPROFILE) += arch/x86/oprofile/
193 193
194ifeq ($(CONFIG_X86_32),y) 194# suspend and hibernation support
195drivers-$(CONFIG_PM) += arch/x86/power/ 195drivers-$(CONFIG_PM) += arch/x86/power/
196
197ifeq ($(CONFIG_X86_32),y)
196drivers-$(CONFIG_FB) += arch/x86/video/ 198drivers-$(CONFIG_FB) += arch/x86/video/
197endif 199endif
198 200
diff --git a/arch/x86/boot/.gitignore b/arch/x86/boot/.gitignore
index 18465143cfa2..b1bdc4c6f9f2 100644
--- a/arch/x86/boot/.gitignore
+++ b/arch/x86/boot/.gitignore
@@ -3,3 +3,5 @@ bzImage
3setup 3setup
4setup.bin 4setup.bin
5setup.elf 5setup.elf
6cpustr.h
7mkcpustr
diff --git a/arch/x86/boot/printf.c b/arch/x86/boot/printf.c
index 1a09f9309d3c..7e7e890699be 100644
--- a/arch/x86/boot/printf.c
+++ b/arch/x86/boot/printf.c
@@ -33,8 +33,8 @@ static int skip_atoi(const char **s)
33#define PLUS 4 /* show plus */ 33#define PLUS 4 /* show plus */
34#define SPACE 8 /* space if plus */ 34#define SPACE 8 /* space if plus */
35#define LEFT 16 /* left justified */ 35#define LEFT 16 /* left justified */
36#define SPECIAL 32 /* 0x */ 36#define SMALL 32 /* Must be 32 == 0x20 */
37#define LARGE 64 /* use 'ABCDEF' instead of 'abcdef' */ 37#define SPECIAL 64 /* 0x */
38 38
39#define do_div(n,base) ({ \ 39#define do_div(n,base) ({ \
40int __res; \ 40int __res; \
@@ -45,12 +45,16 @@ __res; })
45static char *number(char *str, long num, int base, int size, int precision, 45static char *number(char *str, long num, int base, int size, int precision,
46 int type) 46 int type)
47{ 47{
48 char c, sign, tmp[66]; 48 /* we are called with base 8, 10 or 16, only, thus don't need "G..." */
49 const char *digits = "0123456789abcdefghijklmnopqrstuvwxyz"; 49 static const char digits[16] = "0123456789ABCDEF"; /* "GHIJKLMNOPQRSTUVWXYZ"; */
50
51 char tmp[66];
52 char c, sign, locase;
50 int i; 53 int i;
51 54
52 if (type & LARGE) 55 /* locase = 0 or 0x20. ORing digits or letters with 'locase'
53 digits = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ"; 56 * produces same digits or (maybe lowercased) letters */
57 locase = (type & SMALL);
54 if (type & LEFT) 58 if (type & LEFT)
55 type &= ~ZEROPAD; 59 type &= ~ZEROPAD;
56 if (base < 2 || base > 36) 60 if (base < 2 || base > 36)
@@ -81,7 +85,7 @@ static char *number(char *str, long num, int base, int size, int precision,
81 tmp[i++] = '0'; 85 tmp[i++] = '0';
82 else 86 else
83 while (num != 0) 87 while (num != 0)
84 tmp[i++] = digits[do_div(num, base)]; 88 tmp[i++] = (digits[do_div(num, base)] | locase);
85 if (i > precision) 89 if (i > precision)
86 precision = i; 90 precision = i;
87 size -= precision; 91 size -= precision;
@@ -95,7 +99,7 @@ static char *number(char *str, long num, int base, int size, int precision,
95 *str++ = '0'; 99 *str++ = '0';
96 else if (base == 16) { 100 else if (base == 16) {
97 *str++ = '0'; 101 *str++ = '0';
98 *str++ = digits[33]; 102 *str++ = ('X' | locase);
99 } 103 }
100 } 104 }
101 if (!(type & LEFT)) 105 if (!(type & LEFT))
@@ -244,9 +248,9 @@ int vsprintf(char *buf, const char *fmt, va_list args)
244 base = 8; 248 base = 8;
245 break; 249 break;
246 250
247 case 'X':
248 flags |= LARGE;
249 case 'x': 251 case 'x':
252 flags |= SMALL;
253 case 'X':
250 base = 16; 254 base = 16;
251 break; 255 break;
252 256
diff --git a/arch/x86/configs/i386_defconfig b/arch/x86/configs/i386_defconfig
index 77562e7cdab6..3df340b54e57 100644
--- a/arch/x86/configs/i386_defconfig
+++ b/arch/x86/configs/i386_defconfig
@@ -1421,7 +1421,6 @@ CONFIG_DEBUG_BUGVERBOSE=y
1421# CONFIG_DEBUG_VM is not set 1421# CONFIG_DEBUG_VM is not set
1422# CONFIG_DEBUG_LIST is not set 1422# CONFIG_DEBUG_LIST is not set
1423# CONFIG_FRAME_POINTER is not set 1423# CONFIG_FRAME_POINTER is not set
1424# CONFIG_FORCED_INLINING is not set
1425# CONFIG_RCU_TORTURE_TEST is not set 1424# CONFIG_RCU_TORTURE_TEST is not set
1426# CONFIG_LKDTM is not set 1425# CONFIG_LKDTM is not set
1427# CONFIG_FAULT_INJECTION is not set 1426# CONFIG_FAULT_INJECTION is not set
diff --git a/arch/x86/configs/x86_64_defconfig b/arch/x86/configs/x86_64_defconfig
index 9e2b0ef851de..eef98cb00c62 100644
--- a/arch/x86/configs/x86_64_defconfig
+++ b/arch/x86/configs/x86_64_defconfig
@@ -1346,7 +1346,6 @@ CONFIG_DEBUG_BUGVERBOSE=y
1346# CONFIG_DEBUG_VM is not set 1346# CONFIG_DEBUG_VM is not set
1347# CONFIG_DEBUG_LIST is not set 1347# CONFIG_DEBUG_LIST is not set
1348# CONFIG_FRAME_POINTER is not set 1348# CONFIG_FRAME_POINTER is not set
1349# CONFIG_FORCED_INLINING is not set
1350# CONFIG_RCU_TORTURE_TEST is not set 1349# CONFIG_RCU_TORTURE_TEST is not set
1351# CONFIG_LKDTM is not set 1350# CONFIG_LKDTM is not set
1352# CONFIG_FAULT_INJECTION is not set 1351# CONFIG_FAULT_INJECTION is not set
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index 21dc1a061bf1..76ec0f8f138a 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -84,8 +84,6 @@ ifeq ($(CONFIG_X86_64),y)
84 obj-y += genapic_64.o genapic_flat_64.o 84 obj-y += genapic_64.o genapic_flat_64.o
85 obj-$(CONFIG_X86_PM_TIMER) += pmtimer_64.o 85 obj-$(CONFIG_X86_PM_TIMER) += pmtimer_64.o
86 obj-$(CONFIG_AUDIT) += audit_64.o 86 obj-$(CONFIG_AUDIT) += audit_64.o
87 obj-$(CONFIG_PM) += suspend_64.o
88 obj-$(CONFIG_HIBERNATION) += suspend_asm_64.o
89 87
90 obj-$(CONFIG_GART_IOMMU) += pci-gart_64.o aperture_64.o 88 obj-$(CONFIG_GART_IOMMU) += pci-gart_64.o aperture_64.o
91 obj-$(CONFIG_CALGARY_IOMMU) += pci-calgary_64.o tce_64.o 89 obj-$(CONFIG_CALGARY_IOMMU) += pci-calgary_64.o tce_64.o
diff --git a/arch/x86/kernel/cpu/mcheck/therm_throt.c b/arch/x86/kernel/cpu/mcheck/therm_throt.c
index 24885be5c48c..9b7e01daa1ca 100644
--- a/arch/x86/kernel/cpu/mcheck/therm_throt.c
+++ b/arch/x86/kernel/cpu/mcheck/therm_throt.c
@@ -118,7 +118,7 @@ static __cpuinit int thermal_throttle_add_dev(struct sys_device *sys_dev)
118 118
119static __cpuinit void thermal_throttle_remove_dev(struct sys_device *sys_dev) 119static __cpuinit void thermal_throttle_remove_dev(struct sys_device *sys_dev)
120{ 120{
121 return sysfs_remove_group(&sys_dev->kobj, &thermal_throttle_attr_group); 121 sysfs_remove_group(&sys_dev->kobj, &thermal_throttle_attr_group);
122} 122}
123 123
124/* Mutex protecting device creation against CPU hotplug */ 124/* Mutex protecting device creation against CPU hotplug */
diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S
index be5c31d04884..824e21b80aad 100644
--- a/arch/x86/kernel/entry_32.S
+++ b/arch/x86/kernel/entry_32.S
@@ -409,7 +409,8 @@ restore_nocheck_notrace:
409 RESTORE_REGS 409 RESTORE_REGS
410 addl $4, %esp # skip orig_eax/error_code 410 addl $4, %esp # skip orig_eax/error_code
411 CFI_ADJUST_CFA_OFFSET -4 411 CFI_ADJUST_CFA_OFFSET -4
4121: INTERRUPT_RETURN 412ENTRY(irq_return)
413 INTERRUPT_RETURN
413.section .fixup,"ax" 414.section .fixup,"ax"
414iret_exc: 415iret_exc:
415 pushl $0 # no error code 416 pushl $0 # no error code
@@ -418,7 +419,7 @@ iret_exc:
418.previous 419.previous
419.section __ex_table,"a" 420.section __ex_table,"a"
420 .align 4 421 .align 4
421 .long 1b,iret_exc 422 .long irq_return,iret_exc
422.previous 423.previous
423 424
424 CFI_RESTORE_STATE 425 CFI_RESTORE_STATE
@@ -865,20 +866,16 @@ nmi_espfix_stack:
865 RESTORE_REGS 866 RESTORE_REGS
866 lss 12+4(%esp), %esp # back to espfix stack 867 lss 12+4(%esp), %esp # back to espfix stack
867 CFI_ADJUST_CFA_OFFSET -24 868 CFI_ADJUST_CFA_OFFSET -24
8681: INTERRUPT_RETURN 869 jmp irq_return
869 CFI_ENDPROC 870 CFI_ENDPROC
870.section __ex_table,"a"
871 .align 4
872 .long 1b,iret_exc
873.previous
874KPROBE_END(nmi) 871KPROBE_END(nmi)
875 872
876#ifdef CONFIG_PARAVIRT 873#ifdef CONFIG_PARAVIRT
877ENTRY(native_iret) 874ENTRY(native_iret)
8781: iret 875 iret
879.section __ex_table,"a" 876.section __ex_table,"a"
880 .align 4 877 .align 4
881 .long 1b,iret_exc 878 .long native_iret, iret_exc
882.previous 879.previous
883END(native_iret) 880END(native_iret)
884 881
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index c7341e81941c..6be39a387c5a 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -581,16 +581,24 @@ retint_restore_args: /* return to kernel space */
581 */ 581 */
582 TRACE_IRQS_IRETQ 582 TRACE_IRQS_IRETQ
583restore_args: 583restore_args:
584 RESTORE_ARGS 0,8,0 584 RESTORE_ARGS 0,8,0
585#ifdef CONFIG_PARAVIRT 585
586ENTRY(irq_return)
586 INTERRUPT_RETURN 587 INTERRUPT_RETURN
587#endif 588
589 .section __ex_table, "a"
590 .quad irq_return, bad_iret
591 .previous
592
593#ifdef CONFIG_PARAVIRT
588ENTRY(native_iret) 594ENTRY(native_iret)
589 iretq 595 iretq
590 596
591 .section __ex_table,"a" 597 .section __ex_table,"a"
592 .quad native_iret, bad_iret 598 .quad native_iret, bad_iret
593 .previous 599 .previous
600#endif
601
594 .section .fixup,"ax" 602 .section .fixup,"ax"
595bad_iret: 603bad_iret:
596 /* 604 /*
@@ -804,7 +812,7 @@ paranoid_swapgs\trace:
804 SWAPGS_UNSAFE_STACK 812 SWAPGS_UNSAFE_STACK
805paranoid_restore\trace: 813paranoid_restore\trace:
806 RESTORE_ALL 8 814 RESTORE_ALL 8
807 INTERRUPT_RETURN 815 jmp irq_return
808paranoid_userspace\trace: 816paranoid_userspace\trace:
809 GET_THREAD_INFO(%rcx) 817 GET_THREAD_INFO(%rcx)
810 movl threadinfo_flags(%rcx),%ebx 818 movl threadinfo_flags(%rcx),%ebx
@@ -919,7 +927,7 @@ error_kernelspace:
919 iret run with kernel gs again, so don't set the user space flag. 927 iret run with kernel gs again, so don't set the user space flag.
920 B stepping K8s sometimes report an truncated RIP for IRET 928 B stepping K8s sometimes report an truncated RIP for IRET
921 exceptions returning to compat mode. Check for these here too. */ 929 exceptions returning to compat mode. Check for these here too. */
922 leaq native_iret(%rip),%rbp 930 leaq irq_return(%rip),%rbp
923 cmpq %rbp,RIP(%rsp) 931 cmpq %rbp,RIP(%rsp)
924 je error_swapgs 932 je error_swapgs
925 movl %ebp,%ebp /* zero extend */ 933 movl %ebp,%ebp /* zero extend */
diff --git a/arch/x86/kernel/geode_32.c b/arch/x86/kernel/geode_32.c
index 9c7f7d395968..9dad6ca6cd70 100644
--- a/arch/x86/kernel/geode_32.c
+++ b/arch/x86/kernel/geode_32.c
@@ -163,14 +163,11 @@ EXPORT_SYMBOL_GPL(geode_gpio_setup_event);
163 163
164static int __init geode_southbridge_init(void) 164static int __init geode_southbridge_init(void)
165{ 165{
166 int timers;
167
168 if (!is_geode()) 166 if (!is_geode())
169 return -ENODEV; 167 return -ENODEV;
170 168
171 init_lbars(); 169 init_lbars();
172 timers = geode_mfgpt_detect(); 170 (void) mfgpt_timer_setup();
173 printk(KERN_INFO "geode: %d MFGPT timers available.\n", timers);
174 return 0; 171 return 0;
175} 172}
176 173
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index 5d8c5730686b..74ef4a41f224 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -19,6 +19,10 @@
19#include <asm/thread_info.h> 19#include <asm/thread_info.h>
20#include <asm/asm-offsets.h> 20#include <asm/asm-offsets.h>
21#include <asm/setup.h> 21#include <asm/setup.h>
22#include <asm/processor-flags.h>
23
24/* Physical address */
25#define pa(X) ((X) - __PAGE_OFFSET)
22 26
23/* 27/*
24 * References to members of the new_cpu_data structure. 28 * References to members of the new_cpu_data structure.
@@ -80,10 +84,6 @@ INIT_MAP_BEYOND_END = BOOTBITMAP_SIZE + (PAGE_TABLE_SIZE + ALLOCATOR_SLOP)*PAGE_
80 */ 84 */
81.section .text.head,"ax",@progbits 85.section .text.head,"ax",@progbits
82ENTRY(startup_32) 86ENTRY(startup_32)
83 /* check to see if KEEP_SEGMENTS flag is meaningful */
84 cmpw $0x207, BP_version(%esi)
85 jb 1f
86
87 /* test KEEP_SEGMENTS flag to see if the bootloader is asking 87 /* test KEEP_SEGMENTS flag to see if the bootloader is asking
88 us to not reload segments */ 88 us to not reload segments */
89 testb $(1<<6), BP_loadflags(%esi) 89 testb $(1<<6), BP_loadflags(%esi)
@@ -92,7 +92,7 @@ ENTRY(startup_32)
92/* 92/*
93 * Set segments to known values. 93 * Set segments to known values.
94 */ 94 */
951: lgdt boot_gdt_descr - __PAGE_OFFSET 95 lgdt pa(boot_gdt_descr)
96 movl $(__BOOT_DS),%eax 96 movl $(__BOOT_DS),%eax
97 movl %eax,%ds 97 movl %eax,%ds
98 movl %eax,%es 98 movl %eax,%es
@@ -105,8 +105,8 @@ ENTRY(startup_32)
105 */ 105 */
106 cld 106 cld
107 xorl %eax,%eax 107 xorl %eax,%eax
108 movl $__bss_start - __PAGE_OFFSET,%edi 108 movl $pa(__bss_start),%edi
109 movl $__bss_stop - __PAGE_OFFSET,%ecx 109 movl $pa(__bss_stop),%ecx
110 subl %edi,%ecx 110 subl %edi,%ecx
111 shrl $2,%ecx 111 shrl $2,%ecx
112 rep ; stosl 112 rep ; stosl
@@ -118,31 +118,32 @@ ENTRY(startup_32)
118 * (kexec on panic case). Hence copy out the parameters before initializing 118 * (kexec on panic case). Hence copy out the parameters before initializing
119 * page tables. 119 * page tables.
120 */ 120 */
121 movl $(boot_params - __PAGE_OFFSET),%edi 121 movl $pa(boot_params),%edi
122 movl $(PARAM_SIZE/4),%ecx 122 movl $(PARAM_SIZE/4),%ecx
123 cld 123 cld
124 rep 124 rep
125 movsl 125 movsl
126 movl boot_params - __PAGE_OFFSET + NEW_CL_POINTER,%esi 126 movl pa(boot_params) + NEW_CL_POINTER,%esi
127 andl %esi,%esi 127 andl %esi,%esi
128 jz 1f # No comand line 128 jz 1f # No comand line
129 movl $(boot_command_line - __PAGE_OFFSET),%edi 129 movl $pa(boot_command_line),%edi
130 movl $(COMMAND_LINE_SIZE/4),%ecx 130 movl $(COMMAND_LINE_SIZE/4),%ecx
131 rep 131 rep
132 movsl 132 movsl
1331: 1331:
134 134
135#ifdef CONFIG_PARAVIRT 135#ifdef CONFIG_PARAVIRT
136 cmpw $0x207, (boot_params + BP_version - __PAGE_OFFSET) 136 /* This is can only trip for a broken bootloader... */
137 cmpw $0x207, pa(boot_params + BP_version)
137 jb default_entry 138 jb default_entry
138 139
139 /* Paravirt-compatible boot parameters. Look to see what architecture 140 /* Paravirt-compatible boot parameters. Look to see what architecture
140 we're booting under. */ 141 we're booting under. */
141 movl (boot_params + BP_hardware_subarch - __PAGE_OFFSET), %eax 142 movl pa(boot_params + BP_hardware_subarch), %eax
142 cmpl $num_subarch_entries, %eax 143 cmpl $num_subarch_entries, %eax
143 jae bad_subarch 144 jae bad_subarch
144 145
145 movl subarch_entries - __PAGE_OFFSET(,%eax,4), %eax 146 movl pa(subarch_entries)(,%eax,4), %eax
146 subl $__PAGE_OFFSET, %eax 147 subl $__PAGE_OFFSET, %eax
147 jmp *%eax 148 jmp *%eax
148 149
@@ -170,17 +171,68 @@ num_subarch_entries = (. - subarch_entries) / 4
170 * Mappings are created both at virtual address 0 (identity mapping) 171 * Mappings are created both at virtual address 0 (identity mapping)
171 * and PAGE_OFFSET for up to _end+sizeof(page tables)+INIT_MAP_BEYOND_END. 172 * and PAGE_OFFSET for up to _end+sizeof(page tables)+INIT_MAP_BEYOND_END.
172 * 173 *
173 * Warning: don't use %esi or the stack in this code. However, %esp 174 * Note that the stack is not yet set up!
174 * can be used as a GPR if you really need it...
175 */ 175 */
176page_pde_offset = (__PAGE_OFFSET >> 20); 176#define PTE_ATTR 0x007 /* PRESENT+RW+USER */
177#define PDE_ATTR 0x067 /* PRESENT+RW+USER+DIRTY+ACCESSED */
178#define PGD_ATTR 0x001 /* PRESENT (no other attributes) */
177 179
178default_entry: 180default_entry:
179 movl $(pg0 - __PAGE_OFFSET), %edi 181#ifdef CONFIG_X86_PAE
180 movl $(swapper_pg_dir - __PAGE_OFFSET), %edx 182
181 movl $0x007, %eax /* 0x007 = PRESENT+RW+USER */ 183 /*
184 * In PAE mode swapper_pg_dir is statically defined to contain enough
185 * entries to cover the VMSPLIT option (that is the top 1, 2 or 3
186 * entries). The identity mapping is handled by pointing two PGD
187 * entries to the first kernel PMD.
188 *
189 * Note the upper half of each PMD or PTE are always zero at
190 * this stage.
191 */
192
193#define KPMDS ((0x100000000-__PAGE_OFFSET) >> 30) /* Number of kernel PMDs */
194
195 xorl %ebx,%ebx /* %ebx is kept at zero */
196
197 movl $pa(pg0), %edi
198 movl $pa(swapper_pg_pmd), %edx
199 movl $PTE_ATTR, %eax
20010:
201 leal PDE_ATTR(%edi),%ecx /* Create PMD entry */
202 movl %ecx,(%edx) /* Store PMD entry */
203 /* Upper half already zero */
204 addl $8,%edx
205 movl $512,%ecx
20611:
207 stosl
208 xchgl %eax,%ebx
209 stosl
210 xchgl %eax,%ebx
211 addl $0x1000,%eax
212 loop 11b
213
214 /*
215 * End condition: we must map up to and including INIT_MAP_BEYOND_END
216 * bytes beyond the end of our own page tables.
217 */
218 leal (INIT_MAP_BEYOND_END+PTE_ATTR)(%edi),%ebp
219 cmpl %ebp,%eax
220 jb 10b
2211:
222 movl %edi,pa(init_pg_tables_end)
223
224 /* Do early initialization of the fixmap area */
225 movl $pa(swapper_pg_fixmap)+PDE_ATTR,%eax
226 movl %eax,pa(swapper_pg_pmd+0x1000*KPMDS-8)
227#else /* Not PAE */
228
229page_pde_offset = (__PAGE_OFFSET >> 20);
230
231 movl $pa(pg0), %edi
232 movl $pa(swapper_pg_dir), %edx
233 movl $PTE_ATTR, %eax
18210: 23410:
183 leal 0x007(%edi),%ecx /* Create PDE entry */ 235 leal PDE_ATTR(%edi),%ecx /* Create PDE entry */
184 movl %ecx,(%edx) /* Store identity PDE entry */ 236 movl %ecx,(%edx) /* Store identity PDE entry */
185 movl %ecx,page_pde_offset(%edx) /* Store kernel PDE entry */ 237 movl %ecx,page_pde_offset(%edx) /* Store kernel PDE entry */
186 addl $4,%edx 238 addl $4,%edx
@@ -189,19 +241,20 @@ default_entry:
189 stosl 241 stosl
190 addl $0x1000,%eax 242 addl $0x1000,%eax
191 loop 11b 243 loop 11b
192 /* End condition: we must map up to and including INIT_MAP_BEYOND_END */ 244 /*
193 /* bytes beyond the end of our own page tables; the +0x007 is the attribute bits */ 245 * End condition: we must map up to and including INIT_MAP_BEYOND_END
194 leal (INIT_MAP_BEYOND_END+0x007)(%edi),%ebp 246 * bytes beyond the end of our own page tables; the +0x007 is
247 * the attribute bits
248 */
249 leal (INIT_MAP_BEYOND_END+PTE_ATTR)(%edi),%ebp
195 cmpl %ebp,%eax 250 cmpl %ebp,%eax
196 jb 10b 251 jb 10b
197 movl %edi,(init_pg_tables_end - __PAGE_OFFSET) 252 movl %edi,pa(init_pg_tables_end)
198
199 /* Do an early initialization of the fixmap area */
200 movl $(swapper_pg_dir - __PAGE_OFFSET), %edx
201 movl $(swapper_pg_pmd - __PAGE_OFFSET), %eax
202 addl $0x67, %eax /* 0x67 == _PAGE_TABLE */
203 movl %eax, 4092(%edx)
204 253
254 /* Do early initialization of the fixmap area */
255 movl $pa(swapper_pg_fixmap)+PDE_ATTR,%eax
256 movl %eax,pa(swapper_pg_dir+0xffc)
257#endif
205 jmp 3f 258 jmp 3f
206/* 259/*
207 * Non-boot CPU entry point; entered from trampoline.S 260 * Non-boot CPU entry point; entered from trampoline.S
@@ -241,7 +294,7 @@ ENTRY(startup_32_smp)
241 * NOTE! We have to correct for the fact that we're 294 * NOTE! We have to correct for the fact that we're
242 * not yet offset PAGE_OFFSET.. 295 * not yet offset PAGE_OFFSET..
243 */ 296 */
244#define cr4_bits mmu_cr4_features-__PAGE_OFFSET 297#define cr4_bits pa(mmu_cr4_features)
245 movl cr4_bits,%edx 298 movl cr4_bits,%edx
246 andl %edx,%edx 299 andl %edx,%edx
247 jz 6f 300 jz 6f
@@ -276,10 +329,10 @@ ENTRY(startup_32_smp)
276/* 329/*
277 * Enable paging 330 * Enable paging
278 */ 331 */
279 movl $swapper_pg_dir-__PAGE_OFFSET,%eax 332 movl $pa(swapper_pg_dir),%eax
280 movl %eax,%cr3 /* set the page table pointer.. */ 333 movl %eax,%cr3 /* set the page table pointer.. */
281 movl %cr0,%eax 334 movl %cr0,%eax
282 orl $0x80000000,%eax 335 orl $X86_CR0_PG,%eax
283 movl %eax,%cr0 /* ..and set paging (PG) bit */ 336 movl %eax,%cr0 /* ..and set paging (PG) bit */
284 ljmp $__BOOT_CS,$1f /* Clear prefetch and normalize %eip */ 337 ljmp $__BOOT_CS,$1f /* Clear prefetch and normalize %eip */
2851: 3381:
@@ -552,16 +605,44 @@ ENTRY(_stext)
552 */ 605 */
553.section ".bss.page_aligned","wa" 606.section ".bss.page_aligned","wa"
554 .align PAGE_SIZE_asm 607 .align PAGE_SIZE_asm
608#ifdef CONFIG_X86_PAE
609ENTRY(swapper_pg_pmd)
610 .fill 1024*KPMDS,4,0
611#else
555ENTRY(swapper_pg_dir) 612ENTRY(swapper_pg_dir)
556 .fill 1024,4,0 613 .fill 1024,4,0
557ENTRY(swapper_pg_pmd) 614#endif
615ENTRY(swapper_pg_fixmap)
558 .fill 1024,4,0 616 .fill 1024,4,0
559ENTRY(empty_zero_page) 617ENTRY(empty_zero_page)
560 .fill 4096,1,0 618 .fill 4096,1,0
561
562/* 619/*
563 * This starts the data section. 620 * This starts the data section.
564 */ 621 */
622#ifdef CONFIG_X86_PAE
623.section ".data.page_aligned","wa"
624 /* Page-aligned for the benefit of paravirt? */
625 .align PAGE_SIZE_asm
626ENTRY(swapper_pg_dir)
627 .long pa(swapper_pg_pmd+PGD_ATTR),0 /* low identity map */
628# if KPMDS == 3
629 .long pa(swapper_pg_pmd+PGD_ATTR),0
630 .long pa(swapper_pg_pmd+PGD_ATTR+0x1000),0
631 .long pa(swapper_pg_pmd+PGD_ATTR+0x2000),0
632# elif KPMDS == 2
633 .long 0,0
634 .long pa(swapper_pg_pmd+PGD_ATTR),0
635 .long pa(swapper_pg_pmd+PGD_ATTR+0x1000),0
636# elif KPMDS == 1
637 .long 0,0
638 .long 0,0
639 .long pa(swapper_pg_pmd+PGD_ATTR),0
640# else
641# error "Kernel PMDs should be 1, 2 or 3"
642# endif
643 .align PAGE_SIZE_asm /* needs to be page-sized too */
644#endif
645
565.data 646.data
566ENTRY(stack_start) 647ENTRY(stack_start)
567 .long init_thread_union+THREAD_SIZE 648 .long init_thread_union+THREAD_SIZE
diff --git a/arch/x86/kernel/mfgpt_32.c b/arch/x86/kernel/mfgpt_32.c
index 219f86eb6123..027fc067b399 100644
--- a/arch/x86/kernel/mfgpt_32.c
+++ b/arch/x86/kernel/mfgpt_32.c
@@ -12,48 +12,37 @@
12 */ 12 */
13 13
14/* 14/*
15 * We are using the 32Khz input clock - its the only one that has the 15 * We are using the 32.768kHz input clock - it's the only one that has the
16 * ranges we find desirable. The following table lists the suitable 16 * ranges we find desirable. The following table lists the suitable
17 * divisors and the associated hz, minimum interval 17 * divisors and the associated Hz, minimum interval and the maximum interval:
18 * and the maximum interval:
19 * 18 *
20 * Divisor Hz Min Delta (S) Max Delta (S) 19 * Divisor Hz Min Delta (s) Max Delta (s)
21 * 1 32000 .0005 2.048 20 * 1 32768 .00048828125 2.000
22 * 2 16000 .001 4.096 21 * 2 16384 .0009765625 4.000
23 * 4 8000 .002 8.192 22 * 4 8192 .001953125 8.000
24 * 8 4000 .004 16.384 23 * 8 4096 .00390625 16.000
25 * 16 2000 .008 32.768 24 * 16 2048 .0078125 32.000
26 * 32 1000 .016 65.536 25 * 32 1024 .015625 64.000
27 * 64 500 .032 131.072 26 * 64 512 .03125 128.000
28 * 128 250 .064 262.144 27 * 128 256 .0625 256.000
29 * 256 125 .128 524.288 28 * 256 128 .125 512.000
30 */ 29 */
31 30
32#include <linux/kernel.h> 31#include <linux/kernel.h>
33#include <linux/interrupt.h> 32#include <linux/interrupt.h>
34#include <linux/module.h>
35#include <asm/geode.h> 33#include <asm/geode.h>
36 34
37#define F_AVAIL 0x01
38
39static struct mfgpt_timer_t { 35static struct mfgpt_timer_t {
40 int flags; 36 unsigned int avail:1;
41 struct module *owner;
42} mfgpt_timers[MFGPT_MAX_TIMERS]; 37} mfgpt_timers[MFGPT_MAX_TIMERS];
43 38
44/* Selected from the table above */ 39/* Selected from the table above */
45 40
46#define MFGPT_DIVISOR 16 41#define MFGPT_DIVISOR 16
47#define MFGPT_SCALE 4 /* divisor = 2^(scale) */ 42#define MFGPT_SCALE 4 /* divisor = 2^(scale) */
48#define MFGPT_HZ (32000 / MFGPT_DIVISOR) 43#define MFGPT_HZ (32768 / MFGPT_DIVISOR)
49#define MFGPT_PERIODIC (MFGPT_HZ / HZ) 44#define MFGPT_PERIODIC (MFGPT_HZ / HZ)
50 45
51#ifdef CONFIG_GEODE_MFGPT_TIMER
52static int __init mfgpt_timer_setup(void);
53#else
54#define mfgpt_timer_setup() (0)
55#endif
56
57/* Allow for disabling of MFGPTs */ 46/* Allow for disabling of MFGPTs */
58static int disable; 47static int disable;
59static int __init mfgpt_disable(char *s) 48static int __init mfgpt_disable(char *s)
@@ -85,28 +74,37 @@ __setup("mfgptfix", mfgpt_fix);
85 * In other cases (such as with VSAless OpenFirmware), the system firmware 74 * In other cases (such as with VSAless OpenFirmware), the system firmware
86 * leaves timers available for us to use. 75 * leaves timers available for us to use.
87 */ 76 */
88int __init geode_mfgpt_detect(void) 77
78
79static int timers = -1;
80
81static void geode_mfgpt_detect(void)
89{ 82{
90 int count = 0, i; 83 int i;
91 u16 val; 84 u16 val;
92 85
86 timers = 0;
87
93 if (disable) { 88 if (disable) {
94 printk(KERN_INFO "geode-mfgpt: Skipping MFGPT setup\n"); 89 printk(KERN_INFO "geode-mfgpt: MFGPT support is disabled\n");
95 return 0; 90 goto done;
91 }
92
93 if (!geode_get_dev_base(GEODE_DEV_MFGPT)) {
94 printk(KERN_INFO "geode-mfgpt: MFGPT LBAR is not set up\n");
95 goto done;
96 } 96 }
97 97
98 for (i = 0; i < MFGPT_MAX_TIMERS; i++) { 98 for (i = 0; i < MFGPT_MAX_TIMERS; i++) {
99 val = geode_mfgpt_read(i, MFGPT_REG_SETUP); 99 val = geode_mfgpt_read(i, MFGPT_REG_SETUP);
100 if (!(val & MFGPT_SETUP_SETUP)) { 100 if (!(val & MFGPT_SETUP_SETUP)) {
101 mfgpt_timers[i].flags = F_AVAIL; 101 mfgpt_timers[i].avail = 1;
102 count++; 102 timers++;
103 } 103 }
104 } 104 }
105 105
106 /* set up clock event device, if desired */ 106done:
107 i = mfgpt_timer_setup(); 107 printk(KERN_INFO "geode-mfgpt: %d MFGPT timers available.\n", timers);
108
109 return count;
110} 108}
111 109
112int geode_mfgpt_toggle_event(int timer, int cmp, int event, int enable) 110int geode_mfgpt_toggle_event(int timer, int cmp, int event, int enable)
@@ -183,36 +181,41 @@ int geode_mfgpt_set_irq(int timer, int cmp, int irq, int enable)
183 return 0; 181 return 0;
184} 182}
185 183
186static int mfgpt_get(int timer, struct module *owner) 184static int mfgpt_get(int timer)
187{ 185{
188 mfgpt_timers[timer].flags &= ~F_AVAIL; 186 mfgpt_timers[timer].avail = 0;
189 mfgpt_timers[timer].owner = owner;
190 printk(KERN_INFO "geode-mfgpt: Registered timer %d\n", timer); 187 printk(KERN_INFO "geode-mfgpt: Registered timer %d\n", timer);
191 return timer; 188 return timer;
192} 189}
193 190
194int geode_mfgpt_alloc_timer(int timer, int domain, struct module *owner) 191int geode_mfgpt_alloc_timer(int timer, int domain)
195{ 192{
196 int i; 193 int i;
197 194
198 if (!geode_get_dev_base(GEODE_DEV_MFGPT)) 195 if (timers == -1) {
199 return -ENODEV; 196 /* timers haven't been detected yet */
197 geode_mfgpt_detect();
198 }
199
200 if (!timers)
201 return -1;
202
200 if (timer >= MFGPT_MAX_TIMERS) 203 if (timer >= MFGPT_MAX_TIMERS)
201 return -EIO; 204 return -1;
202 205
203 if (timer < 0) { 206 if (timer < 0) {
204 /* Try to find an available timer */ 207 /* Try to find an available timer */
205 for (i = 0; i < MFGPT_MAX_TIMERS; i++) { 208 for (i = 0; i < MFGPT_MAX_TIMERS; i++) {
206 if (mfgpt_timers[i].flags & F_AVAIL) 209 if (mfgpt_timers[i].avail)
207 return mfgpt_get(i, owner); 210 return mfgpt_get(i);
208 211
209 if (i == 5 && domain == MFGPT_DOMAIN_WORKING) 212 if (i == 5 && domain == MFGPT_DOMAIN_WORKING)
210 break; 213 break;
211 } 214 }
212 } else { 215 } else {
213 /* If they requested a specific timer, try to honor that */ 216 /* If they requested a specific timer, try to honor that */
214 if (mfgpt_timers[timer].flags & F_AVAIL) 217 if (mfgpt_timers[timer].avail)
215 return mfgpt_get(timer, owner); 218 return mfgpt_get(timer);
216 } 219 }
217 220
218 /* No timers available - too bad */ 221 /* No timers available - too bad */
@@ -244,10 +247,11 @@ static int __init mfgpt_setup(char *str)
244} 247}
245__setup("mfgpt_irq=", mfgpt_setup); 248__setup("mfgpt_irq=", mfgpt_setup);
246 249
247static inline void mfgpt_disable_timer(u16 clock) 250static void mfgpt_disable_timer(u16 clock)
248{ 251{
249 u16 val = geode_mfgpt_read(clock, MFGPT_REG_SETUP); 252 /* avoid races by clearing CMP1 and CMP2 unconditionally */
250 geode_mfgpt_write(clock, MFGPT_REG_SETUP, val & ~MFGPT_SETUP_CNTEN); 253 geode_mfgpt_write(clock, MFGPT_REG_SETUP, (u16) ~MFGPT_SETUP_CNTEN |
254 MFGPT_SETUP_CMP1 | MFGPT_SETUP_CMP2);
251} 255}
252 256
253static int mfgpt_next_event(unsigned long, struct clock_event_device *); 257static int mfgpt_next_event(unsigned long, struct clock_event_device *);
@@ -263,7 +267,7 @@ static struct clock_event_device mfgpt_clockevent = {
263 .shift = 32 267 .shift = 32
264}; 268};
265 269
266static inline void mfgpt_start_timer(u16 clock, u16 delta) 270static void mfgpt_start_timer(u16 delta)
267{ 271{
268 geode_mfgpt_write(mfgpt_event_clock, MFGPT_REG_CMP2, (u16) delta); 272 geode_mfgpt_write(mfgpt_event_clock, MFGPT_REG_CMP2, (u16) delta);
269 geode_mfgpt_write(mfgpt_event_clock, MFGPT_REG_COUNTER, 0); 273 geode_mfgpt_write(mfgpt_event_clock, MFGPT_REG_COUNTER, 0);
@@ -278,21 +282,25 @@ static void mfgpt_set_mode(enum clock_event_mode mode,
278 mfgpt_disable_timer(mfgpt_event_clock); 282 mfgpt_disable_timer(mfgpt_event_clock);
279 283
280 if (mode == CLOCK_EVT_MODE_PERIODIC) 284 if (mode == CLOCK_EVT_MODE_PERIODIC)
281 mfgpt_start_timer(mfgpt_event_clock, MFGPT_PERIODIC); 285 mfgpt_start_timer(MFGPT_PERIODIC);
282 286
283 mfgpt_tick_mode = mode; 287 mfgpt_tick_mode = mode;
284} 288}
285 289
286static int mfgpt_next_event(unsigned long delta, struct clock_event_device *evt) 290static int mfgpt_next_event(unsigned long delta, struct clock_event_device *evt)
287{ 291{
288 mfgpt_start_timer(mfgpt_event_clock, delta); 292 mfgpt_start_timer(delta);
289 return 0; 293 return 0;
290} 294}
291 295
292/* Assume (foolishly?), that this interrupt was due to our tick */
293
294static irqreturn_t mfgpt_tick(int irq, void *dev_id) 296static irqreturn_t mfgpt_tick(int irq, void *dev_id)
295{ 297{
298 u16 val = geode_mfgpt_read(mfgpt_event_clock, MFGPT_REG_SETUP);
299
300 /* See if the interrupt was for us */
301 if (!(val & (MFGPT_SETUP_SETUP | MFGPT_SETUP_CMP2 | MFGPT_SETUP_CMP1)))
302 return IRQ_NONE;
303
296 /* Turn off the clock (and clear the event) */ 304 /* Turn off the clock (and clear the event) */
297 mfgpt_disable_timer(mfgpt_event_clock); 305 mfgpt_disable_timer(mfgpt_event_clock);
298 306
@@ -320,13 +328,12 @@ static struct irqaction mfgptirq = {
320 .name = "mfgpt-timer" 328 .name = "mfgpt-timer"
321}; 329};
322 330
323static int __init mfgpt_timer_setup(void) 331int __init mfgpt_timer_setup(void)
324{ 332{
325 int timer, ret; 333 int timer, ret;
326 u16 val; 334 u16 val;
327 335
328 timer = geode_mfgpt_alloc_timer(MFGPT_TIMER_ANY, MFGPT_DOMAIN_WORKING, 336 timer = geode_mfgpt_alloc_timer(MFGPT_TIMER_ANY, MFGPT_DOMAIN_WORKING);
329 THIS_MODULE);
330 if (timer < 0) { 337 if (timer < 0) {
331 printk(KERN_ERR 338 printk(KERN_ERR
332 "mfgpt-timer: Could not allocate a MFPGT timer\n"); 339 "mfgpt-timer: Could not allocate a MFPGT timer\n");
@@ -363,7 +370,7 @@ static int __init mfgpt_timer_setup(void)
363 &mfgpt_clockevent); 370 &mfgpt_clockevent);
364 371
365 printk(KERN_INFO 372 printk(KERN_INFO
366 "mfgpt-timer: registering the MFGT timer as a clock event.\n"); 373 "mfgpt-timer: registering the MFGPT timer as a clock event.\n");
367 clockevents_register_device(&mfgpt_clockevent); 374 clockevents_register_device(&mfgpt_clockevent);
368 375
369 return 0; 376 return 0;
diff --git a/arch/x86/kernel/quirks.c b/arch/x86/kernel/quirks.c
index 6ba33ca8715a..1941482d4ca3 100644
--- a/arch/x86/kernel/quirks.c
+++ b/arch/x86/kernel/quirks.c
@@ -27,7 +27,7 @@ static void __devinit quirk_intel_irqbalance(struct pci_dev *dev)
27 pci_write_config_byte(dev, 0xf4, config|0x2); 27 pci_write_config_byte(dev, 0xf4, config|0x2);
28 28
29 /* read xTPR register */ 29 /* read xTPR register */
30 raw_pci_ops->read(0, 0, 0x40, 0x4c, 2, &word); 30 raw_pci_read(0, 0, 0x40, 0x4c, 2, &word);
31 31
32 if (!(word & (1 << 13))) { 32 if (!(word & (1 << 13))) {
33 dev_info(&dev->dev, "Intel E7520/7320/7525 detected; " 33 dev_info(&dev->dev, "Intel E7520/7320/7525 detected; "
diff --git a/arch/x86/kernel/setup_32.c b/arch/x86/kernel/setup_32.c
index d1d8c347cc0b..691ab4cb167b 100644
--- a/arch/x86/kernel/setup_32.c
+++ b/arch/x86/kernel/setup_32.c
@@ -154,7 +154,11 @@ struct cpuinfo_x86 new_cpu_data __cpuinitdata = { 0, 0, 0, 0, -1, 1, 0, 0, -1 };
154struct cpuinfo_x86 boot_cpu_data __read_mostly = { 0, 0, 0, 0, -1, 1, 0, 0, -1 }; 154struct cpuinfo_x86 boot_cpu_data __read_mostly = { 0, 0, 0, 0, -1, 1, 0, 0, -1 };
155EXPORT_SYMBOL(boot_cpu_data); 155EXPORT_SYMBOL(boot_cpu_data);
156 156
157#ifndef CONFIG_X86_PAE
157unsigned long mmu_cr4_features; 158unsigned long mmu_cr4_features;
159#else
160unsigned long mmu_cr4_features = X86_CR4_PAE;
161#endif
158 162
159/* for MCA, but anyone else can use it if they want */ 163/* for MCA, but anyone else can use it if they want */
160unsigned int machine_id; 164unsigned int machine_id;
diff --git a/arch/x86/kernel/topology.c b/arch/x86/kernel/topology.c
index e6757aaa202b..a40051b71d9b 100644
--- a/arch/x86/kernel/topology.c
+++ b/arch/x86/kernel/topology.c
@@ -53,7 +53,7 @@ EXPORT_SYMBOL(arch_register_cpu);
53 53
54void arch_unregister_cpu(int num) 54void arch_unregister_cpu(int num)
55{ 55{
56 return unregister_cpu(&per_cpu(cpu_devices, num).cpu); 56 unregister_cpu(&per_cpu(cpu_devices, num).cpu);
57} 57}
58EXPORT_SYMBOL(arch_unregister_cpu); 58EXPORT_SYMBOL(arch_unregister_cpu);
59#else 59#else
diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index d1bc04006d16..8106bba41ecb 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -46,6 +46,7 @@
46#include <asm/pgalloc.h> 46#include <asm/pgalloc.h>
47#include <asm/sections.h> 47#include <asm/sections.h>
48#include <asm/paravirt.h> 48#include <asm/paravirt.h>
49#include <asm/setup.h>
49 50
50unsigned int __VMALLOC_RESERVE = 128 << 20; 51unsigned int __VMALLOC_RESERVE = 128 << 20;
51 52
@@ -328,44 +329,38 @@ pteval_t __PAGE_KERNEL_EXEC = _PAGE_KERNEL_EXEC;
328 329
329void __init native_pagetable_setup_start(pgd_t *base) 330void __init native_pagetable_setup_start(pgd_t *base)
330{ 331{
331#ifdef CONFIG_X86_PAE 332 unsigned long pfn, va;
332 int i; 333 pgd_t *pgd;
334 pud_t *pud;
335 pmd_t *pmd;
336 pte_t *pte;
333 337
334 /* 338 /*
335 * Init entries of the first-level page table to the 339 * Remove any mappings which extend past the end of physical
336 * zero page, if they haven't already been set up. 340 * memory from the boot time page table:
337 *
338 * In a normal native boot, we'll be running on a
339 * pagetable rooted in swapper_pg_dir, but not in PAE
340 * mode, so this will end up clobbering the mappings
341 * for the lower 24Mbytes of the address space,
342 * without affecting the kernel address space.
343 */ 341 */
344 for (i = 0; i < USER_PTRS_PER_PGD; i++) 342 for (pfn = max_low_pfn + 1; pfn < 1<<(32-PAGE_SHIFT); pfn++) {
345 set_pgd(&base[i], 343 va = PAGE_OFFSET + (pfn<<PAGE_SHIFT);
346 __pgd(__pa(empty_zero_page) | _PAGE_PRESENT)); 344 pgd = base + pgd_index(va);
347 345 if (!pgd_present(*pgd))
348 /* Make sure kernel address space is empty so that a pagetable 346 break;
349 will be allocated for it. */ 347
350 memset(&base[USER_PTRS_PER_PGD], 0, 348 pud = pud_offset(pgd, va);
351 KERNEL_PGD_PTRS * sizeof(pgd_t)); 349 pmd = pmd_offset(pud, va);
352#else 350 if (!pmd_present(*pmd))
351 break;
352
353 pte = pte_offset_kernel(pmd, va);
354 if (!pte_present(*pte))
355 break;
356
357 pte_clear(NULL, va, pte);
358 }
353 paravirt_alloc_pd(&init_mm, __pa(base) >> PAGE_SHIFT); 359 paravirt_alloc_pd(&init_mm, __pa(base) >> PAGE_SHIFT);
354#endif
355} 360}
356 361
357void __init native_pagetable_setup_done(pgd_t *base) 362void __init native_pagetable_setup_done(pgd_t *base)
358{ 363{
359#ifdef CONFIG_X86_PAE
360 /*
361 * Add low memory identity-mappings - SMP needs it when
362 * starting up on an AP from real-mode. In the non-PAE
363 * case we already have these mappings through head.S.
364 * All user-space mappings are explicitly cleared after
365 * SMP startup.
366 */
367 set_pgd(&base[0], base[USER_PTRS_PER_PGD]);
368#endif
369} 364}
370 365
371/* 366/*
@@ -374,9 +369,8 @@ void __init native_pagetable_setup_done(pgd_t *base)
374 * the boot process. 369 * the boot process.
375 * 370 *
376 * If we're booting on native hardware, this will be a pagetable 371 * If we're booting on native hardware, this will be a pagetable
377 * constructed in arch/i386/kernel/head.S, and not running in PAE mode 372 * constructed in arch/x86/kernel/head_32.S. The root of the
378 * (even if we'll end up running in PAE). The root of the pagetable 373 * pagetable will be swapper_pg_dir.
379 * will be swapper_pg_dir.
380 * 374 *
381 * If we're booting paravirtualized under a hypervisor, then there are 375 * If we're booting paravirtualized under a hypervisor, then there are
382 * more options: we may already be running PAE, and the pagetable may 376 * more options: we may already be running PAE, and the pagetable may
@@ -537,14 +531,6 @@ void __init paging_init(void)
537 531
538 load_cr3(swapper_pg_dir); 532 load_cr3(swapper_pg_dir);
539 533
540#ifdef CONFIG_X86_PAE
541 /*
542 * We will bail out later - printk doesn't work right now so
543 * the user would just see a hanging kernel.
544 */
545 if (cpu_has_pae)
546 set_in_cr4(X86_CR4_PAE);
547#endif
548 __flush_tlb_all(); 534 __flush_tlb_all();
549 535
550 kmap_init(); 536 kmap_init();
@@ -675,13 +661,11 @@ void __init mem_init(void)
675 BUG_ON((unsigned long)high_memory > VMALLOC_START); 661 BUG_ON((unsigned long)high_memory > VMALLOC_START);
676#endif /* double-sanity-check paranoia */ 662#endif /* double-sanity-check paranoia */
677 663
678#ifdef CONFIG_X86_PAE
679 if (!cpu_has_pae)
680 panic("cannot execute a PAE-enabled kernel on a PAE-less CPU!");
681#endif
682 if (boot_cpu_data.wp_works_ok < 0) 664 if (boot_cpu_data.wp_works_ok < 0)
683 test_wp_bit(); 665 test_wp_bit();
684 666
667 cpa_init();
668
685 /* 669 /*
686 * Subtle. SMP is doing it's boot stuff late (because it has to 670 * Subtle. SMP is doing it's boot stuff late (because it has to
687 * fork idle threads) - but it also needs low mappings for the 671 * fork idle threads) - but it also needs low mappings for the
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index 5fe880fc305d..b59fc238151f 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -528,13 +528,15 @@ void __init mem_init(void)
528 reservedpages << (PAGE_SHIFT-10), 528 reservedpages << (PAGE_SHIFT-10),
529 datasize >> 10, 529 datasize >> 10,
530 initsize >> 10); 530 initsize >> 10);
531
532 cpa_init();
531} 533}
532 534
533void free_init_pages(char *what, unsigned long begin, unsigned long end) 535void free_init_pages(char *what, unsigned long begin, unsigned long end)
534{ 536{
535 unsigned long addr; 537 unsigned long addr = begin;
536 538
537 if (begin >= end) 539 if (addr >= end)
538 return; 540 return;
539 541
540 /* 542 /*
@@ -549,7 +551,7 @@ void free_init_pages(char *what, unsigned long begin, unsigned long end)
549#else 551#else
550 printk(KERN_INFO "Freeing %s: %luk freed\n", what, (end - begin) >> 10); 552 printk(KERN_INFO "Freeing %s: %luk freed\n", what, (end - begin) >> 10);
551 553
552 for (addr = begin; addr < end; addr += PAGE_SIZE) { 554 for (; addr < end; addr += PAGE_SIZE) {
553 ClearPageReserved(virt_to_page(addr)); 555 ClearPageReserved(virt_to_page(addr));
554 init_page_count(virt_to_page(addr)); 556 init_page_count(virt_to_page(addr));
555 memset((void *)(addr & ~(PAGE_SIZE-1)), 557 memset((void *)(addr & ~(PAGE_SIZE-1)),
diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
index ee6648fe6b15..a4897a85268a 100644
--- a/arch/x86/mm/ioremap.c
+++ b/arch/x86/mm/ioremap.c
@@ -260,41 +260,46 @@ static int __init early_ioremap_debug_setup(char *str)
260early_param("early_ioremap_debug", early_ioremap_debug_setup); 260early_param("early_ioremap_debug", early_ioremap_debug_setup);
261 261
262static __initdata int after_paging_init; 262static __initdata int after_paging_init;
263static __initdata unsigned long bm_pte[1024] 263static __initdata pte_t bm_pte[PAGE_SIZE/sizeof(pte_t)]
264 __attribute__((aligned(PAGE_SIZE))); 264 __attribute__((aligned(PAGE_SIZE)));
265 265
266static inline unsigned long * __init early_ioremap_pgd(unsigned long addr) 266static inline pmd_t * __init early_ioremap_pmd(unsigned long addr)
267{ 267{
268 return (unsigned long *)swapper_pg_dir + ((addr >> 22) & 1023); 268 pgd_t *pgd = &swapper_pg_dir[pgd_index(addr)];
269 pud_t *pud = pud_offset(pgd, addr);
270 pmd_t *pmd = pmd_offset(pud, addr);
271
272 return pmd;
269} 273}
270 274
271static inline unsigned long * __init early_ioremap_pte(unsigned long addr) 275static inline pte_t * __init early_ioremap_pte(unsigned long addr)
272{ 276{
273 return bm_pte + ((addr >> PAGE_SHIFT) & 1023); 277 return &bm_pte[pte_index(addr)];
274} 278}
275 279
276void __init early_ioremap_init(void) 280void __init early_ioremap_init(void)
277{ 281{
278 unsigned long *pgd; 282 pmd_t *pmd;
279 283
280 if (early_ioremap_debug) 284 if (early_ioremap_debug)
281 printk(KERN_INFO "early_ioremap_init()\n"); 285 printk(KERN_INFO "early_ioremap_init()\n");
282 286
283 pgd = early_ioremap_pgd(fix_to_virt(FIX_BTMAP_BEGIN)); 287 pmd = early_ioremap_pmd(fix_to_virt(FIX_BTMAP_BEGIN));
284 *pgd = __pa(bm_pte) | _PAGE_TABLE;
285 memset(bm_pte, 0, sizeof(bm_pte)); 288 memset(bm_pte, 0, sizeof(bm_pte));
289 pmd_populate_kernel(&init_mm, pmd, bm_pte);
290
286 /* 291 /*
287 * The boot-ioremap range spans multiple pgds, for which 292 * The boot-ioremap range spans multiple pmds, for which
288 * we are not prepared: 293 * we are not prepared:
289 */ 294 */
290 if (pgd != early_ioremap_pgd(fix_to_virt(FIX_BTMAP_END))) { 295 if (pmd != early_ioremap_pmd(fix_to_virt(FIX_BTMAP_END))) {
291 WARN_ON(1); 296 WARN_ON(1);
292 printk(KERN_WARNING "pgd %p != %p\n", 297 printk(KERN_WARNING "pmd %p != %p\n",
293 pgd, early_ioremap_pgd(fix_to_virt(FIX_BTMAP_END))); 298 pmd, early_ioremap_pmd(fix_to_virt(FIX_BTMAP_END)));
294 printk(KERN_WARNING "fix_to_virt(FIX_BTMAP_BEGIN): %08lx\n", 299 printk(KERN_WARNING "fix_to_virt(FIX_BTMAP_BEGIN): %08lx\n",
295 fix_to_virt(FIX_BTMAP_BEGIN)); 300 fix_to_virt(FIX_BTMAP_BEGIN));
296 printk(KERN_WARNING "fix_to_virt(FIX_BTMAP_END): %08lx\n", 301 printk(KERN_WARNING "fix_to_virt(FIX_BTMAP_END): %08lx\n",
297 fix_to_virt(FIX_BTMAP_END)); 302 fix_to_virt(FIX_BTMAP_END));
298 303
299 printk(KERN_WARNING "FIX_BTMAP_END: %d\n", FIX_BTMAP_END); 304 printk(KERN_WARNING "FIX_BTMAP_END: %d\n", FIX_BTMAP_END);
300 printk(KERN_WARNING "FIX_BTMAP_BEGIN: %d\n", 305 printk(KERN_WARNING "FIX_BTMAP_BEGIN: %d\n",
@@ -304,28 +309,29 @@ void __init early_ioremap_init(void)
304 309
305void __init early_ioremap_clear(void) 310void __init early_ioremap_clear(void)
306{ 311{
307 unsigned long *pgd; 312 pmd_t *pmd;
308 313
309 if (early_ioremap_debug) 314 if (early_ioremap_debug)
310 printk(KERN_INFO "early_ioremap_clear()\n"); 315 printk(KERN_INFO "early_ioremap_clear()\n");
311 316
312 pgd = early_ioremap_pgd(fix_to_virt(FIX_BTMAP_BEGIN)); 317 pmd = early_ioremap_pmd(fix_to_virt(FIX_BTMAP_BEGIN));
313 *pgd = 0; 318 pmd_clear(pmd);
314 paravirt_release_pt(__pa(pgd) >> PAGE_SHIFT); 319 paravirt_release_pt(__pa(bm_pte) >> PAGE_SHIFT);
315 __flush_tlb_all(); 320 __flush_tlb_all();
316} 321}
317 322
318void __init early_ioremap_reset(void) 323void __init early_ioremap_reset(void)
319{ 324{
320 enum fixed_addresses idx; 325 enum fixed_addresses idx;
321 unsigned long *pte, phys, addr; 326 unsigned long addr, phys;
327 pte_t *pte;
322 328
323 after_paging_init = 1; 329 after_paging_init = 1;
324 for (idx = FIX_BTMAP_BEGIN; idx >= FIX_BTMAP_END; idx--) { 330 for (idx = FIX_BTMAP_BEGIN; idx >= FIX_BTMAP_END; idx--) {
325 addr = fix_to_virt(idx); 331 addr = fix_to_virt(idx);
326 pte = early_ioremap_pte(addr); 332 pte = early_ioremap_pte(addr);
327 if (*pte & _PAGE_PRESENT) { 333 if (pte_present(*pte)) {
328 phys = *pte & PAGE_MASK; 334 phys = pte_val(*pte) & PAGE_MASK;
329 set_fixmap(idx, phys); 335 set_fixmap(idx, phys);
330 } 336 }
331 } 337 }
@@ -334,7 +340,8 @@ void __init early_ioremap_reset(void)
334static void __init __early_set_fixmap(enum fixed_addresses idx, 340static void __init __early_set_fixmap(enum fixed_addresses idx,
335 unsigned long phys, pgprot_t flags) 341 unsigned long phys, pgprot_t flags)
336{ 342{
337 unsigned long *pte, addr = __fix_to_virt(idx); 343 unsigned long addr = __fix_to_virt(idx);
344 pte_t *pte;
338 345
339 if (idx >= __end_of_fixed_addresses) { 346 if (idx >= __end_of_fixed_addresses) {
340 BUG(); 347 BUG();
@@ -342,9 +349,9 @@ static void __init __early_set_fixmap(enum fixed_addresses idx,
342 } 349 }
343 pte = early_ioremap_pte(addr); 350 pte = early_ioremap_pte(addr);
344 if (pgprot_val(flags)) 351 if (pgprot_val(flags))
345 *pte = (phys & PAGE_MASK) | pgprot_val(flags); 352 set_pte(pte, pfn_pte(phys >> PAGE_SHIFT, flags));
346 else 353 else
347 *pte = 0; 354 pte_clear(NULL, addr, pte);
348 __flush_tlb_one(addr); 355 __flush_tlb_one(addr);
349} 356}
350 357
diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c
index 8493c855582b..440210a2277d 100644
--- a/arch/x86/mm/pageattr.c
+++ b/arch/x86/mm/pageattr.c
@@ -8,6 +8,7 @@
8#include <linux/sched.h> 8#include <linux/sched.h>
9#include <linux/slab.h> 9#include <linux/slab.h>
10#include <linux/mm.h> 10#include <linux/mm.h>
11#include <linux/interrupt.h>
11 12
12#include <asm/e820.h> 13#include <asm/e820.h>
13#include <asm/processor.h> 14#include <asm/processor.h>
@@ -191,7 +192,7 @@ static inline pgprot_t static_protections(pgprot_t prot, unsigned long address)
191 * or when the present bit is not set. Otherwise we would return a 192 * or when the present bit is not set. Otherwise we would return a
192 * pointer to a nonexisting mapping. 193 * pointer to a nonexisting mapping.
193 */ 194 */
194pte_t *lookup_address(unsigned long address, int *level) 195pte_t *lookup_address(unsigned long address, unsigned int *level)
195{ 196{
196 pgd_t *pgd = pgd_offset_k(address); 197 pgd_t *pgd = pgd_offset_k(address);
197 pud_t *pud; 198 pud_t *pud;
@@ -252,10 +253,11 @@ static int
252try_preserve_large_page(pte_t *kpte, unsigned long address, 253try_preserve_large_page(pte_t *kpte, unsigned long address,
253 struct cpa_data *cpa) 254 struct cpa_data *cpa)
254{ 255{
255 unsigned long nextpage_addr, numpages, pmask, psize, flags; 256 unsigned long nextpage_addr, numpages, pmask, psize, flags, addr;
256 pte_t new_pte, old_pte, *tmp; 257 pte_t new_pte, old_pte, *tmp;
257 pgprot_t old_prot, new_prot; 258 pgprot_t old_prot, new_prot;
258 int level, do_split = 1; 259 int i, do_split = 1;
260 unsigned int level;
259 261
260 spin_lock_irqsave(&pgd_lock, flags); 262 spin_lock_irqsave(&pgd_lock, flags);
261 /* 263 /*
@@ -302,6 +304,19 @@ try_preserve_large_page(pte_t *kpte, unsigned long address,
302 new_prot = static_protections(new_prot, address); 304 new_prot = static_protections(new_prot, address);
303 305
304 /* 306 /*
307 * We need to check the full range, whether
308 * static_protection() requires a different pgprot for one of
309 * the pages in the range we try to preserve:
310 */
311 addr = address + PAGE_SIZE;
312 for (i = 1; i < cpa->numpages; i++, addr += PAGE_SIZE) {
313 pgprot_t chk_prot = static_protections(new_prot, addr);
314
315 if (pgprot_val(chk_prot) != pgprot_val(new_prot))
316 goto out_unlock;
317 }
318
319 /*
305 * If there are no changes, return. maxpages has been updated 320 * If there are no changes, return. maxpages has been updated
306 * above: 321 * above:
307 */ 322 */
@@ -335,23 +350,103 @@ out_unlock:
335 return do_split; 350 return do_split;
336} 351}
337 352
353static LIST_HEAD(page_pool);
354static unsigned long pool_size, pool_pages, pool_low;
355static unsigned long pool_used, pool_failed, pool_refill;
356
357static void cpa_fill_pool(void)
358{
359 struct page *p;
360 gfp_t gfp = GFP_KERNEL;
361
362 /* Do not allocate from interrupt context */
363 if (in_irq() || irqs_disabled())
364 return;
365 /*
366 * Check unlocked. I does not matter when we have one more
367 * page in the pool. The bit lock avoids recursive pool
368 * allocations:
369 */
370 if (pool_pages >= pool_size || test_and_set_bit_lock(0, &pool_refill))
371 return;
372
373#ifdef CONFIG_DEBUG_PAGEALLOC
374 /*
375 * We could do:
376 * gfp = in_atomic() ? GFP_ATOMIC : GFP_KERNEL;
377 * but this fails on !PREEMPT kernels
378 */
379 gfp = GFP_ATOMIC | __GFP_NORETRY | __GFP_NOWARN;
380#endif
381
382 while (pool_pages < pool_size) {
383 p = alloc_pages(gfp, 0);
384 if (!p) {
385 pool_failed++;
386 break;
387 }
388 spin_lock_irq(&pgd_lock);
389 list_add(&p->lru, &page_pool);
390 pool_pages++;
391 spin_unlock_irq(&pgd_lock);
392 }
393 clear_bit_unlock(0, &pool_refill);
394}
395
396#define SHIFT_MB (20 - PAGE_SHIFT)
397#define ROUND_MB_GB ((1 << 10) - 1)
398#define SHIFT_MB_GB 10
399#define POOL_PAGES_PER_GB 16
400
401void __init cpa_init(void)
402{
403 struct sysinfo si;
404 unsigned long gb;
405
406 si_meminfo(&si);
407 /*
408 * Calculate the number of pool pages:
409 *
410 * Convert totalram (nr of pages) to MiB and round to the next
411 * GiB. Shift MiB to Gib and multiply the result by
412 * POOL_PAGES_PER_GB:
413 */
414 gb = ((si.totalram >> SHIFT_MB) + ROUND_MB_GB) >> SHIFT_MB_GB;
415 pool_size = POOL_PAGES_PER_GB * gb;
416 pool_low = pool_size;
417
418 cpa_fill_pool();
419 printk(KERN_DEBUG
420 "CPA: page pool initialized %lu of %lu pages preallocated\n",
421 pool_pages, pool_size);
422}
423
338static int split_large_page(pte_t *kpte, unsigned long address) 424static int split_large_page(pte_t *kpte, unsigned long address)
339{ 425{
340 unsigned long flags, pfn, pfninc = 1; 426 unsigned long flags, pfn, pfninc = 1;
341 gfp_t gfp_flags = GFP_KERNEL;
342 unsigned int i, level; 427 unsigned int i, level;
343 pte_t *pbase, *tmp; 428 pte_t *pbase, *tmp;
344 pgprot_t ref_prot; 429 pgprot_t ref_prot;
345 struct page *base; 430 struct page *base;
346 431
347#ifdef CONFIG_DEBUG_PAGEALLOC 432 /*
348 gfp_flags = GFP_ATOMIC | __GFP_NOWARN; 433 * Get a page from the pool. The pool list is protected by the
349#endif 434 * pgd_lock, which we have to take anyway for the split
350 base = alloc_pages(gfp_flags, 0); 435 * operation:
351 if (!base) 436 */
437 spin_lock_irqsave(&pgd_lock, flags);
438 if (list_empty(&page_pool)) {
439 spin_unlock_irqrestore(&pgd_lock, flags);
352 return -ENOMEM; 440 return -ENOMEM;
441 }
442
443 base = list_first_entry(&page_pool, struct page, lru);
444 list_del(&base->lru);
445 pool_pages--;
446
447 if (pool_pages < pool_low)
448 pool_low = pool_pages;
353 449
354 spin_lock_irqsave(&pgd_lock, flags);
355 /* 450 /*
356 * Check for races, another CPU might have split this page 451 * Check for races, another CPU might have split this page
357 * up for us already: 452 * up for us already:
@@ -396,17 +491,24 @@ static int split_large_page(pte_t *kpte, unsigned long address)
396 base = NULL; 491 base = NULL;
397 492
398out_unlock: 493out_unlock:
494 /*
495 * If we dropped out via the lookup_address check under
496 * pgd_lock then stick the page back into the pool:
497 */
498 if (base) {
499 list_add(&base->lru, &page_pool);
500 pool_pages++;
501 } else
502 pool_used++;
399 spin_unlock_irqrestore(&pgd_lock, flags); 503 spin_unlock_irqrestore(&pgd_lock, flags);
400 504
401 if (base)
402 __free_pages(base, 0);
403
404 return 0; 505 return 0;
405} 506}
406 507
407static int __change_page_attr(unsigned long address, struct cpa_data *cpa) 508static int __change_page_attr(unsigned long address, struct cpa_data *cpa)
408{ 509{
409 int level, do_split, err; 510 int do_split, err;
511 unsigned int level;
410 struct page *kpte_page; 512 struct page *kpte_page;
411 pte_t *kpte; 513 pte_t *kpte;
412 514
@@ -598,7 +700,7 @@ static int change_page_attr_set_clr(unsigned long addr, int numpages,
598 * Check whether we really changed something: 700 * Check whether we really changed something:
599 */ 701 */
600 if (!cpa.flushtlb) 702 if (!cpa.flushtlb)
601 return ret; 703 goto out;
602 704
603 /* 705 /*
604 * No need to flush, when we did not set any of the caching 706 * No need to flush, when we did not set any of the caching
@@ -617,6 +719,8 @@ static int change_page_attr_set_clr(unsigned long addr, int numpages,
617 else 719 else
618 cpa_flush_all(cache); 720 cpa_flush_all(cache);
619 721
722out:
723 cpa_fill_pool();
620 return ret; 724 return ret;
621} 725}
622 726
@@ -770,6 +874,12 @@ void kernel_map_pages(struct page *page, int numpages, int enable)
770 * but that can deadlock->flush only current cpu: 874 * but that can deadlock->flush only current cpu:
771 */ 875 */
772 __flush_tlb_all(); 876 __flush_tlb_all();
877
878 /*
879 * Try to refill the page pool here. We can do this only after
880 * the tlb flush.
881 */
882 cpa_fill_pool();
773} 883}
774#endif 884#endif
775 885
diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c
index 52deabc72a6f..b7c67a187b6b 100644
--- a/arch/x86/pci/common.c
+++ b/arch/x86/pci/common.c
@@ -26,16 +26,37 @@ int pcibios_last_bus = -1;
26unsigned long pirq_table_addr; 26unsigned long pirq_table_addr;
27struct pci_bus *pci_root_bus; 27struct pci_bus *pci_root_bus;
28struct pci_raw_ops *raw_pci_ops; 28struct pci_raw_ops *raw_pci_ops;
29struct pci_raw_ops *raw_pci_ext_ops;
30
31int raw_pci_read(unsigned int domain, unsigned int bus, unsigned int devfn,
32 int reg, int len, u32 *val)
33{
34 if (reg < 256 && raw_pci_ops)
35 return raw_pci_ops->read(domain, bus, devfn, reg, len, val);
36 if (raw_pci_ext_ops)
37 return raw_pci_ext_ops->read(domain, bus, devfn, reg, len, val);
38 return -EINVAL;
39}
40
41int raw_pci_write(unsigned int domain, unsigned int bus, unsigned int devfn,
42 int reg, int len, u32 val)
43{
44 if (reg < 256 && raw_pci_ops)
45 return raw_pci_ops->write(domain, bus, devfn, reg, len, val);
46 if (raw_pci_ext_ops)
47 return raw_pci_ext_ops->write(domain, bus, devfn, reg, len, val);
48 return -EINVAL;
49}
29 50
30static int pci_read(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *value) 51static int pci_read(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *value)
31{ 52{
32 return raw_pci_ops->read(pci_domain_nr(bus), bus->number, 53 return raw_pci_read(pci_domain_nr(bus), bus->number,
33 devfn, where, size, value); 54 devfn, where, size, value);
34} 55}
35 56
36static int pci_write(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 value) 57static int pci_write(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 value)
37{ 58{
38 return raw_pci_ops->write(pci_domain_nr(bus), bus->number, 59 return raw_pci_write(pci_domain_nr(bus), bus->number,
39 devfn, where, size, value); 60 devfn, where, size, value);
40} 61}
41 62
diff --git a/arch/x86/pci/direct.c b/arch/x86/pci/direct.c
index 431c9a51b157..42f3e4cad179 100644
--- a/arch/x86/pci/direct.c
+++ b/arch/x86/pci/direct.c
@@ -14,7 +14,7 @@
14#define PCI_CONF1_ADDRESS(bus, devfn, reg) \ 14#define PCI_CONF1_ADDRESS(bus, devfn, reg) \
15 (0x80000000 | (bus << 16) | (devfn << 8) | (reg & ~3)) 15 (0x80000000 | (bus << 16) | (devfn << 8) | (reg & ~3))
16 16
17int pci_conf1_read(unsigned int seg, unsigned int bus, 17static int pci_conf1_read(unsigned int seg, unsigned int bus,
18 unsigned int devfn, int reg, int len, u32 *value) 18 unsigned int devfn, int reg, int len, u32 *value)
19{ 19{
20 unsigned long flags; 20 unsigned long flags;
@@ -45,7 +45,7 @@ int pci_conf1_read(unsigned int seg, unsigned int bus,
45 return 0; 45 return 0;
46} 46}
47 47
48int pci_conf1_write(unsigned int seg, unsigned int bus, 48static int pci_conf1_write(unsigned int seg, unsigned int bus,
49 unsigned int devfn, int reg, int len, u32 value) 49 unsigned int devfn, int reg, int len, u32 value)
50{ 50{
51 unsigned long flags; 51 unsigned long flags;
diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c
index 74d30ff33c49..a5ef5f551373 100644
--- a/arch/x86/pci/fixup.c
+++ b/arch/x86/pci/fixup.c
@@ -215,7 +215,8 @@ static int quirk_aspm_offset[MAX_PCIEROOT << 3];
215 215
216static int quirk_pcie_aspm_read(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *value) 216static int quirk_pcie_aspm_read(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *value)
217{ 217{
218 return raw_pci_ops->read(0, bus->number, devfn, where, size, value); 218 return raw_pci_read(pci_domain_nr(bus), bus->number,
219 devfn, where, size, value);
219} 220}
220 221
221/* 222/*
@@ -231,7 +232,8 @@ static int quirk_pcie_aspm_write(struct pci_bus *bus, unsigned int devfn, int wh
231 if ((offset) && (where == offset)) 232 if ((offset) && (where == offset))
232 value = value & 0xfffffffc; 233 value = value & 0xfffffffc;
233 234
234 return raw_pci_ops->write(0, bus->number, devfn, where, size, value); 235 return raw_pci_write(pci_domain_nr(bus), bus->number,
236 devfn, where, size, value);
235} 237}
236 238
237static struct pci_ops quirk_pcie_aspm_ops = { 239static struct pci_ops quirk_pcie_aspm_ops = {
diff --git a/arch/x86/pci/legacy.c b/arch/x86/pci/legacy.c
index 5565d7016b75..e041ced0ce13 100644
--- a/arch/x86/pci/legacy.c
+++ b/arch/x86/pci/legacy.c
@@ -22,7 +22,7 @@ static void __devinit pcibios_fixup_peer_bridges(void)
22 if (pci_find_bus(0, n)) 22 if (pci_find_bus(0, n))
23 continue; 23 continue;
24 for (devfn = 0; devfn < 256; devfn += 8) { 24 for (devfn = 0; devfn < 256; devfn += 8) {
25 if (!raw_pci_ops->read(0, n, devfn, PCI_VENDOR_ID, 2, &l) && 25 if (!raw_pci_read(0, n, devfn, PCI_VENDOR_ID, 2, &l) &&
26 l != 0x0000 && l != 0xffff) { 26 l != 0x0000 && l != 0xffff) {
27 DBG("Found device at %02x:%02x [%04x]\n", n, devfn, l); 27 DBG("Found device at %02x:%02x [%04x]\n", n, devfn, l);
28 printk(KERN_INFO "PCI: Discovered peer bus %02x\n", n); 28 printk(KERN_INFO "PCI: Discovered peer bus %02x\n", n);
diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c
index 4df637e34f81..8d54df4dfaad 100644
--- a/arch/x86/pci/mmconfig-shared.c
+++ b/arch/x86/pci/mmconfig-shared.c
@@ -22,46 +22,13 @@
22#define MMCONFIG_APER_MIN (2 * 1024*1024) 22#define MMCONFIG_APER_MIN (2 * 1024*1024)
23#define MMCONFIG_APER_MAX (256 * 1024*1024) 23#define MMCONFIG_APER_MAX (256 * 1024*1024)
24 24
25DECLARE_BITMAP(pci_mmcfg_fallback_slots, 32*PCI_MMCFG_MAX_CHECK_BUS);
26
27/* Indicate if the mmcfg resources have been placed into the resource table. */ 25/* Indicate if the mmcfg resources have been placed into the resource table. */
28static int __initdata pci_mmcfg_resources_inserted; 26static int __initdata pci_mmcfg_resources_inserted;
29 27
30/* K8 systems have some devices (typically in the builtin northbridge)
31 that are only accessible using type1
32 Normally this can be expressed in the MCFG by not listing them
33 and assigning suitable _SEGs, but this isn't implemented in some BIOS.
34 Instead try to discover all devices on bus 0 that are unreachable using MM
35 and fallback for them. */
36static void __init unreachable_devices(void)
37{
38 int i, bus;
39 /* Use the max bus number from ACPI here? */
40 for (bus = 0; bus < PCI_MMCFG_MAX_CHECK_BUS; bus++) {
41 for (i = 0; i < 32; i++) {
42 unsigned int devfn = PCI_DEVFN(i, 0);
43 u32 val1, val2;
44
45 pci_conf1_read(0, bus, devfn, 0, 4, &val1);
46 if (val1 == 0xffffffff)
47 continue;
48
49 if (pci_mmcfg_arch_reachable(0, bus, devfn)) {
50 raw_pci_ops->read(0, bus, devfn, 0, 4, &val2);
51 if (val1 == val2)
52 continue;
53 }
54 set_bit(i + 32 * bus, pci_mmcfg_fallback_slots);
55 printk(KERN_NOTICE "PCI: No mmconfig possible on device"
56 " %02x:%02x\n", bus, i);
57 }
58 }
59}
60
61static const char __init *pci_mmcfg_e7520(void) 28static const char __init *pci_mmcfg_e7520(void)
62{ 29{
63 u32 win; 30 u32 win;
64 pci_conf1_read(0, 0, PCI_DEVFN(0,0), 0xce, 2, &win); 31 pci_direct_conf1.read(0, 0, PCI_DEVFN(0,0), 0xce, 2, &win);
65 32
66 win = win & 0xf000; 33 win = win & 0xf000;
67 if(win == 0x0000 || win == 0xf000) 34 if(win == 0x0000 || win == 0xf000)
@@ -86,7 +53,7 @@ static const char __init *pci_mmcfg_intel_945(void)
86 53
87 pci_mmcfg_config_num = 1; 54 pci_mmcfg_config_num = 1;
88 55
89 pci_conf1_read(0, 0, PCI_DEVFN(0,0), 0x48, 4, &pciexbar); 56 pci_direct_conf1.read(0, 0, PCI_DEVFN(0,0), 0x48, 4, &pciexbar);
90 57
91 /* Enable bit */ 58 /* Enable bit */
92 if (!(pciexbar & 1)) 59 if (!(pciexbar & 1))
@@ -151,7 +118,7 @@ static int __init pci_mmcfg_check_hostbridge(void)
151 int i; 118 int i;
152 const char *name; 119 const char *name;
153 120
154 pci_conf1_read(0, 0, PCI_DEVFN(0,0), 0, 4, &l); 121 pci_direct_conf1.read(0, 0, PCI_DEVFN(0,0), 0, 4, &l);
155 vendor = l & 0xffff; 122 vendor = l & 0xffff;
156 device = (l >> 16) & 0xffff; 123 device = (l >> 16) & 0xffff;
157 124
@@ -270,8 +237,6 @@ void __init pci_mmcfg_init(int type)
270 return; 237 return;
271 238
272 if (pci_mmcfg_arch_init()) { 239 if (pci_mmcfg_arch_init()) {
273 if (type == 1)
274 unreachable_devices();
275 if (known_bridge) 240 if (known_bridge)
276 pci_mmcfg_insert_resources(IORESOURCE_BUSY); 241 pci_mmcfg_insert_resources(IORESOURCE_BUSY);
277 pci_probe = (pci_probe & ~PCI_PROBE_MASK) | PCI_PROBE_MMCONF; 242 pci_probe = (pci_probe & ~PCI_PROBE_MASK) | PCI_PROBE_MMCONF;
diff --git a/arch/x86/pci/mmconfig_32.c b/arch/x86/pci/mmconfig_32.c
index 1bf5816d34c8..081816ada057 100644
--- a/arch/x86/pci/mmconfig_32.c
+++ b/arch/x86/pci/mmconfig_32.c
@@ -30,10 +30,6 @@ static u32 get_base_addr(unsigned int seg, int bus, unsigned devfn)
30 struct acpi_mcfg_allocation *cfg; 30 struct acpi_mcfg_allocation *cfg;
31 int cfg_num; 31 int cfg_num;
32 32
33 if (seg == 0 && bus < PCI_MMCFG_MAX_CHECK_BUS &&
34 test_bit(PCI_SLOT(devfn) + 32*bus, pci_mmcfg_fallback_slots))
35 return 0;
36
37 for (cfg_num = 0; cfg_num < pci_mmcfg_config_num; cfg_num++) { 33 for (cfg_num = 0; cfg_num < pci_mmcfg_config_num; cfg_num++) {
38 cfg = &pci_mmcfg_config[cfg_num]; 34 cfg = &pci_mmcfg_config[cfg_num];
39 if (cfg->pci_segment == seg && 35 if (cfg->pci_segment == seg &&
@@ -68,13 +64,13 @@ static int pci_mmcfg_read(unsigned int seg, unsigned int bus,
68 u32 base; 64 u32 base;
69 65
70 if ((bus > 255) || (devfn > 255) || (reg > 4095)) { 66 if ((bus > 255) || (devfn > 255) || (reg > 4095)) {
71 *value = -1; 67err: *value = -1;
72 return -EINVAL; 68 return -EINVAL;
73 } 69 }
74 70
75 base = get_base_addr(seg, bus, devfn); 71 base = get_base_addr(seg, bus, devfn);
76 if (!base) 72 if (!base)
77 return pci_conf1_read(seg,bus,devfn,reg,len,value); 73 goto err;
78 74
79 spin_lock_irqsave(&pci_config_lock, flags); 75 spin_lock_irqsave(&pci_config_lock, flags);
80 76
@@ -107,7 +103,7 @@ static int pci_mmcfg_write(unsigned int seg, unsigned int bus,
107 103
108 base = get_base_addr(seg, bus, devfn); 104 base = get_base_addr(seg, bus, devfn);
109 if (!base) 105 if (!base)
110 return pci_conf1_write(seg,bus,devfn,reg,len,value); 106 return -EINVAL;
111 107
112 spin_lock_irqsave(&pci_config_lock, flags); 108 spin_lock_irqsave(&pci_config_lock, flags);
113 109
@@ -134,15 +130,9 @@ static struct pci_raw_ops pci_mmcfg = {
134 .write = pci_mmcfg_write, 130 .write = pci_mmcfg_write,
135}; 131};
136 132
137int __init pci_mmcfg_arch_reachable(unsigned int seg, unsigned int bus,
138 unsigned int devfn)
139{
140 return get_base_addr(seg, bus, devfn) != 0;
141}
142
143int __init pci_mmcfg_arch_init(void) 133int __init pci_mmcfg_arch_init(void)
144{ 134{
145 printk(KERN_INFO "PCI: Using MMCONFIG\n"); 135 printk(KERN_INFO "PCI: Using MMCONFIG for extended config space\n");
146 raw_pci_ops = &pci_mmcfg; 136 raw_pci_ext_ops = &pci_mmcfg;
147 return 1; 137 return 1;
148} 138}
diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c
index 4095e4d66a1d..9207fd49233c 100644
--- a/arch/x86/pci/mmconfig_64.c
+++ b/arch/x86/pci/mmconfig_64.c
@@ -40,9 +40,7 @@ static char __iomem *get_virt(unsigned int seg, unsigned bus)
40static char __iomem *pci_dev_base(unsigned int seg, unsigned int bus, unsigned int devfn) 40static char __iomem *pci_dev_base(unsigned int seg, unsigned int bus, unsigned int devfn)
41{ 41{
42 char __iomem *addr; 42 char __iomem *addr;
43 if (seg == 0 && bus < PCI_MMCFG_MAX_CHECK_BUS && 43
44 test_bit(32*bus + PCI_SLOT(devfn), pci_mmcfg_fallback_slots))
45 return NULL;
46 addr = get_virt(seg, bus); 44 addr = get_virt(seg, bus);
47 if (!addr) 45 if (!addr)
48 return NULL; 46 return NULL;
@@ -56,13 +54,13 @@ static int pci_mmcfg_read(unsigned int seg, unsigned int bus,
56 54
57 /* Why do we have this when nobody checks it. How about a BUG()!? -AK */ 55 /* Why do we have this when nobody checks it. How about a BUG()!? -AK */
58 if (unlikely((bus > 255) || (devfn > 255) || (reg > 4095))) { 56 if (unlikely((bus > 255) || (devfn > 255) || (reg > 4095))) {
59 *value = -1; 57err: *value = -1;
60 return -EINVAL; 58 return -EINVAL;
61 } 59 }
62 60
63 addr = pci_dev_base(seg, bus, devfn); 61 addr = pci_dev_base(seg, bus, devfn);
64 if (!addr) 62 if (!addr)
65 return pci_conf1_read(seg,bus,devfn,reg,len,value); 63 goto err;
66 64
67 switch (len) { 65 switch (len) {
68 case 1: 66 case 1:
@@ -90,7 +88,7 @@ static int pci_mmcfg_write(unsigned int seg, unsigned int bus,
90 88
91 addr = pci_dev_base(seg, bus, devfn); 89 addr = pci_dev_base(seg, bus, devfn);
92 if (!addr) 90 if (!addr)
93 return pci_conf1_write(seg,bus,devfn,reg,len,value); 91 return -EINVAL;
94 92
95 switch (len) { 93 switch (len) {
96 case 1: 94 case 1:
@@ -126,12 +124,6 @@ static void __iomem * __init mcfg_ioremap(struct acpi_mcfg_allocation *cfg)
126 return addr; 124 return addr;
127} 125}
128 126
129int __init pci_mmcfg_arch_reachable(unsigned int seg, unsigned int bus,
130 unsigned int devfn)
131{
132 return pci_dev_base(seg, bus, devfn) != NULL;
133}
134
135int __init pci_mmcfg_arch_init(void) 127int __init pci_mmcfg_arch_init(void)
136{ 128{
137 int i; 129 int i;
@@ -152,6 +144,6 @@ int __init pci_mmcfg_arch_init(void)
152 return 0; 144 return 0;
153 } 145 }
154 } 146 }
155 raw_pci_ops = &pci_mmcfg; 147 raw_pci_ext_ops = &pci_mmcfg;
156 return 1; 148 return 1;
157} 149}
diff --git a/arch/x86/pci/pci.h b/arch/x86/pci/pci.h
index ac56d3916c50..3431518d921a 100644
--- a/arch/x86/pci/pci.h
+++ b/arch/x86/pci/pci.h
@@ -85,10 +85,17 @@ extern spinlock_t pci_config_lock;
85extern int (*pcibios_enable_irq)(struct pci_dev *dev); 85extern int (*pcibios_enable_irq)(struct pci_dev *dev);
86extern void (*pcibios_disable_irq)(struct pci_dev *dev); 86extern void (*pcibios_disable_irq)(struct pci_dev *dev);
87 87
88extern int pci_conf1_write(unsigned int seg, unsigned int bus, 88struct pci_raw_ops {
89 unsigned int devfn, int reg, int len, u32 value); 89 int (*read)(unsigned int domain, unsigned int bus, unsigned int devfn,
90extern int pci_conf1_read(unsigned int seg, unsigned int bus, 90 int reg, int len, u32 *val);
91 unsigned int devfn, int reg, int len, u32 *value); 91 int (*write)(unsigned int domain, unsigned int bus, unsigned int devfn,
92 int reg, int len, u32 val);
93};
94
95extern struct pci_raw_ops *raw_pci_ops;
96extern struct pci_raw_ops *raw_pci_ext_ops;
97
98extern struct pci_raw_ops pci_direct_conf1;
92 99
93extern int pci_direct_probe(void); 100extern int pci_direct_probe(void);
94extern void pci_direct_init(int type); 101extern void pci_direct_init(int type);
@@ -98,13 +105,6 @@ extern void pcibios_sort(void);
98 105
99/* pci-mmconfig.c */ 106/* pci-mmconfig.c */
100 107
101/* Verify the first 16 busses. We assume that systems with more busses
102 get MCFG right. */
103#define PCI_MMCFG_MAX_CHECK_BUS 16
104extern DECLARE_BITMAP(pci_mmcfg_fallback_slots, 32*PCI_MMCFG_MAX_CHECK_BUS);
105
106extern int __init pci_mmcfg_arch_reachable(unsigned int seg, unsigned int bus,
107 unsigned int devfn);
108extern int __init pci_mmcfg_arch_init(void); 108extern int __init pci_mmcfg_arch_init(void);
109 109
110/* 110/*
diff --git a/arch/x86/pci/visws.c b/arch/x86/pci/visws.c
index 8ecb1c722594..c2df4e97eed6 100644
--- a/arch/x86/pci/visws.c
+++ b/arch/x86/pci/visws.c
@@ -13,9 +13,6 @@
13 13
14#include "pci.h" 14#include "pci.h"
15 15
16
17extern struct pci_raw_ops pci_direct_conf1;
18
19static int pci_visws_enable_irq(struct pci_dev *dev) { return 0; } 16static int pci_visws_enable_irq(struct pci_dev *dev) { return 0; }
20static void pci_visws_disable_irq(struct pci_dev *dev) { } 17static void pci_visws_disable_irq(struct pci_dev *dev) { }
21 18
diff --git a/arch/x86/power/Makefile b/arch/x86/power/Makefile
index d764ec950065..9ff4d5b55ad1 100644
--- a/arch/x86/power/Makefile
+++ b/arch/x86/power/Makefile
@@ -1,2 +1,2 @@
1obj-$(CONFIG_PM) += cpu.o 1obj-$(CONFIG_PM_SLEEP) += cpu_$(BITS).o
2obj-$(CONFIG_HIBERNATION) += swsusp.o suspend.o 2obj-$(CONFIG_HIBERNATION) += hibernate_$(BITS).o hibernate_asm_$(BITS).o
diff --git a/arch/x86/power/cpu.c b/arch/x86/power/cpu_32.c
index efcf620d1439..7f9c6da04a4c 100644
--- a/arch/x86/power/cpu.c
+++ b/arch/x86/power/cpu_32.c
@@ -40,7 +40,7 @@ static void __save_processor_state(struct saved_context *ctxt)
40 savesegment(ss, ctxt->ss); 40 savesegment(ss, ctxt->ss);
41 41
42 /* 42 /*
43 * control registers 43 * control registers
44 */ 44 */
45 ctxt->cr0 = read_cr0(); 45 ctxt->cr0 = read_cr0();
46 ctxt->cr2 = read_cr2(); 46 ctxt->cr2 = read_cr2();
diff --git a/arch/x86/kernel/suspend_64.c b/arch/x86/power/cpu_64.c
index 7ac7130022f1..66bdfb591fd8 100644
--- a/arch/x86/kernel/suspend_64.c
+++ b/arch/x86/power/cpu_64.c
@@ -1,8 +1,9 @@
1/* 1/*
2 * Suspend support specific for i386. 2 * Suspend and hibernation support for x86-64
3 * 3 *
4 * Distribute under GPLv2 4 * Distribute under GPLv2
5 * 5 *
6 * Copyright (c) 2007 Rafael J. Wysocki <rjw@sisk.pl>
6 * Copyright (c) 2002 Pavel Machek <pavel@suse.cz> 7 * Copyright (c) 2002 Pavel Machek <pavel@suse.cz>
7 * Copyright (c) 2001 Patrick Mochel <mochel@osdl.org> 8 * Copyright (c) 2001 Patrick Mochel <mochel@osdl.org>
8 */ 9 */
@@ -14,9 +15,6 @@
14#include <asm/pgtable.h> 15#include <asm/pgtable.h>
15#include <asm/mtrr.h> 16#include <asm/mtrr.h>
16 17
17/* References to section boundaries */
18extern const void __nosave_begin, __nosave_end;
19
20static void fix_processor_context(void); 18static void fix_processor_context(void);
21 19
22struct saved_context saved_context; 20struct saved_context saved_context;
@@ -63,7 +61,7 @@ static void __save_processor_state(struct saved_context *ctxt)
63 mtrr_save_fixed_ranges(NULL); 61 mtrr_save_fixed_ranges(NULL);
64 62
65 /* 63 /*
66 * control registers 64 * control registers
67 */ 65 */
68 rdmsrl(MSR_EFER, ctxt->efer); 66 rdmsrl(MSR_EFER, ctxt->efer);
69 ctxt->cr0 = read_cr0(); 67 ctxt->cr0 = read_cr0();
@@ -166,155 +164,3 @@ static void fix_processor_context(void)
166 loaddebug(&current->thread, 7); 164 loaddebug(&current->thread, 7);
167 } 165 }
168} 166}
169
170#ifdef CONFIG_HIBERNATION
171/* Defined in arch/x86_64/kernel/suspend_asm.S */
172extern int restore_image(void);
173
174/*
175 * Address to jump to in the last phase of restore in order to get to the image
176 * kernel's text (this value is passed in the image header).
177 */
178unsigned long restore_jump_address;
179
180/*
181 * Value of the cr3 register from before the hibernation (this value is passed
182 * in the image header).
183 */
184unsigned long restore_cr3;
185
186pgd_t *temp_level4_pgt;
187
188void *relocated_restore_code;
189
190static int res_phys_pud_init(pud_t *pud, unsigned long address, unsigned long end)
191{
192 long i, j;
193
194 i = pud_index(address);
195 pud = pud + i;
196 for (; i < PTRS_PER_PUD; pud++, i++) {
197 unsigned long paddr;
198 pmd_t *pmd;
199
200 paddr = address + i*PUD_SIZE;
201 if (paddr >= end)
202 break;
203
204 pmd = (pmd_t *)get_safe_page(GFP_ATOMIC);
205 if (!pmd)
206 return -ENOMEM;
207 set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE));
208 for (j = 0; j < PTRS_PER_PMD; pmd++, j++, paddr += PMD_SIZE) {
209 unsigned long pe;
210
211 if (paddr >= end)
212 break;
213 pe = __PAGE_KERNEL_LARGE_EXEC | paddr;
214 pe &= __supported_pte_mask;
215 set_pmd(pmd, __pmd(pe));
216 }
217 }
218 return 0;
219}
220
221static int set_up_temporary_mappings(void)
222{
223 unsigned long start, end, next;
224 int error;
225
226 temp_level4_pgt = (pgd_t *)get_safe_page(GFP_ATOMIC);
227 if (!temp_level4_pgt)
228 return -ENOMEM;
229
230 /* It is safe to reuse the original kernel mapping */
231 set_pgd(temp_level4_pgt + pgd_index(__START_KERNEL_map),
232 init_level4_pgt[pgd_index(__START_KERNEL_map)]);
233
234 /* Set up the direct mapping from scratch */
235 start = (unsigned long)pfn_to_kaddr(0);
236 end = (unsigned long)pfn_to_kaddr(end_pfn);
237
238 for (; start < end; start = next) {
239 pud_t *pud = (pud_t *)get_safe_page(GFP_ATOMIC);
240 if (!pud)
241 return -ENOMEM;
242 next = start + PGDIR_SIZE;
243 if (next > end)
244 next = end;
245 if ((error = res_phys_pud_init(pud, __pa(start), __pa(next))))
246 return error;
247 set_pgd(temp_level4_pgt + pgd_index(start),
248 mk_kernel_pgd(__pa(pud)));
249 }
250 return 0;
251}
252
253int swsusp_arch_resume(void)
254{
255 int error;
256
257 /* We have got enough memory and from now on we cannot recover */
258 if ((error = set_up_temporary_mappings()))
259 return error;
260
261 relocated_restore_code = (void *)get_safe_page(GFP_ATOMIC);
262 if (!relocated_restore_code)
263 return -ENOMEM;
264 memcpy(relocated_restore_code, &core_restore_code,
265 &restore_registers - &core_restore_code);
266
267 restore_image();
268 return 0;
269}
270
271/*
272 * pfn_is_nosave - check if given pfn is in the 'nosave' section
273 */
274
275int pfn_is_nosave(unsigned long pfn)
276{
277 unsigned long nosave_begin_pfn = __pa_symbol(&__nosave_begin) >> PAGE_SHIFT;
278 unsigned long nosave_end_pfn = PAGE_ALIGN(__pa_symbol(&__nosave_end)) >> PAGE_SHIFT;
279 return (pfn >= nosave_begin_pfn) && (pfn < nosave_end_pfn);
280}
281
282struct restore_data_record {
283 unsigned long jump_address;
284 unsigned long cr3;
285 unsigned long magic;
286};
287
288#define RESTORE_MAGIC 0x0123456789ABCDEFUL
289
290/**
291 * arch_hibernation_header_save - populate the architecture specific part
292 * of a hibernation image header
293 * @addr: address to save the data at
294 */
295int arch_hibernation_header_save(void *addr, unsigned int max_size)
296{
297 struct restore_data_record *rdr = addr;
298
299 if (max_size < sizeof(struct restore_data_record))
300 return -EOVERFLOW;
301 rdr->jump_address = restore_jump_address;
302 rdr->cr3 = restore_cr3;
303 rdr->magic = RESTORE_MAGIC;
304 return 0;
305}
306
307/**
308 * arch_hibernation_header_restore - read the architecture specific data
309 * from the hibernation image header
310 * @addr: address to read the data from
311 */
312int arch_hibernation_header_restore(void *addr)
313{
314 struct restore_data_record *rdr = addr;
315
316 restore_jump_address = rdr->jump_address;
317 restore_cr3 = rdr->cr3;
318 return (rdr->magic == RESTORE_MAGIC) ? 0 : -EINVAL;
319}
320#endif /* CONFIG_HIBERNATION */
diff --git a/arch/x86/power/suspend.c b/arch/x86/power/hibernate_32.c
index a0020b913f31..f2b6e3f11bfc 100644
--- a/arch/x86/power/suspend.c
+++ b/arch/x86/power/hibernate_32.c
@@ -1,5 +1,5 @@
1/* 1/*
2 * Suspend support specific for i386 - temporary page tables 2 * Hibernation support specific for i386 - temporary page tables
3 * 3 *
4 * Distribute under GPLv2 4 * Distribute under GPLv2
5 * 5 *
@@ -13,7 +13,7 @@
13#include <asm/page.h> 13#include <asm/page.h>
14#include <asm/pgtable.h> 14#include <asm/pgtable.h>
15 15
16/* Defined in arch/i386/power/swsusp.S */ 16/* Defined in hibernate_asm_32.S */
17extern int restore_image(void); 17extern int restore_image(void);
18 18
19/* References to section boundaries */ 19/* References to section boundaries */
@@ -23,7 +23,7 @@ extern const void __nosave_begin, __nosave_end;
23pgd_t *resume_pg_dir; 23pgd_t *resume_pg_dir;
24 24
25/* The following three functions are based on the analogous code in 25/* The following three functions are based on the analogous code in
26 * arch/i386/mm/init.c 26 * arch/x86/mm/init_32.c
27 */ 27 */
28 28
29/* 29/*
diff --git a/arch/x86/power/hibernate_64.c b/arch/x86/power/hibernate_64.c
new file mode 100644
index 000000000000..b542355e0e34
--- /dev/null
+++ b/arch/x86/power/hibernate_64.c
@@ -0,0 +1,169 @@
1/*
2 * Hibernation support for x86-64
3 *
4 * Distribute under GPLv2
5 *
6 * Copyright (c) 2007 Rafael J. Wysocki <rjw@sisk.pl>
7 * Copyright (c) 2002 Pavel Machek <pavel@suse.cz>
8 * Copyright (c) 2001 Patrick Mochel <mochel@osdl.org>
9 */
10
11#include <linux/smp.h>
12#include <linux/suspend.h>
13#include <asm/proto.h>
14#include <asm/page.h>
15#include <asm/pgtable.h>
16#include <asm/mtrr.h>
17
18/* References to section boundaries */
19extern const void __nosave_begin, __nosave_end;
20
21/* Defined in hibernate_asm_64.S */
22extern int restore_image(void);
23
24/*
25 * Address to jump to in the last phase of restore in order to get to the image
26 * kernel's text (this value is passed in the image header).
27 */
28unsigned long restore_jump_address;
29
30/*
31 * Value of the cr3 register from before the hibernation (this value is passed
32 * in the image header).
33 */
34unsigned long restore_cr3;
35
36pgd_t *temp_level4_pgt;
37
38void *relocated_restore_code;
39
40static int res_phys_pud_init(pud_t *pud, unsigned long address, unsigned long end)
41{
42 long i, j;
43
44 i = pud_index(address);
45 pud = pud + i;
46 for (; i < PTRS_PER_PUD; pud++, i++) {
47 unsigned long paddr;
48 pmd_t *pmd;
49
50 paddr = address + i*PUD_SIZE;
51 if (paddr >= end)
52 break;
53
54 pmd = (pmd_t *)get_safe_page(GFP_ATOMIC);
55 if (!pmd)
56 return -ENOMEM;
57 set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE));
58 for (j = 0; j < PTRS_PER_PMD; pmd++, j++, paddr += PMD_SIZE) {
59 unsigned long pe;
60
61 if (paddr >= end)
62 break;
63 pe = __PAGE_KERNEL_LARGE_EXEC | paddr;
64 pe &= __supported_pte_mask;
65 set_pmd(pmd, __pmd(pe));
66 }
67 }
68 return 0;
69}
70
71static int set_up_temporary_mappings(void)
72{
73 unsigned long start, end, next;
74 int error;
75
76 temp_level4_pgt = (pgd_t *)get_safe_page(GFP_ATOMIC);
77 if (!temp_level4_pgt)
78 return -ENOMEM;
79
80 /* It is safe to reuse the original kernel mapping */
81 set_pgd(temp_level4_pgt + pgd_index(__START_KERNEL_map),
82 init_level4_pgt[pgd_index(__START_KERNEL_map)]);
83
84 /* Set up the direct mapping from scratch */
85 start = (unsigned long)pfn_to_kaddr(0);
86 end = (unsigned long)pfn_to_kaddr(end_pfn);
87
88 for (; start < end; start = next) {
89 pud_t *pud = (pud_t *)get_safe_page(GFP_ATOMIC);
90 if (!pud)
91 return -ENOMEM;
92 next = start + PGDIR_SIZE;
93 if (next > end)
94 next = end;
95 if ((error = res_phys_pud_init(pud, __pa(start), __pa(next))))
96 return error;
97 set_pgd(temp_level4_pgt + pgd_index(start),
98 mk_kernel_pgd(__pa(pud)));
99 }
100 return 0;
101}
102
103int swsusp_arch_resume(void)
104{
105 int error;
106
107 /* We have got enough memory and from now on we cannot recover */
108 if ((error = set_up_temporary_mappings()))
109 return error;
110
111 relocated_restore_code = (void *)get_safe_page(GFP_ATOMIC);
112 if (!relocated_restore_code)
113 return -ENOMEM;
114 memcpy(relocated_restore_code, &core_restore_code,
115 &restore_registers - &core_restore_code);
116
117 restore_image();
118 return 0;
119}
120
121/*
122 * pfn_is_nosave - check if given pfn is in the 'nosave' section
123 */
124
125int pfn_is_nosave(unsigned long pfn)
126{
127 unsigned long nosave_begin_pfn = __pa_symbol(&__nosave_begin) >> PAGE_SHIFT;
128 unsigned long nosave_end_pfn = PAGE_ALIGN(__pa_symbol(&__nosave_end)) >> PAGE_SHIFT;
129 return (pfn >= nosave_begin_pfn) && (pfn < nosave_end_pfn);
130}
131
132struct restore_data_record {
133 unsigned long jump_address;
134 unsigned long cr3;
135 unsigned long magic;
136};
137
138#define RESTORE_MAGIC 0x0123456789ABCDEFUL
139
140/**
141 * arch_hibernation_header_save - populate the architecture specific part
142 * of a hibernation image header
143 * @addr: address to save the data at
144 */
145int arch_hibernation_header_save(void *addr, unsigned int max_size)
146{
147 struct restore_data_record *rdr = addr;
148
149 if (max_size < sizeof(struct restore_data_record))
150 return -EOVERFLOW;
151 rdr->jump_address = restore_jump_address;
152 rdr->cr3 = restore_cr3;
153 rdr->magic = RESTORE_MAGIC;
154 return 0;
155}
156
157/**
158 * arch_hibernation_header_restore - read the architecture specific data
159 * from the hibernation image header
160 * @addr: address to read the data from
161 */
162int arch_hibernation_header_restore(void *addr)
163{
164 struct restore_data_record *rdr = addr;
165
166 restore_jump_address = rdr->jump_address;
167 restore_cr3 = rdr->cr3;
168 return (rdr->magic == RESTORE_MAGIC) ? 0 : -EINVAL;
169}
diff --git a/arch/x86/power/swsusp.S b/arch/x86/power/hibernate_asm_32.S
index 53662e05b393..b95aa6cfe3cb 100644
--- a/arch/x86/power/swsusp.S
+++ b/arch/x86/power/hibernate_asm_32.S
@@ -1,7 +1,6 @@
1.text 1.text
2 2
3/* Originally gcc generated, modified by hand 3/*
4 *
5 * This may not use any stack, nor any variable that is not "NoSave": 4 * This may not use any stack, nor any variable that is not "NoSave":
6 * 5 *
7 * Its rewriting one kernel image with another. What is stack in "old" 6 * Its rewriting one kernel image with another. What is stack in "old"
diff --git a/arch/x86/kernel/suspend_asm_64.S b/arch/x86/power/hibernate_asm_64.S
index aeb9a4d7681e..1deb3244b99b 100644
--- a/arch/x86/kernel/suspend_asm_64.S
+++ b/arch/x86/power/hibernate_asm_64.S
@@ -1,7 +1,12 @@
1/* Copyright 2004,2005 Pavel Machek <pavel@suse.cz>, Andi Kleen <ak@suse.de>, Rafael J. Wysocki <rjw@sisk.pl> 1/*
2 * Hibernation support for x86-64
2 * 3 *
3 * Distribute under GPLv2. 4 * Distribute under GPLv2.
4 * 5 *
6 * Copyright 2007 Rafael J. Wysocki <rjw@sisk.pl>
7 * Copyright 2005 Andi Kleen <ak@suse.de>
8 * Copyright 2004 Pavel Machek <pavel@suse.cz>
9 *
5 * swsusp_arch_resume must not use any stack or any nonlocal variables while 10 * swsusp_arch_resume must not use any stack or any nonlocal variables while
6 * copying pages: 11 * copying pages:
7 * 12 *
@@ -9,7 +14,7 @@
9 * image could very well be data page in "new" image, and overwriting 14 * image could very well be data page in "new" image, and overwriting
10 * your own stack under you is bad idea. 15 * your own stack under you is bad idea.
11 */ 16 */
12 17
13 .text 18 .text
14#include <linux/linkage.h> 19#include <linux/linkage.h>
15#include <asm/segment.h> 20#include <asm/segment.h>
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index 45aa771e73a9..0144395448ae 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -58,7 +58,7 @@
58 58
59xmaddr_t arbitrary_virt_to_machine(unsigned long address) 59xmaddr_t arbitrary_virt_to_machine(unsigned long address)
60{ 60{
61 int level; 61 unsigned int level;
62 pte_t *pte = lookup_address(address, &level); 62 pte_t *pte = lookup_address(address, &level);
63 unsigned offset = address & PAGE_MASK; 63 unsigned offset = address & PAGE_MASK;
64 64
@@ -71,7 +71,7 @@ void make_lowmem_page_readonly(void *vaddr)
71{ 71{
72 pte_t *pte, ptev; 72 pte_t *pte, ptev;
73 unsigned long address = (unsigned long)vaddr; 73 unsigned long address = (unsigned long)vaddr;
74 int level; 74 unsigned int level;
75 75
76 pte = lookup_address(address, &level); 76 pte = lookup_address(address, &level);
77 BUG_ON(pte == NULL); 77 BUG_ON(pte == NULL);
@@ -86,7 +86,7 @@ void make_lowmem_page_readwrite(void *vaddr)
86{ 86{
87 pte_t *pte, ptev; 87 pte_t *pte, ptev;
88 unsigned long address = (unsigned long)vaddr; 88 unsigned long address = (unsigned long)vaddr;
89 int level; 89 unsigned int level;
90 90
91 pte = lookup_address(address, &level); 91 pte = lookup_address(address, &level);
92 BUG_ON(pte == NULL); 92 BUG_ON(pte == NULL);
diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c
index b3721fd6877b..c39e1a5aa241 100644
--- a/arch/x86/xen/time.c
+++ b/arch/x86/xen/time.c
@@ -217,17 +217,17 @@ unsigned long long xen_sched_clock(void)
217/* Get the CPU speed from Xen */ 217/* Get the CPU speed from Xen */
218unsigned long xen_cpu_khz(void) 218unsigned long xen_cpu_khz(void)
219{ 219{
220 u64 cpu_khz = 1000000ULL << 32; 220 u64 xen_khz = 1000000ULL << 32;
221 const struct vcpu_time_info *info = 221 const struct vcpu_time_info *info =
222 &HYPERVISOR_shared_info->vcpu_info[0].time; 222 &HYPERVISOR_shared_info->vcpu_info[0].time;
223 223
224 do_div(cpu_khz, info->tsc_to_system_mul); 224 do_div(xen_khz, info->tsc_to_system_mul);
225 if (info->tsc_shift < 0) 225 if (info->tsc_shift < 0)
226 cpu_khz <<= -info->tsc_shift; 226 xen_khz <<= -info->tsc_shift;
227 else 227 else
228 cpu_khz >>= info->tsc_shift; 228 xen_khz >>= info->tsc_shift;
229 229
230 return cpu_khz; 230 return xen_khz;
231} 231}
232 232
233/* 233/*
diff --git a/arch/xtensa/Kconfig b/arch/xtensa/Kconfig
index fd36764d7fb7..9fc8551a1cf6 100644
--- a/arch/xtensa/Kconfig
+++ b/arch/xtensa/Kconfig
@@ -14,6 +14,7 @@ config ZONE_DMA
14config XTENSA 14config XTENSA
15 bool 15 bool
16 default y 16 default y
17 select HAVE_IDE
17 help 18 help
18 Xtensa processors are 32-bit RISC machines designed by Tensilica 19 Xtensa processors are 32-bit RISC machines designed by Tensilica
19 primarily for embedded systems. These processors are both 20 primarily for embedded systems. These processors are both
generated by cgit v1.2.2 (git 2.25.0) at 2024-12-26 09:21:24 -0500