linux-2.6-block.git
10 months agogenirq: Use cpumask_intersects()
Costa Shulyupin [Fri, 6 Sep 2024 17:01:40 +0000 (20:01 +0300)]
genirq: Use cpumask_intersects()

Replace `cpumask_any_and(a, b) >= nr_cpu_ids` and `cpumask_any_and(a, b) <
nr_cpu_ids` with the more readable `!cpumask_intersects(a, b)` and
`cpumask_intersects(a, b)`

Signed-off-by: Costa Shulyupin <costa.shul@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240906170142.1135207-1-costa.shul@redhat.com
10 months agogenirq/cpuhotplug: Use cpumask_intersects()
Costa Shulyupin [Wed, 4 Sep 2024 13:48:23 +0000 (16:48 +0300)]
genirq/cpuhotplug: Use cpumask_intersects()

Replace `cpumask_any_and(a, b) >= nr_cpu_ids`
with the more readable `!cpumask_intersects(a, b)`.

[ tglx: Massaged change log ]

Signed-off-by: Costa Shulyupin <costa.shul@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/all/20240904134823.777623-2-costa.shul@redhat.com
10 months agoirqchip/apple-aic: Only access system registers on SoCs which provide them
Konrad Dybcio [Sun, 1 Sep 2024 03:40:07 +0000 (11:40 +0800)]
irqchip/apple-aic: Only access system registers on SoCs which provide them

Starting from the A11 (T8015) SoC, Apple introuced system registers for
fast IPI and UNCORE PMC control. These sysregs do not exist on earlier
A7-A10 SoCs and trying to access them results in an instant crash.

Restrict sysreg access within the AIC driver to configurations where
use_fast_ipi is true to allow AIC to function properly on A7-A10 SoCs.

Co-developed-by: Nick Chan <towinchenmi@gmail.com>
Signed-off-by: Konrad Dybcio <konrad.dybcio@somainline.org>
Signed-off-by: Nick Chan <towinchenmi@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Sven Peter <sven@svenpeter.dev>
Link: https://lore.kernel.org/all/20240901034143.12731-5-towinchenmi@gmail.com
10 months agoirqchip/apple-aic: Add a new "Global fast IPIs only" feature level
Nick Chan [Sun, 1 Sep 2024 03:40:06 +0000 (11:40 +0800)]
irqchip/apple-aic: Add a new "Global fast IPIs only" feature level

Starting with the A11 (T8015) SoC, Apple began using arm64 sysregs for
fast IPIs. However, on A11, there is no such things as "Local" fast IPIs,
as the SYS_IMP_APL_IPI_RR_LOCAL_EL1 register does not seem to exist.

Add a new feature level, used by the compatible "apple,t8015-aic",
controlled by a static branch key named use_local_fast_ipi. When
use_fast_ipi is true and use_local_fast_ipi is false, fast IPIs are used
but all IPIs goes through the register SYS_IMP_APL_IPI_RR_GLOBAL_EL1, as
"global" IPIs.

Signed-off-by: Nick Chan <towinchenmi@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Sven Peter <sven@svenpeter.dev>
Link: https://lore.kernel.org/all/20240901034143.12731-4-towinchenmi@gmail.com
10 months agoirqchip/apple-aic: Skip unnecessary enabling of use_fast_ipi
Nick Chan [Sun, 1 Sep 2024 03:40:05 +0000 (11:40 +0800)]
irqchip/apple-aic: Skip unnecessary enabling of use_fast_ipi

use_fast_ipi is true by default and there is no need to "enable" it.

Signed-off-by: Nick Chan <towinchenmi@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Sven Peter <sven@svenpeter.dev>
Link: https://lore.kernel.org/all/20240901034143.12731-3-towinchenmi@gmail.com
10 months agodt-bindings: apple,aic: Document A7-A11 compatibles
Nick Chan [Sun, 1 Sep 2024 03:40:04 +0000 (11:40 +0800)]
dt-bindings: apple,aic: Document A7-A11 compatibles

Document and describe the compatibles for Apple A7-A11 SoCs.
There are three feature levels:

 - apple,aic: No fast IPI, for A7-A10
 - apple,t8015-aic: fast IPI, global only, for A11
 - apple,t8103-aic: fast IPI with local and global support, for M1

Each feature level is an extension of the previous, for example, M1 will
also work with the A7 feature level.

All of A7-M1 gets its own SoC-specific compatible, and the "apple,aic"
compatible as a fallback.

Signed-off-by: Nick Chan <towinchenmi@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Sven Peter <sven@svenpeter.dev>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/all/20240901034143.12731-2-towinchenmi@gmail.com
10 months agoirqdomain: Use IS_ERR_OR_NULL() in irq_domain_trim_hierarchy()
Hongbo Li [Wed, 28 Aug 2024 12:27:24 +0000 (20:27 +0800)]
irqdomain: Use IS_ERR_OR_NULL() in irq_domain_trim_hierarchy()

Use IS_ERR_OR_NULL() instead of open-coding a NULL and a error pointer
check.

Signed-off-by: Hongbo Li <lihongbo22@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240828122724.3697447-1-lihongbo22@huawei.com
10 months agogenirq/msi: Use kmemdup_array() instead of kmemdup()
Jinjie Ruan [Wed, 28 Aug 2024 07:22:19 +0000 (15:22 +0800)]
genirq/msi: Use kmemdup_array() instead of kmemdup()

Let kmemdup_array() take care about sizing instead of doing it open coded.

Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240828072219.1249250-1-ruanjinjie@huawei.com
10 months agogenirq/proc: Change the return value for set affinity permission error
Jeff Xie [Mon, 26 Aug 2024 14:58:05 +0000 (22:58 +0800)]
genirq/proc: Change the return value for set affinity permission error

Currently, when the affinity of an irq cannot be set due to lack of
permission, the write_irq_affinity() returns the error code -EIO.

Change the return value to -EPERM as that reflects the cause of error
correctly.

Signed-off-by: Jeff Xie <jeff.xie@linux.dev>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240826145805.5938-1-jeff.xie@linux.dev
10 months agogenirq/proc: Use irq_move_pending() in show_irq_affinity()
Jinjie Ruan [Thu, 29 Aug 2024 11:15:22 +0000 (19:15 +0800)]
genirq/proc: Use irq_move_pending() in show_irq_affinity()

irq_move_pending() encapsulates irqd_is_setaffinity_pending() depending on
CONFIG_GENERIC_PENDING_IRQ.

Replace the open coded #ifdeffery with it.

Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240829111522.230595-1-ruanjinjie@huawei.com
10 months agogenirq/proc: Correctly set file permissions for affinity control files
Jeff Xie [Sun, 25 Aug 2024 13:19:11 +0000 (21:19 +0800)]
genirq/proc: Correctly set file permissions for affinity control files

The kernel already knows at the time of interrupt allocation whether
affinity of an interrupt can be controlled by userspace or not.

It still creates all related procfs control files with read/write
permissions. That's inconsistent and non-intuitive for system
administrators and tools.

Therefore set the file permissions to read-only for such interrupts.

[ tglx: Massage change log, fixed UP build ]

Signed-off-by: Jeff Xie <jeff.xie@linux.dev>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240825131911.107119-1-jeff.xie@linux.dev
10 months agogenirq: Get rid of global lock in irq_do_set_affinity()
Marc Zyngier [Mon, 26 Aug 2024 08:06:18 +0000 (09:06 +0100)]
genirq: Get rid of global lock in irq_do_set_affinity()

Kunkun Jiang reports that for a workload involving the simultaneous startup
of a large number of VMs (for a total of about 200 vcpus), a lot of CPU
time gets spent on spinning on the tmp_mask_lock that exists as a static
raw spinlock in irq_do_set_affinity(). This lock protects a global cpumask
(tmp_mask) that is used as a temporary variable to compute the resulting
affinity.

While this is triggered by KVM issuing a irq_set_affinity() call each time
a vcpu is about to execute, it is obvious that having a single global
resource is not very scalable.

Since a cpumask can be a fairly large structure on systems with a high core
count, a stack allocation is not really appropriate.  Instead, turn the
global cpumask into a per-CPU variable, removing the need for locking
altogether as the code is executed with preemption and interrupts disabled.

[ tglx: Moved the per CPU variable declaration outside of the function ]

Reported-by: Kunkun Jiang <jiangkunkun@huawei.com>
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Kunkun Jiang <jiangkunkun@huawei.com>
Link: https://lore.kernel.org/all/20240826080618.3886694-1-maz@kernel.org
Link: https://lore.kernel.org/all/a7fc58e4-64c2-77fc-c1dc-f5eb78dbbb01@huawei.com
10 months agogenirq: Fix typo in struct comment
Costa Shulyupin [Thu, 22 Aug 2024 12:31:58 +0000 (15:31 +0300)]
genirq: Fix typo in struct comment

Remove redundant "e" in "assign(e)ments".

Signed-off-by: Costa Shulyupin <costa.shul@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240822123205.2186221-1-costa.shul@redhat.com
10 months agoirqchip/loongarch-avec: Add AVEC irqchip support
Tianyang Zhang [Fri, 23 Aug 2024 10:43:37 +0000 (18:43 +0800)]
irqchip/loongarch-avec: Add AVEC irqchip support

Introduce the advanced extended interrupt controllers (AVECINTC). 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.

The whole topology of irqchips in LoongArch machines looks like this if
AVECINTC is supported:

  +-----+     +-----------------------+     +-------+
  | IPI | --> |        CPUINTC        | <-- | Timer |
  +-----+     +-----------------------+     +-------+
               ^          ^          ^
               |          |          |
        +---------+ +----------+ +---------+     +-------+
        | EIOINTC | | AVECINTC | | LIOINTC | <-- | UARTs |
        +---------+ +----------+ +---------+     +-------+
             ^            ^
             |            |
        +---------+  +---------+
        | PCH-PIC |  | PCH-MSI |
        +---------+  +---------+
          ^     ^           ^
          |     |           |
  +---------+ +---------+ +---------+
  | Devices | | PCH-LPC | | Devices |
  +---------+ +---------+ +---------+
                   ^
                   |
              +---------+
              | Devices |
              +---------+

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>
Co-developed-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Tianyang Zhang <zhangtianyang@loongson.cn>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240823104337.25577-2-zhangtianyang@loongson.cn
10 months agoirqchip/loongson-pch-msi: Prepare get_pch_msi_handle() for AVECINTC
Tianyang Zhang [Fri, 23 Aug 2024 10:43:36 +0000 (18:43 +0800)]
irqchip/loongson-pch-msi: Prepare get_pch_msi_handle() for AVECINTC

On Loongson-3C6000 and higher systems with AVECINTC irqchip, there may
be multiple PCI segments, but only one PCH-MSI irq domain. In this case,
let get_pch_msi_handle() return the first domain handle.

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>
Co-developed-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Tianyang Zhang <zhangtianyang@loongson.cn>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240823104337.25577-1-zhangtianyang@loongson.cn
10 months agoirqchip/loongson-eiointc: Rename CPUHP_AP_IRQ_LOONGARCH_STARTING
Huacai Chen [Fri, 23 Aug 2024 10:39:34 +0000 (18:39 +0800)]
irqchip/loongson-eiointc: Rename CPUHP_AP_IRQ_LOONGARCH_STARTING

Rename CPUHP_AP_IRQ_LOONGARCH_STARTING to CPUHP_AP_IRQ_EIOINTC_STARTING
because the upcoming AVECINTC irqchip driver will introduce a new state
and so both are clearly identifiable.

Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Tianyang Zhang <zhangtianyang@loongson.cn>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240823103936.25092-3-zhangtianyang@loongson.cn
10 months agoLoongArch: Architectural preparation for AVEC irqchip
Huacai Chen [Fri, 23 Aug 2024 10:39:33 +0000 (18:39 +0800)]
LoongArch: Architectural preparation for AVEC irqchip

Add architectural preparation for AVEC irqchip, including:
1. CPUCFG feature bits definition for AVEC;
2. Detection of AVEC irqchip in cpu_probe();
3. New IPI type definition (IPI_CLEAR_VECTOR) for AVEC;
4. Provide arch_probe_nr_irqs() for large NR_IRQS;
5. Other related changes about the number of interrupts.

Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Tianyang Zhang <zhangtianyang@loongson.cn>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240823103936.25092-2-zhangtianyang@loongson.cn
10 months agoLoongArch: Move irqchip function prototypes to irq-loongson.h
Huacai Chen [Fri, 23 Aug 2024 10:39:32 +0000 (18:39 +0800)]
LoongArch: Move irqchip function prototypes to irq-loongson.h

Some irqchip functions are only for internal use by irqchip drivers, so
move their prototypes from asm/irq.h to drivers/irqchip/irq-loongson.h.

All related driver files include the new irq-loongson.h.

Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Tianyang Zhang <zhangtianyang@loongson.cn>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240823103936.25092-1-zhangtianyang@loongson.cn
10 months agoirqchip/loongson-pch-msi: Switch to MSI parent domains
Huacai Chen [Thu, 15 Aug 2024 11:26:07 +0000 (19:26 +0800)]
irqchip/loongson-pch-msi: Switch to MSI parent domains

Remove the global PCI/MSI irqdomain implementation and provide the
required MSI parent functionality by filling in msi_parent_ops, so the
PCI/MSI code can detect the new parent and setup per-device MSI domains.

Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Tianyang Zhang <zhangtianyang@loongson.cn>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240815112608.26925-2-zhangtianyang@loongson.cn
10 months agosoftirq: Remove unused 'action' parameter from action callback
Caleb Sander Mateos [Thu, 15 Aug 2024 17:15:40 +0000 (11:15 -0600)]
softirq: Remove unused 'action' parameter from action callback

When soft interrupt actions are called, they are passed a pointer to the
struct softirq action which contains the action's function pointer.

This pointer isn't useful, as the action callback already knows what
function it is. And since each callback handles a specific soft interrupt,
the callback also knows which soft interrupt number is running.

No soft interrupt action callback actually uses this parameter, so remove
it from the function pointer signature. This clarifies that soft interrupt
actions are global routines and makes it slightly cheaper to call them.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Link: https://lore.kernel.org/all/20240815171549.3260003-1-csander@purestorage.com
10 months agoirqdomain: Always associate interrupts for legacy domains
Matti Vaittinen [Tue, 13 Aug 2024 11:34:27 +0000 (14:34 +0300)]
irqdomain: Always associate interrupts for legacy domains

The unification of irq_domain_create_legacy() missed the fact that
interrupts must be associated even when the Linux interrupt number provided
in the first_irq argument is 0.

This breaks all call sites of irq_domain_create_legacy() which supply 0 as
the first_irq argument.

Enforce the association for legacy domains in __irq_domain_instantiate() to
cure this.

[ tglx: Massaged it slightly. ]

Fixes: 70114e7f7585 ("irqdomain: Simplify simple and legacy domain creation")
Reported-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Signed-off-by Matti Vaittinen <mazziesaccount@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Link: https://lore.kernel.org/all/c3379142-10bc-4f14-b8ac-a46927aeac38@gmail.com
11 months agoirqdomain: Remove stray '-' in the domain name
Andy Shevchenko [Mon, 12 Aug 2024 19:29:40 +0000 (22:29 +0300)]
irqdomain: Remove stray '-' in the domain name

When the domain suffix is not supplied alloc_fwnode_name() unconditionally
adds a separator.

Fix the format strings to get rid of the stray '-' separator.

Fixes: 1e7c05292531 ("irqdomain: Allow giving name suffix for domain")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240812193101.1266625-3-andriy.shevchenko@linux.intel.com
11 months agoirqdomain: Clarify checks for bus_token
Andy Shevchenko [Mon, 12 Aug 2024 19:29:39 +0000 (22:29 +0300)]
irqdomain: Clarify checks for bus_token

The code uses if (bus_token) and if (bus_token == DOMAIN_BUS_ANY).

Since bus_token is an enum, the latter is more robust against changes.

Convert all !bus_token checks to explicitely check for DOMAIN_BUS_ANY.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240812193101.1266625-2-andriy.shevchenko@linux.intel.com
11 months agoMerge tag 'irq-domain-24-08-09' into irq/core
Thomas Gleixner [Fri, 9 Aug 2024 21:04:41 +0000 (23:04 +0200)]
Merge tag 'irq-domain-24-08-09' into irq/core

Merge the irqdomain changes which are required for regmap to apply
depending patches and therefore tagged.

11 months agoirqdomain: Allow giving name suffix for domain
Matti Vaittinen [Thu, 8 Aug 2024 20:23:06 +0000 (22:23 +0200)]
irqdomain: Allow giving name suffix for domain

Devices can provide multiple interrupt lines. One reason for this is that
a device has multiple subfunctions, each providing its own interrupt line.
Another reason is that a device can be designed to be used (also) on a
system where some of the interrupts can be routed to another processor.

A line often further acts as a demultiplex for specific interrupts
and has it's respective set of interrupt (status, mask, ack, ...)
registers.

Regmap supports the handling of these registers and demultiplexing
interrupts, but the interrupt domain code ends up assigning the same name
for the per interrupt line domains. This causes a naming collision in the
debugFS code and leads to confusion, as /proc/interrupts shows two separate
interrupts with the same domain name and hardware interrupt number.

Instead of adding a workaround in regmap or driver code, allow giving a
name suffix for the domain name when the domain is created.

Add a name_suffix field in the irq_domain_info structure and make
irq_domain_instantiate() use this suffix if it is given when a domain is
created.

[ tglx: Adopt it to the cleanup patch and fixup the invalid NULL return ]

Signed-off-by: Matti Vaittinen <mazziesaccount@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/871q2yvk5x.ffs@tglx
11 months agoirqdomain: Cleanup domain name allocation
Thomas Gleixner [Thu, 8 Aug 2024 20:19:41 +0000 (22:19 +0200)]
irqdomain: Cleanup domain name allocation

irq_domain_set_name() is truly unreadable gunk. Clean it up before adding
more.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Matti Vaittinen <mazziesaccount@gmail.com>
Link: https://lore.kernel.org/all/874j7uvkbm.ffs@tglx
11 months agoirqdomain: Simplify simple and legacy domain creation
Matti Vaittinen [Thu, 8 Aug 2024 12:34:02 +0000 (15:34 +0300)]
irqdomain: Simplify simple and legacy domain creation

irq_domain_create_simple() and irq_domain_create_legacy() use
__irq_domain_instantiate(), but have extra handling of allocating interrupt
descriptors and associating interrupts in them. Some of that is duplicated.

There are also call sites which have conditonals to invoke different
interrupt domain creator functions, where one of them is usually
irq_domain_create_legacy(). Alternatively they associate the interrupts for
the legacy case after creating the domain.

Moving the extra logic of irq_domain_create_simple()/legacy() into
__irq_domain_instantiate() allows to consolidate that.

Introduce hwirq_base and virq_base members in the irq_domain_info
structure, which allows to transport the required information and add the
conditional interrupt descriptor allocation and interrupt association into
__irq_domain_instantiate().

This reduces irq_domain_create_legacy() and irq_domain_create_simple() to
trivial wrappers which fill in the info structure and allows call sites
which must support the legacy case along with more modern mechanism to
select the domain type via the parameters of the info struct.

[ tglx: Massaged change log ]

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Matti Vaittinen <mazziesaccount@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/32d07bd79eb2b5416e24da9e9e8fe5955423dcf9.1723120028.git.mazziesaccount@gmail.com
11 months agogenirq: Remove irq_chip_regs:: Polarity
Jiri Slaby (SUSE) [Thu, 8 Aug 2024 10:41:18 +0000 (12:41 +0200)]
genirq: Remove irq_chip_regs:: Polarity

The polarity member of struct irq_chip_regs is unused. Remove it along
with its kernel-doc.

Found by https://github.com/jirislaby/clang-struct.

Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240808104118.430670-3-jirislaby@kernel.org
11 months agogenirq: Remove unused irq_chip_generic:: {type,polarity}_cache
Jiri Slaby (SUSE) [Thu, 8 Aug 2024 10:41:17 +0000 (12:41 +0200)]
genirq: Remove unused irq_chip_generic:: {type,polarity}_cache

The type_cache and polarity_cache members of struct irq_chip_generic are
unused. Remove them both along with their kernel-doc.

Found by https://github.com/jirislaby/clang-struct.

Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240808104118.430670-2-jirislaby@kernel.org
11 months agogenirq/irq_sim: Remove unused irq_sim_work_ctx:: Irq_base
Jiri Slaby (SUSE) [Thu, 8 Aug 2024 10:41:16 +0000 (12:41 +0200)]
genirq/irq_sim: Remove unused irq_sim_work_ctx:: Irq_base

Since commit 337cbeb2c13e ("genirq/irq_sim: Simplify the API"),
irq_sim_work_ctx::irq_base is unused. Drop it.

Found by https://github.com/jirislaby/clang-struct.

Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Link: https://lore.kernel.org/all/20240808104118.430670-1-jirislaby@kernel.org
11 months agoirqchip/mbigen: Simplify code logic with for_each_child_of_node_scoped()
Zhang Zekun [Thu, 8 Aug 2024 03:15:52 +0000 (11:15 +0800)]
irqchip/mbigen: Simplify code logic with for_each_child_of_node_scoped()

for_each_child_of_node_scoped() handles the device_node automaticlly, so
switching over to it removes the device node cleanups and allows to return
directly from the loop.

Signed-off-by: Zhang Zekun <zhangzekun11@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240808031552.3156-1-zhangzekun11@huawei.com
11 months agoirqchip/armada-370-xp: Use mpic_is_ipi_available() in mpic_of_init()
Marek Behún [Wed, 7 Aug 2024 16:41:04 +0000 (18:41 +0200)]
irqchip/armada-370-xp: Use mpic_is_ipi_available() in mpic_of_init()

mpic_of_init() contains the last case where the open coded IPI support
condition needs to be replaced with mpic_is_ipi_available() to keep the
code consistent.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Allow mapping only per-CPU interrupts
Marek Behún [Wed, 7 Aug 2024 16:41:03 +0000 (18:41 +0200)]
irqchip/armada-370-xp: Allow mapping only per-CPU interrupts

On platforms where MPIC is not the top-level interrupt controller the
driver currently only supports handling of the per-CPU interrupts (the
first 29 interrupts). This is obvious from the code of
mpic_handle_cascade_irq(), which reads only one cause register.

Bound the number of available interrupts in the interrupt domain to 29 for
these platforms.

The corresponding device-trees refer only to per-CPU interrupts via MPIC,
the other interrupts are referred to via GIC.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Iterate only valid bits of the per-CPU interrupt cause register
Marek Behún [Wed, 7 Aug 2024 16:41:02 +0000 (18:41 +0200)]
irqchip/armada-370-xp: Iterate only valid bits of the per-CPU interrupt cause register

Use MPIC_PER_CPU_IRQS_NR (29) bound instead of BITS_PER_LONG (32) when
iterating the bits of the per-CPU interrupt cause register, since there
are only 29 per-CPU interrupts. The top 3 bits are always zero anyway.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Fix reenabling last per-CPU interrupt
Marek Behún [Wed, 7 Aug 2024 16:41:01 +0000 (18:41 +0200)]
irqchip/armada-370-xp: Fix reenabling last per-CPU interrupt

The number of per-CPU interrupts is 29 (0 to 28). This is described by
the constant MPIC_MAX_PER_CPU_IRQS, set to 28 (the maximum per-CPU
interrupt).

Commit 0fa4ce746d1d ("irqchip/armada-370-xp: Re-enable per-CPU
interrupts at resume time") used the constant incorrectly in the
for-loop, it used the operator < instead of <=, causing it to iterate
only the first 28 interrupts (0 to 27), ignoring the last, 28th,
per-CPU interrupt.

To avoid this kind of confusions, fix this issue by renaming the constant
to MPIC_PER_CPU_IRQS_NR and set it to 29, the number of per-CPU IRQs.
Update its use in mpic_is_percpu_irq() accordingly.

Fixes: 0fa4ce746d1d ("irqchip/armada-370-xp: Re-enable per-CPU interrupts at resume time")
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: <stable+noautosel@kernel.org> # The 29th interrupt is not used in any device-tree
11 months agoirqchip/armada-370-xp: Dynamically allocate the driver private structure
Marek Behún [Wed, 7 Aug 2024 16:41:00 +0000 (18:41 +0200)]
irqchip/armada-370-xp: Dynamically allocate the driver private structure

Dynamically allocate the driver private structure. This concludes the
conversion of this driver to modern style.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Pass around the driver private structure
Marek Behún [Wed, 7 Aug 2024 16:40:59 +0000 (18:40 +0200)]
irqchip/armada-370-xp: Pass around the driver private structure

In continuation of converting the driver to modern style, drop the
global pointer to the driver private structure and instead pass it
around the functions and callbacks, wherever possible. (There are 3
cases where it is not possible: mpic_cascaded_starting_cpu() and the
syscore operations mpic_suspend() and mpic_resume()).

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Put MSI doorbell limits into the mpic structure
Marek Behún [Wed, 7 Aug 2024 16:40:58 +0000 (18:40 +0200)]
irqchip/armada-370-xp: Put MSI doorbell limits into the mpic structure

Put the MSI doorbell limits msi_doorbell_start, msi_doorbell_size and
msi_doorbell_mask into the driver private structure and get rid of the
corresponding functions.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Put static variables into driver private structure
Marek Behún [Wed, 7 Aug 2024 16:40:57 +0000 (18:40 +0200)]
irqchip/armada-370-xp: Put static variables into driver private structure

In preparation for converting the driver to modern style put all the
interrupt controller private static variables into driver private
structure.

Access to these variables changes as:
  main_int_base mpic->base
  per_cpu_int_base mpic->per_cpu
  mpic_domain mpic->domain
  parent_irq mpic->parent_irq
  ...

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Put __init attribute after return type in mpic_ipi_init()
Marek Behún [Wed, 7 Aug 2024 16:40:56 +0000 (18:40 +0200)]
irqchip/armada-370-xp: Put __init attribute after return type in mpic_ipi_init()

For consistency with the rest of the driver, put the __init attribute
after the return type of the mpic_ipi_init() function.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Add the __init attribute to mpic_msi_init()
Marek Behún [Wed, 7 Aug 2024 16:40:55 +0000 (18:40 +0200)]
irqchip/armada-370-xp: Add the __init attribute to mpic_msi_init()

Add the __init attribute to the mpic_msi_init() function. It is only
called from the device initializer, and so can be dropped after boot is
complete.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Drop msi_doorbell_end()
Marek Behún [Wed, 7 Aug 2024 16:40:54 +0000 (18:40 +0200)]
irqchip/armada-370-xp: Drop msi_doorbell_end()

Drop the msi_doorbell_end() function and related constants, it is not
used anymore.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/armada-370-xp: Drop IPI_DOORBELL_START and rename IPI_DOORBELL_END
Marek Behún [Wed, 7 Aug 2024 16:40:53 +0000 (18:40 +0200)]
irqchip/armada-370-xp: Drop IPI_DOORBELL_START and rename IPI_DOORBELL_END

Drop IPI_DOORBELL_START since it is not used and rename IPI_DOORBELL_END
to IPI_DOORBELL_NR.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
11 months agoirqchip/gic-v4.1: Replace bare number with ID_AA64PFR0_EL1_GIC_V4P1
Anshuman Khandual [Fri, 2 Aug 2024 08:56:01 +0000 (14:26 +0530)]
irqchip/gic-v4.1: Replace bare number with ID_AA64PFR0_EL1_GIC_V4P1

Use ID_AA64PFR0_EL1_GIC_V4P1 instead of '3' in gic_cpuif_has_vsgi() to
check for the GIC version.

Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/all/20240802085601.1824057-1-anshuman.khandual@arm.com
11 months agoirqchip: Remove asmlinkage for handlers registered with set_handle_irq()
Jinjie Ruan [Mon, 29 Jul 2024 11:26:06 +0000 (19:26 +0800)]
irqchip: Remove asmlinkage for handlers registered with set_handle_irq()

All architectures with use set_handle_irq() to set the root chip interrupt
handler call that handler from C code, so there's no need for these
handlers to be marked asmlinkage.

Remove asmlinkage for all handlers registered with set_handle_irq().

Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Link: https://lore.kernel.org/all/20240729112606.1581732-1-ruanjinjie@huawei.com
11 months agoirqchip/armada-370-xp: Print error and return error code on initialization failure
Marek Behún [Thu, 11 Jul 2024 16:09:07 +0000 (18:09 +0200)]
irqchip/armada-370-xp: Print error and return error code on initialization failure

Print error and return error code on main / IPI / MSI domain
initialization failure. Use WARN_ON() instead of BUG_ON().

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240711160907.31012-11-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Refactor initial memory regions mapping
Marek Behún [Thu, 11 Jul 2024 16:09:06 +0000 (18:09 +0200)]
irqchip/armada-370-xp: Refactor initial memory regions mapping

Refactor the initial memory regions mapping:
- put into its own function
- return error numbers on failure
- use WARN_ON() instead of BUG_ON()

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240711160907.31012-10-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Use u32 type instead of unsigned long where possieble
Marek Behún [Thu, 11 Jul 2024 16:09:05 +0000 (18:09 +0200)]
irqchip/armada-370-xp: Use u32 type instead of unsigned long where possieble

For consistency across the driver, use the u32 type instead of unsigned
long for holding register values and return value of cpu_logical_map().

One exception is when the variable is referenced for passing into
for_each_set_bit(), in which case it has to be unsigned long.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240711160907.31012-9-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Rename variable for consistency
Marek Behún [Thu, 11 Jul 2024 16:09:04 +0000 (18:09 +0200)]
irqchip/armada-370-xp: Rename variable for consistency

Rename the variable holding the cause register to "cause" in
mpic_handle_cascade_irq().

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240711160907.31012-8-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Drop redundant continue
Marek Behún [Thu, 11 Jul 2024 16:09:03 +0000 (18:09 +0200)]
irqchip/armada-370-xp: Drop redundant continue

Drop redundant continue from mpic_handle_irq().

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240711160907.31012-7-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Simplify mpic_reenable_percpu() and mpic_resume()
Marek Behún [Thu, 11 Jul 2024 16:09:01 +0000 (18:09 +0200)]
irqchip/armada-370-xp: Simplify mpic_reenable_percpu() and mpic_resume()

Refactor the mpic_reenable_percpu() and mpic_resume() functions to make
them a little bit simpler.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240711160907.31012-5-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Use consistent name for struct irq_data variables
Marek Behún [Thu, 11 Jul 2024 16:09:00 +0000 (18:09 +0200)]
irqchip/armada-370-xp: Use consistent name for struct irq_data variables

Always use variable name "d" for struct irq_data *, for consistency.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240711160907.31012-4-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Use consistent types when iterating interrupts
Marek Behún [Thu, 11 Jul 2024 16:08:59 +0000 (18:08 +0200)]
irqchip/armada-370-xp: Use consistent types when iterating interrupts

When iterating, use either the irq_hw_number_t type or the unsigned int
type for the iterator variable, depending on whether the variable
represents HW IRQ number or whether it is added to a IRQ number.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240711160907.31012-3-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Use consistent variable names for hwirqs
Marek Behún [Thu, 11 Jul 2024 16:08:58 +0000 (18:08 +0200)]
irqchip/armada-370-xp: Use consistent variable names for hwirqs

Use consistent variable names for hwirqs: when iterating, use "i",
otherwise use "hwirq".

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/20240711160907.31012-2-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Refactor handling IPI interrupts
Marek Behún [Thu, 11 Jul 2024 11:57:48 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Refactor handling IPI interrupts

Refactor the handling of IPI interrupts
- put into own function mpic_handle_ipi_irq(), similar to
  mpic_handle_msi_irq()
- rename the variable holding the doorbell cause register to "cause"
- retype and rename the variable holding the IPI HW IRQ number to
  "irq_hw_number_t i"

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-11-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Refactor mpic_handle_msi_irq() code
Marek Behún [Thu, 11 Jul 2024 11:57:47 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Refactor mpic_handle_msi_irq() code

Refactor the mpic_handle_msi_irq() function to make it simpler:
- drop the function arguments, they are not needed
- rename the variable holding the doorbell cause register to "cause"
- rename the iterating variable to "i"
- use for_each_set_bit() (requires retyping "cause" to unsigned long)

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-10-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Use FIELD_GET() and named register constant
Marek Behún [Thu, 11 Jul 2024 11:57:46 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Use FIELD_GET() and named register constant

Use FIELD_GET() and named register mask constant when reading the number
of supported interrupts / current interrupt.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-9-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Don't read number of supported interrupts multiple times
Marek Behún [Thu, 11 Jul 2024 11:57:45 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Don't read number of supported interrupts multiple times

Use mpic_domain::hwirq_max at runtime instead of reading the same value
over and over from the MPIC_INT_CONTROL register.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-8-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Change symbol prefixes to mpic
Marek Behún [Thu, 11 Jul 2024 11:57:44 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Change symbol prefixes to mpic

Change symbol prefixes from armada_370_xp_ or others to mpic_.

The rationale is that it is shorter and more generic (this controller
is called MPIC and is also used on Armada 38x and 39x).

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-7-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Improve indentation
Marek Behún [Thu, 11 Jul 2024 11:57:43 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Improve indentation

Add some blank lines and other indentation improvements.

Checkpatch now stops complaining.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-6-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Simplify ipi_resume() code
Marek Behún [Thu, 11 Jul 2024 11:57:42 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Simplify ipi_resume() code

Refactor the ipi_resume() function to drop one indentation level.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-5-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Use !virq instead of virq == 0 in condition
Marek Behún [Thu, 11 Jul 2024 11:57:41 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Use !virq instead of virq == 0 in condition

Use !virq instead of virq == 0 when checking for availability of the
virq.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-4-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Use unsigned int type for virqs
Marek Behún [Thu, 11 Jul 2024 11:57:40 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Use unsigned int type for virqs

The return type of irq_find_mapping() and irq_linear_revmap() is
unsigned int. Use the unsigned int type for the variables storing the
return value.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-3-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Rename variable for consistency
Marek Behún [Thu, 11 Jul 2024 11:57:39 +0000 (13:57 +0200)]
irqchip/armada-370-xp: Rename variable for consistency

Rename the irq variable to virq in the ipi_resume() function for
consistency with the rest of the code.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240711115748.30268-2-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Declare iterators in for loop
Marek Behún [Mon, 8 Jul 2024 15:18:01 +0000 (17:18 +0200)]
irqchip/armada-370-xp: Declare iterators in for loop

Where possible, declare iterators in for cycle. This is possible since
kernel uses -std=gnu11.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/all/20240708151801.11592-11-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Change to SPDX license identifier
Marek Behún [Mon, 8 Jul 2024 15:18:00 +0000 (17:18 +0200)]
irqchip/armada-370-xp: Change to SPDX license identifier

Change the license identifier to SPDX style.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/all/20240708151801.11592-10-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Simplify is_percpu_irq() code
Marek Behún [Mon, 8 Jul 2024 15:17:59 +0000 (17:17 +0200)]
irqchip/armada-370-xp: Simplify is_percpu_irq() code

Simplify the code in the is_percpu_irq() function. Instead of
  if (condition)
    return true;
  return false;
simply return condition.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240708151801.11592-9-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Use correct type for cpu variable
Marek Behún [Mon, 8 Jul 2024 15:17:58 +0000 (17:17 +0200)]
irqchip/armada-370-xp: Use correct type for cpu variable

Use unsigned int instead of int for variable storing the cpu number.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/all/20240708151801.11592-8-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Change register constants prefix to MPIC_
Marek Behún [Mon, 8 Jul 2024 15:17:57 +0000 (17:17 +0200)]
irqchip/armada-370-xp: Change register constants prefix to MPIC_

Change the long ARMADA_370_XP_ prefix in register constants (ARMADA_375_
in one case) to MPIC_. The rationale is that it is shorter and more
generic (this controller is called MPIC and is also used on Armada 38x
and 39x).

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/all/20240708151801.11592-7-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Cosmetic fix parentheses in register constant definitions
Marek Behún [Mon, 8 Jul 2024 15:17:56 +0000 (17:17 +0200)]
irqchip/armada-370-xp: Cosmetic fix parentheses in register constant definitions

Drop parentheses where not needed and add them where it makes sense in
register constant definitions.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240708151801.11592-6-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Use BIT() and GENMASK() macros
Marek Behún [Mon, 8 Jul 2024 15:17:55 +0000 (17:17 +0200)]
irqchip/armada-370-xp: Use BIT() and GENMASK() macros

Use the BIT() and GENMASK() macros where appropriate.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/all/20240708151801.11592-5-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Change spaces to tabs
Marek Behún [Mon, 8 Jul 2024 15:17:54 +0000 (17:17 +0200)]
irqchip/armada-370-xp: Change spaces to tabs

Change spaces to tabs in register constants definitions.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240708151801.11592-4-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Change register constant suffix from _MSK to _MASK
Marek Behún [Mon, 8 Jul 2024 15:17:53 +0000 (17:17 +0200)]
irqchip/armada-370-xp: Change register constant suffix from _MSK to _MASK

There is one occurrence of suffix _MSK in register constants, others
have _MASK instead. Change the one to _MASK for consistency.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240708151801.11592-3-kabel@kernel.org
11 months agoirqchip/armada-370-xp: Drop _OFFS suffix from some register constants
Marek Behún [Mon, 8 Jul 2024 15:17:52 +0000 (17:17 +0200)]
irqchip/armada-370-xp: Drop _OFFS suffix from some register constants

Some register constants have the _OFFS suffix and some do not. Drop it
to be more consistent.

Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/all/20240708151801.11592-2-kabel@kernel.org
11 months agoLinux 6.11-rc1 v6.11-rc1
Linus Torvalds [Sun, 28 Jul 2024 21:19:55 +0000 (14:19 -0700)]
Linux 6.11-rc1

11 months agoMerge tag 'kbuild-fixes-v6.11' of git://git.kernel.org/pub/scm/linux/kernel/git/masah...
Linus Torvalds [Sun, 28 Jul 2024 21:02:48 +0000 (14:02 -0700)]
Merge tag 'kbuild-fixes-v6.11' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild fixes from Masahiro Yamada:

 - Fix RPM package build error caused by an incorrect locale setup

 - Mark modules.weakdep as ghost in RPM package

 - Fix the odd combination of -S and -c in stack protector scripts,
   which is an error with the latest Clang

* tag 'kbuild-fixes-v6.11' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  kbuild: Fix '-S -c' in x86 stack protector scripts
  kbuild: rpm-pkg: ghost modules.weakdep file
  kbuild: rpm-pkg: Fix C locale setup

11 months agominmax: simplify and clarify min_t()/max_t() implementation
Linus Torvalds [Sun, 28 Jul 2024 20:50:01 +0000 (13:50 -0700)]
minmax: simplify and clarify min_t()/max_t() implementation

This simplifies the min_t() and max_t() macros by no longer making them
work in the context of a C constant expression.

That means that you can no longer use them for static initializers or
for array sizes in type definitions, but there were only a couple of
such uses, and all of them were converted (famous last words) to use
MIN_T/MAX_T instead.

Cc: David Laight <David.Laight@aculab.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 months agominmax: add a few more MIN_T/MAX_T users
Linus Torvalds [Sun, 28 Jul 2024 20:03:48 +0000 (13:03 -0700)]
minmax: add a few more MIN_T/MAX_T users

Commit 3a7e02c040b1 ("minmax: avoid overly complicated constant
expressions in VM code") added the simpler MIN_T/MAX_T macros in order
to avoid some excessive expansion from the rather complicated regular
min/max macros.

The complexity of those macros stems from two issues:

 (a) trying to use them in situations that require a C constant
     expression (in static initializers and for array sizes)

 (b) the type sanity checking

and MIN_T/MAX_T avoids both of these issues.

Now, in the whole (long) discussion about all this, it was pointed out
that the whole type sanity checking is entirely unnecessary for
min_t/max_t which get a fixed type that the comparison is done in.

But that still leaves min_t/max_t unnecessarily complicated due to
worries about the C constant expression case.

However, it turns out that there really aren't very many cases that use
min_t/max_t for this, and we can just force-convert those.

This does exactly that.

Which in turn will then allow for much simpler implementations of
min_t()/max_t().  All the usual "macros in all upper case will evaluate
the arguments multiple times" rules apply.

We should do all the same things for the regular min/max() vs MIN/MAX()
cases, but that has the added complexity of various drivers defining
their own local versions of MIN/MAX, so that needs another level of
fixes first.

Link: https://lore.kernel.org/all/b47fad1d0cf8449886ad148f8c013dae@AcuMS.aculab.com/
Cc: David Laight <David.Laight@aculab.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 months agoMerge tag 'ubifs-for-linus-6.11-rc1-take2' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Sun, 28 Jul 2024 18:51:51 +0000 (11:51 -0700)]
Merge tag 'ubifs-for-linus-6.11-rc1-take2' of git://git./linux/kernel/git/rw/ubifs

Pull UBI and UBIFS updates from Richard Weinberger:

 - Many fixes for power-cut issues by Zhihao Cheng

 - Another ubiblock error path fix

 - ubiblock section mismatch fix

 - Misc fixes all over the place

* tag 'ubifs-for-linus-6.11-rc1-take2' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs:
  ubi: Fix ubi_init() ubiblock_exit() section mismatch
  ubifs: add check for crypto_shash_tfm_digest
  ubifs: Fix inconsistent inode size when powercut happens during appendant writing
  ubi: block: fix null-pointer-dereference in ubiblock_create()
  ubifs: fix kernel-doc warnings
  ubifs: correct UBIFS_DFS_DIR_LEN macro definition and improve code clarity
  mtd: ubi: Restore missing cleanup on ubi_init() failure path
  ubifs: dbg_orphan_check: Fix missed key type checking
  ubifs: Fix unattached inode when powercut happens in creating
  ubifs: Fix space leak when powercut happens in linking tmpfile
  ubifs: Move ui->data initialization after initializing security
  ubifs: Fix adding orphan entry twice for the same inode
  ubifs: Remove insert_dead_orphan from replaying orphan process
  Revert "ubifs: ubifs_symlink: Fix memleak of inode->i_link in error path"
  ubifs: Don't add xattr inode into orphan area
  ubifs: Fix unattached xattr inode if powercut happens after deleting
  mtd: ubi: avoid expensive do_div() on 32-bit machines
  mtd: ubi: make ubi_class constant
  ubi: eba: properly rollback inside self_check_eba

11 months agokbuild: Fix '-S -c' in x86 stack protector scripts
Nathan Chancellor [Fri, 26 Jul 2024 18:05:00 +0000 (11:05 -0700)]
kbuild: Fix '-S -c' in x86 stack protector scripts

After a recent change in clang to stop consuming all instances of '-S'
and '-c' [1], the stack protector scripts break due to the kernel's use
of -Werror=unused-command-line-argument to catch cases where flags are
not being properly consumed by the compiler driver:

  $ echo | clang -o - -x c - -S -c -Werror=unused-command-line-argument
  clang: error: argument unused during compilation: '-c' [-Werror,-Wunused-command-line-argument]

This results in CONFIG_STACKPROTECTOR getting disabled because
CONFIG_CC_HAS_SANE_STACKPROTECTOR is no longer set.

'-c' and '-S' both instruct the compiler to stop at different stages of
the pipeline ('-S' after compiling, '-c' after assembling), so having
them present together in the same command makes little sense. In this
case, the test wants to stop before assembling because it is looking at
the textual assembly output of the compiler for either '%fs' or '%gs',
so remove '-c' from the list of arguments to resolve the error.

All versions of GCC continue to work after this change, along with
versions of clang that do or do not contain the change mentioned above.

Cc: stable@vger.kernel.org
Fixes: 4f7fd4d7a791 ("[PATCH] Add the -fstack-protector option to the CFLAGS")
Fixes: 60a5317ff0f4 ("x86: implement x86_32 stack protector")
Link: https://github.com/llvm/llvm-project/commit/6461e537815f7fa68cef06842505353cf5600e9c
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
11 months agoubi: Fix ubi_init() ubiblock_exit() section mismatch
Richard Weinberger [Sat, 13 Jul 2024 07:35:19 +0000 (09:35 +0200)]
ubi: Fix ubi_init() ubiblock_exit() section mismatch

Since ubiblock_exit() is now called from an init function,
the __exit section no longer makes sense.

Cc: Ben Hutchings <bwh@kernel.org>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202407131403.wZJpd8n2-lkp@intel.com/
Signed-off-by: Richard Weinberger <richard@nod.at>
Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
11 months agoMerge tag 'v6.11-merge' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux
Linus Torvalds [Sun, 28 Jul 2024 17:52:15 +0000 (10:52 -0700)]
Merge tag 'v6.11-merge' of git://git./linux/kernel/git/lenb/linux

Pull turbostat updates from Len Brown:

 - Enable turbostat extensions to add both perf and PMT (Intel
   Platform Monitoring Technology) counters via the cmdline

 - Demonstrate PMT access with built-in support for Meteor Lake's
   Die C6 counter

* tag 'v6.11-merge' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux:
  tools/power turbostat: version 2024.07.26
  tools/power turbostat: Include umask=%x in perf counter's config
  tools/power turbostat: Document PMT in turbostat.8
  tools/power turbostat: Add MTL's PMT DC6 builtin counter
  tools/power turbostat: Add early support for PMT counters
  tools/power turbostat: Add selftests for added perf counters
  tools/power turbostat: Add selftests for SMI, APERF and MPERF counters
  tools/power turbostat: Move verbose counter messages to level 2
  tools/power turbostat: Move debug prints from stdout to stderr
  tools/power turbostat: Fix typo in turbostat.8
  tools/power turbostat: Add perf added counter example to turbostat.8
  tools/power turbostat: Fix formatting in turbostat.8
  tools/power turbostat: Extend --add option with perf counters
  tools/power turbostat: Group SMI counter with APERF and MPERF
  tools/power turbostat: Add ZERO_ARRAY for zero initializing builtin array
  tools/power turbostat: Replace enum rapl_source and cstate_source with counter_source
  tools/power turbostat: Remove anonymous union from rapl_counter_info_t
  tools/power/turbostat: Switch to new Intel CPU model defines

11 months agoMerge tag 'cxl-for-6.11' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl
Linus Torvalds [Sun, 28 Jul 2024 16:33:28 +0000 (09:33 -0700)]
Merge tag 'cxl-for-6.11' of git://git./linux/kernel/git/cxl/cxl

Pull CXL updates from Dave Jiang:
 "Core:

   - A CXL maturity map has been added to the documentation to detail
     the current state of CXL enabling.

     It provides the status of the current state of various CXL features
     to inform current and future contributors of where things are and
     which areas need contribution.

   - A notifier handler has been added in order for a newly created CXL
     memory region to trigger the abstract distance metrics calculation.

     This should bring parity for CXL memory to the same level vs
     hotplugged DRAM for NUMA abstract distance calculation. The
     abstract distance reflects relative performance used for memory
     tiering handling.

   - An addition for XOR math has been added to address the CXL DPA to
     SPA translation.

     CXL address translation did not support address interleave math
     with XOR prior to this change.

  Fixes:

   - Fix to address race condition in the CXL memory hotplug notifier

   - Add missing MODULE_DESCRIPTION() for CXL modules

   - Fix incorrect vendor debug UUID define

  Misc:

   - A warning has been added to inform users of an unsupported
     configuration when mixing CXL VH and RCH/RCD hierarchies

   - The ENXIO error code has been replaced with EBUSY for inject poison
     limit reached via debugfs and cxl-test support

   - Moving the PCI config read in cxl_dvsec_rr_decode() to avoid
     unnecessary PCI config reads

   - A refactor to a common struct for DRAM and general media CXL
     events"

* tag 'cxl-for-6.11' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl:
  cxl/core/pci: Move reading of control register to immediately before usage
  cxl: Remove defunct code calculating host bridge target positions
  cxl/region: Verify target positions using the ordered target list
  cxl: Restore XOR'd position bits during address translation
  cxl/core: Fold cxl_trace_hpa() into cxl_dpa_to_hpa()
  cxl/test: Replace ENXIO with EBUSY for inject poison limit reached
  cxl/memdev: Replace ENXIO with EBUSY for inject poison limit reached
  cxl/acpi: Warn on mixed CXL VH and RCH/RCD Hierarchy
  cxl/core: Fix incorrect vendor debug UUID define
  Documentation: CXL Maturity Map
  cxl/region: Simplify cxl_region_nid()
  cxl/region: Support to calculate memory tier abstract distance
  cxl/region: Fix a race condition in memory hotplug notifier
  cxl: add missing MODULE_DESCRIPTION() macros
  cxl/events: Use a common struct for DRAM and General Media events

11 months agoMerge tag 'unicode-next-6.11' of git://git.kernel.org/pub/scm/linux/kernel/git/krisma...
Linus Torvalds [Sun, 28 Jul 2024 16:14:11 +0000 (09:14 -0700)]
Merge tag 'unicode-next-6.11' of git://git./linux/kernel/git/krisman/unicode

Pull unicode update from Gabriel Krisman Bertazi:
 "Two small fixes to silence the compiler and static analyzers tools
  from Ben Dooks and Jeff Johnson"

* tag 'unicode-next-6.11' of git://git.kernel.org/pub/scm/linux/kernel/git/krisman/unicode:
  unicode: add MODULE_DESCRIPTION() macros
  unicode: make utf8 test count static

11 months agokbuild: rpm-pkg: ghost modules.weakdep file
Jose Ignacio Tornos Martinez [Fri, 26 Jul 2024 09:00:26 +0000 (11:00 +0200)]
kbuild: rpm-pkg: ghost modules.weakdep file

In the same way as for other similar files, mark as ghost the new file
generated by depmod for configured weak dependencies for modules,
modules.weakdep, so that although it is not included in the package,
claim the ownership on it.

Signed-off-by: Jose Ignacio Tornos Martinez <jtornosm@redhat.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
11 months agoMerge tag '6.11-rc-smb-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 28 Jul 2024 03:08:07 +0000 (20:08 -0700)]
Merge tag '6.11-rc-smb-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6

Pull more smb client updates from Steve French:

 - fix for potential null pointer use in init cifs

 - additional dynamic trace points to improve debugging of some common
   scenarios

 - two SMB1 fixes (one addressing reconnect with POSIX extensions, one a
   mount parsing error)

* tag '6.11-rc-smb-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6:
  smb3: add dynamic trace point for session setup key expired failures
  smb3: add four dynamic tracepoints for copy_file_range and reflink
  smb3: add dynamic tracepoint for reflink errors
  cifs: mount with "unix" mount option for SMB1 incorrectly handled
  cifs: fix reconnect with SMB1 UNIX Extensions
  cifs: fix potential null pointer use in destroy_workqueue in init_cifs error path

11 months agoMerge tag 'block-6.11-20240726' of git://git.kernel.dk/linux
Linus Torvalds [Sat, 27 Jul 2024 22:28:53 +0000 (15:28 -0700)]
Merge tag 'block-6.11-20240726' of git://git.kernel.dk/linux

Pull block fixes from Jens Axboe:

 - NVMe pull request via Keith:
     - Fix request without payloads cleanup  (Leon)
     - Use new protection information format (Francis)
     - Improved debug message for lost pci link (Bart)
     - Another apst quirk (Wang)
     - Use appropriate sysfs api for printing chars (Markus)

 - ublk async device deletion fix (Ming)

 - drbd kerneldoc fixups (Simon)

 - Fix deadlock between sd removal and release (Yang)

* tag 'block-6.11-20240726' of git://git.kernel.dk/linux:
  nvme-pci: add missing condition check for existence of mapped data
  ublk: fix UBLK_CMD_DEL_DEV_ASYNC handling
  block: fix deadlock between sd_remove & sd_release
  drbd: Add peer_device to Kernel doc
  nvme-core: choose PIF from QPIF if QPIFS supports and PIF is QTYPE
  nvme-pci: Fix the instructions for disabling power management
  nvme: remove redundant bdev local variable
  nvme-fabrics: Use seq_putc() in __nvmf_concat_opt_tokens()
  nvme/pci: Add APST quirk for Lenovo N60z laptop

11 months agoMerge tag 'io_uring-6.11-20240726' of git://git.kernel.dk/linux
Linus Torvalds [Sat, 27 Jul 2024 22:22:33 +0000 (15:22 -0700)]
Merge tag 'io_uring-6.11-20240726' of git://git.kernel.dk/linux

Pull io_uring fixes from Jens Axboe:

 - Fix a syzbot issue for the msg ring cache added in this release. No
   ill effects from this one, but it did make KMSAN unhappy (me)

 - Sanitize the NAPI timeout handling, by unifying the value handling
   into all ktime_t rather than converting back and forth (Pavel)

 - Fail NAPI registration for IOPOLL rings, it's not supported (Pavel)

 - Fix a theoretical issue with ring polling and cancelations (Pavel)

 - Various little cleanups and fixes (Pavel)

* tag 'io_uring-6.11-20240726' of git://git.kernel.dk/linux:
  io_uring/napi: pass ktime to io_napi_adjust_timeout
  io_uring/napi: use ktime in busy polling
  io_uring/msg_ring: fix uninitialized use of target_req->flags
  io_uring: align iowq and task request error handling
  io_uring: kill REQ_F_CANCEL_SEQ
  io_uring: simplify io_uring_cmd return
  io_uring: fix io_match_task must_hold
  io_uring: don't allow netpolling with SETUP_IOPOLL
  io_uring: tighten task exit cancellations

11 months agoMerge tag 'vfs-6.11-rc1.fixes.3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 27 Jul 2024 22:11:59 +0000 (15:11 -0700)]
Merge tag 'vfs-6.11-rc1.fixes.3' of git://git./linux/kernel/git/vfs/vfs

Pull vfs fixes from Christian Brauner:
 "This contains two fixes for this merge window:

  VFS:

   - I noticed that it is possible for a privileged user to mount most
     filesystems with a non-initial user namespace in sb->s_user_ns.

     When fsopen() is called in a non-init namespace the caller's
     namespace is recorded in fs_context->user_ns. If the returned file
     descriptor is then passed to a process privileged in init_user_ns,
     that process can call fsconfig(fd_fs, FSCONFIG_CMD_CREATE*),
     creating a new superblock with sb->s_user_ns set to the namespace
     of the process which called fsopen().

     This is problematic as only filesystems that raise FS_USERNS_MOUNT
     are known to be able to support a non-initial s_user_ns. Others may
     suffer security issues, on-disk corruption or outright crash the
     kernel. Prevent that by restricting such delegation to filesystems
     that allow FS_USERNS_MOUNT.

     Note, that this delegation requires a privileged process to
     actually create the superblock so either the privileged process is
     cooperaing or someone must have tricked a privileged process into
     operating on a fscontext file descriptor whose origin it doesn't
     know (a stupid idea).

     The bug dates back to about 5 years afaict.

  Misc:

   - Fix hostfs parsing when the mount request comes in via the legacy
     mount api.

     In the legacy mount api hostfs allows to specify the host directory
     mount without any key.

     Restore that behavior"

* tag 'vfs-6.11-rc1.fixes.3' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
  hostfs: fix the host directory parse when mounting.
  fs: don't allow non-init s_user_ns for filesystems without FS_USERNS_MOUNT

11 months agoMerge tag 'rust-6.11' of https://github.com/Rust-for-Linux/linux
Linus Torvalds [Sat, 27 Jul 2024 20:44:54 +0000 (13:44 -0700)]
Merge tag 'rust-6.11' of https://github.com/Rust-for-Linux/linux

Pull Rust updates from Miguel Ojeda:
 "The highlight is the establishment of a minimum version for the Rust
  toolchain, including 'rustc' (and bundled tools) and 'bindgen'.

  The initial minimum will be the pinned version we currently have, i.e.
  we are just widening the allowed versions. That covers three stable
  Rust releases: 1.78.0, 1.79.0, 1.80.0 (getting released tomorrow),
  plus beta, plus nightly.

  This should already be enough for kernel developers in distributions
  that provide recent Rust compiler versions routinely, such as Arch
  Linux, Debian Unstable (outside the freeze period), Fedora Linux,
  Gentoo Linux (especially the testing channel), Nix (unstable) and
  openSUSE Slowroll and Tumbleweed.

  In addition, the kernel is now being built-tested by Rust's pre-merge
  CI. That is, every change that is attempting to land into the Rust
  compiler is tested against the kernel, and it is merged only if it
  passes. Similarly, the bindgen tool has agreed to build the kernel in
  their CI too.

  Thus, with the pre-merge CI in place, both projects hope to avoid
  unintentional changes to Rust that break the kernel. This means that,
  in general, apart from intentional changes on their side (that we will
  need to workaround conditionally on our side), the upcoming Rust
  compiler versions should generally work.

  In addition, the Rust project has proposed getting the kernel into
  stable Rust (at least solving the main blockers) as one of its three
  flagship goals for 2024H2 [1].

  I would like to thank Niko, Sid, Emilio et al. for their help
  promoting the collaboration between Rust and the kernel.

  Toolchain and infrastructure:

   - Support several Rust toolchain versions.

   - Support several bindgen versions.

   - Remove 'cargo' requirement and simplify 'rusttest', thanks to
     'alloc' having been dropped last cycle.

   - Provide proper error reporting for the 'rust-analyzer' target.

  'kernel' crate:

   - Add 'uaccess' module with a safe userspace pointers abstraction.

   - Add 'page' module with a 'struct page' abstraction.

   - Support more complex generics in workqueue's 'impl_has_work!'
     macro.

  'macros' crate:

   - Add 'firmware' field support to the 'module!' macro.

   - Improve 'module!' macro documentation.

  Documentation:

   - Provide instructions on what packages should be installed to build
     the kernel in some popular Linux distributions.

   - Introduce the new kernel.org LLVM+Rust toolchains.

   - Explain '#[no_std]'.

  And a few other small bits"

Link: https://rust-lang.github.io/rust-project-goals/2024h2/index.html#flagship-goals
* tag 'rust-6.11' of https://github.com/Rust-for-Linux/linux: (26 commits)
  docs: rust: quick-start: add section on Linux distributions
  rust: warn about `bindgen` versions 0.66.0 and 0.66.1
  rust: start supporting several `bindgen` versions
  rust: work around `bindgen` 0.69.0 issue
  rust: avoid assuming a particular `bindgen` build
  rust: start supporting several compiler versions
  rust: simplify Clippy warning flags set
  rust: relax most deny-level lints to warnings
  rust: allow `dead_code` for never constructed bindings
  rust: init: simplify from `map_err` to `inspect_err`
  rust: macros: indent list item in `paste!`'s docs
  rust: add abstraction for `struct page`
  rust: uaccess: add typed accessors for userspace pointers
  uaccess: always export _copy_[from|to]_user with CONFIG_RUST
  rust: uaccess: add userspace pointers
  kbuild: rust-analyzer: improve comment documentation
  kbuild: rust-analyzer: better error handling
  docs: rust: no_std is used
  rust: alloc: add __GFP_HIGHMEM flag
  rust: alloc: fix typo in docs for GFP_NOWAIT
  ...

11 months agoMerge tag 'apparmor-pr-2024-07-25' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 27 Jul 2024 20:28:39 +0000 (13:28 -0700)]
Merge tag 'apparmor-pr-2024-07-25' of git://git./linux/kernel/git/jj/linux-apparmor

Pull apparmor updates from John Johansen:
 "Cleanups
   - optimization: try to avoid refing the label in apparmor_file_open
   - remove useless static inline function is_deleted
   - use kvfree_sensitive to free data->data
   - fix typo in kernel doc

  Bug fixes:
   - unpack transition table if dfa is not present
   - test: add MODULE_DESCRIPTION()
   - take nosymfollow flag into account
   - fix possible NULL pointer dereference
   - fix null pointer deref when receiving skb during sock creation"

* tag 'apparmor-pr-2024-07-25' of git://git.kernel.org/pub/scm/linux/kernel/git/jj/linux-apparmor:
  apparmor: unpack transition table if dfa is not present
  apparmor: try to avoid refing the label in apparmor_file_open
  apparmor: test: add MODULE_DESCRIPTION()
  apparmor: take nosymfollow flag into account
  apparmor: fix possible NULL pointer dereference
  apparmor: fix typo in kernel doc
  apparmor: remove useless static inline function is_deleted
  apparmor: use kvfree_sensitive to free data->data
  apparmor: Fix null pointer deref when receiving skb during sock creation

11 months agoMerge tag 'landlock-6.11-rc1-houdini-fix' of git://git.kernel.org/pub/scm/linux/kerne...
Linus Torvalds [Sat, 27 Jul 2024 20:16:53 +0000 (13:16 -0700)]
Merge tag 'landlock-6.11-rc1-houdini-fix' of git://git./linux/kernel/git/mic/linux

Pull landlock fix from Mickaël Salaün:
 "Jann Horn reported a sandbox bypass for Landlock. This includes the
  fix and new tests. This should be backported"

* tag 'landlock-6.11-rc1-houdini-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/mic/linux:
  selftests/landlock: Add cred_transfer test
  landlock: Don't lose track of restrictions on cred_transfer

11 months agoMerge tag 'gpio-fixes-for-v6.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 27 Jul 2024 19:54:06 +0000 (12:54 -0700)]
Merge tag 'gpio-fixes-for-v6.11-rc1' of git://git./linux/kernel/git/brgl/linux

Pull gpio fix from Bartosz Golaszewski:

 - don't use sprintf() with non-constant format string

* tag 'gpio-fixes-for-v6.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
  gpio: virtuser: avoid non-constant format string

11 months agoMerge tag 'devicetree-fixes-for-6.11-1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 27 Jul 2024 19:46:16 +0000 (12:46 -0700)]
Merge tag 'devicetree-fixes-for-6.11-1' of git://git./linux/kernel/git/robh/linux

Pull more devicetree updates from Rob Herring:
 "Most of this is a treewide change to of_property_for_each_u32() which
  was small enough to do in one go before rc1 and avoids the need to
  create of_property_for_each_u32_some_new_name().

   - Treewide conversion of of_property_for_each_u32() to drop internal
     arguments making struct property opaque

   - Add binding for Amlogic A4 SoC watchdog

   - Fix constraints for AD7192 'single-channel' property"

* tag 'devicetree-fixes-for-6.11-1' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
  dt-bindings: iio: adc: ad7192: Fix 'single-channel' constraints
  of: remove internal arguments from of_property_for_each_u32()
  dt-bindings: watchdog: add support for Amlogic A4 SoCs

11 months agoMerge tag 'iommu-fixes-v6.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 27 Jul 2024 19:39:55 +0000 (12:39 -0700)]
Merge tag 'iommu-fixes-v6.11-rc1' of git://git./linux/kernel/git/iommu/linux

Pull iommu fixes from Will Deacon:
 "We're still resolving a regression with the handling of unexpected
  page faults on SMMUv3, but we're not quite there with a fix yet.

   - Fix NULL dereference when freeing domain in Unisoc SPRD driver

   - Separate assignment statements with semicolons in AMD page-table
     code

   - Fix Tegra erratum workaround when the CPU is using 16KiB pages"

* tag 'iommu-fixes-v6.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/iommu/linux:
  iommu: arm-smmu: Fix Tegra workaround for PAGE_SIZE mappings
  iommu/amd: Convert comma to semicolon
  iommu: sprd: Avoid NULL deref in sprd_iommu_hw_en

11 months agoMerge tag 'firewire-fixes-6.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 27 Jul 2024 19:35:12 +0000 (12:35 -0700)]
Merge tag 'firewire-fixes-6.11-rc1' of git://git./linux/kernel/git/ieee1394/linux1394

Pull firewire fixes from Takashi Sakamoto:
 "The recent integration of compiler collections introduced the
  technology to check flexible array length at runtime by providing
  proper annotations. In v6.10 kernel, a patch was merged into firewire
  subsystem to utilize it, however the annotation was inadequate.

  There is also the related change for the flexible array in sound
  subsystem, but it causes a regression where the data in the payload of
  isochronous packet is incorrect for some devices. These bugs are now
  fixed"

* tag 'firewire-fixes-6.11-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394:
  ALSA: firewire-lib: fix wrong value as length of header for CIP_NO_HEADER case
  Revert "firewire: Annotate struct fw_iso_packet with __counted_by()"

11 months agoMerge tag 'spi-fix-v6.11-merge-window' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 27 Jul 2024 19:29:10 +0000 (12:29 -0700)]
Merge tag 'spi-fix-v6.11-merge-window' of git://git./linux/kernel/git/broonie/spi

Pull spi fixes from Mark Brown:
 "The bulk of this is a series of fixes for the microchip-core driver
  mostly originating from one of their customers, I also applied an
  additional patch adding support for controlling the word size which
  came along with it since it's still the merge window and clearly had a
  bunch of fairly thorough testing.

  We also have a fix for the compatible used to bind spidev to the
  BH2228FV"

* tag 'spi-fix-v6.11-merge-window' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi: spidev: add correct compatible for Rohm BH2228FV
  dt-bindings: trivial-devices: fix Rohm BH2228FV compatible string
  spi: microchip-core: add support for word sizes of 1 to 32 bits
  spi: microchip-core: ensure TX and RX FIFOs are empty at start of a transfer
  spi: microchip-core: fix init function not setting the master and motorola modes
  spi: microchip-core: only disable SPI controller when register value change requires it
  spi: microchip-core: defer asserting chip select until just before write to TX FIFO
  spi: microchip-core: fix the issues in the isr

11 months agoMerge tag 'regulator-fix-v6.11-merge-window' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Sat, 27 Jul 2024 19:27:52 +0000 (12:27 -0700)]
Merge tag 'regulator-fix-v6.11-merge-window' of git://git./linux/kernel/git/broonie/regulator

Pull regulator fixes from Mark Brown:
 "These two commits clean up the excessively loose dependencies for the
  RZG2L USB VBCTRL regulator driver, ensuring it shouldn't prompt for
  people who can't use it"

* tag 'regulator-fix-v6.11-merge-window' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: Further restrict RZG2L USB VBCTRL regulator dependencies
  regulator: renesas-usb-vbus-regulator: Update the default

11 months agoMerge tag 'regmap-fix-v6.11-merge-window' of git://git.kernel.org/pub/scm/linux/kerne...
Linus Torvalds [Sat, 27 Jul 2024 19:26:09 +0000 (12:26 -0700)]
Merge tag 'regmap-fix-v6.11-merge-window' of git://git./linux/kernel/git/broonie/regmap

Pull regmap fix from Mark Brown:
 "Arnd sent a workaround for a false positive warning which was showing
  up with GCC 14.1"

* tag 'regmap-fix-v6.11-merge-window' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:
  regmap: maple: work around gcc-14.1 false-positive warning

11 months agoMerge tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux
Linus Torvalds [Sat, 27 Jul 2024 19:07:18 +0000 (12:07 -0700)]
Merge tag 'clk-for-linus' of git://git./linux/kernel/git/clk/linux

Pull clk fixes from Stephen Boyd:
 "A few clk driver fixes for the merge window to fix the build and boot
  on some SoCs.

   - Initialize struct clk_init_data in the TI da8xx-cfgchip driver so
     that stack contents aren't used for things like clk flags leading
     to unexpected behavior

   - Don't leak stack contents in a debug print in the new Sophgo clk
     driver

   - Disable the new T-Head clk driver on 32-bit targets to fix the
     build due to a division

   - Fix Samsung Exynos4 fin_pll wreckage from the clkdev rework done
     last cycle by using a struct clk_hw directly instead of a struct
     clk consumer"

* tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:
  clk: samsung: fix getting Exynos4 fin_pll rate from external clocks
  clk: T-Head: Disable on 32-bit Targets
  clk: sophgo: clk-sg2042-pll: Fix uninitialized variable in debug output
  clk: davinci: da8xx-cfgchip: Initialize clk_init_data before use