This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [RFC v1 00/16] RISC-V glibc port for the 32-bit
- From: Joseph Myers <joseph at codesourcery dot com>
- To: Alistair Francis <alistair dot francis at wdc dot com>
- Cc: <libc-alpha at sourceware dot org>, <alistair23 at gmail dot com>
- Date: Wed, 24 Jul 2019 20:01:21 +0000
- Subject: Re: [RFC v1 00/16] RISC-V glibc port for the 32-bit
- References: <cover.1561177967.git.alistair.francis@wdc.com>
On Fri, 21 Jun 2019, Alistair Francis wrote:
> This patch set contains the glibc port for the 32-bit RISC-V.
The following are general comments that I think are still applicable to
the current version.
See <https://sourceware.org/glibc/wiki/NewPorts>. A lot of things in
there are inapplicable for this case (architecture variant for an existing
port), but some *are* applicable. For example, you're using GLIBC_2.29
symbol versions, which is wrong; the versions have to be those of the
first release to have the new port (so GLIBC_2.31 or later). As another
example, you need to include details of test results when submitting a new
port.
> The last patch is a hack to forcefully rename the remaining system calls
> that are causing compiller failures to the 64-bit versions. This is
> obviously not the correct way to do this. I would like some feedback on
> this series to find the upstream approved way to use the 64-bit
> versions.
Build on top of the Y2038 series, helping to fix the issues with that
series so it can get in sooner. In general, get to a state before adding
this port where each time-related function, for which you think you might
need changes for RV32, is implemented following the pattern where the main
implementation is for 64-bit times and the version for 32-bit times is a
thin wrapper round that, unless there is a good, well-commented reason to
do otherwise, and then you're much less likely to need further changes for
RV32.
--
Joseph S. Myers
joseph@codesourcery.com