Move X86_ELAN (old, AMD based web-boxes) from the subarchitecture
menu to the non-standard-platform section.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Introduce a Y/N Kconfig option for non-PC x86 platforms.

Make VisWS, RDC321 and SGI/UV depend on this.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager has this ISA quirk (because Voyager has no ISA support):

 config ISA
 	bool "ISA support"
	depends on !X86_VOYAGER

There's a ton of x86 hardware that does not support ISA, and because
most ISA drivers cannot auto-detect in a safe way, the convention in
the kernel has always been to not enable ISA drivers if they are not
needed.

Voyager users can do likewise - no need for a Kconfig quirk.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager has this PM/ACPI Kconfig quirk:

 menu "Power management and ACPI options"
	depends on !X86_VOYAGER

Most of the PM features are auto-detect so they should be safe to run
on just about any hardware. (If not, those instances need fixing.)

In any case, if a kernel is built for Voyager, the power management
options can be disabled.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager has this Kconfig quirk:

 config HOTPLUG_CPU
 	bool "Support for hot-pluggable CPUs"
	depends on SMP && HOTPLUG && !X86_VOYAGER

But this exception will be moot once Voyager starts using the
generic x86 code.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

If no MCE code is desired on Voyager hw then the solution
is to turn them off in the .config - and to extend the MCE
code to not initialize on Voyager.

Remove the build-time quirk.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

The lapic/ioapic code properly auto-detects and is safe to run on CPUs that
have no local APIC. (or which have their lapic turned off in the hardware)
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Remove this Kconfig quirk:

 config PARAVIRT
 	bool "Enable paravirtualization code"
	depends on !X86_VOYAGER
 	help

Voyager support built into a kernel does not preclude paravirt support.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager has this quirk currently:

 config KVM_GUEST
 	bool "KVM Guest support"
 	select PARAVIRT
	depends on !X86_VOYAGER

Voyager support built into a kernel image does not exclude
KVM paravirt guest support - so remove this quirk.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager has this build-time quirk to exclude KVM_CLOCK:

 	bool "KVM paravirtualized clock"
 	select PARAVIRT
 	select PARAVIRT_CLOCK
	depends on !X86_VOYAGER

Voyager support built into a kernel image does not exclude
KVM paravirt clock support - so remove this quirk.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

x86/Voyager has this build-time quirk:

 	bool "VMI Guest support"
 	select PARAVIRT
 	depends on X86_32
	depends on !X86_VOYAGER

Since VMI is auto-detected (and Voyager will be auto-detected) there's no
reason for this quirk.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

VisWS has this quirk currently:

 config X86_VISWS
 	bool "SGI 320/540 (Visual Workstation)"
	depends on X86_32 && PCI && !X86_VOYAGER && X86_MPPARSE && PCI_GODIRECT

The !Voyager quirk is unnecessary.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

x86/Voyager had this Kconfig quirk:

 config X86_FIND_SMP_CONFIG
	def_bool y
	depends on X86_MPPARSE || X86_VOYAGER

Which splits off the find_smp_config() callback into a build-time quirk.

Voyager should use the existing x86_quirks.mach_find_smp_config() callback
to introduce SMP-config quirks. NUMAQ-32 and VISWS already use this.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager has this Kconfig quirk:

config X86_BIOS_REBOOT
	bool
	depends on !X86_VOYAGER
	default y

Voyager should use the existing machine_ops.emergency_restart reboot
quirk mechanism instead of a build-time quirk.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager has this Kconfig quirk:

	depends on (X86_32 && !X86_VOYAGER) || X86_64

That is unnecessary as HT support is CPUID driven and explicitly
 enumerated.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

x86/Voyager can boot on non-zero processors. While that can probably
be fixed by properly remapping the physical CPU IDs, keep boot_cpu_id
for now for easier transition - and expand it to all of x86.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

The x86/Voyager subarch used to have this distinction between
 'x86 SMP support' and 'Voyager SMP support':

 config X86_SMP
	bool
	depends on SMP && ((X86_32 && !X86_VOYAGER) || X86_64)

This is a pointless distinction - Voyager can (and already does) use
smp_ops to implement various SMP quirks it has - and it can be extended
more to cover all the specialities of Voyager.

So remove this complication in the Kconfig space.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager has this Kconfig quirk for suspend/resume:

 config ARCH_SUSPEND_POSSIBLE
 	def_bool y
 	depends on !X86_VOYAGER

The proper mechanism to not suspend on a piece of hardware to disable
CONFIG_SUSPEND. Remove the quirk.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager has this hibernation quirk:

 config ARCH_HIBERNATION_POSSIBLE
 	def_bool y
	depends on !SMP || !X86_VOYAGER

Hibernation is a generic facility provided on all x86 platforms. If it
is buggy on Voyager then that bug should be fixed - not worked around.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

x86/Voyager has this KGDB quirk:

	select HAVE_ARCH_KGDB if !X86_VOYAGER

This is completely pointless - there's nothing in KGDB that cannot work
on Voyager. Remove it.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Voyager and other subarchitectures have this Kconfig quirk:

 	select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64)

This is unnecessary, as KVM cleanly detects based on CPUID capabilities.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

x86/Voyager has this quirk for SCx200 support:

 config SCx200
 	tristate "NatSemi SCx200 support"
	depends on !X86_VOYAGER

Remove it - Voyager users can disable drivers they dont need.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Remove Voyager Kconfig quirk: just like any other hardware platform
users of Voyager systems can configure in the hardware drivers they need.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

x86/Voyager does not build right now and it's unclear whether it will
be cleaned up and ported to the subarch-less 32-bit x86 code - so disable
it for now.

If it's fixed we'll re-enable it - or remove it after some time. There's
a very low number of systems running development kernels on x86/Voyager
currently. (one or two on the whole planet)
Signed-off-by: NIngo Molnar <mingo@elte.hu>

CONFIG_BROKEN has been removed from the upstream kernel years ago,
but X86_VOYAGER still had a stale reference to it - remove it.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Impact: standardize all x86 platforms on same setup code

With the preceding changes, Voyager can use the same per-cpu setup
code as all the other x86 platforms.
Signed-off-by: NBrian Gerst <brgerst@gmail.com>
Signed-off-by: NTejun Heo <tj@kernel.org>

Make X86 SGI Ultraviolet support configurable. Saves about 13K of text size
on my modest config.

   text    data     bss     dec     hex filename
6770537 1158680  694356 8623573  8395d5 vmlinux
6757492 1157664  694228 8609384  835e68 vmlinux.nouv
Signed-off-by: NNick Piggin <npiggin@suse.de>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

CONFIG_FRAME_POINTERS=y results in much better debug info for the
kernel (clear and precise backtraces), with the only drawback being
a ~1% increase in kernel size.

So offer it unconditionally and enable it by default.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Impact: adds new Kconfig entry
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

These two IOMMUs can implement the current version of this API. So
select the API if one or both of these IOMMU drivers is selected.
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: improve help text
Signed-off-by: NIngo Molnar <mingo@elte.hu>

New feature - lets disable it by default first - can flip it around
later.
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Impact: reduce kconfig variable scope and clean up

Bartlomiej pointed out that the config dependencies and comments are not right.

update it depend to NUMA, and fix some comments
Reported-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Impact: cleanup
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Signed-off-by: NMike Travis <travis@sgi.com>

Impact: activates new off-stack cpumask code on MAXSMP (non-default) x86 configs

Set MAXSMP to enable CONFIG_CPUMASK_OFFSTACK which moves cpumask's off
the stack (and in structs) when using cpumask_var_t.
Signed-off-by: NMike Travis <travis@sgi.com>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Acked-by: NIngo Molnar <mingo@elte.hy>

Impact: build fix

make intr_remapping.c to include smp.h, so could use boot_cpu_id there

also remove old change that disabling sparseirq with !SMP
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Impact: improve NUMA handling by migrating irq_desc on smp_affinity changes

if CONFIG_NUMA_MIGRATE_IRQ_DESC is set:

-  make irq_desc to go with affinity aka irq_desc moving etc
-  call move_irq_desc in irq_complete_move()
-  legacy irq_desc is not moved, because they are allocated via static array

for logical apic mode, need to add move_desc_in_progress_in_same_domain,
otherwise it will not be moved ==> also could need two phases to get
irq_desc moved.
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Impact: reduce bug table size

This allows reducing the bug table size by half. Perhaps there are
other 64-bit architectures that could also make use of this.
Signed-off-by: NJan Beulich <jbeulich@novell.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Ingo Molnar wrote:

>>>  drivers/pci/intr_remapping.c: In function 'irq_2_iommu_alloc':
>>>  drivers/pci/intr_remapping.c:72: error: 'boot_cpu_id' undeclared (first use in this function)
>>>  drivers/pci/intr_remapping.c:72: error: (Each undeclared identifier is reported only once
>>>  drivers/pci/intr_remapping.c:72: error: for each function it appears in.)

sparseirq should only be used with SMP for now.

Impact: new feature

Problem on distro kernels: irq_desc[NR_IRQS] takes megabytes of RAM with
NR_CPUS set to large values. The goal is to be able to scale up to much
larger NR_IRQS value without impacting the (important) common case.

To solve this, we generalize irq_desc[NR_IRQS] to an (optional) array of
irq_desc pointers.

When CONFIG_SPARSE_IRQ=y is used, we use kzalloc_node to get irq_desc,
this also makes the IRQ descriptors NUMA-local (to the site that calls
request_irq()).

This gets rid of the irq_cfg[] static array on x86 as well: irq_cfg now
uses desc->chip_data for x86 to store irq_cfg.
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>