public inbox for newlib@sourceware.org
 help / color / mirror / Atom feed
From: Torbjorn SVENSSON <torbjorn.svensson@foss.st.com>
To: Chris Packham <judge.packham@gmail.com>
Cc: <newlib@sourceware.org>
Subject: Re: Build error for moxiebox architecture
Date: Wed, 20 Sep 2023 20:24:40 +0200	[thread overview]
Message-ID: <11be429e-0acd-b1e9-ecda-3882c691de34@foss.st.com> (raw)
In-Reply-To: <CAFOYHZCrB7B_fR5eY-COPEA6AxUDYCjzK8D-t_O1Kt5T3XjYZg@mail.gmail.com>



On 2023-09-19 10:04, Chris Packham wrote:
> On Mon, Aug 14, 2023 at 10:24 PM Torbjorn SVENSSON
> <torbjorn.svensson@foss.st.com> wrote:
>>
>> Hello,
>>
>> On 2023-08-13 02:27, Chris Packham wrote:
>>> Hi All,
>>>
>>> I'm getting the next crosstool-ng release together and one build
>>> failure we're having is for the moxiebox architecture:
>>>
>>> moxie-unknown-moxiebox/lib/libc.a(libc_a-closer.o): in function `_close_r':
>>> moxie-unknown-moxiebox/src/newlib/newlib/libc/reent/closer.c:47:
>>> undefined reference to `_close'
>>>
>>> and similar for _lseek, _read, _write and _sbrk
>>
>> As you do not include the link line, I'm assuming that you are not using
>> anything out of the ordinary. In this case, maybe you should give
>> --specs=nosys.specs a go and see if it helps.
>> Regardless if it helps or not, it will likely only provide you with
>> stubs that you need to override in you application to get something working.
> 
> Yeah sorry. I'll get things in an issue for ct-ng so we have a proper
> record to refer back to.
> 
> One thing I have found is that things work fine with 4.2.0.20211231
> but fail with 4.3.0.20230120 (or the tip of main). I did see the
> recent commit 55485616b ("Bring back libsim.a for riscv") and I note
> the moxiebox definition of _close() is in libsim.a so this made me
> wonder if whatever issue was affecting riscv was related to the moxie
> failure. But I also note that there has been a bit of a change in some
> Makefiles as per commit 66dfedc28 ("libgloss: merge bfin into
> top-level Makefile").

I think you should not link with libsim.a when you are not using a 
simulator.

> 
> So as a first step maybe I need to do a similar conversion. It looks
> like those conversions involve some code changes and then the result
> of running automake but I'm not sure what parts are manual and what
> parts are from running automake. Any pointers on doing this for
> another architecture.

Sorry, but I can't help you on that.

      reply	other threads:[~2023-09-20 18:24 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-13  0:27 Chris Packham
2023-08-14 10:24 ` Torbjorn SVENSSON
2023-09-19  8:04   ` Chris Packham
2023-09-20 18:24     ` Torbjorn SVENSSON [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=11be429e-0acd-b1e9-ecda-3882c691de34@foss.st.com \
    --to=torbjorn.svensson@foss.st.com \
    --cc=judge.packham@gmail.com \
    --cc=newlib@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).