IRC log of #maemo-ssu for Wednesday, 2012-05-23

*** mirandir has joined #maemo-ssu00:02
*** mirandir has left #maemo-ssu00:07
*** mirandir has joined #maemo-ssu00:07
*** NIN101 has quit IRC00:12
*** mirandir has left #maemo-ssu00:14
*** mirandir has joined #maemo-ssu00:14
*** mirandir has left #maemo-ssu00:16
*** mirandir has joined #maemo-ssu00:18
*** mase76 has quit IRC00:25
*** FireFly has quit IRC00:37
*** FireFly has joined #maemo-ssu00:37
*** trbs has quit IRC00:38
*** mirandir has left #maemo-ssu00:49
*** smoku has left #maemo-ssu01:17
*** M4rtinK has quit IRC01:34
*** Atarii has quit IRC01:52
*** ZogG has quit IRC02:10
*** ZogG has joined #maemo-ssu02:23
*** Sc0rpius has quit IRC03:51
*** arcean has quit IRC04:04
*** timoph_ has joined #maemo-ssu04:11
*** krayon has joined #maemo-ssu04:18
*** timoph has quit IRC04:19
*** guly has quit IRC04:19
*** guly has joined #maemo-ssu04:21
*** nox- has quit IRC04:23
*** guly has quit IRC04:25
*** guly has joined #maemo-ssu04:30
*** ZogG has quit IRC05:25
*** ZogG has joined #maemo-ssu05:39
*** Sc0rpius has joined #maemo-ssu05:47
*** guly has quit IRC05:49
*** amiconn has quit IRC05:55
*** guly has joined #maemo-ssu05:55
*** amiconn_ has joined #maemo-ssu05:55
*** amiconn_ is now known as amiconn05:55
*** Sc0rpius has quit IRC06:52
*** Sc0rpius has joined #maemo-ssu07:28
*** wmarone has quit IRC07:51
*** wmarone has joined #maemo-ssu07:51
*** LaoLang_cool has joined #maemo-ssu08:32
*** mirandir has joined #maemo-ssu08:45
*** Jade has joined #maemo-ssu08:48
*** timoph_ is now known as timoph09:04
*** M4rtinK has joined #maemo-ssu09:25
*** M4rtinK has quit IRC09:41
*** LaoLang_cool has quit IRC09:55
*** dafox has joined #maemo-ssu10:37
*** arcean has joined #maemo-ssu11:38
*** M4rtinK has joined #maemo-ssu12:49
*** LaoLang_cool has joined #maemo-ssu12:50
*** LaoLang_cool has quit IRC13:14
*** BCMM has joined #maemo-ssu13:14
*** andre__ has joined #maemo-ssu13:42
*** BCMM_ has joined #maemo-ssu13:56
*** BCMM has quit IRC13:57
*** BCMM_ is now known as BCMM13:57
*** BCMM_ has joined #maemo-ssu14:02
*** BCMM has quit IRC14:03
*** BCMM_ is now known as BCMM14:05
*** mirandir has left #maemo-ssu14:07
*** mirandir has joined #maemo-ssu14:08
*** DocScrutinizer has quit IRC14:42
*** DocScrutinizer has joined #maemo-ssu14:42
*** ZogG has quit IRC15:19
*** Pali has joined #maemo-ssu15:30
*** ZogG has joined #maemo-ssu15:33
*** trbs has joined #maemo-ssu16:11
merlin1991freemangordon: ping16:14
*** Pali has quit IRC16:22
*** Pali has joined #maemo-ssu16:29
merlin1991dang now I need him :/16:35
*** arcean has quit IRC16:40
*** BCMM has quit IRC16:41
*** mase76 has joined #maemo-ssu16:45
freemangordonmerlin1991, pong16:59
merlin1991dosfstools had to be built without -O2 ?17:00
merlin1991because you never changed debian/rules17:01
merlin1991freemangordon: ^^17:01
freemangordonmerlin1991, most probably there won' be a problem to add -O217:02
freemangordon*won't17:02
merlin1991freemangordon:  well it's there17:02
merlin1991but I recall you raging about not again like Qt17:02
freemangordonI don't understand then17:02
freemangordonaah17:02
merlin1991and saying that it started to work after you killed -0217:02
merlin1991so with or without -O2 now? :D17:03
freemangordonmerlin1991, no, it was my fault not applying pathches in debian/patches.17:03
freemangordon-O2 is ok17:03
merlin1991ah okay17:03
freemangordongive me a second17:03
freemangordonmerlin1991, debian/rules were missing "include /usr/share/quilt/quilt.make" so the patch in debian/patches have not been applied17:05
freemangordonand it was that patch missing that were resulting in segfault17:06
freemangordonhere is the patch that glues all the pieces together https://gitorious.org/community-ssu/dosfstools/commit/7c10cc567693d31ce14f2738dc0b55ca4f68a06717:06
freemangordonmerlin1991 ^^^17:06
merlin1991perfect :)17:07
*** BCMM has joined #maemo-ssu17:07
freemangordon:)17:07
*** BCMM has quit IRC17:15
*** chainsawbike has quit IRC17:37
*** chainsawbike has joined #maemo-ssu17:39
*** mase76 has quit IRC17:40
Estel_merlin1991, freemangordon, do You know that e2fs is Maemo is broken, and there is way to repair it that could be included in CSSU18:00
*** M4rtinK has quit IRC18:00
Estel_it means, that small filesystem coruptions (like repairable by recovering from journal) works, but bigger repairs result in e2fs segfault18:00
Estel_I've written about it some time ago18:00
Estel_latest version from Debian works well18:00
Estel_and doesn't seem to be causing any regressions18:00
freemangordonEstel_, no, I was not aware of that. Will see if latest debian version could be compiled under SB and will put it in cssu-devel18:09
Estel_thansk a lot. I got something documented "somewhere18:09
Estel_"" will tyr to find that and send You18:09
freemangordonok18:09
Estel_s/tyr/try/18:10
infobotEstel_ meant: "" will try to find that and send You18:10
Estel_freemangordon, found it, You preffer Pm on TMO or zerobin?18:12
Estel_freemangordon, BTw, I forget to mention it - 720p video recording stopped working properly for me :(18:15
Estel_it all stutters, and strange artifacts show on screen18:15
Estel_on viewfinde,r but they're also included in video18:16
Estel_it look like on old TV, when FPs were not synced18:16
Estel_I have such video, if You're interested18:18
Estel_would like to help in finding why. It's reproduceable18:18
freemangordonEstel_, that is strange, could be relatet to too much zoom18:47
Estel_0 zoom18:47
freemangordon*related18:47
freemangordonall resolutions?18:48
Estel_I never use it as it's software zoom, for obvious reasons18:48
Estel_lemme check on non-hd ones...18:48
Estel_I think it started after CSSU *or* kp50 upgrade, can't tell for sure now :/18:48
freemangordonhmm, could be your DSP is too undervolted18:48
freemangordonthere was a change in KP50 re SR voltage calculation18:49
Estel_maybe, because now I can't reproduce it, WTF... Tried many times in ZOO today, when I was filming my sonb18:50
Estel_s/sonb/son/18:50
infobotEstel_ meant: maybe, because now I can't reproduce it, WTF... Tried many times in ZOO today, when I was filming my son18:50
Estel_but honestly, it looked rather like stoping tracker or something like that failed18:51
Estel_will try on fresh reboot18:51
Estel_maybe it will re-appear18:51
freemangordonEstel_, BTW where that "swap fragmentation" thingie comes from?18:53
Estel_You mean ereswap, or problem itself?18:53
freemangordonthe problem18:53
Estel_look. swap, initially, is save din purely sequential manner18:54
Estel_(I've explained it in ereswap readme and thread on TMO :P)18:54
Estel_so, if Yuo save 100 MB18:54
Estel_then free 20 in middle18:54
freemangordonI mean: there is an algorithm in kernel to combine non-sequental sectors to sequental writes18:54
Estel_yea, but it's not that18:54
Estel_damn, shadowjk would explain it better than me but18:54
freemangordonwell, if it is explained I will read it18:55
Estel_as You know things saved in emmc or mmc does have physical write leveling18:55
Estel_ok18:55
Estel_but it's  explained in "human language"18:55
Estel_generally, whole thing is that swaps doesn't care for (logical mapping) freed bits18:55
Estel_it just keep saving sequentialy18:55
Estel_but, when ti reaches end of partition18:56
Estel_it start re-using previously freed bits18:56
Estel_they're, almsot always, not matching size of what You need to write actually18:56
Estel_so swpa performance decrease dramaticaly18:56
Estel_swap*18:56
Estel_and it doesn't matter than our emmc and microSD flash have wear leveling18:57
freemangordonEstel_, I see, but in fact kernel driver tries to find first big enough free space18:57
freemangordonand it should succeed18:57
Estel_Hm, haven't know about that, but I must tell You that, somehow, it's still slower18:57
freemangordonmost of the time18:57
Estel_maybe some bug?18:57
Estel_for sure it's not placebo18:57
Estel_I feel it really, even if I don't know about writes on swap18:57
freemangordoncould be a bug, maybe I should put some tracing to checkwhat is going on18:57
Estel_sometimes, when something trash my swap without me knowing, I tell myself "wtf, why i started to clutter"18:57
Estel_then I check it18:57
Estel_anbd see that I'm, lets say, 20 MB after reaching swap "limit"18:58
Estel_quite possibloe18:58
Estel_possible*18:58
freemangordonEstel_, that deffinitely sounds like a bug18:58
Estel_maybe this kernel-searching is resource hungry?18:58
Estel_You know, even UI is slower18:58
Estel_I always wondered why the hell it seems slower even with no memory huingry things18:58
Estel_OTOH i don't see much CPU usage18:58
Estel_maybe time penalties for searching? or it's really a bug18:59
freemangordonEstel_, don't sound like resource hunger, as if it that was the case, then your limit wouldn't be exactly the partition size18:59
freemangordonBTW better switch to #maemo18:59
Estel_true18:59
*** NIN101 has joined #maemo-ssu19:04
ShadowJKThe nokia modified swap algorithm is: Find largest free block in swap, start writing to that block until end of block, repeat find largest free blck. Over time, as stuff is allocated and freed, the largest free block becomes small, and unaligned. For optimal speed we want somethning like 1Mbyte alignment.19:07
Estel_ShadowJK, thanks a lot19:10
Estel_could You switch to #maemo19:10
*** mirandir has left #maemo-ssu19:25
*** M4rtinK has joined #maemo-ssu19:35
*** Atarii has joined #maemo-ssu19:51
*** Atarii has joined #maemo-ssu19:51
*** dafox has quit IRC20:04
*** M4rtinK has quit IRC20:27
*** M4rtinK has joined #maemo-ssu20:33
*** M4rtinK has quit IRC20:44
*** M4rtinK has joined #maemo-ssu21:00
*** nox- has joined #maemo-ssu22:32
*** M4rtinK has quit IRC22:57
*** NIN101 has quit IRC23:18
*** ekze-nyan has quit IRC23:28
*** ekze-nyan has joined #maemo-ssu23:29
*** smoku has joined #maemo-ssu23:29
*** M4rtinK has joined #maemo-ssu23:31
*** ekze has joined #maemo-ssu23:47
*** ekze-nyan has quit IRC23:48
*** ekze has quit IRC23:51
*** mirandir has joined #maemo-ssu23:56

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!