Hi Dominic, did you found some more info on this? I got similar problems on FreeBSD stable/10.
It just occasionaly crashes with similar memory errors. I want to seek into it more, but would be helpful if you got some results. Robert. On Thu, 05 Sep 2013 09:31:18 +0200 Dominic Fandrey <kamik...@bsdforen.de> wrote: > On 04/09/2013 01:52, Carsten Haitzler (The Rasterman) wrote: > > On Tue, 03 Sep 2013 17:35:03 +0200 Dominic Fandrey > > <kamik...@bsdforen.de> said: > > > >> E17 frequently gets stuck in "PAUSE !", when closing a window. Most > >> frequently, but not exclusively, thunderbird -compose windows. > >> > >> Tail of the .xsession-errors file: > >> ESTART: 0.22645 [0.00001] - E_Shelf Config Update > >> ESTART: 0.22883 [0.00237] - E_Shelf Config Update Done > >> ESTART: 0.22884 [0.00002] - Manage all windows > >> ESTART: 0.23061 [0.00177] - MAIN LOOP AT LAST > >> ESTART: 0.50552 [0.27490] - DELAYED MODULE LOAD: gadman > >> ESTART: 0.50582 [0.00030] - SLEEP > >> EINA: Data at address 0x806c02f74 is invalid. Replacing with zero > >> page. PAUSE ! > > > > that "data address xxx is invalid" line... is a very important > > thing. that's a sigbus. a sigbus is when an app has a perfectly > > valid bit of memory and tries to access it... BUT... it's > > inaccessible. this happens for 2 reasons. 1 - literally out of > > memory. > > Definitely not. The system isn't under any kind of strain, there is > plenty of RAM and SWAP to go around and the system would start > killing processes before something like that happened. > > > the linux kernel overallocates by default - that means > > it hands out memory that it doesn't have, assuming apps dont use it > > all. it can happen where they do and then a valid allocation can > > get a sigbus, but the more common case i know of is... mmaped i/o. > > if the filesystem has corruption ... > > I think at a time when I was frequently trying to reproduce a certain > panic in order to generate a useful kernel dump ... likely cause. > > > now this is actually e_start's problem... it shouldnt consider a > > sigbus a normal crash offense. > > And it actually seems unrelated to the problem. I just had another > round. Only this time I ran startx inside tmux, so I got a complete > record of the session without running into the xdm auth issue when > trying to restart e17. > > There's lots of output by programs like firefox and then there's a > PAUSE ! line followed by a segfault (that one is new) and all the > processes complain and terminate. > ------------------------------------------------------------------------------ Rapidly troubleshoot problems before they affect your business. Most IT organizations don't have a clear picture of how application performance affects their revenue. With AppDynamics, you get 100% visibility into your Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk _______________________________________________ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
