Skip to contentSkip to conversation

genirq: avoid long loops in handle_edge_irq

PATCH Superseded Wei Gong 2 participants 6 messages Sep 25, 2023 – Sep 28, 2023 IRQ

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Wei Gong

author
#0 lore raw

PATCH v2 1/1 genirq: avoid long loops in handle_edge_irq

When there are a large number of interrupts occurring on the tx queue(irq smp_affinity=1) of the network card, changing the CPU affinity of the tx queue (echo 2 > /proc/irq/xx/smp_affinity) will cause handle_edge_irq to loop for a long time in the do {} while() loop.

After setting the IRQ CPU affinity, the next interrupt will only be activated when it arrives. Therefore, the next interrupt will still be on CPU 0. When a new CPU affinity is activated on CPU 0, subsequent interrupts will be processed on CPU 1.

cpu 0 cpu 1
- handle_edge_irq
- apic_ack_irq
- irq_do_set_affinity
- handle_edge_irq
- do {
- handle_irq_event
- istate &= ~IRQS_PENDIN
- IRQD_IRQ_INPROGRESS
- spin_unlock()
- spin_lock()
- istate |= IRQS_PENDIN
- handle_irq_event_percpu - mask_ack_irq()
- spin_unlock()
- spin_unlock

} while(IRQS_PENDIN &&
!irq_disable)

Therefore, when determining whether to continue looping, we add a check to see if the current CPU belongs to the affinity table of the interrupt.

---
 kernel/irq/chip.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/irq/chip.c +2 -1
index dc94e0bf2c94..6da455e1a692 100644
@@ -831,7 +831,8 @@ void handle_edge_irq(struct irq_desc *desc)
 		handle_irq_event(desc);
 
 	} while ((desc->istate & IRQS_PENDING) &&
-		 !irqd_irq_disabled(&desc->irq_data));
+		 !irqd_irq_disabled(&desc->irq_data) &&
+		 cpumask_test_cpu(smp_processor_id(), irq_data_get_affinity_mask(&desc->irq_data)));
 
 out_unlock:
 	raw_spin_unlock(&desc->lock);
Signature · 1 line
2.32.1 (Apple Git-133)
#1 lore raw

kernel/irq/chip.c void handle_edge_irq(struct irq_desc *desc) · line 835

Quoted 11 lines · Wei Gong
kernel/irq/chip.c +2 -1
index dc94e0bf2c94..6da455e1a692 100644
@@ -831,7 +831,8 @@ void handle_edge_irq(struct irq_desc *desc)
 		handle_irq_event(desc);
 
 	} while ((desc->istate & IRQS_PENDING) &&
-		 !irqd_irq_disabled(&desc->irq_data));
+		 !irqd_irq_disabled(&desc->irq_data) &&
+		 cpumask_test_cpu(smp_processor_id(), irq_data_get_affinity_mask(&desc->irq_data)));

Assume affinty mask has CPU0 and CPU1 set and the loop is on CPU0, but the effective affinity is on CPU1 then how is this going to move the interrupt?

Thanks,

tglx

author
#2 lore raw
Quoted 8 lines · Thomas Gleixner
Quoted 7 lines · Wei Gong
kernel/irq/chip.c +0 -0
index dc94e0bf2c94..6da455e1a692 100644
@@ -831,7 +831,8 @@ void handle_edge_irq(struct irq_desc *desc)
 		handle_irq_event(desc);
 
Quoted 8 lines

kernel/irq/chip.c void handle_edge_irq(struct irq_desc *desc) · line 835

Quoted 4 lines
kernel/irq/chip.c +2 -1
 	} while ((desc->istate & IRQS_PENDING) &&
-		 !irqd_irq_disabled(&desc->irq_data));
+		 !irqd_irq_disabled(&desc->irq_data) &&
+		 cpumask_test_cpu(smp_processor_id(), irq_data_get_affinity_mask(&desc->irq_data)));

Assume affinty mask has CPU0 and CPU1 set and the loop is on CPU0, but the effective affinity is on CPU1 then how is this going to move the interrupt?

Loop is on the CPU0 means that the previous effective affinity was on CPU0. When the previous effective affinity is a subset of the new affinity mask, the effective affinity will not be updated.
Therefore, I understand that the scenario you mentioned will not occur?

Quoted 4 lines

Thanks,

tglx

Thanks,

Wei Gong

#3 lore raw
Quoted 18 lines · Wei Gong
Quoted 8 lines · Thomas Gleixner
Quoted 7 lines · Wei Gong
kernel/irq/chip.c +0 -0
index dc94e0bf2c94..6da455e1a692 100644
@@ -831,7 +831,8 @@ void handle_edge_irq(struct irq_desc *desc)
 		handle_irq_event(desc);
 
Quoted 8 lines

kernel/irq/chip.c void handle_edge_irq(struct irq_desc *desc) · line 835

Quoted 4 lines
kernel/irq/chip.c +2 -1
 	} while ((desc->istate & IRQS_PENDING) &&
-		 !irqd_irq_disabled(&desc->irq_data));
+		 !irqd_irq_disabled(&desc->irq_data) &&
+		 cpumask_test_cpu(smp_processor_id(), irq_data_get_affinity_mask(&desc->irq_data)));

Assume affinty mask has CPU0 and CPU1 set and the loop is on CPU0, but the effective affinity is on CPU1 then how is this going to move the interrupt?

Quoted 3 lines

Loop is on the CPU0 means that the previous effective affinity was on CPU0. When the previous effective affinity is a subset of the new affinity mask, the effective affinity will not be updated.

That's an implementation detail of a particular interrupt chip driver, but not a general guaranteed behaviour.

author
#4 lore raw
Quoted 10 lines · Thomas Gleixner
Quoted 9 lines · Wei Gong
Quoted 8 lines · Thomas Gleixner
Quoted 7 lines · Wei Gong
kernel/irq/chip.c +0 -0
index dc94e0bf2c94..6da455e1a692 100644
@@ -831,7 +831,8 @@ void handle_edge_irq(struct irq_desc *desc)
 		handle_irq_event(desc);
 
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

kernel/irq/chip.c void handle_edge_irq(struct irq_desc *desc) · line 835

Quoted 4 lines
kernel/irq/chip.c +2 -1
 	} while ((desc->istate & IRQS_PENDING) &&
-		 !irqd_irq_disabled(&desc->irq_data));
+		 !irqd_irq_disabled(&desc->irq_data) &&
+		 cpumask_test_cpu(smp_processor_id(), irq_data_get_affinity_mask(&desc->irq_data)));

Assume affinty mask has CPU0 and CPU1 set and the loop is on CPU0, but the effective affinity is on CPU1 then how is this going to move the interrupt?

Can replacing irq_data_get_affinity_mask with irq_data_get_effective_affinity_mask solve this issue?

Quoted 8 lines
Quoted 4 lines

Loop is on the CPU0 means that the previous effective affinity was on CPU0. When the previous effective affinity is a subset of the new affinity mask, the effective affinity will not be updated.

That's an implementation detail of a particular interrupt chip driver, but not a general guaranteed behaviour.

#5 lore raw
Quoted 24 lines · Wei Gong
Quoted 23 lines · Thomas Gleixner
Quoted 21 lines · Wei Gong
Quoted 8 lines · Thomas Gleixner
Quoted 7 lines · Wei Gong
kernel/irq/chip.c +0 -0
index dc94e0bf2c94..6da455e1a692 100644
@@ -831,7 +831,8 @@ void handle_edge_irq(struct irq_desc *desc)
 		handle_irq_event(desc);
 
Quoted 8 lines

kernel/irq/chip.c void handle_edge_irq(struct irq_desc *desc) · line 835

Quoted 4 lines
kernel/irq/chip.c +2 -1
 	} while ((desc->istate & IRQS_PENDING) &&
-		 !irqd_irq_disabled(&desc->irq_data));
+		 !irqd_irq_disabled(&desc->irq_data) &&
+		 cpumask_test_cpu(smp_processor_id(), irq_data_get_affinity_mask(&desc->irq_data)));

Assume affinty mask has CPU0 and CPU1 set and the loop is on CPU0, but the effective affinity is on CPU1 then how is this going to move the interrupt?

Loop is on the CPU0 means that the previous effective affinity was on CPU0. When the previous effective affinity is a subset of the new affinity mask, the effective affinity will not be updated.

Quoted 6 lines
Quoted 3 lines

That's an implementation detail of a particular interrupt chip driver, but not a general guaranteed behaviour.

Can replacing irq_data_get_affinity_mask with irq_data_get_effective_affinity_mask solve this issue?

Yes.