Increments and decrements are usually grouped rather than mixed.  We can
optimize the inc and dec functions for that case.

Increment and decrement the counters by 50% more than the threshold in
those cases and set the differential accordingly.  This decreases the need
to update the atomic counters.

The idea came originally from Andrew Morton.  The overstepping alone was
sufficient to address the contention issue found when updating the global
and the per zone counters from 160 processors.

Also remove some code in dec_zone_page_state.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/mthca: Use IRQ safe locks to protect allocation bitmaps

It is supposed to be OK to call mthca_create_ah() and mthca_destroy_ah()
from any context.  However, for mem-full HCAs, these functions use the
mthca_alloc() and mthca_free() bitmap helpers, and those helpers use
non-IRQ-safe spin_lock() internally.  Lockdep correctly warns that
this could lead to a deadlock.  Fix this by changing mthca_alloc() and
mthca_free() to use spin_lock_irqsave().

Signed-off-by: Roland Dreier <rolandd@cisco.com>

  When I tested Linux kernel 2.6.17.7 about statistics
"ipFragFails",found that this counter couldn't increase correctly. The
criteria is RFC2011:
RFC2011
  ipFragFails OBJECT-TYPE
    SYNTAX      Counter32
    MAX-ACCESS  read-only
    STATUS      current
    DESCRIPTION
            "The number of IP datagrams that have been discarded because
            they needed to be fragmented at this entity but could not
            be, e.g., because their Don't Fragment flag was set."
    ::= { ip 18 }

When I send big IP packet to a router with DF bit set to 1 which need to
be fragmented, and router just sends an ICMP error message
ICMP_FRAG_NEEDED but no increments for this counter(in the function
ip_fragment).

Signed-off-by: Wei Dong <weid@nanjing-fnst.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch limits the warning messages when socket allocation failures
happen. It happens under memory pressure.

Signed-off-by: Akinobu Mita <mita@miraclelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Bug noticed by Remi Denis-Courmont <rdenis@simphalempin.com>.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

* 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6:
  [S390] cio: unsolicited interrupts during sense pgid.
  [S390] cio: no path after machine check.
  [S390] cio: kernel stack overflow.
  [S390] dasd: fix device shutdown process.
  [S390] broken copy_in_user function.

Commit 8c749327

 ("i386: Remove
alternative_smp") did not actually compile on x86 with CONFIG_SMP.

This fixes the __build_read/write_lock helpers.  I've boot tested on
SMP.

[ Andi: "Oops, I think that was a quilt unrefreshed patch.  Sorry.  I
  fixed those before testing, but then still send out the old patch." ]

Signed-off-by: Chris Wright <chrisw@sous-sol.org>
Cc: Gerd Hoffmann <kraxel@suse.de>
Acked-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  [POWERPC] Fix return value from memcpy
  [POWERPC] iseries: Define insw et al. so libata/ide will compile
  [POWERPC] Fix irq enable/disable in smp_generic_take_timebase
  [POWERPC] Fix problem with time not advancing on 32-bit platforms
  [POWERPC] Restore copyright notice in arch/powerpc/kernel/fpu.S
  [POWERPC] Fix up ibm_architecture_vec definition
  [POWERPC] Make OF irq map code detect more error cases
  [POWERPC] Support for "weird" MPICs and fixup mpc7448_hpc2
  [POWERPC] Fix MPIC sense codes in documentation
  [POWERPC] Fix performance regression in IRQ radix tree locking
  [POWERPC] Add mpc7448hpc2 device tree source file
  [POWERPC] Add MPC8349E MDS device tree source file to arch/powerpc/boot/dts
  [POWERPC] modify mpc83xx platforms to use new IRQ layer
  [POWERPC] Adapt ipic driver to new host_ops interface, add set_irq_type to set I...

This fixes a hang on ppc32.

The problem was that I was comparing a 32-bit quantity with a 64-bit
quantity, and consequently time wasn't advancing.  This makes us use a
64-bit quantity on all platforms, which ends up simplifying the code
since we can now get rid of the tb_last_stamp variable (which actually
fixes another bug that Ben H and I noticed while going carefully through
the code).

This works fine on my G4 tibook.  Let me know how it goes on your
machines.

Acked-by: Olaf Hering <olaf@aepfle.de>
Acked-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The backlight changes that went in had a bug where they could cause the
kernel to access an unitialized pointer when blanking if there is no
backlight control on a machine.

The bug affects atyfb, aty128fb, nvidiafb and rivafb.  radeonfb seems to
be ok.  This fixes it.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

As pointed out by Herbert Xu <herbert@gondor.apana.org.au>, our
memcpy implementation didn't return the destination pointer as its
return value, and there is code in the kernel that expects that.
This fixes it.

Signed-off-by: Paul Mackerras <paulus@samba.org>

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] Increase default nodes shift to 10, nr_cpus to 1024
  [IA64] remove redundant local_irq_save() calls from sn_sal.h
  [IA64] panic if topology_init kzalloc fails
  [IA64-SGI] Silent data corruption caused by XPC V2.

The radeon requires a VAP state flush when enabling/disabling
vertex programs on the r200 cards.

Signed-off-by: Dave Airlie <airlied@linux.ie>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The following change from -mm is important to 2.6.18 (actually to 2.6.17
but its too late for that). This was contributed over three months ago
by VIA to Bartlomiej and nothing happened. As a result the new chipset
is now out and Linux won't run on it. By the time 2.6.18 is finalised
this will be the defacto standard VIA chipset so support would be a good
plan.

Tested in -mm for a while, its essentially a PCI ident update but for
the bridge chip because VIA do things in weird ways.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>

It's possible to get an invalid page fault in kernel mode when we try to
write out segments from vsyscall32 when dumping core for a 32bit process if
the vsyscall32 DSO is not mapped in its address space (which can happen if,
for example, ulimit -v 100 is run).

Signed-off-by: Suleiman Souhlal <suleiman@google.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The values in init_tss.ist[] can change when an IST event occurs.  Save
the original IST values for checking stack addresses when debugging or
doing stack traces.

Signed-off-by: Keith Owens <kaos@ocs.com.au>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

After all their only point is having them in user space.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

There was a bogus hunk from the genirq merge that essentially
broke stack switching for hard interrupts. Remove it since it isn't
needed.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

After all their only point is having them in user space. On x86-64
they don't even work in kernel space.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

As a replacement for the earlier removal of the e820 MCFG check
we blacklist the Intel SDV with the original BIOS bug that
motivated that check. On those machines don't use MMCONFIG.

This also adds a new pci=mmconf parameter to override the blacklist.

Cc: Greg KH <gregkh@suse.de>
Cc: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The .fill causes miscompilations with some binutils version.

Instead just patch the lock prefix in the lock constructs. That is the
majority of the cost and should be good enough.

Cc: Gerd Hoffmann <kraxel@suse.de>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The .fill causes miscompilations with some binutils version.

Instead just patch the lock prefix in the lock constructs. That is the
majority of the cost and should be good enough.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Noticed by Jan Beulich.

When the kernel was moved from 1MB to 2MB in 2.6.17 the kernel reservation
code wasn't adjusted and it still reserved starting with 1MB. This means 1MB always
were lost.

This patch fixes this by reserving only starting with _text.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The unwinder fallback logic still had potential for falling through to
the legacy stack trace code without printing an indication (at once
serving as a separator) of this.

Further, the stack pointer retrieval for the fallback should be as
restrictive as possible (in order to avoid having the legacy stack
tracer try to access invalid memory). The patch tightens that, but
this could certainly be further improved.

Also making the call_trace command line option now conditional upon
CONFIG_STACK_UNWIND (as it's meaningless otherwise).

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

By hard-coding the cpuid keys for alternative_smp() rather than using
the symbolic constant it turned out that incorrect values were used on
both i386 (0x68 instead of 0x69) and x86-64 (0x66 instead of 0x68).

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

One open question: Should this added push perhaps be made conditional
upon CONFIG_STACK_UNWIND or CONFIG_UNWIND_INFO?
[AK: not needed, these are all very slow paths]

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

One open question: Should these added pushes perhaps be made
conditional upon CONFIG_STACK_UNWIND or CONFIG_UNWIND_INFO?
[AK: Not needed -- these are all very slow paths]

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The check for the MCFG table being reserved in the e820 map was originally
added to detect a broken BIOS in a preproduction Intel SDV. However it also
breaks the Apple x86 Macs, which can't supply this properly, but need
a working MCFG. With this patch they wouldn't use the MCFG and not work.

After some discussion I think it's best to remove the heuristic again.
It also failed on some other boxes (although it didn't cause much
problems there because old style port access for PCI config space
still works as fallback), but the preproduction SDVs can just use
pci=nommcfg. Supporting production machines properly is more
important.

Edgar Hucek did all the debugging work.

Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Edgar Hucek <hostmaster@ed-soft.at>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SUNLANCE]: Fix probing problem.
  [SPARC64]: Fix X server hangs due to large pages.

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [STRIP]: Fix neighbour table refcount leak.
  [IPV6]: ipv6_add_addr should install dstentry earlier
  [NETLINK]: Call panic if nl_table allocation fails
  [TCP]: Two RFC3465 Appropriate Byte Count fixes.
  [IPV6]: SNMPv2 "ipv6IfStatsInAddrErrors" counter error
  [E100]: Add module option to ignore bad EEPROM checksums.
  [SCTP]: Fix sctp_primitive_ABORT() call in sctp_close().

Calls to set a device online with path grouping may get stuck in
some cases because certain device conditions where discarded after
unsolicited interrupts.
Check subchannel activity after unsolicited interrupts and retry
the operation if the subchannel is idle.

Signed-off-by: Stefan Bader <shbader@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

Devices enter no-path state after disabling a channel path
via the SE even though another path has been reenabled at the SE.
The devices are set into no-path state before triggering path
verification even though other paths may have become available.
To fix this trigger path verification before setting a device into
no-path state.

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

Use different kind of assignment to make sure gcc doesn't create code
that creates temp variables on the stack, assigns values to it and
copies the content of the whole temp variable to the destination.
This reduces stack usage of e.g. ccwgroup_driver_register from 976
to 48 bytes instead.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

Fix clear_IO handling (need to wait for interrupt) and
introduced error-handling in shutdown processing.

Signed-off-by: Horst Hummel <horst.hummel@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

The copy_in_user primitive does not work as advertised. If the source
and target area are available copy_in_user copies one byte too much.
If one of the memory areas is not available it does not copy as much
data as it can, but up to 257 bytes less.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

These are build fixes that enable (for example) libata and the ide
code to actually build on iSeries.  The associated hardware will never
be supported on legacy iSeries, so the code paths don't actually need
to work, but it is useful (especially for a combined kernel) if the
code can build.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

Eran Ben-Avi <eranpublic@yahoo.com> pointed out that the arch/ppc version
of smp_generic_take_timebase disables interrupts on entry but exits without
restoring them.  However, both it and the arch/powerpc version have another
problem, which is that they use local_irq_disable/enable rather than
local_irq_save/restore, and they are called with interrupts disabled.

This fixes both problems; it changes a return to a break in the arch/ppc
version, and changes both versions to use local_irq_save/restore.

Signed-off-by: Paul Mackerras <paulus@samba.org>

This fixes a problem introduced in 5db9fa95

.
The last_jiffy per-cpu variable is only 32 bits on 32-bit machines, but it
was being compared with a 64-bit quantity (tb_next_jiffy), which resulted in
time not advancing.

This fixes it by changing last_jiffy to be 64 bits on all platforms.  With
this, we no longer need tb_last_stamp as a 32-bit version of tb_last_jiffy,
so this gets rid of tb_last_stamp and we just use tb_last_jiffy instead.
This also fixes a bug when the boot cpu is not online, because using
tb_last_stamp could have caused the wrong timebase origin value to be used
when calculating the time of day.

Signed-off-by: Paul Mackerras <paulus@samba.org>