Skip to contentSkip to conversation

Add support for errors recovery in the TI SN65DSI83 bridge driver

PATCH Superseded Herve Codina 4 participants 16 messages Feb 3, 2025 – Feb 10, 2025

Subsystems (5)

v5 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

v5 cover Add support for errors recovery in the TI SN65DSI83 bridge driver

Hi,

Usually the TI SN65DSI83 recovers from error by itself but during ESD tests, we have some cases where the TI SN65DSI83 didn't recover.

In order to handle those cases, this series adds support for a recovery mechanism.

Compare to the previous iteration, this v5 series fixes the v4 series because the v4 series fails to apply on top of v6.14-rc1.

The modification are still the same as for the v4 series (v4 series should be simply ignored):
- rebase on top of v6.14-rc1
- disable/re-enable irq in the failure detection and reset process
- Split the patch moving reset_pipe() from VC4 to a new helper
- Rename the new helper to drm_atomic_helper_reset_crtc()
- Remove state duplication and use bridge.encoder->crtc

Best regards,
Hervé Codina

Changes v4 -> v5
v4: https://lore.kernel.org/lkml/[email protected]/

Exact same changes as the v4. The v4 series doesn't apply on top of v6.14-rc1 and should be simply ignore.
This v5 series is the v4 fixed to apply on top of v6.14-rc1.

Changes v3 -> v4
v3: https://lore.kernel.org/lkml/[email protected]/

- Patch 1:
No changes

- Patch 2 and 3 (patch 2 in v3):
Rename the helper to drm_atomic_helper_reset_crtc() Split the patch available in v3 in two patches.

- Patch 4 (patch 3 in v3):
Take into account commit d2b8c6d549570 ("drm/bridge: ti-sn65dsi83:
Add ti,lvds-vod-swing optional properties"), available in v6.14-rc1. Disable irq when a fault is detected and re-enable it after the pipe reset.
Remove state duplication and use bridge.encoder->crtc directly

Changes v2 -> v3
v2: https://lore.kernel.org/lkml/[email protected]/

- Patch 1:
No changes

- Patch 2 (new in v3)
Move reset_pipe() from VC4 HDMI driver to a new atomic helper

- Patch 3
Use the new drm_atomic_helper_reset_pipe()

Patch removed in v3
- Patch 2 in v2
No more needed

Changes v1 -> v2
v1: https://lore.kernel.org/lkml/[email protected]/

- Patch 1:
Add 'Reviewed-by: Laurent Pinchart <[email protected]>' Add 'Acked-by: Conor Dooley <[email protected]>'

- Patch 2 (new patch in v2)
Introduce drm_atomic_helper_disable_connector()

- Patch 3 (patch 2 in v1)
Reset the output path instead of the full pipeline. Update and add more information related to the bridge in commit log.

Herve Codina (4):
dt-bindings: display: bridge: sn65dsi83: Add interrupt drm/atomic-helper: Introduce drm_atomic_helper_reset_crtc() drm/vc4: hdmi: Use drm_atomic_helper_reset_crtc()
drm: bridge: ti-sn65dsi83: Add error recovery mechanism

 .../bindings/display/bridge/ti,sn65dsi83.yaml |   3 +
 drivers/gpu/drm/bridge/ti-sn65dsi83.c         | 131 ++++++++++++++++++
 drivers/gpu/drm/drm_atomic_helper.c           |  41 ++++++
 drivers/gpu/drm/vc4/vc4_hdmi.c                |  30 +---
 include/drm/drm_atomic_helper.h               |   2 +
 5 files changed, 178 insertions(+), 29 deletions(-)
Signature · 1 line
2.47.1
#1 lore raw

PATCH v5 1/4 dt-bindings: display: bridge: sn65dsi83: Add interrupt

Both the TI SN65DSI83 and SN65DSI84 bridges have an IRQ pin to signal errors using interrupt.

This interrupt is not documented in the binding.

Add the missing interrupts property.

---
 .../devicetree/bindings/display/bridge/ti,sn65dsi83.yaml       | 3 +++
 1 file changed, 3 insertions(+)
Documentation/devicetree/bindings/display/bridge/ti,sn65dsi83.yaml +3 -0
index bad6f5c81b06..9b5f3f3eab19 100644
@@ -35,6 +35,9 @@ properties:
   vcc-supply:
     description: A 1.8V power supply (see regulator/regulator.yaml).
 
+  interrupts:
+    maxItems: 1
+
   ports:
     $ref: /schemas/graph.yaml#/properties/ports
 
Signature · 1 line
2.47.1
#2 lore raw

PATCH v5 2/4 drm/atomic-helper: Introduce drm_atomic_helper_reset_crtc()

In reply to Herve Codina

drm_atomic_helper_reset_crtc() allows to reset the CRTC active outputs.

This resets all active components available between the CRTC and connectors.

---
 drivers/gpu/drm/drm_atomic_helper.c | 41 +++++++++++++++++++++++++++++
 include/drm/drm_atomic_helper.h     |  2 ++
 2 files changed, 43 insertions(+)
drivers/gpu/drm/drm_atomic_helper.c +41 -0
index 5186d2114a50..0a636c398578 100644
@@ -3362,6 +3362,47 @@ int drm_atomic_helper_disable_all(struct drm_device *dev,
 }
 EXPORT_SYMBOL(drm_atomic_helper_disable_all);
 
+/**
+ * drm_atomic_helper_reset_crtc - reset the active outputs of a CRTC
+ * @crtc: DRM CRTC
+ * @ctx: lock acquisition context
+ *
+ * Reset the active outputs by indicating that connectors have changed.
+ * This implies a reset of all active components available between the CRTC and
+ * connectors.
+ *
+ * Returns:
+ * 0 on success or a negative error code on failure.
+ */
+int drm_atomic_helper_reset_crtc(struct drm_crtc *crtc,
+				 struct drm_modeset_acquire_ctx *ctx)
+{
+	struct drm_atomic_state *state;
+	struct drm_crtc_state *crtc_state;
+	int ret;
+
+	state = drm_atomic_state_alloc(crtc->dev);
+	if (!state)
+		return -ENOMEM;
+
+	state->acquire_ctx = ctx;
+
+	crtc_state = drm_atomic_get_crtc_state(state, crtc);
+	if (IS_ERR(crtc_state)) {
+		ret = PTR_ERR(crtc_state);
+		goto out;
+	}
+
+	crtc_state->connectors_changed = true;
+
+	ret = drm_atomic_commit(state);
+out:
+	drm_atomic_state_put(state);
+
+	return ret;
+}
+EXPORT_SYMBOL(drm_atomic_helper_reset_crtc);
+
 /**
  * drm_atomic_helper_shutdown - shutdown all CRTC
  * @dev: DRM device
include/drm/drm_atomic_helper.h +2 -0
index 9aa0a05aa072..53382fe93537 100644
@@ -139,6 +139,8 @@ int drm_atomic_helper_set_config(struct drm_mode_set *set,
 
 int drm_atomic_helper_disable_all(struct drm_device *dev,
 				  struct drm_modeset_acquire_ctx *ctx);
+int drm_atomic_helper_reset_crtc(struct drm_crtc *crtc,
+				 struct drm_modeset_acquire_ctx *ctx);
 void drm_atomic_helper_shutdown(struct drm_device *dev);
 struct drm_atomic_state *
 drm_atomic_helper_duplicate_state(struct drm_device *dev,
Signature · 1 line
2.47.1
Reviewed-by on v5 2/4
#3 lore raw

Re: [PATCH v5 2/4] drm/atomic-helper: Introduce drm_atomic_helper_reset_crtc()

Quoted 11 lines · Herve Codina

drm_atomic_helper_reset_crtc() allows to reset the CRTC active outputs.

This resets all active components available between the CRTC and connectors.

---
 drivers/gpu/drm/drm_atomic_helper.c | 41 +++++++++++++++++++++++++++++
 include/drm/drm_atomic_helper.h     |  2 ++
 2 files changed, 43 insertions(+)
Signature · 2 lines
With best wishes
Dmitry
maintainer Reviewed-by on v5 2/4
#4 lore raw

Re: [PATCH v5 2/4] drm/atomic-helper: Introduce drm_atomic_helper_reset_crtc()

In reply to Herve Codina
Quoted 7 lines · Herve Codina

drm_atomic_helper_reset_crtc() allows to reset the CRTC active outputs.

This resets all active components available between the CRTC and connectors.

[ ... ]

Thanks!
Maxime

#5 lore raw

PATCH v5 3/4 drm/vc4: hdmi: Use drm_atomic_helper_reset_crtc()

In reply to Herve Codina

The current code uses a the reset_pipe() local function to reset the CRTC outputs.

drm_atomic_helper_reset_crtc() has been introduced recently and it performs exact same operations.

In order to avoid code duplication, use the new helper instead of the local function.

---
 drivers/gpu/drm/vc4/vc4_hdmi.c | 30 +-----------------------------
 1 file changed, 1 insertion(+), 29 deletions(-)
drivers/gpu/drm/vc4/vc4_hdmi.c +1 -29
index 47d9ada98430..e878eddc9c3f 100644
@@ -270,34 +270,6 @@ static void vc4_hdmi_cec_update_clk_div(struct vc4_hdmi *vc4_hdmi)
 static void vc4_hdmi_cec_update_clk_div(struct vc4_hdmi *vc4_hdmi) {}
 #endif
 
-static int reset_pipe(struct drm_crtc *crtc,
-			struct drm_modeset_acquire_ctx *ctx)
-{
-	struct drm_atomic_state *state;
-	struct drm_crtc_state *crtc_state;
-	int ret;
-
-	state = drm_atomic_state_alloc(crtc->dev);
-	if (!state)
-		return -ENOMEM;
-
-	state->acquire_ctx = ctx;
-
-	crtc_state = drm_atomic_get_crtc_state(state, crtc);
-	if (IS_ERR(crtc_state)) {
-		ret = PTR_ERR(crtc_state);
-		goto out;
-	}
-
-	crtc_state->connectors_changed = true;
-
-	ret = drm_atomic_commit(state);
-out:
-	drm_atomic_state_put(state);
-
-	return ret;
-}
-
 static int vc4_hdmi_reset_link(struct drm_connector *connector,
 			       struct drm_modeset_acquire_ctx *ctx)
 {
@@ -376,7 +348,7 @@ static int vc4_hdmi_reset_link(struct drm_connector *connector,
 	 * would be perfectly happy if were to just reconfigure
 	 * the SCDC settings on the fly.
 	 */
-	return reset_pipe(crtc, ctx);
+	return drm_atomic_helper_reset_crtc(crtc, ctx);
 }
 
 static void vc4_hdmi_handle_hotplug(struct vc4_hdmi *vc4_hdmi,
Signature · 1 line
2.47.1
Acked-by on v5 3/4
#6 lore raw

Re: [PATCH v5 3/4] drm/vc4: hdmi: Use drm_atomic_helper_reset_crtc()

Quoted 6 lines · Herve Codina

The current code uses a the reset_pipe() local function to reset the CRTC outputs.

drm_atomic_helper_reset_crtc() has been introduced recently and it performs exact same operations.

Quoted 8 lines

In order to avoid code duplication, use the new helper instead of the local function.

---
 drivers/gpu/drm/vc4/vc4_hdmi.c | 30 +-----------------------------
 1 file changed, 1 insertion(+), 29 deletions(-)
Signature · 2 lines
With best wishes
Dmitry
maintainer Reviewed-by on v5 3/4
#7 lore raw

Re: [PATCH v5 3/4] drm/vc4: hdmi: Use drm_atomic_helper_reset_crtc()

In reply to Herve Codina
Quoted 7 lines · Herve Codina

The current code uses a the reset_pipe() local function to reset the CRTC outputs.

drm_atomic_helper_reset_crtc() has been introduced recently and it performs exact same operations.

[ ... ]

Thanks!
Maxime

#8 lore raw

PATCH v5 4/4 drm: bridge: ti-sn65dsi83: Add error recovery mechanism

In reply to Herve Codina

In some cases observed during ESD tests, the TI SN65DSI83 cannot recover from errors by itself. A full restart of the bridge is needed in those cases to have the bridge output LVDS signals again.

Also, during tests, cases were observed where reading the status of the bridge was not even possible. Indeed, in those cases, the bridge stops to acknowledge I2C transactions. Only a full reset of the bridge (power off/on) brings back the bridge to a functional state.

The TI SN65DSI83 has some error detection capabilities. Introduce an error recovery mechanism based on this detection.

The errors detected are signaled through an interrupt. On system where this interrupt is not available, the driver uses a polling monitoring fallback to check for errors. When an error is present or when reading the bridge status leads to an I2C failure, the recovery process is launched.

Restarting the bridge needs to redo the initialization sequence. This initialization sequence has to be done with the DSI data lanes driven in LP11 state. In order to do that, the recovery process resets the whole output path (i.e the path from the encoder to the connector) where the bridge is located.

---
 drivers/gpu/drm/bridge/ti-sn65dsi83.c | 131 ++++++++++++++++++++++++++
 1 file changed, 131 insertions(+)
drivers/gpu/drm/bridge/ti-sn65dsi83.c +131 -0
index 336380114eea..26a050b13997 100644
@@ -35,9 +35,12 @@
 #include <linux/of_graph.h>
 #include <linux/regmap.h>
 #include <linux/regulator/consumer.h>
+#include <linux/timer.h>
+#include <linux/workqueue.h>
 
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_bridge.h>
+#include <drm/drm_drv.h> /* DRM_MODESET_LOCK_ALL_BEGIN() needs drm_drv_uses_atomic_modeset() */
 #include <drm/drm_mipi_dsi.h>
 #include <drm/drm_of.h>
 #include <drm/drm_panel.h>
@@ -159,6 +162,9 @@ struct sn65dsi83 {
 	bool				lvds_dual_link_even_odd_swap;
 	int				lvds_vod_swing_conf[2];
 	int				lvds_term_conf[2];
+	int				irq;
+	struct delayed_work		monitor_work;
+	struct work_struct		reset_work;
 };
 
 static const struct regmap_range sn65dsi83_readable_ranges[] = {
@@ -363,6 +369,95 @@ static u8 sn65dsi83_get_dsi_div(struct sn65dsi83 *ctx)
 	return dsi_div - 1;
 }
 
+static int sn65dsi83_reset_pipe(struct sn65dsi83 *sn65dsi83)
+{
+	struct drm_device *dev = sn65dsi83->bridge.dev;
+	struct drm_modeset_acquire_ctx ctx;
+	int err;
+
+	/*
+	 * Reset active outputs of the related CRTC.
+	 *
+	 * This way, drm core will reconfigure each components in the CRTC
+	 * outputs path. In our case, this will force the previous component to
+	 * go back in LP11 mode and so allow the reconfiguration of SN64DSI83
+	 * bridge.
+	 *
+	 * Keep the lock during the whole operation to be atomic.
+	 */
+
+	DRM_MODESET_LOCK_ALL_BEGIN(dev, ctx, 0, err);
+
+	err = drm_atomic_helper_reset_crtc(sn65dsi83->bridge.encoder->crtc, &ctx);
+
+	DRM_MODESET_LOCK_ALL_END(dev, ctx, err);
+
+	return err;
+}
+
+static void sn65dsi83_reset_work(struct work_struct *ws)
+{
+	struct sn65dsi83 *ctx = container_of(ws, struct sn65dsi83, reset_work);
+	int ret;
+
+	dev_warn(ctx->dev, "reset the pipe\n");
+
+	/* Reset the pipe */
+	ret = sn65dsi83_reset_pipe(ctx);
+	if (ret) {
+		dev_err(ctx->dev, "reset pipe failed %pe\n", ERR_PTR(ret));
+		return;
+	}
+	if (ctx->irq)
+		enable_irq(ctx->irq);
+}
+
+static void sn65dsi83_handle_errors(struct sn65dsi83 *ctx)
+{
+	unsigned int irq_stat;
+	int ret;
+
+	/*
+	 * Schedule a reset in case of:
+	 *  - the bridge doesn't answer
+	 *  - the bridge signals an error
+	 */
+
+	ret = regmap_read(ctx->regmap, REG_IRQ_STAT, &irq_stat);
+	if (ret || irq_stat) {
+		/*
+		 * IRQ acknowledged is not always possible (the bridge can be in
+		 * a state where it doesn't answer anymore). To prevent an
+		 * interrupt storm, disable interrupt. The interrupt will be
+		 * after the reset.
+		 */
+		if (ctx->irq)
+			disable_irq_nosync(ctx->irq);
+
+		schedule_work(&ctx->reset_work);
+	}
+}
+
+static void sn65dsi83_monitor_work(struct work_struct *work)
+{
+	struct sn65dsi83 *ctx = container_of(to_delayed_work(work),
+					     struct sn65dsi83, monitor_work);
+
+	sn65dsi83_handle_errors(ctx);
+
+	schedule_delayed_work(&ctx->monitor_work, msecs_to_jiffies(1000));
+}
+
+static void sn65dsi83_monitor_start(struct sn65dsi83 *ctx)
+{
+	schedule_delayed_work(&ctx->monitor_work, msecs_to_jiffies(1000));
+}
+
+static void sn65dsi83_monitor_stop(struct sn65dsi83 *ctx)
+{
+	cancel_delayed_work_sync(&ctx->monitor_work);
+}
+
 static void sn65dsi83_atomic_pre_enable(struct drm_bridge *bridge,
 					struct drm_bridge_state *old_bridge_state)
 {
@@ -549,6 +644,15 @@ static void sn65dsi83_atomic_enable(struct drm_bridge *bridge,
 	regmap_read(ctx->regmap, REG_IRQ_STAT, &pval);
 	if (pval)
 		dev_err(ctx->dev, "Unexpected link status 0x%02x\n", pval);
+
+	if (ctx->irq) {
+		/* Enable irq to detect errors */
+		regmap_write(ctx->regmap, REG_IRQ_GLOBAL, REG_IRQ_GLOBAL_IRQ_EN);
+		regmap_write(ctx->regmap, REG_IRQ_EN, 0xff);
+	} else {
+		/* Use the polling task */
+		sn65dsi83_monitor_start(ctx);
+	}
 }
 
 static void sn65dsi83_atomic_disable(struct drm_bridge *bridge,
@@ -557,6 +661,15 @@ static void sn65dsi83_atomic_disable(struct drm_bridge *bridge,
 	struct sn65dsi83 *ctx = bridge_to_sn65dsi83(bridge);
 	int ret;
 
+	if (ctx->irq) {
+		/* Disable irq */
+		regmap_write(ctx->regmap, REG_IRQ_EN, 0x0);
+		regmap_write(ctx->regmap, REG_IRQ_GLOBAL, 0x0);
+	} else {
+		/* Stop the polling task */
+		sn65dsi83_monitor_stop(ctx);
+	}
+
 	/* Put the chip in reset, pull EN line low, and assure 10ms reset low timing. */
 	gpiod_set_value_cansleep(ctx->enable_gpio, 0);
 	usleep_range(10000, 11000);
@@ -806,6 +919,14 @@ static int sn65dsi83_host_attach(struct sn65dsi83 *ctx)
 	return 0;
 }
 
+static irqreturn_t sn65dsi83_irq(int irq, void *data)
+{
+	struct sn65dsi83 *ctx = data;
+
+	sn65dsi83_handle_errors(ctx);
+	return IRQ_HANDLED;
+}
+
 static int sn65dsi83_probe(struct i2c_client *client)
 {
 	const struct i2c_device_id *id = i2c_client_get_device_id(client);
@@ -819,6 +940,8 @@ static int sn65dsi83_probe(struct i2c_client *client)
 		return -ENOMEM;
 
 	ctx->dev = dev;
+	INIT_WORK(&ctx->reset_work, sn65dsi83_reset_work);
+	INIT_DELAYED_WORK(&ctx->monitor_work, sn65dsi83_monitor_work);
 
 	if (dev->of_node) {
 		model = (enum sn65dsi83_model)(uintptr_t)
@@ -843,6 +966,14 @@ static int sn65dsi83_probe(struct i2c_client *client)
 	if (IS_ERR(ctx->regmap))
 		return dev_err_probe(dev, PTR_ERR(ctx->regmap), "failed to get regmap\n");
 
+	if (client->irq) {
+		ctx->irq = client->irq;
+		ret = devm_request_threaded_irq(ctx->dev, ctx->irq, NULL, sn65dsi83_irq,
+						IRQF_ONESHOT, dev_name(ctx->dev), ctx);
+		if (ret)
+			return dev_err_probe(dev, ret, "failed to request irq\n");
+	}
+
 	dev_set_drvdata(dev, ctx);
 	i2c_set_clientdata(client, ctx);
 
Signature · 1 line
2.47.1
maintainer Reviewed-by on v5 4/4
#9 lore raw

Re: [PATCH v5 4/4] drm: bridge: ti-sn65dsi83: Add error recovery mechanism

Quoted 7 lines · Herve Codina

In some cases observed during ESD tests, the TI SN65DSI83 cannot recover from errors by itself. A full restart of the bridge is needed in those cases to have the bridge output LVDS signals again.

Also, during tests, cases were observed where reading the status of the

[ ... ]

Thanks!
Maxime

#10 lore raw

Re: [PATCH v5 4/4] drm: bridge: ti-sn65dsi83: Add error recovery mechanism

In reply to Herve Codina

Hi,

Am Montag, 3. Februar 2025, 17:16:06 CET schrieb Herve Codina:

Quoted 18 lines

In some cases observed during ESD tests, the TI SN65DSI83 cannot recover from errors by itself. A full restart of the bridge is needed in those cases to have the bridge output LVDS signals again.

Also, during tests, cases were observed where reading the status of the bridge was not even possible. Indeed, in those cases, the bridge stops to acknowledge I2C transactions. Only a full reset of the bridge (power off/on) brings back the bridge to a functional state.

The TI SN65DSI83 has some error detection capabilities. Introduce an error recovery mechanism based on this detection.

The errors detected are signaled through an interrupt. On system where this interrupt is not available, the driver uses a polling monitoring fallback to check for errors. When an error is present or when reading the bridge status leads to an I2C failure, the recovery process is launched.

Quoted 8 lines

Restarting the bridge needs to redo the initialization sequence. This initialization sequence has to be done with the DSI data lanes driven in LP11 state. In order to do that, the recovery process resets the whole output path (i.e the path from the encoder to the connector) where the bridge is located.

---

With interrupt configured I got the following stack trace upon reboot/poweroff:

[   91.317264] sn65dsi83 2-002d: reset the pipe
[   91.344093] Unable to handle ke
** replaying previous printk message **
[   91.344093] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
[   91.344107] Mem abort info:
[   91.344111]   ESR = 0x0000000096000004
[   91.344115]   EC = 0x25: DABT (current EL), IL = 32 bits
[   91.344120]   SET = 0, FnV = 0
[   91.344122]   EA = 0, S1PTW = 0
[   91.344125]   FSC = 0x04: level 0 translation fault
[   91.344128] Data abort info:
[   91.344130]   ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000
[   91.344133]   CM = 0, WnR = 0, TnD = 0, TagAccess = 0
[   91.344136]   GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
[   91.344141] user pgtable: 4k pages, 48-bit VAs, pgdp=0000000102eec000
[   91.344145] [0000000000000000] pgd=0000000000000000, p4d=0000000000000000
[   91.344155] Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP
[   91.420670] Modules linked in: bluetooth 8021q garp mrp stp llc hid_generic hantro_vpu snd_soc_fsl_asoc_card snd_soc_fsl_sai snd_soc_imx_audmux snd_soc_t

lv320aic32x4_i2c imx_pcm_dma v4l2_vp9 snd_soc_simple_card_utils snd_soc_tlv320aic32x4 snd_soc_fsl_utils v4l2_h264 v4l2_jpeg snd_soc_core videobuf2_dma_conti g v4l2_mem2mem videobuf2_memops videobuf2_v4l2 snd_pcm_dmaengine videobuf2_common imx8m_ddrc phy_generic snd_pcm ci_hdrc_imx ti_sn65dsi83 ci_hdrc clk_renesa s_pcie mxsfb snd_timer samsung_dsim usbmisc_imx pwm_imx27 snd soundcore spi_imx imx_sdma imx8mm_thermal panel_simple gpio_aggregator pwm_beeper cfg80211 rfk ill fuse ipv6

[   91.476291] CPU: 0 UID: 0 PID: 83 Comm: kworker/0:3 Not tainted 6.14.0-rc1-next-20250205+ #2906 da26d4e444ec7c54f82096af1ee2c91e843e9303
[   91.488555] Hardware name: TQ-Systems GmbH i.MX8MM TQMa8MxML on MBa8Mx (DT)
[   91.495517] Workqueue: events sn65dsi83_reset_work [ti_sn65dsi83]
[   91.501623] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[   91.508588] pc : drm_atomic_helper_reset_crtc+0x18/0x100
[   91.513906] lr : sn65dsi83_reset_work+0x80/0x118 [ti_sn65dsi83]
[   91.519832] sp : ffff80008271bcc0
[   91.523146] x29: ffff80008271bcc0 x28: 0000000000000000 x27: 0000000000000000
[   91.530288] x26: ffff0000c0d6e340 x25: 0000000000000000 x24: ffff0000c0022005
[   91.537432] x23: ffff0000c4cb4248 x22: ffff0000ff736e40 x21: ffff80008271bcf8
[   91.544576] x20: 0000000000000000 x19: 0000000000000000 x18: 0000000000000000
[   91.551719] x17: 000000040044ffff x16: 0000000000000000 x15: 00000015425993c0
[   91.558860] x14: 0000000000000001 x13: ffff0000c0ece800 x12: 0000000000000020
[   91.566004] x11: ffff0000c0ece800 x10: 0000000000000af0 x9 : ffff80008271ba70
[   91.573147] x8 : ffff0000c0ecf2d0 x7 : ffff0000c0ece800 x6 : dead000000000122
[   91.580291] x5 : 0000000000000000 x4 : ffff0000c0ece780 x3 : 0000000000000000
[   91.587436] x2 : ffff0000c4cb40b8 x1 : ffff80008271bcf8 x0 : 0000000000000000
[   91.594580] Call trace:
[   91.597027]  drm_atomic_helper_reset_crtc+0x18/0x100 (P)
[   91.602346]  sn65dsi83_reset_work+0x80/0x118 [ti_sn65dsi83 d96dc31a1413a92a7c205a475a2357ddacaa4a3b]
[   91.611485]  process_one_work+0x14c/0x3e0
[   91.615503]  worker_thread+0x2d0/0x3f0
[   91.619257]  kthread+0x110/0x1e0
[   91.622488]  ret_from_fork+0x10/0x20
[   91.626073] Code: a90153f3 aa0003f4 f90013f5 aa0103f5 (f9400000) 
[   91.632167] ---[ end trace 0000000000000000 ]---

Running with workqueue does not raise this error.

Best regards,
Alexander

Quoted 198 lines
 drivers/gpu/drm/bridge/ti-sn65dsi83.c | 131 ++++++++++++++++++++++++++
 1 file changed, 131 insertions(+)
drivers/gpu/drm/bridge/ti-sn65dsi83.c +131 -0
index 336380114eea..26a050b13997 100644
@@ -35,9 +35,12 @@
 #include <linux/of_graph.h>
 #include <linux/regmap.h>
 #include <linux/regulator/consumer.h>
+#include <linux/timer.h>
+#include <linux/workqueue.h>
 
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_bridge.h>
+#include <drm/drm_drv.h> /* DRM_MODESET_LOCK_ALL_BEGIN() needs drm_drv_uses_atomic_modeset() */
 #include <drm/drm_mipi_dsi.h>
 #include <drm/drm_of.h>
 #include <drm/drm_panel.h>
@@ -159,6 +162,9 @@ struct sn65dsi83 {
 	bool				lvds_dual_link_even_odd_swap;
 	int				lvds_vod_swing_conf[2];
 	int				lvds_term_conf[2];
+	int				irq;
+	struct delayed_work		monitor_work;
+	struct work_struct		reset_work;
 };
 
 static const struct regmap_range sn65dsi83_readable_ranges[] = {
@@ -363,6 +369,95 @@ static u8 sn65dsi83_get_dsi_div(struct sn65dsi83 *ctx)
 	return dsi_div - 1;
 }
 
+static int sn65dsi83_reset_pipe(struct sn65dsi83 *sn65dsi83)
+{
+	struct drm_device *dev = sn65dsi83->bridge.dev;
+	struct drm_modeset_acquire_ctx ctx;
+	int err;
+
+	/*
+	 * Reset active outputs of the related CRTC.
+	 *
+	 * This way, drm core will reconfigure each components in the CRTC
+	 * outputs path. In our case, this will force the previous component to
+	 * go back in LP11 mode and so allow the reconfiguration of SN64DSI83
+	 * bridge.
+	 *
+	 * Keep the lock during the whole operation to be atomic.
+	 */
+
+	DRM_MODESET_LOCK_ALL_BEGIN(dev, ctx, 0, err);
+
+	err = drm_atomic_helper_reset_crtc(sn65dsi83->bridge.encoder->crtc, &ctx);
+
+	DRM_MODESET_LOCK_ALL_END(dev, ctx, err);
+
+	return err;
+}
+
+static void sn65dsi83_reset_work(struct work_struct *ws)
+{
+	struct sn65dsi83 *ctx = container_of(ws, struct sn65dsi83, reset_work);
+	int ret;
+
+	dev_warn(ctx->dev, "reset the pipe\n");
+
+	/* Reset the pipe */
+	ret = sn65dsi83_reset_pipe(ctx);
+	if (ret) {
+		dev_err(ctx->dev, "reset pipe failed %pe\n", ERR_PTR(ret));
+		return;
+	}
+	if (ctx->irq)
+		enable_irq(ctx->irq);
+}
+
+static void sn65dsi83_handle_errors(struct sn65dsi83 *ctx)
+{
+	unsigned int irq_stat;
+	int ret;
+
+	/*
+	 * Schedule a reset in case of:
+	 *  - the bridge doesn't answer
+	 *  - the bridge signals an error
+	 */
+
+	ret = regmap_read(ctx->regmap, REG_IRQ_STAT, &irq_stat);
+	if (ret || irq_stat) {
+		/*
+		 * IRQ acknowledged is not always possible (the bridge can be in
+		 * a state where it doesn't answer anymore). To prevent an
+		 * interrupt storm, disable interrupt. The interrupt will be
+		 * after the reset.
+		 */
+		if (ctx->irq)
+			disable_irq_nosync(ctx->irq);
+
+		schedule_work(&ctx->reset_work);
+	}
+}
+
+static void sn65dsi83_monitor_work(struct work_struct *work)
+{
+	struct sn65dsi83 *ctx = container_of(to_delayed_work(work),
+					     struct sn65dsi83, monitor_work);
+
+	sn65dsi83_handle_errors(ctx);
+
+	schedule_delayed_work(&ctx->monitor_work, msecs_to_jiffies(1000));
+}
+
+static void sn65dsi83_monitor_start(struct sn65dsi83 *ctx)
+{
+	schedule_delayed_work(&ctx->monitor_work, msecs_to_jiffies(1000));
+}
+
+static void sn65dsi83_monitor_stop(struct sn65dsi83 *ctx)
+{
+	cancel_delayed_work_sync(&ctx->monitor_work);
+}
+
 static void sn65dsi83_atomic_pre_enable(struct drm_bridge *bridge,
 					struct drm_bridge_state *old_bridge_state)
 {
@@ -549,6 +644,15 @@ static void sn65dsi83_atomic_enable(struct drm_bridge *bridge,
 	regmap_read(ctx->regmap, REG_IRQ_STAT, &pval);
 	if (pval)
 		dev_err(ctx->dev, "Unexpected link status 0x%02x\n", pval);
+
+	if (ctx->irq) {
+		/* Enable irq to detect errors */
+		regmap_write(ctx->regmap, REG_IRQ_GLOBAL, REG_IRQ_GLOBAL_IRQ_EN);
+		regmap_write(ctx->regmap, REG_IRQ_EN, 0xff);
+	} else {
+		/* Use the polling task */
+		sn65dsi83_monitor_start(ctx);
+	}
 }
 
 static void sn65dsi83_atomic_disable(struct drm_bridge *bridge,
@@ -557,6 +661,15 @@ static void sn65dsi83_atomic_disable(struct drm_bridge *bridge,
 	struct sn65dsi83 *ctx = bridge_to_sn65dsi83(bridge);
 	int ret;
 
+	if (ctx->irq) {
+		/* Disable irq */
+		regmap_write(ctx->regmap, REG_IRQ_EN, 0x0);
+		regmap_write(ctx->regmap, REG_IRQ_GLOBAL, 0x0);
+	} else {
+		/* Stop the polling task */
+		sn65dsi83_monitor_stop(ctx);
+	}
+
 	/* Put the chip in reset, pull EN line low, and assure 10ms reset low timing. */
 	gpiod_set_value_cansleep(ctx->enable_gpio, 0);
 	usleep_range(10000, 11000);
@@ -806,6 +919,14 @@ static int sn65dsi83_host_attach(struct sn65dsi83 *ctx)
 	return 0;
 }
 
+static irqreturn_t sn65dsi83_irq(int irq, void *data)
+{
+	struct sn65dsi83 *ctx = data;
+
+	sn65dsi83_handle_errors(ctx);
+	return IRQ_HANDLED;
+}
+
 static int sn65dsi83_probe(struct i2c_client *client)
 {
 	const struct i2c_device_id *id = i2c_client_get_device_id(client);
@@ -819,6 +940,8 @@ static int sn65dsi83_probe(struct i2c_client *client)
 		return -ENOMEM;
 
 	ctx->dev = dev;
+	INIT_WORK(&ctx->reset_work, sn65dsi83_reset_work);
+	INIT_DELAYED_WORK(&ctx->monitor_work, sn65dsi83_monitor_work);
 
 	if (dev->of_node) {
 		model = (enum sn65dsi83_model)(uintptr_t)
@@ -843,6 +966,14 @@ static int sn65dsi83_probe(struct i2c_client *client)
 	if (IS_ERR(ctx->regmap))
 		return dev_err_probe(dev, PTR_ERR(ctx->regmap), "failed to get regmap\n");
 
+	if (client->irq) {
+		ctx->irq = client->irq;
+		ret = devm_request_threaded_irq(ctx->dev, ctx->irq, NULL, sn65dsi83_irq,
+						IRQF_ONESHOT, dev_name(ctx->dev), ctx);
+		if (ret)
+			return dev_err_probe(dev, ret, "failed to request irq\n");
+	}
+
 	dev_set_drvdata(dev, ctx);
 	i2c_set_clientdata(client, ctx);
 
Signature · 4 lines
TQ-Systems GmbH | Mühlstraße 2, Gut Delling | 82229 Seefeld, Germany
Amtsgericht München, HRB 105018
Geschäftsführer: Detlef Schneider, Rüdiger Stahl, Stefan Schneider
http://www.tq-group.com/
#11 lore raw

Re: [PATCH v5 4/4] drm: bridge: ti-sn65dsi83: Add error recovery mechanism

Hi Alexander,

On Thu, 06 Feb 2025 15:38:42 +0100
Alexander Stein <[email protected]> wrote:

...

Quoted 7 lines

With interrupt configured I got the following stack trace upon reboot/poweroff:

[   91.317264] sn65dsi83 2-002d: reset the pipe
[   91.344093] Unable to handle ke
** replaying previous printk message **
[   91.344093] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000

...

Ouch :(

One question to help me investigating:
Do you have the issue at init/probe or when you start to display graphics?

Best regards,
Hervé

#12 lore raw

Re: [PATCH v5 4/4] drm: bridge: ti-sn65dsi83: Add error recovery mechanism

Hi Herve,

Am Donnerstag, 6. Februar 2025, 16:20:48 CET schrieb Herve Codina:

Quoted 13 lines

Hi Alexander,

On Thu, 06 Feb 2025 15:38:42 +0100
Alexander Stein <[email protected]> wrote:

...

Quoted 7 lines

With interrupt configured I got the following stack trace upon reboot/poweroff:

[   91.317264] sn65dsi83 2-002d: reset the pipe
[   91.344093] Unable to handle ke
** replaying previous printk message **
[   91.344093] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
Quoted 6 lines

...

Ouch :(

One question to help me investigating:
Do you have the issue at init/probe or when you start to display graphics?

This is during shutdown/poweroff. I assume that regmap_reg() in sn65dsi83_handle_errors() fails and because of that reset_work is scheduled.

Best regards,
Alexander

Signature · 4 lines
TQ-Systems GmbH | Mühlstraße 2, Gut Delling | 82229 Seefeld, Germany
Amtsgericht München, HRB 105018
Geschäftsführer: Detlef Schneider, Rüdiger Stahl, Stefan Schneider
http://www.tq-group.com/
#13 lore raw

Re: [PATCH v5 4/4] drm: bridge: ti-sn65dsi83: Add error recovery mechanism

Hi Alexander,

Quoted 18 lines · On Thu, 06 Feb 2025 16:39:09 +0100 Alexander Stein…

Hi Herve,

Am Donnerstag, 6. Februar 2025, 16:20:48 CET schrieb Herve Codina:

Quoted 15 lines

Hi Alexander,

On Thu, 06 Feb 2025 15:38:42 +0100
Alexander Stein <[email protected]> wrote:

...

Quoted 7 lines

With interrupt configured I got the following stack trace upon reboot/poweroff:

[   91.317264] sn65dsi83 2-002d: reset the pipe
[   91.344093] Unable to handle ke
** replaying previous printk message **
[   91.344093] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000  

...

Quoted 8 lines
Quoted 4 lines

Ouch :(

One question to help me investigating:
Do you have the issue at init/probe or when you start to display graphics?

This is during shutdown/poweroff. I assume that regmap_reg() in sn65dsi83_handle_errors() fails and because of that reset_work is scheduled.

Found the issue.

Can you give me following information so that I can validate what I understood:
- Is your interrupt line connected directly to the SoC?
- Is there any pullup/pulldown on your interrupt line?
- In your devicetree what is the configuration used for this interrupt in terms of level or edge.

Anyway, I will send a fix in the next iteration.

Best regards,
Hervé

#14 lore raw

Re: [PATCH v5 4/4] drm: bridge: ti-sn65dsi83: Add error recovery mechanism

Hi Herve,

Am Freitag, 7. Februar 2025, 19:08:16 CET schrieb Herve Codina:

Quoted 32 lines

Hi Alexander,

Quoted 18 lines · On Thu, 06 Feb 2025 16:39:09 +0100 Alexander Stein…

Hi Herve,

Am Donnerstag, 6. Februar 2025, 16:20:48 CET schrieb Herve Codina:

Quoted 15 lines

Hi Alexander,

On Thu, 06 Feb 2025 15:38:42 +0100
Alexander Stein <[email protected]> wrote:

...

Quoted 7 lines

With interrupt configured I got the following stack trace upon reboot/poweroff:

[   91.317264] sn65dsi83 2-002d: reset the pipe
[   91.344093] Unable to handle ke
** replaying previous printk message **
[   91.344093] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000  

...

Quoted 8 lines
Quoted 4 lines

Ouch :(

One question to help me investigating:
Do you have the issue at init/probe or when you start to display graphics?

This is during shutdown/poweroff. I assume that regmap_reg() in sn65dsi83_handle_errors() fails and because of that reset_work is scheduled.

Quoted 5 lines

Found the issue.

Can you give me following information so that I can validate what I understood:
- Is your interrupt line connected directly to the SoC?

No, unfortunately not. It's connected to a GPIO expander, which in turn delivers an IRQ to the SoC. In between the bridge and the expander is a buffer for voltage switch (1V8 -> 3V3).

Quoted 1 line

- Is there any pullup/pulldown on your interrupt line?

As far a I can tell, there is no pullup/pulldown on the IRQ line to the buffer.

Quoted 2 lines

- In your devicetree what is the configuration used for this interrupt in terms of level or edge.

As this line is connected to an expander (pca9555) only edge triggered interrupts are supported.

Quoted 1 line

Anyway, I will send a fix in the next iteration.

Thanks for digging into this.
Best regards,
Alexander

Signature · 4 lines
TQ-Systems GmbH | Mühlstraße 2, Gut Delling | 82229 Seefeld, Germany
Amtsgericht München, HRB 105018
Geschäftsführer: Detlef Schneider, Rüdiger Stahl, Stefan Schneider
http://www.tq-group.com/
#15 lore raw

Re: [PATCH v5 4/4] drm: bridge: ti-sn65dsi83: Add error recovery mechanism

Hi Alexander,

Quoted 47 lines · On Mon, 10 Feb 2025 11:42:09 +0100 Alexander Stein…

Hi Herve,

Am Freitag, 7. Februar 2025, 19:08:16 CET schrieb Herve Codina:

Quoted 32 lines

Hi Alexander,

Quoted 18 lines · On Thu, 06 Feb 2025 16:39:09 +0100 Alexander Stein…

Hi Herve,

Am Donnerstag, 6. Februar 2025, 16:20:48 CET schrieb Herve Codina:

Quoted 15 lines

Hi Alexander,

On Thu, 06 Feb 2025 15:38:42 +0100
Alexander Stein <[email protected]> wrote:

...

Quoted 7 lines

With interrupt configured I got the following stack trace upon reboot/poweroff:

[   91.317264] sn65dsi83 2-002d: reset the pipe
[   91.344093] Unable to handle ke
** replaying previous printk message **
[   91.344093] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000    

...

Quoted 8 lines
Quoted 4 lines

Ouch :(

One question to help me investigating:
Do you have the issue at init/probe or when you start to display graphics?

This is during shutdown/poweroff. I assume that regmap_reg() in sn65dsi83_handle_errors() fails and because of that reset_work is scheduled.

Quoted 5 lines

Found the issue.

Can you give me following information so that I can validate what I understood:
- Is your interrupt line connected directly to the SoC?

No, unfortunately not. It's connected to a GPIO expander, which in turn delivers an IRQ to the SoC. In between the bridge and the expander is a buffer for voltage switch (1V8 -> 3V3).

Quoted 1 line

- Is there any pullup/pulldown on your interrupt line?

Quoted 11 lines

As far a I can tell, there is no pullup/pulldown on the IRQ line to the buffer.

Quoted 2 lines

- In your devicetree what is the configuration used for this interrupt in terms of level or edge.

As this line is connected to an expander (pca9555) only edge triggered interrupts are supported.

Quoted 1 line

Anyway, I will send a fix in the next iteration.

Thanks for all those details.

I've just sent a new iteration:
https://lore.kernel.org/all/[email protected]/

Can you test it on your side?

Best regards,
Hervé