Date: Tue, 25 Sep 2001 17:48:02 -0500 From: "Matthew Rezny" <mrezny@umr.edu> To: "Jim McGrath" <jmcgrath@sitaranetworks.com> Cc: "net@freebsd.org" <net@freebsd.org> Subject: RE: Alteon gigE NIC (if_ti) problems Message-ID: <20010925224800.6DAB237B401@hub.freebsd.org> In-Reply-To: <31269226357BD211979E00A0C9866DAB01BE7963@rios.sitaranetworks.com>
next in thread | previous in thread | raw e-mail | index | archive | help
The Intel NIC flooded the remote end (computer or switch) with trash such that the device was unable to process anything as long as the link was up. This occured as long as the computer was on, before the driver loaded or the OS booted, so it wasn't a driver problem. I tried both the wx and gx driver and neither seemed to initialize it in any way that allowed it to work properly. Traffic could pass through the link but the remote end couldn't do much since it was flooded with junk that appeared to be flow-control stuff. With this Alteon card, the remote end is just fine and all traffic passes through. The disconnect appears on the computer. The link light on the switch never goes off when the computer says its link is continously going down and up. I suspect that when it thinks the link drops, all trafffic gets delayed until some timeout period after the link is back, which is why I'm seeing slowdowns. On Tue, 25 Sep 2001 18:36:15 -0400, Jim McGrath wrote: >What kind of problems were you having with the Intel GigE NIC? I've got it >running in an Intel box, the only recent problems have been of my own >making. I found a bug in the code for the copper version of the NIC. Were >you running a fiber or copper NIC? > >I'm not as familiar with the ti driver, but it sounds like a watchdog timer >problem. That will cause a reset of the NIC. > >Jim > >-----Original Message----- >From: Matthew Rezny [mailto:mrezny@umr.edu] >Sent: Tuesday, September 25, 2001 2:17 PM >To: net@freebsd.org >Subject: Alteon gigE NIC (if_ti) problems > > >I recent gave up truing to make a Intel gigE NIC work in my Alpha so I >got an Alteon board. It works without the severe problem of the Intel >card. However, its acting a little weird. I see messages saying the >link went down and up quite often. I've tried both 4.3 and 4.4 and have >the same thing. I increased NMBCLUSTERS significantly and still see the >same thing. I made the switch from 4.3 to 4.4 at the same time I >incrased NMBCLUSTERS from 8192 to 32767 so I can't say which, but one >(or both) of those changes increased the rate at which I see the link >status change message. Before I saw it every once in a while >reguardless of whether there was heavy traffic or not. Now it shows up >every few seconds when there is any traffic at all. Anyone have any >ideas what's going on? > >BTW, I would include my dmesg but its full of "ti0: link down" and >"ti0: gigabit link up" so here's there pertinent info. >Alpha 21164PC 533 on PC164SX >Alteon gigE card in 64bit PCI slot with IRQ 11 which is not shared >If any other info would help, just ask and I'll get it. > > > >To Unsubscribe: send mail to majordomo@FreeBSD.org >with "unsubscribe freebsd-net" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010925224800.6DAB237B401>