On Fri, Nov 7, 2025 at 12:00 PM Konrad Dybcio
[off-list ref] wrote:

On 11/7/25 11:29 AM, Bartosz Golaszewski wrote:

quoted

From: Bartosz Golaszewski <redacted>

With match data split into general and power-management sections, let's
now do the same with runtime device data.

Signed-off-by: Bartosz Golaszewski <redacted>
---
 .../ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c    | 46 ++++++++++++----------
 1 file changed, 25 insertions(+), 21 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
index 1f00556bbad997e2ec76b521cffe2eb14fabb79e..09f122062dec87aa11804af2769ddff4964e6596 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-qcom-ethqos.c

@@ -105,17 +105,21 @@ struct ethqos_emac_match_data {
      const struct ethqos_emac_pm_data *pm_data;
 };

+struct ethqos_emac_pm_ctx {
+     struct clk *link_clk;
+     unsigned int link_clk_rate;
+     struct phy *serdes_phy;

What is the benefit of doing this? PHY APIs happily consume a nullptr
and NOP out, and the PHY is already retrieved with _optional(),
similarly with clk

Konrad

Because it clearly divides the driver's logic into the manual and
firmware-driven variants. Just because we could, doesn't necessarily
mean we should just call PHY APIs with a nullptr if readability is
better when we don't.

Bartosz