John Kearney
|
Re: does qnx use gnu glibc?
|
John Kearney
10/30/2013 3:23 PM
post106421
|
Re: does qnx use gnu glibc?
You still need to make sure the those env vars are set correctly.
And you also still need to link against the qnx libc, unless you are doing a bootloader etc...
Sent from my BlackBerry 10 smartphone.
From: Toan Pham
Sent: Wednesday, October 30, 2013 20:05
To: general-toolchain@community.qnx.com
Reply To: general-toolchain@community.qnx.com
Subject: Re: does qnx use gnu glibc?
John. I am not using the makefile method. I try on going away frome qnx recursive make method. Essentially. I just
want to use GCC and binutils and nothing more to build more open source software packages
John Kearney <community-noreply@qnx.com> wrote:
><!-- .EmailQuote { margin-left: 1pt; padding-left: 4pt; border-left: #800000 2px solid; } -->
>
>Have you set
>
>QNX_TARGET and QNX HOST, it sounds like your build environment isn't set correctly.
>
>
>
>Sent from my BlackBerry 10 smartphone.
>
>From: Toan Pha
>
>Sent: Wednesday, October 30, 2013 18:18
>
>To: general-toolchain
>
>Reply To: general-toolchain@community.qnx.com
>
>Subject: Re: RE: does qnx use gnu glibc?
>
>
>
>Thank you,
>
>I understand now,
>
>I am trying to make a build environment on x86-linux platform to cross compile some open source packages to armv7-nto-
qnx. Since, I do not rebuild qnx compiler, and only use the downloaded toolchain along with native/target libraries
provided by this project, I do not have control for the linking process.
>
>The problem i am having right now is that the binary tool (gcc), always want to load in c startup objects in armv7/
armle-v7/lib/crt1.o with a simple hello world program. That path was preconfigured while building the compiler for qnx.
it was probably defined using the preprocessor while building GCC:
>
>#define STANDARD_STARTFILE_PREFIX "/some_cross_platform_library_path/lib/"
>
>
>
>I would like to know if there is anyway I can modify the current precompiled linker (or linker script) to not search
for C-runtime (crt.o) in its standard location at armv7/armle-v7/lib/crt1.o. I tried telling the linker where to find
the c-runtime using -L -I search path arguments and the linker still error out, saying that path to object armv7/armle-
v7/lib/crt1* not found. Certainly, since I am making a build environment, i do not want to be in the target directory
where the linker can find its C-runtime library all the time.
>
>
>Presently, I can only compile a hello world application under my development environment by telling gcc not to use its
linker script and c-runtime libraries by passing the following arguments.
>
>ntoarmv7-gcc -nostdlib -L$root/usr/lib hello.c $root/usr/lib/armv7/armle-v7/lib/{crt*.o,libc.so.3}
>
> This is only a work-around method. To have a truely cross platform toolchain, like the toolchain I made for arm-linux
, the linker must not always try to find its C-runtime library at a fixed location.
>
>
>
>
>_______________________________________________
>
>General
>http://community.qnx.com/sf/go/post106418
>To cancel your subscription to this discussion, please e-mail general-toolchain-unsubscribe@community.qnx.com
>
_______________________________________________
General
http://community.qnx.com/sf/go/post106420
To cancel your subscription to this discussion, please e-mail general-toolchain-unsubscribe@community.qnx.com
|
|
|