/srv/irclogs.linaro.org/2011/09/27/#linaro-meeting.txt

=== plars is now known as plars-zzz
dzinabout to start the infrastructure meeting11:59
dzinwho is here?11:59
dzinhi12:00
deepti1HEllo12:00
salgadohi there12:00
dzinhi deepti1 , salgado12:01
deepti1salgado: how are u doing now ?12:01
salgadomuch better, deepti1.  thanks12:01
dzinjamestunnicliff_, pfalcon12:01
jamestunnicliff_hi12:02
dzinhas anyone seen pfalcon?12:02
dzinmabac is at an appointment12:02
salgadodzin, I think he's off today12:02
dzinok, let's start then.12:02
salgadodzin, getting his visa12:02
dzinplease bear with me, it's the first time I've done this.... :-)12:02
dzin#startmeeting12:02
linarobotMeeting started Tue Sep 27 12:02:58 2011 UTC. The chair is dzin.12:02
linarobotUseful Commands: #action #agreed #help #info #idea #link #topic.12:02
dzin#chair fabo12:03
linarobotCurrent chairs: dzin fabo12:03
dzin#link https://wiki.linaro.org/Platform/Infrastructure/Meetings/2011-09-2712:04
dzin#topic actions from last meeting12:04
dzinmabac and james_w to discuss setting up staging instance of the workitems tracker - TODO12:04
dzinjames_w to ask again for bot access to the roadmap kanban12:05
dzinboth James and Mabac are not here, they will update when they can12:05
dzin#info dzin to aid in getting offspring changes reviewed upstream - INPROGRESS12:06
dzinI've chatted with Steve Magoun and he should be ready to review for Oct 1.12:07
salgadodzin, mabac sent by email the status of the two first action items. want me to paste here?12:07
dzinsure12:07
fabo#info mabac and james_w to discuss setting up staging instance of the workitems tracker - TODO12:08
salgadomabac and james_w to discuss setting up staging instance of the12:08
salgadoworkitems traker12:08
salgadoWe have not done that.12:08
salgado* james_w to ask again for bot access to the roadmap kanban12:08
salgadoHe did and kiko gave me a login. I have no permissions though and have12:08
salgadobugged kiko about that.12:08
salgado* mabac to release svammel on Thursday12:08
salgadoDone12:08
salgado* mabac to work with release team to decide when to release linaro-image-tools12:08
salgadoI think that went well. l-i-t is released and there's a fix for an12:09
salgadoissue that's already been found.12:09
dzinthanks12:09
dzin#info mabac to release svammel on Thursday - DONE12:09
fabo#info james_w to ask again for bot access to the roadmap kanban - DONE12:09
dzin#info mabac to work with release team to decide when to release linaro-image-tools - DONE12:10
dzin#action mabac and james_w to discuss setting up staging instance of the workitems tracker12:11
dzin#info everyone to set up a call with their buddy to talk about work for this month12:11
dzinhow did the buddy calls go?12:11
deepti1dzin: I have been busy with the release work so I have not done any progress on that this week12:12
dzinok, thanks. No one else?12:12
salgadomyself and jamestunnicliff_ are working closely together on private hwpacks, so we chat quite frequently12:13
dzinthanks12:13
salgadowith pfalcon (my other buddy) we're still discussing how best to go about it, because he thinks weekly calls would take too much of his time12:13
dzinsalgado: is that for every week or just around release time?12:14
salgadoevery week12:16
dzinis everyone finding the buddy calls useful or a time hog?12:17
jamestunnicliff_I find them useful because salgado and me just call each other when we have something to discuss.12:19
salgadoI haven't experienced it with a buddy that's working on a different project than myself (which is the scenario it was devised to), so can't really tell12:19
dzin#action dzin to discuss buddy calls with james_w12:20
dzin#topic release12:21
salgadojamestunnicliff_, I think that's more because we're working together on the same project12:21
jamestunnicliff_  salgado, true.12:21
dzinAre there any issues or blockers?12:22
dzinpreventing release?12:22
=== plars-zzz is now known as plars
dzingood :-)12:24
dzinjamestunnicliff_, salgado, progress on private hwpacks?12:25
jamestunnicliff_Not as quick as I would like, but that is the result of illness :-|12:25
salgadoyes, I've made good progress towards supporting private projects at the model layer12:25
dzinhave all the BP's been created and updated?12:26
salgadoyep12:26
salgadowe have 11.09 and 11.10 BPs and I'm updating them12:26
salgadoI think jamestunnicliff_ is updating them as well12:27
dzingreat. Anything else that needs to be addressed?12:27
jamestunnicliff_not from me12:28
deepti1dzin: ah! i am seeing a strange issue with my kernel builds..12:28
dzinis that the reboot bug?12:28
deepti1dzin:  Am seeing that the boot and test are fine .. but it fails to reboot successfully ... yeah12:28
deepti1dzin: am triaging the problem .. but no luck.. anyone has some idea on this ?12:29
dzindeepti1: you can proceed as discussed: triage the bug, and submit the announcement.12:30
deepti1dzin:  I hope to get the root cause .. but if it is not the case then I would include it as a known issue ?12:31
deepti1dzin: salgado jamestunnicliff_ ^^^12:31
jamestunnicliff_deepti1: if you have some instructions I can try and recreate it on my board. It did sound a lot like the issue I was seeing with my old board.12:32
jamestunnicliff_But if it does happen on my board, I am not sure I can add much to our conversation the other day.12:33
deepti1jamestunnicliff_: you just need to use my hwpack and then try to reboot it .. it will hang at the message like the one here http://paste.ubuntu.com/697727/12:33
jamestunnicliff_Probably worth a call out on linaro-dev12:33
jamestunnicliff_deepti1: where do I get your hardware pack from?12:33
salgadodeepti1, I can help reproducing as well, but I only have a beagle xm12:34
deepti1jamestunnicliff_: http://ci.linaro.org/gcc-linaro/hwpack_linaro-panda_20110927-1110_armel_supported.tar.gz12:34
deepti1jamestunnicliff_: ok.. I will give u a hwpack after the meeting12:34
jamestunnicliff_deepti1: thanks. Will take a look12:34
deepti1jamestunnicliff_: thanks a lot for asking .. I need some help here12:34
dzindeepti1: what is the bug ID?12:36
deepti1dzin:  am yet to file12:36
dzinok12:36
deepti1dzin:  will do shortly12:36
dzin#action deepti to send query to linaro-dev concerning kernel reboot bug12:36
dzin#topic any other business?12:37
dzinhave I forgotten anything?12:38
jamestunnicliff_dzin: I think we are good to finish12:39
dzinok, meeting ends in 3,2,12:39
salgadoI don't think so12:39
dzin112:39
dzin#endmeeting12:39
linarobotMeeting ended Tue Sep 27 12:39:53 2011 UTC.12:39
linarobotMinutes:        http://irclogs.linaro.org/meeting-logs/linaro-meeting/2011/linaro-meeting.2011-09-27-12.02.html12:39
linarobotMinutes (text): http://irclogs.linaro.org/meeting-logs/linaro-meeting/2011/linaro-meeting.2011-09-27-12.02.txt12:39
linarobotLog:            http://irclogs.linaro.org/meeting-logs/linaro-meeting/2011/linaro-meeting.2011-09-27-12.02.log.html12:39
linarobotWiki:           http://irclogs.linaro.org/meeting-logs/linaro-meeting/2011/linaro-meeting.2011-09-27-12.02.wiki.txt12:39
dzinthanks all12:40
krtaylorhello13:01
benjiGhi13:01
ronynandyhi13:01
kan_huhi13:01
krtaylorlooks like everyone is here except ibiris and wei_feng13:02
krtaylor#startmeeting13:03
linarobotMeeting started Tue Sep 27 13:03:02 2011 UTC. The chair is krtaylor.13:03
linarobotUseful Commands: #action #agreed #help #info #idea #link #topic.13:03
krtaylor#topic Review Actions13:03
krtaylornot alot of formal actions from last meeting13:03
krtaylorthe one we logged was on CC attribution13:03
krtaylorI have been following the email thread on that, with good guidance from lool13:04
krtaylorI have not seen any problems raised13:05
kan_huso approved?13:05
krtaylorthat should be a quick approval for the TSC13:05
krtaylorkan_hu, not yet, I believe they meet tomorrow13:05
krtaylorit may happen in email, or at the meeting, I'll get that scheduled13:06
kan_huokay13:06
krtaylor#action schedule TSC review for this week for CC approval for samplemedia package13:06
kan_huwayne is trying connect IRC, some network wrong with his home internet connection13:06
krtaylorkan_hu, ok, thanks for the update13:07
krtaylorany other action review comments?13:07
krtaylor#topic Team Status13:07
krtaylorfrom the top13:08
krtaylorbenjiG, interesting comments13:08
krtaylorre: snowball13:08
krtaylor#link https://wiki.linaro.org/WorkingGroups/Middleware/Multimedia/Notes/2011-09-2713:09
benjiGho, snowbal doesn't MMU on hw IP13:09
benjiGso scatter list is useless13:09
robclarkbenjiG, fwiw, scatterlist can have just one node13:09
robclarkwe still need to add some flag in dev->dma_params to indicate that "remapped" buffer for DMA must be contiguous13:10
* wei_feng come13:10
krtaylorhi wei_feng13:11
wei_fenghi13:11
benjiGin my dummy driver it works fine without scatter list13:11
robclarkbenjiG, I think you could use something like sg_is_last() or sg_is_chain() to say that there is only one piece of memory in the scatterlist..13:11
robclarkbut the idea for using scatterlist is that the API could handle both cases.. both devices that need contiguous and those that don't13:12
robclark(that said, I haven't played too much w/ scatterlist myself.. so maybe there is a better way.. but I think somehow or another we need to handle both cases)13:12
benjiGrobclark: scatter list is useful for MMU capable hw IP13:12
krtaylorok, benjiG, so the internal reviews, to gear up for snowball support?13:13
benjiGbut basic data* field in priv struct is easier to deal with for snowball13:13
robclarkI guess you would use something like sg_dma_address() which should give you an address that is contiguous from the purpose of your device..13:13
benjiGkrtaylor: not for snowball, it is for next STE platform13:14
benjiGit seems it is too late for snowball13:14
krtaylorok13:14
krtaylorany other doc edits?13:15
benjiGno13:16
krtaylorany other OMX news?13:16
benjiGno13:16
krtaylorok, anything else?  ;-)13:16
krtaylorkan_hu, your status?13:17
benjiGI'm busy on internal Linaro/UMM/V4L2/ALSA promotion13:17
krtaylorbenjiG, good!13:17
kan_huSetup some compile env for the linaro kernel13:18
kan_huget kernel source from jesse's  branch13:19
kan_huand build for i.mx5 platform13:19
krtaylorkan_hu, all is done on sample media package, just waiting on CC approval right?13:20
* tgall_foo arrives13:20
kan_huIf mpeg4 is ok13:20
kan_huwe can based on david's ppa13:20
krtaylorok, if it can happen tomorrow, we'll see13:21
kan_huor if something is not ok for release, just repack, I think it will not take more than one day to do repack things13:21
kan_huhmm, so it's also ok for 11.09 release?13:22
kan_huI'm not sure about the dealine for 11.09 is ?13:22
kan_hudeadline13:22
krtaylorI just realized I didn't copy/forward the Freescale contact from Jesse13:22
krtaylordeadline is now basically13:23
krtaylorThurs is the release, but this is not rolled up into a LEB13:23
krtaylor29091113:24
krtayloranyway, I'll forward the contact info right after meeting13:24
kan_hucontact of freescale?13:25
kan_huyou means TSC?13:25
krtaylorfor any UMM work13:25
krtaylorfrom plumbers conference, it might help13:26
kan_huokay13:26
krtayloranything else?13:26
kan_huno13:26
krtaylormru, your status?13:26
krtaylorrequirements, perf study and ?13:27
mrunothing you don't know of13:27
krtaylorplease update the wiki so everyone knows  ;-)13:27
mrulooking at libpng to get an estimate of work needed to optimise it13:27
krtaylorgood!13:27
mrufixing those x264 neon functions13:27
krtaylorhow does it look?13:27
krtaylorit=libpng13:28
mruit's lacking any kind of framework for asm opts of course13:28
mruthe actual code isn't terribly difficult13:28
krtaylorok, let me know, but I think that will be one of our "requirements"13:30
mruwhat will?13:30
krtaylorwell, its gathering up a few optimization opportunities for next quarter13:30
krtayloras we are doing in the requirements doc13:31
krtaylorTSC will hopefully have feedback for priority unless it is 2-3 day effort, in which case, we get to self-prioritize13:32
krtaylorok, so anything else mru?13:32
krtaylorronynandy, how is PoC for release?13:33
mruthere are a couple libav neon patches in need of review13:33
mrushould only take an hour or so if they're good13:33
ronynandykrtaylor: its fine.no queries till now13:34
krtaylormru, ok, good13:34
krtaylorronynandy, I expect it will get more active over the next week13:34
ronynandyi am also planning to run the latest images on panda tomorrow..13:34
krtaylorbetween us, we can have 24hr coverage  ;-)13:35
ronynandy:)13:35
krtaylorok, anything to add to wiki status?13:35
ronynandythats all13:36
krtaylorrobclark, your status?13:36
robclarkcouple random things...   extracted dri2 into a separate lib for mesa (since I think this way will be easier to extend dri2 for video)..13:37
robclarkadded omap support in libdrm... still work in progress but when API settles and omapdrm is upstream, then I'll push a patch there..13:38
robclarkrebased omapdrm kernel stuff on linux-next13:38
robclark(linux-next has the dss2 patches I depend on.. so should be getting close to ready to push omapdrm in final (non-RFC) form)13:38
robclarkover weekend started adding tiler support.. which isn't strictly required but made me realize I want to adjust a couple APIs before I push final omapdrm patch..13:39
robclarkand that is it13:39
krtaylorok, thx13:39
krtaylorplease review the requirements doc and add any comments13:40
krtaylorwei_feng, dts decoder test was interesting!13:41
mrudts decoder test, where?13:41
robclarkkrtaylor, did you ever fix the permissions issue w/ requirements link?13:41
wei_fengyeah, i think it's enough on panda13:41
krtaylormru, we are working off the wiki13:41
mruwhat wiki?13:42
krtaylorwei_feng, but imx53 needs work?13:42
krtaylormru, https://wiki.linaro.org/WorkingGroups/Middleware/Multimedia/Notes/2011-09-2713:42
mru"the wiki" is about as vague a description as can be given13:42
wei_feng50% is too high i think13:42
robclarkmru, it's on the internet :-P13:42
wei_fengbut i.mx53 is cortex a8 single core13:42
mruwei_feng: that's very strange13:42
* ronynandy is confused13:42
krtaylormru, same place I asked you to update your status13:42
wei_fengpanda is cortex a9 dual core13:43
mrukrtaylor: there was nothing about dts there then13:43
mruoh hold old13:43
mruis imx53 an a8?13:43
wei_fengyes13:43
krtaylormru, refresh13:43
mrukrtaylor: I did13:43
mruwei_feng: then there's some non-neon floating-point code eating time13:44
krtaylormru, look under Wayne  ;-)13:44
mruyes yes, I see it now13:44
wei_fengi don't know if there's floating-point now13:44
mruI'm 98.3% certain that's the cause13:44
mruhow are you benchmarking it?13:45
wei_fengi can confirm it13:45
wei_fengi use gst-launch13:45
mruso it could be anywhere13:45
mruget a profile13:45
mruand try running avconv with null output13:45
wei_fengfilesrc + mkvparser + dtsdecode + fakesink13:45
mruand the dts decoder is mostly optimised13:46
mruthe missing piece is for pre-transform downmixing of multi-channel audio13:46
wei_fengi can test it with libav directly13:46
wei_fengbefore this Thur13:46
wei_fengno downmixing in my case13:47
mruthen we need that profile13:48
wei_fengok, then the other thing for me is pulseaudio ucm13:48
wei_fengmru, i will give it before this Thurs13:48
krtaylorwei_feng, good13:48
wei_fengseems it won't be integrated to pulseaudio before 1.013:48
krtaylorwei_feng, so you go the latest PA code for UCM13:49
wei_fengi got it from slimlogic13:49
krtaylorok, good13:49
wei_fengwhat i cares is route policy, seems related to ucm13:50
wei_fengneed more investigation for me13:50
krtayloryes, that would be a good thing to discuss with Colin and all in maillist and/or irc13:50
wei_fengcurrently i need studying more code in pulseaudio :-)13:51
wei_fengi will discuss it when got problem13:51
krtaylorI can help somewhat if you have questions13:51
wei_fengok13:51
krtaylorwei_feng, it would be good to at least let them know you are looking into it13:51
wei_fengok, that's all13:52
krtaylorok, thx13:52
krtaylortgall_foo, you around?13:52
tgall_fookrtaylor, yup13:52
krtayloryouve been busy with release I am sure, but still knocking down work items on LJT I see13:53
tgall_fooyup ...  I put my status into the meeting notes ...  tho just about the time the meeting started13:54
tgall_foogenerally speaking ltj should be in oneiric any moment now,  and the release for 11.09 was made last thursday13:54
krtaylorok, keep us posted on the benchmarking - that will be interesting13:54
tgall_footho due to a build hickup depending on which build was picked for test, it might or might not have the update13:54
tgall_fooyup will get a little more time to do some benchmarking soonish today / tomorrow ... knocking off two other blueprints first13:55
krtaylorok13:55
krtayloranything else?13:55
tgall_foonope ...  11.10 stuff is next in queue ... plenty of that13:56
krtaylor#topic New Business13:56
krtayloryes, and that brings me to the next subject13:56
krtaylormore "requirements" documentation is needed13:57
krtayloryou all are the experts in your area, and we need more information13:57
krtaylorplease go over the document I shared with all of you and add any comments, suggestions, concerns13:58
krtaylorlet me know if you cannot get to the doc13:58
krtayloryou can email me as well, whatever you are comfortable with13:58
krtaylorsecondly, thanks everyone for your push last week to get evaluations done13:59
krtaylorlet me know if you have any questions about that13:59
krtaylornext quarter we will have a new system that will hopefully be easier to complete for all of us14:00
krtaylorLastly, book your travel plans for Linaro Connect Orlando if you have not already14:01
krtaylorok, any other new business?14:01
krtaylornew/old business?14:02
krtaylorok, thanks everyone14:02
krtaylor#endmeeting14:02
linarobotMeeting ended Tue Sep 27 14:02:53 2011 UTC.14:02
linarobotMinutes:        http://irclogs.linaro.org/meeting-logs/linaro-meeting/2011/linaro-meeting.2011-09-27-13.03.html14:02
linarobotMinutes (text): http://irclogs.linaro.org/meeting-logs/linaro-meeting/2011/linaro-meeting.2011-09-27-13.03.txt14:02
linarobotLog:            http://irclogs.linaro.org/meeting-logs/linaro-meeting/2011/linaro-meeting.2011-09-27-13.03.log.html14:02
linarobotWiki:           http://irclogs.linaro.org/meeting-logs/linaro-meeting/2011/linaro-meeting.2011-09-27-13.03.wiki.txt14:02
rsalvetiWill be a few mins late18:03

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!