Re: [PATCH 03/15] block: add an API to atomically update queue limits

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 1/22/24 18:36, Christoph Hellwig wrote:
Add a new queue_limits_{start,commit}_update pair of functions that
allows taking an atomic snapshot of queue limits, update it, and
commit it if it passes validity checking.

Signed-off-by: Christoph Hellwig <hch@xxxxxx>
---
  block/blk-core.c       |   1 +
  block/blk-settings.c   | 140 +++++++++++++++++++++++++++++++++++++++++
  block/blk.h            |   1 +
  include/linux/blkdev.h |  23 +++++++
  4 files changed, 165 insertions(+)

diff --git a/block/blk-core.c b/block/blk-core.c
index 11342af420d0c4..09f4a44a4aa3cc 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -424,6 +424,7 @@ struct request_queue *blk_alloc_queue(int node_id)
  	mutex_init(&q->debugfs_mutex);
  	mutex_init(&q->sysfs_lock);
  	mutex_init(&q->sysfs_dir_lock);
+	mutex_init(&q->limits_lock);
  	mutex_init(&q->rq_qos_mutex);
  	spin_lock_init(&q->queue_lock);
diff --git a/block/blk-settings.c b/block/blk-settings.c
index e872b0e168525e..b6692143ccf034 100644
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -96,6 +96,146 @@ static void blk_apply_bdi_limits(struct backing_dev_info *bdi,
  	bdi->io_pages = lim->max_sectors >> (PAGE_SHIFT - 9);
  }
+static int blk_validate_zoned_limits(struct queue_limits *lim)
+{
+	if (!lim->zoned) {
+		if (WARN_ON_ONCE(lim->max_open_zones) ||
+		    WARN_ON_ONCE(lim->max_active_zones) ||
+		    WARN_ON_ONCE(lim->zone_write_granularity) ||
+		    WARN_ON_ONCE(lim->max_zone_append_sectors))
+			return -EINVAL;
+		return 0;
+	}
+
+	if (WARN_ON_ONCE(!IS_ENABLED(CONFIG_BLK_DEV_ZONED)))
+		return -EINVAL;
+
+	if (lim->zone_write_granularity < lim->logical_block_size)
+		lim->zone_write_granularity = lim->logical_block_size;
+
+	if (lim->max_zone_append_sectors) {
+		if (WARN_ON_ONCE(!lim->chunk_sectors))
+			return -EINVAL;
+		lim->max_zone_append_sectors =
+			min3(lim->max_hw_sectors,
+			     lim->max_zone_append_sectors,
+			     lim->chunk_sectors);
+	}
+
+	return 0;
+}
+
+int blk_validate_limits(struct queue_limits *lim)
+{
+	unsigned int max_hw_sectors;
+
+	if (!lim->logical_block_size)
+		lim->logical_block_size = SECTOR_SIZE;
+
+	if (!lim->physical_block_size)
+		lim->physical_block_size = SECTOR_SIZE;
+	if (lim->physical_block_size < lim->logical_block_size)
+		lim->physical_block_size = lim->physical_block_size;
+
+	if (!lim->io_min)
+		lim->io_min = SECTOR_SIZE;
+	if (lim->io_min < lim->physical_block_size)
+		lim->io_min = lim->physical_block_size;
+
+	if (!lim->max_hw_sectors)
+		lim->max_hw_sectors = BLK_SAFE_MAX_SECTORS;
+	if (WARN_ON_ONCE(lim->max_hw_sectors < PAGE_SIZE / SECTOR_SIZE))
+		return -EINVAL;
+
+	lim->max_hw_sectors = round_down(lim->max_hw_sectors,
+			lim->logical_block_size >> SECTOR_SHIFT);
+
+	max_hw_sectors = min_not_zero(lim->max_hw_sectors,
+				lim->max_dev_sectors);
+	if (lim->max_user_sectors) {
+		if (lim->max_user_sectors > max_hw_sectors ||
+		    lim->max_user_sectors < PAGE_SIZE / SECTOR_SIZE)
+			return -EINVAL;
+		lim->max_sectors = min(max_hw_sectors, lim->max_user_sectors);
+	} else {
+		lim->max_sectors = min(max_hw_sectors, BLK_DEF_MAX_SECTORS_CAP);
+	}
+
+	lim->max_sectors = round_down(lim->max_sectors,
+			lim->logical_block_size >> SECTOR_SHIFT);
+
+	if (!lim->max_segments)
+		lim->max_segments = BLK_MAX_SEGMENTS;
+
+	lim->max_discard_sectors = lim->max_hw_discard_sectors;
+	if (!lim->max_discard_segments)
+		lim->max_discard_segments = 1;
+
+	if (lim->discard_granularity < lim->physical_block_size)
+		lim->discard_granularity = lim->physical_block_size;
+
+	if (!lim->seg_boundary_mask)
+		lim->seg_boundary_mask = BLK_SEG_BOUNDARY_MASK;
+	if (WARN_ON_ONCE(lim->seg_boundary_mask < PAGE_SIZE - 1))
+		return -EINVAL;
+
+	if (!lim->max_segment_size)
+		lim->max_segment_size = BLK_MAX_SEGMENT_SIZE;
+	if (WARN_ON_ONCE(lim->max_segment_size < PAGE_SIZE))
+		return -EINVAL;
+
+	/*
+	 * Devices that require a virtual boundary do not support scatter/gather
+	 * I/O natively, but instead require a descriptor list entry for each
+	 * page (which might not be idential to the Linux PAGE_SIZE).  Because
+	 * of that they are not limited by our notion of "segment size".
+	 */
+	if (lim->virt_boundary_mask) {
+		if (WARN_ON_ONCE(lim->max_segment_size &&
+				 lim->max_segment_size != UINT_MAX))
+			return -EINVAL;
+		lim->max_segment_size = UINT_MAX;
+	}
+
+	if (!lim->dma_alignment)
+		lim->dma_alignment = SECTOR_SIZE - 1;
+	if (WARN_ON_ONCE(lim->dma_alignment > PAGE_SIZE))
+		return -EINVAL;
+
+	if (lim->alignment_offset) {
+		lim->alignment_offset &= (lim->physical_block_size - 1);
+		lim->misaligned = 0;
+	}
+
+	return blk_validate_zoned_limits(lim);
+}
+
+/**
+ * queue_limits_commit_update - commit an atomic update of queue limits
+ * @q:		queue to update
+ * @lim:	limits to apply
+ *
+ * Apply the limits in @lim that were obtained from queue_limits_start_update()
+ * and updated by the caller to @q.
+ *
+ * Returns 0 if successful, else a negative error code.
+ */
+int queue_limits_commit_update(struct request_queue *q,
+		struct queue_limits *lim)
+	__releases(q->limits_lock)
+{
+	int error = blk_validate_limits(lim);
+
+	if (!error) {
+		q->limits = *lim;
+		if (q->disk)
+			blk_apply_bdi_limits(q->disk->bdi, lim);
+	}
+	mutex_unlock(&q->limits_lock);
+	return error;
+}
+EXPORT_SYMBOL_GPL(queue_limits_commit_update);
+
  /**
   * blk_queue_bounce_limit - set bounce buffer limit for queue
   * @q: the request queue for the device
diff --git a/block/blk.h b/block/blk.h
index 1ef920f72e0f87..58b5dbac2a487d 100644
--- a/block/blk.h
+++ b/block/blk.h
@@ -447,6 +447,7 @@ static inline void bio_release_page(struct bio *bio, struct page *page)
  		unpin_user_page(page);
  }
+int blk_validate_limits(struct queue_limits *lim);
  struct request_queue *blk_alloc_queue(int node_id);
int disk_scan_partitions(struct gendisk *disk, blk_mode_t mode);
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 4a2e82c7971c86..5b5d3b238de1e7 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -473,6 +473,7 @@ struct request_queue {
struct mutex sysfs_lock;
  	struct mutex		sysfs_dir_lock;
+	struct mutex		limits_lock;
/*
  	 * for reusing dead hctx instance in case of updating
@@ -861,6 +862,28 @@ static inline unsigned int blk_chunk_sectors_left(sector_t offset,
  	return chunk_sectors - (offset & (chunk_sectors - 1));
  }
+/**
+ * queue_limits_start_update - start an atomic update of queue limits
+ * @q:		queue to update
+ *
+ * This functions starts an atomic update of the queue limits.  It takes a lock
+ * to prevent other updates and returns a snapshot of the current limits that
+ * the caller can modify.  The caller must call queue_limits_commit_update()
+ * to finish the update.
+ *
+ * Context: process context.  The caller must have frozen the queue or ensured
+ * that there is outstanding I/O by other means.
+ */
+static inline struct queue_limits
+queue_limits_start_update(struct request_queue *q)
+	__acquires(q->limits_lock)
+{
+	mutex_lock(&q->limits_lock);
+	return q->limits;
+}

I'm slightly confused about the lifetime of the returned structure.
By my understanding, the returned 'struct queue_limit' is allocated
on the stack of the caller, right?
Shouldn't we note this somewhere such that people don't start passing
the structure around, or, worse, calling 'kfree()' on it?

Cheers,

Hannes
--
Dr. Hannes Reinecke                Kernel Storage Architect
hare@xxxxxxx                              +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), GF: Ivo Totev, Andrew McDonald,
Werner Knoblich





[Index of Archives]     [KVM Development]     [Libvirt Development]     [Libvirt Users]     [CentOS Virtualization]     [Netdev]     [Ethernet Bridging]     [Linux Wireless]     [Kernel Newbies]     [Security]     [Linux for Hams]     [Netfilter]     [Bugtraq]     [Yosemite Forum]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux Admin]     [Samba]

  Powered by Linux