• Welcome to OS2World OLD-STATIC-BACKUP Forum.
 

News:

This is an old OS2World backup forum for reference only. IT IS READ ONLY!!!

If you need help with OS/2 - eComStation visit http://www.os2world.com/forum

Main Menu

New kernel

Started by AAA, 2008.11.07, 17:05:07

Previous topic - Next topic

mobybrick

I think that is because 'AAA' posted here about the replacement OS2LDR before it became officially available, and some of us suspect that 'AAA' is the name of a Russian developer by the name of Pavel.

Let's hope, anyway!

Regards,
Moby.

Quote from: warpcafe on 2008.11.10, 21:38:23
Hi all,

I agree with Fahrvenugen... and I am quite shocked to see that apparently nobody has asked before posting their wishes: Where have you got this information from, AAA ?

Cheers,
Thomas

mobybrick

Thinking about it, another new feature (probably as longer term) would be an INI file on the boot drive that contains extensions to the PATH, LIBPATH statements and any other settings you want to set.

A command could then be issued to ask the kernel to re-read this INI file, and append the values contained within to the settings held within CONFIG.SYS for all *new* processes that are started (changing the PSP of processes already running would be dangerous) - although obviously the changes to PATH and LIBPATH would need to be reflected in all DosExecPgm and WinLoad... calls straight away.

This would allow new apps to be installed without the need for re-boots - one of the weaknesses that OS/2 has lived with since the beginning.

Of course, using Run!, sometimes this issue can be worked around - but for others, not being able to install apps without a reboot is a right pain IMHO.

Regards,
Moby

Criguada

I vote for being able to kill those nasty unkillable processes (you know... the "exiting thread 1" issue).

Bye
Cris

Criguada

Hi Blue Warper!

Quote from: The Blue Warper on 2008.11.08, 12:07:11
1) Fixing the SIQ problem at last.  I wonder why this wish wasn't mentioned in any of the previous posts...  Yes, Warp4 FP15 is more stable than Warp4 GA, and eCS is in general more stable than Warp4 FP15 (never had IBM OS/2 Warp 4.52 (MCP) though).  But we should get rid of this somehow.  I once read this shouldn't be too hard to fix or implement.

This is not a problem in the kernel. It can't be fixed in the kernel. And BTW, I haven't seen a SIQ problem in years. A lot of ppl are confusing this issue with the "unkillable process" issue.

Quote from: The Blue Warper on 2008.11.08, 12:07:11
3) The ability to restart PM (Presentation Manager) without rebooting.

Again this is not a problem in the kernel

Quote from: The Blue Warper on 2008.11.08, 12:07:11
My best wishes for your efforts!

I agree with the rest of the points made!

Bye
Cris

Blonde Guy

Right now, the PATH is limited to 1024 chars before OS/2 starts to have problems. Either fix the problems, or truncate the PATH.
Expert Consulting for OS/2 and eComStation

AAA

If somebody want to test ftp://Linking to this site is not permitted on our forums/os2krnlSVN1009.zip ;)

chennecke

This sure looks interesting.

Could you tell us a bit about the technical background? Depending on how this was done, I'm afraid of license issues that would unfortunately make using this nice piece of work a no-go for many people.

mobybrick

If we're all licensed OS/2 or eCS users, what is the no-go problem?

Thanks,
Moby

chennecke

Quote from: mobybrick on 2008.12.14, 16:00:03
If we're all licensed OS/2 or eCS users, what is the no-go problem?

If the new kernel is based on leaked source code, it's illegal to use it.

mobybrick

Quote from: chennecke on 2008.12.14, 17:34:46
Quote from: mobybrick on 2008.12.14, 16:00:03
If we're all licensed OS/2 or eCS users, what is the no-go problem?

If the new kernel is based on leaked source code, it's illegal to use it.

Let's hope that this is not the case. As I see it, there are three possible ways this might have been done:-

1. Access to the original source code. Parts of the original W4 source code was leaked once apparently - but this kernel is light years away from the 104a kernel this OS4kernel claims to be compatible with.

2. A new kernel re-written from scratch. But this is a huge feat and is unlikely - you simply wouldn't build a new kernel with all of the 'old' bugs - which according to the readme still need to be removed with legacy code cleanup.

3. A reverse engineer or decompile - theoretically possible with enough computing power, the debug kernel and the symbol files.

It is also theoretically possible to perform binary patching of the existing kernel, with no access to the source code - but there would be limits as to what could be modified - and the readme to this project does have a source code that it is using.

Any use of illegally obtained source code would render the files commercially and morally useless, at least. What worries me is that the source code for this newly built kernel is closed source - do the authors have something to hide?

Let's hope AAA can put our minds at rest. The work to pre-load critical boot files for AMD chipset ACPI support looks fantastic. If they can fix some of the other urgent problems, such as the 32-bit-only LBA boot support (which may also need changes to OS2DASD.DMD), then that would be brilliant.

Regards,
Moby.

RobertM

Quote from: mobybrick on 2008.12.14, 18:41:54
...

Any use of illegally obtained source code would render the files commercially and morally useless, at least. What worries me is that the source code for this newly built kernel is closed source - do the authors have something to hide?
...


One possibility is it's for an official IBM sanctioned project (ie: perhaps to fill a business requirement of one of their customers), and thus the requirement for everyone to have a copy of eCS/Warp to use it (ie: may not be just a technical reason). It wouldn't be the first project paid for by some IBM customer that IBM and/or others had a hand in writing/fixing/completing and then was subsequently released as an update...

As you said, hopefully we will get clarification at some point. Though, it is also quite possible that for contractual reasons, we wont at this point.

Rob


|
|
Kirk's 5 Year Mission Continues at:
Star Trek New Voyages
|
|


rwklein

With my technicall skill set I'm having real doubts if this kernel is legal. And it would certainly be usefull if they would clarify that.
But I have a very hard time believing that based on DDK code they wrote this kernel and that for first time it works so good.

Besides this its a debug kernel from IBM it seems:

I find this in kernel:

- set COM baud rate/port addr (1 = COM1, 2 = COM2) and more stuff you  get to see in PMDF when you use the kernel debugger. So thats very good kernel to even release for testing (unless the disabled the code). But when you load a debug kernel and it finds and INT 3 it will wait to jump to the seriel port specified in KDB.INI.


Pete

Quote from: AAA on 2008.12.14, 12:42:50
If somebody want to test ftp://Linking to this site is not permitted on our forums/os2krnlSVN1009.zip ;)


Hi AAA

While everyone else is discussing the legality I have downloaded and installed the kernel package.

Seems fairly stable and has not caused any problems in the 6 or so hours it has been in use.

If you want a log sent somewhere then it might be an idea if the url existed; I tried the url in the krnl_en.txt file  http://Linking to this site is not permitted on our forums/tracker/ and got a 404 - probably in Russian.

Keep up the good work

Pete


lpino

This is great news. I hope they continue their work on the updated kernel (it's not really a new kernel). Probably this is just the tip of the iceberg because one thing is to have the source and another is to be able to compile it. Consider all the dependencies, tools and missing parts, it would be like the movie "Jurasic Park", trying to put together the DNA sequence of an extinct dinasour (only on this case is a piece of software).

Well I hope this is the begining of something,,, whatever,, but something :)

djcaetano

#29
Quote from: AAA on 2008.12.14, 12:42:50
If somebody want to test ftp://Linking to this site is not permitted on our forums/os2krnlSVN1009.zip ;)

  These are great news!
  I tested the new kernel, but my results are not-so-good.
  First of all, I would like to tell I am running eCS RC4 using SMP mode, in a Core2Quad machine, 2GB of RAM. The motherboard is a GA-G31M-S2L. I am using Panorama with the onboard Intel video. Network is an RTL8139.
  Also, I would like to tell eCS RC4 kernel usually hangs on boot, just before it executes the RUN statements. After some resets (three or four, usually) it boots ok.
  Well, the behavior with this new kernel is pretty much the same, besides the fact it *always* hangs in that place, no matter what.

  I was only able to boot to command line, where RTL8139 driver is not loaded, but I have not checked if it is the faulty one. I will do the test right now.

***UPDATE***

  I tested the system without RTL8139 driver, but the results were the same. Tried removing several drivers... but it was a no-go.
  Removing the "SMP" parameter on ACPI.PSD (causing the system to boot using only one processor, but still using SMP doscall1), the system boot "ok", but when I loaded OpenChat/2... the system hung... which is not normal. OC/2 has know problems when running which more than one processor (it will not always open without generating a SYS3175), but never presented any problems running on single processor (even with SMP kernels).
   Just to complete the system description, each core is running at 2.4GHz, and I am running ACPI 3.11 and APM 1.26 (they are somewhat old, but software subscription has ended and I had not updated it yet.

*** UPDATE2 ***
 
   I was just thinking... it would be nice to move some Kernel configuration parameters from CONFIG.SYS to OS4KRNL.INI, such as CLOCKSCALE, BUFFERS, EARLYMEMINIT...