Linaro-android

linaro-android@lists.linaro.org

1073 discussions

Default Java version on android-build and imminent Jenkins upgrade

by Paul Sokolovsky

Hello, The latest Jenkins LTS release, 1.625, requires Java7 to run (to give some context, Java9 is in beta now, Java7 is many years old). As learned with some pain on ci.linaro.org, which already upgraded, this requirement applies not just to a Jenkins master, but also to build slaves, just to run slave agent to be connected to a master. We install both Java6 and Java7 on android-build.linaro.org slaves, but the default is still Java6, "for compatibility with older Android versions", as a comment says. Is Java6 being a default still relevant? The easiest way to accommodate the upcoming upgrade is just to switch default to Java7. If there're issues expected with that, we can follow an alternative way of adding adhoc version overrides for slave startup. -- Best Regards, Paul Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog

10 years, 10 months

Re: The static linked 32bit binary size is different between flounder and flo

by John Stultz

On Sun, Nov 1, 2015 at 10:07 PM, YongQin Liu <yongqin.liu(a)linaro.org> wrote: > On 2 November 2015 at 11:59, John Stultz <john.stultz(a)linaro.org> wrote: >> >> You can try to objdump -d the files and diff the output.. That might >> help narrow down where things are different. > > Sorry, I can not find out anything from the output. > > Not able to read the objdump output:( > > Attached the binaries and objdump output files here, in case if someone > likes to check the binary file. Sorry. Yea, objdump isn't super useful after the symbols have been stripped out. Might try to build w/ symbols and try again? thanks -john

10 years, 10 months

Nexus 5X preordering has started...

by Bernhard Rosenkränzer

They're now accepting pre-orders for the Nexus 5X, saying it'll leave their warehouse on November 9 -- so if the past is any indication, we'll see 6.0.0_r2 or 6.0.1_r1 next week to match what's actually on there. Still no sign of the 6P... ttyl bero

10 years, 10 months

LTP Android

by $rik＠nth

Hello Linaro, Does the LTP ported specific to Android by fixing the tests back to use bionic libc or directly cross compiled using tool chain and using them on Android. I am getting below issue while i am running LTP. But nobody user is present under /etc/passwd. Can anyone throw some pointers to fix it or from where i can get the patches available for Android platform. Line 23: access02 1 TBROK : access02.c:125: getpwnam failed: errno=ENOENT(2): No such file or directory Line 33: access04 1 TBROK : access04.c:96: getpwnam failed: errno=ENOENT(2): No such file or directory Line 35: access05 1 TBROK : safe_macros.c:122: access05.c:130: getpwnam(nobody) failed: errno=ENOENT(2): No such file or directory Line 89: capset02 4 TBROK : capset02.c:252: getpwnam failed: errno=ENOENT(2): No such file or directory Line 102: chdir03 1 TBROK : safe_macros.c:122: chdir03.c:168: getpwnam(nobody) failed: errno=ENOENT(2): No such file or directory Line 126: chmod03 1 TBROK : chmod03.c:150: getpwnam failed: errno=ENOENT(2): No such file or directory Line 131: chmod05 1 TBROK : chmod05.c:187: getpwnam("nobody") failed: errno=ENOENT(2): No such file or directory Line 133: chmod06 1 TBROK : safe_macros.c:122: chmod06.c:174: getpwnam(nobody) failed: errno=ENOENT(2): No such file or directory -- Thanks & Regards, M.Srikanth Kumar.

10 years, 10 months

panda-linaro-14.09-release build error

by Anthony Fu

hi, I try to build the android ~linaro-android/panda-linaro-14.09-release by using the script file ( linaro_android_build_cmds.sh) that downloaded from the web page <http://snapshots.linaro.org/android/%7Elinaro-android/panda-linaro-14.09-re…> I got the error message below: Fetching project platform/packages/screensavers/Basic remote: Counting objects: 1, done. remote: Total 1 (delta 0), reused 0 (delta 0) >From git://android.git.linaro.org/platform/packages/screensavers/Basic * [new tag] android-5.1.1_r14 -> android-5.1.1_r14 Fetching projects: 100% (421/421), done. Syncing work tree: 32% (135/420) fatal: failed to unpack tree object HEAD Traceback (most recent call last): File "/home/anthony/pandaboard/linaro-14.09-release/android/.repo/repo/main.py", line 526, in <module> _Main(sys.argv[1:]) File "/home/anthony/pandaboard/linaro-14.09-release/android/.repo/repo/main.py", line 502, in _Main result = repo._Run(argv) or 0 File "/home/anthony/pandaboard/linaro-14.09-release/android/.repo/repo/main.py", line 175, in _Run result = cmd.Execute(copts, cargs) File "/home/anthony/pandaboard/linaro-14.09-release/android/.repo/repo/subcmds/sync.py", line 754, in Execute project.Sync_LocalHalf(syncbuf, force_sync=opt.force_sync) File "/home/anthony/pandaboard/linaro-14.09-release/android/.repo/repo/project.py", line 1246, in Sync_LocalHalf self._InitWorkTree(force_sync=force_sync) File "/home/anthony/pandaboard/linaro-14.09-release/android/.repo/repo/project.py", line 2383, in _InitWorkTree raise GitError("cannot initialize work tree") error.GitError: cannot initialize work tree can somebody help to fix it? thanks, Anthony -- Best Regards, Anthony Fu 傅錫泓 TP(M):+886-928-801-381 CN(M):+86-156-2655-9419

10 years, 12 months

Re: [ANN] Round 2 (and 3) of Gerrit upgrade

by Paul Sokolovsky

Hello, https://dev-private-review.linaro.org/ has been upgraded to Gerrit 2.10.6. Upgrade went smooth, the only issue is that handful open reviews there are now in "Merge Conflict" state, as discussed below. However, I tried to merge a test review with such status and it went OK. If that won't work, a change need to be rebased and re-pushed from command line. To remind, the biggest change in 2.10.x is a new change summary screen. Every user will be notified about it via popup on first access, linking to detailed documentation: https://dev-private-review.linaro.org/Documentation/user-review-ui.html and offering choice to switch back to classic screen (support for which is dropped in 2.11, so use your judgement when you want to learn the new screen - now or later). Please let me know of any issues seen. If nothing big pops up, we'll finish 2.10.6 migration next weekend, upgrading https://review.linaro.org and https://lhg-review.linaro.org as discussed below. Thanks, Paul On Fri, 21 Aug 2015 16:25:58 +0300 Paul Sokolovsky <Paul.Sokolovsky(a)linaro.org> wrote: > Hello, > > As was announced previously, Linaro Systems team is working to upgrade > Gerrit version used on our hosts from 1-year old 2.8 to recent and > supported 2.10. Two weeks ago, we upgraded > https://android-review.linaro.org as a pilot. The upgrade went largely > OK, though as full disclosure, following issues were faced: > > 1. Upgrade uncovered issues with duplicate accounts. This issue is > mostly specific to android-review.linaro.org - it's the oldest Gerrit > system in Linaro which accumulated number of accounts from different > authentication services we used as well as community accounts. Other > systems are unlikely to be affected at all, and even on > android-review.linaro.org only few active users were affected and > issues were resolved proactively. > > 2. 2.10 exposed an AJAX caching issues we experienced intermittently > before - just to allow to nail them down and resolve consistently for > all servers. So, this is off the list. > > 3. The "biggest" issue we saw is that after the upgrade, all pending > open changes in Gerrit were changed to "Merge Conflict" state, > which was not resolvable from UI, with Gerrit suggesting to rebase and > re-push change from command line. Having done that, a reviewed worked > without a problem. We even received a report that this issue may be > related to the new review UI, switching to old screen allowed to > rebase a change via UI button. > > > With this in mind, we think we're ready for the next round of upgrade. > Based on previous discussions, this would be > https://dev-private-review.linaro.org , slated to upgrade next > weekend. We'd like to confirm that this plan works well for them. > > Otherwise, we'd plan to finish Gerrit upgrade (and do any needed > follow-up tweaks) before Connect, so there was productive work there > (and we can consult people on new UI/features they may want to use). > So, if everything goes smooth with dev-private-review.linaro.org, a > weekend after next (Sep, 5) we'd plan to upgrade 2 remaining systems: > https://review.linaro.org and https://lhg-review.linaro.org . Again, > we'd like to be sure that their stakeholders are OK with this. > > > > Thanks, > Paul > > Linaro.org | Open source software for ARM SoCs > Follow Linaro: http://www.facebook.com/pages/Linaro > http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog -- Best Regards, Paul Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog

11 years

Could Android source code build on ARM64 boards?

by Xinliang Liu

I am wonder is that possible to build android source code on ARM64 boards? Currently, it seems the android build system doesn't support arm64 arch as a host. If we want to support this native build, then what are the work should be do? Do you have ever try this? This is import that if we want to deploy ARM64 boards in the build cloud center. Usually, the build cloud should support Android build.

11 years

Using ADB boot options for Android Devices

by Sandeep Chawla

Hello Linaro Android Champions, Thanks for your efforts in making Lava a great automation platform for Android. I use Nexus10 fastboot driver on my devices, which flashes the boot image on the device. This leads to a problem that whenever the dispatcher issues 'adb reboot' command, the device reboot instead of being the fastboot mode as with the regular fastboot driver. Is there an option in the device.conf file to specify a different boot command ? If we could change the command from 'adb reboot' to 'adb reboot-bootloader' , the dispatcher would be able to work fine. Please let me know if you have any further questions. Thanks Sandeep

11 years

Duplicate user accounts on android-review.linaro.org

by Paul Sokolovsky

Hello, We recently upgraded Gerrit on android-review.linaro.org (former review.android.git.linaro.org) to 2.10.6, and it uncovered issue with duplicate accounts (not sure how it happened, there were some account table migrations, maybe it exposed inconsistencies which were latent before). The full list of dup accounts is below. Some folks are no longer with Linaro, other likely not use android-review.linaro.org actively, other are community accounts. And there's no easy way to see which of dup accounts is the right one. So, this mail is cc:ed to few folks I know may use this server, and if you see yourself in the list below, please respond too. So, what we need from you is: 1. Log in to https://android-review.linaro.org 2. Check that your group memberships look as expected at https://android-review.linaro.org/#/settings/group-memberships 3. Look up your numeric account ID at https://android-review.linaro.org/#/settings/ 4. Reply to this mail with the info above (whether your groups are ok, and account ID). Please let some time for resolution (I'm on vacation next week), if you're blocked by dup account issue, please ping Andy Doan or Matt Hart. Thanks, Paul full_name | preferred_email | count ----------------------+--------------------------------------+------ Tony Mansson | tony.mansson(a)linaro.org | 2 Omar Ramirez Luna | omar.luna(a)linaro.org | 2 tinoj antony | tinojantony(a)gmail.com | 2 Georgy Redkozubov | Georgy.Redkozubov(a)linaro.org | 2 Alan Bennett | alan.bennett(a)linaro.org | 2 Paul Beeler | sparksco(a)gmail.com | 2 Yusuf Burmawala | burmawala.yusuf(a)gmail.com | 2 Fahad Kunnathadi | fahad.k(a)linaro.org | 2 Ramesh Chandrasekaran | ramesh.chandrasekaran(a)stericsson.com | 2 Hongbo Zhang | hongbo.zhang(a)linaro.org | 2 Harigopal Gollamudi | harigopal.gollamudi(a)linaro.org | 2 Soumya Basak | soumya.basak(a)linaro.org | 2 Naresh Kamboju | naresh.kamboju(a)linaro.org | 2 Tyler Baker | tyler.baker(a)linaro.org | 2 Angus Ainslie | angus.ainslie(a)linaro.org | 2 Botao Sun | botao.sun(a)linaro.org | 2 Stefan Ekenberg | stefan.ekenberg(a)linaro.org | 2 Anil Kumar Koli | anil.koli.kumar(a)gmail.com | 2 ??Abhishek Paliwal | abhishek.23dec(a)gmail.com | 2 Romain Perier | romain.perier(a)gmail.com | 2 Rajagopal Venkat | rajagopal.venkat(a)linaro.org | 2 David Zinman | david.zinman(a)linaro.org | 2 Amit Khare | amit.khare(a)linaro.org | 2 Bintian Wang | bintian.wang(a)linaro.org | 3 Mathieu Poirier | mathieu.poirier(a)linaro.org | 2 Milosz Wasilewski | milosz.wasilewski(a)linaro.org | 2 -- Best Regards, Paul Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog

11 years

Gerrit 2.10.6 upgrade: All changes are stuck in "Cannot Merge due to Path Conflict" state

by Paul Sokolovsky

Hello, It was just just brought to my attention by Yongqin that after upgrade, all open changes appear to be stuck at "Merge Conflict" state, which is easily visible in the summary table: https://android-review.linaro.org/#/q/status:open Checking any single change, e.g. https://android-review.linaro.org/#/c/15987/ it says "Cannot Merge", and tooltip says "due to a path conflict". There's no "Rebase" button, tooltip says that rebasing should happen locally, and new version uploaded. Yongqin tried that with https://android-review.linaro.org/#/c/15996/ , and it cleared state and change was merged successfully after that. So, there's nothing fatal, but of course it's quite a chore to rebase all patches manually. My googling for what could be the cause of such problem after an upgrade, unfortunately didn't turn up much so far. But there're a lot of older (circa 2010) discussions of Gerrit behavior when harmless, not really conflicting concurrent commits to a repo threw changes into such state. They all relate to the times when Gerrit didn't support automatic conflict resolution or it was conservatively disabled. But on android-review.linaro.org, we exactly have those conservative "off" settings of circa-2.4 times. So, my hypothesis is that's what played role in such situation with upgrade, and i propose to have more process-friendly settings. Specifically, currently we have: Submit Type: Merge if necessary Automatically resolve conflicts: FALSE Require Change-Id in commit message: FALSE I propose to change it to: Submit Type: Rebase if necessary Automatically resolve conflicts: TRUE Require Change-Id in commit message: TRUE Let me know if there're any concerns. Not that these are "avoid problems in the future" changes, current conflicting changes will need to be re-uploaded by their authors manually. -- Best Regards, Paul Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog

11 years

← Newer
1
2
3
4
5
6
7
8
...
108
Older →

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

Linaro-android