hg: jdk7/hotspot-rt/hotspot: 6888526: Linux getCurrentThreadCpuTime is drastically slower than Windows

[John Coomes]

Daniel D. Daugherty (daniel.daugherty at oracle.com) wrote:
> 
> 
> On 7/29/2010 5:38 PM, John Coomes wrote:
> > aph at redhat.com (aph at redhat.com) wrote:
> >   
> >> Changeset: 3d90023429ec
> >> Author:    aph
> >> Date:      2010-07-28 17:38 +0100
> >> URL:       http://hg.openjdk.java.net/jdk7/hotspot-rt/hotspot/rev/3d90023429ec
> >>
> >> 6888526: Linux getCurrentThreadCpuTime is drastically slower than Windows
> >> Reviewed-by: dcubed, dholmes
> >>
> >> ! src/os/linux/vm/globals_linux.hpp
> >> ! src/share/vm/runtime/arguments.cpp
> >
> > Hi Andrew,
> >
> > First, thanks for the fix.
> >
> > I also want to make sure you're aware of our requirements for pushing
> > to hotspot, since this looks like something you pushed directly.  All
> > hotspot changes (even trivial ones) have to go through an automated
> > build and test system that we call JPRT, which pushes your change only
> > if all builds and test pass on all platforms.  As of now that system
> > is only available internally, so in the future when you have a hotspot
> > change ready, ask on the appropriate hotspot-*-dev list and someone
> > will submit the changeset for you.  It's worth noting that all hotspot
> > pushes from oracle engineers go through JPRT.
> >
> > I realize it's a slight inconvenience, but we had much too frequent
> > build and test breakage before we required everyone to use JPRT.
> 
> That's my fault. I gave Andrew permission to push the change to
> RT_Baseline. I guess I've spent too much time in T&L lately where
> external pushes happen without going thru JPRT. In any case,
> Coleen pushed a subsequent job through JPRT after Andrew's and
> I've thoroughly checked the nightly test results.
>
> In the future, I'll shepherd changes through JPRT (as I did for
> the stuff that AMD contributed)...

Sounds good.  And thanks for checking the nightlies; good to know that
all's well.

-John