Re: [PATCH] genl: fix ‘attr_grp.next_len’ may be used uninitialized

Hello Denis

On 17/10/2023 в 10:08 -0500, Denis Kenzior wrote:
> Hi Max,
> 
> On 10/16/23 13:39, MaxLyubimov wrote:
> > make --no-print-directory all-am
> >    CC       ell/genl.lo
> > ell/genl.c: In function ‘parse_cmd_newfamily.part.0’:
> > ell/genl.c:1728:5: error: ‘attr_grp.next_len’ may be used
> > uninitialized
> > in this function [-Werror=maybe-uninitialized]
> >   1728 |  if (!NLA_OK(nla, attr->next_len))
> >        |     ^
> > ell/genl.c:268:22: note: ‘attr_grp.next_len’ was declared here
> >    268 |   struct l_genl_attr attr_grp;
> >        |                      ^~~~~~~~
> > ell/genl.c:1732:14: error: ‘attr_grp.next_data’ may be used
> > uninitialized in this function [-Werror=maybe-uninitialized]
> >   1732 |   *type = nla->nla_type & NLA_TYPE_MASK;
> >        |           ~~~^~~~~~~~~~
> > ell/genl.c:268:22: note: ‘attr_grp.next_data’ was declared here
> >    268 |   struct l_genl_attr attr_grp;
> >        |                      ^~~~~~~~
> > ell/genl.c:1728:5: error: ‘attr_op.next_len’ may be used
> > uninitialized
> > in this function [-Werror=maybe-uninitialized]
> >   1728 |  if (!NLA_OK(nla, attr->next_len))
> >        |     ^
> > ell/genl.c:240:22: note: ‘attr_op.next_len’ was declared here
> >    240 |   struct l_genl_attr attr_op;
> >        |                      ^~~~~~~
> > ell/genl.c:1732:14: error: ‘attr_op.next_data’ may be used
> > uninitialized in this function [-Werror=maybe-uninitialized]
> >   1732 |   *type = nla->nla_type & NLA_TYPE_MASK;
> >        |           ~~~^~~~~~~~~~
> > ell/genl.c:240:22: note: ‘attr_op.next_data’ was declared here
> >    240 |   struct l_genl_attr attr_op;
> >        |                      ^~~~~~~
> > cc1: all warnings being treated as errors
> > make[1]: *** [Makefile:4032: ell/genl.lo] Error 1
> > make: *** [Makefile:2348: all] Error 2
> > ---
> >   ell/genl.c | 6 ++++--
> >   1 file changed, 4 insertions(+), 2 deletions(-)
> > 
> 
> I added a space between 'Max' and 'Lyubimov' in your author
> tag.  Tell me if 
> this was incorrect.  If it was correct, can you please fix your git
> config?

Lyubimov Maxim. Okay, I'll fix the config.

> > diff --git a/ell/genl.c b/ell/genl.c
> > index def3790..33ecc4d 100644
> > --- a/ell/genl.c
> > +++ b/ell/genl.c
> > @@ -240,7 +240,8 @@ static void family_ops(struct
> > l_genl_family_info
> > *info,
> >   		struct l_genl_attr attr_op;
> >   		uint32_t id = 0, flags = 0;
> >   
> > -		l_genl_attr_recurse(attr, &attr_op);
> > +		if (!l_genl_attr_recurse(attr, &attr_op))
> > +			continue;
> >   
> >   		while (l_genl_attr_next(&attr_op, &type, &len, &data))
> > {
> 
> Something in your mail server is messing with the text format.  I had
> to 
> manually edit the patch so it would apply properly (by e.g. removing
> the newline 
> before this '{').  Please fix that since I can't do this for more
> than the 
> simplest patches.

I wanted to ask what git-send-email configuration should be used to
send patches correctly?

> >   			switch (type) {
> > @@ -269,7 +270,8 @@ static void family_mcast_groups(struct
> > l_genl_family_info *info,
> >   		const char *name = NULL;
> >   		uint32_t id = 0;
> >   
> > -		l_genl_attr_recurse(attr, &attr_grp);
> > +		if (!l_genl_attr_recurse(attr, &attr_grp))
> > +			continue;
> >   
> >   		while (l_genl_attr_next(&attr_grp, &type, &len, &data))
> > {
> >   			switch (type) {
> > 
> > 
> 
> Applied, thanks.
> 
> Regards,
> -Denis

Best regards,
Lyubimov Maxim