Re: [GIT PULL] Core block IO changes for 3.14

From: Stephen Rothwell
Date: Tue Jan 28 2014 - 19:05:53 EST

Hi all,

On Tue, 28 Jan 2014 15:52:29 -0700 Jens Axboe <axboe@xxxxxxxxx> wrote:
>
> On Tue, Jan 28 2014, Jens Axboe wrote:
> >
> > This is the pull request for the core block IO changes for 3.14. The
> > major piece in here is the immutable bio_ve series from Kent, the rest
> > is fairly minor. It was supposed to go in last round, but various issues
> > pushed it to this release instead. The pull request contains:
> >
> >
> > - Various smaller blk-mq fixes from different folks. Nothing major here,
> > just minor fixes and cleanups.
> >
> > - Fix for a memory leak in the error path in the block ioctl code from
> > Christian Engelmayer.
> >
> > - Header export fix from CaiZhiyong.
> >
> > - Finally the immutable biovec changes from Kent Overstreet. This
> > enables some nice future work on making arbitrarily sized bios
> > possible, and splitting more efficient. Related fixes to immutable
> > bio_vecs:
> >
> > - dm-cache immutable fixup from Mike Snitzer.
> > - btrfs immutable fixup from Muthu Kumar.
> >
> > - bio-integrity fix from Nic Bellinger, which is also going to stable.
> >
> >
> > Please pull! There will be a bit of merge work for you, but it should be
> > fairly straight forward. It's mostly related to changin:
> >
> > bio->bi_sector -> bio->bi_iter.bi_sector
> > bio->bi_size -> bio->bi_iter.bi_size
> >
> >
> > git://git.kernel.dk/linux-block.git for-3.14/core
>
> BTW, let me know if you want me to merge this. The above has been in
> for-next since forever, and Stephen has carried a fix or two for new
> merges.

The worst bit is the conflicts with the f2fs changes that have already
been merged. My current merge commit looks like this (though I don't
remember getting any comments on my fixes):

da3f6c793c656a022453df8bf458d13e5a353beb
diff --cc drivers/md/dm-thin.c
index 726228b33a01,357eb272dbd9..faaf944597ab
--- a/drivers/md/dm-thin.c
+++ b/drivers/md/dm-thin.c
@@@ -1258,8 -1262,8 +1264,8 @@@ static void process_bio_read_only(struc
r = dm_thin_find_block(tc->td, block, 1, &lookup_result);
switch (r) {
case 0:
- if (lookup_result.shared && (rw == WRITE) && bio->bi_size)
+ if (lookup_result.shared && (rw == WRITE) && bio->bi_iter.bi_size)
- bio_io_error(bio);
+ handle_unserviceable_bio(tc->pool, bio);
else {
inc_all_io_entry(tc->pool, bio);
remap_and_issue(tc, bio, lookup_result.block);
diff --cc drivers/md/raid10.c
index 8d39d63281b9,6d43d88657aa..33fc408e5eac
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@@ -1319,8 -1256,8 +1256,8 @@@ read_again
/* Could not read all from this device, so we will
* need another r10_bio.
*/
- sectors_handled = (r10_bio->sectors + max_sectors
+ sectors_handled = (r10_bio->sector + max_sectors
- - bio->bi_sector);
+ - bio->bi_iter.bi_sector);
r10_bio->sectors = max_sectors;
spin_lock_irq(&conf->device_lock);
if (bio->bi_phys_segments == 0)
diff --cc fs/btrfs/extent_io.c
index fbe501d3bd01,bcb6f1b780d6..85bbd01f1271
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@@ -2375,12 -2332,15 +2375,13 @@@ int end_extent_writepage(struct page *p
*/
static void end_bio_extent_writepage(struct bio *bio, int err)
{
- struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+ struct bio_vec *bvec;
- struct extent_io_tree *tree;
u64 start;
u64 end;
+ int i;

- do {
+ bio_for_each_segment_all(bvec, bio, i) {
struct page *page = bvec->bv_page;
- tree = &BTRFS_I(page->mapping->host)->io_tree;

/* We always issue full-page reads, but if some block
* in a page fails to read, blk_update_request() will
diff --cc fs/btrfs/inode.c
index 1ef056837755,7ab0e94ad492..f0422a5efa78
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@@ -7016,10 -6891,11 +7013,11 @@@ static void btrfs_end_dio_bio(struct bi
struct btrfs_dio_private *dip = bio->bi_private;

if (err) {
- printk(KERN_ERR "btrfs direct IO failed ino %llu rw %lu "
- "sector %#Lx len %u err no %d\n",
+ btrfs_err(BTRFS_I(dip->inode)->root->fs_info,
+ "direct IO failed ino %llu rw %lu sector %#Lx len %u err no %d",
btrfs_ino(dip->inode), bio->bi_rw,
- (unsigned long long)bio->bi_sector, bio->bi_size, err);
+ (unsigned long long)bio->bi_iter.bi_sector,
+ bio->bi_iter.bi_size, err);
dip->errors = 1;

/*
diff --cc fs/f2fs/data.c
index 0ae558723506,a2c8de8ba6ce..25d675e6a138
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@@ -24,195 -24,6 +24,192 @@@
#include "segment.h"
#include <trace/events/f2fs.h>

+static void f2fs_read_end_io(struct bio *bio, int err)
+{
- const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
- struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
++ struct bio_vec *bvec;
++ int i;
+
- do {
++ bio_for_each_segment_all(bvec, bio, i) {
+ struct page *page = bvec->bv_page;
+
- if (--bvec >= bio->bi_io_vec)
- prefetchw(&bvec->bv_page->flags);
-
- if (unlikely(!uptodate)) {
++ if (unlikely(err)) {
+ ClearPageUptodate(page);
+ SetPageError(page);
+ } else {
+ SetPageUptodate(page);
+ }
+ unlock_page(page);
- } while (bvec >= bio->bi_io_vec);
++ }
+
+ bio_put(bio);
+}
+
+static void f2fs_write_end_io(struct bio *bio, int err)
+{
- const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
- struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
- struct f2fs_sb_info *sbi = F2FS_SB(bvec->bv_page->mapping->host->i_sb);
++ struct bio_vec *bvec;
++ struct f2fs_sb_info *sbi = NULL;
++ int i;
+
- do {
++ bio_for_each_segment_all(bvec, bio, i) {
+ struct page *page = bvec->bv_page;
+
- if (--bvec >= bio->bi_io_vec)
- prefetchw(&bvec->bv_page->flags);
++ if (!sbi)
++ sbi = F2FS_SB(bvec->bv_page->mapping->host->i_sb);
+
- if (unlikely(!uptodate)) {
++ if (unlikely(err)) {
+ SetPageError(page);
+ set_bit(AS_EIO, &page->mapping->flags);
+ set_ckpt_flags(sbi->ckpt, CP_ERROR_FLAG);
+ sbi->sb->s_flags |= MS_RDONLY;
+ }
+ end_page_writeback(page);
+ dec_page_count(sbi, F2FS_WRITEBACK);
- } while (bvec >= bio->bi_io_vec);
++ }
+
+ if (bio->bi_private)
+ complete(bio->bi_private);
+
+ if (!get_pages(sbi, F2FS_WRITEBACK) &&
+ !list_empty(&sbi->cp_wait.task_list))
+ wake_up(&sbi->cp_wait);
+
+ bio_put(bio);
+}
+
+/*
+ * Low-level block read/write IO operations.
+ */
+static struct bio *__bio_alloc(struct f2fs_sb_info *sbi, block_t blk_addr,
+ int npages, bool is_read)
+{
+ struct bio *bio;
+
+ /* No failure on bio allocation */
+ bio = bio_alloc(GFP_NOIO, npages);
+
+ bio->bi_bdev = sbi->sb->s_bdev;
- bio->bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
++ bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
+ bio->bi_end_io = is_read ? f2fs_read_end_io : f2fs_write_end_io;
+
+ return bio;
+}
+
+static void __submit_merged_bio(struct f2fs_bio_info *io)
+{
+ struct f2fs_io_info *fio = &io->fio;
+ int rw;
+
+ if (!io->bio)
+ return;
+
+ rw = fio->rw;
+
+ if (is_read_io(rw)) {
+ trace_f2fs_submit_read_bio(io->sbi->sb, rw,
+ fio->type, io->bio);
+ submit_bio(rw, io->bio);
+ } else {
+ trace_f2fs_submit_write_bio(io->sbi->sb, rw,
+ fio->type, io->bio);
+ /*
+ * META_FLUSH is only from the checkpoint procedure, and we
+ * should wait this metadata bio for FS consistency.
+ */
+ if (fio->type == META_FLUSH) {
+ DECLARE_COMPLETION_ONSTACK(wait);
+ io->bio->bi_private = &wait;
+ submit_bio(rw, io->bio);
+ wait_for_completion(&wait);
+ } else {
+ submit_bio(rw, io->bio);
+ }
+ }
+
+ io->bio = NULL;
+}
+
+void f2fs_submit_merged_bio(struct f2fs_sb_info *sbi,
+ enum page_type type, int rw)
+{
+ enum page_type btype = PAGE_TYPE_OF_BIO(type);
+ struct f2fs_bio_info *io;
+
+ io = is_read_io(rw) ? &sbi->read_io : &sbi->write_io[btype];
+
+ mutex_lock(&io->io_mutex);
+
+ /* change META to META_FLUSH in the checkpoint procedure */
+ if (type >= META_FLUSH) {
+ io->fio.type = META_FLUSH;
+ io->fio.rw = WRITE_FLUSH_FUA | REQ_META | REQ_PRIO;
+ }
+ __submit_merged_bio(io);
+ mutex_unlock(&io->io_mutex);
+}
+
+/*
+ * Fill the locked page with data located in the block address.
+ * Return unlocked page.
+ */
+int f2fs_submit_page_bio(struct f2fs_sb_info *sbi, struct page *page,
+ block_t blk_addr, int rw)
+{
+ struct bio *bio;
+
+ trace_f2fs_submit_page_bio(page, blk_addr, rw);
+
+ /* Allocate a new bio */
+ bio = __bio_alloc(sbi, blk_addr, 1, is_read_io(rw));
+
+ if (bio_add_page(bio, page, PAGE_CACHE_SIZE, 0) < PAGE_CACHE_SIZE) {
+ bio_put(bio);
+ f2fs_put_page(page, 1);
+ return -EFAULT;
+ }
+
+ submit_bio(rw, bio);
+ return 0;
+}
+
+void f2fs_submit_page_mbio(struct f2fs_sb_info *sbi, struct page *page,
+ block_t blk_addr, struct f2fs_io_info *fio)
+{
+ enum page_type btype = PAGE_TYPE_OF_BIO(fio->type);
+ struct f2fs_bio_info *io;
+ bool is_read = is_read_io(fio->rw);
+
+ io = is_read ? &sbi->read_io : &sbi->write_io[btype];
+
+ verify_block_addr(sbi, blk_addr);
+
+ mutex_lock(&io->io_mutex);
+
+ if (!is_read)
+ inc_page_count(sbi, F2FS_WRITEBACK);
+
+ if (io->bio && (io->last_block_in_bio != blk_addr - 1 ||
+ io->fio.rw != fio->rw))
+ __submit_merged_bio(io);
+alloc_new:
+ if (io->bio == NULL) {
+ int bio_blocks = MAX_BIO_BLOCKS(max_hw_blocks(sbi));
+
+ io->bio = __bio_alloc(sbi, blk_addr, bio_blocks, is_read);
+ io->fio = *fio;
+ }
+
+ if (bio_add_page(io->bio, page, PAGE_CACHE_SIZE, 0) <
+ PAGE_CACHE_SIZE) {
+ __submit_merged_bio(io);
+ goto alloc_new;
+ }
+
+ io->last_block_in_bio = blk_addr;
+
+ mutex_unlock(&io->io_mutex);
+ trace_f2fs_submit_page_mbio(page, fio->rw, fio->type, blk_addr);
+}
+
/*
* Lock ordering for the change of data block address:
* ->data_page
diff --cc include/trace/events/f2fs.h
index 3b9f28dfc849,bd3ee4fbe7a7..67f38faac589
--- a/include/trace/events/f2fs.h
+++ b/include/trace/events/f2fs.h
@@@ -627,16 -614,16 +627,16 @@@ DECLARE_EVENT_CLASS(f2fs__submit_bio

TP_fast_assign(
__entry->dev = sb->s_dev;
- __entry->btype = btype;
- __entry->sync = sync;
+ __entry->rw = rw;
+ __entry->type = type;
- __entry->sector = bio->bi_sector;
- __entry->size = bio->bi_size;
+ __entry->sector = bio->bi_iter.bi_sector;
+ __entry->size = bio->bi_iter.bi_size;
),

- TP_printk("dev = (%d,%d), type = %s, io = %s, sector = %lld, size = %u",
+ TP_printk("dev = (%d,%d), %s%s, %s, sector = %lld, size = %u",
show_dev(__entry),
- show_block_type(__entry->btype),
- __entry->sync ? "sync" : "no sync",
+ show_bio_type(__entry->rw),
+ show_block_type(__entry->type),
(unsigned long long)__entry->sector,
__entry->size)
);

--
Cheers,
Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx>

Attachment: pgp00000.pgp
Description: PGP signature