Test Marvell Yukon 2 Ethernet Driver

I've posted a test version of my "yge" driver (Marvell Yukon Gigabit Ethernet, but it also supports some 100Mbps devices) up on OpenSolaris.org. Get it here -- get the file called yge-test.tar.gz.

There is a webrev posted too, if you want to beat your head against the wall.

If you do test this, be aware that many possible PCI ids are not in the installer file. If you want to test more, please do so (at your own risk). If you find that other devices either do work, or do not, please let me know -- send mail to garrett.damore at sun dot com.

Thanks, and enjoy! (I'm hoping to get this driver integrated into SNV 123.)

Comments

Dar said…
I've been using the myk driver for some time now on my MBP (many thanks to Masayuki Murayama for that) and while it's been great to use, it's so much better to have a driver directly integrated into OpenSolaris!

Tried the binaries you posted, and so far so good :)
Unknown said…
Switched over from the myk driver for my 4364 device, hasn't blown up horribly yet.

myk used to go nuts when I was using vnics, hopefully yge won't have any such problems.
eg.
myk: [ID 161980 kern.info] NOTICE: myk0: myk_rx_desc_stat: rx error, status 0x3c0080 length 60
last message repeated 907323 times
Unknown said…
Works fine with a 88E8052 by adding a
yge "pciex11ab,4360"
to driver_aliases.

Unfortunately, it causes a panic with a 88E8001 which does work with the 10.0.3 yukonx driver from marvell.

Thanks for the good work.
Unknown said…
Not surprising. The 88E8001 has some substantially different things about it, and our driver doesn't have the code for it. Only Yukon 2 (not Yukon 1) parts are supported by yge at present.
Unknown said…
Hmmmmm. Spoke a little too soon. Saw this in my logs when the bittorrent client was transmitting alot:

Sep 21 02:12:16 router yge: [ID 451854 kern.warning] WARNING: yge0: TX hang detected!
Sep 21 02:12:17 router mac: [ID 486395 kern.info] NOTICE: yge0 link down
Sep 21 02:12:19 router mac: [ID 435574 kern.info] NOTICE: yge0 link up, 1000 Mbps, full duplex


Also noticed that my SNMP stats for the yge interface seems wrong. It shows something, but not sure exactly what.
Unknown said…
If you upgrade to snv_123 or later, you'll get a greatly improved version of the driver. There were some TX hangs in older versions of the code, but the latest should be much more robust.
R F said…
heh.... it's 124 or later. :) But I grabbed the BFU tarball and tried it. Not getting any errors, but the SNMP stats are still off. I think the issue is that the hw counter is being cleared out, but the system doesn't think it's cleared yet. The e1000g driver has a local copy of the stats that it adds the result to. Wonder if the bsd code is wrong in the same way....
Unknown said…
I need to look more closely at the stats problem. It wouldn't surprise me if we aren't doing the right thing here.

Popular posts from this blog

SP (nanomsg) in Pure Go

An important milestone

GNU grep - A Cautionary Tale About GPLv3