On Wed, Jul 26, 2017 at 01:07:34AM +0100, Ard Biesheuvel wrote: > > > On 26 Jul 2017, at 00:36, Giovanni Cabiddu <giovanni.cabiddu@xxxxxxxxx> wrote: > > > > Hi Ard, > > > >> On Fri, Jul 21, 2017 at 04:42:38PM +0100, Ard Biesheuvel wrote: > >> +static int crypto_scomp_init_tfm(struct crypto_tfm *tfm) > >> +{ > >> + int ret; > >> + > >> + mutex_lock(&scomp_lock); > >> + ret = crypto_scomp_alloc_all_scratches(); > >> + mutex_unlock(&scomp_lock); > >> + > >> + return ret; > >> +} > > If you allocate the scratch buffers at init_tfm, don't you end > > up with a situation where if you allocate two tfms of the same algo > > then you get twice the number of scratches? > > If that is the case, we should implement a reference count > > mechanism. > > Am I missing something? > > > > Isn't the mutex supposed to take care of that? Ignore my previous question. There is already a reference count mechanism that prevents that to happen. Your change sounds correct to me. Regards, -- Giovanni