This is not our bug! Sadly some devices cannot cope with the change
of TCP option ordering which was a result of the recent rewrite of
the option code (not that there was some particular reason steming
from the rewrite for the reordering) though any ordering of TCP
options is perfectly legal. Thus we restore the original ordering
to allow interoperability with/through such broken devices and add
some warning about this trap. Since the reordering just happened
without any particular reason, this change shouldn't cost us
anything.

There are already couple of known failure reports (within close
proximity of the last release), so the problem might be more
wide-spread than a single device. And other reports which may
be due to the same problem though the symptoms were less obvious.
Analysis of one of the case revealed (with very high probability)
that sack capability cannot be negotiated as the first option
(SYN never got a response).
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Reported-by: Aldo Maggi <sentiniate@tiscali.it>
Tested-by: Aldo Maggi <sentiniate@tiscali.it>
Signed-off-by: David S. Miller <davem@davemloft.net>

My change

    commit e2a6b852
    net: Enable TSO if supported by at least one device

didn't do what was intended because the netdev_compute_features
function was designed for conjunctions.  So what happened was that
it would simply take the TSO status of the last constituent device.

This patch extends it to support both conjunctions and disjunctions
under the new name of netdev_increment_features.

It also adds a new function netdev_fix_features which does the
sanity checking that usually occurs upon registration.  This ensures
that the computation doesn't result in an illegal combination
since this checking is absent when the change is initiated via
ethtool.

The two users of netdev_compute_features have been converted.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Once an endpoint has reached the SHUTDOWN-RECEIVED state,
it MUST NOT send a SHUTDOWN in response to a ULP request.
The Cumulative TSN Ack of the received SHUTDOWN chunk
MUST be processed.

This patch fix to process Cumulative TSN Ack of the received
SHUTDOWN chunk in SHUTDOWN_RECEIVED state.
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If SHUTDOWN is received in SHUTDOWN-PENDING state, enpoint should enter
the SHUTDOWN-RECEIVED state and check the Cumulative TSN Ack field of
the SHUTDOWN chunk (RFC 4960 Section 9.2). If the SHUTDOWN chunk can
acknowledge all of the send DATA chunks, SHUTDOWN-ACK should be sent.

But now endpoint just silently discarded the SHUTDOWN chunk.

SHUTDOWN received in SHUTDOWN-PENDING state can happend when the last
SACK is lost by network, or the SHUTDOWN chunk can acknowledge all of
the received DATA chunks. The packet sequence(SACK lost) is like this:

Endpoint A                       Endpoint B       ULP
(ESTABLISHED)                    (ESTABLISHED)

               <-----------      DATA
                                             <--- shutdown
                                 Enter SHUTDOWN-PENDING state
  SACK         ----lost---->

  SHUTDOWN(*1) ------------>

               <-----------      SHUTDOWN-ACK

 (*1) silently discarded now.

This patch fix to handle SHUTDOWN in SHUTDOWN-PENDING state as the same
as ESTABLISHED state.
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If SHUTDOWN chunk is received Cumulative TSN Ack beyond the max tsn currently
send, SHUTDOWN chunk be accepted and the association will be broken. New data
is send, but after received SACK it will be drop because TSN in SACK is less
than the Cumulative TSN, data will be retrans again and again even if correct
SACK is received.

The packet sequence is like this:

Endpoint A                       Endpoint B       ULP
(ESTABLISHED)                    (ESTABLISHED)

               <-----------      DATA (TSN=x-1)

               <-----------      DATA (TSN=x)

  SHUTDOWN     ----------->      (Now Cumulative TSN=x+1000)
  (TSN=x+1000)
               <-----------      DATA (TSN=x+1)

  SACK         ----------->      drop the SACK
  (TSN=x+1)
               <-----------      DATA (TSN=x+1)(retrans)

This patch fix this problem by terminating the association and respond to
the sender with an ABORT.
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If ICMP packet too big message is received with MTU larger than current
PMTU, SCTP will still accept this ICMP message and sync the PMTU of assoc
with the wrong MTU.

Endpoing A                 Endpoint B
(ESTABLISHED)              (ESTABLISHED)
ICMP         --------->
(packet too big, MTU too larger)
                           sync PMTU

This patch fixed the problem by drop that ICMP message.
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Badness at /home/proski/src/linux-2.6/net/mac80211/rx.c:2200
NIP: c02bc850 LR: c02ab268 CTR: 00000000
REGS: ef01fcc0 TRAP: 0700   Tainted: G        W  (2.6.27-wl)
MSR: 00029032 <EE,ME,IR,DR>  CR: 22004084  XER: 20000000
TASK = c1a58800[1778] 'p54pci' THREAD: ef01e000
[...]
NIP [c02bc850] __ieee80211_rx+0x17c/0x638
LR [c02ab268] ieee80211_tasklet_handler+0x104/0x120
Call Trace:
[ef01fd70] [c1a0c020] 0xc1a0c020 (unreliable)
[ef01fdb0] [c02ab268] ieee80211_tasklet_handler+0x104/0x120
[...]

the problem was that some older cards are mis-identified and didn't support
5GHz rates, while they have the right MAC & Synth chip.

This patch changes the way how p54 decides if it should enable 11a channels
or not.
Reported-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

orinoco_dl_firmware and symbol_dl_mage allocate large local
variables (1K); at least orinoco fails with panic or hung
kernel if 4K stacks is enabled.

Allocate large buffers dynamically at run time.

Tested-By: Andrey Borzenkov <arvidjaar@mail.ru> for Agere case

Signed-off-by: Andrey Borzenkov < arvidjaar@mail.ru>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Based on a patch by Elias Oltmanns, we call ath5k_init in resume even
if we didn't previously open the device.  Besides starting up the
device unnecessarily, this also causes an oops on rmmod because
mac80211 will not invoke ath5k_stop and softirqs are left running after
the module has been unloaded.  Add a new state bit, ATH_STAT_STARTED,
to indicate that we have been started up.
Reported-by: Toralf Förster <toralf.foerster@gmx.de>
Signed-off-by: Elias Oltmanns <eo@nebensachen.de>
Signed-off-by: Bob Copeland <me@bobcopeland.com>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Implement polling for 5200FEC to make netconsole work. Tested on
Phytec pcm030 and Efika.
Signed-off-by: Jon Smirl <jonsmirl@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

The chip can issue spurious interrupts for single interrupt
modes. We use disable to clear the condition and allow processing to
continue. Also got rid of legacy specific code since it now needs to
be done on MSI single irq also.
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

This patch extends the critical regions covered by lp->lock to make it
safer on SMP. The main failure point was the smc911x_hard_start_xmit
function being called from different CPUs. It was tested on the ARM SMP
platforms.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Platforms like ARM Ltd's RealView require the IRQ polarity bit to be set
for the SMC9118 chip. This patch allows the dynamic configuration via
the smc911x_platdata structure.

This patch also changes the smc91x_platdata structure name to the
correct smc911x_platdata in the smc911x_drv_probe() function.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Since more ARM platforms use this device, it is easier to add a
dependency on ARM rather than individual platforms.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=10994

Contributed by pablomme@googlemail.com, coenraad@wish.org.za
and a few others.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

The 8139 drivers are a source of error messages that confuse users.

Since this device can not be disambiguated by normal PCI device
id's two drivers match the same info. But the module utilities
seem to correctly handle this overlap, they try one driver, then
if that doesn't load try the other. Therefore there is no need for
a message to be logged with error level severity, just using info
level instead. Can't be completely silent because user might have
configure one driver and forgot the other one.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Add a bool IGB_DCA defined to y if IGB and DCA are enabled, but IGB isn't y while DCA=m.  And thus remove the need to select INTEL_IOATDMA when IGB is enabled, so that non-x86 architectures can build the igb driver.
Based on work/patch from Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

The 82575 has an issue in which the DMA will go out of sync if the link
partner goes into an L0s state.  To prevent this we set the pci-e link
partner capability bits to disable the L0s transition on the hw.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

This patch implements the memory notifier to update the busmap
instantly instead of rebuilding the whole map. This is necessary
because walk_memory_resource provides different information than
required during memory hotplug.
Signed-off-by: Hannes Hering <hering2@de.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

remove integer casting in the read/write IO accessors,
because Blackfin now provides those functions
Tested-by: Javier Herrero <jherrero@hvsistemas.es>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Commit 147e70e6 ("cxgb3: Use SKB list interfaces instead of home-grown
implementation.") causes a crash in t3_l2t_send_slow() when an iWARP
connection request is received.  This is because the new l2t_entry.arpq
skb queue is never initialized, and therefore trying to add an skb to
it causes a NULL dereference.  With the old code there was no need to
initialize the queues because the l2t_entry structures were zeroed,
and the code used NULL to mean empty.

Fix this by adding __skb_queue_head_init() when all the l2t_entry
structures get allocated.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Currently SMC911x driver is broken on ARM/PXA builds.
Unbreak such configurations.
Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Disable NAPI if a failure occurs when setting up the interface. Leaving
it enabled could cause a BUG the next time an ifconfig up is issued.
Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

RBTX4939 board has SMC91X chip and there can be other MIPS boards with
that chip.  Make SMC91X selectable on all MIPS board would be better than
enumerating each boards in Kconfig.
Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Cc: jeff@garzik.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

dev->irq is unsigned
Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

platform_get_irq*() returns on -ENXIO when the resource cannot be
found, but this remains unnoticed if stored in an unsigned.
Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

Checking the signature of the eeprom and the validity of the
MAC address should be enough to filter out the bad addresses
observed so far.

Contributed by Ivan Vecera and Martin Capitanio.

Tested on 8102el, 8168b and 8169 for a start.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Edward Hsu <edward_hsu@realtek.com.tw>

I prefer the debug information to be displayed until
the issue is properly handled.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Edward Hsu <edward_hsu@realtek.com.tw>

While looking for the recent "sack issue" I also read all eff_sacks
usage that was played around by some relevant commit. I found
out that there's another thing that is asking for a fix (unrelated
to the "sack issue" though).

This feature has probably very little significance in practice.
Opposite direction timeout with bidirectional tcp comes to me as
the most likely scenario though there might be other cases as
well related to non-data segments we send (e.g., response to the
opposite direction segment). Also some ACK losses or option space
wasted for other purposes is necessary to prevent the earlier
SACK feedback getting to the sender.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit f06febc9 ("timers: fix itimer/
many thread hang") introduced a new task_cputime interface and
subsequently only converted binfmt_elf over to it.  This results in the
build for binfmt_elf_fdpic blowing up given that p->signal->{u,s}time
have disappeared from underneath us.

Apply the same trivial fix from binfmt_elf to binfmt_elf_fdpic.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Removed duplicated #include <linux/vmalloc.h> in mm/vmalloc.c and
"internal.h" in mm/memory.c.
Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

We're trying to keep the !CONFIG_SHMEM tiny-shmem.c (using ramfs without
swap) in synch with CONFIG_SHMEM shmem.c (and mpm is preparing patches
to combine them).  I was glad to see EXPORT_SYMBOL_GPL(shmem_file_setup)
go into shmem.c, but why not support DRM-GEM when !CONFIG_SHMEM too?
But caution says still depend on MMU, since !CONFIG_MMU is.. different.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Acked-by: Matt Mackall <mpm@selenic.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

__FUNCTION__ is gcc-specific, use __func__
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

commit 9b7530cc ("i915: cleanup coding
horrors in i915_gem_gtt_pwrite()")

broke the i386 build for CONFIG_HIGHMEM=y.

Caught by automatic testing http://www.tglx.de/autoqa-logs/000137-0006-0001.logSigned-off-by: Thomas Gleixner <tglx@linutronix.de>
[ My bad. It's the same patch I sent out earlier, nobody noticed then either.. ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

powerpc doesn't use the generic WARN_ON infrastructure.  The newly
introduced WARN() as a result didn't print the message, this patch adds
the printk for this specific case.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

This fixes

  kernel/kexec.c: In function 'crash_save_vmcoreinfo_init':
  kernel/kexec.c:1374: error: 'vmlist' undeclared (first use in this function)
  kernel/kexec.c:1374: error: (Each undeclared identifier is reported only once
  kernel/kexec.c:1374: error: for each function it appears in.)
  kernel/kexec.c:1410: error: invalid use of undefined type 'struct vm_struct'
  make[1]: *** [kernel/kexec.o] Error 1
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

* git://git.kernel.org/pub/scm/linux/kernel/git/kyle/parisc-2.6:
  parisc: convert to generic compat_sys_ptrace
  parisc: add rtc platform driver
  parisc: initialize unwinder much earlier
  parisc: add new syscalls
  parisc: hijack jump to start_kernel
  parisc: add pdc_coproc_cfg_unlocked and set_firmware_width_unlocked
  parisc: move include/asm-parisc to arch/parisc/include/asm
  parisc: move pdc_result to real2.S
  parisc: unify CCIO_COLLECT_STATS implementation
  parisc: add arch/parisc/kernel/.gitignore
  parisc: ropes.h - fix <asm-parisc/*> -> <asm/*>
  parisc: parisc-agp - fix <asm-parisc/*> -> <asm/*>

Resolve remove/rename conflict: include/asm-parisc/a.out.h is no longer
relevant.

This gets rid of an annoying warning in ehci-hcd.c when DEBUG isn't
enabled:

    warning: label 'err_debug' defined but not used

by moving it inside the already-existing #ifdef DEBUG, so that it
matches the goto.  And now my regular build is warning-free again.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Yes, this will probably be switched over to a cleaner model anyway, but
in the meantime I don't want to see the 'unused variable' warnings that
come from the disgusting #ifdef code.  Make the special case be a nice
inlien function of its own, clean up the code, and make the warning go
away.

I wish people didn't write code that gets (valid) warnings from the
compiler, but I'll limit my fixes to code that I actually care about (in
this case just because I see the warning and it annoys me).
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>