[PATCH 7/9] io-throttle instrumentation

[Andrea Righi <righi.andrea@gmail.com>]

Apply the io-throttle controller to the opportune kernel functions.

Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
---
 block/blk-core.c      |    8 ++++++++
 fs/aio.c              |   12 ++++++++++++
 include/linux/sched.h |    7 +++++++
 kernel/fork.c         |    7 +++++++
 mm/readahead.c        |    3 +++
 5 files changed, 37 insertions(+), 0 deletions(-)

diff --git a/block/blk-core.c b/block/blk-core.c
index 07ab754..4d7f9f6 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -26,6 +26,7 @@
 #include <linux/swap.h>
 #include <linux/writeback.h>
 #include <linux/task_io_accounting_ops.h>
+#include <linux/blk-io-throttle.h>
 #include <linux/blktrace_api.h>
 #include <linux/fault-inject.h>
 #include <trace/block.h>
@@ -1547,11 +1548,16 @@ void submit_bio(int rw, struct bio *bio)
 	 * go through the normal accounting stuff before submission.
 	 */
 	if (bio_has_data(bio)) {
+		unsigned long sleep = 0;
+
 		if (rw & WRITE) {
 			count_vm_events(PGPGOUT, count);
+			sleep = cgroup_io_throttle(bio,
+					bio->bi_bdev, bio->bi_size);
 		} else {
 			task_io_account_read(bio->bi_size);
 			count_vm_events(PGPGIN, count);
+			cgroup_io_throttle(NULL, bio->bi_bdev, bio->bi_size);
 		}
 
 		if (unlikely(block_dump)) {
@@ -1562,6 +1568,8 @@ void submit_bio(int rw, struct bio *bio)
 				(unsigned long long)bio->bi_sector,
 				bdevname(bio->bi_bdev, b));
 		}
+		if (sleep && !iothrottle_make_request(bio, jiffies + sleep))
+			return;
 	}
 
 	generic_make_request(bio);
diff --git a/fs/aio.c b/fs/aio.c
index 76da125..ab6c457 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -22,6 +22,7 @@
 #include <linux/sched.h>
 #include <linux/fs.h>
 #include <linux/file.h>
+#include <linux/blk-io-throttle.h>
 #include <linux/mm.h>
 #include <linux/mman.h>
 #include <linux/slab.h>
@@ -1587,6 +1588,7 @@ static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
 {
 	struct kiocb *req;
 	struct file *file;
+	struct block_device *bdev;
 	ssize_t ret;
 
 	/* enforce forwards compatibility on users */
@@ -1609,6 +1611,14 @@ static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
 	if (unlikely(!file))
 		return -EBADF;
 
+	/* check if we're exceeding the IO throttling limits */
+	bdev = as_to_bdev(file->f_mapping);
+	ret = cgroup_io_throttle(NULL, bdev, 0);
+	if (unlikely(ret)) {
+		fput(file);
+		return -EAGAIN;
+	}
+
 	req = aio_get_req(ctx);		/* returns with 2 references to req */
 	if (unlikely(!req)) {
 		fput(file);
@@ -1652,12 +1662,14 @@ static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
 		goto out_put_req;
 
 	spin_lock_irq(&ctx->ctx_lock);
+	set_in_aio();
 	aio_run_iocb(req);
 	if (!list_empty(&ctx->run_list)) {
 		/* drain the run list */
 		while (__aio_run_iocbs(ctx))
 			;
 	}
+	unset_in_aio();
 	spin_unlock_irq(&ctx->ctx_lock);
 	aio_put_req(req);	/* drop extra ref to req */
 	return 0;
diff --git a/include/linux/sched.h b/include/linux/sched.h
index b4c38bc..e0cd710 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1356,6 +1356,13 @@ struct task_struct {
 	unsigned long ptrace_message;
 	siginfo_t *last_siginfo; /* For ptrace use.  */
 	struct task_io_accounting ioac;
+#ifdef CONFIG_CGROUP_IO_THROTTLE
+	atomic_t in_aio;
+	unsigned long long io_throttle_bw_cnt;
+	unsigned long long io_throttle_bw_sleep;
+	unsigned long long io_throttle_iops_cnt;
+	unsigned long long io_throttle_iops_sleep;
+#endif
 #if defined(CONFIG_TASK_XACCT)
 	u64 acct_rss_mem1;	/* accumulated rss usage */
 	u64 acct_vm_mem1;	/* accumulated virtual memory usage */
diff --git a/kernel/fork.c b/kernel/fork.c
index b9e2edd..272c461 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1043,6 +1043,13 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	task_io_accounting_init(&p->ioac);
 	acct_clear_integrals(p);
 
+#ifdef CONFIG_CGROUP_IO_THROTTLE
+	atomic_set(&p->in_aio, 0);
+	p->io_throttle_bw_cnt = 0;
+	p->io_throttle_bw_sleep = 0;
+	p->io_throttle_iops_cnt = 0;
+	p->io_throttle_iops_sleep = 0;
+#endif
 	posix_cpu_timers_init(p);
 
 	p->lock_depth = -1;		/* -1 = no lock */
diff --git a/mm/readahead.c b/mm/readahead.c
index 133b6d5..25cae4c 100644
--- a/mm/readahead.c
+++ b/mm/readahead.c
@@ -14,6 +14,7 @@
 #include <linux/blkdev.h>
 #include <linux/backing-dev.h>
 #include <linux/task_io_accounting_ops.h>
+#include <linux/blk-io-throttle.h>
 #include <linux/pagevec.h>
 #include <linux/pagemap.h>
 
@@ -81,6 +82,7 @@ int read_cache_pages(struct address_space *mapping, struct list_head *pages,
 			int (*filler)(void *, struct page *), void *data)
 {
 	struct page *page;
+	struct block_device *bdev = as_to_bdev(mapping);
 	int ret = 0;
 
 	while (!list_empty(pages)) {
@@ -99,6 +101,7 @@ int read_cache_pages(struct address_space *mapping, struct list_head *pages,
 			break;
 		}
 		task_io_account_read(PAGE_CACHE_SIZE);
+		cgroup_io_throttle(NULL, bdev, PAGE_CACHE_SIZE);
 	}
 	return ret;
 }
-- 
1.5.6.3