Skip to contentSkip to conversation

i2c: core: Improve i2c_new_scanned_device

PATCH In review Heiner Kallweit 2 participants 4 messages Jan 7, 2025 – Jan 14, 2025

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

PATCH 1/1 i2c: core: Improve i2c_new_scanned_device

Simplify the logic in this function. And as we're no longer limited to C89, move the iterator variable declaration to the for loop.

---
 drivers/i2c/i2c-core-base.c | 18 +++++++-----------
 1 file changed, 7 insertions(+), 11 deletions(-)
drivers/i2c/i2c-core-base.c +7 -11
index 4a858b1ae..f3b1106f2 100644
@@ -2507,12 +2507,10 @@ i2c_new_scanned_device(struct i2c_adapter *adap,
 		       unsigned short const *addr_list,
 		       int (*probe)(struct i2c_adapter *adap, unsigned short addr))
 {
-	int i;
-
 	if (!probe)
 		probe = i2c_default_probe;
 
-	for (i = 0; addr_list[i] != I2C_CLIENT_END; i++) {
+	for (int i = 0; addr_list[i] != I2C_CLIENT_END; i++) {
 		/* Check address validity */
 		if (i2c_check_7bit_addr_validity_strict(addr_list[i]) < 0) {
 			dev_warn(&adap->dev, "Invalid 7-bit address 0x%02x\n",
@@ -2529,17 +2527,15 @@ i2c_new_scanned_device(struct i2c_adapter *adap,
 		}
 
 		/* Test address responsiveness */
-		if (probe(adap, addr_list[i]))
-			break;
+		if ((probe(adap, addr_list[i]))) {
+			info->addr = addr_list[i];
+			return i2c_new_client_device(adap, info);
+		}
 	}
 
-	if (addr_list[i] == I2C_CLIENT_END) {
-		dev_dbg(&adap->dev, "Probing failed, no device found\n");
-		return ERR_PTR(-ENODEV);
-	}
+	dev_dbg(&adap->dev, "Probing failed, no device found\n");
 
-	info->addr = addr_list[i];
-	return i2c_new_client_device(adap, info);
+	return ERR_PTR(-ENODEV);
 }
 EXPORT_SYMBOL_GPL(i2c_new_scanned_device);
 
Signature · 1 line
2.47.1
#1 lore raw
Quoted 4 lines · Heiner Kallweit

Simplify the logic in this function. And as we're no longer limited to C89, move the iterator variable declaration to the for loop.

Did you have HW to test this patch?

#2 lore raw
Quoted 8 lines · Wolfram Sang
Quoted 4 lines · Heiner Kallweit

Simplify the logic in this function. And as we're no longer limited to C89, move the iterator variable declaration to the for loop.

Did you have HW to test this patch?

No functional change is intended. I have hw using i801, which calls i2c_register_spd() -> i2c_new_scanned_device(). With this patch tools like decode-dimms still work.

#3 lore raw
Quoted 4 lines

No functional change is intended. I have hw using i801, which calls i2c_register_spd() -> i2c_new_scanned_device(). With this patch tools like decode-dimms still work.

Thanks. Still, thinking more about it, I am not comfortable with this patch. Kernel code where the intended exit of a function is not at the end but somewhere in the middle often enough confused me in the past.

So, although it is not as concise as your suggestion, I would prefer to leave the code as-is for clarity.