1 |
On Wed, 29 May 2019 10:27:34 -0700 (PDT) |
2 |
Palmer Dabbelt <palmer@××××××.com> wrote: |
3 |
|
4 |
> On Mon, 20 May 2019 02:44:18 PDT (-0700), aballier@g.o wrote: |
5 |
> > On Sat, 18 May 2019 20:47:28 +0200 |
6 |
> > Michał Górny <mgorny@g.o> wrote: |
7 |
> > |
8 |
> >> On Fri, 2019-05-03 at 23:34 +0200, Andreas K. Huettel wrote: |
9 |
> >> > * We will initially add two profiles to profile.desc: |
10 |
> >> > default/linux/riscv/17.0/rv64gc/lp64d (non-multilib, 64bit |
11 |
> >> > hardfloat) default/linux/riscv/17.0/rv64gc (multilib lp64d/lp64, |
12 |
> >> > i.e. hard/softfloat) |
13 |
> >> |
14 |
> >> I still don't understand the purpose of this multilib. If you have |
15 |
> >> a hardfloat CPU, why would you ever build some of the software |
16 |
> >> softfloat? |
17 |
> > |
18 |
> > One reason I could imagine is that the hardfloat isn't IEEE 754 |
19 |
> > compliant. Searching through the RISC-V spec, it does not seem to be |
20 |
> > the case here (ie: it is required to be compliant) so I'm also |
21 |
> > wondering what is the point here. |
22 |
> |
23 |
> The RISC-V floating-point extensions are IEEE-754 compliant, but |
24 |
> they're optional. We have chips without floating-point units, but |
25 |
> right now all the Linux capable chips have FPUs. As far as I know |
26 |
> there are no Linux binaries that anyone cares about that are compiled |
27 |
> for systems without hardware floating-point units, but I may be wrong |
28 |
> about that one. |
29 |
|
30 |
|
31 |
It was my understanding that FPU is not optional for rv64gc, is that |
32 |
correct ? |
33 |
|
34 |
> The non-FPU systems are much more interesting in embedded land, where |
35 |
> lots of users don't have FPUs. That's less relevant for Gentoo, but |
36 |
> I do use crossdev embedded toolchains. |
37 |
|
38 |
|
39 |
You'll probably not be using multilib here but rather a specific CHOST |
40 |
and/or flags to enable softfloat everywhere. |
41 |
|
42 |
|
43 |
[...] |
44 |
|
45 |
|
46 |
Alexis. |