#3618: memory-leak detector in +RTS -DS fails to track allocations in
constructors
-------------------------------+--------------------------------------------
Reporter: guest | Type: bug
Status: closed | Priority: normal
Milestone: _|_ | Component: Runtime System
Version: 6.12.1 RC1 | Resolution: fixed
Keywords: | Testcase:
Blockedby: | Difficulty: Unknown
Os: Linux | Blocking:
Architecture: x86_64 (amd64) | Failure: None/Unknown
-------------------------------+--------------------------------------------
Changes (by Lemming):
* cc: g...@… (added)
Comment:
I'd be interested whether the poster of this ticket could solve his
problem.
I use Linux on x86_32.
I also have a bigger project using LLVM, that crashed.
It crashed with GHC-6.10.4 with static linking
and with GHC-6.12.3 with static and dynamic linking.
I am not able to make the crashing program small enough,
since at some point it does not crash anymore when I reduce its size.
Following http://hackage.haskell.org/trac/ghc/wiki/Debugging/CompiledCode,
I used "+RTS -DS" runtime options
as well as no "-threaded" (it always let me jump into handle_tick)
and "-debug -dcore-lint" compiler options.
In GDB I found that the final crash happens in deRefStablePtr
while accessing an address like 0xaaaaaaaa.
(I actually use StablePtrs in my code.)
The GHC-Wiki page says,
that this is the value a deallocated piece of memory is filled with.
I also got the messages:
{{{
Warning: Freeing non-allocated memory at 0x95b2438
Warning: Freeing non-allocated memory at 0x95b1410
Warning: Freeing non-allocated memory at 0x95b1008
Warning: Freeing non-allocated memory at 0x95b3440
Warning: Freeing non-allocated memory at 0x95b4440
Warning: 36 bytes at 0x95bca10 still allocated at shutdown
Warning: 128 bytes at 0x95bc978 still allocated at shutdown
...
}}}
and thought that is better to fix them first,
before continuing with my deRefStablePtr crash,
because the two problems might be related.
However I have seen that the "freeing non-allocated memory" warnings occur
in even the simplest programs.
Now I read here that the warnings are probably pointless.
I'm still uncertain.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/3618#comment:7>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
