Hi,

I'm trying to add a recipe for:
https://github.com/bootlin/libva-v4l2-request package. This library
heavily depends on actual kernel headers (see
PR.https://github.com/bootlin/libva-v4l2-request). I've bumped kernel
to 5.14-rc7 nad in recipe I'm trying to use what is suggested:

# This is how linux-libc-headers says to include custom uapi headers
CFLAGS_append += " -I ${STAGING_KERNEL_DIR}/include/uapi -I
${STAGING_KERNEL_DIR}/include "
do_configure[depends] += "virtual/kernel:do_shared_workdir"

Problem is that I see a lot of errors like:
kernel-source/include/linux/cache.h:6:10: fatal error: asm/cache.h: No
such file or directory
| 6 | #include <asm/cache.h>

or
recipe-sysroot/usr/include/bits/stdint-intn.h:27:19: note: previous
declaration of 'int64_t' was here
| 27 | typedef __int64_t int64_t;

I tried to compile this package manually on the host only difference
is that I used headers from 5.14 by make headers_install and
everything compiled fine. Any ideas on how to fix this issue?

Thanks and BR,

marek


--
as simple and primitive as possible
-------------------------------------------------
Marek Belisko - OPEN-NANDRA
Freelance Developer

Ruska Nova Ves 219 | Presov, 08005 Slovak Republic
Tel: +421 915 052 184
skype: marekwhite
twitter: #opennandra
web: http://open-nandra.com

Hello,

So have anyone been successful at making a pyinstaller recipe in yocto? Or add pyinstaller in the yocto image? 
Please I need help

A number of typo fixes:
- tmp->tpm in the DISTRO_FEATURES
- update the mailing list address as it was out of date
- update the distro name in the subject

Signed-off-by: Marta Rybczynska <rybczynska@...>
---
meta-tpm/README | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/meta-tpm/README b/meta-tpm/README
index 4441dd2..2b32a9f 100644
--- a/meta-tpm/README
+++ b/meta-tpm/README
@@ -5,7 +5,7 @@ The bbappend files for some recipes (e.g. linux-yocto) in this layer need
to have 'tpm' in DISTRO_FEATURES to have effect.
To enable them, add in configuration file the following line.

- DISTRO_FEATURES:append = " tmp"
+ DISTRO_FEATURES:append = " tpm"

If meta-tpm is included, but tpm is not enabled as a
distro feature a warning is printed at parse time:
@@ -57,15 +57,15 @@ other layers needed. e.g.:
Maintenance
-----------

-Send pull requests, patches, comments or questions to yocto@...
+Send pull requests, patches, comments or questions to yocto@...

When sending single patches, please using something like:
-'git send-email -1 --to yocto@... --subject-prefix=meta-security][PATCH'
+'git send-email -1 --to yocto@... --subject-prefix=meta-tpm][PATCH'

These values can be set as defaults for this repository:

-$ git config sendemail.to yocto@...
-$ git config format.subjectPrefix meta-security][PATCH
+$ git config sendemail.to yocto@...
+$ git config format.subjectPrefix meta-tpm][PATCH

Now you can just do 'git send-email origin/master' to send all local patches.

--
2.30.2

A number of typo fixes:
- tmp->tpm in the DISTRO_FEATURES
- update the mailing list address as it was out of date
- update the distro name in the subject

Signed-off-by: Marta Rybczynska <rybczynska@...>
---
meta-tpm/README | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/meta-tpm/README b/meta-tpm/README
index 4441dd2..5722a92 100644
--- a/meta-tpm/README
+++ b/meta-tpm/README
@@ -5,7 +5,7 @@ The bbappend files for some recipes (e.g. linux-yocto) in this layer need
to have 'tpm' in DISTRO_FEATURES to have effect.
To enable them, add in configuration file the following line.

- DISTRO_FEATURES:append = " tmp"
+ DISTRO_FEATURES:append = " tpm"

If meta-tpm is included, but tpm is not enabled as a
distro feature a warning is printed at parse time:
@@ -57,14 +57,14 @@ other layers needed. e.g.:
Maintenance
-----------

-Send pull requests, patches, comments or questions to yocto@...
+Send pull requests, patches, comments or questions to yocto@...

When sending single patches, please using something like:
-'git send-email -1 --to yocto@... --subject-prefix=meta-security][PATCH'
+'git send-email -1 --to yocto@... --subject-prefix=meta-security][PATCH'

These values can be set as defaults for this repository:

-$ git config sendemail.to yocto@...
+$ git config sendemail.to yocto@...
$ git config format.subjectPrefix meta-security][PATCH

Now you can just do 'git send-email origin/master' to send all local patches.
--
2.30.2

Hi! 

I would like to implement an out-of-tree  ASoC Machine driver in my project. It #includes some in-tree header files such ../codecs/codecX.h. These are however not found when I try to build the recipe. I have followed the example project "hello-mod", but that doesn't require any in-tree header files, so it doesn't help in this case. 

What do I need to do in oder for Bitbake to find these headers ?
Thanks!

Hi Sergei,

Thanks for the advice, the latest ofono in oe-core master is 1.32, but
I am in branch Hardknott, so I copied ofono 1.32 from master to my
layer running Hardknott branch, but I could not build ofono 1.32 in
Hardknott:

ERROR: ParseError at
/build/oe-cor/../meta-solar/recipes-connectivity/ofono/ofono_1.32.bb:22:
unparsed line: 'SYSTEMD_SERVICE:${PN} = "ofono.service"'

I used ofono 1.30 and connman 1.37 in Zeus, both worked fine , but
both connman 1.39 and ofono 1.31 failed to work properly in oe-core
Hardknott branch, to narrow down if it is ofono issue or Yocto OE
build or other library issue, I was able to build ofono 1.30 for
Hardknott, I can confirm that ofono 1.30 failed just like the 1.31
stuck at the statement add_serial_device.

................
ofonod[2806]: ../ofono-1.30/plugins/udevng.c:add_serial_device()
Device is missing required OFONO_DRIVER property
ofonod[2806]: ../ofono-1.30/plugins/udevng.c:add_serial_device()
Device is missing required OFONO_DRIVER property
ofonod[2806]: ../ofono-1.30/plugins/udevng.c:add_serial_device()
Device is missing required OFONO_DRIVER property

If I run ofono 1.30 built from Zeus build, the ofonod would pass
missing OFONO_DRIVER property and kept it work:

......
ofonod[3712]: ../ofono-1.30/plugins/udevng.c:add_serial_device()
Device is missing required OFONO_DRIVER property
ofonod[3712]: ../ofono-1.30/plugins/udevng.c:create_modem()
/sys/devices/soc0/soc/2100000.aips-bus/2184200.usb/ci_hdrc.1/usb1/1-1
ofonod[3712]: ../ofono-1.30/plugins/udevng.c:create_modem() driver=ubloxqmi
ofonod[3712]: ../ofono-1.30/src/modem.c:ofono_modem_create() name:
(null), type: ubloxqmi
ofonod[3712]: ../ofono-1.30/plugins/udevng.c:setup_ubloxqmi()
/sys/devices/soc0/soc/2100000.aips-bus/2184200.usb/ci_hdrc.1/usb1/1-1
.....

So I don't think it is an oFono issue, but rather an OE Yocto build or
library issue, can anyone run oFono in oe-core Hardknott branch?

Thank you.

Kind regards,

- jupiter



On 8/22/21, Sergei Golubtsov <s.e.golubtsov@...> wrote:

Hi jupiter,

Please try to update ofono to the latest release. Seems to be a library
incompatibility issue or something like that.

Yours sincerely,
Sergei Golubtsov.


On Sat, Aug 21, 2021 at 4:09 AM Jupiter <jupiter.hce@...> wrote:

Hi,

I build ofono v1.31 from git://git.yoctoproject.org/poky branch
hardknott,

$ cat meta/recipes-connectivity/ofono/ofono_1.31.bb

SUMMARY = "open source telephony"
DESCRIPTION = "oFono is a stack for mobile telephony devices on Linux.
oFono supports speaking to telephony devices through specific drivers,
or with generic AT commands."
HOMEPAGE = "http://www.ofono.org"
BUGTRACKER = "https://01.org/jira/browse/OF"
LICENSE = "GPLv2"
LIC_FILES_CHKSUM = "file://COPYING;md5=eb723b61539feef013de476e68b5c50a \


file://src/ofono.h;beginline=1;endline=20;md5=3ce17d5978ef3445def265b98899c2ee"
DEPENDS = "dbus glib-2.0 udev mobile-broadband-provider-info ell"

SRC_URI = "\
${KERNELORG_MIRROR}/linux/network/${BPN}/${BP}.tar.xz \
file://ofono \
file://0001-mbim-add-an-optional-TEMP_FAILURE_RETRY-macro-copy.patch
\
"
SRC_URI[md5sum] = "1c26340e3c6ed132cc812595081bb3dc"
SRC_URI[sha256sum] =
"a15c5d28096c10eb30e47a68b6dc2e7c4a5a99d7f4cfedf0b69624f33d859e9b"

inherit autotools pkgconfig update-rc.d systemd
gobject-introspection-data

INITSCRIPT_NAME = "ofono"
INITSCRIPT_PARAMS = "defaults 22"
SYSTEMD_SERVICE_${PN} = "ofono.service"

PACKAGECONFIG ??= "\
${@bb.utils.filter('DISTRO_FEATURES', 'systemd', d)} \
${@bb.utils.contains('DISTRO_FEATURES', 'bluetooth', 'bluez', '', d)}
\
"
PACKAGECONFIG[systemd] =
"--with-systemdunitdir=${systemd_unitdir}/system/,--with-systemdunitdir="
PACKAGECONFIG[bluez] = "--enable-bluetooth, --disable-bluetooth, bluez5"

EXTRA_OECONF += "--enable-test --enable-external-ell"

do_install_append() {
install -d ${D}${sysconfdir}/init.d/
install -m 0755 ${WORKDIR}/ofono ${D}${sysconfdir}/init.d/ofono
}

PACKAGES =+ "${PN}-tests"

FILES_${PN} += "${systemd_unitdir}"
FILES_${PN}-tests = "${libdir}/${BPN}/test"

RDEPENDS_${PN} += "dbus"
RDEPENDS_${PN}-tests = "\
python3-core \
python3-dbus \
${@bb.utils.contains('GI_DATA_ENABLED', 'True',
'python3-pygobject', '', d)} \
"

RRECOMMENDS_${PN} += "kernel-module-tun mobile-broadband-provider-info"

The build was fine, but running the v1.31 ofono.service crashed at
udevng.c missing OFONO_DRIVER property, the ofono.service was repeated
to be restarted every 6 seconds:

.....

ofonod[9005]: ../ofono-1.31/plugins/udevng.c:add_serial_device()
Device is missing required OFONO_DRIVER property
ofonod[9005]: ../ofono-1.31/plugins/udevng.c:add_serial_device()
Device is missing required OFONO_DRIVER property
......

I am not quite sure if it is ofono 1.31 problem or oe-core
ofono_1.31.bb problem, could anyone have a look at and advise? Let me
know if you need more debug information.

Thank you.

Kind regards,

- jupiter


On 8/20/21, Jupiter <jupiter.hce@...> wrote:

Hi,

I upgraded from Zeus to Hardknotte, I was able to run dbus to get
interface properties, but I failed on Hardknotte ofono 1.31. Has
anyone got oFono 1.31 dbus work?

Thank you.

Kind regards,

- jupiter

_______________________________________________
ofono mailing list -- ofono@...
To unsubscribe send an email to ofono-leave@...

--
"A man can fail many times, but he isn't a failure until he begins to
blame somebody else."
-- John Burroughs

The address included in the meta-hardening documentation
does not work and was changed in other places in 2019.

Signed-off-by: Marta Rybczynska <rybczynska@...>
---
meta-hardening/README | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/meta-hardening/README b/meta-hardening/README
index 37a0b7e..191253c 100644
--- a/meta-hardening/README
+++ b/meta-hardening/README
@@ -64,14 +64,14 @@ layers: meta-oe
Maintenance
-----------

-Send pull requests, patches, comments or questions to yocto@...
+Send pull requests, patches, comments or questions to yocto@...

When sending single patches, please using something like:
-'git send-email -1 --to yocto@... --subject-prefix=meta-hardening][PATCH'
+'git send-email -1 --to yocto@... --subject-prefix=meta-hardening][PATCH'

These values can be set as defaults for this repository:

-$ git config sendemail.to yocto@...
+$ git config sendemail.to yocto@...
$ git config format.subjectPrefix meta-hardening][PATCH

Now you can just do 'git send-email origin/master' to send all local patches.
--
2.30.2

Compilers and related utils are better restricted on production platforms.
Change permissions of all installed binutils tools to remove access from
users outside of the root group.

This also demonstrates how to restrict file permissions in a hardened
distribution.

Signed-off-by: Marta Rybczynska <marta.rybczynska@...>
---
meta-hardening/recipes-devtools/binutils/binutils_%.bbappend | 3 +++
1 file changed, 3 insertions(+)
create mode 100644 meta-hardening/recipes-devtools/binutils/binutils_%.bbappend

diff --git a/meta-hardening/recipes-devtools/binutils/binutils_%.bbappend b/meta-hardening/recipes-devtools/binutils/binutils_%.bbappend
new file mode 100644
index 0000000..3eb3ad0
--- /dev/null
+++ b/meta-hardening/recipes-devtools/binutils/binutils_%.bbappend
@@ -0,0 +1,3 @@
+do_install_append_class-target () {
+ chmod o-rx ${D}${prefix}/${TARGET_SYS}/bin/*
+}
--
2.30.2

Hello, 
SO I'm working on a new recipe, while bitbaking it , this happened, and I really don't know what it's about, 

There is a new PACKAGECONFIG for grpc to allows us
to selectively enable/disable shared builds.

Signed-off-by: Sinan Kaya <okaya@...>
---
.../openembedded-layers/recipes-devtools/grpc/grpc_%.bbappend | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/dynamic-layers/openembedded-layers/recipes-devtools/grpc/grpc_%.bbappend b/dynamic-layers/openembedded-layers/recipes-devtools/grpc/grpc_%.bbappend
index dc0ea42..4cbd1a4 100644
--- a/dynamic-layers/openembedded-layers/recipes-devtools/grpc/grpc_%.bbappend
+++ b/dynamic-layers/openembedded-layers/recipes-devtools/grpc/grpc_%.bbappend
@@ -1,5 +1,4 @@
# doesn't build and not required
DEPENDS:remove:mingw32 = "libnsl2"

-EXTRA_OECMAKE:remove:mingw32 = "-DBUILD_SHARED_LIBS=ON"
-EXTRA_OECMAKE:append:mingw32 = " -DBUILD_SHARED_LIBS=OFF"
+PACKAGECONFIG:remove:mingw32 = "shared"
--
2.17.1

Using dunfell 3.1.8

Will post this inquiry while doing my own investigation. When outputting to the debug console that involves standard input(ex. busctl, systemctl status), the output seems misaligned and dirty.
But I noticed this can be easily fixed by adding to the "tools-sdk" to the EXTRA_IMAGE_FEATURES. 
I'm curious to know what else tools-sdk seems to besides adding development tools(ex. gcc,make)

Below is the snapshot when tools-sdk isn't included in the image built

Hi,

I updated the bitbake to run git pull in master branch, now it is
broken, what does the following error message mean, how to fix it?

$ bitbake-layers show-layers

NOTE: Starting bitbake server...
ERROR: Variable PROVIDES_prepend contains an operation using the old
override syntax. Please convert this layer/metadata before attempting
to use with a newer bitbake.

Thank you.

Kind regards,

- jupiter

Hi all, hi Khem,

Am Do, 19. Aug 2021, um 00:22, schrieb Khem Raj:

On Wed, Aug 18, 2021 at 3:06 PM Manuel Wagesreither <ManWag@...> wrote:

Hello all,

I'm building an image to run on various SBCs and would like to equip it with a graphical interface.

There are quite a few things very unclear to me. Can someone help me with that?

* Why is X11 enabled by setting an IMAGE_FEATURE (namely x11, x11-base or x11-sato), while Wayland is enabled by IMAGE_INSTALL only (weston-init and weston)?

x11-* features is primarily to control what kind of x11 packages you
want to include in image e.g.
./meta/recipes-sato/packagegroups/packagegroup-core-x11-sato.bb is
pulled in when x11-sato is added to IMAGE_FEATURES
we have many X11 based images and sato is one of them so thats why its
separated out.

Okay, so if I get things right then IMAGE_FEATURES+="x11" is under the hood nothing more than an IMAGE_INSTALL+="packagegroup-core-x11". Is that right? If so, what's the purpose of adding the concept of IMAGE_FEATURE? I mean, it doesn't make things SO much easier. Setting an IMAGE_FEATURE or an IMAGE_INSTALL variable is the same to me.

you should really is looking at DISTRO_FEATURES e.g. wayland distro
feature is needed for core-image-weston to build.

Yepp, I know. I left them out on purpose because I was mainly interested in where the configuration for X11 and wayland differs conceptually. With "conceptually" I mean that one is added through IMAGE_FEATURES while the other is through IMAGE_INSTALL.

* Theory: Is IMAGE_FEATURE +=x11 manipulating IMAGE_INSTALL under the hood so you don't have to do it manually? And as there is no IMAGE_FEATURE "wayland", you have do it manually. Correct?
* Why is Wayland different in that it doesn't need an IMAGE_FEATURE to enable it?

there are not many wayland based compositors or images we have in core
as of now.

And if there would be more wayland based compositors or images then you would turn extract this into an IMAGE_FEATURE as well? Why? How does that make things easier? Again, I feel there's something to IMAGE_FEATURES I didn't yet understand.

* Why does core-image-weston.bb need to enable IMAGE_FEATURE hwcodec, while core-image-x11.bb does not? (Dunfell branch.)

openGL is needed for wayland/weston to work too but hwcodec feature is
infact to pull in machine specific drivers MACHINE_HWCODECS into image
if a given BSP defined it.
e.g. intel bsps define vaapi codecs and mediasdk for specific machines
via MACHINE_HWCODECS
defaults for this image features are empty

Thanks for the explanation on MACHINE_HWCODEC. I'm curious, so is core-image-x11 require DISTRO_FEATURE hwcodec or not? If yes, than it seems to be missing in the core-image-x11.bb (it's in the core-images-weston.bb, after all), if no, then why is it not required for X11?


I know I'm asking quite detailed questions, but I got the feeling I need to understand this once and for all.

Thanks, regards, Manuel