linux-toradex.git/include/linux/interrupt.h, branch v7.0-rc6

Merge tag 'irq-core-2026-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2026-02-10T21:39:37+00:00

Pull irq core updates from Thomas Gleixner:
 "Updates for the interrupt core subsystem:

   - Remove the interrupt timing infrastructure

     This was added seven years ago to be used for power management
     purposes, but that integration never happened.

   - Clean up the remaining setup_percpu_irq() users

     The memory allocator is available when interrupts can be requested
     so there is not need for static irq_action. Move the remaining
     users to request_percpu_irq() and delete the historical cruft.

   - Warn when interrupt flag inconsistencies are detected in
     request*_irq().

     Inconsistent flags can lead to hard to diagnose malfunction. The
     fallout of this new warning has been addressed in next and the
     fixes are coming in via the maintainer trees and the tip
     irq/cleanup pull requests.

   - Invoke affinity notifier when CPU hotplug breaks affinity

     Otherwise the code using the notifier misses the affinity change
     and operates on stale information.

   - The usual cleanups and improvements"

* tag 'irq-core-2026-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  genirq/proc: Replace snprintf with strscpy in register_handler_proc
  genirq/cpuhotplug: Notify about affinity changes breaking the affinity mask
  genirq: Move clear of kstat_irqs to free_desc()
  genirq: Warn about using IRQF_ONESHOT without a threaded handler
  irqdomain: Fix up const problem in irq_domain_set_name()
  genirq: Remove setup_percpu_irq()
  clocksource/drivers/mips-gic-timer: Move GIC timer to request_percpu_irq()
  MIPS: Move IP27 timer to request_percpu_irq()
  MIPS: Move IP30 timer to request_percpu_irq()
  genirq: Remove __request_percpu_irq() helper
  genirq: Remove IRQ timing tracking infrastructure

genirq: Set IRQF_COND_ONESHOT in devm_request_irq().

2026-02-01T16:37:13+00:00

The flag IRQF_COND_ONESHOT was already force-added to request_irq() because
the ACPI SCI interrupt handler is using the IRQF_ONESHOT flag which breaks
all shared handlers.

devm_request_irq() needs the same change since some users, such as
int0002_vgpio, are using this function instead.

Add IRQF_COND_ONESHOT to the flags passed to devm_request_irq().

Fixes: c37927a203fa2 ("genirq: Set IRQF_COND_ONESHOT in request_irq()")
Signed-off-by: Sebastian Andrzej Siewior 
Signed-off-by: Thomas Gleixner 
Link: https://patch.msgid.link/20260128095540.863589-2-bigeasy@linutronix.de

genirq: Remove __request_percpu_irq() helper

2025-12-15T21:20:50+00:00

With the IRQ timing stuff being gone, there is no need to specify a flag
when requesting a percpu interrupt. Not only IRQF_TIMER was the only flag
(set of flags actually) allowed, but nobody ever passed it.

Get rid of __request_percpu_irq(), which was only getting 0 as flags, and
promote request_percpu_irq_affinity() as its replacement.

Signed-off-by: Marc Zyngier 
Signed-off-by: Thomas Gleixner 
Reviewed-by: Jinjie Ruan 
Link: https://patch.msgid.link/20251210082242.360936-3-maz@kernel.org

genirq: Remove IRQ timing tracking infrastructure

2025-12-15T21:20:50+00:00

The IRQ timing tracking infrastructure was merged in 2019, but was never
plumbed in, is not selectable, and is therefore never used.

As Daniel agrees that there is little hope for this infrastructure to be
completed in the near term, drop it altogether.

Signed-off-by: Marc Zyngier 
Signed-off-by: Thomas Gleixner 
Reviewed-by: Jinjie Ruan 
Link: https://lore.kernel.org/r/87zf7vex6h.wl-maz@kernel.org
Link: https://patch.msgid.link/20251210082242.360936-2-maz@kernel.org

genirq: Fix percpu_devid irq affinity documentation

2025-10-31T21:25:34+00:00

Stephen points out that some of the percpu_devid irq affinity
documentation is either missing or not matching the data structures.

Address all the issues in one go.

Fixes: 87b0031f7f73 ("irqdomain: Add firmware info reporting interface")
Fixes: 258e7d28a3dc ("genirq: Add affinity to percpu_devid interrupt requests")
Reported-by: Stephen Rothwell 
Signed-off-by: Marc Zyngier 
Signed-off-by: Thomas Gleixner 
Link: https://patch.msgid.link/20251030143032.2035987-1-maz@kernel.org

genirq: Add request_percpu_irq_affinity() helper

2025-10-27T16:16:35+00:00

While it would be nice to simply make request_percpu_irq() take an affinity
mask, the churn is likely to be on the irritating side given that most
drivers do not give a damn about affinities.

So take the more innocuous path to provide a helper that parallels
request_percpu_irq(), with an affinity as a bonus argument.

Yes, request_percpu_irq_affinity() is a bit of a mouthful.

Signed-off-by: Marc Zyngier 
Signed-off-by: Thomas Gleixner 
Tested-by: Will Deacon 
Link: https://patch.msgid.link/20251020122944.3074811-18-maz@kernel.org

genirq: Update request_percpu_nmi() to take an affinity

2025-10-27T16:16:35+00:00

Continue spreading the notion of affinity to the per CPU interrupt request
code by updating the call sites that use request_percpu_nmi() (all two of
them) to take an affinity pointer. This pointer is firmly NULL for now.

Signed-off-by: Marc Zyngier 
Signed-off-by: Thomas Gleixner 
Tested-by: Will Deacon 
Link: https://patch.msgid.link/20251020122944.3074811-16-maz@kernel.org

genirq: Add affinity to percpu_devid interrupt requests

2025-10-27T16:16:34+00:00

Add an affinity field to both the irqaction structure and the interrupt
request primitives. Nothing is making use of it yet, and the only value
used it NULL, which is used as a shorthand for cpu_possible_mask.

This will shortly get used with actual affinities.

Signed-off-by: Marc Zyngier 
Signed-off-by: Thomas Gleixner 
Tested-by: Will Deacon 
Link: https://patch.msgid.link/20251020122944.3074811-15-maz@kernel.org

genirq: Merge irqaction::{dev_id,percpu_dev_id}

2025-10-27T16:16:34+00:00

When irqaction::percpu_dev_id was introduced, it was hoped that it could be
part of an anonymous union with dev_id, as the two fields are mutually
exclusive.

However, toolchains used at the time were often showing terrible support
for anonymous unions, breaking the build on a number of architectures. It
was therefore decided to keep the two fields separate and address this down
the line.

14 years later, the compiler dark age is over, and there is universal
support for anonymous unions. Get a whole pointer back that can immediately
be spent on something else.

Signed-off-by: Marc Zyngier 
Signed-off-by: Thomas Gleixner 
Tested-by: Will Deacon 
Link: https://patch.msgid.link/20251020122944.3074811-13-maz@kernel.org

genirq: Fix typo in IRQ_NOTCONNECTED comment

2025-04-24T12:31:03+00:00

Fix a minor typo in the comment for IRQ_NOTCONNECTED:
"distingiush" is corrected to "distinguish".

Signed-off-by: Cheng-Yang Chou 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/all/20250410105144.214849-1-yphbchou0911@gmail.com