On Thu, 2007-01-25 at 20:51 +0100, Pavel Machek wrote: > Hi! > > > > > > > This patch set is against acpi-test sysfs branch which already converts > > > > > > ACPI to follow driver model. Now the ACPI procfs functions are > > > > > > duplicated in sysfs step by step. And I want the ACPI interface in sysfs > > > > > > works exactly the same way as in procfs. > > > > > > > > > > No, you can't port same broken interface into sysfs. Some /proc > > > > > interfaces are horrible, and we do not want to create _exactly same_ > > > > > horrible interfaces in /sys. > > > > > > > > Please be specific. > > > > > > /proc/acpi/alarm is horrible mess, as was detailed in another email > > > > ditto /proc/acpi/wakeup, and all those empty /proc/acpi directories > > Agreed, /proc/acpi/wakeup is unusable/strange/mess, too. I don't get > that many empty /proc/acpi directories, but I surely hope I'll not see > them in /sys. Many /proc/acpi/(drivers)/ may be empty, e.g. /proc/acpi/ac, /proc/acpi/battery... This is because that driver entry in procfs is created after driver is successfully registered. But the actual proc interface is added when a device is binded to the driver. So when a driver is loaded but no matchable device is found, we'll leave an empty directory under /proc/acpi/. Anyway, we won't have such problems in sysfs. Thanks, Rui
