Re: [PATCH v4] iproute2: prevent memory leak

[Stephen Hemminger <stephen@networkplumber.org>]

On Fri, 17 Nov 2023 01:45:51 +0100
Andrea Claudi <aclaudi@redhat.com> wrote:

> On Thu, Nov 16, 2023 at 03:05:21PM -0800, Stephen Hemminger wrote:
> > On Thu, 16 Nov 2023 11:13:08 +0800
> > heminhong <heminhong@kylinos.cn> wrote:
> >   
> > > When the return value of rtnl_talk() is not less than 0,
> > > 'answer' will be allocated. The 'answer' should be free
> > > after using, otherwise it will cause memory leak.
> > > 
> > > Signed-off-by: heminhong <heminhong@kylinos.cn>  
> > 
> > I am skeptical, what is the code path through rtn_talk() that
> > returns non zero, and allocates answer.  If so, that should be fixed
> > there.
> > 
> > In current code, the returns are:
> > 	- sendmsg() fails
> > 	- recvmsg() fails
> > 	- truncated message
> > 	
> > The paths that set answer are returning 0  
> 
> IMHO the memory leak is in the same functions this is patching.
> For example, in ip/link_gre.c:122 we are effectively returning after
> having answer allocated correctly by rtnl_talk().
> 
> The confusion here stems from the fact we are jumping into the error
> path of rtnl_talk() after rtnl_talk() executed fine.
> 


So looks like a GRE etc bug introduced by the change to parsing.
Should add:

Fixes: a066cc6623e1 ("gre/gre6: Unify local/remote endpoint address parsing")
Cc: serhe.popovych@gmail.com