Author Topic: OS/4 kernel and extended KEE  (Read 95994 times)

Lars

  • Hero Member
  • *****
  • Posts: 1271
  • Karma: +65/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #45 on: January 07, 2014, 08:26:00 am »
I got the SVN revision 4160 (assuming it's the most available up to date). I started off with a system that had the 14.104a_SMP debug kernel running (and therefore the SMP version of DOSCALL1.DLL installed) and also ACPI.PSD installed:

1) I copied OS4LOGO, OS4KRNL, OS4KRNLR, OS4KRNL.SYM, OS4KRNL.INI, OS2LDR, OS2LDR.INI to root directory

2) I copied CLOCK03B.SY? to \OS2\BOOT

3) I copied OS4APIC.PSD and OS4APIC.SYM to \OS2\BOOT

4) initially I did not copy DOSCALL1.DLL to \OS2\DLL but I tried later

5) I modified OS2LDR.INI (only the changed parts):
;dbport=0x3f8
;ctrlc=1
;fullcable=1
preload = 1
[kernel]
os2krnl      =  OS/2 debug,ALTF2
os4krnl      =  OS/4 debug using config.os4,LOGSIZE=4096,PRELOAD,CFGEXT=OS4,ALTF2,CTRLC
os4krnlr     =  OS/4 retail using config.os4,LOGSIZE=4096,PRELOAD,CFGEXT=OS4,ALTF2

6) in root dir I created a CONFIG.OS4, I removed all of ACPI.PSD, added OS4APIC.PSD /APIC, removed TIMER01.SYS, added BASEDEV=CLOCK03B.SYS

The result was an IPE for os4krnl as well as os4krnlr and a hang for os2krnl. It was the very same after I also replaced DOSCALL1.DLL with the one from the OS4 zip file.

What am I doing wrong ? I took a dump when the IPE for the os4krnlr happened (I have TRAPDUMP=R0,X: set in config.sys or config.os4, however this did not work for the os4krnl) but PMDF will not even let me read the created dump. Looks like PMDF cannot find a proper kernel revision string in the dump, after looking into the dump I identified it to be 15.1_SMP.4160 but I am not sure about that either.
Not being able to read a dump file is a severe limitation for me, I use this system as a MUT for device driver debugging.
By the way, does the debug kernel (os4krnl) support:
1) the full KDB command set ?
2) also work with UDP instead of serial interface ?


Lars

OS4User

  • Sr. Member
  • ****
  • Posts: 406
  • Karma: +10/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #46 on: January 07, 2014, 08:59:24 am »

added BASEDEV=CLOCK03B.SYS


you should not add  and you may leave TIMER0 in your config.sys

but it is not a big problem  - you will just have a  warrning

TRAPDUMP

probably it is no more supported in os/4

By the way, does the debug kernel (os4krnl) support:
1) the full KDB command set ?
2) also work with UDP instead of serial interface ?

1) yes

2) I do not know if anybody test it.


It would be great if you can show comport log.
« Last Edit: January 07, 2014, 09:05:56 am by OS4User »

gamba

  • Guest
Re: OS/4 kernel and extended KEE
« Reply #47 on: January 07, 2014, 09:11:30 am »
Following Doug's suggestion, I tried ACPI.PSD /VW instead of OS2APIC: now I have 3,5 Gb of RAM, 2 processors running and APM. A full boot to WPS takes less than 30 seconds...

Actually, I think that /VW was Lars' suggestion, but it was a good suggestion.

Thanks to Lars then! (I'm getting lost in these threads :-)  And of course a special thanks to walking_x: I'm really impressed by QSINIT!


Quote

Quote
if just a wireless driver existed :-(

You said you have an Intel device. Apparently, Intel will be the first on the list when Multimac gets around to producing a WiFi driver.

I know. Let's hope.

Lars

  • Hero Member
  • *****
  • Posts: 1271
  • Karma: +65/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #48 on: January 07, 2014, 09:17:58 am »

By the way, does the debug kernel (os4krnl) support:
1) the full KDB command set ?
2) also work with UDP instead of serial interface ?

1) yes

2) I do not know if anybody test it.


It would be great if you can show comport log.

I do have a com port on this machine, unfortunately I don't have a com port on any other of my machines available. Therefore I cannot create a log.
That's why I am using ICAT via UDP and not via serial interface ...
ICAT via UDP DOES have a window to capture all KDB output, unfortunately it is limited in the way that the system must be fully up and running (network driver loaded, netbind.exe loaded ...)

Lars

OS4User

  • Sr. Member
  • ****
  • Posts: 406
  • Karma: +10/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #49 on: January 07, 2014, 09:26:01 am »

I do have a com port on this machine, unfortunately I don't have a com port on any other of my machines available. Therefore I cannot create a log.
Lars

you may try to use USB->COM adapter  for machine  with terminal

comport log is essential for such cases

Lars

  • Hero Member
  • *****
  • Posts: 1271
  • Karma: +65/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #50 on: January 07, 2014, 09:42:38 am »
Forgot to say:
1) I have IBM bootmanager installed as I have 2 OS/2 installations on this machine: 1 for debugging, 1 for normal use, both of which use the SMP kernel (debug and retail respectively). I replaced OS2LDR on the debugging partition.

2) the debugging partition where I tried to run OS/4 kernel is formatted to HPFS

OS4User

  • Sr. Member
  • ****
  • Posts: 406
  • Karma: +10/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #51 on: January 07, 2014, 09:58:21 am »
Forgot to say:
1) I have IBM bootmanager installed as I have 2 OS/2 installations on this machine: 1 for debugging, 1 for normal use, both of which use the SMP kernel (debug and retail respectively). I replaced OS2LDR on the debugging partition.

2) the debugging partition where I tried to run OS/4 kernel is formatted to HPFS

nothing special   -  should work

walking_x

  • Full Member
  • ***
  • Posts: 101
  • Karma: +0/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #52 on: January 07, 2014, 02:40:03 pm »
Interestingly, the new boot manager (Air Boot) has a feature where it will power off the machine, by pressing DEL at the menu. That works on all of my systems, except the L530, where it does nothing. I expect that it is related to the oddities found in QSINIT.
Yes, I am using APM to turn off PC, but APM removed from EFI BIOSes (or most of them). Another method to turn off - ACPI, but I avoid of touching it now - to avoid new problems ;)

Quote
I hope the ACPI problem can be resolved.
I hope too :)

Quote
You can have it back, if you want it.   ;D
I think, it will return itself, after this vacation :)
« Last Edit: January 07, 2014, 03:43:12 pm by walking_x »

Doug Bissett

  • Hero Member
  • *****
  • Posts: 1593
  • Karma: +4/-2
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #53 on: January 07, 2014, 07:10:25 pm »
Quote
Yes, I am using APM to turn off PC

QSINIT shell does power the machine off (using the command). It is Air-Boot that doesn't (perhaps it does sometimes). That seems to follow what I see in QSINIT, where things like Alt-X won't always work. I just wanted to point out that your software isn't the only one affected.

Is there something, in the shell, that will trace keyboard scan codes? That may be useful information.

ACPI  3.22.03 is working well with the /VW parameter.

I have a couple of older machines, where I can try QSINIT. It is impossible to use the RAMDISK (not enough memory), but the loader should work.

I did try the loader and the RAMDISK on my main machine (Asus M3A78-EM motherboard, with quad core AMD Phenom processor). It currently has 3 GB of memory, so all of the RAMDISK came from that. It works fine. I have another memory stick that will work in it, but I don't remember if there is another slot for it.

So much to do, and so little time.   :(

walking_x

  • Full Member
  • ***
  • Posts: 101
  • Karma: +0/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #54 on: January 07, 2014, 07:25:34 pm »
I'm afraid it is a garbage in char code for Alt-X keys....
For F1..F10 i can clear it, but Alt-X with non-zero char code is X key itself.
You can create menu with 2 kernels and press keys in it. In "kernel selection" menu key codes are printed to log.

Doug Bissett

  • Hero Member
  • *****
  • Posts: 1593
  • Karma: +4/-2
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #55 on: January 07, 2014, 08:59:24 pm »
Quote
QSINIT shell does power the machine off (using the command)

I take it back. That doesn't work on the L530 (UEFI), or the T510.

Quote
You can create menu with 2 kernels and press keys in it. In "kernel selection" menu key codes are printed to log.

I am not sure what you mean here, but it isn't important. You seem to understand what the problem is.

Now, more snow to move, and then more testing, but I am running out of things to test.   :)

Doug Bissett

  • Hero Member
  • *****
  • Posts: 1593
  • Karma: +4/-2
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #56 on: January 07, 2014, 09:22:00 pm »
Quote
Quote

    QSINIT shell does power the machine off (using the command)


I take it back. That doesn't work on the L530 (UEFI), or the T510.

Again, I take part of that back. It is the pwroff command that does not work on either system (where did I get that command from??? Now I don't see it in the help, but it still pops up a window). Power off (reply y) does power off the T510. The L530 states that there is no APM.

Sigurd Fastenrath

  • Hero Member
  • *****
  • Posts: 576
  • Karma: +27/-0
  • OS/2 Versus Hardware - Maximum Warp!
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #57 on: January 07, 2014, 09:49:02 pm »
It is so interesting to follow this thread!

Encouraged by the success of Doug I started to change my Warp 4/eCS 2.2beta2 clone and add QSINIT. I customized OS2LDR.INI so that I can choose between 3 different kernel

14.106SMP
14.104aSMP
OS/4

and with three different ways in addition for each: "normal boot"; "Boot with drivers listed ALTF2", "Boot to recovery console ALTF1"

-> so alltogether 9 entries in the boot menue.

In addition I have entries for the OS/4 Loader and the eCS Loader, that makes it up to 11 :-)

So I started to play around wich constellation may work.

Is there a way (other than via kernel options) to add different config.sys directly to the boot menue? I need different for i.e. 14106SMP with "Mensys" ACPI, OS/4 kernel with OS/4 ACPI

What I noticed in the short period I was albe to test today:

- when I choose a different loader i.e. OS4 Loader the red OS/4 in the left upper corner shows up, but the menue stays and says "I am waiting for choosing something, otherwise will boot the normal one" -> or is this the right way? Is the Loader then already loaded and do I have to choose the kernel then?

- when using OS/4 kernel and OS4APIC.PSD the system right at the beginning while booting states that the ACPI4.SYS is missing. Unfortunately the systems allways hangs with this constellation, right before FAT32 drivers are loaded. But if I REM the RUN and IFS for Fat32 it still hang....

Well - I have a lot to learn and a lot to test, but it is that impressive to see these Prebootmenues once F9 i pressed....

Well done!

OS4User

  • Sr. Member
  • ****
  • Posts: 406
  • Karma: +10/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #58 on: January 07, 2014, 10:45:01 pm »

- when using OS/4 kernel and OS4APIC.PSD the system right at the beginning while booting states that the ACPI4.SYS is missing. Unfortunately the systems allways hangs with this constellation, right before FAT32 drivers are loaded. But if I REM the RUN and IFS for Fat32 it still hang....


can you show comport log of it ?

walking_x

  • Full Member
  • ***
  • Posts: 101
  • Karma: +0/-0
    • View Profile
Re: OS/4 kernel and extended KEE
« Reply #59 on: January 08, 2014, 11:27:26 am »
It is the pwroff command that does not work on either system (where did I get that command from??? Now I don't see it in the help, but it still pops up a window). Power off (reply y) does power off the T510. The L530 states that there is no APM.
Yes, no APM on EFI :(
But pwroff - is a batch file, it calls the same "power off /q" command - so this is strange.
May be need to add ACPI power off (when you pressing it - there is no requirements in ACPI until next power on ;) and  I can do anything with it).