ACPI315PRE fails to boot

[losepete]

Hi

I updated the installed v3.14 ACPI with the files from ACPI315PRE.ZIP and rebooted.

The system stopped booting with the screen shown in the attached file SVN_stopped.jpg

As the above was using a "Test kernel" I hit the Reset button and selected to boot using the IBM 14.104a_SMP kernel. That trapped; image in the attached file 14104SMP_stopped.jpg - Sorry it is not a very good snap.

Mainboard uses nVidia nForce430/MCP61 chipset.

[losepete]

I just tried the ACPIPRE2.ZIP files.

That *Fails* to boot with exactly the same TRAP screen when using the 14.104_SMP kernel and stops with the same messages onscreen when using any of the OS/4 (Phoenix) kernels.

[pasha]

line is psd=acpi.psd /SMP /APIC ?

[losepete]

Hi pasha

No.

That line is a non-starter in itself...

I had left the config.sys line as PSD=ACPI.PSD /SMP /CD /TMR /APIC /!NOD

I then tried PSD=ACPI.PSD /SMP /CD /TMR /APIC but the boot stops as soon as the USB controllers try to load, does not get as far as when using the /!NOD switch.

[pasha]

move os2logo from root of boot drive, set psd=acpi.psd /SMP /APIC and give me photo of screen. If trap be here, install debug kernel and create photo.

[losepete]

What "debug kernel"?

[losepete]

Replaced

\ecs\bin\iasl.exe \os2\acpidaemon.exe \os2\boot\acpi.psd \os2\boot\apm.add \os2\dll\acpi32.dll

along with the *.sym files.

Copied OS2LOGO to \os2\OS2LOGO then deleted the original from boot. Just wondering if I should have done the same with OS4LOGO?...

Edited CONFIG.SYS line from

PSD=ACPI.PSD /SMP /CD /TMR /APIC /!NOD

To

PSD=ACPI.PSD /SMP /APIC

Rebooted.

Using OS4KRNL the boot stops at USBEHCD.SYS and 10 minutes later is still stuck there.

Using OS2KRNL the boot pauses for 2 to 3 minutes at USBEHCD.SYS and then displays the TRAP screen shown in attached image 315pre_os2.jpg

Replaced the above files with those from ACPIPRE2 and rebooted.

Using OS4KRNL the boot stops at USBEHCD.SYS and 10 minutes later is still stuck there.

Using OS2KRNL the boot pauses for 2 to 3 minutes at USBEHCD.SYS and then displays the TRAP screen shown in attached image pre2_os2.jpg

I then changed the files involved to those supplied with v3.14 and rebooted. I captured that TRAP screen for comparison as 314_os2.jpg - looks like the same Trap as the acpi315pre and acpipre2 files generate.

I finally edited the CONFIG.SYS file and reinstated the "working with v3.14" line

PSD=ACPI.PSD /SMP /CD /TMR /APIC /!NOD

Rebooted - without Traps - and prepared this report.

We already know the above line fails to help when either acpi315pre and acpipre2 files are installed...

Based on the above acpiv315 does not look like it will work at all with nVidia Nforce430/MCP61 chipsets.

[pasha]

install pls debug kernel? I want see message from acpi.psd, instead kernels traps

[losepete]

Let me try that again: WHAT DEBUG KERNEL?

Please point me at whatever kernel you want me to install.

[pasha]

Open  http://betazone.ecomstation.nl and find acpi-kernel-20070123.ZIP

[losepete]

Replying to pasha:

Open  http://betazone.ecomstation.nl and find acpi-kernel-20070123.ZIP

Try suggesting a file that exists!

That filename is NOT in the betazone.

[losepete]

Replying to losepete:

Replying to pasha:

Open  http://betazone.ecomstation.nl and find acpi-kernel-20070123.ZIP

Try suggesting a file that exists! That filename is NOT in the betazone.

Apologies: The file is hidden in the section called zArchive so will now install and report back.

[losepete]

Installed the kernel.

Installed ACPIPRE2 files.

Changed CONFIG.SYS line to

PSD=ACPI.PSD /SMP /APIC

Rebooted and selected to use the debug kernel.

Result: NO BOOT!!!

Boot stops with USBEHCD.SYS; does NOT go any further, NO Trap screen or Debug screen.

Next suggestion?

[pasha]

Photo of screen with removed os2logo

[losepete]

You want a photo of a blank screen showing the text:-

Loading USBEHCD.SYS

???

Now I am convinced you are wasting my time.

I suggest you go find a mainboard with an nVidia nForce430/MCP61 chipset, eg an ASUS M2N, and waste your own time.

As my eCS Subscription is just about due and I currently have No plans to renew that Subscription I will not be testing any further "developments" of ACPI.

I look forward to renewing this discussion when eCS2.0GA arrives and I find that ACPI does not work.

[pasha]

If you use last acpi.psd, you should see some debug message at screen. If you don't see this message, check acpi.psd is only one and in \os2\boot (check \, os2\, \os2\boot). I want see message from acpi.psd. Then I can begin resolve your problem. Also, if you add /O1 to acpi.psd, remove it.

[losepete]

I just discovered why I was not seeing the debug screen: I had a "blank" template of os2logo in the root of my boot drive.

I got rid of that file and reset the system to use ACPIPRE again.

The result is the attached Pre-new.jpg file, not a brilliant bit of photography but I can read the screen image so hope you can.

[pasha]

Thx. Now I understand where can be problem. Do try add /CD ?

[pasha]

Additionals. If /CD don't help. Install pls W4/Uni kernel (need change doscall1.dll) and boot as psd=acpi.psd. Log attach here. I need look log with /NOD. If you want fast this process we can have contact in irc (ecsnet #ecolabs) or icq (32400130)

[losepete]

I have the uni kernel package 14104a

This does not contain a DOSCALL1.DLL

The only versions of that file I seem to have available are all the SMP version.

Where can I locate the correct DOSCALL1.DLL file?

[pasha]

rename to doscall1.dll and use with Uni kernel 104a

[losepete]

Tried the /CD switch

The result is attached as With_CD.jpg

[losepete]

uni kernel 14.104a installed, doscall1.dll changed, config.sys line changed to psd=acpi.psd and rebooted.

No log available as the boot failed.

I have attached UNI.jpg showing where the boot stopped - the value of ExportTable?32: has the last bit missing so here is the complete value: F9F101

[pasha]

Ok. Can you remove all from config.sys? I mean all - all don't need for boot. I need in log with /NOD.

Now, your problem is in Open high IRQ, when already open low IRQ. You have IRQ flood - next is trap. freeze or any destructive doing. When we give log, then we can try change low IRQ for some device.

[losepete]

Installed ACPIPRE2 files

System failed to boot using a "stripped down" config.sys and UNI kernel

System booted to a working Desktop using the kernel from acpi-kernel-20070123.ZIP

ACPI line used: PSD=ACPI.PSD

Logs attached, Hope they help :-)

[pasha]

Can you rem all scsi support? I see os2aspi.dmd in your photo. If don't help - try remming usb drivers. If it will help - then I try add some options to acpi.cfg

[losepete]

The logs supplied were obtained by booting with a much reduced config.sys file - there were NO usb drivers and os2aspi.dmd was not loaded.

If the above logs - obtained with a lot of effort - are not of help then I doubt there is anything else I can do as I do have have all day, every day in which to fanny around with this problem.

V3.14 works fine as long as the config.sys line looks like this:-

PSD=ACPI.PSD /SMP /CD /TMR /APIC /!NOD

I suggest you look at the code changes between V3.14 and the ACPIPRE2 builds as ACPIPRE2 does not work at all.

[pasha]

I know where and why don't work /!NOD now. As I understand - need run with /NOD. In your photo I see about mask/unmask IRQ21. Looking to your acpi.log:

0:2.0 10DE:3F1 'USB0' IRQs: PCI:10 PIC:'LUB0':10 APIC:'LUB0':21

0:8.0 10DE:3F6 'ATA0' IRQs: PCI:15 PIC:'LSA0':15 APIC:'LSA0':21 MSI

As I know, danis506.add don't open high IRQ before end of boot. So it is USB. Check your config.sys more pls.

[losepete]

The log files were generated when booted using a config.sys with NO reference to USB drivers at all.

I have attached config.sys.reduced for your reading pleasure.

Just in case I am misunderstanding something I will re-install the ACPIPRE2 files and boot with a config.sys containing no USB drivers and using this line:-

PSD=ACPI.PSD /SMP /APIC

Will report back in a day or maybe 2.

[losepete]

Found some time to test immediately.

Result was that using a "near normal" config.sys file with No USB drivers and the line PSD=ACPI.PSD /SMP /APIC the boot hangs with

"HDA NVidia at 0xdfef8000 irq23"

Obviously the uniaud114RC6 plus uniaud32.sys build 1.9.2 is causing a problem.

I deleted those from the config.sys file and tried again.

Result: Booted all the way to a working Desktop - but, obviously, No USB or Audio capability.

I have attached the file acpi-log-Model-of-your-PC-01.ziq for your perusal.

[losepete]

I used tmrtest.exe to check whether the /TMR switch is required - it is so I changed the config.sys line to

PSD=ACPI.PSD /SMP /TMR /APIC

This has made no difference to whether uniaud or usb can be used though.

I would send you screenshots of the failure to boot with uniaud and usb but have to reset my system to be able to use USB before I can do that - no USB capability means that I cannot attach either the camera or the flash card using a USB cardreader at the moment. I will post screenshots later.

[pasha]

try use  http://download.smedley.info/uniaud32-1.9.12.zip , in this version I add some stuff for /APIC.

[losepete]

I think the problem is with the uniaud16.sys driver as the system stops when trying to load that file - uniaud32.sys has already loaded as shown by the message identifying the audio hardware.

Will give it a try though and report back soon - need to setup another RC6a installation to experiment on rather than keep changing files on my main working system.

What possible cures are there for the lack of USB capability though?

[losepete]

I have now attached the screenshots:

usbehcd.jpg - showing where boot stops when usb drivers are present in config.sys file; if I REM usbehcd.sys boot then stops with usbd.sys stuck onscreen.

uniaud.jpg - showing where boot stops when the uniaud drivers are in the config.sys file (no usb drivers)

Will report back on success or not with uniaud32-1.9.12 soon.

[losepete]

Just confirming that uniaud 1.9.12 does allow the system to boot to a working Desktop when using the ACPIPRE2 files.

Sadly uniaud 1.9.12 is a very crap audio driver - system sounds play too fast, no audio when playing video/dvd using mplayer/Smplayer, Klondike Solitaire hangs after dealing the 1st card and requires the use of Ctrl-Alt-Del to run Top to "kill" pmshell.exe before the system can be used again.

When using acpi v3.14 the *only* combination of uniaud files that work reasonably well is uniaud16.sys from the 114RC6 package coupled with uniaud32.sys from uniaud32 1.9.2 - none of the above problems.

[pasha]

Please download experimental ACPI build from Mensys site:

• Experimental build for you:

ACPIPRE3.ZIP

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

Check with /!NOD pls

[losepete]

Installed the ACPIPRE3 files

Left the config.sys line as: PSD=ACPI.PSD /SMP /CD /TMR /APIC /!NOD and rebooted.

Result: No Boot - see Not_NOD.jpg

Changed the config.sys line to: PSD=ACPI.PSD /SMP /APIC and rebooted.

Result: No Boot - boot stops at USBEHCD.SYS

[pasha]

Check pls right writes options and acpi.psd is only one in \,os2,os2\boot. Here /CD don't need.

[losepete]

Yes, acpi.psd is the only file by that name in \os2\boot

What do you mean by "right writes options"?

[pasha]

I have 2 computers, which can use /SMP /APIC /!NOD. Both show 16 IRQ. Your photo show 24. From photo, I think you need check /!NOD or file acpi.psd

[losepete]

I'm sure you know what you mean but I can only guess...

The config.sys line used contained /!NOD

The acpi.psd used was that supplied in ACPIPRE3.ZIP

I was surprised to see IRQ23 onscreen as /!NOD should have limited that to IRQ15 - but it did not.

I will retest to check but I suspect we have already found a problem with acpipre3 - maybe /!NOD does not work with this chipset when using this build?

[losepete]

On retest the system booted fine using the ACPIPRE3 files and this config.sys line:-

PSD=ACPI.PSD /SMP /APIC /!NOD

I then used the tmrtest.exe program to check if /TMR is required - it is. I added this to the above config.sys line so the line is now:-

PSD=ACPI.PSD /SMP /TMR /APIC /!NOD

Before rebooting I tried to attach a USB drive. The drive attached fine but when I later Ejected it the system hung completely requiring a press of the system box Reset button.

On reboot, now using the /TMR switch, the system booted fine.

I retested using tmrtest.exe and the Elapsed time and Must be time are now the same.

I attached a USB drive - the same as used previously - without any problems and accessed a few files on that drive.

Worth noting here is that the crash experienced previously when Ejecting the USB drive did not leave the drive dirty. That suggests the crash happened after the USB drive had been Ejected.

Also, when I Ejected the same drive this time there was no problem, no system hang. Would the /TMR switch have made that difference? - Looks like the /CD switch is no longer required for USB support.

I attach the logs created when booted using the PSD=ACPI.PSD /SMP /TMR /APIC /!NOD line as PRE3_acpi-log-Model-of-your-PC.ziq

[losepete]

Having posted the above report I then tried to add the logs as an attachment and Seamonkey stopped responding followed very quickly by a total system hang requiring a press of the Reset button. I've not had this sort of problem with Seamonkey or the system previously when using ACPI V3.14 so would guess that it is PRE3 related.

According to the Attachments list the file got uploaded before the crash happened. You will have to let me know if it needs resending...

[losepete]

Having used the PRE3 files for over a day I have decided to revert to the acpi V3.14 files.

Reason: The PRE3 files seem to cause too many problems

1] System instability/hangs/crashes

2] Audio and/or network drivers sometimes fail to load

[losepete]

I just found (at Mensys download) and installed the files from acpi315.zip

Experimentation reveals that the following line is the only line to use in config.sys with my Asus M2N mainboard (nVidia nForce430/MCP61 chipset):-

PSD=ACPI.PSD /SMP /TMR /APIC /!NOD

Correct!

This mainboard still cannot boot without the /!NOD switch.

I have to report that v315 has some of the "hang" characteristics of the PRE3 files; the system can become unresponsive for a short period of time but, so far, no crashes.

eStyler Shutdown does not work - this is the eStyler fixed to work with ACPI V3.14 - and fails to get to Power Off/Press Ctrl-Alt-Del/Reboot.

Also worth noting is that eCS systems have a place for LOGFILES; mine is:-

SET LOGFILES=L:\var\log

I mention this as the acpid.cfg file supplied with acpi315 has acpidaemon.log set to C:\ !!!

On my system C: is not accessible being a Windows2000 installation using the NTFS filesystem.

You may want to make sure future updates to acpid.cfg do not make that mistake.

I notice that acpi325 includes a kernel. Is it necessary to use this kernel - and where does it come from?

[losepete]

I do wish you would find a more reliable means of bug trcking.

This TRAC system is extremely flaky: it is very common to get an error when trying to submit changes and sometimes fails to locate pages completely.

[pasha]

With /NOD you have IRQ conflict with USBEHCD.SYS. Way to resolve is 2. 1st - make USBEHCD.SYS as normal for acpi, 2nd try by LINK directive change IRQ. 1st under constructions, second - I am in thinking process.

[ecsnl]

Replying to losepete:

I just found (at Mensys download) and installed the files from acpi315.zip I have to report that v315 has some of the "hang" characteristics of the PRE3 files; the system can become unresponsive for a short period of time but, so far, no crashes.

Read ticket 403. Maybe its related these "short" hangs ?

Roderick

[losepete]

Looks very similar to ticket 403 - but this is a Desktop system and does not have eWP/eCenter installed so the cause must be something different.

As v3.14 is more stable and reliable I have reverted to using that build on this system.

I will try v3.15 on another system, Asus A8N-E mainboard using nForce4 chipset, that works very well with v3.14 and post a report in a day or 2.

[losepete]

I have now had v3.15 installed on the A8N-E (nForce4 chipset) system and the users report:-

The system "hangs" for a few seconds on a regular basis.

Attaching USB drives sometimes does not work.

Neither of the above is a problem when using v3.14.

This system uses PSD=ACPI.PSD /SMP /APIC for both v3.14 and v3.15 builds.

I have attached the logs as A8NE-v314-acpi-log-Model-of-your-PC.ziq A8NE-v315-acpi-log-Model-of-your-PC.ziq

[pasha]

I think you are need in /TMR options. Check pls this via tmrtest.exe.

[losepete]

tmrtest.exe does not indicate a need for /TMR with either 3.14 or 3.15

[losepete]

I have tried the latest 315 build - ACPI315-09_05_2009 - on the M2N mainboard and find that it also has a tendancy to "hang" for a very short period of time; between 1 to 2 seconds.

This is very noticeable when starting an application eg Seamonkey, OpenOffice?, Java apps.

[stevenhl]

This acpi build seems to need Polling added to acpi.cfg. It has reduced the delays on some systems.

[losepete]

Following the comment about "Polling" I reinstalled ACPI315-09_05_2009 to test.

Yes, Polling helps - the "hangs" seem to have been cured.

No change from earlier versions of acpi in that the required config.sys line is still PSD=ACPI.PSD /SMP /CD /TMR /APIC /!NOD

A question about the directory structure of ACPI315-09_05_2009: The contents of \util go into \ecs\bin to overwrite existing files but where should the contents of \gui get unzipped to?

Some hint in the docs may help there.

A further suggestion: REMOVE the "c:\" from the LOG file locations in the *.cfg files - not many people have an eCS installation on drive C: as it is usually in use by a Windows variation and is, usually, totally inaccessible from an eCS system.

[losepete]

My previous post is incorrect: Polling helps a little but does not completely cure the "hangs".

Using VirtualPC with this build of acpi will cause the system to crash within 15 minutes. Initially the guest pc becomes unresponsive and then the host pc crashes.

I have reverted to acpi V3.14 as that is far more stable than any of the V3.15 builds.

[losepete]

I found acpi316-05_29 so installed to test.

Sadly this build does not reliably boot here, 9 out of 10 attempts get stuck near the end of config.sys processing, just prior to the call to pmshell.exe.

There are a lot of lines displayed onscreen all stating

AIR CPU0 Vec:0x57 IRQ7 Mask:FFFF20FD

There are a total of 19 of these lines, 14 start with 14:31:25 and the last 5 start with 14:31:26.

The config.sys line in use is PSD=ACPI.PSD /SMP /CD /TMR /APIC /!NOD - without the /!NOD switch boot stops at the (usual) usbehcd.sys driver.

Back to V3.14...

[losepete]

Installed acpi-06-08 - I put the unknown file, ACPIEXEC.EXE, into \ecs\bin as there are no instructions on where to put that file.

I then Shutdown the system using the Power Off option. After a minute I rebooted successfully to a working Desktop. The config.sys line in use was:-

PSD=ACPI.PSD /SMP /CD /TMR /APIC /!NOD

I edited config.sys and changed the line to:-

PSD=ACPI.PSD /SMP /CD /TMR /APIC

I then Shutdown using Power Off, waited a minute and then tried to boot. The boot stopped at usbehcd.sys which seems to indicate that the /!NOD switch is still required. I changed the config.sys line to include that switch again and rebooted successfully.

After several hours of using the system I Powered Off.

When I tried booting the next day the system refused to boot with exactly the same *19* messages onscreen as the acpi316-05_09 build:-

AIR CPU0 Vec:0x57 IRQ7 Mask:FFFF20FD

Seems strange that both these 3.16 builds work when first installed but FAIL to boot when the system has been Powered Off for several hours.

Could this be the result of applying the v3.16 files to a system using v3.14? - do I need to install the v3.15 files before installing v3.16?

Something for your thought dept:- When including *new* files also include a readme that explains what the new files are and where to install them - this is a consistent problem with acpi and needs "fixing".

Back to V3.14...

[losepete]

Installed acpi-06-08

Could not boot with my usual config.sys line: PSD=ACPI.PSD /SMP /CD /TMR /APIC /!NOD

Boot stops with the screen display noted with earlier 316 builds: AIR CPU0 Vec:0x57 IRQ7 Mask:FFFF20FD

I edited the config.sys line to: PSD=ACPI.PSD /SMP /CD /TMR /APIC

Boot stops at the USBEHCD.SYS driver.

I edited config.sys and deleted USB drivers.

System booted to a working Desktop so I collected the logs, attached as 316-06-16_acpi-log-Model-of-your-PC.ziq

I have had to reboot the system several times since and sometimes it boots OK, sometimes it does not.

I do have to repeat my previous question: Should I have installed v3.15 then updated that with v3.16 files? - especially with regard to acpi.cfg and acpid.cfg

[losepete]

Sorry, my last message states "Installed acpi-06-08" - that should have said Installed acpi-06-16

Following an overnight Power Off it was very difficult to get the system to boot.

Booting stopped with different messages onscreen for the first 3 attempts - no digital camera handy to take snaps with at the moment - the 4th attempt resulted in the system getting to the Desktop Startup sound and then hanging with the end of that sound looping, the 5th attempt got to a working Desktop.

As a result I have reinstalled acpi V3.14