public inbox for ecos-patches@sourceware.org
 help / color / mirror / Atom feed
From: Sergei Gavrikov <sergei.gavrikov@gmail.com>
To: John Dallaway <john@dallaway.org.uk>
Cc: Simon Kallweit <simon.kallweit@intefo.ch>,
	    eCos Patches <ecos-patches@ecos.sourceware.org>,
	    Uwe Kindler <uwe_kindler@web.de>
Subject: Re: uSTL 1.4 update
Date: Mon, 28 Jun 2010 19:21:00 -0000	[thread overview]
Message-ID: <alpine.DEB.2.00.1006282204030.29773@sg-laptop> (raw)
In-Reply-To: <alpine.DEB.2.00.1006281624030.9021@sg-desktop.local>

On Mon, 28 Jun 2010, Sergei Gavrikov wrote:

> On Mon, 28 Jun 2010, Simon Kallweit wrote:
>
>> On 06/28/2010 02:59 PM, John Dallaway wrote:
>>> Hi Simon
>>> 
>>> Simon Kallweit wrote:
>>> 
>>>> Here is the updated patch.
>>>> 
>>>> Added a generic implementation for the NextPow2, which is currently
>>>> suboptimal but mimics what the x86 implementation is doing. Also removed
>>>> the introduced change in include/ustl/uctrstrm.h so it should build
>>>> again with -fno-rtti.
>>> 
>>> Thank you. Tests are now building fine for M5272C3 but there's an error
>>> building for the synthetic target on my CentOS 5 (32-bit) box:
>>> 
>>>> gcc -L/var/tmp/ustl-test/install/lib -Ttarget.ld -o 
>>>> /var/tmp/ustl-test/install/tests/language/cxx/ustl/current/tests/bvt23 
>>>> tests/bvt23.o -g -nostdlib -Wl,--gc-sections -Wl,-static
>>>> tests/bvt23.o: In function `ustl::simd::fround<double, 
>>>> int>::operator()(double const&) const':
>>>> /var/tmp/ustl-test/install/include/ustl/simd.h:109: undefined reference 
>>>> to `lrint'
>>>> /var/tmp/ustl-test/install/include/ustl/simd.h:109: undefined reference 
>>>> to `lrint'
>>>> /var/tmp/ustl-test/install/include/ustl/simd.h:109: undefined reference 
>>>> to `lrint'
>>>> tests/bvt23.o: In function `ustl::simd::fround<float, 
>>>> int>::operator()(float const&) const':
>>>> /var/tmp/ustl-test/install/include/ustl/simd.h:107: undefined reference 
>>>> to `lrintf'
>>>> /var/tmp/ustl-test/install/include/ustl/simd.h:107: undefined reference 
>>>> to `lrintf'
>>>> /var/tmp/ustl-test/install/include/ustl/simd.h:107: undefined reference 
>>>> to `lrintf'
>>>> collect2: ld returned 1 exit status
>>>> make[1]: *** 
>>>> [/var/tmp/ustl-test/install/tests/language/cxx/ustl/current/tests/bvt23] 
>>>> Error 1
>>>> make[1]: Leaving directory `/var/tmp/ustl-test/language/cxx/ustl/current'
>>>> make: *** [tests] Error 2
>>> 
>>> Are you seeing this error?
>> 
>> It builds fine on my Ubuntu 10.04 LTS 32-bit :/
>
> Hi Simon
> and John,
>
> I built successfully uSTL tests using the Simon's latest patch for
> arm7tdmi, i386 pc (i386-elf-gcc from eCosCentric) targets.
>
> But, when I tried to build it for i386linux target I got the same error
> likes John got.
>
> Simon, I often run update-manager :-( and my stuff is

[snip]

> I will try to investigate in the issue tonight.

Hi

I could not understand why a build with native GCC wanted 'lrint'
http://www.opengroup.org/onlinepubs/009695399/functions/lrint.html

I tried a hack (added libm.a to linker GROUP), and that helped...

$ sed -i 's,libgcc_eh.a,& libm.a,' install/lib/target.ld

$ grep GROUP install/lib/target.ld
GROUP(libtarget.a libgcc.a libsupc++.a libgcc_eh.a libm.a)

Well, this was brutal hack, but it was possible to build all tests on my
Linux box. Should we implement 'lrint' in eCos libm?

Sergei

  reply	other threads:[~2010-06-28 19:21 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-06-25 14:33 Simon Kallweit
2010-06-28  8:32 ` John Dallaway
2010-06-28  9:19 ` John Dallaway
2010-06-28  9:34   ` Simon Kallweit
2010-06-28 12:09     ` Simon Kallweit
2010-06-28 12:59       ` John Dallaway
2010-06-28 13:15         ` Simon Kallweit
2010-06-28 13:51           ` Sergei Gavrikov
2010-06-28 19:21             ` Sergei Gavrikov [this message]
2010-06-29  6:34               ` Sergei Gavrikov
2010-06-29  7:17                 ` John Dallaway
2010-06-29  7:23                   ` Simon Kallweit
2010-06-29  8:05       ` John Dallaway

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=alpine.DEB.2.00.1006282204030.29773@sg-laptop \
    --to=sergei.gavrikov@gmail.com \
    --cc=ecos-patches@ecos.sourceware.org \
    --cc=john@dallaway.org.uk \
    --cc=simon.kallweit@intefo.ch \
    --cc=uwe_kindler@web.de \
    /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).