Bug 1932066

This package has changed maintainer in Fedora. Reassigning to the new maintainer of this component.

Here is an example of the warnings about strncpy().

> In file included from /usr/include/string.h:519,
>                  from isup.c:32:
> In function 'strncpy',
>     inlined from 'isup_set_calling' at isup.c:2879:4,
>     inlined from 'isup_set_calling' at isup.c:2875:6:
> /usr/include/bits/string_fortified.h:91:10: warning: 'strncpy' specified bound 64 equals destination size [-Wstringop-truncation]
>    91 |   return __builtin___strncpy_chk (__dest, __src, __len, __bos (__dest));
>       |          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

> void isup_set_calling(struct isup_call *c, const char *calling, unsigned char calling_nai, unsigned char presentation_ind, unsigned char screening_ind)
> {
>         if ((calling && calling[0]) || presentation_ind == SS7_PRESENTATION_ADDR_NOT_AVAILABLE) {
>                 if (calling) {
>                         strncpy(c->calling_party_num, calling, sizeof(c->calling_party_num));
>                 } else {
>                         c->calling_party_num[0] = '\0';
>                 }
>                 c->calling_nai = calling_nai;
>                 c->presentation_ind = presentation_ind;
>                 c->screening_ind = screening_ind;
>         }
> }

There are a number of similar warnings about strncpy() for other fields. In this case, the intent seems to be that the argument “calling” is copied into the struct field with null-termination, truncating any overlong string. But strncpy does not guarantee null-termination. It is likely that the code should have been something like:

>                 if (calling) {
>                         strncpy(c->calling_party_num, calling, sizeof(c->calling_party_num));
>                         c->calling_party_num[sizeof(c->calling_party_num) - 1U] = '\0';
>                 } else {

-----

The other category of strncpy() warnings is like this:

> In function 'strncpy',
>     inlined from 'isup_event_iam' at isup.c:4335:2:
> /usr/include/bits/string_fortified.h:91:10: warning: 'strncpy' output may be truncated copying 50 bytes from a string of length 63 [-Wstringop-truncation]
>    91 |   return __builtin___strncpy_chk (__dest, __src, __len, __bos (__dest));
>       |          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

That comes from

> strncpy(e->iam.called_party_num, c->called_party_num, sizeof(e->iam.called_party_num));

where e is an ss7_event discriminated union (from libss7.h), with the ss7_event_iam-type member active. That has “char called_party_num[50];”, which is indeed shorter than the called_party_num in “c”, which is a struct isup_call.

What should be done in this case? Is truncation appropriate? It is already happening, only without null termination.

-----

I intend to consult the upstream developers on IRC (Freenode #asterisk-dev) to see what they think of these warnings.

Upstream issue: https://issues.asterisk.org/jira/browse/SS7-64

Upstream developers have been busy but plan to look at it when they can.