Keyboard does not work, mouse does

[MarkV]

I installed these new drivers and rebooted.

When the new system came up, I had a mouse that worked, but the keyboard was not functional at all.

I'm using a trendnet vga/usb kvm. I'll run another test around the kvm.

[MarkV]

tracelog when using original usb drivers

[David Azarewicz]

The log you provided does not contain any useful information since it shows that you are using extremely old drivers. Please make sure all your drivers are up-to-date before capturing a log or trace. Also, please only attach logs and traces for the reported configuration.

I will need a startup trace for the USBUHCD.SYS driver. Please see: ​http://svn.netlabs.org/usb/wiki/Troubleshooting

[MarkV]

It does not matter if the keyboard goes through the KVM or not. The keyboard doesn't work.

As for providing useful information, I'll have to figure out how to do that without the use of a keyboard. It might take me a bit of time to put something in place.

[David Azarewicz]

There are a number of peculiarities with your ticket. The strangest one is that I have the exact same hardware in my test lab (as reported by PCI.EXE) and it has absolutely no problems. I wonder if you have all of the USB related drivers updated to the current level. Since they all communicate with each other, there is a chance that something will not work if you mix incompatible versions.

Which version of eCS are you using?

Here is a list of the current relevant drivers as reported by the bldlevel command:

IBMKBD.SYS: Revision: 10.165 Date/Time: 15 Jun 2013 15:41:52

USBKBD.SYS: Revision: 10.959

USBMOUSE.SYS: Revision: 10.959 Date/Time: 18 Mar 2013 09:24:31

AMOUSE.SYS: Revision: 2.75.01 Date/Time: 26 Aug 2012 10:29:21

USBHID.SYS: Revision: 10.959

I suggest that you make sure you have the current versions of these drivers installed, particularly IBMKBD.SYS, USBKBD.SYS, and USBHID.SYS.

Also, if you plug the USB keyboard directly in (not through the USB KVM switch), does it show up in USBDOCK?

After booting, if you unplug the keyboard, and plug it back in, does it work then?

After booting, if you unplug the KVM USB cable and plug it back in, does it work then?

[MarkV]

I've managed to create scripts to help me capture the information we need without using the keyboard. Though the formatted trace file may not have the proper file extension. Since I had to select an existing file, I did not remember what file extension the program would have used.

Please find new attachements for the log and trace. I'll review your questions above, and see if I can find answers for you.

[MarkV]

Trace and log files using the 11.04 release drivers

[MarkV]

Here are some answers to your questions. I'm using eCs 2.0 Unplugging and plugging the keyboard back in does not help. It's the same whether through the KVM or directly into a USB port.

Each time I plug the keyboard back in (KVM or not), I do see the USB block indicator on the bottom right, appear just as I would normally see if any USB device were plugged in.

I have the same number of usb devices reported in USBDOCK using either set of drivers. Though I do not think the information for each of them is the same.

The mouse and keyboard USB interface in USBDOCK all report 1.1.

My printer is reported as USB 2.0.

I"m looking for new drivers, but I have not yet found any. The AMOUSE site appears to be under construction, ao I'm not likely to be able to find anything there anytime soon.

I've looked on eComstation and in hobbes, but I',m not finding any new keyboard usb driver. If you know where else to look please share.

[MarkV]

I"m going to look into the package ACPI 3.22.2 for newer drivers.

[MarkV]

I've installed the ACPI 3.22.2 package including all the drivers.

Unfortunately this didn't help.

Is there anything else I can do to help diagnose this problem?

[David Azarewicz]

Well, given these facts:

1. I have the same hardware in my test lab and it works just fine.
2. Your mouse works, which means that the USB 11.04 drivers are working fine.
3. You reported that the keyboard and mouse (and printer) devices are correctly showing up in USBDOCK, which means that the USB 11.04 drivers are working fine.

Everything points to something else wrong with your setup/configuration.

Things you can try:

1. Post your CONFIG.SYS. It is unlikely, but perhaps I can see something wrong in there.
2. Download ​http://88watts.net/download/TestEcs.7z, unzip it and burn it to a CD. Boot it and see if it works. Beware that none of the eCS drivers are actually used until it boots past the eCS logo screen, so make sure it gets to the graphical screens before testing to see of the keyboard works. Beware that this ISO will expire, and all the drivers contained in it will expire in 30 days. Also, even though this ISO looks like an install disk, it will not actually install anything.
3. Download ​http://88watts.net/download/USB-test-11.05-20130918.zip, unzip it and run the update.cmd script inside. This contains current versions of some of the class drivers that were not in other distributions. Beware that the drivers in this package are "Test Builds" and will expire in 30 days. I recommend that you backup all the USB drivers in your \OS2\BOOT directory before installing this package.

[MarkV]

Zipped Config sys file for your review.

[MarkV]

I've booted with the TestEcs ISO. When I get to the install screens, the mouse works, but the keyboard does not. I've selected the management console, and then the command prompt. The keyboard doesn't work in the command window.

I tried to unplug and plug the keyboard back into the system, and it didn't help. I used the KVM port, a USB hub port, and a direct usb port on the machine.

Do you still want me to try the usb test 11.05 files?

[Lars Erdmann]

You have 4 UHCI controllers which should take the job of providing USB legacy support, 1 controller for 2 ports each where you obviously have 8 USB ports over all. For some reason I don't know your BIOS only enables USB legacy support on 1 of those 4 UHCI controllers. For as long as the HC drivers have not taken over USB legacy is the only thing that will make your mouse and keyboard work. Eventually, the HC drivers will take over: the EHCI driver will detect that it's a USB 1.x device and hands it over to the UHCI controller and I can see that one of the UHCI controllers sees a device connected and the port is enabled which looks correct.

Here is what you can test: try all of the 8 USB ports for your keyboard and see on which 2 ports you can get your keyboard to work in USB legacy mode.

[David Azarewicz]

Replying to MarkV:

I've booted with the TestEcs ISO. When I get to the install screens, the mouse works, but the keyboard does not. I've selected the management console, and then the command prompt. The keyboard doesn't work in the command window.

The fact that the mouse works and the fact that both the keyboard and the mouse show up in usbdock proves that the USB driver are working fine. The problem is somewhere else -- not with the HC drivers.

I tried to unplug and plug the keyboard back into the system, and it didn't help. I used the KVM port, a USB hub port, and a direct usb port on the machine.

Do you still want me to try the usb test 11.05 files?

No, not necessary.

[David Azarewicz]

Replying to erdmann:

You have 4 UHCI controllers which should take the job of providing USB legacy support, 1 controller for 2 ports each where you obviously have 8 USB ports over all. For some reason I don't know your BIOS only enables USB legacy support on 1 of those 4 UHCI controllers. For as long as the HC drivers have not taken over USB legacy is the only thing that will make your mouse and keyboard work. Eventually, the HC drivers will take over: the EHCI driver will detect that it's a USB 1.x device and hands it over to the UHCI controller and I can see that one of the UHCI controllers sees a device connected and the port is enabled which looks correct.

Here is what you can test: try all of the 8 USB ports for your keyboard and see on which 2 ports you can get your keyboard to work in USB legacy mode.

Sorry, but his problem has nothing to do with legacy USB, or the handover, or even the HC drivers, and he is not trying to get it to work in legacy mode.

[David Azarewicz]

You mentioned that the detailed report in usbdock was different between using the 11.05 drivers and the old IBM drivers. There really should not be a difference. If there is a difference, it might be useful to look at the differences. Boot using the old IBM drivers and capture the detailed report for both the keyboard and the mouse. Then boot with the 11.05 drivers and capture the detailed report for both the keyboard and mouse. Post the reports to this ticket.

[MarkV]

The documentation clearly tells me how to capture a detailed report for the 11.05 drivers. Should I do the same thing for the old IBM drivers, or is there another procedure to create that detailed report? When I get clarification on this point, I'll work to get the reports.

And for what it's worth (even if the information isn't useful)...

The keyboard doesn't work in any of the 6 ports that I have. If there are supposed to be 8, then two did not make it out of the case. Clearly the keyboard is recognized each time by the USB system since in the lower right corner the usb icon briefly appears after it is attached each time.

And also, the 11.05 drivers work the same for me as the 11.04 drivers.

[David Azarewicz]

To get a detailed report for a USB device open "Local System->Programs->Utilities->USBDock", then right click on the device and click "View report". Select all the text and copy it to a file or directly into the ticket.

Another probably easier way is:

1. To start out clean, go to d:\ecs\system\usbdock and delete all the log files in there. d: is your boot disk. del *.log

2. Make sure all the USB devices you are interested in are plugged in.

3. Open "Local System->Programs->Utilities->USBDock". Just opening the program writes a log file for every attached device to d:\ecs\system\usbdock, where d: is your boot disk.

4. Zip up all the log files in that directory and attach it to this ticket.

Do this for both the old IBM drivers and the 11.05 drivers. This may be kind of difficult without a keyboard, but you can do a lot with just a mouse.

[MarkV]

I believe I've captured what you want. Please find two new attached ZIP files that contain the log files from the USBDOCK folder. One while running under the orignal drivers, and another set of log files created under the new 11.05 drivers. Please let me know if this isn't quite right.

[MarkV]

USBDOCK log files under the original drivers

[MarkV]

USBDOCK log files created under the 11.05 drivers

[David Azarewicz]

Thanks, that is exactly what I wanted.

Both outputs are exactly the same. They both show exactly the same four devices attached with exactly the same characteristics.

It indicates the following devices are attached to your system:

1. Device 046d:c318 is a Logitech USB Illuminated Keyboard. However it is strange that it has both a keyboard endpoint and a mouse endpoint. This implies that it is capable of being both a keyboard and a mouse.

2. Device 04f3:0216 is an Elan Microelectronics Corp mouse.

3. Device 04f9:003d is a Brother printer.

4. Device 10d5:55a4 is some unknown human interface device with 2 endpoints from Delkin Devices, Inc. Delkin makes memory card readers, but it is not defined correctly for that.

I do not see a KVM switch attached to the system.

Still, I don't see anything wrong in the data provided. Everything appears to be working correctly.

Do you have a different keyboard you can try?

Can you unplug the 10d5:55a4 device or is it built in?

There are reports in the Linux forums of the Logitech Illuminated keyboard not playing well with other devices on the same USB bus. The keyboard doesn't work in those cases and the fix is to plug it into a separate port on a separate bus. You can try plugging the keyboard into different ports.

[David Azarewicz]

Just some notes so they are in the record so I can remember...

The fact that USBDock can see the keyboard and query its configuration means that the HC drivers find the device and are talking to it. This eliminates the HC drivers from the list of suspects.

The next questions are:

1. Is the USBKBD.SYS driver claiming the keyboard?
2. Is USBD asking for someone to claim the keyboard?
3. Why is this the one and only system with this problem?

[David Azarewicz]

I just took a look at your CONFIG.SYS and I noticed these things:

1. DEVICE=C:\ECS\SYSTEM\AMouse\AMOUSE.SYS SMP The AMOUSE.SYS driver in this directory may not be the up-to-date one. The SMP parameter does not exist and is not needed in the newest version.

2. DEVICE=C:\ECS\SYSTEM\AMouse\USBMOUSE.SYS The USBMOUS.SYS driver in this directory may not be the up-to-date one.

3. DEVICE=F:\Tame\usbresmg.sys The USBRESMG.SYS in this directory may not be the up-to-date one.

I suggest that you make sure that all of the above drivers are at the 11.05 level and see if that works.

[David Azarewicz]

The latest amouse.sys

[MarkV]

The Delkin device is unknown to me. I do not have any other USB devices other than the printer, keyboard, & mouse.

I've uninstalled Tame, so that driver has gone away.

I have been unable to find the latest AMOUSE driver, so I'm glad you attached one here.

I do not have another keyboard to test with. THis is my first USB keyboard/mouse configuration. I'll see if I can scrounge up another USB keyboard to test with. I've tried all my different USB ports, and the keyboard doesn't work in any of them.

I'll update the AMOUSE and USBMOUSE drivers and get back to you.

[MarkV]

The new AMOUSE driver doesn't work with either the ECS\SYSTEM\AMOUSE\USBMOUSE.SYS or the OS2\BOOT\USBMOUSE.SYS files. Each give me a SYS1201 error at boot time.

SYS1201 The device driver yadda/USBMOUSE.SYS specified in the DEVICE statment on line 43 of the config.sys file was not installed. Lines 43 ignored.

And of course, I don't have a mouse on the desktop, but I do have a keyboard.

Restoring the original AMOUSE driver and the driver line with SMP switch restores my system.

Perhaps I need the full AMOUSE deliverable and not just the SYS driver file?

I can try the new AMOUSE driver with the 11.05 USB drivers if you think it would be worth a try. I'll see if I can arrange that test later tonight.

[MarkV]

Using the new AMOUSE driver with the new 11.05 drivers gives me the same SYS1201 error for the USBMOUSE.SYS driver on boot. Though I think this actually hangs before I get to the desktop.

[MarkV]

10/30 - Using 11.05 drivers...

Ok, I've replaced my Logitec keyboard with a borrowed Cherry brand keyboard. And it is now working. It seems the problem is the keyboard.

What are the chances we can get these drivers to work with the Logitec keyboard? Or am I looking at getting a permenent keyboard replacement that isn't Logitec brand?

I really like my Logitec keyboard. It feels great, and it lights up when plugged in.

If this isn't going to happpen, then I'll need to start looking for a replacement keyboard.

On another note, can I get someone to share a link where I might get the latest public release of the amouse driver? I can't seem to find it on my own.

[David Azarewicz]

I would say that there is a very good chance that it can work. We just need to figure out exactly where the problem is. The fact that a different keyboard works is helpful. It would also be helpful to know if that Logitec keyboard works on a different system. I believe that the problem is related to the fact that that specific keyboard is reporting itself as both a mouse and a keyboard. The problem could also be that the keyboard is conflicting with one of the other USB devices in the system.

The problem is not because it is a Logitec brand keyboard, the problem is with that specific keyboard. Another Logitec keyboard will probably work fine.

Unfortunately, there are currently no tools available for extracting the needed information from the drivers in a remote situation like this. It would be easy if I had the problem in my test lab. I will have to figure out a way to get the required information and then get back to you. It may take some time.

[David Azarewicz]

The Amouse driver that I attached to this ticket is the latest publicly released driver.

[David Azarewicz]

Just out of curiosity, what happens if you use the /DEVICE:1 switch on the USBKBD command line in your config.sys?

DEVICE=D:\OS2\BOOT\USBKBD.SYS /DEVICE:1

Does your Illuminated keyboard work then? You must reboot to make the switch active.

[MarkV]

Use of the switch does not appear to affect the system at all.

Keyboard still works with old drivers, but not new drivers.

[MarkV]

Replying to dazarewicz:

I would say that there is a very good chance that it can work. We just need to figure out exactly where the problem is. The fact that a different keyboard works is helpful. It would also be helpful to know if that Logitec keyboard works on a different system. I believe that the problem is related to the fact that that specific keyboard is reporting itself as both a mouse and a keyboard. The problem could also be that the keyboard is conflicting with one of the other USB devices in the system.

The problem is not because it is a Logitec brand keyboard, the problem is with that specific keyboard. Another Logitec keyboard will probably work fine.

Unfortunately, there are currently no tools available for extracting the needed information from the drivers in a remote situation like this. It would be easy if I had the problem in my test lab. I will have to figure out a way to get the required information and then get back to you. It may take some time.

I can always send you my keyboard if it would help?

[David Azarewicz]

First download ​http://88watts.net/download/USB-test-11.06-20131031.zip and see if your problem is already fixed because of other changes. Use the update.cmd script in the zip to make sure you install everything correctly, including the trace files.

If your keyboard still does not work, please capture a trace for me.

1. Go to ​http://svn.netlabs.org/usb/wiki/Troubleshooting and setup your system to capture a "Startup Trace file". Use trace code 227 instead of any of the trace codes shown on that page. For example, the TRACE=ON command will be: TRACE=ON 227

2. If you create a dummy file before you reboot (like KBD.TFF), you will be able to capture the trace entirely without the need for a keyboard.

3. Reboot your system with the illuminated keyboard attached.

4. Navigate to Local System->Programs->Utilities->Problem Determination Tools->Trace Options->Trace Formatter and capture a formatted trace as described on the troubleshooting page.

5. Attach the formatted trace file to this ticket.

[MarkV]

I've used the 11.06 test provided. This updates the config.sys too. I was not expecting that. And the results included two SYS1201 errors. One for USBMOUSE.SYS and the other for USBRESMG.SYS, both in the os2\boot directory. The mouse worked fine.

It turns out that the formatted trace file expects FTF extension. Minor hiccup but fine anyway. Please find attached the KBD.FTF file for tracing 227 as requested. Let me know if you need anything else.

[MarkV]

Trace using 227, and 11.06 test

[David Azarewicz]

Please try the test build at ​​http://88watts.net/download/USB-test-11.06-20131211.wpi and let me know if changes anything.

[David Azarewicz]

11.06 is the current version and likely fixes your problem.

[MarkV]

I have both files, the one here, and the one on the beta zone. Thinking validating the one on betazone is what would be best, I'll try that one first unless you tell me otherwise here. Thanks for your hard work on this. I appreciate it.

[MarkV]

I've installed the 11.06 GA warpin package. The logitec keyboard model Y-UY95 still does not work. But plugging in a Cherry model does. So we've made some progress.

Since I really like the logitec keyboard, I've backed out the 11.06 update. Is there anything more you want me to provide in order to make progress on this problem?

[Lars Erdmann]

Test drivers (serialization of control transfer calls from USBKBD and USBMOUSE)

[Lars Erdmann]

1) What was the last version of drivers for USBKBD.SYS, USBMOUSE.SYS and USBHID.SYS where the Logitech keyboard was working as expected ?

2) try the contents of ticket67_1.zip (it contains USBKBD.SYS, USBMOUSE.SYS, USBHID.SYS and the associated SYM files). Make sure you backup your existing drivers/sym files and replace the WHOLE set of drivers contained in ticket67_1.zip.

If they do work or at least if you have an additional USB keyboard to properly operate your system : open OS/2 hardware manager (Local System->System configuration->Hardware manager). Make sure you open it in TREE view. Now check below the "USB Keyboard Driver" and also "USB Mouse Driver". Does there appear some device under either or both of these entries ? What device names show up under "USB Keyboad Driver" and/or "USB Mouse Driver" ?

[MarkV]

1) USBKBD & USBHID are the IBM 10.162, the USBMOUSE is Noller & Breining 2.70 driver.

2) Installing your provided test drivers does indeed give me mouse and keyboard functionality.

Please find attached a PNG graphic showing the USB areas of my hardware tree. Perhaps seeing it is better than me trying to describe it.

For now, I'm keeping your updated drivers running in my system since they work.

[MarkV]

USB Hardware tree using ticket67_1 test drivers

[Lars Erdmann]

Thanks for the screen shot. That tells me that you were also using the KVM switch.

Are you sure you also replaced USBMOUSE.SYS ? Also make sure that in config.sys you load my updated USBMOUSE.SYS driver.

Anyway these drivers are not ready for prime time but we have now a good idea of where the problem is. But if they work ok for you for the time being then of course you can continue to use them.

About file size: these drivers were built with full debug info. But when loaded they will occupy about the same amount of memory than the retail drivers.

[MarkV]

Happy New Year!

Bummer. You are correct. I did not use the delivered USBMOUSE.SYS. So I activated it in the config.sys file and rebooted. The system complained on the way up that it could not load the USBMOUSE.SYS driver and asked that I hit the enter key to continue. I used the BLDLEVEL command to confirm that the driver I tried (and failed) to load was indeed your test driver.

After the system finished booting, I did not have a mouse. I was able to navigate using the keyboard just fine to get my config.sys settings restored to the USMOUSE.SYS driver that came with AMOUSE. As noted above, the USBMOUSE.SYS is the Noller & Breining version.

I hope this helps.

[Lars Erdmann]

Something went wrong, the delivered USBMOUSE.SYS loads for sure as I am using it on my test system. Make sure that you modify config.sys to point to the correct directory. AMouse has its own copy of USBMOUSE.SYS at a different location. Also make sure you only have ONE DEVICE=...\USBMOUSE.SYS statement in config.sys. Also, just to accomplish what follows below, replace AMOUSE.SYS with the normal MOUSE.SYS.

Once you get my copy of USBMOUSE.SYS to load correctly, even if you can only get the keyboard to operate correctly, I would be interested in the OS/2 HW manager screenshot with my USBMOUSE.SYS loaded. If it is not too much hassle.

[MarkV]

I took another look at my config.sys and the drivers in the os2\boot directory. All was as I found the last time. But as I looked at the config.sys I noted that the USBMOUSE.SYS driver was loaded after the AMOUSE.SYS driver on my base installation. During my testing, it was reversed. So I moved the OS2\BOOT device line for your test driver after amouse and rebooted. It boots just fine. I guess the order the SYS files are loaded is important though I do not recall how boot decides to load device drivers in any particular order.

Please find a new attachment for the graphic of my hardware tree as it represents my system after the reboot.

[MarkV]

After successful load of the LARS test USBMOUSE.SYS driver.

[Lars Erdmann]

As to load order: there is a loading stage where all device drivers are invoked on their INIT entry point. Then there is a second loading stage where all device drivers are invoked on their INIT_COMPLETE entry point.

All device drivers (loaded via DEVICE=) are called on their INIT entry point in exactly the order as they are listed in config.sys. Then, all device drivers (loaded via DEVICE=) are called on their INIT_COMPLETE entry point in exactly the reverse order as they are listed in config.sys.

There are additional ordering rules regarding drivers loaded via BASEDEV= in contrast to those loaded via DEVICE= (drivers loaded via BASEDEV= are loaded before drivers loaded via DEVICE=) and then there is yet another load order regarding all drivers loaded via BASEDEV= (it depends on the file extension in this order from first to last: .SYS, .ADD, .DMD, .FLT). But in general, the BASEDEV= drivers follow the very same rule for invocation of INIT and INIT_COMPLETE entry point as the DEVICE= drivers.

As to your screenshot: am I assuming correctly that you can only operate a USB keyboard via your KVM switch but not a USB mouse ? Am I also assuming correctly that you had the USB keyboard and USB mouse directly attached to the system and NOT via the KVM switch ?

[MarkV]

Thanks for the information regarding DEVICE loads. I'll tuck that away for the future.

As for the other question...

I always perform my first tests with the keyboard & mouse plugged through the KVM I assume that the KVM won't cause a problem. But if I had a problem with the keyboard, I've tried using direct ports on the box in follow up tests.

In this case, the keyboard & mouse worked properly through the KVM. I did not rearrange their connections and test any further.

The diagram shows what my system hardware tree looks like when the keyboard & mouse are connected through the KVM.

[David Azarewicz]

Please try ​http://88watts.net/download/USB-test-11.07-20140111.wpi and tell me how it works on your system. The USBHID.SYS, USBMOUSE.SYS and USBKBD.SYS are a matched set. You MUST change all 3 of these drivers as a set. You cannot mix versions of these 3 drivers. The WPI installs them as a set for you.

[MarkV]

It did not work. The problem with the keyboard not functioning still remains when using these test drivers.

[Lars Erdmann]

USBHID: on individual client registration DO NOT call client driver to check for service

[Lars Erdmann]

try david_2014_02_12_2.zip. Backup existing drivers and replace ALL 3 drivers (USBHID.SYS, USBMOUSE.SYS, USBKBD.SYS).

[MarkV]

These three drivers work just fine. I'm leaving them installed. Sounds like we are making good progress. I can't thank you enough.