Re: [uml-user] [PATCH] arch: configuration, deleting 'CONFIG_BUG' since always need it.

[Chen G. <gan...@as...>]

On 05/23/2013 05:12 PM, Geert Uytterhoeven wrote:
> On Thu, May 23, 2013 at 11:05 AM, Russell King - ARM Linux
> <li...@ar...> wrote:
>> > On Thu, May 23, 2013 at 10:40:29AM +0200, Geert Uytterhoeven wrote:
>>> >> On Thu, May 23, 2013 at 9:57 AM, Chen Gang <gan...@as...> wrote:
>>>> >> > -config BUG
>>>> >> > -       bool "BUG() support" if EXPERT
>>>> >> > -       default y
>>>> >> > -       help
>>>> >> > -          Disabling this option eliminates support for BUG and WARN, reducing
>>>> >> > -          the size of your kernel image and potentially quietly ignoring
>>>> >> > -          numerous fatal conditions. You should only consider disabling this
>>>> >> > -          option for embedded systems with no facilities for reporting errors.
>>>> >> > -          Just say Y.
>>> >>
>>> >> ... It's about reducing memory size on devices where you can't show bug or
>>> >> warning messages.
>> >
>> > And turning off CONFIG_BUG causes lots of warning messages at compile time
>> > about functions which are returning nothing which shouldn't.
>> >
>> > The problem is: trying to fix that _will_ mean the result is a larger
>> > kernel than if you just do the usual arch-implemented thing of placing
>> > an defined faulting instruction at the BUG() site - which defeats the
>> > purpose of turning off CONFIG_BUG.
> Is __builtin_unreachable() working well these days?

In fact, using __builtin_unreachable() is a standard way for
architectures to implemented their own BUG() (e.g. x86, s390, powerpc,
arm ...)

Before __builtin_unreachable(), must need an inline asm instruction
which architecture specific.

I have test using __builtin_unreachable() without an related asm
instruction before, it prints many unexpected things (please see the
attachment).

So I think, it is not suitable to use it in "asm-generic/bug.h"


Thanks.
-- 
Chen Gang

Asianux Corporation