Bruce,
It's an honor to get help from the author himself!

I'm starting to understand why they have this:

do_configure_prepend() {
     cp "${S}/arch/${ARCH}/configs/${KBUILD_DEFCONFIG}" "${B}/.config" || bbfatal "CONFIG ${KBUILD_DEFCONFIG} NOT FOUND"
}

Since do_kernel_configme() creates the .config file in https://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/classes/kernel-yocto.bbclass?h=dunfell#n399
(I can't figure out why it fails to merge in our defconfig)

Then the ${WORKDIR}/defconfig is not used by kernel_do_configure() https://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/classes/kernel.bbclass#n595
So they do the copy of defconfig in do_configure_prepend(). 

Anyway,  do_kernel_configme() has already been run hence no CONFIG_LOCALVERSION is written to .config

I could just do this:
do_configure_prepend() {
     cp "${S}/arch/${ARCH}/configs/${KBUILD_DEFCONFIG}" "${B}/.config" || bbfatal "CONFIG ${KBUILD_DEFCONFIG} NOT FOUND"
     echo "CONFIG_LOCALVERSION=\"${LINUX_VERSION_EXTENSION}\"" >> "${B}/.config"
}

 

 

 



On Fri, 2021-10-08 at 16:10 +0000, Monsees, Steven C (US) via
lists.yoctoproject.org wrote:

Sorry for my ignorance, but I am missing your point, what exactly
does the recipe provide when built, intel-opencl-icd ?, will it pull
in igc ?

It will pull in igc. There's a separate recipe for igc in meta-intel.
compute-runtime will provide intel-opencl-icd but the package name is
different.

You can try building the recipe.

Thanks,

Anuj

-----Original Message-----
From: Mittal, Anuj <anuj.mittal@...>
Sent: Friday, October 8, 2021 12:01 PM
To: Monsees, Steven C (US) <steven.monsees@...>;
yocto@...
Subject: Re: [yocto] #zeus meta-intel

External Email Alert

This email has been sent from an account outside of the BAE Systems
network.

Please treat the email with caution, especially if you are requested
to click on a link, decrypt/open an attachment, or enable macros. 
For further information on how to spot phishing, access
“Cybersecurity OneSpace Page” and report phishing by clicking the
button “Report Phishing” on the Outlook toolbar.


Sorry, I didn't understand the question. If you want to know which
version of OpenCL is supported, please see the README for version
that you're building. Also see:

https://github.com/intel/compute-runtime/blob/master/opencl/doc/FAQ.md

Thanks,

Anuj

On Fri, 2021-10-08 at 15:30 +0000, Monsees, Steven C (US) wrote:

Thank you, just for clarification ...

"Building the Neo driver", does that mean no OpenCL library support
?

-----Original Message-----
From: Mittal, Anuj <anuj.mittal@...>
Sent: Friday, October 8, 2021 11:07 AM
To: Monsees, Steven C (US) <steven.monsees@...>;
yocto@...
Subject: Re: [yocto] #zeus meta-intel

External Email Alert

This email has been sent from an account outside of the BAE Systems
network.

Please treat the email with caution, especially if you are
requested
to click on a link, decrypt/open an attachment, or enable macros. 
For
further information on how to spot phishing, access “Cybersecurity
OneSpace Page” and report phishing by clicking the button “Report
Phishing” on the Outlook toolbar.


On Thu, 2021-10-07 at 16:20 +0000, Monsees, Steven C (US) via
lists.yoctoproject.org wrote:

 
 
I am looking at “meta-intel/dynamic-layers/clang-layer/recipes-
opencl/compute-runtime”
I have been “told” that this compute-runtime recipe should
basically
build opencl (Neo) with all the required dependencies… I did not
think Yocto had support for OpenCL, and was looking to build Neo
under the SDK.
 
(1)   Is this true, does meta-intell under zeus now support
building
OpenCL and the OpenCL Graphics compiler under Yocto ?

compute-runtime recipe would allow you to build NEO driver. zeus is
no
longer maintained so the version there is old and not tested now
but
it should still build.

(2)   Of so, what would be the proper way to pull this into my
yocto
build ?

Including meta-intel and meta-clang in bblayers should be enough.

Thanks,

Anuj

On Fri, Oct 8, 2021 at 11:52 AM <mrkozmic@...> wrote:

I have defined LINUX_VERSION_EXTENSION in my kernel recipe.

I have no CONFIG_LOCALVERSION or CONFIG_LOCALVERSION_AUTO defined in my defconfig

After i run bitbake -c kernel_configme <my kernel recipe> I can see in my .config file:

CONFIG_LOCALVERSION=""
# CONFIG_LOCALVERSION_AUTO is not set

and at the very end:

CONFIG_LOCALVERSION="<my extention>"
So far so good.

Then I execute bitbake -c configure <my kernel recipe>
and the .config file is changed. Now there is:
CONFIG_LOCALVERSION=""
CONFIG_LOCALVERSION_AUTO=y
and there is no trace of CONFIG_LOCALVERSION="<my extention>"

Assuming what you are using as a kernel recipe is linux-yocto based,
the code that is adding the CONFIG_LOCALVERSION is a task that runs
after do_configure. So just running -c configure, it shouldn't be in
the .config (and the default kernel configure task, will copy your
defconfig -> .config, and then lets the kernel process it, so you will
have symbols in the .config that are not in your defconfig
explicitly).

If you can make your recipe available, I can probably offer more
specific comments.

Bruce

Yocto: dunfell
Kernel 5.4.70

--
- Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end
- "Use the force Harry" - Gandalf, Star Trek II

Sorry for my ignorance, but I am missing your point, what exactly does the recipe provide when built, intel-opencl-icd ?, will it pull in igc ?

Sorry, I didn't understand the question. If you want to know which
version of OpenCL is supported, please see the README for version that
you're building. Also see:

https://github.com/intel/compute-runtime/blob/master/opencl/doc/FAQ.md

Thanks,

Anuj

I have defined LINUX_VERSION_EXTENSION in my kernel recipe.

I have no CONFIG_LOCALVERSION or CONFIG_LOCALVERSION_AUTO defined in my defconfig

After i run bitbake -c kernel_configme <my kernel recipe> I can see in my .config file:

CONFIG_LOCALVERSION=""
# CONFIG_LOCALVERSION_AUTO is not set

CONFIG_LOCALVERSION="<my extention>"
So far so good.

Then I execute  bitbake -c configure <my kernel recipe>
and the .config file is changed. Now there is:
CONFIG_LOCALVERSION=""
CONFIG_LOCALVERSION_AUTO=y
and there is no trace of CONFIG_LOCALVERSION="<my extention>"

Yocto: dunfell
Kernel 5.4.70

 

Thank you, just for clarification ...

"Building the Neo driver", does that mean no OpenCL library support ?

On Thu, 2021-10-07 at 16:20 +0000, Monsees, Steven C (US) via
lists.yoctoproject.org wrote:

 
 
I am looking at “meta-intel/dynamic-layers/clang-layer/recipes-
opencl/compute-runtime”
I have been “told” that this compute-runtime recipe should basically
build opencl (Neo) with all the required dependencies…
I did not think Yocto had support for OpenCL, and was looking to
build Neo under the SDK.
 
(1)   Is this true, does meta-intell under zeus now support building
OpenCL and the OpenCL Graphics compiler under Yocto ?

compute-runtime recipe would allow you to build NEO driver. zeus is no
longer maintained so the version there is old and not tested now but it
should still build.

(2)   Of so, what would be the proper way to pull this into my yocto
build ?

Including meta-intel and meta-clang in bblayers should be enough.

Thanks,

Anuj

Hello Tien,



On 10/7/21 9:02 PM, thloh85@... wrote:

Hi all,

I'm from StarFive and am working on distro for our risc-v based processor.
We currently have a repository
(https://github.com/starfive-tech/meta-starfive) and wish to get the
repository to be part of git.yoctoproject.org.

Since you already have a layer, have you thought about registering it
with the layer-index ? http://layers.openembedded.org

- Armin

I couldn't find any information on the process to get the repository
into git.yoctoproject.org. Can someone help show me the process?

Thanks
Tien Hock

Attached the diff.

Thanks

Hi Alex,

Actually we'd like to have it up and get everything (including meta-riscv by Khem) into git.yoctoproject.org so that everything in consolidated into a central location.
Does that make sense? Or it is expected that vendors do host their own repository? I see meta-intel, meta-xillinx, so I suspect platform vendor should also get the layers into git.yoctoproject.org

I just read that we'll need to join membership to get repos into the git. I will discuss this with the management.

Thanks for the quick response Alex. 

Hello all,

How can I enable graphics acceleration on qemux86-64? My image is using wayland with weston and the GUI is not exactly fast.

* When I do `runqemu slirp kvm gtk`, it throws the following error: `runqemu - ERROR - Failed to run qemu: qemu-system-x86_64: Display 'gtk' is not available.`
* `runqemu slirp kvm sdl` works, but the weston GUI does not seem to be accelerated.
* `runqemu slirp kvm sdl gl` emits the following: `runqemu - ERROR - Failed to run qemu: qemu-system-x86_64: OpenGL support is disabled`

Below I'm posting a snippet of `poky/scripts/runqemu` to have at hand the qemu options enabled by the runqemu options mentioned above:
```
elif arg == 'sdl':
if 'gl' in sys.argv[1:]:
self.qemu_opt_script += ' -vga virtio -display sdl,gl=on'
elif 'gl-es' in sys.argv[1:]:
self.qemu_opt_script += ' -vga virtio -display sdl,gl=es'
else:
self.qemu_opt_script += ' -display sdl'
elif arg == 'gtk':
if 'gl' in sys.argv[1:]:
self.qemu_opt_script += ' -vga virtio -display gtk,gl=on'
elif 'gl-es' in sys.argv[1:]:
self.qemu_opt_script += ' -vga virtio -display gtk,gl=es'
else:
self.qemu_opt_script += ' -display gtk'
```

Kind regards,
Manuel

Add chipsec, tools to dump and analyzing hardware, system firmware
components, like PCH register, ioport or iomem configuration space.

Signed-off-by: Liwei Song <liwei.song@...>
---
recipes-security/chipsec/chipsec_git.bb | 35 +++++++++++++++++++++++++
1 file changed, 35 insertions(+)
create mode 100644 recipes-security/chipsec/chipsec_git.bb

diff --git a/recipes-security/chipsec/chipsec_git.bb b/recipes-security/chipsec/chipsec_git.bb
new file mode 100644
index 000000000000..3339dc142107
--- /dev/null
+++ b/recipes-security/chipsec/chipsec_git.bb
@@ -0,0 +1,35 @@
+SUMMARY = "CHIPSEC: Platform Security Assessment Framework"
+
+DESCRIPTION = "CHIPSEC is a framework for analyzing the security \
+ of PC platforms including hardware, system firmware \
+ (BIOS/UEFI), and platform components."
+
+LICENSE = "GPLv2"
+LIC_FILES_CHKSUM = "file://COPYING;md5=bc2d1f9b427be5fb63f6af9da56f7c5d"
+
+SRC_URI = "git://github.com/chipsec/chipsec.git;branch=master \
+ "
+
+SRCREV = "b2a61684826dc8b9f622a844a40efea579cd7e7d"
+
+COMPATIBLE_HOST = "(i.86|x86_64).*-linux"
+
+S = "${WORKDIR}/git"
+EXTRA_OEMAKE = "CC='${CC}' LDFLAGS='${LDFLAGS}' CFLAGS='${CFLAGS}'"
+
+DEPENDS = "virtual/kernel nasm-native python3-setuptools-native"
+RDEPENDS:${PN} += "python3 python3-modules"
+
+inherit module distutils3
+
+do_compile:append() {
+ cd ${S}/drivers/linux
+ oe_runmake KSRC=${STAGING_KERNEL_BUILDDIR}
+}
+
+do_install:append() {
+ install -m 0644 ${S}/drivers/linux/chipsec.ko ${D}${PYTHON_SITEPACKAGES_DIR}/chipsec/helper/linux
+}
+
+FILES:${PN} += "${exec_prefix} \
+"
--
2.17.1

Hi all,

I'm from StarFive and am working on distro for our risc-v based processor.
We currently have a repository (https://github.com/starfive-tech/meta-starfive) and wish to get the repository to be part of git.yoctoproject.org. 
I couldn't find any information on the process to get the repository into git.yoctoproject.org. Can someone help show me the process?

Thanks
Tien Hock