followup to summary - force a network card to 100

From: Penny Jaye Deeney <>
Date: Thu Feb 15 2001 - 12:49:57 EST
Question and Summary follow this message.

I've appended these lines to /etc/system:

> * hme driver modes
> *
> set hme:hme_adv_autoneg_cap=0
> set hme:hme_adv_100T4_cap=0
> set hme:hme_adv_100fdx_cap=1
> set hme:hme_adv_100hdx_cap=0
> set hme:hme_adv_10fdx_cap=0
> set hme:hme_adv_10hdx_cap=0

but when I reboot, the network is not available.  To verify that the
problem is not with that particular nic card, I switched the ethernet
cable to hme0, and renamed hostname.hme1* (configured for 3 ip addresses)
to hostname.hme0*.

After boot, I used ndd and got this:

% ndd -get /dev/hme instance
% ndd -set /dev/hme instance 0
% ndd -get /dev/hme instance
% ndd -get /dev/hme link_status
% ndd -get /dev/hme link_speed
% ndd -get /dev/hme link_mode

So, the card is configured for 100 full.  However, it's not connected.  I
can't ping from or to this machine using two other machines on the same

I am thinking it's a problem with either the ethernet cable or somewhere
else on the network.  However, those are not under our control.

What do the rest of you think?  I'd be happy to summarize.

---------- Forwarded message ----------
Date: Fri, 9 Feb 2001 15:52:12  -0500 (EST)
From: Penny Jaye Deeney <>
Subject: summary - force a network card to 100

Many many thanks to
Andrew J Caines
Damon LaCaille
Greg Ulyatt
Mike D. Kail
Price, Michael
Rytting, Chad
Brendan Choi
Doug Otto
David Foster
Stefan Molnar
Thomas Jones
Matthew Stier
Frederick Hall
Madison, Ryan

Mostly for not chiding me that the instructions were in the FAQ (my bad)

Many commented that sun hardware does not autonegotiate with some network
hardware, especially cisco.  I do not know what network hardware we have,
and I do not have access to it.  Our responsibility stops at the ethernet
cable.  I do not think my network folks were lying about the 100 Mbps
capability of the switch, though, because we have another box running at
100 half that connects to that same switch.  It should be 100 full, but
that's next.

The settings in my /etc/system file were almost correct.  I need to change
the autonet_cap and 10hdx_cap settings to 0 and then reboot.  This should

Many also told me about the ndd command.  I was able to use this to check
out the status of my network cards (There are 3 in this machine - hme0, 1
and 2.)

However, when I used ndd to set hme1 adv_100fdx_cap to 1 and all the
others (including adv_autoneg_cap) to 0, I lost network connectivity.  I
ran up to the console, since we hadn't scheduled down time for this
afternoon.  Setting adv_autoneg_cap back to 1 did not fix the connectivity
issue, so I rebooted.

I think that Monday, before I reboot, I will edit /etc/system.  If this
doesn't work, then I will need to check the cable and complain to our
network group.

Thanks everyone!

Original question:
> Sun E3500 and UltraSparc 2, both running Solaris 7.
> Both machines have ethernet cards set to auto-negotiate, but they keep
> coming up 10Mbps half duplex.  We have been assured by our Network group
> that the network is capable of 100 full, and another Sun, an Ultra 5
> running Solaris 8, negotiates at 100 half.
> We need these two machines to run at 100Mbps.  Full would be best for
> three, but half is do-able.
> I looked through the OpenBoot documentation, but couldn't find a way of
> forcing this at the ok prompt.
> We found a web page that suggested appending these lines to the end of
> /etc/system:
> * hme driver modes
> *
> set hme:hme_adv_autoneg_cap=1
> set hme:hme_adv_100T4_cap=0
> set hme:hme_adv_100fdx_cap=1
> set hme:hme_adv_100hdx_cap=0
> set hme:hme_adv_10fdx_cap=0
> set hme:hme_adv_10hdx_cap=1
> But upon reboot we got transceiver errors.  Our hardware guy is
> his sun guy about this, but it's taken a couple of days and we need a
> We have a scheduled reboot on Monday, and we'd have time then to try
> something.
Received on Thu Feb 15 17:49:57 2001

This archive was generated by hypermail 2.1.8 : Wed Mar 23 2016 - 17:46:15 EDT