latency improvement in __smp_call_single_queue
PATCH 1/1 latency improvement in __smp_call_single_queue
If an interrupt arrives between llist_add and send_call_function_single_ipi in the following code snippet, then the remote CPU will not receive the IPI in a timely manner and subsequent SMP calls even from other CPUs for other functions will be delayed:
if (llist_add(node, &per_cpu(call_single_queue, cpu))) send_call_function_single_ipi(cpu);
Note: llist_add returns 1 if it was empty before the operation.
CPU 0 | CPU 1 | CPU 2 __smp_call_single_q(2,f1) | __smp_call_single_q(2,f2) | llist_add returns 1 | | interrupted | llist_add returns 0 | ... | branch not taken | ... | | resumed | | send_call_function_single_ipi | |
| | f1 | | f2
The call from CPU 1 for function f2 will be delayed because CPU 0 was interrupted.
---
kernel/smp.c | 4 ++++
1 file changed, 4 insertions(+)
kernel/smp.c +0 -0
@@ -135,6 +135,8 @@ static DEFINE_PER_CPU_SHARED_ALIGNED(call_single_data_t, csd_data);
void __smp_call_single_queue(int cpu, struct llist_node *node)
{
+ unsigned long flags;
+
/*
* The list addition should be visible before sending the IPI
* handler locks the list to pull the entry off it because of
@@ -146,8 +148,10 @@ void __smp_call_single_queue(int cpu, struct llist_node *node)
* locking and barrier primitives. Generic code isn't really
* equipped to do the right thing...
*/
+ local_irq_save(flags);
if (llist_add(node, &per_cpu(call_single_queue, cpu))) send_call_function_single_ipi(cpu);
+ local_irq_restore(flags);
}
/*
Signature · 1 line
2.16.6
Quoted 18 lines · George Prekas
If an interrupt arrives between
llist_addandsend_call_function_single_ipiin the following code snippet, then the remote CPU will not receive the IPI in a timely manner and subsequent SMP calls even from other CPUs for other functions will be delayed:if (
llist_add(node, &per_cpu(call_single_queue, cpu)))send_call_function_single_ipi(cpu);Note:
llist_addreturns 1 if it was empty before the operation.CPU 0 | CPU 1 | CPU 2 __smp_call_single_q(2,f1) |
__smp_call_single_q(2,f2)|llist_addreturns 1 | | interrupted |llist_addreturns 0 | ... | branch not taken | ... | | resumed | |send_call_function_single_ipi| |
Quoted 5 lines
| | f1 | | f2The call from CPU 1 for function f2 will be delayed because CPU 0 was interrupted.
Do you happen to have any actual numbers and a use-case where this was relevant?
Quoted 22 lines · On 9/24/2020 3:42 AM, [email protected] wrote:
Quoted 21 lines · George Prekas
If an interrupt arrives between
llist_addandsend_call_function_single_ipiin the following code snippet, then the remote CPU will not receive the IPI in a timely manner and subsequent SMP calls even from other CPUs for other functions will be delayed:if (
llist_add(node, &per_cpu(call_single_queue, cpu)))send_call_function_single_ipi(cpu);Note:
llist_addreturns 1 if it was empty before the operation.CPU 0 | CPU 1 | CPU 2 __smp_call_single_q(2,f1) |
__smp_call_single_q(2,f2)|llist_addreturns 1 | |
interrupted |llist_addreturns 0 | ... | branch not taken | ... | |
resumed | |send_call_function_single_ipi| || | f1 | | f2
Quoted 5 lines
Quoted 2 lines
The call from CPU 1 for function f2 will be delayed because CPU 0 was interrupted.
Do you happen to have any actual numbers and a use-case where this was relevant?
Hi Peter,
I encountered this problem while developing a device driver that used smp_call_function_single to communicate with other cores. I observed latency spikes and after investigation I figured out the problem described above.
I have written a simple device driver to validate the above fix. It does smp_call_function_single and measures the latency. I can post it here if it is appropriate. The latency impact is equal to the duration of the CPU 0's interruption.
--
George