[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] How does ARM VFP is emulated?
From: |
Peter Maydell |
Subject: |
Re: [Qemu-devel] How does ARM VFP is emulated? |
Date: |
Fri, 17 Aug 2012 13:11:57 +0100 |
On 17 August 2012 12:27, 陳韋任 (Wei-Ren Chen) <address@hidden> wrote:
> On Fri, Aug 17, 2012 at 10:29:24AM +0200, Laurent Desnogues wrote:
>> On Thursday, August 16, 2012, Oi Khote <address@hidden> wrote:
>> > So how exactly does VFP is being emulated.
>>
>> QEMU uses a library for FP computations, based on the softfloat package.
>
> I thought QEMU emulates VFP itself
Depends how you think about it. We do emulate all the VFP
with integer operations...
> But I am wrong, it acutally does this, right?
>
> float64 VFP_HELPER(sqrt, d)(float64 a, CPUARMState *env)
> {
> return float64_sqrt(a, &env->vfp.fp_status);
> }
>
> And float64_sqrt is implemented in fpu/* .
...it just happens that the code that does generic IEEE floating
point emulation lives in fpu/ and is shared between all targets.
(That code is a library called 'softfloat' but we have modified
our copy of it quite a bit over the years.)
-- PMM