Yet another public BSOD
-
@ZiggyFish said:
This is why I love Linux, have you seen a public kernel panic?
Not a kernel panic, but I did see once the "news screen" in my office complex's elevator show Fedora Core booting up. Sideways.
However this was on a Friday, 9pm. Most people would have gone and it usually is maintenance time for these thingies. ;)
Windows, on the other side ... I've seen an NT desktop on an ATM screen showing some bizarre error.
-
@morbiuswilters said:
Prior to Mac OS X (released in 2000), no variant of Mac OS utilized segmentation or paging.
I suppose it depends how you define "paging". I would imagine that what Apple did was to use a single page table globally for the entire memory map, so that paging could be implemented without having to implement virtual memory for individual processes. Virtual memory would prevent older Mac software from writing to system globals, the very existence of which was a big mistake right from the start.
Wikipedia reports that Apple introduced Virtual Memory (page swapping) in System 7, which is what I thought. I don't think any of my Macs supported it though. However, they failed to provide virtual memory (that is, virtual address spaces, not "Virtual Memory" as in page swapping). Hell, even the Wikipedia article for the history of Mac OS links to the page on virtual memory (virtual address space) which is plain wrong.
-
@Daniel Beardsmore said:
I suppose it depends how you define "paging". I would imagine that what Apple did was to use a single page table globally for the entire memory map, so that paging could be implemented without having to implement virtual memory for individual processes. Virtual memory would prevent older Mac software from writing to system globals, the very existence of which was a big mistake right from the start.
Wikipedia reports that Apple introduced Virtual Memory (page swapping) in System 7, which is what I thought. I don't think any of my Macs supported it though. However, they failed to provide virtual memory (that is, virtual address spaces, not "Virtual Memory" as in page swapping). Hell, even the Wikipedia article for the history of Mac OS links to the page on virtual memory (virtual address space) which is plain wrong.
This discussion has been about memory protection all along. Mac OS prior to OS X did not utilize memory protection. You are correct that from OS 7 on the OS would utilize the MMU (assuming it existed) to catch invalid page faults for the purpose of provide a larger address space (assuming the user enabled the feature). It still provided no memory protection. "Paging" and even "virtual memory" generally refer to the combined practice of utilizing page protection and disk swap.
-
@morbiuswilters said:
This discussion has been about memory protection all along.
I started the discussion. I merely objected to the claim that Mac OS 9 looked like Windows 3.11. This was reinforced with the statement that they were the same because both were "fugly". Fugliness has nothing to do with memory management -- see the X11 screenshot MasterPlanSoftware posted.
@morbiuswilters said:
It still provided no memory protection.
I find it amusing that you continually argue for what I already said (even if it was a few mails prior) or not mentioned at all, as though I'd said the opposite.
I also find it amusing that you also disregarded where, right from the start, I detailed actual problems with Mac OS, in particular, ones that had nothing to do with memory protection. If an OS experiences a problem within the kernel, it's dead. BSOD. Kernel panic. Or, on a Mac, a bomb. System stability mandates that the kernel not crap its pants, and that was precisely the problem with Mac OS. One wrong API call and the kernel would blow itself to bits.
Mac crashes due to extension conflicts were also nothing to do with memory protection, as extensions all lived in kernel space and fundamentally altered the system's behaviour. Since all Mac drivers were implemented as extensions, it would be similar to kernel panics and BSODs from bad drivers. The problem wasn't memory protection, it was that even trivial system features were implemented by patching the kernel! By Mac OS 9, a very small number of features had become daemons (but not many) so at least if those crashed, odds are the system wouldn't care (unless they did make a vile API call, or genuinely start corrupting RAM).
With Mac OS, Apple got a lot right and a lot wrong. What I would wish for is for people to non-judgementally learn from Apple's successes and from their mistakes, from what they got right, and what they got wrong. Mac users didn't love their Macs for no reason at all, and we must remember that Apple invested heavily in quality UI design, some of which has slowly but surely made it over into the Open Source world and even Windows.
Mac OS is not perfect, neither System 1, nor 7, nor X, nor honestly is any OS. We make a choice, and we live with it. Using Mac OS taught me a great deal and I appreciate that, and in fact, I still use Mac OS 9 to this day. I must be cheating, as my best uptime for Mac OS 9.1, with some absurd number of programs running including my home Web and FTP servers, BBEdit, Photoshop etc was, I think, around 90 days.
-
@Daniel Beardsmore said:
The problem wasn't memory protection, it was that even trivial system features were implemented by patching the kernel!
It's because you keep posting stuff like this that I keep bringing up memory protection. Without it, a single wrong pointer in any program could take down the whole system. Without preemptive multitasking and memory protection, every application essentially runs in kernel mode.
@Daniel Beardsmore said:
With Mac OS, Apple got a lot right and a lot wrong. What I would wish for is for people to non-judgementally learn from Apple's successes and from their mistakes, from what they got right, and what they got wrong. Mac users didn't love their Macs for no reason at all, and we must remember that Apple invested heavily in quality UI design, some of which has slowly but surely made it over into the Open Source world and even Windows.
UI is subjective, but the vast majority of computer users use Windows. I'm not saying Apple got it all wrong, but they seriously lagged behind everyone else for a decade and a half. Not only did the OS lack the most basic stability and scalability features, Apple actively resisted many useful advances in UI design. I don't understand where this idea that Apple perfected the UI and everyone copied them comes from. Even OS X is obnoxious in this regard. Why do different windows from one app all share the same task bar? Why does closing an app window not close the app itself? Why are you not permitted to maximize a window? Why do Mactards get excited over a feature like "Time Machine" which has been standard in Windows since 1999? I know a lot of people who use Macs and they have applications and the OS crash on them all the time. Seriously, they usually have to reboot at least once a day. But they go on and on about how "at least I'm not using Windows!" I have to wonder what the fuck these people were doing to their Windows machines to make them worse than that.
For years I listened to Mac people rant about how PowerPC was so superior to x86. I actually agree with them in theory, but the day Steve Jobs announced the move to Intel they were all on-board. Why do Mac fans pretend their platform is so open when it's the most locked-down consumer OS there is? Why do Mac fans go on an on about how great it is that you can control so many Apple apps with Applescript like this is something novel? Microsoft was doing this shit over a decade ago. Why do Mac fans gloat about how "everything just works together" but when MS does the same thing they scream "monopoly!" Why in God's name is Apple praised for their elegance when their mice and keyboards are essentially ergonomically torturous (albeit pretty)? Why is it when Apple releases a patch it's a "good thing" but when Microsoft does it's just another sign of their failure?
Yeah, this is a full-on rant. I'm sick and tired of the Mactards. You know what, use whatever the fuck you want, but stop trying to justify it to me. The arguments are baseless and retarded. Stop pretending that because your OS is "built on UNIX" like you know what that means or that somehow it impresses me. UNIX has a history bugginess, proprietary lock-in and security holes. This isn't the "lock-in" Microsoft is accused of either, this is the absolute "buy our hardware, our applications and our support contract or you are SOL" lock-in that people praised Microsoft for freeing us from. Yeah, this isn't really aimed at you, Daniel, but I am sick of this nonsense.
-
@alegr said:
Oh, and while we at the squirting contest, here is a little trivia. In what year/version the following features were supported in Windows NT+ and in Linux or your favorite Unix:
1. Multithreading;
2. SMP.
3. Preemptive (multithreaded) kernel
4. Multiprocessor kernel.
5. Support for RAM larger than 1 GB.
6. Support for files larger than 4GB.
7 Support for RAM larger than 4 GB.
8. Fail-safe filesystem.
9. Filesystem support for transactions (at multi-file level).
10. Volume shadow copies.
11. Asynchronous I/O.
Can I add 2:
12. IPC(interprocess communication)
13. Userspace Filesystems
Most of that stuff, has been done before under a different name. I'm too busy to find out exactly what dates and will look into it over the weekend.
As for Windows 3.1 having memory protection, AFAIK paging was used for memory management and never memory protection.
-
@morbiuswilters said:
Yeah, this is a full-on rant. I'm sick and tired of the Mactards.
Funny -- I seem to recall that, so far, I've not attacked Windows, I've not attacked Linux, and I've explained real problems with Mac OS 9 and earlier, which are in fact still significantly relevant to an OS with protected memory and pre-emptive multitasking! (cf. hard microkernel) So rant away, but this has become too senseless for me.
-
@Daniel Beardsmore said:
Funny -- I seem to recall that, so far, I've not attacked Windows, I've not attacked Linux, and I've explained real problems with Mac OS 9 and earlier, which are in fact still significantly relevant to an OS with protected memory and pre-emptive multitasking! (cf. hard microkernel) So rant away, but this has become too senseless for me.
You do, however, seem to practice selective reading comprehension. You apparently missed morbiuswilters saying:
@morbiuswilters said:
Yeah, this isn't really aimed at you, Daniel, but I am sick of this nonsense.
I quite agree, and kind of like the term "Mactards". <g>
-
@KenW said:
I quite agree, and kind of like the term "Mactards". <g>
Indeed, I think that is going to become a regular part of my vocab. I think we need to start using that as a tag here.
Can we also have something like nixtard? Suggestions on something better?
-
@KenW said:
You do, however, seem to practice selective reading comprehension. You apparently missed morbiuswilters saying ...
I read that perfectly well, but I don't see his point, as he was screaming at me up until then, and I don't see anyone else here with intimate knowledge of classic Mac OS who he could have been referring to. Besides, I'm tired of the senselessness of his stance that the sole determiner of system stability is memory protection. After all, Linux is well-known for kernel panics from defective drivers, and it has excellent memory protection.
System stability comes from many factors. You can put in a lot more stability by going hard microkernel, although as I understand it, hardware still needs direct access to RAM to write incoming data so even if drivers are all isolated in separate address spaces, flawed or, worse, misdirected hardware can still clobber RAM. A driver that sends the PCI card the wrong address range by mistake could easily end up hosing the system.
As it stands, Linux and Windows both run a lot of software within kernel space and memory protection won't save you there. Everyone who's had Windows bluescreen or Linux kernel panic -- i.e. everyone -- will attest that memory protection can't save you. This is not to argue that we've come up a very long way since the bad old days. And regardless of the conflicting views on whether Windows 3.11 had protected memory or not, it sure crashed plenty often! Boy am I glad to be done with that. Trying new software out was a scary, as all too often, you'd just hang the system.
-
@Daniel Beardsmore said:
I read that perfectly well, but I don't see his point, as he was screaming at me up until then, and I don't see anyone else here with intimate knowledge of classic Mac OS who he could have been referring to.
Just a general anti-Mac rant and most of my complaints were about UI annoyances that exist in OS X. It really was nothing personal against you.
@Daniel Beardsmore said:
Besides, I'm tired of the senselessness of his stance that the sole determiner of system stability is memory protection. After all, Linux is well-known for kernel panics from defective drivers, and it has excellent memory protection.
Never said it was. But you cannot have stability without memory protection, that's been my point all along. I think you underestimate (or understate) how significant memory protection is to overall system stability. Kernels and drivers have bugs but they are usually much more stable than application software.
@Daniel Beardsmore said:
Everyone who's had Windows bluescreen or Linux kernel panic -- i.e. everyone -- will attest that memory protection can't save you.
I've had my fair share of both. However, anyone who has run a UNIX command on an active server and seen "segmentation fault" printed on the screen will be damn thankful for memory protection. Just like anyone using Windows will appreciate when it kills a rogue application.
@Daniel Beardsmore said:
And regardless of the conflicting views on whether Windows 3.11 had protected memory or not, it sure crashed plenty often!
True, but I often wonder what people are using as a baseline here. DOS crashed more. I've had more trouble with Mac OS than any flavor of Windows (YMMV). Solaris was pretty damn stable, but it wasn't really a consumer OS. Honestly, Windows has always been damn near the top (if not at the top) of the consumer OS market in terms of features, compatiblity and stability. I am not a Windows user at all, mind you, but I know from personal experience that OS X and Linux are usually worse. Still, Windows, Mac OS and Linux have made great strides in the last 20 years. We owe a lot to the hard work and persistence of OS developers at Microsoft, IBM, Apple, Redhat and others.
-
@morbiuswilters said:
Just a general anti-Mac rant and most of my complaints were about UI annoyances that exist in OS X. It really was nothing personal against you.
My "Mactard" days are long over. Personally, I find the Mac OS 9 UI refreshingly self-consistent, while Mac OS X's UI is pretty hopeless. Once you start hiding programs and minimising windows you'll find yourself very confused, and Mac OS X copied Microsoft's stupid idea of unowned windows -- Mac OS 9 does not permit a document class window to have no owning application. (Globally floating windows can belong to daemons, that's how the Control Strip and Application Switcher work.)
Another great way to crash Mac OS 9 is to write a daemon and tell it to open a focusable window. The window will open in the background. Click on that window, and Mac OS instantly crashes. I would hazard an assumption that when you click on it, the UI tries to bring that program forward, but being a daemon, it's not listed as running and the UI gets in a muddle and dies.
Again, nothing that protected memory would help with. You've just crashed the entire Mac graphical interface. I don't know what happens if you crash the window manager in Linux, but I've certainly triggered a crash of the whole X11 GUI, and of course you lose all your work. However, since the Linux GUI is non-kernel, the kernel is still up and your apps could use the peer reset of the connection to initiate an autosave before they bail out. The Mac GUI is part of the kernel, so if you crash it, there's nothing you can do!
Protected memory is good, but it only protects processes, not the kernel. I can't figure Windows out though -- where exactly IS the Windows GUI? CSRSS? ntoskrnl? I know you can pass confusing parameters into (something like) SetWindowLong which can lead to some very confused-looking windows, like two restore buttons side by side, but you'll never crash it, just confuse it sometimes. However, if the GUI does hang -- now what?
With Linux, it's simple: open another virtual terminal (ctrl-alt-1 or something, I can never remember), log in again, and kill off any offending processes that were screwing up the GUI. Mac OS X doesn't have this either, so if Windows or Mac OS X freeze, and you think it's the GUI, your only option is to have previously installed (Windows) or enabled (OS X) a SSH daemon. Or the Windows Telnet server. With Windows, you then need to figure a way out of the mess. taskkill will kill a process (if you have XP), but how do you list running processes? With Windows 2000, you'll have to install extra commands like kill and tlist for this.
Fortunately, the Windows GUI is pretty robust. OS X's GUI is robust in the sense that it won't segfault, but it's much easier to get to the point where you can't get it to respond any more.
-
@Daniel Beardsmore said:
I don't know what happens if you crash the window manager in Linux
I'm not really arguing, I'm just answering as if that were a question.
Anyways, if you run a DM like GDM, KDM or XDM generally it will call either the desktop environment's session manager or the window manager for the lighter stuff like wmii, fluxbox and ratpoison (there are many more), if the WM is killed usually you can respawn one. However there is a process that dominates X, if that dies so does X, it is normally the DM, for hand-crafted xinitrc files it is usually the WM, or on my box it is root-tail as I switch WMs for testing.
I don't have GDM on this box so I can't give you an exact answer, but as I recall you can kill all X clients other than GDM and be fine. Haven't used KDM or XDM so I can't comment on those.
PS: Just tested GDM with fluxbox, if FB dies GDM logs you out. Sorry about the acronyms and abbreviations.
-
@alegr said:
Oh, and while we at the squirting contest, here is a little trivia. In what year/version the following features were supported in Windows NT+ and in Linux or your favorite Unix:
1. Multithreading;
2. SMP.
3. Preemptive (multithreaded) kernel
4. Multiprocessor kernel.
5. Support for RAM larger than 1 GB.
6. Support for files larger than 4GB.
7 Support for RAM larger than 4 GB.
8. Fail-safe filesystem.
9. Filesystem support for transactions (at multi-file level).
10. Volume shadow copies.
11. Asynchronous I/O.
Sorry, I've been a bit busy lately.
As I said, all of them have been done before (most before windows NT), if you look deep enough you'll find, that UNIX or other OS had done all of them before windows. I will how show you one:
@alegr said:
10. Volume shadow copies.
This took me some time to find information, on it. But in the end, RAID 1 does a similar job. From what understand, volume shadow copies are mainly use to allow backup software to access locked files. UNIX avoids this by keeping configuration and locked files in separate parts of the file system. The other reason why Volume Shadow copies are use is to prevent data loss (i,e if one hard drive fails, the other kicks in). This is what RAID 1 was design for.
As for Windows 3.1 memory protection, I was hoping to test the following out on windows 3.1 (fuck this is a shit forum editor):
- Create program that does the following:
- Wait for the user to press a key
- Display cs and ds.
- Send control to the operating system
- terminate.
- start 2 instances of that program.
- press key on one
- copy the output
- press key on other
- compare this output to the first program, If different, it means that we have a segmented OS.
- Create and run another program that uses ES to access data from another process.
-
@ZiggyFish said:
Sorry, I've been a bit busy lately.
I was actually enjoying the absence of Ziggy. I thought my sign worked.
-
@ZiggyFish said:
blah blah blah
So what is your point here? You really need to lay off the trolling...
-
@ZiggyFish said:
...wtf wtf wtf WTF wtf..
Dude, no way. Just... no. You are so off-base here that none of your points deserve a response. Let it go.
-
@morbiuswilters said:
@ZiggyFish said:
...wtf wtf wtf WTF wtf..
Dude, no way. Just... no. You are so off-base here that none of your points deserve a response. Let it go.
Do you get the same feeling as I do? This is just a ~12 year old kid who read a book or two about computers, tried Linux, fell in love with the uber-1337 feeling, and then started feverishly reading linux/anti-ms forums.
Now he feels he has reached some level of enlightenment and is going on a campaign to spread the gospel?
I mean, just read his replies, it all fits in place perfectly. He has this basic idea of computers, but because of all the retard forums he has read feels he is intimately familiar with everything technology related. After all, how hard could it be to write an OS? He could probably do it in a weekend. (If only mom wouldn't make him clean his room!)
-
@MasterPlanSoftware said:
Do you get the same feeling as I do? This is just a ~12 year old kid who read a book or two about computers, tried Linux, fell in love with the uber-1337 feeling, and then started feverishly reading linux/anti-ms forums.
Now he feels he has reached some level of enlightenment and is going on a campaign to spread the gospel?
I mean, just read his replies, it all fits in place perfectly. He has this basic idea of computers, but because of all the retard forums he has read feels he is intimately familiar with everything technology related. After all, how hard could it be to write an OS? He could probably do it in a weekend. (If only mom wouldn't make him clean his room!)
Yeah, that's the feeling I get too. I hate to jump to conclusions about random people over the Internet, but he has consistently shown that he understands some basic concepts but misses most of the points that are being made. He also seems intent on disregarding anything that doesn't mesh with his pseudo-religious view on operating systems. Like the RAID stuff, he just found the first thing that he could compare volume shadow copies to and proceeded to call RAID a Linux invention. *sigh*
Ziggy, if you're going to be in IT, please educate yourself thoroughly on a topic before commenting on it. Drop the fanatical Linux support that is blinding you and always keep learning. If you're just a kid who has no real interest in IT and just wants to feel 1337 because of your OS, fuck off back to Slashdot.
-
Your all a bunch of Microsoft munchkins, if I had of said anything wrong about Linux, you would of loved me.
Also I never said RAID was a Linux invention. I said that UNIX has ways of getting around the need for this technology. (I like this 'feature' you have in your forum editor, when you delete one letter you delete two)
-
@ZiggyFish said:
Your all a bunch of Microsoft munchkins, if I had of said anything wrong about Linux, you would of loved me.
Actually we would have called you a little idiot for insulting Linux too. It is not the argument you are making, but the way you are making it. You sound like an idiot. Just about everyone here uses Linux.
@ZiggyFish said:
(I like this 'feature' you have in your forum editor, when you delete one letter you delete two)
Actually it only does that for you because it knows that what you are saying is just trolling.
@ZiggyFish said:
Also I never said RAID was a Linux invention. I said that UNIX has ways of getting around the need for this technology.
God, you really are a fucking fool.
-
@ZiggyFish said:
Your all a bunch of Microsoft munchkins, if I had of said anything wrong about Linux, you would of loved me.
Only if what you said was correct. I never claimed Windows is without problems, but I refuse to listen to senseless Microsoft bashing. I've already had Slashdot ruined for me (regular reader since 1998) by that kind of close-mindedness and I don't want to have it overtake another site I enjoy. There are plenty of bugs in Linux, FreeBSD, OS X and Windows. Personally I don't feel the need to bash a perfectly good OS to feel justified in using Linux.
@ZiggyFish said:
(I like this 'feature' you have in your forum editor, when you delete one letter you delete two)
Uh, WTF? Nobody here wrote the forum software (that I am aware of) and it seems fairly stable to me. I've noticed several small bugs, but nothing that ruins the experience for me. I've heard others complain about the backspace bug, but I have never seen it myself. Since belgarion always complains about it I'm going to assume it's a legitimate complaint, but I've used FF2 on Ubuntu, Gentoo and a completely custom compiled Linux install and never had the problem.
@ZiggyFish said:
I said that UNIX has ways of getting around the need for this technology.
RAID and volume shadow copy are NOT the same thing. That was my point. And Windows has supported RAID for an eternity as well, so it's pointless to bring it up. You need to go out and actually educate yourself on what volume shadow copy is. Also, there are implementations available for Linux and have been for quite some time, but they are proprietary. An open source alternative has been available for Linux since 2003 or so (check my previous post where I already answered all of the Linux vs. NT questions to find out when).
I can be a patient man, but you are really getting on my nerves. I think it is better to encourage individuals like yourself to educate yourself than to just outright insult. This is because I believe that one day people like you will be working in IT and I want all of you to have an open, educated mind. I have spent years learning everything I can about computers and I don't consider myself worthy of commenting on something until I have a very good grasp of the concepts involved. I learn something new every day and I think most of the people on here do as well. However, this is the last attempt I'm going to make at reasoning with you. You can engage in informed discussion and I will gladly treat you with respect (as I imagine most people on this forum will). I'm not going to call you an idiot if you are unfamiliar with something or if you ask questions. I get the impression that you are a younger person and I don't want to lash out at you and reinforce the elitism you've already assigned yourself in your own head. If you persist in posting uninformed nonsense, though, I will treat you with the same respect I treat Lysis with. That is all.
-
@MasterPlanSoftware said:
Do you get the same feeling as I do? This is just a ~12 year old kid who read a book or two about computers, tried Linux, fell in love with the uber-1337 feeling, and then started feverishly reading linux/anti-ms forums.
Now he feels he has reached some level of enlightenment and is going on a campaign to spread the gospel?
Same thing I was thinking...
I have an idea. What if we find a way to introduce him to Cruise and Travolta? He could convert to Scientology, and then maybe would move to religious fora instead to post the nonsense!
-
@morbiuswilters said:
Wrong, wrong, wrong, wrong. Windows 1.0, 2.0 and DOS could all run in protected mode.
I'm familiar with DPMI, but let's take a look at what I said again.
@CDarklock said:
The primary variety of memory protection implemented in Windows 3.1 was called the "General Protection Fault". If a process tried to use "protected" memory, it crashed.
So what DOS and Windows 1 or 2 did is simply irrelevant.
You go on to say:
Older versions of Windows would often report a variety of end-user error messages for invalid page faults such as "general fault protection", "invalid page fault" or "an unrecoverable error has occurred".
Why, that's almost exactly what I said:
Thanks to compatibility, if you had a Windows 3 program that could reliably produce a sufficiently severe UAE, moving it to 3.1 and 95 would show you that the exact same problem still crashed Windows handily... it just gave you a different error before you had to reboot.
I fail to see what is "wrong" with that. Just because you call something "protection" doesn't mean it actually protects you.
-
@belgariontheking said:
We've got a Win98 machine here running some alarm software that refuses to run on anything NT-based. It's uptime counter just rolled over the second time.
-
@lolwtf said:
@belgariontheking said:
We've got a Win98 machine here running some alarm software that refuses to run on anything NT-based. It's uptime counter just rolled over the second time.
Hey n00b, you got something to say? Maybe something like "is it possible that you're running a 16 bit version of Win98?"
Also, learn to quote properly. I didn't say what you quoted me as saying.
