From: Gowtham S Bhushan <gowtham398gs@gmail.com>
To: David Howells <dhowells@redhat.com>,
Marc Dionne <marc.dionne@auristor.com>,
"David S . Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@kernel.org>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
syzbot+0cd6f4783442ac67b56d@syzkaller.appspotmail.com
Cc: Simon Horman <horms@kernel.org>,
linux-afs@lists.infradead.org, netdev@vger.kernel.org,
linux-kernel@vger.kernel.org, syzkaller-bugs@googlegroups.com,
Gowtham S Bhushan <gowtham398gs@gmail.com>
Subject: [PATCH v2] rxrpc: Fix slab-out-of-bounds write in rxrpc_send_ACK() during PMTUD probing
Date: Thu, 19 Nov 2026 07:42:02 +0530 [thread overview]
Message-ID: <20261119021202.525077-1-gowtham398gs@gmail.com> (raw)
When rxrpc initializes Path MTU Discovery (PMTUD) parameters for a peer,
it computes the usable payload size as peer->if_mtu - peer->hdrsize. If
an interface has an MTU smaller than the transport header size (such as a
virtual interface configured with MTU 0), this unsigned subtraction
underflows to a value near UINT_MAX.
As a result, peer->pmtud_bad and peer->pmtud_trial become excessively
large. When rxrpc_send_ACK() builds a padded RXRPC_ACK_PING packet for
PMTUD probing, it appends PAGE_SIZE zero-page entries to call->local->kvec
inside a loop until len reaches probe_mtu. Because nr_kv is not checked
against the fixed 47-element capacity of call->local->kvec, nr_kv advances
past index 46 and triggers a KASAN slab-out-of-bounds write past the
1584-byte struct rxrpc_local object in the kmalloc-2k cache.
Fix this by guarding the subtraction against unsigned underflow when
peer->if_mtu <= peer->hdrsize, clamping pmtud_good, pmtud_bad, and
pmtud_trial to the usable payload limit, and only enabling pmtud_pending
when the link supports more than the minimum floor.
Additionally, ensure the padding loop in rxrpc_send_ACK() checks
nr_kv < ARRAY_SIZE(call->local->kvec) before writing to the array.
Reported-by: syzbot+0cd6f4783442ac67b56d@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=0cd6f4783442ac67b56d
Fixes: eeaedc5449d9 ("rxrpc: Implement path-MTU probing using padded PING ACKs (RFC8899)")
Signed-off-by: Gowtham S Bhushan <gowtham398gs@gmail.com>
---
Changes in v2:
- Derive peer->pmtud_bad from max_mtu rather than RXRPC_JUMBO(1) so jumbo
DATA frames remain enabled.
- Cap max_mtu to the local->kvec payload limit so large legitimate MTUs
cannot overflow local->kvec.
- Keep peer->pmtud_good at 500 and move variable declarations to the top of
rxrpc_assess_MTU_size().
- Fix probe_mtu spelling in comments and clarify the PMTUD binary-search
mechanism in the commit message.
v1-Link: https://lore.kernel.org/netdev/20261111144846.838278-1-gowtham398gs@gmail.com/
---
net/rxrpc/output.c | 12 +++++++++++-
net/rxrpc/peer_object.c | 21 ++++++++++++++++++---
2 files changed, 29 insertions(+), 4 deletions(-)
diff --git a/net/rxrpc/output.c b/net/rxrpc/output.c
index 88cad087f13b..a9d083076a16 100644
--- a/net/rxrpc/output.c
+++ b/net/rxrpc/output.c
@@ -325,9 +325,19 @@ void rxrpc_send_ACK(struct rxrpc_call *call, u8 ack_reason,
if (why == rxrpc_propose_ack_ping_for_mtu_probe) {
size_t probe_mtu = call->peer->pmtud_trial + sizeof(struct rxrpc_wire_header);
+ /*
+ * skip padding if the base ACK packets already meets/exceeds probe_mtu
+ */
+
if (len > probe_mtu)
goto skip;
- while (len < probe_mtu) {
+ /*
+ * pad the probe packet with zero pages up to probe_mtu,
+ * ensuring we never write past the fixed capacity of
+ * local->kvec.
+ */
+
+ while (len < probe_mtu && nr_kv < ARRAY_SIZE(call->local->kvec)) {
size_t part = umin(probe_mtu - len, PAGE_SIZE);
kv[nr_kv].iov_base = page_address(ZERO_PAGE(0));
diff --git a/net/rxrpc/peer_object.c b/net/rxrpc/peer_object.c
index fa9a406e1168..32b75a7617f9 100644
--- a/net/rxrpc/peer_object.c
+++ b/net/rxrpc/peer_object.c
@@ -156,6 +156,8 @@ void rxrpc_assess_MTU_size(struct rxrpc_local *local, struct rxrpc_peer *peer)
struct rtable *rt;
struct flowi fl;
struct flowi4 *fl4 = &fl.u.ip4;
+ unsigned int max_mtu, kvec_limit;
+
#ifdef CONFIG_AF_RXRPC_IPV6
struct flowi6 *fl6 = &fl.u.ip6;
#endif
@@ -207,11 +209,24 @@ void rxrpc_assess_MTU_size(struct rxrpc_local *local, struct rxrpc_peer *peer)
peer->tx_seg_max = dst->dev->gso_max_segs;
dst_release(dst);
- peer->max_data = umin(RXRPC_JUMBO(1), peer->if_mtu - peer->hdrsize);
+ /*
+ * Avoid unsigned underflow if the interface MTU is smaller than the
+ * transport header size (e.g., on virtual interfaces with MTU 0).
+ * Also cap the PMTUD upper bound to what local->kvec can hold when
+ * padding probes with PAGE_SIZE zero pages (3 header/SACK/trailer
+ * kvecs are reserved before the zero-page padding loop).
+ */
+
+ max_mtu = peer->if_mtu > peer->hdrsize ? peer->if_mtu - peer->hdrsize : 0;
+ kvec_limit = (ARRAY_SIZE(local->kvec) - 3) * PAGE_SIZE -
+ sizeof(struct rxrpc_wire_header);
+ max_mtu = umin(max_mtu, kvec_limit);
+
+ peer->max_data = umin(RXRPC_JUMBO(1), max_mtu);
peer->pmtud_good = 500;
- peer->pmtud_bad = peer->if_mtu - peer->hdrsize + 1;
+ peer->pmtud_bad = max_mtu + 1;
peer->pmtud_trial = umin(peer->max_data, peer->pmtud_bad - 1);
- peer->pmtud_pending = true;
+ peer->pmtud_pending = max_mtu > peer->pmtud_good;
_leave(" [if_mtu %u]", peer->if_mtu);
}
--
2.56.0
reply other threads:[~2026-11-19 2:12 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261119021202.525077-1-gowtham398gs@gmail.com \
--to=gowtham398gs@gmail.com \
--cc=davem@davemloft.net \
--cc=dhowells@redhat.com \
--cc=edumazet@kernel.org \
--cc=horms@kernel.org \
--cc=kuba@kernel.org \
--cc=linux-afs@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=marc.dionne@auristor.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=syzbot+0cd6f4783442ac67b56d@syzkaller.appspotmail.com \
--cc=syzkaller-bugs@googlegroups.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox