Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 30 Mar 2011 09:58:11 +0400
From:      Andrey Zonov <andrey@zonov.org>
To:        Jack Vogel <jfvogel@gmail.com>
Cc:        freebsd-net@freebsd.org
Subject:   Re: igb(4) won't start with "igb0: Could not setup receive structures"
Message-ID:  <4D92C673.2080107@zonov.org>
In-Reply-To: <AANLkTinZ=Js_pqwtAWX8yqPw==3Fm-7FA50MHDuz0Y86@mail.gmail.com>
References:  <4D923931.2070606@zonov.org> <AANLkTinZ=Js_pqwtAWX8yqPw==3Fm-7FA50MHDuz0Y86@mail.gmail.com>

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

9k jumbos is ubiquitous now. I believe we use 9k jumbos for last four 
years or more.
You've got unworkable system after upgrade from 8.1 to 8.2 and 
documentation doesn't help here. It isn't a good way is made driver (or 
system) which doesn't work with jumbo by default.

My point is if you're using machine with 8 CPUs than 
maxusers/clusters/9k mbufs should have been increased by system, because 
on this machine minimum 2Gb memory is available.

-- 
Andrey Zonov


30.03.2011 1:55, Jack Vogel ?????:
> Our validation group has a default postinstall process, every 
> installed system gets those changes,
> and these mbuf pool sizes are in that set of changes. While I'm not 
> opposed to system default settings
> changing its usually necessary to have local sys changes anyway, after 
> all you don't get 9K jumbos
> without manually specifying them as well :)
>
> Regards,
>
> Jack
>
>
> On Tue, Mar 29, 2011 at 12:55 PM, Andrey Zonov <andrey@zonov.org 
> <mailto:andrey@zonov.org>> wrote:
>
>     Hi,
>
>     New igb driver (and I think em too) is required too much 9k mbufs
>     when it's been configured with mtu = 9000. On machine with 8 CPUs,
>     driver is required 8192 9k mbufs, but by default there is only
>     6400 and network won't start. In previous versions for big mtu it
>     was used 4k mbufs, by default there is 12800 and all worked fine.
>
>     Maybe it's time to think about increasing default
>     kern.maxusers/kern.ipc.nmbclusters? or use mp_ncpus for
>     calculation these values? or just increase amount of
>     mbuf_cluster/mbuf_jumbo_page/mbuf_jumbo_9k from that driver...
>
>     I just want igb to work out-of-the-box.
>
>     -- 
>     Andrey Zonov
>
>     _______________________________________________
>     freebsd-net@freebsd.org <mailto:freebsd-net@freebsd.org> mailing list
>     http://lists.freebsd.org/mailman/listinfo/freebsd-net
>     To unsubscribe, send any mail to
>     "freebsd-net-unsubscribe@freebsd.org
>     <mailto:freebsd-net-unsubscribe@freebsd.org>"
>
>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4D92C673.2080107>