[Buildroot] [PATCH 1/1] linux/linux.mk: Use '-isystem' instead of '-I' in HOST_CFLAGS

Arnout Vandecappelle arnout at mind.be
Mon Sep 3 22:30:49 UTC 2018



On 01/09/2018 09:55, Thomas Petazzoni wrote:
> Hello,
> 
> On Thu, 30 Aug 2018 11:57:48 +0200, David De Grave (Essensium/Mind)
> wrote:
>> Since the last update of the DTC package, linux build breaks at some point
>> with the following error:
>>
>>     In file included from {output}/host/include/libfdt.h:54:0,
>>                      from scripts/dtc/libfdt/fdt.c:54:
>>     {output}/host/include/libfdt_env.h:82:24: error: redefinition of \
>>     ‘fdt16_to_cpu’
>>      static inline uint16_t fdt16_to_cpu(fdt16_t x)
>>                             ^~~~~~~~~~~~
>>     In file included from scripts/dtc/libfdt/fdt.c:51:0:
>>     scripts/dtc/libfdt/libfdt_env.h:81:24: note: previous definition of \
>>     ‘fdt16_to_cpu’ was here
>>
>> This is because the HOST_CFLAGS is added before Linux's makefile add it's
>> own local ones (-Iscripts/dtc/libfdt), resulting in a mix of local and host
>> headers beeing used.
>>
>> So, to fix this the '-I' is substituted by '-isystem' to let it search for
>> the headers in the right location first.
>>
>> Signed-off-by: David De Grave (Essensium/Mind) <david.degrave at mind.be>
>> ---
>>  linux/linux.mk | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> I suspect this one is superseded by your new patch series
> http://patchwork.ozlabs.org/project/buildroot/list/?series=63524, so
> I've marked this patch as Changes Requested.

 Actually, the idea was to apply this patch on master because it fixes a bug,
and the series on next because it is more invasive.

 It turns out, however, that the problem only occurs for the beaglebone kernel
on github, because that one builds its fdtoverlay host tool. "Normal" kernels
don't build libfdt at all for the host, so the error reported doesn't occur.

> 
> Note that
> http://patchwork.ozlabs.org/project/buildroot/list/?series=63524
> doesn't contain PATCH 2/3 due to a patchwork issue when associating
> patches to series, but PATCH 2/3 has been recorded by patchwork.

 Patch 2/3 actually doesn't apply, it's a revert of this patch... So I've marked
that one as N/A.

 Regards,
 Arnout

> 
> Best regards,
> 
> Thomas
> 

-- 
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF



More information about the buildroot mailing list