From: Christoph Hellwig <hch@lst.de> To: Jens Axboe <axboe@kernel.dk>, Alasdair Kergon <agk@redhat.com>, Mike Snitzer <snitzer@kernel.org> Cc: Yu Kuai <yukuai1@huaweicloud.com>, dm-devel@redhat.com, linux-block@vger.kernel.org Subject: [PATCH 5/7] dm: track per-add_disk holder relations in DM Date: Sun, 30 Oct 2022 16:31:17 +0100 [thread overview] Message-ID: <20221030153120.1045101-6-hch@lst.de> (raw) In-Reply-To: <20221030153120.1045101-1-hch@lst.de> dm is a bit special in that it opens the underlying devices. Commit 89f871af1b26 ("dm: delay registering the gendisk") tried to accomodate that by allowing to add the holder to the list before add_gendisk and then just add them to sysfs once add_disk is called. But that leads to really odd lifetime problems and error handling problems as we can't know the state of the kobjects and don't unwind properly. To fix this switch to just registering all existing table_devices with the holder code right after add_disk, and remove them before calling del_gendisk. Fixes: 89f871af1b26 ("dm: delay registering the gendisk") Reported-by: Yu Kuai <yukuai1@huaweicloud.com> Signed-off-by: Christoph Hellwig <hch@lst.de> --- drivers/md/dm.c | 45 +++++++++++++++++++++++++++++++++++++-------- 1 file changed, 37 insertions(+), 8 deletions(-) diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 2917700b1e15c..7b0d6dc957549 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -751,9 +751,16 @@ static struct table_device *open_table_device(struct mapped_device *md, goto out_free_td; } - r = bd_link_disk_holder(bdev, dm_disk(md)); - if (r) - goto out_blkdev_put; + /* + * We can be called before the dm disk is added. In that case we can't + * register the holder relation here. It will be done once add_disk was + * called. + */ + if (md->disk->slave_dir) { + r = bd_link_disk_holder(bdev, md->disk); + if (r) + goto out_blkdev_put; + } td->dm_dev.mode = mode; td->dm_dev.bdev = bdev; @@ -774,7 +781,8 @@ static struct table_device *open_table_device(struct mapped_device *md, */ static void close_table_device(struct table_device *td, struct mapped_device *md) { - bd_unlink_disk_holder(td->dm_dev.bdev, dm_disk(md)); + if (md->disk->slave_dir) + bd_unlink_disk_holder(td->dm_dev.bdev, md->disk); blkdev_put(td->dm_dev.bdev, td->dm_dev.mode | FMODE_EXCL); put_dax(td->dm_dev.dax_dev); list_del(&td->list); @@ -1951,7 +1959,13 @@ static void cleanup_mapped_device(struct mapped_device *md) md->disk->private_data = NULL; spin_unlock(&_minor_lock); if (dm_get_md_type(md) != DM_TYPE_NONE) { + struct table_device *td; + dm_sysfs_exit(md); + list_for_each_entry(td, &md->table_devices, list) { + bd_unlink_disk_holder(td->dm_dev.bdev, + md->disk); + } del_gendisk(md->disk); } dm_queue_destroy_crypto_profile(md->queue); @@ -2284,6 +2298,7 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t) { enum dm_queue_mode type = dm_table_get_type(t); struct queue_limits limits; + struct table_device *td; int r; switch (type) { @@ -2316,13 +2331,27 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t) if (r) return r; - r = dm_sysfs_init(md); - if (r) { - del_gendisk(md->disk); - return r; + /* + * Register the holder relationship for devices added before the disk + * was live. + */ + list_for_each_entry(td, &md->table_devices, list) { + r = bd_link_disk_holder(td->dm_dev.bdev, md->disk); + if (r) + goto out_undo_holders; } + + r = dm_sysfs_init(md); + if (r) + goto out_undo_holders; md->type = type; return 0; + +out_undo_holders: + list_for_each_entry_continue_reverse(td, &md->table_devices, list) + bd_unlink_disk_holder(td->dm_dev.bdev, md->disk); + del_gendisk(md->disk); + return r; } struct mapped_device *dm_get_md(dev_t dev) -- 2.30.2
WARNING: multiple messages have this Message-ID (diff)
From: Christoph Hellwig <hch@lst.de> To: Jens Axboe <axboe@kernel.dk>, Alasdair Kergon <agk@redhat.com>, Mike Snitzer <snitzer@kernel.org> Cc: linux-block@vger.kernel.org, Yu Kuai <yukuai1@huaweicloud.com>, dm-devel@redhat.com Subject: [dm-devel] [PATCH 5/7] dm: track per-add_disk holder relations in DM Date: Sun, 30 Oct 2022 16:31:17 +0100 [thread overview] Message-ID: <20221030153120.1045101-6-hch@lst.de> (raw) In-Reply-To: <20221030153120.1045101-1-hch@lst.de> dm is a bit special in that it opens the underlying devices. Commit 89f871af1b26 ("dm: delay registering the gendisk") tried to accomodate that by allowing to add the holder to the list before add_gendisk and then just add them to sysfs once add_disk is called. But that leads to really odd lifetime problems and error handling problems as we can't know the state of the kobjects and don't unwind properly. To fix this switch to just registering all existing table_devices with the holder code right after add_disk, and remove them before calling del_gendisk. Fixes: 89f871af1b26 ("dm: delay registering the gendisk") Reported-by: Yu Kuai <yukuai1@huaweicloud.com> Signed-off-by: Christoph Hellwig <hch@lst.de> --- drivers/md/dm.c | 45 +++++++++++++++++++++++++++++++++++++-------- 1 file changed, 37 insertions(+), 8 deletions(-) diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 2917700b1e15c..7b0d6dc957549 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -751,9 +751,16 @@ static struct table_device *open_table_device(struct mapped_device *md, goto out_free_td; } - r = bd_link_disk_holder(bdev, dm_disk(md)); - if (r) - goto out_blkdev_put; + /* + * We can be called before the dm disk is added. In that case we can't + * register the holder relation here. It will be done once add_disk was + * called. + */ + if (md->disk->slave_dir) { + r = bd_link_disk_holder(bdev, md->disk); + if (r) + goto out_blkdev_put; + } td->dm_dev.mode = mode; td->dm_dev.bdev = bdev; @@ -774,7 +781,8 @@ static struct table_device *open_table_device(struct mapped_device *md, */ static void close_table_device(struct table_device *td, struct mapped_device *md) { - bd_unlink_disk_holder(td->dm_dev.bdev, dm_disk(md)); + if (md->disk->slave_dir) + bd_unlink_disk_holder(td->dm_dev.bdev, md->disk); blkdev_put(td->dm_dev.bdev, td->dm_dev.mode | FMODE_EXCL); put_dax(td->dm_dev.dax_dev); list_del(&td->list); @@ -1951,7 +1959,13 @@ static void cleanup_mapped_device(struct mapped_device *md) md->disk->private_data = NULL; spin_unlock(&_minor_lock); if (dm_get_md_type(md) != DM_TYPE_NONE) { + struct table_device *td; + dm_sysfs_exit(md); + list_for_each_entry(td, &md->table_devices, list) { + bd_unlink_disk_holder(td->dm_dev.bdev, + md->disk); + } del_gendisk(md->disk); } dm_queue_destroy_crypto_profile(md->queue); @@ -2284,6 +2298,7 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t) { enum dm_queue_mode type = dm_table_get_type(t); struct queue_limits limits; + struct table_device *td; int r; switch (type) { @@ -2316,13 +2331,27 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t) if (r) return r; - r = dm_sysfs_init(md); - if (r) { - del_gendisk(md->disk); - return r; + /* + * Register the holder relationship for devices added before the disk + * was live. + */ + list_for_each_entry(td, &md->table_devices, list) { + r = bd_link_disk_holder(td->dm_dev.bdev, md->disk); + if (r) + goto out_undo_holders; } + + r = dm_sysfs_init(md); + if (r) + goto out_undo_holders; md->type = type; return 0; + +out_undo_holders: + list_for_each_entry_continue_reverse(td, &md->table_devices, list) + bd_unlink_disk_holder(td->dm_dev.bdev, md->disk); + del_gendisk(md->disk); + return r; } struct mapped_device *dm_get_md(dev_t dev) -- 2.30.2 -- dm-devel mailing list dm-devel@redhat.com https://listman.redhat.com/mailman/listinfo/dm-devel
next prev parent reply other threads:[~2022-10-30 15:31 UTC|newest] Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top 2022-10-30 15:31 fix delayed holder tracking v2 Christoph Hellwig 2022-10-30 15:31 ` [dm-devel] " Christoph Hellwig 2022-10-30 15:31 ` [PATCH 1/7] block: clear ->slave_dir when dropping the main slave_dir reference Christoph Hellwig 2022-10-30 15:31 ` [dm-devel] " Christoph Hellwig 2022-10-30 15:31 ` [PATCH 2/7] dm: remove free_table_devices Christoph Hellwig 2022-10-30 15:31 ` [dm-devel] " Christoph Hellwig 2022-10-30 15:31 ` [PATCH 3/7] dm: cleanup open_table_device Christoph Hellwig 2022-10-30 15:31 ` [dm-devel] " Christoph Hellwig 2022-10-30 15:31 ` [PATCH 4/7] dm: cleanup close_table_device Christoph Hellwig 2022-10-30 15:31 ` [dm-devel] " Christoph Hellwig 2022-10-30 15:31 ` Christoph Hellwig [this message] 2022-10-30 15:31 ` [dm-devel] [PATCH 5/7] dm: track per-add_disk holder relations in DM Christoph Hellwig 2022-11-09 2:08 ` Yu Kuai 2022-11-09 2:08 ` [dm-devel] " Yu Kuai 2022-11-09 8:26 ` Christoph Hellwig 2022-11-09 8:26 ` [dm-devel] " Christoph Hellwig 2022-11-10 18:09 ` Mike Snitzer 2022-11-10 18:09 ` [dm-devel] " Mike Snitzer 2022-11-10 19:48 ` Mike Snitzer 2022-11-10 19:48 ` Mike Snitzer 2022-11-12 6:23 ` Yu Kuai 2022-11-12 6:23 ` [dm-devel] " Yu Kuai 2022-10-30 15:31 ` [PATCH 6/7] block: remove delayed holder registration Christoph Hellwig 2022-10-30 15:31 ` [dm-devel] " Christoph Hellwig 2022-10-30 15:31 ` [PATCH 7/7] block: store the holder kobject in bd_holder_disk Christoph Hellwig 2022-10-30 15:31 ` [dm-devel] " Christoph Hellwig 2022-10-31 1:52 ` Yu Kuai 2022-10-31 1:52 ` [dm-devel] " Yu Kuai 2022-11-01 10:49 ` Christoph Hellwig 2022-11-01 10:49 ` [dm-devel] " Christoph Hellwig 2022-11-01 11:12 ` Yu Kuai 2022-11-01 11:12 ` [dm-devel] " Yu Kuai 2022-11-01 11:21 ` Christoph Hellwig 2022-11-01 11:21 ` [dm-devel] " Christoph Hellwig 2022-11-01 11:28 ` Yu Kuai 2022-11-01 11:28 ` [dm-devel] " Yu Kuai 2022-11-01 13:18 ` Christoph Hellwig 2022-11-01 13:18 ` [dm-devel] " Christoph Hellwig 2022-11-01 13:29 ` Yu Kuai 2022-11-01 13:29 ` [dm-devel] " Yu Kuai
Reply instructions: You may reply publicly 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=20221030153120.1045101-6-hch@lst.de \ --to=hch@lst.de \ --cc=agk@redhat.com \ --cc=axboe@kernel.dk \ --cc=dm-devel@redhat.com \ --cc=linux-block@vger.kernel.org \ --cc=snitzer@kernel.org \ --cc=yukuai1@huaweicloud.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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.