[16:33] <Daviey> Has anyone reported losing right click ability on synaptic touchpads?
[16:34] <bryceh> Daviey, dunno but cnd's been working on that chunk of code lately so he'd be the one to chat up
[16:34] <bryceh> Daviey, upgrade or downgrade your -synaptics until you find a version that works.
[16:36] <Sarvatt> Daviey: sorry was responding in #ubuntu-desktop
[16:36] <Sarvatt> Daviey: easiest way is to disable the synaptics clickpad xinput option
[16:36] <Daviey> Sarvatt: yes, sorry for x-posting
[16:36] <Daviey> thanks
[16:37] <Sarvatt> xinput --list, find the id for the bcm5974 device, xinput --list-props <id>, find the number for the clickpad property then xinput --set-prop <id> <clickpad id>
[16:37] <Sarvatt> hmm there was a better bug cnd responded on, trying to find it
[16:40] <Daviey> Sarvatt: sorry is the proprty the "Synaptic clickPad (259): 1) .. is it 1 or 259?
[16:40] <Sarvatt> 259
[16:41] <Daviey> Sarvatt: perfect, thanks!
[16:41] <Daviey> That worked
[16:45] <Sarvatt> guess its easier to link people to http://paste.ubuntu.com/859398/ this is going to come up a lot :P
[16:45] <Sarvatt> not being able to right click at all is nasty
[18:03] <bryceh> hmm
[21:46] <sforshee> There have been multiple reports of desktop freezes on systems with sandy bridge (it doesn't seem to be related to RC6) which I am also seeing (bug 938770). For me they seem to go away if I downgrade mesa to 7.11-0ubuntu4. Can anyone suggest next steps for debugging?
[21:48] <sforshee> one user also reported that going back to a 3.0 kernel gets rid of the issue, but I can't confirm this as my graphics aren't working at all with any pre-3.2 kernel
[21:49] <Sarvatt> sforshee: check your ~/.xsession-errors to see what desktop component failing and killing your session :)
[21:49] <Sarvatt> sforshee: a lot of the dupes you're referring to i believe were the ones fixed in the 8.0.1 mesa upload so be careful looking for dupes until you find out whats actually going wrong
[21:50] <Sarvatt> aka you'll see a lot of noise in the bug reports
[21:50] <sforshee> Sarvatt, I'm not searching for dupes, these are from people reporting problems on the RC6 testing bug 
[21:51] <Sarvatt> since when?
[21:51] <Sarvatt> link?
[21:51] <sforshee> let me find it...
[21:51] <Sarvatt> i've been following the wiki closely with leann and cking and the intel guys since i'm the one that did the sauce patch to reenable it
[21:51] <Sarvatt> haven't see anything
[21:53] <Sarvatt> the rc6 freezes aren't an issue anymore with 3.2.0-17.27, with -17.26 and earlier kernels it was still enabling rc6p and there were some hard system power offs because the bioses put the voltages too the more aggressive rc6 states
[21:53] <sforshee> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/818830/comments/70
[21:53] <Sarvatt> err bioses put the voltages to low in the more aggressive rc6 states i meant
[21:53] <sforshee> that user still sees hangs with -17.27
[21:53] <sforshee> and I do too :)
[21:54] <Sarvatt> libre office intel antialiasing text crash..
[21:54] <sforshee> I see it with chromium
[21:54] <sforshee> it doesn't matter whether rc6 is enabled
[21:54] <sforshee> I went back to older precise kernels too from when I wasn't seeing this issue, and it still happens
[21:54] <sforshee> with rc6 disabled
[21:55] <sforshee> any other reports I see haven't confirmed it with -17.27 though
[21:56] <Sarvatt> compiz stuck polling in your backtrace, you should have something in ~/.xsession-errors saying what crashed and made your desktop unresponsive, especially if restarting compiz fixes it
[21:56]  * Sarvatt is getting confused why you're mentioning rc6 with this userspace bug :)
[21:57] <Sarvatt> you'll probably see some kind of assert killing things in there
[21:57] <sforshee> Sarvatt, I figured when I said sandy bridge the first assumption would be that it was rc6 related :)
[21:57] <Sarvatt> if ya can grab the log over ssh whenever it happens next
[21:58] <sforshee> Sarvatt, http://pastebin.ubuntu.com/859778/
[21:58] <sforshee> don't need ssh, my vts still work
[21:58] <Sarvatt> its stuck right now?
[21:58] <sforshee> that's .xsession-errors when it's hung
[21:58] <sforshee> no, but I copied xsessin-errors while it was hung, and that's the contents
[21:59] <Sarvatt> ** CRITICAL **: syncdaemon_status_info_get_online: assertion `SYNCDAEMON_IS_STATUS_INFO (sinfo)' failed
[21:59] <Sarvatt> hmm
[21:59] <sforshee> i can readily reproduce it, so if there's something else you'd like me to try let me know
[21:59] <Sarvatt> thats part of gnome-settings-daemon
[22:00] <Sarvatt> sforshee: do you have 2 instances of syndaemon running?
[22:00] <sforshee> nope
[22:02] <Sarvatt> sforshee: and no segfaults in dmesg?
[22:02] <sforshee> no, no segfaults, no gpu hangs, nothing in i915_error_state
[22:03] <sforshee> nothing interesting at all in dmesg when this happens
[22:04] <RAOF> If compiz is polling it probably means it's waiting for a swap to complete.
[22:04] <sforshee> I might have to take that back -- there's something interesting this time, but this is the first time I've seen it
[22:04] <Sarvatt> sforshee: by any chance, does apport-collect 938770 add any more info?
[22:04] <sforshee> init: hybrid-gfx main process (2809) terminated with status 127
[22:04] <Sarvatt> i reassigned to xorg so we can get some logs on the bug
[22:05] <RAOF> sforshee: Yeah, that's entirely benign; it's tselliot's “make hybrid nvidia systems less awful” upstart script running.
[22:07] <broder> RAOF: what source package did that script end up in? (curious to skim it)
[22:07] <Sarvatt> sforshee: I see exactly what you're describing when I have clickpad support enabled and scroll with 2 finger scrolling for instance with xserver 1.12 and it's a synaptics/utouch issue, restarting unity from a VT makes it work again and its definitely not video related even though it looks it so that would be worth trying
[22:07] <RAOF> broder: nvidia-common.
[22:08] <sforshee> Sarvatt, I was thinking it might be clickpad related earlier
[22:08] <Sarvatt> does unity --replace from a VT fix the desktop?
[22:08] <sforshee> I couldn't reproduce on another system with radeon graphics and the same clickpad however
[22:08] <Sarvatt> (note I have to unity --replace twice for it to actually work)
[22:09] <sforshee> unity --replace does work, and I only had to do it once
[22:11] <Sarvatt> when it happens to me, I get some warnings from utouch-frame in unity's stderr in the VT that dont show up in ~/.xsession-errors ever, WARNING: failed to get previous touch value
[22:12] <Sarvatt> wonder if its the same issue, i should get off xorg-edgers and go back to stock precise to see if that newer synaptics is still busted there on this machine so i can file a real bug
[22:13] <sforshee> Sarvatt, I don't see those
[22:13] <sforshee> I get some 'invalid cast' warnings
[22:13] <sforshee> and some icons it can't find, that's about it
[22:18] <sforshee> Sarvatt, apport-collect has finished
[22:19] <Sarvatt> sforshee: heh same exactly laptop as me
[22:20] <sforshee> that's convenient :)
[22:20] <Sarvatt> so there might be merit in it indeed being a synaptics/utouch issue i'm hitting on xserver 1.12, i'll downgrade to 1.11 in precise and get more info on this
[22:20] <Sarvatt> there should be no reason i can only hit it in 1.12 since we have the same commits for input in both
[22:20] <sforshee> I'm doing enablement on the machine and also on a macbook pro with the same touchpad but radeon graphics, and I can't reproduce it there
[22:22] <Sarvatt> i'm going to go ahead and move this over to x-x-i-synaptics
[22:22] <Sarvatt> xserver-xorg-input-synaptics              1.5.99+git20111212.9f9b55ab is fine, thats what i've been using for months to work around the issue
[22:23] <Sarvatt> whats apple-bl-gmux btw?
[22:24] <Sarvatt> ah
[22:24] <sforshee> it's for some macbooks that need to control the backlight through the gmux (display mux device)
[22:25] <sforshee> not needed for the air
[22:28] <sforshee> Sarvatt, bug 940771 is a similar freeze with libreoffice, but he isn't using a touchpad. You think that one is different?
[22:30] <Sarvatt> so libreoffice has other issues historically with antialiased text (which is turned on by default) on intel and those are actual crashes, i wouldn't necessarily assume its the same even if its while scrolling too but could be wrong
[22:30] <sforshee> okay, thanks.
[22:31] <Sarvatt> but yeah that is interesting
[22:31] <Sarvatt> i dont *think* i had a screwed up actually enabling rc6p instead of rc6 in the place he linked
[22:32] <Sarvatt> that was like 4 am saturday night though so all bets are off if i actually did that in a daze
[22:33] <Sarvatt> the crashes from libreoffice were memory exhaustion related so it could be he didnt test the other one enough too
[22:35] <Sarvatt> and they definitely show in dmesg which that person didn't attach
[22:35] <sforshee> I think he was running an unofficial rc6 test kernel, so apport refused to collect the logs