The only way the debugger can handle a trap in inside rcu_lock,
notify_die, or atomic_notifier_call_chain without a recursive fault is
to have a low level "first opportunity handler" do_trap_or_bp() handler.

Generally this will be something the vast majority of folks will not
need, but for those who need it, it is added as a kernel .config
option called KGDB_LOW_LEVEL_TRAP.

Also added was a die notification for oops such that kdb can catch an
oops for analysis.

There appeared to be no obvious way to pass the struct pt_regs from
the original exception back to the stack back tracer, so a special
case was added to show_stack() for when kdb is active because you
generally desire to generally look at the back trace of the original
exception.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Acked-by: Ralf Baechle <ralf@linux-mips.org>

The only way the debugger can handle a trap in inside rcu_lock,
notify_die, or atomic_notifier_call_chain without a recursive fault is
to allow the kernel debugger to handle the exception first in
program_check_exception().

The other change here is to make sure that kgdb_handle_exception() is
called with correct parameters when catching an oops, because kdb
needs to know if the entry was an oops, single step, or breakpoint
exception.

[benh@kernel.crashing.org: move debugger_bpt instead of #ifdef]

CC: Paul Mackerras <paulus@samba.org>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

The only way the debugger can handle a trap in inside rcu_lock,
notify_die, or atomic_notifier_call_chain without a triple fault is
to have a low level "first opportunity handler" in the int3 exception
handler.

Generally this will be something the vast majority of folks will not
need, but for those who need it, it is added as a kernel .config
option called KGDB_LOW_LEVEL_TRAP.

CC: Ingo Molnar <mingo@elte.hu>
CC: Thomas Gleixner <tglx@linutronix.de>
CC: H. Peter Anvin <hpa@zytor.com>
CC: x86@kernel.org
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>

Remove all the references to the kgdb_post_primary_code.  This
function serves no useful purpose because you can obtain the same
information from the "struct kgdb_state *ks" from with in the
debugger, if for some reason you want the data.

Also remove the unintentional duplicate assignment for ks->ex_vector.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>

Update the kgdb docs to reflect the new directory structure and API.

Merge in the kdb shell information.

[Randy Dunlap <rdunlap@xenotime.net>: grammatical corrections]

CC: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>

This patch adds in the kdb PS/2 keyboard driver.  This was mostly a
direct port from the original kdb where I cleaned up the code against
checkpatch.pl and added the glue to stitch it into kgdb.

This patch also enables early kdb debug via kgdbwait and the keyboard.

All the access to configure kdb using either a serial console or the
keyboard is done via kgdboc.

If you want to use only the keyboard and want to break in early you
would add to your kernel command arguments:

    kgdboc=kbd kgdbwait

If you wanted serial and or the keyboard access you could use:

    kgdboc=kbd,ttyS0

You can also configure kgdboc as a kernel module or at run time with
the sysfs where you can activate and deactivate kgdb.

Turn it on:
    echo kbd,ttyS0 > /sys/module/kgdboc/parameters/kgdboc

Turn it off:
    echo "" > /sys/module/kgdboc/parameters/kgdboc
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>

One of the driving forces behind integrating another front end (kdb)
to the debug core is to allow front end commands to be accessible via
gdb's monitor command.  It is true that you could write gdb macros to
get certain data, but you may want to just use gdb to access the
commands that are available in the kdb front end.

This patch implements the Rcmd gdb stub packet.  In gdb you access
this with the "monitor" command.  For instance you could type "monitor
help", "monitor lsmod" or "monitor ps A" etc...

There is no error checking or command restrictions on what you can and
cannot access at this point.  Doing something like trying to set
breakpoints with the monitor command is going to cause nothing but
problems.  Perhaps in the future only the commands that are actually
known to work with the gdb monitor command will be available.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>

Allow kgdboc to work on sparc hardware with the Zilog serial chips.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Acked-by: David S. Miller <davem@davemloft.net>

The sci_poll_get_char() needs to return immediately if there is no
input from the chip to process, and must return a value of
NO_POLL_CHAR.

This allows kgdboc to process multiple polled devices while kgdb is
active.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>

The design of the kdb shell requires that every device that can
provide input to kdb have a polling routine that exits immediately if
there is no character available.  This is required in order to get the
page scrolling mechanism working.

Changing the kernel debugger I/O API to require all polling character
routines to exit immediately if there is no data allows the kernel
debugger to process multiple input channels.

NO_POLL_CHAR will be the return code to the polling routine when ever
there is no character available.

CC: linux-serial@vger.kernel.org
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>

These are the minimum changes to the kgdb core in order to enable an
API to connect a new front end (kdb) to the debug core.

This patch introduces the dbg_kdb_mode variable controls where the
user level I/O is routed.  It will be routed to the gdbstub (kgdb) or
to the kdb front end which is a simple shell available over the kgdboc
connection.

You can switch back and forth between kdb or the gdb stub mode of
operation dynamically.  From gdb stub mode you can blindly type
"$3#33", or from the kdb mode you can enter "kgdb" to switch to the
gdb stub.

The logic in the debug core depends on kdb to look for the typical gdb
connection sequences and return immediately with KGDB_PASS_EVENT if a
gdb serial command sequence is detected.  That should allow a
reasonably seamless transition between kdb -> gdb without leaving the
kernel exception state.  The two gdb serial queries that kdb is
responsible for detecting are the "?" and "qSupported" packets.

CC: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Acked-by: Martin Hicks <mort@sgi.com>

This patch contains the hooks and instrumentation into kernel which
live outside the kernel/debug directory, which the kdb core
will call to run commands like lsmod, dmesg, bt etc...

CC: linux-arch@vger.kernel.org
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Martin Hicks <mort@sgi.com>

This patch contains only the kdb core.  Because the change set was
large, it was split.  The next patch in the series includes the
instrumentation into the core kernel which are mainly helper functions
for kdb.

This work is directly derived from kdb v4.4 found at:

ftp://oss.sgi.com/projects/kdb/download/v4.4/

The kdb internals have been re-organized to make them mostly platform
independent and to connect everything to the debug core which is used by
gdbstub (which has long been known as kgdb).

The original version of kdb was 58,000 lines worth of changes to
support x86.  From that implementation only the kdb shell, and basic
commands for memory access, runcontrol, lsmod, and dmesg where carried
forward.

This is a generic implementation which aims to cover all the current
architectures using the kgdb core: ppc, arm, x86, mips, sparc, sh and
blackfin.  More archictectures can be added by implementing the
architecture specific kgdb functions.

[mort@sgi.com: Compile fix with hugepages enabled]
[mort@sgi.com: Clean breakpoint code renaming kdba_ -> kdb_]
[mort@sgi.com: fix new line after printing registers]
[mort@sgi.com: Remove the concept of global vs. local breakpoints]
[mort@sgi.com: Rework kdb_si_swapinfo to use more generic name]
[mort@sgi.com: fix the information dump macros, remove 'arch' from the names]
[sfr@canb.auug.org.au: include fixup to include linux/slab.h]

CC: linux-arch@vger.kernel.org
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Martin Hicks <mort@sgi.com>

The new debug core api requires all architectures that use to debug
core to implement a function to set the program counter.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Acked-by: Sonic Zhang <sonic.zhang@analog.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>

Implement kgdb_arch_pc() which adjusts the pc if it needs to be
adjusted after a software breakpoint trap.

Implement kgdb_arch_set_pc() which is a new required function in the
kgdb debug core.

When processing a single step return zero in the error exception field
so that the debug core can distinguish between a single step trap and
a breakpoint trap generically.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>

The new debug core api requires all architectures that use to debug
core to implement a function to set the program counter.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Acked-by: David S. Miller <davem@davemloft.net>

Update the kgdb section.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>

Split the former kernel/kgdb.c into debug_core.c which contains the
kernel debugger exception logic and to the gdbstub.c which contains
the logic for allowing gdb to talk to the debug core.

This also created a private include file called debug_core.h which
contains all the definitions to glue the debug_core to any other
debugger connections.

CC: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>

Move kgdb.c in preparation to separate the gdbstub from the debug
core and exception handling.

CC: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input: (40 commits)
  Input: psmouse - small formatting changes to better follow coding style
  Input: synaptics - set dimensions as reported by firmware
  Input: elantech - relax signature checks
  Input: elantech - enforce common prefix on messages
  Input: wistron_btns - switch to using kmemdup()
  Input: usbtouchscreen - switch to using kmemdup()
  Input: do not force selecting i8042 on Moorestown
  Input: Documentation/sysrq.txt - update KEY_SYSRQ info
  Input: 88pm860x_onkey - remove invalid irq number assignment
  Input: i8042 - add a PNP entry to the aux device list
  Input: i8042 - add some extra PNP keyboard types
  Input: wm9712 - fix wm97xx_set_gpio() logic
  Input: add keypad driver for keys interfaced to TCA6416
  Input: remove obsolete {corgi,spitz,tosa}kbd.c
  Input: kbtab - do not advertise unsupported events
  Input: kbtab - simplify kbtab_disconnect()
  Input: kbtab - fix incorrect size parameter in usb_buffer_free
  Input: acecad - don't advertise mouse events
  Input: acecad - fix some formatting issues
  Input: acecad - simplify usb_acecad_disconnect()
  ...

Trivial conflict in Documentation/feature-removal-schedule.txt

* 'acpica' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6: (22 commits)
  ACPI: fix early DSDT dmi check warnings on ia64
  ACPICA: Update version to 20100428.
  ACPICA: Update/clarify some parameter names associated with acpi_handle
  ACPICA: Rename acpi_ex_system_do_suspend->acpi_ex_system_do_sleep
  ACPICA: Prevent possible allocation overrun during object copy
  ACPICA: Split large file, evgpeblk
  ACPICA: Add GPE support for dynamically loaded ACPI tables
  ACPICA: Clarify/rename some root table descriptor fields
  ACPICA: Update version to 20100331.
  ACPICA: Minimize the differences between linux GPE code and ACPICA code base
  ACPI: add boot option acpi=copy_dsdt to fix corrupt DSDT
  ACPICA: Update DSDT copy/detection.
  ACPICA: Add subsystem option to force copy of DSDT to local memory
  ACPICA: Add detection of corrupted/replaced DSDT
  ACPICA: Add write support for DataTable operation regions
  ACPICA: Fix for acpi_reallocate_root_table for incorrect root table copy
  ACPICA: Update comments/headers, no functional change
  ACPICA: Update version to 20100304
  ACPICA: Fix for possible fault in acpi_ex_release_mutex
  ACPICA: Standardize integer output for ACPICA warnings/errors
  ...

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6: (250 commits)
  ALSA: hda: Storage class should be before const qualifier
  ASoC: tpa6130a2: Remove CPVSS and HPVdd supplies
  ASoC: tpa6130a2: Define output pins with SND_SOC_DAPM_OUTPUT
  ASoC: sdp4430 - add sdp4430 pcm ops to DAI.
  ASoC: TWL6040: Enable earphone path in codec
  ASoC: SDP4430: Add support for Earphone speaker
  ASoC: SDP4430: Add sdp4430 machine driver
  ASoC: tlv320dac33: Avoid powering off while in BIAS_OFF
  ASoC: tlv320dac33: Use dev_dbg in dac33_hard_power function
  ALSA: sound/pci/asihpi: Use kzalloc
  ALSA: hdmi - dont fail on extra nodes
  ALSA: intelhdmi - add id for the CougarPoint chipset
  ALSA: intelhdmi - user friendly codec name
  ALSA: intelhdmi - add dependency on SND_DYNAMIC_MINORS
  ALSA: asihpi: incorrect range check
  ALSA: asihpi: testing the wrong variable
  ALSA: es1688: add pedantic range checks
  ARM: McBSP: Add support for omap4 in McBSP driver
  ARM: McBSP: Fix request for irq in OMAP4
  OMAP: McBSP: Add 32-bit mode support
  ...

* 'for-linus/i2c-2635' of git://git.fluff.org/bjdooks/linux: (21 commits)
  i2c-highlander: remover superflous variable
  i2c-ibm-iic: drop NO_IRQ
  i2c-cpm: drop NO_IRQ
  i2c-mpc: drop NO_IRQ
  MAINTAINERS: add i2c tree for embedded platforms
  i2c-pxa: only define 'blue_murder'-function if DEBUG is #defined
  i2c-pxa: remove unused macro
  i2c-nomadik: fix operator precedence warning
  i2c-nomadik: release region when removed
  OMAP3: I2C: Clean up Errata 1p153 handling
  OMAP2/3: I2C: Errata ID i207: Clear wrong RDR interrupt
  omap: i2c: add a timeout to the busy waiting
  omap: i2c: make errata 1.153 workaround a separate function
  i2c-omap: add mpu wake up latency constraint in i2c
  omap: i2c: Add i2c support on omap4 platform
  i2c-bfin-twi: return completion in interrupt for smbus quick transfers
  i2c-bfin-twi: remove redundant retry
  i2c-bfin-twi: fix lost interrupts at high speeds
  i2c-bfin-twi: add debug output for error status
  i2c-bfin-twi: integrate timeout timer with completion interface
  ...

* 'v4l_for_2.6.35' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-2.6: (534 commits)
  V4L/DVB (13554a): v4l: Use the video_drvdata function in drivers
  V4L/DVB: vivi and mem2mem_testdev need slab.h to build
  V4L/DVB: tm6000: bugfix image position
  V4L/DVB: IR/imon: remove dead IMON_KEY_RELEASE_OFFSET
  V4L/DVB: tm6000: README - add vbi
  V4L/DVB: Fix unlock logic at medusa_video_init
  V4L/DVB: fix dvb frontend lockup
  V4L/DVB: s2255drv: remove dead code
  V4L/DVB: s2255drv: return if vdev not found
  V4L/DVB: ov511: cleanup: remove unneeded null check
  V4L/DVB: media/mem2mem: dereferencing free memory
  V4L/DVB: media/IR: Add missing include file to rc-map.c
  V4L/DVB: dvb/stv6110x: cleanup error handling
  V4L/DVB: ngene: Add lgdt3303 and mt2131 deps to Kconfig
  V4L/DVB: ngene: start separating out DVB functions into separate file
  V4L/DVB: ngene: split out card specific code into a separate file
  V4L/DVB: ngene: split out i2c code into a separate file
  V4L/DVB: ngene: add initial support for digital side of Avermedia m780
  V4L/DVB: ngene: properly support boards where channel 0 isn't a TS input
  V4L-DVB: ngene: make sure that tuner headers are included
  ...

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev: (49 commits)
  libata-sff: separate out BMDMA qc_issue
  libata-sff: prd is BMDMA specific
  libata-sff: ata_sff_[dumb_]qc_prep are BMDMA specific
  libata-sff: separate out BMDMA EH
  libata-sff: port_task is SFF specific
  libata-sff: ap->[last_]ctl are SFF specific
  libata-sff: rename ap->ops->drain_fifo() to sff_drain_fifo()
  libata-sff: introduce ata_sff_init/exit() and ata_sff_port_init()
  libata-sff: clean up BMDMA initialization
  libata-sff: clean up inheritance in several drivers
  libata-sff: reorder SFF/BMDMA functions
  sata_inic162x: kill PORT_PRD_ADDR initialization
  libata: kill ATA_FLAG_DISABLED
  libata-sff: kill unused prototype and make ata_dev_select() static
  libata-sff: update bmdma host bus error handling
  sata_mv: remove unnecessary initialization
  sata_inic162x: inic162x is not dependent on CONFIG_ATA_SFF
  pata_sch: use ata_pci_sff_init_one()
  pata_sil680: Do our own exec_command posting
  libata: Remove excess delay in the tf_load path
  ...

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial: (44 commits)
  vlynq: make whole Kconfig-menu dependant on architecture
  add descriptive comment for TIF_MEMDIE task flag declaration.
  EEPROM: max6875: Header file cleanup
  EEPROM: 93cx6: Header file cleanup
  EEPROM: Header file cleanup
  agp: use NULL instead of 0 when pointer is needed
  rtc-v3020: make bitfield unsigned
  PCI: make bitfield unsigned
  jbd2: use NULL instead of 0 when pointer is needed
  cciss: fix shadows sparse warning
  doc: inode uses a mutex instead of a semaphore.
  uml: i386: Avoid redefinition of NR_syscalls
  fix "seperate" typos in comments
  cocbalt_lcdfb: correct sections
  doc: Change urls for sparse
  Powerpc: wii: Fix typo in comment
  i2o: cleanup some exit paths
  Documentation/: it's -> its where appropriate
  UML: Fix compiler warning due to missing task_struct declaration
  UML: add kernel.h include to signal.c
  ...

* git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6: (29 commits)
  pcmcia: disable PCMCIA ioctl also for ARM
  drivers/staging/comedi: dev_node removal (quatech_daqp_cs)
  drivers/staging/comedi: dev_node removal (ni_mio_cs)
  drivers/staging/comedi: dev_node removal (ni_labpc_cs)
  drivers/staging/comedi: dev_node removal (ni_daq_dio24)
  drivers/staging/comedi: dev_node removal (ni_daq_700)
  drivers/staging/comedi: dev_node removal (das08_cs)
  drivers/staging/comedi: dev_node removal (cb_das16_cs)
  pata_pcmcia: get rid of extra indirection
  pcmcia: remove suspend-related comment from yenta_socket.c
  pcmcia: call pcmcia_{read,write}_cis_mem with ops_mutex held
  pcmcia: remove pcmcia_add_device_lock
  pcmcia: update gfp/slab.h includes
  pcmcia: remove unused mem_op.h
  pcmcia: do not autoadd root PCI bus resources
  pcmcia: clarify alloc_io_space, move it to resource handlers
  pcmcia: move all pcmcia_resource_ops providers into one module
  pcmcia: move high level CIS access code to separate file
  pcmcia: dev_node removal (core)
  pcmcia: dev_node removal (remaining drivers)
  ...

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/suspend-2.6:
  PM: PM QOS update fix
  Freezer / cgroup freezer: Update stale locking comments
  PM / platform_bus: Allow runtime PM by default
  i2c: Fix bus-level power management callbacks
  PM QOS update
  PM / Hibernate: Fix block_io.c printk warning
  PM / Hibernate: Group swap ops
  PM / Hibernate: Move the first_sector out of swsusp_write
  PM / Hibernate: Separate block_io
  PM / Hibernate: Snapshot cleanup
  FS / libfs: Implement simple_write_to_buffer
  PM / Hibernate: document open(/dev/snapshot) side effects
  PM / Runtime: Add sysfs debug files
  PM: Improve device power management document
  PM: Update device power management document
  PM: Allow runtime_suspend methods to call pm_schedule_suspend()
  PM: pm_wakeup - switch to using bool

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: change cancel_work_sync() to clear work->data
  workqueue: warn about flush_scheduled_work()

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu:
  ia64: add sparse annotation to __ia64_per_cpu_var()
  percpu: implement kernel memory based chunk allocation
  percpu: move vmalloc based chunk management into percpu-vm.c
  percpu: misc preparations for nommu support
  percpu: reorganize chunk creation and destruction
  percpu: factor out pcpu_addr_in_first/reserved_chunk() and update per_cpu_ptr_to_phys()

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/core: Use kmemdup() instead of kmalloc()+memcpy()
  IB/iser: Fix error flow in iser_create_ib_conn_res()
  IB/iser: Enhance disconnection logic for multi-pathing
  IB/iser: Remove buggy back-pointer setting
  IB/iser: Add asynchronous event handler
  MAINTAINERS: Add cxgb4 and iw_cxgb4 entries
  RDMA/cxgb3: Shrink .text with compile-time init of handlers arrays
  IPoIB: Allow disabling/enabling TSO on the fly through ethtool
  IB/mlx4: Add support for masked atomic operations
  IB/core: Add support for masked atomic operations
  RDMA/cma: Randomize local port allocation
  RDMA/nes: Make unnecessarily global functions static
  RDMA/nes: Make nesadapter->phy_lock usage consistent
  RDMA/cxgb4: Add driver for Chelsio T4 RNIC
  IB/mthca: Use the dma state API instead of pci equivalents
  RDMA/amso1100: Use the dma state API instead of pci equivalents
  RDMA/cxgb3: Don't free skbs on NET_XMIT_* indications from LLD
  RDMA/cxgb3: Use the dma state API instead of pci equivalents
  IB: Explicitly rule out llseek to avoid BKL in default_llseek()

Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6: (61 commits)
  KEYS: Return more accurate error codes
  LSM: Add __init to fixup function.
  TOMOYO: Add pathname grouping support.
  ima: remove ACPI dependency
  TPM: ACPI/PNP dependency removal
  security/selinux/ss: Use kstrdup
  TOMOYO: Use stack memory for pending entry.
  Revert "ima: remove ACPI dependency"
  Revert "TPM: ACPI/PNP dependency removal"
  KEYS: Do preallocation for __key_link()
  TOMOYO: Use mutex_lock_interruptible.
  KEYS: Better handling of errors from construct_alloc_key()
  KEYS: keyring_serialise_link_sem is only needed for keyring->keyring links
  TOMOYO: Use GFP_NOFS rather than GFP_KERNEL.
  ima: remove ACPI dependency
  TPM: ACPI/PNP dependency removal
  selinux: generalize disabling of execmem for plt-in-heap archs
  LSM Audit: rename LSM_AUDIT_NO_AUDIT to LSM_AUDIT_DATA_NONE
  CRED: Holding a spinlock does not imply the holding of RCU read lock
  SMACK: Don't #include Ext2 headers
  ...

Conflicts:
	sound/soc/codecs/ad1938.c

The C99 specification states in section 6.11.5:

The placement of a storage-class specifier other than at the beginning
of the declaration specifiers in a declaration is an obsolescent
feature.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: Takashi Iwai <tiwai@suse.de>