irritating USB EHCI riddle

[BuXb]

So here's what has been going on on my laptop:

 

With 8.4.1 kernel sys the computer shuts down and restarts fine. With 8.8.1 kernel sys it only does so when 1 of the two cores is deactivated in BIOS. As soon as the 2nd core is added, OS x86 hangs at "CPU halted" (verbose mode) for sometimes 20 seconds. Sometimes I even have to push the powr button for a second to kick the shutdown or restart process through, sometimes I have to hard-reset. That happens with ALL the 10.4.8 8.4.1 releases from JaS & friends.

 

A couple of days ago I updated my laptop's BIOS to 3.3, and look-a-there: all of a sudde OS x86 would shutdown and restart fine !! BUT - access to voulmes via USB was gone (USB stick, USB HD..). I recognized that, at verbose boot-up the usual message "USB: EHCI caused wake event", was gone, too. Then I remembered orByte's Generic USB2.0 (EHCI) Driver and applied it. At verbose boot up there were then 4 lines about the actions of the kext (addressing ranges, forcing offsets and whatever), and again (look-a-there): USB mass storage access was reestablised! BUT: the RSB (restart- and shutdown bug) was there again ..

 

Disabled one CPU core, instead four lines @ verbose bootup only one saying "USB: EHCI caused wake event" as with Apple's original AppleUSBEHCI.kext, and also the RSB is gone again like expected. Reverted to Apple's AppleUSBEHCI.kext + reenabled 2nd core --> RSB gone but again no USB storage access, both as expected.

 

Reverted back to BIOS 2.4 --> same thing. Applied the latest 10.4.8 USB update from Apple --> no change. Tried both, Apple's and orByte's 10.4.7 USB packages instead but neither is compatible due to kernel dependencies incl. a 10.4.7 kernel.framework (Apple) --> no boot.

---------------------------------------------------------------------------------------------

 

So, obvioulsy the EHCI portion of the 10.4.8 IOUSBFamily.kext (v.2.6.1), no matter if it contains the AppleUSBEHCI.kext or orByte's generic PCGenUSBEHCI.kext, causes an annoying shutdown and restart delay or even total hang on this Toshiba Laptop AS SOON AS BOTH CPU CORES ARE ENABLED!! ever since the 8.8.1 kernel/ kernel framework. What's the logic behind that? Why can "duo coring" influence USB behaviour at all ???

---------------------------------------------------------------------------------------------

 

From my perspective there is two possible (=realistic) approaches to solve this: recoding/ improving/ adapting..

 

1) .. orByte's generic PCGenUSBEHCI.kext

 

2) .. semthex kernel

 

[ 3) rather not: the BIOS - brrrrr ]

 

or maybe both.

 

Does anybody have any helpful input on this?

 

Cheers,

Bugs

Edited February 19, 2007 by BugsBunny

[orByte]

Hi, probably related to clock management without HPET. Also the USB driver registers without clock, so it is possible that with two cores time no longer works as a monotonic increasing function (this was one of the arguments for HPET with dual core...) causing some timeout condition on driver unload. Of course, this is just a guess. Anybody else experience this problem with dual cores?

 

 

Best

[Funky frank]

I have this problem too. Please fix it.