perfasm and the <kernel> region

[Chris Vest]

Yeah, this goes even further than the system call boundary. Pretty cool.

Cheers,
Chris

> On 23 Apr 2015, at 16:10, Aleksey Shipilev <aleksey.shipilev at oracle.com> wrote:
> 
> On 04/23/2015 04:07 PM, Aleksey Shipilev wrote:
>> On 04/23/2015 03:39 PM, Chris Vest wrote:
>>> I’m assuming the time spent in the system calls for the IO and the
>>> blocking/unblocking of threads are both included in the <kernel>
>>> region. Would it be possible to break them out? I tried strace’ing
>>> the benchmark, but the overhead of strace is so big that it changes
>>> the behavioural patterns of the code, ruining the measurement.
>> 
>> "perfasm" processes whatever events perf is providing back to us. At
>> this point, we attribute all events from the negative addresses as the
>> <kernel> region. One can massage perfasm code to parse those properly as
>> well, like this:
> 
> Okay, that's a good improvement on my machines, so I pushed:
>  https://bugs.openjdk.java.net/browse/CODETOOLS-7901405
> 
> The comment in the bug explains what sys options might need to be
> enabled for kernel symbols to be resolveable.
> 
> Thanks,
> -Aleksey
> 
>