linux-kernel - Re: [PATCH 4/5] efi: Export Runtime Configuration Interface table to sysfs

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20191002102159.GA2109@localhost.localdomain>
Date:   Wed, 2 Oct 2019 10:22:10 +0000
From:   <Narendra.K@...l.com>
To:     <geert@...ux-m68k.org>
CC:     <Mario.Limonciello@...l.com>, <ard.biesheuvel@...aro.org>,
        <linux-efi@...r.kernel.org>, <mingo@...nel.org>,
        <tglx@...utronix.de>, <linux-kernel@...r.kernel.org>,
        <james.morse@....com>, <tanxiaofei@...wei.com>,
        <Narendra.K@...l.com>
Subject: Re: [PATCH 4/5] efi: Export Runtime Configuration Interface table to
 sysfs

On Tue, Oct 01, 2019 at 08:23:51PM +0200, Geert Uytterhoeven wrote:
[...]
> > > > > > > > +config EFI_RCI2_TABLE
> > > > > > > > +       bool "EFI Runtime Configuration Interface Table Version 2 Support"
[...]
> > > Is it really a problem to just say n?
> > >
> > > I think this seems like a needless change that would slow down adoption of
> > > !x86 if Dell EMC PowerEdge systems did start going that route, especially
> > > when it comes to distributions that move glacially slow with picking up new
> > > kernel code.
> >
> > Hi Ard/Geert,
> >
> > Any additional thoughts here ?
> 
> Sure ;-)
> 
> A typical platform-specific sarm/arm64 .config file has almost 3000
> config options
> disabled.  Hence that means I have to say "n" almost 3000 times.
> Fortunately I started doing this several years ago, so I can do this
> incrementally ;-)
> 
> Perhaps someone should try to remove all lines like "depends on ... ||
> COMPILE_TEST", run "make oldconfig", read all help texts before saying "n",
> and time the whole operation...
> 
> I hope I managed to convince you of the benefits.

Thank you Geert. The description is helpful. I am working on it. 
As I understand, the issue is 'make oldconfig' provides a prompt to the user 
and user is expecting that a prompt is not needed as the option is not
relevant.

I cloned upstream kernel 5.3.2 as it does not have EFI_RCI2_TABLE option
and generated a .config by calling 'make defconfig'. The .config has
COMPILE_TEST set to n. I copied it to 5.4-rc1 and added 'depends on COMPILE_TEST' 
to drivers/firmware/efi/Kconfig (did not add CONFIG_X86 because it is
set to y by the defconfig from 5.3.2). 'make oldconfig' still provides a
prompt for CONFIG_EFI_RCI2_TABLE. 

I removed 'depends on COMPILE_TEST' from Kconfig and modified it to
include the below change -

config EFI_RCI2_TABLE
	bool 
	prompt "EFI Runtime Configuration Interface Table Version 2 Support" if COMPILE_TEST
	default n
	help

Adding the condition to the 'prompt' section seems to have desired
result. With this change, 'make oldconfig' did not provide a prompt. 

It seems like 'make oldconfig' will provide a prompt to the user if the
CONFIG option is new and providing the prompt does not depend on the
'depends on' section. It seems to be dependent on the 'prompt' section.

Any thoughts ? If the above understanding is correct, I will work to
submit a patch with 'prompt' section modified to contain

prompt "EFI Runtime Configuration Interface Table Version 2 Support" if X86 || COMPILE_TEST

-- 
With regards,
Narendra K