Here is a patch for ehca to use proper flag, ie. GFP_ATOMIC
resp. GFP_KERNEL, when calling get_zeroed_page() to prevent "Bug:
scheduling while atomic...". This error does not cause a kernel panic
but makes ipoib un-usable afterwards.  It is reproducible on
2.6.20-rc4 if one does ifconfig down during a flood ping test.  I have
not observed this error in earlier releases incl. 2.6.20-rc1.

This error occurs when a qp event/irq is received and ehca event
handler allocates a control block/page to obtain HCA error data block.
Use of GFP_ATOMIC when in interrupt context prevents this issue.

Signed-off-by Hoang-Nam Nguyen <hnguyen@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

According to the Tavor and Arbel programmer's reference manuals, the
number of bytes transferred is not provided in the byte_cnt field of
the CQ entry for atomic operation completions.  For atomic operations,
the number of bytes transferred is always 8 (when the status is
"success"), and this constant value should always be used by the
driver in the ib_wc entry returned, rather than using the CQE.
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

There's a problem with how rdma cm events are reported to userspace
that can lead to application crashes.

When a new connection request arrives, a context for the connection is
allocated in the kernel.  The connection event is then reported to
userspace.  The userspace library retrieves the event and allocates
its own context for the connection.  The userspace context is
associated with the kernel's context when accepting.  This allows the
kernel to give userspace context with other events.

A problem occurs if a second event for the same connection occurs
before the user has had a chance to call accept.  The userspace
context has not yet been set, which causes the librdmacm to crash.
(This has been seen when the app takes too long to call accept,
resulting in the remote side timing out and rejecting the connection)

Fix this by ignoring events for new connections until userspace has
set their context.  This can only happen if an error occurs on a new
connection before the user accepts it.  This is okay, since the accept
will just fail later.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

We discard new connection requests while the listen backlog is full,
but leak a struct ucma_event in the process.  Free the structure in
this case.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

The iWARP CM should report timeouts as event RDMA_CM_EVENT_UNREACHABLE,
not event RDMA_CM_EVENT_REJECTED.
Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

iSER limits the number of outstanding PDUs to send. When this threshold
is reached, it should return an error code (-ENOBUFS) instead of setting
the suspend_tx bit (which should be used only by libiscsi).
Signed-off-by: Erez Zilber <erezz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

mthca_table_find() will return the wrong address when the table entry
being searched for is exactly at the beginning of a sglist entry
(other than the first), because it uses >= when it should use >.

Example: assume we have 2 entries in scatterlist, 4K each, offset is
4K.  The current code will return first entry + 4K when we really want
the second entry.

In particular this means mapping an FMR on a memfree HCA may end up
writing the page table into the wrong place, leading to memory
corruption and also causing the HCA to use an incorrect address
translation table.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6:
  ACPI: asus_acpi: new MAINTAINER
  ACPI: fix section mis-match build warning
  ACPI: increase ACPI_MAX_REFERENCE_COUNT for larger systems
  ACPI: EC: move verbose printk to debug build only
  backlight: fix backlight_device_register compile failures

The help text for CONFIG_HID might imply for someone that
it's necessary to enable it for any keyboard or mouse
attached to the system. This is obviously not correct, so
fix it to avoid confusing the users.
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

HID it defaults to 'y'.  When you have input deselected, this
causes the kernel to fail to link.
Fix it by making it depend on INPUT.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

* master.kernel.org:/pub/scm/linux/kernel/git/davej/agpgart:
  [AGPGART] drivers/char/agp/sgi-agp.c: check kmalloc() return value
  [AGPGART] Fix PCI-posting flush typo.
  [AGPGART] fix detection of aperture size versus GTT size on G965
  [AGPGART] Remove unnecessary flushes when inserting and removing pages.
  [AGPGART] K8M890 support for amd-k8.

* master.kernel.org:/pub/scm/linux/kernel/git/davej/cpufreq:
  [CPUFREQ] longhaul: Kill off warnings introduced by recent changes.
  [CPUFREQ] Uninitialized use of cmd.val in arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.c:acpi_cpufreq_target()
  [CPUFREQ] Longhaul - Always guess FSB
  [CPUFREQ] Longhaul - Fix up powersaver assumptions.
  [CPUFREQ] longhaul: Fix up unreachable code.
  [CPUFREQ] speedstep-centrino: missing space and bracket
  [CPUFREQ] Bug fix for acpi-cpufreq and cpufreq_stats oops on frequency change notification
  [CPUFREQ] select consistently

Alan agreed to take over casual maintenance of the ide-cd atapi cdrom
driver, so I'm happy to sign it over to him. Alan, I hope the address is
the one you want to use. I also changed the list to linux-ide as that
seems more appropriate.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

It's a known fact that Windows times out commands after 7 seconds, so
drives generally try and respond if they can before that happens.  We
default to 5 seconds, which sometimes is a bit too short.

Jeremy Higdon reported here:

	http://lkml.org/lkml/2007/1/1/145

that his drive takes longer than 5 seconds for a "read track
information" command, later confirming that it is about 6.7 seconds.

So just do the sane thing and change the default command timeout to 7
seconds to avoid other surprises.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

If caller passed the tsk, we should use it to validate a stack ptr.
Otherwise, sysrq-t and other debugging stuff doesn't work.
Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Bunch of unused vars + one case where gcc isn't smart enough.
Signed-off-by: Dave Jones <davej@redhat.com>

[CPUFREQ] Uninitialized use of cmd.val in arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.c:acpi_cpufreq_target()

cmd.val was used uninitialized on the line below.
Signed-off-by: Guillaume Chazarain <guichaz@yahoo.fr>
Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>

This is patch that solves Ebox mini PC issue and make
FSB code more specification compilant. At start guess_fsb
function is guessing 200MHz FSB too. It is better to
make it in this way because, thanks to this function, driver
will fail for bogus FSB values caused by bogus multiplier
value. For PowerSaver processors we can't depend on Max /
MinMHzFSB because these values are only used for
PowerSaver 2.0 and 3.0. Most processors on which Longhaul
is used are PowerSaver 1.0 only. I'm changing code for older
CPU's too, but not so much as previously, and this code was
already used for Ezra. Using MinMHzBR for Ezra-T is outside
spec. It is for voltage scaling purpose and don't have to
be equal to minmult (but it is). Same for Nehemiah (it
isn't for sure). Added mult - current multiplier value.
Signed-off-by: Rafa³ Bilski <rafalbilski@interia.pl>
Signed-off-by: Dave Jones <davej@redhat.com>

drivers/char/agp/sgi-agp.c: check kmalloc() return value
Signed-off-by: Amit Choudhary <amit2030@gmail.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>

* 'for-linus' of master.kernel.org:/home/rmk/linux-2.6-arm:
  [ARM] 4081/1: Add definition for TI Sync Serial Protocol
  [ARM] 4080/1: Fix for the SSCR0_SlotsPerFrm macro
  [ARM] Fix VFP initialisation issue for SMP systems
  [ARM] 4078/1: Fix ARM copypage cache coherency problems
  [ARM] 4077/1: iop13xx: fix __io() macro
  [ARM] 4074/1: Flat loader stack alignment
  [ARM] 4073/1: Prevent s3c24xx drivers from including asm/arch/hardware.h and asm/arch/irqs.h
  [ARM] 4071/1: S3C24XX: Documentation update
  [ARM] 4066/1: correct a comment about PXA's sched_clock range
  [ARM] 4065/1: S3C24XX: dma printk fixes
  [ARM] 4064/1: make pxa_get_cycles() static
  [ARM] 4063/1: ep93xx: fix IRQ_EP93XX_GPIO?MUX numbering

Of the possible SSP frame formats (FRF bits in SSCR0), only SSCR0_PSP is defined. Other possible formats are Motorola SPI (0<<4), TI SSP (1<<4) and Microwire (2<<4). Attached patch adds a definition SSCR0_TISSP.
This mode is used for the sound codec attached to the PXA272 SSP1 of some HTC PDA phones.
Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

The SSCR0_SlotsPerFrm macro writes a 3-bit value to bits [2:0], while the correct location of FRDC in SSCR0 is at bits [26:24]. This patch adds the missing "<< 24".
Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

When we install the handlers for context switching, we must enable
VFP on all CPU cores, otherwise undefined (and random) effects
occur.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

Commit 968de4f0 ("i386: Relocatable
kernel support") caused problems for people with old binutils versions
that didn't mark ".text.*" sections automatically allocated.

So we should use .section command to specifically mark .text.head
section as AX (allocatable and executable) to solve the problem.

This should be unnecessary with binutils 2.15 and later, which is
already three years old, but it doesn't hurt supporting older toolchains
where possible.
Signed-off-by: Vivek Goyal <vgoyal@in.ibm.com>
Acked-by: Jean Delvare <khali@linux-fr.org>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

do not call a sleeping lock API in an RCU read section.
lock_sock_nested can sleep, its BH counterpart doesn't.
selinux_netlbl_inode_permission() needs to use the BH counterpart
unconditionally.

Compile tested.

From: Ingo Molnar <mingo@elte.hu>

added BH disabling, because this function can be called from non-atomic
contexts too, so a naked bh_lock_sock() would be deadlock-prone.

Boot-tested the resulting kernel.
Signed-off-by: Parag Warudkar <paragw@paragw.zapto.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Two issues:

- The final return 1 should be a return 0, otherwise comparing cfqq is
  a noop.

- bio_sync() only checks the sync flag, while rq_is_sync() checks both
  for READ and sync. The latter is what we want. Expand the bio check
  to include reads, and relax the restriction to allow merging of async
  io into sync requests.

In the future we want to clean up the SYNC logic, right now it means
both sync request (such as READ and O_DIRECT WRITE) and unplug-on-issue.
Leave that for later.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This is a slight variant on the patch I posted December 16th to fix
libata combined mode handling. The only real change is that we now
correctly also reserve BAR1,2,4. That is basically a neatness issue.

Jeff was unhappy about two things

1. That it didn't work in the case of one channel native one channel
legacy.

This is a silly complaint because the SFF layer in libata doesn't handle
this case yet anyway.

2. The case where combined mode is in use and IDE=n.

In this case the libata quirk code reserves the resources in question
correctly already.

Once the combined mode stuff is redone properly (2.6.21) then the entire
mess turns into a single pci_request_regions() for all cases and all the
ugly resource hackery goes away.

I'm sending this now rather than after running full test suites so that
it can get the maximal testing in a short time. I'll be running tests on
this after lunch.
Signed-off-by: Alan Cox <alan@redhat.com>
Cc: Jeff Garzik <jgarzik@pobox.com>
Acked-by: Alessandro Suardi <alessandro.suardi@gmail.com>
Acked-by: Theodore Tso <tytso@mit.edu>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Signed-off-by: Len Brown <len.brown@intel.com>

Dunno why this pops out in only in the allmodconfig build.
Though the warning is accurate, all the callers of the flagged
non __init function are __init, this is not a functional change.

WARNING: vmlinux - Section mismatch: reference to .init.data:acpi_sci_flags from .text between 'acpi_sci_ioapic_setup' (at offset 0xc010f0a
6) and 'acpi_gsi_to_irq'                                                                                                                   WARNING: vmlinux - Section mismatch: reference to .init.text:mp_override_legacy_irq from .text between 'acpi_sci_ioapic_setup' (at offset 0
xc010f0de) and 'acpi_gsi_to_irq'                                                                                                           WARNING: vmlinux - Section mismatch: reference to .init.data:acpi_sci_override_gsi from .text between 'acpi_sci_ioapic_setup' (at offset 0x
c010f0e4) and 'acpi_gsi_to_irq'
Signed-off-by: Len Brown <len.brown@intel.com>

We have some new larger ia64 systems in HP that trip over the
ACPI_MAX_REFERENCE_COUNT limit which triggers a large number of these
debug messages:

ACPI Warning (utdelete-0397): Large Reference Count (XXX) in object e0000a0ff6797ab0  [20060707]

This was increased once in the past as described in this very brief thread:

http://www.mail-archive.com/linux-acpi@vger.kernel.org/msg00890.htmlSigned-off-by: Doug Chapman <doug.chapman@hp.com>
Signed-off-by: Len Brown <len.brown@intel.com>

The recent EC cleanup left a printk enabled on handler evaluation
resulting in a bunch of messages on normal operation, like so:

ACPI: EC: evaluating _Q60
Signed-off-by: Len Brown <len.brown@intel.com>

Fix breakage from commit 519ab5f2 which
didn't update all references to backlight_device_register causing
compile failures.
Signed-off-by: Richard Purdie <rpurdie@rpsys.net>
Signed-off-by: Len Brown <len.brown@intel.com>

Recognize the Realtek ALC883 chip on MSI K9A Platinum motherboards
(model no. MS-7280), enabling full sound capabilities.
Signed-off-by: Leonard Norrgård <leonard.norrgard@refactor.fi>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This reverts commit a9622f62.  Now that
the Calgary code apparently detects itself properly, it's not needed any
more.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

...because it's always a good idea to cut a release *before* you go out
to party and get drunk.

Remember kids: "Don't Drink and Release!"

Commit b2b2cbc4 introduced a user-
visible change: ->pdeath_signal is sent only when the entire thread
group exits.

While this change is imho good, it may break things.  So restore the
old behaviour for now.
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
To: Albert Cahalan <acahalan@gmail.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Andrew Morton <akpm@osdl.org>
Cc: Linus Torvalds <torvalds@osdl.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Qi Yong <qiyong@fc-cn.com>
Cc: Roland McGrath <roland@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC64]: Handle ISA devices with no 'regs' property.
  [SPARC64]: Update defconfig.
  [SPARC64]: Fix of_iounmap() region release.
  [SPARC64]: Fix "mem=xxx" handling.

Installing an IPsec SA using old algorithm names (.compat) does not work
if the algorithm is not already loaded. When not using the PF_KEY
interface, algorithms are not preloaded in xfrm_probe_algs() and
installing a IPsec SA fails.
Signed-off-by: Martin Willi <martin@strongswan.org>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Don't add it there please; add it lower down inside the existing #ifdef
__KERNEL__. You just made the _userspace_ net.h include random.h, which
then fails to compile unless <asm/types.h> was already included.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

When a driver writer calls this, they generally expect that
all previous stores and modifications they've made will be
visible before netif_poll_enable() executes, so ensure this.

Noticed by Ben H.
Signed-off-by: David S. Miller <davem@davemloft.net>