deliverable/linux.git
8 years agocpufreq: pxa2xx: fix pxa_cpufreq_change_voltage prototype
Arnd Bergmann [Mon, 25 Jan 2016 15:44:38 +0000 (16:44 +0100)] 
cpufreq: pxa2xx: fix pxa_cpufreq_change_voltage prototype

There are two definitions of pxa_cpufreq_change_voltage, with slightly
different prototypes after one of them had its argument marked 'const'.
Now the other one (for !CONFIG_REGULATOR) produces a harmless warning:

drivers/cpufreq/pxa2xx-cpufreq.c: In function 'pxa_set_target':
drivers/cpufreq/pxa2xx-cpufreq.c:291:36: warning: passing argument 1 of 'pxa_cpufreq_change_voltage' discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
   ret = pxa_cpufreq_change_voltage(&pxa_freq_settings[idx]);
                                    ^
drivers/cpufreq/pxa2xx-cpufreq.c:205:12: note: expected 'struct pxa_freqs *' but argument is of type 'const struct pxa_freqs *'
 static int pxa_cpufreq_change_voltage(struct pxa_freqs *pxa_freq)
            ^

This changes the prototype in the same way as the other, which
avoids the warning.

Fixes: 03c229906311 (cpufreq: pxa: make pxa_freqs arrays const)
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Cc: 4.2+ <stable@vger.kernel.org> # 4.2+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
8 years agoPM: APM_EMULATION does not depend on PM
Arnd Bergmann [Mon, 25 Jan 2016 15:48:28 +0000 (16:48 +0100)] 
PM: APM_EMULATION does not depend on PM

The APM emulation code does multiple things, and some of them depend on
PM_SLEEP, while the battery management does not. However, selecting
the symbol like SHARPSL_PM does causes a Kconfig warning:

warning: (SHARPSL_PM && PMAC_APM_EMU) selects APM_EMULATION which has unmet direct dependencies (PM && SYS_SUPPORTS_APM_EMULATION)

From all I can tell, this is completely harmless, and we can simply allow
APM_EMULATION to be enabled here, even if PM is not.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
8 years agocpufreq: Use list_is_last() to check last entry of the policy list
Gautham R Shenoy [Wed, 27 Jan 2016 06:32:26 +0000 (12:02 +0530)] 
cpufreq: Use list_is_last() to check last entry of the policy list

Currently next_policy() explicitly checks if a policy is the last
policy in the cpufreq_policy_list. Use the standard list_is_last
primitive instead.

Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
8 years agocpufreq: Fix NULL reference crash while accessing policy->governor_data
Viresh Kumar [Mon, 25 Jan 2016 17:03:46 +0000 (22:33 +0530)] 
cpufreq: Fix NULL reference crash while accessing policy->governor_data

There is a race discovered by Juri, where we are able to:
- create and read a sysfs file before policy->governor_data is being set
  to a non NULL value.
  OR
- set policy->governor_data to NULL, and reading a file before being
  destroyed.

And so such a crash is reported:

Unable to handle kernel NULL pointer dereference at virtual address 0000000c
pgd = edfc8000
[0000000c] *pgd=bfc8c835
Internal error: Oops: 17 [#1] SMP ARM
Modules linked in:
CPU: 4 PID: 1730 Comm: cat Not tainted 4.5.0-rc1+ #463
Hardware name: ARM-Versatile Express
task: ee8e8480 ti: ee930000 task.ti: ee930000
PC is at show_ignore_nice_load_gov_pol+0x24/0x34
LR is at show+0x4c/0x60
pc : [<c058f1bc>]    lr : [<c058ae88>]    psr: a0070013
sp : ee931dd0  ip : ee931de0  fp : ee931ddc
r10: ee4bc290  r9 : 00001000  r8 : ef2cb000
r7 : ee4bc200  r6 : ef2cb000  r5 : c0af57b0  r4 : ee4bc2e0
r3 : 00000000  r2 : 00000000  r1 : c0928df4  r0 : ef2cb000
Flags: NzCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment none
Control: 10c5387d  Table: adfc806a  DAC: 00000051
Process cat (pid: 1730, stack limit = 0xee930210)
Stack: (0xee931dd0 to 0xee932000)
1dc0:                                     ee931dfc ee931de0 c058ae88 c058f1a4
1de0: edce3bc0 c07bfca4 edce3ac0 00001000 ee931e24 ee931e00 c01fcb90 c058ae48
1e00: 00000001 edce3bc0 00000000 00000001 ee931e50 ee8ff480 ee931e34 ee931e28
1e20: c01fb33c c01fcb0c ee931e8c ee931e38 c01a5210 c01fb314 ee931e9c ee931e48
1e40: 00000000 edce3bf0 befe4a00 ee931f78 00000000 00000000 000001e4 00000000
1e60: c00545a8 edce3ac0 00001000 00001000 befe4a00 ee931f78 00000000 00001000
1e80: ee931ed4 ee931e90 c01fbed8 c01a5038 ed085a58 00020000 00000000 00000000
1ea0: c0ad72e4 ee931f78 ee8ff488 ee8ff480 c077f3fc 00001000 befe4a00 ee931f78
1ec0: 00000000 00001000 ee931f44 ee931ed8 c017c328 c01fbdc4 00001000 00000000
1ee0: ee8ff480 00001000 ee931f44 ee931ef8 c017c65c c03deb10 ee931fac ee931f08
1f00: c0009270 c001f290 c0a8d968 ef2cb000 ef2cb000 ee8ff480 00000020 ee8ff480
1f20: ee8ff480 befe4a00 00001000 ee931f78 00000000 00000000 ee931f74 ee931f48
1f40: c017d1ec c017c2f8 c019c724 c019c684 ee8ff480 ee8ff480 00001000 befe4a00
1f60: 00000000 00000000 ee931fa4 ee931f78 c017d2a8 c017d160 00000000 00000000
1f80: 000a9f20 00001000 befe4a00 00000003 c000ffe4 ee930000 00000000 ee931fa8
1fa0: c000fe40 c017d264 000a9f20 00001000 00000003 befe4a00 00001000 00000000
Unable to handle kernel NULL pointer dereference at virtual address 0000000c
1fc0: 000a9f20 00001000 befe4a00 00000003 00000000 00000000 00000003 00000001
pgd = edfc4000
[0000000c] *pgd=bfcac835
1fe0: 00000000 befe49dc 000197f8 b6e35dfc 60070010 00000003 3065b49d 134ac2c9

[<c058f1bc>] (show_ignore_nice_load_gov_pol) from [<c058ae88>] (show+0x4c/0x60)
[<c058ae88>] (show) from [<c01fcb90>] (sysfs_kf_seq_show+0x90/0xfc)
[<c01fcb90>] (sysfs_kf_seq_show) from [<c01fb33c>] (kernfs_seq_show+0x34/0x38)
[<c01fb33c>] (kernfs_seq_show) from [<c01a5210>] (seq_read+0x1e4/0x4e4)
[<c01a5210>] (seq_read) from [<c01fbed8>] (kernfs_fop_read+0x120/0x1a0)
[<c01fbed8>] (kernfs_fop_read) from [<c017c328>] (__vfs_read+0x3c/0xe0)
[<c017c328>] (__vfs_read) from [<c017d1ec>] (vfs_read+0x98/0x104)
[<c017d1ec>] (vfs_read) from [<c017d2a8>] (SyS_read+0x50/0x90)
[<c017d2a8>] (SyS_read) from [<c000fe40>] (ret_fast_syscall+0x0/0x1c)
Code: e5903044 e1a00001 e3081df4 e34c1092 (e593300c)
---[ end trace 5994b9a5111f35ee ]---

Fix that by making sure, policy->governor_data is updated at the right
places only.

Cc: 4.2+ <stable@vger.kernel.org> # 4.2+
Reported-and-tested-by: Juri Lelli <juri.lelli@arm.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
8 years agocpuidle: coupled: remove unused define cpuidle_coupled_lock
Anders Roxell [Wed, 27 Jan 2016 19:26:54 +0000 (20:26 +0100)] 
cpuidle: coupled: remove unused define cpuidle_coupled_lock

This was found with the -RT patch enabled, but the fix should apply to
non-RT also.

Used multi_v7_defconfig+PREEMPT_RT_FULL=y and this caused a compilation
warning without this fix:
../drivers/cpuidle/coupled.c:122:21: warning: 'cpuidle_coupled_lock'
defined but not used [-Wunused-variable]

Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
8 years agoPM / Domains: Fix typo in comment
Moritz Fischer [Wed, 27 Jan 2016 07:29:27 +0000 (08:29 +0100)] 
PM / Domains: Fix typo in comment

Acked-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: Moritz Fischer <moritz.fischer@ettus.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
8 years agoPM / Domains: Fix potential deadlock while adding/removing subdomains
Ulf Hansson [Wed, 27 Jan 2016 08:10:37 +0000 (09:10 +0100)] 
PM / Domains: Fix potential deadlock while adding/removing subdomains

We must preserve the same order of how we acquire and release the lock for
genpd, as otherwise we may encounter deadlocks.

The power on phase of a genpd starts by acquiring its lock. Then it walks
the hierarchy of its parent domains to be able to power on these first, as
per design of genpd.

From a locking perspective this means the locks of the parents becomes
acquired after the lock of the subdomain.

Let's fix pm_genpd_add|remove_subdomain() to maintain the same order of
acquiring/releasing the genpd lock as being applied in the power on/off
sequence.

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
8 years agoMerge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Linus Torvalds [Wed, 27 Jan 2016 19:56:03 +0000 (11:56 -0800)] 
Merge tag 'for_linus' of git://git./linux/kernel/git/mst/vhost

Pull virtio tests and fixes from Michael Tsirkin:
 "This fixes existing tests broken by barrier rework, and adds some new
  tests.

  Plus, there's a fix for an old bug in virtio-pci"

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  tools/virtio: add ringtest utilities
  sh: fix smp_store_mb for !SMP
  tools/virtio: use virt_xxx barriers
  virtio_pci: fix use after free on release

8 years agoRevert "MIPS: bcm63xx: nvram: Remove unused bcm63xx_nvram_get_psi_size() function"
Ralf Baechle [Wed, 27 Jan 2016 19:50:36 +0000 (20:50 +0100)] 
Revert "MIPS: bcm63xx: nvram: Remove unused bcm63xx_nvram_get_psi_size() function"

This reverts commit 5bdb102b3f9785cb88467bc7c75fa0f5cacc8dc5.

Brian Norris <computersforpeace@gmail.com> is reporting:

Ralf,

Please revert this and send it to Linus (or else, I can send it myself).
This is causing build failures, because I didn't take the rest of
Simon's series yet.

drivers/mtd/bcm63xxpart.c: In function 'bcm63xx_parse_cfe_partitions':
drivers/mtd/bcm63xxpart.c:93:2: error: implicit declaration of function
'bcm63xx_nvram_get_psi_size' [-Werror=implicit-function-declaration]

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
References: https://www.linux-mips.org/cgi-bin/mesg.cgi?a=linux-mips&i=20160126191607.GA111152%40google.com

8 years agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Wed, 27 Jan 2016 19:08:30 +0000 (11:08 -0800)] 
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fixes from Will Deacon:
 "The main thing here is to get us booting under qemu again after it
  turned out that not all of the PMU registers are emulated there,
  causing us to die early in boot.

   - Ensure we don't access PMU registers of the PMU is not implemented
     (fixes booting under QEMU)

   - Fix BUG_ON triggered during module loading with DEBUG_SET_MODULE_RONX

   - Ensure the kasan zero page is read-only

   - Hide __efistub_ symbol aliases from kallsyms, since they otherwise
     confuse the backtrace code

   - Ensure !PTE_WRITE kernel ptes are marked as read-only

   - defconfig updates based on requests and patches on the list

   - Other minor fixes (typos, build system)"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: mm: avoid calling apply_to_page_range on empty range
  arm64: defconfig: updates for 4.5
  arm64: errata: Add -mpc-relative-literal-loads to build flags
  Eliminate the .eh_frame sections from the aarch64 vmlinux and kernel modules
  arm64: Fix an enum typo in mm/dump.c
  arm64: Honour !PTE_WRITE in set_pte_at() for kernel mappings
  arm64: kernel: fix architected PMU registers unconditional access
  arm64: kasan: ensure that the KASAN zero page is mapped read-only
  arm64: hide __efistub_ aliases from kallsyms

8 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Wed, 27 Jan 2016 18:50:42 +0000 (10:50 -0800)] 
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull KVM fixes from Paolo Bonzini:
 "s390 and POWER bug fixes, plus enabling the KVM-VFIO interface on
  s390"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM doc: Fix KVM_SMI chapter number
  KVM: s390: fix memory overwrites when vx is disabled
  KVM: s390: Enable the KVM-VFIO device
  KVM: s390: fix guest fprs memory leak
  KVM: PPC: Fix ONE_REG AltiVec support
  KVM: PPC: Increase memslots to 512
  KVM: PPC: Book3S PR: Remove unused variable 'vcpu_book3s'
  KVM: PPC: Fix emulation of H_SET_DABR/X on POWER8
  KVM: PPC: Book3S HV: Handle unexpected traps in guest entry/exit code better

8 years agomailmap: redirect inactive address <ao2@amarulasolutions.com>
Antonio Ospite [Wed, 27 Jan 2016 12:37:34 +0000 (13:37 +0100)] 
mailmap: redirect inactive address <ao2@amarulasolutions.com>

The email address <ao2@amarulasolutions.com> is not active anymore, use
Antonio's personal email address <ao2@ao2.it> in case someone wants to
get in touch for the code wrote for Amarula Solutions.

Signed-off-by: Antonio Ospite <ao2@ao2.it>
Cc: Michael Trimarchi <michael@amarulasolutions.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
8 years agovfio/noiommu: Don't use iommu_present() to track fake groups
Alex Williamson [Wed, 27 Jan 2016 18:22:25 +0000 (11:22 -0700)] 
vfio/noiommu: Don't use iommu_present() to track fake groups

Using iommu_present() to determine whether an IOMMU group is real or
fake has some problems.  First, apparently Power systems don't
register an IOMMU on the device bus, so the groups and containers get
marked as noiommu and then won't bind to their actual IOMMU driver.
Second, I expect we'll run into the same issue as we try to support
vGPUs through vfio, since they're likely to emulate this behavior of
creating an IOMMU group on a virtual device and then providing a vfio
IOMMU backend tailored to the sort of isolation they provide, which
won't necessarily be fully compatible with the IOMMU API.

The solution here is to use the existing iommudata interface to IOMMU
groups, which allows us to easily identify the fake groups we've
created for noiommu purposes.  The iommudata we set is purely
arbitrary since we're only comparing the address, so we use the
address of the noiommu switch itself.

Reported-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Tested-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Tested-by: Anatoly Burakov <anatoly.burakov@intel.com>
Tested-by: Santosh Shukla <sshukla@mvista.com>
Fixes: 03a76b60f8ba ("vfio: Include No-IOMMU mode")
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
8 years agoARM: wire up copy_file_range() syscall
Russell King [Wed, 27 Jan 2016 18:14:13 +0000 (18:14 +0000)] 
ARM: wire up copy_file_range() syscall

Add the copy_file_range() syscall to ARM.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
8 years agodrm/radeon: fix DP audio support for APU with DCE4.1 display engine
Slava Grigorev [Tue, 26 Jan 2016 22:35:57 +0000 (17:35 -0500)] 
drm/radeon: fix DP audio support for APU with DCE4.1 display engine

Properly setup the DFS divider for DP audio for DCE4.1.

Signed-off-by: Slava Grigorev <slava.grigorev@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
8 years agodrm/radeon: Add a common function for DFS handling
Slava Grigorev [Tue, 26 Jan 2016 21:56:25 +0000 (16:56 -0500)] 
drm/radeon: Add a common function for DFS handling

Move encoding of DFS (digital frequency synthesizer) divider into a
separate function and improve calculation precision.

Signed-off-by: Slava Grigorev <slava.grigorev@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
8 years agodrm/radeon: cleaned up VCO output settings for DP audio
Slava Grigorev [Tue, 26 Jan 2016 21:45:10 +0000 (16:45 -0500)] 
drm/radeon: cleaned up VCO output settings for DP audio

This is preparation for the fixes in the following patches.

Signed-off-by: Slava Grigorev <slava.grigorev@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
8 years agodrm/amd/powerplay: Update SMU firmware loading for Stoney
Rex Zhu [Thu, 21 Jan 2016 11:24:44 +0000 (19:24 +0800)] 
drm/amd/powerplay: Update SMU firmware loading for Stoney

Fix firmware init on Stoney when powerplay is enabled.

Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
8 years agoseccomp: always propagate NO_NEW_PRIVS on tsync
Jann Horn [Sat, 26 Dec 2015 05:00:48 +0000 (06:00 +0100)] 
seccomp: always propagate NO_NEW_PRIVS on tsync

Before this patch, a process with some permissive seccomp filter
that was applied by root without NO_NEW_PRIVS was able to add
more filters to itself without setting NO_NEW_PRIVS by setting
the new filter from a throwaway thread with NO_NEW_PRIVS.

Signed-off-by: Jann Horn <jann@thejh.net>
Cc: stable@vger.kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
8 years agobtrfs: don't use GFP_HIGHMEM for free-space-tree bitmap kzalloc
Chris Mason [Wed, 27 Jan 2016 14:38:45 +0000 (06:38 -0800)] 
btrfs: don't use GFP_HIGHMEM for free-space-tree bitmap kzalloc

This was copied incorrectly from the __vmalloc call.

Signed-off-by: Chris Mason <clm@fb.com>
8 years agodebugobjects: Allow bigger number of early boot objects
Christian Borntraeger [Wed, 27 Jan 2016 14:37:58 +0000 (15:37 +0100)] 
debugobjects: Allow bigger number of early boot objects

On my bigger s390 systems  I always get "Out of memory.
ODEBUG disabled". Since the number of objects is needed at
compile time, we can not change the size dynamically before
the caches etc are available. Doubling the size seems to
do the trick. Since it is init data it will be freed anyway,
this should be ok.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Link: http://lkml.kernel.org/r/1453905478-13409-1-git-send-email-borntraeger@de.ibm.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agoinclude/linux/cleancache.h: Clean up code
Chen Gang [Wed, 6 Jan 2016 21:06:13 +0000 (05:06 +0800)] 
include/linux/cleancache.h: Clean up code

Let cleancache_fs_enabled() call cleancache_fs_enabled_mapping()
directly.

Remove redundant variable ret in cleancache_get_page().

Signed-off-by: Chen Gang <gang.chen.5i5j@gmail.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
8 years agocleancache: constify cleancache_ops structure
Julia Lawall [Thu, 21 Jan 2016 15:47:29 +0000 (16:47 +0100)] 
cleancache: constify cleancache_ops structure

The cleancache_ops structure is never modified, so declare it as const.

Done with the help of Coccinelle.

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
8 years agoMerge branch 'dev/fst-followup' of git://git.kernel.org/pub/scm/linux/kernel/git...
Chris Mason [Wed, 27 Jan 2016 13:48:23 +0000 (05:48 -0800)] 
Merge branch 'dev/fst-followup' of git://git./linux/kernel/git/kdave/linux into for-linus-4.5

8 years agobtrfs: sysfs: check initialization state before updating features
David Sterba [Wed, 27 Jan 2016 13:06:29 +0000 (14:06 +0100)] 
btrfs: sysfs: check initialization state before updating features

If the mount phase is not finished, we can't update the sysfs files.

Reported-by: Chris Mason <clm@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
8 years agoirqchip/atmel-aic: Fix wrong bit operation for IRQ priority
Milo Kim [Wed, 13 Jan 2016 07:19:50 +0000 (16:19 +0900)] 
irqchip/atmel-aic: Fix wrong bit operation for IRQ priority

Atmel AIC has common structure for SMR (Source Mode Register).

  bit[6:5] Interrupt source type
  bit[2:0] Priority level
  Other bits are unused.

To update new priority value, bit[2:0] should be cleared first and then
new priority level can be written. However, aic_common_set_priority()
helper clears source type bits instead of priority bits.
This patch fixes wrong mask bit operation.

Fixes: b1479ebb7720 "irqchip: atmel-aic: Add atmel AIC/AIC5 drivers"
Signed-off-by: Milo Kim <milo.kim@ti.com>
Acked-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Cc: Jason Cooper <jason@lakedaemon.net>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Cc: Ludovic Desroches <ludovic.desroches@atmel.com>
Cc: Nicholas Ferre <nicolas.ferre@atmel.com>
Cc: stable@vger.kernel.org #v3.17+
Link: http://lkml.kernel.org/r/1452669592-3401-2-git-send-email-milo.kim@ti.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agotick/nohz: Set the correct expiry when switching to nohz/lowres mode
Wanpeng Li [Wed, 27 Jan 2016 11:26:07 +0000 (19:26 +0800)] 
tick/nohz: Set the correct expiry when switching to nohz/lowres mode

commit 0ff53d096422 sets the next tick interrupt to the last jiffies update,
i.e. in the past, because the forward operation is invoked before the set
operation. There is no resulting damage (yet), but we get an extra pointless
tick interrupt.

Revert the order so we get the next tick interrupt in the future.

Fixes: commit 0ff53d096422 "tick: sched: Force tick interrupt and get rid of softirq magic"
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/1453893967-3458-1-git-send-email-wanpeng.li@hotmail.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agostaging/speakup: Use tty_ldisc_ref() for paste kworker
Peter Hurley [Mon, 11 Jan 2016 06:40:58 +0000 (22:40 -0800)] 
staging/speakup: Use tty_ldisc_ref() for paste kworker

As the function documentation for tty_ldisc_ref_wait() notes, it is
only callable from a tty file_operations routine; otherwise there
is no guarantee the ref won't be NULL.

The key difference with the VT's paste_selection() is that is an ioctl,
where __speakup_paste_selection() is completely async kworker, kicked
off from interrupt context.

Fixes: 28a821c30688 ("Staging: speakup: Update __speakup_paste_selection()
       tty (ab)usage to match vt")
Cc: <stable@vger.kernel.org>
Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 years agon_tty: Fix unsafe reference to "other" ldisc
Peter Hurley [Mon, 11 Jan 2016 06:40:56 +0000 (22:40 -0800)] 
n_tty: Fix unsafe reference to "other" ldisc

Although n_tty_check_unthrottle() has a valid ldisc reference (since
the tty core gets the ldisc ref in tty_read() before calling the line
discipline read() method), it does not have a valid ldisc reference to
the "other" pty of a pty pair. Since getting an ldisc reference for
tty->link essentially open-codes tty_wakeup(), just replace with the
equivalent tty_wakeup().

Cc: <stable@vger.kernel.org>
Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 years agotty: Fix unsafe ldisc reference via ioctl(TIOCGETD)
Peter Hurley [Mon, 11 Jan 2016 06:40:55 +0000 (22:40 -0800)] 
tty: Fix unsafe ldisc reference via ioctl(TIOCGETD)

ioctl(TIOCGETD) retrieves the line discipline id directly from the
ldisc because the line discipline id (c_line) in termios is untrustworthy;
userspace may have set termios via ioctl(TCSETS*) without actually
changing the line discipline via ioctl(TIOCSETD).

However, directly accessing the current ldisc via tty->ldisc is
unsafe; the ldisc ptr dereferenced may be stale if the line discipline
is changing via ioctl(TIOCSETD) or hangup.

Wait for the line discipline reference (just like read() or write())
to retrieve the "current" line discipline id.

Cc: <stable@vger.kernel.org>
Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 years agotty: Retry failed reopen if tty teardown in-progress
Peter Hurley [Sun, 10 Jan 2016 05:13:45 +0000 (21:13 -0800)] 
tty: Retry failed reopen if tty teardown in-progress

A small window exists where a tty reopen will observe the tty
just prior to imminent teardown (tty->count == 0); in this case, open()
returns EIO to userspace.

Instead, retry the open after checking for signals and yielding;
this interruptible retry loop allows teardown to commence and initialize
a new tty on retry. Never retry the BSD master pty reopen; there is no
guarantee the pty pair teardown is imminent since the slave file
descriptors may remain open indefinitely.

Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
Cc: stable <stable@vger.kernel.org> # 4.4
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 years agotty: Wait interruptibly for tty lock on reopen
Peter Hurley [Sun, 10 Jan 2016 05:13:44 +0000 (21:13 -0800)] 
tty: Wait interruptibly for tty lock on reopen

Allow a signal to interrupt the wait for a tty reopen; eg., if
the tty has starting final close and is waiting for the device to
drain.

Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
Cc: stable <stable@vger.kernel.org> # 4.4
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 years agoALSA: Add missing dependency on CONFIG_SND_TIMER
Takashi Iwai [Wed, 27 Jan 2016 06:05:56 +0000 (07:05 +0100)] 
ALSA: Add missing dependency on CONFIG_SND_TIMER

Since the build of PCM timer may be disabled via Kconfig now, each
driver that provides a timer interface needs to set CONFIG_SND_TIMER
explicitly.  Otherwise it may get a build error due to missing
symbol.

Fixes: 90bbaf66ee7b ('ALSA: timer: add config item to export PCM timer disabling for expert')
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Cc: <stable@vger.kernel.org> # v4.4+
Signed-off-by: Takashi Iwai <tiwai@suse.de>
8 years agoMerge remote-tracking branch 'mkp-scsi/4.5/scsi-fixes' into fixes
James Bottomley [Wed, 27 Jan 2016 01:44:42 +0000 (17:44 -0800)] 
Merge remote-tracking branch 'mkp-scsi/4.5/scsi-fixes' into fixes

8 years agoSCSI: fix crashes in sd and sr runtime PM
Alan Stern [Wed, 20 Jan 2016 16:26:01 +0000 (11:26 -0500)] 
SCSI: fix crashes in sd and sr runtime PM

Runtime suspend during driver probe and removal can cause problems.
The driver's runtime_suspend or runtime_resume callbacks may invoked
before the driver has finished binding to the device or after the
driver has unbound from the device.

This problem shows up with the sd and sr drivers, and can cause disk
or CD/DVD drives to become unusable as a result.  The fix is simple.
The drivers store a pointer to the scsi_disk or scsi_cd structure as
their private device data when probing is finished, so we simply have
to be sure to clear the private data during removal and test it during
runtime suspend/resume.

This fixes <https://bugs.debian.org/801925>.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reported-by: Paul Menzel <paul.menzel@giantmonkey.de>
Reported-by: Erich Schubert <erich@debian.org>
Reported-by: Alexandre Rossi <alexandre.rossi@gmail.com>
Tested-by: Paul Menzel <paul.menzel@giantmonkey.de>
Tested-by: Erich Schubert <erich@debian.org>
CC: <stable@vger.kernel.org>
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
8 years agopowerpc/eeh: Fix PE location code
Gavin Shan [Wed, 2 Dec 2015 05:25:32 +0000 (16:25 +1100)] 
powerpc/eeh: Fix PE location code

In eeh_pe_loc_get(), the PE location code is retrieved from the
"ibm,loc-code" property of the device node for the bridge of the
PE's primary bus. It's not correct because the property indicates
the parent PE's location code.

This reads the correct PE location code from "ibm,io-base-loc-code"
or "ibm,slot-location-code" property of PE parent bus's device node.

Cc: stable@vger.kernel.org # v3.16+
Fixes: 357b2f3dd9b7 ("powerpc/eeh: Dump PE location code")
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Tested-by: Russell Currey <ruscur@russell.cc>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
8 years agoARM: 8500/1: fix atags_to_fdt with stack-protector-strong
Kees Cook [Tue, 26 Jan 2016 00:18:13 +0000 (01:18 +0100)] 
ARM: 8500/1: fix atags_to_fdt with stack-protector-strong

Building with CONFIG_CC_STACKPROTECTOR_STRONG triggers protection code
generation under CONFIG_ARM_ATAG_DTB_COMPAT but this is too early for
being able to use any of the stack_chk code. Explicitly disable it for
only the atags_to_fdt bits.

Suggested-by: zhxihu <zhxihu@marvell.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
8 years agox86/mm: Fix types used in pgprot cacheability flags translations
Jan Beulich [Tue, 26 Jan 2016 11:15:18 +0000 (04:15 -0700)] 
x86/mm: Fix types used in pgprot cacheability flags translations

For PAE kernels "unsigned long" is not suitable to hold page protection
flags, since _PAGE_NX doesn't fit there. This is the reason for quite a
few W+X pages getting reported as insecure during boot (observed namely
for the entire initrd range).

Fixes: 281d4078be ("x86: Make page cache mode a real type")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Juergen Gross <JGross@suse.com>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/56A7635602000078000CAFF1@prv-mh.provo.novell.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agodrm/etnaviv: call correct function when trying to vmap a DMABUF
Lucas Stach [Mon, 25 Jan 2016 14:47:28 +0000 (15:47 +0100)] 
drm/etnaviv: call correct function when trying to vmap a DMABUF

When trying to get the vmap address of an imported buffer, we must
call into the appropriate helper function, to allow the exporter to
establish the vmap, instead of trying to vmap the buffer on our own.

Add an indirection through etnaviv_gem_ops to allow the correct
implementation to be called.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agodrm/etnaviv: rename etnaviv_gem_vaddr to etnaviv_gem_vmap
Lucas Stach [Tue, 26 Jan 2016 17:10:32 +0000 (18:10 +0100)] 
drm/etnaviv: rename etnaviv_gem_vaddr to etnaviv_gem_vmap

This function follows the semantics of vmap() by returning
NULL in case of an error. To make things less confusing
rename it to make make both functions more closely related.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agodrm/etnaviv: fix get pages error path in etnaviv_gem_vaddr
Lucas Stach [Mon, 25 Jan 2016 14:37:28 +0000 (15:37 +0100)] 
drm/etnaviv: fix get pages error path in etnaviv_gem_vaddr

In case that etnaviv_gem_get_pages is unable to get the required
pages the object mutex needs to be unlocked. Also return NULL in
this case instead of propagating the error, as callers of this
function might not be prepared to handle a pointer error, but
expect this call to follow the semantics of a plain vmap to return
NULL in case of an error.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agodrm/etnaviv: fix memory leak in IOMMU init path
Lucas Stach [Mon, 25 Jan 2016 11:41:05 +0000 (12:41 +0100)] 
drm/etnaviv: fix memory leak in IOMMU init path

Plug in error handling to free any allocated ressources in the
IOMMU init path.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agodrm/etnaviv: add further minor features and varyings count
Russell King [Sun, 24 Jan 2016 17:36:04 +0000 (17:36 +0000)] 
drm/etnaviv: add further minor features and varyings count

Export further minor feature bitmasks and the varyings count from
the GPU specifications registers to userspace.

Acked-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agodrm/etnaviv: add helper for comparing model/revision IDs
Russell King [Sun, 24 Jan 2016 17:35:59 +0000 (17:35 +0000)] 
drm/etnaviv: add helper for comparing model/revision IDs

Add and use a helper for comparing the model and revision IDs.

Acked-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agodrm/etnaviv: add helper to extract bitfields
Russell King [Sun, 24 Jan 2016 17:35:54 +0000 (17:35 +0000)] 
drm/etnaviv: add helper to extract bitfields

Add a helper to extract etnaviv bitfields from register values.

Acked-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agodrm/etnaviv: use defined constants for the chip model
Russell King [Sun, 24 Jan 2016 17:35:48 +0000 (17:35 +0000)] 
drm/etnaviv: use defined constants for the chip model

Use the defined constants in common.xml.h for the chip model rather
than coding these as hex numbers.

Acked-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agodrm/etnaviv: update common and state_hi xml.h files
Russell King [Sun, 24 Jan 2016 17:35:43 +0000 (17:35 +0000)] 
drm/etnaviv: update common and state_hi xml.h files

Update the common and state_hi xml.h header files from the etnaviv
repository.

Acked-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agodrm/etnaviv: ignore VG GPUs with FE2.0
Russell King [Thu, 21 Jan 2016 15:19:59 +0000 (15:19 +0000)] 
drm/etnaviv: ignore VG GPUs with FE2.0

Ignore GPUs with a 2.0 front end.  These have a different register
layout for the front end, which provokes imprecise aborts from the
register accesses in the 'gpu' debugfs file.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
8 years agobase/platform: Fix platform drivers with no probe callback
Martin Wilck [Mon, 30 Nov 2015 11:50:05 +0000 (12:50 +0100)] 
base/platform: Fix platform drivers with no probe callback

Since b8b2c7d845d5, platform_drv_probe() is called for all platform
devices. If drv->probe is NULL, and dev_pm_domain_attach() fails,
platform_drv_probe() will return the error code from dev_pm_domain_attach().

This causes real_probe() to enter the "probe_failed" path and set
dev->driver to NULL. Before b8b2c7d845d5, real_probe() would assume
success if both dev->bus->probe and drv->probe were missing. As a result,
a device and driver could be "bound" together just by matching their names;
this doesn't work any more after b8b2c7d845d5.

This may cause problems later for certain usage of platform_driver_register()
and platform_device_register_simple(). I observed a panic while loading
the tpm_tis driver with parameter "force=1" (i.e. registering tpm_tis as
a platform driver), because tpm_tis_init's assumption that the device
returned by platform_device_register_simple() was bound didn't hold any more
(tpmm_chip_alloc() dereferences chip->pdev->driver, causing panic).

This patch restores the previous (4.3.0 and earlier) behavior of
platform_drv_probe() in the case when the associated platform driver has
no "probe" function.

Fixes: b8b2c7d845d5 ("base/platform: assert that dev_pm_domain callbacks are called unconditionally")
Signed-off-by: Martin Wilck <Martin.Wilck@ts.fujitsu.com>
Cc: stable <stable@vger.kernel.org> # 4.4
Cc: Martin Fuzzey <mfuzzey@parkeon.com>
Acked-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
8 years agoarm64: mm: avoid calling apply_to_page_range on empty range
Mika Penttilä [Tue, 26 Jan 2016 15:47:25 +0000 (15:47 +0000)] 
arm64: mm: avoid calling apply_to_page_range on empty range

Calling apply_to_page_range with an empty range results in a BUG_ON
from the core code. This can be triggered by trying to load the st_drv
module with CONFIG_DEBUG_SET_MODULE_RONX enabled:

  kernel BUG at mm/memory.c:1874!
  Internal error: Oops - BUG: 0 [#1] PREEMPT SMP
  Modules linked in:
  CPU: 3 PID: 1764 Comm: insmod Not tainted 4.5.0-rc1+ #2
  Hardware name: ARM Juno development board (r0) (DT)
  task: ffffffc9763b8000 ti: ffffffc975af8000 task.ti: ffffffc975af8000
  PC is at apply_to_page_range+0x2cc/0x2d0
  LR is at change_memory_common+0x80/0x108

This patch fixes the issue by making change_memory_common (called by the
set_memory_* functions) a NOP when numpages == 0, therefore avoiding the
erroneous call to apply_to_page_range and bringing us into line with x86
and s390.

Cc: <stable@vger.kernel.org>
Reviewed-by: Laura Abbott <labbott@redhat.com>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Mika Penttilä <mika.penttila@nextfour.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
8 years agoirqchip/gic-v3-its: Recompute the number of pages on page size change
Marc Zyngier [Tue, 26 Jan 2016 14:24:15 +0000 (14:24 +0000)] 
irqchip/gic-v3-its: Recompute the number of pages on page size change

When the programming of a GITS_BASERn register fails because of
an unsupported ITS page size, we retry it with a smaller page size.
Unfortunately, we don't recompute the number of allocated ITS pages,
indicating the wrong value computed in the original allocation.

A convenient fix is to free the pages we allocated, update the
page size, and restart the allocation. This will ensure that
we always allocate the right amount in the case of a device
table, specially if we have to reduce the allocation order
to stay within the boundaries of the ITS maximum allocation.

Reported-and-tested-by: Ma Jun <majun258@huawei.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: Jason Cooper <jason@lakedaemon.net>
Link: http://lkml.kernel.org/r/1453818255-1289-1-git-send-email-marc.zyngier@arm.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agobase: Export platform_msi_domain_[alloc,free]_irqs
Thomas Petazzoni [Tue, 26 Jan 2016 13:52:27 +0000 (13:52 +0000)] 
base: Export platform_msi_domain_[alloc,free]_irqs

The new function platform_msi_domain_{alloc,free}_irqs are meant to be
used in platform drivers, which can be built as modules. Therefore, it
makes sense to export them to be used from kernel modules.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Frank Rowand <frowand.list@gmail.com>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Jiang Liu <jiang.liu@linux.intel.com>
Link: http://lkml.kernel.org/r/1453816347-32720-4-git-send-email-marc.zyngier@arm.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agoKVM doc: Fix KVM_SMI chapter number
Alexey Kardashevskiy [Tue, 19 Jan 2016 05:12:28 +0000 (16:12 +1100)] 
KVM doc: Fix KVM_SMI chapter number

The KVM_SMI capability is following the KVM_S390_SET_IRQ_STATE capability
which is "4.95", this changes the number of the KVM_SMI chapter to 4.96.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
8 years agoMerge tag 'kvm-s390-master-4.5-1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Paolo Bonzini [Tue, 26 Jan 2016 15:28:36 +0000 (16:28 +0100)] 
Merge tag 'kvm-s390-master-4.5-1' of git://git./linux/kernel/git/kvms390/linux into HEAD

KVM: s390: Fixes for kvm/master (targeting 4.5)

1. Fallout of some bigger floating point/vector rework in s390
- memory leak -> stable 4.3+
- memory overwrite -> stable 4.4+

2. enable KVM-VFIO for s390

8 years agoclocksource: Fix dependencies for archs w/o HAS_IOMEM
Richard Weinberger [Mon, 25 Jan 2016 22:24:19 +0000 (23:24 +0100)] 
clocksource: Fix dependencies for archs w/o HAS_IOMEM

Not every arch has io memory.
So, unbreak the build by fixing the dependencies.

Signed-off-by: Richard Weinberger <richard@nod.at>
Cc: user-mode-linux-devel@lists.sourceforge.net
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: http://lkml.kernel.org/r/1453760661-1444-21-git-send-email-richard@nod.at
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agoclocksource: Select CLKSRC_MMIO where needed
Arnd Bergmann [Mon, 25 Jan 2016 16:02:49 +0000 (17:02 +0100)] 
clocksource: Select CLKSRC_MMIO where needed

The Tegra clocksource implementation uses the clocksource_mmio helper
functions, but currently can be configured without them, which fails:

drivers/clocksource/built-in.o: In function `tegra20_init_timer':
:(.init.text+0xac): undefined reference to `clocksource_mmio_init'
:(.init.text+0x140): undefined reference to `clocksource_mmio_readl_up'

The same problem exists for Digicolor:

drivers/clocksource/built-in.o: In function `digicolor_timer_init':
:(.init.text+0xfa): undefined reference to `clocksource_mmio_init'
:(.init.text+0x14c): undefined reference to `clocksource_mmio_readl_down'

I've inspected the Kconfig file to look for other cases that I have not
yet run into, and added an explicit 'select' to each one to ensure we
can successfully link the drivers.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: linux-arm-kernel@lists.infradead.org
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: http://lkml.kernel.org/r/1453737776-1960372-1-git-send-email-arnd@arndb.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agotick/sched: Hide unused oneshot timer code
Arnd Bergmann [Mon, 25 Jan 2016 15:41:49 +0000 (16:41 +0100)] 
tick/sched: Hide unused oneshot timer code

A couple of functions in kernel/time/tick-sched.c are only
relevant for oneshot timer mode, i.e. when hires-timers or
nohz mode are enabled. If both are disabled, we get gcc warnings
about them:

kernel/time/tick-sched.c:98:16: warning: 'tick_init_jiffy_update' defined but not used [-Wunused-function]
 static ktime_t tick_init_jiffy_update(void)
                ^
kernel/time/tick-sched.c:112:13: warning: 'tick_sched_do_timer' defined but not used [-Wunused-function]
 static void tick_sched_do_timer(ktime_t now)
             ^
kernel/time/tick-sched.c:134:13: warning: 'tick_sched_handle' defined but not used [-Wunused-function]
 static void tick_sched_handle(struct tick_sched *ts, struct pt_regs *regs)
             ^

This encloses the whole set of functions in an appropriate ifdef
to avoid the warning and to make it clearer when they are used.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: linux-arm-kernel@lists.infradead.org
Link: http://lkml.kernel.org/r/1453736525-1959191-1-git-send-email-arnd@arndb.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agokselftests: timers: Add adjtimex SETOFFSET validity tests
John Stultz [Thu, 21 Jan 2016 23:03:35 +0000 (15:03 -0800)] 
kselftests: timers: Add adjtimex SETOFFSET validity tests

Add some simple tests to check both valid and invalid
offsets when using adjtimex's ADJ_SETOFFSET method.

Signed-off-by: John Stultz <john.stultz@linaro.org>
Acked-by: Shuah Khan <shuahkh@osg.samsung.com>
Cc: Sasha Levin <sasha.levin@oracle.com>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Harald Hoyer <harald@redhat.com>
Cc: Kay Sievers <kay@vrfy.org>
Cc: David Herrmann <dh.herrmann@gmail.com>
Link: http://lkml.kernel.org/r/1453417415-19110-3-git-send-email-john.stultz@linaro.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agoof: MSI: Simplify irqdomain lookup
Marc Zyngier [Tue, 26 Jan 2016 13:52:26 +0000 (13:52 +0000)] 
of: MSI: Simplify irqdomain lookup

So far, when trying to associate a device with its MSI domain,
we first lookup the domain using a MSI token, and if this
doesn't return anything useful, we pick up any domain matching
the same node.

This logic is broken for two reasons:
1) Only the generic MSI code (PCI or platform) sets this token
   to PCI/MSI or platform MSI. So we're guaranteed that if there
   is something to be found, we will find it with the first call.
2) If we have a convoluted situation where:
   - a single node implements both wired and MSI interrupts
   - MSI support for that HW hasn't been compiled in
   we'll end up using the wired domain for MSIs anyway, and things
   break badly.

So let's just remove __of_get_msi_domain, and replace it by a direct
call to irq_find_matching_host, because that's what we really want.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Acked-by: Rob Herring <robh+dt@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Frank Rowand <frowand.list@gmail.com>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Jiang Liu <jiang.liu@linux.intel.com>
Link: http://lkml.kernel.org/r/1453816347-32720-3-git-send-email-marc.zyngier@arm.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agoirqdomain: Allow domain lookup with DOMAIN_BUS_WIRED token
Marc Zyngier [Tue, 26 Jan 2016 13:52:25 +0000 (13:52 +0000)] 
irqdomain: Allow domain lookup with DOMAIN_BUS_WIRED token

Let's take the (outlandish) example of an interrupt controller
capable of handling both wired interrupts and PCI MSIs.

With the current code, the PCI MSI domain is going to be tagged
with DOMAIN_BUS_PCI_MSI, and the wired domain with DOMAIN_BUS_ANY.

Things get hairy when we start looking up the domain for a wired
interrupt (typically when creating it based on some firmware
information - DT or ACPI).

In irq_create_fwspec_mapping(), we perform the lookup using
DOMAIN_BUS_ANY, which is actually used as a wildcard. This gives
us one chance out of two to end up with the wrong domain, and
we try to configure a wired interrupt with the MSI domain.
Everything grinds to a halt pretty quickly.

What we really need to do is to start looking for a domain that
would uniquely identify a wired interrupt domain, and only use
DOMAIN_BUS_ANY as a fallback.

In order to solve this, let's introduce a new DOMAIN_BUS_WIRED
token, which is going to be used exactly as described above.
Of course, this depends on the irqchip to setup the domain
bus_token, and nobody had to implement this so far.

Only so far.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Frank Rowand <frowand.list@gmail.com>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Jiang Liu <jiang.liu@linux.intel.com>
Link: http://lkml.kernel.org/r/1453816347-32720-2-git-send-email-marc.zyngier@arm.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agoirqchip: Fix dependencies for archs w/o HAS_IOMEM
Richard Weinberger [Mon, 25 Jan 2016 22:24:17 +0000 (23:24 +0100)] 
irqchip: Fix dependencies for archs w/o HAS_IOMEM

Not every arch has io memory. So, unbreak the build by fixing the
dependencies.

Signed-off-by: Richard Weinberger <richard@nod.at>
Cc: user-mode-linux-devel@lists.sourceforge.net
Cc: Jason Cooper <jason@lakedaemon.net>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Link: http://lkml.kernel.org/r/1453760661-1444-19-git-send-email-richard@nod.at
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agoirqchip/s3c24xx: Mark init_eint as __maybe_unused
Arnd Bergmann [Mon, 25 Jan 2016 15:58:08 +0000 (16:58 +0100)] 
irqchip/s3c24xx: Mark init_eint as __maybe_unused

The init_eint array in the s3c24xx irqchip driver is used by
every individual chip variant, but Kconfig allows building
the driver when they are all disabled, and that leads to
a harmless compile-time warning:

drivers/irqchip/irq-s3c24xx.c:608:28: error: 'init_eint' defined but not used [-Werror=unused-variable]

This marks the array as __maybe_unused to avoid the warning.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: Jason Cooper <jason@lakedaemon.net>
Link: http://lkml.kernel.org/r/1453737499-1960073-1-git-send-email-arnd@arndb.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
8 years agoKVM: s390: fix memory overwrites when vx is disabled
David Hildenbrand [Thu, 14 Jan 2016 21:12:47 +0000 (22:12 +0100)] 
KVM: s390: fix memory overwrites when vx is disabled

The kernel now always uses vector registers when available, however KVM
has special logic if support is really enabled for a guest. If support
is disabled, guest_fpregs.fregs will only contain memory for the fpu.
The kernel, however, will store vector registers into that area,
resulting in crazy memory overwrites.

Simply extending that area is not enough, because the format of the
registers also changes. We would have to do additional conversions, making
the code even more complex. Therefore let's directly use one place for
the vector/fpu registers + fpc (in kvm_run). We just have to convert the
data properly when accessing it. This makes current code much easier.

Please note that vector/fpu registers are now always stored to
vcpu->run->s.regs.vrs. Although this data is visible to QEMU and
used for migration, we only guarantee valid values to user space  when
KVM_SYNC_VRS is set. As that is only the case when we have vector
register support, we are on the safe side.

Fixes: b5510d9b68c3 ("s390/fpu: always enable the vector facility if it is available")
Cc: stable@vger.kernel.org # v4.4 d9a3a09af54d s390/kvm: remove dependency on struct save_area definition
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
[adopt to d9a3a09af54d]

8 years agoKVM: s390: Enable the KVM-VFIO device
Dong Jia Shi [Wed, 20 Jan 2016 05:33:42 +0000 (06:33 +0100)] 
KVM: s390: Enable the KVM-VFIO device

The KVM-VFIO device is used by the QEMU VFIO device. It is used to
record the list of in-use VFIO groups so that KVM can manipulate
them.
While we don't need this on s390 currently, let's try to be like
everyone else.

Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
8 years agoKVM: s390: fix guest fprs memory leak
David Hildenbrand [Fri, 22 Jan 2016 13:55:56 +0000 (14:55 +0100)] 
KVM: s390: fix guest fprs memory leak

fprs is never freed, therefore resulting in a memory leak if
kvm_vcpu_init() fails or the vcpu is destroyed.

Fixes: 9977e886cbbc ("s390/kernel: lazy restore fpu registers")
Cc: stable@vger.kernel.org # v4.3+
Reported-by: Eric Farman <farman@linux.vnet.ibm.com>
Signed-off-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Reviewed-by: Eric Farman <farman@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
8 years agoperf stat: Do not clean event's private stats
Jiri Olsa [Wed, 20 Jan 2016 11:56:34 +0000 (12:56 +0100)] 
perf stat: Do not clean event's private stats

Mel reported stddev reporting was broken due to following commit:

106a94a0f8c2 ("perf stat: Introduce read_counters function")

This commit merged interval and overall counters reading into single
read_counters function.

The old interval code cleaned the stddev data for some reason (it's
never displayed in interval mode) and the mentioned commit kept on
cleaning the stddev data in merged function, which resulted in the
stddev not being displayed.

Removing the wrong stddev data cleanup init_stats call.

Reported-and-Tested-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: stable@vger.kernel.org # v4.2+
Fixes: 106a94a0f8c2 ("perf stat: Introduce read_counters function")
Link: http://lkml.kernel.org/r/1453290995-18485-4-git-send-email-jolsa@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
8 years agoperf hists: Fix HISTC_MEM_DCACHELINE width setting
Jiri Olsa [Wed, 20 Jan 2016 11:56:33 +0000 (12:56 +0100)] 
perf hists: Fix HISTC_MEM_DCACHELINE width setting

Set correct width for unresolved mem_dcacheline addr.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Fixes: 9b32ba71ba90 ("perf tools: Add dcacheline sort")
Link: http://lkml.kernel.org/r/1453290995-18485-3-git-send-email-jolsa@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
8 years agoperf annotate browser: Fix behaviour of Shift-Tab with nothing focussed
Markus Trippelsdorf [Mon, 14 Dec 2015 15:44:03 +0000 (16:44 +0100)] 
perf annotate browser: Fix behaviour of Shift-Tab with nothing focussed

The issue was pointed out by gcc-6's -Wmisleading-indentation.

Signed-off-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Acked-by: Ingo Molnar <mingo@kernel.org>
Cc: Ben Hutchings <ben@decadent.org.uk>
Cc: Matt Fleming <matt@codeblueprint.co.uk>
Cc: Peter Zijlstra <peterz@infradead.org>
Fixes: c97cf42219b7 ("perf top: Live TUI Annotation")
Link: http://lkml.kernel.org/r/20151214154403.GB1409@x4
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
8 years agoperf tests: Remove wrong semicolon in while loop in CQM test
Markus Trippelsdorf [Mon, 14 Dec 2015 15:43:35 +0000 (16:43 +0100)] 
perf tests: Remove wrong semicolon in while loop in CQM test

The while loop was spinning. Fix by removing a semicolon.

The issue was pointed out by gcc-6's -Wmisleading-indentation.

Signed-off-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Reviewed-by: Matt Fleming <matt@codeblueprint.co.uk>
Acked-by: Ingo Molnar <mingo@kernel.org>
Cc: Ben Hutchings <ben@decadent.org.uk>
Cc: Peter Zijlstra <peterz@infradead.org>
Fixes: 035827e9f2bd ("perf tests: Add Intel CQM test")
Link: http://lkml.kernel.org/r/20151214154335.GA1409@x4
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
8 years agos390/cio: update measurement characteristics
Sebastian Ott [Mon, 25 Jan 2016 09:32:51 +0000 (10:32 +0100)] 
s390/cio: update measurement characteristics

Per channel path measurement characteristics are obtained during channel
path registration. However if some properties of a channel path change
we don't update the measurement characteristics.

Make sure to update the characteristics when we change the properties of
a channel path or receive a notification from FW about such a change.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/cio: ensure consistent measurement state
Sebastian Ott [Mon, 25 Jan 2016 09:31:33 +0000 (10:31 +0100)] 
s390/cio: ensure consistent measurement state

Make sure that in all cases where we could not obtain measurement
characteristics the associated fields are set to invalid values.

Note: without this change the "shared" capability of a channel path
for which we could not obtain the measurement characteristics was
incorrectly displayed as 0 (not shared). We will now correctly
report "unknown" in this case.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/cio: fix measurement characteristics memleak
Sebastian Ott [Mon, 25 Jan 2016 09:30:27 +0000 (10:30 +0100)] 
s390/cio: fix measurement characteristics memleak

Measurement characteristics are allocated during channel path
registration but not freed during deregistration. Fix this by
embedding these characteristics inside struct channel_path.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/zcrypt: Fix cryptographic device id in kernel messages
Ingo Tuchscherer [Fri, 22 Jan 2016 15:18:29 +0000 (16:18 +0100)] 
s390/zcrypt: Fix cryptographic device id in kernel messages

Currently, on card response failures a combination of card domain and
domain id is recorded in the kernel messages.

According to the message description only the card id will be recorded.
The domain id is not relevant, since the whole card including all domains
is set offline.

Signed-off-by: Ingo Tuchscherer <ingo.tuchscherer@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/pci: remove iomap sanity checks
Sebastian Ott [Fri, 22 Jan 2016 13:11:21 +0000 (14:11 +0100)] 
s390/pci: remove iomap sanity checks

Since each iomap_entry handles only one bar of one pci function
(even when disjunct ranges of a bar are mapped) the sanity check
in pci_iomap_range is not needed and can be removed.

Also convert the remaining BUG_ONs to WARN_ONs.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/pci: set error state for unusable functions
Sebastian Ott [Fri, 22 Jan 2016 13:04:18 +0000 (14:04 +0100)] 
s390/pci: set error state for unusable functions

We receive special notifications from firmware when an error was detected
and a pci function became unusable. Set the error_state accordingly to give
device drivers a hint that they don't need to try error recovery.

Suggested-by: Alexander Schmidt <alexschm@de.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/pci: fix bar check
Sebastian Ott [Fri, 22 Jan 2016 13:03:06 +0000 (14:03 +0100)] 
s390/pci: fix bar check

Fix the check which bar space we should map to allow available bars only.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/pci: resize iomap
Sebastian Ott [Fri, 22 Jan 2016 13:01:44 +0000 (14:01 +0100)] 
s390/pci: resize iomap

On s390 we need to maintain a mapping between iomem addresses
and arch specific function identifiers. Currently the mapping
table is created as such that we could span the whole iomem
address space. Since we can only map each bar space from each
possible function we have an upper bound for the number of
mapping entries.

This reduces the size of the iomap from 256K to less than 4K
(using the defconfig).

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/pci: improve ZPCI_* macros
Sebastian Ott [Fri, 22 Jan 2016 12:59:35 +0000 (13:59 +0100)] 
s390/pci: improve ZPCI_* macros

Most of the constants defined in pci_io.h depend on each other
and thus can be calculated.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/pci: provide ZPCI_ADDR macro
Sebastian Ott [Fri, 22 Jan 2016 12:58:42 +0000 (13:58 +0100)] 
s390/pci: provide ZPCI_ADDR macro

Provide and use a ZPCI_ADDR macro as the complement of ZPCI_IDX
to get rid of some constants in the code.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/pci: adjust IOMAP_MAX_ENTRIES
Sebastian Ott [Fri, 22 Jan 2016 12:57:34 +0000 (13:57 +0100)] 
s390/pci: adjust IOMAP_MAX_ENTRIES

ZPCI_IOMAP_MAX_ENTRIES is off by one. Let's adjust this
for the sake of correctness.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agos390/numa: move numa_init_late() from device to arch_initcall
Michael Holzheu [Wed, 20 Jan 2016 18:22:16 +0000 (19:22 +0100)] 
s390/numa: move numa_init_late() from device to arch_initcall

Commit 3e89e1c5ea ("hugetlb: make mm and fs code explicitly non-modular")
moves hugetlb_init() from module_init to subsys_initcall.

The hugetlb_init()->hugetlb_register_node() code accesses "node->dev.kobj"
which is initialized in numa_init_late().

Since numa_init_late() is a device_initcall which is called *after*
subsys_initcall the above mentioned patch breaks NUMA on s390.

So fix this and move numa_init_late() to arch_initcall.

Fixes: 3e89e1c5ea ("hugetlb: make mm and fs code explicitly non-modular")
Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
8 years agoarm64: defconfig: updates for 4.5
Will Deacon [Mon, 25 Jan 2016 18:25:59 +0000 (18:25 +0000)] 
arm64: defconfig: updates for 4.5

Based on requests, update our defconfig so that:

  - We don't build any modules
  - PL031 is enabled (RTC emulated by qemu)
  - Xen guest support is enabled
  - The Uniphier built-in I2C controller is enabled
  - PCI host controller drivers for the various arm64 SoCs are enabled
  - Device passthrough works on Seattle using SMMU and VFIO
  - The Hisilicon IRQ controller (mbigen) is enabled

Signed-off-by: Will Deacon <will.deacon@arm.com>
8 years agoarm64: errata: Add -mpc-relative-literal-loads to build flags
dann frazier [Mon, 25 Jan 2016 23:52:16 +0000 (16:52 -0700)] 
arm64: errata: Add -mpc-relative-literal-loads to build flags

GCC6 (and Linaro's 2015.12 snapshot of GCC5) has a new default that uses
adrp/ldr or adrp/add to address literal pools. When CONFIG_ARM64_ERRATUM_843419
is enabled, modules built with this toolchain fail to load:

  module libahci: unsupported RELA relocation: 275

This patch fixes the problem by passing '-mpc-relative-literal-loads'
to the compiler.

Cc: stable@vger.kernel.org
Fixes: df057cc7b4fa ("arm64: errata: add module build workaround for erratum #843419")
BugLink: http://bugs.launchpad.net/bugs/1533009
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Suggested-by: Christophe Lyon <christophe.lyon@linaro.org>
Signed-off-by: Dann Frazier <dann.frazier@canonical.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
8 years agoEliminate the .eh_frame sections from the aarch64 vmlinux and kernel modules
William Cohen [Fri, 22 Jan 2016 03:56:26 +0000 (22:56 -0500)] 
Eliminate the .eh_frame sections from the aarch64 vmlinux and kernel modules

By default the aarch64 gcc generates .eh_frame sections.  Unlike
.debug_frame sections, the .eh_frame sections are loaded into memory
when the associated code is loaded.  On an example kernel being built
with this default the .eh_frame section in vmlinux used an extra 1.7MB
of memory.  The x86 disables the creation of the .eh_frame section.
The aarch64 should probably do the same to save some memory.

Signed-off-by: William Cohen <wcohen@redhat.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
8 years agoALSA: bebob: Use a signed return type for get_formation_index
Lucas Tanure [Mon, 25 Jan 2016 21:30:23 +0000 (19:30 -0200)] 
ALSA: bebob: Use a signed return type for get_formation_index

The return type "unsigned int" was used by the get_formation_index function
despite of the aspect that it will eventually return a negative error code.
So, change to signed int and get index by reference in the parameters.

Done with the help of Coccinelle.

[Fix the missing braces suggested by Julia Lawall -- tiwai]

Signed-off-by: Lucas Tanure <tanure@linux.com>
Reviewed-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Tested-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
8 years agortmutex: Make wait_lock irq safe
Thomas Gleixner [Wed, 13 Jan 2016 10:25:38 +0000 (11:25 +0100)] 
rtmutex: Make wait_lock irq safe

Sasha reported a lockdep splat about a potential deadlock between RCU boosting
rtmutex and the posix timer it_lock.

CPU0 CPU1

rtmutex_lock(&rcu->rt_mutex)
  spin_lock(&rcu->rt_mutex.wait_lock)
local_irq_disable()
spin_lock(&timer->it_lock)
spin_lock(&rcu->mutex.wait_lock)
--> Interrupt
    spin_lock(&timer->it_lock)

This is caused by the following code sequence on CPU1

     rcu_read_lock()
     x = lookup();
     if (x)
      spin_lock_irqsave(&x->it_lock);
     rcu_read_unlock();
     return x;

We could fix that in the posix timer code by keeping rcu read locked across
the spinlocked and irq disabled section, but the above sequence is common and
there is no reason not to support it.

Taking rt_mutex.wait_lock irq safe prevents the deadlock.

Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paul McKenney <paulmck@linux.vnet.ibm.com>
8 years agotools/virtio: add ringtest utilities
Michael S. Tsirkin [Thu, 21 Jan 2016 12:44:10 +0000 (14:44 +0200)] 
tools/virtio: add ringtest utilities

This adds micro-benchmarks useful for tuning virtio ring layouts.
Three layouts are currently implemented:

- virtio 0.9 compatible one
- an experimental extension bypassing the ring index, polling ring
  itself instead
- an experimental extension bypassing avail and used ring completely

Typical use:

sh run-on-all.sh perf stat -r 10 --log-fd 1 -- ./ring

It doesn't depend on the kernel directly, but it's handy
to have as much virtio stuff as possible in one tree.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
8 years agosh: fix smp_store_mb for !SMP
Michael S. Tsirkin [Sun, 27 Dec 2015 13:04:42 +0000 (15:04 +0200)] 
sh: fix smp_store_mb for !SMP

sh variant of smp_store_mb() calls xchg() on !SMP which is stronger than
implied by both the name and the documentation.

commit 90a3ccb0be538a914e6a5c51ae919762261563ad ("sh: define __smp_xxx,
fix smp_store_mb for !SMP") was supposed to fix it but
left the bug in place.

Drop smp_store_mb, so that code in asm-generic/barrier.h
will define it correctly depending on CONFIG_SMP.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
8 years agotools/virtio: use virt_xxx barriers
Michael S. Tsirkin [Wed, 20 Jan 2016 19:12:58 +0000 (21:12 +0200)] 
tools/virtio: use virt_xxx barriers

Fix build after API changes.

Reported-by: Kamal Mostafa <kamal@canonical.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
8 years agovirtio_pci: fix use after free on release
Michael S. Tsirkin [Thu, 14 Jan 2016 14:00:41 +0000 (16:00 +0200)] 
virtio_pci: fix use after free on release

KASan detected a use-after-free error in virtio-pci remove code. In
virtio_pci_remove(), vp_dev is still used after being freed in
unregister_virtio_device() (in virtio_pci_release_dev() more
precisely).

To fix, keep a reference until cleanup is done.

Fixes: 63bd62a08ca4 ("virtio_pci: defer kfree until release callback")
Reported-by: Jerome Marchand <jmarchan@redhat.com>
Cc: stable@vger.kernel.org
Cc: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Jerome Marchand <jmarchan@redhat.com>
8 years agoi2c: designware-pci: use IRQF_COND_SUSPEND flag
Andy Shevchenko [Fri, 15 Jan 2016 20:02:12 +0000 (22:02 +0200)] 
i2c: designware-pci: use IRQF_COND_SUSPEND flag

This is effectively reapplies the commit b0898fdaffb2 ("i2c: designware-pci: use
IRQF_COND_SUSPEND flag") after the commit d80d134182ba ("i2c: designware: Move
common probe code into i2c_dw_probe()"). Original message as follows.

The mentioned flag fixes a warning on Intel Edison board since one of the I2C
controller shares IRQ line with watchdog timer.

Fixes: d80d134182ba (i2c: designware: Move common probe code into i2c_dw_probe())
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
8 years agoALSA: usb-audio: Fix TEAC UD-501/UD-503/NT-503 usb delay
Guillaume Fougnies [Mon, 25 Jan 2016 23:28:27 +0000 (00:28 +0100)] 
ALSA: usb-audio: Fix TEAC UD-501/UD-503/NT-503 usb delay

TEAC UD-501/UD-503/NT-503 fail to switch properly between different
rate/format. Similar to 'Playback Design', this patch corrects the
invalid clock source error for TEAC products and avoids complete
freeze of the usb interface of 503 series.

Signed-off-by: Guillaume Fougnies <guillaume@eulerian.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
8 years agoi2c: piix4: Fully initialize SB800 before it is registered
Jean Delvare [Mon, 25 Jan 2016 11:17:07 +0000 (12:17 +0100)] 
i2c: piix4: Fully initialize SB800 before it is registered

This closes a race window where I2C device drivers attempt to access
I2C buses which aren't fully initialized yet.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Tested-by: Christian Fetzer <fetzer.ch@gmail.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
8 years agoi2c: piix4: Fix SB800 locking
Jean Delvare [Fri, 22 Jan 2016 13:12:02 +0000 (14:12 +0100)] 
i2c: piix4: Fix SB800 locking

We need a single mutex for all 4 shared SMBus ports on the SB800. A
per-port mutex doesn't protect us from concurrent access.

In theory the mutex should be per PCI device, however in practice we
know that there's only ever a single instance of the device in a given
system so we can use a global.

Also take the mutex during initialization, as first port may be already
in use when second port is initialized.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Tested-by: Christian Fetzer <fetzer.ch@gmail.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
[wsa: made mutex static]
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
8 years agodrm/amdgpu: don't init fbdev if we don't have any connectors
Alex Deucher [Tue, 26 Jan 2016 05:30:33 +0000 (00:30 -0500)] 
drm/amdgpu: don't init fbdev if we don't have any connectors

Don't init fbdev if we don't have connectors.  E.g., if you have
a PX laptop with the displays attached to an IGP with no driver
support, you may end up with a blank screen rather than falling
back to vesa, etc.

Based on a similar radeon patch from Rob Clark.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
8 years agodrm/radeon: only init fbdev if we have connectors
Rob Clark [Mon, 25 Jan 2016 23:06:48 +0000 (18:06 -0500)] 
drm/radeon: only init fbdev if we have connectors

This fixes an issue that was noticed on an optimus/prime laptop with
a kernel that was old enough to not support the integrated intel gfx
(which was driving all the outputs), but did have support for the
discrete radeon gpu.  The end result was not falling back to VESA and
leaving the user with a black screen.

(Plus it is kind of silly to create an framebuffer device if there
are no outputs hooked up to the gpu.)

Signed-off-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
8 years agoRevert "btrfs: clear PF_NOFREEZE in cleaner_kthread()"
David Sterba [Mon, 25 Jan 2016 10:02:06 +0000 (11:02 +0100)] 
Revert "btrfs: clear PF_NOFREEZE in cleaner_kthread()"

This reverts commit 696249132158014d594896df3a81390616069c5c. The
cleaner thread can block freezing when there's a snapshot cleaning in
progress and the other threads get suspended first. From the logs
provided by Martin we're waiting for reading extent pages:

kernel: PM: Syncing filesystems ... done.
kernel: Freezing user space processes ... (elapsed 0.015 seconds) done.
kernel: Freezing remaining freezable tasks ...
kernel: Freezing of tasks failed after 20.003 seconds (1 tasks refusing to freeze, wq_busy=0):
kernel: btrfs-cleaner   D ffff88033dd13bc0     0   152      2 0x00000000
kernel: ffff88032ebc2e00 ffff88032e750000 ffff88032e74fa50 7fffffffffffffff
kernel: ffffffff814a58df 0000000000000002 ffffea000934d580 ffffffff814a5451
kernel: 7fffffffffffffff ffffffff814a6e8f 0000000000000000 0000000000000020
kernel: Call Trace:
kernel: [<ffffffff814a58df>] ? bit_wait+0x2c/0x2c
kernel: [<ffffffff814a5451>] ? schedule+0x6f/0x7c
kernel: [<ffffffff814a6e8f>] ? schedule_timeout+0x2f/0xd8
kernel: [<ffffffff81076f94>] ? timekeeping_get_ns+0xa/0x2e
kernel: [<ffffffff81077603>] ? ktime_get+0x36/0x44
kernel: [<ffffffff814a4f6c>] ? io_schedule_timeout+0x94/0xf2
kernel: [<ffffffff814a4f6c>] ? io_schedule_timeout+0x94/0xf2
kernel: [<ffffffff814a590b>] ? bit_wait_io+0x2c/0x30
kernel: [<ffffffff814a5694>] ? __wait_on_bit+0x41/0x73
kernel: [<ffffffff8109eba8>] ? wait_on_page_bit+0x6d/0x72
kernel: [<ffffffff8105d718>] ? autoremove_wake_function+0x2a/0x2a
kernel: [<ffffffff811a02d7>] ? read_extent_buffer_pages+0x1bd/0x203
kernel: [<ffffffff8117d9e9>] ? free_root_pointers+0x4c/0x4c
kernel: [<ffffffff8117e831>] ? btree_read_extent_buffer_pages.constprop.57+0x5a/0xe9
kernel: [<ffffffff8117f4f3>] ? read_tree_block+0x2d/0x45
kernel: [<ffffffff8116782a>] ? read_block_for_search.isra.34+0x22a/0x26b
kernel: [<ffffffff811656c3>] ? btrfs_set_path_blocking+0x1e/0x4a
kernel: [<ffffffff8116919b>] ? btrfs_search_slot+0x648/0x736
kernel: [<ffffffff81170559>] ? btrfs_lookup_extent_info+0xb7/0x2c7
kernel: [<ffffffff81170ee5>] ? walk_down_proc+0x9c/0x1ae
kernel: [<ffffffff81171c9d>] ? walk_down_tree+0x40/0xa4
kernel: [<ffffffff8117375f>] ? btrfs_drop_snapshot+0x2da/0x664
kernel: [<ffffffff8104ff21>] ? finish_task_switch+0x126/0x167
kernel: [<ffffffff811850f8>] ? btrfs_clean_one_deleted_snapshot+0xa6/0xb0
kernel: [<ffffffff8117eaba>] ? cleaner_kthread+0x13e/0x17b
kernel: [<ffffffff8117e97c>] ? btrfs_item_end+0x33/0x33
kernel: [<ffffffff8104d256>] ? kthread+0x95/0x9d
kernel: [<ffffffff8104d1c1>] ? kthread_parkme+0x16/0x16
kernel: [<ffffffff814a7b5f>] ? ret_from_fork+0x3f/0x70
kernel: [<ffffffff8104d1c1>] ? kthread_parkme+0x16/0x16

As this affects a released kernel (4.4) we need a minimal fix for
stable kernels.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=108361
Reported-by: Martin Ziegler <ziegler@uni-freiburg.de>
CC: stable@vger.kernel.org # 4.4
CC: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
8 years agobtrfs: async-thread: Fix a use-after-free error for trace
Qu Wenruo [Fri, 22 Jan 2016 01:28:38 +0000 (09:28 +0800)] 
btrfs: async-thread: Fix a use-after-free error for trace

Parameter of trace_btrfs_work_queued() can be freed in its workqueue.
So no one use use that pointer after queue_work().

Fix the user-after-free bug by move the trace line before queue_work().

Reported-by: Dave Jones <davej@codemonkey.org.uk>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
8 years agoBtrfs: fix race between fsync and lockless direct IO writes
Filipe Manana [Thu, 21 Jan 2016 10:17:54 +0000 (10:17 +0000)] 
Btrfs: fix race between fsync and lockless direct IO writes

An fsync, using the fast path, can race with a concurrent lockless direct
IO write and end up logging a file extent item that points to an extent
that wasn't written to yet. This is because the fast fsync path collects
ordered extents into a local list and then collects all the new extent
maps to log file extent items based on them, while the direct IO write
path creates the new extent map before it creates the corresponding
ordered extent (and submitting the respective bio(s)).

So fix this by making the direct IO write path create ordered extents
before the extent maps and make the fast fsync path collect any new
ordered extents after it collects the extent maps.
Note that making the fsync handler call inode_dio_wait() (after acquiring
the inode's i_mutex) would not work and lead to a deadlock when doing
AIO, as through AIO we end up in a path where the fsync handler is called
(through dio_aio_complete_work() -> dio_complete() -> vfs_fsync_range())
before the inode's dio counter is decremented (inode_dio_wait() waits
for this counter to have a value of zero).

Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
8 years agoMerge branch 'fix/fst-sysfs' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave...
Chris Mason [Tue, 26 Jan 2016 00:43:13 +0000 (16:43 -0800)] 
Merge branch 'fix/fst-sysfs' of git://git./linux/kernel/git/kdave/linux into for-linus-4.5

Signed-off-by: Chris Mason <clm@fb.com>
This page took 0.054764 seconds and 5 git commands to generate.