Solarflare Solarflare XtremeScale series, User Manual, Page: 120 / 400

Solarflare Server Adapter User Guide
Solarflare Adapters on Linux
Issue 20 © Solarflare Communications 2017 113
3.26 Interrupt Affinity
Interrupt affinity describes the set of host CPUs that may service a particular
interrupt.
This affinity therefore dictates the CPU context where received packets will be
processed and where transmit packets will be freed once sent. If the application can
process the received packets in the same CPU context by being affinitized to the
relevant CPU, then latency and CPU utilization can be improved. This improvement
is achieved because well tuned affinities reduce inter ‐ CPU communication.
Tuning interrupt affinity is most relevant when MSI ‐ X interrupts and RSS are being
used. The irqbalance service, which typically runs by default in most Linux
distributions, is a service that automatically changes interrupt affinities based on
CPU workload.
In many cases the irqbalance service hinders rather than enhances network
performance. It is therefore necessary to disable it and then set interrupt affinities.
• To disable irqbalance permanently, run:
/sbin/chkconfig ‐ level 12345 irqbalance off
• To see whether irqbalance is currently running, run:
/sbin/service irqbalance status
• To disable irqbalance temporarily, run:
/sbin/service irqbalance stop
Once the irqbalance service has been stopped, the Interrupt affinities can be
configured manually.
NOTE: The Solarflare driver will evenly distribute interrupts across the available
host CPUs (based on the
rss_cpus
module parameter).


To use the Solarflare driver default affinities (recommended), the irqbalance
service must be disabled before the Solarflare driver is loaded (otherwise it will
immediately overwrite the affinity configuration values set by the Solarflare driver).
Example 1:
How affinities should be manually set will depend on the application. For a single
streamed application such as Netperf, one recommendation would be to affinitize
all the Rx queues and the application on the same CPU. This can be achieved with
the following steps:
1 Determine which interrupt line numbers the network interface uses. Assuming
the interface is eth0, this can be done with:
# cat /proc/interrupts | grep eth0 ‐

123: 13302 0 0 0 PCI ‐ MSI ‐ X eth0 ‐ 0

131: 0 24 0 0 PCI ‐ MSI ‐ X eth0 ‐ 1

139: 0 0 32 0 PCI ‐ MSI ‐ X eth0 ‐ 2

147: 0 0 0 21 PCI ‐ MSI ‐ X eth0 ‐ 3