Hugues Fruchet [Mon, 18 Jun 2018 10:29:17 +0000 (06:29 -0400)]
media: ov5640: add HFLIP/VFLIP controls support
Add HFLIP/VFLIP controls support by setting registers REG21/REG20.
Useless values in hardcoded mode sequences are removed and
remaining binning values are now set after mode sequence being set.
Note that due to BSI (Back Side Illuminated) technology, image capture
is physically mirrored, mirror logic is so inversed in REG21 register
to cancel this effect.
Sakari Ailus [Wed, 2 May 2018 20:59:00 +0000 (16:59 -0400)]
media: smiapp: Support the "rotation" property
Use the "rotation" property to tell that the sensor is mounted upside
down. This reverses the behaviour of the VFLIP and HFLIP controls as well
as the pixel order.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.co.uk> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Sakari Ailus [Wed, 2 May 2018 20:47:20 +0000 (16:47 -0400)]
media: dt-bindings: media: Define "rotation" property for sensors
Sensors are occasionally mounted upside down to systems such as mobile
phones or tablets. In order to use such a sensor without having to turn
every image upside down, most camera sensors support reversing the readout
order by setting both horizontal and vertical flipping.
This patch documents the "rotation" property for camera sensors, mirroring
what is defined for displays in
Documentation/devicetree/bindings/display/panel/panel.txt .
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.co.uk> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hans Verkuil [Fri, 15 Jun 2018 13:19:46 +0000 (09:19 -0400)]
media: media.h: remove __NEED_MEDIA_LEGACY_API
The __NEED_MEDIA_LEGACY_API define is 1) ugly and 2) dangerous
since it is all too easy for drivers to define it to get hold of
legacy defines. Instead just define what we need in media-device.c
which is the only place where we need the legacy define
(MEDIA_ENT_T_DEVNODE_UNKNOWN).
The v4l uAPI documentation [0] makes clear that in the case of interlaced
video (i.e: field is V4L2_FIELD_ALTERNATE) the height refers to the number
of lines in the field and not the number of lines in the full frame (which
is twice the field height for interlaced formats).
So the original height calculation was correct, and it shouldn't had been
changed by the mentioned commit.
Fixes: 0866df8dffd5 ("[media] tvp5150: fix pad format frame height") Signed-off-by: Javier Martinez Canillas <javierm@redhat.com> Cc: <stable@vger.kernel.org> # for v4.12 and up Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Commit f1a81afc98e3 ("[media] m2m: fix bad unlock balance")
removed the last use of v4l2_m2m_ops.lock and
v4l2_m2m_ops.unlock hooks. They are not actually
used anymore. Remove them.
Ezequiel Garcia [Fri, 1 Jun 2018 19:49:52 +0000 (15:49 -0400)]
media: rockchip/rga: Remove unrequired wait in .job_abort
As per the documentation, job_abort is not required
to wait until the current job finishes. It is redundant
to do so, as the core will perform the wait operation.
Remove the wait infrastructure completely.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com> Reviewed-by: Jacob Chen <jacob-chen@iotwrt.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Ezequiel Garcia [Fri, 1 Jun 2018 19:49:51 +0000 (15:49 -0400)]
media: rockchip/rga: Fix broken .start_streaming
Currently, rga_buf_start_streaming() is expecting
pm_runtime_get_sync to return zero on success, which
is wrong.
As per the documentation, pm_runtime_get_sync increments
the device's usage counter and return its result.
This means it will typically return a positive integer
on success and a negative error code.
Therefore, rockchip-rga driver is currently unusable
failing to start_streaming in most cases. Fix it and
while here, cleanup the buffer return-to-core logic.
Fixes: f7e7b48e6d79 ("[media] rockchip/rga: v4l2 m2m support") Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com> Reviewed-by: Jacob Chen <jacob-chen@iotwrt.com>
[hans.verkuil@cisco.com: fix line over 80 cols warning] Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: venus: keep resolution when adjusting format
When checking a format for validity, the resolution is reset to 1280x720
whenever the pixel format is not supported. This behavior can mislead
user-space into believing that this is the only resolution supported,
and looks strange considering that if we try/set the same format with
just the pixel format changed to a valid one, the call will this time
succeed without altering the resolution.
Resolution is managed independently of the pixel format, so remove this
reset.
Colin Ian King [Tue, 12 Jun 2018 09:42:00 +0000 (05:42 -0400)]
media: cx18: remove redundant zero check on retval
The check for a zero retval is redundant as all paths that lead to
this point have set retval to an error return value that is non-zero.
Remove the redundant check.
Detected by CoverityScan, CID#102589 ("Logically dead code")
Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Ezequiel Garcia [Tue, 12 Jun 2018 13:22:51 +0000 (09:22 -0400)]
media: mem2mem: Remove excessive try_run call
If there is a schedulable job, v4l2_m2m_try_schedule() calls
v4l2_m2m_try_run(). This makes the unconditional v4l2_m2m_try_run()
called by v4l2_m2m_job_finish superfluous. Remove it.
Trivial fix to spelling mistake in pr_notice message text
Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Corentin Labbe [Thu, 7 Jun 2018 20:02:21 +0000 (16:02 -0400)]
media: cx25821: remove cx25821-audio-upstream.c and cx25821-video-upstream.c
Those two files are unused since commit 486a7a2813c7 ("[media] cx25821: remove cx25821-audio-upstream.c")
and commit b6f21dc3541a ("[media] cx25821: remove video output support")
Remove them from tree.
This does not fix any bug - this is just a code simplification. As
np is not used after passing it to v4l2_fwnode_endpoint_parse() its
refcount can be decremented immediately and at one location.
Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: stm32-dcmi: add mandatory of_node_put() in success path
The endpoint allocated by of_graph_get_next_endpoint() needs an of_node_put()
in both error and success path. As ep is not used the refcount decrement
can be right after the last use of ep.
Fixes: commit 37404f91ef8b ("[media] stm32-dcmi: STM32 DCMI camera interface driver") Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The while(1) is effectively useless as all possible paths within it
return thus there is no way to loop.
Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hugues Fruchet [Mon, 11 Jun 2018 09:50:24 +0000 (05:50 -0400)]
media: stm32-dcmi: do not fall into error on buffer starvation
Return silently instead of falling into error when running
out of available buffers when restarting capture.
Capture will be restarted when new buffers will be
provided by V4L2 client.
Niklas Söderlund [Thu, 17 May 2018 14:30:15 +0000 (10:30 -0400)]
media: v4l2-ioctl: create helper to fill in v4l2_standard for ENUMSTD
Prepare for adding a new IOCTL VIDIOC_SUBDEV_ENUMSTD which would
enumerate the standards for a subdevice by breaking out the code which
could be shared between the video and subdevice versions of this IOCTL.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
[hans.verkuil@cisco.com: fixed 'sdandard' typos in v4l2-ioctl.h] Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Jacopo Mondi [Mon, 28 May 2018 16:37:11 +0000 (12:37 -0400)]
media: arch: sh: ap325rxa: Use new renesas-ceu camera driver
Use the new renesas-ceu camera driver in ap325rxa board file instead of
the soc_camera based sh_mobile_ceu_camera driver.
Get rid of soc_camera specific components, and register CEU0 with a single
video sensor (ov7725).
Memory for the CEU video buffers is now reserved with membocks APIs
and need to be declared as dma_coherent during machine initialization to
remove that architecture specific part from CEU driver.
The ap325rxa board file registers another camera (ncm03j) for which I found no
driver in mainline kernel version, and that was configured/probed by sending
i2c messages (of 'magic blobs) from the board file itself. I removed the
sensor registration from this new version as it used soc_camera components
that will be later removed.
While at there update license to SPDX header and sort headers alphabetically.
Jacopo Mondi [Mon, 28 May 2018 16:37:10 +0000 (12:37 -0400)]
media: arch: sh: ms7724se: Use new renesas-ceu camera driver
Use the new renesas-ceu camera driver is ms7724se board file instead of
the soc_camera based sh_mobile_ceu_camera driver.
Get rid of soc_camera specific components, and register CEU0 and CEU1 with
no active video subdevices.
Memory for the CEU video buffers is now reserved with membocks APIs
and need to be declared as dma_coherent during machine initialization to
remove that architecture specific part from CEU driver.
While at there update license to SPDX header and sort headers
alphabetically.
No need to udapte the clock source names, as
commit c2f9b05fd5c1 ("media: arch: sh: ecovec: Use new renesas-ceu camera driver")
already updated it to the new ceu driver name for all SH7724 boards
(possibly breaking ms7724se before this commit).
Jacopo Mondi [Mon, 28 May 2018 16:37:09 +0000 (12:37 -0400)]
media: arch: sh: kfr2r09: Use new renesas-ceu camera driver
Use the new renesas-ceu camera driver in kfr2r09 board file instead of
the soc_camera based sh_mobile_ceu_camera driver.
Get rid of soc_camera specific components, and move clk and gpio handling
away from board file, registering the clock source and the enable gpios
for driver consumption.
Memory for the CEU video buffers is now reserved with membocks APIs,
and need to be declared as dma_coherent during machine initialization to
remove that architecture specific part from CEU driver.
While at there update license to SPDX header and sort headers alphabetically.
No need to udapte the clock source names, as
commit c2f9b05fd5c1 ("media: arch: sh: ecovec: Use new renesas-ceu camera driver")
already updated it to the new ceu driver name for all SH7724 boards (possibly
breaking kfr2r09 before this commit).
Remove soc_camera framework dependencies from rj54n1 sensor driver.
- Handle clock
- Handle GPIOs (named 'powerup' and 'enable')
- Register the async subdevice
- Remove g/s_mbus_config as they're deprecated.
- Adjust build system
- List the driver as maintained for 'Odd Fixes' as I don't have HW to test.
This commits does not remove the original soc_camera based driver.
Copy the soc_camera based driver in v4l2 sensor driver directory.
This commit just copies the original file without modifying it.
No modification to KConfig and Makefile as soc_camera framework
dependencies need to be removed first in next commit.
Niklas Söderlund [Mon, 28 May 2018 23:26:37 +0000 (19:26 -0400)]
media: rcar-vin: sync which hardware buffer to start capture from
When starting the VIN capture procedure we are not guaranteed that the
first buffer written to is VnMB1 to which we assigned the first buffer
queued. This is problematic for two reasons. Buffers might not be
dequeued in the same order they where queued for capture. Future
features planed for the VIN driver is support for outputting frames in
SEQ_TB/BT format and to do that it's important that capture starts from
the first buffer slot, VnMB1.
We are guaranteed that capturing always happens in sequence (VnMB1 ->
VnMB2 -> VnMB3 -> VnMB1). So drop up to two frames when starting
capturing so that the driver always returns buffers in the same order
they are queued and prepare for SEQ_TB/BT output.
Jacopo Mondi [Wed, 6 Jun 2018 13:21:53 +0000 (09:21 -0400)]
media: renesas-ceu: Add support for YUYV permutations
Add support for YUYV permutations, passing down to the subdevice the media bus
format with components ordering that reflect the output pixel format one.
media: atmel-isi: move of_node_put() to cover success branch as well
The of_node_put() was only covering the error branch but missed the
success branch so the refcount for ep which
of_graph_get_remote_port_parent() incremented on success would was
not being decremented.
Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org> Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
As there is no way this can loop it actually makes no sense to have
a while(1){} around the body - all three possible paths end in a return
statement.
Fixes: commit c1d82b895380 "[media] atmel-isi: move out of soc_camera to atmel" Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org> Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
As the of_node_put() is unconditional here there is no need to have it
twice.
Signed-off-by: Nicholas Mc Guire <hofrat@opentech.at> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Arnd Bergmann [Wed, 30 May 2018 22:07:11 +0000 (18:07 -0400)]
media: v4l: omap: add VIDEO_V4L2 dependency
The omap media driver can be built-in while the v4l2 core is a loadable
module. This is a mistake and leads to link errors:
drivers/media/platform/omap/omap_vout.o: In function `omap_vout_remove':
omap_vout.c:(.text+0xec): undefined reference to `v4l2_device_unregister'
omap_vout.c:(.text+0x140): undefined reference to `video_device_release'
omap_vout.c:(.text+0x150): undefined reference to `video_unregister_device'
omap_vout.c:(.text+0x15c): undefined reference to `v4l2_ctrl_handler_free'
An explicit Kconfig dependency on VIDEO_V4L2 avoids the problem.
I ran into this problem for the first time today during my randconfig
builds, but could not find what caused it.
Recently binding of the Memory Controller has been extended, exposing
the Memory Client reset controls and hence it is now a reset controller.
Tegra video-decoder device is among the Memory Controller reset users,
document the new optional VDE HW reset property.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com> Acked-by: Rob Herring <robh@kernel.org> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Trivial fix to spelling mistake in module description text.
Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: v4l: rcar_fdp1: Change platform dependency to ARCH_RENESAS
The Renesas Fine Display Processor driver is used on Renesas R-Car SoCs
only. Since commit 9b5ba0df4ea4f940 ("ARM: shmobile: Introduce
ARCH_RENESAS") is ARCH_RENESAS a more appropriate platform dependency
than the legacy ARCH_SHMOBILE, hence use the former.
This will allow to drop ARCH_SHMOBILE on ARM and ARM64 in the near
future.
Arnd Bergmann [Wed, 30 May 2018 22:07:10 +0000 (18:07 -0400)]
media: v4l: cadence: add VIDEO_V4L2 dependency
The cadence media drivers can be built-in while the v4l2 core is a loadable
module. This is a mistake and leads to link errors:
drivers/media/v4l2-core/v4l2-fwnode.o: In function `v4l2_async_register_subdev_sensor_common':
v4l2-fwnode.c:(.text+0x12f0): undefined reference to `v4l2_async_subdev_notifier_register'
v4l2-fwnode.c:(.text+0x1304): undefined reference to `v4l2_async_register_subdev'
v4l2-fwnode.c:(.text+0x1318): undefined reference to `v4l2_async_notifier_unregister'
v4l2-fwnode.c:(.text+0x1338): undefined reference to `v4l2_async_notifier_cleanup'
cdns-csi2rx.c:(.text+0x9f8): undefined reference to `v4l2_subdev_init'
cdns-csi2rx.c:(.text+0xa78): undefined reference to `v4l2_async_register_subdev'
drivers/media/platform/cadence/cdns-csi2tx.o: In function `csi2tx_remove':
cdns-csi2tx.c:(.text+0x88): undefined reference to `v4l2_async_unregister_subdev'
drivers/media/platform/cadence/cdns-csi2tx.o: In function `csi2tx_probe':
cdns-csi2tx.c:(.text+0x884): undefined reference to `v4l2_subdev_init'
cdns-csi2tx.c:(.text+0xa9c): undefined reference to `v4l2_async_register_subdev'
An explicit Kconfig dependency on VIDEO_V4L2 avoids the problem.
Arnd Bergmann [Fri, 25 May 2018 15:25:09 +0000 (11:25 -0400)]
media: v4l: cadence: include linux/slab.h
I ran into a randconfig build error with the new driver:
drivers/media/platform/cadence/cdns-csi2tx.c: In function 'csi2tx_probe':
drivers/media/platform/cadence/cdns-csi2tx.c:477:11: error: implicit declaration of function 'kzalloc'; did you mean 'd_alloc'? [-Werror=implicit-function-declaration]
kzalloc() is declared in linux/slab.h, so let's include this to make it
build in all configurations.
Sakari Ailus [Wed, 23 May 2018 09:33:32 +0000 (05:33 -0400)]
media: imx258: Check the rotation property has a value of 180
The driver only supports streaming images flipped horizontally and
vertically. In order to ensure that all current users will be fine if or
when support for upright streaming is added, require the presence of the
"rotation" control now.
Linus Torvalds [Sun, 24 Jun 2018 12:29:15 +0000 (20:29 +0800)]
Merge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf fixes from Thomas Gleixner:
"A pile of perf updates:
Kernel side:
- Remove an incorrect warning in uprobe_init_insn() when
insn_get_length() fails. The error return code is handled at the
call site.
- Move the inline keyword to the right place in the perf ringbuffer
code to address a W=1 build warning.
Tooling:
perf stat:
- Fix metric column header display alignment
- Improve error messages for default attributes, providing better
output for error in command line.
- Add --interval-clear option, to provide a 'watch' like printing
perf script:
- Show hw-cache events too
perf c2c:
- Fix data dependency problem in layout of 'struct c2c_hist_entry'
Core:
- Do not blindly assume that 'struct perf_evsel' can be obtained via
a straight forward container_of() as there are call sites which
hand in a plain 'struct hist' which is not part of a container.
- Fix error index in the PMU event parser, so that error messages can
point to the problematic token"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/core: Move the inline keyword at the beginning of the function declaration
uprobes/x86: Remove incorrect WARN_ON() in uprobe_init_insn()
perf script: Show hw-cache events
perf c2c: Keep struct hist_entry at the end of struct c2c_hist_entry
perf stat: Add event parsing error handling to add_default_attributes
perf stat: Allow to specify specific metric column len
perf stat: Fix metric column header display alignment
perf stat: Use only color_fprintf call in print_metric_only
perf stat: Add --interval-clear option
perf tools: Fix error index for pmu event parser
perf hists: Reimplement hists__has_callchains()
perf hists browser gtk: Use hist_entry__has_callchains()
perf hists: Make hist_entry__has_callchains() work with 'perf c2c'
perf hists: Save the callchain_size in struct hist_entry
Linus Torvalds [Sun, 24 Jun 2018 12:18:19 +0000 (20:18 +0800)]
Merge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull rseq fixes from Thomas Gleixer:
"A pile of rseq related fixups:
- Prevent infinite recursion when delivering SIGSEGV
- Remove the abort of rseq critical section on fork() as syscalls
inside rseq critical sections are explicitely forbidden. So no
point in doing the abort on the child.
- Align the rseq structure on 32 bytes in the ARM selftest code.
- Fix file permissions of the test script"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
rseq: Avoid infinite recursion when delivering SIGSEGV
rseq/cleanup: Do not abort rseq c.s. in child on fork()
rseq/selftests/arm: Align 'struct rseq_cs' on 32 bytes
rseq/selftests: Make run_param_test.sh executable
Linus Torvalds [Sun, 24 Jun 2018 12:16:17 +0000 (20:16 +0800)]
Merge branch 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull EFI fixes from Thomas Gleixner:
"Two fixlets for the EFI maze:
- Properly zero variables to prevent an early boot hang on EFI mixed
mode systems
- Fix the fallout of merging the 32bit and 64bit variants of EFI PCI
related code which ended up chosing the 32bit variant of the actual
EFi call invocation which leads to failures on 64bit"
* 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
efi/x86: Fix incorrect invocation of PciIo->Attributes()
efi/libstub/tpm: Initialize efi_physical_addr_t vars to zero for mixed mode
Linus Torvalds [Sun, 24 Jun 2018 12:06:42 +0000 (20:06 +0800)]
Merge branch 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull core fixes from Thomas Gleixner:
"Two tiny fixes:
- Add the missing machine_real_restart() to objtools noreturn list so
it stops complaining
- Fix a trivial comment typo"
* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
kernel.h: Fix a typo in comment
objtool: Add machine_real_restart() to the noreturn list
Linus Torvalds [Sun, 24 Jun 2018 11:59:52 +0000 (19:59 +0800)]
Merge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Thomas Gleixner:
"A set of fixes for x86:
- Make Xen PV guest deal with speculative store bypass correctly
- Address more fallout from the 5-Level pagetable handling. Undo an
__initdata annotation to avoid section mismatch and malfunction
when post init code would touch the freed variable.
- Handle exception fixup in math_error() before calling notify_die().
The reverse call order incorrectly triggers notify_die() listeners
for soemthing which is handled correctly at the site which issues
the floating point instruction.
- Fix an off by one in the LLC topology calculation on AMD
- Handle non standard memory block sizes gracefully un UV platforms
- Plug a memory leak in the microcode loader
- Sanitize the purgatory build magic
- Add the x86 specific device tree bindings directory to the x86
MAINTAINER file patterns"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/mm: Fix 'no5lvl' handling
Revert "x86/mm: Mark __pgtable_l5_enabled __initdata"
x86/CPU/AMD: Fix LLC ID bit-shift calculation
MAINTAINERS: Add file patterns for x86 device tree bindings
x86/microcode/intel: Fix memleak in save_microcode_patch()
x86/platform/UV: Add kernel parameter to set memory block size
x86/platform/UV: Use new set memory block size function
x86/platform/UV: Add adjustable set memory block size function
x86/build: Remove unnecessary preparation for purgatory
Revert "kexec/purgatory: Add clean-up for purgatory directory"
x86/xen: Add call of speculative_store_bypass_ht_init() to PV paths
x86: Call fixup_exception() before notify_die() in math_error()
Linus Torvalds [Sun, 24 Jun 2018 11:48:30 +0000 (19:48 +0800)]
Merge branch 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 pti fixes from Thomas Gleixner:
"Two small updates for the speculative distractions:
- Make it more clear to the compiler that array_index_mask_nospec()
is not subject for optimizations. It's not perfect, but ...
- Don't report XEN PV guests as vulnerable because their mitigation
state depends on the hypervisor. Report unknown and refer to the
hypervisor requirement"
* 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/spectre_v1: Disable compiler optimizations over array_index_mask_nospec()
x86/pti: Don't report XenPV as vulnerable
Linus Torvalds [Sun, 24 Jun 2018 11:36:16 +0000 (19:36 +0800)]
Merge branch 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull locking fixes from Thomas Gleixner:
"A set of fixes and updates for the locking code:
- Prevent lockdep from updating irq state within its own code and
thereby confusing itself.
- Buid fix for older GCCs which mistreat anonymous unions
- Add a missing lockdep annotation in down_read_non_onwer() which
causes up_read_non_owner() to emit a lockdep splat
- Remove the custom alpha dec_and_lock() implementation which is
incorrect in terms of ordering and use the generic one.
The remaining two commits are not strictly fixes. They provide irqsave
variants of atomic_dec_and_lock() and refcount_dec_and_lock(). These
are required to merge the relevant updates and cleanups into different
maintainer trees for 4.19, so routing them into mainline without
actual users is the sanest approach.
They should have been in -rc1, but last weekend I took the liberty to
just avoid computers in order to regain some mental sanity"
* 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
locking/qspinlock: Fix build for anonymous union in older GCC compilers
locking/lockdep: Do not record IRQ state within lockdep code
locking/rwsem: Fix up_read_non_owner() warning with DEBUG_RWSEMS
locking/refcounts: Implement refcount_dec_and_lock_irqsave()
atomic: Add irqsave variant of atomic_dec_and_lock()
alpha: Remove custom dec_and_lock() implementation
Linus Torvalds [Sun, 24 Jun 2018 11:22:19 +0000 (19:22 +0800)]
Merge branch 'ras-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull ras fixes from Thomas Gleixner:
"A set of fixes for RAS/MCE:
- Improve the error message when the kernel cannot recover from a MCE
so the maximum amount of information gets provided.
- Individually check MCE recovery features on SkyLake CPUs instead of
assuming none when the CAPID0 register does not advertise the
general ability for recovery.
- Prevent MCE to output inconsistent messages which first show an
error location and then claim that the source is unknown.
- Prevent overwriting MCi_STATUS in the attempt to gather more
information when a fatal MCE has alreay been detected. This leads
to empty status values in the printout and failing to react
promptly on the fatal event"
* 'ras-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/mce: Fix incorrect "Machine check from unknown source" message
x86/mce: Do not overwrite MCi_STATUS in mce_no_way_out()
x86/mce: Check for alternate indication of machine check recovery on Skylake
x86/mce: Improve error message when kernel cannot recover
Linus Torvalds [Sun, 24 Jun 2018 11:16:42 +0000 (19:16 +0800)]
Merge branch 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull timer fixes from Thomas Gleixner:
"A small set of fixes for time(r) related issues:
- Fix a long standing conversion issue in jiffies_to_msecs() for odd
HZ values like 1024 or 1200 which resulted in returning 0 for small
jiffies values due to rounding down.
- Use the proper CONFIG symbol in the new Y2038 safe compat code for
posix-timers. Not yet a visible breakage, but this will immediately
trigger when the architecture support for the new interfaces is
merged.
- Return an error code in the STM32 clocksource driver on failure
instead of success.
- Remove the redundant and stale irq disabled check in the posix cpu
timer code. The check is at the wrong place anyway and lockdep
already covers it via the sighand lock locking coverage"
* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
time: Make sure jiffies_to_msecs() preserves non-zero time periods
posix-timers: Fix nanosleep_copyout() for CONFIG_COMPAT_32BIT_TIME
clocksource/drivers/stm32: Fix error return code
posix-cpu-timers: Remove lockdep_assert_irqs_disabled()
Linus Torvalds [Sun, 24 Jun 2018 11:01:18 +0000 (19:01 +0800)]
Merge branch 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull irq fixes from Thomas Gleixner:
"A set of fixes mostly for the ARM/GIC world:
- Fix the MSI affinity handling in the ls-scfg irq chip driver so it
updates and uses the effective affinity mask correctly
- Prevent binding LPIs to offline CPUs and respect the Cavium erratum
which requires that LPIs which belong to an offline NUMA node are
not bound to a CPU on a different NUMA node.
- Free only the amount of allocated interrupts in the GIC-V2M driver
instead of trying to free log2(nrirqs).
- Prevent emitting SYNC and VSYNC targetting non existing interrupt
collections in the GIC-V3 ITS driver
- Ensure that the GIV-V3 interrupt redistributor is correctly
reprogrammed on CPU hotplug
- Remove a stale unused helper function"
* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
irqdesc: Delete irq_desc_get_msi_desc()
irqchip/gic-v3-its: Fix reprogramming of redistributors on CPU hotplug
irqchip/gic-v3-its: Only emit VSYNC if targetting a valid collection
irqchip/gic-v3-its: Only emit SYNC if targetting a valid collection
irqchip/gic-v3-its: Don't bind LPI to unavailable NUMA node
irqchip/gic-v2m: Fix SPI release on error path
irqchip/ls-scfg-msi: Fix MSI affinity handling
genirq/debugfs: Add missing IRQCHIP_SUPPORTS_LEVEL_MSI debug
Linus Torvalds [Sun, 24 Jun 2018 09:19:42 +0000 (17:19 +0800)]
Merge tag 'mips_fixes_4.18_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux
Pull MIPS fixes from Paul Burton:
"A few MIPS fixes for 4.18:
- a GPIO device name fix for a regression in v4.15-rc1.
- an errata workaround for the BCM5300X platform.
- a fix to ftrace function graph tracing, broken for a long time with
the fix applying cleanly back as far as v3.17.
- addition of read barriers to in{b,w,l,q}() functions, matching
behavior of other architectures & mirroring the equivalent addition
to read{b,w,l,q} in v4.17-rc2.
Plus changes to wire up new syscalls introduced in the 4.18 cycle:
- Restartable sequences support is added, including MIPS support in
the selftests.
- io_pgetevents is wired up"
* tag 'mips_fixes_4.18_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
MIPS: Wire up io_pgetevents syscall
rseq/selftests: Implement MIPS support
MIPS: Wire up the restartable sequences (rseq) syscall
MIPS: Add syscall detection for restartable sequences
MIPS: Add support for restartable sequences
MIPS: io: Add barrier after register read in inX()
mips: ftrace: fix static function graph tracing
MIPS: BCM47XX: Enable 74K Core ExternalSync for PCIe erratum
MIPS: pb44: Fix i2c-gpio GPIO descriptor table
Ard Biesheuvel [Sat, 23 Jun 2018 21:19:03 +0000 (23:19 +0200)]
efi/x86: Fix incorrect invocation of PciIo->Attributes()
The following commit:
2c3625cb9fa2 ("efi/x86: Fold __setup_efi_pci32() and __setup_efi_pci64() into one function")
... merged the two versions of __setup_efi_pciXX(), without taking into
account that the 32-bit version used a rather dodgy trick to pass an
immediate 0 constant as argument for a uint64_t parameter.
The issue is caused by the fact that on x86, UEFI protocol method calls
are redirected via struct efi_config::call(), which is a variadic function,
and so the compiler has to infer the types of the parameters from the
arguments rather than from the prototype.
As the 32-bit x86 calling convention passes arguments via the stack,
passing the unqualified constant 0 twice is the same as passing 0ULL,
which is why the 32-bit code in __setup_efi_pci32() contained the
following call:
status = efi_early->call(pci->attributes, pci,
EfiPciIoAttributeOperationGet, 0, 0,
&attributes);
to invoke this UEFI protocol method:
typedef
EFI_STATUS
(EFIAPI *EFI_PCI_IO_PROTOCOL_ATTRIBUTES) (
IN EFI_PCI_IO_PROTOCOL *This,
IN EFI_PCI_IO_PROTOCOL_ATTRIBUTE_OPERATION Operation,
IN UINT64 Attributes,
OUT UINT64 *Result OPTIONAL
);
After the merge, we inadvertently ended up with this version for both
32-bit and 64-bit builds, breaking the latter.
So replace the two zeroes with the explicitly typed constant 0ULL,
which works as expected on both 32-bit and 64-bit builds.
Wilfried tested the 64-bit build, and I checked the generated assembly
of a 32-bit build with and without this patch, and they are identical.
Linus Torvalds [Sat, 23 Jun 2018 22:33:54 +0000 (06:33 +0800)]
Merge tag 'for-linus-20180623' of git://git.kernel.dk/linux-block
Pull block fixes from Jens Axboe:
- Further timeout fixes. We aren't quite there yet, so expect another
round of fixes for that to completely close some of the IRQ vs
completion races. (Christoph/Bart)
- Set of NVMe fixes from the usual suspects, mostly error handling
- Two off-by-one fixes (Dan)
- Another bdi race fix (Jan)
- Fix nbd reconfigure with NBD_DISCONNECT_ON_CLOSE (Doron)
* tag 'for-linus-20180623' of git://git.kernel.dk/linux-block:
blk-mq: Fix timeout handling in case the timeout handler returns BLK_EH_DONE
bdi: Fix another oops in wb_workfn()
lightnvm: Remove depends on HAS_DMA in case of platform dependency
nvme-pci: limit max IO size and segments to avoid high order allocations
nvme-pci: move nvme_kill_queues to nvme_remove_dead_ctrl
nvme-fc: release io queues to allow fast fail
nbd: Add the nbd NBD_DISCONNECT_ON_CLOSE config flag.
block: sed-opal: Fix a couple off by one bugs
blk-mq-debugfs: Off by one in blk_mq_rq_state_name()
nvmet: reset keep alive timer in controller enable
nvme-rdma: don't override opts->queue_size
nvme-rdma: Fix command completion race at error recovery
nvme-rdma: fix possible free of a non-allocated async event buffer
nvme-rdma: fix possible double free condition when failing to create a controller
Revert "block: Add warning for bi_next not NULL in bio_endio()"
block: fix timeout changes for legacy request drivers
Linus Torvalds [Sat, 23 Jun 2018 22:31:54 +0000 (06:31 +0800)]
Merge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Pull crypto fixes from Herbert Xu:
- Fix use after free in chtls
- Fix RBP breakage in sha3
- Fix use after free in hwrng_unregister
- Fix overread in morus640
- Move sleep out of kernel_neon in arm64/aes-blk
* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
hwrng: core - Always drop the RNG in hwrng_unregister()
crypto: morus640 - Fix out-of-bounds access
crypto: don't optimize keccakf()
crypto: arm64/aes-blk - fix and move skcipher_walk_done out of kernel_neon_begin, _end
crypto: chtls - use after free in chtls_pt_recvmsg()
Linus Torvalds [Sat, 23 Jun 2018 22:23:28 +0000 (06:23 +0800)]
Merge tag 'trace-v4.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace
Pull tracing fixes from Steven Rostedt:
"This contains a few fixes and a clean up.
- a bad merge caused an "endif" to go in the wrong place in
scripts/Makefile.build
- softirq tracing fix for tracing that corrupts lockdep and causes a
false splat
- histogram documentation typo fixes
- fix a bad memory reference when passing in no filter to the filter
code
- simplify code by using the swap macro instead of open coding the
swap"
* tag 'trace-v4.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
tracing: Fix SKIP_STACK_VALIDATION=1 build due to bad merge with -mrecord-mcount
tracing: Fix some errors in histogram documentation
tracing: Use swap macro in update_max_tr
softirq: Reorder trace_softirqs_on to prevent lockdep splat
tracing: Check for no filter when processing event filters
Bart Van Assche [Fri, 22 Jun 2018 20:18:09 +0000 (13:18 -0700)]
blk-mq: Fix timeout handling in case the timeout handler returns BLK_EH_DONE
Make sure that RQF_TIMED_OUT is cleared when a request is reused
after a block driver timeout handler has returned BLK_EH_DONE.
Fixes: da6612673988 ("blk-mq: don't time out requests again that are in the timeout handler") Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Jianchao Wang <jianchao.w.wang@oracle.com> Cc: Andrew Randrianasulu <randrianasulu@gmail.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Linus Torvalds [Sat, 23 Jun 2018 13:13:05 +0000 (21:13 +0800)]
Merge tag 'powerpc-4.18-2' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc fixes from Michael Ellerman:
- a fix for hugetlb with 4K pages, broken by our recent changes for
split PMD PTL.
- set the correct assembler machine type on e500mc, needed since
binutils 2.26 introduced two forms for the "wait" instruction.
- a fix for potential missed TLB flushes with MADV_[FREE|DONTNEED] etc.
and THP on Power9 Radix.
- three fixes to try and make our panic handling more robust by hard
disabling interrupts, and not marking stopped CPUs as offline because
they haven't been properly offlined.
- three other minor fixes.
Thanks to: Aneesh Kumar K.V, Michael Jeanson, Nicholas Piggin.
* tag 'powerpc-4.18-2' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
powerpc/mm/hash/4k: Free hugetlb page table caches correctly.
powerpc/64s/radix: Fix radix_kvm_prefetch_workaround paca access of not possible CPU
powerpc/64s: Fix build failures with CONFIG_NMI_IPI=n
powerpc/64: hard disable irqs on the panic()ing CPU
powerpc: smp_send_stop do not offline stopped CPUs
powerpc/64: hard disable irqs in panic_smp_self_stop
powerpc/64s: Fix DT CPU features Power9 DD2.1 logic
powerpc/64s/radix: Fix MADV_[FREE|DONTNEED] TLB flush miss problem with THP
powerpc/e500mc: Set assembler machine type to e500mc
Linus Torvalds [Sat, 23 Jun 2018 13:07:43 +0000 (21:07 +0800)]
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 fixes from Catalin Marinas:
- clear buffers allocated with FORCE_CONTIGUOUS explicitly until the
CMA code honours __GFP_ZERO
- notrace annotation for secondary_start_kernel()
- use early_param() instead of __setup() for "kpti=" as it is needed
for the cpufeature callback remapping swapper to non-global mappings
- ensure writes to swapper are ordered wrt subsequent cache maintenance
in the kpti non-global remapping code
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
arm64: mm: Ensure writes to swapper are ordered wrt subsequent cache maintenance
arm64: kpti: Use early_param for kpti= command-line option
arm64: make secondary_start_kernel() notrace
arm64: dma-mapping: clear buffers allocated with FORCE_CONTIGUOUS flag
Linus Torvalds [Sat, 23 Jun 2018 12:59:00 +0000 (20:59 +0800)]
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull KVM fixes from Radim Krčmář:
"ARM:
- Lazy FPSIMD switching fixes
- Really disable compat ioctls on architectures that don't want it
- Disable compat on arm64 (it was never implemented...)
- Rely on architectural requirements for GICV on GICv3
- Detect bad alignments in unmap_stage2_range
x86:
- Add nested VM entry checks to avoid broken error recovery path
- Minor documentation fix"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
KVM: fix KVM_CAP_HYPERV_TLBFLUSH paragraph number
kvm: vmx: Nested VM-entry prereqs for event inj.
KVM: arm64: Prevent KVM_COMPAT from being selected
KVM: Enforce error in ioctl for compat tasks when !KVM_COMPAT
KVM: arm/arm64: add WARN_ON if size is not PAGE_SIZE aligned in unmap_stage2_range
KVM: arm64: Avoid mistaken attempts to save SVE state for vcpus
KVM: arm64/sve: Fix SVE trap restoration for non-current tasks
KVM: arm64: Don't mask softirq with IRQs disabled in vcpu_put()
arm64: Introduce sysreg_clear_set()
KVM: arm/arm64: Drop resource size check for GICV window
Linus Torvalds [Sat, 23 Jun 2018 12:44:11 +0000 (20:44 +0800)]
Merge tag 'for-linus-4.18-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull xen fixes from Juergen Gross:
"This contains the following fixes/cleanups:
- the removal of a BUG_ON() which wasn't necessary and which could
trigger now due to a recent change
- a correction of a long standing bug happening very rarely in Xen
dom0 when a hypercall buffer from user land was not accessible by
the hypervisor for very short periods of time due to e.g. page
migration or compaction
- usage of EXPORT_SYMBOL_GPL() instead of EXPORT_SYMBOL() in a
Xen-related driver (no breakage possible as using those symbols
without others already exported via EXPORT-SYMBOL_GPL() wouldn't
make any sense)
- a simplification for Xen PVH or Xen ARM guests
- some additional error handling for callers of xenbus_printf()"
* tag 'for-linus-4.18-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen: Remove unnecessary BUG_ON from __unbind_from_irq()
xen: add new hypercall buffer mapping device
xen/scsiback: add error handling for xenbus_printf
scsi: xen-scsifront: add error handling for xenbus_printf
xen/grant-table: Export gnttab_{alloc|free}_pages as GPL
xen: add error handling for xenbus_printf
xen: share start flags between PV and PVH
__pgtable_l5_enabled marked as __initdata, but cpu_init() is not __init.
It's fixable: early code can be isolated into a separate translation unit,
but such change collides with other work in the area. That's too much
hassle to save 4 bytes of memory.
Return __pgtable_l5_enabled back to be __ro_after_init.
The current logic incorrectly calculates the LLC ID from the APIC ID.
Unless specified otherwise, the LLC ID should be calculated by removing
the Core and Thread ID bits from the least significant end of the APIC
ID. For more info, see "ApicId Enumeration Requirements" in any Fam17h
PPR document.
Jan Kara [Mon, 18 Jun 2018 13:46:58 +0000 (15:46 +0200)]
bdi: Fix another oops in wb_workfn()
syzbot is reporting NULL pointer dereference at wb_workfn() [1] due to
wb->bdi->dev being NULL. And Dmitry confirmed that wb->state was
WB_shutting_down after wb->bdi->dev became NULL. This indicates that
unregister_bdi() failed to call wb_shutdown() on one of wb objects.
The problem is in cgwb_bdi_unregister() which does cgwb_kill() and thus
drops bdi's reference to wb structures before going through the list of
wbs again and calling wb_shutdown() on each of them. This way the loop
iterating through all wbs can easily miss a wb if that wb has already
passed through cgwb_remove_from_bdi_list() called from wb_shutdown()
from cgwb_release_workfn() and as a result fully shutdown bdi although
wb_workfn() for this wb structure is still running. In fact there are
also other ways cgwb_bdi_unregister() can race with
cgwb_release_workfn() leading e.g. to use-after-free issues:
We solve these issues by synchronizing writeback structure shutdown from
cgwb_bdi_unregister() with cgwb_release_workfn() using a new mutex. That
way we also no longer need synchronization using WB_shutting_down as the
mutex provides it for CONFIG_CGROUP_WRITEBACK case and without
CONFIG_CGROUP_WRITEBACK wb_shutdown() can be called only once from
bdi_unregister().
Reported-by: syzbot <syzbot+4a7438e774b21ddd8eca@syzkaller.appspotmail.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@kernel.dk>
lightnvm: Remove depends on HAS_DMA in case of platform dependency
Remove dependencies on HAS_DMA where a Kconfig symbol depends on another
symbol that implies HAS_DMA, and, optionally, on "|| COMPILE_TEST".
In most cases this other symbol is an architecture or platform specific
symbol, or PCI.
Generic symbols and drivers without platform dependencies keep their
dependencies on HAS_DMA, to prevent compiling subsystems or drivers that
cannot work anyway.
This simplifies the dependencies, and allows to improve compile-testing.
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> Reviewed-by: Mark Brown <broonie@kernel.org> Acked-by: Robin Murphy <robin.murphy@arm.com> Reviewed-by: Matias Bjørling <mb@lightnvm.io> Signed-off-by: Jens Axboe <axboe@kernel.dk>
Will Deacon [Fri, 22 Jun 2018 10:45:07 +0000 (11:45 +0100)]
rseq: Avoid infinite recursion when delivering SIGSEGV
When delivering a signal to a task that is using rseq, we call into
__rseq_handle_notify_resume() so that the registers pushed in the
sigframe are updated to reflect the state of the restartable sequence
(for example, ensuring that the signal returns to the abort handler if
necessary).
However, if the rseq management fails due to an unrecoverable fault when
accessing userspace or certain combinations of RSEQ_CS_* flags, then we
will attempt to deliver a SIGSEGV. This has the potential for infinite
recursion if the rseq code continuously fails on signal delivery.
Avoid this problem by using force_sigsegv() instead of force_sig(), which
is explicitly designed to reset the SEGV handler to SIG_DFL in the case
of a recursive fault. In doing so, remove rseq_signal_deliver() from the
internal rseq API and have an optional struct ksignal * parameter to
rseq_handle_notify_resume() instead.
Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: peterz@infradead.org Cc: paulmck@linux.vnet.ibm.com Cc: boqun.feng@gmail.com Link: https://lkml.kernel.org/r/1529664307-983-1-git-send-email-will.deacon@arm.com
Will Deacon [Fri, 22 Jun 2018 15:23:45 +0000 (16:23 +0100)]
arm64: mm: Ensure writes to swapper are ordered wrt subsequent cache maintenance
When rewriting swapper using nG mappings, we must performance cache
maintenance around each page table access in order to avoid coherency
problems with the host's cacheable alias under KVM. To ensure correct
ordering of the maintenance with respect to Device memory accesses made
with the Stage-1 MMU disabled, DMBs need to be added between the
maintenance and the corresponding memory access.
This patch adds a missing DMB between writing a new page table entry and
performing a clean+invalidate on the same line.
Fixes: f992b4dfd58b ("arm64: kpti: Add ->enable callback to remap swapper using nG mappings") Cc: <stable@vger.kernel.org> # 4.16.x- Acked-by: Mark Rutland <mark.rutland@arm.com> Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>