Author Topic: Anyone tried the updated USB Controller drivers on hobbes yet?  (Read 4665 times)

Pete

  • Hero Member
  • *****
  • Posts: 865
    • View Profile
Hi All

I have not done any extended testing with these drivers - only just installed them - but can report the usbehcd and usbohcd drivers seem to work fine. However, I've only got a mouse and drive to test with at the moment.

Just in case anyone else wants to try them the current url is  http://hobbes.nmsu.edu/download/pub/incoming/usbxhcd.zip

Thanks to Lars for these updates.

Regards

Pete


DougB

  • Sr. Member
  • ****
  • Posts: 407
    • View Profile
I still need to report it, but they don't work for me on my Asus M3A78-EM board with quad core AMD Phenom processor.

Installed to eCS 2.0 Silver, which hangs on the "original" USBOHCD.SYS when attempting to boot with /SMP /APIC.

I get a TRAP 000d in the "new" USBOHCD.SYS when I attempt to boot with /SMP /APIC.

If I attempt to boot using /SMP /PIC (which works with the "original" USB drivers), the boot stops just before USBSER.SYS reports that it attached COM:2. REMing that line doesn't help.

Pete

  • Hero Member
  • *****
  • Posts: 865
    • View Profile
Hi Doug

I've been testing these drivers using the /PIC switch as that is how my system prefers to work with acpi.psd but I can confirm that these drivers Trap - not sure which 1 causes the trap due to splash screen - when using the /APIC switch.

Guess I'll have to rummage around my memory and remember how to capture the Trap screen to floppy and send it to Lars.

Regards

Pete








DougB

  • Sr. Member
  • ****
  • Posts: 407
    • View Profile
I created an incident, per instructions, on the eCS bug tracker. Ctrl-Numlock-Numlock (or Ctrl-F10-F10) doesn't do anything, so a dump to diskette is not possible. I took a photo of the trap screen, and submitted that.

Joachim

  • Newbie
  • *
  • Posts: 36
    • View Profile
I created an incident, per instructions, on the eCS bug tracker. Ctrl-Numlock-Numlock (or Ctrl-F10-F10) doesn't do anything, so a dump to diskette is not possible. I took a photo of the trap screen, and submitted that.

Thanks Doug,

this is appreciated.

Regards,

Joachim Benjamins

ivan

  • Hero Member
  • *****
  • Posts: 593
    • View Profile
On comp.os.os2.misc news group Lars posted the following.
Quote
Hallo,
 
in contrast to what the readme.txt says, please only report problems here even if you are an eCS
user ! These drivers were released by me for testing purposes.
Raising a BUG in the eComStation BUG tracker is not helpful here as nobody is going to look at it.
Eventually, this code will be handed over to eComStation and merged with improvements/corrections
from other developers.
 
Lars

ivan

DougB

  • Sr. Member
  • ****
  • Posts: 407
    • View Profile
Quote
Raising a BUG in the eComStation BUG tracker is not helpful here as nobody is going to look at it.

I suggest that since the package specifically says to post results at the eCS bug tracker, that Lars had better go there to look. I, personally, don't follow the comp.os.os2.misc news group.

Joachim, I would suggest that you make sure that it happens.

DougB

  • Sr. Member
  • ****
  • Posts: 407
    • View Profile
I just tried USBUHCD.SYS on my antique ThinkPad A22e (only one USB 1.1 controller).That one seems to work good with a new installed eCS 2.0 GA. It is SLOWWW, but that is to be expected...

RobertM

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 2034
    • View Profile
    • A.I.BuiltPC - using OS/2 Warp Server & eComStation for Custom Web and Database Solutions
I just tried USBUHCD.SYS on my antique ThinkPad A22e (only one USB 1.1 controller).That one seems to work good with a new installed eCS 2.0 GA. It is SLOWWW, but that is to be expected...

Heh... the other day I had a W2K3 server to back up... the beast has some decent specs, especially for it's time. But, because it's actually a rather ancient box, it also had only USB 1.1 - I only had 78GB to back up. After four hours, I stopped the backup and simply backed it up through the network through a machine with USB2.0 ports.

And no, I wasn't sitting there waiting... I had other work to do there, and figured in the 8 hours I was going to be there, maybe USB1.1 would have made the grade... at the halfway point I realized I needed a lot more than 8 hours and went to "Plan B"

Ah... painful memories.  ;D
|
|
Kirk's 5 Year Mission Continues at:
Star Trek New Voyages
|
|


Joachim

  • Newbie
  • *
  • Posts: 36
    • View Profile
Quote
Raising a BUG in the eComStation BUG tracker is not helpful here as nobody is going to look at it.

I suggest that since the package specifically says to post results at the eCS bug tracker, that Lars had better go there to look. I, personally, don't follow the comp.os.os2.misc news group.

Joachim, I would suggest that you make sure that it happens.

Sure, I will notify Lars. He actually is a frequent visitor of the bugtracker so I don't really see a problem in using it for bugreports on these drivers, even if they're not mature yet.

Thanks,

Joachim

DougB

  • Sr. Member
  • ****
  • Posts: 407
    • View Profile
More results, using the contents of usbxhcd2.zip.

I still get the hang at boot time, if I attempt to use the updated USBEHCD.SYS. It happens immediately after VVGA.SYS (I use VVGA.SYS in place of VSVGA.SYS, so that I can use Win/OS2), and the next thing that should happen is that USBSER.SYS should inform me about what it configured. REMing that line does not change the hang. The next thing that should happen, is to load HelperW.EXE (GENMAC).

If I revert to the old USBEHCD.SYS, I can boot using the new USBOHCD.SYS when using /SMP /PIC. That seems to work, but no real testing was done. If I try to boot with /SMP /APIC, I no longer get a trap000d in USBOHCD.SYS. Now, it gets a trap000d in RESOURCE.SYS. The eCS bugtracker has been updated with the details.

I am also using USBUHCD.SYS with my antique ThinkPad A22e (only one UHCI adapter). That seems to be working good, but so did the older USBUHCD.SYS. That machine does not use ACPI.

DougB

  • Sr. Member
  • ****
  • Posts: 407
    • View Profile
Even more results, using the contents of usbxhcd2.zip.

I tried USBEHCD.SYS and USBUHCD.SYS on my ThinkPad T43 (1871-W8M). They seem to work fine. This machine is using eCS 2.0 silver and  ACPI 3.18 with /SMP /APIC, even though it is only a single processor.

I also tried my old Gigabyte GA-7IXE motherboard, with AMD Athlon64 processor, eCS 2.0 silver, and ACPI 3.18. The two controllers that are on that board never did work, and still don't. But the generic plug in card with 1 EHCI and 2 OHCI controllers, works good.

So far, everything works, except USBEHCD.SYS on the quad core system causes a hang at boot time, and the USBOHCD driver causes a trap if I use /SMP /APIC with ACPI. It works if I use /SMP /PIC. The problems are not in the SMP kernel, all of the systems are using that. The problem is also not in ACPI, although ACPI may not be setting up the I/O on the quad core correctly, but it works with the old drivers, and /SMP /PIC.

DougB

  • Sr. Member
  • ****
  • Posts: 407
    • View Profile
Now, with usbxhcd3.zip.

I am back to having USBOHCD.SYS crash, if I try to use /SMP /APIC, otherwise no changes.

Blonde Guy

  • Sr. Member
  • ****
  • Posts: 283
    • View Profile
    • Blonde Guy
I have one ThinkCentre desktop computer where USB has worked very poorly. With the latest IBM drivers, the machine locks up in most cases when I use a USB keyboard and mouse. Only by using legacy BIOS support can I use a USB keyboard and mouse. No other USB function works, it doesn't detect any other device.

With usbxhcd3 and above, I have different behavior. The USB works for keyboard and mouse without legacy BIOS support. Clearly this is better. Further, the machine simply stops if I plug in any other USB device. While this is useless for any practical use of the machine, it is quite encouraging from a programming standpoint.

If you want give/get good feedback, contact Lars on comp.os.os2.misc.
Expert Consulting for OS/2 and eComStation