Linus et al, Here are the ocfs2 fixes we have ready for pulling right now. I have a bit of a procedural question, though. I've been busy, so Tao was kind enough to take the fixes I already had and apply all the new fixes. He coordinated testing them and getting them ready to go upstream. That's why I have his git tree in the pull request below. I wasn't sure how the signoffs work for this. He's got signoffs on the stuff he put together, because it's coming through his tree. My naive assumption is that this is the correct way to go. If you'd like another arrangement (same branch in my tree, a merge commit somewhere to stick my SoB on, etc), let me know. Otherwise, please pull. Joel The following changes since commit 2bfc96a127bc1cc94d26bfaa40159966064f9c8c: Linus Torvalds (1): Linux 2.6.36-rc3 are available in the git repository at: git://oss.oracle.com/git/tma/linux-2.6.git fixes Jan Kara (2): ocfs2: Flush drive's caches on fdatasync ocfs2: Fix deadlock when allocating page Mark Fasheh (6): ocfs2: properly set and use inode group alloc hint ocfs2: split out inode alloc code from ocfs2_mknod_locked ocfs2: use ocfs2_alloc_dinode_update_counts() instead of open coding ocfs2: allow return of new inode block location before allocation of the inode ocfs2: split out ocfs2_prepare_orphan_dir() into locking and prep functions ocfs2: Fix orphan add in ocfs2_create_inode_in_orphan Sunil Mushran (2): ocfs2: Fix metaecc error messages ocfs2: Fix incorrect checksum validation error Tao Ma (2): ocfs2: make __ocfs2_page_mkwrite handle file end properly. ocfs2: Use the right group in nfs sync check. Tristan Ye (1): Ocfs2: Fix a regression bug from mainline commit(6b933c8e6f1a2f3118082c455eef25f9b1ac7b45). fs/ocfs2/alloc.c | 2 +- fs/ocfs2/blockcheck.c | 4 +- fs/ocfs2/file.c | 15 ++- fs/ocfs2/inode.c | 6 +- fs/ocfs2/mmap.c | 8 +- fs/ocfs2/namei.c | 302 ++++++++++++++++++++++++++++++++++++----------- fs/ocfs2/refcounttree.c | 5 +- fs/ocfs2/suballoc.c | 219 ++++++++++++++++++++++++++++++---- fs/ocfs2/suballoc.h | 21 ++++ 9 files changed, 475 insertions(+), 107 deletions(-)
On Tue, Sep 7, 2010 at 4:41 PM, Joel Becker <boel.becker at oracle.com> wrote:> > ? ? ? ?Here are the ocfs2 fixes we have ready for pulling right now. ?I > have a bit of a procedural question, though. ?I've been busy, so Tao was > kind enough to take the fixes I already had and apply all the new fixes. > He coordinated testing them and getting them ready to go upstream. > That's why I have his git tree in the pull request below. > ? ? ? ?I wasn't sure how the signoffs work for this. ?He's got signoffs > on the stuff he put together, because it's coming through his tree. ?My > naive assumption is that this is the correct way to go.The sign-off chain should basically stretch from the author to the committer - so a good sign-off chain is one where the first sign-off matches the author of the commit, and the last sign-off matches the committer, and the ones in the middle are obviously the path it took from one to the other. However, Tao has *NOT* added sign-offs to the patches he's taken. I didn't look at all of them, but the first few commits are all authored (and signed off on - so far so good) by Mark Fasheh, but there is no sign-off for Tao. So the sign-off chain is clearly not complete. Linus