[PATCH] sysfs: change uses of f_{dentry, vfsmnt} to use f_path

Change all the uses of f_{dentry,vfsmnt} to f_path.{dentry,mnt} in the sysfs
filesystem code.

Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] VFS: change struct file to use struct path

This patch changes struct file to use struct path instead of having
independent pointers to struct dentry and struct vfsmount, and converts all
users of f_{dentry,vfsmnt} in fs/ to use f_path.{dentry,mnt}.

Additionally, it adds two #define's to make the transition easier for users of
the f_dentry and f_vfsmnt.

Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] struct path: make eCryptfs a user of struct path

Convert eCryptfs dentry-vfsmount pairs in dentry private data to struct
path.

Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Cc: Michael Halcrow <mhalcrow@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] struct path: move struct path from fs/namei.c into include/linux

Moved struct path from fs/namei.c to include/linux/namei.h.  This allows many
places in the VFS, as well as any stackable filesystem to easily keep track of
dentry-vfsmount pairs.

Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] struct path: rename DM's struct path

Rename DM's struct path to struct dm_path to prevent name collision between it
and struct path from fs/namei.c.

Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Acked-by: Alasdair G Kergon <agk@redhat.com>
Cc: <reiserfs-dev@namesys.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] struct path: rename Reiserfs's struct path

Rename Reiserfs's struct path to struct treepath to prevent name collision
between it and struct path from fs/namei.c.

Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Cc: <reiserfs-dev@namesys.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] eCryptfs: Use fsstack's generic copy inode attr functions

Replace eCryptfs specific code & calls with the more generic fsstack
equivalents and remove the eCryptfs specific functions.

Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Cc: Michael Halcrow <mhalcrow@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] fsstack: Introduce fsstack_copy_{attr,inode}_*

Introduce several fsstack_copy_* functions which allow stackable filesystems
(such as eCryptfs and Unionfs) to easily copy over (currently only) inode
attributes.  This prevents code duplication and allows for code reuse.

[akpm@osdl.org: Remove unneeded wrapper]
[bunk@stusta.de: fs/stack.c should #include <linux/fs_stack.h>]
Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Cc: Michael Halcrow <mhalcrow@us.ibm.com>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] isdn/gigaset: use bitrev8

Use bitrev8 for gigaset isdn driver.

Acked-by: Karsten Keil <kkeil@suse.de>
Cc: Kai Germaschewski <kai.germaschewski@gmx.de>
Acked-by: Hansjoerg Lipp <hjlipp@web.de>
Acked-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] atm/ambassador: use bitrev8

Use bitrev8 for ambassador driver.

Cc: Chas Williams <linux-atm-general@lists.sourceforge.net>
Cc: Giuliano Procida at Madge Networks <gprocida@madge.com>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] isdn/hisax: use bitrev8

Use bitrev8 for hisax_st5481 driver.

Acked-by: Karsten Keil <kkeil@suse.de>
Cc: Kai Germaschewski <kai.germaschewski@gmx.de>
Cc: Frode Isaksen <fisaksen@bewan.com>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] video: use bitrev8

Use bitrev8 for nvidiafb, rivafb, and tgafb drivers

Cc: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] crc32: replace bitreverse by bitrev32

This patch replaces bitreverse() by bitrev32.  The only users of bitreverse()
are crc32 itself and via-velocity.

Cc: Jeff Garzik <jgarzik@pobox.com>
Cc: Matt Domsch <Matt_Domsch@dell.com>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] bit reverse library

This patch provides two bit reverse functions and bit reverse table.

- reverse the order of bits in a u32 value

u8 bitrev8(u8 x);

- reverse the order of bits in a u32 value

u32 bitrev32(u32 x);

- byte reverse table

const u8 byte_rev_table[256];

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] Fix generic WARN_ON message

A warning is a warning, not a BUG.

Signed-off-by: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] use generic BUG for ppc

Switch ppc over to using the generic BUG implementation.

Signed-off-by: Judith Lebzelter <judith@osdl.org>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Michael Ellerman <michael@ellerman.id.au>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] UML: add generic BUG support

The BUG changes in -mm3 need some arch support.  This patch adds the UML
support needed.  For the most part, it was stolen from the underlying
architecture.  The exception is the kernel eip < PAGE_OFFSET test, which is
wrong for skas mode UMLs.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] Generic BUG for x86-64

This makes x86-64 use the generic BUG machinery.

The main advantage in using the generic BUG machinery for x86-64 is that
the inlined overhead of BUG is just the ud2a instruction; the file+line
information are no longer inlined into the instruction stream.  This
reduces cache pollution.

Signed-off-by: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Andi Kleen <ak@muc.de>
Cc: Hugh Dickens <hugh@veritas.com>
Cc: Michael Ellerman <michael@ellerman.id.au>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] Generic BUG for i386

This makes i386 use the generic BUG machinery.  There are no functional
changes from the old i386 implementation.

The main advantage in using the generic BUG machinery for i386 is that the
inlined overhead of BUG is just the ud2a instruction; the file+line(+function)
information are no longer inlined into the instruction stream.  This reduces
cache pollution, and makes disassembly work properly.

Signed-off-by: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Andi Kleen <ak@muc.de>
Cc: Hugh Dickens <hugh@veritas.com>
Cc: Michael Ellerman <michael@ellerman.id.au>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] Generic BUG implementation

This patch adds common handling for kernel BUGs, for use by architectures as
they wish.  The code is derived from arch/powerpc.

The advantages of having common BUG handling are:
 - consistent BUG reporting across architectures
 - shared implementation of out-of-line file/line data
 - implement CONFIG_DEBUG_BUGVERBOSE consistently

This means that in inline impact of BUG is just the illegal instruction
itself, which is an improvement for i386 and x86-64.

A BUG is represented in the instruction stream as an illegal instruction,
which has file/line information associated with it.  This extra information is
stored in the __bug_table section in the ELF file.

When the kernel gets an illegal instruction, it first confirms it might
possibly be from a BUG (ie, in kernel mode, the right illegal instruction).
It then calls report_bug().  This searches __bug_table for a matching
instruction pointer, and if found, prints the corresponding file/line
information.  If report_bug() determines that it wasn't a BUG which caused the
trap, it returns BUG_TRAP_TYPE_NONE.

Some architectures (powerpc) implement WARN using the same mechanism; if the
illegal instruction was the result of a WARN, then report_bug(Q) returns
CONFIG_DEBUG_BUGVERBOSE; otherwise it returns BUG_TRAP_TYPE_BUG.

lib/bug.c keeps a list of loaded modules which can be searched for __bug_table
entries.  The architecture must call
module_bug_finalize()/module_bug_cleanup() from its corresponding
module_finalize/cleanup functions.

Unsetting CONFIG_DEBUG_BUGVERBOSE will reduce the kernel size by some amount.
At the very least, filename and line information will not be recorded for each
but, but architectures may decide to store no extra information per BUG at
all.

Unfortunately, gcc doesn't have a general way to mark an asm() as noreturn, so
architectures will generally have to include an infinite loop (or similar) in
the BUG code, so that gcc knows execution won't continue beyond that point.
gcc does have a __builtin_trap() operator which may be useful to achieve the
same effect, unfortunately it cannot be used to actually implement the BUG
itself, because there's no way to get the instruction's address for use in
generating the __bug_table entry.

[randy.dunlap@oracle.com: Handle BUG=n, GENERIC_BUG=n to prevent build errors]
[bunk@stusta.de: include/linux/bug.h must always #include <linux/module.h]
Signed-off-by: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Andi Kleen <ak@muc.de>
Cc: Hugh Dickens <hugh@veritas.com>
Cc: Michael Ellerman <michael@ellerman.id.au>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] bdev: fix ->bd_part_count leak

Don't leak a ->bd_part_count when the partition open fails with -ENXIO.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] lockdep: avoid lockdep warning in md

md_open takes ->reconfig_mutex which causes lockdep to complain.  This
(normally) doesn't have deadlock potential as the possible conflict is with a
reconfig_mutex in a different device.

I say "normally" because if a loop were created in the array->member hierarchy
a deadlock could happen.  However that causes bigger problems than a deadlock
and should be fixed independently.

So we flag the lock in md_open as a nested lock.  This requires defining
mutex_lock_interruptible_nested.

Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] lockdep: use mutex_lock_nested for bd_mutex to avoid lockdep warning

Now that the nesting in blkdev_{get,put} is simpler, adding mutex_lock_nested
is trivial.

Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] lockdep: simplify some aspects of bd_mutex nesting

When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition.  The involves some limited recursion.
This patch tries to simplify some aspects of this.

As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).

The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.

More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).

If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success).  The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.

Note that this slightly changes the semantics of bd_part_count.  Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release.  This is an adequate semantic as it is only
ever tested for "== 0".

Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(...  1).  Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1).  This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.

Finally the mutex_lock on whole->bd_mutex in do_open can be removed.  It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.

The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.

Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] lockdep: remove lock_key approach to managing nested bd_mutex locks

The extra call to get_gendisk is not good.  It causes a ->probe and possible
module load before it is really appropriate to do this.

Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] new bd_mutex lockdep annotation

Use the gendisk partition number to set a lock class.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Neil Brown <neilb@cse.unsw.edu.au>
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Jason Baron <jbaron@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] remove the old bd_mutex lockdep annotation

Remove the old complex and crufty bd_mutex annotation.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Neil Brown <neilb@cse.unsw.edu.au>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Jason Baron <jbaron@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] pktcdvd: add sysfs and debugfs interface

Add a sysfs and debugfs interface to the pktcdvd driver.

Look into the Documentation/ABI/testing/* files in the patch for more info.

Signed-off-by: Thomas Maier <balagi@justmail.de>
Signed-off-by: Peter Osterlund <petero2@telia.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] pktcdvd: bio write congestion using congestion_wait()

This adds a bio write queue congestion control to the pktcdvd driver with
fixed on/off marks.  It prevents that the driver consumes a unlimited
amount of write requests.

[akpm@osdl.org: sync with congestion_wait() renaming]
Signed-off-by: Thomas Maier <balagi@justmail.de>
Cc: Peter Osterlund <petero2@telia.com>
Cc: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] pktcdvd: make procfs interface optional

pktcdvd: Update Kconfig help text.

Signed-off-by: Thomas Maier <balagi@justmail.de>
Signed-off-by: Peter Osterlund <petero2@telia.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] pktcdvd: reusability of procfs functions

This patch makes some of the procfs functions reusable (for
coming sysfs patch e.g.):
pkt_setup_dev()
pkt_remove_dev()
...

Signed-off-by: Thomas Maier <balagi@justmail.de>
Signed-off-by: Peter Osterlund <petero2@telia.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] sys_unshare: remove a broken CLONE_SIGHAND code

sys_unshare(CLONE_SIGHAND) is broken, the code under 'if (new_sigh)' is
never executed but very wrong. Just remove it to avoid a confusion,
task_lock() has nothing to do with ->sighand changing.

Also, change the comment in unshare_sighand(). Yes, CLONE_THREAD implies
CLONE_SIGHAND, but still it looks confusing. Also, we don't need to check
current->sighand != NULL.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] make set_special_pids() static

Make set_special_pids() static, the only caller is daemonize().

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] do_acct_process(): don't take tty_mutex

No need to take the global tty_mutex, signal->tty->driver can't go away while
we are holding ->siglock.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] do_task_stat(): don't take tty_mutex

->signal->tty is protected by ->siglock, no need to take the global tty_mutex.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] tty: ->signal->tty locking

Fix the locking of signal->tty.

Use ->sighand->siglock to protect ->signal->tty; this lock is already used
by most other members of ->signal/->sighand.  And unless we are 'current'
or the tasklist_lock is held we need ->siglock to access ->signal anyway.

(NOTE: sys_unshare() is broken wrt ->sighand locking rules)

Note that tty_mutex is held over tty destruction, so while holding
tty_mutex any tty pointer remains valid.  Otherwise the lifetime of ttys
are governed by their open file handles.  This leaves some holes for tty
access from signal->tty (or any other non file related tty access).

It solves the tty SLAB scribbles we were seeing.

(NOTE: the change from group_send_sig_info to __group_send_sig_info needs to
       be examined by someone familiar with the security framework, I think
       it is safe given the SEND_SIG_PRIV from other __group_send_sig_info
       invocations)

[schwidefsky@de.ibm.com: 3270 fix]
[akpm@osdl.org: various post-viro fixes]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Alan Cox <alan@redhat.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Chris Wright <chrisw@sous-sol.org>
Cc: Roland McGrath <roland@redhat.com>
Cc: Stephen Smalley <sds@tycho.nsa.gov>
Cc: James Morris <jmorris@namei.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jeff Dike <jdike@addtoit.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Jan Kara <jack@ucw.cz>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] time: re-add verify_pmtmr_rate

This patch re-adds the verify_pmtmr_rate functionality from 2.6.17 that
I dropped 2.6.18.

This resolves problems seen on older K6 ASUS boards where the ACPI PM
timer runs too fast.

See:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=211902
http://bugme.osdl.org/show_bug.cgi?id=2375

Thanks to Ian Campbell for re-reporting this and testing the fix!

Signed-off-by: John Stultz <johnstul@us.ibm.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Ian Campbell <ijc@hellion.org.uk>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Roman Zippel <zippel@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] Fix linux banner utsname information

utsname information is shown in the linux banner, which also is used for
/proc/version (which can have different utsname values inside a uts
namespaces).  this patch makes the varying data arguments and changes the
string to a format string, using those arguments.

Signed-off-by: Herbert Poetzl <herbert@13thfloor.at>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] CPEI gets warning at kernel/irq/migration.c:27/move_masked_irq()

While running my MCA test (hardware error injection) on 2.6.19,
I got some warning like following:

> BUG: warning at kernel/irq/migration.c:27/move_masked_irq()
>
> Call Trace:
>  [<a000000100013d20>] show_stack+0x40/0xa0
>                                 sp=e00000006b2578d0 bsp=e00000006b2510b0
>  [<a000000100013db0>] dump_stack+0x30/0x60
>                                 sp=e00000006b257aa0 bsp=e00000006b251098
>  [<a0000001000de430>] move_masked_irq+0xb0/0x240
>                                 sp=e00000006b257aa0 bsp=e00000006b251070
>  [<a0000001000de6a0>] move_native_irq+0xe0/0x180
>                                 sp=e00000006b257aa0 bsp=e00000006b251040
>  [<a00000010004ff50>] iosapic_end_level_irq+0x30/0xe0
>                                 sp=e00000006b257aa0 bsp=e00000006b251020
>  [<a0000001000d94d0>] __do_IRQ+0x170/0x400
>                                 sp=e00000006b257aa0 bsp=e00000006b250fd8
>  [<a0000001000116f0>] ia64_handle_irq+0x1b0/0x260
>                                 sp=e00000006b257aa0 bsp=e00000006b250fa8
>  [<a00000010000c3a0>] ia64_leave_kernel+0x0/0x280
>                                 sp=e00000006b257aa0 bsp=e00000006b250fa8
>  [<a000000100690cf0>] _spin_unlock_irqrestore+0x30/0x60
>                                 sp=e00000006b257c70 bsp=e00000006b250f90

It comes from:

[kernel/irq/migration.c]
  26         if (CHECK_IRQ_PER_CPU(desc->status)) {
  27                 WARN_ON(1);
  28                 return;
  29         }

By putting some printk in kernel, I found that irqbalance is trying to
move CPEI which is handled as PER_CPU irq. That's why.

CPEI(Corrected Platform Error Interrupt) is ia64 specific irq, is
allowed to pin to particular processor which selected by the platform, and
even it is PER_CPU but it has set_affinity handler (=iosapic_set_affinity)
as same as other IO-SAPIC-level interrupts. (I don't know why, but
I guess that there would be typical situation where the handler for
migration is needed, such as hotplug - the processor going to be
offline/hot-removed.)

To shut up this warning, there are 2 way at least:
 a) fix CPEI stuff
 b) prohibit setting affinity to PER_CPU irq

I'm not sure what stuff of CPEI need to be fixed, but I think that
returning error to attempting move PER_CPU irq is useful for all
applications since it will never work.

Following small patch takes b) style.
It works, the warning disappeared and irqbalance still runs well.

Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Acked-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] move kallsyms data to .rodata

Kallsyms data is never written to, so it can as well benefit from
CONFIG_DEBUG_RODATA.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] m32r: bootloader support for OPSPUT platform

This patch supports "m32r-g00ff" bootloader for an OPSPUT platform.

Applying this patch, it is possible to do ATA-boot from an IDE drive or
HTTP-boot from network by m32r-g00ff.

    * arch/m32r/boot/compressed/m32r_sio.c: Fix hangup on OPSPUT at boot.

    * arch/m32r/kernel/io_opsput.c: IDE support for OPSPUT.
    * arch/m32r/kernel/setup_opsput.c: ditto.
    * include/asm-m32r/ide.h: ditto.

Signed-off-by: Kazuhiro Inaoka <inaoka@linux-m32r.org>
Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] m32r: fix ace_handler to pass full 32-bit address

Don't mask the lower 12-bit of the page fault address.

In the current m32r kernel implementation, we use an access exception
to detect page faults.

This patch fixes ace_handler (access exception handler) for m32r.  In order to
check userspace address in do_page_fault, we have to pass full 32-bit address
to do_page_fault.

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] m32r: support a synthesizable M32700 core

This patch is for supporting a synthesizable M32700 core for the Mappi-II FPGA
board.

On the core, location of MFT (Multi-Function Timer) registers is slightly
different from the M32700 chip.

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] m32r: make userspace headers platform-independent

The m32r kernel 2.6.18-rc1 or after cause build errors of "unknown isa
configuration" for userspace application programs, such as glibc, gdb, etc.

This is because the recent kernel do not include linux/config.h not to expose
kernel headers for userspace.

To fix the above compile errors, this patch fixes two headers ptrace.h and
sigcontext.h for m32r and makes them platform-independent.

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] mm: fallback_alloc cpuset_zone_allowed irq fix

fallback_alloc() could end up calling cpuset_zone_allowed() with interrupts
disabled (by code in kmem_cache_alloc_node()), but without __GFP_HARDWALL
set, leading to a possible call of a sleeping function with interrupts
disabled.

This results in the BUG report:

  BUG: sleeping function called from invalid context at kernel/cpuset.c:1520
in_atomic():0, irqs_disabled():1

Thanks to Paul Menage for catching this one.

Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ext4 calls journal_stop

journal_stop() is not defined for ext4; change to ext4_journal_stop().

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Input: ucb1400 - fix compile error

try_to_freeze() was moved into include/linux/freezer.h

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: kill maple_keyb.c driver

The bus for this was removed entirely some time ago, as well as most
of the drivers that referenced it. maple_keyb seems to have been the
odd one out, and was still sitting in the source tree (though not
actually part of the build system). Kill off the rest of it..

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: i8042 - add another Lifebook P7010 to nomux blacklist

We already had entry for Fujitsu Lifebook P7010 in the nomux
blacklist but for some reason Fujitsu decided to fiddle with
DMI data...

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Merge /pub/scm/linux/kernel/git/torvalds/linux-2.6

Conflicts:

drivers/usb/input/hid.h

Merge branch 'for-linus' of git://git./linux/kernel/git/ieee1394/linux1394-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6: (55 commits)
  ieee1394: sbp2: code formatting around work_struct stuff
  ieee1394: nodemgr: remove a kcalloc
  ieee1394: conditionally export ieee1394_bus_type
  ieee1394: Consolidate driver registering
  ieee1394: sbp2: convert from PCI DMA to generic DMA
  ieee1394: nodemgr: spaces to tabs
  ieee1394: nodemgr: fix deadlock in shutdown
  ieee1394: nodemgr: remove duplicate assignment
  sbp2: make 1bit bitfield unsigned
  ieee1394: schedule *_oui sysfs attributes for removal
  ieee1394: schedule unused symbol exports for removal
  ieee1394: dv1394: schedule for feature removal
  ieee1394: raw1394: defer feature removal of old isoch interface
  ieee1394: ohci1394: call PMac code in shutdown only for proper machines
  ieee1394: ohci1394: reformat PPC_PMAC platform code
  ieee1394: ohci1394: add PPC_PMAC platform code to driver probe
  ieee1394: sbp2: wrap two functions into one
  ieee1394: sbp2: update comment on things to do
  ieee1394: sbp2: use list_move_tail()
  ieee1394: sbp2: more concise names for types and variables
  ...

Merge branch 'for-linus' of /home/rmk/linux-2.6-arm

* 'for-linus' of master.kernel.org:/home/rmk/linux-2.6-arm: (76 commits)
  [ARM] 4002/1: S3C24XX: leave parent IRQs unmasked
  [ARM] 4001/1: S3C24XX: shorten reboot time
  [ARM] 3983/2: remove unused argument to __bug()
  [ARM] 4000/1: Osiris: add third serial port in
  [ARM] 3999/1: RX3715: suspend to RAM support
  [ARM] 3998/1: VR1000: LED platform devices
  [ARM] 3995/1: iop13xx: add iop13xx support
  [ARM] 3968/1: iop13xx: add iop13xx_defconfig
  [ARM] Update mach-types
  [ARM] Allow gcc to optimise arm_add_memory a little more
  [ARM] 3991/1: i.MX/MX1 high resolution time source
  [ARM] 3990/1: i.MX/MX1 more precise PLL decode
  [ARM] 3986/1: H1940: suspend to RAM support
  [ARM] 3985/1: ixp4xx clocksource cleanup
  [ARM] 3984/1: ixp4xx/nslu2: Fix disk LED numbering (take 2)
  [ARM] 3994/1: ixp23xx: fix handling of pci master aborts
  [ARM] 3981/1: sched_clock for PXA2xx
  [ARM] 3980/1: extend the ARM Versatile sched_clock implementation from 32 to 63 bit
  [ARM] 3979/1: extend the SA11x0 sched_clock implementation from 32 to 63 bit period
  [ARM] 3978/1: macro to provide a 63-bit value from a 32-bit hardware counter
  ...

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

* 'release' of master.kernel.org:/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] replace kmalloc+memset with kzalloc
  [IA64] resolve name clash by renaming is_available_memory()
  [IA64] Need export for csum_ipv6_magic
  [IA64] Fix DISCONTIGMEM without VIRTUAL_MEM_MAP
  [PATCH] Add support for type argument in PAL_GET_PSTATE
  [IA64] tidy up return value of ip_fast_csum
  [IA64] implement csum_ipv6_magic for ia64.
  [IA64] More Itanium PAL spec updates
  [IA64] Update processor_info features
  [IA64] Add se bit to Processor State Parameter structure
  [IA64] Add dp bit to cache and bus check structs
  [IA64] SN: Correctly update smp_affinty mask
  [IA64] sparse cleanups
  [IA64] IA64 Kexec/kdump

[ARM] Merge individual ARM sub-trees

Merge:
 Atmel AT91RM9200 and AT91SAM9260 changes
 General ARM developments
 Disconfiguous memory cleanups
 64-bit/32-bit division and sched_clock extension patches
 EP93xx support changes
 IOP support changes

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

Merge branch 'intx' of /linux/kernel/git/jgarzik/misc-2.6

* 'intx' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/misc-2.6:
  PCI MSI: always toggle legacy-INTx-enable bit upon MSI entry/exit

[ARM] 4002/1: S3C24XX: leave parent IRQs unmasked

Do not bother masking/unmasking the parent IRQ
for the mulitplexed EINT irqs, as masking the
leaf seems to be fine.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

[ARM] 4001/1: S3C24XX: shorten reboot time

Cut down the time between requesting a reboot
and actually getting the reboot to happen by
a quarter.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

PCI MSI: always toggle legacy-INTx-enable bit upon MSI entry/exit

The current code (prior to this change) would disable the PCI INTx
legacy interrupt when enabling MSI... but only on PCI Express.  We
should do this for all MSI devices, for safety's sake.

Signed-off-by: Jeff Garzik <jeff@garzik.org>

[ARM] 3983/2: remove unused argument to __bug()

It appears that include/asm-arm/bug.h requires include/linux/stddef.h
for the definition of NULL. It seems that stddef.h was always included
indirectly in most cases, and that issue was properly fixed a while ago.

Then commit 5047f09b56d0bc3c21aec9cb16de60283da645c6 incorrectly reverted
change from commit ff10952a547dad934d9ed9afc5cf579ed1ccb53a (bad dwmw2)
and the problem recently resurfaced.

Because the third argument to __bug() is never used anyway, RMK suggested
getting rid of it entirely instead of readding #include <linux/stddef.h>
which this patch does.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

[ARM] 4000/1: Osiris: add third serial port in

Add the mapping for the third S3C2440 serial
port into the default serial devices for the
Osiris.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

[ARM] 3999/1: RX3715: suspend to RAM support

The RX3715 is similar to the H1940 in the way
that suspend to RAM works, so we can use most
of the extant support for the H1940 with only
a few modifictions

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

[ARM] 3998/1: VR1000: LED platform devices

Support for the GPIO attached LEDs on the VR1000

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

ieee1394: sbp2: code formatting around work_struct stuff

Merge is finished, can bring the code in readable style again.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: nodemgr: remove a kcalloc

Was I sleepwalking when I wrote this?

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: conditionally export ieee1394_bus_type

Follow-up to patch "Consolidate driver registering":
Since I plan the lifetime of Linux 2.6.20 to be the deprecation phase
of CONFIG_IEEE1394_EXPORT_FULL_API, it seems fair to keep all previously
exported symbols available with this option until this phase is over.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: Consolidate driver registering

This patch consolidates some bookkeeping for driver registering. It
closely models what pci_register_driver() does. The main addition is
that the owner of the driver is set, so we get a proper symlink
for /sys/bus/ieee1394/driver/*/module.

Also moves setting of name and bus type into nodemgr. Because of this,
we can remove the EXPORT_SYMBOL for ieee1394_bus_type, since it's now
only used in ieee1394.ko.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: convert from PCI DMA to generic DMA

API conversion without change in functionality

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: nodemgr: spaces to tabs

whitespace pedantry

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: nodemgr: fix deadlock in shutdown

If "modprobe ohci1394" was quickly followed by "modprobe -r ohci1394",
say with 1 second pause in between, the modprobe -r got stuck in
uninterruptible sleep in kthread_stop.  At the same time the knodemgrd
slept uninterruptibly in bus_rescan_devices_helper.  That's because
driver_detach took the semaphore of the PCI device and
bus_rescan_devices_helper wanted to take the semaphore of the FireWire
host device's parent, which is the same semaphore. This was a regression
since Linux 2.6.16, commit bf74ad5bc41727d5f2f1c6bedb2c1fac394de731,
"Hold the device's parent's lock during probe and remove".

The fix (or workaround) adds a dummy driver to the hpsb_host device. Now
bus_rescan_devices_helper won't scan the host device anymore.  This
doesn't hurt since we have no drivers which will bind to these devices
and it is unlikely that there will ever be such a driver.  The dummy
driver is befittingly presented as a representation of ieee1394 itself.

Fixes: http://bugzilla.kernel.org/show_bug.cgi?id=6706

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: nodemgr: remove duplicate assignment

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

sbp2: make 1bit bitfield unsigned

A signed single-bit bitfield doesn't make much sense. Make it unsigned.

Signed-off-by: Luca Tettamanti <kronos.it@gmail.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: schedule *_oui sysfs attributes for removal

There is no manpower available to reform oui.db into a library for use
in more kernel subsystems.  The low ratio of usefulness to size and the
occasional need to update oui.db from IEEE's official list suggest to
drop oui.db.  I plan to make a userspace script available which
translates the remaining numeric sysfs attributes to names of
organizations.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: schedule unused symbol exports for removal

This also means that former parts of ieee1394's API will be subject to
change or removal.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: dv1394: schedule for feature removal

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: raw1394: defer feature removal of old isoch interface

Known to be affected:
 - libdc1394: prefers video1394 for now, old-style raw1394 support might
   be dropped eventually
 - OpenH323 PWLib, AVC video input module: uses libraw1394's old API

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: ohci1394: call PMac code in shutdown only for proper machines

There has been an if(...) missing, for ages.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: ohci1394: reformat PPC_PMAC platform code

Adjust whitespace and line lengths

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: ohci1394: add PPC_PMAC platform code to driver probe

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=7431
iBook G3 threw a machine check exception and put the display backlight
to full brightness after ohci1394 was unloaded and reloaded.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: wrap two functions into one

Move the body of a function into its only caller.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: update comment on things to do

Some people actually look at those comments.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: use list_move_tail()

It's OK to reorder list_del() and sbp2util_free_command_dma() here.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: more concise names for types and variables

"struct scsi_id_instance_data" represents a logical unit.
Rename it to "struct sbp2_lu", and "scsi_id" to "lu".
Rename some other variables too.
Wrap almost all lines after at most 80 columns.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: remove unused struct members

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: proper unit in module parameter description

It's 2^20 bit/s, not 0.001 bit.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: clean up sbp2_ namespace

Prepend sbp2*_ to anything globally defined in sbp2.c except for some
macros.  Strip sbp2_ from names of struct members.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: some conditions in queue_command are unlikely

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

[IA64] replace kmalloc+memset with kzalloc

Replace kmalloc+memset with kzalloc

Signed-off-by: Yan Burman <burman.yan@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>

[IA64] resolve name clash by renaming is_available_memory()

There is a name clash with ia64 arch code in Andrew's tree. Rename
is_avialable_memory() to is_memory_available() to avoid the clash.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>

ieee1394: sbp2: remove superfluous comments

And update and reformat remaining comments.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: delayed_work -> work_struct

This work is not delayed.

Also bring the code format in a state which reduces my work to merge
pending sbp2 patchs.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

Merge branch 'master' of /home/trondmy/kernel/linux-2.6/ into merge_linus

[IA64] Need export for csum_ipv6_magic

Now we have our own highly optimized assembly code version of
this routine (Thanks Ken!) we should export it so that it can
be used.

Signed-off-by: Tony Luck <tony.luck@intel.com>

Merge branch 'master' of /home/trondmy/kernel/linux-2.6/ into merge_linus

Conflicts:

include/linux/sunrpc/xprt.h
net/sunrpc/xprtsock.c
Fix up conflicts with the workqueue changes.

ieee1394: sbp2: coding style of some macros

Adjust parentheses, indentation, line lengths.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: remove debug macros

No need to keep them in released sources.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: consolidate log levels

Replace some calls to SBP2_ERR and SBP2_WARN by SBP2_INFO.
Remove logging macros SBP2_NOTICE and SBP2_WARN.
Remove direct usage of HPSB_ logging macros.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: remove duplicate code

The same case is handled further below in sbp2scsi_complete_command.
Note, the second version behaves slightly different but looks
preferable.  It's an extremely unlikely case by the way.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: remove dead code

This has been within #if 0 for a long time and is wrong anyway.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: clean up function declarations

Remove unnecessary function prototypes.
Remove variable names from function prototypes.
Move declarations from sbp2.h to sbp2.c.
Move definitions of driver templates together near the top of sbp2.c.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

ieee1394: sbp2: remove irritating log message

The queue depth can be read from /sys/bus/scsi/devices/*/queue_depth,
so don't log it.  And the hint about speed improvements is misleading,
at least under current kernels.  If serialization is switched off, read
performance is typically increased by less than 10%.  (I did not test
write performance recently.)  On the other hand, serialize_io=0 is not
yet safe due to some implementation issues that are not trivial to fix.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>