Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 8 May 2002 10:46:06 -0500
From:      Damon Anton Permezel <dap@damon.com>
To:        Jamie Heckford <jamie@tridentmicrosystems.co.uk>
Cc:        freebsd-hackers@FreeBSD.ORG, jamie@trident-uk.co.uk
Subject:   Re: Broadcom BCM5701 GigE Ethernet problems??
Message-ID:  <20020508104606.F53377@damon.com>
In-Reply-To: <20020508144540.A33368@mufuf.trident-uk.co.uk>; from jamie@tridentmicrosystems.co.uk on Wed, May 08, 2002 at 02:45:40PM %2B0100
References:  <20020508144540.A33368@mufuf.trident-uk.co.uk>

next in thread | previous in thread | raw e-mail | index | archive | help
Hi.  I just yesterday pulled out one of these because I figured it must
be hardwarily broken, given that the bge driver works so well for the
embedded BCM5700 on the motherboard of another box.

My experience was that the link would come up on the switch as 1000baseTX,
but the ifconfig would report the device as 100baseTX<full-duplex>.

If I "ifconfig bge0 down media 100baseTX mediaopt full-duplex up"
the switch would evidence link down/link up<100baseTX> and a few packets
could be exchanged.

Doing pings, then, I would see <echo req/ echo reply> on the target side,
but many missed <echo replies>.... RX was not working.  Not just an
issue of missing interrupts, as the replies were lost (missing ping sequence
numbers).
"netstat -s" would indicate UDP packets with bad chacksums received.
TCP also.  Not the case for ICMP or IP-header: no checksum error stats recorded.
Essentially unusable though.

If I tried to "ifconfig bge0 down media 1000baseTX" I would get an error
indicating that 1000baseTX was not a suitable option.

Booting with "-v" I would see a whole slew of PHY output indicating
many duplicate instances of 100baseTX phys being found, but no 1000baseTX
that I remember seeing.  Sorry, can't get to those logs right now.

This is what it looks like on the BCM5700:
	miibus1: <MII bus> on bge0
	brgphy0: <BCM5401 10/100/1000baseTX PHY> on miibus1
	brgphy0:  10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseTX, 1000baseTX-FDX, auto

I remember there were brgphyN for some N at least <= 7, some babble
about some 802.XY where Y was 'q' or 'u', perhaps, and I do not
recall seeing "1000baseTX".

If I ifconfig'd down/up too many times (like: two), the system froze.

I really needed some damn interface running so I could CVSUP the box.
In frustration, I almost installed loonix on it.... thats how desperate one
can get these days, what with rumours of FreeBSD's demise and all, but
fortunately, reason prevailed.

If there is reason to suspect that the hardware is not toasted,
I hereby offer my assistance in debugging, testing, or even fixing the
driver.  All I would need in the latter case is the relevant specs and
errata.

Cheers,
Damon.

On Wed, May 08, 2002 at 02:45:40PM +0100, Jamie Heckford wrote:
> Hi,
> 
> Im having strange problems with a Compaq DL360 rackmount machine, which
> has 2x Broadcom BCM5701 Gigabit Ethernet Adapters installed.
> 
> FreeBSD (4.5-R) recognises the cards, but is displaying odd behaviour.
> 
> For example, on boot it correctly assigns the card the IP address etc, and
> the link light is on on both the switch port and NIC. However, as soon as 
> an application tries to use the card (In the case of the bootup sequence 
> ntpdate) the link light on the NIC turns off - but remains lit on the 
> switch port! Attempting to put any sort of traffic across the link then
> fails. 
> 
> Pinging localhost and the cards IP address from the machine works
> fine but not from anywhere else on the network.
> 
> The switch its connected to is a Cisco Cat. 2924. I've tried forcing the
> card and the switch port to 100Mbit FDX, and setting both sides to auto-neg
> but have had no luck.
> 
> Would anyone be able to let me know if there are any known issues with
> the driver/card?
> 
> Thanks in advance,
> 
> -- 
> Jamie Heckford
> Network Manager
> Trident Microsystems Ltd
> Tel:  01737 780790  Fax:  01737 771908
> http://www.tridentmicrosystems.co.uk
> 
> ***********
> 
> This e-mail message contains confidential information for 
> the above addressee only. Any opinion or views contained 
> in this e-mail message are those of the sender and do not 
> necessarily represent those of the Company. Unless otherwise 
> stated this e-mail message is not intended to be contractually 
> binding.  E-mail transmission cannot be guaranteed to be 
> secure or error-free and the sender therefore does not accept 
> liability for any errors or omissions in the contents of this 
> message.
> 
> Trident Microsystems Ltd Group of Companies
> Perrywood Business Park, Honeycrock Lane, Salfords, Redhill, 
> Surrey, RH1 5JQ
> 
> Tel:  (44) (0) 1737 780790   Fax:  (44) (0) 1737 771908
> 
> Registered office:  Abacus House, Bone Lane, Newbury, 
> Berkshire, RG14 5SF
> 
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-hackers" in the body of the message

-- 
--
Damon Permezel
dap@damon.com


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020508104606.F53377>