The powerpc kernel maintains information about logical memory blocks
in the lmb.memory structure, which is initialized and updated at boot
time, but not when memory is added or removed while the kernel is
running.

This adds a hotplug memory notifier which updates lmb.memory when
memory is added or removed.  This information is useful for eHEA
driver to find out the memory layout and holes.

NOTE: No special locking is needed for lmb_add() and lmb_remove().
Calls to these are serialized by caller. (pSeries_reconfig_chain).
Signed-off-by: NBadari Pulavarty <pbadari@us.ibm.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This makes it possible to use separate stacks for hard and soft IRQs
on 32-bit powerpc as well as on 64-bit.  The code for 32-bit is just
the 32-bit analog of the 64-bit code.

* Added allocation and initialization of the irq stacks.  We limit the
  stacks to be in lowmem for ppc32.
* Implemented ppc32 versions of call_do_softirq() and call_handle_irq()
  to switch the stack pointers
* Reworked how we do stack overflow detection.  We now keep around the
  limit of the stack in the thread_struct and compare against the limit
  to see if we've overflowed.  We can now use this on ppc64 if desired.

[ paulus@samba.org: Fixed bug on 6xx where we need to reload r9 with the
  thread_info pointer. ]
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This changes the definitions of the xchg and cmpxchg families of
functions in include/asm-powerpc/system.h to be marked __always_inline
rather than __inline__.  The reason for doing this is that we rely on
the compiler inlining them in order to eliminate the references to
__xchg_called_with_bad_pointer and __cmpxchg_called_with_bad_pointer,
which are deliberately left undefined.  Thus this change will enable
us to make the inline keyword be just a hint rather than a directive.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

* Use 'hwif->dma_base + {4,8}' instead of hwif->dma_prdtable in
  {ide,scc}_dma_setup().

* Remove no longer needed ->dma_prdtable field from ide_hwif_t.

While at it:

* Use ATA_DMA_TABLE_OFS define.
Acked-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Use 'hwif->dma_base + {1,3}' instead of hwif->dma_vendor{1,3} in
  pdc202xx_new host driver.

* Remove no longer needed ->dma_vendor{1,3} fields from ide_hwif_t.
Acked-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Add ide_pad_transfer() helper (which uses ->{in,out}put_data methods
  internally so the transfer is also padded to drive+host requirements)
  and use it instead of ide_atapi_{write_zeros,discard_data}().

* Remove no longer needed ide_atapi_{write_zeros,discard_data}().

Cc: Borislav Petkov <petkovbb@gmail.com>
Acked-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Remove no longer used ->INW and ->OUTW methods.

While at it:

* scc_pata.c: scc_ide_{out,in}w() is called only in scc_tf_{load,read}()
  so inline it there.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Add ->tf_load and ->tf_read methods to ide_hwif_t and set the default
  methods in default_hwif_transport().

* Use ->tf_{load,read} instead o calling ide_tf_{load,read}() directly.

* Make ide_tf_{load,read}() static.

There should be no functional changes caused by this patch.
Acked-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Factor out debugging code from ide_tf_load() to ide_tf_dump() helper
and update ide_tf_load() users accordingly.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Add ide_execute_pkt_cmd() helper for executing PACKET command,
then convert ATAPI device drivers to use it.

As a nice side-effect this fixes ide-{floppy,tape,scsi} w.r.t.
ide_lock taking (ide-cd was OK).
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Use ins{w,l}()/outs{w,l}() and __ide_mm_ins{w,l}()/__ide_mm_outs{w,l}()
  directly in ata_{in,out}put_data() (by using IDE_HFLAG_MMIO host flag to
  decide which I/O ops are required).

* Remove no longer needed ->INS{W,L} and ->OUTS{W,L} methods (ide-h8300,
  au1xxx-ide and scc_pata implement their own ->{in,out}put_data methods).

There should be no functional changes caused by this patch.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Add IDE_HFLAG_MMIO host flag and set it for hosts which use
  default_hwif_mmiops().

v2:
* Fix kernel panic in pmac host driver (',' should be '|').

  Thanks to Kamalesh for reporting it + testing the fix
  and to Andrew for hinting me about the source of the issue.

Cc: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Merge ->atapi_{in,out}put_bytes and ->ata_{in,out}put_data methods
  into new ->{in,out}put_data methods which take number of bytes to
  transfer as an argument and always do padding.

While at it:

* Use 'hwif' or 'drive->hwif' instead of 'HWIF(drive)'.

There should be no functional changes caused by this patch (all users
of ->ata_{in,out}put_data methods were using multiply-of-4 word counts).
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Add ->atapi_{in,out}put_bytes and ->ata_{in,out}put_data methods to
  falconide and q40ide host drivers (->ata_* methods are implemented on
  top of ->atapi_* methods so they also do byte-swapping now).

* Cleanup atapi_{in,out}put_bytes().

v2:
* Add 'struct request *rq' argument to ->ata_{in,out}put_data methods
  and don't byte-swap disk fs requests (we shouldn't un-swap fs requests
  because fs itself is stored byte-swapped on the disk) - this is how
  things were done before the patch (ideally device mapper should be
  used instead but it would break existing setups and would have some
  performance impact).

Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Michael Schmitz <schmitz@debian.org>
Cc: Roman Zippel <zippel@linux-m68k.org>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Richard Zidlicky <rz@linux-m68k.org>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

usb_control_msg() converts arguments to little-endian itself,
doing that in caller means breakage on big-endian boxen.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

psr is not a good name for local variable in macro body when it
has a good chance of being the argument of said macro (actually
is at least in one place)
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Since both the IDE interface and SMC 91C111 Ethernet chip are on-board
devices, not SOC devices, move the platform device registration form the
common to the board specific code.

While at it, remove semicolon (which didn't break compilation only by
chance) from the AU1XXX_ATA_DDMA_REQ macro and do some renaming:

- change 'au1200_ide0_' variable name prefix to the mere 'ide_';

- change 'smc91x_' variable name prefix to 'smc91c111_' since that's the
  name of the chip used on the boards;

- drop 'AU1XXX_' prefix from the names of macros describing IDE and Ethernet
  on-board devices;

- change 'SMC91111_' to 'SMC91C111_', change 'IRQ' to 'INT' in the names of
  the macros describing the Ethernet chip for consistency with the IDE
  macros;

- change 'ATA_' to 'IDE_' and 'OFFSET' to 'SHIFT' (since this value is
  indeed a shift count) in the names of the macros describing the IDE
  interface.
Signed-off-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

asm-mips/mach-ip27/topology.h must #include <asm-generic/topology.h>
This fixes the following compile error:

...
  CC      kernel/sched.o
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c: In function 'find_next_best_node':
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c:7015: error: implicit declaration of function 'node_to_cpumask_ptr'
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c:7015: error: '__tmp__' undeclared (first use in this function)
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c:7015: error: (Each undeclared identifier is reported only once
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c:7015: error: for each function it appears in.)
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c: In function 'sched_domain_node_span':
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c:7047: error: 'nodemask' undeclared (first use in this function)
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c:7048: warning: ISO C90 forbids mixed declarations and code
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c:7059: error: implicit declaration of function 'node_to_cpumask_ptr_next'
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c: In function '__build_sched_domains':
/home/bunk/linux/kernel-2.6/git/linux-2.6/kernel/sched.c:7605: error: 'pnodemask' undeclared (first use in this function)
make[2]: *** [kernel/sched.o] Error 1

<--  snip  -->
Signed-off-by: NAdrian Bunk <bunk@kernel.org>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

* Declare board_bind_eic_interrupt, board_watchpoint_handler in traps.h
* Make msc_bind_eic_interrupt static and fix its argument types.
* Make msc_levelirq_type, msc_edgeirq_type static.
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NYoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Add DECstation I/O ASIC clocksource
Signed-off-by: NYoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

* Do not use non-standard I/O accessors, such as reg_rd08, etc.
* Kill unnecessary wbflush()
* Kill tx4938_mips.h
* Kill unnecessary includes
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Use generic txx9 gpio (and gpiolib) for JMR3927 board.
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Use generic txx9 gpio (and gpiolib) for RBHMA4500 board.
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

This is a board-independent TXx9 gpio API implementation using gpiolib.
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

If gpiolib was selected, asm-generic/gpio.h provides some prototypes
for gpio API and implementation helpers.  With this patch, platform
code can implement its GPIO API using gpiolib without custom gpio.h
file.
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Defer the unmasking of the count/compare interrupt (IRQ5) till the
clockevent driver initialization:

- only enable the cascaded IRQs 0 thru 4 in arch_init_irq(); kill the
  ALLINTS macro -- this change is blessed by AMD as I saw it in their own
  patch; :-)

- do not force IRQ5 enabled in plat_time_init() if PM is enabled and there's
  no 32 KHz crystal.

Update the copyrights (taking into account my prior changes), also removing
Pete Popov's old email...
Signed-off-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: Ndaniel.j.laird <daniel.j.laird@nxp.com>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NChris Dearman <chris@mips.com>
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NChris Dearman <chris@mips.com>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Slightly tacky, but there is a precedent in the sparc archirecture code.
Signed-off-by: NChris Dearman <chris@mips.com>
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NChris Dearman <chris@mips.com>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Still, only the 4K may actually implement it.
Signed-off-by: NChris Dearman <chris@mips.com>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

The on-disk media specification field in FAT is only 8-bits, so testing for
<=0xff is pointless, and can generate a "comparison is always true due to
limited range of data type" warning.

While we're there, convert FAT_VALID_MEDIA() into a C function - the present
implementation is buggy: it generates either one or two references to its
argument.

Cc: Frank Seidel <fseidel@suse.de>
Acked-by: NOGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Currently, free_clusters is not updated until it is trusted, because
Windows doesn't update it correctly.

But if user is using FAT driver of Linux, it updates free_clusters
correctly.  Instead, this updates it even if it's untrusted, so if
free_clustes is correct, now keep correct value.
Signed-off-by: NOGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Normally utime(2) checks current process is owner of the file, or it
has CAP_FOWNER capability.  But FAT filesystem doesn't have uid/gid as
on disk info, so normal check is too unflexible.

With this option you can relax it.
Signed-off-by: NOGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

- Rename fat_notify_change() to fat_setattr()
- check_mode() cleanup
- Change layout of code
Signed-off-by: NOGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Quota files cannot have tails because quota_write and quota_read functions do
not support them.  So far when quota files did have tail, we just refused to
turn quotas on it.  Sadly this check has been wrong and so there are now
plenty installations where quota files don't have NOTAIL flag set and so now
after fixing the check, they suddently fail to turn quotas on.  Since it's
easy to unpack the tail from kernel, do this from reiserfs_quota_on() which
solves the problem and is generally nicer to users anyway.
Signed-off-by: NJan Kara <jack@suse.cz>
Reported-by: <urhausen@urifabi.net>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: Chris Mason <chris.mason@oracle.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Improve write performance by preventing the delayed_list from dumping all its
stripes onto the handle_list in one shot.  Delayed stripes are now further
delayed by being held on the 'hold_list'.  The 'hold_list' is bypassed when:

  * a STRIPE_IO_STARTED stripe is found at the head of 'handle_list'
  * 'handle_list' is empty and i/o is being done to satisfy full stripe-width
    write requests
  * 'bypass_count' is less than 'bypass_threshold'.  By default the threshold
    is 1, i.e. every other stripe handled is a preread stripe provided the
    top two conditions are false.

Benchmark data:
System: 2x Xeon 5150, 4x SATA, mem=1GB
Baseline: 2.6.24-rc7
Configuration: mdadm --create /dev/md0 /dev/sd[b-e] -n 4 -l 5 --assume-clean
Test1: dd if=/dev/zero of=/dev/md0 bs=1024k count=2048
  * patched:  +33% (stripe_cache_size = 256), +25% (stripe_cache_size = 512)

Test2: tiobench --size 2048 --numruns 5 --block 4096 --block 131072 (XFS)
  * patched: +13%
  * patched + preread_bypass_threshold = 0: +37%

Changes since v1:
* reduce bypass_threshold from (chunk_size / sectors_per_chunk) to (1) and
  make it configurable.  This defaults to fairness and modest performance
  gains out of the box.
Changes since v2:
* [neilb@suse.de]: kill STRIPE_PRIO_HI and preread_needed as they are not
  necessary, the important change was clearing STRIPE_DELAYED in
  add_stripe_bio and this has been moved out to make_request for the hang
  fix.
* [neilb@suse.de]: simplify get_priority_stripe
* [dan.j.williams@intel.com]: reset the bypass_count when ->hold_list is
  sampled empty (+11%)
* [dan.j.williams@intel.com]: decrement the bypass_count at the detection
  of stripes being naturally promoted off of hold_list +2%.  Note, resetting
  bypass_count instead of decrementing on these events yields +4% but that is
  probably too aggressive.
Changes since v3:
* cosmetic fixups
Tested-by: NJames W. Laferriere <babydr@baby-dragons.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>
Signed-off-by: NNeil Brown <neilb@suse.de>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>