Merge tag 'dm-3.8-fixes' of git://git.kernel.org/pub/scm/linux/kernel…

…/git/agk/linux-dm

Pull dm update from Alasdair G Kergon:
 "Miscellaneous device-mapper fixes, cleanups and performance
  improvements.

  Of particular note:
   - Disable broken WRITE SAME support in all targets except linear and
     striped.  Use it when kcopyd is zeroing blocks.
   - Remove several mempools from targets by moving the data into the
     bio's new front_pad area(which dm calls 'per_bio_data').
   - Fix a race in thin provisioning if discards are misused.
   - Prevent userspace from interfering with the ioctl parameters and
     use kmalloc for the data buffer if it's small instead of vmalloc.
   - Throttle some annoying error messages when I/O fails."

* tag 'dm-3.8-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-dm: (36 commits)
  dm stripe: add WRITE SAME support
  dm: remove map_info
  dm snapshot: do not use map_context
  dm thin: dont use map_context
  dm raid1: dont use map_context
  dm flakey: dont use map_context
  dm raid1: rename read_record to bio_record
  dm: move target request nr to dm_target_io
  dm snapshot: use per_bio_data
  dm verity: use per_bio_data
  dm raid1: use per_bio_data
  dm: introduce per_bio_data
  dm kcopyd: add WRITE SAME support to dm_kcopyd_zero
  dm linear: add WRITE SAME support
  dm: add WRITE SAME support
  dm: prepare to support WRITE SAME
  dm ioctl: use kmalloc if possible
  dm ioctl: remove PF_MEMALLOC
  dm persistent data: improve improve space map block alloc failure message
  dm thin: use DMERR_LIMIT for errors
  ...

drivers/md/dm-bio-prison.c

@@ -207,31 +207,6 @@ void dm_cell_release(struct dm_bio_prison_cell *cell, struct bio_list *bios)
 }
 EXPORT_SYMBOL_GPL(dm_cell_release);
 
-/*
- * There are a couple of places where we put a bio into a cell briefly
- * before taking it out again.  In these situations we know that no other
- * bio may be in the cell.  This function releases the cell, and also does
- * a sanity check.
- */
-static void __cell_release_singleton(struct dm_bio_prison_cell *cell, struct bio *bio)
-{
-	BUG_ON(cell->holder != bio);
-	BUG_ON(!bio_list_empty(&cell->bios));
-
-	__cell_release(cell, NULL);
-}
-
-void dm_cell_release_singleton(struct dm_bio_prison_cell *cell, struct bio *bio)
-{
-	unsigned long flags;
-	struct dm_bio_prison *prison = cell->prison;
-
-	spin_lock_irqsave(&prison->lock, flags);
-	__cell_release_singleton(cell, bio);
-	spin_unlock_irqrestore(&prison->lock, flags);
-}
-EXPORT_SYMBOL_GPL(dm_cell_release_singleton);
-
 /*
  * Sometimes we don't want the holder, just the additional bios.
  */

drivers/md/dm-bio-prison.h

@@ -44,7 +44,6 @@ int dm_bio_detain(struct dm_bio_prison *prison, struct dm_cell_key *key,
 		  struct bio *inmate, struct dm_bio_prison_cell **ref);
 
 void dm_cell_release(struct dm_bio_prison_cell *cell, struct bio_list *bios);
-void dm_cell_release_singleton(struct dm_bio_prison_cell *cell, struct bio *bio); // FIXME: bio arg not needed
 void dm_cell_release_no_holder(struct dm_bio_prison_cell *cell, struct bio_list *inmates);
 void dm_cell_error(struct dm_bio_prison_cell *cell);
 

drivers/md/dm-crypt.c

@@ -1689,8 +1689,7 @@ static int crypt_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 	return ret;
 }
 
-static int crypt_map(struct dm_target *ti, struct bio *bio,
-		     union map_info *map_context)
+static int crypt_map(struct dm_target *ti, struct bio *bio)
 {
 	struct dm_crypt_io *io;
 	struct crypt_config *cc = ti->private;
@@ -1846,7 +1845,7 @@ static int crypt_iterate_devices(struct dm_target *ti,
 
 static struct target_type crypt_target = {
 	.name   = "crypt",
-	.version = {1, 11, 0},
+	.version = {1, 12, 0},
 	.module = THIS_MODULE,
 	.ctr    = crypt_ctr,
 	.dtr    = crypt_dtr,

drivers/md/dm-delay.c

@@ -274,8 +274,7 @@ static void delay_resume(struct dm_target *ti)
 	atomic_set(&dc->may_delay, 1);
 }
 
-static int delay_map(struct dm_target *ti, struct bio *bio,
-		     union map_info *map_context)
+static int delay_map(struct dm_target *ti, struct bio *bio)
 {
 	struct delay_c *dc = ti->private;
 
@@ -338,7 +337,7 @@ static int delay_iterate_devices(struct dm_target *ti,
 
 static struct target_type delay_target = {
 	.name	     = "delay",
-	.version     = {1, 1, 0},
+	.version     = {1, 2, 0},
 	.module      = THIS_MODULE,
 	.ctr	     = delay_ctr,
 	.dtr	     = delay_dtr,

drivers/md/dm-flakey.c

@@ -39,6 +39,10 @@ enum feature_flag_bits {
 	DROP_WRITES
 };
 
+struct per_bio_data {
+	bool bio_submitted;
+};
+
 static int parse_features(struct dm_arg_set *as, struct flakey_c *fc,
 			  struct dm_target *ti)
 {
@@ -214,6 +218,7 @@ static int flakey_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 
 	ti->num_flush_requests = 1;
 	ti->num_discard_requests = 1;
+	ti->per_bio_data_size = sizeof(struct per_bio_data);
 	ti->private = fc;
 	return 0;
 
@@ -265,19 +270,20 @@ static void corrupt_bio_data(struct bio *bio, struct flakey_c *fc)
 	}
 }
 
-static int flakey_map(struct dm_target *ti, struct bio *bio,
-		      union map_info *map_context)
+static int flakey_map(struct dm_target *ti, struct bio *bio)
 {
 	struct flakey_c *fc = ti->private;
 	unsigned elapsed;
+	struct per_bio_data *pb = dm_per_bio_data(bio, sizeof(struct per_bio_data));
+	pb->bio_submitted = false;
 
 	/* Are we alive ? */
 	elapsed = (jiffies - fc->start_time) / HZ;
 	if (elapsed % (fc->up_interval + fc->down_interval) >= fc->up_interval) {
 		/*
 		 * Flag this bio as submitted while down.
 		 */
-		map_context->ll = 1;
+		pb->bio_submitted = true;
 
 		/*
 		 * Map reads as normal.
@@ -314,17 +320,16 @@ static int flakey_map(struct dm_target *ti, struct bio *bio,
 	return DM_MAPIO_REMAPPED;
 }
 
-static int flakey_end_io(struct dm_target *ti, struct bio *bio,
-			 int error, union map_info *map_context)
+static int flakey_end_io(struct dm_target *ti, struct bio *bio, int error)
 {
 	struct flakey_c *fc = ti->private;
-	unsigned bio_submitted_while_down = map_context->ll;
+	struct per_bio_data *pb = dm_per_bio_data(bio, sizeof(struct per_bio_data));
 
 	/*
 	 * Corrupt successful READs while in down state.
 	 * If flags were specified, only corrupt those that match.
 	 */
-	if (fc->corrupt_bio_byte && !error && bio_submitted_while_down &&
+	if (fc->corrupt_bio_byte && !error && pb->bio_submitted &&
 	    (bio_data_dir(bio) == READ) && (fc->corrupt_bio_rw == READ) &&
 	    all_corrupt_bio_flags_match(bio, fc))
 		corrupt_bio_data(bio, fc);
@@ -406,7 +411,7 @@ static int flakey_iterate_devices(struct dm_target *ti, iterate_devices_callout_
 
 static struct target_type flakey_target = {
 	.name   = "flakey",
-	.version = {1, 2, 0},
+	.version = {1, 3, 0},
 	.module = THIS_MODULE,
 	.ctr    = flakey_ctr,
 	.dtr    = flakey_dtr,

drivers/md/dm-io.c

@@ -287,7 +287,8 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where,
 	unsigned num_bvecs;
 	sector_t remaining = where->count;
 	struct request_queue *q = bdev_get_queue(where->bdev);
-	sector_t discard_sectors;
+	unsigned short logical_block_size = queue_logical_block_size(q);
+	sector_t num_sectors;
 
 	/*
 	 * where->count may be zero if rw holds a flush and we need to
@@ -297,7 +298,7 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where,
 		/*
 		 * Allocate a suitably sized-bio.
 		 */
-		if (rw & REQ_DISCARD)
+		if ((rw & REQ_DISCARD) || (rw & REQ_WRITE_SAME))
 			num_bvecs = 1;
 		else
 			num_bvecs = min_t(int, bio_get_nr_vecs(where->bdev),
@@ -310,9 +311,21 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where,
 		store_io_and_region_in_bio(bio, io, region);
 
 		if (rw & REQ_DISCARD) {
-			discard_sectors = min_t(sector_t, q->limits.max_discard_sectors, remaining);
-			bio->bi_size = discard_sectors << SECTOR_SHIFT;
-			remaining -= discard_sectors;
+			num_sectors = min_t(sector_t, q->limits.max_discard_sectors, remaining);
+			bio->bi_size = num_sectors << SECTOR_SHIFT;
+			remaining -= num_sectors;
+		} else if (rw & REQ_WRITE_SAME) {
+			/*
+			 * WRITE SAME only uses a single page.
+			 */
+			dp->get_page(dp, &page, &len, &offset);
+			bio_add_page(bio, page, logical_block_size, offset);
+			num_sectors = min_t(sector_t, q->limits.max_write_same_sectors, remaining);
+			bio->bi_size = num_sectors << SECTOR_SHIFT;
+
+			offset = 0;
+			remaining -= num_sectors;
+			dp->next_page(dp);
 		} else while (remaining) {
 			/*
 			 * Try and add as many pages as possible.

drivers/md/dm-ioctl.c

@@ -1543,7 +1543,21 @@ static int check_version(unsigned int cmd, struct dm_ioctl __user *user)
 	return r;
 }
 
-static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl **param)
+#define DM_PARAMS_VMALLOC	0x0001	/* Params alloced with vmalloc not kmalloc */
+#define DM_WIPE_BUFFER		0x0010	/* Wipe input buffer before returning from ioctl */
+
+static void free_params(struct dm_ioctl *param, size_t param_size, int param_flags)
+{
+	if (param_flags & DM_WIPE_BUFFER)
+		memset(param, 0, param_size);
+
+	if (param_flags & DM_PARAMS_VMALLOC)
+		vfree(param);
+	else
+		kfree(param);
+}
+
+static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl **param, int *param_flags)
 {
 	struct dm_ioctl tmp, *dmi;
 	int secure_data;
@@ -1556,7 +1570,21 @@ static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl **param)
 
 	secure_data = tmp.flags & DM_SECURE_DATA_FLAG;
 
-	dmi = vmalloc(tmp.data_size);
+	*param_flags = secure_data ? DM_WIPE_BUFFER : 0;
+
+	/*
+	 * Try to avoid low memory issues when a device is suspended.
+	 * Use kmalloc() rather than vmalloc() when we can.
+	 */
+	dmi = NULL;
+	if (tmp.data_size <= KMALLOC_MAX_SIZE)
+		dmi = kmalloc(tmp.data_size, GFP_NOIO | __GFP_NORETRY | __GFP_NOMEMALLOC | __GFP_NOWARN);
+
+	if (!dmi) {
+		dmi = __vmalloc(tmp.data_size, GFP_NOIO | __GFP_REPEAT | __GFP_HIGH, PAGE_KERNEL);
+		*param_flags |= DM_PARAMS_VMALLOC;
+	}
+
 	if (!dmi) {
 		if (secure_data && clear_user(user, tmp.data_size))
 			return -EFAULT;
@@ -1566,6 +1594,14 @@ static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl **param)
 	if (copy_from_user(dmi, user, tmp.data_size))
 		goto bad;
 
+	/*
+	 * Abort if something changed the ioctl data while it was being copied.
+	 */
+	if (dmi->data_size != tmp.data_size) {
+		DMERR("rejecting ioctl: data size modified while processing parameters");
+		goto bad;
+	}
+
 	/* Wipe the user buffer so we do not return it to userspace */
 	if (secure_data && clear_user(user, tmp.data_size))
 		goto bad;
@@ -1574,9 +1610,8 @@ static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl **param)
 	return 0;
 
 bad:
-	if (secure_data)
-		memset(dmi, 0, tmp.data_size);
-	vfree(dmi);
+	free_params(dmi, tmp.data_size, *param_flags);
+
 	return -EFAULT;
 }
 
@@ -1613,7 +1648,7 @@ static int validate_params(uint cmd, struct dm_ioctl *param)
 static int ctl_ioctl(uint command, struct dm_ioctl __user *user)
 {
 	int r = 0;
-	int wipe_buffer;
+	int param_flags;
 	unsigned int cmd;
 	struct dm_ioctl *uninitialized_var(param);
 	ioctl_fn fn = NULL;
@@ -1648,25 +1683,15 @@ static int ctl_ioctl(uint command, struct dm_ioctl __user *user)
 		return -ENOTTY;
 	}
 
-	/*
-	 * Trying to avoid low memory issues when a device is
-	 * suspended.
-	 */
-	current->flags |= PF_MEMALLOC;
-
 	/*
 	 * Copy the parameters into kernel space.
 	 */
-	r = copy_params(user, &param);
-
-	current->flags &= ~PF_MEMALLOC;
+	r = copy_params(user, &param, &param_flags);
 
 	if (r)
 		return r;
 
 	input_param_size = param->data_size;
-	wipe_buffer = param->flags & DM_SECURE_DATA_FLAG;
-
 	r = validate_params(cmd, param);
 	if (r)
 		goto out;
@@ -1681,10 +1706,7 @@ static int ctl_ioctl(uint command, struct dm_ioctl __user *user)
 		r = -EFAULT;
 
 out:
-	if (wipe_buffer)
-		memset(param, 0, input_param_size);
-
-	vfree(param);
+	free_params(param, input_param_size, param_flags);
 	return r;
 }
 

drivers/md/dm-kcopyd.c

@@ -349,7 +349,7 @@ static void complete_io(unsigned long error, void *context)
 	struct dm_kcopyd_client *kc = job->kc;
 
 	if (error) {
-		if (job->rw == WRITE)
+		if (job->rw & WRITE)
 			job->write_err |= error;
 		else
 			job->read_err = 1;
@@ -361,7 +361,7 @@ static void complete_io(unsigned long error, void *context)
 		}
 	}
 
-	if (job->rw == WRITE)
+	if (job->rw & WRITE)
 		push(&kc->complete_jobs, job);
 
 	else {
@@ -432,7 +432,7 @@ static int process_jobs(struct list_head *jobs, struct dm_kcopyd_client *kc,
 
 		if (r < 0) {
 			/* error this rogue job */
-			if (job->rw == WRITE)
+			if (job->rw & WRITE)
 				job->write_err = (unsigned long) -1L;
 			else
 				job->read_err = 1;
@@ -585,6 +585,7 @@ int dm_kcopyd_copy(struct dm_kcopyd_client *kc, struct dm_io_region *from,
 		   unsigned int flags, dm_kcopyd_notify_fn fn, void *context)
 {
 	struct kcopyd_job *job;
+	int i;
 
 	/*
 	 * Allocate an array of jobs consisting of one master job
@@ -611,7 +612,16 @@ int dm_kcopyd_copy(struct dm_kcopyd_client *kc, struct dm_io_region *from,
 		memset(&job->source, 0, sizeof job->source);
 		job->source.count = job->dests[0].count;
 		job->pages = &zero_page_list;
-		job->rw = WRITE;
+
+		/*
+		 * Use WRITE SAME to optimize zeroing if all dests support it.
+		 */
+		job->rw = WRITE | REQ_WRITE_SAME;
+		for (i = 0; i < job->num_dests; i++)
+			if (!bdev_write_same(job->dests[i].bdev)) {
+				job->rw = WRITE;
+				break;
+			}
 	}
 
 	job->fn = fn;

drivers/md/dm-linear.c

@@ -55,6 +55,7 @@ static int linear_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 
 	ti->num_flush_requests = 1;
 	ti->num_discard_requests = 1;
+	ti->num_write_same_requests = 1;
 	ti->private = lc;
 	return 0;
 
@@ -87,8 +88,7 @@ static void linear_map_bio(struct dm_target *ti, struct bio *bio)
 		bio->bi_sector = linear_map_sector(ti, bio->bi_sector);
 }
 
-static int linear_map(struct dm_target *ti, struct bio *bio,
-		      union map_info *map_context)
+static int linear_map(struct dm_target *ti, struct bio *bio)
 {
 	linear_map_bio(ti, bio);
 
@@ -155,7 +155,7 @@ static int linear_iterate_devices(struct dm_target *ti,
 
 static struct target_type linear_target = {
 	.name   = "linear",
-	.version = {1, 1, 0},
+	.version = {1, 2, 0},
 	.module = THIS_MODULE,
 	.ctr    = linear_ctr,
 	.dtr    = linear_dtr,