Skip to contentSkip to conversation

When complaining about attempting to set the irq affinity to multiple cpus,

PATCH In review Mark Mason 2 participants 3 messages Sep 21, 2009 – Sep 23, 2009 1 message not in the archive MIPS

lore
Share & tools
lore

Mark Mason

author
#0 lore raw

PATCH 1/1 When complaining about attempting to set the irq affinity to multiple cpus,

---
 arch/mips/sibyte/bcm1480/irq.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletions(-)
arch/mips/sibyte/bcm1480/irq.c +5 -1
index ba59839..fc87ea4 100644
@@ -118,7 +118,11 @@ static int bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask)
 	unsigned int irq_dirty;
 
 	if (cpumask_weight(mask) != 1) {
-		printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq);
+		printk("attempted to set irq affinity for irq %d to multiple CPUs:", irq);
+		/* Print the mask */
+		for_each_cpu(i, mask)
+			printk(" %d", i);
+		printk("\n");
 		return -1;
 	}
 	i = cpumask_first(mask);
Signature · 1 line
1.6.0.2
#1 lore raw
Quoted 8 lines · Mark Mason
 arch/mips/sibyte/bcm1480/irq.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletions(-)
arch/mips/sibyte/bcm1480/irq.c +0 -0
index ba59839..fc87ea4 100644

arch/mips/sibyte/bcm1480/irq.c static int bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask) · line 125

Quoted 9 lines
arch/mips/sibyte/bcm1480/irq.c +5 -1
@@ -118,7 +118,11 @@ static int bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask)
 	unsigned int irq_dirty;
 
 	if (cpumask_weight(mask) != 1) {
-		printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq);
+		printk("attempted to set irq affinity for irq %d to multiple CPUs:", irq);
+		/* Print the mask */
+		for_each_cpu(i, mask)
+			printk(" %d", i);
+		printk("\n");

The whole if block is a zombie that was supposed to have been removed by commit 14275ccdb1e4b487cca745aba994699c426a31ee but then returned in 92241940be501f798cb21db344bbb3d1ec3c4f1c. So I went for a slightly different patch, see below.

Thanks,

Ralf

MIPS: BCM1480: Re-apply patch lost due to bad resolution of merge conflict.

Patch 14275ccdb1e4b487cca745aba994699c426a31ee and d5dedd4507d307eb3f35f21b6e16f336fdc0d82a are conflicting and the conflict was resolved badly in merge 92241940be501f798cb21db344bbb3d1ec3c4f1c resulting in the BCM1480 changes of 14275ccdb1e4b487cca745aba994699c426a31ee getting lost. Sort out the damage.

Reported and initial patch by Mark Mason <[email protected]>.

arch/mips/sibyte/bcm1480/irq.c +0 -4
index ba59839..4070268 100644
@@ -117,10 +117,6 @@ static int bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask)
 	unsigned long flags;
 	unsigned int irq_dirty;
 
-	if (cpumask_weight(mask) != 1) {
-		printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq);
-		return -1;
-	}
 	i = cpumask_first(mask);
 
 	/* Convert logical CPU to physical CPU */

A message that is not in the archive (sent off-list or not yet received): BD3F7F1EFBA6D54DB056C4FFA4514008037D6641C8@SJEXCHCCR01.corp.ad.broadcom.com

#3 lore raw
Quoted 1 line · Mark E Mason

Did we ever figure out why the kernel was attempting to set the affinity to more than one CPU? The hardware certainly supports doing that (but we're not at the moment).

The hardware semantics is a bit ususual. On Sibyte the hardware will route an interrupt to all CPUs set in the affinity mask. On most other systems that I'm aware of it will route the interrupt to only one of CPUs selected by irq_chip->set_affinity(). We don't want such an stampede so the Sibyte interrupt code clears all but the lowest set bit from the set_affinity() argument. Either way, attempting to set the mask to an arbitrary non-empty set is entirely legal so the warning itself was a bug and the volume potencially made them a real problem for a few users.

Ralf