pf: Fix incorrect fragment handling

A sequence of overlapping IPv4 fragments could crash the kernel in
pf due to an assertion.

Reported by:	Alexander Bluhm
Obtained from:	OpenBSD
MFC after:	3 days
Sponsored by:	Rubicon Communications, LLC ("Netgate")
This commit is contained in:
Kristof Provost 2021-02-25 08:07:36 +01:00
parent d2eb5dbd7d
commit 5f1b1f184b

View File

@ -549,6 +549,7 @@ pf_fillup_fragment(struct pf_fragment_cmp *key, struct pf_frent *frent,
struct pf_frent *after, *next, *prev;
struct pf_fragment *frag;
uint16_t total;
int old_index, new_index;
PF_FRAG_ASSERT();
@ -660,8 +661,30 @@ pf_fillup_fragment(struct pf_fragment_cmp *key, struct pf_frent *frent,
DPFPRINTF(("adjust overlap %d\n", aftercut));
if (aftercut < after->fe_len) {
m_adj(after->fe_m, aftercut);
old_index = pf_frent_index(after);
after->fe_off += aftercut;
after->fe_len -= aftercut;
new_index = pf_frent_index(after);
if (old_index != new_index) {
DPFPRINTF(("frag index %d, new %d",
old_index, new_index));
/* Fragment switched queue as fe_off changed */
after->fe_off -= aftercut;
after->fe_len += aftercut;
/* Remove restored fragment from old queue */
pf_frent_remove(frag, after);
after->fe_off += aftercut;
after->fe_len -= aftercut;
/* Insert into correct queue */
if (pf_frent_insert(frag, after, prev)) {
DPFPRINTF(
("fragment requeue limit exceeded"));
m_freem(after->fe_m);
uma_zfree(V_pf_frent_z, after);
/* There is not way to recover */
goto bad_fragment;
}
}
break;
}