netdev - Re: [PATCH net-next] net: generalize skb freeing deferral to per-cpu lists

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CANn89iKqLtyz5H6D1e9Yxd30FTJvbASfNckMAq63UJ+gvauu-Q@mail.gmail.com>
Date:   Fri, 22 Apr 2022 10:25:40 -0700
From:   Eric Dumazet <edumazet@...gle.com>
To:     Jakub Kicinski <kuba@...nel.org>
Cc:     Eric Dumazet <eric.dumazet@...il.com>,
        "David S . Miller" <davem@...emloft.net>,
        Paolo Abeni <pabeni@...hat.com>,
        netdev <netdev@...r.kernel.org>
Subject: Re: [PATCH net-next] net: generalize skb freeing deferral to per-cpu lists

On Fri, Apr 22, 2022 at 9:50 AM Eric Dumazet <edumazet@...gle.com> wrote:
>
> On Fri, Apr 22, 2022 at 9:40 AM Jakub Kicinski <kuba@...nel.org> wrote:

> >
> > If so maybe we can avoid some dirtying and use a single-linked list?
> > No point modifying the cache line of the skb already on the list.
>
> Good idea, I can think about it.
>


My first implementation was using an llist (as current per-socket llist),
but then I needed the count as well, so I converted to standard sk_buff_head

It seems we can hand code to:

     spinlock_t  lock;
     struct sk_buff *skb_head;
     int count;

We also could keep an llist,  and an atomic_t for the count, but that
would require two atomic ops, so no good.