On 11 January 2017 at 18:29, Burton, Ross <[email protected]> wrote:
> > On 6 January 2017 at 16:54, Khem Raj <[email protected]> wrote: > >> regardless, I posted a patch for ppp for fixing same issue, seee >> >> https://patchwork.openembedded.org/patch/135536/ >> >> may be the same or similar patch can be applied to connman as well. >> > > Similar but not the same. > Several core connman files do include linux/if.h (and xtables includes net/if.h) but I don't think these can be avoided. connman/src/iptables.c includes <xtables.h> and nothing else relevant as > far as I can tell. However the fragment of code that is throwing an error > (pulled in via xtables.h) is only enabled because the kernel headers don't > think that the libc is defining IFF_LOWER_UP, but musl is. > https://git.busybox.net/buildroot/commit/?id= > 4470dd9b1bdbb9b39e9fb13f27bcaa044719de6d is how buildroot work around > this, by defining a _UAPI symbol that the kernel header listens to. > I did pretty much the same "fix" inside connman (defined __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO 0 everywhere the problem shows up) and it works... but should the hack be done somewhere in the toolchain instead? - Jussi > Ross > > -- > _______________________________________________ > Openembedded-core mailing list > [email protected] > http://lists.openembedded.org/mailman/listinfo/openembedded-core > >
-- _______________________________________________ Openembedded-core mailing list [email protected] http://lists.openembedded.org/mailman/listinfo/openembedded-core
