RE: [RESEND v2 4/5] PCI: imx6: Fix the clock reference handling unbalance when link never came up

-----Original Message-----
From: Lucas Stach <l.stach@pengutronix.de>
Sent: Saturday, October 16, 2021 2:25 AM
To: Richard Zhu <hongxing.zhu@nxp.com>; bhelgaas@google.com;
lorenzo.pieralisi@arm.com
Cc: linux-pci@vger.kernel.org; dl-linux-imx <redacted>;
linux-arm-kernel@lists.infradead.org; linux-kernel@vger.kernel.org;
kernel@pengutronix.de
Subject: Re: [RESEND v2 4/5] PCI: imx6: Fix the clock reference handling
unbalance when link never came up

Am Freitag, dem 15.10.2021 um 14:05 +0800 schrieb Richard Zhu:

quoted

When link never came up, driver probe would be failed with error -110.
To keep usage counter balance of the clocks, disable the previous
enabled clocks when link is down.
Move definitions of the imx6_pcie_clk_disable() function to the proper
place. Because it wouldn't be used in imx6_pcie_suspend_noirq() only.

Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
---
 drivers/pci/controller/dwc/pci-imx6.c | 47
++++++++++++++-------------
 1 file changed, 24 insertions(+), 23 deletions(-)

diff --git a/drivers/pci/controller/dwc/pci-imx6.c

b/drivers/pci/controller/dwc/pci-imx6.c
index cc837f8bf6d4..d6a5d99ffa52 100644

--- a/drivers/pci/controller/dwc/pci-imx6.c
+++ b/drivers/pci/controller/dwc/pci-imx6.c

@@ -514,6 +514,29 @@ static int imx6_pcie_clk_enable(struct imx6_pcie

*imx6_pcie)

quoted

 	return ret;
 }

+static void imx6_pcie_clk_disable(struct imx6_pcie *imx6_pcie) {
+	clk_disable_unprepare(imx6_pcie->pcie);
+	clk_disable_unprepare(imx6_pcie->pcie_phy);
+	clk_disable_unprepare(imx6_pcie->pcie_bus);
+
+	switch (imx6_pcie->drvdata->variant) {
+	case IMX6SX:
+		clk_disable_unprepare(imx6_pcie->pcie_inbound_axi);
+		break;
+	case IMX7D:
+		regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR12,
+				   IMX7D_GPR12_PCIE_PHY_REFCLK_SEL,
+				   IMX7D_GPR12_PCIE_PHY_REFCLK_SEL);
+		break;
+	case IMX8MQ:
+		clk_disable_unprepare(imx6_pcie->pcie_aux);
+		break;
+	default:
+		break;
+	}
+}
+
 static void imx7d_pcie_wait_for_phy_pll_lock(struct imx6_pcie
*imx6_pcie)  {
 	u32 val;

@@ -853,6 +876,7 @@ static int imx6_pcie_start_link(struct dw_pcie *pci)
 		dw_pcie_readl_dbi(pci, PCIE_PORT_DEBUG0),
 		dw_pcie_readl_dbi(pci, PCIE_PORT_DEBUG1));
 	imx6_pcie_reset_phy(imx6_pcie);
+	imx6_pcie_clk_disable(imx6_pcie);

Same comment as with the previous patch. We should not cram in more error
handling in the imx6_pcie_start_link function, but rather move out all the
error handling to be after dw_pcie_host_init. Even the already existing phy
reset here seems misplaced and should be moved out.

[Richard Zhu] About the clk disabled, please see my explains under the addressed comments in the #3 patch.
Agree to move the existing phy reset to the error handling after dw_pcie_host_init.
Thanks.
BR
Richard

Regards,
Lucas

quoted

 	if (imx6_pcie->vpcie && regulator_is_enabled(imx6_pcie->vpcie) > 0)
 		regulator_disable(imx6_pcie->vpcie);
 	return ret;

@@ -941,29 +965,6 @@ static void imx6_pcie_pm_turnoff(struct

imx6_pcie *imx6_pcie)

quoted

 	usleep_range(1000, 10000);
 }

-static void imx6_pcie_clk_disable(struct imx6_pcie *imx6_pcie) -{
-	clk_disable_unprepare(imx6_pcie->pcie);
-	clk_disable_unprepare(imx6_pcie->pcie_phy);
-	clk_disable_unprepare(imx6_pcie->pcie_bus);
-
-	switch (imx6_pcie->drvdata->variant) {
-	case IMX6SX:
-		clk_disable_unprepare(imx6_pcie->pcie_inbound_axi);
-		break;
-	case IMX7D:
-		regmap_update_bits(imx6_pcie->iomuxc_gpr, IOMUXC_GPR12,
-				   IMX7D_GPR12_PCIE_PHY_REFCLK_SEL,
-				   IMX7D_GPR12_PCIE_PHY_REFCLK_SEL);
-		break;
-	case IMX8MQ:
-		clk_disable_unprepare(imx6_pcie->pcie_aux);
-		break;
-	default:
-		break;
-	}
-}
-
 static int imx6_pcie_suspend_noirq(struct device *dev)  {
 	struct imx6_pcie *imx6_pcie = dev_get_drvdata(dev);