Message ID | 20191025114436.9746-1-i.maximets@ovn.org |
---|---|
State | Accepted |
Headers | show |
Series | [ovs-dev] dpif-netdev: Fix time delta overflow in case of race for meter lock. | expand |
On Fri, Oct 25, 2019 at 4:44 AM Ilya Maximets <i.maximets@ovn.org> wrote: > > There is a race window between getting the time and getting the meter > lock. This could lead to situation where the thread with larger > current time (this thread called time_{um}sec() later than others) > will acquire meter lock first and update meter->used to the large > value. Next threads will try to calculate time delta by subtracting > the large meter->used from their lower time getting the negative value > which will be converted to a big unsigned delta. > > Fix that by assuming that all these threads received packets in the > same time in this case, i.e. dropping negative delta to 0. > > CC: Jarno Rajahalme <jarno@ovn.org> > Fixes: 4b27db644a8c ("dpif-netdev: Simple DROP meter implementation.") > Reported-at: https://mail.openvswitch.org/pipermail/ovs-dev/2019-September/363126.html > Signed-off-by: Ilya Maximets <i.maximets@ovn.org> > --- LGTM. Thanks for the fix Acked-by: William Tu <u9012063@gmail.com>
On 25.10.2019 17:55, William Tu wrote: > On Fri, Oct 25, 2019 at 4:44 AM Ilya Maximets <i.maximets@ovn.org> wrote: >> >> There is a race window between getting the time and getting the meter >> lock. This could lead to situation where the thread with larger >> current time (this thread called time_{um}sec() later than others) >> will acquire meter lock first and update meter->used to the large >> value. Next threads will try to calculate time delta by subtracting >> the large meter->used from their lower time getting the negative value >> which will be converted to a big unsigned delta. >> >> Fix that by assuming that all these threads received packets in the >> same time in this case, i.e. dropping negative delta to 0. >> >> CC: Jarno Rajahalme <jarno@ovn.org> >> Fixes: 4b27db644a8c ("dpif-netdev: Simple DROP meter implementation.") >> Reported-at: https://mail.openvswitch.org/pipermail/ovs-dev/2019-September/363126.html >> Signed-off-by: Ilya Maximets <i.maximets@ovn.org> >> --- > > LGTM. > Thanks for the fix > Acked-by: William Tu <u9012063@gmail.com> > Thanks, William! Applied to master and backported. Best regards, Ilya Maximets.
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c index 4546b55e8..13586206a 100644 --- a/lib/dpif-netdev.c +++ b/lib/dpif-netdev.c @@ -5646,6 +5646,14 @@ dp_netdev_run_meter(struct dp_netdev *dp, struct dp_packet_batch *packets_, /* All packets will hit the meter at the same time. */ long_delta_t = now / 1000 - meter->used / 1000; /* msec */ + if (long_delta_t < 0) { + /* This condition means that we have several threads fighting for a + meter lock, and the one who received the packets a bit later wins. + Assuming that all racing threads received packets at the same time + to avoid overflow. */ + long_delta_t = 0; + } + /* Make sure delta_t will not be too large, so that bucket will not * wrap around below. */ delta_t = (long_delta_t > (long long int)meter->max_delta_t)
There is a race window between getting the time and getting the meter lock. This could lead to situation where the thread with larger current time (this thread called time_{um}sec() later than others) will acquire meter lock first and update meter->used to the large value. Next threads will try to calculate time delta by subtracting the large meter->used from their lower time getting the negative value which will be converted to a big unsigned delta. Fix that by assuming that all these threads received packets in the same time in this case, i.e. dropping negative delta to 0. CC: Jarno Rajahalme <jarno@ovn.org> Fixes: 4b27db644a8c ("dpif-netdev: Simple DROP meter implementation.") Reported-at: https://mail.openvswitch.org/pipermail/ovs-dev/2019-September/363126.html Signed-off-by: Ilya Maximets <i.maximets@ovn.org> --- lib/dpif-netdev.c | 8 ++++++++ 1 file changed, 8 insertions(+)