Patchwork [v2,3/3] PCI: imx: Add workaround for e10728, IMX7d PCIe PLL failure

login
register
mail settings
Submitter Trent Piepho
Date Feb. 5, 2019, 12:17 a.m.
Message ID <20190205001721.18639-4-tpiepho@impinj.com>
Download mbox | patch
Permalink /patch/717987/
State New
Headers show

Comments

Trent Piepho - Feb. 5, 2019, 12:17 a.m.
This implements the workound described in the NXP IMX7d erratum e10728.

Initial VCO oscillation may fail under corner conditions such as cold
temperature.  It causes PCIe PLL to fail to lock in the initialization
phase, which results in the PCIe link failing to come up.

The workaround is to disable Duty-cycle Corrector (DCC) calibration
after G_RST.

To do this it is necessary to gain access to the undocumented and
currently unused PCIe PHY register bank.  A new device tree node of type
"fsl,imx7d-pcie-phy" is created for the PHY block and the existing PCIe
device uses a phandle named "fsl,imx7d-pcie-phy" to point to it.

Signed-off-by: Trent Piepho <tpiepho@impinj.com>
---
 drivers/pci/controller/dwc/pci-imx6.c | 57 +++++++++++++++++++++++++++++++++++
 1 file changed, 57 insertions(+)
Lucas Stach - Feb. 5, 2019, 9:48 a.m.
Am Dienstag, den 05.02.2019, 00:17 +0000 schrieb Trent Piepho:
> This implements the workound described in the NXP IMX7d erratum e10728.
> 
> Initial VCO oscillation may fail under corner conditions such as cold
> temperature.  It causes PCIe PLL to fail to lock in the initialization
> phase, which results in the PCIe link failing to come up.
> 
> The workaround is to disable Duty-cycle Corrector (DCC) calibration
> after G_RST.
> 
> To do this it is necessary to gain access to the undocumented and
> currently unused PCIe PHY register bank.  A new device tree node of type
> "fsl,imx7d-pcie-phy" is created for the PHY block and the existing PCIe
> device uses a phandle named "fsl,imx7d-pcie-phy" to point to it.
> 
> Signed-off-by: Trent Piepho <tpiepho@impinj.com>

Reviewed-by: Lucas Stach <l.stach@pengutronix.de>

> ---
>  drivers/pci/controller/dwc/pci-imx6.c | 57 +++++++++++++++++++++++++++++++++++
>  1 file changed, 57 insertions(+)
> 
> diff --git a/drivers/pci/controller/dwc/pci-imx6.c b/drivers/pci/controller/dwc/pci-imx6.c
> index 80f843030e36..06dd6aa927d4 100644
> --- a/drivers/pci/controller/dwc/pci-imx6.c
> +++ b/drivers/pci/controller/dwc/pci-imx6.c
> @@ -18,6 +18,7 @@
>  #include <linux/module.h>
>  #include <linux/of_gpio.h>
>  #include <linux/of_device.h>
> +#include <linux/of_address.h>
>  #include <linux/pci.h>
>  #include <linux/platform_device.h>
>  #include <linux/regmap.h>
> @@ -61,6 +62,7 @@ struct imx6_pcie {
> > >  	u32			tx_swing_low;
> > >  	int			link_gen;
> > >  	struct regulator	*vpcie;
> > > +	void __iomem		*phy_base;
>  
> >  	/* power domain for pcie */
> > >  	struct device		*pd_pcie;
> @@ -117,6 +119,23 @@ struct imx6_pcie {
>  #define PCIE_PHY_RX_ASIC_OUT 0x100D
> >  #define PCIE_PHY_RX_ASIC_OUT_VALID	(1 << 0)
>  
> +/* iMX7 PCIe PHY registers */
> > +#define PCIE_PHY_CMN_REG4		0x14
> +/* These are probably the bits that *aren't* DCC_FB_EN */
> > +#define PCIE_PHY_CMN_REG4_DCC_FB_EN	0x29
> +
> > +#define PCIE_PHY_CMN_REG15	        0x54
> > +#define PCIE_PHY_CMN_REG15_DLY_4	BIT(2)
> > +#define PCIE_PHY_CMN_REG15_PLL_PD	BIT(5)
> > +#define PCIE_PHY_CMN_REG15_OVRD_PLL_PD	BIT(7)
> +
> > +#define PCIE_PHY_CMN_REG24		0x90
> > +#define PCIE_PHY_CMN_REG24_RX_EQ	BIT(6)
> > +#define PCIE_PHY_CMN_REG24_RX_EQ_SEL	BIT(3)
> +
> > +#define PCIE_PHY_CMN_REG26		0x98
> > +#define PCIE_PHY_CMN_REG26_ATT_MODE	0xBC
> +
>  #define PHY_RX_OVRD_IN_LO 0x1005
>  #define PHY_RX_OVRD_IN_LO_RX_DATA_EN (1 << 5)
>  #define PHY_RX_OVRD_IN_LO_RX_PLL_EN (1 << 3)
> @@ -490,6 +509,26 @@ static void imx6_pcie_deassert_core_reset(struct imx6_pcie *imx6_pcie)
> >  	switch (imx6_pcie->variant) {
> >  	case IMX7D:
> >  		reset_control_deassert(imx6_pcie->pciephy_reset);
> +
> > +		/* Workaround for ERR010728, failure of PCI-e PLL VCO to
> > +		 * oscillate, especially when cold.  This turns off "Duty-cycle
> > +		 * Corrector" and other mysterious undocumented things.
> > +		 */
> > +		if (likely(imx6_pcie->phy_base)) {
> > +			/* De-assert DCC_FB_EN */
> > +			writel(PCIE_PHY_CMN_REG4_DCC_FB_EN,
> > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG4);
> > +			/* Assert RX_EQS and RX_EQS_SEL */
> > +			writel(PCIE_PHY_CMN_REG24_RX_EQ_SEL
> > +				| PCIE_PHY_CMN_REG24_RX_EQ,
> > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG24);
> > +			/* Assert ATT_MODE */
> > +			writel(PCIE_PHY_CMN_REG26_ATT_MODE,
> > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG26);
> > +		} else {
> > +			dev_warn(dev, "DT lacks imx7d-pcie-phy, unable to apply ERR010728 workaround\n");
> > +		}
> +
> >  		imx7d_pcie_wait_for_phy_pll_lock(imx6_pcie);
> >  		break;
> >  	case IMX6SX:
> @@ -919,6 +958,7 @@ static int imx6_pcie_probe(struct platform_device *pdev)
> >  	struct device *dev = &pdev->dev;
> >  	struct dw_pcie *pci;
> >  	struct imx6_pcie *imx6_pcie;
> > +	struct device_node *np;
> >  	struct resource *dbi_base;
> >  	struct device_node *node = dev->of_node;
> >  	int ret;
> @@ -939,6 +979,23 @@ static int imx6_pcie_probe(struct platform_device *pdev)
> >  	imx6_pcie->variant =
> >  		(enum imx6_pcie_variants)of_device_get_match_data(dev);
>  
> > +	/* Find the PHY if one is defined, only imx7d uses it */
> > +	np = of_parse_phandle(node, "fsl,imx7d-pcie-phy", 0);
> > +	if (np) {
> > +		struct resource res;
> +
> > +		ret = of_address_to_resource(np, 0, &res);
> > +		if (ret) {
> > +			dev_err(dev, "Unable to map PCIe PHY\n");
> > +			return ret;
> > +		}
> > +		imx6_pcie->phy_base = devm_ioremap_resource(dev, &res);
> > +		if (IS_ERR(imx6_pcie->phy_base)) {
> > +			dev_err(dev, "Unable to map PCIe PHY\n");
> > +			return PTR_ERR(imx6_pcie->phy_base);
> > +		}
> > +	}
> +
> >  	dbi_base = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> >  	pci->dbi_base = devm_ioremap_resource(dev, dbi_base);
> >  	if (IS_ERR(pci->dbi_base))
Lorenzo Pieralisi - Feb. 7, 2019, 12:31 p.m.
On Tue, Feb 05, 2019 at 12:17:41AM +0000, Trent Piepho wrote:
> This implements the workound described in the NXP IMX7d erratum e10728.
> 
> Initial VCO oscillation may fail under corner conditions such as cold
> temperature.  It causes PCIe PLL to fail to lock in the initialization
> phase, which results in the PCIe link failing to come up.
> 
> The workaround is to disable Duty-cycle Corrector (DCC) calibration
> after G_RST.
> 
> To do this it is necessary to gain access to the undocumented and
> currently unused PCIe PHY register bank.  A new device tree node of type
> "fsl,imx7d-pcie-phy" is created for the PHY block and the existing PCIe
> device uses a phandle named "fsl,imx7d-pcie-phy" to point to it.
> 
> Signed-off-by: Trent Piepho <tpiepho@impinj.com>
> ---
>  drivers/pci/controller/dwc/pci-imx6.c | 57 +++++++++++++++++++++++++++++++++++
>  1 file changed, 57 insertions(+)
> 
> diff --git a/drivers/pci/controller/dwc/pci-imx6.c b/drivers/pci/controller/dwc/pci-imx6.c
> index 80f843030e36..06dd6aa927d4 100644
> --- a/drivers/pci/controller/dwc/pci-imx6.c
> +++ b/drivers/pci/controller/dwc/pci-imx6.c
> @@ -18,6 +18,7 @@
>  #include <linux/module.h>
>  #include <linux/of_gpio.h>
>  #include <linux/of_device.h>
> +#include <linux/of_address.h>
>  #include <linux/pci.h>
>  #include <linux/platform_device.h>
>  #include <linux/regmap.h>
> @@ -61,6 +62,7 @@ struct imx6_pcie {
>  	u32			tx_swing_low;
>  	int			link_gen;
>  	struct regulator	*vpcie;
> +	void __iomem		*phy_base;
>  
>  	/* power domain for pcie */
>  	struct device		*pd_pcie;
> @@ -117,6 +119,23 @@ struct imx6_pcie {
>  #define PCIE_PHY_RX_ASIC_OUT 0x100D
>  #define PCIE_PHY_RX_ASIC_OUT_VALID	(1 << 0)
>  
> +/* iMX7 PCIe PHY registers */
> +#define PCIE_PHY_CMN_REG4		0x14
> +/* These are probably the bits that *aren't* DCC_FB_EN */
> +#define PCIE_PHY_CMN_REG4_DCC_FB_EN	0x29
> +
> +#define PCIE_PHY_CMN_REG15	        0x54
> +#define PCIE_PHY_CMN_REG15_DLY_4	BIT(2)
> +#define PCIE_PHY_CMN_REG15_PLL_PD	BIT(5)
> +#define PCIE_PHY_CMN_REG15_OVRD_PLL_PD	BIT(7)
> +
> +#define PCIE_PHY_CMN_REG24		0x90
> +#define PCIE_PHY_CMN_REG24_RX_EQ	BIT(6)
> +#define PCIE_PHY_CMN_REG24_RX_EQ_SEL	BIT(3)
> +
> +#define PCIE_PHY_CMN_REG26		0x98
> +#define PCIE_PHY_CMN_REG26_ATT_MODE	0xBC
> +
>  #define PHY_RX_OVRD_IN_LO 0x1005
>  #define PHY_RX_OVRD_IN_LO_RX_DATA_EN (1 << 5)
>  #define PHY_RX_OVRD_IN_LO_RX_PLL_EN (1 << 3)
> @@ -490,6 +509,26 @@ static void imx6_pcie_deassert_core_reset(struct imx6_pcie *imx6_pcie)
>  	switch (imx6_pcie->variant) {
>  	case IMX7D:
>  		reset_control_deassert(imx6_pcie->pciephy_reset);
> +
> +		/* Workaround for ERR010728, failure of PCI-e PLL VCO to
> +		 * oscillate, especially when cold.  This turns off "Duty-cycle
> +		 * Corrector" and other mysterious undocumented things.
> +		 */
> +		if (likely(imx6_pcie->phy_base)) {

Nit: this is not a fast-path, so this branch performance is hardly
relevant.

> +			/* De-assert DCC_FB_EN */
> +			writel(PCIE_PHY_CMN_REG4_DCC_FB_EN,
> +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG4);
> +			/* Assert RX_EQS and RX_EQS_SEL */
> +			writel(PCIE_PHY_CMN_REG24_RX_EQ_SEL
> +				| PCIE_PHY_CMN_REG24_RX_EQ,
> +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG24);
> +			/* Assert ATT_MODE */
> +			writel(PCIE_PHY_CMN_REG26_ATT_MODE,
> +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG26);
> +		} else {
> +			dev_warn(dev, "DT lacks imx7d-pcie-phy, unable to apply ERR010728 workaround\n");

It is a nit but the warning log is not necessarily true ie you may not
have a mapped address for other reasons as well. It is up to you but I
would change the log message.

Thanks,
Lorenzo

> +		}
> +
>  		imx7d_pcie_wait_for_phy_pll_lock(imx6_pcie);
>  		break;
>  	case IMX6SX:
> @@ -919,6 +958,7 @@ static int imx6_pcie_probe(struct platform_device *pdev)
>  	struct device *dev = &pdev->dev;
>  	struct dw_pcie *pci;
>  	struct imx6_pcie *imx6_pcie;
> +	struct device_node *np;
>  	struct resource *dbi_base;
>  	struct device_node *node = dev->of_node;
>  	int ret;
> @@ -939,6 +979,23 @@ static int imx6_pcie_probe(struct platform_device *pdev)
>  	imx6_pcie->variant =
>  		(enum imx6_pcie_variants)of_device_get_match_data(dev);
>  
> +	/* Find the PHY if one is defined, only imx7d uses it */
> +	np = of_parse_phandle(node, "fsl,imx7d-pcie-phy", 0);
> +	if (np) {
> +		struct resource res;
> +
> +		ret = of_address_to_resource(np, 0, &res);
> +		if (ret) {
> +			dev_err(dev, "Unable to map PCIe PHY\n");
> +			return ret;
> +		}
> +		imx6_pcie->phy_base = devm_ioremap_resource(dev, &res);
> +		if (IS_ERR(imx6_pcie->phy_base)) {
> +			dev_err(dev, "Unable to map PCIe PHY\n");
> +			return PTR_ERR(imx6_pcie->phy_base);
> +		}
> +	}
> +
>  	dbi_base = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>  	pci->dbi_base = devm_ioremap_resource(dev, dbi_base);
>  	if (IS_ERR(pci->dbi_base))
> -- 
> 2.14.4
>
Trent Piepho - Feb. 7, 2019, 6:15 p.m.
On Thu, 2019-02-07 at 12:31 +0000, Lorenzo Pieralisi wrote:
> +			/* De-assert DCC_FB_EN */

> > +			writel(PCIE_PHY_CMN_REG4_DCC_FB_EN,

> > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG4);

> > +			/* Assert RX_EQS and RX_EQS_SEL */

> > +			writel(PCIE_PHY_CMN_REG24_RX_EQ_SEL

> > +				| PCIE_PHY_CMN_REG24_RX_EQ,

> > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG24);

> > +			/* Assert ATT_MODE */

> > +			writel(PCIE_PHY_CMN_REG26_ATT_MODE,

> > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG26);

> > +		} else {

> > +			dev_warn(dev, "DT lacks imx7d-pcie-phy, unable to apply ERR010728 workaround\n");

> 

> It is a nit but the warning log is not necessarily true ie you may not

> have a mapped address for other reasons as well. It is up to you but I

> would change the log message.


True, but I hardly want to enumerate every possible failure point in a
log message.  How about:

Unable to apply ERR010728 workaround.  DT lacks imx7d-pcie-phy?
Lorenzo Pieralisi - Feb. 8, 2019, 10:18 a.m.
On Thu, Feb 07, 2019 at 06:15:52PM +0000, Trent Piepho wrote:
> On Thu, 2019-02-07 at 12:31 +0000, Lorenzo Pieralisi wrote:
> > +			/* De-assert DCC_FB_EN */
> > > +			writel(PCIE_PHY_CMN_REG4_DCC_FB_EN,
> > > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG4);
> > > +			/* Assert RX_EQS and RX_EQS_SEL */
> > > +			writel(PCIE_PHY_CMN_REG24_RX_EQ_SEL
> > > +				| PCIE_PHY_CMN_REG24_RX_EQ,
> > > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG24);
> > > +			/* Assert ATT_MODE */
> > > +			writel(PCIE_PHY_CMN_REG26_ATT_MODE,
> > > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG26);
> > > +		} else {
> > > +			dev_warn(dev, "DT lacks imx7d-pcie-phy, unable to apply ERR010728 workaround\n");
> > 
> > It is a nit but the warning log is not necessarily true ie you may not
> > have a mapped address for other reasons as well. It is up to you but I
> > would change the log message.
> 
> True, but I hardly want to enumerate every possible failure point in a
> log message.  How about:
> 
> Unable to apply ERR010728 workaround.  DT lacks imx7d-pcie-phy?

I would just log "Unable to apply ERR010728 workaround" but it is up
to you. I can make the change myself with no need for reposting.

Lorenzo
Trent Piepho - Feb. 8, 2019, 6:19 p.m.
On Fri, 2019-02-08 at 10:18 +0000, Lorenzo Pieralisi wrote:
> On Thu, Feb 07, 2019 at 06:15:52PM +0000, Trent Piepho wrote:

> > On Thu, 2019-02-07 at 12:31 +0000, Lorenzo Pieralisi wrote:

> > > +			/* De-assert DCC_FB_EN */

> > > > +			writel(PCIE_PHY_CMN_REG4_DCC_FB_EN,

> > > > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG4);

> > > > +			/* Assert RX_EQS and RX_EQS_SEL */

> > > > +			writel(PCIE_PHY_CMN_REG24_RX_EQ_SEL

> > > > +				| PCIE_PHY_CMN_REG24_RX_EQ,

> > > > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG24);

> > > > +			/* Assert ATT_MODE */

> > > > +			writel(PCIE_PHY_CMN_REG26_ATT_MODE,

> > > > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG26);

> > > > +		} else {

> > > > +			dev_warn(dev, "DT lacks imx7d-pcie-phy, unable to apply ERR010728 workaround\n");

> > > 

> > > It is a nit but the warning log is not necessarily true ie you may not

> > > have a mapped address for other reasons as well. It is up to you but I

> > > would change the log message.

> > 

> > True, but I hardly want to enumerate every possible failure point in a

> > log message.  How about:

> > 

> > Unable to apply ERR010728 workaround.  DT lacks imx7d-pcie-phy?

> 

> I would just log "Unable to apply ERR010728 workaround" but it is up

> to you. I can make the change myself with no need for reposting.


I was following Lucas's comment on the v1 version:

   the user. It would be better to have something like "No access to
   PHY registers, can't apply workaround for ERR010728." Maybe even
   with a hint to update the DT.

I think the pretty much the only way any of the errors conditions will
get hit, is when someone with an old DT runs a new kernel.  So the hint
about the DT change might be actually useful to some users.

But it's not a big deal so just change it to whatever when applying the
patch.
Lorenzo Pieralisi - Feb. 11, 2019, 12:18 p.m.
On Fri, Feb 08, 2019 at 06:19:10PM +0000, Trent Piepho wrote:
> On Fri, 2019-02-08 at 10:18 +0000, Lorenzo Pieralisi wrote:
> > On Thu, Feb 07, 2019 at 06:15:52PM +0000, Trent Piepho wrote:
> > > On Thu, 2019-02-07 at 12:31 +0000, Lorenzo Pieralisi wrote:
> > > > +			/* De-assert DCC_FB_EN */
> > > > > +			writel(PCIE_PHY_CMN_REG4_DCC_FB_EN,
> > > > > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG4);
> > > > > +			/* Assert RX_EQS and RX_EQS_SEL */
> > > > > +			writel(PCIE_PHY_CMN_REG24_RX_EQ_SEL
> > > > > +				| PCIE_PHY_CMN_REG24_RX_EQ,
> > > > > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG24);
> > > > > +			/* Assert ATT_MODE */
> > > > > +			writel(PCIE_PHY_CMN_REG26_ATT_MODE,
> > > > > +			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG26);
> > > > > +		} else {
> > > > > +			dev_warn(dev, "DT lacks imx7d-pcie-phy, unable to apply ERR010728 workaround\n");
> > > > 
> > > > It is a nit but the warning log is not necessarily true ie you may not
> > > > have a mapped address for other reasons as well. It is up to you but I
> > > > would change the log message.
> > > 
> > > True, but I hardly want to enumerate every possible failure point in a
> > > log message.  How about:
> > > 
> > > Unable to apply ERR010728 workaround.  DT lacks imx7d-pcie-phy?
> > 
> > I would just log "Unable to apply ERR010728 workaround" but it is up
> > to you. I can make the change myself with no need for reposting.
> 
> I was following Lucas's comment on the v1 version:
> 
>    the user. It would be better to have something like "No access to
>    PHY registers, can't apply workaround for ERR010728." Maybe even
>    with a hint to update the DT.
> 
> I think the pretty much the only way any of the errors conditions will
> get hit, is when someone with an old DT runs a new kernel.  So the hint
> about the DT change might be actually useful to some users.
> 
> But it's not a big deal so just change it to whatever when applying the
> patch.

Applied the series to pci/dwc for v5.1, please have a look to check
if we are in agreement (slightly tweaked the log string), thanks.

Lorenzo

Patch

diff --git a/drivers/pci/controller/dwc/pci-imx6.c b/drivers/pci/controller/dwc/pci-imx6.c
index 80f843030e36..06dd6aa927d4 100644
--- a/drivers/pci/controller/dwc/pci-imx6.c
+++ b/drivers/pci/controller/dwc/pci-imx6.c
@@ -18,6 +18,7 @@ 
 #include <linux/module.h>
 #include <linux/of_gpio.h>
 #include <linux/of_device.h>
+#include <linux/of_address.h>
 #include <linux/pci.h>
 #include <linux/platform_device.h>
 #include <linux/regmap.h>
@@ -61,6 +62,7 @@  struct imx6_pcie {
 	u32			tx_swing_low;
 	int			link_gen;
 	struct regulator	*vpcie;
+	void __iomem		*phy_base;
 
 	/* power domain for pcie */
 	struct device		*pd_pcie;
@@ -117,6 +119,23 @@  struct imx6_pcie {
 #define PCIE_PHY_RX_ASIC_OUT 0x100D
 #define PCIE_PHY_RX_ASIC_OUT_VALID	(1 << 0)
 
+/* iMX7 PCIe PHY registers */
+#define PCIE_PHY_CMN_REG4		0x14
+/* These are probably the bits that *aren't* DCC_FB_EN */
+#define PCIE_PHY_CMN_REG4_DCC_FB_EN	0x29
+
+#define PCIE_PHY_CMN_REG15	        0x54
+#define PCIE_PHY_CMN_REG15_DLY_4	BIT(2)
+#define PCIE_PHY_CMN_REG15_PLL_PD	BIT(5)
+#define PCIE_PHY_CMN_REG15_OVRD_PLL_PD	BIT(7)
+
+#define PCIE_PHY_CMN_REG24		0x90
+#define PCIE_PHY_CMN_REG24_RX_EQ	BIT(6)
+#define PCIE_PHY_CMN_REG24_RX_EQ_SEL	BIT(3)
+
+#define PCIE_PHY_CMN_REG26		0x98
+#define PCIE_PHY_CMN_REG26_ATT_MODE	0xBC
+
 #define PHY_RX_OVRD_IN_LO 0x1005
 #define PHY_RX_OVRD_IN_LO_RX_DATA_EN (1 << 5)
 #define PHY_RX_OVRD_IN_LO_RX_PLL_EN (1 << 3)
@@ -490,6 +509,26 @@  static void imx6_pcie_deassert_core_reset(struct imx6_pcie *imx6_pcie)
 	switch (imx6_pcie->variant) {
 	case IMX7D:
 		reset_control_deassert(imx6_pcie->pciephy_reset);
+
+		/* Workaround for ERR010728, failure of PCI-e PLL VCO to
+		 * oscillate, especially when cold.  This turns off "Duty-cycle
+		 * Corrector" and other mysterious undocumented things.
+		 */
+		if (likely(imx6_pcie->phy_base)) {
+			/* De-assert DCC_FB_EN */
+			writel(PCIE_PHY_CMN_REG4_DCC_FB_EN,
+			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG4);
+			/* Assert RX_EQS and RX_EQS_SEL */
+			writel(PCIE_PHY_CMN_REG24_RX_EQ_SEL
+				| PCIE_PHY_CMN_REG24_RX_EQ,
+			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG24);
+			/* Assert ATT_MODE */
+			writel(PCIE_PHY_CMN_REG26_ATT_MODE,
+			       imx6_pcie->phy_base + PCIE_PHY_CMN_REG26);
+		} else {
+			dev_warn(dev, "DT lacks imx7d-pcie-phy, unable to apply ERR010728 workaround\n");
+		}
+
 		imx7d_pcie_wait_for_phy_pll_lock(imx6_pcie);
 		break;
 	case IMX6SX:
@@ -919,6 +958,7 @@  static int imx6_pcie_probe(struct platform_device *pdev)
 	struct device *dev = &pdev->dev;
 	struct dw_pcie *pci;
 	struct imx6_pcie *imx6_pcie;
+	struct device_node *np;
 	struct resource *dbi_base;
 	struct device_node *node = dev->of_node;
 	int ret;
@@ -939,6 +979,23 @@  static int imx6_pcie_probe(struct platform_device *pdev)
 	imx6_pcie->variant =
 		(enum imx6_pcie_variants)of_device_get_match_data(dev);
 
+	/* Find the PHY if one is defined, only imx7d uses it */
+	np = of_parse_phandle(node, "fsl,imx7d-pcie-phy", 0);
+	if (np) {
+		struct resource res;
+
+		ret = of_address_to_resource(np, 0, &res);
+		if (ret) {
+			dev_err(dev, "Unable to map PCIe PHY\n");
+			return ret;
+		}
+		imx6_pcie->phy_base = devm_ioremap_resource(dev, &res);
+		if (IS_ERR(imx6_pcie->phy_base)) {
+			dev_err(dev, "Unable to map PCIe PHY\n");
+			return PTR_ERR(imx6_pcie->phy_base);
+		}
+	}
+
 	dbi_base = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	pci->dbi_base = devm_ioremap_resource(dev, dbi_base);
 	if (IS_ERR(pci->dbi_base))