Various mtk_eth_soc cleanups
Subsystems (4)
cover Various mtk_eth_soc cleanups
Hi,
These are a number of patches that do a bit of cleanup to mtk_eth_soc, including one (the last) which seems to indicate that I somehow missed adding RMII and REVMII to the supported_interfaces - but no one has complained, and checking the DT files in the kernel, no one uses these modes (so the driver is probably untested for these.)
The first patch cleans up mtk_gmac0_rgmii_adjust(), which is the troublesome function preventing the driver becoming a post-March2020 phylink driver. It doesn't solve that problem, merely makes the code easier to follow by getting rid of repeated tenary operators.
The second patch moves the check for DDR2 memory to the initialisation of phylink's supported_interfaces - if TRGMII is not possible for some reason, we should not be erroring out in phylink MAC operations when that can be determined prior to phylink creation.
The third patch removes checks from mtk_mac_config() that are done when initialising supported_interfaces - phylink will not call mtk_mac_config() with an interface that was not marked as supported, so these checks are redundant.
The last patch adds comments for REVMII and RMII. As I note, these seem to be unused in the DT files, and as no one has reported that these don't work, I suggest that no one uses them. Should we drop support for these modes, or add them to supported_interfaces?
Please review, and I'll collect attributations for resending after net-next has re-opened.
Thanks.
drivers/net/ethernet/mediatek/mtk_eth_soc.c | 71 ++++++++++++++--------------- 1 file changed, 34 insertions(+), 37 deletions(-)
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
RFC PATCH 1/4 net: mtk_eth_soc: tidy mtk_gmac0_rgmii_adjust()
Get rid of the multiple tenary operators in mtk_gmac0_rgmii_adjust() replacing them with a single if(), thus making the code easier to read.
--- drivers/net/ethernet/mediatek/mtk_eth_soc.c | 34 ++++++++++++--------- 1 file changed, 19 insertions(+), 15 deletions(-)
drivers/net/ethernet/mediatek/mtk_eth_soc.c +19 -15
@@ -397,38 +397,42 @@ static int mt7621_gmac0_rgmii_adjust(struct mtk_eth *eth,static void mtk_gmac0_rgmii_adjust (struct mtk_eth *eth,phy_interface_t interface,int speed) { -u32 val; +unsigned long rate; +u32 tck, rck, intf;int ret;if (interface ==PHY_INTERFACE_MODE_TRGMII ) {mtk_w32 (eth,TRGMII_MODE ,INTF_MODE ); - val =500000000 ; - ret =clk_set_rate (eth->clks[MTK_CLK_TRGPLL ], val); + ret =clk_set_rate (eth->clks[MTK_CLK_TRGPLL ],500000000 );if (ret)dev_err (eth->dev,"Failed to set trgmii pll: %d\n" , ret);return ; } - val = (speed ==SPEED_1000 ) ? -INTF_MODE_RGMII_1000 :INTF_MODE_RGMII_10_100 ; -mtk_w32 (eth, val,INTF_MODE ); +if (speed ==SPEED_1000 ) { + intf =INTF_MODE_RGMII_1000 ; + rate =250000000 ; + rck =RCK_CTRL_RGMII_1000 ; + tck =TCK_CTRL_RGMII_1000 ; + }else { + intf =INTF_MODE_RGMII_10_100 ; + rate =500000000 ; + rck =RCK_CTRL_RGMII_10_100 ; + tck =TCK_CTRL_RGMII_10_100 ; + } + +mtk_w32 (eth, intf,INTF_MODE );regmap_update_bits (eth->ethsys,ETHSYS_CLKCFG0 ,ETHSYS_TRGMII_CLK_SEL362_5 ,ETHSYS_TRGMII_CLK_SEL362_5 ); - val = (speed ==SPEED_1000 ) ?250000000 :500000000 ; - ret =clk_set_rate (eth->clks[MTK_CLK_TRGPLL ], val); + ret =clk_set_rate (eth->clks[MTK_CLK_TRGPLL ], rate);if (ret)dev_err (eth->dev,"Failed to set trgmii pll: %d\n" , ret); - val = (speed ==SPEED_1000 ) ? -RCK_CTRL_RGMII_1000 :RCK_CTRL_RGMII_10_100 ; -mtk_w32 (eth, val,TRGMII_RCK_CTRL ); - - val = (speed ==SPEED_1000 ) ? -TCK_CTRL_RGMII_1000 :TCK_CTRL_RGMII_10_100 ; -mtk_w32 (eth, val,TRGMII_TCK_CTRL ); +mtk_w32 (eth, rck,TRGMII_RCK_CTRL ); +mtk_w32 (eth, tck,TRGMII_TCK_CTRL ); } static struct phylink_pcs *mtk_mac_select_pcs(struct phylink_config *config,
Signature · 1 line
2.30.2
Re: [PATCH RFC net-next 1/4] net: mtk_eth_soc: tidy mtk_gmac0_rgmii_adjust()
Quoted 4 lines · Russell King (Oracle)
Get rid of the multiple tenary operators in
mtk_gmac0_rgmii_adjust()replacing them with a singleif(), thus making the code easier to read.Signed-off-by Russell King (Oracle) <[email protected]>
RFC PATCH 2/4 net: mtk_eth_soc: move trgmii ddr2 check to probe function
In reply to Russell King (Oracle)If TRGMII mode is not permitted when using DDR2 mode, we should handle that when setting up phylink's ->supported_interfaces so phylink knows that this is not supported by the hardware. Move this check to mtk_add_mac().
--- drivers/net/ethernet/mediatek/mtk_eth_soc.c | 21 ++++++++++----------- 1 file changed, 10 insertions(+), 11 deletions(-)
drivers/net/ethernet/mediatek/mtk_eth_soc.c +10 -11
@@ -374,17 +374,6 @@ static int mt7621_gmac0_rgmii_adjust(struct mtk_eth *eth, {u32 val; -/* Check DDR memory type. -* Currently TRGMII mode with DDR2 memory is not supported. -*/ -regmap_read (eth->ethsys,ETHSYS_SYSCFG , &val); -if (interface ==PHY_INTERFACE_MODE_TRGMII && - val &SYSCFG_DRAM_TYPE_DDR2 ) { -dev_err (eth->dev, -"TRGMII mode with DDR2 memory is not supported!\n" ); -return -EOPNOTSUPP ; - } - val = (interface ==PHY_INTERFACE_MODE_TRGMII ) ?ETHSYS_TRGMII_MT7621_DDR_PLL :0 ;
@@ -4333,6 +4322,7 @@ static int mtk_add_mac(struct mtk_eth *eth, struct device_node *np)struct mtk_mac *mac;int id, err;int txqs =1 ; +u32 val;if (!_id) {dev_err (eth->dev,"missing mac id\n" );
@@ -4409,6 +4399,15 @@ static int mtk_add_mac(struct mtk_eth *eth, struct device_node *np)__set_bit (PHY_INTERFACE_MODE_TRGMII , mac->phylink_config.supported_interfaces); +/* TRGMII is not permitted on MT7621 if using DDR2 */ +if (MTK_HAS_CAPS (mac->hw->soc->caps,MTK_GMAC1_TRGMII ) && +MTK_HAS_CAPS (mac->hw->soc->caps,MTK_TRGMII_MT7621_CLK )) { +regmap_read (eth->ethsys,ETHSYS_SYSCFG , &val); +if (val &SYSCFG_DRAM_TYPE_DDR2 ) +__clear_bit (PHY_INTERFACE_MODE_TRGMII , + mac->phylink_config.supported_interfaces); + } +if (MTK_HAS_CAPS (mac->hw->soc->caps,MTK_SGMII )) {__set_bit (PHY_INTERFACE_MODE_SGMII , mac->phylink_config.supported_interfaces);
Signature · 1 line
2.30.2
Re: [PATCH RFC net-next 2/4] net: mtk_eth_soc: move trgmii ddr2 check to probe function
Quoted 6 lines · Russell King (Oracle)
If TRGMII mode is not permitted when using DDR2 mode, we should handle that when setting up phylink's ->supported_interfaces so phylink knows that this is not supported by the hardware. Move this check to
mtk_add_mac().Signed-off-by Russell King (Oracle) <[email protected]>
RFC PATCH 3/4 net: mtk_eth_soc: remove unnecessary checks in mtk_mac_config()
In reply to Russell King (Oracle)mtk_mac_config() checks that the interface mode is permitted for the capabilities, but we already do these checks in mtk_add_mac() when initialising phylink's supported_interfaces bitmap. Remove the unnecessary tests.
--- drivers/net/ethernet/mediatek/mtk_eth_soc.c | 14 +++----------- 1 file changed, 3 insertions(+), 11 deletions(-)
drivers/net/ethernet/mediatek/mtk_eth_soc.c +3 -11
@@ -458,12 +458,6 @@ static void mtk_mac_config(struct phylink_config *config, unsigned int mode,/* Setup soc pin functions */ switch (state->interface) {case PHY_INTERFACE_MODE_TRGMII : -if (mac->id) - goto err_phy; -if (!MTK_HAS_CAPS (mac->hw->soc->caps, -MTK_GMAC1_TRGMII )) - goto err_phy; - fallthrough;case PHY_INTERFACE_MODE_RGMII_TXID :case PHY_INTERFACE_MODE_RGMII_RXID :case PHY_INTERFACE_MODE_RGMII_ID :
@@ -480,11 +474,9 @@ static void mtk_mac_config(struct phylink_config *config, unsigned int mode,case PHY_INTERFACE_MODE_1000BASEX :case PHY_INTERFACE_MODE_2500BASEX :case PHY_INTERFACE_MODE_SGMII : -if (MTK_HAS_CAPS (eth->soc->caps,MTK_SGMII )) { - err =mtk_gmac_sgmii_path_setup (eth, mac->id); -if (err) - goto init_err; - } + err =mtk_gmac_sgmii_path_setup (eth, mac->id); +if (err) + goto init_err;break ;case PHY_INTERFACE_MODE_GMII :if (MTK_HAS_CAPS (eth->soc->caps,MTK_GEPHY )) {
Signature · 1 line
2.30.2
Re: [PATCH RFC net-next 3/4] net: mtk_eth_soc: remove unnecessary checks in mtk_mac_config()
Quoted 6 lines · Russell King (Oracle)
mtk_mac_config()checks that the interface mode is permitted for the capabilities, but we already do these checks inmtk_add_mac()when initialising phylink'ssupported_interfacesbitmap. Remove the unnecessary tests.Signed-off-by Russell King (Oracle) <[email protected]>
RFC PATCH 4/4 net: mtk_eth_soc: note interface modes not set in supported_interfaces
In reply to Russell King (Oracle)--- drivers/net/ethernet/mediatek/mtk_eth_soc.c | 2 ++ 1 file changed, 2 insertions(+)
drivers/net/ethernet/mediatek/mtk_eth_soc.c +2 -0
@@ -530,9 +530,11 @@ static void mtk_mac_config(struct phylink_config *config, unsigned int mode,case PHY_INTERFACE_MODE_GMII : ge_mode =1 ;break ; +/* FIXME: RMII is not set in the phylink capabilities */ case PHY_INTERFACE_MODE_REVMII : ge_mode =2 ;break ; +/* FIXME: RMII is not set in the phylink capabilities */ case PHY_INTERFACE_MODE_RMII :if (mac->id) goto err_phy;
Signature · 1 line
2.30.2
Re: [PATCH RFC net-next 4/4] net: mtk_eth_soc: note interface modes not set in supported_interfaces
On Fri, Feb 24, 2023 at 12:36:26PM +0000, Russell King (Oracle) wrote:
Hi Russell,
I think it would be good to add a patch description here.
Code change looks good to me.
Quoted 4 lines
Signed-off-by Russell King (Oracle) <[email protected]>--- drivers/net/ethernet/mediatek/mtk_eth_soc.c | 2 ++ 1 file changed, 2 insertions(+)
...
Re: [PATCH RFC net-next 4/4] net: mtk_eth_soc: note interface modes not set in supported_interfaces
Quoted 7 lines · Simon Horman
On Fri, Feb 24, 2023 at 12:36:26PM +0000, Russell King (Oracle) wrote:
Hi Russell,
I think it would be good to add a patch description here.
Code change looks good to me.
As noted in the cover message, this is to highlight the issue to hopefully get folk to think what we should do about RMII and REVMII in this driver - basically, should we continue to support them, or remove it completely.
Either way, this patch won't hit net-next in its current form.
Essentially, the choice is either we remove these two switch cases, or we add these interface modes to the supported_interfaces array.
I'd rather those with mtk_eth_soc made the decision, even though it is highly unlikely that these modes are used on the hardware they have - as I don't have any mediatek hardware.
Thanks.
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Re: [PATCH RFC net-next 4/4] net: mtk_eth_soc: note interface modes not set in supported_interfaces
Quoted 16 lines · Russell King (Oracle)
Quoted 7 lines · Simon Horman
On Fri, Feb 24, 2023 at 12:36:26PM +0000, Russell King (Oracle) wrote:
Hi Russell,
I think it would be good to add a patch description here.
Code change looks good to me.
As noted in the cover message, this is to highlight the issue to hopefully get folk to think what we should do about RMII and REVMII in this driver - basically, should we continue to support them, or remove it completely.
Either way, this patch won't hit net-next in its current form.
Quoted 8 lines
Essentially, the choice is either we remove these two switch cases, or we add these interface modes to the
supported_interfacesarray.I'd rather those with
mtk_eth_socmade the decision, even though it is highly unlikely that these modes are used on the hardware they have - as I don't have any mediatek hardware.Thanks.
Thanks, understood.
Sorry for missing this earlier.
Re: [PATCH RFC net-next 4/4] net: mtk_eth_soc: note interface modes not set in supported_interfaces
In reply to Russell King (Oracle)To any of those with Mediatek hardware... Any opinions on what we should do concerning these modes?
If I don't get a reply, then I'll delete these. There's no point having this code if it's been unreachable for a while and no one's complained.
Thanks.
Quoted 25 lines · Russell King (Oracle)
Signed-off-by Russell King (Oracle) <[email protected]>--- drivers/net/ethernet/mediatek/mtk_eth_soc.c | 2 ++ 1 file changed, 2 insertions(+)drivers/net/ethernet/mediatek/mtk_eth_soc.c +2
-0@@ -530,9 +530,11 @@ static void mtk_mac_config(struct phylink_config *config, unsigned int mode,case PHY_INTERFACE_MODE_GMII : ge_mode =1 ;break ; +/* FIXME: RMII is not set in the phylink capabilities */ case PHY_INTERFACE_MODE_REVMII : ge_mode =2 ;break ; +/* FIXME: RMII is not set in the phylink capabilities */ case PHY_INTERFACE_MODE_RMII :if (mac->id) goto err_phy;Signature · 1 line
2.30.2
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Re: [PATCH RFC net-next 4/4] net: mtk_eth_soc: note interface modes not set in supported_interfaces
Quoted 5 lines · Russell King (Oracle)
To any of those with Mediatek hardware... Any opinions on what we should do concerning these modes?
If I don't get a reply, then I'll delete these. There's no point having this code if it's been unreachable for a while and no one's complained.
A quick grep through the device trees of the more than 650 ramips and mediatek boards we support in OpenWrt has revealed that *none* of them uses either reduced-MII or reverse-MII PHY modes. I could imaging that some more specialized ramips boards may use the RMII 100M PHY mode to connect with exotic PHYs for industrial or automotive applications (think: for 100BASE-T1 PHY connected via RMII). I have never seen or touched such boards, but there are hints that they do exist.
For reverse-MII there are cases in which the Ralink SoC (Rt305x, for example) is used in iNIC mode, ie. connected as a PHY to another SoC, and running only a minimal firmware rather than running Linux. Due to the lack of external DRAM for the Ralink SoC on this kind of boards, the Ralink SoC there will anyway never be able to boot Linux. I've seen this e.g. in multimedia devices like early WiFi-connected not-yet-so-smart TVs.
Tl;dr: I'd drop them. If anyone really needs them, it would be easy to add them again and then also add them to the phylink capability mask.
Quoted 33 lines
Thanks.
Quoted 25 lines · Russell King (Oracle)
Signed-off-by Russell King (Oracle) <[email protected]>--- drivers/net/ethernet/mediatek/mtk_eth_soc.c | 2 ++ 1 file changed, 2 insertions(+)drivers/net/ethernet/mediatek/mtk_eth_soc.c +2
-0@@ -530,9 +530,11 @@ static void mtk_mac_config(struct phylink_config *config, unsigned int mode,case PHY_INTERFACE_MODE_GMII : ge_mode =1 ;break ; +/* FIXME: RMII is not set in the phylink capabilities */ case PHY_INTERFACE_MODE_REVMII : ge_mode =2 ;break ; +/* FIXME: RMII is not set in the phylink capabilities */ case PHY_INTERFACE_MODE_RMII :if (mac->id) goto err_phy;Signature · 1 line
2.30.2Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Re: [PATCH RFC net-next 4/4] net: mtk_eth_soc: note interface modes not set in supported_interfaces
Quoted 16 lines · Daniel Golle
A quick grep through the device trees of the more than 650 ramips and mediatek boards we support in OpenWrt has revealed that *none* of them uses either reduced-MII or reverse-MII PHY modes. I could imaging that some more specialized ramips boards may use the RMII 100M PHY mode to connect with exotic PHYs for industrial or automotive applications (think: for 100BASE-T1 PHY connected via RMII). I have never seen or touched such boards, but there are hints that they do exist.
For reverse-MII there are cases in which the Ralink SoC (Rt305x, for example) is used in iNIC mode, ie. connected as a PHY to another SoC, and running only a minimal firmware rather than running Linux. Due to the lack of external DRAM for the Ralink SoC on this kind of boards, the Ralink SoC there will anyway never be able to boot Linux. I've seen this e.g. in multimedia devices like early WiFi-connected not-yet-so-smart TVs.
Quoted 2 lines
Tl;dr: I'd drop them. If anyone really needs them, it would be easy to add them again and then also add them to the phylink capability mask.
Thanks! That seems to be well reasoned. Would you have any objection to using the above as part of the commit message removing these modes?
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Re: [PATCH RFC net-next 4/4] net: mtk_eth_soc: note interface modes not set in supported_interfaces
Quoted 17 lines · Russell King (Oracle)
Quoted 16 lines · Daniel Golle
A quick grep through the device trees of the more than 650 ramips and mediatek boards we support in OpenWrt has revealed that *none* of them uses either reduced-MII or reverse-MII PHY modes. I could imaging that some more specialized ramips boards may use the RMII 100M PHY mode to connect with exotic PHYs for industrial or automotive applications (think: for 100BASE-T1 PHY connected via RMII). I have never seen or touched such boards, but there are hints that they do exist.
For reverse-MII there are cases in which the Ralink SoC (Rt305x, for example) is used in iNIC mode, ie. connected as a PHY to another SoC, and running only a minimal firmware rather than running Linux. Due to the lack of external DRAM for the Ralink SoC on this kind of boards, the Ralink SoC there will anyway never be able to boot Linux. I've seen this e.g. in multimedia devices like early WiFi-connected not-yet-so-smart TVs.
Quoted 5 lines
Quoted 2 lines
Tl;dr: I'd drop them. If anyone really needs them, it would be easy to add them again and then also add them to the phylink capability mask.
Thanks! That seems to be well reasoned. Would you have any objection to using the above as part of the commit message removing these modes?
Sure, go ahead, sounds good to me.