[PATCH] arm: add buildsys support for arm926ej-s

[Khem Raj]

On Tue, Jan 4, 2011 at 3:05 AM, Carmelo AMOROSO <carmelo.amoroso at st.com> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 1/4/2011 12:03 PM, Carmelo AMOROSO wrote:
>> ARM926EJ-S belongs to the ARM9 processor family.
>> It implements the ARMv5TEJ instruction set
>> (v5TE architecture with Jazelle technology)
>>
>> Signed-off-by: Carmelo Amoroso <carmelo.amoroso at st.com>
>> ---
>>  Rules.mak                |    1 +
>>  extra/Configs/Config.arm |    4 ++++
>>  2 files changed, 5 insertions(+), 0 deletions(-)
>>
>> diff --git a/Rules.mak b/Rules.mak
>> index eecdc64..e99ad26 100644
>> --- a/Rules.mak
>> +++ b/Rules.mak
>> @@ -341,6 +341,7 @@ ifeq ($(TARGET_ARCH),arm)
>>       CPU_CFLAGS-$(CONFIG_ARM922T)+=-mtune=arm9tdmi -march=armv4t
>>       CPU_CFLAGS-$(CONFIG_ARM926T)+=-mtune=arm9e -march=armv5te
>>       CPU_CFLAGS-$(CONFIG_ARM10T)+=-mtune=arm10tdmi -march=armv5t
>> +     CPU_CFLAGS-$(CONFIG_ARM926EJ_S) += -mtune=arm926ej-s -march=armv5te
>>       CPU_CFLAGS-$(CONFIG_ARM1136JF_S)+=-mtune=arm1136jf-s -march=armv6
>>       CPU_CFLAGS-$(CONFIG_ARM1176JZ_S)+=-mtune=arm1176jz-s -march=armv6
>>       CPU_CFLAGS-$(CONFIG_ARM1176JZF_S)+=-mtune=arm1176jzf-s -march=armv6
>> diff --git a/extra/Configs/Config.arm b/extra/Configs/Config.arm
>> index b060ace..7cd132d 100644
>> --- a/extra/Configs/Config.arm
>> +++ b/extra/Configs/Config.arm
>> @@ -88,6 +88,10 @@ config CONFIG_ARM926T
>>       bool "Arm 926T"
>>       select ARCH_HAS_MMU
>>
>> +config CONFIG_ARM926EJ_S
>> +     bool "Arm 926EJ-S"
>> +     select ARCH_HAS_MMU
>> +
>>  config CONFIG_ARM10T
>>       bool "Arm 10T"
>>       select ARCH_HAS_MMU
>
>
> Khem,
> as ARM expert, could you please ack this patch ? thanks. I'm starting
> playing with ARM right now ;-)
>



the patch is in line with the methods we have although I would prefer
to get rid of gcc tuning and arch setting as this is something one
can add to extra cflags as well if the compiler is not tuned for a
given SOC. Ideally we should only keep the options which denote
anamolies
and need to take care in assembly e.g. if some instruction is not
present in one arch etc.

> Carmelo
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (MingW32)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>
> iEYEARECAAYFAk0i/woACgkQoRq/3BrK1s+IfgCg0Vn4rnMUHz0k57Grrw/YH0Va
> CNkAniItDywIQVESFWIoat7QclXbFW04
> =nl2Q
> -----END PGP SIGNATURE-----
>