driver core: device_rename's new_name can be const

The new_name argument to device_rename() can be
const as kobject_rename's new_name argument is.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

sysfs: Remove owner field from sysfs struct attribute

Signed-off-by: Guenter Roeck <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

powerpc/pci: Remove owner field from attribute initialization in PCI bridge init

Signed-off-by: Guenter Roeck <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

regulator: Remove owner field from attribute initialization in regulator core driver

Signed-off-by: Guenter Roeck <[email protected]>
Acked-by: Mark Brown <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

leds: Remove owner field from attribute initialization in bd2802 driver

Signed-off-by: Guenter Roeck <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

scsi: Remove owner field from attribute initialization in ARCMSR driver

Signed-off-by: Guenter Roeck <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

scsi: Remove owner field from attribute initialization in LPFC driver

Signed-off-by: Guenter Roeck <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

cgroupfs: create /sys/fs/cgroup to mount cgroupfs on

We really shouldn't be asking userspace to create new root filesystems.
So follow along with all of the other in-kernel filesystems, and provide
a mount point in sysfs.

For cgroupfs, this should be in /sys/fs/cgroup/  This change provides
that mount point when the cgroup filesystem is registered in the kernel.

Acked-by: Paul Menage <[email protected]>
Acked-by: Dhaval Giani <[email protected]>
Cc: Li Zefan <[email protected]>
Cc: Lennart Poettering <[email protected]>
Cc: Kay Sievers <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

Driver core: Add BUS_NOTIFY_BIND_DRIVER

Add BUS_NOTIFY_BIND_DRIVER as a bus notifier event.

For driver binding/unbinding we with this in
place have the following bus notifier events:
 - BUS_NOTIFY_BIND_DRIVER - before ->probe()
 - BUS_NOTIFY_BOUND_DRIVER - after ->probe()
 - BUS_NOTIFY_UNBIND_DRIVER - before ->remove()
 - BUS_NOTIFY_UNBOUND_DRIVER - after ->remove()

The event BUS_NOTIFY_BIND_DRIVER allows bus code
to be notified that ->probe() is about to be called.

Useful for bus code that needs to setup hardware before
the driver gets to run. With this in place platform
drivers can be loaded and unloaded as modules and the
new BIND event allows bus code to control for instance
device clocks that must be enabled before the driver
can be executed.

Without this patch there is no way for the bus code to
get notified that a modular driver is about to be probed.

Signed-off-by: Magnus Damm <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

driver core: fix memory leak on one error path in bus_register()

Reported-by: [email protected]
Signed-off-by: Jike Song <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

debugfs: no longer needs to depend on SYSFS

debugfs no longer uses 'kernel_subsys' (which is gone), and other
kernel/ksysfs.c code is always built, so DEBUG_FS does not need
to depend on SYSFS.

Fixes this kconfig warning:

warning: (TREE_RCU_TRACE || AMD_IOMMU_STATS && AMD_IOMMU || MTD_UBI_DEBUG && MTD && SYSFS && MTD_UBI || UBIFS_FS_DEBUG && MISC_FILESYSTEMS && UBIFS_FS || DEBUG_KMEMLEAK && DEBUG_KERNEL && EXPERIMENTAL && !MEMORY_HOTPLUG && (X86 || ARM || PPC || S390 || SPARC64 || SUPERH || MICROBLAZE) && SYSFS || TRACING || X86_PTDUMP && DEBUG_KERNEL || BLK_DEV_IO_TRACE && TRACING_SUPPORT && FTRACE && SYSFS && BLOCK) selects DEBUG_FS which has unmet direct dependencies (SYSFS)

Signed-off-by: Randy Dunlap <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

sysfs: Fix one more signature discrepancy between sysfs implementation and docs.

Signed-off-by: Ira Weiny <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

sysfs: fix discrepancies between implementation and documentation

Fix all discrepancies I know of between the sysfs implementation and its
documentation.

Signed-off-by: Bart Van Assche <[email protected]>
Cc: Randy Dunlap <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

dcdbas: remove a redundant smi_data_buf_free in dcdbas_exit

smi_data_buf_free is called twice in current implementation.
The second call simply return because smi_data_buf is set to NULL in first call.
This patch removes the second smi_data_buf_free call.

Signed-off-by: Axel Lin <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

dmi-id: fix a memory leak in dmi_id_init error path

This patch adds a missing kfree(dmi_dev) in dmi_id_init error path.

Signed-off-by: Axel Lin <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

sysfs: sysfs_chmod_file's attr can be const

sysfs_chmod_file doesn't change the attribute it operates on, so this
attribute can be marked const.

Signed-off-by: Jean Delvare <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

firmware: Update hotplug script

Update the in-kernel hotplug example script to work
properly with recent kernels. Without this fix the
script may load the firmware twice - both at "add"
and "remove" time.

The second load only triggers in the case when multiple
firmware images are used. A good example is the b43
driver which does not work properly without this fix.

Signed-off-by: Magnus Damm <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

Driver core: move platform device creation helpers to .init.text (if MODULE=n)

Platform devices should only be called by init code, so it should be
possible to move creation helpers to .init.text -- at least if modules
are disabled.

Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

Driver core: reduce duplicated code for platform_device creation

This makes the two similar functions platform_device_register_simple
and platform_device_register_data one line inline functions using a new
generic function platform_device_register_resndata.

Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

Driver core: use kmemdup in platform_device_add_resources

This makes platform_device_add_resources look like
platform_device_add_data.

Signed-off-by: Uwe Kleine-König <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

firmware loader: embed device into firmware_priv structure

Both these structures have the same lifetime rules so instead of allocating
and managing them separately embed struct device into struct firmware_priv.
Also make sure to delete sysfs attributes ourselves instead of expecting
sysfs to clean up our mess.

Signed-off-by: Dmitry Torokhov <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

firmware loader: use statically initialized data attribute

There is no reason why we are using a template for binary attribute
and copying it into per-firmware data before registering. Using the
original works as well.

Signed-off-by: Dmitry Torokhov <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

Driver core: Drop __must_check from bus_for_each_drv()

There is little rationale for marking bus_for_each_drv() __must_check.
It is more of an iteration helper than a real function. You don't know
in advance which callback it will be used on, so you have no clue how
important it can be to check the returned value. In practice, this
helper function can be used for best-effort tasks.

As a matter of fact, bus_for_each_dev() is not marked __must_check.
So remove it from bus_for_each_drv() as well. This is the same that
was done back in October 2006 by Russell King for
device_for_each_child(), for exactly the same reasons.

Signed-off-by: Jean Delvare <[email protected]>
Cc: Andrew Morton <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

Driver core: internal struct dma_coherent_mem, change type of a member.

struct dma_coherent_mem in drivers/base/dma-coherent.c
has member 'device_base' that is of type u32,
but is assigned value of type dma_addr_t, which may be
64 bits for x86_64. Change the type to dma_addr_t.

Signed-off-by: Marin Mitov <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

hotplug: Support kernel/hotplug sysctl variable when !CONFIG_NET

The kernel/hotplug sysctl variable (/proc/sys/kernel/hotplug file) was
made conditional on CONFIG_NET by commit
f743ca5e10f4145e0b3e6d11b9b46171e16af7ce (applied in 2.6.18) to fix
problems with undefined references in 2.6.16 when CONFIG_HOTPLUG=y &&
!CONFIG_NET, but this restriction is no longer needed.

This patch makes the kernel/hotplug sysctl variable depend only on
CONFIG_HOTPLUG.

Signed-off-by: Ian Abbott <[email protected]>
Acked-by: Randy Dunlap <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

uio: Remove IRQF_DISABLED flag from uio_cif.c

Remove IRQF_DISABLED since it is deprecated and a no-op in the
current kernel.

Signed-off-by: Hans J. Koch <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

uio: Remove IRQF_DISABLED from uio_sercos3.c

Remove IRQF_DISABLED since it is deprecated and a no-op in the
current kernel.

Signed-off-by: Hans J. Koch <[email protected]>
Acked-by: John Ogness <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

uio: Remove IRQF_DISABLED flag from uio_pdrv_genirq.c

Remove IRQF_DISABLED flag since it is deprecated and a no-op in the
current kernel.

Signed-off-by: Hans J. Koch <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

Merge branch 'next' of git://git./linux/kernel/git/benh/powerpc

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (79 commits)
  powerpc/8xx: Add support for the MPC8xx based boards from TQC
  powerpc/85xx: Introduce support for the Freescale P1022DS reference board
  powerpc/85xx: Adding DTS for the STx GP3-SSA MPC8555 board
  powerpc/85xx: Change deprecated binding for 85xx-based boards
  powerpc/tqm85xx: add a quirk for ti1520 PCMCIA bridge
  powerpc/tqm85xx: update PCI interrupt-map attribute
  powerpc/mpc8308rdb: support for MPC8308RDB board from Freescale
  powerpc/fsl_pci: add quirk for mpc8308 pcie bridge
  powerpc/85xx: Cleanup QE initialization for MPC85xxMDS boards
  powerpc/85xx: Fix booting for P1021MDS boards
  powerpc/85xx: Fix SWIOTLB initalization for MPC85xxMDS boards
  powerpc/85xx: kexec for SMP 85xx BookE systems
  powerpc/5200/i2c: improve i2c bus error recovery
  of/xilinxfb: update tft compatible versions
  powerpc/fsl-diu-fb: Support setting display mode using EDID
  powerpc/5121: doc/dts-bindings: update doc of FSL DIU bindings
  powerpc/5121: shared DIU framebuffer support
  powerpc/5121: move fsl-diu-fb.h to include/linux
  powerpc/5121: fsl-diu-fb: fix issue with re-enabling DIU area descriptor
  powerpc/512x: add clock structure for Video-IN (VIU) unit
  ...

Merge branch 'for-linus' of git://git.monstr.eu/linux-2.6-microblaze

* 'for-linus' of git://git.monstr.eu/linux-2.6-microblaze: (49 commits)
  microblaze: Add KGDB support
  microblaze: Support brki rX, 0x18 for user application debugging
  microblaze: Remove nop after MSRCLR/SET, MTS, MFS instructions
  microblaze: Simplify syscall rutine
  microblaze: Move PT_MODE saving to delay slot
  microblaze: Fix _interrupt function
  microblaze: Fix _user_exception function
  microblaze: Put together addik instructions
  microblaze: Use delay slot in syscall macros
  microblaze: Save kernel mode in delay slot
  microblaze: Do not mix register saving and mode setting
  microblaze: Move SAVE_STATE upward
  microblaze: entry.S: Macro optimization
  microblaze: Optimize hw exception rutine
  microblaze: Implement clear_ums macro and fix SAVE_STATE macro
  microblaze: Remove additional setup for kernel_mode
  microblaze: Optimize SAVE_STATE macro
  microblaze: Remove additional loading
  microblaze: Completely remove working with R11 register
  microblaze: Do not setup BIP in _debug_exception
  ...

Merge git://git./linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  arch/sparc/mm: Use GFP_KERNEL
  MAINTAINERS: Add trailing slash to SBUS path.
  sbus: autoconvert trivial BKL users to private mutex

Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] increase ia64 static per cpu area
  [IA64] Put ia64 config files on the Uwe Kleine-König diet
  [IA64] perfmon: convert to unlocked_ioctl
  [IA64] beautify vmlinux.lds.h

Merge branch 'upstream' of git://git.linux-mips.org/upstream-linus

* 'upstream' of git://git.linux-mips.org/pub/scm/upstream-linus: (150 commits)
  MIPS: PowerTV: Separate PowerTV USB support from non-USB code
  MIPS: strip the un-needed sections of vmlinuz
  MIPS: Clean up the calculation of VMLINUZ_LOAD_ADDRESS
  MIPS: Clean up arch/mips/boot/compressed/decompress.c
  MIPS: Clean up arch/mips/boot/compressed/ld.script
  MIPS: Unify the suffix of compressed vmlinux.bin
  MIPS: PowerTV: Add Gaia platform definitions.
  MIPS: BCM47xx: Fix nvram_getenv return value.
  MIPS: Octeon: Allow more than 3.75GB of memory with PCIe
  MIPS: Clean up notify_die() usage.
  MIPS: Remove unused task_struct.trap_no field.
  Documentation: Mention that KProbes is supported on MIPS
  SAMPLES: kprobe_example: Make it print something on MIPS.
  MIPS: kprobe: Add support.
  MIPS: Add instrunction format for BREAK and SYSCALL
  MIPS: kprobes: Define regs_return_value()
  MIPS: Ritually kill stupid printk.
  MIPS: Octeon: Disallow MSI-X interrupt and fall back to MSI interrupts.
  MIPS: Octeon: Support 256 MSI on PCIe
  MIPS: Decode core number for R2 CPUs.
  ...

Merge git://git./linux/kernel/git/czankel/xtensa-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/czankel/xtensa-2.6:
  xtensa: Disable PCI and nfsroot on simulation target
  xtensa: Add missing include in coprocessor.h
  xtensa: Fix the network driver for the simulator target
  xtensa: Shuffle include statements to fix linker script
  xtensa: Add -mforce-no-pic option is supported
  xtensa: Fixes due to bss boundary symbol name changes.
  xtensa: Fix linker script patch-up
  xtensa: Fix FLUSH_DCACHE macro for some variants.

nick piggin: change email address

Signed-off-by: Nick Piggin <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

MIPS: PowerTV: Separate PowerTV USB support from non-USB code

Separate USB code into a file separate from asic/asic_devices.

Separating the USB code from everything else in asic/asic_devices.c goes
a long way toward reducing the use of that file as a dumping ground for
everything that didn't seem to fit anywhere else.

Signed-off-by: David VomLehn <[email protected]>
To: [email protected]
Cc: [email protected]
Cc: [email protected]
Patchwork: http://patchwork.linux-mips.org/patch/1522/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: strip the un-needed sections of vmlinuz

This patch use "strip -s" to strip the .symtab and .strtab sections of
vmlinuz.

Note: This patch is based on http://patchwork.linux-mips.org/patch/1324/

Signed-off-by: Wu Zhangjin <[email protected]>
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1383/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Clean up the calculation of VMLINUZ_LOAD_ADDRESS

We have calculated VMLINUZ_LOAD_ADDRESS in shell, which is indecipherable. This
patch rewrites it in C.

Signed-off-by: Wu Zhangjin <[email protected]>
To: linux-mips <[email protected]>
Cc: Alexander Clouter <[email protected]>
Cc: Manuel Lauss <[email protected]>
Cc: Sam Ravnborg <[email protected]>
Acked-by: Sam Ravnborg <[email protected]>
Reviewed-by: Alexander Clouter <[email protected]>
Patchwork: https://patchwork.linux-mips.org/patch/1324/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Clean up arch/mips/boot/compressed/decompress.c

- Remove several outdated comments
- Clearify the definition of zimage_start and zimage_size and the their
  usage

Signed-off-by: Wu Zhangjin <[email protected]>
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1382/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Clean up arch/mips/boot/compressed/ld.script

- Remove unused symbols: _fdata, _text; only _edata and _end are needed by
  head.S
- Remove unused sections: .sbss, .stab, .gptab.sdata, .gptab.sbss
- Change the alignment to 16 bytes to ensure it is greater than any
  fundamental type of a MIPS compiler.
- Clean up comments

Signed-off-by: Wu Zhangjin <[email protected]>
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1381/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Unify the suffix of compressed vmlinux.bin

The compressed vmlinux.bin is only a temp file so it's ok to use the same
suffix .z for them (.gz,.lzo,.lzma...) to remove several lines and simpify
the maintenance (no need to add the "suffix_$(xxx) := suffix" line).

Signed-off-by: Wu Zhangjin <[email protected]>
To: linux-mips <[email protected]>
Cc: Alexander Clouter <[email protected]>
Cc: Manuel Lauss <[email protected]>
Cc: Sam Ravnborg <[email protected]>
Patchwork: https://patchwork.linux-mips.org/patch/1323/
Signed-off-by: Ralf Baechle <[email protected]>
---

MIPS: PowerTV: Add Gaia platform definitions.

Define ASIC address, memory preallocations, and initialization code for the
Gaia platform.

Signed-off-by: David VomLehn <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1519/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: BCM47xx: Fix nvram_getenv return value.

Nvram_getenv should behave like cfe_getenv. cfe_getenv returns 0 on
success and -9 if the value was not found. If the input was wrong -8
will be returned by cfe_getenv.  Change nvram_getenv to do the same.

Signed-off-by: Hauke Mehrtens <[email protected]>
Cc: Waldemar Brodkorb <[email protected]>
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1520/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Allow more than 3.75GB of memory with PCIe

We reserve the 3.75GB - 4GB region of PCIe address space for device to
device transfers, making the corresponding physical memory under
direct mapping unavailable for DMA.

To allow for PCIe DMA to all physical memory we map this chunk of
physical memory with BAR1.  Because of the resulting discontinuity in
the mapping function, we remove a page of memory at each end of the
range so multi-page DMA buffers can never be allocated that span the
range.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1535/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Clean up notify_die() usage.

The sixth argument of notify_die() is a signal number, the fifth is a
trap number.

Instead of passing a signal number in a randomly selected argument,
pass it in the sixth.  Extract the exception code from regs and pass
that as the trap number.

Get rid of redundant cast, and remove some gratuitous spaces.

Nobody actually does anything with the signal number or trap number,
but we might as well populate them with sensible values.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1532/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Remove unused task_struct.trap_no field.

It is initialized to zero and only ever read.  Remove it, and pass zero in
its place.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1531/
Signed-off-by: Ralf Baechle <[email protected]>

Documentation: Mention that KProbes is supported on MIPS

MIPS now has KProbes support, so kprobes.txt should reflect it.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1527/
Signed-off-by: Ralf Baechle <[email protected]>

SAMPLES: kprobe_example: Make it print something on MIPS.

This KProbes example is a little useless if it doesn't print anything.
For MIPS print similar messages to those produced on x86 and PPC.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1528/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: kprobe: Add support.

This patch is based on previous work by Sony and Himanshu Chauhan.

I have done some cleanup and implemented JProbes and KRETPROBES.  The
KRETPROBES part is pretty much copied verbatim from powerpc.  A possible
future enhance might be to factor out the common code.

Signed-off-by: David Daney <[email protected]>
Cc: Himanshu Chauhan <[email protected]>
To: [email protected]
To: [email protected],
To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1525/
Patchwork: https://patchwork.linux-mips.org/patch/1530/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Add instrunction format for BREAK and SYSCALL

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1524/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: kprobes: Define regs_return_value()

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
To: [email protected]
To: [email protected]
To: [email protected]
Cc: [email protected],
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1529/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Ritually kill stupid printk.

This belongs into userland.

Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Disallow MSI-X interrupt and fall back to MSI interrupts.

MSI-X interrupts are not supported yet for Octeon, return error if
MSI-X interrupts are requested by driver so that the driver will fall
back to use MSI interrupts.

Signed-off-by: Chandrakala Chavva <[email protected]>
To: [email protected]
Cc: David Daney <[email protected]>
Patchwork: https://patchwork.linux-mips.org/patch/1506/
Signed-off-by: Ralf Baechle <[email protected]>
Signed-off-by: David Daney <[email protected]>

MIPS: Octeon: Support 256 MSI on PCIe

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: http://patchwork.linux-mips.org/patch/1507/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Decode core number for R2 CPUs.

The struct cpuinfo_mips.core field should be populated with the
physical core number.  For R2 CPUs, this is carried in the low 10 bits
of Ebase.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1505/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: SMTC: Use %p to format pointers

While at it, drop 0x prefix.

Signed-off-by: Kulikov Vasiliy <[email protected]>
To: [email protected]
Cc: Chris Dearman <[email protected]>
Cc: "Robert P. J. Day" <[email protected]>
Cc: Rusty Russell <[email protected]>
Cc: André Goddard Rosa <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1458/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Loongson: Remove unused macro LOONGSON_PERFCNT_IRQ

LOONGSON2_PERFCNT_IRQ is used for the irq number of the performance
overflow interrupts; LOONGSON_PERFCNT_IRQ is unused so remove it.

Signed-off-by: Wu Zhangjin <[email protected]>
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1494/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Loongson: Oprofile: add a new do_perfcnt_IRQ()

On FuLoong-2F IP6 is shared by the performance counter overflow interrupt
and the Bonito northbridge interrupt.  To reduce overhead only call
do_IRQ() when oprofile is enabled to reduce overhead.

This patch adds an inline function do_perfcnt_IRQ() to hide the #if's ,
which can be shared by the other Loongson machines, i.e. gdium.

Signed-off-by: Wu Zhangjin <[email protected]>
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1492/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Loongson: Remove set_irq_trigger_mode()

set_irq_trigger_mode() is not needed on all platforms so remove it
and move the related source code to mach_init_irq().

This will allow gdium to share the common irq.c without adding an empty
set_irq_trigger_mode().

Signed-off-by: Wu Zhangjin <[email protected]>
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1493/
Signed-off-by: Ralf Baechle <[email protected]>

WATCHDOG: Add watchdog driver for OCTEON SOCs

The OCTEON is a MIPS64 based SOC family with an on chip watchdog unit.

The driver is split into two source files one for the C code and one
for assembly.  Assembly is needed to handle the NMI and then print the
machine state before the reboot is triggered.

Signed-off-by: David Daney <[email protected]>
Cc: Wim Van Sebroeck <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Russell King <[email protected]>
Cc: Tony Lindgren <[email protected]>
Cc: Marc Zyngier <[email protected]>
Cc: Thierry Reding <[email protected]>
Cc: Sam Ravnborg <[email protected]>
To: [email protected]
Cc: [email protected],
Patchwork: https://patchwork.linux-mips.org/patch/1503/
Signed-off-by: Wim Van Sebroeck <[email protected]>
Signed-off-by: Ralf Baechle <[email protected]>
 create mode 100644 drivers/watchdog/octeon-wdt-main.c
 create mode 100644 drivers/watchdog/octeon-wdt-nmi.S

MIPS: Define ST0_NMI in asm/mipsregs.h

This is used by the forthcoming OCTEON watchdog patch.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1498/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Export __cpu_number_map and __cpu_logical_map.

The forthcoming Octeon watchdog driver will use them.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1499/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Export prom_putchar().

The forthcoming watchdog driver will use it.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1499/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: uasm: Add option to export uasm API.

A 'select EXPORT_UASM' in Kconfig will cause the uasm to be exported
for use in modules.  When it is exported, all the uasm data and code
cease to be __init and __initdata.

Also daddiu_bug cannot be __cpuinitdata if uasm is exported.  The
cleanest thing is to just make it normal data.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1500/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: uasm: Add BBIT0 and BBIT1 instructions

These are OCTEON specific instructions.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1496/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: uasm: Add drotr32 and uasm_i_drotr_safe.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
To: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1495/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Implement delays with cycle counter.

Power throttling make deterministic delay loops impossible.
Re-implement delays using the cycle counter.  This also allows us to
get rid of the code that calculates loops per jiffy.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1317/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add qi_lb60 board support

Add support for the qi_lb60 (a.k.a QI Ben NanoNote) clamshell device.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1472/
Signed-off-by: Ralf Baechle <[email protected]>

POWER: Add JZ4740 battery driver.

Add support for the battery voltage measurement part of the JZ4740 ADC unit.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Acked-by: Anton Vorontsov <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1416/
Signed-off-by: Ralf Baechle <[email protected]>

HWMON: Add JZ4740 ADC driver

Add support for reading the ADCIN pin of the ADC unit on JZ4740 SoCs.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Acked-by: Jean Delvare <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1425/
Signed-off-by: Axel Lin <[email protected]>
Signed-off-by: Ralf Baechle <[email protected]>

USB: Add JZ4740 OHCI support

Add OHCI glue code for JZ4740 SoCs OHCI module.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Cc: David Brownell <[email protected]>
Cc: [email protected]
Acked-by: Greg Kroah-Hartman <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1411/
Signed-off-by: Ralf Baechle <[email protected]>

MMC: Add support for the controller on JZ4740 SoCs.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Acked-by: Matt Fleming <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Matt Fleming <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1463/
Patchwork: https://patchwork.linux-mips.org/patch/1523/
Signed-off-by: Ralf Baechle <[email protected]>

MTD: Nand: Add JZ4740 NAND driver

Add support for the NAND controller on JZ4740 SoCs.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: David Woodhouse <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1470/
Signed-off-by: Ralf Baechle <[email protected]>

FBDEV: JZ4740: Add framebuffer driver

Add support for the LCD controller on JZ4740 SoCs.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1470/
Signed-off-by: Ralf Baechle <[email protected]>

RTC: Add JZ4740 RTC driver

Add support for the RTC unit on JZ4740 SoCs.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: Alessandro Zummo <[email protected]>
Cc: Paul Gortmaker <[email protected]>
Cc: [email protected]
Acked-by: Wan ZongShun <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: Alessandro Zummo <[email protected]>,
Patchwork: https://patchwork.linux-mips.org/patch/1424/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add Kbuild files

Add the Kbuild files for the JZ4740 architecture and adds JZ4740 support
to the MIPS Kbuild files.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1406/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add platform devices

Add platform devices for all the JZ4740 platform drivers.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1469/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add prom support

Add support for initializing arcs_cmdline on JZ4740 based machines and
provides a prom_putchar implementation.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1404/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add serial support

The JZ4740 UART interface is almost 16550 compatible.
The UART module needs to be enabled by setting a bit in the FCR register
and it has support for receive timeout interrupts.  Instead of adding yet
another machine specific quirk to the 8250 serial driver we provide a
serial_out implementation which sets the required additional flags.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1403/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add PWM support

Add support for the PWM part of the timer unit on a JZ4740 SoC.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1468/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add DMA support.

Add support for DMA transfers on JZ4740 SoCs.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1401/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add GPIO support

Add gpiolib support for JZ4740 SoCs.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1467/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add setup code

Add plat_mem_setup and get_system_type for JZ4740 SoCs.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1399/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add power-management and system reset support

Add support for suspend/resume and poweroff/reboot on a JZ4740 SoC.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1398/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add clocksource/clockevent support.

Add clocksource and clockevent support for the timer/counter unit on
JZ4740 SoCs.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1397/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add timer support

Add support for the timer/counter unit on a JZ4740 SoC.  This code is used
as a common base for the JZ4740 clocksource/clockevent implementation and
PWM support.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1396/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add IRQ handler code

Add support for IRQ handling on a JZ4740 SoC.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1465/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add clock API support.

Add support for managing the clocks found on JZ4740 SoC through the
Linux clock API.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1466/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: JZ4740: Add base support for Ingenic JZ4740 System-on-a-Chip

Adds a new cpu type for the JZ4740 to the Linux MIPS architecture code.
It also adds the iomem addresses for the different components found on
a JZ4740 SoC.

Signed-off-by: Lars-Peter Clausen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1464/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: HOTPLUG_CPU fixes.

* Rename camel-case InitTLBStart_addr to octeon_bootloader_entry_addr.

* Convert calls to cvmx_read64_uint32(), to simple pointer
  dereferences.

* Set proper ebase.

* Don't confuse coreid and cpu numbers.

* Try to maintain consistent bootloader coremask.

* Update the signature and boot_init_vector of supported bootloaders.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1491/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Simplify hotcpu_notifier registration.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1490/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Clean up SMP CPU numbering.

Also number offline CPUs that could potentially be brought on-line
later.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1489/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Make MSI use handle_simple_irq().

The use of handle_percpu_irq() is not really what we want for MSI, use
handle_simple_irq() instead.  This is probably the prototypical case
for using handle_simple_irq(), because all the MSIs are dispatched from
the root interrupt service routine.

Also since the base IRQ is not shared, don't pass IRQF_SHARED.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1488/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Get rid of a bunch of MSI IRQ number definitions.

MSI IRQ numbers are allocated dynamically, so there is no reason to
have all these static definitions.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1487/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Fix fixup_irqs for HOTPLUG_CPU

The original version went behind the back of everything, leaving
things in an inconsistent state.

Now we use the irq_set_affinity() to do the work for us.  This has the
advantage that the IRQ core's view of the affinity stays consistent.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1486/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Improve interrupt handling.

The main change is to change most of the IRQs from handle_percpu_irq
to handle_fasteoi_irq.  This necessitates extracting all the .ack code
to common functions that are not exposed to the irq core.

The affinity code now acts more sanely, by doing round-robin
distribution instead of broadcasting.

Because of the change to handle_fasteoi_irq and affinity, some of the
IRQs had to be split into separate groups with their own struct
irq_chip to prevent undefined operations on specific IRQ lines.

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1485/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Octeon: Move MSI code out of octeon-irq.c.

Put all the MSI code in one place (msi-octeon.c).  This simplifies
octeon-irq.c and gets rid of some ugly #ifdefs

Signed-off-by: David Daney <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1484/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: Alchemy: Add basic support for the GPR board

From: Wolfgang Grandegger <[email protected]>

Add basic support for the General Purpose Router (GPR) board from
Trapeze ITS.

Signed-off-by: Wolfgang Grandegger <[email protected]>
To: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1460/
Signed-off-by: Ralf Baechle <[email protected]>

MIPS: au1000_eth: Get ethernet address from platform_data

au1000_eth uses firmware calls to get a valid MAC address, and changes
it depending on platform device id.  This patch moves this logic out of
the driver into the platform device registration part, where boards with
supported chips can use whatever firmware interface they need; the default
implementation maintains compatibility with existing, YAMON-based firmware.

Tested-by: Wolfgang Grandegger <[email protected]>
Acked-by: Florian Fainelli <[email protected]>
Signed-off-by: Manuel Lauss <[email protected]>
To: Linux-MIPS <[email protected]>
Cc: [email protected]
Patchwork: https://patchwork.linux-mips.org/patch/1481/
Acked-by: David S. Miller <[email protected]>
Signed-off-by: Ralf Baechle <[email protected]>

SERIAL: 8250: Remove SERIAL_8250_AU1X00

Remove the SERIAL_8250_AU1X00 config symbol.  Instead, use the MIPS_ALCHEMY
one which is always defined when building an Au1x00-based platform.

Signed-off-by: Manuel Lauss <[email protected]>
To: Linux-MIPS <[email protected]>
Cc: Linux-serial <[email protected]>
Patchwork: https://patchwork.linux-mips.org/patch/1461/
Signed-off-by: Ralf Baechle <[email protected]>
This one depends on a previous patch (which removes SOC_AU1X00 and changes
MACH_ALCHEMY) to apply cleanly (and then actually work), so I'd love for
this to go in via the mips tree.