This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [patch] BFD ieee.c fix
- To: cgd at sibyte dot com
- Subject: Re: [patch] BFD ieee.c fix
- From: Nick Clifton <nickc at redhat dot com>
- Date: Wed, 18 Oct 2000 12:22:44 -0700
- CC: binutils at sourceware dot cygnus dot com
Hi Guys,
: Alan Modra <alan@linuxcare.com.au> writes:
: > I recommend this patch _not_ be applied, nor any future patch mentioning
: > the word trivial associated with bfd_release!
:
: Heh. Based on your comment, and the semantics of bfd_release (of
: which I wasn't aware) i'm not convinced that there's really an error
: in that function at all.
Well hold on there, there may still be a memory leak. In the function
we are looking at, the first instance of "goto got_wrong_format_error"
does not call bfd_release(), but the second one does. (And the memory
being released was allocated at the start of the function, before the
first goto). This implies that either the call to bfd_release by the
second goto is wrong, or that it is OK and the first goto also ought
to call bfd_release.
The only function call between the bfd_alloc that allocates the memory
that is the argument to the bfd_release and the first goto is a call
to bfd_read. bfd_read does not allocate any memory, nor do any of the
functions that it calls, (I checked). read_id does allocate memory,
but that is called after the first goto, not before it, so there is no
problem there.
Thus I actually think that that patch is OK. I am not going to check
it in yet, in case there are any more comments, but if no has any more
objections then it will go in.
Cheers
Nick