[net-next,0/2] net: phy: aquantia: enable firmware loading for aqr105 on PCIe cards

Message

Hans-Frieder Vogt Aug. 24, 2024, 5:33 p.m. UTC

This patch series adds support for firmware loading from the filesystem for
Aquantia PHYs in a non-device-tree environment and activates firmware
loading
for the AQR105 PHY.

Firmware loading for the Aquantia PHYs cannot use a static file name,
because
the firmware differs depending on the MAC it is connected to.
Therefore, the firmware name is created using the PHY and MAC (rather: MDIO)
names at runtime.

Activating firmware loading on the AQR105 is achieved by using the probe
function which is already used for the AQR107 and later PHY families.

The patch was tested on a Tehuti TN9510 card, with the improved
aqr_wait_reset_complete handling suggested by Vladimir Oltean
(discussion of patch by
Bartosz Golaszewski)
https://lore.kernel.org/netdev/20240806112747.soclko5vex2f2c64@skbuf/
(without these changes, the driver would timeout in
aqr_wait_reset_complete and
would return an error instead of loading firmware)

Hans-Frieder Vogt (2):
    net: phy: aquantia: create firmware name for aqr PHYs at runtime
    net: phy: aquantia: add firmware loading for aqr105

  drivers/net/phy/aquantia/aquantia_firmware.c | 78 ++++++++++++++++++++
  drivers/net/phy/aquantia/aquantia_main.c     |  1 +
  2 file changed, 79 insertions(+)

Comments

Andrew Lunn Aug. 26, 2024, 1:20 a.m. UTC | #1

On Sat, Aug 24, 2024 at 07:33:26PM +0200, Hans-Frieder Vogt wrote:
> This patch series adds support for firmware loading from the filesystem for
> Aquantia PHYs in a non-device-tree environment and activates firmware
> loading
> for the AQR105 PHY.


These patches are not threaded correctly. git format-patch/git
send-email should get that correct, so i wounder what you did to break
it?

	Andrew