2022-02-02 15:43:20

by Joe Perches

[permalink] [raw]
Subject: Re: [PATCH RESEND v2] scripts/checkpatch.pl: remove _deferred and _deferred_once false warning

On Wed, 2022-02-02 at 15:22 +0530, Maninder Singh wrote:
> Hi,
>
>
> > > ?printk_deferred?and?printk_deferred_once?requires?LOGLEVEL?in?argument,
> > > ?but?checkpatch.pl?reports?it?as?warning:
> ?
> > When?did?that?occur???Please?reference?the?specific?commit.
> > When?printk_sched?(now?printk_deferred)?was?created?it?did?not
> > allow?KERN_<LEVEL>.
> ?
> > see?commit?3ccf3e830615?("printk/sched:?Introduce?special?printk_sched()?for?those?awkward?moments")
> ?
> I think with below commit:
>
> 98e35f5894cf208084688ec0c7bb7b713efc997f (printk: git rid of [sched_delayed] message for printk_deferred)
>
> earlier it was passing hardcoded KERN_WARNING to all printk_deferred messages, but now it switched
> to normal printk behavior.
>
> - if (in_sched)
> - text_len = scnprintf(text, sizeof(textbuf),
> - KERN_WARNING "[sched_delayed] ");
> -
> - text_len += vscnprintf(text + text_len,
> - sizeof(textbuf) - text_len, fmt, args);
> + text_len = vscnprintf(text, sizeof(textbuf), fmt, args);
>
>
> I did not search that earlier, because we were fixing some issue in our local module code with printk_deferred
> and then checkatch reported the issues, so thought of fixing it, becaue without level
> pritk_deferred messages was shifted to default loglevel as done by normal printk.
>
> So I thought it was designed to pass loglevel from starting stage and checjpatch reports falsely.
> but seems checkpatch needs to fixed with that commit itself, because earlier it was not required to pass loglevel.

Most of that should be in the commit message.



2022-02-03 17:21:05

by Maninder Singh

[permalink] [raw]
Subject: RE: [PATCH RESEND v2] scripts/checkpatch.pl: remove _deferred and _deferred_once false warning

Hi,

> > > > printk_deferred and printk_deferred_once requires LOGLEVEL in argument,
> > > > but checkpatch.pl reports it as warning:
> >
> > > When did that occur? Please reference the specific commit.
> > > When printk_sched (now printk_deferred) was created it did not
> > > allow KERN_<LEVEL>.
> >
> > > see commit 3ccf3e830615 ("printk/sched: Introduce special printk_sched() for those awkward moments")
> >
> > I think with below commit:
> >
> > 98e35f5894cf208084688ec0c7bb7b713efc997f (printk: git rid of [sched_delayed] message for printk_deferred)
> >
> > earlier it was passing hardcoded KERN_WARNING to all printk_deferred messages, but now it switched
> > to normal printk behavior.
> >
> > - if (in_sched)
> > - text_len = scnprintf(text, sizeof(textbuf),
> > - KERN_WARNING "[sched_delayed] ");
> > -
> > - text_len += vscnprintf(text + text_len,
> > - sizeof(textbuf) - text_len, fmt, args);
> > + text_len = vscnprintf(text, sizeof(textbuf), fmt, args);
> >
> >
> > I did not search that earlier, because we were fixing some issue in our local module code with printk_deferred
> > and then checkatch reported the issues, so thought of fixing it, becaue without level
> > pritk_deferred messages was shifted to default loglevel as done by normal printk.
> >
> > So I thought it was designed to pass loglevel from starting stage and checjpatch reports falsely.
> > but seems checkpatch needs to fixed with that commit itself, because earlier it was not required to pass loglevel.
>
> Most of that should be in the commit message.

Yes, Initially I thought it's added in checkpatch by mistake,
I Should have checked history of printk_deferred also.

I will send new version with updated comit message.

Thanks


Attachments:
rcptInfo.txt (1.28 kB)