[yocto] ac_nonexistent.h error in do_configure

Muhlenkamp, Lewis lewis.muhlenkamp at stryker.com
Tue Oct 2 08:53:22 PDT 2018


Hello,

I get a failure when trying to build the core-image-x11 image with the intel-corei7-64 MACHINE.  The error message that appears in the build/tmp-glibc/corei7-64-oe-linux/xf86-video-intel/2_2.99.917+gitAUTOINC+e4fe79cf0d-r0/build/config.log file is as follows:

=== Start excerpt ===
configure:4381: checking how to run the C preprocessor
configure:4451: result: x86_64-oe-linux-gcc -E --sysroot=/home/lmuhlenkamp/build-target02/tmp-glibc/work/corei7-64-oe-linux/xf86-video-intel/2_2.99.917+gitAUTOINC+e4fe79cf0d-r0/recipe-sysroot  -m64 -march=nehalem -mtune=generic -mfpmath=sse -msse4.2
configure:4471: x86_64-oe-linux-gcc -E --sysroot=/home/lmuhlenkamp/build-target02/tmp-glibc/work/corei7-64-oe-linux/xf86-video-intel/2_2.99.917+gitAUTOINC+e4fe79cf0d-r0/recipe-sysroot  -m64 -march=nehalem -mtune=generic -mfpmath=sse -msse4.2  conftest.c
configure:4471: $? = 0
configure:4485: x86_64-oe-linux-gcc -E --sysroot=/home/lmuhlenkamp/build-target02/tmp-glibc/work/corei7-64-oe-linux/xf86-video-intel/2_2.99.917+gitAUTOINC+e4fe79cf0d-r0/recipe-sysroot  -m64 -march=nehalem -mtune=generic -mfpmath=sse -msse4.2  conftest.c
conftest.c:11:10: fatal error: ac_nonexistent.h: No such file or directory
#include <ac_nonexistent.h>
          ^~~~~~~~~~~~~~~~~~
compilation terminated.
configure:4485: $? = 1
configure: failed program was:
| /* confdefs.h */
| #define PACKAGE_NAME "xf86-video-intel"
| #define PACKAGE_TARNAME "xf86-video-intel"
| #define PACKAGE_VERSION "2.99.917"
| #define PACKAGE_STRING "xf86-video-intel 2.99.917"
| #define PACKAGE_BUGREPORT "https://bugs.freedesktop.org/enter_bug.cgi?product=xorg"
| #define PACKAGE_URL ""
| #define PACKAGE "xf86-video-intel"
| #define VERSION "2.99.917"
| /* end confdefs.h.  */
| #include <ac_nonexistent.h>
=== End excerpt ===

I understand that ac_nonexistent.h is not supposed to exist, and that the autoconf stuff is supposed to handle this failure as a successful test if the gcc compiler handles things properly, but it does not look like the gcc compiler is handling things properly, and I do not know why.

What I am trying to do is to get an image using the minimal amount of meta layers as possible, and to not use poky.  I am still learning Yocto.  This is a step in my learning process, but I do not understand Yocto enough to determine how to proceed.

I did the following


  *   cd /home/lmuhlenkamp
  *   git clone git://git.yoctoproject.org/meta-intel
  *   git clone git://git.openembedded.org/meta-openembedded
  *   git clone git://git.openembedded.org/openembedded-core oe-core
  *   cd oe-core/
  *   git clone git://git.openembedded.org/bitbake
  *   . ./oe-init-build-env ~/build-target02

I made the following modifications to bblayers.conf:


  *   diff bblayers.conf .#bblayers.conf.orig
10,16d9
<   /home/lmuhlenkamp/meta-openembedded/meta-python \
<   /home/lmuhlenkamp/meta-openembedded/meta-gnome \
<   /home/lmuhlenkamp/meta-openembedded/meta-filesystems \
<   /home/lmuhlenkamp/meta-openembedded/meta-oe \
<   /home/lmuhlenkamp/meta-openembedded/meta-networking \
<   /home/lmuhlenkamp/meta-openembedded/meta-initramfs \
<   /home/lmuhlenkamp/meta-intel \

I made the following modifications to local.conf:


  *   diff local.conf .#local.conf.orig
28,31d27
< # From the meta-intel BSP layer
< #MACHINE ?= "intel-core2-32"
< MACHINE ?= "intel-corei7-64"
< #
46c42
< DL_DIR ?= "/downloads"
---
> #DL_DIR ?= "${TOPDIR}/downloads"
221,224d216
< PACKAGE_CLASSES = "package_rpm"
< VIRTUAL-RUNTIME_init_manager = "systemd"
< DISTRO_FEATURES_append = " systemd"
< DISTRO_FEATURES_BACKFILL_CONSIDERED_append = " sysvinit"


As you may have guessed by some of the above changes, I eventually want to be able to use the meta-intel BSP with the meta-anaconda layer.  My end goal is to be able to create a bootable DVD that can be given to our field engineers to be able to do either fresh installs in the field, or upgrade existing hardware in the field.  I am familiar with Anaconda from doing a lot of work with Red Hat derived OSes.  So, seemed liked a good way to go.

I'm thinking I need to make some other change to local.conf to get the ac_nonexistent.h test "succeed".  I am not sure what that change is.

I've looked through a lot of online documentation.  I have done a lot of Google searches.  I have not found anything helpful.

Any help anyone can provide would be greatly appreciated, mainly on the ac_nonexistent.h error, but if anyone has any helpful pointers or such for combining meta-intel and meta-anaconda, I would really appreciate that too.

Thank you

Lewis Muhlenkamp

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.yoctoproject.org/pipermail/yocto/attachments/20181002/28747ce8/attachment.html>


More information about the yocto mailing list