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: <CAMuHMdWALc_hneRaiwQbMWUXe=LnVqU7dkkWibV0cqb8Gc5e0g@mail.gmail.com>
Date:   Sat, 12 Oct 2019 19:04:15 +0200
From:   Geert Uytterhoeven <geert@...ux-m68k.org>
To:     Narendra K <Narendra.K@...l.com>
Cc:     Ard Biesheuvel <ard.biesheuvel@...aro.org>,
        linux-efi <linux-efi@...r.kernel.org>,
        Mario Limonciello <Mario.Limonciello@...l.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        James Morse <james.morse@....com>,
        Ingo Molnar <mingo@...nel.org>
Subject: Re: [PATCH] Ask user input only when CONFIG_X86 or
 CONFIG_COMPILE_TEST is set to y

Hi Narendra,

On Fri, Oct 11, 2019 at 2:55 PM <Narendra.K@...l.com> wrote:
> On Fri, Oct 11, 2019 at 12:01:25PM +0200, Geert Uytterhoeven wrote:
> > > > > > > -       bool "EFI Runtime Configuration Interface Table Version 2 Support"
> > > > > > > +       bool
> > > > > > > +       prompt "EFI RCI Table Version 2 Support" if X86 || COMPILE_TEST
> > > >
> > > > Why the split of bool and prompt?
> > > > Why not simply add a single line "depends on X86 || COMPILE_TEST"?
> > >
> > > It is because of the findings shared in [1]. Please let me know your
> > > thoughts on the findings.
> >
> > So you want to prevent the user from seeing a prompt for an option he may
> > or may not need to enable, when running "make oldconfig"?
>
> Geert,
>
> > The code in question is entirely architecture agnostic, and defaults
> > to 'n', so I am not convinced this is needed. (It came up in the
> > review as well)
>
> >> "make oldconfig" still asks me the question on e.g. arm64, where it is
> >> irrelevant, until arm64 variants of the hardware show up.
>
> >> So IMHO it should have "depends on X86 || COMPILE_TEST".
>
> From the discussion in [1] and [2](pasted a part of it above), my understanding
> of the issue you reported is that 'make oldconfig' asks the user a question for arm64
> though the EFI_RCI2_TABLE is not relevant for arm64. From the tests,
> it seemed like adding "depends on X86 || COMPILE_TEST" does not fix the
> issue, splitting bool into bool + prompt fixes it.
>
> Please let me know if I am missing any detail in the issue you reported.

Adding a "depends on X86 || COMPILE_TEST" should fix the issue, as
X86 is never set on arm64, nor on any other architecture than X86.
If COMPILE_TEST=y, it's normal expected behavior to show the question.

> With the way EFI_RCI2_TABLE is currently defined, my understanding is
> that 'make oldconfig' does not set the EFI_RCI2_TABLE to 'y' by default
> on arm64, but it asks the user the question. User has to say 'y' if he
> wants it to be set to 'y', else by default 'n' is set. This behavior is
> as expected.

If the option doesn't make sense on arm64 (more broadly: on non-X86),
it should depend on X86 || COMPILE_TEST, to avoid spamming the user
with (zillions of) options that do not matter for his platform.

> > One common approach is to let the Kconfig symbol for the platform (not for
> > all of X86!) select EFI_RCI2_TABLE.
> > That way it will be enabled automatically when needed.
>
> We did not intend to enable EFI_RCI2_TABLE option by default even on all
> X86 systems from the begining. As a result, we chose to set it to 'n' by
> default and added the guidance in 'help' section to say 'y' for Dell EMC
> PowerEdge systems.

Good.

>> > Another approach is to not force the option on, but guide the user towards
> > enabling it, by adding "default y if <platform_symbol>".
>
> As mentioned above, we want to keep the default to n.

OK.

> > Without the "|| COMPILE_TEST", you cannot enable compile-testing of
> > the driver on non-x86 platforms with EFI.
>
> Ok. We could keep the check. Could we make it independent of platforms
> by adding 'defbool y if COMPILE_TEST' ?

Please don't do that, as it with always enable the driver if COMPILE_TEST=y,
without providing a way to opt-out for the user.

Thanks!

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ