Re: [PATCH v2 3/3] power: process: use explicit levels for printk continuations

[Petr Mladek]

On Fri 2022-11-25 21:41:55, Thomas Weißschuh wrote:
> On 2022-11-25 11:53-0800, Joe Perches wrote:
> > On Fri, 2022-11-25 at 20:09 +0100, Thomas Weißschuh wrote:
> >> Many of the printk messages emitted during suspend and resume are
> >> emitted in fragments using pr_cont()/KERN_CONT.
> >> 
> >> As during suspend and resume a lot of operations are happing in the
> >> kernel the chances are high that the fragments are interspersed with
> >> unrelated messages.
> >> 
> >> In this case if no explicit level is specified for the fragments the
> >> standard level is applied, which by default is KERN_WARNING.
> >> 
> >> If the user is only observing KERN_WARNING and *not* KERN_INFO messages
> >> they will see incomplete message fragments.
> >> 
> >> By specifing the correct printk level also with the continuations this
> >> mismatch can be avoided.
> >> Also it reduces the amount of false-positive KERN_WARNING messages.
> >> 
> >> Signed-off-by: Thomas Weißschuh <linux@xxxxxxxxxxxxxx>
> >> ---
> >>  kernel/power/process.c | 18 +++++++++---------
> >>  1 file changed, 9 insertions(+), 9 deletions(-)
> >> 
> >> diff --git a/kernel/power/process.c b/kernel/power/process.c
> > []
> >> @@ -82,7 +82,7 @@ static int try_to_freeze_tasks(bool user_only)
> >>  	elapsed_msecs = ktime_to_ms(elapsed);
> >>  
> >>  	if (todo) {
> >> -		pr_cont("\n");
> >> +		pr_info_cont("\n");
> > 
> > I think this isn't needed because of the immediately following pr_err.

Great catch.


> The pr_cont() itself or the conversion to pr_info_cont() is not needed?

The pr_cont() was needed before the commit 4bcc595ccd80decb42450
("printk: reinstate KERN_CONT for printing continuation lines").

Before this commit, lines were appended even without KERN_CONT
when the previous line did not end with "\n".

The above commit caused that only lines with KERN_CONT will be
appended.

We have the code:

		pr_cont("\n");
		pr_err("Freezing of tasks %s after %d.%03d seconds "

The pr_cont() is not needed here because pr_err() does not have
KERN_CONT. It will always start on a new line.

> Personally I would prefer to keep the patch as is.
> 
> If only the conversion is not needed for consistency with the rest of the file.
> If the pr_cont() in general is not needed it should be changed in a dedicated
> patch (by somebody who knows this code better).

I agree that it should be removed in a separate patch. The commit
message should mention the commit that modified the KERN_CONT
handling.

Best Regards,
Petr