1. yocto@lists.yoctoproject.org
  2. Messages
   Date   
Re: Yocto SDK: how to deploy application with step "make install" to target?

Philip Balister
 

On 12/01/2017 10:15 AM, Jerry Lian wrote:
Thanks Eric for your explanation!
* (I know I can build the app within yocto build, but sometimes I prefer
the SDK method outside yocto build.)
* So just copy manually? I thought there are some tricks that I don't know.

Now I have follow-up question:
* It seems that "make install" will copy different files to different
folders
* Can I do like these:?
--- make
--- copy (via scp)
--- make install
* Or it needs to be:
--- make
--- make install
--- copy (via scp) -----then how do I know: what files to go which folder?
--- can you explain:
"... you may supply a custom install prefix directory to configure."
--- how?
I've used sshfs to mount the target on the machine with the sdk and then
something like:

make install DESTDIR=~/mydevice

From:

https://wiki.gnuradio.org/index.php/Cross_compile_an_OOT_and_install_on_target

Philip


Thanks!





On Fri, Dec 1, 2017 at 9:34 AM, Eric Schwarz <eas@...>
wrote:

Am 01.12.2017 15:16, schrieb Eric Schwarz:

Hi Jerry,

Am 01.12.2017 15:07, schrieb Jerry Lian:

I am new to embedded linux, and some concepts confuse me:
* I have downloaded an Autotools-based application to run on my board.
* If I boot the board with original image/original SDK, I normally
build/install it with steps:
----- ./bootstrap
----- ./configure
----- make
----- make install
You may build your app within yocto build process so your application is
contained in your rootFS directly.

* Now if I boot the board with yocto image without SDK, how can I deploy
the application?
---- (surely I do "bitbake -c populate_sdk some-image", and source the
environment!)

----- ./bootstrap (Yocto SDK on host)
----- ./configure (Yocto SDK on host)
----- make (Yocto SDK on host)
----- make install (???)
If you build it w/ the SDK outside of the yocto build process just
install it to a local directory and copy all the stuff w/ e.g. scp onto
your boards rootFS.
... you may supply a custom install prefix directory to configure.



* How to do "make install"? (I want to install it on target, but how?)

Thanks!
Cheers
Eric
include and a library dependency in a recipe which uses a Makefile

Rail Shafigulin <rail@...>
 

Somewhat new to Yocto.

I have a relatively simple Makefile which builds a library. The file
is shown below:

LIB=PyAnokiWave

It is shown below
# Add any other object files to this list below
LIB_OBJS = libPyAnokiWave.o
LIB_HEADERS=libPyAnokiWave.h

CFLAGS+=-I/usr/include/python2.7
LDFLAGS+=-lpython2.7 -lAnokiWave
all: build

.PHONY: build
build: $(LIB).a $(LIB).so

$(LIB).a:$(LIB_OBJS)
$(AR) rcs $@ $(LIB_OBJS)

$(LIB).so:$(LIB_OBJS)
$(CC) $(LDFLAGS) -shared -g -o $@ $(LIB_OBJS)

%.o: %.c
$(CC) -c $(CFLAGS) -o $@ $<

I need to use this Makefile in a recipe which but I'm struggling to
figure out how to make it point to correct include folder as well as a
library folder.

Here is what I need to change:

1. CFLAGS+=-I/usr/include/python2.7

Clearly it is not pointing to the right location. How do I make it
point to the include directory of my build (at least I think it should
point to this location)

2. LDFLAGS+=-lpython2.7 -lAnokiWave

I do have a meta-python layer and it builds python2.7 library. I also
have a custom layer which builds a libAnokiWave.so. How do I make
compiler/linker point to the right folders so that these libraries
would be used? I do understand that it has something to do with the
EXTRA_OEMAKE variable, but I wasn't able to find a good example that
would show me how to use it.

At the moment my do_compile task looks like the following:

do_compile () {
oe_runmake
}

I would greatly appreciate any help with this.
--
Rail Shafigulin
Software Engineer
Esencia Technologies

--




*ESENCIA TECHNOLOGIES, INC.*3945 Freedom Circle, Suite #360,
Santa Clara CA 95054
________________________________________________________

Phone: +1 408 736 8284 Fax: +1 408 519 3475
http://www.esenciatech.com | http://www.lnttechservices.com
Re: dev package non-sym link

Rail Shafigulin <rail@...>
 

On Fri, Dec 1, 2017 at 12:12 AM, Alexander Kanavin
<alexander.kanavin@...> wrote:
On 11/30/2017 11:46 PM, Rail Shafigulin wrote:

do_install () {
install -d ${D}${libdir}/
install -m 0755 ${S}/libAnokiWave.so.1.0.1 ${D}${libdir}
ln -s ${D}${libdir}/libAnokiWave.so.1.0.1
${D}${libdir}/libAnokiwave.so.1
ln -s ${D}${libdir}/libAnokiWave.so.1 ${D}${libdir}/libAnokiwave.so
}

You need to call 'ln -sr', then you'll get relative symlinks. Better yet,
add some build system support to the project (meson, cmake), then it'll take
care of these things for you.

Alex
Thanks for the reply. It worked. I will definitely will look into
meson and cmake at a later time.


--
Rail Shafigulin
Software Engineer
Esencia Technologies

--




*ESENCIA TECHNOLOGIES, INC.*3945 Freedom Circle, Suite #360,
Santa Clara CA 95054
________________________________________________________

Phone: +1 408 736 8284 Fax: +1 408 519 3475
http://www.esenciatech.com | http://www.lnttechservices.com
Re: [meta-raspberrypi][PATCH] rpi-config: Setup lcd_rotate for Raspberry Pi display

Andrei Gherzan
 


On Thu, Nov 30, 2017 at 5:43 PM, Paul Barker <pbarker@...> wrote:
On Thu, Nov 30, 2017 at 5:16 PM, Paul Barker <pbarker@...> wrote:
> On Wed, Nov 29, 2017 at 10:20 PM,  <drew.moseley@...> wrote:
>> From: Drew Moseley <drew.moseley@...>
>>
>> Signed-off-by: Drew Moseley <drew.moseley@...>
>> ---
>>  docs/extra-build-config.md              | 6 ++++++
>>  recipes-bsp/bootfiles/rpi-config_git.bb | 6 ++++++
>>  2 files changed, 12 insertions(+)
>>
>> diff --git a/docs/extra-build-config.md b/docs/extra-build-config.md
>> index 380d969..0390eee 100644
>> --- a/docs/extra-build-config.md
>> +++ b/docs/extra-build-config.md
>> @@ -170,6 +170,12 @@ Screen LCD, HDMI interface (<http://www.waveshare.com/7inch-HDMI-LCD-C.htm>) Rev
>>
>>      WAVESHARE_1024X600_C_2_1 = "1"
>>
>> +If you would like to use the Raspberry Pi, 7 inch Capacitive Touch Screen
>> +LCD (https://www.raspberrypi.org/products/raspberry-pi-touch-display/), please
>> +set the following in your local.conf:
>> +
>> +    RASPBERRYPI_DISPLAY = "1"
>> +
>>  ## Enable UART
>>
>>  RaspberryPi 0, 1, 2 and CM will have UART console enabled by default.
>> diff --git a/recipes-bsp/bootfiles/rpi-config_git.bb b/recipes-bsp/bootfiles/rpi-config_git.bb
>> index ca4160b..e9a4fe2 100644
>> --- a/recipes-bsp/bootfiles/rpi-config_git.bb
>> +++ b/recipes-bsp/bootfiles/rpi-config_git.bb
>> @@ -127,6 +127,12 @@ do_deploy() {
>>          echo "hdmi_cvt 1024 600 60 6 0 0 0" >> ${DEPLOYDIR}/bcm2835-bootfiles/config.txt
>>          echo "hdmi_drive=1" >> ${DEPLOYDIR}/bcm2835-bootfiles/config.txt
>>      fi
>> +
>> +    # Raspberry Pi 7" Display/Touch screen (https://www.raspberrypi.org/products/raspberry-pi-touch-display/)
>> +    if [ "${RASPBERRYPI_DISPLAY}" = "1" ]; then
>> +        echo "# Raspberry Pi 7\" display/touch screen" >> ${DEPLOYDIR}/bcm2835-bootfiles/config.txt
>> +        echo "lcd_rotate=2" >> ${DEPLOYDIR}/bcm2835-bootfiles/config.txt
>> +    fi
>>  }
>>
>>  do_deploy_append_raspberrypi3-64() {
>
> This all looks good in general. Just a quick check though - are all
> users of this display going to want that lcd_rotate value? I've never
> used this display so not sure if that's specific to particular use
> cases or not.
>
> If you're happy that this is applicable for everyone could you submit
> this as a pull request on github
> (https://github.com/agherzan/meta-raspberrypi).
>

Someone has come back off-list and said they use that display without
the lcd_rotate setting. I think we'll have to drop this then.

It may be good to have an RPI_EXTRA_CONFIG variable or similar to add
arbitrary text to the config file, so you could do something like
this:

RPI_EXTRA_CONFIG = " \
    # Raspberry Pi 7\" display/touch screen \
    lcd_rotate=2 \
    "

How does that sound?



I like this.



--
Andrei Gherzan


--
Andrei Gherzan

On Thu, Nov 30, 2017 at 5:43 PM, Paul Barker <pbarker@...> wrote:
On Thu, Nov 30, 2017 at 5:16 PM, Paul Barker <pbarker@...> wrote:
> On Wed, Nov 29, 2017 at 10:20 PM,  <drew.moseley@...> wrote:
>> From: Drew Moseley <drew.moseley@...>
>>
>> Signed-off-by: Drew Moseley <drew.moseley@...>
>> ---
>>  docs/extra-build-config.md              | 6 ++++++
>>  recipes-bsp/bootfiles/rpi-config_git.bb | 6 ++++++
>>  2 files changed, 12 insertions(+)
>>
>> diff --git a/docs/extra-build-config.md b/docs/extra-build-config.md
>> index 380d969..0390eee 100644
>> --- a/docs/extra-build-config.md
>> +++ b/docs/extra-build-config.md
>> @@ -170,6 +170,12 @@ Screen LCD, HDMI interface (<http://www.waveshare.com/7inch-HDMI-LCD-C.htm>) Rev
>>
>>      WAVESHARE_1024X600_C_2_1 = "1"
>>
>> +If you would like to use the Raspberry Pi, 7 inch Capacitive Touch Screen
>> +LCD (https://www.raspberrypi.org/products/raspberry-pi-touch-display/), please
>> +set the following in your local.conf:
>> +
>> +    RASPBERRYPI_DISPLAY = "1"
>> +
>>  ## Enable UART
>>
>>  RaspberryPi 0, 1, 2 and CM will have UART console enabled by default.
>> diff --git a/recipes-bsp/bootfiles/rpi-config_git.bb b/recipes-bsp/bootfiles/rpi-config_git.bb
>> index ca4160b..e9a4fe2 100644
>> --- a/recipes-bsp/bootfiles/rpi-config_git.bb
>> +++ b/recipes-bsp/bootfiles/rpi-config_git.bb
>> @@ -127,6 +127,12 @@ do_deploy() {
>>          echo "hdmi_cvt 1024 600 60 6 0 0 0" >> ${DEPLOYDIR}/bcm2835-bootfiles/config.txt
>>          echo "hdmi_drive=1" >> ${DEPLOYDIR}/bcm2835-bootfiles/config.txt
>>      fi
>> +
>> +    # Raspberry Pi 7" Display/Touch screen (https://www.raspberrypi.org/products/raspberry-pi-touch-display/)
>> +    if [ "${RASPBERRYPI_DISPLAY}" = "1" ]; then
>> +        echo "# Raspberry Pi 7\" display/touch screen" >> ${DEPLOYDIR}/bcm2835-bootfiles/config.txt
>> +        echo "lcd_rotate=2" >> ${DEPLOYDIR}/bcm2835-bootfiles/config.txt
>> +    fi
>>  }
>>
>>  do_deploy_append_raspberrypi3-64() {
>
> This all looks good in general. Just a quick check though - are all
> users of this display going to want that lcd_rotate value? I've never
> used this display so not sure if that's specific to particular use
> cases or not.
>
> If you're happy that this is applicable for everyone could you submit
> this as a pull request on github
> (https://github.com/agherzan/meta-raspberrypi).
>

Someone has come back off-list and said they use that display without
the lcd_rotate setting. I think we'll have to drop this then.

It may be good to have an RPI_EXTRA_CONFIG variable or similar to add
arbitrary text to the config file, so you could do something like
this:

RPI_EXTRA_CONFIG = " \
    # Raspberry Pi 7\" display/touch screen \
    lcd_rotate=2 \
    "

How does that sound?

--
Paul Barker
Togán Labs Ltd
--
_______________________________________________
yocto mailing list
yocto@...
https://lists.yoctoproject.org/listinfo/yocto
Re: how to know Poky kernel version before running it

Vincent Prince
 

Something like following command will give you every variables for kernel recipe:

bitbake -e virtual/kernel | egrep '^[A-Z][A-Z_]*' 

You will find prefered version and source url variables.

2017-12-01 18:11 GMT+01:00 Steve Pavao <stevep@...>:

Is there a definitive way to know at/after build time what kernel version is being built via a particular bitbake of Poky?

I don’t have the board on hand to boot the result of my work right now; it is a scout build for a board I’m getting soon.

I’m trying to figure it out by browsing the console-latest.log and looking for the highest kernel version I can observe from the recipes used.  Is there a better way to find this out at build time?

Knowing the kernel version would help me decide what version of a patch to apply to go along with it.

Steve Pavao
Korg R&D

--
_______________________________________________
yocto mailing list
yocto@...
https://lists.yoctoproject.org/listinfo/yocto
how to know Poky kernel version before running it

Steve Pavao
 

Is there a definitive way to know at/after build time what kernel version is being built via a particular bitbake of Poky?

I don’t have the board on hand to boot the result of my work right now; it is a scout build for a board I’m getting soon.

I’m trying to figure it out by browsing the console-latest.log and looking for the highest kernel version I can observe from the recipes used. Is there a better way to find this out at build time?

Knowing the kernel version would help me decide what version of a patch to apply to go along with it.

Steve Pavao
Korg R&D
Re: Yocto SDK: how to deploy application with step "make install" to target?

Jerry Lian <jerry.lian@...>
 

OK, that is what confused me: I DID try "make modules_install" on host!

So according to what you said, the right procedure is:
* do          "make install"                on host, then copy to target:
* don't do "make modules_install" on host, but make a package first, then copy to target and install.
* right?

* then how to make a package?
* how do I know which are "relevant other files"?

sorry for my newbie questions, thanks!





On Fri, Dec 1, 2017 at 11:30 AM, Eric Schwarz <eas@...> wrote:
Am 01.12.2017 16:58, schrieb Jerry Lian:

Thanks Eric again.
* So you mean "copy -Rfa" can match folder-by-folder, so long as the folder-tree are the same?
(if yes, that is super great!!!)

 Yes.

* I check that "make install" copy files to one folder, so that is easy for deployment!

Another question:
* My application actually has another step: "make modules_install" (surely step "make modules" also)
* But I got error (see below):
what could be wrong with "SSL error"? isn't build already passed successfully?

 You are doing the 'make modules_install' on the target or where else? You need to build the modules within the Yocto environment and may create a package e.g. *.deb you may install later on the target or copy the modules by hand including relevant other files needed. Obviously the system tries to recompile the kernel modules. That's what 'make modules_install' does.



----------------------------------------------------------------------------------------------------
jerry@yct:~/ecat-33b922$ make modules_install
make -C "/usr/src/linux-headers-4.10.0-28-generic" M="/home/jerry/ecat-33b922" \
INSTALL_MOD_DIR="ethercat" modules_install
make[1]: Entering directory '/usr/src/linux-headers-4.10.0-28-generic'
mkdir: cannot create directory '/lib/modules/4.10.0-28-generic/ethercat': Permission denied
Makefile:1536: recipe for target '_emodinst_' failed
make[1]: *** [_emodinst_] Error 1
make[1]: Leaving directory '/usr/src/linux-headers-4.10.0-28-generic'
Makefile:934: recipe for target 'modules_install' failed
make: *** [modules_install] Error 2

jerry@yct:~/ecat-33b922$ sudo make modules_install
make -C "/usr/src/linux-headers-4.10.0-28-generic" M="/home/jerry/ecat-33b922" \
INSTALL_MOD_DIR="ethercat" modules_install
make[1]: Entering directory '/usr/src/linux-headers-4.10.0-28-generic'
INSTALL /home/jerry/ecat-33b922/devices/ec_generic.ko
At main.c:158:
- SSL error:02001002:system library:fopen:No such file or directory: bss_file.c:175
- SSL error:2006D080:BIO routines:BIO_new_file:no such file: bss_file.c:178
sign-file: certs/signing_key.pem: No such file or directory
INSTALL /home/jerry/ecat-33b922/examples/mini/ec_mini.ko
At main.c:158:
- SSL error:02001002:system library:fopen:No such file or directory: bss_file.c:175
- SSL error:2006D080:BIO routines:BIO_new_file:no such file: bss_file.c:178
sign-file: certs/signing_key.pem: No such file or directory
INSTALL /home/jerry/ecat-33b922/master/ec_master.ko
At main.c:158:
- SSL error:02001002:system library:fopen:No such file or directory: bss_file.c:175
- SSL error:2006D080:BIO routines:BIO_new_file:no such file: bss_file.c:178
sign-file: certs/signing_key.pem: No such file or directory
DEPMOD  4.10.0-28-generic
make[1]: Leaving directory '/usr/src/linux-headers-4.10.0-28-generic'

jerry@yct:~/ecat-33b922$
-------------------------------------------------------------------------------------------------------

Thanks!

On Fri, Dec 1, 2017 at 10:37 AM, Eric Schwarz <eas@...> wrote:
Am 01.12.2017 16:15, schrieb Jerry Lian:

Thanks Eric for your explanation!
* (I know I can build the app within yocto build, but sometimes I prefer the SDK method outside yocto build.)

* So just copy manually? I thought there are some tricks  that I don't know.

Acutally that's how we are doing it ...

Now I have follow-up question:
* It seems that "make install" will copy different files to different folders
* Can I do like these:?
--- make
--- copy (via scp)
--- make install
* Or it needs to be:
--- make
--- make install Install the cross-compiled stuff on your local disc e.g. into a "transfer" directory. Then copy it to the target.
On the target 'cp -Rfa <transfer-dir>/* /' should do it if the basic directory structure fits to your rootFS.

--- copy (via scp) -----then how do I know: what files to go which folder?
--- can you explain:
"... you may supply a custom install prefix directory to configure." --- how?

There is IMHO no standardized variable available. Just look into the configure script how the variable is named and if it is actually available at all.

Thanks!

On Fri, Dec 1, 2017 at 9:34 AM, Eric Schwarz <eas@...> wrote:
Am 01.12.2017 15:16, schrieb Eric Schwarz:

Hi Jerry,

Am 01.12.2017 15:07, schrieb Jerry Lian:

I am new to embedded linux, and some concepts confuse me:
* I have downloaded an Autotools-based application to run on my board.
* If I boot the board with original image/original SDK, I normally build/install it with steps:
----- ./bootstrap
----- ./configure
----- make
----- make install
You may build your app within yocto build process so your application is contained in your rootFS directly.

* Now if I boot the board with yocto image without SDK, how can I deploy the application?
---- (surely I do "bitbake -c populate_sdk some-image", and source the environment!)

----- ./bootstrap (Yocto SDK on host)
----- ./configure (Yocto SDK on host)
----- make         (Yocto SDK on host)
----- make install (???)
If you build it w/ the SDK outside of the yocto build process just install it to a local directory and copy all the stuff w/ e.g. scp onto your boards rootFS.  ... you may supply a custom install prefix directory to configure.

* How to do "make install"? (I want to install it on target, but how?)

Thanks!
Cheers
Eric
Re: Yocto SDK: how to deploy application with step "make install" to target?

Eric Schwarz
 

Am 01.12.2017 16:58, schrieb Jerry Lian:

Thanks Eric again.
* So you mean "copy -Rfa" can match folder-by-folder, so long as the folder-tree are the same?
(if yes, that is super great!!!)
Yes.

* I check that "make install" copy files to one folder, so that is easy for deployment!
Another question:
* My application actually has another step: "make modules_install" (surely step "make modules" also)
* But I got error (see below):
what could be wrong with "SSL error"? isn't build already passed successfully?
You are doing the 'make modules_install' on the target or where else? You need to build the modules within the Yocto environment and may create a package e.g. *.deb you may install later on the target or copy the modules by hand including relevant other files needed. Obviously the system tries to recompile the kernel modules. That's what 'make modules_install' does.

----------------------------------------------------------------------------------------------------
jerry@yct:~/ecat-33b922$ make modules_install
make -C "/usr/src/linux-headers-4.10.0-28-generic" M="/home/jerry/ecat-33b922" \
INSTALL_MOD_DIR="ethercat" modules_install
make[1]: Entering directory '/usr/src/linux-headers-4.10.0-28-generic'
mkdir: cannot create directory '/lib/modules/4.10.0-28-generic/ethercat': Permission denied
Makefile:1536: recipe for target '_emodinst_' failed
make[1]: *** [_emodinst_] Error 1
make[1]: Leaving directory '/usr/src/linux-headers-4.10.0-28-generic'
Makefile:934: recipe for target 'modules_install' failed
make: *** [modules_install] Error 2
jerry@yct:~/ecat-33b922$ sudo make modules_install
make -C "/usr/src/linux-headers-4.10.0-28-generic" M="/home/jerry/ecat-33b922" \
INSTALL_MOD_DIR="ethercat" modules_install
make[1]: Entering directory '/usr/src/linux-headers-4.10.0-28-generic'
INSTALL /home/jerry/ecat-33b922/devices/ec_generic.ko
At main.c:158:
- SSL error:02001002:system library:fopen:No such file or directory: bss_file.c:175
- SSL error:2006D080:BIO routines:BIO_new_file:no such file: bss_file.c:178
sign-file: certs/signing_key.pem: No such file or directory
INSTALL /home/jerry/ecat-33b922/examples/mini/ec_mini.ko
At main.c:158:
- SSL error:02001002:system library:fopen:No such file or directory: bss_file.c:175
- SSL error:2006D080:BIO routines:BIO_new_file:no such file: bss_file.c:178
sign-file: certs/signing_key.pem: No such file or directory
INSTALL /home/jerry/ecat-33b922/master/ec_master.ko
At main.c:158:
- SSL error:02001002:system library:fopen:No such file or directory: bss_file.c:175
- SSL error:2006D080:BIO routines:BIO_new_file:no such file: bss_file.c:178
sign-file: certs/signing_key.pem: No such file or directory
DEPMOD 4.10.0-28-generic
make[1]: Leaving directory '/usr/src/linux-headers-4.10.0-28-generic'
jerry@yct:~/ecat-33b922$
-------------------------------------------------------------------------------------------------------
Thanks!
On Fri, Dec 1, 2017 at 10:37 AM, Eric Schwarz <eas@...> wrote:
Am 01.12.2017 16:15, schrieb Jerry Lian:
Thanks Eric for your explanation!
* (I know I can build the app within yocto build, but sometimes I prefer the SDK method outside yocto build.)
* So just copy manually? I thought there are some tricks that I don't know.
Acutally that's how we are doing it ...
Now I have follow-up question:
* It seems that "make install" will copy different files to different folders
* Can I do like these:?
--- make
--- copy (via scp)
--- make install
* Or it needs to be:
--- make
--- make install Install the cross-compiled stuff on your local disc e.g. into a "transfer" directory. Then copy it to the target.
On the target 'cp -Rfa <transfer-dir>/* /' should do it if the basic directory structure fits to your rootFS.
--- copy (via scp) -----then how do I know: what files to go which folder?
--- can you explain:
"... you may supply a custom install prefix directory to configure." --- how?
There is IMHO no standardized variable available. Just look into the configure script how the variable is named and if it is actually available at all.
Thanks!
On Fri, Dec 1, 2017 at 9:34 AM, Eric Schwarz <eas@...> wrote:
Am 01.12.2017 15:16, schrieb Eric Schwarz:
Hi Jerry,
Am 01.12.2017 15:07, schrieb Jerry Lian:
I am new to embedded linux, and some concepts confuse me:
* I have downloaded an Autotools-based application to run on my board.
* If I boot the board with original image/original SDK, I normally build/install it with steps:
----- ./bootstrap
----- ./configure
----- make
----- make install
You may build your app within yocto build process so your application is contained in your rootFS directly.
* Now if I boot the board with yocto image without SDK, how can I deploy the application?
---- (surely I do "bitbake -c populate_sdk some-image", and source the environment!)
----- ./bootstrap (Yocto SDK on host)
----- ./configure (Yocto SDK on host)
----- make (Yocto SDK on host)
----- make install (???)
If you build it w/ the SDK outside of the yocto build process just install it to a local directory and copy all the stuff w/ e.g. scp onto your boards rootFS. ... you may supply a custom install prefix directory to configure.
* How to do "make install"? (I want to install it on target, but how?)
Thanks!
Cheers
Eric
Re: Yocto SDK: how to deploy application with step "make install" to target?

Jerry Lian <jerry.lian@...>
 

Thanks Eric again.
* So you mean "copy -Rfa" can match folder-by-folder, so long as the folder-tree are the same?
                (if yes, that is super great!!!)
* I check that "make install" copy files to one folder, so that is easy for deployment!

Another question:
* My application actually has another step: "make modules_install" (surely step "make modules" also)
* But I got error (see below):
               what could be wrong with "SSL error"? isn't build already passed successfully?

----------------------------------------------------------------------------------------------------
jerry@yct:~/ecat-33b922$ make modules_install
make -C "/usr/src/linux-headers-4.10.0-28-generic" M="/home/jerry/ecat-33b922" \
    INSTALL_MOD_DIR="ethercat" modules_install
make[1]: Entering directory '/usr/src/linux-headers-4.10.0-28-generic'
mkdir: cannot create directory ‘/lib/modules/4.10.0-28-generic/ethercat’: Permission denied
Makefile:1536: recipe for target '_emodinst_' failed
make[1]: *** [_emodinst_] Error 1
make[1]: Leaving directory '/usr/src/linux-headers-4.10.0-28-generic'
Makefile:934: recipe for target 'modules_install' failed
make: *** [modules_install] Error 2

jerry@yct:~/ecat-33b922$ sudo make modules_install
make -C "/usr/src/linux-headers-4.10.0-28-generic" M="/home/jerry/ecat-33b922" \
    INSTALL_MOD_DIR="ethercat" modules_install
make[1]: Entering directory '/usr/src/linux-headers-4.10.0-28-generic'
  INSTALL /home/jerry/ecat-33b922/devices/ec_generic.ko
At main.c:158:
- SSL error:02001002:system library:fopen:No such file or directory: bss_file.c:175
- SSL error:2006D080:BIO routines:BIO_new_file:no such file: bss_file.c:178
sign-file: certs/signing_key.pem: No such file or directory
  INSTALL /home/jerry/ecat-33b922/examples/mini/ec_mini.ko
At main.c:158:
- SSL error:02001002:system library:fopen:No such file or directory: bss_file.c:175
- SSL error:2006D080:BIO routines:BIO_new_file:no such file: bss_file.c:178
sign-file: certs/signing_key.pem: No such file or directory
  INSTALL /home/jerry/ecat-33b922/master/ec_master.ko
At main.c:158:
- SSL error:02001002:system library:fopen:No such file or directory: bss_file.c:175
- SSL error:2006D080:BIO routines:BIO_new_file:no such file: bss_file.c:178
sign-file: certs/signing_key.pem: No such file or directory
  DEPMOD  4.10.0-28-generic
make[1]: Leaving directory '/usr/src/linux-headers-4.10.0-28-generic'

jerry@yct:~/ecat-33b922$
-------------------------------------------------------------------------------------------------------

Thanks!






On Fri, Dec 1, 2017 at 10:37 AM, Eric Schwarz <eas@...> wrote:
Am 01.12.2017 16:15, schrieb Jerry Lian:

Thanks Eric for your explanation!
* (I know I can build the app within yocto build, but sometimes I prefer the SDK method outside yocto build.)

* So just copy manually? I thought there are some tricks  that I don't know.


 Acutally that's how we are doing it ...

Now I have follow-up question:
* It seems that "make install" will copy different files to different folders
* Can I do like these:?
--- make
--- copy (via scp)
--- make install
* Or it needs to be:
--- make
--- make install

 Install the cross-compiled stuff on your local disc e.g. into a "transfer" directory. Then copy it to the target.
On the target 'cp -Rfa <transfer-dir>/* /' should do it if the basic directory structure fits to your rootFS.

--- copy (via scp) -----then how do I know: what files to go which folder?
--- can you explain:
"... you may supply a custom install prefix directory to configure." --- how?


 There is IMHO no standardized variable available. Just look into the configure script how the variable is named and if it is actually available at all.


Thanks!

On Fri, Dec 1, 2017 at 9:34 AM, Eric Schwarz <eas@...> wrote:
Am 01.12.2017 15:16, schrieb Eric Schwarz:

Hi Jerry,

Am 01.12.2017 15:07, schrieb Jerry Lian:

I am new to embedded linux, and some concepts confuse me:
* I have downloaded an Autotools-based application to run on my board.
* If I boot the board with original image/original SDK, I normally build/install it with steps:
----- ./bootstrap
----- ./configure
----- make
----- make install
You may build your app within yocto build process so your application is contained in your rootFS directly.

* Now if I boot the board with yocto image without SDK, how can I deploy the application?
---- (surely I do "bitbake -c populate_sdk some-image", and source the environment!)

----- ./bootstrap (Yocto SDK on host)
----- ./configure (Yocto SDK on host)
----- make         (Yocto SDK on host)
----- make install (???)
If you build it w/ the SDK outside of the yocto build process just install it to a local directory and copy all the stuff w/ e.g. scp onto your boards rootFS.
 ... you may supply a custom install prefix directory to configure.

* How to do "make install"? (I want to install it on target, but how?)

Thanks!

Cheers
Eric
Re: Yocto SDK: how to deploy application with step "make install" to target?

Eric Schwarz
 

Am 01.12.2017 16:15, schrieb Jerry Lian:

Thanks Eric for your explanation!
* (I know I can build the app within yocto build, but sometimes I prefer the SDK method outside yocto build.)
* So just copy manually? I thought there are some tricks that I don't know.
Acutally that's how we are doing it ...

Now I have follow-up question:
* It seems that "make install" will copy different files to different folders
* Can I do like these:?
--- make
--- copy (via scp)
--- make install
* Or it needs to be:
--- make
--- make install
Install the cross-compiled stuff on your local disc e.g. into a "transfer" directory. Then copy it to the target.
On the target 'cp -Rfa <transfer-dir>/* /' should do it if the basic directory structure fits to your rootFS.

--- copy (via scp) -----then how do I know: what files to go which folder?
--- can you explain:
"... you may supply a custom install prefix directory to configure." --- how?
There is IMHO no standardized variable available. Just look into the configure script how the variable is named and if it is actually available at all.

Thanks!
On Fri, Dec 1, 2017 at 9:34 AM, Eric Schwarz <eas@...> wrote:
Am 01.12.2017 15:16, schrieb Eric Schwarz:
Hi Jerry,
Am 01.12.2017 15:07, schrieb Jerry Lian:
I am new to embedded linux, and some concepts confuse me:
* I have downloaded an Autotools-based application to run on my board.
* If I boot the board with original image/original SDK, I normally build/install it with steps:
----- ./bootstrap
----- ./configure
----- make
----- make install
You may build your app within yocto build process so your application is contained in your rootFS directly.
* Now if I boot the board with yocto image without SDK, how can I deploy the application?
---- (surely I do "bitbake -c populate_sdk some-image", and source the environment!)
----- ./bootstrap (Yocto SDK on host)
----- ./configure (Yocto SDK on host)
----- make (Yocto SDK on host)
----- make install (???)
If you build it w/ the SDK outside of the yocto build process just install it to a local directory and copy all the stuff w/ e.g. scp onto your boards rootFS.
... you may supply a custom install prefix directory to configure.

* How to do "make install"? (I want to install it on target, but how?)
Thanks!
Cheers
Eric
Re: Yocto SDK: how to deploy application with step "make install" to target?

Jerry Lian <jerry.lian@...>
 

Thanks Eric for your explanation! 
* (I know I can build the app within yocto build, but sometimes I prefer the SDK method outside yocto build.)
* So just copy manually? I thought there are some tricks  that I don't know.

Now I have follow-up question:
* It seems that "make install" will copy different files to different folders
* Can I do like these:?
--- make
--- copy (via scp) 
--- make install
* Or it needs to be:
--- make
--- make install
--- copy (via scp) -----then how do I know: what files to go which folder?
--- can you explain:
    "... you may supply a custom install prefix directory to configure." --- how?

Thanks!





On Fri, Dec 1, 2017 at 9:34 AM, Eric Schwarz <eas@...> wrote:
Am 01.12.2017 15:16, schrieb Eric Schwarz:

Hi Jerry,

Am 01.12.2017 15:07, schrieb Jerry Lian:

I am new to embedded linux, and some concepts confuse me:
* I have downloaded an Autotools-based application to run on my board.
* If I boot the board with original image/original SDK, I normally build/install it with steps:
----- ./bootstrap
----- ./configure
----- make
----- make install

You may build your app within yocto build process so your application is contained in your rootFS directly.

* Now if I boot the board with yocto image without SDK, how can I deploy the application?
---- (surely I do "bitbake -c populate_sdk some-image", and source the environment!)

----- ./bootstrap (Yocto SDK on host)
----- ./configure (Yocto SDK on host)
----- make         (Yocto SDK on host)
----- make install (???)

If you build it w/ the SDK outside of the yocto build process just install it to a local directory and copy all the stuff w/ e.g. scp onto your boards rootFS.

 ... you may supply a custom install prefix directory to configure.



* How to do "make install"? (I want to install it on target, but how?)

Thanks!

Cheers
Eric
Re: Yocto SDK: how to deploy application with step "make install" to target?

Eric Schwarz
 

Am 01.12.2017 15:16, schrieb Eric Schwarz:

Hi Jerry,
Am 01.12.2017 15:07, schrieb Jerry Lian:

I am new to embedded linux, and some concepts confuse me:
* I have downloaded an Autotools-based application to run on my board.
* If I boot the board with original image/original SDK, I normally build/install it with steps:
----- ./bootstrap
----- ./configure
----- make
----- make install
You may build your app within yocto build process so your application is contained in your rootFS directly.

* Now if I boot the board with yocto image without SDK, how can I deploy the application?
---- (surely I do "bitbake -c populate_sdk some-image", and source the environment!)
----- ./bootstrap (Yocto SDK on host)
----- ./configure (Yocto SDK on host)
----- make (Yocto SDK on host)
----- make install (???)
If you build it w/ the SDK outside of the yocto build process just install it to a local directory and copy all the stuff w/ e.g. scp onto your boards rootFS.
... you may supply a custom install prefix directory to configure.


* How to do "make install"? (I want to install it on target, but how?)
Thanks!
Cheers
Eric
Re: Yocto SDK: how to deploy application with step "make install" to target?

Eric Schwarz
 

Hi Jerry,

Am 01.12.2017 15:07, schrieb Jerry Lian:

I am new to embedded linux, and some concepts confuse me:
* I have downloaded an Autotools-based application to run on my board.
* If I boot the board with original image/original SDK, I normally build/install it with steps:
----- ./bootstrap
----- ./configure
----- make
----- make install
You may build your app within yocto build process so your application is contained in your rootFS directly.

* Now if I boot the board with yocto image without SDK, how can I deploy the application?
---- (surely I do "bitbake -c populate_sdk some-image", and source the environment!)
----- ./bootstrap (Yocto SDK on host)
----- ./configure (Yocto SDK on host)
----- make (Yocto SDK on host)
----- make install (???)
If you build it w/ the SDK outside of the yocto build process just install it to a local directory and copy all the stuff w/ e.g. scp onto your boards rootFS.

* How to do "make install"? (I want to install it on target, but how?)
Thanks!
Cheers
Eric
Yocto SDK: how to deploy application with step "make install" to target?

Jerry Lian <jerry.lian@...>
 

I am new to embedded linux, and some concepts confuse me:
* I have downloaded an Autotools-based application to run on my board.
* If I boot the board with original image/original SDK, I normally build/install it with steps:
----- ./bootstrap
----- ./configure
----- make
----- make install
* Now if I boot the board with yocto image without SDK, how can I deploy the application?
---- (surely I do "bitbake -c populate_sdk some-image", and source the environment!)
----- ./bootstrap (Yocto SDK on host)
----- ./configure (Yocto SDK on host)
----- make         (Yocto SDK on host)
----- make install (???)                   

* How to do "make install"? (I want to install it on target, but how?)

Thanks!
Re: dev package non-sym link

Alexander Kanavin <alexander.kanavin@...>
 

On 11/30/2017 11:46 PM, Rail Shafigulin wrote:
do_install () {
install -d ${D}${libdir}/
install -m 0755 ${S}/libAnokiWave.so.1.0.1 ${D}${libdir}
ln -s ${D}${libdir}/libAnokiWave.so.1.0.1 ${D}${libdir}/libAnokiwave.so.1
ln -s ${D}${libdir}/libAnokiWave.so.1 ${D}${libdir}/libAnokiwave.so
}
You need to call 'ln -sr', then you'll get relative symlinks. Better yet, add some build system support to the project (meson, cmake), then it'll take care of these things for you.

Alex
Re: dev package non-sym link

Rail Shafigulin <rail@...>
 

Check whether the situation is the same in .../image/usr/lib. If so, then
the upstream is installing .so files incorrectly and you need to figure out
what it does and how to fix it so that it symlinks .so instead of copying or
hardlinking it.

Alex
Thanks for the reply. I changed do_install of my recipe

do_install () {
install -d ${D}${libdir}/
install -m 0755 ${S}/libAnokiWave.so.1.0.1 ${D}${libdir}
}

Now the error is gone and there are no warnings but if I check

/tmp/work/cortexa9hf-neon-xilinx-linux-gnueabi/anokiwavelib/1.0+gitAUTOINC+16df62468c-r0/package/usr/lib/

all I see is libAnokiWave.so.1.0.1

If I change my install task to

do_install () {
install -d ${D}${libdir}/
install -m 0755 ${S}/libAnokiWave.so.1.0.1 ${D}${libdir}
ln -s ${D}${libdir}/libAnokiWave.so.1.0.1 ${D}${libdir}/libAnokiwave.so.1
ln -s ${D}${libdir}/libAnokiWave.so.1 ${D}${libdir}/libAnokiwave.so
}

I get the following warnings

WARNING: anokiwavelib-1.0+gitAUTOINC+16df62468c-r0 do_package_qa: QA
Issue: Symlink /usr/lib/libAnokiwave.so.1 in anokiwavelib points to
TMPDIR [symlink-to-sysroot]
WARNING: anokiwavelib-1.0+gitAUTOINC+16df62468c-r0 do_package_qa: QA
Issue: Symlink /usr/lib/libAnokiwave.so in anokiwavelib-dev points to
TMPDIR [symlink-to-sysroot]

I'm not sure how to fix these issues. The links, libAnokiWave.so.1 and
libAnokiWave.so, must exist but right now they don't point to the
right location or doen't exist at all.

Would appreciate any help with this.

Rail Shafigulin
Software Engineer
Esencia Technologies

--




*ESENCIA TECHNOLOGIES, INC.*3945 Freedom Circle, Suite #360,
Santa Clara CA 95054
________________________________________________________

Phone: +1 408 736 8284 Fax: +1 408 519 3475
http://www.esenciatech.com | http://www.lnttechservices.com
Kernel Crash when moving from Pyro to Rocko (or master): change in kernel compile process or flags?

Michael
 

Hi,
I am experiencing a Kernel crash when moving my system from Yocto 2.3 (Pyro) to Rocko or Master.
Target system is an ARMv7 (NXP Layerscape LS1021a, built on Cortex-A7).
This is what I did:
1) switched poky version
2) re-built same image in a new, blank directory

Results:
Root-FS works fine, the kernel crashes with an undefined instruction oops.
Just exchanging the kernel zImage with the one compiled with Pyro (same recipe), my system works fine again.

What I tried so far:
1) checked and compared the kernel configs generated: 100% identical
2) switched gcc back to the version used in Pyro (6.3) instead of 7.2: kernel still crashes, same oops, no changes. So it's not related to GCC 7.2
3) checked on different boards: same issue for the same board type, but the switch to rocko worked fine for a system based on Cortex-A9.

I am using separate tmp directories and am not sharing sstate. I always started builds in blank build directories.
The error is reproducible (100%)
It was also present with all of the various snapshots of the master branch I tried

What I suspect: there are some changes in the ways that gcc is called from pyro to rocko. Maybe one of the switches, or features, etc
Before deep-diving into the trial and error phase of reproducing the CFLAGS, etc:
Is anyone experiencing the same issues?
* What exactly has changed in the way the kernel build is called (methods, compile/link flags, etc) between pyro and rocko? *
* Is there a way to (easily) disable some of the features to test which one causes the issue? *
* Any suspicions about what could cause this? *

Pls. find a log of the crash below, and an excerpt of my kernel recipe and machine configuration.

Thanks in advance for your help,
Best regards,
Michael

-------------
VFS: Mounted root (ext4 filesystem) on device 179:1.
devtmpfs: mounted
Freeing unused kernel memory: 2048K
Internal error: Oops - undefined instruction: 0 [#1] SMP THUMB2
Modules linked in:
CPU: 1 PID: 1 Comm: init Not tainted 4.14.2-avnet-nxp #2
Hardware name: Freescale LS1021A
task: bf0a0000 task.stack: bf09a000
PC is at ret_fast_syscall+0x2/0x52
LR is at SyS_brk+0x109/0x128
pc : [<80205fc2>] lr : [<802c513d>] psr: 60000013
sp : bf09bfa8 ip : 70c5387d fp : 00001000
r10: 00000000 r9 : bf09a000 r8 : 802061a4
r7 : 0000002d r6 : 76f71bac r5 : 00010034 r4 : 00000009
r3 : 000273c8 r2 : ffff0001 r1 : 00000000 r0 : 00028000
Flags: nZCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none
Control: 70c5387d Table: be7d1000 DAC: fffbfff3
Process init (pid: 1, stack limit = 0xbf09a210)
Stack: (0xbf09bfa8 to 0xbf09c000)
bfa0: 00000009 00010034 00000000 7ee86fea 00000000 00000001
bfc0: 00000009 00010034 76f71bac 0000002d 00000000 00000001 00000001 00001000
bfe0: 7ee86f14 7ee86e4c 76f87a58 76f88e88 60000010 00000000 00000000 00000000
Code: 20012000 0000bd08 00000000 b6720000 (2008f8d9)
---[ end trace 7bab6dceda641aab ]---
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b

CPU0: stopping
CPU: 0 PID: 0 Comm: swapper/0 Tainted: G D 4.14.2-avnet-nxp #2
Hardware name: Freescale LS1021A
[<8020bbe9>] (unwind_backtrace) from [<80208e77>] (show_stack+0xb/0xc)
[<80208e77>] (show_stack) from [<8066af19>] (dump_stack+0x69/0x78)
[<8066af19>] (dump_stack) from [<8020b1d5>] (handle_IPI+0x261/0x274)
[<8020b1d5>] (handle_IPI) from [<802012b1>] (gic_handle_irq+0x61/0x64)
[<802012b1>] (gic_handle_irq) from [<8067d953>] (__irq_svc+0x53/0x7c)
Exception stack(0x80e01f38 to 0x80e01f80)
1f20: 00000000 000027d0
1f40: 3eb79000 80212641 ffffe000 80e03cd0 80e03c74 80e7f3ab 8096e480 80e03c40
1f60: 80e8ce00 80c42a30 01400000 80e01f88 80206a3d 80206a3e 40000033 ffffffff
[<8067d953>] (__irq_svc) from [<80206a3e>] (arch_cpu_idle+0x22/0x24)
[<80206a3e>] (arch_cpu_idle) from [<802405bf>] (do_idle+0x9f/0xe4)
[<802405bf>] (do_idle) from [<802407bb>] (cpu_startup_entry+0x13/0x14)
[<802407bb>] (cpu_startup_entry) from [<80c00a05>] (start_kernel+0x305/0x310)
---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
-----------------------

Excerpt from linux_xxx.bb:
----
DEPENDS_kernel-base += "kernel-devicetree"

LINUX_VERSION = "4.14-stable"
S = "${WORKDIR}/git"
PR = "r1"
SRCREV = "8292fd8d726105abc01dae26d0a2cddcf53d4e0f"
PV = "4.14.2"
SRC_URI = " \
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git;branch=master;protocol=git;nobranch=1 \
"
SRC_URI_append_ls1021a = " file://defconfig "
SRC_URI_append_ls1021a = " file://pcie.cfg file://caam.cfg file://usb.cfg file://bt.cfg "
----

Excerpt from machine.conf
---
SOC_FAMILY = "ls1021a"
TARGET_ARCH = "arm"
DEFAULTTUNE = "cortexa7hf-neon-vfpv4"
IMAGE_INSTALL_append = " kernel-modules kernel-devicetree kernel-image"
KERNEL_IMAGETYPE = "zImage"

KERNEL_EXTRA_ARGS += "LOADADDR=${UBOOT_ENTRYPOINT}"
EXTRA_IMAGEDEPENDS += "virtual/bootloader"
----
Re: Security upgrade stategy

Andre McCurdy <armccurdy@...>
 

On Wed, Nov 29, 2017 at 11:24 AM, Khem Raj <raj.khem@...> wrote:
On Wed, Nov 29, 2017 at 11:06 AM, Brian Smucker <bds@...> wrote:
Hello,

I have an older yocto-based image built with Danny.

We have a need to update two components, php and dropbear, to the most
modern versions for security purposes.

What is the preferred way to do this without moving the whole image to the
latest version of poky/yocto? That would be a big job, which would
necessitate migrating my linux dev environment to something newer, in
addition to other things.

Or is there a preferred way? How do people handle this?
We usually backport the needed fixes back into release branches.
Version upgrades
are usually not the norm. However, in your case where are on danny which might
have stopped to receive any security fixes already
To avoid and doubt, danny has definitely stopped receiving any security fixes!

, it might be
something you can
undertake in your distro to upgrade the versions. Take the new version
from latest
release or master and retrofit it into the release you are on then
fix/integrate it into
rest of your distro. However here you have forked and can keep
following the same
procedure for subsequent upgrades for these packages. Obviously, you might see
integration issues but thats expected.

Thanks,

Brian
--
_______________________________________________
yocto mailing list
yocto@...
https://lists.yoctoproject.org/listinfo/yocto
--
_______________________________________________
yocto mailing list
yocto@...
https://lists.yoctoproject.org/listinfo/yocto
Re: dev package non-sym link

Alexander Kanavin <alexander.kanavin@...>
 

On 11/30/2017 08:10 PM, Rail Shafigulin wrote:
Somewhat new to Yocto. Trying to build a library. It compiled but I'm
having issues during the package task. The error output is
ERROR: anokiwavelib-1.0+gitAUTOINC+16df62468c-r0 do_package_qa: QA
Issue: -dev package contains non-symlink .so: anokiwavelib-dev path
'/work/cortexa9hf-neon-xilinx-linux-gnueabi/anokiwavelib/1.0+gitAUTOINC+16df62468c-r0/packages-split/anokiwavelib-dev/usr/lib/libAnokiWave.so'
[dev-elf]
ERROR: anokiwavelib-1.0+gitAUTOINC+16df62468c-r0 do_package_qa: QA run
found fatal errors. Please consider fixing them.
ERROR: anokiwavelib-1.0+gitAUTOINC+16df62468c-r0 do_package_qa:
Function failed: do_package_qa
ERROR: Logfile of failure stored in:
/home/someuser/projects/petalinux-build-system/petalinux-build-env/tmp/work/cortexa9hf-neon-xilinx-linux-gnueabi/anokiwavelib/1.0+gitAUTOINC+16df62468c-r0/temp/log.do_package_qa.5765
ERROR: Task (/home/someuser/projects/petalinux-build-system/sources/core/../meta-customlayer/recipes-anokiwave/lib/anokiwavelib_git.bb:do_package_qa)
failed with exit code '1'
I do understand the issue. It doesn't create links for libAnokiWave.so
and libAnokiWave.so.1. I do see the following in
/home/someuser/projects/petalinux-build-system/petalinux-build-env/tmp/work/cortexa9hf-neon-xilinx-linux-gnueabi/anokiwavelib/1.0+gitAUTOINC+16df62468c-r0/package/usr/lib
Check whether the situation is the same in .../image/usr/lib. If so, then the upstream is installing .so files incorrectly and you need to figure out what it does and how to fix it so that it symlinks .so instead of copying or hardlinking it.

Alex
dev package non-sym link

Rail Shafigulin <rail@...>
 

Somewhat new to Yocto. Trying to build a library. It compiled but I'm
having issues during the package task. The error output is

ERROR: anokiwavelib-1.0+gitAUTOINC+16df62468c-r0 do_package_qa: QA
Issue: -dev package contains non-symlink .so: anokiwavelib-dev path
'/work/cortexa9hf-neon-xilinx-linux-gnueabi/anokiwavelib/1.0+gitAUTOINC+16df62468c-r0/packages-split/anokiwavelib-dev/usr/lib/libAnokiWave.so'
[dev-elf]
ERROR: anokiwavelib-1.0+gitAUTOINC+16df62468c-r0 do_package_qa: QA run
found fatal errors. Please consider fixing them.
ERROR: anokiwavelib-1.0+gitAUTOINC+16df62468c-r0 do_package_qa:
Function failed: do_package_qa
ERROR: Logfile of failure stored in:
/home/someuser/projects/petalinux-build-system/petalinux-build-env/tmp/work/cortexa9hf-neon-xilinx-linux-gnueabi/anokiwavelib/1.0+gitAUTOINC+16df62468c-r0/temp/log.do_package_qa.5765
ERROR: Task (/home/someuser/projects/petalinux-build-system/sources/core/../meta-customlayer/recipes-anokiwave/lib/anokiwavelib_git.bb:do_package_qa)
failed with exit code '1'


I do understand the issue. It doesn't create links for libAnokiWave.so
and libAnokiWave.so.1. I do see the following in
/home/someuser/projects/petalinux-build-system/petalinux-build-env/tmp/work/cortexa9hf-neon-xilinx-linux-gnueabi/anokiwavelib/1.0+gitAUTOINC+16df62468c-r0/package/usr/lib

drwxr-xr-x 3 someuser someuser 4096 Nov 29 14:10 ./
drwxr-xr-x 4 someuser someuser 4096 Nov 29 14:10 ../
drwxr-xr-x 2 someuser someuser 4096 Nov 29 14:10 .debug/
-rw-r--r-- 2 someuser someuser 45084 Nov 29 14:10 libAnokiWave.so
-rw-r--r-- 2 someuser someuser 45088 Nov 29 14:10 libAnokiWave.so.1
-rwxr-xr-x 2 someuser someuser 45092 Nov 29 14:10 libAnokiWave.so.1.0.1*

and .debug contains

drwxr-xr-x 2 someuser someuser 4096 Nov 29 14:10 ./
drwxr-xr-x 3 someuser someuser 4096 Nov 29 14:10 ../
-rwxr-xr-x 2 someuser someuser 81500 Nov 29 14:10 libAnokiWave.so*
-rwxr-xr-x 2 someuser someuser 81500 Nov 29 14:10 libAnokiWave.so.1*
-rwxr-xr-x 2 someuser someuser 81500 Nov 29 14:10 libAnokiWave.so.1.0.1*

Which is incorrect, libAnokiWave.so and libAnokiWave.so.1 are not links.

I'm not really sure how to fix this. Would appreciate any help with it.

--
Rail Shafigulin
Software Engineer
Esencia Technologies

--




*ESENCIA TECHNOLOGIES, INC.*3945 Freedom Circle, Suite #360,
Santa Clara CA 95054
________________________________________________________

Phone: +1 408 736 8284 Fax: +1 408 519 3475
http://www.esenciatech.com | http://www.lnttechservices.com
18821 - 18840 of 57786