Re: [PATCH v2] apparmor/file: Removing unnecessary initial values for variable pointers

[Serge E. Hallyn]

On Thu, Sep 28, 2023 at 10:36:16AM -0700, John Johansen wrote:
> On 9/18/23 18:56, Li kunyu wrote:
> > These variable pointers are assigned during use and do not need to be
> > initialized for assignment.
> > 
> > Signed-off-by: Li kunyu <kunyu@xxxxxxxxxxxx>
> > ---
> >   v2: Fix timestamp issues
> > 
> >   security/apparmor/file.c | 4 ++--
> >   1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/security/apparmor/file.c b/security/apparmor/file.c
> > index 698b124e649f..12eafdf18fc0 100644
> > --- a/security/apparmor/file.c
> > +++ b/security/apparmor/file.c
> > @@ -264,7 +264,7 @@ int aa_path_perm(const char *op, struct aa_label *label,
> >   {
> >   	struct aa_perms perms = {};
> >   	struct aa_profile *profile;
> > -	char *buffer = NULL;
> > +	char *buffer;
> 
> this is okay
> 
> >   	int error;
> >   	flags |= PATH_DELEGATE_DELETED | (S_ISDIR(cond->mode) ? PATH_IS_DIR :
> > @@ -412,7 +412,7 @@ int aa_path_link(struct aa_label *label, struct dentry *old_dentry,
> >   		d_backing_inode(old_dentry)->i_uid,
> >   		d_backing_inode(old_dentry)->i_mode
> >   	};
> > -	char *buffer = NULL, *buffer2 = NULL;
> > +	char *buffer, *buffer2;
> 
> this can cause an oops if buffer2 allocation fails. There are a couple of ways I can
> see to fix this, do you want to take a crack at it.
> 
> 
> >   	struct aa_profile *profile;
> >   	int error;

I don't whether this kind of thing has become in vogue, but while indeed
the first case is okay right now, it becomes more likely that a future patch
to this function will inadvertently goto aa_put_buffer(buffer) before the
aa_get_buffer call.  I would not have NACKed an original version of this fn
without the = NULL, but I'm not in favor of dropping it.

-serge