sparse irq_desc[] array: core kernel and x86 changes
Impact: new feature
Problem on distro kernels: irq_desc[NR_IRQS] takes megabytes of RAM with
NR_CPUS set to large values. The goal is to be able to scale up to much
larger NR_IRQS value without impacting the (important) common case.
To solve this, we generalize irq_desc[NR_IRQS] to an (optional) array of
irq_desc pointers.
When CONFIG_SPARSE_IRQ=y is used, we use kzalloc_node to get irq_desc,
this also makes the IRQ descriptors NUMA-local (to the site that calls
request_irq()).
This gets rid of the irq_cfg[] static array on x86 as well: irq_cfg now
uses desc->chip_data for x86 to store irq_cfg.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c
index 9043251..9de17f5 100644
--- a/arch/x86/kernel/io_apic.c
+++ b/arch/x86/kernel/io_apic.c
@@ -108,55 +108,6 @@
early_param("noapic", parse_noapic);
struct irq_pin_list;
-struct irq_cfg {
- unsigned int irq;
- struct irq_pin_list *irq_2_pin;
- cpumask_t domain;
- cpumask_t old_domain;
- unsigned move_cleanup_count;
- u8 vector;
- u8 move_in_progress : 1;
-};
-
-/* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */
-static struct irq_cfg irq_cfgx[NR_IRQS] = {
- [0] = { .irq = 0, .domain = CPU_MASK_ALL, .vector = IRQ0_VECTOR, },
- [1] = { .irq = 1, .domain = CPU_MASK_ALL, .vector = IRQ1_VECTOR, },
- [2] = { .irq = 2, .domain = CPU_MASK_ALL, .vector = IRQ2_VECTOR, },
- [3] = { .irq = 3, .domain = CPU_MASK_ALL, .vector = IRQ3_VECTOR, },
- [4] = { .irq = 4, .domain = CPU_MASK_ALL, .vector = IRQ4_VECTOR, },
- [5] = { .irq = 5, .domain = CPU_MASK_ALL, .vector = IRQ5_VECTOR, },
- [6] = { .irq = 6, .domain = CPU_MASK_ALL, .vector = IRQ6_VECTOR, },
- [7] = { .irq = 7, .domain = CPU_MASK_ALL, .vector = IRQ7_VECTOR, },
- [8] = { .irq = 8, .domain = CPU_MASK_ALL, .vector = IRQ8_VECTOR, },
- [9] = { .irq = 9, .domain = CPU_MASK_ALL, .vector = IRQ9_VECTOR, },
- [10] = { .irq = 10, .domain = CPU_MASK_ALL, .vector = IRQ10_VECTOR, },
- [11] = { .irq = 11, .domain = CPU_MASK_ALL, .vector = IRQ11_VECTOR, },
- [12] = { .irq = 12, .domain = CPU_MASK_ALL, .vector = IRQ12_VECTOR, },
- [13] = { .irq = 13, .domain = CPU_MASK_ALL, .vector = IRQ13_VECTOR, },
- [14] = { .irq = 14, .domain = CPU_MASK_ALL, .vector = IRQ14_VECTOR, },
- [15] = { .irq = 15, .domain = CPU_MASK_ALL, .vector = IRQ15_VECTOR, },
-};
-
-#define for_each_irq_cfg(irq, cfg) \
- for (irq = 0, cfg = irq_cfgx; irq < nr_irqs; irq++, cfg++)
-
-static struct irq_cfg *irq_cfg(unsigned int irq)
-{
- return irq < nr_irqs ? irq_cfgx + irq : NULL;
-}
-
-static struct irq_cfg *irq_cfg_alloc(unsigned int irq)
-{
- return irq_cfg(irq);
-}
-
-/*
- * Rough estimation of how many shared IRQs there are, can be changed
- * anytime.
- */
-#define MAX_PLUS_SHARED_IRQS NR_IRQS
-#define PIN_MAP_SIZE (MAX_PLUS_SHARED_IRQS + NR_IRQS)
/*
* This is performance-critical, we want to do it O(1)
@@ -170,32 +121,116 @@
struct irq_pin_list *next;
};
-static struct irq_pin_list irq_2_pin_head[PIN_MAP_SIZE];
-static struct irq_pin_list *irq_2_pin_ptr;
-
-static void __init irq_2_pin_init(void)
+static struct irq_pin_list *get_one_free_irq_2_pin(int cpu)
{
- struct irq_pin_list *pin = irq_2_pin_head;
- int i;
+ struct irq_pin_list *pin;
+ int node;
- for (i = 1; i < PIN_MAP_SIZE; i++)
- pin[i-1].next = &pin[i];
+ node = cpu_to_node(cpu);
- irq_2_pin_ptr = &pin[0];
-}
+ pin = kzalloc_node(sizeof(*pin), GFP_ATOMIC, node);
+ printk(KERN_DEBUG " alloc irq_2_pin on cpu %d node %d\n", cpu, node);
-static struct irq_pin_list *get_one_free_irq_2_pin(void)
-{
- struct irq_pin_list *pin = irq_2_pin_ptr;
-
- if (!pin)
- panic("can not get more irq_2_pin\n");
-
- irq_2_pin_ptr = pin->next;
- pin->next = NULL;
return pin;
}
+struct irq_cfg {
+ struct irq_pin_list *irq_2_pin;
+ cpumask_t domain;
+ cpumask_t old_domain;
+ unsigned move_cleanup_count;
+ u8 vector;
+ u8 move_in_progress : 1;
+};
+
+/* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */
+#ifdef CONFIG_SPARSE_IRQ
+static struct irq_cfg irq_cfgx[] = {
+#else
+static struct irq_cfg irq_cfgx[NR_IRQS] = {
+#endif
+ [0] = { .domain = CPU_MASK_ALL, .vector = IRQ0_VECTOR, },
+ [1] = { .domain = CPU_MASK_ALL, .vector = IRQ1_VECTOR, },
+ [2] = { .domain = CPU_MASK_ALL, .vector = IRQ2_VECTOR, },
+ [3] = { .domain = CPU_MASK_ALL, .vector = IRQ3_VECTOR, },
+ [4] = { .domain = CPU_MASK_ALL, .vector = IRQ4_VECTOR, },
+ [5] = { .domain = CPU_MASK_ALL, .vector = IRQ5_VECTOR, },
+ [6] = { .domain = CPU_MASK_ALL, .vector = IRQ6_VECTOR, },
+ [7] = { .domain = CPU_MASK_ALL, .vector = IRQ7_VECTOR, },
+ [8] = { .domain = CPU_MASK_ALL, .vector = IRQ8_VECTOR, },
+ [9] = { .domain = CPU_MASK_ALL, .vector = IRQ9_VECTOR, },
+ [10] = { .domain = CPU_MASK_ALL, .vector = IRQ10_VECTOR, },
+ [11] = { .domain = CPU_MASK_ALL, .vector = IRQ11_VECTOR, },
+ [12] = { .domain = CPU_MASK_ALL, .vector = IRQ12_VECTOR, },
+ [13] = { .domain = CPU_MASK_ALL, .vector = IRQ13_VECTOR, },
+ [14] = { .domain = CPU_MASK_ALL, .vector = IRQ14_VECTOR, },
+ [15] = { .domain = CPU_MASK_ALL, .vector = IRQ15_VECTOR, },
+};
+
+void __init arch_early_irq_init(void)
+{
+ struct irq_cfg *cfg;
+ struct irq_desc *desc;
+ int count;
+ int i;
+
+ cfg = irq_cfgx;
+ count = ARRAY_SIZE(irq_cfgx);
+
+ for (i = 0; i < count; i++) {
+ desc = irq_to_desc(i);
+ desc->chip_data = &cfg[i];
+ }
+}
+
+#ifdef CONFIG_SPARSE_IRQ
+static struct irq_cfg *irq_cfg(unsigned int irq)
+{
+ struct irq_cfg *cfg = NULL;
+ struct irq_desc *desc;
+
+ desc = irq_to_desc(irq);
+ if (desc)
+ cfg = desc->chip_data;
+
+ return cfg;
+}
+
+static struct irq_cfg *get_one_free_irq_cfg(int cpu)
+{
+ struct irq_cfg *cfg;
+ int node;
+
+ node = cpu_to_node(cpu);
+
+ cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node);
+ printk(KERN_DEBUG " alloc irq_cfg on cpu %d node %d\n", cpu, node);
+
+ return cfg;
+}
+
+void arch_init_chip_data(struct irq_desc *desc, int cpu)
+{
+ struct irq_cfg *cfg;
+
+ cfg = desc->chip_data;
+ if (!cfg) {
+ desc->chip_data = get_one_free_irq_cfg(cpu);
+ if (!desc->chip_data) {
+ printk(KERN_ERR "can not alloc irq_cfg\n");
+ BUG_ON(1);
+ }
+ }
+}
+
+#else
+static struct irq_cfg *irq_cfg(unsigned int irq)
+{
+ return irq < nr_irqs ? irq_cfgx + irq : NULL;
+}
+
+#endif
+
struct io_apic {
unsigned int index;
unsigned int unused[3];
@@ -397,16 +432,19 @@
* shared ISA-space IRQs, so we have to support them. We are super
* fast in the common case, and fast for shared ISA-space IRQs.
*/
-static void add_pin_to_irq(unsigned int irq, int apic, int pin)
+static void add_pin_to_irq_cpu(unsigned int irq, int cpu, int apic, int pin)
{
- struct irq_cfg *cfg;
struct irq_pin_list *entry;
+ struct irq_cfg *cfg = irq_cfg(irq);
- /* first time to refer irq_cfg, so with new */
- cfg = irq_cfg_alloc(irq);
entry = cfg->irq_2_pin;
if (!entry) {
- entry = get_one_free_irq_2_pin();
+ entry = get_one_free_irq_2_pin(cpu);
+ if (!entry) {
+ printk(KERN_ERR "can not alloc irq_2_pin to add %d - %d\n",
+ apic, pin);
+ return;
+ }
cfg->irq_2_pin = entry;
entry->apic = apic;
entry->pin = pin;
@@ -421,7 +459,7 @@
entry = entry->next;
}
- entry->next = get_one_free_irq_2_pin();
+ entry->next = get_one_free_irq_2_pin(cpu);
entry = entry->next;
entry->apic = apic;
entry->pin = pin;
@@ -430,7 +468,7 @@
/*
* Reroute an IRQ to a different pin.
*/
-static void __init replace_pin_at_irq(unsigned int irq,
+static void __init replace_pin_at_irq(unsigned int irq, int cpu,
int oldapic, int oldpin,
int newapic, int newpin)
{
@@ -451,7 +489,7 @@
/* why? call replace before add? */
if (!replaced)
- add_pin_to_irq(irq, newapic, newpin);
+ add_pin_to_irq_cpu(irq, cpu, newapic, newpin);
}
static inline void io_apic_modify_irq(unsigned int irq,
@@ -1162,9 +1200,13 @@
/* This function must be called with vector_lock held */
int irq, vector;
struct irq_cfg *cfg;
+ struct irq_desc *desc;
/* Mark the inuse vectors */
- for_each_irq_cfg(irq, cfg) {
+ for_each_irq_desc(irq, desc) {
+ if (!desc)
+ continue;
+ cfg = desc->chip_data;
if (!cpu_isset(cpu, cfg->domain))
continue;
vector = cfg->vector;
@@ -1356,6 +1398,8 @@
{
int apic, pin, idx, irq;
int notcon = 0;
+ struct irq_desc *desc;
+ int cpu = boot_cpu_id;
apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n");
@@ -1387,7 +1431,12 @@
if (multi_timer_check(apic, irq))
continue;
#endif
- add_pin_to_irq(irq, apic, pin);
+ desc = irq_to_desc_alloc_cpu(irq, cpu);
+ if (!desc) {
+ printk(KERN_INFO "can not get irq_desc for %d\n", irq);
+ continue;
+ }
+ add_pin_to_irq_cpu(irq, cpu, apic, pin);
setup_IO_APIC_irq(apic, pin, irq,
irq_trigger(idx), irq_polarity(idx));
@@ -1448,6 +1497,7 @@
union IO_APIC_reg_03 reg_03;
unsigned long flags;
struct irq_cfg *cfg;
+ struct irq_desc *desc;
unsigned int irq;
if (apic_verbosity == APIC_QUIET)
@@ -1537,8 +1587,13 @@
}
}
printk(KERN_DEBUG "IRQ to pin mappings:\n");
- for_each_irq_cfg(irq, cfg) {
- struct irq_pin_list *entry = cfg->irq_2_pin;
+ for_each_irq_desc(irq, desc) {
+ struct irq_pin_list *entry;
+
+ if (!desc)
+ continue;
+ cfg = desc->chip_data;
+ entry = cfg->irq_2_pin;
if (!entry)
continue;
printk(KERN_DEBUG "IRQ%d ", irq);
@@ -2022,6 +2077,7 @@
{
int was_pending = 0;
unsigned long flags;
+ struct irq_cfg *cfg;
spin_lock_irqsave(&ioapic_lock, flags);
if (irq < 16) {
@@ -2029,6 +2085,7 @@
if (i8259A_irq_pending(irq))
was_pending = 1;
}
+ cfg = irq_cfg(irq);
__unmask_IO_APIC_irq(irq);
spin_unlock_irqrestore(&ioapic_lock, flags);
@@ -2178,6 +2235,9 @@
struct irq_desc *desc;
for_each_irq_desc(irq, desc) {
+ if (!desc)
+ continue;
+
if (desc->status & IRQ_MOVE_PENDING) {
unsigned long flags;
@@ -2229,6 +2289,9 @@
struct irq_cfg *cfg;
irq = __get_cpu_var(vector_irq)[vector];
+ if (irq == -1)
+ continue;
+
desc = irq_to_desc(irq);
if (!desc)
continue;
@@ -2430,8 +2493,12 @@
* Also, we've got to be careful not to trash gate
* 0x80, because int 0x80 is hm, kind of importantish. ;)
*/
- for_each_irq_cfg(irq, cfg) {
- if (IO_APIC_IRQ(irq) && !cfg->vector) {
+ for_each_irq_desc(irq, desc) {
+ if (!desc)
+ continue;
+
+ cfg = desc->chip_data;
+ if (IO_APIC_IRQ(irq) && cfg && !cfg->vector) {
/*
* Hmm.. We don't have an entry for this,
* so default to an old-fashioned 8259
@@ -2439,11 +2506,9 @@
*/
if (irq < 16)
make_8259A_irq(irq);
- else {
- desc = irq_to_desc(irq);
+ else
/* Strange. Oh, well.. */
desc->chip = &no_irq_chip;
- }
}
}
}
@@ -2654,7 +2719,7 @@
* Ok, does IRQ0 through the IOAPIC work?
*/
if (no_pin1) {
- add_pin_to_irq(0, apic1, pin1);
+ add_pin_to_irq_cpu(0, boot_cpu_id, apic1, pin1);
setup_timer_IRQ0_pin(apic1, pin1, cfg->vector);
}
unmask_IO_APIC_irq(0);
@@ -2683,7 +2748,7 @@
/*
* legacy devices should be connected to IO APIC #0
*/
- replace_pin_at_irq(0, apic1, pin1, apic2, pin2);
+ replace_pin_at_irq(0, boot_cpu_id, apic1, pin1, apic2, pin2);
setup_timer_IRQ0_pin(apic2, pin2, cfg->vector);
unmask_IO_APIC_irq(0);
enable_8259A_irq(0);
@@ -2902,21 +2967,25 @@
unsigned int irq;
unsigned int new;
unsigned long flags;
- struct irq_cfg *cfg_new;
-
- irq_want = nr_irqs - 1;
+ struct irq_cfg *cfg_new = NULL;
+ int cpu = boot_cpu_id;
+ struct irq_desc *desc_new = NULL;
irq = 0;
spin_lock_irqsave(&vector_lock, flags);
for (new = irq_want; new > 0; new--) {
if (platform_legacy_irq(new))
continue;
- cfg_new = irq_cfg(new);
- if (cfg_new && cfg_new->vector != 0)
+
+ desc_new = irq_to_desc_alloc_cpu(new, cpu);
+ if (!desc_new) {
+ printk(KERN_INFO "can not get irq_desc for %d\n", new);
continue;
- /* check if need to create one */
- if (!cfg_new)
- cfg_new = irq_cfg_alloc(new);
+ }
+ cfg_new = desc_new->chip_data;
+
+ if (cfg_new->vector != 0)
+ continue;
if (__assign_irq_vector(new, TARGET_CPUS) == 0)
irq = new;
break;
@@ -2925,6 +2994,9 @@
if (irq > 0) {
dynamic_irq_init(irq);
+ /* restore it, in case dynamic_irq_init clear it */
+ if (desc_new)
+ desc_new->chip_data = cfg_new;
}
return irq;
}
@@ -2944,8 +3016,16 @@
void destroy_irq(unsigned int irq)
{
unsigned long flags;
+ struct irq_cfg *cfg;
+ struct irq_desc *desc;
+ /* store it, in case dynamic_irq_cleanup clear it */
+ desc = irq_to_desc(irq);
+ cfg = desc->chip_data;
dynamic_irq_cleanup(irq);
+ /* connect back irq_cfg */
+ if (desc)
+ desc->chip_data = cfg;
#ifdef CONFIG_INTR_REMAP
free_irte(irq);
@@ -3195,26 +3275,13 @@
return 0;
}
-static unsigned int build_irq_for_pci_dev(struct pci_dev *dev)
-{
- unsigned int irq;
-
- irq = dev->bus->number;
- irq <<= 8;
- irq |= dev->devfn;
- irq <<= 12;
-
- return irq;
-}
-
-int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc)
+int arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *msidesc)
{
unsigned int irq;
int ret;
unsigned int irq_want;
- irq_want = build_irq_for_pci_dev(dev) + 0x100;
-
+ irq_want = nr_irqs - 1;
irq = create_irq_nr(irq_want);
if (irq == 0)
return -1;
@@ -3228,7 +3295,7 @@
goto error;
no_ir:
#endif
- ret = setup_msi_irq(dev, desc, irq);
+ ret = setup_msi_irq(dev, msidesc, irq);
if (ret < 0) {
destroy_irq(irq);
return ret;
@@ -3246,7 +3313,7 @@
{
unsigned int irq;
int ret, sub_handle;
- struct msi_desc *desc;
+ struct msi_desc *msidesc;
unsigned int irq_want;
#ifdef CONFIG_INTR_REMAP
@@ -3254,10 +3321,11 @@
int index = 0;
#endif
- irq_want = build_irq_for_pci_dev(dev) + 0x100;
+ irq_want = nr_irqs - 1;
sub_handle = 0;
- list_for_each_entry(desc, &dev->msi_list, list) {
- irq = create_irq_nr(irq_want--);
+ list_for_each_entry(msidesc, &dev->msi_list, list) {
+ irq = create_irq_nr(irq_want);
+ irq_want--;
if (irq == 0)
return -1;
#ifdef CONFIG_INTR_REMAP
@@ -3289,7 +3357,7 @@
}
no_ir:
#endif
- ret = setup_msi_irq(dev, desc, irq);
+ ret = setup_msi_irq(dev, msidesc, irq);
if (ret < 0)
goto error;
sub_handle++;
@@ -3707,17 +3775,29 @@
int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int polarity)
{
+ struct irq_desc *desc;
+ struct irq_cfg *cfg;
+ int cpu = boot_cpu_id;
+
if (!IO_APIC_IRQ(irq)) {
apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n",
ioapic);
return -EINVAL;
}
+ desc = irq_to_desc_alloc_cpu(irq, cpu);
+ if (!desc) {
+ printk(KERN_INFO "can not get irq_desc %d\n", irq);
+ return 0;
+ }
+
/*
* IRQs < 16 are already in the irq_2_pin[] map
*/
- if (irq >= 16)
- add_pin_to_irq(irq, ioapic, pin);
+ if (irq >= 16) {
+ cfg = desc->chip_data;
+ add_pin_to_irq_cpu(irq, cpu, ioapic, pin);
+ }
setup_IO_APIC_irq(ioapic, pin, irq, triggering, polarity);
@@ -3773,7 +3853,8 @@
* when you have too many devices, because at that time only boot
* cpu is online.
*/
- cfg = irq_cfg(irq);
+ desc = irq_to_desc(irq);
+ cfg = desc->chip_data;
if (!cfg->vector) {
setup_IO_APIC_irq(ioapic, pin, irq,
irq_trigger(irq_entry),
@@ -3785,7 +3866,6 @@
/*
* Honour affinities which have been set in early boot
*/
- desc = irq_to_desc(irq);
if (desc->status &
(IRQ_NO_BALANCING | IRQ_AFFINITY_SET))
mask = desc->affinity;
@@ -3846,7 +3926,6 @@
struct resource *ioapic_res;
int i;
- irq_2_pin_init();
ioapic_res = ioapic_setup_resources();
for (i = 0; i < nr_ioapics; i++) {
if (smp_found_config) {