On Tuesday, February 04, 2014 04:24:13 PM Sebastian Capella wrote:
Quoting Rafael J. Wysocki (2014-02-04 16:28:13)
On Tuesday, February 04, 2014 04:06:42 PM Sebastian Capella wrote:
Quoting Rafael J. Wysocki (2014-02-04 16:03:29)
On Tuesday, February 04, 2014 03:22:22 PM Sebastian Capella wrote:
Quoting Sebastian Capella (2014-02-04 14:37:33)
Quoting Rafael J. Wysocki (2014-02-04 13:36:29) > > static int __init resumedelay_setup(char *str) > > { > > - resume_delay = simple_strtoul(str, NULL, 0); > > + int ret = kstrtoint(str, 0, &resume_delay); > > + /* mask must_check warn; on failure, leaves resume_delay unchanged */ > > + (void)ret;
One unintended consequence of this change is that it'll now accept a negative integer parameter.
Well, what about using kstrtouint(), then?
I was thinking of doing something like:
int delay, res; res = kstrtoint(str, 0, &delay); if (!res && delay >= 0) resume_delay = delay; return 1;It uses simple_strtoul() for a reason. You can change the type of resume_delay to match, but the basic question is:
Why exactly do you want to change that thing?
This entire patch is a result of a single checkpatch warning from a printk that I indented.
I was hoping to be helpful by removing all of the warnings from this file, since I was going to have a separate cleanup patch for the printk.
I can see this is not a good direction.
Would it be better also to leave the file's printks as they were and drop the cleanup patch completely?
Well, I had considered changing them to pr_something, but decided that it wasn't worth the effort. Quite frankly, I'd leave the code as is. :-)
Thanks!