On Mon, Feb 26, 2024 at 10:48:14AM +0100, Philippe Mathieu-Daudé wrote:
> Hi,
>
> On 26/2/24 10:06, dinglimin wrote:
> > Signed-off-by: dinglimin <dingli...@cmss.chinamobile.com>
> > ---
> > semihosting/uaccess.c | 6 +++---
> > 1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/semihosting/uaccess.c b/semihosting/uaccess.c
> > index dc587d73bc..7788ead9b2 100644
> > --- a/semihosting/uaccess.c
> > +++ b/semihosting/uaccess.c
> > @@ -14,10 +14,10 @@
> > void *uaccess_lock_user(CPUArchState *env, target_ulong addr,
> > target_ulong len, bool copy)
> > {
> > - void *p = malloc(len);
> > + void *p = g_try_malloc(len);
> > if (p && copy) {
> > if (cpu_memory_rw_debug(env_cpu(env), addr, p, len, 0)) {
> > - free(p);
> > + g_free(p);
> > p = NULL;
> > }
> > }
>
> This seems dangerous, now all users of uaccess_lock_user() must
> use g_free(), in particular lock_user_string() which is used more
> than a hundred of times:
This is not true for many years now.
GLib is hardcoded to always use the system allocator, so g_malloc
can be freely mixed with free, and vica-verca.
Using 'g_free' is stylistically preferred, but not functionally
required.
>
> $ git grep -w lock_user_string | wc -l
> 116
>
> > @@ -87,5 +87,5 @@ void uaccess_unlock_user(CPUArchState *env, void *p,
> > if (len) {
> > cpu_memory_rw_debug(env_cpu(env), addr, p, len, 1);
> > }
> > - free(p);
> > + g_free(p);
> > }
>
>
With regards,
Daniel
--
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org -o- https://fstop138.berrange.com :|
|: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
