[Buildroot] [PATCH 1/3] package/Makefile.in: Use '-isystem' instead of '-I' in HOST_CFLAGS globally
Yann E. MORIN
yann.morin.1998 at free.fr
Sun Oct 21 14:18:22 UTC 2018
David, All,
On 2018-09-10 15:00 +0200, David De Grave spake thusly:
> Hello Thomas, Arnout,
> On Mon, Sep 10, 2018 at 11:51 AM, Arnout Vandecappelle < [1]arnout at mind.be> wrote:
>
>
> > Could you give a few more details about the motivation for this change.
> > Is it just because it is "cleaner" to refer to Buildroot
> > $(HOST_DIR)/include using -isystem, or does this fix some actual
> > problem/build issue ?
>
> More than making things cleaner/logic, it fixes a build issue I've with the beagle bone
> black's Linux kernel: [2]https://github.com/beagleboard/linux.git
> They implemented a driver who handle DTBs overlays and what they call "capes"
> (PCB extensions to the bbb mother board). Since the update of the DTC package few
> weeks ago, my build is throwing compilation errors in linux when it starts to compile
> their own version of the dtc. Due to the way they include their headers, some of them
> were taken from the host while some others are taken from local dir. Moreover, their
> dtc version use different names in their defines (a missing underscore) and this why
> at a certain point, it throws a "function redefinition error". Maybe other kernels are
> suffering from the same problem, I don't know...
> I first fixed the issue by adding a patch to my project to change the involved sources
> in the Kernel tree...
> But then I thought that other people may suffers from this problem too and decided to
> make it more generic by patching buildroot itself, so that I could save them few hours
> of investigations. This is why I submitted my first patch who applied the same kind of
> change in the [3]linux.mk file.
> Then, I discussed this with Arnout and we thought it would be even more benefit to
> make it global as it's already a requirement for uboot and now linux too. Moreover it
> sounds more logic/cleaner to make it that way, but that was not the main idea.
> So, that's why I submitted this patch series.
The -isystem has already been tried, and lately resurfaced in another
thread, in which it was noticed that it broke at least a few packages,
please see:
http://lists.busybox.net/pipermail/buildroot/2018-October/232615.html
So, before we reinstate the use of -isystem, a few packages will have to
be fixed first.
Thanks!
Regards,
Yann E. MORIN.
> > Doing such a change is pretty invasive, it could potentially cause
> > some build failures, so I'd like to understand the motivation.
>
> I understand, no problem :-)
> Regards,
> David.
> --
>
> +--------------------------------------------------------------------------------------------------------+
> | | David De Grave |
> | | Senior Embedded Software Developer |
> | | Gsm : +32(0)496.364.960 | Tel : +32-16-28.65.00 | Fax : +32-16-28a.65.01 |
> | | [4]Essensium-Mind - Gaston Geenslaan 9, B-3001 Leuven, Belgium |
> +--------------------------------------------------------------------------------------------------------+
>
> Links:
> 1. mailto:arnout at mind.be
> 2. https://github.com/beagleboard/linux.git
> 3. http://linux.mk
> 4. https://www.mind.be/
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
--
.-----------------.--------------------.------------------.--------------------.
| Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ |
| +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. |
'------------------------------^-------^------------------^--------------------'
More information about the buildroot
mailing list