From: Theodore Ts'o Subject: [PATCH-v2 01/20] ext4 crypto: add ext4_mpage_readpages() Date: Sun, 12 Apr 2015 23:16:17 -0400 Message-ID: <1428894996-7852-2-git-send-email-tytso@mit.edu> References: <1428894996-7852-1-git-send-email-tytso@mit.edu> Cc: mhalcrow@google.com, Theodore Ts'o To: Ext4 Developers List Return-path: Received: from imap.thunk.org ([74.207.234.97]:35616 "EHLO imap.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752714AbbDMDRN (ORCPT ); Sun, 12 Apr 2015 23:17:13 -0400 In-Reply-To: <1428894996-7852-1-git-send-email-tytso@mit.edu> Sender: linux-ext4-owner@vger.kernel.org List-ID: This takes code from fs/mpage.c and optimizes it for ext4. Its primary reason is to allow us to more easily add encryption to ext4's read path in an efficient manner. Change-Id: I0a115e90ab13861ab3e96641a0eb82b951198ecc Signed-off-by: Theodore Ts'o --- fs/ext4/Makefile | 2 +- fs/ext4/ext4.h | 4 + fs/ext4/inode.c | 4 +- fs/ext4/readpage.c | 264 +++++++++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 271 insertions(+), 3 deletions(-) create mode 100644 fs/ext4/readpage.c diff --git a/fs/ext4/Makefile b/fs/ext4/Makefile index 0310fec..cd6f50f 100644 --- a/fs/ext4/Makefile +++ b/fs/ext4/Makefile @@ -8,7 +8,7 @@ ext4-y := balloc.o bitmap.o dir.o file.o fsync.o ialloc.o inode.o page-io.o \ ioctl.o namei.o super.o symlink.o hash.o resize.o extents.o \ ext4_jbd2.o migrate.o mballoc.o block_validity.o move_extent.o \ mmp.o indirect.o extents_status.o xattr.o xattr_user.o \ - xattr_trusted.o inline.o + xattr_trusted.o inline.o readpage.o ext4-$(CONFIG_EXT4_FS_POSIX_ACL) += acl.o ext4-$(CONFIG_EXT4_FS_SECURITY) += xattr_security.o diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index a75fba6..06e8add 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -2683,6 +2683,10 @@ static inline void ext4_set_de_type(struct super_block *sb, de->file_type = ext4_type_by_mode[(mode & S_IFMT)>>S_SHIFT]; } +/* readpages.c */ +extern int ext4_mpage_readpages(struct address_space *mapping, + struct list_head *pages, struct page *page, + unsigned nr_pages); /* symlink.c */ extern const struct inode_operations ext4_symlink_inode_operations; diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 5653fa4..a68cacc 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2798,7 +2798,7 @@ static int ext4_readpage(struct file *file, struct page *page) ret = ext4_readpage_inline(inode, page); if (ret == -EAGAIN) - return mpage_readpage(page, ext4_get_block); + return ext4_mpage_readpages(page->mapping, NULL, page, 1); return ret; } @@ -2813,7 +2813,7 @@ ext4_readpages(struct file *file, struct address_space *mapping, if (ext4_has_inline_data(inode)) return 0; - return mpage_readpages(mapping, pages, nr_pages, ext4_get_block); + return ext4_mpage_readpages(mapping, pages, NULL, nr_pages); } static void ext4_invalidatepage(struct page *page, unsigned int offset, diff --git a/fs/ext4/readpage.c b/fs/ext4/readpage.c new file mode 100644 index 0000000..fff9fe6 --- /dev/null +++ b/fs/ext4/readpage.c @@ -0,0 +1,264 @@ +/* + * linux/fs/ext4/readpage.c + * + * Copyright (C) 2002, Linus Torvalds. + * Copyright (C) 2015, Google, Inc. + * + * This was originally taken from fs/mpage.c + * + * The intent is the ext4_mpage_readpages() function here is intended + * to replace mpage_readpages() in the general case, not just for + * encrypted files. It has some limitations (see below), where it + * will fall back to read_block_full_page(), but these limitations + * should only be hit when page_size != block_size. + * + * This will allow us to attach a callback function to support ext4 + * encryption. + * + * If anything unusual happens, such as: + * + * - encountering a page which has buffers + * - encountering a page which has a non-hole after a hole + * - encountering a page with non-contiguous blocks + * + * then this code just gives up and calls the buffer_head-based read function. + * It does handle a page which has holes at the end - that is a common case: + * the end-of-file on blocksize < PAGE_CACHE_SIZE setups. + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "ext4.h" + +/* + * I/O completion handler for multipage BIOs. + * + * The mpage code never puts partial pages into a BIO (except for end-of-file). + * If a page does not map to a contiguous run of blocks then it simply falls + * back to block_read_full_page(). + * + * Why is this? If a page's completion depends on a number of different BIOs + * which can complete in any order (or at the same time) then determining the + * status of that page is hard. See end_buffer_async_read() for the details. + * There is no point in duplicating all that complexity. + */ +static void mpage_end_io(struct bio *bio, int err) +{ + struct bio_vec *bv; + int i; + + bio_for_each_segment_all(bv, bio, i) { + struct page *page = bv->bv_page; + + if (!err) { + SetPageUptodate(page); + } else { + ClearPageUptodate(page); + SetPageError(page); + } + unlock_page(page); + } + + bio_put(bio); +} + +int ext4_mpage_readpages(struct address_space *mapping, + struct list_head *pages, struct page *page, + unsigned nr_pages) +{ + struct bio *bio = NULL; + unsigned page_idx; + sector_t last_block_in_bio = 0; + + struct inode *inode = mapping->host; + const unsigned blkbits = inode->i_blkbits; + const unsigned blocks_per_page = PAGE_CACHE_SIZE >> blkbits; + const unsigned blocksize = 1 << blkbits; + sector_t block_in_file; + sector_t last_block; + sector_t last_block_in_file; + sector_t blocks[MAX_BUF_PER_PAGE]; + unsigned page_block; + struct block_device *bdev = inode->i_sb->s_bdev; + int length; + unsigned relative_block = 0; + struct ext4_map_blocks map; + + map.m_pblk = 0; + map.m_lblk = 0; + map.m_len = 0; + map.m_flags = 0; + + for (page_idx = 0; nr_pages; page_idx++, nr_pages--) { + int fully_mapped = 1; + unsigned first_hole = blocks_per_page; + + prefetchw(&page->flags); + if (pages) { + page = list_entry(pages->prev, struct page, lru); + list_del(&page->lru); + if (add_to_page_cache_lru(page, mapping, + page->index, GFP_KERNEL)) + goto next_page; + } + + if (page_has_buffers(page)) + goto confused; + + block_in_file = (sector_t)page->index << (PAGE_CACHE_SHIFT - blkbits); + last_block = block_in_file + nr_pages * blocks_per_page; + last_block_in_file = (i_size_read(inode) + blocksize - 1) >> blkbits; + if (last_block > last_block_in_file) + last_block = last_block_in_file; + page_block = 0; + + /* + * Map blocks using the previous result first. + */ + if ((map.m_flags & EXT4_MAP_MAPPED) && + block_in_file > map.m_lblk && + block_in_file < (map.m_lblk + map.m_len)) { + unsigned map_offset = block_in_file - map.m_lblk; + unsigned last = map.m_len - map_offset; + + for (relative_block = 0; ; relative_block++) { + if (relative_block == last) { + /* needed? */ + map.m_flags &= ~EXT4_MAP_MAPPED; + break; + } + if (page_block == blocks_per_page) + break; + blocks[page_block] = map.m_pblk + map_offset + + relative_block; + page_block++; + block_in_file++; + } + } + + /* + * Then do more ext4_map_blocks() calls until we are + * done with this page. + */ + while (page_block < blocks_per_page) { + if (block_in_file < last_block) { + map.m_lblk = block_in_file; + map.m_len = last_block - block_in_file; + + if (ext4_map_blocks(NULL, inode, &map, 0) < 0) { + set_error_page: + SetPageError(page); + zero_user_segment(page, 0, + PAGE_CACHE_SIZE); + unlock_page(page); + goto next_page; + } + } + if ((map.m_flags & EXT4_MAP_MAPPED) == 0) { + fully_mapped = 0; + if (first_hole == blocks_per_page) + first_hole = page_block; + page_block++; + block_in_file++; + continue; + } + if (first_hole != blocks_per_page) + goto confused; /* hole -> non-hole */ + + /* Contiguous blocks? */ + if (page_block && blocks[page_block-1] != map.m_pblk-1) + goto confused; + for (relative_block = 0; ; relative_block++) { + if (relative_block == map.m_len) { + /* needed? */ + map.m_flags &= ~EXT4_MAP_MAPPED; + break; + } else if (page_block == blocks_per_page) + break; + blocks[page_block] = map.m_pblk+relative_block; + page_block++; + block_in_file++; + } + } + if (first_hole != blocks_per_page) { + zero_user_segment(page, first_hole << blkbits, + PAGE_CACHE_SIZE); + if (first_hole == 0) { + SetPageUptodate(page); + unlock_page(page); + goto next_page; + } + } else if (fully_mapped) { + SetPageMappedToDisk(page); + } + if (fully_mapped && blocks_per_page == 1 && + !PageUptodate(page) && cleancache_get_page(page) == 0) { + SetPageUptodate(page); + goto confused; + } + + /* + * This page will go to BIO. Do we need to send this + * BIO off first? + */ + if (bio && (last_block_in_bio != blocks[0] - 1)) { + submit_and_realloc: + submit_bio(READ, bio); + bio = NULL; + } + if (bio == NULL) { + bio = bio_alloc(GFP_KERNEL, + min_t(int, nr_pages, bio_get_nr_vecs(bdev))); + if (!bio) + goto set_error_page; + bio->bi_bdev = bdev; + bio->bi_iter.bi_sector = blocks[0] << (blkbits - 9); + bio->bi_end_io = mpage_end_io; + } + + length = first_hole << blkbits; + if (bio_add_page(bio, page, length, 0) < length) + goto submit_and_realloc; + + if (((map.m_flags & EXT4_MAP_BOUNDARY) && + (relative_block == map.m_len)) || + (first_hole != blocks_per_page)) { + submit_bio(READ, bio); + bio = NULL; + } else + last_block_in_bio = blocks[blocks_per_page - 1]; + goto next_page; + confused: + if (bio) { + submit_bio(READ, bio); + bio = NULL; + } + if (!PageUptodate(page)) + block_read_full_page(page, ext4_get_block); + else + unlock_page(page); + next_page: + if (pages) + page_cache_release(page); + } + BUG_ON(pages && !list_empty(pages)); + if (bio) + submit_bio(READ, bio); + return 0; +} -- 2.3.0