Date   

Re: Integration of Docker in Read-only Rootfile system through Yocto

Poornesh <poornesh.g@...>
 

Greetings !

I have added the docker in yocto (zeus) and it is working with the Rootfile system which is having "Read-write" permission. When I trying to run the "Hello-world" image with the Rootfile system having "Read-only" permission through docker  I am facing issue . Please find the attached error logs .

Requesting your suggestions to overcome this issue.

--

Thanks and Regards

Poornesh G


Re: #archiver Archiver not archiving sources from Mirror #archiver

 

On Fri, 9 Jul 2021 10:21:56 +0000
"Seipel, Christoph" <Christoph.Seipel@...> wrote:

Hello,

I have a problem with the archiver. I want to use it to export the sources used in our image for license compliance (GPL etc.).
We have a BSP supplier that supplies the Yocto system and the sources to build it in a mirror directory.

Now, when I use the archiver (mode: original) the exported sources don't contain the sources from the mirror, but only other files like patches.

To analyse this problem I put some extra logging into the archiver.bbclass to inspect the variable contents, and it seems the "local" variable where the archiver is looking for the sources to export doesn't fit with the path of the actual sources. This leads to the archiver just skipping that URL and proceeding.

For example for the systemd recipe it has the variable URL="git://github.com/systemd/systemd.git;protocol=git" and therefore it decides to look at local="<builddir>/downloads/git2/github.com.systemd.systemd.git"
But that file doesn't exist.
I inspected the do_fetch and do_unpack logs to find that the fetcher resolves "git://github.com/systemd/systemd.git;protocol=git" with returning the corresponding file from the mirror file:///<mirrorpath>/gitshallow_github.com.systemd.systemd.git_46659f7-1_master.tar.gz<file:///%3cmirrorpath%3e/gitshallow_github.com.systemd.systemd.git_46659f7-1_master.tar.gz>
The do_unpack task then unpacks tar -xzf <builddir>/downloads/gitshallow_github.com.systemd.systemd.git_46659f7-1_master.tar.gz
Which is possible because there is a symlink, I guess that was created by the fetcher
<builddir>/downloads/gitshallow_github.com.systemd.systemd.git_46659f7-1_master.tar.gz -> <mirror>/gitshallow_github.com.systemd.systemd.git_46659f7-1_master.tar.gz

So the problem seems to be that the archiver expects the sources in downloads/git2 but the fetcher creates the symlink in downloads and with additional pre- and postfiexes on the filename.

Sadly we are still on sumo, but at least for the archiver class there isn't any relevant change to hardknott. At least a dunfell upgrade is incoming in the next weeks.
If the archives you've been given have names starting with
'gitshallow_' then it's likely that the following variables were set in
local.conf or in a distro conf file to generate them:

BB_GENERATE_MIRROR_TARBALLS = "1"
BB_GENERATE_SHALLOW_TARBALLS = "1"
BB_GIT_SHALLOW = "1"

Could you make sure those are set in your configuration? If they're not
then the archiver may not be looking for the file names you have.

If things still don't work after that you could try backporting the
archiver class from hardknott and using the mirror archiver mode. That
mode is written to pick up the mirror tarballs from the downloads
directory so it should be able to handle your situation.

Thanks,

--
Paul Barker
https://pbarker.dev/


#archiver Archiver not archiving sources from Mirror #archiver

Seipel, Christoph
 

Hello,

 

I have a problem with the archiver. I want to use it to export the sources used in our image for license compliance (GPL etc.).

We have a BSP supplier that supplies the Yocto system and the sources to build it in a mirror directory.

 

Now, when I use the archiver (mode: original) the exported sources don’t contain the sources from the mirror, but only other files like patches.

 

To analyse this problem I put some extra logging into the archiver.bbclass to inspect the variable contents, and it seems the “local” variable where the archiver is looking for the sources to export doesn’t fit with the path of the actual sources. This leads to the archiver just skipping that URL and proceeding.

 

For example for the systemd recipe it has the variable URL=”git://github.com/systemd/systemd.git;protocol=git” and therefore it decides to look at local=”<builddir>/downloads/git2/github.com.systemd.systemd.git”

But that file doesn’t exist.

I inspected the do_fetch and do_unpack logs to find that the fetcher resolves “git://github.com/systemd/systemd.git;protocol=git” with returning the corresponding file from the mirror file:///<mirrorpath>/gitshallow_github.com.systemd.systemd.git_46659f7-1_master.tar.gz

The do_unpack task then unpacks tar -xzf <builddir>/downloads/gitshallow_github.com.systemd.systemd.git_46659f7-1_master.tar.gz

Which is possible because there is a symlink, I guess that was created by the fetcher

<builddir>/downloads/gitshallow_github.com.systemd.systemd.git_46659f7-1_master.tar.gz -> <mirror>/gitshallow_github.com.systemd.systemd.git_46659f7-1_master.tar.gz

 

So the problem seems to be that the archiver expects the sources in downloads/git2 but the fetcher creates the symlink in downloads and with additional pre- and postfiexes on the filename.

 

Sadly we are still on sumo, but at least for the archiver class there isn’t any relevant change to hardknott. At least a dunfell upgrade is incoming in the next weeks.

 

Best regards,

 

Christoph

_____________________________________________________

Christoph Seipel
 
CLAAS E-Systems GmbH

System & Software Displays, Operator Panels & Mobile Apps
Sommerkämpen 11
49201 Dissen a.T.W.
Germany
 
Phone +49 5421 9311-8762
Mail christoph.seipel@...

 


Re: #linux #dunfell How to use eclipse with autotools without yocto plugin #linux #dunfell

Sebastian Holzgreve
 

Hi,

we were currently upgrading our yocto system to v3.1. In the older yocto system there was the "Yocto Eclipse Plugin" available. I saw in the changelog that unfortunately that this plugin is not maintained anymore and removed.

From the past we got whole bunch of autotools based projects. Is it somehow posible to use eclipse for development and debugging keeping the projects in the autotools world?

I tried using the eSDK (Following: https://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#sdk-working-projects)

Rebuilding the project from eclipse is now possible. It still complains that is can run the executable but I can start debugging. I tried to setup the debugger configuration like it was set up, when the "Yocto Eclipse Plugin" generated it. But when I try to debug the code, eclipse opens up the "Debug Perspective", also attaches to the remotly running gdbserver, but it never stops on breakpoints that were set in the eclipse GUI.

Is there any chance to use eclipse without the plugin, without converting all the packages to CMake or plain makefile projects?

Any hints or pointings to the right direction is appreciated.

I found the solution: My problem was that the default project configuration which created by "./configure {CONFIGURE_FLAGS}" doesn't do a build with debug information.

So ./configure {CONFIGURE_FLAGS} CXXFLAGS="-g -O0" has to be called before, and then debug is working.


Sorry for the stupid question before.


Kind regards,

Sebastian


Re: [error-report-web][PATCH] report-error.bbclass: Add layer and bitbake version info to error report

Milan Shah
 

Thanks, Richard

Closed the Bug as Resolved Fixed Status and commit link as a Comment.

Thanks & Regards,
Milan Shah

Milan Shah | Software Engineer
a: MontaVista Software, LLC | Bangalore, India
e: info@... | w: www.mvista.com/
p: +91-80-4939-5000


On Fri, Jul 9, 2021 at 2:23 PM Richard Purdie <richard.purdie@...> wrote:
On Fri, 2021-07-09 at 14:07 +0530, Milan Shah wrote:
> Well, there was no update on the bug as well as on the patch so I was 
> not aware of it.
>
> Thanks, Richard for responding. I'll definitely follow your suggestion 
> next time.

We'd normally expect the bug owner to update the status of the bug since it
isn't really possible for me or the triage team to track and update them all.
We can close this one with a link to the commit:

http://git.yoctoproject.org/cgit.cgi/poky/commit/meta/classes/report-error.bbclass?id=8b299a62c55aa8d1583aff4b8428f16def7231b7

Happy for you to, or I can...

Cheers,

Richard



Re: [error-report-web][PATCH] report-error.bbclass: Add layer and bitbake version info to error report

Richard Purdie
 

On Fri, 2021-07-09 at 14:07 +0530, Milan Shah wrote:
Well, there was no update on the bug as well as on the patch so I was 
not aware of it.

Thanks, Richard for responding. I'll definitely follow your suggestion 
next time.
We'd normally expect the bug owner to update the status of the bug since it
isn't really possible for me or the triage team to track and update them all.
We can close this one with a link to the commit:

http://git.yoctoproject.org/cgit.cgi/poky/commit/meta/classes/report-error.bbclass?id=8b299a62c55aa8d1583aff4b8428f16def7231b7

Happy for you to, or I can...

Cheers,

Richard


Re: [error-report-web][PATCH] report-error.bbclass: Add layer and bitbake version info to error report

Milan Shah
 

Well, there was no update on the bug as well as on the patch so I was not aware of it.

Thanks, Richard for responding. I'll definitely follow your suggestion next time.

Thanks & Regards,
Milan Shah

Milan Shah | Software Engineer
a: MontaVista Software, LLC | Bangalore, India
e: info@... | w: www.mvista.com/
p: +91-80-4939-5000


On Fri, Jul 9, 2021 at 2:03 PM Richard Purdie <richard.purdie@...> wrote:
On Fri, 2021-07-09 at 10:51 +0530, Milan Shah wrote:
> Hello All,
>
> My Patch has not been reviewed for the past 7 months. Looking forward 
> to see some updates on this.
>
> Please find related links below.
> https://lists.yoctoproject.org/g/yocto/message/51891
> https://bugzilla.yoctoproject.org/show_bug.cgi?id=9700
>

http://git.yoctoproject.org/cgit.cgi/poky/commit/meta/classes/report-error.bbclass?id=8b299a62c55aa8d1583aff4b8428f16def7231b7

It appears to have merged 5 months ago?

Also, looking at the patch that inherit base should not be there and needs
to be removed since everything always inherits base.

In the future please send OE-Core changes to the openembedded-core mailing
list as that will speed up review and patch handling (as per the repo's README).

Cheers,

Richard


Re: [error-report-web][PATCH] report-error.bbclass: Add layer and bitbake version info to error report

Richard Purdie
 

On Fri, 2021-07-09 at 10:51 +0530, Milan Shah wrote:
Hello All,

My Patch has not been reviewed for the past 7 months. Looking forward 
to see some updates on this.

Please find related links below.
https://lists.yoctoproject.org/g/yocto/message/51891
https://bugzilla.yoctoproject.org/show_bug.cgi?id=9700
http://git.yoctoproject.org/cgit.cgi/poky/commit/meta/classes/report-error.bbclass?id=8b299a62c55aa8d1583aff4b8428f16def7231b7

It appears to have merged 5 months ago?

Also, looking at the patch that inherit base should not be there and needs
to be removed since everything always inherits base.

In the future please send OE-Core changes to the openembedded-core mailing
list as that will speed up review and patch handling (as per the repo's README).

Cheers,

Richard


#linux #dunfell How to use eclipse with autotools without yocto plugin #linux #dunfell

Sebastian Holzgreve
 

Hi,

we were currently upgrading our yocto system to v3.1. In the older yocto system there was the "Yocto Eclipse Plugin" available. I saw in the changelog that unfortunately that this plugin is not maintained anymore and removed.

From the past we got whole bunch of autotools based projects. Is it somehow posible to use eclipse for development and debugging keeping the projects in the autotools world?

I tried using the eSDK (Following: https://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#sdk-working-projects)

Rebuilding the project from eclipse is now possible. It still complains that is can run the executable but I can start debugging. I tried to setup the debugger configuration like it was set up, when the "Yocto Eclipse Plugin" generated it. But when I try to debug the code, eclipse opens up the "Debug Perspective", also attaches to the remotly running gdbserver, but it never stops on breakpoints that were set in the eclipse GUI.

Is there any chance to use eclipse without the plugin, without converting all the packages to CMake or plain makefile projects?

Any hints or pointings to the right direction is appreciated.


Re: [error-report-web][PATCH] report-error.bbclass: Add layer and bitbake version info to error report

Milan Shah
 

Hello All,

My Patch has not been reviewed for the past 7 months. Looking forward to see some updates on this.

Please find related links below.

Thanks & Regards,
Milan Shah

Milan Shah | Software Engineer
a: MontaVista Software, LLC | Bangalore, India
e: info@... | w: www.mvista.com/
p: +91-80-4939-5000


On Fri, Mar 26, 2021 at 5:16 PM Milan Shah via lists.yoctoproject.org <mshah=mvista.com@...> wrote:
Thanks, Richard for responding.

I completely understood your point and appreciate the effort you people make to maintain these repositories.

But as a contributor, It encourages me to contribute more when my work gets noticed/criticized/accepted by the community.

Anyways, I hope that the community resolves these issues as soon as possible so that there won't be much delay in updating the patches sent by the contributors.

Thanks & Regards,
Milan Shah

Milan Shah | Software Engineer
a: MontaVista Software, LLC | Bangalore, India
e: info@... | w: www.mvista.com/
p: +91-80-4939-5000


On Fri, Mar 26, 2021 at 4:21 PM Richard Purdie <richard.purdie@...> wrote:
On Fri, 2021-03-26 at 16:04 +0530, Milan Shah wrote:
> It seems like patches sent for error-report-web are not much important to the community as this patch has
> not been reviewed/commented on by anyone since 6th January 2021 (more than 2.5 months).
>
> Still, I'm sending a reminder one last time for [YOCTO #9700].
> https://lists.yoctoproject.org/g/yocto/message/51891
>
> Please review/comment on it so that things can move further.

Sorry, that isn't true, they are important and appreciated. We're 
just struggling to handle everything.

The challenge/problem has been that there are much older patches which
were half merged and broke things. People have rightly been complaining
about those issues and Khem for example has been carrying a patch for 3
years to work arond that problem. I did figure out a fix for that which
has now been merged.

We also needed to migrate the service from py2 to py3 and deal with a
set of patches that had been applied locally that we had been unaware of
until very recently but which complicate things. The people who wrote 
them are no longer around to help with that.

The original authors of this code are no longer around, the people who
stepped up to try and help also haven't had time so the support/review
has struggled. The service is in daily use however so it is important
and that is why I'm now trying to unravel and fix things.

I do want to get to your patch but you can appreciate that adding it
directly on top of an already failing set of changes wouldn't be good
for anyone.

We are making progress in resolving the issues so I hope to get some
kind of better answer to you soon.

Cheers,

Richard






Re: binman support in u-boot

Tom Rini
 

On Thu, Jul 08, 2021 at 04:10:32AM +0000, Rebecca Chang Swee Fun wrote:

Hi,

I was having a bbappend for uboot v2021.04 that calls binman command from u-boot source tree but it doesn't seems to work with bitbake with an error of unknown location of libfdt.
I have added dependency to dtc-native or did I missed anything? Is there any plan to enable binman support within u-boot-tools recipe?

| running build
| running build_py
| package init file '__init__.py' not found (or not a regular file)
| package init file 'etype/__init__.py' not found (or not a regular file)
| package init file '__init__.py' not found (or not a regular file)
| package init file 'etype/__init__.py' not found (or not a regular file)
| running build_scripts
| binman: cannot import name 'QUIET_NOTFOUND' from 'libfdt' (unknown location)
| WARNING: /build/yocto-stable/agilex-gsrd-rootfs/tmp/work/agilex-poky-linux/u-boot-socfpga/1_v2021.04+gitAUTOINC+396ee271f2-r0/temp/run.do_compile.8579:158 exit 1 from './tools/binman/binman build -u -d /build/yocto-stable/agilex-gsrd-rootfs/tmp/work/agilex-poky-linux/u-boot-socfpga/1_v2021.04+gitAUTOINC+396ee271f2-r0/build/${config}/u-boot.dtb -O . -i u-boot'
| WARNING: Backtrace (BB generated script):
| #1: do_compile, /build/yocto-stable/agilex-gsrd-rootfs/tmp/work/agilex-poky-linux/u-boot-socfpga/1_v2021.04+gitAUTOINC+396ee271f2-r0/temp/run.do_compile.8579, line 158
| #2: main, /build/yocto-stable/agilex-gsrd-rootfs/tmp/work/agilex-poky-linux/u-boot-socfpga/1_v2021.04+gitAUTOINC+396ee271f2-r0/temp/run.do_compile.8579, line 184
|
| Backtrace (metadata-relative locations):
| #1: do_compile, autogenerated, line 7
ERROR: Task (/build/yocto-stable/agilex-gsrd-rootfs/../meta-intel-fpga/recipes-bsp/u-boot/u-boot-socfpga_v2021.04.bb:do_compile) failed with exit code '1'
NOTE: Tasks Summary: Attempted 2114 tasks of which 2111 didn't need to be rerun and 1 failed.

bbappend:
DEPENDS += "bc-native dtc-native python3-setuptools-native swig-native"

# This is require so that tools/binman/setup.py is discoverable during the build
DISTUTILS_SETUP_PATH = "${S}/tools/binman"
distutils3_do_compile() {
cd ${DISTUTILS_SETUP_PATH}
NO_FETCH_BUILD=1 \
STAGING_INCDIR=${STAGING_INCDIR} \
STAGING_LIBDIR=${STAGING_LIBDIR} \
${STAGING_BINDIR_NATIVE}/${PYTHON_PN}-native/${PYTHON_PN} setup.py \
build --build-base=${B} ${DISTUTILS_BUILD_ARGS} || \
bbfatal_log "'${PYTHON_PN} setup.py build ${DISTUTILS_BUILD_ARGS}' execution failed."
}

do_compile_append() {
cp ${DEPLOY_DIR_IMAGE}/bl31.bin ${B}/${config}/bl31.bin
cd ${S}
./tools/binman/binman build -u -d ${B}/${config}/u-boot.dtb -O . -i u-boot
}
Off-hand, I suspect it might be a little tricky to use binman outside of
the U-Boot context directly. Simon, any thoughts?

--
Tom


valgrind failed do_compile during bitbake

Michael
 

I'm using Yocto "warrior" version.

It is setup for powerpc-nf
DEFAULTTUNE = "powerpc-nf"
DISTRO = "poky"

Everything seems to work find until I enable "valgrind".  'bitbake' fails to compile valgrind

************************
:
| collect2: error: ld returned 1 exit status
| make[3]: *** [Makefile:1034: memcheck-ppc32-linux] Error 1
| make[3]: Leaving directory '${WORK_DIR}/powerpc-nf-poky-linux/valgrind/3.14.0-r0/build/memcheck'
| make[2]: *** [Makefile:1345: all-recursive] Error 1
| make[2]: Leaving directory '${WORK_DIR}/powerpc-nf-poky-linux/valgrind/3.14.0-r0/build/memcheck'
| make[1]: *** [Makefile:834: all-recursive] Error 1
| make[1]: Leaving directory '${WORK_DIR}/work/powerpc-nf-poky-linux/valgrind/3.14.0-r0/build'
| make: *** [Makefile:703: all] Error 2
| make: Leaving directory '${WORK_DIR}/powerpc-nf-poky-linux/valgrind/3.14.0-r0/build'
| ERROR: oe_runmake failed
| WARNING: ${WORK_DIR}/powerpc-nf-poky-linux/valgrind/3.14.0-r0/temp/run.do_compile.23806:1 exit 1 from 'exit 1'
| ERROR: Function failed: do_compile (log file is located at ${WORK_DIR}/powerpc-nf-poky-linux/valgrind/3.14.0-r0/temp/log.do_compile.23806)
ERROR: Task (${REPODIR}fips-build/meta/recipes-devtools/valgrind/valgrind_3.14.0.bb:do_compile) failed with exit code '1'
NOTE: Tasks Summary: Attempted 4091 tasks of which 4089 didn't need to be rerun and 1 failed.

Summary: 1 task failed:
  ${REPO_DIR}/fips-build/meta/recipes-devtools/valgrind/valgrind_3.14.0.bb:do_compile
Summary: There were 2 ERROR messages shown, returning a non-zero exit code.
*********************************************

Many link failure...
**********************************
 undefined reference to `__floatsidf'
undefined reference to `__muldf3'
undefined reference to `__ltdf2'
undefined reference to `__fixdfsi'
undefined reference to `__divdf3'
undefined reference to `__floatunsidf'
undefined reference to `__floatundidf'
undefined reference to `__adddf3'
undefined reference to `__fixunsdfdi'
undefined reference to `__nedf2'
undefined reference to `__gtdf2'
undefined reference to `__subdf3'
undefined reference to `__eqsf2'
undefined reference to `__eqdf2'
undefined reference to `__floatunsisf'
undefined reference to `__divsf3'

***********************

Seems like many people were running into this issue for many reason.  Saw a lot of suggestion in google, but seem like there is no real solution...
I try to create valgrind_%.bbappen with the following contents.  And did not help...
#####################
#valgrind will not compile even with below changes
CFLAGS_prepend = " -lc -lm "
CFLAGS_append = " -ffast-math -static-libgcc -lgcc -ffreestanding -lc -lm "
#BASE_WORKDIR="/tmp/youhour/fips-build/m20210707a/fips-ppcnf/build/tmp/work"
#./powerpc-nf-poky-linux/valgrind/3.14.0-r0/recipe-sysroot/usr/lib/powerpc-poky-linux/8.3.0/libgcc.a
BUILD_LDFLAGS_append = " -L${BASE_WORKDIR}/powerpc-nf-poky-linux/valgrind/3.14.0-r0/recipe-sysroot/usr/lib/powerpc-poky-linux/8.3.0 "
#####################

Any suggestion is appreciated.



Installing a .deb produced by npm during build

Matt Bernhard
 

Hi,

I've been struggling to add a debian package to my build. I'm trying to build this app and install it on the target: https://github.com/votingworks/kiosk-browser

It's a node app built with yarn, and I've found the instructions in the manual to be a bit off the mark for what I'm trying to do. I want to:
  1. Fetch and build the app using its own makefile
  2. Install the output debian package in the target during buildtime
My recipe is below, but as best I can tell all it does right now is package the source directory and put it on the image. DPKG reports that it's installed, but I can't seem to run the app. I'm sure I'm confused a bit about the build process, and perhaps I'm not invoking a task when I should?

Any pointers would be greatly appreciated.

All the best,
Matt

P.S. Here's the recipe:

# Recipe created by recipetool
# This is the basis of a recipe and may need further editing in order to be fully functional.
# (Feel free to remove these comments when editing.)

SUMMARY = "Generic kiosk-mode browser."
# WARNING ommitted for brevity ...

LICENSE = "Unknown & MIT & ISC"
LIC_FILES_CHKSUM = "file://LICENSE;md5=e49f4652534af377a713df3d9dec60cb \
... Licensing stuff that I'm omitting for brevity ... file://node_modules/xrandr-parse/package.json;md5=671eb8501a73fbcd413813c4832bb560" SRC_URI = " \ git://github.com/votingworks/kiosk-browser.git;protocol=https \ npmsw://${THISDIR}/${BPN}/npm-shrinkwrap.json \ " # Modify these as desired PV = "1.0.0+git${SRCPV}" SRCREV = "c85708f50747c595878d4bc6da6edf72203f87ae" S = "${WORKDIR}/git" inherit npm

RDEPENDS_${PN} = "\
bash \
"

DEPENDS = "\
libusb \
dpkg \
" LICENSE_${PN} = "Unknown"
... Licensing stuff that I'm omitting for brevity ...
EOF


Yocto Autobuilder: Latency Monitor and AB-INT - Meeting notes: July 8, 2021

Randy MacLeod
 

YP AB Intermittent failures meeting
===================================
July 1, 2021, 9 AM ET
https://windriver.zoom.us/j/3696693975

Attendees: Tony, Richard, Trevor, Randy


Summary:

========



The autobuilder RCU hang is still fixed ;-),
master branch builds greener.

ptest failures are the top problem now.



Add Michael Halstead, see questions below in section 4.



If anyone wants to help, we could use more eyes on the logs,

particularly the summary logs and understanding iostat #

when the dd test times out.





Plans for the week:



===================





Richard: glibc upgrade, etc.



Alex: ?



Sakib: pub/non-release link upgrade, script clean-up.



Trevor: make job server test. Try it on YP AB!!! What type of build?



Tony: fix/work-around valgrind ptest bug:
none/tests/amd64/fb_test_amd64
Saul: nothing this week for YP.



Randy: vacation, then email catch-up!




Meeting Notes:

==============


1, runqemu

Tony having trouble with runqemu on some Wind River machine.

Richard has a fix for a race in runqemu in master-next.

These might be related but if not Tony should debug the
issue/ collect logs.

2. job server

- Trevor has conclusive evidence that the 'make' job server is useful.
email summary to come. Need to fix some assumptoins in code that
parses PARALLEL_MAKE then send patch to yocto-autobuilder-helper.

- ninja will have to be done next.


3. AB status

generally better but...

ptests are having some recurring problems.
- parted - only on arm?
- valgrind - none/tests/amd64/fb_test_amd64
- gdb test failing again. - Randy!

4. Richard reported
- something really flaky going on with serial ports.
- particularly bad on qemuppc but also x86.
- related to Saul's QMP data dump?

5. Sakib needs to send patch to make testimage failures
generate summary logs.

6. Richard says that we may need to redesign the data collection system
that Sakib's AB INT tests are based on.





Still relevant parts of
Previous Meeting Notes:
=======================

1. The qemu RCU hang has been fixed to not deadlock anymore!

It still hangs at times but this dramatically reduces the
AB failures.



4. bitbake server timeout.

"Timeout while waiting for a reply from the bitbake server (60s)"

Randy mentioned that the bitbake server timeouts seen in the
Wind River build cluster have gone away after upgrading to
a newer version of docker.

Old: Docker Version: Docker version 18.09.4, build d14af54266
New: Docker Version: Docker version 20.10.7, build f0df350


Clearly the YP ABs aren't running in docker but what
about firmware and kernel tunings.

Michael,

Is the BIOS/firmware kept up to date on most nodes?

It seems that we are running stock kernels which makes sense but
given that we don't have concerns about privacy since system access
is controlled and the nodes are being used to test open
source software, we might consider optimizing for performance
rather than security.

Alex pointed at: https://make-linux-fast-again.com/
Which just lists a set of kernel boot options:
noibrs noibpb nopti nospectre_v2 nospectre_v1 \
l1tf=off nospec_store_bypass_disable no_stf_barrier \
mds=off tsx=on tsx_async_abort=off mitigations=off

Can we enable some or all of these on a node to see what the
performance difference is?


5. io stalls

Richard said that it would make sense to write an ftrace utility
/ script to monitor io latency and we could install it with sudo
Ch^W mentioned ftrace on IRC.
Sakib and Randy will work on that but not for a week or two.

6. Switch the pub/non-release links from full log to summary.
The host data links on:
https://autobuilder.yocto.io/pub/non-release/
should include links to the summary data. I think we have room to
include both links like this:
0 1 2 3 (Full: 0 1 2 3 )





../Randy


Re: Question regarding custom device tree update

Sohil Shah <sohils@...>
 

Hi Robert,

Thank you for your reply.
This works and I got my dtb generated in my build directory!

Best wishes,
Sohil


On Wed, Jul 7, 2021 at 10:29 PM Robert Calhoun <rcalhoun@...> wrote:
> From: yocto@... <yocto@...> on behalf of Sohil Shah <sohils@...>
> Sent: Wednesday, June 30, 2021 09:46
> To: yocto@... <yocto@...>
> Subject: [yocto] Question regarding custom device tree update
>
> (...)
> I want to build the image using my custom dts file where I enable certain peripherals and disable the ones not required. (A test to update dtb's in future).
>
> I tried different methods found here
> https://stackoverflow.com/questions/38917745/quick-rebuild-of-device-tree-only-with-yocto-bitbake
>
> But, I seem to run into some errors when I try to build the image.
>
> Please help and let me know if I missed any required information from my side.
>
>Thank you and Regards,
> Sohil

Hi Sohil,

First make a custom layer for you modifications and add it to bblayers.conf. Then modify your machine definition of KERNEL_DEVICETREE (in meta-yourlayer/conf/machine/machinname.conf) to specify multiple device trees, e.g.

KERNEL_DEVICETREE = "at91-sama5d27_wlsom1_ek.dtb at91-sama5d27_wlsom1_ek_custom.dtb"

Note this should specify the "dtb" (compiled device tree name), not the "dts" (source.)

Next, make a .bbappend for your linux recipe that will provide the .dts source, something like:

FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:"

SRC_URI += " \
 file://at91-sama5d27_wlsom1_ek_custom.dts;subdir=git/arch/arm/boot/dts \
"

Make a subdirectory with package name (e.g. linux-yocto) and put your dts in there. The "subdir" directive tells bitbake to drop the dts into the appropriate directory for the linux build process to find it. The device tree(s) will be built along with linux and put in the image's /boot directory. You can select which device tree you want to boot with via u-boot, or by manipulating the symbolic links in the /boot directory.

Best wishes,

Rob Calhoun


Re: [meta-mingw][PATCH] openssl: support for building nativesdk of mingw

Changqing Li
 


On 7/7/21 1:23 AM, Joshua Watt wrote:

[Please note: This e-mail is from an EXTERNAL e-mail address]


On 1/4/21 9:31 PM, Changqing Li wrote:
* add support for mingw32
* Engines are installed in a slightly different path, which is
  urgly, patch it to make the path shorter
* remove runtime dependency from perl for mingw nativesdk

since commit 70da1f956bfbb627691c47eba7451182aca758e3 of oe-core
'openssl: Add c_rehash to misc package and add perl runtime dependency'

package openssl-misc have runtime dependency on perl, and perl then
have depenency on another 3 recipes, db/gdbm/libxcrypt. according to
http://arsv.github.io/perl-cross/usage.html, perl don't support
cross-compile build for mingw32 and another 3 recipes also don't
support mingw well. so remove the dependency of perl, don't support
c_rehash for mingw.


It would appear that some or all of this patch is unnecessary. OE-core 166bb89f6d97495b6522786182b4f9623acd7ff4 implements part of this patch, which makes me think it's working there without any changes necessary. It would be worth following up to see if that is the case.

Thanks.  I will check the oe-core commit.



Signed-off-by: Changqing Li <changqing.li@...>
---
 ...ile.tmpl-don-t-add-prefix-for-libdir.patch | 32 +++++++++++++++++++
 .../openssl/openssl_%.bbappend                | 31 ++++++++++++++++++
 2 files changed, 63 insertions(+)
 create mode 100644 recipes-connectivity/openssl/files/0001-unix-Makefile.tmpl-don-t-add-prefix-for-libdir.patch
 create mode 100644 recipes-connectivity/openssl/openssl_%.bbappend

diff --git a/recipes-connectivity/openssl/files/0001-unix-Makefile.tmpl-don-t-add-prefix-for-libdir.patch b/recipes-connectivity/openssl/files/0001-unix-Makefile.tmpl-don-t-add-prefix-for-libdir.patch
new file mode 100644
index 0000000..028431b
--- /dev/null
+++ b/recipes-connectivity/openssl/files/0001-unix-Makefile.tmpl-don-t-add-prefix-for-libdir.patch
@@ -0,0 +1,32 @@
+From 8fe5c9421acfaff35b637e7ad55d1df598bb7081 Mon Sep 17 00:00:00 2001
+From: Changqing Li <changqing.li@...>
+Date: Tue, 22 Dec 2020 09:22:10 +0800
+Subject: [PATCH] unix-Makefile.tmpl: don't add prefix for libdir
+
+we had pass libdir to Configure, don't use prefix again to
+avoid engineer dir set to:
+/opt/poky/3.2+snapshot/sysroots/x86_64-w64-mingw32/usr/opt/poky/3.2+snapshot/sysroots/x86_64-w64-mingw32/usr/lib/engines-1_1
+
+Upstream-Status: Inappropriate[oe-specific]
+
+Signed-off-by: Changqing Li <changqing.li@...>
+---
+ Configurations/unix-Makefile.tmpl | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/Configurations/unix-Makefile.tmpl b/Configurations/unix-Makefile.tmpl
+index bbafb98..eecb63e 100644
+--- a/Configurations/unix-Makefile.tmpl
++++ b/Configurations/unix-Makefile.tmpl
+@@ -244,7 +244,7 @@ LIBDIR={- our $libdir = $config{libdir} || "lib";
+           File::Spec::Win32->file_name_is_absolute($libdir) ? "" : $libdir -}
+ ENGINESDIR_dev={- use File::Spec::Win32;
+                   our $enginesdir =
+-                      File::Spec::Win32->catdir($prefix,$libdir,
++                      File::Spec::Win32->catdir($libdir,
+                                                 "engines-$sover_dirname");
+                   our ($enginesdir_dev, $enginesdir_dir, $enginesdir_file) =
+                       File::Spec::Win32->splitpath($enginesdir, 1);
+-- 
+2.17.1
+
diff --git a/recipes-connectivity/openssl/openssl_%.bbappend b/recipes-connectivity/openssl/openssl_%.bbappend
new file mode 100644
index 0000000..7fd82f1
--- /dev/null
+++ b/recipes-connectivity/openssl/openssl_%.bbappend
@@ -0,0 +1,31 @@
+FILESEXTRAPATHS_prepend := "${THISDIR}/files:"
+
+SRC_URI_append_mingw32_class-nativesdk = " \
+           file://0001-unix-Makefile.tmpl-don-t-add-prefix-for-libdir.patch \
+"
+
+do_configure_mingw32 () {
+	os=${HOST_OS}
+	target="$os-${HOST_ARCH}"
+	case $target in
+        mingw32-x86_64)
+                target=mingw64
+                ;;
+        mingw32-i686)
+                target=mingw
+                ;;
+        esac
+
+        useprefix=${prefix}
+        if [ "x$useprefix" = "x" ]; then
+                useprefix=/
+        fi
+        # WARNING: do not set compiler/linker flags (-I/-D etc.) in EXTRA_OECONF, as they will fully replace the
+        # environment variables set by bitbake. Adjust the environment variables instead.
+        HASHBANGPERL="/usr/bin/env perl" PERL=perl PERL5LIB="${S}/external/perl/Text-Template-1.46/lib/" \
+        perl ${S}/Configure ${EXTRA_OECONF} ${PACKAGECONFIG_CONFARGS} --prefix=$useprefix --openssldir=${libdir}/ssl-1.1 --libdir=${libdir} $target
+        perl ${B}/configdata.pm --dump
+}
+
+FILES_${PN}-engines_mingw32_class-nativesdk = "${libdir}/engines-1_1"
+RDEPENDS_${PN}-misc_remove_mingw32_class-nativesdk = "perl"



[meta-raspberrypi][PATCH] userland: add dtc dependency

massimo toscanelli
 

When using dtparam, dtoverlay converts /proc/device-tree to a .dtb and
loads it. However, if dtc is not installed, this operation cannot be
done and every call to dtparam fails.

Therefore, dtc needs to be added to the list of userland dependencies
in order to properly call dtparam.

Signed-off-by: massimo toscanelli <massimo.toscanelli@...>
---
recipes-graphics/userland/userland_git.bb | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/recipes-graphics/userland/userland_git.bb b/recipes-graphics/userland/userland_git.bb
index ead81f0..9ec57c9 100644
--- a/recipes-graphics/userland/userland_git.bb
+++ b/recipes-graphics/userland/userland_git.bb
@@ -104,5 +104,5 @@ FILES_${PN}-dev += "${includedir} \
FILES_${PN}-doc += "${datadir}/install"
FILES_${PN}-dbg += "${libdir}/plugins/.debug"

-RDEPENDS_${PN} += "bash"
+RDEPENDS_${PN} += "bash dtc"
RDEPENDS_${PN} += "${@bb.utils.contains("MACHINE_FEATURES", "vc4graphics", "libegl-mesa", "", d)}"
--
2.17.1


binman support in u-boot

Rebecca Chang Swee Fun
 

Hi,

 

I was having a bbappend for uboot v2021.04 that calls binman command from u-boot source tree but it doesn’t seems to work with bitbake with an error of unknown location of libfdt.

I have added dependency to dtc-native or did I missed anything? Is there any plan to enable binman support within u-boot-tools recipe?

 

| running build

| running build_py

| package init file '__init__.py' not found (or not a regular file)

| package init file 'etype/__init__.py' not found (or not a regular file)

| package init file '__init__.py' not found (or not a regular file)

| package init file 'etype/__init__.py' not found (or not a regular file)

| running build_scripts

| binman: cannot import name 'QUIET_NOTFOUND' from 'libfdt' (unknown location)

| WARNING: /build/yocto-stable/agilex-gsrd-rootfs/tmp/work/agilex-poky-linux/u-boot-socfpga/1_v2021.04+gitAUTOINC+396ee271f2-r0/temp/run.do_compile.8579:158 exit 1 from './tools/binman/binman build -u -d /build/yocto-stable/agilex-gsrd-rootfs/tmp/work/agilex-poky-linux/u-boot-socfpga/1_v2021.04+gitAUTOINC+396ee271f2-r0/build/${config}/u-boot.dtb -O . -i u-boot'

| WARNING: Backtrace (BB generated script):

|       #1: do_compile, /build/yocto-stable/agilex-gsrd-rootfs/tmp/work/agilex-poky-linux/u-boot-socfpga/1_v2021.04+gitAUTOINC+396ee271f2-r0/temp/run.do_compile.8579, line 158

|       #2: main, /build/yocto-stable/agilex-gsrd-rootfs/tmp/work/agilex-poky-linux/u-boot-socfpga/1_v2021.04+gitAUTOINC+396ee271f2-r0/temp/run.do_compile.8579, line 184

|

| Backtrace (metadata-relative locations):

|       #1: do_compile, autogenerated, line 7

ERROR: Task (/build/yocto-stable/agilex-gsrd-rootfs/../meta-intel-fpga/recipes-bsp/u-boot/u-boot-socfpga_v2021.04.bb:do_compile) failed with exit code '1'

NOTE: Tasks Summary: Attempted 2114 tasks of which 2111 didn't need to be rerun and 1 failed.

 

bbappend:

DEPENDS += "bc-native dtc-native python3-setuptools-native swig-native"

 

# This is require so that tools/binman/setup.py is discoverable during the build

DISTUTILS_SETUP_PATH = "${S}/tools/binman"

distutils3_do_compile() {

        cd ${DISTUTILS_SETUP_PATH}

        NO_FETCH_BUILD=1 \

        STAGING_INCDIR=${STAGING_INCDIR} \

        STAGING_LIBDIR=${STAGING_LIBDIR} \

        ${STAGING_BINDIR_NATIVE}/${PYTHON_PN}-native/${PYTHON_PN} setup.py \

        build --build-base=${B} ${DISTUTILS_BUILD_ARGS} || \

        bbfatal_log "'${PYTHON_PN} setup.py build ${DISTUTILS_BUILD_ARGS}' execution failed."

}

 

do_compile_append() {

                cp ${DEPLOY_DIR_IMAGE}/bl31.bin ${B}/${config}/bl31.bin

                cd ${S}

                ./tools/binman/binman build -u -d ${B}/${config}/u-boot.dtb -O . -i u-boot

}


[yocto-autobuilder-helper][dunfell 0/1] Pull request (cover letter only)

Steve Sakoman
 

The following changes since commit 1e5c23dea6d989f0e5c30ebac2c95a17e78ffdf9:

config.json: Add check-layer-nightly with only meta-intel at this point (2021-06-02 12:33:34 +0100)

are available in the Git repository at:

git://git.yoctoproject.org/yocto-autobuilder-helper contrib/sakoman
http://git.yoctoproject.org/cgit.cgi/yocto-autobuilder-helper/log/?h=contrib/sakoman

Richard Purdie (1):
config.json: Switch to newer buildtools-extended-tarball

config.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

--
2.25.1


Re: Question regarding custom device tree update

Rob Calhoun
 

From: yocto@... <yocto@...> on behalf of Sohil Shah <sohils@...>
Sent: Wednesday, June 30, 2021 09:46
To: yocto@... <yocto@...>
Subject: [yocto] Question regarding custom device tree update

(...)
I want to build the image using my custom dts file where I enable certain peripherals and disable the ones not required. (A test to update dtb's in future).

I tried different methods found here
https://stackoverflow.com/questions/38917745/quick-rebuild-of-device-tree-only-with-yocto-bitbake

But, I seem to run into some errors when I try to build the image.

Please help and let me know if I missed any required information from my side.

Thank you and Regards,
Sohil
Hi Sohil,

First make a custom layer for you modifications and add it to bblayers.conf. Then modify your machine definition of KERNEL_DEVICETREE (in meta-yourlayer/conf/machine/machinname.conf) to specify multiple device trees, e.g.

KERNEL_DEVICETREE = "at91-sama5d27_wlsom1_ek.dtb at91-sama5d27_wlsom1_ek_custom.dtb"

Note this should specify the "dtb" (compiled device tree name), not the "dts" (source.)

Next, make a .bbappend for your linux recipe that will provide the .dts source, something like:

FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:"

SRC_URI += " \
file://at91-sama5d27_wlsom1_ek_custom.dts;subdir=git/arch/arm/boot/dts \
"

Make a subdirectory with package name (e.g. linux-yocto) and put your dts in there. The "subdir" directive tells bitbake to drop the dts into the appropriate directory for the linux build process to find it. The device tree(s) will be built along with linux and put in the image's /boot directory. You can select which device tree you want to boot with via u-boot, or by manipulating the symbolic links in the /boot directory.

Best wishes,

Rob Calhoun

3721 - 3740 of 57773