Skip to contentSkip to conversation

Added absolute paths to checks when copying the device tree

PATCH In review Werner Grift 2 participants 9 messages Feb 11, 2015 – Feb 17, 2015

lore
Share & tools
lore
#0 lore raw

PATCH 73/73 Added absolute paths to checks when copying the device tree

---
 classes/image_types_fsl.bbclass | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
classes/image_types_fsl.bbclass +3 -3
index 53f8767..91c418a 100644
@@ -168,9 +168,9 @@ generate_imx_sdcard () {
 	if test -n "${KERNEL_DEVICETREE}"; then
 		for DTS_FILE in ${KERNEL_DEVICETREE}; do
 			DTS_BASE_NAME=`basename ${DTS_FILE} | awk -F "." '{print $1}'`
-			if [ -e "${KERNEL_IMAGETYPE}-${DTS_BASE_NAME}.dtb" ]; then
-				kernel_bin="`readlink ${KERNEL_IMAGETYPE}-${MACHINE}.bin`"
-				kernel_bin_for_dtb="`readlink ${KERNEL_IMAGETYPE}-${DTS_BASE_NAME}.dtb | sed "s,$DTS_BASE_NAME,${MACHINE},g;s,\.dtb$,.bin,g"`"
+			if [ -e "${DEPLOY_DIR_IMAGE}/${KERNEL_IMAGETYPE}-${DTS_BASE_NAME}.dtb" ]; then
+				kernel_bin="`readlink ${DEPLOY_DIR_IMAGE}/${KERNEL_IMAGETYPE}-${MACHINE}.bin`"
+				kernel_bin_for_dtb="`readlink ${DEPLOY_DIR_IMAGE}/${KERNEL_IMAGETYPE}-${DTS_BASE_NAME}.dtb | sed "s,$DTS_BASE_NAME,${MACHINE},g;s,\.dtb$,.bin,g"`"
 				if [ $kernel_bin = $kernel_bin_for_dtb ]; then
 					mcopy -i ${WORKDIR}/boot.img -s ${DEPLOY_DIR_IMAGE}/${KERNEL_IMAGETYPE}-${DTS_BASE_NAME}.dtb ::/${DTS_BASE_NAME}.dtb
 				fi
Signature · 1 line
2.0.0
#1 lore raw

Hello Werner,

Quoted 1 line · Werner Grift

I am wondering how this has been working previously. Any idea what changed which causes the behaviour change?

Signature · 3 lines
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750
#2 lore raw

Hi Otavio.

I received the recipes (and a working version of it on our build server) from a colleague that only introduced the following patch
<http://c7b9611ad0ead17a624fc73a60c321ff249c2214> to oe at rev c7b9611ad0ead17a624fc73a60c321ff249c2214. The rest of the system was untouched.

I made a fresh build system on my development machine. My build ran a couple of times without above mentioned patch, but broke because it was not applied. After applying the patch my build worked, but the copy device tree issue remained. (But not on the build server) I then ran ./oebb.sh update to make sure everything is cozy (which removed the patch mentioned above without me realizing) and miraculously the build continued to succeed, but the device tree issue remained. I suspect I broke sstate by doing this. My sstate is not shared between dev and build.

Therefor, I suspect that the issue was generated in house. Regardless, the rest of create_image.sdcard does not make assumptions as to pwd, so I figured the patch is needed for completeness.

Regards,

Werner.

Quoted 14 lines · Otavio Salvador

Hello Werner,

Quoted 1 line · Werner Grift

I am wondering how this has been working previously. Any idea what changed which causes the behaviour change?

--
Otavio Salvador O.S. Systems
http://www.ossystems.com.br http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854 Mobile: +1 (347) 903-9750

#3 lore raw

I screwed up the patch link in my previous email:
http://hastebin.com/yaxisokuqo.py

Quoted 45 lines · Werner Grift

Hi Otavio.

I received the recipes (and a working version of it on our build server) from a colleague that only introduced the following patch
<http://c7b9611ad0ead17a624fc73a60c321ff249c2214> to oe at rev c7b9611ad0ead17a624fc73a60c321ff249c2214. The rest of the system was untouched.

I made a fresh build system on my development machine. My build ran a couple of times without above mentioned patch, but broke because it was not applied. After applying the patch my build worked, but the copy device tree issue remained. (But not on the build server) I then ran ./oebb.sh update to make sure everything is cozy (which removed the patch mentioned above without me realizing) and miraculously the build continued to succeed, but the device tree issue remained. I suspect I broke sstate by doing this. My sstate is not shared between dev and build.

Therefor, I suspect that the issue was generated in house. Regardless, the rest of create_image.sdcard does not make assumptions as to pwd, so I figured the patch is needed for completeness.

Regards,

Werner.

Quoted 14 lines · Otavio Salvador

Hello Werner,

Quoted 1 line · Werner Grift

I am wondering how this has been working previously. Any idea what changed which causes the behaviour change?

--
Otavio Salvador O.S. Systems
http://www.ossystems.com.br http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854 Mobile: +1 (347) 903-9750

#4 lore raw
In reply to Werner Grift

On Tue, Feb 17, 2015 at 5:45 AM, Werner Grift <[email protected]> wrote:
...

Quoted 3 lines

without me realizing) and miraculously the build continued to succeed, but the device tree issue remained. I suspect I broke sstate by doing this. My sstate is not shared between dev and build.

...

Which device tree issue?

We need to understand where the issue come from as this is a code used by all people using meta-fsl-arm and I'd like to avoid to apply something we don't fully understand the need.

Signature · 3 lines
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750
#5 lore raw

The device tree issue I refer too is the device tree file that is not being copied to the boot partition because the generated script create_image.sdcard contains if statements that assume that the current working directory is ${DEPLOY_DIR_IMAGE} which it is not. In my case it was ${OE_BASE} which makes sense since it is where bitbake is called from.

Quoted 21 lines · Otavio Salvador

On Tue, Feb 17, 2015 at 5:45 AM, Werner Grift <[email protected]> wrote:
...

Quoted 1 line

without me realizing) and miraculously the build continued to succeed,

but

Quoted 1 line

the device tree issue remained. I suspect I broke sstate by doing this.

My

Quoted 1 line

sstate is not shared between dev and build.

...

Which device tree issue?

We need to understand where the issue come from as this is a code used by all people using meta-fsl-arm and I'd like to avoid to apply something we don't fully understand the need.

--
Otavio Salvador O.S. Systems
http://www.ossystems.com.br http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854 Mobile: +1 (347) 903-9750

#6 lore raw
Quoted 5 lines · Werner Grift

The device tree issue I refer too is the device tree file that is not being copied to the boot partition because the generated script create_image.sdcard contains if statements that assume that the current working directory is ${DEPLOY_DIR_IMAGE} which it is not. In my case it was ${OE_BASE} which makes sense since it is where bitbake is called from.

How is your platform setup?

FSL Community BSP usually looks like:

.

├── build
│   └── conf
├── README
├── setup-environment
└── sources
    ├── base
    ├── Documentation
    ├── meta-fsl-arm
    ├── meta-fsl-arm-extra
    ├── meta-fsl-demos
    ├── meta-openembedded
    └── poky

Is yours same layout?

Signature · 3 lines
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750
#7 lore raw

We use the Angstrom distro 1.7

Our structure does not look like that. Whats this Yocto business?

Looks like script is being called from the wrong directory then.

Quoted 35 lines · Otavio Salvador
Quoted 1 line · Werner Grift

The device tree issue I refer too is the device tree file that is not

being

Quoted 3 lines

copied to the boot partition because the generated script create_image.sdcard contains if statements that assume that the current working directory is ${DEPLOY_DIR_IMAGE} which it is not. In my case it

was

Quoted 1 line

${OE_BASE} which makes sense since it is where bitbake is called from.

How is your platform setup?

FSL Community BSP usually looks like:

.

├── build
│   └── conf
├── README
├── setup-environment
└── sources
    ├── base
    ├── Documentation
    ├── meta-fsl-arm
    ├── meta-fsl-arm-extra
    ├── meta-fsl-demos
    ├── meta-openembedded
    └── poky

Is yours same layout?

--
Otavio Salvador O.S. Systems
http://www.ossystems.com.br http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854 Mobile: +1 (347) 903-9750

#8 lore raw
Quoted 1 line · Werner Grift

We use the Angstrom distro 1.7

Ok.

Quoted 1 line

Our structure does not look like that. Whats this Yocto business?

This is the FSL Community BSP. You can find more information about it in http://freescale.github.io

Quoted 1 line

Looks like script is being called from the wrong directory then.

Nothing wrong. The system must to behave right and to support all different type of platform setups.

Can you give me (might be private - if preferable) more information about your platform layout?

Signature · 3 lines
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750