I've been chasing a weird SELinux bug which shows up mostly when doing
installs of a dev-* rpm (ie. creating and overwriting lots of block
device inodes), but which I've also seen when doing mkinitrd.

It turned out not to be an SELinux problem at all, but a core VFS
S_ISBLK bug.  It seems that SELinux simply widens the race window.

The code at fault is fs/fs-writeback.c:__mark_inode_dirty():

		/*
		 * Only add valid (hashed) inodes to the superblock's
		 * dirty list.  Add blockdev inodes as well.
		 */
		if (!S_ISBLK(inode->i_mode)) {
			if (hlist_unhashed(&inode->i_hash))
				goto out;
			if (inode->i_state & (I_FREEING|I_CLEAR))
				goto out;
		}

The "I_FREEING|I_CLEAR" condition was added after the ISBLK/unhashed
tests were already in the source, but I can't see any reason why we'd
want the I_FREEING test not to apply to block devices.  And indeed, this
results in all sorts of inode list corruptions.  Simply moving the
I_FREEING|I_CLEAR test out of the protection of the S_ISBLK() condition
fixes things entirely.

The existing 2.6 kernel will reliably fail on me in about 2 seconds once
"rpm -Uvh --force dev*.rpm" starts its actual installation of the new
inodes.  With the patch below I can't reproduce it at all.

From: Roland McGrath <roland@redhat.com>

The #include is the part of this patch that matters, so the #ifdef below
works.

The rest of the patch removes gratuitous duplication due to some strange
aversion to concision in the presence of #ifdef, the kind that is all too
common, utterly pointless, and error prone.

From: Pat Gefre <pfg@sgi.com>

rch/ia64/sn/io/machvec/pci_bus_cvlink.c
    Some pretty-print mods
    Use pin instead of slot for dma_flush init

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/io/machvec/pci_bus_cvlink.c
    Couple of checks for kmalloc <= 0 were fixed
    Some __init and static fixes

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/kernel/irq.c
    Need to get the cpu from the passed in pcibr struct
    Made the interrupt list static and gave it a better name - credit jes
    Some lindent'isms
    Took out some code that isn't used ..... yet

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/io/sn2/pcibr/pcibr_intr.c
    Kill pcibr_intr_func()

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/kernel/setup.c

    If generic enabled legacy VGA or kbd - disable them
    Slightly different check for work arounds and only do it once
    If there is no klconfig info and we are in the simulator - ignore it
    Update the pxm_to_nasid() routine. It failed for SP configurations and some
    SMP configurations where M-bricks used pxm numbers lower that the first
    c-brick.
    If we don't find the cpu for pxm, search the memblks for it.

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/io/sn2/ml_iograph.c
    Better code for early_probe_for_widget()

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/io/sn2/pcibr/pcibr_ate.c
    check for pcibr_invalidate_ate 0 used for debugging

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/io/platform_init/sgi_io_init.c
   use numionodes instead of numnodes

arch/ia64/sn/io/sn2/klconflib.c
    find_lboard changes - generalized a number of the interface funcs

arch/ia64/sn/io/sn2/klgraph.c
    call the more general find_lboard funcs

arch/ia64/sn/io/sn2/ml_iograph.c
    call the more general lboard funcs

arch/ia64/sn/io/sn2/module.c
    lboard changes
    mod for headless/memless nodes

arch/ia64/sn/io/sn2/pcibr/pcibr_dvr.c
    isIO9 mod

arch/ia64/sn/kernel/setup.c
    headless/memless mod

include/asm-ia64/sn/klconfig.h
    generalized find_lboard funs

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/io/machvec/pci_dma.c
    Add MINIMAL_ATE_FLAGS() macro usage

include/asm-ia64/sn/pci/pcibr.h
    Add MINIMAL_ATE_FLAG() macro

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/io/sn2/pcibr/pcibr_dvr.c
    extern for pcibr_rrb_alloc_more()

include/asm-ia64/sn/pci/pcibr_private.h
    more unsigned to unsigned int

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/io/io.c@1.12
    misc code cleanup
    no parens on returns
    remove __psunsigned_t

arch/ia64/sn/io/sn2/ml_iograph.c@1.18
    no parens on returns

arch/ia64/sn/io/sn2/pcibr/pcibr_dvr.c@1.25
    no parens on returns
    use -errno

arch/ia64/sn/io/sn2/pcibr/pcibr_hints.c@1.8
    no parens on returns

arch/ia64/sn/io/sn2/pcibr/pcibr_intr.c@1.15
    no parens on returns
    unsigned to unsigned int

arch/ia64/sn/io/sn2/pcibr/pcibr_rrb.c@1.13
    no parens on returns
    use -errno

arch/ia64/sn/io/sn2/pcibr/pcibr_slot.c@1.18
    no parens on returns
    use -errno

arch/ia64/sn/io/sn2/pciio.c@1.14
    oom handling
    removed ASSERT and funcs not needed

arch/ia64/sn/io/sn2/pic.c@1.13
    use -errno

arch/ia64/sn/io/sn2/shuberror.c@1.12
    add delay.h

arch/ia64/sn/io/sn2/xbow.c@1.10
    include file clean up

arch/ia64/sn/io/sn2/xtalk.c@1.9
    misc code cleanup
    no parens on returns

arch/ia64/sn/io/xswitch.c@1.11
    misc code cleanup

arch/ia64/sn/kernel/bte.c@1.5
    include file clean up

arch/ia64/sn/kernel/mca.c@1.8
    include file clean up

arch/ia64/sn/kernel/probe.c@1.5
    include file clean up

arch/ia64/sn/kernel/sn2/prominfo_proc.c@1.3
    include file clean up

arch/ia64/sn/kernel/sn2/sn2_smp.c@1.9
    include file clean up

arch/ia64/sn/kernel/sn2/sn_proc_fs.c@1.4
    include file clean up

drivers/char/sn_serial.c@1.3
    include file clean up

include/asm-ia64/sn/addrs.h@1.8
    include file clean up
    remove __psunsigned_t

include/asm-ia64/sn/alenlist.h@1.8
    unsigned to unsigned int

include/asm-ia64/sn/arch.h@1.7
    include file clean up

include/asm-ia64/sn/bte.h@1.6
    include file clean up

include/asm-ia64/sn/clksupport.h@1.8
    include file clean up

include/asm-ia64/sn/driver.h@1.5
    remove __psunsigned_t

include/asm-ia64/sn/hcl.h@1.13
    include file clean up

include/asm-ia64/sn/hcl_util.h@1.7
    include file clean up

include/asm-ia64/sn/hwgfs.h@1.5
    include file clean up

include/asm-ia64/sn/iograph.h@1.9
    include file clean up

include/asm-ia64/sn/klconfig.h@1.14
    remove __psunsigned_t

include/asm-ia64/sn/kldir.h@1.5
    remove __psunsigned_t
    include file clean up

include/asm-ia64/sn/module.h@1.12
    include file clean up

include/asm-ia64/sn/nodepda.h@1.14
    include file clean up

include/asm-ia64/sn/pci/bridge.h@1.13
    uchar_t to unsigned char

include/asm-ia64/sn/pci/pcibr_private.h@1.20
    move PVnnnnnn's

include/asm-ia64/sn/pci/pciio.h@1.14
    unsigned to unsigned int

include/asm-ia64/sn/pci/pciio_private.h@1.10
    unsigned to unsigned int

include/asm-ia64/sn/pda.h@1.9
    include file clean up

include/asm-ia64/sn/pio.h@1.7
    include file clean up
    ulong to unsigned long

include/asm-ia64/sn/sgi.h@1.11
    include file clean up
    move PVnnnnnn's to here

include/asm-ia64/sn/sn2/arch.h@1.6
    include file clean up

include/asm-ia64/sn/sn2/sn_private.h@1.12
    include file clean up
    remove __psunsigned_t

include/asm-ia64/sn/sn_cpuid.h@1.8
    include file clean up

include/asm-ia64/sn/sn_private.h@1.6
    include file clean up

include/asm-ia64/sn/types.h@1.6
    include file clean up
    remove __psunsigned_t

include/asm-ia64/sn/vector.h@1.6
    include file clean up

include/asm-ia64/sn/xtalk/xbow.h@1.9
    include file clean up
    misc code cleanup

include/asm-ia64/sn/xtalk/xtalk.h@1.12
    unsigned to unsigned int

include/asm-ia64/sn/xtalk/xwidget.h@1.8
    unsigned to unsigned int

From: Pat Gefre <pfg@sgi.com>

arch/ia64/sn/io/machvec/pci_bus_cvlink.c
    Changes for new pcireg_ interfaces
    pcibr reorg
    Some code cleanup/reorg

arch/ia64/sn/io/machvec/pci_dma.c
    IS_PCIA64() not needed

arch/ia64/sn/io/sn2/ml_iograph.c
    new pcireg_ interface

arch/ia64/sn/io/sn2/pcibr/pcibr_ate.c
    code reorg/clean up

arch/ia64/sn/io/sn2/pcibr/pcibr_config.c
    code reorg/cleanup

arch/ia64/sn/io/sn2/pcibr/pcibr_dvr.c
    reorg/cleanup

arch/ia64/sn/io/sn2/pcibr/pcibr_error.c
    reorg/cleanup

arch/ia64/sn/io/sn2/pcibr/pcibr_intr.c
    reorg/cleanup

arch/ia64/sn/io/sn2/pcibr/pcibr_reg.c
    Fixed the interface to these functions - one call/data type

arch/ia64/sn/io/sn2/pcibr/pcibr_rrb.c
    reorg/cleanup

arch/ia64/sn/io/sn2/pcibr/pcibr_slot.c
    reorg/cleanup

arch/ia64/sn/io/sn2/pciio.c
    removed unused functions

arch/ia64/sn/io/sn2/pic.c
    reorg/cleanup

arch/ia64/sn/kernel/irq.c
    IS_PIC_SOFT not needed
    mod for new pcireg_ interfaces

include/asm-ia64/sn/module.h
    nodes/geoid[] -> MAX_SLABS

include/asm-ia64/sn/pci/bridge.h
    IS_[X]BRIDGE not needed

include/asm-ia64/sn/pci/pci_bus_cvlink.h
    SET_PCIA64 and IS_PCIA64 not needed
    isa64, dma_buf_sync, xbow_buf_sync gone

include/asm-ia64/sn/pci/pcibr.h
    mostly cleanup
    some reorg mods

include/asm-ia64/sn/pci/pcibr_private.h
    some reorg code
    protos for new pcireg_ interfaces

include/asm-ia64/sn/pci/pciio.h
    cleanup

include/asm-ia64/sn/pci/pic.h
    cleanup

include/asm-ia64/sn/sn2/intr.h
    changed IA64_SN2_FIRST_DEVICE_VECTOR and IA64_SN2_LAST_DEVICE_VECTOR

From: Andi Kleen <ak@muc.de>

Fix a bug introduced with the last merge that prevented booting with
CONFIG_DEBUG_INFO on on x86-64.  It would corrupt registers in interrupts.
This has hit a few people, so I would consider it as a critical fix.

From: IWAMOTO Toshihiro <iwamoto@valinux.co.jp>

This patch fixes a memory leak that happens when a core file hits the
process's resource limit.

From: Geert Uytterhoeven <geert@linux-m68k.org>

Spinlock code needs <linux/sched.h>

From: Geert Uytterhoeven <geert@linux-m68k.org>

Never include <asm/delay.h> directly

From: Geert Uytterhoeven <geert@linux-m68k.org>

Never include <asm/delay.h> directly

From: Geert Uytterhoeven <geert@linux-m68k.org>

Compile fix: add missing #include <linux/init.h>

From: Geert Uytterhoeven <geert@linux-m68k.org>

Compile fixes for the Moxa serial drivers:

- Add missing #include <linux/init.h>

- Kill warning if CONFIG_PCI is not set

From: Geert Uytterhoeven <geert@linux-m68k.org>

Fix compilation if CONFIG_PCI is not set

From: Junio C Hamano <junkio@cox.net>

Compiling neofb.c without MTRR results in two "unused variable" warnings.

We're using a 4-order allocation in there and it can fail.  Change it to just
keep reclaiming memory until something gives.

From: Martin Hicks <mort@wildopensource.com>

This patch removes kmalloc_percpu_init() from include/linux/percpu.h

It is unused and doesn't seem to be required.

From: Patrick Mansfield <patmans@us.ibm.com>

On NUMAQ, any module which calls __alloc_pages() needs cpu_2_node() (via
cpu_to_node()).

This patch exports cpu_2_node.

From: James Morris <jmorris@redhat.com>

It seems to me that this dentry assignment in open_namei is not needed per
the patch below.  On this path, dentry is not referenced.

From: Jens Axboe <axboe@suse.de>

2.6 imposes a 64 frame limit where 2.4 does not (just relies on kmalloc()
failing and limiting frames from that).  That breaks at least on guys app.
With MSF adressing, it's much simpler to be able to ask for a full second
at the time, so I think we should just allow that.  So bump the limit from
64 to CD_FRAMES (which is 75).

From: Jake Moilanen <moilanen@austin.ibm.com>

On a ppc64 logically partitioned system, there can be a setup where function
0 of a PCI-PCI bridge is assigned to one partition and (for example) function
2 is assigned to a second partition.  On the second partition, it would
appear that function 0 does not exist, but function 2 does.  If all the
functions are not scanned, everything under function 2 would not be detected.

This patch allows devices that don't respond to function 0, but do respond to
other functions to be marked with a quirk and have all of their functions
scanned.

From: Jack Steiner <steiner@sgi.com>

The cpu_vm_mask use to be close to the mmu_context_t field in the mm
struct.  Recently some large members were added between "cpu_vm_mask" and
"context".  I suspect that was an oversight.

Here is a patch that puts the fields close together.  This makes it likely
that both fields are in the same cache line.  Since both fields are likely
to be updated at the same time, this may improve performance.

From: Matt Mackall <mpm@selenic.com>

I've been fiddling with cleaning up some old code here and suggest the
following to make Lindent match actual practice more closely. This does:

a) (no -psl)

void *foo(void)
{

 instead of

void *
foo(void) {

b) (no -bs) "sizeof(foo)" rather than "sizeof (foo)"

c) (-ncs) "(void *)foo" rather than "(void *) foo"

From: Louis Zhuang <louis_zhuang@linux.co.intel.com>

enable  fast  symbol lookup via an inverted index.

From: Dave Olien <dmo@osdl.org>

acked by npiggin.

From: Nikita Danilov <Nikita@Namesys.COM>

shrink_list() checks PageSwapCache() before calling add_to_swap(), this
means that anonymous page that is going to be added to the swap right
now these checks return false and:

 (*) it will be unaccounted for in nr_mapped, and

 (*) it won't be written to the swap if gfp_flags include __GFP_IO but
     not __GFP_FS.

(Both will happen only on the next round of scanning.)

Patch below just moves may_enter_fs initialization down. I am not sure
about (*nr_mapped) increase though.

From: Andi Kleen <ak@suse.de>

When the user gave an address hint in mmap use it as starting point for the
search for !MAP_FIXED.

Currently it is only checked directly and when already used the free area
cache is used as starting point.  With this change you can use mmap(4096,
....) to e.g.  get the lowest free address in your address space, which is
sometimes useful.  For example on x86-64 glibc wants to preferably allocate
thread local data in the first 4GB but use higher addresses when this is
not possible.

This can be a bit more costly in CPU time because it may have to skip over
more VMAs, but gives better semantics for most cases.  Most programs pass
NULL as hint anyways so it won't make any difference for them.

I did it for the generic mmap and for x86-64 for now.  Also minor white
space fixes for x86-64.

From: Jes Sorensen <jes@trained-monkey.org>

nr_free_pages() is expensive, especially on large SMP machines.  The patch
changes the memory overcommit code so that it only calls nr_free_pages() is
we're about to fail the allocation attempt.

From: Jens Axboe <axboe@suse.de>

It's from Pascal Schmidt and adds write protect handling to ide-cd along
with support for non-2kb block sizes.

From: Maneesh Soni <maneesh@in.ibm.com>

The patch fixes locking in proc_check_root().  It brings is_subdir() call
under vfsmount_lock.  Holding vfsmount_lock will ensure mnt_mountpoint
dentry is intact and the dentry does not go away while it is being checked
in is_subdir().

From: Maneesh Soni <maneesh@in.ibm.com>

o The following patch fixes is_subdir() races with d_move. Due to concurrent
  d_move, in is_subdir() we can end up accessing freed d_parent pointer in
  case of pre-emptible kernel. To avoid this we can use rcu_read_lock() and
  rcu_read_unlock().

o This also fixes the seqlock uses in is_subdir() as we need to restart the
  the inner loop with the origianl new_dentry passed to the routine in case
  of any rename occured while we are traversing d_parent links.

From: Andrew Zabolotny <zap@homelink.ru>

ac97_unregister_driver() is nulling out the ->driver field for all codecs.
It should only null the codecs which are using this driver.