fdtfile versus fdt_file

[jsrc27]

OSTree has long since adopted the use of fdt_file as the variable to track device tree names across deployments. In the Yocto-built Linux distribution that my company maintains however, we have been maintaining a patch for OSTree that adds support for fdtfile as that is what we use instead of fdt_file.

Ideally we'd like to drop this patch to have less maintenance burden. I was looking into why OSTree initially adopted the use of fdt_file, and as far as I can tell it was somehow determined to be the more prevalent variable name in U-Boot: ce2995e

That said, U-Boot actually documents fdtfile as the standard variable name to use: https://source.denx.de/u-boot/u-boot/-/blob/master/doc/usage/environment.rst

It seems like this has been documented as such for quite a while: https://source.denx.de/u-boot/u-boot/-/commit/dc0b7b0e619e3f1413e626dde193fe21427ac831

Given this, I wanted to ask if it would be reasonable for OSTree to consider the use of fdtfile. Or if there was some other reason unknown to me that fdt_file is used. More realistically I imagine you'd want to have fdtfile exist in parallel alongside fdt_file. As changing fdt_file at this point risks breaking compatibility on other distros/setups.

If the idea to add support for fdtfile has no issues, then we would be happy to submit a PR for this as well.

[cgwalters]

If the idea to add support for fdtfile has no issues, then we would be happy to submit a PR for this as well.

I'll be honest, my experience and knowledge is generally very datacenter/PC focused and I personally do not have extensive experience with devicetree specifically.

I can say first at a procedural level: Please please post any downstream patches that might exist upstream - at the very least in draft/RFC form - so they're more visible and discoverable and give a chance for upstream review. This is an aside but...and this one has long probably slipped out of most people's memory but I suspect e.g. the Debian openssl CVE would have more likely not happened had at least upstream had a chance to see the patch, but AFAIK it was never sent, just patched downstream.

Anyways, yes, at the very least let's get a PR up and discuss it there. Thanks!