[meta-freescale] [meta-fsl-arm][PATCH] linux-mfgtool: Prevent overwriting linux-imx deployments

[Sven Ebenfeld]

Hi Daiane, Thanks for your comments.

Am 19.11.2014 um 16:07 schrieb Daiane Angolini:
> Hi Sven Ebenfeld, sorry for delay.
> 
> On Tue, Nov 11, 2014 at 7:56 AM, Sven Ebenfeld <sven.ebenfeld at gmail.com> wrote:
>> When I try building fsl-image-mfgtool-initramfs it builds both,
> 
> Could you, please, use "When building blablabla" instead of using
> "When I try building blablabla"?
> 
> Who is the "person" behind "I" in a commit log?

Okay, I will not use this wording in my next commit logs.

> 
>> linux-imx and linux-mfgtool. Both recipes are trying to write
>> the README... file and KERNEL_IMAGETYPE-file into deploy dir.
> 
> I don´t understand why you add the ... after the README file name.
> Would it be possibly avoided?

The README file name is quite a long name I tried to avoid mentioning
the whole name.
But if it's better to mention the whole name for better understanding, I
will mention it. Or even leave the whole name away.

> 
>> This leads to an error, that one recipe wants to write files that
>> already exist. This patch prevents linux-mfgtool from overwriting
>> linux-imx's results.
>>
>> Signed-off-by: Sven Ebenfeld <sven.ebenfeld at gmail.com>
>> ---
>>  recipes-kernel/linux/linux-mfgtool.inc | 26 ++++++++++++++++++++++++++
>>  1 file changed, 26 insertions(+)
>>
>> diff --git a/recipes-kernel/linux/linux-mfgtool.inc b/recipes-kernel/linux/linux-mfgtool.inc
>> index 106a7d9..eaccb44 100644
>> --- a/recipes-kernel/linux/linux-mfgtool.inc
>> +++ b/recipes-kernel/linux/linux-mfgtool.inc
>> @@ -79,3 +79,29 @@ do_install_append() {
>>  sysroot_stage_all() {
>>      :
>>  }
>> +
>> +# FIXME: Overwrite kernel_do_deploy as it tries to overwrite (u/z)Image symlink
>> +#        and DO_NOT...DELETE.txt file. This leads to an error that this recipe
> 
> Here again you use ...
> Would you be trying to replace a long file name?
> 
> I understand this is only a workaround for the problem, am I right?

Yes, it's only a workaround. Unfortunately I don't know a better way for
fixing this, yet.
The same problem exists with u-boot.

Do you think it's possible to get it merged with a better commit log or
do you have a proposal where I should look for a better fix?

> 
> 
> Daiane
> 
>> +#        is trying to install files into an area where those files already exist.
>> +kernel_do_deploy() {
>> +       install -m 0644 ${KERNEL_OUTPUT} ${DEPLOYDIR}/${KERNEL_IMAGE_BASE_NAME}.bin
>> +       if [ ${MODULE_TARBALL_DEPLOY} = "1" ] && (grep -q -i -e '^CONFIG_MODULES=y$' .config); then
>> +               mkdir -p ${D}/lib
>> +               tar -cvzf ${DEPLOYDIR}/${MODULE_TARBALL_BASE_NAME} -C ${D} lib
>> +               ln -sf ${MODULE_TARBALL_BASE_NAME} ${DEPLOYDIR}/${MODULE_TARBALL_SYMLINK_NAME}
>> +       fi
>> +
>> +       ln -sf ${KERNEL_IMAGE_BASE_NAME}.bin ${DEPLOYDIR}/${KERNEL_IMAGE_SYMLINK_NAME}.bin
>> +       ln -sf ${KERNEL_IMAGE_BASE_NAME}.bin ${DEPLOYDIR}/mfgtool-${KERNEL_IMAGETYPE}
>> +
>> +       cd ${B}
>> +       # Update deploy directory
>> +       if [ -e "${KERNEL_OUTPUT}.initramfs" ]; then
>> +               echo "Copying deploy kernel-initramfs image and setting up links..."
>> +               initramfs_base_name=${INITRAMFS_BASE_NAME}
>> +               initramfs_symlink_name=mfgtool-${KERNEL_IMAGETYPE}-initramfs-${MACHINE}
>> +               install -m 0644 ${KERNEL_OUTPUT}.initramfs ${DEPLOYDIR}/${initramfs_base_name}.bin
>> +               cd ${DEPLOYDIR}
>> +               ln -sf ${initramfs_base_name}.bin ${initramfs_symlink_name}.bin
>> +       fi
>> +}
>> --
>> 1.8.1.rc3
>>
>> --
>> _______________________________________________
>> meta-freescale mailing list
>> meta-freescale at yoctoproject.org
>> https://lists.yoctoproject.org/listinfo/meta-freescale