Linux-BTRFS Archive on lore.kernel.org
 help / color / Atom feed
From: Johannes Thumshirn <jth@kernel.org>
To: David Sterba <dsterba@suse.com>
Cc: Nikolay Borisov <nborisov@suse.com>, Qu Wenruo <wqu@suse.com>,
	Linux BTRFS Mailinglist <linux-btrfs@vger.kernel.org>,
	Johannes Thumshirn <jthumshirn@suse.de>
Subject: [PATCH v5 2/2] btrfs: reset device back to allocation state when removing
Date: Wed,  4 Dec 2019 14:36:39 +0100
Message-ID: <20191204133639.2382-3-jth@kernel.org> (raw)
In-Reply-To: <20191204133639.2382-1-jth@kernel.org>

From: Johannes Thumshirn <jthumshirn@suse.de>

When closing a device, btrfs_close_one_device() first allocates a new
device, copies the device to close's name, replaces it in the dev_list
with the copy and then finally frees it.

This involves two memory allocation, which can potentially fail. As this
code path is tricky to unwind, the allocation failures where handled by
BUG_ON()s.

But this copying isn't strictly needed, all that is needed is resetting
the device in question to it's state it had after the allocation.

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>

---
Changes to v4:
- Clear dev_stat_ccnt on removal (Dave)
- Don't clear BTRFS_DEV_STATE_MISSING and BTRFS_DEV_STATE_FS_METADATA as
  they'll be handled elsewhere
- Release extent_io_tree (fstests)

Changes to v3:
- Clear DEV_STATE_WRITABLE _after_ btrfs_close_bdev() (Nik)
---
 fs/btrfs/volumes.c | 28 +++++++++++-----------------
 1 file changed, 11 insertions(+), 17 deletions(-)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index ae3980ba3a87..eef0b9ed9ea4 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -1066,8 +1066,6 @@ static void btrfs_close_bdev(struct btrfs_device *device)
 static void btrfs_close_one_device(struct btrfs_device *device)
 {
 	struct btrfs_fs_devices *fs_devices = device->fs_devices;
-	struct btrfs_device *new_device;
-	struct rcu_string *name;
 
 	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state) &&
 	    device->devid != BTRFS_DEV_REPLACE_DEVID) {
@@ -1079,25 +1077,21 @@ static void btrfs_close_one_device(struct btrfs_device *device)
 		fs_devices->missing_devices--;
 
 	btrfs_close_bdev(device);
-	if (device->bdev)
+	if (device->bdev) {
 		fs_devices->open_devices--;
-
-	new_device = btrfs_alloc_device(NULL, &device->devid,
-					device->uuid);
-	BUG_ON(IS_ERR(new_device)); /* -ENOMEM */
-
-	/* Safe because we are under uuid_mutex */
-	if (device->name) {
-		name = rcu_string_strdup(device->name->str, GFP_NOFS);
-		BUG_ON(!name); /* -ENOMEM */
-		rcu_assign_pointer(new_device->name, name);
+		device->bdev = NULL;
 	}
+	clear_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 
-	list_replace_rcu(&device->dev_list, &new_device->dev_list);
-	new_device->fs_devices = device->fs_devices;
+	atomic_set(&device->dev_stats_ccnt, 0);
+	extent_io_tree_release(&device->alloc_state);
 
-	synchronize_rcu();
-	btrfs_free_device(device);
+	/* Verify the device is back in a pristine state  */
+	ASSERT(!test_bit(BTRFS_DEV_STATE_FLUSH_SENT, &device->dev_state));
+	ASSERT(!test_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state));
+	ASSERT(list_empty(&device->dev_alloc_list));
+	ASSERT(list_empty(&device->post_commit_list));
+	ASSERT(atomic_read(&device->reada_in_flight) == 0);
 }
 
 static int close_fs_devices(struct btrfs_fs_devices *fs_devices)
-- 
2.20.1 (Apple Git-117)


      parent reply index

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-04 13:36 [PATCH v5 0/2] remove BUG_ON()s in btrfs_close_one_device() Johannes Thumshirn
2019-12-04 13:36 ` [PATCH v5 1/2] btrfs: decrement number of open devices after closing the device not before Johannes Thumshirn
2019-12-04 13:36 ` Johannes Thumshirn [this message]

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20191204133639.2382-3-jth@kernel.org \
    --to=jth@kernel.org \
    --cc=dsterba@suse.com \
    --cc=jthumshirn@suse.de \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=nborisov@suse.com \
    --cc=wqu@suse.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-BTRFS Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-btrfs/0 linux-btrfs/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-btrfs linux-btrfs/ https://lore.kernel.org/linux-btrfs \
		linux-btrfs@vger.kernel.org
	public-inbox-index linux-btrfs

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-btrfs


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git