[PATCH] 1/2 Add exposure of the irq delivery mask on x86

[PATCH] 1/2 Add exposure of the irq delivery mask on x86

[James Bottomley]

[-- Attachment #1: Type: text/plain, Size: 213 bytes --]


This exposes a delivery mask in /proc/<irq>/mask for the userspace irq
balancer to use in its calculations.

The only current consumer of this on x86 would be voyager I think (patch
for voyager follows).

James


[-- Attachment #2: tmp.diff --]
[-- Type: text/plain, Size: 2377 bytes --]

# This is a BitKeeper generated patch for the following project:
# Project Name: Linux kernel tree
# This patch format is intended for GNU patch command version 2.5 or higher.
# This patch includes the following deltas:
#	           ChangeSet	1.1209  -> 1.1210 
#	arch/i386/kernel/irq.c	1.36    -> 1.37   
#	include/asm-i386/irq.h	1.9     -> 1.10   
#
# The following is the BitKeeper ChangeSet Log
# --------------------------------------------
# 03/05/22	jejb@oldfenric.sc.steeleye.com	1.1210
# Add irq masks into /proc/<irq> for x86
# 
# This exposes the mask both as an external variable and through the /proc
# interface.  Machine specific code is still required to reflect the true
# value.
# --------------------------------------------
#
diff -Nru a/arch/i386/kernel/irq.c b/arch/i386/kernel/irq.c
--- a/arch/i386/kernel/irq.c	Thu May 22 13:13:57 2003
+++ b/arch/i386/kernel/irq.c	Thu May 22 13:13:57 2003
@@ -868,8 +868,20 @@
 #ifdef CONFIG_SMP
 
 static struct proc_dir_entry * smp_affinity_entry [NR_IRQS];
+static struct proc_dir_entry * smp_mask_entry [NR_IRQS];
 
 unsigned long irq_affinity [NR_IRQS] = { [0 ... NR_IRQS-1] = ~0UL };
+unsigned long irq_mask [NR_IRQS] = { [0 ... NR_IRQS-1] = ~0UL };
+
+static int irq_mask_read_proc(char *page, char **start, off_t off,
+			int count, int *eof, void *data)
+{
+	if (count < HEX_DIGITS+1)
+		return -EINVAL;
+	return sprintf (page, "%08lx\n", irq_mask [(long)data]);
+}
+
+
 static int irq_affinity_read_proc (char *page, char **start, off_t off,
 			int count, int *eof, void *data)
 {
@@ -959,6 +971,18 @@
 		}
 
 		smp_affinity_entry[irq] = entry;
+
+		entry = create_proc_entry("mask", 0400, irq_dir[irq]);
+
+		if (entry) {
+			entry->nlink = 1;
+			entry->data = (void *)(long)irq;
+			entry->read_proc = irq_mask_read_proc;
+			entry->write_proc = NULL;
+		}
+
+		smp_mask_entry[irq] = entry;
+
 	}
 #endif
 }
diff -Nru a/include/asm-i386/irq.h b/include/asm-i386/irq.h
--- a/include/asm-i386/irq.h	Thu May 22 13:13:57 2003
+++ b/include/asm-i386/irq.h	Thu May 22 13:13:57 2003
@@ -25,6 +25,8 @@
 extern void enable_irq(unsigned int);
 extern void release_x86_irqs(struct task_struct *);
 
+extern unsigned long irq_mask [NR_IRQS];
+
 #ifdef CONFIG_X86_LOCAL_APIC
 #define ARCH_HAS_NMI_WATCHDOG		/* See include/linux/nmi.h */
 #endif

Re: [PATCH] 1/2 Add exposure of the irq delivery mask on x86

[Keith Mannthey]

On Thu, 2003-05-22 at 15:50, Keith Mannthey wrote:
> >>This exposes a delivery mask in /proc/<irq>/mask for the userspace irq
> >>balancer to use in its calculations.
> 
>   What exactly are you trying to expose? These patches look to just show
> what cpus are on-line.  Wouldn't this just be the cpu_online_map in all
> cases?  I guess I am a little confuses by what a "delivery mask" is.
> 
> Thanks,  
>   Keith  
> 
Sorry I read the voyager.txt file and I understand what you are trying
to do now. 

Keith 

Re: [PATCH] 1/2 Add exposure of the irq delivery mask on x86

[Keith Mannthey]

>>This exposes a delivery mask in /proc/<irq>/mask for the userspace irq
>>balancer to use in its calculations.

  What exactly are you trying to expose? These patches look to just show
what cpus are on-line.  Wouldn't this just be the cpu_online_map in all
cases?  I guess I am a little confuses by what a "delivery mask" is.

Thanks,  
  Keith