Author Topic: Well ArcaOS 5.0 is probably the last version for me.  (Read 5583 times)

Roderick Klein

  • Sr. Member
  • ****
  • Thank You
  • -Given: 0
  • -Receive: 26
  • Posts: 425
  • Karma: +6/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #45 on: June 01, 2021, 10:44:04 pm »
I suspect I already know the answer to this, but I seem to recall reading years ago (probably 15 to 20 years ago - back when IBM still had a few people working on OS/2's kernel) that someone at IBM was looking into updates / modifications to OS/2's kernel to allow it to enable and use PAE mode - at least from the standpoint of allowing memory above the 4GB barrier to be fully used by the system.  Does anyone have any idea what might have become of this?

I'm guessing the current ramdisk option we have is using PAE to some extent to allow that memory to be accessed.  I'm just wondering if my memory of the supposed IBM work on PAE in the kernel is correct/incorrect, or if it was tried and just didn't work reliably which is why we now have the ramdisk option (as it does work well).

It depends on what you consider kernel development.  From my point of view most work on the OS/2 kernel was done last in WSeB (Warp Server for Ebusiness).  These kernels also later made it into the Warp 4 fixpaks starting at fixpak 13 (with single core support).  After that things went downhill best I can memory when it came to expanding the functionality of OS/2. Scott Garfunkle I think left his OS/2 kernel development in 2002.
Its hard to say how much enhancements he made to the kernel but from my point of view it was IBM's its policy to only maintain OS/2 under the SWC program. I have never heard anything about PAE support being put in by IBM.

Geting PAE into the kernel without sources based on the people I talked is a big big big job. Consider it practically speakign impossible. Some much would have to modified because you are dealing with memory allocation and memory relase rountines.... When a program wants to allocate memory such a web browser it asks the in the end the kernel for chunk of memory. In so many places adding PAE memory address range in the kernel would be a very big job.

The current PAE support for the ramdisc is depended of how the OS/2 kernel works.

It might however be possible with other technology that LIBC might be able use the memory the ram disc offers. Allocate the memory directly.
But this is still under discussion with BWW.

Roderick

Dave Yeo

  • Hero Member
  • *****
  • Thank You
  • -Given: 13
  • -Receive: 224
  • Posts: 3136
  • Karma: +56/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #46 on: June 02, 2021, 01:57:31 am »
While I never heard anything about IBM trying to support PAE, PAE I believe can have problems with device drivers, at least that is one problem on Windows.
The ramdisk does use PAE and I believe there is an API so other programs could use the memory. Those programs would need to be written to use that memory, compared to now where it is simply malloced (or similar).
Something like a disk cache using PAE might be relatively simple to implement, something like a browser likely not so much.

Lars

  • Hero Member
  • *****
  • Thank You
  • -Given: 2
  • -Receive: 77
  • Posts: 830
  • Karma: +24/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #47 on: June 02, 2021, 11:46:50 am »
The QSINIT API allows you to ask for 512 byte sectors of memory. It can only be used by exactly one app at a time and I doubt that there is some protection against multiple users. Also, you will then lose all ability to use it as a RAM disc. The mem cannot be shared between RAM disc and other use. It's all or nothing.

andreas

  • Newbie
  • *
  • Thank You
  • -Given: 0
  • -Receive: 1
  • Posts: 36
  • Karma: +3/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #48 on: June 08, 2021, 06:44:06 pm »
I know this might not be realistic (as i mentioned before, i'm not a programmer), but i just would like to know if this could be a way to get a 64-bit system:

Can there be some sort of merger between a host and a virtual machine? Maybe similar like WINOS/2 that got part of OS/2?

Maybe we could get a basic new 64-OS with only everything that is necessary to load our 32-bit-OS/2 in some sort of virtual machine in some way that you don't even realize, that we have a virtual machine.
Then we'd also need some sort of connection between that host and the guest - also in some way, that the user wouln't even realize we have actually 2 systems.
Would that be feasible or are there any technical restrictions?
Since i know, we have virtual machines worlking on 64-bit platforms, this might work. I'm just not sure, if an interaction between both worlds can be achieved...

Ibrahim Hakeem

  • Jr. Member
  • **
  • Thank You
  • -Given: 3
  • -Receive: 4
  • Posts: 87
  • Karma: +2/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #49 on: June 08, 2021, 06:47:31 pm »
I know this might not be realistic (as i mentioned before, i'm not a programmer), but i just would like to know if this could be a way to get a 64-bit system:

Can there be some sort of merger between a host and a virtual machine? Maybe similar like WINOS/2 that got part of OS/2?

Maybe we could get a basic new 64-OS with only everything that is necessary to load our 32-bit-OS/2 in some sort of virtual machine in some way that you don't even realize, that we have a virtual machine.
Then we'd also need some sort of connection between that host and the guest - also in some way, that the user wouln't even realize we have actually 2 systems.
Would that be feasible or are there any technical restrictions?
Since i know, we have virtual machines worlking on 64-bit platforms, this might work. I'm just not sure, if an interaction between both worlds can be achieved...

This has been spoken about quite a bit, even in the most recent Warpstock it was stated to be wishful thinking.

andreas

  • Newbie
  • *
  • Thank You
  • -Given: 0
  • -Receive: 1
  • Posts: 36
  • Karma: +3/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #50 on: June 08, 2021, 07:06:36 pm »
This has been spoken about quite a bit, even in the most recent Warpstock it was stated to be wishful thinking.

i know this might be wishful thinking. i just wonder what the obstacles might be.
It was possible to run 16-bit-programs from another OS on OS/2. So i wonder what is the difference now.
Maybe we could even use an existing 64-OS like Linux and adjust it to that concept.

Have there ever been any trials to get OS/2-programs running seamless on Linux? If we achieved that, we would get out of wishful thinking towards a more realistic option. Of course i know that would still be a hard way to go from there though.
« Last Edit: June 08, 2021, 07:15:06 pm by andreas »

Roderick Klein

  • Sr. Member
  • ****
  • Thank You
  • -Given: 0
  • -Receive: 26
  • Posts: 425
  • Karma: +6/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #51 on: June 08, 2021, 09:52:33 pm »
I know this might not be realistic (as i mentioned before, i'm not a programmer), but i just would like to know if this could be a way to get a 64-bit system:

Can there be some sort of merger between a host and a virtual machine? Maybe similar like WINOS/2 that got part of OS/2?

Maybe we could get a basic new 64-OS with only everything that is necessary to load our 32-bit-OS/2 in some sort of virtual machine in some way that you don't even realize, that we have a virtual machine.
Then we'd also need some sort of connection between that host and the guest - also in some way, that the user wouln't even realize we have actually 2 systems.
Would that be feasible or are there any technical restrictions?
Since i know, we have virtual machines worlking on 64-bit platforms, this might work. I'm just not sure, if an interaction between both worlds can be achieved...

This has been spoken about quite a bit, even in the most recent Warpstock it was stated to be wishful thinking.

The discussion has been running for 20 years... Its nothing new the idea. The question how realistic is it and achieve critical mass...

Roderick

Dave Yeo

  • Hero Member
  • *****
  • Thank You
  • -Given: 13
  • -Receive: 224
  • Posts: 3136
  • Karma: +56/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #52 on: June 09, 2021, 01:25:11 am »
This has been spoken about quite a bit, even in the most recent Warpstock it was stated to be wishful thinking.

i know this might be wishful thinking. i just wonder what the obstacles might be.
It was possible to run 16-bit-programs from another OS on OS/2. So i wonder what is the difference now.
Maybe we could even use an existing 64-OS like Linux and adjust it to that concept.

Have there ever been any trials to get OS/2-programs running seamless on Linux? If we achieved that, we would get out of wishful thinking towards a more realistic option. Of course i know that would still be a hard way to go from there though.

There's https://www.patreon.com/posts/2ine-16-bit-exe-19337541 and a couple of other posts in the same thread. This is more along the way WINE works. While possible, there are a lot of differences between any *nix and OS/2. A better course would be to add an OS/2 personality to ReactOS. It is a reimplementation of the NT kernel and NT did start out as OS/2 V3 NT with NT supporting 16 bit OS/2 text apps out of the box and a 16 bit Presentation Manager add on. I also have/had a Byte magazine from the mid 90's with a news article about Microsoft getting the 32 bit Presentation Manager running on NT. The OS/2 subsystem was removed after Win2k.
I understand that at the beginning of ReactOS development, they did reach out to the OS/2 community about supporting OS/2, but didn't get a positive response. OS/2 programmers are/were like cats and herding them is not easy.
At this point, with so few users left, I can't imagine a serious attempt to support OS/2 on any other kernel.
While wishing, what is really a shame is that the OS/2 PPC source seems to be lost, there was a time when releasing it was considered, though I doubt it would have ever happened. Porting it to i386 would have been relatively easy as it was written to be portable, unlike regular OS/2. Still would have had the drivers problem.

Martin Iturbide

  • OS2World NewsMaster
  • Global Moderator
  • Hero Member
  • *****
  • Thank You
  • -Given: 603
  • -Receive: 122
  • Posts: 3286
  • Karma: +27/-0
  • Your Friend Wil Declares...
    • View Profile
    • Martin's Personal Blog
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #53 on: June 09, 2021, 01:48:14 am »
Hi
There's https://www.patreon.com/posts/2ine-16-bit-exe-19337541 and a couple of other posts in the same thread. ....
2ine was my 2018 feel good moment when I saw that one little document that I asked IBM permission to repost was actually used by someone.

Quote
After much head scratching about why it was trying to jump back from 16-bit land to a totally bogus 32-bit code segment, I found an ancient IBM CourseWare document on the Internet Archive that explained this.
  (source)

The dream will be to add "OS/2 personality" to a microkernel project. Instead of suffering for our own kernel, lets grab it from some stronger community that worries about that. Some people says the kernel made the OS, I believe that the "OS experience" is the OS.
I would prefer Zircon, based in LK (made by the same guy from HaikuOS), since it is an open source kernel where Google is investing and creating a community for it.  It will stop being OS/2, since compatibility will be interpreted over a different kernel, but it may open the doors for a better future for the platform.

Regards
Martin Iturbide
OS2World NewsMaster
... just share the dream.

andreas

  • Newbie
  • *
  • Thank You
  • -Given: 0
  • -Receive: 1
  • Posts: 36
  • Karma: +3/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #54 on: June 09, 2021, 01:33:20 pm »

The dream will be to add "OS/2 personality" to a microkernel project. Instead of suffering for our own kernel, lets grab it from some stronger community that worries about that. Some people says the kernel made the OS, I believe that the "OS experience" is the OS.
I would prefer Zircon, based in LK (made by the same guy from HaikuOS), since it is an open source kernel where Google is investing and creating a community for it.  It will stop being OS/2, since compatibility will be interpreted over a different kernel, but it may open the doors for a better future for the platform.

Regards

Martin, i agree with you. I don't really care, where the kernel comes from, as long as it work and I can keep the "OS/2-feeling".
From the technical side i have no clue which kernel is the best. But having Google in the background would be a huge advantage. Maybe they could even revive OS/2 (or whatever they might call it then).
I think in the end it's just a matter of money. Google has it. The OS/2 community doesn't. And it seems ReactOS is pretty the same.

Kékróka Klán

  • Newbie
  • *
  • Thank You
  • -Given: 0
  • -Receive: 0
  • Posts: 5
  • Karma: +0/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #55 on: June 09, 2021, 04:33:57 pm »

The "easy way" of doing this (if there is such a thing) to me seems like if you could wedge a hypervisor under OS/2 (or inside of it as the case my be) so you could run multiple instances of it, as if it were on KVM of ESXi but have it be totally transparent (or mostly transparently) to the user so it is all managed by native ArcaOS / OS/2 tooling in the UI and it appears like you are running one OS (think of the way virtual DOS machines currently work in OS/2).

I totally agree with this approach especially as multicore machines cries out machine partitioning as well.
If legacy 32-bit OS/2 would settle in VMs as DOS now (VOM = Virtual OS/2 Machine), then such old apps can use that, and new environment can be 64-bit, so memory management can move on enabling using physical memory above 4.00 GB.

Storage capacity itself can be resolved with network drives. That way OS/2 users can use any kind of filesystems those for there are no installable filesystem solution actually - e.g. XFS, OpenZFS, etc.
     Also if you want RAID there, no problem if OS/2 does not support natively any raid device -- there you can have it underlying on the other machine. However I rather suggest LVM possibilities (on Linux) to add additional drive / copies of physical partitions when a new LV created for a filesystem. I got raid error on my home PC server once, and I could mirror / sripe on LVM layer more reliable than the risky raid stuff. In case RAID you really need a good, working backup/restore solution.
You don't need to use up many drive letters as you can mount them into directory as well. So the max. storage can be up above 48 GB that way.
 

Dave Yeo

  • Hero Member
  • *****
  • Thank You
  • -Given: 13
  • -Receive: 224
  • Posts: 3136
  • Karma: +56/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #56 on: June 09, 2021, 05:37:03 pm »
The problem is the business case. All I can do is guess at Arca Noae's business, and they're the ones with a license to resell OS/2. According to discussions here, there's only about 500 hobbyists, with not all having the money to spend but even 500 paying customers doesn't seem enough to support Arca Noae which leads to the assumption that they also have a fair number of enterprise customers. Any enterprise customers who can virtualize their environment likely already have, likely back in the eCS days without much need to upgrade. This leaves people/businesses that need real hardware which seems to be what Arca Noae is concentrating on.
There's lots of stuff that just can't be done in a virtual environment, a while back there was someone needed lots of com ports, which didn't work under a virtual machine for example.

Roderick Klein

  • Sr. Member
  • ****
  • Thank You
  • -Given: 0
  • -Receive: 26
  • Posts: 425
  • Karma: +6/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #57 on: June 10, 2021, 12:49:18 am »
The problem is the business case. All I can do is guess at Arca Noae's business, and they're the ones with a license to resell OS/2. According to discussions here, there's only about 500 hobbyists, with not all having the money to spend but even 500 paying customers doesn't seem enough to support Arca Noae which leads to the assumption that they also have a fair number of enterprise customers. Any enterprise customers who can virtualize their environment likely already have, likely back in the eCS days without much need to upgrade. This leaves people/businesses that need real hardware which seems to be what Arca Noae is concentrating on.
There's lots of stuff that just can't be done in a virtual environment, a while back there was someone needed lots of com ports, which didn't work under a virtual machine for example.

I never said that there are 500 users (if you are basing that number on my statement). I think we have a few thousand private users left in total.
I worked at Mensys until 2013 on eCS. I get the impression a lot of the private users sticked around. True users left but I get the impression more users are left based on time at Mensys. I can not imagen we are left with 500 private users.

Roderick

Roderick

Dave Yeo

  • Hero Member
  • *****
  • Thank You
  • -Given: 13
  • -Receive: 224
  • Posts: 3136
  • Karma: +56/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #58 on: June 10, 2021, 02:32:32 am »
OK, thanks for clarifying the number of users. Be interesting to know Arca Noae's numbers, especially regular users though the number of enterprise users would also be interesting.

Roderick Klein

  • Sr. Member
  • ****
  • Thank You
  • -Given: 0
  • -Receive: 26
  • Posts: 425
  • Karma: +6/-0
    • View Profile
Re: Well ArcaOS 5.0 is probably the last version for me.
« Reply #59 on: June 10, 2021, 03:39:13 pm »
The problem is the business case. All I can do is guess at Arca Noae's business, and they're the ones with a license to resell OS/2. According to discussions here, there's only about 500 hobbyists, with not all having the money to spend but even 500 paying customers doesn't seem enough to support Arca Noae which leads to the assumption that they also have a fair number of enterprise customers. Any enterprise customers who can virtualize their environment likely already have, likely back in the eCS days without much need to upgrade. This leaves people/businesses that need real hardware which seems to be what Arca Noae is concentrating on.
There's lots of stuff that just can't be done in a virtual environment, a while back there was someone needed lots of com ports, which didn't work under a virtual machine for example.

I never said that there are 500 users (if you are basing that number on my statement). I think we have a few thousand private users left in total.
I worked at Mensys until 2013 on eCS. I get the impression a lot of the private users sticked around. True users left but I get the impression more users are left based on time at Mensys. I can not imagen we are left with 500 private users.

Roderick

Again I am looking for numbers to estimate how big our community roughly but it will always be a gamble to a large extend. It was not much different when I worked at Mensys that people want to buy OS/2/eCS after many years again.

The 500 users was based on the IP numbers from Netlabs. But how accurate that is ? But I never stated we have 500 private users left. Read the other postings for the rest of the details.

Roderick