Re: [GHC] #13531: GHC fails with "Dynamic linker not initialised" when -j is on and trying to load nonexistent .so file

#13531: GHC fails with "Dynamic linker not initialised" when -j is on and trying to load nonexistent .so file -------------------------------------+------------------------------------- Reporter: nh2 | Owner: (none) Type: bug | Status: patch Priority: normal | Milestone: Component: Compiler | Version: 8.0.2 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Compile-time | Unknown/Multiple crash or panic | Test Case: Blocked By: | Blocking: Related Tickets: #13137, #9868, | Differential Rev(s): Phab:D5012 #10355, #13607, #10919 | Wiki Page: | -------------------------------------+------------------------------------- Comment (by Ben Gamari ): In [changeset:"4fc6524a2a4a0003495a96c8b84783286f65c198/ghc" 4fc6524/ghc]: {{{ #!CommitTicketReference repository="ghc" revision="4fc6524a2a4a0003495a96c8b84783286f65c198" Stop the linker panic If we fail to initialize the liker properly, we still set the `initLinkerDone`. In fact we even set that prior to actually initializing the linker. However if the linker initialization fails, we the `Done` state is still true. As such we run into the `Dynamic Linker not initialised` error. Which while technically corret is confusing as it pulls the attation away from the real issue. This change puts the Done state into an MVar, and as such ensureing that all parallel access needs to wait for the linker to be actually initialized, or try to re-initilize if it fails. Reviewers: bgamari, RyanGlScott, simonmar, hvr Reviewed By: bgamari Subscribers: rwbarton, thomie, carter GHC Trac Issues: #9868, #10355, #13137, #13607, #13531 Differential Revision: https://phabricator.haskell.org/D5012 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13531#comment:6 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler