Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 30 May 2007 17:26:02 +0200
From:      "=?ISO-8859-1?Q?Ermal_Lu=E7i?=" <ermal.luci@gmail.com>
To:        "Alexander Motin" <mav@freebsd.org>
Cc:        freebsd-net@freebsd.org, freebsd-pf@freebsd.org
Subject:   Re: [PATCH] ng_pf and l7 filtering possibility with PF
Message-ID:  <9a542da30705300826g30e281bal6a68d5e17902fc21@mail.gmail.com>
In-Reply-To: <465D5D50.2020009@freebsd.org>
References:  <1180481012.00748860.1180470003@10.7.7.3> <465D5D50.2020009@freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Alexander,

the only reason i made them available only for debugging is cause of
int32_t types of those counter  and these could overflow easily on
busy environments.
For 64bit counters on 32bit archs you need atomic operations and i
don't know how much overhead it will be!?(correct me if i am wrong)


On 5/30/07, Alexander Motin <mav@freebsd.org> wrote:
> Hi.
>
> Ermal Lu=E7i wrote:
> > 4- The node has these messages:
> > #ifdef NG_PF_DEBUG
> >       NGM_PF_GET_STATS,       (number of packets in/out)
> >       NGM_PF_CLR_STATS,
> >       NGM_PF_GETCLR_STATS,
> > #endif
>
> What for comment them out? To save CPU on several addition operations?
> This stats could be used not only by developers but also by users and
> they may not be to rebuild kernel module to get that stats.
>
> --
> Alexander Motin
>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9a542da30705300826g30e281bal6a68d5e17902fc21>