quoted hunk ↗ jump to hunk
mptcp_active_should_disable() calculates the active-disable interval as
multiplier * blackhole_timeout * HZ. Since all operands are 32-bit, a
large blackhole_timeout can wrap the delay even on 64-bit kernels and
make active MPTCP available again while the blackhole is still present.
For example, with HZ=1000, a timeout of 6912213 seconds and the maximum
multiplier of 64 should produce 442381632000 jiffies, but the current
calculation truncates it to 512 jiffies.
This issue was found using a static analysis tool. A loopback reproducer
induced 12 MPTCP blackholes and waited 1.2 seconds before opening another
connection. The unpatched kernel retried MP_CAPABLE, while the patched
kernel kept MPTCP disabled and used TCP fallback directly. The reproducer
required neither kernel modifications nor fault injection.
Compute the interval in u64 and cap it at MAX_JIFFY_OFFSET before adding
it to the saved timestamp. This preserves the existing exponential
backoff for representable intervals and prevents large values from
wrapping into short timeouts.
Fixes: 27069e7cb3d1 ("mptcp: disable active MPTCP in case of blackhole")
Cc: <redacted>
Signed-off-by: Yingjie Wang <redacted>
---
net/mptcp/ctrl.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
index 63c5747f0f63..5224cfb3ed1f 100644
--- a/net/mptcp/ctrl.c
+++ b/net/mptcp/ctrl.c
@@ -487,6 +487,7 @@ bool mptcp_active_should_disable(struct sock *ssk)
unsigned long timeout;
int disable_times;
int multiplier;
+ u64 delay;
pernet = mptcp_get_pernet(net);
blackhole_timeout = READ_ONCE(pernet->blackhole_timeout);
@@ -503,10 +504,11 @@ bool mptcp_active_should_disable(struct sock *ssk)
/* Limit timeout to max: 2^6 * initial timeout */
multiplier = 1 << min(disable_times - 1, 6);
+ delay = (u64)multiplier * blackhole_timeout * HZ;
Hi Yingjie,
Thanks for the patch. Basically LGTM,
Just two small questions:
quoted hunk ↗ jump to hunk
/* Paired with the WRITE_ONCE() in mptcp_active_disable(). */
timeout = READ_ONCE(pernet->active_disable_stamp) +
- multiplier * blackhole_timeout * HZ;
+ min(delay, MAX_JIFFY_OFFSET);
1. AFAICS, on 64bit machine, the largest possible delay is 64 * INT_MAX * HZ, far
less than MAX_JIFFY_OFFSET ((LONG_MAX >> 1) - 1 = 2^62 - 2). So is the min() intended for the
32-bit machine?
2. If so, should the comment be updated? Because there is a second timeout on 32-bit machine.
Thanks
Gang
Reviewed-by: Gang Yan <redacted>
return time_before(jiffies, timeout);
}
--
2.43.0