Re: possible bug in arm's cpu_setmcontext()?

[minux <minux.ma%gmail.com@localhost>]

On Tue, Mar 5, 2013 at 7:15 AM, Matt Thomas <matt%3am-software.com@localhost> 
wrote:
> On Mar 4, 2013, at 10:45 AM, minux <minux.ma%gmail.com@localhost> wrote:
>
>> #ifdef FPU_VFP
>>        if ((flags & _UC_FPU)
>>            && (curcpu()->ci_vfp_id || (flags & _UC_ARM_VFP) == 0)) {
>>                return EINVAL;
>>        }
>> #endif
>
> Actually, it should be curcpu()->ci_vfp_id == 0.  And I've committed
> that.
>
> The point is that you don't set _UC_FPU if there is no FPU.
> It won't be set by getcontext if there is no FPU.
Makes sense. Thank you.
As long as setcontext(2) doesn't give EINVAL for unmodified
ucontext from kernel, i'm happy.