When determining the resend timer value, we have a value in nsec but the
timer is in jiffies which may be a million or more times more coarse.
nsecs_to_jiffies() rounds down - which means that the resend timeout
expressed as jiffies is very likely earlier than the one expressed as
nanoseconds from which it was derived.
The problem is that rxrpc_resend() gets triggered by the timer, but can't
then find anything to resend yet. It sets the timer again - but gets
kicked off immediately again and again until the nanosecond-based expiry
time is reached and we actually retransmit.
Fix this by adding 1 to the jiffies-based resend_at value to counteract the
rounding and make sure that the timer happens after the nanosecond-based
expiry is passed.
Alternatives would be to adjust the timestamp on the packets to align
with the jiffie scale or to switch back to using jiffie-timestamps.
Signed-off-by: David Howells <dhowe...@redhat.com>
---
net/rxrpc/call_event.c | 10 ++++++++--
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/net/rxrpc/call_event.c b/net/rxrpc/call_event.c
index a78a92fe5d77..d5bf9ce7ec6f 100644
--- a/net/rxrpc/call_event.c
+++ b/net/rxrpc/call_event.c
@@ -200,8 +200,14 @@ static void rxrpc_resend(struct rxrpc_call *call)
ktime_to_ns(ktime_sub(skb->tstamp,
max_age)));
}
- resend_at = ktime_sub(ktime_add_ms(oldest, rxrpc_resend_timeout), now);
- call->resend_at = jiffies + nsecs_to_jiffies(ktime_to_ns(resend_at));
+ resend_at = ktime_add_ms(oldest, rxrpc_resend_timeout);
+ call->resend_at = jiffies +
+ nsecs_to_jiffies(ktime_to_ns(ktime_sub(resend_at, now))) +
+ 1; /* We have to make sure that the calculated jiffies value
+ * falls at or after the nsec value, or we shall loop
+ * ceaselessly because the timer times out, but we haven't
+ * reached the nsec timeout yet.
+ */
/* Now go through the Tx window and perform the retransmissions. We
* have to drop the lock for each send. If an ACK comes in whilst the
