#include <linux/time.h>
#include <linux/init.h>
#include <linux/string.h>
-#include <linux/smp_lock.h>
#include <linux/backing-dev.h>
#include <linux/mpage.h>
#include <linux/swap.h>
#include <linux/writeback.h>
#include <linux/bit_spinlock.h>
-#include <linux/version.h>
+#include <linux/slab.h>
+#include "compat.h"
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
#include "btrfs_inode.h"
#include "volumes.h"
#include "ordered-data.h"
-#include "compat.h"
#include "compression.h"
#include "extent_io.h"
#include "extent_map.h"
/* IO errors */
int errors;
+ int mirror_num;
/* for reads, this is the bio we are copying the data into */
struct bio *orig_bio;
+
+ /*
+ * the start of a variable length array of checksums only
+ * used by reads
+ */
+ u32 sums;
};
+static inline int compressed_bio_size(struct btrfs_root *root,
+ unsigned long disk_size)
+{
+ u16 csum_size = btrfs_super_csum_size(&root->fs_info->super_copy);
+ return sizeof(struct compressed_bio) +
+ ((disk_size + root->sectorsize - 1) / root->sectorsize) *
+ csum_size;
+}
+
static struct bio *compressed_bio_alloc(struct block_device *bdev,
u64 first_byte, gfp_t gfp_flags)
{
return bio;
}
+static int check_compressed_csum(struct inode *inode,
+ struct compressed_bio *cb,
+ u64 disk_start)
+{
+ int ret;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct page *page;
+ unsigned long i;
+ char *kaddr;
+ u32 csum;
+ u32 *cb_sum = &cb->sums;
+
+ if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)
+ return 0;
+
+ for (i = 0; i < cb->nr_pages; i++) {
+ page = cb->compressed_pages[i];
+ csum = ~(u32)0;
+
+ kaddr = kmap_atomic(page, KM_USER0);
+ csum = btrfs_csum_data(root, kaddr, csum, PAGE_CACHE_SIZE);
+ btrfs_csum_final(csum, (char *)&csum);
+ kunmap_atomic(kaddr, KM_USER0);
+
+ if (csum != *cb_sum) {
+ printk(KERN_INFO "btrfs csum failed ino %lu "
+ "extent %llu csum %u "
+ "wanted %u mirror %d\n", inode->i_ino,
+ (unsigned long long)disk_start,
+ csum, *cb_sum, cb->mirror_num);
+ ret = -EIO;
+ goto fail;
+ }
+ cb_sum++;
+
+ }
+ ret = 0;
+fail:
+ return ret;
+}
+
/* when we finish reading compressed pages from the disk, we
* decompress them and then run the bio end_io routines on the
* decompressed pages (in the inode address space).
if (!atomic_dec_and_test(&cb->pending_bios))
goto out;
+ inode = cb->inode;
+ ret = check_compressed_csum(inode, cb, (u64)bio->bi_sector << 9);
+ if (ret)
+ goto csum_failed;
+
/* ok, we're the last bio for this extent, lets start
* the decompression.
*/
- inode = cb->inode;
tree = &BTRFS_I(inode)->io_tree;
ret = btrfs_zlib_decompress_biovec(cb->compressed_pages,
cb->start,
cb->orig_bio->bi_io_vec,
cb->orig_bio->bi_vcnt,
cb->compressed_len);
+csum_failed:
if (ret)
cb->errors = 1;
}
/* do io completion on the original bio */
- if (cb->errors)
+ if (cb->errors) {
bio_io_error(cb->orig_bio);
- else
+ } else {
+ int bio_index = 0;
+ struct bio_vec *bvec = cb->orig_bio->bi_io_vec;
+
+ /*
+ * we have verified the checksum already, set page
+ * checked so the end_io handlers know about it
+ */
+ while (bio_index < cb->orig_bio->bi_vcnt) {
+ SetPageChecked(bvec->bv_page);
+ bvec++;
+ bio_index++;
+ }
bio_endio(cb->orig_bio, 0);
+ }
/* finally free the cb struct */
kfree(cb->compressed_pages);
int i;
int ret;
- while(nr_pages > 0) {
+ while (nr_pages > 0) {
ret = find_get_pages_contig(inode->i_mapping, index,
- min(nr_pages, ARRAY_SIZE(pages)), pages);
+ min_t(unsigned long,
+ nr_pages, ARRAY_SIZE(pages)), pages);
if (ret == 0) {
nr_pages -= 1;
index += 1;
int ret;
WARN_ON(start & ((u64)PAGE_CACHE_SIZE - 1));
- cb = kmalloc(sizeof(*cb), GFP_NOFS);
+ cb = kmalloc(compressed_bio_size(root, compressed_len), GFP_NOFS);
atomic_set(&cb->pending_bios, 0);
cb->errors = 0;
cb->inode = inode;
cb->start = start;
cb->len = len;
+ cb->mirror_num = 0;
cb->compressed_pages = compressed_pages;
cb->compressed_len = compressed_len;
cb->orig_bio = NULL;
bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
- ret = btrfs_csum_file_bytes(root, inode, start, len);
- BUG_ON(ret);
-
bio = compressed_bio_alloc(bdev, first_byte, GFP_NOFS);
bio->bi_private = cb;
bio->bi_end_io = end_compressed_bio_write;
PAGE_CACHE_SIZE) {
bio_get(bio);
+ /*
+ * inc the count before we submit the bio so
+ * we know the end IO handler won't happen before
+ * we inc the count. Otherwise, the cb might get
+ * freed before we're done setting it up
+ */
+ atomic_inc(&cb->pending_bios);
ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
BUG_ON(ret);
+ ret = btrfs_csum_one_bio(root, inode, bio, start, 1);
+ BUG_ON(ret);
+
ret = btrfs_map_bio(root, WRITE, bio, 0, 1);
BUG_ON(ret);
bio_put(bio);
bio = compressed_bio_alloc(bdev, first_byte, GFP_NOFS);
- atomic_inc(&cb->pending_bios);
bio->bi_private = cb;
bio->bi_end_io = end_compressed_bio_write;
bio_add_page(bio, page, PAGE_CACHE_SIZE, 0);
}
bytes_left -= PAGE_CACHE_SIZE;
first_byte += PAGE_CACHE_SIZE;
+ cond_resched();
}
bio_get(bio);
ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
BUG_ON(ret);
+ ret = btrfs_csum_one_bio(root, inode, bio, start, 1);
+ BUG_ON(ret);
+
ret = btrfs_map_bio(root, WRITE, bio, 0, 1);
BUG_ON(ret);
return 0;
}
+static noinline int add_ra_bio_pages(struct inode *inode,
+ u64 compressed_end,
+ struct compressed_bio *cb)
+{
+ unsigned long end_index;
+ unsigned long page_index;
+ u64 last_offset;
+ u64 isize = i_size_read(inode);
+ int ret;
+ struct page *page;
+ unsigned long nr_pages = 0;
+ struct extent_map *em;
+ struct address_space *mapping = inode->i_mapping;
+ struct extent_map_tree *em_tree;
+ struct extent_io_tree *tree;
+ u64 end;
+ int misses = 0;
+
+ page = cb->orig_bio->bi_io_vec[cb->orig_bio->bi_vcnt - 1].bv_page;
+ last_offset = (page_offset(page) + PAGE_CACHE_SIZE);
+ em_tree = &BTRFS_I(inode)->extent_tree;
+ tree = &BTRFS_I(inode)->io_tree;
+
+ if (isize == 0)
+ return 0;
+
+ end_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT;
+
+ while (last_offset < compressed_end) {
+ page_index = last_offset >> PAGE_CACHE_SHIFT;
+
+ if (page_index > end_index)
+ break;
+
+ rcu_read_lock();
+ page = radix_tree_lookup(&mapping->page_tree, page_index);
+ rcu_read_unlock();
+ if (page) {
+ misses++;
+ if (misses > 4)
+ break;
+ goto next;
+ }
+
+ page = __page_cache_alloc(mapping_gfp_mask(mapping) &
+ ~__GFP_FS);
+ if (!page)
+ break;
+
+ if (add_to_page_cache_lru(page, mapping, page_index,
+ GFP_NOFS)) {
+ page_cache_release(page);
+ goto next;
+ }
+
+ end = last_offset + PAGE_CACHE_SIZE - 1;
+ /*
+ * at this point, we have a locked page in the page cache
+ * for these bytes in the file. But, we have to make
+ * sure they map to this compressed extent on disk.
+ */
+ set_page_extent_mapped(page);
+ lock_extent(tree, last_offset, end, GFP_NOFS);
+ read_lock(&em_tree->lock);
+ em = lookup_extent_mapping(em_tree, last_offset,
+ PAGE_CACHE_SIZE);
+ read_unlock(&em_tree->lock);
+
+ if (!em || last_offset < em->start ||
+ (last_offset + PAGE_CACHE_SIZE > extent_map_end(em)) ||
+ (em->block_start >> 9) != cb->orig_bio->bi_sector) {
+ free_extent_map(em);
+ unlock_extent(tree, last_offset, end, GFP_NOFS);
+ unlock_page(page);
+ page_cache_release(page);
+ break;
+ }
+ free_extent_map(em);
+
+ if (page->index == end_index) {
+ char *userpage;
+ size_t zero_offset = isize & (PAGE_CACHE_SIZE - 1);
+
+ if (zero_offset) {
+ int zeros;
+ zeros = PAGE_CACHE_SIZE - zero_offset;
+ userpage = kmap_atomic(page, KM_USER0);
+ memset(userpage + zero_offset, 0, zeros);
+ flush_dcache_page(page);
+ kunmap_atomic(userpage, KM_USER0);
+ }
+ }
+
+ ret = bio_add_page(cb->orig_bio, page,
+ PAGE_CACHE_SIZE, 0);
+
+ if (ret == PAGE_CACHE_SIZE) {
+ nr_pages++;
+ page_cache_release(page);
+ } else {
+ unlock_extent(tree, last_offset, end, GFP_NOFS);
+ unlock_page(page);
+ page_cache_release(page);
+ break;
+ }
+next:
+ last_offset += PAGE_CACHE_SIZE;
+ }
+ return 0;
+}
+
/*
* for a compressed read, the bio we get passed has all the inode pages
* in it. We don't actually do IO on those pages but allocate new ones
struct block_device *bdev;
struct bio *comp_bio;
u64 cur_disk_byte = (u64)bio->bi_sector << 9;
+ u64 em_len;
+ u64 em_start;
struct extent_map *em;
int ret;
+ u32 *sums;
tree = &BTRFS_I(inode)->io_tree;
em_tree = &BTRFS_I(inode)->extent_tree;
/* we need the actual starting offset of this extent in the file */
- spin_lock(&em_tree->lock);
+ read_lock(&em_tree->lock);
em = lookup_extent_mapping(em_tree,
page_offset(bio->bi_io_vec->bv_page),
PAGE_CACHE_SIZE);
- spin_unlock(&em_tree->lock);
+ read_unlock(&em_tree->lock);
- cb = kmalloc(sizeof(*cb), GFP_NOFS);
+ compressed_len = em->block_len;
+ cb = kmalloc(compressed_bio_size(root, compressed_len), GFP_NOFS);
atomic_set(&cb->pending_bios, 0);
cb->errors = 0;
cb->inode = inode;
+ cb->mirror_num = mirror_num;
+ sums = &cb->sums;
+
+ cb->start = em->orig_start;
+ em_len = em->len;
+ em_start = em->start;
- cb->start = em->start;
- compressed_len = em->block_len;
free_extent_map(em);
+ em = NULL;
cb->len = uncompressed_len;
cb->compressed_len = compressed_len;
}
cb->nr_pages = nr_pages;
+ add_ra_bio_pages(inode, em_start + em_len, cb);
+
+ /* include any pages we added in add_ra-bio_pages */
+ uncompressed_len = bio->bi_vcnt * PAGE_CACHE_SIZE;
+ cb->len = uncompressed_len;
+
comp_bio = compressed_bio_alloc(bdev, cur_disk_byte, GFP_NOFS);
comp_bio->bi_private = cb;
comp_bio->bi_end_io = end_compressed_bio_read;
for (page_index = 0; page_index < nr_pages; page_index++) {
page = cb->compressed_pages[page_index];
page->mapping = inode->i_mapping;
+ page->index = em_start >> PAGE_CACHE_SHIFT;
+
if (comp_bio->bi_size)
ret = tree->ops->merge_bio_hook(page, 0,
PAGE_CACHE_SIZE,
ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0);
BUG_ON(ret);
- ret = btrfs_map_bio(root, READ, comp_bio, 0, 0);
+ /*
+ * inc the count before we submit the bio so
+ * we know the end IO handler won't happen before
+ * we inc the count. Otherwise, the cb might get
+ * freed before we're done setting it up
+ */
+ atomic_inc(&cb->pending_bios);
+
+ if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) {
+ btrfs_lookup_bio_sums(root, inode, comp_bio,
+ sums);
+ }
+ sums += (comp_bio->bi_size + root->sectorsize - 1) /
+ root->sectorsize;
+
+ ret = btrfs_map_bio(root, READ, comp_bio,
+ mirror_num, 0);
BUG_ON(ret);
bio_put(comp_bio);
comp_bio = compressed_bio_alloc(bdev, cur_disk_byte,
GFP_NOFS);
- atomic_inc(&cb->pending_bios);
- bio->bi_private = cb;
- bio->bi_end_io = end_compressed_bio_write;
- bio_add_page(bio, page, PAGE_CACHE_SIZE, 0);
+ comp_bio->bi_private = cb;
+ comp_bio->bi_end_io = end_compressed_bio_read;
+
+ bio_add_page(comp_bio, page, PAGE_CACHE_SIZE, 0);
}
cur_disk_byte += PAGE_CACHE_SIZE;
}
ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0);
BUG_ON(ret);
- ret = btrfs_map_bio(root, READ, comp_bio, 0, 0);
+ if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM))
+ btrfs_lookup_bio_sums(root, inode, comp_bio, sums);
+
+ ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0);
BUG_ON(ret);
bio_put(comp_bio);