Hi Russell,

On Thu, Sep 4, 2025 at 10:31 PM Russell King (Oracle)
[off-list ref] wrote:

On Thu, Sep 04, 2025 at 10:10:32PM +0100, Lad, Prabhakar wrote:

quoted

Hi Russell,

On Thu, Sep 4, 2025 at 9:49 PM Russell King (Oracle)
[off-list ref] wrote:

quoted

On Thu, Sep 04, 2025 at 09:39:48PM +0100, Prabhakar wrote:

quoted

      plat_dat->init = renesas_gbeth_init;
      plat_dat->exit = renesas_gbeth_exit;
-     plat_dat->flags |= STMMAC_FLAG_HWTSTAMP_CORRECT_LATENCY |
-                        STMMAC_FLAG_EN_TX_LPI_CLK_PHY_CAP |
-                        STMMAC_FLAG_SPH_DISABLE;
+     plat_dat->flags |= gbeth->of_data->stmmac_flags;

You include the first two flags in your new device. I would like to see
at least STMMAC_FLAG_EN_TX_LPI_CLK_PHY_CAP always being set. The only
reason we have the STMMAC_FLAG_EN_TX_LPI_CLK_PHY_CAP flag is to avoid
changing existing behaviour and causing regressions. New stuff should
always set this.

Me confused, STMMAC_FLAG_EN_TX_LPI_CLK_PHY_CAP flag is set in the new
device [0]. The reason STMMAC_FLAG_SPH_DISABLE flag being dropped in
the new device is SPHEN=1 in MAC HW feature reg for the new device.

What I'm saying is I'd like to see:

        plat_dat->flags |= STMMAC_FLAG_EN_TX_LPI_CLK_PHY_CAP |
                           gbeth->of_data->stmmac_flags;

iow, it is set unconditionally, even if forgotten in a future patch.

Ah got you. Thank you for the clarification.

Cheers,
Prabhakar