Oleksii reported that he had seen an oops similar to this:

BUG: unable to handle kernel NULL pointer dereference at 0000000000000088
IP: [<ffffffff814dcc13>] sock_sendmsg+0x93/0xd0
PGD 0
Oops: 0000 [#1] PREEMPT SMP
Modules linked in: ipt_MASQUERADE xt_REDIRECT xt_tcpudp iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack ip_tables x_tables carl9170 ath usb_storage f2fs nfnetlink_log nfnetlink md4 cifs dns_resolver hid_generic usbhid hid af_packet uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_core videodev rfcomm btusb bnep bluetooth qmi_wwan qcserial cdc_wdm usb_wwan usbnet usbserial mii snd_hda_codec_hdmi snd_hda_codec_realtek iwldvm mac80211 coretemp intel_powerclamp kvm_intel kvm iwlwifi snd_hda_intel cfg80211 snd_hda_codec xhci_hcd e1000e ehci_pci snd_hwdep sdhci_pci snd_pcm ehci_hcd microcode psmouse sdhci thinkpad_acpi mmc_core i2c_i801 pcspkr usbcore hwmon snd_timer snd_page_alloc snd ptp rfkill pps_core soundcore evdev usb_common vboxnetflt(O) vboxdrv(O)Oops#2 Part8
 loop tun binfmt_misc fuse msr acpi_call(O) ipv6 autofs4
CPU: 0 PID: 21612 Comm: kworker/0:1 Tainted: G        W  O 3.10.1SIGN #28
Hardware name: LENOVO 2306CTO/2306CTO, BIOS G2ET92WW (2.52 ) 02/22/2013
Workqueue: cifsiod cifs_echo_request [cifs]
task: ffff8801e1f416f0 ti: ffff880148744000 task.ti: ffff880148744000
RIP: 0010:[<ffffffff814dcc13>]  [<ffffffff814dcc13>] sock_sendmsg+0x93/0xd0
RSP: 0000:ffff880148745b00  EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff880148745b78 RCX: 0000000000000048
RDX: ffff880148745c90 RSI: ffff880181864a00 RDI: ffff880148745b78
RBP: ffff880148745c48 R08: 0000000000000048 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: ffff880181864a00
R13: ffff880148745c90 R14: 0000000000000048 R15: 0000000000000048
FS:  0000000000000000(0000) GS:ffff88021e200000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000088 CR3: 000000020c42c000 CR4: 00000000001407b0
Oops#2 Part7
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Stack:
 ffff880148745b30 ffffffff810c4af9 0000004848745b30 ffff880181864a00
 ffffffff81ffbc40 0000000000000000 ffff880148745c90 ffffffff810a5aab
 ffff880148745bc0 ffffffff81ffbc40 ffff880148745b60 ffffffff815a9fb8
Call Trace:
 [<ffffffff810c4af9>] ? finish_task_switch+0x49/0xe0
 [<ffffffff810a5aab>] ? lock_timer_base.isra.36+0x2b/0x50
 [<ffffffff815a9fb8>] ? _raw_spin_unlock_irqrestore+0x18/0x40
 [<ffffffff810a673f>] ? try_to_del_timer_sync+0x4f/0x70
 [<ffffffff815aa38f>] ? _raw_spin_unlock_bh+0x1f/0x30
 [<ffffffff814dcc87>] kernel_sendmsg+0x37/0x50
 [<ffffffffa081a0e0>] smb_send_kvec+0xd0/0x1d0 [cifs]
 [<ffffffffa081a263>] smb_send_rqst+0x83/0x1f0 [cifs]
 [<ffffffffa081ab6c>] cifs_call_async+0xec/0x1b0 [cifs]
 [<ffffffffa08245e0>] ? free_rsp_buf+0x40/0x40 [cifs]
Oops#2 Part6
 [<ffffffffa082606e>] SMB2_echo+0x8e/0xb0 [cifs]
 [<ffffffffa0808789>] cifs_echo_request+0x79/0xa0 [cifs]
 [<ffffffff810b45b3>] process_one_work+0x173/0x4a0
 [<ffffffff810b52a1>] worker_thread+0x121/0x3a0
 [<ffffffff810b5180>] ? manage_workers.isra.27+0x2b0/0x2b0
 [<ffffffff810bae00>] kthread+0xc0/0xd0
 [<ffffffff810bad40>] ? kthread_create_on_node+0x120/0x120
 [<ffffffff815b199c>] ret_from_fork+0x7c/0xb0
 [<ffffffff810bad40>] ? kthread_create_on_node+0x120/0x120
Code: 84 24 b8 00 00 00 4c 89 f1 4c 89 ea 4c 89 e6 48 89 df 4c 89 60 18 48 c7 40 28 00 00 00 00 4c 89 68 30 44 89 70 14 49 8b 44 24 28 <ff> 90 88 00 00 00 3d ef fd ff ff 74 10 48 8d 65 e0 5b 41 5c 41
 RIP  [<ffffffff814dcc13>] sock_sendmsg+0x93/0xd0
 RSP <ffff880148745b00>
CR2: 0000000000000088

The client was in the middle of trying to send a frame when the
server->ssocket pointer got zeroed out. In most places, that we access
that pointer, the srv_mutex is held. There's only one spot that I see
that the server->ssocket pointer gets set and the srv_mutex isn't held.
This patch corrects that.

The upstream bug report was here:

    https://bugzilla.kernel.org/show_bug.cgi?id=60557

Cc: <stable@vger.kernel.org>
Reported-by: Oleksii Shevchuk <alxchk@gmail.com>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@gmail.com>

Pull NVM Express driver update from Matthew Wilcox.

* git://git.infradead.org/users/willy/linux-nvme:
  NVMe: Merge issue on character device bring-up
  NVMe: Handle ioremap failure
  NVMe: Add pci suspend/resume driver callbacks
  NVMe: Use normal shutdown
  NVMe: Separate controller init from disk discovery
  NVMe: Separate queue alloc/free from create/delete
  NVMe: Group pci related actions in functions
  NVMe: Disk stats for read/write commands only
  NVMe: Bring up cdev on set feature failure
  NVMe: Fix checkpatch issues
  NVMe: Namespace IDs are unsigned
  NVMe: Update nvme_id_power_state with latest spec
  NVMe: Split header file into user-visible and kernel-visible pieces
  NVMe: Call nvme_process_cq from submission path
  NVMe: Remove "process_cq did something" message
  NVMe: Return correct value from interrupt handler
  NVMe: Disk IO statistics
  NVMe: Restructure MSI / MSI-X setup
  NVMe: Use kzalloc instead of kmalloc+memset

Pull NTB (non-transparent bridge) updates from Jon Mason:
 "NTB driver bug fixes to address issues in NTB-RP enablement, spad,
  debugfs, and USD/DSD identification.

  Add a workaround on Xeon NTB devices for b2bdoorbell errata.  Also,
  add new NTB driver features to support 32bit x86, DMA engine support,
  and NTB-RP support.

  Finally, a few clean-ups and update to MAINTAINERS for the NTB git
  tree and wiki location"

* tag 'ntb-3.12' of git://github.com/jonmason/ntb:
  ntb: clean up unnecessary MSI/MSI-X capability find
  MAINTAINERS: Add Website and Git Tree for NTB
  NTB: Update Version
  NTB: Comment Fix
  NTB: Remove unused variable
  NTB: Remove References of non-B2B BWD HW
  NTB: NTB-RP support
  NTB: Rename Variables for NTB-RP
  NTB: Use DMA Engine to Transmit and Receive
  NTB: Enable 32bit Support
  NTB: Update Device IDs
  NTB: BWD Link Recovery
  NTB: Xeon Errata Workaround
  NTB: Correct debugfs to work with more than 1 NTB Device
  NTB: Correct USD/DSD Identification
  NTB: Correct Number of Scratch Pad Registers
  NTB: Add Error Handling in ntb_device_setup

Pull MFD (multi-function device) updates from Samuel Ortiz:
 "For the 3.12 merge window we have one new driver for the DA9063 PMIC
  from Dialog Semiconductor.

  Besides that driver we also have:

   - Device tree support for the s2mps11 driver

   - More devm_* conversion for the pm8921, max89xx, menelaus, tps65010,
     wl1273 and pcf50633-adc drivers.

   - A conversion to threaded IRQ and IRQ domain for the twl6030 driver.

   - A fairly big update for the rtsx driver: Better power saving
     support, better vendor settings handling, and a few fixes.

   - Support for a couple more boards (COMe-bHL6 and COMe-cTH6) for the
     Kontron driver.

   - A conversion to the dev_get_platdata() API for all MFD drivers.

   - A removal of non-DT (legacy) support for the twl6040 driver.

   - A few fixes and additions (Mic detect level) to the wm5110 register
     tables.

   - Regmap support for the davinci_voicecodec driver.

   - The usual bunch of minor cleanups and janitorial fixes"

* tag 'mfd-3.12-1' of git://git.kernel.org/pub/scm/linux/kernel/git/sameo/mfd-next: (81 commits)
  mfd: ucb1x00-core: Rewrite ucb1x00_add_dev()
  mfd: ab8500-debugfs: Apply a check for -ENOMEM after allocating memory for event name
  mfd: ab8500-debugfs: Apply a check for -ENOMEM after allocating memory for sysfs
  mfd: timberdale: Use module_pci_driver
  mfd: timberdale: Remove redundant break
  mfd: timberdale: Staticize local variables
  mfd: ab8500-debugfs: Staticize local variables
  mfd: db8500-prcmu: Staticize clk_mgt
  mfd: db8500-prcmu: Use ANSI function declaration
  mfd: omap-usb-host: Staticize usbhs_driver_name
  mfd: 88pm805: Fix potential NULL pdata dereference
  mfd: 88pm800: Fix potential NULL pdata dereference
  mfd: twl6040: Use regmap for register cache
  mfd: davinci_voicecodec: Provide a regmap for register I/O
  mfd: davinci_voicecodec: Remove unused read and write functions
  mmc: memstick: rtsx: Modify copyright comments
  mmc: rtsx: Clear SD_CLK toggle enable bit if switching voltage fail
  mfd: mmc: rtsx: Change default tx phase
  mfd: pcf50633-adc: Use devm_*() functions
  mfd: rtsx: Copyright modifications
  ...

Pull misc kbuild updates from Michal Marek:
 "In the kbuild misc branch, I have:
   - make rpm-pkg updates, most importantly the rpm package now calls
     /sbin/installkernel
   - make deb-pkg: debuginfo split, correct kernel image path for
     parisc, mips and powerpc and a couple more minor fixes
   - New coccinelle check"

* 'misc' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
  scripts/checkkconfigsymbols.sh: replace echo -e with printf
  Provide version number for Debian firmware package
  coccinelle: replace 0/1 with false/true in functions returning bool
  deb-pkg: add a hook argument to match debian hooks parameters
  deb-pkg: fix installed image path on parisc, mips and powerpc
  deb-pkg: split debug symbols in their own package
  deb-pkg: use KCONFIG_CONFIG instead of .config file directly
  rpm-pkg: add generation of kernel-devel
  rpm-pkg: install firmware files in kernel relative directory
  rpm-pkg: add %post section to create initramfs and grub hooks

Pull kbuild update from Michal Marek:
 "Only these two commits are in the kbuild branch this time:
   - Using filechk for include/config/kernel.release
   - Cleanup in scripts/sortextable.c"

* 'kbuild' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
  kbuild: Do not overwrite include/config/kernel.release needlessly
  scripts: remove unused function in sortextable.c

The only actual current lockref user (dcache) uses zero reference counts
even for perfectly live dentries, because it's a cache: there may not be
any users, but that doesn't mean that we want to throw away the dentry.

At the same time, the dentry cache does have a notion of a truly "dead"
dentry that we must not even increment the reference count of, because
we have pruned it and it is not valid.

Currently that distinction is not visible in the lockref itself, and the
dentry cache validation uses "lockref_get_or_lock()" to either get a new
reference to a dentry that already had existing references (and thus
cannot be dead), or get the dentry lock so that we can then verify the
dentry and increment the reference count under the lock if that
verification was successful.

That's all somewhat complicated.

This adds the concept of being "dead" to the lockref itself, by simply
using a count that is negative.  This allows a usage scenario where we
can increment the refcount of a dentry without having to validate it,
and pushing the special "we killed it" case into the lockref code.

The dentry code itself doesn't actually use this yet, and it's probably
too late in the merge window to do that code (the dentry_kill() code
with its "should I decrement the count" logic really is pretty complex
code), but let's introduce the concept at the lockref level now.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The code got rewritten, but the comments got copied as-is from older
versions, and as a result the argument name in the comment didn't
actually match the code any more.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull vfs pile 2 (of many) from Al Viro:
 "Mostly Miklos' series this time"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  constify dcache.c inlined helpers where possible
  fuse: drop dentry on failed revalidate
  fuse: clean up return in fuse_dentry_revalidate()
  fuse: use d_materialise_unique()
  sysfs: use check_submounts_and_drop()
  nfs: use check_submounts_and_drop()
  gfs2: use check_submounts_and_drop()
  afs: use check_submounts_and_drop()
  vfs: check unlinked ancestors before mount
  vfs: check submounts and drop atomically
  vfs: add d_walk()
  vfs: restructure d_genocide()

Pull namespace changes from Eric Biederman:
 "This is an assorted mishmash of small cleanups, enhancements and bug
  fixes.

  The major theme is user namespace mount restrictions.  nsown_capable
  is killed as it encourages not thinking about details that need to be
  considered.  A very hard to hit pid namespace exiting bug was finally
  tracked and fixed.  A couple of cleanups to the basic namespace
  infrastructure.

  Finally there is an enhancement that makes per user namespace
  capabilities usable as capabilities, and an enhancement that allows
  the per userns root to nice other processes in the user namespace"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace:
  userns:  Kill nsown_capable it makes the wrong thing easy
  capabilities: allow nice if we are privileged
  pidns: Don't have unshare(CLONE_NEWPID) imply CLONE_THREAD
  userns: Allow PR_CAPBSET_DROP in a user namespace.
  namespaces: Simplify copy_namespaces so it is clear what is going on.
  pidns: Fix hang in zap_pid_ns_processes by sending a potentially extra wakeup
  sysfs: Restrict mounting sysfs
  userns: Better restrictions on when proc and sysfs can be mounted
  vfs: Don't copy mount bind mounts of /proc/<pid>/ns/mnt between namespaces
  kernel/nsproxy.c: Improving a snippet of code.
  proc: Restrict mounting the proc filesystem
  vfs: Lock in place mounts from more privileged users

Pull security subsystem updates from James Morris:
 "Nothing major for this kernel, just maintenance updates"

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security: (21 commits)
  apparmor: add the ability to report a sha1 hash of loaded policy
  apparmor: export set of capabilities supported by the apparmor module
  apparmor: add the profile introspection file to interface
  apparmor: add an optional profile attachment string for profiles
  apparmor: add interface files for profiles and namespaces
  apparmor: allow setting any profile into the unconfined state
  apparmor: make free_profile available outside of policy.c
  apparmor: rework namespace free path
  apparmor: update how unconfined is handled
  apparmor: change how profile replacement update is done
  apparmor: convert profile lists to RCU based locking
  apparmor: provide base for multiple profiles to be replaced at once
  apparmor: add a features/policy dir to interface
  apparmor: enable users to query whether apparmor is enabled
  apparmor: remove minimum size check for vmalloc()
  Smack: parse multiple rules per write to load2, up to PAGE_SIZE-1 bytes
  Smack: network label match fix
  security: smack: add a hash table to quicken smk_find_entry()
  security: smack: fix memleak in smk_write_rules_list()
  xattr: Constify ->name member of "struct xattr".
  ...

Pull crypto update from Herbert Xu:
 "Here is the crypto update for 3.12:

   - Added MODULE_SOFTDEP to allow pre-loading of modules.
   - Reinstated crct10dif driver using the module softdep feature.
   - Allow via rng driver to be auto-loaded.

   - Split large input data when necessary in nx.
   - Handle zero length messages correctly for GCM/XCBC in nx.
   - Handle SHA-2 chunks bigger than block size properly in nx.

   - Handle unaligned lengths in omap-aes.
   - Added SHA384/SHA512 to omap-sham.
   - Added OMAP5/AM43XX SHAM support.
   - Added OMAP4 TRNG support.

   - Misc fixes"

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (66 commits)
  Reinstate "crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework"
  hwrng: via - Add MODULE_DEVICE_TABLE
  crypto: fcrypt - Fix bitoperation for compilation with clang
  crypto: nx - fix SHA-2 for chunks bigger than block size
  crypto: nx - fix GCM for zero length messages
  crypto: nx - fix XCBC for zero length messages
  crypto: nx - fix limits to sg lists for AES-CCM
  crypto: nx - fix limits to sg lists for AES-XCBC
  crypto: nx - fix limits to sg lists for AES-GCM
  crypto: nx - fix limits to sg lists for AES-CTR
  crypto: nx - fix limits to sg lists for AES-CBC
  crypto: nx - fix limits to sg lists for AES-ECB
  crypto: nx - add offset to nx_build_sg_lists()
  padata - Register hotcpu notifier after initialization
  padata - share code between CPU_ONLINE and CPU_DOWN_FAILED, same to CPU_DOWN_PREPARE and CPU_UP_CANCELED
  hwrng: omap - reorder OMAP TRNG driver code
  crypto: omap-sham - correct dma burst size
  crypto: omap-sham - Enable Polling mode if DMA fails
  crypto: tegra-aes - bitwise vs logical and
  crypto: sahara - checking the wrong variable
  ...

Pull networking fixes from David Miller:
 "A quick set of fixes, some to deal with fallout from yesterday's
  net-next merge.

   1) Fix compilation of bnx2x driver with CONFIG_BNX2X_SRIOV disabled,
      from Dmitry Kravkov.

   2) Fix a bnx2x regression caused by one of Dave Jones's mistaken
      braces changes, from Eilon Greenstein.

   3) Add some protective filtering in the netlink tap code, from Daniel
      Borkmann.

   4) Fix TCP congestion window growth regression after timeouts, from
      Yuchung Cheng.

   5) Correctly adjust TCP's rcv_ssthresh for out of order packets, from
      Eric Dumazet"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  tcp: properly increase rcv_ssthresh for ofo packets
  net: add documentation for BQL helpers
  mlx5: remove unused MLX5_DEBUG param in Kconfig
  bnx2x: Restore a call to config_init
  bnx2x: fix broken compilation with CONFIG_BNX2X_SRIOV is not set
  tcp: fix no cwnd growth after timeout
  net: netlink: filter particular protocols from analyzers

Pull hwmon fixes from Jean Delvare.

* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  hwmon: (emc6w201) Do not declare enum variable
  hwmon: (w83792d) Update module author

Pull GPIO updates from Linus Walleij:
 "This is the bulk of GPIO changes for the v3.12 series:

   - A new driver for the TZ1090 PDC which is used on the metag
     architecture.

   - A new driver for the Kontron ETX or COMexpress GPIO block.  This is
     found on some ETX x86 devices.

   - A new driver for the Fintek Super-I/O chips, used on some x86
     boards.

   - Added device tree probing on a few select GPIO blocks.

   - Drop the Exynos support from the Samsung GPIO driver.

     The Samsung maintainers have moved over to use the modernized pin
     control driver to provide GPIO for the modern platforms instead.

   - The usual bunch of non-critical fixes and cleanups"

* tag 'gpio-v3.12-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio: (36 commits)
  gpio: return -ENOTSUPP if debounce cannot be set
  gpio: improve error path in gpiolib
  gpio: add GPIO support for F71882FG and F71889F
  of: add vendor prefix for Microchip Technology Inc
  gpio: mcp23s08: rename the device tree property
  gpio: samsung: Drop support for Exynos SoCs
  gpio: pcf857x: Remove pdata argument to pcf857x_irq_domain_init()
  gpio: pcf857x: Sort headers alphabetically
  gpio: max7301: Reverting "Do not force SPI speed when using OF Platform"
  gpio: Fix bit masking in Kontron PLD GPIO driver
  gpio: pca953x: fix gpio input on gpio offsets >= 8
  drivers/gpio: simplify use of devm_ioremap_resource
  drivers/gpio/gpio-omap.c: convert comma to semicolon
  gpio-lynxpoint: Fix warning about unbalanced pm_runtime_enable
  gpio: Fix platform driver name in Kontron PLD GPIO driver
  gpio: adnp: Fix segfault if request_threaded_irq fails
  gpio: msm: Staticize local variable 'msm_gpio'
  gpio: gpiolib-of.c: make error message more meaningful by adding the node name and index
  gpio: use dev_get_platdata()
  gpio/mxc: add chained_irq_enter/exit() to mx2_gpio_irq_handler
  ...

Pull input updates from Dmitry Torokhov:
 "A new driver for slidebar on Ideapad laptops and a bunch of assorted
  driver fixes"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input: (32 commits)
  Input: add SYN_MAX and SYN_CNT constants
  Input: max11801_ts - convert to devm
  Input: egalax-ts - fix typo and improve text
  Input: MAINTAINERS - change maintainer for cyttsp driver
  Input: cyttsp4 - kill 'defined but not used' compiler warnings
  Input: add driver for slidebar on Lenovo IdeaPad laptops
  Input: omap-keypad - set up irq type from DT
  Input: omap-keypad - enable wakeup capability for keypad.
  Input: omap-keypad - clear interrupts on open
  Input: omap-keypad - convert to threaded IRQ
  Input: omap-keypad - use bitfiled instead of hardcoded values
  Input: cyttsp4 - remove useless NULL test from cyttsp4_watchdog_timer()
  Input: wacom - fix error return code in wacom_probe()
  Input: as5011 - fix error return code in as5011_probe()
  Input: keyboard, serio - simplify use of devm_ioremap_resource
  Input: tegra-kbc - simplify use of devm_ioremap_resource
  Input: htcpen - fix incorrect placement of __initdata
  Input: qt1070 - add power management ops
  Input: wistron_btns - add MODULE_DEVICE_TABLE
  Input: wistron_btns - mark the Medion MD96500 keymap as tested
  ...

This reverts commits 61e00655, 73f8645d and 8e22ecb6:
  "Input: introduce BTN/ABS bits for drums and guitars"
  "HID: wiimote: add support for Guitar-Hero drums"
  "HID: wiimote: add support for Guitar-Hero guitars"

The extra new ABS_xx values resulted in ABS_MAX no longer being a
power-of-two, which broke the comparison logic.  It also caused the
ioctl numbers to overflow into the next byte, causing problems for that.

We'll try again for 3.13.
Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Acked-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge first round of changes for 3.12 merge window.

This patch reinstates commits
	67822649
	39761214
	0b95a7f8
	31d93962
	2d31e518

Now that module softdeps are in the kernel we can use that to resolve
the boot issue which cause the revert.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Merge upstream tree in order to reinstate crct10dif.

Pull ARM SoC board updates from Olof Johansson:
 "Board updates for 3.12.  Again, a bit of domain overlap with SoC and
  DT branches, but most of this is around legacy code and board support.
  We've found that platform maintainers have a hard time separating all
  of these out and might move towards fewer branches for next release.

   - Removal of a number of Marvell Kirkwood board files, since contents
     is now common and mostly configured via DT.
   - Device-tree updates for Marvell Dove, including irqchip and
     clocksource setup.
   - Defconfig updates.  Gotta go somewhere.  One new one for Renesas
     Lager.
   - New backlight drivers for backlights used on Renesas shmobile
     platforms.
   - Removal of Renesas leds driver.
   - Shuffling of some of the new Broadcom platforms to give room for
     others in the same mach directory.  More in 3.13"

* tag 'boards-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (67 commits)
  mmc: sdhci-bcm-kona: Staticize sdhci_bcm_kona_card_event
  mmc: sdhci-bcm-kona: Remove unneeded version.h inclusion
  ARM: bcm: Make secure API call optional
  ARM: DT: binding fixup to align with vendor-prefixes.txt (drivers)
  ARM: mmc: fix NONREMOVABLE test in sdhci-bcm-kona
  ARM: bcm: Rename board_bcm
  mmc: sdhci-bcm-kona: make linker-section warning go away
  ARM: tegra: defconfig updates
  ARM: dove: add initial DT file for Globalscale D2Plug
  ARM: dove: add GPIO IR receiver node to SolidRun CuBox
  ARM: dove: add common pinmux functions to DT
  ARM: dove: add cpu device tree node
  ARM: dove: update dove_defconfig with SI5351, PCI, and xHCI
  arch/arm/mach-kirkwood: Avoid using ARRAY_AND_SIZE(e) as a function argument
  ARM: kirkwood: fix DT building and update defconfig
  ARM: kirkwood: Remove all remaining trace of DNS-320/325 platform code
  ARM: configs: disable DEBUG_LL in bcm_defconfig
  ARM: bcm281xx: Board specific reboot code
  ARM bcm281xx: Turn on socket & network support.
  ARM: bcm281xx: Turn on L2 cache.
  ...

Pull ARM SoC platform changes from Olof Johansson:
 "This branch contains mostly additions and changes to platform
  enablement and SoC-level drivers.  Since there's sometimes a
  dependency on device-tree changes, there's also a fair amount of
  those in this branch.

  Pieces worth mentioning are:

   - Mbus driver for Marvell platforms, allowing kernel configuration
     and resource allocation of on-chip peripherals.
   - Enablement of the mbus infrastructure from Marvell PCI-e drivers.
   - Preparation of MSI support for Marvell platforms.
   - Addition of new PCI-e host controller driver for Tegra platforms
   - Some churn caused by sharing of macro names between i.MX 6Q and 6DL
     platforms in the device tree sources and header files.
   - Various suspend/PM updates for Tegra, including LP1 support.
   - Versatile Express support for MCPM, part of big little support.
   - Allwinner platform support for A20 and A31 SoCs (dual and quad
     Cortex-A7)
   - OMAP2+ support for DRA7, a new Cortex-A15-based SoC.

  The code that touches other architectures are patches moving MSI
  arch-specific functions over to weak symbols and removal of
  ARCH_SUPPORTS_MSI, acked by PCI maintainers"

* tag 'soc-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (266 commits)
  tegra-cpuidle: provide stub when !CONFIG_CPU_IDLE
  PCI: tegra: replace devm_request_and_ioremap by devm_ioremap_resource
  ARM: tegra: Drop ARCH_SUPPORTS_MSI and sort list
  ARM: dts: vf610-twr: enable i2c0 device
  ARM: dts: i.MX51: Add one more I2C2 pinmux entry
  ARM: dts: i.MX51: Move pins configuration under "iomuxc" label
  ARM: dtsi: imx6qdl-sabresd: Add USB OTG vbus pin to pinctrl_hog
  ARM: dtsi: imx6qdl-sabresd: Add USB host 1 VBUS regulator
  ARM: dts: imx27-phytec-phycore-som: Enable AUDMUX
  ARM: dts: i.MX27: Disable AUDMUX in the template
  ARM: dts: wandboard: Add support for SDIO bcm4329
  ARM: i.MX5 clocks: Remove optional clock setup (CKIH1) from i.MX51 template
  ARM: dts: imx53-qsb: Make USBH1 functional
  ARM i.MX6Q: dts: Enable I2C1 with EEPROM and PMIC on Phytec phyFLEX-i.MX6 Ouad module
  ARM i.MX6Q: dts: Enable SPI NOR flash on Phytec phyFLEX-i.MX6 Ouad module
  ARM: dts: imx6qdl-sabresd: Add touchscreen support
  ARM: imx: add ocram clock for imx53
  ARM: dts: imx: ocram size is different between imx6q and imx6dl
  ARM: dts: imx27-phytec-phycore-som: Fix regulator settings
  ARM: dts: i.MX27: Remove clock name from CPU node
  ...

A recent patch made it possible to bring up the character handle when the
device is responsive but not accepting a set-features command. Another
recent patch moved the initialization that requires we move where the
checks for this condition occur. This patch merges these two ideas so
it works much as before.
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>

Pull ARM SoC DT updates from Olof Johansson:
 "Device tree and bindings updates for 3.12.

  General additions of various on-chip and on-board peripherals on
  various platforms as support gets added.  Some of the bigger changes
  are:

   - Addition of (new) PCI-e support on Tegra.
   - More Tegra4 support, including PMC configuration for Dalmore.
   - Addition of a new board for Exynos4 (trats2) and more bindings for
     4x12 IP.
   - Addition of Allwinner A20 and A31 SoC and board files.
   - Move of the ST Ericsson device tree files to now use ste-* prefix.
   - More move of hardware description of shmobile platforms to DT.
   - Two new board dts files for Freescale MXs"

* tag 'dt-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (177 commits)
  dts: Rename DW APB timer compatible strings
  dts: Deprecate ALTR as a vendor prefix
  of: add vendor prefix for Altera Corp.
  ARM: at91/dt: sam9x5ek: add sound configuration
  ARM: at91/dt: sam9x5ek: enable SSC
  ARM: at91/dt: sam9x5ek: add WM8731 codec
  ARM: at91/dt: sam9x5: add SSC DMA parameters
  ARM: at91/dt: add at91rm9200 PQFP package version
  ARM: at91: at91rm9200: set default mmc0 pinctrl-names
  ARM: at91: at91sam9n12: correct pin number of gpio-key
  ARM: at91: at91sam9n12: add qt1070 support
  ARM: at91: at91sam9n12: add pinctrl of TWI
  ARM: at91: Add PMU support for sama5d3
  ARM: at91: at91sam9260: add missing pinctrl-names on mmc
  ARM: tegra: configure power off for Dalmore
  ARM: DT: binding fixup to align with vendor-prefixes.txt (DT)
  ARM: dts: add sdio blocks to bcm28155-ap board
  ARM: dts: align sdio numbers to HW definition
  ARM: sun7i: Add Olimex A20-Olinuxino-Micro support
  ARM: sun7i: Add Allwinner A20 DTSI
  ...

Pull ARM SoC cleanups from Olof Johansson:
 "This branch contains code cleanups, moves and removals for 3.12.

  There's a large number of various cleanups, and a nice net removal of
  13500 lines of code.

  Highlights worth mentioning are:

   - A series of patches from Stephen Boyd removing the ARM local timer
     API.
   - Move of Qualcomm MSM IOMMU code to drivers/iommu.
   - Samsung PWM driver cleanups from Tomasz Figa, removing legacy PWM
     driver and switching over to the drivers/pwm one.
   - Removal of some unusued auto-generated headers for OMAP2+ (PRM/CM).

  There's also a move of a header file out of include/linux/i2c/ to
  platform_data, where it really belongs.  It touches mostly ARM
  platform code for include changes so we took it through our tree"

* tag 'cleanup-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (83 commits)
  ARM: OMAP2+: Add back the define for AM33XX_RST_GLOBAL_WARM_SW_MASK
  gpio: (gpio-pca953x) move header to linux/platform_data/
  arm: zynq: hotplug: Remove unreachable code
  ARM: SAMSUNG: Remove unnecessary exynos4_default_sdhci*()
  tegra: simplify use of devm_ioremap_resource
  ARM: SAMSUNG: Remove plat/regs-timer.h header
  ARM: SAMSUNG: Remove remaining uses of plat/regs-timer.h header
  ARM: SAMSUNG: Remove pwm-clock infrastructure
  ARM: SAMSUNG: Remove old PWM timer platform devices
  pwm: Remove superseded pwm-samsung-legacy driver
  ARM: SAMSUNG: Modify board files to use new PWM platform device
  ARM: SAMSUNG: Rework private data handling in dev-backlight
  pwm: Add new pwm-samsung driver
  ARM: mach-mvebu: remove redundant DT parsing and validation
  ARM: msm: Only compile io.c on platforms that use it
  iommu/msm: Move mach includes to iommu directory
  ARM: msm: Remove devices-iommu.c
  ARM: msm: Move mach/board.h contents to common.h
  ARM: msm: Migrate msm_timer to CLOCKSOURCE_OF_DECLARE
  ARM: msm: Remove TMR and TMR0 static mappings
  ...

Pull ARM SoC low-priority fixes from Olof Johansson:
 "This branch contains a handful of fixes for various platforms that
  weren't serious enough to be included in late 3.11-rc releases.  Most
  of them are for minor cleanups and cosmetic fixes.

  There's also a bit of code removal here, one board file removal for
  clps711x, and removal of some legacy device creation on OMAP2+"

* tag 'fixes-nc-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (25 commits)
  ARM: OMAP2: use 'int' instead of 'unsigned' for variable 'gpmc_irq_start'
  ARM: OMAP2: remove useless variable 'ret'
  ARM: OMAP: dma: fix error return code in omap_system_dma_probe()
  ARM: OMAP2+: fix wrong address when loading PRM_FRAC_INCREMENTOR_DENUMERATOR_RELOAD
  ARM: OMAP2+: am33xx-restart: trigger warm reset on omap2+ boards
  ARM: OMAP2: Use a consistent AM33XX SoC option description
  ARM: OMAP2+: Remove legacy device creation for McPDM and DMIC
  ARM: clps711x: edb7211: Remove extra iotable_init() call
  ARM: clps711x: autcpu12: Remove incorrect config checking
  ARM: clps711x: Drop fortunet board support
  ARM: clps711x: Remove the special name for the syscon driver
  ARM: dts: Fix memory node in skeleton64.dtsi
  ARM: Keystone: Convert device tree file to use IRQ defines
  ARM: keystone: use #include to include skeleton.dtsi
  ARM: keystone: Drop the un-necessary dsb from keystone_cpu_smc()
  ARM: Keystone: No need to preserve r12 across smc call
  ARM: keystone: remove redundant smp_init_cpus definition
  ARM: keystone: drop useless HAVE_SCHED_CLOCK
  ARM: dove: fix missing __init section of dove_mpp_gpio_mode
  ARM: shmobile: armadillo800eva-reference: fix compiler warning
  ...

Pull ia64 fixes from Tony Luck:
 "Couple of small cleanups for ia64"

* tag 'please-pull-misc-3.12' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux:
  [IA64] Use asm-generic/bitops/builtin-ffs.h
  [IA64] dmi.h: Make dmi_alloc use kzalloc

TCP receive window handling is multi staged.

A socket has a memory budget, static or dynamic, in sk_rcvbuf.

Because we do not really know how this memory budget translates to
a TCP window (payload), TCP announces a small initial window
(about 20 MSS).

When a packet is received, we increase TCP rcv_win depending
on the payload/truesize ratio of this packet. Good citizen
packets give a hint that it's reasonable to have rcv_win = sk_rcvbuf/2

This heuristic takes place in tcp_grow_window()

Problem is : We currently call tcp_grow_window() only for in-order
packets.

This means that reorders or packet losses stop proper grow of
rcv_win, and senders are unable to benefit from fast recovery,
or proper reordering level detection.

Really, a packet being stored in OFO queue is not a bad citizen.
It should be part of the game as in-order packets.

In our traces, we very often see sender is limited by linux small
receive windows, even if linux hosts use autotuning (DRS) and should
allow rcv_win to grow to ~3MB.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Provide a kernel-doc comment documentation for the BQL helpers:
- netdev_sent_queue
- netdev_completed_queue
- netdev_reset_queue

Similarly to how it is done for the other functions, the documentation
only covers the function operating on struct net_device and not struct
netdev_queue.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch proposes to remove the MLX5_DEBUG kernel configuration
parameter defined in drivers/net/ethernet/mellanox/mlx5/core/Kconfig,
but used nowhere in the makefiles and source code.

This could also be fixed by using this parameter,
but this may be a leftover from driver development...
Signed-off-by: Michael Opdenacker <michael.opdenacker@free-electrons.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit c0a77ec7 'bnx2x: Add missing braces in
bnx2x:bnx2x_link_initialize' identified indentation problem, but resolved it
by adding braces instead of fixing the indentation. The braces now prevents a
config_init call in some cases, though it should be called regardless of that
condition. This patch removes the braces and fix the confusing indentation
that caused this mess.
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
CC: Dave Jones <davej@redhat.com>
Tested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Since commit 60cad4e6
"bnx2x: VF RSS support - VF side" fails to compile w/o
CONFIG_BNX2X_SRIOV option.
Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In commit 0f7cc9a3 "tcp: increase throughput when reordering is high",
it only allows cwnd to increase in Open state. This mistakenly disables
slow start after timeout (CA_Loss). Moreover cwnd won't grow if the
state moves from Disorder to Open later in tcp_fastretrans_alert().

Therefore the correct logic should be to allow cwnd to grow as long
as the data is received in order in Open, Loss, or even Disorder state.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix finer-grained control and let only a whitelist of allowed netlink
protocols pass, in our case related to networking. If later on, other
subsystems decide they want to add their protocol as well to the list
of allowed protocols they shall simply add it. While at it, we also
need to tell what protocol is in use otherwise BPF_S_ANC_PROTOCOL can
not pick it up (as it's not filled out).
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull Tile arch updates from Chris Metcalf:
 "These changes bring in a bunch of new functionality that has been
  maintained internally at Tilera over the last year, plus other stray
  bits of work that I've taken into the tile tree from other folks.

  The changes include some PCI root complex work, interrupt-driven
  console support, support for performing fast-path unaligned data
  fixups by kernel-based JIT code generation, CONFIG_PREEMPT support,
  vDSO support for gettimeofday(), a serial driver for the tilegx
  on-chip UART, KGDB support, more optimized string routines, support
  for ftrace and kprobes, improved ASLR, and many bug fixes.

  We also remove support for the old TILE64 chip, which is no longer
  buildable"

* git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile: (85 commits)
  tile: refresh tile defconfig files
  tile: rework <asm/cmpxchg.h>
  tile PCI RC: make default consistent DMA mask 32-bit
  tile: add null check for kzalloc in tile/kernel/setup.c
  tile: make __write_once a synonym for __read_mostly
  tile: remove support for TILE64
  tile: use asm-generic/bitops/builtin-*.h
  tile: eliminate no-op "noatomichash" boot argument
  tile: use standard tile_bundle_bits type in traps.c
  tile: simplify code referencing hypervisor API addresses
  tile: change <asm/system.h> to <asm/switch_to.h> in comments
  tile: mark pcibios_init() as __init
  tile: check for correct compiler earlier in asm-offsets.c
  tile: use standard 'generic-y' model for <asm/hw_irq.h>
  tile: use asm-generic version of <asm/local64.h>
  tile PCI RC: add comment about "PCI hole" problem
  tile: remove DEBUG_EXTRA_FLAGS kernel config option
  tile: add virt_to_kpte() API and clean up and document behavior
  tile: support FRAME_POINTER
  tile: support reporting Tilera hypervisor statistics
  ...

Pull ARM64 update from Catalin Marinas:
 - User tagged pointers support (top 8-bit of user pointers
   automatically ignored by the CPU).
 - Kernel mode NEON (no users for arm64 yet but work in progress).
 - arm64 kernel Image header extended to accommodate future EFI stub.
 - Remove BogoMIPS reporting (not relevant, it's just the timer
   frequency).
 - Clean-up (EM_AARCH64/EM_ARM to elf-em.h, ELF notes in read-only
   segment, unused variable).
 - Bug-fixes (RAM boundaries not 2MB aligned, perf, includes).

* tag 'arm64-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cmarinas/linux-aarch64:
  Documentation/arm64: clarify requirements for DTB placement
  arm64: mm: permit use of tagged pointers at EL0
  Move the EM_ARM and EM_AARCH64 definitions to uapi/linux/elf-em.h
  arm64: Remove unused cpu_name ascii in arch/arm64/mm/proc.S
  arm64: delay: don't bother reporting bogomips in /proc/cpuinfo
  arm64: Fix mapping of memory banks not ending on a PMD_SIZE boundary
  arm64: move elf notes into readonly segment
  arm64: Enable interrupts in the EL0 undef handler
  arm64: Expand arm64 image header
  ARM64: include: asm: include "asm/types.h" in "pgtable-2level-types.h" and "pgtable-3level-types.h"
  arm64: add support for kernel mode NEON
  arm64: perf: fix ARMv8 EVTYPE_MASK to include NSH bit
  arm64: perf: fix group validation when using enable_on_exec

Pull Microblaze patches from Michal Simek:
 - PCI fixes
 - Selfmod code removing
 - Intc and timer fixes
 - Adding new MB versions
 - Minor fixes

* tag 'microblaze-3.12-rc1' of git://git.monstr.eu/linux-2.6-microblaze:
  microblaze: Show message when reset gpio is not present
  microblaze: Add linux.bin.ub target
  microblaze: Add PVR version string for MB v9.0 and v9.1
  microblaze: timer: Replace microblaze_ prefix by xilinx_
  microblaze: timer: Update header
  microblaze: timer: Remove unused header
  microblaze: timer: Clear driver init function
  microblaze: timer: Use CLKSRC_OF initialization
  microblaze: intc: Remove unused header
  microblaze: intc: Clean driver init function
  microblaze: intc: Using irqchip
  microblaze: intc: Update header
  microblaze: intc: Remove unused headers
  microblaze: Remove selfmodified feature
  of/pci: Use of_pci_range_parser

Pull powerpc updates from Ben Herrenschmidt:
 "Here's the powerpc batch for this merge window.  Some of the
  highlights are:

   - A bunch of endian fixes ! We don't have full LE support yet in that
     release but this contains a lot of fixes all over arch/powerpc to
     use the proper accessors, call the firmware with the right endian
     mode, etc...

   - A few updates to our "powernv" platform (non-virtualized, the one
     to run KVM on), among other, support for bridging the P8 LPC bus
     for UARTs, support and some EEH fixes.

   - Some mpc51xx clock API cleanups in preparation for a clock API
     overhaul

   - A pile of cleanups of our old math emulation code, including better
     support for using it to emulate optional FP instructions on
     embedded chips that otherwise have a HW FPU.

   - Some infrastructure in selftest, for powerpc now, but could be
     generalized, initially used by some tests for our perf instruction
     counting code.

   - A pile of fixes for hotplug on pseries (that was seriously
     bitrotting)

   - The usual slew of freescale embedded updates, new boards, 64-bit
     hiberation support, e6500 core PMU support, etc..."

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (146 commits)
  powerpc: Correct FSCR bit definitions
  powerpc/xmon: Fix printing of set of CPUs in xmon
  powerpc/pseries: Move lparcfg.c to platforms/pseries
  powerpc/powernv: Return secondary CPUs to firmware on kexec
  powerpc/btext: Fix CONFIG_PPC_EARLY_DEBUG_BOOTX on ppc32
  powerpc: Cleanup handling of the DSCR bit in the FSCR register
  powerpc/pseries: Child nodes are not detached by dlpar_detach_node
  powerpc/pseries: Add mising of_node_put in delete_dt_node
  powerpc/pseries: Make dlpar_configure_connector parent node aware
  powerpc/pseries: Do all node initialization in dlpar_parse_cc_node
  powerpc/pseries: Fix parsing of initial node path in update_dt_node
  powerpc/pseries: Pack update_props_workarea to map correctly to rtas buffer header
  powerpc/pseries: Fix over writing of rtas return code in update_dt_node
  powerpc/pseries: Fix creation of loop in device node property list
  powerpc: Skip emulating & leave interrupts off for kernel program checks
  powerpc: Add more exception trampolines for hypervisor exceptions
  powerpc: Fix location and rename exception trampolines
  powerpc: Add more trap names to xmon
  powerpc/pseries: Add a warning in the case of cross-cpu VPA registration
  powerpc: Update the 00-Index in Documentation/powerpc
  ...

These are based on the current shipping versions of the config files
from Tilera, as synced up to the tip, so are a better starting point
for folks who want a default configuration.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>

The macrology in cmpxchg.h was designed to allow arbitrary pointer
and integer values to be passed through the routines.  To support
cmpxchg() on 64-bit values on the 32-bit tilepro architecture, we
used the idiom "(typeof(val))(typeof(val-val))".  This way, in the
"size 8" branch of the switch, when the underlying cmpxchg routine
returns a 64-bit quantity, we cast it first to a typeof(val-val)
quantity (i.e. size_t if "val" is a pointer) with no warnings about
casting between pointers and integers of different sizes, then cast
onwards to typeof(val), again with no warnings.  If val is not a
pointer type, the additional cast is a no-op.  We can't replace the
typeof(val-val) cast with (for example) unsigned long, since then if
"val" is really a 64-bit type, we cast away the high bits.

HOWEVER, this fails with current gcc (through 4.7 at least) if "val"
is a pointer to an incomplete type.  Unfortunately gcc isn't smart
enough to realize that "val - val" will always be a size_t type
even if it's an incomplete type pointer.

Accordingly, I've reworked the way we handle the casting.  We have
given up the ability to use cmpxchg() on 64-bit values on tilepro,
which is OK in the kernel since we should use cmpxchg64() explicitly
on such values anyway.  As a result, I can just use simple "unsigned
long" casts internally.

As I reworked it, I realized it would be cleaner to move the
architecture-specific conditionals for cmpxchg and xchg out of the
atomic.h headers and into cmpxchg, and then use the cmpxchg() and
xchg() primitives directly in atomic.h and elsewhere.  This allowed
the cmpxchg.h header to stand on its own without relying on the
implicit include of it that is performed by <asm/atomic.h>.
It also allowed collapsing the atomic_xchg/atomic_cmpxchg routines
from atomic_{32,64}.h into atomic.h.

I improved the tests that guard the allowed size of the arguments
to the routines to use a __compiletime_error() test.  (By avoiding
the use of BUILD_BUG, I could include cmpxchg.h into bitops.h as
well and use the macros there, which is otherwise impossible due
to include order dependency issues.)

The tilepro _atomic_xxx internal methods were previously set up to
take atomic_t and atomic64_t arguments, which isn't as convenient
with the new model, so I modified them to take int or u64 arguments,
which is consistent with how they used the arguments internally
anyway, so provided some nice simplification there too.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>