https://bugzilla.suse.com/show_bug.cgi?id=1196744 https://bugzilla.suse.com/show_bug.cgi?id=1196744#c5 Dead Mozay <dead_mozay@opensuse.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |WONTFIX --- Comment #5 from Dead Mozay <dead_mozay@opensuse.org> --- (In reply to Takashi Iwai from comment #4)

Well, the bottom line is that those drivers can be built as modules. But the currently those are only built as built-in kernel, and for allowing them to be built as modules, we need to patch the kernel code. I asked to hack it by yourself and check whether it works.

But after writing this, I remembered that I chatted with Ben Hutchings some time ago about this topic. There are a couple of downstream patches for Ubuntu to build those as modules:

https://salsa.debian.org/kernel-team/linux/blob/master/debian/patches/debian... android-enable-building-ashmem-and-binder-as-modules.patch

https://salsa.debian.org/kernel-team/linux/blob/master/debian/patches/debian... export-symbols-needed-by-android-drivers.patch

However, the problem is that the latter one (to add exported symbols to core code) would be VERY unlikely accepted by the upstream. The first attempt failed miserably:

https://lore.kernel.org/linux-fsdevel/20180730143710.14413-1- christian@brauner.io/T/#m32f56e7f2f3dbb8105009e2fb62ffc71243e89bd

That is, the binder stuff can almost never be a module, hence we can't go for it for our standard kernel.

OTOH, the ashmem seems to have a better chance, as it has little dependencies on other core parts, but I have no idea whether ashmem alone would make any sense.

Now it���s clear, then there���s no point in ashmem either. Thank you for the clarification. I think the bug can be closed. -- You are receiving this mail because: You are the assignee for the bug.