This brings in a preliminary patch ("net: bcmgenet: Clear RGMII_LINK
upon link down") to make sure that ("net: bcmgenet: synchronize
EXT_RGMII_OOB_CTRL access") applies to the correct context.

Doug Berger (3):
  net: bcmgenet: synchronize EXT_RGMII_OOB_CTRL access
  net: bcmgenet: synchronize use of bcmgenet_set_rx_mode()
  net: bcmgenet: synchronize UMAC_CMD access

Florian Fainelli (1):
  net: bcmgenet: Clear RGMII_LINK upon link down

 .../net/ethernet/broadcom/genet/bcmgenet.c    | 16 ++++++++++--
 .../net/ethernet/broadcom/genet/bcmgenet.h    |  4 ++-
 .../ethernet/broadcom/genet/bcmgenet_wol.c    |  8 +++++-
 drivers/net/ethernet/broadcom/genet/bcmmii.c  | 25 ++++++++++++++-----
 4 files changed, 43 insertions(+), 10 deletions(-)

-- 
2.34.1

On Wed, May 15, 2024 at 10:02:23AM -0700, Florian Fainelli wrote:
> This brings in a preliminary patch ("net: bcmgenet: Clear RGMII_LINK
> upon link down") to make sure that ("net: bcmgenet: synchronize
> EXT_RGMII_OOB_CTRL access") applies to the correct context.

That seems to resolve the issue on 6.1 stable for me.

On Wed, May 15, 2024 at 06:35:55PM +0100, Mark Brown wrote:
> On Wed, May 15, 2024 at 10:02:23AM -0700, Florian Fainelli wrote:
> > This brings in a preliminary patch ("net: bcmgenet: Clear RGMII_LINK
> > upon link down") to make sure that ("net: bcmgenet: synchronize
> > EXT_RGMII_OOB_CTRL access") applies to the correct context.
> 
> That seems to resolve the issue on 6.1 stable for me.

Great, thanks for testing and letting us know.  I've queued these up and
will push out a -rc3 now.

greg k-h