Here's the first version of my patchset which intended to bring huge pages to ext4. It's not yet ready for applying or serious use, but good enough to show the approach. The basics are the same as with tmpfs[1] which is in -mm tree and ext4 built on of it. The main difference is that we need to handle read out from and write-back to backing storage. Head page links buffers for whole huge page. Dirty/writeback tracking happens on per-hugepage level. We read out whole huge page at once. It required bumping BIO_MAX_PAGES to 512 to get it work on x86-64, which is hack. I'm not sure how to handle it properly. Readahead doesn't play with huge pages well too: 128k max readahead window, assumption on page size, PageReadahead() to track hit/miss. I've got it to allocate huge pages, but it doesn't provide any readahead as such. I don't know how to do this right. Unlike tmpfs, ext4 makes use of tags in radix-tree. The approach I used for tmpfs -- 512 entries in radix-tree per-hugepages -- doesn't work well if we want to have coherent view on tags. So the first 8 patches of the patchset converts tmpfs to use multi-order entries in radix-tree. The same infrastructure used for ext4. Writeback works for simple cases, but xfstests manages to trigger BUG_ON() eventually. That's what I work on currently. My understanding of writeback process is still rather limited and any help would be appreciated. For now I try to make xfstests run smoothly on filesystem with huge=always and 4k block size. Once it will be done, I'll widen testing to 1k blocks, encryption and bigalloc. Any comments? [1] http://lkml.kernel.org/r/1465222029-45942-1-git-send-email-kirill.shutemov@xxxxxxxxxxxxxxx TODO: - stabilize writeback; - make ext4_move_extents() work with huge pages (split them?); - check if memory reclaim process is adequate for huge pages with backing storage (unnecessary split_huge_page() ?); - handle shadow entries properly; - encryption, 1k blocks, bigalloc, ... Kirill A. Shutemov (27): mm, shmem: swich huge tmpfs to multi-order radix-tree entries Revert "radix-tree: implement radix_tree_maybe_preload_order()" page-flags: relax page flag poliry for PG_error and PG_writeback mm, rmap: account file thp pages thp: allow splitting non-shmem file-backed THPs truncate: make sure invalidate_mapping_pages() can discard huge pages filemap: allocate huge page in page_cache_read(), if allowed filemap: handle huge pages in do_generic_file_read() filemap: allocate huge page in pagecache_get_page(), if allowed filemap: handle huge pages in filemap_fdatawait_range() HACK: readahead: alloc huge pages, if allowed HACK: block: bump BIO_MAX_PAGES mm: make write_cache_pages() work on huge pages thp: introduce hpage_size() and hpage_mask() fs: make block_read_full_page() be able to read huge page fs: make block_write_{begin,end}() be able to handle huge pages fs: make block_page_mkwrite() aware about huge pages truncate: make truncate_inode_pages_range() aware about huge pages ext4: make ext4_mpage_readpages() hugepage-aware ext4: make ext4_writepage() work on huge pages ext4: handle huge pages in ext4_page_mkwrite() ext4: handle huge pages in __ext4_block_zero_page_range() ext4: handle huge pages in ext4_da_write_end() ext4: relax assert in ext4_da_page_release_reservation() WIP: ext4: handle writeback with huge pages mm, fs, ext4: expand use of page_mapping() and page_to_pgoff() ext4, vfs: add huge= mount option Matthew Wilcox (6): tools: Add WARN_ON_ONCE radix tree test suite: Allow GFP_ATOMIC allocations to fail radix-tree: Add radix_tree_join radix-tree: Add radix_tree_split radix-tree: Add radix_tree_split_preload() radix-tree: Handle multiorder entries being deleted by replace_clear_tags drivers/base/node.c | 6 + fs/buffer.c | 89 ++++--- fs/ext4/ext4.h | 5 + fs/ext4/inode.c | 106 +++++--- fs/ext4/page-io.c | 11 +- fs/ext4/readpage.c | 38 ++- fs/ext4/super.c | 19 ++ fs/proc/meminfo.c | 4 + fs/proc/task_mmu.c | 5 +- include/linux/bio.h | 2 +- include/linux/buffer_head.h | 9 +- include/linux/fs.h | 5 + include/linux/huge_mm.h | 16 ++ include/linux/mm.h | 1 + include/linux/mmzone.h | 2 + include/linux/page-flags.h | 8 +- include/linux/pagemap.h | 22 +- include/linux/radix-tree.h | 10 +- lib/radix-tree.c | 357 ++++++++++++++++++-------- mm/filemap.c | 458 +++++++++++++++++++++++----------- mm/huge_memory.c | 51 +++- mm/khugepaged.c | 26 +- mm/memory.c | 4 +- mm/page-writeback.c | 19 +- mm/page_alloc.c | 5 + mm/readahead.c | 16 +- mm/rmap.c | 12 +- mm/shmem.c | 36 +-- mm/truncate.c | 106 +++++++- mm/vmstat.c | 2 + tools/include/asm/bug.h | 11 + tools/testing/radix-tree/Makefile | 2 +- tools/testing/radix-tree/linux.c | 7 +- tools/testing/radix-tree/linux/bug.h | 2 +- tools/testing/radix-tree/linux/gfp.h | 24 +- tools/testing/radix-tree/linux/slab.h | 5 - tools/testing/radix-tree/multiorder.c | 82 ++++++ tools/testing/radix-tree/test.h | 9 + 38 files changed, 1162 insertions(+), 430 deletions(-) -- 2.8.1 -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
