SLIDE 1
Threaded Network Interrupts Steven Rostedt srostedt@redhat.com - - PowerPoint PPT Presentation
Threaded Network Interrupts Steven Rostedt srostedt@redhat.com - - PowerPoint PPT Presentation
Threaded Network Interrupts Steven Rostedt srostedt@redhat.com <rostedt@goodmis.org> http://people.redhat.com/srostedt network-thread-irq.odp Disclaimer I'm not a network guru I had to fight for the hardware I have Threaded Interrupts
SLIDE 2
SLIDE 3
Threaded Interrupts
- Similar to NAPI code
- Interrupt comes in, disable the device
- Thread is awoken
- Thread polls on device, processing packets
- Finished and enables device
SLIDE 4
Benchmarking
- On my 1Gb NICs, all kernels had same result
in netperf
– Saturated network
- 10 Gig NIC
– Could not get working with the threaded
infrastructure
- I'm not a network device expert
– Decided to just give up and test with the RT
patch
- This is far from optimizing the design
SLIDE 5
Benchmark Machine
- Two Nehalem machines
- 2 x 4 core Intel Xeons
- Chelsio Communications Inc T310 10GbE
Single Port Adapter
SLIDE 6
Benchmark
- Ran netperf -H host
- Red Hat Enterprise Linux 5.4
- Full Preempt Real Time Patch (PREEMPT_RT)
– 2.6.31-rt10
- Hard and Soft interrupts as threads
– RT patch without PREEMPT_RT enabled
- Soft interrupts as threads
– RT patch with only PREEMPT_SOFTIRQ set
- Vanilla Linus Kernel
– 2.6.31 downloaded from kernel.org
SLIDE 7
SLIDE 8
SLIDE 9
SLIDE 10
SLIDE 11
SLIDE 12
SLIDE 13
SLIDE 14
SLIDE 15
SLIDE 16
SLIDE 17
SLIDE 18
Cons
- Horrible latency on server config (no kernel
preemption)
- Redesign of network infrastructure
SLIDE 19
Pros
- Can schedule in handler
- Easier locking algorithms
- Prioritize interrupts
- Smoother desktop experience
SLIDE 20
Discussion
SLIDE 21
1
Threaded Network Interrupts
Steven Rostedt srostedt@redhat.com <rostedt@goodmis.org> http://people.redhat.com/srostedt network-thread-irq.odp
SLIDE 22
2
Disclaimer
I'm not a network guru I had to fight for the hardware I have
SLIDE 23
3
Threaded Interrupts
- Similar to NAPI code
- Interrupt comes in, disable the device
- Thread is awoken
- Thread polls on device, processing packets
- Finished and enables device
SLIDE 24
4
Benchmarking
- On my 1Gb NICs, all kernels had same result
in netperf
– Saturated network
- 10 Gig NIC
– Could not get working with the threaded
infrastructure
- I'm not a network device expert
– Decided to just give up and test with the RT
patch
- This is far from optimizing the design
SLIDE 25
5
Benchmark Machine
- Two Nehalem machines
- 2 x 4 core Intel Xeons
- Chelsio Communications Inc T310 10GbE
Single Port Adapter
SLIDE 26
6
Benchmark
- Ran netperf -H host
- Red Hat Enterprise Linux 5.4
- Full Preempt Real Time Patch (PREEMPT_RT)
– 2.6.31-rt10
- Hard and Soft interrupts as threads
– RT patch without PREEMPT_RT enabled
- Soft interrupts as threads
– RT patch with only PREEMPT_SOFTIRQ set
- Vanilla Linus Kernel
– 2.6.31 downloaded from kernel.org
SLIDE 27
7
SLIDE 28
8
SLIDE 29
9
SLIDE 30
10
SLIDE 31
11
SLIDE 32
12
SLIDE 33
13
SLIDE 34
14
SLIDE 35
15
SLIDE 36
16
SLIDE 37
17
SLIDE 38
18
Cons
- Horrible latency on server config (no kernel
preemption)
- Redesign of network infrastructure
SLIDE 39
19
Pros
- Can schedule in handler
- Easier locking algorithms
- Prioritize interrupts
- Smoother desktop experience
SLIDE 40
20