Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 11 Sep 2013 12:51:29 -0400
From: Rich Felker <dalias@...ifal.cx>
To: musl@...ts.openwall.com
Subject: Re: Arm AArch64 port - search for interested people

On Wed, Sep 11, 2013 at 02:58:51PM +0200, Szabolcs Nagy wrote:
> * Bortis Kevin <Kevin.Bortis@....ch> [2013-09-11 13:38:03 +0200]:
> > Is someone interessted to help out with a musl ARM Aarch64 port? I got
> > the message, that I will receive a dev board Q2/2014 for a new project,
> > so I hope to have a working musl toolchain by then :)
> > 
> > I have started a wiki page where I gather all Information needed for the
> > port. First step will be, learn A64 assembler/registers and get the ARM
> > Foundation Model running with a Linux distribution. I also hope that the
> > musl documentation with information on musl internals and porting will
> > be finished by then ;)
> > 
> > http://wiki.musl-libc.org/wiki/Porting/AArch64
> 
> it seems long double is IEEE binary128 in the abi,
> but there are no quad precision fp instructions..

It's really unfortunate that AArch64 does not have instructions for
ld128; even if they were unimplemented, they could be emulated to
avoid the hideousness of softfloat.

> so good that i did not remove the ld128 code from libm,
> however complete support will take some time.. and it will
> depend on the soft-float implementation in gcc which has
> fenv issues

I would hope on archs with hardfloat but where long double is
implemented with softfloat, gcc is smart enough to do something with
the fpu to raise the appropriate flags. Any idea if this is the case,
or if the AArch64 ABI mentions the issue at all?

Rich

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.