Memory leak in FFI callback: GHC 6.6

[SevenThunders]

I'm not sure if this is a bug, but it sure seems that way to me.  When using the "wrapper" technique for creating a Haskell callback function it seems that freeHaskellFunPtr is not properly freeing the resource.

Here is an example program
<leaky.hs>
module Main where
import Foreign
import Foreign.Ptr
import Foreign.Storable

-- | A convenient type synonym for storable arrays
type Darr = Ptr (Double)
-- | Function type for mapping doubles to doubles
type Dfunc = Double -> Double
-- | A convenient type synonym for  monad containing storable arrays
type IODarr = IO (Darr)


foreign import ccall "wrapper" mkDfunc :: Dfunc -> IO (FunPtr Dfunc)
foreign import ccall "cleaky.h cfunc" cfunc :: (FunPtr Dfunc) -> IO ( Double )


dadd :: Dfunc
dadd x = x + 1.0

getleaky :: Dfunc -> IO ()
getleaky cf = do
        pcf <- mkDfunc cf
        -- pd <- cfunc pcf
        print pcf
        freeHaskellFunPtr pcf


main = sequence_ [getleaky dadd | q <- [1..500000]]


It's compiled with
ghc -fglasgow-exts -fffi -prof -auto -I. --make  leaky.hs cleaky.o

A possible cleaky.c and cleaky.h are
#include <math.h>
#include <stdio.h>
#include "cleaky.h"


double state = 1.0 ;

double cfunc(DFptr fptr)
{
// printf("In cfunc, fptr: %p\n", fptr) ;
state = (*fptr) (state) ;
// printf("state: %g\n", state) ;
return(state) ;
}

typedef double (*DFptr) (double) ;
double cfunc(DFptr) ;

though they are not used.

leaky.exe grows the heap linearly without bound on windows XP 64.  The profiler never reports this correctly either.
calling
leaky +RTS -hc -RTS
does not show the increase in the memory usage.

[Simon Peyton-Jones]

Dear SevenThunders (if that is how you like to addressed)

Thanks for extracting a small program that exhibits the leak; that is
really helpful.  We'll look into it. Would you like to create a Trac bug
and attach your files?

Simon

| -----Original Message-----
| From: glasgow-haskell-users-bounces@...
[mailto:glasgow-haskell-users-bounces@...]
| On Behalf Of SevenThunders
| Sent: 03 November 2006 02:54
| To: glasgow-haskell-users@...
| Subject: Memory leak in FFI callback: GHC 6.6
|
|
| I'm not sure if this is a bug, but it sure seems that way to me.  When
using
| the "wrapper" technique for creating a Haskell callback function it
seems
| that freeHaskellFunPtr is not properly freeing the resource.
|
| Here is an example program
| <leaky.hs>
| module Main where
| import Foreign
| import Foreign.Ptr
| import Foreign.Storable
|
| -- | A convenient type synonym for storable arrays
| type Darr = Ptr (Double)
| -- | Function type for mapping doubles to doubles
| type Dfunc = Double -> Double
| -- | A convenient type synonym for  monad containing storable arrays
| type IODarr = IO (Darr)
|
|
| foreign import ccall "wrapper" mkDfunc :: Dfunc -> IO (FunPtr Dfunc)
| foreign import ccall "cleaky.h cfunc" cfunc :: (FunPtr Dfunc) -> IO (
Double
| )
|
|
| dadd :: Dfunc
| dadd x = x + 1.0
|
| getleaky :: Dfunc -> IO ()
| getleaky cf = do
| pcf <- mkDfunc cf
| -- pd <- cfunc pcf
| print pcf
| freeHaskellFunPtr pcf
|
|
| main = sequence_ [getleaky dadd | q <- [1..500000]]
|
|
| It's compiled with
| ghc -fglasgow-exts -fffi -prof -auto -I. --make  leaky.hs cleaky.o
|
| A possible cleaky.c and cleaky.h are
| #include <math.h>
| #include <stdio.h>
| #include "cleaky.h"
|
|
| double state = 1.0 ;
|
| double cfunc(DFptr fptr)
| {
| // printf("In cfunc, fptr: %p\n", fptr) ;
| state = (*fptr) (state) ;
| // printf("state: %g\n", state) ;
| return(state) ;
| }
|
| typedef double (*DFptr) (double) ;
| double cfunc(DFptr) ;
|
| though they are not used.
|
| leaky.exe grows the heap linearly without bound on windows XP 64.  The
| profiler never reports this correctly either.
| calling
| leaky +RTS -hc -RTS
| does not show the increase in the memory usage.
|
| --
| View this message in context:
http://www.nabble.com/Memory-leak-in-FFI-callback%3A-GHC-6.6-
| tf2565446.html#a7150757
| Sent from the Haskell - Glasgow-haskell-users mailing list archive at
Nabble.com.
|
| _______________________________________________
| Glasgow-haskell-users mailing list
| Glasgow-haskell-users@...
| http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
_______________________________________________
Glasgow-haskell-users mailing list
Glasgow-haskell-users@...
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users

[SevenThunders]

Simon Peyton-Jones wrote:

Dear SevenThunders (if that is how you like to addressed)

Thanks for extracting a small program that exhibits the leak; that is
really helpful.  We'll look into it. Would you like to create a Trac bug
and attach your files?

Simon

You can call me Matthew.  The bug, if that's what it is,  is now ticket number 985.  By the way, why can't I get the callback to actually work in this example?  If I for example uncomment the line
        -- pd <- cfunc pcf
 
and say print out the value of pd,  I get garbage values.

 Should I have used CDouble?  I think for windows HsDouble = double in C land.

[Simon Peyton-Jones]

Matthew

Thanks for submitting the Trac bug.  I don't know what the answer to
your qn below is, but maybe someone else on ghc-users does.  Meanwhile,
I added the qn to the bug report.  

Simon

| -----Original Message-----
| From: glasgow-haskell-users-bounces@...
[mailto:glasgow-haskell-users-bounces@...]
| On Behalf Of SevenThunders
| Sent: 03 November 2006 19:04
| To: glasgow-haskell-users@...
| Subject: RE: Memory leak in FFI callback: GHC 6.6
|
|
|
| Simon Peyton-Jones wrote:
| >
| > Dear SevenThunders (if that is how you like to addressed)
| >
| > Thanks for extracting a small program that exhibits the leak; that
is
| > really helpful.  We'll look into it. Would you like to create a Trac
bug
| > and attach your files?
| >
| > Simon
| >
| >
| >
|
| You can call me Matthew.  The bug, if that's what it is,  is now
ticket
| number 985.  By the way, why can't I get the callback to actually work
in
| this example?  If I for example uncomment the line
| -- pd <- cfunc pcf
|
| and say print out the value of pd,  I get garbage values.
|
|  Should I have used CDouble?  I think for windows HsDouble = double in
C
| land.
| --
| View this message in context:
http://www.nabble.com/Memory-leak-in-FFI-callback%3A-GHC-6.6-
| tf2565446.html#a7165550
| Sent from the Haskell - Glasgow-haskell-users mailing list archive at
Nabble.com.
|
| _______________________________________________
| Glasgow-haskell-users mailing list
| Glasgow-haskell-users@...
| http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
_______________________________________________
Glasgow-haskell-users mailing list
Glasgow-haskell-users@...
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users