Skip to contentSkip to conversation

regmap: debugfs: Fix compilation warning

PATCH Mark Brown 2 participants 5 messages Jan 22, 2013 – Jan 26, 2013 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Mark Brown

author
#1 lore raw

Review

Quoted 4 lines · Vincent Stehlé

This fixes the following compilation warning:

(quoted diff) +1 -1
-	unsigned int i, ret;
+	unsigned int i, ret = 0;

This sort of fix is not a good idea, you're just shutting the warning up without any sort of analysis explaining why it's generated in error. If it's generating a spurious error that's a compiler bug.

#2 lore raw

On 01/22/2013 07:41 AM, Mark Brown wrote:
(..)

Quoted 4 lines

This sort of fix is not a good idea, you're just shutting the warning up without any sort of analysis explaining why it's generated in error. If it's generating a spurious error that's a compiler bug.

In the regmap_debugfs_get_dump_start() function control flow, I think the compiler cannot know for sure that the list_for_each_entry() iterates at least once.:

	regmap_debugfs_get_dump_start()
	{
		unsigned int ret;
		<snip some code, which does not touch ret>

		list_for_each_entry() {
			<snip some code>
			ret = <value>;
		}

		return ret;
	}

Do you think there is a way to "mark" the list_for_each_entry() as iterating at least once? an __attribute__ maybe?

Best regards,

V.

author
#3 lore raw
Quoted 2 lines · Vincent Stehlé

Do you think there is a way to "mark" the list_for_each_entry() as iterating at least once? an __attribute__ maybe?

No - but are you sure that's true?

#4 lore raw
Quoted 6 lines · Mark Brown
Quoted 2 lines · Vincent Stehlé

Do you think there is a way to "mark" the list_for_each_entry() as iterating at least once? an __attribute__ maybe?

No - but are you sure that's true?

If you mean "am I sure the loop iterates at least once", then yes, as we have an explicit check just before the concerned list_for_each_entry():

	/*
	 * This should never happen; we return above if we fail to
	 * allocate and we should never be in this code if there are
	 * no registers at all.
	 */
	if (list_empty(&map->debugfs_off_cache)) {
		WARN_ON(list_empty(&map->debugfs_off_cache));
		return base;
	}

Best regards,

V.

author
#5 lore raw
Quoted 2 lines · Vincent Stehlé

If you mean "am I sure the loop iterates at least once", then yes, as we have an explicit check just before the concerned list_for_each_entry():

So, I hadn't realised this was triggered by a new compiler version - Arnd just tried the same bodge. I've committed a workaround for the compiler which avoids the problems that setting 0 will have and allows the flow analyis to run where it can.