On Mon, 2019-05-06 at 16:06 -0500, Rob Herring wrote: > +Lubomir > > On Mon, May 6, 2019 at 2:31 PM Randy Dunlap <[email protected]> wrote: > > From: Randy Dunlap <[email protected]> > > > > Fix section mismatch warning: > > > > WARNING: vmlinux.o(.text+0x36e00): Section mismatch in reference from the > > function olpc_dt_compatible_match() to the function > > .init.text:olpc_dt_getproperty() > > The function olpc_dt_compatible_match() references > > the function __init olpc_dt_getproperty(). > > This is often because olpc_dt_compatible_match lacks a __init > > annotation or the annotation of olpc_dt_getproperty is wrong. > > > > All calls to olpc_dt_compatible_match() are from __init functions, > > so it can be marked __init also. > > > > Signed-off-by: Randy Dunlap <[email protected]> > > Cc: [email protected] > > Cc: Rob Herring <[email protected]> > > Fixes: a7a9bacb9a32 ("x86/platform/olpc: Use a correct version when > making up a battery node") > Acked-by: Rob Herring <[email protected]>
Thanks for this. Which tree does this apply to? I can't see the patch that introduce the problem in x86's for-next? I've a mostly equivalent patch lined up with an intent to send it over to x86 once the faulty commit reaches it: https://lists.01.org/pipermail/kbuild-all/2019-April/060269.html In any case; Reviewed-by: Lubomir Rintel <[email protected]> > -int olpc_dt_compatible_match(phandle node, const char *compat) > +int __init olpc_dt_compatible_match(phandle node, const char *compat) My patch also marks olpc_dt_compatible_match() static. It should still be done if this one ends up being applied. Lubo
