#14912: UnsafeReenter test fails with threaded1 and threaded2 -----------------------------------+-------------------------------------- Reporter: alpmestan | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Runtime System | Version: 8.5 Resolution: | Keywords: Operating System: Linux | Architecture: x86_64 (amd64) Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: #14743 | Differential Rev(s): Wiki Page: | -----------------------------------+-------------------------------------- Changes (by osa1): * cc: simonmar (added) * related: => #14743 Comment: This still happens with GHC HEAD, but only with threaded runtime. {{{ $ ghc-stage2 UnsafeReenter.hs UnsafeReenterC.c -fforce-recomp -threaded [1 of 1] Compiling Main ( UnsafeReenter.hs, UnsafeReenter.o ) Linking UnsafeReenter ... $ ./UnsafeReenter In Haskell in C ^C^C }}} Interestingly, if I add `-N2` the program terminates, but not with the error message as expected! {{{ $ ./UnsafeReenter +RTS -N2 In Haskell in C Back in Haskell Finished }}} I checked Haskell 2010 and this program actually has an undefined behavior: (section 8.4.3)
safe call is less efficient, but guarantees to leave the Haskell system in a state that allows callbacks from the external code. In contrast, an unsafe call, while carrying less overhead, must not trigger a callback into the Haskell system. If it does, the system behaviour is undefined.
So both outputs above are actually fine. But perhaps we still want to catch Haskell calls from unsafe FFI calls in the current implementation? Simon, any ideas? -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14912#comment:2 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler