On Fri, 27 Oct 2023 at 18:27, Caleb Connolly caleb.connolly@linaro.org wrote:
[...]
diff --git a/arch/arm/mach-ipq40xx/clock-ipq4019.c b/drivers/clk/qcom/clock-ipq4019.c similarity index 56% rename from arch/arm/mach-ipq40xx/clock-ipq4019.c rename to drivers/clk/qcom/clock-ipq4019.c index c1d5c4ecdd81..04c99964df15 100644 --- a/arch/arm/mach-ipq40xx/clock-ipq4019.c +++ b/drivers/clk/qcom/clock-ipq4019.c @@ -12,12 +12,9 @@
[...]
-static const struct udevice_id msm_clk_ids[] = {
{ .compatible = "qcom,gcc-ipq4019" },This compatible should be moved to "clock-qcom.c".
Should all of the clock drivers have their udevice_id tables in clock-qcom.c or just this one?
Actually this comment was with respect to this patch only being incomplete in itself. You are trying to reuse compatibles defined in clock-qcom.c while moving this driver to drivers/clk/qcom/.
I am fine with the later patch to have udevice_id table in corresponding SoC specific driver. There are lots of moving parts in this series :).
-Sumit
I'll note that I forgot to adjust the U_BOOT_DRIVER definition below to be apq4019 specific, it shouldn't say "clk_msm".
-Sumit
{ }-};
-U_BOOT_DRIVER(clk_msm) = {
.name = "clk_msm",.id = UCLASS_CLK,.of_match = msm_clk_ids,.ops = &msm_clk_ops,.priv_auto = sizeof(struct msm_clk_priv),.probe = msm_clk_probe,-};
-- 2.42.0
-- // Caleb (they/them)