The problems are that, with the ACPI vs timer overring issue _fixed_,
after using the box for some time (between several seconds and 1 hour, at
random) processes get very high CPU loads (once I've got X using 107% of
the CPU, for example) and the system becomes unresponsive, as though there
were interrupts lost or something similar.

Andreas Herrman reproduced similar problems:

> Ok, now I've reproduced the stability problem.
> - Using tip/master,
> - reverting e38502eb8aa82314d5ab0eba45f50e6790dadd88 and
> - applying your patch from this posting
>   http://marc.info/?l=linux-kernel&m=121539354224562&w=4


>
> Starting X, firefox, gimp, tuxpaint and doing some drawing in tuxpaint
> results in a slow system. Drawing is almost not possible anymore --
> Selections of new colors, cursors etc. is performed with huge delay
> if it's performed at all.
>
> BTW, the code sets up timer IRQ as Virtual Wire IRQ:
>
> Jul  8 14:57:58 kodscha IO-APIC (apicid-pin) 2-22, 2-23 not connected.
> Jul  8 14:57:58 kodscha ..TIMER: vector=0x30 apic1=0 pin1=2 apic2=-1 pin2=-1
> Jul  8 14:57:58 kodscha ...trying to set up timer as Virtual Wire IRQ... works.
>
> and both INT0 and INT2 of IOAPIC are masked:
>
> Jul  8 14:57:58 kodscha NR Dst Mask Trig IRR Pol Stat Dmod Deli Vect:
> Jul  8 14:57:58 kodscha 00 000 1    0    0   0   0    0    0    00
> Jul  8 14:57:58 kodscha 01 003 0    0    0   0   0    1    1    31
> Jul  8 14:57:58 kodscha 02 003 1    0    0   0   0    0    0    30
>
> I've also seen strange CPU utilization -- with syslog-ng:
>
> top - 15:33:06 up 35 min,  4 users,  load average: 1.70, 0.68, 0.37
> Tasks:  64 total,   4 running,  60 sleeping,   0 stopped,   0 zombie
> Cpu0  :  0.0%us,100.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
> Cpu1  :  6.4%us, 87.2%sy,  0.0%ni,  5.8%id,  0.0%wa,  0.6%hi,  0.0%si,  0.0%st
> Mem:    895384k total,   283568k used,   611816k free,    35492k buffers
> Swap:  1959920k total,        0k used,  1959920k free,   163044k cached
>
>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
>  4632 root      20   0 17216  800  580 S  104  0.1   0:34.22 syslog-ng
> 28505 root      20   0  205m  11m 4024 S    6  1.3   0:21.16 X
> 28518 root      20   0 56292 5652 4492 S    1  0.6   0:01.80 fluxbox
>     1 root      20   0  3724  608  508 S    0  0.1   0:00.36 init
>
> So far I have no clue why C1E-idle in conjunction with virtual wire
> mode causes this strange behaviour.
>
> ... and I start to think about the root cause of all this.
>
> I've performed similar tests under X with the IRQ0/INT0 configuration and
> I did not see above symptoms.

So lets fall back to the IRQ0/INT0 configuration on this box.

This basically restores the dont-use-the-lapic-timer exception mechanism
that was unconditional on this box prior commit 8750bf5 ("x86: add C1E
aware idle function").

Signed-off-by: Ingo Molnar <mingo@elte.hu>

handle head and tail that are not aligned to big pages (2MB/1GB boundary).

with this patch, on system that support gbpages, change:

  last_map_addr: 1080000000 end: 1078000000

to:

  last_map_addr: 1078000000 end: 1078000000

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

When system have 4g less ram installed, and acpi table sit
near end of ram, make max_pfn cover them too,
so 64bit kernel don't need to mess up fixmap.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: "Suresh Siddha" <suresh.b.siddha@intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: "Nick Piggin" <npiggin@suse.de>
Cc: "Mark McLoughlin" <markmc@redhat.com>
Cc: xen-devel <xen-devel@lists.xensource.com>
Cc: "Eduardo Habkost" <ehabkost@redhat.com>
Cc: "Vegard Nossum" <vegard.nossum@gmail.com>
Cc: "Stephen Tweedie" <sct@redhat.com>
Cc: "Jeremy Fitzhardinge" <jeremy@goop.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

prepare for overmapped patch

also printout last_map_addr together with end

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

fix:

 In file included from arch/x86/kernel/tlb_uv.c:14:
 include/asm/uv/uv_mmrs.h:986: error: redefinition of ‘union uvh_rh_gam_cfg_overlay_config_mmr_u’
 include/asm/uv/uv_mmrs.h:988: error: redefinition of ‘struct uvh_rh_gam_cfg_overlay_config_mmr_s’
 include/asm/uv/uv_mmrs.h:1064: error: redefinition of ‘union uvh_rh_gam_mmioh_overlay_config_mmr_u’
 include/asm/uv/uv_mmrs.h:1066: error: redefinition of ‘struct uvh_rh_gam_mmioh_overlay_config_mmr_s’

caused by another duplicate section (cut & paste error) in commit
5d061e39

 "x86, uv: update x86 mmr list for SGI uv".

Signed-off-by: Ingo Molnar <mingo@elte.hu>

fix:

In file included from arch/x86/kernel/genx2apic_uv_x.c:25:
include/asm/uv/uv_mmrs.h:986: error: redefinition of ‘union uvh_rh_gam_cfg_overlay_config_mmr_u’
include/asm/uv/uv_mmrs.h:988: error: redefinition of ‘struct uvh_rh_gam_cfg_overlay_config_mmr_s’
include/asm/uv/uv_mmrs.h:1064: error: redefinition of ‘union uvh_rh_gam_mmioh_overlay_config_mmr_u’
include/asm/uv/uv_mmrs.h:1066: error: redefinition of ‘struct uvh_rh_gam_mmioh_overlay_config_mmr_s’

caused by duplicate section (cut & paste error) in commit
5d061e39

 "x86, uv: update x86 mmr list for SGI uv".

Signed-off-by: Ingo Molnar <mingo@elte.hu>

arch/x86/mm/pgtable_32.c:144: warning: 'fixmaps' defined but not used

Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

arch/x86/kernel/smpboot.c: In function 'do_boot_cpu':
arch/x86/kernel/smpboot.c:943: warning: label 'restore_state' defined but not used

Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

 - order of local variable declarations
 - minor code changes

Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
Acked-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

 - local caching of smp_processor_id() in default_do_nmi()
 - v2: do not split default_do_nmi over two lines

On Wed, Jul 02, 2008 at 08:12:20PM +0400, Cyrill Gorcunov wrote:
> | -static notrace __kprobes void default_do_nmi(struct pt_regs *regs)
> | +static notrace __kprobes void
> | +default_do_nmi(struct pt_regs *regs)
> | [ ... ]
> | -asmlinkage notrace  __kprobes void default_do_nmi(struct pt_regs *regs)
> | +asmlinkage notrace __kprobes void
> | +default_do_nmi(struct pt_regs *regs)
>
> Hi Alexander, good done, thanks! But why did you split default_do_nmi
> definition by two lines? I think it would be better to keep them as it
> was before, ie by a single line
>
> 	static notrace __kprobes void default_do_nmi(struct pt_regs *regs)

Thanks! Here is the replacement patch with default_do_nmi left on
a single line.

Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
Acked-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

 - if (cond) block -> if (!cond) goto end_of_block
 - local caching of current

Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
Acked-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

if (cond) block -> if (!cond) goto end_of_block

Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
Acked-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

if (cond) block -> if (!cond) goto end_of_block

Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
Acked-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Reorder headers and collect globals in traps_32.c and traps_64.c

Code size and data size are unaffected by the changes. Code
itself is changed due to different ordering of data and bss.
The bss segment changed size due to a change in the packing
of the variables.

Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
Acked-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

This patch does not change the generated object files.

Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Rename the paravirtualized calculate_cpu_khz to calibrate_tsc.
In all cases, we actually calibrate_tsc and use that as the cpu_khz value.

Signed-off-by: Alok N Kataria <akataria@vmware.com>
Signed-off-by: Dan Hecht <dhecht@vmware.com>
Cc: Dan Hecht <dhecht@vmware.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Unify the clocksource code.
Unify the tsc_init code.

Signed-off-by: Alok N Kataria <akataria@vmware.com>
Signed-off-by: Dan Hecht <dhecht@vmware.com>
Cc: Dan Hecht <dhecht@vmware.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Unify the TSC cpufreq code.

Signed-off-by: Alok N Kataria <akataria@vmware.com>
Signed-off-by: Dan Hecht <dhecht@vmware.com>
Cc: Dan Hecht <dhecht@vmware.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Merge the tsc calibration code for the 32bit and 64bit kernel.
The paravirtualized calculate_cpu_khz for 64bit now points to the correct
tsc_calibrate code as in 32bit.
Original native_calculate_cpu_khz for 64 bit is now called as calibrate_cpu.

Also moved the recalibrate_cpu_khz function in the common file.
Note that this function is called only from powernow K7 cpu freq driver.

Signed-off-by: Alok N Kataria <akataria@vmware.com>
Signed-off-by: Dan Hecht <dhecht@vmware.com>
Cc: Dan Hecht <dhecht@vmware.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Move the basic global variable definitions and sched_clock handling in the
common "tsc.c" file.

 - Unify notsc kernel command line handling for 32 bit and 64bit.
 - Functional changes for 64bit.
        - "tsc_disabled" is updated if "notsc" is passed at boottime.
        - Fallback to jiffies for sched_clock, incase notsc is passed on
	  commandline.

Signed-off-by: Alok N Kataria <akataria@vmware.com>
Signed-off-by: Dan Hecht <dhecht@vmware.com>
Cc: Dan Hecht <dhecht@vmware.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Add lapic resource into kernel resource map and mark it as busy

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: "Maciej W. Rozycki" <macro@linux-mips.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
CC: Maciej W. Rozycki <macro@linux-mips.org>

This patch updates the X86 mmr list for SGI uv.

Signed-off-by: Dimitri Sivanich <sivanich@sgi.com>
Cc: Jack Steiner <steiner@sgi.com>
Cc: Russ Anderson <rja@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Create page table entries to map the SGI UV chipset GRU. local MMR &
global MMR ranges.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Cc: linux-mm@kvack.org
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Add boot-time function for creating additional 2MB page table entries for
mapping chipset specific cached/uncached ranges.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Cc: linux-mm@kvack.org
Signed-off-by: Ingo Molnar <mingo@elte.hu>

so when it is called after early_param, e820_saved get updated too.
esp for mpc update.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: Bernhard Walle <bwalle@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

seperate reserve_setup_data into e820_reserved_setup_data,
and reserve_early_setup_data.

So could use e820_reserved_setup_data to backup e820 with setup_data.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: Bernhard Walle <bwalle@suse.de>
Cc: Ying Huang <ying.huang@intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

so other path that will override memory_setup or
machine_specific_memory_setup could have e820_saved too.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Bernhard Walle <bwalle@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

This patch uses the /sys/firmware/memmap interface provided in the last patch
on the x86 architecture when E820 is used. The patch copies the E820
memory map very early, and registers the E820 map afterwards via
firmware_map_add_early().

Signed-off-by: Bernhard Walle <bwalle@suse.de>
Acked-by: Greg KH <gregkh@suse.de>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: kexec@lists.infradead.org
Cc: yhlu.kernel@gmail.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>

This patch adds /sys/firmware/memmap interface that represents the BIOS
(or Firmware) provided memory map. The tree looks like:

    /sys/firmware/memmap/0/start   (hex number)
                           end     (hex number)
                           type    (string)
    ...                 /1/start
                           end
                           type

With the following shell snippet one can print the memory map in the same form
the kernel prints itself when booting on x86 (the E820 map).

  --------- 8< --------------------------
    #!/bin/sh
    cd /sys/firmware/memmap
    for dir in * ; do
        start=$(cat $dir/start)
        end=$(cat $dir/end)
        type=$(cat $dir/type)
        printf "%016x-%016x (%s)\n" $start $[ $end +1] "$type"
    done
  --------- >8 --------------------------

That patch only provides the needed interface:

 1. The sysfs interface.
 2. The structure and enumeration definition.
 3. The function firmware_map_add() and firmware_map_add_early()
    that should be called from architecture code (E820/EFI, for
    example) to add the contents to the interface.

If the kernel is compiled without CONFIG_FIRMWARE_MEMMAP, the interface does
nothing without cluttering the architecture-specific code with #ifdef's.

The purpose of the new interface is kexec: While /proc/iomem represents
the *used* memory map (e.g. modified via kernel parameters like 'memmap'
and 'mem'), the /sys/firmware/memmap tree represents the unmodified memory
map provided via the firmware. So kexec can:

 - use the original memory map for rebooting,
 - use the /proc/iomem for setting up the ELF core headers for kdump
   case that should only represent the memory of the system.

The patch has been tested on i386 and x86_64.

Signed-off-by: Bernhard Walle <bwalle@suse.de>
Acked-by: Greg KH <gregkh@suse.de>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: kexec@lists.infradead.org
Cc: yhlu.kernel@gmail.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>

use ACPI_NUMA directly

and move srat_32.c to mm/

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

we already have the same srat handling interface for 32bit.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

this symbol is needed by drivers/video/xen-fbfront.ko.

[ cherry-picked from tip/core/printk ]

Signed-off-by: Ingo Molnar <mingo@elte.hu>

Rather than using _PAGE_GLOBAL - which not all CPUs support - to test
CPA, use one of the reserved-for-software-use PTE flags instead.  This
allows CPA testing to work on CPUs which don't support PGD.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

From: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>

Older x86-32 processors do not support global mappings (PGD), so must
only use it if the processor supports it.

The _PAGE_KERNEL* flags always have _PAGE_KERNEL set, since logically
we always want it set.

This is OK even on processors which do not support PGD, since all
_PAGE flags are masked with __supported_pte_mask before being turned
into a real in-pagetable pte.  On 32-bit systems, __supported_pte_mask
is initialized to not contain _PAGE_GLOBAL, and it is then added if
the CPU is found to support it.

The x86-32 code used to use __PAGE_KERNEL/__PAGE_KERNEL_EXEC for this
purpose, but they're now redundant and can be removed.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Consistently set _PAGE_GLOBAL in _PAGE_KERNEL flags.  This makes 32-
and 64-bit code consistent, and removes some special cases where
__PAGE_KERNEL* did not have _PAGE_GLOBAL set, causing confusion as a
result of the inconsistencies.

This patch only affects x86-64, which generally always supports PGD.
The x86-32 patch is next.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

When allocating a new pud, unconditionally populate the pgd (why did
we bother to create a new pud if we weren't going to populate it?).

This will only happen if the pgd slot was empty, since any existing
pud will be reused.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

add back this line lost from trap_init():

        set_trap_gate(0,  &divide_error);

Signed-off-by: Ingo Molnar <mingo@elte.hu>

fix:

arch/x86/kernel/built-in.o: In function `setup_arch':
: undefined reference to `prefill_possible_map'

Signed-off-by: Ingo Molnar <mingo@elte.hu>

call it right after we are done with MADT/mptable handling, instead of
doing that in setup_per_cpu_areas() later on...

this way for_possible_cpu() can be used early.

Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>