Douglas Atique wrote: >There are also messages that warn about "shared interrupts" and "USB OHCI >controller unusable". Are these also manifestations of the same problem? > > > The warning about "shared interrutps" is harmless. That message should have been removed. Not sure about the OHCI issue since the header files may be different.
>About using Nevada, that's out of question for this machine. I have to stick >to the supported Solaris 10 with its online updates. So... can anyone on this >list tell when S10 Update 3 will be released and whether the cardbus driver is >expected for that update? > > > S10u3 may be available in Nov (Not quite sure). But cardbus will not be in u3 (Sure). Please understand that this is not trivial task, when we try to put something like this in an update release. There're related bugs in PCI and boot, and a large set of machines need to be tested before we can make a commitment to our customers. >As an aside, I don't understand one thing: when I just slip the card into the >slot, /var/adm/messages issues nothing and no system activity seems to take >place except for a few HD reads. The real thing starts when I do a reconfigure >reboot with the card inserted. Shouldn't the system detect that the card was >inserted automatically? I mean, not even prtconf seems to notice the card was >inserted unless I do the reconfigure reboot. What is happening to the hotplug >behaviour in this case? (sorry for the newbieness...) > > The insertion interrupt is not triggered in this case. If this happens in Nevada with the native cardbus driver, then we can be sure that there's a bug. The problem can be in cardbus, pci, or even the ACPI tables in BIOS. We've seen in the past that alot of reasons can cause this. We cannot claim that we support all x86 boxes, only the set of machines that we or someone have tested. And you can help enlarge that set. Vincent.
