]> git.ipfire.org Git - thirdparty/kernel/linux.git/log
thirdparty/kernel/linux.git
5 months agohwmon: (core) Use device name as a fallback in devm_hwmon_device_register_with_info
Heiner Kallweit [Thu, 9 Jan 2025 19:53:55 +0000 (20:53 +0100)] 
hwmon: (core) Use device name as a fallback in devm_hwmon_device_register_with_info

A number of network PHY drivers use the following code:

name = devm_hwmon_sanitize_name(dev, dev_name(dev));
if (IS_ERR(name))
return PTR_ERR(name);
devm_hwmon_device_register_with_info(dev, name, ..);

Make this a generic fallback option and use the device name if no name
is provided to devm_hwmon_device_register_with_info(). This would allow
to simplify the affected drivers.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://lore.kernel.org/r/1ebe6961-6445-4408-bfb4-b56173af9db5@gmail.com
[groeck: Update API document]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
5 months agohwmon: (pmbus/max15301) Add support for MAX15303
Nuno Sa [Wed, 8 Jan 2025 16:44:23 +0000 (16:44 +0000)] 
hwmon: (pmbus/max15301) Add support for MAX15303

The MAX15303 is a full-featured, flexible, efficient, 6A digital
point-of-load (PoL) controller with integrated switching MOSFETs. It
contains advanced power management and telemetry features.

Tested-by: Robin Getz <rgetz503@gmail.com>
Signed-off-by: Nuno Sa <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20250108-dev-max15303-v1-1-9154eb2c9d9e@analog.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
5 months agohwmon: (pmbus/adm1275) add adm1273 support
John Erasmus Mari Geronimo [Mon, 6 Jan 2025 13:17:40 +0000 (21:17 +0800)] 
hwmon: (pmbus/adm1275) add adm1273 support

Add support for adm1273 which is similar to adm1275 and other chips
of the series.

Signed-off-by: John Erasmus Mari Geronimo <johnerasmusmari.geronimo@analog.com>
Link: https://lore.kernel.org/r/20250106131740.305988-3-johnerasmusmari.geronimo@analog.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
5 months agodt-bindings: hwmon: adm1275: add adm1273
John Erasmus Mari Geronimo [Mon, 6 Jan 2025 13:17:39 +0000 (21:17 +0800)] 
dt-bindings: hwmon: adm1275: add adm1273

Add support for the adm1273 Hot-Swap Controller and Digital Power
and Energy Monitor

Signed-off-by: John Erasmus Mari Geronimo <johnerasmusmari.geronimo@analog.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20250106131740.305988-2-johnerasmusmari.geronimo@analog.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
5 months agohwmon: (nct6683) Add another customer ID for MSI
John Audia [Wed, 1 Jan 2025 13:27:59 +0000 (08:27 -0500)] 
hwmon: (nct6683) Add another customer ID for MSI

The new customer ID was found on a MPG X870E CARBON WIFI (MS-7E49) with
a NCT6687D chip.

Signed-off-by: John Audia <therealgraysky@proton.me>
[groeck: Resolved conflicts in Documentation/hwmon/nct6683.rst]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
5 months agohwmon: (pwm-fan): Make use of device properties everywhere
Peter Korsgaard [Thu, 2 Jan 2025 17:04:29 +0000 (18:04 +0100)] 
hwmon: (pwm-fan): Make use of device properties everywhere

Commit 255ab27a0743 ("hwmon: (pwm-fan) Introduce start from stopped state
handling") added two of_property_read_u32() calls after the driver was
reworked to use device_property_* in commit dfd977d85b15 ("hwmon: (pwm-fan)
Make use of device properties"), so convert those as well.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Link: https://lore.kernel.org/r/20250102170429.791912-1-peter@korsgaard.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
5 months agohwmon: (lm75) add I3C support for P3T1755
Wolfram Sang [Fri, 20 Dec 2024 09:36:34 +0000 (10:36 +0100)] 
hwmon: (lm75) add I3C support for P3T1755

Introduce I3C support by defining I3C accessors for regmap and
implementing an I3C driver. Enable I3C for the NXP P3T1755.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20241220093635.11218-1-wsa+renesas@sang-engineering.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (lm75) separate probe into common and I2C parts
Wolfram Sang [Thu, 19 Dec 2024 22:55:26 +0000 (23:55 +0100)] 
hwmon: (lm75) separate probe into common and I2C parts

Put generic probe functionality into a separate function and let the I2C
driver call it. This is a preparation for adding I3C support which will
also use the generic probe function.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20241219225522.3490-11-wsa+renesas@sang-engineering.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (lm75) Remove superfluous 'client' member from private struct
Wolfram Sang [Thu, 19 Dec 2024 22:55:25 +0000 (23:55 +0100)] 
hwmon: (lm75) Remove superfluous 'client' member from private struct

The regmap-only conversion allows us to store the client-pointer as the
'context' parameter for regmap. This not only makes the private struct
smaller, but also allows proper separation of I2C and I3C in the future.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20241219225522.3490-10-wsa+renesas@sang-engineering.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (lm75) simplify regulator handling
Wolfram Sang [Thu, 19 Dec 2024 22:55:24 +0000 (23:55 +0100)] 
hwmon: (lm75) simplify regulator handling

devm_regulator_get_enable() was introduced exactly to avoid open coding
regulator handling like in this driver. Make use of this helper.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20241219225522.3490-9-wsa+renesas@sang-engineering.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (lm75) simplify lm75_write_config()
Wolfram Sang [Thu, 19 Dec 2024 22:55:23 +0000 (23:55 +0100)] 
hwmon: (lm75) simplify lm75_write_config()

After previous refactoring, it is now possible to make
lm75_write_config() a simple inline function.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20241219225522.3490-8-wsa+renesas@sang-engineering.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (lm75) Hide register size differences in regmap access functions
Guenter Roeck [Tue, 17 Dec 2024 17:29:10 +0000 (09:29 -0800)] 
hwmon: (lm75) Hide register size differences in regmap access functions

Hide register size differences in regmap access functions to simplify
runtime code and to simplify adding support for I3C devices. Also
use regmap API functions for bit operations where possible.

For this to work, the 16-bit and 8-bit configuration register has to be
mapped to a 16-bit value. Unlike other registers, this register is a
low-byte-first register, presumably for compatibility with chips with
8-bit wide configuration registers. Hide the differences in the regmap
access code.

While at it, enable alarm attribute support for TMP112.

Cc: Wolfram Sang <wsa+renesas@sang-engineering.com>
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>
6 months agohwmon: (pmbus/crps) Add Intel CRPS185 power supply
Ninad Palsule [Tue, 17 Dec 2024 17:35:33 +0000 (11:35 -0600)] 
hwmon: (pmbus/crps) Add Intel CRPS185 power supply

Add the driver to monitor Intel common redundant power supply (crps185)
with hwmon over pmbus.

Signed-off-by: Ninad Palsule <ninad@linux.ibm.com>
Link: https://lore.kernel.org/r/20241217173537.192331-3-ninad@linux.ibm.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agodt-bindings: hwmon: intel,crps185: Add to trivial
Ninad Palsule [Tue, 17 Dec 2024 17:35:34 +0000 (11:35 -0600)] 
dt-bindings: hwmon: intel,crps185: Add to trivial

Add INTEL Common Redundant Power Supply Versions crps185 bindings as
trivial. The hardware does not have any resources like clocks which are
required to be included in the device tree.

Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Ninad Palsule <ninad@linux.ibm.com>
Link: https://lore.kernel.org/r/20241217173537.192331-4-ninad@linux.ibm.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (lm75) Fix LM75B document link
Wolfram Sang [Wed, 18 Dec 2024 07:41:34 +0000 (08:41 +0100)] 
hwmon: (lm75) Fix LM75B document link

NXP reorganized their website. Update the link for the LM75B datasheet.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20241218074131.4351-8-wsa+renesas@sang-engineering.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (lm75) Add NXP P3T1755 support
Wolfram Sang [Wed, 18 Dec 2024 07:41:33 +0000 (08:41 +0100)] 
hwmon: (lm75) Add NXP P3T1755 support

Add this LM75 compatible sensor which needs a separate entry because of
its default sampling time and SMBusAlert handling.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20241218074131.4351-7-wsa+renesas@sang-engineering.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agodt-bindings: hwmon: lm75: Add NXP P3T1755
Wolfram Sang [Wed, 18 Dec 2024 07:41:32 +0000 (08:41 +0100)] 
dt-bindings: hwmon: lm75: Add NXP P3T1755

Add this LM75 compatible sensor which needs a separate entry because of
its default sampling time and SMBusAlert handling.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20241218074131.4351-6-wsa+renesas@sang-engineering.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (pmbus/core) Add PMBUS_REVISION in debugfs
Ninad Palsule [Mon, 16 Dec 2024 17:50:39 +0000 (11:50 -0600)] 
hwmon: (pmbus/core) Add PMBUS_REVISION in debugfs

Add debugfs file for the PMBUS_REVISION command. This command provides
information about PMBus protocol revision number.

Signed-off-by: Ninad Palsule <ninad@linux.ibm.com>
Link: https://lore.kernel.org/r/20241216175044.4144442-2-ninad@linux.ibm.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (occ/p9_sbe) Constify 'struct bin_attribute'
Thomas Weißschuh [Sun, 15 Dec 2024 15:19:04 +0000 (16:19 +0100)] 
hwmon: (occ/p9_sbe) Constify 'struct bin_attribute'

The sysfs core now allows instances of 'struct bin_attribute' to be
moved into read-only memory. Make use of that to protect them against
accidental or malicious modifications.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/r/20241215-sysfs-const-bin_attr-hwmon-v1-1-ea72a6a46c36@weissschuh.net
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (nct6683) Add customer ID for ASRock B650 Steel Legend WiFi
Max Ammann [Sat, 14 Dec 2024 10:22:01 +0000 (11:22 +0100)] 
hwmon: (nct6683) Add customer ID for ASRock B650 Steel Legend WiFi

This value was found on an ASRock B650 Steel Legend WiFi with an
NCT6686D chip.

Signed-off-by: Max Ammann <max@maxammann.org>
Link: https://lore.kernel.org/r/20241214102201.122851-1-max@maxammann.org
[groeck: htmldocs fix]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (isl28022) Drop explicit initialization of struct i2c_device_id::driver_data...
Uwe Kleine-König [Wed, 4 Dec 2024 14:08:57 +0000 (15:08 +0100)] 
hwmon: (isl28022) Drop explicit initialization of struct i2c_device_id::driver_data to 0

This driver doesn't use the driver_data member of struct i2c_device_id,
so don't explicitly initialize it.

This prepares putting driver_data in an anonymous union which requires
either no initialization or named designators. But it's also a nice
cleanup on its own.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@baylibre.com>
Link: https://lore.kernel.org/r/20241204140857.1691402-2-u.kleine-koenig@baylibre.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (tmp108) Drop of_match_ptr() protection
Guenter Roeck [Wed, 4 Dec 2024 14:17:54 +0000 (06:17 -0800)] 
hwmon: (tmp108) Drop of_match_ptr() protection

Limiting the scope of devicetree support to CONFIG_OF prevents use of this
driver with ACPI via PRP0001. Drop the dependency.

While at it,
- Switch of.h for mod_devicetable.h include given the use of struct
  of_device_id which is defined in that header
- Add I2C device ID for p3t1085
- Drop the unnecessary include of hwmon-sysfs.h.

Cc: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20241204141754.4051186-1-linux@roeck-us.net
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (tmp108) Add basic regulator support
Stanislav Jakubek [Tue, 26 Nov 2024 08:40:35 +0000 (09:40 +0100)] 
hwmon: (tmp108) Add basic regulator support

TMP108/P3T1085 are powered by the V+/VCC regulator, add support for it.

Signed-off-by: Stanislav Jakubek <stano.jakubek@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (asus-ec-sensors) Add TUF GAMING X670E PLUS
Li XingYang [Sun, 1 Dec 2024 03:47:45 +0000 (11:47 +0800)] 
hwmon: (asus-ec-sensors) Add TUF GAMING X670E PLUS

Add support for TUF GAMING X670E PLUS.

Signed-off-by: Li XingYang <yanhuoguifan@gmail.com>
[groeck: Massaged description]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (asus-ec-sensors) Add support for fan cpu opt on AMD 600 motherboards
Li XingYang [Sun, 1 Dec 2024 03:47:44 +0000 (11:47 +0800)] 
hwmon: (asus-ec-sensors) Add support for fan cpu opt on AMD 600 motherboards

According to the LibreHardwareMonitor project, LibreHardwareMonitorLib/
Hardware/Motherboard/Lpc/EC/EmbeddedController.cs file, fan cpu opt
is supported by the AMD600 EC. Registers are described as follows.

BoardFamily.Amd600, new Dictionary<ECSensor, EmbeddedControllerSource>
{
{ ECSensor.FanCPUOpt,  new EmbeddedControllerSource
("CPU Optional Fan", SensorType.Fan, 0x00b0, 2) },
}

Add FanCPUOpt support for AMD600.

Signed-off-by: Li XingYang <yanhuoguifan@gmail.com>
[groeck: Massaged subject and description]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (nct6775): Actually make use of the HWMON_NCT6775 symbol namespace
Uwe Kleine-König [Tue, 3 Dec 2024 17:31:49 +0000 (18:31 +0100)] 
hwmon: (nct6775): Actually make use of the HWMON_NCT6775 symbol namespace

DEFAULT_SYMBOL_NAMESPACE must already be defined when <linux/export.h>
is included. So move the define above the include block.

Fixes: c3963bc0a0cf ("hwmon: (nct6775) Split core and platform driver")
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@baylibre.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (pmbus/tps25990) Add initial support
Jerome Brunet [Mon, 2 Dec 2024 10:28:02 +0000 (11:28 +0100)] 
hwmon: (pmbus/tps25990) Add initial support

Add initial support for the Texas Instruments TPS25990 eFuse.
This adds the basic PMBUS telemetry support for the device.

From Karol Przybylski:
  The tps25990_read_word_data function contains a block of unreachable code
  caused by the syntactic structure in the PMBUS_VIRT_READ_IIN_MAX case.

  Specifically, the return TPS25990_READ_IIN_PEAK; statement immediately
  exits the function, making the next lines unreachable.

  This patch removes the return statement, leaving the expected handling.
  Discovered in coverity: CID 1602227

Tested-by: Vaishnav Achath <vaishnav.a@ti.com>
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
[groeck: Adjust to MODULE_IMPORT_NS API change]
[karprzy7@gmail.com: Fix unreachable code in tps25990_read_word_data]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (pmbus/core) add wp module param
Jerome Brunet [Mon, 2 Dec 2024 10:28:01 +0000 (11:28 +0100)] 
hwmon: (pmbus/core) add wp module param

Add a module parameter to force the write protection mode of pmbus chips.

4 protections modes are provided to start with:
* 0: Remove the write protection
* 1: Disable all writes except to the WRITE_PROTECT, OPERATION,
     PAGE, ON_OFF_CONFIG and VOUT_COMMAND commands
* 2: Disable all writes except to the WRITE_PROTECT, OPERATION and
     PAGE commands
* 3: Disable all writes except to the WRITE_PROTECT command

Of course, if the parameter is not provided, the default write protection
status of the pmbus chips is left untouched.

Suggested-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (pmbus/core) improve handling of write protected regulators
Jerome Brunet [Mon, 2 Dec 2024 10:28:00 +0000 (11:28 +0100)] 
hwmon: (pmbus/core) improve handling of write protected regulators

Writing PMBus protected registers does succeed from the smbus perspective,
even if the write is ignored by the device and a communication fault is
raised. This fault will silently be caught and cleared by pmbus irq if one
has been registered.

This means that the regulator call may return succeed although the
operation was ignored.

With this change, the operation which are not supported will be properly
flagged as such and the regulator framework won't even try to execute them.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
[groeck: Adjust to EXPORT_SYMBOL_NS_GPL API change]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (raspberrypi) Add PM suspend/resume support
Stefan Wahren [Mon, 2 Dec 2024 11:58:32 +0000 (12:58 +0100)] 
hwmon: (raspberrypi) Add PM suspend/resume support

Add suspend/resume handler in order to stop firmware polling
during s2idle. This was just waking-up the system without a real
benefit.

Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (core) Avoid ifdef CONFIG_THERMAL in C source file
Thomas Weißschuh [Mon, 18 Nov 2024 06:15:59 +0000 (07:15 +0100)] 
hwmon: (core) Avoid ifdef CONFIG_THERMAL in C source file

Using an #ifdef in a C source files to have different definitions
of the same symbol makes the code harder to read and understand.
Furthermore it makes it harder to test compilation of the different
branches.

Replace the ifdeffery with IS_ENABLED() which is just a normal
conditional.
The resulting binary is still the same as before as the compiler
optimizes away all the unused code and definitions.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agothermal: core: Add stub for thermal_zone_device_update()
Thomas Weißschuh [Mon, 18 Nov 2024 06:15:58 +0000 (07:15 +0100)] 
thermal: core: Add stub for thermal_zone_device_update()

To simplify the !CONFIG_THERMAL case in the hwmon core,
add a !CONFIG_THERMAL stub for thermal_zone_device_update().

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Rafael J. Wysocki <rafael@kernel.org>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (chipcap2) Switch to guard() for mutext handling
Javier Carrasco [Thu, 21 Nov 2024 21:44:56 +0000 (22:44 +0100)] 
hwmon: (chipcap2) Switch to guard() for mutext handling

Switch to guard() for mutex handling to simplify the code, getting rid
of the 'ret = x, break; return ret;' construct and return the result
of the operation instead.

Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agoMAINTAINERS: Drop IIO from the title of the Chipcap 2 hwmon driver
Javier Carrasco [Thu, 21 Nov 2024 21:44:55 +0000 (22:44 +0100)] 
MAINTAINERS: Drop IIO from the title of the Chipcap 2 hwmon driver

The Chipcap 2 driver belongs to hwmon, and not to iio. When at it,
drop the sensor type description, as that is documented elsewhere.

Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (isl28022) apply coding style to module init/exit
Wolfram Sang [Wed, 20 Nov 2024 08:33:52 +0000 (09:33 +0100)] 
hwmon: (isl28022) apply coding style to module init/exit

Function declarations can be in one line. module_init|exit macros should
be tied to the function.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (isl28022) document shunt voltage channel
Wolfram Sang [Wed, 20 Nov 2024 08:33:51 +0000 (09:33 +0100)] 
hwmon: (isl28022) document shunt voltage channel

During the upstreaming process, the driver has been extended to report
the shunt voltage as well. Document that.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agohwmon: (isl28022) use proper path for DT bindings
Wolfram Sang [Wed, 20 Nov 2024 08:33:50 +0000 (09:33 +0100)] 
hwmon: (isl28022) use proper path for DT bindings

Vendor is "Renesas" not "ISL".

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
6 months agoLinux 6.13-rc2 v6.13-rc2
Linus Torvalds [Sun, 8 Dec 2024 22:03:39 +0000 (14:03 -0800)] 
Linux 6.13-rc2

6 months agoMerge tag 'kbuild-fixes-v6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/masah...
Linus Torvalds [Sun, 8 Dec 2024 20:01:06 +0000 (12:01 -0800)] 
Merge tag 'kbuild-fixes-v6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild fixes from Masahiro Yamada:

 - Fix a section mismatch warning in modpost

 - Fix Debian package build error with the O= option

* tag 'kbuild-fixes-v6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  kbuild: deb-pkg: fix build error with O=
  modpost: Add .irqentry.text to OTHER_SECTIONS

6 months agoMerge tag 'irq_urgent_for_v6.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 8 Dec 2024 19:54:04 +0000 (11:54 -0800)] 
Merge tag 'irq_urgent_for_v6.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull irq fixes from Borislav Petkov:

 - Fix a /proc/interrupts formatting regression

 - Have the BCM2836 interrupt controller enter power management states
   properly

 - Other fixlets

* tag 'irq_urgent_for_v6.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  irqchip/stm32mp-exti: CONFIG_STM32MP_EXTI should not default to y when compile-testing
  genirq/proc: Add missing space separator back
  irqchip/bcm2836: Enable SKIP_SET_WAKE and MASK_ON_SUSPEND
  irqchip/gic-v3: Fix irq_complete_ack() comment

6 months agoMerge tag 'timers_urgent_for_v6.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 8 Dec 2024 19:51:29 +0000 (11:51 -0800)] 
Merge tag 'timers_urgent_for_v6.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull timer fix from Borislav Petkov:

 - Handle the case where clocksources with small counter width can,
   in conjunction with overly long idle sleeps, falsely trigger the
   negative motion detection of clocksources

* tag 'timers_urgent_for_v6.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  clocksource: Make negative motion detection more robust

6 months agoMerge tag 'x86_urgent_for_v6.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 8 Dec 2024 19:38:56 +0000 (11:38 -0800)] 
Merge tag 'x86_urgent_for_v6.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 fixes from Borislav Petkov:

 - Have the Automatic IBRS setting check on AMD does not falsely fire in
   the guest when it has been set already on the host

 - Make sure cacheinfo structures memory is allocated to address a boot
   NULL ptr dereference on Intel Meteor Lake which has different numbers
   of subleafs in its CPUID(4) leaf

 - Take care of the GDT restoring on the kexec path too, as expected by
   the kernel

 - Make sure SMP is not disabled when IO-APIC is disabled on the kernel
   cmdline

 - Add a PGD flag _PAGE_NOPTISHADOW to instruct machinery not to
   propagate changes to the kernelmode page tables, to the user portion,
   in PTI

 - Mark Intel Lunar Lake as affected by an issue where MONITOR wakeups
   can get lost and thus user-visible delays happen

 - Make sure PKRU is properly restored with XRSTOR on AMD after a PRKU
   write of 0 (WRPKRU) which will mark PKRU in its init state and thus
   lose the actual buffer

* tag 'x86_urgent_for_v6.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/CPU/AMD: WARN when setting EFER.AUTOIBRS if and only if the WRMSR fails
  x86/cacheinfo: Delete global num_cache_leaves
  cacheinfo: Allocate memory during CPU hotplug if not done from the primary CPU
  x86/kexec: Restore GDT on return from ::preserve_context kexec
  x86/cpu/topology: Remove limit of CPUs due to disabled IO/APIC
  x86/mm: Add _PAGE_NOPTISHADOW bit to avoid updating userspace page tables
  x86/cpu: Add Lunar Lake to list of CPUs with a broken MONITOR implementation
  x86/pkeys: Ensure updated PKRU value is XRSTOR'd
  x86/pkeys: Change caller of update_pkru_in_sigframe()

6 months agoMerge tag 'mm-hotfixes-stable-2024-12-07-22-39' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Sun, 8 Dec 2024 19:26:13 +0000 (11:26 -0800)] 
Merge tag 'mm-hotfixes-stable-2024-12-07-22-39' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm

Pull misc fixes from Andrew Morton:
 "24 hotfixes.  17 are cc:stable.  15 are MM and 9 are non-MM.

  The usual bunch of singletons - please see the relevant changelogs for
  details"

* tag 'mm-hotfixes-stable-2024-12-07-22-39' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (24 commits)
  iio: magnetometer: yas530: use signed integer type for clamp limits
  sched/numa: fix memory leak due to the overwritten vma->numab_state
  mm/damon: fix order of arguments in damos_before_apply tracepoint
  lib: stackinit: hide never-taken branch from compiler
  mm/filemap: don't call folio_test_locked() without a reference in next_uptodate_folio()
  scatterlist: fix incorrect func name in kernel-doc
  mm: correct typo in MMAP_STATE() macro
  mm: respect mmap hint address when aligning for THP
  mm: memcg: declare do_memsw_account inline
  mm/codetag: swap tags when migrate pages
  ocfs2: update seq_file index in ocfs2_dlm_seq_next
  stackdepot: fix stack_depot_save_flags() in NMI context
  mm: open-code page_folio() in dump_page()
  mm: open-code PageTail in folio_flags() and const_folio_flags()
  mm: fix vrealloc()'s KASAN poisoning logic
  Revert "readahead: properly shorten readahead when falling back to do_page_cache_ra()"
  selftests/damon: add _damon_sysfs.py to TEST_FILES
  selftest: hugetlb_dio: fix test naming
  ocfs2: free inode when ocfs2_get_init_inode() fails
  nilfs2: fix potential out-of-bounds memory access in nilfs_find_entry()
  ...

6 months agokbuild: deb-pkg: fix build error with O=
Masahiro Yamada [Sun, 8 Dec 2024 07:56:45 +0000 (16:56 +0900)] 
kbuild: deb-pkg: fix build error with O=

Since commit 13b25489b6f8 ("kbuild: change working directory to external
module directory with M="), the Debian package build fails if a relative
path is specified with the O= option.

  $ make O=build bindeb-pkg
    [ snip ]
  dpkg-deb: building package 'linux-image-6.13.0-rc1' in '../linux-image-6.13.0-rc1_6.13.0-rc1-6_amd64.deb'.
  Rebuilding host programs with x86_64-linux-gnu-gcc...
  make[6]: Entering directory '/home/masahiro/linux/build'
  /home/masahiro/linux/Makefile:190: *** specified kernel directory "build" does not exist.  Stop.

This occurs because the sub_make_done flag is cleared, even though the
working directory is already in the output directory.

Passing KBUILD_OUTPUT=. resolves the issue.

Fixes: 13b25489b6f8 ("kbuild: change working directory to external module directory with M=")
Reported-by: Charlie Jenkins <charlie@rivosinc.com>
Closes: https://lore.kernel.org/all/Z1DnP-GJcfseyrM3@ghost/
Tested-by: Charlie Jenkins <charlie@rivosinc.com>
Reviewed-by: Charlie Jenkins <charlie@rivosinc.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
6 months agomodpost: Add .irqentry.text to OTHER_SECTIONS
Thomas Gleixner [Sun, 1 Dec 2024 11:17:30 +0000 (12:17 +0100)] 
modpost: Add .irqentry.text to OTHER_SECTIONS

The compiler can fully inline the actual handler function of an interrupt
entry into the .irqentry.text entry point. If such a function contains an
access which has an exception table entry, modpost complains about a
section mismatch:

  WARNING: vmlinux.o(__ex_table+0x447c): Section mismatch in reference ...

  The relocation at __ex_table+0x447c references section ".irqentry.text"
  which is not in the list of authorized sections.

Add .irqentry.text to OTHER_SECTIONS to cure the issue.

Reported-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org # needed for linux-5.4-y
Link: https://lore.kernel.org/all/20241128111844.GE10431@google.com/
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
6 months agoMerge tag '6.13-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 8 Dec 2024 01:27:25 +0000 (17:27 -0800)] 
Merge tag '6.13-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull smb client fixes from Steve French:

 - DFS fix (for race with tree disconnect and dfs cache worker)

 - Four fixes for SMB3.1.1 posix extensions:
      - improve special file support e.g. to Samba, retrieving the file
        type earlier
      - reduce roundtrips (e.g. on ls -l, in some cases)

* tag '6.13-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  smb: client: fix potential race in cifs_put_tcon()
  smb3.1.1: fix posix mounts to older servers
  fs/smb/client: cifs_prime_dcache() for SMB3 POSIX reparse points
  fs/smb/client: Implement new SMB3 POSIX type
  fs/smb/client: avoid querying SMB2_OP_QUERY_WSL_EA for SMB3 POSIX

6 months agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Sun, 8 Dec 2024 01:17:38 +0000 (17:17 -0800)] 
Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "Large number of small fixes, all in drivers"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (32 commits)
  scsi: scsi_debug: Fix hrtimer support for ndelay
  scsi: storvsc: Do not flag MAINTENANCE_IN return of SRB_STATUS_DATA_OVERRUN as an error
  scsi: ufs: core: Add missing post notify for power mode change
  scsi: sg: Fix slab-use-after-free read in sg_release()
  scsi: ufs: core: sysfs: Prevent div by zero
  scsi: qla2xxx: Update version to 10.02.09.400-k
  scsi: qla2xxx: Supported speed displayed incorrectly for VPorts
  scsi: qla2xxx: Fix NVMe and NPIV connect issue
  scsi: qla2xxx: Remove check req_sg_cnt should be equal to rsp_sg_cnt
  scsi: qla2xxx: Fix use after free on unload
  scsi: qla2xxx: Fix abort in bsg timeout
  scsi: mpi3mr: Update driver version to 8.12.0.3.50
  scsi: mpi3mr: Handling of fault code for insufficient power
  scsi: mpi3mr: Start controller indexing from 0
  scsi: mpi3mr: Fix corrupt config pages PHY state is switched in sysfs
  scsi: mpi3mr: Synchronize access to ioctl data buffer
  scsi: mpt3sas: Update driver version to 51.100.00.00
  scsi: mpt3sas: Diag-Reset when Doorbell-In-Use bit is set during driver load time
  scsi: ufs: pltfrm: Dellocate HBA during ufshcd_pltfrm_remove()
  scsi: ufs: pltfrm: Drop PM runtime reference count after ufshcd_remove()
  ...

6 months agoMerge tag 'block-6.13-20241207' of git://git.kernel.dk/linux
Linus Torvalds [Sat, 7 Dec 2024 18:07:05 +0000 (10:07 -0800)] 
Merge tag 'block-6.13-20241207' of git://git.kernel.dk/linux

Pull block fixes from Jens Axboe:

 - NVMe pull request via Keith:
      - Target fix using incorrect zero buffer (Nilay)
      - Device specifc deallocate quirk fixes (Christoph, Keith)
      - Fabrics fix for handling max command target bugs (Maurizio)
      - Cocci fix usage for kzalloc (Yu-Chen)
      - DMA size fix for host memory buffer feature (Christoph)
      - Fabrics queue cleanup fixes (Chunguang)

 - CPU hotplug ordering fixes

 - Add missing MODULE_DESCRIPTION for rnull

 - bcache error value fix

 - virtio-blk queue freeze fix

* tag 'block-6.13-20241207' of git://git.kernel.dk/linux:
  blk-mq: move cpuhp callback registering out of q->sysfs_lock
  blk-mq: register cpuhp callback after hctx is added to xarray table
  virtio-blk: don't keep queue frozen during system suspend
  nvme-tcp: simplify nvme_tcp_teardown_io_queues()
  nvme-tcp: no need to quiesce admin_q in nvme_tcp_teardown_io_queues()
  nvme-rdma: unquiesce admin_q before destroy it
  nvme-tcp: fix the memleak while create new ctrl failed
  nvme-pci: don't use dma_alloc_noncontiguous with 0 merge boundary
  nvmet: replace kmalloc + memset with kzalloc for data allocation
  nvme-fabrics: handle zero MAXCMD without closing the connection
  bcache: revert replacing IS_ERR_OR_NULL with IS_ERR again
  nvme-pci: remove two deallocate zeroes quirks
  block: rnull: add missing MODULE_DESCRIPTION
  nvme: don't apply NVME_QUIRK_DEALLOCATE_ZEROES when DSM is not supported
  nvmet: use kzalloc instead of ZERO_PAGE in nvme_execute_identify_ns_nvm()

6 months agoMerge tag 'io_uring-6.13-20241207' of git://git.kernel.dk/linux
Linus Torvalds [Sat, 7 Dec 2024 18:01:13 +0000 (10:01 -0800)] 
Merge tag 'io_uring-6.13-20241207' of git://git.kernel.dk/linux

Pull io_uring fix from Jens Axboe:
 "A single fix for a parameter type which affects 32-bit"

* tag 'io_uring-6.13-20241207' of git://git.kernel.dk/linux:
  io_uring: Change res2 parameter type in io_uring_cmd_done

6 months agoMerge tag 'ubifs-for-linus-6.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 7 Dec 2024 17:57:38 +0000 (09:57 -0800)] 
Merge tag 'ubifs-for-linus-6.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs

Pull jffs2 fix from Richard Weinberger:

 - Fixup rtime compressor bounds checking

* tag 'ubifs-for-linus-6.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs:
  jffs2: Fix rtime decompressor

6 months agoMerge tag 'bpf-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Linus Torvalds [Fri, 6 Dec 2024 23:07:48 +0000 (15:07 -0800)] 
Merge tag 'bpf-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf

Pull bpf fixes from Daniel Borkmann::

 - Fix several issues for BPF LPM trie map which were found by syzbot
   and during addition of new test cases (Hou Tao)

 - Fix a missing process_iter_arg register type check in the BPF
   verifier (Kumar Kartikeya Dwivedi, Tao Lyu)

 - Fix several correctness gaps in the BPF verifier when interacting
   with the BPF stack without CAP_PERFMON (Kumar Kartikeya Dwivedi,
   Eduard Zingerman, Tao Lyu)

 - Fix OOB BPF map writes when deleting elements for the case of xsk map
   as well as devmap (Maciej Fijalkowski)

 - Fix xsk sockets to always clear DMA mapping information when
   unmapping the pool (Larysa Zaremba)

 - Fix sk_mem_uncharge logic in tcp_bpf_sendmsg to only uncharge after
   sent bytes have been finalized (Zijian Zhang)

 - Fix BPF sockmap with vsocks which was missing a queue check in poll
   and sockmap cleanup on close (Michal Luczaj)

 - Fix tools infra to override makefile ARCH variable if defined but
   empty, which addresses cross-building tools. (Björn Töpel)

 - Fix two resolve_btfids build warnings on unresolved bpf_lsm symbols
   (Thomas Weißschuh)

 - Fix a NULL pointer dereference in bpftool (Amir Mohammadi)

 - Fix BPF selftests to check for CONFIG_PREEMPTION instead of
   CONFIG_PREEMPT (Sebastian Andrzej Siewior)

* tag 'bpf-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf: (31 commits)
  selftests/bpf: Add more test cases for LPM trie
  selftests/bpf: Move test_lpm_map.c to map_tests
  bpf: Use raw_spinlock_t for LPM trie
  bpf: Switch to bpf mem allocator for LPM trie
  bpf: Fix exact match conditions in trie_get_next_key()
  bpf: Handle in-place update for full LPM trie correctly
  bpf: Handle BPF_EXIST and BPF_NOEXIST for LPM trie
  bpf: Remove unnecessary kfree(im_node) in lpm_trie_update_elem
  bpf: Remove unnecessary check when updating LPM trie
  selftests/bpf: Add test for narrow spill into 64-bit spilled scalar
  selftests/bpf: Add test for reading from STACK_INVALID slots
  selftests/bpf: Introduce __caps_unpriv annotation for tests
  bpf: Fix narrow scalar spill onto 64-bit spilled scalar slots
  bpf: Don't mark STACK_INVALID as STACK_MISC in mark_stack_slot_misc
  samples/bpf: Remove unnecessary -I flags from libbpf EXTRA_CFLAGS
  bpf: Zero index arg error string for dynptr and iter
  selftests/bpf: Add tests for iter arg check
  bpf: Ensure reg is PTR_TO_STACK in process_iter_arg
  tools: Override makefile ARCH variable if defined, but empty
  selftests/bpf: Add apply_bytes test to test_txmsg_redir_wait_sndmem in test_sockmap
  ...

6 months agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Fri, 6 Dec 2024 21:47:55 +0000 (13:47 -0800)] 
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux

Pull arm64 fixes from Catalin Marinas:
 "Nothing major, some left-overs from the recent merging window (MTE,
  coco) and some newly found issues like the ptrace() ones.

   - MTE/hugetlbfs:

      - Set VM_MTE_ALLOWED in the arch code and remove it from the core
        code for hugetlbfs mappings

      - Fix copy_highpage() warning when the source is a huge page but
        not MTE tagged, taking the wrong small page path

   - drivers/virt/coco:

      - Add the pKVM and Arm CCA drivers under the arm64 maintainership

      - Fix the pkvm driver to fall back to ioremap() (and warn) if the
        MMIO_GUARD hypercall fails

      - Keep the Arm CCA driver default 'n' rather than 'm'

   - A series of fixes for the arm64 ptrace() implementation,
     potentially leading to the kernel consuming uninitialised stack
     variables when PTRACE_SETREGSET is invoked with a length of 0

   - Fix zone_dma_limit calculation when RAM starts below 4GB and
     ZONE_DMA is capped to this limit

   - Fix early boot warning with CONFIG_DEBUG_VIRTUAL=y triggered by a
     call to page_to_phys() (from patch_map()) which checks pfn_valid()
     before vmemmap has been set up

   - Do not clobber bits 15:8 of the ASID used for TTBR1_EL1 and TLBI
     ops when the kernel assumes 8-bit ASIDs but running under a
     hypervisor on a system that implements 16-bit ASIDs (found running
     Linux under Parallels on Apple M4)

   - ACPI/IORT: Add PMCG platform information for HiSilicon HIP09A as it
     is using the same SMMU PMCG as HIP09 and suffers from the same
     errata

   - Add GCS to cpucap_is_possible(), missed in the recent merge"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: ptrace: fix partial SETREGSET for NT_ARM_GCS
  arm64: ptrace: fix partial SETREGSET for NT_ARM_POE
  arm64: ptrace: fix partial SETREGSET for NT_ARM_FPMR
  arm64: ptrace: fix partial SETREGSET for NT_ARM_TAGGED_ADDR_CTRL
  arm64: cpufeature: Add GCS to cpucap_is_possible()
  coco: virt: arm64: Do not enable cca guest driver by default
  arm64: mte: Fix copy_highpage() warning on hugetlb folios
  arm64: Ensure bits ASID[15:8] are masked out when the kernel uses 8-bit ASIDs
  ACPI/IORT: Add PMCG platform information for HiSilicon HIP09A
  MAINTAINERS: Add CCA and pKVM CoCO guest support to the ARM64 entry
  drivers/virt: pkvm: Don't fail ioremap() call if MMIO_GUARD fails
  arm64: patching: avoid early page_to_phys()
  arm64: mm: Fix zone_dma_limit calculation
  arm64: mte: set VM_MTE_ALLOWED for hugetlbfs at correct place

6 months agoMerge tag 'fixes-2024-12-06' of git://git.kernel.org/pub/scm/linux/kernel/git/rppt...
Linus Torvalds [Fri, 6 Dec 2024 21:42:03 +0000 (13:42 -0800)] 
Merge tag 'fixes-2024-12-06' of git://git.kernel.org/pub/scm/linux/kernel/git/rppt/memblock

Pull memblock fixes from Mike Rapoport:
 "Restore check for node validity in arch_numa.

  The rework of NUMA initialization in arch_numa dropped a check that
  refused to accept configurations with invalid node IDs.

  Restore that check to ensure that when firmware passes invalid nodes,
  such configuration is rejected and kernel gracefully falls back to
  dummy NUMA"

* tag 'fixes-2024-12-06' of git://git.kernel.org/pub/scm/linux/kernel/git/rppt/memblock:
  arch_numa: Restore nid checks before registering a memblock with a node
  memblock: allow zero threshold in validate_numa_converage()

6 months agoMerge tag 'drm-fixes-2024-12-06' of https://gitlab.freedesktop.org/drm/kernel
Linus Torvalds [Fri, 6 Dec 2024 21:16:41 +0000 (13:16 -0800)] 
Merge tag 'drm-fixes-2024-12-06' of https://gitlab.freedesktop.org/drm/kernel

Pull more drm fixes from Simona Vetter:
 "Due to mailing list unreliability we missed the amdgpu pull, hence
  part two with that now included:

   - amdgu: mostly display fixes + jpeg vcn 1.0, sriov, dcn4.0 resume
     fixes

   - amdkfd fixes"

* tag 'drm-fixes-2024-12-06' of https://gitlab.freedesktop.org/drm/kernel:
  drm/amdgpu: rework resume handling for display (v2)
  drm/amd/pm: fix and simplify workload handling
  Revert "drm/amd/pm: correct the workload setting"
  drm/amdgpu: fix sriov reinit late orders
  drm/amdgpu: Fix ISP hw init issue
  drm/amd/display: Add hblank borrowing support
  drm/amd/display: Limit VTotal range to max hw cap minus fp
  drm/amd/display: Correct prefetch calculation
  drm/amd/display: Add option to retrieve detile buffer size
  drm/amd/display: Add a left edge pixel if in YCbCr422 or YCbCr420 and odm
  drm/amdkfd: hard-code cacheline for gc943,gc944
  drm/amdkfd: add MEC version that supports no PCIe atomics for GFX12
  drm/amd/display: Fix programming backlight on OLED panels
  drm/amd: Sanity check the ACPI EDID
  drm/amdgpu/hdp7.0: do a posting read when flushing HDP
  drm/amdgpu/hdp6.0: do a posting read when flushing HDP
  drm/amdgpu/hdp5.2: do a posting read when flushing HDP
  drm/amdgpu/hdp5.0: do a posting read when flushing HDP
  drm/amdgpu/hdp4.0: do a posting read when flushing HDP
  drm/amdgpu/jpeg1.0: fix idle work handler

6 months agoMerge tag 'amd-drm-fixes-6.13-2024-12-04' of https://gitlab.freedesktop.org/agd5f...
Simona Vetter [Fri, 6 Dec 2024 20:54:04 +0000 (21:54 +0100)] 
Merge tag 'amd-drm-fixes-6.13-2024-12-04' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-6.13-2024-12-04:

amdgpu:
- Jpeg work handler fix for VCN 1.0
- HDP flush fixes
- ACPI EDID sanity check
- OLED panel backlight fix
- DC YCbCr fix
- DC Detile buffer size debugging
- DC prefetch calculation fix
- DC VTotal handling fix
- DC HBlank fix
- ISP fix
- SR-IOV fix
- Workload profile fixes
- DCN 4.0.1 resume fix

amdkfd:
- GC 12.x fix
- GC 9.4.x fix

Signed-off-by: Simona Vetter <simona.vetter@ffwll.ch>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241206190452.2571042-1-alexander.deucher@amd.com
6 months agoMerge tag 'drm-fixes-2024-12-07' of https://gitlab.freedesktop.org/drm/kernel
Linus Torvalds [Fri, 6 Dec 2024 19:52:15 +0000 (11:52 -0800)] 
Merge tag 'drm-fixes-2024-12-07' of https://gitlab.freedesktop.org/drm/kernel

Pull drm fixes from Dave Airlie:
 "Pretty quiet week which is probably expected after US holidays, the
  dma-fence and displayport MST message handling fixes make up the bulk
  of this, along with a couple of minor xe and other driver fixes.

  dma-fence:
   - Fix reference leak on fence-merge failure path
   - Simplify fence merging with kernel's sort()
   - Fix dma_fence_array_signaled() to ensure forward progress

  dp_mst:
   - Fix MST sideband message body length check
   - Fix a bunch of locking/state handling with DP MST msgs

  sti:
   - Add __iomem for mixer_dbg_mxn()'s parameter

  xe:
   - Missing init value and 64-bit write-order check
   - Fix a memory allocation issue causing lockdep violation

  v3d:
   - Performance counter fix"

* tag 'drm-fixes-2024-12-07' of https://gitlab.freedesktop.org/drm/kernel:
  drm/v3d: Enable Performance Counters before clearing them
  drm/dp_mst: Use reset_msg_rx_state() instead of open coding it
  drm/dp_mst: Reset message rx state after OOM in drm_dp_mst_handle_up_req()
  drm/dp_mst: Ensure mst_primary pointer is valid in drm_dp_mst_handle_up_req()
  drm/dp_mst: Fix down request message timeout handling
  drm/dp_mst: Simplify error path in drm_dp_mst_handle_down_rep()
  drm/dp_mst: Verify request type in the corresponding down message reply
  drm/dp_mst: Fix resetting msg rx state after topology removal
  drm/xe: Move the coredump registration to the worker thread
  drm/xe/guc: Fix missing init value and add register order check
  drm/sti: Add __iomem for mixer_dbg_mxn's parameter
  drm/dp_mst: Fix MST sideband message body length check
  dma-buf: fix dma_fence_array_signaled v4
  dma-fence: Use kernel's sort for merging fences
  dma-fence: Fix reference leak on fence merge failure path

6 months agoMerge tag 'sound-6.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Fri, 6 Dec 2024 19:46:39 +0000 (11:46 -0800)] 
Merge tag 'sound-6.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "A collection of small fixes that have been gathered in the week.

   - Fix the missing XRUN handling in USB-audio low latency mode

   - Fix regression by the previous USB-audio hadening change

   - Clean up old SH sound driver to use the standard helpers

   - A few further fixes for MIDI 2.0 UMP handling

   - Various HD-audio and USB-audio quirks

   - Fix jack handling at PM on ASoC Intel AVS

   - Misc small fixes for ASoC SOF and Mediatek"

* tag 'sound-6.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda/realtek: Fix spelling mistake "Firelfy" -> "Firefly"
  ASoC: mediatek: mt8188-mt6359: Remove hardcoded dmic codec
  ALSA: hda/realtek: fix micmute LEDs don't work on HP Laptops
  ALSA: usb-audio: Add extra PID for RME Digiface USB
  ALSA: usb-audio: Fix a DMA to stack memory bug
  ASoC: SOF: ipc3-topology: fix resource leaks in sof_ipc3_widget_setup_comp_dai()
  ALSA: hda/realtek: Add support for Samsung Galaxy Book3 360 (NP730QFG)
  ASoC: Intel: avs: da7219: Remove suspend_pre() and resume_post()
  ALSA: hda/tas2781: Fix error code tas2781_read_acpi()
  ALSA: hda/realtek: Enable mute and micmute LED on HP ProBook 430 G8
  ALSA: usb-audio: add mixer mapping for Corsair HS80
  ALSA: ump: Shut up truncated string warning
  ALSA: sh: Use standard helper for buffer accesses
  ALSA: usb-audio: Notify xrun for low-latency mode
  ALSA: hda/conexant: fix Z60MR100 startup pop issue
  ALSA: ump: Update legacy substream names upon FB info update
  ALSA: ump: Indicate the inactive group in legacy substream names
  ALSA: ump: Don't open legacy substream for an inactive group
  ALSA: seq: ump: Fix seq port updates per FB info notify

6 months agoMerge tag 'regmap-fix-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 6 Dec 2024 19:43:22 +0000 (11:43 -0800)] 
Merge tag 'regmap-fix-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap

Pull regmap fixes from Mark Brown:
 "A couple of small fixes, fixing an incorrect format specifier in a log
  message and adding missing cleanup of the devres data used to support
  dev_get_regmap() when a device is unregistered"

* tag 'regmap-fix-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:
  regmap: detach regmap from dev on regmap_exit
  regmap: Use correct format specifier for logging range errors

6 months agoMerge tag 'spi-fix-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brooni...
Linus Torvalds [Fri, 6 Dec 2024 19:36:48 +0000 (11:36 -0800)] 
Merge tag 'spi-fix-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi

Pull spi fixes from Mark Brown:
 "A few small driver specific fixes and device ID updates for SPI.

  The Apple change flags the driver as being compatible with the core's
  GPIO chip select support, fixing support for some systems"

* tag 'spi-fix-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi: omap2-mcspi: Fix the IS_ERR() bug for devm_clk_get_optional_enabled()
  spi: intel: Add Panther Lake SPI controller support
  spi: apple: Set use_gpio_descriptors to true
  spi: mpc52xx: Add cancel_work_sync before module remove

6 months agoMerge tag 'mmc-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc
Linus Torvalds [Fri, 6 Dec 2024 19:27:10 +0000 (11:27 -0800)] 
Merge tag 'mmc-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc

Pull MMC fixes from Ulf Hansson:
 "Core:
   - Further prevent card detect during shutdown

  Host drivers:
   - sdhci-pci: Add DMI quirk for missing CD GPIO on Vexia Edu Atla 10
     tablet"

* tag 'mmc-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc:
  mmc: core: Further prevent card detect during shutdown
  mmc: sdhci-pci: Add DMI quirk for missing CD GPIO on Vexia Edu Atla 10 tablet

6 months agoMerge tag 'pmdomain-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh...
Linus Torvalds [Fri, 6 Dec 2024 19:24:00 +0000 (11:24 -0800)] 
Merge tag 'pmdomain-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/linux-pm

Pull pmdomain fixes from Ulf Hansson:
 "Core:
   - Fix a couple of memory-leaks during genpd init/remove

  Providers:
   - imx: Adjust delay for gpcv2 to fix power up handshake
   - mediatek: Fix DT bindings by adding another nested power-domain
     layer"

* tag 'pmdomain-v6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/linux-pm:
  pmdomain: imx: gpcv2: Adjust delay after power up handshake
  pmdomain: core: Fix error path in pm_genpd_init() when ida alloc fails
  pmdomain: core: Add missing put_device()
  dt-bindings: power: mediatek: Add another nested power-domain layer

6 months agox86/CPU/AMD: WARN when setting EFER.AUTOIBRS if and only if the WRMSR fails
Sean Christopherson [Fri, 6 Dec 2024 16:20:06 +0000 (08:20 -0800)] 
x86/CPU/AMD: WARN when setting EFER.AUTOIBRS if and only if the WRMSR fails

When ensuring EFER.AUTOIBRS is set, WARN only on a negative return code
from msr_set_bit(), as '1' is used to indicate the WRMSR was successful
('0' indicates the MSR bit was already set).

Fixes: 8cc68c9c9e92 ("x86/CPU/AMD: Make sure EFER[AIBRSE] is set")
Reported-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/Z1MkNofJjt7Oq0G6@google.com
Closes: https://lore.kernel.org/all/20241205220604.GA2054199@thelio-3990X
6 months agoMerge branch 'fixes-for-lpm-trie'
Alexei Starovoitov [Fri, 6 Dec 2024 17:14:26 +0000 (09:14 -0800)] 
Merge branch 'fixes-for-lpm-trie'

Hou Tao says:

====================
This patch set fixes several issues for LPM trie. These issues were
found during adding new test cases or were reported by syzbot.

The patch set is structured as follows:

Patch #1~#2 are clean-ups for lpm_trie_update_elem().
Patch #3 handles BPF_EXIST and BPF_NOEXIST correctly for LPM trie.
Patch #4 fixes the accounting of n_entries when doing in-place update.
Patch #5 fixes the exact match condition in trie_get_next_key() and it
may skip keys when the passed key is not found in the map.
Patch #6~#7 switch from kmalloc() to bpf memory allocator for LPM trie
to fix several lock order warnings reported by syzbot. It also enables
raw_spinlock_t for LPM trie again. After these changes, the LPM trie will
be closer to being usable in any context (though the reentrance check of
trie->lock is still missing, but it is on my todo list).
Patch #8: move test_lpm_map to map_tests to make it run regularly.
Patch #9: add test cases for the issues fixed by patch #3~#5.

Please see individual patches for more details. Comments are always
welcome.

Change Log:
v3:
  * patch #2: remove the unnecessary NULL-init for im_node
  * patch #6: alloc the leaf node before disabling IRQ to low
    the possibility of -ENOMEM when leaf_size is large; Free
    these nodes outside the trie lock (Suggested by Alexei)
  * collect review and ack tags (Thanks for Toke & Daniel)

v2: https://lore.kernel.org/bpf/20241127004641.1118269-1-houtao@huaweicloud.com/
  * collect review tags (Thanks for Toke)
  * drop "Add bpf_mem_cache_is_mergeable() helper" patch
  * patch #3~#4: add fix tag
  * patch #4: rename the helper to trie_check_add_elem() and increase
    n_entries in it.
  * patch #6: use one bpf mem allocator and update commit message to
    clarify that using bpf mem allocator is more appropriate.
  * patch #7: update commit message to add the possible max running time
    for update operation.
  * patch #9: update commit message to specify the purpose of these test
    cases.

v1: https://lore.kernel.org/bpf/20241118010808.2243555-1-houtao@huaweicloud.com/
====================

Link: https://lore.kernel.org/all/20241206110622.1161752-1-houtao@huaweicloud.com/
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agoselftests/bpf: Add more test cases for LPM trie
Hou Tao [Fri, 6 Dec 2024 11:06:22 +0000 (19:06 +0800)] 
selftests/bpf: Add more test cases for LPM trie

Add more test cases for LPM trie in test_maps:

1) test_lpm_trie_update_flags
It constructs various use cases for BPF_EXIST and BPF_NOEXIST and check
whether the return value of update operation is expected.

2) test_lpm_trie_update_full_maps
It tests the update operations on a full LPM trie map. Adding new node
will fail and overwriting the value of existed node will succeed.

3) test_lpm_trie_iterate_strs and test_lpm_trie_iterate_ints
There two test cases test whether the iteration through get_next_key is
sorted and expected. These two test cases delete the minimal key after
each iteration and check whether next iteration returns the second
minimal key. The only difference between these two test cases is the
former one saves strings in the LPM trie and the latter saves integers.
Without the fix of get_next_key, these two cases will fail as shown
below:
  test_lpm_trie_iterate_strs(1091):FAIL:iterate #2 got abc exp abS
  test_lpm_trie_iterate_ints(1142):FAIL:iterate #1 got 0x2 exp 0x1

Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20241206110622.1161752-10-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agoselftests/bpf: Move test_lpm_map.c to map_tests
Hou Tao [Fri, 6 Dec 2024 11:06:21 +0000 (19:06 +0800)] 
selftests/bpf: Move test_lpm_map.c to map_tests

Move test_lpm_map.c to map_tests/ to include LPM trie test cases in
regular test_maps run. Most code remains unchanged, including the use of
assert(). Only reduce n_lookups from 64K to 512, which decreases
test_lpm_map runtime from 37s to 0.7s.

Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20241206110622.1161752-9-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agobpf: Use raw_spinlock_t for LPM trie
Hou Tao [Fri, 6 Dec 2024 11:06:20 +0000 (19:06 +0800)] 
bpf: Use raw_spinlock_t for LPM trie

After switching from kmalloc() to the bpf memory allocator, there will be
no blocking operation during the update of LPM trie. Therefore, change
trie->lock from spinlock_t to raw_spinlock_t to make LPM trie usable in
atomic context, even on RT kernels.

The max value of prefixlen is 2048. Therefore, update or deletion
operations will find the target after at most 2048 comparisons.
Constructing a test case which updates an element after 2048 comparisons
under a 8 CPU VM, and the average time and the maximal time for such
update operation is about 210us and 900us.

Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20241206110622.1161752-8-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agobpf: Switch to bpf mem allocator for LPM trie
Hou Tao [Fri, 6 Dec 2024 11:06:19 +0000 (19:06 +0800)] 
bpf: Switch to bpf mem allocator for LPM trie

Multiple syzbot warnings have been reported. These warnings are mainly
about the lock order between trie->lock and kmalloc()'s internal lock.
See report [1] as an example:

======================================================
WARNING: possible circular locking dependency detected
6.10.0-rc7-syzkaller-00003-g4376e966ecb7 #0 Not tainted
------------------------------------------------------
syz.3.2069/15008 is trying to acquire lock:
ffff88801544e6d8 (&n->list_lock){-.-.}-{2:2}, at: get_partial_node ...

but task is already holding lock:
ffff88802dcc89f8 (&trie->lock){-.-.}-{2:2}, at: trie_update_elem ...

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (&trie->lock){-.-.}-{2:2}:
       __raw_spin_lock_irqsave
       _raw_spin_lock_irqsave+0x3a/0x60
       trie_delete_elem+0xb0/0x820
       ___bpf_prog_run+0x3e51/0xabd0
       __bpf_prog_run32+0xc1/0x100
       bpf_dispatcher_nop_func
       ......
       bpf_trace_run2+0x231/0x590
       __bpf_trace_contention_end+0xca/0x110
       trace_contention_end.constprop.0+0xea/0x170
       __pv_queued_spin_lock_slowpath+0x28e/0xcc0
       pv_queued_spin_lock_slowpath
       queued_spin_lock_slowpath
       queued_spin_lock
       do_raw_spin_lock+0x210/0x2c0
       __raw_spin_lock_irqsave
       _raw_spin_lock_irqsave+0x42/0x60
       __put_partials+0xc3/0x170
       qlink_free
       qlist_free_all+0x4e/0x140
       kasan_quarantine_reduce+0x192/0x1e0
       __kasan_slab_alloc+0x69/0x90
       kasan_slab_alloc
       slab_post_alloc_hook
       slab_alloc_node
       kmem_cache_alloc_node_noprof+0x153/0x310
       __alloc_skb+0x2b1/0x380
       ......

-> #0 (&n->list_lock){-.-.}-{2:2}:
       check_prev_add
       check_prevs_add
       validate_chain
       __lock_acquire+0x2478/0x3b30
       lock_acquire
       lock_acquire+0x1b1/0x560
       __raw_spin_lock_irqsave
       _raw_spin_lock_irqsave+0x3a/0x60
       get_partial_node.part.0+0x20/0x350
       get_partial_node
       get_partial
       ___slab_alloc+0x65b/0x1870
       __slab_alloc.constprop.0+0x56/0xb0
       __slab_alloc_node
       slab_alloc_node
       __do_kmalloc_node
       __kmalloc_node_noprof+0x35c/0x440
       kmalloc_node_noprof
       bpf_map_kmalloc_node+0x98/0x4a0
       lpm_trie_node_alloc
       trie_update_elem+0x1ef/0xe00
       bpf_map_update_value+0x2c1/0x6c0
       map_update_elem+0x623/0x910
       __sys_bpf+0x90c/0x49a0
       ...

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&trie->lock);
                               lock(&n->list_lock);
                               lock(&trie->lock);
  lock(&n->list_lock);

 *** DEADLOCK ***

[1]: https://syzkaller.appspot.com/bug?extid=9045c0a3d5a7f1b119f7

A bpf program attached to trace_contention_end() triggers after
acquiring &n->list_lock. The program invokes trie_delete_elem(), which
then acquires trie->lock. However, it is possible that another
process is invoking trie_update_elem(). trie_update_elem() will acquire
trie->lock first, then invoke kmalloc_node(). kmalloc_node() may invoke
get_partial_node() and try to acquire &n->list_lock (not necessarily the
same lock object). Therefore, lockdep warns about the circular locking
dependency.

Invoking kmalloc() before acquiring trie->lock could fix the warning.
However, since BPF programs call be invoked from any context (e.g.,
through kprobe/tracepoint/fentry), there may still be lock ordering
problems for internal locks in kmalloc() or trie->lock itself.

To eliminate these potential lock ordering problems with kmalloc()'s
internal locks, replacing kmalloc()/kfree()/kfree_rcu() with equivalent
BPF memory allocator APIs that can be invoked in any context. The lock
ordering problems with trie->lock (e.g., reentrance) will be handled
separately.

Three aspects of this change require explanation:

1. Intermediate and leaf nodes are allocated from the same allocator.
Since the value size of LPM trie is usually small, using a single
alocator reduces the memory overhead of the BPF memory allocator.

2. Leaf nodes are allocated before disabling IRQs. This handles cases
where leaf_size is large (e.g., > 4KB - 8) and updates require
intermediate node allocation. If leaf nodes were allocated in
IRQ-disabled region, the free objects in BPF memory allocator would not
be refilled timely and the intermediate node allocation may fail.

3. Paired migrate_{disable|enable}() calls for node alloc and free. The
BPF memory allocator uses per-CPU struct internally, these paired calls
are necessary to guarantee correctness.

Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20241206110622.1161752-7-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agobpf: Fix exact match conditions in trie_get_next_key()
Hou Tao [Fri, 6 Dec 2024 11:06:18 +0000 (19:06 +0800)] 
bpf: Fix exact match conditions in trie_get_next_key()

trie_get_next_key() uses node->prefixlen == key->prefixlen to identify
an exact match, However, it is incorrect because when the target key
doesn't fully match the found node (e.g., node->prefixlen != matchlen),
these two nodes may also have the same prefixlen. It will return
expected result when the passed key exist in the trie. However when a
recently-deleted key or nonexistent key is passed to
trie_get_next_key(), it may skip keys and return incorrect result.

Fix it by using node->prefixlen == matchlen to identify exact matches.
When the condition is true after the search, it also implies
node->prefixlen equals key->prefixlen, otherwise, the search would
return NULL instead.

Fixes: b471f2f1de8b ("bpf: implement MAP_GET_NEXT_KEY command for LPM_TRIE map")
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20241206110622.1161752-6-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agobpf: Handle in-place update for full LPM trie correctly
Hou Tao [Fri, 6 Dec 2024 11:06:17 +0000 (19:06 +0800)] 
bpf: Handle in-place update for full LPM trie correctly

When a LPM trie is full, in-place updates of existing elements
incorrectly return -ENOSPC.

Fix this by deferring the check of trie->n_entries. For new insertions,
n_entries must not exceed max_entries. However, in-place updates are
allowed even when the trie is full.

Fixes: b95a5c4db09b ("bpf: add a longest prefix match trie map implementation")
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20241206110622.1161752-5-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agobpf: Handle BPF_EXIST and BPF_NOEXIST for LPM trie
Hou Tao [Fri, 6 Dec 2024 11:06:16 +0000 (19:06 +0800)] 
bpf: Handle BPF_EXIST and BPF_NOEXIST for LPM trie

Add the currently missing handling for the BPF_EXIST and BPF_NOEXIST
flags. These flags can be specified by users and are relevant since LPM
trie supports exact matches during update.

Fixes: b95a5c4db09b ("bpf: add a longest prefix match trie map implementation")
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20241206110622.1161752-4-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agobpf: Remove unnecessary kfree(im_node) in lpm_trie_update_elem
Hou Tao [Fri, 6 Dec 2024 11:06:15 +0000 (19:06 +0800)] 
bpf: Remove unnecessary kfree(im_node) in lpm_trie_update_elem

There is no need to call kfree(im_node) when updating element fails,
because im_node must be NULL. Remove the unnecessary kfree() for
im_node.

Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20241206110622.1161752-3-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agobpf: Remove unnecessary check when updating LPM trie
Hou Tao [Fri, 6 Dec 2024 11:06:14 +0000 (19:06 +0800)] 
bpf: Remove unnecessary check when updating LPM trie

When "node->prefixlen == matchlen" is true, it means that the node is
fully matched. If "node->prefixlen == key->prefixlen" is false, it means
the prefix length of key is greater than the prefix length of node,
otherwise, matchlen will not be equal with node->prefixlen. However, it
also implies that the prefix length of node must be less than
max_prefixlen.

Therefore, "node->prefixlen == trie->max_prefixlen" will always be false
when the check of "node->prefixlen == key->prefixlen" returns false.
Remove this unnecessary comparison.

Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20241206110622.1161752-2-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
6 months agoblk-mq: move cpuhp callback registering out of q->sysfs_lock
Ming Lei [Fri, 6 Dec 2024 11:16:07 +0000 (19:16 +0800)] 
blk-mq: move cpuhp callback registering out of q->sysfs_lock

Registering and unregistering cpuhp callback requires global cpu hotplug lock,
which is used everywhere. Meantime q->sysfs_lock is used in block layer
almost everywhere.

It is easy to trigger lockdep warning[1] by connecting the two locks.

Fix the warning by moving blk-mq's cpuhp callback registering out of
q->sysfs_lock. Add one dedicated global lock for covering registering &
unregistering hctx's cpuhp, and it is safe to do so because hctx is
guaranteed to be live if our request_queue is live.

[1] https://lore.kernel.org/lkml/Z04pz3AlvI4o0Mr8@agluck-desk3/

Cc: Reinette Chatre <reinette.chatre@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Peter Newman <peternewman@google.com>
Cc: Babu Moger <babu.moger@amd.com>
Reported-by: Luck Tony <tony.luck@intel.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Tony Luck <tony.luck@intel.com>
Link: https://lore.kernel.org/r/20241206111611.978870-3-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
6 months agoblk-mq: register cpuhp callback after hctx is added to xarray table
Ming Lei [Fri, 6 Dec 2024 11:16:06 +0000 (19:16 +0800)] 
blk-mq: register cpuhp callback after hctx is added to xarray table

We need to retrieve 'hctx' from xarray table in the cpuhp callback, so the
callback should be registered after this 'hctx' is added to xarray table.

Cc: Reinette Chatre <reinette.chatre@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Peter Newman <peternewman@google.com>
Cc: Babu Moger <babu.moger@amd.com>
Cc: Luck Tony <tony.luck@intel.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Tony Luck <tony.luck@intel.com>
Link: https://lore.kernel.org/r/20241206111611.978870-2-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
6 months agosmb: client: fix potential race in cifs_put_tcon()
Paulo Alcantara [Fri, 6 Dec 2024 14:49:07 +0000 (11:49 -0300)] 
smb: client: fix potential race in cifs_put_tcon()

dfs_cache_refresh() delayed worker could race with cifs_put_tcon(), so
make sure to call list_replace_init() on @tcon->dfs_ses_list after
kworker is cancelled or finished.

Fixes: 4f42a8b54b5c ("smb: client: fix DFS interlink failover")
Signed-off-by: Paulo Alcantara (Red Hat) <pc@manguebit.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
6 months agosmb3.1.1: fix posix mounts to older servers
Steve French [Wed, 4 Dec 2024 23:46:00 +0000 (17:46 -0600)] 
smb3.1.1: fix posix mounts to older servers

Some servers which implement the SMB3.1.1 POSIX extensions did not
set the file type in the mode in the infolevel 100 response.
With the recent changes for checking the file type via the mode field,
this can cause the root directory to be reported incorrectly and
mounts (e.g. to ksmbd) to fail.

Fixes: 6a832bc8bbb2 ("fs/smb/client: Implement new SMB3 POSIX type")
Cc: stable@vger.kernel.org
Acked-by: Paulo Alcantara (Red Hat) <pc@manguebit.com>
Cc: Ralph Boehme <slow@samba.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
6 months agox86/cacheinfo: Delete global num_cache_leaves
Ricardo Neri [Thu, 28 Nov 2024 00:22:47 +0000 (16:22 -0800)] 
x86/cacheinfo: Delete global num_cache_leaves

Linux remembers cpu_cachinfo::num_leaves per CPU, but x86 initializes all
CPUs from the same global "num_cache_leaves".

This is erroneous on systems such as Meteor Lake, where each CPU has a
distinct num_leaves value. Delete the global "num_cache_leaves" and
initialize num_leaves on each CPU.

init_cache_level() no longer needs to set num_leaves. Also, it never had to
set num_levels as it is unnecessary in x86. Keep checking for zero cache
leaves. Such condition indicates a bug.

  [ bp: Cleanup. ]

Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: stable@vger.kernel.org # 6.3+
Link: https://lore.kernel.org/r/20241128002247.26726-3-ricardo.neri-calderon@linux.intel.com
6 months agocacheinfo: Allocate memory during CPU hotplug if not done from the primary CPU
Ricardo Neri [Thu, 28 Nov 2024 00:22:46 +0000 (16:22 -0800)] 
cacheinfo: Allocate memory during CPU hotplug if not done from the primary CPU

Commit

  5944ce092b97 ("arch_topology: Build cacheinfo from primary CPU")

adds functionality that architectures can use to optionally allocate and
build cacheinfo early during boot. Commit

  6539cffa9495 ("cacheinfo: Add arch specific early level initializer")

lets secondary CPUs correct (and reallocate memory) cacheinfo data if
needed.

If the early build functionality is not used and cacheinfo does not need
correction, memory for cacheinfo is never allocated. x86 does not use
the early build functionality. Consequently, during the cacheinfo CPU
hotplug callback, last_level_cache_is_valid() attempts to dereference
a NULL pointer:

  BUG: kernel NULL pointer dereference, address: 0000000000000100
  #PF: supervisor read access in kernel mode
  #PF: error_code(0x0000) - not present page
  PGD 0 P4D 0
  Oops: 0000 [#1] PREEPMT SMP NOPTI
  CPU: 0 PID 19 Comm: cpuhp/0 Not tainted 6.4.0-rc2 #1
  RIP: 0010: last_level_cache_is_valid+0x95/0xe0a

Allocate memory for cacheinfo during the cacheinfo CPU hotplug callback
if not done earlier.

Moreover, before determining the validity of the last-level cache info,
ensure that it has been allocated. Simply checking for non-zero
cache_leaves() is not sufficient, as some architectures (e.g., Intel
processors) have non-zero cache_leaves() before allocation.

Dereferencing NULL cacheinfo can occur in update_per_cpu_data_slice_size().
This function iterates over all online CPUs. However, a CPU may have come
online recently, but its cacheinfo may not have been allocated yet.

While here, remove an unnecessary indentation in allocate_cache_info().

  [ bp: Massage. ]

Fixes: 6539cffa9495 ("cacheinfo: Add arch specific early level initializer")
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Radu Rendec <rrendec@redhat.com>
Reviewed-by: Nikolay Borisov <nik.borisov@suse.com>
Reviewed-by: Andreas Herrmann <aherrmann@suse.de>
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Cc: stable@vger.kernel.org # 6.3+
Link: https://lore.kernel.org/r/20241128002247.26726-2-ricardo.neri-calderon@linux.intel.com
6 months agox86/kexec: Restore GDT on return from ::preserve_context kexec
David Woodhouse [Thu, 5 Dec 2024 15:05:07 +0000 (15:05 +0000)] 
x86/kexec: Restore GDT on return from ::preserve_context kexec

The restore_processor_state() function explicitly states that "the asm code
that gets us here will have restored a usable GDT". That wasn't true in the
case of returning from a ::preserve_context kexec. Make it so.

Without this, the kernel was depending on the called function to reload a
GDT which is appropriate for the kernel before returning.

Test program:

 #include <unistd.h>
 #include <errno.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <linux/kexec.h>
 #include <linux/reboot.h>
 #include <sys/reboot.h>
 #include <sys/syscall.h>

 int main (void)
 {
        struct kexec_segment segment = {};
unsigned char purgatory[] = {
0x66, 0xba, 0xf8, 0x03, // mov $0x3f8, %dx
0xb0, 0x42, // mov $0x42, %al
0xee, // outb %al, (%dx)
0xc3, // ret
};
int ret;

segment.buf = &purgatory;
segment.bufsz = sizeof(purgatory);
segment.mem = (void *)0x400000;
segment.memsz = 0x1000;
ret = syscall(__NR_kexec_load, 0x400000, 1, &segment, KEXEC_PRESERVE_CONTEXT);
if (ret) {
perror("kexec_load");
exit(1);
}

ret = syscall(__NR_reboot, LINUX_REBOOT_MAGIC1, LINUX_REBOOT_MAGIC2, LINUX_REBOOT_CMD_KEXEC);
if (ret) {
perror("kexec reboot");
exit(1);
}
printf("Success\n");
return 0;
 }

Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20241205153343.3275139-2-dwmw2@infradead.org
6 months agoiio: magnetometer: yas530: use signed integer type for clamp limits
Jakob Hauser [Fri, 29 Nov 2024 21:25:07 +0000 (22:25 +0100)] 
iio: magnetometer: yas530: use signed integer type for clamp limits

In the function yas537_measure() there is a clamp_val() with limits of
-BIT(13) and BIT(13) - 1.  The input clamp value h[] is of type s32.  The
BIT() is of type unsigned long integer due to its define in
include/vdso/bits.h.  The lower limit -BIT(13) is recognized as -8192 but
expressed as an unsigned long integer.  The size of an unsigned long
integer differs between 32-bit and 64-bit architectures.  Converting this
to type s32 may lead to undesired behavior.

Additionally, in the calculation lines h[0], h[1] and h[2] the unsigned
long integer divisor BIT(13) causes an unsigned division, shifting the
left-hand side of the equation back and forth, possibly ending up in large
positive values instead of negative values on 32-bit architectures.

To solve those two issues, declare a signed integer with a value of
BIT(13).

There is another omission in the clamp line: clamp_val() returns a value
and it's going nowhere here.  Self-assign it to h[i] to make use of the
clamp macro.

Finally, replace clamp_val() macro by clamp() because after changing the
limits from type unsigned long integer to signed integer it's fine that
way.

Link: https://lkml.kernel.org/r/11609b2243c295d65ab4d47e78c239d61ad6be75.1732914810.git.jahau@rocketmail.com
Fixes: 65f79b501030 ("iio: magnetometer: yas530: Add YAS537 variant")
Signed-off-by: Jakob Hauser <jahau@rocketmail.com>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202411230458.dhZwh3TT-lkp@intel.com/
Closes: https://lore.kernel.org/oe-kbuild-all/202411282222.oF0B4110-lkp@intel.com/
Reviewed-by: David Laight <david.laight@aculab.com>
Acked-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Lars-Peter Clausen <lars@metafoo.de>
Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agosched/numa: fix memory leak due to the overwritten vma->numab_state
Adrian Huang [Wed, 13 Nov 2024 10:21:46 +0000 (18:21 +0800)] 
sched/numa: fix memory leak due to the overwritten vma->numab_state

[Problem Description]
When running the hackbench program of LTP, the following memory leak is
reported by kmemleak.

  # /opt/ltp/testcases/bin/hackbench 20 thread 1000
  Running with 20*40 (== 800) tasks.

  # dmesg | grep kmemleak
  ...
  kmemleak: 480 new suspected memory leaks (see /sys/kernel/debug/kmemleak)
  kmemleak: 665 new suspected memory leaks (see /sys/kernel/debug/kmemleak)

  # cat /sys/kernel/debug/kmemleak
  unreferenced object 0xffff888cd8ca2c40 (size 64):
    comm "hackbench", pid 17142, jiffies 4299780315
    hex dump (first 32 bytes):
      ac 74 49 00 01 00 00 00 4c 84 49 00 01 00 00 00  .tI.....L.I.....
      00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
    backtrace (crc bff18fd4):
      [<ffffffff81419a89>] __kmalloc_cache_noprof+0x2f9/0x3f0
      [<ffffffff8113f715>] task_numa_work+0x725/0xa00
      [<ffffffff8110f878>] task_work_run+0x58/0x90
      [<ffffffff81ddd9f8>] syscall_exit_to_user_mode+0x1c8/0x1e0
      [<ffffffff81dd78d5>] do_syscall_64+0x85/0x150
      [<ffffffff81e0012b>] entry_SYSCALL_64_after_hwframe+0x76/0x7e
  ...

This issue can be consistently reproduced on three different servers:
  * a 448-core server
  * a 256-core server
  * a 192-core server

[Root Cause]
Since multiple threads are created by the hackbench program (along with
the command argument 'thread'), a shared vma might be accessed by two or
more cores simultaneously. When two or more cores observe that
vma->numab_state is NULL at the same time, vma->numab_state will be
overwritten.

Although current code ensures that only one thread scans the VMAs in a
single 'numa_scan_period', there might be a chance for another thread
to enter in the next 'numa_scan_period' while we have not gotten till
numab_state allocation [1].

Note that the command `/opt/ltp/testcases/bin/hackbench 50 process 1000`
cannot the reproduce the issue. It is verified with 200+ test runs.

[Solution]
Use the cmpxchg atomic operation to ensure that only one thread executes
the vma->numab_state assignment.

[1] https://lore.kernel.org/lkml/1794be3c-358c-4cdc-a43d-a1f841d91ef7@amd.com/

Link: https://lkml.kernel.org/r/20241113102146.2384-1-ahuang12@lenovo.com
Fixes: ef6a22b70f6d ("sched/numa: apply the scan delay to every new vma")
Signed-off-by: Adrian Huang <ahuang12@lenovo.com>
Reported-by: Jiwei Sun <sunjw10@lenovo.com>
Reviewed-by: Raghavendra K T <raghavendra.kt@amd.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Ben Segall <bsegall@google.com>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Valentin Schneider <vschneid@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agomm/damon: fix order of arguments in damos_before_apply tracepoint
Akinobu Mita [Fri, 15 Nov 2024 18:20:23 +0000 (10:20 -0800)] 
mm/damon: fix order of arguments in damos_before_apply tracepoint

Since the order of the scheme_idx and target_idx arguments in TP_ARGS is
reversed, they are stored in the trace record in reverse.

Link: https://lkml.kernel.org/r/20241115182023.43118-1-sj@kernel.org
Link: https://patch.msgid.link/20241112154828.40307-1-akinobu.mita@gmail.com
Fixes: c603c630b509 ("mm/damon/core: add a tracepoint for damos apply target regions")
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agolib: stackinit: hide never-taken branch from compiler
Kees Cook [Sun, 17 Nov 2024 11:38:13 +0000 (03:38 -0800)] 
lib: stackinit: hide never-taken branch from compiler

The never-taken branch leads to an invalid bounds condition, which is by
design. To avoid the unwanted warning from the compiler, hide the
variable from the optimizer.

../lib/stackinit_kunit.c: In function 'do_nothing_u16_zero':
../lib/stackinit_kunit.c:51:49: error: array subscript 1 is outside array bounds of 'u16[0]' {aka 'short unsigned int[]'} [-Werror=array-bounds=]
   51 | #define DO_NOTHING_RETURN_SCALAR(ptr)           *(ptr)
      |                                                 ^~~~~~
../lib/stackinit_kunit.c:219:24: note: in expansion of macro 'DO_NOTHING_RETURN_SCALAR'
  219 |                 return DO_NOTHING_RETURN_ ## which(ptr + 1);    \
      |                        ^~~~~~~~~~~~~~~~~~

Link: https://lkml.kernel.org/r/20241117113813.work.735-kees@kernel.org
Signed-off-by: Kees Cook <kees@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agomm/filemap: don't call folio_test_locked() without a reference in next_uptodate_folio()
David Hildenbrand [Fri, 29 Nov 2024 12:53:03 +0000 (13:53 +0100)] 
mm/filemap: don't call folio_test_locked() without a reference in next_uptodate_folio()

The folio can get freed + buddy-merged + reallocated in the meantime,
resulting in us calling folio_test_locked() possibly on a tail page.

This makes const_folio_flags VM_BUG_ON_PGFLAGS() when stumbling over the
tail page.

Could this result in other issues?  Doesn't look like it.  False positives
and false negatives don't really matter, because this folio would get
skipped either way when detecting that they have been reallocated in the
meantime.

Fix it by performing the folio_test_locked() checked after grabbing a
reference.  If this ever becomes a real problem, we could add a special
helper that racily checks if the bit is set even on tail pages ...  but
let's hope that's not required so we can just handle it cleaner: work on
the folio after we hold a reference.

Do we really need the folio_test_locked() check if we are going to trylock
briefly after?  Well, we can at least avoid a xas_reload().

It's a bit unclear which exact change introduced that issue.  Likely, ever
since we made PG_locked obey to the PF_NO_TAIL policy it could have been
triggered in some way.

Link: https://lkml.kernel.org/r/20241129125303.4033164-1-david@redhat.com
Fixes: 48c935ad88f5 ("page-flags: define PG_locked behavior on compound pages")
Signed-off-by: David Hildenbrand <david@redhat.com>
Reported-by: syzbot+9f9a7f73fb079b2387a6@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/lkml/674184c9.050a0220.1cc393.0001.GAE@google.com/
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: Hillf Danton <hdanton@sina.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agoscatterlist: fix incorrect func name in kernel-doc
Randy Dunlap [Sat, 30 Nov 2024 02:24:06 +0000 (18:24 -0800)] 
scatterlist: fix incorrect func name in kernel-doc

Fix a kernel-doc warning by making the kernel-doc function description
match the function name:

include/linux/scatterlist.h:323: warning: expecting prototype for sg_unmark_bus_address(). Prototype was for sg_dma_unmark_bus_address() instead

Link: https://lkml.kernel.org/r/20241130022406.537973-1-rdunlap@infradead.org
Fixes: 42399301203e ("lib/scatterlist: add flag for indicating P2PDMA segments in an SGL")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agomm: correct typo in MMAP_STATE() macro
Lorenzo Stoakes [Mon, 18 Nov 2024 17:54:14 +0000 (17:54 +0000)] 
mm: correct typo in MMAP_STATE() macro

We mistakenly refer to len rather than len_ here.  The only existing
caller passes len to the len_ parameter so this has no impact on the code,
but it is obviously incorrect to do this, so fix it.

Link: https://lkml.kernel.org/r/20241118175414.390827-1-lorenzo.stoakes@oracle.com
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Reviewed-by: Liam R. Howlett <Liam.Howlett@Oracle.com>
Reviewed-by: Wei Yang <richard.weiyang@gmail.com>
Cc: Jann Horn <jannh@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agomm: respect mmap hint address when aligning for THP
Kalesh Singh [Mon, 18 Nov 2024 21:46:48 +0000 (13:46 -0800)] 
mm: respect mmap hint address when aligning for THP

Commit efa7df3e3bb5 ("mm: align larger anonymous mappings on THP
boundaries") updated __get_unmapped_area() to align the start address for
the VMA to a PMD boundary if CONFIG_TRANSPARENT_HUGEPAGE=y.

It does this by effectively looking up a region that is of size,
request_size + PMD_SIZE, and aligning up the start to a PMD boundary.

Commit 4ef9ad19e176 ("mm: huge_memory: don't force huge page alignment on
32 bit") opted out of this for 32bit due to regressions in mmap base
randomization.

Commit d4148aeab412 ("mm, mmap: limit THP alignment of anonymous mappings
to PMD-aligned sizes") restricted this to only mmap sizes that are
multiples of the PMD_SIZE due to reported regressions in some performance
benchmarks -- which seemed mostly due to the reduced spatial locality of
related mappings due to the forced PMD-alignment.

Another unintended side effect has emerged: When a user specifies an mmap
hint address, the THP alignment logic modifies the behavior, potentially
ignoring the hint even if a sufficiently large gap exists at the requested
hint location.

Example Scenario:

Consider the following simplified virtual address (VA) space:

    ...

    0x200000-0x400000 --- VMA A
    0x400000-0x600000 --- Hole
    0x600000-0x800000 --- VMA B

    ...

A call to mmap() with hint=0x400000 and len=0x200000 behaves differently:

  - Before THP alignment: The requested region (size 0x200000) fits into
    the gap at 0x400000, so the hint is respected.

  - After alignment: The logic searches for a region of size
    0x400000 (len + PMD_SIZE) starting at 0x400000.
    This search fails due to the mapping at 0x600000 (VMA B), and the hint
    is ignored, falling back to arch_get_unmapped_area[_topdown]().

In general the hint is effectively ignored, if there is any existing
mapping in the below range:

     [mmap_hint + mmap_size, mmap_hint + mmap_size + PMD_SIZE)

This changes the semantics of mmap hint; from ""Respect the hint if a
sufficiently large gap exists at the requested location" to "Respect the
hint only if an additional PMD-sized gap exists beyond the requested
size".

This has performance implications for allocators that allocate their heap
using mmap but try to keep it "as contiguous as possible" by using the end
of the exisiting heap as the address hint.  With the new behavior it's
more likely to get a much less contiguous heap, adding extra fragmentation
and performance overhead.

To restore the expected behavior; don't use
thp_get_unmapped_area_vmflags() when the user provided a hint address, for
anonymous mappings.

Note: As Yang Shi pointed out: the issue still remains for filesystems
which are using thp_get_unmapped_area() for their get_unmapped_area() op.
It is unclear what worklaods will regress for if we ignore THP alignment
when the hint address is provided for such file backed mappings -- so this
fix will be handled separately.

Link: https://lkml.kernel.org/r/20241118214650.3667577-1-kaleshsingh@google.com
Fixes: efa7df3e3bb5 ("mm: align larger anonymous mappings on THP boundaries")
Signed-off-by: Kalesh Singh <kaleshsingh@google.com>
Reviewed-by: Rik van Riel <riel@surriel.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Kefeng Wang <wangkefeng.wang@huawei.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Yang Shi <yang@os.amperecomputing.com>
Cc: Rik van Riel <riel@surriel.com>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Hans Boehm <hboehm@google.com>
Cc: Lokesh Gidra <lokeshgidra@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agomm: memcg: declare do_memsw_account inline
John Sperbeck [Thu, 28 Nov 2024 20:39:59 +0000 (12:39 -0800)] 
mm: memcg: declare do_memsw_account inline

In commit 66d60c428b23 ("mm: memcg: move legacy memcg event code into
memcontrol-v1.c"), the static do_memsw_account() function was moved from a
.c file to a .h file.  Unfortunately, the traditional inline keyword
wasn't added.  If a file (e.g., a unit test) includes the .h file, but
doesn't refer to do_memsw_account(), it will get a warning like:

mm/memcontrol-v1.h:41:13: warning: unused function 'do_memsw_account' [-Wunused-function]
   41 | static bool do_memsw_account(void)
      |             ^~~~~~~~~~~~~~~~

Link: https://lkml.kernel.org/r/20241128203959.726527-1-jsperbeck@google.com
Fixes: 66d60c428b23 ("mm: memcg: move legacy memcg event code into memcontrol-v1.c")
Signed-off-by: John Sperbeck <jsperbeck@google.com>
Acked-by: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agomm/codetag: swap tags when migrate pages
David Wang [Fri, 29 Nov 2024 02:52:13 +0000 (10:52 +0800)] 
mm/codetag: swap tags when migrate pages

Current solution to adjust codetag references during page migration is
done in 3 steps:

1. sets the codetag reference of the old page as empty (not pointing
   to any codetag);

2. subtracts counters of the new page to compensate for its own
   allocation;

3. sets codetag reference of the new page to point to the codetag of
   the old page.

This does not work if CONFIG_MEM_ALLOC_PROFILING_DEBUG=n because
set_codetag_empty() becomes NOOP.  Instead, let's simply swap codetag
references so that the new page is referencing the old codetag and the old
page is referencing the new codetag.  This way accounting stays valid and
the logic makes more sense.

Link: https://lkml.kernel.org/r/20241129025213.34836-1-00107082@163.com
Fixes: e0a955bf7f61 ("mm/codetag: add pgalloc_tag_copy()")
Signed-off-by: David Wang <00107082@163.com>
Closes: https://lore.kernel.org/lkml/20241124074318.399027-1-00107082@163.com/
Acked-by: Suren Baghdasaryan <surenb@google.com>
Suggested-by: Suren Baghdasaryan <surenb@google.com>
Acked-by: Yu Zhao <yuzhao@google.com>
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agoocfs2: update seq_file index in ocfs2_dlm_seq_next
Wengang Wang [Tue, 19 Nov 2024 17:45:00 +0000 (09:45 -0800)] 
ocfs2: update seq_file index in ocfs2_dlm_seq_next

The following INFO level message was seen:

seq_file: buggy .next function ocfs2_dlm_seq_next [ocfs2] did not
update position index

Fix:
Update *pos (so m->index) to make seq_read_iter happy though the index its
self makes no sense to ocfs2_dlm_seq_next.

Link: https://lkml.kernel.org/r/20241119174500.9198-1-wen.gang.wang@oracle.com
Signed-off-by: Wengang Wang <wen.gang.wang@oracle.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Jun Piao <piaojun@huawei.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agostackdepot: fix stack_depot_save_flags() in NMI context
Marco Elver [Fri, 22 Nov 2024 15:39:47 +0000 (16:39 +0100)] 
stackdepot: fix stack_depot_save_flags() in NMI context

Per documentation, stack_depot_save_flags() was meant to be usable from
NMI context if STACK_DEPOT_FLAG_CAN_ALLOC is unset.  However, it still
would try to take the pool_lock in an attempt to save a stack trace in the
current pool (if space is available).

This could result in deadlock if an NMI is handled while pool_lock is
already held.  To avoid deadlock, only try to take the lock in NMI context
and give up if unsuccessful.

The documentation is fixed to clearly convey this.

Link: https://lkml.kernel.org/r/Z0CcyfbPqmxJ9uJH@elver.google.com
Link: https://lkml.kernel.org/r/20241122154051.3914732-1-elver@google.com
Fixes: 4434a56ec209 ("stackdepot: make fast paths lock-less again")
Signed-off-by: Marco Elver <elver@google.com>
Reported-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agomm: open-code page_folio() in dump_page()
Matthew Wilcox (Oracle) [Mon, 25 Nov 2024 20:17:19 +0000 (20:17 +0000)] 
mm: open-code page_folio() in dump_page()

page_folio() calls page_fixed_fake_head() which will misidentify this page
as being a fake head and load off the end of 'precise'.  We may have a
pointer to a fake head, but that's OK because it contains the right
information for dump_page().

gcc-15 is smart enough to catch this with -Warray-bounds:

In function 'page_fixed_fake_head',
    inlined from '_compound_head' at ../include/linux/page-flags.h:251:24,
    inlined from '__dump_page' at ../mm/debug.c:123:11:
../include/asm-generic/rwonce.h:44:26: warning: array subscript 9 is outside
+array bounds of 'struct page[1]' [-Warray-bounds=]

Link: https://lkml.kernel.org/r/20241125201721.2963278-2-willy@infradead.org
Fixes: fae7d834c43c ("mm: add __dump_folio()")
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reported-by: Kees Cook <kees@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agomm: open-code PageTail in folio_flags() and const_folio_flags()
Matthew Wilcox (Oracle) [Mon, 25 Nov 2024 20:17:18 +0000 (20:17 +0000)] 
mm: open-code PageTail in folio_flags() and const_folio_flags()

It is unsafe to call PageTail() in dump_page() as page_is_fake_head() will
almost certainly return true when called on a head page that is copied to
the stack.  That will cause the VM_BUG_ON_PGFLAGS() in const_folio_flags()
to trigger when it shouldn't.  Fortunately, we don't need to call
PageTail() here; it's fine to have a pointer to a virtual alias of the
page's flag word rather than the real page's flag word.

Link: https://lkml.kernel.org/r/20241125201721.2963278-1-willy@infradead.org
Fixes: fae7d834c43c ("mm: add __dump_folio()")
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Kees Cook <kees@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agomm: fix vrealloc()'s KASAN poisoning logic
Andrii Nakryiko [Tue, 26 Nov 2024 00:52:06 +0000 (16:52 -0800)] 
mm: fix vrealloc()'s KASAN poisoning logic

When vrealloc() reuses already allocated vmap_area, we need to re-annotate
poisoned and unpoisoned portions of underlying memory according to the new
size.

This results in a KASAN splat recorded at [1].  A KASAN mis-reporting
issue where there is none.

Note, hard-coding KASAN_VMALLOC_PROT_NORMAL might not be exactly correct,
but KASAN flag logic is pretty involved and spread out throughout
__vmalloc_node_range_noprof(), so I'm using the bare minimum flag here and
leaving the rest to mm people to refactor this logic and reuse it here.

Link: https://lkml.kernel.org/r/20241126005206.3457974-1-andrii@kernel.org
Link: https://lore.kernel.org/bpf/67450f9b.050a0220.21d33d.0004.GAE@google.com/
Fixes: 3ddc2fefe6f3 ("mm: vmalloc: implement vrealloc()")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agoRevert "readahead: properly shorten readahead when falling back to do_page_cache_ra()"
Jan Kara [Tue, 26 Nov 2024 14:52:08 +0000 (15:52 +0100)] 
Revert "readahead: properly shorten readahead when falling back to do_page_cache_ra()"

This reverts commit 7c877586da3178974a8a94577b6045a48377ff25.

Anders and Philippe have reported that recent kernels occasionally hang
when used with NFS in readahead code.  The problem has been bisected to
7c877586da3 ("readahead: properly shorten readahead when falling back to
do_page_cache_ra()").  The cause of the problem is that ra->size can be
shrunk by read_pages() call and subsequently we end up calling
do_page_cache_ra() with negative (read huge positive) number of pages.
Let's revert 7c877586da3 for now until we can find a proper way how the
logic in read_pages() and page_cache_ra_order() can coexist.  This can
lead to reduced readahead throughput due to readahead window confusion but
that's better than outright hangs.

Link: https://lkml.kernel.org/r/20241126145208.985-1-jack@suse.cz
Fixes: 7c877586da31 ("readahead: properly shorten readahead when falling back to do_page_cache_ra()")
Reported-by: Anders Blomdell <anders.blomdell@gmail.com>
Reported-by: Philippe Troin <phil@fifi.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Tested-by: Philippe Troin <phil@fifi.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agoselftests/damon: add _damon_sysfs.py to TEST_FILES
Maximilian Heyne [Wed, 27 Nov 2024 12:08:53 +0000 (12:08 +0000)] 
selftests/damon: add _damon_sysfs.py to TEST_FILES

When running selftests I encountered the following error message with
some damon tests:

 # Traceback (most recent call last):
 #   File "[...]/damon/./damos_quota.py", line 7, in <module>
 #     import _damon_sysfs
 # ModuleNotFoundError: No module named '_damon_sysfs'

Fix this by adding the _damon_sysfs.py file to TEST_FILES so that it
will be available when running the respective damon selftests.

Link: https://lkml.kernel.org/r/20241127-picks-visitor-7416685b-mheyne@amazon.de
Fixes: 306abb63a8ca ("selftests/damon: implement a python module for test-purpose DAMON sysfs controls")
Signed-off-by: Maximilian Heyne <mheyne@amazon.de>
Reviewed-by: SeongJae Park <sj@kernel.org>
Cc: Shuah Khan <shuah@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agoselftest: hugetlb_dio: fix test naming
Mark Brown [Wed, 27 Nov 2024 16:14:22 +0000 (16:14 +0000)] 
selftest: hugetlb_dio: fix test naming

The string logged when a test passes or fails is used by the selftest
framework to identify which test is being reported.  The hugetlb_dio test
not only uses the same strings for every test that is run but it also uses
different strings for test passes and failures which means that test
automation is unable to follow what the test is doing at all.

Pull the existing duplicated logging of the number of free huge pages
before and after the test out of the conditional and replace that and the
logging of the result with a single ksft_print_result() which incorporates
the parameters passed into the test into the output.

Link: https://lkml.kernel.org/r/20241127-kselftest-mm-hugetlb-dio-names-v1-1-22aab01bf550@kernel.org
Fixes: fae1980347bf ("selftests: hugetlb_dio: fixup check for initial conditions to skip in the start")
Signed-off-by: Mark Brown <broonie@kernel.org>
Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Cc: Donet Tom <donettom@linux.ibm.com>
Cc: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agoocfs2: free inode when ocfs2_get_init_inode() fails
Tetsuo Handa [Sat, 23 Nov 2024 13:28:34 +0000 (22:28 +0900)] 
ocfs2: free inode when ocfs2_get_init_inode() fails

syzbot is reporting busy inodes after unmount, for commit 9c89fe0af826
("ocfs2: Handle error from dquot_initialize()") forgot to call iput() when
new_inode() succeeded and dquot_initialize() failed.

Link: https://lkml.kernel.org/r/e68c0224-b7c6-4784-b4fa-a9fc8c675525@I-love.SAKURA.ne.jp
Fixes: 9c89fe0af826 ("ocfs2: Handle error from dquot_initialize()")
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Reported-by: syzbot+0af00f6a2cba2058b5db@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=0af00f6a2cba2058b5db
Tested-by: syzbot+0af00f6a2cba2058b5db@syzkaller.appspotmail.com
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Jun Piao <piaojun@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agonilfs2: fix potential out-of-bounds memory access in nilfs_find_entry()
Ryusuke Konishi [Tue, 19 Nov 2024 17:23:37 +0000 (02:23 +0900)] 
nilfs2: fix potential out-of-bounds memory access in nilfs_find_entry()

Syzbot reported that when searching for records in a directory where the
inode's i_size is corrupted and has a large value, memory access outside
the folio/page range may occur, or a use-after-free bug may be detected if
KASAN is enabled.

This is because nilfs_last_byte(), which is called by nilfs_find_entry()
and others to calculate the number of valid bytes of directory data in a
page from i_size and the page index, loses the upper 32 bits of the 64-bit
size information due to an inappropriate type of local variable to which
the i_size value is assigned.

This caused a large byte offset value due to underflow in the end address
calculation in the calling nilfs_find_entry(), resulting in memory access
that exceeds the folio/page size.

Fix this issue by changing the type of the local variable causing the bit
loss from "unsigned int" to "u64".  The return value of nilfs_last_byte()
is also of type "unsigned int", but it is truncated so as not to exceed
PAGE_SIZE and no bit loss occurs, so no change is required.

Link: https://lkml.kernel.org/r/20241119172403.9292-1-konishi.ryusuke@gmail.com
Fixes: 2ba466d74ed7 ("nilfs2: directory entry operations")
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Reported-by: syzbot+96d5d14c47d97015c624@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=96d5d14c47d97015c624
Tested-by: syzbot+96d5d14c47d97015c624@syzkaller.appspotmail.com
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
6 months agokasan: make report_lock a raw spinlock
Jared Kangas [Tue, 19 Nov 2024 21:02:34 +0000 (13:02 -0800)] 
kasan: make report_lock a raw spinlock

If PREEMPT_RT is enabled, report_lock is a sleeping spinlock and must not
be locked when IRQs are disabled.  However, KASAN reports may be triggered
in such contexts.  For example:

        char *s = kzalloc(1, GFP_KERNEL);
        kfree(s);
        local_irq_disable();
        char c = *s;  /* KASAN report here leads to spin_lock() */
        local_irq_enable();

Make report_spinlock a raw spinlock to prevent rescheduling when
PREEMPT_RT is enabled.

Link: https://lkml.kernel.org/r/20241119210234.1602529-1-jkangas@redhat.com
Fixes: 342a93247e08 ("locking/spinlock: Provide RT variant header: <linux/spinlock_rt.h>")
Signed-off-by: Jared Kangas <jkangas@redhat.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>