Re: [PATCH 2/5] x86: ia32_signal: introduce COPY_SEG_STRICT

[Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>]

Américo Wang wrote:
> On Sat, Nov 15, 2008 at 3:26 AM, Hiroshi Shimamoto
> <h-shimamoto@ct.jp.nec.com> wrote:
>> From: Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
>>
>> Impact: cleanup
>>
>> Introduce COPY_SEG_STRICT for ia32_restore_sigcontext().
>>
>> Signed-off-by: Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
>> ---
>>  arch/x86/ia32/ia32_signal.c |   12 ++++++++----
>>  1 files changed, 8 insertions(+), 4 deletions(-)
>>
>> diff --git a/arch/x86/ia32/ia32_signal.c b/arch/x86/ia32/ia32_signal.c
>> index a28790a..f74178e 100644
>> --- a/arch/x86/ia32/ia32_signal.c
>> +++ b/arch/x86/ia32/ia32_signal.c
>> @@ -201,6 +201,12 @@ struct rt_sigframe
>>        err |= __get_user(regs->x, &sc->x);     \
>>  }
>>
>> +#define COPY_SEG_STRICT(seg)   {                       \
>> +               unsigned short tmp;                     \
>> +               err |= __get_user(tmp, &sc->seg);       \
>> +               regs->seg = tmp | 3;                    \
>> +}
>> +
> 
> Since your first patch is to kill the temporary variables, then why do you
> introduce a temporary variable here? It can be avoided like it was.
> 
> Can you explain the reason? :)

The first reason is to make this macro same as arch/x86/kernel/signal_{32|64}.c.
And second, I guess gcc may generate little better code.

thanks,
Hiroshi Shimamoto