On Fri, Jul 24, 2020 at 01:22:58PM +0300, Mike Rapoport wrote: > On Fri, Jul 24, 2020 at 08:05:48AM +0300, Jarkko Sakkinen wrote: > > Add lock_modules() and unlock_modules() wrappers for acquiring module_mutex > > in order to remove the compile time dependency to it. > > > > Cc: linux-mm@xxxxxxxxx > > Cc: Andi Kleen <ak@xxxxxxxxxxxxxxx> > > Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx> > > Suggested-by: Masami Hiramatsu <mhiramat@xxxxxxxxxx> > > Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx> > > --- > > include/linux/module.h | 18 ++++++++++++++++++ > > kernel/kprobes.c | 4 ++-- > > kernel/trace/trace_kprobe.c | 4 ++-- > > 3 files changed, 22 insertions(+), 4 deletions(-) > > Any reason to convert only kprobes to the new API and leave others with > opencoded implementation? Not anything particular. Lets see: $ git --no-pager grep "mutex_lock(&module_mutex)" arch/powerpc/platforms/powernv/pci-cxl.c: mutex_lock(&module_mutex); drivers/gpu/drm/drm_fb_helper.c: mutex_lock(&module_mutex); include/linux/module.h: mutex_lock(&module_mutex); kernel/livepatch/core.c: mutex_lock(&module_mutex); kernel/livepatch/core.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); kernel/module.c: mutex_lock(&module_mutex); I could refine this commit to patch these sites. Or should I split it into multiple? /Jarkko
