Skip to contentSkip to conversation

updates for i3c error printing

RFC In review Ben Dooks 2 participants 9 messages Jun 21, 2023 – Jul 11, 2023

Subsystems (2)
lore
Share & tools
lore

Ben Dooks

author
#0 lore raw

cover updates for i3c error printing

During work on an i3c compatible system some of the probing does not produce much useful error output. This series is an attempt to add error reporting where it would be useful to track down the culprit.

Ben Dooks (5):
i3c: show error with node for invalid reg property
i3c: add error print to show device failing during populate bus
i3c: show node when printing unsupported 10-bit i2c dev
i3c: show error messages in of_i3c_master_add_i3c_boardinfo
i3c: dw; add print if cannot get resources

 drivers/i3c/master.c               | 28 +++++++++++++++++++++-------
 drivers/i3c/master/dw-i3c-master.c | 12 +++++++++---
 2 files changed, 30 insertions(+), 10 deletions(-)
Signature · 1 line
2.40.1
author
#1 lore raw

RFC 1/5 i3c: show error with node for invalid reg property

When adding i3c devices via fdt, it is useful to get an error if the reg node is wrong (say, you thought it was an i2c device and only needed one value here) and have the error show the OF node which was the cause of the problem (in case of many devices in the system),

---
 drivers/i3c/master.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/i3c/master.c +3 -1
index 08aeb69a7800..559fc2781a81 100644
@@ -2094,8 +2094,10 @@ static int of_i3c_master_add_dev(struct i3c_master_controller *master,
 		return -EINVAL;
 
 	ret = of_property_read_u32_array(node, "reg", reg, ARRAY_SIZE(reg));
-	if (ret)
+	if (ret) {
+		dev_err(&master->dev, "%pOF: invalid reg property", node);
 		return ret;
+	}
 
 	/*
 	 * The manufacturer ID can't be 0. If reg[1] == 0 that means we're
Signature · 1 line
2.40.1
author
#2 lore raw

RFC 2/5 i3c: add error print to show device failing during populate bus

In reply to Ben Dooks

The of_populate_i3c_bus() does not produce much helpful output when a device fails to add, so addd an explicit dev_err() showing the device node that failed and the error code it failed with. This should make finding bad device-tree entries easier.

---
 drivers/i3c/master.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/i3c/master.c +2 -0
index 559fc2781a81..6316f3fc914a 100644
@@ -2125,6 +2125,8 @@ static int of_populate_i3c_bus(struct i3c_master_controller *master)
 	for_each_available_child_of_node(i3cbus_np, node) {
 		ret = of_i3c_master_add_dev(master, node);
 		if (ret) {
+			dev_err(dev, "%pOF: failed to add device (%d)\n",
+				node, ret);
 			of_node_put(node);
 			return ret;
 		}
Signature · 1 line
2.40.1
author
#3 lore raw

RFC 3/5 i3c: show node when printing unsupported 10-bit i2c dev

In reply to Ben Dooks

When an i2c device is errored due to not being supported, the OF node is not show (unlike errors from of_i2c_get_board_info). Make the failed node explict in the of_i3c_master_add_i2c_boardinfo() to aid in tracking down incorrect entries in the device tree.

---
 drivers/i3c/master.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/i3c/master.c +1 -1
index 6316f3fc914a..bc42669f5c6d 100644
@@ -2023,7 +2023,7 @@ of_i3c_master_add_i2c_boardinfo(struct i3c_master_controller *master,
 	 * DEFSLVS command.
 	 */
 	if (boardinfo->base.flags & I2C_CLIENT_TEN) {
-		dev_err(dev, "I2C device with 10 bit address not supported.");
+		dev_err(dev, "%pOF: I2C device with 10 bit address not supported.", node);
 		return -ENOTSUPP;
 	}
 
Signature · 1 line
2.40.1
author
#4 lore raw

RFC 4/5 i3c: show error messages in of_i3c_master_add_i3c_boardinfo

In reply to Ben Dooks

If of_i3c_master_add_i3c_boardinfo() fails, then there's no much to say what the issue was (most of the error returns are -EINVAL), so add some printing of the errors using dev_err() showing the node that failed and what the issue was.

This should help with finding which device-tree node was causing the issue and also mirrors the i2c case where it shows the node and the error.

---
 drivers/i3c/master.c | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)
drivers/i3c/master.c +15 -5
index bc42669f5c6d..2a9ebb1d9d57 100644
@@ -2050,32 +2050,42 @@ of_i3c_master_add_i3c_boardinfo(struct i3c_master_controller *master,
 		return -ENOMEM;
 
 	if (reg[0]) {
-		if (reg[0] > I3C_MAX_ADDR)
+		if (reg[0] > I3C_MAX_ADDR) {
+			dev_err(dev, "%pOF: address too big\n", node);
 			return -EINVAL;
+		}
 
 		addrstatus = i3c_bus_get_addr_slot_status(&master->bus,
 							  reg[0]);
-		if (addrstatus != I3C_ADDR_SLOT_FREE)
+		if (addrstatus != I3C_ADDR_SLOT_FREE) {
+			dev_err(dev, "%pOF: slot in use\n", node);
 			return -EINVAL;
+		}
 	}
 
 	boardinfo->static_addr = reg[0];
 
 	if (!of_property_read_u32(node, "assigned-address", &init_dyn_addr)) {
-		if (init_dyn_addr > I3C_MAX_ADDR)
+		if (init_dyn_addr > I3C_MAX_ADDR) {
+			dev_err(dev, "%pOF: cannot assign address\n", node);
 			return -EINVAL;
+		}
 
 		addrstatus = i3c_bus_get_addr_slot_status(&master->bus,
 							  init_dyn_addr);
-		if (addrstatus != I3C_ADDR_SLOT_FREE)
+		if (addrstatus != I3C_ADDR_SLOT_FREE) {
+			dev_err(dev, "%pOF: slot in use\n", node);
 			return -EINVAL;
+		}
 	}
 
 	boardinfo->pid = ((u64)reg[1] << 32) | reg[2];
 
 	if ((boardinfo->pid & GENMASK_ULL(63, 48)) ||
-	    I3C_PID_RND_LOWER_32BITS(boardinfo->pid))
+	    I3C_PID_RND_LOWER_32BITS(boardinfo->pid)) {
+		dev_err(dev, "%pOF: bad PID\n", node);
 		return -EINVAL;
+	}
 
 	boardinfo->init_dyn_addr = init_dyn_addr;
 	boardinfo->of_node = of_node_get(node);
Signature · 1 line
2.40.1
author
#5 lore raw

RFC 5/5 i3c: dw; add print if cannot get resources

In reply to Ben Dooks

The devm_reset_control_get_optional_exclusive() call does not print any errors, neiterh does the clk_prepare_enable or devm_request_irq() call.

Add some basic error printing to make the probe failures easier to debug.

---
 drivers/i3c/master/dw-i3c-master.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/i3c/master/dw-i3c-master.c +9 -3
index 9332ae5f6419..ffc84ff6225c 100644
@@ -1429,12 +1429,16 @@ int dw_i3c_common_probe(struct dw_i3c_master *master,
 
 	master->core_rst = devm_reset_control_get_optional_exclusive(&pdev->dev,
 								    "core_rst");
-	if (IS_ERR(master->core_rst))
+	if (IS_ERR(master->core_rst)) {
+		dev_err(&pdev->dev, "cannot get core_rst\n");
 		return PTR_ERR(master->core_rst);
+	}
 
 	ret = clk_prepare_enable(master->core_clk);
-	if (ret)
+	if (ret) {
+		dev_err(&pdev->dev, "cannot enable core_clk\n");
 		goto err_disable_core_clk;
+	}
 
 	reset_control_deassert(master->core_rst);
 
@@ -1446,8 +1450,10 @@ int dw_i3c_common_probe(struct dw_i3c_master *master,
 	ret = devm_request_irq(&pdev->dev, irq,
 			       dw_i3c_master_irq_handler, 0,
 			       dev_name(&pdev->dev), master);
-	if (ret)
+	if (ret) {
+		dev_err(&pdev->dev, "cannot get irq\n");
 		goto err_assert_rst;
+	}
 
 	platform_set_drvdata(pdev, master);
 
Signature · 1 line
2.40.1
#6 lore raw

Re: [RFC 5/5] i3c: dw; add print if cannot get resources

Hello,

Please fix the typo in the subject.

Quoted 2 lines · Ben Dooks

The devm_reset_control_get_optional_exclusive() call does not print any errors, neiterh does the clk_prepare_enable

Also fix this one.

Quoted 4 lines

or devm_request_irq() call.

Add some basic error printing to make the probe failures easier to debug.

I guess all those dev_err could be dev_dbg so we don't litter the driver with strings that will only ever be useful during bring up.

Quoted 44 lines
---
 drivers/i3c/master/dw-i3c-master.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/i3c/master/dw-i3c-master.c +9 -3
index 9332ae5f6419..ffc84ff6225c 100644
@@ -1429,12 +1429,16 @@ int dw_i3c_common_probe(struct dw_i3c_master *master,
 
 	master->core_rst = devm_reset_control_get_optional_exclusive(&pdev->dev,
 								    "core_rst");
-	if (IS_ERR(master->core_rst))
+	if (IS_ERR(master->core_rst)) {
+		dev_err(&pdev->dev, "cannot get core_rst\n");
 		return PTR_ERR(master->core_rst);
+	}
 
 	ret = clk_prepare_enable(master->core_clk);
-	if (ret)
+	if (ret) {
+		dev_err(&pdev->dev, "cannot enable core_clk\n");
 		goto err_disable_core_clk;
+	}
 
 	reset_control_deassert(master->core_rst);
 
@@ -1446,8 +1450,10 @@ int dw_i3c_common_probe(struct dw_i3c_master *master,
 	ret = devm_request_irq(&pdev->dev, irq,
 			       dw_i3c_master_irq_handler, 0,
 			       dev_name(&pdev->dev), master);
-	if (ret)
+	if (ret) {
+		dev_err(&pdev->dev, "cannot get irq\n");
 		goto err_assert_rst;
+	}
 
 	platform_set_drvdata(pdev, master);
 
Signature · 1 line
2.40.1
Signature · 3 lines
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
author
#7 lore raw

Re: [RFC 5/5] i3c: dw; add print if cannot get resources

Quoted 9 lines · Alexandre Belloni

Hello,

Please fix the typo in the subject.

Quoted 2 lines · Ben Dooks

The devm_reset_control_get_optional_exclusive() call does not print any errors, neiterh does the clk_prepare_enable

Also fix this one.

Quoted 7 lines
Quoted 4 lines

or devm_request_irq() call.

Add some basic error printing to make the probe failures easier to debug.

I guess all those dev_err could be dev_dbg so we don't litter the driver with strings that will only ever be useful during bring up.

I think dev_err_probe() is probably the right one as we aren't going to want to warn on any sort of probe defer errors. However most drivers are printing errors if there are resources missing so I think an error is the best way to report issues.

I'll correct the spelling when the discussions about the correct way to print errors is.

Quoted 44 lines
Quoted 43 lines
---
  drivers/i3c/master/dw-i3c-master.c | 12 +++++++++---
  1 file changed, 9 insertions(+), 3 deletions(-)
drivers/i3c/master/dw-i3c-master.c +9 -3
index 9332ae5f6419..ffc84ff6225c 100644
@@ -1429,12 +1429,16 @@ int dw_i3c_common_probe(struct dw_i3c_master *master,
 
  	master->core_rst = devm_reset_control_get_optional_exclusive(&pdev->dev,
  								    "core_rst");
-	if (IS_ERR(master->core_rst))
+	if (IS_ERR(master->core_rst)) {
+		dev_err(&pdev->dev, "cannot get core_rst\n");
  		return PTR_ERR(master->core_rst);
+	}
 
  	ret = clk_prepare_enable(master->core_clk);
-	if (ret)
+	if (ret) {
+		dev_err(&pdev->dev, "cannot enable core_clk\n");
  		goto err_disable_core_clk;
+	}
 
  	reset_control_deassert(master->core_rst);
 
@@ -1446,8 +1450,10 @@ int dw_i3c_common_probe(struct dw_i3c_master *master,
  	ret = devm_request_irq(&pdev->dev, irq,
  			       dw_i3c_master_irq_handler, 0,
  			       dev_name(&pdev->dev), master);
-	if (ret)
+	if (ret) {
+		dev_err(&pdev->dev, "cannot get irq\n");
  		goto err_assert_rst;
+	}
 
  	platform_set_drvdata(pdev, master);
 
Signature · 1 line
2.40.1
Signature · 10 lines
Ben Dooks				http://www.codethink.co.uk/
Senior Engineer				Codethink - Providing Genius

https://www.codethink.co.uk/privacy.html


-- 
linux-i3c mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-i3c
author
#8 lore raw
In reply to Ben Dooks
Quoted 12 lines · Ben Dooks

During work on an i3c compatible system some of the probing does not produce much useful error output. This series is an attempt to add error reporting where it would be useful to track down the culprit.

Ben Dooks (5):
i3c: show error with node for invalid reg property
i3c: add error print to show device failing during populate bus
i3c: show node when printing unsupported 10-bit i2c dev
i3c: show error messages in of_i3c_master_add_i3c_boardinfo
i3c: dw; add print if cannot get resources

Quoted 3 lines
  drivers/i3c/master.c               | 28 +++++++++++++++++++++-------
  drivers/i3c/master/dw-i3c-master.c | 12 +++++++++---
  2 files changed, 30 insertions(+), 10 deletions(-)

Has anyone had a chance to review these?

Signature · 4 lines
Ben Dooks				http://www.codethink.co.uk/
Senior Engineer				Codethink - Providing Genius

https://www.codethink.co.uk/privacy.html