Add support for flush VXLAN FDB entries by nexthop ID.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for flush VXLAN FDB entries by source VNI.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The merge commit 9271686937 ("Merge branch 'br-flush-filtering'")
added support for FDB flushing in bridge driver only, the VXLAN driver does
not support such flushing. Extend VXLAN driver to support FDB flushing.
In this commit, add support for flushing with state and flags, which are
the fields that supported in the bridge driver.
Note that bridge driver supports 'NTF_USE' flag, but there is no point to
support this flag for flushing as it is ignored when flags are stored.
'NTF_STICKY' is not relevant for VXLAN driver.
'NTF_ROUTER' is not supported in bridge driver for flush as it is not
relevant for bridge, add it for VXLAN.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, the function vxlan_flush() does not flush the default FDB entry
(an entry with all_zeros_mac and default VNI), as it is deleted at
vxlan_uninit(). When this function will be used for flushing FDB entries
from user space, it will have to flush also the default entry in case that
other parameters match (e.g., VNI, flags).
Extend 'struct vxlan_fdb_flush_desc' to include an indication whether
the default entry should be flushed or not. The default value (false)
indicates to flush it, adjust all the existing callers to set
'.ignore_default_entry' to true, so the current behavior will not be
changed.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The function vxlan_flush() gets a boolean called 'do_all' and in case
that it is false, it does not flush entries with state 'NUD_PERMANENT'
or 'NUD_NOARP'. The following patches will add support for FDB flush
with parameters from user space. Make the function more generic, so it
can be used later.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The merge commit 9271686937 ("Merge branch 'br-flush-filtering'")
added support for FDB flushing in bridge driver. The following patches
will extend VXLAN driver to support FDB flushing as well. The netlink
message for bulk delete is shared between the drivers. With the existing
implementation, there is no way to prevent user from flushing with
attributes that are not supported per driver. For example, when VNI will
be added, user will not get an error for flush FDB entries in bridge
with VNI, although this attribute is not relevant for bridge.
As preparation for support of FDB flush in VXLAN driver, move the policy
to be handled in bridge driver, later a new policy for VXLAN will be
added in VXLAN driver. Do not pass 'vid' as part of ndo_fdb_del_bulk(),
as this field is relevant only for bridge.
Signed-off-by: Amit Cohen <amcohen@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Similar to other drivers, we need to make sure that the clock is
disabled during suspend and re-enabled during resume.
Reported-by: Angus Clark <angus.clark@broadcom.com>
Signed-off-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
The suspend/resume functions currently utilize
clk_disable()/clk_enable() respectively which may be no-ops with certain
clock providers such as SCMI. Fix this to use clk_disable_unprepare()
and clk_prepare_enable() respectively as we should.
Fixes: 3a9f595702 ("pwm: Add Broadcom BCM7038 PWM controller support")
Signed-off-by: Florian Fainelli <florian.fainelli@broadcom.com>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
The DT of_device.h and of_platform.h date back to the separate
of_platform_bus_type before it as merged into the regular platform bus.
As part of that merge prepping Arm DT support 13 years ago, they
"temporarily" include each other. They also include platform_device.h
and of.h. As a result, there's a pretty much random mix of those include
files used throughout the tree. In order to detangle these headers and
replace the implicit includes with struct declarations, users need to
explicitly include the correct includes.
Signed-off-by: Rob Herring <robh@kernel.org>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Using devm_pwmchip_add() allows to drop pwmchip_remove() from the remove
function which makes this function empty. Then there is no user of
drvdata left and platform_set_drvdata() can be dropped, too.
Further simplify and improve error returning using dev_err_probe().
Link: https://lore.kernel.org/r/20230929161918.2410424-12-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
With devm_clk_get_prepared() the call to clk_unprepare() can be dropped
from the error path and the remove callback. With devm_pwmchip_add()
pwmchip_remove() can be dropped. Then the remove callback is empty and
can go away, too. With vt8500_pwm_remove() the last user of
platform_get_drvdata() is gone and so platform_set_drvdata() can be
dropped, too.
Also use dev_err_probe() for simplified (and improved) error reporting.
Link: https://lore.kernel.org/r/20230929161918.2410424-10-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Using devm_pwmchip_add() allows to drop pwmchip_remove() from the remove
function which makes this function empty. Then there is no user of
drvdata left and platform_set_drvdata() can be dropped, too.
Further simplify and improve error returning using dev_err_probe().
Link: https://lore.kernel.org/r/20230929161918.2410424-9-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
With devm_clk_get_prepared() the call to clk_unprepare() can be dropped
from the error path and the remove callback. With devm_pwmchip_add()
pwmchip_remove() can be dropped. Then the remove callback is empty and
can go away, too. With spear_pwm_remove() the last user of
platform_get_drvdata() is gone and so platform_set_drvdata() can be
dropped, too.
Also use dev_err_probe() for simplified (and improved) error reporting.
Link: https://lore.kernel.org/r/20230929161918.2410424-7-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
With devm_pwmchip_add() pwmchip_remove() can be dropped from the remove
callback. Then the remove callback is empty and can go away, too. With
mtk_disp_pwm_remove() the last user of platform_get_drvdata() is gone and
so platform_set_drvdata() can be dropped, too.
Also use dev_err_probe() for simplified (and improved) error reporting.
Link: https://lore.kernel.org/r/20230929161918.2410424-6-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
With devm_clk_get_enabled() the call to clk_disable_unprepare() can be
dropped from the error path and the remove callback. With
devm_pwmchip_add() pwmchip_remove() can be dropped. Then the remove
callback is empty and can go away, too.
Also use dev_err_probe() for simplified (and improved) error reporting.
Link: https://lore.kernel.org/r/20230929161918.2410424-5-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
With devm_clk_get_enabled() the call to clk_disable_unprepare() can be
dropped from the error path and the remove callback. With
devm_pwmchip_add() pwmchip_remove() can be dropped. Then the remove
callback is empty and can go away, too.
Also use dev_err_probe() for simplified (and improved) error reporting.
Link: https://lore.kernel.org/r/20230929161918.2410424-4-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
With devm_clk_get_enabled() the call to clk_disable_unprepare() can be
dropped from the error path and the remove callback. With
devm_pwmchip_add() pwmchip_remove() can be dropped. Then the remove
callback is empty and can go away, too. With bcm2835_pwm_remove() the only
user of platform_get_drvdata() is gone and so platform_set_drvdata() can
be dropped from .probe(), too.
Also use dev_err_probe() for simplified (and improved) error reporting.
Link: https://lore.kernel.org/r/20230929161918.2410424-3-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
With devm_clk_get_enabled() the call to clk_disable_unprepare() can be
dropped from the error path and the remove callback. With
devm_pwmchip_add() pwmchip_remove() can be dropped. Then the remove
callback is empty and can go away, too.
Also use dev_err_probe() for simplified (and improved) error reporting.
Link: https://lore.kernel.org/r/20230929161918.2410424-2-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Most low-level PWM drivers support duty_cycle == period, and so does the
sysfs API. Also polarity can be changed for enabled PWMs since commit
39100ceea7 ("pwm: Switch to the atomic API").
Reported-by: Jens Gehrlein <J.Gehrlein@eckelmann.de>
Link: https://lore.kernel.org/r/20230911154454.675057-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
If we are not in PWM mode, then the output is technically a 50%
output based on a single timer instead of the high-low based on
the two counters. Add a check for the PWM mode in dwc_pwm_get_state()
and if DWC_TIM_CTRL_PWM is not set, then return a 50% cycle.
This may only be an issue on initialisation, as the rest of the
code currently assumes we're always going to have the extended
PWM mode using two counters.
Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20230907161242.67190-4-ben.dooks@codethink.co.uk
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Add a configurable clock base rate for the pwm as when being built
for non-PCI the block may be sourced from an internal clock.
Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20230907161242.67190-3-ben.dooks@codethink.co.uk
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Moving towards adding non-pci support for the driver, move the pci
parts out of the core into their own module. This is partly due to
the module_driver() code only being allowed once in a module and also
to avoid a number of #ifdef if we build a single file in a system
without pci support.
Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20230907161242.67190-2-ben.dooks@codethink.co.uk
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
With devm_clk_get_enabled() the call to clk_disable_unprepare() can be
dropped from the error path and the remove callback. With
devm_pwmchip_add() pwmchip_remove() can be dropped. Then the remove
callback is empty and can go away, too.
Link: https://lore.kernel.org/r/20230718175545.3946935-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
The semantic of chip_data is a bit surprising as it's cleared when
pwm_put() is called. Also there is a big overlap with the standard driver
data.
All drivers were adapted to not make use of chip_data any more, so it can
go away.
Link: https://lore.kernel.org/r/20230705080650.2353391-9-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Instead of an allocation of a single u16 per channel, allocate them all in
a single chunk which greatly reduces memory fragmentation and also the
overhead to track the allocated memory. Also put the channel data in
driver data where it's cheaper to determine the address (no function call
involved, just a trivial pointer addition).
This also allows to get rid of the request and free callbacks.
The only cost is that the channel data is allocated early, and even for
unused channels.
Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org>
Link: https://lore.kernel.org/r/20230705080650.2353391-8-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Instead of using one allocation per capture channel, use a single one. Also
store it in driver data instead of chip data.
This has several advantages:
- driver data isn't cleared when pwm_put() is called
- Reduces memory fragmentation
Also register the pwm chip only after the per capture channel data is
initialized as the capture callback relies on this initialization and it
might be called even before pwmchip_add() returns.
It would be still better to have struct sti_pwm_compat_data and the
per-channel data struct sti_cpt_ddata in a single memory chunk, but that's
not easily possible because the number of capture channels isn't known yet
when the driver data struct is allocated.
Fixes: e926b12c61 ("pwm: Clear chip_data in pwm_put()")
Reported-by: George Stark <gnstark@sberdevices.ru>
Fixes: c97267ae83 ("pwm: sti: Add PWM capture callback")
Link: https://lore.kernel.org/r/20230705080650.2353391-7-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Instead of distributing the driver's bookkeeping over 5 (i.e.
TPU_CHANNEL_MAX + 1) separately allocated memory chunks, put all together
in struct tpu_device. This reduces the number of memory allocations and
so fragmentation and maybe even the number of cache misses. Also
&tpu->tpd[pwm->hwpwm] is cheaper to evaluate than pwm_get_chip_data(pwm)
as the former is just an addition in machine code while the latter involves
a function call.
Link: https://lore.kernel.org/r/20230705080650.2353391-6-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Instead of distributing the driver's bookkeeping over 3 (i.e.
LP3943_NUM_PWMS + 1) separately allocated memory chunks, put all together
in struct lp3943_pwm. This reduces the number of memory allocations and
so fragmentation and maybe even the number of cache misses. Also
&lp3943_pwm->pwm_map[pwm->hwpwm] is cheaper to evaluate than
pwm_get_chip_data(pwm) as the former is just an addition in machine code
while the latter involves a function call.
Link: https://lore.kernel.org/r/20230705080650.2353391-5-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Stop using chip_data which is about to go away. Instead track the
per-channel clk in struct jz4740_pwm_chip.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Paul Cercueil <paul@crapouillou.net>
Link: https://lore.kernel.org/r/20230705080650.2353391-4-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Instead of allocating extra data in .request() provide the needed memory
in struct samsung_pwm_chip. This reduces the number of allocations. Even
though now all 5 channel structs are allocated this is probably
outweighed by the reduced overhead to track up to 6 smaller allocations.
Link: https://lore.kernel.org/r/20230705080650.2353391-3-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Instead of allocating extra data in .request() provide the needed memory
in struct berlin_pwm_chip. This reduces the number of allocations. A
side effect is that on suspend and resume the state for all four
channels is always saved and restored. This is easier (and probably
quicker) than looking up the matching pwm_device and checking its
PWMF_REQUESTED bit.
Link: https://lore.kernel.org/r/20230705080650.2353391-2-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
The driver compiles just fine as a module. The parent driver's Kconfig
symbol already depends on X86 || COMPILE_TEST, so X86 can just be
dropped from the dependencies allowing compilation on other platforms
than x86.
Link: https://lore.kernel.org/r/20230804142707.412137-3-u.kleine-koenig@pengutronix.de
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Instead of requiring each driver to care for assigning the owner member
of struct pwm_ops, handle that implicitly using a macro. Note that the
owner member has to be moved to struct pwm_chip, as the ops structure
usually lives in read-only memory and so cannot be modified.
The upside is that new low level drivers cannot forget the assignment and
save one line each. The pwm-crc driver didn't assign .owner, that's not
a problem in practice though as the driver cannot be compiled as a
module.
Acked-by: Andy Shevchenko <andy.shevchenko@gmail.com> # Intel LPSS
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com> # pwm-{bcm,brcm}*.c
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com> # sun4i
Acked-by: Andi Shyti <andi.shyti@kernel.org>
Acked-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp> # pwm-visconti
Acked-by: Heiko Stuebner <heiko@sntech.de> # pwm-rockchip
Acked-by: Michael Walle <michael@walle.cc> # pwm-sl28cpld
Acked-by: Neil Armstrong <neil.armstrong@linaro.org> # pwm-meson
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20230804142707.412137-2-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Most but not all pointers to driver data are already called "jz":
$ git grep 'struct jz4740_pwm_chip \*' v6.5-rc1 -- drivers/pwm/pwm-jz4740.c
v6.5-rc1:drivers/pwm/pwm-jz4740.c:static inline struct jz4740_pwm_chip *to_jz4740(struct pwm_chip *chip)
v6.5-rc1:drivers/pwm/pwm-jz4740.c:static bool jz4740_pwm_can_use_chn(struct jz4740_pwm_chip *jz,
v6.5-rc1:drivers/pwm/pwm-jz4740.c: struct jz4740_pwm_chip *jz = to_jz4740(chip);
v6.5-rc1:drivers/pwm/pwm-jz4740.c: struct jz4740_pwm_chip *jz = to_jz4740(chip);
v6.5-rc1:drivers/pwm/pwm-jz4740.c: struct jz4740_pwm_chip *jz = to_jz4740(chip);
v6.5-rc1:drivers/pwm/pwm-jz4740.c: struct jz4740_pwm_chip *jz4740 = to_jz4740(pwm->chip);
v6.5-rc1:drivers/pwm/pwm-jz4740.c: struct jz4740_pwm_chip *jz4740;
Adapt the two variables called "jz4740" to use the same name for
consistency.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Link: https://lore.kernel.org/r/20230808062608.897710-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
We are using a wrong mcbsp functional clock. The interconnect target module
driver provided clock for mcbsp is not same as the mcbsp functional clock
known as the gfclk main_clk. The mcbsp functional clocks for mcbsp should
have been added before we dropped the legacy platform data.
Additionally we are also missing the clock aliases for the clocks used by
the audio driver if reparenting is needed. This causes audio driver errors
like "CLKS: could not clk_get() prcm_fck" for mcbsp as reported by Andreas.
The mcbsp clock aliases too should have been added before we dropped the
legacy platform data.
Let's add the clocks and aliases with a single patch to fix the issue
similar to omap4. On omap5, there is no mcbsp4 instance on the l4_per
interconnect.
Fixes: b1da0fa21b ("ARM: OMAP2+: Drop legacy platform data for omap5 mcbsp")
Cc: H. Nikolaus Schaller <hns@goldelico.com>
Reported-by: Andreas Kemnade <andreas@kemnade.info>
Reported-by: Péter Ujfalusi <peter.ujfalusi@gmail.com>
Acked-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
We are using a wrong mcbsp functional clock. The interconnect target module
driver provided clock for mcbsp is not same as the mcbsp functional clock
known as the gfclk main_clk. The mcbsp functional clocks for mcbsp should
have been added before we dropped the legacy platform data.
Additionally we are also missing the clock aliases for the clocks used by
the audio driver if reparenting is needed. This causes audio driver errors
like "CLKS: could not clk_get() prcm_fck" for mcbsp as reported by Andreas.
The mcbsp clock aliases too should have been added before we dropped the
legacy platform data.
Let's add the clocks and aliases with a single patch to fix the issue.
Fixes: 349355ce3a ("ARM: OMAP2+: Drop legacy platform data for omap4 mcbsp")
Reported-by: Andreas Kemnade <andreas@kemnade.info>
Reported-by: Péter Ujfalusi <peter.ujfalusi@gmail.com>
Acked-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Regulator drivers were modified to use asynchronous device probe. Since
then, the board .init_late hook fails to acquire a GPIO based fixed
regulator needed by an on-board voice MODEM device, and unregisters the
MODEM. That in turn triggers a so far not discovered bug of device
unregister function called for a device with no associated release() op.
serial8250 serial8250.1: incomplete constraints, dummy supplies not allowed
WARNING: CPU: 0 PID: 1 at drivers/base/core.c:2486 device_release+0x98/0xa8
Device 'serial8250.1' does not have a release() function, it is broken and
must be fixed. See Documentation/core-api/kobject.rst.
...
put_device from platform_device_put+0x1c/0x24
platform_device_put from ams_delta_init_late+0x4c/0x68
ams_delta_init_late from init_machine_late+0x1c/0x94
init_machine_late from do_one_initcall+0x60/0x1d4
As a consequence, ASoC CODEC driver is no longer able to control its
device over the voice MODEM's tty interface.
cx20442-codec: ASoC: error at soc_component_write_no_lock
on cx20442-codec for register: [0x00000000] -5
cx20442-codec: ASoC: error at snd_soc_component_update_bits_legacy
on cx20442-codec for register: [0x00000000] -5
cx20442-codec: ASoC: error at snd_soc_component_update_bits
on cx20442-codec for register: [0x00000000] -5
The regulator hangs of a GPIO pin controlled by basic-mmio-gpio driver.
Unlike most GPIO drivers, that driver doesn't probe for devices before
device_initcall, then GPIO pins under its control are not availabele to
majority of devices probed at that phase, including regulators. On the
other hand, serial8250 driver used by the MODEM device neither accepts via
platform data nor handles regulators, then the board file is not able to
teach that driver to return -EPROBE_DEFER when the regulator is not ready
so the failed probe is retried after late_initcall.
Resolve the issue by extending description of the MODEM device with a
dedicated power management domain. Acquire the regulator from the
domain's .activate hook and return -EPROBE_DEFER if the regulator is not
available. Having that under control, add the regulator device
description to the list of platform devices initialized from .init_machine
and drop the no longer needed custom .init_late hook.
v2: Trim down the warning for prettier git log output (Tony).
Fixes: 259b93b21a ("regulator: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in 4.14")
Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com>
Cc: stable@vger.kernel.org # v6.4+
Message-ID: <20231011175038.1907629-1-jmkrzyszt@gmail.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
The PELT acronym definition can be found right at the top of
kernel/sched/pelt.c (of course), but it cannot be found through use of
grep -r PELT kernel/sched/
Add the acronym "(PELT)" after "Per Entity Load Tracking" at the top of
the source file.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20231012125824.1260774-1-mathieu.desnoyers@efficios.com
Kuyo reported sporadic failures on a sched_setaffinity() vs CPU
hotplug stress-test -- notably affine_move_task() remains stuck in
wait_for_completion(), leading to a hung-task detector warning.
Specifically, it was reported that stop_one_cpu_nowait(.fn =
migration_cpu_stop) returns false -- this stopper is responsible for
the matching complete().
The race scenario is:
CPU0 CPU1
// doing _cpu_down()
__set_cpus_allowed_ptr()
task_rq_lock();
takedown_cpu()
stop_machine_cpuslocked(take_cpu_down..)
<PREEMPT: cpu_stopper_thread()
MULTI_STOP_PREPARE
...
__set_cpus_allowed_ptr_locked()
affine_move_task()
task_rq_unlock();
<PREEMPT: cpu_stopper_thread()\>
ack_state()
MULTI_STOP_RUN
take_cpu_down()
__cpu_disable();
stop_machine_park();
stopper->enabled = false;
/>
/>
stop_one_cpu_nowait(.fn = migration_cpu_stop);
if (stopper->enabled) // false!!!
That is, by doing stop_one_cpu_nowait() after dropping rq-lock, the
stopper thread gets a chance to preempt and allows the cpu-down for
the target CPU to complete.
OTOH, since stop_one_cpu_nowait() / cpu_stop_queue_work() needs to
issue a wakeup, it must not be ran under the scheduler locks.
Solve this apparent contradiction by keeping preemption disabled over
the unlock + queue_stopper combination:
preempt_disable();
task_rq_unlock(...);
if (!stop_pending)
stop_one_cpu_nowait(...)
preempt_enable();
This respects the lock ordering contraints while still avoiding the
above race. That is, if we find the CPU is online under rq-lock, the
targeted stop_one_cpu_nowait() must succeed.
Apply this pattern to all similar stop_one_cpu_nowait() invocations.
Fixes: 6d337eab04 ("sched: Fix migrate_disable() vs set_cpus_allowed_ptr()")
Reported-by: "Kuyo Chang (張建文)" <Kuyo.Chang@mediatek.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: "Kuyo Chang (張建文)" <Kuyo.Chang@mediatek.com>
Link: https://lkml.kernel.org/r/20231010200442.GA16515@noisy.programming.kicks-ass.net
We could bail out early when psi was disabled.
Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Chengming Zhou <zhouchengming@bytedance.com>
Link: https://lore.kernel.org/r/20230926115722.467833-1-haifeng.xu@shopee.com
Enable drivers for the hardware blocks present on the Amlogic Meson SoC
based boards: Khadas VIM3 and Hardkernel Odroid N2 to increase testing
coverage.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://lore.kernel.org/r/20231012103600.3381340-1-m.szyprowski@samsung.com
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20231009083856.222030-20-u.kleine-koenig@pengutronix.de
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Add basic support for RZ/G3S to be able to boot from SD card, have a
running console port, and use GPIOs. RZ/G3S has 82 general-purpose IO
ports. Support for the remaining pin functions (e.g. Ethernet, XSPI)
will be added along with controller-specific support.
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20230929053915.1530607-22-claudiu.beznea@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Add documentation for the pin controller found on the Renesas RZ/G3S
(R9A08G045) SoC. Compared to RZ/G2{L,UL}, RZ/G3S has 82 general-purpose
IOs, no slew rate and output impedance support, and more values for
drive strength which needs to be expressed in microamp.
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20230929053915.1530607-21-claudiu.beznea@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
RZ/G3S supports different drive strength values for different power
sources and pin groups (A, B, C). On each group there could be up to 4
drive strength values per power source. Available power sources are
1v8, 2v5, 3v3. Drive strength values are more fine tuned than what was
previously available on the driver thus the necessity of having
micro-amp support. As drive strength and power source values are linked
together the hardware setup for these was moved at the end of
rzg2l_pinctrl_pinconf_set() to ensure proper validation of the new
values.
The drive strength values are expected to be initialized though the
SoC-specific hardware configuration data structure.
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20230929053915.1530607-19-claudiu.beznea@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Move drive strength and output impedance values to the SoC-specific
configuration data structure (struct rzg2l_hwcfg). This allows
extending the drive strength support for RZ/G3S. Along with this the DS
values were converted to uA for simple integration with RZ/G3S support.
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20230929053915.1530607-18-claudiu.beznea@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
On RZ/G3S PFC register allow setting 8 functions for individual ports
(function1 to function8). For function1 the register need to be
configured with 0, for function8 the register need to be configured with
7. We cannot use zero based addressing when requesting functions from
different code places as the documentation
(RZG3S_pinfunction_List_r1.0.xlsx) states explicitly that function0 is
GPIO.
Add a new member to struct rzg2l_hwcfg that will keep the offset that
needs to be substracted before applying a value to a PFC register.
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20230929053915.1530607-17-claudiu.beznea@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
