zvol processing should use struct bio

Internally, zvols are files exposed through the block device API. This
is intended to reduce overhead when things require block devices.
However, the ZoL zvol code emulates a traditional block device in that
it has a top half and a bottom half. This is an unnecessary source of
overhead that does not exist on any other OpenZFS platform does this.
This patch removes it. Early users of this patch reported double digit
performance gains in IOPS on zvols in the range of 50% to 80%.

Comments in the code suggest that the current implementation was done to
obtain IO merging from Linux's IO elevator. However, the DMU already
does write merging while arc_read() should implicitly merge read IOs
because only 1 thread is permitted to fetch the buffer into ARC. In
addition, commercial ZFSOnLinux distributions report that regular files
are more performant than zvols under the current implementation, and the
main consumers of zvols are VMs and iSCSI targets, which have their own
elevators to merge IOs.

Some minor refactoring allows us to register zfs_request() as our
->make_request() handler in place of the generic_make_request()
function. This eliminates the layer of code that broke IO requests on
zvols into a top half and a bottom half. This has several benefits:

1. No per zvol spinlocks.
2. No redundant IO elevator processing.
3. Interrupts are disabled only when actually necessary.
4. No redispatching of IOs when all taskq threads are busy.
5. Linux's page out routines will properly block.
6. Many autotools checks become obsolete.

An unfortunate consequence of eliminating the layer that
generic_make_request() is that we no longer calls the instrumentation
hooks for block IO accounting. Those hooks are GPL-exported, so we
cannot call them ourselves and consequently, we lose the ability to do
IO monitoring via iostat.  Since zvols are internally files mapped as
block devices, this should be okay. Anyone who is willing to accept the
performance penalty for the block IO layer's accounting could use the
loop device in between the zvol and its consumer. Alternatively, perf
and ftrace likely could be used. Also, tools like latencytop will still
work. Tools such as latencytop sometimes provide a better view of
performance bottlenecks than the traditional block IO accounting tools
do.

Lastly, if direct reclaim occurs during spacemap loading and swap is on
a zvol, this code will deadlock. That deadlock could already occur with
sync=always on zvols. Given that swap on zvols is not yet production
ready, this is not a blocker.

Signed-off-by: Richard Yao <ryao@gentoo.org>

include/linux/blkdev_compat.h

@@ -295,10 +295,16 @@ struct req_iterator {
 #define	BIO_BI_SECTOR(bio)	(bio)->bi_iter.bi_sector
 #define	BIO_BI_SIZE(bio)	(bio)->bi_iter.bi_size
 #define	BIO_BI_IDX(bio)		(bio)->bi_iter.bi_idx
+#define	bio_for_each_segment4(bv, bvp, b, i)	\
+	bio_for_each_segment((bv), (b), (i))
+typedef struct bvec_iter bvec_iterator_t;
 #else
 #define	BIO_BI_SECTOR(bio)	(bio)->bi_sector
 #define	BIO_BI_SIZE(bio)	(bio)->bi_size
 #define	BIO_BI_IDX(bio)		(bio)->bi_idx
+#define	bio_for_each_segment4(bv, bvp, b, i)	\
+	bio_for_each_segment((bvp), (b), (i))
+typedef int bvec_iterator_t;
 #endif
 
 /*
@@ -457,17 +463,30 @@ bio_set_flags_failfast(struct block_device *bdev, int *flags)
 #define	VDEV_REQ_FUA			REQ_FUA
 #else
 #define	VDEV_WRITE_FLUSH_FUA		WRITE_BARRIER
+#ifdef HAVE_BIO_RW_BARRIER
+#define	VDEV_REQ_FLUSH			(1 << BIO_RW_BARRIER)
+#define	VDEV_REQ_FUA			(1 << BIO_RW_BARRIER)
+#else
 #define	VDEV_REQ_FLUSH			REQ_HARDBARRIER
 #define	VDEV_REQ_FUA			REQ_FUA
 #endif
+#endif
 
 /*
  * 2.6.32 API change
  * Use the normal I/O patch for discards.
  */
-#ifdef REQ_DISCARD
+#ifdef QUEUE_FLAG_DISCARD
+#ifdef HAVE_BIO_RW_DISCARD
+#define	VDEV_REQ_DISCARD		(1 << BIO_RW_DISCARD)
+#else
 #define	VDEV_REQ_DISCARD		REQ_DISCARD
 #endif
+#else
+#error	"Allowing the build will cause discard requests to become writes "
+	"potentially triggering the DMU_MAX_ACCESS assertion. Please file a "
+	"an issue report at: https://github.com/zfsonlinux/zfs/issues/new"
+#endif
 
 /*
  * 2.6.33 API change

include/sys/dmu.h

@@ -710,8 +710,8 @@ void dmu_prealloc(objset_t *os, uint64_t object, uint64_t offset, uint64_t size,
 	dmu_tx_t *tx);
 #ifdef _KERNEL
 #include <linux/blkdev_compat.h>
-int dmu_read_req(objset_t *os, uint64_t object, struct request *req);
-int dmu_write_req(objset_t *os, uint64_t object, struct request *req,
+int dmu_read_bio(objset_t *os, uint64_t object, struct bio *bio);
+int dmu_write_bio(objset_t *os, uint64_t object, struct bio *bio,
 	dmu_tx_t *tx);
 int dmu_read_uio(objset_t *os, uint64_t object, struct uio *uio, uint64_t size);
 int dmu_read_uio_dbuf(dmu_buf_t *zdb, struct uio *uio, uint64_t size);