Comment by EvanAnderson | Hacker News Reader

Comment by EvanAnderson | Hacker News Reader

Parent Full thread

The Intel PRO/100 NICs using the 82557, 82558, and 82559 controllers were a joy to work with on Windows. The NICs performed well and there was some backwards compatibility in the drivers. I have a vague memory of a lockup bug in some Linux driver versions that would knock machines off the network but it was eventually fixed.

I really enjoyed Intel's foray into networking in the late 90s. Their Ethernet hubs and switches were economical and reliable. It really made no sense for them to be in that market, and they ultimately ditched it, but I liked the simple and inexpensive gear.

The first chassis switch I ever used was the XLNT Millennium that Intel acquired and rebranded as the Intel "6000 Series". At a time when gigabit Ethernet was still exotic getting to play with one in a lab was a treat.

Aside: Videos like this leave me with this "Gee, I could've done that" feeling. It would probably be fun to divest myself of the old junk I have sitting around by way or making videos about the stuff before I liquidate it.

In this era 3Com was still the gold standard, but intel eventually dominated the market because they integrated 100mbit and then gigabit on the PCH relegating any margin in the expansion card market to servers. As someone that has worked on a lot of Intel drivers I always find Intel being put on a pedestal to be a little amusing, they have a tendency to make a lot of gaffes in silicon and sometimes these reappear in followup designs. It's certainly not bad, but I would place Intel networking about mid-tier historically and today.

Realtek then dominated by being cheaper and simpler than Intel; they had the majority of the NIC marketshare (70%+) in the late 90s and early 2000s.

Most realtek is F tier, they basically repeat the bad part of this NE2000 saga, and the simplicity comes at the cost of CPU cycles. In that sense the Intel PCH controllers were great.

"Good enough for most people and cheap" is what Realtek do.

I had some EtherExpress cards, bought new circa 1993. They had a lifetime warranty which Intel honoured sometime around 2005.

Rock-solid drivers - can’t think of a single platform they didn’t run on.

> I really enjoyed Intel's foray into networking in the late 90s. Their Ethernet hubs and switches were economical and reliable. It really made no sense for them to be in that market, and they ultimately ditched it, but I liked the simple and inexpensive gear.

Intel was always looking for new markets to enter. It's definitely not unreasonable to build NICs and things for NICs to talk to.

The bug on Linux still exists today on this family (e100, e1000, e1000e). Maybe the root cause is different, but the solution is still to disable TCP segmentation offload and I suppose this is a firmware bug. At some point, Mellanox took over the "top" card segment for servers.

I've assumed it worked fine on Linux but your comment brought me to this after a quick search: https://bugzilla.kernel.org/show_bug.cgi?id=118721. I've not seen any major issues on Linux e1000e defaults but never pushed it harder than laptop usage. I've occasionally tried to make em(4) TSO work on FreeBSD (it works fine on igb(4)) but gave up due to stalls. Linux has GSO (maybe invented partially due to those issues?) so the benefit is a lot smaller there.

There's a comical change where Intel downclocked DMA to fix some issue and it makes TSO max out at ~700mbps: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin... but they later disable TSO there with https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin...

Those Intel cards were the number 1 choice on openbsd back then too.

Reply on news.ycombinator.com