| Message ID | Zw-OCSv7SldjB7iU@shell.armlinux.org.uk (mailing list archive) |
|---|---|
| Headers | show |
| Series | Removing more phylink cruft | expand |
Hi, On Wed, 16 Oct 2024 10:57:29 +0100 "Russell King (Oracle)" <linux@armlinux.org.uk> wrote: > Hi, > > Continuing on with the cleanups, this patch series removes > dsa_port_phylink_mac_select_pcs() which is no longer required. This > will have no functional effect as phylink does this: > > bool using_mac_select_pcs = false; > > if (mac_ops->mac_select_pcs && > mac_ops->mac_select_pcs(config, PHY_INTERFACE_MODE_NA) != > ERR_PTR(-EOPNOTSUPP)) > using_mac_select_pcs = true; > > and no mac_select_pcs() method is equivalent to a mac_select_pcs() > that returns -EOPNOTSUPP. > > We then make mv88e6xxx_mac_select_pcs() return NULL, as we don't want > to invoke this old behaviour anymore - mv88e6xxx doesn't require it. > > Then, allow phylink to remove PCS, which has been a long standing > behavioural oddity. > > Remove the use of pl->pcs when validating as this will never be > non-NULL unless "using_mac_select_pcs" was set. > > This then clears the way to removing using_mac_select_pcs from phylink > and the check. I don't have any setup with such a switch or that could trigger the PCS removal, so I couldn't fully test it. However it does look good to me :) For the series, Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Maxime
Hello: This series was applied to netdev/net-next.git (main) by Andrew Lunn <andrew@lunn.ch>: On Wed, 16 Oct 2024 10:57:29 +0100 you wrote: > Hi, > > Continuing on with the cleanups, this patch series removes > dsa_port_phylink_mac_select_pcs() which is no longer required. This > will have no functional effect as phylink does this: > > bool using_mac_select_pcs = false; > > [...] Here is the summary with links: - [net-next,v2,1/5] net: dsa: remove dsa_port_phylink_mac_select_pcs() https://git.kernel.org/netdev/net-next/c/ecb595ebba0e - [net-next,v2,2/5] net: dsa: mv88e6xxx: return NULL when no PCS is present https://git.kernel.org/netdev/net-next/c/14ca726ada7f - [net-next,v2,3/5] net: phylink: allow mac_select_pcs() to remove a PCS https://git.kernel.org/netdev/net-next/c/486dc391ef43 - [net-next,v2,4/5] net: phylink: remove use of pl->pcs in phylink_validate_mac_and_pcs() https://git.kernel.org/netdev/net-next/c/6c48cd044cc8 - [net-next,v2,5/5] net: phylink: remove "using_mac_select_pcs" https://git.kernel.org/netdev/net-next/c/7530ea26c810 You are awesome, thank you!
Hi, Continuing on with the cleanups, this patch series removes dsa_port_phylink_mac_select_pcs() which is no longer required. This will have no functional effect as phylink does this: bool using_mac_select_pcs = false; if (mac_ops->mac_select_pcs && mac_ops->mac_select_pcs(config, PHY_INTERFACE_MODE_NA) != ERR_PTR(-EOPNOTSUPP)) using_mac_select_pcs = true; and no mac_select_pcs() method is equivalent to a mac_select_pcs() that returns -EOPNOTSUPP. We then make mv88e6xxx_mac_select_pcs() return NULL, as we don't want to invoke this old behaviour anymore - mv88e6xxx doesn't require it. Then, allow phylink to remove PCS, which has been a long standing behavioural oddity. Remove the use of pl->pcs when validating as this will never be non-NULL unless "using_mac_select_pcs" was set. This then clears the way to removing using_mac_select_pcs from phylink and the check. drivers/net/dsa/mv88e6xxx/chip.c | 2 +- drivers/net/phy/phylink.c | 18 +++----------- net/dsa/port.c | 8 -------- 3 files changed, 5 insertions(+), 23 deletions(-)