More hotspot tests and some backported fixes
Kelly.Ohair at Sun.COM
Tue Sep 30 13:00:17 PDT 2008
Historically, it was the j2se (now jdk) repository that utilized the
test/ directory in the repository to hold what we called 'jtreg' or
regression tests, usually each test mapped to a bug or set of bugs.
Hotspot started using the test/ directory a little later, so that's
why there are so many more of these tests in the jdk repository
(and langtools repository) than the other repositories in the forest.
As to the quality of these jtreg tests, many are good and solid tests,
but some are not as solid or reliable, just a warning.
The Hotspot tests were primarily not opened up due to a lack of time
in verifying ownership of the files, as Chuck has stated in his response.
We still have quite a few jdk tests in the same state, I think.
It takes time to go through the history of the files, and also make sure
they weren't copied in from some other non-opensource friendly places.
One of the issues with these jtreg tests and Hotspot was the issue of
jtreg needing to know how to compile native code, which it did not
learn to do as far as I know (but not sure it should).
Not sure Hotspot has any tests that required native compilation, but
it was an issue at one time.
Volker Simonis wrote:
> I think these kind of tests are intersting and usefull but
> unfortuantely, there are jus a small handfull of them available under
> "hotspot/test/compiler". Does anybody from SUN know if there are more
> of these hotspot/compiler kind of test avaialble within SUN which are
> just waiting to get open sourced?
> I think for porters they would be very interesting and helpfull.
> Especially tests which cover corner cases of the HotSpot like for
> - Java to Native transitions
> - Deoptimzations and OSR in various situations
> - above a compiled frame
> - above an interpreted frame
> - the various adapters (I2C, C2I, I2N,...)
> - the various calls (static, virtual, leaf, ...)
> - stack overflow situations
> - exception handling in various situations
> - JIT inlining
> - Synchronization (lightweight-, heavyweight-, biased locki
> Does anybody know if such kind of tests exist somewhere?
> On 9/26/08, Mark Wielaard <mark at klomp.org> wrote:
>> I noticed openjdk6 and openjdk7 seem to have a mutually exclusive set of
>> tests under hotspot/test/compiler. Since I believe having more tests is
>> a good thing I merged them in icedtea6. That uncovered two crasher bugs
>> in the hotspot form openjdk6 that we use. So I tracked down the fixes
>> for those and backported them (patches attached). All 20 tests now PASS.
>> It is probably a good idea to also merge the tests from 6 to 7. But I
>> haven't done that yet.
>> Note that again I did only backport for the hotspot from openjdk6, not
>> the one that zero/shark uses (7b24).
More information about the hotspot-compiler-dev