linux-stable/kernel/irq/irqdesc.c

1053 lines
25 KiB
C
Raw Permalink Normal View History

// SPDX-License-Identifier: GPL-2.0
/*
* Copyright (C) 1992, 1998-2006 Linus Torvalds, Ingo Molnar
* Copyright (C) 2005-2006, Thomas Gleixner, Russell King
*
* This file contains the interrupt descriptor management code. Detailed
* information is available in Documentation/core-api/genericirq.rst
*
*/
#include <linux/irq.h>
#include <linux/slab.h>
#include <linux/export.h>
#include <linux/interrupt.h>
#include <linux/kernel_stat.h>
#include <linux/maple_tree.h>
genirq: Add irq_domain-aware core IRQ handler Calling irq_find_mapping from outside a irq_{enter,exit} section is unsafe and produces ugly messages if CONFIG_PROVE_RCU is enabled: If coming from the idle state, the rcu_read_lock call in irq_find_mapping will generate an unpleasant warning: <quote> =============================== [ INFO: suspicious RCU usage. ] 3.16.0-rc1+ #135 Not tainted ------------------------------- include/linux/rcupdate.h:871 rcu_read_lock() used illegally while idle! other info that might help us debug this: RCU used illegally from idle CPU! rcu_scheduler_active = 1, debug_locks = 0 RCU used illegally from extended quiescent state! 1 lock held by swapper/0/0: #0: (rcu_read_lock){......}, at: [<ffffffc00010206c>] irq_find_mapping+0x4c/0x198 </quote> As this issue is fairly widespread and involves at least three different architectures, a possible solution is to add a new handle_domain_irq entry point into the generic IRQ code that the interrupt controller code can call. This new function takes an irq_domain, and calls into irq_find_domain inside the irq_{enter,exit} block. An additional "lookup" parameter is used to allow non-domain architecture code to be replaced by this as well. Interrupt controllers can then be updated to use the new mechanism. This code is sitting behind a new CONFIG_HANDLE_DOMAIN_IRQ, as not all architectures implement set_irq_regs (yes, mn10300, I'm looking at you...). Reported-by: Vladimir Murzin <vladimir.murzin@arm.com> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Link: https://lkml.kernel.org/r/1409047421-27649-2-git-send-email-marc.zyngier@arm.com Signed-off-by: Jason Cooper <jason@lakedaemon.net>
2014-08-26 10:03:16 +00:00
#include <linux/irqdomain.h>
#include <linux/sysfs.h>
#include "internals.h"
/*
* lockdep: we want to handle all irq_desc locks as a single lock-class:
*/
static struct lock_class_key irq_desc_lock_class;
#if defined(CONFIG_SMP)
static int __init irq_affinity_setup(char *str)
{
alloc_bootmem_cpumask_var(&irq_default_affinity);
cpulist_parse(str, irq_default_affinity);
/*
* Set at least the boot cpu. We don't want to end up with
* bugreports caused by random commandline masks
*/
cpumask_set_cpu(smp_processor_id(), irq_default_affinity);
return 1;
}
__setup("irqaffinity=", irq_affinity_setup);
static void __init init_irq_default_affinity(void)
{
if (!cpumask_available(irq_default_affinity))
zalloc_cpumask_var(&irq_default_affinity, GFP_NOWAIT);
if (cpumask_empty(irq_default_affinity))
cpumask_setall(irq_default_affinity);
}
#else
static void __init init_irq_default_affinity(void)
{
}
#endif
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
#ifdef CONFIG_SMP
static int alloc_masks(struct irq_desc *desc, int node)
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
{
if (!zalloc_cpumask_var_node(&desc->irq_common_data.affinity,
GFP_KERNEL, node))
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
return -ENOMEM;
#ifdef CONFIG_GENERIC_IRQ_EFFECTIVE_AFF_MASK
if (!zalloc_cpumask_var_node(&desc->irq_common_data.effective_affinity,
GFP_KERNEL, node)) {
free_cpumask_var(desc->irq_common_data.affinity);
return -ENOMEM;
}
#endif
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
#ifdef CONFIG_GENERIC_PENDING_IRQ
if (!zalloc_cpumask_var_node(&desc->pending_mask, GFP_KERNEL, node)) {
#ifdef CONFIG_GENERIC_IRQ_EFFECTIVE_AFF_MASK
free_cpumask_var(desc->irq_common_data.effective_affinity);
#endif
free_cpumask_var(desc->irq_common_data.affinity);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
return -ENOMEM;
}
#endif
return 0;
}
static void desc_smp_init(struct irq_desc *desc, int node,
const struct cpumask *affinity)
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
{
if (!affinity)
affinity = irq_default_affinity;
cpumask_copy(desc->irq_common_data.affinity, affinity);
#ifdef CONFIG_GENERIC_PENDING_IRQ
cpumask_clear(desc->pending_mask);
#endif
#ifdef CONFIG_NUMA
desc->irq_common_data.node = node;
#endif
}
static void free_masks(struct irq_desc *desc)
{
#ifdef CONFIG_GENERIC_PENDING_IRQ
free_cpumask_var(desc->pending_mask);
#endif
free_cpumask_var(desc->irq_common_data.affinity);
#ifdef CONFIG_GENERIC_IRQ_EFFECTIVE_AFF_MASK
free_cpumask_var(desc->irq_common_data.effective_affinity);
#endif
}
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
#else
static inline int
alloc_masks(struct irq_desc *desc, int node) { return 0; }
static inline void
desc_smp_init(struct irq_desc *desc, int node, const struct cpumask *affinity) { }
static inline void free_masks(struct irq_desc *desc) { }
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
#endif
static void desc_set_defaults(unsigned int irq, struct irq_desc *desc, int node,
const struct cpumask *affinity, struct module *owner)
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
{
int cpu;
desc->irq_common_data.handler_data = NULL;
desc->irq_common_data.msi_desc = NULL;
desc->irq_data.common = &desc->irq_common_data;
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
desc->irq_data.irq = irq;
desc->irq_data.chip = &no_irq_chip;
desc->irq_data.chip_data = NULL;
irq_settings_clr_and_set(desc, ~0, _IRQ_DEFAULT_INIT_FLAGS);
irqd_set(&desc->irq_data, IRQD_IRQ_DISABLED);
irqd_set(&desc->irq_data, IRQD_IRQ_MASKED);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
desc->handle_irq = handle_bad_irq;
desc->depth = 1;
desc->irq_count = 0;
desc->irqs_unhandled = 0;
genirq: Avoid summation loops for /proc/stat Waiman reported that on large systems with a large amount of interrupts the readout of /proc/stat takes a long time to sum up the interrupt statistics. In principle this is not a problem. but for unknown reasons some enterprise quality software reads /proc/stat with a high frequency. The reason for this is that interrupt statistics are accounted per cpu. So the /proc/stat logic has to sum up the interrupt stats for each interrupt. This can be largely avoided for interrupts which are not marked as 'PER_CPU' interrupts by simply adding a per interrupt summation counter which is incremented along with the per interrupt per cpu counter. The PER_CPU interrupts need to avoid that and use only per cpu accounting because they share the interrupt number and the interrupt descriptor and concurrent updates would conflict or require unwanted synchronization. Reported-by: Waiman Long <longman@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Waiman Long <longman@redhat.com> Reviewed-by: Marc Zyngier <marc.zyngier@arm.com> Reviewed-by: Davidlohr Bueso <dbueso@suse.de> Cc: Matthew Wilcox <willy@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Kees Cook <keescook@chromium.org> Cc: linux-fsdevel@vger.kernel.org Cc: Davidlohr Bueso <dave@stgolabs.net> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Daniel Colascione <dancol@google.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Randy Dunlap <rdunlap@infradead.org> Link: https://lkml.kernel.org/r/20190208135020.925487496@linutronix.de 8<------------- v2: Undo the unintentional layout change of struct irq_desc. include/linux/irqdesc.h | 1 + kernel/irq/chip.c | 12 ++++++++++-- kernel/irq/internals.h | 8 +++++++- kernel/irq/irqdesc.c | 7 ++++++- 4 files changed, 24 insertions(+), 4 deletions(-)
2019-02-08 13:48:03 +00:00
desc->tot_count = 0;
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
desc->name = NULL;
desc->owner = owner;
for_each_possible_cpu(cpu)
*per_cpu_ptr(desc->kstat_irqs, cpu) = (struct irqstat) { };
desc_smp_init(desc, node, affinity);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
}
int nr_irqs = NR_IRQS;
EXPORT_SYMBOL_GPL(nr_irqs);
static DEFINE_MUTEX(sparse_irq_lock);
static struct maple_tree sparse_irqs = MTREE_INIT_EXT(sparse_irqs,
MT_FLAGS_ALLOC_RANGE |
MT_FLAGS_LOCK_EXTERN |
MT_FLAGS_USE_RCU,
sparse_irq_lock);
static int irq_find_free_area(unsigned int from, unsigned int cnt)
{
MA_STATE(mas, &sparse_irqs, 0, 0);
if (mas_empty_area(&mas, from, MAX_SPARSE_IRQS, cnt))
return -ENOSPC;
return mas.index;
}
static unsigned int irq_find_at_or_after(unsigned int offset)
{
unsigned long index = offset;
struct irq_desc *desc;
guard(rcu)();
desc = mt_find(&sparse_irqs, &index, nr_irqs);
return desc ? irq_desc_get_irq(desc) : nr_irqs;
}
static void irq_insert_desc(unsigned int irq, struct irq_desc *desc)
{
MA_STATE(mas, &sparse_irqs, irq, irq);
WARN_ON(mas_store_gfp(&mas, desc, GFP_KERNEL) != 0);
}
static void delete_irq_desc(unsigned int irq)
{
MA_STATE(mas, &sparse_irqs, irq, irq);
mas_erase(&mas);
}
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
#ifdef CONFIG_SPARSE_IRQ
static const struct kobj_type irq_kobj_type;
#endif
static int init_desc(struct irq_desc *desc, int irq, int node,
unsigned int flags,
const struct cpumask *affinity,
struct module *owner)
{
desc->kstat_irqs = alloc_percpu(struct irqstat);
if (!desc->kstat_irqs)
return -ENOMEM;
if (alloc_masks(desc, node)) {
free_percpu(desc->kstat_irqs);
return -ENOMEM;
}
raw_spin_lock_init(&desc->lock);
lockdep_set_class(&desc->lock, &irq_desc_lock_class);
mutex_init(&desc->request_mutex);
init_waitqueue_head(&desc->wait_for_threads);
desc_set_defaults(irq, desc, node, affinity, owner);
irqd_set(&desc->irq_data, flags);
irq_resend_init(desc);
#ifdef CONFIG_SPARSE_IRQ
kobject_init(&desc->kobj, &irq_kobj_type);
init_rcu_head(&desc->rcu);
#endif
return 0;
}
#ifdef CONFIG_SPARSE_IRQ
static void irq_kobj_release(struct kobject *kobj);
#ifdef CONFIG_SYSFS
static struct kobject *irq_kobj_base;
#define IRQ_ATTR_RO(_name) \
static struct kobj_attribute _name##_attr = __ATTR_RO(_name)
static ssize_t per_cpu_count_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
struct irq_desc *desc = container_of(kobj, struct irq_desc, kobj);
ssize_t ret = 0;
char *p = "";
int cpu;
for_each_possible_cpu(cpu) {
unsigned int c = irq_desc_kstat_cpu(desc, cpu);
ret += scnprintf(buf + ret, PAGE_SIZE - ret, "%s%u", p, c);
p = ",";
}
ret += scnprintf(buf + ret, PAGE_SIZE - ret, "\n");
return ret;
}
IRQ_ATTR_RO(per_cpu_count);
static ssize_t chip_name_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
struct irq_desc *desc = container_of(kobj, struct irq_desc, kobj);
ssize_t ret = 0;
raw_spin_lock_irq(&desc->lock);
if (desc->irq_data.chip && desc->irq_data.chip->name) {
ret = scnprintf(buf, PAGE_SIZE, "%s\n",
desc->irq_data.chip->name);
}
raw_spin_unlock_irq(&desc->lock);
return ret;
}
IRQ_ATTR_RO(chip_name);
static ssize_t hwirq_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
struct irq_desc *desc = container_of(kobj, struct irq_desc, kobj);
ssize_t ret = 0;
raw_spin_lock_irq(&desc->lock);
if (desc->irq_data.domain)
ret = sprintf(buf, "%lu\n", desc->irq_data.hwirq);
raw_spin_unlock_irq(&desc->lock);
return ret;
}
IRQ_ATTR_RO(hwirq);
static ssize_t type_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
struct irq_desc *desc = container_of(kobj, struct irq_desc, kobj);
ssize_t ret = 0;
raw_spin_lock_irq(&desc->lock);
ret = sprintf(buf, "%s\n",
irqd_is_level_type(&desc->irq_data) ? "level" : "edge");
raw_spin_unlock_irq(&desc->lock);
return ret;
}
IRQ_ATTR_RO(type);
static ssize_t wakeup_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
struct irq_desc *desc = container_of(kobj, struct irq_desc, kobj);
ssize_t ret = 0;
raw_spin_lock_irq(&desc->lock);
ret = sprintf(buf, "%s\n",
irqd_is_wakeup_set(&desc->irq_data) ? "enabled" : "disabled");
raw_spin_unlock_irq(&desc->lock);
return ret;
}
IRQ_ATTR_RO(wakeup);
static ssize_t name_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
struct irq_desc *desc = container_of(kobj, struct irq_desc, kobj);
ssize_t ret = 0;
raw_spin_lock_irq(&desc->lock);
if (desc->name)
ret = scnprintf(buf, PAGE_SIZE, "%s\n", desc->name);
raw_spin_unlock_irq(&desc->lock);
return ret;
}
IRQ_ATTR_RO(name);
static ssize_t actions_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
struct irq_desc *desc = container_of(kobj, struct irq_desc, kobj);
struct irqaction *action;
ssize_t ret = 0;
char *p = "";
raw_spin_lock_irq(&desc->lock);
for_each_action_of_desc(desc, action) {
ret += scnprintf(buf + ret, PAGE_SIZE - ret, "%s%s",
p, action->name);
p = ",";
}
raw_spin_unlock_irq(&desc->lock);
if (ret)
ret += scnprintf(buf + ret, PAGE_SIZE - ret, "\n");
return ret;
}
IRQ_ATTR_RO(actions);
static struct attribute *irq_attrs[] = {
&per_cpu_count_attr.attr,
&chip_name_attr.attr,
&hwirq_attr.attr,
&type_attr.attr,
&wakeup_attr.attr,
&name_attr.attr,
&actions_attr.attr,
NULL
};
ATTRIBUTE_GROUPS(irq);
static const struct kobj_type irq_kobj_type = {
.release = irq_kobj_release,
.sysfs_ops = &kobj_sysfs_ops,
.default_groups = irq_groups,
};
static void irq_sysfs_add(int irq, struct irq_desc *desc)
{
if (irq_kobj_base) {
/*
* Continue even in case of failure as this is nothing
genirq/irqdesc: Don't try to remove non-existing sysfs files Fault injection tests trigger warnings like this: kernfs: can not remove 'chip_name', no directory WARNING: CPU: 0 PID: 253 at fs/kernfs/dir.c:1616 kernfs_remove_by_name_ns+0xce/0xe0 RIP: 0010:kernfs_remove_by_name_ns+0xce/0xe0 Call Trace: <TASK> remove_files.isra.1+0x3f/0xb0 sysfs_remove_group+0x68/0xe0 sysfs_remove_groups+0x41/0x70 __kobject_del+0x45/0xc0 kobject_del+0x29/0x40 free_desc+0x42/0x70 irq_free_descs+0x5e/0x90 The reason is that the interrupt descriptor sysfs handling does not roll back on a failing kobject_add() during allocation. If the descriptor is freed later on, kobject_del() is invoked with a not added kobject resulting in the above warnings. A proper rollback in case of a kobject_add() failure would be the straight forward solution. But this is not possible due to the way how interrupt descriptor sysfs handling works. Interrupt descriptors are allocated before sysfs becomes available. So the sysfs files for the early allocated descriptors are added later in the boot process. At this point there can be nothing useful done about a failing kobject_add(). For consistency the interrupt descriptor allocation always treats kobject_add() failures as non-critical and just emits a warning. To solve this problem, keep track in the interrupt descriptor whether kobject_add() was successful or not and make the invocation of kobject_del() conditional on that. [ tglx: Massage changelog, comments and use a state bit. ] Fixes: ecb3f394c5db ("genirq: Expose interrupt information through sysfs") Signed-off-by: Yang Yingliang <yangyingliang@huawei.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Link: https://lore.kernel.org/r/20221128151612.1786122-1-yangyingliang@huawei.com
2022-11-28 15:16:12 +00:00
* crucial and failures in the late irq_sysfs_init()
* cannot be rolled back.
*/
if (kobject_add(&desc->kobj, irq_kobj_base, "%d", irq))
pr_warn("Failed to add kobject for irq %d\n", irq);
genirq/irqdesc: Don't try to remove non-existing sysfs files Fault injection tests trigger warnings like this: kernfs: can not remove 'chip_name', no directory WARNING: CPU: 0 PID: 253 at fs/kernfs/dir.c:1616 kernfs_remove_by_name_ns+0xce/0xe0 RIP: 0010:kernfs_remove_by_name_ns+0xce/0xe0 Call Trace: <TASK> remove_files.isra.1+0x3f/0xb0 sysfs_remove_group+0x68/0xe0 sysfs_remove_groups+0x41/0x70 __kobject_del+0x45/0xc0 kobject_del+0x29/0x40 free_desc+0x42/0x70 irq_free_descs+0x5e/0x90 The reason is that the interrupt descriptor sysfs handling does not roll back on a failing kobject_add() during allocation. If the descriptor is freed later on, kobject_del() is invoked with a not added kobject resulting in the above warnings. A proper rollback in case of a kobject_add() failure would be the straight forward solution. But this is not possible due to the way how interrupt descriptor sysfs handling works. Interrupt descriptors are allocated before sysfs becomes available. So the sysfs files for the early allocated descriptors are added later in the boot process. At this point there can be nothing useful done about a failing kobject_add(). For consistency the interrupt descriptor allocation always treats kobject_add() failures as non-critical and just emits a warning. To solve this problem, keep track in the interrupt descriptor whether kobject_add() was successful or not and make the invocation of kobject_del() conditional on that. [ tglx: Massage changelog, comments and use a state bit. ] Fixes: ecb3f394c5db ("genirq: Expose interrupt information through sysfs") Signed-off-by: Yang Yingliang <yangyingliang@huawei.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Link: https://lore.kernel.org/r/20221128151612.1786122-1-yangyingliang@huawei.com
2022-11-28 15:16:12 +00:00
else
desc->istate |= IRQS_SYSFS;
}
}
static void irq_sysfs_del(struct irq_desc *desc)
{
/*
genirq/irqdesc: Don't try to remove non-existing sysfs files Fault injection tests trigger warnings like this: kernfs: can not remove 'chip_name', no directory WARNING: CPU: 0 PID: 253 at fs/kernfs/dir.c:1616 kernfs_remove_by_name_ns+0xce/0xe0 RIP: 0010:kernfs_remove_by_name_ns+0xce/0xe0 Call Trace: <TASK> remove_files.isra.1+0x3f/0xb0 sysfs_remove_group+0x68/0xe0 sysfs_remove_groups+0x41/0x70 __kobject_del+0x45/0xc0 kobject_del+0x29/0x40 free_desc+0x42/0x70 irq_free_descs+0x5e/0x90 The reason is that the interrupt descriptor sysfs handling does not roll back on a failing kobject_add() during allocation. If the descriptor is freed later on, kobject_del() is invoked with a not added kobject resulting in the above warnings. A proper rollback in case of a kobject_add() failure would be the straight forward solution. But this is not possible due to the way how interrupt descriptor sysfs handling works. Interrupt descriptors are allocated before sysfs becomes available. So the sysfs files for the early allocated descriptors are added later in the boot process. At this point there can be nothing useful done about a failing kobject_add(). For consistency the interrupt descriptor allocation always treats kobject_add() failures as non-critical and just emits a warning. To solve this problem, keep track in the interrupt descriptor whether kobject_add() was successful or not and make the invocation of kobject_del() conditional on that. [ tglx: Massage changelog, comments and use a state bit. ] Fixes: ecb3f394c5db ("genirq: Expose interrupt information through sysfs") Signed-off-by: Yang Yingliang <yangyingliang@huawei.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Link: https://lore.kernel.org/r/20221128151612.1786122-1-yangyingliang@huawei.com
2022-11-28 15:16:12 +00:00
* Only invoke kobject_del() when kobject_add() was successfully
* invoked for the descriptor. This covers both early boot, where
* sysfs is not initialized yet, and the case of a failed
* kobject_add() invocation.
*/
genirq/irqdesc: Don't try to remove non-existing sysfs files Fault injection tests trigger warnings like this: kernfs: can not remove 'chip_name', no directory WARNING: CPU: 0 PID: 253 at fs/kernfs/dir.c:1616 kernfs_remove_by_name_ns+0xce/0xe0 RIP: 0010:kernfs_remove_by_name_ns+0xce/0xe0 Call Trace: <TASK> remove_files.isra.1+0x3f/0xb0 sysfs_remove_group+0x68/0xe0 sysfs_remove_groups+0x41/0x70 __kobject_del+0x45/0xc0 kobject_del+0x29/0x40 free_desc+0x42/0x70 irq_free_descs+0x5e/0x90 The reason is that the interrupt descriptor sysfs handling does not roll back on a failing kobject_add() during allocation. If the descriptor is freed later on, kobject_del() is invoked with a not added kobject resulting in the above warnings. A proper rollback in case of a kobject_add() failure would be the straight forward solution. But this is not possible due to the way how interrupt descriptor sysfs handling works. Interrupt descriptors are allocated before sysfs becomes available. So the sysfs files for the early allocated descriptors are added later in the boot process. At this point there can be nothing useful done about a failing kobject_add(). For consistency the interrupt descriptor allocation always treats kobject_add() failures as non-critical and just emits a warning. To solve this problem, keep track in the interrupt descriptor whether kobject_add() was successful or not and make the invocation of kobject_del() conditional on that. [ tglx: Massage changelog, comments and use a state bit. ] Fixes: ecb3f394c5db ("genirq: Expose interrupt information through sysfs") Signed-off-by: Yang Yingliang <yangyingliang@huawei.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Link: https://lore.kernel.org/r/20221128151612.1786122-1-yangyingliang@huawei.com
2022-11-28 15:16:12 +00:00
if (desc->istate & IRQS_SYSFS)
kobject_del(&desc->kobj);
}
static int __init irq_sysfs_init(void)
{
struct irq_desc *desc;
int irq;
/* Prevent concurrent irq alloc/free */
irq_lock_sparse();
irq_kobj_base = kobject_create_and_add("irq", kernel_kobj);
if (!irq_kobj_base) {
irq_unlock_sparse();
return -ENOMEM;
}
/* Add the already allocated interrupts */
for_each_irq_desc(irq, desc)
irq_sysfs_add(irq, desc);
irq_unlock_sparse();
return 0;
}
postcore_initcall(irq_sysfs_init);
#else /* !CONFIG_SYSFS */
static const struct kobj_type irq_kobj_type = {
.release = irq_kobj_release,
};
static void irq_sysfs_add(int irq, struct irq_desc *desc) {}
static void irq_sysfs_del(struct irq_desc *desc) {}
#endif /* CONFIG_SYSFS */
struct irq_desc *irq_to_desc(unsigned int irq)
{
return mtree_load(&sparse_irqs, irq);
}
#ifdef CONFIG_KVM_BOOK3S_64_HV_MODULE
EXPORT_SYMBOL_GPL(irq_to_desc);
#endif
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
void irq_lock_sparse(void)
{
mutex_lock(&sparse_irq_lock);
}
void irq_unlock_sparse(void)
{
mutex_unlock(&sparse_irq_lock);
}
static struct irq_desc *alloc_desc(int irq, int node, unsigned int flags,
const struct cpumask *affinity,
struct module *owner)
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
{
struct irq_desc *desc;
int ret;
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
desc = kzalloc_node(sizeof(*desc), GFP_KERNEL, node);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
if (!desc)
return NULL;
ret = init_desc(desc, irq, node, flags, affinity, owner);
if (unlikely(ret)) {
kfree(desc);
return NULL;
}
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
return desc;
}
static void irq_kobj_release(struct kobject *kobj)
{
struct irq_desc *desc = container_of(kobj, struct irq_desc, kobj);
free_masks(desc);
free_percpu(desc->kstat_irqs);
kfree(desc);
}
static void delayed_free_desc(struct rcu_head *rhp)
{
struct irq_desc *desc = container_of(rhp, struct irq_desc, rcu);
kobject_put(&desc->kobj);
}
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
static void free_desc(unsigned int irq)
{
struct irq_desc *desc = irq_to_desc(irq);
genirq/debugfs: Add proper debugfs interface Debugging (hierarchical) interupt domains is tedious as there is no information about the hierarchy and no information about states of interrupts in the various domain levels. Add a debugfs directory 'irq' and subdirectories 'domains' and 'irqs'. The domains directory contains the domain files. The content is information about the domain. If the domain is part of a hierarchy then the parent domains are printed as well. # ls /sys/kernel/debug/irq/domains/ default INTEL-IR-2 INTEL-IR-MSI-2 IO-APIC-IR-2 PCI-MSI DMAR-MSI INTEL-IR-3 INTEL-IR-MSI-3 IO-APIC-IR-3 unknown-1 INTEL-IR-0 INTEL-IR-MSI-0 IO-APIC-IR-0 IO-APIC-IR-4 VECTOR INTEL-IR-1 INTEL-IR-MSI-1 IO-APIC-IR-1 PCI-HT # cat /sys/kernel/debug/irq/domains/VECTOR name: VECTOR size: 0 mapped: 216 flags: 0x00000041 # cat /sys/kernel/debug/irq/domains/IO-APIC-IR-0 name: IO-APIC-IR-0 size: 24 mapped: 19 flags: 0x00000041 parent: INTEL-IR-3 name: INTEL-IR-3 size: 65536 mapped: 167 flags: 0x00000041 parent: VECTOR name: VECTOR size: 0 mapped: 216 flags: 0x00000041 Unfortunately there is no per cpu information about the VECTOR domain (yet). The irqs directory contains detailed information about mapped interrupts. # cat /sys/kernel/debug/irq/irqs/3 handler: handle_edge_irq status: 0x00004000 istate: 0x00000000 ddepth: 1 wdepth: 0 dstate: 0x01018000 IRQD_IRQ_DISABLED IRQD_SINGLE_TARGET IRQD_MOVE_PCNTXT node: 0 affinity: 0-143 effectiv: 0 pending: domain: IO-APIC-IR-0 hwirq: 0x3 chip: IR-IO-APIC flags: 0x10 IRQCHIP_SKIP_SET_WAKE parent: domain: INTEL-IR-3 hwirq: 0x20000 chip: INTEL-IR flags: 0x0 parent: domain: VECTOR hwirq: 0x3 chip: APIC flags: 0x0 This was developed to simplify the debugging of the managed affinity changes. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Marc Zyngier <marc.zyngier@arm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Keith Busch <keith.busch@intel.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Christoph Hellwig <hch@lst.de> Link: http://lkml.kernel.org/r/20170619235444.537566163@linutronix.de Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2017-06-19 23:37:17 +00:00
irq_remove_debugfs_entry(desc);
unregister_irq_proc(irq, desc);
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
/*
* sparse_irq_lock protects also show_interrupts() and
* kstat_irq_usr(). Once we deleted the descriptor from the
* sparse tree we can free it. Access in proc will fail to
* lookup the descriptor.
*
* The sysfs entry must be serialized against a concurrent
* irq_sysfs_init() as well.
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
*/
irq_sysfs_del(desc);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
delete_irq_desc(irq);
/*
* We free the descriptor, masks and stat fields via RCU. That
* allows demultiplex interrupts to do rcu based management of
* the child interrupts.
* This also allows us to use rcu in kstat_irqs_usr().
*/
call_rcu(&desc->rcu, delayed_free_desc);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
}
static int alloc_descs(unsigned int start, unsigned int cnt, int node,
genirq/core: Introduce struct irq_affinity_desc The interrupt affinity management uses straight cpumask pointers to convey the automatically assigned affinity masks for managed interrupts. The core interrupt descriptor allocation also decides based on the pointer being non NULL whether an interrupt is managed or not. Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. To remedy that situation it's required to convey more information than the cpumasks through various interfaces related to interrupt descriptor allocation. Instead of adding yet another argument, create a new data structure 'irq_affinity_desc' which for now just contains the cpumask. This struct can be expanded to convey auxilliary information in the next step. No functional change, just preparatory work. [ tglx: Simplified logic and clarified changelog ] Suggested-by: Thomas Gleixner <tglx@linutronix.de> Suggested-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: kashyap.desai@broadcom.com Cc: shivasharan.srikanteshwara@broadcom.com Cc: sumit.saxena@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-2-douliyangs@gmail.com
2018-12-04 15:51:20 +00:00
const struct irq_affinity_desc *affinity,
struct module *owner)
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
{
struct irq_desc *desc;
int i;
/* Validate affinity mask(s) */
if (affinity) {
for (i = 0; i < cnt; i++) {
genirq/core: Introduce struct irq_affinity_desc The interrupt affinity management uses straight cpumask pointers to convey the automatically assigned affinity masks for managed interrupts. The core interrupt descriptor allocation also decides based on the pointer being non NULL whether an interrupt is managed or not. Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. To remedy that situation it's required to convey more information than the cpumasks through various interfaces related to interrupt descriptor allocation. Instead of adding yet another argument, create a new data structure 'irq_affinity_desc' which for now just contains the cpumask. This struct can be expanded to convey auxilliary information in the next step. No functional change, just preparatory work. [ tglx: Simplified logic and clarified changelog ] Suggested-by: Thomas Gleixner <tglx@linutronix.de> Suggested-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: kashyap.desai@broadcom.com Cc: shivasharan.srikanteshwara@broadcom.com Cc: sumit.saxena@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-2-douliyangs@gmail.com
2018-12-04 15:51:20 +00:00
if (cpumask_empty(&affinity[i].mask))
return -EINVAL;
}
}
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
for (i = 0; i < cnt; i++) {
genirq/core: Introduce struct irq_affinity_desc The interrupt affinity management uses straight cpumask pointers to convey the automatically assigned affinity masks for managed interrupts. The core interrupt descriptor allocation also decides based on the pointer being non NULL whether an interrupt is managed or not. Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. To remedy that situation it's required to convey more information than the cpumasks through various interfaces related to interrupt descriptor allocation. Instead of adding yet another argument, create a new data structure 'irq_affinity_desc' which for now just contains the cpumask. This struct can be expanded to convey auxilliary information in the next step. No functional change, just preparatory work. [ tglx: Simplified logic and clarified changelog ] Suggested-by: Thomas Gleixner <tglx@linutronix.de> Suggested-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: kashyap.desai@broadcom.com Cc: shivasharan.srikanteshwara@broadcom.com Cc: sumit.saxena@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-2-douliyangs@gmail.com
2018-12-04 15:51:20 +00:00
const struct cpumask *mask = NULL;
genirq/affinity: Add is_managed to struct irq_affinity_desc Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. That limitation was reported by Kashyap and Sumit. Expand struct irq_affinity_desc with a new bit 'is_managed' which is set for truly managed interrupts (queue interrupts) and cleared for the general device interrupts. [ tglx: Simplify code and massage changelog ] Reported-by: Kashyap Desai <kashyap.desai@broadcom.com> Reported-by: Sumit Saxena <sumit.saxena@broadcom.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: shivasharan.srikanteshwara@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: bhelgaas@google.com Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-3-douliyangs@gmail.com
2018-12-04 15:51:21 +00:00
unsigned int flags = 0;
genirq/core: Introduce struct irq_affinity_desc The interrupt affinity management uses straight cpumask pointers to convey the automatically assigned affinity masks for managed interrupts. The core interrupt descriptor allocation also decides based on the pointer being non NULL whether an interrupt is managed or not. Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. To remedy that situation it's required to convey more information than the cpumasks through various interfaces related to interrupt descriptor allocation. Instead of adding yet another argument, create a new data structure 'irq_affinity_desc' which for now just contains the cpumask. This struct can be expanded to convey auxilliary information in the next step. No functional change, just preparatory work. [ tglx: Simplified logic and clarified changelog ] Suggested-by: Thomas Gleixner <tglx@linutronix.de> Suggested-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: kashyap.desai@broadcom.com Cc: shivasharan.srikanteshwara@broadcom.com Cc: sumit.saxena@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-2-douliyangs@gmail.com
2018-12-04 15:51:20 +00:00
if (affinity) {
genirq/affinity: Add is_managed to struct irq_affinity_desc Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. That limitation was reported by Kashyap and Sumit. Expand struct irq_affinity_desc with a new bit 'is_managed' which is set for truly managed interrupts (queue interrupts) and cleared for the general device interrupts. [ tglx: Simplify code and massage changelog ] Reported-by: Kashyap Desai <kashyap.desai@broadcom.com> Reported-by: Sumit Saxena <sumit.saxena@broadcom.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: shivasharan.srikanteshwara@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: bhelgaas@google.com Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-3-douliyangs@gmail.com
2018-12-04 15:51:21 +00:00
if (affinity->is_managed) {
flags = IRQD_AFFINITY_MANAGED |
IRQD_MANAGED_SHUTDOWN;
}
genirq/core: Introduce struct irq_affinity_desc The interrupt affinity management uses straight cpumask pointers to convey the automatically assigned affinity masks for managed interrupts. The core interrupt descriptor allocation also decides based on the pointer being non NULL whether an interrupt is managed or not. Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. To remedy that situation it's required to convey more information than the cpumasks through various interfaces related to interrupt descriptor allocation. Instead of adding yet another argument, create a new data structure 'irq_affinity_desc' which for now just contains the cpumask. This struct can be expanded to convey auxilliary information in the next step. No functional change, just preparatory work. [ tglx: Simplified logic and clarified changelog ] Suggested-by: Thomas Gleixner <tglx@linutronix.de> Suggested-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: kashyap.desai@broadcom.com Cc: shivasharan.srikanteshwara@broadcom.com Cc: sumit.saxena@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-2-douliyangs@gmail.com
2018-12-04 15:51:20 +00:00
mask = &affinity->mask;
genirq/affinity: Add is_managed to struct irq_affinity_desc Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. That limitation was reported by Kashyap and Sumit. Expand struct irq_affinity_desc with a new bit 'is_managed' which is set for truly managed interrupts (queue interrupts) and cleared for the general device interrupts. [ tglx: Simplify code and massage changelog ] Reported-by: Kashyap Desai <kashyap.desai@broadcom.com> Reported-by: Sumit Saxena <sumit.saxena@broadcom.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: shivasharan.srikanteshwara@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: bhelgaas@google.com Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-3-douliyangs@gmail.com
2018-12-04 15:51:21 +00:00
node = cpu_to_node(cpumask_first(mask));
affinity++;
}
genirq/affinity: Add is_managed to struct irq_affinity_desc Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. That limitation was reported by Kashyap and Sumit. Expand struct irq_affinity_desc with a new bit 'is_managed' which is set for truly managed interrupts (queue interrupts) and cleared for the general device interrupts. [ tglx: Simplify code and massage changelog ] Reported-by: Kashyap Desai <kashyap.desai@broadcom.com> Reported-by: Sumit Saxena <sumit.saxena@broadcom.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: shivasharan.srikanteshwara@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: bhelgaas@google.com Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-3-douliyangs@gmail.com
2018-12-04 15:51:21 +00:00
desc = alloc_desc(start + i, node, flags, mask, owner);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
if (!desc)
goto err;
irq_insert_desc(start + i, desc);
irq_sysfs_add(start + i, desc);
irq_add_debugfs_entry(start + i, desc);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
}
return start;
err:
for (i--; i >= 0; i--)
free_desc(start + i);
return -ENOMEM;
}
static int irq_expand_nr_irqs(unsigned int nr)
{
if (nr > MAX_SPARSE_IRQS)
return -ENOMEM;
nr_irqs = nr;
return 0;
}
int __init early_irq_init(void)
{
int i, initcnt, node = first_online_node;
struct irq_desc *desc;
init_irq_default_affinity();
/* Let arch update nr_irqs and return the nr of preallocated irqs */
initcnt = arch_probe_nr_irqs();
printk(KERN_INFO "NR_IRQS: %d, nr_irqs: %d, preallocated irqs: %d\n",
NR_IRQS, nr_irqs, initcnt);
if (WARN_ON(nr_irqs > MAX_SPARSE_IRQS))
nr_irqs = MAX_SPARSE_IRQS;
genirq: Prevent access beyond allocated_irqs bitmap Lars-Peter Clausen pointed out: I stumbled upon this while looking through the existing archs using SPARSE_IRQ. Even with SPARSE_IRQ the NR_IRQS is still the upper limit for the number of IRQs. Both PXA and MMP set NR_IRQS to IRQ_BOARD_START, with IRQ_BOARD_START being the number of IRQs used by the core. In various machine files the nr_irqs field of the ARM machine defintion struct is then set to "IRQ_BOARD_START + NR_BOARD_IRQS". As a result "nr_irqs" will greater then NR_IRQS which then again causes the "allocated_irqs" bitmap in the core irq code to be accessed beyond its size overwriting unrelated data. The core code really misses a sanity check there. This went unnoticed so far as by chance the compiler/linker places data behind that bitmap which gets initialized later on those affected platforms. So the obvious fix would be to add a sanity check in early_irq_init() and break all affected platforms. Though that check wants to be backported to stable as well, which will require to fix all known problematic platforms and probably some more yet not known ones as well. Lots of churn. A way simpler solution is to allocate a slightly larger bitmap and avoid the whole churn w/o breaking anything. Add a few warnings when an arch returns utter crap. Reported-by: Lars-Peter Clausen <lars@metafoo.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@kernel.org # .37 Cc: Haojian Zhuang <haojian.zhuang@marvell.com> Cc: Eric Miao <eric.y.miao@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org>
2011-02-17 16:45:15 +00:00
if (WARN_ON(initcnt > MAX_SPARSE_IRQS))
initcnt = MAX_SPARSE_IRQS;
genirq: Prevent access beyond allocated_irqs bitmap Lars-Peter Clausen pointed out: I stumbled upon this while looking through the existing archs using SPARSE_IRQ. Even with SPARSE_IRQ the NR_IRQS is still the upper limit for the number of IRQs. Both PXA and MMP set NR_IRQS to IRQ_BOARD_START, with IRQ_BOARD_START being the number of IRQs used by the core. In various machine files the nr_irqs field of the ARM machine defintion struct is then set to "IRQ_BOARD_START + NR_BOARD_IRQS". As a result "nr_irqs" will greater then NR_IRQS which then again causes the "allocated_irqs" bitmap in the core irq code to be accessed beyond its size overwriting unrelated data. The core code really misses a sanity check there. This went unnoticed so far as by chance the compiler/linker places data behind that bitmap which gets initialized later on those affected platforms. So the obvious fix would be to add a sanity check in early_irq_init() and break all affected platforms. Though that check wants to be backported to stable as well, which will require to fix all known problematic platforms and probably some more yet not known ones as well. Lots of churn. A way simpler solution is to allocate a slightly larger bitmap and avoid the whole churn w/o breaking anything. Add a few warnings when an arch returns utter crap. Reported-by: Lars-Peter Clausen <lars@metafoo.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@kernel.org # .37 Cc: Haojian Zhuang <haojian.zhuang@marvell.com> Cc: Eric Miao <eric.y.miao@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org>
2011-02-17 16:45:15 +00:00
if (initcnt > nr_irqs)
nr_irqs = initcnt;
for (i = 0; i < initcnt; i++) {
desc = alloc_desc(i, node, 0, NULL, NULL);
irq_insert_desc(i, desc);
}
return arch_early_irq_init();
}
#else /* !CONFIG_SPARSE_IRQ */
struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = {
[0 ... NR_IRQS-1] = {
.handle_irq = handle_bad_irq,
.depth = 1,
.lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc->lock),
}
};
int __init early_irq_init(void)
{
int count, i, node = first_online_node;
int ret;
init_irq_default_affinity();
printk(KERN_INFO "NR_IRQS: %d\n", NR_IRQS);
count = ARRAY_SIZE(irq_desc);
for (i = 0; i < count; i++) {
ret = init_desc(irq_desc + i, i, node, 0, NULL, NULL);
if (unlikely(ret))
goto __free_desc_res;
}
return arch_early_irq_init();
__free_desc_res:
while (--i >= 0) {
free_masks(irq_desc + i);
free_percpu(irq_desc[i].kstat_irqs);
}
return ret;
}
struct irq_desc *irq_to_desc(unsigned int irq)
{
return (irq < NR_IRQS) ? irq_desc + irq : NULL;
}
EXPORT_SYMBOL(irq_to_desc);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
static void free_desc(unsigned int irq)
{
struct irq_desc *desc = irq_to_desc(irq);
unsigned long flags;
raw_spin_lock_irqsave(&desc->lock, flags);
desc_set_defaults(irq, desc, irq_desc_get_node(desc), NULL, NULL);
raw_spin_unlock_irqrestore(&desc->lock, flags);
delete_irq_desc(irq);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
}
static inline int alloc_descs(unsigned int start, unsigned int cnt, int node,
genirq/core: Introduce struct irq_affinity_desc The interrupt affinity management uses straight cpumask pointers to convey the automatically assigned affinity masks for managed interrupts. The core interrupt descriptor allocation also decides based on the pointer being non NULL whether an interrupt is managed or not. Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. To remedy that situation it's required to convey more information than the cpumasks through various interfaces related to interrupt descriptor allocation. Instead of adding yet another argument, create a new data structure 'irq_affinity_desc' which for now just contains the cpumask. This struct can be expanded to convey auxilliary information in the next step. No functional change, just preparatory work. [ tglx: Simplified logic and clarified changelog ] Suggested-by: Thomas Gleixner <tglx@linutronix.de> Suggested-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: kashyap.desai@broadcom.com Cc: shivasharan.srikanteshwara@broadcom.com Cc: sumit.saxena@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-2-douliyangs@gmail.com
2018-12-04 15:51:20 +00:00
const struct irq_affinity_desc *affinity,
struct module *owner)
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
{
u32 i;
for (i = 0; i < cnt; i++) {
struct irq_desc *desc = irq_to_desc(start + i);
desc->owner = owner;
irq_insert_desc(start + i, desc);
}
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
return start;
}
static int irq_expand_nr_irqs(unsigned int nr)
{
return -ENOMEM;
}
void irq_mark_irq(unsigned int irq)
{
mutex_lock(&sparse_irq_lock);
irq_insert_desc(irq, irq_desc + irq);
mutex_unlock(&sparse_irq_lock);
}
#ifdef CONFIG_GENERIC_IRQ_LEGACY
void irq_init_desc(unsigned int irq)
{
free_desc(irq);
}
#endif
#endif /* !CONFIG_SPARSE_IRQ */
int handle_irq_desc(struct irq_desc *desc)
{
struct irq_data *data;
if (!desc)
return -EINVAL;
data = irq_desc_get_irq_data(desc);
if (WARN_ON_ONCE(!in_hardirq() && handle_enforce_irqctx(data)))
return -EPERM;
generic_handle_irq_desc(desc);
return 0;
}
/**
* generic_handle_irq - Invoke the handler for a particular irq
* @irq: The irq number to handle
*
* Returns: 0 on success, or -EINVAL if conversion has failed
*
* This function must be called from an IRQ context with irq regs
* initialized.
*/
int generic_handle_irq(unsigned int irq)
{
return handle_irq_desc(irq_to_desc(irq));
}
EXPORT_SYMBOL_GPL(generic_handle_irq);
/**
* generic_handle_irq_safe - Invoke the handler for a particular irq from any
* context.
* @irq: The irq number to handle
*
* Returns: 0 on success, a negative value on error.
*
* This function can be called from any context (IRQ or process context). It
* will report an error if not invoked from IRQ context and the irq has been
* marked to enforce IRQ-context only.
*/
int generic_handle_irq_safe(unsigned int irq)
{
unsigned long flags;
int ret;
local_irq_save(flags);
ret = handle_irq_desc(irq_to_desc(irq));
local_irq_restore(flags);
return ret;
}
EXPORT_SYMBOL_GPL(generic_handle_irq_safe);
#ifdef CONFIG_IRQ_DOMAIN
genirq: Add irq_domain-aware core IRQ handler Calling irq_find_mapping from outside a irq_{enter,exit} section is unsafe and produces ugly messages if CONFIG_PROVE_RCU is enabled: If coming from the idle state, the rcu_read_lock call in irq_find_mapping will generate an unpleasant warning: <quote> =============================== [ INFO: suspicious RCU usage. ] 3.16.0-rc1+ #135 Not tainted ------------------------------- include/linux/rcupdate.h:871 rcu_read_lock() used illegally while idle! other info that might help us debug this: RCU used illegally from idle CPU! rcu_scheduler_active = 1, debug_locks = 0 RCU used illegally from extended quiescent state! 1 lock held by swapper/0/0: #0: (rcu_read_lock){......}, at: [<ffffffc00010206c>] irq_find_mapping+0x4c/0x198 </quote> As this issue is fairly widespread and involves at least three different architectures, a possible solution is to add a new handle_domain_irq entry point into the generic IRQ code that the interrupt controller code can call. This new function takes an irq_domain, and calls into irq_find_domain inside the irq_{enter,exit} block. An additional "lookup" parameter is used to allow non-domain architecture code to be replaced by this as well. Interrupt controllers can then be updated to use the new mechanism. This code is sitting behind a new CONFIG_HANDLE_DOMAIN_IRQ, as not all architectures implement set_irq_regs (yes, mn10300, I'm looking at you...). Reported-by: Vladimir Murzin <vladimir.murzin@arm.com> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Link: https://lkml.kernel.org/r/1409047421-27649-2-git-send-email-marc.zyngier@arm.com Signed-off-by: Jason Cooper <jason@lakedaemon.net>
2014-08-26 10:03:16 +00:00
/**
* generic_handle_domain_irq - Invoke the handler for a HW irq belonging
* to a domain.
* @domain: The domain where to perform the lookup
* @hwirq: The HW irq number to convert to a logical one
*
* Returns: 0 on success, or -EINVAL if conversion has failed
*
* This function must be called from an IRQ context with irq regs
* initialized.
*/
int generic_handle_domain_irq(struct irq_domain *domain, unsigned int hwirq)
{
return handle_irq_desc(irq_resolve_mapping(domain, hwirq));
}
EXPORT_SYMBOL_GPL(generic_handle_domain_irq);
/**
* generic_handle_irq_safe - Invoke the handler for a HW irq belonging
* to a domain from any context.
* @domain: The domain where to perform the lookup
* @hwirq: The HW irq number to convert to a logical one
*
* Returns: 0 on success, a negative value on error.
*
* This function can be called from any context (IRQ or process
* context). If the interrupt is marked as 'enforce IRQ-context only' then
* the function must be invoked from hard interrupt context.
*/
int generic_handle_domain_irq_safe(struct irq_domain *domain, unsigned int hwirq)
{
unsigned long flags;
int ret;
local_irq_save(flags);
ret = handle_irq_desc(irq_resolve_mapping(domain, hwirq));
local_irq_restore(flags);
return ret;
}
EXPORT_SYMBOL_GPL(generic_handle_domain_irq_safe);
irq: add a (temporary) CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY Going forward we want architecture/entry code to perform all the necessary work to enter/exit IRQ context, with irqchip code merely handling the mapping of the interrupt to any handler(s). Among other reasons, this is necessary to consistently fix some longstanding issues with the ordering of lockdep/RCU/tracing instrumentation which many architectures get wrong today in their entry code. Importantly, rcu_irq_{enter,exit}() must be called precisely once per IRQ exception, so that rcu_is_cpu_rrupt_from_idle() can correctly identify when an interrupt was taken from an idle context which must be explicitly preempted. Currently handle_domain_irq() calls rcu_irq_{enter,exit}() via irq_{enter,exit}(), but entry code needs to be able to call rcu_irq_{enter,exit}() earlier for correct ordering across lockdep/RCU/tracing updates for sequences such as: lockdep_hardirqs_off(CALLER_ADDR0); rcu_irq_enter(); trace_hardirqs_off_finish(); To permit each architecture to be converted to the new style in turn, this patch adds a new CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY selected by all current users of HANDLE_DOMAIN_IRQ, which gates the existing behaviour. When CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY is not selected, handle_domain_irq() requires entry code to perform the irq_{enter,exit}() work, with an explicit check for this matching the style of handle_domain_nmi(). Subsequent patches will: 1) Add the necessary IRQ entry accounting to each architecture in turn, dropping CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY from that architecture's Kconfig. 2) Remove CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY once it is no longer selected. 3) Convert irqchip drivers to consistently use generic_handle_domain_irq() rather than handle_domain_irq(). 4) Remove handle_domain_irq() and CONFIG_HANDLE_DOMAIN_IRQ. ... which should leave us with a clear split of responsiblity across the entry and irqchip code, making it possible to perform additional cleanups and fixes for the aforementioned longstanding issues with entry code. There should be no functional change as a result of this patch. Signed-off-by: Mark Rutland <mark.rutland@arm.com> Reviewed-by: Marc Zyngier <maz@kernel.org> Cc: Thomas Gleixner <tglx@linutronix.de>
2021-10-19 10:12:31 +00:00
/**
* generic_handle_domain_nmi - Invoke the handler for a HW nmi belonging
* to a domain.
irq: add a (temporary) CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY Going forward we want architecture/entry code to perform all the necessary work to enter/exit IRQ context, with irqchip code merely handling the mapping of the interrupt to any handler(s). Among other reasons, this is necessary to consistently fix some longstanding issues with the ordering of lockdep/RCU/tracing instrumentation which many architectures get wrong today in their entry code. Importantly, rcu_irq_{enter,exit}() must be called precisely once per IRQ exception, so that rcu_is_cpu_rrupt_from_idle() can correctly identify when an interrupt was taken from an idle context which must be explicitly preempted. Currently handle_domain_irq() calls rcu_irq_{enter,exit}() via irq_{enter,exit}(), but entry code needs to be able to call rcu_irq_{enter,exit}() earlier for correct ordering across lockdep/RCU/tracing updates for sequences such as: lockdep_hardirqs_off(CALLER_ADDR0); rcu_irq_enter(); trace_hardirqs_off_finish(); To permit each architecture to be converted to the new style in turn, this patch adds a new CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY selected by all current users of HANDLE_DOMAIN_IRQ, which gates the existing behaviour. When CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY is not selected, handle_domain_irq() requires entry code to perform the irq_{enter,exit}() work, with an explicit check for this matching the style of handle_domain_nmi(). Subsequent patches will: 1) Add the necessary IRQ entry accounting to each architecture in turn, dropping CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY from that architecture's Kconfig. 2) Remove CONFIG_HANDLE_DOMAIN_IRQ_IRQENTRY once it is no longer selected. 3) Convert irqchip drivers to consistently use generic_handle_domain_irq() rather than handle_domain_irq(). 4) Remove handle_domain_irq() and CONFIG_HANDLE_DOMAIN_IRQ. ... which should leave us with a clear split of responsiblity across the entry and irqchip code, making it possible to perform additional cleanups and fixes for the aforementioned longstanding issues with entry code. There should be no functional change as a result of this patch. Signed-off-by: Mark Rutland <mark.rutland@arm.com> Reviewed-by: Marc Zyngier <maz@kernel.org> Cc: Thomas Gleixner <tglx@linutronix.de>
2021-10-19 10:12:31 +00:00
* @domain: The domain where to perform the lookup
* @hwirq: The HW irq number to convert to a logical one
*
* Returns: 0 on success, or -EINVAL if conversion has failed
*
* This function must be called from an NMI context with irq regs
* initialized.
**/
int generic_handle_domain_nmi(struct irq_domain *domain, unsigned int hwirq)
{
WARN_ON_ONCE(!in_nmi());
return handle_irq_desc(irq_resolve_mapping(domain, hwirq));
}
#endif
genirq: Add irq_domain-aware core IRQ handler Calling irq_find_mapping from outside a irq_{enter,exit} section is unsafe and produces ugly messages if CONFIG_PROVE_RCU is enabled: If coming from the idle state, the rcu_read_lock call in irq_find_mapping will generate an unpleasant warning: <quote> =============================== [ INFO: suspicious RCU usage. ] 3.16.0-rc1+ #135 Not tainted ------------------------------- include/linux/rcupdate.h:871 rcu_read_lock() used illegally while idle! other info that might help us debug this: RCU used illegally from idle CPU! rcu_scheduler_active = 1, debug_locks = 0 RCU used illegally from extended quiescent state! 1 lock held by swapper/0/0: #0: (rcu_read_lock){......}, at: [<ffffffc00010206c>] irq_find_mapping+0x4c/0x198 </quote> As this issue is fairly widespread and involves at least three different architectures, a possible solution is to add a new handle_domain_irq entry point into the generic IRQ code that the interrupt controller code can call. This new function takes an irq_domain, and calls into irq_find_domain inside the irq_{enter,exit} block. An additional "lookup" parameter is used to allow non-domain architecture code to be replaced by this as well. Interrupt controllers can then be updated to use the new mechanism. This code is sitting behind a new CONFIG_HANDLE_DOMAIN_IRQ, as not all architectures implement set_irq_regs (yes, mn10300, I'm looking at you...). Reported-by: Vladimir Murzin <vladimir.murzin@arm.com> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Link: https://lkml.kernel.org/r/1409047421-27649-2-git-send-email-marc.zyngier@arm.com Signed-off-by: Jason Cooper <jason@lakedaemon.net>
2014-08-26 10:03:16 +00:00
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
/* Dynamic interrupt handling */
/**
* irq_free_descs - free irq descriptors
* @from: Start of descriptor range
* @cnt: Number of consecutive irqs to free
*/
void irq_free_descs(unsigned int from, unsigned int cnt)
{
int i;
if (from >= nr_irqs || (from + cnt) > nr_irqs)
return;
mutex_lock(&sparse_irq_lock);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
for (i = 0; i < cnt; i++)
free_desc(from + i);
mutex_unlock(&sparse_irq_lock);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
}
EXPORT_SYMBOL_GPL(irq_free_descs);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
/**
* __irq_alloc_descs - allocate and initialize a range of irq descriptors
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
* @irq: Allocate for specific irq number if irq >= 0
* @from: Start the search from this irq number
* @cnt: Number of consecutive irqs to allocate.
* @node: Preferred node on which the irq descriptor should be allocated
* @owner: Owning module (can be NULL)
* @affinity: Optional pointer to an affinity mask array of size @cnt which
* hints where the irq descriptors should be allocated and which
* default affinities to use
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
*
* Returns the first irq number or error code
*/
int __ref
__irq_alloc_descs(int irq, unsigned int from, unsigned int cnt, int node,
genirq/core: Introduce struct irq_affinity_desc The interrupt affinity management uses straight cpumask pointers to convey the automatically assigned affinity masks for managed interrupts. The core interrupt descriptor allocation also decides based on the pointer being non NULL whether an interrupt is managed or not. Devices which use managed interrupts usually have two classes of interrupts: - Interrupts for multiple device queues - Interrupts for general device management Currently both classes are treated the same way, i.e. as managed interrupts. The general interrupts get the default affinity mask assigned while the device queue interrupts are spread out over the possible CPUs. Treating the general interrupts as managed is both a limitation and under certain circumstances a bug. Assume the following situation: default_irq_affinity = 4..7 So if CPUs 4-7 are offlined, then the core code will shut down the device management interrupts because the last CPU in their affinity mask went offline. It's also a limitation because it's desired to allow manual placement of the general device interrupts for various reasons. If they are marked managed then the interrupt affinity setting from both user and kernel space is disabled. To remedy that situation it's required to convey more information than the cpumasks through various interfaces related to interrupt descriptor allocation. Instead of adding yet another argument, create a new data structure 'irq_affinity_desc' which for now just contains the cpumask. This struct can be expanded to convey auxilliary information in the next step. No functional change, just preparatory work. [ tglx: Simplified logic and clarified changelog ] Suggested-by: Thomas Gleixner <tglx@linutronix.de> Suggested-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Dou Liyang <douliyangs@gmail.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: linux-pci@vger.kernel.org Cc: kashyap.desai@broadcom.com Cc: shivasharan.srikanteshwara@broadcom.com Cc: sumit.saxena@broadcom.com Cc: ming.lei@redhat.com Cc: hch@lst.de Cc: douliyang1@huawei.com Link: https://lkml.kernel.org/r/20181204155122.6327-2-douliyangs@gmail.com
2018-12-04 15:51:20 +00:00
struct module *owner, const struct irq_affinity_desc *affinity)
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
{
int start, ret;
if (!cnt)
return -EINVAL;
if (irq >= 0) {
if (from > irq)
return -EINVAL;
from = irq;
genirq: x86: Ensure that dynamic irq allocation does not conflict On x86 the allocation of irq descriptors may allocate interrupts which are in the range of the GSI interrupts. That's wrong as those interrupts are hardwired and we don't have the irq domain translation like PPC. So one of these interrupts can be hooked up later to one of the devices which are hard wired to it and the io_apic init code for that particular interrupt line happily reuses that descriptor with a completely different configuration so hell breaks lose. Inside x86 we allocate dynamic interrupts from above nr_gsi_irqs, except for a few usage sites which have not yet blown up in our face for whatever reason. But for drivers which need an irq range, like the GPIO drivers, we have no limit in place and we don't want to expose such a detail to a driver. To cure this introduce a function which an architecture can implement to impose a lower bound on the dynamic interrupt allocations. Implement it for x86 and set the lower bound to nr_gsi_irqs, which is the end of the hardwired interrupt space, so all dynamic allocations happen above. That not only allows the GPIO driver to work sanely, it also protects the bogus callsites of create_irq_nr() in hpet, uv, irq_remapping and htirq code. They need to be cleaned up as well, but that's a separate issue. Reported-by: Jin Yao <yao.jin@linux.intel.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Mika Westerberg <mika.westerberg@linux.intel.com> Cc: Mathias Nyman <mathias.nyman@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Grant Likely <grant.likely@linaro.org> Cc: H. Peter Anvin <hpa@linux.intel.com> Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Cc: Krogerus Heikki <heikki.krogerus@intel.com> Cc: Linus Walleij <linus.walleij@linaro.org> Link: http://lkml.kernel.org/r/alpine.DEB.2.02.1404241617360.28206@ionos.tec.linutronix.de Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2014-04-24 07:50:53 +00:00
} else {
/*
* For interrupts which are freely allocated the
* architecture can force a lower bound to the @from
* argument. x86 uses this to exclude the GSI space.
*/
from = arch_dynirq_lower_bound(from);
}
mutex_lock(&sparse_irq_lock);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
start = irq_find_free_area(from, cnt);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
ret = -EEXIST;
if (irq >=0 && start != irq)
goto unlock;
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
if (start + cnt > nr_irqs) {
ret = irq_expand_nr_irqs(start + cnt);
if (ret)
goto unlock;
}
ret = alloc_descs(start, cnt, node, affinity, owner);
unlock:
mutex_unlock(&sparse_irq_lock);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
return ret;
}
EXPORT_SYMBOL_GPL(__irq_alloc_descs);
genirq: Implement a sane sparse_irq allocator The current sparse_irq allocator has several short comings due to failures in the design or the lack of it: - Requires iteration over the number of active irqs to find a free slot (Some architectures have grown their own workarounds for this) - Removal of entries is not possible - Racy between create_irq_nr and destroy_irq (plugged by horrible callbacks) - Migration of active irq descriptors is not possible - No bulk allocation of irq ranges - Sprinkeled irq_desc references all over the place outside of kernel/irq/ (The previous chip functions series is addressing this issue) Implement a sane allocator which fixes the above short comings (though migration of active descriptors needs a full tree wide cleanup of the direct and mostly unlocked access to irq_desc). The new allocator still uses a radix_tree, but uses a bitmap for keeping track of allocated irq numbers. That allows: - Fast lookup of a free slot - Allows the removal of descriptors - Prevents the create/destroy race - Bulk allocation of consecutive irq ranges - Basic design is ready for migration of life descriptors after further cleanups The bitmap is also used in the SPARSE_IRQ=n case for lookup and raceless (de)allocation of irq numbers. So it removes the requirement for looping through the descriptor array to find slots. Right now it uses sparse_irq_lock to protect the bitmap and the radix tree, but after cleaning up all users we should be able convert that to a mutex and to switch the radix_tree and decriptor allocations to GFP_KERNEL. [ Folded in a bugfix from Yinghai Lu ] Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Ingo Molnar <mingo@elte.hu>
2010-09-27 15:48:26 +00:00
/**
* irq_get_next_irq - get next allocated irq number
* @offset: where to start the search
*
* Returns next irq number after offset or nr_irqs if none is found.
*/
unsigned int irq_get_next_irq(unsigned int offset)
{
return irq_find_at_or_after(offset);
}
struct irq_desc *
genirq: Add support for per-cpu dev_id interrupts The ARM GIC interrupt controller offers per CPU interrupts (PPIs), which are usually used to connect local timers to each core. Each CPU has its own private interface to the GIC, and only sees the PPIs that are directly connect to it. While these timers are separate devices and have a separate interrupt line to a core, they all use the same IRQ number. For these devices, request_irq() is not the right API as it assumes that an IRQ number is visible by a number of CPUs (through the affinity setting), but makes it very awkward to express that an IRQ number can be handled by all CPUs, and yet be a different interrupt line on each CPU, requiring a different dev_id cookie to be passed back to the handler. The *_percpu_irq() functions is designed to overcome these limitations, by providing a per-cpu dev_id vector: int request_percpu_irq(unsigned int irq, irq_handler_t handler, const char *devname, void __percpu *percpu_dev_id); void free_percpu_irq(unsigned int, void __percpu *); int setup_percpu_irq(unsigned int irq, struct irqaction *new); void remove_percpu_irq(unsigned int irq, struct irqaction *act); void enable_percpu_irq(unsigned int irq); void disable_percpu_irq(unsigned int irq); The API has a number of limitations: - no interrupt sharing - no threading - common handler across all the CPUs Once the interrupt is requested using setup_percpu_irq() or request_percpu_irq(), it must be enabled by each core that wishes its local interrupt to be delivered. Based on an initial patch by Thomas Gleixner. Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Cc: linux-arm-kernel@lists.infradead.org Link: http://lkml.kernel.org/r/1316793788-14500-2-git-send-email-marc.zyngier@arm.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2011-09-23 16:03:06 +00:00
__irq_get_desc_lock(unsigned int irq, unsigned long *flags, bool bus,
unsigned int check)
{
struct irq_desc *desc = irq_to_desc(irq);
if (desc) {
genirq: Add support for per-cpu dev_id interrupts The ARM GIC interrupt controller offers per CPU interrupts (PPIs), which are usually used to connect local timers to each core. Each CPU has its own private interface to the GIC, and only sees the PPIs that are directly connect to it. While these timers are separate devices and have a separate interrupt line to a core, they all use the same IRQ number. For these devices, request_irq() is not the right API as it assumes that an IRQ number is visible by a number of CPUs (through the affinity setting), but makes it very awkward to express that an IRQ number can be handled by all CPUs, and yet be a different interrupt line on each CPU, requiring a different dev_id cookie to be passed back to the handler. The *_percpu_irq() functions is designed to overcome these limitations, by providing a per-cpu dev_id vector: int request_percpu_irq(unsigned int irq, irq_handler_t handler, const char *devname, void __percpu *percpu_dev_id); void free_percpu_irq(unsigned int, void __percpu *); int setup_percpu_irq(unsigned int irq, struct irqaction *new); void remove_percpu_irq(unsigned int irq, struct irqaction *act); void enable_percpu_irq(unsigned int irq); void disable_percpu_irq(unsigned int irq); The API has a number of limitations: - no interrupt sharing - no threading - common handler across all the CPUs Once the interrupt is requested using setup_percpu_irq() or request_percpu_irq(), it must be enabled by each core that wishes its local interrupt to be delivered. Based on an initial patch by Thomas Gleixner. Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Cc: linux-arm-kernel@lists.infradead.org Link: http://lkml.kernel.org/r/1316793788-14500-2-git-send-email-marc.zyngier@arm.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2011-09-23 16:03:06 +00:00
if (check & _IRQ_DESC_CHECK) {
if ((check & _IRQ_DESC_PERCPU) &&
!irq_settings_is_per_cpu_devid(desc))
return NULL;
if (!(check & _IRQ_DESC_PERCPU) &&
irq_settings_is_per_cpu_devid(desc))
return NULL;
}
if (bus)
chip_bus_lock(desc);
raw_spin_lock_irqsave(&desc->lock, *flags);
}
return desc;
}
void __irq_put_desc_unlock(struct irq_desc *desc, unsigned long flags, bool bus)
__releases(&desc->lock)
{
raw_spin_unlock_irqrestore(&desc->lock, flags);
if (bus)
chip_bus_sync_unlock(desc);
}
int irq_set_percpu_devid_partition(unsigned int irq,
const struct cpumask *affinity)
genirq: Add support for per-cpu dev_id interrupts The ARM GIC interrupt controller offers per CPU interrupts (PPIs), which are usually used to connect local timers to each core. Each CPU has its own private interface to the GIC, and only sees the PPIs that are directly connect to it. While these timers are separate devices and have a separate interrupt line to a core, they all use the same IRQ number. For these devices, request_irq() is not the right API as it assumes that an IRQ number is visible by a number of CPUs (through the affinity setting), but makes it very awkward to express that an IRQ number can be handled by all CPUs, and yet be a different interrupt line on each CPU, requiring a different dev_id cookie to be passed back to the handler. The *_percpu_irq() functions is designed to overcome these limitations, by providing a per-cpu dev_id vector: int request_percpu_irq(unsigned int irq, irq_handler_t handler, const char *devname, void __percpu *percpu_dev_id); void free_percpu_irq(unsigned int, void __percpu *); int setup_percpu_irq(unsigned int irq, struct irqaction *new); void remove_percpu_irq(unsigned int irq, struct irqaction *act); void enable_percpu_irq(unsigned int irq); void disable_percpu_irq(unsigned int irq); The API has a number of limitations: - no interrupt sharing - no threading - common handler across all the CPUs Once the interrupt is requested using setup_percpu_irq() or request_percpu_irq(), it must be enabled by each core that wishes its local interrupt to be delivered. Based on an initial patch by Thomas Gleixner. Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Cc: linux-arm-kernel@lists.infradead.org Link: http://lkml.kernel.org/r/1316793788-14500-2-git-send-email-marc.zyngier@arm.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2011-09-23 16:03:06 +00:00
{
struct irq_desc *desc = irq_to_desc(irq);
if (!desc || desc->percpu_enabled)
genirq: Add support for per-cpu dev_id interrupts The ARM GIC interrupt controller offers per CPU interrupts (PPIs), which are usually used to connect local timers to each core. Each CPU has its own private interface to the GIC, and only sees the PPIs that are directly connect to it. While these timers are separate devices and have a separate interrupt line to a core, they all use the same IRQ number. For these devices, request_irq() is not the right API as it assumes that an IRQ number is visible by a number of CPUs (through the affinity setting), but makes it very awkward to express that an IRQ number can be handled by all CPUs, and yet be a different interrupt line on each CPU, requiring a different dev_id cookie to be passed back to the handler. The *_percpu_irq() functions is designed to overcome these limitations, by providing a per-cpu dev_id vector: int request_percpu_irq(unsigned int irq, irq_handler_t handler, const char *devname, void __percpu *percpu_dev_id); void free_percpu_irq(unsigned int, void __percpu *); int setup_percpu_irq(unsigned int irq, struct irqaction *new); void remove_percpu_irq(unsigned int irq, struct irqaction *act); void enable_percpu_irq(unsigned int irq); void disable_percpu_irq(unsigned int irq); The API has a number of limitations: - no interrupt sharing - no threading - common handler across all the CPUs Once the interrupt is requested using setup_percpu_irq() or request_percpu_irq(), it must be enabled by each core that wishes its local interrupt to be delivered. Based on an initial patch by Thomas Gleixner. Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Cc: linux-arm-kernel@lists.infradead.org Link: http://lkml.kernel.org/r/1316793788-14500-2-git-send-email-marc.zyngier@arm.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2011-09-23 16:03:06 +00:00
return -EINVAL;
desc->percpu_enabled = kzalloc(sizeof(*desc->percpu_enabled), GFP_KERNEL);
if (!desc->percpu_enabled)
return -ENOMEM;
desc->percpu_affinity = affinity ? : cpu_possible_mask;
genirq: Add support for per-cpu dev_id interrupts The ARM GIC interrupt controller offers per CPU interrupts (PPIs), which are usually used to connect local timers to each core. Each CPU has its own private interface to the GIC, and only sees the PPIs that are directly connect to it. While these timers are separate devices and have a separate interrupt line to a core, they all use the same IRQ number. For these devices, request_irq() is not the right API as it assumes that an IRQ number is visible by a number of CPUs (through the affinity setting), but makes it very awkward to express that an IRQ number can be handled by all CPUs, and yet be a different interrupt line on each CPU, requiring a different dev_id cookie to be passed back to the handler. The *_percpu_irq() functions is designed to overcome these limitations, by providing a per-cpu dev_id vector: int request_percpu_irq(unsigned int irq, irq_handler_t handler, const char *devname, void __percpu *percpu_dev_id); void free_percpu_irq(unsigned int, void __percpu *); int setup_percpu_irq(unsigned int irq, struct irqaction *new); void remove_percpu_irq(unsigned int irq, struct irqaction *act); void enable_percpu_irq(unsigned int irq); void disable_percpu_irq(unsigned int irq); The API has a number of limitations: - no interrupt sharing - no threading - common handler across all the CPUs Once the interrupt is requested using setup_percpu_irq() or request_percpu_irq(), it must be enabled by each core that wishes its local interrupt to be delivered. Based on an initial patch by Thomas Gleixner. Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Cc: linux-arm-kernel@lists.infradead.org Link: http://lkml.kernel.org/r/1316793788-14500-2-git-send-email-marc.zyngier@arm.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2011-09-23 16:03:06 +00:00
irq_set_percpu_devid_flags(irq);
return 0;
}
int irq_set_percpu_devid(unsigned int irq)
{
return irq_set_percpu_devid_partition(irq, NULL);
}
int irq_get_percpu_devid_partition(unsigned int irq, struct cpumask *affinity)
{
struct irq_desc *desc = irq_to_desc(irq);
if (!desc || !desc->percpu_enabled)
return -EINVAL;
if (affinity)
cpumask_copy(affinity, desc->percpu_affinity);
return 0;
}
EXPORT_SYMBOL_GPL(irq_get_percpu_devid_partition);
void kstat_incr_irq_this_cpu(unsigned int irq)
{
kstat_incr_irqs_this_cpu(irq_to_desc(irq));
}
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
/**
* kstat_irqs_cpu - Get the statistics for an interrupt on a cpu
* @irq: The interrupt number
* @cpu: The cpu number
*
* Returns the sum of interrupt counts on @cpu since boot for
* @irq. The caller must ensure that the interrupt is not removed
* concurrently.
*/
unsigned int kstat_irqs_cpu(unsigned int irq, int cpu)
{
struct irq_desc *desc = irq_to_desc(irq);
return desc && desc->kstat_irqs ? per_cpu(desc->kstat_irqs->cnt, cpu) : 0;
}
unsigned int kstat_irqs_desc(struct irq_desc *desc, const struct cpumask *cpumask)
{
unsigned int sum = 0;
genirq: Avoid summation loops for /proc/stat Waiman reported that on large systems with a large amount of interrupts the readout of /proc/stat takes a long time to sum up the interrupt statistics. In principle this is not a problem. but for unknown reasons some enterprise quality software reads /proc/stat with a high frequency. The reason for this is that interrupt statistics are accounted per cpu. So the /proc/stat logic has to sum up the interrupt stats for each interrupt. This can be largely avoided for interrupts which are not marked as 'PER_CPU' interrupts by simply adding a per interrupt summation counter which is incremented along with the per interrupt per cpu counter. The PER_CPU interrupts need to avoid that and use only per cpu accounting because they share the interrupt number and the interrupt descriptor and concurrent updates would conflict or require unwanted synchronization. Reported-by: Waiman Long <longman@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Waiman Long <longman@redhat.com> Reviewed-by: Marc Zyngier <marc.zyngier@arm.com> Reviewed-by: Davidlohr Bueso <dbueso@suse.de> Cc: Matthew Wilcox <willy@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Kees Cook <keescook@chromium.org> Cc: linux-fsdevel@vger.kernel.org Cc: Davidlohr Bueso <dave@stgolabs.net> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Daniel Colascione <dancol@google.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Randy Dunlap <rdunlap@infradead.org> Link: https://lkml.kernel.org/r/20190208135020.925487496@linutronix.de 8<------------- v2: Undo the unintentional layout change of struct irq_desc. include/linux/irqdesc.h | 1 + kernel/irq/chip.c | 12 ++++++++++-- kernel/irq/internals.h | 8 +++++++- kernel/irq/irqdesc.c | 7 ++++++- 4 files changed, 24 insertions(+), 4 deletions(-)
2019-02-08 13:48:03 +00:00
int cpu;
genirq: Avoid summation loops for /proc/stat Waiman reported that on large systems with a large amount of interrupts the readout of /proc/stat takes a long time to sum up the interrupt statistics. In principle this is not a problem. but for unknown reasons some enterprise quality software reads /proc/stat with a high frequency. The reason for this is that interrupt statistics are accounted per cpu. So the /proc/stat logic has to sum up the interrupt stats for each interrupt. This can be largely avoided for interrupts which are not marked as 'PER_CPU' interrupts by simply adding a per interrupt summation counter which is incremented along with the per interrupt per cpu counter. The PER_CPU interrupts need to avoid that and use only per cpu accounting because they share the interrupt number and the interrupt descriptor and concurrent updates would conflict or require unwanted synchronization. Reported-by: Waiman Long <longman@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Waiman Long <longman@redhat.com> Reviewed-by: Marc Zyngier <marc.zyngier@arm.com> Reviewed-by: Davidlohr Bueso <dbueso@suse.de> Cc: Matthew Wilcox <willy@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Kees Cook <keescook@chromium.org> Cc: linux-fsdevel@vger.kernel.org Cc: Davidlohr Bueso <dave@stgolabs.net> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Daniel Colascione <dancol@google.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Randy Dunlap <rdunlap@infradead.org> Link: https://lkml.kernel.org/r/20190208135020.925487496@linutronix.de 8<------------- v2: Undo the unintentional layout change of struct irq_desc. include/linux/irqdesc.h | 1 + kernel/irq/chip.c | 12 ++++++++++-- kernel/irq/internals.h | 8 +++++++- kernel/irq/irqdesc.c | 7 ++++++- 4 files changed, 24 insertions(+), 4 deletions(-)
2019-02-08 13:48:03 +00:00
if (!irq_settings_is_per_cpu_devid(desc) &&
!irq_settings_is_per_cpu(desc) &&
!irq_is_nmi(desc))
return data_race(desc->tot_count);
genirq: Avoid summation loops for /proc/stat Waiman reported that on large systems with a large amount of interrupts the readout of /proc/stat takes a long time to sum up the interrupt statistics. In principle this is not a problem. but for unknown reasons some enterprise quality software reads /proc/stat with a high frequency. The reason for this is that interrupt statistics are accounted per cpu. So the /proc/stat logic has to sum up the interrupt stats for each interrupt. This can be largely avoided for interrupts which are not marked as 'PER_CPU' interrupts by simply adding a per interrupt summation counter which is incremented along with the per interrupt per cpu counter. The PER_CPU interrupts need to avoid that and use only per cpu accounting because they share the interrupt number and the interrupt descriptor and concurrent updates would conflict or require unwanted synchronization. Reported-by: Waiman Long <longman@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Waiman Long <longman@redhat.com> Reviewed-by: Marc Zyngier <marc.zyngier@arm.com> Reviewed-by: Davidlohr Bueso <dbueso@suse.de> Cc: Matthew Wilcox <willy@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Kees Cook <keescook@chromium.org> Cc: linux-fsdevel@vger.kernel.org Cc: Davidlohr Bueso <dave@stgolabs.net> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Daniel Colascione <dancol@google.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Randy Dunlap <rdunlap@infradead.org> Link: https://lkml.kernel.org/r/20190208135020.925487496@linutronix.de 8<------------- v2: Undo the unintentional layout change of struct irq_desc. include/linux/irqdesc.h | 1 + kernel/irq/chip.c | 12 ++++++++++-- kernel/irq/internals.h | 8 +++++++- kernel/irq/irqdesc.c | 7 ++++++- 4 files changed, 24 insertions(+), 4 deletions(-)
2019-02-08 13:48:03 +00:00
for_each_cpu(cpu, cpumask)
sum += data_race(per_cpu(desc->kstat_irqs->cnt, cpu));
return sum;
}
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
static unsigned int kstat_irqs(unsigned int irq)
{
struct irq_desc *desc = irq_to_desc(irq);
if (!desc || !desc->kstat_irqs)
return 0;
return kstat_irqs_desc(desc, cpu_possible_mask);
}
#ifdef CONFIG_GENERIC_IRQ_STAT_SNAPSHOT
void kstat_snapshot_irqs(void)
{
struct irq_desc *desc;
unsigned int irq;
for_each_irq_desc(irq, desc) {
if (!desc->kstat_irqs)
continue;
this_cpu_write(desc->kstat_irqs->ref, this_cpu_read(desc->kstat_irqs->cnt));
}
}
unsigned int kstat_get_irq_since_snapshot(unsigned int irq)
{
struct irq_desc *desc = irq_to_desc(irq);
if (!desc || !desc->kstat_irqs)
return 0;
return this_cpu_read(desc->kstat_irqs->cnt) - this_cpu_read(desc->kstat_irqs->ref);
}
#endif
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
/**
* kstat_irqs_usr - Get the statistics for an interrupt from thread context
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
* @irq: The interrupt number
*
* Returns the sum of interrupt counts on all cpus since boot for @irq.
*
* It uses rcu to protect the access since a concurrent removal of an
* interrupt descriptor is observing an rcu grace period before
* delayed_free_desc()/irq_kobj_release().
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
*/
unsigned int kstat_irqs_usr(unsigned int irq)
{
unsigned int sum;
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
rcu_read_lock();
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
sum = kstat_irqs(irq);
rcu_read_unlock();
genirq: Prevent proc race against freeing of irq descriptors Since the rework of the sparse interrupt code to actually free the unused interrupt descriptors there exists a race between the /proc interfaces to the irq subsystem and the code which frees the interrupt descriptor. CPU0 CPU1 show_interrupts() desc = irq_to_desc(X); free_desc(desc) remove_from_radix_tree(); kfree(desc); raw_spinlock_irq(&desc->lock); /proc/interrupts is the only interface which can actively corrupt kernel memory via the lock access. /proc/stat can only read from freed memory. Extremly hard to trigger, but possible. The interfaces in /proc/irq/N/ are not affected by this because the removal of the proc file is serialized in procfs against concurrent readers/writers. The removal happens before the descriptor is freed. For architectures which have CONFIG_SPARSE_IRQ=n this is a non issue as the descriptor is never freed. It's merely cleared out with the irq descriptor lock held. So any concurrent proc access will either see the old correct value or the cleared out ones. Protect the lookup and access to the irq descriptor in show_interrupts() with the sparse_irq_lock. Provide kstat_irqs_usr() which is protecting the lookup and access with sparse_irq_lock and switch /proc/stat to use it. Document the existing kstat_irqs interfaces so it's clear that the caller needs to take care about protection. The users of these interfaces are either not affected due to SPARSE_IRQ=n or already protected against removal. Fixes: 1f5a5b87f78f "genirq: Implement a sane sparse_irq allocator" Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org
2014-12-11 22:01:41 +00:00
return sum;
}
#ifdef CONFIG_LOCKDEP
void __irq_set_lockdep_class(unsigned int irq, struct lock_class_key *lock_class,
struct lock_class_key *request_class)
{
struct irq_desc *desc = irq_to_desc(irq);
if (desc) {
lockdep_set_class(&desc->lock, lock_class);
lockdep_set_class(&desc->request_mutex, request_class);
}
}
EXPORT_SYMBOL_GPL(__irq_set_lockdep_class);
#endif