Skip to contentSkip to conversation

mfd: wm831x: Fix typo "convertor" in comment

PATCH New Hemanth Selam 2 participants 2 messages Sep 8

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

PATCH v3 1/1 mfd: wm831x: Fix typo "convertor" in comment

Correct "convertor" to "converter", reported by scripts/checkpatch.pl using the misspelling list in scripts/spelling.txt. Only touches a comment, no code changes.

Since v1 the subject carries the driver name and a capitalised description, to match the MFD subsystem convention.

Assisted-by Cursor:claude-opus-5

---

Changes since v2:
- sent on its own instead of as part of a cirrus series. Mark Brown asked for patches touching different subsystems not to be grouped into one series when there is no dependency between them.

 drivers/mfd/wm831x-auxadc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/mfd/wm831x-auxadc.c +1 -1
index 744e1d161c5a..6280d4f38981 100644
@@ -132,7 +132,7 @@ static irqreturn_t wm831x_auxadc_irq(int irq, void *irq_data)
 			1 << input, 0);
 	wm831x->auxadc_active &= ~(1 << input);
 
-	/* Turn off the entire convertor if idle */
+	/* Turn off the entire converter if idle */
 	if (!wm831x->auxadc_active)
 		wm831x_reg_write(wm831x, WM831X_AUXADC_CONTROL, 0);
 
Signature · 1 line
2.48.1
#1 lore raw
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

Pre-existing issues:
- [High] This problem wasn't introduced by this patch, but memory corruption can occur due to freeing a list element without removing it from the list in wm831x_auxadc_read_irq.
- [High] This problem wasn't introduced by this patch, but a broken polling loop in wm831x_auxadc_read_polled causes premature timeout.
--

commit c4685699a90d5d341f26100b93279945ad18d7cc
Author: Hemanth Selam <[email protected]>

mfd: wm831x: Fix typo "convertor" in comment

Correct "convertor" to "converter", reported by scripts/checkpatch.pl using the misspelling list in scripts/spelling.txt. Only touches a comment, no code changes.

drivers/mfd/wm831x-auxadc.c static irqreturn_t wm831x_auxadc_irq(int irq, void *irq_data) · line 143

Quoted 12 lines
drivers/mfd/wm831x-auxadc.c +1 -1
index 6ee3018d8653f..94b8d24611e43 100644
@@ -138,7 +138,7 @@ static irqreturn_t wm831x_auxadc_irq(int irq, void *irq_data)
 			1 << input, 0);
 	wm831x->auxadc_active &= ~(1 << input);
 
-	/* Turn off the entire convertor if idle */
+	/* Turn off the entire converter if idle */
 	if (!wm831x->auxadc_active)
 		wm831x_reg_write(wm831x, WM831X_AUXADC_CONTROL, 0);

While reviewing this file, I noticed a couple of pre-existing issues in the surrounding code.

In drivers/mfd/wm831x-auxadc.c:wm831x_auxadc_read_irq():

Quoted 18 lines
drivers/mfd/wm831x-auxadc.c +0 -0
 	/* Enqueue the request */
	list_add(&req->list, &wm831x->auxadc_pending);

	ena = !wm831x->auxadc_active;

	if (ena) {
		ret = wm831x_set_bits(wm831x, WM831X_AUXADC_CONTROL,
				      WM831X_AUX_ENA, WM831X_AUX_ENA);
		if (ret != 0) {
			dev_err(wm831x->dev, "Failed to enable AUXADC: %d\n",
				ret);
			goto out;
		}
	}
[ ... ]
	list_del(&req->list);
	ret = req->val;
Quoted 6 lines

out:

	mutex_unlock(&wm831x->auxadc_lock);

	kfree(req);

	return ret;

[Severity: High]
This isn't a bug introduced by this patch, but memory corruption can occur due to freeing a list element without removing it from the list in wm831x_auxadc_read_irq.

If wm831x_set_bits() fails and jumps to the out label, the dynamically allocated req structure is freed via kfree(req) while it is still linked in the wm831x->auxadc_pending list.

Could this leave a dangling pointer in the pending list that might be dereferenced during subsequent calls to the interrupt handler wm831x_auxadc_irq() or other wm831x_auxadc_read_irq() calls?

Also in drivers/mfd/wm831x-auxadc.c:wm831x_auxadc_read_polled():

Quoted 19 lines
	/* If we're not using interrupts then poll the
	 * interrupt status register */
	timeout = 5;
	while (timeout) {
		msleep(1);

		ret = wm831x_reg_read(wm831x,
				      WM831X_INTERRUPT_STATUS_1);
		if (ret < 0) {
			dev_err(wm831x->dev,
				"ISR 1 read failed: %d\n", ret);
			goto disable;
		}

		/* Did it complete? */
		if (ret & WM831X_AUXADC_DATA_EINT) {
			wm831x_reg_write(wm831x,
					 WM831X_INTERRUPT_STATUS_1,
					 WM831X_AUXADC_DATA_EINT);
Quoted 8 lines
			break;
		} else {
			dev_err(wm831x->dev,
				"AUXADC conversion timeout\n");
			ret = -EBUSY;
			goto disable;
		}
	}

[Severity: High]
This isn't a bug introduced by this patch, but a broken polling loop in wm831x_auxadc_read_polled causes premature timeout.

When the driver falls back to polled mode, the else block unconditionally aborts the polling loop on the first failure after a 1ms delay. The timeout variable is also never decremented.

Does this structural issue prevent the while (timeout) loop from actually functioning as a polling loop, unconditionally returning -EBUSY if the ADC takes slightly longer than 1ms to finish?

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1