[Patch v3 net-next 0/3] Add support for mdb offload failure notification

Joseph Huang posted 3 patches 10 months, 1 week ago
There is a newer version of this series
include/uapi/linux/if_bridge.h | 10 ++++++----
net/bridge/br.c                |  5 +++++
net/bridge/br_mdb.c            | 28 +++++++++++++++++++++++-----
net/bridge/br_private.h        | 30 +++++++++++++++++++++++++-----
net/bridge/br_switchdev.c      | 13 +++++++++----
5 files changed, 68 insertions(+), 18 deletions(-)
[Patch v3 net-next 0/3] Add support for mdb offload failure notification
Posted by Joseph Huang 10 months, 1 week ago
Currently the bridge does not provide real-time feedback to user space
on whether or not an attempt to offload an mdb entry was successful.

This patch set adds support to notify user space about failed offload
attempts, and is controlled by a new knob mdb_offload_fail_notification.

A break-down of the patches in the series:

Patch 1 adds offload failed flag to indicate that the offload attempt
has failed. The flag is reflected in netlink mdb entry flags.

Patch 2 adds the new bridge bool option mdb_offload_fail_notification.

Patch 3 notifies user space when the result is known, controlled by
mdb_offload_fail_notification setting.

Joseph Huang (3):
  net: bridge: mcast: Add offload failed mdb flag
  net: bridge: Add offload_fail_notification bopt
  net: bridge: mcast: Notify on mdb offload failure

 include/uapi/linux/if_bridge.h | 10 ++++++----
 net/bridge/br.c                |  5 +++++
 net/bridge/br_mdb.c            | 28 +++++++++++++++++++++++-----
 net/bridge/br_private.h        | 30 +++++++++++++++++++++++++-----
 net/bridge/br_switchdev.c      | 13 +++++++++----
 5 files changed, 68 insertions(+), 18 deletions(-)

---
v1: https://lore.kernel.org/netdev/20250318224255.143683-1-Joseph.Huang@garmin.com/
    iproute2 link:
    https://lore.kernel.org/netdev/20250318225026.145501-1-Joseph.Huang@garmin.com/
v2: https://lore.kernel.org/netdev/20250403234412.1531714-1-Joseph.Huang@garmin.com/
    iproute2 link:
    https://lore.kernel.org/netdev/20250403235452.1534269-1-Joseph.Huang@garmin.com/
    Add br_multicast_pg_set_offload_flags helper to set offload flags
    Change multi-valued option mdb_notify_on_flag_change to bool option
    mdb_offload_fail_notification
    Change _br_mdb_notify to __br_mdb_notify
    Drop all #ifdef CONFIG_NET_SWITCHDEV
    Add br_mdb_should_notify helper and reorganize code in
    br_switch_mdb_complete
v3: Patch 1/3 Do not set offload flags when switchdev returns -EOPNOTSUPP
-- 
2.49.0
Re: [Patch v3 net-next 0/3] Add support for mdb offload failure notification
Posted by Jakub Kicinski 10 months ago
On Fri, 4 Apr 2025 17:29:32 -0400 Joseph Huang wrote:
> Currently the bridge does not provide real-time feedback to user space
> on whether or not an attempt to offload an mdb entry was successful.
> 
> This patch set adds support to notify user space about failed offload
> attempts, and is controlled by a new knob mdb_offload_fail_notification.
> 
> A break-down of the patches in the series:
> 
> Patch 1 adds offload failed flag to indicate that the offload attempt
> has failed. The flag is reflected in netlink mdb entry flags.
> 
> Patch 2 adds the new bridge bool option mdb_offload_fail_notification.
> 
> Patch 3 notifies user space when the result is known, controlled by
> mdb_offload_fail_notification setting.

You submitted this during the merge window, when the net-next tree
was closed. See: 
https://www.kernel.org/doc/html/next/process/maintainer-netdev.html#development-cycle
Could you repost so that the series will be re-enqueued? 

Thanks!
-- 
pw-bot: defer
Re: [Patch v3 net-next 0/3] Add support for mdb offload failure notification
Posted by Joseph Huang 10 months ago
On 4/7/2025 1:29 PM, Jakub Kicinski wrote:
> On Fri, 4 Apr 2025 17:29:32 -0400 Joseph Huang wrote:
>> Currently the bridge does not provide real-time feedback to user space
>> on whether or not an attempt to offload an mdb entry was successful.
>>
>> This patch set adds support to notify user space about failed offload
>> attempts, and is controlled by a new knob mdb_offload_fail_notification.
>>
>> A break-down of the patches in the series:
>>
>> Patch 1 adds offload failed flag to indicate that the offload attempt
>> has failed. The flag is reflected in netlink mdb entry flags.
>>
>> Patch 2 adds the new bridge bool option mdb_offload_fail_notification.
>>
>> Patch 3 notifies user space when the result is known, controlled by
>> mdb_offload_fail_notification setting.
> 
> You submitted this during the merge window, when the net-next tree
> was closed. See:
> https://www.kernel.org/doc/html/next/process/maintainer-netdev.html#development-cycle
> Could you repost so that the series will be re-enqueued?
> 
> Thanks!

Sure thing!

A couple of questions:

- Should the re-post be v3 (no change) or v4 (bump)?
- Do I re-post after 6.15 is released? Around what time frame (so that I 
can set a reminder)?

Thanks,
Joseph
Re: [Patch v3 net-next 0/3] Add support for mdb offload failure notification
Posted by Jakub Kicinski 10 months ago
On Mon, 7 Apr 2025 14:15:31 -0400 Joseph Huang wrote:
> - Should the re-post be v3 (no change) or v4 (bump)?

Doesn't matter much, but probably v4 is less confusing.

> - Do I re-post after 6.15 is released? Around what time frame (so that I 
> can set a reminder)?

No, no, I mean very soon. Like tomorrow. The merge window is when
maintainers merge their tress up, rather than when we merge code
from contributors. It's a bit confusing. Merge window open ==
normal contribution closed, and vice versa.
Re: [Patch v3 net-next 0/3] Add support for mdb offload failure notification
Posted by Joseph Huang 10 months ago
On 4/7/2025 2:37 PM, Jakub Kicinski wrote:
> On Mon, 7 Apr 2025 14:15:31 -0400 Joseph Huang wrote:
>> - Should the re-post be v3 (no change) or v4 (bump)?
> 
> Doesn't matter much, but probably v4 is less confusing.
> 
>> - Do I re-post after 6.15 is released? Around what time frame (so that I
>> can set a reminder)?
> 
> No, no, I mean very soon. Like tomorrow. The merge window is when
> maintainers merge their tress up, rather than when we merge code
> from contributors. It's a bit confusing. Merge window open ==
> normal contribution closed, and vice versa.

Got it. Thanks for the clarification. Will re-post tomorrow.

Thanks,
Joseph