On Wed, 29 Mar 2017, Michal Hocko wrote:
> On Wed 29-03-17 11:32:02, Maninder Singh wrote:
> > This patch checks if any module which is going to be unloaded
> > is doing vmalloc memory leak or not.
>
> Hmm, how can you track _all_ vmalloc allocations done on behalf of the
> module? It is quite some time since I've checked kernel/module.c but
> from my vague understading your check is basically only about statically
> vmalloced areas by module loader. Is that correct? If yes then is this
> actually useful? Were there any bugs in the loader code recently? What
> led you to prepare this patch? All this should be part of the changelog!
Moreover, I don't understand one thing:
> > Logs:-
> > [ 129.336368] Module [test_module] is getting unloaded before doing vfree
ok, but...
> > +static void check_memory_leak(struct module *mod)
> > +{
> > + struct vmap_area *va;
> > +
> > + rcu_read_lock();
> > + list_for_each_entry_rcu(va, &vmap_area_list, list) {
> > + if (!(va->flags & VM_VM_AREA))
> > + continue;
> > + if ((mod->core_layout.base < va->vm->caller) &&
> > + (mod->core_layout.base + mod->core_layout.size) >
> > va->vm->caller) {
> > + pr_err("Module [%s] is getting unloaded before doing
> > vfree\n", mod->name);
> > + pr_err("Memory still allocated: addr:0x%lx - 0x%lx,
> > pages %u\n",
> > + va->va_start, va->va_end, va->vm->nr_pages);
> > + pr_err("Allocating function %pS\n", va->vm->caller);
> > + }
> > +
> > + }
> > + rcu_read_unlock();
> > +}
> > +
> > /* Free a module, remove from lists, etc. */
> > static void free_module(struct module *mod)
> > {
> > + check_memory_leak(mod);
> > +
Of course, vfree() has not been called yet. It is the beginning of
free_module(). vfree() is one of the last things you need to do. See
module_memfree(). If I am not missing something, you get pr_err()
everytime a module is unloaded.
Regards,
Miroslav
