Regression in USBHCD184, after shutdown the system restart immediately.

[Anchieri]

I have installed your last usbhcd184 and have the regression in USBEHCD signaled. With the last usb 181, 182 and 183 my system shutdown with power off without problem, but with 184 the system shutdown and after a few second reboot without my intervention. I have replaced the USBEHCD of 184 with that of 183 and system shutdown and power off normally. My system is an ASUS with a P5LD2-X/1333 mother board dual core2 duo with 2 USBOHCD, 4 USBUHCD AND 2 USBEHCD.

[Anchieri]

Please I have installed your new version usbhcd185 but this problem is still present, also I have another problems, my usb printer attached to a usb port version 1.1 don't print anymore if I want print I have to use a usb version 2.0 port. And a last problem, I have two external disk drive one with 160GB formatted in JFS and the other with 512GB that I use as backup. I was used to attach this drive when they need, but now with this two version I was forced to attach it before to boot the system, otherwise is impossible to attach they when the wps is up.

[Lars Erdmann]

Try:

BASEDEV=USBEHCD.SYS /S:8

[Anchieri]

I have tried this switch but nothing happened, the system restart immediately after shutdown, it is impossible to attach my removable disk (only sd card come attached) and the printer don't print using 1.1 port.

[Lars Erdmann]

Try brandy.6.zip from ticket #26

Check if the printer now works ok.
Did you update ACPI.PSD when the reboot problem started to occur ?

[Anchieri]

I have tried the brandy.6.zip an printer now work using usb 1.1 port, but during shutdown I have a repeating message that I have to skip for the shutdown can continue, and after the shutdown complete the system behave as comment3 (excluding printer).

[Anchieri]

I have installed acpi.psd version 3.20.01 in may after that I have installed 3.20.04 a few week ago.

[Lars Erdmann]

The problems you describe have nothing to do with the USB drivers at all.
There is no such thing as a usb 1.1 port. Each port is both, USB 2.0 and 1.1, depending on what device (speed) you attach to that port.

I had accidentally introduced an error in USBOHCD.SYS with versions 10.184,10.185. Since this error is now fixed in brandy.6.zip, that is the reason why your USB printer now works ok.

The USB drivers have no repercussion on shutdown whatsoever (unless they hang the system on shutdown but this is not the case). Your problem must be somewhere else. I also suggest upgrading ACPI to version 30.21.2 which is the most current version today.
From the problems you describe I believe that you have misconfigured eWorkplace/XWorkplace. Make sure that you have automatic restart disabled in eWorkplace/XWorkplace and that you mark all applications to auto-close which you want to have automatically closed on shutdown. This is all configurable via the eWorkplace/XWorkplace Shutdown configuration page (right click on desktop, select "settings" and find the eWorkplace/XWorkplace Shutdown page).

One hint (see also ticket #26): do yourself a favor and remove the USB widget from eCenter / XCenter if you are currently using it. It's a disastrous piece of Software that can have all kinds of negative effects. Instead, use \ecs\boot\usbmsdd.exe.

Lars

[Anchieri]

Ok, I have executed 3 more test: 1) deleted from eWPS the usb widget, without success on shutdown. 2) reinstalled your usbhcd183.zip file and tested shutdown, the system shutdown and power off and REMAIN IN OFF, without rebooting after few second as the last version. 3) reinstalled the brandy.6.zip and over that i copied the USBEHCD.sys of usbhcd183 version, the system shutdown and power off AND REMAIN IN OFF indefinitely. I have to push the power on button to reboot the system, therefore if you don't have modified this program from 183 version to 184 and up I don't now how this can happen.

[Lars Erdmann]

And why did you not use ALL of brandy.6.zip to test?
In short: I cannot help you with your problem as I see no relation to USB whatsoever.

[Lars Erdmann]

Can you backlevel ACPI.PSD to version 3.20.01 and see if the shutdown and reboot problem goes away ?

Can you please check your eWorkplace / XWorkplace shutdown settings ? You can make a screen shot of the relevant settings page and attach here. You should make sure that the auto close feature is properly configured and that the "restart after shutdown" feature is DISABLED.

Also, please attach your config.sys.

Finally, run "\ecs\install\detectei\pciscan.exe -D >output.txt" and attach the generated file output.txt

[Anchieri]

The test that I have executed are on ACPI.PSD version 3.21.02. This is the result: 1) using ALL files contained in the USBHCD183.ZIP file shutdown POWER OFF my system and it remain for ever. 2) using ALL files contained in the brandy.6.zip file shutdown don't function (no steady power off) 2) if I substituted the single phase USBEHCD.SYS taken from USBHCD183.zip, and copied it over that of brandy.6.zip, shutdown function perfectly and the system remain in POWER OFF indefinitely. The setting in eWPS is that I have in image captured001.jpg. For your request I insert also my config.sys file and output.txt from pciscan.

[Anchieri]

output from pciscan

[Lars Erdmann]

Change eWorkplace to power of via APM (that will put APM.ADD into use) and turn off "Delay".
Also: did you restore a Desktop archive recently ?

Other than that I have no idea why shutdown won't work anymore. I have changed NOTHING about the shutdown code. Powering off the system is entirely managed via ACPI.PSD.

The only thing that I could image is that something traps on shutdown and you have automatic reboot enabled. But judging from your config.sys, this is not the case.

Other than that: does all USB stuff work as expected (memory sticks, keyboard, mouse) ?

[Lars Erdmann]

By the way: can you provide me with a valid email address for notifications ? I can then add it as CC so that you get notified.

[Lars Erdmann]

Try anch.zip and see if it makes a difference. That's the only thing I could think of that would make a difference.

[Anchieri]

1) changing the setting of eWorkplace to use APM with or without delay give the same result: POWER OFF don't remain in off but after few second(1:3) the system reboot.

2) the same result after installing the phase from anch.zip, but if I copy USBEHCD.SYS from USBHCD.183 zip file over that of ANCH.ZIP the system POWER OFF and remain in OFF state for ever. If you don't have managed to change something from version 183 to 184 and up, in the USBEHCD.SYS phase I don't now how this can happen.

3) I don't have usb keyboard or usb mouse and my usb sdd disk attach immediately.

4) I have problem to attach my 3 usb hard disk, I have to plug in this disk in different port before the system can detect and attach it. (I have 2 port in front of my PC desktop, 4 port in the rear of pc natively, and other 4 port attached via usb pc card added successively). Before deleting usb widget from eWPS, this disk can be attached only if they were on and plugged in before the boot of the system.

5) my eMail address is: giuliomario.anchieri@…

[Lars Erdmann]

Do you have anything in POPUPLOG.OS2 ?

These drivers were tested by the ACPI developer on about 10 different systems, I have tested them on my 2 systems. I have no problems to power down my system.

I am sorry but I don't see any relation to the USB drivers. You should backlevel ACPI.PSD and see if that makes a difference. The new USB drivers might reveal problems that may have been hidden before. That's all I can say. Again: I have changed NOTHING about the code in USBEHCD.SYS that is called on shutdown.

About disks not attaching properly: please take a trace with this setting in config.sys:
TRACEBUF=512 /M=NW,Q,NDTI /D=ALL

and trace major code 226 (USBEHCD.SYS).

[Lars Erdmann]

USBEHCD,USBUHCD: change shutdown to include a HC reset, don't stop lists for EHCI

[Lars Erdmann]

try anch.2.zip.

[Lars Erdmann]

Forgot: try anch.2.zip with and without the /FS flag. Let me know for either case if it solves your shutdown problem or not.

[Lars Erdmann]

USBEHCD,USBUHCD,USBOHCD: change shutdown

[Lars Erdmann]

Try anch.3.zip with and without /FS flag. For either case, tell me if shutdown works or not. Please use ALL drivers of this package.

[Anchieri]

After testing anch.3.zip with /FS switch, my pc POWER OFF and remain in off, deleting this switch the system reboot after the usual 1:3 second. This means that you have found the mistake, thank for your work on resolving this ticket. I test now the attach detach of my external USB disk, at a preliminary test the problem exist and seem to me that I can attach only one disk at a time after detaching the previous. It is most difficult reattach a disk previously detached.

[Anchieri]

Testing attach/detach(Eject) my external disk I have this result today:

1) is impossible to attach this disk

2) I can attach and detach(eject) only my SDD card

3) Yesterday when one disk come attached after I eject it the LVM phase remain active and unkillable, the system power off and rebooted immediately (1;3 seconds).

4) inserting in config.sys the TRACE statement:

TRACEBUF=512 /M=NW,Q,NDTI /D=ALL

and rebooting the system, it is impossible to attach one disk also using the refreshing removable media and the trace formatter give no data traced. I have activated the trace (TRACE ON) before this test.

5) If this disks is in power on and attached before the boot of the system, it function perfectly, icon in disk folder is for HARD DISK and is not present the eject function in pull down menu, the system shutdown normally and stay in off.

[Lars Erdmann]

For tracing to work you will have to put this in config.sys:

TRACEBUF=512 /M=NW,Q,NDTI /D=ALL
TRACE=ON 226

This is for USBEHCD.SYS. For USBUHCD.SYS you would specify 224, for USBOHCD you would specify 225

Did all this work with 10.183 ? Do not use USB widget, use USBMSDD.EXE in order to manage memory sticks and USB disks.

[Lars Erdmann]

By the way: use the USB resource manager / USB dock and gather information for the problematic external USB disk. Attach here.

[Lars Erdmann]

Another note: you can still eject a disk that has no "eject" function menu entry. You just run "eject.exe X:" from the commandline (assuming that X: is your drive letter for the USB disk).

[Anchieri]

I don't know how to use usbmsdd.exe I inserted it in startup folder without success, I actually are unable to attach my external hard disk after WPS is up. I attach all output from USBDOCK with information on all USB device that come attached to my desktop. The name of this file is: dlink, hama, iomega, lacie, printer, seagate (log). Moreover I don't know how I can send you the trace date, because when I formatted it with trace formatter that is in the trace option folder and closing the window, the file disappear from the directory.

[Lars Erdmann]

1) please read the file "readme.txt" on how to take and save a trace file !

2) how to rebuild the USB mass device monitor:

/* REXX script */
if RxFuncAdd('SysLoadFuncs','REXXUTIL','SysLoadFuncs') then do
   rc = SysLoadFuncs()
end
bootDrive = SysBootDrive()
class='WPProgram'
title='USB-Monitor V1.2'
location='<WP_START>'
app='EXENAME='bootDrive'\ecs\boot\usbmsdd.exe'
startupdir='STARTUPDIR='bootDrive'\ecs\boot'
parameters='PARAMETERS='bootDrive'\ecs\boot\usbmsdd.ini'
icon='ICONFILE='bootDrive'\os2\boot\usbmon.ico'
iconpos='ICONPOS=0,0'
open='OPEN=RUNNING'
minimized='MINIMIZED=YES'
progtype='PROGTYPE=WINDOWABLEVIO'
autoclose='NOAUTOCLOSE=NO'
objectid='OBJECTID=<USB_MON>'
rc = SysCreateObject(class,title,location,app';'startupdir';'parameters';'icon';'iconpos';'open';'minimized';'progtype';'autoclose';'objectid';','R')
if rc==0 then
   ret = 0
else
   ret = 1
return ret

Save this into a file with file extension .cmd and run it from a commandline. It's a REXX script that will create a Program object in the startup folder with the correct parameters.

3) Exactly WHICH USB device is giving you problem ? Iomega ? Lacie ? Seagate ? Why do you force me into guessing each time ?

[Lars Erdmann]

try 10.186, it's on Hobbes.

[Anchieri]

I have installed all files in your 186 version but it's a disaster, I can't attach any of my external disk, if attach it before to boot the system only one or two of this three disk come attache, in many case nothing of this disk come attached also if attached before the boot time. I have noted that if I switch of the pc after it power off in the next boot two disk come attached, never all three. For testing porpose I have copied over the USBEHCD.sys of 186 the USBEHCD.sys of 183 at boot time all three disk come attached and after the wps are up I receive three window messages that inform me that disk J K and M are attached. Never receive this window messages using the USBEHCD.sys of 186 version.

As for your request I attach 2 trace formatted dump:

trace05-3-of-3-ok-usbehcd183.ftf that contain the trace from USBEHCD.SYS of version 183 and trace04-2-of-3-nolacie.ftf that contain trace from USBEHCD version 186 that attach only two disk of three. (the Lacie disk is not attached) but in many case is not attached Iomega disk. In one case nothing of this disk come attached.

I was at your disposal for other test that you want I can do.

[Lars Erdmann]

You also have to update the corresponding .tff files for the trace to be readable:
USBUHCD.SYS -> TRC00E0.TFF
USBOHCD.SYS -> TRC00E1.TFF
USBEHCD.SYS -> TRC00E2.TFF

The .tff files need to go into directory \os2\system\trace. They contain formatting templates that allow the trace formatting utility tracefmt.exe to display traced data in human readable form.

Don't forget that if you backlevel a driver that then you will also need to backlevel the corresponding .tff file !

[Anchieri]

new test:

1) I have recopied all file in version 186 to os2\boot and trc file in os2\system\trace I attach all my three disk and power on it, then rebooted the system obtaining the trace02.ftf file that as more than 800kb and so I can't attach directly I can only send you via eMail or using other way.(it is not possible obtain the non formatted file because the trecefmt give me an error during save)

2) I copied over USBEHCD 186 those of 183 and so for trace TRC00E2.tff and repeat point 1) obtaining trace01.ftf that I can attach due to small amount of date.

[Anchieri]

I forgot to say you that during point 2) above all three disk come attached and on point 1) only one (the J-lacie disk).

[Lars Erdmann]

It does not help if you send me a trace for something that is known to work.

Now use this in config.sys TRACEBUF=64 /M=NW,Q,NDTI /D=ALL TRACE=ON 226

and take the trace for the 10.186 USBEHCD.SYS driver. This will reduce the trace buffer to 64k and therefore the trace record will be significantly shorter.

[Lars Erdmann]

I can see that you have one EHCI (USB 2.0) controller that supports 5 ports and another one which supports 8 ports.

Question: does the NEC plug-in card have those 5 ports and on-board (Intel controller) you have 8 ports ?

If it fails on the NEC card, can you plug in your disks to the on-board ports or vice versa ? I need to know WHICH one of the EHCI controllers has problems.

[Lars Erdmann]

USBEHCD: remving port power on directly after reset (copy and paste error)

[Lars Erdmann]

try anch.4.zip. It's enough to just replace USBEHCD.SYS and TRC00E2.TFF.

[Lars Erdmann]

Please, only trace major code 226 (USBEHCD.SYS). Otherwise the trace log will become too large.

[Anchieri]

I respond to your question:

1) The NEC pc-card has 4 port on the rear and 1 port internal, the internal INTEL has four port on the rear and 2 port on front of my PC

2) I have downloaded anch4 file and installed only USBEHCD with his trace file, and deleted 225,224 from config.sys statement.

3) I obtained the TRACE03.FTF file that I have zipped as TRACE03.ZIP that I attach (it is more than 256k, after I have deleted more than 3000 line from it).

4) only 2 of 3 disk come attached: J (LACIE) and K (IOMEGA) M (SEAGATE is not attached) all this disk is attached to the rear port of internal INTEL mother board. The window messages that give me information that disk x: is attached is never showed.

5) I has never tested if NEC pc-card can attach my disk. I have tested with front port with mixed result one time ok the disk attach, other time does not attach.

[Lars Erdmann]

Ok, from trace03.ftf:

1) I can see 5 USB 2.0 devices attaching to the Intel controller. They properly attach. Please check with USB Dock that indeed you can see 5 USB 2.0 devices.

2) You did not attach any devices to the NEC controller as far as I can tell.

Are the observations correct ?

Did you initiate a data transmission on any of the USB 2.0 devices ? The repetitive commands that you saw in the trace are data transmissions.

By the way: you will ONLY get notification that a disk attaches if it is NOT plugged in on boot. If a disk is plugged in on boot you will get NO notification but you will get drive letters showing them as fixed disks and they will have a drive letter. Therefore what you observe is expected behaviour. If you get 3 drive letters (J: K: and M:) that means that 3 disks attached properly and I don't understand what problem you have.
If they attached properly (check with USB dock) you might need to run chkdsk on the disks. That's what I have to do if I do not properly detach a disk (for example if I unplug it without ejecting or if the system does not shutdown properly as was your case).
Running chkdsk on a drive is especially necessary for JFS formatted disks if they were not ejected properly before detaching them.

Please now try and attach a disk AFTER the system has booted up (and as I said: run chkdsk on all your disks LACIE,IOMEGA and SEAGATE).

[Anchieri]

When I affirm that a disk is not attached is because in disk folder is not present his letter and so I was not able to give the command: chkdsk x: /F to check his integrity. In the previous messages I confirm that only 2 of three letter are present in disk folder, the M: (seagate disk) is not present and for me is not attached, if you confirm me that in trace all three disk are attached I don't know how the third letter M: is not present in disk folder. In the NEC pci card I have plugged in only my usb printer that is not power on during the boot process and so is not visible in the trace. I use OS/2 from 1998 version 1.2 and now I have installed eCS 2.1GA I am a client of Sinergy system by 2001 with eCS1.0 and up and if I have difficulty on use of os/2 trace system is because I never had the necessity to use it before now. I proceed to test if the disks come attached after the wps is up and running.

[Anchieri]

New test: The file used is the same as previous test, I have attached my iomega disk to one port of NEC pci card the other two remain on INTEL internal mother board.

ALL three disk come attached but not as hard disk but as removable disk with all the capability of removable ie to eject it. During the chkdsk of iomega disk (K:) I received the strange error that I document you in the image usb01.jpg that attach. moreover after the wps is up I received three Windows messages that inform me that the (J, K, M) disk are been attached.

I presume that if not all disk come attached I don't receive this messages, Or in other case if the disk are attached as Hard disk this message don't are issued as in the previous comment.

[Anchieri]

TRACE04 contain the trace at boot with thre disk one of that (IOMEGA) attached to NEC pci card

[Lars Erdmann]

1) usb01.jpg: the JFS filesystem on the USB drive is corrupted. You need to run "chkdsk K: /F" to fix it.

2) trace04.ftf: I can see one USB 2.0 device attaching to the NEC controller. I can also see 4 USB 2.0 devices attaching to the Intel controller. Is this correct ? Can you see these 5 devices in USB dock / USB resource manager ? You always refuse to tell me if all these devices properly show up in USB dock / USB resource manager ...

3) USBD.SYS is known to cause problems if you have too many USB drives attached on boot. Sorry but I cannot fix that, I have the same problem with one USB stick that I have that will not properly attach during boot. If a device won't attach just detach it and reattach it and it should work ok.

Sorry but I don't see anything else I could do. What happens if you have all 3 USB disks attached on bootup when you use the 10.162 drivers ? Does that work ok ?

To reiterate: if a USB disk properly attaches already on boot, it will get a non-removable disk icon and NO attach message will show. If a USB disk is attached after the bootup completes, it will get a removable disk icon and an attach message will be displayed. It has nothing to do with "not all disks ...".

[Lars Erdmann]

USBEHCD: modifying end of HC reset (remove some waits, reorder schedule enablement)

[Lars Erdmann]

Try anch.5.zip.

[Lars Erdmann]

USBEHCD: on HC reset only wait for POTPGT when ports are not permanently powered

[Lars Erdmann]

Go directly for anch.zip (the newest file so to say).

[Anchieri]

I respond to your question in comment40:

I have five device attached: one disk IOMEGA attached to NEC port, on INTEL port I have two disk LACIE and SEAGATE and other two card reader (1 internal and 1 external) this is displayed using USBDOCK.

If I unplug the external card reader, I can attach without problem one or two disk, very difficult to attach the third disk.

I have tested all 2 zip file ANCH.5 and ANCH but either stop booting before reaching WPS if I have one or more disk attached and in power on, to boot I have to power off all disk before booting the system.

I proceed to test with the ibm 162 version to verify if it function properly or not for give you the result.

[Anchieri]

I have reinstalled version 162 and i find a big problem on your version, if you analyze the usb01.jpg you see this messages:

JFS0148: CHKDSK Unrecoverable error reading M from K:. CHKDSK CANNOT CONTINUE

I have attached the image usb02,jpg that contain the pm window that I used to check the same disk K: and using the original version chkdsk came to end without any error can you give me why your version cannot?.

[Lars Erdmann]

No. I don't know and I cannot help you. I suggest you use the original 10.162 drivers.

[Anchieri]

I have reinstalled the v. 162 and I can reformat my disk. You can close this ticket.

[Lars Erdmann]

shutdown problem fixed with 10.187, there exists a new ticket for the remaining problems