RE: [PATCH net-next 1/2] rtase: Add support for RTL907XD-VA PCIe port

[Justin Lai <justinlai0215@realtek.com>]

> 
> On Mon, Nov 11, 2024 at 10:55:31AM +0800, Justin Lai wrote:
> > Add RTL907XD-VA hardware version and modify the speed reported by
> > .get_link_ksettings in ethtool_ops.
> >
> > Signed-off-by: Justin Lai <justinlai0215@realtek.com>
> 
> Hi Justin,
> 
> this seems to be doing several things:
> 
> 1) Adding defines for existing values
> 2) Correcting the speed for RTL907XD-V1
> 3) Adding support for RTL907XD-VA
> 
> I think these would be best handled as 3 patches.
> And I wonder if 2) is a bug fix for net rather than an enhancement for net-next.

Ok, I'll try to break down the patch into more detailed parts to 
make it clearer.
> 
> > ---
> >  drivers/net/ethernet/realtek/rtase/rtase.h    | 10 +++++--
> >  .../net/ethernet/realtek/rtase/rtase_main.c   | 26 ++++++++++++++-----
> >  2 files changed, 28 insertions(+), 8 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/realtek/rtase/rtase.h
> > b/drivers/net/ethernet/realtek/rtase/rtase.h
> > index 583c33930f88..2bbfcad613ab 100644
> > --- a/drivers/net/ethernet/realtek/rtase/rtase.h
> > +++ b/drivers/net/ethernet/realtek/rtase/rtase.h
> > @@ -9,7 +9,11 @@
> >  #ifndef RTASE_H
> >  #define RTASE_H
> >
> > -#define RTASE_HW_VER_MASK 0x7C800000
> > +#define RTASE_HW_VER_MASK     0x7C800000
> > +#define RTASE_HW_VER_906X_7XA 0x00800000 #define
> > +RTASE_HW_VER_906X_7XC 0x04000000 #define
> RTASE_HW_VER_907XD_V1
> > +0x04800000 #define RTASE_HW_VER_907XD_VA 0x08000000
> >
> >  #define RTASE_RX_DMA_BURST_256       4
> >  #define RTASE_TX_DMA_BURST_UNLIMITED 7 @@ -170,7 +174,7 @@
> enum
> > rtase_registers {
> >       RTASE_INT_MITI_TX = 0x0A00,
> >       RTASE_INT_MITI_RX = 0x0A80,
> >
> > -     RTASE_VLAN_ENTRY_0     = 0xAC80,
> > +     RTASE_VLAN_ENTRY_0 = 0xAC80,
> 
> This change doesn't seem related to the rest of the patch.

I'll separate this into an additional patch and upload it.
> 
> >  };
> >
> >  enum rtase_desc_status_bit {
> > @@ -327,6 +331,8 @@ struct rtase_private {
> >       u16 int_nums;
> >       u16 tx_int_mit;
> >       u16 rx_int_mit;
> > +
> > +     u32 hw_ver;
> >  };
> >
> >  #define RTASE_LSO_64K 64000
> > diff --git a/drivers/net/ethernet/realtek/rtase/rtase_main.c
> > b/drivers/net/ethernet/realtek/rtase/rtase_main.c
> > index f8777b7663d3..73ebdf0bc376 100644
> > --- a/drivers/net/ethernet/realtek/rtase/rtase_main.c
> > +++ b/drivers/net/ethernet/realtek/rtase/rtase_main.c
> > @@ -1714,10 +1714,22 @@ static int rtase_get_settings(struct net_device
> *dev,
> >                             struct ethtool_link_ksettings *cmd)  {
> >       u32 supported = SUPPORTED_MII | SUPPORTED_Pause |
> > SUPPORTED_Asym_Pause;
> > +     const struct rtase_private *tp = netdev_priv(dev);
> >
> >
> ethtool_convert_legacy_u32_to_link_mode(cmd->link_modes.supported,
> >                                               supported);
> > -     cmd->base.speed = SPEED_5000;
> > +
> > +     switch (tp->hw_ver) {
> > +     case RTASE_HW_VER_906X_7XA:
> > +     case RTASE_HW_VER_906X_7XC:
> > +             cmd->base.speed = SPEED_5000;
> > +             break;
> > +     case RTASE_HW_VER_907XD_V1:
> > +     case RTASE_HW_VER_907XD_VA:
> > +             cmd->base.speed = SPEED_10000;
> > +             break;
> > +     }
> > +
> >       cmd->base.duplex = DUPLEX_FULL;
> >       cmd->base.port = PORT_MII;
> >       cmd->base.autoneg = AUTONEG_DISABLE;
> 
> > @@ -1974,13 +1986,15 @@ static void
> > rtase_init_software_variable(struct pci_dev *pdev,
> >
> >  static bool rtase_check_mac_version_valid(struct rtase_private *tp)
> > {
> > -     u32 hw_ver = rtase_r32(tp, RTASE_TX_CONFIG_0) &
> RTASE_HW_VER_MASK;
> >       bool known_ver = false;
> >
> > -     switch (hw_ver) {
> > -     case 0x00800000:
> > -     case 0x04000000:
> > -     case 0x04800000:
> > +     tp->hw_ver = rtase_r32(tp, RTASE_TX_CONFIG_0) &
> > + RTASE_HW_VER_MASK;
> 
> Now that this is setting tp->hw_ver perhaps the name of the function should be
> changed? Perhaps rtase_set_mac_version() ? Perhaps a single patch can be
> created that reworks this function, preparing for other work, by:
> 
> * Changes the name of the function
> * Sets tp->hw_ver
> * Changes the return type from bool to int
>   (as is currently done as part of patch 2/2)

This function is not simply used to set tp->hw_ver. Its primary purpose
is to validate the MAC version. Since hw_ver is also used elsewhere, it
is stored in tp->hw_ver. Therefore, I don't believe the function name
needs to be changed. However, I will group the remaining two items into
a separate patch and include it in this patch set.
> 
> Although a refactor, perhaps that could be part of a series for net that also
> includes two more patches that depend on it and:
> 
> * Correct the speed for RTL907XD-V1
> * Corrects error handling in the case where the version is invalid
>   (as is currently done as part of patch 2/2)

Thank you for your valuable suggestions. I will upload the three patches
as discussed to the net.
> 
> And then any remaning enhancements can be addressed as follow-up patches
> for net-next.

Ok, I will do that.
> 
> 
> > +
> > +     switch (tp->hw_ver) {
> > +     case RTASE_HW_VER_906X_7XA:
> > +     case RTASE_HW_VER_906X_7XC:
> > +     case RTASE_HW_VER_907XD_V1:
> > +     case RTASE_HW_VER_907XD_VA:
> >               known_ver = true;
> >               break;
> >       }
> > --
> > 2.34.1
> >