* [dpdk-dev] [PATCH] ip_frag : Fix double-free of chained mbufs
@ 2016-04-21 17:28 cychong
2016-04-22 9:54 ` Ananyev, Konstantin
0 siblings, 1 reply; 3+ messages in thread
From: cychong @ 2016-04-21 17:28 UTC (permalink / raw)
To: dev; +Cc: cychong
If any fragment hole is found in ipv4_frag_reassemble() and ipv6_frag_reassemble(),
whole ip_frag_pkt mbufs are moved to death-row. Any mbufs already chained to
another mbuf are freed multiple times as there are still in ip_frag_pkt array.
Signed-off-by: cychong <cychong@gmail.com>
---
lib/librte_ip_frag/ip_frag_common.h | 4 ++--
lib/librte_ip_frag/rte_ipv4_reassembly.c | 8 +++++++-
lib/librte_ip_frag/rte_ipv6_reassembly.c | 8 +++++++-
3 files changed, 16 insertions(+), 4 deletions(-)
diff --git a/lib/librte_ip_frag/ip_frag_common.h b/lib/librte_ip_frag/ip_frag_common.h
index cde6ed4..5542a38 100644
--- a/lib/librte_ip_frag/ip_frag_common.h
+++ b/lib/librte_ip_frag/ip_frag_common.h
@@ -76,8 +76,8 @@ struct ip_frag_pkt * ip_frag_lookup(struct rte_ip_frag_tbl *tbl,
struct ip_frag_pkt **free, struct ip_frag_pkt **stale);
/* these functions need to be declared here as ip_frag_process relies on them */
-struct rte_mbuf * ipv4_frag_reassemble(const struct ip_frag_pkt *fp);
-struct rte_mbuf * ipv6_frag_reassemble(const struct ip_frag_pkt *fp);
+struct rte_mbuf *ipv4_frag_reassemble(struct ip_frag_pkt *fp);
+struct rte_mbuf *ipv6_frag_reassemble(struct ip_frag_pkt *fp);
diff --git a/lib/librte_ip_frag/rte_ipv4_reassembly.c b/lib/librte_ip_frag/rte_ipv4_reassembly.c
index 26d07f9..e084ca5 100644
--- a/lib/librte_ip_frag/rte_ipv4_reassembly.c
+++ b/lib/librte_ip_frag/rte_ipv4_reassembly.c
@@ -41,11 +41,12 @@
* Reassemble fragments into one packet.
*/
struct rte_mbuf *
-ipv4_frag_reassemble(const struct ip_frag_pkt *fp)
+ipv4_frag_reassemble(struct ip_frag_pkt *fp)
{
struct ipv4_hdr *ip_hdr;
struct rte_mbuf *m, *prev;
uint32_t i, n, ofs, first_len;
+ uint32_t curr_idx = 0;
first_len = fp->frags[IP_FIRST_FRAG_IDX].len;
n = fp->last_idx - 1;
@@ -53,6 +54,7 @@ ipv4_frag_reassemble(const struct ip_frag_pkt *fp)
/*start from the last fragment. */
m = fp->frags[IP_LAST_FRAG_IDX].mb;
ofs = fp->frags[IP_LAST_FRAG_IDX].ofs;
+ curr_idx = IP_LAST_FRAG_IDX;
while (ofs != first_len) {
@@ -67,6 +69,10 @@ ipv4_frag_reassemble(const struct ip_frag_pkt *fp)
rte_pktmbuf_adj(m, (uint16_t)(m->l2_len + m->l3_len));
rte_pktmbuf_chain(fp->frags[i].mb, m);
+ /* this mbuf should not be accessed directly */
+ fp->frags[curr_idx].mb = NULL;
+ curr_idx = i;
+
/* update our last fragment and offset. */
m = fp->frags[i].mb;
ofs = fp->frags[i].ofs;
diff --git a/lib/librte_ip_frag/rte_ipv6_reassembly.c b/lib/librte_ip_frag/rte_ipv6_reassembly.c
index d29cb1d..21a5ef5 100644
--- a/lib/librte_ip_frag/rte_ipv6_reassembly.c
+++ b/lib/librte_ip_frag/rte_ipv6_reassembly.c
@@ -59,13 +59,14 @@ ip_frag_memmove(char *dst, char *src, int len)
* Reassemble fragments into one packet.
*/
struct rte_mbuf *
-ipv6_frag_reassemble(const struct ip_frag_pkt *fp)
+ipv6_frag_reassemble(struct ip_frag_pkt *fp)
{
struct ipv6_hdr *ip_hdr;
struct ipv6_extension_fragment *frag_hdr;
struct rte_mbuf *m, *prev;
uint32_t i, n, ofs, first_len;
uint32_t last_len, move_len, payload_len;
+ uint32_t curr_idx = 0;
first_len = fp->frags[IP_FIRST_FRAG_IDX].len;
n = fp->last_idx - 1;
@@ -74,6 +75,7 @@ ipv6_frag_reassemble(const struct ip_frag_pkt *fp)
m = fp->frags[IP_LAST_FRAG_IDX].mb;
ofs = fp->frags[IP_LAST_FRAG_IDX].ofs;
last_len = fp->frags[IP_LAST_FRAG_IDX].len;
+ curr_idx = IP_LAST_FRAG_IDX;
payload_len = ofs + last_len;
@@ -90,6 +92,10 @@ ipv6_frag_reassemble(const struct ip_frag_pkt *fp)
rte_pktmbuf_adj(m, (uint16_t)(m->l2_len + m->l3_len));
rte_pktmbuf_chain(fp->frags[i].mb, m);
+ /* this mbuf should not be accessed directly */
+ fp->frags[curr_idx].mb = NULL;
+ curr_idx = i;
+
/* update our last fragment and offset. */
m = fp->frags[i].mb;
ofs = fp->frags[i].ofs;
--
1.9.1
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [dpdk-dev] [PATCH] ip_frag : Fix double-free of chained mbufs
2016-04-21 17:28 [dpdk-dev] [PATCH] ip_frag : Fix double-free of chained mbufs cychong
@ 2016-04-22 9:54 ` Ananyev, Konstantin
2016-04-29 16:04 ` Thomas Monjalon
0 siblings, 1 reply; 3+ messages in thread
From: Ananyev, Konstantin @ 2016-04-22 9:54 UTC (permalink / raw)
To: cychong, dev
> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of cychong
> Sent: Thursday, April 21, 2016 6:28 PM
> To: dev@dpdk.org
> Cc: cychong@gmail.com
> Subject: [dpdk-dev] [PATCH] ip_frag : Fix double-free of chained mbufs
>
> If any fragment hole is found in ipv4_frag_reassemble() and ipv6_frag_reassemble(),
> whole ip_frag_pkt mbufs are moved to death-row. Any mbufs already chained to
> another mbuf are freed multiple times as there are still in ip_frag_pkt array.
>
> Signed-off-by: cychong <cychong@gmail.com>
> ---
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [dpdk-dev] [PATCH] ip_frag : Fix double-free of chained mbufs
2016-04-22 9:54 ` Ananyev, Konstantin
@ 2016-04-29 16:04 ` Thomas Monjalon
0 siblings, 0 replies; 3+ messages in thread
From: Thomas Monjalon @ 2016-04-29 16:04 UTC (permalink / raw)
To: cychong; +Cc: dev, Ananyev, Konstantin
> > If any fragment hole is found in ipv4_frag_reassemble() and ipv6_frag_reassemble(),
> > whole ip_frag_pkt mbufs are moved to death-row. Any mbufs already chained to
> > another mbuf are freed multiple times as there are still in ip_frag_pkt array.
> >
> > Signed-off-by: cychong <cychong@gmail.com>
>
> Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Applied, thanks
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2016-04-29 16:04 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-04-21 17:28 [dpdk-dev] [PATCH] ip_frag : Fix double-free of chained mbufs cychong
2016-04-22 9:54 ` Ananyev, Konstantin
2016-04-29 16:04 ` Thomas Monjalon
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).