The SLB miss handler uses r3 for the faulting address but r12 is
mostly able to be freed up to save r3 in. It just requires SRR1
be reloaded again on error.
It would be more conventional to use r12 for SRR1 (and use r11 to
save r3), but slb_allocate_realmode clobbers r11 and not r12.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The EXCEPTION_PROLOG_1 used by SLB miss already saves CTR when the
kernel is built with CONFIG_RELOCATABLE. So it does not have to be
saved and reloaded when branching to slb_miss_realmode. It can be
restored from the PACA as usual.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The EX_DAR save area is only used in exceptional cases. With r3 no
longer clobbered by slb_allocate_realmode, saving faulting address to
EX_DAR can be deferred to those cases.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This moves the PLL SETUP code from stv0367ter_init() into a dedicated
function, and also make it possible to configure 58Mhz IC speed at
27MHz Xtal (used on STV0367-based DDB cards/modules in QAM mode).
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Move the *ter and *cab st_register tables into a separate header file and
additionally organize them via a multidimensional array, allowing to add
more tables with differing init values, and also prepare for a base init
table which should contain general setup values. Also add a state var to
store the table triplet to be used.
Also fixes three minor style problems reported by checkpatch.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Change defaults table writing so tables can be of dynamic length without
having to keep track of their lengths by adding and evaluating an end
marker (reg 0x0000), also move table writing to a dedicated function to
remove code duplication. Additionally mark st_register tables const since
they're used read-only.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The CPAMP log lines generated in stv0367_ter_check_cpamp() are printed
everytime tuning succeeds or fails, quite cluttering the normal kernel log.
Use dprintk() instead of printk(KERN_ERR...) so that if the information is
needed, it'll be printed when the stv_debug modparam is enabled.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
One fewer registers clobbered by this function means the SLB miss
handler can save one fewer.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Some hardware and bridges (namely ddbridge) require that tuner access is
limited to one concurrent access and wrap i2c gate control with a
mutex_lock when attaching frontends. According to vendor information, this
is required as concurrent tuner reconfiguration can interfere each other
and at worst cause tuning fails or bad reception quality.
If the demod driver does gate_ctrl before setting up tuner parameters, and
the tuner does another I2C enable, it will deadlock forever when gate_ctrl
is wrapped into the mutex_lock. This adds a flag and a conditional before
triggering gate_ctrl in the demodulator driver.
Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
I would like to replace Kyungmin Park, who is no longer interested in
maintaining S5P-CEC driver. I have access to various Exynos boards. I also
already did some tests of this driver and helped enabling it on various
Exynos boards. The driver itself is no longer in staging, so lets fix the
path too and add DT bindings file pattern match. Also change the mailing
list from ARM generic to Samsung SoC specific to get more attention and
easier review in the future.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
There are multiple places where arrays or otherwise variable sized
buffer are allocated through V4L2 core code, including things like
controls, memory pages, staging buffers for ioctls and so on. Such
allocations can potentially require an order > 0 allocation from the
page allocator, which is not guaranteed to be fulfilled and is likely to
fail on a system with severe memory fragmentation (e.g. a system with
very long uptime).
Since the memory being allocated is intended to be used by the CPU
exclusively, we can consider using vmalloc() as a fallback and this is
exactly what the recently merged kvmalloc() helpers do. A kmalloc() call
is still attempted, even for order > 0 allocations, but it is done
with __GFP_NORETRY and __GFP_NOWARN, with expectation of failing if
requested memory is not available instantly. Only then the vmalloc()
fallback is used. This should give us fast and more reliable allocations
even on systems with higher memory pressure and/or more fragmentation,
while still retaining the same performance level on systems not
suffering from such conditions.
While at it, replace explicit array size calculations on changed
allocations with kvmalloc_array().
Purposedly not touching videobuf1, as it is deprecated, has only few
users remaining and would rather be seen removed instead.
Signed-off-by: Tomasz Figa <tfiga@chromium.org>
Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This is a wrapper around the media entity get_fwnode_pad operation.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The optional operation can be used by entities to report how it maps its
fwnode endpoints to media pad numbers. This is useful for devices which
require advanced mappings of pads.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Add a check for v4l2_dev to v4l2_async_notifier_register() as to fail as
early as possible since this will fail later in v4l2_async_test_notify().
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The new Renesas R-Car H3 ES2.0 platforms have a new hw version register.
Update the driver accordingly, defaulting to the new hw revision, and
differentiating the older revision as ES1
Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Remove useless variable assignment in function tc358743_isr().
The value stored in variable _intstatus_ at line 1299 is
overwritten at line 1302, just before it can be used.
Addresses-Coverity-ID: 1397678
Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Updating my personal email address in solo6x10.
Signed-off-by: Andrey Utkin <andrey.utkin@corp.bluecherry.net>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Anton Sviridenko is now in charge of drivers in Bluecherry.
Signed-off-by: Andrey Utkin <andrey.utkin@corp.bluecherry.net>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The davinci VPIF is a single hardware block, but the existing driver
is broken up into a common library (vpif.c), output (vpif_display.c) and
intput (vpif_capture.c).
When migrating to DT, to better model the hardware, and because
registers, interrupts, etc. are all common,it was decided to
have a single VPIF hardware node[1].
Because davinci uses legacy, non-DT boot on several SoCs still, the
platform_drivers need to remain. But they are also needed in DT boot.
Since there are no DT nodes for the display/capture parts in DT
boot (there is a single node for the parent/common device) we need to
create platform_devices somewhere to instantiate the platform_drivers.
When VPIF display/capture are needed for a DT boot, the VPIF node
will have endpoints defined for its subdevs. Therefore, vpif_probe()
checks for the presence of endpoints, and if detected manually creates
the platform_devices for the display and capture platform_drivers.
[1] Documentation/devicetree/bindings/media/ti,da850-vpif.txt
Signed-off-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
'call_ptr_memop' can return NULL, so we must test its return value with
'IS_ERR_OR_NULL'. Otherwise, the test 'if (mem_priv)' is meaningless.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
As long as only one CODA context is running we get alternating device_run()
and wait_for_completion() calls, but when more then one CODA context is
active, other VDOA slots can be inserted between those calls for one context.
Make sure to wait on job completion before running a different context and
before destroying the currently active context.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This implements a simple handler for the case where decode did not finish
sucessfully. This might be helpful during normal streaming, but for now it
only handles the case where the context would deadlock with userspace,
i.e. userspace issued DEC_CMD_STOP and waits for EOS, but after the failed
decode run we would hold the context and wait for userspace to queue more
buffers.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The mvcol buffer needs to be placed behind the chroma plane(s), so
use the real offset including any required rounding.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This makes firmware name and path part of venus_resources
structure and initialize it properly depending on the SoC and
firmware version.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This adds support for V4L2_CID_MIN_BUFFERS_FOR_CAPTURE get control
in venus decoder, it is usable in case when the userspace wants
to know minimum capture buffers before calling request_buf for
capture queue in mem2mem drivers. Also this will fix an issue
found gstreamer v4l2videodec element, i.e. the video decoder
element cannot continue because the buffers are insufficient.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This fixes the following compile error ocured when building
with gcc7:
drivers/media/platform/qcom/venus/venc.c:1150
venc_close() error: dereferencing freed memory 'inst'
by moving kfree as a last call.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This fixes the following compile error ocured when building
with gcc7:
drivers/media/platform/qcom/venus/vdec.c:1022
vdec_close() error: dereferencing freed memory 'inst'
by moving kfree as a last call.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This fixes a warning found when building with gcc7:
drivers/media/platform/qcom/venus/hfi_msgs.c:465:40:
warning: variable 'domain' set but not used [-Wunused-but-set-variable]
u32 rem_bytes, num_props, codecs = 0, domain = 0;
^~~~~~
drivers/media/platform/qcom/venus/hfi_msgs.c:465:28:
warning: variable 'codecs' set but not used [-Wunused-but-set-variable]
u32 rem_bytes, num_props, codecs = 0, domain = 0;
The warning is avoided by deleting the variables declaration.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This fixes a warning found when building with gcc7:
drivers/media/platform/qcom/venus/hfi_venus.c:998
venus_isr_thread() warn: variable dereferenced before check
'hdev' (see line 994)
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This fixes a warning found when building the driver with gcc7:
drivers/media/platform/qcom/venus/helpers.c:157
load_per_instance() warn: variable dereferenced before check
'inst' (see line 153)
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This fixes a warning found when building the driver with gcc7:
drivers/media/platform/qcom/venus/hfi_cmds.c:415
pkt_session_set_property_1x() warn: variable dereferenced before
check 'pkt' (see line 412)
drivers/media/platform/qcom/venus/hfi_cmds.c:1177
pkt_session_set_property_3xx() warn: variable dereferenced before
check 'pkt' (see line 1174)
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This fixed a warning when build driver with gcc7:
drivers/media/platform/qcom/venus/hfi.c:171
hfi_core_ping() warn: inconsistent returns 'mutex:&core->lock'.
Locked on: line 159
Unlocked on: line 171
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This adds Venus driver Makefile and changes v4l2 platform
Makefile/Kconfig in order to enable building of the driver.
Note that in this initial version the COMPILE_TEST-ing is not
supported because the drivers specific to ARM builds are still
in process of enabling the aforementioned compile testing.
Once that disadvantage is fixed the Venus driver compile testing
will be possible with follow-up changes.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Here is the implementation of Venus video accelerator low-level
functionality. It contanins code which setup the registers and
startup uthe processor, allocate and manipulates with the shared
memory used for sending commands and receiving messages.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This is the implementation of HFI. It is charged with the
responsibility to comunicate with the firmware through an
interface commands and messages.
- hfi.c has interface functions used by the core, decoder
and encoder parts to comunicate with the firmware. For example
there are functions for session and core initialisation.
- hfi_cmds has packetization operations which preparing
packets to be send from host to firmware.
- hfi_msgs takes care of messages sent from firmware to the
host.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This adds encoder part of the driver plus encoder controls.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This consists of video decoder implementation plus decoder
controls.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
* core.c has implemented the platform driver methods, file
operations and v4l2 registration.
* helpers.c has implemented common helper functions for:
- buffer management
- vb2_ops and functions for format propagation,
- functions for allocating and freeing buffers for
internal usage. The buffer parameters describing internal
buffers depends on current format, resolution and codec.
- functions for calculation of current load of the
hardware. Depending on the count of instances and
resolutions it selects the best clock rate for the video
core.
* firmware loader
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
this add functions for:
- remove buffers from src/dst queue by index
- remove exact buffer from src/dst queue
also extends m2m API to iterate over a list of src/dst buffers
in safely and non-safely manner.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
-----BEGIN PGP SIGNATURE-----
iQEbBAABAgAGBQJZR92EAAoJEHm+PkMAQRiGT1oH+KW2FLrRaYxtut+KyGA6l7tc
R/hFx1n9BibkjXeqD+y6/4SjRTe6/pT8Zkihv3/19eZ5algUWeQa0Hm+/455sl58
IdIXx/pzuCO3kqR3//fP9ZFD657GNDsuQ0fYnZESItFwiWQtO1TNfZD6KQjkqBdI
L3MVhDUVBZA2ZtPwC4ERei5/sToV9woykKoJ/A3+OkWjgX6w4SBimqgkSEFk4uE8
xS0pycyDZci93rJlECi1UueewdODTKSmhwdC80qvGEiYXzsC6UFtaF0Fj66XO1AL
UMjxqI/gkm5ZuCIjRsmPmJjgL5q1RT6UX/qtw9yn71XTmcgMiPW2/DF/v/OaTg==
=XbW2
-----END PGP SIGNATURE-----
Merge tag 'v4.12-rc6' into patchwork
Linux 4.12-rc6
* tag 'v4.12-rc6': (813 commits)
Linux 4.12-rc6
mm: larger stack guard gap, between vmas
virtio_balloon: disable VIOMMU support
mm: correct the comment when reclaimed pages exceed the scanned pages
userfaultfd: shmem: handle coredumping in handle_userfault()
mm: numa: avoid waiting on freed migrated pages
swap: cond_resched in swap_cgroup_prepare()
mm/memory-failure.c: use compound_head() flags for huge pages
perf unwind: Report module before querying isactivation in dwfl unwind
fs: pass on flags in compat_writev
objtool: Add fortify_panic as __noreturn function
powerpc/debug: Add missing warn flag to WARN_ON's non-builtin path
USB: gadgetfs, dummy-hcd, net2280: fix locking for callbacks
drm: mxsfb_crtc: Reset the eLCDIF controller
drm/mgag200: Fix to always set HiPri for G200e4 V2
i2c: ismt: fix wrong device address when unmap the data buffer
i2c: rcar: use correct length when unmapping DMA
powerpc/xive: Fix offset for store EOI MMIOs
drm/tegra: Correct idr_alloc() minimum id
drm/tegra: Fix lockup on a use of staging API
...
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This patch factors out the allocator for signal frame optional
records into a separate function, to ensure consistency and
facilitate later expansion.
No overrun checking is currently done, because the allocation is in
user memory and anyway the kernel never tries to allocate enough
space in the signal frame yet for an overrun to occur. This
behaviour will be refined in future patches.
The approach taken in this patch to allocation of the terminator
record is not very clean: this will also be replaced in subsequent
patches.
For future extension, a comment is added in sigcontext.h
documenting the current static allocations in __reserved[]. This
will be important for determining under what circumstances
userspace may or may not see an expanded signal frame.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Dave Martin <Dave.Martin@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
In preparation for expanding the signal frame, this patch refactors
the signal frame setup code in setup_sigframe() into two separate
passes.
The first pass, setup_sigframe_layout(), determines the size of the
signal frame and its internal layout, including the presence and
location of optional records. The resulting knowledge is used to
allocate and locate the user stack space required for the signal
frame and to determine which optional records to include.
The second pass, setup_sigframe(), is called once the stack frame
is allocated in order to populate it with the necessary context
information.
As a result of these changes, it becomes more natural to represent
locations in the signal frame by a base pointer and an offset,
since the absolute address of each location is not known during the
layout pass. To be more consistent with this logic,
parse_user_sigframe() is refactored to describe signal frame
locations in a similar way.
This change has no effect on the signal ABI, but will make it
easier to expand the signal frame in future patches.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Dave Martin <Dave.Martin@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Currently, rt_sigreturn does very limited checking on the
sigcontext coming from userspace.
Future additions to the sigcontext data will increase the potential
for surprises. Also, it is not clear whether the sigcontext
extension records are supposed to occur in a particular order.
To allow the parsing code to be extended more easily, this patch
factors out the sigcontext parsing into a separate function, and
adds extra checks to validate the well-formedness of the sigcontext
structure.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Dave Martin <Dave.Martin@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
In order to be able to increase the amount of the data currently
written to the __reserved[] array in the signal frame, it is
necessary to overwrite the locations currently occupied by the
{fp,lr} frame link record pushed at the top of the signal stack.
In order for this to work, this patch detaches the frame link
record from struct rt_sigframe and places it separately at the top
of the signal stack. This will allow subsequent patches to insert
data between it and __reserved[].
This change relies on the non-ABI status of the placement of the
frame record with respect to struct sigframe: this status is
undocumented, but the placement is not declared or described in the
user headers, and known unwinder implementations (libgcc,
libunwind, gdb) appear not to rely on it.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Dave Martin <Dave.Martin@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
To avoid issues with the /proc/kcore code getting confused about the
kernels block mappings in the VMALLOC region, enable the existing
facility that describes the [_text, _end) interval as a separate
KCORE_TEXT region, which supersedes the KCORE_VMALLOC region that
it intersects with on arm64.
Reported-by: Tan Xiaojun <tanxiaojun@huawei.com>
Tested-by: Tan Xiaojun <tanxiaojun@huawei.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Laura Abbott <labbott@redhat.com>
Reviewed-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Instead of passing each start address into is_vmalloc_or_module_addr()
to decide whether it falls into either the VMALLOC or the MODULES region,
we can simply check the type field of the current kcore_list entry, since
it will be set to KCORE_VMALLOC based on exactly the same conditions.
As a bonus, when reading the KCORE_TEXT region on architectures that have
one, this will avoid using vread() on the region if it happens to intersect
with a KCORE_VMALLOC region. This is due the fact that the KCORE_TEXT
region is the first one to be added to the kcore region list.
Reported-by: Tan Xiaojun <tanxiaojun@huawei.com>
Tested-by: Tan Xiaojun <tanxiaojun@huawei.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Laura Abbott <labbott@redhat.com>
Reviewed-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
As it turns out, arm64 deviates from other architectures in the way it
maps the VMALLOC region: on most (all?) other architectures, it resides
strictly above the kernel's direct mapping of DRAM, but on arm64, this
is the other way around. For instance, for a 48-bit VA configuration,
we have
modules : 0xffff000000000000 - 0xffff000008000000 ( 128 MB)
vmalloc : 0xffff000008000000 - 0xffff7dffbfff0000 (129022 GB)
...
vmemmap : 0xffff7e0000000000 - 0xffff800000000000 ( 2048 GB maximum)
0xffff7e0000000000 - 0xffff7e0003ff0000 ( 63 MB actual)
memory : 0xffff800000000000 - 0xffff8000ffc00000 ( 4092 MB)
This has mostly gone unnoticed until now, but it does appear that it
breaks an assumption in the kmem read/write code, which does something
like
if (p < (unsigned long) high_memory) {
... use straight copy_[to|from]_user() using p as virtual address ...
}
...
if (count > 0) {
... use vread/vwrite for accesses past high_memory ...
}
The first condition will inadvertently hold for the VMALLOC region if
VMALLOC_START < PAGE_OFFSET [which is the case on arm64], but the read
or write will subsequently fail the virt_addr_valid() check, resulting
in a -ENXIO return value.
Given how kmem seems to be living in borrowed time anyway, and given
the fact that nobody noticed that the read/write interface is broken
on arm64 in the first place, let's not bother trying to fix it, but
simply disable the /dev/kmem interface entirely for arm64.
Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>