USB interrupt endpoint limitations

1) The endpoint descriptor has a minimum value of 1 ms polling period.

Therefore it would seem that I can get a maximum of 1000 packets per
second with each packet
containing a maximum of 64 bytes of payload data.

Maximum throughput then limited to 64000 bytes per second.

Anyone disagree with this assessment?

2) On some computers it appears that though my USB device requests 1 ms
interrupt endpoint
polling the actual polling rate is 2 ms.

Can anyone confirm this experience and explain it? Or better yet, tell
me how to correct it.

Thanks,
J.R. Heisey
Synaptics, Inc.



.

Relevant Pages

  • Re: Polling tuning and performance
    ... network traffic and that only by polling). ... I am unhappy to report that my fine-tuned interrupt handling ... sucking in packets as fast as possible with minimal loss. ... of power) unless the system can rarely be idle anyway, ...
    (freebsd-performance)
  • Re: Freebsd IP Forwarding performance (question, and some info) [7-stable, current, em, smp]
    ... Turning on polling.. ... packets errs bytes packets errs bytes colls ... All indications are that you are running out of CPU and memory (DMA ...
    (freebsd-net)
  • Re: Device polling performance
    ... Please post your "polling stats and userland benchmarking" results. ... "EVIDENCE ... ... I started a cpu intensive userland task. ... is that you lose packets when it becomes overmatched, ...
    (freebsd-questions)
  • Re: Freebsd IP Forwarding performance (question, and some info) [7-stable, current, em, smp]
    ... packets errs bytes packets errs bytes colls ... Polling turned on provided better performance on 32 bit, but it gets strange errors on 64 bit.. ... kpps load% ips kpps load% ips ...
    (freebsd-net)
  • Re: Proposed 6.2 em RELEASE patch
    ... I am guessing its pretty well the same patch. ... Polling will ... Not sure if I am just hitting the limits of the PCIe bus, or interrupt moderation is not kicking in, or this is a case of "Doctor, it hurts when I send a lot of packets through"... ... Since my box does not take too much load and to me a slightly more loaded machine is better than an unstable one i re-complied the kernel without SMP so I have a dual CPU system with only one of the CPU's working. ...
    (freebsd-stable)