[SRU][o:azure 0/2][n:azure 0/3][PATCH] Miscellaneous MANA fixes

Vinicius Peixoto vinicius.peixoto at canonical.com
Thu Nov 7 23:08:37 UTC 2024


[Impact]

Microsoft has requested backports for the following bugfixes:

58a63729c957 ("net: mana: Fix doorbell out of order violation and avoid unnecessary doorbell rings")
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=58a63729c957

3410d0e14f9a ("net: mana: Implement get_ringparam/set_ringparam for mana")
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=3410d0e14f9a

1705341485ff ("net: mana: Improve mana_set_channels() in low mem conditions")
https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=1705341485ff

[Fix]

58a63729c957 ("net: mana: Fix doorbell out of order violation and avoid unnecessary doorbell rings")
- oracular: already in tree
- noble: cherry-pick

3410d0e14f9a ("net: mana: Implement get_ringparam/set_ringparam for mana")
- oracular, noble: cherry-pick

1705341485ff ("net: mana: Improve mana_set_channels() in low mem conditions")
- oracular, noble: cherry-pick

[Test plan]

Compile tested only.

[Where problems could occur]

Although the diff for this patchset is sizeable (due to the introduction
of the get_ringparam/set_ringparam ethtool_ops), the changes are
relatively self-contained, so the regression risk seems low; however,
issues with them would cause degraded network operation on machines
using the MANA ethernet drivers.

[Other info]

SF #00396267

Long Li (1):
  net: mana: Fix doorbell out of order violation and avoid unnecessary
    doorbell rings

Shradha Gupta (2):
  net: mana: Implement get_ringparam/set_ringparam for mana
  net: mana: Improve mana_set_channels() in low mem conditions

 drivers/net/ethernet/microsoft/mana/mana_en.c | 53 ++++++----
 .../ethernet/microsoft/mana/mana_ethtool.c    | 96 +++++++++++++++++--
 include/net/mana/mana.h                       | 24 ++++-
 3 files changed, 139 insertions(+), 34 deletions(-)

-- 
2.43.0




More information about the kernel-team mailing list