diff mbox series

[RFC] usb: dwc3: dwc3-octeon: Fix USB PHY High-Speed PLL Initialization

Message ID ZShGmL4mph91Ncib@lenoch (mailing list archive)
State New, archived
Headers show
Series [RFC] usb: dwc3: dwc3-octeon: Fix USB PHY High-Speed PLL Initialization | expand

Commit Message

Ladislav Michl Oct. 12, 2023, 7:18 p.m. UTC
From: Ladislav Michl <ladis@linux-mips.org>

Implement workaround for Octeon Known Issue Id 29206:
| The USB high speed logic contains a PLL that must lock during
| initialization for correct operation. In very rare circumstances,
| it is possible for the PLL to fail to start correctly.
| Workaround
| After initialization, check the USB PLL lock register via the
| UPHY CR interface. If the PLL is not running, power it down and
| back up and restart the initialization.

PLL initialization code taken from Cavium's vendor bootloader:
u-boot/drivers/usb/host/xhci-octeon.c:octeon3_usb_clocks_start

Signed-off-by: Ladislav Michl <ladis@linux-mips.org>
---
 NOTE:
 This patch fixes initialization issue found on some CN7020 based boards.
 Without this patch, controller sometimes fails to detect devices connected.
 Original code comes from Cavium released u-boot monster patch, which seems
 to suffer from mistakes made while resolving merge conflicts when upgrading
 to newer u-boot.
 Testing revealed that only single reinit is needed to properly lock PLL,
 this agrees with comment in Cavium's u-boot code, which is claiming the
 same. However, same as in u-boot code, reinit is attempted three times.
 (in could be done using while loop instead of goto, just let me know
 which way do you prefer)
 SoCs suffering from this problem would fail to initialize PHY about
 several tens times of thousand boots. This patch always restored
 functional state.

 drivers/usb/dwc3/dwc3-octeon.c | 285 ++++++++++++++++++++++++++++++++-
 1 file changed, 284 insertions(+), 1 deletion(-)

Comments

Thinh Nguyen Oct. 12, 2023, 10:23 p.m. UTC | #1
On Thu, Oct 12, 2023, Ladislav Michl wrote:
> From: Ladislav Michl <ladis@linux-mips.org>
> 
> Implement workaround for Octeon Known Issue Id 29206:
> | The USB high speed logic contains a PLL that must lock during
> | initialization for correct operation. In very rare circumstances,
> | it is possible for the PLL to fail to start correctly.
> | Workaround
> | After initialization, check the USB PLL lock register via the
> | UPHY CR interface. If the PLL is not running, power it down and
> | back up and restart the initialization.

Minor nit: Can we replace "|" with tabs. I think it's easier to read.

> 
> PLL initialization code taken from Cavium's vendor bootloader:
> u-boot/drivers/usb/host/xhci-octeon.c:octeon3_usb_clocks_start
> 
> Signed-off-by: Ladislav Michl <ladis@linux-mips.org>
> ---
>  NOTE:
>  This patch fixes initialization issue found on some CN7020 based boards.
>  Without this patch, controller sometimes fails to detect devices connected.
>  Original code comes from Cavium released u-boot monster patch, which seems
>  to suffer from mistakes made while resolving merge conflicts when upgrading
>  to newer u-boot.
>  Testing revealed that only single reinit is needed to properly lock PLL,
>  this agrees with comment in Cavium's u-boot code, which is claiming the
>  same. However, same as in u-boot code, reinit is attempted three times.
>  (in could be done using while loop instead of goto, just let me know
>  which way do you prefer)
>  SoCs suffering from this problem would fail to initialize PHY about
>  several tens times of thousand boots. This patch always restored
>  functional state.

What kernel version did you use? Perhaps it has the same issue due to
the commit e835c0a4e23c ("usb: dwc3: don't reset device side if dwc3 was
configured as host-only")

Did you test this against Greg's usb-linus branch with the fix for the
above?

Thanks,
Thinh
Ladislav Michl Oct. 13, 2023, 5:16 a.m. UTC | #2
On Thu, Oct 12, 2023 at 10:23:30PM +0000, Thinh Nguyen wrote:
> On Thu, Oct 12, 2023, Ladislav Michl wrote:
> > From: Ladislav Michl <ladis@linux-mips.org>
> > 
> > Implement workaround for Octeon Known Issue Id 29206:
> > | The USB high speed logic contains a PLL that must lock during
> > | initialization for correct operation. In very rare circumstances,
> > | it is possible for the PLL to fail to start correctly.
> > | Workaround
> > | After initialization, check the USB PLL lock register via the
> > | UPHY CR interface. If the PLL is not running, power it down and
> > | back up and restart the initialization.
> 
> Minor nit: Can we replace "|" with tabs. I think it's easier to read.

Ok.
 
> > PLL initialization code taken from Cavium's vendor bootloader:
> > u-boot/drivers/usb/host/xhci-octeon.c:octeon3_usb_clocks_start
> > 
> > Signed-off-by: Ladislav Michl <ladis@linux-mips.org>
> > ---
> >  NOTE:
> >  This patch fixes initialization issue found on some CN7020 based boards.
> >  Without this patch, controller sometimes fails to detect devices connected.
> >  Original code comes from Cavium released u-boot monster patch, which seems
> >  to suffer from mistakes made while resolving merge conflicts when upgrading
> >  to newer u-boot.
> >  Testing revealed that only single reinit is needed to properly lock PLL,
> >  this agrees with comment in Cavium's u-boot code, which is claiming the
> >  same. However, same as in u-boot code, reinit is attempted three times.
> >  (in could be done using while loop instead of goto, just let me know
> >  which way do you prefer)
> >  SoCs suffering from this problem would fail to initialize PHY about
> >  several tens times of thousand boots. This patch always restored
> >  functional state.
> 
> What kernel version did you use? Perhaps it has the same issue due to
> the commit e835c0a4e23c ("usb: dwc3: don't reset device side if dwc3 was
> configured as host-only")

Production devices are running 6.1.y, patch was developed and tested against
6.1.38, then ported to 6.6-rc5. The purpose of this RFC was to figure out
how to handle situation when informations are comming from NDA document
and code is based on the one coming from unpublished vendor tree.

> Did you test this against Greg's usb-linus branch with the fix for the
> above?

Do you mean "usb: dwc3: Soft reset phy on probe for host" patch? I had
noticed it, however this patch was not tested with that fix.

Cavium's u-boot is also resetting PHYs similar way it is done there,
but I ommited this change as testing revealed it is not needed.

Please see dwc3_uphy_pll_reset function and comment above it.
Without hardware documentation I can only guess whenever it makes sense.

Regards,
	ladis
Thinh Nguyen Oct. 20, 2023, 11:11 p.m. UTC | #3
Sorry for the delay response.

On Fri, Oct 13, 2023, Ladislav Michl wrote:
> On Thu, Oct 12, 2023 at 10:23:30PM +0000, Thinh Nguyen wrote:
> > On Thu, Oct 12, 2023, Ladislav Michl wrote:
> > > From: Ladislav Michl <ladis@linux-mips.org>
> > > 
> > > Implement workaround for Octeon Known Issue Id 29206:
> > > | The USB high speed logic contains a PLL that must lock during
> > > | initialization for correct operation. In very rare circumstances,
> > > | it is possible for the PLL to fail to start correctly.
> > > | Workaround
> > > | After initialization, check the USB PLL lock register via the
> > > | UPHY CR interface. If the PLL is not running, power it down and
> > > | back up and restart the initialization.
> > 
> > Minor nit: Can we replace "|" with tabs. I think it's easier to read.
> 
> Ok.
>  
> > > PLL initialization code taken from Cavium's vendor bootloader:
> > > u-boot/drivers/usb/host/xhci-octeon.c:octeon3_usb_clocks_start
> > > 
> > > Signed-off-by: Ladislav Michl <ladis@linux-mips.org>
> > > ---
> > >  NOTE:
> > >  This patch fixes initialization issue found on some CN7020 based boards.
> > >  Without this patch, controller sometimes fails to detect devices connected.
> > >  Original code comes from Cavium released u-boot monster patch, which seems
> > >  to suffer from mistakes made while resolving merge conflicts when upgrading
> > >  to newer u-boot.
> > >  Testing revealed that only single reinit is needed to properly lock PLL,
> > >  this agrees with comment in Cavium's u-boot code, which is claiming the
> > >  same. However, same as in u-boot code, reinit is attempted three times.
> > >  (in could be done using while loop instead of goto, just let me know
> > >  which way do you prefer)
> > >  SoCs suffering from this problem would fail to initialize PHY about
> > >  several tens times of thousand boots. This patch always restored
> > >  functional state.
> > 
> > What kernel version did you use? Perhaps it has the same issue due to
> > the commit e835c0a4e23c ("usb: dwc3: don't reset device side if dwc3 was
> > configured as host-only")
> 
> Production devices are running 6.1.y, patch was developed and tested against
> 6.1.38, then ported to 6.6-rc5. The purpose of this RFC was to figure out
> how to handle situation when informations are comming from NDA document
> and code is based on the one coming from unpublished vendor tree.
> 
> > Did you test this against Greg's usb-linus branch with the fix for the
> > above?
> 
> Do you mean "usb: dwc3: Soft reset phy on probe for host" patch? I had
> noticed it, however this patch was not tested with that fix.

Ok.

> 
> Cavium's u-boot is also resetting PHYs similar way it is done there,
> but I ommited this change as testing revealed it is not needed.
> 
> Please see dwc3_uphy_pll_reset function and comment above it.
> Without hardware documentation I can only guess whenever it makes sense.
> 

I have much less knowledge of this platform than you. As long as you
document your findings and how you came to the logic for the patch as
you did. It should be fine.

I notice that there are debug comments, and this looks like a
work-in-progress. When it's ready, please clean it up and remove the RFC
so we can merge it.

Thanks,
Thinh
diff mbox series

Patch

diff --git a/drivers/usb/dwc3/dwc3-octeon.c b/drivers/usb/dwc3/dwc3-octeon.c
index d45d62c72b2d..497c18e19c73 100644
--- a/drivers/usb/dwc3/dwc3-octeon.c
+++ b/drivers/usb/dwc3/dwc3-octeon.c
@@ -122,8 +122,32 @@ 
 #define USBDRD_UCTL_INTSTAT			0x30
 #define USBDRD_UCTL_PORT_CFG_HS(port)		(0x40 + (0x20 * port))
 #define USBDRD_UCTL_PORT_CFG_SS(port)		(0x48 + (0x20 * port))
+
+/*
+ * UCTL Port Debug Configuration Registers
+ */
 #define USBDRD_UCTL_PORT_CR_DBG_CFG(port)	(0x50 + (0x20 * port))
+/* Rising edge triggers a register write operation of the captured
+ * address with the captured data
+ */
+# define USBDRD_UCTL_PORT_CR_DBG_CFG_WRITE	BIT_ULL(0)
+/* Rising edge triggers a register read operation of the capture address */
+# define USBDRD_UCTL_PORT_CR_DBG_CFG_READ	BIT_ULL(1)
+/* Rising edge triggers the [DATA_IN] field to be captured as the write data */
+# define USBDRD_UCTL_PORT_CR_DBG_CFG_CAP_DATA	BIT_ULL(2)
+/* Rising edge triggers the [DATA_IN] field to be captured as the address */
+# define USBDRD_UCTL_PORT_CR_DBG_CFG_CAP_ADDR	BIT_ULL(3)
+/* Address or data to be written to the CR interface */
+# define USBDRD_UCTL_PORT_CR_DBG_CFG_DATA_IN	GENMASK_ULL(47, 32)
+
+/*
+ * UCTL Port Debug Status Registers
+ */
 #define USBDRD_UCTL_PORT_CR_DBG_STATUS(port)	(0x58 + (0x20 * port))
+/* Acknowledge that the CAP_ADDR, CAP_DATA, READ, WRITE commands have completed */
+# define USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK	BIT_ULL(0)
+/* Last data read from the CR interface */
+# define USBDRD_UCTL_PORT_CR_DBG_STATUS_DATA_OUT GENMASK_ULL(47, 32)
 
 /*
  * UCTL Configuration Register
@@ -410,6 +434,252 @@  static void dwc3_octeon_phy_reset(struct dwc3_octeon *octeon)
 	dwc3_octeon_writeq(uctl_ctl_reg, val);
 }
 
+/* Internal indirect register that reports if the phy PLL has lock.
+ * This will be 1 if lock, 0 if no lock.
+ */
+#define DWC3_INT_IND_PLL_LOCK_REG			0x200b
+
+/* Internal indirect UPHY register that controls the power to the UPHY PLL. */
+#define DWC3_INT_IND_UPHY_PLL_PU			0x2012
+/* Write enable bit for DWC3_INT_IND_PLL_POWER_CTL */
+# define DWC3_INT_IND_UPHY_PLL_PU_WE			BIT(5)
+/* Power enable bit for DWC3_INT_IND_PLL_POWER_CTL */
+# define DWC3_INT_IND_UPHY_PLL_PU_POWER_EN		BIT(2)
+
+/* Internal indirect UPHY PLL register */
+#define DWC3_INT_IND_UPHY_PLL_RESET			0x201C
+/* Write enable bit */
+# define DWC3_INT_IND_UPHY_PLL_RESET_WE			BIT(4)
+/* VCO reset bit */
+# define DWC3_INT_IND_UPHY_PLL_RESET_VCO_RST		BIT(0)
+
+static int dwc3_octeon_indirect_read(struct dwc3_octeon *octeon, u32 addr)
+{
+	int ret;
+	u64 val;
+	unsigned long timeout;
+	struct device *dev = octeon->dev;
+	void __iomem *cfg_reg = octeon->base + USBDRD_UCTL_PORT_CR_DBG_CFG(0);
+	void __iomem *status_reg = octeon->base + USBDRD_UCTL_PORT_CR_DBG_STATUS(0);
+
+	/* 1 */
+	val = FIELD_PREP(USBDRD_UCTL_PORT_CR_DBG_CFG_DATA_IN, addr);
+	dwc3_octeon_writeq(cfg_reg, val);
+	/* 2 */
+	val |= USBDRD_UCTL_PORT_CR_DBG_CFG_CAP_ADDR;
+	dwc3_octeon_writeq(cfg_reg, val);
+	/* 3 */
+	timeout = jiffies + msecs_to_jiffies(1000);
+	while (!(dwc3_octeon_readq(status_reg) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK)) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "set read address timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+	/* 4 */
+	dwc3_octeon_writeq(cfg_reg, 0);
+	/* 5 */
+	timeout = jiffies + msecs_to_jiffies(1000);
+	while (dwc3_octeon_readq(status_reg) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "read ack address clear timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+	/* 6 */
+	dwc3_octeon_writeq(cfg_reg, USBDRD_UCTL_PORT_CR_DBG_CFG_READ);
+	/* 7 */
+	timeout = jiffies + msecs_to_jiffies(1000);
+	while (!((val = dwc3_octeon_readq(status_reg)) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK)) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "read data timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+	/* 8 */
+	ret = FIELD_GET(USBDRD_UCTL_PORT_CR_DBG_STATUS_DATA_OUT, val);
+	/* 9 */
+	dwc3_octeon_writeq(cfg_reg, 0);
+	/* 10 */
+	while (dwc3_octeon_readq(status_reg) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "read ack data clear timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+
+	return ret;
+}
+
+static int dwc3_octeon_indirect_write(struct dwc3_octeon *octeon, u32 addr, u16 value)
+{
+	u64 val;
+	unsigned long timeout;
+	struct device *dev = octeon->dev;
+	void __iomem *cfg_reg = octeon->base + USBDRD_UCTL_PORT_CR_DBG_CFG(0);
+	void __iomem *status_reg = octeon->base + USBDRD_UCTL_PORT_CR_DBG_STATUS(0);
+
+	/* 1 */
+	val = FIELD_PREP(USBDRD_UCTL_PORT_CR_DBG_CFG_DATA_IN, addr);
+	dwc3_octeon_writeq(cfg_reg, val);
+	/* 2 */
+	val |= USBDRD_UCTL_PORT_CR_DBG_CFG_CAP_ADDR;
+	dwc3_octeon_writeq(cfg_reg, val);
+	/* 3 */
+	timeout = jiffies + msecs_to_jiffies(1000);
+	while (!(dwc3_octeon_readq(status_reg) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK)) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "set write address timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+	/* 4 */
+	dwc3_octeon_writeq(cfg_reg, 0);
+	/* 5 */
+	timeout = jiffies + msecs_to_jiffies(1000);
+	while (dwc3_octeon_readq(status_reg) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "write ack address clear timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+	/* 6 */
+	val = FIELD_PREP(USBDRD_UCTL_PORT_CR_DBG_CFG_DATA_IN, value);
+	dwc3_octeon_writeq(cfg_reg, FIELD_PREP(USBDRD_UCTL_PORT_CR_DBG_CFG_DATA_IN, value));
+	/* 7 */
+	val |= USBDRD_UCTL_PORT_CR_DBG_CFG_CAP_DATA;
+	dwc3_octeon_writeq(cfg_reg, val);
+	/* 8 */
+	timeout = jiffies + msecs_to_jiffies(1000);
+	while (!(dwc3_octeon_readq(status_reg) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK)) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "write set data timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+	/* 9 */
+	dwc3_octeon_writeq(cfg_reg, 0);
+	/* 10 */
+	timeout = jiffies + msecs_to_jiffies(1000);
+	while (dwc3_octeon_readq(status_reg) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "write ack clear timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+	/* 11 */
+	dwc3_octeon_writeq(cfg_reg, USBDRD_UCTL_PORT_CR_DBG_CFG_WRITE);
+	/* 12 */
+	timeout = jiffies + msecs_to_jiffies(1000);
+	while (!(dwc3_octeon_readq(status_reg) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK)) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "write data timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+	/* 13 */
+	dwc3_octeon_writeq(cfg_reg, 0);
+	/* 14 */
+	timeout = jiffies + msecs_to_jiffies(1000);
+	while (dwc3_octeon_readq(status_reg) & USBDRD_UCTL_PORT_CR_DBG_STATUS_ACK) {
+                if (time_after(jiffies, timeout)) {
+			dev_warn(dev, "write ack clear timeout (%x)\n", addr);
+                        return -ETIMEDOUT;
+		}
+                cpu_relax();
+        }
+
+	return 0;
+}
+
+static int dwc3_octeon_pll_locked(struct dwc3_octeon *octeon)
+{
+	int ret = dwc3_octeon_indirect_read(octeon, DWC3_INT_IND_PLL_LOCK_REG);
+
+	if (ret < 0)
+		return ret;
+	return ret & 1;
+}
+
+/**
+ * Performs a full reset of the UPHY PLL. Note that this is normally done
+ * internally by a state machine when the UPHY is brought out of reset but this
+ * version gives far more time for things to settle before continuing.
+ */
+static int dwc3_uphy_pll_reset(struct dwc3_octeon *octeon)
+{
+	u16 ctrl, pwr;
+
+	/* 1. Turn on write enable so we can assert reset to the PLL VCO */
+	ctrl = dwc3_octeon_indirect_read(octeon, DWC3_INT_IND_UPHY_PLL_RESET);
+	ctrl |= DWC3_INT_IND_UPHY_PLL_RESET_WE;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_RESET, ctrl);
+
+	/* 2. Turn on write enable for PLL power control */
+	pwr = dwc3_octeon_indirect_read(octeon, DWC3_INT_IND_UPHY_PLL_PU);
+	pwr |= DWC3_INT_IND_UPHY_PLL_PU_WE;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_PU, pwr);
+
+	/* 3. Assert VCO reset */
+	ctrl |= DWC3_INT_IND_UPHY_PLL_RESET_VCO_RST;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_RESET, ctrl);
+
+	/* 4. Power off the PLL */
+	pwr &= ~DWC3_INT_IND_UPHY_PLL_PU_POWER_EN;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_PU, pwr);
+	usleep_range(1000, 2000);
+
+	/* 5. Power on the PLL while VCO is held in reset */
+	pwr |= DWC3_INT_IND_UPHY_PLL_PU_POWER_EN;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_PU, pwr);
+
+	/* Wait for things to stabilize before taking VCO out of reset */
+	usleep_range(1000, 2000);
+
+	/* 6. Take the VCO out of reset */
+	ctrl &= ~DWC3_INT_IND_UPHY_PLL_RESET_VCO_RST;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_RESET, ctrl);
+	usleep_range(1000, 2000);
+
+	/* 7. Put the VCO back in reset */
+	ctrl |= ~DWC3_INT_IND_UPHY_PLL_RESET_VCO_RST;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_RESET, ctrl);
+
+	/* 8. Power off the PLL */
+	pwr &= ~DWC3_INT_IND_UPHY_PLL_PU_POWER_EN;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_PU, pwr);
+	usleep_range(1000, 2000);
+
+	/* 9. Power on the PLL while VCO is held in reset */
+	pwr |= DWC3_INT_IND_UPHY_PLL_PU_POWER_EN;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_PU, pwr);
+
+	/* 10. Take the VCO out of reset */
+	ctrl &= ~DWC3_INT_IND_UPHY_PLL_RESET_VCO_RST;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_RESET, ctrl);
+
+	/* 11. Turn off write enables */
+	pwr &= ~DWC3_INT_IND_UPHY_PLL_PU_WE;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_PU, pwr);
+
+	ctrl &= ~DWC3_INT_IND_UPHY_PLL_RESET_WE;
+	dwc3_octeon_indirect_write(octeon, DWC3_INT_IND_UPHY_PLL_RESET, ctrl);
+
+	usleep_range(1000, 2000);
+
+	/* Return if we have lock or not */
+	return dwc3_octeon_pll_locked(octeon);
+}
+
 static int dwc3_octeon_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
@@ -418,7 +688,7 @@  static int dwc3_octeon_probe(struct platform_device *pdev)
 	const char *hs_clock_type, *ss_clock_type;
 	int ref_clk_sel, ref_clk_fsel, mpll_mul;
 	int power_active_low, power_gpio;
-	int err, len;
+	int err, len, tries = 0;
 	u32 clock_rate;
 
 	if (of_property_read_u32(node, "refclk-frequency", &clock_rate)) {
@@ -503,6 +773,7 @@  static int dwc3_octeon_probe(struct platform_device *pdev)
 	if (IS_ERR(octeon->base))
 		return PTR_ERR(octeon->base);
 
+retry:
 	err = dwc3_octeon_setup(octeon, ref_clk_sel, ref_clk_fsel, mpll_mul,
 				power_gpio, power_active_low);
 	if (err)
@@ -511,6 +782,18 @@  static int dwc3_octeon_probe(struct platform_device *pdev)
 	dwc3_octeon_set_endian_mode(octeon);
 	dwc3_octeon_phy_reset(octeon);
 
+	usleep_range(50, 100);
+	if (dwc3_octeon_pll_locked(octeon) == 0) {
+		dev_warn(dev, "PLL unlocked, reseting (%d of 3)\n", ++tries);
+		err = dwc3_uphy_pll_reset(octeon);
+		if (err < 0)
+			return err;
+		if (tries < 3)
+			goto retry;
+		dev_err(dev, "PLL lock failed\n");
+		return -EIO;
+	}
+
 	platform_set_drvdata(pdev, octeon);
 
 	return of_platform_populate(node, NULL, NULL, dev);