[agl-discussions] Raspi can't compile: Can't get CWD
Stephane Desneux
stephane.desneux at iot.bzh
Fri Jan 27 23:42:02 UTC 2017
Good catch :)
About pushd/popd and bash/dash problem: using "ln -s" correctly doesn't require
to use "cd" first: it's easy to pass the correct arguments and this is the fix
to propose probably.
But I wonder why in most cases, the problem doesn't show up: this may be related
to bitbake scheduling, but given the error, it should occur quite often and
that's not what we observe...
I don't want to blame anyone, but "git blame" tells that the line "cd ..." in
the recipe was introduced on 2/2/16, that is 1 year ago approx. So we shouldn't
be the first ones to see it (or is it due to a slow derivative of Murphy's law
:) ? )
Best,
---
Stephane Desneux - CTO - IoT.bzh
stephane.desneux at iot.bzh - www.iot.bzh
On 27/01/17 21:00, Blayne.Dennis at elektrobit.com wrote:
> I have created a bug ticket https://bugzilla.yoctoproject.org/show_bug.cgi?id=10985
>
> Blayne Dennis
> Software Engineer
> Elektrobit Automotive Americas, Inc
> Blayne.Dennis at elektrobit.com
>
>
>
> -----Original Message-----
> From: Jan-Simon Möller [mailto:jsmoeller at linuxfoundation.org]
> Sent: Friday, January 27, 2017 2:22 PM
> To: Dennis Blayne <Blayne.Dennis at elektrobit.com>
> Cc: stephane.desneux at iot.bzh; automotive-discussions at lists.linuxfoundation.org
> Subject: Re: [agl-discussions] Raspi can't compile: Can't get CWD
>
> Am Freitag, 27. Januar 2017, 19:06:51 schrieb Blayne.Dennis at elektrobit.com:
>> I change poky/meta/recipes-core/dbus/dbus_1.10.6.bb from:
>> 125 cd ${D}${systemd_system_unitdir}/dbus.target.wants/
>> 126 ln -fs ../dbus.socket
>> ${D}${systemd_system_unitdir}/dbus.target.wants/dbus.socket 127 ln
>> -fs ../dbus.socket
>> ${D}${systemd_system_unitdir}/sockets.target.wants/dbus.socket 128
>> ln -fs ../dbus.service
>> ${D}${systemd_system_unitdir}/multi-user.target.wants/dbus.service
>>
>> To:
>> 125 OLDPATH=`pwd`
>> 126 cd ${D}${systemd_system_unitdir}/dbus.target.wants/
>> 127 ln -fs ../dbus.socket
>> ${D}${systemd_system_unitdir}/dbus.target.wants/dbus.socket 128 ln
>> -fs ../dbus.socket
>> ${D}${systemd_system_unitdir}/sockets.target.wants/dbus.socket 129
>> ln -fs ../dbus.service
>> ${D}${systemd_system_unitdir}/multi-user.target.wants/dbus.services 130
>> cd $OLDPATH
>>
>
>
>> And it didn't error. Not sure if it actually worked, but comment if it
>> comes up again.
>
> Yep, that is the solution then.
>
>> Note: I tried pushd/popd, but the shell couldn't find command 'pushd'
>> so I improvised
>
> pushd/popd is bash, so if you're defaulting to dash, this might not work.
>
> We should enter this into http://bugzilla.yoctoproject.org/ .
>
> Best,
> Jan-Simon Möller
>
>
>
>> -----Original Message-----
>> From: Jan-Simon Möller [mailto:jsmoeller at linuxfoundation.org]
>> Sent: Friday, January 27, 2017 1:53 PM
>> To: Dennis Blayne <Blayne.Dennis at elektrobit.com>
>> Cc: stephane.desneux at iot.bzh;
>> automotive-discussions at lists.linuxfoundation.org Subject: Re:
>> [agl-discussions] Raspi can't compile: Can't get CWD
>>
>> Am Freitag, 27. Januar 2017, 18:41:22 schrieb Blayne.Dennis at elektrobit.com:
>>> I reran source aglsetup.sh command with "agl-sota" at the end, and
>>> the same error occurred in dbus.
>>>
>>> Stephane, where is the specific recipe in question?
>>
>> the poky version is in:
>> poky/meta/recipes-core/dbus/dbus_1.10.6.bb
>>
>>> -----Original Message-----
>>> From: Stephane Desneux [mailto:stephane.desneux at iot.bzh]
>>> Sent: Friday, January 27, 2017 12:53 PM
>>> To: Dennis Blayne <Blayne.Dennis at elektrobit.com>;
>>> jsmoeller at linuxfoundation.org;
>>> automotive-discussions at lists.linuxfoundation.org Subject: Re:
>>> [agl-discussions] Raspi can't compile: Can't get CWD
>>>
>>> As Jan-Simon said, we reproduced the problem on Ubuntu.
>>>
>>> Blayne's logs and ours seem to be the same and we can see this in
>>> both
>>> cases: in the dbus recipe (upstream poky), we can see that if
>>> systemd is enabled, then the do_install task will "cd" into the
>>> sysroot. Then at the end, the same directory gets deleted, then the
>>> CWD error pops up and things get worse and worse.
>>>
>>> Also there's a core dump somewhere (which process ? pseudo ?).
>>>
>>> So maybe that fixing the dbus recipe using pushd/popd instead of "cd"
>>> would do the trick ?
>>>
>>> Best regards,
>>> ---
>>> Stephane Desneux - CTO - IoT.bzh
>>> stephane.desneux at iot.bzh - www.iot.bzh
>>>
>>> On 27/01/17 18:28, Blayne.Dennis at elektrobit.com wrote:
>>>> When you say include "agl-sota" do you mean in the source command
>>>> for
>>>> aglsetup.sh?:
>>>>
>>>> Source $AGL_TOP/meta-agl/scripts/aglsetup.sh -f -m raspberrypi3
>>>> agl-demo agl-netboot agl-appfw-smack agl-sota
>>>>
>>>>
>>>>
>>>> I attached the dbus build log. Below is the output from the
>>>> requested
>>>> commands:
>>>>
>>>> [17:50:36]
>>>> devel at agl-worker-sdldev-VirtualBox-0-sdldev:~/workspace_agl$ uname
>>>> -a Linux agl-worker-sdldev-VirtualBox-0-sdldev 4.4.0-59-generic
>>>> #80-Ubuntu SMP Fri Jan 6 17:47:47 UTC 2017 x86_64 GNU/Linux
>>>>
>>>> [18:07:15] devel at agl-worker-sdldev-VirtualBox-0-sdldev:~/workspace_agl$
>>>> ulimit -a core file size (blocks, -c) unlimited
>>>> data seg size (kbytes, -d) unlimited
>>>> scheduling priority (-e) 0
>>>> file size (blocks, -f) unlimited
>>>> pending signals (-i) 15653
>>>> max locked memory (kbytes, -l) 64
>>>> max memory size (kbytes, -m) unlimited
>>>> open files (-n) 65536
>>>> pipe size (512 bytes, -p) 8
>>>> POSIX message queues (bytes, -q) 819200
>>>> real-time priority (-r) 0
>>>> stack size (kbytes, -s) 8192
>>>> cpu time (seconds, -t) unlimited
>>>> max user processes (-u) unlimited
>>>> virtual memory (kbytes, -v) unlimited
>>>> file locks (-x) unlimited
>>>>
>>>> -----Original Message-----
>>>> From: Jan-Simon Möller [mailto:jsmoeller at linuxfoundation.org]
>>>> Sent: Friday, January 27, 2017 11:51 AM
>>>> To: automotive-discussions at lists.linuxfoundation.org
>>>> Cc: Dennis Blayne <Blayne.Dennis at elektrobit.com>
>>>> Subject: Re: [agl-discussions] Raspi can't compile: Can't get CWD
>>>>
>>>> Hi all!
>>>>
>>>> Update on the issue:
>>>>
>>>> it has been reproduced on one more kubuntu system.
>>>>
>>>> We're investigating if it is related to upstream's version of
>>>> pseudo, the dbus recipe or if it is *fixed* by the pseudo version
>>>> uprev in meta-sota.
>>>>
>>>> In any case: try to build with "agl-sota" enabled. This might help
>>>> as we use a newer version of pseudo there.
>>>>
>>>> Opened SPEC-430.
>>>>
>>>> Please comment/continue there.
>>>>
>>>> Best,
>>>> Jan-Simon
>>>>
>>>> Am Freitag, 27. Januar 2017, 17:23:09 schrieb Jan-Simon Möller:
>>>>> Also:
>>>>>
>>>>> uname -a
>>>>> ulimit -a
>>>>>
>>>>> distro / version
>>>>>
>>>>> Best,
>>>>> JS
>>>>>
>>>>> Am Freitag, 27. Januar 2017, 10:18:42 schrieb Jan-Simon Möller:
>>>>>> Hi Dennis,
>>>>>>
>>>>>> I ran the build here with your manifest and it worked. So let's
>>>>>> get more context from the build: Please run
>>>>>>
>>>>>> bitbake -vv -D dbus | tee -a dbus.build.log 2>&1
>>>>>>
>>>>>> and send the logfile.
>>>>>>
>>>>>> Best,
>>>>>> Jan-Simon
>>>>>>
>>>>>> Am Donnerstag, 26. Januar 2017, 16:55:40 schrieb
>>>>>
>>>>> Blayne.Dennis at elektrobit.com:
>>>>>>> Alright, I reran everything and this time I saved the output in
>>>>>>> individual files. Is this what is needed?
>>>>>>>
>>>>>>> -----Original Message-----
>>>>>>> From: Jan-Simon Möller [mailto:jsmoeller at linuxfoundation.org]
>>>>>>> Sent: Thursday, January 26, 2017 10:46 AM
>>>>>>> To: automotive-discussions at lists.linuxfoundation.org
>>>>>>> Cc: Dennis Blayne <Blayne.Dennis at elektrobit.com>
>>>>>>> Subject: Re: [agl-discussions] Raspi can't compile: Can't get
>>>>>>> CWD
>>>>>>>
>>>>>>> manifest paste is incomplete (end of lines missing).
>>>>>>>
>>>>>>> JS
>>>>>>>
>>>>>>> Am Donnerstag, 26. Januar 2017, 15:29:10 schrieb
>>>>>>
>>>>>> Blayne.Dennis at elektrobit.com:
>>>>>>>> Any thoughts based on this informstion?
>>>>>>>>
>>>>>>>> -----Original Message-----
>>>>>>>> From: Dennis Blayne
>>>>>>>> Sent: Tuesday, January 24, 2017 12:53 PM
>>>>>>>> To: 'Jan-Simon Möller' <dl9pf at gmx.de>
>>>>>>>> Cc: patrick.ohly at intel.com;
>>>>>>>> automotive-discussions at lists.linuxfoundation.org
>>>>>>>> Subject: RE: [agl-discussions] Raspi can't compile: Can't get
>>>>>>>> CWD
>>>>>>>>
>>>>>>>> I tried building in the container, in the host, and tried
>>>>>>>> changing to the latest in the Chinook branch (-b chinook -u
>>>>>>>> https://gerrit.automotivelinux.org/gerrit/AGL/AGL-repo), all
>>>>>>>> had the same error. See attached local.conf.
>>>>>>>>
>>>>>>>>
>>>>>>>> [18:49:32]
>>>>>>>> devel at agl-worker-sdldev-VirtualBox-0-sdldev:~/workspace_agl$
>>>>>>>> repo manifest -r <?xml version="1.0" encoding="UTF-8"?>
>>>>>>>> <manifest> <remote fetch="https://github.com/01org/"
>>>>>>>> name="01org"/>>
>>>>>>>>
>>>>>>>> <remote fetch="https://gerrit.automotivelinux.org/gerrit/"
>>>>>>>> name="agl"
>>>>>>>>
>>>>>>>> pushurl="ssh://gerrit.automotivelinux.org:29418 <remote
>>>>>>>> fetch="git://git.freescale.com/imx" name="fsl-release"/>
>>>>>>>> <remote fetch="https://github.com/" name="github"/>
>>>>>>>>
>>>>>>>> <remote fetch="git://git.openembedded.org/" name="openembedded"/>
>>>>>>>> <remote fetch="git://git.yoctoproject.org/" name="yocto"/>
>>>>>>>>
>>>>>>>> <default remote="agl" revision="chinook" sync-j="4"/>
>>>>>>>>
>>>>>>>> <project name="AGL/meta-agl" path="meta-agl"
>>>>>>>>
>>>>>>>> revision="5a85901e245aa2605721c7dae645a00156a67dde"
>>>>>>>> upstream="chinook"/> <project name="AGL/meta-agl-demo"
>>>>>>>> path="meta-agl-demo"
>>>>>>>> revision="fb92e316727a1ad0012ac90ed52c5d9cecdcf433" upstream="
>>>>>>>> <project name="AGL/meta-agl-devel" path="meta-agl-devel"
>>>>>>>> revision="134e5f446b52f6a02f38212b9bde076adad798f8" upstream
>>>>>>>> <project name="AGL/meta-agl-extra" path="meta-agl-extra"
>>>>>>>> revision="812bc8ab9490e2418dedcf3a868a0ef9393ab01f" upstream
>>>>>>>> <project name="AGL/meta-renesas" path="meta-renesas"
>>>>>>>> revision="90122dd63d72061b93bd3cbe3a7b520aaa531a19"
>>>>>>>> upstream="ch <project name="CogentEmbedded/meta-rcar" path="meta-rcar"
>>>>>>>> remote="github"
>>>>>>>> revision="cc6c4ee3f61b4dd864804b6b34c295cd0 <project
>>>>>>>> name="Freescale/meta-fsl-arm-extra" path="meta-fsl-arm-extra"
>>>>>>>> remote="github" revision="e95f4ae61fdaf6452d6d <project
>>>>>>>> name="meta-fsl-arm"
>>>>>>>> remote="yocto" revision="e2254e7b2ded0c2b66b1226f879b3a6d52037b2d"
>>>>>>>> upstream="krogoth"/> <project name="meta-intel" remote="yocto"
>>>>>>>> revision="1f8dd1b00ce9c72d73583c405ec392690d9b08b7"
>>>>>>>> upstream="krogoth"/> <project name="meta-intel-iot-security"
>>>>>>>> remote="01org"
>>>>>>>> revision="20bbb97f6d5400b126ae96ef446c3e60c7e16285"/> <project
>>>>>>>> name="meta-oic" remote="yocto"
>>>>>>>> revision="0bdd959e0dee4f63098702c302d1d214ac3d808a"/> <project
>>>>>>>> name="meta-openembedded" remote="openembedded"
>>>>>>>> revision="55c8a76da5dc099a7bc3838495c672140cedb78e" upstream=
>>>>>>>> <project name="meta-qcom" remote="yocto"
>>>>>>>> revision="3bfe83535358289aa7f7342ed0977c076e7550c8"
>>>>>>>> upstream="krogoth"/> <project name="meta-qt5/meta-qt5"
>>>>>>>> path="meta-qt5"
>>>>>>>> remote="github" revision="da138c99142bec136e43eba35ce09d86f00eca2c"
>>>>>>>> <project
>>>>>>>> name="meta-raspberrypi" remote="yocto"
>>>>>>>> revision="a5f9b07a820d50ae5fb62e07306cd4e72d8638a9"
>>>>>>>> upstream="krogoth <project name="meta-rust/meta-rust"
>>>>>>>> path="meta-rust"
>>>>>>>> remote="github"
>>>>>>>> revision="d0663639a08ed60bb83fd6eb99e3e2045b21b5
>>>>>>>> <project
>>>>>>>> name="meta-security-isafw" remote="01org"
>>>>>>>> revision="f44f15bb54057dd60399bb5dc7e0ecc7822168a2"
>>>>>>>> upstream="mast <project name="meta-ti" remote="yocto"
>>>>>>>> revision="d9a81e3be2945e11bd9da58b6de2ac31d51e9286"
>>>>>>>> upstream="krogoth"/> <project name="poky" remote="yocto"
>>>>>>>> revision="ae9b341ecfcc60e970f29cfe04306411ad26c0cf"
>>>>>>>> upstream="krogoth"/> <project name="tripzero/meta-amb"
>>>>>>>> path="meta-amb"
>>>>>>>> remote="github" revision="ef3495bb8d6543709f6d1f7b657cb894d32c1757"
>>>>>>>> </manifest>
>>>>>>>>
>>>>>>>> [18:49:27]
>>>>>>>>
>>>>>>>> devel at agl-worker-sdldev-VirtualBox-0-sdldev:~/workspace_agl$
>>>>>>>> repo status nothing to commit (working directory clean)
>>>>>>>>
>>>>>>>>
>>>>>>>> -----Original Message-----
>>>>>>>> From: Jan-Simon Möller [mailto:dl9pf at gmx.de]
>>>>>>>> Sent: Tuesday, January 24, 2017 11:55 AM
>>>>>>>> To: Dennis Blayne <Blayne.Dennis at elektrobit.com>
>>>>>>>> Cc: patrick.ohly at intel.com;
>>>>>>>> automotive-discussions at lists.linuxfoundation.org
>>>>>>>> Subject: Re: [agl-discussions] Raspi can't compile: Can't get
>>>>>>>> CWD
>>>>>>>>
>>>>>>>> To look into that I need to reproduce your env. Please send:
>>>>>>>>
>>>>>>>> output of:
>>>>>>>>
>>>>>>>> - repo manifest -r
>>>>>>>> - repo status
>>>>>>>>
>>>>>>>> attach also :
>>>>>>>> conf/local.conf
>>>>>>>>
>>>>>>>> Best,
>>>>>>>> Jan-Simon
>>>>>>>>
>>>>>>>> Am Dienstag, 24. Januar 2017, 16:03:09 schrieb
>>>>>>
>>>>>> Blayne.Dennis at elektrobit.com:
>>>>>>>>> Is it possible for yocto to run multiple processes against
>>>>>>>>> the same recipe?
>>>>>>>>> In which case there might be a conflict from one process
>>>>>>>>> deletes files it no longer needs but the other requires,
>>>>>>>>> causing the race condition?
>>>>>>>>>
>>>>>>>>> -----Original Message-----
>>>>>>>>> From: Patrick Ohly [mailto:patrick.ohly at intel.com]
>>>>>>>>> Sent: Tuesday, January 24, 2017 4:39 AM
>>>>>>>>> To: Dennis Blayne <Blayne.Dennis at elektrobit.com>
>>>>>>>>> Cc: dl9pf at gmx.de;
>>>>>>>>> automotive-discussions at lists.linuxfoundation.org
>>>>>>>>> Subject: Re: [agl-discussions] Raspi can't compile: Can't get
>>>>>>>>> CWD
>>>>>>>>>
>>>>>>>>> On Mon, 2017-01-23 at 21:39 +0000,
>>>>>>>>> Blayne.Dennis at elektrobit.com
>>>>
>>>> wrote:
>>>>>>>>>> I just reran the commands in a new terminal:
>>>>>>>>>>
>>>>>>>>>> devel at agl-worker-sdldev-VirtualBox-0-sdldev:~/workspace_agl/
>>>>>>>>>> build$ source meta-agl/scripts/aglsetup.sh -f -m
>>>>>>>>>> raspberrypi3 agl-demo agl-netboot agl-appfw-smack
>>>>>>>>>>
>>>>>>>>>> devel at agl-worker-sdldev-VirtualBox-0-sdldev:~/workspace_agl/
>>>>>>>>>> build$
>>>>>>>>>> bitbake agl-demo-platform
>>>>>>>>>>
>>>>>>>>>> The same error occurs:
>>>>>>>>>> | Can't get CWD: No such file or directory Couldn't stash
>>>>>>>>>> | directory before opening socket: Stale file handlepseudo:
>>>>>>>>>> | server connection persistently failed, aborting.
>>>>>>>>
>>>>>>>> Aborted (core dumped)
>>>>>>>>
>>>>>>>>>> | WARNING: exit code 134 from a shell command.
>>>>>>>>>> | ERROR: Function failed: do_install (log file is located at
>>>>>>>>>> | /home/devel/workspace_agl/build/tmp/work/cortexa7hf-neon-v
>>>>>>>>>> | fpv4-a
>>>>>>>>>> | gl
>>>>>>>>>> | -l
>>>>>>>>>> | inux-gnueabi/dbus/1.10.6-r0/temp/log.do_install.24665)
>>>>>>>>>>
>>>>>>>>>> ERROR: Task 3212
>>>>>>>>>> (/home/devel/workspace_agl/poky/meta/recipes-core/dbus/dbus_
>>>>>>>>>> 1.10.6
>>>>>>>>>> .b
>>>>>>>>>> b,
>>>>>>>>>> do_install) failed with exit code '1'
>>>>>>>>>
>>>>>>>>> To me, that looks like the current working directory of the
>>>>>>>>> process got deleted while the process is still using it, i.e.
>>>>>>>>> it's some kind of race condition. That this then leads to an
>>>>>>>>> error from pseudo would be just a follow-up error.
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Best Regards, Patrick Ohly
>>>>>>>>>
>>>>>>>>> The content of this message is my personal opinion only and
>>>>>>>>> although I am an employee of Intel, the statements I make
>>>>>>>>> here in no way represent Intel's position on the issue, nor
>>>>>>>>> am I authorized to speak on behalf of Intel on this matter.
>>>>>>>>
>>>>>>>> --
>>>>>>>> --
>>>>>>>> Jan-Simon Möller
>>>>>>>> dl9pf at gmx.de
>>>>>>>> _______________________________________________
>>>>>>>> automotive-discussions mailing list
>>>>>>>> automotive-discussions at lists.linuxfoundation.org
>>>>>>>> https://lists.linuxfoundation.org/mailman/listinfo/automotive-
>>>>>>>> di
>>>>>>>> scussi
>>>>>>>> ons
>>>>
>>>> _______________________________________________
>>>> automotive-discussions mailing list
>>>> automotive-discussions at lists.linuxfoundation.org
>>>> https://lists.linuxfoundation.org/mailman/listinfo/automotive-disc
>>>> us
>>>> si
>>>> ons
>
More information about the automotive-discussions
mailing list