buffer, writeback: make __block_write_full_page() honor cgroup writeback
authorTejun Heo <tj@kernel.org>
Tue, 2 Jun 2015 14:37:23 +0000 (08:37 -0600)
committerJens Axboe <axboe@fb.com>
Tue, 2 Jun 2015 14:37:23 +0000 (08:37 -0600)
[__]block_write_full_page() is used to implement ->writepage in
various filesystems.  All writeback logic is now updated to handle
cgroup writeback and the block cgroup to issue IOs for is encoded in
writeback_control and can be retrieved from the inode; however,
[__]block_write_full_page() currently ignores the blkcg indicated by
inode and issues all bio's without explicit blkcg association.

This patch adds submit_bh_blkcg() which associates the bio with the
specified blkio cgroup before issuing and uses it in
__block_write_full_page() so that the issued bio's are associated with
inode_to_wb_blkcg_css(inode).

v2: Updated for per-inode wb association.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Jan Kara <jack@suse.cz>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <axboe@fb.com>
fs/buffer.c
include/linux/backing-dev.h

index 23b640d..b85e941 100644 (file)
@@ -30,6 +30,7 @@
 #include <linux/quotaops.h>
 #include <linux/highmem.h>
 #include <linux/export.h>
+#include <linux/backing-dev.h>
 #include <linux/writeback.h>
 #include <linux/hash.h>
 #include <linux/suspend.h>
@@ -44,6 +45,9 @@
 #include <trace/events/block.h>
 
 static int fsync_buffers_list(spinlock_t *lock, struct list_head *list);
+static int submit_bh_blkcg(int rw, struct buffer_head *bh,
+                          unsigned long bio_flags,
+                          struct cgroup_subsys_state *blkcg_css);
 
 #define BH_ENTRY(list) list_entry((list), struct buffer_head, b_assoc_buffers)
 
@@ -1704,8 +1708,8 @@ static int __block_write_full_page(struct inode *inode, struct page *page,
        struct buffer_head *bh, *head;
        unsigned int blocksize, bbits;
        int nr_underway = 0;
-       int write_op = (wbc->sync_mode == WB_SYNC_ALL ?
-                       WRITE_SYNC : WRITE);
+       int write_op = (wbc->sync_mode == WB_SYNC_ALL ? WRITE_SYNC : WRITE);
+       struct cgroup_subsys_state *blkcg_css = inode_to_wb_blkcg_css(inode);
 
        head = create_page_buffers(page, inode,
                                        (1 << BH_Dirty)|(1 << BH_Uptodate));
@@ -1794,7 +1798,7 @@ static int __block_write_full_page(struct inode *inode, struct page *page,
        do {
                struct buffer_head *next = bh->b_this_page;
                if (buffer_async_write(bh)) {
-                       submit_bh(write_op, bh);
+                       submit_bh_blkcg(write_op, bh, 0, blkcg_css);
                        nr_underway++;
                }
                bh = next;
@@ -1848,7 +1852,7 @@ recover:
                struct buffer_head *next = bh->b_this_page;
                if (buffer_async_write(bh)) {
                        clear_buffer_dirty(bh);
-                       submit_bh(write_op, bh);
+                       submit_bh_blkcg(write_op, bh, 0, blkcg_css);
                        nr_underway++;
                }
                bh = next;
@@ -3013,7 +3017,9 @@ void guard_bio_eod(int rw, struct bio *bio)
        }
 }
 
-int _submit_bh(int rw, struct buffer_head *bh, unsigned long bio_flags)
+static int submit_bh_blkcg(int rw, struct buffer_head *bh,
+                          unsigned long bio_flags,
+                          struct cgroup_subsys_state *blkcg_css)
 {
        struct bio *bio;
 
@@ -3035,6 +3041,9 @@ int _submit_bh(int rw, struct buffer_head *bh, unsigned long bio_flags)
         */
        bio = bio_alloc(GFP_NOIO, 1);
 
+       if (blkcg_css)
+               bio_associate_blkcg(bio, blkcg_css);
+
        bio->bi_iter.bi_sector = bh->b_blocknr * (bh->b_size >> 9);
        bio->bi_bdev = bh->b_bdev;
        bio->bi_io_vec[0].bv_page = bh->b_page;
@@ -3059,11 +3068,16 @@ int _submit_bh(int rw, struct buffer_head *bh, unsigned long bio_flags)
        submit_bio(rw, bio);
        return 0;
 }
+
+int _submit_bh(int rw, struct buffer_head *bh, unsigned long bio_flags)
+{
+       return submit_bh_blkcg(rw, bh, bio_flags, NULL);
+}
 EXPORT_SYMBOL_GPL(_submit_bh);
 
 int submit_bh(int rw, struct buffer_head *bh)
 {
-       return _submit_bh(rw, bh, 0);
+       return submit_bh_blkcg(rw, bh, 0, NULL);
 }
 EXPORT_SYMBOL(submit_bh);
 
index 9cc11e5..e9d7373 100644 (file)
@@ -393,6 +393,12 @@ static inline struct bdi_writeback *inode_to_wb(struct inode *inode)
        return inode->i_wb;
 }
 
+static inline struct cgroup_subsys_state *
+inode_to_wb_blkcg_css(struct inode *inode)
+{
+       return inode_to_wb(inode)->blkcg_css;
+}
+
 struct wb_iter {
        int                     start_blkcg_id;
        struct radix_tree_iter  tree_iter;
@@ -510,6 +516,12 @@ static inline void wb_blkcg_offline(struct blkcg *blkcg)
 {
 }
 
+static inline struct cgroup_subsys_state *
+inode_to_wb_blkcg_css(struct inode *inode)
+{
+       return blkcg_root_css;
+}
+
 struct wb_iter {
        int             next_id;
 };