netdev - Re: [net-next PATCH 3/7] net: bulk alloc and reuse of SKBs in NAPI context

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]

Date:	Mon, 14 Mar 2016 07:55:37 +0100
From:	Jesper Dangaard Brouer <brouer@...hat.com>
To:	Rana Shahout <rana.shahot@...il.com>
Cc:	netdev@...r.kernel.org, "David S. Miller" <davem@...emloft.net>,
	eugenia@...lanox.com, Alexander Duyck <alexander.duyck@...il.com>,
	alexei.starovoitov@...il.com, saeedm@...lanox.com,
	gerlitz.or@...il.com, brouer@...hat.com
Subject: Re: [net-next PATCH 3/7] net: bulk alloc and reuse of SKBs in NAPI
 context

On Sun, 13 Mar 2016 16:06:17 +0200
Rana Shahout <rana.shahot@...il.com> wrote:

> On Fri, Mar 4, 2016 at 3:01 PM, Jesper Dangaard Brouer
> <brouer@...hat.com> wrote:
> 
> >  /* build_skb() is wrapper over __build_skb(), that specifically
> >   * takes care of skb->head and skb->pfmemalloc
> >   * This means that if @frag_size is not zero, then @data must be backed
> > @@ -490,8 +500,8 @@ struct sk_buff *__napi_alloc_skb(struct napi_struct *napi, unsigned int len,
> >
> >         len += NET_SKB_PAD + NET_IP_ALIGN;
> >
> > -       if ((len > SKB_WITH_OVERHEAD(PAGE_SIZE)) ||
> > -           (gfp_mask & (__GFP_DIRECT_RECLAIM | GFP_DMA))) {
> > +       if (unlikely((len > SKB_WITH_OVERHEAD(PAGE_SIZE)) ||
> > +                    (gfp_mask & (__GFP_DIRECT_RECLAIM | GFP_DMA)))) {  
> 
> Why unlikely? I know it is better for the common case where most
> likely linear SKBs are << SKB_WITH_OVERHEAD(PAGE_SIZE)), but what
> about the case of Hardware LRO,  where linear SKB is likely to be >>
> SKB_WITH_OVERHEAD(PAGE_SIZE)).

You said it yourself, this is better for the common case.  With
unlikely() I'm asking the compiler to layout the code for the common
case.  This helps the CPU instruction cache prefetcher.

-- 
Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  Author of http://www.iptv-analyzer.org
  LinkedIn: http://www.linkedin.com/in/brouer