Ooh, good catch.

fusefs: add a intr/nointr mount option

fusefs: fix another semi-infinite loop bug regarding signal handling

fusefs: multiple interruptility improvements

style changes to getvfsbyname

Revert r346608

Build fixes

Remerge from the projects branch.

Do any of you intend to review this? I can understand if it's simply too much to review. But if so, I'd like to know. In that case I'll go ahead and commit, relying on the post-commit review that some of you have already given for my commits to the project branch.

LGTM, though ngie sometimes objects to changes to these files because of the upstream divergence. It's a shame that neither the compiler nor Coverity caught this error. However, it's worth noting that Coverity _does_ catch 43 other errors in contrib/netbsd-tests; it might be worth proactively fixing those. BTW, what "out-of-tree ABIs" generate a fault?

Abandoning in favor of D20940 , which omits the useless mergeinfo.

Yeah, I need to svn revert the useless mergeinfo. As for a single massive merge versus many tiny ones, I think it's better to do the large merge. Because even though the branch's history appears very incremental, there are also a lot of bugs in that history. I can't easily separate the sequentially added features without introducing bugs in intermediate revisions.

What's this about st_birthtime and ext2? I thought ext2 didn't support it at all.

Grr, Phabricator seems to ignore newly added files underneath a newly added directory. There's actually a lot of stuff in tests/sys/fs/fusefs

inline vn_ioc_bmap2

@piotr.konopelko_moosefs.com does MooseFS use FUSE_POLL for regular files? poll(2) already works just fine for unix-domain sockets stored on fuse file systems. The new feature is just to make it work with regular files. Currently, poll(2) will immediately return ready for regular files stored on fusefs.

@piotr.konopelko_moosefs.com done.

Update: I have finished readahead and cache invalidation. I'm close to finishing writeback cache improvements. I don't know if I'll have time for the others. posix_fallocate at least should be very easy, so I may do that one anyway.

Wrap in vfs_vnops.c at line 502. Currently, the largest possible overflow is about 2x. So using an int16_t makes it safe.

FYI, I discovered that f_seqcount needs to be int16_t rather than int8_t in order to detect an overflow. Otherwise the slightest overflow of IO_SEQMAX will wrap.