Constant reboot w ACPI 3.14 and /APIC switch

[Peter Kruithof]

Mb ASAS P5K-WS - ecom 2.0 RC5 -U160 SCSI - JFS bootable Works well as long as i do not use /APIC Previous versions only hangs the system during LOGO display Tried start from CD and check disk Acpi.psd is the first statement in config unable to collect logs with /APIC (used debug and moved the logo) logs attached with no switch and with /SMP

[eco]

e-co:

you have SCSI adapter. so you can use SMP PIC only

​http://ecomstation.ru/acpitools -> Documents -> ACPI and SCSI

[Peter Kruithof]

Right ,i have read docs earlier but did not quite got the right picture.

But now i have temperary installed a SATA disk ,took the SCSI out and booted .....

Now i have the exact problem ,restart ,restart ,etc

On this disk is ecom RC6a

????

[eco]

e-co:

Can you give more information? When does the computer reboots/resets/restart? in which moment? what do you see? Do you see desktop? etc.

[eco]

e-co:

ecs RC5 - OK, ecs RC6a - problems?

Do you talk about shutdown problem? read more how to setup eWorkPlace: http://svn.netlabs.org/acpi/ticket/342

[Peter Kruithof]

Replying to eco:

e-co:

Can you give more information? When does the computer reboots/resets/restart? in which moment? what do you see? Do you see desktop? etc.

Reboot occors just after the upper-left "ecs" appears ,there is just enough time to press atf-F1 However when leaving it restart over and over again many times ,it sometimes goes a little further and the ECS logo appears but then only hard reset helps I will make a detaled report later on the shutdown issue

[Peter Kruithof]

Replying to eco:

e-co:

ecs RC5 - OK, ecs RC6a - problems?

Do you talk about shutdown problem? read more how to setup eWorkPlace: http://svn.netlabs.org/acpi/ticket/342

Here is an extensive report regarding shutdown problem (RC6a): Shutdown has inproved in RC6 but not all is o.k Reboot is in all cases now working , in RC5 and lower if failed with the never dissapearing "releasing filesystems " box while shutdown/poweroff worked in some combinations of settings.

I have now tested the following : ecom RC6 with acpi 3.14 ,checked "power offf" and "delay" - used eshutdown

---

reboot o.k. (3 times tested) shutdown ,power off :

1st time -hang - empty desktop forever 2rd time - o.k. 3rd time hang - empty desktop forever 4th time hang - empty desktop forever etc ..... all cases the message "releasing filesystems" dissapeared but i think it left the FS flag dirty -not sure though-

---

settings box "animate before shutdown" not checked and "before reboot" checked - result:

reboot o.k. (3 times tested) shutdown ,power off : 1st time - o.k 2nd time -hang - empty desktop forever 3rd time - o.k. 4th time hang - empty desktop forever 5th time - o.k. 6th time - o.k etc .....

---

settings box "animate before shutdown" checked and "before reboot" not checked - result: reboot o.k. (3 times tested) shutdown ,power off : all tries "message "releasing filesystems" stay on forever

---

settings box "animate before shutdown" checked and "before reboot" checked - result: reboot o.k. (3 times tested) shutdown ,power off : all tries "message "releasing filesystems" stay on forever

In the situations where power off works ,it not tuns the computer completely off , the USB power stays on.

Anything else i can test ?

[eco]

e-co:

"SMP APIC, infinity reboot"

Can you retell us BIOS setup settings?

[Peter Kruithof]

Replying to eco:

e-co:

"SMP APIC, infinity reboot"

Can you retell us BIOS setup settings?

Pictures of bios added

[Peter Kruithof]

The pictures are now here (files too large to put here) : ​http://www.microdrome.nl/biosscreen.zip

[eco]

e-co:

IMHO.. No critical settings in BIOS Setup..

Please check the kernel. OS2KRNL + DOSCALL1.DLL from one package.

Does this script find problems? ​http://ecomstation.ru/projects/acpitools/download/acpi-test.cmd

[eco]

e-co:

You have many USB devices, please try detach all devices

[Peter Kruithof]

output of acpitest : Test eComStation configuration, is it ready for ACPI.PSD? (c) 2008 eCo Software Version of script: 3.13

Test 1 - Checking snoop.lst Test 1, PROBLEM FOUND: L:\OS2\BOOT\SNOOP.LST, Keep ibmkbd.snp, remove other snoopers

Test 2 - Checking CONFIG.SYS Test 2 - OK

Test 3 - Checking RESOURCE.SYS Test 3 - PROBLEM FOUND: Wrong RESOURCE.SYS (sub-test 2)

Test 4 - Checking OS2KRNL Test 4 - OK, the system is equipped with supported version of kernel

14.104a#@_SMP

Test 5 - Checking danis506 Test 5 - OK, the system is equipped with modern DANIS506.ADD

Found version: 1.8.5

Test 6 - Checking duplicates Test 6 - OK, can't find duplicates of ACPI.PSD on bootable disk

Test 7 - Checking ACPI files version Test 7 - PROBLEM FOUND L:\OS2\BOOT\ACPI.PSD ver. 3.14 , should be 3.13 Test 7 - PROBLEM FOUND L:\OS2\BOOT\APM.ADD ver. 1.28 , should be 1.27

Removing usb devices or even disabling it in the bios makes no difference What do you mean by : OS2KRNL + DOSCALL1.DLL from one package. ? They come from the ecom RC6a cd

[ecsnl]

A more logical approach, lets install APCI.PSD debug version, debug kernel and remove os2logo, then we have SOME output. The moment the machine crashes could you please take a picture of the screen and post the results here.

I send the customer an email with more details.

[ecsnl]

Using the debug kernel is more preferred.

[Peter Kruithof]

o.k. here is the output from the debug (no screen output though) so i captured the serial com1 output:

OS/2

(C) Copyright IBM Corp. 1987, 2001. All rights reserved.

US Government Users Restricted Rights - Use, duplication or disclosure

restricted by GSA ADP Schedule Contract with IBM Corp.

Internal revision 14.104a_SMP Symbols linked (acpi)

System Debugger 8/8/97 [80?86]

54EB 4990 4D42 3420 352E 0030 0002 0000 0000 0000 F800 0000 003F 00FF 003F 0000 8A82 12A1 8980 BF29 57D5 5651 4C4F 4D55 2045 2038 2020 464A 2053 2020 2020 0000 0000 0000 0000 0010 0020 0000 0000 0000 0000 0000 0000 33FA 8EC0 BCD0 7C00 BBFB 07C0 C38E CDFB 2D12 0054 F025 C1FF 06E0 D88E 3366 66FF DB33 2666 1CA1 6600 0126 3E06 6600 1126 421E 1E00 20B8 3300 E8F6 005A 661F A126 003E A366 003E 2666 42A1 6600 42A3 2600 24A0 A200 0024 3E81 0200 1961 0E74 C0B8 8E07 A1D8 0200 C8BE E800 00A0 B81E 199C CB50 6E49 6176 696C 2064 6F63 6564 6620 726F 4A20 5346 0020 202D 5953 3053 3032 3732 2D20 0020 521E DA8C 1F06 48A3 8900 4C16 8900 4A36 BE00 0046 A166 003E A366 004E A166 0042 A366 0052 0166 4E3E 6600 1E11 0052 42B4 168A 0024 00B0 13CD 0473 E40A 0375 1F5A 50C3 A166 0052 C166 10E8 5DE8 6600 52A1 E800 0056 A166 004E C166 10E8 4BE8 6600 4EA1 E800 0044 48A1 6600 E0C1 5810 C28A DEBE E800 0000 66FC AC50 FFA8 0974 0EB4 07BB CD00 EB10 FBF2 5866 5066 2566 0000 FFFF C166 10E8 11E8 B000 B43A BB0E 0007 10CD 5866 03E8 EB00 C3FE 8A50 24C4 B1F0 D204 E8E8 001D 5058 C48A 0F24 14E8 5800 2450 B1F0 D204 E8E8 0009 5058 0F24 02E8 5800 04C3 3C30 7E39 0402 B407 BB0E 0007 10CD 4FC3 3253 444C 0052 534F 4232 4F4F 0054 6328 2029 2E50 6853 6574 656D 6B6E 206F 3032 3230 322C 3030 0034 0000 0000 0000 AA55 MaxProcesses = 1025 Symbols linked (os2krnl)

[pasha]

place kdb.ini to root of boot drive . Add /O1 to acpi.psd. Better, if acpi.sym copy to \os2\boot too.

Setting for terminal:

115200,n,8,1

[Peter Kruithof]

result of debug /O1

[Peter Kruithof]

result added

[pasha]

Ok. I see where is problem. Need day two for resolving

[pasha]

try acpi20012008.zip. Attach log in any results

[eco]

e-co:

How to download special build? ​http://ecomstation.ru/projects/acpitools/?action=testcase

[Peter Kruithof]

result acpi 20012008

[Peter Kruithof]

tested the acpi 20012008 , no boot as we hoped for ,picture added , used switch /smp /apic and debug kernal

[pasha]

acpi21012009.zip, need log. If trap is here, install pls debug kernel. I need in full message at screen

[Peter Kruithof]

acpi result working with /smp /apic

[Peter Kruithof]

Looking good ,full startup with high irq's , logs attached , great work !!!

pls let me know if there is more to test

Perhaps SCSI later on as well ?? very impotant to me , is remap working ?

[pasha]

What is driver for SCSI?

[Peter Kruithof]

it a adaptec 39160 card ,driver is the AICU160.add

[pasha]

Do want boot from it? If no - it don't need to remap. If yes, you can try, but I afraid will conflict with USB

[Peter Kruithof]

I do not boot from scsi at this test , but without remap it goes in a loop at startup

When using remap it boots o.k. but the network card cannot be used.

[eco]

e-co:

correct me if I wrong,

Ticket #362, SVN revision: 167

Pasha,

Again, we don't have notes what is the problem? how it was fixed?

[pasha]

Please download experimental ACPI build from Mensys site:

• Experimental build for you:

ACPI-APIC-BAT.ZIP

• Read instructions: ​http://ecomstation.ru/projects/acpitools/?action=testcase

[Peter Kruithof]

After trying the exprimental build ,i got the orginal problem back - constant reboot...

[pasha]

acpi log? Better with debug kernel

[Peter Kruithof]

Sorry for the delay ,but i had to reinstall due to some other experiments which failed.

I am not sure if i have done all as before but with the debug kernel i get no results at all.

No output on screeen nor at the serial port.

??

[pasha]

check comport number, by default kernel use com2. Or getting new loader from hobbes.nmsu.edu and write to os2ldr.ini.

[Peter Kruithof]

Just a question ,pasha : what is different between the last test version and the one at which acpi seems to work o.k. ?

[Steven Levine]

Pasha seems to not be able to answer this kind of question. See my comment to ticket #313. This is the major change. Pasha also integrated a new version of the ACPICA sources from Intel. This fixed some thing and broke some things.

[Peter Kruithof]

Allright ,i the meantime i have discovered that things did go wrong while testing after my reinstall of ecom RC6a.

Now i have found out that the acpi actually works ,for unknown reason the testing failed.

Logs attached

[Peter Kruithof]

working with /apic and /smp acpi file name:ACPI-APIC-BAT.ZIP

[Steven Levine]

Please verify that 3.15, when it is released, continues to work on this system. Thanks.