Re: [PATCH 6/6] nbd: reduce the nbd_index_mutex scope

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

 



On 8/11/21 8:44 AM, Christoph Hellwig wrote:
nbd_index_mutex is currently held over add_disk and inside ->open, which
leads to lock order reversals.  Refactor the device creation code path
so that nbd_dev_add is called without nbd_index_mutex lock held and
only takes it for the IDR insertation.

Signed-off-by: Christoph Hellwig <hch@xxxxxx>
---
  drivers/block/nbd.c | 55 +++++++++++++++++++++++----------------------
  1 file changed, 28 insertions(+), 27 deletions(-)

diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index 0b46a608f879..4054cc33fc1e 100644
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -1681,7 +1681,7 @@ static const struct blk_mq_ops nbd_mq_ops = {
  	.timeout	= nbd_xmit_timeout,
  };
-static struct nbd_device *nbd_dev_add(int index)
+static struct nbd_device *nbd_dev_add(int index, unsigned int refs)
  {
  	struct nbd_device *nbd;
  	struct gendisk *disk;
@@ -1707,6 +1707,7 @@ static struct nbd_device *nbd_dev_add(int index)
  	if (err)
  		goto out_free_nbd;
+ mutex_lock(&nbd_index_mutex);
  	if (index >= 0) {
  		err = idr_alloc(&nbd_index_idr, nbd, index, index + 1,
  				GFP_KERNEL);
@@ -1717,6 +1718,7 @@ static struct nbd_device *nbd_dev_add(int index)
  		if (err >= 0)
  			index = err;
  	}
+	mutex_unlock(&nbd_index_mutex);
  	if (err < 0)
  		goto out_free_tags;
  	nbd->index = index;
@@ -1743,7 +1745,7 @@ static struct nbd_device *nbd_dev_add(int index)
mutex_init(&nbd->config_lock);
  	refcount_set(&nbd->config_refs, 0);
-	refcount_set(&nbd->refs, 1);
+	refcount_set(&nbd->refs, refs);
  	INIT_LIST_HEAD(&nbd->list);
  	disk->major = NBD_MAJOR;
  	disk->first_minor = index << part_shift;
@@ -1847,34 +1849,35 @@ static int nbd_genl_connect(struct sk_buff *skb, struct genl_info *info)
  		nbd = idr_find(&nbd_index_idr, index);
  	}
- if (!nbd) {
-		nbd = nbd_dev_add(index);
-		if (IS_ERR(nbd)) {
+	if (nbd) {
+		if (test_bit(NBD_DESTROY_ON_DISCONNECT, &nbd->flags) &&
+		    test_bit(NBD_DISCONNECT_REQUESTED, &nbd->flags)) {
+			nbd->destroy_complete = &destroy_complete;
  			mutex_unlock(&nbd_index_mutex);
-			pr_err("nbd: failed to add new device\n");
-			return PTR_ERR(nbd);
+
+			/* wait until the nbd device is completely destroyed */
+			wait_for_completion(&destroy_complete);
+			goto again;
  		}
-	}
- if (test_bit(NBD_DESTROY_ON_DISCONNECT, &nbd->flags) &&
-	    test_bit(NBD_DISCONNECT_REQUESTED, &nbd->flags)) {
-		nbd->destroy_complete = &destroy_complete;
+		if (!refcount_inc_not_zero(&nbd->refs)) {
+			mutex_unlock(&nbd_index_mutex);
+			if (index == -1)
+				goto again;
+			pr_err("nbd: device at index %d is going down\n",
+		       		index);

Errant whitespace here.  Thanks,

Josef



[Index of Archives]     [Linux RAID]     [Linux SCSI]     [Linux ATA RAID]     [IDE]     [Linux Wireless]     [Linux Kernel]     [ATH6KL]     [Linux Bluetooth]     [Linux Netdev]     [Kernel Newbies]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Device Mapper]

  Powered by Linux