linux-kernel - Re: [PATCH 1/4] Add a basic debugging framework for memory initialisation

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] [day] [month] [year] [list]

Message-ID: <20080422112151.GB30798@csn.ul.ie>
Date:	Tue, 22 Apr 2008 12:21:52 +0100
From:	Mel Gorman <mel@....ul.ie>
To:	Ingo Molnar <mingo@...e.hu>
Cc:	linux-mm@...ck.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/4] Add a basic debugging framework for memory initialisation

On (21/04/08 17:14), Ingo Molnar didst pronounce:
> 
> * Mel Gorman <mel@....ul.ie> wrote:
> 
> > +config DEBUG_MEMORY_INIT
> > +	bool "Debug memory initialisation"
> > +	depends on DEBUG_KERNEL
> > +	help
> > +	  Enable this to turn on debug checks during memory initialisation. By
> > +	  default, sanity checks will be made on the memory model and
> > +	  information provided by the architecture. What level of checking
> > +	  made and verbosity during boot can be set with the
> > +	  mminit_debug_level= command-line option.
> > +
> > +	  If unsure, say N
> 
> should be "default y" - and perhaps only disable-able on 
> CONFIG_EMBEDDED.

Ok, that would be something like the following?

       bool "Debug memory initialisation" if DEBUG_KERNEL && EMBEDDED
       depends on DEBUG_KERNEL
       default !EMBEDDED

This will slow up boot slightly on debug kernels as the additional checks
are made. It'll remain to be seen as to whether this is a problem for people
or not. I doubt it'll be noticed.

> We generally want such bugs to pop up as soon as 
> possible, and the sanity checks should only go away if someone 
> specifically aims for lowest system footprint.
> 

Seems fair and it's the second time this has been suggested (off-list reviewer
again). The only potential gotcha is if a sanity check is introduced that is
itself broken. It should be very obvious when this type of bug occurs though.

> the default loglevel for debug printouts might deserve another debug 
> option - but the core checks should always be included, and _errors_ 
> should always be printed out.
> 

I'll replace mminit_debug_level with mminit_loglevel that determines whether
information is printed at KERN_DEBUG or not. This matches what other similar
debug-frameworks are doing. I'll make sure errors always get printed.

Thanks

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/