ACK/Cmnt: [SRU][G/F][PATCH 0/1] net/mlx5e: Fix multicast counter not up-to-date in "ip -s"

Stefan Bader stefan.bader at canonical.com
Wed Jan 20 09:00:07 UTC 2021


On 15.01.21 14:46, William Breathitt Gray wrote:
> [Impact]
> On a mlx5 device rx-mcast counter is not increased in ip -s after
> receiving mcast packets. Rx-packets counter is increased normally.
> 
> [Test case]
> Send multicast packets to a mlx5 device using mcast address (like 224.0.0.1)
> # iperf --client 224.0.0.1 -u --bind 15.194.5.1
> 
> See that before and after sending the mcast counter remains the same
> 
> before:
> # ip -s l show ens5f0
> 10: ens5f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
>     link/ether ec:0d:9a:d4:2d:fc brd ff:ff:ff:ff:ff:ff
>     RX: bytes packets errors dropped overrun mcast
>     1350216 893 0 0 0 0
>     TX: bytes packets errors dropped carrier collsns
>     16221 94 0 0 0 0
> 
> after
> # ip -s l show ens5f0
> 10: ens5f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
>     link/ether ec:0d:9a:d4:2d:fc brd ff:ff:ff:ff:ff:ff
>     RX: bytes packets errors dropped overrun mcast
>     2700432 1786 0 0 0 0
>     TX: bytes packets errors dropped carrier collsns
>     16221 94 0 0 0 0
> 
> [Fix]
> The issue is solved with upstream commit 47c97e6b10a1
> ("net/mlx5e: Fix multicast counter not up-to-date in "ip -s"").
> 
> After applying this patch the mcast counter is increased.
> 
> Before sending mcast packets:
> # ip -s l show ens5f0
> 10: ens5f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
>     link/ether ec:0d:9a:d4:2d:fc brd ff:ff:ff:ff:ff:ff
>     RX: bytes packets errors dropped overrun mcast
>     0 0 0 0 0 0
>     TX: bytes packets errors dropped carrier collsns
>     7182 41 0 0 0 0
> 
> After sending mcast packets:
> # ip -s l show ens5f0
> 10: ens5f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
>     link/ether ec:0d:9a:d4:2d:fc brd ff:ff:ff:ff:ff:ff
>     RX: bytes packets errors dropped overrun mcast
>     1350216 893 0 0 0 893
>     TX: bytes packets errors dropped carrier collsns
>     9860 58 0 0 0 0
> 
> The patch is applied with 2 rejected hunks that need to be inserted
> manually with no changes.
> 
> [Regression Potential]
> Regression risk is low. No change in the code of upstream path, only
> context adjustments.

Again, as a reminder, the regression potential in the sru template for the bug
report does not aim for a liklyhood. Rather an outlook of what could be going
wrong and how that would manifest. Like a case where the counter worked before
for some reason, no longer works after the change.

Acked-by: Stefan Bader <stefan.bader at canonical.com>
> 
> Ron Diskin (1):
>   net/mlx5e: Fix multicast counter not up-to-date in "ip -s"
> 
>  drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h  | 5 +++++
>  drivers/net/ethernet/mellanox/mlx5/core/en_main.c  | 8 +-------
>  drivers/net/ethernet/mellanox/mlx5/core/en_rx.c    | 4 ++++
>  drivers/net/ethernet/mellanox/mlx5/core/en_stats.h | 2 ++
>  4 files changed, 12 insertions(+), 7 deletions(-)
> 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20210120/9d1c54f0/attachment-0001.sig>


More information about the kernel-team mailing list