Thread (25 messages) 25 messages, 7 authors, 2022-01-26

Re: [PATCH v7 8/8] PCI: imx: Add the imx8mm pcie support

From: Krzysztof Wilczyński <hidden>
Date: 2021-12-16 16:51:57
Also in: linux-arm-kernel, linux-pci, linux-phy, lkml

Hi Richard,

Apologies for a very late review!  Especially since Lorenzo already took
patches as per:

  https://lore.kernel.org/linux-pci/163965080404.20006.5241609551643501749.b4-ty@arm.com/ (local)

However, perhaps it's not too late.

[...]
quoted hunk ↗ jump to hunk
@@ -446,6 +452,13 @@ static int imx6_pcie_enable_ref_clk(struct imx6_pcie *imx6_pcie)
 		break;
 	case IMX7D:
 		break;
+	case IMX8MM:
+		ret = clk_prepare_enable(imx6_pcie->pcie_aux);
+		if (ret) {
+			dev_err(dev, "unable to enable pcie_aux clock\n");
+			break;
+		}
+		break;
You can drop the inner break, it wouldn't do much here, unless this was
intended to be a return?
quoted hunk ↗ jump to hunk
@@ -538,6 +559,10 @@ static void imx6_pcie_deassert_core_reset(struct imx6_pcie *imx6_pcie)
 	case IMX8MQ:
 		reset_control_deassert(imx6_pcie->pciephy_reset);
 		break;
+	case IMX8MM:
+		if (phy_init(imx6_pcie->phy) != 0)
+			dev_err(dev, "Waiting for PHY ready timeout!\n");
+		break;
If the above, you can keep the same style as used throughout the file
already, so it would just simply be:

  if (phy_init(imx6_pcie->phy))

Also, a nitpick: to be consistent with other such messages here, the error
message would be all lower-case letters.

[...]
quoted hunk ↗ jump to hunk
@@ -614,6 +639,8 @@ static void imx6_pcie_configure_type(struct imx6_pcie *imx6_pcie)
 static void imx6_pcie_init_phy(struct imx6_pcie *imx6_pcie)
 {
 	switch (imx6_pcie->drvdata->variant) {
+	case IMX8MM:
+		break;
 	case IMX8MQ:
Would it warrant a comment that adds a note there to this single bare
break?  Perhaps this version is not support, lack this particular
functionality, etc.

[...]
quoted hunk ↗ jump to hunk
@@ -1089,10 +1122,39 @@ static int imx6_pcie_probe(struct platform_device *pdev)
 			dev_err(dev, "Failed to get PCIE APPS reset control\n");
 			return PTR_ERR(imx6_pcie->apps_reset);
 		}
+		break;
+	case IMX8MM:
+		imx6_pcie->pcie_aux = devm_clk_get(dev, "pcie_aux");
+		if (IS_ERR(imx6_pcie->pcie_aux))
+			return dev_err_probe(dev, PTR_ERR(imx6_pcie->pcie_aux),
+					     "pcie_aux clock source missing or invalid\n");
+		imx6_pcie->apps_reset = devm_reset_control_get_exclusive(dev,
+									 "apps");
+		if (IS_ERR(imx6_pcie->apps_reset)) {
+			dev_err(dev, "Failed to get PCIE APPS reset control\n");
+			return PTR_ERR(imx6_pcie->apps_reset);
+		}
+
+		imx6_pcie->phy = devm_phy_get(dev, "pcie-phy");
+		if (IS_ERR(imx6_pcie->phy)) {
+			if (PTR_ERR(imx6_pcie->phy) == -EPROBE_DEFER)
+				return -EPROBE_DEFER;
+			dev_err(dev, "Failed to get PCIE PHY\n");
+			return PTR_ERR(imx6_pcie->phy);
+		}
A question about handling of the -EPROBE_DEFER above: why not to use the
dev_err_probe() helper similarly to the code above and below?  Would there
be something different preventing the use of dev_err_probe() here too?
 		break;
 	default:
 		break;
 	}
+	/* Don't fetch the pcie_phy clock, if it has abstract PHY driver */
+	if (imx6_pcie->phy == NULL) {
+		imx6_pcie->pcie_phy = devm_clk_get(dev, "pcie_phy");
+		if (IS_ERR(imx6_pcie->pcie_phy))
+			return dev_err_probe(dev, PTR_ERR(imx6_pcie->pcie_phy),
+					     "pcie_phy clock source missing or invalid\n");
+	}
Thank you for another amazing patch!

	Krzysztof
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help