提交 e273d140 编写于 作者: E Eric W. Biederman 提交者: Linus Torvalds

[PATCH] x86_64 irq: Use NR_IRQS not NR_IRQ_VECTORS

NR_IRQ_VECTORS is currently a compatiblity define set to NR_IRQs.
This patch updates the users of NR_IRQ_VECTORS to use NR_IRQs instead
so that NR_IRQ_VECTORS can be removed.

There is still shared code with arch/i386 that uses NR_IRQ_VECTORS
so we can't remove the #define just yet :(
Signed-off-by: NEric W. Biederman <ebiederm@xmission.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
上级 f45bcd70
...@@ -74,7 +74,7 @@ int nr_ioapic_registers[MAX_IO_APICS]; ...@@ -74,7 +74,7 @@ int nr_ioapic_registers[MAX_IO_APICS];
* Rough estimation of how many shared IRQs there are, can * Rough estimation of how many shared IRQs there are, can
* be changed anytime. * be changed anytime.
*/ */
#define MAX_PLUS_SHARED_IRQS NR_IRQ_VECTORS #define MAX_PLUS_SHARED_IRQS NR_IRQS
#define PIN_MAP_SIZE (MAX_PLUS_SHARED_IRQS + NR_IRQS) #define PIN_MAP_SIZE (MAX_PLUS_SHARED_IRQS + NR_IRQS)
/* /*
...@@ -615,7 +615,7 @@ static int pin_2_irq(int idx, int apic, int pin) ...@@ -615,7 +615,7 @@ static int pin_2_irq(int idx, int apic, int pin)
/* irq_vectors is indexed by the sum of all RTEs in all I/O APICs. */ /* irq_vectors is indexed by the sum of all RTEs in all I/O APICs. */
static u8 irq_vector[NR_IRQ_VECTORS] __read_mostly = { static u8 irq_vector[NR_IRQS] __read_mostly = {
[0] = FIRST_EXTERNAL_VECTOR + 0, [0] = FIRST_EXTERNAL_VECTOR + 0,
[1] = FIRST_EXTERNAL_VECTOR + 1, [1] = FIRST_EXTERNAL_VECTOR + 1,
[2] = FIRST_EXTERNAL_VECTOR + 2, [2] = FIRST_EXTERNAL_VECTOR + 2,
...@@ -634,7 +634,7 @@ static u8 irq_vector[NR_IRQ_VECTORS] __read_mostly = { ...@@ -634,7 +634,7 @@ static u8 irq_vector[NR_IRQ_VECTORS] __read_mostly = {
[15] = FIRST_EXTERNAL_VECTOR + 15, [15] = FIRST_EXTERNAL_VECTOR + 15,
}; };
static cpumask_t irq_domain[NR_IRQ_VECTORS] __read_mostly = { static cpumask_t irq_domain[NR_IRQS] __read_mostly = {
[0] = CPU_MASK_ALL, [0] = CPU_MASK_ALL,
[1] = CPU_MASK_ALL, [1] = CPU_MASK_ALL,
[2] = CPU_MASK_ALL, [2] = CPU_MASK_ALL,
...@@ -671,7 +671,7 @@ static int __assign_irq_vector(int irq, cpumask_t mask, cpumask_t *result) ...@@ -671,7 +671,7 @@ static int __assign_irq_vector(int irq, cpumask_t mask, cpumask_t *result)
int old_vector = -1; int old_vector = -1;
int cpu; int cpu;
BUG_ON((unsigned)irq >= NR_IRQ_VECTORS); BUG_ON((unsigned)irq >= NR_IRQS);
/* Only try and allocate irqs on cpus that are present */ /* Only try and allocate irqs on cpus that are present */
cpus_and(mask, mask, cpu_online_map); cpus_and(mask, mask, cpu_online_map);
...@@ -758,7 +758,7 @@ void __setup_vector_irq(int cpu) ...@@ -758,7 +758,7 @@ void __setup_vector_irq(int cpu)
int irq, vector; int irq, vector;
/* Mark the inuse vectors */ /* Mark the inuse vectors */
for (irq = 0; irq < NR_IRQ_VECTORS; ++irq) { for (irq = 0; irq < NR_IRQS; ++irq) {
if (!cpu_isset(cpu, irq_domain[irq])) if (!cpu_isset(cpu, irq_domain[irq]))
continue; continue;
vector = irq_vector[irq]; vector = irq_vector[irq];
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册