On Wednesday 10 December 2014 at 13:52:08, Edward Shishkin wrote: > > On 12/10/2014 01:27 PM, Ivan Shapovalov wrote: > > On Thursday 23 October 2014 at 11:18:02, Ivan Shapovalov wrote: > >> These are mostly equivalent transformations (aside from the first patch) > >> I've done when I was experimenting with rwsems at zero-th iteration of > >> batch discard support. > >> > >> v2: - disable grab in current context unconditionally, not only if !BA_FORCE > >> > >> Ivan Shapovalov (3): > >> reiser4: block_alloc: improve error handling in reiser4_grab_reserved(). > >> reiser4: block_alloc: sanitize grab_enabled modifications. > >> reiser4: do not mess with grab_enabled; instead, use BA_FORCE. > >> > >> fs/reiser4/block_alloc.c | 57 +++++++++++++++++--------------- > >> fs/reiser4/plugin/file/file.c | 12 +++---- > >> fs/reiser4/plugin/file/file_conversion.c | 3 +- > >> fs/reiser4/plugin/file/tail_conversion.c | 11 +++--- > >> fs/reiser4/plugin/item/extent_file_ops.c | 3 +- > >> fs/reiser4/plugin/item/tail.c | 3 +- > >> fs/reiser4/safe_link.c | 4 +-- > >> 7 files changed, 44 insertions(+), 49 deletions(-) > > ping, what's the status? > > > This is in "upstream" already. Hm.. I didn't see it.. Well, then a follow-up fix should be applied, as described in one of previous messages, or this should be reverted. > > /* > > I'll send another round of discard patches shortly... That is, once I figure > > out how to explain in comments the sheer amount of happily coinciding subtle > > details, whose coincidence makes everything actually work. :/ > > > Note, that basic discard stuff is also in "upstream". > So, please, make sure that your patches are "re-based" Yes, I'll ensure that everything can be applied cleanly. > > BTW, could you please check that I've rebased your "precise discard" patch > > correctly? > > https://github.com/intelfx/linux/commit/47f27446d5ae7b796a842735811c48cc07615dd6 > > */ > > > > > I suggest to start with implementing the bitmap primitives that > I talked about. Just to avoid extra-work.. If I understand you correctly, then I've already done that. But let's first deal with this (current) patchset; it's broken. Sorry for creating more problems than solving. Thanks, -- Ivan Shapovalov / intelfx /
Attachment:
signature.asc
Description: This is a digitally signed message part.
