News Raptor Lake Motherboards Allegedly Hit With Ethernet Controller Flaw

Toggle sidebar Toggle sidebar
First they skimped on 10Gb/s and replaced them with cheaper 2.5Gb/s.

Now it becomes clear they also skimped on software.

Next, another tech company will ask why sales are in the gutter and why they have to fire thousands of employees.
 
  • Like
Reactions: Kridian and RichardtST
I've been using a 4 port i226-V based box as a pfsense router for months and have never had an issue. Internet connection never drops out (1gbps) and the connection to the box never drops out (2.5gbps). I'm not saying that there isn't a problem at work, but if it was just the controller then I'd have been having issues as well. I can't imagine the implementation is fundamentally different on the effected motherboards and hopefully whatever is going on can be resolved with software.
 
  • Like
Reactions: jp7189 and cyrusfox
Lovely. Thanks. I ran into stability problems on the 12th gen intel board ethernet interfaces. Resolved them eventually, but it was painful. I actively look for non-intel ethernet interfaces now. Lesson learned and reinforced.
 
  • Like
Reactions: bit_user and peterf28
I started using Aquantia AQC 10Gbe lan years ago for windows Linux and Macintosh… and I never looked back. I’ve had many zoom and teams calls and other real time applications with no dropouts.
Too bad intel’s i225-v1 and i226 are buggy.

Another recent buggy intel hardware solution is maple ridge thunderbolt 4… what happened to intel ? Early versions of the controller firmware result in not being able to hotplug early TB3 devices built on alpine ridge (JHL6240). Newer versions of the firmware rectified that issue, but introduced another issue where the controller takes 20 seconds to wake up when a TB device is connected and the pc enters and resumes from sleep. Then the controller enters a non-functional state, is not detected, and the only way to recover is to reboot or cold boot. Nvm36 and nvm38 are susceptible to the problem.
 
I'm using the Intel 1.1.3.28 driver supplied by Asus for a z790 board with an I226-V.

e2fexpress Information EventID 32, and Warning Event ID 27 are both generated any time the driver reloads, such as when you change a setting in the driver properties and hit apply. They are also generated after a system reboot for the same reason, so you may have zero connection issues but will still find those events
I had 212 of both combined after using this machine for a few weeks, and shut down daily and often end up rebooting for installations or registry changes too.

Both contain "The message resource is present but the message was not found in the message table", Intel just hasn't provided a message for them. and at least one of those two eventID's is probably being generated solely because of that missing message.

I often play games that have high UDP throughput and have not had any drop outs or packet loss here, my router is only 1Gbps ethernet however.
 
  • Like
Reactions: bit_user
It was clearly stated that this is a hardware level problem, and is not software related.
Click to expand...

Can you actually keep up before commenting?

The network problems had to be fixed with a software update.

The latest Windows drivers and I225 firmware available for download (driver version 1.0.0.30 or later NVM version 1.38 or later) can detect when the IPG packet drops occur and will automatically reset the link speed to 1Gpbs.
Click to expand...
 
Last edited:
PlaneInTheSky said:
Can you actually keep up before commenting?

The network problems had to be fixed with a software update.
Click to expand...
That is for the 225 the idiots did it again with the newer 226 and there is no patch of any kind yet. Even the 225 firmware you talk about is not a actual fix they just disable your ability to run the port at 2.5g.
I guess you can call it a software fix if you consider disabling the hardware feature a fix.
 
I'm not sure why anyone's going back and forth regarding the i225-V as whatever is going on here with the i226-V is not the same thing. The controller itself does not have some hardware bug causing disconnects which means it's down to a bad batch, implementation and/or bad software.
 
According to some posts this is to be related 2,5Gbit but from what I seen the problem is visible using 1Gbit too?
But I could not see it on 100mbit. Is it possible that the problem manifest itself a bit different for people? Going 1Gbit solves it for some but not for others?
Maybe it is more like a manufacturing/signal processing defect on an analog level that could hurt the ICs to a different degree so some will get a quite ok IC and not even notice this? And that some (most?) ICs might even work fine at 2,5Gbit and that is why they did not discover it at product tests?
Another thing that crossed my mind is, does the device in the other end matter? I did a quick test of this by connecting to a laptop at 1Gbit with enabled internet connection sharing and did not experience drops as when connected to a Turris omnia at 1Gbit.
I also have a slight feeling that some network load is required to experience the drops, of course you won't notice it when not using the network but checking if the LAN connection time have been reset did not reveal any drops when idle. That too makes it feel like a signal processing problem maybe causing high BER (bit error rate) because of poor signal quality and ultimately it forces a speed renegotiation?
 
Back in 2021Q4 i was building an amd htpc based around the 5700g. I bought several asrock and asus boards with the i225v nic, all rev 3. All suffered from network instability as soon as the ethernet cable was plugged in.

Tried all sorts of drivers, attempted a fw update (which failed as the chip was on the most recent fw available). The failure was unusual indeed. When the system became unstable, pressing the reset or holding down the power button failed. System had to be manually unplugged from the wall or switch on the back of PS flipped off.

After 3 or 4 different boards I gave up and went with something based on the rtl8125 nic - an asus tuf gaming B550-plus (wifi). No issues at all.

More recently, microcenter had deal on a 12700k + TUF GAMING Z690-PLUS WIFI D4. This board uses the i225v chip. I cringed when I saw that but figured I'd give it a shot anyway. While bench testing, ethernet functionality was the first thing tested after OS install. No issues at all, even with the latest driver.

So, is it some issue between amd chipset/cpu and the nic??? However, there are other reports of similar instability with the nic and intel cpu's. The conclusion is inconclusive!
 
According to https://www.techpowerup.com/forums/...op-issue-no-fix-available.303854/post-4935294 it seems the other connected device does matter. Also #13 seems to hint a similar thing.
And people on reddit say that forcing 2,5Gbit instead of auto negotiation works View: https://www.reddit.com/r/intel/comments/zpgy7j/i226v_ethernet_controller_has_the_exact_same/?utm_term=41677149142&utm_medium=comment_embed&utm_source=embed&utm_name=462ebd6c-9a46-11ed-95ac-1a7e672bd4eb&utm_content=header

Have a hard time to believe this is the whole truth as I cant even get a stable 1Gbit.
 
  • Like
Reactions: bit_user
Intel probably tested the snot out of the ethernet silicon. Intel only sells that silicon to the vendor, they don't design Asus' motherboards. The end usability is based on the individual OEMs design. Your silicon manufacturer (Intel) does not test the motherboard design of Asus, Gigabyte, MSI, etc... that is the OEM's responsibility to validate their product and design. The I226-V may work perfectly fine in one board, yet fail in another board from the same vendor. It depends on how well the motherboard design engineer did his job. The real question is "ASUS why didn't you validate this board?", or "Gigabyte why didn't you test this?".
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom