linux-kernel - Re: [PATCH 2/2] Enable 'make CONFIG

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]

Date:	Sun, 31 Jul 2011 01:21:27 -0400
From:	Arnaud Lacombe <lacombar@...il.com>
To:	"H. Peter Anvin" <hpa@...or.com>
Cc:	David Woodhouse <dwmw2@...radead.org>,
	linux-kernel@...r.kernel.org, linux-kbuild@...r.kernel.org,
	Roman Zippel <zippel@...ux-m68k.org>
Subject: Re: [PATCH 2/2] Enable 'make CONFIG_FOO=y oldconfig'

Hi,

On Sat, Jul 30, 2011 at 10:09 PM, Arnaud Lacombe <lacombar@...il.com> wrote:
> Hi,
>
> On Sat, Jul 30, 2011 at 9:28 PM, H. Peter Anvin <hpa@...or.com> wrote:
>> On 07/30/2011 06:06 PM, Arnaud Lacombe wrote:
>>>>
>>> The principle of least surprise is broken anyway as the proposed patch
>>> has absolutely no dependency checking and verification. You can `make
>>> CONFIG_SATA_MV=y allnoconfig', you will _not_ get it set.
>>>
>> This gives you exactly the same way as a .config file or a ALLCONFIG
>> file with the same options, which is what one realistically should expect.
>>
> no, I would expect it to have the same effect as a `select
> CONFIG_SATA_MV', have it forcibly set to the given value, and be
> warned if there was any problem.
>
> Btw, `make CONFIG_GENERIC_BUG=n oldconfig' or `make CONFIG_64BIT=n
> oldconfig'[0] does not even work, and I'm getting no error. So either
> you make it work for all possible uses, or you warn the user he tried
> something illegal, but you don't just fail silently.
>
ok, the issue is that you will only be allowed to change visible
symbols. CONFIG_64BIT is conditionally visible (when ARCH=x86), so
right now, you can not do on x86-64:

% make ARCH=i386 defconfig
% make CONFIG_64BIT=n oldconfig # [0]

and expect it to work.

Instead of allowing explicitly a symbol to be overridden, the behavior
is to implicitly, silently and un-deterministicaly[1] deny an
override. Strange ...

 - Arnaud

[0]: this match David's use-case described in
<1311986969.20983.52.camel@...infradead.org>

[1]: ie. it depends on the state of the tree when ran
--
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/