On Monday 28 July 2014 at 13:46:01, Edward Shishkin wrote: > Thanks, looks OK (I haven't tested it yet though..) > > I suggest to release a version without garbage collection for now > (just pass the sorted and merged extents to blkdev_issue_discard). Yeah, seems reasonable. I'll submit v7 of "initial implementation" shortly with simplified discard_extent(), then we could apply two patchsets and that'll be it. ---- Regarding the "ideal" solution ---- It seems that reiser4_alloc_blocks() can't be told to "allocate this exact count of blocks or fail". It allocates any number of blocks from 1 to needed, and actual count is returned in *len. Is it better to add an "exact" flag to reiser4_blocknr_hint or to add an "allocate" flag to reiser4_check_blocks()? -- Ivan Shapovalov / intelfx / > > Thanks, > Edward. > > On 07/21/2014 08:19 PM, Ivan Shapovalov wrote: > > This applies on top of PATCHv6 of "initial implementation". > > The first patch has been already sent some time ago, but I've included it > > here for sake of clarity. > > > > v2: fix disk space leak in error path, fix comments and commit messages, etc. > > > > Ivan Shapovalov (3): > > reiser4: fix reiser4_post_{commit,write_back}_hook() and their invocations. > > reiser4: discard support: use reiser4_post_write_back_hook() for discarding and completing deferred deallocations. > > reiser4: discard support: perform discard before all deallocations. > > >
Attachment:
signature.asc
Description: This is a digitally signed message part.