linux-block.git
12 months agoirqdomain: Fix the kernel-doc and plug it into Documentation
Jiri Slaby (SUSE) [Fri, 12 Jul 2024 06:41:48 +0000 (08:41 +0200)]
irqdomain: Fix the kernel-doc and plug it into Documentation

There were several undocumented fields in structs irq_domain_ops and
irq_domain_info. Document them.

irq_domain_ops::revmap_size contained "[]" in the description, which is not
allowed in sphinx. Remove that.

Finally, plug the whole header (irqdomain.h) into genericirq.rst, so that
the docs is autogenerated and hyperlinks to these structure are created.

Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Link: https://lore.kernel.org/r/20240712064148.157040-1-jirislaby@kernel.org
12 months agogenirq: Set IRQF_COND_ONESHOT in request_irq()
Rafael J. Wysocki [Thu, 11 Jul 2024 10:20:04 +0000 (12:20 +0200)]
genirq: Set IRQF_COND_ONESHOT in request_irq()

The callers of request_irq() don't care about IRQF_ONESHOT because they
don't provide threaded handlers, but if they happen to share the IRQ with
the ACPI SCI, which has a threaded handler and sets IRQF_ONESHOT,
request_irq() will fail for them due to a flags mismatch.

Address this by making request_irq() add IRQF_COND_ONESHOT to the flags
passed to request_threaded_irq() for all of its callers.

Fixes: 7a36b901a6eb ("ACPI: OSL: Use a threaded interrupt handler for SCI")
Reported-by: Stefan Seyfried <stefan.seyfried@googlemail.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Stefan Seyfried <stefan.seyfried@googlemail.com>
Cc: stable@vger.kerel.org
Link: https://lore.kernel.org/r/5800834.DvuYhMxLoT@rjwysocki.net
Closes: https://lore.kernel.org/lkml/205bd84a-fe8e-4963-968e-0763285f35ba@message-id.googlemail.com

12 months agoirqchip/imx-irqsteer: Handle runtime power management correctly
Shenwei Wang [Wed, 3 Jul 2024 16:32:50 +0000 (11:32 -0500)]
irqchip/imx-irqsteer: Handle runtime power management correctly

The power domain is automatically activated from clk_prepare(). However, on
certain platforms like i.MX8QM and i.MX8QXP, the power-on handling invokes
sleeping functions, which triggers the 'scheduling while atomic' bug in the
context switch path during device probing:

 BUG: scheduling while atomic: kworker/u13:1/48/0x00000002
 Call trace:
  __schedule_bug+0x54/0x6c
  __schedule+0x7f0/0xa94
  schedule+0x5c/0xc4
  schedule_preempt_disabled+0x24/0x40
  __mutex_lock.constprop.0+0x2c0/0x540
  __mutex_lock_slowpath+0x14/0x20
  mutex_lock+0x48/0x54
  clk_prepare_lock+0x44/0xa0
  clk_prepare+0x20/0x44
  imx_irqsteer_resume+0x28/0xe0
  pm_generic_runtime_resume+0x2c/0x44
  __genpd_runtime_resume+0x30/0x80
  genpd_runtime_resume+0xc8/0x2c0
  __rpm_callback+0x48/0x1d8
  rpm_callback+0x6c/0x78
  rpm_resume+0x490/0x6b4
  __pm_runtime_resume+0x50/0x94
  irq_chip_pm_get+0x2c/0xa0
  __irq_do_set_handler+0x178/0x24c
  irq_set_chained_handler_and_data+0x60/0xa4
  mxc_gpio_probe+0x160/0x4b0

Cure this by implementing the irq_bus_lock/sync_unlock() interrupt chip
callbacks and handle power management in them as they are invoked from
non-atomic context.

[ tglx: Rewrote change log, added Fixes tag ]

Fixes: 0136afa08967 ("irqchip: Add driver for imx-irqsteer controller")
Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20240703163250.47887-1-shenwei.wang@nxp.com
12 months agoirqchip/gic-v3: Pass #redistributor-regions to gic_of_setup_kvm_info()
Geert Uytterhoeven [Tue, 2 Jul 2024 09:24:14 +0000 (11:24 +0200)]
irqchip/gic-v3: Pass #redistributor-regions to gic_of_setup_kvm_info()

The caller of gic_of_setup_kvm_info() already queried DT for the value
of the #redistributor-regions property.  So just pass this value,
instead of doing the DT look-up again in the callee.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/808286a3ac08f60585ae7e2c848e0f9b3cb79cf8.1719912215.git.geert+renesas@glider.be
12 months agoirqchip/bcm2835: Enable SKIP_SET_WAKE and MASK_ON_SUSPEND
Stefan Wahren [Sun, 30 Jun 2024 15:36:46 +0000 (17:36 +0200)]
irqchip/bcm2835: Enable SKIP_SET_WAKE and MASK_ON_SUSPEND

The BCM2835 ARMCTRL interrupt controller doesn't provide any facility to
configure the wakeup sources. That's the reason why the driver lacks the
irq_set_wake() callback for the interrupt chip.

But this prevent to properly enter power management states like "suspend to
idle".

Enable the flags IRQCHIP_SKIP_SET_WAKE and IRQCHIP_MASK_ON_SUSPEND so the
interrupt suspend logic can handle the chip correctly.

Signed-off-by: Stefan Wahren <wahrenst@gmx.net>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
12 months agoirqchip/gic-v4: Make sure a VPE is locked when VMAPP is issued
Marc Zyngier [Fri, 5 Jul 2024 09:31:55 +0000 (10:31 +0100)]
irqchip/gic-v4: Make sure a VPE is locked when VMAPP is issued

In order to make sure that vpe->col_idx is correctly sampled when a VMAPP
command is issued, the vpe_lock must be held for the VPE. This is now
possible since the introduction of the per-VM vmapp_lock, which can be
taken before vpe_lock in the correct locking order.

Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Nianyao Tang <tangnianyao@huawei.com>
Link: https://lore.kernel.org/r/20240705093155.871070-4-maz@kernel.org
12 months agoirqchip/gic-v4: Substitute vmovp_lock for a per-VM lock
Marc Zyngier [Fri, 5 Jul 2024 09:31:54 +0000 (10:31 +0100)]
irqchip/gic-v4: Substitute vmovp_lock for a per-VM lock

vmovp_lock is abused in a number of cases to serialise updates
to vlpi_count[] and deal with map/unmap of a VM to ITSs.

Instead, provide a per-VM lock and revisit the use of vlpi_count[]
so that it is always wrapped in this per-VM vmapp_lock.

This reduces the potential contention on a concurrent VMOVP command,
and paves the way for subsequent VPE locking that holding vmovp_lock
actively prevents due to the lock ordering.

Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Nianyao Tang <tangnianyao@huawei.com>
Link: https://lore.kernel.org/r/20240705093155.871070-3-maz@kernel.org
12 months agoirqchip/gic-v4: Always configure affinity on VPE activation
Marc Zyngier [Fri, 5 Jul 2024 09:31:53 +0000 (10:31 +0100)]
irqchip/gic-v4: Always configure affinity on VPE activation

There are currently two paths to set the initial affinity of a VPE:

 - at activation time on GICv4 without the stupid VMOVP list, and
   on GICv4.1

 - at map time for GICv4 with VMOVP list

The latter location may end-up modifying the affinity of VPE that is
currently running, making the results unpredictible.

Instead, unify the two paths, making sure to set the initial affinity only
at activation time.

Reported-by: Nianyao Tang <tangnianyao@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Nianyao Tang <tangnianyao@huawei.com>
Link: https://lore.kernel.org/r/20240705093155.871070-2-maz@kernel.org
12 months agoRevert "irqchip/dw-apb-ictl: Support building as module"
Thomas Gleixner [Tue, 25 Jun 2024 19:30:48 +0000 (21:30 +0200)]
Revert "irqchip/dw-apb-ictl: Support building as module"

This reverts commit 7cc4f309c933ec5d64eea31066fe86bbf9e48819.

Causes build fails.

Reported-by: Mark Brown <broonie@kernel.org>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Jisheng Zhang <jszhang@kernel.org>
https://lore.kernel.org/oe-kbuild-all/202406250214.WZEjWnnU-lkp@intel.com/

12 months agoRevert "Loongarch: Support loongarch avec"
Thomas Gleixner [Tue, 25 Jun 2024 05:40:08 +0000 (07:40 +0200)]
Revert "Loongarch: Support loongarch avec"

This reverts commit 760d7e719499d64beea62bfcf53938fb233bb6e7.

This results in build failures and has other issues according to Tianyang.

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Tianyang Zhang <zhangtianyang@loongson.cn>
Closes: https://lore.kernel.org/oe-kbuild-all/202406240451.ygBFNyJ3-lkp@intel.com/

12 months agoarm64: Kconfig: Allow build irq-stm32mp-exti driver as module
Antonio Borneo [Thu, 20 Jun 2024 08:31:15 +0000 (10:31 +0200)]
arm64: Kconfig: Allow build irq-stm32mp-exti driver as module

Drop auto-selecting the driver, so it can be built either as a
module or built-in.

Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240620083115.204362-9-antonio.borneo@foss.st.com
12 months agoARM: stm32: Allow build irq-stm32mp-exti driver as module
Antonio Borneo [Thu, 20 Jun 2024 08:31:14 +0000 (10:31 +0200)]
ARM: stm32: Allow build irq-stm32mp-exti driver as module

Drop auto-selecting the driver, so it can be built either as a module or
built-in.

Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240620083115.204362-8-antonio.borneo@foss.st.com
12 months agoirqchip/stm32mp-exti: Allow building as module
Antonio Borneo [Thu, 20 Jun 2024 08:31:13 +0000 (10:31 +0200)]
irqchip/stm32mp-exti: Allow building as module

Allow to build the driver as a module by adding the necessarily hooks in
Kconfig and in the driver's code.

Since all the probe dependencies linked to this driver have already been
fixed, remove the not longer relevant 'arch_initcall'.

Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240620083115.204362-7-antonio.borneo@foss.st.com
12 months agoirqchip/stm32mp-exti: Rename internal symbols
Antonio Borneo [Thu, 20 Jun 2024 08:31:12 +0000 (10:31 +0200)]
irqchip/stm32mp-exti: Rename internal symbols

Rename all the internal symbols accordingly to the new name of the
driver.

Renaming done automatically through sed rules:
s/stm32_exti_set_type/stm32mp_exti_convert_type/g
s/stm32_exti_h_/stm32mp_exti_/g
s/stm32_exti/stm32mp_exti/g
s/stm32_bank/bank/g
s/stm32_/stm32mp_/g
s/STM32_/STM32MP_/g
s/STM32MP1_/STM32MP_/g
s/stm32mp1_exti_/stm32mp_exti_/g
s/stm32-exti-h/stm32mp-exti/g

Manually fix some indentation after the rename.

[ tglx: Mop up more coding style issues while at it ]

Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240620083115.204362-6-antonio.borneo@foss.st.com
12 months agoirqchip/stm32-exti: Split MCU and MPU code
Antonio Borneo [Thu, 20 Jun 2024 08:31:11 +0000 (10:31 +0200)]
irqchip/stm32-exti: Split MCU and MPU code

Keep only the code for ARMv7m STM32 MCUs in in stm32-exti.c and split out
the code for ARMv7a & ARMv8a STM32MPxxx MPUs into stm32mp-exti.c

Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240620083115.204362-5-antonio.borneo@foss.st.com
12 months agoarm64: Kconfig: Select STM32MP_EXTI on STM32 platforms
Antonio Borneo [Thu, 20 Jun 2024 08:31:10 +0000 (10:31 +0200)]
arm64: Kconfig: Select STM32MP_EXTI on STM32 platforms

Use the new config flag to build the correct driver that will be
extracted from the old code.

Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240620083115.204362-4-antonio.borneo@foss.st.com
12 months agoARM: stm32: Use different EXTI driver on ARMv7m and ARMv7a
Antonio Borneo [Thu, 20 Jun 2024 08:31:09 +0000 (10:31 +0200)]
ARM: stm32: Use different EXTI driver on ARMv7m and ARMv7a

Build the proper driver by selecting the appropriate config flag.

Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240620083115.204362-3-antonio.borneo@foss.st.com
12 months agoirqchip/stm32-exti: Add CONFIG_STM32MP_EXTI
Antonio Borneo [Thu, 20 Jun 2024 08:31:08 +0000 (10:31 +0200)]
irqchip/stm32-exti: Add CONFIG_STM32MP_EXTI

To guarantee bisect-ability during the split of stm32-exti in MCU and MPU
code, introduce CONFIG_STM32MP_EXTI.

Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240620083115.204362-2-antonio.borneo@foss.st.com
12 months agoirqchip/dw-apb-ictl: Support building as module
Jisheng Zhang [Fri, 14 Jun 2024 15:34:49 +0000 (23:34 +0800)]
irqchip/dw-apb-ictl: Support building as module

The driver is now always built in. In some synaptics ARM64 SoCs it is used
as a second level interrupt controller hanging off the ARM GIC and is
therefore loadable during boot.

Enable it to be built as a module and handle built-in usage correctly, so
that it continues working on systems where it is the main interrupt
controller.

[ tglx: Massage changelog ]

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614153449.2083-1-jszhang@kernel.org
12 months agoirqchip/riscv-aplic: Simplify the initialization code
Jinjie Ruan [Mon, 3 Jun 2024 12:56:52 +0000 (12:56 +0000)]
irqchip/riscv-aplic: Simplify the initialization code

The initialization code has an is_of_node() check and invokes to_of_node()
for every of_property_*() invocation.

to_of_node() has a is_of_node() check already, so simplify the code by
invoking to_of_node() and checking that for NULL. If not NULL hand in the
node pointer to of_property_*().

The same applies to of_property_*() which fails when invoked with a NULL
node pointer.

[ tglx: Massaged change log ]

Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240603125652.791601-1-ruanjinjie@huawei.com
12 months agoLoongarch: Support loongarch avec
Tianyang Zhang [Tue, 4 Jun 2024 12:50:26 +0000 (20:50 +0800)]
Loongarch: Support loongarch avec

Introduce the advanced extended interrupt controllers. This feature will
allow each core to have 256 independent interrupt vectors and MSI
interrupts can be independently routed to any vector on any CPU.

[ tglx: Fixed up coding style. Made on/offline functions void ]

Co-developed-by: Jianmin Lv <lvjianmin@loongson.cn>
Signed-off-by: Jianmin Lv <lvjianmin@loongson.cn>
Co-developed-by: Liupu Wang <wangliupu@loongson.cn>
Signed-off-by: Liupu Wang <wangliupu@loongson.cn>
Signed-off-by: Tianyang Zhang <zhangtianyang@loongson.cn>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240604125026.18745-1-zhangtianyang@loongson.cn
12 months agoirqchip/armada-370-xp: Add support for 32 MSI interrupts on non-IPI platforms
Pali Rohár [Fri, 21 Jun 2024 09:38:31 +0000 (11:38 +0200)]
irqchip/armada-370-xp: Add support for 32 MSI interrupts on non-IPI platforms

The doorbell interrupts have the following layout on IPI vs no-IPI
platforms:

                    |  0...7  |  8...15  |       16...31       |
  ------------------+---------+----------+---------------------+
       IPI platform |   IPI   |   n/a    |         MSI         |
  ------------------+---------+----------+---------------------+
   non-IPI platform |                   MSI                    |
  ------------------+------------------------------------------+

Currently the driver only allows for the upper 16...31 interrupts for
MSI domain (i.e. the MSI domain has only 16 interrupts).

On platforms where IPI is not available, we can use whole 32 MSI
interrupts.

Implement support also for the lower 16 MSI interrupts on non-IPI
platforms.

[ Marek: refactored, changed commit message ]

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
12 months agoirqchip/armada-370-xp: Do not touch IPI registers on platforms without IPI
Pali Rohár [Fri, 21 Jun 2024 09:38:30 +0000 (11:38 +0200)]
irqchip/armada-370-xp: Do not touch IPI registers on platforms without IPI

On platforms where IPI is not available in the MPIC, the IPI registers
instead represent an additional set of MSI interrupt registers (currently
unused by the driver).

Do not touch these registers on platforms where IPI is not available in the
MPIC.

[ Marek: refactored, changed commit message ]

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
12 months agoirqchip/armada-370-xp: Only call ipi_resume() if IPI is available
Pali Rohár [Fri, 21 Jun 2024 09:38:29 +0000 (11:38 +0200)]
irqchip/armada-370-xp: Only call ipi_resume() if IPI is available

IPI is available only on systems where the mpic controller does not have a
parent interrupt defined (e.g. on Armada XP). If a parent interrupt is
defined, inter-processor interrupts are handled by an interrupt controller
higher in the hierarchy (most probably a parent GIC).

Only call ipi_resume() on systems where IPI is available in the mpic
controller.

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
[ refactored a little and changed commit message ]

12 months agoirqchip/armada-370-xp: Do not allow mapping IRQ 0 and 1
Pali Rohár [Fri, 21 Jun 2024 09:38:28 +0000 (11:38 +0200)]
irqchip/armada-370-xp: Do not allow mapping IRQ 0 and 1

IRQs 0 (IPI) and 1 (MSI) are handled internally by this driver,
generic_handle_domain_irq() is never called for these IRQs.

Disallow mapping these IRQs.

[ Marek: changed commit message ]

Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
12 months agoirqchip/armada-370-xp: Use atomic_io_modify() instead of another spinlock
Marek Behún [Thu, 20 Jun 2024 09:52:33 +0000 (11:52 +0200)]
irqchip/armada-370-xp: Use atomic_io_modify() instead of another spinlock

Use the dedicated atomic_io_modify() instead of a open coded spin_lock() +
readl() + writel() + spin_unlock() sequence.

This allows to drop the irq_controller_lock spinlock from the driver.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
12 months agoirqchip: Add missing MODULE_DESCRIPTION() macros
Jeff Johnson [Sat, 8 Jun 2024 16:14:37 +0000 (09:14 -0700)]
irqchip: Add missing MODULE_DESCRIPTION() macros

On x86, make allmodconfig && make W=1 C=1 reports:
WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/irqchip/irq-ts4800.o
WARNING: modpost: missing MODULE_DESCRIPTION() in drivers/irqchip/irq-meson-gpio.o

Add the missing invocation of the MODULE_DESCRIPTION() macro to all
files which have a MODULE_LICENSE().  This includes a 3rd file,
irq-mvebu-pic.c, which did not produce a warning with the x86
allmodconfig, but which may cause this warning with other kernel
configurations.

Signed-off-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20240608-md-drivers-irqchip-v1-1-dd02c3229277@quicinc.com
12 months agoirqchip/renesas-rzg2l: Reorder function calls in rzg2l_irqc_irq_disable()
Lad Prabhakar [Thu, 6 Jun 2024 19:48:13 +0000 (20:48 +0100)]
irqchip/renesas-rzg2l: Reorder function calls in rzg2l_irqc_irq_disable()

The order of function calls in the disable operation should be the reverse
of that in the enable operation. Thus, reorder the function calls to first
disable the parent IRQ chip before disabling the TINT IRQ.

Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com> # on RZ/G3S
Link: https://lore.kernel.org/r/20240606194813.676823-1-prabhakar.mahadev-lad.rj@bp.renesas.com
13 months agoirqchip/riscv-intc: Remove asmlinkage
Jisheng Zhang [Fri, 14 Jun 2024 15:19:55 +0000 (23:19 +0800)]
irqchip/riscv-intc: Remove asmlinkage

The two functions riscv_intc_aia_irq() and riscv_intc_irq()
are only called by C functions.

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Anup Patel <anup@brainfault.org>
Link: https://lore.kernel.org/r/20240614151955.1949-1-jszhang@kernel.org
13 months agoirqdomain: Make build work for CONFIG_GENERIC_IRQ_CHIP=n
Thomas Gleixner [Mon, 17 Jun 2024 18:31:31 +0000 (20:31 +0200)]
irqdomain: Make build work for CONFIG_GENERIC_IRQ_CHIP=n

ld: kernel/irq/irqdomain.o: in function `irq_domain_instantiate':
kernel/irq/irqdomain.c:296:(.text+0x10dd): undefined reference to `irq_domain_alloc_generic_chips'
ld: kernel/irq/irqdomain.c:313:(.text+0x1218): undefined reference to `irq_domain_remove_generic_chips'
ld: kernel/irq/irqdomain.o: in function `irq_domain_remove':
kernel/irq/irqdomain.c:349:(.text+0x1ddf): undefined reference to `irq_domain_remove_generic_chips'

Provide the required stubs.

Fixes: e6f67ce32e8e ("irqdomain: Add support for generic irq chips creation before publishing a domain")
Reported-by: Borislav Betkov <bp@alien8.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 months agoMAINTAINERS: Add the Microchip LAN966x OIC driver entry
Herve Codina [Fri, 14 Jun 2024 17:32:24 +0000 (19:32 +0200)]
MAINTAINERS: Add the Microchip LAN966x OIC driver entry

After contributing the driver, add myself as the maintainer for the
Microchip LAN966x OIC driver.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-24-herve.codina@bootlin.com
13 months agoirqchip: Add support for LAN966x OIC
Herve Codina [Fri, 14 Jun 2024 17:32:23 +0000 (19:32 +0200)]
irqchip: Add support for LAN966x OIC

The Microchip LAN966x outband interrupt controller (OIC) maps the
internal interrupt sources of the LAN966x device to an external
interrupt.
When the LAN966x device is used as a PCI device, the external interrupt
is routed to the PCI interrupt.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-23-herve.codina@bootlin.com
13 months agodt-bindings: interrupt-controller: Add support for Microchip LAN966x OIC
Herve Codina [Fri, 14 Jun 2024 17:32:22 +0000 (19:32 +0200)]
dt-bindings: interrupt-controller: Add support for Microchip LAN966x OIC

The Microchip LAN966x outband interrupt controller (OIC) maps the
internal interrupt sources of the LAN966x device to an external
interrupt.
When the LAN966x device is used as a PCI device, the external interrupt
is routed to the PCI interrupt.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
Link: https://lore.kernel.org/r/20240614173232.1184015-22-herve.codina@bootlin.com
13 months agoirqdomain: Remove __irq_domain_add()
Herve Codina [Fri, 14 Jun 2024 17:32:21 +0000 (19:32 +0200)]
irqdomain: Remove __irq_domain_add()

__irq_domain_add() has been replaced by irq_domain_instanciate() and so,
it is no more used.

Simply remove it.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-21-herve.codina@bootlin.com
13 months ago_PATCH_19_23_um_virt_pci_Use_irq_domain_instantiate_
Herve Codina [Fri, 14 Jun 2024 17:32:20 +0000 (19:32 +0200)]
_PATCH_19_23_um_virt_pci_Use_irq_domain_instantiate_

um_pci_init() uses __irq_domain_add(). With the introduction of
irq_domain_instantiate(), __irq_domain_add() becomes obsolete.

In order to fully remove __irq_domain_add(), use directly
irq_domain_instantiate().

[ tglx: Fixup struct initializer ]

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-20-herve.codina@bootlin.com
13 months agoirqdomain: Convert domain creation functions to irq_domain_instantiate()
Herve Codina [Fri, 14 Jun 2024 17:32:19 +0000 (19:32 +0200)]
irqdomain: Convert domain creation functions to irq_domain_instantiate()

Domain creation functions use __irq_domain_add(). With the introduction
of irq_domain_instantiate(), __irq_domain_add() becomes obsolete.

In order to fully remove __irq_domain_add(), convert domain
creation function to irq_domain_instantiate()

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-19-herve.codina@bootlin.com
13 months agoirqdomain: Convert __irq_domain_add() wrappers to irq_domain_instantiate()
Herve Codina [Fri, 14 Jun 2024 17:32:18 +0000 (19:32 +0200)]
irqdomain: Convert __irq_domain_add() wrappers to irq_domain_instantiate()

__irq_domain_add() wrappers use directly __irq_domain_add(). With the
introduction of irq_domain_instantiate(), __irq_domain_add() becomes
obsolete.

In order to fully remove __irq_domain_add(), convert wrappers to
irq_domain_instantiate()

[ tglx: Fixup struct initializers ]

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-18-herve.codina@bootlin.com
13 months agoirqdomain: Add a resource managed version of irq_domain_instantiate()
Herve Codina [Fri, 14 Jun 2024 17:32:17 +0000 (19:32 +0200)]
irqdomain: Add a resource managed version of irq_domain_instantiate()

Add a devres version of irq_domain_instantiate().

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-17-herve.codina@bootlin.com
13 months agoirqdomain: Add support for generic irq chips creation before publishing a domain
Herve Codina [Fri, 14 Jun 2024 17:32:16 +0000 (19:32 +0200)]
irqdomain: Add support for generic irq chips creation before publishing a domain

The current API functions create an irq_domain and also publish this
newly created to domain. Once an irq_domain is published, consumers can
request IRQ in order to use them.

Some interrupt controller drivers have to perform some more operations
with the created irq_domain in order to have it ready to be used.
For instance:
   - Allocate generic irq chips with irq_alloc_domain_generic_chips()
   - Retrieve the generic irq chips with irq_get_domain_generic_chip()
   - Initialize retrieved chips: set register base address and offsets,
     set several hooks such as irq_mask, irq_unmask, ...

With the newly introduced irq_domain_alloc_generic_chips(), an interrupt
controller driver can use the irq_domain_chip_generic_info structure and
set the init() hook to perform its generic chips initialization.

In order to avoid a window where the domain is published but not yet
ready to be used, handle the generic chip creation (i.e the
irq_domain_alloc_generic_chips() call) before the domain is published.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-16-herve.codina@bootlin.com
13 months agogenirq/generic_chip: Introduce init() and exit() hooks
Herve Codina [Fri, 14 Jun 2024 17:32:15 +0000 (19:32 +0200)]
genirq/generic_chip: Introduce init() and exit() hooks

Most of generic chip drivers need to perform some more additional
initializations on the generic chips allocated before they can be fully
ready.

These additional initializations need to be performed before the IRQ
domain is published to avoid a race condition between IRQ consumers and
suppliers.

Introduce the init() hook to perform these initializations at the right
place just after the generic chip creation. Also introduce the exit() hook
to allow reverting operations done by the init() hook just before the
generic chip is destroyed.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-15-herve.codina@bootlin.com
13 months agogenirq/generic_chip: Introduce irq_domain_{alloc,remove}_generic_chips()
Herve Codina [Fri, 14 Jun 2024 17:32:14 +0000 (19:32 +0200)]
genirq/generic_chip: Introduce irq_domain_{alloc,remove}_generic_chips()

The existing __irq_alloc_domain_generic_chips() uses a bunch of parameters
to describe the generic chips that need to be allocated.

Adding more parameters and wrappers to hide new parameters in the existing
code leads to more and more code without any relevant values and without
any flexibility.

Introduce irq_domain_alloc_generic_chips() where the generic chips
description is done using the irq_domain_chip_generic_info structure
instead of the bunch of parameters to allow flexibility and easy evolution.

Also introduce irq_domain_remove_generic_chips() to revert the operations
done by irq_domain_alloc_generic_chips().

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-14-herve.codina@bootlin.com
13 months agoirqdomain: Introduce init() and exit() hooks
Herve Codina [Fri, 14 Jun 2024 17:32:13 +0000 (19:32 +0200)]
irqdomain: Introduce init() and exit() hooks

The current API does not allow additional initialization before the
domain is published. This can lead to a race condition between consumers
and supplier as a domain can be available for consumers before being
fully ready.

Introduce the init() hook to allow additional initialization before
plublishing the domain. Also introduce the exit() hook to revert
operations done in init() on domain removal.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-13-herve.codina@bootlin.com
13 months agoirqdomain: Handle domain bus token in irq_domain_create()
Herve Codina [Fri, 14 Jun 2024 17:32:12 +0000 (19:32 +0200)]
irqdomain: Handle domain bus token in irq_domain_create()

irq_domain_update_bus_token() is the only way to set the domain bus
token. This is sub-optimal as irq_domain_update_bus_token() can be called
only once the domain is created and needs to revert some operations, change
the domain name and redo the operations.

In order to avoid this revert/change/redo sequence, take the domain bus
into account token during the domain creation.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-12-herve.codina@bootlin.com
13 months agoirqdomain: Make __irq_domain_create() return an error code
Herve Codina [Fri, 14 Jun 2024 17:32:11 +0000 (19:32 +0200)]
irqdomain: Make __irq_domain_create() return an error code

__irq_domain_create() can fail for several reasons. When it fails it
returns a NULL pointer and so filters out the exact failure reason.
The only user of __irq_domain_create() is irq_domain_instantiate() which
can return a PTR_ERR value. On __irq_domain_create() failure, it uses an
arbitrary error code.

Rather than using this arbitrary error value, make __irq_domain_create()
return is own error code and use that one.

[ tglx: Remove the pointless ERR_CAST. domain is a valid return pointer ]

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-11-herve.codina@bootlin.com
13 months agoirqdomain: Use irq_domain_instantiate() for hierarchy domain creation
Herve Codina [Fri, 14 Jun 2024 17:32:10 +0000 (19:32 +0200)]
irqdomain: Use irq_domain_instantiate() for hierarchy domain creation

irq_domain_instantiate() handles all needs to be used in
irq_domain_create_hierarchy()

Avoid code duplication and use directly irq_domain_instantiate() for
hierarchy domain creation.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-10-herve.codina@bootlin.com
13 months agoirqdomain: Handle domain hierarchy parent in irq_domain_instantiate()
Herve Codina [Fri, 14 Jun 2024 17:32:09 +0000 (19:32 +0200)]
irqdomain: Handle domain hierarchy parent in irq_domain_instantiate()

To use irq_domain_instantiate() from irq_domain_create_hierarchy(),
irq_domain_instantiate() needs to handle the domain hierarchy parent.

Add the required functionality.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-9-herve.codina@bootlin.com
13 months agoirqdomain: Handle additional domain flags in irq_domain_instantiate()
Herve Codina [Fri, 14 Jun 2024 17:32:08 +0000 (19:32 +0200)]
irqdomain: Handle additional domain flags in irq_domain_instantiate()

In order to use irq_domain_instantiate() from several places such as
irq_domain_create_hierarchy(), irq_domain_instantiate() needs to handle
additional domain flags.

Add the required infrastructure.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-8-herve.codina@bootlin.com
13 months agoirqdomain: Convert __irq_domain_create() to use struct irq_domain_info
Herve Codina [Fri, 14 Jun 2024 17:32:07 +0000 (19:32 +0200)]
irqdomain: Convert __irq_domain_create() to use struct irq_domain_info

The existing __irq_domain_create() use a bunch of parameters to create
an irq domain.

With the introduction of irq_domain_info structure, these parameters are
available in the information structure itself.
Using directly this information structure allows future flexibility to
add other parameters in a simple way without the need to change the
__irq_domain_create() prototype.

Convert __irq_domain_create() to use the information structure.

[ tglx: Fixup struct initializer ]

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-7-herve.codina@bootlin.com
13 months agoirqdomain: Use a dedicated function to set the domain name
Herve Codina [Fri, 14 Jun 2024 17:32:06 +0000 (19:32 +0200)]
irqdomain: Use a dedicated function to set the domain name

The interrupt domain name computation and setting is directly done in
__irq_domain_create(). This leads to a quite long __irq_domain_create()
function.

In order to simplify __irq_domain_create() and isolate the domain name
computation and setting, move the related operations to a dedicated
function.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-6-herve.codina@bootlin.com
13 months agoirqdomain: Constify parameter in is_fwnode_irqchip()
Herve Codina [Fri, 14 Jun 2024 17:32:05 +0000 (19:32 +0200)]
irqdomain: Constify parameter in is_fwnode_irqchip()

The fwnode parameter has no reason to be a pointer to an un-const struct
fwnode_handle. Indeed, struct fwnode_handle is not supposed to be modified
by the function.

Be consistent with other function performing the same kind of operation
such as is_of_node(), is_acpi_device_node() or is_software_node(): constify
the fwnode parameter.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-5-herve.codina@bootlin.com
13 months agoirqdomain: Introduce irq_domain_instantiate()
Herve Codina [Fri, 14 Jun 2024 17:32:03 +0000 (19:32 +0200)]
irqdomain: Introduce irq_domain_instantiate()

The existing irq_domain_add_*() functions used to instantiate an IRQ
domain are wrappers built on top of __irq_domain_add() and describe the
domain properties using a bunch of parameters.

Adding more parameters and wrappers to hide new parameters in the
existing code lead to more and more code without any relevant value and
without any flexibility.

Introduce irq_domain_instantiate() where the interrupt domain properties
are given using a irq_domain_info structure instead of the bunch of
parameters to allow flexibility and easy evolution.

irq_domain_instantiate() performs the same operation as the one done by
__irq_domain_add(). For compatibility reason with existing code, keep
__irq_domain_add() but convert it to irq_domain_instantiate().

[ tglx: Fixed up struct initializer coding style ]

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-3-herve.codina@bootlin.com
13 months agoirqdomain: Introduce irq_domain_free()
Herve Codina [Fri, 14 Jun 2024 17:32:02 +0000 (19:32 +0200)]
irqdomain: Introduce irq_domain_free()

In preparation of the introduction of the irq domain instantiation,
introduce irq_domain_free() to avoid code duplication on later
modifications.

This new function is an extraction of the current operations performed
to free the irq domain. No functional change intended.

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240614173232.1184015-2-herve.codina@bootlin.com
13 months agoirqdomain: Fixed unbalanced fwnode get and put
Herve Codina [Fri, 14 Jun 2024 17:32:04 +0000 (19:32 +0200)]
irqdomain: Fixed unbalanced fwnode get and put

fwnode_handle_get(fwnode) is called when a domain is created with fwnode
passed as a function parameter. fwnode_handle_put(domain->fwnode) is called
when the domain is destroyed but during the creation a path exists that
does not set domain->fwnode.

If this path is taken, the fwnode get will never be put.

To avoid the unbalanced get and put, set domain->fwnode unconditionally.

Fixes: d59f6617eef0 ("genirq: Allow fwnode to carry name information only")
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20240614173232.1184015-4-herve.codina@bootlin.com
13 months agoirqchip/gic-v3: Enable non-coherent redistributors/ITSes ACPI probing
Lorenzo Pieralisi [Thu, 6 Jun 2024 09:42:38 +0000 (11:42 +0200)]
irqchip/gic-v3: Enable non-coherent redistributors/ITSes ACPI probing

The GIC architecture specification defines a set of registers for
redistributors and ITSes that control the sharebility and cacheability
attributes of redistributors/ITSes initiator ports on the interconnect
(GICR_[V]PROPBASER, GICR_[V]PENDBASER, GITS_BASER<n>).

Architecturally the GIC provides a means to drive shareability and
cacheability attributes signals but it is not mandatory for designs to
wire up the corresponding interconnect signals that control the
cacheability/shareability of transactions.

Redistributors and ITSes interconnect ports can be connected to
non-coherent interconnects that are not able to manage the
shareability/cacheability attributes; this implicitly makes the
redistributors and ITSes non-coherent observers.

To enable non-coherent GIC designs on ACPI based systems, parse the MADT
GICC/GICR/ITS subtables non-coherent flags to determine whether the
respective components are non-coherent observers and force the
shareability attributes to be programmed into the redistributors and
ITSes registers.

An ACPI global function (acpi_get_madt_revision()) is added to retrieve
the MADT revision, in that it is essential to check the MADT revision
before checking for flags that were added with MADT revision 7 so that
if the kernel is booted with an ACPI MADT table with revision < 7 it
skips parsing the newly added flags (that should be zeroed reserved
values for MADT versions < 7 but they could turn out to be buggy and
should be ignored).

Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Link: https://lore.kernel.org/r/20240606094238.757649-2-lpieralisi@kernel.org
13 months agoirqchip/renesas-rzg2l: Add support for RZ/Five SoC
Lad Prabhakar [Tue, 4 Jun 2024 17:37:10 +0000 (18:37 +0100)]
irqchip/renesas-rzg2l: Add support for RZ/Five SoC

The IX45 block has additional mask registers (NMSK/IMSK/TMSK) compared
to the RZ/G2L (family) SoC.

A new rzfive_irqc_chip irq_chip is introduced for RZ/Five, where function
pointers for irq_[un]mask() and irq_[dis|en]able() handle the ([un]masking
of the interrupts. The irq_chip pointer is now passed as an init callback
and stored in the priv pointer to differentiate between RZ/G2L and RZ/Five.

Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20240604173710.534132-3-prabhakar.mahadev-lad.rj@bp.renesas.com
13 months agodt-bindings: interrupt-controller: renesas,rzg2l-irqc: Document RZ/Five SoC
Lad Prabhakar [Tue, 4 Jun 2024 17:37:09 +0000 (18:37 +0100)]
dt-bindings: interrupt-controller: renesas,rzg2l-irqc: Document RZ/Five SoC

Document RZ/Five (R9A07G043F) IRQC bindings. The IRQC block on the RZ/Five
SoC is almost identical to the one found on the RZ/G2L SoC, with the only
difference being that it has additional mask control registers for
NMI/IRQ/TINT.

Hence new compatible string "renesas,r9a07g043f-irqc" is added for RZ/Five
SoC.

Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20240604173710.534132-2-prabhakar.mahadev-lad.rj@bp.renesas.com
13 months agoirqdomain: Add missing parameter descriptions in kernel-doc comments
Herve Codina [Mon, 27 May 2024 16:14:36 +0000 (18:14 +0200)]
irqdomain: Add missing parameter descriptions in kernel-doc comments

During compilation, several warning of the following form were raised:
  Function parameter or struct member 'x' not described in 'yyy'

Add the missing function parameter descriptions.

Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240527161450.326615-10-herve.codina@bootlin.com
13 months agogenirq/debugfs: Print irqdomain flags as human-readable strings
Jinjie Ruan [Wed, 29 May 2024 09:16:28 +0000 (09:16 +0000)]
genirq/debugfs: Print irqdomain flags as human-readable strings

Improve the readability of irqdomain debugging information in debugfs by
printing the flags field of domain files as human-readable strings instead
of a raw bitmask, which aligned with the existing style used for irqchip
flags in the irq debug files.

Before:
#cat :cpus:cpu@0:interrupt-controller
name:   :cpus:cpu@0:interrupt-controller
 size:   0
 mapped: 2
 flags:  0x00000003

After:
#cat :cpus:cpu@0:interrupt-controller
name:   :cpus:cpu@0:interrupt-controller
 size:   0
 mapped: 3
 flags:  0x00000003
            IRQ_DOMAIN_FLAG_HIERARCHY
            IRQ_DOMAIN_NAME_ALLOCATED

Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20240529091628.3666379-1-ruanjinjie@huawei.com
13 months agogenirq/proc: Simplify irqdesc::kstat_irqs handling further
Adrian Huang [Wed, 15 May 2024 10:06:32 +0000 (18:06 +0800)]
genirq/proc: Simplify irqdesc::kstat_irqs handling further

Interrupts which have no action and chained interrupts can be
ignored due to the following reasons (as per tglx's comment):

  1) Interrupts which have no action are completely uninteresting as
     there is no real information attached.

  2) Chained interrupts do not have a count at all.

So there is no point to evaluate the number of accounted interrupts before
checking for non-requested or chained interrupts.

Remove the any_count logic and simply check whether the interrupt
descriptor has the kstat_irqs member populated.

[ tglx: Adapted to upstream changes ]

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Adrian Huang <ahuang12@lenovo.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Jiwei Sun <sunjw10@lenovo.com>
Link: https://lore.kernel.org/r/20240515100632.1419-1-ahuang12@lenovo.com
Link: https://lore.kernel.org/lkml/87h6f0knau.ffs@tglx/
13 months agoLinux 6.10-rc2 v6.10-rc2
Linus Torvalds [Sun, 2 Jun 2024 22:44:56 +0000 (15:44 -0700)]
Linux 6.10-rc2

13 months agoMerge tag 'ata-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/libata...
Linus Torvalds [Sun, 2 Jun 2024 20:30:53 +0000 (13:30 -0700)]
Merge tag 'ata-6.10-rc2' of git://git./linux/kernel/git/libata/linux

Pull ata fixes from Niklas Cassel:

 - Add a quirk for three different devices that have shown issues with
   LPM (link power management). These devices appear to not implement
   LPM properly, since we see command timeouts when enabling LPM. The
   quirk disables LPM for these problematic devices. (Me)

 - Do not apply the Intel PCS quirk on Alder Lake. The quirk is not
   needed and was originally added by mistake when LPM support was
   enabled for this AHCI controller. Enabling the quirk when not needed
   causes the the controller to not be able to detect the connected
   devices on some platforms.

* tag 'ata-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/libata/linux:
  ata: libata-core: Add ATA_HORKAGE_NOLPM for Apacer AS340
  ata: libata-core: Add ATA_HORKAGE_NOLPM for AMD Radeon S3 SSD
  ata: libata-core: Add ATA_HORKAGE_NOLPM for Crucial CT240BX500SSD1
  ata: ahci: Do not apply Intel PCS quirk on Intel Alder Lake

13 months agoMerge tag 'x86-urgent-2024-06-02' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 2 Jun 2024 16:32:34 +0000 (09:32 -0700)]
Merge tag 'x86-urgent-2024-06-02' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar:
 "Miscellaneous topology parsing fixes:

   - Fix topology parsing regression on older CPUs in the new AMD/Hygon
     parser

   - Fix boot crash on odd Intel Quark and similar CPUs that do not fill
     out cpuinfo_x86::x86_clflush_size and zero out
     cpuinfo_x86::x86_cache_alignment as a result.

     Provide 32 bytes as a general fallback value.

   - Fix topology enumeration on certain rare CPUs where the BIOS locks
     certain CPUID leaves and the kernel unlocked them late, which broke
     with the new topology parsing code. Factor out this unlocking logic
     and move it earlier in the parsing sequence"

* tag 'x86-urgent-2024-06-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/topology/intel: Unlock CPUID before evaluating anything
  x86/cpu: Provide default cache line size if not enumerated
  x86/topology/amd: Evaluate SMT in CPUID leaf 0x8000001e only on family 0x17 and greater

13 months agoMerge tag 'sched-urgent-2024-06-02' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 2 Jun 2024 16:23:35 +0000 (09:23 -0700)]
Merge tag 'sched-urgent-2024-06-02' of git://git./linux/kernel/git/tip/tip

Pull scheduler fix from Ingo Molnar:
 "Export a symbol to make life easier for instrumentation/debugging"

* tag 'sched-urgent-2024-06-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/x86: Export 'percpu arch_freq_scale'

13 months agoMerge tag 'perf-urgent-2024-06-02' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 2 Jun 2024 16:20:37 +0000 (09:20 -0700)]
Merge tag 'perf-urgent-2024-06-02' of git://git./linux/kernel/git/tip/tip

Pull perf events fix from Ingo Molnar:
 "Add missing MODULE_DESCRIPTION() lines"

* tag 'perf-urgent-2024-06-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/intel: Add missing MODULE_DESCRIPTION() lines
  perf/x86/rapl: Add missing MODULE_DESCRIPTION() line

13 months agoMerge tag 'hardening-v6.10-rc2-take2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 2 Jun 2024 16:15:28 +0000 (09:15 -0700)]
Merge tag 'hardening-v6.10-rc2-take2' of git://git./linux/kernel/git/kees/linux

Pull hardening fixes from Kees Cook:

 - scsi: mpt3sas: Avoid possible run-time warning with long manufacturer
   strings

 - mailmap: update entry for Kees Cook

 - kunit/fortify: Remove __kmalloc_node() test

* tag 'hardening-v6.10-rc2-take2' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
  kunit/fortify: Remove __kmalloc_node() test
  mailmap: update entry for Kees Cook
  scsi: mpt3sas: Avoid possible run-time warning with long manufacturer strings

13 months agoMerge tag 'powerpc-6.10-2' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Sun, 2 Jun 2024 00:34:35 +0000 (17:34 -0700)]
Merge tag 'powerpc-6.10-2' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc fixes from Michael Ellerman:

 - Enforce full ordering for ATOMIC operations with BPF_FETCH

 - Fix uaccess build errors seen with GCC 13/14

 - Fix build errors on ppc32 due to ARCH_HAS_KERNEL_FPU_SUPPORT

 - Drop error message from lparcfg guest name lookup

Thanks to Christophe Leroy, Guenter Roeck, Nathan Lynch, Naveen N Rao,
Puranjay Mohan, and Samuel Holland.

* tag 'powerpc-6.10-2' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc: Limit ARCH_HAS_KERNEL_FPU_SUPPORT to PPC64
  powerpc/uaccess: Use YZ asm constraint for ld
  powerpc/uaccess: Fix build errors seen with GCC 13/14
  powerpc/pseries/lparcfg: drop error message from guest name lookup
  powerpc/bpf: enforce full ordering for ATOMIC operations with BPF_FETCH

13 months agoMerge tag 'firewire-fixes-6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 2 Jun 2024 00:05:00 +0000 (17:05 -0700)]
Merge tag 'firewire-fixes-6.10-rc1' of git://git./linux/kernel/git/ieee1394/linux1394

Pull firewire fix from Takashi Sakamoto:
 "After merging a commit 1fffe7a34c89 ("script: modpost: emit a warning
  when the description is missing"), MODULE_DESCRIPTOR seems to be
  mandatory for kernel modules. In FireWire subsystem, the most of
  practical kernel modules have the field, while KUnit test modules do
  not. A single patch is applied to fix them"

* tag 'firewire-fixes-6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394:
  firewire: add missing MODULE_DESCRIPTION() to test modules

13 months agoMerge tag '6.10-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sat, 1 Jun 2024 21:35:57 +0000 (14:35 -0700)]
Merge tag '6.10-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull smb client fixes from Steve French:
 "Two small smb3 fixes:

   - Fix socket creation with sfu mount option (spotted by test generic/423)

   - Minor cleanup: fix missing description in two files"

* tag '6.10-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: fix creating sockets when using sfu mount options
  fs: smb: common: add missing MODULE_DESCRIPTION() macros

13 months agoMerge tag 'kbuild-fixes-v6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/masah...
Linus Torvalds [Sat, 1 Jun 2024 16:33:55 +0000 (09:33 -0700)]
Merge tag 'kbuild-fixes-v6.10' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild fixes from Masahiro Yamada:

 - Fix a Kconfig bug regarding comparisons to 'm' or 'n'

 - Replace missed $(srctree)/$(src)

 - Fix unneeded kallsyms step 3

 - Remove incorrect "compatible" properties from image nodes in
   image.fit

 - Improve gen_kheaders.sh

 - Fix 'make dt_binding_check'

 - Clean up unnecessary code

* tag 'kbuild-fixes-v6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  dt-bindings: kbuild: Fix dt_binding_check on unconfigured build
  kheaders: use `command -v` to test for existence of `cpio`
  kheaders: explicitly define file modes for archived headers
  scripts/make_fit: Drop fdt image entry compatible string
  kbuild: remove a stale comment about cleaning in link-vmlinux.sh
  kbuild: fix short log for AS in link-vmlinux.sh
  kbuild: change scripts/mksysmap into sed script
  kbuild: avoid unneeded kallsyms step 3
  kbuild: scripts/gdb: Replace missed $(srctree)/$(src) w/ $(src)
  kconfig: remove redundant check in expr_join_or()
  kconfig: fix comparison to constant symbols, 'm', 'n'
  kconfig: remove unused expr_is_no()

13 months agoMerge tag 'xfs-6.10-fixes-1' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Linus Torvalds [Sat, 1 Jun 2024 15:59:04 +0000 (08:59 -0700)]
Merge tag 'xfs-6.10-fixes-1' of git://git./fs/xfs/xfs-linux

Pull xfs fixes from Chandan Babu:

 - Fix a livelock by dropping an xfarray sortinfo folio when an error
   is encountered

 - During extended attribute operations, Initialize transaction
   reservation computation based on attribute operation code

 - Relax symbolic link's ondisk verification code to allow symbolic
   links with short remote targets

 - Prevent soft lockups when unmapping file ranges and also during
   remapping blocks during a reflink operation

 - Fix compilation warnings when XFS is built with W=1 option

* tag 'xfs-6.10-fixes-1' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  xfs: Add cond_resched to block unmap range and reflink remap path
  xfs: don't open-code u64_to_user_ptr
  xfs: allow symlinks with short remote targets
  xfs: fix xfs_init_attr_trans not handling explicit operation codes
  xfs: drop xfarray sortinfo folio on error
  xfs: Stop using __maybe_unused in xfs_alloc.c
  xfs: Clear W=1 warning in xfs_iwalk_run_callbacks()

13 months agoMerge tag 'tty-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sat, 1 Jun 2024 15:53:39 +0000 (08:53 -0700)]
Merge tag 'tty-6.10-rc2' of git://git./linux/kernel/git/gregkh/tty

Pull tty fix from Greg KH:
 "Here is a single revert for a much-reported regression in 6.10-rc1
  when it comes to a few older architectures.

  Turns out that the VT ioctls don't work the same across all cpu types
  because of some old compatibility requrements for stuff like alpha and
  powerpc. So revert the change that attempted to have them use the
  _IO() macros and go back to the known-working values instead.

  This has NOT been in linux-next but has had many reports that it fixes
  the issue with 6.10-rc1"

* tag 'tty-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  Revert "VT: Use macros to define ioctls"

13 months agoMerge tag 'landlock-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/mic...
Linus Torvalds [Sat, 1 Jun 2024 15:28:24 +0000 (08:28 -0700)]
Merge tag 'landlock-6.10-rc2' of git://git./linux/kernel/git/mic/linux

Pull landlock fix from Mickaël Salaün:
 "This fixes a wrong path walk triggered by syzkaller"

* tag 'landlock-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/mic/linux:
  selftests/landlock: Add layout1.refer_mount_root
  landlock: Fix d_parent walk

13 months agoRevert "VT: Use macros to define ioctls"
Greg Kroah-Hartman [Sat, 1 Jun 2024 05:28:21 +0000 (07:28 +0200)]
Revert "VT: Use macros to define ioctls"

This reverts commit 8c467f3300591a206fa8dcc6988d768910799872.

Turns out this breaks many architectures as the vt ioctls do not all
match up everywhere due to historical reasons, so the original commit is
invalid for many values.

Reported-by: Nick Bowler <nbowler@draconx.ca>
Reported-by: Arnd Bergmann <arnd@kernel.org>
Reported-by: Jiri Slaby <jirislaby@kernel.org>
Reported-by: Christian Zigotzky <chzigotzky@xenosoft.de>
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Alexey Gladkov <legion@kernel.org>
Link: https://lore.kernel.org/r/ad4e561c-1d49-4f25-882c-7a36c6b1b5c0@draconx.ca
Link: https://lore.kernel.org/r/0da9785e-ba44-4718-9d08-4e96c1ba7ab2@kernel.org
Link: https://lore.kernel.org/all/34d848f4-670b-4493-bf21-130ef862521b@xenosoft.de/
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
13 months agoMerge tag 'drm-fixes-2024-06-01' of https://gitlab.freedesktop.org/drm/kernel
Linus Torvalds [Fri, 31 May 2024 23:26:48 +0000 (16:26 -0700)]
Merge tag 'drm-fixes-2024-06-01' of https://gitlab.freedesktop.org/drm/kernel

Pull drm fixes from Dave Airlie:
 "This is the weekly fixes. Lots of small fixes across the board, one
  BUG_ON fix in shmem seems most important, otherwise amdgpu, i915, xe
  mostly with small fixes to all the other drivers.

  shmem:
   - fix BUG_ON in COW handling
   - warn when trying to pin imported objects

  buddy:
   - fix page size handling

  dma-buf:
   - sw-sync: Don't interfere with IRQ handling
   - fix kthreads-handling error path

  i915:
   - fix a race in audio component by registering it later
   - make DPT object unshrinkable to avoid shrinking when framebuffer
     has not shrunk
   - fix CCS id calculation to fix a perf regression
   - fix selftest caching mode
   - fix FIELD_PREP compiler warnings
   - fix indefinite wait for GT wakeref release
   - revert overeager multi-gt pm reference removal

  xe:
   - pcode polling timeout change
   - fix for deadlocks for faulting VMs
   - error-path lock imbalance fix

  amdgpu:
   - RAS fix
   - fix colorspace property for MST connectors
   - fix for PCIe DPM
   - silence UBSAN warning
   - GPUVM robustness fix
   - partition fix
   - drop deprecated I2C_CLASS_SPD

  amdkfd:
   - revert unused changes for certain 11.0.3 devices
   - simplify APU VRAM handling

  lima:
   - fix dma_resv-related deadlock in object pin

  msm:
   - remove build-time dependency on Python 3.9

  nouveau:
   - nvif: Fix possible integer overflow

  panel:
   - lg-sw43408: Select DP helpers; Declare backlight ops as static
   - sitronix-st7789v: Various fixes for jt240mhqs_hwt_ek_e3 panel

  panfrost:
   - fix dma_resv-related deadlock in object pin"

* tag 'drm-fixes-2024-06-01' of https://gitlab.freedesktop.org/drm/kernel: (35 commits)
  drm/msm: remove python 3.9 dependency for compiling msm
  drm/panel: sitronix-st7789v: fix display size for jt240mhqs_hwt_ek_e3 panel
  drm/panel: sitronix-st7789v: tweak timing for jt240mhqs_hwt_ek_e3 panel
  drm/panel: sitronix-st7789v: fix timing for jt240mhqs_hwt_ek_e3 panel
  drm/amd/pm: remove deprecated I2C_CLASS_SPD support from newly added SMU_14_0_2
  drm/amdgpu: Make CPX mode auto default in NPS4
  drm/amdkfd: simplify APU VRAM handling
  Revert "drm/amdkfd: fix gfx_target_version for certain 11.0.3 devices"
  drm/amdgpu: fix dereference null return value for the function amdgpu_vm_pt_parent
  drm/amdgpu: silence UBSAN warning
  drm/amdgpu: Adjust logic in amdgpu_device_partner_bandwidth()
  drm/i915: Fix audio component initialization
  drm/i915/dpt: Make DPT object unshrinkable
  drm/i915/gt: Fix CCS id's calculation for CCS mode setting
  drm/panel/lg-sw43408: mark sw43408_backlight_ops as static
  drm/i915/selftests: Set always_coherent to false when reading from CPU
  drm/panel/lg-sw43408: select CONFIG_DRM_DISPLAY_DP_HELPER
  drm/i915/guc: avoid FIELD_PREP warning
  drm/i915/gt: Disarm breadcrumbs if engines are already idle
  Revert "drm/i915: Remove extra multi-gt pm-references"
  ...

13 months agoMerge tag 'hwmon-for-v6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 31 May 2024 23:24:11 +0000 (16:24 -0700)]
Merge tag 'hwmon-for-v6.10-rc2' of git://git./linux/kernel/git/groeck/linux-staging

Pull hwmon fixes from Guenter Roeck:

 - sttcs: Fix property spelling

 - intel-m10-bmc-hwmon: Fix multiplier for N6000 board power sensor

 - ltc2992: Fix memory leak

 - dell-smm: Add Dell G15 5511 to fan control whitelist

* tag 'hwmon-for-v6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (shtc1) Fix property misspelling
  hwmon: (intel-m10-bmc-hwmon) Fix multiplier for N6000 board power sensor
  hwmon: (ltc2992) Fix memory leak in ltc2992_parse_dt()
  hwmon: (dell-smm) Add Dell G15 5511 to fan control whitelist

13 months agoMerge tag 'mailbox-fixes-v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 31 May 2024 23:21:00 +0000 (16:21 -0700)]
Merge tag 'mailbox-fixes-v6.10-rc1' of git://git./linux/kernel/git/jassibrar/mailbox

Pull mailbox fix from Jassi Brar:

 - zynqmp-ipi: fix linker error on some configurations

* tag 'mailbox-fixes-v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jassibrar/mailbox:
  mailbox: zynqmp-ipi: drop irq_to_desc() call

13 months agoMerge tag 'spi-fix-v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brooni...
Linus Torvalds [Fri, 31 May 2024 23:17:40 +0000 (16:17 -0700)]
Merge tag 'spi-fix-v6.10-rc1' of git://git./linux/kernel/git/broonie/spi

Pull spi fixes from Mark Brown:
 "A series of fixes that came in since the merge window, the main thing
  being the fixes Andy did for DMA sync where we were calling into the
  DMA API in suprising ways and causing issues as a result, the main
  thing being confusing the IOMMU code.

  We've also got some fairly important fixes for the stm32 driver, it
  supports a wide range of hardware and some optimisations that were
  done recently have broken on some systems, and a fix to prevent
  glitched signals on the bus in the cadence driver"

* tag 'spi-fix-v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi: stm32: Don't warn about spurious interrupts
  spi: Assign dummy scatterlist to unidirectional transfers
  spi: cadence: Ensure data lines set to low during dummy-cycle period
  spi: stm32: Revert change that enabled controller before asserting CS
  spi: Check if transfer is mapped before calling DMA sync APIs
  spi: Don't mark message DMA mapped when no transfer in it is

13 months agoMerge tag 'regulator-fix-v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 31 May 2024 23:12:54 +0000 (16:12 -0700)]
Merge tag 'regulator-fix-v6.10-rc1' of git://git./linux/kernel/git/broonie/regulator

Pull regulator fix from Mark Brown:
 "One fix that came in since -rc1, fixing misuse of a local variable in
  the DT parsing code in the RTQ2208 driver"

* tag 'regulator-fix-v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: rtq2208: Fix invalid memory access when devm_of_regulator_put_matches is called

13 months agoMerge tag 'regmap-fix-v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 31 May 2024 23:09:27 +0000 (16:09 -0700)]
Merge tag 'regmap-fix-v6.10-rc1' of git://git./linux/kernel/git/broonie/regmap

Pull regmap fix from Mark Brown:
 "The I2C bus was not taking account of the register and any padding
  bytes when handling maximum write sizes supported by an I2C adaptor,
  this patch from Jim Wylder fixes that"

* tag 'regmap-fix-v6.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:
  regmap-i2c: Subtract reg size from max_write

13 months agoMerge tag 'block-6.10-20240530' of git://git.kernel.dk/linux
Linus Torvalds [Fri, 31 May 2024 22:31:27 +0000 (15:31 -0700)]
Merge tag 'block-6.10-20240530' of git://git.kernel.dk/linux

Pull block fixes from Jens Axboe:

 - NVMe fixes via Keith:
      - Removing unused fields (Kanchan)
      - Large folio offsets support (Kundan)
      - Multipath NUMA node initialiazation fix (Nilay)
      - Multipath IO stats accounting fixes (Keith)
      - Circular lockdep fix (Keith)
      - Target race condition fix (Sagi)
      - Target memory leak fix (Sagi)

 - bcache fixes

 - null_blk fixes (Damien)

 - Fix regression in io.max due to throttle low removal (Waiman)

 - DM limit table fixes (Christoph)

 - SCSI and block limit fixes (Christoph)

 - zone fixes (Damien)

 - Misc fixes (Christoph, Hannes, hexue)

* tag 'block-6.10-20240530' of git://git.kernel.dk/linux: (25 commits)
  blk-throttle: Fix incorrect display of io.max
  block: Fix zone write plugging handling of devices with a runt zone
  block: Fix validation of zoned device with a runt zone
  null_blk: Do not allow runt zone with zone capacity smaller then zone size
  nvmet: fix a possible leak when destroy a ctrl during qp establishment
  nvme: use srcu for iterating namespace list
  bcache: code cleanup in __bch_bucket_alloc_set()
  bcache: call force_wake_up_gc() if necessary in check_should_bypass()
  bcache: allow allocator to invalidate bucket in gc
  block: check for max_hw_sectors underflow
  block: stack max_user_sectors
  sd: also set max_user_sectors when setting max_sectors
  null_blk: Print correct max open zones limit in null_init_zoned_dev()
  block: delete redundant function declaration
  null_blk: Fix return value of nullb_device_power_store()
  dm: make dm_set_zones_restrictions work on the queue limits
  dm: remove dm_check_zoned
  dm: move setting zoned_enabled to dm_table_set_restrictions
  block: remove blk_queue_max_integrity_segments
  nvme: adjust multiples of NVME_CTRL_PAGE_SIZE in offset
  ...

13 months agoMerge tag 'io_uring-6.10-20240530' of git://git.kernel.dk/linux
Linus Torvalds [Fri, 31 May 2024 22:22:58 +0000 (15:22 -0700)]
Merge tag 'io_uring-6.10-20240530' of git://git.kernel.dk/linux

Pull io_uring fixes from Jens Axboe:
 "A couple of minor fixes for issues introduced in the 6.10 merge window:

   - Ensure that all read/write ops have an appropriate cleanup handler
     set (Breno)

   - Regression for applications still doing multiple mmaps even if
     FEAT_SINGLE_MMAP is set (me)

   - Move kmsg inquiry setting above any potential failure point,
     avoiding a spurious NONEMPTY flag setting on early error (me)"

* tag 'io_uring-6.10-20240530' of git://git.kernel.dk/linux:
  io_uring/net: assign kmsg inq/flags before buffer selection
  io_uring/rw: Free iovec before cleaning async data
  io_uring: don't attempt to mmap larger than what the user asks for

13 months agokunit/fortify: Remove __kmalloc_node() test
Kees Cook [Fri, 31 May 2024 18:57:07 +0000 (11:57 -0700)]
kunit/fortify: Remove __kmalloc_node() test

__kmalloc_node() is considered an "internal" function to the Slab, so
drop it from explicit testing.

Link: https://lore.kernel.org/r/20240531185703.work.588-kees@kernel.org
Signed-off-by: Kees Cook <kees@kernel.org>
13 months agoMerge tag 'dma-mapping-6.10-2024-05-31' of git://git.infradead.org/users/hch/dma...
Linus Torvalds [Fri, 31 May 2024 19:14:55 +0000 (12:14 -0700)]
Merge tag 'dma-mapping-6.10-2024-05-31' of git://git.infradead.org/users/hch/dma-mapping

Pull dma-mapping fixes from Christoph Hellwig:

 - dma-mapping benchmark error handling fixes (Fedor Pchelkin)

 - correct a config symbol reference in the DMA API documentation (Lukas
   Bulwahn)

* tag 'dma-mapping-6.10-2024-05-31' of git://git.infradead.org/users/hch/dma-mapping:
  Documentation/core-api: correct reference to SWIOTLB_DYNAMIC
  dma-mapping: benchmark: handle NUMA_NO_NODE correctly
  dma-mapping: benchmark: fix node id validation
  dma-mapping: benchmark: avoid needless copy_to_user if benchmark fails
  dma-mapping: benchmark: fix up kthread-related error handling

13 months agoMerge tag 'sound-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Fri, 31 May 2024 19:11:44 +0000 (12:11 -0700)]
Merge tag 'sound-6.10-rc2' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "Lots of small fixes:

   - A race fix for debugfs handling in ALSA core

   - A series of corrections for MIDI2 core format conversions

   - ASoC Intel fixes for 16 bit DMIC config

   - Updates for missing module parameters in ASoC code

   - HD-audio quirk, Cirrus codec fix, etc minor fixes"

* tag 'sound-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound: (26 commits)
  ALSA: seq: ump: Fix swapped song position pointer data
  ASoC: SOF: ipc4-topology: Adjust the params based on DAI formats
  ASoC: SOF: ipc4-topology: Improve readability of sof_ipc4_prepare_dai_copier()
  ASoC: SOF: ipc4-topology/pcm: Rename sof_ipc4_copier_is_single_format()
  ASoC: SOF: ipc4-topology: Print out the channel count in sof_ipc4_dbg_audio_format
  ASoC: SOF: ipc4-topology: Add support for NHLT with 16-bit only DMIC blob
  ALSA: seq: Fix yet another spot for system message conversion
  ALSA: ump: Set default protocol when not given explicitly
  ALSA: ump: Don't accept an invalid UMP protocol number
  ASoC: SOF: ipc4-topology: Fix input format query of process modules without base extension
  ASoC: Intel: sof-sdw: fix missing SPI_MASTER dependency
  ALSA: pcm: fix typo in comment
  ALSA: ump: Don't clear bank selection after sending a program change
  ALSA: seq: Fix incorrect UMP type for system messages
  ALSA/hda: intel-dsp-config: reduce log verbosity
  ALSA: seq: Don't clear bank selection at event -> UMP MIDI2 conversion
  ALSA: seq: Fix missing bank setup between MIDI1/MIDI2 UMP conversion
  ASoC: SOF: add missing MODULE_DESCRIPTION()
  ASoC: SOF: reorder MODULE_ definitions
  ASoC: SOF: AMD: group all module related information
  ...

13 months agoMerge tag 'platform-drivers-x86-v6.10-2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 31 May 2024 19:03:28 +0000 (12:03 -0700)]
Merge tag 'platform-drivers-x86-v6.10-2' of git://git./linux/kernel/git/pdx86/platform-drivers-x86

Pull x86 platform driver fixes from Hans de Goede:

 - a use-after-free bugfix

 - Kconfig fixes for randconfig builds

 - allow setting touchscreen_dmi quirks from the cmdline for debugging

 - touchscreen_dmi quirks for two new laptop/tablet models

* tag 'platform-drivers-x86-v6.10-2' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86:
  platform/x86: touchscreen_dmi: Add info for the EZpad 6s Pro
  platform/x86: touchscreen_dmi: Add info for GlobalSpace SolT IVW 11.6" tablet
  platform/x86: touchscreen_dmi: Add support for setting touchscreen properties from cmdline
  platform/x86: thinkpad_acpi: Select INPUT_SPARSEKMAP in Kconfig
  platform/x86: x86-android-tablets: Add "select LEDS_CLASS"
  platform/x86: ISST: fix use-after-free in tpmi_sst_dev_remove()

13 months agoMerge tag 'riscv-for-linus-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 31 May 2024 18:52:06 +0000 (11:52 -0700)]
Merge tag 'riscv-for-linus-6.10-rc2' of git://git./linux/kernel/git/riscv/linux

Pull RISC-V fixes from Palmer Dabbelt:

 - A fix to avoid pt_regs aliasing with idle thread stacks on secondary
   harts.

 - HAVE_ARCH_HUGE_VMAP is enabled on XIP kernels, which fixes boot
   issues on XIP systems with huge pages.

 - An update to the uABI documentation clarifying that only scalar
   misaligned accesses were grandfathered in as supported, as the vector
   extension did not exist at the time the uABI was frozen.

 - A fix for the recently-added byte/half atomics to avoid losing the
   fully ordered decorations.

* tag 'riscv-for-linus-6.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
  riscv: Fix fully ordered LR/SC xchg[8|16]() implementations
  Documentation: RISC-V: uabi: Only scalar misaligned loads are supported
  riscv: enable HAVE_ARCH_HUGE_VMAP for XIP kernel
  riscv: prevent pt_regs corruption for secondary idle threads

13 months agoMerge tag 'bcachefs-2024-05-30' of https://evilpiepirate.org/git/bcachefs
Linus Torvalds [Fri, 31 May 2024 18:45:41 +0000 (11:45 -0700)]
Merge tag 'bcachefs-2024-05-30' of https://evilpiepirate.org/git/bcachefs

Pull bcachefs fixes from Kent Overstreet:
 "Assorted odds and ends...

   - two downgrade fixes

   - a couple snapshot deletion and repair fixes, thanks to noradtux for
     finding these and providing the image to debug them

   - a couple assert fixes

   - convert to folio helper, from Matthew

   - some improved error messages

   - bit of code reorganization (just moving things around); doing this
     while things are quiet so I'm not rebasing fixes past reorgs

   - don't return -EROFS on inconsistency error in recovery, this
     confuses util-linux and has it retry the mount

   - fix failure to return error on misaligned dio write; reported as an
     issue with coreutils shred"

* tag 'bcachefs-2024-05-30' of https://evilpiepirate.org/git/bcachefs: (21 commits)
  bcachefs: Fix failure to return error on misaligned dio write
  bcachefs: Don't return -EROFS from mount on inconsistency error
  bcachefs: Fix uninitialized var warning
  bcachefs: Split out sb-errors_format.h
  bcachefs: Split out journal_seq_blacklist_format.h
  bcachefs: Split out replicas_format.h
  bcachefs: Split out disk_groups_format.h
  bcachefs: split out sb-downgrade_format.h
  bcachefs: split out sb-members_format.h
  bcachefs: Better fsck error message for key version
  bcachefs: btree_gc can now handle unknown btrees
  bcachefs: add missing MODULE_DESCRIPTION()
  bcachefs: Fix setting of downgrade recovery passes/errors
  bcachefs: Run check_key_has_snapshot in snapshot_delete_keys()
  bcachefs: Refactor delete_dead_snapshots()
  bcachefs: Fix locking assert
  bcachefs: Fix lookup_first_inode() when inode_generations are present
  bcachefs: Plumb bkey into __btree_err()
  bcachefs: Use copy_folio_from_iter_atomic()
  bcachefs: Fix sb-downgrade validation
  ...

13 months agox86/topology/intel: Unlock CPUID before evaluating anything
Thomas Gleixner [Thu, 30 May 2024 15:29:18 +0000 (17:29 +0200)]
x86/topology/intel: Unlock CPUID before evaluating anything

Intel CPUs have a MSR bit to limit CPUID enumeration to leaf two. If
this bit is set by the BIOS then CPUID evaluation including topology
enumeration does not work correctly as the evaluation code does not try
to analyze any leaf greater than two.

This went unnoticed before because the original topology code just
repeated evaluation several times and managed to overwrite the initial
limited information with the correct one later. The new evaluation code
does it once and therefore ends up with the limited and wrong
information.

Cure this by unlocking CPUID right before evaluating anything which
depends on the maximum CPUID leaf being greater than two instead of
rereading stuff after unlock.

Fixes: 22d63660c35e ("x86/cpu: Use common topology code for Intel")
Reported-by: Peter Schneider <pschneider1968@googlemail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Tested-by: Peter Schneider <pschneider1968@googlemail.com>
Cc: <stable@kernel.org>
Link: https://lore.kernel.org/r/fd3f73dc-a86f-4bcf-9c60-43556a21eb42@googlemail.com
13 months agomailbox: zynqmp-ipi: drop irq_to_desc() call
Arnd Bergmann [Tue, 28 May 2024 12:02:35 +0000 (14:02 +0200)]
mailbox: zynqmp-ipi: drop irq_to_desc() call

irq_to_desc() is not exported to loadable modules, so this driver now
fails to link in some configurations:

ERROR: modpost: "irq_to_desc" [drivers/mailbox/zynqmp-ipi-mailbox.ko] undefined!

I can't see a purpose for this call, since the return value is unused
and probably left over from some code refactoring.

Address the link failure by just removing the line.

Fixes: 6ffb1635341b ("mailbox: zynqmp: handle SGI for shared IPI")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Tanmay Shah <tanmay.shah@amd.com>
Signed-off-by: Jassi Brar <jassisinghbrar@gmail.com>
13 months agomailmap: update entry for Kees Cook
Kees Cook [Fri, 31 May 2024 15:54:52 +0000 (08:54 -0700)]
mailmap: update entry for Kees Cook

I'm tired of gmail breaking DKIM. Switch everything over to my
@kernel.org alias instead.

Signed-off-by: Kees Cook <kees@kernel.org>
13 months agoscsi: mpt3sas: Avoid possible run-time warning with long manufacturer strings
Kees Cook [Wed, 10 Apr 2024 02:31:52 +0000 (19:31 -0700)]
scsi: mpt3sas: Avoid possible run-time warning with long manufacturer strings

The prior strscpy() replacement of strncpy() here expected the
manufacture_reply strings to be NUL-terminated, but it is possible
they are not, as the code pattern here shows, e.g., edev->vendor_id
being exactly 1 character larger than manufacture_reply->vendor_id,
and the replaced strncpy() was copying only up to the size of the
source character array. Replace this with memtostr(), which is the
unambiguous way to convert a maybe not-NUL-terminated character array
into a NUL-terminated string.

Fixes: b7e9712a02e8 ("scsi: mpt3sas: Replace deprecated strncpy() with strscpy()")
Signed-off-by: Kees Cook <keescook@chromium.org>
Tested-by: Marco Patalano <mpatalan@redhat.com>
Reviewed-by: Ewan D. Milne <emilne@redhat.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Link: https://lore.kernel.org/r/20240410023155.2100422-3-keescook@chromium.org
Signed-off-by: Kees Cook <kees@kernel.org>
13 months agocifs: fix creating sockets when using sfu mount options
Steve French [Wed, 29 May 2024 23:16:56 +0000 (18:16 -0500)]
cifs: fix creating sockets when using sfu mount options

When running fstest generic/423 with sfu mount option, it
was being skipped due to inability to create sockets:

  generic/423  [not run] cifs does not support mknod/mkfifo

which can also be easily reproduced with their af_unix tool:

  ./src/af_unix /mnt1/socket-two bind: Operation not permitted

Fix sfu mount option to allow creating and reporting sockets.

Cc: stable@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
13 months agoselftests/landlock: Add layout1.refer_mount_root
Mickaël Salaün [Thu, 16 May 2024 18:19:35 +0000 (20:19 +0200)]
selftests/landlock: Add layout1.refer_mount_root

Add tests to check error codes when linking or renaming a mount root
directory.  This previously triggered a kernel warning, but it is fixed
with the previous commit.

Cc: Günther Noack <gnoack@google.com>
Cc: Paul Moore <paul@paul-moore.com>
Link: https://lore.kernel.org/r/20240516181935.1645983-3-mic@digikod.net
Signed-off-by: Mickaël Salaün <mic@digikod.net>
13 months agolandlock: Fix d_parent walk
Mickaël Salaün [Thu, 16 May 2024 18:19:34 +0000 (20:19 +0200)]
landlock: Fix d_parent walk

The WARN_ON_ONCE() in collect_domain_accesses() can be triggered when
trying to link a root mount point.  This cannot work in practice because
this directory is mounted, but the VFS check is done after the call to
security_path_link().

Do not use source directory's d_parent when the source directory is the
mount point.

Cc: Günther Noack <gnoack@google.com>
Cc: Paul Moore <paul@paul-moore.com>
Cc: stable@vger.kernel.org
Reported-by: syzbot+bf4903dc7e12b18ebc87@syzkaller.appspotmail.com
Fixes: b91c3e4ea756 ("landlock: Add support for file reparenting with LANDLOCK_ACCESS_FS_REFER")
Closes: https://lore.kernel.org/r/000000000000553d3f0618198200@google.com
Link: https://lore.kernel.org/r/20240516181935.1645983-2-mic@digikod.net
[mic: Fix commit message]
Signed-off-by: Mickaël Salaün <mic@digikod.net>
13 months agoata: libata-core: Add ATA_HORKAGE_NOLPM for Apacer AS340
Niklas Cassel [Thu, 30 May 2024 21:27:04 +0000 (23:27 +0200)]
ata: libata-core: Add ATA_HORKAGE_NOLPM for Apacer AS340

Commit 7627a0edef54 ("ata: ahci: Drop low power policy board type")
dropped the board_ahci_low_power board type, and instead enables LPM if:
-The AHCI controller reports that it supports LPM (Partial/Slumber), and
-CONFIG_SATA_MOBILE_LPM_POLICY != 0, and
-The port is not defined as external in the per port PxCMD register, and
-The port is not defined as hotplug capable in the per port PxCMD
 register.

Partial and Slumber LPM states can either be initiated by HIPM or DIPM.

For HIPM (host initiated power management) to get enabled, both the AHCI
controller and the drive have to report that they support HIPM.

For DIPM (device initiated power management) to get enabled, only the
drive has to report that it supports DIPM. However, the HBA will reject
device requests to enter LPM states which the HBA does not support.

The problem is that Apacer AS340 drives do not handle low power modes
correctly. The problem was most likely not seen before because no one
had used this drive with a AHCI controller with LPM enabled.

Add a quirk so that we do not enable LPM for this drive, since we see
command timeouts if we do (even though the drive claims to support DIPM).

Fixes: 7627a0edef54 ("ata: ahci: Drop low power policy board type")
Cc: stable@vger.kernel.org
Reported-by: Tim Teichmann <teichmanntim@outlook.de>
Closes: https://lore.kernel.org/linux-ide/87bk4pbve8.ffs@tglx/
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Niklas Cassel <cassel@kernel.org>
13 months agoata: libata-core: Add ATA_HORKAGE_NOLPM for AMD Radeon S3 SSD
Niklas Cassel [Thu, 30 May 2024 21:32:44 +0000 (23:32 +0200)]
ata: libata-core: Add ATA_HORKAGE_NOLPM for AMD Radeon S3 SSD

Commit 7627a0edef54 ("ata: ahci: Drop low power policy board type")
dropped the board_ahci_low_power board type, and instead enables LPM if:
-The AHCI controller reports that it supports LPM (Partial/Slumber), and
-CONFIG_SATA_MOBILE_LPM_POLICY != 0, and
-The port is not defined as external in the per port PxCMD register, and
-The port is not defined as hotplug capable in the per port PxCMD
 register.

Partial and Slumber LPM states can either be initiated by HIPM or DIPM.

For HIPM (host initiated power management) to get enabled, both the AHCI
controller and the drive have to report that they support HIPM.

For DIPM (device initiated power management) to get enabled, only the
drive has to report that it supports DIPM. However, the HBA will reject
device requests to enter LPM states which the HBA does not support.

The problem is that AMD Radeon S3 SSD drives do not handle low power modes
correctly. The problem was most likely not seen before because no one
had used this drive with a AHCI controller with LPM enabled.

Add a quirk so that we do not enable LPM for this drive, since we see
command timeouts if we do (even though the drive claims to support both
HIPM and DIPM).

Fixes: 7627a0edef54 ("ata: ahci: Drop low power policy board type")
Cc: stable@vger.kernel.org
Reported-by: Doru Iorgulescu <doru.iorgulescu1@gmail.com>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218832
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Niklas Cassel <cassel@kernel.org>
13 months agoata: libata-core: Add ATA_HORKAGE_NOLPM for Crucial CT240BX500SSD1
Niklas Cassel [Thu, 30 May 2024 21:28:17 +0000 (23:28 +0200)]
ata: libata-core: Add ATA_HORKAGE_NOLPM for Crucial CT240BX500SSD1

Commit 7627a0edef54 ("ata: ahci: Drop low power policy board type")
dropped the board_ahci_low_power board type, and instead enables LPM if:
-The AHCI controller reports that it supports LPM (Partial/Slumber), and
-CONFIG_SATA_MOBILE_LPM_POLICY != 0, and
-The port is not defined as external in the per port PxCMD register, and
-The port is not defined as hotplug capable in the per port PxCMD
 register.

Partial and Slumber LPM states can either be initiated by HIPM or DIPM.

For HIPM (host initiated power management) to get enabled, both the AHCI
controller and the drive have to report that they support HIPM.

For DIPM (device initiated power management) to get enabled, only the
drive has to report that it supports DIPM. However, the HBA will reject
device requests to enter LPM states which the HBA does not support.

The problem is that Crucial CT240BX500SSD1 drives do not handle low power
modes correctly. The problem was most likely not seen before because no
one had used this drive with a AHCI controller with LPM enabled.

Add a quirk so that we do not enable LPM for this drive, since we see
command timeouts if we do (even though the drive claims to support DIPM).

Fixes: 7627a0edef54 ("ata: ahci: Drop low power policy board type")
Cc: stable@vger.kernel.org
Reported-by: Aarrayy <lp610mh@gmail.com>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218832
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Niklas Cassel <cassel@kernel.org>
13 months agosched/x86: Export 'percpu arch_freq_scale'
Phil Auld [Thu, 30 May 2024 18:15:48 +0000 (14:15 -0400)]
sched/x86: Export 'percpu arch_freq_scale'

Commit:

  7bc263840bc3 ("sched/topology: Consolidate and clean up access to a CPU's max compute capacity")

removed rq->cpu_capacity_orig in favor of using arch_scale_freq_capacity()
calls. Export the underlying percpu symbol on x86 so that external trace
point helper modules can be made to work again.

Signed-off-by: Phil Auld <pauld@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240530181548.2039216-1-pauld@redhat.com
13 months agoperf/x86/intel: Add missing MODULE_DESCRIPTION() lines
Jeff Johnson [Thu, 30 May 2024 20:42:51 +0000 (13:42 -0700)]
perf/x86/intel: Add missing MODULE_DESCRIPTION() lines

Fix the 'make W=1 C=1' warnings:

  WARNING: modpost: missing MODULE_DESCRIPTION() in arch/x86/events/intel/intel-uncore.o
  WARNING: modpost: missing MODULE_DESCRIPTION() in arch/x86/events/intel/intel-cstate.o

Signed-off-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Link: https://lore.kernel.org/r/20240530-md-arch-x86-events-intel-v1-1-8252194ed20a@quicinc.com
13 months agoperf/x86/rapl: Add missing MODULE_DESCRIPTION() line
Jeff Johnson [Thu, 30 May 2024 20:12:03 +0000 (13:12 -0700)]
perf/x86/rapl: Add missing MODULE_DESCRIPTION() line

Fix the warning from 'make C=1 W=1':

  WARNING: modpost: missing MODULE_DESCRIPTION() in arch/x86/events/rapl.o

Signed-off-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Link: https://lore.kernel.org/r/20240530-md-arch-x86-events-v1-1-e45ffa8af99f@quicinc.com