On 12/7/16, Barry Kauler <[email protected]> wrote:
> Oh, I have just about completed compiling my selection of packages on
> my Raspberry Pi3, hit a failure with package audio/ladspa.
>
> In T2 config, I chose CPU "cortex-a8" and enabled thumb instructions.
> I did NOT tick the "soft floating point" option.
>
> However, I have just discovered that all my packages have compiled
> with soft floating point, not using the VFP or NEON.
>
> I found this excellent page, that explains why:
>
> https://community.arm.com/groups/tools/blog/2013/04/15/arm-cortex-a-processors-and-gcc-command-lines
>
> Just specifying "cortex-a8" does not turn on the vfp or neon, they
> have to be explicitly turned on.
>
So, all arm builds in T2 are partially crippled, as they all use
software floating point.
The "-msoft-float" option is redundant, you get that whether it is there or not.
As a temporary fix, to cover my choice of "cortex-a8", I have appended
this line to architecture/arm/gcc-options.in:
[ "$SDECFG_SOFTFLOAT" = 1 ] ||
var_append GCC_WRAPPER_INSERT " " "-mfpu=vfpv3 -mfloat-abi=hard"
However, it really needs a case statement, to put in the appropriate
options for each type of arm architecture..
Regards,
Barry Kauler
-----------------------------------------------------------
If you wish to unsubscribe from this mailing, send mail to
[email protected] with a subject of: unsubscribe t2
