RFR(s): 8214180: Need better granularity for sleeping

Robbin Ehn robbin.ehn at oracle.com
Sat Nov 24 20:38:26 UTC 2018


Hi Dan,

On 2018-11-24 15:28, Daniel D. Daugherty wrote:
> src/hotspot/os/linux/os_linux.cpp
>      old L4030 // Short sleep, direct OS call.
>      old L4031 //
>      old L4032 // Note: certain versions of Linux CFS scheduler (since 2.6.23) 
> do not guarantee
>      old L4033 // sched_yield(2) will actually give up the CPU:
>      old L4034 //
>      old L4035 //   * Alone on this pariticular CPU, keeps running.
>      old L4036 //   * Before the introduction of "skip_buddy" with 
> "compat_yield" disabled
>      old L4037 //     (pre 2.6.39).
>      old L4038 //
>      old L4039 // So calling this with 0 is an alternative.
>      old L4040 //
>          I noticed this comment didn't get copied over to os_posix.cpp.
>          Is the comment about an old enough Linux kernel that it is no
>          longer useful?
> 

At least to me it is not useful. Having it in the generic os_posix.cpp seems
miss-placed ? Let me know if you want it somewhere?

> 
> Thumbs up.

Thanks Dan!

/Robbin


> 
> Dan
> 
> 
>>
>> Passes t1-3.
>>
>> Thanks, Robbin
> 


More information about the hotspot-runtime-dev mailing list