The interrupt handler uses a magic number to check that the doorbell
was rung. Better replace this number with official Broadcom define.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1621105859-30215-9-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
By converting the first and the third if statement into continue early
the function poll_services_of_group() can avoid 2 indention levels.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1621105859-30215-8-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The function poll_services() has too many indention levels. So keep only
the group iteration loop and move the rest into a new function
poll_services_of_group().
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1621105859-30215-7-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Recent commit "staging: vchiq_core: drop vchiq_status from vchiq_init_state"
missed to change the return type in the definition. Let's fix this now.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1621105859-30215-2-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replace the custom set of return values with proper Linux error codes for
vchiq_init_state().
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-12-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replace the custom set of return values with proper Linux error codes for
vchiq_initialise().
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-11-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replace the custom set of return values with proper Linux error codes for
vchiq_set_service_option(). Now we can use the result directly as return
value for vchiq_ioctl().
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-10-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replace the custom set of return values with proper Linux error codes
for the following functions:
vchiq_use_internal()
vchiq_release_internal()
vchiq_use_service_internal()
vchiq_release_service_internal()
Now we can use the result directly as return value for vchiq_ioctl().
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-9-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replace the custom set of return values with proper Linux error codes.
As a result we need to initialize vchiq_status in vchiq_init_state.
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-8-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The function vchiq_arm_init_state() cannot fail. So drop the return
value and the unnecessary code.
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-7-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is no modification to the vchiq_status before the first if, so drop
this unnecessary check.
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-6-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Nesting multiple ternary operators over multiple lines isn't easy to
read. Move this logic into a separate inline function.
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-5-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Avoid the multiline statement for notify_bulks by return early in
error case.
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-4-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The logic in VCHIQ_SRVSTATE_OPEN* is unnecessary complex. Handle the error
case of queue_message() first makes it easier to read.
Btw we get the rid of the checkpatch warning:
WARNING: else is not generally useful after a break or return
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-3-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Using BUG_ON in a non-essential driver isn't recommend. So better
trigger a stacktrace and bailout.
Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1619347863-16080-2-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Updates for SoC specific drivers include a few subsystems that
have their own maintainers but send them through the soc tree:
TEE/OP-TEE:
- Add tracepoints around calls to secure world
Memory controller drivers:
- Minor fixes for Renesas, Exynos, Mediatek and Tegra platforms
- Add debug statistics to Tegra20 memory controller
- Update Tegra bindings and convert to dtschema
ARM SCMI Firmware:
- Support for modular SCMI protocols and vendor specific extensions
- New SCMI IIO driver
- Per-cpu DVFS
The other driver changes are all from the platform maintainers
directly and reflect the drivers that don't fit into any other
subsystem as well as treewide changes for a particular platform.
SoCFPGA:
- Various cleanups contributed by Krzysztof Kozlowski
Mediatek:
- add MT8183 support to mutex driver
- MMSYS: use per SoC array to describe the possible routing
- add MMSYS support for MT8183 and MT8167
- add support for PMIC wrapper with integrated arbiter
- add support for MT8192/MT6873
Tegra:
- Bug fixes to PMC and clock drivers
NXP/i.MX:
- Update SCU power domain driver to keep console domain power on.
- Add missing ADC1 power domain to SCU power domain driver.
- Update comments for single global power domain in SCU power domain
driver.
- Add i.MX51/i.MX53 unique id support to i.MX SoC driver.
NXP/FSL SoC driver updates for v5.13
- Add ACPI support for RCPM driver
- Use generic io{read,write} for QE drivers after performance optimized
for PowerPC
- Fix QBMAN probe to cleanup HW states correctly for kexec
- Various cleanup and style fix for QBMAN/QE/GUTS drivers
OMAP:
- Preparation to use devicetree for genpd
- ti-sysc needs iorange check improved when the interconnect target module
has no control registers listed
- ti-sysc needs to probe l4_wkup and l4_cfg interconnects first to avoid
issues with missing resources and unnecessary deferred probe
- ti-sysc debug option can now detect more devices
- ti-sysc now warns if an old incomplete devicetree data is found as we
now rely on it being complete for am3 and 4
- soc init code needs to check for prcm and prm nodes for omap4/5 and dra7
- omap-prm driver needs to enable autoidle retention support for omap4
- omap5 clocks are missing gpmc and ocmc clock registers
- pci-dra7xx now needs to use builtin_platform_driver instead of using
builtin_platform_driver_probe for deferred probe to work
Raspberry Pi:
- Fix-up all RPi firmware drivers so as for unbind to happen in an
orderly fashion
- Support for RPi's PoE hat PWM bus
Qualcomm
- Improved detection for SCM calling conventions
- Support for OEM specific wifi firmware path
- Added drivers for SC7280/SM8350: RPMH, LLCC< AOSS QMP
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
-----BEGIN PGP SIGNATURE-----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=iidq
-----END PGP SIGNATURE-----
Merge tag 'arm-drivers-5.13' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Pull ARM SoC driver updates from Arnd Bergmann:
"Updates for SoC specific drivers include a few subsystems that have
their own maintainers but send them through the soc tree:
TEE/OP-TEE:
- Add tracepoints around calls to secure world
Memory controller drivers:
- Minor fixes for Renesas, Exynos, Mediatek and Tegra platforms
- Add debug statistics to Tegra20 memory controller
- Update Tegra bindings and convert to dtschema
ARM SCMI Firmware:
- Support for modular SCMI protocols and vendor specific extensions
- New SCMI IIO driver
- Per-cpu DVFS
The other driver changes are all from the platform maintainers
directly and reflect the drivers that don't fit into any other
subsystem as well as treewide changes for a particular platform.
SoCFPGA:
- Various cleanups contributed by Krzysztof Kozlowski
Mediatek:
- add MT8183 support to mutex driver
- MMSYS: use per SoC array to describe the possible routing
- add MMSYS support for MT8183 and MT8167
- add support for PMIC wrapper with integrated arbiter
- add support for MT8192/MT6873
Tegra:
- Bug fixes to PMC and clock drivers
NXP/i.MX:
- Update SCU power domain driver to keep console domain power on.
- Add missing ADC1 power domain to SCU power domain driver.
- Update comments for single global power domain in SCU power domain
driver.
- Add i.MX51/i.MX53 unique id support to i.MX SoC driver.
NXP/FSL SoC driver updates for v5.13
- Add ACPI support for RCPM driver
- Use generic io{read,write} for QE drivers after performance
optimized for PowerPC
- Fix QBMAN probe to cleanup HW states correctly for kexec
- Various cleanup and style fix for QBMAN/QE/GUTS drivers
OMAP:
- Preparation to use devicetree for genpd
- ti-sysc needs iorange check improved when the interconnect target
module has no control registers listed
- ti-sysc needs to probe l4_wkup and l4_cfg interconnects first to
avoid issues with missing resources and unnecessary deferred probe
- ti-sysc debug option can now detect more devices
- ti-sysc now warns if an old incomplete devicetree data is found as
we now rely on it being complete for am3 and 4
- soc init code needs to check for prcm and prm nodes for omap4/5 and
dra7
- omap-prm driver needs to enable autoidle retention support for
omap4
- omap5 clocks are missing gpmc and ocmc clock registers
- pci-dra7xx now needs to use builtin_platform_driver instead of
using builtin_platform_driver_probe for deferred probe to work
Raspberry Pi:
- Fix-up all RPi firmware drivers so as for unbind to happen in an
orderly fashion
- Support for RPi's PoE hat PWM bus
Qualcomm
- Improved detection for SCM calling conventions
- Support for OEM specific wifi firmware path
- Added drivers for SC7280/SM8350: RPMH, LLCC< AOSS QMP"
* tag 'arm-drivers-5.13' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (165 commits)
soc: aspeed: fix a ternary sign expansion bug
memory: mtk-smi: Add device-link between smi-larb and smi-common
memory: samsung: exynos5422-dmc: handle clk_set_parent() failure
memory: renesas-rpc-if: fix possible NULL pointer dereference of resource
clk: socfpga: fix iomem pointer cast on 64-bit
soc: aspeed: Adapt to new LPC device tree layout
pinctrl: aspeed-g5: Adapt to new LPC device tree layout
ipmi: kcs: aspeed: Adapt to new LPC DTS layout
ARM: dts: Remove LPC BMC and Host partitions
dt-bindings: aspeed-lpc: Remove LPC partitioning
soc: fsl: enable acpi support in RCPM driver
soc: qcom: mdt_loader: Detect truncated read of segments
soc: qcom: mdt_loader: Validate that p_filesz < p_memsz
soc: qcom: pdr: Fix error return code in pdr_register_listener
firmware: qcom_scm: Fix kernel-doc function names to match
firmware: qcom_scm: Suppress sysfs bind attributes
firmware: qcom_scm: Workaround lack of "is available" call on SC7180
firmware: qcom_scm: Reduce locking section for __get_convention()
firmware: qcom_scm: Make __qcom_scm_is_call_available() return bool
Revert "soc: fsl: qe: introduce qe_io{read,write}* wrappers"
...
Storing the index for poll services in a local var will increase the
readability of the second loop. Also we get the rid off the checkpatch
issue about the line ending with a square bracket.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1618164700-21150-7-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In case there is no service pointer provided, we can skip these polling.
Use a goto to reduce the indentation, which is necessary for the following
patches. Btw fix the brace alignment of the loops.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1618164700-21150-6-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Exit early allow us to reduce the indention in vchiq_open_service_internal()
and vchiq_set_service_option(). Btw we can avoid the multi-line assignments
of quota.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1618164700-21150-5-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Additional curly braces around cases are a bit harder to read. So
change the scope of service quota to get the rid off those braces.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Link: https://lore.kernel.org/r/1618164700-21150-4-git-send-email-stefan.wahren@i2se.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use devm_rpi_firmware_get() so as to make sure we release RPi's firmware
interface when unbinding the device.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When arm_state is NULL, no error return code of vchiq_release_internal()
and vchiq_use_internal() is assigned.
To fix this bug, ret is assigned with VCHIQ_ERROR.
Reported-by: TOTE Robot <oslab@tsinghua.edu.cn>
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Link: https://lore.kernel.org/r/20210306132245.16811-1-baijiaju1990@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix checkpatch.pl warning regarding space indents and replace it with
tabs to comply with kernel coding style.
Signed-off-by: Bastien Maureille <bastien.maureille@gmail.com>
Link: https://lore.kernel.org/r/20210222181812.268909-1-bastien.maureille@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Smatch found a local variable that can get copied to another
local variable without an initializion in the error case:
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c:1056 vchiq_get_user_ptr() error: uninitialized symbol 'ptr'.
This seems harmless, as the function should normally get inlined, with
the output directly written or not. In any case, the uninitialized data
is never used after get_user() fails.
As Dan mentions, it could still trigger an UBSAN runtime error, and it
is of course a bad idea to copy uninitialized variables, so just
bail out early.
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20210105135256.1810337-1-arnd@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Record in the TODO file that the address of "&waiter->bulk_waiter"
should never be returned to userspace.
Acked-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Link: https://lore.kernel.org/r/20210105162030.1415213-4-phil@raspberrypi.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The recent change to the bulk transfer compat function missed the fact
the relevant ioctl command is VCHIQ_IOC_QUEUE_BULK_TRANSMIT32, not
VCHIQ_IOC_QUEUE_BULK_TRANSMIT, as any attempt to send a bulk block
to the VPU would have shown.
Fixes: a4367cd2b2 ("staging: vchiq: convert compat bulk transfer")
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Link: https://lore.kernel.org/r/20210105162030.1415213-3-phil@raspberrypi.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The addition of the local 'userdata' pointer to
vchiq_irq_queue_bulk_tx_rx omitted the case where neither BLOCKING nor
WAITING modes are used, in which case the value provided by the
caller is not returned to them as expected, but instead it is replaced
with a NULL. This lack of a suitable context may cause the application
to crash or otherwise malfunction.
Fixes: 4184da4f31 ("staging: vchiq: fix __user annotations")
Tested-by: Stefan Wahren <stefan.wahren@i2se.com>
Acked-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Link: https://lore.kernel.org/r/20210105162030.1415213-2-phil@raspberrypi.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This comment describes a security problem which was fixed in commit
1c954540c0 ("staging: vchiq: avoid mixing kernel and user pointers").
The bug is fixed now so the FIXME can be removed.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/X/RnUjY3XkZohk7w@mwanda
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mutex lock can be initialized automatically with DEFINE_MUTEX()
rather than explicitly calling mutex_init().
Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Link: https://lore.kernel.org/r/20201224132528.31558-1-zhengyongjun3@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
After a list_for_each_entry() loop, the list iterator is always non-NULL
so these conditions don't work. If the "waiter" is not found then this
results in an out of bounds access.
I have fixed it by introducing a new "found" variable. In one case, I
used an else statement for readability.
Fixes: 46e4b9ec4f ("staging: vchiq_arm: use list_for_each_entry when accessing bulk_waiter_list")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/20201006134748.GA2076872@mwanda
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Smatch complains that "userdata" can be passed to vchiq_bulk_transfer()
without being initialized. This leads to a potential information leak
later on.
Fixes: a4367cd2b2 ("staging: vchiq: convert compat bulk transfer")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200930123036.GC4282@kadam
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As found earlier, there is a problem in the create_pagelist() function
that takes a pointer argument that either points into vmalloc space or
into user space, with the pointer value controlled by user space allowing
a malicious user to trick the driver into accessing the kernel instead.
Avoid this problem by adding another function argument and passing
kernel pointers separately from user pointers. This makes it possible
to rely on sparse to point out invalid conversions, and it prevents
user space from faking a kernel pointer.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200925114424.2647144-2-arnd@arndb.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
My earlier patches caused some new sparse warnings, but it turns out
that a number of those are actual bugs, or at least suspicous code.
Adding __user annotations to the data structures that are defined in
uapi headers helps avoid the new warnings, but that causes a different
set of warnings to show up, as some of these structures are used both
inside of the kernel and at the user interface but storing pointers to
different things there.
Duplicating the vchiq_service_params and vchiq_completion_data structures
in turn takes care of most of those, and then it turns out that there
is a 'data' pointer that can be any of a __user address, a dmd_addr_t
and a kernel pointer in vmalloc space at times.
I'm trying to annotate these as best I can without changing behavior,
but there still seems to be a serious bug when user space passes
a valid vmalloc space address instead of a user pointer. Adding
comments in the code there, and leaving the warnings in place that
seem to correspond to actual bugs.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200925114424.2647144-1-arnd@arndb.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Split out the ioctl implementation for VCHIQ_IOC_QUEUE_BULK_TRANSMIT
into a separate function so it can be shared with the compat
implementation.
This one is the trickiest conversion, as the compat implementation
is already quite different from the native one. By using a common
handler, the behavior is changed to be the same again: The
indirect __user pointer accesses are now handled through helper
functions that check for compat mode internally.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200918095441.1446041-6-arnd@arndb.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Split out the ioctl implementation for VCHIQ_IOC_QUEUE_BULK_TRANSMIT
into a separate function so it can be shared with the compat
implementation.
Here, the input data is converted separately in the compat
handler, while the output data is passed as a __user pointer
to thec vchiq_queue_bulk_transfer->mode word that is
compatible.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200918095441.1446041-5-arnd@arndb.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Split out the ioctl implementation for VCHIQ_IOC_DEQUEUE_MESSAGE
into a separate function so it can be shared with the compat
implementation.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200918095441.1446041-4-arnd@arndb.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Split out the ioctl implementation for VCHIQ_IOC_CREATE_SERVICE
into a separate function so it can be shared with the compat
implementation.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200918095441.1446041-3-arnd@arndb.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The compat handlers for VCHIQ_IOC_QUEUE_MESSAGE32 and
VCHIQ_IOC_GET_CONFIG32 can simply call the underlying implementations
that are already separate functions rather than using copy_in_user to
simulate the native 64-bit interface for the full ioctl handler.
vchiq_ioc_queue_message gets a small update to the calling
conventions to simplify the compat version by directly
returning a normal errno value.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200918095441.1446041-2-arnd@arndb.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes the checkpatch.pl warning:
WARNING: Prefer 'unsigned int' to bare use of 'unsigned'
Signed-off-by: Baidyanath Kundu <kundubaidya99@gmail.com>
Link: https://lore.kernel.org/r/20200713143653.32332-1-kundubaidya99@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To make the separation clear between vchiq's header files and vchiq.h,
which is to be used by services and is the 'public' API, move it into a
dedicated includes directory.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-48-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Those are only used in the core vchiq code, while present in vchiq's
'public' API header. Move them into the right place.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-47-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This used to be vchiq_if.h but vchiq.h is more concise for an include
file that will hopefully be in the future in the includes directory.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-46-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The barrier is only used by core code. So keep the barrier definition in
the core header.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-45-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
All the functions that vchi currently provides are a 1:1 mapping to its
vchiq counterparts. Get rid of vchi altogether and use vchiq's on all
services.
In the process also get rid of the vchi directory, as the only remaining
file was a TODO file, which now lives in the parent directory.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-44-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We can't really merge it with vchiq_queue_message() as it has internal
users that will not benefit from the retry mechanism
vchiq_queue_kernel_message() uses. So, for the sake of getting rid of
vchi, move it into vchiq.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-43-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Its vchiq counterpart, vchiq_bulk_transmit() is only used by vchi. We
can then merge both functions by moving vchi_bulk_queue_transmit()'s
retry mechanism into vchiq_bulk_transmit() and let services call the
later.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-42-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Its vchiq counterpart, vchiq_bulk_receive() is only used by vchi. We can
then merge both functions by moving vchi_bulk_queue_receive()'s retry
mechanism into vchiq_bulk_receive() and let services call the later.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-41-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vchiq_if.h is used by service consumers and only the exported API should
be present. Move the rest of definitions, which were not exported by the
way, to vchiq_core.h where they belong.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-40-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The function is being exported although there is no use for it outside
of vchiq's core code. Keep it local then.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-39-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is already enough duplication with regard to fourcc generation in
the kernel. Let's at least use the same macro all over the vchiq
ecosystem.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-35-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The services have access to struct vchiq_header's internals, so we can
let them get the data pointer. This pretty much makes both functions
exactly the same, which will allow us to make a switch to
vchiq_msg_hold() further down the road.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-34-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vchi created an opaque structure to be held by services while they
process callback messages. The contents of this opaque structure are the
service handle, which all services already maintain, and a pointer to
vchiq's message structure, struct vchiq_header.
Let's get rid of this opaque structure and directly pass struct
vchiq_header, which is what ultimately vchiq consumes when handling
callback messages.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-33-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The structure only contains a single parameter, which is the underlying
vchiq handle. Get rid of the struct and directly pass the handle around.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-32-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For initialization, vchi has its own params structure, which is then
translated to vchiq's params structure. They are essentially the same,
so lets directly use vchiq's.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-31-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As vchiq_shim's callback does nothing aside from pushing messages into
the service's queue, let's bypass it and jump directly to the service's
callbacks, letting them choose whether to use the message queue.
It turns out most services don't need to use the message queue, which
makes for simpler code in the end.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-29-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vchiq consumer drivers may need to use this function in order to get the
benefits of vchiq's per service message queueing mechanism.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-28-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The header file only provides other includes. Move the relevant includes
to their respective C files and delete it for good.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-25-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This has historically been handled by vchi, but there is no reason why
this couldn't be handled directly in vchiq.
The patch tries to avoid altering any behavior, with the exception of
the msg_queue size, which is now fixed to VCHIQ_MAX_SLOTS (it was set to
VCHIQ_MAX_SLOTS / 2). This is done to match vchiq's user_service message
queue, which could be merged with this one in the future.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-24-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When adding a new service validate the configuration parameters
provided, and remove unnecessary checks in vchi, now that we have
validated service's config.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-23-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It was probably there to trick compilers into ignoring unused variables,
which isn't needed in Linux.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-22-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
enum vchi_callback_reason maps 1:1 to enum vchiq_reason, in an effort to
simplify things, let's use the later, and get rid of the extra
indirection.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-21-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vchi has a set of transfer flags which almost map 1:1 to vchiq's own
transfer modes. For the sake of simplicity let's use the later and
delete vchi's.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-20-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
All users are ignoring the flags argument. So for the sake of
simplicity delete it.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-19-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The idea behind struct vchi_service_handle is to create an opaque handle
to struct shim_service. This can be achieved by doing a forward
declaration of struct shim_service, which will avoid unwarranted casts
and pointer play.
Ultimately as a rename is due all over the vchi user space, rename
struct shim_service into struvt vchi_service, which is more consistent
with the rest of the exposed API.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-13-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The idea behind this was to create an opaque handle to struct
vchiq_instance. This can be achieved without creating a new type by
means of a forward declaration of struct vchiq_instance.
This saves us from a lot of useless casting and overall simplifies code.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-12-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There already is a function that covers most of the functionality
vchi_msg_peek() provides: vchi_msg_hold(). The main difference being
that the later removes the message from vchu's queue while the other
does it later on, while releasing the message.
There are no users of this function that can't be trivially converted to
vchi_msg_hold(). So, for the sake of removing duplicate code, get rid of
vchi_msg_peek().
Note that the opposite change could be performed as well. But
vchi_msg_peek()'s implementation was deemed less robust as messages have
to be released in order.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-11-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
All the registered children of vchiq have a corresponding call to
platform_device_unregister except bcm2835_audio. Fix that.
Fixes: 25c7597af2 ("staging: vchiq_arm: Register a platform device for audio")
Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-9-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cleans up some bitshifts by using the BIT macro.
Signed-off-by: Garrit Franke <garritfranke@gmail.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/20200625093237.2640-1-garritfranke@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In 2019, we introduced pin_user_pages*() and now we are converting
get_user_pages*() to the new API as appropriate. [1] & [2] could
be referred for more information.
[1] Documentation/core-api/pin_user_pages.rst
[2] "Explicit pinning of user-space pages":
https://lwn.net/Articles/807108/
Signed-off-by: Souptick Joarder <jrdr.linux@gmail.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Link: https://lore.kernel.org/r/1591124057-27696-1-git-send-email-jrdr.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For messages with a reason different from VCHIQ_MESSAGE_AVAILABLE the
responsibility for releasing them is kept in vchi, in other words,
services don't need to worry about it. As we're trying to unify vchi and
vchiq, move the release code into vchiq.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200527115400.31391-10-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is an include only used by the Linux kernel, so no need to worry
about C++ compatibility.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200527115400.31391-9-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vchiq functions return an enum and vchi's ints. Those are compatible, no
need to explicitly cast them.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200527115400.31391-7-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There are no gains from that extra indirection level. Also, get rid of
the function description, the whole file will disappear soon.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200527115400.31391-5-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
All vchi users use the kernel variant of the copy callback. The only
user for the user space variant of the copy callback is in the ioctl
implementation. So move all this copying logic into vchiq, and expose a
new function that explicitly passes kernel messages.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200527115400.31391-4-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In several cases where a pointer marked as __user is
(intentionally) assigned or passed to a non-marked target,
cast to the target pointer type with a __force directive
to quiet warnings from sparse.
Signed-off-by: Mitchell Tasman <tasman@leaflabs.com>
Link: https://lore.kernel.org/r/20200519004531.33158-1-tasman@leaflabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Coding style issue reported by checkpatch.pl
This patch clears the checkpatch.pl "Block comments should align
the * on each line" warning.
Also cleared /****** and blank line.
Signed-off-by: John Oldman <john.oldman@polehill.co.uk>
Link: https://lore.kernel.org/r/20200513125405.28242-1-john.oldman@polehill.co.uk
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix the following gcc warning:
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c:2356:16:
warning: variable ‘local_entity_uc’ set but not used
[-Wunused-but-set-variable]
int local_uc, local_entity_uc;
^~~~~~~~~~~~~~~
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Link: https://lore.kernel.org/r/20200409085339.47255-1-yanaijie@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Correct long lines of comments to respect 80 character per line
limit and make them C89 style one.
Found using checkpatch.pl.
Signed-off-by: R Veera Kumar <vkor@vkten.in>
Link: https://lore.kernel.org/r/20200320024337.4754-1-vkor@vkten.in
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If kref_get_unless_zero() fails, we should keep looking for the
next service, since the callers of this function expect that a NULL
return value means there are no more.
Signed-off-by: Marcelo Diop-Gonzalez <marcgonzalez@google.com>
Link: https://lore.kernel.org/r/20200213194001.130110-1-marcgonzalez@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There are a few places where a service's reference count is incremented,
something quick is done, and the refcount is dropped. This can be made
a little simpler/faster by not grabbing a reference in these cases.
Signed-off-by: Marcelo Diop-Gonzalez <marcgonzalez@google.com>
Link: https://lore.kernel.org/r/ac6186ac888f1acf489b5b504efcba8b0d6a8b25.1581532523.git.marcgonzalez@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently reference counts are implemented by locking service_spinlock
and then incrementing the service's ->ref_count field, calling
kfree() when the last reference has been dropped. But at the same
time, there's code in multiple places that dereferences pointers
to services without having a reference, so there could be a race there.
It should be possible to avoid taking any lock in unlock_service()
or service_release() because we are setting a single array element
to NULL, and on service creation, a mutex is locked before looking
for a NULL spot to put the new service in.
Using a struct kref and RCU-delaying the freeing of services fixes
this race condition while still making it possible to skip
grabbing a reference in many places. Also it avoids the need to
acquire a single spinlock when e.g. taking a reference on
state->services[i] when somebody else is in the middle of taking
a reference on state->services[j].
Signed-off-by: Marcelo Diop-Gonzalez <marcgonzalez@google.com>
Link: https://lore.kernel.org/r/3bf6f1ec6ace64d7072025505e165b8dd18b25ca.1581532523.git.marcgonzalez@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The only remaining state doesn't provide any information anymore nor the
completion that depends on it, which is set to complete_all() before any
user may wait on it.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200131103836.14312-18-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
After the first states cleanup pass some set_suspend/resume_state()
calls disappeared which opened up even more opportunities for deletions.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200131103836.14312-17-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vchiq_platform_suspend() and vchiq_platform_resume() do nothing, get rid
of them.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200131103836.14312-9-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The function always returns 0, delete the function and all code
conditional to it, namely the suspend timer.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200131103836.14312-7-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The boolean value is never set to true, hence remove it and all the code that
depends on it.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200131103836.14312-3-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix up some relative paths in #include "..." directives, and remove
the include path to drivers/staging/vc04_services.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Link: https://lore.kernel.org/r/20200110135615.11617-1-masahiroy@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This undoes the previous call to alloc_chrdev_region() on failure,
and is probably what was meant originally given the label name.
Signed-off-by: Marcelo Diop-Gonzalez <marcgonzalez@google.com>
Cc: stable <stable@vger.kernel.org>
Fixes: 187ac53e59 ("staging: vchiq_arm: rework probe and init functions")
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20191203153921.70540-1-marcgonzalez@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
These functions currently modify the struct dump_context passed
to them, and set context->actual to -EFAULT in case of error.
The issue is that this is never returned to the user (except
accidentally when things align so that that happens). So, have
these functions return 0 on success and the appropriate error
code otherwise, and return nonzero errors to the user.
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Marcelo Diop-Gonzalez <marcgonzalez@google.com>
Reviewed-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20191120202102.249121-5-marcgonzalez@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Doing this helps with readability, and makes
the logic easier to follow.
Signed-off-by: Marcelo Diop-Gonzalez <marcgonzalez@google.com>
Reviewed-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20191120202102.249121-4-marcgonzalez@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reducing the indentation level helps a bit with
the readability of this function. There's also a checkpatch
fix here, moving the first argument to kthread_create() onto
the same line, as well as a relocation of the statement
"char threadname[16];" to the top of the function to avoid
a declaration in the middle of code.
Signed-off-by: Marcelo Diop-Gonzalez <marcgonzalez@google.com>
Link: https://lore.kernel.org/r/20191114232801.71458-1-marcgonzalez@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It could be confusing to return 1 on success and 0 on error
given the style elswhere.
Signed-off-by: Marcelo Diop-Gonzalez <marcgonzalez@google.com>
Link: https://lore.kernel.org/r/20191105222756.25369-1-marcgonzalez@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replaces VCHIQ_INSTANCE_T typedef with struct vchiq_instance to match
kernel code style. Issue found by checkpatch.
Additionally, as part of the process renames "struct vchiq_instance_struct" to "struct vchiq_instance".
Signed-off-by: Jamal Shareef <jamal.k.shareef@gmail.com>
Link: https://lore.kernel.org/r/ed2b3076f93a920149716687b48e0c5e3ddf0569.1572994235.git.jamal.k.shareef@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>