| Message ID | E1rtn25-0065p0-2C@rmk-PC.armlinux.org.uk (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | net: dsa: allow phylink_mac_ops in DSA drivers | expand |
On Mon, Apr 08, 2024 at 12:19:25PM +0100, Russell King (Oracle) wrote: > Rather than having a shim for each and every phylink MAC operation, > allow DSA switch drivers to provide their own ops structure. > > Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Andrew
On 4/8/24 04:19, Russell King (Oracle) wrote: > Rather than having a shim for each and every phylink MAC operation, > allow DSA switch drivers to provide their own ops structure. > > Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Hi Russell,
kernel test robot noticed the following build warnings:
[auto build test WARNING on net-next/main]
url: https://github.com/intel-lab-lkp/linux/commits/Russell-King-Oracle/net-dsa-introduce-dsa_phylink_to_port/20240408-232316
base: net-next/main
patch link: https://lore.kernel.org/r/E1rtn25-0065p0-2C%40rmk-PC.armlinux.org.uk
patch subject: [PATCH net-next 2/3] net: dsa: allow DSA switch drivers to provide their own phylink mac ops
config: arm-defconfig (https://download.01.org/0day-ci/archive/20240409/202404091147.pdi8izJ3-lkp@intel.com/config)
compiler: clang version 14.0.6 (https://github.com/llvm/llvm-project.git f28c006a5895fc0e329fe15fead81e37457cb1d1)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240409/202404091147.pdi8izJ3-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202404091147.pdi8izJ3-lkp@intel.com/
All warnings (new ones prefixed by >>):
>> net/dsa/port.c:1981:6: warning: unused variable 'port' [-Wunused-variable]
int port = dp->index;
^
1 warning generated.
vim +/port +1981 net/dsa/port.c
3a506cb7012a98 Russell King (Oracle 2024-04-08 1975)
770375ff331111 Vladimir Oltean 2022-08-18 1976 int dsa_shared_port_link_register_of(struct dsa_port *dp)
57ab1ca2159717 Vivien Didelot 2017-10-26 1977 {
0e27921816ad99 Ioana Ciornei 2019-05-28 1978 struct dsa_switch *ds = dp->ds;
e09e9873152e3f Vladimir Oltean 2022-08-18 1979 bool missing_link_description;
e09e9873152e3f Vladimir Oltean 2022-08-18 1980 bool missing_phy_mode;
3be98b2d5fbca3 Andrew Lunn 2020-04-14 @1981 int port = dp->index;
0e27921816ad99 Ioana Ciornei 2019-05-28 1982
e09e9873152e3f Vladimir Oltean 2022-08-18 1983 dsa_shared_port_validate_of(dp, &missing_phy_mode,
e09e9873152e3f Vladimir Oltean 2022-08-18 1984 &missing_link_description);
e09e9873152e3f Vladimir Oltean 2022-08-18 1985
e09e9873152e3f Vladimir Oltean 2022-08-18 1986 if ((missing_phy_mode || missing_link_description) &&
e09e9873152e3f Vladimir Oltean 2022-08-18 1987 !of_device_compatible_match(ds->dev->of_node,
e09e9873152e3f Vladimir Oltean 2022-08-18 1988 dsa_switches_apply_workarounds))
e09e9873152e3f Vladimir Oltean 2022-08-18 1989 return -EINVAL;
e09e9873152e3f Vladimir Oltean 2022-08-18 1990
a20f997010c4ec Andrew Lunn 2020-03-11 1991 if (!ds->ops->adjust_link) {
e09e9873152e3f Vladimir Oltean 2022-08-18 1992 if (missing_link_description) {
e09e9873152e3f Vladimir Oltean 2022-08-18 1993 dev_warn(ds->dev,
e09e9873152e3f Vladimir Oltean 2022-08-18 1994 "Skipping phylink registration for %s port %d\n",
e09e9873152e3f Vladimir Oltean 2022-08-18 1995 dsa_port_is_cpu(dp) ? "CPU" : "DSA", dp->index);
e09e9873152e3f Vladimir Oltean 2022-08-18 1996 } else {
3a506cb7012a98 Russell King (Oracle 2024-04-08 1997) dsa_shared_port_link_down(dp);
e09e9873152e3f Vladimir Oltean 2022-08-18 1998
770375ff331111 Vladimir Oltean 2022-08-18 1999 return dsa_shared_port_phylink_register(dp);
3be98b2d5fbca3 Andrew Lunn 2020-04-14 2000 }
a20f997010c4ec Andrew Lunn 2020-03-11 2001 return 0;
a20f997010c4ec Andrew Lunn 2020-03-11 2002 }
0e27921816ad99 Ioana Ciornei 2019-05-28 2003
0e27921816ad99 Ioana Ciornei 2019-05-28 2004 dev_warn(ds->dev,
0e27921816ad99 Ioana Ciornei 2019-05-28 2005 "Using legacy PHYLIB callbacks. Please migrate to PHYLINK!\n");
0e27921816ad99 Ioana Ciornei 2019-05-28 2006
33615367f378fe Sebastian Reichel 2018-01-23 2007 if (of_phy_is_fixed_link(dp->dn))
770375ff331111 Vladimir Oltean 2022-08-18 2008 return dsa_shared_port_fixed_link_register_of(dp);
33615367f378fe Sebastian Reichel 2018-01-23 2009 else
770375ff331111 Vladimir Oltean 2022-08-18 2010 return dsa_shared_port_setup_phy_of(dp, true);
33615367f378fe Sebastian Reichel 2018-01-23 2011 }
57ab1ca2159717 Vivien Didelot 2017-10-26 2012
On Tue, Apr 09, 2024 at 11:15:01AM +0800, kernel test robot wrote: > All warnings (new ones prefixed by >>): > > >> net/dsa/port.c:1981:6: warning: unused variable 'port' [-Wunused-variable] > int port = dp->index; > ^ > 1 warning generated. Sigh... I don't get this warning when I build it. I guess that's because unused variable warnings are normally turned off. I'll respin dropping this.
On Mon, Apr 08, 2024 at 12:19:25PM +0100, Russell King (Oracle) wrote: > +static void dsa_shared_port_link_down(struct dsa_port *dp) > +{ > + struct dsa_switch *ds = dp->ds; > + > + if (ds->phylink_mac_ops) { > + if (ds->phylink_mac_ops->mac_link_down) > + ds->phylink_mac_ops->mac_link_down(&dp->pl_config, > + MLO_AN_FIXED, > + PHY_INTERFACE_MODE_NA); > + } else { > + if (ds->ops->phylink_mac_link_down) > + ds->ops->phylink_mac_link_down(ds, dp->index, > + MLO_AN_FIXED, PHY_INTERFACE_MODE_NA); > + } > +} Please roll this other change into the patch when respinning: else { if { } } becomes else if {} Also please align the arguments of the phylink_mac_link_down() call with the open parenthesis.
On Tue, Apr 09, 2024 at 03:37:31PM +0300, Vladimir Oltean wrote: > On Mon, Apr 08, 2024 at 12:19:25PM +0100, Russell King (Oracle) wrote: > > +static void dsa_shared_port_link_down(struct dsa_port *dp) > > +{ > > + struct dsa_switch *ds = dp->ds; > > + > > + if (ds->phylink_mac_ops) { > > + if (ds->phylink_mac_ops->mac_link_down) > > + ds->phylink_mac_ops->mac_link_down(&dp->pl_config, > > + MLO_AN_FIXED, > > + PHY_INTERFACE_MODE_NA); > > + } else { > > + if (ds->ops->phylink_mac_link_down) > > + ds->ops->phylink_mac_link_down(ds, dp->index, > > + MLO_AN_FIXED, PHY_INTERFACE_MODE_NA); > > + } > > +} > > Please roll this other change into the patch when respinning: > > else { > if { } > } > > becomes > > else if {} > > Also please align the arguments of the phylink_mac_link_down() call with > the open parenthesis. Something like this: static void dsa_shared_port_link_down(struct dsa_port *dp) { struct dsa_switch *ds = dp->ds; if (ds->phylink_mac_ops && ds->phylink_mac_ops->mac_link_down) { ds->phylink_mac_ops->mac_link_down(&dp->pl_config, MLO_AN_FIXED, PHY_INTERFACE_MODE_NA); } else if (ds->ops->phylink_mac_link_down) { ds->ops->phylink_mac_link_down(ds, dp->index, MLO_AN_FIXED, PHY_INTERFACE_MODE_NA); } }
On Tue, Apr 09, 2024 at 06:33:46PM +0300, Vladimir Oltean wrote: > On Tue, Apr 09, 2024 at 03:37:31PM +0300, Vladimir Oltean wrote: > > On Mon, Apr 08, 2024 at 12:19:25PM +0100, Russell King (Oracle) wrote: > > > +static void dsa_shared_port_link_down(struct dsa_port *dp) > > > +{ > > > + struct dsa_switch *ds = dp->ds; > > > + > > > + if (ds->phylink_mac_ops) { > > > + if (ds->phylink_mac_ops->mac_link_down) > > > + ds->phylink_mac_ops->mac_link_down(&dp->pl_config, > > > + MLO_AN_FIXED, > > > + PHY_INTERFACE_MODE_NA); > > > + } else { > > > + if (ds->ops->phylink_mac_link_down) > > > + ds->ops->phylink_mac_link_down(ds, dp->index, > > > + MLO_AN_FIXED, PHY_INTERFACE_MODE_NA); > > > + } > > > +} > > > > Please roll this other change into the patch when respinning: > > > > else { > > if { } > > } > > > > becomes > > > > else if {} This would destroy the symmetry that I think aids readability. I did consider it at the time and decided against it. > Something like this: > > static void dsa_shared_port_link_down(struct dsa_port *dp) > { > struct dsa_switch *ds = dp->ds; > > if (ds->phylink_mac_ops && ds->phylink_mac_ops->mac_link_down) { > ds->phylink_mac_ops->mac_link_down(&dp->pl_config, MLO_AN_FIXED, > PHY_INTERFACE_MODE_NA); > } else if (ds->ops->phylink_mac_link_down) { > ds->ops->phylink_mac_link_down(ds, dp->index, MLO_AN_FIXED, > PHY_INTERFACE_MODE_NA); > } This changes the logic - it allows driver authors to provide the MAC operations, omit the mac_link_down() op _and_ an ops->phylink_mac_link_down() function. This could lead to buggy drivers since this will only happen in this path and none of the others. I want this to be an "either provide phylink_mac_ops, and thus none of the phylink_mac_* ops in dsa_switch_ops will be called" or "don't provide phylink_mac_ops and the phylink_mac_* ops in dsa_switch_ops will be called". It's then completely clear cut that it's one or the other, whereas the code above makes it unclear.
On Tue, Apr 09, 2024 at 05:29:23PM +0100, Russell King (Oracle) wrote: > This changes the logic - it allows driver authors to provide the > MAC operations, omit the mac_link_down() op _and_ an > ops->phylink_mac_link_down() function. This could lead to buggy > drivers since this will only happen in this path and none of the > others. > > I want this to be an "either provide phylink_mac_ops, and thus > none of the phylink_mac_* ops in dsa_switch_ops will be called" or > "don't provide phylink_mac_ops and the phylink_mac_* ops in > dsa_switch_ops will be called". It's then completely clear cut > that it's one or the other, whereas the code above makes it > unclear. If you want for the API transition to be self-documenting and clear, it would be good to do that validation separately and more comprehensively rather than just a fall-through for one single operation here. If phylink_mac_link_ops is provided, the following ds->ops methods are obsoleted and can't be provided at the same time (fail probing otherwise): - phylink_mac_select_pcs() - phylink_mac_prepare() - phylink_mac_config() - phylink_mac_finish() - phylink_mac_link_down() - phylink_mac_link_up() Hopefully it makes it more clear that the following are _not_ obsoleted by the dedicated phylink mac_ops: - phylink_get_caps() - phylink_fixed_state() Then (after this validation), the simplified "if (ops && ops->mac_link_down) else (ds->ops->phylink_mac_link_down)" would be equivalent, because we've errored out on the case which has a mix of old and new API.
On Tue, Apr 09, 2024 at 07:52:30PM +0300, Vladimir Oltean wrote: > On Tue, Apr 09, 2024 at 05:29:23PM +0100, Russell King (Oracle) wrote: > > This changes the logic - it allows driver authors to provide the > > MAC operations, omit the mac_link_down() op _and_ an > > ops->phylink_mac_link_down() function. This could lead to buggy > > drivers since this will only happen in this path and none of the > > others. > > > > I want this to be an "either provide phylink_mac_ops, and thus > > none of the phylink_mac_* ops in dsa_switch_ops will be called" or > > "don't provide phylink_mac_ops and the phylink_mac_* ops in > > dsa_switch_ops will be called". It's then completely clear cut > > that it's one or the other, whereas the code above makes it > > unclear. > > If you want for the API transition to be self-documenting and clear, > it would be good to do that validation separately and more comprehensively > rather than just a fall-through for one single operation here. > > If phylink_mac_link_ops is provided, the following ds->ops methods are > obsoleted and can't be provided at the same time (fail probing otherwise): > > - phylink_mac_select_pcs() > - phylink_mac_prepare() > - phylink_mac_config() > - phylink_mac_finish() > - phylink_mac_link_down() > - phylink_mac_link_up() > > Hopefully it makes it more clear that the following are _not_ obsoleted > by the dedicated phylink mac_ops: > > - phylink_get_caps() > - phylink_fixed_state() > > Then (after this validation), the simplified > "if (ops && ops->mac_link_down) else (ds->ops->phylink_mac_link_down)" > would be equivalent, because we've errored out on the case which has a > mix of old and new API. Okay. However, I can't predict when I'll get around to doing these changes as my time is very limited over this week and next week - so may be right before the merge window is due. Maybe we should've done that for the ops->adjust_link vs ops->phylink_mac_link_down/ops->phylink_mac_link_up as well.
diff --git a/include/net/dsa.h b/include/net/dsa.h index f228b479a5fd..7edfd8de8882 100644 --- a/include/net/dsa.h +++ b/include/net/dsa.h @@ -457,6 +457,11 @@ struct dsa_switch { */ const struct dsa_switch_ops *ops; + /* + * Allow a DSA switch driver to override the phylink MAC ops + */ + const struct phylink_mac_ops *phylink_mac_ops; + /* * User mii_bus and devices for the individual ports. */ diff --git a/net/dsa/port.c b/net/dsa/port.c index 02bf1c306bdc..bfacf41344df 100644 --- a/net/dsa/port.c +++ b/net/dsa/port.c @@ -1662,6 +1662,7 @@ static const struct phylink_mac_ops dsa_port_phylink_mac_ops = { int dsa_port_phylink_create(struct dsa_port *dp) { + const struct phylink_mac_ops *mac_ops; struct dsa_switch *ds = dp->ds; phy_interface_t mode; struct phylink *pl; @@ -1685,8 +1686,12 @@ int dsa_port_phylink_create(struct dsa_port *dp) } } - pl = phylink_create(&dp->pl_config, of_fwnode_handle(dp->dn), - mode, &dsa_port_phylink_mac_ops); + mac_ops = &dsa_port_phylink_mac_ops; + if (ds->phylink_mac_ops) + mac_ops = ds->phylink_mac_ops; + + pl = phylink_create(&dp->pl_config, of_fwnode_handle(dp->dn), mode, + mac_ops); if (IS_ERR(pl)) { pr_err("error creating PHYLINK: %ld\n", PTR_ERR(pl)); return PTR_ERR(pl); @@ -1952,6 +1957,22 @@ static void dsa_shared_port_validate_of(struct dsa_port *dp, dn, dsa_port_is_cpu(dp) ? "CPU" : "DSA", dp->index); } +static void dsa_shared_port_link_down(struct dsa_port *dp) +{ + struct dsa_switch *ds = dp->ds; + + if (ds->phylink_mac_ops) { + if (ds->phylink_mac_ops->mac_link_down) + ds->phylink_mac_ops->mac_link_down(&dp->pl_config, + MLO_AN_FIXED, + PHY_INTERFACE_MODE_NA); + } else { + if (ds->ops->phylink_mac_link_down) + ds->ops->phylink_mac_link_down(ds, dp->index, + MLO_AN_FIXED, PHY_INTERFACE_MODE_NA); + } +} + int dsa_shared_port_link_register_of(struct dsa_port *dp) { struct dsa_switch *ds = dp->ds; @@ -1973,9 +1994,7 @@ int dsa_shared_port_link_register_of(struct dsa_port *dp) "Skipping phylink registration for %s port %d\n", dsa_port_is_cpu(dp) ? "CPU" : "DSA", dp->index); } else { - if (ds->ops->phylink_mac_link_down) - ds->ops->phylink_mac_link_down(ds, port, - MLO_AN_FIXED, PHY_INTERFACE_MODE_NA); + dsa_shared_port_link_down(dp); return dsa_shared_port_phylink_register(dp); }
Rather than having a shim for each and every phylink MAC operation, allow DSA switch drivers to provide their own ops structure. Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> --- include/net/dsa.h | 5 +++++ net/dsa/port.c | 29 ++++++++++++++++++++++++----- 2 files changed, 29 insertions(+), 5 deletions(-)