ldrb vs ldrh in gcc-3.4.4 (ARM)

[Steven Scholz]

Richard Earnshaw wrote:

>>>>The arm9tdmi is the main 'core' of the CPU.  The arm920t is the core
>>>>plus the associated cache, mmu, and system registers.   Different models
>>>>all have substantially the same 'core' but may have different (or
>>>>missing) additional parts...
>>>
>>>
>>>Thanks. So when I am using Freescale i.MX and Atmel AT91RM9200 which are 
>>>both arm920t I rather use
>>>
>>>    GCC_EXTRA_CONFIG="--with-float=soft --with-cpu=arm920t \
>>>                          --enable-cxx-flags=-mcpu=arm920t"
>>>
>>>instead of
>>>
>>>    GCC_EXTRA_CONFIG="--with-float=soft --with-cpu=arm9tdmi \
>>>                          --enable-cxx-flags=-mcpu=arm9tdmi"
>>>
>>>when building the toolchain. Right?
>>
>>ARGH! There's no "--with-cpu=arm920t" !
>>
> 
> This has been fixed in gcc-4.0 and it's also been fixed in the
> CodeSourcery version of 3.4
> (http://www.codesourcery.com/gnu_toolchains/arm/).
> 
> 
>>gcc/config/arm/arm.h only knows
>>
>>#define TARGET_CPU_arm9         0x0080
>>#define TARGET_CPU_arm9tdmi     0x0080

I suppose it's defined as 0x0080 as well!?

>>So when arm9tdmi is only thre core without caches will "--with-cpu=arm9tdmi" use 
>>the caches in a sane way ?
> 
> It probably makes no difference in 3.4 and earlier.  It might have a
> small effect in later versions, but gcc doesn't really understand caches
> and their layout yet.
Ok.

Thanks.

--
Steven

------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sources.redhat.com