bind mount for /downloads not happening

[chombourger@gmail.com]

[-- Attachment #1.1: Type: text/plain, Size: 1994 bytes --]

Hi Jan,

I was in the process of rebasing to the latest Isar/next to discard some of 
the local commits we had.
In particular we had a change from Christian to make git trees available to 
the build process.
This change should no longer be required thanks to 
9a8b7fe77cbd006b6053f85e593301c60ea0cfbd (Make sure git repos are usable 
inside buildchroot)

It unfortunately did not quite work for me. I believe I understand why and 
wanted to run my theory by you to double-check!.

# 
https://github.com/ilbers/isar/blob/next/meta/classes/dpkg-base.bbclass#L44
do_build() { 
     mkdir -p ${BUILDROOT} 
     sudo mount --bind ${WORKDIR} ${BUILDROOT} 

     sudo flock ${MOUNT_LOCKFILE} -c ' \ 
         if ! grep -q ${BUILDCHROOT_DIR}/isar-apt /proc/mounts; then \ 
            mount --bind ${DEPLOY_DIR_APT}/${DISTRO} 
${BUILDCHROOT_DIR}/isar-apt; \ 
            mount --bind ${DL_DIR} ${BUILDCHROOT_DIR}/downloads; \ 
            mount -t devtmpfs -o mode=0755,nosuid devtmpfs 
${BUILDCHROOT_DIR}/dev; \ 
            mount -t proc none ${BUILDCHROOT_DIR}/proc; \ 
        fi' 

We check if isar-apt was mounted to mount everything else. OK.

We however have the following code also mounting isar-apt:

# 
https://github.com/ilbers/isar/blob/next/meta/recipes-devtools/buildchroot/buildchroot.bb#L68
do_build() {
    ...
    sudo mount --bind ${DEPLOY_DIR_APT}/${DISTRO} 
${BUILDCHROOT_DIR}/isar-apt 
    sudo mount -t devtmpfs -o mode=0755,nosuid devtmpfs 
${BUILDCHROOT_DIR}/dev 
    sudo mount -t proc none ${BUILDCHROOT_DIR}/proc
    ...

If you do a full platform build, isar-apt will get mounted from the 
buildchroot recipe and our downloads folder will not be mounted from 
dpkg-base.bb (and cause our swiupdate recipe to fail)

I would be tempted to add a bind mount for downloads in the buildchroot 
recipe. The other option would be add a check in dpkg-base.bb for downloads 
being mounted (we would then have 2 "if" statements in the code block 
passed to flock.

Thoughts?

Cedric

[-- Attachment #1.2: Type: text/html, Size: 5260 bytes --]