Re: [PATCH v2 net] af_unix: Do not call kmemdup() for init_net's sysctl table.

[Kuniyuki Iwashima <kuniyu@amazon.com>]

From:   Eric Dumazet <edumazet@google.com>
Date:   Mon, 27 Jun 2022 20:40:24 +0200
> On Mon, Jun 27, 2022 at 8:30 PM Kuniyuki Iwashima <kuniyu@amazon.com> wrote:
> >
> > From:   Jakub Kicinski <kuba@kernel.org>
> > Date:   Mon, 27 Jun 2022 10:58:59 -0700
> > > On Sun, 26 Jun 2022 11:43:27 -0500 Eric W. Biederman wrote:
> > > > Kuniyuki Iwashima <kuniyu@amazon.com> writes:
> > > >
> > > > > While setting up init_net's sysctl table, we need not duplicate the global
> > > > > table and can use it directly.
> > > >
> > > > Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
> > > >
> > > > I am not quite certain the savings of a single entry table justivies
> > > > the complexity.  But the looks correct.
> > >
> > > Yeah, the commit message is a little sparse. The "why" is not addressed.
> > > Could you add more details to explain the motivation?
> >
> > I was working on a series which converts UDP/TCP hash tables into per-netns
> > ones like AF_UNIX to speed up looking up sockets.  It will consume much
> > memory on a host with thousands of netns, but it can be waste if we do not
> > have its protocol family's sockets.
> 
> For the record, I doubt we will accept such a patch (per net-ns
> TCP/UDP hash tables)

Is it because it's risky?
IIRC, you said we need per netns table for TCP in the future.


> > So, I'm now working on a follow-up series for AF_UNIX per-netns hash table
> > so that we can change the size for a child netns by a sysctl knob:
> >
> >   # sysctl -w net.unix.child_hash_entries=128
> >   # ip net add test  # created with the hash table size 128
> >   # ip net exec test sh
> >   # sysctl net.unix.hash_entries  # read-only
> >   128
> >
> >   (The size for init_net can be changed via a new boot parameter
> >    xhash_entries like uhash_entries/thash_entries.)
> >
> > While implementing that, I found that kmemdup() is called for init_net but
> > TCP/UDP does not (See: ipv4_sysctl_init_net()).  Unlike IPv4, AF_UNIX does
> > not have a huge sysctl table, so it cannot be a problem though, this patch
> > is for consuming less memory and kind of consistency.  The reason I submit
> > this seperately is that it might be better to have a Fixes tag.
> 
> I think that af_unix module can be unloaded.
> 
> Your patch will break the module unload operation.

Thank you!
I had to take of kfree() in unix_sysctl_unregister().