linux-2.6-microblaze.git
9 years agoMerge branch 'kconfig' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild
Linus Torvalds [Thu, 19 Feb 2015 18:36:45 +0000 (10:36 -0800)]
Merge branch 'kconfig' of git://git./linux/kernel/git/mmarek/kbuild

Pull kconfig updates from Michal Marek:
 "Yann E Morin was supposed to take over kconfig maintainership, but
  this hasn't happened.  So I'm sending a few kconfig patches that I
  collected:

   - Fix for missing va_end in kconfig
   - merge_config.sh displays used if given too few arguments
   - s/boolean/bool/ in Kconfig files for consistency, with the plan to
     only support bool in the future"

* 'kconfig' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
  kconfig: use va_end to match corresponding va_start
  merge_config.sh: Display usage if given too few arguments
  kconfig: use bool instead of boolean for type definition attributes

9 years agoMerge branch 'misc' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild
Linus Torvalds [Thu, 19 Feb 2015 18:31:37 +0000 (10:31 -0800)]
Merge branch 'misc' of git://git./linux/kernel/git/mmarek/kbuild

Pull misc kbuild changes from Michal Marek:
 "Just a few non-critical kbuild changes:

   - builddeb adds the actual distribution name in the changelog
   - documentation fixes"

* 'misc' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
  kbuild: trivial - fix the help doc of CONFIG_CC_OPTIMIZE_FOR_SIZE
  kbuild: Update documentation of clean-files and clean-dirs
  builddeb: Try to determine distribution
  builddeb: Update year and git repository URL in debian/copyright

9 years agoMAINTAINERS: update Ceph and RBD maintainers
Sage Weil [Thu, 19 Feb 2015 18:10:40 +0000 (10:10 -0800)]
MAINTAINERS: update Ceph and RBD maintainers

- add Ilya, drop Yehuda as an RBD maintainer
- add Zheng as a Ceph maintainer
- update Yehuda and Sage's emails

Signed-off-by: Sage Weil <sage@redhat.com>
9 years agoMerge branch 'kbuild' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild
Linus Torvalds [Thu, 19 Feb 2015 18:07:08 +0000 (10:07 -0800)]
Merge branch 'kbuild' of git://git./linux/kernel/git/mmarek/kbuild

Pull kbuild updates from Michal Marek:

 - several cleanups in kbuild

 - serialize multiple *config targets so that 'make defconfig kvmconfig'
   works

 - The cc-ifversion macro got support for an else-branch

* 'kbuild' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
  kbuild,gcov: simplify kernel/gcov/Makefile more
  kbuild: allow cc-ifversion to have the argument for false condition
  kbuild,gcov: simplify kernel/gcov/Makefile
  kbuild,gcov: remove unnecessary workaround
  kbuild: do not add $(call ...) to invoke cc-version or cc-fullversion
  kbuild: fix cc-ifversion macro
  kbuild: drop $(version_h) from MRPROPER_FILES
  kbuild: use mixed-targets when two or more config targets are given
  kbuild: remove redundant line from bounds.h/asm-offsets.h
  kbuild: merge bounds.h and asm-offsets.h rules
  kbuild: Drop support for clean-rule

9 years agos390/spinlock: disabled compare-and-delay by default
Martin Schwidefsky [Thu, 19 Feb 2015 16:53:16 +0000 (17:53 +0100)]
s390/spinlock: disabled compare-and-delay by default

Until we have hard performance data about the effects of CAD in the
spinlock loop disable the instruction by default.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
9 years agoi2c: designware-baytrail: another fixup for proper Kconfig dependencies
Wolfram Sang [Thu, 19 Feb 2015 16:02:59 +0000 (17:02 +0100)]
i2c: designware-baytrail: another fixup for proper Kconfig dependencies

IOSF_MBI is tristate. Baytrail driver isn't.

Reported-by: Randy Dunlap <rdunlap@infradead.org>
Acked-by: David E. Box <david.e.box@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
9 years agoi2c: fix reference to functionality constants definition
Baruch Siach [Thu, 19 Feb 2015 10:51:30 +0000 (12:51 +0200)]
i2c: fix reference to functionality constants definition

Since commit 607ca46e97 ('UAPI: (Scripted) Disintegrate include/linux') the
list of functionality constants moved to include/uapi/linux/i2c.h. Update the
reference accordingly.

Fixes: 607ca46e97 ('UAPI: (Scripted) Disintegrate include/linux')
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
9 years agoMerge branch 'mipsr6-for-3.20' of git://git.linux-mips.org/pub/scm/mchandras/linux...
Ralf Baechle [Thu, 19 Feb 2015 15:00:34 +0000 (16:00 +0100)]
Merge branch 'mipsr6-for-3.20' of git://git.linux-mips.org/mchandras/linux into mips-for-linux-next

9 years agoMIPS: Export MSA functions used by lose_fpu(1) for KVM
James Hogan [Tue, 10 Feb 2015 10:03:00 +0000 (10:03 +0000)]
MIPS: Export MSA functions used by lose_fpu(1) for KVM

Export the _save_msa asm function used by the lose_fpu(1) macro to GPL
modules so that KVM can make use of it when it is built as a module.

This fixes the following build error when CONFIG_KVM=m and
CONFIG_CPU_HAS_MSA=y due to commit f798217dfd03 ("KVM: MIPS: Don't leak
FPU/DSP to guest"):

ERROR: "_save_msa" [arch/mips/kvm/kvm.ko] undefined!

Fixes: f798217dfd03 (KVM: MIPS: Don't leak FPU/DSP to guest)
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Gleb Natapov <gleb@kernel.org>
Cc: kvm@vger.kernel.org
Cc: linux-mips@linux-mips.org
Cc: <stable@vger.kernel.org> # 3.15+
Patchwork: https://patchwork.linux-mips.org/patch/9261/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Export FP functions used by lose_fpu(1) for KVM
James Hogan [Tue, 10 Feb 2015 10:02:59 +0000 (10:02 +0000)]
MIPS: Export FP functions used by lose_fpu(1) for KVM

Export the _save_fp asm function used by the lose_fpu(1) macro to GPL
modules so that KVM can make use of it when it is built as a module.

This fixes the following build error when CONFIG_KVM=m due to commit
f798217dfd03 ("KVM: MIPS: Don't leak FPU/DSP to guest"):

ERROR: "_save_fp" [arch/mips/kvm/kvm.ko] undefined!

Signed-off-by: James Hogan <james.hogan@imgtec.com>
Fixes: f798217dfd03 (KVM: MIPS: Don't leak FPU/DSP to guest)
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Gleb Natapov <gleb@kernel.org>
Cc: kvm@vger.kernel.org
Cc: linux-mips@linux-mips.org
Cc: <stable@vger.kernel.org> # 3.10+
Patchwork: https://patchwork.linux-mips.org/patch/9260/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: BCM3384: Fix outdated use of mips_cpu_intc_init()
Kevin Cernekee [Thu, 25 Dec 2014 17:48:56 +0000 (09:48 -0800)]
MIPS: BCM3384: Fix outdated use of mips_cpu_intc_init()

This function was renamed to mips_cpu_irq_of_init(), so fix it to avoid
a compile error.

Signed-off-by: Kevin Cernekee <cernekee@gmail.com>
Cc: f.fainelli@gmail.com
Cc: jaedon.shin@gmail.com
Cc: abrestic@chromium.org
Cc: tglx@linutronix.de
Cc: jason@lakedaemon.net
Cc: jogo@openwrt.org
Cc: arnd@arndb.de
Cc: computersforpeace@gmail.com
Cc: linux-mips@linux-mips.org
Cc: devicetree@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/8834/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Provide correct siginfo_t.si_stime
Petr Malat [Fri, 12 Dec 2014 14:28:01 +0000 (15:28 +0100)]
MIPS: Provide correct siginfo_t.si_stime

Provide correct siginfo_t.si_stime on MIPS64

Bug description:
MIPS version of copy_siginfo() is not aware of alignment on platforms with
64-bit long integers, which leads to an incorrect si_stime passed to signal
handlers, because the last element (si_stime) of _sifields._sigchld is not
copied. If _MIPS_SZLONG is 64, then the _sifields starts at the offset of
4 * sizeof(int).

Patch description:
Use the generic copy_siginfo, which doesn't have this problem.

Signed-off-by: Petr Malat <oss@malat.biz>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/8671/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Makefile: Move the ASEs checks after setting the core's CFLAGS
Markos Chandras [Mon, 2 Feb 2015 15:41:01 +0000 (15:41 +0000)]
MIPS: Makefile: Move the ASEs checks after setting the core's CFLAGS

We need to check the ASEs support against the core's CFLAGS instead
of depending to the default -march option from the toolchain.

Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Cc: Maciej W. Rozycki <macro@linux-mips.org>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/9180/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMIPS: Makefile: Pass -march option on Loongson3A cores
Ralf Baechle [Thu, 19 Feb 2015 12:47:20 +0000 (13:47 +0100)]
MIPS: Makefile: Pass -march option on Loongson3A cores

The loongson 3A cores do not select a suitable -march option so the build
system uses the default one from the toolchain. This may or may not be
suitable for a loongson 3A build. In order to avoid that, we explicitly set
a suitable -march option for that core.  Furthermore, some very old
compilers don't support -march= at all and there is the possibility of
toolchain combinations such as GCC 4.9 and binutils 2.24 for which
-march=loongson3a will result in MIPS64 R2 code being generated but then
rejected by GAS.  So treat the Longsoon 3A as an R2 CPU.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
9 years agoMerge tag 'microcode_fixes_for-3.21' of git://git.kernel.org/pub/scm/linux/kernel...
Ingo Molnar [Thu, 19 Feb 2015 12:32:42 +0000 (13:32 +0100)]
Merge tag 'microcode_fixes_for-3.21' of git://git./linux/kernel/git/bp/bp into x86/urgent

Pull microcode fixes from Borislav Petkov:

  - Two fixes hardening microcode data handling. (Quentin Casasnovas)

Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agox86/microcode/intel: Handle truncated microcode images more robustly
Quentin Casasnovas [Tue, 3 Feb 2015 12:00:24 +0000 (13:00 +0100)]
x86/microcode/intel: Handle truncated microcode images more robustly

We do not check the input data bounds containing the microcode before
copying a struct microcode_intel_header from it. A specially crafted
microcode could cause the kernel to read invalid memory and lead to a
denial-of-service.

Signed-off-by: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Link: http://lkml.kernel.org/r/1422964824-22056-3-git-send-email-quentin.casasnovas@oracle.com
[ Made error message differ from the next one and flipped comparison. ]
Signed-off-by: Borislav Petkov <bp@suse.de>
9 years agox86/microcode/intel: Guard against stack overflow in the loader
Quentin Casasnovas [Tue, 3 Feb 2015 12:00:22 +0000 (13:00 +0100)]
x86/microcode/intel: Guard against stack overflow in the loader

mc_saved_tmp is a static array allocated on the stack, we need to make
sure mc_saved_count stays within its bounds, otherwise we're overflowing
the stack in _save_mc(). A specially crafted microcode header could lead
to a kernel crash or potentially kernel execution.

Signed-off-by: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Link: http://lkml.kernel.org/r/1422964824-22056-1-git-send-email-quentin.casasnovas@oracle.com
Signed-off-by: Borislav Petkov <bp@suse.de>
9 years agoMerge branch 'tip-x86-kaslr' of git://git.kernel.org/pub/scm/linux/kernel/git/bp...
Ingo Molnar [Thu, 19 Feb 2015 11:06:04 +0000 (12:06 +0100)]
Merge branch 'tip-x86-kaslr' of git://git./linux/kernel/git/bp/bp into x86/urgent

Pull ASLR and kASLR fixes from Borislav Petkov:

  - Add a global flag announcing KASLR state so that relevant code can do
    informed decisions based on its setting. (Jiri Kosina)

  - Fix a stack randomization entropy decrease bug. (Hector Marco-Gisbert)

Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agolibceph: kfree() in put_osd() shouldn't depend on authorizer
Ilya Dryomov [Mon, 16 Feb 2015 08:49:42 +0000 (11:49 +0300)]
libceph: kfree() in put_osd() shouldn't depend on authorizer

a255651d4cad ("ceph: ensure auth ops are defined before use") made
kfree() in put_osd() conditional on the authorizer.  A mechanical
mistake most likely - fix it.

Cc: Alex Elder <elder@linaro.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Alex Elder <elder@linaro.org>
9 years agolibceph: fix double __remove_osd() problem
Ilya Dryomov [Tue, 17 Feb 2015 16:37:15 +0000 (19:37 +0300)]
libceph: fix double __remove_osd() problem

It turns out it's possible to get __remove_osd() called twice on the
same OSD.  That doesn't sit well with rb_erase() - depending on the
shape of the tree we can get a NULL dereference, a soft lockup or
a random crash at some point in the future as we end up touching freed
memory.  One scenario that I was able to reproduce is as follows:

            <osd3 is idle, on the osd lru list>
<con reset - osd3>
con_fault_finish()
  osd_reset()
                              <osdmap - osd3 down>
                              ceph_osdc_handle_map()
                                <takes map_sem>
                                kick_requests()
                                  <takes request_mutex>
                                  reset_changed_osds()
                                    __reset_osd()
                                      __remove_osd()
                                  <releases request_mutex>
                                <releases map_sem>
    <takes map_sem>
    <takes request_mutex>
    __kick_osd_requests()
      __reset_osd()
        __remove_osd() <-- !!!

A case can be made that osd refcounting is imperfect and reworking it
would be a proper resolution, but for now Sage and I decided to fix
this by adding a safe guard around __remove_osd().

Fixes: http://tracker.ceph.com/issues/8087

Cc: Sage Weil <sage@redhat.com>
Cc: stable@vger.kernel.org # 3.9+: 7c6e6fc53e73: libceph: assert both regular and lingering lists in __remove_osd()
Cc: stable@vger.kernel.org # 3.9+: cc9f1f518cec: libceph: change from BUG to WARN for __remove_osd() asserts
Cc: stable@vger.kernel.org # 3.9+
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Alex Elder <elder@linaro.org>
9 years agorbd: convert to blk-mq
Christoph Hellwig [Tue, 13 Jan 2015 16:20:04 +0000 (17:20 +0100)]
rbd: convert to blk-mq

This converts the rbd driver to use the blk-mq infrastructure.  Except
for switching to a per-request work item this is almost mechanical.

This was tested by Alexandre DERUMIER in November, and found to give
him 120000 iops, although the only comparism available was an old
3.10 kernel which gave 80000iops.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <elder@linaro.org>
[idryomov@gmail.com: context, blk_mq_init_queue() EH]
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
9 years agox86, mm/ASLR: Fix stack randomization on 64-bit systems
Hector Marco-Gisbert [Sat, 14 Feb 2015 17:33:50 +0000 (09:33 -0800)]
x86, mm/ASLR: Fix stack randomization on 64-bit systems

The issue is that the stack for processes is not properly randomized on
64 bit architectures due to an integer overflow.

The affected function is randomize_stack_top() in file
"fs/binfmt_elf.c":

  static unsigned long randomize_stack_top(unsigned long stack_top)
  {
           unsigned int random_variable = 0;

           if ((current->flags & PF_RANDOMIZE) &&
                   !(current->personality & ADDR_NO_RANDOMIZE)) {
                   random_variable = get_random_int() & STACK_RND_MASK;
                   random_variable <<= PAGE_SHIFT;
           }
           return PAGE_ALIGN(stack_top) + random_variable;
           return PAGE_ALIGN(stack_top) - random_variable;
  }

Note that, it declares the "random_variable" variable as "unsigned int".
Since the result of the shifting operation between STACK_RND_MASK (which
is 0x3fffff on x86_64, 22 bits) and PAGE_SHIFT (which is 12 on x86_64):

  random_variable <<= PAGE_SHIFT;

then the two leftmost bits are dropped when storing the result in the
"random_variable". This variable shall be at least 34 bits long to hold
the (22+12) result.

These two dropped bits have an impact on the entropy of process stack.
Concretely, the total stack entropy is reduced by four: from 2^28 to
2^30 (One fourth of expected entropy).

This patch restores back the entropy by correcting the types involved
in the operations in the functions randomize_stack_top() and
stack_maxrandom_size().

The successful fix can be tested with:

  $ for i in `seq 1 10`; do cat /proc/self/maps | grep stack; done
  7ffeda566000-7ffeda587000 rw-p 00000000 00:00 0                          [stack]
  7fff5a332000-7fff5a353000 rw-p 00000000 00:00 0                          [stack]
  7ffcdb7a1000-7ffcdb7c2000 rw-p 00000000 00:00 0                          [stack]
  7ffd5e2c4000-7ffd5e2e5000 rw-p 00000000 00:00 0                          [stack]
  ...

Once corrected, the leading bytes should be between 7ffc and 7fff,
rather than always being 7fff.

Signed-off-by: Hector Marco-Gisbert <hecmargi@upv.es>
Signed-off-by: Ismael Ripoll <iripoll@upv.es>
[ Rebased, fixed 80 char bugs, cleaned up commit message, added test example and CVE ]
Signed-off-by: Kees Cook <keescook@chromium.org>
Cc: <stable@vger.kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Fixes: CVE-2015-1593
Link: http://lkml.kernel.org/r/20150214173350.GA18393@www.outflux.net
Signed-off-by: Borislav Petkov <bp@suse.de>
9 years agoMerge branch 'tip-x86-mm' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp...
Ingo Molnar [Thu, 19 Feb 2015 10:59:18 +0000 (11:59 +0100)]
Merge branch 'tip-x86-mm' of git://git./linux/kernel/git/bp/bp into x86/urgent

Pull boot printout fix from Borislav Petkov.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agox86/mm/init: Fix incorrect page size in init_memory_mapping() printks
Dave Hansen [Tue, 10 Feb 2015 21:20:30 +0000 (13:20 -0800)]
x86/mm/init: Fix incorrect page size in init_memory_mapping() printks

With 32-bit non-PAE kernels, we have 2 page sizes available
(at most): 4k and 4M.

Enabling PAE replaces that 4M size with a 2M one (which 64-bit
systems use too).

But, when booting a 32-bit non-PAE kernel, in one of our
early-boot printouts, we say:

  init_memory_mapping: [mem 0x00000000-0x000fffff]
   [mem 0x00000000-0x000fffff] page 4k
  init_memory_mapping: [mem 0x37000000-0x373fffff]
   [mem 0x37000000-0x373fffff] page 2M
  init_memory_mapping: [mem 0x00100000-0x36ffffff]
   [mem 0x00100000-0x003fffff] page 4k
   [mem 0x00400000-0x36ffffff] page 2M
  init_memory_mapping: [mem 0x37400000-0x377fdfff]
   [mem 0x37400000-0x377fdfff] page 4k

Which is obviously wrong.  There is no 2M page available.  This
is probably because of a badly-named variable: in the map_range
code: PG_LEVEL_2M.

Instead of renaming all the PG_LEVEL_2M's.  This patch just
fixes the printout:

  init_memory_mapping: [mem 0x00000000-0x000fffff]
   [mem 0x00000000-0x000fffff] page 4k
  init_memory_mapping: [mem 0x37000000-0x373fffff]
   [mem 0x37000000-0x373fffff] page 4M
  init_memory_mapping: [mem 0x00100000-0x36ffffff]
   [mem 0x00100000-0x003fffff] page 4k
   [mem 0x00400000-0x36ffffff] page 4M
  init_memory_mapping: [mem 0x37400000-0x377fdfff]
   [mem 0x37400000-0x377fdfff] page 4k
  BRK [0x03206000, 0x03206fff] PGTABLE

Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Yinghai Lu <yinghai@kernel.org>
Link: http://lkml.kernel.org/r/20150210212030.665EC267@viggo.jf.intel.com
Signed-off-by: Borislav Petkov <bp@suse.de>
9 years agox86/mm/ASLR: Propagate base load address calculation
Jiri Kosina [Fri, 13 Feb 2015 15:04:55 +0000 (16:04 +0100)]
x86/mm/ASLR: Propagate base load address calculation

Commit:

  e2b32e678513 ("x86, kaslr: randomize module base load address")

makes the base address for module to be unconditionally randomized in
case when CONFIG_RANDOMIZE_BASE is defined and "nokaslr" option isn't
present on the commandline.

This is not consistent with how choose_kernel_location() decides whether
it will randomize kernel load base.

Namely, CONFIG_HIBERNATION disables kASLR (unless "kaslr" option is
explicitly specified on kernel commandline), which makes the state space
larger than what module loader is looking at. IOW CONFIG_HIBERNATION &&
CONFIG_RANDOMIZE_BASE is a valid config option, kASLR wouldn't be applied
by default in that case, but module loader is not aware of that.

Instead of fixing the logic in module.c, this patch takes more generic
aproach. It introduces a new bootparam setup data_type SETUP_KASLR and
uses that to pass the information whether kaslr has been applied during
kernel decompression, and sets a global 'kaslr_enabled' variable
accordingly, so that any kernel code (module loading, livepatching, ...)
can make decisions based on its value.

x86 module loader is converted to make use of this flag.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Acked-by: Kees Cook <keescook@chromium.org>
Cc: "H. Peter Anvin" <hpa@linux.intel.com>
Link: https://lkml.kernel.org/r/alpine.LNX.2.00.1502101411280.10719@pobox.suse.cz
[ Always dump correct kaslr status when panicking ]
Signed-off-by: Borislav Petkov <bp@suse.de>
9 years agoceph: return error for traceless reply race
Yan, Zheng [Wed, 4 Feb 2015 07:10:48 +0000 (15:10 +0800)]
ceph: return error for traceless reply race

When we receives traceless reply for request that created new inode,
we re-send a lookup request to MDS get information of the newly created
inode. (VFS expects FS' callback return an inode in create case)
This breaks one request into two requests. Other client may modify or
move to the new inode in the middle.

When the race happens, ceph_handle_notrace_create() unconditionally
links the dentry for 'create' operation to the inode returned by lookup.
This may confuse VFS when the inode is a directory (VFS does not allow
multiple linkages for directory inode).

This patch makes ceph_handle_notrace_create() when it detect a race.
This event should be rare and it happens only when we talk to old MDS.
Recent MDS does not send traceless reply for request that creates new
inode.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: fix dentry leaks
Yan, Zheng [Mon, 2 Feb 2015 03:27:56 +0000 (11:27 +0800)]
ceph: fix dentry leaks

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: re-send requests when MDS enters reconnecting stage
Yan, Zheng [Wed, 4 Feb 2015 06:26:22 +0000 (14:26 +0800)]
ceph: re-send requests when MDS enters reconnecting stage

So that MDS can check if any request is already completed and process
completed requests in clientreplay stage. When completed requests are
processed in clientreplay stage, MDS can avoid sending traceless
replies.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: show nocephx_require_signatures and notcp_nodelay options
Ilya Dryomov [Mon, 2 Feb 2015 14:54:17 +0000 (17:54 +0300)]
ceph: show nocephx_require_signatures and notcp_nodelay options

Signed-off-by: Ilya Dryomov <idryomov@redhat.com>
9 years agolibceph: tcp_nodelay support
Chaitanya Huilgol [Fri, 23 Jan 2015 11:11:25 +0000 (16:41 +0530)]
libceph: tcp_nodelay support

TCP_NODELAY socket option set on connection sockets,
disables Nagle’s algorithm and improves latency characteristics.
tcp_nodelay(default)/notcp_nodelay option flags provided to
enable/disable setting the socket option.

Signed-off-by: Chaitanya Huilgol <chaitanya.huilgol@sandisk.com>
[idryomov@redhat.com: NO_TCP_NODELAY -> TCP_NODELAY, minor adjustments]
Signed-off-by: Ilya Dryomov <idryomov@redhat.com>
9 years agorbd: do not treat standalone as flatten
Ilya Dryomov [Mon, 19 Jan 2015 19:57:39 +0000 (22:57 +0300)]
rbd: do not treat standalone as flatten

If the clone is resized down to 0, it becomes standalone.  If such
resize is carried over while an image is mapped we would detect this
and call rbd_dev_parent_put() which means "let go of all parent state,
including the spec(s) of parent images(s)".  This leads to a mismatch
between "rbd info" and sysfs parent fields, so a fix is in order.

    # rbd create --image-format 2 --size 1 foo
    # rbd snap create foo@snap
    # rbd snap protect foo@snap
    # rbd clone foo@snap bar
    # DEV=$(rbd map bar)
    # rbd resize --allow-shrink --size 0 bar
    # rbd resize --size 1 bar
    # rbd info bar | grep parent
            parent: rbd/foo@snap

Before:

    # cat /sys/bus/rbd/devices/0/parent
    (no parent image)

After:

    # cat /sys/bus/rbd/devices/0/parent
    pool_id 0
    pool_name rbd
    image_id 10056b8b4567
    image_name foo
    snap_id 2
    snap_name snap
    overlap 0

Signed-off-by: Ilya Dryomov <idryomov@redhat.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
Reviewed-by: Alex Elder <elder@linaro.org>
9 years agoceph: fix atomic_open snapdir
Yan, Zheng [Mon, 19 Jan 2015 05:23:20 +0000 (13:23 +0800)]
ceph: fix atomic_open snapdir

ceph_handle_snapdir() checks ceph_mdsc_do_request()'s return value
and creates snapdir inode if it's -ENOENT

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: properly mark empty directory as complete
Yan, Zheng [Mon, 19 Jan 2015 05:12:24 +0000 (13:12 +0800)]
ceph: properly mark empty directory as complete

ceph_add_cap() calls __check_cap_issue(), which clears directory
inode' complete flag. so we should set the complete flag for empty
directory should be set after calling ceph_add_cap().

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoclient: include kernel version in client metadata
Yan, Zheng [Fri, 16 Jan 2015 02:54:43 +0000 (10:54 +0800)]
client: include kernel version in client metadata

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: provide seperate {inode,file}_operations for snapdir
Yan, Zheng [Wed, 14 Jan 2015 05:46:04 +0000 (13:46 +0800)]
ceph: provide seperate {inode,file}_operations for snapdir

remove all unsupported operations from {inode,file}_operations.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: fix request time stamp encoding
Yan, Zheng [Tue, 13 Jan 2015 07:20:52 +0000 (15:20 +0800)]
ceph: fix request time stamp encoding

struct timespec uses 'long' to present second and nanosecond. 'long'
is 64 bits on 64bits machine. ceph MDS expects time stamp to be
encoded as struct ceph_timespec, which uses 'u32' to present second
and nanosecond.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: fix reading inline data when i_size > PAGE_SIZE
Yan, Zheng [Sat, 10 Jan 2015 03:43:12 +0000 (11:43 +0800)]
ceph: fix reading inline data when i_size > PAGE_SIZE

when inode has inline data but its size > PAGE_SIZE (it was truncated
to larger size), previous direct read code return -EIO. This patch adds
code to return zeros for data whose offset > PAGE_SIZE.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: avoid block operation when !TASK_RUNNING (ceph_mdsc_close_sessions)
Yan, Zheng [Fri, 9 Jan 2015 09:00:42 +0000 (17:00 +0800)]
ceph: avoid block operation when !TASK_RUNNING (ceph_mdsc_close_sessions)

use an atomic variable to track number of sessions, this can avoid block
operation inside wait loops.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: avoid block operation when !TASK_RUNNING (ceph_get_caps)
Yan, Zheng [Fri, 9 Jan 2015 07:56:18 +0000 (15:56 +0800)]
ceph: avoid block operation when !TASK_RUNNING (ceph_get_caps)

we should not do block operation in wait_event_interruptible()'s condition
check function, but reading inline data can block. so move the read inline
data code to ceph_get_caps()

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: avoid block operation when !TASK_RUNNING (ceph_mdsc_sync)
Yan, Zheng [Thu, 8 Jan 2015 13:30:12 +0000 (21:30 +0800)]
ceph: avoid block operation when !TASK_RUNNING (ceph_mdsc_sync)

check_cap_flush() calls mutex_lock(), which may block. So we can't
use it as condition check function for wait_event();

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agorbd: fix error paths in rbd_dev_refresh()
Ilya Dryomov [Thu, 8 Jan 2015 17:18:22 +0000 (20:18 +0300)]
rbd: fix error paths in rbd_dev_refresh()

header_rwsem should be released on errors.  Also remove useless
rbd_dev->mapping.size != rbd_dev->header.image_size test.

Signed-off-by: Ilya Dryomov <idryomov@redhat.com>
9 years agoceph: improve reference tracking for snaprealm
Yan, Zheng [Tue, 23 Dec 2014 07:30:54 +0000 (15:30 +0800)]
ceph: improve reference tracking for snaprealm

When snaprealm is created, its initial reference count is zero.
But in some rare cases, the newly created snaprealm is not referenced
by anyone. This causes snaprealm with zero reference count not freed.

The fix is set reference count of newly snaprealm to 1. The reference
is return the function who requests to create the snaprealm. When the
function finishes its job, it releases the reference.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: properly zero data pages for file holes.
Yan, Zheng [Tue, 6 Jan 2015 07:29:14 +0000 (15:29 +0800)]
ceph: properly zero data pages for file holes.

A bug is found in striped_read() of fs/ceph/file.c. striped_read() calls
ceph_zero_pape_vector_range().  The first argument, page_align + read + ret,
passed to ceph_zero_pape_vector_range() is wrong.

When a file has holes, this wrong parameter may cause memory corruption
either in kernal space or user space. Kernel space memory may be corrupted in
the case of non direct IO; user space memory may be corrupted in the case of
direct IO. In the latter case, the application doing direct IO may crash due
to memory corruption, as we have experienced.

The correct value should be initial_align + read + ret, where intial_align =
o_direct ? buf_align : io_align.  Compared with page_align, the current page
offest, initial_align is the initial page offest, which should be used to
calculate the page and offset in ceph_zero_pape_vector_range().

Reported-by: caifeng zhu <zhucaifeng@unissoft-nj.com>
Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agoceph: acl: Remove unused function
Rickard Strandqvist [Sat, 3 Jan 2015 23:44:54 +0000 (00:44 +0100)]
ceph: acl: Remove unused function

Remove the function ceph_get_cached_acl() that is not used anywhere.

This was partially found by using a static code analysis program called cppcheck.

Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Reviewed-by: Yan, Zheng <zyan@redhat.com>
9 years agorbd: nuke copy_token()
Rickard Strandqvist [Thu, 1 Jan 2015 16:58:32 +0000 (17:58 +0100)]
rbd: nuke copy_token()

It's been largely superseded by dup_token() and unused for over
2 years, identified by cppcheck.

Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
[idryomov@redhat.com: changelog]
Signed-off-by: Ilya Dryomov <idryomov@redhat.com>
9 years agoceph: handle SESSION_FORCE_RO message
Yan, Zheng [Mon, 5 Jan 2015 03:04:04 +0000 (11:04 +0800)]
ceph: handle SESSION_FORCE_RO message

mark session as readonly and wake up all cap waiters.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
9 years agolibceph: use mon_client.c/put_generic_request() more
Ilya Dryomov [Mon, 22 Dec 2014 16:35:17 +0000 (19:35 +0300)]
libceph: use mon_client.c/put_generic_request() more

Signed-off-by: Ilya Dryomov <idryomov@redhat.com>
9 years agolibceph: nuke pool op infrastructure
Ilya Dryomov [Mon, 22 Dec 2014 16:14:26 +0000 (19:14 +0300)]
libceph: nuke pool op infrastructure

On Mon, Dec 22, 2014 at 5:35 PM, Sage Weil <sage@newdream.net> wrote:
> On Mon, 22 Dec 2014, Ilya Dryomov wrote:
>> Actually, pool op stuff has been unused for over two years - looks like
>> it was added for rbd create_snap and that got ripped out in 2012.  It's
>> unlikely we'd ever need to manage pools or snaps from the kernel client
>> so I think it makes sense to nuke it.  Sage?
>
> Yep!

Signed-off-by: Ilya Dryomov <idryomov@redhat.com>
9 years agos390/mm: align 64-bit PIE binaries to 4GB
Martin Schwidefsky [Thu, 12 Feb 2015 13:17:52 +0000 (14:17 +0100)]
s390/mm: align 64-bit PIE binaries to 4GB

The base address (STACK_TOP / 3 * 2) for a 64-bit program is two thirds
into the 4GB segment at 0x2aa00000000. The randomization added on z13
can eat another 1GB of the remaining 1.33GB to the next 4GB boundary.
In the worst case 300MB are left for the executable + bss which may
cross into the next 4GB segment. This is bad for branch prediction,
therefore align the base address to 4GB to give the program more room
before it crosses the 4GB boundary.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
9 years agox86/intel/quark: Fix simple_return.cocci warnings
Fengguang Wu [Thu, 19 Feb 2015 08:14:32 +0000 (16:14 +0800)]
x86/intel/quark: Fix simple_return.cocci warnings

arch/x86/platform/intel-quark/imr.c:129:1-4: WARNING: end returns can be simpified

 Simplify a trivial if-return sequence.  Possibly combine with a preceding function call.

Generated by: scripts/coccinelle/misc/simple_return.cocci

Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Cc: Andy Shevchenko <andy.schevchenko@gmail.com>
Cc: Ong, Boon Leong <boon.leong.ong@intel.com>
Cc: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Darren Hart <dvhart@linux.intel.com>
Cc: kbuild-all@01.org
Link: http://lkml.kernel.org/r/20150219081432.GA21996@waimea
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agox86/intel/quark: Fix ptr_ret.cocci warnings
Fengguang Wu [Thu, 19 Feb 2015 08:14:32 +0000 (16:14 +0800)]
x86/intel/quark: Fix ptr_ret.cocci warnings

arch/x86/platform/intel-quark/imr.c:280:1-3: WARNING: PTR_ERR_OR_ZERO can be used

 Use PTR_ERR_OR_ZERO rather than if(IS_ERR(...)) + PTR_ERR

Generated by: scripts/coccinelle/api/ptr_ret.cocci

Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Cc: Andy Shevchenko <andy.schevchenko@gmail.com>
Cc: Ong, Boon Leong <boon.leong.ong@intel.com>
Cc: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Darren Hart <dvhart@linux.intel.com>
Cc: kbuild-all@01.org
Link: http://lkml.kernel.org/r/20150219081432.GA21983@waimea
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agoARM: make of_device_ids const
Uwe Kleine-König [Wed, 18 Feb 2015 20:19:56 +0000 (21:19 +0100)]
ARM: make of_device_ids const

of_device_ids (i.e. compatible strings and the respective data) are not
supposed to change at runtime. All functions working with of_device_ids
provided by <linux/of.h> work with const of_device_ids. So mark the
non-const structs in arch/arm as const, too.

While at it also add some __initconst annotations.

Acked-by: Jason Cooper <jason@lakedameon.net>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
9 years agoARM: make arrays containing machine compatible strings const
Uwe Kleine-König [Wed, 18 Feb 2015 20:01:45 +0000 (21:01 +0100)]
ARM: make arrays containing machine compatible strings const

The definition

static const char *axxia_dt_match[] __initconst = {
...

defines a changable array of constant strings. That is you must not do:

*axxia_dt_match[0] = 'k';

but

axxia_dt_match[0] = "different string";

is fine. So the annotation __initconst is wrong and yields a compiler
error when other really const variables are added with __initconst.

As the struct machine_desc member dt_compat is declared as

const char *const *dt_compat;

making the arrays const is the better alternative over changing all
annotations to __initdata.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
9 years agocpufreq: s3c: remove last use of resume_clocks callback
Arnd Bergmann [Wed, 18 Feb 2015 20:55:53 +0000 (21:55 +0100)]
cpufreq: s3c: remove last use of resume_clocks callback

Commit 32726d2d550 ("ARM: SAMSUNG: Remove legacy clock code")
already removed the callback pointer, but there was one remaining
user:

drivers/cpufreq/s3c24xx-cpufreq.c: In function 's3c_cpufreq_resume_clocks':
drivers/cpufreq/s3c24xx-cpufreq.c:149:14: error: 'struct s3c_cpufreq_info' has no member named 'resume_clocks'
  cpu_cur.info->resume_clocks();
              ^

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 32726d2d550 ("ARM: SAMSUNG: Remove legacy clock code")
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Cc: 3.17+ <stable@vger.kernel.org> # v3.17+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
9 years agocpufreq: s3c: remove incorrect __init annotations
Arnd Bergmann [Wed, 18 Feb 2015 20:55:03 +0000 (21:55 +0100)]
cpufreq: s3c: remove incorrect __init annotations

The two functions s3c2416_cpufreq_driver_init and s3c_cpufreq_register
are marked init but are called from a context that might be run after
the __init sections are discarded, as the compiler points out:

WARNING: vmlinux.o(.data+0x1ad9dc): Section mismatch in reference from the variable s3c2416_cpufreq_driver to the function .init.text:s3c2416_cpufreq_driver_init()
WARNING: drivers/built-in.o(.text+0x35b5dc): Section mismatch in reference from the function s3c2410a_cpufreq_add() to the function .init.text:s3c_cpufreq_register()

This removes the __init markings.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Cc: All applicable <stable@vger.kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
9 years agox86/intel/quark: Add Intel Quark platform support
Bryan O'Donoghue [Fri, 30 Jan 2015 16:29:39 +0000 (16:29 +0000)]
x86/intel/quark: Add Intel Quark platform support

Add Intel Quark platform support. Quark needs to pull down all
unlocked IMRs to ensure agreement with the EFI memory map post
boot.

This patch adds an entry in Kconfig for Quark as a platform and
makes IMR support mandatory if selected.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Suggested-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Tested-by: Ong, Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Reviewed-by: Andy Shevchenko <andy.schevchenko@gmail.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Reviewed-by: Ong, Boon Leong <boon.leong.ong@intel.com>
Cc: dvhart@infradead.org
Link: http://lkml.kernel.org/r/1422635379-12476-3-git-send-email-pure.logic@nexus-software.ie
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agox86/intel/quark: Add Isolated Memory Regions for Quark X1000
Bryan O'Donoghue [Fri, 30 Jan 2015 16:29:38 +0000 (16:29 +0000)]
x86/intel/quark: Add Isolated Memory Regions for Quark X1000

Intel's Quark X1000 SoC contains a set of registers called
Isolated Memory Regions. IMRs are accessed over the IOSF mailbox
interface. IMRs are areas carved out of memory that define
read/write access rights to the various system agents within the
Quark system. For a given agent in the system it is possible to
specify if that agent may read or write an area of memory
defined by an IMR with a granularity of 1 KiB.

Quark_SecureBootPRM_330234_001.pdf section 4.5 details the
concept of IMRs quark-x1000-datasheet.pdf section 12.7.4 details
the implementation of IMRs in silicon.

eSRAM flush, CPU Snoop write-only, CPU SMM Mode, CPU non-SMM
mode, RMU and PCIe Virtual Channels (VC0 and VC1) can have
individual read/write access masks applied to them for a given
memory region in Quark X1000. This enables IMRs to treat each
memory transaction type listed above on an individual basis and
to filter appropriately based on the IMR access mask for the
memory region. Quark supports eight IMRs.

Since all of the DMA capable SoC components in the X1000 are
mapped to VC0 it is possible to define sections of memory as
invalid for DMA write operations originating from Ethernet, USB,
SD and any other DMA capable south-cluster component on VC0.
Similarly it is possible to mark kernel memory as non-SMM mode
read/write only or to mark BIOS runtime memory as SMM mode
accessible only depending on the particular memory footprint on
a given system.

On an IMR violation Quark SoC X1000 systems are configured to
reset the system, so ensuring that the IMR memory map is
consistent with the EFI provided memory map is critical to
ensure no IMR violations reset the system.

The API for accessing IMRs is based on MTRR code but doesn't
provide a /proc or /sys interface to manipulate IMRs. Defining
the size and extent of IMRs is exclusively the domain of
in-kernel code.

Quark firmware sets up a series of locked IMRs around pieces of
memory that firmware owns such as ACPI runtime data. During boot
a series of unlocked IMRs are placed around items in memory to
guarantee no DMA modification of those items can take place.
Grub also places an unlocked IMR around the kernel boot params
data structure and compressed kernel image. It is necessary for
the kernel to tear down all unlocked IMRs in order to ensure
that the kernel's view of memory passed via the EFI memory map
is consistent with the IMR memory map. Without tearing down all
unlocked IMRs on boot transitory IMRs such as those used to
protect the compressed kernel image will cause IMR violations and system reboots.

The IMR init code tears down all unlocked IMRs and sets a
protective IMR around the kernel .text and .rodata as one
contiguous block. This sanitizes the IMR memory map with respect
to the EFI memory map and protects the read-only portions of the
kernel from unwarranted DMA access.

Tested-by: Ong, Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Reviewed-by: Andy Shevchenko <andy.schevchenko@gmail.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Reviewed-by: Ong, Boon Leong <boon.leong.ong@intel.com>
Cc: andy.shevchenko@gmail.com
Cc: dvhart@infradead.org
Link: http://lkml.kernel.org/r/1422635379-12476-2-git-send-email-pure.logic@nexus-software.ie
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agoDocumentation/x86: Fix path in zero-page.txt
Alexander Kuleshov [Sat, 31 Jan 2015 07:23:24 +0000 (13:23 +0600)]
Documentation/x86: Fix path in zero-page.txt

Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>
Cc: Martin Mares <mj@ucw.cz>
Link: http://lkml.kernel.org/r/1422689004-13318-1-git-send-email-kuleshovmail@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agox86/apic: Fix the devicetree build in certain configs
Ricardo Ribalda Delgado [Mon, 2 Feb 2015 19:27:11 +0000 (20:27 +0100)]
x86/apic: Fix the devicetree build in certain configs

Without this patch:

  LD      init/built-in.o
  arch/x86/built-in.o: In function `dtb_lapic_setup': kernel/devicetree.c:155:
  undefined reference to `apic_force_enable'
  Makefile:923: recipe for target 'vmlinux' failed
  make: *** [vmlinux] Error 1

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Reviewed-by: Maciej W. Rozycki <macro@linux-mips.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Jan Beulich <JBeulich@suse.com>
Link: http://lkml.kernel.org/r/1422905231-16067-1-git-send-email-ricardo.ribalda@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agoNFSv4.1: Clean up bind_conn_to_session
Trond Myklebust [Wed, 18 Feb 2015 17:27:18 +0000 (09:27 -0800)]
NFSv4.1: Clean up bind_conn_to_session

We don't need to fake up an entire session in order retrieve the arguments.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
9 years agokprobes/x86: Mark 2 bytes NOP as boostable
Wang Nan [Tue, 10 Feb 2015 01:34:05 +0000 (09:34 +0800)]
kprobes/x86: Mark 2 bytes NOP as boostable

Currently, x86 kprobes is unable to boost 2 bytes nop like:

  nopl 0x0(%rax,%rax,1)

which is 0x0f 0x1f 0x44 0x00 0x00.

Such nops have exactly 5 bytes to hold a relative jmp
instruction. Boosting them should be obviously safe.

This patch enable boosting such nops by simply updating
twobyte_is_boostable[] array.

Signed-off-by: Wang Nan <wangnan0@huawei.com>
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: <lizefan@huawei.com>
Link: http://lkml.kernel.org/r/1423532045-41049-1-git-send-email-wangnan0@huawei.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agoNFSv4.1: Always set up a forward channel when binding the session
Trond Myklebust [Wed, 18 Feb 2015 20:07:19 +0000 (12:07 -0800)]
NFSv4.1: Always set up a forward channel when binding the session

Currently, the client requests a back channel or a bidirectional
connection when binding a new TCP channel to an existing session.
Fix that to ask for a forward channel or bidirectional.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
9 years agoNFSv4.1: Don't set up a backchannel if the server didn't agree to do so
Trond Myklebust [Wed, 18 Feb 2015 19:34:58 +0000 (11:34 -0800)]
NFSv4.1: Don't set up a backchannel if the server didn't agree to do so

If the server doesn't agree to out backchannel setup request, then
don't set one up.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
9 years agoNFSv4.1: Clean up create_session
Trond Myklebust [Wed, 18 Feb 2015 19:30:18 +0000 (11:30 -0800)]
NFSv4.1: Clean up create_session

Don't decode directly into the shared struct session

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
9 years agouprobes/x86: Fix 2-byte opcode table
Denys Vlasenko [Thu, 12 Feb 2015 19:18:52 +0000 (20:18 +0100)]
uprobes/x86: Fix 2-byte opcode table

Enabled probing of lar, lsl, popcnt, lddqu, prefetch insns.
They should be safe to probe, they throw no exceptions.

Enabled probing of 3-byte opcodes 0f 38-3f xx - these are
vector isns, so should be safe.

Enabled probing of many currently undefined 0f xx insns.
At the rate new vector instructions are getting added,
we don't want to constantly enable more bits.
We want to only occasionally *disable* ones which
for some reason can't be probed.
This includes 0f 24,26 opcodes, which are undefined
since Pentium. On 486, they were "mov to/from test register".

Explained more fully what 0f 78,79 opcodes are.

Explained what 0f ae opcode is. (It's unclear why we don't allow
probing it, but let's not change it for now).

Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Jim Keniston <jkenisto@us.ibm.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1423768732-32194-3-git-send-email-dvlasenk@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agouprobes/x86: Fix 1-byte opcode tables
Denys Vlasenko [Thu, 12 Feb 2015 19:18:51 +0000 (20:18 +0100)]
uprobes/x86: Fix 1-byte opcode tables

This change fixes 1-byte opcode tables so that only insns
for which we have real reasons to disallow probing are marked
with unset bits.

To that end:

Set bits for all prefix bytes. Their setting is ignored anyway -
we check the bitmap against OPCODE1(insn), not against first
byte. Keeping them set to 0 only confuses code reader with
"why we don't support that opcode" question.

Thus: enable bytes c4,c5 in 64-bit mode (VEX prefixes).
Byte 62 (EVEX prefix) is not yet enabled since insn decoder
does not support that yet.

For 32-bit mode, enable probing of opcodes 63 (arpl) and d6
(salc). They don't require any special handling.

For 64-bit mode, disable 9a and ea - these undefined opcodes
were mistakenly left enabled.

Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Jim Keniston <jkenisto@us.ibm.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1423768732-32194-2-git-send-email-dvlasenk@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agouprobes/x86: Add comment with insn opcodes, mnemonics and why we dont support them
Denys Vlasenko [Thu, 12 Feb 2015 19:18:50 +0000 (20:18 +0100)]
uprobes/x86: Add comment with insn opcodes, mnemonics and why we dont support them

After adding these, it's clear we have some awkward choices
there. Some valid instructions are prohibited from uprobing
while several invalid ones are allowed.

Hopefully future edits to the good-opcode tables will fix wrong
bits or explain why those bits are not wrong.

No actual code changes.

Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Jim Keniston <jkenisto@us.ibm.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1423768732-32194-1-git-send-email-dvlasenk@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agoMerge tag 'locks-v3.20-2' of git://git.samba.org/jlayton/linux
Linus Torvalds [Wed, 18 Feb 2015 18:21:47 +0000 (10:21 -0800)]
Merge tag 'locks-v3.20-2' of git://git.samba.org/jlayton/linux

Pull file locking fixes from Jeff Layton:
 "A small set of patches to fix problems with the recent file locking
  changes that we discussed earlier this week"
"

* tag 'locks-v3.20-2' of git://git.samba.org/jlayton/linux:
  locks: fix list insertion when lock is split in two
  locks: remove conditional lock release in middle of flock_lock_file
  locks: only remove leases associated with the file being closed
  Revert "locks: keep a count of locks on the flctx lists"

9 years agoRevert "clk: mxs: Fix invalid 32-bit access to frac registers"
Stefan Wahren [Thu, 12 Feb 2015 20:30:07 +0000 (20:30 +0000)]
Revert "clk: mxs: Fix invalid 32-bit access to frac registers"

Revert commit 039e59707507 (clk: mxs: Fix invalid 32-bit access to frac
registers), because it leads to a faulty spi communication on mx28evk.

Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Reported-by: Fabio Estevam <fabio.estevam@freescale.com>
Tested-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoMerge tag 'asm-generic-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 18 Feb 2015 18:02:24 +0000 (10:02 -0800)]
Merge tag 'asm-generic-for-linus' of git://git./linux/kernel/git/arnd/asm-generic

Pull asm-generic uaccess.h cleanup from Arnd Bergmann:
 "Like in 3.19, I once more have a multi-stage cleanup for one
  asm-generic header file, this time the work was done by Michael
  Tsirkin and cleans up the uaccess.h file in asm-generic, as well as
  all architectures for which the respective maintainers did not pick up
  his patches directly"

* tag 'asm-generic-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic: (37 commits)
  sparc32: nocheck uaccess coding style tweaks
  sparc64: nocheck uaccess coding style tweaks
  xtensa: macro whitespace fixes
  sh: macro whitespace fixes
  parisc: macro whitespace fixes
  m68k: macro whitespace fixes
  m32r: macro whitespace fixes
  frv: macro whitespace fixes
  cris: macro whitespace fixes
  avr32: macro whitespace fixes
  arm64: macro whitespace fixes
  arm: macro whitespace fixes
  alpha: macro whitespace fixes
  blackfin: macro whitespace fixes
  sparc64: uaccess_64 macro whitespace fixes
  sparc32: uaccess_32 macro whitespace fixes
  avr32: whitespace fix
  sh: fix put_user sparse errors
  metag: fix put_user sparse errors
  ia64: fix put_user sparse errors
  ...

9 years agoclk: qoriq: Add support for the platform PLL
Emil Medve [Wed, 21 Jan 2015 10:03:31 +0000 (04:03 -0600)]
clk: qoriq: Add support for the platform PLL

Change-Id: Iac11ed95f274485a86d2c11f32a3dc502bcd020f
Signed-off-by: Emil Medve <Emilian.Medve@Freescale.com>
Acked-by: Tang Yuantian <Yuantian.Tang@freescale.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agopowerpc/corenet: Enable CLK_QORIQ
Emil Medve [Wed, 21 Jan 2015 10:03:30 +0000 (04:03 -0600)]
powerpc/corenet: Enable CLK_QORIQ

Change-Id: I1a80ad7b9f6854791bd270b746f93a91439155a6
Signed-off-by: Emil Medve <Emilian.Medve@Freescale.com>
Acked-by: Tang Yuantian <Yuantian.Tang@freescale.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoACPI / LPSS: Deassert resets for SPI host controllers on Braswell
Mika Westerberg [Wed, 18 Feb 2015 11:50:17 +0000 (13:50 +0200)]
ACPI / LPSS: Deassert resets for SPI host controllers on Braswell

On some Braswell systems BIOS leaves resets for SPI host controllers
active. This prevents the SPI driver from transferring messages on wire.

Fix this in similar way that we do for I2C already by deasserting resets
for the SPI host controllers.

Reported-by: Yang A Fang <yang.a.fang@intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: 3.17+ <stable@vger.kernel.org> # 3.17+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
9 years agoACPI / LPSS: Always disable I2C host controllers
Mika Westerberg [Wed, 18 Feb 2015 11:50:16 +0000 (13:50 +0200)]
ACPI / LPSS: Always disable I2C host controllers

On Baytrail and Braswell the BIOS might leave the I2C host controllers
enabled, probably because it uses them for its own purposes. This is fine
in normal cases because the I2C driver will disable the hardware when it
is probed anyway.

However, in case of suspend to disk it is different story. If the driver
happens to be compiled as a module the boot kernel never loads the driver
thus leaving host controllers enabled upon loading the hibernation image.

The I2C host controller interrupt mask register has default value of 0x8ff,
in other words it has most of the interrupts unmasked. When combined with
the fact that the host controller is enabled, the driver immediately starts
getting interrupts even before its resume hook is called (once IO-APIC is
resumed). Since the driver is not prepared for this it will crash the
kernel due to NULL pointer derefence because dev->msgs is NULL.

Unfortunately we were not able to get full backtrace to from the console
which could be reproduced here.

In order to fix this even when the driver is compiled as module, we disable
the I2C host controllers in byt_i2c_setup() before devices are created.

Reported-by: Yu Chen <yu.c.chen@intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: 3.17+ <stable@vger.kernel.org> # 3.17+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
9 years agoMerge tag 'pci-v3.20-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaa...
Linus Torvalds [Wed, 18 Feb 2015 17:43:46 +0000 (09:43 -0800)]
Merge tag 'pci-v3.20-fixes-1' of git://git./linux/kernel/git/helgaas/pci

Pull PCI resource management fix from Bjorn Helgaas:
 "This fixes a double kfree in code we merged for v3.20:

  Remove duplicate kfree in of_pci_get_host_bridge_resources() (Lorenzo
  Pieralisi)"

* tag 'pci-v3.20-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci:
  of/pci: Remove duplicate kfree in of_pci_get_host_bridge_resources()

9 years agoclk: Replace explicit clk assignment with __clk_hw_set_clk
Javier Martinez Canillas [Thu, 12 Feb 2015 13:58:30 +0000 (14:58 +0100)]
clk: Replace explicit clk assignment with __clk_hw_set_clk

The change in the clk API to return a per-user clock instance, moved
the clock state to struct clk_core so now the struct clk_hw .core field
is used instead of .clk for most operations.

So for hardware clocks that needs to share the same clock state, both
the .core and .clk pointers have to be assigned but currently only the
.clk is set. This leads to NULL pointer dereference when the operations
try to access the hw clock .core. For example, the composite clock rate
and mux components didn't have a .core set which leads to this error:

Unable to handle kernel NULL pointer dereference at virtual address 00000034
pgd = c0004000
[00000034] *pgd=00000000
Internal error: Oops: 5 [#1] PREEMPT SMP ARM
Modules linked in:
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.19.0-next-20150211-00002-g1fb7f0e1150d #423
Hardware name: SAMSUNG EXYNOS (Flattened Device Tree)
task: ee480000 ti: ee488000 task.ti: ee488000
PC is at clk_mux_determine_rate_flags+0x14/0x19c
LR is at __clk_mux_determine_rate+0x24/0x2c
pc : [<c03a355c>]    lr : [<c03a3734>]    psr: a0000113
sp : ee489ce8  ip : ee489d84  fp : ee489d84
r10: 0000005c  r9 : 00000001  r8 : 016e3600
r7 : 00000000  r6 : 00000000  r5 : ee442200  r4 : ee440c98
r3 : ffffffff  r2 : 00000000  r1 : 016e3600  r0 : ee440c98
Flags: NzCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment kernel
Control: 10c5387d  Table: 4000406a  DAC: 00000015
Process swapper/0 (pid: 1, stack limit = 0xee488210)
Stack: (0xee489ce8 to 0xee48a000)
9ce0:                   00000000 ffffffff 60000113 ee440c98 ee442200 00000000
9d00: 016e3600 ffffffff 00000001 0000005c ee489d84 c03a3734 ee489d80 ee489d84
9d20: 00000000 c048b130 00000400 c03a5798 ee489d80 ee489d84 c0607f60 ffffffea
9d40: 00000001 00000001 ee489d5c c003f844 c06e3340 ee402680 ee440d0c ed935000
9d60: 016e3600 00000003 00000001 0000005c eded3700 c03a11a0 ee489d80 ee489d84
9d80: 016e3600 ee402680 c05b413a eddc9900 016e3600 c03a1228 00000000 ffffffff
9da0: ffffffff eddc9900 016e3600 c03a1c1c ffffffff 016e3600 ed8c6710 c03d6ce4
9dc0: eded3400 00000000 00000000 c03c797c 00000001 0000005c eded3700 eded3700
9de0: 000005e0 00000001 0000005c c03db8ac c06e7e54 c03c8f08 00000000 c06e7e64
9e00: c06b6e74 c06e7f64 000005e0 c06e7df8 c06e5100 00000000 c06e7e6c c06e7f54
9e20: 00000000 00000000 eebd9550 00000000 c06e7da0 c06e7e54 ee7b5010 c06e7da0
9e40: eddc9690 c06e7db4 c06b6e74 00000097 00000000 c03d4398 00000000 ee7b5010
9e60: eebd9550 c06e7da0 00000000 c03db824 ee7b5010 fffffffe c06e7db4 c0299c7c
9e80: ee7b5010 c072a05c 00000000 c0298858 ee7b5010 c06e7db4 ee7b5044 00000000
9ea0: eddc9580 c0298a04 c06e7db4 00000000 c0298978 c02971d4 ee405c78 ee732b40
9ec0: c06e7db4 eded3800 c06d6738 c0298044 c0608300 c06e7db4 00000000 c06e7db4
9ee0: 00000000 c06beb58 c06beb58 c0299024 00000000 c068dd00 00000000 c0008944
9f00: 00000038 c049013c ee462200 c0711920 ee480000 60000113 c06c2cb0 00000000
9f20: 00000000 c06c2cb0 60000113 00000000 ef7fcafc 00000000 c0640194 c00389ec
9f40: c05ec3a8 c063f824 00000006 00000006 c06c2c50 c0696444 00000006 c0696424
9f60: c06ee1c0 c066b588 c06b6e74 00000097 00000000 c066bd44 00000006 00000006
9f80: c066b588 c003d684 00000000 c0481938 00000000 00000000 00000000 00000000
9fa0: 00000000 c0481940 00000000 c000e680 00000000 00000000 00000000 00000000
9fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
9fe0: 00000000 00000000 00000000 00000000 00000013 00000000 00000000 00000000
[<c03a355c>] (clk_mux_determine_rate_flags) from [<c03a3734>] (__clk_mux_determine_rate+0x24/0x2c)
[<c03a3734>] (__clk_mux_determine_rate) from [<c03a5798>] (clk_composite_determine_rate+0xbc/0x238)
[<c03a5798>] (clk_composite_determine_rate) from [<c03a11a0>] (clk_core_round_rate_nolock+0x5c/0x9c)
[<c03a11a0>] (clk_core_round_rate_nolock) from [<c03a1228>] (__clk_round_rate+0x38/0x40)
[<c03a1228>] (__clk_round_rate) from [<c03a1c1c>] (clk_round_rate+0x20/0x38)
[<c03a1c1c>] (clk_round_rate) from [<c03d6ce4>] (max98090_dai_set_sysclk+0x34/0x118)
[<c03d6ce4>] (max98090_dai_set_sysclk) from [<c03c797c>] (snd_soc_dai_set_sysclk+0x38/0x80)
[<c03c797c>] (snd_soc_dai_set_sysclk) from [<c03db8ac>] (snow_late_probe+0x24/0x48)
[<c03db8ac>] (snow_late_probe) from [<c03c8f08>] (snd_soc_register_card+0xf04/0x1070)
[<c03c8f08>] (snd_soc_register_card) from [<c03d4398>] (devm_snd_soc_register_card+0x30/0x64)
[<c03d4398>] (devm_snd_soc_register_card) from [<c03db824>] (snow_probe+0x68/0xcc)
[<c03db824>] (snow_probe) from [<c0299c7c>] (platform_drv_probe+0x48/0x98)
[<c0299c7c>] (platform_drv_probe) from [<c0298858>] (driver_probe_device+0x114/0x234)
[<c0298858>] (driver_probe_device) from [<c0298a04>] (__driver_attach+0x8c/0x90)
[<c0298a04>] (__driver_attach) from [<c02971d4>] (bus_for_each_dev+0x54/0x88)
[<c02971d4>] (bus_for_each_dev) from [<c0298044>] (bus_add_driver+0xd8/0x1cc)
[<c0298044>] (bus_add_driver) from [<c0299024>] (driver_register+0x78/0xf4)
[<c0299024>] (driver_register) from [<c0008944>] (do_one_initcall+0x80/0x1d0)
[<c0008944>] (do_one_initcall) from [<c066bd44>] (kernel_init_freeable+0x10c/0x1d8)
[<c066bd44>] (kernel_init_freeable) from [<c0481940>] (kernel_init+0x8/0xe4)
[<c0481940>] (kernel_init) from [<c000e680>] (ret_from_fork+0x14/0x34)
Code: e24dd00c e5907000 e1a08001 e88d000c (e5970034)

The changes were made using the following cocinelle semantic patch:

@i@
@@

@depends on i@
identifier dst;
@@

- dst->clk = hw->clk;
+ __clk_hw_set_clk(dst, hw);

@depends on i@
identifier dst;
@@

- dst->hw.clk = hw->clk;
+ __clk_hw_set_clk(&dst->hw, hw);

Fixes: 035a61c314eb3 ("clk: Make clk API return per-user struct clk instances")
Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: Add __clk_hw_set_clk helper function
Javier Martinez Canillas [Thu, 12 Feb 2015 13:58:29 +0000 (14:58 +0100)]
clk: Add __clk_hw_set_clk helper function

After the clk API change to return a per-user clock instance, both the
struct clk_core and struct clk pointers from the hw clock needs to be
assigned to clock that share the same state.

In the future the struct clk_core will be removed and this is going to
change again so to avoid having to change the assignments twice in all
the drivers, add a helper function to have an indirection level.

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: Don't dereference parent clock if is NULL
Javier Martinez Canillas [Thu, 12 Feb 2015 13:58:28 +0000 (14:58 +0100)]
clk: Don't dereference parent clock if is NULL

The clock passed as an argument to clk_mux_determine_rate_flags()
has the CLK_SET_RATE_PARENT flag set but it has no parent, then a
NULL pointer will tried to be dereferenced.

This shouldn't happen since setting that flag for a clock with no
parent is a bug but the core should be robust to handle that case.

Fixes: 035a61c314eb3 ("clk: Make clk API return per-user struct clk instances")
Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoMIPS: Alchemy: Remove bogus args from alchemy_clk_fgcs_detr
Tomeu Vizoso [Fri, 13 Feb 2015 13:34:54 +0000 (14:34 +0100)]
MIPS: Alchemy: Remove bogus args from alchemy_clk_fgcs_detr

They were added to this function by mistake when they were added to the
clk_ops.determine_rate callback.

Fixes: 1c8e600440c7 ("clk: Add rate constraints to clocks")
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoMerge tag 'virtio-next-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 18 Feb 2015 17:24:01 +0000 (09:24 -0800)]
Merge tag 'virtio-next-for-linus' of git://git./linux/kernel/git/rusty/linux

Pull virtio updates from Rusty Russell:
 "OK, this has the big virtio 1.0 implementation, as specified by OASIS.

  On top of tht is the major rework of lguest, to use PCI and virtio
  1.0, to double-check the implementation.

  Then comes the inevitable fixes and cleanups from that work"

* tag 'virtio-next-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux: (80 commits)
  virtio: don't set VIRTIO_CONFIG_S_DRIVER_OK twice.
  virtio_net: unconditionally define struct virtio_net_hdr_v1.
  tools/lguest: don't use legacy definitions for net device in example launcher.
  virtio: Don't expose legacy net features when VIRTIO_NET_NO_LEGACY defined.
  tools/lguest: use common error macros in the example launcher.
  tools/lguest: give virtqueues names for better error messages
  tools/lguest: more documentation and checking of virtio 1.0 compliance.
  lguest: don't look in console features to find emerg_wr.
  tools/lguest: don't start devices until DRIVER_OK status set.
  tools/lguest: handle indirect partway through chain.
  tools/lguest: insert driver references from the 1.0 spec (4.1 Virtio Over PCI)
  tools/lguest: insert device references from the 1.0 spec (4.1 Virtio Over PCI)
  tools/lguest: rename virtio_pci_cfg_cap field to match spec.
  tools/lguest: fix features_accepted logic in example launcher.
  tools/lguest: handle device reset correctly in example launcher.
  virtual: Documentation: simplify and generalize paravirt_ops.txt
  lguest: remove NOTIFY call and eventfd facility.
  lguest: remove NOTIFY facility from demonstration launcher.
  lguest: use the PCI console device's emerg_wr for early boot messages.
  lguest: always put console in PCI slot #1.
  ...

9 years agoMerge tag 'mfd-for-linus-3.20' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd
Linus Torvalds [Wed, 18 Feb 2015 17:05:48 +0000 (09:05 -0800)]
Merge tag 'mfd-for-linus-3.20' of git://git./linux/kernel/git/lee/mfd

Pull MFD updates from Lee Jones:
 "Changes to existing drivers:
   - fixr platform device collision; da9052, wm8994-core
   - regmap configuration amendments; tps65218
   - fix runtime PM deadlock; rtsx_usb
   - remove unused/superfluous code; db8500-prcmu, omap-usb-host
   - enable watchdog timer; lpc_sch
   - add start/stop RX URBs helpers; dln2
   - remove platform device (DT only); max77686, max77802
   - support suspend and resume; dln2
   - add Device Tree support; da9063
   - extra error checking; intel_soc_pmic
   - const'ify all the things; 88pm860x, hi6421-pmic, intel_soc_pmic,
                               max77686, lm3533, retu, pcf50633,
                               davinci_voicecodec, smsc-ece1099,
                               tps65218, mc13xxx, tps65217, twl-core,
                               twl6040

  New drivers/supported devices:
   - new driver for Richtek RT5033
   - new driver for DA9150 Charger and FuelGauge
   - new driver for Qualcomm Resource Power Manager (RPM)
   - add support for the ir-clk into sun6i-prcm
   - add support for FuelGauge into axp20x"

* tag 'mfd-for-linus-3.20' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd: (32 commits)
  mfd: intel_soc_pmic: Add missing error check for devm_kzalloc
  mfd: rtsx_usb: Defer autosuspend while card exists
  mfd: devicetree: Add bindings for DA9063
  mfd: da9063: Add device tree support
  regulator: qcom-rpm: Add missing state flag in call to RPM
  mfd: qcom-rpm: Driver for the Qualcomm RPM
  mfd: devicetree: bindings: Add Qualcomm RPM DT binding
  mfd: max77686/802: Remove support for board files
  mfd: omap-usb-host: Remove some unused functions
  mfd: twl6040: Constify struct regmap_config and reg_default array
  mfd: twl-core: Constify struct regmap_config and reg_default array
  mfd: tps65217: Constify struct regmap_config
  mfd: mc13xxx: i2c/spi: Constify struct regmap_config
  mfd: tps65218: Constify struct regmap_config
  mfd: smsc-ece1099: Constify struct regmap_config
  mfd: davinci_voicecodec: Constify struct regmap_config
  mfd: pcf50633: Constify struct regmap_config
  mfd: retu: Constify struct regmap_config
  mfd: lm3533: Constify struct regmap_config
  mfd: max77686: Constify struct regmap_config
  ...

9 years agoMerge tag 'pwm/for-3.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/thierry...
Linus Torvalds [Wed, 18 Feb 2015 17:00:53 +0000 (09:00 -0800)]
Merge tag 'pwm/for-3.20-rc1' of git://git./linux/kernel/git/thierry.reding/linux-pwm

Pull pwm updates from Thierry Reding:
 "This contains two new drivers, one for Allwinner SoCs and the other
  for Imagination Technologies' Pistachio SoC.

  Complementing this are a couple of fixes to the Atmel HLCDC PWM and
  STi PWM drivers as well as minor cleanups to the core and the Tegra
  driver"

* tag 'pwm/for-3.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/thierry.reding/linux-pwm:
  pwm: tegra: Use NSEC_PER_SEC
  pwm: Remove unnecessary check before of_node_put()
  pwm: Add device tree binding document for IMG PWM DAC
  pwm: Imagination Technologies PWM DAC driver
  pwm: sti: Maintain a bitmap of configured devices
  pwm: sunxi: document OF bindings
  pwm: Add Allwinner SoC support
  pwm: atmel-hlcdc: Prevent division by zero
  pwm: atmel-hlcdc: Depend on HAVE_CLK

9 years agoMerge branch 'rcu/next' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck...
Ingo Molnar [Wed, 18 Feb 2015 16:59:20 +0000 (17:59 +0100)]
Merge branch 'rcu/next' of git://git./linux/kernel/git/paulmck/linux-rcu into core/urgent

Pull RCU fix from Paul E. McKenney.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
9 years agoMerge branch 'for-linus' of git://git.infradead.org/users/vkoul/slave-dma
Linus Torvalds [Wed, 18 Feb 2015 16:49:20 +0000 (08:49 -0800)]
Merge branch 'for-linus' of git://git.infradead.org/users/vkoul/slave-dma

Pull dmaengine updates from Vinod Koul:
 "This update brings:

   - the big cleanup up by Maxime for device control and slave
     capabilities.  This makes the API much cleaner.

   - new IMG MDC driver by Andrew

   - new Renesas R-Car Gen2 DMA Controller driver by Laurent along with
     bunch of fixes on rcar drivers

   - odd fixes and updates spread over driver"

* 'for-linus' of git://git.infradead.org/users/vkoul/slave-dma: (130 commits)
  dmaengine: pl330: add DMA_PAUSE feature
  dmaengine: pl330: improve pl330_tx_status() function
  dmaengine: rcar-dmac: Disable channel 0 when using IOMMU
  dmaengine: rcar-dmac: Work around descriptor mode IOMMU errata
  dmaengine: rcar-dmac: Allocate hardware descriptors with DMAC device
  dmaengine: rcar-dmac: Fix oops due to unintialized list in error ISR
  dmaengine: rcar-dmac: Fix spinlock issues in interrupt
  dmaenegine: edma: fix sparse warnings
  dmaengine: rcar-dmac: Fix uninitialized variable usage
  dmaengine: shdmac: extend PM methods
  dmaengine: shdmac: use SET_RUNTIME_PM_OPS()
  dmaengine: pl330: fix bug that cause start the same descs in cyclic
  dmaengine: at_xdmac: allow muliple dwidths when doing slave transfers
  dmaengine: at_xdmac: simplify channel configuration stuff
  dmaengine: at_xdmac: introduce save_cc field
  dmaengine: at_xdmac: wait for in-progress transaction to complete after pausing a channel
  ioat: fail self-test if wait_for_completion times out
  dmaengine: dw: define DW_DMA_MAX_NR_MASTERS
  dmaengine: dw: amend description of dma_dev field
  dmatest: move src_off, dst_off, len inside loop
  ...

9 years agoMerge git://www.linux-watchdog.org/linux-watchdog
Linus Torvalds [Wed, 18 Feb 2015 16:42:47 +0000 (08:42 -0800)]
Merge git://www.linux-watchdog.org/linux-watchdog

Pull watchdog updates from Wim Van Sebroeck:
 "This adds the following new drivers:

   - ImgTec PDC Watchdog Timer Driver,
   - Mediatek SoC integrated watchdog

  Add support for BCM5301X, IT8783, NCT6791 and NCT6792 WDT's

  Add bcm47xx_wdt and da9063 restart handlers and contains overall
  improvements and fixes"

* git://www.linux-watchdog.org/linux-watchdog:
  watchdog: bcm47xx_wdt.c: allow enabling on BCM5301X arch
  watchdog: jz4740: Add DT support
  dt: watchdog: Add DT binding documentation for jz4740 watchdog timer
  watchdog: dw_wdt: Try to get a 30 second watchdog by default
  watchdog: dw_wdt: pat the watchdog before enabling it
  watchdog: w83627hf_wdt: Add support for NCT6791 and NCT6792
  watchdog: bcm47xx_wdt.c: add restart handler support
  watchdog: gpio_wdt: Add "always_running" feature to GPIO watchdog
  watchdog: da9063: Add restart handler support
  ARM: mediatek: dts: Add bindings for watchdog
  watchdog: Add driver for Mediatek watchdog
  watchdog: Fix omap watchdogs to enable the magic close bit
  watchdog: rt2880_wdt: minor clean up
  watchdog: hpwdt: Fix initialization message in hpwdt.c
  watchdog: it87_wdt: add IT8783 ID
  watchdog: imx2: Constify struct regmap_config and watchdog_ops
  DT: watchdog: Add ImgTec PDC Watchdog Timer binding documentation
  watchdog: ImgTec PDC Watchdog Timer Driver

9 years agoIB/mlx5: Enable the ODP capability query verb
Haggai Eran [Sun, 8 Feb 2015 11:28:52 +0000 (13:28 +0200)]
IB/mlx5: Enable the ODP capability query verb

Re-enable the on-demand paging capability query through the
extended query device verb.

Signed-off-by: Haggai Eran <haggaie@mellanox.com>
Reviewed-by: Yann Droneaud <ydroneaud@opteya.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoIB/core: Add on demand paging caps to ib_uverbs_ex_query_device
Haggai Eran [Sun, 8 Feb 2015 11:28:51 +0000 (13:28 +0200)]
IB/core: Add on demand paging caps to ib_uverbs_ex_query_device

Add on-demand paging capabilities reporting to the extended query device verb.

Yann Droneaud writes:

    Note: as offsetof() is used to retrieve the size of the lower chunk
    of the response, beware that it only works if the upper chunk
    is right after, without any implicit padding. And, as the size of
    the latter chunk is added to the base size, implicit padding at the
    end of the structure is not taken in account. Both point must be
    taken in account when extending the uverbs functionalities.

Signed-off-by: Haggai Eran <haggaie@mellanox.com>
Reviewed-by: Yann Droneaud <ydroneaud@opteya.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoIB/core: Add support for extended query device caps
Eli Cohen [Sun, 8 Feb 2015 11:28:50 +0000 (13:28 +0200)]
IB/core: Add support for extended query device caps

Add extensible query device capabilities verb to allow adding new features.
ib_uverbs_ex_query_device is added and copy_query_dev_fields is used to copy
capability fields to be used by both ib_uverbs_query_device and
ib_uverbs_ex_query_device.

Following the discussion about this patch [1], the code now validates
the command's comp_mask is zero, returning -EINVAL for unknown values,
in order to allow extending the verb in the future.

The verb also checks the user-space provided response buffer size and
only fills in capabilities that will fit in the buffer. In attempt to
follow the spirit of presentation [2] by Tzahi Oved that was presented
during OpenFabrics Alliance International Developer Workshop 2013, the
comp_mask bits will only describe which fields are valid.  Furthermore,
fields that can simply be cleared when they are not supported, do not
require a comp_mask bit at all.  The verb returns a response_length
field containing the actual number of bytes written by the kernel, so
that a newer version running on an older kernel can tell which fields
were actually returned.

[1] [PATCH v1 0/5] IB/core: extended query device caps cleanup for v3.19
    http://thread.gmane.org/gmane.linux.kernel.api/7889/

[2] https://www.openfabrics.org/images/docs/2013_Dev_Workshop/Tues_0423/2013_Workshop_Tues_0830_Tzahi_Oved-verbs_extensions_ofa_2013-tzahio.pdf

Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Haggai Eran <haggaie@mellanox.com>
Reviewed-by: Yann Droneaud <ydroneaud@opteya.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/cxgb4: Don't hang threads forever waiting on WR replies
Hariprasad S [Wed, 17 Dec 2014 08:41:03 +0000 (14:11 +0530)]
RDMA/cxgb4: Don't hang threads forever waiting on WR replies

In c4iw_wait_for_reply(), if a FW6_MSG WR reply is not received after
C4IW_WR_TO seconds, fail the WR operation and mark the device as fatally
dead.  Further, if the device is marked fatally dead, then fail the WR
wait immediately.

Also change the timeout to 60 seconds.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: Fix off by one in ocrdma_query_gid()
Dan Carpenter [Mon, 16 Feb 2015 10:01:36 +0000 (13:01 +0300)]
RDMA/ocrdma: Fix off by one in ocrdma_query_gid()

The ->sgid_tbl[] array has OCRDMA_MAX_SGID number of elements so this
test is off by one.  ->sgid_tbl is allocated in ocrdma_alloc_resources().

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Selvin Xavier <selvin.xavier@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: Use unsigned for bit index
Rasmus Villemoes [Fri, 16 Jan 2015 14:39:56 +0000 (15:39 +0100)]
RDMA/ocrdma: Use unsigned for bit index

In the expressions idx/32 and idx%32, both idx and 32 have signed
type, and unfortunately the C standard prescribes rounding to 0, so
unless gcc can prove that idx is non-negative, these cannot be
implemented as simple shift respectively mask operations. Help gcc by
changing the type of idx to unsigned - this cuts another few
instructions from the generated code.

Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Acked-by: Selvin Xavier <selvin.xavier@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: Help gcc generate better code for ocrdma_srq_toggle_bit
Rasmus Villemoes [Fri, 16 Jan 2015 14:39:55 +0000 (15:39 +0100)]
RDMA/ocrdma: Help gcc generate better code for ocrdma_srq_toggle_bit

gcc emits a surprising amount of code in order to flip a bit. One
would think that a single instruction is enough.

$ scripts/bloat-o-meter /tmp/ocrdma_verbs.o drivers/infiniband/hw/ocrdma/ocrdma_verbs.o
add/remove: 0/0 grow/shrink: 0/3 up/down: 0/-142 (-142)
function                                     old     new   delta
ocrdma_post_srq_recv                         498     460     -38
ocrdma_poll_cq                              2010    1962     -48
ocrdma_discard_cqes                          495     439     -56

All three calls of ocrdma_srq_toggle_bit happen within spinlocks, so
saving a few useless instructions might be worthwhile.

Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Acked-by: Selvin Xavier <selvin.xavier@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: Update the ocrdma module version string
Mitesh Ahuja [Thu, 18 Dec 2014 08:43:08 +0000 (14:13 +0530)]
RDMA/ocrdma: Update the ocrdma module version string

Signed-off-by: Mitesh Ahuja <mitesh.ahuja@emulex.com>
Signed-off-by: Devesh Sharma <devesh.sharma@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: set vlan present bit for user AH
Devesh Sharma [Thu, 18 Dec 2014 08:43:07 +0000 (14:13 +0530)]
RDMA/ocrdma: set vlan present bit for user AH

For the AH that describs a VLAN interface details, vlan present bit
needs to be set during posting a WQE. This patch adds the code to
allow it happening.

Signed-off-by: Mitesh Ahuja <mitesh.ahuja@emulex.com>
Signed-off-by: Devesh Sharma <devesh.sharma@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: remove reference of ocrdma_dev out of ocrdma_qp structure
Mitesh Ahuja [Thu, 18 Dec 2014 08:43:06 +0000 (14:13 +0530)]
RDMA/ocrdma: remove reference of ocrdma_dev out of ocrdma_qp structure

Use get_ocrdma_dev(ocrdma_qp->ibqp.device) function to access ocrdma
device pointer.

Signed-off-by: Mitesh Ahuja <mitesh.ahuja@emulex.com>
Signed-off-by: Devesh Sharma <devesh.sharma@emulex.com>
Signed-off-by: Selvin Xavier <selvin.xavier@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: Add support for interrupt moderation
Mitesh Ahuja [Thu, 18 Dec 2014 08:43:05 +0000 (14:13 +0530)]
RDMA/ocrdma: Add support for interrupt moderation

Add support for interrupt moderation for ocrdma device.  Thresholds
for high interrupt rates are static values derived based on experimental
results.

Signed-off-by: Mitesh Ahuja <mitesh.ahuja@emulex.com>
Signed-off-by: Devesh Sharma <devesh.sharma@emulex.com>
Signed-off-by: Selvin Xavier <selvin.xavier@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: Honor return value of ocrdma_resolve_dmac
Devesh Sharma [Thu, 18 Dec 2014 08:43:04 +0000 (14:13 +0530)]
RDMA/ocrdma: Honor return value of ocrdma_resolve_dmac

Check for return value for ocrdma_resolve_dmac while setting AV params.

Signed-off-by: Devesh Sharma <devesh.sharma@emulex.com>
Signed-off-by: Mitesh Ahuja <mitesh.ahuja@emulex.com>
Signed-off-by: Selvin Xavier <selvin.xavier@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: Allow expansion of the SQ CQEs via buddy CQ expansion of the QP
Selvin Xavier [Thu, 18 Dec 2014 08:43:03 +0000 (14:13 +0530)]
RDMA/ocrdma: Allow expansion of the SQ CQEs via buddy CQ expansion of the QP

If the SQ and RQ of the QP in error state uses separate CQs, traverse
the list of QPs using each CQs and invoke the buddy CQ handler for
both SQ and RQ.

Signed-off-by: Selvin Xavier <selvin.xavier@emulex.com>
Signed-off-by: Devesh Sharma <devesh.sharma@emulex.com>
Signed-off-by: Mitesh Ahuja <mitesh.ahuja@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: Discontinue support of RDMA-READ-WITH-INVALIDATE
Devesh Sharma [Thu, 18 Dec 2014 08:43:02 +0000 (14:13 +0530)]
RDMA/ocrdma: Discontinue support of RDMA-READ-WITH-INVALIDATE

Remove support for RDMA-READ-WITH-INVALIDATE from ocrdma driver.

Signed-off-by: Devesh Sharma <devesh.sharma@emulex.com>
Signed-off-by: Mitesh Ahuja <mitesh.ahuja@emulex.com>
Signed-off-by: Selvin Xavier <selvin.xavier@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
9 years agoRDMA/ocrdma: Host crash on destroying device resources
Mitesh Ahuja [Thu, 18 Dec 2014 08:43:01 +0000 (14:13 +0530)]
RDMA/ocrdma: Host crash on destroying device resources

1. Cleanup sequence in ocrdma_remove(). The device should be
   unregistered from IB stack before any device specific cleanup.
2. Always return success in the resource destroy path. In case destroy
   command returns error, IB stack will trigger cleanup again while
   closing the uverbs device causing kernel panic BUG_ON().

Signed-off-by: Selvin Xavier <selvin.xavier@emulex.com>
Signed-off-by: Mitesh Ahuja <mitesh.ahuja@emulex.com>
Signed-off-by: Devesh Sharma <devesh.sharma@emulex.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>