When calling debugfs_lookup() the result must have dput() called on it,
otherwise the memory will leak over time. To solve this, remove the
lookup and create the directory on the first device found, and then
remove it when the module is unloaded.
Cc: Tudor Ambarus <tudor.ambarus@microchip.com>
Cc: Pratyush Yadav <pratyush@kernel.org>
Cc: Miquel Raynal <miquel.raynal@bootlin.com>
Cc: Richard Weinberger <richard@nod.at>
Cc: Vignesh Raghavendra <vigneshr@ti.com>
Cc: linux-mtd@lists.infradead.org
Reviewed-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20230208160230.2179905-1-gregkh@linuxfoundation.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
CONFIG_SYSFS_DEPRECATED was added in commit 88a22c985e
("CONFIG_SYSFS_DEPRECATED") in 2006 to allow systems with older versions
of some tools (i.e. Fedora 3's version of udev) to boot properly. Four
years later, in 2010, the option was attempted to be removed as most of
userspace should have been fixed up properly by then, but some kernel
developers clung to those old systems and refused to update, so we added
CONFIG_SYSFS_DEPRECATED_V2 in commit e52eec13cd ("SYSFS: Allow boot
time switching between deprecated and modern sysfs layout") to allow
them to continue to boot properly, and we allowed a boot time parameter
to be used to switch back to the old format if needed.
Over time, the logic that was covered under these config options was
slowly removed from individual driver subsystems successfully, removed,
and the only thing that is now left in the kernel are some changes in
the block layer's representation in sysfs where real directories are
used instead of symlinks like normal.
Because the original changes were done to userspace tools in 2006, and
all distros that use those tools are long end-of-life, and older
non-udev-based systems do not care about the block layer's sysfs
representation, it is time to finally remove this old logic and the
config entries from the kernel.
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: linux-block@vger.kernel.org
Cc: linux-doc@vger.kernel.org
Acked-by: Jens Axboe <axboe@kernel.dk>
Link: https://lore.kernel.org/r/20230223073326.2073220-1-gregkh@linuxfoundation.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Zero-length arrays are deprecated and we are moving towards adopting
C99 flexible-array members, instead. So, replace zero-length arrays
declarations in anonymous union with the new __DECLARE_FLEX_ARRAY()
helper macro.
This helper allows for flexible-array members in unions.
Link: https://github.com/KSPP/linux/issues/193
Link: https://github.com/KSPP/linux/issues/227
Link: https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Amadeusz Sławiński <amadeuszx.slawinski@linux.intel.com>
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
LVDS' lpi2c and pwm power domains definitions are missed.
Add them here.
Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
With commit b203e6f1e8 ("arm64: dts: ls1028a: sl28: get MAC addresses
from VPD"), the network adapter now depends on the nvmem device to be
present, which isn't the case and thus breaks networking on this board.
Revert it.
Fixes: b203e6f1e8 ("arm64: dts: ls1028a: sl28: get MAC addresses from VPD")
Signed-off-by: Michael Walle <michael@walle.cc>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
"ranges" is a standard property and we have common helper functions for
parsing it, so let's use them.
Signed-off-by: Rob Herring <robh@kernel.org>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
The correct clock order is "fspi_en" and "fspi". As they are identical
just reordering the names is sufficient.
Fixes: 6276d66984 ("arm64: dts: imx8dxl: add flexspi0 support")
Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
This simplifies the code and silences -517 error messages. Also
the reason is listed in /sys/kernel/debug/devices_deferred.
Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Reviewed-by: Mukesh Ojha <quic_mojha@quicinc.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Without this patch we have relatively high amount of dropped packets.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
IMX6UL_CLK_ENETx_REF is behind of CLK_ENETx_REF_SEL:
FEC MAC <---------- CLK_ENETx_REF_SEL <--------- CLK_ENETx_REF
\
^------<-> CLK_ENETx_REF_PAD
We should point to the clock selector instead. So, we will be able to
use external clock source from CLK_ENETx_REF_PAD as well.
At same time, remove enet_out clk. It is using always the same clock as
enet_clk_ref and do not help to solve any challenges of this HW.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
On this board the PHY is the ref clock provider. So, configure ethernet
reference clock as input.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Old imx6q machine code makes RGMII/RMII clock direction decision based on
configuration of "ptp" clock. "enet_out" is not used and make no real
sense, since we can't configure it as output or use it as clock
provider.
Instead of "enet_out" use "enet_clk_ref" which is actual selector to
choose between internal and external clock source:
FEC MAC <---------- enet_clk_ref <--------- SoC PLL
\
^------<-> refclock PAD (bi directional)
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
The i.MX7d contains a Pixel Pipeline in version 3.0. Add the device tree
node to make it available.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Ethernet refclock direction is board specific and should be configurable
by devicetree. In fact there are board not working with this defaults,
which will be fixed by separate patch.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Current mach-imx6q code has following logic:
- if ptp clock of the ethernet controller node is attached to the SoC
internal enet_ref clock, then we configure RMII reference clock pin as
output by setting IOMUXC_GPR1 BIT(21).
In this case - MAC (SoC) is the clock provider, PHY is the clock consumer.
- if ptp clock of the ethernet controller node is not attached to the
enet_ref clock, then we configure RMII reference clock pin as input by
clearing IOMUXC_GPR1 BIT(21).
In this case - PHY is the clock provider, MAC is the clock consumer.
According to the Freescale MX6SDL ReferenceManual v4, IOMUXC_GPR1 BIT(21)
(page 2033) this configuration bit is not related to the PTP (IEEE1588)
clock:
21 ENET_CLK_SEL - choose enet reference clk mode:
0 - get enet tx reference clk from pad (external OSC for both external
PHY and Internal Controller)
1 - get enet tx reference clk from internal clock from anatop (loopback
through pad), this clock also sent out to external PHY.
According to the Documentation/devicetree/bindings/net/fsl,fec.yaml:
The "ptp"(option), for IEEE1588 timer clock that requires the clock.
The "enet_clk_ref"(option), for MAC transmit/receiver reference clock
like RGMII TXC clock or RMII reference clock. It depends on board
design, the clock is required if RGMII TXC and RMII reference clock
source from SOC internal PLL.
The "enet_out"(option), output clock for external device, like supply
clock for PHY. The clock is required if PHY clock source from SOC.
We can see, that "enet_clk_ref" clock is the best fit for this purpose.
Other properties like "ptp" is designed for IEEE1588 and "enet_out" do
not have real functionality within imx related clock infrastructure.
Since the "enet_clk_ref" is not used by the imx6qdl devicetrees, we can
use it as indicator of potentially properly configured DT. At same time
we can keep ptp clock based logic as the fallback for old DTs.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
While I'm not aware of any problems that have occurred running these
at 100 MHz, the official word from ASRock is that 50 MHz is the
correct speed to use, so let's be safe and use that instead.
Signed-off-by: Zev Weiss <zev@bewilderbeest.net>
Cc: stable@vger.kernel.org
Fixes: 2b81613ce4 ("ARM: dts: aspeed: Add ASRock E3C246D4I BMC")
Fixes: a9a3d60b93 ("ARM: dts: aspeed: Add ASRock ROMED8HM3 BMC")
Link: https://lore.kernel.org/r/20230224000400.12226-4-zev@bewilderbeest.net
Signed-off-by: Joel Stanley <joel@jms.id.au>
When send SMB_COM_NT_CANCEL and RFC1002_SESSION_REQUEST, the
in_send statistic was lost.
Let's move the in_send statistic to the send function to avoid
this scenario.
Fixes: 7ee1af765d ("[CIFS]")
Signed-off-by: Zhang Xiaoxu <zhangxiaoxu5@huawei.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Group some variables based on their sizes to reduce hole and avoid padding.
On x86_64, this shrinks the size from 112 to 96 bytes.
This should have no real impact on memory allocation because 'struct
spi_message' is mostly used on stack, but it can save a few cycles
when the structure is initialized with spi_message_init() and co.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Tested-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Link: https://lore.kernel.org/r/c112aad16eb47808e1ec10abd87b3d273c969a68.1677704283.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Mark Brown <broonie@kernel.org>
variable slave in spi_alloc_master() or spi_alloc_slave()
has been assigned. it is not necessary to be assigned again
Signed-off-by: Lizhe <sensor1010@163.com>
Link: https://lore.kernel.org/r/20230226063334.7489-1-sensor1010@163.com
Signed-off-by: Mark Brown <broonie@kernel.org>
spi_pcpu_stats_totalize() is a rather large macro, and is instantiated
28 times, causing a large amount of duplication in the amount of
generated code.
Reduce the duplication by replacing spi_pcpu_stats_totalize() by a real
C function, and absorb all other common code from
spi_statistics_##name##_show(). As (a) the old "field" parameter was
the name of a structure member, which cannot be passed to a function,
and (b) passing a pointer to the member is also not an option, due to
the loop over all possible CPUs, the "field" parameter is replaced by an
"offset" parameter, pointing to a location within the structure.
This reduces kernel size by ca. 4 KiB (on arm32 and arm64).
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/cb7690d9d04c06eec23dbb98fbb5444082125cff.1677594432.git.geert+renesas@glider.be
Signed-off-by: Mark Brown <broonie@kernel.org>
There is no build time dependency on the Qualcomm platform support so add
an || COMPILE_TEST so we've got better build coverage of the driver.
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230221-spi-arch-deps-v1-6-83d1566474cf@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
There is no build time dependency on the platform support so add an
|| COMPILE_TEST so we've got better build coverage of the driver.
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230221-spi-arch-deps-v1-5-83d1566474cf@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
There is no build time dependency on the DaVicni or Keystone architecture
support so add an || COMPILE_TEST so we've got better build coverage of the
driver.
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230221-spi-arch-deps-v1-4-83d1566474cf@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Currently the NXP Flex SPI driver has a dependency on ARCH_LAYERSCAPE ||
HAS_IOMEM which means that the dependency is almost always true and the
driver available. Really these should be two separate dependencies, with
an || COMPILE_TEST dependency for the architecture to ensure build coverage
is maintained.
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230221-spi-arch-deps-v1-3-83d1566474cf@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
If base support for Rockchip SoCs has been disabled then the SPI driver
won't be terribly useful, add a dependency on ARCH_ROCKCHIP || COMPILE_TEST
to avoid it appearing when not needed.
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230221-spi-arch-deps-v1-2-83d1566474cf@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
The mpc52xx-psc and mpc512x-psc drivers use DT property parsing
functions for 'reg' and 'interrupts', but those are available as
platform device resources. Convert probe functions to use them and
simplify probe to a single function. For 'cell-index', also use the
preferred typed property function.
Signed-off-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20230217-dt-mpc5xxx-spi-v1-3-3be8602fce1e@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Convert the mpc52xx-psc and mpc512x-psc drivers to use the managed
devres variants of functions in probe. Also use dev_err_probe() as
appropriate. With this, the error handling can be simplified.
Signed-off-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20230217-dt-mpc5xxx-spi-v1-2-3be8602fce1e@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Add QSPI reset operation in device probe and add RISCV support to
QUAD SPI Kconfig.
Co-developed-by: Ziv Xu <ziv.xu@starfivetech.com>
Signed-off-by: Ziv Xu <ziv.xu@starfivetech.com>
Signed-off-by: William Qiu <william.qiu@starfivetech.com>
Link: https://lore.kernel.org/r/20230302105221.197421-3-william.qiu@starfivetech.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The QSPI controller needs three reset items to work properly on JH7110 SoC,
so there is need to change the maxItems's value to 3 and add minItems
whose value is equal to 2. Other platforms do not have this constraint.
Signed-off-by: William Qiu <william.qiu@starfivetech.com>
Link: https://lore.kernel.org/r/20230302105221.197421-2-william.qiu@starfivetech.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The RT5739 is a step-down switching voltage regulator that supports
output voltage ragne from 300mV to 1300mV with the wide input supply
voltage range from 2.5V to 5.5V.
Signed-off-by: ChiYuan Huang <cy_huang@richtek.com>
Link: https://lore.kernel.org/r/1676599618-24819-3-git-send-email-cy_huang@richtek.com
Signed-off-by: Mark Brown <broonie@kernel.org>
It seems that a couple of members got lost theirorder, put them back.
Besides that, split field descriptions into groups in the same way
as it's done in the structure definition.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20230220153334.87049-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Some devices lack status registers, yet expect to handle interrupts.
Introduce a no_status flag to indicate such a configuration, where
rather than read a status register to verify, all interrupts received
are assumed to be active.
Cc: Mark Brown <broonie@kernel.org>
Signed-off-by: William Breathitt Gray <william.gray@linaro.org>
Link: https://lore.kernel.org/r/bd501b4b5ff88da24d467f75e8c71b4e0e6f21e2.1677515341.git.william.gray@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Some SoundWire devices have larger width device specific register
maps, in addition to the standard SoundWire 8-bit map. Update the
helpers to allow accessing arbitrarily sized register values and remove
the explicit 8-bit restriction from regmap_sdw_config_check.
Signed-off-by: Lucas Tanure <tanureal@opensource.cirrus.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230112171840.2098463-3-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In the regmap config reg_bits represents the number of address bits not
the number of value bits. Correct the misleading comment which looks a
lot like it suggests the register value itself is 32-bits wide.
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230112171840.2098463-2-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
