On Saturday 12 April 2008, Pavel Machek wrote: > + WARN_ON(!irqs_disabled()); > ret = drv->suspend(sysdev, state); Wouldn't it be better to assert the warning AFTER the driver had a chance to screw it up? (On the resume side of things too.) We know IRQs were disabled on entry to the loop. If something went wrong it was some goofy driver. Best warn about it ASAP instead of *maybe* getting a warning before the next driver (if there is one). And if you're concerned about such issues, I'd think similar warnings would be done in suspend_late() and resume_early() support... - Dave _______________________________________________ linux-pm mailing list linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/linux-pm
