http://www.bughost.org/pipermail/power/2008-January/001234.html notes that syndaemon wakes up every 20ms (50 times a second!) to see if it's idle timer (nominally 2.0 seconds) has expired yet. That's a little much.
I'd like to apply the patch noted in the URL above (which opensuse is carrying), plus a patch to make it poll every 200ms, dropping the wakeups to 5/sec.
Objections? Patches inline below.
Thanks, Matt
http://www.bughost.org/pipermail/power/2008-January/001234.html notes that syndaemon wakes up every 20ms (50 times a second!) to see if it's idle timer (nominally 2.0 seconds) has expired yet. That's a little much.
I'd like to apply the patch noted in the URL above (which opensuse is carrying), plus a patch to make it poll every 200ms, dropping the wakeups to 5/sec.
Objections? Patches inline below.
Similar problem in ksynaptic's syndaemon (https://bugzilla.redhat.com/show_bug.cgi?id=252258).
I guess I was waiting for the perfect solution rather than just doing the good in the mean time. Upstream has mostly abandoned ksynaptics as well as become unreachable via email. However, there does seem to be some work there on a new package "TochFreeze" that while not providing all of the functionality of ksynaptics, does appear to be event driven instead of polling. See http://qsynaptics.sourceforge.net/
- Orion
On Mon, Mar 17, 2008 at 10:12:54PM -0600, orion@cora.nwra.com wrote:
http://www.bughost.org/pipermail/power/2008-January/001234.html notes that syndaemon wakes up every 20ms (50 times a second!) to see if it's idle timer (nominally 2.0 seconds) has expired yet. That's a little much.
I'd like to apply the patch noted in the URL above (which opensuse is carrying), plus a patch to make it poll every 200ms, dropping the wakeups to 5/sec.
Objections? Patches inline below.
I've patched and built this now.
Similar problem in ksynaptic's syndaemon (https://bugzilla.redhat.com/show_bug.cgi?id=252258).
I guess I was waiting for the perfect solution rather than just doing the good in the mean time. Upstream has mostly abandoned ksynaptics as well as become unreachable via email. However, there does seem to be some work there on a new package "TochFreeze" that while not providing all of the functionality of ksynaptics, does appear to be event driven instead of polling. See http://qsynaptics.sourceforge.net/
I took a stab at rewriting syndaemon to catch X events and sleep otherwise, but was unable to figure out how to get events when a) the app is not a window, and b) events go to windows. Given that I know nothing about really writing such an app, it was an interesting couple-hour exercise, but ultimately fruitless. Someone who knows more about X could probably recreate my worthless attempt in 5 minutes, and may even know enough to do it right.
On Wed, 2008-03-19 at 15:06 -0500, Matt Domsch wrote:
On Mon, Mar 17, 2008 at 10:12:54PM -0600, orion@cora.nwra.com wrote:
I guess I was waiting for the perfect solution rather than just doing the good in the mean time. Upstream has mostly abandoned ksynaptics as well as become unreachable via email. However, there does seem to be some work there on a new package "TochFreeze" that while not providing all of the functionality of ksynaptics, does appear to be event driven instead of polling. See http://qsynaptics.sourceforge.net/
I took a stab at rewriting syndaemon to catch X events and sleep otherwise, but was unable to figure out how to get events when a) the app is not a window, and b) events go to windows. Given that I know nothing about really writing such an app, it was an interesting couple-hour exercise, but ultimately fruitless. Someone who knows more about X could probably recreate my worthless attempt in 5 minutes, and may even know enough to do it right.
Wow, what a hack. That really seems like something better done in the server itself.
- ajax
-
Adam Jackson -
Jeremy Katz -
Matt Domsch -
Orion Poplawski