Re: [PATCH v1 1/1] checkpatch: Add dev_err_probe() to the list of Log Functions

From: Guenter Roeck
Date: Fri Dec 01 2023 - 11:01:38 EST


On 12/1/23 07:14, Andy Shevchenko wrote:
dev_err_probe() is missing in the list of Log Functions and hence
checkpatch issues a warning in the cases when any other function
in use won't trigger it. Add dev_err_probe() to the list to behave
consistently.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
---
scripts/checkpatch.pl | 1 +
1 file changed, 1 insertion(+)

diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
index a94ed6c46a6d..c40f3f784f7e 100755
--- a/scripts/checkpatch.pl
+++ b/scripts/checkpatch.pl
@@ -593,6 +593,7 @@ our $zero_initializer = qr{(?:(?:0[xX])?0+$Int_type?|NULL|false)\b};
our $logFunctions = qr{(?x:
printk(?:_ratelimited|_once|_deferred_once|_deferred|)|
(?:[a-z0-9]+_){1,2}(?:printk|emerg|alert|crit|err|warning|warn|notice|info|debug|dbg|vdbg|devel|cont|WARN)(?:_ratelimited|_once|)|
+ dev_err_probe|
TP_printk|
WARN(?:_RATELIMIT|_ONCE|)|
panic|


Not sure if I agree. The difference here is that dev_err_probe()
has two additional parameters ahead of the string. I would very much prefer
to have those two additional parameters on a separate line if the string is
too long to fit in 100 columns with those two parameters on the same line.
In other words, I very much prefer

dev_err_probe(dev, -ESOMETHING,
"very long string");
over
dev_err_probe(dev, -ESOMETHING, "very long string");

and I don't really think that the latter has any benefits.

Also note that other dev_xxx() log functions are not included in the above test
and would still generate warnings. Accepting

dev_err_probe(dev, -ESOMETHING, "very long string");
but not
dev_err(dev, "very long string");

doesn't really make sense to me.

Guenter