aboutsummaryrefslogtreecommitdiffstats
path: root/arch/i386
diff options
context:
space:
mode:
authorSteven Whitehouse <swhiteho@redhat.com>2006-09-28 08:29:59 -0400
committerSteven Whitehouse <swhiteho@redhat.com>2006-09-28 08:29:59 -0400
commit185a257f2f73bcd89050ad02da5bedbc28fc43fa (patch)
tree5e32586114534ed3f2165614cba3d578f5d87307 /arch/i386
parent3f1a9aaeffd8d1cbc5ab9776c45cbd66af1c9699 (diff)
parenta77c64c1a641950626181b4857abb701d8f38ccc (diff)
Merge branch 'master' into gfs2
Diffstat (limited to 'arch/i386')
-rw-r--r--arch/i386/Kconfig35
-rw-r--r--arch/i386/Makefile8
-rw-r--r--arch/i386/boot/edd.S97
-rw-r--r--arch/i386/boot/setup.S4
-rw-r--r--arch/i386/defconfig1063
-rw-r--r--arch/i386/kernel/Makefile3
-rw-r--r--arch/i386/kernel/acpi/Makefile2
-rw-r--r--arch/i386/kernel/acpi/boot.c181
-rw-r--r--arch/i386/kernel/acpi/earlyquirk.c6
-rw-r--r--arch/i386/kernel/apic.c31
-rw-r--r--arch/i386/kernel/apm.c26
-rw-r--r--arch/i386/kernel/cpu/amd.c7
-rw-r--r--arch/i386/kernel/cpu/centaur.c24
-rw-r--r--arch/i386/kernel/cpu/common.c8
-rw-r--r--arch/i386/kernel/cpu/cpu.h2
-rw-r--r--arch/i386/kernel/cpu/cyrix.c42
-rw-r--r--arch/i386/kernel/cpu/intel.c3
-rw-r--r--arch/i386/kernel/cpu/mcheck/Makefile2
-rw-r--r--arch/i386/kernel/cpu/mcheck/p4.c26
-rw-r--r--arch/i386/kernel/cpu/mcheck/therm_throt.c180
-rw-r--r--arch/i386/kernel/cpu/mtrr/generic.c4
-rw-r--r--arch/i386/kernel/cpu/nexgen.c9
-rw-r--r--arch/i386/kernel/cpu/proc.c4
-rw-r--r--arch/i386/kernel/cpu/rise.c4
-rw-r--r--arch/i386/kernel/cpu/transmeta.c7
-rw-r--r--arch/i386/kernel/cpu/umc.c7
-rw-r--r--arch/i386/kernel/crash.c22
-rw-r--r--arch/i386/kernel/efi_stub.S1
-rw-r--r--arch/i386/kernel/entry.S110
-rw-r--r--arch/i386/kernel/head.S67
-rw-r--r--arch/i386/kernel/i8259.c6
-rw-r--r--arch/i386/kernel/io_apic.c125
-rw-r--r--arch/i386/kernel/machine_kexec.c140
-rw-r--r--arch/i386/kernel/mca.c8
-rw-r--r--arch/i386/kernel/microcode.c774
-rw-r--r--arch/i386/kernel/mpparse.c70
-rw-r--r--arch/i386/kernel/nmi.c940
-rw-r--r--arch/i386/kernel/process.c14
-rw-r--r--arch/i386/kernel/ptrace.c10
-rw-r--r--arch/i386/kernel/reboot.c12
-rw-r--r--arch/i386/kernel/relocate_kernel.S162
-rw-r--r--arch/i386/kernel/semaphore.c134
-rw-r--r--arch/i386/kernel/setup.c408
-rw-r--r--arch/i386/kernel/smp.c66
-rw-r--r--arch/i386/kernel/smpboot.c25
-rw-r--r--arch/i386/kernel/srat.c102
-rw-r--r--arch/i386/kernel/stacktrace.c98
-rw-r--r--arch/i386/kernel/syscall_table.S1
-rw-r--r--arch/i386/kernel/time.c73
-rw-r--r--arch/i386/kernel/time_hpet.c37
-rw-r--r--arch/i386/kernel/topology.c21
-rw-r--r--arch/i386/kernel/traps.c242
-rw-r--r--arch/i386/kernel/tsc.c2
-rw-r--r--arch/i386/kernel/vmlinux.lds.S12
-rw-r--r--arch/i386/lib/Makefile2
-rw-r--r--arch/i386/lib/semaphore.S217
-rw-r--r--arch/i386/mach-generic/bigsmp.c1
-rw-r--r--arch/i386/mach-generic/es7000.c1
-rw-r--r--arch/i386/mach-generic/probe.c60
-rw-r--r--arch/i386/mach-generic/summit.c1
-rw-r--r--arch/i386/mach-voyager/voyager_thread.c1
-rw-r--r--arch/i386/mm/boot_ioremap.c7
-rw-r--r--arch/i386/mm/discontig.c105
-rw-r--r--arch/i386/mm/extable.c2
-rw-r--r--arch/i386/mm/fault.c25
-rw-r--r--arch/i386/mm/highmem.c2
-rw-r--r--arch/i386/mm/init.c82
-rw-r--r--arch/i386/mm/pgtable.c30
-rw-r--r--arch/i386/oprofile/nmi_int.c88
-rw-r--r--arch/i386/oprofile/nmi_timer_int.c35
-rw-r--r--arch/i386/oprofile/op_model_athlon.c54
-rw-r--r--arch/i386/oprofile/op_model_p4.c152
-rw-r--r--arch/i386/oprofile/op_model_ppro.c65
-rw-r--r--arch/i386/oprofile/op_x86_model.h1
-rw-r--r--arch/i386/pci/Makefile2
-rw-r--r--arch/i386/pci/common.c4
-rw-r--r--arch/i386/pci/direct.c25
-rw-r--r--arch/i386/pci/early.c52
-rw-r--r--arch/i386/pci/init.c9
-rw-r--r--arch/i386/pci/mmconfig.c41
-rw-r--r--arch/i386/pci/pci.h7
-rw-r--r--arch/i386/power/swsusp.S2
82 files changed, 3966 insertions, 2574 deletions
diff --git a/arch/i386/Kconfig b/arch/i386/Kconfig
index b2751eadbc56..3fd2f256f2be 100644
--- a/arch/i386/Kconfig
+++ b/arch/i386/Kconfig
@@ -166,7 +166,6 @@ config X86_VISWS
166 166
167config X86_GENERICARCH 167config X86_GENERICARCH
168 bool "Generic architecture (Summit, bigsmp, ES7000, default)" 168 bool "Generic architecture (Summit, bigsmp, ES7000, default)"
169 depends on SMP
170 help 169 help
171 This option compiles in the Summit, bigsmp, ES7000, default subarchitectures. 170 This option compiles in the Summit, bigsmp, ES7000, default subarchitectures.
172 It is intended for a generic binary kernel. 171 It is intended for a generic binary kernel.
@@ -263,7 +262,7 @@ source "kernel/Kconfig.preempt"
263 262
264config X86_UP_APIC 263config X86_UP_APIC
265 bool "Local APIC support on uniprocessors" 264 bool "Local APIC support on uniprocessors"
266 depends on !SMP && !(X86_VISWS || X86_VOYAGER) 265 depends on !SMP && !(X86_VISWS || X86_VOYAGER || X86_GENERICARCH)
267 help 266 help
268 A local APIC (Advanced Programmable Interrupt Controller) is an 267 A local APIC (Advanced Programmable Interrupt Controller) is an
269 integrated interrupt controller in the CPU. If you have a single-CPU 268 integrated interrupt controller in the CPU. If you have a single-CPU
@@ -288,12 +287,12 @@ config X86_UP_IOAPIC
288 287
289config X86_LOCAL_APIC 288config X86_LOCAL_APIC
290 bool 289 bool
291 depends on X86_UP_APIC || ((X86_VISWS || SMP) && !X86_VOYAGER) 290 depends on X86_UP_APIC || ((X86_VISWS || SMP) && !X86_VOYAGER) || X86_GENERICARCH
292 default y 291 default y
293 292
294config X86_IO_APIC 293config X86_IO_APIC
295 bool 294 bool
296 depends on X86_UP_IOAPIC || (SMP && !(X86_VISWS || X86_VOYAGER)) 295 depends on X86_UP_IOAPIC || (SMP && !(X86_VISWS || X86_VOYAGER)) || X86_GENERICARCH
297 default y 296 default y
298 297
299config X86_VISWS_APIC 298config X86_VISWS_APIC
@@ -402,6 +401,7 @@ config X86_REBOOTFIXUPS
402 401
403config MICROCODE 402config MICROCODE
404 tristate "/dev/cpu/microcode - Intel IA32 CPU microcode support" 403 tristate "/dev/cpu/microcode - Intel IA32 CPU microcode support"
404 select FW_LOADER
405 ---help--- 405 ---help---
406 If you say Y here and also to "/dev file system support" in the 406 If you say Y here and also to "/dev file system support" in the
407 'File systems' section, you will be able to update the microcode on 407 'File systems' section, you will be able to update the microcode on
@@ -417,6 +417,11 @@ config MICROCODE
417 To compile this driver as a module, choose M here: the 417 To compile this driver as a module, choose M here: the
418 module will be called microcode. 418 module will be called microcode.
419 419
420config MICROCODE_OLD_INTERFACE
421 bool
422 depends on MICROCODE
423 default y
424
420config X86_MSR 425config X86_MSR
421 tristate "/dev/cpu/*/msr - Model-specific register support" 426 tristate "/dev/cpu/*/msr - Model-specific register support"
422 help 427 help
@@ -494,7 +499,7 @@ config HIGHMEM64G
494endchoice 499endchoice
495 500
496choice 501choice
497 depends on EXPERIMENTAL && !X86_PAE 502 depends on EXPERIMENTAL
498 prompt "Memory split" if EMBEDDED 503 prompt "Memory split" if EMBEDDED
499 default VMSPLIT_3G 504 default VMSPLIT_3G
500 help 505 help
@@ -516,6 +521,7 @@ choice
516 config VMSPLIT_3G 521 config VMSPLIT_3G
517 bool "3G/1G user/kernel split" 522 bool "3G/1G user/kernel split"
518 config VMSPLIT_3G_OPT 523 config VMSPLIT_3G_OPT
524 depends on !HIGHMEM
519 bool "3G/1G user/kernel split (for full 1G low memory)" 525 bool "3G/1G user/kernel split (for full 1G low memory)"
520 config VMSPLIT_2G 526 config VMSPLIT_2G
521 bool "2G/2G user/kernel split" 527 bool "2G/2G user/kernel split"
@@ -598,12 +604,10 @@ config ARCH_SELECT_MEMORY_MODEL
598 def_bool y 604 def_bool y
599 depends on ARCH_SPARSEMEM_ENABLE 605 depends on ARCH_SPARSEMEM_ENABLE
600 606
601source "mm/Kconfig" 607config ARCH_POPULATES_NODE_MAP
608 def_bool y
602 609
603config HAVE_ARCH_EARLY_PFN_TO_NID 610source "mm/Kconfig"
604 bool
605 default y
606 depends on NUMA
607 611
608config HIGHPTE 612config HIGHPTE
609 bool "Allocate 3rd-level pagetables from highmem" 613 bool "Allocate 3rd-level pagetables from highmem"
@@ -740,8 +744,7 @@ config SECCOMP
740source kernel/Kconfig.hz 744source kernel/Kconfig.hz
741 745
742config KEXEC 746config KEXEC
743 bool "kexec system call (EXPERIMENTAL)" 747 bool "kexec system call"
744 depends on EXPERIMENTAL
745 help 748 help
746 kexec is a system call that implements the ability to shutdown your 749 kexec is a system call that implements the ability to shutdown your
747 current kernel, and to start another kernel. It is like a reboot 750 current kernel, and to start another kernel. It is like a reboot
@@ -762,6 +765,13 @@ config CRASH_DUMP
762 depends on HIGHMEM 765 depends on HIGHMEM
763 help 766 help
764 Generate crash dump after being started by kexec. 767 Generate crash dump after being started by kexec.
768 This should be normally only set in special crash dump kernels
769 which are loaded in the main kernel with kexec-tools into
770 a specially reserved region and then later executed after
771 a crash by kdump/kexec. The crash dump kernel must be compiled
772 to a memory address not used by the main kernel or BIOS using
773 PHYSICAL_START.
774 For more details see Documentation/kdump/kdump.txt
765 775
766config PHYSICAL_START 776config PHYSICAL_START
767 hex "Physical address where the kernel is loaded" if (EMBEDDED || CRASH_DUMP) 777 hex "Physical address where the kernel is loaded" if (EMBEDDED || CRASH_DUMP)
@@ -794,6 +804,7 @@ config HOTPLUG_CPU
794config COMPAT_VDSO 804config COMPAT_VDSO
795 bool "Compat VDSO support" 805 bool "Compat VDSO support"
796 default y 806 default y
807 depends on !PARAVIRT
797 help 808 help
798 Map the VDSO to the predictable old-style address too. 809 Map the VDSO to the predictable old-style address too.
799 ---help--- 810 ---help---
diff --git a/arch/i386/Makefile b/arch/i386/Makefile
index 3e4adb1e2244..7cc0b189b82b 100644
--- a/arch/i386/Makefile
+++ b/arch/i386/Makefile
@@ -46,6 +46,14 @@ cflags-y += -ffreestanding
46# a lot more stack due to the lack of sharing of stacklots: 46# a lot more stack due to the lack of sharing of stacklots:
47CFLAGS += $(shell if [ $(call cc-version) -lt 0400 ] ; then echo $(call cc-option,-fno-unit-at-a-time); fi ;) 47CFLAGS += $(shell if [ $(call cc-version) -lt 0400 ] ; then echo $(call cc-option,-fno-unit-at-a-time); fi ;)
48 48
49# do binutils support CFI?
50cflags-y += $(call as-instr,.cfi_startproc\n.cfi_endproc,-DCONFIG_AS_CFI=1,)
51AFLAGS += $(call as-instr,.cfi_startproc\n.cfi_endproc,-DCONFIG_AS_CFI=1,)
52
53# is .cfi_signal_frame supported too?
54cflags-y += $(call as-instr,.cfi_startproc\n.cfi_endproc,-DCONFIG_AS_CFI=1,)
55AFLAGS += $(call as-instr,.cfi_startproc\n.cfi_endproc,-DCONFIG_AS_CFI=1,)
56
49CFLAGS += $(cflags-y) 57CFLAGS += $(cflags-y)
50 58
51# Default subarch .c files 59# Default subarch .c files
diff --git a/arch/i386/boot/edd.S b/arch/i386/boot/edd.S
index 4b84ea216f2b..34321368011a 100644
--- a/arch/i386/boot/edd.S
+++ b/arch/i386/boot/edd.S
@@ -15,42 +15,95 @@
15#include <asm/setup.h> 15#include <asm/setup.h>
16 16
17#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE) 17#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
18
19# It is assumed that %ds == INITSEG here
20
18 movb $0, (EDD_MBR_SIG_NR_BUF) 21 movb $0, (EDD_MBR_SIG_NR_BUF)
19 movb $0, (EDDNR) 22 movb $0, (EDDNR)
20 23
21# Check the command line for two options: 24# Check the command line for options:
22# edd=of disables EDD completely (edd=off) 25# edd=of disables EDD completely (edd=off)
23# edd=sk skips the MBR test (edd=skipmbr) 26# edd=sk skips the MBR test (edd=skipmbr)
27# edd=on re-enables EDD (edd=on)
28
24 pushl %esi 29 pushl %esi
25 cmpl $0, %cs:cmd_line_ptr 30 movw $edd_mbr_sig_start, %di # Default to edd=on
26 jz done_cl 31
27 movl %cs:(cmd_line_ptr), %esi 32 movl %cs:(cmd_line_ptr), %esi
28# ds:esi has the pointer to the command line now 33 andl %esi, %esi
29 movl $(COMMAND_LINE_SIZE-7), %ecx 34 jz old_cl # Old boot protocol?
30# loop through kernel command line one byte at a time 35
31cl_loop: 36# Convert to a real-mode pointer in fs:si
32 cmpl $EDD_CL_EQUALS, (%si) 37 movl %esi, %eax
38 shrl $4, %eax
39 movw %ax, %fs
40 andw $0xf, %si
41 jmp have_cl_pointer
42
43# Old-style boot protocol?
44old_cl:
45 push %ds # aka INITSEG
46 pop %fs
47
48 cmpw $0xa33f, (0x20)
49 jne done_cl # No command line at all?
50 movw (0x22), %si # Pointer relative to INITSEG
51
52# fs:si has the pointer to the command line now
53have_cl_pointer:
54
55# Loop through kernel command line one byte at a time. Just in
56# case the loader is buggy and failed to null-terminate the command line
57# terminate if we get close enough to the end of the segment that we
58# cannot fit "edd=XX"...
59cl_atspace:
60 cmpw $-5, %si # Watch for segment wraparound
61 jae done_cl
62 movl %fs:(%si), %eax
63 andb %al, %al # End of line?
64 jz done_cl
65 cmpl $EDD_CL_EQUALS, %eax
33 jz found_edd_equals 66 jz found_edd_equals
34 incl %esi 67 cmpb $0x20, %al # <= space consider whitespace
35 loop cl_loop 68 ja cl_skipword
36 jmp done_cl 69 incw %si
70 jmp cl_atspace
71
72cl_skipword:
73 cmpw $-5, %si # Watch for segment wraparound
74 jae done_cl
75 movb %fs:(%si), %al # End of string?
76 andb %al, %al
77 jz done_cl
78 cmpb $0x20, %al
79 jbe cl_atspace
80 incw %si
81 jmp cl_skipword
82
37found_edd_equals: 83found_edd_equals:
38# only looking at first two characters after equals 84# only looking at first two characters after equals
39 addl $4, %esi 85# late overrides early on the command line, so keep going after finding something
40 cmpw $EDD_CL_OFF, (%si) # edd=of 86 movw %fs:4(%si), %ax
41 jz do_edd_off 87 cmpw $EDD_CL_OFF, %ax # edd=of
42 cmpw $EDD_CL_SKIP, (%si) # edd=sk 88 je do_edd_off
43 jz do_edd_skipmbr 89 cmpw $EDD_CL_SKIP, %ax # edd=sk
44 jmp done_cl 90 je do_edd_skipmbr
91 cmpw $EDD_CL_ON, %ax # edd=on
92 je do_edd_on
93 jmp cl_skipword
45do_edd_skipmbr: 94do_edd_skipmbr:
46 popl %esi 95 movw $edd_start, %di
47 jmp edd_start 96 jmp cl_skipword
48do_edd_off: 97do_edd_off:
49 popl %esi 98 movw $edd_done, %di
50 jmp edd_done 99 jmp cl_skipword
100do_edd_on:
101 movw $edd_mbr_sig_start, %di
102 jmp cl_skipword
103
51done_cl: 104done_cl:
52 popl %esi 105 popl %esi
53 106 jmpw *%di
54 107
55# Read the first sector of each BIOS disk device and store the 4-byte signature 108# Read the first sector of each BIOS disk device and store the 4-byte signature
56edd_mbr_sig_start: 109edd_mbr_sig_start:
diff --git a/arch/i386/boot/setup.S b/arch/i386/boot/setup.S
index d2b684cd620a..3aec4538a113 100644
--- a/arch/i386/boot/setup.S
+++ b/arch/i386/boot/setup.S
@@ -494,12 +494,12 @@ no_voyager:
494 movw %cs, %ax # aka SETUPSEG 494 movw %cs, %ax # aka SETUPSEG
495 subw $DELTA_INITSEG, %ax # aka INITSEG 495 subw $DELTA_INITSEG, %ax # aka INITSEG
496 movw %ax, %ds 496 movw %ax, %ds
497 movw $0, (0x1ff) # default is no pointing device 497 movb $0, (0x1ff) # default is no pointing device
498 int $0x11 # int 0x11: equipment list 498 int $0x11 # int 0x11: equipment list
499 testb $0x04, %al # check if mouse installed 499 testb $0x04, %al # check if mouse installed
500 jz no_psmouse 500 jz no_psmouse
501 501
502 movw $0xAA, (0x1ff) # device present 502 movb $0xAA, (0x1ff) # device present
503no_psmouse: 503no_psmouse:
504 504
505#if defined(CONFIG_X86_SPEEDSTEP_SMI) || defined(CONFIG_X86_SPEEDSTEP_SMI_MODULE) 505#if defined(CONFIG_X86_SPEEDSTEP_SMI) || defined(CONFIG_X86_SPEEDSTEP_SMI_MODULE)
diff --git a/arch/i386/defconfig b/arch/i386/defconfig
index 89ebb7a316ab..1a29bfa26d0c 100644
--- a/arch/i386/defconfig
+++ b/arch/i386/defconfig
@@ -1,41 +1,51 @@
1# 1#
2# Automatically generated make config: don't edit 2# Automatically generated make config: don't edit
3# Linux kernel version: 2.6.18-git5
4# Tue Sep 26 09:30:47 2006
3# 5#
4CONFIG_X86_32=y 6CONFIG_X86_32=y
7CONFIG_GENERIC_TIME=y
8CONFIG_LOCKDEP_SUPPORT=y
9CONFIG_STACKTRACE_SUPPORT=y
5CONFIG_SEMAPHORE_SLEEPERS=y 10CONFIG_SEMAPHORE_SLEEPERS=y
6CONFIG_X86=y 11CONFIG_X86=y
7CONFIG_MMU=y 12CONFIG_MMU=y
8CONFIG_GENERIC_ISA_DMA=y 13CONFIG_GENERIC_ISA_DMA=y
9CONFIG_GENERIC_IOMAP=y 14CONFIG_GENERIC_IOMAP=y
15CONFIG_GENERIC_HWEIGHT=y
10CONFIG_ARCH_MAY_HAVE_PC_FDC=y 16CONFIG_ARCH_MAY_HAVE_PC_FDC=y
11CONFIG_DMI=y 17CONFIG_DMI=y
18CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
12 19
13# 20#
14# Code maturity level options 21# Code maturity level options
15# 22#
16CONFIG_EXPERIMENTAL=y 23CONFIG_EXPERIMENTAL=y
17CONFIG_BROKEN_ON_SMP=y 24CONFIG_LOCK_KERNEL=y
18CONFIG_INIT_ENV_ARG_LIMIT=32 25CONFIG_INIT_ENV_ARG_LIMIT=32
19 26
20# 27#
21# General setup 28# General setup
22# 29#
23CONFIG_LOCALVERSION="" 30CONFIG_LOCALVERSION=""
24# CONFIG_LOCALVERSION_AUTO is not set 31CONFIG_LOCALVERSION_AUTO=y
25CONFIG_SWAP=y 32CONFIG_SWAP=y
26CONFIG_SYSVIPC=y 33CONFIG_SYSVIPC=y
27# CONFIG_POSIX_MQUEUE is not set 34CONFIG_POSIX_MQUEUE=y
28# CONFIG_BSD_PROCESS_ACCT is not set 35# CONFIG_BSD_PROCESS_ACCT is not set
29CONFIG_SYSCTL=y 36# CONFIG_TASKSTATS is not set
30# CONFIG_AUDIT is not set 37# CONFIG_AUDIT is not set
31CONFIG_IKCONFIG=y 38CONFIG_IKCONFIG=y
32CONFIG_IKCONFIG_PROC=y 39CONFIG_IKCONFIG_PROC=y
40# CONFIG_CPUSETS is not set
41# CONFIG_RELAY is not set
33CONFIG_INITRAMFS_SOURCE="" 42CONFIG_INITRAMFS_SOURCE=""
34CONFIG_UID16=y
35CONFIG_VM86=y
36CONFIG_CC_OPTIMIZE_FOR_SIZE=y 43CONFIG_CC_OPTIMIZE_FOR_SIZE=y
37# CONFIG_EMBEDDED is not set 44# CONFIG_EMBEDDED is not set
45CONFIG_UID16=y
46CONFIG_SYSCTL=y
38CONFIG_KALLSYMS=y 47CONFIG_KALLSYMS=y
48CONFIG_KALLSYMS_ALL=y
39# CONFIG_KALLSYMS_EXTRA_PASS is not set 49# CONFIG_KALLSYMS_EXTRA_PASS is not set
40CONFIG_HOTPLUG=y 50CONFIG_HOTPLUG=y
41CONFIG_PRINTK=y 51CONFIG_PRINTK=y
@@ -45,11 +55,9 @@ CONFIG_BASE_FULL=y
45CONFIG_FUTEX=y 55CONFIG_FUTEX=y
46CONFIG_EPOLL=y 56CONFIG_EPOLL=y
47CONFIG_SHMEM=y 57CONFIG_SHMEM=y
48CONFIG_CC_ALIGN_FUNCTIONS=0
49CONFIG_CC_ALIGN_LABELS=0
50CONFIG_CC_ALIGN_LOOPS=0
51CONFIG_CC_ALIGN_JUMPS=0
52CONFIG_SLAB=y 58CONFIG_SLAB=y
59CONFIG_VM_EVENT_COUNTERS=y
60CONFIG_RT_MUTEXES=y
53# CONFIG_TINY_SHMEM is not set 61# CONFIG_TINY_SHMEM is not set
54CONFIG_BASE_SMALL=0 62CONFIG_BASE_SMALL=0
55# CONFIG_SLOB is not set 63# CONFIG_SLOB is not set
@@ -60,41 +68,45 @@ CONFIG_BASE_SMALL=0
60CONFIG_MODULES=y 68CONFIG_MODULES=y
61CONFIG_MODULE_UNLOAD=y 69CONFIG_MODULE_UNLOAD=y
62CONFIG_MODULE_FORCE_UNLOAD=y 70CONFIG_MODULE_FORCE_UNLOAD=y
63CONFIG_OBSOLETE_MODPARM=y
64# CONFIG_MODVERSIONS is not set 71# CONFIG_MODVERSIONS is not set
65# CONFIG_MODULE_SRCVERSION_ALL is not set 72# CONFIG_MODULE_SRCVERSION_ALL is not set
66# CONFIG_KMOD is not set 73# CONFIG_KMOD is not set
74CONFIG_STOP_MACHINE=y
67 75
68# 76#
69# Block layer 77# Block layer
70# 78#
71# CONFIG_LBD is not set 79CONFIG_LBD=y
80# CONFIG_BLK_DEV_IO_TRACE is not set
81# CONFIG_LSF is not set
72 82
73# 83#
74# IO Schedulers 84# IO Schedulers
75# 85#
76CONFIG_IOSCHED_NOOP=y 86CONFIG_IOSCHED_NOOP=y
77# CONFIG_IOSCHED_AS is not set 87CONFIG_IOSCHED_AS=y
78# CONFIG_IOSCHED_DEADLINE is not set 88CONFIG_IOSCHED_DEADLINE=y
79CONFIG_IOSCHED_CFQ=y 89CONFIG_IOSCHED_CFQ=y
80# CONFIG_DEFAULT_AS is not set 90CONFIG_DEFAULT_AS=y
81# CONFIG_DEFAULT_DEADLINE is not set 91# CONFIG_DEFAULT_DEADLINE is not set
82CONFIG_DEFAULT_CFQ=y 92# CONFIG_DEFAULT_CFQ is not set
83# CONFIG_DEFAULT_NOOP is not set 93# CONFIG_DEFAULT_NOOP is not set
84CONFIG_DEFAULT_IOSCHED="cfq" 94CONFIG_DEFAULT_IOSCHED="anticipatory"
85 95
86# 96#
87# Processor type and features 97# Processor type and features
88# 98#
89CONFIG_X86_PC=y 99CONFIG_SMP=y
100# CONFIG_X86_PC is not set
90# CONFIG_X86_ELAN is not set 101# CONFIG_X86_ELAN is not set
91# CONFIG_X86_VOYAGER is not set 102# CONFIG_X86_VOYAGER is not set
92# CONFIG_X86_NUMAQ is not set 103# CONFIG_X86_NUMAQ is not set
93# CONFIG_X86_SUMMIT is not set 104# CONFIG_X86_SUMMIT is not set
94# CONFIG_X86_BIGSMP is not set 105# CONFIG_X86_BIGSMP is not set
95# CONFIG_X86_VISWS is not set 106# CONFIG_X86_VISWS is not set
96# CONFIG_X86_GENERICARCH is not set 107CONFIG_X86_GENERICARCH=y
97# CONFIG_X86_ES7000 is not set 108# CONFIG_X86_ES7000 is not set
109CONFIG_X86_CYCLONE_TIMER=y
98# CONFIG_M386 is not set 110# CONFIG_M386 is not set
99# CONFIG_M486 is not set 111# CONFIG_M486 is not set
100# CONFIG_M586 is not set 112# CONFIG_M586 is not set
@@ -102,11 +114,11 @@ CONFIG_X86_PC=y
102# CONFIG_M586MMX is not set 114# CONFIG_M586MMX is not set
103# CONFIG_M686 is not set 115# CONFIG_M686 is not set
104# CONFIG_MPENTIUMII is not set 116# CONFIG_MPENTIUMII is not set
105# CONFIG_MPENTIUMIII is not set 117CONFIG_MPENTIUMIII=y
106# CONFIG_MPENTIUMM is not set 118# CONFIG_MPENTIUMM is not set
107# CONFIG_MPENTIUM4 is not set 119# CONFIG_MPENTIUM4 is not set
108# CONFIG_MK6 is not set 120# CONFIG_MK6 is not set
109CONFIG_MK7=y 121# CONFIG_MK7 is not set
110# CONFIG_MK8 is not set 122# CONFIG_MK8 is not set
111# CONFIG_MCRUSOE is not set 123# CONFIG_MCRUSOE is not set
112# CONFIG_MEFFICEON is not set 124# CONFIG_MEFFICEON is not set
@@ -117,10 +129,10 @@ CONFIG_MK7=y
117# CONFIG_MGEODE_LX is not set 129# CONFIG_MGEODE_LX is not set
118# CONFIG_MCYRIXIII is not set 130# CONFIG_MCYRIXIII is not set
119# CONFIG_MVIAC3_2 is not set 131# CONFIG_MVIAC3_2 is not set
120# CONFIG_X86_GENERIC is not set 132CONFIG_X86_GENERIC=y
121CONFIG_X86_CMPXCHG=y 133CONFIG_X86_CMPXCHG=y
122CONFIG_X86_XADD=y 134CONFIG_X86_XADD=y
123CONFIG_X86_L1_CACHE_SHIFT=6 135CONFIG_X86_L1_CACHE_SHIFT=7
124CONFIG_RWSEM_XCHGADD_ALGORITHM=y 136CONFIG_RWSEM_XCHGADD_ALGORITHM=y
125CONFIG_GENERIC_CALIBRATE_DELAY=y 137CONFIG_GENERIC_CALIBRATE_DELAY=y
126CONFIG_X86_WP_WORKS_OK=y 138CONFIG_X86_WP_WORKS_OK=y
@@ -131,26 +143,28 @@ CONFIG_X86_CMPXCHG64=y
131CONFIG_X86_GOOD_APIC=y 143CONFIG_X86_GOOD_APIC=y
132CONFIG_X86_INTEL_USERCOPY=y 144CONFIG_X86_INTEL_USERCOPY=y
133CONFIG_X86_USE_PPRO_CHECKSUM=y 145CONFIG_X86_USE_PPRO_CHECKSUM=y
134CONFIG_X86_USE_3DNOW=y
135CONFIG_X86_TSC=y 146CONFIG_X86_TSC=y
136# CONFIG_HPET_TIMER is not set 147CONFIG_HPET_TIMER=y
137# CONFIG_SMP is not set 148CONFIG_HPET_EMULATE_RTC=y
138CONFIG_PREEMPT_NONE=y 149CONFIG_NR_CPUS=32
139# CONFIG_PREEMPT_VOLUNTARY is not set 150CONFIG_SCHED_SMT=y
151CONFIG_SCHED_MC=y
152# CONFIG_PREEMPT_NONE is not set
153CONFIG_PREEMPT_VOLUNTARY=y
140# CONFIG_PREEMPT is not set 154# CONFIG_PREEMPT is not set
141CONFIG_X86_UP_APIC=y 155CONFIG_PREEMPT_BKL=y
142CONFIG_X86_UP_IOAPIC=y
143CONFIG_X86_LOCAL_APIC=y 156CONFIG_X86_LOCAL_APIC=y
144CONFIG_X86_IO_APIC=y 157CONFIG_X86_IO_APIC=y
145CONFIG_X86_MCE=y 158CONFIG_X86_MCE=y
146CONFIG_X86_MCE_NONFATAL=y 159CONFIG_X86_MCE_NONFATAL=y
147# CONFIG_X86_MCE_P4THERMAL is not set 160CONFIG_X86_MCE_P4THERMAL=y
161CONFIG_VM86=y
148# CONFIG_TOSHIBA is not set 162# CONFIG_TOSHIBA is not set
149# CONFIG_I8K is not set 163# CONFIG_I8K is not set
150# CONFIG_X86_REBOOTFIXUPS is not set 164# CONFIG_X86_REBOOTFIXUPS is not set
151# CONFIG_MICROCODE is not set 165CONFIG_MICROCODE=y
152# CONFIG_X86_MSR is not set 166CONFIG_X86_MSR=y
153# CONFIG_X86_CPUID is not set 167CONFIG_X86_CPUID=y
154 168
155# 169#
156# Firmware Drivers 170# Firmware Drivers
@@ -158,68 +172,67 @@ CONFIG_X86_MCE_NONFATAL=y
158# CONFIG_EDD is not set 172# CONFIG_EDD is not set
159# CONFIG_DELL_RBU is not set 173# CONFIG_DELL_RBU is not set
160# CONFIG_DCDBAS is not set 174# CONFIG_DCDBAS is not set
161CONFIG_NOHIGHMEM=y 175# CONFIG_NOHIGHMEM is not set
162# CONFIG_HIGHMEM4G is not set 176CONFIG_HIGHMEM4G=y
163# CONFIG_HIGHMEM64G is not set 177# CONFIG_HIGHMEM64G is not set
164CONFIG_VMSPLIT_3G=y
165# CONFIG_VMSPLIT_3G_OPT is not set
166# CONFIG_VMSPLIT_2G is not set
167# CONFIG_VMSPLIT_1G is not set
168CONFIG_PAGE_OFFSET=0xC0000000 178CONFIG_PAGE_OFFSET=0xC0000000
169CONFIG_ARCH_FLATMEM_ENABLE=y 179CONFIG_HIGHMEM=y
170CONFIG_ARCH_SPARSEMEM_ENABLE=y
171CONFIG_ARCH_SELECT_MEMORY_MODEL=y
172CONFIG_SELECT_MEMORY_MODEL=y 180CONFIG_SELECT_MEMORY_MODEL=y
173CONFIG_FLATMEM_MANUAL=y 181CONFIG_FLATMEM_MANUAL=y
174# CONFIG_DISCONTIGMEM_MANUAL is not set 182# CONFIG_DISCONTIGMEM_MANUAL is not set
175# CONFIG_SPARSEMEM_MANUAL is not set 183# CONFIG_SPARSEMEM_MANUAL is not set
176CONFIG_FLATMEM=y 184CONFIG_FLATMEM=y
177CONFIG_FLAT_NODE_MEM_MAP=y 185CONFIG_FLAT_NODE_MEM_MAP=y
178CONFIG_SPARSEMEM_STATIC=y 186# CONFIG_SPARSEMEM_STATIC is not set
179CONFIG_SPLIT_PTLOCK_CPUS=4 187CONFIG_SPLIT_PTLOCK_CPUS=4
188CONFIG_RESOURCES_64BIT=y
189# CONFIG_HIGHPTE is not set
180# CONFIG_MATH_EMULATION is not set 190# CONFIG_MATH_EMULATION is not set
181CONFIG_MTRR=y 191CONFIG_MTRR=y
182# CONFIG_EFI is not set 192# CONFIG_EFI is not set
193# CONFIG_IRQBALANCE is not set
183CONFIG_REGPARM=y 194CONFIG_REGPARM=y
184# CONFIG_SECCOMP is not set 195CONFIG_SECCOMP=y
185CONFIG_HZ_100=y 196# CONFIG_HZ_100 is not set
186# CONFIG_HZ_250 is not set 197CONFIG_HZ_250=y
187# CONFIG_HZ_1000 is not set 198# CONFIG_HZ_1000 is not set
188CONFIG_HZ=100 199CONFIG_HZ=250
189# CONFIG_KEXEC is not set 200# CONFIG_KEXEC is not set
201# CONFIG_CRASH_DUMP is not set
190CONFIG_PHYSICAL_START=0x100000 202CONFIG_PHYSICAL_START=0x100000
191CONFIG_DOUBLEFAULT=y 203# CONFIG_HOTPLUG_CPU is not set
204CONFIG_COMPAT_VDSO=y
205CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
192 206
193# 207#
194# Power management options (ACPI, APM) 208# Power management options (ACPI, APM)
195# 209#
196CONFIG_PM=y 210CONFIG_PM=y
197# CONFIG_PM_LEGACY is not set 211CONFIG_PM_LEGACY=y
198# CONFIG_PM_DEBUG is not set 212# CONFIG_PM_DEBUG is not set
199CONFIG_SOFTWARE_SUSPEND=y
200CONFIG_PM_STD_PARTITION=""
201 213
202# 214#
203# ACPI (Advanced Configuration and Power Interface) Support 215# ACPI (Advanced Configuration and Power Interface) Support
204# 216#
205CONFIG_ACPI=y 217CONFIG_ACPI=y
206# CONFIG_ACPI_SLEEP is not set 218CONFIG_ACPI_AC=y
207# CONFIG_ACPI_AC is not set 219CONFIG_ACPI_BATTERY=y
208# CONFIG_ACPI_BATTERY is not set 220CONFIG_ACPI_BUTTON=y
209# CONFIG_ACPI_BUTTON is not set
210# CONFIG_ACPI_VIDEO is not set 221# CONFIG_ACPI_VIDEO is not set
211# CONFIG_ACPI_HOTKEY is not set 222# CONFIG_ACPI_HOTKEY is not set
212# CONFIG_ACPI_FAN is not set 223CONFIG_ACPI_FAN=y
213# CONFIG_ACPI_PROCESSOR is not set 224# CONFIG_ACPI_DOCK is not set
225CONFIG_ACPI_PROCESSOR=y
226CONFIG_ACPI_THERMAL=y
214# CONFIG_ACPI_ASUS is not set 227# CONFIG_ACPI_ASUS is not set
215# CONFIG_ACPI_IBM is not set 228# CONFIG_ACPI_IBM is not set
216# CONFIG_ACPI_TOSHIBA is not set 229# CONFIG_ACPI_TOSHIBA is not set
217CONFIG_ACPI_BLACKLIST_YEAR=0 230CONFIG_ACPI_BLACKLIST_YEAR=2001
218# CONFIG_ACPI_DEBUG is not set 231CONFIG_ACPI_DEBUG=y
219CONFIG_ACPI_EC=y 232CONFIG_ACPI_EC=y
220CONFIG_ACPI_POWER=y 233CONFIG_ACPI_POWER=y
221CONFIG_ACPI_SYSTEM=y 234CONFIG_ACPI_SYSTEM=y
222# CONFIG_X86_PM_TIMER is not set 235CONFIG_X86_PM_TIMER=y
223# CONFIG_ACPI_CONTAINER is not set 236# CONFIG_ACPI_CONTAINER is not set
224 237
225# 238#
@@ -230,7 +243,41 @@ CONFIG_ACPI_SYSTEM=y
230# 243#
231# CPU Frequency scaling 244# CPU Frequency scaling
232# 245#
233# CONFIG_CPU_FREQ is not set 246CONFIG_CPU_FREQ=y
247CONFIG_CPU_FREQ_TABLE=y
248CONFIG_CPU_FREQ_DEBUG=y
249CONFIG_CPU_FREQ_STAT=y
250# CONFIG_CPU_FREQ_STAT_DETAILS is not set
251CONFIG_CPU_FREQ_DEFAULT_GOV_PERFORMANCE=y
252# CONFIG_CPU_FREQ_DEFAULT_GOV_USERSPACE is not set
253CONFIG_CPU_FREQ_GOV_PERFORMANCE=y
254# CONFIG_CPU_FREQ_GOV_POWERSAVE is not set
255CONFIG_CPU_FREQ_GOV_USERSPACE=y
256CONFIG_CPU_FREQ_GOV_ONDEMAND=y
257# CONFIG_CPU_FREQ_GOV_CONSERVATIVE is not set
258
259#
260# CPUFreq processor drivers
261#
262CONFIG_X86_ACPI_CPUFREQ=y
263# CONFIG_X86_POWERNOW_K6 is not set
264# CONFIG_X86_POWERNOW_K7 is not set
265CONFIG_X86_POWERNOW_K8=y
266CONFIG_X86_POWERNOW_K8_ACPI=y
267# CONFIG_X86_GX_SUSPMOD is not set
268# CONFIG_X86_SPEEDSTEP_CENTRINO is not set
269# CONFIG_X86_SPEEDSTEP_ICH is not set
270# CONFIG_X86_SPEEDSTEP_SMI is not set
271# CONFIG_X86_P4_CLOCKMOD is not set
272# CONFIG_X86_CPUFREQ_NFORCE2 is not set
273# CONFIG_X86_LONGRUN is not set
274# CONFIG_X86_LONGHAUL is not set
275
276#
277# shared options
278#
279CONFIG_X86_ACPI_CPUFREQ_PROC_INTF=y
280# CONFIG_X86_SPEEDSTEP_LIB is not set
234 281
235# 282#
236# Bus options (PCI, PCMCIA, EISA, MCA, ISA) 283# Bus options (PCI, PCMCIA, EISA, MCA, ISA)
@@ -244,12 +291,13 @@ CONFIG_PCI_BIOS=y
244CONFIG_PCI_DIRECT=y 291CONFIG_PCI_DIRECT=y
245CONFIG_PCI_MMCONFIG=y 292CONFIG_PCI_MMCONFIG=y
246# CONFIG_PCIEPORTBUS is not set 293# CONFIG_PCIEPORTBUS is not set
247# CONFIG_PCI_MSI is not set 294CONFIG_PCI_MSI=y
248# CONFIG_PCI_LEGACY_PROC is not set 295# CONFIG_PCI_DEBUG is not set
249CONFIG_ISA_DMA_API=y 296CONFIG_ISA_DMA_API=y
250# CONFIG_ISA is not set 297# CONFIG_ISA is not set
251# CONFIG_MCA is not set 298# CONFIG_MCA is not set
252# CONFIG_SCx200 is not set 299# CONFIG_SCx200 is not set
300CONFIG_K8_NB=y
253 301
254# 302#
255# PCCARD (PCMCIA/CardBus) support 303# PCCARD (PCMCIA/CardBus) support
@@ -278,93 +326,54 @@ CONFIG_NET=y
278# 326#
279# CONFIG_NETDEBUG is not set 327# CONFIG_NETDEBUG is not set
280CONFIG_PACKET=y 328CONFIG_PACKET=y
281CONFIG_PACKET_MMAP=y 329# CONFIG_PACKET_MMAP is not set
282CONFIG_UNIX=y 330CONFIG_UNIX=y
331CONFIG_XFRM=y
332# CONFIG_XFRM_USER is not set
333# CONFIG_XFRM_SUB_POLICY is not set
283# CONFIG_NET_KEY is not set 334# CONFIG_NET_KEY is not set
284CONFIG_INET=y 335CONFIG_INET=y
285# CONFIG_IP_MULTICAST is not set 336CONFIG_IP_MULTICAST=y
286# CONFIG_IP_ADVANCED_ROUTER is not set 337# CONFIG_IP_ADVANCED_ROUTER is not set
287CONFIG_IP_FIB_HASH=y 338CONFIG_IP_FIB_HASH=y
288# CONFIG_IP_PNP is not set 339CONFIG_IP_PNP=y
340CONFIG_IP_PNP_DHCP=y
341# CONFIG_IP_PNP_BOOTP is not set
342# CONFIG_IP_PNP_RARP is not set
289# CONFIG_NET_IPIP is not set 343# CONFIG_NET_IPIP is not set
290# CONFIG_NET_IPGRE is not set 344# CONFIG_NET_IPGRE is not set
345# CONFIG_IP_MROUTE is not set
291# CONFIG_ARPD is not set 346# CONFIG_ARPD is not set
292# CONFIG_SYN_COOKIES is not set 347# CONFIG_SYN_COOKIES is not set
293# CONFIG_INET_AH is not set 348# CONFIG_INET_AH is not set
294# CONFIG_INET_ESP is not set 349# CONFIG_INET_ESP is not set
295# CONFIG_INET_IPCOMP is not set 350# CONFIG_INET_IPCOMP is not set
351# CONFIG_INET_XFRM_TUNNEL is not set
296# CONFIG_INET_TUNNEL is not set 352# CONFIG_INET_TUNNEL is not set
297# CONFIG_INET_DIAG is not set 353CONFIG_INET_XFRM_MODE_TRANSPORT=y
354CONFIG_INET_XFRM_MODE_TUNNEL=y
355CONFIG_INET_DIAG=y
356CONFIG_INET_TCP_DIAG=y
298# CONFIG_TCP_CONG_ADVANCED is not set 357# CONFIG_TCP_CONG_ADVANCED is not set
299CONFIG_TCP_CONG_BIC=y 358CONFIG_TCP_CONG_CUBIC=y
300 359CONFIG_DEFAULT_TCP_CONG="cubic"
301# 360CONFIG_IPV6=y
302# IP: Virtual Server Configuration 361# CONFIG_IPV6_PRIVACY is not set
303# 362# CONFIG_IPV6_ROUTER_PREF is not set
304# CONFIG_IP_VS is not set 363# CONFIG_INET6_AH is not set
305# CONFIG_IPV6 is not set 364# CONFIG_INET6_ESP is not set
306CONFIG_NETFILTER=y 365# CONFIG_INET6_IPCOMP is not set
307# CONFIG_NETFILTER_DEBUG is not set 366# CONFIG_IPV6_MIP6 is not set
308 367# CONFIG_INET6_XFRM_TUNNEL is not set
309# 368# CONFIG_INET6_TUNNEL is not set
310# Core Netfilter Configuration 369CONFIG_INET6_XFRM_MODE_TRANSPORT=y
311# 370CONFIG_INET6_XFRM_MODE_TUNNEL=y
312# CONFIG_NETFILTER_NETLINK is not set 371# CONFIG_INET6_XFRM_MODE_ROUTEOPTIMIZATION is not set
313CONFIG_NETFILTER_XTABLES=y 372# CONFIG_IPV6_TUNNEL is not set
314# CONFIG_NETFILTER_XT_TARGET_CLASSIFY is not set 373# CONFIG_IPV6_SUBTREES is not set
315# CONFIG_NETFILTER_XT_TARGET_MARK is not set 374# CONFIG_IPV6_MULTIPLE_TABLES is not set
316# CONFIG_NETFILTER_XT_TARGET_NFQUEUE is not set 375# CONFIG_NETWORK_SECMARK is not set
317# CONFIG_NETFILTER_XT_MATCH_COMMENT is not set 376# CONFIG_NETFILTER is not set
318# CONFIG_NETFILTER_XT_MATCH_CONNTRACK is not set
319# CONFIG_NETFILTER_XT_MATCH_DCCP is not set
320# CONFIG_NETFILTER_XT_MATCH_HELPER is not set
321# CONFIG_NETFILTER_XT_MATCH_LENGTH is not set
322CONFIG_NETFILTER_XT_MATCH_LIMIT=y
323CONFIG_NETFILTER_XT_MATCH_MAC=y
324# CONFIG_NETFILTER_XT_MATCH_MARK is not set
325# CONFIG_NETFILTER_XT_MATCH_PKTTYPE is not set
326# CONFIG_NETFILTER_XT_MATCH_REALM is not set
327# CONFIG_NETFILTER_XT_MATCH_SCTP is not set
328CONFIG_NETFILTER_XT_MATCH_STATE=y
329# CONFIG_NETFILTER_XT_MATCH_STRING is not set
330# CONFIG_NETFILTER_XT_MATCH_TCPMSS is not set
331
332#
333# IP: Netfilter Configuration
334#
335CONFIG_IP_NF_CONNTRACK=y
336# CONFIG_IP_NF_CT_ACCT is not set
337# CONFIG_IP_NF_CONNTRACK_MARK is not set
338# CONFIG_IP_NF_CONNTRACK_EVENTS is not set
339# CONFIG_IP_NF_CT_PROTO_SCTP is not set
340CONFIG_IP_NF_FTP=y
341# CONFIG_IP_NF_IRC is not set
342# CONFIG_IP_NF_NETBIOS_NS is not set
343# CONFIG_IP_NF_TFTP is not set
344# CONFIG_IP_NF_AMANDA is not set
345# CONFIG_IP_NF_PPTP is not set
346# CONFIG_IP_NF_QUEUE is not set
347CONFIG_IP_NF_IPTABLES=y
348# CONFIG_IP_NF_MATCH_IPRANGE is not set
349# CONFIG_IP_NF_MATCH_MULTIPORT is not set
350# CONFIG_IP_NF_MATCH_TOS is not set
351# CONFIG_IP_NF_MATCH_RECENT is not set
352# CONFIG_IP_NF_MATCH_ECN is not set
353# CONFIG_IP_NF_MATCH_DSCP is not set
354# CONFIG_IP_NF_MATCH_AH_ESP is not set
355# CONFIG_IP_NF_MATCH_TTL is not set
356# CONFIG_IP_NF_MATCH_OWNER is not set
357# CONFIG_IP_NF_MATCH_ADDRTYPE is not set
358# CONFIG_IP_NF_MATCH_HASHLIMIT is not set
359CONFIG_IP_NF_FILTER=y
360# CONFIG_IP_NF_TARGET_REJECT is not set
361CONFIG_IP_NF_TARGET_LOG=y
362# CONFIG_IP_NF_TARGET_ULOG is not set
363# CONFIG_IP_NF_TARGET_TCPMSS is not set
364# CONFIG_IP_NF_NAT is not set
365# CONFIG_IP_NF_MANGLE is not set
366# CONFIG_IP_NF_RAW is not set
367# CONFIG_IP_NF_ARPTABLES is not set
368 377
369# 378#
370# DCCP Configuration (EXPERIMENTAL) 379# DCCP Configuration (EXPERIMENTAL)
@@ -389,7 +398,6 @@ CONFIG_IP_NF_TARGET_LOG=y
389# CONFIG_ATALK is not set 398# CONFIG_ATALK is not set
390# CONFIG_X25 is not set 399# CONFIG_X25 is not set
391# CONFIG_LAPB is not set 400# CONFIG_LAPB is not set
392# CONFIG_NET_DIVERT is not set
393# CONFIG_ECONET is not set 401# CONFIG_ECONET is not set
394# CONFIG_WAN_ROUTER is not set 402# CONFIG_WAN_ROUTER is not set
395 403
@@ -402,6 +410,7 @@ CONFIG_IP_NF_TARGET_LOG=y
402# Network testing 410# Network testing
403# 411#
404# CONFIG_NET_PKTGEN is not set 412# CONFIG_NET_PKTGEN is not set
413# CONFIG_NET_TCPPROBE is not set
405# CONFIG_HAMRADIO is not set 414# CONFIG_HAMRADIO is not set
406# CONFIG_IRDA is not set 415# CONFIG_IRDA is not set
407# CONFIG_BT is not set 416# CONFIG_BT is not set
@@ -416,7 +425,9 @@ CONFIG_IP_NF_TARGET_LOG=y
416# 425#
417CONFIG_STANDALONE=y 426CONFIG_STANDALONE=y
418CONFIG_PREVENT_FIRMWARE_BUILD=y 427CONFIG_PREVENT_FIRMWARE_BUILD=y
419# CONFIG_FW_LOADER is not set 428CONFIG_FW_LOADER=y
429# CONFIG_DEBUG_DRIVER is not set
430# CONFIG_SYS_HYPERVISOR is not set
420 431
421# 432#
422# Connector - unified userspace <-> kernelspace linker 433# Connector - unified userspace <-> kernelspace linker
@@ -431,13 +442,7 @@ CONFIG_PREVENT_FIRMWARE_BUILD=y
431# 442#
432# Parallel port support 443# Parallel port support
433# 444#
434CONFIG_PARPORT=y 445# CONFIG_PARPORT is not set
435CONFIG_PARPORT_PC=y
436# CONFIG_PARPORT_SERIAL is not set
437# CONFIG_PARPORT_PC_FIFO is not set
438# CONFIG_PARPORT_PC_SUPERIO is not set
439# CONFIG_PARPORT_GSC is not set
440CONFIG_PARPORT_1284=y
441 446
442# 447#
443# Plug and Play support 448# Plug and Play support
@@ -447,8 +452,7 @@ CONFIG_PARPORT_1284=y
447# 452#
448# Block devices 453# Block devices
449# 454#
450# CONFIG_BLK_DEV_FD is not set 455CONFIG_BLK_DEV_FD=y
451# CONFIG_PARIDE is not set
452# CONFIG_BLK_CPQ_DA is not set 456# CONFIG_BLK_CPQ_DA is not set
453# CONFIG_BLK_CPQ_CISS_DA is not set 457# CONFIG_BLK_CPQ_CISS_DA is not set
454# CONFIG_BLK_DEV_DAC960 is not set 458# CONFIG_BLK_DEV_DAC960 is not set
@@ -459,8 +463,11 @@ CONFIG_BLK_DEV_LOOP=y
459# CONFIG_BLK_DEV_NBD is not set 463# CONFIG_BLK_DEV_NBD is not set
460# CONFIG_BLK_DEV_SX8 is not set 464# CONFIG_BLK_DEV_SX8 is not set
461# CONFIG_BLK_DEV_UB is not set 465# CONFIG_BLK_DEV_UB is not set
462# CONFIG_BLK_DEV_RAM is not set 466CONFIG_BLK_DEV_RAM=y
463CONFIG_BLK_DEV_RAM_COUNT=16 467CONFIG_BLK_DEV_RAM_COUNT=16
468CONFIG_BLK_DEV_RAM_SIZE=4096
469CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
470CONFIG_BLK_DEV_INITRD=y
464# CONFIG_CDROM_PKTCDVD is not set 471# CONFIG_CDROM_PKTCDVD is not set
465# CONFIG_ATA_OVER_ETH is not set 472# CONFIG_ATA_OVER_ETH is not set
466 473
@@ -476,7 +483,7 @@ CONFIG_BLK_DEV_IDE=y
476# CONFIG_BLK_DEV_IDE_SATA is not set 483# CONFIG_BLK_DEV_IDE_SATA is not set
477# CONFIG_BLK_DEV_HD_IDE is not set 484# CONFIG_BLK_DEV_HD_IDE is not set
478CONFIG_BLK_DEV_IDEDISK=y 485CONFIG_BLK_DEV_IDEDISK=y
479# CONFIG_IDEDISK_MULTI_MODE is not set 486CONFIG_IDEDISK_MULTI_MODE=y
480CONFIG_BLK_DEV_IDECD=y 487CONFIG_BLK_DEV_IDECD=y
481# CONFIG_BLK_DEV_IDETAPE is not set 488# CONFIG_BLK_DEV_IDETAPE is not set
482# CONFIG_BLK_DEV_IDEFLOPPY is not set 489# CONFIG_BLK_DEV_IDEFLOPPY is not set
@@ -486,10 +493,10 @@ CONFIG_BLK_DEV_IDECD=y
486# 493#
487# IDE chipset support/bugfixes 494# IDE chipset support/bugfixes
488# 495#
489# CONFIG_IDE_GENERIC is not set 496CONFIG_IDE_GENERIC=y
490# CONFIG_BLK_DEV_CMD640 is not set 497# CONFIG_BLK_DEV_CMD640 is not set
491CONFIG_BLK_DEV_IDEPCI=y 498CONFIG_BLK_DEV_IDEPCI=y
492CONFIG_IDEPCI_SHARE_IRQ=y 499# CONFIG_IDEPCI_SHARE_IRQ is not set
493# CONFIG_BLK_DEV_OFFBOARD is not set 500# CONFIG_BLK_DEV_OFFBOARD is not set
494# CONFIG_BLK_DEV_GENERIC is not set 501# CONFIG_BLK_DEV_GENERIC is not set
495# CONFIG_BLK_DEV_OPTI621 is not set 502# CONFIG_BLK_DEV_OPTI621 is not set
@@ -500,7 +507,7 @@ CONFIG_IDEDMA_PCI_AUTO=y
500# CONFIG_IDEDMA_ONLYDISK is not set 507# CONFIG_IDEDMA_ONLYDISK is not set
501# CONFIG_BLK_DEV_AEC62XX is not set 508# CONFIG_BLK_DEV_AEC62XX is not set
502# CONFIG_BLK_DEV_ALI15X3 is not set 509# CONFIG_BLK_DEV_ALI15X3 is not set
503# CONFIG_BLK_DEV_AMD74XX is not set 510CONFIG_BLK_DEV_AMD74XX=y
504# CONFIG_BLK_DEV_ATIIXP is not set 511# CONFIG_BLK_DEV_ATIIXP is not set
505# CONFIG_BLK_DEV_CMD64X is not set 512# CONFIG_BLK_DEV_CMD64X is not set
506# CONFIG_BLK_DEV_TRIFLEX is not set 513# CONFIG_BLK_DEV_TRIFLEX is not set
@@ -511,7 +518,7 @@ CONFIG_IDEDMA_PCI_AUTO=y
511# CONFIG_BLK_DEV_HPT34X is not set 518# CONFIG_BLK_DEV_HPT34X is not set
512# CONFIG_BLK_DEV_HPT366 is not set 519# CONFIG_BLK_DEV_HPT366 is not set
513# CONFIG_BLK_DEV_SC1200 is not set 520# CONFIG_BLK_DEV_SC1200 is not set
514# CONFIG_BLK_DEV_PIIX is not set 521CONFIG_BLK_DEV_PIIX=y
515# CONFIG_BLK_DEV_IT821X is not set 522# CONFIG_BLK_DEV_IT821X is not set
516# CONFIG_BLK_DEV_NS87415 is not set 523# CONFIG_BLK_DEV_NS87415 is not set
517# CONFIG_BLK_DEV_PDC202XX_OLD is not set 524# CONFIG_BLK_DEV_PDC202XX_OLD is not set
@@ -521,7 +528,7 @@ CONFIG_IDEDMA_PCI_AUTO=y
521# CONFIG_BLK_DEV_SIS5513 is not set 528# CONFIG_BLK_DEV_SIS5513 is not set
522# CONFIG_BLK_DEV_SLC90E66 is not set 529# CONFIG_BLK_DEV_SLC90E66 is not set
523# CONFIG_BLK_DEV_TRM290 is not set 530# CONFIG_BLK_DEV_TRM290 is not set
524CONFIG_BLK_DEV_VIA82CXXX=y 531# CONFIG_BLK_DEV_VIA82CXXX is not set
525# CONFIG_IDE_ARM is not set 532# CONFIG_IDE_ARM is not set
526CONFIG_BLK_DEV_IDEDMA=y 533CONFIG_BLK_DEV_IDEDMA=y
527# CONFIG_IDEDMA_IVB is not set 534# CONFIG_IDEDMA_IVB is not set
@@ -533,6 +540,7 @@ CONFIG_IDEDMA_AUTO=y
533# 540#
534# CONFIG_RAID_ATTRS is not set 541# CONFIG_RAID_ATTRS is not set
535CONFIG_SCSI=y 542CONFIG_SCSI=y
543CONFIG_SCSI_NETLINK=y
536# CONFIG_SCSI_PROC_FS is not set 544# CONFIG_SCSI_PROC_FS is not set
537 545
538# 546#
@@ -541,8 +549,9 @@ CONFIG_SCSI=y
541CONFIG_BLK_DEV_SD=y 549CONFIG_BLK_DEV_SD=y
542# CONFIG_CHR_DEV_ST is not set 550# CONFIG_CHR_DEV_ST is not set
543# CONFIG_CHR_DEV_OSST is not set 551# CONFIG_CHR_DEV_OSST is not set
544# CONFIG_BLK_DEV_SR is not set 552CONFIG_BLK_DEV_SR=y
545# CONFIG_CHR_DEV_SG is not set 553# CONFIG_BLK_DEV_SR_VENDOR is not set
554CONFIG_CHR_DEV_SG=y
546# CONFIG_CHR_DEV_SCH is not set 555# CONFIG_CHR_DEV_SCH is not set
547 556
548# 557#
@@ -553,29 +562,44 @@ CONFIG_BLK_DEV_SD=y
553# CONFIG_SCSI_LOGGING is not set 562# CONFIG_SCSI_LOGGING is not set
554 563
555# 564#
556# SCSI Transport Attributes 565# SCSI Transports
557# 566#
558# CONFIG_SCSI_SPI_ATTRS is not set 567CONFIG_SCSI_SPI_ATTRS=y
559# CONFIG_SCSI_FC_ATTRS is not set 568CONFIG_SCSI_FC_ATTRS=y
560# CONFIG_SCSI_ISCSI_ATTRS is not set 569# CONFIG_SCSI_ISCSI_ATTRS is not set
561# CONFIG_SCSI_SAS_ATTRS is not set 570# CONFIG_SCSI_SAS_ATTRS is not set
571# CONFIG_SCSI_SAS_LIBSAS is not set
562 572
563# 573#
564# SCSI low-level drivers 574# SCSI low-level drivers
565# 575#
566# CONFIG_ISCSI_TCP is not set 576# CONFIG_ISCSI_TCP is not set
567# CONFIG_BLK_DEV_3W_XXXX_RAID is not set 577CONFIG_BLK_DEV_3W_XXXX_RAID=y
568# CONFIG_SCSI_3W_9XXX is not set 578# CONFIG_SCSI_3W_9XXX is not set
569# CONFIG_SCSI_ACARD is not set 579# CONFIG_SCSI_ACARD is not set
570# CONFIG_SCSI_AACRAID is not set 580# CONFIG_SCSI_AACRAID is not set
571# CONFIG_SCSI_AIC7XXX is not set 581CONFIG_SCSI_AIC7XXX=y
582CONFIG_AIC7XXX_CMDS_PER_DEVICE=32
583CONFIG_AIC7XXX_RESET_DELAY_MS=5000
584CONFIG_AIC7XXX_DEBUG_ENABLE=y
585CONFIG_AIC7XXX_DEBUG_MASK=0
586CONFIG_AIC7XXX_REG_PRETTY_PRINT=y
572# CONFIG_SCSI_AIC7XXX_OLD is not set 587# CONFIG_SCSI_AIC7XXX_OLD is not set
573# CONFIG_SCSI_AIC79XX is not set 588CONFIG_SCSI_AIC79XX=y
589CONFIG_AIC79XX_CMDS_PER_DEVICE=32
590CONFIG_AIC79XX_RESET_DELAY_MS=4000
591# CONFIG_AIC79XX_ENABLE_RD_STRM is not set
592# CONFIG_AIC79XX_DEBUG_ENABLE is not set
593CONFIG_AIC79XX_DEBUG_MASK=0
594# CONFIG_AIC79XX_REG_PRETTY_PRINT is not set
595# CONFIG_SCSI_AIC94XX is not set
574# CONFIG_SCSI_DPT_I2O is not set 596# CONFIG_SCSI_DPT_I2O is not set
597# CONFIG_SCSI_ADVANSYS is not set
598# CONFIG_SCSI_ARCMSR is not set
575# CONFIG_MEGARAID_NEWGEN is not set 599# CONFIG_MEGARAID_NEWGEN is not set
576# CONFIG_MEGARAID_LEGACY is not set 600# CONFIG_MEGARAID_LEGACY is not set
577# CONFIG_MEGARAID_SAS is not set 601# CONFIG_MEGARAID_SAS is not set
578# CONFIG_SCSI_SATA is not set 602# CONFIG_SCSI_HPTIOP is not set
579# CONFIG_SCSI_BUSLOGIC is not set 603# CONFIG_SCSI_BUSLOGIC is not set
580# CONFIG_SCSI_DMX3191D is not set 604# CONFIG_SCSI_DMX3191D is not set
581# CONFIG_SCSI_EATA is not set 605# CONFIG_SCSI_EATA is not set
@@ -584,11 +608,9 @@ CONFIG_BLK_DEV_SD=y
584# CONFIG_SCSI_IPS is not set 608# CONFIG_SCSI_IPS is not set
585# CONFIG_SCSI_INITIO is not set 609# CONFIG_SCSI_INITIO is not set
586# CONFIG_SCSI_INIA100 is not set 610# CONFIG_SCSI_INIA100 is not set
587# CONFIG_SCSI_PPA is not set 611# CONFIG_SCSI_STEX is not set
588# CONFIG_SCSI_IMM is not set
589# CONFIG_SCSI_SYM53C8XX_2 is not set 612# CONFIG_SCSI_SYM53C8XX_2 is not set
590# CONFIG_SCSI_IPR is not set 613# CONFIG_SCSI_IPR is not set
591# CONFIG_SCSI_QLOGIC_FC is not set
592# CONFIG_SCSI_QLOGIC_1280 is not set 614# CONFIG_SCSI_QLOGIC_1280 is not set
593# CONFIG_SCSI_QLA_FC is not set 615# CONFIG_SCSI_QLA_FC is not set
594# CONFIG_SCSI_LPFC is not set 616# CONFIG_SCSI_LPFC is not set
@@ -598,22 +620,114 @@ CONFIG_BLK_DEV_SD=y
598# CONFIG_SCSI_DEBUG is not set 620# CONFIG_SCSI_DEBUG is not set
599 621
600# 622#
623# Serial ATA (prod) and Parallel ATA (experimental) drivers
624#
625CONFIG_ATA=y
626CONFIG_SATA_AHCI=y
627CONFIG_SATA_SVW=y
628CONFIG_ATA_PIIX=y
629# CONFIG_SATA_MV is not set
630CONFIG_SATA_NV=y
631# CONFIG_PDC_ADMA is not set
632# CONFIG_SATA_QSTOR is not set
633# CONFIG_SATA_PROMISE is not set
634# CONFIG_SATA_SX4 is not set
635CONFIG_SATA_SIL=y
636# CONFIG_SATA_SIL24 is not set
637# CONFIG_SATA_SIS is not set
638# CONFIG_SATA_ULI is not set
639CONFIG_SATA_VIA=y
640# CONFIG_SATA_VITESSE is not set
641CONFIG_SATA_INTEL_COMBINED=y
642# CONFIG_PATA_ALI is not set
643# CONFIG_PATA_AMD is not set
644# CONFIG_PATA_ARTOP is not set
645# CONFIG_PATA_ATIIXP is not set
646# CONFIG_PATA_CMD64X is not set
647# CONFIG_PATA_CS5520 is not set
648# CONFIG_PATA_CS5530 is not set
649# CONFIG_PATA_CS5535 is not set
650# CONFIG_PATA_CYPRESS is not set
651# CONFIG_PATA_EFAR is not set
652# CONFIG_ATA_GENERIC is not set
653# CONFIG_PATA_HPT366 is not set
654# CONFIG_PATA_HPT37X is not set
655# CONFIG_PATA_HPT3X2N is not set
656# CONFIG_PATA_HPT3X3 is not set
657# CONFIG_PATA_IT821X is not set
658# CONFIG_PATA_JMICRON is not set
659# CONFIG_PATA_LEGACY is not set
660# CONFIG_PATA_TRIFLEX is not set
661# CONFIG_PATA_MPIIX is not set
662# CONFIG_PATA_OLDPIIX is not set
663# CONFIG_PATA_NETCELL is not set
664# CONFIG_PATA_NS87410 is not set
665# CONFIG_PATA_OPTI is not set
666# CONFIG_PATA_OPTIDMA is not set
667# CONFIG_PATA_PDC_OLD is not set
668# CONFIG_PATA_QDI is not set
669# CONFIG_PATA_RADISYS is not set
670# CONFIG_PATA_RZ1000 is not set
671# CONFIG_PATA_SC1200 is not set
672# CONFIG_PATA_SERVERWORKS is not set
673# CONFIG_PATA_PDC2027X is not set
674# CONFIG_PATA_SIL680 is not set
675# CONFIG_PATA_SIS is not set
676# CONFIG_PATA_VIA is not set
677# CONFIG_PATA_WINBOND is not set
678
679#
601# Multi-device support (RAID and LVM) 680# Multi-device support (RAID and LVM)
602# 681#
603# CONFIG_MD is not set 682CONFIG_MD=y
683# CONFIG_BLK_DEV_MD is not set
684CONFIG_BLK_DEV_DM=y
685# CONFIG_DM_CRYPT is not set
686# CONFIG_DM_SNAPSHOT is not set
687# CONFIG_DM_MIRROR is not set
688# CONFIG_DM_ZERO is not set
689# CONFIG_DM_MULTIPATH is not set
604 690
605# 691#
606# Fusion MPT device support 692# Fusion MPT device support
607# 693#
608# CONFIG_FUSION is not set 694CONFIG_FUSION=y
609# CONFIG_FUSION_SPI is not set 695CONFIG_FUSION_SPI=y
610# CONFIG_FUSION_FC is not set 696# CONFIG_FUSION_FC is not set
611# CONFIG_FUSION_SAS is not set 697# CONFIG_FUSION_SAS is not set
698CONFIG_FUSION_MAX_SGE=128
699# CONFIG_FUSION_CTL is not set
612 700
613# 701#
614# IEEE 1394 (FireWire) support 702# IEEE 1394 (FireWire) support
615# 703#
616# CONFIG_IEEE1394 is not set 704CONFIG_IEEE1394=y
705
706#
707# Subsystem Options
708#
709# CONFIG_IEEE1394_VERBOSEDEBUG is not set
710# CONFIG_IEEE1394_OUI_DB is not set
711# CONFIG_IEEE1394_EXTRA_CONFIG_ROMS is not set
712# CONFIG_IEEE1394_EXPORT_FULL_API is not set
713
714#
715# Device Drivers
716#
717
718#
719# Texas Instruments PCILynx requires I2C
720#
721CONFIG_IEEE1394_OHCI1394=y
722
723#
724# Protocol Drivers
725#
726# CONFIG_IEEE1394_VIDEO1394 is not set
727# CONFIG_IEEE1394_SBP2 is not set
728# CONFIG_IEEE1394_ETH1394 is not set
729# CONFIG_IEEE1394_DV1394 is not set
730CONFIG_IEEE1394_RAWIO=y
617 731
618# 732#
619# I2O device support 733# I2O device support
@@ -652,46 +766,63 @@ CONFIG_MII=y
652# 766#
653# Tulip family network device support 767# Tulip family network device support
654# 768#
655# CONFIG_NET_TULIP is not set 769CONFIG_NET_TULIP=y
770# CONFIG_DE2104X is not set
771CONFIG_TULIP=y
772# CONFIG_TULIP_MWI is not set
773# CONFIG_TULIP_MMIO is not set
774# CONFIG_TULIP_NAPI is not set
775# CONFIG_DE4X5 is not set
776# CONFIG_WINBOND_840 is not set
777# CONFIG_DM9102 is not set
778# CONFIG_ULI526X is not set
656# CONFIG_HP100 is not set 779# CONFIG_HP100 is not set
657CONFIG_NET_PCI=y 780CONFIG_NET_PCI=y
658# CONFIG_PCNET32 is not set 781# CONFIG_PCNET32 is not set
659# CONFIG_AMD8111_ETH is not set 782# CONFIG_AMD8111_ETH is not set
660# CONFIG_ADAPTEC_STARFIRE is not set 783# CONFIG_ADAPTEC_STARFIRE is not set
661# CONFIG_B44 is not set 784CONFIG_B44=y
662# CONFIG_FORCEDETH is not set 785CONFIG_FORCEDETH=y
786# CONFIG_FORCEDETH_NAPI is not set
663# CONFIG_DGRS is not set 787# CONFIG_DGRS is not set
664# CONFIG_EEPRO100 is not set 788# CONFIG_EEPRO100 is not set
665CONFIG_E100=y 789CONFIG_E100=y
666# CONFIG_FEALNX is not set 790# CONFIG_FEALNX is not set
667# CONFIG_NATSEMI is not set 791# CONFIG_NATSEMI is not set
668# CONFIG_NE2K_PCI is not set 792# CONFIG_NE2K_PCI is not set
669# CONFIG_8139CP is not set 793CONFIG_8139CP=y
670# CONFIG_8139TOO is not set 794CONFIG_8139TOO=y
795# CONFIG_8139TOO_PIO is not set
796# CONFIG_8139TOO_TUNE_TWISTER is not set
797# CONFIG_8139TOO_8129 is not set
798# CONFIG_8139_OLD_RX_RESET is not set
671# CONFIG_SIS900 is not set 799# CONFIG_SIS900 is not set
672# CONFIG_EPIC100 is not set 800# CONFIG_EPIC100 is not set
673# CONFIG_SUNDANCE is not set 801# CONFIG_SUNDANCE is not set
674# CONFIG_TLAN is not set 802# CONFIG_TLAN is not set
675# CONFIG_VIA_RHINE is not set 803# CONFIG_VIA_RHINE is not set
676# CONFIG_NET_POCKET is not set
677 804
678# 805#
679# Ethernet (1000 Mbit) 806# Ethernet (1000 Mbit)
680# 807#
681# CONFIG_ACENIC is not set 808# CONFIG_ACENIC is not set
682# CONFIG_DL2K is not set 809# CONFIG_DL2K is not set
683# CONFIG_E1000 is not set 810CONFIG_E1000=y
811# CONFIG_E1000_NAPI is not set
812# CONFIG_E1000_DISABLE_PACKET_SPLIT is not set
684# CONFIG_NS83820 is not set 813# CONFIG_NS83820 is not set
685# CONFIG_HAMACHI is not set 814# CONFIG_HAMACHI is not set
686# CONFIG_YELLOWFIN is not set 815# CONFIG_YELLOWFIN is not set
687# CONFIG_R8169 is not set 816CONFIG_R8169=y
817# CONFIG_R8169_NAPI is not set
688# CONFIG_SIS190 is not set 818# CONFIG_SIS190 is not set
689# CONFIG_SKGE is not set 819# CONFIG_SKGE is not set
690# CONFIG_SKY2 is not set 820CONFIG_SKY2=y
691# CONFIG_SK98LIN is not set 821# CONFIG_SK98LIN is not set
692# CONFIG_VIA_VELOCITY is not set 822# CONFIG_VIA_VELOCITY is not set
693# CONFIG_TIGON3 is not set 823CONFIG_TIGON3=y
694# CONFIG_BNX2 is not set 824CONFIG_BNX2=y
825# CONFIG_QLA3XXX is not set
695 826
696# 827#
697# Ethernet (10000 Mbit) 828# Ethernet (10000 Mbit)
@@ -699,6 +830,7 @@ CONFIG_E100=y
699# CONFIG_CHELSIO_T1 is not set 830# CONFIG_CHELSIO_T1 is not set
700# CONFIG_IXGB is not set 831# CONFIG_IXGB is not set
701# CONFIG_S2IO is not set 832# CONFIG_S2IO is not set
833# CONFIG_MYRI10GE is not set
702 834
703# 835#
704# Token Ring devices 836# Token Ring devices
@@ -716,14 +848,15 @@ CONFIG_E100=y
716# CONFIG_WAN is not set 848# CONFIG_WAN is not set
717# CONFIG_FDDI is not set 849# CONFIG_FDDI is not set
718# CONFIG_HIPPI is not set 850# CONFIG_HIPPI is not set
719# CONFIG_PLIP is not set
720# CONFIG_PPP is not set 851# CONFIG_PPP is not set
721# CONFIG_SLIP is not set 852# CONFIG_SLIP is not set
722# CONFIG_NET_FC is not set 853# CONFIG_NET_FC is not set
723# CONFIG_SHAPER is not set 854# CONFIG_SHAPER is not set
724# CONFIG_NETCONSOLE is not set 855CONFIG_NETCONSOLE=y
725# CONFIG_NETPOLL is not set 856CONFIG_NETPOLL=y
726# CONFIG_NET_POLL_CONTROLLER is not set 857# CONFIG_NETPOLL_RX is not set
858# CONFIG_NETPOLL_TRAP is not set
859CONFIG_NET_POLL_CONTROLLER=y
727 860
728# 861#
729# ISDN subsystem 862# ISDN subsystem
@@ -745,8 +878,8 @@ CONFIG_INPUT=y
745# 878#
746CONFIG_INPUT_MOUSEDEV=y 879CONFIG_INPUT_MOUSEDEV=y
747CONFIG_INPUT_MOUSEDEV_PSAUX=y 880CONFIG_INPUT_MOUSEDEV_PSAUX=y
748CONFIG_INPUT_MOUSEDEV_SCREEN_X=1280 881CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
749CONFIG_INPUT_MOUSEDEV_SCREEN_Y=1024 882CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
750# CONFIG_INPUT_JOYDEV is not set 883# CONFIG_INPUT_JOYDEV is not set
751# CONFIG_INPUT_TSDEV is not set 884# CONFIG_INPUT_TSDEV is not set
752CONFIG_INPUT_EVDEV=y 885CONFIG_INPUT_EVDEV=y
@@ -776,7 +909,6 @@ CONFIG_SERIO=y
776CONFIG_SERIO_I8042=y 909CONFIG_SERIO_I8042=y
777# CONFIG_SERIO_SERPORT is not set 910# CONFIG_SERIO_SERPORT is not set
778# CONFIG_SERIO_CT82C710 is not set 911# CONFIG_SERIO_CT82C710 is not set
779# CONFIG_SERIO_PARKBD is not set
780# CONFIG_SERIO_PCIPS2 is not set 912# CONFIG_SERIO_PCIPS2 is not set
781CONFIG_SERIO_LIBPS2=y 913CONFIG_SERIO_LIBPS2=y
782# CONFIG_SERIO_RAW is not set 914# CONFIG_SERIO_RAW is not set
@@ -788,14 +920,15 @@ CONFIG_SERIO_LIBPS2=y
788CONFIG_VT=y 920CONFIG_VT=y
789CONFIG_VT_CONSOLE=y 921CONFIG_VT_CONSOLE=y
790CONFIG_HW_CONSOLE=y 922CONFIG_HW_CONSOLE=y
923# CONFIG_VT_HW_CONSOLE_BINDING is not set
791# CONFIG_SERIAL_NONSTANDARD is not set 924# CONFIG_SERIAL_NONSTANDARD is not set
792 925
793# 926#
794# Serial drivers 927# Serial drivers
795# 928#
796CONFIG_SERIAL_8250=y 929CONFIG_SERIAL_8250=y
797# CONFIG_SERIAL_8250_CONSOLE is not set 930CONFIG_SERIAL_8250_CONSOLE=y
798# CONFIG_SERIAL_8250_ACPI is not set 931CONFIG_SERIAL_8250_PCI=y
799CONFIG_SERIAL_8250_NR_UARTS=4 932CONFIG_SERIAL_8250_NR_UARTS=4
800CONFIG_SERIAL_8250_RUNTIME_UARTS=4 933CONFIG_SERIAL_8250_RUNTIME_UARTS=4
801# CONFIG_SERIAL_8250_EXTENDED is not set 934# CONFIG_SERIAL_8250_EXTENDED is not set
@@ -804,14 +937,11 @@ CONFIG_SERIAL_8250_RUNTIME_UARTS=4
804# Non-8250 serial port support 937# Non-8250 serial port support
805# 938#
806CONFIG_SERIAL_CORE=y 939CONFIG_SERIAL_CORE=y
940CONFIG_SERIAL_CORE_CONSOLE=y
807# CONFIG_SERIAL_JSM is not set 941# CONFIG_SERIAL_JSM is not set
808CONFIG_UNIX98_PTYS=y 942CONFIG_UNIX98_PTYS=y
809CONFIG_LEGACY_PTYS=y 943CONFIG_LEGACY_PTYS=y
810CONFIG_LEGACY_PTY_COUNT=256 944CONFIG_LEGACY_PTY_COUNT=256
811CONFIG_PRINTER=y
812# CONFIG_LP_CONSOLE is not set
813# CONFIG_PPDEV is not set
814# CONFIG_TIPAR is not set
815 945
816# 946#
817# IPMI 947# IPMI
@@ -822,8 +952,12 @@ CONFIG_PRINTER=y
822# Watchdog Cards 952# Watchdog Cards
823# 953#
824# CONFIG_WATCHDOG is not set 954# CONFIG_WATCHDOG is not set
825# CONFIG_HW_RANDOM is not set 955CONFIG_HW_RANDOM=y
826CONFIG_NVRAM=y 956CONFIG_HW_RANDOM_INTEL=y
957CONFIG_HW_RANDOM_AMD=y
958CONFIG_HW_RANDOM_GEODE=y
959CONFIG_HW_RANDOM_VIA=y
960# CONFIG_NVRAM is not set
827CONFIG_RTC=y 961CONFIG_RTC=y
828# CONFIG_DTLK is not set 962# CONFIG_DTLK is not set
829# CONFIG_R3964 is not set 963# CONFIG_R3964 is not set
@@ -833,31 +967,28 @@ CONFIG_RTC=y
833# 967#
834# Ftape, the floppy tape device driver 968# Ftape, the floppy tape device driver
835# 969#
836# CONFIG_FTAPE is not set
837CONFIG_AGP=y 970CONFIG_AGP=y
838# CONFIG_AGP_ALI is not set 971# CONFIG_AGP_ALI is not set
839# CONFIG_AGP_ATI is not set 972# CONFIG_AGP_ATI is not set
840# CONFIG_AGP_AMD is not set 973# CONFIG_AGP_AMD is not set
841# CONFIG_AGP_AMD64 is not set 974CONFIG_AGP_AMD64=y
842# CONFIG_AGP_INTEL is not set 975CONFIG_AGP_INTEL=y
843# CONFIG_AGP_NVIDIA is not set 976# CONFIG_AGP_NVIDIA is not set
844# CONFIG_AGP_SIS is not set 977# CONFIG_AGP_SIS is not set
845# CONFIG_AGP_SWORKS is not set 978# CONFIG_AGP_SWORKS is not set
846CONFIG_AGP_VIA=y 979# CONFIG_AGP_VIA is not set
847# CONFIG_AGP_EFFICEON is not set 980# CONFIG_AGP_EFFICEON is not set
848CONFIG_DRM=y 981# CONFIG_DRM is not set
849# CONFIG_DRM_TDFX is not set
850# CONFIG_DRM_R128 is not set
851CONFIG_DRM_RADEON=y
852# CONFIG_DRM_MGA is not set
853# CONFIG_DRM_SIS is not set
854# CONFIG_DRM_VIA is not set
855# CONFIG_DRM_SAVAGE is not set
856# CONFIG_MWAVE is not set 982# CONFIG_MWAVE is not set
983# CONFIG_PC8736x_GPIO is not set
984# CONFIG_NSC_GPIO is not set
857# CONFIG_CS5535_GPIO is not set 985# CONFIG_CS5535_GPIO is not set
858# CONFIG_RAW_DRIVER is not set 986CONFIG_RAW_DRIVER=y
859# CONFIG_HPET is not set 987CONFIG_MAX_RAW_DEVS=256
860# CONFIG_HANGCHECK_TIMER is not set 988CONFIG_HPET=y
989# CONFIG_HPET_RTC_IRQ is not set
990CONFIG_HPET_MMAP=y
991CONFIG_HANGCHECK_TIMER=y
861 992
862# 993#
863# TPM devices 994# TPM devices
@@ -868,59 +999,7 @@ CONFIG_DRM_RADEON=y
868# 999#
869# I2C support 1000# I2C support
870# 1001#
871CONFIG_I2C=y 1002# CONFIG_I2C is not set
872CONFIG_I2C_CHARDEV=y
873
874#
875# I2C Algorithms
876#
877CONFIG_I2C_ALGOBIT=y
878# CONFIG_I2C_ALGOPCF is not set
879# CONFIG_I2C_ALGOPCA is not set
880
881#
882# I2C Hardware Bus support
883#
884# CONFIG_I2C_ALI1535 is not set
885# CONFIG_I2C_ALI1563 is not set
886# CONFIG_I2C_ALI15X3 is not set
887# CONFIG_I2C_AMD756 is not set
888# CONFIG_I2C_AMD8111 is not set
889# CONFIG_I2C_I801 is not set
890# CONFIG_I2C_I810 is not set
891# CONFIG_I2C_PIIX4 is not set
892CONFIG_I2C_ISA=y
893# CONFIG_I2C_NFORCE2 is not set
894# CONFIG_I2C_PARPORT is not set
895# CONFIG_I2C_PARPORT_LIGHT is not set
896# CONFIG_I2C_PROSAVAGE is not set
897# CONFIG_I2C_SAVAGE4 is not set
898# CONFIG_SCx200_ACB is not set
899# CONFIG_I2C_SIS5595 is not set
900# CONFIG_I2C_SIS630 is not set
901# CONFIG_I2C_SIS96X is not set
902# CONFIG_I2C_STUB is not set
903# CONFIG_I2C_VIA is not set
904CONFIG_I2C_VIAPRO=y
905# CONFIG_I2C_VOODOO3 is not set
906# CONFIG_I2C_PCA_ISA is not set
907
908#
909# Miscellaneous I2C Chip support
910#
911# CONFIG_SENSORS_DS1337 is not set
912# CONFIG_SENSORS_DS1374 is not set
913# CONFIG_SENSORS_EEPROM is not set
914# CONFIG_SENSORS_PCF8574 is not set
915# CONFIG_SENSORS_PCA9539 is not set
916# CONFIG_SENSORS_PCF8591 is not set
917# CONFIG_SENSORS_RTC8564 is not set
918# CONFIG_SENSORS_MAX6875 is not set
919# CONFIG_RTC_X1205_I2C is not set
920# CONFIG_I2C_DEBUG_CORE is not set
921# CONFIG_I2C_DEBUG_ALGO is not set
922# CONFIG_I2C_DEBUG_BUS is not set
923# CONFIG_I2C_DEBUG_CHIP is not set
924 1003
925# 1004#
926# SPI support 1005# SPI support
@@ -931,51 +1010,12 @@ CONFIG_I2C_VIAPRO=y
931# 1010#
932# Dallas's 1-wire bus 1011# Dallas's 1-wire bus
933# 1012#
934# CONFIG_W1 is not set
935 1013
936# 1014#
937# Hardware Monitoring support 1015# Hardware Monitoring support
938# 1016#
939CONFIG_HWMON=y 1017# CONFIG_HWMON is not set
940CONFIG_HWMON_VID=y 1018# CONFIG_HWMON_VID is not set
941# CONFIG_SENSORS_ADM1021 is not set
942# CONFIG_SENSORS_ADM1025 is not set
943# CONFIG_SENSORS_ADM1026 is not set
944# CONFIG_SENSORS_ADM1031 is not set
945# CONFIG_SENSORS_ADM9240 is not set
946# CONFIG_SENSORS_ASB100 is not set
947# CONFIG_SENSORS_ATXP1 is not set
948# CONFIG_SENSORS_DS1621 is not set
949# CONFIG_SENSORS_F71805F is not set
950# CONFIG_SENSORS_FSCHER is not set
951# CONFIG_SENSORS_FSCPOS is not set
952# CONFIG_SENSORS_GL518SM is not set
953# CONFIG_SENSORS_GL520SM is not set
954CONFIG_SENSORS_IT87=y
955# CONFIG_SENSORS_LM63 is not set
956# CONFIG_SENSORS_LM75 is not set
957# CONFIG_SENSORS_LM77 is not set
958# CONFIG_SENSORS_LM78 is not set
959# CONFIG_SENSORS_LM80 is not set
960# CONFIG_SENSORS_LM83 is not set
961# CONFIG_SENSORS_LM85 is not set
962# CONFIG_SENSORS_LM87 is not set
963# CONFIG_SENSORS_LM90 is not set
964# CONFIG_SENSORS_LM92 is not set
965# CONFIG_SENSORS_MAX1619 is not set
966# CONFIG_SENSORS_PC87360 is not set
967# CONFIG_SENSORS_SIS5595 is not set
968# CONFIG_SENSORS_SMSC47M1 is not set
969# CONFIG_SENSORS_SMSC47B397 is not set
970# CONFIG_SENSORS_VIA686A is not set
971# CONFIG_SENSORS_VT8231 is not set
972# CONFIG_SENSORS_W83781D is not set
973# CONFIG_SENSORS_W83792D is not set
974# CONFIG_SENSORS_W83L785TS is not set
975# CONFIG_SENSORS_W83627HF is not set
976# CONFIG_SENSORS_W83627EHF is not set
977# CONFIG_SENSORS_HDAPS is not set
978# CONFIG_HWMON_DEBUG_CHIP is not set
979 1019
980# 1020#
981# Misc devices 1021# Misc devices
@@ -983,117 +1023,31 @@ CONFIG_SENSORS_IT87=y
983# CONFIG_IBM_ASM is not set 1023# CONFIG_IBM_ASM is not set
984 1024
985# 1025#
986# Multimedia Capabilities Port drivers
987#
988
989#
990# Multimedia devices 1026# Multimedia devices
991# 1027#
992CONFIG_VIDEO_DEV=y 1028# CONFIG_VIDEO_DEV is not set
993 1029CONFIG_VIDEO_V4L2=y
994#
995# Video For Linux
996#
997
998#
999# Video Adapters
1000#
1001# CONFIG_VIDEO_ADV_DEBUG is not set
1002# CONFIG_VIDEO_BT848 is not set
1003# CONFIG_VIDEO_BWQCAM is not set
1004# CONFIG_VIDEO_CQCAM is not set
1005# CONFIG_VIDEO_W9966 is not set
1006# CONFIG_VIDEO_CPIA is not set
1007# CONFIG_VIDEO_SAA5246A is not set
1008# CONFIG_VIDEO_SAA5249 is not set
1009# CONFIG_TUNER_3036 is not set
1010# CONFIG_VIDEO_STRADIS is not set
1011# CONFIG_VIDEO_ZORAN is not set
1012CONFIG_VIDEO_SAA7134=y
1013# CONFIG_VIDEO_SAA7134_ALSA is not set
1014# CONFIG_VIDEO_MXB is not set
1015# CONFIG_VIDEO_DPC is not set
1016# CONFIG_VIDEO_HEXIUM_ORION is not set
1017# CONFIG_VIDEO_HEXIUM_GEMINI is not set
1018# CONFIG_VIDEO_CX88 is not set
1019# CONFIG_VIDEO_EM28XX is not set
1020# CONFIG_VIDEO_OVCAMCHIP is not set
1021# CONFIG_VIDEO_AUDIO_DECODER is not set
1022# CONFIG_VIDEO_DECODER is not set
1023
1024#
1025# Radio Adapters
1026#
1027# CONFIG_RADIO_GEMTEK_PCI is not set
1028# CONFIG_RADIO_MAXIRADIO is not set
1029# CONFIG_RADIO_MAESTRO is not set
1030 1030
1031# 1031#
1032# Digital Video Broadcasting Devices 1032# Digital Video Broadcasting Devices
1033# 1033#
1034# CONFIG_DVB is not set 1034# CONFIG_DVB is not set
1035CONFIG_VIDEO_TUNER=y 1035# CONFIG_USB_DABUSB is not set
1036CONFIG_VIDEO_BUF=y
1037CONFIG_VIDEO_IR=y
1038 1036
1039# 1037#
1040# Graphics support 1038# Graphics support
1041# 1039#
1042CONFIG_FB=y 1040CONFIG_FIRMWARE_EDID=y
1043CONFIG_FB_CFB_FILLRECT=y 1041# CONFIG_FB is not set
1044CONFIG_FB_CFB_COPYAREA=y
1045CONFIG_FB_CFB_IMAGEBLIT=y
1046# CONFIG_FB_MACMODES is not set
1047CONFIG_FB_MODE_HELPERS=y
1048# CONFIG_FB_TILEBLITTING is not set
1049# CONFIG_FB_CIRRUS is not set
1050# CONFIG_FB_PM2 is not set
1051# CONFIG_FB_CYBER2000 is not set
1052# CONFIG_FB_ARC is not set
1053# CONFIG_FB_ASILIANT is not set
1054# CONFIG_FB_IMSTT is not set
1055# CONFIG_FB_VGA16 is not set
1056# CONFIG_FB_VESA is not set
1057CONFIG_VIDEO_SELECT=y
1058# CONFIG_FB_HGA is not set
1059# CONFIG_FB_S1D13XXX is not set
1060# CONFIG_FB_NVIDIA is not set
1061# CONFIG_FB_RIVA is not set
1062# CONFIG_FB_I810 is not set
1063# CONFIG_FB_INTEL is not set
1064# CONFIG_FB_MATROX is not set
1065# CONFIG_FB_RADEON_OLD is not set
1066CONFIG_FB_RADEON=y
1067CONFIG_FB_RADEON_I2C=y
1068# CONFIG_FB_RADEON_DEBUG is not set
1069# CONFIG_FB_ATY128 is not set
1070# CONFIG_FB_ATY is not set
1071# CONFIG_FB_SAVAGE is not set
1072# CONFIG_FB_SIS is not set
1073# CONFIG_FB_NEOMAGIC is not set
1074# CONFIG_FB_KYRO is not set
1075# CONFIG_FB_3DFX is not set
1076# CONFIG_FB_VOODOO1 is not set
1077# CONFIG_FB_CYBLA is not set
1078# CONFIG_FB_TRIDENT is not set
1079# CONFIG_FB_GEODE is not set
1080# CONFIG_FB_VIRTUAL is not set
1081 1042
1082# 1043#
1083# Console display driver support 1044# Console display driver support
1084# 1045#
1085CONFIG_VGA_CONSOLE=y 1046CONFIG_VGA_CONSOLE=y
1047CONFIG_VGACON_SOFT_SCROLLBACK=y
1048CONFIG_VGACON_SOFT_SCROLLBACK_SIZE=128
1049CONFIG_VIDEO_SELECT=y
1086CONFIG_DUMMY_CONSOLE=y 1050CONFIG_DUMMY_CONSOLE=y
1087CONFIG_FRAMEBUFFER_CONSOLE=y
1088# CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set
1089# CONFIG_FONTS is not set
1090CONFIG_FONT_8x8=y
1091CONFIG_FONT_8x16=y
1092
1093#
1094# Logo configuration
1095#
1096# CONFIG_LOGO is not set
1097# CONFIG_BACKLIGHT_LCD_SUPPORT is not set 1051# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
1098 1052
1099# 1053#
@@ -1104,97 +1058,30 @@ CONFIG_SOUND=y
1104# 1058#
1105# Advanced Linux Sound Architecture 1059# Advanced Linux Sound Architecture
1106# 1060#
1107CONFIG_SND=y 1061# CONFIG_SND is not set
1108CONFIG_SND_TIMER=y
1109CONFIG_SND_PCM=y
1110CONFIG_SND_RAWMIDI=y
1111CONFIG_SND_SEQUENCER=y
1112# CONFIG_SND_SEQ_DUMMY is not set
1113# CONFIG_SND_MIXER_OSS is not set
1114# CONFIG_SND_PCM_OSS is not set
1115# CONFIG_SND_SEQUENCER_OSS is not set
1116CONFIG_SND_RTCTIMER=y
1117CONFIG_SND_SEQ_RTCTIMER_DEFAULT=y
1118# CONFIG_SND_DYNAMIC_MINORS is not set
1119# CONFIG_SND_SUPPORT_OLD_API is not set
1120# CONFIG_SND_VERBOSE_PRINTK is not set
1121# CONFIG_SND_DEBUG is not set
1122
1123#
1124# Generic devices
1125#
1126CONFIG_SND_MPU401_UART=y
1127CONFIG_SND_AC97_CODEC=y
1128CONFIG_SND_AC97_BUS=y
1129# CONFIG_SND_DUMMY is not set
1130# CONFIG_SND_VIRMIDI is not set
1131# CONFIG_SND_MTPAV is not set
1132# CONFIG_SND_SERIAL_U16550 is not set
1133# CONFIG_SND_MPU401 is not set
1134
1135#
1136# PCI devices
1137#
1138# CONFIG_SND_AD1889 is not set
1139# CONFIG_SND_ALS4000 is not set
1140# CONFIG_SND_ALI5451 is not set
1141# CONFIG_SND_ATIIXP is not set
1142# CONFIG_SND_ATIIXP_MODEM is not set
1143# CONFIG_SND_AU8810 is not set
1144# CONFIG_SND_AU8820 is not set
1145# CONFIG_SND_AU8830 is not set
1146# CONFIG_SND_AZT3328 is not set
1147# CONFIG_SND_BT87X is not set
1148# CONFIG_SND_CA0106 is not set
1149# CONFIG_SND_CMIPCI is not set
1150# CONFIG_SND_CS4281 is not set
1151# CONFIG_SND_CS46XX is not set
1152# CONFIG_SND_CS5535AUDIO is not set
1153# CONFIG_SND_EMU10K1 is not set
1154# CONFIG_SND_EMU10K1X is not set
1155# CONFIG_SND_ENS1370 is not set
1156# CONFIG_SND_ENS1371 is not set
1157# CONFIG_SND_ES1938 is not set
1158# CONFIG_SND_ES1968 is not set
1159# CONFIG_SND_FM801 is not set
1160# CONFIG_SND_HDA_INTEL is not set
1161# CONFIG_SND_HDSP is not set
1162# CONFIG_SND_HDSPM is not set
1163# CONFIG_SND_ICE1712 is not set
1164# CONFIG_SND_ICE1724 is not set
1165# CONFIG_SND_INTEL8X0 is not set
1166# CONFIG_SND_INTEL8X0M is not set
1167# CONFIG_SND_KORG1212 is not set
1168# CONFIG_SND_MAESTRO3 is not set
1169# CONFIG_SND_MIXART is not set
1170# CONFIG_SND_NM256 is not set
1171# CONFIG_SND_PCXHR is not set
1172# CONFIG_SND_RME32 is not set
1173# CONFIG_SND_RME96 is not set
1174# CONFIG_SND_RME9652 is not set
1175# CONFIG_SND_SONICVIBES is not set
1176# CONFIG_SND_TRIDENT is not set
1177CONFIG_SND_VIA82XX=y
1178# CONFIG_SND_VIA82XX_MODEM is not set
1179# CONFIG_SND_VX222 is not set
1180# CONFIG_SND_YMFPCI is not set
1181
1182#
1183# USB devices
1184#
1185# CONFIG_SND_USB_AUDIO is not set
1186# CONFIG_SND_USB_USX2Y is not set
1187 1062
1188# 1063#
1189# Open Sound System 1064# Open Sound System
1190# 1065#
1191# CONFIG_SOUND_PRIME is not set 1066CONFIG_SOUND_PRIME=y
1067CONFIG_OSS_OBSOLETE_DRIVER=y
1068# CONFIG_SOUND_BT878 is not set
1069# CONFIG_SOUND_EMU10K1 is not set
1070# CONFIG_SOUND_FUSION is not set
1071# CONFIG_SOUND_ES1371 is not set
1072CONFIG_SOUND_ICH=y
1073# CONFIG_SOUND_TRIDENT is not set
1074# CONFIG_SOUND_MSNDCLAS is not set
1075# CONFIG_SOUND_MSNDPIN is not set
1076# CONFIG_SOUND_VIA82CXXX is not set
1077# CONFIG_SOUND_OSS is not set
1192 1078
1193# 1079#
1194# USB support 1080# USB support
1195# 1081#
1196CONFIG_USB_ARCH_HAS_HCD=y 1082CONFIG_USB_ARCH_HAS_HCD=y
1197CONFIG_USB_ARCH_HAS_OHCI=y 1083CONFIG_USB_ARCH_HAS_OHCI=y
1084CONFIG_USB_ARCH_HAS_EHCI=y
1198CONFIG_USB=y 1085CONFIG_USB=y
1199# CONFIG_USB_DEBUG is not set 1086# CONFIG_USB_DEBUG is not set
1200 1087
@@ -1213,17 +1100,19 @@ CONFIG_USB_DEVICEFS=y
1213CONFIG_USB_EHCI_HCD=y 1100CONFIG_USB_EHCI_HCD=y
1214# CONFIG_USB_EHCI_SPLIT_ISO is not set 1101# CONFIG_USB_EHCI_SPLIT_ISO is not set
1215# CONFIG_USB_EHCI_ROOT_HUB_TT is not set 1102# CONFIG_USB_EHCI_ROOT_HUB_TT is not set
1103# CONFIG_USB_EHCI_TT_NEWSCHED is not set
1216# CONFIG_USB_ISP116X_HCD is not set 1104# CONFIG_USB_ISP116X_HCD is not set
1217# CONFIG_USB_OHCI_HCD is not set 1105CONFIG_USB_OHCI_HCD=y
1106# CONFIG_USB_OHCI_BIG_ENDIAN is not set
1107CONFIG_USB_OHCI_LITTLE_ENDIAN=y
1218CONFIG_USB_UHCI_HCD=y 1108CONFIG_USB_UHCI_HCD=y
1219# CONFIG_USB_SL811_HCD is not set 1109# CONFIG_USB_SL811_HCD is not set
1220 1110
1221# 1111#
1222# USB Device Class drivers 1112# USB Device Class drivers
1223# 1113#
1224# CONFIG_OBSOLETE_OSS_USB_DRIVER is not set
1225# CONFIG_USB_ACM is not set 1114# CONFIG_USB_ACM is not set
1226# CONFIG_USB_PRINTER is not set 1115CONFIG_USB_PRINTER=y
1227 1116
1228# 1117#
1229# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support' 1118# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
@@ -1248,21 +1137,17 @@ CONFIG_USB_STORAGE=y
1248# 1137#
1249# USB Input Devices 1138# USB Input Devices
1250# 1139#
1251# CONFIG_USB_HID is not set 1140CONFIG_USB_HID=y
1252 1141CONFIG_USB_HIDINPUT=y
1253# 1142# CONFIG_USB_HIDINPUT_POWERBOOK is not set
1254# USB HID Boot Protocol drivers 1143# CONFIG_HID_FF is not set
1255# 1144# CONFIG_USB_HIDDEV is not set
1256# CONFIG_USB_KBD is not set
1257# CONFIG_USB_MOUSE is not set
1258# CONFIG_USB_AIPTEK is not set 1145# CONFIG_USB_AIPTEK is not set
1259# CONFIG_USB_WACOM is not set 1146# CONFIG_USB_WACOM is not set
1260# CONFIG_USB_ACECAD is not set 1147# CONFIG_USB_ACECAD is not set
1261# CONFIG_USB_KBTAB is not set 1148# CONFIG_USB_KBTAB is not set
1262# CONFIG_USB_POWERMATE is not set 1149# CONFIG_USB_POWERMATE is not set
1263# CONFIG_USB_MTOUCH is not set 1150# CONFIG_USB_TOUCHSCREEN is not set
1264# CONFIG_USB_ITMTOUCH is not set
1265# CONFIG_USB_EGALAX is not set
1266# CONFIG_USB_YEALINK is not set 1151# CONFIG_USB_YEALINK is not set
1267# CONFIG_USB_XPAD is not set 1152# CONFIG_USB_XPAD is not set
1268# CONFIG_USB_ATI_REMOTE is not set 1153# CONFIG_USB_ATI_REMOTE is not set
@@ -1277,21 +1162,6 @@ CONFIG_USB_STORAGE=y
1277# CONFIG_USB_MICROTEK is not set 1162# CONFIG_USB_MICROTEK is not set
1278 1163
1279# 1164#
1280# USB Multimedia devices
1281#
1282# CONFIG_USB_DABUSB is not set
1283# CONFIG_USB_VICAM is not set
1284# CONFIG_USB_DSBR is not set
1285# CONFIG_USB_ET61X251 is not set
1286# CONFIG_USB_IBMCAM is not set
1287# CONFIG_USB_KONICAWC is not set
1288# CONFIG_USB_OV511 is not set
1289# CONFIG_USB_SE401 is not set
1290# CONFIG_USB_SN9C102 is not set
1291# CONFIG_USB_STV680 is not set
1292# CONFIG_USB_PWC is not set
1293
1294#
1295# USB Network Adapters 1165# USB Network Adapters
1296# 1166#
1297# CONFIG_USB_CATC is not set 1167# CONFIG_USB_CATC is not set
@@ -1299,12 +1169,11 @@ CONFIG_USB_STORAGE=y
1299# CONFIG_USB_PEGASUS is not set 1169# CONFIG_USB_PEGASUS is not set
1300# CONFIG_USB_RTL8150 is not set 1170# CONFIG_USB_RTL8150 is not set
1301# CONFIG_USB_USBNET is not set 1171# CONFIG_USB_USBNET is not set
1302# CONFIG_USB_MON is not set 1172CONFIG_USB_MON=y
1303 1173
1304# 1174#
1305# USB port drivers 1175# USB port drivers
1306# 1176#
1307# CONFIG_USB_USS720 is not set
1308 1177
1309# 1178#
1310# USB Serial Converter support 1179# USB Serial Converter support
@@ -1321,10 +1190,12 @@ CONFIG_USB_STORAGE=y
1321# CONFIG_USB_LEGOTOWER is not set 1190# CONFIG_USB_LEGOTOWER is not set
1322# CONFIG_USB_LCD is not set 1191# CONFIG_USB_LCD is not set
1323# CONFIG_USB_LED is not set 1192# CONFIG_USB_LED is not set
1193# CONFIG_USB_CYPRESS_CY7C63 is not set
1324# CONFIG_USB_CYTHERM is not set 1194# CONFIG_USB_CYTHERM is not set
1325# CONFIG_USB_PHIDGETKIT is not set 1195# CONFIG_USB_PHIDGETKIT is not set
1326# CONFIG_USB_PHIDGETSERVO is not set 1196# CONFIG_USB_PHIDGETSERVO is not set
1327# CONFIG_USB_IDMOUSE is not set 1197# CONFIG_USB_IDMOUSE is not set
1198# CONFIG_USB_APPLEDISPLAY is not set
1328# CONFIG_USB_SISUSBVGA is not set 1199# CONFIG_USB_SISUSBVGA is not set
1329# CONFIG_USB_LD is not set 1200# CONFIG_USB_LD is not set
1330# CONFIG_USB_TEST is not set 1201# CONFIG_USB_TEST is not set
@@ -1344,56 +1215,96 @@ CONFIG_USB_STORAGE=y
1344# CONFIG_MMC is not set 1215# CONFIG_MMC is not set
1345 1216
1346# 1217#
1218# LED devices
1219#
1220# CONFIG_NEW_LEDS is not set
1221
1222#
1223# LED drivers
1224#
1225
1226#
1227# LED Triggers
1228#
1229
1230#
1347# InfiniBand support 1231# InfiniBand support
1348# 1232#
1349# CONFIG_INFINIBAND is not set 1233# CONFIG_INFINIBAND is not set
1350 1234
1351# 1235#
1352# SN Devices 1236# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
1237#
1238# CONFIG_EDAC is not set
1239
1240#
1241# Real Time Clock
1353# 1242#
1243# CONFIG_RTC_CLASS is not set
1354 1244
1355# 1245#
1356# EDAC - error detection and reporting (RAS) 1246# DMA Engine support
1247#
1248# CONFIG_DMA_ENGINE is not set
1249
1250#
1251# DMA Clients
1252#
1253
1254#
1255# DMA Devices
1357# 1256#
1358# CONFIG_EDAC is not set
1359 1257
1360# 1258#
1361# File systems 1259# File systems
1362# 1260#
1363CONFIG_EXT2_FS=y 1261CONFIG_EXT2_FS=y
1364# CONFIG_EXT2_FS_XATTR is not set 1262CONFIG_EXT2_FS_XATTR=y
1263CONFIG_EXT2_FS_POSIX_ACL=y
1264# CONFIG_EXT2_FS_SECURITY is not set
1365# CONFIG_EXT2_FS_XIP is not set 1265# CONFIG_EXT2_FS_XIP is not set
1366# CONFIG_EXT3_FS is not set 1266CONFIG_EXT3_FS=y
1367# CONFIG_REISERFS_FS is not set 1267CONFIG_EXT3_FS_XATTR=y
1268CONFIG_EXT3_FS_POSIX_ACL=y
1269# CONFIG_EXT3_FS_SECURITY is not set
1270CONFIG_JBD=y
1271# CONFIG_JBD_DEBUG is not set
1272CONFIG_FS_MBCACHE=y
1273CONFIG_REISERFS_FS=y
1274# CONFIG_REISERFS_CHECK is not set
1275# CONFIG_REISERFS_PROC_INFO is not set
1276CONFIG_REISERFS_FS_XATTR=y
1277CONFIG_REISERFS_FS_POSIX_ACL=y
1278# CONFIG_REISERFS_FS_SECURITY is not set
1368# CONFIG_JFS_FS is not set 1279# CONFIG_JFS_FS is not set
1369# CONFIG_FS_POSIX_ACL is not set 1280CONFIG_FS_POSIX_ACL=y
1370# CONFIG_XFS_FS is not set 1281# CONFIG_XFS_FS is not set
1371# CONFIG_OCFS2_FS is not set 1282# CONFIG_OCFS2_FS is not set
1372# CONFIG_MINIX_FS is not set 1283# CONFIG_MINIX_FS is not set
1373# CONFIG_ROMFS_FS is not set 1284# CONFIG_ROMFS_FS is not set
1374# CONFIG_INOTIFY is not set 1285CONFIG_INOTIFY=y
1286CONFIG_INOTIFY_USER=y
1375# CONFIG_QUOTA is not set 1287# CONFIG_QUOTA is not set
1376CONFIG_DNOTIFY=y 1288CONFIG_DNOTIFY=y
1377# CONFIG_AUTOFS_FS is not set 1289# CONFIG_AUTOFS_FS is not set
1378# CONFIG_AUTOFS4_FS is not set 1290CONFIG_AUTOFS4_FS=y
1379# CONFIG_FUSE_FS is not set 1291# CONFIG_FUSE_FS is not set
1380 1292
1381# 1293#
1382# CD-ROM/DVD Filesystems 1294# CD-ROM/DVD Filesystems
1383# 1295#
1384CONFIG_ISO9660_FS=y 1296CONFIG_ISO9660_FS=y
1385CONFIG_JOLIET=y 1297# CONFIG_JOLIET is not set
1386CONFIG_ZISOFS=y 1298# CONFIG_ZISOFS is not set
1387CONFIG_ZISOFS_FS=y
1388# CONFIG_UDF_FS is not set 1299# CONFIG_UDF_FS is not set
1389 1300
1390# 1301#
1391# DOS/FAT/NT Filesystems 1302# DOS/FAT/NT Filesystems
1392# 1303#
1393CONFIG_FAT_FS=y 1304CONFIG_FAT_FS=y
1394# CONFIG_MSDOS_FS is not set 1305CONFIG_MSDOS_FS=y
1395CONFIG_VFAT_FS=y 1306CONFIG_VFAT_FS=y
1396CONFIG_FAT_DEFAULT_CODEPAGE=850 1307CONFIG_FAT_DEFAULT_CODEPAGE=437
1397CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1" 1308CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
1398# CONFIG_NTFS_FS is not set 1309# CONFIG_NTFS_FS is not set
1399 1310
@@ -1404,10 +1315,9 @@ CONFIG_PROC_FS=y
1404CONFIG_PROC_KCORE=y 1315CONFIG_PROC_KCORE=y
1405CONFIG_SYSFS=y 1316CONFIG_SYSFS=y
1406CONFIG_TMPFS=y 1317CONFIG_TMPFS=y
1407# CONFIG_HUGETLBFS is not set 1318CONFIG_HUGETLBFS=y
1408# CONFIG_HUGETLB_PAGE is not set 1319CONFIG_HUGETLB_PAGE=y
1409CONFIG_RAMFS=y 1320CONFIG_RAMFS=y
1410# CONFIG_RELAYFS_FS is not set
1411# CONFIG_CONFIGFS_FS is not set 1321# CONFIG_CONFIGFS_FS is not set
1412 1322
1413# 1323#
@@ -1430,13 +1340,26 @@ CONFIG_RAMFS=y
1430# 1340#
1431# Network File Systems 1341# Network File Systems
1432# 1342#
1433# CONFIG_NFS_FS is not set 1343CONFIG_NFS_FS=y
1434# CONFIG_NFSD is not set 1344CONFIG_NFS_V3=y
1345# CONFIG_NFS_V3_ACL is not set
1346# CONFIG_NFS_V4 is not set
1347# CONFIG_NFS_DIRECTIO is not set
1348CONFIG_NFSD=y
1349CONFIG_NFSD_V3=y
1350# CONFIG_NFSD_V3_ACL is not set
1351# CONFIG_NFSD_V4 is not set
1352CONFIG_NFSD_TCP=y
1353CONFIG_ROOT_NFS=y
1354CONFIG_LOCKD=y
1355CONFIG_LOCKD_V4=y
1356CONFIG_EXPORTFS=y
1357CONFIG_NFS_COMMON=y
1358CONFIG_SUNRPC=y
1359# CONFIG_RPCSEC_GSS_KRB5 is not set
1360# CONFIG_RPCSEC_GSS_SPKM3 is not set
1435# CONFIG_SMB_FS is not set 1361# CONFIG_SMB_FS is not set
1436CONFIG_CIFS=y 1362# CONFIG_CIFS is not set
1437# CONFIG_CIFS_STATS is not set
1438# CONFIG_CIFS_XATTR is not set
1439# CONFIG_CIFS_EXPERIMENTAL is not set
1440# CONFIG_NCP_FS is not set 1363# CONFIG_NCP_FS is not set
1441# CONFIG_CODA_FS is not set 1364# CONFIG_CODA_FS is not set
1442# CONFIG_AFS_FS is not set 1365# CONFIG_AFS_FS is not set
@@ -1445,33 +1368,18 @@ CONFIG_CIFS=y
1445# 1368#
1446# Partition Types 1369# Partition Types
1447# 1370#
1448CONFIG_PARTITION_ADVANCED=y 1371# CONFIG_PARTITION_ADVANCED is not set
1449# CONFIG_ACORN_PARTITION is not set
1450# CONFIG_OSF_PARTITION is not set
1451# CONFIG_AMIGA_PARTITION is not set
1452# CONFIG_ATARI_PARTITION is not set
1453# CONFIG_MAC_PARTITION is not set
1454CONFIG_MSDOS_PARTITION=y 1372CONFIG_MSDOS_PARTITION=y
1455# CONFIG_BSD_DISKLABEL is not set
1456# CONFIG_MINIX_SUBPARTITION is not set
1457# CONFIG_SOLARIS_X86_PARTITION is not set
1458# CONFIG_UNIXWARE_DISKLABEL is not set
1459# CONFIG_LDM_PARTITION is not set
1460# CONFIG_SGI_PARTITION is not set
1461# CONFIG_ULTRIX_PARTITION is not set
1462# CONFIG_SUN_PARTITION is not set
1463# CONFIG_KARMA_PARTITION is not set
1464# CONFIG_EFI_PARTITION is not set
1465 1373
1466# 1374#
1467# Native Language Support 1375# Native Language Support
1468# 1376#
1469CONFIG_NLS=y 1377CONFIG_NLS=y
1470CONFIG_NLS_DEFAULT="iso8859-15" 1378CONFIG_NLS_DEFAULT="iso8859-1"
1471# CONFIG_NLS_CODEPAGE_437 is not set 1379CONFIG_NLS_CODEPAGE_437=y
1472# CONFIG_NLS_CODEPAGE_737 is not set 1380# CONFIG_NLS_CODEPAGE_737 is not set
1473# CONFIG_NLS_CODEPAGE_775 is not set 1381# CONFIG_NLS_CODEPAGE_775 is not set
1474CONFIG_NLS_CODEPAGE_850=y 1382# CONFIG_NLS_CODEPAGE_850 is not set
1475# CONFIG_NLS_CODEPAGE_852 is not set 1383# CONFIG_NLS_CODEPAGE_852 is not set
1476# CONFIG_NLS_CODEPAGE_855 is not set 1384# CONFIG_NLS_CODEPAGE_855 is not set
1477# CONFIG_NLS_CODEPAGE_857 is not set 1385# CONFIG_NLS_CODEPAGE_857 is not set
@@ -1491,7 +1399,7 @@ CONFIG_NLS_CODEPAGE_850=y
1491# CONFIG_NLS_ISO8859_8 is not set 1399# CONFIG_NLS_ISO8859_8 is not set
1492# CONFIG_NLS_CODEPAGE_1250 is not set 1400# CONFIG_NLS_CODEPAGE_1250 is not set
1493# CONFIG_NLS_CODEPAGE_1251 is not set 1401# CONFIG_NLS_CODEPAGE_1251 is not set
1494# CONFIG_NLS_ASCII is not set 1402CONFIG_NLS_ASCII=y
1495CONFIG_NLS_ISO8859_1=y 1403CONFIG_NLS_ISO8859_1=y
1496# CONFIG_NLS_ISO8859_2 is not set 1404# CONFIG_NLS_ISO8859_2 is not set
1497# CONFIG_NLS_ISO8859_3 is not set 1405# CONFIG_NLS_ISO8859_3 is not set
@@ -1510,20 +1418,50 @@ CONFIG_NLS_UTF8=y
1510# 1418#
1511# Instrumentation Support 1419# Instrumentation Support
1512# 1420#
1513# CONFIG_PROFILING is not set 1421CONFIG_PROFILING=y
1514# CONFIG_KPROBES is not set 1422CONFIG_OPROFILE=y
1423CONFIG_KPROBES=y
1515 1424
1516# 1425#
1517# Kernel hacking 1426# Kernel hacking
1518# 1427#
1428CONFIG_TRACE_IRQFLAGS_SUPPORT=y
1519# CONFIG_PRINTK_TIME is not set 1429# CONFIG_PRINTK_TIME is not set
1520CONFIG_MAGIC_SYSRQ=y 1430CONFIG_MAGIC_SYSRQ=y
1521# CONFIG_DEBUG_KERNEL is not set 1431CONFIG_UNUSED_SYMBOLS=y
1522CONFIG_LOG_BUF_SHIFT=14 1432CONFIG_DEBUG_KERNEL=y
1433CONFIG_LOG_BUF_SHIFT=18
1434CONFIG_DETECT_SOFTLOCKUP=y
1435# CONFIG_SCHEDSTATS is not set
1436# CONFIG_DEBUG_SLAB is not set
1437# CONFIG_DEBUG_RT_MUTEXES is not set
1438# CONFIG_RT_MUTEX_TESTER is not set
1439# CONFIG_DEBUG_SPINLOCK is not set
1440# CONFIG_DEBUG_MUTEXES is not set
1441# CONFIG_DEBUG_RWSEMS is not set
1442# CONFIG_DEBUG_LOCK_ALLOC is not set
1443# CONFIG_PROVE_LOCKING is not set
1444# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
1445# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
1446# CONFIG_DEBUG_KOBJECT is not set
1447# CONFIG_DEBUG_HIGHMEM is not set
1523CONFIG_DEBUG_BUGVERBOSE=y 1448CONFIG_DEBUG_BUGVERBOSE=y
1449# CONFIG_DEBUG_INFO is not set
1450# CONFIG_DEBUG_FS is not set
1451# CONFIG_DEBUG_VM is not set
1452# CONFIG_FRAME_POINTER is not set
1453CONFIG_UNWIND_INFO=y
1454CONFIG_STACK_UNWIND=y
1455# CONFIG_FORCED_INLINING is not set
1456# CONFIG_RCU_TORTURE_TEST is not set
1524CONFIG_EARLY_PRINTK=y 1457CONFIG_EARLY_PRINTK=y
1458CONFIG_DEBUG_STACKOVERFLOW=y
1459# CONFIG_DEBUG_STACK_USAGE is not set
1460# CONFIG_DEBUG_RODATA is not set
1461# CONFIG_4KSTACKS is not set
1525CONFIG_X86_FIND_SMP_CONFIG=y 1462CONFIG_X86_FIND_SMP_CONFIG=y
1526CONFIG_X86_MPPARSE=y 1463CONFIG_X86_MPPARSE=y
1464CONFIG_DOUBLEFAULT=y
1527 1465
1528# 1466#
1529# Security options 1467# Security options
@@ -1537,10 +1475,6 @@ CONFIG_X86_MPPARSE=y
1537# CONFIG_CRYPTO is not set 1475# CONFIG_CRYPTO is not set
1538 1476
1539# 1477#
1540# Hardware crypto devices
1541#
1542
1543#
1544# Library routines 1478# Library routines
1545# 1479#
1546# CONFIG_CRC_CCITT is not set 1480# CONFIG_CRC_CCITT is not set
@@ -1548,7 +1482,12 @@ CONFIG_X86_MPPARSE=y
1548CONFIG_CRC32=y 1482CONFIG_CRC32=y
1549# CONFIG_LIBCRC32C is not set 1483# CONFIG_LIBCRC32C is not set
1550CONFIG_ZLIB_INFLATE=y 1484CONFIG_ZLIB_INFLATE=y
1485CONFIG_PLIST=y
1551CONFIG_GENERIC_HARDIRQS=y 1486CONFIG_GENERIC_HARDIRQS=y
1552CONFIG_GENERIC_IRQ_PROBE=y 1487CONFIG_GENERIC_IRQ_PROBE=y
1488CONFIG_GENERIC_PENDING_IRQ=y
1489CONFIG_X86_SMP=y
1490CONFIG_X86_HT=y
1553CONFIG_X86_BIOS_REBOOT=y 1491CONFIG_X86_BIOS_REBOOT=y
1492CONFIG_X86_TRAMPOLINE=y
1554CONFIG_KTIME_SCALAR=y 1493CONFIG_KTIME_SCALAR=y
diff --git a/arch/i386/kernel/Makefile b/arch/i386/kernel/Makefile
index 5427a842e841..1a884b6e6e5c 100644
--- a/arch/i386/kernel/Makefile
+++ b/arch/i386/kernel/Makefile
@@ -4,7 +4,7 @@
4 4
5extra-y := head.o init_task.o vmlinux.lds 5extra-y := head.o init_task.o vmlinux.lds
6 6
7obj-y := process.o semaphore.o signal.o entry.o traps.o irq.o \ 7obj-y := process.o signal.o entry.o traps.o irq.o \
8 ptrace.o time.o ioport.o ldt.o setup.o i8259.o sys_i386.o \ 8 ptrace.o time.o ioport.o ldt.o setup.o i8259.o sys_i386.o \
9 pci-dma.o i386_ksyms.o i387.o bootflag.o \ 9 pci-dma.o i386_ksyms.o i387.o bootflag.o \
10 quirks.o i8237.o topology.o alternative.o i8253.o tsc.o 10 quirks.o i8237.o topology.o alternative.o i8253.o tsc.o
@@ -81,4 +81,5 @@ $(obj)/vsyscall-syms.o: $(src)/vsyscall.lds \
81 $(call if_changed,syscall) 81 $(call if_changed,syscall)
82 82
83k8-y += ../../x86_64/kernel/k8.o 83k8-y += ../../x86_64/kernel/k8.o
84stacktrace-y += ../../x86_64/kernel/stacktrace.o
84 85
diff --git a/arch/i386/kernel/acpi/Makefile b/arch/i386/kernel/acpi/Makefile
index 7e9ac99354f4..7f7be01f44e6 100644
--- a/arch/i386/kernel/acpi/Makefile
+++ b/arch/i386/kernel/acpi/Makefile
@@ -1,5 +1,7 @@
1obj-$(CONFIG_ACPI) += boot.o 1obj-$(CONFIG_ACPI) += boot.o
2ifneq ($(CONFIG_PCI),)
2obj-$(CONFIG_X86_IO_APIC) += earlyquirk.o 3obj-$(CONFIG_X86_IO_APIC) += earlyquirk.o
4endif
3obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup.o 5obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup.o
4 6
5ifneq ($(CONFIG_ACPI_PROCESSOR),) 7ifneq ($(CONFIG_ACPI_PROCESSOR),)
diff --git a/arch/i386/kernel/acpi/boot.c b/arch/i386/kernel/acpi/boot.c
index ee003bc0e8b1..1aaea6ab8c46 100644
--- a/arch/i386/kernel/acpi/boot.c
+++ b/arch/i386/kernel/acpi/boot.c
@@ -26,9 +26,12 @@
26#include <linux/init.h> 26#include <linux/init.h>
27#include <linux/acpi.h> 27#include <linux/acpi.h>
28#include <linux/efi.h> 28#include <linux/efi.h>
29#include <linux/cpumask.h>
29#include <linux/module.h> 30#include <linux/module.h>
30#include <linux/dmi.h> 31#include <linux/dmi.h>
31#include <linux/irq.h> 32#include <linux/irq.h>
33#include <linux/bootmem.h>
34#include <linux/ioport.h>
32 35
33#include <asm/pgtable.h> 36#include <asm/pgtable.h>
34#include <asm/io_apic.h> 37#include <asm/io_apic.h>
@@ -36,11 +39,17 @@
36#include <asm/io.h> 39#include <asm/io.h>
37#include <asm/mpspec.h> 40#include <asm/mpspec.h>
38 41
39#ifdef CONFIG_X86_64 42static int __initdata acpi_force = 0;
40 43
41extern void __init clustered_apic_check(void); 44#ifdef CONFIG_ACPI
45int acpi_disabled = 0;
46#else
47int acpi_disabled = 1;
48#endif
49EXPORT_SYMBOL(acpi_disabled);
50
51#ifdef CONFIG_X86_64
42 52
43extern int gsi_irq_sharing(int gsi);
44#include <asm/proto.h> 53#include <asm/proto.h>
45 54
46static inline int acpi_madt_oem_check(char *oem_id, char *oem_table_id) { return 0; } 55static inline int acpi_madt_oem_check(char *oem_id, char *oem_table_id) { return 0; }
@@ -506,16 +515,76 @@ EXPORT_SYMBOL(acpi_register_gsi);
506#ifdef CONFIG_ACPI_HOTPLUG_CPU 515#ifdef CONFIG_ACPI_HOTPLUG_CPU
507int acpi_map_lsapic(acpi_handle handle, int *pcpu) 516int acpi_map_lsapic(acpi_handle handle, int *pcpu)
508{ 517{
509 /* TBD */ 518 struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
510 return -EINVAL; 519 union acpi_object *obj;
520 struct acpi_table_lapic *lapic;
521 cpumask_t tmp_map, new_map;
522 u8 physid;
523 int cpu;
524
525 if (ACPI_FAILURE(acpi_evaluate_object(handle, "_MAT", NULL, &buffer)))
526 return -EINVAL;
527
528 if (!buffer.length || !buffer.pointer)
529 return -EINVAL;
530
531 obj = buffer.pointer;
532 if (obj->type != ACPI_TYPE_BUFFER ||
533 obj->buffer.length < sizeof(*lapic)) {
534 kfree(buffer.pointer);
535 return -EINVAL;
536 }
537
538 lapic = (struct acpi_table_lapic *)obj->buffer.pointer;
539
540 if ((lapic->header.type != ACPI_MADT_LAPIC) ||
541 (!lapic->flags.enabled)) {
542 kfree(buffer.pointer);
543 return -EINVAL;
544 }
545
546 physid = lapic->id;
547
548 kfree(buffer.pointer);
549 buffer.length = ACPI_ALLOCATE_BUFFER;
550 buffer.pointer = NULL;
551
552 tmp_map = cpu_present_map;
553 mp_register_lapic(physid, lapic->flags.enabled);
554
555 /*
556 * If mp_register_lapic successfully generates a new logical cpu
557 * number, then the following will get us exactly what was mapped
558 */
559 cpus_andnot(new_map, cpu_present_map, tmp_map);
560 if (cpus_empty(new_map)) {
561 printk ("Unable to map lapic to logical cpu number\n");
562 return -EINVAL;
563 }
564
565 cpu = first_cpu(new_map);
566
567 *pcpu = cpu;
568 return 0;
511} 569}
512 570
513EXPORT_SYMBOL(acpi_map_lsapic); 571EXPORT_SYMBOL(acpi_map_lsapic);
514 572
515int acpi_unmap_lsapic(int cpu) 573int acpi_unmap_lsapic(int cpu)
516{ 574{
517 /* TBD */ 575 int i;
518 return -EINVAL; 576
577 for_each_possible_cpu(i) {
578 if (x86_acpiid_to_apicid[i] == x86_cpu_to_apicid[cpu]) {
579 x86_acpiid_to_apicid[i] = -1;
580 break;
581 }
582 }
583 x86_cpu_to_apicid[cpu] = -1;
584 cpu_clear(cpu, cpu_present_map);
585 num_processors--;
586
587 return (0);
519} 588}
520 589
521EXPORT_SYMBOL(acpi_unmap_lsapic); 590EXPORT_SYMBOL(acpi_unmap_lsapic);
@@ -579,6 +648,8 @@ static int __init acpi_parse_sbf(unsigned long phys_addr, unsigned long size)
579static int __init acpi_parse_hpet(unsigned long phys, unsigned long size) 648static int __init acpi_parse_hpet(unsigned long phys, unsigned long size)
580{ 649{
581 struct acpi_table_hpet *hpet_tbl; 650 struct acpi_table_hpet *hpet_tbl;
651 struct resource *hpet_res;
652 resource_size_t res_start;
582 653
583 if (!phys || !size) 654 if (!phys || !size)
584 return -EINVAL; 655 return -EINVAL;
@@ -594,12 +665,26 @@ static int __init acpi_parse_hpet(unsigned long phys, unsigned long size)
594 "memory.\n"); 665 "memory.\n");
595 return -1; 666 return -1;
596 } 667 }
668
669#define HPET_RESOURCE_NAME_SIZE 9
670 hpet_res = alloc_bootmem(sizeof(*hpet_res) + HPET_RESOURCE_NAME_SIZE);
671 if (hpet_res) {
672 memset(hpet_res, 0, sizeof(*hpet_res));
673 hpet_res->name = (void *)&hpet_res[1];
674 hpet_res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
675 snprintf((char *)hpet_res->name, HPET_RESOURCE_NAME_SIZE,
676 "HPET %u", hpet_tbl->number);
677 hpet_res->end = (1 * 1024) - 1;
678 }
679
597#ifdef CONFIG_X86_64 680#ifdef CONFIG_X86_64
598 vxtime.hpet_address = hpet_tbl->addr.addrl | 681 vxtime.hpet_address = hpet_tbl->addr.addrl |
599 ((long)hpet_tbl->addr.addrh << 32); 682 ((long)hpet_tbl->addr.addrh << 32);
600 683
601 printk(KERN_INFO PREFIX "HPET id: %#x base: %#lx\n", 684 printk(KERN_INFO PREFIX "HPET id: %#x base: %#lx\n",
602 hpet_tbl->id, vxtime.hpet_address); 685 hpet_tbl->id, vxtime.hpet_address);
686
687 res_start = vxtime.hpet_address;
603#else /* X86 */ 688#else /* X86 */
604 { 689 {
605 extern unsigned long hpet_address; 690 extern unsigned long hpet_address;
@@ -607,9 +692,17 @@ static int __init acpi_parse_hpet(unsigned long phys, unsigned long size)
607 hpet_address = hpet_tbl->addr.addrl; 692 hpet_address = hpet_tbl->addr.addrl;
608 printk(KERN_INFO PREFIX "HPET id: %#x base: %#lx\n", 693 printk(KERN_INFO PREFIX "HPET id: %#x base: %#lx\n",
609 hpet_tbl->id, hpet_address); 694 hpet_tbl->id, hpet_address);
695
696 res_start = hpet_address;
610 } 697 }
611#endif /* X86 */ 698#endif /* X86 */
612 699
700 if (hpet_res) {
701 hpet_res->start = res_start;
702 hpet_res->end += res_start;
703 insert_resource(&iomem_resource, hpet_res);
704 }
705
613 return 0; 706 return 0;
614} 707}
615#else 708#else
@@ -860,8 +953,6 @@ static void __init acpi_process_madt(void)
860 return; 953 return;
861} 954}
862 955
863extern int acpi_force;
864
865#ifdef __i386__ 956#ifdef __i386__
866 957
867static int __init disable_acpi_irq(struct dmi_system_id *d) 958static int __init disable_acpi_irq(struct dmi_system_id *d)
@@ -1163,3 +1254,75 @@ int __init acpi_boot_init(void)
1163 1254
1164 return 0; 1255 return 0;
1165} 1256}
1257
1258static int __init parse_acpi(char *arg)
1259{
1260 if (!arg)
1261 return -EINVAL;
1262
1263 /* "acpi=off" disables both ACPI table parsing and interpreter */
1264 if (strcmp(arg, "off") == 0) {
1265 disable_acpi();
1266 }
1267 /* acpi=force to over-ride black-list */
1268 else if (strcmp(arg, "force") == 0) {
1269 acpi_force = 1;
1270 acpi_ht = 1;
1271 acpi_disabled = 0;
1272 }
1273 /* acpi=strict disables out-of-spec workarounds */
1274 else if (strcmp(arg, "strict") == 0) {
1275 acpi_strict = 1;
1276 }
1277 /* Limit ACPI just to boot-time to enable HT */
1278 else if (strcmp(arg, "ht") == 0) {
1279 if (!acpi_force)
1280 disable_acpi();
1281 acpi_ht = 1;
1282 }
1283 /* "acpi=noirq" disables ACPI interrupt routing */
1284 else if (strcmp(arg, "noirq") == 0) {
1285 acpi_noirq_set();
1286 } else {
1287 /* Core will printk when we return error. */
1288 return -EINVAL;
1289 }
1290 return 0;
1291}
1292early_param("acpi", parse_acpi);
1293
1294/* FIXME: Using pci= for an ACPI parameter is a travesty. */
1295static int __init parse_pci(char *arg)
1296{
1297 if (arg && strcmp(arg, "noacpi") == 0)
1298 acpi_disable_pci();
1299 return 0;
1300}
1301early_param("pci", parse_pci);
1302
1303#ifdef CONFIG_X86_IO_APIC
1304static int __init parse_acpi_skip_timer_override(char *arg)
1305{
1306 acpi_skip_timer_override = 1;
1307 return 0;
1308}
1309early_param("acpi_skip_timer_override", parse_acpi_skip_timer_override);
1310#endif /* CONFIG_X86_IO_APIC */
1311
1312static int __init setup_acpi_sci(char *s)
1313{
1314 if (!s)
1315 return -EINVAL;
1316 if (!strcmp(s, "edge"))
1317 acpi_sci_flags.trigger = 1;
1318 else if (!strcmp(s, "level"))
1319 acpi_sci_flags.trigger = 3;
1320 else if (!strcmp(s, "high"))
1321 acpi_sci_flags.polarity = 1;
1322 else if (!strcmp(s, "low"))
1323 acpi_sci_flags.polarity = 3;
1324 else
1325 return -EINVAL;
1326 return 0;
1327}
1328early_param("acpi_sci", setup_acpi_sci);
diff --git a/arch/i386/kernel/acpi/earlyquirk.c b/arch/i386/kernel/acpi/earlyquirk.c
index 1649a175a206..fe799b11ac0a 100644
--- a/arch/i386/kernel/acpi/earlyquirk.c
+++ b/arch/i386/kernel/acpi/earlyquirk.c
@@ -48,7 +48,11 @@ void __init check_acpi_pci(void)
48 int num, slot, func; 48 int num, slot, func;
49 49
50 /* Assume the machine supports type 1. If not it will 50 /* Assume the machine supports type 1. If not it will
51 always read ffffffff and should not have any side effect. */ 51 always read ffffffff and should not have any side effect.
52 Actually a few buggy systems can machine check. Allow the user
53 to disable it by command line option at least -AK */
54 if (!early_pci_allowed())
55 return;
52 56
53 /* Poor man's PCI discovery */ 57 /* Poor man's PCI discovery */
54 for (num = 0; num < 32; num++) { 58 for (num = 0; num < 32; num++) {
diff --git a/arch/i386/kernel/apic.c b/arch/i386/kernel/apic.c
index 8c844d07862f..90faae5c5d30 100644
--- a/arch/i386/kernel/apic.c
+++ b/arch/i386/kernel/apic.c
@@ -52,7 +52,18 @@ static cpumask_t timer_bcast_ipi;
52/* 52/*
53 * Knob to control our willingness to enable the local APIC. 53 * Knob to control our willingness to enable the local APIC.
54 */ 54 */
55int enable_local_apic __initdata = 0; /* -1=force-disable, +1=force-enable */ 55static int enable_local_apic __initdata = 0; /* -1=force-disable, +1=force-enable */
56
57static inline void lapic_disable(void)
58{
59 enable_local_apic = -1;
60 clear_bit(X86_FEATURE_APIC, boot_cpu_data.x86_capability);
61}
62
63static inline void lapic_enable(void)
64{
65 enable_local_apic = 1;
66}
56 67
57/* 68/*
58 * Debug level 69 * Debug level
@@ -586,8 +597,7 @@ void __devinit setup_local_APIC(void)
586 printk("No ESR for 82489DX.\n"); 597 printk("No ESR for 82489DX.\n");
587 } 598 }
588 599
589 if (nmi_watchdog == NMI_LOCAL_APIC) 600 setup_apic_nmi_watchdog(NULL);
590 setup_apic_nmi_watchdog();
591 apic_pm_activate(); 601 apic_pm_activate();
592} 602}
593 603
@@ -1373,3 +1383,18 @@ int __init APIC_init_uniprocessor (void)
1373 1383
1374 return 0; 1384 return 0;
1375} 1385}
1386
1387static int __init parse_lapic(char *arg)
1388{
1389 lapic_enable();
1390 return 0;
1391}
1392early_param("lapic", parse_lapic);
1393
1394static int __init parse_nolapic(char *arg)
1395{
1396 lapic_disable();
1397 return 0;
1398}
1399early_param("nolapic", parse_nolapic);
1400
diff --git a/arch/i386/kernel/apm.c b/arch/i386/kernel/apm.c
index 8591f2fa920c..ff9ce4b5eaa8 100644
--- a/arch/i386/kernel/apm.c
+++ b/arch/i386/kernel/apm.c
@@ -1154,9 +1154,11 @@ out:
1154 1154
1155static void set_time(void) 1155static void set_time(void)
1156{ 1156{
1157 struct timespec ts;
1157 if (got_clock_diff) { /* Must know time zone in order to set clock */ 1158 if (got_clock_diff) { /* Must know time zone in order to set clock */
1158 xtime.tv_sec = get_cmos_time() + clock_cmos_diff; 1159 ts.tv_sec = get_cmos_time() + clock_cmos_diff;
1159 xtime.tv_nsec = 0; 1160 ts.tv_nsec = 0;
1161 do_settimeofday(&ts);
1160 } 1162 }
1161} 1163}
1162 1164
@@ -1232,13 +1234,8 @@ static int suspend(int vetoable)
1232 restore_processor_state(); 1234 restore_processor_state();
1233 1235
1234 local_irq_disable(); 1236 local_irq_disable();
1235 write_seqlock(&xtime_lock);
1236 spin_lock(&i8253_lock);
1237 reinit_timer();
1238 set_time(); 1237 set_time();
1239 1238 reinit_timer();
1240 spin_unlock(&i8253_lock);
1241 write_sequnlock(&xtime_lock);
1242 1239
1243 if (err == APM_NO_ERROR) 1240 if (err == APM_NO_ERROR)
1244 err = APM_SUCCESS; 1241 err = APM_SUCCESS;
@@ -1365,9 +1362,7 @@ static void check_events(void)
1365 ignore_bounce = 1; 1362 ignore_bounce = 1;
1366 if ((event != APM_NORMAL_RESUME) 1363 if ((event != APM_NORMAL_RESUME)
1367 || (ignore_normal_resume == 0)) { 1364 || (ignore_normal_resume == 0)) {
1368 write_seqlock_irq(&xtime_lock);
1369 set_time(); 1365 set_time();
1370 write_sequnlock_irq(&xtime_lock);
1371 device_resume(); 1366 device_resume();
1372 pm_send_all(PM_RESUME, (void *)0); 1367 pm_send_all(PM_RESUME, (void *)0);
1373 queue_event(event, NULL); 1368 queue_event(event, NULL);
@@ -1383,9 +1378,7 @@ static void check_events(void)
1383 break; 1378 break;
1384 1379
1385 case APM_UPDATE_TIME: 1380 case APM_UPDATE_TIME:
1386 write_seqlock_irq(&xtime_lock);
1387 set_time(); 1381 set_time();
1388 write_sequnlock_irq(&xtime_lock);
1389 break; 1382 break;
1390 1383
1391 case APM_CRITICAL_SUSPEND: 1384 case APM_CRITICAL_SUSPEND:
@@ -2339,6 +2332,7 @@ static int __init apm_init(void)
2339 ret = kernel_thread(apm, NULL, CLONE_KERNEL | SIGCHLD); 2332 ret = kernel_thread(apm, NULL, CLONE_KERNEL | SIGCHLD);
2340 if (ret < 0) { 2333 if (ret < 0) {
2341 printk(KERN_ERR "apm: disabled - Unable to start kernel thread.\n"); 2334 printk(KERN_ERR "apm: disabled - Unable to start kernel thread.\n");
2335 remove_proc_entry("apm", NULL);
2342 return -ENOMEM; 2336 return -ENOMEM;
2343 } 2337 }
2344 2338
@@ -2348,7 +2342,13 @@ static int __init apm_init(void)
2348 return 0; 2342 return 0;
2349 } 2343 }
2350 2344
2351 misc_register(&apm_device); 2345 /*
2346 * Note we don't actually care if the misc_device cannot be registered.
2347 * this driver can do its job without it, even if userspace can't
2348 * control it. just log the error
2349 */
2350 if (misc_register(&apm_device))
2351 printk(KERN_WARNING "apm: Could not register misc device.\n");
2352 2352
2353 if (HZ != 100) 2353 if (HZ != 100)
2354 idle_period = (idle_period * HZ) / 100; 2354 idle_period = (idle_period * HZ) / 100;
diff --git a/arch/i386/kernel/cpu/amd.c b/arch/i386/kernel/cpu/amd.c
index e6a2d6b80cda..e4758095d87a 100644
--- a/arch/i386/kernel/cpu/amd.c
+++ b/arch/i386/kernel/cpu/amd.c
@@ -22,7 +22,7 @@
22extern void vide(void); 22extern void vide(void);
23__asm__(".align 4\nvide: ret"); 23__asm__(".align 4\nvide: ret");
24 24
25static void __init init_amd(struct cpuinfo_x86 *c) 25static void __cpuinit init_amd(struct cpuinfo_x86 *c)
26{ 26{
27 u32 l, h; 27 u32 l, h;
28 int mbytes = num_physpages >> (20-PAGE_SHIFT); 28 int mbytes = num_physpages >> (20-PAGE_SHIFT);
@@ -246,7 +246,7 @@ static void __init init_amd(struct cpuinfo_x86 *c)
246 num_cache_leaves = 3; 246 num_cache_leaves = 3;
247} 247}
248 248
249static unsigned int amd_size_cache(struct cpuinfo_x86 * c, unsigned int size) 249static unsigned int __cpuinit amd_size_cache(struct cpuinfo_x86 * c, unsigned int size)
250{ 250{
251 /* AMD errata T13 (order #21922) */ 251 /* AMD errata T13 (order #21922) */
252 if ((c->x86 == 6)) { 252 if ((c->x86 == 6)) {
@@ -259,7 +259,7 @@ static unsigned int amd_size_cache(struct cpuinfo_x86 * c, unsigned int size)
259 return size; 259 return size;
260} 260}
261 261
262static struct cpu_dev amd_cpu_dev __initdata = { 262static struct cpu_dev amd_cpu_dev __cpuinitdata = {
263 .c_vendor = "AMD", 263 .c_vendor = "AMD",
264 .c_ident = { "AuthenticAMD" }, 264 .c_ident = { "AuthenticAMD" },
265 .c_models = { 265 .c_models = {
@@ -275,7 +275,6 @@ static struct cpu_dev amd_cpu_dev __initdata = {
275 }, 275 },
276 }, 276 },
277 .c_init = init_amd, 277 .c_init = init_amd,
278 .c_identify = generic_identify,
279 .c_size_cache = amd_size_cache, 278 .c_size_cache = amd_size_cache,
280}; 279};
281 280
diff --git a/arch/i386/kernel/cpu/centaur.c b/arch/i386/kernel/cpu/centaur.c
index bd75629dd262..8c25047975c0 100644
--- a/arch/i386/kernel/cpu/centaur.c
+++ b/arch/i386/kernel/cpu/centaur.c
@@ -9,7 +9,7 @@
9 9
10#ifdef CONFIG_X86_OOSTORE 10#ifdef CONFIG_X86_OOSTORE
11 11
12static u32 __init power2(u32 x) 12static u32 __cpuinit power2(u32 x)
13{ 13{
14 u32 s=1; 14 u32 s=1;
15 while(s<=x) 15 while(s<=x)
@@ -22,7 +22,7 @@ static u32 __init power2(u32 x)
22 * Set up an actual MCR 22 * Set up an actual MCR
23 */ 23 */
24 24
25static void __init centaur_mcr_insert(int reg, u32 base, u32 size, int key) 25static void __cpuinit centaur_mcr_insert(int reg, u32 base, u32 size, int key)
26{ 26{
27 u32 lo, hi; 27 u32 lo, hi;
28 28
@@ -40,7 +40,7 @@ static void __init centaur_mcr_insert(int reg, u32 base, u32 size, int key)
40 * Shortcut: We know you can't put 4Gig of RAM on a winchip 40 * Shortcut: We know you can't put 4Gig of RAM on a winchip
41 */ 41 */
42 42
43static u32 __init ramtop(void) /* 16388 */ 43static u32 __cpuinit ramtop(void) /* 16388 */
44{ 44{
45 int i; 45 int i;
46 u32 top = 0; 46 u32 top = 0;
@@ -91,7 +91,7 @@ static u32 __init ramtop(void) /* 16388 */
91 * Compute a set of MCR's to give maximum coverage 91 * Compute a set of MCR's to give maximum coverage
92 */ 92 */
93 93
94static int __init centaur_mcr_compute(int nr, int key) 94static int __cpuinit centaur_mcr_compute(int nr, int key)
95{ 95{
96 u32 mem = ramtop(); 96 u32 mem = ramtop();
97 u32 root = power2(mem); 97 u32 root = power2(mem);
@@ -166,7 +166,7 @@ static int __init centaur_mcr_compute(int nr, int key)
166 return ct; 166 return ct;
167} 167}
168 168
169static void __init centaur_create_optimal_mcr(void) 169static void __cpuinit centaur_create_optimal_mcr(void)
170{ 170{
171 int i; 171 int i;
172 /* 172 /*
@@ -189,7 +189,7 @@ static void __init centaur_create_optimal_mcr(void)
189 wrmsr(MSR_IDT_MCR0+i, 0, 0); 189 wrmsr(MSR_IDT_MCR0+i, 0, 0);
190} 190}
191 191
192static void __init winchip2_create_optimal_mcr(void) 192static void __cpuinit winchip2_create_optimal_mcr(void)
193{ 193{
194 u32 lo, hi; 194 u32 lo, hi;
195 int i; 195 int i;
@@ -227,7 +227,7 @@ static void __init winchip2_create_optimal_mcr(void)
227 * Handle the MCR key on the Winchip 2. 227 * Handle the MCR key on the Winchip 2.
228 */ 228 */
229 229
230static void __init winchip2_unprotect_mcr(void) 230static void __cpuinit winchip2_unprotect_mcr(void)
231{ 231{
232 u32 lo, hi; 232 u32 lo, hi;
233 u32 key; 233 u32 key;
@@ -239,7 +239,7 @@ static void __init winchip2_unprotect_mcr(void)
239 wrmsr(MSR_IDT_MCR_CTRL, lo, hi); 239 wrmsr(MSR_IDT_MCR_CTRL, lo, hi);
240} 240}
241 241
242static void __init winchip2_protect_mcr(void) 242static void __cpuinit winchip2_protect_mcr(void)
243{ 243{
244 u32 lo, hi; 244 u32 lo, hi;
245 245
@@ -257,7 +257,7 @@ static void __init winchip2_protect_mcr(void)
257#define RNG_ENABLED (1 << 3) 257#define RNG_ENABLED (1 << 3)
258#define RNG_ENABLE (1 << 6) /* MSR_VIA_RNG */ 258#define RNG_ENABLE (1 << 6) /* MSR_VIA_RNG */
259 259
260static void __init init_c3(struct cpuinfo_x86 *c) 260static void __cpuinit init_c3(struct cpuinfo_x86 *c)
261{ 261{
262 u32 lo, hi; 262 u32 lo, hi;
263 263
@@ -303,7 +303,7 @@ static void __init init_c3(struct cpuinfo_x86 *c)
303 display_cacheinfo(c); 303 display_cacheinfo(c);
304} 304}
305 305
306static void __init init_centaur(struct cpuinfo_x86 *c) 306static void __cpuinit init_centaur(struct cpuinfo_x86 *c)
307{ 307{
308 enum { 308 enum {
309 ECX8=1<<1, 309 ECX8=1<<1,
@@ -442,7 +442,7 @@ static void __init init_centaur(struct cpuinfo_x86 *c)
442 } 442 }
443} 443}
444 444
445static unsigned int centaur_size_cache(struct cpuinfo_x86 * c, unsigned int size) 445static unsigned int __cpuinit centaur_size_cache(struct cpuinfo_x86 * c, unsigned int size)
446{ 446{
447 /* VIA C3 CPUs (670-68F) need further shifting. */ 447 /* VIA C3 CPUs (670-68F) need further shifting. */
448 if ((c->x86 == 6) && ((c->x86_model == 7) || (c->x86_model == 8))) 448 if ((c->x86 == 6) && ((c->x86_model == 7) || (c->x86_model == 8)))
@@ -457,7 +457,7 @@ static unsigned int centaur_size_cache(struct cpuinfo_x86 * c, unsigned int size
457 return size; 457 return size;
458} 458}
459 459
460static struct cpu_dev centaur_cpu_dev __initdata = { 460static struct cpu_dev centaur_cpu_dev __cpuinitdata = {
461 .c_vendor = "Centaur", 461 .c_vendor = "Centaur",
462 .c_ident = { "CentaurHauls" }, 462 .c_ident = { "CentaurHauls" },
463 .c_init = init_centaur, 463 .c_init = init_centaur,
diff --git a/arch/i386/kernel/cpu/common.c b/arch/i386/kernel/cpu/common.c
index 70c87de582c7..2799baaadf45 100644
--- a/arch/i386/kernel/cpu/common.c
+++ b/arch/i386/kernel/cpu/common.c
@@ -36,7 +36,7 @@ struct cpu_dev * cpu_devs[X86_VENDOR_NUM] = {};
36 36
37extern int disable_pse; 37extern int disable_pse;
38 38
39static void default_init(struct cpuinfo_x86 * c) 39static void __cpuinit default_init(struct cpuinfo_x86 * c)
40{ 40{
41 /* Not much we can do here... */ 41 /* Not much we can do here... */
42 /* Check if at least it has cpuid */ 42 /* Check if at least it has cpuid */
@@ -49,7 +49,7 @@ static void default_init(struct cpuinfo_x86 * c)
49 } 49 }
50} 50}
51 51
52static struct cpu_dev default_cpu = { 52static struct cpu_dev __cpuinitdata default_cpu = {
53 .c_init = default_init, 53 .c_init = default_init,
54 .c_vendor = "Unknown", 54 .c_vendor = "Unknown",
55}; 55};
@@ -265,7 +265,7 @@ static void __init early_cpu_detect(void)
265 } 265 }
266} 266}
267 267
268void __cpuinit generic_identify(struct cpuinfo_x86 * c) 268static void __cpuinit generic_identify(struct cpuinfo_x86 * c)
269{ 269{
270 u32 tfms, xlvl; 270 u32 tfms, xlvl;
271 int ebx; 271 int ebx;
@@ -675,7 +675,7 @@ old_gdt:
675#endif 675#endif
676 676
677 /* Clear %fs and %gs. */ 677 /* Clear %fs and %gs. */
678 asm volatile ("xorl %eax, %eax; movl %eax, %fs; movl %eax, %gs"); 678 asm volatile ("movl %0, %%fs; movl %0, %%gs" : : "r" (0));
679 679
680 /* Clear all 6 debug registers: */ 680 /* Clear all 6 debug registers: */
681 set_debugreg(0, 0); 681 set_debugreg(0, 0);
diff --git a/arch/i386/kernel/cpu/cpu.h b/arch/i386/kernel/cpu/cpu.h
index 5a1d4f163e84..2f6432cef6ff 100644
--- a/arch/i386/kernel/cpu/cpu.h
+++ b/arch/i386/kernel/cpu/cpu.h
@@ -24,7 +24,5 @@ extern struct cpu_dev * cpu_devs [X86_VENDOR_NUM];
24extern int get_model_name(struct cpuinfo_x86 *c); 24extern int get_model_name(struct cpuinfo_x86 *c);
25extern void display_cacheinfo(struct cpuinfo_x86 *c); 25extern void display_cacheinfo(struct cpuinfo_x86 *c);
26 26
27extern void generic_identify(struct cpuinfo_x86 * c);
28
29extern void early_intel_workaround(struct cpuinfo_x86 *c); 27extern void early_intel_workaround(struct cpuinfo_x86 *c);
30 28
diff --git a/arch/i386/kernel/cpu/cyrix.c b/arch/i386/kernel/cpu/cyrix.c
index f03b7f94c304..c0c3b59de32c 100644
--- a/arch/i386/kernel/cpu/cyrix.c
+++ b/arch/i386/kernel/cpu/cyrix.c
@@ -12,7 +12,7 @@
12/* 12/*
13 * Read NSC/Cyrix DEVID registers (DIR) to get more detailed info. about the CPU 13 * Read NSC/Cyrix DEVID registers (DIR) to get more detailed info. about the CPU
14 */ 14 */
15static void __init do_cyrix_devid(unsigned char *dir0, unsigned char *dir1) 15static void __cpuinit do_cyrix_devid(unsigned char *dir0, unsigned char *dir1)
16{ 16{
17 unsigned char ccr2, ccr3; 17 unsigned char ccr2, ccr3;
18 unsigned long flags; 18 unsigned long flags;
@@ -52,25 +52,25 @@ static void __init do_cyrix_devid(unsigned char *dir0, unsigned char *dir1)
52 * Actually since bugs.h doesn't even reference this perhaps someone should 52 * Actually since bugs.h doesn't even reference this perhaps someone should
53 * fix the documentation ??? 53 * fix the documentation ???
54 */ 54 */
55static unsigned char Cx86_dir0_msb __initdata = 0; 55static unsigned char Cx86_dir0_msb __cpuinitdata = 0;
56 56
57static char Cx86_model[][9] __initdata = { 57static char Cx86_model[][9] __cpuinitdata = {
58 "Cx486", "Cx486", "5x86 ", "6x86", "MediaGX ", "6x86MX ", 58 "Cx486", "Cx486", "5x86 ", "6x86", "MediaGX ", "6x86MX ",
59 "M II ", "Unknown" 59 "M II ", "Unknown"
60}; 60};
61static char Cx486_name[][5] __initdata = { 61static char Cx486_name[][5] __cpuinitdata = {
62 "SLC", "DLC", "SLC2", "DLC2", "SRx", "DRx", 62 "SLC", "DLC", "SLC2", "DLC2", "SRx", "DRx",
63 "SRx2", "DRx2" 63 "SRx2", "DRx2"
64}; 64};
65static char Cx486S_name[][4] __initdata = { 65static char Cx486S_name[][4] __cpuinitdata = {
66 "S", "S2", "Se", "S2e" 66 "S", "S2", "Se", "S2e"
67}; 67};
68static char Cx486D_name[][4] __initdata = { 68static char Cx486D_name[][4] __cpuinitdata = {
69 "DX", "DX2", "?", "?", "?", "DX4" 69 "DX", "DX2", "?", "?", "?", "DX4"
70}; 70};
71static char Cx86_cb[] __initdata = "?.5x Core/Bus Clock"; 71static char Cx86_cb[] __cpuinitdata = "?.5x Core/Bus Clock";
72static char cyrix_model_mult1[] __initdata = "12??43"; 72static char cyrix_model_mult1[] __cpuinitdata = "12??43";
73static char cyrix_model_mult2[] __initdata = "12233445"; 73static char cyrix_model_mult2[] __cpuinitdata = "12233445";
74 74
75/* 75/*
76 * Reset the slow-loop (SLOP) bit on the 686(L) which is set by some old 76 * Reset the slow-loop (SLOP) bit on the 686(L) which is set by some old
@@ -82,7 +82,7 @@ static char cyrix_model_mult2[] __initdata = "12233445";
82 82
83extern void calibrate_delay(void) __init; 83extern void calibrate_delay(void) __init;
84 84
85static void __init check_cx686_slop(struct cpuinfo_x86 *c) 85static void __cpuinit check_cx686_slop(struct cpuinfo_x86 *c)
86{ 86{
87 unsigned long flags; 87 unsigned long flags;
88 88
@@ -107,7 +107,7 @@ static void __init check_cx686_slop(struct cpuinfo_x86 *c)
107} 107}
108 108
109 109
110static void __init set_cx86_reorder(void) 110static void __cpuinit set_cx86_reorder(void)
111{ 111{
112 u8 ccr3; 112 u8 ccr3;
113 113
@@ -122,7 +122,7 @@ static void __init set_cx86_reorder(void)
122 setCx86(CX86_CCR3, ccr3); 122 setCx86(CX86_CCR3, ccr3);
123} 123}
124 124
125static void __init set_cx86_memwb(void) 125static void __cpuinit set_cx86_memwb(void)
126{ 126{
127 u32 cr0; 127 u32 cr0;
128 128
@@ -137,7 +137,7 @@ static void __init set_cx86_memwb(void)
137 setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x14 ); 137 setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x14 );
138} 138}
139 139
140static void __init set_cx86_inc(void) 140static void __cpuinit set_cx86_inc(void)
141{ 141{
142 unsigned char ccr3; 142 unsigned char ccr3;
143 143
@@ -158,7 +158,7 @@ static void __init set_cx86_inc(void)
158 * Configure later MediaGX and/or Geode processor. 158 * Configure later MediaGX and/or Geode processor.
159 */ 159 */
160 160
161static void __init geode_configure(void) 161static void __cpuinit geode_configure(void)
162{ 162{
163 unsigned long flags; 163 unsigned long flags;
164 u8 ccr3, ccr4; 164 u8 ccr3, ccr4;
@@ -184,14 +184,14 @@ static void __init geode_configure(void)
184 184
185 185
186#ifdef CONFIG_PCI 186#ifdef CONFIG_PCI
187static struct pci_device_id __initdata cyrix_55x0[] = { 187static struct pci_device_id __cpuinitdata cyrix_55x0[] = {
188 { PCI_DEVICE(PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5510) }, 188 { PCI_DEVICE(PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5510) },
189 { PCI_DEVICE(PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5520) }, 189 { PCI_DEVICE(PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5520) },
190 { }, 190 { },
191}; 191};
192#endif 192#endif
193 193
194static void __init init_cyrix(struct cpuinfo_x86 *c) 194static void __cpuinit init_cyrix(struct cpuinfo_x86 *c)
195{ 195{
196 unsigned char dir0, dir0_msn, dir0_lsn, dir1 = 0; 196 unsigned char dir0, dir0_msn, dir0_lsn, dir1 = 0;
197 char *buf = c->x86_model_id; 197 char *buf = c->x86_model_id;
@@ -346,7 +346,7 @@ static void __init init_cyrix(struct cpuinfo_x86 *c)
346/* 346/*
347 * Handle National Semiconductor branded processors 347 * Handle National Semiconductor branded processors
348 */ 348 */
349static void __init init_nsc(struct cpuinfo_x86 *c) 349static void __cpuinit init_nsc(struct cpuinfo_x86 *c)
350{ 350{
351 /* There may be GX1 processors in the wild that are branded 351 /* There may be GX1 processors in the wild that are branded
352 * NSC and not Cyrix. 352 * NSC and not Cyrix.
@@ -394,7 +394,7 @@ static inline int test_cyrix_52div(void)
394 return (unsigned char) (test >> 8) == 0x02; 394 return (unsigned char) (test >> 8) == 0x02;
395} 395}
396 396
397static void cyrix_identify(struct cpuinfo_x86 * c) 397static void __cpuinit cyrix_identify(struct cpuinfo_x86 * c)
398{ 398{
399 /* Detect Cyrix with disabled CPUID */ 399 /* Detect Cyrix with disabled CPUID */
400 if ( c->x86 == 4 && test_cyrix_52div() ) { 400 if ( c->x86 == 4 && test_cyrix_52div() ) {
@@ -427,10 +427,9 @@ static void cyrix_identify(struct cpuinfo_x86 * c)
427 local_irq_restore(flags); 427 local_irq_restore(flags);
428 } 428 }
429 } 429 }
430 generic_identify(c);
431} 430}
432 431
433static struct cpu_dev cyrix_cpu_dev __initdata = { 432static struct cpu_dev cyrix_cpu_dev __cpuinitdata = {
434 .c_vendor = "Cyrix", 433 .c_vendor = "Cyrix",
435 .c_ident = { "CyrixInstead" }, 434 .c_ident = { "CyrixInstead" },
436 .c_init = init_cyrix, 435 .c_init = init_cyrix,
@@ -453,11 +452,10 @@ static int __init cyrix_exit_cpu(void)
453 452
454late_initcall(cyrix_exit_cpu); 453late_initcall(cyrix_exit_cpu);
455 454
456static struct cpu_dev nsc_cpu_dev __initdata = { 455static struct cpu_dev nsc_cpu_dev __cpuinitdata = {
457 .c_vendor = "NSC", 456 .c_vendor = "NSC",
458 .c_ident = { "Geode by NSC" }, 457 .c_ident = { "Geode by NSC" },
459 .c_init = init_nsc, 458 .c_init = init_nsc,
460 .c_identify = generic_identify,
461}; 459};
462 460
463int __init nsc_init_cpu(void) 461int __init nsc_init_cpu(void)
diff --git a/arch/i386/kernel/cpu/intel.c b/arch/i386/kernel/cpu/intel.c
index 5a2e270924b1..94a95aa5227e 100644
--- a/arch/i386/kernel/cpu/intel.c
+++ b/arch/i386/kernel/cpu/intel.c
@@ -198,7 +198,7 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c)
198} 198}
199 199
200 200
201static unsigned int intel_size_cache(struct cpuinfo_x86 * c, unsigned int size) 201static unsigned int __cpuinit intel_size_cache(struct cpuinfo_x86 * c, unsigned int size)
202{ 202{
203 /* Intel PIII Tualatin. This comes in two flavours. 203 /* Intel PIII Tualatin. This comes in two flavours.
204 * One has 256kb of cache, the other 512. We have no way 204 * One has 256kb of cache, the other 512. We have no way
@@ -263,7 +263,6 @@ static struct cpu_dev intel_cpu_dev __cpuinitdata = {
263 }, 263 },
264 }, 264 },
265 .c_init = init_intel, 265 .c_init = init_intel,
266 .c_identify = generic_identify,
267 .c_size_cache = intel_size_cache, 266 .c_size_cache = intel_size_cache,
268}; 267};
269 268
diff --git a/arch/i386/kernel/cpu/mcheck/Makefile b/arch/i386/kernel/cpu/mcheck/Makefile
index 30808f3d6715..f1ebe1c1c17a 100644
--- a/arch/i386/kernel/cpu/mcheck/Makefile
+++ b/arch/i386/kernel/cpu/mcheck/Makefile
@@ -1,2 +1,2 @@
1obj-y = mce.o k7.o p4.o p5.o p6.o winchip.o 1obj-y = mce.o k7.o p4.o p5.o p6.o winchip.o therm_throt.o
2obj-$(CONFIG_X86_MCE_NONFATAL) += non-fatal.o 2obj-$(CONFIG_X86_MCE_NONFATAL) += non-fatal.o
diff --git a/arch/i386/kernel/cpu/mcheck/p4.c b/arch/i386/kernel/cpu/mcheck/p4.c
index b95f1b3d53aa..504434a46011 100644
--- a/arch/i386/kernel/cpu/mcheck/p4.c
+++ b/arch/i386/kernel/cpu/mcheck/p4.c
@@ -13,6 +13,8 @@
13#include <asm/msr.h> 13#include <asm/msr.h>
14#include <asm/apic.h> 14#include <asm/apic.h>
15 15
16#include <asm/therm_throt.h>
17
16#include "mce.h" 18#include "mce.h"
17 19
18/* as supported by the P4/Xeon family */ 20/* as supported by the P4/Xeon family */
@@ -44,25 +46,12 @@ static void unexpected_thermal_interrupt(struct pt_regs *regs)
44/* P4/Xeon Thermal transition interrupt handler */ 46/* P4/Xeon Thermal transition interrupt handler */
45static void intel_thermal_interrupt(struct pt_regs *regs) 47static void intel_thermal_interrupt(struct pt_regs *regs)
46{ 48{
47 u32 l, h; 49 __u64 msr_val;
48 unsigned int cpu = smp_processor_id();
49 static unsigned long next[NR_CPUS];
50 50
51 ack_APIC_irq(); 51 ack_APIC_irq();
52 52
53 if (time_after(next[cpu], jiffies)) 53 rdmsrl(MSR_IA32_THERM_STATUS, msr_val);
54 return; 54 therm_throt_process(msr_val & 0x1);
55
56 next[cpu] = jiffies + HZ*5;
57 rdmsr(MSR_IA32_THERM_STATUS, l, h);
58 if (l & 0x1) {
59 printk(KERN_EMERG "CPU%d: Temperature above threshold\n", cpu);
60 printk(KERN_EMERG "CPU%d: Running in modulated clock mode\n",
61 cpu);
62 add_taint(TAINT_MACHINE_CHECK);
63 } else {
64 printk(KERN_INFO "CPU%d: Temperature/speed normal\n", cpu);
65 }
66} 55}
67 56
68/* Thermal interrupt handler for this CPU setup */ 57/* Thermal interrupt handler for this CPU setup */
@@ -122,10 +111,13 @@ static void intel_init_thermal(struct cpuinfo_x86 *c)
122 111
123 rdmsr (MSR_IA32_MISC_ENABLE, l, h); 112 rdmsr (MSR_IA32_MISC_ENABLE, l, h);
124 wrmsr (MSR_IA32_MISC_ENABLE, l | (1<<3), h); 113 wrmsr (MSR_IA32_MISC_ENABLE, l | (1<<3), h);
125 114
126 l = apic_read (APIC_LVTTHMR); 115 l = apic_read (APIC_LVTTHMR);
127 apic_write_around (APIC_LVTTHMR, l & ~APIC_LVT_MASKED); 116 apic_write_around (APIC_LVTTHMR, l & ~APIC_LVT_MASKED);
128 printk (KERN_INFO "CPU%d: Thermal monitoring enabled\n", cpu); 117 printk (KERN_INFO "CPU%d: Thermal monitoring enabled\n", cpu);
118
119 /* enable thermal throttle processing */
120 atomic_set(&therm_throt_en, 1);
129 return; 121 return;
130} 122}
131#endif /* CONFIG_X86_MCE_P4THERMAL */ 123#endif /* CONFIG_X86_MCE_P4THERMAL */
diff --git a/arch/i386/kernel/cpu/mcheck/therm_throt.c b/arch/i386/kernel/cpu/mcheck/therm_throt.c
new file mode 100644
index 000000000000..4f43047de406
--- /dev/null
+++ b/arch/i386/kernel/cpu/mcheck/therm_throt.c
@@ -0,0 +1,180 @@
1/*
2 * linux/arch/i386/kerne/cpu/mcheck/therm_throt.c
3 *
4 * Thermal throttle event support code (such as syslog messaging and rate
5 * limiting) that was factored out from x86_64 (mce_intel.c) and i386 (p4.c).
6 * This allows consistent reporting of CPU thermal throttle events.
7 *
8 * Maintains a counter in /sys that keeps track of the number of thermal
9 * events, such that the user knows how bad the thermal problem might be
10 * (since the logging to syslog and mcelog is rate limited).
11 *
12 * Author: Dmitriy Zavin (dmitriyz@google.com)
13 *
14 * Credits: Adapted from Zwane Mwaikambo's original code in mce_intel.c.
15 * Inspired by Ross Biro's and Al Borchers' counter code.
16 */
17
18#include <linux/percpu.h>
19#include <linux/sysdev.h>
20#include <linux/cpu.h>
21#include <asm/cpu.h>
22#include <linux/notifier.h>
23#include <asm/therm_throt.h>
24
25/* How long to wait between reporting thermal events */
26#define CHECK_INTERVAL (300 * HZ)
27
28static DEFINE_PER_CPU(__u64, next_check) = INITIAL_JIFFIES;
29static DEFINE_PER_CPU(unsigned long, thermal_throttle_count);
30atomic_t therm_throt_en = ATOMIC_INIT(0);
31
32#ifdef CONFIG_SYSFS
33#define define_therm_throt_sysdev_one_ro(_name) \
34 static SYSDEV_ATTR(_name, 0444, therm_throt_sysdev_show_##_name, NULL)
35
36#define define_therm_throt_sysdev_show_func(name) \
37static ssize_t therm_throt_sysdev_show_##name(struct sys_device *dev, \
38 char *buf) \
39{ \
40 unsigned int cpu = dev->id; \
41 ssize_t ret; \
42 \
43 preempt_disable(); /* CPU hotplug */ \
44 if (cpu_online(cpu)) \
45 ret = sprintf(buf, "%lu\n", \
46 per_cpu(thermal_throttle_##name, cpu)); \
47 else \
48 ret = 0; \
49 preempt_enable(); \
50 \
51 return ret; \
52}
53
54define_therm_throt_sysdev_show_func(count);
55define_therm_throt_sysdev_one_ro(count);
56
57static struct attribute *thermal_throttle_attrs[] = {
58 &attr_count.attr,
59 NULL
60};
61
62static struct attribute_group thermal_throttle_attr_group = {
63 .attrs = thermal_throttle_attrs,
64 .name = "thermal_throttle"
65};
66#endif /* CONFIG_SYSFS */
67
68/***
69 * therm_throt_process - Process thermal throttling event from interrupt
70 * @curr: Whether the condition is current or not (boolean), since the
71 * thermal interrupt normally gets called both when the thermal
72 * event begins and once the event has ended.
73 *
74 * This function is called by the thermal interrupt after the
75 * IRQ has been acknowledged.
76 *
77 * It will take care of rate limiting and printing messages to the syslog.
78 *
79 * Returns: 0 : Event should NOT be further logged, i.e. still in
80 * "timeout" from previous log message.
81 * 1 : Event should be logged further, and a message has been
82 * printed to the syslog.
83 */
84int therm_throt_process(int curr)
85{
86 unsigned int cpu = smp_processor_id();
87 __u64 tmp_jiffs = get_jiffies_64();
88
89 if (curr)
90 __get_cpu_var(thermal_throttle_count)++;
91
92 if (time_before64(tmp_jiffs, __get_cpu_var(next_check)))
93 return 0;
94
95 __get_cpu_var(next_check) = tmp_jiffs + CHECK_INTERVAL;
96
97 /* if we just entered the thermal event */
98 if (curr) {
99 printk(KERN_CRIT "CPU%d: Temperature above threshold, "
100 "cpu clock throttled (total events = %lu)\n", cpu,
101 __get_cpu_var(thermal_throttle_count));
102
103 add_taint(TAINT_MACHINE_CHECK);
104 } else {
105 printk(KERN_CRIT "CPU%d: Temperature/speed normal\n", cpu);
106 }
107
108 return 1;
109}
110
111#ifdef CONFIG_SYSFS
112/* Add/Remove thermal_throttle interface for CPU device */
113static __cpuinit int thermal_throttle_add_dev(struct sys_device * sys_dev)
114{
115 sysfs_create_group(&sys_dev->kobj, &thermal_throttle_attr_group);
116 return 0;
117}
118
119#ifdef CONFIG_HOTPLUG_CPU
120static __cpuinit int thermal_throttle_remove_dev(struct sys_device * sys_dev)
121{
122 sysfs_remove_group(&sys_dev->kobj, &thermal_throttle_attr_group);
123 return 0;
124}
125
126/* Mutex protecting device creation against CPU hotplug */
127static DEFINE_MUTEX(therm_cpu_lock);
128
129/* Get notified when a cpu comes on/off. Be hotplug friendly. */
130static __cpuinit int thermal_throttle_cpu_callback(struct notifier_block *nfb,
131 unsigned long action,
132 void *hcpu)
133{
134 unsigned int cpu = (unsigned long)hcpu;
135 struct sys_device *sys_dev;
136
137 sys_dev = get_cpu_sysdev(cpu);
138 mutex_lock(&therm_cpu_lock);
139 switch (action) {
140 case CPU_ONLINE:
141 thermal_throttle_add_dev(sys_dev);
142 break;
143 case CPU_DEAD:
144 thermal_throttle_remove_dev(sys_dev);
145 break;
146 }
147 mutex_unlock(&therm_cpu_lock);
148 return NOTIFY_OK;
149}
150
151static struct notifier_block thermal_throttle_cpu_notifier =
152{
153 .notifier_call = thermal_throttle_cpu_callback,
154};
155#endif /* CONFIG_HOTPLUG_CPU */
156
157static __init int thermal_throttle_init_device(void)
158{
159 unsigned int cpu = 0;
160
161 if (!atomic_read(&therm_throt_en))
162 return 0;
163
164 register_hotcpu_notifier(&thermal_throttle_cpu_notifier);
165
166#ifdef CONFIG_HOTPLUG_CPU
167 mutex_lock(&therm_cpu_lock);
168#endif
169 /* connect live CPUs to sysfs */
170 for_each_online_cpu(cpu)
171 thermal_throttle_add_dev(get_cpu_sysdev(cpu));
172#ifdef CONFIG_HOTPLUG_CPU
173 mutex_unlock(&therm_cpu_lock);
174#endif
175
176 return 0;
177}
178
179device_initcall(thermal_throttle_init_device);
180#endif /* CONFIG_SYSFS */
diff --git a/arch/i386/kernel/cpu/mtrr/generic.c b/arch/i386/kernel/cpu/mtrr/generic.c
index 169ac8e0db68..0b61eed8bbd8 100644
--- a/arch/i386/kernel/cpu/mtrr/generic.c
+++ b/arch/i386/kernel/cpu/mtrr/generic.c
@@ -243,7 +243,7 @@ static DEFINE_SPINLOCK(set_atomicity_lock);
243 * has been called. 243 * has been called.
244 */ 244 */
245 245
246static void prepare_set(void) 246static void prepare_set(void) __acquires(set_atomicity_lock)
247{ 247{
248 unsigned long cr0; 248 unsigned long cr0;
249 249
@@ -274,7 +274,7 @@ static void prepare_set(void)
274 mtrr_wrmsr(MTRRdefType_MSR, deftype_lo & 0xf300UL, deftype_hi); 274 mtrr_wrmsr(MTRRdefType_MSR, deftype_lo & 0xf300UL, deftype_hi);
275} 275}
276 276
277static void post_set(void) 277static void post_set(void) __releases(set_atomicity_lock)
278{ 278{
279 /* Flush TLBs (no need to flush caches - they are disabled) */ 279 /* Flush TLBs (no need to flush caches - they are disabled) */
280 __flush_tlb(); 280 __flush_tlb();
diff --git a/arch/i386/kernel/cpu/nexgen.c b/arch/i386/kernel/cpu/nexgen.c
index ad87fa58058d..8bf23cc80c63 100644
--- a/arch/i386/kernel/cpu/nexgen.c
+++ b/arch/i386/kernel/cpu/nexgen.c
@@ -10,7 +10,7 @@
10 * to have CPUID. (Thanks to Herbert Oppmann) 10 * to have CPUID. (Thanks to Herbert Oppmann)
11 */ 11 */
12 12
13static int __init deep_magic_nexgen_probe(void) 13static int __cpuinit deep_magic_nexgen_probe(void)
14{ 14{
15 int ret; 15 int ret;
16 16
@@ -27,21 +27,20 @@ static int __init deep_magic_nexgen_probe(void)
27 return ret; 27 return ret;
28} 28}
29 29
30static void __init init_nexgen(struct cpuinfo_x86 * c) 30static void __cpuinit init_nexgen(struct cpuinfo_x86 * c)
31{ 31{
32 c->x86_cache_size = 256; /* A few had 1 MB... */ 32 c->x86_cache_size = 256; /* A few had 1 MB... */
33} 33}
34 34
35static void __init nexgen_identify(struct cpuinfo_x86 * c) 35static void __cpuinit nexgen_identify(struct cpuinfo_x86 * c)
36{ 36{
37 /* Detect NexGen with old hypercode */ 37 /* Detect NexGen with old hypercode */
38 if ( deep_magic_nexgen_probe() ) { 38 if ( deep_magic_nexgen_probe() ) {
39 strcpy(c->x86_vendor_id, "NexGenDriven"); 39 strcpy(c->x86_vendor_id, "NexGenDriven");
40 } 40 }
41 generic_identify(c);
42} 41}
43 42
44static struct cpu_dev nexgen_cpu_dev __initdata = { 43static struct cpu_dev nexgen_cpu_dev __cpuinitdata = {
45 .c_vendor = "Nexgen", 44 .c_vendor = "Nexgen",
46 .c_ident = { "NexGenDriven" }, 45 .c_ident = { "NexGenDriven" },
47 .c_models = { 46 .c_models = {
diff --git a/arch/i386/kernel/cpu/proc.c b/arch/i386/kernel/cpu/proc.c
index f54a15268ed7..76aac088a323 100644
--- a/arch/i386/kernel/cpu/proc.c
+++ b/arch/i386/kernel/cpu/proc.c
@@ -46,8 +46,8 @@ static int show_cpuinfo(struct seq_file *m, void *v)
46 46
47 /* Intel-defined (#2) */ 47 /* Intel-defined (#2) */
48 "pni", NULL, NULL, "monitor", "ds_cpl", "vmx", "smx", "est", 48 "pni", NULL, NULL, "monitor", "ds_cpl", "vmx", "smx", "est",
49 "tm2", NULL, "cid", NULL, NULL, "cx16", "xtpr", NULL, 49 "tm2", "ssse3", "cid", NULL, NULL, "cx16", "xtpr", NULL,
50 NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, 50 NULL, NULL, "dca", NULL, NULL, NULL, NULL, NULL,
51 NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, 51 NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
52 52
53 /* VIA/Cyrix/Centaur-defined */ 53 /* VIA/Cyrix/Centaur-defined */
diff --git a/arch/i386/kernel/cpu/rise.c b/arch/i386/kernel/cpu/rise.c
index d08d5a2811c8..9317f7414989 100644
--- a/arch/i386/kernel/cpu/rise.c
+++ b/arch/i386/kernel/cpu/rise.c
@@ -5,7 +5,7 @@
5 5
6#include "cpu.h" 6#include "cpu.h"
7 7
8static void __init init_rise(struct cpuinfo_x86 *c) 8static void __cpuinit init_rise(struct cpuinfo_x86 *c)
9{ 9{
10 printk("CPU: Rise iDragon"); 10 printk("CPU: Rise iDragon");
11 if (c->x86_model > 2) 11 if (c->x86_model > 2)
@@ -28,7 +28,7 @@ static void __init init_rise(struct cpuinfo_x86 *c)
28 set_bit(X86_FEATURE_CX8, c->x86_capability); 28 set_bit(X86_FEATURE_CX8, c->x86_capability);
29} 29}
30 30
31static struct cpu_dev rise_cpu_dev __initdata = { 31static struct cpu_dev rise_cpu_dev __cpuinitdata = {
32 .c_vendor = "Rise", 32 .c_vendor = "Rise",
33 .c_ident = { "RiseRiseRise" }, 33 .c_ident = { "RiseRiseRise" },
34 .c_models = { 34 .c_models = {
diff --git a/arch/i386/kernel/cpu/transmeta.c b/arch/i386/kernel/cpu/transmeta.c
index 7214c9b577ab..4056fb7d2cdf 100644
--- a/arch/i386/kernel/cpu/transmeta.c
+++ b/arch/i386/kernel/cpu/transmeta.c
@@ -5,7 +5,7 @@
5#include <asm/msr.h> 5#include <asm/msr.h>
6#include "cpu.h" 6#include "cpu.h"
7 7
8static void __init init_transmeta(struct cpuinfo_x86 *c) 8static void __cpuinit init_transmeta(struct cpuinfo_x86 *c)
9{ 9{
10 unsigned int cap_mask, uk, max, dummy; 10 unsigned int cap_mask, uk, max, dummy;
11 unsigned int cms_rev1, cms_rev2; 11 unsigned int cms_rev1, cms_rev2;
@@ -85,10 +85,9 @@ static void __init init_transmeta(struct cpuinfo_x86 *c)
85#endif 85#endif
86} 86}
87 87
88static void __init transmeta_identify(struct cpuinfo_x86 * c) 88static void __cpuinit transmeta_identify(struct cpuinfo_x86 * c)
89{ 89{
90 u32 xlvl; 90 u32 xlvl;
91 generic_identify(c);
92 91
93 /* Transmeta-defined flags: level 0x80860001 */ 92 /* Transmeta-defined flags: level 0x80860001 */
94 xlvl = cpuid_eax(0x80860000); 93 xlvl = cpuid_eax(0x80860000);
@@ -98,7 +97,7 @@ static void __init transmeta_identify(struct cpuinfo_x86 * c)
98 } 97 }
99} 98}
100 99
101static struct cpu_dev transmeta_cpu_dev __initdata = { 100static struct cpu_dev transmeta_cpu_dev __cpuinitdata = {
102 .c_vendor = "Transmeta", 101 .c_vendor = "Transmeta",
103 .c_ident = { "GenuineTMx86", "TransmetaCPU" }, 102 .c_ident = { "GenuineTMx86", "TransmetaCPU" },
104 .c_init = init_transmeta, 103 .c_init = init_transmeta,
diff --git a/arch/i386/kernel/cpu/umc.c b/arch/i386/kernel/cpu/umc.c
index 2cd988f6dc55..1bf3f87e9c5b 100644
--- a/arch/i386/kernel/cpu/umc.c
+++ b/arch/i386/kernel/cpu/umc.c
@@ -5,12 +5,8 @@
5 5
6/* UMC chips appear to be only either 386 or 486, so no special init takes place. 6/* UMC chips appear to be only either 386 or 486, so no special init takes place.
7 */ 7 */
8static void __init init_umc(struct cpuinfo_x86 * c)
9{
10
11}
12 8
13static struct cpu_dev umc_cpu_dev __initdata = { 9static struct cpu_dev umc_cpu_dev __cpuinitdata = {
14 .c_vendor = "UMC", 10 .c_vendor = "UMC",
15 .c_ident = { "UMC UMC UMC" }, 11 .c_ident = { "UMC UMC UMC" },
16 .c_models = { 12 .c_models = {
@@ -21,7 +17,6 @@ static struct cpu_dev umc_cpu_dev __initdata = {
21 } 17 }
22 }, 18 },
23 }, 19 },
24 .c_init = init_umc,
25}; 20};
26 21
27int __init umc_init_cpu(void) 22int __init umc_init_cpu(void)
diff --git a/arch/i386/kernel/crash.c b/arch/i386/kernel/crash.c
index 5b96f038367f..67d297dc1003 100644
--- a/arch/i386/kernel/crash.c
+++ b/arch/i386/kernel/crash.c
@@ -22,6 +22,8 @@
22#include <asm/nmi.h> 22#include <asm/nmi.h>
23#include <asm/hw_irq.h> 23#include <asm/hw_irq.h>
24#include <asm/apic.h> 24#include <asm/apic.h>
25#include <asm/kdebug.h>
26
25#include <mach_ipi.h> 27#include <mach_ipi.h>
26 28
27 29
@@ -93,16 +95,25 @@ static void crash_save_self(struct pt_regs *regs)
93#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC) 95#if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC)
94static atomic_t waiting_for_crash_ipi; 96static atomic_t waiting_for_crash_ipi;
95 97
96static int crash_nmi_callback(struct pt_regs *regs, int cpu) 98static int crash_nmi_callback(struct notifier_block *self,
99 unsigned long val, void *data)
97{ 100{
101 struct pt_regs *regs;
98 struct pt_regs fixed_regs; 102 struct pt_regs fixed_regs;
103 int cpu;
104
105 if (val != DIE_NMI_IPI)
106 return NOTIFY_OK;
107
108 regs = ((struct die_args *)data)->regs;
109 cpu = raw_smp_processor_id();
99 110
100 /* Don't do anything if this handler is invoked on crashing cpu. 111 /* Don't do anything if this handler is invoked on crashing cpu.
101 * Otherwise, system will completely hang. Crashing cpu can get 112 * Otherwise, system will completely hang. Crashing cpu can get
102 * an NMI if system was initially booted with nmi_watchdog parameter. 113 * an NMI if system was initially booted with nmi_watchdog parameter.
103 */ 114 */
104 if (cpu == crashing_cpu) 115 if (cpu == crashing_cpu)
105 return 1; 116 return NOTIFY_STOP;
106 local_irq_disable(); 117 local_irq_disable();
107 118
108 if (!user_mode_vm(regs)) { 119 if (!user_mode_vm(regs)) {
@@ -125,13 +136,18 @@ static void smp_send_nmi_allbutself(void)
125 send_IPI_allbutself(NMI_VECTOR); 136 send_IPI_allbutself(NMI_VECTOR);
126} 137}
127 138
139static struct notifier_block crash_nmi_nb = {
140 .notifier_call = crash_nmi_callback,
141};
142
128static void nmi_shootdown_cpus(void) 143static void nmi_shootdown_cpus(void)
129{ 144{
130 unsigned long msecs; 145 unsigned long msecs;
131 146
132 atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1); 147 atomic_set(&waiting_for_crash_ipi, num_online_cpus() - 1);
133 /* Would it be better to replace the trap vector here? */ 148 /* Would it be better to replace the trap vector here? */
134 set_nmi_callback(crash_nmi_callback); 149 if (register_die_notifier(&crash_nmi_nb))
150 return; /* return what? */
135 /* Ensure the new callback function is set before sending 151 /* Ensure the new callback function is set before sending
136 * out the NMI 152 * out the NMI
137 */ 153 */
diff --git a/arch/i386/kernel/efi_stub.S b/arch/i386/kernel/efi_stub.S
index d3ee73a3eee3..ef00bb77d7e4 100644
--- a/arch/i386/kernel/efi_stub.S
+++ b/arch/i386/kernel/efi_stub.S
@@ -7,7 +7,6 @@
7 7
8#include <linux/linkage.h> 8#include <linux/linkage.h>
9#include <asm/page.h> 9#include <asm/page.h>
10#include <asm/pgtable.h>
11 10
12/* 11/*
13 * efi_call_phys(void *, ...) is a function with variable parameters. 12 * efi_call_phys(void *, ...) is a function with variable parameters.
diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
index 87f9f60b803b..5a63d6fdb70e 100644
--- a/arch/i386/kernel/entry.S
+++ b/arch/i386/kernel/entry.S
@@ -76,8 +76,15 @@ DF_MASK = 0x00000400
76NT_MASK = 0x00004000 76NT_MASK = 0x00004000
77VM_MASK = 0x00020000 77VM_MASK = 0x00020000
78 78
79/* These are replaces for paravirtualization */
80#define DISABLE_INTERRUPTS cli
81#define ENABLE_INTERRUPTS sti
82#define ENABLE_INTERRUPTS_SYSEXIT sti; sysexit
83#define INTERRUPT_RETURN iret
84#define GET_CR0_INTO_EAX movl %cr0, %eax
85
79#ifdef CONFIG_PREEMPT 86#ifdef CONFIG_PREEMPT
80#define preempt_stop cli; TRACE_IRQS_OFF 87#define preempt_stop DISABLE_INTERRUPTS; TRACE_IRQS_OFF
81#else 88#else
82#define preempt_stop 89#define preempt_stop
83#define resume_kernel restore_nocheck 90#define resume_kernel restore_nocheck
@@ -176,18 +183,21 @@ VM_MASK = 0x00020000
176 183
177#define RING0_INT_FRAME \ 184#define RING0_INT_FRAME \
178 CFI_STARTPROC simple;\ 185 CFI_STARTPROC simple;\
186 CFI_SIGNAL_FRAME;\
179 CFI_DEF_CFA esp, 3*4;\ 187 CFI_DEF_CFA esp, 3*4;\
180 /*CFI_OFFSET cs, -2*4;*/\ 188 /*CFI_OFFSET cs, -2*4;*/\
181 CFI_OFFSET eip, -3*4 189 CFI_OFFSET eip, -3*4
182 190
183#define RING0_EC_FRAME \ 191#define RING0_EC_FRAME \
184 CFI_STARTPROC simple;\ 192 CFI_STARTPROC simple;\
193 CFI_SIGNAL_FRAME;\
185 CFI_DEF_CFA esp, 4*4;\ 194 CFI_DEF_CFA esp, 4*4;\
186 /*CFI_OFFSET cs, -2*4;*/\ 195 /*CFI_OFFSET cs, -2*4;*/\
187 CFI_OFFSET eip, -3*4 196 CFI_OFFSET eip, -3*4
188 197
189#define RING0_PTREGS_FRAME \ 198#define RING0_PTREGS_FRAME \
190 CFI_STARTPROC simple;\ 199 CFI_STARTPROC simple;\
200 CFI_SIGNAL_FRAME;\
191 CFI_DEF_CFA esp, OLDESP-EBX;\ 201 CFI_DEF_CFA esp, OLDESP-EBX;\
192 /*CFI_OFFSET cs, CS-OLDESP;*/\ 202 /*CFI_OFFSET cs, CS-OLDESP;*/\
193 CFI_OFFSET eip, EIP-OLDESP;\ 203 CFI_OFFSET eip, EIP-OLDESP;\
@@ -233,10 +243,11 @@ ret_from_intr:
233check_userspace: 243check_userspace:
234 movl EFLAGS(%esp), %eax # mix EFLAGS and CS 244 movl EFLAGS(%esp), %eax # mix EFLAGS and CS
235 movb CS(%esp), %al 245 movb CS(%esp), %al
236 testl $(VM_MASK | 3), %eax 246 andl $(VM_MASK | SEGMENT_RPL_MASK), %eax
237 jz resume_kernel 247 cmpl $USER_RPL, %eax
248 jb resume_kernel # not returning to v8086 or userspace
238ENTRY(resume_userspace) 249ENTRY(resume_userspace)
239 cli # make sure we don't miss an interrupt 250 DISABLE_INTERRUPTS # make sure we don't miss an interrupt
240 # setting need_resched or sigpending 251 # setting need_resched or sigpending
241 # between sampling and the iret 252 # between sampling and the iret
242 movl TI_flags(%ebp), %ecx 253 movl TI_flags(%ebp), %ecx
@@ -247,7 +258,7 @@ ENTRY(resume_userspace)
247 258
248#ifdef CONFIG_PREEMPT 259#ifdef CONFIG_PREEMPT
249ENTRY(resume_kernel) 260ENTRY(resume_kernel)
250 cli 261 DISABLE_INTERRUPTS
251 cmpl $0,TI_preempt_count(%ebp) # non-zero preempt_count ? 262 cmpl $0,TI_preempt_count(%ebp) # non-zero preempt_count ?
252 jnz restore_nocheck 263 jnz restore_nocheck
253need_resched: 264need_resched:
@@ -267,6 +278,7 @@ need_resched:
267 # sysenter call handler stub 278 # sysenter call handler stub
268ENTRY(sysenter_entry) 279ENTRY(sysenter_entry)
269 CFI_STARTPROC simple 280 CFI_STARTPROC simple
281 CFI_SIGNAL_FRAME
270 CFI_DEF_CFA esp, 0 282 CFI_DEF_CFA esp, 0
271 CFI_REGISTER esp, ebp 283 CFI_REGISTER esp, ebp
272 movl TSS_sysenter_esp0(%esp),%esp 284 movl TSS_sysenter_esp0(%esp),%esp
@@ -275,7 +287,7 @@ sysenter_past_esp:
275 * No need to follow this irqs on/off section: the syscall 287 * No need to follow this irqs on/off section: the syscall
276 * disabled irqs and here we enable it straight after entry: 288 * disabled irqs and here we enable it straight after entry:
277 */ 289 */
278 sti 290 ENABLE_INTERRUPTS
279 pushl $(__USER_DS) 291 pushl $(__USER_DS)
280 CFI_ADJUST_CFA_OFFSET 4 292 CFI_ADJUST_CFA_OFFSET 4
281 /*CFI_REL_OFFSET ss, 0*/ 293 /*CFI_REL_OFFSET ss, 0*/
@@ -320,7 +332,7 @@ sysenter_past_esp:
320 jae syscall_badsys 332 jae syscall_badsys
321 call *sys_call_table(,%eax,4) 333 call *sys_call_table(,%eax,4)
322 movl %eax,EAX(%esp) 334 movl %eax,EAX(%esp)
323 cli 335 DISABLE_INTERRUPTS
324 TRACE_IRQS_OFF 336 TRACE_IRQS_OFF
325 movl TI_flags(%ebp), %ecx 337 movl TI_flags(%ebp), %ecx
326 testw $_TIF_ALLWORK_MASK, %cx 338 testw $_TIF_ALLWORK_MASK, %cx
@@ -330,8 +342,7 @@ sysenter_past_esp:
330 movl OLDESP(%esp), %ecx 342 movl OLDESP(%esp), %ecx
331 xorl %ebp,%ebp 343 xorl %ebp,%ebp
332 TRACE_IRQS_ON 344 TRACE_IRQS_ON
333 sti 345 ENABLE_INTERRUPTS_SYSEXIT
334 sysexit
335 CFI_ENDPROC 346 CFI_ENDPROC
336 347
337 348
@@ -356,7 +367,7 @@ syscall_call:
356 call *sys_call_table(,%eax,4) 367 call *sys_call_table(,%eax,4)
357 movl %eax,EAX(%esp) # store the return value 368 movl %eax,EAX(%esp) # store the return value
358syscall_exit: 369syscall_exit:
359 cli # make sure we don't miss an interrupt 370 DISABLE_INTERRUPTS # make sure we don't miss an interrupt
360 # setting need_resched or sigpending 371 # setting need_resched or sigpending
361 # between sampling and the iret 372 # between sampling and the iret
362 TRACE_IRQS_OFF 373 TRACE_IRQS_OFF
@@ -371,8 +382,8 @@ restore_all:
371 # See comments in process.c:copy_thread() for details. 382 # See comments in process.c:copy_thread() for details.
372 movb OLDSS(%esp), %ah 383 movb OLDSS(%esp), %ah
373 movb CS(%esp), %al 384 movb CS(%esp), %al
374 andl $(VM_MASK | (4 << 8) | 3), %eax 385 andl $(VM_MASK | (SEGMENT_TI_MASK << 8) | SEGMENT_RPL_MASK), %eax
375 cmpl $((4 << 8) | 3), %eax 386 cmpl $((SEGMENT_LDT << 8) | USER_RPL), %eax
376 CFI_REMEMBER_STATE 387 CFI_REMEMBER_STATE
377 je ldt_ss # returning to user-space with LDT SS 388 je ldt_ss # returning to user-space with LDT SS
378restore_nocheck: 389restore_nocheck:
@@ -381,11 +392,11 @@ restore_nocheck_notrace:
381 RESTORE_REGS 392 RESTORE_REGS
382 addl $4, %esp 393 addl $4, %esp
383 CFI_ADJUST_CFA_OFFSET -4 394 CFI_ADJUST_CFA_OFFSET -4
3841: iret 3951: INTERRUPT_RETURN
385.section .fixup,"ax" 396.section .fixup,"ax"
386iret_exc: 397iret_exc:
387 TRACE_IRQS_ON 398 TRACE_IRQS_ON
388 sti 399 ENABLE_INTERRUPTS
389 pushl $0 # no error code 400 pushl $0 # no error code
390 pushl $do_iret_error 401 pushl $do_iret_error
391 jmp error_code 402 jmp error_code
@@ -409,7 +420,7 @@ ldt_ss:
409 * dosemu and wine happy. */ 420 * dosemu and wine happy. */
410 subl $8, %esp # reserve space for switch16 pointer 421 subl $8, %esp # reserve space for switch16 pointer
411 CFI_ADJUST_CFA_OFFSET 8 422 CFI_ADJUST_CFA_OFFSET 8
412 cli 423 DISABLE_INTERRUPTS
413 TRACE_IRQS_OFF 424 TRACE_IRQS_OFF
414 movl %esp, %eax 425 movl %esp, %eax
415 /* Set up the 16bit stack frame with switch32 pointer on top, 426 /* Set up the 16bit stack frame with switch32 pointer on top,
@@ -419,7 +430,7 @@ ldt_ss:
419 TRACE_IRQS_IRET 430 TRACE_IRQS_IRET
420 RESTORE_REGS 431 RESTORE_REGS
421 lss 20+4(%esp), %esp # switch to 16bit stack 432 lss 20+4(%esp), %esp # switch to 16bit stack
4221: iret 4331: INTERRUPT_RETURN
423.section __ex_table,"a" 434.section __ex_table,"a"
424 .align 4 435 .align 4
425 .long 1b,iret_exc 436 .long 1b,iret_exc
@@ -434,7 +445,7 @@ work_pending:
434 jz work_notifysig 445 jz work_notifysig
435work_resched: 446work_resched:
436 call schedule 447 call schedule
437 cli # make sure we don't miss an interrupt 448 DISABLE_INTERRUPTS # make sure we don't miss an interrupt
438 # setting need_resched or sigpending 449 # setting need_resched or sigpending
439 # between sampling and the iret 450 # between sampling and the iret
440 TRACE_IRQS_OFF 451 TRACE_IRQS_OFF
@@ -490,7 +501,7 @@ syscall_exit_work:
490 testb $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SINGLESTEP), %cl 501 testb $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SINGLESTEP), %cl
491 jz work_pending 502 jz work_pending
492 TRACE_IRQS_ON 503 TRACE_IRQS_ON
493 sti # could let do_syscall_trace() call 504 ENABLE_INTERRUPTS # could let do_syscall_trace() call
494 # schedule() instead 505 # schedule() instead
495 movl %esp, %eax 506 movl %esp, %eax
496 movl $1, %edx 507 movl $1, %edx
@@ -591,11 +602,9 @@ ENTRY(name) \
591/* The include is where all of the SMP etc. interrupts come from */ 602/* The include is where all of the SMP etc. interrupts come from */
592#include "entry_arch.h" 603#include "entry_arch.h"
593 604
594ENTRY(divide_error) 605KPROBE_ENTRY(page_fault)
595 RING0_INT_FRAME 606 RING0_EC_FRAME
596 pushl $0 # no error code 607 pushl $do_page_fault
597 CFI_ADJUST_CFA_OFFSET 4
598 pushl $do_divide_error
599 CFI_ADJUST_CFA_OFFSET 4 608 CFI_ADJUST_CFA_OFFSET 4
600 ALIGN 609 ALIGN
601error_code: 610error_code:
@@ -645,6 +654,7 @@ error_code:
645 call *%edi 654 call *%edi
646 jmp ret_from_exception 655 jmp ret_from_exception
647 CFI_ENDPROC 656 CFI_ENDPROC
657KPROBE_END(page_fault)
648 658
649ENTRY(coprocessor_error) 659ENTRY(coprocessor_error)
650 RING0_INT_FRAME 660 RING0_INT_FRAME
@@ -669,7 +679,7 @@ ENTRY(device_not_available)
669 pushl $-1 # mark this as an int 679 pushl $-1 # mark this as an int
670 CFI_ADJUST_CFA_OFFSET 4 680 CFI_ADJUST_CFA_OFFSET 4
671 SAVE_ALL 681 SAVE_ALL
672 movl %cr0, %eax 682 GET_CR0_INTO_EAX
673 testl $0x4, %eax # EM (math emulation bit) 683 testl $0x4, %eax # EM (math emulation bit)
674 jne device_not_available_emulate 684 jne device_not_available_emulate
675 preempt_stop 685 preempt_stop
@@ -702,9 +712,15 @@ device_not_available_emulate:
702 jne ok; \ 712 jne ok; \
703label: \ 713label: \
704 movl TSS_sysenter_esp0+offset(%esp),%esp; \ 714 movl TSS_sysenter_esp0+offset(%esp),%esp; \
715 CFI_DEF_CFA esp, 0; \
716 CFI_UNDEFINED eip; \
705 pushfl; \ 717 pushfl; \
718 CFI_ADJUST_CFA_OFFSET 4; \
706 pushl $__KERNEL_CS; \ 719 pushl $__KERNEL_CS; \
707 pushl $sysenter_past_esp 720 CFI_ADJUST_CFA_OFFSET 4; \
721 pushl $sysenter_past_esp; \
722 CFI_ADJUST_CFA_OFFSET 4; \
723 CFI_REL_OFFSET eip, 0
708 724
709KPROBE_ENTRY(debug) 725KPROBE_ENTRY(debug)
710 RING0_INT_FRAME 726 RING0_INT_FRAME
@@ -720,7 +736,8 @@ debug_stack_correct:
720 call do_debug 736 call do_debug
721 jmp ret_from_exception 737 jmp ret_from_exception
722 CFI_ENDPROC 738 CFI_ENDPROC
723 .previous .text 739KPROBE_END(debug)
740
724/* 741/*
725 * NMI is doubly nasty. It can happen _while_ we're handling 742 * NMI is doubly nasty. It can happen _while_ we're handling
726 * a debug fault, and the debug fault hasn't yet been able to 743 * a debug fault, and the debug fault hasn't yet been able to
@@ -729,7 +746,7 @@ debug_stack_correct:
729 * check whether we got an NMI on the debug path where the debug 746 * check whether we got an NMI on the debug path where the debug
730 * fault happened on the sysenter path. 747 * fault happened on the sysenter path.
731 */ 748 */
732ENTRY(nmi) 749KPROBE_ENTRY(nmi)
733 RING0_INT_FRAME 750 RING0_INT_FRAME
734 pushl %eax 751 pushl %eax
735 CFI_ADJUST_CFA_OFFSET 4 752 CFI_ADJUST_CFA_OFFSET 4
@@ -754,6 +771,7 @@ ENTRY(nmi)
754 cmpl $sysenter_entry,12(%esp) 771 cmpl $sysenter_entry,12(%esp)
755 je nmi_debug_stack_check 772 je nmi_debug_stack_check
756nmi_stack_correct: 773nmi_stack_correct:
774 /* We have a RING0_INT_FRAME here */
757 pushl %eax 775 pushl %eax
758 CFI_ADJUST_CFA_OFFSET 4 776 CFI_ADJUST_CFA_OFFSET 4
759 SAVE_ALL 777 SAVE_ALL
@@ -764,9 +782,12 @@ nmi_stack_correct:
764 CFI_ENDPROC 782 CFI_ENDPROC
765 783
766nmi_stack_fixup: 784nmi_stack_fixup:
785 RING0_INT_FRAME
767 FIX_STACK(12,nmi_stack_correct, 1) 786 FIX_STACK(12,nmi_stack_correct, 1)
768 jmp nmi_stack_correct 787 jmp nmi_stack_correct
788
769nmi_debug_stack_check: 789nmi_debug_stack_check:
790 /* We have a RING0_INT_FRAME here */
770 cmpw $__KERNEL_CS,16(%esp) 791 cmpw $__KERNEL_CS,16(%esp)
771 jne nmi_stack_correct 792 jne nmi_stack_correct
772 cmpl $debug,(%esp) 793 cmpl $debug,(%esp)
@@ -777,8 +798,10 @@ nmi_debug_stack_check:
777 jmp nmi_stack_correct 798 jmp nmi_stack_correct
778 799
779nmi_16bit_stack: 800nmi_16bit_stack:
780 RING0_INT_FRAME 801 /* We have a RING0_INT_FRAME here.
781 /* create the pointer to lss back */ 802 *
803 * create the pointer to lss back
804 */
782 pushl %ss 805 pushl %ss
783 CFI_ADJUST_CFA_OFFSET 4 806 CFI_ADJUST_CFA_OFFSET 4
784 pushl %esp 807 pushl %esp
@@ -799,12 +822,13 @@ nmi_16bit_stack:
799 call do_nmi 822 call do_nmi
800 RESTORE_REGS 823 RESTORE_REGS
801 lss 12+4(%esp), %esp # back to 16bit stack 824 lss 12+4(%esp), %esp # back to 16bit stack
8021: iret 8251: INTERRUPT_RETURN
803 CFI_ENDPROC 826 CFI_ENDPROC
804.section __ex_table,"a" 827.section __ex_table,"a"
805 .align 4 828 .align 4
806 .long 1b,iret_exc 829 .long 1b,iret_exc
807.previous 830.previous
831KPROBE_END(nmi)
808 832
809KPROBE_ENTRY(int3) 833KPROBE_ENTRY(int3)
810 RING0_INT_FRAME 834 RING0_INT_FRAME
@@ -816,7 +840,7 @@ KPROBE_ENTRY(int3)
816 call do_int3 840 call do_int3
817 jmp ret_from_exception 841 jmp ret_from_exception
818 CFI_ENDPROC 842 CFI_ENDPROC
819 .previous .text 843KPROBE_END(int3)
820 844
821ENTRY(overflow) 845ENTRY(overflow)
822 RING0_INT_FRAME 846 RING0_INT_FRAME
@@ -881,7 +905,7 @@ KPROBE_ENTRY(general_protection)
881 CFI_ADJUST_CFA_OFFSET 4 905 CFI_ADJUST_CFA_OFFSET 4
882 jmp error_code 906 jmp error_code
883 CFI_ENDPROC 907 CFI_ENDPROC
884 .previous .text 908KPROBE_END(general_protection)
885 909
886ENTRY(alignment_check) 910ENTRY(alignment_check)
887 RING0_EC_FRAME 911 RING0_EC_FRAME
@@ -890,13 +914,14 @@ ENTRY(alignment_check)
890 jmp error_code 914 jmp error_code
891 CFI_ENDPROC 915 CFI_ENDPROC
892 916
893KPROBE_ENTRY(page_fault) 917ENTRY(divide_error)
894 RING0_EC_FRAME 918 RING0_INT_FRAME
895 pushl $do_page_fault 919 pushl $0 # no error code
920 CFI_ADJUST_CFA_OFFSET 4
921 pushl $do_divide_error
896 CFI_ADJUST_CFA_OFFSET 4 922 CFI_ADJUST_CFA_OFFSET 4
897 jmp error_code 923 jmp error_code
898 CFI_ENDPROC 924 CFI_ENDPROC
899 .previous .text
900 925
901#ifdef CONFIG_X86_MCE 926#ifdef CONFIG_X86_MCE
902ENTRY(machine_check) 927ENTRY(machine_check)
@@ -949,6 +974,19 @@ ENTRY(arch_unwind_init_running)
949ENDPROC(arch_unwind_init_running) 974ENDPROC(arch_unwind_init_running)
950#endif 975#endif
951 976
977ENTRY(kernel_thread_helper)
978 pushl $0 # fake return address for unwinder
979 CFI_STARTPROC
980 movl %edx,%eax
981 push %edx
982 CFI_ADJUST_CFA_OFFSET 4
983 call *%ebx
984 push %eax
985 CFI_ADJUST_CFA_OFFSET 4
986 call do_exit
987 CFI_ENDPROC
988ENDPROC(kernel_thread_helper)
989
952.section .rodata,"a" 990.section .rodata,"a"
953#include "syscall_table.S" 991#include "syscall_table.S"
954 992
diff --git a/arch/i386/kernel/head.S b/arch/i386/kernel/head.S
index a6b8bd89aa27..be9d883c62ce 100644
--- a/arch/i386/kernel/head.S
+++ b/arch/i386/kernel/head.S
@@ -371,8 +371,65 @@ rp_sidt:
371 addl $8,%edi 371 addl $8,%edi
372 dec %ecx 372 dec %ecx
373 jne rp_sidt 373 jne rp_sidt
374
375.macro set_early_handler handler,trapno
376 lea \handler,%edx
377 movl $(__KERNEL_CS << 16),%eax
378 movw %dx,%ax
379 movw $0x8E00,%dx /* interrupt gate - dpl=0, present */
380 lea idt_table,%edi
381 movl %eax,8*\trapno(%edi)
382 movl %edx,8*\trapno+4(%edi)
383.endm
384
385 set_early_handler handler=early_divide_err,trapno=0
386 set_early_handler handler=early_illegal_opcode,trapno=6
387 set_early_handler handler=early_protection_fault,trapno=13
388 set_early_handler handler=early_page_fault,trapno=14
389
374 ret 390 ret
375 391
392early_divide_err:
393 xor %edx,%edx
394 pushl $0 /* fake errcode */
395 jmp early_fault
396
397early_illegal_opcode:
398 movl $6,%edx
399 pushl $0 /* fake errcode */
400 jmp early_fault
401
402early_protection_fault:
403 movl $13,%edx
404 jmp early_fault
405
406early_page_fault:
407 movl $14,%edx
408 jmp early_fault
409
410early_fault:
411 cld
412#ifdef CONFIG_PRINTK
413 movl $(__KERNEL_DS),%eax
414 movl %eax,%ds
415 movl %eax,%es
416 cmpl $2,early_recursion_flag
417 je hlt_loop
418 incl early_recursion_flag
419 movl %cr2,%eax
420 pushl %eax
421 pushl %edx /* trapno */
422 pushl $fault_msg
423#ifdef CONFIG_EARLY_PRINTK
424 call early_printk
425#else
426 call printk
427#endif
428#endif
429hlt_loop:
430 hlt
431 jmp hlt_loop
432
376/* This is the default interrupt "handler" :-) */ 433/* This is the default interrupt "handler" :-) */
377 ALIGN 434 ALIGN
378ignore_int: 435ignore_int:
@@ -386,6 +443,9 @@ ignore_int:
386 movl $(__KERNEL_DS),%eax 443 movl $(__KERNEL_DS),%eax
387 movl %eax,%ds 444 movl %eax,%ds
388 movl %eax,%es 445 movl %eax,%es
446 cmpl $2,early_recursion_flag
447 je hlt_loop
448 incl early_recursion_flag
389 pushl 16(%esp) 449 pushl 16(%esp)
390 pushl 24(%esp) 450 pushl 24(%esp)
391 pushl 32(%esp) 451 pushl 32(%esp)
@@ -431,9 +491,16 @@ ENTRY(stack_start)
431 491
432ready: .byte 0 492ready: .byte 0
433 493
494early_recursion_flag:
495 .long 0
496
434int_msg: 497int_msg:
435 .asciz "Unknown interrupt or fault at EIP %p %p %p\n" 498 .asciz "Unknown interrupt or fault at EIP %p %p %p\n"
436 499
500fault_msg:
501 .ascii "Int %d: CR2 %p err %p EIP %p CS %p flags %p\n"
502 .asciz "Stack: %p %p %p %p %p %p %p %p\n"
503
437/* 504/*
438 * The IDT and GDT 'descriptors' are a strange 48-bit object 505 * The IDT and GDT 'descriptors' are a strange 48-bit object
439 * only used by the lidt and lgdt instructions. They are not 506 * only used by the lidt and lgdt instructions. They are not
diff --git a/arch/i386/kernel/i8259.c b/arch/i386/kernel/i8259.c
index d4756d154f47..ea5f4e7958d8 100644
--- a/arch/i386/kernel/i8259.c
+++ b/arch/i386/kernel/i8259.c
@@ -45,6 +45,8 @@ static void end_8259A_irq (unsigned int irq)
45 45
46#define shutdown_8259A_irq disable_8259A_irq 46#define shutdown_8259A_irq disable_8259A_irq
47 47
48static int i8259A_auto_eoi;
49
48static void mask_and_ack_8259A(unsigned int); 50static void mask_and_ack_8259A(unsigned int);
49 51
50unsigned int startup_8259A_irq(unsigned int irq) 52unsigned int startup_8259A_irq(unsigned int irq)
@@ -253,7 +255,7 @@ static void save_ELCR(char *trigger)
253 255
254static int i8259A_resume(struct sys_device *dev) 256static int i8259A_resume(struct sys_device *dev)
255{ 257{
256 init_8259A(0); 258 init_8259A(i8259A_auto_eoi);
257 restore_ELCR(irq_trigger); 259 restore_ELCR(irq_trigger);
258 return 0; 260 return 0;
259} 261}
@@ -301,6 +303,8 @@ void init_8259A(int auto_eoi)
301{ 303{
302 unsigned long flags; 304 unsigned long flags;
303 305
306 i8259A_auto_eoi = auto_eoi;
307
304 spin_lock_irqsave(&i8259A_lock, flags); 308 spin_lock_irqsave(&i8259A_lock, flags);
305 309
306 outb(0xff, PIC_MASTER_IMR); /* mask all of 8259A-1 */ 310 outb(0xff, PIC_MASTER_IMR); /* mask all of 8259A-1 */
diff --git a/arch/i386/kernel/io_apic.c b/arch/i386/kernel/io_apic.c
index 4fb32c551fe0..fd0df75cfbda 100644
--- a/arch/i386/kernel/io_apic.c
+++ b/arch/i386/kernel/io_apic.c
@@ -40,6 +40,7 @@
40#include <asm/nmi.h> 40#include <asm/nmi.h>
41 41
42#include <mach_apic.h> 42#include <mach_apic.h>
43#include <mach_apicdef.h>
43 44
44#include "io_ports.h" 45#include "io_ports.h"
45 46
@@ -65,7 +66,7 @@ int sis_apic_bug = -1;
65 */ 66 */
66int nr_ioapic_registers[MAX_IO_APICS]; 67int nr_ioapic_registers[MAX_IO_APICS];
67 68
68int disable_timer_pin_1 __initdata; 69static int disable_timer_pin_1 __initdata;
69 70
70/* 71/*
71 * Rough estimation of how many shared IRQs there are, can 72 * Rough estimation of how many shared IRQs there are, can
@@ -93,6 +94,34 @@ int vector_irq[NR_VECTORS] __read_mostly = { [0 ... NR_VECTORS - 1] = -1};
93#define vector_to_irq(vector) (vector) 94#define vector_to_irq(vector) (vector)
94#endif 95#endif
95 96
97
98union entry_union {
99 struct { u32 w1, w2; };
100 struct IO_APIC_route_entry entry;
101};
102
103static struct IO_APIC_route_entry ioapic_read_entry(int apic, int pin)
104{
105 union entry_union eu;
106 unsigned long flags;
107 spin_lock_irqsave(&ioapic_lock, flags);
108 eu.w1 = io_apic_read(apic, 0x10 + 2 * pin);
109 eu.w2 = io_apic_read(apic, 0x11 + 2 * pin);
110 spin_unlock_irqrestore(&ioapic_lock, flags);
111 return eu.entry;
112}
113
114static void ioapic_write_entry(int apic, int pin, struct IO_APIC_route_entry e)
115{
116 unsigned long flags;
117 union entry_union eu;
118 eu.entry = e;
119 spin_lock_irqsave(&ioapic_lock, flags);
120 io_apic_write(apic, 0x10 + 2*pin, eu.w1);
121 io_apic_write(apic, 0x11 + 2*pin, eu.w2);
122 spin_unlock_irqrestore(&ioapic_lock, flags);
123}
124
96/* 125/*
97 * The common case is 1:1 IRQ<->pin mappings. Sometimes there are 126 * The common case is 1:1 IRQ<->pin mappings. Sometimes there are
98 * shared ISA-space IRQs, so we have to support them. We are super 127 * shared ISA-space IRQs, so we have to support them. We are super
@@ -200,13 +229,9 @@ static void unmask_IO_APIC_irq (unsigned int irq)
200static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin) 229static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin)
201{ 230{
202 struct IO_APIC_route_entry entry; 231 struct IO_APIC_route_entry entry;
203 unsigned long flags;
204 232
205 /* Check delivery_mode to be sure we're not clearing an SMI pin */ 233 /* Check delivery_mode to be sure we're not clearing an SMI pin */
206 spin_lock_irqsave(&ioapic_lock, flags); 234 entry = ioapic_read_entry(apic, pin);
207 *(((int*)&entry) + 0) = io_apic_read(apic, 0x10 + 2 * pin);
208 *(((int*)&entry) + 1) = io_apic_read(apic, 0x11 + 2 * pin);
209 spin_unlock_irqrestore(&ioapic_lock, flags);
210 if (entry.delivery_mode == dest_SMI) 235 if (entry.delivery_mode == dest_SMI)
211 return; 236 return;
212 237
@@ -215,10 +240,7 @@ static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin)
215 */ 240 */
216 memset(&entry, 0, sizeof(entry)); 241 memset(&entry, 0, sizeof(entry));
217 entry.mask = 1; 242 entry.mask = 1;
218 spin_lock_irqsave(&ioapic_lock, flags); 243 ioapic_write_entry(apic, pin, entry);
219 io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry) + 0));
220 io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry) + 1));
221 spin_unlock_irqrestore(&ioapic_lock, flags);
222} 244}
223 245
224static void clear_IO_APIC (void) 246static void clear_IO_APIC (void)
@@ -1283,9 +1305,8 @@ static void __init setup_IO_APIC_irqs(void)
1283 if (!apic && (irq < 16)) 1305 if (!apic && (irq < 16))
1284 disable_8259A_irq(irq); 1306 disable_8259A_irq(irq);
1285 } 1307 }
1308 ioapic_write_entry(apic, pin, entry);
1286 spin_lock_irqsave(&ioapic_lock, flags); 1309 spin_lock_irqsave(&ioapic_lock, flags);
1287 io_apic_write(apic, 0x11+2*pin, *(((int *)&entry)+1));
1288 io_apic_write(apic, 0x10+2*pin, *(((int *)&entry)+0));
1289 set_native_irq_info(irq, TARGET_CPUS); 1310 set_native_irq_info(irq, TARGET_CPUS);
1290 spin_unlock_irqrestore(&ioapic_lock, flags); 1311 spin_unlock_irqrestore(&ioapic_lock, flags);
1291 } 1312 }
@@ -1301,7 +1322,6 @@ static void __init setup_IO_APIC_irqs(void)
1301static void __init setup_ExtINT_IRQ0_pin(unsigned int apic, unsigned int pin, int vector) 1322static void __init setup_ExtINT_IRQ0_pin(unsigned int apic, unsigned int pin, int vector)
1302{ 1323{
1303 struct IO_APIC_route_entry entry; 1324 struct IO_APIC_route_entry entry;
1304 unsigned long flags;
1305 1325
1306 memset(&entry,0,sizeof(entry)); 1326 memset(&entry,0,sizeof(entry));
1307 1327
@@ -1331,10 +1351,7 @@ static void __init setup_ExtINT_IRQ0_pin(unsigned int apic, unsigned int pin, in
1331 /* 1351 /*
1332 * Add it to the IO-APIC irq-routing table: 1352 * Add it to the IO-APIC irq-routing table:
1333 */ 1353 */
1334 spin_lock_irqsave(&ioapic_lock, flags); 1354 ioapic_write_entry(apic, pin, entry);
1335 io_apic_write(apic, 0x11+2*pin, *(((int *)&entry)+1));
1336 io_apic_write(apic, 0x10+2*pin, *(((int *)&entry)+0));
1337 spin_unlock_irqrestore(&ioapic_lock, flags);
1338 1355
1339 enable_8259A_irq(0); 1356 enable_8259A_irq(0);
1340} 1357}
@@ -1444,10 +1461,7 @@ void __init print_IO_APIC(void)
1444 for (i = 0; i <= reg_01.bits.entries; i++) { 1461 for (i = 0; i <= reg_01.bits.entries; i++) {
1445 struct IO_APIC_route_entry entry; 1462 struct IO_APIC_route_entry entry;
1446 1463
1447 spin_lock_irqsave(&ioapic_lock, flags); 1464 entry = ioapic_read_entry(apic, i);
1448 *(((int *)&entry)+0) = io_apic_read(apic, 0x10+i*2);
1449 *(((int *)&entry)+1) = io_apic_read(apic, 0x11+i*2);
1450 spin_unlock_irqrestore(&ioapic_lock, flags);
1451 1465
1452 printk(KERN_DEBUG " %02x %03X %02X ", 1466 printk(KERN_DEBUG " %02x %03X %02X ",
1453 i, 1467 i,
@@ -1666,10 +1680,7 @@ static void __init enable_IO_APIC(void)
1666 /* See if any of the pins is in ExtINT mode */ 1680 /* See if any of the pins is in ExtINT mode */
1667 for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) { 1681 for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) {
1668 struct IO_APIC_route_entry entry; 1682 struct IO_APIC_route_entry entry;
1669 spin_lock_irqsave(&ioapic_lock, flags); 1683 entry = ioapic_read_entry(apic, pin);
1670 *(((int *)&entry) + 0) = io_apic_read(apic, 0x10 + 2 * pin);
1671 *(((int *)&entry) + 1) = io_apic_read(apic, 0x11 + 2 * pin);
1672 spin_unlock_irqrestore(&ioapic_lock, flags);
1673 1684
1674 1685
1675 /* If the interrupt line is enabled and in ExtInt mode 1686 /* If the interrupt line is enabled and in ExtInt mode
@@ -1726,7 +1737,6 @@ void disable_IO_APIC(void)
1726 */ 1737 */
1727 if (ioapic_i8259.pin != -1) { 1738 if (ioapic_i8259.pin != -1) {
1728 struct IO_APIC_route_entry entry; 1739 struct IO_APIC_route_entry entry;
1729 unsigned long flags;
1730 1740
1731 memset(&entry, 0, sizeof(entry)); 1741 memset(&entry, 0, sizeof(entry));
1732 entry.mask = 0; /* Enabled */ 1742 entry.mask = 0; /* Enabled */
@@ -1743,12 +1753,7 @@ void disable_IO_APIC(void)
1743 /* 1753 /*
1744 * Add it to the IO-APIC irq-routing table: 1754 * Add it to the IO-APIC irq-routing table:
1745 */ 1755 */
1746 spin_lock_irqsave(&ioapic_lock, flags); 1756 ioapic_write_entry(ioapic_i8259.apic, ioapic_i8259.pin, entry);
1747 io_apic_write(ioapic_i8259.apic, 0x11+2*ioapic_i8259.pin,
1748 *(((int *)&entry)+1));
1749 io_apic_write(ioapic_i8259.apic, 0x10+2*ioapic_i8259.pin,
1750 *(((int *)&entry)+0));
1751 spin_unlock_irqrestore(&ioapic_lock, flags);
1752 } 1757 }
1753 disconnect_bsp_APIC(ioapic_i8259.pin != -1); 1758 disconnect_bsp_APIC(ioapic_i8259.pin != -1);
1754} 1759}
@@ -2213,17 +2218,13 @@ static inline void unlock_ExtINT_logic(void)
2213 int apic, pin, i; 2218 int apic, pin, i;
2214 struct IO_APIC_route_entry entry0, entry1; 2219 struct IO_APIC_route_entry entry0, entry1;
2215 unsigned char save_control, save_freq_select; 2220 unsigned char save_control, save_freq_select;
2216 unsigned long flags;
2217 2221
2218 pin = find_isa_irq_pin(8, mp_INT); 2222 pin = find_isa_irq_pin(8, mp_INT);
2219 apic = find_isa_irq_apic(8, mp_INT); 2223 apic = find_isa_irq_apic(8, mp_INT);
2220 if (pin == -1) 2224 if (pin == -1)
2221 return; 2225 return;
2222 2226
2223 spin_lock_irqsave(&ioapic_lock, flags); 2227 entry0 = ioapic_read_entry(apic, pin);
2224 *(((int *)&entry0) + 1) = io_apic_read(apic, 0x11 + 2 * pin);
2225 *(((int *)&entry0) + 0) = io_apic_read(apic, 0x10 + 2 * pin);
2226 spin_unlock_irqrestore(&ioapic_lock, flags);
2227 clear_IO_APIC_pin(apic, pin); 2228 clear_IO_APIC_pin(apic, pin);
2228 2229
2229 memset(&entry1, 0, sizeof(entry1)); 2230 memset(&entry1, 0, sizeof(entry1));
@@ -2236,10 +2237,7 @@ static inline void unlock_ExtINT_logic(void)
2236 entry1.trigger = 0; 2237 entry1.trigger = 0;
2237 entry1.vector = 0; 2238 entry1.vector = 0;
2238 2239
2239 spin_lock_irqsave(&ioapic_lock, flags); 2240 ioapic_write_entry(apic, pin, entry1);
2240 io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry1) + 1));
2241 io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry1) + 0));
2242 spin_unlock_irqrestore(&ioapic_lock, flags);
2243 2241
2244 save_control = CMOS_READ(RTC_CONTROL); 2242 save_control = CMOS_READ(RTC_CONTROL);
2245 save_freq_select = CMOS_READ(RTC_FREQ_SELECT); 2243 save_freq_select = CMOS_READ(RTC_FREQ_SELECT);
@@ -2258,10 +2256,7 @@ static inline void unlock_ExtINT_logic(void)
2258 CMOS_WRITE(save_freq_select, RTC_FREQ_SELECT); 2256 CMOS_WRITE(save_freq_select, RTC_FREQ_SELECT);
2259 clear_IO_APIC_pin(apic, pin); 2257 clear_IO_APIC_pin(apic, pin);
2260 2258
2261 spin_lock_irqsave(&ioapic_lock, flags); 2259 ioapic_write_entry(apic, pin, entry0);
2262 io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry0) + 1));
2263 io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry0) + 0));
2264 spin_unlock_irqrestore(&ioapic_lock, flags);
2265} 2260}
2266 2261
2267int timer_uses_ioapic_pin_0; 2262int timer_uses_ioapic_pin_0;
@@ -2461,17 +2456,12 @@ static int ioapic_suspend(struct sys_device *dev, pm_message_t state)
2461{ 2456{
2462 struct IO_APIC_route_entry *entry; 2457 struct IO_APIC_route_entry *entry;
2463 struct sysfs_ioapic_data *data; 2458 struct sysfs_ioapic_data *data;
2464 unsigned long flags;
2465 int i; 2459 int i;
2466 2460
2467 data = container_of(dev, struct sysfs_ioapic_data, dev); 2461 data = container_of(dev, struct sysfs_ioapic_data, dev);
2468 entry = data->entry; 2462 entry = data->entry;
2469 spin_lock_irqsave(&ioapic_lock, flags); 2463 for (i = 0; i < nr_ioapic_registers[dev->id]; i ++)
2470 for (i = 0; i < nr_ioapic_registers[dev->id]; i ++, entry ++ ) { 2464 entry[i] = ioapic_read_entry(dev->id, i);
2471 *(((int *)entry) + 1) = io_apic_read(dev->id, 0x11 + 2 * i);
2472 *(((int *)entry) + 0) = io_apic_read(dev->id, 0x10 + 2 * i);
2473 }
2474 spin_unlock_irqrestore(&ioapic_lock, flags);
2475 2465
2476 return 0; 2466 return 0;
2477} 2467}
@@ -2493,11 +2483,9 @@ static int ioapic_resume(struct sys_device *dev)
2493 reg_00.bits.ID = mp_ioapics[dev->id].mpc_apicid; 2483 reg_00.bits.ID = mp_ioapics[dev->id].mpc_apicid;
2494 io_apic_write(dev->id, 0, reg_00.raw); 2484 io_apic_write(dev->id, 0, reg_00.raw);
2495 } 2485 }
2496 for (i = 0; i < nr_ioapic_registers[dev->id]; i ++, entry ++ ) {
2497 io_apic_write(dev->id, 0x11+2*i, *(((int *)entry)+1));
2498 io_apic_write(dev->id, 0x10+2*i, *(((int *)entry)+0));
2499 }
2500 spin_unlock_irqrestore(&ioapic_lock, flags); 2486 spin_unlock_irqrestore(&ioapic_lock, flags);
2487 for (i = 0; i < nr_ioapic_registers[dev->id]; i ++)
2488 ioapic_write_entry(dev->id, i, entry[i]);
2501 2489
2502 return 0; 2490 return 0;
2503} 2491}
@@ -2694,9 +2682,8 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int edge_level, int a
2694 if (!ioapic && (irq < 16)) 2682 if (!ioapic && (irq < 16))
2695 disable_8259A_irq(irq); 2683 disable_8259A_irq(irq);
2696 2684
2685 ioapic_write_entry(ioapic, pin, entry);
2697 spin_lock_irqsave(&ioapic_lock, flags); 2686 spin_lock_irqsave(&ioapic_lock, flags);
2698 io_apic_write(ioapic, 0x11+2*pin, *(((int *)&entry)+1));
2699 io_apic_write(ioapic, 0x10+2*pin, *(((int *)&entry)+0));
2700 set_native_irq_info(use_pci_vector() ? entry.vector : irq, TARGET_CPUS); 2687 set_native_irq_info(use_pci_vector() ? entry.vector : irq, TARGET_CPUS);
2701 spin_unlock_irqrestore(&ioapic_lock, flags); 2688 spin_unlock_irqrestore(&ioapic_lock, flags);
2702 2689
@@ -2704,3 +2691,25 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int edge_level, int a
2704} 2691}
2705 2692
2706#endif /* CONFIG_ACPI */ 2693#endif /* CONFIG_ACPI */
2694
2695static int __init parse_disable_timer_pin_1(char *arg)
2696{
2697 disable_timer_pin_1 = 1;
2698 return 0;
2699}
2700early_param("disable_timer_pin_1", parse_disable_timer_pin_1);
2701
2702static int __init parse_enable_timer_pin_1(char *arg)
2703{
2704 disable_timer_pin_1 = -1;
2705 return 0;
2706}
2707early_param("enable_timer_pin_1", parse_enable_timer_pin_1);
2708
2709static int __init parse_noapic(char *arg)
2710{
2711 /* disable IO-APIC */
2712 disable_ioapic_setup();
2713 return 0;
2714}
2715early_param("noapic", parse_noapic);
diff --git a/arch/i386/kernel/machine_kexec.c b/arch/i386/kernel/machine_kexec.c
index 6b1ae6ba76f0..91966bafb3dc 100644
--- a/arch/i386/kernel/machine_kexec.c
+++ b/arch/i386/kernel/machine_kexec.c
@@ -9,6 +9,7 @@
9#include <linux/mm.h> 9#include <linux/mm.h>
10#include <linux/kexec.h> 10#include <linux/kexec.h>
11#include <linux/delay.h> 11#include <linux/delay.h>
12#include <linux/init.h>
12#include <asm/pgtable.h> 13#include <asm/pgtable.h>
13#include <asm/pgalloc.h> 14#include <asm/pgalloc.h>
14#include <asm/tlbflush.h> 15#include <asm/tlbflush.h>
@@ -20,70 +21,13 @@
20#include <asm/system.h> 21#include <asm/system.h>
21 22
22#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE))) 23#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE)))
23 24static u32 kexec_pgd[1024] PAGE_ALIGNED;
24#define L0_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY) 25#ifdef CONFIG_X86_PAE
25#define L1_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY) 26static u32 kexec_pmd0[1024] PAGE_ALIGNED;
26#define L2_ATTR (_PAGE_PRESENT) 27static u32 kexec_pmd1[1024] PAGE_ALIGNED;
27
28#define LEVEL0_SIZE (1UL << 12UL)
29
30#ifndef CONFIG_X86_PAE
31#define LEVEL1_SIZE (1UL << 22UL)
32static u32 pgtable_level1[1024] PAGE_ALIGNED;
33
34static void identity_map_page(unsigned long address)
35{
36 unsigned long level1_index, level2_index;
37 u32 *pgtable_level2;
38
39 /* Find the current page table */
40 pgtable_level2 = __va(read_cr3());
41
42 /* Find the indexes of the physical address to identity map */
43 level1_index = (address % LEVEL1_SIZE)/LEVEL0_SIZE;
44 level2_index = address / LEVEL1_SIZE;
45
46 /* Identity map the page table entry */
47 pgtable_level1[level1_index] = address | L0_ATTR;
48 pgtable_level2[level2_index] = __pa(pgtable_level1) | L1_ATTR;
49
50 /* Flush the tlb so the new mapping takes effect.
51 * Global tlb entries are not flushed but that is not an issue.
52 */
53 load_cr3(pgtable_level2);
54}
55
56#else
57#define LEVEL1_SIZE (1UL << 21UL)
58#define LEVEL2_SIZE (1UL << 30UL)
59static u64 pgtable_level1[512] PAGE_ALIGNED;
60static u64 pgtable_level2[512] PAGE_ALIGNED;
61
62static void identity_map_page(unsigned long address)
63{
64 unsigned long level1_index, level2_index, level3_index;
65 u64 *pgtable_level3;
66
67 /* Find the current page table */
68 pgtable_level3 = __va(read_cr3());
69
70 /* Find the indexes of the physical address to identity map */
71 level1_index = (address % LEVEL1_SIZE)/LEVEL0_SIZE;
72 level2_index = (address % LEVEL2_SIZE)/LEVEL1_SIZE;
73 level3_index = address / LEVEL2_SIZE;
74
75 /* Identity map the page table entry */
76 pgtable_level1[level1_index] = address | L0_ATTR;
77 pgtable_level2[level2_index] = __pa(pgtable_level1) | L1_ATTR;
78 set_64bit(&pgtable_level3[level3_index],
79 __pa(pgtable_level2) | L2_ATTR);
80
81 /* Flush the tlb so the new mapping takes effect.
82 * Global tlb entries are not flushed but that is not an issue.
83 */
84 load_cr3(pgtable_level3);
85}
86#endif 28#endif
29static u32 kexec_pte0[1024] PAGE_ALIGNED;
30static u32 kexec_pte1[1024] PAGE_ALIGNED;
87 31
88static void set_idt(void *newidt, __u16 limit) 32static void set_idt(void *newidt, __u16 limit)
89{ 33{
@@ -127,16 +71,6 @@ static void load_segments(void)
127#undef __STR 71#undef __STR
128} 72}
129 73
130typedef asmlinkage NORET_TYPE void (*relocate_new_kernel_t)(
131 unsigned long indirection_page,
132 unsigned long reboot_code_buffer,
133 unsigned long start_address,
134 unsigned int has_pae) ATTRIB_NORET;
135
136extern const unsigned char relocate_new_kernel[];
137extern void relocate_new_kernel_end(void);
138extern const unsigned int relocate_new_kernel_size;
139
140/* 74/*
141 * A architecture hook called to validate the 75 * A architecture hook called to validate the
142 * proposed image and prepare the control pages 76 * proposed image and prepare the control pages
@@ -169,25 +103,29 @@ void machine_kexec_cleanup(struct kimage *image)
169 */ 103 */
170NORET_TYPE void machine_kexec(struct kimage *image) 104NORET_TYPE void machine_kexec(struct kimage *image)
171{ 105{
172 unsigned long page_list; 106 unsigned long page_list[PAGES_NR];
173 unsigned long reboot_code_buffer; 107 void *control_page;
174
175 relocate_new_kernel_t rnk;
176 108
177 /* Interrupts aren't acceptable while we reboot */ 109 /* Interrupts aren't acceptable while we reboot */
178 local_irq_disable(); 110 local_irq_disable();
179 111
180 /* Compute some offsets */ 112 control_page = page_address(image->control_code_page);
181 reboot_code_buffer = page_to_pfn(image->control_code_page) 113 memcpy(control_page, relocate_kernel, PAGE_SIZE);
182 << PAGE_SHIFT; 114
183 page_list = image->head; 115 page_list[PA_CONTROL_PAGE] = __pa(control_page);
184 116 page_list[VA_CONTROL_PAGE] = (unsigned long)relocate_kernel;
185 /* Set up an identity mapping for the reboot_code_buffer */ 117 page_list[PA_PGD] = __pa(kexec_pgd);
186 identity_map_page(reboot_code_buffer); 118 page_list[VA_PGD] = (unsigned long)kexec_pgd;
187 119#ifdef CONFIG_X86_PAE
188 /* copy it out */ 120 page_list[PA_PMD_0] = __pa(kexec_pmd0);
189 memcpy((void *)reboot_code_buffer, relocate_new_kernel, 121 page_list[VA_PMD_0] = (unsigned long)kexec_pmd0;
190 relocate_new_kernel_size); 122 page_list[PA_PMD_1] = __pa(kexec_pmd1);
123 page_list[VA_PMD_1] = (unsigned long)kexec_pmd1;
124#endif
125 page_list[PA_PTE_0] = __pa(kexec_pte0);
126 page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
127 page_list[PA_PTE_1] = __pa(kexec_pte1);
128 page_list[VA_PTE_1] = (unsigned long)kexec_pte1;
191 129
192 /* The segment registers are funny things, they have both a 130 /* The segment registers are funny things, they have both a
193 * visible and an invisible part. Whenever the visible part is 131 * visible and an invisible part. Whenever the visible part is
@@ -206,6 +144,28 @@ NORET_TYPE void machine_kexec(struct kimage *image)
206 set_idt(phys_to_virt(0),0); 144 set_idt(phys_to_virt(0),0);
207 145
208 /* now call it */ 146 /* now call it */
209 rnk = (relocate_new_kernel_t) reboot_code_buffer; 147 relocate_kernel((unsigned long)image->head, (unsigned long)page_list,
210 (*rnk)(page_list, reboot_code_buffer, image->start, cpu_has_pae); 148 image->start, cpu_has_pae);
149}
150
151/* crashkernel=size@addr specifies the location to reserve for
152 * a crash kernel. By reserving this memory we guarantee
153 * that linux never sets it up as a DMA target.
154 * Useful for holding code to do something appropriate
155 * after a kernel panic.
156 */
157static int __init parse_crashkernel(char *arg)
158{
159 unsigned long size, base;
160 size = memparse(arg, &arg);
161 if (*arg == '@') {
162 base = memparse(arg+1, &arg);
163 /* FIXME: Do I want a sanity check
164 * to validate the memory range?
165 */
166 crashk_res.start = base;
167 crashk_res.end = base + size - 1;
168 }
169 return 0;
211} 170}
171early_param("crashkernel", parse_crashkernel);
diff --git a/arch/i386/kernel/mca.c b/arch/i386/kernel/mca.c
index cd5456f14af4..eb57a851789d 100644
--- a/arch/i386/kernel/mca.c
+++ b/arch/i386/kernel/mca.c
@@ -42,6 +42,7 @@
42#include <linux/errno.h> 42#include <linux/errno.h>
43#include <linux/kernel.h> 43#include <linux/kernel.h>
44#include <linux/mca.h> 44#include <linux/mca.h>
45#include <linux/kprobes.h>
45#include <asm/system.h> 46#include <asm/system.h>
46#include <asm/io.h> 47#include <asm/io.h>
47#include <linux/proc_fs.h> 48#include <linux/proc_fs.h>
@@ -414,7 +415,8 @@ subsys_initcall(mca_init);
414 415
415/*--------------------------------------------------------------------*/ 416/*--------------------------------------------------------------------*/
416 417
417static void mca_handle_nmi_device(struct mca_device *mca_dev, int check_flag) 418static __kprobes void
419mca_handle_nmi_device(struct mca_device *mca_dev, int check_flag)
418{ 420{
419 int slot = mca_dev->slot; 421 int slot = mca_dev->slot;
420 422
@@ -444,7 +446,7 @@ static void mca_handle_nmi_device(struct mca_device *mca_dev, int check_flag)
444 446
445/*--------------------------------------------------------------------*/ 447/*--------------------------------------------------------------------*/
446 448
447static int mca_handle_nmi_callback(struct device *dev, void *data) 449static int __kprobes mca_handle_nmi_callback(struct device *dev, void *data)
448{ 450{
449 struct mca_device *mca_dev = to_mca_device(dev); 451 struct mca_device *mca_dev = to_mca_device(dev);
450 unsigned char pos5; 452 unsigned char pos5;
@@ -462,7 +464,7 @@ static int mca_handle_nmi_callback(struct device *dev, void *data)
462 return 0; 464 return 0;
463} 465}
464 466
465void mca_handle_nmi(void) 467void __kprobes mca_handle_nmi(void)
466{ 468{
467 /* First try - scan the various adapters and see if a specific 469 /* First try - scan the various adapters and see if a specific
468 * adapter was responsible for the error. 470 * adapter was responsible for the error.
diff --git a/arch/i386/kernel/microcode.c b/arch/i386/kernel/microcode.c
index 40b44cc0d14b..9b9479768d5e 100644
--- a/arch/i386/kernel/microcode.c
+++ b/arch/i386/kernel/microcode.c
@@ -2,6 +2,7 @@
2 * Intel CPU Microcode Update Driver for Linux 2 * Intel CPU Microcode Update Driver for Linux
3 * 3 *
4 * Copyright (C) 2000-2004 Tigran Aivazian 4 * Copyright (C) 2000-2004 Tigran Aivazian
5 * 2006 Shaohua Li <shaohua.li@intel.com>
5 * 6 *
6 * This driver allows to upgrade microcode on Intel processors 7 * This driver allows to upgrade microcode on Intel processors
7 * belonging to IA-32 family - PentiumPro, Pentium II, 8 * belonging to IA-32 family - PentiumPro, Pentium II,
@@ -82,6 +83,9 @@
82#include <linux/spinlock.h> 83#include <linux/spinlock.h>
83#include <linux/mm.h> 84#include <linux/mm.h>
84#include <linux/mutex.h> 85#include <linux/mutex.h>
86#include <linux/cpu.h>
87#include <linux/firmware.h>
88#include <linux/platform_device.h>
85 89
86#include <asm/msr.h> 90#include <asm/msr.h>
87#include <asm/uaccess.h> 91#include <asm/uaccess.h>
@@ -91,9 +95,6 @@ MODULE_DESCRIPTION("Intel CPU (IA-32) Microcode Update Driver");
91MODULE_AUTHOR("Tigran Aivazian <tigran@veritas.com>"); 95MODULE_AUTHOR("Tigran Aivazian <tigran@veritas.com>");
92MODULE_LICENSE("GPL"); 96MODULE_LICENSE("GPL");
93 97
94static int verbose;
95module_param(verbose, int, 0644);
96
97#define MICROCODE_VERSION "1.14a" 98#define MICROCODE_VERSION "1.14a"
98 99
99#define DEFAULT_UCODE_DATASIZE (2000) /* 2000 bytes */ 100#define DEFAULT_UCODE_DATASIZE (2000) /* 2000 bytes */
@@ -120,55 +121,40 @@ static DEFINE_SPINLOCK(microcode_update_lock);
120/* no concurrent ->write()s are allowed on /dev/cpu/microcode */ 121/* no concurrent ->write()s are allowed on /dev/cpu/microcode */
121static DEFINE_MUTEX(microcode_mutex); 122static DEFINE_MUTEX(microcode_mutex);
122 123
123static void __user *user_buffer; /* user area microcode data buffer */
124static unsigned int user_buffer_size; /* it's size */
125
126typedef enum mc_error_code {
127 MC_SUCCESS = 0,
128 MC_IGNORED = 1,
129 MC_NOTFOUND = 2,
130 MC_MARKED = 3,
131 MC_ALLOCATED = 4,
132} mc_error_code_t;
133
134static struct ucode_cpu_info { 124static struct ucode_cpu_info {
125 int valid;
135 unsigned int sig; 126 unsigned int sig;
136 unsigned int pf, orig_pf; 127 unsigned int pf;
137 unsigned int rev; 128 unsigned int rev;
138 unsigned int cksum;
139 mc_error_code_t err;
140 microcode_t *mc; 129 microcode_t *mc;
141} ucode_cpu_info[NR_CPUS]; 130} ucode_cpu_info[NR_CPUS];
142
143static int microcode_open (struct inode *unused1, struct file *unused2)
144{
145 return capable(CAP_SYS_RAWIO) ? 0 : -EPERM;
146}
147 131
148static void collect_cpu_info (void *unused) 132static void collect_cpu_info(int cpu_num)
149{ 133{
150 int cpu_num = smp_processor_id();
151 struct cpuinfo_x86 *c = cpu_data + cpu_num; 134 struct cpuinfo_x86 *c = cpu_data + cpu_num;
152 struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; 135 struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num;
153 unsigned int val[2]; 136 unsigned int val[2];
154 137
155 uci->sig = uci->pf = uci->rev = uci->cksum = 0; 138 /* We should bind the task to the CPU */
156 uci->err = MC_NOTFOUND; 139 BUG_ON(raw_smp_processor_id() != cpu_num);
140 uci->pf = uci->rev = 0;
157 uci->mc = NULL; 141 uci->mc = NULL;
142 uci->valid = 1;
158 143
159 if (c->x86_vendor != X86_VENDOR_INTEL || c->x86 < 6 || 144 if (c->x86_vendor != X86_VENDOR_INTEL || c->x86 < 6 ||
160 cpu_has(c, X86_FEATURE_IA64)) { 145 cpu_has(c, X86_FEATURE_IA64)) {
161 printk(KERN_ERR "microcode: CPU%d not a capable Intel processor\n", cpu_num); 146 printk(KERN_ERR "microcode: CPU%d not a capable Intel "
147 "processor\n", cpu_num);
148 uci->valid = 0;
162 return; 149 return;
163 } else { 150 }
164 uci->sig = cpuid_eax(0x00000001);
165 151
166 if ((c->x86_model >= 5) || (c->x86 > 6)) { 152 uci->sig = cpuid_eax(0x00000001);
167 /* get processor flags from MSR 0x17 */ 153
168 rdmsr(MSR_IA32_PLATFORM_ID, val[0], val[1]); 154 if ((c->x86_model >= 5) || (c->x86 > 6)) {
169 uci->pf = 1 << ((val[1] >> 18) & 7); 155 /* get processor flags from MSR 0x17 */
170 } 156 rdmsr(MSR_IA32_PLATFORM_ID, val[0], val[1]);
171 uci->orig_pf = uci->pf; 157 uci->pf = 1 << ((val[1] >> 18) & 7);
172 } 158 }
173 159
174 wrmsr(MSR_IA32_UCODE_REV, 0, 0); 160 wrmsr(MSR_IA32_UCODE_REV, 0, 0);
@@ -180,218 +166,159 @@ static void collect_cpu_info (void *unused)
180 uci->sig, uci->pf, uci->rev); 166 uci->sig, uci->pf, uci->rev);
181} 167}
182 168
183static inline void mark_microcode_update (int cpu_num, microcode_header_t *mc_header, int sig, int pf, int cksum) 169static inline int microcode_update_match(int cpu_num,
170 microcode_header_t *mc_header, int sig, int pf)
184{ 171{
185 struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; 172 struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num;
186 173
187 pr_debug("Microcode Found.\n"); 174 if (!sigmatch(sig, uci->sig, pf, uci->pf)
188 pr_debug(" Header Revision 0x%x\n", mc_header->hdrver); 175 || mc_header->rev <= uci->rev)
189 pr_debug(" Loader Revision 0x%x\n", mc_header->ldrver); 176 return 0;
190 pr_debug(" Revision 0x%x \n", mc_header->rev); 177 return 1;
191 pr_debug(" Date %x/%x/%x\n",
192 ((mc_header->date >> 24 ) & 0xff),
193 ((mc_header->date >> 16 ) & 0xff),
194 (mc_header->date & 0xFFFF));
195 pr_debug(" Signature 0x%x\n", sig);
196 pr_debug(" Type 0x%x Family 0x%x Model 0x%x Stepping 0x%x\n",
197 ((sig >> 12) & 0x3),
198 ((sig >> 8) & 0xf),
199 ((sig >> 4) & 0xf),
200 ((sig & 0xf)));
201 pr_debug(" Processor Flags 0x%x\n", pf);
202 pr_debug(" Checksum 0x%x\n", cksum);
203
204 if (mc_header->rev < uci->rev) {
205 if (uci->err == MC_NOTFOUND) {
206 uci->err = MC_IGNORED;
207 uci->cksum = mc_header->rev;
208 } else if (uci->err == MC_IGNORED && uci->cksum < mc_header->rev)
209 uci->cksum = mc_header->rev;
210 } else if (mc_header->rev == uci->rev) {
211 if (uci->err < MC_MARKED) {
212 /* notify the caller of success on this cpu */
213 uci->err = MC_SUCCESS;
214 }
215 } else if (uci->err != MC_ALLOCATED || mc_header->rev > uci->mc->hdr.rev) {
216 pr_debug("microcode: CPU%d found a matching microcode update with "
217 " revision 0x%x (current=0x%x)\n", cpu_num, mc_header->rev, uci->rev);
218 uci->cksum = cksum;
219 uci->pf = pf; /* keep the original mc pf for cksum calculation */
220 uci->err = MC_MARKED; /* found the match */
221 for_each_online_cpu(cpu_num) {
222 if (ucode_cpu_info + cpu_num != uci
223 && ucode_cpu_info[cpu_num].mc == uci->mc) {
224 uci->mc = NULL;
225 break;
226 }
227 }
228 if (uci->mc != NULL) {
229 vfree(uci->mc);
230 uci->mc = NULL;
231 }
232 }
233 return;
234} 178}
235 179
236static int find_matching_ucodes (void) 180static int microcode_sanity_check(void *mc)
237{ 181{
238 int cursor = 0; 182 microcode_header_t *mc_header = mc;
239 int error = 0; 183 struct extended_sigtable *ext_header = NULL;
240 184 struct extended_signature *ext_sig;
241 while (cursor + MC_HEADER_SIZE < user_buffer_size) { 185 unsigned long total_size, data_size, ext_table_size;
242 microcode_header_t mc_header; 186 int sum, orig_sum, ext_sigcount = 0, i;
243 void *newmc = NULL; 187
244 int i, sum, cpu_num, allocated_flag, total_size, data_size, ext_table_size; 188 total_size = get_totalsize(mc_header);
189 data_size = get_datasize(mc_header);
190 if (data_size + MC_HEADER_SIZE > total_size) {
191 printk(KERN_ERR "microcode: error! "
192 "Bad data size in microcode data file\n");
193 return -EINVAL;
194 }
245 195
246 if (copy_from_user(&mc_header, user_buffer + cursor, MC_HEADER_SIZE)) { 196 if (mc_header->ldrver != 1 || mc_header->hdrver != 1) {
247 printk(KERN_ERR "microcode: error! Can not read user data\n"); 197 printk(KERN_ERR "microcode: error! "
248 error = -EFAULT; 198 "Unknown microcode update format\n");
249 goto out; 199 return -EINVAL;
200 }
201 ext_table_size = total_size - (MC_HEADER_SIZE + data_size);
202 if (ext_table_size) {
203 if ((ext_table_size < EXT_HEADER_SIZE)
204 || ((ext_table_size - EXT_HEADER_SIZE) % EXT_SIGNATURE_SIZE)) {
205 printk(KERN_ERR "microcode: error! "
206 "Small exttable size in microcode data file\n");
207 return -EINVAL;
250 } 208 }
251 209 ext_header = mc + MC_HEADER_SIZE + data_size;
252 total_size = get_totalsize(&mc_header); 210 if (ext_table_size != exttable_size(ext_header)) {
253 if ((cursor + total_size > user_buffer_size) || (total_size < DEFAULT_UCODE_TOTALSIZE)) { 211 printk(KERN_ERR "microcode: error! "
254 printk(KERN_ERR "microcode: error! Bad data in microcode data file\n"); 212 "Bad exttable size in microcode data file\n");
255 error = -EINVAL; 213 return -EFAULT;
256 goto out;
257 } 214 }
215 ext_sigcount = ext_header->count;
216 }
258 217
259 data_size = get_datasize(&mc_header); 218 /* check extended table checksum */
260 if ((data_size + MC_HEADER_SIZE > total_size) || (data_size < DEFAULT_UCODE_DATASIZE)) { 219 if (ext_table_size) {
261 printk(KERN_ERR "microcode: error! Bad data in microcode data file\n"); 220 int ext_table_sum = 0;
262 error = -EINVAL; 221 int *ext_tablep = (int *)ext_header;
263 goto out; 222
223 i = ext_table_size / DWSIZE;
224 while (i--)
225 ext_table_sum += ext_tablep[i];
226 if (ext_table_sum) {
227 printk(KERN_WARNING "microcode: aborting, "
228 "bad extended signature table checksum\n");
229 return -EINVAL;
264 } 230 }
231 }
265 232
266 if (mc_header.ldrver != 1 || mc_header.hdrver != 1) { 233 /* calculate the checksum */
267 printk(KERN_ERR "microcode: error! Unknown microcode update format\n"); 234 orig_sum = 0;
268 error = -EINVAL; 235 i = (MC_HEADER_SIZE + data_size) / DWSIZE;
269 goto out; 236 while (i--)
237 orig_sum += ((int *)mc)[i];
238 if (orig_sum) {
239 printk(KERN_ERR "microcode: aborting, bad checksum\n");
240 return -EINVAL;
241 }
242 if (!ext_table_size)
243 return 0;
244 /* check extended signature checksum */
245 for (i = 0; i < ext_sigcount; i++) {
246 ext_sig = (struct extended_signature *)((void *)ext_header
247 + EXT_HEADER_SIZE + EXT_SIGNATURE_SIZE * i);
248 sum = orig_sum
249 - (mc_header->sig + mc_header->pf + mc_header->cksum)
250 + (ext_sig->sig + ext_sig->pf + ext_sig->cksum);
251 if (sum) {
252 printk(KERN_ERR "microcode: aborting, bad checksum\n");
253 return -EINVAL;
270 } 254 }
255 }
256 return 0;
257}
271 258
272 for_each_online_cpu(cpu_num) { 259/*
273 struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; 260 * return 0 - no update found
274 261 * return 1 - found update
275 if (sigmatch(mc_header.sig, uci->sig, mc_header.pf, uci->orig_pf)) 262 * return < 0 - error
276 mark_microcode_update(cpu_num, &mc_header, mc_header.sig, mc_header.pf, mc_header.cksum); 263 */
277 } 264static int get_maching_microcode(void *mc, int cpu)
265{
266 struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
267 microcode_header_t *mc_header = mc;
268 struct extended_sigtable *ext_header;
269 unsigned long total_size = get_totalsize(mc_header);
270 int ext_sigcount, i;
271 struct extended_signature *ext_sig;
272 void *new_mc;
273
274 if (microcode_update_match(cpu, mc_header,
275 mc_header->sig, mc_header->pf))
276 goto find;
277
278 if (total_size <= get_datasize(mc_header) + MC_HEADER_SIZE)
279 return 0;
280
281 ext_header = (struct extended_sigtable *)(mc +
282 get_datasize(mc_header) + MC_HEADER_SIZE);
283 ext_sigcount = ext_header->count;
284 ext_sig = (struct extended_signature *)((void *)ext_header
285 + EXT_HEADER_SIZE);
286 for (i = 0; i < ext_sigcount; i++) {
287 if (microcode_update_match(cpu, mc_header,
288 ext_sig->sig, ext_sig->pf))
289 goto find;
290 ext_sig++;
291 }
292 return 0;
293find:
294 pr_debug("microcode: CPU %d found a matching microcode update with"
295 " version 0x%x (current=0x%x)\n", cpu, mc_header->rev,uci->rev);
296 new_mc = vmalloc(total_size);
297 if (!new_mc) {
298 printk(KERN_ERR "microcode: error! Can not allocate memory\n");
299 return -ENOMEM;
300 }
278 301
279 ext_table_size = total_size - (MC_HEADER_SIZE + data_size); 302 /* free previous update file */
280 if (ext_table_size) { 303 vfree(uci->mc);
281 struct extended_sigtable ext_header;
282 struct extended_signature ext_sig;
283 int ext_sigcount;
284 304
285 if ((ext_table_size < EXT_HEADER_SIZE) 305 memcpy(new_mc, mc, total_size);
286 || ((ext_table_size - EXT_HEADER_SIZE) % EXT_SIGNATURE_SIZE)) { 306 uci->mc = new_mc;
287 printk(KERN_ERR "microcode: error! Bad data in microcode data file\n"); 307 return 1;
288 error = -EINVAL;
289 goto out;
290 }
291 if (copy_from_user(&ext_header, user_buffer + cursor
292 + MC_HEADER_SIZE + data_size, EXT_HEADER_SIZE)) {
293 printk(KERN_ERR "microcode: error! Can not read user data\n");
294 error = -EFAULT;
295 goto out;
296 }
297 if (ext_table_size != exttable_size(&ext_header)) {
298 printk(KERN_ERR "microcode: error! Bad data in microcode data file\n");
299 error = -EFAULT;
300 goto out;
301 }
302
303 ext_sigcount = ext_header.count;
304
305 for (i = 0; i < ext_sigcount; i++) {
306 if (copy_from_user(&ext_sig, user_buffer + cursor + MC_HEADER_SIZE + data_size + EXT_HEADER_SIZE
307 + EXT_SIGNATURE_SIZE * i, EXT_SIGNATURE_SIZE)) {
308 printk(KERN_ERR "microcode: error! Can not read user data\n");
309 error = -EFAULT;
310 goto out;
311 }
312 for_each_online_cpu(cpu_num) {
313 struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num;
314
315 if (sigmatch(ext_sig.sig, uci->sig, ext_sig.pf, uci->orig_pf)) {
316 mark_microcode_update(cpu_num, &mc_header, ext_sig.sig, ext_sig.pf, ext_sig.cksum);
317 }
318 }
319 }
320 }
321 /* now check if any cpu has matched */
322 allocated_flag = 0;
323 sum = 0;
324 for_each_online_cpu(cpu_num) {
325 if (ucode_cpu_info[cpu_num].err == MC_MARKED) {
326 struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num;
327 if (!allocated_flag) {
328 allocated_flag = 1;
329 newmc = vmalloc(total_size);
330 if (!newmc) {
331 printk(KERN_ERR "microcode: error! Can not allocate memory\n");
332 error = -ENOMEM;
333 goto out;
334 }
335 if (copy_from_user(newmc + MC_HEADER_SIZE,
336 user_buffer + cursor + MC_HEADER_SIZE,
337 total_size - MC_HEADER_SIZE)) {
338 printk(KERN_ERR "microcode: error! Can not read user data\n");
339 vfree(newmc);
340 error = -EFAULT;
341 goto out;
342 }
343 memcpy(newmc, &mc_header, MC_HEADER_SIZE);
344 /* check extended table checksum */
345 if (ext_table_size) {
346 int ext_table_sum = 0;
347 int * ext_tablep = (((void *) newmc) + MC_HEADER_SIZE + data_size);
348 i = ext_table_size / DWSIZE;
349 while (i--) ext_table_sum += ext_tablep[i];
350 if (ext_table_sum) {
351 printk(KERN_WARNING "microcode: aborting, bad extended signature table checksum\n");
352 vfree(newmc);
353 error = -EINVAL;
354 goto out;
355 }
356 }
357
358 /* calculate the checksum */
359 i = (MC_HEADER_SIZE + data_size) / DWSIZE;
360 while (i--) sum += ((int *)newmc)[i];
361 sum -= (mc_header.sig + mc_header.pf + mc_header.cksum);
362 }
363 ucode_cpu_info[cpu_num].mc = newmc;
364 ucode_cpu_info[cpu_num].err = MC_ALLOCATED; /* mc updated */
365 if (sum + uci->sig + uci->pf + uci->cksum != 0) {
366 printk(KERN_ERR "microcode: CPU%d aborting, bad checksum\n", cpu_num);
367 error = -EINVAL;
368 goto out;
369 }
370 }
371 }
372 cursor += total_size; /* goto the next update patch */
373 } /* end of while */
374out:
375 return error;
376} 308}
377 309
378static void do_update_one (void * unused) 310static void apply_microcode(int cpu)
379{ 311{
380 unsigned long flags; 312 unsigned long flags;
381 unsigned int val[2]; 313 unsigned int val[2];
382 int cpu_num = smp_processor_id(); 314 int cpu_num = raw_smp_processor_id();
383 struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; 315 struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num;
384 316
385 if (uci->mc == NULL) { 317 /* We should bind the task to the CPU */
386 if (verbose) { 318 BUG_ON(cpu_num != cpu);
387 if (uci->err == MC_SUCCESS) 319
388 printk(KERN_INFO "microcode: CPU%d already at revision 0x%x\n", 320 if (uci->mc == NULL)
389 cpu_num, uci->rev);
390 else
391 printk(KERN_INFO "microcode: No new microcode data for CPU%d\n", cpu_num);
392 }
393 return; 321 return;
394 }
395 322
396 /* serialize access to the physical write to MSR 0x79 */ 323 /* serialize access to the physical write to MSR 0x79 */
397 spin_lock_irqsave(&microcode_update_lock, flags); 324 spin_lock_irqsave(&microcode_update_lock, flags);
@@ -408,68 +335,107 @@ static void do_update_one (void * unused)
408 /* get the current revision from MSR 0x8B */ 335 /* get the current revision from MSR 0x8B */
409 rdmsr(MSR_IA32_UCODE_REV, val[0], val[1]); 336 rdmsr(MSR_IA32_UCODE_REV, val[0], val[1]);
410 337
411 /* notify the caller of success on this cpu */
412 uci->err = MC_SUCCESS;
413 spin_unlock_irqrestore(&microcode_update_lock, flags); 338 spin_unlock_irqrestore(&microcode_update_lock, flags);
414 printk(KERN_INFO "microcode: CPU%d updated from revision " 339 if (val[1] != uci->mc->hdr.rev) {
340 printk(KERN_ERR "microcode: CPU%d updated from revision "
341 "0x%x to 0x%x failed\n", cpu_num, uci->rev, val[1]);
342 return;
343 }
344 pr_debug("microcode: CPU%d updated from revision "
415 "0x%x to 0x%x, date = %08x \n", 345 "0x%x to 0x%x, date = %08x \n",
416 cpu_num, uci->rev, val[1], uci->mc->hdr.date); 346 cpu_num, uci->rev, val[1], uci->mc->hdr.date);
417 return; 347 uci->rev = val[1];
418} 348}
419 349
420static int do_microcode_update (void) 350#ifdef CONFIG_MICROCODE_OLD_INTERFACE
421{ 351static void __user *user_buffer; /* user area microcode data buffer */
422 int i, error; 352static unsigned int user_buffer_size; /* it's size */
423 353
424 if (on_each_cpu(collect_cpu_info, NULL, 1, 1) != 0) { 354static long get_next_ucode(void **mc, long offset)
425 printk(KERN_ERR "microcode: Error! Could not run on all processors\n"); 355{
426 error = -EIO; 356 microcode_header_t mc_header;
427 goto out; 357 unsigned long total_size;
358
359 /* No more data */
360 if (offset >= user_buffer_size)
361 return 0;
362 if (copy_from_user(&mc_header, user_buffer + offset, MC_HEADER_SIZE)) {
363 printk(KERN_ERR "microcode: error! Can not read user data\n");
364 return -EFAULT;
428 } 365 }
429 366 total_size = get_totalsize(&mc_header);
430 if ((error = find_matching_ucodes())) { 367 if (offset + total_size > user_buffer_size) {
431 printk(KERN_ERR "microcode: Error in the microcode data\n"); 368 printk(KERN_ERR "microcode: error! Bad total size in microcode "
432 goto out_free; 369 "data file\n");
370 return -EINVAL;
433 } 371 }
434 372 *mc = vmalloc(total_size);
435 if (on_each_cpu(do_update_one, NULL, 1, 1) != 0) { 373 if (!*mc)
436 printk(KERN_ERR "microcode: Error! Could not run on all processors\n"); 374 return -ENOMEM;
437 error = -EIO; 375 if (copy_from_user(*mc, user_buffer + offset, total_size)) {
376 printk(KERN_ERR "microcode: error! Can not read user data\n");
377 vfree(*mc);
378 return -EFAULT;
438 } 379 }
380 return offset + total_size;
381}
382
383static int do_microcode_update (void)
384{
385 long cursor = 0;
386 int error = 0;
387 void *new_mc;
388 int cpu;
389 cpumask_t old;
390
391 old = current->cpus_allowed;
439 392
440out_free: 393 while ((cursor = get_next_ucode(&new_mc, cursor)) > 0) {
441 for_each_online_cpu(i) { 394 error = microcode_sanity_check(new_mc);
442 if (ucode_cpu_info[i].mc) { 395 if (error)
443 int j; 396 goto out;
444 void *tmp = ucode_cpu_info[i].mc; 397 /*
445 vfree(tmp); 398 * It's possible the data file has multiple matching ucode,
446 for_each_online_cpu(j) { 399 * lets keep searching till the latest version
447 if (ucode_cpu_info[j].mc == tmp) 400 */
448 ucode_cpu_info[j].mc = NULL; 401 for_each_online_cpu(cpu) {
449 } 402 struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
403
404 if (!uci->valid)
405 continue;
406 set_cpus_allowed(current, cpumask_of_cpu(cpu));
407 error = get_maching_microcode(new_mc, cpu);
408 if (error < 0)
409 goto out;
410 if (error == 1)
411 apply_microcode(cpu);
450 } 412 }
451 if (ucode_cpu_info[i].err == MC_IGNORED && verbose) 413 vfree(new_mc);
452 printk(KERN_WARNING "microcode: CPU%d not 'upgrading' to earlier revision"
453 " 0x%x (current=0x%x)\n", i, ucode_cpu_info[i].cksum, ucode_cpu_info[i].rev);
454 } 414 }
455out: 415out:
416 if (cursor > 0)
417 vfree(new_mc);
418 if (cursor < 0)
419 error = cursor;
420 set_cpus_allowed(current, old);
456 return error; 421 return error;
457} 422}
458 423
424static int microcode_open (struct inode *unused1, struct file *unused2)
425{
426 return capable(CAP_SYS_RAWIO) ? 0 : -EPERM;
427}
428
459static ssize_t microcode_write (struct file *file, const char __user *buf, size_t len, loff_t *ppos) 429static ssize_t microcode_write (struct file *file, const char __user *buf, size_t len, loff_t *ppos)
460{ 430{
461 ssize_t ret; 431 ssize_t ret;
462 432
463 if (len < DEFAULT_UCODE_TOTALSIZE) {
464 printk(KERN_ERR "microcode: not enough data\n");
465 return -EINVAL;
466 }
467
468 if ((len >> PAGE_SHIFT) > num_physpages) { 433 if ((len >> PAGE_SHIFT) > num_physpages) {
469 printk(KERN_ERR "microcode: too much data (max %ld pages)\n", num_physpages); 434 printk(KERN_ERR "microcode: too much data (max %ld pages)\n", num_physpages);
470 return -EINVAL; 435 return -EINVAL;
471 } 436 }
472 437
438 lock_cpu_hotplug();
473 mutex_lock(&microcode_mutex); 439 mutex_lock(&microcode_mutex);
474 440
475 user_buffer = (void __user *) buf; 441 user_buffer = (void __user *) buf;
@@ -480,6 +446,7 @@ static ssize_t microcode_write (struct file *file, const char __user *buf, size_
480 ret = (ssize_t)len; 446 ret = (ssize_t)len;
481 447
482 mutex_unlock(&microcode_mutex); 448 mutex_unlock(&microcode_mutex);
449 unlock_cpu_hotplug();
483 450
484 return ret; 451 return ret;
485} 452}
@@ -496,7 +463,7 @@ static struct miscdevice microcode_dev = {
496 .fops = &microcode_fops, 463 .fops = &microcode_fops,
497}; 464};
498 465
499static int __init microcode_init (void) 466static int __init microcode_dev_init (void)
500{ 467{
501 int error; 468 int error;
502 469
@@ -508,6 +475,280 @@ static int __init microcode_init (void)
508 return error; 475 return error;
509 } 476 }
510 477
478 return 0;
479}
480
481static void __exit microcode_dev_exit (void)
482{
483 misc_deregister(&microcode_dev);
484}
485
486MODULE_ALIAS_MISCDEV(MICROCODE_MINOR);
487#else
488#define microcode_dev_init() 0
489#define microcode_dev_exit() do { } while(0)
490#endif
491
492static long get_next_ucode_from_buffer(void **mc, void *buf,
493 unsigned long size, long offset)
494{
495 microcode_header_t *mc_header;
496 unsigned long total_size;
497
498 /* No more data */
499 if (offset >= size)
500 return 0;
501 mc_header = (microcode_header_t *)(buf + offset);
502 total_size = get_totalsize(mc_header);
503
504 if (offset + total_size > size) {
505 printk(KERN_ERR "microcode: error! Bad data in microcode data file\n");
506 return -EINVAL;
507 }
508
509 *mc = vmalloc(total_size);
510 if (!*mc) {
511 printk(KERN_ERR "microcode: error! Can not allocate memory\n");
512 return -ENOMEM;
513 }
514 memcpy(*mc, buf + offset, total_size);
515 return offset + total_size;
516}
517
518/* fake device for request_firmware */
519static struct platform_device *microcode_pdev;
520
521static int cpu_request_microcode(int cpu)
522{
523 char name[30];
524 struct cpuinfo_x86 *c = cpu_data + cpu;
525 const struct firmware *firmware;
526 void *buf;
527 unsigned long size;
528 long offset = 0;
529 int error;
530 void *mc;
531
532 /* We should bind the task to the CPU */
533 BUG_ON(cpu != raw_smp_processor_id());
534 sprintf(name,"intel-ucode/%02x-%02x-%02x",
535 c->x86, c->x86_model, c->x86_mask);
536 error = request_firmware(&firmware, name, &microcode_pdev->dev);
537 if (error) {
538 pr_debug("ucode data file %s load failed\n", name);
539 return error;
540 }
541 buf = (void *)firmware->data;
542 size = firmware->size;
543 while ((offset = get_next_ucode_from_buffer(&mc, buf, size, offset))
544 > 0) {
545 error = microcode_sanity_check(mc);
546 if (error)
547 break;
548 error = get_maching_microcode(mc, cpu);
549 if (error < 0)
550 break;
551 /*
552 * It's possible the data file has multiple matching ucode,
553 * lets keep searching till the latest version
554 */
555 if (error == 1) {
556 apply_microcode(cpu);
557 error = 0;
558 }
559 vfree(mc);
560 }
561 if (offset > 0)
562 vfree(mc);
563 if (offset < 0)
564 error = offset;
565 release_firmware(firmware);
566
567 return error;
568}
569
570static void microcode_init_cpu(int cpu)
571{
572 cpumask_t old;
573 struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
574
575 old = current->cpus_allowed;
576
577 set_cpus_allowed(current, cpumask_of_cpu(cpu));
578 mutex_lock(&microcode_mutex);
579 collect_cpu_info(cpu);
580 if (uci->valid)
581 cpu_request_microcode(cpu);
582 mutex_unlock(&microcode_mutex);
583 set_cpus_allowed(current, old);
584}
585
586static void microcode_fini_cpu(int cpu)
587{
588 struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
589
590 mutex_lock(&microcode_mutex);
591 uci->valid = 0;
592 vfree(uci->mc);
593 uci->mc = NULL;
594 mutex_unlock(&microcode_mutex);
595}
596
597static ssize_t reload_store(struct sys_device *dev, const char *buf, size_t sz)
598{
599 struct ucode_cpu_info *uci = ucode_cpu_info + dev->id;
600 char *end;
601 unsigned long val = simple_strtoul(buf, &end, 0);
602 int err = 0;
603 int cpu = dev->id;
604
605 if (end == buf)
606 return -EINVAL;
607 if (val == 1) {
608 cpumask_t old;
609
610 old = current->cpus_allowed;
611
612 lock_cpu_hotplug();
613 set_cpus_allowed(current, cpumask_of_cpu(cpu));
614
615 mutex_lock(&microcode_mutex);
616 if (uci->valid)
617 err = cpu_request_microcode(cpu);
618 mutex_unlock(&microcode_mutex);
619 unlock_cpu_hotplug();
620 set_cpus_allowed(current, old);
621 }
622 if (err)
623 return err;
624 return sz;
625}
626
627static ssize_t version_show(struct sys_device *dev, char *buf)
628{
629 struct ucode_cpu_info *uci = ucode_cpu_info + dev->id;
630
631 return sprintf(buf, "0x%x\n", uci->rev);
632}
633
634static ssize_t pf_show(struct sys_device *dev, char *buf)
635{
636 struct ucode_cpu_info *uci = ucode_cpu_info + dev->id;
637
638 return sprintf(buf, "0x%x\n", uci->pf);
639}
640
641static SYSDEV_ATTR(reload, 0200, NULL, reload_store);
642static SYSDEV_ATTR(version, 0400, version_show, NULL);
643static SYSDEV_ATTR(processor_flags, 0400, pf_show, NULL);
644
645static struct attribute *mc_default_attrs[] = {
646 &attr_reload.attr,
647 &attr_version.attr,
648 &attr_processor_flags.attr,
649 NULL
650};
651
652static struct attribute_group mc_attr_group = {
653 .attrs = mc_default_attrs,
654 .name = "microcode",
655};
656
657static int mc_sysdev_add(struct sys_device *sys_dev)
658{
659 int cpu = sys_dev->id;
660 struct ucode_cpu_info *uci = ucode_cpu_info + cpu;
661
662 if (!cpu_online(cpu))
663 return 0;
664 pr_debug("Microcode:CPU %d added\n", cpu);
665 memset(uci, 0, sizeof(*uci));
666 sysfs_create_group(&sys_dev->kobj, &mc_attr_group);
667
668 microcode_init_cpu(cpu);
669 return 0;
670}
671
672static int mc_sysdev_remove(struct sys_device *sys_dev)
673{
674 int cpu = sys_dev->id;
675
676 if (!cpu_online(cpu))
677 return 0;
678 pr_debug("Microcode:CPU %d removed\n", cpu);
679 microcode_fini_cpu(cpu);
680 sysfs_remove_group(&sys_dev->kobj, &mc_attr_group);
681 return 0;
682}
683
684static int mc_sysdev_resume(struct sys_device *dev)
685{
686 int cpu = dev->id;
687
688 if (!cpu_online(cpu))
689 return 0;
690 pr_debug("Microcode:CPU %d resumed\n", cpu);
691 /* only CPU 0 will apply ucode here */
692 apply_microcode(0);
693 return 0;
694}
695
696static struct sysdev_driver mc_sysdev_driver = {
697 .add = mc_sysdev_add,
698 .remove = mc_sysdev_remove,
699 .resume = mc_sysdev_resume,
700};
701
702#ifdef CONFIG_HOTPLUG_CPU
703static __cpuinit int
704mc_cpu_callback(struct notifier_block *nb, unsigned long action, void *hcpu)
705{
706 unsigned int cpu = (unsigned long)hcpu;
707 struct sys_device *sys_dev;
708
709 sys_dev = get_cpu_sysdev(cpu);
710 switch (action) {
711 case CPU_ONLINE:
712 case CPU_DOWN_FAILED:
713 mc_sysdev_add(sys_dev);
714 break;
715 case CPU_DOWN_PREPARE:
716 mc_sysdev_remove(sys_dev);
717 break;
718 }
719 return NOTIFY_OK;
720}
721
722static struct notifier_block mc_cpu_notifier = {
723 .notifier_call = mc_cpu_callback,
724};
725#endif
726
727static int __init microcode_init (void)
728{
729 int error;
730
731 error = microcode_dev_init();
732 if (error)
733 return error;
734 microcode_pdev = platform_device_register_simple("microcode", -1,
735 NULL, 0);
736 if (IS_ERR(microcode_pdev)) {
737 microcode_dev_exit();
738 return PTR_ERR(microcode_pdev);
739 }
740
741 lock_cpu_hotplug();
742 error = sysdev_driver_register(&cpu_sysdev_class, &mc_sysdev_driver);
743 unlock_cpu_hotplug();
744 if (error) {
745 microcode_dev_exit();
746 platform_device_unregister(microcode_pdev);
747 return error;
748 }
749
750 register_hotcpu_notifier(&mc_cpu_notifier);
751
511 printk(KERN_INFO 752 printk(KERN_INFO
512 "IA-32 Microcode Update Driver: v" MICROCODE_VERSION " <tigran@veritas.com>\n"); 753 "IA-32 Microcode Update Driver: v" MICROCODE_VERSION " <tigran@veritas.com>\n");
513 return 0; 754 return 0;
@@ -515,9 +756,16 @@ static int __init microcode_init (void)
515 756
516static void __exit microcode_exit (void) 757static void __exit microcode_exit (void)
517{ 758{
518 misc_deregister(&microcode_dev); 759 microcode_dev_exit();
760
761 unregister_hotcpu_notifier(&mc_cpu_notifier);
762
763 lock_cpu_hotplug();
764 sysdev_driver_unregister(&cpu_sysdev_class, &mc_sysdev_driver);
765 unlock_cpu_hotplug();
766
767 platform_device_unregister(microcode_pdev);
519} 768}
520 769
521module_init(microcode_init) 770module_init(microcode_init)
522module_exit(microcode_exit) 771module_exit(microcode_exit)
523MODULE_ALIAS_MISCDEV(MICROCODE_MINOR);
diff --git a/arch/i386/kernel/mpparse.c b/arch/i386/kernel/mpparse.c
index a70b5fa0ef06..442aaf8c77eb 100644
--- a/arch/i386/kernel/mpparse.c
+++ b/arch/i386/kernel/mpparse.c
@@ -30,6 +30,7 @@
30#include <asm/io_apic.h> 30#include <asm/io_apic.h>
31 31
32#include <mach_apic.h> 32#include <mach_apic.h>
33#include <mach_apicdef.h>
33#include <mach_mpparse.h> 34#include <mach_mpparse.h>
34#include <bios_ebda.h> 35#include <bios_ebda.h>
35 36
@@ -68,7 +69,7 @@ unsigned int def_to_bigsmp = 0;
68/* Processor that is doing the boot up */ 69/* Processor that is doing the boot up */
69unsigned int boot_cpu_physical_apicid = -1U; 70unsigned int boot_cpu_physical_apicid = -1U;
70/* Internal processor count */ 71/* Internal processor count */
71static unsigned int __devinitdata num_processors; 72unsigned int __cpuinitdata num_processors;
72 73
73/* Bitmask of physically existing CPUs */ 74/* Bitmask of physically existing CPUs */
74physid_mask_t phys_cpu_present_map; 75physid_mask_t phys_cpu_present_map;
@@ -228,12 +229,14 @@ static void __init MP_bus_info (struct mpc_config_bus *m)
228 229
229 mpc_oem_bus_info(m, str, translation_table[mpc_record]); 230 mpc_oem_bus_info(m, str, translation_table[mpc_record]);
230 231
232#if MAX_MP_BUSSES < 256
231 if (m->mpc_busid >= MAX_MP_BUSSES) { 233 if (m->mpc_busid >= MAX_MP_BUSSES) {
232 printk(KERN_WARNING "MP table busid value (%d) for bustype %s " 234 printk(KERN_WARNING "MP table busid value (%d) for bustype %s "
233 " is too large, max. supported is %d\n", 235 " is too large, max. supported is %d\n",
234 m->mpc_busid, str, MAX_MP_BUSSES - 1); 236 m->mpc_busid, str, MAX_MP_BUSSES - 1);
235 return; 237 return;
236 } 238 }
239#endif
237 240
238 if (strncmp(str, BUSTYPE_ISA, sizeof(BUSTYPE_ISA)-1) == 0) { 241 if (strncmp(str, BUSTYPE_ISA, sizeof(BUSTYPE_ISA)-1) == 0) {
239 mp_bus_id_to_type[m->mpc_busid] = MP_BUS_ISA; 242 mp_bus_id_to_type[m->mpc_busid] = MP_BUS_ISA;
@@ -293,19 +296,6 @@ static void __init MP_lintsrc_info (struct mpc_config_lintsrc *m)
293 m->mpc_irqtype, m->mpc_irqflag & 3, 296 m->mpc_irqtype, m->mpc_irqflag & 3,
294 (m->mpc_irqflag >> 2) &3, m->mpc_srcbusid, 297 (m->mpc_irqflag >> 2) &3, m->mpc_srcbusid,
295 m->mpc_srcbusirq, m->mpc_destapic, m->mpc_destapiclint); 298 m->mpc_srcbusirq, m->mpc_destapic, m->mpc_destapiclint);
296 /*
297 * Well it seems all SMP boards in existence
298 * use ExtINT/LVT1 == LINT0 and
299 * NMI/LVT2 == LINT1 - the following check
300 * will show us if this assumptions is false.
301 * Until then we do not have to add baggage.
302 */
303 if ((m->mpc_irqtype == mp_ExtINT) &&
304 (m->mpc_destapiclint != 0))
305 BUG();
306 if ((m->mpc_irqtype == mp_NMI) &&
307 (m->mpc_destapiclint != 1))
308 BUG();
309} 299}
310 300
311#ifdef CONFIG_X86_NUMAQ 301#ifdef CONFIG_X86_NUMAQ
@@ -822,8 +812,7 @@ int es7000_plat;
822 812
823#ifdef CONFIG_ACPI 813#ifdef CONFIG_ACPI
824 814
825void __init mp_register_lapic_address ( 815void __init mp_register_lapic_address(u64 address)
826 u64 address)
827{ 816{
828 mp_lapic_addr = (unsigned long) address; 817 mp_lapic_addr = (unsigned long) address;
829 818
@@ -835,13 +824,10 @@ void __init mp_register_lapic_address (
835 Dprintk("Boot CPU = %d\n", boot_cpu_physical_apicid); 824 Dprintk("Boot CPU = %d\n", boot_cpu_physical_apicid);
836} 825}
837 826
838 827void __devinit mp_register_lapic (u8 id, u8 enabled)
839void __devinit mp_register_lapic (
840 u8 id,
841 u8 enabled)
842{ 828{
843 struct mpc_config_processor processor; 829 struct mpc_config_processor processor;
844 int boot_cpu = 0; 830 int boot_cpu = 0;
845 831
846 if (MAX_APICS - id <= 0) { 832 if (MAX_APICS - id <= 0) {
847 printk(KERN_WARNING "Processor #%d invalid (max %d)\n", 833 printk(KERN_WARNING "Processor #%d invalid (max %d)\n",
@@ -878,11 +864,9 @@ static struct mp_ioapic_routing {
878 u32 pin_programmed[4]; 864 u32 pin_programmed[4];
879} mp_ioapic_routing[MAX_IO_APICS]; 865} mp_ioapic_routing[MAX_IO_APICS];
880 866
881 867static int mp_find_ioapic (int gsi)
882static int mp_find_ioapic (
883 int gsi)
884{ 868{
885 int i = 0; 869 int i = 0;
886 870
887 /* Find the IOAPIC that manages this GSI. */ 871 /* Find the IOAPIC that manages this GSI. */
888 for (i = 0; i < nr_ioapics; i++) { 872 for (i = 0; i < nr_ioapics; i++) {
@@ -895,15 +879,11 @@ static int mp_find_ioapic (
895 879
896 return -1; 880 return -1;
897} 881}
898
899 882
900void __init mp_register_ioapic ( 883void __init mp_register_ioapic(u8 id, u32 address, u32 gsi_base)
901 u8 id,
902 u32 address,
903 u32 gsi_base)
904{ 884{
905 int idx = 0; 885 int idx = 0;
906 int tmpid; 886 int tmpid;
907 887
908 if (nr_ioapics >= MAX_IO_APICS) { 888 if (nr_ioapics >= MAX_IO_APICS) {
909 printk(KERN_ERR "ERROR: Max # of I/O APICs (%d) exceeded " 889 printk(KERN_ERR "ERROR: Max # of I/O APICs (%d) exceeded "
@@ -949,16 +929,10 @@ void __init mp_register_ioapic (
949 mp_ioapics[idx].mpc_apicver, mp_ioapics[idx].mpc_apicaddr, 929 mp_ioapics[idx].mpc_apicver, mp_ioapics[idx].mpc_apicaddr,
950 mp_ioapic_routing[idx].gsi_base, 930 mp_ioapic_routing[idx].gsi_base,
951 mp_ioapic_routing[idx].gsi_end); 931 mp_ioapic_routing[idx].gsi_end);
952
953 return;
954} 932}
955 933
956 934void __init
957void __init mp_override_legacy_irq ( 935mp_override_legacy_irq(u8 bus_irq, u8 polarity, u8 trigger, u32 gsi)
958 u8 bus_irq,
959 u8 polarity,
960 u8 trigger,
961 u32 gsi)
962{ 936{
963 struct mpc_config_intsrc intsrc; 937 struct mpc_config_intsrc intsrc;
964 int ioapic = -1; 938 int ioapic = -1;
@@ -996,15 +970,13 @@ void __init mp_override_legacy_irq (
996 mp_irqs[mp_irq_entries] = intsrc; 970 mp_irqs[mp_irq_entries] = intsrc;
997 if (++mp_irq_entries == MAX_IRQ_SOURCES) 971 if (++mp_irq_entries == MAX_IRQ_SOURCES)
998 panic("Max # of irq sources exceeded!\n"); 972 panic("Max # of irq sources exceeded!\n");
999
1000 return;
1001} 973}
1002 974
1003void __init mp_config_acpi_legacy_irqs (void) 975void __init mp_config_acpi_legacy_irqs (void)
1004{ 976{
1005 struct mpc_config_intsrc intsrc; 977 struct mpc_config_intsrc intsrc;
1006 int i = 0; 978 int i = 0;
1007 int ioapic = -1; 979 int ioapic = -1;
1008 980
1009 /* 981 /*
1010 * Fabricate the legacy ISA bus (bus #31). 982 * Fabricate the legacy ISA bus (bus #31).
@@ -1073,12 +1045,12 @@ void __init mp_config_acpi_legacy_irqs (void)
1073 1045
1074#define MAX_GSI_NUM 4096 1046#define MAX_GSI_NUM 4096
1075 1047
1076int mp_register_gsi (u32 gsi, int triggering, int polarity) 1048int mp_register_gsi(u32 gsi, int triggering, int polarity)
1077{ 1049{
1078 int ioapic = -1; 1050 int ioapic = -1;
1079 int ioapic_pin = 0; 1051 int ioapic_pin = 0;
1080 int idx, bit = 0; 1052 int idx, bit = 0;
1081 static int pci_irq = 16; 1053 static int pci_irq = 16;
1082 /* 1054 /*
1083 * Mapping between Global System Interrups, which 1055 * Mapping between Global System Interrups, which
1084 * represent all possible interrupts, and IRQs 1056 * represent all possible interrupts, and IRQs
diff --git a/arch/i386/kernel/nmi.c b/arch/i386/kernel/nmi.c
index acb351478e42..dbda706fdd14 100644
--- a/arch/i386/kernel/nmi.c
+++ b/arch/i386/kernel/nmi.c
@@ -21,83 +21,174 @@
21#include <linux/sysdev.h> 21#include <linux/sysdev.h>
22#include <linux/sysctl.h> 22#include <linux/sysctl.h>
23#include <linux/percpu.h> 23#include <linux/percpu.h>
24#include <linux/dmi.h>
25#include <linux/kprobes.h>
24 26
25#include <asm/smp.h> 27#include <asm/smp.h>
26#include <asm/nmi.h> 28#include <asm/nmi.h>
29#include <asm/kdebug.h>
27#include <asm/intel_arch_perfmon.h> 30#include <asm/intel_arch_perfmon.h>
28 31
29#include "mach_traps.h" 32#include "mach_traps.h"
30 33
31unsigned int nmi_watchdog = NMI_NONE; 34/* perfctr_nmi_owner tracks the ownership of the perfctr registers:
32extern int unknown_nmi_panic; 35 * evtsel_nmi_owner tracks the ownership of the event selection
33static unsigned int nmi_hz = HZ; 36 * - different performance counters/ event selection may be reserved for
34static unsigned int nmi_perfctr_msr; /* the MSR to reset in NMI handler */ 37 * different subsystems this reservation system just tries to coordinate
35static unsigned int nmi_p4_cccr_val; 38 * things a little
36extern void show_registers(struct pt_regs *regs); 39 */
40static DEFINE_PER_CPU(unsigned long, perfctr_nmi_owner);
41static DEFINE_PER_CPU(unsigned long, evntsel_nmi_owner[3]);
37 42
38/* 43/* this number is calculated from Intel's MSR_P4_CRU_ESCR5 register and it's
39 * lapic_nmi_owner tracks the ownership of the lapic NMI hardware: 44 * offset from MSR_P4_BSU_ESCR0. It will be the max for all platforms (for now)
40 * - it may be reserved by some other driver, or not
41 * - when not reserved by some other driver, it may be used for
42 * the NMI watchdog, or not
43 *
44 * This is maintained separately from nmi_active because the NMI
45 * watchdog may also be driven from the I/O APIC timer.
46 */ 45 */
47static DEFINE_SPINLOCK(lapic_nmi_owner_lock); 46#define NMI_MAX_COUNTER_BITS 66
48static unsigned int lapic_nmi_owner;
49#define LAPIC_NMI_WATCHDOG (1<<0)
50#define LAPIC_NMI_RESERVED (1<<1)
51 47
52/* nmi_active: 48/* nmi_active:
53 * +1: the lapic NMI watchdog is active, but can be disabled 49 * >0: the lapic NMI watchdog is active, but can be disabled
54 * 0: the lapic NMI watchdog has not been set up, and cannot 50 * <0: the lapic NMI watchdog has not been set up, and cannot
55 * be enabled 51 * be enabled
56 * -1: the lapic NMI watchdog is disabled, but can be enabled 52 * 0: the lapic NMI watchdog is disabled, but can be enabled
57 */ 53 */
58int nmi_active; 54atomic_t nmi_active = ATOMIC_INIT(0); /* oprofile uses this */
59 55
60#define K7_EVNTSEL_ENABLE (1 << 22) 56unsigned int nmi_watchdog = NMI_DEFAULT;
61#define K7_EVNTSEL_INT (1 << 20) 57static unsigned int nmi_hz = HZ;
62#define K7_EVNTSEL_OS (1 << 17)
63#define K7_EVNTSEL_USR (1 << 16)
64#define K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING 0x76
65#define K7_NMI_EVENT K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING
66 58
67#define P6_EVNTSEL0_ENABLE (1 << 22) 59struct nmi_watchdog_ctlblk {
68#define P6_EVNTSEL_INT (1 << 20) 60 int enabled;
69#define P6_EVNTSEL_OS (1 << 17) 61 u64 check_bit;
70#define P6_EVNTSEL_USR (1 << 16) 62 unsigned int cccr_msr;
71#define P6_EVENT_CPU_CLOCKS_NOT_HALTED 0x79 63 unsigned int perfctr_msr; /* the MSR to reset in NMI handler */
72#define P6_NMI_EVENT P6_EVENT_CPU_CLOCKS_NOT_HALTED 64 unsigned int evntsel_msr; /* the MSR to select the events to handle */
65};
66static DEFINE_PER_CPU(struct nmi_watchdog_ctlblk, nmi_watchdog_ctlblk);
73 67
74#define MSR_P4_MISC_ENABLE 0x1A0 68/* local prototypes */
75#define MSR_P4_MISC_ENABLE_PERF_AVAIL (1<<7) 69static int unknown_nmi_panic_callback(struct pt_regs *regs, int cpu);
76#define MSR_P4_MISC_ENABLE_PEBS_UNAVAIL (1<<12)
77#define MSR_P4_PERFCTR0 0x300
78#define MSR_P4_CCCR0 0x360
79#define P4_ESCR_EVENT_SELECT(N) ((N)<<25)
80#define P4_ESCR_OS (1<<3)
81#define P4_ESCR_USR (1<<2)
82#define P4_CCCR_OVF_PMI0 (1<<26)
83#define P4_CCCR_OVF_PMI1 (1<<27)
84#define P4_CCCR_THRESHOLD(N) ((N)<<20)
85#define P4_CCCR_COMPLEMENT (1<<19)
86#define P4_CCCR_COMPARE (1<<18)
87#define P4_CCCR_REQUIRED (3<<16)
88#define P4_CCCR_ESCR_SELECT(N) ((N)<<13)
89#define P4_CCCR_ENABLE (1<<12)
90/* Set up IQ_COUNTER0 to behave like a clock, by having IQ_CCCR0 filter
91 CRU_ESCR0 (with any non-null event selector) through a complemented
92 max threshold. [IA32-Vol3, Section 14.9.9] */
93#define MSR_P4_IQ_COUNTER0 0x30C
94#define P4_NMI_CRU_ESCR0 (P4_ESCR_EVENT_SELECT(0x3F)|P4_ESCR_OS|P4_ESCR_USR)
95#define P4_NMI_IQ_CCCR0 \
96 (P4_CCCR_OVF_PMI0|P4_CCCR_THRESHOLD(15)|P4_CCCR_COMPLEMENT| \
97 P4_CCCR_COMPARE|P4_CCCR_REQUIRED|P4_CCCR_ESCR_SELECT(4)|P4_CCCR_ENABLE)
98 70
99#define ARCH_PERFMON_NMI_EVENT_SEL ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL 71extern void show_registers(struct pt_regs *regs);
100#define ARCH_PERFMON_NMI_EVENT_UMASK ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK 72extern int unknown_nmi_panic;
73
74/* converts an msr to an appropriate reservation bit */
75static inline unsigned int nmi_perfctr_msr_to_bit(unsigned int msr)
76{
77 /* returns the bit offset of the performance counter register */
78 switch (boot_cpu_data.x86_vendor) {
79 case X86_VENDOR_AMD:
80 return (msr - MSR_K7_PERFCTR0);
81 case X86_VENDOR_INTEL:
82 if (cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON))
83 return (msr - MSR_ARCH_PERFMON_PERFCTR0);
84
85 switch (boot_cpu_data.x86) {
86 case 6:
87 return (msr - MSR_P6_PERFCTR0);
88 case 15:
89 return (msr - MSR_P4_BPU_PERFCTR0);
90 }
91 }
92 return 0;
93}
94
95/* converts an msr to an appropriate reservation bit */
96static inline unsigned int nmi_evntsel_msr_to_bit(unsigned int msr)
97{
98 /* returns the bit offset of the event selection register */
99 switch (boot_cpu_data.x86_vendor) {
100 case X86_VENDOR_AMD:
101 return (msr - MSR_K7_EVNTSEL0);
102 case X86_VENDOR_INTEL:
103 if (cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON))
104 return (msr - MSR_ARCH_PERFMON_EVENTSEL0);
105
106 switch (boot_cpu_data.x86) {
107 case 6:
108 return (msr - MSR_P6_EVNTSEL0);
109 case 15:
110 return (msr - MSR_P4_BSU_ESCR0);
111 }
112 }
113 return 0;
114}
115
116/* checks for a bit availability (hack for oprofile) */
117int avail_to_resrv_perfctr_nmi_bit(unsigned int counter)
118{
119 BUG_ON(counter > NMI_MAX_COUNTER_BITS);
120
121 return (!test_bit(counter, &__get_cpu_var(perfctr_nmi_owner)));
122}
123
124/* checks the an msr for availability */
125int avail_to_resrv_perfctr_nmi(unsigned int msr)
126{
127 unsigned int counter;
128
129 counter = nmi_perfctr_msr_to_bit(msr);
130 BUG_ON(counter > NMI_MAX_COUNTER_BITS);
131
132 return (!test_bit(counter, &__get_cpu_var(perfctr_nmi_owner)));
133}
134
135int reserve_perfctr_nmi(unsigned int msr)
136{
137 unsigned int counter;
138
139 counter = nmi_perfctr_msr_to_bit(msr);
140 BUG_ON(counter > NMI_MAX_COUNTER_BITS);
141
142 if (!test_and_set_bit(counter, &__get_cpu_var(perfctr_nmi_owner)))
143 return 1;
144 return 0;
145}
146
147void release_perfctr_nmi(unsigned int msr)
148{
149 unsigned int counter;
150
151 counter = nmi_perfctr_msr_to_bit(msr);
152 BUG_ON(counter > NMI_MAX_COUNTER_BITS);
153
154 clear_bit(counter, &__get_cpu_var(perfctr_nmi_owner));
155}
156
157int reserve_evntsel_nmi(unsigned int msr)
158{
159 unsigned int counter;
160
161 counter = nmi_evntsel_msr_to_bit(msr);
162 BUG_ON(counter > NMI_MAX_COUNTER_BITS);
163
164 if (!test_and_set_bit(counter, &__get_cpu_var(evntsel_nmi_owner)[0]))
165 return 1;
166 return 0;
167}
168
169void release_evntsel_nmi(unsigned int msr)
170{
171 unsigned int counter;
172
173 counter = nmi_evntsel_msr_to_bit(msr);
174 BUG_ON(counter > NMI_MAX_COUNTER_BITS);
175
176 clear_bit(counter, &__get_cpu_var(evntsel_nmi_owner)[0]);
177}
178
179static __cpuinit inline int nmi_known_cpu(void)
180{
181 switch (boot_cpu_data.x86_vendor) {
182 case X86_VENDOR_AMD:
183 return ((boot_cpu_data.x86 == 15) || (boot_cpu_data.x86 == 6));
184 case X86_VENDOR_INTEL:
185 if (cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON))
186 return 1;
187 else
188 return ((boot_cpu_data.x86 == 15) || (boot_cpu_data.x86 == 6));
189 }
190 return 0;
191}
101 192
102#ifdef CONFIG_SMP 193#ifdef CONFIG_SMP
103/* The performance counters used by NMI_LOCAL_APIC don't trigger when 194/* The performance counters used by NMI_LOCAL_APIC don't trigger when
@@ -125,7 +216,18 @@ static int __init check_nmi_watchdog(void)
125 unsigned int *prev_nmi_count; 216 unsigned int *prev_nmi_count;
126 int cpu; 217 int cpu;
127 218
128 if (nmi_watchdog == NMI_NONE) 219 /* Enable NMI watchdog for newer systems.
220 Actually it should be safe for most systems before 2004 too except
221 for some IBM systems that corrupt registers when NMI happens
222 during SMM. Unfortunately we don't have more exact information
223 on these and use this coarse check. */
224 if (nmi_watchdog == NMI_DEFAULT && dmi_get_year(DMI_BIOS_DATE) >= 2004)
225 nmi_watchdog = NMI_LOCAL_APIC;
226
227 if ((nmi_watchdog == NMI_NONE) || (nmi_watchdog == NMI_DEFAULT))
228 return 0;
229
230 if (!atomic_read(&nmi_active))
129 return 0; 231 return 0;
130 232
131 prev_nmi_count = kmalloc(NR_CPUS * sizeof(int), GFP_KERNEL); 233 prev_nmi_count = kmalloc(NR_CPUS * sizeof(int), GFP_KERNEL);
@@ -149,25 +251,45 @@ static int __init check_nmi_watchdog(void)
149 if (!cpu_isset(cpu, cpu_callin_map)) 251 if (!cpu_isset(cpu, cpu_callin_map))
150 continue; 252 continue;
151#endif 253#endif
254 if (!per_cpu(nmi_watchdog_ctlblk, cpu).enabled)
255 continue;
152 if (nmi_count(cpu) - prev_nmi_count[cpu] <= 5) { 256 if (nmi_count(cpu) - prev_nmi_count[cpu] <= 5) {
153 endflag = 1;
154 printk("CPU#%d: NMI appears to be stuck (%d->%d)!\n", 257 printk("CPU#%d: NMI appears to be stuck (%d->%d)!\n",
155 cpu, 258 cpu,
156 prev_nmi_count[cpu], 259 prev_nmi_count[cpu],
157 nmi_count(cpu)); 260 nmi_count(cpu));
158 nmi_active = 0; 261 per_cpu(nmi_watchdog_ctlblk, cpu).enabled = 0;
159 lapic_nmi_owner &= ~LAPIC_NMI_WATCHDOG; 262 atomic_dec(&nmi_active);
160 kfree(prev_nmi_count);
161 return -1;
162 } 263 }
163 } 264 }
265 if (!atomic_read(&nmi_active)) {
266 kfree(prev_nmi_count);
267 atomic_set(&nmi_active, -1);
268 return -1;
269 }
164 endflag = 1; 270 endflag = 1;
165 printk("OK.\n"); 271 printk("OK.\n");
166 272
167 /* now that we know it works we can reduce NMI frequency to 273 /* now that we know it works we can reduce NMI frequency to
168 something more reasonable; makes a difference in some configs */ 274 something more reasonable; makes a difference in some configs */
169 if (nmi_watchdog == NMI_LOCAL_APIC) 275 if (nmi_watchdog == NMI_LOCAL_APIC) {
276 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
277
170 nmi_hz = 1; 278 nmi_hz = 1;
279 /*
280 * On Intel CPUs with ARCH_PERFMON only 32 bits in the counter
281 * are writable, with higher bits sign extending from bit 31.
282 * So, we can only program the counter with 31 bit values and
283 * 32nd bit should be 1, for 33.. to be 1.
284 * Find the appropriate nmi_hz
285 */
286 if (wd->perfctr_msr == MSR_ARCH_PERFMON_PERFCTR0 &&
287 ((u64)cpu_khz * 1000) > 0x7fffffffULL) {
288 u64 count = (u64)cpu_khz * 1000;
289 do_div(count, 0x7fffffffUL);
290 nmi_hz = count + 1;
291 }
292 }
171 293
172 kfree(prev_nmi_count); 294 kfree(prev_nmi_count);
173 return 0; 295 return 0;
@@ -181,124 +303,70 @@ static int __init setup_nmi_watchdog(char *str)
181 303
182 get_option(&str, &nmi); 304 get_option(&str, &nmi);
183 305
184 if (nmi >= NMI_INVALID) 306 if ((nmi >= NMI_INVALID) || (nmi < NMI_NONE))
185 return 0; 307 return 0;
186 if (nmi == NMI_NONE)
187 nmi_watchdog = nmi;
188 /* 308 /*
189 * If any other x86 CPU has a local APIC, then 309 * If any other x86 CPU has a local APIC, then
190 * please test the NMI stuff there and send me the 310 * please test the NMI stuff there and send me the
191 * missing bits. Right now Intel P6/P4 and AMD K7 only. 311 * missing bits. Right now Intel P6/P4 and AMD K7 only.
192 */ 312 */
193 if ((nmi == NMI_LOCAL_APIC) && 313 if ((nmi == NMI_LOCAL_APIC) && (nmi_known_cpu() == 0))
194 (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) && 314 return 0; /* no lapic support */
195 (boot_cpu_data.x86 == 6 || boot_cpu_data.x86 == 15)) 315 nmi_watchdog = nmi;
196 nmi_watchdog = nmi;
197 if ((nmi == NMI_LOCAL_APIC) &&
198 (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) &&
199 (boot_cpu_data.x86 == 6 || boot_cpu_data.x86 == 15))
200 nmi_watchdog = nmi;
201 /*
202 * We can enable the IO-APIC watchdog
203 * unconditionally.
204 */
205 if (nmi == NMI_IO_APIC) {
206 nmi_active = 1;
207 nmi_watchdog = nmi;
208 }
209 return 1; 316 return 1;
210} 317}
211 318
212__setup("nmi_watchdog=", setup_nmi_watchdog); 319__setup("nmi_watchdog=", setup_nmi_watchdog);
213 320
214static void disable_intel_arch_watchdog(void);
215
216static void disable_lapic_nmi_watchdog(void) 321static void disable_lapic_nmi_watchdog(void)
217{ 322{
218 if (nmi_active <= 0) 323 BUG_ON(nmi_watchdog != NMI_LOCAL_APIC);
324
325 if (atomic_read(&nmi_active) <= 0)
219 return; 326 return;
220 switch (boot_cpu_data.x86_vendor) {
221 case X86_VENDOR_AMD:
222 wrmsr(MSR_K7_EVNTSEL0, 0, 0);
223 break;
224 case X86_VENDOR_INTEL:
225 if (cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) {
226 disable_intel_arch_watchdog();
227 break;
228 }
229 switch (boot_cpu_data.x86) {
230 case 6:
231 if (boot_cpu_data.x86_model > 0xd)
232 break;
233 327
234 wrmsr(MSR_P6_EVNTSEL0, 0, 0); 328 on_each_cpu(stop_apic_nmi_watchdog, NULL, 0, 1);
235 break;
236 case 15:
237 if (boot_cpu_data.x86_model > 0x4)
238 break;
239 329
240 wrmsr(MSR_P4_IQ_CCCR0, 0, 0); 330 BUG_ON(atomic_read(&nmi_active) != 0);
241 wrmsr(MSR_P4_CRU_ESCR0, 0, 0);
242 break;
243 }
244 break;
245 }
246 nmi_active = -1;
247 /* tell do_nmi() and others that we're not active any more */
248 nmi_watchdog = 0;
249} 331}
250 332
251static void enable_lapic_nmi_watchdog(void) 333static void enable_lapic_nmi_watchdog(void)
252{ 334{
253 if (nmi_active < 0) { 335 BUG_ON(nmi_watchdog != NMI_LOCAL_APIC);
254 nmi_watchdog = NMI_LOCAL_APIC;
255 setup_apic_nmi_watchdog();
256 }
257}
258 336
259int reserve_lapic_nmi(void) 337 /* are we already enabled */
260{ 338 if (atomic_read(&nmi_active) != 0)
261 unsigned int old_owner; 339 return;
262
263 spin_lock(&lapic_nmi_owner_lock);
264 old_owner = lapic_nmi_owner;
265 lapic_nmi_owner |= LAPIC_NMI_RESERVED;
266 spin_unlock(&lapic_nmi_owner_lock);
267 if (old_owner & LAPIC_NMI_RESERVED)
268 return -EBUSY;
269 if (old_owner & LAPIC_NMI_WATCHDOG)
270 disable_lapic_nmi_watchdog();
271 return 0;
272}
273 340
274void release_lapic_nmi(void) 341 /* are we lapic aware */
275{ 342 if (nmi_known_cpu() <= 0)
276 unsigned int new_owner; 343 return;
277 344
278 spin_lock(&lapic_nmi_owner_lock); 345 on_each_cpu(setup_apic_nmi_watchdog, NULL, 0, 1);
279 new_owner = lapic_nmi_owner & ~LAPIC_NMI_RESERVED; 346 touch_nmi_watchdog();
280 lapic_nmi_owner = new_owner;
281 spin_unlock(&lapic_nmi_owner_lock);
282 if (new_owner & LAPIC_NMI_WATCHDOG)
283 enable_lapic_nmi_watchdog();
284} 347}
285 348
286void disable_timer_nmi_watchdog(void) 349void disable_timer_nmi_watchdog(void)
287{ 350{
288 if ((nmi_watchdog != NMI_IO_APIC) || (nmi_active <= 0)) 351 BUG_ON(nmi_watchdog != NMI_IO_APIC);
352
353 if (atomic_read(&nmi_active) <= 0)
289 return; 354 return;
290 355
291 unset_nmi_callback(); 356 disable_irq(0);
292 nmi_active = -1; 357 on_each_cpu(stop_apic_nmi_watchdog, NULL, 0, 1);
293 nmi_watchdog = NMI_NONE; 358
359 BUG_ON(atomic_read(&nmi_active) != 0);
294} 360}
295 361
296void enable_timer_nmi_watchdog(void) 362void enable_timer_nmi_watchdog(void)
297{ 363{
298 if (nmi_active < 0) { 364 BUG_ON(nmi_watchdog != NMI_IO_APIC);
299 nmi_watchdog = NMI_IO_APIC; 365
366 if (atomic_read(&nmi_active) == 0) {
300 touch_nmi_watchdog(); 367 touch_nmi_watchdog();
301 nmi_active = 1; 368 on_each_cpu(setup_apic_nmi_watchdog, NULL, 0, 1);
369 enable_irq(0);
302 } 370 }
303} 371}
304 372
@@ -308,15 +376,20 @@ static int nmi_pm_active; /* nmi_active before suspend */
308 376
309static int lapic_nmi_suspend(struct sys_device *dev, pm_message_t state) 377static int lapic_nmi_suspend(struct sys_device *dev, pm_message_t state)
310{ 378{
311 nmi_pm_active = nmi_active; 379 /* only CPU0 goes here, other CPUs should be offline */
312 disable_lapic_nmi_watchdog(); 380 nmi_pm_active = atomic_read(&nmi_active);
381 stop_apic_nmi_watchdog(NULL);
382 BUG_ON(atomic_read(&nmi_active) != 0);
313 return 0; 383 return 0;
314} 384}
315 385
316static int lapic_nmi_resume(struct sys_device *dev) 386static int lapic_nmi_resume(struct sys_device *dev)
317{ 387{
318 if (nmi_pm_active > 0) 388 /* only CPU0 goes here, other CPUs should be offline */
319 enable_lapic_nmi_watchdog(); 389 if (nmi_pm_active > 0) {
390 setup_apic_nmi_watchdog(NULL);
391 touch_nmi_watchdog();
392 }
320 return 0; 393 return 0;
321} 394}
322 395
@@ -336,7 +409,13 @@ static int __init init_lapic_nmi_sysfs(void)
336{ 409{
337 int error; 410 int error;
338 411
339 if (nmi_active == 0 || nmi_watchdog != NMI_LOCAL_APIC) 412 /* should really be a BUG_ON but b/c this is an
413 * init call, it just doesn't work. -dcz
414 */
415 if (nmi_watchdog != NMI_LOCAL_APIC)
416 return 0;
417
418 if ( atomic_read(&nmi_active) < 0 )
340 return 0; 419 return 0;
341 420
342 error = sysdev_class_register(&nmi_sysclass); 421 error = sysdev_class_register(&nmi_sysclass);
@@ -354,138 +433,269 @@ late_initcall(init_lapic_nmi_sysfs);
354 * Original code written by Keith Owens. 433 * Original code written by Keith Owens.
355 */ 434 */
356 435
357static void clear_msr_range(unsigned int base, unsigned int n) 436static void write_watchdog_counter(unsigned int perfctr_msr, const char *descr)
358{
359 unsigned int i;
360
361 for(i = 0; i < n; ++i)
362 wrmsr(base+i, 0, 0);
363}
364
365static void write_watchdog_counter(const char *descr)
366{ 437{
367 u64 count = (u64)cpu_khz * 1000; 438 u64 count = (u64)cpu_khz * 1000;
368 439
369 do_div(count, nmi_hz); 440 do_div(count, nmi_hz);
370 if(descr) 441 if(descr)
371 Dprintk("setting %s to -0x%08Lx\n", descr, count); 442 Dprintk("setting %s to -0x%08Lx\n", descr, count);
372 wrmsrl(nmi_perfctr_msr, 0 - count); 443 wrmsrl(perfctr_msr, 0 - count);
373} 444}
374 445
375static void setup_k7_watchdog(void) 446/* Note that these events don't tick when the CPU idles. This means
447 the frequency varies with CPU load. */
448
449#define K7_EVNTSEL_ENABLE (1 << 22)
450#define K7_EVNTSEL_INT (1 << 20)
451#define K7_EVNTSEL_OS (1 << 17)
452#define K7_EVNTSEL_USR (1 << 16)
453#define K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING 0x76
454#define K7_NMI_EVENT K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING
455
456static int setup_k7_watchdog(void)
376{ 457{
458 unsigned int perfctr_msr, evntsel_msr;
377 unsigned int evntsel; 459 unsigned int evntsel;
460 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
461
462 perfctr_msr = MSR_K7_PERFCTR0;
463 evntsel_msr = MSR_K7_EVNTSEL0;
464 if (!reserve_perfctr_nmi(perfctr_msr))
465 goto fail;
378 466
379 nmi_perfctr_msr = MSR_K7_PERFCTR0; 467 if (!reserve_evntsel_nmi(evntsel_msr))
468 goto fail1;
380 469
381 clear_msr_range(MSR_K7_EVNTSEL0, 4); 470 wrmsrl(perfctr_msr, 0UL);
382 clear_msr_range(MSR_K7_PERFCTR0, 4);
383 471
384 evntsel = K7_EVNTSEL_INT 472 evntsel = K7_EVNTSEL_INT
385 | K7_EVNTSEL_OS 473 | K7_EVNTSEL_OS
386 | K7_EVNTSEL_USR 474 | K7_EVNTSEL_USR
387 | K7_NMI_EVENT; 475 | K7_NMI_EVENT;
388 476
389 wrmsr(MSR_K7_EVNTSEL0, evntsel, 0); 477 /* setup the timer */
390 write_watchdog_counter("K7_PERFCTR0"); 478 wrmsr(evntsel_msr, evntsel, 0);
479 write_watchdog_counter(perfctr_msr, "K7_PERFCTR0");
391 apic_write(APIC_LVTPC, APIC_DM_NMI); 480 apic_write(APIC_LVTPC, APIC_DM_NMI);
392 evntsel |= K7_EVNTSEL_ENABLE; 481 evntsel |= K7_EVNTSEL_ENABLE;
393 wrmsr(MSR_K7_EVNTSEL0, evntsel, 0); 482 wrmsr(evntsel_msr, evntsel, 0);
483
484 wd->perfctr_msr = perfctr_msr;
485 wd->evntsel_msr = evntsel_msr;
486 wd->cccr_msr = 0; //unused
487 wd->check_bit = 1ULL<<63;
488 return 1;
489fail1:
490 release_perfctr_nmi(perfctr_msr);
491fail:
492 return 0;
493}
494
495static void stop_k7_watchdog(void)
496{
497 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
498
499 wrmsr(wd->evntsel_msr, 0, 0);
500
501 release_evntsel_nmi(wd->evntsel_msr);
502 release_perfctr_nmi(wd->perfctr_msr);
394} 503}
395 504
396static void setup_p6_watchdog(void) 505#define P6_EVNTSEL0_ENABLE (1 << 22)
506#define P6_EVNTSEL_INT (1 << 20)
507#define P6_EVNTSEL_OS (1 << 17)
508#define P6_EVNTSEL_USR (1 << 16)
509#define P6_EVENT_CPU_CLOCKS_NOT_HALTED 0x79
510#define P6_NMI_EVENT P6_EVENT_CPU_CLOCKS_NOT_HALTED
511
512static int setup_p6_watchdog(void)
397{ 513{
514 unsigned int perfctr_msr, evntsel_msr;
398 unsigned int evntsel; 515 unsigned int evntsel;
516 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
517
518 perfctr_msr = MSR_P6_PERFCTR0;
519 evntsel_msr = MSR_P6_EVNTSEL0;
520 if (!reserve_perfctr_nmi(perfctr_msr))
521 goto fail;
399 522
400 nmi_perfctr_msr = MSR_P6_PERFCTR0; 523 if (!reserve_evntsel_nmi(evntsel_msr))
524 goto fail1;
401 525
402 clear_msr_range(MSR_P6_EVNTSEL0, 2); 526 wrmsrl(perfctr_msr, 0UL);
403 clear_msr_range(MSR_P6_PERFCTR0, 2);
404 527
405 evntsel = P6_EVNTSEL_INT 528 evntsel = P6_EVNTSEL_INT
406 | P6_EVNTSEL_OS 529 | P6_EVNTSEL_OS
407 | P6_EVNTSEL_USR 530 | P6_EVNTSEL_USR
408 | P6_NMI_EVENT; 531 | P6_NMI_EVENT;
409 532
410 wrmsr(MSR_P6_EVNTSEL0, evntsel, 0); 533 /* setup the timer */
411 write_watchdog_counter("P6_PERFCTR0"); 534 wrmsr(evntsel_msr, evntsel, 0);
535 write_watchdog_counter(perfctr_msr, "P6_PERFCTR0");
412 apic_write(APIC_LVTPC, APIC_DM_NMI); 536 apic_write(APIC_LVTPC, APIC_DM_NMI);
413 evntsel |= P6_EVNTSEL0_ENABLE; 537 evntsel |= P6_EVNTSEL0_ENABLE;
414 wrmsr(MSR_P6_EVNTSEL0, evntsel, 0); 538 wrmsr(evntsel_msr, evntsel, 0);
539
540 wd->perfctr_msr = perfctr_msr;
541 wd->evntsel_msr = evntsel_msr;
542 wd->cccr_msr = 0; //unused
543 wd->check_bit = 1ULL<<39;
544 return 1;
545fail1:
546 release_perfctr_nmi(perfctr_msr);
547fail:
548 return 0;
549}
550
551static void stop_p6_watchdog(void)
552{
553 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
554
555 wrmsr(wd->evntsel_msr, 0, 0);
556
557 release_evntsel_nmi(wd->evntsel_msr);
558 release_perfctr_nmi(wd->perfctr_msr);
415} 559}
416 560
561/* Note that these events don't tick when the CPU idles. This means
562 the frequency varies with CPU load. */
563
564#define MSR_P4_MISC_ENABLE_PERF_AVAIL (1<<7)
565#define P4_ESCR_EVENT_SELECT(N) ((N)<<25)
566#define P4_ESCR_OS (1<<3)
567#define P4_ESCR_USR (1<<2)
568#define P4_CCCR_OVF_PMI0 (1<<26)
569#define P4_CCCR_OVF_PMI1 (1<<27)
570#define P4_CCCR_THRESHOLD(N) ((N)<<20)
571#define P4_CCCR_COMPLEMENT (1<<19)
572#define P4_CCCR_COMPARE (1<<18)
573#define P4_CCCR_REQUIRED (3<<16)
574#define P4_CCCR_ESCR_SELECT(N) ((N)<<13)
575#define P4_CCCR_ENABLE (1<<12)
576#define P4_CCCR_OVF (1<<31)
577/* Set up IQ_COUNTER0 to behave like a clock, by having IQ_CCCR0 filter
578 CRU_ESCR0 (with any non-null event selector) through a complemented
579 max threshold. [IA32-Vol3, Section 14.9.9] */
580
417static int setup_p4_watchdog(void) 581static int setup_p4_watchdog(void)
418{ 582{
583 unsigned int perfctr_msr, evntsel_msr, cccr_msr;
584 unsigned int evntsel, cccr_val;
419 unsigned int misc_enable, dummy; 585 unsigned int misc_enable, dummy;
586 unsigned int ht_num;
587 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
420 588
421 rdmsr(MSR_P4_MISC_ENABLE, misc_enable, dummy); 589 rdmsr(MSR_IA32_MISC_ENABLE, misc_enable, dummy);
422 if (!(misc_enable & MSR_P4_MISC_ENABLE_PERF_AVAIL)) 590 if (!(misc_enable & MSR_P4_MISC_ENABLE_PERF_AVAIL))
423 return 0; 591 return 0;
424 592
425 nmi_perfctr_msr = MSR_P4_IQ_COUNTER0;
426 nmi_p4_cccr_val = P4_NMI_IQ_CCCR0;
427#ifdef CONFIG_SMP 593#ifdef CONFIG_SMP
428 if (smp_num_siblings == 2) 594 /* detect which hyperthread we are on */
429 nmi_p4_cccr_val |= P4_CCCR_OVF_PMI1; 595 if (smp_num_siblings == 2) {
596 unsigned int ebx, apicid;
597
598 ebx = cpuid_ebx(1);
599 apicid = (ebx >> 24) & 0xff;
600 ht_num = apicid & 1;
601 } else
430#endif 602#endif
603 ht_num = 0;
431 604
432 if (!(misc_enable & MSR_P4_MISC_ENABLE_PEBS_UNAVAIL)) 605 /* performance counters are shared resources
433 clear_msr_range(0x3F1, 2); 606 * assign each hyperthread its own set
434 /* MSR 0x3F0 seems to have a default value of 0xFC00, but current 607 * (re-use the ESCR0 register, seems safe
435 docs doesn't fully define it, so leave it alone for now. */ 608 * and keeps the cccr_val the same)
436 if (boot_cpu_data.x86_model >= 0x3) { 609 */
437 /* MSR_P4_IQ_ESCR0/1 (0x3ba/0x3bb) removed */ 610 if (!ht_num) {
438 clear_msr_range(0x3A0, 26); 611 /* logical cpu 0 */
439 clear_msr_range(0x3BC, 3); 612 perfctr_msr = MSR_P4_IQ_PERFCTR0;
613 evntsel_msr = MSR_P4_CRU_ESCR0;
614 cccr_msr = MSR_P4_IQ_CCCR0;
615 cccr_val = P4_CCCR_OVF_PMI0 | P4_CCCR_ESCR_SELECT(4);
440 } else { 616 } else {
441 clear_msr_range(0x3A0, 31); 617 /* logical cpu 1 */
618 perfctr_msr = MSR_P4_IQ_PERFCTR1;
619 evntsel_msr = MSR_P4_CRU_ESCR0;
620 cccr_msr = MSR_P4_IQ_CCCR1;
621 cccr_val = P4_CCCR_OVF_PMI1 | P4_CCCR_ESCR_SELECT(4);
442 } 622 }
443 clear_msr_range(0x3C0, 6); 623
444 clear_msr_range(0x3C8, 6); 624 if (!reserve_perfctr_nmi(perfctr_msr))
445 clear_msr_range(0x3E0, 2); 625 goto fail;
446 clear_msr_range(MSR_P4_CCCR0, 18); 626
447 clear_msr_range(MSR_P4_PERFCTR0, 18); 627 if (!reserve_evntsel_nmi(evntsel_msr))
448 628 goto fail1;
449 wrmsr(MSR_P4_CRU_ESCR0, P4_NMI_CRU_ESCR0, 0); 629
450 wrmsr(MSR_P4_IQ_CCCR0, P4_NMI_IQ_CCCR0 & ~P4_CCCR_ENABLE, 0); 630 evntsel = P4_ESCR_EVENT_SELECT(0x3F)
451 write_watchdog_counter("P4_IQ_COUNTER0"); 631 | P4_ESCR_OS
632 | P4_ESCR_USR;
633
634 cccr_val |= P4_CCCR_THRESHOLD(15)
635 | P4_CCCR_COMPLEMENT
636 | P4_CCCR_COMPARE
637 | P4_CCCR_REQUIRED;
638
639 wrmsr(evntsel_msr, evntsel, 0);
640 wrmsr(cccr_msr, cccr_val, 0);
641 write_watchdog_counter(perfctr_msr, "P4_IQ_COUNTER0");
452 apic_write(APIC_LVTPC, APIC_DM_NMI); 642 apic_write(APIC_LVTPC, APIC_DM_NMI);
453 wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0); 643 cccr_val |= P4_CCCR_ENABLE;
644 wrmsr(cccr_msr, cccr_val, 0);
645 wd->perfctr_msr = perfctr_msr;
646 wd->evntsel_msr = evntsel_msr;
647 wd->cccr_msr = cccr_msr;
648 wd->check_bit = 1ULL<<39;
454 return 1; 649 return 1;
650fail1:
651 release_perfctr_nmi(perfctr_msr);
652fail:
653 return 0;
455} 654}
456 655
457static void disable_intel_arch_watchdog(void) 656static void stop_p4_watchdog(void)
458{ 657{
459 unsigned ebx; 658 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
460 659
461 /* 660 wrmsr(wd->cccr_msr, 0, 0);
462 * Check whether the Architectural PerfMon supports 661 wrmsr(wd->evntsel_msr, 0, 0);
463 * Unhalted Core Cycles Event or not. 662
464 * NOTE: Corresponding bit = 0 in ebp indicates event present. 663 release_evntsel_nmi(wd->evntsel_msr);
465 */ 664 release_perfctr_nmi(wd->perfctr_msr);
466 ebx = cpuid_ebx(10);
467 if (!(ebx & ARCH_PERFMON_UNHALTED_CORE_CYCLES_PRESENT))
468 wrmsr(MSR_ARCH_PERFMON_EVENTSEL0, 0, 0);
469} 665}
470 666
667#define ARCH_PERFMON_NMI_EVENT_SEL ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL
668#define ARCH_PERFMON_NMI_EVENT_UMASK ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK
669
471static int setup_intel_arch_watchdog(void) 670static int setup_intel_arch_watchdog(void)
472{ 671{
672 unsigned int ebx;
673 union cpuid10_eax eax;
674 unsigned int unused;
675 unsigned int perfctr_msr, evntsel_msr;
473 unsigned int evntsel; 676 unsigned int evntsel;
474 unsigned ebx; 677 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
475 678
476 /* 679 /*
477 * Check whether the Architectural PerfMon supports 680 * Check whether the Architectural PerfMon supports
478 * Unhalted Core Cycles Event or not. 681 * Unhalted Core Cycles Event or not.
479 * NOTE: Corresponding bit = 0 in ebp indicates event present. 682 * NOTE: Corresponding bit = 0 in ebx indicates event present.
480 */ 683 */
481 ebx = cpuid_ebx(10); 684 cpuid(10, &(eax.full), &ebx, &unused, &unused);
482 if ((ebx & ARCH_PERFMON_UNHALTED_CORE_CYCLES_PRESENT)) 685 if ((eax.split.mask_length < (ARCH_PERFMON_UNHALTED_CORE_CYCLES_INDEX+1)) ||
483 return 0; 686 (ebx & ARCH_PERFMON_UNHALTED_CORE_CYCLES_PRESENT))
687 goto fail;
688
689 perfctr_msr = MSR_ARCH_PERFMON_PERFCTR0;
690 evntsel_msr = MSR_ARCH_PERFMON_EVENTSEL0;
484 691
485 nmi_perfctr_msr = MSR_ARCH_PERFMON_PERFCTR0; 692 if (!reserve_perfctr_nmi(perfctr_msr))
693 goto fail;
486 694
487 clear_msr_range(MSR_ARCH_PERFMON_EVENTSEL0, 2); 695 if (!reserve_evntsel_nmi(evntsel_msr))
488 clear_msr_range(MSR_ARCH_PERFMON_PERFCTR0, 2); 696 goto fail1;
697
698 wrmsrl(perfctr_msr, 0UL);
489 699
490 evntsel = ARCH_PERFMON_EVENTSEL_INT 700 evntsel = ARCH_PERFMON_EVENTSEL_INT
491 | ARCH_PERFMON_EVENTSEL_OS 701 | ARCH_PERFMON_EVENTSEL_OS
@@ -493,51 +703,145 @@ static int setup_intel_arch_watchdog(void)
493 | ARCH_PERFMON_NMI_EVENT_SEL 703 | ARCH_PERFMON_NMI_EVENT_SEL
494 | ARCH_PERFMON_NMI_EVENT_UMASK; 704 | ARCH_PERFMON_NMI_EVENT_UMASK;
495 705
496 wrmsr(MSR_ARCH_PERFMON_EVENTSEL0, evntsel, 0); 706 /* setup the timer */
497 write_watchdog_counter("INTEL_ARCH_PERFCTR0"); 707 wrmsr(evntsel_msr, evntsel, 0);
708 write_watchdog_counter(perfctr_msr, "INTEL_ARCH_PERFCTR0");
498 apic_write(APIC_LVTPC, APIC_DM_NMI); 709 apic_write(APIC_LVTPC, APIC_DM_NMI);
499 evntsel |= ARCH_PERFMON_EVENTSEL0_ENABLE; 710 evntsel |= ARCH_PERFMON_EVENTSEL0_ENABLE;
500 wrmsr(MSR_ARCH_PERFMON_EVENTSEL0, evntsel, 0); 711 wrmsr(evntsel_msr, evntsel, 0);
712
713 wd->perfctr_msr = perfctr_msr;
714 wd->evntsel_msr = evntsel_msr;
715 wd->cccr_msr = 0; //unused
716 wd->check_bit = 1ULL << (eax.split.bit_width - 1);
501 return 1; 717 return 1;
718fail1:
719 release_perfctr_nmi(perfctr_msr);
720fail:
721 return 0;
502} 722}
503 723
504void setup_apic_nmi_watchdog (void) 724static void stop_intel_arch_watchdog(void)
505{ 725{
506 switch (boot_cpu_data.x86_vendor) { 726 unsigned int ebx;
507 case X86_VENDOR_AMD: 727 union cpuid10_eax eax;
508 if (boot_cpu_data.x86 != 6 && boot_cpu_data.x86 != 15) 728 unsigned int unused;
509 return; 729 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
510 setup_k7_watchdog(); 730
511 break; 731 /*
512 case X86_VENDOR_INTEL: 732 * Check whether the Architectural PerfMon supports
513 if (cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) { 733 * Unhalted Core Cycles Event or not.
514 if (!setup_intel_arch_watchdog()) 734 * NOTE: Corresponding bit = 0 in ebx indicates event present.
735 */
736 cpuid(10, &(eax.full), &ebx, &unused, &unused);
737 if ((eax.split.mask_length < (ARCH_PERFMON_UNHALTED_CORE_CYCLES_INDEX+1)) ||
738 (ebx & ARCH_PERFMON_UNHALTED_CORE_CYCLES_PRESENT))
739 return;
740
741 wrmsr(wd->evntsel_msr, 0, 0);
742 release_evntsel_nmi(wd->evntsel_msr);
743 release_perfctr_nmi(wd->perfctr_msr);
744}
745
746void setup_apic_nmi_watchdog (void *unused)
747{
748 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
749
750 /* only support LOCAL and IO APICs for now */
751 if ((nmi_watchdog != NMI_LOCAL_APIC) &&
752 (nmi_watchdog != NMI_IO_APIC))
753 return;
754
755 if (wd->enabled == 1)
756 return;
757
758 /* cheap hack to support suspend/resume */
759 /* if cpu0 is not active neither should the other cpus */
760 if ((smp_processor_id() != 0) && (atomic_read(&nmi_active) <= 0))
761 return;
762
763 if (nmi_watchdog == NMI_LOCAL_APIC) {
764 switch (boot_cpu_data.x86_vendor) {
765 case X86_VENDOR_AMD:
766 if (boot_cpu_data.x86 != 6 && boot_cpu_data.x86 != 15)
515 return; 767 return;
516 break; 768 if (!setup_k7_watchdog())
517 }
518 switch (boot_cpu_data.x86) {
519 case 6:
520 if (boot_cpu_data.x86_model > 0xd)
521 return; 769 return;
522
523 setup_p6_watchdog();
524 break; 770 break;
525 case 15: 771 case X86_VENDOR_INTEL:
526 if (boot_cpu_data.x86_model > 0x4) 772 if (cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) {
527 return; 773 if (!setup_intel_arch_watchdog())
774 return;
775 break;
776 }
777 switch (boot_cpu_data.x86) {
778 case 6:
779 if (boot_cpu_data.x86_model > 0xd)
780 return;
781
782 if (!setup_p6_watchdog())
783 return;
784 break;
785 case 15:
786 if (boot_cpu_data.x86_model > 0x4)
787 return;
528 788
529 if (!setup_p4_watchdog()) 789 if (!setup_p4_watchdog())
790 return;
791 break;
792 default:
530 return; 793 return;
794 }
531 break; 795 break;
532 default: 796 default:
533 return; 797 return;
534 } 798 }
535 break; 799 }
536 default: 800 wd->enabled = 1;
801 atomic_inc(&nmi_active);
802}
803
804void stop_apic_nmi_watchdog(void *unused)
805{
806 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
807
808 /* only support LOCAL and IO APICs for now */
809 if ((nmi_watchdog != NMI_LOCAL_APIC) &&
810 (nmi_watchdog != NMI_IO_APIC))
811 return;
812
813 if (wd->enabled == 0)
537 return; 814 return;
815
816 if (nmi_watchdog == NMI_LOCAL_APIC) {
817 switch (boot_cpu_data.x86_vendor) {
818 case X86_VENDOR_AMD:
819 stop_k7_watchdog();
820 break;
821 case X86_VENDOR_INTEL:
822 if (cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) {
823 stop_intel_arch_watchdog();
824 break;
825 }
826 switch (boot_cpu_data.x86) {
827 case 6:
828 if (boot_cpu_data.x86_model > 0xd)
829 break;
830 stop_p6_watchdog();
831 break;
832 case 15:
833 if (boot_cpu_data.x86_model > 0x4)
834 break;
835 stop_p4_watchdog();
836 break;
837 }
838 break;
839 default:
840 return;
841 }
538 } 842 }
539 lapic_nmi_owner = LAPIC_NMI_WATCHDOG; 843 wd->enabled = 0;
540 nmi_active = 1; 844 atomic_dec(&nmi_active);
541} 845}
542 846
543/* 847/*
@@ -579,7 +883,7 @@ EXPORT_SYMBOL(touch_nmi_watchdog);
579 883
580extern void die_nmi(struct pt_regs *, const char *msg); 884extern void die_nmi(struct pt_regs *, const char *msg);
581 885
582void nmi_watchdog_tick (struct pt_regs * regs) 886__kprobes int nmi_watchdog_tick(struct pt_regs * regs, unsigned reason)
583{ 887{
584 888
585 /* 889 /*
@@ -588,11 +892,23 @@ void nmi_watchdog_tick (struct pt_regs * regs)
588 * smp_processor_id(). 892 * smp_processor_id().
589 */ 893 */
590 unsigned int sum; 894 unsigned int sum;
895 int touched = 0;
591 int cpu = smp_processor_id(); 896 int cpu = smp_processor_id();
897 struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
898 u64 dummy;
899 int rc=0;
900
901 /* check for other users first */
902 if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT)
903 == NOTIFY_STOP) {
904 rc = 1;
905 touched = 1;
906 }
592 907
593 sum = per_cpu(irq_stat, cpu).apic_timer_irqs; 908 sum = per_cpu(irq_stat, cpu).apic_timer_irqs;
594 909
595 if (last_irq_sums[cpu] == sum) { 910 /* if the apic timer isn't firing, this cpu isn't doing much */
911 if (!touched && last_irq_sums[cpu] == sum) {
596 /* 912 /*
597 * Ayiee, looks like this CPU is stuck ... 913 * Ayiee, looks like this CPU is stuck ...
598 * wait a few IRQs (5 seconds) before doing the oops ... 914 * wait a few IRQs (5 seconds) before doing the oops ...
@@ -607,27 +923,59 @@ void nmi_watchdog_tick (struct pt_regs * regs)
607 last_irq_sums[cpu] = sum; 923 last_irq_sums[cpu] = sum;
608 alert_counter[cpu] = 0; 924 alert_counter[cpu] = 0;
609 } 925 }
610 if (nmi_perfctr_msr) { 926 /* see if the nmi watchdog went off */
611 if (nmi_perfctr_msr == MSR_P4_IQ_COUNTER0) { 927 if (wd->enabled) {
612 /* 928 if (nmi_watchdog == NMI_LOCAL_APIC) {
613 * P4 quirks: 929 rdmsrl(wd->perfctr_msr, dummy);
614 * - An overflown perfctr will assert its interrupt 930 if (dummy & wd->check_bit){
615 * until the OVF flag in its CCCR is cleared. 931 /* this wasn't a watchdog timer interrupt */
616 * - LVTPC is masked on interrupt and must be 932 goto done;
617 * unmasked by the LVTPC handler. 933 }
934
935 /* only Intel P4 uses the cccr msr */
936 if (wd->cccr_msr != 0) {
937 /*
938 * P4 quirks:
939 * - An overflown perfctr will assert its interrupt
940 * until the OVF flag in its CCCR is cleared.
941 * - LVTPC is masked on interrupt and must be
942 * unmasked by the LVTPC handler.
943 */
944 rdmsrl(wd->cccr_msr, dummy);
945 dummy &= ~P4_CCCR_OVF;
946 wrmsrl(wd->cccr_msr, dummy);
947 apic_write(APIC_LVTPC, APIC_DM_NMI);
948 }
949 else if (wd->perfctr_msr == MSR_P6_PERFCTR0 ||
950 wd->perfctr_msr == MSR_ARCH_PERFMON_PERFCTR0) {
951 /* P6 based Pentium M need to re-unmask
952 * the apic vector but it doesn't hurt
953 * other P6 variant.
954 * ArchPerfom/Core Duo also needs this */
955 apic_write(APIC_LVTPC, APIC_DM_NMI);
956 }
957 /* start the cycle over again */
958 write_watchdog_counter(wd->perfctr_msr, NULL);
959 rc = 1;
960 } else if (nmi_watchdog == NMI_IO_APIC) {
961 /* don't know how to accurately check for this.
962 * just assume it was a watchdog timer interrupt
963 * This matches the old behaviour.
618 */ 964 */
619 wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0); 965 rc = 1;
620 apic_write(APIC_LVTPC, APIC_DM_NMI);
621 } 966 }
622 else if (nmi_perfctr_msr == MSR_P6_PERFCTR0 ||
623 nmi_perfctr_msr == MSR_ARCH_PERFMON_PERFCTR0) {
624 /* Only P6 based Pentium M need to re-unmask
625 * the apic vector but it doesn't hurt
626 * other P6 variant */
627 apic_write(APIC_LVTPC, APIC_DM_NMI);
628 }
629 write_watchdog_counter(NULL);
630 } 967 }
968done:
969 return rc;
970}
971
972int do_nmi_callback(struct pt_regs * regs, int cpu)
973{
974#ifdef CONFIG_SYSCTL
975 if (unknown_nmi_panic)
976 return unknown_nmi_panic_callback(regs, cpu);
977#endif
978 return 0;
631} 979}
632 980
633#ifdef CONFIG_SYSCTL 981#ifdef CONFIG_SYSCTL
@@ -637,36 +985,46 @@ static int unknown_nmi_panic_callback(struct pt_regs *regs, int cpu)
637 unsigned char reason = get_nmi_reason(); 985 unsigned char reason = get_nmi_reason();
638 char buf[64]; 986 char buf[64];
639 987
640 if (!(reason & 0xc0)) { 988 sprintf(buf, "NMI received for unknown reason %02x\n", reason);
641 sprintf(buf, "NMI received for unknown reason %02x\n", reason); 989 die_nmi(regs, buf);
642 die_nmi(regs, buf);
643 }
644 return 0; 990 return 0;
645} 991}
646 992
647/* 993/*
648 * proc handler for /proc/sys/kernel/unknown_nmi_panic 994 * proc handler for /proc/sys/kernel/nmi
649 */ 995 */
650int proc_unknown_nmi_panic(ctl_table *table, int write, struct file *file, 996int proc_nmi_enabled(struct ctl_table *table, int write, struct file *file,
651 void __user *buffer, size_t *length, loff_t *ppos) 997 void __user *buffer, size_t *length, loff_t *ppos)
652{ 998{
653 int old_state; 999 int old_state;
654 1000
655 old_state = unknown_nmi_panic; 1001 nmi_watchdog_enabled = (atomic_read(&nmi_active) > 0) ? 1 : 0;
1002 old_state = nmi_watchdog_enabled;
656 proc_dointvec(table, write, file, buffer, length, ppos); 1003 proc_dointvec(table, write, file, buffer, length, ppos);
657 if (!!old_state == !!unknown_nmi_panic) 1004 if (!!old_state == !!nmi_watchdog_enabled)
658 return 0; 1005 return 0;
659 1006
660 if (unknown_nmi_panic) { 1007 if (atomic_read(&nmi_active) < 0) {
661 if (reserve_lapic_nmi() < 0) { 1008 printk( KERN_WARNING "NMI watchdog is permanently disabled\n");
662 unknown_nmi_panic = 0; 1009 return -EIO;
663 return -EBUSY; 1010 }
664 } else { 1011
665 set_nmi_callback(unknown_nmi_panic_callback); 1012 if (nmi_watchdog == NMI_DEFAULT) {
666 } 1013 if (nmi_known_cpu() > 0)
1014 nmi_watchdog = NMI_LOCAL_APIC;
1015 else
1016 nmi_watchdog = NMI_IO_APIC;
1017 }
1018
1019 if (nmi_watchdog == NMI_LOCAL_APIC) {
1020 if (nmi_watchdog_enabled)
1021 enable_lapic_nmi_watchdog();
1022 else
1023 disable_lapic_nmi_watchdog();
667 } else { 1024 } else {
668 release_lapic_nmi(); 1025 printk( KERN_WARNING
669 unset_nmi_callback(); 1026 "NMI watchdog doesn't know what hardware to touch\n");
1027 return -EIO;
670 } 1028 }
671 return 0; 1029 return 0;
672} 1030}
@@ -675,7 +1033,11 @@ int proc_unknown_nmi_panic(ctl_table *table, int write, struct file *file,
675 1033
676EXPORT_SYMBOL(nmi_active); 1034EXPORT_SYMBOL(nmi_active);
677EXPORT_SYMBOL(nmi_watchdog); 1035EXPORT_SYMBOL(nmi_watchdog);
678EXPORT_SYMBOL(reserve_lapic_nmi); 1036EXPORT_SYMBOL(avail_to_resrv_perfctr_nmi);
679EXPORT_SYMBOL(release_lapic_nmi); 1037EXPORT_SYMBOL(avail_to_resrv_perfctr_nmi_bit);
1038EXPORT_SYMBOL(reserve_perfctr_nmi);
1039EXPORT_SYMBOL(release_perfctr_nmi);
1040EXPORT_SYMBOL(reserve_evntsel_nmi);
1041EXPORT_SYMBOL(release_evntsel_nmi);
680EXPORT_SYMBOL(disable_timer_nmi_watchdog); 1042EXPORT_SYMBOL(disable_timer_nmi_watchdog);
681EXPORT_SYMBOL(enable_timer_nmi_watchdog); 1043EXPORT_SYMBOL(enable_timer_nmi_watchdog);
diff --git a/arch/i386/kernel/process.c b/arch/i386/kernel/process.c
index 8657c739656a..8c190ca7ae44 100644
--- a/arch/i386/kernel/process.c
+++ b/arch/i386/kernel/process.c
@@ -37,6 +37,7 @@
37#include <linux/kallsyms.h> 37#include <linux/kallsyms.h>
38#include <linux/ptrace.h> 38#include <linux/ptrace.h>
39#include <linux/random.h> 39#include <linux/random.h>
40#include <linux/personality.h>
40 41
41#include <asm/uaccess.h> 42#include <asm/uaccess.h>
42#include <asm/pgtable.h> 43#include <asm/pgtable.h>
@@ -320,15 +321,6 @@ void show_regs(struct pt_regs * regs)
320 * the "args". 321 * the "args".
321 */ 322 */
322extern void kernel_thread_helper(void); 323extern void kernel_thread_helper(void);
323__asm__(".section .text\n"
324 ".align 4\n"
325 "kernel_thread_helper:\n\t"
326 "movl %edx,%eax\n\t"
327 "pushl %edx\n\t"
328 "call *%ebx\n\t"
329 "pushl %eax\n\t"
330 "call do_exit\n"
331 ".previous");
332 324
333/* 325/*
334 * Create a kernel thread 326 * Create a kernel thread
@@ -346,7 +338,7 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
346 regs.xes = __USER_DS; 338 regs.xes = __USER_DS;
347 regs.orig_eax = -1; 339 regs.orig_eax = -1;
348 regs.eip = (unsigned long) kernel_thread_helper; 340 regs.eip = (unsigned long) kernel_thread_helper;
349 regs.xcs = __KERNEL_CS; 341 regs.xcs = __KERNEL_CS | get_kernel_rpl();
350 regs.eflags = X86_EFLAGS_IF | X86_EFLAGS_SF | X86_EFLAGS_PF | 0x2; 342 regs.eflags = X86_EFLAGS_IF | X86_EFLAGS_SF | X86_EFLAGS_PF | 0x2;
351 343
352 /* Ok, create the new process.. */ 344 /* Ok, create the new process.. */
@@ -905,7 +897,7 @@ asmlinkage int sys_get_thread_area(struct user_desc __user *u_info)
905 897
906unsigned long arch_align_stack(unsigned long sp) 898unsigned long arch_align_stack(unsigned long sp)
907{ 899{
908 if (randomize_va_space) 900 if (!(current->personality & ADDR_NO_RANDOMIZE) && randomize_va_space)
909 sp -= get_random_int() % 8192; 901 sp -= get_random_int() % 8192;
910 return sp & ~0xf; 902 return sp & ~0xf;
911} 903}
diff --git a/arch/i386/kernel/ptrace.c b/arch/i386/kernel/ptrace.c
index d3db03f4085d..775f50e9395b 100644
--- a/arch/i386/kernel/ptrace.c
+++ b/arch/i386/kernel/ptrace.c
@@ -185,17 +185,17 @@ static unsigned long convert_eip_to_linear(struct task_struct *child, struct pt_
185 return addr; 185 return addr;
186} 186}
187 187
188static inline int is_at_popf(struct task_struct *child, struct pt_regs *regs) 188static inline int is_setting_trap_flag(struct task_struct *child, struct pt_regs *regs)
189{ 189{
190 int i, copied; 190 int i, copied;
191 unsigned char opcode[16]; 191 unsigned char opcode[15];
192 unsigned long addr = convert_eip_to_linear(child, regs); 192 unsigned long addr = convert_eip_to_linear(child, regs);
193 193
194 copied = access_process_vm(child, addr, opcode, sizeof(opcode), 0); 194 copied = access_process_vm(child, addr, opcode, sizeof(opcode), 0);
195 for (i = 0; i < copied; i++) { 195 for (i = 0; i < copied; i++) {
196 switch (opcode[i]) { 196 switch (opcode[i]) {
197 /* popf */ 197 /* popf and iret */
198 case 0x9d: 198 case 0x9d: case 0xcf:
199 return 1; 199 return 1;
200 /* opcode and address size prefixes */ 200 /* opcode and address size prefixes */
201 case 0x66: case 0x67: 201 case 0x66: case 0x67:
@@ -247,7 +247,7 @@ static void set_singlestep(struct task_struct *child)
247 * don't mark it as being "us" that set it, so that we 247 * don't mark it as being "us" that set it, so that we
248 * won't clear it by hand later. 248 * won't clear it by hand later.
249 */ 249 */
250 if (is_at_popf(child, regs)) 250 if (is_setting_trap_flag(child, regs))
251 return; 251 return;
252 252
253 child->ptrace |= PT_DTRACE; 253 child->ptrace |= PT_DTRACE;
diff --git a/arch/i386/kernel/reboot.c b/arch/i386/kernel/reboot.c
index 54cfeabbc5e4..84278e0093a2 100644
--- a/arch/i386/kernel/reboot.c
+++ b/arch/i386/kernel/reboot.c
@@ -145,14 +145,10 @@ real_mode_gdt_entries [3] =
145 0x000092000100ffffULL /* 16-bit real-mode 64k data at 0x00000100 */ 145 0x000092000100ffffULL /* 16-bit real-mode 64k data at 0x00000100 */
146}; 146};
147 147
148static struct 148static struct Xgt_desc_struct
149{ 149real_mode_gdt = { sizeof (real_mode_gdt_entries) - 1, (long)real_mode_gdt_entries },
150 unsigned short size __attribute__ ((packed)); 150real_mode_idt = { 0x3ff, 0 },
151 unsigned long long * base __attribute__ ((packed)); 151no_idt = { 0, 0 };
152}
153real_mode_gdt = { sizeof (real_mode_gdt_entries) - 1, real_mode_gdt_entries },
154real_mode_idt = { 0x3ff, NULL },
155no_idt = { 0, NULL };
156 152
157 153
158/* This is 16-bit protected mode code to disable paging and the cache, 154/* This is 16-bit protected mode code to disable paging and the cache,
diff --git a/arch/i386/kernel/relocate_kernel.S b/arch/i386/kernel/relocate_kernel.S
index d312616effa1..f151d6fae462 100644
--- a/arch/i386/kernel/relocate_kernel.S
+++ b/arch/i386/kernel/relocate_kernel.S
@@ -7,16 +7,138 @@
7 */ 7 */
8 8
9#include <linux/linkage.h> 9#include <linux/linkage.h>
10#include <asm/page.h>
11#include <asm/kexec.h>
12
13/*
14 * Must be relocatable PIC code callable as a C function
15 */
16
17#define PTR(x) (x << 2)
18#define PAGE_ALIGNED (1 << PAGE_SHIFT)
19#define PAGE_ATTR 0x63 /* _PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_DIRTY */
20#define PAE_PGD_ATTR 0x01 /* _PAGE_PRESENT */
21
22 .text
23 .align PAGE_ALIGNED
24 .globl relocate_kernel
25relocate_kernel:
26 movl 8(%esp), %ebp /* list of pages */
27
28#ifdef CONFIG_X86_PAE
29 /* map the control page at its virtual address */
30
31 movl PTR(VA_PGD)(%ebp), %edi
32 movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
33 andl $0xc0000000, %eax
34 shrl $27, %eax
35 addl %edi, %eax
36
37 movl PTR(PA_PMD_0)(%ebp), %edx
38 orl $PAE_PGD_ATTR, %edx
39 movl %edx, (%eax)
40
41 movl PTR(VA_PMD_0)(%ebp), %edi
42 movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
43 andl $0x3fe00000, %eax
44 shrl $18, %eax
45 addl %edi, %eax
46
47 movl PTR(PA_PTE_0)(%ebp), %edx
48 orl $PAGE_ATTR, %edx
49 movl %edx, (%eax)
50
51 movl PTR(VA_PTE_0)(%ebp), %edi
52 movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
53 andl $0x001ff000, %eax
54 shrl $9, %eax
55 addl %edi, %eax
56
57 movl PTR(PA_CONTROL_PAGE)(%ebp), %edx
58 orl $PAGE_ATTR, %edx
59 movl %edx, (%eax)
60
61 /* identity map the control page at its physical address */
62
63 movl PTR(VA_PGD)(%ebp), %edi
64 movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
65 andl $0xc0000000, %eax
66 shrl $27, %eax
67 addl %edi, %eax
68
69 movl PTR(PA_PMD_1)(%ebp), %edx
70 orl $PAE_PGD_ATTR, %edx
71 movl %edx, (%eax)
72
73 movl PTR(VA_PMD_1)(%ebp), %edi
74 movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
75 andl $0x3fe00000, %eax
76 shrl $18, %eax
77 addl %edi, %eax
78
79 movl PTR(PA_PTE_1)(%ebp), %edx
80 orl $PAGE_ATTR, %edx
81 movl %edx, (%eax)
82
83 movl PTR(VA_PTE_1)(%ebp), %edi
84 movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
85 andl $0x001ff000, %eax
86 shrl $9, %eax
87 addl %edi, %eax
88
89 movl PTR(PA_CONTROL_PAGE)(%ebp), %edx
90 orl $PAGE_ATTR, %edx
91 movl %edx, (%eax)
92#else
93 /* map the control page at its virtual address */
94
95 movl PTR(VA_PGD)(%ebp), %edi
96 movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
97 andl $0xffc00000, %eax
98 shrl $20, %eax
99 addl %edi, %eax
100
101 movl PTR(PA_PTE_0)(%ebp), %edx
102 orl $PAGE_ATTR, %edx
103 movl %edx, (%eax)
104
105 movl PTR(VA_PTE_0)(%ebp), %edi
106 movl PTR(VA_CONTROL_PAGE)(%ebp), %eax
107 andl $0x003ff000, %eax
108 shrl $10, %eax
109 addl %edi, %eax
110
111 movl PTR(PA_CONTROL_PAGE)(%ebp), %edx
112 orl $PAGE_ATTR, %edx
113 movl %edx, (%eax)
114
115 /* identity map the control page at its physical address */
116
117 movl PTR(VA_PGD)(%ebp), %edi
118 movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
119 andl $0xffc00000, %eax
120 shrl $20, %eax
121 addl %edi, %eax
122
123 movl PTR(PA_PTE_1)(%ebp), %edx
124 orl $PAGE_ATTR, %edx
125 movl %edx, (%eax)
126
127 movl PTR(VA_PTE_1)(%ebp), %edi
128 movl PTR(PA_CONTROL_PAGE)(%ebp), %eax
129 andl $0x003ff000, %eax
130 shrl $10, %eax
131 addl %edi, %eax
132
133 movl PTR(PA_CONTROL_PAGE)(%ebp), %edx
134 orl $PAGE_ATTR, %edx
135 movl %edx, (%eax)
136#endif
10 137
11 /*
12 * Must be relocatable PIC code callable as a C function, that once
13 * it starts can not use the previous processes stack.
14 */
15 .globl relocate_new_kernel
16relocate_new_kernel: 138relocate_new_kernel:
17 /* read the arguments and say goodbye to the stack */ 139 /* read the arguments and say goodbye to the stack */
18 movl 4(%esp), %ebx /* page_list */ 140 movl 4(%esp), %ebx /* page_list */
19 movl 8(%esp), %ebp /* reboot_code_buffer */ 141 movl 8(%esp), %ebp /* list of pages */
20 movl 12(%esp), %edx /* start address */ 142 movl 12(%esp), %edx /* start address */
21 movl 16(%esp), %ecx /* cpu_has_pae */ 143 movl 16(%esp), %ecx /* cpu_has_pae */
22 144
@@ -24,11 +146,26 @@ relocate_new_kernel:
24 pushl $0 146 pushl $0
25 popfl 147 popfl
26 148
27 /* set a new stack at the bottom of our page... */ 149 /* get physical address of control page now */
28 lea 4096(%ebp), %esp 150 /* this is impossible after page table switch */
151 movl PTR(PA_CONTROL_PAGE)(%ebp), %edi
29 152
30 /* store the parameters back on the stack */ 153 /* switch to new set of page tables */
31 pushl %edx /* store the start address */ 154 movl PTR(PA_PGD)(%ebp), %eax
155 movl %eax, %cr3
156
157 /* setup a new stack at the end of the physical control page */
158 lea 4096(%edi), %esp
159
160 /* jump to identity mapped page */
161 movl %edi, %eax
162 addl $(identity_mapped - relocate_kernel), %eax
163 pushl %eax
164 ret
165
166identity_mapped:
167 /* store the start address on the stack */
168 pushl %edx
32 169
33 /* Set cr0 to a known state: 170 /* Set cr0 to a known state:
34 * 31 0 == Paging disabled 171 * 31 0 == Paging disabled
@@ -113,8 +250,3 @@ relocate_new_kernel:
113 xorl %edi, %edi 250 xorl %edi, %edi
114 xorl %ebp, %ebp 251 xorl %ebp, %ebp
115 ret 252 ret
116relocate_new_kernel_end:
117
118 .globl relocate_new_kernel_size
119relocate_new_kernel_size:
120 .long relocate_new_kernel_end - relocate_new_kernel
diff --git a/arch/i386/kernel/semaphore.c b/arch/i386/kernel/semaphore.c
deleted file mode 100644
index 98352c374c76..000000000000
--- a/arch/i386/kernel/semaphore.c
+++ /dev/null
@@ -1,134 +0,0 @@
1/*
2 * i386 semaphore implementation.
3 *
4 * (C) Copyright 1999 Linus Torvalds
5 *
6 * Portions Copyright 1999 Red Hat, Inc.
7 *
8 * This program is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU General Public License
10 * as published by the Free Software Foundation; either version
11 * 2 of the License, or (at your option) any later version.
12 *
13 * rw semaphores implemented November 1999 by Benjamin LaHaise <bcrl@kvack.org>
14 */
15#include <asm/semaphore.h>
16
17/*
18 * The semaphore operations have a special calling sequence that
19 * allow us to do a simpler in-line version of them. These routines
20 * need to convert that sequence back into the C sequence when
21 * there is contention on the semaphore.
22 *
23 * %eax contains the semaphore pointer on entry. Save the C-clobbered
24 * registers (%eax, %edx and %ecx) except %eax whish is either a return
25 * value or just clobbered..
26 */
27asm(
28".section .sched.text\n"
29".align 4\n"
30".globl __down_failed\n"
31"__down_failed:\n\t"
32#if defined(CONFIG_FRAME_POINTER)
33 "pushl %ebp\n\t"
34 "movl %esp,%ebp\n\t"
35#endif
36 "pushl %edx\n\t"
37 "pushl %ecx\n\t"
38 "call __down\n\t"
39 "popl %ecx\n\t"
40 "popl %edx\n\t"
41#if defined(CONFIG_FRAME_POINTER)
42 "movl %ebp,%esp\n\t"
43 "popl %ebp\n\t"
44#endif
45 "ret"
46);
47
48asm(
49".section .sched.text\n"
50".align 4\n"
51".globl __down_failed_interruptible\n"
52"__down_failed_interruptible:\n\t"
53#if defined(CONFIG_FRAME_POINTER)
54 "pushl %ebp\n\t"
55 "movl %esp,%ebp\n\t"
56#endif
57 "pushl %edx\n\t"
58 "pushl %ecx\n\t"
59 "call __down_interruptible\n\t"
60 "popl %ecx\n\t"
61 "popl %edx\n\t"
62#if defined(CONFIG_FRAME_POINTER)
63 "movl %ebp,%esp\n\t"
64 "popl %ebp\n\t"
65#endif
66 "ret"
67);
68
69asm(
70".section .sched.text\n"
71".align 4\n"
72".globl __down_failed_trylock\n"
73"__down_failed_trylock:\n\t"
74#if defined(CONFIG_FRAME_POINTER)
75 "pushl %ebp\n\t"
76 "movl %esp,%ebp\n\t"
77#endif
78 "pushl %edx\n\t"
79 "pushl %ecx\n\t"
80 "call __down_trylock\n\t"
81 "popl %ecx\n\t"
82 "popl %edx\n\t"
83#if defined(CONFIG_FRAME_POINTER)
84 "movl %ebp,%esp\n\t"
85 "popl %ebp\n\t"
86#endif
87 "ret"
88);
89
90asm(
91".section .sched.text\n"
92".align 4\n"
93".globl __up_wakeup\n"
94"__up_wakeup:\n\t"
95 "pushl %edx\n\t"
96 "pushl %ecx\n\t"
97 "call __up\n\t"
98 "popl %ecx\n\t"
99 "popl %edx\n\t"
100 "ret"
101);
102
103/*
104 * rw spinlock fallbacks
105 */
106#if defined(CONFIG_SMP)
107asm(
108".section .sched.text\n"
109".align 4\n"
110".globl __write_lock_failed\n"
111"__write_lock_failed:\n\t"
112 LOCK_PREFIX "addl $" RW_LOCK_BIAS_STR ",(%eax)\n"
113"1: rep; nop\n\t"
114 "cmpl $" RW_LOCK_BIAS_STR ",(%eax)\n\t"
115 "jne 1b\n\t"
116 LOCK_PREFIX "subl $" RW_LOCK_BIAS_STR ",(%eax)\n\t"
117 "jnz __write_lock_failed\n\t"
118 "ret"
119);
120
121asm(
122".section .sched.text\n"
123".align 4\n"
124".globl __read_lock_failed\n"
125"__read_lock_failed:\n\t"
126 LOCK_PREFIX "incl (%eax)\n"
127"1: rep; nop\n\t"
128 "cmpl $1,(%eax)\n\t"
129 "js 1b\n\t"
130 LOCK_PREFIX "decl (%eax)\n\t"
131 "js __read_lock_failed\n\t"
132 "ret"
133);
134#endif
diff --git a/arch/i386/kernel/setup.c b/arch/i386/kernel/setup.c
index f1682206d304..814cdebf7377 100644
--- a/arch/i386/kernel/setup.c
+++ b/arch/i386/kernel/setup.c
@@ -53,6 +53,7 @@
53#include <asm/apic.h> 53#include <asm/apic.h>
54#include <asm/e820.h> 54#include <asm/e820.h>
55#include <asm/mpspec.h> 55#include <asm/mpspec.h>
56#include <asm/mmzone.h>
56#include <asm/setup.h> 57#include <asm/setup.h>
57#include <asm/arch_hooks.h> 58#include <asm/arch_hooks.h>
58#include <asm/sections.h> 59#include <asm/sections.h>
@@ -89,18 +90,6 @@ EXPORT_SYMBOL(boot_cpu_data);
89 90
90unsigned long mmu_cr4_features; 91unsigned long mmu_cr4_features;
91 92
92#ifdef CONFIG_ACPI
93 int acpi_disabled = 0;
94#else
95 int acpi_disabled = 1;
96#endif
97EXPORT_SYMBOL(acpi_disabled);
98
99#ifdef CONFIG_ACPI
100int __initdata acpi_force = 0;
101extern acpi_interrupt_flags acpi_sci_flags;
102#endif
103
104/* for MCA, but anyone else can use it if they want */ 93/* for MCA, but anyone else can use it if they want */
105unsigned int machine_id; 94unsigned int machine_id;
106#ifdef CONFIG_MCA 95#ifdef CONFIG_MCA
@@ -148,7 +137,6 @@ EXPORT_SYMBOL(ist_info);
148struct e820map e820; 137struct e820map e820;
149 138
150extern void early_cpu_init(void); 139extern void early_cpu_init(void);
151extern void generic_apic_probe(char *);
152extern int root_mountflags; 140extern int root_mountflags;
153 141
154unsigned long saved_videomode; 142unsigned long saved_videomode;
@@ -700,238 +688,150 @@ static inline void copy_edd(void)
700} 688}
701#endif 689#endif
702 690
703static void __init parse_cmdline_early (char ** cmdline_p) 691static int __initdata user_defined_memmap = 0;
704{
705 char c = ' ', *to = command_line, *from = saved_command_line;
706 int len = 0;
707 int userdef = 0;
708 692
709 /* Save unparsed command line copy for /proc/cmdline */ 693/*
710 saved_command_line[COMMAND_LINE_SIZE-1] = '\0'; 694 * "mem=nopentium" disables the 4MB page tables.
695 * "mem=XXX[kKmM]" defines a memory region from HIGH_MEM
696 * to <mem>, overriding the bios size.
697 * "memmap=XXX[KkmM]@XXX[KkmM]" defines a memory region from
698 * <start> to <start>+<mem>, overriding the bios size.
699 *
700 * HPA tells me bootloaders need to parse mem=, so no new
701 * option should be mem= [also see Documentation/i386/boot.txt]
702 */
703static int __init parse_mem(char *arg)
704{
705 if (!arg)
706 return -EINVAL;
711 707
712 for (;;) { 708 if (strcmp(arg, "nopentium") == 0) {
713 if (c != ' ') 709 clear_bit(X86_FEATURE_PSE, boot_cpu_data.x86_capability);
714 goto next_char; 710 disable_pse = 1;
715 /* 711 } else {
716 * "mem=nopentium" disables the 4MB page tables. 712 /* If the user specifies memory size, we
717 * "mem=XXX[kKmM]" defines a memory region from HIGH_MEM 713 * limit the BIOS-provided memory map to
718 * to <mem>, overriding the bios size. 714 * that size. exactmap can be used to specify
719 * "memmap=XXX[KkmM]@XXX[KkmM]" defines a memory region from 715 * the exact map. mem=number can be used to
720 * <start> to <start>+<mem>, overriding the bios size. 716 * trim the existing memory map.
721 *
722 * HPA tells me bootloaders need to parse mem=, so no new
723 * option should be mem= [also see Documentation/i386/boot.txt]
724 */ 717 */
725 if (!memcmp(from, "mem=", 4)) { 718 unsigned long long mem_size;
726 if (to != command_line)
727 to--;
728 if (!memcmp(from+4, "nopentium", 9)) {
729 from += 9+4;
730 clear_bit(X86_FEATURE_PSE, boot_cpu_data.x86_capability);
731 disable_pse = 1;
732 } else {
733 /* If the user specifies memory size, we
734 * limit the BIOS-provided memory map to
735 * that size. exactmap can be used to specify
736 * the exact map. mem=number can be used to
737 * trim the existing memory map.
738 */
739 unsigned long long mem_size;
740
741 mem_size = memparse(from+4, &from);
742 limit_regions(mem_size);
743 userdef=1;
744 }
745 }
746
747 else if (!memcmp(from, "memmap=", 7)) {
748 if (to != command_line)
749 to--;
750 if (!memcmp(from+7, "exactmap", 8)) {
751#ifdef CONFIG_CRASH_DUMP
752 /* If we are doing a crash dump, we
753 * still need to know the real mem
754 * size before original memory map is
755 * reset.
756 */
757 find_max_pfn();
758 saved_max_pfn = max_pfn;
759#endif
760 from += 8+7;
761 e820.nr_map = 0;
762 userdef = 1;
763 } else {
764 /* If the user specifies memory size, we
765 * limit the BIOS-provided memory map to
766 * that size. exactmap can be used to specify
767 * the exact map. mem=number can be used to
768 * trim the existing memory map.
769 */
770 unsigned long long start_at, mem_size;
771 719
772 mem_size = memparse(from+7, &from); 720 mem_size = memparse(arg, &arg);
773 if (*from == '@') { 721 limit_regions(mem_size);
774 start_at = memparse(from+1, &from); 722 user_defined_memmap = 1;
775 add_memory_region(start_at, mem_size, E820_RAM); 723 }
776 } else if (*from == '#') { 724 return 0;
777 start_at = memparse(from+1, &from); 725}
778 add_memory_region(start_at, mem_size, E820_ACPI); 726early_param("mem", parse_mem);
779 } else if (*from == '$') {
780 start_at = memparse(from+1, &from);
781 add_memory_region(start_at, mem_size, E820_RESERVED);
782 } else {
783 limit_regions(mem_size);
784 userdef=1;
785 }
786 }
787 }
788
789 else if (!memcmp(from, "noexec=", 7))
790 noexec_setup(from + 7);
791 727
728static int __init parse_memmap(char *arg)
729{
730 if (!arg)
731 return -EINVAL;
792 732
793#ifdef CONFIG_X86_SMP 733 if (strcmp(arg, "exactmap") == 0) {
794 /* 734#ifdef CONFIG_CRASH_DUMP
795 * If the BIOS enumerates physical processors before logical, 735 /* If we are doing a crash dump, we
796 * maxcpus=N at enumeration-time can be used to disable HT. 736 * still need to know the real mem
737 * size before original memory map is
738 * reset.
797 */ 739 */
798 else if (!memcmp(from, "maxcpus=", 8)) { 740 find_max_pfn();
799 extern unsigned int maxcpus; 741 saved_max_pfn = max_pfn;
800
801 maxcpus = simple_strtoul(from + 8, NULL, 0);
802 }
803#endif 742#endif
804 743 e820.nr_map = 0;
805#ifdef CONFIG_ACPI 744 user_defined_memmap = 1;
806 /* "acpi=off" disables both ACPI table parsing and interpreter */ 745 } else {
807 else if (!memcmp(from, "acpi=off", 8)) { 746 /* If the user specifies memory size, we
808 disable_acpi(); 747 * limit the BIOS-provided memory map to
809 } 748 * that size. exactmap can be used to specify
810 749 * the exact map. mem=number can be used to
811 /* acpi=force to over-ride black-list */ 750 * trim the existing memory map.
812 else if (!memcmp(from, "acpi=force", 10)) { 751 */
813 acpi_force = 1; 752 unsigned long long start_at, mem_size;
814 acpi_ht = 1; 753
815 acpi_disabled = 0; 754 mem_size = memparse(arg, &arg);
816 } 755 if (*arg == '@') {
817 756 start_at = memparse(arg+1, &arg);
818 /* acpi=strict disables out-of-spec workarounds */ 757 add_memory_region(start_at, mem_size, E820_RAM);
819 else if (!memcmp(from, "acpi=strict", 11)) { 758 } else if (*arg == '#') {
820 acpi_strict = 1; 759 start_at = memparse(arg+1, &arg);
821 } 760 add_memory_region(start_at, mem_size, E820_ACPI);
822 761 } else if (*arg == '$') {
823 /* Limit ACPI just to boot-time to enable HT */ 762 start_at = memparse(arg+1, &arg);
824 else if (!memcmp(from, "acpi=ht", 7)) { 763 add_memory_region(start_at, mem_size, E820_RESERVED);
825 if (!acpi_force) 764 } else {
826 disable_acpi(); 765 limit_regions(mem_size);
827 acpi_ht = 1; 766 user_defined_memmap = 1;
828 }
829
830 /* "pci=noacpi" disable ACPI IRQ routing and PCI scan */
831 else if (!memcmp(from, "pci=noacpi", 10)) {
832 acpi_disable_pci();
833 }
834 /* "acpi=noirq" disables ACPI interrupt routing */
835 else if (!memcmp(from, "acpi=noirq", 10)) {
836 acpi_noirq_set();
837 } 767 }
768 }
769 return 0;
770}
771early_param("memmap", parse_memmap);
838 772
839 else if (!memcmp(from, "acpi_sci=edge", 13)) 773#ifdef CONFIG_PROC_VMCORE
840 acpi_sci_flags.trigger = 1; 774/* elfcorehdr= specifies the location of elf core header
841 775 * stored by the crashed kernel.
842 else if (!memcmp(from, "acpi_sci=level", 14)) 776 */
843 acpi_sci_flags.trigger = 3; 777static int __init parse_elfcorehdr(char *arg)
844 778{
845 else if (!memcmp(from, "acpi_sci=high", 13)) 779 if (!arg)
846 acpi_sci_flags.polarity = 1; 780 return -EINVAL;
847 781
848 else if (!memcmp(from, "acpi_sci=low", 12)) 782 elfcorehdr_addr = memparse(arg, &arg);
849 acpi_sci_flags.polarity = 3; 783 return 0;
784}
785early_param("elfcorehdr", parse_elfcorehdr);
786#endif /* CONFIG_PROC_VMCORE */
850 787
851#ifdef CONFIG_X86_IO_APIC 788/*
852 else if (!memcmp(from, "acpi_skip_timer_override", 24)) 789 * highmem=size forces highmem to be exactly 'size' bytes.
853 acpi_skip_timer_override = 1; 790 * This works even on boxes that have no highmem otherwise.
791 * This also works to reduce highmem size on bigger boxes.
792 */
793static int __init parse_highmem(char *arg)
794{
795 if (!arg)
796 return -EINVAL;
854 797
855 if (!memcmp(from, "disable_timer_pin_1", 19)) 798 highmem_pages = memparse(arg, &arg) >> PAGE_SHIFT;
856 disable_timer_pin_1 = 1; 799 return 0;
857 if (!memcmp(from, "enable_timer_pin_1", 18)) 800}
858 disable_timer_pin_1 = -1; 801early_param("highmem", parse_highmem);
859 802
860 /* disable IO-APIC */ 803/*
861 else if (!memcmp(from, "noapic", 6)) 804 * vmalloc=size forces the vmalloc area to be exactly 'size'
862 disable_ioapic_setup(); 805 * bytes. This can be used to increase (or decrease) the
863#endif /* CONFIG_X86_IO_APIC */ 806 * vmalloc area - the default is 128m.
864#endif /* CONFIG_ACPI */ 807 */
808static int __init parse_vmalloc(char *arg)
809{
810 if (!arg)
811 return -EINVAL;
865 812
866#ifdef CONFIG_X86_LOCAL_APIC 813 __VMALLOC_RESERVE = memparse(arg, &arg);
867 /* enable local APIC */ 814 return 0;
868 else if (!memcmp(from, "lapic", 5)) 815}
869 lapic_enable(); 816early_param("vmalloc", parse_vmalloc);
870 817
871 /* disable local APIC */ 818/*
872 else if (!memcmp(from, "nolapic", 6)) 819 * reservetop=size reserves a hole at the top of the kernel address space which
873 lapic_disable(); 820 * a hypervisor can load into later. Needed for dynamically loaded hypervisors,
874#endif /* CONFIG_X86_LOCAL_APIC */ 821 * so relocating the fixmap can be done before paging initialization.
822 */
823static int __init parse_reservetop(char *arg)
824{
825 unsigned long address;
875 826
876#ifdef CONFIG_KEXEC 827 if (!arg)
877 /* crashkernel=size@addr specifies the location to reserve for 828 return -EINVAL;
878 * a crash kernel. By reserving this memory we guarantee
879 * that linux never set's it up as a DMA target.
880 * Useful for holding code to do something appropriate
881 * after a kernel panic.
882 */
883 else if (!memcmp(from, "crashkernel=", 12)) {
884 unsigned long size, base;
885 size = memparse(from+12, &from);
886 if (*from == '@') {
887 base = memparse(from+1, &from);
888 /* FIXME: Do I want a sanity check
889 * to validate the memory range?
890 */
891 crashk_res.start = base;
892 crashk_res.end = base + size - 1;
893 }
894 }
895#endif
896#ifdef CONFIG_PROC_VMCORE
897 /* elfcorehdr= specifies the location of elf core header
898 * stored by the crashed kernel.
899 */
900 else if (!memcmp(from, "elfcorehdr=", 11))
901 elfcorehdr_addr = memparse(from+11, &from);
902#endif
903 829
904 /* 830 address = memparse(arg, &arg);
905 * highmem=size forces highmem to be exactly 'size' bytes. 831 reserve_top_address(address);
906 * This works even on boxes that have no highmem otherwise. 832 return 0;
907 * This also works to reduce highmem size on bigger boxes.
908 */
909 else if (!memcmp(from, "highmem=", 8))
910 highmem_pages = memparse(from+8, &from) >> PAGE_SHIFT;
911
912 /*
913 * vmalloc=size forces the vmalloc area to be exactly 'size'
914 * bytes. This can be used to increase (or decrease) the
915 * vmalloc area - the default is 128m.
916 */
917 else if (!memcmp(from, "vmalloc=", 8))
918 __VMALLOC_RESERVE = memparse(from+8, &from);
919
920 next_char:
921 c = *(from++);
922 if (!c)
923 break;
924 if (COMMAND_LINE_SIZE <= ++len)
925 break;
926 *(to++) = c;
927 }
928 *to = '\0';
929 *cmdline_p = command_line;
930 if (userdef) {
931 printk(KERN_INFO "user-defined physical RAM map:\n");
932 print_memory_map("user");
933 }
934} 833}
834early_param("reservetop", parse_reservetop);
935 835
936/* 836/*
937 * Callback for efi_memory_walk. 837 * Callback for efi_memory_walk.
@@ -1170,6 +1070,14 @@ static unsigned long __init setup_memory(void)
1170 } 1070 }
1171 printk(KERN_NOTICE "%ldMB HIGHMEM available.\n", 1071 printk(KERN_NOTICE "%ldMB HIGHMEM available.\n",
1172 pages_to_mb(highend_pfn - highstart_pfn)); 1072 pages_to_mb(highend_pfn - highstart_pfn));
1073 num_physpages = highend_pfn;
1074 high_memory = (void *) __va(highstart_pfn * PAGE_SIZE - 1) + 1;
1075#else
1076 num_physpages = max_low_pfn;
1077 high_memory = (void *) __va(max_low_pfn * PAGE_SIZE - 1) + 1;
1078#endif
1079#ifdef CONFIG_FLATMEM
1080 max_mapnr = num_physpages;
1173#endif 1081#endif
1174 printk(KERN_NOTICE "%ldMB LOWMEM available.\n", 1082 printk(KERN_NOTICE "%ldMB LOWMEM available.\n",
1175 pages_to_mb(max_low_pfn)); 1083 pages_to_mb(max_low_pfn));
@@ -1181,22 +1089,20 @@ static unsigned long __init setup_memory(void)
1181 1089
1182void __init zone_sizes_init(void) 1090void __init zone_sizes_init(void)
1183{ 1091{
1184 unsigned long zones_size[MAX_NR_ZONES] = {0, 0, 0};
1185 unsigned int max_dma, low;
1186
1187 max_dma = virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT;
1188 low = max_low_pfn;
1189
1190 if (low < max_dma)
1191 zones_size[ZONE_DMA] = low;
1192 else {
1193 zones_size[ZONE_DMA] = max_dma;
1194 zones_size[ZONE_NORMAL] = low - max_dma;
1195#ifdef CONFIG_HIGHMEM 1092#ifdef CONFIG_HIGHMEM
1196 zones_size[ZONE_HIGHMEM] = highend_pfn - low; 1093 unsigned long max_zone_pfns[MAX_NR_ZONES] = {
1094 virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT,
1095 max_low_pfn,
1096 highend_pfn};
1097 add_active_range(0, 0, highend_pfn);
1098#else
1099 unsigned long max_zone_pfns[MAX_NR_ZONES] = {
1100 virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT,
1101 max_low_pfn};
1102 add_active_range(0, 0, max_low_pfn);
1197#endif 1103#endif
1198 } 1104
1199 free_area_init(zones_size); 1105 free_area_init_nodes(max_zone_pfns);
1200} 1106}
1201#else 1107#else
1202extern unsigned long __init setup_memory(void); 1108extern unsigned long __init setup_memory(void);
@@ -1258,7 +1164,7 @@ void __init setup_bootmem_allocator(void)
1258 */ 1164 */
1259 find_smp_config(); 1165 find_smp_config();
1260#endif 1166#endif
1261 1167 numa_kva_reserve();
1262#ifdef CONFIG_BLK_DEV_INITRD 1168#ifdef CONFIG_BLK_DEV_INITRD
1263 if (LOADER_TYPE && INITRD_START) { 1169 if (LOADER_TYPE && INITRD_START) {
1264 if (INITRD_START + INITRD_SIZE <= (max_low_pfn << PAGE_SHIFT)) { 1170 if (INITRD_START + INITRD_SIZE <= (max_low_pfn << PAGE_SHIFT)) {
@@ -1499,17 +1405,15 @@ void __init setup_arch(char **cmdline_p)
1499 data_resource.start = virt_to_phys(_etext); 1405 data_resource.start = virt_to_phys(_etext);
1500 data_resource.end = virt_to_phys(_edata)-1; 1406 data_resource.end = virt_to_phys(_edata)-1;
1501 1407
1502 parse_cmdline_early(cmdline_p); 1408 parse_early_param();
1503 1409
1504#ifdef CONFIG_EARLY_PRINTK 1410 if (user_defined_memmap) {
1505 { 1411 printk(KERN_INFO "user-defined physical RAM map:\n");
1506 char *s = strstr(*cmdline_p, "earlyprintk="); 1412 print_memory_map("user");
1507 if (s) {
1508 setup_early_printk(strchr(s, '=') + 1);
1509 printk("early console enabled\n");
1510 }
1511 } 1413 }
1512#endif 1414
1415 strlcpy(command_line, saved_command_line, COMMAND_LINE_SIZE);
1416 *cmdline_p = command_line;
1513 1417
1514 max_low_pfn = setup_memory(); 1418 max_low_pfn = setup_memory();
1515 1419
@@ -1538,7 +1442,7 @@ void __init setup_arch(char **cmdline_p)
1538 dmi_scan_machine(); 1442 dmi_scan_machine();
1539 1443
1540#ifdef CONFIG_X86_GENERICARCH 1444#ifdef CONFIG_X86_GENERICARCH
1541 generic_apic_probe(*cmdline_p); 1445 generic_apic_probe();
1542#endif 1446#endif
1543 if (efi_enabled) 1447 if (efi_enabled)
1544 efi_map_memmap(); 1448 efi_map_memmap();
@@ -1550,9 +1454,11 @@ void __init setup_arch(char **cmdline_p)
1550 acpi_boot_table_init(); 1454 acpi_boot_table_init();
1551#endif 1455#endif
1552 1456
1457#ifdef CONFIG_PCI
1553#ifdef CONFIG_X86_IO_APIC 1458#ifdef CONFIG_X86_IO_APIC
1554 check_acpi_pci(); /* Checks more than just ACPI actually */ 1459 check_acpi_pci(); /* Checks more than just ACPI actually */
1555#endif 1460#endif
1461#endif
1556 1462
1557#ifdef CONFIG_ACPI 1463#ifdef CONFIG_ACPI
1558 acpi_boot_init(); 1464 acpi_boot_init();
diff --git a/arch/i386/kernel/smp.c b/arch/i386/kernel/smp.c
index c10789d7a9d3..465188e2d701 100644
--- a/arch/i386/kernel/smp.c
+++ b/arch/i386/kernel/smp.c
@@ -634,3 +634,69 @@ fastcall void smp_call_function_interrupt(struct pt_regs *regs)
634 } 634 }
635} 635}
636 636
637/*
638 * this function sends a 'generic call function' IPI to one other CPU
639 * in the system.
640 *
641 * cpu is a standard Linux logical CPU number.
642 */
643static void
644__smp_call_function_single(int cpu, void (*func) (void *info), void *info,
645 int nonatomic, int wait)
646{
647 struct call_data_struct data;
648 int cpus = 1;
649
650 data.func = func;
651 data.info = info;
652 atomic_set(&data.started, 0);
653 data.wait = wait;
654 if (wait)
655 atomic_set(&data.finished, 0);
656
657 call_data = &data;
658 wmb();
659 /* Send a message to all other CPUs and wait for them to respond */
660 send_IPI_mask(cpumask_of_cpu(cpu), CALL_FUNCTION_VECTOR);
661
662 /* Wait for response */
663 while (atomic_read(&data.started) != cpus)
664 cpu_relax();
665
666 if (!wait)
667 return;
668
669 while (atomic_read(&data.finished) != cpus)
670 cpu_relax();
671}
672
673/*
674 * smp_call_function_single - Run a function on another CPU
675 * @func: The function to run. This must be fast and non-blocking.
676 * @info: An arbitrary pointer to pass to the function.
677 * @nonatomic: Currently unused.
678 * @wait: If true, wait until function has completed on other CPUs.
679 *
680 * Retrurns 0 on success, else a negative status code.
681 *
682 * Does not return until the remote CPU is nearly ready to execute <func>
683 * or is or has executed.
684 */
685
686int smp_call_function_single(int cpu, void (*func) (void *info), void *info,
687 int nonatomic, int wait)
688{
689 /* prevent preemption and reschedule on another processor */
690 int me = get_cpu();
691 if (cpu == me) {
692 WARN_ON(1);
693 put_cpu();
694 return -EBUSY;
695 }
696 spin_lock_bh(&call_lock);
697 __smp_call_function_single(cpu, func, info, nonatomic, wait);
698 spin_unlock_bh(&call_lock);
699 put_cpu();
700 return 0;
701}
702EXPORT_SYMBOL(smp_call_function_single);
diff --git a/arch/i386/kernel/smpboot.c b/arch/i386/kernel/smpboot.c
index f948419c888a..020d873b7d21 100644
--- a/arch/i386/kernel/smpboot.c
+++ b/arch/i386/kernel/smpboot.c
@@ -177,6 +177,9 @@ static void __devinit smp_store_cpu_info(int id)
177 */ 177 */
178 if ((c->x86_vendor == X86_VENDOR_AMD) && (c->x86 == 6)) { 178 if ((c->x86_vendor == X86_VENDOR_AMD) && (c->x86 == 6)) {
179 179
180 if (num_possible_cpus() == 1)
181 goto valid_k7;
182
180 /* Athlon 660/661 is valid. */ 183 /* Athlon 660/661 is valid. */
181 if ((c->x86_model==6) && ((c->x86_mask==0) || (c->x86_mask==1))) 184 if ((c->x86_model==6) && ((c->x86_mask==0) || (c->x86_mask==1)))
182 goto valid_k7; 185 goto valid_k7;
@@ -642,9 +645,13 @@ static void map_cpu_to_logical_apicid(void)
642{ 645{
643 int cpu = smp_processor_id(); 646 int cpu = smp_processor_id();
644 int apicid = logical_smp_processor_id(); 647 int apicid = logical_smp_processor_id();
648 int node = apicid_to_node(apicid);
649
650 if (!node_online(node))
651 node = first_online_node;
645 652
646 cpu_2_logical_apicid[cpu] = apicid; 653 cpu_2_logical_apicid[cpu] = apicid;
647 map_cpu_to_node(cpu, apicid_to_node(apicid)); 654 map_cpu_to_node(cpu, node);
648} 655}
649 656
650static void unmap_cpu_to_logical_apicid(int cpu) 657static void unmap_cpu_to_logical_apicid(int cpu)
@@ -1372,7 +1379,8 @@ int __cpu_disable(void)
1372 */ 1379 */
1373 if (cpu == 0) 1380 if (cpu == 0)
1374 return -EBUSY; 1381 return -EBUSY;
1375 1382 if (nmi_watchdog == NMI_LOCAL_APIC)
1383 stop_apic_nmi_watchdog(NULL);
1376 clear_local_APIC(); 1384 clear_local_APIC();
1377 /* Allow any queued timer interrupts to get serviced */ 1385 /* Allow any queued timer interrupts to get serviced */
1378 local_irq_enable(); 1386 local_irq_enable();
@@ -1486,3 +1494,16 @@ void __init smp_intr_init(void)
1486 /* IPI for generic function call */ 1494 /* IPI for generic function call */
1487 set_intr_gate(CALL_FUNCTION_VECTOR, call_function_interrupt); 1495 set_intr_gate(CALL_FUNCTION_VECTOR, call_function_interrupt);
1488} 1496}
1497
1498/*
1499 * If the BIOS enumerates physical processors before logical,
1500 * maxcpus=N at enumeration-time can be used to disable HT.
1501 */
1502static int __init parse_maxcpus(char *arg)
1503{
1504 extern unsigned int maxcpus;
1505
1506 maxcpus = simple_strtoul(arg, NULL, 0);
1507 return 0;
1508}
1509early_param("maxcpus", parse_maxcpus);
diff --git a/arch/i386/kernel/srat.c b/arch/i386/kernel/srat.c
index b1809c9a0899..32413122c4c2 100644
--- a/arch/i386/kernel/srat.c
+++ b/arch/i386/kernel/srat.c
@@ -42,7 +42,7 @@
42#define PXM_BITMAP_LEN (MAX_PXM_DOMAINS / 8) 42#define PXM_BITMAP_LEN (MAX_PXM_DOMAINS / 8)
43static u8 pxm_bitmap[PXM_BITMAP_LEN]; /* bitmap of proximity domains */ 43static u8 pxm_bitmap[PXM_BITMAP_LEN]; /* bitmap of proximity domains */
44 44
45#define MAX_CHUNKS_PER_NODE 4 45#define MAX_CHUNKS_PER_NODE 3
46#define MAXCHUNKS (MAX_CHUNKS_PER_NODE * MAX_NUMNODES) 46#define MAXCHUNKS (MAX_CHUNKS_PER_NODE * MAX_NUMNODES)
47struct node_memory_chunk_s { 47struct node_memory_chunk_s {
48 unsigned long start_pfn; 48 unsigned long start_pfn;
@@ -54,8 +54,6 @@ struct node_memory_chunk_s {
54static struct node_memory_chunk_s node_memory_chunk[MAXCHUNKS]; 54static struct node_memory_chunk_s node_memory_chunk[MAXCHUNKS];
55 55
56static int num_memory_chunks; /* total number of memory chunks */ 56static int num_memory_chunks; /* total number of memory chunks */
57static int zholes_size_init;
58static unsigned long zholes_size[MAX_NUMNODES * MAX_NR_ZONES];
59 57
60extern void * boot_ioremap(unsigned long, unsigned long); 58extern void * boot_ioremap(unsigned long, unsigned long);
61 59
@@ -135,50 +133,6 @@ static void __init parse_memory_affinity_structure (char *sratp)
135 "enabled and removable" : "enabled" ) ); 133 "enabled and removable" : "enabled" ) );
136} 134}
137 135
138#if MAX_NR_ZONES != 4
139#error "MAX_NR_ZONES != 4, chunk_to_zone requires review"
140#endif
141/* Take a chunk of pages from page frame cstart to cend and count the number
142 * of pages in each zone, returned via zones[].
143 */
144static __init void chunk_to_zones(unsigned long cstart, unsigned long cend,
145 unsigned long *zones)
146{
147 unsigned long max_dma;
148 extern unsigned long max_low_pfn;
149
150 int z;
151 unsigned long rend;
152
153 /* FIXME: MAX_DMA_ADDRESS and max_low_pfn are trying to provide
154 * similarly scoped information and should be handled in a consistant
155 * manner.
156 */
157 max_dma = virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT;
158
159 /* Split the hole into the zones in which it falls. Repeatedly
160 * take the segment in which the remaining hole starts, round it
161 * to the end of that zone.
162 */
163 memset(zones, 0, MAX_NR_ZONES * sizeof(long));
164 while (cstart < cend) {
165 if (cstart < max_dma) {
166 z = ZONE_DMA;
167 rend = (cend < max_dma)? cend : max_dma;
168
169 } else if (cstart < max_low_pfn) {
170 z = ZONE_NORMAL;
171 rend = (cend < max_low_pfn)? cend : max_low_pfn;
172
173 } else {
174 z = ZONE_HIGHMEM;
175 rend = cend;
176 }
177 zones[z] += rend - cstart;
178 cstart = rend;
179 }
180}
181
182/* 136/*
183 * The SRAT table always lists ascending addresses, so can always 137 * The SRAT table always lists ascending addresses, so can always
184 * assume that the first "start" address that you see is the real 138 * assume that the first "start" address that you see is the real
@@ -223,7 +177,6 @@ static int __init acpi20_parse_srat(struct acpi_table_srat *sratp)
223 177
224 memset(pxm_bitmap, 0, sizeof(pxm_bitmap)); /* init proximity domain bitmap */ 178 memset(pxm_bitmap, 0, sizeof(pxm_bitmap)); /* init proximity domain bitmap */
225 memset(node_memory_chunk, 0, sizeof(node_memory_chunk)); 179 memset(node_memory_chunk, 0, sizeof(node_memory_chunk));
226 memset(zholes_size, 0, sizeof(zholes_size));
227 180
228 num_memory_chunks = 0; 181 num_memory_chunks = 0;
229 while (p < end) { 182 while (p < end) {
@@ -287,6 +240,7 @@ static int __init acpi20_parse_srat(struct acpi_table_srat *sratp)
287 printk("chunk %d nid %d start_pfn %08lx end_pfn %08lx\n", 240 printk("chunk %d nid %d start_pfn %08lx end_pfn %08lx\n",
288 j, chunk->nid, chunk->start_pfn, chunk->end_pfn); 241 j, chunk->nid, chunk->start_pfn, chunk->end_pfn);
289 node_read_chunk(chunk->nid, chunk); 242 node_read_chunk(chunk->nid, chunk);
243 add_active_range(chunk->nid, chunk->start_pfn, chunk->end_pfn);
290 } 244 }
291 245
292 for_each_online_node(nid) { 246 for_each_online_node(nid) {
@@ -395,57 +349,7 @@ int __init get_memcfg_from_srat(void)
395 return acpi20_parse_srat((struct acpi_table_srat *)header); 349 return acpi20_parse_srat((struct acpi_table_srat *)header);
396 } 350 }
397out_err: 351out_err:
352 remove_all_active_ranges();
398 printk("failed to get NUMA memory information from SRAT table\n"); 353 printk("failed to get NUMA memory information from SRAT table\n");
399 return 0; 354 return 0;
400} 355}
401
402/* For each node run the memory list to determine whether there are
403 * any memory holes. For each hole determine which ZONE they fall
404 * into.
405 *
406 * NOTE#1: this requires knowledge of the zone boundries and so
407 * _cannot_ be performed before those are calculated in setup_memory.
408 *
409 * NOTE#2: we rely on the fact that the memory chunks are ordered by
410 * start pfn number during setup.
411 */
412static void __init get_zholes_init(void)
413{
414 int nid;
415 int c;
416 int first;
417 unsigned long end = 0;
418
419 for_each_online_node(nid) {
420 first = 1;
421 for (c = 0; c < num_memory_chunks; c++){
422 if (node_memory_chunk[c].nid == nid) {
423 if (first) {
424 end = node_memory_chunk[c].end_pfn;
425 first = 0;
426
427 } else {
428 /* Record any gap between this chunk
429 * and the previous chunk on this node
430 * against the zones it spans.
431 */
432 chunk_to_zones(end,
433 node_memory_chunk[c].start_pfn,
434 &zholes_size[nid * MAX_NR_ZONES]);
435 }
436 }
437 }
438 }
439}
440
441unsigned long * __init get_zholes_size(int nid)
442{
443 if (!zholes_size_init) {
444 zholes_size_init++;
445 get_zholes_init();
446 }
447 if (nid >= MAX_NUMNODES || !node_online(nid))
448 printk("%s: nid = %d is invalid/offline. num_online_nodes = %d",
449 __FUNCTION__, nid, num_online_nodes());
450 return &zholes_size[nid * MAX_NR_ZONES];
451}
diff --git a/arch/i386/kernel/stacktrace.c b/arch/i386/kernel/stacktrace.c
deleted file mode 100644
index e62a037ab399..000000000000
--- a/arch/i386/kernel/stacktrace.c
+++ /dev/null
@@ -1,98 +0,0 @@
1/*
2 * arch/i386/kernel/stacktrace.c
3 *
4 * Stack trace management functions
5 *
6 * Copyright (C) 2006 Red Hat, Inc., Ingo Molnar <mingo@redhat.com>
7 */
8#include <linux/sched.h>
9#include <linux/stacktrace.h>
10
11static inline int valid_stack_ptr(struct thread_info *tinfo, void *p)
12{
13 return p > (void *)tinfo &&
14 p < (void *)tinfo + THREAD_SIZE - 3;
15}
16
17/*
18 * Save stack-backtrace addresses into a stack_trace buffer:
19 */
20static inline unsigned long
21save_context_stack(struct stack_trace *trace, unsigned int skip,
22 struct thread_info *tinfo, unsigned long *stack,
23 unsigned long ebp)
24{
25 unsigned long addr;
26
27#ifdef CONFIG_FRAME_POINTER
28 while (valid_stack_ptr(tinfo, (void *)ebp)) {
29 addr = *(unsigned long *)(ebp + 4);
30 if (!skip)
31 trace->entries[trace->nr_entries++] = addr;
32 else
33 skip--;
34 if (trace->nr_entries >= trace->max_entries)
35 break;
36 /*
37 * break out of recursive entries (such as
38 * end_of_stack_stop_unwind_function):
39 */
40 if (ebp == *(unsigned long *)ebp)
41 break;
42
43 ebp = *(unsigned long *)ebp;
44 }
45#else
46 while (valid_stack_ptr(tinfo, stack)) {
47 addr = *stack++;
48 if (__kernel_text_address(addr)) {
49 if (!skip)
50 trace->entries[trace->nr_entries++] = addr;
51 else
52 skip--;
53 if (trace->nr_entries >= trace->max_entries)
54 break;
55 }
56 }
57#endif
58
59 return ebp;
60}
61
62/*
63 * Save stack-backtrace addresses into a stack_trace buffer.
64 * If all_contexts is set, all contexts (hardirq, softirq and process)
65 * are saved. If not set then only the current context is saved.
66 */
67void save_stack_trace(struct stack_trace *trace,
68 struct task_struct *task, int all_contexts,
69 unsigned int skip)
70{
71 unsigned long ebp;
72 unsigned long *stack = &ebp;
73
74 WARN_ON(trace->nr_entries || !trace->max_entries);
75
76 if (!task || task == current) {
77 /* Grab ebp right from our regs: */
78 asm ("movl %%ebp, %0" : "=r" (ebp));
79 } else {
80 /* ebp is the last reg pushed by switch_to(): */
81 ebp = *(unsigned long *) task->thread.esp;
82 }
83
84 while (1) {
85 struct thread_info *context = (struct thread_info *)
86 ((unsigned long)stack & (~(THREAD_SIZE - 1)));
87
88 ebp = save_context_stack(trace, skip, context, stack, ebp);
89 stack = (unsigned long *)context->previous_esp;
90 if (!all_contexts || !stack ||
91 trace->nr_entries >= trace->max_entries)
92 break;
93 trace->entries[trace->nr_entries++] = ULONG_MAX;
94 if (trace->nr_entries >= trace->max_entries)
95 break;
96 }
97}
98
diff --git a/arch/i386/kernel/syscall_table.S b/arch/i386/kernel/syscall_table.S
index dd63d4775398..7e639f78b0b9 100644
--- a/arch/i386/kernel/syscall_table.S
+++ b/arch/i386/kernel/syscall_table.S
@@ -317,3 +317,4 @@ ENTRY(sys_call_table)
317 .long sys_tee /* 315 */ 317 .long sys_tee /* 315 */
318 .long sys_vmsplice 318 .long sys_vmsplice
319 .long sys_move_pages 319 .long sys_move_pages
320 .long sys_getcpu
diff --git a/arch/i386/kernel/time.c b/arch/i386/kernel/time.c
index edd00f6cee37..86944acfb647 100644
--- a/arch/i386/kernel/time.c
+++ b/arch/i386/kernel/time.c
@@ -130,18 +130,33 @@ static int set_rtc_mmss(unsigned long nowtime)
130 130
131int timer_ack; 131int timer_ack;
132 132
133#if defined(CONFIG_SMP) && defined(CONFIG_FRAME_POINTER)
134unsigned long profile_pc(struct pt_regs *regs) 133unsigned long profile_pc(struct pt_regs *regs)
135{ 134{
136 unsigned long pc = instruction_pointer(regs); 135 unsigned long pc = instruction_pointer(regs);
137 136
138 if (!user_mode_vm(regs) && in_lock_functions(pc)) 137#ifdef CONFIG_SMP
138 if (!user_mode_vm(regs) && in_lock_functions(pc)) {
139#ifdef CONFIG_FRAME_POINTER
139 return *(unsigned long *)(regs->ebp + 4); 140 return *(unsigned long *)(regs->ebp + 4);
140 141#else
142 unsigned long *sp;
143 if ((regs->xcs & 3) == 0)
144 sp = (unsigned long *)&regs->esp;
145 else
146 sp = (unsigned long *)regs->esp;
147 /* Return address is either directly at stack pointer
148 or above a saved eflags. Eflags has bits 22-31 zero,
149 kernel addresses don't. */
150 if (sp[0] >> 22)
151 return sp[0];
152 if (sp[1] >> 22)
153 return sp[1];
154#endif
155 }
156#endif
141 return pc; 157 return pc;
142} 158}
143EXPORT_SYMBOL(profile_pc); 159EXPORT_SYMBOL(profile_pc);
144#endif
145 160
146/* 161/*
147 * This is the same as the above, except we _also_ save the current 162 * This is the same as the above, except we _also_ save the current
@@ -270,16 +285,19 @@ void notify_arch_cmos_timer(void)
270 mod_timer(&sync_cmos_timer, jiffies + 1); 285 mod_timer(&sync_cmos_timer, jiffies + 1);
271} 286}
272 287
273static long clock_cmos_diff, sleep_start; 288static long clock_cmos_diff;
289static unsigned long sleep_start;
274 290
275static int timer_suspend(struct sys_device *dev, pm_message_t state) 291static int timer_suspend(struct sys_device *dev, pm_message_t state)
276{ 292{
277 /* 293 /*
278 * Estimate time zone so that set_time can update the clock 294 * Estimate time zone so that set_time can update the clock
279 */ 295 */
280 clock_cmos_diff = -get_cmos_time(); 296 unsigned long ctime = get_cmos_time();
297
298 clock_cmos_diff = -ctime;
281 clock_cmos_diff += get_seconds(); 299 clock_cmos_diff += get_seconds();
282 sleep_start = get_cmos_time(); 300 sleep_start = ctime;
283 return 0; 301 return 0;
284} 302}
285 303
@@ -287,18 +305,29 @@ static int timer_resume(struct sys_device *dev)
287{ 305{
288 unsigned long flags; 306 unsigned long flags;
289 unsigned long sec; 307 unsigned long sec;
290 unsigned long sleep_length; 308 unsigned long ctime = get_cmos_time();
291 309 long sleep_length = (ctime - sleep_start) * HZ;
310 struct timespec ts;
311
312 if (sleep_length < 0) {
313 printk(KERN_WARNING "CMOS clock skew detected in timer resume!\n");
314 /* The time after the resume must not be earlier than the time
315 * before the suspend or some nasty things will happen
316 */
317 sleep_length = 0;
318 ctime = sleep_start;
319 }
292#ifdef CONFIG_HPET_TIMER 320#ifdef CONFIG_HPET_TIMER
293 if (is_hpet_enabled()) 321 if (is_hpet_enabled())
294 hpet_reenable(); 322 hpet_reenable();
295#endif 323#endif
296 setup_pit_timer(); 324 setup_pit_timer();
297 sec = get_cmos_time() + clock_cmos_diff; 325
298 sleep_length = (get_cmos_time() - sleep_start) * HZ; 326 sec = ctime + clock_cmos_diff;
327 ts.tv_sec = sec;
328 ts.tv_nsec = 0;
329 do_settimeofday(&ts);
299 write_seqlock_irqsave(&xtime_lock, flags); 330 write_seqlock_irqsave(&xtime_lock, flags);
300 xtime.tv_sec = sec;
301 xtime.tv_nsec = 0;
302 jiffies_64 += sleep_length; 331 jiffies_64 += sleep_length;
303 wall_jiffies += sleep_length; 332 wall_jiffies += sleep_length;
304 write_sequnlock_irqrestore(&xtime_lock, flags); 333 write_sequnlock_irqrestore(&xtime_lock, flags);
@@ -334,10 +363,11 @@ extern void (*late_time_init)(void);
334/* Duplicate of time_init() below, with hpet_enable part added */ 363/* Duplicate of time_init() below, with hpet_enable part added */
335static void __init hpet_time_init(void) 364static void __init hpet_time_init(void)
336{ 365{
337 xtime.tv_sec = get_cmos_time(); 366 struct timespec ts;
338 xtime.tv_nsec = (INITIAL_JIFFIES % HZ) * (NSEC_PER_SEC / HZ); 367 ts.tv_sec = get_cmos_time();
339 set_normalized_timespec(&wall_to_monotonic, 368 ts.tv_nsec = (INITIAL_JIFFIES % HZ) * (NSEC_PER_SEC / HZ);
340 -xtime.tv_sec, -xtime.tv_nsec); 369
370 do_settimeofday(&ts);
341 371
342 if ((hpet_enable() >= 0) && hpet_use_timer) { 372 if ((hpet_enable() >= 0) && hpet_use_timer) {
343 printk("Using HPET for base-timer\n"); 373 printk("Using HPET for base-timer\n");
@@ -349,6 +379,7 @@ static void __init hpet_time_init(void)
349 379
350void __init time_init(void) 380void __init time_init(void)
351{ 381{
382 struct timespec ts;
352#ifdef CONFIG_HPET_TIMER 383#ifdef CONFIG_HPET_TIMER
353 if (is_hpet_capable()) { 384 if (is_hpet_capable()) {
354 /* 385 /*
@@ -359,10 +390,10 @@ void __init time_init(void)
359 return; 390 return;
360 } 391 }
361#endif 392#endif
362 xtime.tv_sec = get_cmos_time(); 393 ts.tv_sec = get_cmos_time();
363 xtime.tv_nsec = (INITIAL_JIFFIES % HZ) * (NSEC_PER_SEC / HZ); 394 ts.tv_nsec = (INITIAL_JIFFIES % HZ) * (NSEC_PER_SEC / HZ);
364 set_normalized_timespec(&wall_to_monotonic, 395
365 -xtime.tv_sec, -xtime.tv_nsec); 396 do_settimeofday(&ts);
366 397
367 time_init_hook(); 398 time_init_hook();
368} 399}
diff --git a/arch/i386/kernel/time_hpet.c b/arch/i386/kernel/time_hpet.c
index 14a1376fedd1..6bf14a4e995e 100644
--- a/arch/i386/kernel/time_hpet.c
+++ b/arch/i386/kernel/time_hpet.c
@@ -301,23 +301,25 @@ int hpet_rtc_timer_init(void)
301 hpet_rtc_int_freq = DEFAULT_RTC_INT_FREQ; 301 hpet_rtc_int_freq = DEFAULT_RTC_INT_FREQ;
302 302
303 local_irq_save(flags); 303 local_irq_save(flags);
304
304 cnt = hpet_readl(HPET_COUNTER); 305 cnt = hpet_readl(HPET_COUNTER);
305 cnt += ((hpet_tick*HZ)/hpet_rtc_int_freq); 306 cnt += ((hpet_tick*HZ)/hpet_rtc_int_freq);
306 hpet_writel(cnt, HPET_T1_CMP); 307 hpet_writel(cnt, HPET_T1_CMP);
307 hpet_t1_cmp = cnt; 308 hpet_t1_cmp = cnt;
308 local_irq_restore(flags);
309 309
310 cfg = hpet_readl(HPET_T1_CFG); 310 cfg = hpet_readl(HPET_T1_CFG);
311 cfg &= ~HPET_TN_PERIODIC; 311 cfg &= ~HPET_TN_PERIODIC;
312 cfg |= HPET_TN_ENABLE | HPET_TN_32BIT; 312 cfg |= HPET_TN_ENABLE | HPET_TN_32BIT;
313 hpet_writel(cfg, HPET_T1_CFG); 313 hpet_writel(cfg, HPET_T1_CFG);
314 314
315 local_irq_restore(flags);
316
315 return 1; 317 return 1;
316} 318}
317 319
318static void hpet_rtc_timer_reinit(void) 320static void hpet_rtc_timer_reinit(void)
319{ 321{
320 unsigned int cfg, cnt; 322 unsigned int cfg, cnt, ticks_per_int, lost_ints;
321 323
322 if (unlikely(!(PIE_on | AIE_on | UIE_on))) { 324 if (unlikely(!(PIE_on | AIE_on | UIE_on))) {
323 cfg = hpet_readl(HPET_T1_CFG); 325 cfg = hpet_readl(HPET_T1_CFG);
@@ -332,10 +334,33 @@ static void hpet_rtc_timer_reinit(void)
332 hpet_rtc_int_freq = DEFAULT_RTC_INT_FREQ; 334 hpet_rtc_int_freq = DEFAULT_RTC_INT_FREQ;
333 335
334 /* It is more accurate to use the comparator value than current count.*/ 336 /* It is more accurate to use the comparator value than current count.*/
335 cnt = hpet_t1_cmp; 337 ticks_per_int = hpet_tick * HZ / hpet_rtc_int_freq;
336 cnt += hpet_tick*HZ/hpet_rtc_int_freq; 338 hpet_t1_cmp += ticks_per_int;
337 hpet_writel(cnt, HPET_T1_CMP); 339 hpet_writel(hpet_t1_cmp, HPET_T1_CMP);
338 hpet_t1_cmp = cnt; 340
341 /*
342 * If the interrupt handler was delayed too long, the write above tries
343 * to schedule the next interrupt in the past and the hardware would
344 * not interrupt until the counter had wrapped around.
345 * So we have to check that the comparator wasn't set to a past time.
346 */
347 cnt = hpet_readl(HPET_COUNTER);
348 if (unlikely((int)(cnt - hpet_t1_cmp) > 0)) {
349 lost_ints = (cnt - hpet_t1_cmp) / ticks_per_int + 1;
350 /* Make sure that, even with the time needed to execute
351 * this code, the next scheduled interrupt has been moved
352 * back to the future: */
353 lost_ints++;
354
355 hpet_t1_cmp += lost_ints * ticks_per_int;
356 hpet_writel(hpet_t1_cmp, HPET_T1_CMP);
357
358 if (PIE_on)
359 PIE_count += lost_ints;
360
361 printk(KERN_WARNING "rtc: lost some interrupts at %ldHz.\n",
362 hpet_rtc_int_freq);
363 }
339} 364}
340 365
341/* 366/*
diff --git a/arch/i386/kernel/topology.c b/arch/i386/kernel/topology.c
index e2e281d4bcc8..07d6da36a825 100644
--- a/arch/i386/kernel/topology.c
+++ b/arch/i386/kernel/topology.c
@@ -28,6 +28,7 @@
28#include <linux/init.h> 28#include <linux/init.h>
29#include <linux/smp.h> 29#include <linux/smp.h>
30#include <linux/nodemask.h> 30#include <linux/nodemask.h>
31#include <linux/mmzone.h>
31#include <asm/cpu.h> 32#include <asm/cpu.h>
32 33
33static struct i386_cpu cpu_devices[NR_CPUS]; 34static struct i386_cpu cpu_devices[NR_CPUS];
@@ -55,34 +56,18 @@ EXPORT_SYMBOL(arch_register_cpu);
55EXPORT_SYMBOL(arch_unregister_cpu); 56EXPORT_SYMBOL(arch_unregister_cpu);
56#endif /*CONFIG_HOTPLUG_CPU*/ 57#endif /*CONFIG_HOTPLUG_CPU*/
57 58
58
59
60#ifdef CONFIG_NUMA
61#include <linux/mmzone.h>
62
63static int __init topology_init(void) 59static int __init topology_init(void)
64{ 60{
65 int i; 61 int i;
66 62
63#ifdef CONFIG_NUMA
67 for_each_online_node(i) 64 for_each_online_node(i)
68 register_one_node(i); 65 register_one_node(i);
66#endif /* CONFIG_NUMA */
69 67
70 for_each_present_cpu(i) 68 for_each_present_cpu(i)
71 arch_register_cpu(i); 69 arch_register_cpu(i);
72 return 0; 70 return 0;
73} 71}
74 72
75#else /* !CONFIG_NUMA */
76
77static int __init topology_init(void)
78{
79 int i;
80
81 for_each_present_cpu(i)
82 arch_register_cpu(i);
83 return 0;
84}
85
86#endif /* CONFIG_NUMA */
87
88subsys_initcall(topology_init); 73subsys_initcall(topology_init);
diff --git a/arch/i386/kernel/traps.c b/arch/i386/kernel/traps.c
index 7e9edafffd8a..a13037fe0ee3 100644
--- a/arch/i386/kernel/traps.c
+++ b/arch/i386/kernel/traps.c
@@ -28,6 +28,7 @@
28#include <linux/kprobes.h> 28#include <linux/kprobes.h>
29#include <linux/kexec.h> 29#include <linux/kexec.h>
30#include <linux/unwind.h> 30#include <linux/unwind.h>
31#include <linux/uaccess.h>
31 32
32#ifdef CONFIG_EISA 33#ifdef CONFIG_EISA
33#include <linux/ioport.h> 34#include <linux/ioport.h>
@@ -40,7 +41,6 @@
40 41
41#include <asm/processor.h> 42#include <asm/processor.h>
42#include <asm/system.h> 43#include <asm/system.h>
43#include <asm/uaccess.h>
44#include <asm/io.h> 44#include <asm/io.h>
45#include <asm/atomic.h> 45#include <asm/atomic.h>
46#include <asm/debugreg.h> 46#include <asm/debugreg.h>
@@ -51,6 +51,7 @@
51#include <asm/smp.h> 51#include <asm/smp.h>
52#include <asm/arch_hooks.h> 52#include <asm/arch_hooks.h>
53#include <asm/kdebug.h> 53#include <asm/kdebug.h>
54#include <asm/stacktrace.h>
54 55
55#include <linux/module.h> 56#include <linux/module.h>
56 57
@@ -118,26 +119,16 @@ static inline int valid_stack_ptr(struct thread_info *tinfo, void *p)
118 p < (void *)tinfo + THREAD_SIZE - 3; 119 p < (void *)tinfo + THREAD_SIZE - 3;
119} 120}
120 121
121/*
122 * Print one address/symbol entries per line.
123 */
124static inline void print_addr_and_symbol(unsigned long addr, char *log_lvl)
125{
126 printk(" [<%08lx>] ", addr);
127
128 print_symbol("%s\n", addr);
129}
130
131static inline unsigned long print_context_stack(struct thread_info *tinfo, 122static inline unsigned long print_context_stack(struct thread_info *tinfo,
132 unsigned long *stack, unsigned long ebp, 123 unsigned long *stack, unsigned long ebp,
133 char *log_lvl) 124 struct stacktrace_ops *ops, void *data)
134{ 125{
135 unsigned long addr; 126 unsigned long addr;
136 127
137#ifdef CONFIG_FRAME_POINTER 128#ifdef CONFIG_FRAME_POINTER
138 while (valid_stack_ptr(tinfo, (void *)ebp)) { 129 while (valid_stack_ptr(tinfo, (void *)ebp)) {
139 addr = *(unsigned long *)(ebp + 4); 130 addr = *(unsigned long *)(ebp + 4);
140 print_addr_and_symbol(addr, log_lvl); 131 ops->address(data, addr);
141 /* 132 /*
142 * break out of recursive entries (such as 133 * break out of recursive entries (such as
143 * end_of_stack_stop_unwind_function): 134 * end_of_stack_stop_unwind_function):
@@ -150,30 +141,37 @@ static inline unsigned long print_context_stack(struct thread_info *tinfo,
150 while (valid_stack_ptr(tinfo, stack)) { 141 while (valid_stack_ptr(tinfo, stack)) {
151 addr = *stack++; 142 addr = *stack++;
152 if (__kernel_text_address(addr)) 143 if (__kernel_text_address(addr))
153 print_addr_and_symbol(addr, log_lvl); 144 ops->address(data, addr);
154 } 145 }
155#endif 146#endif
156 return ebp; 147 return ebp;
157} 148}
158 149
150struct ops_and_data {
151 struct stacktrace_ops *ops;
152 void *data;
153};
154
159static asmlinkage int 155static asmlinkage int
160show_trace_unwind(struct unwind_frame_info *info, void *log_lvl) 156dump_trace_unwind(struct unwind_frame_info *info, void *data)
161{ 157{
158 struct ops_and_data *oad = (struct ops_and_data *)data;
162 int n = 0; 159 int n = 0;
163 160
164 while (unwind(info) == 0 && UNW_PC(info)) { 161 while (unwind(info) == 0 && UNW_PC(info)) {
165 n++; 162 n++;
166 print_addr_and_symbol(UNW_PC(info), log_lvl); 163 oad->ops->address(oad->data, UNW_PC(info));
167 if (arch_unw_user_mode(info)) 164 if (arch_unw_user_mode(info))
168 break; 165 break;
169 } 166 }
170 return n; 167 return n;
171} 168}
172 169
173static void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, 170void dump_trace(struct task_struct *task, struct pt_regs *regs,
174 unsigned long *stack, char *log_lvl) 171 unsigned long *stack,
172 struct stacktrace_ops *ops, void *data)
175{ 173{
176 unsigned long ebp; 174 unsigned long ebp = 0;
177 175
178 if (!task) 176 if (!task)
179 task = current; 177 task = current;
@@ -181,54 +179,116 @@ static void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
181 if (call_trace >= 0) { 179 if (call_trace >= 0) {
182 int unw_ret = 0; 180 int unw_ret = 0;
183 struct unwind_frame_info info; 181 struct unwind_frame_info info;
182 struct ops_and_data oad = { .ops = ops, .data = data };
184 183
185 if (regs) { 184 if (regs) {
186 if (unwind_init_frame_info(&info, task, regs) == 0) 185 if (unwind_init_frame_info(&info, task, regs) == 0)
187 unw_ret = show_trace_unwind(&info, log_lvl); 186 unw_ret = dump_trace_unwind(&info, &oad);
188 } else if (task == current) 187 } else if (task == current)
189 unw_ret = unwind_init_running(&info, show_trace_unwind, log_lvl); 188 unw_ret = unwind_init_running(&info, dump_trace_unwind, &oad);
190 else { 189 else {
191 if (unwind_init_blocked(&info, task) == 0) 190 if (unwind_init_blocked(&info, task) == 0)
192 unw_ret = show_trace_unwind(&info, log_lvl); 191 unw_ret = dump_trace_unwind(&info, &oad);
193 } 192 }
194 if (unw_ret > 0) { 193 if (unw_ret > 0) {
195 if (call_trace == 1 && !arch_unw_user_mode(&info)) { 194 if (call_trace == 1 && !arch_unw_user_mode(&info)) {
196 print_symbol("DWARF2 unwinder stuck at %s\n", 195 ops->warning_symbol(data, "DWARF2 unwinder stuck at %s\n",
197 UNW_PC(&info)); 196 UNW_PC(&info));
198 if (UNW_SP(&info) >= PAGE_OFFSET) { 197 if (UNW_SP(&info) >= PAGE_OFFSET) {
199 printk("Leftover inexact backtrace:\n"); 198 ops->warning(data, "Leftover inexact backtrace:\n");
200 stack = (void *)UNW_SP(&info); 199 stack = (void *)UNW_SP(&info);
200 if (!stack)
201 return;
202 ebp = UNW_FP(&info);
201 } else 203 } else
202 printk("Full inexact backtrace again:\n"); 204 ops->warning(data, "Full inexact backtrace again:\n");
203 } else if (call_trace >= 1) 205 } else if (call_trace >= 1)
204 return; 206 return;
205 else 207 else
206 printk("Full inexact backtrace again:\n"); 208 ops->warning(data, "Full inexact backtrace again:\n");
207 } else 209 } else
208 printk("Inexact backtrace:\n"); 210 ops->warning(data, "Inexact backtrace:\n");
211 }
212 if (!stack) {
213 unsigned long dummy;
214 stack = &dummy;
215 if (task && task != current)
216 stack = (unsigned long *)task->thread.esp;
209 } 217 }
210 218
211 if (task == current) { 219#ifdef CONFIG_FRAME_POINTER
212 /* Grab ebp right from our regs */ 220 if (!ebp) {
213 asm ("movl %%ebp, %0" : "=r" (ebp) : ); 221 if (task == current) {
214 } else { 222 /* Grab ebp right from our regs */
215 /* ebp is the last reg pushed by switch_to */ 223 asm ("movl %%ebp, %0" : "=r" (ebp) : );
216 ebp = *(unsigned long *) task->thread.esp; 224 } else {
225 /* ebp is the last reg pushed by switch_to */
226 ebp = *(unsigned long *) task->thread.esp;
227 }
217 } 228 }
229#endif
218 230
219 while (1) { 231 while (1) {
220 struct thread_info *context; 232 struct thread_info *context;
221 context = (struct thread_info *) 233 context = (struct thread_info *)
222 ((unsigned long)stack & (~(THREAD_SIZE - 1))); 234 ((unsigned long)stack & (~(THREAD_SIZE - 1)));
223 ebp = print_context_stack(context, stack, ebp, log_lvl); 235 ebp = print_context_stack(context, stack, ebp, ops, data);
236 /* Should be after the line below, but somewhere
237 in early boot context comes out corrupted and we
238 can't reference it -AK */
239 if (ops->stack(data, "IRQ") < 0)
240 break;
224 stack = (unsigned long*)context->previous_esp; 241 stack = (unsigned long*)context->previous_esp;
225 if (!stack) 242 if (!stack)
226 break; 243 break;
227 printk("%s =======================\n", log_lvl);
228 } 244 }
229} 245}
246EXPORT_SYMBOL(dump_trace);
247
248static void
249print_trace_warning_symbol(void *data, char *msg, unsigned long symbol)
250{
251 printk(data);
252 print_symbol(msg, symbol);
253 printk("\n");
254}
255
256static void print_trace_warning(void *data, char *msg)
257{
258 printk("%s%s\n", (char *)data, msg);
259}
260
261static int print_trace_stack(void *data, char *name)
262{
263 return 0;
264}
265
266/*
267 * Print one address/symbol entries per line.
268 */
269static void print_trace_address(void *data, unsigned long addr)
270{
271 printk("%s [<%08lx>] ", (char *)data, addr);
272 print_symbol("%s\n", addr);
273}
274
275static struct stacktrace_ops print_trace_ops = {
276 .warning = print_trace_warning,
277 .warning_symbol = print_trace_warning_symbol,
278 .stack = print_trace_stack,
279 .address = print_trace_address,
280};
281
282static void
283show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
284 unsigned long * stack, char *log_lvl)
285{
286 dump_trace(task, regs, stack, &print_trace_ops, log_lvl);
287 printk("%s =======================\n", log_lvl);
288}
230 289
231void show_trace(struct task_struct *task, struct pt_regs *regs, unsigned long * stack) 290void show_trace(struct task_struct *task, struct pt_regs *regs,
291 unsigned long * stack)
232{ 292{
233 show_trace_log_lvl(task, regs, stack, ""); 293 show_trace_log_lvl(task, regs, stack, "");
234} 294}
@@ -291,8 +351,9 @@ void show_registers(struct pt_regs *regs)
291 ss = regs->xss & 0xffff; 351 ss = regs->xss & 0xffff;
292 } 352 }
293 print_modules(); 353 print_modules();
294 printk(KERN_EMERG "CPU: %d\nEIP: %04x:[<%08lx>] %s VLI\n" 354 printk(KERN_EMERG "CPU: %d\n"
295 "EFLAGS: %08lx (%s %.*s) \n", 355 KERN_EMERG "EIP: %04x:[<%08lx>] %s VLI\n"
356 KERN_EMERG "EFLAGS: %08lx (%s %.*s)\n",
296 smp_processor_id(), 0xffff & regs->xcs, regs->eip, 357 smp_processor_id(), 0xffff & regs->xcs, regs->eip,
297 print_tainted(), regs->eflags, system_utsname.release, 358 print_tainted(), regs->eflags, system_utsname.release,
298 (int)strcspn(system_utsname.version, " "), 359 (int)strcspn(system_utsname.version, " "),
@@ -313,6 +374,8 @@ void show_registers(struct pt_regs *regs)
313 */ 374 */
314 if (in_kernel) { 375 if (in_kernel) {
315 u8 __user *eip; 376 u8 __user *eip;
377 int code_bytes = 64;
378 unsigned char c;
316 379
317 printk("\n" KERN_EMERG "Stack: "); 380 printk("\n" KERN_EMERG "Stack: ");
318 show_stack_log_lvl(NULL, regs, (unsigned long *)esp, KERN_EMERG); 381 show_stack_log_lvl(NULL, regs, (unsigned long *)esp, KERN_EMERG);
@@ -320,9 +383,12 @@ void show_registers(struct pt_regs *regs)
320 printk(KERN_EMERG "Code: "); 383 printk(KERN_EMERG "Code: ");
321 384
322 eip = (u8 __user *)regs->eip - 43; 385 eip = (u8 __user *)regs->eip - 43;
323 for (i = 0; i < 64; i++, eip++) { 386 if (eip < (u8 __user *)PAGE_OFFSET || __get_user(c, eip)) {
324 unsigned char c; 387 /* try starting at EIP */
325 388 eip = (u8 __user *)regs->eip;
389 code_bytes = 32;
390 }
391 for (i = 0; i < code_bytes; i++, eip++) {
326 if (eip < (u8 __user *)PAGE_OFFSET || __get_user(c, eip)) { 392 if (eip < (u8 __user *)PAGE_OFFSET || __get_user(c, eip)) {
327 printk(" Bad EIP value."); 393 printk(" Bad EIP value.");
328 break; 394 break;
@@ -343,7 +409,7 @@ static void handle_BUG(struct pt_regs *regs)
343 409
344 if (eip < PAGE_OFFSET) 410 if (eip < PAGE_OFFSET)
345 return; 411 return;
346 if (__get_user(ud2, (unsigned short __user *)eip)) 412 if (probe_kernel_address((unsigned short __user *)eip, ud2))
347 return; 413 return;
348 if (ud2 != 0x0b0f) 414 if (ud2 != 0x0b0f)
349 return; 415 return;
@@ -356,7 +422,8 @@ static void handle_BUG(struct pt_regs *regs)
356 char *file; 422 char *file;
357 char c; 423 char c;
358 424
359 if (__get_user(line, (unsigned short __user *)(eip + 2))) 425 if (probe_kernel_address((unsigned short __user *)(eip + 2),
426 line))
360 break; 427 break;
361 if (__get_user(file, (char * __user *)(eip + 4)) || 428 if (__get_user(file, (char * __user *)(eip + 4)) ||
362 (unsigned long)file < PAGE_OFFSET || __get_user(c, file)) 429 (unsigned long)file < PAGE_OFFSET || __get_user(c, file))
@@ -629,18 +696,24 @@ gp_in_kernel:
629 } 696 }
630} 697}
631 698
632static void mem_parity_error(unsigned char reason, struct pt_regs * regs) 699static __kprobes void
700mem_parity_error(unsigned char reason, struct pt_regs * regs)
633{ 701{
634 printk(KERN_EMERG "Uhhuh. NMI received. Dazed and confused, but trying " 702 printk(KERN_EMERG "Uhhuh. NMI received for unknown reason %02x on "
635 "to continue\n"); 703 "CPU %d.\n", reason, smp_processor_id());
636 printk(KERN_EMERG "You probably have a hardware problem with your RAM " 704 printk(KERN_EMERG "You probably have a hardware problem with your RAM "
637 "chips\n"); 705 "chips\n");
706 if (panic_on_unrecovered_nmi)
707 panic("NMI: Not continuing");
708
709 printk(KERN_EMERG "Dazed and confused, but trying to continue\n");
638 710
639 /* Clear and disable the memory parity error line. */ 711 /* Clear and disable the memory parity error line. */
640 clear_mem_error(reason); 712 clear_mem_error(reason);
641} 713}
642 714
643static void io_check_error(unsigned char reason, struct pt_regs * regs) 715static __kprobes void
716io_check_error(unsigned char reason, struct pt_regs * regs)
644{ 717{
645 unsigned long i; 718 unsigned long i;
646 719
@@ -656,7 +729,8 @@ static void io_check_error(unsigned char reason, struct pt_regs * regs)
656 outb(reason, 0x61); 729 outb(reason, 0x61);
657} 730}
658 731
659static void unknown_nmi_error(unsigned char reason, struct pt_regs * regs) 732static __kprobes void
733unknown_nmi_error(unsigned char reason, struct pt_regs * regs)
660{ 734{
661#ifdef CONFIG_MCA 735#ifdef CONFIG_MCA
662 /* Might actually be able to figure out what the guilty party 736 /* Might actually be able to figure out what the guilty party
@@ -666,15 +740,18 @@ static void unknown_nmi_error(unsigned char reason, struct pt_regs * regs)
666 return; 740 return;
667 } 741 }
668#endif 742#endif
669 printk("Uhhuh. NMI received for unknown reason %02x on CPU %d.\n", 743 printk(KERN_EMERG "Uhhuh. NMI received for unknown reason %02x on "
670 reason, smp_processor_id()); 744 "CPU %d.\n", reason, smp_processor_id());
671 printk("Dazed and confused, but trying to continue\n"); 745 printk(KERN_EMERG "Do you have a strange power saving mode enabled?\n");
672 printk("Do you have a strange power saving mode enabled?\n"); 746 if (panic_on_unrecovered_nmi)
747 panic("NMI: Not continuing");
748
749 printk(KERN_EMERG "Dazed and confused, but trying to continue\n");
673} 750}
674 751
675static DEFINE_SPINLOCK(nmi_print_lock); 752static DEFINE_SPINLOCK(nmi_print_lock);
676 753
677void die_nmi (struct pt_regs *regs, const char *msg) 754void __kprobes die_nmi(struct pt_regs *regs, const char *msg)
678{ 755{
679 if (notify_die(DIE_NMIWATCHDOG, msg, regs, 0, 2, SIGINT) == 756 if (notify_die(DIE_NMIWATCHDOG, msg, regs, 0, 2, SIGINT) ==
680 NOTIFY_STOP) 757 NOTIFY_STOP)
@@ -706,7 +783,7 @@ void die_nmi (struct pt_regs *regs, const char *msg)
706 do_exit(SIGSEGV); 783 do_exit(SIGSEGV);
707} 784}
708 785
709static void default_do_nmi(struct pt_regs * regs) 786static __kprobes void default_do_nmi(struct pt_regs * regs)
710{ 787{
711 unsigned char reason = 0; 788 unsigned char reason = 0;
712 789
@@ -723,12 +800,12 @@ static void default_do_nmi(struct pt_regs * regs)
723 * Ok, so this is none of the documented NMI sources, 800 * Ok, so this is none of the documented NMI sources,
724 * so it must be the NMI watchdog. 801 * so it must be the NMI watchdog.
725 */ 802 */
726 if (nmi_watchdog) { 803 if (nmi_watchdog_tick(regs, reason))
727 nmi_watchdog_tick(regs);
728 return; 804 return;
729 } 805 if (!do_nmi_callback(regs, smp_processor_id()))
730#endif 806#endif
731 unknown_nmi_error(reason, regs); 807 unknown_nmi_error(reason, regs);
808
732 return; 809 return;
733 } 810 }
734 if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT) == NOTIFY_STOP) 811 if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT) == NOTIFY_STOP)
@@ -744,14 +821,7 @@ static void default_do_nmi(struct pt_regs * regs)
744 reassert_nmi(); 821 reassert_nmi();
745} 822}
746 823
747static int dummy_nmi_callback(struct pt_regs * regs, int cpu) 824fastcall __kprobes void do_nmi(struct pt_regs * regs, long error_code)
748{
749 return 0;
750}
751
752static nmi_callback_t nmi_callback = dummy_nmi_callback;
753
754fastcall void do_nmi(struct pt_regs * regs, long error_code)
755{ 825{
756 int cpu; 826 int cpu;
757 827
@@ -761,25 +831,11 @@ fastcall void do_nmi(struct pt_regs * regs, long error_code)
761 831
762 ++nmi_count(cpu); 832 ++nmi_count(cpu);
763 833
764 if (!rcu_dereference(nmi_callback)(regs, cpu)) 834 default_do_nmi(regs);
765 default_do_nmi(regs);
766 835
767 nmi_exit(); 836 nmi_exit();
768} 837}
769 838
770void set_nmi_callback(nmi_callback_t callback)
771{
772 vmalloc_sync_all();
773 rcu_assign_pointer(nmi_callback, callback);
774}
775EXPORT_SYMBOL_GPL(set_nmi_callback);
776
777void unset_nmi_callback(void)
778{
779 nmi_callback = dummy_nmi_callback;
780}
781EXPORT_SYMBOL_GPL(unset_nmi_callback);
782
783#ifdef CONFIG_KPROBES 839#ifdef CONFIG_KPROBES
784fastcall void __kprobes do_int3(struct pt_regs *regs, long error_code) 840fastcall void __kprobes do_int3(struct pt_regs *regs, long error_code)
785{ 841{
@@ -1119,20 +1175,6 @@ void __init trap_init_f00f_bug(void)
1119} 1175}
1120#endif 1176#endif
1121 1177
1122#define _set_gate(gate_addr,type,dpl,addr,seg) \
1123do { \
1124 int __d0, __d1; \
1125 __asm__ __volatile__ ("movw %%dx,%%ax\n\t" \
1126 "movw %4,%%dx\n\t" \
1127 "movl %%eax,%0\n\t" \
1128 "movl %%edx,%1" \
1129 :"=m" (*((long *) (gate_addr))), \
1130 "=m" (*(1+(long *) (gate_addr))), "=&a" (__d0), "=&d" (__d1) \
1131 :"i" ((short) (0x8000+(dpl<<13)+(type<<8))), \
1132 "3" ((char *) (addr)),"2" ((seg) << 16)); \
1133} while (0)
1134
1135
1136/* 1178/*
1137 * This needs to use 'idt_table' rather than 'idt', and 1179 * This needs to use 'idt_table' rather than 'idt', and
1138 * thus use the _nonmapped_ version of the IDT, as the 1180 * thus use the _nonmapped_ version of the IDT, as the
@@ -1141,7 +1183,7 @@ do { \
1141 */ 1183 */
1142void set_intr_gate(unsigned int n, void *addr) 1184void set_intr_gate(unsigned int n, void *addr)
1143{ 1185{
1144 _set_gate(idt_table+n,14,0,addr,__KERNEL_CS); 1186 _set_gate(n, DESCTYPE_INT, addr, __KERNEL_CS);
1145} 1187}
1146 1188
1147/* 1189/*
@@ -1149,22 +1191,22 @@ void set_intr_gate(unsigned int n, void *addr)
1149 */ 1191 */
1150static inline void set_system_intr_gate(unsigned int n, void *addr) 1192static inline void set_system_intr_gate(unsigned int n, void *addr)
1151{ 1193{
1152 _set_gate(idt_table+n, 14, 3, addr, __KERNEL_CS); 1194 _set_gate(n, DESCTYPE_INT | DESCTYPE_DPL3, addr, __KERNEL_CS);
1153} 1195}
1154 1196
1155static void __init set_trap_gate(unsigned int n, void *addr) 1197static void __init set_trap_gate(unsigned int n, void *addr)
1156{ 1198{
1157 _set_gate(idt_table+n,15,0,addr,__KERNEL_CS); 1199 _set_gate(n, DESCTYPE_TRAP, addr, __KERNEL_CS);
1158} 1200}
1159 1201
1160static void __init set_system_gate(unsigned int n, void *addr) 1202static void __init set_system_gate(unsigned int n, void *addr)
1161{ 1203{
1162 _set_gate(idt_table+n,15,3,addr,__KERNEL_CS); 1204 _set_gate(n, DESCTYPE_TRAP | DESCTYPE_DPL3, addr, __KERNEL_CS);
1163} 1205}
1164 1206
1165static void __init set_task_gate(unsigned int n, unsigned int gdt_entry) 1207static void __init set_task_gate(unsigned int n, unsigned int gdt_entry)
1166{ 1208{
1167 _set_gate(idt_table+n,5,0,0,(gdt_entry<<3)); 1209 _set_gate(n, DESCTYPE_TASK, (void *)0, (gdt_entry<<3));
1168} 1210}
1169 1211
1170 1212
diff --git a/arch/i386/kernel/tsc.c b/arch/i386/kernel/tsc.c
index 7e0d8dab2075..b8fa0a8b2e47 100644
--- a/arch/i386/kernel/tsc.c
+++ b/arch/i386/kernel/tsc.c
@@ -192,7 +192,7 @@ int recalibrate_cpu_khz(void)
192 192
193EXPORT_SYMBOL(recalibrate_cpu_khz); 193EXPORT_SYMBOL(recalibrate_cpu_khz);
194 194
195void tsc_init(void) 195void __init tsc_init(void)
196{ 196{
197 if (!cpu_has_tsc || tsc_disable) 197 if (!cpu_has_tsc || tsc_disable)
198 return; 198 return;
diff --git a/arch/i386/kernel/vmlinux.lds.S b/arch/i386/kernel/vmlinux.lds.S
index 2d4f1386e2b1..1e7ac1c44ddc 100644
--- a/arch/i386/kernel/vmlinux.lds.S
+++ b/arch/i386/kernel/vmlinux.lds.S
@@ -13,6 +13,12 @@ OUTPUT_FORMAT("elf32-i386", "elf32-i386", "elf32-i386")
13OUTPUT_ARCH(i386) 13OUTPUT_ARCH(i386)
14ENTRY(phys_startup_32) 14ENTRY(phys_startup_32)
15jiffies = jiffies_64; 15jiffies = jiffies_64;
16
17PHDRS {
18 text PT_LOAD FLAGS(5); /* R_E */
19 data PT_LOAD FLAGS(7); /* RWE */
20 note PT_NOTE FLAGS(4); /* R__ */
21}
16SECTIONS 22SECTIONS
17{ 23{
18 . = __KERNEL_START; 24 . = __KERNEL_START;
@@ -26,7 +32,7 @@ SECTIONS
26 KPROBES_TEXT 32 KPROBES_TEXT
27 *(.fixup) 33 *(.fixup)
28 *(.gnu.warning) 34 *(.gnu.warning)
29 } = 0x9090 35 } :text = 0x9090
30 36
31 _etext = .; /* End of text section */ 37 _etext = .; /* End of text section */
32 38
@@ -48,7 +54,7 @@ SECTIONS
48 .data : AT(ADDR(.data) - LOAD_OFFSET) { /* Data */ 54 .data : AT(ADDR(.data) - LOAD_OFFSET) { /* Data */
49 *(.data) 55 *(.data)
50 CONSTRUCTORS 56 CONSTRUCTORS
51 } 57 } :data
52 58
53 . = ALIGN(4096); 59 . = ALIGN(4096);
54 __nosave_begin = .; 60 __nosave_begin = .;
@@ -184,4 +190,6 @@ SECTIONS
184 STABS_DEBUG 190 STABS_DEBUG
185 191
186 DWARF_DEBUG 192 DWARF_DEBUG
193
194 NOTES
187} 195}
diff --git a/arch/i386/lib/Makefile b/arch/i386/lib/Makefile
index 914933e9ec3d..d86a548b8d54 100644
--- a/arch/i386/lib/Makefile
+++ b/arch/i386/lib/Makefile
@@ -4,6 +4,6 @@
4 4
5 5
6lib-y = checksum.o delay.o usercopy.o getuser.o putuser.o memcpy.o strstr.o \ 6lib-y = checksum.o delay.o usercopy.o getuser.o putuser.o memcpy.o strstr.o \
7 bitops.o 7 bitops.o semaphore.o
8 8
9lib-$(CONFIG_X86_USE_3DNOW) += mmx.o 9lib-$(CONFIG_X86_USE_3DNOW) += mmx.o
diff --git a/arch/i386/lib/semaphore.S b/arch/i386/lib/semaphore.S
new file mode 100644
index 000000000000..01f80b5c45d2
--- /dev/null
+++ b/arch/i386/lib/semaphore.S
@@ -0,0 +1,217 @@
1/*
2 * i386 semaphore implementation.
3 *
4 * (C) Copyright 1999 Linus Torvalds
5 *
6 * Portions Copyright 1999 Red Hat, Inc.
7 *
8 * This program is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU General Public License
10 * as published by the Free Software Foundation; either version
11 * 2 of the License, or (at your option) any later version.
12 *
13 * rw semaphores implemented November 1999 by Benjamin LaHaise <bcrl@kvack.org>
14 */
15
16#include <linux/config.h>
17#include <linux/linkage.h>
18#include <asm/rwlock.h>
19#include <asm/alternative-asm.i>
20#include <asm/frame.i>
21#include <asm/dwarf2.h>
22
23/*
24 * The semaphore operations have a special calling sequence that
25 * allow us to do a simpler in-line version of them. These routines
26 * need to convert that sequence back into the C sequence when
27 * there is contention on the semaphore.
28 *
29 * %eax contains the semaphore pointer on entry. Save the C-clobbered
30 * registers (%eax, %edx and %ecx) except %eax whish is either a return
31 * value or just clobbered..
32 */
33 .section .sched.text
34ENTRY(__down_failed)
35 CFI_STARTPROC
36 FRAME
37 pushl %edx
38 CFI_ADJUST_CFA_OFFSET 4
39 CFI_REL_OFFSET edx,0
40 pushl %ecx
41 CFI_ADJUST_CFA_OFFSET 4
42 CFI_REL_OFFSET ecx,0
43 call __down
44 popl %ecx
45 CFI_ADJUST_CFA_OFFSET -4
46 CFI_RESTORE ecx
47 popl %edx
48 CFI_ADJUST_CFA_OFFSET -4
49 CFI_RESTORE edx
50 ENDFRAME
51 ret
52 CFI_ENDPROC
53 END(__down_failed)
54
55ENTRY(__down_failed_interruptible)
56 CFI_STARTPROC
57 FRAME
58 pushl %edx
59 CFI_ADJUST_CFA_OFFSET 4
60 CFI_REL_OFFSET edx,0
61 pushl %ecx
62 CFI_ADJUST_CFA_OFFSET 4
63 CFI_REL_OFFSET ecx,0
64 call __down_interruptible
65 popl %ecx
66 CFI_ADJUST_CFA_OFFSET -4
67 CFI_RESTORE ecx
68 popl %edx
69 CFI_ADJUST_CFA_OFFSET -4
70 CFI_RESTORE edx
71 ENDFRAME
72 ret
73 CFI_ENDPROC
74 END(__down_failed_interruptible)
75
76ENTRY(__down_failed_trylock)
77 CFI_STARTPROC
78 FRAME
79 pushl %edx
80 CFI_ADJUST_CFA_OFFSET 4
81 CFI_REL_OFFSET edx,0
82 pushl %ecx
83 CFI_ADJUST_CFA_OFFSET 4
84 CFI_REL_OFFSET ecx,0
85 call __down_trylock
86 popl %ecx
87 CFI_ADJUST_CFA_OFFSET -4
88 CFI_RESTORE ecx
89 popl %edx
90 CFI_ADJUST_CFA_OFFSET -4
91 CFI_RESTORE edx
92 ENDFRAME
93 ret
94 CFI_ENDPROC
95 END(__down_failed_trylock)
96
97ENTRY(__up_wakeup)
98 CFI_STARTPROC
99 FRAME
100 pushl %edx
101 CFI_ADJUST_CFA_OFFSET 4
102 CFI_REL_OFFSET edx,0
103 pushl %ecx
104 CFI_ADJUST_CFA_OFFSET 4
105 CFI_REL_OFFSET ecx,0
106 call __up
107 popl %ecx
108 CFI_ADJUST_CFA_OFFSET -4
109 CFI_RESTORE ecx
110 popl %edx
111 CFI_ADJUST_CFA_OFFSET -4
112 CFI_RESTORE edx
113 ENDFRAME
114 ret
115 CFI_ENDPROC
116 END(__up_wakeup)
117
118/*
119 * rw spinlock fallbacks
120 */
121#ifdef CONFIG_SMP
122ENTRY(__write_lock_failed)
123 CFI_STARTPROC simple
124 FRAME
1252: LOCK_PREFIX
126 addl $ RW_LOCK_BIAS,(%eax)
1271: rep; nop
128 cmpl $ RW_LOCK_BIAS,(%eax)
129 jne 1b
130 LOCK_PREFIX
131 subl $ RW_LOCK_BIAS,(%eax)
132 jnz 2b
133 ENDFRAME
134 ret
135 CFI_ENDPROC
136 END(__write_lock_failed)
137
138ENTRY(__read_lock_failed)
139 CFI_STARTPROC
140 FRAME
1412: LOCK_PREFIX
142 incl (%eax)
1431: rep; nop
144 cmpl $1,(%eax)
145 js 1b
146 LOCK_PREFIX
147 decl (%eax)
148 js 2b
149 ENDFRAME
150 ret
151 CFI_ENDPROC
152 END(__read_lock_failed)
153
154#endif
155
156/* Fix up special calling conventions */
157ENTRY(call_rwsem_down_read_failed)
158 CFI_STARTPROC
159 push %ecx
160 CFI_ADJUST_CFA_OFFSET 4
161 CFI_REL_OFFSET ecx,0
162 push %edx
163 CFI_ADJUST_CFA_OFFSET 4
164 CFI_REL_OFFSET edx,0
165 call rwsem_down_read_failed
166 pop %edx
167 CFI_ADJUST_CFA_OFFSET -4
168 pop %ecx
169 CFI_ADJUST_CFA_OFFSET -4
170 ret
171 CFI_ENDPROC
172 END(call_rwsem_down_read_failed)
173
174ENTRY(call_rwsem_down_write_failed)
175 CFI_STARTPROC
176 push %ecx
177 CFI_ADJUST_CFA_OFFSET 4
178 CFI_REL_OFFSET ecx,0
179 calll rwsem_down_write_failed
180 pop %ecx
181 CFI_ADJUST_CFA_OFFSET -4
182 ret
183 CFI_ENDPROC
184 END(call_rwsem_down_write_failed)
185
186ENTRY(call_rwsem_wake)
187 CFI_STARTPROC
188 decw %dx /* do nothing if still outstanding active readers */
189 jnz 1f
190 push %ecx
191 CFI_ADJUST_CFA_OFFSET 4
192 CFI_REL_OFFSET ecx,0
193 call rwsem_wake
194 pop %ecx
195 CFI_ADJUST_CFA_OFFSET -4
1961: ret
197 CFI_ENDPROC
198 END(call_rwsem_wake)
199
200/* Fix up special calling conventions */
201ENTRY(call_rwsem_downgrade_wake)
202 CFI_STARTPROC
203 push %ecx
204 CFI_ADJUST_CFA_OFFSET 4
205 CFI_REL_OFFSET ecx,0
206 push %edx
207 CFI_ADJUST_CFA_OFFSET 4
208 CFI_REL_OFFSET edx,0
209 call rwsem_downgrade_wake
210 pop %edx
211 CFI_ADJUST_CFA_OFFSET -4
212 pop %ecx
213 CFI_ADJUST_CFA_OFFSET -4
214 ret
215 CFI_ENDPROC
216 END(call_rwsem_downgrade_wake)
217
diff --git a/arch/i386/mach-generic/bigsmp.c b/arch/i386/mach-generic/bigsmp.c
index ef7a6e6fcb9f..33d9f93557ba 100644
--- a/arch/i386/mach-generic/bigsmp.c
+++ b/arch/i386/mach-generic/bigsmp.c
@@ -5,6 +5,7 @@
5#define APIC_DEFINITION 1 5#define APIC_DEFINITION 1
6#include <linux/threads.h> 6#include <linux/threads.h>
7#include <linux/cpumask.h> 7#include <linux/cpumask.h>
8#include <asm/smp.h>
8#include <asm/mpspec.h> 9#include <asm/mpspec.h>
9#include <asm/genapic.h> 10#include <asm/genapic.h>
10#include <asm/fixmap.h> 11#include <asm/fixmap.h>
diff --git a/arch/i386/mach-generic/es7000.c b/arch/i386/mach-generic/es7000.c
index 845cdd0b3593..aa144d82334d 100644
--- a/arch/i386/mach-generic/es7000.c
+++ b/arch/i386/mach-generic/es7000.c
@@ -4,6 +4,7 @@
4#define APIC_DEFINITION 1 4#define APIC_DEFINITION 1
5#include <linux/threads.h> 5#include <linux/threads.h>
6#include <linux/cpumask.h> 6#include <linux/cpumask.h>
7#include <asm/smp.h>
7#include <asm/mpspec.h> 8#include <asm/mpspec.h>
8#include <asm/genapic.h> 9#include <asm/genapic.h>
9#include <asm/fixmap.h> 10#include <asm/fixmap.h>
diff --git a/arch/i386/mach-generic/probe.c b/arch/i386/mach-generic/probe.c
index bcd1bcfaa723..94b1fd9cbe3c 100644
--- a/arch/i386/mach-generic/probe.c
+++ b/arch/i386/mach-generic/probe.c
@@ -9,6 +9,7 @@
9#include <linux/kernel.h> 9#include <linux/kernel.h>
10#include <linux/ctype.h> 10#include <linux/ctype.h>
11#include <linux/init.h> 11#include <linux/init.h>
12#include <linux/errno.h>
12#include <asm/fixmap.h> 13#include <asm/fixmap.h>
13#include <asm/mpspec.h> 14#include <asm/mpspec.h>
14#include <asm/apicdef.h> 15#include <asm/apicdef.h>
@@ -29,7 +30,24 @@ struct genapic *apic_probe[] __initdata = {
29 NULL, 30 NULL,
30}; 31};
31 32
32static int cmdline_apic; 33static int cmdline_apic __initdata;
34static int __init parse_apic(char *arg)
35{
36 int i;
37
38 if (!arg)
39 return -EINVAL;
40
41 for (i = 0; apic_probe[i]; i++) {
42 if (!strcmp(apic_probe[i]->name, arg)) {
43 genapic = apic_probe[i];
44 cmdline_apic = 1;
45 return 0;
46 }
47 }
48 return -ENOENT;
49}
50early_param("apic", parse_apic);
33 51
34void __init generic_bigsmp_probe(void) 52void __init generic_bigsmp_probe(void)
35{ 53{
@@ -48,40 +66,20 @@ void __init generic_bigsmp_probe(void)
48 } 66 }
49} 67}
50 68
51void __init generic_apic_probe(char *command_line) 69void __init generic_apic_probe(void)
52{ 70{
53 char *s; 71 if (!cmdline_apic) {
54 int i; 72 int i;
55 int changed = 0; 73 for (i = 0; apic_probe[i]; i++) {
56 74 if (apic_probe[i]->probe()) {
57 s = strstr(command_line, "apic=");
58 if (s && (s == command_line || isspace(s[-1]))) {
59 char *p = strchr(s, ' '), old;
60 if (!p)
61 p = strchr(s, '\0');
62 old = *p;
63 *p = 0;
64 for (i = 0; !changed && apic_probe[i]; i++) {
65 if (!strcmp(apic_probe[i]->name, s+5)) {
66 changed = 1;
67 genapic = apic_probe[i]; 75 genapic = apic_probe[i];
76 break;
68 } 77 }
69 } 78 }
70 if (!changed) 79 /* Not visible without early console */
71 printk(KERN_ERR "Unknown genapic `%s' specified.\n", s); 80 if (!apic_probe[i])
72 *p = old; 81 panic("Didn't find an APIC driver");
73 cmdline_apic = changed;
74 }
75 for (i = 0; !changed && apic_probe[i]; i++) {
76 if (apic_probe[i]->probe()) {
77 changed = 1;
78 genapic = apic_probe[i];
79 }
80 } 82 }
81 /* Not visible without early console */
82 if (!changed)
83 panic("Didn't find an APIC driver");
84
85 printk(KERN_INFO "Using APIC driver %s\n", genapic->name); 83 printk(KERN_INFO "Using APIC driver %s\n", genapic->name);
86} 84}
87 85
@@ -119,7 +117,9 @@ int __init acpi_madt_oem_check(char *oem_id, char *oem_table_id)
119 return 0; 117 return 0;
120} 118}
121 119
120#ifdef CONFIG_SMP
122int hard_smp_processor_id(void) 121int hard_smp_processor_id(void)
123{ 122{
124 return genapic->get_apic_id(*(unsigned long *)(APIC_BASE+APIC_ID)); 123 return genapic->get_apic_id(*(unsigned long *)(APIC_BASE+APIC_ID));
125} 124}
125#endif
diff --git a/arch/i386/mach-generic/summit.c b/arch/i386/mach-generic/summit.c
index b73501ddd653..f7e5d66648dc 100644
--- a/arch/i386/mach-generic/summit.c
+++ b/arch/i386/mach-generic/summit.c
@@ -4,6 +4,7 @@
4#define APIC_DEFINITION 1 4#define APIC_DEFINITION 1
5#include <linux/threads.h> 5#include <linux/threads.h>
6#include <linux/cpumask.h> 6#include <linux/cpumask.h>
7#include <asm/smp.h>
7#include <asm/mpspec.h> 8#include <asm/mpspec.h>
8#include <asm/genapic.h> 9#include <asm/genapic.h>
9#include <asm/fixmap.h> 10#include <asm/fixmap.h>
diff --git a/arch/i386/mach-voyager/voyager_thread.c b/arch/i386/mach-voyager/voyager_thread.c
index 50f6de6ff64d..f39887359e8e 100644
--- a/arch/i386/mach-voyager/voyager_thread.c
+++ b/arch/i386/mach-voyager/voyager_thread.c
@@ -130,7 +130,6 @@ thread(void *unused)
130 init_timer(&wakeup_timer); 130 init_timer(&wakeup_timer);
131 131
132 sigfillset(&current->blocked); 132 sigfillset(&current->blocked);
133 current->signal->tty = NULL;
134 133
135 printk(KERN_NOTICE "Voyager starting monitor thread\n"); 134 printk(KERN_NOTICE "Voyager starting monitor thread\n");
136 135
diff --git a/arch/i386/mm/boot_ioremap.c b/arch/i386/mm/boot_ioremap.c
index 5d44f4f5ff59..4de11f508c3a 100644
--- a/arch/i386/mm/boot_ioremap.c
+++ b/arch/i386/mm/boot_ioremap.c
@@ -29,8 +29,11 @@
29 */ 29 */
30 30
31#define BOOT_PTE_PTRS (PTRS_PER_PTE*2) 31#define BOOT_PTE_PTRS (PTRS_PER_PTE*2)
32#define boot_pte_index(address) \ 32
33 (((address) >> PAGE_SHIFT) & (BOOT_PTE_PTRS - 1)) 33static unsigned long boot_pte_index(unsigned long vaddr)
34{
35 return __pa(vaddr) >> PAGE_SHIFT;
36}
34 37
35static inline boot_pte_t* boot_vaddr_to_pte(void *address) 38static inline boot_pte_t* boot_vaddr_to_pte(void *address)
36{ 39{
diff --git a/arch/i386/mm/discontig.c b/arch/i386/mm/discontig.c
index 7c392dc553b8..51e3739dd227 100644
--- a/arch/i386/mm/discontig.c
+++ b/arch/i386/mm/discontig.c
@@ -117,7 +117,8 @@ void set_pmd_pfn(unsigned long vaddr, unsigned long pfn, pgprot_t flags);
117 117
118void *node_remap_end_vaddr[MAX_NUMNODES]; 118void *node_remap_end_vaddr[MAX_NUMNODES];
119void *node_remap_alloc_vaddr[MAX_NUMNODES]; 119void *node_remap_alloc_vaddr[MAX_NUMNODES];
120 120static unsigned long kva_start_pfn;
121static unsigned long kva_pages;
121/* 122/*
122 * FLAT - support for basic PC memory model with discontig enabled, essentially 123 * FLAT - support for basic PC memory model with discontig enabled, essentially
123 * a single node with all available processors in it with a flat 124 * a single node with all available processors in it with a flat
@@ -156,21 +157,6 @@ static void __init find_max_pfn_node(int nid)
156 BUG(); 157 BUG();
157} 158}
158 159
159/* Find the owning node for a pfn. */
160int early_pfn_to_nid(unsigned long pfn)
161{
162 int nid;
163
164 for_each_node(nid) {
165 if (node_end_pfn[nid] == 0)
166 break;
167 if (node_start_pfn[nid] <= pfn && node_end_pfn[nid] >= pfn)
168 return nid;
169 }
170
171 return 0;
172}
173
174/* 160/*
175 * Allocate memory for the pg_data_t for this node via a crude pre-bootmem 161 * Allocate memory for the pg_data_t for this node via a crude pre-bootmem
176 * method. For node zero take this from the bottom of memory, for 162 * method. For node zero take this from the bottom of memory, for
@@ -226,6 +212,8 @@ static unsigned long calculate_numa_remap_pages(void)
226 unsigned long pfn; 212 unsigned long pfn;
227 213
228 for_each_online_node(nid) { 214 for_each_online_node(nid) {
215 unsigned old_end_pfn = node_end_pfn[nid];
216
229 /* 217 /*
230 * The acpi/srat node info can show hot-add memroy zones 218 * The acpi/srat node info can show hot-add memroy zones
231 * where memory could be added but not currently present. 219 * where memory could be added but not currently present.
@@ -275,6 +263,7 @@ static unsigned long calculate_numa_remap_pages(void)
275 263
276 node_end_pfn[nid] -= size; 264 node_end_pfn[nid] -= size;
277 node_remap_start_pfn[nid] = node_end_pfn[nid]; 265 node_remap_start_pfn[nid] = node_end_pfn[nid];
266 shrink_active_range(nid, old_end_pfn, node_end_pfn[nid]);
278 } 267 }
279 printk("Reserving total of %ld pages for numa KVA remap\n", 268 printk("Reserving total of %ld pages for numa KVA remap\n",
280 reserve_pages); 269 reserve_pages);
@@ -286,7 +275,6 @@ unsigned long __init setup_memory(void)
286{ 275{
287 int nid; 276 int nid;
288 unsigned long system_start_pfn, system_max_low_pfn; 277 unsigned long system_start_pfn, system_max_low_pfn;
289 unsigned long reserve_pages;
290 278
291 /* 279 /*
292 * When mapping a NUMA machine we allocate the node_mem_map arrays 280 * When mapping a NUMA machine we allocate the node_mem_map arrays
@@ -298,14 +286,23 @@ unsigned long __init setup_memory(void)
298 find_max_pfn(); 286 find_max_pfn();
299 get_memcfg_numa(); 287 get_memcfg_numa();
300 288
301 reserve_pages = calculate_numa_remap_pages(); 289 kva_pages = calculate_numa_remap_pages();
302 290
303 /* partially used pages are not usable - thus round upwards */ 291 /* partially used pages are not usable - thus round upwards */
304 system_start_pfn = min_low_pfn = PFN_UP(init_pg_tables_end); 292 system_start_pfn = min_low_pfn = PFN_UP(init_pg_tables_end);
305 293
306 system_max_low_pfn = max_low_pfn = find_max_low_pfn() - reserve_pages; 294 kva_start_pfn = find_max_low_pfn() - kva_pages;
307 printk("reserve_pages = %ld find_max_low_pfn() ~ %ld\n", 295
308 reserve_pages, max_low_pfn + reserve_pages); 296#ifdef CONFIG_BLK_DEV_INITRD
297 /* Numa kva area is below the initrd */
298 if (LOADER_TYPE && INITRD_START)
299 kva_start_pfn = PFN_DOWN(INITRD_START) - kva_pages;
300#endif
301 kva_start_pfn -= kva_start_pfn & (PTRS_PER_PTE-1);
302
303 system_max_low_pfn = max_low_pfn = find_max_low_pfn();
304 printk("kva_start_pfn ~ %ld find_max_low_pfn() ~ %ld\n",
305 kva_start_pfn, max_low_pfn);
309 printk("max_pfn = %ld\n", max_pfn); 306 printk("max_pfn = %ld\n", max_pfn);
310#ifdef CONFIG_HIGHMEM 307#ifdef CONFIG_HIGHMEM
311 highstart_pfn = highend_pfn = max_pfn; 308 highstart_pfn = highend_pfn = max_pfn;
@@ -313,6 +310,11 @@ unsigned long __init setup_memory(void)
313 highstart_pfn = system_max_low_pfn; 310 highstart_pfn = system_max_low_pfn;
314 printk(KERN_NOTICE "%ldMB HIGHMEM available.\n", 311 printk(KERN_NOTICE "%ldMB HIGHMEM available.\n",
315 pages_to_mb(highend_pfn - highstart_pfn)); 312 pages_to_mb(highend_pfn - highstart_pfn));
313 num_physpages = highend_pfn;
314 high_memory = (void *) __va(highstart_pfn * PAGE_SIZE - 1) + 1;
315#else
316 num_physpages = system_max_low_pfn;
317 high_memory = (void *) __va(system_max_low_pfn * PAGE_SIZE - 1) + 1;
316#endif 318#endif
317 printk(KERN_NOTICE "%ldMB LOWMEM available.\n", 319 printk(KERN_NOTICE "%ldMB LOWMEM available.\n",
318 pages_to_mb(system_max_low_pfn)); 320 pages_to_mb(system_max_low_pfn));
@@ -323,7 +325,7 @@ unsigned long __init setup_memory(void)
323 (ulong) pfn_to_kaddr(max_low_pfn)); 325 (ulong) pfn_to_kaddr(max_low_pfn));
324 for_each_online_node(nid) { 326 for_each_online_node(nid) {
325 node_remap_start_vaddr[nid] = pfn_to_kaddr( 327 node_remap_start_vaddr[nid] = pfn_to_kaddr(
326 highstart_pfn + node_remap_offset[nid]); 328 kva_start_pfn + node_remap_offset[nid]);
327 /* Init the node remap allocator */ 329 /* Init the node remap allocator */
328 node_remap_end_vaddr[nid] = node_remap_start_vaddr[nid] + 330 node_remap_end_vaddr[nid] = node_remap_start_vaddr[nid] +
329 (node_remap_size[nid] * PAGE_SIZE); 331 (node_remap_size[nid] * PAGE_SIZE);
@@ -338,7 +340,6 @@ unsigned long __init setup_memory(void)
338 } 340 }
339 printk("High memory starts at vaddr %08lx\n", 341 printk("High memory starts at vaddr %08lx\n",
340 (ulong) pfn_to_kaddr(highstart_pfn)); 342 (ulong) pfn_to_kaddr(highstart_pfn));
341 vmalloc_earlyreserve = reserve_pages * PAGE_SIZE;
342 for_each_online_node(nid) 343 for_each_online_node(nid)
343 find_max_pfn_node(nid); 344 find_max_pfn_node(nid);
344 345
@@ -348,48 +349,30 @@ unsigned long __init setup_memory(void)
348 return max_low_pfn; 349 return max_low_pfn;
349} 350}
350 351
352void __init numa_kva_reserve(void)
353{
354 reserve_bootmem(PFN_PHYS(kva_start_pfn),PFN_PHYS(kva_pages));
355}
356
351void __init zone_sizes_init(void) 357void __init zone_sizes_init(void)
352{ 358{
353 int nid; 359 int nid;
354 360 unsigned long max_zone_pfns[MAX_NR_ZONES] = {
355 361 virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT,
356 for_each_online_node(nid) { 362 max_low_pfn,
357 unsigned long zones_size[MAX_NR_ZONES] = {0, 0, 0}; 363 highend_pfn
358 unsigned long *zholes_size; 364 };
359 unsigned int max_dma; 365
360 366 /* If SRAT has not registered memory, register it now */
361 unsigned long low = max_low_pfn; 367 if (find_max_pfn_with_active_regions() == 0) {
362 unsigned long start = node_start_pfn[nid]; 368 for_each_online_node(nid) {
363 unsigned long high = node_end_pfn[nid]; 369 if (node_has_online_mem(nid))
364 370 add_active_range(nid, node_start_pfn[nid],
365 max_dma = virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT; 371 node_end_pfn[nid]);
366
367 if (node_has_online_mem(nid)){
368 if (start > low) {
369#ifdef CONFIG_HIGHMEM
370 BUG_ON(start > high);
371 zones_size[ZONE_HIGHMEM] = high - start;
372#endif
373 } else {
374 if (low < max_dma)
375 zones_size[ZONE_DMA] = low;
376 else {
377 BUG_ON(max_dma > low);
378 BUG_ON(low > high);
379 zones_size[ZONE_DMA] = max_dma;
380 zones_size[ZONE_NORMAL] = low - max_dma;
381#ifdef CONFIG_HIGHMEM
382 zones_size[ZONE_HIGHMEM] = high - low;
383#endif
384 }
385 }
386 } 372 }
387
388 zholes_size = get_zholes_size(nid);
389
390 free_area_init_node(nid, NODE_DATA(nid), zones_size, start,
391 zholes_size);
392 } 373 }
374
375 free_area_init_nodes(max_zone_pfns);
393 return; 376 return;
394} 377}
395 378
@@ -409,7 +392,7 @@ void __init set_highmem_pages_init(int bad_ppro)
409 zone_end_pfn = zone_start_pfn + zone->spanned_pages; 392 zone_end_pfn = zone_start_pfn + zone->spanned_pages;
410 393
411 printk("Initializing %s for node %d (%08lx:%08lx)\n", 394 printk("Initializing %s for node %d (%08lx:%08lx)\n",
412 zone->name, zone->zone_pgdat->node_id, 395 zone->name, zone_to_nid(zone),
413 zone_start_pfn, zone_end_pfn); 396 zone_start_pfn, zone_end_pfn);
414 397
415 for (node_pfn = zone_start_pfn; node_pfn < zone_end_pfn; node_pfn++) { 398 for (node_pfn = zone_start_pfn; node_pfn < zone_end_pfn; node_pfn++) {
diff --git a/arch/i386/mm/extable.c b/arch/i386/mm/extable.c
index de03c5430abc..0ce4f22a2635 100644
--- a/arch/i386/mm/extable.c
+++ b/arch/i386/mm/extable.c
@@ -11,7 +11,7 @@ int fixup_exception(struct pt_regs *regs)
11 const struct exception_table_entry *fixup; 11 const struct exception_table_entry *fixup;
12 12
13#ifdef CONFIG_PNPBIOS 13#ifdef CONFIG_PNPBIOS
14 if (unlikely((regs->xcs & ~15) == (GDT_ENTRY_PNPBIOS_BASE << 3))) 14 if (unlikely(SEGMENT_IS_PNP_CODE(regs->xcs)))
15 { 15 {
16 extern u32 pnp_bios_fault_eip, pnp_bios_fault_esp; 16 extern u32 pnp_bios_fault_eip, pnp_bios_fault_esp;
17 extern u32 pnp_bios_is_utter_crap; 17 extern u32 pnp_bios_is_utter_crap;
diff --git a/arch/i386/mm/fault.c b/arch/i386/mm/fault.c
index f7279468323a..5e17a3f43b41 100644
--- a/arch/i386/mm/fault.c
+++ b/arch/i386/mm/fault.c
@@ -27,21 +27,24 @@
27#include <asm/uaccess.h> 27#include <asm/uaccess.h>
28#include <asm/desc.h> 28#include <asm/desc.h>
29#include <asm/kdebug.h> 29#include <asm/kdebug.h>
30#include <asm/segment.h>
30 31
31extern void die(const char *,struct pt_regs *,long); 32extern void die(const char *,struct pt_regs *,long);
32 33
33#ifdef CONFIG_KPROBES 34static ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain);
34ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain); 35
35int register_page_fault_notifier(struct notifier_block *nb) 36int register_page_fault_notifier(struct notifier_block *nb)
36{ 37{
37 vmalloc_sync_all(); 38 vmalloc_sync_all();
38 return atomic_notifier_chain_register(&notify_page_fault_chain, nb); 39 return atomic_notifier_chain_register(&notify_page_fault_chain, nb);
39} 40}
41EXPORT_SYMBOL_GPL(register_page_fault_notifier);
40 42
41int unregister_page_fault_notifier(struct notifier_block *nb) 43int unregister_page_fault_notifier(struct notifier_block *nb)
42{ 44{
43 return atomic_notifier_chain_unregister(&notify_page_fault_chain, nb); 45 return atomic_notifier_chain_unregister(&notify_page_fault_chain, nb);
44} 46}
47EXPORT_SYMBOL_GPL(unregister_page_fault_notifier);
45 48
46static inline int notify_page_fault(enum die_val val, const char *str, 49static inline int notify_page_fault(enum die_val val, const char *str,
47 struct pt_regs *regs, long err, int trap, int sig) 50 struct pt_regs *regs, long err, int trap, int sig)
@@ -55,14 +58,6 @@ static inline int notify_page_fault(enum die_val val, const char *str,
55 }; 58 };
56 return atomic_notifier_call_chain(&notify_page_fault_chain, val, &args); 59 return atomic_notifier_call_chain(&notify_page_fault_chain, val, &args);
57} 60}
58#else
59static inline int notify_page_fault(enum die_val val, const char *str,
60 struct pt_regs *regs, long err, int trap, int sig)
61{
62 return NOTIFY_DONE;
63}
64#endif
65
66 61
67/* 62/*
68 * Unlock any spinlocks which will prevent us from getting the 63 * Unlock any spinlocks which will prevent us from getting the
@@ -119,10 +114,10 @@ static inline unsigned long get_segment_eip(struct pt_regs *regs,
119 } 114 }
120 115
121 /* The standard kernel/user address space limit. */ 116 /* The standard kernel/user address space limit. */
122 *eip_limit = (seg & 3) ? USER_DS.seg : KERNEL_DS.seg; 117 *eip_limit = user_mode(regs) ? USER_DS.seg : KERNEL_DS.seg;
123 118
124 /* By far the most common cases. */ 119 /* By far the most common cases. */
125 if (likely(seg == __USER_CS || seg == __KERNEL_CS)) 120 if (likely(SEGMENT_IS_FLAT_CODE(seg)))
126 return eip; 121 return eip;
127 122
128 /* Check the segment exists, is within the current LDT/GDT size, 123 /* Check the segment exists, is within the current LDT/GDT size,
@@ -436,11 +431,7 @@ good_area:
436 write = 0; 431 write = 0;
437 switch (error_code & 3) { 432 switch (error_code & 3) {
438 default: /* 3: write, present */ 433 default: /* 3: write, present */
439#ifdef TEST_VERIFY_AREA 434 /* fall through */
440 if (regs->cs == KERNEL_CS)
441 printk("WP fault at %08lx\n", regs->eip);
442#endif
443 /* fall through */
444 case 2: /* write, not present */ 435 case 2: /* write, not present */
445 if (!(vma->vm_flags & VM_WRITE)) 436 if (!(vma->vm_flags & VM_WRITE))
446 goto bad_area; 437 goto bad_area;
diff --git a/arch/i386/mm/highmem.c b/arch/i386/mm/highmem.c
index b6eb4dcb8777..ba44000b9069 100644
--- a/arch/i386/mm/highmem.c
+++ b/arch/i386/mm/highmem.c
@@ -54,7 +54,7 @@ void kunmap_atomic(void *kvaddr, enum km_type type)
54 unsigned long vaddr = (unsigned long) kvaddr & PAGE_MASK; 54 unsigned long vaddr = (unsigned long) kvaddr & PAGE_MASK;
55 enum fixed_addresses idx = type + KM_TYPE_NR*smp_processor_id(); 55 enum fixed_addresses idx = type + KM_TYPE_NR*smp_processor_id();
56 56
57 if (vaddr < FIXADDR_START) { // FIXME 57 if (vaddr >= PAGE_OFFSET && vaddr < (unsigned long)high_memory) {
58 dec_preempt_count(); 58 dec_preempt_count();
59 preempt_check_resched(); 59 preempt_check_resched();
60 return; 60 return;
diff --git a/arch/i386/mm/init.c b/arch/i386/mm/init.c
index 89e8486aac34..4a5a914b3432 100644
--- a/arch/i386/mm/init.c
+++ b/arch/i386/mm/init.c
@@ -435,16 +435,22 @@ u64 __supported_pte_mask __read_mostly = ~_PAGE_NX;
435 * on Enable 435 * on Enable
436 * off Disable 436 * off Disable
437 */ 437 */
438void __init noexec_setup(const char *str) 438static int __init noexec_setup(char *str)
439{ 439{
440 if (!strncmp(str, "on",2) && cpu_has_nx) { 440 if (!str || !strcmp(str, "on")) {
441 __supported_pte_mask |= _PAGE_NX; 441 if (cpu_has_nx) {
442 disable_nx = 0; 442 __supported_pte_mask |= _PAGE_NX;
443 } else if (!strncmp(str,"off",3)) { 443 disable_nx = 0;
444 }
445 } else if (!strcmp(str,"off")) {
444 disable_nx = 1; 446 disable_nx = 1;
445 __supported_pte_mask &= ~_PAGE_NX; 447 __supported_pte_mask &= ~_PAGE_NX;
446 } 448 } else
449 return -EINVAL;
450
451 return 0;
447} 452}
453early_param("noexec", noexec_setup);
448 454
449int nx_enabled = 0; 455int nx_enabled = 0;
450#ifdef CONFIG_X86_PAE 456#ifdef CONFIG_X86_PAE
@@ -552,18 +558,6 @@ static void __init test_wp_bit(void)
552 } 558 }
553} 559}
554 560
555static void __init set_max_mapnr_init(void)
556{
557#ifdef CONFIG_HIGHMEM
558 num_physpages = highend_pfn;
559#else
560 num_physpages = max_low_pfn;
561#endif
562#ifdef CONFIG_FLATMEM
563 max_mapnr = num_physpages;
564#endif
565}
566
567static struct kcore_list kcore_mem, kcore_vmalloc; 561static struct kcore_list kcore_mem, kcore_vmalloc;
568 562
569void __init mem_init(void) 563void __init mem_init(void)
@@ -590,14 +584,6 @@ void __init mem_init(void)
590 } 584 }
591#endif 585#endif
592 586
593 set_max_mapnr_init();
594
595#ifdef CONFIG_HIGHMEM
596 high_memory = (void *) __va(highstart_pfn * PAGE_SIZE - 1) + 1;
597#else
598 high_memory = (void *) __va(max_low_pfn * PAGE_SIZE - 1) + 1;
599#endif
600
601 /* this will put all low memory onto the freelists */ 587 /* this will put all low memory onto the freelists */
602 totalram_pages += free_all_bootmem(); 588 totalram_pages += free_all_bootmem();
603 589
@@ -629,6 +615,48 @@ void __init mem_init(void)
629 (unsigned long) (totalhigh_pages << (PAGE_SHIFT-10)) 615 (unsigned long) (totalhigh_pages << (PAGE_SHIFT-10))
630 ); 616 );
631 617
618#if 1 /* double-sanity-check paranoia */
619 printk("virtual kernel memory layout:\n"
620 " fixmap : 0x%08lx - 0x%08lx (%4ld kB)\n"
621#ifdef CONFIG_HIGHMEM
622 " pkmap : 0x%08lx - 0x%08lx (%4ld kB)\n"
623#endif
624 " vmalloc : 0x%08lx - 0x%08lx (%4ld MB)\n"
625 " lowmem : 0x%08lx - 0x%08lx (%4ld MB)\n"
626 " .init : 0x%08lx - 0x%08lx (%4ld kB)\n"
627 " .data : 0x%08lx - 0x%08lx (%4ld kB)\n"
628 " .text : 0x%08lx - 0x%08lx (%4ld kB)\n",
629 FIXADDR_START, FIXADDR_TOP,
630 (FIXADDR_TOP - FIXADDR_START) >> 10,
631
632#ifdef CONFIG_HIGHMEM
633 PKMAP_BASE, PKMAP_BASE+LAST_PKMAP*PAGE_SIZE,
634 (LAST_PKMAP*PAGE_SIZE) >> 10,
635#endif
636
637 VMALLOC_START, VMALLOC_END,
638 (VMALLOC_END - VMALLOC_START) >> 20,
639
640 (unsigned long)__va(0), (unsigned long)high_memory,
641 ((unsigned long)high_memory - (unsigned long)__va(0)) >> 20,
642
643 (unsigned long)&__init_begin, (unsigned long)&__init_end,
644 ((unsigned long)&__init_end - (unsigned long)&__init_begin) >> 10,
645
646 (unsigned long)&_etext, (unsigned long)&_edata,
647 ((unsigned long)&_edata - (unsigned long)&_etext) >> 10,
648
649 (unsigned long)&_text, (unsigned long)&_etext,
650 ((unsigned long)&_etext - (unsigned long)&_text) >> 10);
651
652#ifdef CONFIG_HIGHMEM
653 BUG_ON(PKMAP_BASE+LAST_PKMAP*PAGE_SIZE > FIXADDR_START);
654 BUG_ON(VMALLOC_END > PKMAP_BASE);
655#endif
656 BUG_ON(VMALLOC_START > VMALLOC_END);
657 BUG_ON((unsigned long)high_memory > VMALLOC_START);
658#endif /* double-sanity-check paranoia */
659
632#ifdef CONFIG_X86_PAE 660#ifdef CONFIG_X86_PAE
633 if (!cpu_has_pae) 661 if (!cpu_has_pae)
634 panic("cannot execute a PAE-enabled kernel on a PAE-less CPU!"); 662 panic("cannot execute a PAE-enabled kernel on a PAE-less CPU!");
@@ -657,7 +685,7 @@ void __init mem_init(void)
657int arch_add_memory(int nid, u64 start, u64 size) 685int arch_add_memory(int nid, u64 start, u64 size)
658{ 686{
659 struct pglist_data *pgdata = &contig_page_data; 687 struct pglist_data *pgdata = &contig_page_data;
660 struct zone *zone = pgdata->node_zones + MAX_NR_ZONES-1; 688 struct zone *zone = pgdata->node_zones + ZONE_HIGHMEM;
661 unsigned long start_pfn = start >> PAGE_SHIFT; 689 unsigned long start_pfn = start >> PAGE_SHIFT;
662 unsigned long nr_pages = size >> PAGE_SHIFT; 690 unsigned long nr_pages = size >> PAGE_SHIFT;
663 691
diff --git a/arch/i386/mm/pgtable.c b/arch/i386/mm/pgtable.c
index bd98768d8764..10126e3f8174 100644
--- a/arch/i386/mm/pgtable.c
+++ b/arch/i386/mm/pgtable.c
@@ -12,6 +12,7 @@
12#include <linux/slab.h> 12#include <linux/slab.h>
13#include <linux/pagemap.h> 13#include <linux/pagemap.h>
14#include <linux/spinlock.h> 14#include <linux/spinlock.h>
15#include <linux/module.h>
15 16
16#include <asm/system.h> 17#include <asm/system.h>
17#include <asm/pgtable.h> 18#include <asm/pgtable.h>
@@ -60,7 +61,9 @@ void show_mem(void)
60 printk(KERN_INFO "%lu pages writeback\n", 61 printk(KERN_INFO "%lu pages writeback\n",
61 global_page_state(NR_WRITEBACK)); 62 global_page_state(NR_WRITEBACK));
62 printk(KERN_INFO "%lu pages mapped\n", global_page_state(NR_FILE_MAPPED)); 63 printk(KERN_INFO "%lu pages mapped\n", global_page_state(NR_FILE_MAPPED));
63 printk(KERN_INFO "%lu pages slab\n", global_page_state(NR_SLAB)); 64 printk(KERN_INFO "%lu pages slab\n",
65 global_page_state(NR_SLAB_RECLAIMABLE) +
66 global_page_state(NR_SLAB_UNRECLAIMABLE));
64 printk(KERN_INFO "%lu pages pagetables\n", 67 printk(KERN_INFO "%lu pages pagetables\n",
65 global_page_state(NR_PAGETABLE)); 68 global_page_state(NR_PAGETABLE));
66} 69}
@@ -137,6 +140,12 @@ void set_pmd_pfn(unsigned long vaddr, unsigned long pfn, pgprot_t flags)
137 __flush_tlb_one(vaddr); 140 __flush_tlb_one(vaddr);
138} 141}
139 142
143static int fixmaps;
144#ifndef CONFIG_COMPAT_VDSO
145unsigned long __FIXADDR_TOP = 0xfffff000;
146EXPORT_SYMBOL(__FIXADDR_TOP);
147#endif
148
140void __set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t flags) 149void __set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t flags)
141{ 150{
142 unsigned long address = __fix_to_virt(idx); 151 unsigned long address = __fix_to_virt(idx);
@@ -146,6 +155,25 @@ void __set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t flags)
146 return; 155 return;
147 } 156 }
148 set_pte_pfn(address, phys >> PAGE_SHIFT, flags); 157 set_pte_pfn(address, phys >> PAGE_SHIFT, flags);
158 fixmaps++;
159}
160
161/**
162 * reserve_top_address - reserves a hole in the top of kernel address space
163 * @reserve - size of hole to reserve
164 *
165 * Can be used to relocate the fixmap area and poke a hole in the top
166 * of kernel address space to make room for a hypervisor.
167 */
168void reserve_top_address(unsigned long reserve)
169{
170 BUG_ON(fixmaps > 0);
171#ifdef CONFIG_COMPAT_VDSO
172 BUG_ON(reserve != 0);
173#else
174 __FIXADDR_TOP = -reserve - PAGE_SIZE;
175 __VMALLOC_RESERVE += reserve;
176#endif
149} 177}
150 178
151pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address) 179pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
diff --git a/arch/i386/oprofile/nmi_int.c b/arch/i386/oprofile/nmi_int.c
index 5f8dc8a21bd7..3700eef78743 100644
--- a/arch/i386/oprofile/nmi_int.c
+++ b/arch/i386/oprofile/nmi_int.c
@@ -17,14 +17,15 @@
17#include <asm/nmi.h> 17#include <asm/nmi.h>
18#include <asm/msr.h> 18#include <asm/msr.h>
19#include <asm/apic.h> 19#include <asm/apic.h>
20#include <asm/kdebug.h>
20 21
21#include "op_counter.h" 22#include "op_counter.h"
22#include "op_x86_model.h" 23#include "op_x86_model.h"
23 24
24static struct op_x86_model_spec const * model; 25static struct op_x86_model_spec const * model;
25static struct op_msrs cpu_msrs[NR_CPUS]; 26static struct op_msrs cpu_msrs[NR_CPUS];
26static unsigned long saved_lvtpc[NR_CPUS]; 27static unsigned long saved_lvtpc[NR_CPUS];
27 28
28static int nmi_start(void); 29static int nmi_start(void);
29static void nmi_stop(void); 30static void nmi_stop(void);
30 31
@@ -82,13 +83,24 @@ static void exit_driverfs(void)
82#define exit_driverfs() do { } while (0) 83#define exit_driverfs() do { } while (0)
83#endif /* CONFIG_PM */ 84#endif /* CONFIG_PM */
84 85
85 86static int profile_exceptions_notify(struct notifier_block *self,
86static int nmi_callback(struct pt_regs * regs, int cpu) 87 unsigned long val, void *data)
87{ 88{
88 return model->check_ctrs(regs, &cpu_msrs[cpu]); 89 struct die_args *args = (struct die_args *)data;
90 int ret = NOTIFY_DONE;
91 int cpu = smp_processor_id();
92
93 switch(val) {
94 case DIE_NMI:
95 if (model->check_ctrs(args->regs, &cpu_msrs[cpu]))
96 ret = NOTIFY_STOP;
97 break;
98 default:
99 break;
100 }
101 return ret;
89} 102}
90 103
91
92static void nmi_cpu_save_registers(struct op_msrs * msrs) 104static void nmi_cpu_save_registers(struct op_msrs * msrs)
93{ 105{
94 unsigned int const nr_ctrs = model->num_counters; 106 unsigned int const nr_ctrs = model->num_counters;
@@ -98,15 +110,19 @@ static void nmi_cpu_save_registers(struct op_msrs * msrs)
98 unsigned int i; 110 unsigned int i;
99 111
100 for (i = 0; i < nr_ctrs; ++i) { 112 for (i = 0; i < nr_ctrs; ++i) {
101 rdmsr(counters[i].addr, 113 if (counters[i].addr){
102 counters[i].saved.low, 114 rdmsr(counters[i].addr,
103 counters[i].saved.high); 115 counters[i].saved.low,
116 counters[i].saved.high);
117 }
104 } 118 }
105 119
106 for (i = 0; i < nr_ctrls; ++i) { 120 for (i = 0; i < nr_ctrls; ++i) {
107 rdmsr(controls[i].addr, 121 if (controls[i].addr){
108 controls[i].saved.low, 122 rdmsr(controls[i].addr,
109 controls[i].saved.high); 123 controls[i].saved.low,
124 controls[i].saved.high);
125 }
110 } 126 }
111} 127}
112 128
@@ -170,27 +186,29 @@ static void nmi_cpu_setup(void * dummy)
170 apic_write(APIC_LVTPC, APIC_DM_NMI); 186 apic_write(APIC_LVTPC, APIC_DM_NMI);
171} 187}
172 188
189static struct notifier_block profile_exceptions_nb = {
190 .notifier_call = profile_exceptions_notify,
191 .next = NULL,
192 .priority = 0
193};
173 194
174static int nmi_setup(void) 195static int nmi_setup(void)
175{ 196{
197 int err=0;
198
176 if (!allocate_msrs()) 199 if (!allocate_msrs())
177 return -ENOMEM; 200 return -ENOMEM;
178 201
179 /* We walk a thin line between law and rape here. 202 if ((err = register_die_notifier(&profile_exceptions_nb))){
180 * We need to be careful to install our NMI handler
181 * without actually triggering any NMIs as this will
182 * break the core code horrifically.
183 */
184 if (reserve_lapic_nmi() < 0) {
185 free_msrs(); 203 free_msrs();
186 return -EBUSY; 204 return err;
187 } 205 }
206
188 /* We need to serialize save and setup for HT because the subset 207 /* We need to serialize save and setup for HT because the subset
189 * of msrs are distinct for save and setup operations 208 * of msrs are distinct for save and setup operations
190 */ 209 */
191 on_each_cpu(nmi_save_registers, NULL, 0, 1); 210 on_each_cpu(nmi_save_registers, NULL, 0, 1);
192 on_each_cpu(nmi_cpu_setup, NULL, 0, 1); 211 on_each_cpu(nmi_cpu_setup, NULL, 0, 1);
193 set_nmi_callback(nmi_callback);
194 nmi_enabled = 1; 212 nmi_enabled = 1;
195 return 0; 213 return 0;
196} 214}
@@ -205,15 +223,19 @@ static void nmi_restore_registers(struct op_msrs * msrs)
205 unsigned int i; 223 unsigned int i;
206 224
207 for (i = 0; i < nr_ctrls; ++i) { 225 for (i = 0; i < nr_ctrls; ++i) {
208 wrmsr(controls[i].addr, 226 if (controls[i].addr){
209 controls[i].saved.low, 227 wrmsr(controls[i].addr,
210 controls[i].saved.high); 228 controls[i].saved.low,
229 controls[i].saved.high);
230 }
211 } 231 }
212 232
213 for (i = 0; i < nr_ctrs; ++i) { 233 for (i = 0; i < nr_ctrs; ++i) {
214 wrmsr(counters[i].addr, 234 if (counters[i].addr){
215 counters[i].saved.low, 235 wrmsr(counters[i].addr,
216 counters[i].saved.high); 236 counters[i].saved.low,
237 counters[i].saved.high);
238 }
217 } 239 }
218} 240}
219 241
@@ -234,6 +256,7 @@ static void nmi_cpu_shutdown(void * dummy)
234 apic_write(APIC_LVTPC, saved_lvtpc[cpu]); 256 apic_write(APIC_LVTPC, saved_lvtpc[cpu]);
235 apic_write(APIC_LVTERR, v); 257 apic_write(APIC_LVTERR, v);
236 nmi_restore_registers(msrs); 258 nmi_restore_registers(msrs);
259 model->shutdown(msrs);
237} 260}
238 261
239 262
@@ -241,8 +264,7 @@ static void nmi_shutdown(void)
241{ 264{
242 nmi_enabled = 0; 265 nmi_enabled = 0;
243 on_each_cpu(nmi_cpu_shutdown, NULL, 0, 1); 266 on_each_cpu(nmi_cpu_shutdown, NULL, 0, 1);
244 unset_nmi_callback(); 267 unregister_die_notifier(&profile_exceptions_nb);
245 release_lapic_nmi();
246 free_msrs(); 268 free_msrs();
247} 269}
248 270
@@ -284,6 +306,14 @@ static int nmi_create_files(struct super_block * sb, struct dentry * root)
284 struct dentry * dir; 306 struct dentry * dir;
285 char buf[4]; 307 char buf[4];
286 308
309 /* quick little hack to _not_ expose a counter if it is not
310 * available for use. This should protect userspace app.
311 * NOTE: assumes 1:1 mapping here (that counters are organized
312 * sequentially in their struct assignment).
313 */
314 if (unlikely(!avail_to_resrv_perfctr_nmi_bit(i)))
315 continue;
316
287 snprintf(buf, sizeof(buf), "%d", i); 317 snprintf(buf, sizeof(buf), "%d", i);
288 dir = oprofilefs_mkdir(sb, root, buf); 318 dir = oprofilefs_mkdir(sb, root, buf);
289 oprofilefs_create_ulong(sb, dir, "enabled", &counter_config[i].enabled); 319 oprofilefs_create_ulong(sb, dir, "enabled", &counter_config[i].enabled);
diff --git a/arch/i386/oprofile/nmi_timer_int.c b/arch/i386/oprofile/nmi_timer_int.c
index 930a1127bb30..abf0ba52a635 100644
--- a/arch/i386/oprofile/nmi_timer_int.c
+++ b/arch/i386/oprofile/nmi_timer_int.c
@@ -17,34 +17,49 @@
17#include <asm/nmi.h> 17#include <asm/nmi.h>
18#include <asm/apic.h> 18#include <asm/apic.h>
19#include <asm/ptrace.h> 19#include <asm/ptrace.h>
20#include <asm/kdebug.h>
20 21
21static int nmi_timer_callback(struct pt_regs * regs, int cpu) 22static int profile_timer_exceptions_notify(struct notifier_block *self,
23 unsigned long val, void *data)
22{ 24{
23 oprofile_add_sample(regs, 0); 25 struct die_args *args = (struct die_args *)data;
24 return 1; 26 int ret = NOTIFY_DONE;
27
28 switch(val) {
29 case DIE_NMI:
30 oprofile_add_sample(args->regs, 0);
31 ret = NOTIFY_STOP;
32 break;
33 default:
34 break;
35 }
36 return ret;
25} 37}
26 38
39static struct notifier_block profile_timer_exceptions_nb = {
40 .notifier_call = profile_timer_exceptions_notify,
41 .next = NULL,
42 .priority = 0
43};
44
27static int timer_start(void) 45static int timer_start(void)
28{ 46{
29 disable_timer_nmi_watchdog(); 47 if (register_die_notifier(&profile_timer_exceptions_nb))
30 set_nmi_callback(nmi_timer_callback); 48 return 1;
31 return 0; 49 return 0;
32} 50}
33 51
34 52
35static void timer_stop(void) 53static void timer_stop(void)
36{ 54{
37 enable_timer_nmi_watchdog(); 55 unregister_die_notifier(&profile_timer_exceptions_nb);
38 unset_nmi_callback();
39 synchronize_sched(); /* Allow already-started NMIs to complete. */ 56 synchronize_sched(); /* Allow already-started NMIs to complete. */
40} 57}
41 58
42 59
43int __init op_nmi_timer_init(struct oprofile_operations * ops) 60int __init op_nmi_timer_init(struct oprofile_operations * ops)
44{ 61{
45 extern int nmi_active; 62 if ((nmi_watchdog != NMI_IO_APIC) || (atomic_read(&nmi_active) <= 0))
46
47 if (nmi_active <= 0)
48 return -ENODEV; 63 return -ENODEV;
49 64
50 ops->start = timer_start; 65 ops->start = timer_start;
diff --git a/arch/i386/oprofile/op_model_athlon.c b/arch/i386/oprofile/op_model_athlon.c
index 693bdea4a52b..3057a19e4641 100644
--- a/arch/i386/oprofile/op_model_athlon.c
+++ b/arch/i386/oprofile/op_model_athlon.c
@@ -21,10 +21,12 @@
21#define NUM_COUNTERS 4 21#define NUM_COUNTERS 4
22#define NUM_CONTROLS 4 22#define NUM_CONTROLS 4
23 23
24#define CTR_IS_RESERVED(msrs,c) (msrs->counters[(c)].addr ? 1 : 0)
24#define CTR_READ(l,h,msrs,c) do {rdmsr(msrs->counters[(c)].addr, (l), (h));} while (0) 25#define CTR_READ(l,h,msrs,c) do {rdmsr(msrs->counters[(c)].addr, (l), (h));} while (0)
25#define CTR_WRITE(l,msrs,c) do {wrmsr(msrs->counters[(c)].addr, -(unsigned int)(l), -1);} while (0) 26#define CTR_WRITE(l,msrs,c) do {wrmsr(msrs->counters[(c)].addr, -(unsigned int)(l), -1);} while (0)
26#define CTR_OVERFLOWED(n) (!((n) & (1U<<31))) 27#define CTR_OVERFLOWED(n) (!((n) & (1U<<31)))
27 28
29#define CTRL_IS_RESERVED(msrs,c) (msrs->controls[(c)].addr ? 1 : 0)
28#define CTRL_READ(l,h,msrs,c) do {rdmsr(msrs->controls[(c)].addr, (l), (h));} while (0) 30#define CTRL_READ(l,h,msrs,c) do {rdmsr(msrs->controls[(c)].addr, (l), (h));} while (0)
29#define CTRL_WRITE(l,h,msrs,c) do {wrmsr(msrs->controls[(c)].addr, (l), (h));} while (0) 31#define CTRL_WRITE(l,h,msrs,c) do {wrmsr(msrs->controls[(c)].addr, (l), (h));} while (0)
30#define CTRL_SET_ACTIVE(n) (n |= (1<<22)) 32#define CTRL_SET_ACTIVE(n) (n |= (1<<22))
@@ -40,15 +42,21 @@ static unsigned long reset_value[NUM_COUNTERS];
40 42
41static void athlon_fill_in_addresses(struct op_msrs * const msrs) 43static void athlon_fill_in_addresses(struct op_msrs * const msrs)
42{ 44{
43 msrs->counters[0].addr = MSR_K7_PERFCTR0; 45 int i;
44 msrs->counters[1].addr = MSR_K7_PERFCTR1; 46
45 msrs->counters[2].addr = MSR_K7_PERFCTR2; 47 for (i=0; i < NUM_COUNTERS; i++) {
46 msrs->counters[3].addr = MSR_K7_PERFCTR3; 48 if (reserve_perfctr_nmi(MSR_K7_PERFCTR0 + i))
47 49 msrs->counters[i].addr = MSR_K7_PERFCTR0 + i;
48 msrs->controls[0].addr = MSR_K7_EVNTSEL0; 50 else
49 msrs->controls[1].addr = MSR_K7_EVNTSEL1; 51 msrs->counters[i].addr = 0;
50 msrs->controls[2].addr = MSR_K7_EVNTSEL2; 52 }
51 msrs->controls[3].addr = MSR_K7_EVNTSEL3; 53
54 for (i=0; i < NUM_CONTROLS; i++) {
55 if (reserve_evntsel_nmi(MSR_K7_EVNTSEL0 + i))
56 msrs->controls[i].addr = MSR_K7_EVNTSEL0 + i;
57 else
58 msrs->controls[i].addr = 0;
59 }
52} 60}
53 61
54 62
@@ -59,19 +67,23 @@ static void athlon_setup_ctrs(struct op_msrs const * const msrs)
59 67
60 /* clear all counters */ 68 /* clear all counters */
61 for (i = 0 ; i < NUM_CONTROLS; ++i) { 69 for (i = 0 ; i < NUM_CONTROLS; ++i) {
70 if (unlikely(!CTRL_IS_RESERVED(msrs,i)))
71 continue;
62 CTRL_READ(low, high, msrs, i); 72 CTRL_READ(low, high, msrs, i);
63 CTRL_CLEAR(low); 73 CTRL_CLEAR(low);
64 CTRL_WRITE(low, high, msrs, i); 74 CTRL_WRITE(low, high, msrs, i);
65 } 75 }
66 76
67 /* avoid a false detection of ctr overflows in NMI handler */ 77 /* avoid a false detection of ctr overflows in NMI handler */
68 for (i = 0; i < NUM_COUNTERS; ++i) { 78 for (i = 0; i < NUM_COUNTERS; ++i) {
79 if (unlikely(!CTR_IS_RESERVED(msrs,i)))
80 continue;
69 CTR_WRITE(1, msrs, i); 81 CTR_WRITE(1, msrs, i);
70 } 82 }
71 83
72 /* enable active counters */ 84 /* enable active counters */
73 for (i = 0; i < NUM_COUNTERS; ++i) { 85 for (i = 0; i < NUM_COUNTERS; ++i) {
74 if (counter_config[i].enabled) { 86 if ((counter_config[i].enabled) && (CTR_IS_RESERVED(msrs,i))) {
75 reset_value[i] = counter_config[i].count; 87 reset_value[i] = counter_config[i].count;
76 88
77 CTR_WRITE(counter_config[i].count, msrs, i); 89 CTR_WRITE(counter_config[i].count, msrs, i);
@@ -98,6 +110,8 @@ static int athlon_check_ctrs(struct pt_regs * const regs,
98 int i; 110 int i;
99 111
100 for (i = 0 ; i < NUM_COUNTERS; ++i) { 112 for (i = 0 ; i < NUM_COUNTERS; ++i) {
113 if (!reset_value[i])
114 continue;
101 CTR_READ(low, high, msrs, i); 115 CTR_READ(low, high, msrs, i);
102 if (CTR_OVERFLOWED(low)) { 116 if (CTR_OVERFLOWED(low)) {
103 oprofile_add_sample(regs, i); 117 oprofile_add_sample(regs, i);
@@ -132,12 +146,27 @@ static void athlon_stop(struct op_msrs const * const msrs)
132 /* Subtle: stop on all counters to avoid race with 146 /* Subtle: stop on all counters to avoid race with
133 * setting our pm callback */ 147 * setting our pm callback */
134 for (i = 0 ; i < NUM_COUNTERS ; ++i) { 148 for (i = 0 ; i < NUM_COUNTERS ; ++i) {
149 if (!reset_value[i])
150 continue;
135 CTRL_READ(low, high, msrs, i); 151 CTRL_READ(low, high, msrs, i);
136 CTRL_SET_INACTIVE(low); 152 CTRL_SET_INACTIVE(low);
137 CTRL_WRITE(low, high, msrs, i); 153 CTRL_WRITE(low, high, msrs, i);
138 } 154 }
139} 155}
140 156
157static void athlon_shutdown(struct op_msrs const * const msrs)
158{
159 int i;
160
161 for (i = 0 ; i < NUM_COUNTERS ; ++i) {
162 if (CTR_IS_RESERVED(msrs,i))
163 release_perfctr_nmi(MSR_K7_PERFCTR0 + i);
164 }
165 for (i = 0 ; i < NUM_CONTROLS ; ++i) {
166 if (CTRL_IS_RESERVED(msrs,i))
167 release_evntsel_nmi(MSR_K7_EVNTSEL0 + i);
168 }
169}
141 170
142struct op_x86_model_spec const op_athlon_spec = { 171struct op_x86_model_spec const op_athlon_spec = {
143 .num_counters = NUM_COUNTERS, 172 .num_counters = NUM_COUNTERS,
@@ -146,5 +175,6 @@ struct op_x86_model_spec const op_athlon_spec = {
146 .setup_ctrs = &athlon_setup_ctrs, 175 .setup_ctrs = &athlon_setup_ctrs,
147 .check_ctrs = &athlon_check_ctrs, 176 .check_ctrs = &athlon_check_ctrs,
148 .start = &athlon_start, 177 .start = &athlon_start,
149 .stop = &athlon_stop 178 .stop = &athlon_stop,
179 .shutdown = &athlon_shutdown
150}; 180};
diff --git a/arch/i386/oprofile/op_model_p4.c b/arch/i386/oprofile/op_model_p4.c
index 7c61d357b82b..47925927b12f 100644
--- a/arch/i386/oprofile/op_model_p4.c
+++ b/arch/i386/oprofile/op_model_p4.c
@@ -32,7 +32,7 @@
32#define NUM_CONTROLS_HT2 (NUM_ESCRS_HT2 + NUM_CCCRS_HT2) 32#define NUM_CONTROLS_HT2 (NUM_ESCRS_HT2 + NUM_CCCRS_HT2)
33 33
34static unsigned int num_counters = NUM_COUNTERS_NON_HT; 34static unsigned int num_counters = NUM_COUNTERS_NON_HT;
35 35static unsigned int num_controls = NUM_CONTROLS_NON_HT;
36 36
37/* this has to be checked dynamically since the 37/* this has to be checked dynamically since the
38 hyper-threadedness of a chip is discovered at 38 hyper-threadedness of a chip is discovered at
@@ -40,8 +40,10 @@ static unsigned int num_counters = NUM_COUNTERS_NON_HT;
40static inline void setup_num_counters(void) 40static inline void setup_num_counters(void)
41{ 41{
42#ifdef CONFIG_SMP 42#ifdef CONFIG_SMP
43 if (smp_num_siblings == 2) 43 if (smp_num_siblings == 2){
44 num_counters = NUM_COUNTERS_HT2; 44 num_counters = NUM_COUNTERS_HT2;
45 num_controls = NUM_CONTROLS_HT2;
46 }
45#endif 47#endif
46} 48}
47 49
@@ -97,15 +99,6 @@ static struct p4_counter_binding p4_counters [NUM_COUNTERS_NON_HT] = {
97 99
98#define NUM_UNUSED_CCCRS NUM_CCCRS_NON_HT - NUM_COUNTERS_NON_HT 100#define NUM_UNUSED_CCCRS NUM_CCCRS_NON_HT - NUM_COUNTERS_NON_HT
99 101
100/* All cccr we don't use. */
101static int p4_unused_cccr[NUM_UNUSED_CCCRS] = {
102 MSR_P4_BPU_CCCR1, MSR_P4_BPU_CCCR3,
103 MSR_P4_MS_CCCR1, MSR_P4_MS_CCCR3,
104 MSR_P4_FLAME_CCCR1, MSR_P4_FLAME_CCCR3,
105 MSR_P4_IQ_CCCR0, MSR_P4_IQ_CCCR1,
106 MSR_P4_IQ_CCCR2, MSR_P4_IQ_CCCR3
107};
108
109/* p4 event codes in libop/op_event.h are indices into this table. */ 102/* p4 event codes in libop/op_event.h are indices into this table. */
110 103
111static struct p4_event_binding p4_events[NUM_EVENTS] = { 104static struct p4_event_binding p4_events[NUM_EVENTS] = {
@@ -372,6 +365,8 @@ static struct p4_event_binding p4_events[NUM_EVENTS] = {
372#define CCCR_OVF_P(cccr) ((cccr) & (1U<<31)) 365#define CCCR_OVF_P(cccr) ((cccr) & (1U<<31))
373#define CCCR_CLEAR_OVF(cccr) ((cccr) &= (~(1U<<31))) 366#define CCCR_CLEAR_OVF(cccr) ((cccr) &= (~(1U<<31)))
374 367
368#define CTRL_IS_RESERVED(msrs,c) (msrs->controls[(c)].addr ? 1 : 0)
369#define CTR_IS_RESERVED(msrs,c) (msrs->counters[(c)].addr ? 1 : 0)
375#define CTR_READ(l,h,i) do {rdmsr(p4_counters[(i)].counter_address, (l), (h));} while (0) 370#define CTR_READ(l,h,i) do {rdmsr(p4_counters[(i)].counter_address, (l), (h));} while (0)
376#define CTR_WRITE(l,i) do {wrmsr(p4_counters[(i)].counter_address, -(u32)(l), -1);} while (0) 371#define CTR_WRITE(l,i) do {wrmsr(p4_counters[(i)].counter_address, -(u32)(l), -1);} while (0)
377#define CTR_OVERFLOW_P(ctr) (!((ctr) & 0x80000000)) 372#define CTR_OVERFLOW_P(ctr) (!((ctr) & 0x80000000))
@@ -401,29 +396,34 @@ static unsigned long reset_value[NUM_COUNTERS_NON_HT];
401static void p4_fill_in_addresses(struct op_msrs * const msrs) 396static void p4_fill_in_addresses(struct op_msrs * const msrs)
402{ 397{
403 unsigned int i; 398 unsigned int i;
404 unsigned int addr, stag; 399 unsigned int addr, cccraddr, stag;
405 400
406 setup_num_counters(); 401 setup_num_counters();
407 stag = get_stagger(); 402 stag = get_stagger();
408 403
409 /* the counter registers we pay attention to */ 404 /* initialize some registers */
410 for (i = 0; i < num_counters; ++i) { 405 for (i = 0; i < num_counters; ++i) {
411 msrs->counters[i].addr = 406 msrs->counters[i].addr = 0;
412 p4_counters[VIRT_CTR(stag, i)].counter_address;
413 } 407 }
414 408 for (i = 0; i < num_controls; ++i) {
415 /* FIXME: bad feeling, we don't save the 10 counters we don't use. */ 409 msrs->controls[i].addr = 0;
416
417 /* 18 CCCR registers */
418 for (i = 0, addr = MSR_P4_BPU_CCCR0 + stag;
419 addr <= MSR_P4_IQ_CCCR5; ++i, addr += addr_increment()) {
420 msrs->controls[i].addr = addr;
421 } 410 }
422 411
412 /* the counter & cccr registers we pay attention to */
413 for (i = 0; i < num_counters; ++i) {
414 addr = p4_counters[VIRT_CTR(stag, i)].counter_address;
415 cccraddr = p4_counters[VIRT_CTR(stag, i)].cccr_address;
416 if (reserve_perfctr_nmi(addr)){
417 msrs->counters[i].addr = addr;
418 msrs->controls[i].addr = cccraddr;
419 }
420 }
421
423 /* 43 ESCR registers in three or four discontiguous group */ 422 /* 43 ESCR registers in three or four discontiguous group */
424 for (addr = MSR_P4_BSU_ESCR0 + stag; 423 for (addr = MSR_P4_BSU_ESCR0 + stag;
425 addr < MSR_P4_IQ_ESCR0; ++i, addr += addr_increment()) { 424 addr < MSR_P4_IQ_ESCR0; ++i, addr += addr_increment()) {
426 msrs->controls[i].addr = addr; 425 if (reserve_evntsel_nmi(addr))
426 msrs->controls[i].addr = addr;
427 } 427 }
428 428
429 /* no IQ_ESCR0/1 on some models, we save a seconde time BSU_ESCR0/1 429 /* no IQ_ESCR0/1 on some models, we save a seconde time BSU_ESCR0/1
@@ -431,47 +431,57 @@ static void p4_fill_in_addresses(struct op_msrs * const msrs)
431 if (boot_cpu_data.x86_model >= 0x3) { 431 if (boot_cpu_data.x86_model >= 0x3) {
432 for (addr = MSR_P4_BSU_ESCR0 + stag; 432 for (addr = MSR_P4_BSU_ESCR0 + stag;
433 addr <= MSR_P4_BSU_ESCR1; ++i, addr += addr_increment()) { 433 addr <= MSR_P4_BSU_ESCR1; ++i, addr += addr_increment()) {
434 msrs->controls[i].addr = addr; 434 if (reserve_evntsel_nmi(addr))
435 msrs->controls[i].addr = addr;
435 } 436 }
436 } else { 437 } else {
437 for (addr = MSR_P4_IQ_ESCR0 + stag; 438 for (addr = MSR_P4_IQ_ESCR0 + stag;
438 addr <= MSR_P4_IQ_ESCR1; ++i, addr += addr_increment()) { 439 addr <= MSR_P4_IQ_ESCR1; ++i, addr += addr_increment()) {
439 msrs->controls[i].addr = addr; 440 if (reserve_evntsel_nmi(addr))
441 msrs->controls[i].addr = addr;
440 } 442 }
441 } 443 }
442 444
443 for (addr = MSR_P4_RAT_ESCR0 + stag; 445 for (addr = MSR_P4_RAT_ESCR0 + stag;
444 addr <= MSR_P4_SSU_ESCR0; ++i, addr += addr_increment()) { 446 addr <= MSR_P4_SSU_ESCR0; ++i, addr += addr_increment()) {
445 msrs->controls[i].addr = addr; 447 if (reserve_evntsel_nmi(addr))
448 msrs->controls[i].addr = addr;
446 } 449 }
447 450
448 for (addr = MSR_P4_MS_ESCR0 + stag; 451 for (addr = MSR_P4_MS_ESCR0 + stag;
449 addr <= MSR_P4_TC_ESCR1; ++i, addr += addr_increment()) { 452 addr <= MSR_P4_TC_ESCR1; ++i, addr += addr_increment()) {
450 msrs->controls[i].addr = addr; 453 if (reserve_evntsel_nmi(addr))
454 msrs->controls[i].addr = addr;
451 } 455 }
452 456
453 for (addr = MSR_P4_IX_ESCR0 + stag; 457 for (addr = MSR_P4_IX_ESCR0 + stag;
454 addr <= MSR_P4_CRU_ESCR3; ++i, addr += addr_increment()) { 458 addr <= MSR_P4_CRU_ESCR3; ++i, addr += addr_increment()) {
455 msrs->controls[i].addr = addr; 459 if (reserve_evntsel_nmi(addr))
460 msrs->controls[i].addr = addr;
456 } 461 }
457 462
458 /* there are 2 remaining non-contiguously located ESCRs */ 463 /* there are 2 remaining non-contiguously located ESCRs */
459 464
460 if (num_counters == NUM_COUNTERS_NON_HT) { 465 if (num_counters == NUM_COUNTERS_NON_HT) {
461 /* standard non-HT CPUs handle both remaining ESCRs*/ 466 /* standard non-HT CPUs handle both remaining ESCRs*/
462 msrs->controls[i++].addr = MSR_P4_CRU_ESCR5; 467 if (reserve_evntsel_nmi(MSR_P4_CRU_ESCR5))
463 msrs->controls[i++].addr = MSR_P4_CRU_ESCR4; 468 msrs->controls[i++].addr = MSR_P4_CRU_ESCR5;
469 if (reserve_evntsel_nmi(MSR_P4_CRU_ESCR4))
470 msrs->controls[i++].addr = MSR_P4_CRU_ESCR4;
464 471
465 } else if (stag == 0) { 472 } else if (stag == 0) {
466 /* HT CPUs give the first remainder to the even thread, as 473 /* HT CPUs give the first remainder to the even thread, as
467 the 32nd control register */ 474 the 32nd control register */
468 msrs->controls[i++].addr = MSR_P4_CRU_ESCR4; 475 if (reserve_evntsel_nmi(MSR_P4_CRU_ESCR4))
476 msrs->controls[i++].addr = MSR_P4_CRU_ESCR4;
469 477
470 } else { 478 } else {
471 /* and two copies of the second to the odd thread, 479 /* and two copies of the second to the odd thread,
472 for the 22st and 23nd control registers */ 480 for the 22st and 23nd control registers */
473 msrs->controls[i++].addr = MSR_P4_CRU_ESCR5; 481 if (reserve_evntsel_nmi(MSR_P4_CRU_ESCR5)) {
474 msrs->controls[i++].addr = MSR_P4_CRU_ESCR5; 482 msrs->controls[i++].addr = MSR_P4_CRU_ESCR5;
483 msrs->controls[i++].addr = MSR_P4_CRU_ESCR5;
484 }
475 } 485 }
476} 486}
477 487
@@ -544,7 +554,6 @@ static void p4_setup_ctrs(struct op_msrs const * const msrs)
544{ 554{
545 unsigned int i; 555 unsigned int i;
546 unsigned int low, high; 556 unsigned int low, high;
547 unsigned int addr;
548 unsigned int stag; 557 unsigned int stag;
549 558
550 stag = get_stagger(); 559 stag = get_stagger();
@@ -557,59 +566,24 @@ static void p4_setup_ctrs(struct op_msrs const * const msrs)
557 566
558 /* clear the cccrs we will use */ 567 /* clear the cccrs we will use */
559 for (i = 0 ; i < num_counters ; i++) { 568 for (i = 0 ; i < num_counters ; i++) {
569 if (unlikely(!CTRL_IS_RESERVED(msrs,i)))
570 continue;
560 rdmsr(p4_counters[VIRT_CTR(stag, i)].cccr_address, low, high); 571 rdmsr(p4_counters[VIRT_CTR(stag, i)].cccr_address, low, high);
561 CCCR_CLEAR(low); 572 CCCR_CLEAR(low);
562 CCCR_SET_REQUIRED_BITS(low); 573 CCCR_SET_REQUIRED_BITS(low);
563 wrmsr(p4_counters[VIRT_CTR(stag, i)].cccr_address, low, high); 574 wrmsr(p4_counters[VIRT_CTR(stag, i)].cccr_address, low, high);
564 } 575 }
565 576
566 /* clear cccrs outside our concern */
567 for (i = stag ; i < NUM_UNUSED_CCCRS ; i += addr_increment()) {
568 rdmsr(p4_unused_cccr[i], low, high);
569 CCCR_CLEAR(low);
570 CCCR_SET_REQUIRED_BITS(low);
571 wrmsr(p4_unused_cccr[i], low, high);
572 }
573
574 /* clear all escrs (including those outside our concern) */ 577 /* clear all escrs (including those outside our concern) */
575 for (addr = MSR_P4_BSU_ESCR0 + stag; 578 for (i = num_counters; i < num_controls; i++) {
576 addr < MSR_P4_IQ_ESCR0; addr += addr_increment()) { 579 if (unlikely(!CTRL_IS_RESERVED(msrs,i)))
577 wrmsr(addr, 0, 0); 580 continue;
578 } 581 wrmsr(msrs->controls[i].addr, 0, 0);
579
580 /* On older models clear also MSR_P4_IQ_ESCR0/1 */
581 if (boot_cpu_data.x86_model < 0x3) {
582 wrmsr(MSR_P4_IQ_ESCR0, 0, 0);
583 wrmsr(MSR_P4_IQ_ESCR1, 0, 0);
584 }
585
586 for (addr = MSR_P4_RAT_ESCR0 + stag;
587 addr <= MSR_P4_SSU_ESCR0; ++i, addr += addr_increment()) {
588 wrmsr(addr, 0, 0);
589 }
590
591 for (addr = MSR_P4_MS_ESCR0 + stag;
592 addr <= MSR_P4_TC_ESCR1; addr += addr_increment()){
593 wrmsr(addr, 0, 0);
594 }
595
596 for (addr = MSR_P4_IX_ESCR0 + stag;
597 addr <= MSR_P4_CRU_ESCR3; addr += addr_increment()){
598 wrmsr(addr, 0, 0);
599 } 582 }
600 583
601 if (num_counters == NUM_COUNTERS_NON_HT) {
602 wrmsr(MSR_P4_CRU_ESCR4, 0, 0);
603 wrmsr(MSR_P4_CRU_ESCR5, 0, 0);
604 } else if (stag == 0) {
605 wrmsr(MSR_P4_CRU_ESCR4, 0, 0);
606 } else {
607 wrmsr(MSR_P4_CRU_ESCR5, 0, 0);
608 }
609
610 /* setup all counters */ 584 /* setup all counters */
611 for (i = 0 ; i < num_counters ; ++i) { 585 for (i = 0 ; i < num_counters ; ++i) {
612 if (counter_config[i].enabled) { 586 if ((counter_config[i].enabled) && (CTRL_IS_RESERVED(msrs,i))) {
613 reset_value[i] = counter_config[i].count; 587 reset_value[i] = counter_config[i].count;
614 pmc_setup_one_p4_counter(i); 588 pmc_setup_one_p4_counter(i);
615 CTR_WRITE(counter_config[i].count, VIRT_CTR(stag, i)); 589 CTR_WRITE(counter_config[i].count, VIRT_CTR(stag, i));
@@ -696,12 +670,32 @@ static void p4_stop(struct op_msrs const * const msrs)
696 stag = get_stagger(); 670 stag = get_stagger();
697 671
698 for (i = 0; i < num_counters; ++i) { 672 for (i = 0; i < num_counters; ++i) {
673 if (!reset_value[i])
674 continue;
699 CCCR_READ(low, high, VIRT_CTR(stag, i)); 675 CCCR_READ(low, high, VIRT_CTR(stag, i));
700 CCCR_SET_DISABLE(low); 676 CCCR_SET_DISABLE(low);
701 CCCR_WRITE(low, high, VIRT_CTR(stag, i)); 677 CCCR_WRITE(low, high, VIRT_CTR(stag, i));
702 } 678 }
703} 679}
704 680
681static void p4_shutdown(struct op_msrs const * const msrs)
682{
683 int i;
684
685 for (i = 0 ; i < num_counters ; ++i) {
686 if (CTR_IS_RESERVED(msrs,i))
687 release_perfctr_nmi(msrs->counters[i].addr);
688 }
689 /* some of the control registers are specially reserved in
690 * conjunction with the counter registers (hence the starting offset).
691 * This saves a few bits.
692 */
693 for (i = num_counters ; i < num_controls ; ++i) {
694 if (CTRL_IS_RESERVED(msrs,i))
695 release_evntsel_nmi(msrs->controls[i].addr);
696 }
697}
698
705 699
706#ifdef CONFIG_SMP 700#ifdef CONFIG_SMP
707struct op_x86_model_spec const op_p4_ht2_spec = { 701struct op_x86_model_spec const op_p4_ht2_spec = {
@@ -711,7 +705,8 @@ struct op_x86_model_spec const op_p4_ht2_spec = {
711 .setup_ctrs = &p4_setup_ctrs, 705 .setup_ctrs = &p4_setup_ctrs,
712 .check_ctrs = &p4_check_ctrs, 706 .check_ctrs = &p4_check_ctrs,
713 .start = &p4_start, 707 .start = &p4_start,
714 .stop = &p4_stop 708 .stop = &p4_stop,
709 .shutdown = &p4_shutdown
715}; 710};
716#endif 711#endif
717 712
@@ -722,5 +717,6 @@ struct op_x86_model_spec const op_p4_spec = {
722 .setup_ctrs = &p4_setup_ctrs, 717 .setup_ctrs = &p4_setup_ctrs,
723 .check_ctrs = &p4_check_ctrs, 718 .check_ctrs = &p4_check_ctrs,
724 .start = &p4_start, 719 .start = &p4_start,
725 .stop = &p4_stop 720 .stop = &p4_stop,
721 .shutdown = &p4_shutdown
726}; 722};
diff --git a/arch/i386/oprofile/op_model_ppro.c b/arch/i386/oprofile/op_model_ppro.c
index 5c3ab4b027ad..f88e05ba8eb3 100644
--- a/arch/i386/oprofile/op_model_ppro.c
+++ b/arch/i386/oprofile/op_model_ppro.c
@@ -22,10 +22,12 @@
22#define NUM_COUNTERS 2 22#define NUM_COUNTERS 2
23#define NUM_CONTROLS 2 23#define NUM_CONTROLS 2
24 24
25#define CTR_IS_RESERVED(msrs,c) (msrs->counters[(c)].addr ? 1 : 0)
25#define CTR_READ(l,h,msrs,c) do {rdmsr(msrs->counters[(c)].addr, (l), (h));} while (0) 26#define CTR_READ(l,h,msrs,c) do {rdmsr(msrs->counters[(c)].addr, (l), (h));} while (0)
26#define CTR_WRITE(l,msrs,c) do {wrmsr(msrs->counters[(c)].addr, -(u32)(l), -1);} while (0) 27#define CTR_WRITE(l,msrs,c) do {wrmsr(msrs->counters[(c)].addr, -(u32)(l), -1);} while (0)
27#define CTR_OVERFLOWED(n) (!((n) & (1U<<31))) 28#define CTR_OVERFLOWED(n) (!((n) & (1U<<31)))
28 29
30#define CTRL_IS_RESERVED(msrs,c) (msrs->controls[(c)].addr ? 1 : 0)
29#define CTRL_READ(l,h,msrs,c) do {rdmsr((msrs->controls[(c)].addr), (l), (h));} while (0) 31#define CTRL_READ(l,h,msrs,c) do {rdmsr((msrs->controls[(c)].addr), (l), (h));} while (0)
30#define CTRL_WRITE(l,h,msrs,c) do {wrmsr((msrs->controls[(c)].addr), (l), (h));} while (0) 32#define CTRL_WRITE(l,h,msrs,c) do {wrmsr((msrs->controls[(c)].addr), (l), (h));} while (0)
31#define CTRL_SET_ACTIVE(n) (n |= (1<<22)) 33#define CTRL_SET_ACTIVE(n) (n |= (1<<22))
@@ -41,11 +43,21 @@ static unsigned long reset_value[NUM_COUNTERS];
41 43
42static void ppro_fill_in_addresses(struct op_msrs * const msrs) 44static void ppro_fill_in_addresses(struct op_msrs * const msrs)
43{ 45{
44 msrs->counters[0].addr = MSR_P6_PERFCTR0; 46 int i;
45 msrs->counters[1].addr = MSR_P6_PERFCTR1; 47
48 for (i=0; i < NUM_COUNTERS; i++) {
49 if (reserve_perfctr_nmi(MSR_P6_PERFCTR0 + i))
50 msrs->counters[i].addr = MSR_P6_PERFCTR0 + i;
51 else
52 msrs->counters[i].addr = 0;
53 }
46 54
47 msrs->controls[0].addr = MSR_P6_EVNTSEL0; 55 for (i=0; i < NUM_CONTROLS; i++) {
48 msrs->controls[1].addr = MSR_P6_EVNTSEL1; 56 if (reserve_evntsel_nmi(MSR_P6_EVNTSEL0 + i))
57 msrs->controls[i].addr = MSR_P6_EVNTSEL0 + i;
58 else
59 msrs->controls[i].addr = 0;
60 }
49} 61}
50 62
51 63
@@ -56,6 +68,8 @@ static void ppro_setup_ctrs(struct op_msrs const * const msrs)
56 68
57 /* clear all counters */ 69 /* clear all counters */
58 for (i = 0 ; i < NUM_CONTROLS; ++i) { 70 for (i = 0 ; i < NUM_CONTROLS; ++i) {
71 if (unlikely(!CTRL_IS_RESERVED(msrs,i)))
72 continue;
59 CTRL_READ(low, high, msrs, i); 73 CTRL_READ(low, high, msrs, i);
60 CTRL_CLEAR(low); 74 CTRL_CLEAR(low);
61 CTRL_WRITE(low, high, msrs, i); 75 CTRL_WRITE(low, high, msrs, i);
@@ -63,12 +77,14 @@ static void ppro_setup_ctrs(struct op_msrs const * const msrs)
63 77
64 /* avoid a false detection of ctr overflows in NMI handler */ 78 /* avoid a false detection of ctr overflows in NMI handler */
65 for (i = 0; i < NUM_COUNTERS; ++i) { 79 for (i = 0; i < NUM_COUNTERS; ++i) {
80 if (unlikely(!CTR_IS_RESERVED(msrs,i)))
81 continue;
66 CTR_WRITE(1, msrs, i); 82 CTR_WRITE(1, msrs, i);
67 } 83 }
68 84
69 /* enable active counters */ 85 /* enable active counters */
70 for (i = 0; i < NUM_COUNTERS; ++i) { 86 for (i = 0; i < NUM_COUNTERS; ++i) {
71 if (counter_config[i].enabled) { 87 if ((counter_config[i].enabled) && (CTR_IS_RESERVED(msrs,i))) {
72 reset_value[i] = counter_config[i].count; 88 reset_value[i] = counter_config[i].count;
73 89
74 CTR_WRITE(counter_config[i].count, msrs, i); 90 CTR_WRITE(counter_config[i].count, msrs, i);
@@ -81,6 +97,8 @@ static void ppro_setup_ctrs(struct op_msrs const * const msrs)
81 CTRL_SET_UM(low, counter_config[i].unit_mask); 97 CTRL_SET_UM(low, counter_config[i].unit_mask);
82 CTRL_SET_EVENT(low, counter_config[i].event); 98 CTRL_SET_EVENT(low, counter_config[i].event);
83 CTRL_WRITE(low, high, msrs, i); 99 CTRL_WRITE(low, high, msrs, i);
100 } else {
101 reset_value[i] = 0;
84 } 102 }
85 } 103 }
86} 104}
@@ -93,6 +111,8 @@ static int ppro_check_ctrs(struct pt_regs * const regs,
93 int i; 111 int i;
94 112
95 for (i = 0 ; i < NUM_COUNTERS; ++i) { 113 for (i = 0 ; i < NUM_COUNTERS; ++i) {
114 if (!reset_value[i])
115 continue;
96 CTR_READ(low, high, msrs, i); 116 CTR_READ(low, high, msrs, i);
97 if (CTR_OVERFLOWED(low)) { 117 if (CTR_OVERFLOWED(low)) {
98 oprofile_add_sample(regs, i); 118 oprofile_add_sample(regs, i);
@@ -118,18 +138,38 @@ static int ppro_check_ctrs(struct pt_regs * const regs,
118static void ppro_start(struct op_msrs const * const msrs) 138static void ppro_start(struct op_msrs const * const msrs)
119{ 139{
120 unsigned int low,high; 140 unsigned int low,high;
121 CTRL_READ(low, high, msrs, 0); 141
122 CTRL_SET_ACTIVE(low); 142 if (reset_value[0]) {
123 CTRL_WRITE(low, high, msrs, 0); 143 CTRL_READ(low, high, msrs, 0);
144 CTRL_SET_ACTIVE(low);
145 CTRL_WRITE(low, high, msrs, 0);
146 }
124} 147}
125 148
126 149
127static void ppro_stop(struct op_msrs const * const msrs) 150static void ppro_stop(struct op_msrs const * const msrs)
128{ 151{
129 unsigned int low,high; 152 unsigned int low,high;
130 CTRL_READ(low, high, msrs, 0); 153
131 CTRL_SET_INACTIVE(low); 154 if (reset_value[0]) {
132 CTRL_WRITE(low, high, msrs, 0); 155 CTRL_READ(low, high, msrs, 0);
156 CTRL_SET_INACTIVE(low);
157 CTRL_WRITE(low, high, msrs, 0);
158 }
159}
160
161static void ppro_shutdown(struct op_msrs const * const msrs)
162{
163 int i;
164
165 for (i = 0 ; i < NUM_COUNTERS ; ++i) {
166 if (CTR_IS_RESERVED(msrs,i))
167 release_perfctr_nmi(MSR_P6_PERFCTR0 + i);
168 }
169 for (i = 0 ; i < NUM_CONTROLS ; ++i) {
170 if (CTRL_IS_RESERVED(msrs,i))
171 release_evntsel_nmi(MSR_P6_EVNTSEL0 + i);
172 }
133} 173}
134 174
135 175
@@ -140,5 +180,6 @@ struct op_x86_model_spec const op_ppro_spec = {
140 .setup_ctrs = &ppro_setup_ctrs, 180 .setup_ctrs = &ppro_setup_ctrs,
141 .check_ctrs = &ppro_check_ctrs, 181 .check_ctrs = &ppro_check_ctrs,
142 .start = &ppro_start, 182 .start = &ppro_start,
143 .stop = &ppro_stop 183 .stop = &ppro_stop,
184 .shutdown = &ppro_shutdown
144}; 185};
diff --git a/arch/i386/oprofile/op_x86_model.h b/arch/i386/oprofile/op_x86_model.h
index 123b7e90a9ee..abb1aa95b979 100644
--- a/arch/i386/oprofile/op_x86_model.h
+++ b/arch/i386/oprofile/op_x86_model.h
@@ -40,6 +40,7 @@ struct op_x86_model_spec {
40 struct op_msrs const * const msrs); 40 struct op_msrs const * const msrs);
41 void (*start)(struct op_msrs const * const msrs); 41 void (*start)(struct op_msrs const * const msrs);
42 void (*stop)(struct op_msrs const * const msrs); 42 void (*stop)(struct op_msrs const * const msrs);
43 void (*shutdown)(struct op_msrs const * const msrs);
43}; 44};
44 45
45extern struct op_x86_model_spec const op_ppro_spec; 46extern struct op_x86_model_spec const op_ppro_spec;
diff --git a/arch/i386/pci/Makefile b/arch/i386/pci/Makefile
index 62ad75c57e6a..1594d2f55c8f 100644
--- a/arch/i386/pci/Makefile
+++ b/arch/i386/pci/Makefile
@@ -11,4 +11,4 @@ pci-y += legacy.o irq.o
11pci-$(CONFIG_X86_VISWS) := visws.o fixup.o 11pci-$(CONFIG_X86_VISWS) := visws.o fixup.o
12pci-$(CONFIG_X86_NUMAQ) := numa.o irq.o 12pci-$(CONFIG_X86_NUMAQ) := numa.o irq.o
13 13
14obj-y += $(pci-y) common.o 14obj-y += $(pci-y) common.o early.o
diff --git a/arch/i386/pci/common.c b/arch/i386/pci/common.c
index 0a362e3aeac5..68bce194e688 100644
--- a/arch/i386/pci/common.c
+++ b/arch/i386/pci/common.c
@@ -242,6 +242,10 @@ char * __devinit pcibios_setup(char *str)
242 acpi_noirq_set(); 242 acpi_noirq_set();
243 return NULL; 243 return NULL;
244 } 244 }
245 else if (!strcmp(str, "noearly")) {
246 pci_probe |= PCI_PROBE_NOEARLY;
247 return NULL;
248 }
245#ifndef CONFIG_X86_VISWS 249#ifndef CONFIG_X86_VISWS
246 else if (!strcmp(str, "usepirqmask")) { 250 else if (!strcmp(str, "usepirqmask")) {
247 pci_probe |= PCI_USE_PIRQ_MASK; 251 pci_probe |= PCI_USE_PIRQ_MASK;
diff --git a/arch/i386/pci/direct.c b/arch/i386/pci/direct.c
index 5d81fb510375..5acf0b4743cf 100644
--- a/arch/i386/pci/direct.c
+++ b/arch/i386/pci/direct.c
@@ -254,7 +254,16 @@ static int __init pci_check_type2(void)
254 return works; 254 return works;
255} 255}
256 256
257void __init pci_direct_init(void) 257void __init pci_direct_init(int type)
258{
259 printk(KERN_INFO "PCI: Using configuration type %d\n", type);
260 if (type == 1)
261 raw_pci_ops = &pci_direct_conf1;
262 else
263 raw_pci_ops = &pci_direct_conf2;
264}
265
266int __init pci_direct_probe(void)
258{ 267{
259 struct resource *region, *region2; 268 struct resource *region, *region2;
260 269
@@ -264,19 +273,16 @@ void __init pci_direct_init(void)
264 if (!region) 273 if (!region)
265 goto type2; 274 goto type2;
266 275
267 if (pci_check_type1()) { 276 if (pci_check_type1())
268 printk(KERN_INFO "PCI: Using configuration type 1\n"); 277 return 1;
269 raw_pci_ops = &pci_direct_conf1;
270 return;
271 }
272 release_resource(region); 278 release_resource(region);
273 279
274 type2: 280 type2:
275 if ((pci_probe & PCI_PROBE_CONF2) == 0) 281 if ((pci_probe & PCI_PROBE_CONF2) == 0)
276 return; 282 return 0;
277 region = request_region(0xCF8, 4, "PCI conf2"); 283 region = request_region(0xCF8, 4, "PCI conf2");
278 if (!region) 284 if (!region)
279 return; 285 return 0;
280 region2 = request_region(0xC000, 0x1000, "PCI conf2"); 286 region2 = request_region(0xC000, 0x1000, "PCI conf2");
281 if (!region2) 287 if (!region2)
282 goto fail2; 288 goto fail2;
@@ -284,10 +290,11 @@ void __init pci_direct_init(void)
284 if (pci_check_type2()) { 290 if (pci_check_type2()) {
285 printk(KERN_INFO "PCI: Using configuration type 2\n"); 291 printk(KERN_INFO "PCI: Using configuration type 2\n");
286 raw_pci_ops = &pci_direct_conf2; 292 raw_pci_ops = &pci_direct_conf2;
287 return; 293 return 2;
288 } 294 }
289 295
290 release_resource(region2); 296 release_resource(region2);
291 fail2: 297 fail2:
292 release_resource(region); 298 release_resource(region);
299 return 0;
293} 300}
diff --git a/arch/i386/pci/early.c b/arch/i386/pci/early.c
new file mode 100644
index 000000000000..713d6c866cae
--- /dev/null
+++ b/arch/i386/pci/early.c
@@ -0,0 +1,52 @@
1#include <linux/kernel.h>
2#include <linux/pci.h>
3#include <asm/pci-direct.h>
4#include <asm/io.h>
5#include "pci.h"
6
7/* Direct PCI access. This is used for PCI accesses in early boot before
8 the PCI subsystem works. */
9
10#define PDprintk(x...)
11
12u32 read_pci_config(u8 bus, u8 slot, u8 func, u8 offset)
13{
14 u32 v;
15 outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8);
16 v = inl(0xcfc);
17 if (v != 0xffffffff)
18 PDprintk("%x reading 4 from %x: %x\n", slot, offset, v);
19 return v;
20}
21
22u8 read_pci_config_byte(u8 bus, u8 slot, u8 func, u8 offset)
23{
24 u8 v;
25 outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8);
26 v = inb(0xcfc + (offset&3));
27 PDprintk("%x reading 1 from %x: %x\n", slot, offset, v);
28 return v;
29}
30
31u16 read_pci_config_16(u8 bus, u8 slot, u8 func, u8 offset)
32{
33 u16 v;
34 outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8);
35 v = inw(0xcfc + (offset&2));
36 PDprintk("%x reading 2 from %x: %x\n", slot, offset, v);
37 return v;
38}
39
40void write_pci_config(u8 bus, u8 slot, u8 func, u8 offset,
41 u32 val)
42{
43 PDprintk("%x writing to %x: %x\n", slot, offset, val);
44 outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8);
45 outl(val, 0xcfc);
46}
47
48int early_pci_allowed(void)
49{
50 return (pci_probe & (PCI_PROBE_CONF1|PCI_PROBE_NOEARLY)) ==
51 PCI_PROBE_CONF1;
52}
diff --git a/arch/i386/pci/init.c b/arch/i386/pci/init.c
index 51087a9d9172..d028e1b05c36 100644
--- a/arch/i386/pci/init.c
+++ b/arch/i386/pci/init.c
@@ -6,8 +6,13 @@
6 in the right sequence from here. */ 6 in the right sequence from here. */
7static __init int pci_access_init(void) 7static __init int pci_access_init(void)
8{ 8{
9 int type = 0;
10
11#ifdef CONFIG_PCI_DIRECT
12 type = pci_direct_probe();
13#endif
9#ifdef CONFIG_PCI_MMCONFIG 14#ifdef CONFIG_PCI_MMCONFIG
10 pci_mmcfg_init(); 15 pci_mmcfg_init(type);
11#endif 16#endif
12 if (raw_pci_ops) 17 if (raw_pci_ops)
13 return 0; 18 return 0;
@@ -21,7 +26,7 @@ static __init int pci_access_init(void)
21 * fails. 26 * fails.
22 */ 27 */
23#ifdef CONFIG_PCI_DIRECT 28#ifdef CONFIG_PCI_DIRECT
24 pci_direct_init(); 29 pci_direct_init(type);
25#endif 30#endif
26 return 0; 31 return 0;
27} 32}
diff --git a/arch/i386/pci/mmconfig.c b/arch/i386/pci/mmconfig.c
index 972180f738d9..05be8db58a8c 100644
--- a/arch/i386/pci/mmconfig.c
+++ b/arch/i386/pci/mmconfig.c
@@ -151,6 +151,38 @@ static struct pci_raw_ops pci_mmcfg = {
151 .write = pci_mmcfg_write, 151 .write = pci_mmcfg_write,
152}; 152};
153 153
154
155static __init void pci_mmcfg_insert_resources(void)
156{
157#define PCI_MMCFG_RESOURCE_NAME_LEN 19
158 int i;
159 struct resource *res;
160 char *names;
161 unsigned num_buses;
162
163 res = kcalloc(PCI_MMCFG_RESOURCE_NAME_LEN + sizeof(*res),
164 pci_mmcfg_config_num, GFP_KERNEL);
165
166 if (!res) {
167 printk(KERN_ERR "PCI: Unable to allocate MMCONFIG resources\n");
168 return;
169 }
170
171 names = (void *)&res[pci_mmcfg_config_num];
172 for (i = 0; i < pci_mmcfg_config_num; i++, res++) {
173 num_buses = pci_mmcfg_config[i].end_bus_number -
174 pci_mmcfg_config[i].start_bus_number + 1;
175 res->name = names;
176 snprintf(names, PCI_MMCFG_RESOURCE_NAME_LEN, "PCI MMCONFIG %u",
177 pci_mmcfg_config[i].pci_segment_group_number);
178 res->start = pci_mmcfg_config[i].base_address;
179 res->end = res->start + (num_buses << 20) - 1;
180 res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
181 insert_resource(&iomem_resource, res);
182 names += PCI_MMCFG_RESOURCE_NAME_LEN;
183 }
184}
185
154/* K8 systems have some devices (typically in the builtin northbridge) 186/* K8 systems have some devices (typically in the builtin northbridge)
155 that are only accessible using type1 187 that are only accessible using type1
156 Normally this can be expressed in the MCFG by not listing them 188 Normally this can be expressed in the MCFG by not listing them
@@ -187,7 +219,9 @@ static __init void unreachable_devices(void)
187 } 219 }
188} 220}
189 221
190void __init pci_mmcfg_init(void) 222
223
224void __init pci_mmcfg_init(int type)
191{ 225{
192 if ((pci_probe & PCI_PROBE_MMCONF) == 0) 226 if ((pci_probe & PCI_PROBE_MMCONF) == 0)
193 return; 227 return;
@@ -198,7 +232,9 @@ void __init pci_mmcfg_init(void)
198 (pci_mmcfg_config[0].base_address == 0)) 232 (pci_mmcfg_config[0].base_address == 0))
199 return; 233 return;
200 234
201 if (!e820_all_mapped(pci_mmcfg_config[0].base_address, 235 /* Only do this check when type 1 works. If it doesn't work
236 assume we run on a Mac and always use MCFG */
237 if (type == 1 && !e820_all_mapped(pci_mmcfg_config[0].base_address,
202 pci_mmcfg_config[0].base_address + MMCONFIG_APER_MIN, 238 pci_mmcfg_config[0].base_address + MMCONFIG_APER_MIN,
203 E820_RESERVED)) { 239 E820_RESERVED)) {
204 printk(KERN_ERR "PCI: BIOS Bug: MCFG area at %x is not E820-reserved\n", 240 printk(KERN_ERR "PCI: BIOS Bug: MCFG area at %x is not E820-reserved\n",
@@ -212,4 +248,5 @@ void __init pci_mmcfg_init(void)
212 pci_probe = (pci_probe & ~PCI_PROBE_MASK) | PCI_PROBE_MMCONF; 248 pci_probe = (pci_probe & ~PCI_PROBE_MASK) | PCI_PROBE_MMCONF;
213 249
214 unreachable_devices(); 250 unreachable_devices();
251 pci_mmcfg_insert_resources();
215} 252}
diff --git a/arch/i386/pci/pci.h b/arch/i386/pci/pci.h
index bf4e79335388..1814f74569c6 100644
--- a/arch/i386/pci/pci.h
+++ b/arch/i386/pci/pci.h
@@ -17,6 +17,7 @@
17#define PCI_PROBE_CONF2 0x0004 17#define PCI_PROBE_CONF2 0x0004
18#define PCI_PROBE_MMCONF 0x0008 18#define PCI_PROBE_MMCONF 0x0008
19#define PCI_PROBE_MASK 0x000f 19#define PCI_PROBE_MASK 0x000f
20#define PCI_PROBE_NOEARLY 0x0010
20 21
21#define PCI_NO_SORT 0x0100 22#define PCI_NO_SORT 0x0100
22#define PCI_BIOS_SORT 0x0200 23#define PCI_BIOS_SORT 0x0200
@@ -81,7 +82,9 @@ extern int pci_conf1_write(unsigned int seg, unsigned int bus,
81extern int pci_conf1_read(unsigned int seg, unsigned int bus, 82extern int pci_conf1_read(unsigned int seg, unsigned int bus,
82 unsigned int devfn, int reg, int len, u32 *value); 83 unsigned int devfn, int reg, int len, u32 *value);
83 84
84extern void pci_direct_init(void); 85extern int pci_direct_probe(void);
86extern void pci_direct_init(int type);
85extern void pci_pcbios_init(void); 87extern void pci_pcbios_init(void);
86extern void pci_mmcfg_init(void); 88extern void pci_mmcfg_init(int type);
87extern void pcibios_sort(void); 89extern void pcibios_sort(void);
90
diff --git a/arch/i386/power/swsusp.S b/arch/i386/power/swsusp.S
index c893b897217f..8a2b50a0aaad 100644
--- a/arch/i386/power/swsusp.S
+++ b/arch/i386/power/swsusp.S
@@ -32,7 +32,7 @@ ENTRY(swsusp_arch_resume)
32 movl $swsusp_pg_dir-__PAGE_OFFSET, %ecx 32 movl $swsusp_pg_dir-__PAGE_OFFSET, %ecx
33 movl %ecx, %cr3 33 movl %ecx, %cr3
34 34
35 movl pagedir_nosave, %edx 35 movl restore_pblist, %edx
36 .p2align 4,,7 36 .p2align 4,,7
37 37
38copy_loop: 38copy_loop:
generated by cgit v1.2.2 (git 2.25.0) at 2026-01-01 04:15:17 -0500