hello-world build and run on Debian 5.6.14 kernel 5.6.0-2-amd64 #219
Comments
|
@BartMassey sorry I was so slow to look at this -- I think all of the type issues you had to fix should be resolved on |
|
#227 adds So it looks like the remaining diffs are
|
|
Current master Thanks much for this project, and for adding the requisite stuff to get it working again. |
|
What's your build failure on lower clang versions? (I can probably reproduce this myself, but I'm currently tracking down an issue where Ubuntu Bionic + clang 9 + kernel 4.15 fails on Travis but not on EC2 and I have yet to figure out what the relevant difference is, so it seems safer to ask...) |
4.5 added * copy_file_range (torvalds/linux@2973293) * clone_file_range (torvalds/linux@04b38d6) * dedupe_file_range (torvalds/linux@54dbc15) 4.7 added iterate_shared (torvalds/linux@6192269), a variant of iterate that takes a non-exclusive lock. It sounds like it'd be safe to provide an implementation of iterate_shared as iterate. 4.9 removed aio_fsync (torvalds/linux@723c038), which was never actually used, so we can set it to None in older kernels and not think about it. 4.13 added wait_for_random_bytes (torvalds/linux@e297a78). For now, conditionalize the entire random module on it. All test cases other than random pass on Ubuntu 16.04's kernel. Update the README to say we expect 4.4 onwards to work, possibly with newer Clang for really new kernels (refs #219).
4.5 added * copy_file_range (torvalds/linux@2973293) * clone_file_range (torvalds/linux@04b38d6) * dedupe_file_range (torvalds/linux@54dbc15) 4.7 added iterate_shared (torvalds/linux@6192269), a variant of iterate that takes a non-exclusive lock. It sounds like it'd be safe to provide an implementation of iterate_shared as iterate. 4.9 removed aio_fsync (torvalds/linux@723c038), which was never actually used, so we can set it to None in older kernels and not think about it. 4.13 added wait_for_random_bytes (torvalds/linux@e297a78). For now, conditionalize the entire random module on it. All test cases other than random pass on Ubuntu 16.04's kernel. Update the README to say we expect 4.4 onwards to work, possibly with newer Clang for really new kernels (refs #219).
|
Are you getting something like ? If so, that's probably https://reviews.llvm.org/D75563 yeah. |
|
Yes, that's the one. |
|
OK, I updated the README to mention it. Though.. we could work around it by adding a |
|
I guess it depends on how close we are to the clang-11 release, which I'm not really tracking. I suspect pretty close, since Debian is packaging it in unstable at this point. |
This branch in my repo fork makes the
hello-worldexample build and work on on Debian-5.6.14-1 with Linux kernel 5.6.0-2-amd64. I did not put in a pull request, because this was sort of kludged together. Thought someone might want it, though. I built usingclang-11:clang-9did not work for me, but I don't know whetherclang-10would have.The text was updated successfully, but these errors were encountered: