Hi,
2012/7/6 Albert ARIBAUD <albert.u.b...@aribaud.net>:
> Hi Nobuhiro,
>
> On Fri, 6 Jul 2012 09:07:30 +0900, Nobuhiro Iwamatsu
> <nobuhiro.iwamatsu...@renesas.com> wrote:
>> cmp_loop function was used from the structure of register.
>> And for the member of this structure, volatile is defined.
>> Since a volatile declaration of cmp_loop function is not made
>> by argument, made warning by compiler. This fixes this problem.
>>
>> Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu...@renesas.com>
>> ---
>> board/kmc/kzm9g/kzm9g.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/board/kmc/kzm9g/kzm9g.c b/board/kmc/kzm9g/kzm9g.c
>> index ad72c8d..8d7c61a 100644
>> --- a/board/kmc/kzm9g/kzm9g.c
>> +++ b/board/kmc/kzm9g/kzm9g.c
>> @@ -48,7 +48,7 @@ DECLARE_GLOBAL_DATA_PTR;
>> #define PORT34CR (0xE6051022)
>> #define PORT35CR (0xE6051023)
>>
>> -static int cmp_loop(u32 *addr, u32 data, u32 cmp)
>> +static int cmp_loop(volatile u32 *addr, u32 data, u32 cmp)
>> {
>> int err = -1;
>> int timeout = 100;
>
> I doubt cmp_loop needs this change. Volatile qualifiers do not make
> much sense in anything other than actual read or write instructions;
> if cmp_loop needs any volatile, it's as casts in its body, not as
> qualifiers in its signature. Fix the compiler warning by casting the
> actual argument in the *call* to cmp_loop to non-volatile instead.
>
> (incidentally, you should ask yourself why is the structure, or
> structure member, is qualified volatile in the first place; usually
> this is the sign that some code should instead use readl or writel
> and/or have isb() barriers at strategic places.)
>
Thank you for your comment.
You are right. I will fix structure member.
Best regards,
Nobuhiro
--
Nobuhiro Iwamatsu
iwamatsu at {nigauri.org / debian.org}
GPG ID: 40AD1FA6
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
