lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <YUDIOLTCe0/kzVSW@bismarck.dyn.berto.se>
Date:   Tue, 14 Sep 2021 18:05:12 +0200
From:   Niklas Söderlund 
        <niklas.soderlund@...igine.com>
To:     Joe Perches <joe@...ches.com>
Cc:     Andy Whitcroft <apw@...onical.com>, linux-kernel@...r.kernel.org
Subject: False positive for 'Possible unnecessary KERN_ERR' warning in
 checkpatch

Hi Joe,

Maybe you are already aware of this, but in case you are not.

The issue is the checkpatch check for unnecessary KERN_<LEVEL> for log 
functions. If a single line contains a statement that match a log 
function name from $logFunctions, such as foo_err() and the same line 
contains a KERN_<LEVEL> statement the 'WARNING: Possible unnecessary 
KERN_ERR' is triggered. This is true even if the KERN_<LEVEL> statement 
is not part of the arguments to the foo_err() definition that triggers 
the first part of the check.

This can be demonstrated by,

    ./scripts/checkpatch.pl --mailback --git c821e617896e99b8

Where we get (among others) the warning,

    WARNING: Possible unnecessary KERN_ERR
    #38: FILE: drivers/net/ethernet/netronome/nfp/nfp_net.h:63:
    +#define nn_err(nn, fmt, args...)	nn_pr(nn, KERN_ERR, fmt, ## args)

Looking at the code in checkpatch.pl we have,

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|)|
        TP_printk|
        WARN(?:_RATELIMIT|_ONCE|)|
        panic|
        MODULE_[A-Z_]+|
        seq_vprintf|seq_printf|seq_puts
)};

...

# check for logging functions with KERN_<LEVEL>
                if ($line !~ /printk(?:_ratelimited|_once)?\s*\(/ &&
                    $line =~ /\b$logFunctions\s*\(.*\b(KERN_[A-Z]+)\b/) {
                        my $level = $1;
                        if (WARN("UNNECESSARY_KERN_LEVEL",
                                 "Possible unnecessary $level\n" . $herecurr) &&
                            $fix) {
                                $fixed[$fixlinenr] =~ s/\s*$level\s*//;
                        }
                }

Looking at the line from above that triggers the warning,

	#define nn_err(nn, fmt, args...)   nn_pr(nn, KERN_ERR, fmt, ## args)

We see that the warning is triggers by the regexp but that it matches 
the first part on nn_err( and then the second part of on the second 
argument to nn_pr, KERN_ERR. I believe this to be a false positive.

Unfortunately my Perl skills are not good enough to fix the check to only 
look for KERN_[A-Z]+ inside the argument list to the log function name 
that matches the first part of the regexp.

-- 
Regards,
Niklas Söderlund

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ