haskell ports failing with signal 10 in poudriere:

[Eitan Adler]

On 13 June 2018 at 11:01, Gleb Popov <6yearold at gmail.com> wrote:
>
>
> On Wed, Jun 13, 2018 at 8:11 PM, Eitan Adler <lists at eitanadler.com> wrote:
>>
>> On 13 June 2018 at 07:43, Gleb Popov <6yearold at gmail.com> wrote:
>> >
>> >
>> > On Wed, Jun 13, 2018 at 3:26 PM, Eitan Adler <lists at eitanadler.com>
>> > wrote:
>> >>
>> >> On 12 June 2018 at 09:50, Gleb Popov <6yearold at gmail.com> wrote:
>> >> >
>> >> >
>> >> > On Tue, Jun 12, 2018 at 7:41 PM, Eitan Adler <lists at eitanadler.com>
>> >> > wrote:
>> >> >>
>> >> >> On 12 June 2018 at 03:54, Gleb Popov <6yearold at gmail.com> wrote:
>> >> >> >
>> >> >> >
>> >> >> > On Mon, Jun 11, 2018 at 6:53 PM, Eitan Adler
>> >> >> > <lists at eitanadler.com>
>> >> >> > wrote:
>> >> >> >>
>> >> >> >> On 11 June 2018 at 00:28, Gleb Popov <6yearold at gmail.com> wrote:
>> >> >> >> >
>> >> >> >> >
>> >> >> >> > On Mon, Jun 11, 2018 at 2:52 AM, Eitan Adler
>> >> >> >> > <lists at eitanadler.com>
>> >> >> >> > wrote:
>> >> >> >> >>
>> >> >> >> >> On 10 June 2018 at 12:59, Gleb Popov <6yearold at gmail.com>
>> >> >> >> >> wrote:
>> >> >> >> >> >
>> >> >> >> >> >
>> >> >> >> >> > On Sun, Jun 10, 2018 at 10:30 PM, Eitan Adler
>> >> >> >> >> > <lists at eitanadler.com>
>> >> >> >> >> > wrote:
>> >> >> >> >> >>
>> >> >> >> >> >> On 10 June 2018 at 02:03, Gleb Popov <6yearold at gmail.com>
>> >> >> >> >> >> wrote:
>> >> >> >> >> >> >
>> >> >> >> >> >> >
>> >> >> >> >> >> > On Sun, Jun 10, 2018 at 12:00 PM, Eitan Adler
>> >> >> >> >> >> > <lists at eitanadler.com>
>> >> >> >> >> >> > wrote:
>> >> >> >> >> >> >>
>> >> >> >> >> >> >> On 10 June 2018 at 01:30, Gleb Popov
>> >> >> >> >> >> >> <6yearold at gmail.com>
>> >> >> >> >> >> >> wrote:
>> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >
>> >> >> >> >> >> >> > On Sat, Jun 9, 2018 at 8:06 AM, Eitan Adler
>> >> >> >> >> >> >> > <lists at eitanadler.com>
>> >> >> >> >> >> >> > wrote:
>> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >> Any idea what may be wrong?
>> >> >> >> >> >> >> > Can't reproduce. What OS revision are you running?
>> >> >> >> >> >> >>
>> >> >> >> >> >> >> FreeBSD fasteagle 12.0-CURRENT FreeBSD 12.0-CURRENT #12
>> >> >> >> >> >> >> r334863:
>> >> >> >> >> >> >> Fri
>> >> >> >> >> >> >> Jun  8 23:04:30 UTC 2018
>> >> >> >> >> >> >>
>> >> >> >> >> >> >> eax at fasteagle:/srv/obj/fbsd/usr/src/amd64.amd64/sys/EADLER
>> >> >> >> >> >> >> amd64
>> >> >> >> >> >> >
>> >> >> >> >> >> >
>> >> >> >> >> >> > Mine's a bit older. Let me rebuild everything first.
>> >> >> >> >> >>
>> >> >> >> >> >> Not well researched but perhaps related to
>> >> >> >> >> >> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=214864 ?
>> >> >> >> >> >
>> >> >> >> >> >
>> >> >> >> >> > This should be fixed by
>> >> >> >> >> > https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=226872
>> >> >> >> >>
>> >> >> >> >> This as in the SIGBUS or this as in the uninformed guess as to
>> >> >> >> >> the
>> >> >> >> >> cause I referenced?
>> >> >> >> >
>> >> >> >> >
>> >> >> >> > My English parser got a bit lost in this sentence. This is
>> >> >> >> > regarding
>> >> >> >> > switching ld to lld. GHC had problems with lld, but these were
>> >> >> >> > fixed
>> >> >> >> > in
>> >> >> >> > the
>> >> >> >> > PR I mentioned.
>> >> >> >>
>> >> >> >> So my issues are unrelated. Thanks!
>> >> >> >
>> >> >> >
>> >> >> > I've rebuilt everything from scratch using r334922 of base and
>> >> >> > hs-random
>> >> >> > builds fine for me. Do you still have the problem?
>> >> >>
>> >> >> Yes.
>> >> >
>> >> >
>> >> > Hum, maybe you can send me the ghc package you are using?
>> >>
>> >> freebsd.org/~eadler/files/ghc-8.4.2_1.txz
>> >
>> >
>> > Ok, your package indeed produces broken executables. Even simple Hello
>> > world
>> > in haskell crashes with bus error. Debugging it gives
>> >
>> > Program received signal SIGBUS, Bus error.
>> > 0x000000000029e60a in new_gc_thread (n=<error reading variable: Cannot
>> > access memory at address 0x0>, t=<optimized out>) at rts/sm/GC.c:879
>> > 879     rts/sm/GC.c: No such file or directory.
>> > (gdb) bt
>> > #0  0x000000000029e60a in new_gc_thread (n=<error reading variable:
>> > Cannot
>> > access memory at address 0x0>, t=<optimized out>) at rts/sm/GC.c:879
>> > #1  initGcThreads (from=<optimized out>, to=<optimized out>) at
>> > rts/sm/GC.c:918
>> > #2  0x0000000000297040 in hs_init_ghc (argc=0x7fffffffe544,
>> > argv=<optimized
>> > out>, rts_config=...) at rts/RtsStartup.c:242
>> > #3  0x000000000028d15d in hs_main (argc=-7392, argv=0x8,
>> > main_closure=<optimized out>, rts_config=...) at rts/RtsMain.c:57
>> > #4  0x00000000002174a5 in main ()
>> >
>> > However, I am unable to reproduce the issue on my side. You could try
>> > building ghc WITH_DEBUG=1 so we can try figure out what's wrong with
>> > RTS.
>> >
>> > I also noted that OSVERSION of your package is 68, which is even newer
>> > that
>> > my recently built 67. Just to rule out base version factor, did you also
>> > have this problem on older version?
>>
>> Yes. I keep my system fairly up to date, so its a bit of a moving
>> target, but its been this way for two weeks or so now.
>
>
> Stupid me! This has to be this issue:
> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=226059
>
> You are probably compiling with CPYTUPE=native ?

Its an explicit CPUTYPE rather than native. Easy enough to fix though.

Exact file here:
https://github.com/unixdotfiles/global/blob/master/roles/develop/files/poudriere_default_make.conf#L61

-- 
Eitan Adler