linux-kernel - Re: [PATCH 2/2] x86/e820: Use pr_debug to avoid spamming dmesg log with debug messages

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-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <e0116c3e-21a3-50a4-e9fd-cb227ef0b63b@akamai.com>
Date:   Fri, 14 May 2021 15:56:01 -0400
From:   Jason Baron <jbaron@...mai.com>
To:     Joe Perches <joe@...ches.com>,
        Heiner Kallweit <hkallweit1@...il.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>
Cc:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 2/2] x86/e820: Use pr_debug to avoid spamming dmesg log
 with debug messages



On 5/14/21 1:38 PM, Joe Perches wrote:
> On Tue, 2021-05-11 at 17:31 -0400, Jason Baron wrote:
> 
>> That said, I do see the value in not having to open code the branch stuff, and
>> making pr_debug() consistent with printk which does return a value. So that
>> makes sense to me.
> 
> IMO: printk should not return a value.
> 

Ok, the issue we are trying to resolve is to control whether a 'pr_debug()' statement
is enabled and thus use that to control subsequent output. The proposed patch does:


+	printed = pr_debug("e820: update [mem %#010Lx-%#010Lx] ", start, end - 1);
+	if (printed > 0) {
+		e820_print_type(old_type);
+		pr_cont(" ==> ");
+		e820_print_type(new_type);
+		pr_cont("\n");
+	}

I do think pr_debug() here is different from printk() b/c it can be explicitly
toggled.

I also suggested an alternative, which is possible with the current code which
is to use DYNAMIC_DEBUG_BRANCH().

if (DYNAMIC_DEBUG_BRANCH(e820_debg)) {
    printk(KERN_DEBUG "e820: update [mem %#010Lx-%#010Lx] ", start, end - 1);
    e820_print_type(old_type);
    pr_cont(" ==> ");
    e820_print_type(new_type);
    pr_cont("\n");
}

That however does require one to do something like this first:

DEFINE_DYNAMIC_DEBUG_METADATA(e820_dbg, "e820 verbose mode");

So I don't feel too strongly either way, but maybe others do?

Thanks,

-Jason