As was foretold, we've added advertisements to the forums! If you have questions, or if you encounter any bugs, please visit this thread: https://forums.penny-arcade.com/discussion/240191/forum-advertisement-faq-and-reports-thread/
Options
Vista, IPX and Diablo 1
LoneIgadzra
Registered User regular
Registered User regular
Okay there has got to be a way to get this shit to work. For LAN multiplayer, Diablo 1 only supports IPX. Is there some way - any way - to get this to work on 64-bit Vista?
LoneIgadzra on
0
Posts
pretty sure it's still around.
http://www.starbase01.com/site/index.php?aid=87
http://forums.microsoft.com/TechNet/ShowPost.aspx?PostID=1200750&SiteID=17
For Vista 64x versions you would have to disable the forced driver signing or what ever the name is to get IPX installed and working until a third party app from Natware or someone else comes out.
That's pretty damned annoying. Who the hell is microsoft to tell me what the fuck I can or can't do with my hardware?
Welcome to the world of closed source software. If you want to use Windows or OSX you have to put up with their conditions that you agree to when the OS is installed. This is why I went with Ultimate x86 for now. We're just going to have to wait until official software is written for IPX to be used on Vista x64.
Something like 70% of all crashes in Windows are the fault of crappy third party drivers. In Vista x64 they decided to make all drivers go through a certification process to get signed, so they'd be less likely to crash the box and cause the user to complain about Microsoft's unreliable OS.
Bullshit.
90% of people stick with the drivers that came with their PC until the day they buy a new one.
I reiterate:
BULLSHIT.
Well, many of those default drivers have sometimes seemed pretty shitty to me. So the signing won't help much, but I can see the 70%.
Default drivers will never be third party drivers. Ever. Third party drivers are things like Omega drivers for your ATI card. How many gamers even install third party drivers? There's no way 70% of crashes are caused by something that maybe 1% of computer users install. Usually third party drivers are installed because those drivers are better than the ones that manufacturers provide.
Most drivers that come with your windows are not made by Microsoft. A lot of them are old and crappy.
Hell, every logitech wireless keyboard or webcam I've installed on work computers the last 2 years, the drivers on the disc that came with it caused bluescreens.
http://en.wikipedia.org/wiki/Omega_drivers
Percentages don't mean shit unless you have hard data. That goes for the both of you. And anecdotel evidence is NOT evidence.
The VAST MAJORITY of people stick with the drivers that come with their systems. These drivers are not third party drivers. Therefore I contend that the majority of windows crashes, let alone 70% of them, are NOT caused by third party device drivers, and so such reasoning is bullshit for microsoft closing up driver freedom.
edit: also anecdotal evidence IS evidence, of an anecdotal nature.
Third party drivers are drivers not written by Microsoft. The drivers for your onboard sound, for instance, were almost certainly not written by Microsoft. If you have an AMD or NVidia graphics card, those drivers were not written by Microsoft. The same is likely true for your network drivers, they're specific to the motherboard.
There are third party drivers on virtually 100% of any stock, out of the box PC.
(edit) For proof, open up your Device Manager. You can get to it through System (windows key+break)->Hardware->Device Manager. Right click on a device, go to Properties, then the Driver tab. The Driver Provider is the company who wrote that driver. If it doesn't say "Microsoft" it's a third party driver. For instance, I have drivers from Broadcom, SigmaTel, Intel and ATI Technologies Inc. running on my box that has absolutely no accessories attached. Good places to look are Network adapters, Display adapters, Sound controllers and System devices.
(edit2) Also the 70% was arrived at through analyzing a lot of crash dumps sent to Microsoft through Watson. If you load the dump in something like windbg you can often tell exactly what caused the crash. The total breakdown was ~70% third party drivers, ~10% hardware failure, ~5% bugs in Microsoft code and ~15% the memory was too corrupted to find the root cause. This was on XP, not Vista, but it's rather informative of their decision to enforce better quality control on third party drivers for Vista x64.
Point number two is that Microsoft has nothing to gain by "closing up your driver freedoms." They're only lightly into PC hardware, and they routinely approve and sign drivers made by competitors there such as Logitech.
This whole thing is just another vapid theory borne of ignorance.
I meant in the context the previous guy was using them. I mean, yes, there are drivers that are third-party to the hardware, but in terms of windows, most everything is third-party.
However if the guy said third party meaning 'not written by the OS vendor', his argument is kind of invalidated by the fact that MS default drivers suck a lot more than manufacturer drivers, and typically third party drivers (real third party drivers, as in written by someone not directly tied to windows or the hardware in question) are used because they're better than either. Otherwise they don't get used. Either way it's no reason to force people to go through a signing process which we can already see was ineffectual at making Vista launch drivers reliable.
My argument is based on the fact that, whether or not you know the meaning of 'third party', only allowing signed drivers does not improve stability. Grok?
I'm calling them third party drivers because they are drivers that interact with Windows, but were not written by Microsoft. They are third party drivers in the OS environment. Still, feel free to yell "BULLSHIT" because our definitions differ.
The "fact" that MS drivers crash more? Where did you get this data? Anyway, the signing process is about how the driver interacts with the kernel, not about how well it knows its hardware. It can be tricky writing device drivers because you're dealing with asynchronous device events that can be difficult to debug. You do something wrong, you crash the OS. The quality testing and signing is all about trying to enforce some kind of standards here.
And about the Vista drivers, where are you getting your statistics about driver reliability at launch? Second, if they were less reliable, why would it make any sense at all to blame the driver signing process for not completely eliminating all the churn from moving to a new OS? Do you think Microsoft would have been able to simply deny people the use of their graphics cards when ATi and Nvidia dropped the ball on their Vista drivers?
(edit) Also if you are still curious about where my numbers came from, check out slide 7 of this powerpoint deck.
ATI and nVidia launch drivers both 'illegally' (as in against MS driver policy) access the kernel. And were signed drivers. Thumbs up for quality testing and standards enforcement.
Go look up any gaming benchmarks of launch Vista. Vista launch drivers were one of the reasons for a slow adoption of the OS. I build systems regularly and I've watched a steady improvement in driver quality since launch, eventually culminating in a point where (beyond a minimum spec) Vista keeps up with XP pretty well. Mostly thanks to first party drivers written by hardware vendors, with a lot of progress spearheaded by third party driver writers.
I love your last two sentences.
Microsoft can't enforce quality by refusing to sign shitty drivers, otherwise they'd be denying people the use of their soon to be unreliable hardware. The solution therefore is to have a driver signing process, where microsoft only signs reliable drivers. But Microsoft can't enforce quality by refusing to sign shitty drivers, otherwise they'd be denying people the use of their soon to be...
oh shit, logic short bus derailed!
Dear lord. I guess if we have two reasonable definitions which both have precedent, yours is right and everyone else should learn better? But fine, I'll call them non-Microsoft drivers since any other term confuses you horribly.
When someone is releasing a new driver or a new version of a driver for the same OS, they can pretty easily reject the drivers because there's already a working version, or there are no consumers out there with the device yet. When you have millions of people with a new OS who cannot use their graphics cards at all until some kind of driver is released, maybe you decide to allow a driver, even if it's not perfect, so those millions of people are at least not royally screwed. Besides, the bad performance on the drivers at launch is not a reliability issue. You still have not shown that they were any less reliable.
I assume you have a gaming machine, I invite you to reinstall your default vista video drivers and try an extended session of WoW or UT2k3/2k4. Both don't just run badly, they freeze or crash out to the desktop with default Vista ATI drivers, and nVidia drivers for at least WoW (not sure on UTs).
If that's insufficient, I should ask why I'm the one providing evidence when you're the one making the claim that signed drivers improve stability (except when they don't, which is also totally okay, apparently.) But instead I will suggest that you could try a google search on 'ATI and nVidia kernel security hole.' Goooo microsoft signing process! Finally, approved drivers that can be exploited to rewrite my registry.
edit: and MS approved those drivers last year, when theoretically your signing process should no longer have been corrupted by the desire to shove a much-delayed operating system out the door as quickly as possible.
So it's impossible for the same term to have different meanings in different contexts? Maybe "third-party" means one thing when talking about what driver you want for your graphics card, but might mean something different when breaking down causes of OS crashes? Rather than accept this, you'd rather just yell that your definition is correct in all circumstances? You're just being an ignorant prick now.
The quality testing and signing process is meant to improve them and make them more reliable, not make them perfect. Pointing out a single bug as "evidence" that quality testing does not work is like pointing out one criminal on the street and saying "look, the police didn't catch him, obviously the police are a complete waste of money and should be disbanded." Something that can catch all bugs in a given piece of code does not exist.
So a signing process was useless at launch, allowing the introduction of severe security flaws, has been useless since then, allowing the introduction of severe security flaws in drivers for the two most popular video card manufacturers, and this evidence is insufficient for forced driver signing being a worse idea than letting the user worry about their drivers? Since, apparently signed or not, the user's going to have to do that anyway?
Microsoft is missing huge security flaws in some of the most popular hardware drivers. It seems to me that this signing process isn't just a wasted effort, but something that does more harm than good by restricting an end-users control over their own hardware through unsigned third party driver options. A warning that you're installing an unverified and uncertified driver (and I've watched MANY users stop at the WDM certification warning for XP) and MS can't guarantee stability or performance would be enough to keep average computer users away from such drivers while providing power users with the option to expand or tweak hardware to whatever extent they desired.
Did you know the same word can have different meaning in different contexts? Consider the following two sentences: "I drank out of that cup." and "I am going to cup my hands together." The very same word takes on two different meanings, depending on what surrounds it! Wow, it's the magic of language!
You have no evidence that the signing process has "become useless" since launch. All you do is point out individual bugs, which means absolutely nothing when you want to evaluate how rigorous a certification process is. A point in my last post which you apparently ignored. Any certification process will let through some bugs. What matters is how many you catch and the defect rates of certified vs uncertified. You don't have this, you have assertions that you are pulling out of your ass.
As for the unsigned driver warning, I'm sure some people stop there, but plenty of people are going to click through anyway, especially if it's the only way to use a device on their computer. Making it annoying to run unsigned drivers not only makes it harder for someone to casually make their system unstable, it's leverage to get hardware companies to get their drivers certified. If not getting drivers certified means users have to set a boot option every time they turn on their computer, they are sure as hell going to get them certified. If their users just have to click through a warning, it's less of a motivator.
I do understand the frustration of advanced users and the downside of making it more difficult to tinker with your system. I'm sure this disadvantage was weighed against the advantages of making their platform more stable for the average user. They went with stability and I think that's a reasonable choice. I know it rubs you the wrong way but hopefully you understand the other side a bit better now.
No more out of my ass than your assertion that signing drivers DOES somehow make them more stable. You said it first, and you haven't provided any sort of evidence for it.
http://arstechnica.com/news.ars/post/20080325-vista-capable-lawsuit-paints-picture-of-buggy-nvidia-drivers.html
Here. Looks like driver signing did great. That some unspecified point in '07.
Yeah, that definitely smells like it came out of your ass. Where's your evidence that vista driver signing has helped anything? I haven't seen it, anywhere.
I thought we already covered the whole graphics drivers at launch thing?
I think it's a reasonable assumption that drivers that pass a quality test are going to be higher quality. But since it has to be laid out for you, here's an example. It's talking about ATi's run-ins with WHQL and their efforts to comply with it. And from what I can read a few of ATis candidate drivers have failed WHQL.
That link, like all your other links, isn't proof in any way that signed drivers improve stability. It's from 2004. Geez. And no, it's not a reasonable assumption that signed drivers will be higher quality. Unsigned drivers are often used because they can do things that signed drivers can't, or do the same things better. Typical official video drivers, just as an example, are optimized for synthetic benchmarks. Not being able to use unsigned drivers means being stuck with a system that isn't actually optimized for what you're doing with it.
Since it apparently has to be laid out for you, find me some proof that signed drivers are anything other than a clever way to help enforce DRM.
http://www.microsoft.com/whdc/winlogo/drvsign/kmsigning.mspx
As I understand it, SP1 has also patched the signed driver rule into x32.
http://blogs.msdn.com/windowsvistasecurity/archive/2007/08/03/x64-driver-signing-update.aspx Dumb as it may sound, that really is a big deal. Remember that this only applies to kernel-mode drivers... which really shouldn't be much of your hardware.
I only run x86, so forgive me if I'm wrong: can't you run unsigned drivers if you really want just by pressing F8 at boot?
That author-providing bit is a good idea.
If there's no quality testing at all, I definitely can't see advantages of the signed driver policy. It looks like we get the same driver quality we always got (totally dependent on the hardware manufacturer) and lose out on third party driver choices.
What? The article talks about what a big deal WHQL certification is and how committed ATi is to meeting it. It's pretty darn obvious they're improving their drivers to meet it. And even if it was in 2004, it still debunks your claim that WHQL testing never improves quality. Also, an Nvidia driver fails WHQL in 2007.
And no, Vista x86 did not inherit x64's signing requirements in SP1. If the x64's signing requirement were meant for DRM, it would be stupid to only protect one version of the OS because the pirates would just use the other.
God, you're helpless. Here:
http://en.wikipedia.org/wiki/WHQL_Testing
You pass WHQL testing, you get to install your device driver on x64.
Can you read, RandomEngy? Are you unable to grasp the implications of the above for the discussion at hand?
I'd love to see where I made this claim, or a post I make in this thread where I ever use the term 'WHQL.' I'm talking about the restriction to signed drivers. As I always have been. I believe I've said, over and over, that driver signing doesn't do anything but screw over users by stopping them from using third party drivers. While you kept replying to me to tell me different. The only person who actually pointed out a way driver signing can improve stability was ecchi.
Remember that? The post that prompted your response, and the whole discussion? Or did you forget what we're talking about?
Actually leaving it out of x86 is kind of a smart idea, x64 is essentially predestined for adoption with the 3gb memory limit of 32 bit operating systems, while leaving x86 vista closer to XP's driver security model means that 32 bit users won't have one more reason NOT to migrate to Vista.
edit: and microsoft has said that this 'feature' will be a part of all new windows operating systems from now on.
This is the first thing you wrote. It's as wrong now as it was then. The signing process is not about certification and is therefore unrelated to driver reliability. It just restricts the end user to using signed drivers, which means they can do less with their hardware.
Done? Can you hold the entirety of that paragraph in your head, and understand?
Relevant part:
However, about the DRM, it's just a conspiracy theory. Enforced signing for kernel mode code in x64 is about being able to identify who wrote the code and being able to revoke certificates that sign malware code. This gives the user transparency into what code is running on their box and makes it easier to contact authors of drivers that cause OS crashes. If you can prevent malware from getting into the kernel mode, you can prevent the nasty, hidden malware, the rootkits. This article talks about the approach a bit. That's what this is all about. It's a strong enough reason on its own to warrant the move.
I know you're still going to be convinced that Microsoft did this because they want to control your hardware and they love DRM, but at least the facts are out there now.
Also, you can test-sign any unsigned drivers you are working on with your own certificate. It will only be valid on your box, but you won't need to hit a boot option every time.
And a side note, I found that the ATi vulnerability that would allow unsigned code to execute in kernel mode was never exploited. It was found ahead of time and patched without incident. As far as I could tell there has been no malware that has actually managed to run in kernel mode in x64.
http://www.pagetable.com/?p=12
Take a look. Guess I'm not the only one who came to the DRM conclusion. Given Microsoft's forays into digital delivery (MS's political stance in the last five years, too) this seems pretty sensible. Being the OS that's considered 'safe' for content delivery would do a lot for their desktop market share as digital movie/tv show/music delivery gets more popular.
Anyways, malware is big business, and even if it took several hundred dollars spent on a certificate to write it, the organized crime that backs Big Bad Malware would happily pay. They make millions off it. Not that they'll have to, they'll stick to user-mode software.
Also, I'm well aware that other people also think x64 driver signing is all about DRM. That doesn't make them correct.
As for the theory "they'll just pay a few hundred dollars and get their kernel-mode malware signed", I don't think it's borne out in practice. As far as I can tell, the only revocations they've had to do were with the ATi security hole and with Atsiv. Perhaps the fact that you need to provide proof of a business with a physical address to get a signing certificate is sufficient incentive to keep malware authors to stick to user-mode, which means easier to be removed. Though I'd be thrilled if you could find another example. Also, being able to kill the malware by revoking its certificate is quite a handy tool. I guess in theory you could have a criminal ring that forges a lot of fake businesses and addresses to try and create new certificates faster than they're revoked, but that just hasn't happened.