hi everyone, i am puneeth aditya (mugiwaraluffy56 github), an undergrad student interested in contributing to erofs as part of GSoC 2026. i just sent my first patch to the list today (blobchunk: fix 48-bit format detection to use final remapped block addresses, message id [email protected]) and have been spending the past week going through the erofs-utils codebase to understand how things fit together.
i went through the roadmap at erofs.docs.kernel.org and i am particularly interested in the multi-threaded userspace decompression work. the current single-threaded decompression path is a real bottleneck on large images and i think there is good scope for a structured GSoC project around it. would love any guidance on what a good proposal would look like, and whether there are specific parts of the decompression pipeline you would want a GSoC contributor to focus on first. thanks, puneeth aditya
