Date: Wed, 16 Aug 2006 19:17:59 -0400 (EDT) From: "Brian A. Seklecki" <lavalamp@spiritual-machines.org> To: Gleb Smirnoff <glebius@FreeBSD.org> Cc: freebsd-current@freebsd.org, netops@collaborativefusion.com Subject: Re: Dell PowerEdge 850 bge(4) RELENG_6 (WAS: Re: bge(4) problem) Message-ID: <20060816191152.J69548@arbitor.digitalfreaks.org> In-Reply-To: <20060810160126.E55918@arbitor.digitalfreaks.org> References: <43767.150.101.159.26.1140420612.squirrel@mailbox.TU-Berlin.DE> <20060720104238.L8726@arbitor.digitalfreaks.org> <20060807100622.GY96644@cell.sick.ru> <20060810160126.E55918@arbitor.digitalfreaks.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 10 Aug 2006, Brian A. Seklecki wrote:
>
> I will experiment with it tomorrow.
Sorry about the delay.
The problem persists with a fresh 6.1-STABLE kernel (rev 1.91.2.16
2006/08/10).
When I force the duplex on both sides, I still have to manually bounce the
interface state post-boot. I still get "bge1: 2 link states coalesced".
It seems to perform well once that happens.
If I leave it in auto-media, it comes up at 100/half duplex, which leads
to unusable performance. 100/half duplex is technically impossible
anyway.
I doubt the cause of this problem is the driver at this point. It can't
be the switch. This Broadcom chipset must just suck.
I will just throw a dual-port em(4) in this machine instead. I will make
sure to let our Dell sales rep know we're not interested in purchasing any
more systems with Broadcom chips onboard.
The PE 1950/2950 better not have this problem.
~BAS
>
> Meanwhile it should be pointed out that the new Dell 9th generation
> PE{1,2}950 uses the 5708 integrated dual port.
>
> However the Addon-NIC option is still the dual port 5721.
>
> ~BAS
>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060816191152.J69548>