If the last clkids defined have no hw pointer associated with them in the hw_onecell_data array, the initialization is going to crash badly because the array is not big enough.
The patchset makes sure the array is big enough, and that each clkid is associated with a valid hw pointer, or NULL. Doing it in two patches to ease backport, since we are not fixing the same commit on meson8b and gxbb Jerome Brunet (2): clk: meson: meson8b: fix protection against undefined clks clk: meson: gxbb: fix protection against undefined clks drivers/clk/meson/gxbb.c | 2 ++ drivers/clk/meson/meson8b.c | 1 + 2 files changed, 3 insertions(+) -- 2.9.4
