usbhcd188 on ancient hardware

[dbongo]

In the readme you said "I'd be especially interested in reports on older systems (also without ACPI.PSD) and on systems using ACPI.PSD."

Well, I installed usbhcd188 on my old Pentium (socket 7, vintage 1995-ish, with an underclocked Pentium MMX 200 running at 166MHz with an el cheapo $5 PCI USB card) and I popped in a USB flash drive (ironically, a 32GB flash drive...on a system that only supports 8GB hard drives) and it read just fine. (I didn't try writing to the drive because I didn't need to. Testing the driver didn't even cross my mind.)

If you want me to provide any more specific information or run any tests let me know. Or mark this closed and be content in the knowledge that the driver works are hardware that's too old to justifiably need it.

[Lars Erdmann]

If it's not too much work: I'd appreciate a PCI dump of the system (pci.exe) and also a trace for the relevant drivers (USBUHCD.SYS,USBOHCD.SYS,USBEHCD.SYS, whatever is present in the system). The readme.txt says how to do it.

What I am after is if my CPU frequency calculation also properly works on older systems. I am using the CPU frequency in timing loops using the Pentium TSC (time stamp counter) so as to avoid using the unreliable/too coarse timing facilities that the OS provides otherwise.

[dbongo]

PCI dump will be attached immediately after sending this. I don't have time right now to do the trace - probably tomorrow. But since I do have your proverbial ear, the USB card shows (as I recall) 2 OHCI and 1 EHCI so would TRACE=ON 225,226 be correct, or would I use TRACE=ON 225,225,226 to cover the 2 OHCI controllers? (I have no problem doing this for you, but I want to do it right the first time so I only have to do it one time.)

[dbongo]

PCI dump

[Lars Erdmann]

All OHCI driver instances will use same trace code 225, therefore:

TRACE=ON 225,226 is correct (in order to trace all OHCI and EHCI controller instances)

[dbongo]

PCI Dump with -D switch

[dbongo]

Trace file

[dbongo]

The correct PCI dump (I forgot the -D switch when I did the first one) and the trace file are attached. If you need anything else, please ask. (Your updated drivers helped out immensely on my primary PC, so I'm very happy to help.)

And, just a thought, I did mention that I have the CPU in this dinosaur underclocked. Would it be helpful to set it correctly and run another trace? I'm going to run it at its native frequency, I just haven't gotten around to fixing the jumpers. If it would help, just ask and I'll move on it. (This is a back-burner project that I only turn on in my spare time.)

[Lars Erdmann]

No you don't have to do anything else. The drivers correctly compute your CPU frequency to be executing at 166 MHz which meets the expectation. Thanks for your help. I am now closing this bug.