> >> With these changes the P2B-LS might be able to switch to suspended state > >> now; it writes "Saving system state. Please wait..." and spins down the > >> hdd. > >> I can't break into kmdb any more, though. The PS/2 keyboard seems to be > >> powered off, too. But the box is still powered on. Question is: how > >> do we wake up the box from suspended state? > >> > > > > Hmm... the uadmin 3 20 test point probably was intended to have a timer set > > somewhere... Randy is the expert on that code, and he's not online today. > > 'uadmin 3 20' is not one of the test points, but 'uadmin 3 22' is (there > is also '3 24', but that could leave your system in an undesireable state > - that is a "suspend even if a driver fails" entry): > > A_FREEZE commands: > 20 - Suspend to ram (power off the box) > 21 - Check if s2r is supported > 22 - s2r, but don't power off, and resume with a success > 23 - s2r, but fail - usefull for making sure devices resume if > suspend fails. > 24 - s2r, ignoring drivers that return DDI_FAILURE > 25 - suspend a single driver, with a 3rd argument that is the major > number of the driver > > From the "failed to turn off" perspective, we have seen misbehaving USB > devices keep the machine from actually shutting down. Dana Myers knows > something of this failure mode (which is mostly at the acpica/BIOS level), > and *might* be able to provide more information, but one question I have > is what USB drivers are actually loaded?
This ASUS box has an Intel chipset with an USB 1.x UHCI controller, and I also have an ALi USB 2.0 PCI controller card in it, which has an EHCI+OHCI controller. I was running it with both uhci and ohci disabled. I didn't disable the ehci driver. On a different test system - a Toshiba Tecra S1 notebook - uadmin 3 20 is now actually running detach(DDI_SUSPEND) for all loaded Solaris drivers, and powers off the notebook; the power led is flashing yellow, which should be an indication that we're in some S3 suspended-to-ram state (looks the same when suspending Windows XP). But when pressing the power button, the system does a full boot, indead of the expected resume from S3 suspend-to-ram. On the Tecra S1, I needed my "ata" pci-ide DDI_CTLOPS_DETACH fix to prevent the endless loop in pci_save_config_regs, and I had to disable the "uhci", "iprb" and "ipw" drivers. _______________________________________________ driver-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/driver-discuss
