| Message ID | 20250207092920.543458-1-liuhangbin@gmail.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | bonding: fix incorrect mac address setting | expand |
Please hold on this patch. Our QE reported that with bare NIC, the backup NIC can't receive the NS messages even after joining the multicast MAC group. But after remove the backup NIC from bond, the NIC interface could receive the NS message. This is weird, it looks the backup NIC dropped the NS message somewhere, even using tcpdump (the NIC will be in promisc mode) I can't capture the NS message on backup slave. I need to debug more. Thanks Hangbin On Fri, Feb 07, 2025 at 09:29:18AM +0000, Hangbin Liu wrote: > The mac address on backup slave should be convert from Solicited-Node > Multicast address, not from bonding unicast target address. > > v3: also fix the mac setting for slave_set_ns_maddr. (Jay) > Add function description for slave_set_ns_maddr/slave_set_ns_maddrs (Jay) > v2: fix patch 01's subject > > Hangbin Liu (2): > bonding: fix incorrect MAC address setting to receive NS messages > selftests: bonding: fix incorrect mac address > > drivers/net/bonding/bond_options.c | 55 ++++++++++++++++--- > .../drivers/net/bonding/bond_options.sh | 4 +- > 2 files changed, 49 insertions(+), 10 deletions(-) > > -- > 2.46.0 >
Hi Jay, On Sat, Feb 08, 2025 at 06:34:21AM +0000, Hangbin Liu wrote: > Please hold on this patch. Our QE reported that with bare NIC, the > backup NIC can't receive the NS messages even after joining the multicast > MAC group. But after remove the backup NIC from bond, the NIC interface > could receive the NS message. > > This is weird, it looks the backup NIC dropped the NS message somewhere, > even using tcpdump (the NIC will be in promisc mode) I can't capture the > NS message on backup slave. > > I need to debug more. After debug, I find it's a driver issue. The issue exists with ice dirver NIC. I tried with a mlx5 NIC and the patch works good for me. So I think you can start review this patch when you have time. I will debug the ice driver later. Thanks Hangbin
Hi Jay, Any comments? Thanks Hangbin On Tue, Feb 11, 2025 at 07:31:32AM +0000, Hangbin Liu wrote: > Hi Jay, > On Sat, Feb 08, 2025 at 06:34:21AM +0000, Hangbin Liu wrote: > > Please hold on this patch. Our QE reported that with bare NIC, the > > backup NIC can't receive the NS messages even after joining the multicast > > MAC group. But after remove the backup NIC from bond, the NIC interface > > could receive the NS message. > > > > This is weird, it looks the backup NIC dropped the NS message somewhere, > > even using tcpdump (the NIC will be in promisc mode) I can't capture the > > NS message on backup slave. > > > > I need to debug more. > > After debug, I find it's a driver issue. The issue exists with ice dirver NIC. > I tried with a mlx5 NIC and the patch works good for me. > > So I think you can start review this patch when you have time. I will debug > the ice driver later. > > Thanks > Hangbin
Hangbin Liu <liuhangbin@gmail.com> wrote: >The mac address on backup slave should be convert from Solicited-Node >Multicast address, not from bonding unicast target address. > >v3: also fix the mac setting for slave_set_ns_maddr. (Jay) > Add function description for slave_set_ns_maddr/slave_set_ns_maddrs (Jay) >v2: fix patch 01's subject > >Hangbin Liu (2): > bonding: fix incorrect MAC address setting to receive NS messages > selftests: bonding: fix incorrect mac address For the series: Acked-by: Jay Vosburgh <jv@jvosburgh.net> -J > drivers/net/bonding/bond_options.c | 55 ++++++++++++++++--- > .../drivers/net/bonding/bond_options.sh | 4 +- > 2 files changed, 49 insertions(+), 10 deletions(-) > >-- >2.46.0 > --- -Jay Vosburgh, jv@jvosburgh.net
On 2/7/25 11:29, Hangbin Liu wrote: > The mac address on backup slave should be convert from Solicited-Node > Multicast address, not from bonding unicast target address. > > v3: also fix the mac setting for slave_set_ns_maddr. (Jay) > Add function description for slave_set_ns_maddr/slave_set_ns_maddrs (Jay) > v2: fix patch 01's subject > > Hangbin Liu (2): > bonding: fix incorrect MAC address setting to receive NS messages > selftests: bonding: fix incorrect mac address > > drivers/net/bonding/bond_options.c | 55 ++++++++++++++++--- > .../drivers/net/bonding/bond_options.sh | 4 +- > 2 files changed, 49 insertions(+), 10 deletions(-) > For the set: Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org>
Hi David, This patch set was marked as "Changes Requested" due to my initial reply. The series has now been approved by Jay and Nikolay. Could you help process it, or should I re-post it? Thanks Hangbin On Fri, Feb 07, 2025 at 09:29:18AM +0000, Hangbin Liu wrote: > The mac address on backup slave should be convert from Solicited-Node > Multicast address, not from bonding unicast target address. > > v3: also fix the mac setting for slave_set_ns_maddr. (Jay) > Add function description for slave_set_ns_maddr/slave_set_ns_maddrs (Jay) > v2: fix patch 01's subject > > Hangbin Liu (2): > bonding: fix incorrect MAC address setting to receive NS messages > selftests: bonding: fix incorrect mac address > > drivers/net/bonding/bond_options.c | 55 ++++++++++++++++--- > .../drivers/net/bonding/bond_options.sh | 4 +- > 2 files changed, 49 insertions(+), 10 deletions(-) > > -- > 2.46.0 >
On Wed, 5 Mar 2025 08:25:07 +0000 Hangbin Liu wrote:
> Could you help process it, or should I re-post it?
Repost.
The mac address on backup slave should be convert from Solicited-Node Multicast address, not from bonding unicast target address. v3: also fix the mac setting for slave_set_ns_maddr. (Jay) Add function description for slave_set_ns_maddr/slave_set_ns_maddrs (Jay) v2: fix patch 01's subject Hangbin Liu (2): bonding: fix incorrect MAC address setting to receive NS messages selftests: bonding: fix incorrect mac address drivers/net/bonding/bond_options.c | 55 ++++++++++++++++--- .../drivers/net/bonding/bond_options.sh | 4 +- 2 files changed, 49 insertions(+), 10 deletions(-)