Opened 13 years ago

Closed 13 years ago

#13 closed defect (wontfix)

Just 4 usbuhcd.sys entries

Reported by: hanno Owned by: Lars Erdmann
Priority: minor Component: basedrv
Version: Keywords:
Cc:

Description

Hi,

and thanks a lot for your work on usb drivers!

I have the following (little) problem: My computers (Fujitsu-Siemens Esprimo P5916) have 2 EHCI and 5 UHCI ports installed. So I should have 2 and 5 corresponding entries within config.sys. But doing so, system does not start. I found out, that stating BASEDEV=USBUHCD.SYS only 4 times solves the problem.

Kind regards Hanno

PS: This is true for 10.178 earlier version.

Attachments (3)

pci_hanno.txt (45.3 KB ) - added by hanno 13 years ago.
hanno.zip (212.7 KB ) - added by Lars Erdmann 13 years ago.
updated drivers for hanno, updated tracing
hanno_ftf.zip (3.5 KB ) - added by hanno 13 years ago.

Download all attachments as: .zip

Change History (15)

comment:1 by Lars Erdmann, 13 years ago

Hm, it's seems very odd to have 2 EHCI and 5 UHCI controllers. I suspect that 4 of those 5 UHCI controllers have an associated EHCI controller but the remaining 1 UHCI controller is a standalone controller (whatever problem that might cause). Can you do a "pci.exe -D" and add as an attachment here ?

Last edited 13 years ago by Lars Erdmann (previous) (diff)

by hanno, 13 years ago

Attachment: pci_hanno.txt added

in reply to:  1 comment:2 by hanno, 13 years ago

Replying to erdmann:

Hm, it's seems very odd to have 2 EHCI and 5 UHCI controllers. I suspect that 4 of those 5 UHCI controllers have an associated EHCI controller but the remaining 1 UHCI controller is a standalone controller (whatever problem that might cause). Can you do a "pci.exe -D" and add as an attachment here ?

Hi, and thanks for your answer. Here is the pci output - hope it helps :-)

comment:3 by Lars Erdmann, 13 years ago

Was this pci output taken with my drivers in place ?

comment:4 by Lars Erdmann, 13 years ago

Please try out the updated drivers. Find attached. You might also want to try with 5 USBUHCD.SYS driver instances. Maybe something got fixed in the meantime ...

1.) take a trace for major trace code 224 (USBUHCD.SYS) only. For how to set up trace, see readme.txt from usbhcd178.zip. Save trace as FORMATTED and attach here.

2.) Then take a trace for major trace code 226 (USBEHCD.SYS) only. For how to set up trace, see readme.txt from usbhcd178.zip. Save trace as FORMATTED and attach here.

3.) Would you also be willing to test suspend/resume ? If yes I would also provide "beta version" of updated USBD.SYS.

4.) in case of trap screen, take picture of trap screen and attach here.

Last edited 13 years ago by Lars Erdmann (previous) (diff)

comment:5 by Lars Erdmann, 13 years ago

Owner: changed from somebody to Lars Erdmann
Status: newaccepted

comment:6 by Lars Erdmann, 13 years ago

Note: UHCI Controller #3 (BDF:0:29.2) is operated without associated driver -> PCILEG Sup reg in PCI config space offset 0xC0 = 0x0010 (USBSMIEN: SMI enabled on USB IRQ)

by Lars Erdmann, 13 years ago

Attachment: hanno.zip added

updated drivers for hanno, updated tracing

comment:7 by Lars Erdmann, 13 years ago

If you also want to give suspend/resume a try, also try out the contained usbd.sys.

in reply to:  7 comment:8 by hanno, 13 years ago

Replying to erdmann:

to comment 3:

Was this pci output taken with my drivers in place?

Yes

Thanks for your drivers! Same result for both versions "hanno.zip": System starts and works with 4 UBSUHCD entries but hangs with 5 entries.

I have really spent a lot of time (with x-restarts and check-disks ...) to take a trace/dump but without success. System freezes when hitting CTRL-ALT-NumLock-NumLock (There is absolutely no action when pressing it on systemstart (when system is stopped) and there are only two lines with info about needed disk-space on a black screen when hitting it on a running system.). I believe, the problem is in conjunction with the dump-partition?!?

Perhaps it is of interest, that I use PS/2 keyboard and mouse.

As system basically runs, I would suggest to stop further testing to save your and my time ...? But if my tests will help, I can do it of course.

Kind regards Hanno

comment:9 by Lars Erdmann, 13 years ago

I would appreciate if you could do the requested traces for major code 224 and 226 (separately) with 4 USBUHCD.SYS drivers in place (so that it won't hang).
There is some chance that the traces will reveal if the HCs are properly "pulled away" from the BIOS'es grip (legacy USB).
You can also try and disable "legacy USB" in your BIOS and check if you can then use 5 USBUHCD.SYS drivers.
What ACPI.PSD version are you using ? Make sure you use the latest one (3.19.14 at least).

by hanno, 13 years ago

Attachment: hanno_ftf.zip added

comment:10 by hanno, 13 years ago

Hi,

I have uploaded the 2 trace-files (took a long time as I never did a trace and I had to learn about these things ;-) ).

Make sure you use the latest one (3.19.14 at least)

Yes, I am using 3.19.14

You can also try and disable "legacy USB" in your BIOS and check ...

BIOS does not give me the possibility to change this issue. All I can do is choose between:

  • USB1.1 or USB1.1+USB2.0
  • Enable all USB-ports or just Keyboard+Mouse, No Storage devices and Hubs, ....
  • Some other possibilities I do not remember at the moment, if you are interested in, I can post a complete list.

comment:11 by Lars Erdmann, 13 years ago

Would you please take a trace for major code 224 and 226 with the latest usbhcd183.zip ? Please make sure that you also properly update the .TFF files so that the formatting of the trace data will be OK !

comment:12 by Lars Erdmann, 13 years ago

Resolution: wontfix
Status: acceptedclosed
Note: See TracTickets for help on using tickets.