Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Wed, 30 Nov 2022 15:51:08 +0100
From:   Petr Mladek <pmladek@suse.com>
To:     Joe Perches <joe@perches.com>
Cc:     Thomas =?iso-8859-1?Q?Wei=DFschuh?= <linux@weissschuh.net>,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        Pavel Machek <pavel@ucw.cz>, Len Brown <len.brown@intel.com>,
        linux-pm@vger.kernel.org,
        Sergey Senozhatsky <senozhatsky@chromium.org>,
        Andy Whitcroft <apw@canonical.com>,
        linux-kernel@vger.kernel.org, Steven Rostedt <rostedt@goodmis.org>,
        Dwaipayan Ray <dwaipayanray1@gmail.com>,
        Lukas Bulwahn <lukas.bulwahn@gmail.com>
Subject: Re: [PATCH v2 2/3] checkpatch: handle new pr_<level>_cont macros
Message-ID: <Y4dt3IMYKYnufHg5@alley>
References: <20221125190948.2062-1-linux@weissschuh.net>
 <20221125190948.2062-3-linux@weissschuh.net>
 <39f687f7c71b162eb8caefbec939fa2ad74fcdfc.camel@perches.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <39f687f7c71b162eb8caefbec939fa2ad74fcdfc.camel@perches.com>
Precedence: bulk

On Fri 2022-11-25 12:17:05, Joe Perches wrote:
> On Fri, 2022-11-25 at 20:09 +0100, Thomas Wei?schuh wrote:
> > These new macros from include/linux/printk.h replace the usage of plain
> > pr_cont().
> []
> > --- a/scripts/checkpatch.pl
> > +++ b/scripts/checkpatch.pl
> > @@ -590,7 +590,7 @@ our $zero_initializer = qr{(?:(?:0[xX])?0+$Int_type?|NULL|false)\b};
> > +# check for logging continuations without explicit level
> > +		if ($line =~ /\bpr_cont\s*\(/) {
> > +			WARN("LOGGING_CONTINUATION_WITHOUT_LEVEL",
> > +			     "Avoid logging continuation without level\n" . $herecurr);
> > +		}
> > +
> 
> Not so sure about this one.
> 
> I think relatively few situations are going to require interleaving avoidance.

Well, the problem is generic and any pr_cont() is affected except for
NMI context on single CPU system.

I though about a generic solution. We could store the last used printk
log level per-process and per-CPU context. But it does not solve
the situation when an unrelated printk() is printed by a nested
function.

It is exactly the case with try_to_freeze_tasks() in the 3rd patch.
Simplified code:

int freeze_processes(void)
{
	pr_info("Freezing user space processes ... ");
	try_to_freeze_tasks(true);
	pr_info_cont("done.");
}

, where

static int try_to_freeze_tasks(bool user_only)
{
[...]
	if (todo) {
		pr_err("Freezing of tasks %s after %d.%03d seconds "
[...]
}


I would personally add this check into checkpatch.pl. It might help
to make people aware about that pr_cont() is just the best effort.

Best Regards,
Petr