PERFORCE change 37154 for review
Nate Lawson
njl at FreeBSD.org
Thu Aug 28 22:16:32 PDT 2003
http://perforce.freebsd.org/chv.cgi?CH=37154
Change 37154 by njl at njl_laptop on 2003/08/28 22:16:21
Add my 6.0 wishlist for ACPI and newbus
Affected files ...
.. //depot/doc/6.0-TODO#4 edit
Differences ...
==== //depot/doc/6.0-TODO#4 (text+ko) ====
@@ -74,5 +74,37 @@
pass. Modules loaded after boot inherently always run at the last
pass.
+# njl Wishlist
+- Device suspend/resume tree sweep. Many drivers do not properly support
+ device_suspend and device_resume. On my laptop alone, uhci(4) and
+ snd_ich (pcm(4)) both fail to work after a suspend. We would maintain
+ a list of drivers and their suspend/resume level of support, similar to
+ the SMP status page. The goal would be to get maintainers to fix and
+ mark their devices as working.
+
+- Fix and/or implement VGA register save/restore so that syscons is not
+ garbled after a resume.
+
+- Add a device_suspend_prep() entry that is called before device_suspend().
+ ACPI and APM could traverse the bus tree and allow each device to decide
+ whether or not to veto the suspend request.
+
+- Add a device_eject() method to allow devices to prepare to eject (i.e.
+ PC card, Hotplug PCI). Could this be provided by device_shutdown()?
+ Call ACPI's _EJx methods for the given device as well.
+
+- Hook newbus so that ACPI Notifies are delivered to the destination
+ device if it wishes to be aware of them.
+
+- Add support for ACPI 2.0 to our implementation. This includes implementing
+ processor power down states (C1, C2, ...), evaluating other namespaces,
+ etc. Linux has partial 2.0 support.
+
+- Add support for vendor-specific ACPI devices (like ASUS and Toshiba
+ laptop buttons). This can be studied in Linux.
+
+- Much ACPI bugfixing. I'm uncomfortable with the current level of locking,
+ for instance.
+
< need to insert phk's wishlist here? >
More information about the p4-releng
mailing list