platform/kernel/linux-rpi.git
4 years agonet: sched: act_csum: Fix csum calc for tagged packets
Eli Britstein [Tue, 26 Feb 2019 09:57:34 +0000 (09:57 +0000)]
net: sched: act_csum: Fix csum calc for tagged packets

[ Upstream commit 2ecba2d1e45b24620a7c3df9531895cf68d5dec6 ]

The csum calculation is different for IPv4/6. For VLAN packets,
tc_skb_protocol returns the VLAN protocol rather than the packet's one
(e.g. IPv4/6), so csum is not calculated. Furthermore, VLAN may not be
stripped so csum is not calculated in this case too. Calculate the
csum for those cases.

Fixes: d8b9605d2697 ("net: sched: fix skb->protocol use in case of accelerated vlan path")
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonetfilter: nft_set_hash: bogus element self comparison from deactivation path
Pablo Neira Ayuso [Mon, 25 Feb 2019 13:13:42 +0000 (14:13 +0100)]
netfilter: nft_set_hash: bogus element self comparison from deactivation path

[ Upstream commit a01cbae57ec29b161d42ee1caa4ffffda5d519c2 ]

Use the element from the loop iteration, not the same element we want to
deactivate otherwise this branch always evaluates true.

Fixes: 6c03ae210ce3 ("netfilter: nft_set_hash: add non-resizable hashtable implementation")
Reported-by: Florian Westphal <fw@strlen.de>
Tested-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonetfilter: nft_set_hash: fix lookups with fixed size hash on big endian
Pablo Neira Ayuso [Mon, 25 Feb 2019 13:13:41 +0000 (14:13 +0100)]
netfilter: nft_set_hash: fix lookups with fixed size hash on big endian

[ Upstream commit 3b02b0adc242a72b5e46019b6a9e4f84823592f6 ]

Call jhash_1word() for the 4-bytes key case from the insertion and
deactivation path, otherwise big endian arch set lookups fail.

Fixes: 446a8268b7f5 ("netfilter: nft_set_hash: add lookup variant for fixed size hashtable")
Reported-by: Florian Westphal <fw@strlen.de>
Tested-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoath10k: Fix length of wmi tlv command for protected mgmt frames
Surabhi Vishnoi [Fri, 15 Feb 2019 15:19:37 +0000 (20:49 +0530)]
ath10k: Fix length of wmi tlv command for protected mgmt frames

[ Upstream commit 761156ff573d1002983416e4fd1fe8d3489c4bd8 ]

The length of wmi tlv command for management tx send is calculated
incorrectly in case of protected management frames as there is addition
of IEEE80211_CCMP_MIC_LEN twice. This leads to improper behaviour of
firmware as the wmi tlv mgmt tx send command for protected mgmt frames
is formed wrongly.

Fix the length calculation of wmi tlv command for mgmt tx send in case
of protected management frames by adding the IEEE80211_CCMP_MIC_LEN only
once.

Tested HW: WCN3990
Tested FW: WLAN.HL.3.1-00784-QCAHLSWMTPLZ-1

Fixes: 1807da49733e "ath10k: wmi: add management tx by reference support over wmi"
Signed-off-by: Surabhi Vishnoi <svishnoi@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoregulator: wm831x-dcdc: Fix list of wm831x_dcdc_ilim from mA to uA
Axel Lin [Sun, 24 Feb 2019 13:16:51 +0000 (21:16 +0800)]
regulator: wm831x-dcdc: Fix list of wm831x_dcdc_ilim from mA to uA

[ Upstream commit c25d47888f0fb3d836d68322d4aea2caf31a75a6 ]

The wm831x_dcdc_ilim entries needs to be uA because it is used to compare
with min_uA and max_uA.
While at it also make the array const and change to use unsigned int.

Fixes: e4ee831f949a ("regulator: Add WM831x DC-DC buck convertor support")
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoARM: 8849/1: NOMMU: Fix encodings for PMSAv8's PRBAR4/PRLAR4
Vladimir Murzin [Wed, 20 Feb 2019 14:00:53 +0000 (15:00 +0100)]
ARM: 8849/1: NOMMU: Fix encodings for PMSAv8's PRBAR4/PRLAR4

[ Upstream commit d410a8a49e3e00e07d43037e90f776d522b25a6a ]

To access PRBARn, where n is referenced as a binary number:

MRC p15, 0, <Rt>, c6, c8+n[3:1], 4*n[0] ; Read PRBARn into Rt
MCR p15, 0, <Rt>, c6, c8+n[3:1], 4*n[0] ; Write Rt into PRBARn

To access PRLARn, where n is referenced as a binary number:

MRC p15, 0, <Rt>, c6, c8+n[3:1], 4*n[0]+1 ; Read PRLARn into Rt
MCR p15, 0, <Rt>, c6, c8+n[3:1], 4*n[0]+1 ; Write Rt into PRLARn

For PR{B,L}AR4, n is 4, n[0] is 0, n[3:1] is 2, while current encoding
done with n[0] set to 1 which is wrong. Use proper encoding instead.

Fixes: 046835b4aa22b9ab6aa0bb274e3b71047c4b887d ("ARM: 8757/1: NOMMU: Support PMSAv8 MPU")
Signed-off-by: Vladimir Murzin <vladimir.murzin@arm.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoARM: 8848/1: virt: Align GIC version check with arm64 counterpart
Vladimir Murzin [Wed, 20 Feb 2019 14:00:13 +0000 (15:00 +0100)]
ARM: 8848/1: virt: Align GIC version check with arm64 counterpart

[ Upstream commit 9db043d36bd379f4cc99054c079de0dabfc38d03 ]

arm64 has got relaxation on GIC version check at early boot stage due
to update of the GIC architecture let's align ARM with that.

To help backports (even though the code was correct at the time of writing)
Fixes: e59941b9b381 ("ARM: 8527/1: virt: enable GICv3 system registers")
Signed-off-by: Vladimir Murzin <vladimir.murzin@arm.com>
Reviewed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoARM: 8847/1: pm: fix HYP/SVC mode mismatch when MCPM is used
Marek Szyprowski [Mon, 18 Feb 2019 08:31:41 +0000 (09:31 +0100)]
ARM: 8847/1: pm: fix HYP/SVC mode mismatch when MCPM is used

[ Upstream commit ca70ea43f80c98582f5ffbbd1e6f4da2742da0c4 ]

MCPM does a soft reset of the CPUs and uses common cpu_resume() routine to
perform low-level platform initialization. This results in a try to install
HYP stubs for the second time for each CPU and results in false HYP/SVC
mode mismatch detection. The HYP stubs are already installed at the
beginning of the kernel initialization on the boot CPU (head.S) or in the
secondary_startup() for other CPUs. To fix this issue MCPM code should use
a cpu_resume() routine without HYP stubs installation.

This change fixes HYP/SVC mode mismatch on Samsung Exynos5422-based Odroid
XU3/XU4/HC1 boards.

Fixes: 3721924c8154 ("ARM: 8081/1: MCPM: provide infrastructure to allow for MCPM loopback")
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Tested-by: Anand Moon <linux.amoon@gmail.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoiommu: Fix IOMMU debugfs fallout
Geert Uytterhoeven [Wed, 20 Feb 2019 13:05:05 +0000 (14:05 +0100)]
iommu: Fix IOMMU debugfs fallout

[ Upstream commit 18b3af4492a0aa6046b86d712f6ba4cbb66100fb ]

A change made in the final version of IOMMU debugfs support replaced the
public function iommu_debugfs_new_driver_dir() by the public dentry
iommu_debugfs_dir in <linux/iommu.h>, but forgot to update both the
implementation in iommu-debugfs.c, and the patch description.

Fix this by exporting iommu_debugfs_dir, and removing the reference to
and implementation of iommu_debugfs_new_driver_dir().

Fixes: bad614b24293ae46 ("iommu: Enable debugfs exposure of IOMMU driver internals")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Gary R Hook <gary.hook@amd.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agommc: sdhci-brcmstb: handle mmc_of_parse() errors during probe
Stefan Wahren [Sun, 23 Dec 2018 20:59:18 +0000 (21:59 +0100)]
mmc: sdhci-brcmstb: handle mmc_of_parse() errors during probe

[ Upstream commit 1e20186e706da8446f9435f2924cd65ab1397e73 ]

We need to handle mmc_of_parse() errors during probe otherwise the
MMC driver could start without proper initialization (e.g. power sequence).

Fixes: 476bf3d62d5c ("mmc: sdhci-brcmstb: Add driver for Broadcom BRCMSTB SoCs")
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoNFS/pnfs: Bulk destroy of layouts needs to be safe w.r.t. umount
Trond Myklebust [Fri, 22 Feb 2019 19:20:27 +0000 (14:20 -0500)]
NFS/pnfs: Bulk destroy of layouts needs to be safe w.r.t. umount

[ Upstream commit 5085607d209102b37b169bc94d0aa39566a9842a ]

If a bulk layout recall or a metadata server reboot coincides with a
umount, then holding a reference to an inode is unsafe unless we
also hold a reference to the super block.

Fixes: fd9a8d7160937 ("NFSv4.1: Fix bulk recall and destroy of layouts")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoplatform/x86: wmi: fix potential null pointer dereference
Mattias Jacobsson [Wed, 30 Jan 2019 15:14:24 +0000 (16:14 +0100)]
platform/x86: wmi: fix potential null pointer dereference

[ Upstream commit c355ec651a8941864549f2586f969d0eb7bf499a ]

In the function wmi_dev_match() the variable id is dereferenced without
first performing a NULL check. The variable can for example be NULL if
a WMI driver is registered without specifying the id_table field in
struct wmi_driver.

Add a NULL check and return that the driver can't handle the device if
the variable is NULL.

Fixes: 844af950da94 ("platform/x86: wmi: Turn WMI into a bus driver")
Signed-off-by: Mattias Jacobsson <2pi@mok.nu>
Signed-off-by: Darren Hart (VMware) <dvhart@infradead.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoclocksource/drivers/exynos_mct: Fix error path in timer resources initialization
Marek Szyprowski [Thu, 18 Oct 2018 09:57:04 +0000 (11:57 +0200)]
clocksource/drivers/exynos_mct: Fix error path in timer resources initialization

[ Upstream commit b9307420196009cdf18bad55e762ac49fb9a80f4 ]

While freeing interrupt handlers in error path, don't assume that all
requested interrupts are per-processor interrupts and properly release
standard interrupts too.

Reported-by: Krzysztof Kozlowski <krzk@kernel.org>
Fixes: 56a94f13919c ("clocksource: exynos_mct: Avoid blocking calls in the cpu hotplug notifier")
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Krzysztof Kozlowski <krzk@kernel.org>
Reviewed-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoclocksource/drivers/sun5i: Fail gracefully when clock rate is unavailable
Chen-Yu Tsai [Thu, 10 Jan 2019 06:22:07 +0000 (14:22 +0800)]
clocksource/drivers/sun5i: Fail gracefully when clock rate is unavailable

[ Upstream commit e7e7e0d7beafebd11b0c065cd5fbc1e5759c5aab ]

If the clock tree is not fully populated when the timer-sun5i init code
is called, attempts to get the clock rate for the timer would fail and
return 0.

Make the init code for both clock events and clocksource check the
returned clock rate and fail gracefully if the result is 0, instead of
causing a divide by 0 exception later on.

Fixes: 4a59058f0b09 ("clocksource/drivers/sun5i: Refactor the current code")
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoperf, pt, coresight: Fix address filters for vmas with non-zero offset
Alexander Shishkin [Fri, 15 Feb 2019 11:56:55 +0000 (13:56 +0200)]
perf, pt, coresight: Fix address filters for vmas with non-zero offset

[ Upstream commit c60f83b813e5b25ccd5de7e8c8925c31b3aebcc1 ]

Currently, the address range calculation for file-based filters works as
long as the vma that maps the matching part of the object file starts
from offset zero into the file (vm_pgoff==0). Otherwise, the resulting
filter range would be off by vm_pgoff pages. Another related problem is
that in case of a partially matching vma, that is, a vma that matches
part of a filter region, the filter range size wouldn't be adjusted.

Fix the arithmetics around address filter range calculations, taking
into account vma offset, so that the entire calculation is done before
the filter configuration is passed to the PMU drivers instead of having
those drivers do the final bit of arithmetics.

Based on the patch by Adrian Hunter <adrian.hunter.intel.com>.

Reported-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Tested-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Fixes: 375637bc5249 ("perf/core: Introduce address range filtering")
Link: http://lkml.kernel.org/r/20190215115655.63469-3-alexander.shishkin@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoperf: Copy parent's address filter offsets on clone
Alexander Shishkin [Fri, 15 Feb 2019 11:56:54 +0000 (13:56 +0200)]
perf: Copy parent's address filter offsets on clone

[ Upstream commit 18736eef12137c59f60cc9f56dc5bea05c92e0eb ]

When a child event is allocated in the inherit_event() path, the VMA
based filter offsets are not copied from the parent, even though the
address space mapping of the new task remains the same, which leads to
no trace for the new task until exec.

Reported-by: Mansour Alharthi <malharthi9@gatech.edu>
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Tested-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Fixes: 375637bc5249 ("perf/core: Introduce address range filtering")
Link: http://lkml.kernel.org/r/20190215115655.63469-2-alexander.shishkin@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoNFS: Fix a soft lockup in the delegation recovery code
Trond Myklebust [Thu, 21 Feb 2019 19:51:25 +0000 (14:51 -0500)]
NFS: Fix a soft lockup in the delegation recovery code

[ Upstream commit 6f9449be53f3ce383caed797708b332ede8d952c ]

Fix a soft lockup when NFS client delegation recovery is attempted
but the inode is in the process of being freed. When the
igrab(inode) call fails, and we have to restart the recovery process,
we need to ensure that we won't attempt to recover the same delegation
again.

Fixes: 45870d6909d5a ("NFSv4.1: Test delegation stateids when server...")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agopowerpc/64s: Fix logic when handling unknown CPU features
Michael Ellerman [Mon, 11 Feb 2019 00:20:01 +0000 (11:20 +1100)]
powerpc/64s: Fix logic when handling unknown CPU features

[ Upstream commit 8cfaf106918a8c13abb24c641556172afbb9545c ]

In cpufeatures_process_feature(), if a provided CPU feature is unknown and
enable_unknown is false, we erroneously print that the feature is being
enabled and return true, even though no feature has been enabled, and
may also set feature bits based on the last entry in the match table.

Fix this so that we only set feature bits from the match table if we have
actually enabled a feature from that table, and when failing to enable an
unknown feature, always print the "not enabling" message and return false.

Coincidentally, some older gccs (<GCC 7), when invoked with
-fsanitize-coverage=trace-pc, cause a spurious uninitialised variable
warning in this function:

  arch/powerpc/kernel/dt_cpu_ftrs.c: In function ‘cpufeatures_process_feature’:
  arch/powerpc/kernel/dt_cpu_ftrs.c:686:7: warning: ‘m’ may be used uninitialized in this function [-Wmaybe-uninitialized]
    if (m->cpu_ftr_bit_mask)

An upcoming patch will enable support for kcov, which requires this option.
This patch avoids the warning.

Fixes: 5a61ef74f269 ("powerpc/64s: Support new device tree binding for discovering CPU features")
Reported-by: Segher Boessenkool <segher@kernel.crashing.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[ajd: add commit message]
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agostaging: rtlwifi: Use proper enum for return in halmac_parse_psd_data_88xx
Nathan Chancellor [Thu, 21 Feb 2019 05:25:24 +0000 (22:25 -0700)]
staging: rtlwifi: Use proper enum for return in halmac_parse_psd_data_88xx

[ Upstream commit e8edc32d70a4e09160835792eb5d1af71a0eec14 ]

Clang warns:

drivers/staging/rtlwifi/halmac/halmac_88xx/halmac_func_88xx.c:2472:11:
warning: implicit conversion from enumeration type 'enum
halmac_cmd_process_status' to different enumeration type 'enum
halmac_ret_status' [-Wenum-conversion]
                        return HALMAC_CMD_PROCESS_ERROR;
                        ~~~~~~ ^~~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.

Fix this by using the proper enum for allocation failures,
HALMAC_RET_MALLOC_FAIL, which is used in the rest of this file.

Fixes: e4b08e16b7d9 ("staging: r8822be: check kzalloc return or bail")
Link: https://github.com/ClangBuiltLinux/linux/issues/375
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Reviewed-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agofs/nfs: Fix nfs_parse_devname to not modify it's argument
Eric W. Biederman [Wed, 30 Jan 2019 13:58:38 +0000 (07:58 -0600)]
fs/nfs: Fix nfs_parse_devname to not modify it's argument

[ Upstream commit 40cc394be1aa18848b8757e03bd8ed23281f572e ]

In the rare and unsupported case of a hostname list nfs_parse_devname
will modify dev_name.  There is no need to modify dev_name as the all
that is being computed is the length of the hostname, so the computed
length can just be shorted.

Fixes: dc04589827f7 ("NFS: Use common device name parsing logic for NFSv4 and NFSv2/v3")
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet: dsa: fix unintended change of bridge interface STP state
Russell King [Wed, 20 Feb 2019 10:32:52 +0000 (10:32 +0000)]
net: dsa: fix unintended change of bridge interface STP state

[ Upstream commit 9c2054a5cf415a9dc32c91ffde78399955deb571 ]

When a DSA port is added to a bridge and brought up, the resulting STP
state programmed into the hardware depends on the order that these
operations are performed.  However, the Linux bridge code believes that
the port is in disabled mode.

If the DSA port is first added to a bridge and then brought up, it will
be in blocking mode.  If it is brought up and then added to the bridge,
it will be in disabled mode.

This difference is caused by DSA always setting the STP mode in
dsa_port_enable() whether or not this port is part of a bridge.  Since
bridge always sets the STP state when the port is added, brought up or
taken down, it is unnecessary for us to manipulate the STP state.

Apparently, this code was copied from Rocker, and the very next day a
similar fix for Rocker was merged but was not propagated to DSA.  See
e47172ab7e41 ("rocker: put port in FORWADING state after leaving bridge")

Fixes: b73adef67765 ("net: dsa: integrate with SWITCHDEV for HW bridging")
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Vivien Didelot <vivien.didelot@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoASoC: qcom: Fix of-node refcount unbalance in apq8016_sbc_parse_of()
Takashi Iwai [Tue, 19 Feb 2019 15:46:50 +0000 (16:46 +0100)]
ASoC: qcom: Fix of-node refcount unbalance in apq8016_sbc_parse_of()

[ Upstream commit 8d1667200850f8753c0265fa4bd25c9a6e5f94ce ]

The apq8016 driver leaves the of-node refcount at aborting from the
loop of for_each_child_of_node() in the error path.  Not only the
iterator node of for_each_child_of_node(), the children nodes referred
from it for codec and cpu have to be properly unreferenced.

Fixes: bdb052e81f62 ("ASoC: qcom: add apq8016 sound card support")
Cc: Patrick Lai <plai@codeaurora.org>
Cc: Banajit Goswami <bgoswami@codeaurora.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodriver core: Fix PM-runtime for links added during consumer probe
Rafael J. Wysocki [Tue, 19 Feb 2019 16:53:26 +0000 (17:53 +0100)]
driver core: Fix PM-runtime for links added during consumer probe

[ Upstream commit 36003d4cf57ca431fb3f94d317bcca426a2394d6 ]

Commit 4c06c4e6cf63 ("driver core: Fix possible supplier PM-usage
counter imbalance") introduced a regression that causes suppliers
to be suspended prematurely for device links added during consumer
driver probe if the initial PM-runtime status of the consumer is
"suspended" and the consumer is resumed after adding the link and
before pm_runtime_put_suppliers() is called.  In that case,
pm_runtime_put_suppliers() will drop the rpm_active refcount for
the link by one and (since rpm_active is equal to two after the
preceding consumer resume) the supplier's PM-runtime usage counter
will be decremented, which may cause the supplier to suspend even
though the consumer's PM-runtime status is "active".

For this reason, partially revert commit 4c06c4e6cf63 as the problem
it tried to fix needs to be addressed somewhat differently, and
change pm_runtime_get_suppliers() and pm_runtime_put_suppliers() so
that the latter only drops rpm_active references acquired by the
former.  [This requires adding a new field to struct device_link,
but I coulnd't find a cleaner way to address the issue that would
work in all cases.]

This causes pm_runtime_put_suppliers() to effectively ignore device
links added during consumer probe, so device_link_add() doesn't need
to worry about ensuring that suppliers will remain active after
pm_runtime_put_suppliers() for links created with DL_FLAG_RPM_ACTIVE
set and it only needs to bump up rpm_active by one for those links,
so pm_runtime_active_link() is not necessary any more.

Fixes: 4c06c4e6cf63 ("driver core: Fix possible supplier PM-usage counter imbalance")
Reported-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodrm/nouveau: fix missing break in switch statement
Colin Ian King [Mon, 8 Oct 2018 20:47:36 +0000 (21:47 +0100)]
drm/nouveau: fix missing break in switch statement

[ Upstream commit 785cf1eeafa23ec63f426d322401054d13abe2a3 ]

The NOUVEAU_GETPARAM_PCI_DEVICE case is missing a break statement and falls
through to the following NOUVEAU_GETPARAM_BUS_TYPE case and may end up
re-assigning the getparam->value to an undesired value. Fix this by adding
in the missing break.

Detected by CoverityScan, CID#1460507 ("Missing break in switch")

Fixes: 359088d5b8ec ("drm/nouveau: remove trivial cases of nvxx_device() usage")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodrm/nouveau/pmu: don't print reply values if exec is false
Colin Ian King [Wed, 19 Dec 2018 15:29:49 +0000 (15:29 +0000)]
drm/nouveau/pmu: don't print reply values if exec is false

[ Upstream commit b1d03fc36ec9834465a08c275c8d563e07f6f6bf ]

Currently the uninitialized values in the array reply are printed out
when exec is false and nvkm_pmu_send has not updated the array. Avoid
confusion by only dumping out these values if they have been actually
updated.

Detected by CoverityScan, CID#1271291 ("Uninitialized scaler variable")
Fixes: ebb58dc2ef8c ("drm/nouveau/pmu: rename from pwr (no binary change)")

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodrm/nouveau/bios/ramcfg: fix missing parentheses when calculating RON
Colin Ian King [Sun, 25 Nov 2018 17:09:18 +0000 (17:09 +0000)]
drm/nouveau/bios/ramcfg: fix missing parentheses when calculating RON

[ Upstream commit 13649101a25c53c87f4ab98a076dfe61f3636ab1 ]

Currently, the expression for calculating RON is always going to result
in zero no matter the value of ram->mr[1] because the ! operator has
higher precedence than the shift >> operator.  I believe the missing
parentheses around the expression before appying the ! operator will
result in the desired result.

[ Note, not tested ]

Detected by CoveritScan, CID#1324005 ("Operands don't affect result")

Fixes: c25bf7b6155c ("drm/nouveau/bios/ramcfg: Separate out RON pull value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet/mlx5: Delete unused FPGA QPN variable
Leon Romanovsky [Sun, 17 Feb 2019 13:21:28 +0000 (15:21 +0200)]
net/mlx5: Delete unused FPGA QPN variable

[ Upstream commit 566428375a53619196e31803130dd1a7010c4d7f ]

fpga_qpn was assigned but never used and compilation with W=1
produced the following warning:

drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c: In function _mlx5_fpga_event_:
drivers/net/ethernet/mellanox/mlx5/core/fpga/core.c:320:6: warning:
variable _fpga_qpn_ set but not used [-Wunused-but-set-variable]
  u32 fpga_qpn;
      ^~~~~~~~

Fixes: 98db16bab59f ("net/mlx5: FPGA, Handle QP error event")
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet: dsa: qca8k: Enable delay for RGMII_ID mode
Vinod Koul [Tue, 19 Feb 2019 06:59:43 +0000 (12:29 +0530)]
net: dsa: qca8k: Enable delay for RGMII_ID mode

[ Upstream commit a968b5e9d5879f9535d6099505f9e14abcafb623 ]

RGMII_ID specifies that we should have internal delay, so resurrect the
delay addition routine but under the RGMII_ID mode.

Fixes: 40269aa9f40a ("net: dsa: qca8k: disable delay for RGMII mode")
Tested-by: Michal Vokáč <michal.vokac@ysoft.com>
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoregulator: pv88090: Fix array out-of-bounds access
Axel Lin [Tue, 19 Feb 2019 10:00:02 +0000 (18:00 +0800)]
regulator: pv88090: Fix array out-of-bounds access

[ Upstream commit a5455c9159414748bed4678184bf69989a4f7ba3 ]

Fix off-by-one while iterating current_limits array.
The valid index should be 0 ~ n_current_limits -1.

Fixes: c90456e36d9c ("regulator: pv88090: new regulator driver")
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoregulator: pv88080: Fix array out-of-bounds access
Axel Lin [Tue, 19 Feb 2019 10:00:01 +0000 (18:00 +0800)]
regulator: pv88080: Fix array out-of-bounds access

[ Upstream commit 3c413f594c4f9df40061445667ca11a12bc8ee34 ]

Fix off-by-one while iterating current_limits array.
The valid index should be 0 ~ n_current_limits -1.

Fixes: 99cf3af5e2d5 ("regulator: pv88080: new regulator driver")
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoregulator: pv88060: Fix array out-of-bounds access
Axel Lin [Tue, 19 Feb 2019 10:00:00 +0000 (18:00 +0800)]
regulator: pv88060: Fix array out-of-bounds access

[ Upstream commit 7cd415f875591bc66c5ecb49bf84ef97e80d7b0e ]

Fix off-by-one while iterating current_limits array.
The valid index should be 0 ~ n_current_limits -1.

Fixes: f307a7e9b7af ("regulator: pv88060: new regulator driver")
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agobrcmfmac: create debugfs files for bus-specific layer
Arend van Spriel [Thu, 14 Feb 2019 12:43:49 +0000 (13:43 +0100)]
brcmfmac: create debugfs files for bus-specific layer

[ Upstream commit aaf6a5e86e36766abbeedf220462bde8031f9a72 ]

Since we moved the drivers debugfs directory under ieee80211 debugfs the
debugfs entries need to be added after wiphy_register() has been called.
For most part that has been done accordingly, but for the debugfs entries
added by SDIO it was not and failed silently. This patch fixes that by
adding a bus-layer callback for it.

Fixes: 856d5a011c86 ("brcmfmac: allocate struct brcmf_pub instance using wiphy_new()")
Reported-by: Russel King <linux@armlinux.org.uk>
Reviewed-by: Franky Lin <franky.lin@broadcom.com>
Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agocdc-wdm: pass return value of recover_from_urb_loss
YueHaibing [Mon, 18 Feb 2019 14:34:51 +0000 (22:34 +0800)]
cdc-wdm: pass return value of recover_from_urb_loss

[ Upstream commit 0742a338f5b3446a26de551ad8273fb41b2787f2 ]

'rv' is the correct return value, pass it upstream instead of 0

Fixes: 17d80d562fd7 ("USB: autosuspend for cdc-wdm")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodmaengine: mv_xor: Use correct device for DMA API
Robin Murphy [Mon, 18 Feb 2019 18:27:06 +0000 (18:27 +0000)]
dmaengine: mv_xor: Use correct device for DMA API

[ Upstream commit 3e5daee5ecf314da33a890fabaa2404244cd2a36 ]

Using dma_dev->dev for mappings before it's assigned with the correct
device is unlikely to work as expected, and with future dma-direct
changes, passing a NULL device may end up crashing entirely. I don't
know enough about this hardware or the mv_xor_prep_dma_interrupt()
operation to implement the appropriate error-handling logic that would
have revealed those dma_map_single() calls failing on arm64 for as long
as the driver has been enabled there, but moving the assignment earlier
will at least make the current code operate as intended.

Fixes: 22843545b200 ("dma: mv_xor: Add support for DMA_INTERRUPT")
Reported-by: John David Anglin <dave.anglin@bell.net>
Tested-by: John David Anglin <dave.anglin@bell.net>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Tested-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agostaging: r8822be: check kzalloc return or bail
Nicholas Mc Guire [Fri, 15 Feb 2019 09:24:22 +0000 (10:24 +0100)]
staging: r8822be: check kzalloc return or bail

[ Upstream commit e4b08e16b7d9d030b6475ef48f94d734a39f3c81 ]

The kzalloc() in halmac_parse_psd_data_88xx() can fail and return NULL
so check the psd_set->data after allocation and if allocation failed
return HALMAC_CMD_PROCESS_ERROR.

Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Fixes: 938a0447f094 ("staging: r8822be: Add code for halmac sub-drive")
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoKVM: PPC: Release all hardware TCE tables attached to a group
Alexey Kardashevskiy [Tue, 12 Feb 2019 04:37:45 +0000 (15:37 +1100)]
KVM: PPC: Release all hardware TCE tables attached to a group

[ Upstream commit a67614cc05a5052b265ea48196dab2fce11f5f2e ]

The SPAPR TCE KVM device references all hardware IOMMU tables assigned to
some IOMMU group to ensure that in-kernel KVM acceleration of H_PUT_TCE
can work. The tables are references when an IOMMU group gets registered
with the VFIO KVM device by the KVM_DEV_VFIO_GROUP_ADD ioctl;
KVM_DEV_VFIO_GROUP_DEL calls into the dereferencing code
in kvm_spapr_tce_release_iommu_group() which walks through the list of
LIOBNs, finds a matching IOMMU table and calls kref_put() when found.

However that code stops after the very first successful derefencing
leaving other tables referenced till the SPAPR TCE KVM device is destroyed
which normally happens on guest reboot or termination so if we do hotplug
and unplug in a loop, we are leaking IOMMU tables here.

This removes a premature return to let kvm_spapr_tce_release_iommu_group()
find and dereference all attached tables.

Fixes: 121f80ba68f ("KVM: PPC: VFIO: Add in-kernel acceleration for VFIO")
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agomdio_bus: Fix PTR_ERR() usage after initialization to constant
YueHaibing [Sat, 16 Feb 2019 02:59:35 +0000 (10:59 +0800)]
mdio_bus: Fix PTR_ERR() usage after initialization to constant

[ Upstream commit 780feae7eb69388c8d8b661cda6706b0dc0f642b ]

Fix coccinelle warning:

./drivers/net/phy/mdio_bus.c:51:5-12: ERROR: PTR_ERR applied after initialization to constant on line 44
./drivers/net/phy/mdio_bus.c:52:5-12: ERROR: PTR_ERR applied after initialization to constant on line 44

fix this by using IS_ERR before PTR_ERR

Fixes: bafbdd527d56 ("phylib: Add device reset GPIO support")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agohwmon: (pmbus/tps53679) Fix driver info initialization in probe routine
Vadim Pasternak [Mon, 18 Feb 2019 19:54:40 +0000 (19:54 +0000)]
hwmon: (pmbus/tps53679) Fix driver info initialization in probe routine

[ Upstream commit ff066653aeed8ee2d4dadb1e35774dd91ecbb19f ]

Fix tps53679_probe() by using dynamically allocated "pmbus_driver_info"
structure instead of static. Usage of static structures causes
overwritten of the field "vrm_version", in case the system is equipped
with several tps53679 devices with the different "vrm_version".
In such case the last probed device overwrites this field for all
others.

Fixes: 610526527a13 ("hwmon: (pmbus) Add support for Texas Instruments tps53679 device")
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agovfio_pci: Enable memory accesses before calling pci_map_rom
Eric Auger [Fri, 15 Feb 2019 16:16:06 +0000 (17:16 +0100)]
vfio_pci: Enable memory accesses before calling pci_map_rom

[ Upstream commit 0cfd027be1d6def4a462cdc180c055143af24069 ]

pci_map_rom/pci_get_rom_size() performs memory access in the ROM.
In case the Memory Space accesses were disabled, readw() is likely
to trigger a synchronous external abort on some platforms.

In case memory accesses were disabled, re-enable them before the
call and disable them back again just after.

Fixes: 89e1f7d4c66d ("vfio: Add PCI device driver")
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Suggested-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agomedia: sh: migor: Include missing dma-mapping header
Jacopo Mondi [Tue, 5 Feb 2019 19:37:42 +0000 (14:37 -0500)]
media: sh: migor: Include missing dma-mapping header

[ Upstream commit 5c88ee02932a964096cbbcc7c9f38b78d230bacb ]

Since the removal of the stale soc_camera headers, Migo-R board fails to
build due to missing dma-mapping include directive.

Include missing dma-mapping.h header in Migo-R board file to fix the build
error.

Fixes: a50c7738e8ae ("media: sh: migor: Remove stale soc_camera include")

Signed-off-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agomt76: usb: fix possible memory leak in mt76u_buf_free
Lorenzo Bianconi [Sun, 10 Feb 2019 21:49:14 +0000 (22:49 +0100)]
mt76: usb: fix possible memory leak in mt76u_buf_free

[ Upstream commit cb83585e1121bd6d6c039cf09fa32380bf8b6258 ]

Move q->ndesc initialization before the for loop in mt76u_alloc_rx
since otherwise allocated urbs will not be freed in mt76u_buf_free
Double-check scatterlist pointer in mt76u_buf_free

Fixes: b40b15e1521f ("mt76: add usb support to mt76 layer")
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet: dsa: b53: Do not program CPU port's PVID
Florian Fainelli [Fri, 15 Feb 2019 20:16:53 +0000 (12:16 -0800)]
net: dsa: b53: Do not program CPU port's PVID

[ Upstream commit 10163aaee9671b01b2f4737922e1a4f43581047a ]

The CPU port is special and does not need to obey VLAN restrictions as
far as untagged traffic goes, also, having the CPU port be part of a
particular PVID is against the idea of keeping it tagged in all VLANs.

Fixes: ca8931948344 ("net: dsa: b53: Keep CPU port as tagged in all VLANs")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet: dsa: b53: Properly account for VLAN filtering
Florian Fainelli [Fri, 15 Feb 2019 20:16:50 +0000 (12:16 -0800)]
net: dsa: b53: Properly account for VLAN filtering

[ Upstream commit dad8d7c6452b5b9f9828c9e2c7ca143205fd40c7 ]

VLAN filtering can be built into the kernel, and also dynamically turned
on/off through the bridge master device. Allow re-configuring the switch
appropriately to account for that by deciding whether VLAN table
(v_table) misses should lead to a drop or forward.

Fixes: a2482d2ce349 ("net: dsa: b53: Plug in VLAN support")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet: dsa: b53: Fix default VLAN ID
Florian Fainelli [Fri, 15 Feb 2019 20:16:49 +0000 (12:16 -0800)]
net: dsa: b53: Fix default VLAN ID

[ Upstream commit fea83353177a55540c71c140887737c282137aa2 ]

We were not consistent in how the default VID of a given port was
defined, b53_br_leave() would make sure the VLAN ID would be either 0/1
depending on the switch generation, but b53_configure_vlan(), which is
the default configuration would unconditionally set it to 1. The correct
value is 1 for 5325/5365 series and 0 otherwise. To avoid repeating that
mistake ever again, introduce a helper function: b53_default_pvid() to
factor that out.

Fixes: 967dd82ffc52 ("net: dsa: b53: Add support for Broadcom RoboSwitch")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agokeys: Timestamp new keys
David Howells [Thu, 14 Feb 2019 16:20:37 +0000 (16:20 +0000)]
keys: Timestamp new keys

[ Upstream commit 7c1857bdbdf1e4c541e45eab477ee23ed4333ea4 ]

Set the timestamp on new keys rather than leaving it unset.

Fixes: 31d5a79d7f3d ("KEYS: Do LRU discard in full keyrings")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <james.morris@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoblock: don't use bio->bi_vcnt to figure out segment number
Ming Lei [Fri, 15 Feb 2019 11:13:08 +0000 (19:13 +0800)]
block: don't use bio->bi_vcnt to figure out segment number

[ Upstream commit 1a67356e9a4829da2935dd338630a550c59c8489 ]

It is wrong to use bio->bi_vcnt to figure out how many segments
there are in the bio even though CLONED flag isn't set on this bio,
because this bio may be splitted or advanced.

So always use bio_segments() in blk_recount_segments(), and it shouldn't
cause any performance loss now because the physical segment number is figured
out in blk_queue_split() and BIO_SEG_VALID is set meantime since
bdced438acd83ad83a6c ("block: setup bi_phys_segments after splitting").

Reviewed-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Fixes: 76d8137a3113 ("blk-merge: recaculate segment if it isn't less than max segments")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agousb: phy: twl6030-usb: fix possible use-after-free on remove
Sven Van Asbroeck [Mon, 11 Feb 2019 15:04:26 +0000 (10:04 -0500)]
usb: phy: twl6030-usb: fix possible use-after-free on remove

[ Upstream commit 5895d311d28f2605e2f71c1a3e043ed38f3ac9d2 ]

In remove(), use cancel_delayed_work_sync() to cancel the
delayed work. Otherwise there's a chance that this work
will continue to run until after the device has been removed.

This issue was detected with the help of Coccinelle.

Cc: Tony Lindgren <tony@atomide.com>
Cc: Bin Liu <b-liu@ti.com>
Fixes: b6a619a883c3 ("usb: phy: Check initial state for twl6030")
Signed-off-by: Sven Van Asbroeck <TheSven73@gmail.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoPCI: endpoint: functions: Use memcpy_fromio()/memcpy_toio()
Wen Yang [Mon, 11 Feb 2019 09:04:00 +0000 (17:04 +0800)]
PCI: endpoint: functions: Use memcpy_fromio()/memcpy_toio()

[ Upstream commit 726dabfde6aa35a4f1508e235ae37edbbf9fbc65 ]

Functions copying from/to IO addresses should use the
memcpy_fromio()/memcpy_toio() API rather than plain memcpy().

Fix the issue detected through the sparse tool.

Fixes: 349e7a85b25f ("PCI: endpoint: functions: Add an EP function to test PCI")
Suggested-by: Kishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: Wen Yang <wen.yang99@zte.com.cn>
[lorenzo.pieralisi@arm.com: updated log]
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Kishon Vijay Abraham I <kishon@ti.com>
CC: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
CC: Bjorn Helgaas <bhelgaas@google.com>
CC: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
CC: Niklas Cassel <niklas.cassel@axis.com>
CC: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
CC: Cyrille Pitchen <cyrille.pitchen@free-electrons.com>
CC: linux-pci@vger.kernel.org
CC: linux-kernel@vger.kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodriver core: Fix possible supplier PM-usage counter imbalance
Rafael J. Wysocki [Tue, 12 Feb 2019 12:08:10 +0000 (13:08 +0100)]
driver core: Fix possible supplier PM-usage counter imbalance

[ Upstream commit 4c06c4e6cf63d7f3d5dfe62593a073253d750a59 ]

If a stateless device link to a certain supplier with
DL_FLAG_PM_RUNTIME set in the flags is added and then removed by the
consumer driver's probe callback, the supplier's PM-runtime usage
counter will be nonzero after that which effectively causes the
supplier to remain "always on" going forward.

Namely, device_link_add() called to add the link invokes
device_link_rpm_prepare() which notices that the consumer driver is
probing, so it increments the supplier's PM-runtime usage counter
with the assumption that the link will stay around until
pm_runtime_put_suppliers() is called by driver_probe_device(),
but if the link goes away before that point, the supplier's
PM-runtime usage counter will remain nonzero.

To prevent that from happening, first rework pm_runtime_get_suppliers()
and pm_runtime_put_suppliers() to use the rpm_active refounts of device
links and make the latter only drop rpm_active and the supplier's
PM-runtime usage counter for each link by one, unless rpm_active is
one already for it.  Next, modify device_link_add() to bump up the
new link's rpm_active refcount and the suppliers PM-runtime usage
counter by two, to prevent pm_runtime_put_suppliers(), if it is
called subsequently, from suspending the supplier prematurely (in
case its PM-runtime usage counter goes down to 0 in there).

Due to the way rpm_put_suppliers() works, this change does not
affect runtime suspend of the consumer ends of new device links (or,
generally, device links for which DL_FLAG_PM_RUNTIME has just been
set).

Fixes: e2f3cd831a28 ("driver core: Fix handling of runtime PM flags in device_link_add()")
Reported-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoRDMA/mlx5: Fix memory leak in case we fail to add an IB device
Mark Bloch [Mon, 11 Feb 2019 15:40:54 +0000 (17:40 +0200)]
RDMA/mlx5: Fix memory leak in case we fail to add an IB device

[ Upstream commit fc9e4477f924e84d7798f7a1d41401d699de1219 ]

Make sure the IB device is freed on failure.

Fixes: b5ca15ad7e61 ("IB/mlx5: Add proper representors support")
Signed-off-by: Mark Bloch <markb@mellanox.com>
Reviewed-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Håkon Bugge <haakon.bugge@oracle.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agopinctrl: sh-pfc: sh73a0: Fix fsic_spdif pin groups
Geert Uytterhoeven [Wed, 23 Jan 2019 15:51:21 +0000 (16:51 +0100)]
pinctrl: sh-pfc: sh73a0: Fix fsic_spdif pin groups

[ Upstream commit 0e6e448bdcf896d001a289a6112a704542d51516 ]

There are two pin groups for the FSIC SPDIF signal, but the FSIC pin
group array lists only one, and it refers to a nonexistent group.

Fixes: 2ecd4154c906b7d6 ("sh-pfc: sh73a0: Add FSI pin groups and functions")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agopinctrl: sh-pfc: r8a7792: Fix vin1_data18_b pin group
Geert Uytterhoeven [Wed, 23 Jan 2019 16:14:07 +0000 (17:14 +0100)]
pinctrl: sh-pfc: r8a7792: Fix vin1_data18_b pin group

[ Upstream commit b9fd50488b4939ce5b3a026d29e752e17c2d1800 ]

The vin1_data18_b pin group itself is present, but it is not listed in
the VIN1 pin group array, and thus cannot be selected.

Fixes: 7dd74bb1f058786e ("pinctrl: sh-pfc: r8a7792: Add VIN pin groups")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agopinctrl: sh-pfc: r8a7791: Fix scifb2_data_c pin group
Geert Uytterhoeven [Wed, 23 Jan 2019 16:07:43 +0000 (17:07 +0100)]
pinctrl: sh-pfc: r8a7791: Fix scifb2_data_c pin group

[ Upstream commit a4b0350047f1b10207e25e72d7cd3f7826e93769 ]

The entry for "scifb2_data_c" in the SCIFB2 pin group array contains a
typo, thus the group cannot be selected.

Fixes: 5088451962389924 ("pinctrl: sh-pfc: r8a7791 PFC support")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agopinctrl: sh-pfc: emev2: Add missing pinmux functions
Geert Uytterhoeven [Thu, 24 Jan 2019 12:04:52 +0000 (13:04 +0100)]
pinctrl: sh-pfc: emev2: Add missing pinmux functions

[ Upstream commit 1ecd8c9cb899ae277e6986ae134635cb1a50f5de ]

The err_rst_reqb, ext_clki, lowpwr, and ref_clko pin groups are present,
but no pinmux functions refer to them, hence they can not be selected.

Fixes: 1e7d5d849cf4f0c5 ("sh-pfc: Add emev2 pinmux support")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agontb_hw_switchtec: NT req id mapping table register entry number should be 512
Wesley Sheng [Thu, 6 Dec 2018 13:30:51 +0000 (21:30 +0800)]
ntb_hw_switchtec: NT req id mapping table register entry number should be 512

[ Upstream commit d123fab71f63aae129aebe052664fda73131921a ]

The number of available NT req id mapping table entries per NTB control
register is 512. The driver mistakenly limits the number to 256.

Fix the array size of NT req id mapping table.

Fixes: c082b04c9d40 ("NTB: switchtec: Add NTB hardware register definitions")
Signed-off-by: Wesley Sheng <wesley.sheng@microchip.com>
Reviewed-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agontb_hw_switchtec: debug print 64bit aligned crosslink BAR Numbers
Paul Selles [Thu, 6 Dec 2018 13:30:50 +0000 (21:30 +0800)]
ntb_hw_switchtec: debug print 64bit aligned crosslink BAR Numbers

[ Upstream commit cce8e04cf79e47809455215744685e8eb56f94bb ]

Switchtec NTB crosslink BARs are 64bit addressed but they are printed as
32bit addressed BARs. Fix debug log to increment the BAR numbers by 2 to
reflect the 64bit address alignment.

Fixes: 017525018202 ("ntb_hw_switchtec: Add initialization code for crosslink")
Signed-off-by: Paul Selles <paul.selles@microchip.com>
Signed-off-by: Wesley Sheng <wesley.sheng@microchip.com>
Reviewed-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodrm/etnaviv: potential NULL dereference
Dan Carpenter [Tue, 5 Feb 2019 09:08:19 +0000 (12:08 +0300)]
drm/etnaviv: potential NULL dereference

[ Upstream commit 9e05352340d3a3e68c144136db9810b26ebb88c3 ]

The etnaviv_gem_prime_get_sg_table() is supposed to return error
pointers.  Otherwise it can lead to a NULL dereference when it's called
from drm_gem_map_dma_buf().

Fixes: 5f4a4a73f437 ("drm/etnaviv: fix gem_prime_get_sg_table to return new SG table")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoxsk: add missing smp_rmb() in xsk_mmap
Magnus Karlsson [Fri, 8 Feb 2019 13:13:50 +0000 (14:13 +0100)]
xsk: add missing smp_rmb() in xsk_mmap

[ Upstream commit e6762c8bcf982821935a2b1cb33cf8335d0eefae ]

All the setup code in AF_XDP is protected by a mutex with the
exception of the mmap code that cannot use it. To make sure that a
process banging on the mmap call at the same time as another process
is setting up the socket, smp_wmb() calls were added in the umem
registration code and the queue creation code, so that the published
structures that xsk_mmap needs would be consistent. However, the
corresponding smp_rmb() calls were not added to the xsk_mmap
code. This patch adds these calls.

Fixes: 37b076933a8e3 ("xsk: add missing write- and data-dependency barrier")
Fixes: c0c77d8fb787c ("xsk: add user memory registration support sockopt")
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoipmi: kcs_bmc: handle devm_kasprintf() failure case
Nicholas Mc Guire [Wed, 21 Nov 2018 15:08:25 +0000 (16:08 +0100)]
ipmi: kcs_bmc: handle devm_kasprintf() failure case

[ Upstream commit 42c7c6ef1e6fa5fc0425120f06f045190b1dda2d ]

devm_kasprintf() may return NULL if internal allocation failed so this
assignment is not safe. Moved the error exit path and added the !NULL
which then allows the devres manager to take care of cleanup.

Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Fixes: cd2315d471f4 ("ipmi: kcs_bmc: don't change device name")
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Reviewed-by: Haiyue Wang <haiyue.wang@linux.intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoiw_cxgb4: use tos when finding ipv6 routes
Steve Wise [Fri, 1 Feb 2019 20:44:53 +0000 (12:44 -0800)]
iw_cxgb4: use tos when finding ipv6 routes

[ Upstream commit c8a7eb554a83214c3d8ee5cb322da8c72810d2dc ]

When IPv6 support was added, the correct tos was not passed to
cxgb_find_route6(). This potentially results in the wrong route entry.

Fixes: 830662f6f032 ("RDMA/cxgb4: Add support for active and passive open connection with IPv6 address")
Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoiw_cxgb4: use tos when importing the endpoint
Steve Wise [Fri, 1 Feb 2019 20:44:41 +0000 (12:44 -0800)]
iw_cxgb4: use tos when importing the endpoint

[ Upstream commit cb3ba0bde881f0cb7e3945d2a266901e2bd18c92 ]

import_ep() is passed the correct tos, but doesn't use it correctly.

Fixes: ac8e4c69a021 ("cxgb4/iw_cxgb4: TOS support")
Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agofbdev: chipsfb: remove set but not used variable 'size'
YueHaibing [Fri, 8 Feb 2019 18:24:45 +0000 (19:24 +0100)]
fbdev: chipsfb: remove set but not used variable 'size'

[ Upstream commit 8e71fa5e4d86bedfd26df85381d65d6b4c860020 ]

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/video/fbdev/chipsfb.c: In function 'chipsfb_pci_init':
drivers/video/fbdev/chipsfb.c:352:22: warning:
 variable 'size' set but not used [-Wunused-but-set-variable]

Fixes: 8c8709334cec ("[PATCH] ppc32: Remove CONFIG_PMAC_PBOOK").
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Michael Ellerman <mpe@ellerman.id.au>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Christophe Leroy <christophe.leroy@c-s.fr>
[b.zolnierkie: minor commit summary and description fixups]
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agortc: pm8xxx: fix unintended sign extension
Colin Ian King [Wed, 6 Feb 2019 10:31:02 +0000 (10:31 +0000)]
rtc: pm8xxx: fix unintended sign extension

[ Upstream commit e42280886018c6f77f0a90190f7cba344b0df3e0 ]

Shifting a u8 by 24 will cause the value to be promoted to an integer. If
the top bit of the u8 is set then the following conversion to an unsigned
long will sign extend the value causing the upper 32 bits to be set in
the result.

Fix this by casting the u8 value to an unsigned long before the shift.

Detected by CoverityScan, CID#1309693 ("Unintended sign extension")

Fixes: 9a9a54ad7aa2 ("drivers/rtc: add support for Qualcomm PMIC8xxx RTC")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agortc: 88pm80x: fix unintended sign extension
Colin Ian King [Wed, 6 Feb 2019 10:08:11 +0000 (10:08 +0000)]
rtc: 88pm80x: fix unintended sign extension

[ Upstream commit fb0b322537a831b5b0cb948c56f8f958ce493d3a ]

Shifting a u8 by 24 will cause the value to be promoted to an integer. If
the top bit of the u8 is set then the following conversion to an unsigned
long will sign extend the value causing the upper 32 bits to be set in
the result.

Fix this by casting the u8 value to an unsigned long before the shift.

Detected by CoverityScan, CID#714646-714649 ("Unintended sign extension")

Fixes: 2985c29c1964 ("rtc: Add rtc support to 88PM80X PMIC")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agortc: 88pm860x: fix unintended sign extension
Colin Ian King [Wed, 6 Feb 2019 09:50:53 +0000 (09:50 +0000)]
rtc: 88pm860x: fix unintended sign extension

[ Upstream commit dc9e47160626cdb58d5c39a4f43dcfdb27a5c004 ]

Shifting a u8 by 24 will cause the value to be promoted to an integer. If
the top bit of the u8 is set then the following conversion to an unsigned
long will sign extend the value causing the upper 32 bits to be set in
the result.

Fix this by casting the u8 value to an unsigned long before the shift.

Detected by CoverityScan, CID#144925-144928 ("Unintended sign extension")

Fixes: 008b30408c40 ("mfd: Add rtc support to 88pm860x")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet/smc: original socket family in inet_sock_diag
Karsten Graul [Thu, 7 Feb 2019 14:56:20 +0000 (15:56 +0100)]
net/smc: original socket family in inet_sock_diag

[ Upstream commit 232dc8ef647658a5352da807d9e994e0e03b43cd ]

Commit ed75986f4aae ("net/smc: ipv6 support for smc_diag.c") changed the
value of the diag_family field. The idea was to indicate the family of
the IP address in the inet_diag_sockid field. But the change makes it
impossible to distinguish an inet_sock_diag response message from SMC
sock_diag response. This patch restores the original behaviour and sends
AF_SMC as value of the diag_family field.

Fixes: ed75986f4aae ("net/smc: ipv6 support for smc_diag.c")
Reported-by: Eugene Syromiatnikov <esyr@redhat.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: Ursula Braun <ubraun@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agortc: ds1307: rx8130: Fix alarm handling
Uwe Kleine-König [Fri, 25 Jan 2019 14:35:58 +0000 (15:35 +0100)]
rtc: ds1307: rx8130: Fix alarm handling

[ Upstream commit 3f929cad943380370b6db31fcb7a38d898d91089 ]

When the EXTENSION.WADA bit is set, register 0x19 contains a bitmap of
week days, not a day of month. As Linux only handles a single alarm
without repetition using day of month is more flexible, so clear this
bit. (Otherwise a value depending on time.tm_wday would have to be
written to register 0x19.)

Also optimize setting the AIE bit to use a single register write instead
of a bulk write of three registers.

Fixes: ee0981be7704 ("rtc: ds1307: Add support for Epson RX8130CE")
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet: phy: fixed_phy: Fix fixed_phy not checking GPIO
Moritz Fischer [Thu, 7 Feb 2019 05:45:29 +0000 (21:45 -0800)]
net: phy: fixed_phy: Fix fixed_phy not checking GPIO

[ Upstream commit 8f289805616e81f7c1690931aa8a586c76f4fa88 ]

Fix fixed_phy not checking GPIO if no link_update callback
is registered.

In the original version all users registered a link_update
callback so the issue was masked.

Fixes: a5597008dbc2 ("phy: fixed_phy: Add gpio to determine link up/down.")
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Moritz Fischer <mdf@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoath10k: fix dma unmap direction for management frames
Rakesh Pillai [Fri, 25 Jan 2019 04:21:06 +0000 (09:51 +0530)]
ath10k: fix dma unmap direction for management frames

[ Upstream commit 6e8a8991e2103dcb6a9cff28f460390e8e360848 ]

The management frames transmitted are dma mapped with
direction TO_DEVICE, but incorrectly mapped with
direction FROM_DEVICE during tx complete and error cases.

Fix the direction of dma during dma unmap of the
transmitted management frames.

Tested HW: WCN3990
Tested FW: WLAN.HL.2.0-01188-QCAHLSWMTPLZ-1

Fixes: 38a1390e02b7 ("ath10k: dma unmap mgmt tx buffer if wmi cmd send fails")
Signed-off-by: Rakesh Pillai <pillair@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoarm64: dts: msm8916: remove bogus argument to the cpu clock
Niklas Cassel [Thu, 24 Jan 2019 12:45:08 +0000 (13:45 +0100)]
arm64: dts: msm8916: remove bogus argument to the cpu clock

[ Upstream commit e4f045ef38e61ba37aa4afc916fce4fc1b37aa19 ]

The apcs node has #clock-cells = <0>, which means that those who
references it should specify 0 arguments.

The apcs reference in the cpu node incorrectly specifies an argument,
remove this bogus argument.

Fixes: 65afdf458360 ("arm64: dts: qcom: msm8916: Add CPU frequency scaling support")
Signed-off-by: Niklas Cassel <niklas.cassel@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Amit Kucheria <amit.kucheria@linaro.org>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agothermal: mediatek: fix register index error
Michael Kao [Fri, 1 Feb 2019 07:38:07 +0000 (15:38 +0800)]
thermal: mediatek: fix register index error

[ Upstream commit eb9aecd90d1a39601e91cd08b90d5fee51d321a6 ]

The index of msr and adcpnp should match the sensor
which belongs to the selected bank in the for loop.

Fixes: b7cf0053738c ("thermal: Add Mediatek thermal driver for mt2701.")
Signed-off-by: Michael Kao <michael.kao@mediatek.com>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agortc: ds1672: fix unintended sign extension
Colin Ian King [Tue, 5 Feb 2019 18:04:49 +0000 (18:04 +0000)]
rtc: ds1672: fix unintended sign extension

[ Upstream commit f0c04c276739ed8acbb41b4868e942a55b128dca ]

Shifting a u8 by 24 will cause the value to be promoted to an integer. If
the top bit of the u8 is set then the following conversion to an unsigned
long will sign extend the value causing the upper 32 bits to be set in
the result.

Fix this by casting the u8 value to an unsigned long before the shift.

Detected by CoverityScan, CID#138801 ("Unintended sign extension")

Fixes: edf1aaa31fc5 ("[PATCH] RTC subsystem: DS1672 driver")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoclk: ingenic: jz4740: Fix gating of UDC clock
Paul Cercueil [Fri, 25 Jan 2019 15:34:36 +0000 (12:34 -0300)]
clk: ingenic: jz4740: Fix gating of UDC clock

[ Upstream commit b7e29924a1a628aec60d18651b493fa1601bf944 ]

The UDC clock is gated when the bit is cleared, not when it is set.

Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Tested-by: Artur Rojek <contact@artur-rojek.eu>
Fixes: 2b555a4b9cae ("clk: ingenic: Add missing flag for UDC clock")
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agostaging: most: cdev: add missing check for cdev_add failure
Colin Ian King [Sat, 2 Feb 2019 22:34:49 +0000 (22:34 +0000)]
staging: most: cdev: add missing check for cdev_add failure

[ Upstream commit 5ae890780e1b4d08f2c0c5d4ea96fc3928fc0ee9 ]

Currently the call to cdev_add is missing a check for failure. Fix this by
checking for failure and exiting via a new error path that ensures the
allocated comp_channel struct is kfree'd.

Detected by CoverityScan, CID#1462359 ("Unchecked return value")

Fixes: 9bc79bbcd0c5 ("Staging: most: add MOST driver's aim-cdev module")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoiwlwifi: mvm: fix RSS config command
Sara Sharon [Wed, 12 Dec 2018 07:45:11 +0000 (09:45 +0200)]
iwlwifi: mvm: fix RSS config command

[ Upstream commit 608dce95db10b8ee1a26dbce3f60204bb69812a5 ]

The hash mask is a bitmap, so we should use BIT() on
the enum values.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Fixes: 43413a975d06 ("iwlwifi: mvm: support rss queues configuration command")
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodrm/xen-front: Fix mmap attributes for display buffers
Oleksandr Andrushchenko [Tue, 29 Jan 2019 15:04:22 +0000 (17:04 +0200)]
drm/xen-front: Fix mmap attributes for display buffers

[ Upstream commit 24ded292a5c2ed476f01c77fee65f8320552cd27 ]

When GEM backing storage is allocated those are normal pages,
so there is no point using pgprot_writecombine while mmaping.
This fixes mismatch of buffer pages' memory attributes between
the frontend and backend which may cause screen artifacts.

Fixes: c575b7eeb89f ("drm/xen-front: Add support for Xen PV display frontend")

Signed-off-by: Oleksandr Andrushchenko <oleksandr_andrushchenko@epam.com>
Suggested-by: Julien Grall <julien.grall@arm.com>
Acked-by: Julien Grall <julien.grall@arm.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190129150422.19867-1-andr2000@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoARM: dts: lpc32xx: phy3250: fix SD card regulator voltage
Vladimir Zapolskiy [Tue, 29 Jan 2019 19:20:41 +0000 (21:20 +0200)]
ARM: dts: lpc32xx: phy3250: fix SD card regulator voltage

[ Upstream commit dc141b99fc36cf910a1d8d5ee30f43f2442fd1bd ]

The fixed voltage regulator on Phytec phyCORE-LPC3250 board, which
supplies SD/MMC card's power, has a constant output voltage level
of either 3.15V or 3.3V, the actual value depends on JP4 position,
the power rail is referenced as VCC_SDIO in the board hardware manual.

Fixes: d06670e96267 ("arm: dts: phy3250: add SD fixed regulator")
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoARM: dts: lpc32xx: fix ARM PrimeCell LCD controller clocks property
Vladimir Zapolskiy [Tue, 29 Jan 2019 19:20:40 +0000 (21:20 +0200)]
ARM: dts: lpc32xx: fix ARM PrimeCell LCD controller clocks property

[ Upstream commit 30fc01bae3cda747e7d9c352b1aa51ca113c8a9d ]

The originally added ARM PrimeCell PL111 clocks property misses
the required "clcdclk" clock, which is the same as a clock to enable
the LCD controller on NXP LPC3230 and NXP LPC3250 SoCs.

Fixes: 93898eb775e5 ("arm: dts: lpc32xx: add clock properties to device nodes")
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoARM: dts: lpc32xx: fix ARM PrimeCell LCD controller variant
Vladimir Zapolskiy [Tue, 29 Jan 2019 19:20:39 +0000 (21:20 +0200)]
ARM: dts: lpc32xx: fix ARM PrimeCell LCD controller variant

[ Upstream commit 7a0790a4121cbcd111cc537cdc801c46ccb789ee ]

ARM PrimeCell PL111 LCD controller is found on On NXP LPC3230
and LPC3250 SoCs variants, the original reference in compatible
property to an older one ARM PrimeCell PL110 is invalid.

Fixes: e04920d9efcb3 ("ARM: LPC32xx: DTS files for device tree conversion")
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoARM: dts: lpc32xx: reparent keypad controller to SIC1
Vladimir Zapolskiy [Sat, 26 Jan 2019 14:29:21 +0000 (16:29 +0200)]
ARM: dts: lpc32xx: reparent keypad controller to SIC1

[ Upstream commit 489261c45f0ebbc1c2813f337bbdf858267f5033 ]

After switching to a new interrupt controller scheme by separating SIC1
and SIC2 from MIC interrupt controller just one SoC keypad controller
was not taken into account, fix it now:

  WARNING: CPU: 0 PID: 1 at kernel/irq/irqdomain.c:524 irq_domain_associate+0x50/0x1b0
  error: hwirq 0x36 is too large for interrupt-controller@40008000
  ...
  lpc32xx_keys 40050000.key: failed to get platform irq
  lpc32xx_keys: probe of 40050000.key failed with error -22

Fixes: 9b8ad3fb81ae ("ARM: dts: lpc32xx: reparent SIC1 and SIC2 interrupts from MIC")
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoARM: dts: lpc32xx: add required clocks property to keypad device node
Vladimir Zapolskiy [Sat, 26 Jan 2019 14:29:20 +0000 (16:29 +0200)]
ARM: dts: lpc32xx: add required clocks property to keypad device node

[ Upstream commit 3e88bc38b9f6fe4b69cecf81badd3c19fde97f97 ]

NXP LPC32xx keypad controller requires a clock property to be defined.

The change fixes the driver initialization problem:

  lpc32xx_keys 40050000.key: failed to get clock
  lpc32xx_keys: probe of 40050000.key failed with error -2

Fixes: 93898eb775e5 ("arm: dts: lpc32xx: add clock properties to device nodes")
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodriver core: Do not call rpm_put_suppliers() in pm_runtime_drop_link()
Rafael J. Wysocki [Fri, 1 Feb 2019 00:52:45 +0000 (01:52 +0100)]
driver core: Do not call rpm_put_suppliers() in pm_runtime_drop_link()

[ Upstream commit a1fdbfbb1da2063ba98a12eb6f1bdd07451c7145 ]

Calling rpm_put_suppliers() from pm_runtime_drop_link() is excessive
as it affects all suppliers of the consumer device and not just the
one pointed to by the device link being dropped.  Worst case it may
cause the consumer device to stop working unexpectedly.  Moreover, in
principle it is racy with respect to runtime PM of the consumer
device.

To avoid these problems drop runtime PM references on the particular
supplier pointed to by the link in question only and do that after
the link has been dropped from the consumer device's list of links to
suppliers, which is in device_link_free().

Fixes: a0504aecba76 ("PM / runtime: Drop usage count for suppliers at device link removal")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodriver core: Fix handling of runtime PM flags in device_link_add()
Rafael J. Wysocki [Fri, 1 Feb 2019 00:49:14 +0000 (01:49 +0100)]
driver core: Fix handling of runtime PM flags in device_link_add()

[ Upstream commit e2f3cd831a280fc226118d9369bf3f77aab58c56 ]

After commit ead18c23c263 ("driver core: Introduce device links
reference counting"), if there is a link between the given supplier
and the given consumer already, device_link_add() will refcount it
and return it unconditionally without updating its flags.  It is
possible, however, that the second (or any subsequent) caller of
device_link_add() for the same consumer-supplier pair will pass
DL_FLAG_PM_RUNTIME, possibly along with DL_FLAG_RPM_ACTIVE, in flags
to it and the existing link may not behave as expected then.

First, if DL_FLAG_PM_RUNTIME is not set in the existing link's flags
at all, it needs to be set like during the original initialization of
the link.

Second, if DL_FLAG_RPM_ACTIVE is passed to device_link_add() in flags
(in addition to DL_FLAG_PM_RUNTIME), the existing link should to be
updated to reflect the "active" runtime PM configuration of the
consumer-supplier pair and extra care must be taken here to avoid
possible destructive races with runtime PM of the consumer.

To that end, redefine the rpm_active field in struct device_link
as a refcount, initialize it to 1 and make rpm_resume() (for the
consumer) and device_link_add() increment it whenever they acquire
a runtime PM reference on the supplier device.  Accordingly, make
rpm_suspend() (for the consumer) and pm_runtime_clean_up_links()
decrement it and drop runtime PM references to the supplier
device in a loop until rpm_active becones 1 again.

Fixes: ead18c23c263 ("driver core: Introduce device links reference counting")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodriver core: Do not resume suppliers under device_links_write_lock()
Rafael J. Wysocki [Fri, 1 Feb 2019 00:47:53 +0000 (01:47 +0100)]
driver core: Do not resume suppliers under device_links_write_lock()

[ Upstream commit 5db25c9eb893df8f6b93c1d97b8006d768e1b6f5 ]

It is incorrect to call pm_runtime_get_sync() under
device_links_write_lock(), because it may end up trying to take
device_links_read_lock() while resuming the target device and that
will deadlock in the non-SRCU case, so avoid that by resuming the
supplier device in device_link_add() before calling
device_links_write_lock().

Fixes: 21d5c57b3726 ("PM / runtime: Use device links")
Fixes: baa8809f6097 ("PM / runtime: Optimize the use of device links")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodriver core: Avoid careless re-use of existing device links
Rafael J. Wysocki [Fri, 1 Feb 2019 00:46:54 +0000 (01:46 +0100)]
driver core: Avoid careless re-use of existing device links

[ Upstream commit f265df550a4350dce0a4d721a77c52e4b847ea40 ]

After commit ead18c23c263 ("driver core: Introduce device links
reference counting"), if there is a link between the given supplier
and the given consumer already, device_link_add() will refcount it
and return it unconditionally.  However, if the flags passed to
it on the second (or any subsequent) attempt to create a device
link between the same consumer-supplier pair are not compatible with
the existing link's flags, that is incorrect.

First off, if the existing link is stateless and the next caller of
device_link_add() for the same consumer-supplier pair wants a
stateful one, or the other way around, the existing link cannot be
returned, because it will not match the expected behavior, so make
device_link_add() dump the stack and return NULL in that case.

Moreover, if the DL_FLAG_AUTOREMOVE_CONSUMER flag is passed to
device_link_add(), its caller will expect its reference to the link
to be dropped automatically on consumer driver removal, which will
not happen if that flag is not set in the link's flags (and
analogously for DL_FLAG_AUTOREMOVE_SUPPLIER).  For this reason, make
device_link_add() update the existing link's flags accordingly
before returning it to the caller.

Fixes: ead18c23c263 ("driver core: Introduce device links reference counting")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodriver core: Fix DL_FLAG_AUTOREMOVE_SUPPLIER device link flag handling
Rafael J. Wysocki [Fri, 1 Feb 2019 00:45:55 +0000 (01:45 +0100)]
driver core: Fix DL_FLAG_AUTOREMOVE_SUPPLIER device link flag handling

[ Upstream commit c8d50986da5d74ddfc233b13b91d0a13369fa164 ]

Change the list walk in device_links_driver_cleanup() to a safe one
to avoid use-after-free when dropping a link from the list during the
walk.

Also, while at it, fix device_link_add() to refuse to create
stateless device links with DL_FLAG_AUTOREMOVE_SUPPLIER set, which is
an invalid combination (setting that flag means that the driver core
should manage the link, so it cannot be stateless), and extend the
kerneldoc comment of device_link_add() to cover the
DL_FLAG_AUTOREMOVE_SUPPLIER flag properly too.

Fixes: 1689cac5b32a ("driver core: Add flag to autoremove device link on supplier unbind")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agocrypto: crypto4xx - Fix wrong ppc4xx_trng_probe()/ppc4xx_trng_remove() arguments
Corentin Labbe [Wed, 23 Jan 2019 11:24:18 +0000 (11:24 +0000)]
crypto: crypto4xx - Fix wrong ppc4xx_trng_probe()/ppc4xx_trng_remove() arguments

[ Upstream commit 6e88098ca43a3d80ae86908f7badba683c8a0d84 ]

When building without CONFIG_HW_RANDOM_PPC4XX, I hit the following build failure:
drivers/crypto/amcc/crypto4xx_core.c: In function 'crypto4xx_probe':
drivers/crypto/amcc/crypto4xx_core.c:1407:20: error: passing argument 1 of 'ppc4xx_trng_probe' from incompatible pointer type [-Werror=incompatible-pointer-types]
In file included from drivers/crypto/amcc/crypto4xx_core.c:50:0:
drivers/crypto/amcc/crypto4xx_trng.h:28:20: note: expected 'struct crypto4xx_device *' but argument is of type 'struct crypto4xx_core_device *'
drivers/crypto/amcc/crypto4xx_core.c: In function 'crypto4xx_remove':
drivers/crypto/amcc/crypto4xx_core.c:1434:21: error: passing argument 1 of 'ppc4xx_trng_remove' from incompatible pointer type [-Werror=incompatible-pointer-types]
In file included from drivers/crypto/amcc/crypto4xx_core.c:50:0:
drivers/crypto/amcc/crypto4xx_trng.h:30:20: note: expected 'struct crypto4xx_device *' but argument is of type 'struct crypto4xx_core_device *'

This patch fix the needed argument of ppc4xx_trng_probe()/ppc4xx_trng_remove() in that case.

Fixes: 5343e674f32f ("crypto4xx: integrate ppc4xx-rng into crypto4xx")
Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodriver: uio: fix possible use-after-free in __uio_register_device
Liu Jian [Tue, 22 Jan 2019 22:45:38 +0000 (06:45 +0800)]
driver: uio: fix possible use-after-free in __uio_register_device

[ Upstream commit 221a1f4ac12d2ab46246c160b2e00d1b1160d5d9 ]

In uio_dev_add_attributes() error handing case, idev is used after
device_unregister(), in which 'idev' has been released, touch idev cause
use-after-free.

Fixes: a93e7b331568 ("uio: Prevent device destruction while fds are open")
Signed-off-by: Liu Jian <liujian56@huawei.com>
Reviewed-by: Hamish Martin <hamish.martin@alliedtelesis.co.nz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agodriver: uio: fix possible memory leak in __uio_register_device
Liu Jian [Tue, 22 Jan 2019 22:45:37 +0000 (06:45 +0800)]
driver: uio: fix possible memory leak in __uio_register_device

[ Upstream commit 1a392b3de7c5747506b38fc14b2e79977d3c7770 ]

'idev' is malloced in __uio_register_device() and leak free it before
leaving from the uio_get_minor() error handing case, it will cause
memory leak.

Fixes: a93e7b331568 ("uio: Prevent device destruction while fds are open")
Signed-off-by: Liu Jian <liujian56@huawei.com>
Reviewed-by: Hamish Martin <hamish.martin@alliedtelesis.co.nz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agotty: ipwireless: Fix potential NULL pointer dereference
YueHaibing [Wed, 30 Jan 2019 10:30:51 +0000 (18:30 +0800)]
tty: ipwireless: Fix potential NULL pointer dereference

[ Upstream commit 7dd50e205b3348dc7784efbdf85723551de64a25 ]

There is a potential NULL pointer dereference in case
alloc_ctrl_packet() fails and returns NULL.

Fixes: 099dc4fb6265 ("ipwireless: driver for PC Card 3G/UMTS modem")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agobus: ti-sysc: Fix timer handling with drop pm_runtime_irq_safe()
Tony Lindgren [Tue, 22 Jan 2019 17:03:08 +0000 (09:03 -0800)]
bus: ti-sysc: Fix timer handling with drop pm_runtime_irq_safe()

[ Upstream commit 9bd34c63f5536c490c152833c77fa47f59aeade3 ]

Commit 84badc5ec5fc ("ARM: dts: omap4: Move l4 child devices to probe
them with ti-sysc") started producing a warning for pwm-omap-dmtimer:

WARNING: CPU: 0 PID: 77 at drivers/bus/omap_l3_noc.c:147
l3_interrupt_handler+0x2f8/0x388
44000000.ocp:L3 Custom Error: MASTER MPU TARGET L4PER2 (Idle):
Data Access in Supervisor mode during Functional access
...
__pm_runtime_idle
omap_dm_timer_disable
pwm_omap_dmtimer_start
pwm_omap_dmtimer_enable
pwm_apply_state
pwm_vibrator_start
pwm_vibrator_play_work

This is because the timer that pwm-omap-dmtimer is using is now being
probed with ti-sysc interconnect target module instead of omap_device
and the ti-sysc quirk for SYSC_QUIRK_LEGACY_IDLE is not fully
compatible with what omap_device has been doing.

We could fix this by reverting the timer changes and have the timer
probe again with omap_device. Or we could add more quirk handling to
ti-sysc driver. But as these options don't work nicely as longer term
solutions, let's just make timers probe with ti-sysc without any
quirks.

To do this, all we need to do is remove quirks for timers for ti-sysc,
and drop the bogus pm_runtime_irq_safe() flag for timer-ti-dm.

We should not use pm_runtime_irq_safe() anyways for drivers as it will
take a permanent use count on the parent device blocking the parent
devices from idling and has been forcing ti-sysc driver to use a
quirk flag.

Note that we will move the timer data to DEBUG section later on in
clean-up patches.

Fixes: 84badc5ec5fc ("ARM: dts: omap4: Move l4 child devices to probe them with ti-sysc")
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: H. Nikolaus Schaller <hns@goldelico.com>
Cc: Keerthy <j-keerthy@ti.com>
Cc: Ladislav Michl <ladis@linux-mips.org>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Sebastian Reichel <sre@kernel.org>
Cc: Tero Kristo <t-kristo@ti.com>
Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Reported-by: H. Nikolaus Schaller <hns@goldelico.com>
Tested-By: Andreas Kemnade <andreas@kemnade.info>
Tested-By: H. Nikolaus Schaller <hns@goldelico.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoiwlwifi: mvm: fix A-MPDU reference assignment
Johannes Berg [Thu, 8 Nov 2018 08:51:56 +0000 (09:51 +0100)]
iwlwifi: mvm: fix A-MPDU reference assignment

[ Upstream commit 1f7698abedeeb3fef3cbcf78e16f925df675a179 ]

The current code assigns the reference, and then goes to increment
it if the toggle bit has changed. That way, we get

Toggle  0  0  0  0  1  1  1  1
ID      1  1  1  1  1  2  2  2

Fix that by assigning the post-toggle ID to get

Toggle  0  0  0  0  1  1  1  1
ID      1  1  1  1  2  2  2  2

Reported-by: Danny Alexander <danny.alexander@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Fixes: fbe4112791b8 ("iwlwifi: mvm: update mpdu metadata API")
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoarm64: dts: allwinner: h6: Move GIC device node fix base address ordering
Chen-Yu Tsai [Sun, 27 Jan 2019 16:39:30 +0000 (00:39 +0800)]
arm64: dts: allwinner: h6: Move GIC device node fix base address ordering

[ Upstream commit 52d9bcb3d0de3fa1e07aff3800f857836d30410d ]

The GIC device node was placed out of order in the initial device tree
submission. Move it so the nodes are correctly sorted by base address
again.

Fixes: e54be32d0273 ("arm64: allwinner: h6: add the basical Allwinner H6 DTSI file")
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoip_tunnel: Fix route fl4 init in ip_md_tunnel_xmit
wenxu [Tue, 22 Jan 2019 10:39:51 +0000 (18:39 +0800)]
ip_tunnel: Fix route fl4 init in ip_md_tunnel_xmit

[ Upstream commit 6e6b904ad4f9aed43ec320afbd5a52ed8461ab41 ]

Init the gre_key from tuninfo->key.tun_id and init the mark
from the skb->mark, set the oif to zero in the collect metadata
mode.

Fixes: cfc7381b3002 ("ip_tunnel: add collect_md mode to IPIP tunnel")
Signed-off-by: wenxu <wenxu@ucloud.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet/mlx5: Take lock with IRQs disabled to avoid deadlock
Moni Shoua [Wed, 26 Dec 2018 17:21:21 +0000 (19:21 +0200)]
net/mlx5: Take lock with IRQs disabled to avoid deadlock

[ Upstream commit 33814e5d127e21f53b52e17b0722c1b57d4f4d29 ]

The lock in qp_table might be taken from process context or from
interrupt context. This may lead to a deadlock unless it is taken with
IRQs disabled.

Discovered by lockdep

================================
WARNING: inconsistent lock state
4.20.0-rc6
--------------------------------
inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W}

python/12572 [HC1[1]:SC0[0]:HE0:SE1] takes:
00000000052a4df4 (&(&table->lock)->rlock#2){?.+.}, /0x50 [mlx5_core]
{HARDIRQ-ON-W} state was registered at:
  _raw_spin_lock+0x33/0x70
  mlx5_get_rsc+0x1a/0x50 [mlx5_core]
  mlx5_ib_eqe_pf_action+0x493/0x1be0 [mlx5_ib]
  process_one_work+0x90c/0x1820
  worker_thread+0x87/0xbb0
  kthread+0x320/0x3e0
  ret_from_fork+0x24/0x30
irq event stamp: 103928
hardirqs last  enabled at (103927): [] nk+0x1a/0x1c
hardirqs last disabled at (103928): [] unk+0x1a/0x1c
softirqs last  enabled at (103924): [] tcp_sendmsg+0x31/0x40
softirqs last disabled at (103922): [] 80

other info that might help us debug this:
 Possible unsafe locking scenario:

       CPU0
       ----
  lock(&(&table->lock)->rlock#2);

    lock(&(&table->lock)->rlock#2);

 *** DEADLOCK ***

Fixes: 032080ab43ac ("IB/mlx5: Lock QP during page fault handling")
Signed-off-by: Moni Shoua <monis@mellanox.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoiwlwifi: mvm: avoid possible access out of array.
Mordechay Goodstein [Sun, 21 Oct 2018 15:27:26 +0000 (18:27 +0300)]
iwlwifi: mvm: avoid possible access out of array.

[ Upstream commit b0d795a9ae558209656b18930c2b4def5f8fdfb8 ]

The value in txq_id can be out of array scope,
validate it before accessing the array.

Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com>
Fixes: cf961e16620f ("iwlwifi: mvm: support dqa-mode agg on non-shared queue")
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoclk: sunxi-ng: sun8i-a23: Enable PLL-MIPI LDOs when ungating it
Chen-Yu Tsai [Fri, 25 Jan 2019 03:23:04 +0000 (11:23 +0800)]
clk: sunxi-ng: sun8i-a23: Enable PLL-MIPI LDOs when ungating it

[ Upstream commit 108a459ef4cd17a28711d81092044e597b5c7618 ]

The PLL-MIPI clock is somewhat special as it has its own LDOs which
need to be turned on for this PLL to actually work and output a clock
signal.

Add the 2 LDO enable bits to the gate bits.

Fixes: 5690879d93e8 ("clk: sunxi-ng: Add A23 CCU")
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agoARM: dts: sun8i-a23-a33: Move NAND controller device node to sort by address
Chen-Yu Tsai [Fri, 25 Jan 2019 03:23:10 +0000 (11:23 +0800)]
ARM: dts: sun8i-a23-a33: Move NAND controller device node to sort by address

[ Upstream commit d027521497592773cd23d016d36975574d3452db ]

The NAND controller device node was inserted into the wrong position,
probably due to a rebase or merge, as the file's structure does not
provide enough context for git to accurately match the previous device
node block.

Fixes: d7b843df13ea ("ARM: dts: sun8i: add NAND controller node for A23/A33")
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agonet: hns3: fix bug of ethtool_ops.get_channels for VF
Huazhong Tan [Tue, 22 Jan 2019 23:39:32 +0000 (07:39 +0800)]
net: hns3: fix bug of ethtool_ops.get_channels for VF

[ Upstream commit 8be7362186bd5ccb5f6f72be49751ad2778e2636 ]

The current code returns the number of all queues that can be used and
the number of queues that have been allocated, which is incorrect.
What should be returned is the number of queues allocated for each enabled
TC and the number of queues that can be allocated.

This patch fixes it.

Fixes: 849e46077689 ("net: hns3: add ethtool_ops.get_channels support for VF")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
4 years agospi/topcliff_pch: Fix potential NULL dereference on allocation error
YueHaibing [Wed, 23 Jan 2019 12:00:22 +0000 (20:00 +0800)]
spi/topcliff_pch: Fix potential NULL dereference on allocation error

[ Upstream commit e902cdcb5112b89ee445588147964723fd69ffb4 ]

In pch_spi_handle_dma, it doesn't check for NULL returns of kcalloc
so it would result in an Oops.

Fixes: c37f3c2749b5 ("spi/topcliff_pch: DMA support")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>