Skip to contentSkip to conversation

usb: dwc2: skip polling for the soft reset clear bit and add after reset delay

PATCH In review [email protected] 2 participants 4 messages Jun 23 – Aug 26

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 usb: dwc2: skip polling for the soft reset clear bit and add after reset delay

From: Chali Anis <[email protected]>

The RPi4 SoC do not deassert GRSTCTL_CSFTRST within the expected window but continue to operate correctly, causing probe to fail with ETIMEDOUT.

Per the datasheet, GRSTCTL_CSFTRST is self-clearing but requires at least 3 PHY clocks after reset before any PHY domain access. Add a 1us delay to satisfy this requirement.

---
 drivers/usb/dwc2/core.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
drivers/usb/dwc2/core.c +9 -4
index 60cc690fdbc0..6d4b7930da9e 100644
@@ -809,11 +809,16 @@ int dwc2_core_reset(struct dwc2 *dwc2)
 	greset |= GRSTCTL_CSFTRST;
 	dwc2_writel(dwc2, greset, GRSTCTL);
 
-	ret = dwc2_wait_bit_clear(dwc2, GRSTCTL, GRSTCTL_CSFTRST, 10000);
-	if (ret) {
-		dwc2_warn(dwc2, "%s: Timeout! Waiting for Core Soft Reset\n",
+	/* Wait for at least 3 PHY Clocks */
+	udelay(1);
+
+	if (!of_machine_is_compatible("brcm,bcm2711")) {
+		ret = dwc2_wait_bit_clear(dwc2, GRSTCTL, GRSTCTL_CSFTRST, 10000);
+		if (ret) {
+			dwc2_warn(dwc2, "%s: Timeout! Waiting for Core Soft Reset\n",
 				__func__);
-		return ret;
+			return ret;
+		}
 	}
 
 	if (wait_for_host_mode)
#1 lore raw
Quoted 5 lines · On 2026-06-23 10:23, chalianis1 wrote:

From: Chali Anis <[email protected]>

The RPi4 SoC do not deassert GRSTCTL_CSFTRST within the expected window but continue to operate correctly, causing probe to fail with ETIMEDOUT.

Does "within the expected window" imply that there is a bigger window in which it is de-asserted?

Quoted 4 lines

Per the datasheet, GRSTCTL_CSFTRST is self-clearing but requires at least 3 PHY clocks after reset before any PHY domain access. Add a 1us delay to satisfy this requirement.

Which datasheet, the SoC datasheet or a DWC2 datasheet?

I'm just asking if there is a chance to not need a SoC check here.

Sascha

Quoted 32 lines
---
 drivers/usb/dwc2/core.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
drivers/usb/dwc2/core.c +9 -4
index 60cc690fdbc0..6d4b7930da9e 100644
@@ -809,11 +809,16 @@ int dwc2_core_reset(struct dwc2 *dwc2)
 	greset |= GRSTCTL_CSFTRST;
 	dwc2_writel(dwc2, greset, GRSTCTL);
 
-	ret = dwc2_wait_bit_clear(dwc2, GRSTCTL, GRSTCTL_CSFTRST, 10000);
-	if (ret) {
-		dwc2_warn(dwc2, "%s: Timeout! Waiting for Core Soft Reset\n",
+	/* Wait for at least 3 PHY Clocks */
+	udelay(1);
+
+	if (!of_machine_is_compatible("brcm,bcm2711")) {
+		ret = dwc2_wait_bit_clear(dwc2, GRSTCTL, GRSTCTL_CSFTRST, 10000);
+		if (ret) {
+			dwc2_warn(dwc2, "%s: Timeout! Waiting for Core Soft Reset\n",
 				__func__);
-		return ret;
+			return ret;
+		}
 	}
 
 	if (wait_for_host_mode)
Signature · 4 lines
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
#2 lore raw
Quoted 10 lines · Sascha Hauer
Quoted 5 lines · On 2026-06-23 10:23, chalianis1 wrote:

From: Chali Anis <[email protected]>

The RPi4 SoC do not deassert GRSTCTL_CSFTRST within the expected window but continue to operate correctly, causing probe to fail with ETIMEDOUT.

Does "within the expected window" imply that there is a bigger window in which it is de-asserted?

What I mean is within the expected window according to the SoC datasheet of RK3588 which I just realized that I wrongly thought that it has a DWC2. Rockchip RK3588 TRM V1.0-Part2 20220309 page 2221

Quoted 7 lines
Quoted 4 lines

Per the datasheet, GRSTCTL_CSFTRST is self-clearing but requires at least 3 PHY clocks after reset before any PHY domain access. Add a 1us delay to satisfy this requirement.

Which datasheet, the SoC datasheet or a DWC2 datasheet?

the SOC one.

Quoted 34 lines

I'm just asking if there is a chance to not need a SoC check here.

Sascha

Quoted 30 lines
---
 drivers/usb/dwc2/core.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
drivers/usb/dwc2/core.c +9 -4
index 60cc690fdbc0..6d4b7930da9e 100644
@@ -809,11 +809,16 @@ int dwc2_core_reset(struct dwc2 *dwc2)
 	greset |= GRSTCTL_CSFTRST;
 	dwc2_writel(dwc2, greset, GRSTCTL);
 
-	ret = dwc2_wait_bit_clear(dwc2, GRSTCTL, GRSTCTL_CSFTRST, 10000);
-	if (ret) {
-		dwc2_warn(dwc2, "%s: Timeout! Waiting for Core Soft Reset\n",
+	/* Wait for at least 3 PHY Clocks */
+	udelay(1);
+
+	if (!of_machine_is_compatible("brcm,bcm2711")) {
+		ret = dwc2_wait_bit_clear(dwc2, GRSTCTL, GRSTCTL_CSFTRST, 10000);
+		if (ret) {
+			dwc2_warn(dwc2, "%s: Timeout! Waiting for Core Soft Reset\n",
 				__func__);
-		return ret;
+			return ret;
+		}
 	}
 
Quoted 11 lines
Quoted 2 lines
drivers/usb/dwc2/core.c +0 -0
 	if (wait_for_host_mode)
Signature · 4 lines
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

Best regards.
Anis

#3 lore raw
Quoted 68 lines · On 2026-08-26 13:54 +0200, [email protected] wrote:
Quoted 10 lines · Sascha Hauer
Quoted 5 lines · On 2026-06-23 10:23, chalianis1 wrote:

From: Chali Anis <[email protected]>

The RPi4 SoC do not deassert GRSTCTL_CSFTRST within the expected window but continue to operate correctly, causing probe to fail with ETIMEDOUT.

Does "within the expected window" imply that there is a bigger window in which it is de-asserted?

What I mean is within the expected window according to the SoC datasheet of RK3588 which I just realized that I wrongly thought that it has a DWC2. Rockchip RK3588 TRM V1.0-Part2 20220309 page 2221

Quoted 7 lines
Quoted 4 lines

Per the datasheet, GRSTCTL_CSFTRST is self-clearing but requires at least 3 PHY clocks after reset before any PHY domain access. Add a 1us delay to satisfy this requirement.

Which datasheet, the SoC datasheet or a DWC2 datasheet?

the SOC one.

Quoted 34 lines

I'm just asking if there is a chance to not need a SoC check here.

Sascha

Quoted 30 lines
---
 drivers/usb/dwc2/core.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
drivers/usb/dwc2/core.c +9 -4
index 60cc690fdbc0..6d4b7930da9e 100644
@@ -809,11 +809,16 @@ int dwc2_core_reset(struct dwc2 *dwc2)
 	greset |= GRSTCTL_CSFTRST;
 	dwc2_writel(dwc2, greset, GRSTCTL);
 
-	ret = dwc2_wait_bit_clear(dwc2, GRSTCTL, GRSTCTL_CSFTRST, 10000);
-	if (ret) {
-		dwc2_warn(dwc2, "%s: Timeout! Waiting for Core Soft Reset\n",
+	/* Wait for at least 3 PHY Clocks */
+	udelay(1);
+
+	if (!of_machine_is_compatible("brcm,bcm2711")) {
+		ret = dwc2_wait_bit_clear(dwc2, GRSTCTL, GRSTCTL_CSFTRST, 10000);
+		if (ret) {
+			dwc2_warn(dwc2, "%s: Timeout! Waiting for Core Soft Reset\n",
 				__func__);
-		return ret;
+			return ret;
+		}
 	}
 
Quoted 11 lines
Quoted 2 lines
drivers/usb/dwc2/core.c +0 -0
 	if (wait_for_host_mode)
Signature · 4 lines
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
Quoted 6 lines

Best regards.
Anis

I reply again on the thread because I tested something that works by adding a delay on the the top of dwc2_core_reset the issue is resolved and corresponds more for what the SoC even if it is not the good datasheet stipulates see below.

drivers/usb/dwc2/core.c +1 -0
index 60cc690fdb..1fb8e2c6ac 100644
@@ -776,6 +776,7 @@ int dwc2_core_reset(struct dwc2 *dwc2)
 
        dwc2_dbg(dwc2, "%s(%p)\n", __func__, dwc2);
 
+       udelay(10);
        /* Wait for AHB master IDLE state. */
        ret = dwc2_wait_bit_set(dwc2, GRSTCTL, GRSTCTL_AHBIDLE, 10000);
        if (ret) {

Best regards,
Anis