Re: Possible password left in buffer

On Thu, 15 Apr 2010 01:06:55 -0400
Scott Lovenberg <[email protected]> wrote:

> Just wondering if SAFE_FREE(char*) has to be called on temp_val in
> open_cred_file() of mount.cifs.c.  Unless I'm missing something, a parsed
> password will be left in temp_val because it never gets zeroed out.
> 

Yeah, looks like something that should be fixed. In fact, that whole
function needs some overhaul -- the magic 4096/4086/4087 numbers in it
are pretty yucky and make it hard to tell what's going on. That should
use a #define'd constant. Also, the user= field should probably be
fixed so that it uses parse_username() too.

Care to send a patch? If not, I'll plan to fix that sometime before the
next release.

Thanks,
-- 
Jeff Layton <[email protected]>