Could you please provide .jtr files along with the jtreg test results?

Volker Simonis volker.simonis at gmail.com
Tue Jan 10 10:48:44 UTC 2017


Hi Rory,

thanks for the confirmation. Vladimir just confirmed this as well on
the hotspot list.
The RTM feature is not on by default so this problem is not as sever
as I initially thought.

The unfortunate thing is that these tests are showing up "Passed"
instead of "Not executed". I think this is not easy to fix in jtreg
because the tests are actually executed and only at runtime the test
decides that its pay load can't be executed on the actual hardware.

Maybe we need some special return codes such that tests can express
that a tests wasn't really executed. However that would probably
require the redesign of many test because many of them execute several
subtests depending on the platform. But I think it would be definitely
valuable if the jtreg results would be more accurate.

Regards,
Volker


On Tue, Jan 10, 2017 at 10:26 AM, Rory O'Donnell
<rory.odonnell at oracle.com> wrote:
> Hi Volker,
>
> Finally got the info on the test machine used to run the tests, as you
> expected the machine doesn't
> support transactional memory.
>
> Rgds,Rory
>
> This is the output of the command :
> [0.011s][info][os,cpu] flags
> : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36
> clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt
> rdtscp lm 3dnowext 3dnow rep_good extd_apicid pni cx16 lahf_lm cmp_legacy
> svm extapic cr8_legacy
>
>
>
> On 05/01/2017 10:57, Volker Simonis wrote:
>
> Hi,
>
> when looking at the jdk9 test results at:
>
> http://download.java.net/openjdk/testresults/9/testresults.html
>
> it looks like all the transactional memory tests
> (/hotspot/test/compiler/rtm) passed.
>
> However when running them locally, I always get 8 failures (and 30
> tests passing).
>
> The problem is that the RTM tests are flagged as "Passed" if they run
> on a machine which does not support transactional memory (because
> either the CPU or the OS is too old). It would therefore help if you
> could offer the test result files (i.e. the .jtr files) for every test
> such that one could see if a test really passed or if it just passed
> because it was not properly executed.
>
> If you can't offer the .jtr files in the near term, could you please
> let me know if your test machine supports transactional memory? You
> can easily check this by verifying that the following java command:
>
> java -Xlog:os+cpu -version | grep rtm
>
> outputs something like:
>
> [0.007s][info][os,cpu] flags            : fpu vme de pse tsc msr pae
> mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr
> sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc
> arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf
> eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 fma
> cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt
> tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm arat epb pln
> pts dtherm tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust
> bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm xsaveopt cqm_llc
> cqm_occup_llc
>
>
> Thank you and best regards,
> Volker
>
>
> --
> Rgds,Rory O'Donnell
> Quality Engineering Manager
> Oracle EMEA , Dublin, Ireland


More information about the quality-discuss mailing list