On Mon, 4 Sept 2023 at 20:46, Andrey Konovalov <andreyknvl@xxxxxxxxx> wrote: > > On Wed, Aug 30, 2023 at 11:33 AM Marco Elver <elver@xxxxxxxxxx> wrote: > > > > If someone doesn't use stack_depot_evict(), and the refcount eventually > > overflows, it'll do a WARN (per refcount_warn_saturate()). > > > > I think the interface needs to be different: > > > > stack_depot_get(): increments refcount (could be inline if just > > wrapper around refcount_inc()) > > > > stack_depot_put(): what stack_depot_evict() currently does > > > > Then it's clear that if someone uses either stack_depot_get() or _put() > > that these need to be balanced. Not using either will result in the old > > behaviour of never evicting an entry. > > So you mean the exported interface needs to be different? And the > users will need to call both stack_depot_save+stack_depot_get for > saving? Hm, this seems odd. > > WDYT about adding a new flavor of stack_depot_save called > stack_depot_save_get that would increment the refcount? And renaming > stack_depot_evict to stack_depot_put. If there are no other uses of stack_depot_get(), which seems likely, just stack_depot_save_get() seems ok. > I'm not sure though if the overflow is actually an issue. Hitting that > would require calling stack_depot_save INT_MAX times. With a long-running kernel it's possible.
