Embedding percpu first chunk allocator can now handle very sparse unit
mapping.  Use embedding allocator instead of lpage for 64bit NUMA.
This removes extra TLB pressure and the need to do complex and fragile
dancing when changing page attributes.

For 32bit, using very sparse unit mapping isn't a good idea because
the vmalloc space is very constrained.  32bit NUMA machines aren't
exactly the focus of optimization and it isn't very clear whether
lpage performs better than page.  Use page first chunk allocator for
32bit NUMAs.

As this leaves setup_pcpu_*() functions pretty much empty, fold them
into setup_per_cpu_areas().
Signed-off-by: NTejun Heo <tj@kernel.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Andi Kleen <andi@firstfloor.org>

Now that percpu core can handle very sparse units, given that vmalloc
space is large enough, embedding first chunk allocator can use any
memory to build the first chunk.  This patch teaches
pcpu_embed_first_chunk() about distances between cpus and to use
alloc/free callbacks to allocate node specific areas for each group
and use them for the first chunk.

This brings the benefits of embedding allocator to NUMA configurations
- no extra TLB pressure with the flexibility of unified dynamic
allocator and no need to restructure arch code to build memory layout
suitable for percpu.  With units put into atom_size aligned groups
according to cpu distances, using large page for dynamic chunks is
also easily possible with falling back to reuglar pages if large
allocation fails.

Embedding allocator users are converted to specify NULL
cpu_distance_fn, so this patch doesn't cause any visible behavior
difference.  Following patches will convert them.
Signed-off-by: NTejun Heo <tj@kernel.org>

Currently units are mapped sequentially into address space.  This
patch adds pcpu_unit_offsets[] which allows units to be mapped to
arbitrary offsets from the chunk base address.  This is necessary to
allow sparse embedding which might would need to allocate address
ranges and memory areas which aren't aligned to unit size but
allocation atom size (page or large page size).  This also simplifies
things a bit by removing the need to calculate offset from unit
number.

With this change, there's no need for the arch code to know
pcpu_unit_size.  Update pcpu_setup_first_chunk() and first chunk
allocators to return regular 0 or -errno return code instead of unit
size or -errno.
Signed-off-by: NTejun Heo <tj@kernel.org>
Cc: David S. Miller <davem@davemloft.net>

Till now, non-linear cpu->unit map was expressed using an integer
array which maps each cpu to a unit and used only by lpage allocator.
Although how many units have been placed in a single contiguos area
(group) is known while building unit_map, the information is lost when
the result is recorded into the unit_map array.  For lpage allocator,
as all allocations are done by lpages and whether two adjacent lpages
are in the same group or not is irrelevant, this didn't cause any
problem.  Non-linear cpu->unit mapping will be used for sparse
embedding and this grouping information is necessary for that.

This patch introduces pcpu_alloc_info which contains all the
information necessary for initializing percpu allocator.
pcpu_alloc_info contains array of pcpu_group_info which describes how
units are grouped and mapped to cpus.  pcpu_group_info also has
base_offset field to specify its offset from the chunk's base address.
pcpu_build_alloc_info() initializes this field as if all groups are
allocated back-to-back as is currently done but this will be used to
sparsely place groups.

pcpu_alloc_info is a rather complex data structure which contains a
flexible array which in turn points to nested cpu_map arrays.

* pcpu_alloc_alloc_info() and pcpu_free_alloc_info() are provided to
  help dealing with pcpu_alloc_info.

* pcpu_lpage_build_unit_map() is updated to build pcpu_alloc_info,
  generalized and renamed to pcpu_build_alloc_info().
  @cpu_distance_fn may be NULL indicating that all cpus are of
  LOCAL_DISTANCE.

* pcpul_lpage_dump_cfg() is updated to process pcpu_alloc_info,
  generalized and renamed to pcpu_dump_alloc_info().  It now also
  prints which group each alloc unit belongs to.

* pcpu_setup_first_chunk() now takes pcpu_alloc_info instead of the
  separate parameters.  All first chunk allocators are updated to use
  pcpu_build_alloc_info() to build alloc_info and call
  pcpu_setup_first_chunk() with it.  This has the side effect of
  packing units for sparse possible cpus.  ie. if cpus 0, 2 and 4 are
  possible, they'll be assigned unit 0, 1 and 2 instead of 0, 2 and 4.

* x86 setup_pcpu_lpage() is updated to deal with alloc_info.

* sparc64 setup_per_cpu_areas() is updated to build alloc_info.

Although the changes made by this patch are pretty pervasive, it
doesn't cause any behavior difference other than packing of sparse
cpus.  It mostly changes how information is passed among
initialization functions and makes room for more flexibility.
Signed-off-by: NTejun Heo <tj@kernel.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: David Miller <davem@davemloft.net>

pcpu_fc_alloc_fn_t is about to see more interesting usage, add @align
parameter.
Signed-off-by: NTejun Heo <tj@kernel.org>

First chunk allocators assume percpu areas have been linked using one
of PERCPU_*() macros and depend on __per_cpu_load symbol defined by
those macros, so there isn't much point in passing in static area size
explicitly when it can be easily calculated from __per_cpu_start and
__per_cpu_end.  Drop @static_size from all percpu first chunk
allocators and helpers.
Signed-off-by: NTejun Heo <tj@kernel.org>

Now that all first chunk allocators are in mm/percpu.c, it makes sense
to make generalize percpu_alloc kernel parameter.  Define PCPU_FC_*
and set pcpu_chosen_fc using early_param() in mm/percpu.c.  Arch code
can use the set value to determine which first chunk allocator to use.
Signed-off-by: NTejun Heo <tj@kernel.org>

There's no need to build unused first chunk allocators in.  Define
CONFIG_NEED_PER_CPU_*_FIRST_CHUNK and let archs enable them
selectively.
Signed-off-by: NTejun Heo <tj@kernel.org>

Page size isn't always 4k depending on arch and configuration.  Rename
4k first chunk allocator to page.
Signed-off-by: NTejun Heo <tj@kernel.org>
Cc: David Howells <dhowells@redhat.com>

percpu code has been assuming num_possible_cpus() == nr_cpu_ids which
is incorrect if cpu_possible_map contains holes.  This causes percpu
code to access beyond allocated memories and vmalloc areas.  On a
sparc64 machine with cpus 0 and 2 (u60), this triggers the following
warning or fails boot.

 WARNING: at /devel/tj/os/work/mm/vmalloc.c:106 vmap_page_range_noflush+0x1f0/0x240()
 Modules linked in:
 Call Trace:
  [00000000004b17d0] vmap_page_range_noflush+0x1f0/0x240
  [00000000004b1840] map_vm_area+0x20/0x60
  [00000000004b1950] __vmalloc_area_node+0xd0/0x160
  [0000000000593434] deflate_init+0x14/0xe0
  [0000000000583b94] __crypto_alloc_tfm+0xd4/0x1e0
  [00000000005844f0] crypto_alloc_base+0x50/0xa0
  [000000000058b898] alg_test_comp+0x18/0x80
  [000000000058dad4] alg_test+0x54/0x180
  [000000000058af00] cryptomgr_test+0x40/0x60
  [0000000000473098] kthread+0x58/0x80
  [000000000042b590] kernel_thread+0x30/0x60
  [0000000000472fd0] kthreadd+0xf0/0x160
 ---[ end trace 429b268a213317ba ]---

This patch fixes generic percpu functions and sparc64
setup_per_cpu_areas() so that they handle sparse cpu_possible_map
properly.

Please note that on x86, cpu_possible_map() doesn't contain holes and
thus num_possible_cpus() == nr_cpu_ids and this patch doesn't cause
any behavior difference.
Signed-off-by: NTejun Heo <tj@kernel.org>
Acked-by: NDavid S. Miller <davem@davemloft.net>
Cc: Ingo Molnar <mingo@elte.hu>

Johannes Stezenbach reported that his Pentium-M based
laptop does not have the local APIC enabled by default,
and hence perfcounters do not get initialized.

Add a fallback for this case: allow non-sampled counters
and return with an error on sampled counters. This allows
'perf stat' to work out of box - and allows 'perf top'
and 'perf record' to fall back on a hrtimer based sampling
method.

( Passing 'lapic' on the boot line will allow hardware
  sampling to occur - but if the APIC is disabled
  permanently by the hardware then this fallback still
  allows more systems to use perfcounters. )

Also decouple perfcounter support from X86_LOCAL_APIC.

-v2: fix typo breaking counters on all other systems ...
Reported-by: NJohannes Stezenbach <js@sig21.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <new-submission>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Kernel is broken for x86 CPUs without CPUID since 2.6.28. It
crashes with NULL pointer dereference in identify_cpu():

766        generic_identify(c);
767
768-->     if (this_cpu->c_identify)
769               this_cpu->c_identify(c);

this_cpu is NULL. This is because it's only initialized in
get_cpu_vendor() function, which is not called if the CPU has
no CPUID instruction.
Signed-off-by: NOndrej Zary <linux@rainbow-software.org>
LKML-Reference: <200908112000.15993.linux@rainbow-software.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Due to an erratum with certain AMD Athlon 64 processors, the
BIOS may need to force enable the LAHF_LM capability.
Unfortunately, in at least one case, the BIOS does this even
for processors that do not support the functionality.

Add a specific check that will clear the feature bit for
processors known not to support the LAHF/SAHF instructions.
Signed-off-by: NKevin Winchester <kjwinchester@gmail.com>
Acked-by: NBorislav Petkov <petkovbb@googlemail.com>
LKML-Reference: <4A80A5AD.2000209@gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Johannes Stezenbach reported that 'perf stat' does not count
cache-miss and cache-references events on his Pentium-M based
laptop.

This is because we left them blank in p6_perfmon_event_map[],
fill them in.
Reported-by: NJohannes Stezenbach <js@sig21.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <new-submission>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Instead of this garbled bootup on UP Pentium-M systems:

[    0.015048] Performance Counters:
[    0.016004] no Local APIC, try rebooting with lapicno PMU driver, software counters only.

Print:

[    0.015050] Performance Counters:
[    0.016004] no APIC, boot with the "lapic" boot parameter to force-enable it.
[    0.017003] no PMU driver, software counters only.

Cf: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <new-submission>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

My Latitude d630 seems to be handling thermal events in SMI by
lowering the max frequency of the CPU till it cools down but
still leaks the "everything is normal" events.

This spams the console and with high priority printks.

Adjust therm_throt driver to only print messages about the fact
that temperatire returned back to normal when leaving the
throttling state.

Also lower the severity of "back to normal" message from
KERN_CRIT to KERN_INFO.
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
LKML-Reference: <20090810051513.0558F526EC9@mailhub.coreip.homeip.net>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Reboot does not work on my MacBook Pro 13 inch (MacBookPro5,5)
too. It seems all unibody MacBook and MacBookPro require
PCI reboot handling, i guess.

Following model/machine ID list shows unibody MacBook/Pro have
the 5 series of model number:

   http://www.everymac.com/systems/by_capability/macs-by-machine-model-machine-id.htmlSigned-off-by: NShunichi Fuji <palglowr@gmail.com>
Cc: Ozan Çağlayan <ozan@pardus.org.tr>
LKML-Reference: <30046e3b0908101134p6487ddbftd8776e4ddef204be@mail.gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Wei Chong Tan reported a fast-PIT-calibration corner-case:

| pit_expect_msb() is vulnerable to SMI disturbance corner case
| in some platforms which causes /proc/cpuinfo to show wrong
| CPU MHz value when quick_pit_calibrate() jumps to success
| section.

I think that the real issue isn't even an SMI - but the fact
that in the very last iteration of the loop, there's no
serializing instruction _after_ the last 'rdtsc'. So even in
the absense of SMI's, we do have a situation where the cycle
counter was read without proper serialization.

The last check should be done outside the outer loop, since
_inside_ the outer loop, we'll be testing that the PIT has
the right MSB value has the right value in the next iteration.

So only the _last_ iteration is special, because that's the one
that will not check the PIT MSB value any more, and because the
final 'get_cycles()' isn't serialized.

In other words:

 - I'd like to move the PIT MSB check to after the last
   iteration, rather than in every iteration

 - I think we should comment on the fact that it's also a
   serializing instruction and so 'fences in' the TSC read.

Here's a suggested replacement.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
Reported-by: N"Tan, Wei Chong" <wei.chong.tan@intel.com>
Tested-by: N"Tan, Wei Chong" <wei.chong.tan@intel.com>
LKML-Reference: <B28277FD4E0F9247A3D55704C440A140D5D683F3@pgsmsx504.gar.corp.intel.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

If the vendor name (from c16) can be longer than 100 bytes (or missing a
terminating null), then the null is written past the end of vendor[].

Found with Parfait, http://research.sun.com/projects/parfait/Signed-off-by: NRoel Kluin <roel.kluin@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NH. Peter Anvin <hpa@zytor.com>
Cc: Huang Ying <ying.huang@intel.com>

MacBookPro5,1 is not able to reboot unless reboot=pci is set.
This patch forces it through a DMI quirk specific to this
device.
Signed-off-by: NOzan Çağlayan <ozan@pardus.org.tr>
LKML-Reference: <1249403971-6543-1-git-send-email-ozan@pardus.org.tr>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

found a system where x2apic reports an MSI-X irq initialization
failure:

[  302.859446] igbvf 0000:81:10.4: enabling device (0000 -> 0002)
[  302.874369] igbvf 0000:81:10.4: using 64bit DMA mask
[  302.879023] igbvf 0000:81:10.4: using 64bit consistent DMA mask
[  302.894386] igbvf 0000:81:10.4: enabling bus mastering
[  302.898171] igbvf 0000:81:10.4: setting latency timer to 64
[  302.914050] reserve_memtype added 0xefb08000-0xefb0c000, track uncached-minus, req uncached-minus, ret uncached-minus
[  302.933839] reserve_memtype added 0xefb28000-0xefb29000, track uncached-minus, req uncached-minus, ret uncached-minus
[  302.940367]   alloc irq_desc for 265 on node 4
[  302.956874]   alloc kstat_irqs on node 4
[  302.959452] alloc irq_2_iommu on node 0
[  302.974328] igbvf 0000:81:10.4: irq 265 for MSI/MSI-X
[  302.977778]   alloc irq_desc for 266 on node 4
[  302.980347]   alloc kstat_irqs on node 4
[  302.995312] free_memtype request 0xefb28000-0xefb29000
[  302.998816] igbvf 0000:81:10.4: Failed to initialize MSI-X interrupts.

... it turns out that when trying to enable MSI-X,
__assign_irq_vector(new, cfg_new, apic->target_cpus()) can not
get vector because for x2apic target-cpus returns cpumask_of(0)

Update that to online_mask like xapic.
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
Acked-by: NSuresh Siddha <suresh.b.siddha@intel.com>
LKML-Reference: <4A785AFF.3050902@kernel.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Otherwise the host can spend too long traversing an rmap chain, which
happens under a spinlock.

Cc: stable@kernel.org
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

We have to disable preemption and IRQs on every exit from
handle_invalid_guest_state, otherwise we generate at least a
preempt_disable imbalance.
Signed-off-by: NJan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Release and re-acquire preemption and IRQ lock in the same order as
vcpu_enter_guest does.
Signed-off-by: NJan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

kvm_mmu_change_mmu_pages mishandles the case where n_alloc_mmu_pages is
smaller then n_free_mmu_pages, by not checking if the result of
the subtraction is negative.

Its a valid condition which can happen if a large number of pages has
been recently freed.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

If a migrated vcpu matches the asid_generation value of the target pcpu,
there will be no TLB flush via TLB_CONTROL_FLUSH_ALL_ASID.

The check for vcpu.cpu in pre_svm_run is meaningless since svm_vcpu_load
already updated it on schedule in.

Such vcpu will VMRUN with stale TLB entries.

Based on original patch from Joerg Roedel (http://patchwork.kernel.org/patch/10021/)
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Acked-by: NJoerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Do not allow invalid memory types in MTRR/PAT (generating a #GP
otherwise).
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Fix division by zero triggered by latch count command on uninitialized
counter.

Cc: stable@kernel.org
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

So far, KVM copied the emulated_msrs (only MSR_IA32_MISC_ENABLE) to a
wrong address in user space due to broken pointer arithmetic. This
caused subtle corruption up there (missing MSR_IA32_MISC_ENABLE had
probably no practical relevance). Moreover, the size check for the
user-provided kvm_msr_list forgot about emulated MSRs.

Cc: stable@kernel.org
Signed-off-by: NJan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

With CONFIG_STACK_PROTECTOR turned on, VMI doesn't boot with
more than one processor. The problem is with the gs value not
being initialized correctly when registering the secondary
processor for VMI's case.

The patch below initializes the gs value for the AP to
__KERNEL_STACK_CANARY. Without this the secondary processor
keeps on taking a GP on every gs access.
Signed-off-by: NAlok N Kataria <akataria@vmware.com>
Cc: <stable@kernel.org> # for v2.6.30.x
LKML-Reference: <1249425262.18955.40.camel@ank32.eng.vmware.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

The following fix was initially inspired by David Howells fix
few days back:

  http://lkml.org/lkml/2009/7/9/109

However, Ingo disapproves such fixes as it's dangerous (it can
hide future, relevant warnings) - in something as
performance-uncritical.

So, initialize 'err' to '0' to work around a GCC false positive
warning:

  http://lkml.org/lkml/2009/7/18/89

Signed-off-by: Subrata Modak<subrata@linux.vnet.ibm.com>
Cc: Sachin P Sant <sachinp@linux.vnet.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
LKML-Reference: <20090721023226.31855.67236.sendpatchset@subratamodak.linux.ibm.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

In uv_setup_irq(), the call to create_irq() initially assigns
IRQ vectors to cpu 0. The subsequent call to
assign_irq_vector() in arch_enable_uv_irq() migrates the IRQ to
another cpu and frees the cpu 0 vector - at least it will be
freed as soon as the "IRQ move" completes.

arch_enable_uv_irq() needs to send a cleanup IPI to complete
the IRQ move. Otherwise, assignment of GRU interrupts on large
systems (>200 cpus) will exhaust the cpu 0 interrupt vectors
and initialization of the GRU driver will fail.
Signed-off-by: NJack Steiner <steiner@sgi.com>
LKML-Reference: <20090720142840.GA8885@sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

With VMALLOC_END included in the calculation of MAXMEM (as of
2.6.28) it is no longer correct to also bump __VMALLOC_RESERVE
in reserve_top_address(). Doing so results in needlessly small
lowmem.
Signed-off-by: NJan Beulich <jbeulich@novell.com>
LKML-Reference: <4A71DD2A020000780000D482@vpn.id2.novell.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

One system has socket 1 come up as BSP.

kexeced kernel reports BSP as:

[    1.524550] Initializing cgroup subsys cpuacct
[    1.536064] initial_apicid:20
[    1.537135] ht_mask_width:1
[    1.538128] core_select_mask:f
[    1.539126] core_plus_mask_width:5
[    1.558479] CPU: Physical Processor ID: 0
[    1.559501] CPU: Processor Core ID: 0
[    1.560539] CPU: L1 I cache: 32K, L1 D cache: 32K
[    1.579098] CPU: L2 cache: 256K
[    1.580085] CPU: L3 cache: 24576K
[    1.581108] CPU 0/0x20 -> Node 0
[    1.596193] CPU 0 microcode level: 0xffff0008

It doesn't have correct physical processor id and will get an
error:

[   38.840859] CPU0 attaching sched-domain:
[   38.848287]  domain 0: span 0,8,72 level SIBLING
[   38.851151]   groups: 0 8 72
[   38.858137]   domain 1: span 0,8-15,72-79 level MC
[   38.868944]    groups: 0,8,72 9,73 10,74 11,75 12,76 13,77 14,78 15,79
[   38.881383] ERROR: parent span is not a superset of domain->span
[   38.890724]    domain 2: span 0-7,64-71 level CPU
[   38.899237] ERROR: domain->groups does not contain CPU0
[   38.909229]     groups: 8-15,72-79
[   38.912547] ERROR: groups don't span domain->span
[   38.919665]     domain 3: span 0-127 level NODE
[   38.930739]      groups: 0-7,64-71 8-15,72-79 16-23,80-87 24-31,88-95 32-39,96-103 40-47,104-111 48-55,112-119 56-63,120-127

it turns out: we can not use current_cpu_data in phys_pgd_id
for x2apic.

identify_boot_cpu() is called by check_bugs() before
smp_prepare_cpus() and till smp_prepare_cpus() current_cpu_data
for bsp is assigned with boot_cpu_data.

Just make phys_pkg_id for x2apic is aligned to xapic.
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
Acked-by: NSuresh Siddha <suresh.b.siddha@intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
LKML-Reference: <4A6ADD0D.10002@kernel.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Change SGI UV default apicid mode to "physical". This is
required to match settings in the UV hub chip.
Signed-off-by: NJack Steiner <steiner@sgi.com>
LKML-Reference: <20090727143856.GA8905@sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

The UV chipset automatically supplies the upper bits on nodes
being referenced by MMR accesses. These bit can be deleted from
the hub addressing macros.
Signed-off-by: NJack Steiner <steiner@sgi.com>
LKML-Reference: <20090727143808.GA8076@sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

The UV BIOS has added additional MMR ranges that are mapped via
EFI virtual mode mappings. These ranges should be deleted from
ranges mapped by uv_system_init().
Signed-off-by: NJack Steiner <steiner@sgi.com>
Cc: linux-mm@kvack.org
LKML-Reference: <20090727143656.GA7698@sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

UV blades may not have any blade-local memory. Add a field
(nid) to the UV blade structure to indicates whether the node
has local memory. This is needed by the GRU driver (pushed
separately).
Signed-off-by: NJack Steiner <steiner@sgi.com>
Cc: linux-mm@kvack.org
LKML-Reference: <20090727143507.GA7006@sgi.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

From Gabe Black in bugzilla 13888:

native_save_fl is implemented as follows:

  11static inline unsigned long native_save_fl(void)
  12{
  13        unsigned long flags;
  14
  15        asm volatile("# __raw_save_flags\n\t"
  16                     "pushf ; pop %0"
  17                     : "=g" (flags)
  18                     : /* no input */
  19                     : "memory");
  20
  21        return flags;
  22}

If gcc chooses to put flags on the stack, for instance because this is
inlined into a larger function with more register pressure, the offset
of the flags variable from the stack pointer will change when the
pushf is performed. gcc doesn't attempt to understand that fact, and
address used for pop will still be the same. It will write to
somewhere near flags on the stack but not actually into it and
overwrite some other value.

I saw this happen in the ide_device_add_all function when running in a
simulator I work on. I'm assuming that some quirk of how the simulated
hardware is set up caused the code path this is on to be executed when
it normally wouldn't.

A simple fix might be to change "=g" to "=r".
Reported-by: NGabe Black <spamforgabe@umich.edu>
Signed-off-by: NH. Peter Anvin <hpa@zytor.com>
Cc: Stable Team <stable@kernel.org>

Make rdmsr_on_cpus/wrmsr_on_cpus execute on the current CPU only if it
is in the supplied bitmask.
Signed-off-by: NBorislav Petkov <borislav.petkov@amd.com>
Signed-off-by: NH. Peter Anvin <hpa@zytor.com>