segfault when building ghc head on os x 10.8 with -fllvm but not -fasm

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

segfault when building ghc head on os x 10.8 with -fllvm but not -fasm

Carter Schonwald
Hey All,
i've opened a ticket for this build problem
http://hackage.haskell.org/trac/ghc/ticket/7874

ghc seems to build successful when i do quick, but not quick-llvm. Which
seems worrisome maybe?

was with current ghc head, commit
hash 2fa44a8ab9d6c68c14ccb25ac26a556d1d82d70d

any thoughts would be appreciated on what could be the culprit and how i
can track it down / help fix it.

thanks
-Carter

heres's the results of my make test run

OVERALL SUMMARY for test run started at Tue Apr 30 16:06:51 EDT 2013
    3628 total tests, which gave rise to
   14337 test cases, of which
   11044 were skipped

      28 had missing libraries
    3182 expected passes
      56 expected failures

       3 caused framework failures
       2 unexpected passes
      22 unexpected failures

Unexpected passes:
   driver           T5313 (normal)
   perf/should_run  T149 (normal)

Unexpected failures:
   codeGen/should_run        cgrun071 [bad exit code] (normal)
   driver                    T2507 [bad stderr] (normal)
   ghc-api/T7478             T7478 [bad exit code] (normal)
   perf/compiler             T1969 [stat not good enough] (normal)
   perf/compiler             T3064 [stat not good enough] (normal)
   perf/compiler             T3294 [stat not good enough] (normal)
   perf/compiler             T4801 [stat not good enough] (normal)
   perf/compiler             T5030 [stat not good enough] (normal)
   perf/compiler             T5321FD [stat not good enough] (normal)
   perf/compiler             T5321Fun [stat not good enough] (normal)
   perf/compiler             T5631 [stat not good enough] (normal)
   perf/compiler             T5642 [stat not good enough] (normal)
   perf/compiler             T5837 [stat not good enough] (normal)
   perf/compiler             T6048 [stat not good enough] (optasm)
   perf/compiler             T783 [stat not good enough] (normal)
   perf/compiler             parsing001 [stat not good enough] (normal)
   perf/should_run           T4474a [stat not good enough] (normal)
   perf/should_run           T4474b [stat not good enough] (normal)
   perf/should_run           T4474c [stat not good enough] (normal)
   simplCore/should_compile  T4908 [stderr mismatch] (optasm)
   simplCore/should_compile  T7360 [stderr mismatch] (optasm)
   simplCore/should_compile  spec-inline [stderr mismatch] (optasm)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/ghc-devs/attachments/20130430/94bea4dd/attachment.htm>

Reply | Threaded
Open this post in threaded view
|

segfault when building ghc head on os x 10.8 with -fllvm but not -fasm

Carter Schonwald
using ghc build via the -fasm route and then
cding  into ./testsuite/tests and running
 make WAYS=llvm test

i get the following results (also added to the ticket i opened)

Unexpected results from:
TEST="T5313 conc058 conc068 conc065 conc066 conc020 conc035 conc015 conc017
T7571 dynamic004 T3016 qq007 qq008 ffi009"

OVERALL SUMMARY for test run started at Tue Apr 30 17:14:35 EDT 2013
    3628 total tests, which gave rise to
   16632 test cases, of which
   14905 were skipped

      24 had missing libraries
    1672 expected passes
      16 expected failures

       0 caused framework failures
       1 unexpected passes
      14 unexpected failures

Unexpected passes:
   driver  T5313 (llvm)

Unexpected failures:
   ../../libraries/base/tests  dynamic004 [bad exit code] (llvm)
   concurrent/should_run       conc015 [exit code non-0] (llvm)
   concurrent/should_run       conc017 [exit code non-0] (llvm)
   concurrent/should_run       conc020 [exit code non-0] (llvm)
   concurrent/should_run       conc035 [exit code non-0] (llvm)
   concurrent/should_run       conc058 [exit code non-0] (llvm)
   concurrent/should_run       conc065 [exit code non-0] (llvm)
   concurrent/should_run       conc066 [exit code non-0] (llvm)
   concurrent/should_run       conc068 [exit code non-0] (llvm)
   ffi/should_run              ffi009 [bad exit code] (llvm)
   llvm/should_compile         T7571 [stderr mismatch] (llvm)
   quasiquotation/qq007        qq007 [exit code non-0] (llvm)
   quasiquotation/qq008        qq008 [exit code non-0] (llvm)
   simplCore/should_compile    T3016 [exit code non-0] (llvm)


On Tue, Apr 30, 2013 at 4:56 PM, Carter Schonwald <
carter.schonwald at gmail.com> wrote:

> Hey All,
> i've opened a ticket for this build problem
> http://hackage.haskell.org/trac/ghc/ticket/7874
>
> ghc seems to build successful when i do quick, but not quick-llvm. Which
> seems worrisome maybe?
>
> was with current ghc head, commit
> hash 2fa44a8ab9d6c68c14ccb25ac26a556d1d82d70d
>
> any thoughts would be appreciated on what could be the culprit and how i
> can track it down / help fix it.
>
> thanks
> -Carter
>
> heres's the results of my make test run
>
> OVERALL SUMMARY for test run started at Tue Apr 30 16:06:51 EDT 2013
>     3628 total tests, which gave rise to
>    14337 test cases, of which
>    11044 were skipped
>
>       28 had missing libraries
>     3182 expected passes
>       56 expected failures
>
>        3 caused framework failures
>        2 unexpected passes
>       22 unexpected failures
>
> Unexpected passes:
>    driver           T5313 (normal)
>    perf/should_run  T149 (normal)
>
> Unexpected failures:
>    codeGen/should_run        cgrun071 [bad exit code] (normal)
>    driver                    T2507 [bad stderr] (normal)
>    ghc-api/T7478             T7478 [bad exit code] (normal)
>    perf/compiler             T1969 [stat not good enough] (normal)
>    perf/compiler             T3064 [stat not good enough] (normal)
>    perf/compiler             T3294 [stat not good enough] (normal)
>    perf/compiler             T4801 [stat not good enough] (normal)
>    perf/compiler             T5030 [stat not good enough] (normal)
>    perf/compiler             T5321FD [stat not good enough] (normal)
>    perf/compiler             T5321Fun [stat not good enough] (normal)
>    perf/compiler             T5631 [stat not good enough] (normal)
>    perf/compiler             T5642 [stat not good enough] (normal)
>    perf/compiler             T5837 [stat not good enough] (normal)
>    perf/compiler             T6048 [stat not good enough] (optasm)
>    perf/compiler             T783 [stat not good enough] (normal)
>    perf/compiler             parsing001 [stat not good enough] (normal)
>    perf/should_run           T4474a [stat not good enough] (normal)
>    perf/should_run           T4474b [stat not good enough] (normal)
>    perf/should_run           T4474c [stat not good enough] (normal)
>    simplCore/should_compile  T4908 [stderr mismatch] (optasm)
>    simplCore/should_compile  T7360 [stderr mismatch] (optasm)
>    simplCore/should_compile  spec-inline [stderr mismatch] (optasm)
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/ghc-devs/attachments/20130430/490a9e1d/attachment.htm>