Re: The idea about scheduler test module(STM)

2012-10-25 Thread Michael Wang
On 10/26/2012 10:27 AM, Charles Wang wrote: > Yes, it's a new way to do scheduler test. But why use kernel > threads? The info, total time, run time, wait time, preempt number, all > can be collected from tasks' sched info from /proc/pid/sched and > /proc/pid/stat. That's right, next version's ST

Re: The idea about scheduler test module(STM)

2012-10-25 Thread Charles Wang
Yes, it's a new way to do scheduler test. But why use kernel threads? The info, total time, run time, wait time, preempt number, all can be collected from tasks' sched info from /proc/pid/sched and /proc/pid/stat. I don't understand clearly about "pure scheduler performance" here. In order to

The idea about scheduler test module(STM)

2012-10-24 Thread Michael Wang
Hi, Folks Charles has raised a problem that we don't have any tool yet for testing the scheduler with out any disturb from other subsystem, and I also found it's hard to test scheduler optimize patch, since the improvement could be easily eaten by other subsystem like IO. So Let's check the tools