mailing list of musl libc
 help / color / mirror / code / Atom feed
From: Andrew Bradford <andrew@bradfordembedded.com>
To: musl@lists.openwall.com
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>,
	Gregor Richards <gr@purdue.edu>,
	CLFS development discussion <clfs-dev@lists.cross-lfs.org>,
	CLFS Support <clfs-support@lists.cross-lfs.org>
Subject: Re: [musl] Re: [Clfs-dev] CLFS embedded book switches to musl-libc
Date: Wed, 23 Oct 2013 07:07:24 -0400	[thread overview]
Message-ID: <5267ADEC.9000107@bradfordembedded.com> (raw)
In-Reply-To: <20131022210846.GZ20515@brightrain.aerifal.cx>

On 10/22/13 17:08, Rich Felker wrote:
> On Tue, Oct 22, 2013 at 02:58:24PM -0400, Kevyn-Alexandre Paré wrote:
>> For the musl installation [2], the make install have an error that
>> seem to be fix with the next command you propose:
>>
>> ../tools/install.sh -D -m 644 lib/crt1.o
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/crt1.o
>> ../tools/install.sh -D -m 644 lib/Scrt1.o
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/Scrt1.o
>> ../tools/install.sh -D -m 644 lib/crti.o
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/crti.o
>> ../tools/install.sh -D -m 644 lib/crtn.o
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/crtn.o
>> ../tools/install.sh -D -m 644 lib/libc.a
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libc.a
>> ../tools/install.sh -D -m 755 lib/libc.so
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libc.so
>> ../tools/install.sh -D -m 644 lib/libm.a
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libm.a
>> ../tools/install.sh -D -m 644 lib/librt.a
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/librt.a
>> ../tools/install.sh -D -m 644 lib/libpthread.a
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libpthread.a
>> ../tools/install.sh -D -m 644 lib/libcrypt.a
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libcrypt.a
>> ../tools/install.sh -D -m 644 lib/libutil.a
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libutil.a
>> ../tools/install.sh -D -m 644 lib/libxnet.a
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libxnet.a
>> ../tools/install.sh -D -m 644 lib/libresolv.a
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libresolv.a
>> ../tools/install.sh -D -m 644 lib/libdl.a
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libdl.a
>> ../tools/install.sh -D -l
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/lib/libc.so
>> /lib/ld-musl-arm.so.1 || true
>> ln: failed to create symbolic link `/lib/ld-musl-arm.so.1.tmp.47054':
>> Permission denied
>> ../tools/install.sh -D -m 644 arch/arm/bits/alltypes.h
>> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/include/bits/alltypes.h
> 
> This (attempting to install the dynamic linker symlink in /lib) is
> caused by specifying the installation location
> /mnt/clfs/cross-tools/armv7a-unknown-linux-musleabi/ as part of the
> prefix rather than as the DESTDIR at "make install" time. If this is
> part of the installation into the new filesystem root that will be
> later used as chroot or the root for booting, DESTDIR is the proper
> approach. If this is just a temporary musl build for use in
> bootstrapping, the current --prefix is fine, but you also need either
> --syslibdir or --disable-shared.

Yes, I had run into this issue as well.  I had just missed that
permission denied failure in haste and thus my brute force symlink.

The goal is to build a cross toolchain which is stand alone but from
which the libs are copied into the actual target which will boot.  This
is roughly what buildroot does when using an external toolchain (I believe).

In this case, using --disable-shared does not seem reasonable as I'd
like to have shared libs.  I'll take a look at --syslibdir and DESTDIR
and see if they accomplish what I think I want.  The symlink from ld to
libc.so was my quick work around and produces a system which boots and
runs fine as far as I can tell but which now seems as though it may not
be the best method (besides my symlink command not covering all cases
for the archs used in the CLFS book or supported by musl).

Thanks,
Andrew
_______________________________________________
Clfs-support mailing list
Clfs-support@lists.cross-lfs.org
http://lists.cross-lfs.org/listinfo.cgi/clfs-support-cross-lfs.org

  reply	other threads:[~2013-10-23 11:07 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-10-16 19:59 Andrew Bradford
2013-10-16 20:38 ` John Spencer
2013-10-16 20:46   ` Andrew Bradford
2013-10-16 21:22     ` John Spencer
2013-10-17 12:50       ` Andrew Bradford
2013-10-22 18:58 ` [Clfs-dev] " Kevyn-Alexandre Paré
2013-10-22 21:08   ` Rich Felker
2013-10-23 11:07     ` Andrew Bradford [this message]
2013-10-23 14:13       ` Rich Felker

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=5267ADEC.9000107@bradfordembedded.com \
    --to=andrew@bradfordembedded.com \
    --cc=clfs-dev@lists.cross-lfs.org \
    --cc=clfs-support@lists.cross-lfs.org \
    --cc=gr@purdue.edu \
    --cc=musl@lists.openwall.com \
    --cc=thomas.petazzoni@free-electrons.com \
    /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.
Code repositories for project(s) associated with this public inbox

	https://git.vuxu.org/mirror/musl/

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 NNTP newsgroup(s).