This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: RISC-V glibc port v2
- From: Joseph Myers <joseph at codesourcery dot com>
- To: Christoph Hellwig <hch at lst dot de>
- Cc: Palmer Dabbelt <palmer at dabbelt dot com>, <libc-alpha at sourceware dot org>, Andrew Waterman <andrew at sifive dot com>, Darius Rad <darius at bluespec dot com>, <dj at redhat dot com>
- Date: Wed, 3 Jan 2018 13:42:35 +0000
- Subject: Re: RISC-V glibc port v2
- Authentication-results: sourceware.org; auth=none
- References: <mhng-c7e28359-5ed1-4fa0-92a3-263d4a0eb70a@palmer-si-x1c4> <alpine.DEB.2.20.1801010106030.28505@digraph.polyomino.org.uk> <20180103133714.GA14077@lst.de>
On Wed, 3 Jan 2018, Christoph Hellwig wrote:
> On Mon, Jan 01, 2018 at 01:19:48AM +0000, Joseph Myers wrote:
> > Do I understand correctly from this that RV64I processors cannot execute
> > RV32I code (unlike e.g. x86_64 where execution of 32-bit code is always
> > supported by the processor, or AArch64 where processors may or may not
> > support execution of AArch32 code)? If so, then indeed
> > bits/environments.h is not relevant at present.
>
> RV64I processors might be able to execute RV32I code (that support
> is optional in the architecture by writing to the sstatus register). But
> the Linux port currently does not support that (yet).
If suppported in future, then all the multi-ABI pieces I mentioned
(ldconfig / ldd support etc.) will become relevant for pieces built with
glibc for one of RV64I and RV32I working properly in the presence on the
system of libraries for the other one of those two, as well as
bits/environments.h.
For now, in the absence of such support, those pieces are only relevant
for differences in floating-point ABI.
--
Joseph S. Myers
joseph@codesourcery.com