Jurassicc wrote: OS4 has vunerabilities that 3.X didn't and we should be prepared to deal with malware.
On a classic you had to open the case to get at the roms. It can be done with a system friendly script under OS4, even unprotecting said write-protected files.
I am sorry, but this is an absolutely constructed argument. FWIW, the AmigaOS 4 bootloader creates a ROM copy that does nothing but reset the PowerPC and goes to HALT. It's easy to remap the ROM and patch it, or use SetFunction to override functions - as a matter of fact, this is exactly what SetPatch does. And if in doubt, you can just install a ROMTAG with a higher version number.
So this is not a vulnerability of AmigaOS 4.x more than it ever was on 3.x
Quote:
I was thinking something of an option in DOS prefs or Workbench prefs to protect system settings. Malware or the user would not be able to delete, modify essential system files, allow to write to rdb or remove the slb or alter filesystems.
How would that help? Without memory protection, you cannot protect anything because everything is open.
Seriously, if you do want to contact me write me a mail. You're more likely to get a reply then.
the ability of the Mach kernel to run multiple OSes at the same time.
That sound where interesting, and I think it might be where useful, one thing I hope, is that OS4 Mach becomes more Amiga Exec like, and hope most parts of OS4, becomes ported to Amiga Exec mach, so what we get low IPC legacy as possible, I hope mach features be open for any one to use, for PPC VM?s
(NutsAboutAmiga)
Basilisk II for AmigaOS4 AmigaInputAnywhere Excalibur and other tools and apps.
OS4 has vunerabilities that 3.X didn't and we should be prepared to deal with malware.
On a classic you had to open the case to get at the roms. It can be done with a system friendly script under OS4, even unprotecting said write-protected files.
Quote:
Rogue wrote:
I am sorry, but this is an absolutely constructed argument. FWIW, the AmigaOS 4 bootloader creates a ROM copy that does nothing but reset the PowerPC and goes to HALT. It's easy to remap the ROM and patch it, or use SetFunction to override functions - as a matter of fact, this is exactly what SetPatch does. And if in doubt, you can just install a ROMTAG with a higher version number.
So this is not a vulnerability of AmigaOS 4.x more than it ever was on 3.x
Quote:
Jurassicc wrote:
I was thinking something of an option in DOS prefs or Workbench prefs to protect system settings. Malware or the user would not be able to delete, modify essential system files, allow to write to rdb or remove the slb or alter filesystems.
Quote:
Rogue wrote:
How would that help? Without memory protection, you cannot protect anything because everything is open.Hi Rogue,
Hi Rogue,
I guess it's just the "romantic" in me.... I like it the way it was, we'd have the LAST "open" OS. Let's face it, you can't fight virii in any open system anyway* and that's what I was hoping we'd have (an open system).
I find it refreshing that even with SMP, STILL MP is not a requirement.
Why am I against MP? Because I believe that it's possible to write SW in a no MP environment that isn't possible in an MP enviroment, and barring virii/malware, anything that is made in an MP environment can be done in a non-MP OS.
Now imagine AOS4.x-UAE, all the SW has no idea that there is an "outside world".
When I sit down to compute, I want to be in "super-user" mode always (no passwords, levels of permisson, etc.).... access to anything, anytime. That is how it was and I don't see any real reason to have it otherwise.
Besides, THAT has already been DONE (crappily, I might add).
But I have no intention to disturb the (last) ship.... so I'll remain a (quiet) passenger.
* Closed "safe" system being, a cellphone, or a dumb data entry/retrieval terminal attached to a mainframe in another location.
Support Amiga Fantasy cases!!! How to program: 1. Start with lots and lots of 0's. 10. Add 1's, liberally. "Details for OS 5 will be made public in the fourth quarter of 2007, ..." - Bill McEwen Whoah!!! He spoke, a bit late.
We keep adding this, that, and the other, then we realize, "oh!, that's why they did this, that and the other (registry, say)" and before too long (at all) we have what THEY have, and I don't want to see that happen.... because they were there already, and I rejected that. Heck, we've already got the huge disadvantage of OS bloat simply because PPC code is 50, to 250% bigger than the old compiles.
In todays world of huge memories and HD space, so what, right? But, what about when we go 64 bit, won't the size balloon up again? Even the very same commands like "dir" and "list"? I don't know, I'm asking.
How am I going to get AOS4.x running in a wristwatch if it needs 250 Megs of ram to start (hey, vista hit 3 to 6 gigs, right? and I'm sure every last assembler byte is necessary!!! )
Support Amiga Fantasy cases!!! How to program: 1. Start with lots and lots of 0's. 10. Add 1's, liberally. "Details for OS 5 will be made public in the fourth quarter of 2007, ..." - Bill McEwen Whoah!!! He spoke, a bit late.
I guess it's just the "romantic" in me.... I like it the way it was, we'd have the LAST "open" OS. Let's face it, you can't fight virii in any open system anyway* and that's what I was hoping we'd have (an open system).
This sounds like a "just because" argument, and it doesn't convince me.
For the record, I don't think being the last "open" OS is a quality but rather a problem. Why would it be desirable to have an "open" OS when "open" means "vulnerable to both malicious and dumb code"?
Quote:
I find it refreshing that even with SMP, STILL MP is not a requirement.
The reason is that neither has anything to do with the other. SMP is multi-processing, and can be done without regards to the memory protection; it just double the chance to run into a fatal flaw of software.
Quote:
Why am I against MP? Because I believe that it's possible to write SW in a no MP environment that isn't possible in an MP enviroment, and barring virii/malware, anything that is made in an MP environment can be done in a non-MP OS.
Well, belief is for Religion. I would rather have you name a useful program that would be made impossible by a system with memory protection.
Yes, it is true that you can do everything that you can do in an MP-environment in a Non-MP environment. You can also implement any algorithm conceivable on a Turing Machine, there are mathematical proofs for that as well. But at what cost? Would anybody in his right mind try to write any complex algorithm on a Turing or Random Access Machine? I don't think so.
MP does not offer any additional functionality in itself. In a world where there is only correct code and no malicious code, MP is not required, just as e.g. E-Mail security was not required in a time where the only people sending E-mail around where university professors and the odd student. In an internet like today where most of the bandwidth is used up by porn downloads and Spam mail, the security becomes an issue. The same goes for MP.
Quote:
Now imagine AOS4.x-UAE, all the SW has no idea that there is an "outside world".
When I sit down to compute, I want to be in "super-user" mode always (no passwords, levels of permisson, etc.).... access to anything, anytime. That is how it was and I don't see any real reason to have it otherwise.
Well I can tell you a reason. Any time you click on a link in a web browser, some idiot script kiddie might try to shove some malicious code onto your harddisk. A friend of mine recently had his kids looking at a porn advertisement page because a virus had changed his browser homepage to some porn website. That isn't funny. Now, usually changing your homepage can be done without a password, but a virus will not stop there. It will install a program that will misuse your machine to send out spam mail to others, and this is possible because you are in "superuser" mode and it can install any crap on your machine without even so much as a notification.
Now, if the system would ask for your permission to install a new program, you would be suspicious and would look at what it wants to install, and that would improve your system's stability as well as preventing others from getting spam mail from your machine.
Things go further. You might be the only one in your household using a computer. Others might not. If security is compromised on one system, chances are good that it spreads. What if your next-door neighbor suddenly chokes your internet connection because the house you live in has a common access point?
Quote:
Besides, THAT has already been DONE (crappily, I might add).
You should not judge a concept by its implementation. besides, the only reason why it does not happen frequently on the Amiga is that the Amiga is insignificant. Compared to the number of Windows system, it is simply not worth the effort.
So, again, what good argument is there AGAINST MP, after so many arguments for MP?
Seriously, if you do want to contact me write me a mail. You're more likely to get a reply then.
Atheist wrote: We keep adding this, that, and the other, then we realize, "oh!, that's why they did this, that and the other (registry, say)" and before too long (at all) we have what THEY have, and I don't want to see that happen....
There is a huge difference between blindly adding features and adding features that are deemed essential. We could argue that "other systems already have a graphics user interface and web browser, so why should we have that", to which you would certainly say "because these things are essential"
Quote:
Heck, we've already got the huge disadvantage of OS bloat simply because PPC code is 50, to 250% bigger than the old compiles.
That is hardly an issue. The difference in code size isn't that big (admittedly there is a difference), and code size is rarely what makes the difference - a good PNG image collection is already bigger than most programs, not to mention game data files and textures.
The OS is bigger now than it was during 3.1, granted, but there is a WHOLE lot of more functionality in it that wasn't even REMOTELY available.
Quote:
In todays world of huge memories and HD space, so what, right? But, what about when we go 64 bit, won't the size balloon up again? Even the very same commands like "dir" and "list"? I don't know, I'm asking.
No it won't. And the point is that memory and harddisk space today IS bigger than it was (if you had a harddisk at all). The point is that the expectations have gone up too, today programs are more colorful, they look better, and they have much more functionality. I had used TASWORD II on a Sinclair ZX Spectrum to write texts, yeah that was possible but if you compare it to OpenOffice.org, the functionality does not even come remotely close to it.
Quote:
How am I going to get AOS4.x running in a wristwatch if it needs 250 Megs of ram to start (hey, vista hit 3 to 6 gigs, right? and I'm sure every last assembler byte is necessary!!! )
The question is not whether something is possible or not, but rather, whether something is necessary or not. What purpose does it have to be able to run in a few kilobytes when all embedded devices come with several megabytes of memory? If there is no difference in price? The Arctic had 64 MB, plus a 32 MB ROM where the OS and a RAM disk was located with OS 4 on it and a few programs (we even had Turbo Calc and WBsteroids running on the thing). What good does it make to only use a fraction of the ROM? There was still plenty of free space anyway.
So the question is whether you think that adding a feature like MP will raise this footprint significantly. The answer is no it wouldn't. The additional features that the OS might get might, but certainly not the MP.
Seriously, if you do want to contact me write me a mail. You're more likely to get a reply then.
does the PPC "ExecNG" have faclities for handling Virtual Machining in current state? just yes/no would be fine to answer this,
I see nothing wrong with defaulting the UMA to being read-write and anything *allocated* becomes write-only by the allocator... executable code being "write-enabled" exclusive of "executable" seems fair also...
I hope to code something dependant on such functionality once I lay hands on an AOS4 capable device, ...
I wish you all the best in the current mess
@Athiest:
Ive seen enough of this thread to actually get a proper grasp at how "Memory Protection" actually needs to work... and with what I learnt of the AOS3 and AOS2 systems I originally learned programming with I will be looking forward to the Hyperion produced AOS4 becoming generally available.
there is *nothing* stopping a program asking exec nicely to walk *all* virtual memory spaces as "read-only" to find strings... being able to "write" changes over the found string is another matter...
So, again, what good argument is there AGAINST MP, after so many arguments for MP?
I only have one: Make it within the next release/version and please do everything to get AOS4 released as soon as possible with no new features for the existent version.
An OS with everything a user/developer wants or can imagine is worth nothing if no potential user/developer is able to obtain it!
Why am I against MP? Because I believe that it's possible to write SW in a no MP environment that isn't possible in an MP enviroment, and barring virii/malware, anything that is made in an MP environment can be done in a non-MP OS.
Well, belief is for Religion.
Hi Rogue,
If I were Billsey*, I'd complain I was the victim of a cheap shot.
* But I'm not.
Support Amiga Fantasy cases!!! How to program: 1. Start with lots and lots of 0's. 10. Add 1's, liberally. "Details for OS 5 will be made public in the fourth quarter of 2007, ..." - Bill McEwen Whoah!!! He spoke, a bit late.
there is *nothing* stopping a program asking exec nicely to walk *all* virtual memory spaces as "read-only" to find strings...
I don't agree here, even ability to "read-only" is a security hole. just imagine I'm a malware program, I'll then ask read permission for the whole memory space and will snoop for YAM or SimpleMail signature, then I'll be able to retreive all your email accounts with usernames *and* passwords Thus i'll then be able to impersonate you on the network easily, "just" with this "innocent read-only" permission So in a MP system, each process can only see its own memory space except for memory spaces other processes *explicitly* tagged as public (or shared).
true enough ... I wasnt thinking about such security measures inside I was more thinking that protection would happen "at the door" not after entry is successful
Does the PPC "ExecNG" have faclities for handling Virtual Machining in current state? just yes/no would be fine to answer this,
I think the answer to that is almost, it does have virtualized memory.
This what TonyW wrote:
Quote:
True, but LiveForIt can't do that himself, only the Exec can write to the MMU registers, only the Exec knows where LiveForIt's program is and what the mapping for his program is
It looks like it not possible for me have my own memory map for a possess running because only the Exec can write to MMU register and configure it.
So we are where close to being able to write Virtual Machines on OS4.
Hehe you talk like I did 1994. Back then I was a fanatic Amiga fan aswell, "everything Amiga do is better". Too bad that isn't true anymore, and probably hasn't been for atleast 15 years. I grew up and got a more open mind to things. Windows doesn't suck anymore (without having tested it before). XP is actually usable and supports everything I want. It simply works. That doesn't mean I want better things or dislike Amiga. I actually like the control Amiga OS gives me.
You can't seriously mean that we shouldn't use MP because we should be different??? If MP doesn't cause bloat or make the OS less responsive and offer you a virtual enviroment for classic software, so what's the problem???
Sounds to me you're more into retro computing and not OS4 and more modern stuff. I mean keeping the hardware and software mainly the same forever. Am I wrong?
Edited by samwel on 2007/9/7 2:57:55 Edited by samwel on 2007/9/7 2:59:26
with the virtual machining I was thinking of launching non-native material within a wrapping process of some kind...
Me too, but I?m having some problems testing that out.
Quote:
anyway... I will have to leave things there until I can sort out an AOS4 machine to test some ideas out on...
Me too, my Amiga kind bits and peaces right now, CPU in Italy, motherboard taken out of the cabinet.
Quote:
donations or use of a machine to get something past "idea" into a functional toolkit would be appreciated, I can be found with this same username on UtilityBase for donation info
Sorry I?m not going to do that I simply can?t live whit out my AmigaOne.
(NutsAboutAmiga)
Basilisk II for AmigaOS4 AmigaInputAnywhere Excalibur and other tools and apps.
So, again, what good argument is there AGAINST MP, after so many arguments for MP?
I only have one: Make it within the next release/version and please do everything to get AOS4 released as soon as possible with no new features for the existent version.
An OS with everything a user/developer wants or can imagine is worth nothing if no potential user/developer is able to obtain it!
Bye
We are talking future here. Right now, any change of that magnitude is out of question.
Seriously, if you do want to contact me write me a mail. You're more likely to get a reply then.
If I were Billsey*, I'd complain I was the victim of a cheap shot.
Huh? Why would someone take offense on that? IMO (although this is totally off-topic) Religion is ALL about belief, belief without proof. Technical facts are about technical facts, and do not have anything to do with belief.
Seriously, if you do want to contact me write me a mail. You're more likely to get a reply then.