Commit Graph

347980 Commits

Author SHA1 Message Date
Max Filippov
72100ed7ef xtensa: provide endianness macro for sparse
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:25 -08:00
Max Filippov
ec747b21c7 xtensa: fix RASID SR initialization
set_rasid_register accepts new RASID SR value, but ASID_USER_FIRST is
ASID value for the ring 1; RASID value is made by ASID_INSERT macro.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:25 -08:00
Max Filippov
eab5e7a79d xtensa: initialize CPENABLE SR when core has one
XCHAL_CP_NUM is defined in variant/tie.h and it is not included by
head.S, leaving CPENABLE register uninitialised. XCHAL_HAVE_CP is
defined in variant/core.h to 1 when core has CPENABLE SR.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:25 -08:00
Max Filippov
79fcf52ba0 xtensa: reset all timers on initialization
There are XCHAL_NUM_TIMERS, reset them all.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:25 -08:00
Wei Yongjun
90e3bc798b Use for_each_compatible_node() macro.
Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Acked-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:25 -08:00
Max Filippov
5584b4da78 xtensa: add XTFPGA DTS
Add common XTFPGA parts as *.dtsi (base board, flash) and DTS for LX60
and for ML605.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:24 -08:00
Max Filippov
0d456bad36 xtensa: add support for the XTFPGA boards
The Avnet LX60/LX110/LX200 board is an FPGA board that can be configured with
an Xtensa processor and an OpenCores Ethernet device.

Signed-off-by: Chris Zankel <chris@zankel.net>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:24 -08:00
Max Filippov
da844a8177 xtensa: add device trees support
Device trees allow specification of hardware topology and device
parameters at runtime instead of hard-coding them in platform setup
code. This allows running single binary kernel on a range of compatible
boards.

New boot parameters tag BP_TAG_FDT is allocated and a pointer to flat
device tree is passed in it.

Note that current interrupt mapping scheme uses single cell for
interrupt identification. That means that IRQ numbers used in DTS must
be CPU internal IRQ numbers, not external. It is possible to extend
interrupt identification to two cells, and use second cell to tell
external IRQ numbers form internal. That would allow to use single DTS
on multiple boards with different mapping of external IRQ numbers.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:23 -08:00
Max Filippov
2206d5dd9a xtensa: add IRQ domains support
IRQ domains provide a mechanism for conversion of linux IRQ numbers to
hardware IRQ numbers and vice versus. It is used by OpenFirmware for
linking device tree objects to their respective interrupt controllers.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:23 -08:00
Max Filippov
0322cabd39 xtensa: add U-Boot image support (uImage).
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:23 -08:00
Max Filippov
3f5ec298e5 xtensa: clean up boot make rules
- remove duplicate rules for binary and packed image
- use predefined macros for ld/objcopy/gzip
- remove build-id section from bootable elf image

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:23 -08:00
Max Filippov
599bf77a0d xtensa: fix mb and wmb definitions
Define mb and wmb as memw to force memory barrier.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:23 -08:00
Max Filippov
71872b5fb2 xtensa: add s32c1i-based spinlock implementations
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:22 -08:00
Max Filippov
e5a9f6adba xtensa: add s32c1i-based bitops implementations
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:22 -08:00
Max Filippov
219b1e4c61 xtensa: add s32c1i-based atomic ops implementations
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:22 -08:00
Max Filippov
00273125c3 xtensa: add s32c1i sanity check
Add a brief sanity test of S32C1I functionality.  This instruction
is needed by the kernel and userland as part of the base ABI
(including GCC atomic builtins, certain threading packages, future
atomic support in the C++ standard, etc).  However, correct operation
of this instruction requires some cooperation by hardware external to
the processor (such as bus bridge, bus fabric, or memory controller).
Minimally exercising this mechanism and reporting explicit status
early in the boot process is helpful to chip vendors using the Linux
kernel as a benchmark of correctness of hardware.

As it turns out, S32C1I is not exercised by the kernel and by uClibc
based userland as of early June 2008.  This is expected to change
soon as both incorporate more recent open source developments.

Signed-off-by: Marc Gauthier <marc@tensilica.com>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:22 -08:00
Max Filippov
28570e8dac xtensa: add trap_set_handler function
trap_set_handler sets new C-handler in the exception table and returns
previous handler.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:22 -08:00
Max Filippov
c622b29d1f xtensa: initialize atomctl SR
In order to use S32C1I instruction on cores with ATOMCTL SR the register
must be properly initialized.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:22 -08:00
Max Filippov
733536b865 xtensa: save and restore scompare1 SR on kernel entry
Although scompare1 may be saved/restored by xchal_ncp_{load,store}
macros, explicit save/restore of registers manipulated by the kernel
itself is considered more correct.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:22 -08:00
Max Filippov
2f6ea6a767 xtensa: display s32c1i feature flag in cpuinfo
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:21 -08:00
Max Filippov
415217efc1 xtensa: fix CPU cache flags formatting
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:21 -08:00
Max Filippov
288dc2b68c xtensa: properly fix missing compiler barrier in simcall
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:21 -08:00
Max Filippov
382cb5b917 xtensa: fix build warning for arch/xtensa/mm/tlb.c
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:21 -08:00
Max Filippov
35b16a9a09 xtensa: provide DMA_ERROR_CODE definition
This fixes the following allmodconfig build error:

drivers/uio/uio_dmem_genirq.c:95:18: error: 'DMA_ERROR_CODE' undeclared (first use in this function)
drivers/uio/uio_dmem_genirq.c:238:18: error: 'DMA_ERROR_CODE' undeclared (first use in this function)
make[3]: *** [drivers/uio/uio_dmem_genirq.o] Error 1

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:21 -08:00
Max Filippov
94d6c61b97 xtensa: ISS: add BASE_BAUD definition to serial.h
This fixes the following build error in allyesconfig:

drivers/tty/serial/8250/8250_early.c: In function 'parse_options':
drivers/tty/serial/8250/8250_early.c:160:18: error: 'BASE_BAUD' undeclared (first use in this function)

Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:21 -08:00
Chris Zankel
d1538c4675 xtensa: provide proper assembler function boundaries with ENDPROC()
Use ENDPROC() to mark the end of assembler functions.

Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:20 -08:00
Max Filippov
c0226e34a4 xtensa: make DoubleExceptionVector literals fit the gap
Manually load references to exc_table from the explicit literal in order
to fit DoubleExceptionVector.literals into the available 16-byte gap
before DoubleExceptionVector.text in the absence of link time
relaxation. Without this fix DoubleExceptionVector.literal section
overlaps DoubleExceptionVector.text section in the linked vmlinux image.

Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:20 -08:00
Chris Zankel
6550162200 xtensa: add config option to disable linker relaxation
The default linker behavior is to optimize identical literal values and
remove unnecessary overhead from assembler-generated "longcall" sequences
to reduce code size. Provide an option to disable this behavior to improve
compile time.

Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:20 -08:00
Nicolas Kaiser
02b25d811f xtensa: unbalanced parentheses
Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:20 -08:00
Wanlong Gao
09378d7c21 xtensa:fix the incompatible pointer type warning in time.c
Fix the definition of the function ccount_read to be compatible
to the member read of the structure clocksource.

Signed-off-by: Wanlong Gao <wanlong.gao@gmail.com>
Signed-off-by: Chris Zankel <chris@zankel.net>
2012-12-18 21:10:20 -08:00
Michael Spang
d5c1b541ba ARM: SAMSUNG: Add missing include guard to gpio-core.h
Signed-off-by: Michael Spang <spang@chromium.org>
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
2012-12-18 19:00:25 -08:00
Axel Lin
d5fd5da290 pinctrl: exynos5440/samsung: Staticize pcfgs
I got below build error with random config if CONFIG_PINCTRL_SAMSUNG=y &&
CONFIG_PINCTRL_EXYNOS5440=y.

Fix the build error by making pcfgs static.

  LD      drivers/pinctrl/built-in.o
drivers/pinctrl/pinctrl-exynos5440.o: In function `.LANCHOR0':
pinctrl-exynos5440.c:(.data+0x54): multiple definition of `pcfgs'
drivers/pinctrl/pinctrl-samsung.o:pinctrl-samsung.c:(.data+0x54): first defined here
make[2]: *** [drivers/pinctrl/built-in.o] Error 1
make[1]: *** [drivers/pinctrl] Error 2
make: *** [drivers] Error 2

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Thomas Abraham <thomas.abraham@linaro.org>
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
2012-12-18 19:00:25 -08:00
Sachin Kamat
88f2324fa9 pinctrl: samsung: Fix a typo in pinctrl-samsung.h
struct samsung_pin_bank does not have a member called reg_offset.
It should be pctl_offset instead.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
2012-12-18 19:00:25 -08:00
Kukjin Kim
83e877a334 ARM: EXYNOS: fix skip scu_enable() for EXYNOS5440
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
2012-12-18 19:00:25 -08:00
Kukjin Kim
12fee1949e ARM: EXYNOS: fix GIC using for EXYNOS5440
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
2012-12-18 19:00:25 -08:00
Kukjin Kim
ade7ffbb9b ARM: EXYNOS: fix build error when MFC is not selected
This fixes following:
arch/arm/mach-exynos/built-in.o: In function `exynos5_reserve':
arch/arm/mach-exynos/mach-exynos5-dt.c:177: undefined reference to `s5p_fdt_find_mfc_mem'
arch/arm/mach-exynos/mach-exynos5-dt.c:177: undefined reference to `s5p_fdt_find_mfc_mem'
arch/arm/mach-exynos/mach-exynos5-dt.c:178: undefined reference to `s5p_mfc_reserve_mem'

Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
2012-12-18 19:00:25 -08:00
Linus Torvalds
752451f01c Merge branch 'i2c-embedded/for-next' of git://git.pengutronix.de/git/wsa/linux
Pull i2c-embedded changes from Wolfram Sang:
 - CBUS driver (an I2C variant)
 - continued rework of the omap driver
 - s3c2410 gets lots of fixes and gains pinctrl support
 - at91 gains DMA support
 - the GPIO muxer gains devicetree probing
 - typical fixes and additions all over

* 'i2c-embedded/for-next' of git://git.pengutronix.de/git/wsa/linux: (45 commits)
  i2c: omap: Remove the OMAP_I2C_FLAG_RESET_REGS_POSTIDLE flag
  i2c: at91: add dma support
  i2c: at91: change struct members indentation
  i2c: at91: fix compilation warning
  i2c: mxs: Do not disable the I2C SMBus quick mode
  i2c: mxs: Handle i2c DMA failure properly
  i2c: s3c2410: Remove recently introduced performance overheads
  i2c: ocores: Move grlib set/get functions into #ifdef CONFIG_OF block
  i2c: s3c2410: Add fix for i2c suspend/resume
  i2c: s3c2410: Fix code to free gpios
  i2c: i2c-cbus-gpio: introduce driver
  i2c: ocores: Add support for the GRLIB port of the controller and use function pointers for getreg and setreg functions
  i2c: ocores: Add irq support for sparc
  i2c: omap: Move the remove constraint
  ARM: dts: cfa10049: Add the i2c muxer buses to the CFA-10049
  i2c: s3c2410: do not special case HDMIPHY stuck bus detection
  i2c: s3c2410: use exponential back off while polling for bus idle
  i2c: s3c2410: do not generate STOP for QUIRK_HDMIPHY
  i2c: s3c2410: grab adapter lock while changing i2c clock
  i2c: s3c2410: Add support for pinctrl
  ...
2012-12-18 16:51:10 -08:00
Chris Mason
57ba86c00f Revert "Btrfs: reorder tree mod log operations in deleting a pointer"
This reverts commit 6a7a665d78.

This was bug was fixed differently in 3.6, so this commit
isn't needed.

Conflicts:
	fs/btrfs/ctree.c

Signed-off-by: Chris Mason <chris.mason@fusionio.com>
2012-12-18 19:35:32 -05:00
Shawn Guo
c1e37ea287 net: fec: forbid FEC_PTP on SoCs that do not support
Beside imx6q, the kernel built from imx_v6_v7_defconfig is also
supposed to be running on other IMX SoCs that do not have the PTP
block.  Before fec driver gets fixed to run-time detect target hardware
rather than conditional compiling with #ifdef CONFIG_FEC_PTP, let's
give it a quick fix in Kconfig to forbid FEC_PTP on those IMX SoCs that
do not support PTP.

Reported-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2012-12-18 16:26:18 -08:00
Sathya Perla
d23e946cb6 be2net: fix wrong frag_idx reported by RX CQ
The RX CQ can report completions with invalid frag_idx when the RXQ that
was *previously* using it, was not cleaned up properly. This hits
a BUG_ON() in be2net.

When completion coalescing is enabled on a CQ, an explicit CQ-notify
(with rearm) is needed for each compl, to flush partially coalesced CQ
entries that are pending DMA.

In be_close(), this fix now notifies CQ for each compl, waits explicitly
for the flush compl to arrive and complains if it doesn't arrive.

Also renaming be_crit_error() to be_hw_error() as it's the more
appropriate name and to convey that we don't wait for the flush compl
only when a HW error has occurred.

Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2012-12-18 16:18:39 -08:00
Sathya Perla
a323d9bf83 be2net: fix be_close() to ensure all events are ack'ed
In be_close(), be_eq_clean() must be called after all RX/TX/MCC queues
have been cleaned to ensure that any events caused while cleaning up
completions are notified/acked. Not clearing all events can cause
upredictable behaviour when RX rings are re-created in the subsequent
be_open().

Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2012-12-18 16:18:39 -08:00
David S. Miller
4a9d1946b0 sparc64: Define pte_accessible()
We can elide flush_tlb_*() calls when _PAGE_VALID is clear
as that is the test used to determine whether or not to
queue up a TLB flush in set_pte_at().

Signed-off-by: David S. Miller <davem@davemloft.net>
2012-12-18 16:06:16 -08:00
Dave Kleikamp
6cb9c36975 sparc: huge_ptep_set_* functions need to call set_huge_pte_at()
Modifying the huge pte's requires that all the underlying pte's be
modified.

Version 2: added missing flush_tlb_page()

Signed-off-by: Dave Kleikamp <dave.kleikamp@oracle.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: sparclinux@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
2012-12-18 15:36:20 -08:00
Linus Torvalds
673ab8783b Merge branch 'akpm' (more patches from Andrew)
Merge patches from Andrew Morton:
 "Most of the rest of MM, plus a few dribs and drabs.

  I still have quite a few irritating patches left around: ones with
  dubious testing results, lack of review, ones which should have gone
  via maintainer trees but the maintainers are slack, etc.

  I need to be more activist in getting these things wrapped up outside
  the merge window, but they're such a PITA."

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (48 commits)
  mm/vmscan.c: avoid possible deadlock caused by too_many_isolated()
  vmscan: comment too_many_isolated()
  mm/kmemleak.c: remove obsolete simple_strtoul
  mm/memory_hotplug.c: improve comments
  mm/hugetlb: create hugetlb cgroup file in hugetlb_init
  mm/mprotect.c: coding-style cleanups
  Documentation: ABI: /sys/devices/system/node/
  slub: drop mutex before deleting sysfs entry
  memcg: add comments clarifying aspects of cache attribute propagation
  kmem: add slab-specific documentation about the kmem controller
  slub: slub-specific propagation changes
  slab: propagate tunable values
  memcg: aggregate memcg cache values in slabinfo
  memcg/sl[au]b: shrink dead caches
  memcg/sl[au]b: track all the memcg children of a kmem_cache
  memcg: destroy memcg caches
  sl[au]b: allocate objects from memcg cache
  sl[au]b: always get the cache from its page in kmem_cache_free()
  memcg: skip memcg kmem allocations in specified code regions
  memcg: infrastructure to match an allocation to the right cache
  ...
2012-12-18 15:08:12 -08:00
Linus Torvalds
d7b96ca5d0 Fix fallout from __devexit removal
-----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1.4.11 (GNU/Linux)
 
 iQIcBAABAgAGBQJQzoH6AAoJEMsfJm/On5mBk8MQAJgBifZmikVrhleYIaCW7p4g
 MwyL3+2TFE1Hdgr8j2cblaOm4rfSbmAAOrPHcZQHY6lAeEJpwClce1SzyAL7C97T
 GLYfPegUjN1Itvs+jL3dEv9Y9xmIcIvSG18IScjLhaazIYz8sWprsGs62BVX9BnC
 UZMZBHqSpyOuh2NZSJSGj2J3ZGkDBeHI82N2FidLP0Xh2bki/Q/H4AY+f3+oomLC
 DTXiF9URH2qbS0oSpKgpfVzxA2sS4JF/i70AvBLYFlG5YjtGOQBtPxoPHNAgBdLH
 vZ7a/hEdExhqP/2l7m7yRaLVwCKi5o/Mf8xCFpygNu/vo66u2sw7Mo3TjFQ18iVl
 9hDwTz1Uq7TlBsORHx6cngHIq8V6s6dxt1zOe68txjdH82LFEJEB4ybRdXoAPHc/
 GRuCe2RnYTmMLwRnVmAuNTfq32fXPu6lq4zwFrK5jxMSMBkqsQWtBkJRKOy09vGb
 mp71PUnv5N1/SEcS4+ddjKnMQ/WBl4pB5qkkGer/vSxeQSWbyfaLfruR0KdEBhnY
 bK/JFZVN9jII040yk3gxQmOmZQLnsHy6U8mOVDAeDAIII2gpPL2X+IQOmmnHY47g
 pUF6zh70eRMD37+OzB4VuxmR9zS7PDGWPO3i1VuAGLN0FNdD1b9Hhdnmmf1xQM7a
 7gQL4V/S3epB3GtvXJO7
 =uh1S
 -----END PGP SIGNATURE-----

Merge tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging

Pull hwmon fixlet from Guenter Roeck:
 "Fix fallout from __devexit removal"

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (twl4030-madc-hwmon) Fix warning message caused by removal of __devexit
2012-12-18 15:06:11 -08:00
Linus Torvalds
5b3040a48b Merge git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile
Pull tile updates from Chris Metcalf:
 "These are a smattering of minor changes from Tilera and other folks,
  mostly in the ptrace area."

* git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile:
  arch/tile: set CORE_DUMP_USE_REGSET on tile
  arch/tile: implement arch_ptrace using user_regset on tile
  arch/tile: implement user_regset interface on tile
  arch/tile: clean up tile-specific PTRACE_SETOPTIONS
  arch/tile: provide PT_FLAGS_COMPAT value in pt_regs
  tile/PCI: use for_each_pci_dev to simplify the code
  tilegx: remove __init from pci fixup hook
2012-12-18 15:05:30 -08:00
Fengguang Wu
3cf23841b4 mm/vmscan.c: avoid possible deadlock caused by too_many_isolated()
Neil found that if too_many_isolated() returns true while performing
direct reclaim we can end up waiting for other threads to complete their
direct reclaim.  If those threads are allowed to enter the FS or IO to
free memory, but this thread is not, then it is possible that those
threads will be waiting on this thread and so we get a circular deadlock.

some task enters direct reclaim with GFP_KERNEL
  => too_many_isolated() false
    => vmscan and run into dirty pages
      => pageout()
        => take some FS lock
          => fs/block code does GFP_NOIO allocation
            => enter direct reclaim again
              => too_many_isolated() true
                => waiting for others to progress, however the other
                   tasks may be circular waiting for the FS lock..

The fix is to let !__GFP_IO and !__GFP_FS direct reclaims enjoy higher
priority than normal ones, by lowering the throttle threshold for the
latter.

Allowing ~1/8 isolated pages in normal is large enough.  For example, for
a 1GB LRU list, that's ~128MB isolated pages, or 1k blocked tasks (each
isolates 32 4KB pages), or 64 blocked tasks per logical CPU (assuming 16
logical CPUs per NUMA node).  So it's not likely some CPU goes idle
waiting (when it could make progress) because of this limit: there are
much more sleeping reclaim tasks than the number of CPU, so the task may
well be blocked by some low level queue/lock anyway.

Now !GFP_IOFS reclaims won't be waiting for GFP_IOFS reclaims to progress.
 They will be blocked only when there are too many concurrent !GFP_IOFS
reclaims, however that's very unlikely because the IO-less direct reclaims
is able to progress much more faster, and they won't deadlock each other.
The threshold is raised high enough for them, so that there can be
sufficient parallel progress of !GFP_IOFS reclaims.

[akpm@linux-foundation.org: tweak comment]
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: Torsten Kaiser <just.for.lkml@googlemail.com>
Tested-by: NeilBrown <neilb@suse.de>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 15:02:15 -08:00
Fengguang Wu
d37dd5dcb9 vmscan: comment too_many_isolated()
Comment "Why it's doing so" rather than "What it does" as proposed by
Andrew Morton.

Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 15:02:15 -08:00
Abhijit Pawar
dc053733ea mm/kmemleak.c: remove obsolete simple_strtoul
Replace the obsolete simple_strtoul() with kstrtoul().

Signed-off-by: Abhijit Pawar <abhi.c.pawar@gmail.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 15:02:15 -08:00
Tang Chen
79a4dcefd3 mm/memory_hotplug.c: improve comments
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 15:02:15 -08:00