Change ehca's Kconfig to activates scaling code as default.  After
several measurements we saw that this feature prevents dropped packets
(UD) in stress situation. Thus, enabling it helps to improve ehca's
bandwidth through IPoIB.
Signed-off-by: NHoang-Nam Nguyen <hnguyen@de.ibm.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Define and use a constant EHCA_MAX_MTU instead hardcoded value.
Signed-off-by: NHoang-Nam Nguyen <hnguyen@de.ibm.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Assure 4K alignment for firmware control blocks in 64K page mode,
because kzalloc()'s result address might not be 4K aligned if 64K
pages are enabled. Thus, we introduce wrappers called
ehca_{alloc,free}_fw_ctrlblock(), which use a slab cache for objects
with 4K length and 4K alignment in order to alloc/free firmware
control blocks in 64K page mode. In 4K page mode those wrappers just
are defines of get_zeroed_page() and free_page().
Signed-off-by: NHoang-Nam Nguyen <hnguyen@de.ibm.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
  RDMA/addr: Use client registration to fix module unload race
  IB/mthca: Fix MAD extended header format for MAD_IFC firmware command
  IB/uverbs: Return sq_draining value in query_qp response
  IB/amso1100: Fix incorrect pr_debug()
  IB/amso1100: Use dma_alloc_coherent() instead of kmalloc/dma_map_single
  IB/ehca: Fix eHCA driver compilation for uniprocessor
  RDMA/cma: rdma_bind_addr() leaks a cma_dev reference count
  IB/iser: Start connection after enabling iSER

Require registration with ib_addr module to prevent caller from
unloading while a callback is in progress.
Signed-off-by: NSean Hefty <sean.hefty@intel.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] Do not use -msym32 option for modules.
  [MIPS] Don't use R10000 llsc workaround version for all llsc-full processors.
  [MIPS] Ocelot G: Fix : "CURRENTLY_UNUSED" is not defined warning.
  [MIPS] Fix warning about init_initrd() call if !CONFIG_BLK_DEV_INITRD.
  [MIPS] IP27: Allow SMP ;-)  Another changeset messed up by patch.
  [MIPS] Fix merge screwup by patch(1)
  Revert "[MIPS] Make SPARSEMEM selectable on QEMU."

On 64-bit kernel, modules are loaded into XKSEG for now.  While XKSEG
address is not a sign-extended 32-bit address, we can not use -msym32
option.
Signed-off-by: NAtsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Found and original patch by bile@landofbile.com.
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

  CC      arch/mips/momentum/ocelot_g/gt-irq.o
arch/mips/momentum/ocelot_g/gt-irq.c:30:5: warning: "CURRENTLY_UNUSED" is not defined
arch/mips/momentum/ocelot_g/gt-irq.c:199:5: warning: "CURRENTLY_UNUSED" is not defined
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

When lmo commit 4ef893e0515e8bf336dfbd200884f244869fbb43 was merged to
kernel.org as e73ea273 patch happily
applied the IP27 segment to IP22.  f63f36c18b11e166d0f362ac04dbcd7e6ea23f9e
did fix the effects partially - and with a wrong log message.  Now fixed
for real (tm).
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Patch happily applied an Ocelot G patch to Ocelot C when merging
linux-mips.org changeset 91ee9a801e65d2981dfe327d2519c7fc6ab02e6b into
kernel.org as 6ceb6d3a.
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

This reverts commit 31473747.

Another amazing example of patch(1) messing up - lmo changeset
66e8560d11d02bcadc261498471831a6375ad046 was merged twice to kernel.org
and ended up doing this rubbish job.
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/shaggy/jfs-2.6:
  jfs: Add splice support

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [NETFILTER]: silence a warning in ebtables
  [IPV6]: File the fingerprints off ah6->spi/esp6->spi
  [TCP]: Set default congestion control when no sysctl.
  [TIPC] net/tipc/port.c: fix NULL dereference

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC64]: Fix futex_atomic_cmpxchg_inatomic implementation.

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  Revert "[PATCH] Add 0x7110 piix to ata_piix.c"
  [libata] sata_nv: Add PCI IDs
  [PATCH] ahci: fix status register check in ahci_softreset

This reverts commit f833229c:

According to reviewers and the lspci data provided in commit message
itself, PCI ID 0x7110 should not have been added to ata_piix.
Signed-off-by: NJeff Garzik <jeff@garzik.org>

net/bridge/netfilter/ebtables.c: In function 'ebt_dev_check':
net/bridge/netfilter/ebtables.c:89: warning: initialization discards qualifiers from pointer target type

So make the char* a const char * and the warning is gone.
Signed-off-by: NMeelis Roos <mroos@linux.ee>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

I copied the logic from ll/sc arch implementations, but that
was wrong and makes no sense at all.  Just do a straight
compare-exchange instruction, just like x86.

Based upon bug reports from Dennis Gilmore and Fabio Massimo.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In theory these are opaque 32bit values.  However, we end up
allocating them sequentially in host-endian and stick unchanged
on the wire.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The setting of the default congestion control was buried in
the sysctl code so it would not be done properly if SYSCTL was
not enabled.
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The correct order is: NULL check before dereference

Spotted by the Coverity checker.
Signed-off-by: NAdrian Bunk <bunk@stusta.de>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It turns out that the linker warnings on 64-bit powerpc about "section
blah exceeds stub group size" were being triggered by conditional
branches in head_64.S branching to global symbols, whether in
head_64.S or in other files.  This eliminates the warnings by making
some global symbols in head_64.S no longer global, and by rearranging
some branches.
Signed-off-by: NPaul Mackerras <paulus@samba.org>
[ Yee-haa. Maybe I'll notice newly introduced real warnings now - Linus ]
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] Fix warning in mips-boards generic PCI
  [MIPS] SMTC: Synchronize cp0 counters on bootup.
  [MIPS] SMTC: Fix crash if # of TC's > # of VPE's after pt_regs irq cleanup.
  [MIPS] 16K & 64K page size fixes

Since the "mask" bit is in the low word, when we write a new entry, we
need to write the high word first, before we potentially unmask it.

The exception is when we actually want to mask the interrupt, in which
case we want to write the low word first to make sure that the high word
doesn't change while the interrupt routing is still active.
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

arch/mips/mips-boards/generic/pci.c: In function `mips_pcibios_init':
arch/mips/mips-boards/generic/pci.c:227: warning: comparison of distinct pointer types lacks a cast
arch/mips/mips-boards/generic/pci.c:228: warning: comparison of distinct pointer types lacks a cast
Signed-off-by: NYoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Derived from Peter Watkins <treestem@gmail.com>'s work.
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

This is preparation for fixing the ordering of the accesses that
got broken by the commit cf4c6a2f when
factoring out the "common" io apic routing entry accesses.

Move the accessor function (that were only used by io_apic.c) out
of a header file, and use proper memory-mapped accesses rather than
making up our own "volatile" pointers.
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  [POWERPC] Make alignment exception always check exception table
  [POWERPC] Disallow kprobes on emulate_step and branch_taken
  [POWERPC] Make mmiowb's io_sync preempt safe
  [POWERPC] Make high hugepage areas preempt safe
  [POWERPC] Make current preempt-safe
  [POWERPC] qe_lib: qe_issue_cmd writes wrong value to CECDR
  [POWERPC] Use 4kB iommu pages even on 64kB-page systems
  [POWERPC] Fix oprofile support for e500 in arch/powerpc
  [POWERPC] Fix rmb() for e500-based machines it
  [POWERPC] Fix various offb issues

Signed-off-by: NJeff Garzik <jeff@garzik.org>

ahci_softreset() used to use ahci_tf_read() which reads D2H_REG area
to check for the Status register.  However, this area is zeroed on
initialization and not set by initial signature FIS.  Replace it with
ahci_check_status().

This bug prevented CLO code from being activated whenever BSY and/or
DRQ is set prior to softreset.  This fix makes
AHCI_FLAG_RESET_NEEDS_CLO flag redundant.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  [PATCH] ata_piix: allow 01b MAP for both ICH6M and ICH7M
  [PATCH] libata: unexport ata_dev_revalidate()
  [PATCH] Add 0x7110 piix to ata_piix.c
  [PATCH] sata_sis: fix flags handling for the secondary port

The alignment exception used to only check the exception table for
-EFAULT, not for other errors. That opens an oops window if we can
coerce the kernel into getting an alignment exception for other reasons
in what would normally be a user-protected accessor, which can be done
via some of the futex ops. This fixes it by always checking the
exception tables.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

On powerpc, probing on emulate_step function will crash 2.6.18.1 when
it is triggered.

When kprobe is triggered, emulate_step() is on its kernel path and
will cause recursive kprobe fault.  And branch_taken() is called
in emulate_step().  This disallows kprobes on both of them.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

If mmiowb() is always used prior to releasing spinlock as Doc suggests,
then it's safe against preemption; but I'm not convinced that's always
the case.  If preemption occurs between sync and get_paca()->io_sync = 0,
I believe there's no problem.  But in the unlikely event that gcc does
the store relative to another register than r13 (as it did with current),
then there's a small danger of setting another cpu's io_sync to 0, after
it had just set it to 1.  Rewrite ppc64 mmiowb to prevent that.

The remaining io_sync assignments in io.h all get_paca()->io_sync = 1,
which is harmless even if preempted to the wrong cpu (the context switch
itself syncs); and those in spinlock.h are while preemption is disabled.
Signed-off-by: NHugh Dickins <hugh@veritas.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Checking source for other get_paca()->field preemption dangers found that
open_high_hpage_areas does a structure copy into its paca while preemption
is enabled: unsafe however gcc accomplishes it.  Just remove that copy:
it's done safely afterwards by on_each_cpu, as in open_low_hpage_areas.
Signed-off-by: NHugh Dickins <hugh@veritas.com>
Acked-by: NDavid Gibson <dwg@au1.ibm.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Repeated -j20 kernel builds on a G5 Quad running an SMP PREEMPT kernel
would often collapse within a day, some exec failing with "Bad address".
In each case examined, load_elf_binary was doing a kernel_read, but
generic_file_aio_read's access_ok saw current->thread.fs.seg as USER_DS
instead of KERNEL_DS.

objdump of filemap.o shows gcc 4.1.0 emitting "mr r5,r13 ... ld r9,416(r5)"
here for get_paca()->__current, instead of the expected and much more usual
"ld r9,416(r13)"; I've seen other gcc4s do the same, but perhaps not gcc3s.

So, if the task is preempted and rescheduled on a different cpu in between
the mr and the ld, r5 will be looking at a different paca_struct from the
one it's now on, pick up the wrong __current, and perhaps the wrong seg.
Presumably much worse could happen elsewhere, though that split is rare.

Other architectures appear to be safe (x86_64's read_pda is more limiting
than get_paca), but ppc64 needs to force "current" into one instruction.
Signed-off-by: NHugh Dickins <hugh@veritas.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>