diff options
| author | Donald Sharp <sharpd@nvidia.com> | 2023-04-21 07:38:11 -0400 |
|---|---|---|
| committer | Donald Sharp <sharpd@nvidia.com> | 2023-04-21 07:38:11 -0400 |
| commit | 35801e6234fec8c62094e4c9cc8400b5942d013e (patch) | |
| tree | 15921c087044b4849835b2d2bb26de0cea5570ea /pimd/pim6_mld.c | |
| parent | db865bf6ed708e087eaef18d89f458044582cac4 (diff) | |
pimd: Tell coverity what is really going on
Fix a code path that coverity has decided a variable
is NULL when it never can be.
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Diffstat (limited to 'pimd/pim6_mld.c')
| -rw-r--r-- | pimd/pim6_mld.c | 13 |
1 files changed, 13 insertions, 0 deletions
diff --git a/pimd/pim6_mld.c b/pimd/pim6_mld.c index e74707b522..74cd729d44 100644 --- a/pimd/pim6_mld.c +++ b/pimd/pim6_mld.c @@ -2471,6 +2471,19 @@ static void gm_show_if_vrf(struct vty *vty, struct vrf *vrf, const char *ifname, if (js) { js_if = json_object_new_object(); + /* + * If we have js as true and detail as false + * and if Coverity thinks that js_if is NULL + * because of a failed call to new then + * when we call gm_show_if_one below + * the tt can be deref'ed and as such + * FRR will crash. But since we know + * that json_object_new_object never fails + * then let's tell Coverity that this assumption + * is true. I'm not worried about fast path + * here at all. + */ + assert(js_if); json_object_object_add(js_vrf, ifp->name, js_if); } |
