Andreas pointed out that original version of the last patch in this series would break the 16k/64k blocksize regressiont tests on MacOS 10.5. When I looked into addressing this, I realized that we didn't need to request 16k or 64k alignment, and in general we often requested aligned buffers when they were not necessary, and/or requested a larger alignment factor than necessary. This patch series cleans up this mess by refactoring the code so that DIO alignment is calculated in one place, and so we have a common routine for allocating I/O buffers used by e2fsprogs's I/O layer. - Ted Theodore Ts'o (5): libext2fs: move the alignment field from unix_io to the io_manager libext2fs: refactor Direct I/O alignment requirement calculations libext2fs: make read_bitmaps() more efficient when using direct I/O libext2fs: factor out I/O buffer allocation Support systems without posix_memalign() and memalign() lib/ext2fs/Makefile.in | 9 +++++++- lib/ext2fs/ext2_io.h | 3 +++ lib/ext2fs/ext2fs.h | 1 + lib/ext2fs/getsectsize.c | 26 +++++++++++++++++++++ lib/ext2fs/inline.c | 56 ++++++++++++++++++++++++++++++++++++++++++++-- lib/ext2fs/inode.c | 4 ++-- lib/ext2fs/io_manager.c | 17 ++++++++++++++ lib/ext2fs/mmp.c | 37 ++++++++---------------------- lib/ext2fs/openfs.c | 2 +- lib/ext2fs/rw_bitmaps.c | 24 ++++++++------------ lib/ext2fs/test_io.c | 7 +++++- lib/ext2fs/unix_io.c | 42 +++++++++++++++------------------- 12 files changed, 154 insertions(+), 74 deletions(-) -- 1.7.10.rc3 -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
