From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/6287 Path: news.gmane.org!not-for-mail From: stephen Turner Newsgroups: gmane.linux.lib.musl.general Subject: Re: libgcc errors Date: Thu, 9 Oct 2014 17:08:35 -0400 Message-ID: References: <20141007195156.GV21835@port70.net> <5434704E.8080504@sholland.net> <54367BF6.80203@bradfordembedded.com> <263E3217-740D-44F2-BC2A-BDCA4AFB38B8@sholland.net> Reply-To: musl@lists.openwall.com NNTP-Posting-Host: plane.gmane.org Mime-Version: 1.0 Content-Type: multipart/alternative; boundary=089e0163412ceef1a8050503d588 X-Trace: ger.gmane.org 1412888936 2988 80.91.229.3 (9 Oct 2014 21:08:56 GMT) X-Complaints-To: usenet@ger.gmane.org NNTP-Posting-Date: Thu, 9 Oct 2014 21:08:56 +0000 (UTC) To: musl@lists.openwall.com Original-X-From: musl-return-6300-gllmg-musl=m.gmane.org@lists.openwall.com Thu Oct 09 23:08:50 2014 Return-path: Envelope-to: gllmg-musl@plane.gmane.org Original-Received: from mother.openwall.net ([195.42.179.200]) by plane.gmane.org with smtp (Exim 4.69) (envelope-from ) id 1XcKx8-0005Sm-DB for gllmg-musl@plane.gmane.org; Thu, 09 Oct 2014 23:08:50 +0200 Original-Received: (qmail 5342 invoked by uid 550); 9 Oct 2014 21:08:48 -0000 Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: Original-Received: (qmail 5331 invoked from network); 9 Oct 2014 21:08:47 -0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; bh=ZBqMggX6ezBejjcKcXRvaZJ/83FDqTOs53Ber5OTbCo=; b=w2iMAgRUMyBqEXUUmjsoCvJ271SvgukdZX1KZouKZQiVuTmlc47/xfcDqmE2VqoUK2 8D/5wAIQXxAZ1nvk1/PLCAYx2GjYNfqJ/akEEOfLkS/uu8Ls2wUHsw3/885UW/BOrvoY 1rla+iO5wOpICbP76sp459E+L4WnyFvrXQAFmHjIeJTEncdmNqIwZ6m4QC4cy75DQiWg 5jkmDPJCWpCpgBPYcbr8q4izeDh71qyMmQVNiMJmEjkXTP/XrkcwtJVMhaC5EoxCnzC8 pYQ4QMGZQQX82QOA/0PYnsLr6aVt+4xETE98j1HOK7p5kz4Cuu5OzFTA0naW4K6v5tbo xbtg== X-Received: by 10.220.49.10 with SMTP id t10mr966315vcf.34.1412888915799; Thu, 09 Oct 2014 14:08:35 -0700 (PDT) In-Reply-To: <263E3217-740D-44F2-BC2A-BDCA4AFB38B8@sholland.net> Xref: news.gmane.org gmane.linux.lib.musl.general:6287 Archived-At: --089e0163412ceef1a8050503d588 Content-Type: text/plain; charset=UTF-8 Thanks Samuel, I would use the musl-gcc wrapper (and have) however once gcc is compiled from it then it does not support dynamic linking due to the lack of --host (suggested) or --target (what seems more appropriate to me) specified in the configure flags. If i do specify it then i get all kinds of crazy errors which is why i consulted this mailing list in the first place since i have tried googling my way through documentation for a month prior to asking for help. Its my inexperienced opinion that simply switching libs from glibc6 to musl should be relatively easy compared to the full cross compile. To me it seems like it would be as easy as specifying a new lib when rebuilding gcc and that would be all. let me ask you one quick question. I went to the musl-cross website on bitbucket/gregorr and grabbed the patch for 4.7.3, is that a good working patch and gcc version? (all my errors seem to stem from compiling gcc) and is there any special way to download it or just click the raw button and wget the link then "patch -Np1 -i file.name.diff"? On Thu, Oct 9, 2014 at 3:38 PM, Samuel Holland wrote: > On October 9, 2014 7:13:42 AM CDT, Andrew Bradford < > andrew@bradfordembedded.com> wrote: > >Could you please provide more info on why you think the embedded Cross > >LFS way [1] is more complicated than is needed? How could it be made > >more simple? I'm happy to take suggestions for improvement. > > > >[1]:http://cross-lfs.org/view/clfs-embedded/arm/ > > > >For the most part the toolchain building portion of embedded CLFS > >follows Gregor's musl-cross. > > Because he doesn't need a cross compiler, only library isolation. Even the > LFS method is more work than absolutely necessary. You could compile musl > with your host GCC, then compile binutils and GCC with musl-gcc, and be > done with it. I recommend the LFS way because 1) it works with C++ and 2) > some people claim musl-gcc makes unreliable GCC builds. > > Cross LFS makes sense when you're planning to use musl on another > machine/architecture, but not (in my opinion) for making a native > toolchain. I have no problems with CLFS; I just think it's the wrong tool > for the job. > > >Thanks, > >Andrew > > -- > Regards, > Samuel Holland > --089e0163412ceef1a8050503d588 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Thanks Samuel,
=C2=A0I would use the musl-gcc wrapper = (and have) however once gcc is compiled from it then it does not support dy= namic linking due to the lack of --host (suggested) or --target (what seems= more appropriate to me) specified in the configure flags. If i do specify = it then i get all kinds of crazy errors which is why i consulted this maili= ng list in the first place since i have tried googling my way through docum= entation for a month prior to asking for help. =C2=A0

<= div>Its my inexperienced opinion that simply switching libs from glibc6 to = musl should be relatively easy compared to the full cross compile. To me it= seems like it would be as easy as specifying a new lib when rebuilding gcc= and that would be all. =C2=A0

let me ask you = one quick question. I went to the musl-cross website on bitbucket/gregorr a= nd grabbed the patch for 4.7.3, is that a good working patch and gcc versio= n? (all my errors seem to stem from compiling gcc) and is there any special= way to download it or just click the raw button and wget the link =C2=A0th= en "patch -Np1 -i file.name.diff"?

On Thu, Oct 9, 2014 at 3:38 PM, Samu= el Holland <samuel@sholland.net> wrote:
On October 9, 2014 7:13:42 AM CDT, Andrew = Bradford <andrew@bradford= embedded.com> wrote:
>Could you please provide more info on why you think the embedded Cross<= br> >LFS way [1] is more complicated than is needed?=C2=A0 How could it be m= ade
>more simple?=C2=A0 I'm happy to take suggestions for improvement. >
>[1]:http://cross-lfs.org/view/clfs-embedded/arm/
>
>For the most part the toolchain building portion of embedded CLFS
>follows Gregor's musl-cross.

Because he doesn't need a cross compiler, only library isolation= . Even the LFS method is more work than absolutely necessary. You could com= pile musl with your host GCC, then compile binutils and GCC with musl-gcc, = and be done with it. I recommend the LFS way because 1) it works with C++ a= nd 2) some people claim musl-gcc makes unreliable GCC builds.

Cross LFS makes sense when you're planning to use musl on another machi= ne/architecture, but not (in my opinion) for making a native toolchain. I h= ave no problems with CLFS; I just think it's the wrong tool for the job= .

>Thanks,
>Andrew

--
Regards,
Samuel Holland <samuel@sholland.n= et>

--089e0163412ceef1a8050503d588--