What is Digg?58 Comments
- Nephersir7, on 10/03/2008, -4/+36Intel rocks
- symx, on 10/03/2008, -0/+24it's good too see they have taken responsibility and fixed the issue . Makes you trust linux even more.
- yevkasem, on 10/03/2008, -0/+21so mistakes happen. with linux they get fixed a lot quicker than proprietary operating systems.
- ThatsUnpossible, on 10/03/2008, -3/+24What does Intel's bug have to do with Linux, and why is Linux getting credit for fixing anything when it was someone at Intel?
Confused. - recruz, on 10/03/2008, -2/+21Please write in complete sentences.
- billessig, on 10/03/2008, -0/+16The hardware isn't broken. The data in the NVRAM in the hardware got corrupted.
- inactive, on 10/03/2008, -11/+25Linux is better than windows
- jjmac, on 10/03/2008, -0/+13Note: Intel did not fix a bug with Linux they fixed a bug in their network card. Description was a bit misleading / confusing.
- merreborn, on 10/03/2008, -0/+13Dude, install adblock and take your lithium. Everything's gonna be okay.
- inactive, on 10/03/2008, -0/+13What are you talking about? A large amount of Fortune 500 companies use Linux almost exclusively.
- ileftfark, on 10/03/2008, -0/+11The newest Linux kernel contained code that could corrupt the eeprom in the networking chip, rendering it useless (bricked). The Linux community, headed by Linus Torvalds, spoke out and said that Intel's product was fault- no hardware should able to be wiped out by standard kernel coding. So it was an Intel mistake, pointed out by Linux devs, and since, the Linux kernel has applied a patch to prevent this from happening, and Intel has promised to provide a fix for those who have already had their nvram wiped out.
I pretty much just paraphrased the story which it seems you have also read, but that is the gist of it: Intel problem, Linux devs point it out, Linux devs provide patch, Intel devs promise repairs. Fin. - Mutiny32, on 10/02/2008, -0/+10"Later we will post patches to help users who have had this
problem restore their eeprom from either a saved image from
ethtool -e or from another identical system." - Intel
I wonder if the eeprom stored the hardware's MAC address. If so, how does a restored backup from an identical system not make the problem worse? - thcobbs, on 10/03/2008, -0/+10If it wasn't for the openness of linux, the problem like would never have been root-caused and intel would have happily sold more cards/motherboards to replaced fried EEProms.
- wigren, on 10/03/2008, -1/+11I think he means, and I agree, that at least they were quick with a fix.
- Hello1024, on 10/03/2008, -0/+9Lock bits are pretty standard for anything that could permanently damage hardware. When operating systems (such as the linux kernel) go "wrong" they can frequently read and write random data to random addresses and devices before the fault is detected and a blue screen/kernel panic is shown. Using a lock bit massively reduces the chances of hardware damage should the software go horrendously wrong.
One example of a case where this could happen is code that calculates the IO port address (such as COM1 COM2 COM3 etc.). If someone were to try to open a non-existant COM999 and there was no check to see it really existed then it's possible the code that should be accessing the COM port will actually be accessing an entirely random bit of IO address space, that could, for example, be writing to the ethernet cards NVRAM.
The above is a very simple example, but illustrates how bugs like this can happen. - DamnMan, on 10/03/2008, -2/+10No, Intel is just bad at making network chipsets.
- andycr512, on 10/04/2008, -0/+7If you click the headline, it takes you to an article. Once you're at the article, you read it. After you have read it, if you still have questions, you post here. You skipped a few steps.
- chewbie, on 10/03/2008, -0/+7intel's fault in the first place! Read Hello1024 's comment please!
- armo, on 10/03/2008, -0/+7I disagree completely. It was Intels failed QA that caused this problem in the first place. Well done to them for getting a patch out fairly quickly but with proper testing before they submitted the driver and a better hardware design this would never have been a problem.
- MattBD, on 10/03/2008, -1/+8The thing is, you would never hear about it if this happened in OS X or Windows (as long as it got caught in the first place) as they are proprietary OS's. This has now been fixed, so if you subsequently use a Linux distro that runs this kernel you won't have a problem anyway.
- init100, on 10/04/2008, -0/+6"There is a reason we have WHQL on Windows."
WHQL is simply a way for Microsoft to make more money.
"Wanna have that WHQL certification on your driver? You need to pay us a lot of money first." - inactive, on 10/03/2008, -1/+7Um wtf are u talking about?
- thcobbs, on 10/03/2008, -0/+6Yes, it does store the mac address... and if you know what you are doing, you can alter it. That is if you have the right tools and/or your manufacturer didn't lock down the eeprom.
The problem is not the mac address... the problem is all the other settings that are stored in that eeprom. - linuxinsidev2, on 10/04/2008, -0/+6as Linus said, the lock bit should be set on the eeprom rather than being writable all the time, this coupled with something in the kernel that inadvertently tries to write to that memory location.
Whether it be a fault of the programmer or not, it shouldn't be possible to overwrite the NVRAM in a running system after the modules have been loaded. - Spire3660, on 10/03/2008, -0/+6Im much better now. How is adblock going to stop me from having to see an intermediary page before the actual page loads?
- mrsteveman1, on 10/03/2008, -0/+5Is Google a large enterprise? You might want to notify them of their huge mistake, they've been walking down this horrible linux-using path for their important stuff since the beginning.
- brisbin33, on 10/03/2008, -0/+4i'm not sure, but the way i read it; the bug allowed a bit of code to fry some firmware. the fix prevents that from happening. so the _retroactive_ fix would be to restore the firmware from an [un-fried] saved image or identical system back to an un-fried state. then all is well.
that's just my superficial take, i could be way off. - oobuntu, on 10/04/2008, -0/+4FYI the article is so-called because he did a previous article on this subject called "when linux goes bad".
however it should be remembered that this was alpha release of a distro using a pre-release kernel. i guess quite a few linux fans run alphas under VMs and maybe try the betas on their PCs natively to check hardware compatibility.
i expect stuff to break when running alpha/beta software, probably not this badly though... - MtheoryX, on 10/03/2008, -3/+7You didn't read the article, did you?
1.) The bug was in the linux kernel, and affected Ethernet firmware in systems with intel chips. (Thus the Linux/Intel connection).
2.) Torvalds urged Intel to release a fix (Another connection)
3.) A patch was released by a Linux kernel developer and Intel engineer (another connection)
It's not Linux getting credit...it's Linux getting a large company to fix a problem.
Clear as mud? - Giga, on 10/03/2008, -0/+4***** happens. It is good to see that they put the effort in to rectify the problem rather than deny it, unlike some other companies out there...
- TheWindBlows, on 10/04/2008, -0/+4Yeah i know my post was kind of mean but at least it was resolved sadly this bad driver bug puts a tarnish on linux in general oh well fortunately it'll be forgotten soon enough.
- ileftfark, on 10/03/2008, -0/+3Not that I'm aware of- it was a conflict with the Linux kernel.
- Gavagai80, on 10/04/2008, -0/+3Just you.
- LinuxGalore, on 10/05/2008, -0/+3One small problem, Intel wrote the buggy code not the Linux community, Intel also wrote the patch too. Also no one raised the question "why was the Intel hardware so poorly designed that it would flush its firmware at the drop of a hat".
- mrsteveman1, on 10/04/2008, -0/+3Special servers only then huh? What makes them special? They're servers, i see no reason (and you gave none) why Linux is somehow not usable on rack servers, blade servers, smb servers, etc.
Either way, Redhat and Novell would love to know about the apparently lack of a market they have both been operating in, they must have lost millions on these "normal servers". - mrsteveman1, on 10/03/2008, -0/+3So has this problem also happened on non-linux systems?
- jmichaelx, on 10/04/2008, -1/+3I could not agree more. Furthermore, he sometimes has no clue what he is talking about.
- eruanno, on 10/04/2008, -2/+3This guy always writes the worst articles. Everything is buried in overly verbose and speculative paragraphs.
- enantiodromia, on 10/07/2008, -0/+1lol @ special servers
google is FAMOUS for using the cheapest servers possible, not the best possible, and it totally changed how other companies built their enterprise - clickwir, on 10/03/2008, -1/+2No *****? Damnit, that's why I can't get the damn ethernet link on my work pc with an Intel board. Crap.
I wonder if I can load windows xp and just do a BIOS flash if it will update the NIC as well with a good copy. - chewbie, on 10/03/2008, -1/+2thanks Hello1024!
- computershack, on 10/04/2008, -1/+1Are they going to compensate all the people who had their NICs fried?
- justsomedigger, on 10/03/2008, -2/+2kinda unrelated, but i read this a day or two ago.. and thought this was kinda interesting...
this is a link that shows the investment company that bill gates owns and runs..
http://en.wikipedia.org/wiki/Cascade_Investments_L ...
anyone else think its shady that mr. microsoft controls the funds and investments of one of the premier Linux companies?
maybe its just me.. - Matt2k, on 10/03/2008, -5/+5So, what again? The driver overwrites the firmware. And the firmware doesn't properly make this difficult enough through a lock bit. Is this what I'm understanding? Amazing how this can be turned into a pro-Linux story. Good job, everyone.
- fatas, on 10/05/2008, -1/+1Linux stole my baby.
- vpshockwave, on 10/03/2008, -3/+3Trust intel* even more.
- MMaster23, on 10/03/2008, -4/+3Ok .. so Linux does well when they crap up your NIC and Intel has to come in and save your damaged hardware?
Oh yes.. I can clearly see how Linux is doing well. There is a reason we have WHQL on Windows. - Giga, on 10/04/2008, -2/+1"Is Google a large enterprise? ..."
Yeah, but I think their stuff typically consists of "special servers". - Gavagai80, on 10/04/2008, -3/+2Having used no operating system but linux since 2003, I can fairly say that hardware support for linux really sucks to the point where you have to be a complete hardware expert or buy a pre-installed PC which you're absolutely sure you won't want to add anything to. Even if something works now, it may stop on your next dist-upgrade.
- TheWindBlows, on 10/03/2008, -10/+8It was intel's fault in the first place for allowing such a possible thing to happen in their driver much less their hardware.
-
Show 51 - 58 of 58 discussions
Digg is coming to a city (and computer) near you! Check out all the details on our 
© Digg Inc. 2009
— Content created and posted by Digg users is