| by admin | No comments


If you don’t take the semaphore, then blam. Apaprently the Suse dev farm have some with dead ethernet’s, a Linux dev has had his laptop screwed, and worryingly there’s a report of something similar happening with a RealTek ethernet card with the same symptoms. Many server motherboards, all vPro-branded desktops and all current Intel motherboards also come with these, so many more people are at risk than just those who have bought discrete Intel Ethernet adapters. Kerminator Seniorius Lurkius Registered: Only 5 confirmed reports of the flaw so far, or something? Canonical has released a kernel update for Ubuntu 8. Mon Sep 22,

Uploader: Vozilkree
Date Added: 24 June 2009
File Size: 39.61 Mb
Operating Systems: Windows NT/2000/XP/2003/2003/7/8/10 MacOS 10/X
Downloads: 44454
Price: Free* [*Free Regsitration Required]

List of Supported Retail Intel® Ethernet Adapters

It will allow additional upstream developers to examine the issue. Maybe it helps in edge cases like booting off a network with a big MTU etc.

In this situation, neither set of kernel patches currently available will fix the issue. These things are all intel based, and just giving a FYI out there for Epc folks. Everything is working fine, but every now and than the network activity stops at any time without any error or anything.

Linux Kernel Driver DataBase: CONFIG_EE: Intel(R) PRO/ PCI-Express Gigabit Ethernet support

Good thing this isn’t AMD people would be all over them claiming Intel would never make such a mistake. Once you’ve tested the upstream kernel, please remove the ‘needs- upstream- testing’ tag. 88086 unlinked ee for the time being. Thu Sep 25, 8: Kerminator Seniorius Lurkius Registered: Mon Nov 17, 3: You are not directly subscribed to this bug’s notifications. Changed in linux Ubuntu: Wed Sep 24, 4: Elsewhere he suggests that the issues are with the and devices, i.


One hopes that the kernel teams of the distros adopt these patches quickly.

Penalver penalvch wrote on Originally posted by stmiller: Likely, because I’ve been bugging them about having a warning page, and some over there are tied much more deeply into the developers than I will ever be. Making things worse is that 10db least for the 100bd chipset-integrated parts, the only tools available to reflash the EEPROM will destroy the functionality, making it so the devices are no longer even enumerated on the PCI Express bus.

It seems like test methods just aren’t keeping up with the technology.

PCI\VEN_&DEV_10BD – DM-2 Gigabit Network… | Device Hunt

More speculation is that it isn’t kernel code at allbut rather these alpha and beta distro snapshots use a buggy pre-release version of X or its driverswhich of course have their own page table entries memory-mapping devices, and that X drivers may be inappropriately mapping and performing MMIO to the EEPROM NVM, destroying its contents. Ars Legatus Legionis et 10bc. The kernel doesn’t seem to perform any sort of range checking or validation of 10bbd memory map requests, which means a bad mapping could include the EEPROM NVM without kernel interference.


Thu Oct 02, 6: I have recently installed Ubuntu Server Thanks for the warning. Intel has also provided patches 1 2 3 that map the EEPROM read-only into the address space, likely eliminating the problem from the root.

Thank you for this post, I was going to do something positively retarded in about two hours.

Intel [HDA Intel], device 0: Waqar Ali Aziz, please be sure to confirm this issue exists with the latest development release of Ubuntu. Tue Sep 23, 3: Tue Sep 23, 6: Suburbs have little in common with the city, so why do we manage them the same way? If the mainline kernel does not fix this bug, please add the tag: What I’d really like to know is why this new driver actually triggers the behavior so much more than previously.