[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CANubcdWwHXqC4gUztNseoConYHp+G4nUSYFaO-Te3jFVqT=1sw@mail.gmail.com>
Date: Wed, 27 Apr 2022 09:57:33 +0800
From: Stephen Zhang <starzhangzsd@...il.com>
To: "Maciej W. Rozycki" <macro@...am.me.uk>
Cc: Thomas Bogendoerfer <tsbogend@...ha.franken.de>,
liam.howlett@...cle.com, ebiederm@...ssion.com, dbueso@...e.de,
alobakin@...me, f.fainelli@...il.com,
zhangshida <zhangshida@...inos.cn>, linux-kernel@...r.kernel.org,
linux-mips@...r.kernel.org
Subject: Re: [PATCH] arch/mips/kernel/traps: add CONFIG_MIPS_FP_SUPPORT when
using handle_fpe
Maciej W. Rozycki <macro@...am.me.uk> 于2022年4月27日周三 08:40写道:
>
> No need to conditionalise this either, because `cpu_has_fpu' is forced 0
> (in arch/mips/include/asm/cpu-features.h) if !CONFIG_MIPS_FP_SUPPORT. So
> this code translates to:
>
> if (0 && !0)
> set_except_vector(15, handle_fpe);
>
> in the preprocessor if CONFIG_MIPS_FP_SUPPORT is unset and is optimised
> away. Otherwise it should be written as:
>
> if (IS_ENABLED(CONFIG_MIPS_FP_SUPPORT) && ...
>
> so as not to clutter C code with #ifdef, as per our coding style.
>
> Maciej
Thanks for your comment. Do you mean the following code:
if (0 && !0)
set_except_vector(15, handle_fpe);
will be optimised away if !CONFIG_MIPS_FP_SUPPORT?
But we did get “undefined reference to `handle_fpe” error when compiled with
!CONFIG_MIPS_FP_SUPPORT.
Powered by blists - more mailing lists