Four reasonably small fixes to the core for scsi host allocation
failure paths. The root problem is that we're not freeing the memory
allocated by dev_set_name(), which involves a rejig of may of the free
on error paths to do put_device() instead of kfree which, in turn, has
several other knock on ramifications and inspection turned up a few
other lurking bugs.
Signed-off-by: James E.J. Bottomley <jejb@linux.ibm.com>
-----BEGIN PGP SIGNATURE-----
iJwEABMIAEQWIQTnYEDbdso9F2cI+arnQslM7pishQUCYMYEVCYcamFtZXMuYm90
dG9tbGV5QGhhbnNlbnBhcnRuZXJzaGlwLmNvbQAKCRDnQslM7pishbP/AP4oyLA5
h7T5v7z29prQWn0P3TApcDVvXjOnqPNUzZlvkAEAifnVHLMehlzrJDeaSR0OUf8u
U+SKrsxkiov5XYvwGGU=
=0Vfx
-----END PGP SIGNATURE-----
Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI fixes from James Bottomley:
"Four reasonably small fixes to the core for scsi host allocation
failure paths.
The root problem is that we're not freeing the memory allocated by
dev_set_name(), which involves a rejig of may of the free on error
paths to do put_device() instead of kfree which, in turn, has several
other knock on ramifications and inspection turned up a few other
lurking bugs"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: core: Only put parent device if host state differs from SHOST_CREATED
scsi: core: Put .shost_dev in failure path if host state changes to RUNNING
scsi: core: Fix failure handling of scsi_add_host_with_dma()
scsi: core: Fix error handling of scsi_host_alloc()
All users of arm_pm_restart() have been converted to use the kernel
restart handler.
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
All users of arm_pm_restart() have been converted to use the kernel
restart handler.
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Tested-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
By making use of the kernel restart handler, board specific restart
handlers can be prioritized amongst available mechanisms for a particular
board or system.
Select the default priority of 128 to indicate that the restart callback
in the machine description is the default restart mechanism.
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Register with kernel restart handler instead of setting arm_pm_restart
directly. This enables support for replacing the PSCI restart handler
with a different handler if necessary for a specific board.
Select a priority of 129 to indicate a higher than default priority, but
keep it as low as possible since PSCI reset is known to fail on some
boards.
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Tested-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Register with kernel restart handler instead of setting arm_pm_restart
directly.
Select a high priority of 192 to ensure that default restart handlers
are replaced if Xen is running.
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
This reverts commit e220ba6022.
The VERIFY_PERMISSION_FAULT is introduced since 2009 but no
one use it, just revert it and clean unused comment.
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Using our knowledge of where the physical kernel sections start
and end we can split mapping of lowmem and kernel apart.
This is helpful when you want to place the kernel independently
from lowmem and not be limited to putting it into lowmem only,
but also into places such as the VMALLOC area.
We extensively rewrite the lowmem mapping code to account for
all cases where the kernel image overlaps with the lowmem in
different ways. This is helpful to handle situations which
occur when the kernel is loaded in different places and makes
it possible to place the kernel in a more random manner
which is done with e.g. KASLR.
We sprinkle some comments with illustrations and pr_debug()
over it so it is also very evident to readers what is happening.
We now use the kernel_sec_start and kernel_sec_end instead
of relying on __pa() (phys_to_virt) to provide this. This
is helpful if we want to resolve physical-to-virtual and
virtual-to-physical mappings at runtime rather than
compiletime, especially if we are not using patch phys to
virt.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
When we are mapping the initial sections in head.S we
know very well where the start and end of the kernel image
in physical memory is placed. Later on it gets hard
to determine this.
Save the information into two variables named
kernel_sec_start and kernel_sec_end for convenience
for later work involving the physical start and end
of the kernel. These variables are section-aligned
corresponding to the early section mappings set up
in head.S.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
We want to be able to compile the kernel into an address different
from PAGE_OFFSET (start of lowmem) + TEXT_OFFSET, so start to pry
apart the address of where the kernel is located from the address
where the lowmem is located by defining and using KERNEL_OFFSET in
a few key places.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Add the power controller node and the correspondending qos nodes for
RK322x.
Also add the power-domain property to the nodes that are already
present.
Signed-off-by: Alex Bee <knaerzche@gmail.com>
Link: https://lore.kernel.org/r/20210527154455.358869-10-knaerzche@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Add the power controller node and the correspondending qos nodes for
RK3036.
Also add the power-domain property to the nodes that are already
present.
Note: Since the regiser offsets of the axi interconnect QoS are missing
in the TRM (RK3036 TRM V1.0), they have been taken from vendor kernel.
Signed-off-by: Alex Bee <knaerzche@gmail.com>
Link: https://lore.kernel.org/r/20210527154455.358869-9-knaerzche@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
While the kernel doesn't care so much right now, boot loaders like
u-boot need to refine the node on their side, so to make life easier
for everyone add the labels to the timer nodes on rk3066a.
Signed-off-by: Johan Jonker <jbx6244@gmail.com>
Link: https://lore.kernel.org/r/20210612184733.2331-1-jbx6244@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The struct counter_synapse actions_list member expects a const enum
counter_synapse_action array. This patch renames
interrupt_cnt_synapse_actionss to interrupt_cnt_synapse_actions and adds
a const qualifier to match actions_list.
Cc: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: William Breathitt Gray <vilhelm.gray@gmail.com>
Reviewed-by: <o.rempel@pengutronix.de>
Link: https://lore.kernel.org/r/20210610013642.149961-1-vilhelm.gray@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
This patch marks the struct ltr501_chip_info as constant.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Tested-by: Nikita Travkin <nikita@trvn.ru> # ltr559
Link: https://lore.kernel.org/r/20210610134619.2101372-5-mkl@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The PS ADC Channel data is spread over 2 registers in little-endian
form. This patch adds the missing endianness conversion.
Fixes: 2690be9051 ("iio: Add Lite-On ltr501 ambient light / proximity sensor driver")
Signed-off-by: Oliver Lang <Oliver.Lang@gossenmetrawatt.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Tested-by: Nikita Travkin <nikita@trvn.ru> # ltr559
Link: https://lore.kernel.org/r/20210610134619.2101372-4-mkl@pengutronix.de
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The ltr559 chip uses only the lowest bit of the ALS_CONTR register to
configure between active and stand-by mode. In the original driver
BIT(1) is used, which does a software reset instead.
This patch fixes the problem by using BIT(0) as als_mode_active for
the ltr559 chip.
Fixes: 8592a7eefa ("iio: ltr501: Add support for ltr559 chip")
Signed-off-by: Oliver Lang <Oliver.Lang@gossenmetrawatt.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Tested-by: Nikita Travkin <nikita@trvn.ru> # ltr559
Link: https://lore.kernel.org/r/20210610134619.2101372-3-mkl@pengutronix.de
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The regmap is configured for 8 bit registers, uses a RB-Tree cache and
marks several registers as volatile (i.e. do not cache).
The ALS and PS data registers in the chip are 16 bit wide and spans
two regmap registers. In the current driver only the base register is
marked as volatile, resulting in the upper register only read once.
Further the data sheet notes:
| When the I2C read operation starts, all four ALS data registers are
| locked until the I2C read operation of register 0x8B is completed.
Which results in the registers never update after the 2nd read.
This patch fixes the problem by marking the upper 8 bits of the ALS
and PS registers as volatile, too.
Fixes: 2f2c96338a ("iio: ltr501: Add regmap support.")
Reported-by: Oliver Lang <Oliver.Lang@gossenmetrawatt.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Tested-by: Nikita Travkin <nikita@trvn.ru> # ltr559
Link: https://lore.kernel.org/r/20210610134619.2101372-2-mkl@pengutronix.de
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Commit c1d1c4a62d ("iio: accel: bma180: BMA254 support") added
BMA254 support to the bma180 driver and changed some naming to BMA25x
to make it easier to add support for BMA253 and BMA255.
Unfortunately, there is quite some overlap between the bma180 driver
and the bmc150-accel driver. Back when the commit was made, the
bmc150-accel driver actually already had support for BMA255, and
adding support for BMA254 would have been as simple as adding a new
compatible to bmc150-accel.
The bmc150-accel driver is a bit better for BMA254 since it also
supports the motion trigger/interrupt functionality. Fortunately,
moving BMA254 support over to bmc150-accel is fairly simple because
the drivers have compatible device tree bindings.
Revert most of the changes for BMA254 support in bma180 and move
BMA254 over to bmc150-accel. This has the following advantages:
- Support for motion trigger/interrupt
- Fix incorrect scale values (BMA254 currently uses the same as
BMA250 but actually they're different because of 10 vs 12 bits
data size)
- Less code than before :)
BMA250 could be potentially also moved but it's more complicated
because its chip_id conflicts with the one for BMA222 in bmc150-accel.
Perhaps there are also other register differences, I did not investigate
further yet (and I have no way to test it).
Cc: Peter Meerwald <pmeerw@pmeerw.net>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-11-stephan@gerhold.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
BMA254 is very similar to BMA253/BMA255 which are both supported by
the bmc150-accel driver. In general, there is quite some overlap between
the bma180 and bmc150-accel driver, but the bmc150-accel driver has a few
more features (e.g. motion trigger/interrupt).
Let's move bma254 over to the bma255 schema (bmc150-accel driver).
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-10-stephan@gerhold.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
BMA253 has two interrupt pins (INT1 and INT2) that can be configured
independently. At the moment the bmc150-accel driver does not make use
of them but it might be able to in the future, so it's useful to already
specify all available interrupts in the device tree.
Set maxItems: 2 for interrupts to allow specifying a second one.
This is necessary as preparation to move the bosch,bma254 compatible
from bosch,bma180.yaml to bosch,bma255.yaml since bma180 allows two
interrupts, but BMA254 is better supported by the bmc150-accel driver.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-9-stephan@gerhold.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
BMA253 is mostly like BMA255 and has exactly the same register layout
as used by the bmc150-accel driver as far I can tell. Making it work
is as simple as adding new device IDs for it since it has the same
chip_id = 0xFA (250) as BMA255 and others.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-8-stephan@gerhold.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
BMA253 is mostly like BMA255 that is already supported by the
bmc150-accel driver. Document an extra bosch,bma253 compatible for it.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-7-stephan@gerhold.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Right now all the device IDs are listed in seemingly random order,
make this consistent by ordering those alphabetically. Also, order
bmc150_accel_chip_info_tbl by chip ID for the same reason.
Suggested-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-6-stephan@gerhold.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The chips supported by the bmc150-accel driver are clearly documented
in Kconfig, in the bmc150_accel_chip_info_tbl as well as in all the
device ID tables in the I2C/SPI drivers. It's easy to forget to update
the lists in the file header. Drop those entirely to reduce the amount
of changes required to add new chip variants.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-5-stephan@gerhold.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Commit 0ad4bf3701 ("iio:accel:bmc150-accel: Use the chip ID to detect
sensor variant") stopped using the I2C/ACPI match data to look up the
bmc150_accel_chip_info. However, the bmc150_accel_chip_info_tbl remained
as-is, with multiple entries with the same chip_id (e.g. 0xFA for
BMC150/BMI055/BMA255). This is redundant now because actually the driver
will always select the first entry with a matching chip_id.
So even if a device probes e.g. with BMA0255 it will end up using the
chip_info for BMC150. And in general that's fine for now, the entries
for BMC150/BMI055/BMA255 were exactly the same anyway (except for the
name, which is replaced with the more accurate one later).
But in this case it's misleading because it suggests that one should
add even more entries with the same chip_id when adding support for
new variants. Let's make that more clear by removing the enum with
the chip identifiers entirely and instead have only one entry per
chip_id.
Note that we may need to bring back some mechanism to differentiate
between different chips with the same chip_id in the future.
For example, BMA250 (currently supported by the bma180 driver) has the
same chip_id = 0x03 as BMA222 even though they have different channel
sizes (8 bits vs 10 bits). But in any case, that mechanism would
need to look quite different from what we have right now.
Cc: Bastien Nocera <hadess@hadess.net>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-4-stephan@gerhold.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The Kconfig option currently says that all Bosch accelerometers
supported by the bmc150-accel driver are combo chips with both
accelerometer and magnetometer. This is wrong: actually only BMC150
is such a combo. The BMA* variants only contain an accelerometer
and the BMI055 actually is a accelerometer + gyroscope combo.
Clarify this in the help text and also make the list of supported
variants complete and sorted for consistency.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-3-stephan@gerhold.net
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
According to sysfs-bus-iio documentation the unit for accelerometer
values after applying scale/offset should be m/s^2, not g, which explains
why the scale values for the other variants in bmc150-accel do not match
exactly the values given in the datasheet.
To get the correct values, we need to multiply the BMA222 scale values
by g = 9.80665 m/s^2.
Fixes: a1a210bf29 ("iio: accel: bmc150-accel: Add support for BMA222")
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210611080903.14384-2-stephan@gerhold.net
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Gcc reports build error when CONFIG_IIO_TRIGGERED_BUFFER is not set:
riscv64-linux-gnu-ld: drivers/iio/dummy/iio_simple_dummy_buffer.o: in function `iio_simple_dummy_configure_buffer':
iio_simple_dummy_buffer.c:(.text+0x2b0): undefined reference to `iio_triggered_buffer_setup_ext'
riscv64-linux-gnu-ld: drivers/iio/dummy/iio_simple_dummy_buffer.o: in function `.L0 ':
iio_simple_dummy_buffer.c:(.text+0x2fc): undefined reference to `iio_triggered_buffer_cleanup'
Fix it by select IIO_TRIGGERED_BUFFER for config IIO_SIMPLE_DUMMY_BUFFER.
Fixes: 738f6ba118 ("iio: dummy: iio_simple_dummy_buffer: use triggered buffer core calls")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Disable reg and clk when devm_gpiod_get_optional() fails in adf4350_probe().
Fixes:4a89d2f47ccd ("iio: adf4350: Convert to use GPIO descriptor")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20210601142605.3613605-1-yangyingliang@huawei.com
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
According to the BMA253 datasheet [1] and BMA250 datasheet [2] the
bandwidth value for BMA25x should be set as 01xxx:
"Settings 00xxx result in a bandwidth of 7.81 Hz; [...]
It is recommended [...] to use the range from ´01000b´ to ´01111b´
only in order to be compatible with future products."
However, at the moment the drivers sets bandwidth values from 0 to 6,
which is not recommended and always results into 7.81 Hz bandwidth
according to the datasheet.
Fix this by introducing a bw_offset = 8 = 01000b for BMA25x,
so the additional bit is always set for BMA25x.
[1]: https://www.bosch-sensortec.com/media/boschsensortec/downloads/datasheets/bst-bma253-ds000.pdf
[2]: https://datasheet.octopart.com/BMA250-Bosch-datasheet-15540103.pdf
Cc: Peter Meerwald <pmeerw@pmeerw.net>
Fixes: 2017cff24c ("iio:bma180: Add BMA250 chip support")
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20210526094408.34298-2-stephan@gerhold.net
Cc: <Stable@vger.kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Use to_pci_dev() helper instead of container_of(d, struct pci_dev, dev);
Suggested-by: Jonathan Cameron <jic23@kernel.org>
Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com>
Link: https://lore.kernel.org/r/20210611115558.796338-2-jarkko.nikula@linux.intel.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Remove CONFIG_PM ifdef and mark PM callbacks with __maybe_unused.
Suggested-by: Jonathan Cameron <jic23@kernel.org>
Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com>
Link: https://lore.kernel.org/r/20210611115558.796338-1-jarkko.nikula@linux.intel.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
These were usually used before the conversion to devm_ functions, so that
the remove hook would be able to retrieve the pointer and do cleanups on
remove.
When the conversion happened, they should have been removed, but were
omitted.
Some drivers were copied from drivers that fit the criteria described
above. In any case, in order to prevent more drivers from being used as
example (and have spi_set_drvdata() needlessly set), this change removes it
from the IIO ADC group.
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513111035.77950-1-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
These were usually used before the conversion to devm_ functions, so that
the remove hook would be able to retrieve the pointer and do cleanups on
remove.
When the conversion happened, they should have been removed, but were
omitted.
Some drivers were copied from drivers that fit the criteria described
above. In any case, in order to prevent more drivers from being used as
example (and have spi_set_drvdata() needlessly set), this change removes it
from the IIO IMU group.
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513122512.93187-1-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Since all AD Sigma-Delta drivers now use the
devm_ad_sd_setup_buffer_and_trigger() function, we can remove the old
ad_sd_{setup,cleanup}_buffer_and_trigger() functions.
This way we can discourage new drivers that use the ad_sigma_delta
lib-driver to use these (older functions).
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513120752.90074-13-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
As not many steps were not already devm_ managed, use
devm_add_action_or_reset() to handle the rest.
This also uses the new devm_ad_sd_setup_buffer_and_trigger() function.
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513120752.90074-12-aardelean@deviqon.com
With the devm_ad_sd_setup_buffer_and_trigger() helper, it's a bit easier
now to convert the probe of the AD7192 driver to use device-managed
functions.
The regulators and the mclk requires devm_add_action_or_reset() callbacks
though.
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513120752.90074-11-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The devm_clk_get_optional() helper returns NULL when devm_clk_get() returns
-ENOENT.
This makes things slightly cleaner. The added benefit is mostly cosmetic.
Also, a minor detail with this call, is that the reference for the parent
device is taken as `spi->dev` instead of `&st->sd.spi->dev` (which looks a
little quirky).
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513120752.90074-10-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
With the devm_ad_sd_setup_buffer_and_trigger() helper, it's a bit easier
now to convert the probe of the AD7780 driver to use device-managed
functions.
Only the regulator disable requires a devm_add_action_or_reset() callback.
This change does that, cleaning up the driver a bit.
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513120752.90074-9-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
With the devm_ad_sd_setup_buffer_and_trigger() helper, it's a bit easier
now to convert the probe of the AD7791 driver to use device-managed
functions.
Only the regulator disable requires a devm_add_action_or_reset() callback.
This change does that, cleaning up the driver a bit.
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513120752.90074-8-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
With the devm_ad_sd_setup_buffer_and_trigger() helper, it's a bit easier
now to convert the probe of the AD7793 driver to use device-managed
functions.
Only the regulator disable requires a devm_add_action_or_reset() callback.
This change does that, cleaning up the driver a bit.
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513120752.90074-7-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
This is a version of ad_sd_setup_buffer_and_trigger() with all underlying
functions (that are used) being replaced with their device-managed
variants.
One thing to take care here is with {devm_}iio_trigger_alloc(), where both
functions take a parent-device object as the first parameter.
To make sure nothing quirky is happening, the devm_ad_sd_probe_trigger()
function is checking that the provided 'dev' reference is the same as the
one stored on the 'struct ad_sigma_delta' driver data.
Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210513120752.90074-6-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The SOC_SIFIVE Kconfig entry unconditionally selects ERRATA_SIFIVE.
However, ERRATA_SIFIVE depends on RISCV_ERRATA_ALTERNATIVE, which is
not set, so SOC_SIFIVE should either depend on or select
RISCV_ERRATA_ALTERNATIVE. Use 'select' here to quieten the Kconfig
warning.
WARNING: unmet direct dependencies detected for ERRATA_SIFIVE
Depends on [n]: RISCV_ERRATA_ALTERNATIVE [=n]
Selected by [y]:
- SOC_SIFIVE [=y]
Fixes: 1a0e5dbd37 ("riscv: sifive: Add SiFive alternative ports")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Albert Ou <aou@eecs.berkeley.edu>
Cc: linux-riscv@lists.infradead.org
Cc: Vincent Chen <vincent.chen@sifive.com>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
When CONFIG_CMODEL_MEDLOW is used it ends up generating riscv_hi20_rela
relocations in modules which are not resolved during runtime and
following errors would be seen
[ 4.802714] virtio_input: target 00000000c1539090 can not be addressed by the 32-bit offset from PC = 39148b7b
[ 4.854800] virtio_input: target 00000000c1539090 can not be addressed by the 32-bit offset from PC = 9774456d
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@google.com>
* A pair of XIP fixes: one to fix alternatives, and one to turn off the
rest of the features that require code modification.
* A fix to a type that was causing some alternatives to break.
* A build fix for BUILTIN_DTB.
-----BEGIN PGP SIGNATURE-----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=c3op
-----END PGP SIGNATURE-----
Merge tag 'riscv-for-linus-5.13-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux
Pull RISC-V fixes from Palmer Dabbelt:
- A pair of XIP fixes: one to fix alternatives, and one to turn off the
rest of the features that require code modification
- A fix to a type that was causing some alternatives to break
- A build fix for BUILTIN_DTB
* tag 'riscv-for-linus-5.13-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
riscv: Fix BUILTIN_DTB for sifive and microchip soc
riscv: alternative: fix typo in macro name
riscv: code patching only works on !XIP_KERNEL
riscv: xip: support runtime trap patching
0day robot reported a 9.2% regression for will-it-scale mmap1 test
case[1], caused by commit 57efa1fe59 ("mm/gup: prevent gup_fast from
racing with COW during fork").
Further debug shows the regression is due to that commit changes the
offset of hot fields 'mmap_lock' inside structure 'mm_struct', thus some
cache alignment changes.
From the perf data, the contention for 'mmap_lock' is very severe and
takes around 95% cpu cycles, and it is a rw_semaphore
struct rw_semaphore {
atomic_long_t count; /* 8 bytes */
atomic_long_t owner; /* 8 bytes */
struct optimistic_spin_queue osq; /* spinner MCS lock */
...
Before commit 57efa1fe59 adds the 'write_protect_seq', it happens to
have a very optimal cache alignment layout, as Linus explained:
"and before the addition of the 'write_protect_seq' field, the
mmap_sem was at offset 120 in 'struct mm_struct'.
Which meant that count and owner were in two different cachelines,
and then when you have contention and spend time in
rwsem_down_write_slowpath(), this is probably *exactly* the kind
of layout you want.
Because first the rwsem_write_trylock() will do a cmpxchg on the
first cacheline (for the optimistic fast-path), and then in the
case of contention, rwsem_down_write_slowpath() will just access
the second cacheline.
Which is probably just optimal for a load that spends a lot of
time contended - new waiters touch that first cacheline, and then
they queue themselves up on the second cacheline."
After the commit, the rw_semaphore is at offset 128, which means the
'count' and 'owner' fields are now in the same cacheline, and causes
more cache bouncing.
Currently there are 3 "#ifdef CONFIG_XXX" before 'mmap_lock' which will
affect its offset:
CONFIG_MMU
CONFIG_MEMBARRIER
CONFIG_HAVE_ARCH_COMPAT_MMAP_BASES
The layout above is on 64 bits system with 0day's default kernel config
(similar to RHEL-8.3's config), in which all these 3 options are 'y'.
And the layout can vary with different kernel configs.
Relayouting a structure is usually a double-edged sword, as sometimes it
can helps one case, but hurt other cases. For this case, one solution
is, as the newly added 'write_protect_seq' is a 4 bytes long seqcount_t
(when CONFIG_DEBUG_LOCK_ALLOC=n), placing it into an existing 4 bytes
hole in 'mm_struct' will not change other fields' alignment, while
restoring the regression.
Link: https://lore.kernel.org/lkml/20210525031636.GB7744@xsang-OptiPlex-9020/ [1]
Reported-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Feng Tang <feng.tang@intel.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Cc: Peter Xu <peterx@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Alex Elder says:
====================
net: qualcomm: rmnet: MAPv4 download checksum cleanup, part 2
This is part 2 of a large series that reworks some code that handles
downloaded packets when MAPv4 checksum offload is enabled. The
first part, which includes an overview, is here:
https://lore.kernel.org/netdev/20210611190529.3085813-1-elder@linaro.org/
This second part of the series completes the simplification of this
handling code, removing unnecessary byte swaps and bitwise inversions
of checksum values, and along the way avoids the need for almost all
of the forced type casts. The checksum field in an RMNet download
trailer is given __sum16_t type to accurately reflect the meaning of
that field.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>