On Tue, Dec 29, 2009 at 05:38:17PM +0200, Michael S. Tsirkin wrote:
> On Mon, Dec 28, 2009 at 04:49:00PM +0100, Aurelien Jarno wrote:
> > realloc(ptr, 0) is always allowed by the standard. The return value is
> > either NULL or a pointer that can be freed with free().
> >
> > Allow usage of qemu_realloc(ptr, 0), and return NULL in that case, as
> > free(NULL) should always be a nop.
> >
> > This fixes -kernel with stripped kernels.
> >
> > Signed-off-by: Aurelien Jarno <aurel...@aurel32.net>
>
> So this is superceded by
> loader: don't call realloc(non_null, 0)
> right?
Oh, you said as much already. Sorry about the noise.
> > ---
> > qemu-malloc.c | 6 +++---
> > 1 files changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/qemu-malloc.c b/qemu-malloc.c
> > index 5d9e34d..cf6a1f1 100644
> > --- a/qemu-malloc.c
> > +++ b/qemu-malloc.c
> > @@ -63,10 +63,10 @@ void *qemu_realloc(void *ptr, size_t size)
> > {
> > if (size) {
> > return oom_check(realloc(ptr, size));
> > - } else if (allow_zero_malloc()) {
> > - return oom_check(realloc(ptr, size ? size : 1));
> > + } else if (ptr) {
> > + qemu_free(ptr);
> > }
> > - abort();
> > + return NULL;
> > }
> >
> > void *qemu_mallocz(size_t size)
> > --
> > 1.6.5.3
> >
> >
>
