Hi all, Here's a couple of patches that make fallocate's ZERO_RANGE the preferred implementation of unix_zeroout, and rearrange unix_zeroout and unix_discard to try fallocate first always. fallocate for block devices has existed in the kernel for a few years now, and satisfies libext2fs' requirement that reads following a zero/discard must return zeroes. The existing fallbacks for both functions are preserved. The third patch fixes some build problems due to the MKINSTALLDIRS -> MKDIR_P transition. Questions? Comments? This series targets e2fsprogs 1.45. --D