Hi Carlo,

On 26.06.26 18:59, Carlo Szelinsky wrote:

This is v3 of Corey's series [1]. It takes the PSE controller lookup out
of the MDIO probe path, so a modular PSE driver no longer makes the
PHY/DSA probe spin on -EPROBE_DEFER until the PSE module loads.

v2 was four patches. The first one (a regulator handle fix) is a
self-contained bug fix, so on Jakub's suggestion it is going to net on
its own [2] and is not part of this series. The three patches here are
the notifier rework and target net-next. net-next was closed for the
merge window when v2 was posted; it is open again now, so here they are.

How it works: pse_core gets a notifier chain (REGISTERED / UNREGISTERED).
The phy layer subscribes, owns phydev->psec, and attaches the PSE handle
when the controller shows up instead of during probe. fwnode_mdio loses
its PSE awareness, so no -EPROBE_DEFER leaves it and the probe-retry loop
is gone.

Tested on a Realtek rtl93xx PoE switch with two HS104 PSE controllers on
i2c:

 - clean boot, no probe-retry loop, no watchdog reset
 - 10G SFP+ port: module hotplug works, no deadlock
 - ethtool --set-pse enable/disable cuts and restores power to a PD
 - i2c unbind -> rmmod -> modprobe: PSE detaches on unbind and re-attaches
   on reload with power restored, no reboot. No lockdep splats.

Tested-by: Carlo Szelinsky <redacted>

Changes in v3:
 - Drop patch 1 (regulator handle fix); it goes to net separately [2].
 - Rebase on current net-next. No code changes to the three patches.

v1 was an RFC by Corey [3].

[1] https://lore.kernel.org/netdev/20260620112440.1734404-1-github@szelinsky.de/ (local)
[2] https://lore.kernel.org/netdev/20260624204017.2752934-1-github@szelinsky.de/ (local)
[3] https://lore.kernel.org/netdev/20260423-pse-notifier-decouple-v1-0-86ed750a9d62@leavitt.info/ (local)

Corey Leavitt (3):
  net: pse-pd: add notifier chain for controller lifecycle events
  net: pse-pd: fire lifecycle events on controller register/unregister
  net: phy: own phydev->psec via PSE notifier and remove fwnode_mdio
    hook

 drivers/net/mdio/fwnode_mdio.c |  34 -------
 drivers/net/phy/phy_device.c   | 168 +++++++++++++++++++++++++++++++--
 drivers/net/phy/sfp.c          |   2 +-
 drivers/net/pse-pd/pse_core.c  |  54 +++++++++++
 include/linux/phy.h            |   2 +
 include/linux/pse-pd/pse.h     |  41 ++++++++
 6 files changed, 258 insertions(+), 43 deletions(-)


base-commit: 805185b7c7a1069e407b6f7b3bc98e44d415f484

Thanks a lot for continuing the work!

since you need to do a v4 anyway because net-next is still closed, feel
free to include:

Tested-by: Jonas Jelonek <jelonek.jonas@gmail.com>

The issue I reported before, specifically with RTL8241FC, is fixed now.
No deadlock anymore.

Best,
Jonas