So testing that libcamera PR is opening a can of worms.
A large number of things have changed with the streams API, and upstreaming drivers. We're in a slight state of flux with a load of downstream patches that no longer work with upstream, and general weirdness.
That libcamera PR matches with a backport to kernel 6.6, and is taking a load changes that we potentially don't want right now (changing of control IDs that aren't quite in upstream, so could change yet again).
We need to arrange a similar kernel PR for 6.12 to fix things up. In theory it should just be a subset of those for 6.6, however things aren't always that trivial. It's also chicken and egg in that the 6.12 does work with the current libcamera tree due to all the downstream patches, so we need to change all 3 (libcamera, 6.6, and 6.12) in sync. It'll be fun!
I have managed to get my FPDlink devices nominally streaming, but only by hardcoding the formats, and the images are corrupt.
I'm recompiling my 6.6 branch to ensure that there isn't a hardware issue with my setup, but I think this is that we need to resolve the differences between mainline and downstream.
A large number of things have changed with the streams API, and upstreaming drivers. We're in a slight state of flux with a load of downstream patches that no longer work with upstream, and general weirdness.
That libcamera PR matches with a backport to kernel 6.6, and is taking a load changes that we potentially don't want right now (changing of control IDs that aren't quite in upstream, so could change yet again).
We need to arrange a similar kernel PR for 6.12 to fix things up. In theory it should just be a subset of those for 6.6, however things aren't always that trivial. It's also chicken and egg in that the 6.12 does work with the current libcamera tree due to all the downstream patches, so we need to change all 3 (libcamera, 6.6, and 6.12) in sync. It'll be fun!
I have managed to get my FPDlink devices nominally streaming, but only by hardcoding the formats, and the images are corrupt.
I'm recompiling my 6.6 branch to ensure that there isn't a hardware issue with my setup, but I think this is that we need to resolve the differences between mainline and downstream.
Statistics: Posted by 6by9 — Mon Nov 25, 2024 4:08 pm