Re: PR31145, potential memory leak in binutils/ld

[Alan Modra <amodra@gmail.com>]

Revert most of this patch, it isn't correct to free the BFD_IN_MEMORY
iostream in io_reinit.

	PR 31145
	* format.c (io_reinit): Revert last change.  Comment.
	* opncls.c (_bfd_delete_bfd): Likewise.

diff --git a/bfd/format.c b/bfd/format.c
index 31aeb52861a..0790a3a2ee5 100644
--- a/bfd/format.c
+++ b/bfd/format.c
@@ -163,16 +163,15 @@ io_reinit (bfd *abfd, struct bfd_preserve *preserve)
   if (abfd->iovec != preserve->iovec)
     {
       /* Handle file backed to in-memory transition.  bfd_cache_close
-	 won't do anything unless abfd->iovec is the cache_iovec.  */
+	 won't do anything unless abfd->iovec is the cache_iovec.
+	 Don't be tempted to call iovec->bclose here.  We don't want
+	 to call memory_bclose, which would free the bim.  The bim
+	 must be kept if bfd_check_format_matches is going to decide
+	 later that the PE format needing it is in fact the correct
+	 target match.  */
       bfd_cache_close (abfd);
       abfd->iovec = preserve->iovec;
-
-      if (abfd->iostream != preserve->iostream)
-	{
-	  if ((abfd->flags & BFD_IN_MEMORY) != 0)
-	    free (abfd->iostream);
-	  abfd->iostream = preserve->iostream;
-	}
+      abfd->iostream = preserve->iostream;
 
       /* Handle in-memory to file backed transition.  */
       if ((abfd->flags & BFD_CLOSED_BY_CACHE) != 0
diff --git a/bfd/opncls.c b/bfd/opncls.c
index e7b39593608..32c95d6ba52 100644
--- a/bfd/opncls.c
+++ b/bfd/opncls.c
@@ -176,8 +176,6 @@ _bfd_delete_bfd (bfd *abfd)
   else
     free ((char *) bfd_get_filename (abfd));
 
-  if ((abfd->flags & BFD_IN_MEMORY) != 0)
-    free (abfd->iostream);
   free (abfd->arelt_data);
   free (abfd);
 }

-- 
Alan Modra
Australia Development Lab, IBM