Initially, we decided to make am33xx family of device to fall
under omap3 class (cpu_is_omap34xx() = true), since it carries
Cortex-A8 core. But while adding complete baseport support
(like, clock, power and hwmod) support, it is observed that,
we are creating more and more problems by treating am33xx device
as omap3 family, as nothing matches between them
(except cortex-A8 mpu).

So,  after long discussion we have came to the conclusion that,
we should not consider am33xx device as omap3 family, instead
create separate class (SOC_AM33XX) under OMAP2PLUS.
This means, for am33xx device, cpu_is_omap34xx() will return false,
and only cpu_is_am33xx() will be true.

Please refer to the link below, for mailing-list discussion on this -

http://www.spinics.net/lists/linux-omap/msg69439.htmlSigned-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
[tony@atomide.com: fixed typo, updated for soc_is changes]
Signed-off-by: Tony Lindgren <tony@atomide.com>

In order to remove unnecessary idefs, move noncore and core
dpll ops to dpll3xxx.c file (where it should have been already).

The clkops (clkops_omap3_core_dpll_ops & clkops_omap3_noncore_dpll_ops)
is used in clock data files, and dependency is already handled by
Makefile rule.
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Acked-by: Paul Walmsley <paul@pwsan.com>
Cc: Kevin Hilman <khilman@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>

All OMAP2PLUS based devices, builds omap-device.o target;
so just add one entry so that there is no need to patch this file
for any future OMAP2+ devices.
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>

Replace the regular kzalloc and ioremap with the devm_ equivalent
to simplify error handling. We don't need kree() any more in
omap_dm_timer_remove().

Also added *dev* pointer to reference pdev->dev which makes the
usage shorter in code.

Cc: Cousson, Benoit <b-cousson@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: Rajendra Nayak <rnayak@ti.com>
Cc: Santosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: Tarun Kanti DebBarma <tarun.kanti@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>

Merge tag 'omap-devel-f-for-3.6' of git://git.kernel.org/pub/scm/linux/kernel/git/pjw/omap-pending into cleanup-part2

Miscellaneous OMAP clock, hwmod, clockdomain, and powerdomain patches
for 3.6.  Mostly small infrastructure improvements, and preparation
for OMAP5 and AM33xx code.

Conflicts:
	arch/arm/mach-omap2/omap_hwmod.c
	arch/arm/plat-omap/include/plat/omap_hwmod.h

Merge branches 'hwmod_am335x_support_3.6', 'clkdm_pwrdm_devel_a_3.6' and 'misc_devel_3.6' into omap_devel_f_3.6

The DMADISABLE bit is a semi-automatic bit present in sysconfig register
of some modules. When the DMA must perform read/write accesses, the
DMADISABLE bit is cleared by the hardware. But when the DMA must stop for power
management, software must set the DMADISABLE bit back to 1.

In cases where the ROMCODE/BOOTLOADER uses dma, the hardware clears the
DMADISABLE bit (but the romcode/bootloader might not set it back to 1).
In order for the kernel to start in a clean state, it is
necessary for the kernel to set DMADISABLE bit back to 1 (irrespective
of whether it's been set to 1 in romcode or bootloader).

During _reset of the (hwmod)device, the DMADISABLE bit is set so that it
does not prevent idling of the system. (NOTE: having DMADISABLE to 0,
prevents the system to idle)

DMADISABLE bit is present in usbotgss module of omap5.

Cc: Benoit Cousson <b-cousson@ti.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
[paul@pwsan.com: updated to apply; fixed checkpatch warnings]
Signed-off-by: Paul Walmsley <paul@pwsan.com>

Some prm and cm registers read/write and status functions
are built only for some custom OMAP2+ builds and are stubbed
in header files for other builds under ifdef statements.
But this results in adding new CONFIG_ARCH_OMAPXXX
checks when SOCs are added in the future. So move them
to a common place for OMAP2+ and make them 'weak' implementations.

This way no new ifdefs would be required in the future and also
cleans up the existing code.
Signed-off-by: R Sricharan <r.sricharan@ti.com>
[paul@pwsan.com: unsplit quoted strings; moved PRM functions to
 mach-omap2/prm_common.c; resolved sparse warnings]
Signed-off-by: Paul Walmsley <paul@pwsan.com>

Add an API to get main clock name associated with a given @oh.
This will avoid the need to construct fclk names during early
initialization in order to get fclk handle using clk_get().
Signed-off-by: Tarun Kanti DebBarma <tarun.kanti@ti.com>
Cc: Benoit Cousson <b-cousson@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Tony Lindgren <tony@atomide.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: Rajendra Nayak <rnayak@ti.com>
Cc: Santosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: Benoit Cousson <b-cousson@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>

If the dpll is already locked, code can be optimized
to return much earlier than doing redundent set of lock mode
and wait on idlest.

Cc: Tony Lindgren <tony@atomide.com>
Cc: Jon Hunter <jon-hunter@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Mike Turquette <mturquette@ti.com>
Signed-off-by: Vikram Pandita <vikram.pandita@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>

The register is used to configure the behaviour of the CSI-2 and CCP-2
receivers. This register is available only in OMAP3630.

The original patch was submitted by Vimarsh Zutshi.
Signed-off-by: Sakari Ailus <sakari.ailus@iki.fi>
Cc: Vimarsh Zutshi <vimarsh.zutshi@gmail.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>

The wake-up power domain is an alway-on power domain and so this power domain
does not have a power state status (PM_PWSTST_xxx) register that indicates the
current state. However, during the registering of the wake-up power domain the
state of the domain is queried by calling pwrdm_read_pwrst(). This actually
tries to read a register that does not exist and returns a value of 0 that
indicates that the current state is OFF. The OFF state count of the wake-up
power domain is then set to 1 and the current state to OFF. Both of which are
incorrect.

To fix this, if a power domain only supports the ON state, do not attempt to
read the power state status register and simply return ON as the current power
state.

This is based upon Tony's current linux-omap master branch.

Testing:
- Boot tested on OMAP4460 panda.
- Boot tested on OMAP3430 beagle and validated CORE RET still working (using
  Paul's 32k timer patch [1]).

[1] http://marc.info/?l=linux-omap&m=134000053229888&w=2Signed-off-by: Jon Hunter <jon-hunter@ti.com>
Acked-by: Kevin Hilman <khilman@ti.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
[paul@pwsan.com: edited commit message slightly]
Signed-off-by: Paul Walmsley <paul@pwsan.com>

For OMAP3+ devices, the clock domains (CLKDMs) support one or more of the
following transition modes ...

NO_SLEEP (0x0) - A clock domain sleep transition is never initiated,
		 irrespective of the hardware conditions.
SW_SLEEP (0x1) - A software-forced sleep transition. The transition is initiated
		 when the associated hardware conditions are satisfied
SW_WKUP  (0x2) - A software-forced clock domain wake-up transition is initiated,
		 irrespective of the hardware conditions.
HW_AUTO  (0x3) - Hardware-controlled automatic sleep and wake-up transition is
		 initiated by the PRCM module when the associated hardware
		 conditions are satisfied.

For OMAP4 devices, SW_SLEEP is equivalent to HW_AUTO and NO_SLEEP is equivalent
to SW_WKUP. The only difference between HW_AUTO and SW_SLEEP for OMAP4 devices
is that the PRM_IRQSTATUS_MPU.TRANSITION_ST interrupt status is set in case of
SW_SLEEP transition, and not set in case of HW_AUTO transition.

For OMAP4 devices, all CLKDMs support HW_AUTO and therefore we can place the
CLKDMs in the HW_AUTO state instead of the SW_SLEEP mode. Hence, we do not
need to use the SW_SLEEP mode. With regard to NO_SLEEP and SW_WKUP it is
preferred to use SW_WKUP mode if the CLKDM supports it and so use this mode
instead of NO_SLEEP where possible.

For a software perspective the above 4 modes are represented by the following
flags to indicate what modes are supported by each of the CLKDMs.

CLKDM_CAN_DISABLE_AUTO	--> NO_SLEEP
CLKDM_CAN_ENABLE_AUTO	--> HW_AUTO
CLKDM_CAN_FORCE_SLEEP	--> SW_SLEEP
CLKDM_CAN_FORCE_WAKEUP	--> SW_WKUP

By eliminating the SW_SLEEP mode the the mapping of the flags for OMAP4 devices
can becomes ...

CLKDM_CAN_DISABLE_AUTO	--> NO_SLEEP
CLKDM_CAN_ENABLE_AUTO	--> HW_AUTO
CLKDM_CAN_FORCE_SLEEP	--> HW_AUTO
CLKDM_CAN_FORCE_WAKEUP	--> SW_WKUP

Cc: Ming Lei <ming.lei@canonical.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Benoit Cousson <b-cousson@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Kevin Hilman <khilman@ti.com>
Reviewed-by: Benoit Cousson <b-cousson@ti.com>
Reviewed-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: Jon Hunter <jon-hunter@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>

In case of AM33XX device, XXX_RSTST register offset is not
consistent across PRM modules/instances,

PRM_XXX            RSTST
=========================
PRM_PER_MOD:       0x04
PRM_WKUP_MOD:      0x0C
PRM_MPU_MOD:       NA
PRM_DEVICE_MOD:    0x08

This means, we need to pass on XXX_RSTST register offset
information through omap_hwmod data, similar to XXX_RSTCTRL.

Currently, this field is only applicable and used for AM33XX
devices.
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Cc: Benoit Cousson <b-cousson@ti.com>
Cc: Tony Lindgren <tony@atomide.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Rajendra Nayak <rnayak@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>

In case of AM33xx family of devices (like cpsw) have different sysc
bit field offsets defined,

sysc_type3:
|  3     2  |  1    0  |
| STDBYMODE | IDLEMODE |

So introduce new sysc_type3 in omap_hwmod common data.
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Vaibhav Bedia <vaibhav.bedia@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>

Merge tags 'omap-cleanup-for-v3.6', 'omap-devel-dmtimer-for-v3.6' and 'omap-devel-am33xx-for-v3.6' into devel-am33xx-part2

Pull ARM SoC fixes from Olof Johansson:
 "Another week, another batch of fixes.

  All are small, contained, targeted fixes for explicit problems --
  mostly build and boot failures across i.MX, OMAP, Renesas/Shmobile and
  Samsung."

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: imx6q: fix suspend regression caused by common clk migration
  ARM: OMAP4470: Fix OMAP4470 boot failure
  ARM: EXYNOS: Fix EXYNOS_DEV_DMA Kconfig entry
  ARM: OMAP2+: nand: fix build error when CONFIG_MTD_ONENAND_OMAP2=n
  ARM: shmobile: r8a7779: Route all interrupts to ARM
  ARM: shmobile: kzm9d: use late init machine hook
  ARM: shmobile: kzm9g: use late init machine hook
  ARM: mach-shmobile: armadillo800eva: Use late init machine hook
  ARM: SAMSUNG: Fix for S3C2412 EBI memory mapping
  ARM: mach-shmobile: add missing GPIO IRQ configuration on mackerel
  ARM: mach-shmobile: Fix build when SMP is enabled and EMEV2 is not enabled
  ARM: shmobile: sh7372: bugfix: chclr_offset base
  ARM: shmobile: sh73a0: bugfix: SY-DMAC number
  ARM: SAMSUNG: Should check for IS_ERR(clk) instead of NULL

Fix kernel-doc warnings in printk.c: use correct parameter name.

  Warning(kernel/printk.c:2429): No description found for parameter 'buf'
  Warning(kernel/printk.c:2429): Excess function parameter 'line' description in 'kmsg_dump_get_buffer'
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Fix kernel-doc warning.  This struct member was removed in commit
87568264 ("irq: Remove irq_chip->release()") so remove its
associated kernel-doc entry also.

  Warning(include/linux/irq.h:338): Excess struct/union/enum/typedef member 'release' description in 'irq_chip'
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Cc: Richard Weinberger <richard@nod.at>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge branch 'v3.5-samsung-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung into fixes

* 'v3.5-samsung-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: EXYNOS: Fix EXYNOS_DEV_DMA Kconfig entry
  ARM: SAMSUNG: Fix for S3C2412 EBI memory mapping
  ARM: SAMSUNG: Should check for IS_ERR(clk) instead of NULL

When moving to common clk framework, the imx6q clks rom and mmdc_ch1_axi
get different on/off states than old clk driver, which breaks suspend
function.  There might be a better way to manage these clocks, but let's
takes the old clk driver approach to fix the regression first.
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Olof Johansson <olof@lixom.net>

Merge tag 'omap-fixes-for-v3.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap into fixes

From Tony Lindgren:
"Here's one more regression fix that I missed earlier, and a
 trivial fix to get omap4470 booting."

* tag 'omap-fixes-for-v3.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap:
  ARM: OMAP4470: Fix OMAP4470 boot failure
  ARM: OMAP2+: nand: fix build error when CONFIG_MTD_ONENAND_OMAP2=n

Pull ACPI & Power Management patches from Len Brown.

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux:
  acpi_pad: fix power_saving thread deadlock
  ACPI video: Still use ACPI backlight control if _DOS doesn't exist
  ACPI, APEI, Avoid too much error reporting in runtime
  ACPI: Add a quirk for "AMILO PRO V2030" to ignore the timer overriding
  ACPI: Remove one board specific WARN when ignoring timer overriding
  ACPI: Make acpi_skip_timer_override cover all source_irq==0 cases
  ACPI, x86: fix Dell M6600 ACPI reboot regression via DMI
  ACPI sysfs.c strlen fix

Pull driver Core fixes from Greg Kroah-Hartman:
 "Here is a number of printk() fixes, specifically a few reported by the
  crazy blog program that ships in SUSE releases (that's "boot log" and
  not "web log", it predates the general "blog" terminology by many
  years), and the restoration of the continuation line functionality
  reported by Stephen and others.  Yes, the changes seem a bit big this
  late in the cycle, but I've been beating on them for a while now, and
  Stephen has even optimized it a bit, so all looks good to me.

  The other change in here is a Documentation update for the stable
  kernel rules describing how some distro patches should be backported,
  to hopefully drive a bit more response from the distros to the stable
  kernel releases.

  Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"

* tag 'driver-core-3.5-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  printk: Optimize if statement logic where newline exists
  printk: flush continuation lines immediately to console
  syslog: fill buffer with more than a single message for SYSLOG_ACTION_READ
  Revert "printk: return -EINVAL if the message len is bigger than the buf size"
  printk: fix regression in SYSLOG_ACTION_CLEAR
  stable: Allow merging of backports for serious user-visible performance issues

Merge branches 'acpi_pad-bugzilla-42981', 'apei-bugzilla-43282', 'video-bugzilla-43168', 'bugzilla-40002' and 'bugfix-misc' into release

bug fixes

The acpi_pad driver can get stuck in destroy_power_saving_task()
waiting for kthread_stop() to stop a power_saving thread.  The problem
is that the isolated_cpus_lock mutex is owned when
destroy_power_saving_task() calls kthread_stop(), which waits for a
power_saving thread to end, and the power_saving thread tries to
acquire the isolated_cpus_lock when it calls round_robin_cpu().  This
patch fixes the issue by making round_robin_cpu() use its own mutex.

https://bugzilla.kernel.org/show_bug.cgi?id=42981

Cc: stable@vger.kernel.org
Signed-off-by: Stuart Hayes <Stuart_Hayes@Dell.com>
Signed-off-by: Len Brown <len.brown@intel.com>

This fixes a regression in 3.4-rc1 caused by commit
ea9f8856
(ACPI video: Harden video bus adding.)

Some platforms don't have _DOS control method, but the ACPI
backlight still works.
We should not invoke _DOS for these platforms.

https://bugzilla.kernel.org/show_bug.cgi?id=43168

Cc: Igor Murzov <intergalactic.anonymous@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>

Pull power management fixes from Rafael J. Wysocki:

 * Fix for a bug in async suspend error code path causing parents to
   wait forever for their children in case of a suspend error from
   Mandeep Singh Baines (-stable metarial).

 * Fix for a suspend regression related to earlier changes in the ACPI
   cpuidle driver from Deepthi Dharwar.

* tag 'pm-for-3.5-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  PM / ACPI: Fix suspend/resume regression caused by cpuidle cleanup.
  PM / Sleep: Prevent waiting forever on asynchronous suspend after abort

In reviewing Kay's fix up patch: "printk: Have printk() never buffer its
data", I found two if statements that could be combined and optimized.

Put together the two 'cont.len && cont.owner == current' if statements
into a single one, and check if we need to call cont_add(). This also
removes the unneeded double cont_flush() calls.

Link: http://lkml.kernel.org/r/1340869133.876.10.camel@mopSigned-off-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Kay Sievers <kay@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Pull powerpc fixes from Benjamin Herrenschmidt:
 "Here are a few powerpc fixes.  Arguably some of this should have come
  to you earlier but I'm only just catching up after my medical leave.

  Mostly these fixes regressions, a couple are long standing bugs."

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc/pseries: Fix software invalidate TCE
  powerpc: check_and_cede_processor() never cedes
  powerpc/ftrace: Do not trace restore_interrupts()
  powerpc: Fix Section mismatch warnings in prom_init.c
  ppc64: fix missing to check all bits of _TIF_USER_WORK_MASK in preempt
  powerpc: Fix uninitialised error in numa.c
  powerpc: Fix BPF_JIT code to link with multiple TOCs

Pull x86 fixes from Ingo Molnar.

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86, cpufeature: Remove stray %s, add -w to mkcapflags.pl
  x86, cpufeature: Catch duplicate CPU feature strings
  x86, cpufeature: Rename X86_FEATURE_DTS to X86_FEATURE_DTHERM
  x86: Fix kernel-doc warnings
  x86, compat: Use test_thread_flag(TIF_IA32) in compat signal delivery

Pull oprofile fixlet from Ingo Molnar.

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  oprofile: perf: use NR_CPUS instead or nr_cpumask_bits for static array

Pull RCU fix from Ingo Molnar.

Fixes a bug introduced in this merge window by commit b1420f1c ("Make
rcu_barrier() less disruptive")

* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  rcu: Stop rcu_do_batch() from multiplexing the "count" variable

Continuation lines are buffered internally, intended to merge the
chunked printk()s into a single record, and to isolate potentially
racy continuation users from usual terminated line users.

This though, has the effect that partial lines are not printed to
the console in the moment they are emitted. In case the kernel
crashes in the meantime, the potentially interesting printed
information would never reach the consoles.

Here we share the continuation buffer with the console copy logic,
and partial lines are always immediately flushed to the available
consoles. They are still buffered internally to improve the
readability and integrity of the messages and minimize the amount
of needed record headers to store.
Signed-off-by: Kay Sievers <kay@vrfy.org>
Tested-by: Steven Rostedt <rostedt@goodmis.org>
Acked-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The following added support for powernv but broke pseries/BML:
 1f1616e8 powerpc/powernv: Add TCE SW invalidation support

TCE_PCI_SW_INVAL was split into FREE and CREATE flags but the tests in
the pseries code were not updated to reflect this.
Signed-off-by: Michael Neuling <mikey@neuling.org>
cc: stable@kernel.org [v3.3+]
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

Commit f948501b ("Make hard_irq_disable() actually hard-disable
interrupts") caused check_and_cede_processor to stop working.
->irq_happened will never be zero right after a hard_irq_disable
so the compiler removes the call to cede_processor completely.

The bug was introduced back in the lazy interrupt handling rework
of 3.4 but was hidden until recently because hard_irq_disable did
nothing.

This issue will eventually appear in 3.4 stable since the
hard_irq_disable fix is marked stable, so mark this one for stable
too.
Signed-off-by: Anton Blanchard <anton@samba.org>
Cc: stable@vger.kernel.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

As I was adding code that affects all archs, I started testing function
tracer against PPC64 and found that it currently locks up with 3.4
kernel. I figured it was due to tracing a function that shouldn't be, so
I went through the following process to bisect to find the culprit:

 cat /debug/tracing/available_filter_functions > t
 num=`wc -l t`
 sed -ne "1,${num}p" t > t1
 let num=num+1
 sed -ne "${num},$p" t > t2
 cat t1 > /debug/tracing/set_ftrace_filter
 echo function /debug/tracing/current_tracer
 <failed? bisect t1, if not bisect t2>

It finally came down to this function: restore_interrupts()

I'm not sure why this locks up the system. It just seems to prevent
scheduling from occurring. Interrupts seem to still work, as I can ping
the box. But all user processes freeze.

When restore_interrupts() is not traced, function tracing works fine.

Cc: stable@kernel.org
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

This patches tries to fix a couple of Section mismatch warnings like
following one:

WARNING: arch/powerpc/kernel/built-in.o(.text+0x2923c): Section mismatch
in reference from the function .prom_query_opal() to the
function .init.text:.call_prom()
The function .prom_query_opal() references
the function __init .call_prom().
This is often because .prom_query_opal lacks a __init
annotation or the annotation of .call_prom is wrong.
Signed-off-by: Li Zhong <zhong@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

In entry_64.S version of ret_from_except_lite, you'll notice that
in the !preempt case, after we've checked MSR_PR we test for any
TIF flag in _TIF_USER_WORK_MASK to decide whether to go to do_work
or not. However, in the preempt case, we do a convoluted trick to
test SIGPENDING only if PR was set and always test NEED_RESCHED ...
but we forget to test any other bit of _TIF_USER_WORK_MASK !!! So
that means that with preempt, we completely fail to test for things
like single step, syscall tracing, etc...

This should be fixed as the following path:

 - Test PR. If not set, go to resume_kernel, else continue.

 - If go resume_kernel, to do that original do_work.

 - If else, then always test for _TIF_USER_WORK_MASK to decide to do
that original user_work, else restore directly.
Signed-off-by: Tiejun Chen <tiejun.chen@windriver.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>