On Mon, 2 Jul 2018, Mathieu Malaterre wrote: > On Mon, Jul 2, 2018 at 10:25 AM Finn Thain <fth...@telegraphics.com.au> > wrote: > > > > The pmu_init() function has the __init qualifier, but the ops struct > > that holds a pointer to it does not. This causes a build warning. The > > driver works fine because the pointer is only dereferenced early. > > > > The function is so small that there's negligible benefit from using > > the __init qualifier. Remove it to fix the warning, consistent with > > the other ADB drivers. > > Would you mind copy/pasting the warning you are seeing. > > Make sure you have: > > 58935176ad17 powerpc/via-pmu: Fix section mismatch warning > > Thanks >
It's true, the section mismatch warning from 'make' has disappeared since I wrote this patch, but that doesn't mean it is wrong. Before this patch: $ powerpc-linux-gnu-objdump -xda vmlinux |egrep -w "via_pmu_driver|pmu_init" c0711c84 l F .init.text 0000001c pmu_init c05eb408 g O .rodata 00000028 via_pmu_driver c0711c84 <pmu_init>: $ After: $ powerpc-linux-gnu-objdump -xda vmlinux |egrep -w "via_pmu_driver|pmu_init" c038e42c l F .text 0000001c pmu_init c05e1e58 g O .rodata 00000028 via_pmu_driver c038e42c <pmu_init>: $ I gather that commit 58935176ad17 ("powerpc/via-pmu: Fix section mismatch warning") has moved via_pmu_driver from .data to .rodata, but I'm afraid I don't see the point of that change. The commit log entry doesn't explain it either. If .rodata is not discarded then the dangling pointer remains, right? --