|
Minutes: Yocto Technical Team (Tuesday, August 02, 2011 8:00 AM-9:00 AM (UTC-08:00) Pacific Time (US & Canada))
Liu, Song <song.liu@...>
Messages to the team:
 * Thanks and kudos to the team, Yocto 1.1 has reached feature complete. The team is working on completing the build and bug fixes now. Many have been working really hard to make this happen, especially people working on multi-lib, HOB, kernel works, builds and QA, etc. Thanks to Nitin for completing x32 while his fingers are still hurting :) * The current plan is to have M3 RC2 build today or tomorrow, then QA will run weekly test on RC2. Will have another build RC3 with critical fixes over the weekend and have QA run full test next week on RC3.  Attendees: Mark, Paul, Saul, Josh, Jefro, Dave, Scott, Jessica, Tom, Richard, Jason, Song  New Action Item List: ------------------------------------------------------------------------------------------------------------------------------------ * The team to review ALL bugs they own and do some estimation. Please put the estimation on the white board by next Thursday (8/4/11) ------------------------------------------------------------------------------------------------------------------------------------ * ScottR will follow up with Paul on doc auto-generation and revision inclusion in the doc. The goal is to have docs refer to the release? * Song will follow up with Tracey to make sure marketing planning happens between milestones so that there is no disruption to the development schedule.    Agenda:  1. Opens collection - 5 min (Song) 2. M3 release criteria and plan - 20 min (team) https://wiki.yoctoproject.org/wiki/Yocto_Project_v1.1_Release_Criteria * Yocto 1.1 reached feature complete last week. * 2 tasks remaining for M3 (Beth) 3. M4: bug effort estimate (see messages above) 4. Opens - 10 min i. Yocto 1.2 feature request ii. Saul: The state of pull request, and consolidated pull request: * Sending out to consolidated pull request, as M3 is wrapping up. Saul will look at the remaining pull requests, things will move forward a little slower, 24 hours response time for pull request. Saul will be out next week. Richard will be handling the pull requests. iii. ScottR: ADT/HOB scripts: 2 videos for ADT and HOB. *Jessica/Josh creating the video piece and script. Scott needs the scripts. Just a heads up. Jessica and Josh are ok with this. This is scheduled for M4.  5. Action item Review - 5 min  Action Item List: ------------------------------------------------------------------------------------------------------------------------------------ * The team to review ALL bugs they own and do some estimation. Please put the estimation on the white board by next Thursday (8/4/11) ------------------------------------------------------------------------------------------------------------------------------------ * Saul: talk to Jianjun, QA should work with developers to make sure we can cover more distro testing (informal) before major release. Joshua and Darren volunteers for some of these testing. (Done) * Beth/Darren/ScottR: document the central location in the release, and put any doc changes in the location after release. - git repos should match the website - Include the revision in the doc when we generate the doc from the source. See how we can auto generate the doc. Scott will follow up Paul on this. The goal is to have doc refer to the release. * Song will follow up with Tracey to make sure marketing planning happens between milestones so that there is no disruption to the development schedule. - Tracey was on vacation. Back this week
|
|
|
|
Re: kernel panic - not syncing: No init found
Bruce Ashfield <bruce.ashfield@...>
On 08/02/11 01:38, Francis Meyvis wrote:
Hello,Interesting, since this is failing on multiple image types, there should be different init scripts at play, each with different requirements that should be met by the rootfs generation. Have you tried updating master and re-building ? Some of the issues from last week should be resolved at this point. The boards are known to boot, and boot out of the box, so either this is something transient, or unique to your setup at the moment. Bruce
|
|
|
|
Re: examples / docs on utilizing an external toolchain
Rifenbark, Scott M <scott.m.rifenbark@...>
Is the problem here that we need better documentation on using a prebuilt toolchain? I will have to learn more about the context of the problem from someone. The bug could just be against documentation in general and specify the need for "better documentation using a prebuilt toolchain."
toggle quoted messageShow quoted text
ScottR
-----Original Message-----
From: Kumar Gala [mailto:galak@...] Sent: Tuesday, August 02, 2011 9:10 AM To: Stewart, David C Cc: 'yocto@...'; Rifenbark, Scott M Subject: Re: [yocto] examples / docs on utilizing an external toolchain I can, but not sure what the bug is. - k On Jul 28, 2011, at 11:39 PM, Stewart, David C wrote: Scott - I have had a couple of questions about this exact topic. Can you please submit a bug on this? Thanks.
|
|
|
|
Re: examples / docs on utilizing an external toolchain
Kumar Gala <galak@...>
I can, but not sure what the bug is.
toggle quoted messageShow quoted text
- k
On Jul 28, 2011, at 11:39 PM, Stewart, David C wrote:
Scott - I have had a couple of questions about this exact topic. Can you please submit a bug on this? Thanks.
|
|
|
|
Re: Magic/File problems
Jeff Mitchell <jmitchell@...>
Hi there,
On 08/01/2011 06:18 PM, Saul Wold wrote: Which version of poky are you working with?Git master as of that point. Revision 46cf540e63a848512617b20fd8492f81bfb2f704 There was a problem that weI do have a tmp/sysroots/x86_64-linux/usr/bin/file.real -- although I'm building for BeagleBoard, and don't have one in that directory. Thanks, Jeff
|
|
|
|
Re: kernel panic - not syncing: No init found
Francis Meyvis <francis.meyvis@...>
Hello, When I mounted the root image on loop back device, I actually ran the init program. It executed and complained about missing params. I tried with the 5, like for network connected inittab. Then init complained there was no /dev/initctl. So combining this with the replies I got, I think the latest /sbin/init on yocto master does not support a default run level? I think the architecture of my root image is for x86 cause my pc could exec it. Any suggestions on how to continue? Groetjes,
On Aug 1, 2011 6:04 AM, "Bruce Ashfield" <bruce.ashfield@...> wrote: > On 11-07-31 11:30 PM, Tom Zanussi wrote: >> On Sun, 2011-07-31 at 19:42 -0700, Bruce Ashfield wrote: >>> On 11-07-31 3:44 PM, Francis Meyvis wrote: >>>> Hello, >>>> >>>> I probably miss something trivial. >>>> I cloned the git://git.yoctoproject.org/poky.git >>>> I build on a 64 bit machine a qemux86-64 >>>> (. ./oe-init-build-env qemux86-64 and changing the conf/local.conf) >>>> I build the core-image-sato& core-image-minimal. >>>> Then I try to run with >>>> runqemu qemux86-64 core-image-sato ext3 >>>> >>>> But both sato and minimal give me this message: >>>> Kernel panic - not syncing: No init found. >>>> >>>> As runqemu showed me the full command line I tried to add the --append >>>> command line option >>>> init=/sbin/init and /sbin/init.sysvinit >>>> But that did not help (there's a message saying Failed to execute >>>> /sbin/init. Attempting defaults...) >>>> I verified by mounting these ext3 images on a loop device that there's >>>> really a /sbin/init present. >>>> >>>> Can somebody tell what I'm doing wrong? >>> >>> Can you send a full bootlog, or paste it somewhere accessible ? It's hard >>> to say what's with the information you've given. >>> >>> Is this the latest yocto master ? >>> >>> If you are seeing that message it typically means that the the >>> device that is being used as the root isn't ready (fixed by >>> rootwait/rootdelay) or isn't supported (i.e. NFS root without >>> the right ethernet device). Changing what init is, won't change >>> the result if either one of those is the case. >>> >> >> I'm also seeing this on both sugarbay and jasperforest with the latest >> master. rootwait doesn't help. >> >> rtc_cmos 00:07: setting system clock to 2010-01-02 05:35:34 UTC >> (1262410534) >> Freeing unused kernel memory: 720k freed >> Failed to execute /init >> Kernel panic - not syncing: No init found. Try passing init= option to >> kernel. >> See Linux Documentation/init.txt for guidance. >> Pid: 1, comm: swapper Not tainted 3.0.0-rc7-yocto-standard+ #1 >> Call Trace: >> [<ffffffff81541b77>] panic+0x9b/0x191 >> [<ffffffff81540862>] init_post+0xc0/0xc0 >> [<ffffffff8188bcef>] kernel_init+0x17b/0x17b >> [<ffffffff8154b234>] kernel_thread_helper+0x4/0x10 >> [<ffffffff8188bb74>] ? start_kernel+0x377/0x377 >> [<ffffffff8154b230>] ? gs_change+0xb/0xb >> >> I thought initially it had something to do with the -live image changes, >> but reverting the two -live patches didn't help. >> >> The problem seems to coincide with the tune file changes, but that may >> be a red herring. > > Nope. I don't think it's a red herring: > > Both of these: > > Freeing unused kernel memory: 720k freed > Failed to execute /init > > Indicate that the device came up and init was loaded, the kernel > passed control to userspace and then things went bad. i.e. you just > ran init of the wrong arch or something incompatible with the > kernel support, etc. > > Since I'm having no trouble with old rootfs and new kernels, that's > another sign. I'm assuming that old userspace's boot for you ? > > Bruce > >> >> Tom >> >> >>> Cheers, >>> >>> Bruce >>> >>>> >>>> BTW is there any way to not have to run qemu with root permissions? >>>> I ran the android emulator and it does not require me to be root. >>>> Should I configure something on my Ubuntu machine to get qemu to function? >>>> >>>> Thanks, >>>> francis >>>> _______________________________________________ >>>> yocto mailing list >>>> yocto@... >>>> https://lists.yoctoproject.org/listinfo/yocto >>> >>> _______________________________________________ >>> yocto mailing list >>> yocto@... >>> https://lists.yoctoproject.org/listinfo/yocto >> >> >
|
|
|
|
Re: Magic/File problems
Saul Wold <sgw@...>
On 08/01/2011 10:55 AM, Jeff Mitchell wrote:
On my host system I have the "file" utility version 5.05. The version inJeff, Which version of poky are you working with? There was a problem that we had with file at one point and thought was fixed. If you check in your tmp/sysroots directory for the host machine, do you have a file.real? We needed to create a cover script which correctly pointed file to the right magic.mgc file. Hope that helped. Sau!
|
|
|
|
Rescheduled: mailing lists are moving to a new host
Michael Halstead <michaelx.d.halstead@...>
The move originally scheduled for today will now happen on Wednesday August 3rd from 10pm to midnight PDT.
For reference, we will move mailing lists and related services to new hardware hosted at the OSU Open Source Lab. Archive access will be available during the entire process but changes using the web interface at lists.yoctoproject.org may be lost. E-mail sent to lists will be delayed until the transition is complete. If you send an e-mail to the list during the move expect that it will go through after a few hours. There should be no need to resend e-mail. Michael Halstead Intel OTC Yocto
|
|
|
|
Magic/File problems
Jeff Mitchell <jmitchell@...>
On my host system I have the "file" utility version 5.05. The version in poky/in my build directories is version 5.04. I'm seeing a lot of builds fail with the following:
| error: magic_load(ms, /usr/share/misc/magic) failed: File 5.4 supports only version 7 magic files. `/usr/share/misc/magic.mgc' is version 8 Note that the version there is 5.4; this doesn't match anything as far as I can tell. Any ideas? Thanks, Jeff
|
|
|
|
Re: POKY_DISTRO_VERSION or OECORE_DISTRO_VERSION?
Zhang, Jessica
Lianhao,
toggle quoted messageShow quoted text
For DISTRO_VERSION, we still need to use POKY_DISTRO_VERSION since our distro is "POKY" as to other rename to OECORE, are those variables that're distro neutral. Hope you get the idea. I've asked RP the same question and his answer is before we can figure out a mechanism to make everything distro neutral, we have stay with the existing POKY naming convention for certain variables that are specific to the distro. Thanks, Jessica
-----Original Message-----
From: Lu, Lianhao Sent: Monday, August 01, 2011 12:45 AM To: Zhang, Jessica; Ke, Liping Cc: yocto@... Subject: POKY_DISTRO_VERSION or OECORE_DISTRO_VERSION? Hi Jessica, I noticed the we still use POKY_DISTRO_VERSION in the environment while some other variables have changed their prefix to OECORE_ (e.g. OECORE_NATIVE_SYSROOT, etc.) Which prefix should we use for those version related variables in the environment files? OECORE_ or POKY_? -Lianhao
|
|
|
|
Re: crownbay-noemgd poky-image-minimal fails to build
Andre Haupt <andre@...>
On Fri, Jul 29, 2011 at 10:32:03AM -0400, Bruce Ashfield wrote:
Thank you very much, Bruce. The patch is now picked up, andI pushed out the change. If your SRCREV isArgh. That explains it. You can always cherry pick / applyAs for the second, there's a patch in the 2.6.34 kernel treeThis commit is not in the crownbay-standard branch. the build runs successfully. cheers, Andre
|
|
|
|
POKY_DISTRO_VERSION or OECORE_DISTRO_VERSION?
Lu, Lianhao <lianhao.lu@...>
Hi Jessica,
I noticed the we still use POKY_DISTRO_VERSION in the environment while some other variables have changed their prefix to OECORE_ (e.g. OECORE_NATIVE_SYSROOT, etc.) Which prefix should we use for those version related variables in the environment files? OECORE_ or POKY_? -Lianhao
|
|
|
|
Re: kernel panic - not syncing: No init found
Bruce Ashfield <bruce.ashfield@...>
On 11-07-31 11:30 PM, Tom Zanussi wrote:
On Sun, 2011-07-31 at 19:42 -0700, Bruce Ashfield wrote:Nope. I don't think it's a red herring:On 11-07-31 3:44 PM, Francis Meyvis wrote:I'm also seeing this on both sugarbay and jasperforest with the latestHello,Can you send a full bootlog, or paste it somewhere accessible ? It's hard Both of these: Freeing unused kernel memory: 720k freed Failed to execute /init Indicate that the device came up and init was loaded, the kernel passed control to userspace and then things went bad. i.e. you just ran init of the wrong arch or something incompatible with the kernel support, etc. Since I'm having no trouble with old rootfs and new kernels, that's another sign. I'm assuming that old userspace's boot for you ? Bruce
|
|
|
|
Re: kernel panic - not syncing: No init found
Tom Zanussi <tom.zanussi@...>
On Sun, 2011-07-31 at 19:42 -0700, Bruce Ashfield wrote:
On 11-07-31 3:44 PM, Francis Meyvis wrote:I'm also seeing this on both sugarbay and jasperforest with the latestHello,Can you send a full bootlog, or paste it somewhere accessible ? It's hard master. rootwait doesn't help. rtc_cmos 00:07: setting system clock to 2010-01-02 05:35:34 UTC (1262410534) Freeing unused kernel memory: 720k freed Failed to execute /init Kernel panic - not syncing: No init found. Try passing init= option to kernel. See Linux Documentation/init.txt for guidance. Pid: 1, comm: swapper Not tainted 3.0.0-rc7-yocto-standard+ #1 Call Trace: [<ffffffff81541b77>] panic+0x9b/0x191 [<ffffffff81540862>] init_post+0xc0/0xc0 [<ffffffff8188bcef>] kernel_init+0x17b/0x17b [<ffffffff8154b234>] kernel_thread_helper+0x4/0x10 [<ffffffff8188bb74>] ? start_kernel+0x377/0x377 [<ffffffff8154b230>] ? gs_change+0xb/0xb I thought initially it had something to do with the -live image changes, but reverting the two -live patches didn't help. The problem seems to coincide with the tune file changes, but that may be a red herring. Tom Cheers,
|
|
|
|
Re: kernel panic - not syncing: No init found
Bruce Ashfield <bruce.ashfield@...>
On 11-07-31 3:44 PM, Francis Meyvis wrote:
Hello,Can you send a full bootlog, or paste it somewhere accessible ? It's hard to say what's with the information you've given. Is this the latest yocto master ? If you are seeing that message it typically means that the the device that is being used as the root isn't ready (fixed by rootwait/rootdelay) or isn't supported (i.e. NFS root without the right ethernet device). Changing what init is, won't change the result if either one of those is the case. Cheers, Bruce
|
|
|
|
kernel panic - not syncing: No init found
Francis Meyvis <francis.meyvis@...>
Hello,
I probably miss something trivial. I cloned the git://git.yoctoproject.org/poky.git I build on a 64 bit machine a qemux86-64 (. ./oe-init-build-env qemux86-64 and changing the conf/local.conf) I build the core-image-sato & core-image-minimal. Then I try to run with runqemu qemux86-64 core-image-sato ext3 But both sato and minimal give me this message: Kernel panic - not syncing: No init found. As runqemu showed me the full command line I tried to add the --append command line option init=/sbin/init and /sbin/init.sysvinit But that did not help (there's a message saying Failed to execute /sbin/init. Attempting defaults...) I verified by mounting these ext3 images on a loop device that there's really a /sbin/init present. Can somebody tell what I'm doing wrong? BTW is there any way to not have to run qemu with root permissions? I ran the android emulator and it does not require me to be root. Should I configure something on my Ubuntu machine to get qemu to function? Thanks, francis
|
|
|
|
libpam-1.1.4-r1 strange behavior
Francis Meyvis <francis.meyvis@...>
Hello,
My download directory is on a CIFS mounted drive. The idea was that other PCs could access the already downloaded files on the local network. The package libpam however failed processing. After changing the path on which I had mounted the CIFS partition, to a PC local download directory, the package libpam succeeded. What I saw: It seemed to get in a busy loop related to SMB when using files name "*.lock" and "*.done" I found this by starting a wireshark session to see what was going on. It seems to try to QUERY_PATH_INFO for *.lock in the download directory. The CIFS server responds STATUS_OBJECT_NAME_INVALID After I changed to a local download directory, I found these files after processing the package in the download directory: -rw-r--r-- 1 franchan franchan 0 2011-07-31 18:38 99_pam.done -rw-r--r-- 1 franchan franchan 0 2011-07-31 18:38 *.done -rw-r--r-- 1 franchan franchan 0 2011-07-31 18:38 libpam-xtests.patch.done -rw-r--r-- 1 franchan franchan 1123186 2011-06-24 13:42 Linux-PAM-1.1.4.tar.bz2 -rw-r--r-- 1 franchan franchan 0 2011-07-31 18:38 Linux-PAM-1.1.4.tar.bz2.done According to me CIFS has troubles with these *.lock and *.done flag files. Could the libpam package be done in another way so that it works for CIFS files systems? Or should I not have used CIFS fs when using bitbake? Thanks, francis
|
|
|
|
Yocto weekly bug trend charts -- WW31
Xu, Jiajun <jiajun.xu@...>
Hi all,
Yocto 1.1 M2 released last week. The new-submit vs. fixed bug is 28 vs. 36. A lot of bugs are fixed, which make both open bug number and WDD number decreased a lot -- Open bug number from 185 to 174; WDD number from 943 to 887. Bug status of WW30 could be found on https://wiki.pokylinux.org/wiki/Yocto_Bug_Trend. Best Regards, Jiajun
|
|
|
|
Re: [PATCH][linux-yocto-3.0] drivers/misc/pch_phub.c: don't oops if dmi_get_system_info returns NULL
Bruce Ashfield <bruce.ashfield@...>
On 11-07-30 2:34 PM, Khem Raj wrote:
On Thursday, July 28, 2011 04:51:41 PM Darren Hart wrote:Could very well be. I grabbed this one directly from linus' tree,Bruce,May be it could be just if ( dmi_get_system_info(DMI_BOARD_NAME)&& so I've got it as-is for now, but I've tagged it to be considered later if we do decide that we want to send a tweak of this upstream Thanks! Bruce pch_phub_read_modify_write_reg(chip,
|
|
|
|
Re: [PATCH][linux-yocto-3.0] drivers/misc/pch_phub.c: don't oops if dmi_get_system_info returns NULL
On Thursday, July 28, 2011 04:51:41 PM Darren Hart wrote:
Bruce,May be it could be just if ( dmi_get_system_info(DMI_BOARD_NAME) && strstr(board_name, "CM-iTC")) although I think compiler will already do it internally pch_phub_read_modify_write_reg(chip,-- Khem Raj
|
|
|