[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:                    |
-------------------------------------+-------------------------------------

Comment (by bgamari):

 The previous debugging session ended up looking a bit messy. Starting
 afresh, this time with `+RTS -DS`. Things are looking much clearer now.

 We get the following segfault:
 {{{
 Thread 2 received signal SIGSEGV, Segmentation fault.
 0x00000000007c8370 in stg_IND_STATIC_info () at
 rts/StgMiscClosures.cmm:270
 270     {
 >>> bt
 #0  0x00000000007c8370 in stg_IND_STATIC_info () at
 rts/StgMiscClosures.cmm:270
 #1  0x00000000006ae1f0 in
 bytestringzm0zi10zi8zi2_DataziByteStringziLazzy_fromChunkszugo_info () at
 libraries/bytestring/Data/ByteString/Lazy.hs:267
 #2  0x00000000007c5388 in ?? () at rts/Updates.cmm:31
 #3  0x00000000006adc50 in
 bytestringzm0zi10zi8zi2_DataziByteStringziLazzy_toChunkszugo_info () at
 libraries/bytestring/Data/ByteString/Lazy.hs:271
 #4  0x00000000007c5388 in ?? () at rts/Updates.cmm:31
 #5  0x0000000000624d20 in s7zn_info ()
 #6  0x0000000000000000 in ?? ()
 }}}
 where `$rbx` points to memory cleared by the RTS:
 {{{
 >>> x/8a $rbx
 0x4200365960:   0xaaaaaaaaaaaaaaaa      0xaaaaaaaaaaaaaaaa
 0x4200365970:   0xaaaaaaaaaaaaaaaa      0xaaaaaaaaaaaaaaaa
 0x4200365980:   0xaaaaaaaaaaaaaaaa      0xaaaaaaaaaaaaaaaa
 0x4200365990:   0xaaaaaaaaaaaaaaaa      0xaaaaaaaaaaaaaaaa
 }}}
 Tracing this back it looks like this once resided in a nursery that has
 since been freed. It sounds like something is being freed too soon
 (perhaps something like #14346/#15260). Looking at the package
 implementation it looks like there are a few `withForeignPtr` uses. This
 is quite suspicious.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15544#comment:7>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler