Tool chain build - gcc won't compile

[sumeetkpatel]

Trying to build the toolchain on an OrangePi Lite using Armbian nightly.

Main compilation error:

checking for suffix of object files... configure: error: in `/root/open-ath9k-htc-firmware-master/toolchain/build/gcc-6.2.0/xtensa-elf/libgcc':
configure: error: cannot compute suffix of object files: cannot compile
See `config.log' for more details.
Makefile:10972: recipe for target 'configure-target-libgcc' failed
make[2]: *** [configure-target-libgcc] Error 1
make[2]: Leaving directory '/root/open-ath9k-htc-firmware-master/toolchain/build/gcc-6.2.0'
Makefile:856: recipe for target 'all' failed
make[1]: *** [all] Error 2
make[1]: Leaving directory '/root/open-ath9k-htc-firmware-master/toolchain/build/gcc-6.2.0'
Makefile:149: recipe for target '/root/open-ath9k-htc-firmware-master/toolchain/build/gcc-6.2.0/.built' failed
make: *** [/root/open-ath9k-htc-firmware-master/toolchain/build/gcc-6.2.0/.built] Error 2

/toolchain/build/gcc-6.2.0/xtensa-elf/libgcc/config.log config.tar.gz

Currently installed GCC: gcc version 5.4.0 20160609 (Ubuntu/Linaro 5.4.0-6ubuntu1~16.04.4)

See: https://gcc.gnu.org/wiki/FAQ#configure_suffix https://aur.archlinux.org/packages/ath9k-htc-firmware/?comments=all

Apparently it has problems on an ARM based system?

Update: I was able to compile it on an Ubuntu x86 VM, just needed to apt-get install m4 cmake.

[paulfertser]

Hello,

On Wed, May 24, 2017 at 03:39:33PM -0700, Sumeet Patel wrote:

checking for suffix of object files... configure: error: in /root/open-ath9k-htc-firmware-master/toolchain/build/gcc-6.2.0/xtensa-elf/libgcc': configure: error: cannot compute suffix of object files: cannot compile Seeconfig.log' for more details.

So please attach full config.log.

[paulfertser]

Eh, sorry, Github mail confused me as it didn't say anything explicit about the attached files.

Are you following the procedure as described an README to the letter? It probably looks like unpatched xtensa binutils are used.

[olerem]

as far as i can see, we can't build currently it on armhf: https://tests.reproducible-builds.org/debian/rb-pkg/unstable/armhf/open-ath9k-htc-firmware.html

[paulfertser]

On Thu, May 25, 2017 at 04:01:30AM -0700, Oleksij Rempel wrote:

as far as i can see, we can't build currently it on armhf:

Reproduced on my armhf laptop, and apparently it's the xtensa binutils build itself that's failing straight away. Will try digging later...

[paulfertser]

On Thu, May 25, 2017 at 04:01:30AM -0700, Oleksij Rempel wrote:

as far as i can see, we can't build currently it on armhf: [1]https://tests.reproducible-builds.org/debian/rb-pkg/unstable/armhf/open-ath9k-htc-firmware.html

I suggest those affected by this build failure to test adding this patch to gcc build:

https://paulfertser.info/gcc-xtensa-fix-entry.patch

[jcmvbkbc]

I believe the following patch is the right solution to this issue: https://github.com/jcmvbkbc/gcc-xtensa/commit/82cf9f2bd81e4821c6a076a44a3bb47a04b42520.patch

[paulfertser]

On Mon, May 29, 2017 at 08:45:49PM -0700, Max Filippov wrote:

I believe the following patch is the right solution to this issue: [1]https://github.com/jcmvbkbc/gcc-xtensa/commit/f89616b932b9460ebff3167e363f1d6ac2fa1445.patch

Thank you so much Max! It is the right solution indeed, I confirm it's working as intended. And I've actually noticed the gcc warnings regarding incorrect format strings when building the unpatched xtensa.c but ignored those, silly me.

I guess it would probably make sense to have -Werror- for those, as codegeneration depends on fprintf working correctly in this case.

As a sidenote, for those building with non-HEAD gcc version, the ChangeLog part of the patch needs to be stripped, but other than that, it applies nicely to 6.3.0.