Re: [PATCH v7 01/22] net/tcp: Prepare tcp_md5sig_pool for TCP-AO

[Steen Hegelund]

Hi Dmitry,

On Thu, 2023-06-15 at 17:44 +0100, Dmitry Safonov wrote:
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the
> content is safe
> 
> Hi Steen,
> 
> On 6/15/23 11:45, Steen Hegelund wrote:
> > Hi Dmitry,
> > 
> > On Thu, 2023-06-15 at 00:09 +0100, Dmitry Safonov wrote:
> [..]
> > > +/**
> > > + * tcp_sigpool_alloc_ahash - allocates pool for ahash requests
> > > + * @alg: name of async hash algorithm
> > > + * @scratch_size: reserve a tcp_sigpool::scratch buffer of this size
> > > + */
> > > +int tcp_sigpool_alloc_ahash(const char *alg, size_t scratch_size)
> > > +{
> > > +       int i, ret;
> > > +
> > > +       /* slow-path */
> > > +       mutex_lock(&cpool_mutex);
> > > +       ret = sigpool_reserve_scratch(scratch_size);
> > > +       if (ret)
> > > +               goto out;
> > > +       for (i = 0; i < cpool_populated; i++) {
> > > +               if (!cpool[i].alg)
> > > +                       continue;
> > > +               if (strcmp(cpool[i].alg, alg))
> > > +                       continue;
> > > +
> > > +               if (kref_read(&cpool[i].kref) > 0)
> > > +                       kref_get(&cpool[i].kref);
> > > +               else
> > > +                       kref_init(&cpool[i].kref);
> > > +               ret = i;
> > > +               goto out;
> > > +       }
> > 
> > Here it looks to me like you will never get to this part of the code since
> > you
> > always end up going to the out label in the previous loop.
> 
> Well, not exactly: this part is looking if the crypto algorithm is
> already in this pool, so that it can increment refcounter rather than
> initialize a new tfm. In case strcmp(cpool[i].alg, alg) fails, this loop
> will never goto out.

Ah, right, you never find any algo and then get out at the end of the list.

> 
> I.e., you issued previously setsockopt()s for TCP-MD5 and TCP-AO with
> HMAC-SHA1, so in this pool there'll be two algorithms: "md5" and
> "hmac(sha1)". Now if you want to use TCP-AO with "cmac(aes128)" or
> "hmac(sha256)", you won't find them in the pool yet.
> 
> > 
> > > +
> > > +       for (i = 0; i < cpool_populated; i++) {
> > > +               if (!cpool[i].alg)
> > > +                       break;
> > > +       }
> > > +       if (i >= CPOOL_SIZE) {
> > > +               ret = -ENOSPC;
> > > +               goto out;
> > > +       }
> > > +
> > > +       ret = __cpool_alloc_ahash(&cpool[i], alg);
> > > +       if (!ret) {
> > > +               ret = i;
> > > +               if (i == cpool_populated)
> > > +                       cpool_populated++;
> > > +       }
> > > +out:
> > > +       mutex_unlock(&cpool_mutex);
> > > +       return ret;
> > > +}
> > > +EXPORT_SYMBOL_GPL(tcp_sigpool_alloc_ahash);
> > > +
> > 
> > ... snip ...
> > 
> > 
> > >  clear_hash:
> > > -       tcp_put_md5sig_pool();
> > > -clear_hash_noput:
> > > +       tcp_sigpool_end(&hp);
> > > +clear_hash_nostart:
> > >         memset(md5_hash, 0, 16);
> > >         return 1;
> > >  }
> Thanks,
>             Dmitry
> 

Reviewed-by: Steen Hegelund <Steen.Hegelund@xxxxxxxxxxxxx>

BR
Steen