Skip to contentSkip to conversation

fix apalis_t30 optional pcie operation

PATCH Superseded Marcel Ziswiler 3 participants 7 messages Aug 9, 2017

Subsystems (2)

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

v2 cover fix apalis_t30 optional pcie operation

This series addresses a PCIe reliability issue as observed on Apalis T30 related to a PCIe reset timing violation.

This series depends on Simon's work available at u-boot-dm/master plus my previous series "move apalis t30/tk1, colibri t20/t30 to livetree" and "fix apalis-tk1 pcie gigabit ethernet operation".

This series is available at http://git.toradex.com/cgit/u-boot-toradex.git/log/?h=for-next

Changes in v2:
- Leave resp. enable all port 0 pins input drivers as a customer may optionally want to use some of those MXM3 pins as inputs as well.
- Stick to struct tegra_pcie_port as suggested by Stephen.
- Introduce proper CONFIG_APALIS_T30_PCIE_EVALBOARD_INIT Kconfig option as suggested by Stephen.
- Improved the ifdef vs. if curly braces sequencing as suggested by Stephen.
- Keep PCIe port reset status in order to safeguard for future changes to the port reset order or even allow for re-initialisation should that ever be implemented in the higher levels of the driver model.

Marcel Ziswiler (3):
apalis_t30: describe pcie ports
apalis_t30: fix pcie port 0 and 1 pin muxing
apalis_t30: fix optional pcie port reset for reliable pcie operation

 arch/arm/dts/tegra30-apalis.dts                    |  3 ++
 board/toradex/apalis_t30/Kconfig                   |  9 ++++
 board/toradex/apalis_t30/apalis_t30.c              | 54 ++++++++++++++++++++++
 .../toradex/apalis_t30/pinmux-config-apalis_t30.h  | 16 ++++---
 4 files changed, 75 insertions(+), 7 deletions(-)
Signature · 1 line
2.9.4
#1 lore raw

PATCH v2 1/3 apalis_t30: describe pcie ports

From: Marcel Ziswiler <[email protected]>

Add some more comments describing the various PCIe ports available.

---

Changes in v2: None

 arch/arm/dts/tegra30-apalis.dts | 3 +++
 1 file changed, 3 insertions(+)
arch/arm/dts/tegra30-apalis.dts +3 -0
index 0b84dae..0852d8d 100644
@@ -43,16 +43,19 @@
 		vddio-pex-ctl-supply = <&sys_3v3_reg>;
 		hvdd-pex-supply = <&sys_3v3_reg>;
 
+		/* Apalis Type Specific 4 Lane PCIe */
 		pci at 1,0 {
 			/* TS_DIFF1/2/3/4 left disabled */
 			nvidia,num-lanes = <4>;
 		};
 
+		/* Apalis PCIe */
 		pci at 2,0 {
 			/* PCIE1_RX/TX left disabled */
 			nvidia,num-lanes = <1>;
 		};
 
+		/* I210 Gigabit Ethernet Controller (On-module) */
 		pci at 3,0 {
 			status = "okay";
 			nvidia,num-lanes = <1>;
Signature · 1 line
2.9.4
#2 lore raw

PATCH v2 2/3 apalis_t30: fix pcie port 0 and 1 pin muxing

In reply to Marcel Ziswiler

From: Marcel Ziswiler <[email protected]>

Fix optional Apalis type specific 4 lane PCIe port 0 and Apalis PCIe port 1 pin muxing.

---

Changes in v2:
- Leave resp. enable all port 0 pins input drivers as a customer may optionally want to use some of those MXM3 pins as inputs as well.

 board/toradex/apalis_t30/pinmux-config-apalis_t30.h | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)
board/toradex/apalis_t30/pinmux-config-apalis_t30.h +9 -7
index e0b00ea..a1e3bb6 100644
@@ -285,17 +285,19 @@ static struct pmux_pingrp_config tegra3_pinmux_common[] = {
 	DEFAULT_PINMUX(SPI1_CS0_N_PX6, SPI1, NORMAL, NORMAL, INPUT),
 	DEFAULT_PINMUX(SPI1_MISO_PX7, SPI1, NORMAL, NORMAL, INPUT),
 
-	DEFAULT_PINMUX(PEX_L0_PRSNT_N_PDD0, PCIE, NORMAL, NORMAL, INPUT),
-	DEFAULT_PINMUX(PEX_L0_RST_N_PDD1, PCIE, NORMAL, NORMAL, OUTPUT),
-	DEFAULT_PINMUX(PEX_L0_CLKREQ_N_PDD2, PCIE, NORMAL, NORMAL, INPUT),
-	DEFAULT_PINMUX(PEX_WAKE_N_PDD3, PCIE, NORMAL, NORMAL, INPUT),
+	DEFAULT_PINMUX(PEX_L0_PRSNT_N_PDD0, RSVD2, NORMAL, NORMAL, INPUT),
+	DEFAULT_PINMUX(PEX_L0_RST_N_PDD1, RSVD2, NORMAL, NORMAL, INPUT),
+	DEFAULT_PINMUX(PEX_L0_CLKREQ_N_PDD2, RSVD2, NORMAL, NORMAL, INPUT),
+
+	DEFAULT_PINMUX(PEX_L1_PRSNT_N_PDD4, RSVD2, DOWN, TRISTATE, OUTPUT), /* NC */
+	DEFAULT_PINMUX(PEX_L1_RST_N_PDD5, RSVD2, DOWN, TRISTATE, OUTPUT), /* NC */
+	DEFAULT_PINMUX(PEX_L1_CLKREQ_N_PDD6, RSVD2, DOWN, TRISTATE, OUTPUT), /* NC */
 
-	DEFAULT_PINMUX(PEX_L1_PRSNT_N_PDD4, PCIE, DOWN, TRISTATE, OUTPUT), /* NC */
-	DEFAULT_PINMUX(PEX_L1_RST_N_PDD5, PCIE, DOWN, TRISTATE, OUTPUT), /* NC */
-	DEFAULT_PINMUX(PEX_L1_CLKREQ_N_PDD6, PCIE, DOWN, TRISTATE, OUTPUT), /* NC */
 	DEFAULT_PINMUX(PEX_L2_PRSNT_N_PDD7, PCIE, NORMAL, NORMAL, INPUT),
 	DEFAULT_PINMUX(PEX_L2_RST_N_PCC6, PCIE, NORMAL, NORMAL, OUTPUT),
 	DEFAULT_PINMUX(PEX_L2_CLKREQ_N_PCC7, PCIE, NORMAL, NORMAL, INPUT),
+	DEFAULT_PINMUX(PEX_WAKE_N_PDD3, PCIE, NORMAL, NORMAL, INPUT),
+
 	DEFAULT_PINMUX(HDMI_CEC_PEE3, CEC, NORMAL, NORMAL, INPUT),
 	DEFAULT_PINMUX(HDMI_INT_PN7, RSVD1, NORMAL, NORMAL, INPUT),
 
Signature · 1 line
2.9.4
#3 lore raw

PATCH v2 3/3 apalis_t30: fix optional pcie port reset for reliable pcie operation

In reply to Marcel Ziswiler

From: Marcel Ziswiler <[email protected]>

Allow optionally bringing up the Apalis type specific 4 lane PCIe port as well as the PCIe switch as found on the Apalis Evaluation board. In order to avoid violating the PCIe reset timing do this by overriding the tegra_pcie_board_port_reset() function. Note however that both the Apalis type specific 4 lane PCIe port as well as the regular Apalis PCIe port are also left disabled in the device tree by default.

---

Changes in v2:
- Stick to struct tegra_pcie_port as suggested by Stephen.
- Introduce proper CONFIG_APALIS_T30_PCIE_EVALBOARD_INIT Kconfig option as suggested by Stephen.
- Improved the ifdef vs. if curly braces sequencing as suggested by Stephen.
- Keep PCIe port reset status in order to safeguard for future changes to the port reset order or even allow for re-initialisation should that ever be implemented in the higher levels of the driver model.

 board/toradex/apalis_t30/Kconfig      |  9 ++++++
 board/toradex/apalis_t30/apalis_t30.c | 54 +++++++++++++++++++++++++++++++++++
 2 files changed, 63 insertions(+)
board/toradex/apalis_t30/Kconfig +9 -0
index 16224da..9cd4970 100644
@@ -25,6 +25,15 @@ config TDX_CFG_BLOCK_PART
 config TDX_CFG_BLOCK_OFFSET
 	default "-512"
 
+config APALIS_T30_PCIE_EVALBOARD_INIT
+	bool "Apalis Evaluation Board PCIe Initialisation"
+	help
+	  Bring up the Apalis type specific 4 lane PCIe port as well as the
+	  Apalis PCIe port with the PCIe switch as found on the Apalis
+	  Evaluation board. Note that by default both those ports are also left
+	  disabled in the device tree which needs changing as well for this to
+	  actually work.
+
 source "board/toradex/common/Kconfig"
 
 endif
board/toradex/apalis_t30/apalis_t30.c +54 -0

72 lines not on this page · raw message

Signature · 1 line
2.9.4
#4 lore raw

PATCH v2 3/3 apalis_t30: fix optional pcie port reset for reliable pcie operation

Quoted 23 lines · Marcel Ziswiler

From: Marcel Ziswiler <[email protected]>

Allow optionally bringing up the Apalis type specific 4 lane PCIe port as well as the PCIe switch as found on the Apalis Evaluation board. In order to avoid violating the PCIe reset timing do this by overriding the tegra_pcie_board_port_reset() function. Note however that both the Apalis type specific 4 lane PCIe port as well as the regular Apalis PCIe port are also left disabled in the device tree by default.

board/toradex/apalis_t30/apalis_t30.c +11 -0
 
+void tegra_pcie_board_port_reset(struct tegra_pcie_port *port)
 
+			/*
+			 * Reset PLX PEX 8605 PCIe Switch plus PCIe devices on
+			 * Apalis Evaluation Board
+			 */
+			gpio_direction_output(PEX_PERST_N, 0);
+			gpio_direction_output(RESET_MOCI_CTRL, 0);
+
+			/*
+			 * Must be asserted for 100 ms after power and clocks
+			 * are stable

board/toradex/apalis_t30/apalis_t30.c int tegra_pcie_board_init(void) · line 160

Quoted 10 lines
board/toradex/apalis_t30/apalis_t30.c +10 -0
+			 */
+			mdelay(100);
+
+			gpio_set_value(PEX_PERST_N, 1);
+			/*
+			 * Err_5: PEX_REFCLK_OUTpx/nx Clock Outputs is not
+			 * Guaranteed Until 900 us After PEX_PERST# De-assertion
+			 */
+			mdelay(1);
+			gpio_set_value(RESET_MOCI_CTRL, 1);

Don't we need to also call the core tegra_pcie_port_reset() function here, so that the driver gets to do any reset of the PCIe controller HW that's required? I think that part should happen irrespective of whether CONFIG_APALIS_T30_PCIE_EVALBOARD_INIT is enabled?

#5 lore raw

PATCH v2 3/3 apalis_t30: fix optional pcie port reset for reliable pcie operation

Quoted 39 lines · Stephen Warren
Quoted 38 lines · Marcel Ziswiler

From: Marcel Ziswiler <[email protected]>

Allow optionally bringing up the Apalis type specific 4 lane PCIe port
as well as the PCIe switch as found on the Apalis Evaluation board. In
order to avoid violating the PCIe reset timing do this by overriding the
tegra_pcie_board_port_reset() function. Note however that both the Apalis type specific 4 lane PCIe port as well as the regular Apalis PCIe
port are also left disabled in the device tree by default.

board/toradex/apalis_t30/apalis_t30.c +3 -0
+void tegra_pcie_board_port_reset(struct tegra_pcie_port *port)
+			/*
+			 * Reset PLX PEX 8605 PCIe Switch plus

PCIe devices on

board/toradex/apalis_t30/apalis_t30.c +7 -0
+			 * Apalis Evaluation Board
+			 */
+			gpio_direction_output(PEX_PERST_N, 0);
+			gpio_direction_output(RESET_MOCI_CTRL, 0);
+
+			/*
+			 * Must be asserted for 100 ms after power

and clocks

board/toradex/apalis_t30/apalis_t30.c +7 -0
+			 * are stable
+			 */
+			mdelay(100);
+
+			gpio_set_value(PEX_PERST_N, 1);
+			/*
+			 * Err_5: PEX_REFCLK_OUTpx/nx Clock

Outputs is not

board/toradex/apalis_t30/apalis_t30.c +1 -0
+			 * Guaranteed Until 900 us After

PEX_PERST# De-assertion

board/toradex/apalis_t30/apalis_t30.c +2 -0
+			 */
+			mdelay(1);
Quoted 8 lines

board/toradex/apalis_t30/apalis_t30.c int tegra_pcie_board_init(void) · line 160

Quoted 1 line
board/toradex/apalis_t30/apalis_t30.c +1 -0
+			gpio_set_value(RESET_MOCI_CTRL, 1);

Don't we need to also call the core tegra_pcie_port_reset() function here, so that the driver gets to do any reset of the PCIe controller HW
that's required? I think that part should happen irrespective of whether
CONFIG_APALIS_T30_PCIE_EVALBOARD_INIT is enabled?

I really don't think so. All it really would do is just resetting by the means of the regular PEX_CTL signals which our design is not using like that anyway.

#6 lore raw

PATCH v2 3/3 apalis_t30: fix optional pcie port reset for reliable pcie operation

Quoted 49 lines · Marcel Ziswiler
Quoted 39 lines · Stephen Warren
Quoted 38 lines · Marcel Ziswiler

From: Marcel Ziswiler <[email protected]>

Allow optionally bringing up the Apalis type specific 4 lane PCIe port
as well as the PCIe switch as found on the Apalis Evaluation board. In
order to avoid violating the PCIe reset timing do this by overriding the
tegra_pcie_board_port_reset() function. Note however that both the Apalis type specific 4 lane PCIe port as well as the regular Apalis PCIe
port are also left disabled in the device tree by default.

board/toradex/apalis_t30/apalis_t30.c +3 -0
+void tegra_pcie_board_port_reset(struct tegra_pcie_port *port)
+			/*
+			 * Reset PLX PEX 8605 PCIe Switch plus

PCIe devices on

board/toradex/apalis_t30/apalis_t30.c +7 -0
+			 * Apalis Evaluation Board
+			 */
+			gpio_direction_output(PEX_PERST_N, 0);
+			gpio_direction_output(RESET_MOCI_CTRL, 0);
+
+			/*
+			 * Must be asserted for 100 ms after power

and clocks

board/toradex/apalis_t30/apalis_t30.c +7 -0
+			 * are stable
+			 */
+			mdelay(100);
+
+			gpio_set_value(PEX_PERST_N, 1);
+			/*
+			 * Err_5: PEX_REFCLK_OUTpx/nx Clock

Outputs is not

board/toradex/apalis_t30/apalis_t30.c +1 -0
+			 * Guaranteed Until 900 us After

PEX_PERST# De-assertion

board/toradex/apalis_t30/apalis_t30.c +2 -0
+			 */
+			mdelay(1);
Quoted 8 lines

board/toradex/apalis_t30/apalis_t30.c int tegra_pcie_board_init(void) · line 160

Quoted 1 line
board/toradex/apalis_t30/apalis_t30.c +1 -0
+			gpio_set_value(RESET_MOCI_CTRL, 1);

Don't we need to also call the core tegra_pcie_port_reset() function here, so that the driver gets to do any reset of the PCIe controller HW
that's required? I think that part should happen irrespective of whether
CONFIG_APALIS_T30_PCIE_EVALBOARD_INIT is enabled?

Quoted 1 line

I really don't think so. All it really would do is just resetting by the means of the regular PEX_CTL signals which our design is not using like that anyway.

OK, I guess whatever the outcome, we could always fix it later since it's purely board-specific code. So, this patch,