[PATCH V2 0/9] Enhancements to pcie-tegra194 driver

[PATCH V2 0/9] Enhancements to pcie-tegra194 driver

[Vidya Sagar]

This patch series contains enhancements to the pcie-tegra194.c driver
that works for both Tegra194 and Tegra234 SoCs and for both RootPort
and Endpoint modes.

V2:
* Addressed review comments from test bot and Vinod

Vidya Sagar (9):
  PCI: tegra194: Use devm_gpiod_get_optional() to parse
    "nvidia,refclk-select"
  PCI: tegra194: Drive CLKREQ signal low explicitly
  PCI: tegra194: Fix polling delay for L2 state
  PCI: tegra194: Handle errors in BPMP response
  PCI: tegra194: Apply pinctrl settings for both PCIe RP and EP
  PCI: tegra194: Refactor LTSSM state polling on surprise down
  PCI: tegra194: Disable direct speed change for EP
  phy: tegra: p2u: Set ENABLE_L2_EXIT_RATE_CHANGE in calibration
  PCI: tegra194: Calibrate P2U for endpoint mode

 drivers/pci/controller/dwc/pcie-tegra194.c | 126 +++++++++++++++------
 drivers/phy/tegra/phy-tegra194-p2u.c       |  14 +++
 2 files changed, 104 insertions(+), 36 deletions(-)

-- 
2.17.1

[PATCH V2 1/9] PCI: tegra194: Use devm_gpiod_get_optional() to parse "nvidia,refclk-select"

[Vidya Sagar]

The GPIO DT property "nvidia,refclk-select" to select the PCIe reference
clock is optional. Use devm_gpiod_get_optional() to get it.

Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
---
V2:
* None

 drivers/pci/controller/dwc/pcie-tegra194.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c
index 2600304522eb..941fdb23e02f 100644
--- a/drivers/pci/controller/dwc/pcie-tegra194.c
+++ b/drivers/pci/controller/dwc/pcie-tegra194.c
@@ -1176,9 +1176,9 @@ static int tegra_pcie_dw_parse_dt(struct tegra_pcie_dw *pcie)
 		return err;
 	}
 
-	pcie->pex_refclk_sel_gpiod = devm_gpiod_get(pcie->dev,
-						    "nvidia,refclk-select",
-						    GPIOD_OUT_HIGH);
+	pcie->pex_refclk_sel_gpiod = devm_gpiod_get_optional(pcie->dev,
+							     "nvidia,refclk-select",
+							     GPIOD_OUT_HIGH);
 	if (IS_ERR(pcie->pex_refclk_sel_gpiod)) {
 		int err = PTR_ERR(pcie->pex_refclk_sel_gpiod);
 		const char *level = KERN_ERR;
-- 
2.17.1

[PATCH V2 2/9] PCI: tegra194: Drive CLKREQ signal low explicitly

[Vidya Sagar]

Currently, the default setting is that CLKREQ signal of a Root Port
is internally overridden to '0' to enable REFCLK flowing out to the slot.
It is observed that one of the PCIe switches (case in point Broadcom PCIe
Gen4 switch) is propagating the CLKREQ signal of the root port to the
downstream side of the switch and expecting the endpoints to pull it low
so that it (PCIe switch) can give out the REFCLK although the Switch as
such doesn't support CLK-PM or ASPM-L1SS. So, as a work-around, this patch
drives the CLKREQ of the Root Port itself low to avoid link up issues
between PCIe switch downstream port and endpoints. This is not a wrong
thing to do after all the CLKREQ is anyway being overridden to '0'
internally and now it is just that the same is being propagated outside
also.

Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
---
V2:
* None

 drivers/pci/controller/dwc/pcie-tegra194.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c
index 941fdb23e02f..7721f920dd74 100644
--- a/drivers/pci/controller/dwc/pcie-tegra194.c
+++ b/drivers/pci/controller/dwc/pcie-tegra194.c
@@ -46,6 +46,7 @@
 #define APPL_PINMUX_CLKREQ_OVERRIDE		BIT(3)
 #define APPL_PINMUX_CLK_OUTPUT_IN_OVERRIDE_EN	BIT(4)
 #define APPL_PINMUX_CLK_OUTPUT_IN_OVERRIDE	BIT(5)
+#define APPL_PINMUX_CLKREQ_DEFAULT_VALUE	BIT(13)
 
 #define APPL_CTRL				0x4
 #define APPL_CTRL_SYS_PRE_DET_STATE		BIT(6)
@@ -1453,6 +1454,7 @@ static int tegra_pcie_config_controller(struct tegra_pcie_dw *pcie,
 		val = appl_readl(pcie, APPL_PINMUX);
 		val |= APPL_PINMUX_CLKREQ_OVERRIDE_EN;
 		val &= ~APPL_PINMUX_CLKREQ_OVERRIDE;
+		val &= ~APPL_PINMUX_CLKREQ_DEFAULT_VALUE;
 		appl_writel(pcie, val, APPL_PINMUX);
 	}
 
-- 
2.17.1

[PATCH V2 3/9] PCI: tegra194: Fix polling delay for L2 state

[Vidya Sagar]

Current L2 state polling with 1us interval is too aggressive. Based on
the experiments with different endpoints, LTSSM state transisition to L2
is happening between 6us ~ 40us. Hence, update the polling delay for L2
state from 1us to 100us for a better utilization of CPU cycles.

Fixes: 56e15a238d92 ("PCI: tegra: Add Tegra194 PCIe support")
Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
---
V2:
* None

 drivers/pci/controller/dwc/pcie-tegra194.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c
index 7721f920dd74..7d6e54a12eff 100644
--- a/drivers/pci/controller/dwc/pcie-tegra194.c
+++ b/drivers/pci/controller/dwc/pcie-tegra194.c
@@ -206,7 +206,8 @@
 #define CAP_SPCIE_CAP_OFF_USP_TX_PRESET0_MASK	GENMASK(11, 8)
 #define CAP_SPCIE_CAP_OFF_USP_TX_PRESET0_SHIFT	8
 
-#define PME_ACK_TIMEOUT 10000
+#define PME_ACK_DELAY		100   /* 100 us */
+#define PME_ACK_TIMEOUT		10000 /* 10 ms */
 
 #define LTSSM_TIMEOUT 50000	/* 50ms */
 
@@ -1556,7 +1557,7 @@ static int tegra_pcie_try_link_l2(struct tegra_pcie_dw *pcie)
 
 	return readl_poll_timeout_atomic(pcie->appl_base + APPL_DEBUG, val,
 				 val & APPL_DEBUG_PM_LINKST_IN_L2_LAT,
-				 1, PME_ACK_TIMEOUT);
+				 PME_ACK_DELAY, PME_ACK_TIMEOUT);
 }
 
 static void tegra_pcie_dw_pme_turnoff(struct tegra_pcie_dw *pcie)
-- 
2.17.1

[PATCH V2 4/9] PCI: tegra194: Handle errors in BPMP response

[Vidya Sagar]

The return value from tegra_bpmp_transfer indicates the success or
failure of the IPC transaction with BPMP. If the transaction
succeeded, we also need to check the actual command's result code.
Add code to do this.

Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
---
V2:
* None

 drivers/pci/controller/dwc/pcie-tegra194.c | 18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)

diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c
index 7d6e54a12eff..0268eacdae48 100644
--- a/drivers/pci/controller/dwc/pcie-tegra194.c
+++ b/drivers/pci/controller/dwc/pcie-tegra194.c
@@ -1203,6 +1203,7 @@ static int tegra_pcie_bpmp_set_ctrl_state(struct tegra_pcie_dw *pcie,
 	struct mrq_uphy_response resp;
 	struct tegra_bpmp_message msg;
 	struct mrq_uphy_request req;
+	int err;
 
 	/*
 	 * Controller-5 doesn't need to have its state set by BPMP-FW in
@@ -1225,7 +1226,13 @@ static int tegra_pcie_bpmp_set_ctrl_state(struct tegra_pcie_dw *pcie,
 	msg.rx.data = &resp;
 	msg.rx.size = sizeof(resp);
 
-	return tegra_bpmp_transfer(pcie->bpmp, &msg);
+	err = tegra_bpmp_transfer(pcie->bpmp, &msg);
+	if (err)
+		return err;
+	if (msg.rx.ret)
+		return -EINVAL;
+
+	return 0;
 }
 
 static int tegra_pcie_bpmp_set_pll_state(struct tegra_pcie_dw *pcie,
@@ -1234,6 +1241,7 @@ static int tegra_pcie_bpmp_set_pll_state(struct tegra_pcie_dw *pcie,
 	struct mrq_uphy_response resp;
 	struct tegra_bpmp_message msg;
 	struct mrq_uphy_request req;
+	int err;
 
 	memset(&req, 0, sizeof(req));
 	memset(&resp, 0, sizeof(resp));
@@ -1253,7 +1261,13 @@ static int tegra_pcie_bpmp_set_pll_state(struct tegra_pcie_dw *pcie,
 	msg.rx.data = &resp;
 	msg.rx.size = sizeof(resp);
 
-	return tegra_bpmp_transfer(pcie->bpmp, &msg);
+	err = tegra_bpmp_transfer(pcie->bpmp, &msg);
+	if (err)
+		return err;
+	if (msg.rx.ret)
+		return -EINVAL;
+
+	return 0;
 }
 
 static void tegra_pcie_downstream_dev_to_D0(struct tegra_pcie_dw *pcie)
-- 
2.17.1

[PATCH V2 5/9] PCI: tegra194: Apply pinctrl settings for both PCIe RP and EP

[Vidya Sagar]

PERST# and CLKREQ# pinctrl settings should be applied for both root port
and endpoint mode. Move pinctrl_pm_select_default_state() function call
from root port specific configuration function to probe().

Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
---
V2:
* None

 drivers/pci/controller/dwc/pcie-tegra194.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)

diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c
index 0268eacdae48..4ba2a17d92d2 100644
--- a/drivers/pci/controller/dwc/pcie-tegra194.c
+++ b/drivers/pci/controller/dwc/pcie-tegra194.c
@@ -1659,12 +1659,6 @@ static int tegra_pcie_config_rp(struct tegra_pcie_dw *pcie)
 		goto fail_pm_get_sync;
 	}
 
-	ret = pinctrl_pm_select_default_state(dev);
-	if (ret < 0) {
-		dev_err(dev, "Failed to configure sideband pins: %d\n", ret);
-		goto fail_pm_get_sync;
-	}
-
 	ret = tegra_pcie_init_controller(pcie);
 	if (ret < 0) {
 		dev_err(dev, "Failed to initialize controller: %d\n", ret);
@@ -2120,6 +2114,19 @@ static int tegra_pcie_dw_probe(struct platform_device *pdev)
 	pp = &pci->pp;
 	pp->num_vectors = MAX_MSI_IRQS;
 
+	ret = pinctrl_pm_select_default_state(dev);
+	if (ret < 0) {
+		const char *level = KERN_ERR;
+
+		if (ret == -EPROBE_DEFER)
+			level = KERN_DEBUG;
+
+		dev_printk(level, dev,
+			   "Failed to configure sideband pins: %d\n",
+			   ret);
+		return ret;
+	}
+
 	ret = tegra_pcie_dw_parse_dt(pcie);
 	if (ret < 0) {
 		const char *level = KERN_ERR;
-- 
2.17.1

[PATCH V2 6/9] PCI: tegra194: Refactor LTSSM state polling on surprise down

[Vidya Sagar]

On surprise down LTSSM state transisition from L0 -> Recovery.RcvrLock ->
Recovery.RcvrSpeed -> Gen1 Recovery.RcvrLock -> Detect.
Recovery.RcvrLock and Recovery.RcvrSpeed time is 24 msec and 48 msec
respectively. It takes ~96 msec to move from L0 to detect state, hence,
increase the poll time to 120 msec. Disable the LTSSM state after it moves
to detect to avoid LTSSM toggle between polling and detect.

Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
---
V2:
* None

 drivers/pci/controller/dwc/pcie-tegra194.c | 69 ++++++++++++++--------
 1 file changed, 46 insertions(+), 23 deletions(-)

diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c
index 4ba2a17d92d2..702c55deb747 100644
--- a/drivers/pci/controller/dwc/pcie-tegra194.c
+++ b/drivers/pci/controller/dwc/pcie-tegra194.c
@@ -140,7 +140,11 @@
 #define APPL_DEBUG_PM_LINKST_IN_L0		0x11
 #define APPL_DEBUG_LTSSM_STATE_MASK		GENMASK(8, 3)
 #define APPL_DEBUG_LTSSM_STATE_SHIFT		3
-#define LTSSM_STATE_PRE_DETECT			5
+#define LTSSM_STATE_DETECT_QUIET		0x00
+#define LTSSM_STATE_DETECT_ACT			0x08
+#define LTSSM_STATE_PRE_DETECT_QUIET		0x28
+#define LTSSM_STATE_DETECT_WAIT			0x30
+#define LTSSM_STATE_L2_IDLE			0xa8
 
 #define APPL_RADM_STATUS			0xE4
 #define APPL_PM_XMT_TURNOFF_STATE		BIT(0)
@@ -209,7 +213,8 @@
 #define PME_ACK_DELAY		100   /* 100 us */
 #define PME_ACK_TIMEOUT		10000 /* 10 ms */
 
-#define LTSSM_TIMEOUT 50000	/* 50ms */
+#define LTSSM_DELAY		10000	/* 10 ms */
+#define LTSSM_TIMEOUT		120000	/* 120 ms */
 
 #define GEN3_GEN4_EQ_PRESET_INIT	5
 
@@ -1606,23 +1611,31 @@ static void tegra_pcie_dw_pme_turnoff(struct tegra_pcie_dw *pcie)
 		data &= ~APPL_PINMUX_PEX_RST;
 		appl_writel(pcie, data, APPL_PINMUX);
 
-		/*
-		 * Some cards do not go to detect state even after de-asserting
-		 * PERST#. So, de-assert LTSSM to bring link to detect state.
-		 */
-		data = readl(pcie->appl_base + APPL_CTRL);
-		data &= ~APPL_CTRL_LTSSM_EN;
-		writel(data, pcie->appl_base + APPL_CTRL);
-
 		err = readl_poll_timeout_atomic(pcie->appl_base + APPL_DEBUG,
 						data,
 						((data &
-						APPL_DEBUG_LTSSM_STATE_MASK) >>
-						APPL_DEBUG_LTSSM_STATE_SHIFT) ==
-						LTSSM_STATE_PRE_DETECT,
-						1, LTSSM_TIMEOUT);
+						APPL_DEBUG_LTSSM_STATE_MASK) ==
+						LTSSM_STATE_DETECT_QUIET) ||
+						((data &
+						APPL_DEBUG_LTSSM_STATE_MASK) ==
+						LTSSM_STATE_DETECT_ACT) ||
+						((data &
+						APPL_DEBUG_LTSSM_STATE_MASK) ==
+						LTSSM_STATE_PRE_DETECT_QUIET) ||
+						((data &
+						APPL_DEBUG_LTSSM_STATE_MASK) ==
+						LTSSM_STATE_DETECT_WAIT),
+						LTSSM_DELAY, LTSSM_TIMEOUT);
 		if (err)
 			dev_info(pcie->dev, "Link didn't go to detect state\n");
+
+		/*
+		 * Deassert LTSSM state to stop the state toggling between
+		 * polling and detect.
+		 */
+		data = readl(pcie->appl_base + APPL_CTRL);
+		data &= ~APPL_CTRL_LTSSM_EN;
+		writel(data, pcie->appl_base + APPL_CTRL);
 	}
 	/*
 	 * DBI registers may not be accessible after this as PLL-E would be
@@ -1698,19 +1711,29 @@ static void pex_ep_event_pex_rst_assert(struct tegra_pcie_dw *pcie)
 	if (pcie->ep_state == EP_STATE_DISABLED)
 		return;
 
-	/* Disable LTSSM */
+	ret = readl_poll_timeout(pcie->appl_base + APPL_DEBUG, val,
+				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
+				 LTSSM_STATE_DETECT_QUIET) ||
+				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
+				 LTSSM_STATE_DETECT_ACT) ||
+				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
+				 LTSSM_STATE_PRE_DETECT_QUIET) ||
+				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
+				 LTSSM_STATE_DETECT_WAIT) ||
+				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
+				 LTSSM_STATE_L2_IDLE),
+				 LTSSM_DELAY, LTSSM_TIMEOUT);
+	if (ret)
+		dev_err(pcie->dev, "LTSSM state: 0x%x timeout: %d\n", val, ret);
+
+	/*
+	 * Deassert LTSSM state to stop the state toggling between
+	 * polling and detect.
+	 */
 	val = appl_readl(pcie, APPL_CTRL);
 	val &= ~APPL_CTRL_LTSSM_EN;
 	appl_writel(pcie, val, APPL_CTRL);
 
-	ret = readl_poll_timeout(pcie->appl_base + APPL_DEBUG, val,
-				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) >>
-				 APPL_DEBUG_LTSSM_STATE_SHIFT) ==
-				 LTSSM_STATE_PRE_DETECT,
-				 1, LTSSM_TIMEOUT);
-	if (ret)
-		dev_err(pcie->dev, "Failed to go Detect state: %d\n", ret);
-
 	reset_control_assert(pcie->core_rst);
 
 	tegra_pcie_disable_phy(pcie);
-- 
2.17.1

Re: [PATCH V2 6/9] PCI: tegra194: Refactor LTSSM state polling on surprise down

[Bjorn Helgaas]

On Mon, Sep 26, 2022 at 05:20:35PM +0530, Vidya Sagar wrote:
> On surprise down LTSSM state transisition from L0 -> Recovery.RcvrLock ->
> Recovery.RcvrSpeed -> Gen1 Recovery.RcvrLock -> Detect.
> Recovery.RcvrLock and Recovery.RcvrSpeed time is 24 msec and 48 msec
> respectively. It takes ~96 msec to move from L0 to detect state, hence,
> increase the poll time to 120 msec. Disable the LTSSM state after it moves
> to detect to avoid LTSSM toggle between polling and detect.

s/transisition/transition/

>  		err = readl_poll_timeout_atomic(pcie->appl_base + APPL_DEBUG,
>  						data,
>  						((data &
> -						APPL_DEBUG_LTSSM_STATE_MASK) >>
> -						APPL_DEBUG_LTSSM_STATE_SHIFT) ==
> -						LTSSM_STATE_PRE_DETECT,
> -						1, LTSSM_TIMEOUT);
> +						APPL_DEBUG_LTSSM_STATE_MASK) ==
> +						LTSSM_STATE_DETECT_QUIET) ||
> +						((data &
> +						APPL_DEBUG_LTSSM_STATE_MASK) ==
> +						LTSSM_STATE_DETECT_ACT) ||
> +						((data &
> +						APPL_DEBUG_LTSSM_STATE_MASK) ==
> +						LTSSM_STATE_PRE_DETECT_QUIET) ||
> +						((data &
> +						APPL_DEBUG_LTSSM_STATE_MASK) ==
> +						LTSSM_STATE_DETECT_WAIT),
> +						LTSSM_DELAY, LTSSM_TIMEOUT);

I know we usually line up the function parameters with that opening
paren, but I think overall it would be nicer to drop a couple of the
leading tabs so these don't wrap so awkwardly.

> +	ret = readl_poll_timeout(pcie->appl_base + APPL_DEBUG, val,
> +				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
> +				 LTSSM_STATE_DETECT_QUIET) ||
> +				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
> +				 LTSSM_STATE_DETECT_ACT) ||
> +				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
> +				 LTSSM_STATE_PRE_DETECT_QUIET) ||
> +				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
> +				 LTSSM_STATE_DETECT_WAIT) ||
> +				 ((val & APPL_DEBUG_LTSSM_STATE_MASK) ==
> +				 LTSSM_STATE_L2_IDLE),
> +				 LTSSM_DELAY, LTSSM_TIMEOUT);

Ditto.

[PATCH V2 7/9] PCI: tegra194: Disable direct speed change for EP

[Vidya Sagar]

Disable direct speed change for endpoint to prevent it from doing
speed change by itself post physical layer link up at gen1 and also
to leave link speed in control of the host.

Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
---
V2:
* None

 drivers/pci/controller/dwc/pcie-tegra194.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c
index 702c55deb747..3e4868ab8416 100644
--- a/drivers/pci/controller/dwc/pcie-tegra194.c
+++ b/drivers/pci/controller/dwc/pcie-tegra194.c
@@ -1872,6 +1872,10 @@ static void pex_ep_event_pex_rst_deassert(struct tegra_pcie_dw *pcie)
 
 	reset_control_deassert(pcie->core_rst);
 
+	val = dw_pcie_readl_dbi(pci, PCIE_LINK_WIDTH_SPEED_CONTROL);
+	val &= ~PORT_LOGIC_SPEED_CHANGE;
+	dw_pcie_writel_dbi(pci, PCIE_LINK_WIDTH_SPEED_CONTROL, val);
+
 	if (pcie->update_fc_fixup) {
 		val = dw_pcie_readl_dbi(pci, CFG_TIMER_CTRL_MAX_FUNC_NUM_OFF);
 		val |= 0x1 << CFG_TIMER_CTRL_ACK_NAK_SHIFT;
-- 
2.17.1

[PATCH V2 8/9] phy: tegra: p2u: Set ENABLE_L2_EXIT_RATE_CHANGE in calibration

[Vidya Sagar]

Set ENABLE_L2_EXIT_RATE_CHANGE to request UPHY PLL rate change to Gen1
during initialization. This helps in the below surprise down cases,
  - Surprise down happens at Gen3/Gen4 link speed
  - Surprise down happens and external REFCLK is cut off which causes
UPHY PLL rate to deviate to an invalid rate

ENABLE_L2_EXIT_RATE_CHANGE needs to be set to bring the UPHY PLL rate
back to Gen1 during controller initialization for the link up.

Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
Reported-by: kernel test robot <lkp@intel.com>
---
V2:
* Addressed review comment from test bot and Vinod

 drivers/phy/tegra/phy-tegra194-p2u.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/drivers/phy/tegra/phy-tegra194-p2u.c b/drivers/phy/tegra/phy-tegra194-p2u.c
index 1415ca71de38..633e6b747275 100644
--- a/drivers/phy/tegra/phy-tegra194-p2u.c
+++ b/drivers/phy/tegra/phy-tegra194-p2u.c
@@ -15,6 +15,7 @@
 #include <linux/phy/phy.h>
 
 #define P2U_CONTROL_CMN			0x74
+#define P2U_CONTROL_CMN_ENABLE_L2_EXIT_RATE_CHANGE		BIT(13)
 #define P2U_CONTROL_CMN_SKP_SIZE_PROTECTION_EN			BIT(20)
 
 #define P2U_PERIODIC_EQ_CTRL_GEN3	0xc0
@@ -85,8 +86,21 @@ static int tegra_p2u_power_on(struct phy *x)
 	return 0;
 }
 
+static int tegra_p2u_calibrate(struct phy *x)
+{
+	struct tegra_p2u *phy = phy_get_drvdata(x);
+	u32 val;
+
+	val = p2u_readl(phy, P2U_CONTROL_CMN);
+	val |= P2U_CONTROL_CMN_ENABLE_L2_EXIT_RATE_CHANGE;
+	p2u_writel(phy, val, P2U_CONTROL_CMN);
+
+	return 0;
+}
+
 static const struct phy_ops ops = {
 	.power_on = tegra_p2u_power_on,
+	.calibrate = tegra_p2u_calibrate,
 	.owner = THIS_MODULE,
 };
 
-- 
2.17.1

Re: [PATCH V2 8/9] phy: tegra: p2u: Set ENABLE_L2_EXIT_RATE_CHANGE in calibration

[Bjorn Helgaas]

On Mon, Sep 26, 2022 at 05:20:37PM +0530, Vidya Sagar wrote:
> Set ENABLE_L2_EXIT_RATE_CHANGE to request UPHY PLL rate change to Gen1
> during initialization. This helps in the below surprise down cases,
>   - Surprise down happens at Gen3/Gen4 link speed
>   - Surprise down happens and external REFCLK is cut off which causes
> UPHY PLL rate to deviate to an invalid rate
> 
> ENABLE_L2_EXIT_RATE_CHANGE needs to be set to bring the UPHY PLL rate
> back to Gen1 during controller initialization for the link up.
> 
> Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
> Reported-by: kernel test robot <lkp@intel.com>

I doubt the kernel test robot reported the issue being fixed by this
patch.  More likely it reported a syntax or similar issue in v1, and I
think the "reported-by" tag is meaningless in cases like that.

[PATCH V2 9/9] PCI: tegra194: Calibrate P2U for endpoint mode

[Vidya Sagar]

Calibrate P2U for endpoint to request UPHY PLL rate change to Gen1 during
initialization.

Signed-off-by: Vidya Sagar <vidyas@nvidia.com>
---
V2:
* None

 drivers/pci/controller/dwc/pcie-tegra194.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c
index 3e4868ab8416..cfe32a98b2f3 100644
--- a/drivers/pci/controller/dwc/pcie-tegra194.c
+++ b/drivers/pci/controller/dwc/pcie-tegra194.c
@@ -1078,6 +1078,9 @@ static int tegra_pcie_enable_phy(struct tegra_pcie_dw *pcie)
 		ret = phy_power_on(pcie->phys[i]);
 		if (ret < 0)
 			goto phy_exit;
+
+		if (pcie->of_data->mode == DW_PCIE_EP_TYPE)
+			phy_calibrate(pcie->phys[i]);
 	}
 
 	return 0;
-- 
2.17.1