Merge tag 'for-linus-20190809' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:

 - Revert of a bcache patch that caused an oops for some (Coly)

 - ata rb532 unused warning fix (Gustavo)

 - AoE kernel crash fix (He)

 - Error handling fixup for blkdev_get() (Jan)

 - libata read/write translation and SFF PIO fix (me)

 - Use after free and error handling fix for O_DIRECT fragments. There's
   still a nowait + sync oddity in there, we'll nail that start next
   week. If all else fails, I'll queue a revert of the NOWAIT change.
   (me)

 - Loop GFP_KERNEL -> GFP_NOIO deadlock fix (Mikulas)

 - Two BFQ regression fixes that caused crashes (Paolo)

* tag 'for-linus-20190809' of git://git.kernel.dk/linux-block:
  bcache: Revert "bcache: use sysfs_match_string() instead of __sysfs_match_string()"
  loop: set PF_MEMALLOC_NOIO for the worker thread
  bdev: Fixup error handling in blkdev_get()
  block, bfq: handle NULL return value by bfq_init_rq()
  block, bfq: move update of waker and woken list to queue freeing
  block, bfq: reset last_completed_rq_bfqq if the pointed queue is freed
  block: aoe: Fix kernel crash due to atomic sleep when exiting
  libata: add SG safety checks in SFF pio transfers
  libata: have ata_scsi_rw_xlat() fail invalid passthrough requests
  block: fix O_DIRECT error handling for bio fragments
  ata: rb532_cf: Fix unused variable warning in rb532_pata_driver_probe

8 files changed, 113 insertions, 51 deletions

diff --git a/block/bfq-iosched.c b/block/bfq-iosched.c
index 586fcfe227eae6..b33be928d164fe 100644
--- a/block/bfq-iosched.c
+++ b/block/bfq-iosched.c
@@ -1924,12 +1924,13 @@ static void bfq_add_request(struct request *rq)
 		 * confirmed no later than during the next
 		 * I/O-plugging interval for bfqq.
 		 */
-		if (!bfq_bfqq_has_short_ttime(bfqq) &&
+		if (bfqd->last_completed_rq_bfqq &&
+		    !bfq_bfqq_has_short_ttime(bfqq) &&
 		    ktime_get_ns() - bfqd->last_completion <
 		    200 * NSEC_PER_USEC) {
 			if (bfqd->last_completed_rq_bfqq != bfqq &&
-				   bfqd->last_completed_rq_bfqq !=
-				   bfqq->waker_bfqq) {
+			    bfqd->last_completed_rq_bfqq !=
+			    bfqq->waker_bfqq) {
 				/*
 				 * First synchronization detected with
 				 * a candidate waker queue, or with a
@@ -2250,9 +2251,14 @@ static void bfq_request_merged(struct request_queue *q, struct request *req,
 	    blk_rq_pos(container_of(rb_prev(&req->rb_node),
 				    struct request, rb_node))) {
 		struct bfq_queue *bfqq = bfq_init_rq(req);
-		struct bfq_data *bfqd = bfqq->bfqd;
+		struct bfq_data *bfqd;
 		struct request *prev, *next_rq;
 
+		if (!bfqq)
+			return;
+
+		bfqd = bfqq->bfqd;
+
 		/* Reposition request in its sort_list */
 		elv_rb_del(&bfqq->sort_list, req);
 		elv_rb_add(&bfqq->sort_list, req);
@@ -2299,6 +2305,9 @@ static void bfq_requests_merged(struct request_queue *q, struct request *rq,
 	struct bfq_queue *bfqq = bfq_init_rq(rq),
 		*next_bfqq = bfq_init_rq(next);
 
+	if (!bfqq)
+		return;
+
 	/*
 	 * If next and rq belong to the same bfq_queue and next is older
 	 * than rq, then reposition rq in the fifo (by substituting next
@@ -4764,6 +4773,8 @@ static struct request *bfq_dispatch_request(struct blk_mq_hw_ctx *hctx)
  */
 void bfq_put_queue(struct bfq_queue *bfqq)
 {
+	struct bfq_queue *item;
+	struct hlist_node *n;
 #ifdef CONFIG_BFQ_GROUP_IOSCHED
 	struct bfq_group *bfqg = bfqq_group(bfqq);
 #endif
@@ -4808,6 +4819,36 @@ void bfq_put_queue(struct bfq_queue *bfqq)
 			bfqq->bfqd->burst_size--;
 	}
 
+	/*
+	 * bfqq does not exist any longer, so it cannot be woken by
+	 * any other queue, and cannot wake any other queue. Then bfqq
+	 * must be removed from the woken list of its possible waker
+	 * queue, and all queues in the woken list of bfqq must stop
+	 * having a waker queue. Strictly speaking, these updates
+	 * should be performed when bfqq remains with no I/O source
+	 * attached to it, which happens before bfqq gets freed. In
+	 * particular, this happens when the last process associated
+	 * with bfqq exits or gets associated with a different
+	 * queue. However, both events lead to bfqq being freed soon,
+	 * and dangling references would come out only after bfqq gets
+	 * freed. So these updates are done here, as a simple and safe
+	 * way to handle all cases.
+	 */
+	/* remove bfqq from woken list */
+	if (!hlist_unhashed(&bfqq->woken_list_node))
+		hlist_del_init(&bfqq->woken_list_node);
+
+	/* reset waker for all queues in woken list */
+	hlist_for_each_entry_safe(item, n, &bfqq->woken_list,
+				  woken_list_node) {
+		item->waker_bfqq = NULL;
+		bfq_clear_bfqq_has_waker(item);
+		hlist_del_init(&item->woken_list_node);
+	}
+
+	if (bfqq->bfqd && bfqq->bfqd->last_completed_rq_bfqq == bfqq)
+		bfqq->bfqd->last_completed_rq_bfqq = NULL;
+
 	kmem_cache_free(bfq_pool, bfqq);
 #ifdef CONFIG_BFQ_GROUP_IOSCHED
 	bfqg_and_blkg_put(bfqg);
@@ -4835,9 +4876,6 @@ static void bfq_put_cooperator(struct bfq_queue *bfqq)
 
 static void bfq_exit_bfqq(struct bfq_data *bfqd, struct bfq_queue *bfqq)
 {
-	struct bfq_queue *item;
-	struct hlist_node *n;
-
 	if (bfqq == bfqd->in_service_queue) {
 		__bfq_bfqq_expire(bfqd, bfqq, BFQQE_BUDGET_TIMEOUT);
 		bfq_schedule_dispatch(bfqd);
@@ -4847,18 +4885,6 @@ static void bfq_exit_bfqq(struct bfq_data *bfqd, struct bfq_queue *bfqq)
 
 	bfq_put_cooperator(bfqq);
 
-	/* remove bfqq from woken list */
-	if (!hlist_unhashed(&bfqq->woken_list_node))
-		hlist_del_init(&bfqq->woken_list_node);
-
-	/* reset waker for all queues in woken list */
-	hlist_for_each_entry_safe(item, n, &bfqq->woken_list,
-				  woken_list_node) {
-		item->waker_bfqq = NULL;
-		bfq_clear_bfqq_has_waker(item);
-		hlist_del_init(&item->woken_list_node);
-	}
-
 	bfq_put_queue(bfqq); /* release process reference */
 }
 
@@ -5436,12 +5462,12 @@ static void bfq_insert_request(struct blk_mq_hw_ctx *hctx, struct request *rq,
 
 	spin_lock_irq(&bfqd->lock);
 	bfqq = bfq_init_rq(rq);
-	if (at_head || blk_rq_is_passthrough(rq)) {
+	if (!bfqq || at_head || blk_rq_is_passthrough(rq)) {
 		if (at_head)
 			list_add(&rq->queuelist, &bfqd->dispatch);
 		else
 			list_add_tail(&rq->queuelist, &bfqd->dispatch);
-	} else { /* bfqq is assumed to be non null here */
+	} else {
 		idle_timer_disabled = __bfq_insert_request(bfqd, rq);
 		/*
 		 * Update bfqq, because, if a queue merge has occurred
diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
index 391ac0503dc075..76d0f9de767bce 100644
--- a/drivers/ata/libata-scsi.c
+++ b/drivers/ata/libata-scsi.c
@@ -1786,6 +1786,21 @@ nothing_to_do:
 	return 1;
 }
 
+static bool ata_check_nblocks(struct scsi_cmnd *scmd, u32 n_blocks)
+{
+	struct request *rq = scmd->request;
+	u32 req_blocks;
+
+	if (!blk_rq_is_passthrough(rq))
+		return true;
+
+	req_blocks = blk_rq_bytes(rq) / scmd->device->sector_size;
+	if (n_blocks > req_blocks)
+		return false;
+
+	return true;
+}
+
 /**
  *	ata_scsi_rw_xlat - Translate SCSI r/w command into an ATA one
  *	@qc: Storage for translated ATA taskfile
@@ -1830,6 +1845,8 @@ static unsigned int ata_scsi_rw_xlat(struct ata_queued_cmd *qc)
 		scsi_10_lba_len(cdb, &block, &n_block);
 		if (cdb[1] & (1 << 3))
 			tf_flags |= ATA_TFLAG_FUA;
+		if (!ata_check_nblocks(scmd, n_block))
+			goto invalid_fld;
 		break;
 	case READ_6:
 	case WRITE_6:
@@ -1844,6 +1861,8 @@ static unsigned int ata_scsi_rw_xlat(struct ata_queued_cmd *qc)
 		 */
 		if (!n_block)
 			n_block = 256;
+		if (!ata_check_nblocks(scmd, n_block))
+			goto invalid_fld;
 		break;
 	case READ_16:
 	case WRITE_16:
@@ -1854,6 +1873,8 @@ static unsigned int ata_scsi_rw_xlat(struct ata_queued_cmd *qc)
 		scsi_16_lba_len(cdb, &block, &n_block);
 		if (cdb[1] & (1 << 3))
 			tf_flags |= ATA_TFLAG_FUA;
+		if (!ata_check_nblocks(scmd, n_block))
+			goto invalid_fld;
 		break;
 	default:
 		DPRINTK("no-byte command\n");
diff --git a/drivers/ata/libata-sff.c b/drivers/ata/libata-sff.c
index 10aa2788214279..4f115adb4ee83b 100644
--- a/drivers/ata/libata-sff.c
+++ b/drivers/ata/libata-sff.c
@@ -658,6 +658,10 @@ static void ata_pio_sector(struct ata_queued_cmd *qc)
 	unsigned int offset;
 	unsigned char *buf;
 
+	if (!qc->cursg) {
+		qc->curbytes = qc->nbytes;
+		return;
+	}
 	if (qc->curbytes == qc->nbytes - qc->sect_size)
 		ap->hsm_task_state = HSM_ST_LAST;
 
@@ -683,6 +687,8 @@ static void ata_pio_sector(struct ata_queued_cmd *qc)
 
 	if (qc->cursg_ofs == qc->cursg->length) {
 		qc->cursg = sg_next(qc->cursg);
+		if (!qc->cursg)
+			ap->hsm_task_state = HSM_ST_LAST;
 		qc->cursg_ofs = 0;
 	}
 }
diff --git a/drivers/ata/pata_rb532_cf.c b/drivers/ata/pata_rb532_cf.c
index 7c37f2ff09e416..deae466395de1a 100644
--- a/drivers/ata/pata_rb532_cf.c
+++ b/drivers/ata/pata_rb532_cf.c
@@ -158,7 +158,6 @@ static int rb532_pata_driver_probe(struct platform_device *pdev)
 static int rb532_pata_driver_remove(struct platform_device *pdev)
 {
 	struct ata_host *ah = platform_get_drvdata(pdev);
-	struct rb532_cf_info *info = ah->private_data;
 
 	ata_host_detach(ah);
 
diff --git a/drivers/block/aoe/aoedev.c b/drivers/block/aoe/aoedev.c
index 5b49f1b33ebec4..e2ea2356da0610 100644
--- a/drivers/block/aoe/aoedev.c
+++ b/drivers/block/aoe/aoedev.c
@@ -323,10 +323,14 @@ flush(const char __user *str, size_t cnt, int exiting)
 	}
 
 	flush_scheduled_work();
-	/* pass one: without sleeping, do aoedev_downdev */
+	/* pass one: do aoedev_downdev, which might sleep */
+restart1:
 	spin_lock_irqsave(&devlist_lock, flags);
 	for (d = devlist; d; d = d->next) {
 		spin_lock(&d->lock);
+		if (d->flags & DEVFL_TKILL)
+			goto cont;
+
 		if (exiting) {
 			/* unconditionally take each device down */
 		} else if (specified) {
@@ -338,8 +342,11 @@ flush(const char __user *str, size_t cnt, int exiting)
 		|| d->ref)
 			goto cont;
 
+		spin_unlock(&d->lock);
+		spin_unlock_irqrestore(&devlist_lock, flags);
 		aoedev_downdev(d);
 		d->flags |= DEVFL_TKILL;
+		goto restart1;
 cont:
 		spin_unlock(&d->lock);
 	}
@@ -348,7 +355,7 @@ cont:
 	/* pass two: call freedev, which might sleep,
 	 * for aoedevs marked with DEVFL_TKILL
 	 */
-restart:
+restart2:
 	spin_lock_irqsave(&devlist_lock, flags);
 	for (d = devlist; d; d = d->next) {
 		spin_lock(&d->lock);
@@ -357,7 +364,7 @@ restart:
 			spin_unlock(&d->lock);
 			spin_unlock_irqrestore(&devlist_lock, flags);
 			freedev(d);
-			goto restart;
+			goto restart2;
 		}
 		spin_unlock(&d->lock);
 	}
diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 3036883fc9f878..ab7ca5989097ac 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -885,7 +885,7 @@ static void loop_unprepare_queue(struct loop_device *lo)
 
 static int loop_kthread_worker_fn(void *worker_ptr)
 {
-	current->flags |= PF_LESS_THROTTLE;
+	current->flags |= PF_LESS_THROTTLE | PF_MEMALLOC_NOIO;
 	return kthread_worker_fn(worker_ptr);
 }
 
diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c
index 9f082671284505..e2059af9079181 100644
--- a/drivers/md/bcache/sysfs.c
+++ b/drivers/md/bcache/sysfs.c
@@ -23,24 +23,28 @@ static const char * const bch_cache_modes[] = {
 	"writethrough",
 	"writeback",
 	"writearound",
-	"none"
+	"none",
+	NULL
 };
 
 /* Default is 0 ("auto") */
 static const char * const bch_stop_on_failure_modes[] = {
 	"auto",
-	"always"
+	"always",
+	NULL
 };
 
 static const char * const cache_replacement_policies[] = {
 	"lru",
 	"fifo",
-	"random"
+	"random",
+	NULL
 };
 
 static const char * const error_actions[] = {
 	"unregister",
-	"panic"
+	"panic",
+	NULL
 };
 
 write_attribute(attach);
@@ -338,7 +342,7 @@ STORE(__cached_dev)
 	}
 
 	if (attr == &sysfs_cache_mode) {
-		v = sysfs_match_string(bch_cache_modes, buf);
+		v = __sysfs_match_string(bch_cache_modes, -1, buf);
 		if (v < 0)
 			return v;
 
@@ -349,7 +353,7 @@ STORE(__cached_dev)
 	}
 
 	if (attr == &sysfs_stop_when_cache_set_failed) {
-		v = sysfs_match_string(bch_stop_on_failure_modes, buf);
+		v = __sysfs_match_string(bch_stop_on_failure_modes, -1, buf);
 		if (v < 0)
 			return v;
 
@@ -816,7 +820,7 @@ STORE(__bch_cache_set)
 			    0, UINT_MAX);
 
 	if (attr == &sysfs_errors) {
-		v = sysfs_match_string(error_actions, buf);
+		v = __sysfs_match_string(error_actions, -1, buf);
 		if (v < 0)
 			return v;
 
@@ -1088,7 +1092,7 @@ STORE(__bch_cache)
 	}
 
 	if (attr == &sysfs_cache_replacement_policy) {
-		v = sysfs_match_string(cache_replacement_policies, buf);
+		v = __sysfs_match_string(cache_replacement_policies, -1, buf);
 		if (v < 0)
 			return v;
 
diff --git a/fs/block_dev.c b/fs/block_dev.c
index a6f7c892cb4a39..eb657ab9406062 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -349,7 +349,7 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
 	loff_t pos = iocb->ki_pos;
 	blk_qc_t qc = BLK_QC_T_NONE;
 	gfp_t gfp;
-	ssize_t ret;
+	int ret;
 
 	if ((pos | iov_iter_alignment(iter)) &
 	    (bdev_logical_block_size(bdev) - 1))
@@ -386,8 +386,6 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
 
 	ret = 0;
 	for (;;) {
-		int err;
-
 		bio_set_dev(bio, bdev);
 		bio->bi_iter.bi_sector = pos >> 9;
 		bio->bi_write_hint = iocb->ki_hint;
@@ -395,10 +393,8 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
 		bio->bi_end_io = blkdev_bio_end_io;
 		bio->bi_ioprio = iocb->ki_ioprio;
 
-		err = bio_iov_iter_get_pages(bio, iter);
-		if (unlikely(err)) {
-			if (!ret)
-				ret = err;
+		ret = bio_iov_iter_get_pages(bio, iter);
+		if (unlikely(ret)) {
 			bio->bi_status = BLK_STS_IOERR;
 			bio_endio(bio);
 			break;
@@ -421,7 +417,6 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
 		if (nowait)
 			bio->bi_opf |= (REQ_NOWAIT | REQ_NOWAIT_INLINE);
 
-		dio->size += bio->bi_iter.bi_size;
 		pos += bio->bi_iter.bi_size;
 
 		nr_pages = iov_iter_npages(iter, BIO_MAX_PAGES);
@@ -433,13 +428,13 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
 				polled = true;
 			}
 
+			dio->size += bio->bi_iter.bi_size;
 			qc = submit_bio(bio);
 			if (qc == BLK_QC_T_EAGAIN) {
-				if (!ret)
-					ret = -EAGAIN;
+				dio->size -= bio->bi_iter.bi_size;
+				ret = -EAGAIN;
 				goto error;
 			}
-			ret = dio->size;
 
 			if (polled)
 				WRITE_ONCE(iocb->ki_cookie, qc);
@@ -460,18 +455,17 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
 			atomic_inc(&dio->ref);
 		}
 
+		dio->size += bio->bi_iter.bi_size;
 		qc = submit_bio(bio);
 		if (qc == BLK_QC_T_EAGAIN) {
-			if (!ret)
-				ret = -EAGAIN;
+			dio->size -= bio->bi_iter.bi_size;
+			ret = -EAGAIN;
 			goto error;
 		}
-		ret = dio->size;
 
 		bio = bio_alloc(gfp, nr_pages);
 		if (!bio) {
-			if (!ret)
-				ret = -EAGAIN;
+			ret = -EAGAIN;
 			goto error;
 		}
 	}
@@ -496,6 +490,8 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
 out:
 	if (!ret)
 		ret = blk_status_to_errno(dio->bio.bi_status);
+	if (likely(!ret))
+		ret = dio->size;
 
 	bio_put(&dio->bio);
 	return ret;
@@ -1754,7 +1750,10 @@ int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder)
 
 		/* finish claiming */
 		mutex_lock(&bdev->bd_mutex);
-		bd_finish_claiming(bdev, whole, holder);
+		if (!res)
+			bd_finish_claiming(bdev, whole, holder);
+		else
+			bd_abort_claiming(bdev, whole, holder);
 		/*
 		 * Block event polling for write claims if requested.  Any
 		 * write holder makes the write_holder state stick until