Opened 12 years ago

Closed 11 years ago

#186 closed defect (unknown)

ACPI 3.05 not work on Intel Rack server

Reported by: eco Owned by: pasha
Priority: major Milestone: Release version 3.13
Component: ACPI PSD Version: 3.06
Keywords: Cc:

Description

Stephen:

I am trying to install eCS R4 on an Intel rack server, equipped with dual Xeon processors.

eComStation 1.2MR installs and boots just fine.

eComStation 2.0RC4 completes install stage 1, but then will not boot. The boot process hangs in a random device driver just after the logo has appeared - sometimes SCREEN01.SYS, sometimtes USBEHCI.SYS, sometimes ISAPNP.SYS etc..

I do not think it is any of these drivers, but merely a timing/random early problem during the boot.

I have tried

ACPI.PSD /!NOD ACPI.PSD /APIC ACPI.PSD /SMP /R /CD ACPI.PSD /R /CD

The BIOS of the server has options for 'Enable extended CPUID functions' but I have not tried this with all combinations of above and am not sure it would make any difference anyway.

Any help you can give would be appreciated.

Attachments (18)

ACPI.LOG (10.2 KB) - added by eco 12 years ago.
Debug1.jpg (70.8 KB) - added by guest 12 years ago.
ACPI.PSD Debug Output
Debug2.jpg (154.2 KB) - added by guest 12 years ago.
ACPI.PSD /SMP Output
Debug3.jpg (49.9 KB) - added by guest 12 years ago.
After removing USBUHCD.SYS driver - still fail
Debug4.jpg (50.9 KB) - added by guest 12 years ago.
After removing *ALL* USB drivers - then fails on disk driver
Debug5.jpg (47.1 KB) - added by guest 12 years ago.
Change of kernel - no difference
Debug6.jpg (49.3 KB) - added by guest 12 years ago.
After DOSCALL1.DLL Change - No Difference in Behaviour
Debug8.jpg (46.9 KB) - added by guest 12 years ago.
After Chuck's DOSCALL1.DLL - The situation has not improved :)
Debug9.jpg (48.9 KB) - added by guest 12 years ago.
After 'Enable Extended CPUID Functions' in BIOS - seems to get a little further, but still hang in random driver
acpi-kdb1.txt (13.9 KB) - added by guest 12 years ago.
KDB Output - hang on 'Loading TIMER0.SYS'
acpi-306.log (13.9 KB) - added by guest 12 years ago.
Terminal output of 3.06
acpi-2601.log (7.1 KB) - added by guest 12 years ago.
Trap 14 in latest build
acpi-2601-01.log (14.0 KB) - added by guest 12 years ago.
Back to hang whilst load ISAPNP.SNP
acpi2601-01-nonod.log (10.9 KB) - added by guest 12 years ago.
/SMP /!NOD /O1 - Hang Loading PCIBUS.SNP
acpi2601-01-fs-nonod.log (245.2 KB) - added by guest 12 years ago.
/SMP /!NOD /FS /O1 - Progress, hang at UNMASK PIC
acpi2601-01-fs-apic.log (151.0 KB) - added by guest 12 years ago.
/SMP /APIC /FS /O1 - IPE whilst loading USBEHCD.SYS
acpi2601-01-fs.log (248.1 KB) - added by guest 12 years ago.
/SMP /FS /O1 - Hang 'Installed OS2LVM.DMD' / UNMASK PIC
acpi31012008.log (247.9 KB) - added by guest 12 years ago.
Testcase - Hang on UNMASK / Installed OS2LVM.DMD as before

Download all attachments as: .zip

Change History (79)

Changed 12 years ago by eco

comment:1 Changed 12 years ago by guest

Make and model?

A couple of tips:

Go into the BIOS and set all of the PCI IRQ assignments to "Auto select" (if possible).

Make sure that ACPI support is enabled (if possible).

Install using Advanced path, use SNAP for the video driver, disable APM support.

Chuck

Changed 12 years ago by guest

ACPI.PSD Debug Output

comment:2 Changed 12 years ago by guest

Model: Intel SR1400. Dual Xeon 2.8 with 4Gb of RAM.

All IRQ assignments already set to auto and am already using SNAP driver. Have already tried removing APM.ADD from CONFIG.SYS - no difference.

No matter what switches to ACPI.PSD, always early hang in random driver.

Have now uploaded screenshot of ACPI.PSD debug output.

comment:3 Changed 12 years ago by guest

  • Version changed from 1.0 to 3.05

comment:4 Changed 12 years ago by pasha

Boot with:

psd=acpi.psd /SMP, if problem - attach photo of screen

Changed 12 years ago by guest

ACPI.PSD /SMP Output

comment:5 Changed 12 years ago by guest

ACPI.PSD /SMP output attached. None of the various combination of switches seems to work.

See DEBUG2.JPG (sorry large scale) for debug info.

Thanks, Stephen

comment:6 follow-up: Changed 12 years ago by pasha

Now ACPI is right, problem in USb drivers, try rem basedev=USBUHCD.SYS

Changed 12 years ago by guest

After removing USBUHCD.SYS driver - still fail

Changed 12 years ago by guest

After removing *ALL* USB drivers - then fails on disk driver

comment:7 in reply to: ↑ 6 Changed 12 years ago by guest

Replying to pasha:

Now ACPI is right, problem in USb drivers, try rem basedev=USBUHCD.SYS

OK, removed USBHCD.SYS driver. Then fails on USBMSD driver (DEBUG3.JPG) attached. Then failed on next USB driver in CONFIG.SYS - so I removed *ALL* USB driver.

In the boot order, the next driver is the disk controller. And it fails on that.

How is it that all of these drivers work fine (and first time) under eCS 1.2MR reliably, if ACPI is OK as you say? Or are any type of drivers at all no longer allowed or supported?

comment:8 Changed 12 years ago by guest

Uploaded DEBUG4.JPG showing fail on LSIMPT.ADD. Guess that driver is no good as well, yes?

comment:9 Changed 12 years ago by guest

Thanks for any help you can offer Pasha - guess you guys must regret asking for 'Deep testing' ;-))

comment:10 follow-up: Changed 12 years ago by pasha

Get from http://fortuna.odessa.ua:8080/smp20050811.zip , unzip with all override to root os boot disk and try boot.

Changed 12 years ago by guest

Change of kernel - no difference

comment:11 in reply to: ↑ 10 Changed 12 years ago by guest

Thanks Pasha. No improvement, I'm afraid. See attached DEBUG5.JPG

Sometimes it gets further - even as far as LSIMPT.ADD - but usually it hangs on any early driver.

Thanks, Stephen

Replying to pasha:

Get from http://fortuna.odessa.ua:8080/smp20050811.zip , unzip with all override to root os boot disk and try boot.

comment:12 follow-up: Changed 12 years ago by pasha

Check pls doscal1.dll, must be from SMP version

comment:13 in reply to: ↑ 12 Changed 12 years ago by guest

Do you know where I can get a copy which is definately the right one? I guess the one installed is the only put on by the eCS installer... I've no idea which one that is...

Replying to pasha:

Check pls doscal1.dll, must be from SMP version

Changed 12 years ago by guest

After DOSCALL1.DLL Change - No Difference in Behaviour

comment:15 Changed 12 years ago by guest

Hi Pasha,

Copied on the replacement DOSCALL1.DLL (new = 144387 bytes, old = 192100 bytes) but no difference. First boot, hang in LSIMPT.ADD. DEBUG6.JPG attached.

Thanks, Stephen

comment:16 follow-up: Changed 12 years ago by guest

I downloaded and checked the referenced DOSCALL1.DLL. It does not match any that I use to install SMP support using eCS MT (DOSCALL1.DLL is language dependent). I tried to attach the file that I use in eCS MT, but it was too large. It can be downloaded from www.ecomstation.com. Log on and search the downloads for SMP (sm?c005.zip, where ? is a language code).

Is there a problem with LSI SCSI controllers and ACPI? I seem to recall some mention of them in the documentation. See ACPI-SMP-DRV.doc.

Chuck McKinnis?

comment:17 in reply to: ↑ 16 Changed 12 years ago by guest

I think the point is that it hangs in any driver at random - it was just with LSIMPT in the last example above. But in the other attempts, screenshots and examples it hangs in SCREEN01.SYS, USBMSD.SYS, USBUHCD.SYS, ISAPNP.SNP (which I can't find referenced in the config anywhere!), IBM1FLPY.ADD - unless of course, these are *all* incompatible with ACPI.

I'll try another download of DOSCALL1.DLL...

Regards, Stephen

Replying to guest:

I downloaded and checked the referenced DOSCALL1.DLL. It does not match any that I use to install SMP support using eCS MT (DOSCALL1.DLL is language dependent). I tried to attach the file that I use in eCS MT, but it was too large. It can be downloaded from www.ecomstation.com. Log on and search the downloads for SMP (sm?c005.zip, where ? is a language code).

Is there a problem with LSI SCSI controllers and ACPI? I seem to recall some mention of them in the documentation. See ACPI-SMP-DRV.doc.

Chuck McKinnis?

Changed 12 years ago by guest

After Chuck's DOSCALL1.DLL - The situation has not improved :)

comment:18 Changed 12 years ago by guest

DEBUG8.JPG attached showing Chuck's DOSCALL1.DLL attempt. No difference. Random hang in random driver in early start - in IBMKBD.SNP in DEBUG8.JPG.

Changed 12 years ago by guest

After 'Enable Extended CPUID Functions' in BIOS - seems to get a little further, but still hang in random driver

comment:19 Changed 12 years ago by guest

DEBUG9.JPG: After 'Enable Extended CPUID Functions' in BIOS - seems to get a little further, but still hang in random driver - e.g. 'Installing OS2LVM.DMD' or 'Installed OS2LVM.DMD' or 'Loading DANAPTAP.FLT' - i.e. all drivers or managers slightly later in the boot phase.

Regards, Stephen

comment:20 follow-up: Changed 12 years ago by guest

Funny though - whilst it was hung on 'Installed OS2LVM.DMD' as soon as I pressed the PowerButton?, the 'PowerButton?' string was put onto the screen - and for every time I pressed it. It wouldn't power off tho - had to hold in for five seconds :)

comment:21 in reply to: ↑ 20 Changed 12 years ago by pasha

Replying to guest:

Funny though - whilst it was hung on 'Installed OS2LVM.DMD' as soon as I pressed the PowerButton?, the 'PowerButton?' string was put onto the screen - and for every time I pressed it. It wouldn't power off tho - had to hold in for five seconds :)

In debug ACPI each press to Power Button will loging "Power Button". Any operation with this event possible after run AcpuDaemon?.exe

comment:22 follow-up: Changed 12 years ago by pasha

Have you comport cable and second computer for use as terminal?

comment:23 in reply to: ↑ description ; follow-ups: Changed 12 years ago by guest

Replying to eco:

Stephen:

I am trying to install eCS R4 on an Intel rack server, equipped with dual Xeon processors.

eComStation 1.2MR installs and boots just fine.

eComStation 2.0RC4 completes install stage 1, but then will not boot. The boot process hangs in a random device driver just after the logo has appeared - sometimes SCREEN01.SYS, sometimtes USBEHCI.SYS, sometimes ISAPNP.SYS etc..

I do not think it is any of these drivers, but merely a timing/random early problem during the boot.

I have tried

ACPI.PSD /!NOD ACPI.PSD /APIC ACPI.PSD /SMP /R /CD ACPI.PSD /R /CD

The BIOS of the server has options for 'Enable extended CPUID functions' but I have not tried this with all combinations of above and am not sure it would make any difference anyway.

Any help you can give would be appreciated.

Try with changing the data transfer rate of the installation HD in the LSI Bios from 160 MT/sec to lower values.

comment:24 in reply to: ↑ 23 Changed 12 years ago by guest

Originally the server was equipped with a SATA backplane when I first tried. When I came across this problem I switched to a SCSI backplane (by physically changing it). I don't think the type or speed of disks has therefore anything to do with it.

Replying to guest:

Replying to eco:

Stephen:

I am trying to install eCS R4 on an Intel rack server, equipped with dual Xeon processors.

eComStation 1.2MR installs and boots just fine.

eComStation 2.0RC4 completes install stage 1, but then will not boot. The boot process hangs in a random device driver just after the logo has appeared - sometimes SCREEN01.SYS, sometimtes USBEHCI.SYS, sometimes ISAPNP.SYS etc..

I do not think it is any of these drivers, but merely a timing/random early problem during the boot.

I have tried

ACPI.PSD /!NOD ACPI.PSD /APIC ACPI.PSD /SMP /R /CD ACPI.PSD /R /CD

The BIOS of the server has options for 'Enable extended CPUID functions' but I have not tried this with all combinations of above and am not sure it would make any difference anyway.

Any help you can give would be appreciated.

Try with changing the data transfer rate of the installation HD in the LSI Bios from 160 MT/sec to lower values.

comment:25 in reply to: ↑ 23 ; follow-up: Changed 12 years ago by guest

Besides, if it was a SCSI problem, eCS 1.2MR wouldn't boot either with it?

Replying to guest:

Replying to eco:

Stephen:

I am trying to install eCS R4 on an Intel rack server, equipped with dual Xeon processors.

eComStation 1.2MR installs and boots just fine.

eComStation 2.0RC4 completes install stage 1, but then will not boot. The boot process hangs in a random device driver just after the logo has appeared - sometimes SCREEN01.SYS, sometimtes USBEHCI.SYS, sometimes ISAPNP.SYS etc..

I do not think it is any of these drivers, but merely a timing/random early problem during the boot.

I have tried

ACPI.PSD /!NOD ACPI.PSD /APIC ACPI.PSD /SMP /R /CD ACPI.PSD /R /CD

The BIOS of the server has options for 'Enable extended CPUID functions' but I have not tried this with all combinations of above and am not sure it would make any difference anyway.

Any help you can give would be appreciated.

Try with changing the data transfer rate of the installation HD in the LSI Bios from 160 MT/sec to lower values.

comment:26 in reply to: ↑ 25 Changed 12 years ago by guest

Replying to guest:

Besides, if it was a SCSI problem, eCS 1.2MR wouldn't boot either with it?

Replying to guest:

Replying to eco:

Stephen:

I am trying to install eCS R4 on an Intel rack server, equipped with dual Xeon processors.

eComStation 1.2MR installs and boots just fine.

eComStation 2.0RC4 completes install stage 1, but then will not boot. The boot process hangs in a random device driver just after the logo has appeared - sometimes SCREEN01.SYS, sometimtes USBEHCI.SYS, sometimes ISAPNP.SYS etc..

I do not think it is any of these drivers, but merely a timing/random early problem during the boot.

I have tried

ACPI.PSD /!NOD ACPI.PSD /APIC ACPI.PSD /SMP /R /CD ACPI.PSD /R /CD

The BIOS of the server has options for 'Enable extended CPUID functions' but I have not tried this with all combinations of above and am not sure it would make any difference anyway.

Any help you can give would be appreciated.

Try with changing the data transfer rate of the installation HD in the LSI Bios from 160 MT/sec to lower values.

With my motherboard Supermicro PDSGE, processor Intel Pentium D 940, U320-HBA hp DZ554A with chip LSI20320A, HD Hitachi HUS151436VL3600, driver LSIMPT.ADD a fault-free installation is only possible with reduced SCSI-speed.

comment:27 in reply to: ↑ 22 ; follow-up: Changed 12 years ago by guest

What baud/bits/stop/parity will the terminal output on? The server has a COM1... just got to try and get the right cable for it (it is RJ45 serial)...

Replying to pasha:

Have you comport cable and second computer for use as terminal?

comment:28 in reply to: ↑ 27 Changed 12 years ago by pasha

Replying to guest:

What baud/bits/stop/parity will the terminal output on? The server has a COM1... just got to try and get the right cable for it (it is RJ45 serial)...

Replying to pasha:

Have you comport cable and second computer for use as terminal?

115200,n,1 with HW flow control if you have >3 wire comport cable. Also need replace kernel to debug kernel http://fortuna.odessa.ua:8080/smp20050811d.zip, then create file kdb.ini at root of boot disk with:

.b 115200t

G

ln

If you don;t see trap in terminal, try press CtrlC in terminal window

comment:29 Changed 12 years ago by guest

OK, got the serial up and working.

I get a copyright message on the terminal and then it tells me the revision number. It then says symbols linked (screen01) and that's it.

On the MUT, it has hung *very* early with OemHlpXHandler: CallACPI16: 49008DF, CallACPI32: F92C5182, ExportTable32: F94002CC

Loading PNP.SYS

And that's it.

Can you verify that KDB.INI should read:


.b 115200t

G

ln


together with spaces between lines?

Thanks, Stephen

comment:30 Changed 12 years ago by guest

When I hit the reset button on the server, I got on the terminal:

IPI: IPIResponse timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0

Oooh. Any help?

comment:31 Changed 12 years ago by guest

Server is an Intel SE520JR, if that helps.

comment:32 follow-up: Changed 12 years ago by pasha

Attach log from terminal please

comment:33 in reply to: ↑ 32 Changed 12 years ago by guest

There wasn't anything, apart from the copyright message and the screen01.sys message. Maybe I am missing something?

Replying to pasha:

Attach log from terminal please

comment:34 Changed 12 years ago by pasha

Do you add to options /Ox, where x is number of using comport? 1 for COM1, 2 for COM2

comment:35 Changed 12 years ago by eco

how to setup debug output:

http://ecomstation.ru/projects/acpitools/?action=faq-common

Q21: Does debug version of ACPI.PSD support logging to COM port? A21: Yes, logging to COM port is possible.

Get debug version of ACPI.PSD and modify the CONFIG.SYS:

psd=acpi.psd /O1 (O1 - logging to COM1, O2 - logging to COM2)

Terminal settings:

115200,8,N,1 HW flow support. (ZOC: [x] RTS, [x] CD)

Changed 12 years ago by guest

KDB Output - hang on 'Loading TIMER0.SYS'

comment:36 Changed 12 years ago by guest

Now got KDB output. Biggest problem was getting serial cable that worked, and then remembering settings for PSD.

Attached is acpi-kdb1.txt. The system hung on 'Loading TIMER0.SYS' - this equated to the following line in the terminal:

UNMASK PIC IRQ1 mask 0xFFFFDCF9 Real:FFEB7CAC

Thereafter there was no terminal output for about a minute, and then it produced:

IPI: IPIResponse timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgrer number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 ProcInit? CPU2 LApic ID:0x0: ver 0x50014 Error!!! Warning! Debugger stack overflow eax=ffffffff ebx=ffdaee9d ecx=0000441d edx=bfebfbff esi=fcb9d9c8 edi=f9261804 eip=fff0f5c6 esp=00004e70 ebp=00004f20 iopl=0 -- -- -- nv up di ng nz na pe nc cs=0178 ss=1530 ds=0170 es=0170 fs=0000 gs=0000 cr2=00000000 cr3=00eb6000 p=01s os2krnl:DOSHIGH32CODE:MPSpinLoop: 0178:fff0f5c6 803b00 cmp byte ptr [ebx],00 ds:ffdaee9d=ff 0178:fff0f5c6 os2krnl:DOSHIGH32CODE:MPSpinLoop

Hope this helps, Stephen

comment:37 Changed 12 years ago by pasha

Download pls 3.6 version and get log from it. You need acpi.psd from debug directory, boot and attach terminal log

Changed 12 years ago by guest

Terminal output of 3.06

comment:38 Changed 12 years ago by guest

  • Version changed from 3.05 to 3.06

Downloaded and copied v3.06.

Hung on 'Loading ISAPNP.SNP'

Attached terminal log in 'ACPI-306.LOG'

Last lines of hang were: OemHlpXHandler: ACPI_START_EVENTS 4 OemHlpXHandler: CallACPI16: 49008DF, CallACPI32: F92C5402, ExportTable32: F94001 CC

I waited a few minutes, and then pressed CTRL-C on the terminal. I then got:

IPI: IPIResponse timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 IPI: IPIInProgress timeout from processor number 0 ProcInit? CPU2 LApic ID:0x0: ver 0x50014 CPUINIT: Error!!! Warning! Debugger stack overflow eax=ffffffff ebx=ffdaee9d ecx=0000441d edx=bfebfbff esi=fcb9d9c4 edi=f9261804 eip=fff0f5c6 esp=00004e70 ebp=00004f20 iopl=0 -- -- -- nv up di ng nz na pe nc cs=0178 ss=1530 ds=0170 es=0170 fs=0000 gs=0000 cr2=00000000 cr3=00eb9000 p=01s os2krnl:DOSHIGH32CODE:MPSpinLoop: 0178:fff0f5c6 803b00 cmp byte ptr [ebx],00 ds:ffdaee9d=ff 0178:fff0f5c6 os2krnl:DOSHIGH32CODE:MPSpinLoop

Hope this helps, Regards,

Stephen

comment:39 Changed 12 years ago by pasha

acpi26012007.zip, need terminal log

Changed 12 years ago by guest

Trap 14 in latest build

comment:40 Changed 12 years ago by guest

Hi Pasha,

Trap E in the build above:

Trap 14 (0EH) - Page Fault 0002, Not Present, Write Access, Supervisor eax=00000020 ebx=00000000 ecx=00000040 edx=00000018 esi=00000047 edi=00000002 eip=f92ca409 esp=ffe4ffa8 ebp=ffe4ffc4 iopl=0 rf -- -- nv up di pl nz na po nc cs=0178 ss=1520 ds=0170 es=0170 fs=0108 gs=0000 cr2=00000000 cr3=00225000 p=00 0178:f92ca409 890b mov dword ptr [ebx],ecx ds:00000000=invalid 0178:f92ca1b9 acpi:CODE32:SetupIOAPIC + 250 0178:f92ca4d1 AcpiSetAPIC - c8 ##

Didn't get far through booting (not that it ever does :) ) but didn't even start base driver loading.

Regards, Stephen

comment:41 Changed 12 years ago by pasha

acpi26012007-01.zip, need terminal log

Changed 12 years ago by guest

Back to hang whilst load ISAPNP.SNP

comment:42 Changed 12 years ago by guest

Hi Pasha,

New attachment - acpi-2601-01.log

Back to same behaviour as before. There might be a difference in the terminal log, but after a very quick look I couldn't see anything obvious.

Regards, Stephen

comment:43 Changed 12 years ago by pasha

try psd=acpi.psd /SMP /!NOD /O1

Changed 12 years ago by guest

/SMP /!NOD /O1 - Hang Loading PCIBUS.SNP

comment:44 Changed 12 years ago by guest

acpi2601-01-nonod.log attached. Hang as before, this time PCIBUS.SNP (which is likely just happens to be the file loading as something else goes wrong).

I notice that there are some errors in the log concerning no response from hardware for a sleep button...

What theories are you working on at the moment?

Thanks, Stephen

comment:45 Changed 12 years ago by pasha

Ok. try add /FS, if don't help - need think. After last call acpi, kernel must unmask keybord IRQ and load resource.sys. BTW, try remming all in snoop.lst

Changed 12 years ago by guest

/SMP /!NOD /FS /O1 - Progress, hang at UNMASK PIC

comment:46 Changed 12 years ago by guest

Hi Pasha,

Adding /FS seemed to allow a bit of progress (note: no changes made to snoop.lst - decided to make one change at a time).

Hang 'Loading OS2LVM.DMD' at

evxfevnt-176 [1] Enable: Transition to ACPI mode successful UNMASK PIC IRQ11 mask 0xFFFF54B9 Real:FFEB7CAC UNMASK PIC IRQ3 mask 0xFFFF54B1 Real:FFEB7CAC

in terminal window. After a minute, nothing happended, so I pressed CTRL-C and got:

eax=ffffffff ebx=ffdaee9d ecx=0000441d edx=bfebfbff esi=fcb9d9bc edi=f9261804 eip=fff0f5c6 esp=00004e70 ebp=00004f20 iopl=0 -- -- -- nv up di ng nz na pe nc cs=0178 ss=1530 ds=0170 es=0170 fs=0000 gs=0000 cr2=00000000 cr3=00eb1000 p=01s os2krnl:DOSHIGH32CODE:MPSpinLoop: 0178:fff0f5c6 803b00 cmp byte ptr [ebx],00 ds:ffdaee9d=ff ##

Thanks, Stephen

comment:47 Changed 12 years ago by pasha

It is a progress ;-) Now set psd=acpi.psd /SMP /APIC /FS /O1

comment:48 Changed 12 years ago by pasha

BTW, have you irc client or icq? Find me can in irc.netlabs.org channel #netlabs or ICQ 32400130

Changed 12 years ago by guest

/SMP /APIC /FS /O1 - IPE whilst loading USBEHCD.SYS

comment:49 Changed 12 years ago by guest

Hi,

With /SMP /APIC /FS /O1, I got an IPE whilst loading USBEHCD.SYS

Terminal log in acpi2601-01-fs-apic.log

The last lines in the terminal were:

evxfevnt-176 [1] Enable: Transition to ACPI mode successful IRQMask SetVirtualWireMode? SetVirtualWireMode? Apic 0 PIC 0 PCI:200 PICMask:DCF9 IRQs:56 Shutdown PIC mask FFFFDCF9 The system detected an internal processing error at location ##0178:fff13c65 - 0011:00015c65. 60004, 9103 0986090f - 02/25 02:51:45.18 Internal revision 14.104a_SMP The system is stopped. Record all of the above information and contact your service representative. eax=ffe4ffb0 ebx=ffe4ffb0 ecx=00000002 edx=00000009 esi=ffeb7ac0 edi=00000030 eip=fff13c65 esp=00004d28 ebp=00004d28 iopl=0 -- -- -- nv up di pl zr na pe nc cs=0178 ss=1530 ds=0170 es=0170 fs=0000 gs=0000 cr2=ffdb3000 cr3=00225000 p=00 0178:fff13c65 e88a6fffff call panicHISHIGH32CODE:_PSDter - 9

Must get myself on IRC one day :)

Gotta get some rest, I will try anything you need after some sleep!

Thanks, Stephen

comment:50 follow-up: Changed 12 years ago by pasha

ACPI26012007-02.ZIP

Need remap IRQ55 and IRQ54 to low. try add to acpi.cfg 2 line. "REMAP 54 to 24" and "REMAP 55 to 25", then attach acpi.cfg for my check. I beleive your boot drive is IDE/SATA. If SCSI - this will don't work and need find another way.

comment:51 in reply to: ↑ 50 Changed 12 years ago by guest

Hi Pasha,

Boot drive is LSIMPT U320 SCSI. There is an IDE/SATA controller and CD drive, but no HDD.

Uh-oh!

Regards, Stephen

Replying to pasha:

ACPI26012007-02.ZIP

Need remap IRQ55 and IRQ54 to low. try add to acpi.cfg 2 line. "REMAP 54 to 24" and "REMAP 55 to 25", then attach acpi.cfg for my check. I beleive your boot drive is IDE/SATA. If SCSI - this will don't work and need find another way.

comment:52 Changed 12 years ago by pasha

Ok. Now we see - need /FS. try boot with psd=acpi.psd /SMP /FS /O1. This will w/o APIC so your SCSI driver must understand this.

Changed 12 years ago by guest

/SMP /FS /O1 - Hang 'Installed OS2LVM.DMD' / UNMASK PIC

comment:53 Changed 12 years ago by guest

Hi Pasha,

Log attached - acpi2601-fs.log.

Hung on 'Installed OS2LVM.DMD' with two UNMASK lines last on terminal:

evxfevnt-176 [1] Enable: Transition to ACPI mode successful evxfevnt-225 [1] Disable: ACPI mode disabled evxfevnt-176 [1]mode successful evxfevnt-225 [1] Disable: ACPI mode disabled evxfevnt-176 [1] Enable: Transition to ACPI mode successful UNMASK PIC IRQ11 mask 0xFFFF54B9 Real:FFEB7CAC UNMASK PIC IRQ3 mask 0xFFFF54B1 Real:FFEB7CAC

After two minutes, nothing happened, so I pressed CTRL-C and got a register dump and KDB prompt.

eax=ffffffff ebx=ffdaee9d ecx=0000441d edx=bfebfbff esi=fcb9d9c0 edi=f9261804 eip=fff0f5c6 esp=00004e70 ebp=00004f20 iopl=0 -- -- -- nv up di ng nz na pe nc cs=0178 ss=1530 ds=0170 es=0170 fs=0000 gs=0000 cr2=00000000 cr3=00eb1000 p=01s os2krnl:DOSHIGH32CODE:MPSpinLoop: 0178:fff0f5c6 803b00 cmp byte ptr [ebx],00 ds:ffdaee9d=ff ##

Regards, Stephen

comment:54 follow-up: Changed 12 years ago by guest

Did LSIMPT.ADD get installed in \OS2\BOOT and CONFIG.SYS? The IBM distribution of this driver has a note that looks interesting (a sort of CONFIG.SYS might eat your lunch):

3.2.4 Driver Order in the CONFIG.SYS File

Because of the way OS/2 assigns drive letters, the order in which drivers appear in the CONFIG.SYS file is important. The drivers must appear in the order in which the drive letters are to be assigned. In particular:

OS/2 will install the BASEDEV= line at the beginning of the target system's CONFIG.SYS file regardless of where the line is located in that file on Diskette 1. You may need to rearrange the order in which drivers appear in the CONFIG.SYS file for the target system.

The driver for the host adapter to which the boot device is attached must appear before all other BASEDEV= drivers in the CONFIG.SYS file.

You can find a copy of the driver here:

http://www.os2site.com/sw/drivers/scsi/74p4792.exe

Chuck McKinnis? mckinnis@…

comment:55 in reply to: ↑ 54 Changed 12 years ago by guest

Hi Chuck,

Yes, thanks for this. I am aware of this requirement. For instance, under OS/2 Warp 2/3/4, OS/2 would actually change the drive it was booting off half way thru (if you had two copies of the OS installed) if your BASEDEV statements were in the wrong order. I would argue that this is/was no longer the case under an LVM aware system (despite what IBM say), whereby the 'C' drive is as per the volume information found by OS2LVM.DMD, not the first logically found primary partition.

Nevertheless, I'd already moved LSIMPT.ADD ahead of DANIS506.ADD to rule that out and all of the dumps above are of that order.

I've just tried moving it ahead to nearest the very top (again, according to IBM's instructions, the first boot DASD driver should be the very first BASEDEV, but really stuff like IBMKBD.SYS should make no difference) but as suspected, it made no difference.

Thanks for the suggestion tho!

Regards, Stephen

Replying to guest:

Did LSIMPT.ADD get installed in \OS2\BOOT and CONFIG.SYS? The IBM distribution of this driver has a note that looks interesting (a sort of CONFIG.SYS might eat your lunch):

3.2.4 Driver Order in the CONFIG.SYS File

Because of the way OS/2 assigns drive letters, the order in which drivers appear in the CONFIG.SYS file is important. The drivers must appear in the order in which the drive letters are to be assigned. In particular:

OS/2 will install the BASEDEV= line at the beginning of the target system's CONFIG.SYS file regardless of where the line is located in that file on Diskette 1. You may need to rearrange the order in which drivers appear in the CONFIG.SYS file for the target system.

The driver for the host adapter to which the boot device is attached must appear before all other BASEDEV= drivers in the CONFIG.SYS file.

You can find a copy of the driver here:

http://www.os2site.com/sw/drivers/scsi/74p4792.exe

Chuck McKinnis? mckinnis@…

comment:56 Changed 12 years ago by guest

Any more ideas Pasha?

Incidentally, I'm finding also on this box that I have to use DANIS506.ADD under eCS 1.2MR with the /FORCEGBM switch (even tho there is just a single EIDE device - a CDRW) otherwise the machine will also hang.

I've also previously swapped out the motherboard - no difference.

Regards, Stephen

comment:57 follow-up: Changed 12 years ago by pasha

Please download experimental ACPI build from Mensys site:

  • Experimental build for you:

acpi31012008.zip

Changed 12 years ago by guest

Testcase - Hang on UNMASK / Installed OS2LVM.DMD as before

comment:58 in reply to: ↑ 57 Changed 12 years ago by guest

Hi Pasha,

Thanks very much for this. Unfortunately, the system still seems to hang in the same place. I've uploaded the log file from the COM port.

Please let me know if you want me to try anything else... & once again, thanks for your efforts :)

Regards, Stephen

Replying to pasha:

Please download experimental ACPI build from Mensys site:

  • Experimental build for you:

acpi31012008.zip

comment:59 Changed 12 years ago by guest

  • Milestone set to eCS 2.0

comment:60 Changed 11 years ago by stevenhl

  • Milestone changed from eCS 2.0 to Feedback pending

comment:61 Changed 11 years ago by stevenhl

  • Milestone changed from Feedback pending to Release version 3.13
  • Resolution set to unknown
  • Status changed from new to closed

Waited long enough for feedback.

Note: See TracTickets for help on using tickets.