pfefferz_joey, +100:46
=== plars-afk is now known as plars
plarsmwhudson: wow, colors01:23
plarsmwhudson: hear anything from zyga tonight?01:28
mwhudsonplars: no01:28
mwhudsonwell, maybe he's sent mail, but i'm having an "ignore my inbox" day :-)01:28
plarsmwhudson: heh, yeah I have days like that too.  Reading it with the delete key held down is another strategy :)01:29
plarssome people just call it "skimming", I call it "expedited inbox reduction"01:30
mwhudsonwell, i will read it, and soon01:30
mwhudson(monday i guess)01:30
mwhudsonjust not today01:30
mwhudsonhttp://people.linaro.org/~mwh/ci-mockups/nested-js.html has an update02:43
springzmwhudson, seems complex02:58
mwhudsonspringz: well yes03:00
mwhudsonspringz: it's a lot of data03:00
=== springz is now known as springz-afk
=== Amaranthus is now known as Amaranth
=== springz-afk is now known as springz
jcrigbyrsalveti, are you around?04:27
rsalvetijcrigby: hey, yes04:27
jcrigbyworking on u-boot all day, finally got final fix commit just now04:27
jcrigbywith our extra code u-boot-spl was a little too big and did not fit in the allow part of sram04:28
rsalvetijcrigby: ouch04:28
rsalvetijcrigby: which extra code?04:28
jcrigbyactually its not that much bigger04:28
jcrigbyit is already over with upstream04:28
jcrigbyand then we put it over somemore04:28
rsalvetioh, ok04:29
jcrigbyI need to go back and see if some code can be omitted from spl04:29
jcrigbythat is in some library 04:29
jcrigbylike the pxe uuid stuff04:29
jcrigbyand maybe more04:29
rsalvetihm, was using the upstream yesterday04:29
rsalvetiwith spl and such04:29
jcrigbythe problem is that aneesh upped the size a while back but did not change the base04:30
rsalvetibut didn't check if it was oversized 04:30
rsalvetioh, interesting 04:30
jcrigbythe boot script checks the size that he defines in omap4_common.h04:30
jcrigbybut that size is actually too large04:30
jcrigbyI will send him an email tomorrow04:30
jcrigbyanyway I just pushed it to my -next04:31
jcrigbyif you get a chance try it on 446004:31
rsalvetijcrigby: oh, it's on next04:31
rsalvetiwill test it right away04:31
jcrigbythere was also a problem with upstream at least for me04:33
jcrigbyyou will see a revert under the "fixes" branch that causes kernel fault on my board04:33
jcrigbyand the other problem is that usb does not work in the kernel if you don't turn on usb clocks on u-boot04:34
jcrigbythe external phy gets in a bad state04:34
rsalvetijcrigby: yeah, that I noticed too04:35
rsalvetiwhen working with the upstream code I had to use your 2 patches to enable usb04:35
jcrigbyOMAP4 Panda: clock config for USB04:35
rsalvetiso I could have usb working on it04:35
jcrigbyfixes that04:35
jcrigbyI have not yet tried usb in u-boot04:35
rsalvetibut didn't know if it was entirely my fault or not04:35
jcrigbychasing all these other problems first04:36
rsalvetijcrigby: I didn't saw this kernel fault you said, which kernel did you use?04:36
rsalveticould be because I was only testing the tilt one04:36
jcrigbylatest hwpack04:36
rsalvetiso I believe there's something else at the tilt tree that fixes at the kernel side04:36
jcrigbythe patch looks innocent but when I remove it the problem goes away04:37
rsalvetidon't know, just a guess04:37
jcrigbythank goodness for git bisect04:37
rsalvetithere was a bunch of i2c related changes at the tilt tree04:37
rsalvetijcrigby: http://git.linaro.org/gitweb?p=people/rsalveti/linux-linaro-3.0.git;a=shortlog;h=refs/heads/tilt-linux-linaro-3.0-rsalveti04:37
jcrigbyit could be a kernel bug that only happens when the i2c is in a certain state04:37
rsalvetilike when checking on 4460 and tilt-3.0 I had to port a few i2c related patches04:38
rsalvetiso I could get it working04:38
rsalvetiI noticed that because when working against the tracking tree, the bug didn't show up04:38
rsalvetiso I just cherry-picked them and the problem went away04:38
rsalvetitracking is on top of 3.1-x04:38
rsalvetiguess that is the patch that fix the issue04:40
rsalvetibut I applied the others as it was depending on them04:40
=== michaelh1 is now known as michaelh1|away
rsalvetijcrigby: ok, got it here, will build and test04:43
rsalvetibut the tree sounds sane04:45
rsalvetijcrigby: https://code.launchpad.net/~rsalveti/+recipe/u-boot-daily04:46
jcrigbyjust tried usb/net04:46
rsalvetidaily build for u-boot upstream04:46
jcrigbybootp works after a timeout04:46
rsalvetibuilding same packages we build for u-boot-linaro04:46
rsalvetijcrigby: cool, at least it works :-)04:47
jcrigbybut looks like I have not enable pxe04:47
rsalvetioh great, now my 4460 doesn't boot anymore04:49
jcrigbyoh crap04:49
rsalvetiguess it's missing a proper config for the pmu module04:49
rsalvetihappens from time to time04:49
rsalvetiso, sometimes when you just reboots after a working boot, everything goes fine04:49
rsalvetibut if you shutdown it completely, by removing the power cable, and turn it on again, it'll not boot04:50
rsalvetiboot the kernel, I mean04:50
rsalvetithis should probably be the reason04:52
rsalvetiso it's quite annoying 04:52
rsalvetibecause kernel can reset gpio1 during boot04:53
rsalvetiand gpio1 should be set at the boot loader04:53
rsalvetibut then you'll tight a boot loader version with a kernel04:53
rsalvetiand so on04:53
=== fangbao_ is now known as fangbao
bill4carsonIs anybody working on A15 KVM support?05:35
plarsspringz, mwhudson, liuyq: released everything except lava-server and lava-dashboard, packaging for dispatcher done, and it's been installed on validation.linaro.org05:41
plarsspringz, mwhudson, liuyq: oh, and I did the config file stuff for all the machines we have connected to validation.linaro.org.  So they should all work now, up through panda30 (once it's physically connected and setup in conmux).  I'm doing a quick sanity test right now on panda07, but don't expect any problems05:44
=== Quintasan_ is now known as Quintasan
plarszyga: I've released everything except -server and -dispatcher06:06
plarszyga: package updated and built locally for dispatcher, installed it on the server too06:07
springzplars, great! thanks06:07
plarszyga: I'll probably work on getting lava-android-test packaged and installed tomorrow, and continue with the rest06:07
plarsspringz: no, thank YOU :)06:07
zygaplars, thanks06:10
zygaplars, I lost my internet last night, I was copying 500MB of data to a virtual machine, pings went up to forever and IRC disconnected06:10
zygaplars, I saw you talking to mwhudson about dashboard and server stuff06:11
zygaplars, the upload too about two hours, I just went to sleep, sorry :/06:11
zygaplars, getting back to work now06:11
zyga(VM was in the cloud, not locally, so I taxed my outgoing connection06:11
plarszyga: ouch, yeah I almost wondered if you had some kind of connection problem06:12
deeptispringz: Hello .. 06:45
deeptizyga: springz: I am seeing some error with lava dispatcher.. can you confirm if its a problem with lava dispatcher ?  check for error here : http://paste.ubuntu.com/695479/06:46
janimoericm|ubuntu, around?06:49
* springz reading06:54
plarsdeepti: ugh, looks like the console got messed up somehow06:55
springzdeepti, what's the system, panda android?06:55
plarsdeepti: we used to have issues with that, there was a fix that should have helped things a looong ways back, and haven't seen problems with it since06:55
springzdeepti, '(echo \$?)' should replace by an integer06:56
faboplars: what's the plan for lava-server and lava-dashboard? someone else is going to release them?06:56
faboand aren't you supposed to sleep now ;)06:57
plarsfabo: zyga had internet problems, I think he's working on finishing up testing for them now06:57
plarsfabo: nah, it's only 2 AM06:57
plarsdeepti: I'll pull that board, see if it's something crazy with it.  It would be interesting to see if it's something in the kernel that triggered it though.  Any chance you could resubmit it?06:58
plarsdeepti: I can do it by hand if not06:59
plarsdeepti: also, is it ok with you if I cancel the job? I think it's certainly not going to go any further at this point06:59
=== fangbao_ is now known as fangbao
plarszyga: ugh... looks like this is going to be a late nigte07:10
plarszyga: this lava-test broke some of the tests07:10
zygaplars, the new branch?07:15
zygaplars, tell me more, perhaps I can help07:16
deeptiplars: http://validation.linaro.org/lava-server/scheduler/job/185407:17
deeptiplars: sure I have cancelled the job .. 07:18
davepigottdeepti: How long ago did you cancel it?07:18
deeptiplars: let me see if I can resubmit the job.. but even if I did it would be in the queue and ll take quite a long time for it to run as there are jobs already running and none of the panda boards free07:19
deeptidavepigott: just a while back07:19
davepigottdeepti: It says canceled. Just took a while to catch up07:19
zygaofftopic: it's nice to see a statement like "none of the panda boards are free" :-)07:19
davepigottzyga: :D07:20
deeptizyga: ah! yeah .. I am sure its a feel good factor that the farm is being used extensively07:20
plarsdeepti: ok, that console looks fine after the reboot, going to set it back online07:21
davepigottdeepti: The dispatcher bug means we can't get more up and running until that's fixed and deployed. Pity as I currently have 16 more connected and another 4 will be available today.07:21
davepigottdeepti: mind you, whether the server will take the load or not is a different question.07:22
plarsdavepigott: no bug there, that's a wishlist item :)07:22
plarsdavepigott: I just brought up a bunch more boards07:22
davepigottplars: You did?? Oh. You added them to the dispatcher list?07:22
plarsI have a bigger problem right now though07:22
plarsdavepigott: no, I installed the new dispatcher, and we use proper config files now07:22
plarsdavepigott: see /etc/xdg/lava-dispatcher/*07:23
davepigottplars: Ahhhh! OK! :)07:23
davepigottplars: I have the go ahead from Stephen to get the new server asap, btw07:23
plarszyga: hmm, trying to understand what this new TestParser.parse is doing exactly.  there are some tests that make their own version of it, but they seem to be doing something totally different from what it does now07:24
plarsdavepigott: awesome!07:24
deeptiplars: oh! I see one more error with job 1856 http://paste.ubuntu.com/695496/07:24
deeptiplars: hmm07:24
davepigottplars: He flinched when I told him how much it would be. ;)07:25
plarsdeepti: well, if it's the fact that a couple of the tests have a non-working parser, then that's what I'm looking at right now07:25
plarsdeepti: if it's the SIGILL you got, that's a kernel problem, not a lava problem :)07:25
plarsdavepigott: oh, did you get a quote already, or were you just throwing out ballpark numbers?07:26
plarsdeepti: also, your test kernel doesn't seem to be rebooting properly07:26
deeptiplars: ah! which job is that 1856 ?07:27
plarsdeepti: yes07:27
davepigottplars: Gave him ballpark. Am going to get a quote from CAE today and then comparison shop07:27
deeptiplars: ok 07:27
deeptiplars: check this http://paste.ubuntu.com/695499/ .. beagle test executions are failing bcs of the invalid choice for the --dev.. 07:34
deeptiplars: this is for job 185907:35
plarsdeepti: got it, was a config issue, fixed now07:46
deeptiplars: thanks08:03
deeptiplars: when can we expect the offline panda boards to be up ?08:04
jk-for some reason I read "panda boards" as "pants"08:07
mabacfabo, https://launchpad.net/linaro-image-tools/11.11/2011.09 l-i-t tarball is there and would like to be packaged :08:18
=== fangbao_ is now known as fangbao
* zyga starts the onslaught of lava-server branches and merge proposals08:25
fabomabac: thanks, /me grabs the tarball08:31
mabacfabo, yay!08:31
zygaI get permission denied on pushes to launchpad08:31
zygais launchpad read only now?08:31
mabaczyga, I think it's time for the daily service windoe08:31
mabaclot's of oopses atm08:31
plarszyga: yeah, I just hit it too :(08:31
zygaplars, I'm getting OOPS on code.launchpad.net/lava-server08:32
plarszyga: yeah, I got one when trying to update a bug08:32
plarszyga: good news is I'm sorting through the bugs pretty quickly08:32
zygalp does not oops on upgrades, right?08:32
plarszyga: I would hope not08:32
mabacIt does now. The updates now are supposed to be so quick that noone would notive them.08:33
plarszyga: seems to be back online08:33
mabacSadly I do every other day.08:33
mabacBut it usually only takes a few minutes.08:33
zygayup, it works for me again :)08:34
hrw~curse fedex PL badly08:43
zygaI keep getting oopses when trying to propose branches for merging08:57
plarszyga: still?08:58
zygaplars, yes08:58
plarsdeepti, zyga: ok, I fixed the parser errors with gtkperf and posixtestsuite too08:58
zygaplars, what was the cause of the lava-test issues? my rewrite?08:59
plarszyga: yep, but to be fair, it should have been tested with more of the testsuites clearly09:00
zygaplars, hmm09:00
zygaplars, I'm pretty sure I tested them with both that you reported09:00
plarszyga: some of them were fine, but some of them had not had the necessary changes because they implemented their own parser09:01
zygaplars, were there any other changes?09:01
zygaplars, I took great care not to break compatibility09:01
plarszyga: also, there's a file that hardcodes the list of in-tree tests... and it was missing lots of tests :(09:01
zygaplars, I actually not touched any of the original tests to check this09:01
zygaplars, that should have been updated once new tests were added before this got merged09:01
zygaplars, I talked with ChiThu about this one actually09:01
plarszyga: compatibility broke for gtkperf and openposixtest because they user their own parser class09:01
zygaplars, even with custom parser classes, strange09:02
plarszyga: to be fair, I probably could have just done def parse(self, artifact=None) instead09:02
plarszyga: but I also went ahead and added support for line numbers in the logfile while I was in there09:03
plarszyga: those two bypassed that09:03
zygaplars, ok09:03
plarszyga: anyway... it's all fixed up now09:03
zygaplars, still that's strange, look at how legacy parser checks for presence of custom parse method09:03
zygaplars, yeah, great work and sorry I had to keep you up at night :/09:03
plarszyga: this is why I keep pushing for us to land these branches by 2 days before release though09:03
plarszyga: you didn't, I did :)09:04
plarszyga: I couldn't go to sleep with crap broken09:04
plarszyga: even if it was just a few little things09:04
plarszyga: how's lava-server and lava-dashboard?09:05
zygaplars, lava server has two more patches, one is ready, other is just a few lines away09:06
zygaplars, I'll try proposing merges again in a second09:06
zygaplars, (you can see my branches on lp)09:06
zygaplars, dashboard is about three patches away, one is ready, two need checking and perhaps some updates as I wrote them a while ago09:07
plarszyga: heh, I'll take a look but I'm not sure I'm clearheaded enough for a code review right now.  I'll trust your testing more than me looking at it at 4am :)09:08
plarszyga: I can spot check it though09:08
zygaplars, no, go to bed please09:09
zygaplars, I trust the code is fine, I look forward to talking to you in the evening09:09
plarszyga: I'm debating between doing that, or working on packaging for lava-android-test09:09
zygaplars, how is your familiy?09:09
plarsI really want to see lava-android-test working in android CI :)09:09
plarszyga: good, just crazy schedules lately.  My wife and I have hardly seen each other all week because we usually have different kids at different events09:10
plarszyga: make sure to see the PM I sent if you haven't already09:11
alf_ibiris: FYI, glmark2 and glcompbench released, packaged, tested and copied to overlay09:11
ibirisalf_: \o/09:12
=== cltang_ is now known as cltang
zygaplars, PM?09:12
ibirisalf_: great stuff :D09:12
plarszyga: private message09:14
liuyqplars, zyga about the packaging of lava-android-test, I am sorry I can't do it myself until now:(. I had talked it with zyga. zyga, do you like to do it yourself, or instruct me to do it.09:17
plarsliuyq: I was planning on taking care of it later today09:17
plarsliuyq: no worries09:17
plarsliuyq: if there's time on monday during our sync, I can show you what goes into that09:18
fabomabac: see my message about the tarball :)09:18
plarsliuyq: in this case, it will be nearly identical to lava-test or lava-tool09:19
liuyqOK, thanks, I followed https://wiki.linaro.org/Platform/DevPlatform/ComponentReleasesWithBzr to do, but failed at step "debuild -uc -us -S -i"09:19
plarsliuyq: and hopefully at some point we'll have an adb package, and we can just depend on it :)09:19
janimomabac, fabo do you think that l-m-c works in oneiric and should be uploaded in the archives?09:19
liuyqyes, it's need adb. I have seen the package source about  lava-test, but not very clear about that.09:20
liuyqBTW, which is the job you defined for android by using lava-android-test on http://validation.linaro.org/?09:21
springzplars, I suggest we can split release procedure and components next milestone, with more members involved09:22
plarsspringz: well, the tarball release itself is actually easier if one person does it.  I've got scripts in my +junk that just take a list of everything you want to release and generate all the pieces for it.  The parts leading up to it (merging branches early, testing, testing, more testing) could use some help for sure though09:24
fabojanimo: yes, we discussed about it yesterday. james_w should look into this.09:24
ndecrsalveti: i saw that you created powervr-omap4 package in the linaro overlay PPA. we probably need to align on that.09:24
janimofabo, ah good09:24
plarsspringz: I still have a pipe dream that we should have everything for the release checked in 2 days before.  Spend wednesday and part of thursday with testing and bug fixes, then it would all be really easy09:24
springzplars, we can consider a previous week before release day as a pre-release week and start to integrate and test09:25
springzplars, code should be frozen that time, only bug fixes allowed09:26
plarsspringz: 25% of our cycle spent in freeze is a bit extreme I think.  In practice I think 2 days would be sufficient09:27
springzplars, after one week test and bug fix, should be ok for formal release09:27
plarsspringz: eventually we would prefer to get to more of a continuous integration, testing, release cycle09:27
deeptizyga: did plars changes that he did some time back fixes the http://paste.ubuntu.com/695543/ issue ?09:41
czajkowskimorning all 09:41
plarsdeepti: that is too small of a snippet to tell much, the whole log would have been more useful, but yes I think that should be fixed now09:43
deeptiplars: oops u r still awake.. 09:43
plarsdeepti: yup09:43
deeptiplars: ok let me get you the log 09:43
deeptiplars: ah! I am not able to get the log .. but here is the job http://validation.linaro.org/lava-server/dashboard/streams/anonymous/ci-linux-arm-soc-for-next/bundles/663b801afbc2aed508ab339448be99c52031e205/09:44
plarsdeepti: just the job number is plenty09:44
deeptiplars: sorry not the job but the bundle09:44
plarsdeepti: the log is here: http://validation.linaro.org/lava-server/dashboard/streams/anonymous/ci-linux-arm-soc-for-next/bundles/663b801afbc2aed508ab339448be99c52031e205/c3f1999c-e5ab-11e0-9729-68b599be548c/attachments/10387/09:44
plarsdeepti: "ERROR: There is no test with the specified ID" yes, that's fixed, see bug #85713709:45
ubot2Launchpad bug 857137 in lava-test "many tests not available in lava-test 0.3.0" [High,Fix committed] https://launchpad.net/bugs/85713709:45
plarsdeepti: and I've uploaded it too, so next time it runs it should be fine09:46
deeptiplars: ok thanks09:46
deeptiplars: ok..09:46
hrwdoko_: I see that gcc-multiarch.diff got simplified 09:54
=== davepigott_ is now known as davepigott
asacdavepigott: how is board landing in the lab going? i see many new offline boards :)10:42
=== agreen is now known as agreen-away
=== agreen-away is now known as agreen
botao_sunsangwook: ping10:52
sangwookbotao_sun: pong10:56
botao_sunsangwook: I'm writing an email to you, hold on.10:56
sangwookbotao_sun: ok10:56
* hrw -> customs office11:00
botao_sunsangwook: mail sent.11:00
* davidgiluk imagines a confused customs officer trying to balance the law on shipping endangered animals with the fact a panda couldn't possibly fit in a box that small11:01
janimoalf_, can you add some reasoning here, upload was rejected as it needs a FFE (now filed) and explanation. https://bugs.launchpad.net/ubuntu/+source/glmark2/+bug/85727911:02
ubot2Ubuntu bug 857279 in glmark2 "FFE for Oneiric" [Undecided,New]11:02
mabacjanimo, I can't tell since I only have natty to test on. but I see no reason why it shouldn't work11:10
mabacfabo, oh how embarrassing... I have uploaded the correct tarball now11:11
botao_sunsangwook: This path "drivers/net/wireless/ath" is only for origen ubuntu, not origen android, right?11:19
botao_sunsangwook: Because in origen android, it's /system/modules/ath6kl.ko11:20
sangwookbotao_sun: I think that is for both Android and ubuntu11:20
botao_sunsangwook: but I don't have this path "drivers/net/wireless/ath" on my board.11:21
sangwookbotao_sun: do you mean that you don't have ath6kl  directory ?11:21
sangwookbotao_sun: driver/net/wireles/ath/ath6kl ?11:22
sangwookbotao_sun: back in 10 minutes later.11:22
botao_sunsangwook: the path on board I created is "/lib/firmware/ath6k/AR6003/hw2.0".11:23
botao_sunsangwook: Totally no "driver" path on my origen android.11:23
deeptijamestunnicliffe: Hey James .. 11:25
deeptimabac: Hey Mattias11:25
mabacdeepti, hi deepti 11:26
deeptimabac: I am seeing issue with the hwpacks on snapshot.linaro.org .. http://paste.ubuntu.com/695524/11:26
deeptimabac: The hwpacks dated sep20 seems to be working 11:27
deeptimabac: but not yesterdays and today11:27
mabacdeepti, do they work with the same sd card that failed?11:27
mabacdeepti, the only time I've seen that is with defective sd cards11:27
deeptimabac: yes they work with the same SD card... and with the same nano image11:27
mabacdeepti, ok. which board are you using? I booted my Panda with 20110923 this morning11:28
deeptimabac: Panda board11:28
mabacdeepti, weird11:29
deeptimabac: that is strange11:29
mabacdeepti, yup. I used hwpack_linaro-panda_20110923-0_armel_supported.tar.gz with nano-n-tar-20110923-0.tar.gz11:29
mabacdeepti, I used the tip of l-i-t which is now the 2011.09 release11:30
deeptimabac: yes I pulled the latest l-i-t too11:30
mabacdeepti, uhm. I'll just try again and see what happens11:31
botao_sunsangwook: I find this on my source code: kernel/drivers/net/wireless/ath/ath6kl/init.c11:31
deeptimabac: ok.. meanwhile I will try the Sep 23 hwpack once more now11:31
botao_sunsangwook: You can write the full path directory, then I don't need to find and grep...11:32
sangwookbotao_sun: 11:40
sangwookbotao_sun: you have the right source code11:40
botao_sunsangwook: Hold on, I'm writing email to you.11:40
botao_sunsangwook: mail sent.11:40
james_wjanimo, fabo: it needs an FFe, if you wanted to push that through then it would speed things up11:42
mabacdeepti, I'm unable to reproduce the error, still boots fine (I did re-download the hwpack). Can you try it with a fresh sd card just to be sure11:46
deeptimabac: cool.. am trying .. not yet complete11:47
botao_sunsangwook: Got your email.11:47
mabacdeepti, ok let me know what happens11:47
mabachi james_w 11:47
deeptimabac: sure11:48
botao_sunsangwook:cfg80211.ko and ath6kl.ko has been wrote in the init.rc, so I have to launch a build again in order to change the sequence.11:48
botao_sunsangwook: Hi11:49
mabacjames_w, fabo has uploaded l-i-t to the ppa, just waiting for it to build. we're a bit after schedule since I first gave him svammel to package as linaro-image-tools. :)11:49
sangwookbotao_sun: OK11:49
botao_sunsangwook: Is there anything to change in the BoardConfig.mk? I ask because there is only 1 ko file mentioned there:11:49
botao_sunsangwook: WIFI_DRIVER_MODULE_NAME := "ath6kl.ko"11:49
botao_sunsangwook: Only this one in the BoardConfig.mk, no the other 2.11:50
botao_sunsangwook: what's  your opinion?11:50
sangwookbotao_sun: I am not clear Boardconfig.mk ,but how about making11:51
sangwookbotao_sun:  cfg80211.ko and ath.ko as a built-in11:51
sangwookbotao_sun: in kernel11:51
sangwookbotao_sun: if so, you can need only one .ko file 11:51
sangwookbotao_sun: which can be used in BoradConfig.mk11:51
james_wmabac, haha, great :-)11:51
botao_sunsangwook: They are already built in the kernel, that's why we can find their ko files on the board /system/modules path11:52
botao_sunsangwook: I will write an email to you.11:53
sangwookbotao_sun: understoond11:54
sangwookbotao_sun: with init.rc11:54
sangwookbotao_sun: fine11:54
mabacjames_w, fabo ok then, according to the knowledge base, either builds will start to fail within 10 minutes or we're ok. :) http://paste.ubuntu.com/695590/11:54
mabac(where do we get alerts about failed hwpack builds?)11:54
botao_sunsangwook: mail sent11:57
james_wmabac, you would have to subscribe to every single project in ubuntu-build.linaro.org unfortunately11:57
=== agreen is now known as agreen-away
james_wso my alert is usually fabo pinging me :-)11:58
james_win 10 minutes you can trigger a build and check that it completes11:58
=== agreen-away is now known as agreen
botao_sunsangwook: insmod sequence has been changed as your said, new build is on the way.12:00
mabacfabo alerts should work too. :) triggering a new build won't ruin the existing artifacts right? it will stay in <url>/0/ and we'll get a new one in /1/ ?12:01
mabacjames_w, ^12:01
james_wmabac, exactly12:01
mabacjames_w, Offspring seems to have a problem with me... "Sorry, you do not have permission to request project builds."12:04
james_wyou are logged in?12:05
mabacjames_w, never mind, I'm not12:05
mabaclet's try that12:05
james_wprobably shouldn't show the link at all in that case12:05
deeptimabac: surprisingly the same todays hwpack worked 12:05
sangwookbotao_sun: good12:06
mabacdeepti, cool, that's a nice suprise12:06
mabacjames_w, that worked better12:06
deeptimabac: indeed12:06
alf_janimo: done12:25
* ramana muses12:31
janimoericm|ubuntu, ping?12:42
=== revital_ is now known as revital
=== mayday_jay is now known as mayday_jay-work
mabacfabo, james_w all natty hwpacks that I have tried have built ok on offspring. the Panda hwpack booted on my board. Igep and Omap3 oneiric hwpacks fail to build since the hwpack format has been updated. I have two merges pending to fix those configs12:45
james_wmabac, great, thanks12:47
james_wmabac, are you free to do our call now?12:47
fabomabac: thanks. I tested l-i-t on natty/oneiric/maverick using only natty based image for panda12:47
mabacyeah sure. I'll just need to disconnect the boards, screen etc and go into another room. I'm in the office today.12:48
mabacjames_w, 5 minutes?12:48
james_wmabac, sure thing12:48
mabacjames_w, my mumble won't work with me. a quick reboot and I'll be back12:57
james_wmabac, do you hear me?13:02
mabacjames_w, no I don't13:02
mabaccan you hear me?13:03
mabacmumble keeps freezing on the audio settings wizard so there is a problem on my end13:04
james_wdo you have skype or similar?13:04
mabacI have it installed on my phone. we can try that13:04
james_wotherwise I can call a phone13:04
mabactry mattias.backman on skype?13:05
mabacI'm not on wifi so I have no idea about the quality13:06
ams_csrsandifo: remind me, why is it that the compiler does not use vmov to load DFmode constants, even 0.0?13:09
ramanaams_cs: That is a bad thing on the A9. 13:09
ams_csramana: oh, why is vmov worse that loading from a constant pool?13:09
ramanayou don't want the interleaving of VFP and neon code as the TRM suggests. 13:10
ams_csramana: so, vmov is interpreted as neon, even in vfp code?13:10
ramanawell that is a vmov.i32 that you are using right ? 13:11
ramanaThe instruction that we used was an Advanced SIMD only instruction. 13:11
ramanaand the TRM states that you shouldn't really interleave adv SIMD only instructions with VFP instructions. 13:12
ramanaIIRC . 13:12
ams_csso VFP has no fconstd anymore?13:12
mruiirc all vmov immediate are adv simd13:13
ramanaams_cs: vfp has fconstd . I don't think you can represent 0.0 in that :(  13:13
mrueven if float13:13
ramanaIIRC . 13:13
ams_csmru: the ARM has two encodings for vmov, and vmov.f64 says it is VFPv3. is that not so?13:15
ramanavmov.f64 can't be adv simd. 13:15
mruthe f64 is vfp, yes13:15
ams_cs*ARM = manual13:15
mrubut not f3213:15
mruscratch that13:16
mruthere's scalar .f32 and .f64 which are vfp13:16
mruvfpv3, not vfpv213:17
ams_cshmm, looks like .f32 and .f64 are the same encoding, but interpreted differently in different modes13:18
ramanafconstd is essentially vmov.f6413:18
ams_csno wait, I'm wrong, there's and 'sz' bit snuck in there13:18
mruneither .f32 nor .f64 can encode a zero though13:25
mrusee table A7-1813:25
ramanayes. that's why the idea was to use vmov.i64 0 or vmov.i32 0 to encode a 0.0 initialization. it would work on an A8 but it's not good on an A9 , plus the compiler has to be careful not to try and conditionalize it. 13:27
ramana(theoretically in T2 you could have conditional neon in IT blocks but just don't do it.) 13:28
mruhow does vmov from arm regs behave?13:31
ramanahmmm that might be interesting. 13:33
ramanahmmm mov rn, #0 ; vmov d1, r0, r0 ? 13:35
pm215t=t2 is UNPREDICTABLE13:36
ramanahmmm one more situation where GAS should warn then i suppose . 13:36
pm215er, whoops, only in the other direction. sorry13:36
ramanapm215 : was just about to ask that. 13:38
mabacrsalveti, any change you can have a look at https://code.launchpad.net/~mabac/linaro-images/hwpack.oneiric.linaro-omap3/+merge/76730 and https://code.launchpad.net/~mabac/linaro-images/hwpack.oneiric.linaro-igep/+merge/76574 ? if those changes can land, we might have a couple of working V2 hwpacks today.13:46
mabacrsalveti, I'm only bugging you since I'd like to see it work before clocking off for the week. which I hope to do soon. ;)13:46
plarsfabo: hi, can we move our sync today? the kernel ci call moved on top of it for this week13:51
faboplars: sure14:00
zygafinally back14:10
rsandifodoko: ping14:10
dokorsandifo, hi14:10
* zyga reorganized entire attic to accomodate for his primary workstation arriving from poland :)14:10
rsandifodoko: Hi.  Just seen your update to bug 641125.  Does that mean you were able to remove the chromium workaround and still get a good package?14:11
ubot2Launchpad bug 641125 in mysql-dfsg-5.1 "package mysql-server-5.1 5.1.37-1ubuntu5.4 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,Invalid] https://launchpad.net/bugs/64112514:11
dokorsandifo, not checked myself. will ask hrw again14:12
dokohrw, ^^^14:12
dokorsandifo, hrw, ahh, wait. still building on the buildds14:13
dokojust building for 24h14:13
rsandifohrw doko: Sorry, I meant bug 64112614:13
ubot2Launchpad bug 641126 in binutils-linaro "unresolvable R_ARM_THM_CALL relocation" [Medium,Confirmed] https://launchpad.net/bugs/64112614:13
dokoright, but chromium-browser is still building, see https://launchpad.net/ubuntu/+source/chromium-browser/13.0.782.215~r97094-0ubuntu2/+build/279998214:15
rsandifodoko: ah, ok, thanks14:15
deeptiplars: zyga It is around more than 2 hrs I dont see the job http://validation.linaro.org/lava-server/scheduler/job/1882 moving ahead. it is stuck at the l-m-c step14:16
deeptiplars: zyga can you please check why this is happening ?14:17
hrwrsandifo: last time when I built chromium-browser it was with binutils patched by doko - it succedded so I think latest package will also buil14:19
plarsasac: http://validation.linaro.org/lava-server/scheduler/job/187214:19
asacplars: http://validation.linaro.org/lava-server/scheduler/14:19
ericm|ubuntujanimo, pong14:20
hrwdoko: how fast is canonical powerpc machine? I would like to check gcc-linaro ppc failure one day but it can take hours14:21
mrulinaro ppc wtf?14:21
dokoG4 dual14:22
hrwmru: launchpad ppa are built on amd64/i386/armel/powerpc14:22
* mru has such a machine... not fast14:22
hrwmru: I do not have ppc at home14:22
mruexpand your collection14:22
suihkulokkione machine of every architecture supported by Linux kernel14:23
dokonext winter will be cold, get an alpha14:23
* mru has two alphas14:23
mruworkstation models14:24
mruboth in working order14:24
mrunot currently running14:24
mruthe sgi probably gives off more heat14:24
mruit is certainly noisier14:25
rsandifohrw: ok, great14:25
janimoI was wondering where I can get accelerated GLES drivers for imx5314:26
janimobut was just pointed at a link by pmcgowan_ a few minutes ago14:27
janimoI'd like to test a GLES2 app with hw acceleration14:27
mabacjames_w, fabo If you don't mind, I'll pack my gear and head home. I'll check my emails tonight for signs of progress or failure with the hwpack v2 builds on offspring.14:30
james_wmabac, sounds good.14:33
james_wthanks for your work on this14:33
mabacmy pleasure. :)14:35
=== agreen is now known as agreen-away
=== agreen-away is now known as agreen
hrwhave a nice rest of day14:42
dokorsandifo, hrw: chromium did build, should be available in the archive in one hour14:56
asacjcrigby: !!14:57
asacjcrigby: i wanted to enable you!!14:57
rsandifodoko: ok, thanks.  Sounds like the patch has had a fair bit of testing then, so I'll go ahead and submit it.14:57
asacjcrigby: my understanding was that you had a script you wanted to be able to run on ci.linaro.org for CI ubuntu kernel testing?14:57
jcrigbyasac, thanks but I am doing a kernel release right now and the script is still not complete so I would like to finish it before spending your time14:59
jcrigbyasac, can we revisit this monday?15:00
jcrigbyasac, the part that is missing is it does not do anything with results of the build.  I was planning on having it just push to a git tree.  Does that sound ok?15:01
jcrigbyor is there a better way15:01
asacjcrigby: ok. what are your requirements on jenkins? what we do with other trees is that they have ONE upstream tree that gets tracked. your script will always find that tree checked out in the working directory15:08
asacjcrigby: is that what you need? I guess in theory you want your build to be triggered by two trees? e.g. changes to your packaging tree and changes to the upstream tree?15:08
jcrigbyyes, I track two trees15:09
asacjcrigby: did you develop it with jenkins? or locally?15:09
jcrigbyasac, should I run jenkins locally and get it working and then go from there15:10
asacjcrigby: well, i can set you up a build that always pulls the latest bzr and then you can try live :)15:10
jcrigbyasac, I actually had originally an idea of a state tree that would have pointers to the two upstream trees15:10
asacjcrigby: whats a state tree?15:11
jcrigbyits just info about the last build so you know if you need to do anything.  Actually there are two kinds of info.  Config which would be git urls pointing to the upstream and packaged trees.  And state which is a particular commit id on each tree.  When the script runs if looks at head of each tree and if it is new/different then there is something to do.15:13
jcrigbyit then merges the trees and build them15:14
jcrigbythen it would push the merge result and the results of the build15:14
jcrigbyto a results tree or trees (one for the merged source tree, one for the results of the build)15:14
jcrigbyso as I said, there is configuration  and state15:15
jcrigbydoes any of this make sense?15:15
jcrigbyand since I'm a git head I was going to just do it all with git trees15:16
asacyeah makes sense. let me think15:17
jcrigbyactually the state could be in the results tree15:17
jcrigbyand the first time (results tree has nothing in it) then it just runs 15:18
jcrigbyasac, let me finish the script over the weekend and do some jenkins homework then we can revisit Monday.  Is that ok?15:19
jcrigbycan you point me at existing use cases?15:19
asacjcrigby: sounds good. 15:32
jcrigbyasac, one more thing15:32
rsalvetindec: yes, I also did some changes at the package, to be similar with the one for omap3 we already have at the archive15:33
rsalvetindec: let me know when you have something for the new driver, still waiting :-)15:33
jcrigbyhow do I login to ci.linaro.org?  And how do actually see the recipe or whatever is run for the various jobs?15:33
jcrigbyasac, and its ok to tell me to rtfm:)15:34
asacjcrigby: what happenes if you push login button?15:35
jcrigbyasac, duh that was easy15:35
asacjcrigby: can you see anything?15:35
jcrigbyI logged in with canonical credentials15:36
asacjcrigby: openid i think?15:36
asacjcrigby: do you have a NewJob thing at the top left?15:36
rsalvetifabo: will enable v2 and test with the hardwares I have here15:37
=== agreen is now known as agreen-away
jcrigbyasac, I have People, Build History, My Views   no NewJob15:37
=== agreen-away is now known as agreen
asacjcrigby: i hav eadded you to a team15:39
asacjcrigby: can you relogin and also submit that team (check box on openid page)15:39
faborsalveti: ok15:39
jcrigbyasac, ok I see New Job now15:40
asacjcrigby: right. so there you basically copy an existing copy15:40
asacjcrigby: copy 100%15:40
asacjcrigby: copy linux-linaro-3.0_panda-omap2plus15:41
asacthen change the git url to the base tree you want to track ( i think keep the linux linaro one)15:41
asacjcrigby: in the execute command box you can change the bzr branch and your script name15:42
asacjcrigby: i think the rest probably should be fine15:42
asacjust that you then for now only track changes from upstream tree ... remember to keep the working tree ina good shape15:42
asacjenkins will try to reuse it next time it builds15:42
jcrigbyso its not a fresh checkout each time?15:44
asacjcrigby: maybe linux-linaro-3.0_panda-ubuntu as name ? :)15:44
asacsomethig like that15:44
asacjcrigby: no15:44
asacjcrigby: at least not if ew dont want that15:44
asacalso remember that the ../ directory might be shared15:44
asacso try to not pollute that too much :)15:44
jcrigbyso I should put build results under my tree not ../something15:45
asacjcrigby: first time to clone a git it will take long15:45
asacjcrigby: well. if the veresion is unique enough15:45
asacand you remove only the produced files its fine15:45
asacwe had cases where we overrote files of a different job in a mid-air-collision15:46
jcrigbyso I can point jenkins at my ci script on git.linaro.org and the script itself can git clone the necessary other trees?15:48
jcrigbyand since it stays around thats good so subsequent git clones will be quick15:48
asacjcrigby: your bzr script will be cloned fresh everytime if yoou use the same execute command boxes15:49
asacso in theory you dont need to touch the job, just hit 'build now' button15:49
asacjcrigby: your ci script we assume is quick enough to clone15:49
jcrigbyok, I see the bzr branch in the window15:49
asacjcrigby: iwould propose to start with that branch15:50
asacjcrigby: and have goal to add your script there eventually15:50
asacso infrastructure can take over maintenance etc.15:50
jcrigbyI see15:50
jcrigbyok that makes sense15:50
jcrigbyI think you have "enabled" me:)15:50
jcrigbyasac, thanks, I understand a lot better now15:51
asaccool :-P15:52
asacjcrigby: so for now i think you best really just track upstream branch and ignore that your own branch changes dont retrigger a merge except when you manually hit build now15:52
asacjcrigby: later we can find a way to do that better :)15:53
=== nytowl is now known as nytowl_away
ramanahi does anyone know if we have windows binaries already available for the toolchain ? 15:58
ramanaI had someone come up and ask me today about it and I thought I'd just check. 15:58
mruwhat a weird thing to want :)15:59
asacjcrigby: one thing16:01
asacjcrigby: can you move that build away fromthe official CI Linaro 3.0 view on jenkins?16:01
asacjcrigby: you can put it on All CI for now16:01
asacand there is also a personal view in jenkins (but not sure how that works exactly)16:01
ramanaweirdness is in the eye of the beholder. 16:04
ramanauweigand: but did you put mfloat-abi=softfp ? 16:05
uweigandramana: let me try ...16:06
jcrigbyasac, sorry how do I move it to a different view?16:08
rsalvetifabo: asac: oh, and I just found out yesterday that launchpad is now importing git branches \o/16:09
rsalvetiso it'll make our ci a lot easier :-)16:09
uweigandramana: ok, with -mfloat-abi=softfp *and* -mfpu=neon the test passes even with an unmodified compiler16:09
uweigandramana: if I leave off either option, it fails16:09
faborsalveti: hehe great16:10
asacjcrigby: manage views ... somewhere on the left if you are looking at the job i think16:12
ramanauweigand: that's what I did see as well. 16:12
ramanayeah I think it doesn't trigger because with neon on the testcase has a different addressing form at that point but that doesn't mean that the bug isn't real. 16:13
asacjcrigby: seems you changed the git branch location?16:14
asacjcrigby: i think you could have kept that :) ... it was already the linux-linaro-3.0 upstream branch16:14
jcrigbyI didn't know I changed it16:14
asacjcrigby: anyway ... you will figure. seems that you might want to rove the emails fromthe configuration as well.... otherwise deepti gets spammed with your attempts16:14
jcrigbyI really just want to delete it and start over16:14
asacjcrigby: heh. yeah16:15
asacjcrigby: deleting is only possible for admins16:15
asacthe one should be fine16:15
jcrigbystill have not figured out how to move the view16:16
asacjcrigby: i removed the email 16:16
asaclet me see for moving16:16
asacjcrigby: ah ... go to the view and hit edit view16:16
asacso here: https://ci.linaro.org/jenkins/16:16
asacthen enable it here: https://ci.linaro.org/jenkins/view/All%20CI/16:17
janimojcrigby, how to avoid building the tools for the kernel while cross compiling. I  see do_tools is set to false in the scripts, but fdr binary tries to install a build/tools-/tools/perf/perf file ?16:18
jcrigbyjanimo, I probably don't run fdr binary so I have never run into that16:20
jcrigbyasac when I go to the tab that has my build in it there is no edit views16:21
jcrigbyasac, when I go to other tabs there is a delete view but not in the tab I need to change16:21
=== agreen is now known as agreen-away
=== agreen-away is now known as agreen
rsalvetiagreen: hey, are you also updating your tilt-linux-linaro-3.0 branch today?16:24
rsalvetifabo: guess we also need a tag from the TILT kernel16:24
agreenrsalveti: yes on it now16:24
ramanauweigand: oh this thing is so broken. 16:25
agreenthanks for the cool patches btw16:25
rsalvetiagreen: latest 4460 patches at http://git.linaro.org/gitweb?p=people/rsalveti/linux-linaro-3.0.git;a=shortlog;h=refs/heads/tilt-linux-linaro-3.0-rsalveti16:25
agreenyeah I have them16:25
ramanauweigand : It's not a problem with your patch but I think there's another problem here. 16:25
uweigandramana: hmm, ok ... that's well possible ...16:25
uweigandramana: what else you do think is broken?16:26
rsalvetiagreen: both "DRM: OMAP: Fixing PVR kernel module name" and "Staging: OMAPDRM: Changing driver's name to be compatible with userspace" should be just part of the packaging patches16:26
rsalvetiagreen: the others you can just apply on top of your tree16:26
ramanauweigand: We do if (TARGET_ARM) arm_legitimize_reload_address () else thumb_legitimize_reload_address () and we handle a number of things in arm_legitimize_reload_address for TARGET_THUMB2 16:26
rsalvetiI need this drm patches to make it work with the currently public SGX binaries 16:26
rsalvetiit expects the kernel module to be called as omap_gpu instead of omapdrm16:26
agreenrsalveti: that output tree is actually generated from another tree, I put them there16:27
ramanauweigand: your patch doesn't make anything worse than the way it is today. 16:27
jcrigbyasac, can you just delete it for me and I will create a new one somewhere else16:27
agreenyes I saw16:27
agreenyou found quite a few tricky things16:27
uweigandramana:   if (!TARGET_ARM)16:27
uweigand    {16:27
uweigand      /* TODO: legitimize_address for Thumb2.  */16:27
uweigand      if (TARGET_THUMB2)16:27
uweigand        return x;16:27
uweigandramana: it would appear for THUMB2 there's just no optimization at all ?16:28
rsalvetiagreen: cool, let me know when you have your branch ready and I'll test it at my 4460 board for you16:28
agreenperfect will do16:28
uweigandramana: oops, sorry, wrong function :-/16:29
ramanauweigand: yes. 16:29
ramanauweigand: I was referring to the definition of LEGITIMIZE_RELOAD_ADDRESS in arm.h 16:29
ramanauweigand: though what you point out is also something we need to get right. 16:30
uweigandramana: so the TARGET_THUMB2 stuff in arm_legitimize_reload_address seems to be dead ...16:31
ramanauweigand: yes. 16:31
ramanauweigand: It seems to have been put in there in a well intentioned form to future proof the last time someone was there but it's never been tested .16:32
uweigandramana: huh, comment says: /* ??? This should be safe for thumb2, but we may be able to do better.  */16:32
rsalvetiagreen: most annoying, led shutting down the board and the gpio1 reset thing 16:33
rsalvetiwe set the gpio1 on u-boot now, and then ask the kernel to not reset it16:33
rsalvetias it'd bring the voltage down16:34
ramanauweigand: it does but why handle TARGET_THUMB2 in arm_legitimize_reload_address. 16:34
rsalvetibut it's annoying as now we one specific bootloader to make the kernel work properly 16:34
rsalveti*we have16:34
rsalvetifine for the ubuntu leb thing, but not good for upstream16:34
uweigandramana: seems someone got confused16:35
rsalvetiguess aneesh is on top of that for upstream, as he was the one proposing the first version of the patch16:35
asacjcrigby: just dont bother16:35
asaci will move it16:35
ramanauweigand: yes. 16:36
* ramana is off for the day 17:38
pfefferz_rsalveti, did you have any issues with __linux__18:03
pfefferz_in your kernel tree?18:03
pfefferz_rsalveti, ping18:19
pfefferz_jcrigby, rsalveti I'm trying to rebuild u-boot with the patches rsalveti listed but I'm getting a conflict on 0002 git://git.linaro.org/boot/u-boot-linaro-stable18:47
jcrigbypfefferz_, the patches are probably already there18:47
jcrigbylet me know if anything is broken18:47
jcrigbyother than this that is18:48
pfefferz_patch 0001 went in on master okay18:49
jcrigbywhich patch was that?18:50
pfefferz_these patches to enable 446018:50
pfefferz_> U-Boot patches (on top of 2011.09-rc2):18:50
pfefferz_> http://people.linaro.org/~rsalveti/4460/u-boot/18:50
ibirisJesseBarker: ping - alf_ tells me there was something you wanted to clarify about blueprint names?18:52
jcrigbylooks like I am missing that one18:53
jcrigbypfefferz_, late last night he said that my tree worked for him18:54
pfefferz_giving it a shot now with my Android kernel18:54
jcrigbyok, looks like http://people.linaro.org/~rsalveti/4460/u-boot/ is for vanilla upstream like you said18:55
jcrigbyit has a couple of my fixes as well18:55
pfefferz_okay cool, thanks for checking18:56
pfefferz_jcrigby, hmmm19:04
pfefferz_somethings not working for me19:04
pfefferz_looks like SGX isn't up19:04
pfefferz_I'll try the tip u-boot19:04
jcrigbypfefferz_, ok checked again and rsalveti has 5 patches but only 4 are in my -stable  that first patch that applied cleanly is the missing one19:08
jcrigbypfefferz_, but not sure if it is the problem19:09
pfefferz_I see 7 patches19:09
jcrigbysorry, I was not counting the usb pin config and clock patches, those came from me19:10
pfefferz_okay I'll try with the first patch19:10
pfefferz_whats the config again?19:14
=== agreen is now known as agreen-away
=== agreen-away is now known as agreen
rsalvetipfefferz_: ping19:38
rsalvetiops, pong :-)19:38
rsalvetipfefferz_: jcrigby: yes, that patchset is on top of upstream19:38
rsalvetibecause there was still one extra patch for next19:38
rsalvetithat one with the gpio1 fix19:39
rsalvetibut if you included that already, the next tree should work fine19:39
rsalvetipfefferz_: which kernel tree are you using?19:39
pfefferz_rsalveti, sending an email...19:40
jcrigbyI have not yet done a u-boot package upload, should that patch be added before that19:40
jcrigbyrsalveti, ^19:40
rsalvetiagreen: will try your tree now19:41
rsalvetijcrigby: yes, please19:41
rsalvetijcrigby: that one is useful to make it work all the time with 446019:41
rsalvetijcrigby: guess it was the one you thought it was for kernel19:42
agreenrsalveti ok I am going to bed tho, good luck19:42
rsalvetiagreen: cool, thanks, will reply your email once I'm done testing19:42
jcrigbyrsalveti, no I have that one19:42
jcrigbythe one I don't have is19:42
jcrigby[PATCH 1/7] omap4460: tps initialization changes19:42
=== agreen is now known as agreen-away
rsalvetijcrigby: oh, forgot to point you that19:43
rsalvetisorry, that was the one aneesh sent by email19:43
jcrigbyok, fine no problem19:43
jcrigbyI remember now, I just looked on mailing list for patches you had sent19:43
jcrigbyso this one was missed since it was from aneesh19:44
rsalvetibut I had to change it to apply on top of upstream19:44
rsalvetiforgot to point you the changed on later19:45
rsalvetibut that should be it19:45
jcrigbyok, I'll go do that now since I promised dzin the u-boot would go to ppa today19:45
jcrigbystill working on debug in kernel19:45
jcrigbywill do that next19:46
jcrigbytilt will probably not happen until late or tomorrow19:46
rsalvetijcrigby: sure, push it to the tree and make the tarball later today19:46
rsalvetijcrigby: sure, I'm testing tilt now19:46
rsalvetiso once I'm done testing we just need to package it19:47
rsalvetiand should be fine19:47
jcrigbyI actually already did the u-boot tar so I will update it now19:47
rsalvetioh, ok19:47
jcrigbybut first I need some lunch...19:47
rsalvetisure :-)19:47
rsalvetijust got back from mine19:47
pfefferz_rsalveti, okay, mega email sent19:55
pfefferz_mpoirier, rebooting19:56
rsalvetijcrigby: guess we can try creating the build using the recipe thing already20:48
rsalvetiand if it works we just copy the resulted package to the overlay20:49
rsalvetilet me check if the stable-trunk bzr got updated properly20:49
jcrigbyyes, I could do that20:49
prpplagueanyone know an irc channel related to rtc drivers?20:50
cjbprpplague: there wouldn't be any reason to have a channel that specific.20:50
rsalvetijcrigby: requested another sync at stable-trunk20:51
prpplaguecjb: yea i figured it wouldn't be specific for that, just wondering if there was one that would be general enough and have some of the folks who are on the rtc mailing list20:51
rsalvetijcrigby: just push a new changelog entry at the u-boot-linaro bzr branch, saying it's based on 2011.0920:51
rsalvetiand then write a recipe for it, quite similar with the one I created for u-boot upstream20:52
jcrigbyrsalveti, that reminds me shoule u-boot-linaro stay insync with ubuntu so should I do ppa only stuff on a branch?20:53
rsalvetijcrigby: hm, could be20:53
rsalvetiguess it'd make sense to have a branch just for our ppa/overlay20:53
rsalvetiand another one for ubuntu20:53
rsalvetiso we can track the history for both20:54
rsalvetijcrigby: when creating the new one, just remember removing debian/source/format20:54
rsalvetiseems daily ppa build doesn't handle format 3 well20:54
jcrigbyoh I wondered what that was about20:55
rsalvetiformat 3 requires you to have the orig.tar.gz20:55
jcrigbydid you also remove the original source rule?20:55
rsalvetidaily ppa build will basically create the source package from the merged bzr branch20:55
jcrigbyfrom rules20:55
rsalvetijcrigby: I just removed because I'm not using it at the u-boot bzr branch20:56
rsalvetias it's just used for the daily stuff20:56
jcrigbyright, so it didn't break anything just not needed20:56
rsalvetijcrigby: yes20:56
rsalvetijcrigby: if you just clone your current u-boot-linaro bzr branch and remove the source/format it should be enough already20:57
rsalvetithen just increase the changelog entry 20:57
rsalvetisaying it's based on 2011.0920:57
jcrigbyso I just bzr branch, then what url do I use for the push?20:58
jcrigbyI'm no bzr guru20:58
jcrigbyand the ~jcrigby stuff confuses me20:58
rsalvetijcrigby: http://paste.ubuntu.com/695833/20:58
rsalvetijcrigby: this should do it for the recipe20:58
rsalvetijcrigby: at bzr you can store a branch at your own account or at a group20:59
rsalvetias you're part of linaro-foundations, you could also push it to lp:~linaro-foundations/u-boot-linaro/branch20:59
rsalvetifor example20:59
rsalvetijust to say the owner of the branch20:59
rsalvetithe second part is the project 20:59
rsalvetiand the third the name of your branch21:00
jcrigbyI'll give it a go21:00
rsalvetijcrigby: great, if it works fine then you don't need to push anything to the overlay :-)21:00
rsalvetiwe just copy it from staging21:01
pfefferzrsalveti, is that last patch going to be pulled over?21:40
rsalvetipfefferz: it was applied already21:40
rsalvetiat least last time I checked21:41
pfefferzjcrigby, said it wasn't so I was asking21:41
jcrigbypfefferz, that was hours ago!21:42
pfefferzANdroid 's looking good with the latest TI LT kernel21:42
jcrigbyswell, 21:43
jcrigbyanything noticably better or are we just happy to have it running on this bleeding edge hw21:43
jcrigbyrsalveti, ok I have a branch21:44
pfefferzjcrigby, things don't seem that much better21:44
pfefferzI don't think we're at 1.5 GHz21:44
pfefferzomapzoom seemed a little faster21:44
pfefferzbut of course it crashed every second21:45
rsalvetijcrigby: cool, should be enough it seems21:45
rsalvetijcrigby: just create the recipe and fire up a build :-)21:45
jcrigbyso I just sorta copy your recipe?21:45
rsalvetipfefferz: this tree is still lacking a bunch of 4460 related patches21:45
rsalvetibut they are all broken21:45
pfefferzthat's okay21:46
rsalvetijcrigby: http://paste.ubuntu.com/695833/21:46
pfefferzits up and working21:46
rsalvetipfefferz: what was the sgx thing?21:46
rsalvetipfefferz: and it's expected to work the same way at 443021:46
rsalvetisame image I mean21:46
pfefferzno idea, I think it was fixed by a branch sebjan21:46
rsalvetiok, great21:47
pfefferz" It has latest stuff from Ricardo and sebjan."21:47
pfefferzthat's what I'm running21:47
rsalveticool, just tested it here and it's also working fine21:47
dzinanyone have some spare cycles to help me with an Ubuntu problem?21:48
rsalvetidzin: sure, just ask :-)21:48
rsalvetipfefferz can also help, he's an ubuntu expert ;-)21:49
pfefferzI am21:49
dzinI had to replace my vid card in my desktop, now video is completely messed up.21:49
dzinneed to boot to generic vid drivers21:49
rsalvetidzin: which card are you using now?21:49
dzinasus hd6450 (amd/ati)21:49
rsalvetihm, ati drivers kind of sucks 21:50
rsalvetiwell, I'm using the open source one for my board and it's working nicely21:50
rsalvetithe closed one is the one that sucks21:50
rsalvetiimproves the gl performance, but then I get a lot of regressions21:50
dzinit's not even usable at this point21:51
rsalvetiproblems with dual, kms and such21:51
dzinI cannot even log in since I cannot see anything.21:51
rsalvetidzin: can you paste me the output of your lspci?21:51
rsalvetidzin: paste.ubuntu.com21:51
dzinI don't know what that is21:52
rsalvetijust run it at your terminal, should dump the info about your boards21:52
dzinI can't get to a terminal21:52
dzinas soon as it boots, vid gets messed.21:52
rsalvetiouch, you may need to remove the drivers then21:53
rsalvetitry pressing control + alt + f2 21:53
rsalvetiand see if you're able to use the console session21:53
dzindoesn't work, tried that21:53
rsalveti:-( then booting at recover mode should help21:54
jcrigbyrsalveti, builds pending21:54
dzinI can get to grub edit, but I don't know how to proceed from there,21:54
dzinnope, recover not working either21:54
jcrigbyrsalveti, being called to dinner by family, bbl21:54
rsalvetijcrigby: great, saw it here, thanks21:55
jcrigbyrsalveti, thanks for the help21:55
rsalvetihave a nice dinner 21:55
dzintried to replace "quiet splash" with "text" but still no go21:56
dzinI even nuked x11 conf but that didn't help21:56
rsalvetidzin: guess the problem is that the broken module is loaded automatically 21:57
rsalvetidzin: can you try a live cd or similar?21:57
rsalvetithen you can chroot at your partition and remove the package by hand21:57
dzinhow can I boot to a command line so it does not load graphic mode?21:58
rsalvetitext was the one21:58
rsalvetibut I believe the driver is loaded by default21:58
rsalvetias it could be related with the kernel21:58
rsalvetimaybe using single?21:59
rsalvetior init=/bin/bash21:59
dzinI used the init line and that worked. What do I remove after I'm there?21:59
rsalvetidzin: do you remember installing any specific driver for your board?22:00
dzinno but it was an nvidia card22:01
rsalvetifglrx should be the one for ati22:01
rsalvetithen check if you have any nvidia related package installed22:01
rsalvetidpkg -l | grep nvidia for example22:02
rsalvetidzin: but would still be good to check the specific version of your current board22:02
rsalvetican you run lspci and paste the id of your ati card?22:02
rsalveti01:00.0 VGA compatible controller: ATI Technologies Inc Mobility Radeon HD 3400 Series22:03
rsalvetithis is mine22:03
dzinI can type it22:03
rsalvetijust to check if the kernel support is there already 22:03
rsalvetifor things like kms and such22:03
dzinATI Tech, NI Caicos [AMD RADEON HD6450]22:03
rsalvetidzin: are you using natty or oneiric?22:05
dzinif I nuke xorg.conf in /etc/X11 what will happen?22:06
rsalvetiin theory you shouldn't need any xorg.conf at /etc/X1122:07
dzinoh, there's not much in there anyway22:07
rsalvetidzin: could be that this was installed by the nvidia driver22:07
dzinwhat is the cmdline to get  the an ati or generic vid driver?22:08
rsalvetidzin: in theory if you remove your xorg.conf file, and the nvidia driver you have installed, you should not need anything to use the generic vid driver22:09
rsalvetixserver-xorg-video-radeon is the open source driver for ati boards22:10
rsalvetibut I'm looking if it supports your board already22:10
dzinI still need to remove the nvidia driver, how do I do that?22:13
rsalvetidzin: find the package you installed, dpkg -l | grep nvidia should point you some22:16
dzingot it, removing now22:16
rsalvetidzin: can you run lspci -nn and get me the id from your board?22:16
rsalveti01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Mobility Radeon HD 3400 Series [1002:95c4]22:16
rsalvetithis is from my board22:16
rsalvetithe id is [1002:95c4]22:17
dzinI've removed the driver and tried 'reboot' and 'halt' but shutdown reports unable to shutdown system22:18
dzinI want to sync my changes and reboot22:18
rsalvetidzin: weird22:19
rsalvetitype sync and hard reboot22:19
rsalvetihm, the open source driver doesn't seems to support your card22:20
rsalvetiso you'd need to install the fglrx proprietary driver22:21
dzinok, thanks.22:21
dzindidn't work anyway22:22
dzinI still have the messed up screen22:22
dzingot to go, but thanks for your help. I'll probably have to reinstall.22:23
rsalvetitry first with a live cd22:23
rsalvetigive it a try and try to install the proprietary driver22:23
rsalvetiwould recommend ubuntu oneiric for you, as your board is quite new22:23
rsalvetibut it's up to you :-)22:23
kirklanddoanac: pfefferz: fyi, I have a table outside, in the corner22:38
pfefferzcool, omw22:39
rsalvetijcrigby: http://git.linaro.org/gitweb?p=people/rsalveti/linux-linaro-3.0.git;a=shortlog;h=refs/heads/tilt-linux-linaro-3.0-packaged23:09
rsalvetijcrigby: this is the current tilt-3.0 tree from andy + your sauce from previous tilt package23:09
rsalvetiworked fine for both 4430 and 446023:09
jcrigbyrsalveti, so nothing to do it just works!23:15
rsalvetijcrigby: :-)23:15
rsalvetijcrigby: just don't know how you'll rebase your tree23:16
rsalvetiwhich sauce patches to use and such23:16
rsalvetibut at least the original tree itself works fine :-)23:16
jcrigbyrsalveti: so you built that with packaging? so the config in debian.linaro is good?23:21
rsalvetijcrigby: I updated the config23:21
rsalvetijcrigby: I basically did format-patch to grab all the sauce from your current lt tree23:22
rsalvetithen applied them on top of the tilt-3.0 tree23:22
rsalvetiupdated the config, reverted 1 patch and got it working with packaging :-)23:22
jcrigbyok great you did all the hard part, I should be able to finish23:23
rsalvetihttp://git.linaro.org/gitweb?p=people/rsalveti/linux-linaro-3.0.git;a=commitdiff;h=09c51c625edc4128e265f87cfa037bd2a91562d3;hp=5c32c965bfab2b6ac45a044fd87dd18ad5db6dc3 is the config sync23:23
jcrigbythats what I wanted the working config23:24
rsalvetimay just need to add that config aviksil asked for, and enable the debug package23:24

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