Hi, Jack The principle of my interrupt latency driver is similar to Realfeel.
Realfeel Test of the Preemptible Kernel Patch http://www.linuxjournal.com/article/6405 Download Realfeel: http://brain.mcmaster.ca/~hahn/realfeel.c Realfeel is a user-space program, which can use the standard RTC driver. Regards, Wu Zhangjin On 10/12/10, Jack Daniel <wanders.thi...@gmail.com> wrote: > On Tue, Oct 12, 2010 at 6:56 AM, wu zhangjin <wuzhang...@gmail.com> wrote: >> On Mon, Oct 11, 2010 at 10:00 PM, Jack Daniel <wanders.thi...@gmail.com> >> wrote: >>> On Thu, Oct 7, 2010 at 8:30 PM, wu zhangjin <wuzhang...@gmail.com> wrote: >>>> Hi, Jack >>>> >>>> On 10/7/10, Jack Daniel <wanders.thi...@gmail.com> wrote: >>>>> Hi Wu, >>>>> >>>>> I was going through your latency tracing implementation for loongson. >>>>> Could you please assist me in understanding the following. >>>>> >>>>> 1) How is the latency tracing implemented here different from ftrace >>>>> tracer function "irqsoff" ? "irqsoff" calculates the time when >>>>> interrupts are disabled. >>>> >>>> What do you mean "the latency tracing implemented here"? In reality, >>>> the latency tracers I have used are the ones you said above: irqsoff >>>> tracer, preemptoff tracer and preemptirqsoff tracer, wakeup tracer ... >>> >>> My apologies for not articulating my question properly. What I meant >>> was the latency tracing support available at >>> Device Drivers ---> [*] PREEMPT_RT Specific Device Drivers ---> >>> <*> Interrupt Latency Driver >>> I see that you are the only one to support this in your tree for >>> loongson. How different is this from "irqsoff" supported by Ftrace? >>> >> >> The "Interrupt latency driver" differs from the irqsoff tracer, it >> tries to measure the time between the interrupt emitted and the >> interrupt is really handled, it reflects the response time of an >> interrupt. >> >> Perhaps you can write such one driver for octeon via the RTC driver >> (if your board has such a driver) for we can use RTC to generate >> interrupt, For Loongson, I have used the MFGPT timer. > > Thank you for taking the time to explain this concept. I will try and > implement this tracing method for Octeon by using RTC. First I will > need to study your code. > Thank you once again. > > Regards, > Jack > -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To post to this group, send email to loongson-...@googlegroups.com. To unsubscribe from this group, send email to loongson-dev+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/loongson-dev?hl=en.
