Konrad Wilk reported that the new delayed calibration crashes with a
divide by zero on Xen. The reason is that Xen sets the pmtimer
address, but reading from it returns 0xffffff. That results in the
ref_start and ref_stop value being the same, so the delta is zero
which causes the divide by zero later in the calculation.

The conditional (!hpet && !ref_start && !ref_stop) which sanity checks
the calibration reference values doesn't really make sense. If the
refs are null, but hpet is on, we still want to break out.

The div by zero would be possible to trigger by chance if both reads
from the hardware provided the exact same value (due to hardware
wrapping).

So checking if both the ref values are the same should handle if we
don't have hardware (both null) or if they are the same value (either by
invalid hardware, or by chance), avoiding the div by zero issue.

[ tglx: Applied the same fix to native_calibrate_tsc() where this
  	check was copied from ]
Reported-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Tested-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: NJohn Stultz <johnstul@us.ibm.com>
LKML-Reference: <1295024788-15619-1-git-send-email-johnstul@us.ibm.com>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

The x86 fixmaps need to be all together... unfortunately the
VRTC one was misplaced.

This patch makes sure the MRST VRTC fixmap is put prior to the
__end_of_permanent_fixed_addresses marker.
Signed-off-by: NArjan van de Ven <arjan@linux.intel.com>
Signed-off-by: NAlan Cox <alan@linux.intel.com>
LKML-Reference: <20110111105544.24448.27607.stgit@bob.linux.org.uk>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

We need this for x86 MID platforms where GPIO interrupts are
used. No special magic is needed so the default 1:1 behaviour
will do nicely.
Signed-off-by: NAlek Du <alek.du@intel.com>
Signed-off-by: NJacob Pan <jacob.jun.pan@intel.com>
Signed-off-by: NAlan Cox <alan@linux.intel.com>
LKML-Reference: <20110111105439.24448.69863.stgit@bob.linux.org.uk>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Westmere processors use a different algorithm for
assigning APICIDs on SGI UV systems. The location of the
node number within the apicid is now a function of the
processor type.
Signed-off-by: NJack Steiner <steiner@sgi.com>
LKML-Reference: <20110110195210.GA18737@sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

While both methods should work equivalently well for the native
case, the Xen Dom0 case can't reliably work with the MSR one,
since there's no guarantee that the virtual CPUs it has
available fully cover all necessary physical ones.

As per the suggestion of Robert Richter the patch only adds the
PCI method, but leaves the MSR one as a fallback to cover new
systems the PCI IDs of which may not have got added to the code
base yet.

The only change in v2 is the breaking out of the new CPI
initialization method into a separate function, as requested by
Ingo.
Signed-off-by: NJan Beulich <jbeulich@novell.com>
Acked-by: NRobert Richter <robert.richter@amd.com>
Cc: Andreas Herrmann3 <Andreas.Herrmann3@amd.com>
Cc: Joerg Roedel <joerg.roedel@amd.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
LKML-Reference: <4D2B3FD7020000780002B67D@vpn.id2.novell.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

commit a8760eca (x86: Check tsc available/disabled in the delayed init
function) missed to prevent the setup of the delayed init function in
case the initial tsc calibration failed. This results in the same
divide by zero bug as we have seen without the tsc disabled check.

Skip the delayed work setup when tsc_khz (the initial calibration
value) is 0.
Bisected-and-tested-by: NKirill A. Shutemov <kas@openvz.org>
Cc: John Stultz <john.stultz@linaro.org>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Latest atom socs(penwell) does not have hpet timer.

As their local APIC timer is clocked at 400KHZ, and the current
code limit their Initial Counter register to 23 bits, they
cannot sleep more than 1.34 seconds which leads to ~2 spurious
wakeup per second (1 per thread)

These SOCs support 32bit timer so we change the max_delta to at
least 31bits. So we can at least sleep for 300 seconds.

We could not find any previous chip errata where lapic would
only have 23 bit precision As powertop is suggesting to activate
HPET to "sleep longer", this could mean this problem is already
known.

Problem is here since very first implementation of lapic timer
as a clock event e9e2cdb4 [PATCH] clockevents: i386 drivers.
Signed-off-by: NSuresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: NPierre Tardy <pierre.tardy@intel.com>
Acked-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Adrian Bunk <bunk@stusta.de>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Cc: john stultz <johnstul@us.ibm.com>
Cc: Roman Zippel <zippel@linux-m68k.org>
Cc: Andi Kleen <ak@suse.de>
LKML-Reference: <1294327409-19426-1-git-send-email-pierre.tardy@intel.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Fix sparse warning for non-ANSI function declaration:

  arch/x86/kernel/smpboot.c:100:30: warning: non-ANSI function declaration of function 'cpu_hotplug_driver_lock'
  arch/x86/kernel/smpboot.c:105:32: warning: non-ANSI function declaration of function 'cpu_hotplug_driver_unlock'
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
LKML-Reference: <20110108195914.95d366ea.randy.dunlap@oracle.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

"x86, numa: Fake node-to-cpumask for NUMA emulation" broke the
build when CONFIG_DEBUG_PER_CPU_MAPS is set and CONFIG_NUMA_EMU
is not.  This is because it is possible to map a cpu to multiple
nodes when NUMA emulation is used; the patch required a physical
node address table to find those nodes that was only available
when CONFIG_NUMA_EMU was enabled.

This extracts the common debug functionality to its own function
for CONFIG_DEBUG_PER_CPU_MAPS and uses it regardless of whether
CONFIG_NUMA_EMU is set or not.

NUMA emulation will now iterate over the set of possible nodes
for each cpu and call the new debug function whereas only the
cpu's node will be used without NUMA emulation enabled.
Reported-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NDavid Rientjes <rientjes@google.com>
Acked-by: NYinghai Lu <yinghai@kernel.org>
LKML-Reference: <alpine.DEB.2.00.1012301053590.12995@chino.kir.corp.google.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Prevent the long delay in io_check_error making NMI watchdog
timeout.
Signed-off-by: NHuang Ying <ying.huang@intel.com>
Signed-off-by: NDon Zickus <dzickus@redhat.com>
LKML-Reference: <1294198689-15447-3-git-send-email-dzickus@redhat.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

The spin_lock_debug/rcu_cpu_stall detector uses
trigger_all_cpu_backtrace() to dump cpu backtrace.
Therefore it is possible that trigger_all_cpu_backtrace()
could be called at the same time on different CPUs, which
triggers and 'unknown reason NMI' warning. The following case
illustrates the problem:

      CPU1                    CPU2                     ...   CPU N
                       trigger_all_cpu_backtrace()
                       set "backtrace_mask" to cpu mask
                               |
generate NMI interrupts  generate NMI interrupts       ...
    \                          |                               /
     \                         |                              /

The "backtrace_mask" will be cleaned by the first NMI interrupt
at nmi_watchdog_tick(), then the following NMI interrupts
generated by other cpus's arch_trigger_all_cpu_backtrace() will
be taken as unknown reason NMI interrupts.

This patch uses a test_and_set to avoid the problem, and stop
the arch_trigger_all_cpu_backtrace() from calling to avoid
dumping a double cpu backtrace info when there is already a
trigger_all_cpu_backtrace() in progress.
Signed-off-by: NDongdong Deng <dongdong.deng@windriver.com>
Reviewed-by: NBruce Ashfield <bruce.ashfield@windriver.com>
Cc: fweisbec@gmail.com
LKML-Reference: <1294198689-15447-2-git-send-email-dzickus@redhat.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NDon Zickus <dzickus@redhat.com>

There are some paths that walk the die_chain with preemption on.
Make sure we are in an NMI call before we start doing anything.

This was triggered by do_general_protection calling notify_die
with DIE_GPF.
Reported-by: NJan Kiszka <jan.kiszka@web.de>
Signed-off-by: NDon Zickus <dzickus@redhat.com>
LKML-Reference: <1294198689-15447-1-git-send-email-dzickus@redhat.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Found one x2apic pre-enabled system, x2apic_mode suddenly get
corrupted after register some cpus, when compiled
CONFIG_NR_CPUS=255 instead of 512.

It turns out that generic_processor_info() ==> phyid_set(apicid,
phys_cpu_present_map) causes the problem.

phys_cpu_present_map is sized by MAX_APICS bits, and pre-enabled
system some cpus have an apic id > 255.

The variable after phys_cpu_present_map may get corrupted
silently:

 ffffffff828e8420 B phys_cpu_present_map
 ffffffff828e8440 B apic_verbosity
 ffffffff828e8444 B local_apic_timer_c2_ok
 ffffffff828e8448 B disable_apic
 ffffffff828e844c B x2apic_mode
 ffffffff828e8450 B x2apic_disabled
 ffffffff828e8454 B num_processors
 ...

Actually phys_cpu_present_map is referenced via apic id, instead
index. We should use MAX_LOCAL_APIC instead MAX_APICS.

For 64-bit it will be 32768 in all cases. BSS will increase by 4k bytes
on 64-bit:

	text		data		bss		dec		filename
	21696943	4193748		12787712	38678403	vmlinux.before
	21696943	4193748		12791808	38682499	vmlinux.after

No change on 32bit.

Finally we can remove MAX_APCIS that was rather confusing.
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
LKML-Reference: <4D23BD9C.3070102@kernel.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

v2.6.36-rc8-54-gb40827fa (x86-32, mm: Add an initial page table
for core bootstrapping) made x86 boot using initial_page_table
and broke lguest.

For 2.6.37 we simply cut & paste the initialization code into
lguest (da32dac1 "lguest: populate initial_page_table"), now
we fix it properly by doing that initialization before the
paravirt jump.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Acked-by: NJeremy Fitzhardinge <jeremy@goop.org>
Cc: lguest <lguest@ozlabs.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <201101041720.54535.rusty@rustcorp.com.au>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Add these new power trace events:

 power:cpu_idle
 power:cpu_frequency
 power:machine_suspend

The old C-state/idle accounting events:
  power:power_start
  power:power_end

Have now a replacement (but we are still keeping the old
tracepoints for compatibility):

  power:cpu_idle

and
  power:power_frequency

is replaced with:
  power:cpu_frequency

power:machine_suspend is newly introduced.

Jean Pihet has a patch integrated into the generic layer
(kernel/power/suspend.c) which will make use of it.

the type= field got removed from both, it was never
used and the type is differed by the event type itself.

perf timechart userspace tool gets adjusted in a separate patch.
Signed-off-by: NThomas Renninger <trenn@suse.de>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Acked-by: NArjan van de Ven <arjan@linux.intel.com>
Acked-by: NJean Pihet <jean.pihet@newoldbits.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: rjw@sisk.pl
LKML-Reference: <1294073445-14812-3-git-send-email-trenn@suse.de>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
LKML-Reference: <1290072314-31155-2-git-send-email-trenn@suse.de>

Addresses https://bugzilla.kernel.org/show_bug.cgi?id=25702Reported-by: NMartin Ettl <ettl.martin@gmx.de>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix a hard-coded limit of a maximum of 16 cpu's per socket.

The UV Broadcast Assist Unit code initializes by scanning the
cpu topology of the system and assigning a master cpu for each
socket and UV hub. That scan had an assumption of a limit of 16
cpus per socket. With Westmere we are going over that limit.
The UV hub hardware will allow up to 32.

If the scan finds the system has gone over that limit it returns
an error and we print a warning and fall back to doing TLB
shootdowns without the BAU.
Signed-off-by: NCliff Wickman <cpw@sgi.com>
Cc: <stable@kernel.org> # .37.x
LKML-Reference: <E1PZol7-0000mM-77@eag09.americas.sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

This patch adds code to therm_throt.c to notify core thermal threshold
events. These thresholds are supported by the IA32_THERM_INTERRUPT register.
The status/log for the same is monitored using the IA32_THERM_STATUS register.
The necessary #defines are in msr-index.h. A call back is added to mce.h, to
further notify the thermal stack, about the threshold events.
Signed-off-by: NDurgadoss R <durgadoss.r@intel.com>
LKML-Reference: <D6D887BA8C9DFF48B5233887EF04654105C1251710@bgsmsx502.gar.corp.intel.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

Before this patch, the following error would sometimes occur after a
resume on pxa3xx:

    /path/to/mm/memory.c:144: bad pmd 8040542e.

The problem was that a temporary page table mapping was being improperly
restored.

The PXA3xx resume code creates a temporary mapping of resume_turn_on_mmu
to avoid a prefetch abort.  The pxa3xx_resume_after_mmu code requires
that the r1 register holding the address of this mapping not be
modified, however, resume_turn_on_mmu does modify it. It is mostly
correct in that r1 receives the base table address, but it may also
get other bits in 13:0.  This results in pxa3xx_resume_after_mmu
restoring the original mapping to the wrong place, corrupting memory
and leaving the temporary mapping in place.
Signed-off-by: NMatt Reimer <mreimer@sdgsystems.com>
Signed-off-by: NEric Miao <eric.y.miao@gmail.com>

The commit 6ac6b817 (ARM: pxa: encode
IRQ number into .nr_irqs) removed definition of ITE_LAST_IRQ which
caused the following build error:

CC      arch/arm/common/it8152.o
arch/arm/common/it8152.c: In function 'it8152_init_irq':
arch/arm/common/it8152.c:86: error: 'IT8152_LAST_IRQ' undeclared (first use in this function)
arch/arm/common/it8152.c:86: error: (Each undeclared identifier is reported only once
arch/arm/common/it8152.c:86: error: for each function it appears in.)
make[2]: *** [arch/arm/common/it8152.o] Error 1

Defining the IT8152_LAST_IRQ in the arch/arm/include/hardware/it8152.c
fixes the build.
Signed-off-by: NMike Rapoport <mike@compulab.co.il>
Signed-off-by: NEric Miao <eric.y.miao@gmail.com>

As arch/arm/mach-pxa/eseries.c references w100fb_gpio_{read,write}()
directly.
Signed-off-by: NLennert Buytenhek <buytenh@secretlab.ca>
Signed-off-by: NEric Miao <eric.y.miao@gmail.com>

Disable preemption in init_ibs(). The function only checks the
ibs capabilities and sets up pci devices (if necessary). It runs
only on one cpu but operates with the local APIC and some MSRs,
thus it is better to disable preemption.

[    7.034377] BUG: using smp_processor_id() in preemptible [00000000] code: modprobe/483
[    7.034385] caller is setup_APIC_eilvt+0x155/0x180
[    7.034389] Pid: 483, comm: modprobe Not tainted 2.6.37-rc1-20101110+ #1
[    7.034392] Call Trace:
[    7.034400]  [<ffffffff812a2b72>] debug_smp_processor_id+0xd2/0xf0
[    7.034404]  [<ffffffff8101e985>] setup_APIC_eilvt+0x155/0x180
[ ... ]

Addresses https://bugzilla.kernel.org/show_bug.cgi?id=22812

Reported-by: <atswartz@gmail.com>
Signed-off-by: NRobert Richter <robert.richter@amd.com>
Cc: oprofile-list@lists.sourceforge.net <oprofile-list@lists.sourceforge.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Dan Carpenter <error27@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: <stable@kernel.org>         [2.6.37.x]
LKML-Reference: <20110103111514.GM4739@erda.amd.com>
[ small cleanups ]
Signed-off-by: NIngo Molnar <mingo@elte.hu>

This patch fixes below build error by adding the missing asm/memory.h,
which is needed for arch_is_coherent().

$ make pxa3xx_defconfig; make
  CC      init/do_mounts_rd.o
In file included from include/linux/list_bl.h:5,
                 from include/linux/rculist_bl.h:7,
                 from include/linux/dcache.h:7,
                 from include/linux/fs.h:381,
                 from init/do_mounts_rd.c:3:
include/linux/bit_spinlock.h: In function 'bit_spin_unlock':
include/linux/bit_spinlock.h:61: error: implicit declaration of function 'arch_is_coherent'
make[1]: *** [init/do_mounts_rd.o] Error 1
make: *** [init] Error 2
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

isr_ack is never initialized.  So, until the first PIC reset, interrupts
may fail to be injected.  This can cause Windows XP to fail to boot, as
reported in the fallout from the fix to
https://bugzilla.kernel.org/show_bug.cgi?id=21962.
Reported-and-tested-by: NNicolas Prochazka <prochazka.nicolas@gmail.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

We use the physical address instead of the base gfn for the four
PAE page directories we use in unpaged mode.  When the guest accesses
an address above 1GB that is backed by a large host page, a BUG_ON()
in kvm_mmu_set_gfn() triggers.

Resolves: https://bugzilla.kernel.org/show_bug.cgi?id=21962Reported-and-tested-by: NNicolas Prochazka <prochazka.nicolas@gmail.com>
KVM-Stable-Tag.
Signed-off-by: NAvi Kivity <avi@redhat.com>

arm: export dma_set_coherent_mask

While a regression was fixed with commit 710224fa (arm: fix
"arm: fix pci_set_consistent_dma_mask for dmabounce devices"), a
new one was introduced as dma_set_coherent_mask wasn't exported
for modules. This patch takes care for this issue.
Signed-off-by: NImre Kaloz <kaloz@openwrt.org>
Signed-off-by: NKrzysztof Hałasa <khc@pm.waw.pl>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

halt() should use native_halt()
safe_halt() uses native_safe_halt()

If CONFIG_PARAVIRT=y, halt() is defined in arch/x86/include/asm/paravirt.h as

static inline void halt(void)
{
        PVOP_VCALL0(pv_irq_ops.safe_halt);
}

Otherwise (no CONFIG_PARAVIRT) halt() in arch/x86/include/asm/irqflags.h is

static inline void halt(void)
{
        native_halt();
}

So it looks to me like the CONFIG_PARAVIRT case of using native_safe_halt()
for a halt() is an oversight.
Am I missing something?

It probably hasn't shown up as a problem because the local apic is disabled
on a shutdown or restart.  But if we disable interrupts and call halt()
we shouldn't expect that the halt() will re-enable interrupts.
Signed-off-by: NCliff Wickman <cpw@sgi.com>
LKML-Reference: <E1PSBcz-0001g1-FM@eag09.americas.sgi.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

In arch/x86/kernel/microcode_intel.c::generic_load_microcode()
we have  this:

	while (leftover) {
		...
		if (get_ucode_data(mc, ucode_ptr, mc_size) ||
		    microcode_sanity_check(mc) < 0) {
			vfree(mc);
			break;
		}
		...
	}

	if (mc)
		vfree(mc);

This will cause a double free of 'mc'. This patch fixes that by
just  removing the vfree() call in the loop since 'mc' will be
freed nicely just  after we break out of the loop.

There's also a second change in the patch. I noticed a lot of
checks for  pointers being NULL before passing them to vfree().
That's completely  redundant since vfree() deals gracefully with
being passed a NULL pointer.  Removing the redundant checks
yields a nice size decrease for the object  file.

Size before the patch:
   text    data     bss     dec     hex filename
   4578     240    1032    5850    16da arch/x86/kernel/microcode_intel.o
Size after the patch:
   text    data     bss     dec     hex filename
   4489     240     984    5713    1651 arch/x86/kernel/microcode_intel.o
Signed-off-by: NJesper Juhl <jj@chaosbits.net>
Acked-by: NTigran Aivazian <tigran@aivazian.fsnet.co.uk>
Cc: Shaohua Li <shaohua.li@intel.com>
LKML-Reference: <alpine.LNX.2.00.1012251946100.10759@swampdragon.chaosbits.net>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

If the irqsoff tracer is in use, stop tracing the interrupt disable
interval when returning to userspace.  Tracing userspace execution time
as interrupts disabled time is not helpful for kernel performance
analysis purposes.  Only do so if the irqsoff tracer is enabled, to
avoid overhead for lockdep, which doesn't care.
Signed-off-by: NTodd Poynor <toddpoynor@google.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

... and also remove misleading comment stating that this header is
auto-generated.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Acked-by: NUwe Kleine-Knig <u.kleine-koenig@pengutronix.de>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

The master clock initialization for SH7201 was wholly bogus. Users of the
legacy API must initialize the clock rate through the struct clk itself
rather than returning the clock frequency. Given that the init function
itself is void, returning the frequency isn't terribly effective.
Signed-off-by: NPaul Mundt <lethal@linux-sh.org>

With some recent tidying of duplicate register definitions the se7206 IRQ
code broke:

arch/sh/boards/mach-se/7206/irq.c: error: 'INTC_ICR' undeclared (first use in this function)
arch/sh/boards/mach-se/7206/irq.c: error: (Each undeclared identifier is reported only once
arch/sh/boards/mach-se/7206/irq.c: error: for each function it appears in.)

Fix it up.
Signed-off-by: NPaul Mundt <lethal@linux-sh.org>

Some of the SH4-202 code was overlooked in the set_rate() API conversion,
resulting in:

arch/sh/kernel/cpu/sh4/clock-sh4-202.c: error: too many arguments to function 'clk->ops->set_rate'

Fix it up.
Signed-off-by: NPaul Mundt <lethal@linux-sh.org>

NUMA boot code assumes that physical node ids start at 0, but the DIMMs
that the apic id represents may not be reachable.  If this is the case,
node 0 is never online and cpus never end up getting appropriately
assigned to a node.  This causes the cpumask of all online nodes to be
empty and machines crash with kernel code assuming online nodes have
valid cpus.

The fix is to appropriately map all the address ranges for physical nodes
and ensure the cpu to node mapping function checks all possible nodes (up
to MAX_NUMNODES) instead of simply checking nodes 0-N, where N is the
number of physical nodes, for valid address ranges.

This requires no longer "compressing" the address ranges of nodes in the
physical node map from 0-N, but rather leave indices in physnodes[] to
represent the actual node id of the physical node.  Accordingly, the
topology exported by both amd_get_nodes() and acpi_get_nodes() no longer
must return the number of nodes to iterate through; all such iterations
will now be to MAX_NUMNODES.

This change also passes the end address of system RAM (which may be
different from normal operation if mem= is specified on the command line)
before the physnodes[] array is populated.  ACPI parsed nodes are
truncated to fit within the address range that respect the mem=
boundaries and even some physical nodes may become unreachable in such
cases.

When NUMA emulation does succeed, any apicid to node mapping that exists
for unreachable nodes are given default values so that proximity domains
can still be assigned.  This is important for node_distance() to
function as desired.
Signed-off-by: NDavid Rientjes <rientjes@google.com>
LKML-Reference: <alpine.DEB.2.00.1012221702090.3701@chino.kir.corp.google.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

It's necessary to fake the node-to-cpumask mapping so that an emulated
node ID returns a cpumask that includes all cpus that have affinity to
the memory it represents.

This is a little intrusive because it requires knowledge of the physical
topology of the system.  setup_physnodes() gives us that information, but
since NUMA emulation ends up altering the physnodes array, it's necessary
to reset it before cpus are brought online.

Accordingly, the physnodes array is moved out of init.data and into
cpuinit.data since it will be needed on cpuup callbacks.

This works regardless of whether numa=fake is used on the command line,
or the setup of the fake node succeeds or fails.  The physnodes array
always contains the physical topology of the machine if CONFIG_NUMA_EMU
is enabled and can be used to setup the correct node-to-cpumask mappings
in all cases since setup_physnodes() is called whenever the array needs
to be repopulated with the correct data.

To fake the actual mappings, numa_add_cpu() and numa_remove_cpu() are
rewritten for CONFIG_NUMA_EMU so that we first find the physical node to
which each cpu has local affinity, then iterate through all online nodes
to find the emulated nodes that have local affinity to that physical
node, and then finally map the cpu to each of those emulated nodes.
Signed-off-by: NDavid Rientjes <rientjes@google.com>
LKML-Reference: <alpine.DEB.2.00.1012221701520.3701@chino.kir.corp.google.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

This patch adds the equivalent of acpi_fake_nodes() for AMD Northbridge
platforms.  The goal is to fake the apicid-to-node mappings for NUMA
emulation so the physical topology of the machine is correctly maintained
within the kernel.

This change also fakes proximity domains for both ACPI and k8 code so the
physical distance between emulated nodes is maintained via
node_distance().  This exports the correct distances via
/sys/devices/system/node/.../distance based on the underlying topology.

A new helper function, fake_physnodes(), is introduced to correctly
invoke the correct NUMA code to fake these two mappings based on the
system type.  If there is no underlying NUMA configuration, all cpus are
mapped to node 0 for local distance.

Since acpi_fake_nodes() is no longer called with CONFIG_ACPI_NUMA, it's
prototype can be removed from the header file for such a configuration.
Signed-off-by: NDavid Rientjes <rientjes@google.com>
LKML-Reference: <alpine.DEB.2.00.1012221701360.3701@chino.kir.corp.google.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

Both acpi_get_nodes() and amd_get_nodes() are only necessary when
CONFIG_NUMA_EMU is enabled, so avoid compiling them when the option is
disabled.
Signed-off-by: NDavid Rientjes <rientjes@google.com>
LKML-Reference: <alpine.DEB.2.00.1012221701210.3701@chino.kir.corp.google.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

This patch changes the minimum fake node size from 64MB to 32MB so it is
possible to test NUMA code at a greater scale on smaller machines
(64 nodes on a 2G machine, 1024 nodes on 32G machine with
CONFIG_NODES_SHIFT=10).
Signed-off-by: NDavid Rientjes <rientjes@google.com>
LKML-Reference: <alpine.DEB.2.00.1012221700590.3701@chino.kir.corp.google.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

Recent Intel new system have different order in MADT, aka will list all thread0
at first, then all thread1.
But SRAT table still old order, it will list cpus in one socket all together.

If the user have compiled limited NR_CPUS or boot with nr_cpus=, could have missed
to put some cpus apic id to node mapping into apicid_to_node[].

for example for 4 sockets system with 64 cpus with nr_cpus=32 will get crash...

[    9.106288] Total of 32 processors activated (136190.88 BogoMIPS).
[    9.235021] divide error: 0000 [#1] SMP
[    9.235315] last sysfs file:
[    9.235481] CPU 1
[    9.235592] Modules linked in:
[    9.245398]
[    9.245478] Pid: 2, comm: kthreadd Not tainted 2.6.37-rc1-tip-yh-01782-ge92ef79-dirty #274      /Sun Fire x4800
[    9.265415] RIP: 0010:[<ffffffff81075a8f>]  [<ffffffff81075a8f>] select_task_rq_fair+0x4f0/0x623
...
[    9.645938] RIP  [<ffffffff81075a8f>] select_task_rq_fair+0x4f0/0x623
[    9.665356]  RSP <ffff88103f8d1c40>
[    9.665568] ---[ end trace 2296156d35fdfc87 ]---

So let just parse all cpu entries in SRAT.

Also add apicid checking with MAX_LOCAL_APIC, in case We could out of boundaries of
apicid_to_node[].

it fixes following bug too.
https://bugzilla.kernel.org/show_bug.cgi?id=22662

-v2: expand to 32bit according to hpa
   need to add MAX_LOCAL_APIC for 32bit
Reported-and-Tested-by: NWu Fengguang <fengguang.wu@intel.com>
Reported-by: NBjorn Helgaas <bjorn.helgaas@hp.com>
Tested-by: NMyron Stowe <myron.stowe@hp.com>
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
LKML-Reference: <4D0AD486.9020704@kernel.org>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

We should use MAX_LOCAL_APIC for max apic ids and MAX_APICS as number
of local apics.

Also apic_version[] array should use MAX_LOCAL_APICs.
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
LKML-Reference: <4D0AD464.2020408@kernel.org>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>