The [user space] interface does not filter out offline cpus. It merily
guarantees that the mask contains at least one online cpu.

So the selector in the irq chip implementation needs to make sure to
pick only an online cpu because otherwise:

     Offline Core 1
     Set affinity to 0xe (is valid due to online mask 0xd)
     cpumask_first will pick core 1, which is offline
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-mips@linux-mips.org
Link: http://lkml.kernel.org/r/20140304203100.744800502@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

The [user space] interface does not filter out offline cpus. It merily
guarantees that the mask contains at least one online cpu.

So the selector in the irq chip implementation needs to make sure to
pick only an online cpu because otherwise:

     Offline Core 1
     Set affinity to 0xe (is valid due to online mask 0xd)
     cpumask_first will pick core 1, which is offline
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: ia64 <linux-ia64@vger.kernel.org>
Link: http://lkml.kernel.org/r/20140304203100.650414633@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

Reported-by: NIngo Molnar <mingo@kernel.org>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Compiling last minute changes without setting the proper config
options is not really clever.
Reported-by: NFengguang Wu <fengguang.wu@intel.com>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

commit: 8f945a33 (genirq: Move kstat_incr_irqs_this_cpu() to core)
unearthed the following:

arch/m68k/kernel/ints.c:34:15: error: variable 'auto_irq_chip' has initializer but incomplete type
arch/m68k/kernel/ints.c:35:2: error: unknown field 'name' specified in initializer
arch/m68k/kernel/ints.c:35:2: warning: excess elements in struct initializer [enabled by default]

The reason is that this file requires linux/irq.h and magically
pulled that in via linux/kernel_stat.h

The commit above got rid of the pointless include of linux/irq.h in
linux/kernel_stat.h and therefor broke the build.

Include linux/irq.h

Reported-by: fengguang.wu@intel.com
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Commit b8a9a11b (powerpc: eeh: Kill another abuse of irq_desc) is
missing some brackets .....

It's not a good idea to write patches in grumpy mode and then forget
to at least compile test them or rely on the few eyeballs discussing
that patch to spot it.....

Reported-by: fengguang.wu@intel.com
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Gavin Shan <shangw@linux.vnet.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: ppc <linuxppc-dev@lists.ozlabs.org>

commit: 8f945a33 (genirq: Move kstat_incr_irqs_this_cpu() to core)
unearthed the following:

   arch/s390/kernel/irq.c: In function 'init_IRQ':
>> arch/s390/kernel/irq.c:93:2: error: implicit declaration of function 'irq_reserve_irqs'
[-Werror=implicit-function-declaration]
....
   cc1: some warnings being treated as errors
--
   drivers/s390/cio/cio.c: In function 'init_cio_interrupts':
>> drivers/s390/cio/cio.c:594:2: error: implicit declaration of function
'irq_set_chip_and_handler' [-Werror=implicit-function-declaration]
[-Werror=implicit-function-declaration]
....
   cc1: some warnings being treated as errors

The reason is that those files require linux/irq.h and magically
pulled that in via linux/kernel_stat.h

The commit above got rid of the pointless include of linux/irq.h in
linux/kernel_stat.h and therefor broke the build.

Include linux/irq.h

Reported-by: fengguang.wu@intel.com
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: s390 <linux-s390@vger.kernel.org>

Reported-by: fengguang.wu@intel.com
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: K. Y. Srinivasan <kys@microsoft.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linuxdrivers <devel@linuxdriverproject.org>
Cc: x86 <x86@kernel.org>

Commit 1aec1696 (x86: Hyperv: Cleanup the irq mess) removed the
ability to build the hyperv stuff as a module. Bring it back.

Reported-by: fengguang.wu@intel.com
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: K. Y. Srinivasan <kys@microsoft.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linuxdrivers <devel@linuxdriverproject.org>
Cc: x86 <x86@kernel.org>

This patch removes the IRQF_DISABLED flag from x86 architecture
code. It's a NOOP since 2.6.35 and it will be removed one day.
Signed-off-by: NMichael Opdenacker <michael.opdenacker@free-electrons.com>
Cc: venki@google.com
Link: http://lkml.kernel.org/r/1393965305-17248-1-git-send-email-michael.opdenacker@free-electrons.comSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

This patch removes the IRQF_DISABLED flag from mn10300 architecture
code. It's a NOOP since 2.6.35 and it will be removed one day.
Signed-off-by: NMichael Opdenacker <michael.opdenacker@free-electrons.com>
Cc: srivatsa.bhat@linux.vnet.ibm.com
Cc: linux-am33-list@redhat.com
Cc: dhowells@redhat.com
Cc: yasutake.koichi@jp.panasonic.com
Link: http://lkml.kernel.org/r/1393965111-17092-1-git-send-email-michael.opdenacker@free-electrons.comSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

This patch removes the IRQF_DISABLED flag from ia64 architecture
code. It's a NOOP since 2.6.35 and it will be removed one day.
Signed-off-by: NMichael Opdenacker <michael.opdenacker@free-electrons.com>
Cc: paul.gortmaker@windriver.com
Cc: viro@zeniv.linux.org.uk
Cc: srivatsa.bhat@linux.vnet.ibm.com
Cc: andriy.shevchenko@linux.intel.com
Cc: fenghua.yu@intel.com
Cc: tony.luck@intel.com
Link: http://lkml.kernel.org/r/1393964953-17002-1-git-send-email-michael.opdenacker@free-electrons.comSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

There is no point in having an incomplete copy of irq_set_affinity()
for the hotplug irq migration code.

Use the core function instead and while at it switch to
for_each_active_irq()
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Chris Zankel <chris@zankel.net>
Cc: xtensa <linux-xtensa@linux-xtensa.org>
Link: http://lkml.kernel.org/r/20140223212736.664624945@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

There is no point in having an incomplete copy of irq_set_affinity()
for the hotplug irq migration code.

Use the core function instead.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: sh <linux-sh@vger.kernel.org>
Link: http://lkml.kernel.org/r/20140223212736.774961401@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

The vmbus/hyperv interrupt handling is another complete trainwreck and
probably the worst of all currently in tree.

If CONFIG_HYPERV=y then the interrupt delivery to the vmbus happens
via the direct HYPERVISOR_CALLBACK_VECTOR. So far so good, but:

  The driver requests first a normal device interrupt. The only reason
  to do so is to increment the interrupt stats of that device
  interrupt. For no reason it also installs a private flow handler.

  We have proper accounting mechanisms for direct vectors, but of
  course it's too much effort to add that 5 lines of code.

  Aside of that the alloc_intr_gate() is not protected against
  reallocation which makes module reload impossible.

Solution to the problem is simple to rip out the whole mess and
implement it correctly.

First of all move all that code to arch/x86/kernel/cpu/mshyperv.c and
merily install the HYPERVISOR_CALLBACK_VECTOR with proper reallocation
protection and use the proper direct vector accounting mechanism.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Acked-by: NK. Y. Srinivasan <kys@microsoft.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linuxdrivers <devel@linuxdriverproject.org>
Cc: x86 <x86@kernel.org>
Link: http://lkml.kernel.org/r/20140223212739.028307673@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

HyperV abuses a device interrupt to account for the
HYPERVISOR_CALLBACK_VECTOR.

Provide proper accounting as we have for the other vectors as well.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: K. Y. Srinivasan <kys@microsoft.com>
Cc: x86 <x86@kernel.org>
Link: http://lkml.kernel.org/r/20140223212738.681855582@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

Let the core do the irq_desc resolution.

No functional change.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: ia64 <linux-ia64@vger.kernel.org>
Link: http://lkml.kernel.org/r/20140223212738.099977064@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

Let the core do the irq_desc resolution.

No functional change.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: David Howells <dhowells@redhat.com>
Cc: mn10300 <linux-am33-list@redhat.com>
Link: http://lkml.kernel.org/r/20140223212737.751487689@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

Let the core do the irq_desc resolution.

No functional change.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Reviewed-by: NDavid Vrabel <david.vrabel@citrix.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Xen <xen-devel@lists.xenproject.org>
Cc: x86 <x86@kernel.org>
Link: http://lkml.kernel.org/r/20140223212737.869264085@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

Let the core do the irq_desc resolution.

No functional change.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: David S. Miller <davem@davemloft.net>
Link: http://lkml.kernel.org/r/20140223212737.635609567@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

Let the core do the irq_desc resolution.

No functional change.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: mips <inux-mips@linux-mips.org>
Link: http://lkml.kernel.org/r/20140223212737.517340416@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

The pm-mmp2 and pm-pxa910 power management related irq_set_wake
callbacks fiddle pointlessly with the irq actions for no reason except
for lack of understanding how the wakeup mechanism works.

On supsend the core disables all interrupts lazily, i.e. it does not
mask them at the irq controller level. So any interrupt which is
firing during suspend will mark the corresponding interrupt line as
pending. Just before the core powers down it checks whether there are
interrupts pending from interrupt lines which are marked as wakeup
sources and if so it aborts the suspend and resends the interrupts.
If there was no interrupt at this point, the cpu goes into suspend
with these interrupts unmasked.

The IRQF_NO_SUSPEND flag for interrupt actions is a totally different
mechanism. That allows the device driver to prevent the core from
disabling the interrupt despite the fact that it is not marked as a
wakeup source. This has nothing to do with the case at hand. It was
introduced for special cases where lazy disable is not possible.

Remove the nonsense along with the braindamaged boundary check. The
core code does NOT call these functions out of boundary.

Add a FIXME comment to an unhandled error path which merily printks
some useless blurb instead of returning a proper error code.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: arm <linux-arm-kernel@lists.infradead.org>
Cc: Eric Miao <eric.y.miao@gmail.com>
Cc: Haojian Zhuang <haojian.zhuang@gmail.com>
Cc: Russell King <linux@arm.linux.org.uk>
Link: http://lkml.kernel.org/r/20140223212737.214342433@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

Use the proper functions. There is no need to fiddle with irq_desc.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Acked-by: Shawn Guo <shawn.guo@linaro.org>C
Acked-by: NTony Lindgren <tony@atomide.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: arm <linux-arm-kernel@lists.infradead.org>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Link: http://lkml.kernel.org/r/20140223212737.099151500@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

commit 91150af3 (powerpc/eeh: Fix unbalanced enable for IRQ) is
another brilliant example of trainwreck engineering.

The patch "fixes" the issue of an unbalanced call to irq_enable()
which causes a prominent warning by checking the disabled state of the
interrupt line and call conditionally into the core code.

This is wrong in two aspects:

1) The warning is there to tell users, that they need to fix their
   asymetric enable/disable patterns by finding the root cause and
   solving it there.

   It's definitely not meant to work around it by conditionally
   calling into the core code depending on the random state of the irq
   line.

   Asymetric irq_disable/enable calls are a clear sign of wrong usage
   of the interfaces which have to be cured at the root and not by
   somehow hacking around it.

2) The abuse of core internal data structure instead of using the
   proper interfaces for retrieving the information for the 'hack
   around'

   irq_desc is core internal and it's clear enough stated.

Replace at least the irq_desc abuse with the proper functions and add
a big fat comment why this is absurd and completely wrong.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Gavin Shan <shangw@linux.vnet.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: ppc <linuxppc-dev@lists.ozlabs.org>
Link: http://lkml.kernel.org/r/20140223212736.562906212@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

No functional change
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: ppc <linuxppc-dev@lists.ozlabs.org>
Link: http://lkml.kernel.org/r/20140223212736.333718121@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

I'm really grumpy about this one. The line:

#include "../../../kernel/irq/settings.h"

should have been an alarm sign for all people who added their SOB to
this trainwreck.

When I cleaned up the mess people made with interrupt descriptors a
few years ago, I warned that I'm going to hunt down new offenders and
treat them with stinking trouts. In this case I'll use frozen shark
for a better educational value.

The whole idiocy which was done there could have been avoided with two
lines of perfectly fine code. And do not complain about the lack of
correct examples in tree.

The solution is simple:

  Remove the brainfart and use the proper functions, which should
  have been used in the first place
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Ashish Kalra <ashish.kalra@freescale.com>
Cc: Timur Tabi <timur@freescale.com>
Cc: Kumar Gala <galak@kernel.crashing.org>
Cc: ppc <linuxppc-dev@lists.ozlabs.org>
Link: http://lkml.kernel.org/r/20140223212736.451970660@linutronix.deSigned-off-by: NThomas Gleixner <tglx@linutronix.de>

This reverts commit 01ab1167, it is
incorrect, a future patch will fix this up properly.

Cc: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This fixes the build breakage introduced by
c07a0191 and adds support for the device
control API and save/restore of the VGIC state for ARMv8.

The defines were simply missing from the arm64 header files and
uaccess.h must be implicitly imported from somewhere else on arm.
Signed-off-by: NChristoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Madper reported seeing the following crash,

  BUG: unable to handle kernel paging request at ffffffffff340003
  IP: [<ffffffff81d85ba4>] efi_bgrt_init+0x9d/0x133
  Call Trace:
   [<ffffffff81d8525d>] efi_late_init+0x9/0xb
   [<ffffffff81d68f59>] start_kernel+0x436/0x450
   [<ffffffff81d6892c>] ? repair_env_string+0x5c/0x5c
   [<ffffffff81d68120>] ? early_idt_handlers+0x120/0x120
   [<ffffffff81d685de>] x86_64_start_reservations+0x2a/0x2c
   [<ffffffff81d6871e>] x86_64_start_kernel+0x13e/0x14d

This is caused because the layout of the ACPI BGRT header on this system
doesn't match the definition from the ACPI spec, and so we get a bogus
physical address when dereferencing ->image_address in efi_bgrt_init().

Luckily the status field in the BGRT header clearly marks it as invalid,
so we can check that field and skip BGRT initialisation.
Reported-by: NMadper Xie <cxie@redhat.com>
Suggested-by: NToshi Kani <toshi.kani@hp.com>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Matthew Garrett <mjg59@srcf.ucam.org>
Cc: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: NMatt Fleming <matt.fleming@intel.com>

We do not enable the new efi memmap on 32-bit and thus we need to run
runtime_code_page_mkexec() unconditionally there. Fix that.
Reported-and-tested-by: NLejun Zhu <lejun.zhu@intel.com>
Signed-off-by: NBorislav Petkov <bp@suse.de>
Signed-off-by: NMatt Fleming <matt.fleming@intel.com>

If CONFIG_X86_SMAP is disabled, smap_violation() tests for conditions
which are incorrect (as the AC flag doesn't matter), causing spurious
faults.

The dynamic disabling of SMAP (nosmap on the command line) is fine
because it disables X86_FEATURE_SMAP, therefore causing the
static_cpu_has() to return false.

Found by Fengguang Wu's test system.

[ v3: move all predicates into smap_violation() ]
[ v2: use IS_ENABLED() instead of #ifdef ]
Reported-by: NFengguang Wu <fengguang.wu@intel.com>
Link: http://lkml.kernel.org/r/20140213124550.GA30497@localhostSigned-off-by: NH. Peter Anvin <hpa@linux.intel.com>
Cc: <stable@vger.kernel.org> # v3.7+

If SMAP support is not compiled into the kernel, don't enable SMAP in
CR4 -- in fact, we should clear it, because the kernel doesn't contain
the proper STAC/CLAC instructions for SMAP support.

Found by Fengguang Wu's test system.
Reported-by: NFengguang Wu <fengguang.wu@intel.com>
Link: http://lkml.kernel.org/r/20140213124550.GA30497@localhostSigned-off-by: NH. Peter Anvin <hpa@linux.intel.com>
Cc: <stable@vger.kernel.org> # v3.7+

When the conversion was made to remove stop machine and use the breakpoint
logic instead, the modification of the function graph caller is still
done directly as though it was being done under stop machine.

As it is not converted via stop machine anymore, there is a possibility
that the code could be layed across cache lines and if another CPU is
accessing that function graph call when it is being updated, it could
cause a General Protection Fault.

Convert the update of the function graph caller to use the breakpoint
method as well.

Cc: H. Peter Anvin <hpa@zytor.com>
Cc: stable@vger.kernel.org # 3.5+
Fixes: 08d636b6 "ftrace/x86: Have arch x86_64 use breakpoints instead of stop machine"
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>

Commit 70b41abc
"ARM: ux500: move MSP pin control to the device tree"
accidentally activated MSP2, giving rise to a boot scroll
scream as the kernel attempts to probe a driver for it and
fails to obtain DMA channel 14.

Fix this up by marking the node disabled again.
Acked-by: NLee Jones <lee.jones@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NKevin Hilman <khilman@linaro.org>

This patch adds the support for to create a direct iommu "bypass"
window on IODA2 bridges (such as Power8) allowing to bypass iommu
page translation completely for 64-bit DMA capable devices, thus
significantly improving DMA performances.

Additionally, this adds a hook to the struct iommu_table so that
the IOMMU API / VFIO can disable the bypass when external ownership
is requested, since in that case, the device will be used by an
environment such as userspace or a KVM guest which must not be
allowed to bypass translations.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

We expose a number of OF properties in the kexec and crash dump code
and these need to be big endian.

Cc: stable@vger.kernel.org # v3.13
Signed-off-by: NAnton Blanchard <anton@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

We would allocate one specific exception stack for each kind of
non-base exceptions for every CPU. For ppc32 the CPU hard ID is
used as the subscript to get the specific exception stack for
one CPU. But for an UP kernel, there is only one element in the
each kind of exception stack array. We would get stuck if the
CPU hard ID is not equal to '0'. So in this case we should use the
subscript '0' no matter what the CPU hard ID is.
Signed-off-by: NKevin Hao <haokexin@gmail.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Currently we set our cpu's bit in cpus_in_xmon, and then we take the
output lock and print the exception information.

This can race with the master cpu entering the command loop and printing
the backtrace. The result is that the backtrace gets garbled with
another cpu's exception print out.

Fix it by delaying the set of cpus_in_xmon until we are finished
printing.
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

As far as I can tell, our 70s era timeout loop in get_output_lock() is
generating no code.

This leads to the hostile takeover happening more or less simultaneously
on all cpus. The result is "interesting", some example output that is
more readable than most:

    cpu 0x1: Vector: 100 (Scypsut e0mx bR:e setV)e catto xc0p:u[ c 00
    c0:0  000t0o0V0erc0td:o5 rfc28050000]0c00 0 0  0 6t(pSrycsV1ppuot
    uxe 1m 2 0Rx21e3:0s0ce000c00000t00)00 60602oV2SerucSayt0y 0p 1sxs

Fix it by using udelay() in the timeout loop. The wait time and check
frequency are arbitrary, but seem to work OK. We already rely on
udelay() working so this is not a new dependency.
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

If we enter with xmon_speaker != 0 we skip the first cmpxchg(), we also
skip the while loop because xmon_speaker != last_speaker (0) - meaning we
skip the second cmpxchg() also.

Following that code path the compiler sees no memory barriers and so is
within its rights to never reload xmon_speaker. The end result is we loop
forever.

This manifests as all cpus being in xmon ('c' command), but they refuse
to take control when you switch to them ('c x' for cpu # x).

I have seen this deadlock in practice and also checked the generated code to
confirm this is what's happening.

The simplest fix is just to always try the cmpxchg().
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>