/srv/irclogs.linaro.org/2011/09/21/#linaro-android.txt

=== fgiff_ is now known as fgiff
beropaulliu: let me know when you're ready for the iMX53 session07:56
paulliubero: yeah, I'm ready..07:58
asaccool :)08:00
asacseems netsplit is over :)08:00
paulliubero: so I'm debugging the hang right now. It hangs very early.. Not sure if it is because the serial driver dies or the clk driver dies.. I'm checking..08:02
beropaulliu: Yes, I'm also not sure what's going on there... And I think this hang is the only topic we really have right now08:03
beroWithout a fix for that, everything else is pointless08:03
paulliubero: yeah.. let me do some debug...08:03
beroOf course if we can't fix it properly in time for the 2011.09 release, the other topic to talk about is possible workarounds (such as going back to 2.6.38 or just applying the Android patches on top of your kernel instead of trying to merge your work into the jstultz tree)08:04
paulliubero: yeah.. that's what we are thinking about..08:15
paulliubero: I mean, it seems that we can run Ubuntu on top of the Android kernel..08:15
paulliubero: That is we can include android patches into our Landing team kernel anyway..08:16
beroYes08:16
beroThere's also Android patches in the Ark Linux kernel these days (even though it's a "normal" Linux)08:17
paulliubero: So it is ok for you to not using 3.0 kernel?08:17
beroWell, in the longer run, we want every device we support to be based on the same kernel source, which is the jstultz tree08:18
paulliubero: hmm.. Not each landing team provide an android kernel so you can run on top of it??08:18
berobut I think given you already have 3.1 running, that's an option for the short term08:18
berowith the longer term plan being to merge the changes into the jstultz tree when he updates to 3.1 -- should be less trouble08:19
paulliubero: Because only landing team has the drivers needed but not upstreamed yet..08:19
beroI think everyone would be ok with that08:19
asacpaulliu: bero: we very much would like LT kernels to have jstultz changes inside08:19
asacif this doesnt work this time (because LT kernels work off ubuntu for now) then please lets work on having this in the LT process to test08:20
beroThe other landing teams do provide the kernels for the boards08:20
asacah08:20
asacon i think we have 3.1-rc for fsl?08:20
paulliuasac: yes08:20
asacbut android team is running a 3.0 backport?08:21
beroasac: yes08:21
asachas anyone tried to use 3.1  :)?08:21
=== Guest97181 is now known as mabac
asachmm. guess we dont have android patches for that08:21
asacbecause aosp is down08:21
asacthats really gettinng painful :(08:22
paulliubero: http://paste.ubuntu.com/694308/08:22
beroasac: I expect forward-porting the Android patches to 3.1-rc will probably be a lot easier than backporting all the LT bits and pieces to jstultz-3.008:23
paulliuHangs after TrustZone Interrupt Controller (TZIC) initialized08:23
beropaulliu: Ok, that should give us a better starting point...08:24
asacbero: hmm. anyone might go to #android and ask if they can help out by pushing the common-3.1 kernel somewhere to open :)?08:25
asacexplain the situation etc. :)08:25
asacwho knows08:25
paulliuasac: what patches is in jstultz-3.0??08:25
paulliuasac: hmm... we expect it to be in the upstream?? not true?08:25
asacpaulliu: jstultz_vm works on consolidating android ... he did some work replacing android parts with upstream parts etc.08:25
paulliuOK...08:26
asacpaulliu: and he is actively working with android folks on getting those adopted08:26
asaci dont know exactly whats in there now08:26
asacjstultz_vm: where do you track your android work atm?08:27
asace.g. is there some place where i can go and see what is done/inprogress etc.?08:27
beropaulliu: It also has some more drivers for the various boards we're supporting, so it would be another kernel to upstream landing team changes to08:27
paulliubero: hmm.... that could be double effort..08:28
asacbero: interesting. i guess those come from the backports done on linux-linaro by KWG?08:28
beroasac: I'm not sure where they come from, I just saw a couple of additions when backporting the LT patches08:29
asacbero: right. so whats the plan on backporting? are we focussing on one or two hardware parts? or trying to pull everything?08:29
asace.g. just wifi/bluetooth or so?08:29
asaci guess there are other significant changes iin their tree that are not drivers? :/08:30
beroI think there's some DeviceTree integration bits08:31
paulliubero: that's funny actually.. Because I ask the author of MX5 device tree about the progress of upstreaming.. Actually those are still not good for upstream yet...08:32
asacpaulliu: does fsl have a open wifi/bluetooth driver? i think if we focus on anything we should pick one of those topics, because we did that for panda and are working on origen for that08:33
paulliubero: current patch is just proof-of-concept..08:33
paulliuasac, not wifi and bluetooth on MX53 QuickStart board..08:33
beroasac: The iMX53 QSB doesn't have wifi or bluetooth08:33
asacso device tree we can skip. we have device tree as a roadmap entry for next quarter iirc08:33
asac(for android)08:33
paulliuasac: I think the driver we have here which is important is Display, USB and PMIC...08:34
asacpaulliu: hmm. was there any improvement on hardware enablement compared to what we have in last leb?08:34
asacah ok ... so those improved?08:34
paulliuasac, Just prepare for upstreaming so it works as well as the last leb..08:35
paulliuasac: runs on new versions of kernel..08:35
asachmm08:35
paulliuasac: There are some bug fixing also.. But not critical bugs.08:37
asachmm.08:37
beropaulliu: Are the power management issues (e.g. shutting down power to the root filesystem) fixed?08:37
asacanything that would improve our android leb at all?08:37
paulliubero, yes..08:37
beroasac: So there's the answer -- power management would definitely improve the leb08:38
asacright08:38
asacdoes that also bring us more power management features tracked by amit?08:38
asachttps://wiki.linaro.org/WorkingGroups/PowerManagement/Matrix110508:38
asacor just "basics" like shutdown etc.08:38
asacdoes suspend etc. work?08:39
paulliufor 2.6.38 we believe it is fixed. For 3.1 we are still waiting the test report...08:39
asacsuspend or the whole matrix?08:39
paulliuasac: just suspend..08:40
asacbero: maybe we can just cherry pick pm fixes? or are those typically spread across everywhhere?08:40
beroasac: typically not08:41
asacpaulliu: heh. do you know the status of the matrix?08:41
paulliuasac: no...08:41
asackk08:41
beropaulliu: Any details on where the pm fixes went? I didn't see any pushes to my tree... So it's probably on a branch in the LT git?08:41
paulliubero: let me check the bug report..08:42
paulliuwe did assign someone to look into that problem..08:42
paulliuCause I'm still not able to enable my current board for PM stuff..08:43
asacliuyq: hi!08:45
paulliubero: hmm.. seems not fixed yet.. Do you have the kernel and configuration for that bug to reproduce?08:46
liuyqasac, hi08:46
paulliubero: as my hardware is new with different PMIC. The driver is still not available yet..08:48
beropaulliu: Yes, that's the kernel in my old repository... git://git.linaro.org/people/bernhardrosenkranzer/android-kernel-iMX53.git, lt-android-38 branch (It's a clone of your team's lt-android-38 branch with necessary fixes applied on top of it)08:48
paulliubero: I think you need to enable CONFIG_EARLY_PRINTK=y for debugging the current 3.0 kernel.08:50
paulliubero: and check if it hangs at the same place..08:51
beropaulliu: Oh, it's disabled in the config! That's why I didn't see anything useful when adding the earlyprintk command line...08:51
beroYes, will do that right away08:51
mansson_Man måste ha en hel git på -u, inte en xml-fil! Och när jag synkade fick jag merge-konflikt i default.xml kanske för att jag inte synkat på länge?. Jag får klona ett nytt repo. Jobbigt.09:38
paulliubero: http://paste.ubuntu.com/694348/09:38
paulliubero: After disabling UART clk, crash is found here.09:38
paulliuSeems some devices are register twice?09:39
beroThat's quite possible... It's a typical mistake you'd get from automerges (merging a change from one tree when the same change has already been made in the other tree in a slightly different location...)09:40
beroIs it supposed to create imx31-gpio.0 at all? The name seems rather odd on mx5309:40
paulliubero, yeah.. quite odd..09:41
beroI think the duplicate initializations are related to the multiple board support bits...09:44
beroimx31-gpio seems to be created in arch/arm/mach-mx5/mm.c in both imx51_soc_init() and imx53_soc_init(), and the errors look like both of those are being called09:44
beroMaybe we should be a bit more conservative in the config and just disable MX51 support etc., the only board we have to care about right now is the iMX53 QSB09:45
* bero tries09:45
=== pfefferz is now known as blarg
=== blarg is now known as __oo
=== __oo is now known as pfefferz
=== pfefferz is now known as Guest42774
Guest42774https://bugs.launchpad.net/linaro-android12:51
pabhisheknext12:53
Guest42774pabhishek, https://android-build.linaro.org/builds/~linaro-android/stage-snowball-11.09-release/#build=512:56
=== Guest42774 is now known as pfefferz-x
=== pfefferz-x is now known as _pfefferz
=== _pfefferz is now known as pfefferz_
=== pfefferz_ is now known as help
=== help is now known as pfefferz_
jstultz_vmasac:  https://blueprints.launchpad.net/linux-linaro/+spec/linaro-kernel-o-android-upstreaming14:59
bhojmpoirier, ping15:40
mpoirierbhoj: I'll get back to you in 5 mins15:40
bhojmpoirier, okay15:40
beroPacking up this computer for the move next... I'll be back in a bit from my notebook15:41
mpoirierbhoj: all yours15:45
bhojmpoirier, : )15:45
bhojmpoirier, does the notification icon for ecm enabled ?15:45
mpoirierbhoj: notification icon ?15:46
bhojmpoirier, ethernet is working but I don't see any icon....15:46
bhojmpoirier, ecm has a notification icon similar to wifi ....15:46
mpoirierbhoj: ya, it's just dawning in me...15:46
mpoirierbhoj: yes yes...15:46
mpoirierbhoj: humm...15:47
mpoirierbhoj: I don't know if there is one in the x86 distro.15:47
bhojmpoirier, hmmmm ...15:47
mpoirierbhoj: I'm assuming it wouln't compile had I missed it...15:47
mpoirierbhoj: so, either it's not there at all or the icon, for some reason, isn't showing.15:48
bhojmpoirier, I have not tried x86 anytime recently . Ok .... I have seen it in rowboat...let me see if I can find a screenshot for that15:48
mpoirierbhoj: rowboat ? educate me...15:49
bhojrowboat is a TI inititiative .open source android distribution for TI boards...like beagle,beaglexm,omap3EVM15:49
bhojmpoirier, http://code.google.com/p/rowboat/15:50
mpoirierbhoj: by the way, I never got x86-droid to compile...15:50
mpoirierbhoj: ok, I'll check it out...15:50
mpoirierbhoj: you're on the panda right now ?15:51
bhojmpoirier, I am testing beagle-xm...panda on its way .Do you need anything on that ?15:51
mpoirierbhoj: no, just for my info...15:51
mpoirierI'll fire up my board and see...15:52
bhojmpoirier, check this video http://www.youtube.com/watch?v=AIWo1YSi6ng15:52
mpoirierI hve the battery icon as an example.15:52
bhojmpoirier, there is Ethernet Icon on the right15:52
mpoirierbhoj: ok, watching...15:52
bhojmpoirier, video at 1:1015:53
bhojmpoirier, I am not sure if this icon is an addon from them ....15:54
mpoirierbhoj: isn't the icon on the left ?15:55
bhojmpoirier, sorry you are right .Its left :)15:55
bhojmpoirier, http://news.softpedia.com/newsImage/Run-Android-for-Your-PC-2.jpg/15:55
bhojandroid-x86 had the ethernet icon15:56
mpoirierbhoj: having the x86-android code to compile would solve the matter right away...15:56
mpoirierbhoj: I see...15:56
bhojmpoirier, hmmmm yes.May be the latest code doesn't have the icon .15:57
mpoirierbhoj: ok, I'll look into it.  Thanks for pointing it out.15:57
bhojmpoirier, np.I just wanted to verify with you if it was an issue only on my board :015:57
mpoirierbhoj: it's not an issue... I don't see the icon either.15:57
mpoirierbhoj: but I'm adamant...  Had I missed the icon the code wouldn't compile.15:58
bhojmpoirier, okay  ...I think this is necessary to indicate user if he is connected or not.15:58
mpoirierbhoj: ya, I'll look at it.15:59
bhojmpoirier, Have you looked at status bar code ?? if that needs to be enhanced ??15:59
mpoirierbhoj: hook something up ya.16:00
mpoirierbhoj: I haven't - this was my very first foray in userspace.16:01
mpoirierbhoj: I'm usually in ther kernel.16:01
bhojmpoirier, okay .16:01
mpoirierbhoj: where is that status bar thing you're referring to ?16:02
mpoirierbhoj: location in the code I mean..16:02
bhojmpoirier, just a minute16:02
bhojframeworks/base/packages/SystemUI16:04
bhojmpoirier, frameworks/base/packages/SystemUI16:04
mpoirierbhoj: cool thanks.16:04
bhojmpoirier, np16:05
mpoirierpfefferz_: mumble ?16:27
pfefferz_mpoirier, would you like to mumble now, sorry, for not being available sooner18:13
mpoirierpfefferz_: sure !18:13
pfefferz_mpoirier, here's the BP: https://blueprints.launchpad.net/linaro-android/+spec/linaro-android-rebase-snowball-onto-upstream-tree18:46
mpoirierpfefferz_: got it thanks18:47
mpoirierpfefferz_: you filed it for 11.10 - that's awesome18:47
mkedwardspfefferz_: do you know whether the Samsung LT has a branch that tracks the CMA / dma-mapping / iommu work that has been done for exynos4?18:47
pfefferz_yeah, 11.09 is basically done18:47
pfefferz_just bug fixes etc..18:48
pfefferz_mkedwards, I don't know18:48
pfefferz_mkedwards, I can get an email with Angus to ask him about it18:48
pfefferz_one sec18:48
pfefferz_mkedwards, sent18:51
mkedwardspfefferz_: I am trying to get a feel for what tree I could base a multi-vendor iommu example on18:53
pfefferz_:)18:53
pfefferz_I <3 iommu stuff18:53
pfefferz_Qualcomm did an example implementation already18:53
mkedwardsOmar Ramirez Luna at TI did some work in this area for the zoom3 last month18:54
pfefferz_would you like to talk to them18:54
mkedwardsmostly I'm trying to figure out how this would interact with SELinux labeling for inter-webapp resource isolation (and permission delegation) on Boot to Gecko18:55
mkedwardspurely exploratory18:55
pfefferz_mkedwards, sure...18:56
mkedwardsvery high-level use case: screen scraping for WebEx-like functionality18:56
pfefferz_:)18:56
mkedwardshow deep can we plumb the barriers that protect browser tabs / webapps that you don't want to share inadvertently18:57
mkedwardscertainly something we would want to work together with WebKit-based browser teams on ;-)18:58
mkedwardsdo you have a pointer to the Qualcomm work?  I only know about what has hit linaro-mm-sig18:59
mkedwardscan anyone work with inSignal / origenboard.org to fix their pwned-by-spammers web presence, "sold out" status, and general air of abandonment?20:54
mkedwardsnot the ideal state for a high-profile Linaro-affiliated project20:55
pfefferz_mkedwards, hmmm21:16
mkedwardsin addition, I'd like to order a stack of 'em21:17

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