[GHC] #15544: Non-deterministic segmentation fault in cryptohash-sha256 testsuite

[GHC]

#15544: Non-deterministic segmentation fault in cryptohash-sha256 testsuite
-------------------------------------+-------------------------------------
        Reporter:  bgamari           |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  highest           |            Milestone:  8.6.1
       Component:  Compiler          |              Version:  8.4.3
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Old description:

> {{{
> $ cabal get cryptohash-sha256-0.11.101.0
> $ cabal new-run -w ghc-8.6.1 --enable-test --allow-newer=cryptohash-
> sha256:base,*:stm,*:tasty,async:base test:test-sha256 -- -j 8
> --quickcheck-tests 9999
> }}}
>
> Eventually the program will start spamming stderr with `test-sha256: lost
> signal due to full pipe: 11` repeatedly. This apparently only started
> with 8.6.1.

New description:

 {{{
 $ cabal get cryptohash-sha256-0.11.101.0
 $ cd cryptohash-sha256-0.11.101.0
 $ cabal new-run -w ghc-8.6.1 --enable-test --allow-
 newer=*:base,*:stm,*:tasty test:test-sha256 -- -j 8 --quickcheck-tests
 9999
 }}}

 Eventually the program will start spamming stderr with `test-sha256: lost
 signal due to full pipe: 11` repeatedly. This apparently only started with
 8.6.1.

--

Comment (by sjakobi):

 Another variation I'm seeing when I add `--timeout 1s` is

 {{{
     XL-vec
       inc:             test-sha256: internal error: evacuate: strange
 closure type 0
     (GHC version 8.6.0.20180810 for x86_64_unknown_linux)
 }}}

 Strangely I can't reproduce the bug when I try to run the apparently
 problematic testcase with `--pattern KATs.XL-vec.inc`.

 Also note that a similar issue has been reported for GHC-8.2.2 on 32-mips
 at https://github.com/haskell-hvr/cryptohash-sha256/issues/3.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15544#comment:4>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler