Revert commit f1a9ee75:

  Author: Rik van Riel <riel@redhat.com>
  Date:   Thu Feb 7 00:14:08 2008 -0800

    kswapd should only wait on IO if there is IO

    The current kswapd (and try_to_free_pages) code has an oddity where the
    code will wait on IO, even if there is no IO in flight.  This problem is
    notable especially when the system scans through many unfreeable pages,
    causing unnecessary stalls in the VM.

    Additionally, tasks without __GFP_FS or __GFP_IO in the direct reclaim path
    will sleep if a significant number of pages are encountered that should be
    written out.  This gives kswapd a chance to write out those pages, while
    the direct reclaim task sleeps.
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Because of large latencies and interactivity problems reported by Carlos,
here: http://lkml.org/lkml/2008/3/22/211

Cc: Rik van Riel <riel@redhat.com>
Cc: "Carlos R.  Mafra" <crmafra2@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Update documentation for the hw_random support to be current:

 - Documentation/hw_random.txt has been updated to reflect the
   current code:  it's a framework now, a "core" with a small
   sysfs interface, that hardware-specific drivers plug in to.
   Text specific to Intel hardware is now at the end.

 - Kconfig now references the Documentation/hw_random.txt file
   and better explains what this really does.

Both chunks of documentation now higlight the fact that the kernel entropy
pool is maintained by "rngd", and this driver has nothing directly to do with
that important task.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Older smackfs was parsing MAC rules by characters, thus a need of locking
write sessions on open() was needed.  This lock is no longer useful now since
each rule is handled by a single write() call.

This is also a bugfix since seq_open() was not called if an open() O_RDWR flag
was given, leading to a seq_read() without an initialized seq_file, thus an
Oops.
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Reported-by: Jonathan Corbet <corbet@lwn.net>
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

As Paul pointed out, the ACCESS_ONCE are not needed because we already have
the explicit surrounding memory barriers.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Mason <mmlnx@us.ibm.com>
Cc: Dipankar Sarma <dipankar@in.ibm.com>
Cc: David Smith <dsmith@redhat.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Adrian Bunk <adrian.bunk@movial.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Add comments requested by Andrew.

Updated comments about synchronize_sched().  Since we use call_rcu and
rcu_barrier now, these comments were out of sync with the code.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Mason <mmlnx@us.ibm.com>
Cc: Dipankar Sarma <dipankar@in.ibm.com>
Cc: David Smith <dsmith@redhat.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Adrian Bunk <adrian.bunk@movial.fi>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

With numa enabled, some callers could have a range of memory on one node
but try to free that on other node.  This can cause some pages to be
freed wrongly.

For example: when we try to allocate 128g boot ram early for
gart/swiotlb, and free that range later so gart/swiotlb can get some
range afterwards.

With this patch, we don't need to care which node holds the range, just
loop to call free_bootmem_node for all online nodes.

This patch makes free_bootmem_core() more robust by trimming the sidx
and eidx according the ram range that the node has.

And make the free_bootmem_core handle this out of range case.  We could
use bdata_list to make sure the range can be freed for sure.  So next
time, we don't need to loop online nodes and could use free_bootmem
directly.
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Tested-by: Ingo Molnar <mingo@elte.hu>
Cc: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The block2mtd driver (drivers/mtd/devices/block2mtd.c) will kfree an on-stack
pointer when handling an invalid argument line (e.g.
block2mtd=/dev/loop0,xxx).

The kfree was added some time ago when "name" was dynamically allocated.
Signed-off-by: Ingo van Lil <inguin@gmx.de>
Acked-by: Joern Engel <joern@lazybastard.org>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: <stable@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Provide example for memmap exclude option (it is slightly strange and
non-trivial) and provide nice small HOWTO for people with bad memory.
Signed-off-by: Jan-Simon Moeller <dl9pf@gmx.de>
Signed-off-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  [POWERPC] Fix Oops with TQM5200 on TQM5200
  [POWERPC] mpc5200: Fix null dereference if bestcomm fails to initialize
  [POWERPC] mpc5200-fec: Fix possible NULL dereference in mdio driver
  [POWERPC] Fix crash in init_ipic_sysfs on efika
  [POWERPC] Don't use 64k pages for ioremap on pSeries

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC64]: exec PT_DTRACE
  [SPARC64]: Use shorter list_splice_init() for brevity.
  [SPARC64]: Remove most limitations to kernel image size.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
  sch_htb: fix "too many events" situation
  connector: convert to single-threaded workqueue
  [ATM]: When proc_create() fails, do some error handling work and return -ENOMEM.
  [SUNGEM]: Fix NAPI assertion failure.
  BNX2X: prevent ethtool from setting port type
  [9P] net/9p/trans_fd.c: remove unused variable
  [IPV6] net/ipv6/ndisc.c: remove unused variable
  [IPV4] fib_trie: fix warning from rcu_assign_poinger
  [TCP]: Let skbs grow over a page on fast peers
  [DLCI]: Fix tiny race between module unload and sock_ioctl.
  [SCTP]: Fix build warnings with IPV6 disabled.
  [IPV4]: Fix null dereference in ip_defrag

The iWARP protocol limits RDMA read requests to a single scatter
entry.  NFS/RDMA has code in rdma_read_max_sge() that is supposed to
limit the sge_count for RDMA read requests to 1, but the code to do
that is inside an #ifdef RDMA_TRANSPORT_IWARP block.  In the mainline
kernel at least, RDMA_TRANSPORT_IWARP is an enum and not a
preprocessor #define, so the #ifdef'ed code is never compiled.

In my test of a kernel build with -j8 on an NFS/RDMA mount, this
problem eventually leads to trouble starting with:

    svcrdma: Error posting send = -22
    svcrdma : RDMA_READ error = -22

and things go downhill from there.

The trivial fix is to delete the #ifdef guard.  The check seems to be
a remnant of when the NFS/RDMA code was not merged and needed to
compile against multiple kernel versions, although I don't think it
ever worked as intended.  In any case now that the code is upstream
there's no need to test whether the RDMA_TRANSPORT_IWARP constant is
defined or not.

Without this patch, my kernel build on an NFS/RDMA mount using NetEffect
adapters quickly and 100% reproducibly failed with an error like:

    ld: final link failed: Software caused connection abort

With the patch applied I was able to complete a kernel build on the
same setup.

(Tom Tucker says this is "actually an _ancient_ remnant when it had to
compile against iWARP vs. non-iWARP enabled OFA trees.")
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Acked-by: Tom Tucker <tom@opengridcomputing.com>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

It appears that 64-bit PCI resources cannot possibly ever have worked on
x86-32 even when the RESOURCES_64BIT config option was set, because any
driver that tried to [pci_]ioremap() the resource would have been unable
to do so because the high 32 bits would have been silently dropped on
the floor by the ioremap() routines that only used "unsigned long".

Change them to use "resource_size_t" instead, which properly encodes the
whole 64-bit resource data if RESOURCES_64BIT is enabled.
Acked-by: H. Peter Anvin <hpa@kernel.org>
Acked-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The printk() can deadlock because it can wake up klogd(), and
task enqueueing will try to read the time in order to set a hrtimer.
Reported-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Debugged-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The "bestcomm-core" driver defines its of_match table as follows

static struct of_device_id mpc52xx_bcom_of_match[] = {
	{ .type = "dma-controller", .compatible = "fsl,mpc5200-bestcomm", },
	{ .type = "dma-controller", .compatible = "mpc5200-bestcomm", },
	{},
};

so while registering the driver, the driver's probe function won't be
called, because the device tree node doesn't have a device_type
property.  Thus the driver's bcom_engine structure won't be allocated.
Referencing this structure later causes observed Oops.

Checking bcom_eng pointer for NULL before referencing data pointed
by it prevents oopsing, but fec driver still doesn't work (because
of the lost bestcomm match and resulted task allocation failure).
Actually the compatible property exists and should match and so
the fec driver should work.

This removes .type = "dma-controller" from the bestcomm driver's
mpc52xx_bcom_of_match table to solve the problem.
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>

If the bestcomm initialization fails, calls to the task allocate
function should fail gracefully instead of oopsing with a NULL deref.
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>

If the reg property is missing from the phy node (unlikely, but possible),
then the kernel will oops with a NULL pointer dereference.  This fixes
it by checking the pointer first.
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>

The global primary_ipic in arch/powerpc/sysdev/ipic.c can remain NULL
if ipic_init() fails, which will happen on machines that don't have an
ipic interrupt controller.  init_ipic_sysfs() will crash in that case.
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>

On pSeries, the hypervisor doesn't let us map in the eHEA ethernet
adapter using 64k pages, and thus the ehea driver will fail if 64k
pages are configured.  This works around the problem by always
using 4k pages for ioremap on pSeries (but not on other platforms).
A better fix would be to check whether the partition could ever
have an eHEA adapter, and only force 4k pages if it could, but this
will do for 2.6.25.

This is based on an earlier patch by Tony Breeds.
Signed-off-by: Paul Mackerras <paulus@samba.org>

The PT_DTRACE flag is meaningless and obsolete.
Don't touch it.
Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>

HTB is event driven algorithm and part of its work is to apply
scheduled events at proper times. It tried to defend itself from
livelock by processing only limited number of events per dequeue.
Because of faster computers some users already hit this hardcoded
limit.

This patch limits processing up to 2 jiffies (why not 1 jiffie ?
because it might stop prematurely when only fraction of jiffie
remains).
Signed-off-by: Martin Devera <devik@cdi.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>

From: Evgeniy Polyakov <johnpol@2ka.mipt.ru>

We don't need one cqueue thread for each CPU.  cqueue is used for
receiving userspace datagrams, which are very rare and thus will
happily live with a single queue.
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Wang Chen <wangchen@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

As reported by Johannes Berg:

I started getting this warning with recent kernels:

[  773.908927] ------------[ cut here ]------------
[  773.908954] Badness at net/core/dev.c:2204
 ...

If we loop more than once in gem_poll(), we'll
use more than the real budget in our gem_rx()
calls, thus eventually trigger the caller's
assertions in net_rx_action().

Subtract "work_done" from "budget" for the second
arg to gem_rx() to fix the bug.
Signed-off-by: David S. Miller <davem@davemloft.net>

On 10GBaseT boards setting the type to TP will cause the driver to try
to configure 1GBaseT.
Since there are currently no boards that support setting of the port
type, disable this for now.
Signed-off-by: Eliezer Tamir <eliezert@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The variable cb is initialized but never used otherwise.

The semantic patch that makes this change is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@@
type T;
identifier i;
constant C;
@@

(
extern T i;
|
- T i;
  <+... when != i
- i = C;
  ...+>
)
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: David S. Miller <davem@davemloft.net>

The variable hlen is initialized but never used otherwise.

The semantic patch that makes this change is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@@
type T;
identifier i;
constant C;
@@

(
extern T i;
|
- T i;
  <+... when != i
- i = C;
  ...+>
)
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: David S. Miller <davem@davemloft.net>

This gets rid of a warning caused by the test in rcu_assign_pointer.
I tried to fix rcu_assign_pointer, but that devolved into a long set
of discussions about doing it right that came to no real solution.
Since the test in rcu_assign_pointer for constant NULL would never
succeed in fib_trie, just open code instead.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  RDMA/nes: Fix MSS calculation on RDMA path

* git://git.kernel.org/pub/scm/linux/kernel/git/bart/ide-2.6:
  Revert "ide-tape: schedule driver for removal after 6 months"
  ide: mark "hdx=remap" and "hdx=remap63" kernel parameters as obsoleted
  ide: mark "hdx=[driver_name]" and "hdx=scsi" kernel parameters as obsoleted
  ide: Documentation/ide/ide.txt fixes
  ide: mark special "ide0=" kernel parameters as obsoleted
  ide: remove commented out entries from ide_pio_blacklist[]

* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
  [CIFS] Fix mem leak on dfs referral
  [CIFS] file create with acl support enabled is slow
  [CIFS] Fix mtime on cp -p when file data cached but written out too late
  [CIFS] Fix build problem
  [CIFS] cifs: replace remaining __FUNCTION__ occurrences
  [CIFS]  DFS patch that connects inode with dfs handling ops

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-x86:
  x86: revert: reserve dma32 early for gart

Change pagemap output format to allow for future reporting of huge pages.

(Format comment and minor cleanups: mpm@selenic.com)
Signed-off-by: Hans Rosenfeld <hans.rosenfeld@amd.com>
Signed-off-by: Matt Mackall <mpm@selenic.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Suppressing uevents turned out to be a bad idea as it screws up the
order of events, making user space very confused. Change the system to
use sysfs groups instead.

This is a regression that, for some odd reason, has gone unnoticed for
some time. It confuses hal so that the block devices (which have the
mmc device as a parent) are not registered. End result being that
desktop magic when cards are inserted won't work.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Increase the number of PnP memory resources from 12 to 24.

This removes an "exceeded the max num of mem resources" warning on boot. I
also noticed the reservation of two more iomem ranges on the computer on
which this was tested.
Signed-off-by: Darren Salt <linux@youmustbejoking.demon.co.uk>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

PNP_MAX_MEM and PNP_MAX_PORT are mainly used to size tables of PNP
device resources.  In 2.6.24, we increased their values to accomodate
ACPI devices that have many resources:

                 2.6.23    2.6.24
                 ------    ------
  PNP_MAX_MEM       4         12
  PNP_MAX_PORT      8         40

However, ISAPNP also used these constants as the size of parts of the
logical device register set.  This register set is fixed by hardware,
so increasing the constants meant that we were reading and writing
unintended parts of the register set.

This patch changes ISAPNP to use the correct register set sizes (the
same values we used prior to 2.6.24).
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: Igor Mammedov <niallain@gmail.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>

While testing the virtio-net driver on KVM with TSO I noticed
that TSO performance with a 1500 MTU is significantly worse
compared to the performance of non-TSO with a 16436 MTU.  The
packet dump shows that most of the packets sent are smaller
than a page.

Looking at the code this actually is quite obvious as it always
stop extending the packet if it's the first packet yet to be
sent and if it's larger than the MSS.  Since each extension is
bound by the page size, this means that (given a 1500 MTU) we're
very unlikely to construct packets greater than a page, provided
that the receiver and the path is fast enough so that packets can
always be sent immediately.

The fix is also quite obvious.  The push calls inside the loop
is just an optimisation so that we don't end up doing all the
sending at the end of the loop.  Therefore there is no specific
reason why it has to do so at MSS boundaries.  For TSO, the
most natural extension of this optimisation is to do the pushing
once the skb exceeds the TSO size goal.

This is what the patch does and testing with KVM shows that the
TSO performance with a 1500 MTU easily surpasses that of a 16436
MTU and indeed the packet sizes sent are generally larger than
16436.

I don't see any obvious downsides for slower peers or connections,
but it would be prudent to test this extensively to ensure that
those cases don't regress.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Revert

commit f62f1fc9
Author: Yinghai Lu <yhlu.kernel@gmail.com>
Date:   Fri Mar 7 15:02:50 2008 -0800

    x86: reserve dma32 early for gart

The patch has a dependency on bootmem modifications which are not .25
material that late in the -rc cycle. The problem which is addressed by
the patch is limited to machines with 256G and more memory booted with
NUMA disabled. This is not a .25 regression and the audience which is
affected by this problem is very limited, so it's safer to do the
revert than pulling in intrusive bootmem changes right now.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>