The current logic to deal with old DT missing the LVDS properties doesn't
take into account whether the LVDS output is supported in the first place,
resulting in spurious error messages on SoCs where it doesn't even matter.
Introduce a new TCON flag to list if LVDS is supported at all to prevent
this from happening.
Reviewed-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180221125703.4595-1-maxime.ripard@bootlin.com
ashmem_mutex may create a chain of dependencies like:
CPU0 CPU1
mmap syscall ioctl syscall
-> mmap_sem (acquired) -> ashmem_ioctl
-> ashmem_mmap -> ashmem_mutex (acquired)
-> ashmem_mutex (try to acquire) -> copy_from_user
-> mmap_sem (try to acquire)
There is a lock odering problem between mmap_sem and ashmem_mutex causing
a lockdep splat[1] during a syzcaller test. This patch fixes the problem
by move copy_from_user out of ashmem_mutex.
[1] https://www.spinics.net/lists/kernel/msg2733200.html
Fixes: ce8a3a9e76 (staging: android: ashmem: Fix a race condition in pin ioctls)
Reported-by: syzbot+d7a918a7a8e1c952bc36@syzkaller.appspotmail.com
Signed-off-by: Yisheng Xie <xieyisheng1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A rounding error was causing comedi_nsamples_left to
return the wrong value when nsamples was not a multiple
of the scan length.
Cc: <stable@vger.kernel.org> # v4.4+
Signed-off-by: Frank Mori Hess <fmh6jj@gmail.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
* Fix CSA issues with count 0 and 1;
* Some firmware debugging fixes;
* Removed a wrong error message when removing keys;
* Fix a firmware sysassert most usually triggered in IBSS;
* A couple of fixes on multicast queues;
* A fix with CCMP 256;
-----BEGIN PGP SIGNATURE-----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=6xo0
-----END PGP SIGNATURE-----
Merge tag 'iwlwifi-for-kalle-2018-03-02' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-fixes
Second batch of iwlwifi fixes intended for 4.16:
* Fix CSA issues with count 0 and 1;
* Some firmware debugging fixes;
* Removed a wrong error message when removing keys;
* Fix a firmware sysassert most usually triggered in IBSS;
* A couple of fixes on multicast queues;
* A fix with CCMP 256;
We actually can have the unbind/rebind logic in hid-core.c, leaving
only the match function in hid-generic.
This makes hid-generic simpler and the whole logic simpler too.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Currently when debugfs_create_dir() fails we receive a warning message
that provides no indication as to what was the directory entry that
failed to be created.
Improve the warning message by printing the directory name that failed
in order to help debugging.
Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The Synopsys PCIe Root Complex supports up to MSI 256 IRQs distributed
over 8 controller registers, therefore the maximum number of MSI IRQs
can be changed to 256. The number of controllers can be calculated based
on the number of vectors used by the specific SoC driver.
Update the dwc host bridge driver maximum number of supported MSI
IRQs.
Signed-off-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Tested-by: Niklas Cassel <niklas.cassel@axis.com>
Tested-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Remove the unused old MSI IRQs API from pcie-designware based on
struct msi_controller that should now be considered obsolete.
Signed-off-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Tested-by: Niklas Cassel <niklas.cassel@axis.com>
Tested-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
trigger_on() means that the trigger is available but not ready, however
trigger_on() was making it ready. That can segfault if the signal comes
before trigger_ready(). e.g. (USR2 signal delivery not shown)
$ perf record -e intel_pt//u -S sleep 1
perf: Segmentation fault
Obtained 16 stack frames.
/home/ahunter/bin/perf(sighandler_dump_stack+0x40) [0x4ec550]
/lib/x86_64-linux-gnu/libc.so.6(+0x36caf) [0x7fa76411acaf]
/home/ahunter/bin/perf(perf_evsel__disable+0x26) [0x4b9dd6]
/home/ahunter/bin/perf() [0x43a45b]
/lib/x86_64-linux-gnu/libc.so.6(+0x36caf) [0x7fa76411acaf]
/lib/x86_64-linux-gnu/libc.so.6(__xstat64+0x15) [0x7fa7641d2cc5]
/home/ahunter/bin/perf() [0x4ec6c9]
/home/ahunter/bin/perf() [0x4ec73b]
/home/ahunter/bin/perf() [0x4ec73b]
/home/ahunter/bin/perf() [0x4ec73b]
/home/ahunter/bin/perf() [0x4eca15]
/home/ahunter/bin/perf(machine__create_kernel_maps+0x257) [0x4f0b77]
/home/ahunter/bin/perf(perf_session__new+0xc0) [0x4f86f0]
/home/ahunter/bin/perf(cmd_record+0x722) [0x43c132]
/home/ahunter/bin/perf() [0x4a11ae]
/home/ahunter/bin/perf(main+0x5d4) [0x427fb4]
Note, for testing purposes, this is hard to hit unless you add some sleep()
in builtin-record.c before record__open().
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
Cc: stable@vger.kernel.org
Fixes: 3dcc4436fa ("perf tools: Introduce trigger class")
Link: http://lkml.kernel.org/r/1519807144-30694-1-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Implement a multiplexed IRQ domain hierarchy API in the pcie-designware
host bridge driver that funnels all MSI IRQs into a single parent
interrupt, moving away from the obsolete struct msi_controller based
API.
Although the old implementation API is still available, pcie-designware
will now use the multiplexed IRQ domains hierarchical API.
Remove all existing dwc based host bridges MSI IRQs handlers, in that the
hierarchical API now handles MSI IRQs through the hierarchical/chained
MSI domain implementation.
Signed-off-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Tested-by: Niklas Cassel <niklas.cassel@axis.com>
Tested-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Jingoo Han <jingoohan1@gmail.com>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Sony's NSG-MR5U and NSG-MR7U remote controls have a full keyboard and a
touchpad. The keyboard is already supported by the existing Linux
kernel and drivers but the touchpad is not recognized. This patch adds
the coded needed to bring full functionality to the touchpad.
Note that these remotes use the vendor code for SMK even though they are
Sony branded.
Known limitations
- The built-in accelerometers are not supported by these changes
- When the Drag (Fn) key is used as a mouse button, the button is
automatically released when the key begins repeating. There are two
workarounds for this 1) Use the button behind the touchpad instead of
the Drag (Fn) key or 2) Disable the key repeat functionality or
increase the key repeat delay.
Signed-off-by: Todd Kelner <tsopdump@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Free map->debugfs_name when debugfs_create_dir() failed to avoid memory
leak.
Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Assign true or false to boolean variables instead of an integer value.
This issue was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Acked-by: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Assign true or false to boolean variables instead of an integer value.
This issue was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Assign true or false to boolean variables instead of an integer value.
This issue was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
This patch enables the 6th button on the ELECOM EX-G M-XT4DRBK
trackball mouse.
This is a left handed EX-G variant which only comes in a wireless (D)
model. It has a total of 6 buttons but one of these is not available
because of how the HID descriptor is configured.
Signed-off-by: Tomasz Kramkowski <tk@the-tk.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
This patch changes all references to ELECOM trackballs using their
series name to refer to them by their model name.
ELECOM provides multiple series of mice such as EX-G, HUGE, and DEFT.
Although it has not caused conflicts in the driver, there can be more
than one iteration of mice in each series. For example, there are 7
variants of EX-G trackballs but only three (M-XT3URBK, M-XT3DRBK, and
M-XT4DRBK) need a driver to work correctly.
There are also 4 DEFT series trackballs but two of them have the same
VID:PID as the other two. I picked the earlier model for the naming of
the PID macros.
Signed-off-by: Tomasz Kramkowski <tk@the-tk.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Move the ts bus control function moved higher, enabling it
after configuring frontend and removeing ts_bus_ctrl callback.
While here, also add an error checking and re-add a comment
that were removed by commit 445877742c ("media: si2168:
Add ts bus coontrol, turn off bus on sleep").
[mchehab@s-opensource.com: I ended by applying the first version,
instead of the right one. So, this patch contains the diff and
the v2 changelog instead]
Fixes: 445877742c ("media: si2168: Add ts bus coontrol, turn off bus on sleep")
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Prevent auxtrace_queues__process_index() from queuing AUX area data for
decoding when the --no-itrace option has been used.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1520327598-1317-3-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Rework both DiSEqC send functions (send_master_cmd() and send_burst()) to
utilise the new SET_REG() and SET_FIELD() macros. Esp. due to SET_FIELD(),
this makes sure that not all bits (with unrelated purposes) are always
rewritten, but only those needed for sending DiSEqC commands. In
send_burst(), this makes sure that DISEQC_MODE isn't changed from 3 to 2
inbetween when sending SEC_MINI_A. Also, change both functions to write
DISEQC_MODE first before setting DIS_PRECHARGE. This makes diseqc control
work more reliable for "fullblown" DiSEqC strings in VDR's diseqc.conf in
combination with certain multiswitches.
Fixes: 448461af0e ("media: dvb-frontends/stv0910: implement diseqc_send_burst")
Reported-by: Helmut Auer <post@helmutauer.de>
Cc: Ralph Metzler <rjkm@metzlerbros.de>
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Tested-by: Helmut Auer <post@helmutauer.de>
Tested-by: Richard Scobie <rascobie@slingshot.co.nz>
Tested-by: Dietmar Spingler <d_spingler@freenet.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
When printing stats in CSV mode, 'perf stat' appends extra separators
when a counter is not supported:
<not supported>,,L1-dcache-store-misses,mesos/bd442f34-2b4a-47df-b966-9b281f9f56fc,0,100.00,,,,
Which causes a failure when parsing fields. The numbers of separators
should be the same for each line, no matter if the counter is or not
supported.
Signed-off-by: Ilya Pronin <ipronin@twitter.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/20180306064353.31930-1-xiyou.wangcong@gmail.com
Fixes: 92a61f6412 ("perf stat: Implement CSV metrics output")
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
According to the TODO file, this driver only landed in staging because of
the way device nodes and data transfers are handled. Besides that this way
(use of secX devices) has become sort of standard to date (ie. VDR
supports this literally since ages via the ddci plugin, TVHeadend received
this functionality lately, and minisatip being currently worked on
regarding this), most importantly this I2C client only driver isn't even
responsible for setting up device nodes, not for handling data
transfer and so on, but only serves as interface for the dvb_ca_en50221
subsystem, just like every other DVB card out in the wild, with hard-wired
or such flexible CA interfaces. And, it would even work with cards having
the cxd2099 controller hard-wired.
Also, this driver received quite some love and even is a proper I2C client
driver by now. So, as this driver acts as a EN50221 frontend device, move
it to dvb-frontends. There is no need to keep it buried in staging.
This commit also updates all affected Kconfig and Makefile's, and adds
MEDIA_AUTOSELECT depends to ddbridge and ngene.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Jasmin Jessich <jasmin@anw.at>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The dock line-out pin (NID 0x17 of ALC3254 codec) on Dell Precision
7520 may route to three different DACs, 0x02, 0x03 and 0x06. The
first two DACS have the volume amp controls while the last one
doesn't. And unfortunately, the auto-parser assigns this pin to DAC3,
resulting in the non-working volume control for the line out.
Fix it by disabling the routing to DAC3 on the corresponding pin.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=199029
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Change the way the cxd2099 hardware is being attached to the new I2C
client interface way.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Jasmin Jessich <jasmin@anw.at>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Convert the cxd2099 driver to use regmap for I2C accesses, removing all
own i2c_*() functions. With that, make the driver a proper I2C client
driver. This also adds the benefit of having a proper cleanup function
(cxd2099_remove() in this case) that takes care of resource cleanup
upon I2C client deregistration.
At this point, keep the static inline declared cxd2099_attach()
function so that drivers using the legacy/proprietary style attach way
still compile, albeit lacking the cxd2099 driver functionality. This
is taken care of in the next two patches.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Jasmin Jessich <jasmin@anw.at>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Deduplicate the checks for a valid ptr in port->en, and also handle the
default case to also catch eventually yet unsupported CI hardware.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Even if we don't have extended SCA support, we can have more than 64 CPUs
if we don't enable any HW features that might use the SCA entries.
Now, this works just fine, but we missed a return, which is why we
would actually store the SCA entries. If we have more than 64 CPUs, this
means writing outside of the basic SCA - bad.
Let's fix this. This allows > 64 CPUs when running nested (under vSIE)
without random crashes.
Fixes: a6940674c3 ("KVM: s390: allow 255 VCPUs when sca entries aren't used")
Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180306132758.21034-1-david@redhat.com>
Cc: stable@vger.kernel.org
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Due to a typo, the mask was destroyed by a comparison instead of a bit
shift.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The PLAT_S5P symbol was removed in commit d78c16ccde ("ARM: SAMSUNG:
Remove remaining legacy code").
Remove the PLAT_S5P dependency from VIDEO_SAMSUNG_S5P_CEC.
Discovered with the
https://github.com/ulfalizer/Kconfiglib/blob/master/examples/list_undefined.py
script.
Signed-off-by: Ulf Magnusson <ulfalizer@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Thoses files are unused since commit b3b961448f ("V4L/DVB (13795): [Mantis/Hopper] Code overhaul, add Hopper devices into the PCI ID list")
8 year after, we could remove it.
Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
bsp_i2c.h is unused since commit ffe7c4f921 ("[media] drx-j: Get rid of drx39xyj/bsp_tuner.h")
Remove it from tree.
Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Instead of directly accessing to dmadev->device_prep_interleaved_dma() use
the dmaengine_prep_interleaved_dma() wrapper instead.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hauppauge Solo/Dual HD DVB models use a si2157 tuner, which is set to
produce inverted spectrum. This configures the si2168 DVB demod for
inverted spectrum on both affected models.
[mchehab@s-opensource.com: rebased on the top of upstream]
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Some tuners produce inverted spectrum, but the si2168 is not
currently set up to accept it. This adds an optional parameter
to set the frontend up to receive inverted spectrum.
Parameter is optional and only boards who enable inversion
will utilize this.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
While experimenting with older compiler versions, I ran
into a warning that no longer shows up on gcc-4.8 or newer:
drivers/media/platform/s3c-camif/camif-capture.c: In function '__camif_subdev_try_format':
drivers/media/platform/s3c-camif/camif-capture.c:1265:25: error: array subscript is below array bounds
This is an off-by-one bug, leading to an access before the start of the
array, while newer compilers silently assume this undefined behavior
cannot happen and leave the loop at index 0 if no other entry matches.
As Sylvester explains, we actually need to ensure that the
value is within the range, so this reworks the loop to be
easier to parse correctly, and an additional check to fall
back on the first format value for any unexpected input.
I found an existing gcc bug for it and added a reduced version
of the function there.
Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69249#c3
Fixes: babde1c243 ("[media] V4L: Add driver for S3C24XX/S3C64XX SoC series camera interface")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The driver outputs on success, but is silent on failure. Give
one message that probe failed.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The driver is near silent, this adds a simple announcement at the
end of probe after the chip has been detected and upgrades a debug
message to error if probe has failed.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Includes a function to set TS MODE property os si2168. The function
either disables the TS output bus, or sets mode to config option.
When going to sleep the TS bus is turned off, this makes the driver
compatible with multiple frontend usage.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Remove yield() call. In this case it's use is considered broken, since
it is being assumed that yield() will let another process run that will
make the event true.
Signed-off-by: Thomas Avery <tavery321@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This reverts commit 16f1eeb660.
The reason for this patch was that lustre used copy_from_user_page.
Commit 76133e66b1 ("staging/lustre: Replace jobid acquiring with per
node setting") removed that usage.
So the arch restrictions can go.
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove restriction the lustre must be built
as modules. It now works as a monolithic build.
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When the ptlrpc module is loaded, it starts the pinger thread and
calls LNetNIInit which starts various threads.
We don't need these threads until the module is actually being
used, such as when a lustre filesystem is mounted.
So move the thread creation into new ptlrpc_inc_ref() (modeled on
ptlrpcd_inc_ref()), and call that when needed, such as at mount time.
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>