[RFC 00/18] drbd: part 4 of adding multiple volume support to drbd

[RFC 00/18] drbd: part 4 of adding multiple volume support to drbd

[Philipp Reisner]

This the first request for review of drbd-8.4. The complete set has 
492 patches. This is the fourth installment containing 18 patches.

The whole set is available here:
  git://git.drbd.org/linux-2.6-drbd.git for-jens

and is jens_for-3.2_drivers...for-jens
and this part is cb3bd33...5d96817

The most noticeable change is the support for multiple replicated volumes in
a single DRBD connection.  Write-ordering is obeyed among all writes in all
volumes in a single connection.  This feature is really important for users
who use DRBD for mirroring over longer distances. (Protocol A).

[PATCH 01/18] drbd: default to detach on-io-error

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Old default behaviour was "pass-on",
which is not useful in production at all.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 include/linux/drbd_limits.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/include/linux/drbd_limits.h b/include/linux/drbd_limits.h
index 75f05af..22920a8 100644
--- a/include/linux/drbd_limits.h
+++ b/include/linux/drbd_limits.h
@@ -125,7 +125,7 @@
 #define DRBD_DISK_SIZE_SECT_MAX  (1 * (2LLU << 40))
 #define DRBD_DISK_SIZE_SECT_DEF  0 /* = disabled = no user size... */
 
-#define DRBD_ON_IO_ERROR_DEF EP_PASS_ON
+#define DRBD_ON_IO_ERROR_DEF EP_DETACH
 #define DRBD_FENCING_DEF FP_DONT_CARE
 #define DRBD_AFTER_SB_0P_DEF ASB_DISCONNECT
 #define DRBD_AFTER_SB_1P_DEF ASB_DISCONNECT
-- 
1.7.4.1

[PATCH 02/18] drbd: only wakeup if something changed in update_peer_seq

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

This commit got it wrong:
    drbd: Make the peer_seq updating code more obvious

    Make it more clear that update_peer_seq() is supposed to wake up the
    seq_wait queue whenever the sequence number changes.

We don't need to wake up everytime we receive a sequence number
that is _different_ from our currently stored "newest" sequence number,
but only if we receive a sequence number _newer_ than what we already
have, when we actually change mdev->peer_seq.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_receiver.c |    9 +++++----
 1 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index 620f27f..64f658b 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -1737,14 +1737,15 @@ static bool need_peer_seq(struct drbd_conf *mdev)
 
 static void update_peer_seq(struct drbd_conf *mdev, unsigned int peer_seq)
 {
-	unsigned int old_peer_seq;
+	unsigned int newest_peer_seq;
 
 	if (need_peer_seq(mdev)) {
 		spin_lock(&mdev->peer_seq_lock);
-		old_peer_seq = mdev->peer_seq;
-		mdev->peer_seq = seq_max(mdev->peer_seq, peer_seq);
+		newest_peer_seq = seq_max(mdev->peer_seq, peer_seq);
+		mdev->peer_seq = newest_peer_seq;
 		spin_unlock(&mdev->peer_seq_lock);
-		if (old_peer_seq != peer_seq)
+		/* wake up only if we actually changed mdev->peer_seq */
+		if (peer_seq == newest_peer_seq)
 			wake_up(&mdev->seq_wait);
 	}
 }
-- 
1.7.4.1

[PATCH 03/18] drbd: add page pool to be used for meta data IO

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_int.h  |   23 ++++++++++++++++++++++-
 drivers/block/drbd/drbd_main.c |    9 +++++++++
 2 files changed, 31 insertions(+), 1 deletions(-)

diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index e3f542c..4debab7 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -1463,11 +1463,32 @@ extern struct kmem_cache *drbd_al_ext_cache;	/* activity log extents */
 extern mempool_t *drbd_request_mempool;
 extern mempool_t *drbd_ee_mempool;
 
-extern struct page *drbd_pp_pool; /* drbd's page pool */
+/* drbd's page pool, used to buffer data received from the peer,
+ * or data requested by the peer.
+ *
+ * This does not have an emergency reserve.
+ *
+ * When allocating from this pool, it first takes pages from the pool.
+ * Only if the pool is depleted will try to allocate from the system.
+ *
+ * The assumption is that pages taken from this pool will be processed,
+ * and given back, "quickly", and then can be recycled, so we can avoid
+ * frequent calls to alloc_page(), and still will be able to make progress even
+ * under memory pressure.
+ */
+extern struct page *drbd_pp_pool;
 extern spinlock_t   drbd_pp_lock;
 extern int	    drbd_pp_vacant;
 extern wait_queue_head_t drbd_pp_wait;
 
+/* We also need a standard (emergency-reserve backed) page pool
+ * for meta data IO (activity log, bitmap).
+ * We can keep it global, as long as it is used as "N pages at a time".
+ * 128 should be plenty, currently we probably can get away with as few as 1.
+ */
+#define DRBD_MIN_POOL_PAGES	128
+extern mempool_t *drbd_md_io_page_pool;
+
 extern rwlock_t global_state_lock;
 
 extern int conn_lowest_minor(struct drbd_tconn *tconn);
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 6d761cb..cb1636c 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -129,6 +129,7 @@ struct kmem_cache *drbd_bm_ext_cache;	/* bitmap extents */
 struct kmem_cache *drbd_al_ext_cache;	/* activity log extents */
 mempool_t *drbd_request_mempool;
 mempool_t *drbd_ee_mempool;
+mempool_t *drbd_md_io_page_pool;
 
 /* I do not use a standard mempool, because:
    1) I want to hand out the pre-allocated objects first.
@@ -1952,6 +1953,8 @@ static void drbd_destroy_mempools(void)
 
 	/* D_ASSERT(atomic_read(&drbd_pp_vacant)==0); */
 
+	if (drbd_md_io_page_pool)
+		mempool_destroy(drbd_md_io_page_pool);
 	if (drbd_ee_mempool)
 		mempool_destroy(drbd_ee_mempool);
 	if (drbd_request_mempool)
@@ -1965,6 +1968,7 @@ static void drbd_destroy_mempools(void)
 	if (drbd_al_ext_cache)
 		kmem_cache_destroy(drbd_al_ext_cache);
 
+	drbd_md_io_page_pool = NULL;
 	drbd_ee_mempool      = NULL;
 	drbd_request_mempool = NULL;
 	drbd_ee_cache        = NULL;
@@ -1988,6 +1992,7 @@ static int drbd_create_mempools(void)
 	drbd_bm_ext_cache    = NULL;
 	drbd_al_ext_cache    = NULL;
 	drbd_pp_pool         = NULL;
+	drbd_md_io_page_pool = NULL;
 
 	/* caches */
 	drbd_request_cache = kmem_cache_create(
@@ -2011,6 +2016,10 @@ static int drbd_create_mempools(void)
 		goto Enomem;
 
 	/* mempools */
+	drbd_md_io_page_pool = mempool_create_page_pool(DRBD_MIN_POOL_PAGES, 0);
+	if (drbd_md_io_page_pool == NULL)
+		goto Enomem;
+
 	drbd_request_mempool = mempool_create(number,
 		mempool_alloc_slab, mempool_free_slab, drbd_request_cache);
 	if (drbd_request_mempool == NULL)
-- 
1.7.4.1

[PATCH 04/18] drbd: use the newly introduced page pool for bitmap IO

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_bitmap.c |    9 ++++-----
 1 files changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/block/drbd/drbd_bitmap.c b/drivers/block/drbd/drbd_bitmap.c
index 2fc3dfa..412ca94 100644
--- a/drivers/block/drbd/drbd_bitmap.c
+++ b/drivers/block/drbd/drbd_bitmap.c
@@ -957,9 +957,8 @@ static void bm_async_io_complete(struct bio *bio, int error)
 
 	bm_page_unlock_io(mdev, idx);
 
-	/* FIXME give back to page pool */
 	if (ctx->flags & BM_AIO_COPY_PAGES)
-		put_page(bio->bi_io_vec[0].bv_page);
+		mempool_free(bio->bi_io_vec[0].bv_page, drbd_md_io_page_pool);
 
 	bio_put(bio);
 
@@ -993,10 +992,8 @@ static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must
 	bm_set_page_unchanged(b->bm_pages[page_nr]);
 
 	if (ctx->flags & BM_AIO_COPY_PAGES) {
-		/* FIXME alloc_page is good enough for now, but actually needs
-		 * to use pre-allocated page pool */
 		void *src, *dest;
-		page = alloc_page(__GFP_HIGHMEM|__GFP_WAIT);
+		page = mempool_alloc(drbd_md_io_page_pool, __GFP_HIGHMEM|__GFP_WAIT);
 		dest = kmap_atomic(page, KM_USER0);
 		src = kmap_atomic(b->bm_pages[page_nr], KM_USER1);
 		memcpy(dest, src, PAGE_SIZE);
@@ -1008,6 +1005,8 @@ static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must
 
 	bio->bi_bdev = mdev->ldev->md_bdev;
 	bio->bi_sector = on_disk_sector;
+	/* bio_add_page of a single page to an empty bio will always succeed,
+	 * according to api.  Do we want to assert that? */
 	bio_add_page(bio, page, len, 0);
 	bio->bi_private = ctx;
 	bio->bi_end_io = bm_async_io_complete;
-- 
1.7.4.1

[PATCH 05/18] drbd: introduce a bio_set to allocate housekeeping bios from

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Don't rely on availability of bios from the global fs_bio_set,
we should use our own bio_set for meta data IO.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_actlog.c   |    2 +-
 drivers/block/drbd/drbd_bitmap.c   |    3 +--
 drivers/block/drbd/drbd_int.h      |    6 ++++++
 drivers/block/drbd/drbd_main.c     |   28 ++++++++++++++++++++++++++++
 drivers/block/drbd/drbd_receiver.c |    6 +++++-
 5 files changed, 41 insertions(+), 4 deletions(-)

diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c
index 55f5a7a..9ab6365 100644
--- a/drivers/block/drbd/drbd_actlog.c
+++ b/drivers/block/drbd/drbd_actlog.c
@@ -125,7 +125,7 @@ static int _drbd_md_sync_page_io(struct drbd_conf *mdev,
 		rw |= REQ_FUA | REQ_FLUSH;
 	rw |= REQ_SYNC;
 
-	bio = bio_alloc(GFP_NOIO, 1);
+	bio = bio_alloc_drbd(GFP_NOIO);
 	bio->bi_bdev = bdev->md_bdev;
 	bio->bi_sector = sector;
 	ok = (bio_add_page(bio, page, size, 0) == size);
diff --git a/drivers/block/drbd/drbd_bitmap.c b/drivers/block/drbd/drbd_bitmap.c
index 412ca94..a1d72db 100644
--- a/drivers/block/drbd/drbd_bitmap.c
+++ b/drivers/block/drbd/drbd_bitmap.c
@@ -968,8 +968,7 @@ static void bm_async_io_complete(struct bio *bio, int error)
 
 static void bm_page_io_async(struct bm_aio_ctx *ctx, int page_nr, int rw) __must_hold(local)
 {
-	/* we are process context. we always get a bio */
-	struct bio *bio = bio_alloc(GFP_KERNEL, 1);
+	struct bio *bio = bio_alloc_drbd(GFP_KERNEL);
 	struct drbd_conf *mdev = ctx->mdev;
 	struct drbd_bitmap *b = mdev->bitmap;
 	struct page *page;
diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index 4debab7..5d90b9d 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -1489,6 +1489,12 @@ extern wait_queue_head_t drbd_pp_wait;
 #define DRBD_MIN_POOL_PAGES	128
 extern mempool_t *drbd_md_io_page_pool;
 
+/* We also need to make sure we get a bio
+ * when we need it for housekeeping purposes */
+extern struct bio_set *drbd_md_io_bio_set;
+/* to allocate from that set */
+extern struct bio *bio_alloc_drbd(gfp_t gfp_mask);
+
 extern rwlock_t global_state_lock;
 
 extern int conn_lowest_minor(struct drbd_tconn *tconn);
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index cb1636c..26bcd13 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -130,6 +130,7 @@ struct kmem_cache *drbd_al_ext_cache;	/* activity log extents */
 mempool_t *drbd_request_mempool;
 mempool_t *drbd_ee_mempool;
 mempool_t *drbd_md_io_page_pool;
+struct bio_set *drbd_md_io_bio_set;
 
 /* I do not use a standard mempool, because:
    1) I want to hand out the pre-allocated objects first.
@@ -150,6 +151,25 @@ static const struct block_device_operations drbd_ops = {
 	.release = drbd_release,
 };
 
+static void bio_destructor_drbd(struct bio *bio)
+{
+	bio_free(bio, drbd_md_io_bio_set);
+}
+
+struct bio *bio_alloc_drbd(gfp_t gfp_mask)
+{
+	struct bio *bio;
+
+	if (!drbd_md_io_bio_set)
+		return bio_alloc(gfp_mask, 1);
+
+	bio = bio_alloc_bioset(gfp_mask, 1, drbd_md_io_bio_set);
+	if (!bio)
+		return NULL;
+	bio->bi_destructor = bio_destructor_drbd;
+	return bio;
+}
+
 #ifdef __CHECKER__
 /* When checking with sparse, and this is an inline function, sparse will
    give tons of false positives. When this is a real functions sparse works.
@@ -1953,6 +1973,8 @@ static void drbd_destroy_mempools(void)
 
 	/* D_ASSERT(atomic_read(&drbd_pp_vacant)==0); */
 
+	if (drbd_md_io_bio_set)
+		bioset_free(drbd_md_io_bio_set);
 	if (drbd_md_io_page_pool)
 		mempool_destroy(drbd_md_io_page_pool);
 	if (drbd_ee_mempool)
@@ -1968,6 +1990,7 @@ static void drbd_destroy_mempools(void)
 	if (drbd_al_ext_cache)
 		kmem_cache_destroy(drbd_al_ext_cache);
 
+	drbd_md_io_bio_set   = NULL;
 	drbd_md_io_page_pool = NULL;
 	drbd_ee_mempool      = NULL;
 	drbd_request_mempool = NULL;
@@ -1993,6 +2016,7 @@ static int drbd_create_mempools(void)
 	drbd_al_ext_cache    = NULL;
 	drbd_pp_pool         = NULL;
 	drbd_md_io_page_pool = NULL;
+	drbd_md_io_bio_set   = NULL;
 
 	/* caches */
 	drbd_request_cache = kmem_cache_create(
@@ -2016,6 +2040,10 @@ static int drbd_create_mempools(void)
 		goto Enomem;
 
 	/* mempools */
+	drbd_md_io_bio_set = bioset_create(DRBD_MIN_POOL_PAGES, 0);
+	if (drbd_md_io_bio_set == NULL)
+		goto Enomem;
+
 	drbd_md_io_page_pool = mempool_create_page_pool(DRBD_MIN_POOL_PAGES, 0);
 	if (drbd_md_io_page_pool == NULL)
 		goto Enomem;
diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index 64f658b..12b1533 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -1127,7 +1127,11 @@ int drbd_submit_peer_request(struct drbd_conf *mdev,
 	/* In most cases, we will only need one bio.  But in case the lower
 	 * level restrictions happen to be different at this offset on this
 	 * side than those of the sending peer, we may need to submit the
-	 * request in more than one bio. */
+	 * request in more than one bio.
+	 *
+	 * Plain bio_alloc is good enough here, this is no DRBD internally
+	 * generated bio, but a bio allocated on behalf of the peer.
+	 */
 next_bio:
 	bio = bio_alloc(GFP_NOIO, nr_pages);
 	if (!bio) {
-- 
1.7.4.1

[PATCH 06/18] drbd: fix drbd_delete_device: remove vnr from volumes; idr_remove(); synchronize_rcu(); before cleanup

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Still missing: rcu_readlock() on the various call sites that
access/iterate over those idrs.

We don't need a specific write lock, as we only modify from
configuration context, which is already strictly serialized.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_main.c |   42 ++++++++++++++++++++++-----------------
 1 files changed, 24 insertions(+), 18 deletions(-)

diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 26bcd13..21914f4 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -2124,7 +2124,9 @@ void drbd_delete_device(unsigned int minor)
 	if (!mdev)
 		return;
 
-	idr_remove(&mdev->tconn->volumes, minor);
+	idr_remove(&mdev->tconn->volumes, mdev->vnr);
+	idr_remove(&minors, minor);
+	synchronize_rcu();
 
 	/* paranoia asserts */
 	D_ASSERT(mdev->open_cnt == 0);
@@ -2153,7 +2155,6 @@ void drbd_delete_device(unsigned int minor)
 	 * allocated from drbd_new_device
 	 * and actually free the mdev itself */
 	drbd_free_mdev(mdev);
-	idr_remove(&minors, minor);
 }
 
 static void drbd_cleanup(void)
@@ -2331,15 +2332,6 @@ enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor,
 		return ERR_NOMEM;
 
 	mdev->tconn = tconn;
-	if (!idr_pre_get(&tconn->volumes, GFP_KERNEL))
-		goto out_no_idr;
-	if (idr_get_new(&tconn->volumes, mdev, &vnr_got))
-		goto out_no_idr;
-	if (vnr_got != vnr) {
-		dev_err(DEV, "vnr_got (%d) != vnr (%d)\n", vnr_got, vnr);
-		goto out_no_q;
-	}
-
 	mdev->minor = minor;
 
 	drbd_init_set_defaults(mdev);
@@ -2395,19 +2387,35 @@ enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor,
 	INIT_LIST_HEAD(&mdev->current_epoch->list);
 	mdev->epochs = 1;
 
+	if (!idr_pre_get(&tconn->volumes, GFP_KERNEL))
+		goto out_no_vol_idr;
+	if (idr_get_new(&tconn->volumes, mdev, &vnr_got))
+		goto out_no_vol_idr;
+	if (vnr_got != vnr) {
+		dev_err(DEV, "vnr_got (%d) != vnr (%d)\n", vnr_got, vnr);
+		goto out_idr_remove_vol;
+	}
+
 	if (!idr_pre_get(&minors, GFP_KERNEL))
-		goto out_no_minor_idr;
+		goto out_idr_remove_vol;
 	if (idr_get_new(&minors, mdev, &minor_got))
-		goto out_no_minor_idr;
+		goto out_idr_remove_vol;
 	if (minor_got != minor) {
-		idr_remove(&minors, minor_got);
-		goto out_no_minor_idr;
+		/* minor exists, or other idr strangeness? */
+		dev_err(DEV, "available minor (%d) != requested minor (%d)\n",
+				minor_got, minor);
+		goto out_idr_remove_minor;
 	}
 	add_disk(disk);
 
 	return NO_ERROR;
 
-out_no_minor_idr:
+out_idr_remove_minor:
+	idr_remove(&minors, minor_got);
+out_idr_remove_vol:
+	idr_remove(&tconn->volumes, vnr_got);
+	synchronize_rcu();
+out_no_vol_idr:
 	kfree(mdev->current_epoch);
 out_no_epoch:
 	drbd_bm_cleanup(mdev);
@@ -2418,8 +2426,6 @@ out_no_io_page:
 out_no_disk:
 	blk_cleanup_queue(q);
 out_no_q:
-	idr_remove(&tconn->volumes, vnr_got);
-out_no_idr:
 	kfree(mdev);
 	return ERR_NOMEM;
 }
-- 
1.7.4.1

[PATCH 07/18] drbd: get rid of drbd_bcast_ee, it is of no use anymore

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

This function was used to broadcast the (leading part of the)
bio payload in case we see a data integrity error.  It could be received
from userland with the drbdsetup events subcommand,
to have a peek into the payload that caused the checksum mismatch,
and guess from there what may have caused the mismatch,
mainly to guess wether it was modification of in-flight data,
or data corruption by broken hardware or software bugs.

Meanwhile we support bios that are larger than the maximum payload a
netlink datagram can carry.
And we have means to reliably detect modification of in-flight data by
calculating, and comparing, the checksum before and after sendmsg.
There is no need to carry this around anymore.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_receiver.c |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index 12b1533..6dfea2f 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -1356,8 +1356,6 @@ read_in_block(struct drbd_conf *mdev, u64 id, sector_t sector,
 		if (memcmp(dig_in, dig_vv, dgs)) {
 			dev_err(DEV, "Digest integrity check FAILED: %llus +%u\n",
 				(unsigned long long)sector, data_size);
-			drbd_bcast_ee(mdev, "digest failed",
-					dgs, dig_in, dig_vv, peer_req);
 			drbd_free_ee(mdev, peer_req);
 			return NULL;
 		}
-- 
1.7.4.1

[PATCH 08/18] drbd: prepare the transition from connector to genetlink

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

This adds the new API header and helper files.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 include/linux/drbd_genl.h         |  349 +++++++++++++++++++++++++++++++
 include/linux/drbd_genl_api.h     |   55 +++++
 include/linux/genl_magic_func.h   |  417 +++++++++++++++++++++++++++++++++++++
 include/linux/genl_magic_struct.h |  260 +++++++++++++++++++++++
 4 files changed, 1081 insertions(+), 0 deletions(-)
 create mode 100644 include/linux/drbd_genl.h
 create mode 100644 include/linux/drbd_genl_api.h
 create mode 100644 include/linux/genl_magic_func.h
 create mode 100644 include/linux/genl_magic_struct.h

diff --git a/include/linux/drbd_genl.h b/include/linux/drbd_genl.h
new file mode 100644
index 0000000..84e1684
--- /dev/null
+++ b/include/linux/drbd_genl.h
@@ -0,0 +1,349 @@
+/*
+ * General overview:
+ * full generic netlink message:
+ * |nlmsghdr|genlmsghdr|<payload>
+ *
+ * payload:
+ * |optional fixed size family header|<sequence of netlink attributes>
+ *
+ * sequence of netlink attributes:
+ * I chose to have all "top level" attributes NLA_NESTED,
+ * corresponding to some real struct.
+ * So we have a sequence of |tla, len|<nested nla sequence>
+ *
+ * nested nla sequence:
+ * may be empty, or contain a sequence of netlink attributes
+ * representing the struct fields.
+ *
+ * The tag number of any field (regardless of containing struct)
+ * will be available as T_ ## field_name,
+ * so you cannot have the same field name in two differnt structs.
+ *
+ * The tag numbers themselves are per struct, though,
+ * so should always begin at 1 (not 0, that is the special "NLA_UNSPEC" type,
+ * which we won't use here).
+ * The tag numbers are used as index in the respective nla_policy array.
+ *
+ * GENL_struct(tag_name, tag_number, struct name, struct fields) - struct and policy
+ *	genl_magic_struct.h
+ *		generates the struct declaration,
+ *		generates an entry in the tla enum,
+ *	genl_magic_func.h
+ *		generates an entry in the static tla policy
+ *		with .type = NLA_NESTED
+ *		generates the static <struct_name>_nl_policy definition,
+ *		and static conversion functions
+ *
+ *	genl_magic_func.h
+ *
+ * GENL_mc_group(group)
+ *	genl_magic_struct.h
+ *		does nothing
+ *	genl_magic_func.h
+ *		defines and registers the mcast group,
+ *		and provides a send helper
+ *
+ * GENL_notification(op_name, op_num, mcast_group, tla list)
+ *	These are notifications to userspace.
+ *
+ *	genl_magic_struct.h
+ *		generates an entry in the genl_ops enum,
+ *	genl_magic_func.h
+ *		does nothing
+ *
+ *	mcast group: the name of the mcast group this notification should be
+ *	expected on
+ *	tla list: the list of expected top level attributes,
+ *	for documentation and sanity checking.
+ *
+ * GENL_op(op_name, op_num, flags and handler, tla list) - "genl operations"
+ *	These are requests from userspace.
+ *
+ *	_op and _notification share the same "number space",
+ *	op_nr will be assigned to "genlmsghdr->cmd"
+ *
+ *	genl_magic_struct.h
+ *		generates an entry in the genl_ops enum,
+ *	genl_magic_func.h
+ *		generates an entry in the static genl_ops array,
+ *		and static register/unregister functions to
+ *		genl_register_family_with_ops().
+ *
+ *	flags and handler:
+ *		GENL_op_init( .doit = x, .dumpit = y, .flags = something)
+ *		GENL_doit(x) => .dumpit = NULL, .flags = GENL_ADMIN_PERM
+ *	tla list: the list of expected top level attributes,
+ *	for documentation and sanity checking.
+ */
+
+/*
+ * STRUCTS
+ */
+
+/* this is sent kernel -> userland on various error conditions, and contains
+ * informational textual info, which is supposedly human readable.
+ * The computer relevant return code is in the drbd_genlmsghdr.
+ */
+GENL_struct(DRBD_NLA_CFG_REPLY, 1, drbd_cfg_reply,
+		/* "arbitrary" size strings, nla_policy.len = 0 */
+	__str_field(1, GENLA_F_MANDATORY,	info_text, 0)
+)
+
+/* Configuration requests typically need a context to operate on.
+ * Possible keys are device minor (fits in the drbd_genlmsghdr),
+ * the replication link (aka connection) name,
+ * and/or the replication group (aka resource) name,
+ * and the volume id within the resource. */
+GENL_struct(DRBD_NLA_CFG_CONTEXT, 2, drbd_cfg_context,
+		/* currently only 256 volumes per group,
+		 * but maybe we still change that */
+	__u32_field(1, GENLA_F_MANDATORY,	ctx_volume)
+	__str_field(2, GENLA_F_MANDATORY,	ctx_conn_name, 128)
+)
+
+GENL_struct(DRBD_NLA_DISK_CONF, 3, disk_conf,
+	__u64_field(1, GENLA_F_MANDATORY,	disk_size)
+	__str_field(2, GENLA_F_REQUIRED,	backing_dev,	128)
+	__str_field(3, GENLA_F_REQUIRED,	meta_dev,	128)
+	__u32_field(4, GENLA_F_REQUIRED,	meta_dev_idx)
+	__u32_field(5, GENLA_F_MANDATORY,	max_bio_bvecs)
+	__u32_field(6, GENLA_F_MANDATORY,	on_io_error)
+	__u32_field(7, GENLA_F_MANDATORY,	fencing)
+	__flg_field(8, GENLA_F_MANDATORY,	no_disk_barrier)
+	__flg_field(9, GENLA_F_MANDATORY,	no_disk_flush)
+	__flg_field(10, GENLA_F_MANDATORY,	no_disk_drain)
+	__flg_field(11, GENLA_F_MANDATORY,	no_md_flush)
+	__flg_field(12, GENLA_F_MANDATORY,	use_bmbv)
+)
+
+GENL_struct(DRBD_NLA_SYNCER_CONF, 4, syncer_conf,
+	__u32_field(1,	GENLA_F_MANDATORY,	rate)
+	__u32_field(2,	GENLA_F_MANDATORY,	after)
+	__u32_field(3,	GENLA_F_MANDATORY,	al_extents)
+	__str_field(4,	GENLA_F_MANDATORY,	cpu_mask,       32)
+	__str_field(5,	GENLA_F_MANDATORY,	verify_alg,     SHARED_SECRET_MAX)
+	__str_field(6,	GENLA_F_MANDATORY,	csums_alg,	SHARED_SECRET_MAX)
+	__flg_field(7,	GENLA_F_MANDATORY,	use_rle)
+	__u32_field(8,	GENLA_F_MANDATORY,	on_no_data)
+	__u32_field(9,	GENLA_F_MANDATORY,	c_plan_ahead)
+	__u32_field(10,	GENLA_F_MANDATORY,	c_delay_target)
+	__u32_field(11,	GENLA_F_MANDATORY,	c_fill_target)
+	__u32_field(12,	GENLA_F_MANDATORY,	c_max_rate)
+	__u32_field(13,	GENLA_F_MANDATORY,	c_min_rate)
+)
+
+GENL_struct(DRBD_NLA_NET_CONF, 5, net_conf,
+	__str_field(1,	GENLA_F_MANDATORY | GENLA_F_SENSITIVE,
+						shared_secret,	SHARED_SECRET_MAX)
+	__str_field(2,	GENLA_F_MANDATORY,	cram_hmac_alg,	SHARED_SECRET_MAX)
+	__str_field(3,	GENLA_F_MANDATORY,	integrity_alg,	SHARED_SECRET_MAX)
+	__str_field(4,	GENLA_F_REQUIRED,	my_addr,	128)
+	__str_field(5,	GENLA_F_REQUIRED,	peer_addr,	128)
+	__u32_field(6,	GENLA_F_REQUIRED,	wire_protocol)
+	__u32_field(7,	GENLA_F_MANDATORY,	try_connect_int)
+	__u32_field(8,	GENLA_F_MANDATORY,	timeout)
+	__u32_field(9,	GENLA_F_MANDATORY,	ping_int)
+	__u32_field(10,	GENLA_F_MANDATORY,	ping_timeo)
+	__u32_field(11,	GENLA_F_MANDATORY,	sndbuf_size)
+	__u32_field(12,	GENLA_F_MANDATORY,	rcvbuf_size)
+	__u32_field(13,	GENLA_F_MANDATORY,	ko_count)
+	__u32_field(14,	GENLA_F_MANDATORY,	max_buffers)
+	__u32_field(15,	GENLA_F_MANDATORY,	max_epoch_size)
+	__u32_field(16,	GENLA_F_MANDATORY,	unplug_watermark)
+	__u32_field(17,	GENLA_F_MANDATORY,	after_sb_0p)
+	__u32_field(18,	GENLA_F_MANDATORY,	after_sb_1p)
+	__u32_field(19,	GENLA_F_MANDATORY,	after_sb_2p)
+	__u32_field(20,	GENLA_F_MANDATORY,	rr_conflict)
+	__u32_field(21,	GENLA_F_MANDATORY,	on_congestion)
+	__u32_field(22,	GENLA_F_MANDATORY,	cong_fill)
+	__u32_field(23,	GENLA_F_MANDATORY,	cong_extents)
+	__flg_field(24, GENLA_F_MANDATORY,	two_primaries)
+	__flg_field(25, GENLA_F_MANDATORY,	want_lose)
+	__flg_field(26, GENLA_F_MANDATORY,	no_cork)
+	__flg_field(27, GENLA_F_MANDATORY,	always_asbp)
+	__flg_field(28, GENLA_F_MANDATORY,	dry_run)
+)
+
+GENL_struct(DRBD_NLA_SET_ROLE_PARMS, 6, set_role_parms,
+	__flg_field(1, GENLA_F_MANDATORY,	assume_uptodate)
+)
+
+GENL_struct(DRBD_NLA_RESIZE_PARMS, 7, resize_parms,
+	__u64_field(1, GENLA_F_MANDATORY,	resize_size)
+	__flg_field(2, GENLA_F_MANDATORY,	resize_force)
+	__flg_field(3, GENLA_F_MANDATORY,	no_resync)
+)
+
+GENL_struct(DRBD_NLA_STATE_INFO, 8, state_info,
+	/* the reason of the broadcast,
+	 * if this is an event triggered broadcast. */
+	__u32_field(1, GENLA_F_MANDATORY,	sib_reason)
+	__u32_field(2, GENLA_F_REQUIRED,	current_state)
+	__u64_field(3, GENLA_F_MANDATORY,	capacity)
+	__u64_field(4, GENLA_F_MANDATORY,	ed_uuid)
+
+	/* These are for broadcast from after state change work.
+	 * prev_state and new_state are from the moment the state change took
+	 * place, new_state is not neccessarily the same as current_state,
+	 * there may have been more state changes since.  Which will be
+	 * broadcasted soon, in their respective after state change work.  */
+	__u32_field(5, GENLA_F_MANDATORY,	prev_state)
+	__u32_field(6, GENLA_F_MANDATORY,	new_state)
+
+	/* if we have a local disk: */
+	__bin_field(7, GENLA_F_MANDATORY,	uuids, (UI_SIZE*sizeof(__u64)))
+	__u32_field(8, GENLA_F_MANDATORY,	disk_flags)
+	__u64_field(9, GENLA_F_MANDATORY,	bits_total)
+	__u64_field(10, GENLA_F_MANDATORY,	bits_oos)
+	/* and in case resync or online verify is active */
+	__u64_field(11, GENLA_F_MANDATORY,	bits_rs_total)
+	__u64_field(12, GENLA_F_MANDATORY,	bits_rs_failed)
+
+	/* for pre and post notifications of helper execution */
+	__str_field(13, GENLA_F_MANDATORY,	helper, 32)
+	__u32_field(14, GENLA_F_MANDATORY,	helper_exit_code)
+)
+
+GENL_struct(DRBD_NLA_START_OV_PARMS, 9, start_ov_parms,
+	__u64_field(1, GENLA_F_MANDATORY,	ov_start_sector)
+)
+
+GENL_struct(DRBD_NLA_NEW_C_UUID_PARMS, 10, new_c_uuid_parms,
+	__flg_field(1, GENLA_F_MANDATORY, clear_bm)
+)
+
+GENL_struct(DRBD_NLA_TIMEOUT_PARMS, 11, timeout_parms,
+	__u32_field(1,	GENLA_F_REQUIRED,	timeout_type)
+)
+
+GENL_struct(DRBD_NLA_DISCONNECT_PARMS, 12, disconnect_parms,
+	__flg_field(1, GENLA_F_MANDATORY,	force_disconnect)
+)
+
+/*
+ * Notifications and commands (genlmsghdr->cmd)
+ */
+GENL_mc_group(events)
+
+	/* kernel -> userspace announcement of changes */
+GENL_notification(
+	DRBD_EVENT, 1, events,
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_STATE_INFO, GENLA_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_NET_CONF, GENLA_F_MANDATORY)
+	GENL_tla_expected(DRBD_NLA_DISK_CONF, GENLA_F_MANDATORY)
+	GENL_tla_expected(DRBD_NLA_SYNCER_CONF, GENLA_F_MANDATORY)
+)
+
+	/* query kernel for specific or all info */
+GENL_op(
+	DRBD_ADM_GET_STATUS, 2,
+	GENL_op_init(
+		.doit = drbd_adm_get_status,
+		.dumpit = drbd_adm_get_status_all,
+		/* anyone may ask for the status,
+		 * it is broadcasted anyways */
+	),
+	/* To select the object .doit.
+	 * Or a subset of objects in .dumpit. */
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_MANDATORY)
+)
+
+#if 0
+	/* TO BE DONE */
+	/* create or destroy resources, aka replication groups */
+GENL_op(DRBD_ADM_CREATE_RESOURCE, 3, GENL_doit(drbd_adm_create_resource),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_DELETE_RESOURCE, 4, GENL_doit(drbd_adm_delete_resource),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+#endif
+
+	/* add DRBD minor devices as volumes to resources */
+GENL_op(DRBD_ADM_ADD_MINOR, 5, GENL_doit(drbd_adm_add_minor),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_DEL_MINOR, 6, GENL_doit(drbd_adm_delete_minor),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+
+	/* add or delete replication links to resources */
+GENL_op(DRBD_ADM_ADD_LINK, 7, GENL_doit(drbd_adm_create_connection),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_DEL_LINK, 8, GENL_doit(drbd_adm_delete_connection),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+
+	/* operates on replication links */
+GENL_op(DRBD_ADM_SYNCER, 9,
+	GENL_doit(drbd_adm_syncer),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_SYNCER_CONF, GENLA_F_MANDATORY)
+)
+
+GENL_op(
+	DRBD_ADM_CONNECT, 10,
+	GENL_doit(drbd_adm_connect),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_NET_CONF, GENLA_F_REQUIRED)
+)
+
+GENL_op(DRBD_ADM_DISCONNECT, 11, GENL_doit(drbd_adm_disconnect),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+
+	/* operates on minors */
+GENL_op(DRBD_ADM_ATTACH, 12,
+	GENL_doit(drbd_adm_attach),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_DISK_CONF, GENLA_F_REQUIRED)
+)
+
+GENL_op(
+	DRBD_ADM_RESIZE, 13,
+	GENL_doit(drbd_adm_resize),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_RESIZE_PARMS, GENLA_F_MANDATORY)
+)
+
+	/* operates on all volumes within a resource */
+GENL_op(
+	DRBD_ADM_PRIMARY, 14,
+	GENL_doit(drbd_adm_set_role),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_SET_ROLE_PARMS, GENLA_F_REQUIRED)
+)
+
+GENL_op(
+	DRBD_ADM_SECONDARY, 15,
+	GENL_doit(drbd_adm_set_role),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_SET_ROLE_PARMS, GENLA_F_REQUIRED)
+)
+
+GENL_op(
+	DRBD_ADM_NEW_C_UUID, 16,
+	GENL_doit(drbd_adm_new_c_uuid),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED)
+	GENL_tla_expected(DRBD_NLA_NEW_C_UUID_PARMS, GENLA_F_MANDATORY)
+)
+
+GENL_op(
+	DRBD_ADM_START_OV, 17,
+	GENL_doit(drbd_adm_start_ov),
+	GENL_tla_expected(DRBD_NLA_START_OV_PARMS, GENLA_F_MANDATORY)
+)
+
+GENL_op(DRBD_ADM_DETACH,	18, GENL_doit(drbd_adm_detach),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_INVALIDATE,	19, GENL_doit(drbd_adm_invalidate),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_INVAL_PEER,	20, GENL_doit(drbd_adm_invalidate_peer),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_PAUSE_SYNC,	21, GENL_doit(drbd_adm_pause_sync),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_RESUME_SYNC,	22, GENL_doit(drbd_adm_resume_sync),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_SUSPEND_IO,	23, GENL_doit(drbd_adm_suspend_io),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_RESUME_IO,	24, GENL_doit(drbd_adm_resume_io),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_OUTDATE,	25, GENL_doit(drbd_adm_outdate),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_GET_TIMEOUT_TYPE, 26, GENL_doit(drbd_adm_get_timeout_type),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
diff --git a/include/linux/drbd_genl_api.h b/include/linux/drbd_genl_api.h
new file mode 100644
index 0000000..9ef50d5
--- /dev/null
+++ b/include/linux/drbd_genl_api.h
@@ -0,0 +1,55 @@
+#ifndef DRBD_GENL_STRUCT_H
+#define DRBD_GENL_STRUCT_H
+
+/**
+ * struct drbd_genlmsghdr - DRBD specific header used in NETLINK_GENERIC requests
+ * @minor:
+ *     For admin requests (user -> kernel): which minor device to operate on.
+ *     For (unicast) replies or informational (broadcast) messages
+ *     (kernel -> user): which minor device the information is about.
+ *     If we do not operate on minors, but on connections or resources,
+ *     the minor value shall be (~0), and the attribute DRBD_NLA_CFG_CONTEXT
+ *     is used instead.
+ * @flags: possible operation modifiers (relevant only for user->kernel):
+ *     DRBD_GENL_F_SET_DEFAULTS
+ * @volume:
+ *     When creating a new minor (adding it to a resource), the resource needs
+ *     to know which volume number within the resource this is supposed to be.
+ *     The volume number corresponds to the same volume number on the remote side,
+ *     whereas the minor number on the remote side may be different
+ *     (union with flags).
+ * @ret_code: kernel->userland unicast cfg reply return code (union with flags);
+ */
+struct drbd_genlmsghdr {
+	__u32 minor;
+	union {
+	__u32 flags;
+	__s32 ret_code;
+	};
+};
+
+/* To be used in drbd_genlmsghdr.flags */
+enum {
+	DRBD_GENL_F_SET_DEFAULTS = 1,
+};
+
+enum drbd_state_info_bcast_reason {
+	SIB_GET_STATUS_REPLY = 1,
+	SIB_STATE_CHANGE = 2,
+	SIB_HELPER_PRE = 3,
+	SIB_HELPER_POST = 4,
+	SIB_SYNC_PROGRESS = 5,
+};
+
+/* hack around predefined gcc/cpp "linux=1",
+ * we cannot possibly include <1/drbd_genl.h> */
+#undef linux
+
+#include <linux/drbd.h>
+#define GENL_MAGIC_VERSION	API_VERSION
+#define GENL_MAGIC_FAMILY	drbd
+#define GENL_MAGIC_FAMILY_HDRSZ	sizeof(struct drbd_genlmsghdr)
+#define GENL_MAGIC_INCLUDE_FILE <linux/drbd_genl.h>
+#include <linux/genl_magic_struct.h>
+
+#endif
diff --git a/include/linux/genl_magic_func.h b/include/linux/genl_magic_func.h
new file mode 100644
index 0000000..8a86f65
--- /dev/null
+++ b/include/linux/genl_magic_func.h
@@ -0,0 +1,417 @@
+#ifndef GENL_MAGIC_FUNC_H
+#define GENL_MAGIC_FUNC_H
+
+#include <linux/genl_magic_struct.h>
+
+/*
+ * Extension of genl attribute validation policies			{{{1
+ *									{{{2
+ */
+
+/**
+ * nla_is_required - return true if this attribute is required
+ * @nla: netlink attribute
+ */
+static inline int nla_is_required(const struct nlattr *nla)
+{
+        return nla->nla_type & GENLA_F_REQUIRED;
+}
+
+/**
+ * nla_is_mandatory - return true if understanding this attribute is mandatory
+ * @nla: netlink attribute
+ * Note: REQUIRED attributes are implicitly MANDATORY as well
+ */
+static inline int nla_is_mandatory(const struct nlattr *nla)
+{
+        return nla->nla_type & (GENLA_F_MANDATORY | GENLA_F_REQUIRED);
+}
+
+/* Functionality to be integrated into nla_parse(), and validate_nla(),
+ * respectively.
+ *
+ * Enforcing the "mandatory" bit is done here,
+ * by rejecting unknown mandatory attributes.
+ *
+ * Part of enforcing the "required" flag would mean to embed it into
+ * nla_policy.type, and extending validate_nla(), which currently does
+ * BUG_ON(pt->type > NLA_TYPE_MAX); we have to work on existing kernels,
+ * so we cannot do that.  Thats why enforcing "required" is done in the
+ * generated assignment functions below. */
+static int nla_check_unknown(int maxtype, struct nlattr *head, int len)
+{
+	struct nlattr *nla;
+	int rem;
+        nla_for_each_attr(nla, head, len, rem) {
+		__u16 type = nla_type(nla);
+		if (type > maxtype && nla_is_mandatory(nla))
+			return -EOPNOTSUPP;
+	}
+	return 0;
+}
+
+/*
+ * Magic: declare tla policy						{{{1
+ * Magic: declare nested policies
+ *									{{{2
+ */
+#undef GENL_mc_group
+#define GENL_mc_group(group)
+
+#undef GENL_notification
+#define GENL_notification(op_name, op_num, mcast_group, tla_list)
+
+#undef GENL_op
+#define GENL_op(op_name, op_num, handler, tla_list)
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)		\
+	[tag_name] = { .type = NLA_NESTED },
+
+static struct nla_policy CONCAT_(GENL_MAGIC_FAMILY, _tla_nl_policy)[] = {
+#include GENL_MAGIC_INCLUDE_FILE
+};
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)		\
+static struct nla_policy s_name ## _nl_policy[] __read_mostly =		\
+{ s_fields };
+
+#undef __field
+#define __field(attr_nr, attr_flag, name, nla_type, _type, __get, __put) \
+	[__nla_type(attr_nr)] = { .type = nla_type },
+
+#undef __array
+#define __array(attr_nr, attr_flag, name, nla_type, _type, maxlen,	\
+		__get, __put)						\
+	[__nla_type(attr_nr)] = { .type = nla_type,			\
+		.len = maxlen - (nla_type == NLA_NUL_STRING) },
+
+#include GENL_MAGIC_INCLUDE_FILE
+
+#ifndef __KERNEL__
+#ifndef pr_info
+#define pr_info(args...)	fprintf(stderr, args);
+#endif
+#endif
+
+#if 1
+static void dprint_field(const char *dir, int nla_type,
+		const char *name, void *valp)
+{
+	__u64 val = valp ? *(__u32 *)valp : 1;
+	switch (nla_type) {
+	case NLA_U8:  val = (__u8)val;
+	case NLA_U16: val = (__u16)val;
+	case NLA_U32: val = (__u32)val;
+		pr_info("%s attr %s: %d 0x%08x\n", dir,
+			name, (int)val, (unsigned)val);
+		break;
+	case NLA_U64:
+		val = *(__u64*)valp;
+		pr_info("%s attr %s: %lld 0x%08llx\n", dir,
+			name, (long long)val, (unsigned long long)val);
+		break;
+	case NLA_FLAG:
+		if (val)
+			pr_info("%s attr %s: set\n", dir, name);
+		break;
+	}
+}
+
+static void dprint_array(const char *dir, int nla_type,
+		const char *name, const char *val, unsigned len)
+{
+	switch (nla_type) {
+	case NLA_NUL_STRING:
+		if (len && val[len-1] == '\0')
+			len--;
+		pr_info("%s attr %s: [len:%u] '%s'\n", dir, name, len, val);
+		break;
+	default:
+		/* we can always show 4 byte,
+		 * thats what nlattr are aligned to. */
+		pr_info("%s attr %s: [len:%u] %02x%02x%02x%02x ...\n",
+			dir, name, len, val[0], val[1], val[2], val[3]);
+	}
+}
+
+#define DPRINT_TLA(a, op, b) pr_info("%s %s %s\n", a, op, b);
+
+/* Name is a member field name of the struct s.
+ * If s is NULL (only parsing, no copy requested in *_from_attrs()),
+ * nla is supposed to point to the attribute containing the information
+ * corresponding to that struct member. */
+#define DPRINT_FIELD(dir, nla_type, name, s, nla)			\
+	do {								\
+		if (s)							\
+			dprint_field(dir, nla_type, #name, &s->name);	\
+		else if (nla)						\
+			dprint_field(dir, nla_type, #name,		\
+				(nla_type == NLA_FLAG) ? NULL		\
+						: nla_data(nla));	\
+	} while (0)
+
+#define	DPRINT_ARRAY(dir, nla_type, name, s, nla)			\
+	do {								\
+		if (s)							\
+			dprint_array(dir, nla_type, #name,		\
+					s->name, s->name ## _len);	\
+		else if (nla)						\
+			dprint_array(dir, nla_type, #name,		\
+					nla_data(nla), nla_len(nla));	\
+	} while (0)
+#else
+#define DPRINT_TLA(a, op, b) do {} while (0)
+#define DPRINT_FIELD(dir, nla_type, name, s, nla) do {} while (0)
+#define	DPRINT_ARRAY(dir, nla_type, name, s, nla) do {} while (0)
+#endif
+
+/*
+ * Magic: provide conversion functions					{{{1
+ * populate struct from attribute table:
+ *									{{{2
+ */
+
+/* processing of generic netlink messages is serialized.
+ * use one static buffer for parsing of nested attributes */
+static struct nlattr *nested_attr_tb[128];
+
+#ifndef BUILD_BUG_ON
+/* Force a compilation error if condition is true */
+#define BUILD_BUG_ON(condition) ((void)BUILD_BUG_ON_ZERO(condition))
+/* Force a compilation error if condition is true, but also produce a
+   result (of value 0 and type size_t), so the expression can be used
+   e.g. in a structure initializer (or where-ever else comma expressions
+   aren't permitted). */
+#define BUILD_BUG_ON_ZERO(e) (sizeof(struct { int:-!!(e); }))
+#define BUILD_BUG_ON_NULL(e) ((void *)sizeof(struct { int:-!!(e); }))
+#endif
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)		\
+	/* static, potentially unused */				\
+int s_name ## _from_attrs(struct s_name *s, struct nlattr *tb[])	\
+{									\
+	const int maxtype = ARRAY_SIZE(s_name ## _nl_policy)-1;		\
+	struct nlattr *tla = tb[tag_number];				\
+	struct nlattr **ntb = nested_attr_tb;				\
+	struct nlattr *nla;						\
+	int err;							\
+	BUILD_BUG_ON(ARRAY_SIZE(s_name ## _nl_policy) > ARRAY_SIZE(nested_attr_tb));	\
+	if (!tla)							\
+		return -ENOMSG;						\
+	DPRINT_TLA(#s_name, "<=-", #tag_name);				\
+	err = nla_parse_nested(ntb, maxtype, tla, s_name ## _nl_policy); \
+	if (err)							\
+		return err;						\
+	err = nla_check_unknown(maxtype, nla_data(tla), nla_len(tla));	\
+	if (err)							\
+	      return err;						\
+									\
+	s_fields							\
+	return 0;							\
+}
+
+#undef __field
+#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put)	\
+		nla = ntb[__nla_type(attr_nr)];				\
+		if (nla) {						\
+			if (s)						\
+				s->name = __get(nla);			\
+			DPRINT_FIELD("<<", nla_type, name, s, nla);	\
+		} else if ((attr_flag) & GENLA_F_REQUIRED) {		\
+			pr_info("<< missing attr: %s\n", #name);	\
+			return -ENOMSG;					\
+		}
+
+/* validate_nla() already checked nla_len <= maxlen appropriately. */
+#undef __array
+#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, __get, __put) \
+		nla = ntb[__nla_type(attr_nr)];				\
+		if (nla) {						\
+			if (s)						\
+				s->name ## _len =			\
+					__get(s->name, nla, maxlen);	\
+			DPRINT_ARRAY("<<", nla_type, name, s, nla);	\
+		} else if ((attr_flag) & GENLA_F_REQUIRED) {		\
+			pr_info("<< missing attr: %s\n", #name);	\
+			return -ENOMSG;					\
+		}							\
+
+#include GENL_MAGIC_INCLUDE_FILE
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)
+
+/*
+ * Magic: define op number to op name mapping				{{{1
+ *									{{{2
+ */
+const char *CONCAT_(GENL_MAGIC_FAMILY, _genl_cmd_to_str)(__u8 cmd)
+{
+	switch (cmd) {
+#undef GENL_op
+#define GENL_op(op_name, op_num, handler, tla_list)		\
+	case op_num: return #op_name;
+#include GENL_MAGIC_INCLUDE_FILE
+	default:
+		     return "unknown";
+	}
+}
+
+#ifdef __KERNEL__
+#include <linux/stringify.h>
+/*
+ * Magic: define genl_ops						{{{1
+ *									{{{2
+ */
+
+#undef GENL_op
+#define GENL_op(op_name, op_num, handler, tla_list)		\
+{								\
+	handler							\
+	.cmd = op_name,						\
+	.policy	= CONCAT_(GENL_MAGIC_FAMILY, _tla_nl_policy),	\
+},
+
+#define ZZZ_genl_ops		CONCAT_(GENL_MAGIC_FAMILY, _genl_ops)
+static struct genl_ops ZZZ_genl_ops[] __read_mostly = {
+#include GENL_MAGIC_INCLUDE_FILE
+};
+
+#undef GENL_op
+#define GENL_op(op_name, op_num, handler, tla_list)
+
+/*
+ * Define the genl_family, multicast groups,				{{{1
+ * and provide register/unregister functions.
+ *									{{{2
+ */
+#define ZZZ_genl_family		CONCAT_(GENL_MAGIC_FAMILY, _genl_family)
+static struct genl_family ZZZ_genl_family __read_mostly = {
+	.id = GENL_ID_GENERATE,
+	.name = __stringify(GENL_MAGIC_FAMILY),
+	.version = GENL_MAGIC_VERSION,
+#ifdef GENL_MAGIC_FAMILY_HDRSZ
+	.hdrsize = NLA_ALIGN(GENL_MAGIC_FAMILY_HDRSZ),
+#endif
+	.maxattr = ARRAY_SIZE(drbd_tla_nl_policy)-1,
+};
+
+/*
+ * Magic: define multicast groups
+ * Magic: define multicast group registration helper
+ */
+#undef GENL_mc_group
+#define GENL_mc_group(group)						\
+static struct genl_multicast_group					\
+CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group) __read_mostly = {		\
+	.name = #group,							\
+};									\
+static int CONCAT_(GENL_MAGIC_FAMILY, _genl_multicast_ ## group)(	\
+	struct sk_buff *skb, gfp_t flags)				\
+{									\
+	unsigned int group_id =						\
+		CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group).id;	\
+	if (!group_id)							\
+		return -EINVAL;						\
+	return genlmsg_multicast(skb, 0, group_id, flags);		\
+}
+
+#include GENL_MAGIC_INCLUDE_FILE
+
+int CONCAT_(GENL_MAGIC_FAMILY, _genl_register)(void)
+{
+	int err = genl_register_family_with_ops(&ZZZ_genl_family,
+		ZZZ_genl_ops, ARRAY_SIZE(ZZZ_genl_ops));
+	if (err)
+		return err;
+#undef GENL_mc_group
+#define GENL_mc_group(group)						\
+	err = genl_register_mc_group(&ZZZ_genl_family,			\
+		&CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group));		\
+	if (err)							\
+		goto fail;						\
+	else								\
+		pr_info("%s: mcg %s: %u\n", #group,			\
+			__stringify(GENL_MAGIC_FAMILY),			\
+			CONCAT_(GENL_MAGIC_FAMILY, _mcg_ ## group).id);
+
+#include GENL_MAGIC_INCLUDE_FILE
+
+#undef GENL_mc_group
+#define GENL_mc_group(group)
+	return 0;
+fail:
+	genl_unregister_family(&ZZZ_genl_family);
+	return err;
+}
+
+void CONCAT_(GENL_MAGIC_FAMILY, _genl_unregister)(void)
+{
+	genl_unregister_family(&ZZZ_genl_family);
+}
+
+/*
+ * Magic: provide conversion functions					{{{1
+ * populate skb from struct.
+ *									{{{2
+ */
+
+#undef GENL_op
+#define GENL_op(op_name, op_num, handler, tla_list)
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)		\
+static int s_name ## _to_skb(struct sk_buff *skb, struct s_name *s,	\
+		const bool exclude_sensitive)				\
+{									\
+	struct nlattr *tla = nla_nest_start(skb, tag_number);		\
+	if (!tla)							\
+		goto nla_put_failure;					\
+	DPRINT_TLA(#s_name, "-=>", #tag_name);				\
+	s_fields							\
+	nla_nest_end(skb, tla);						\
+	return 0;							\
+									\
+nla_put_failure:							\
+	if (tla)							\
+		nla_nest_cancel(skb, tla);				\
+        return -EMSGSIZE;						\
+}									\
+static inline int s_name ## _to_priv_skb(struct sk_buff *skb,		\
+		struct s_name *s)					\
+{									\
+	return s_name ## _to_skb(skb, s, 0);				\
+}									\
+static inline int s_name ## _to_unpriv_skb(struct sk_buff *skb,		\
+		struct s_name *s)					\
+{									\
+	return s_name ## _to_skb(skb, s, 1);				\
+}
+
+
+#undef __field
+#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put)	\
+	if (!exclude_sensitive || !((attr_flag) & GENLA_F_SENSITIVE)) {	\
+		DPRINT_FIELD(">>", nla_type, name, s, NULL);		\
+		__put(skb, attr_nr, s->name);				\
+	}
+
+#undef __array
+#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, __get, __put) \
+	if (!exclude_sensitive || !((attr_flag) & GENLA_F_SENSITIVE)) {	\
+		DPRINT_ARRAY(">>",nla_type, name, s, NULL);		\
+		__put(skb, attr_nr, min_t(int, maxlen,			\
+			s->name ## _len + (nla_type == NLA_NUL_STRING)),\
+						s->name);		\
+	}
+
+#include GENL_MAGIC_INCLUDE_FILE
+
+#endif /* __KERNEL__ */
+
+/* }}}1 */
+#endif /* GENL_MAGIC_FUNC_H */
+/* vim: set foldmethod=marker foldlevel=1 nofoldenable : */
diff --git a/include/linux/genl_magic_struct.h b/include/linux/genl_magic_struct.h
new file mode 100644
index 0000000..745ebfd
--- /dev/null
+++ b/include/linux/genl_magic_struct.h
@@ -0,0 +1,260 @@
+#ifndef GENL_MAGIC_STRUCT_H
+#define GENL_MAGIC_STRUCT_H
+
+#ifndef GENL_MAGIC_FAMILY
+# error "you need to define GENL_MAGIC_FAMILY before inclusion"
+#endif
+
+#ifndef GENL_MAGIC_VERSION
+# error "you need to define GENL_MAGIC_VERSION before inclusion"
+#endif
+
+#ifndef GENL_MAGIC_INCLUDE_FILE
+# error "you need to define GENL_MAGIC_INCLUDE_FILE before inclusion"
+#endif
+
+#include <linux/genetlink.h>
+#include <linux/types.h>
+
+#define CONCAT__(a,b)	a ## b
+#define CONCAT_(a,b)	CONCAT__(a,b)
+
+extern int CONCAT_(GENL_MAGIC_FAMILY, _genl_register)(void);
+extern void CONCAT_(GENL_MAGIC_FAMILY, _genl_unregister)(void);
+
+/*
+ * Extension of genl attribute validation policies			{{{2
+ */
+
+/**
+ * GENLA_F_FLAGS - policy type flags to ease compatible ABI evolvement
+ *
+ * @GENLA_F_REQUIRED: attribute has to be present, or message is considered invalid.
+ * Adding new REQUIRED attributes breaks ABI compatibility, so don't do that.
+ *
+ * @GENLA_F_MANDATORY: if present, receiver _must_ understand it.
+ * Without this, unknown attributes (> maxtype) are _silently_ ignored
+ * by validate_nla().
+ *
+ * To be used for API extensions, so older kernel can reject requests for not
+ * yet implemented features, if newer userland tries to use them even though
+ * the genl_family version clearly indicates they are not available.
+ *
+ * @GENLA_F_MAY_IGNORE: To clearly document the fact, for good measure.
+ * To be used for API extensions for things that have sane defaults,
+ * so newer userland can still talk to older kernel, knowing it will
+ * silently ignore these attributes if not yet known.
+ *
+ * NOTE: These flags overload
+ *   NLA_F_NESTED		(1 << 15)
+ *   NLA_F_NET_BYTEORDER	(1 << 14)
+ * from linux/netlink.h, which are not useful for validate_nla():
+ * NET_BYTEORDER is not used anywhere, and NESTED would be specified by setting
+ * .type = NLA_NESTED in the appropriate policy.
+ *
+ * See also: nla_type()
+ */
+enum {
+	GENLA_F_MAY_IGNORE	= 0,
+	GENLA_F_MANDATORY	= 1 << 14,
+	GENLA_F_REQUIRED	= 1 << 15,
+
+	/* This will not be present in the __u16 .nla_type, but can be
+	 * triggered on in <struct>_to_skb, to exclude "sensitive"
+	 * information from broadcasts, or on unpriviledged get requests.
+	 * This is useful because genetlink multicast groups can be listened in
+	 * on by anyone.  */
+	GENLA_F_SENSITIVE	= 1 << 16,
+};
+
+#define __nla_type(x)	((__u16)((__u16)(x) & (__u16)NLA_TYPE_MASK))
+
+/*									}}}1
+ * MAGIC
+ * multi-include macro expansion magic starts here
+ */
+
+/* MAGIC helpers							{{{2 */
+
+/* possible field types */
+#define __flg_field(attr_nr, attr_flag, name) \
+	__field(attr_nr, attr_flag, name, NLA_FLAG, char, \
+			nla_get_flag, __nla_put_flag)
+#define __u8_field(attr_nr, attr_flag, name)	\
+	__field(attr_nr, attr_flag, name, NLA_U8, unsigned char, \
+			nla_get_u8, NLA_PUT_U8)
+#define __u16_field(attr_nr, attr_flag, name)	\
+	__field(attr_nr, attr_flag, name, NLA_U16, __u16, \
+			nla_get_u16, NLA_PUT_U16)
+#define __u32_field(attr_nr, attr_flag, name)	\
+	__field(attr_nr, attr_flag, name, NLA_U32, __u32, \
+			nla_get_u32, NLA_PUT_U32)
+#define __u64_field(attr_nr, attr_flag, name)	\
+	__field(attr_nr, attr_flag, name, NLA_U64, __u64, \
+			nla_get_u64, NLA_PUT_U64)
+#define __str_field(attr_nr, attr_flag, name, maxlen) \
+	__array(attr_nr, attr_flag, name, NLA_NUL_STRING, char, maxlen, \
+			nla_strlcpy, NLA_PUT)
+#define __bin_field(attr_nr, attr_flag, name, maxlen) \
+	__array(attr_nr, attr_flag, name, NLA_BINARY, char, maxlen, \
+			nla_memcpy, NLA_PUT)
+
+#define __nla_put_flag(skb, attrtype, value)		\
+	do {						\
+		if (value)				\
+			NLA_PUT_FLAG(skb, attrtype);	\
+	} while (0)
+
+#define GENL_op_init(args...)	args
+#define GENL_doit(handler)		\
+	.doit = handler,		\
+	.flags = GENL_ADMIN_PERM,
+#define GENL_dumpit(handler)		\
+	.dumpit = handler,		\
+	.flags = GENL_ADMIN_PERM,
+
+/*									}}}1
+ * Magic: define the enum symbols for genl_ops
+ * Magic: define the enum symbols for top level attributes
+ * Magic: define the enum symbols for nested attributes
+ *									{{{2
+ */
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)
+
+#undef GENL_mc_group
+#define GENL_mc_group(group)
+
+#undef GENL_notification
+#define GENL_notification(op_name, op_num, mcast_group, tla_list)	\
+	op_name = op_num,
+
+#undef GENL_op
+#define GENL_op(op_name, op_num, handler, tla_list)			\
+	op_name = op_num,
+
+enum {
+#include GENL_MAGIC_INCLUDE_FILE
+};
+
+#undef GENL_notification
+#define GENL_notification(op_name, op_num, mcast_group, tla_list)
+
+#undef GENL_op
+#define GENL_op(op_name, op_num, handler, attr_list)
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields) \
+		tag_name = tag_number,
+
+enum {
+#include GENL_MAGIC_INCLUDE_FILE
+};
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)	\
+enum {								\
+	s_fields						\
+};
+
+#undef __field
+#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put)	\
+	T_ ## name = (__u16)(attr_nr | attr_flag),
+
+#undef __array
+#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, __get, __put) \
+	T_ ## name = (__u16)(attr_nr | attr_flag),
+
+#include GENL_MAGIC_INCLUDE_FILE
+
+/*									}}}1
+ * Magic: compile time assert unique numbers for operations
+ * Magic: -"- unique numbers for top level attributes
+ * Magic: -"- unique numbers for nested attributes
+ *									{{{2
+ */
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)
+
+#undef GENL_op
+#define GENL_op(op_name, op_num, handler, attr_list)	\
+	case op_name:
+
+#undef GENL_notification
+#define GENL_notification(op_name, op_num, mcast_group, tla_list)	\
+	case op_name:
+
+static inline void ct_assert_unique_operations(void)
+{
+	switch (0) {
+#include GENL_MAGIC_INCLUDE_FILE
+		;
+	}
+}
+
+#undef GENL_op
+#define GENL_op(op_name, op_num, handler, attr_list)
+
+#undef GENL_notification
+#define GENL_notification(op_name, op_num, mcast_group, tla_list)
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)		\
+		case tag_number:
+
+static inline void ct_assert_unique_top_level_attributes(void)
+{
+	switch (0) {
+#include GENL_MAGIC_INCLUDE_FILE
+		;
+	}
+}
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)		\
+static inline void ct_assert_unique_ ## s_name ## _attributes(void)	\
+{									\
+	switch (0) {							\
+		s_fields						\
+			;						\
+	}								\
+}
+
+#undef __field
+#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put)	\
+	case attr_nr:
+
+#undef __array
+#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, __get, __put) \
+	case attr_nr:
+
+#include GENL_MAGIC_INCLUDE_FILE
+
+/*									}}}1
+ * Magic: declare structs
+ * struct <name> {
+ *	fields
+ * };
+ *									{{{2
+ */
+
+#undef GENL_struct
+#define GENL_struct(tag_name, tag_number, s_name, s_fields)		\
+struct s_name { s_fields };
+
+#undef __field
+#define __field(attr_nr, attr_flag, name, nla_type, type, __get, __put) \
+	type name;
+
+#undef __array
+#define __array(attr_nr, attr_flag, name, nla_type, type, maxlen, __get, __put) \
+	type name[maxlen];	\
+	__u32 name ## _len;
+
+#include GENL_MAGIC_INCLUDE_FILE
+
+/* }}}1 */
+#endif /* GENL_MAGIC_STRUCT_H */
+/* vim: set foldmethod=marker nofoldenable : */
-- 
1.7.4.1

[PATCH 09/18] drbd: switch configuration interface from connector to genetlink

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_actlog.c |    3 +-
 drivers/block/drbd/drbd_int.h    |   36 +-
 drivers/block/drbd/drbd_main.c   |   27 +-
 drivers/block/drbd/drbd_nl.c     | 1569 +++++++++++++++++++-------------------
 drivers/block/drbd/drbd_state.c  |    7 +-
 include/linux/drbd.h             |   35 +-
 6 files changed, 839 insertions(+), 838 deletions(-)

diff --git a/drivers/block/drbd/drbd_actlog.c b/drivers/block/drbd/drbd_actlog.c
index 9ab6365..6d1e892 100644
--- a/drivers/block/drbd/drbd_actlog.c
+++ b/drivers/block/drbd/drbd_actlog.c
@@ -692,6 +692,7 @@ static int w_update_odbm(struct drbd_work *w, int unused)
 {
 	struct update_odbm_work *udw = container_of(w, struct update_odbm_work, w);
 	struct drbd_conf *mdev = w->mdev;
+	struct sib_info sib = { .sib_reason = SIB_SYNC_PROGRESS, };
 
 	if (!get_ldev(mdev)) {
 		if (__ratelimit(&drbd_ratelimit_state))
@@ -715,7 +716,7 @@ static int w_update_odbm(struct drbd_work *w, int unused)
 			break;
 		}
 	}
-	drbd_bcast_sync_progress(mdev);
+	drbd_bcast_event(mdev, &sib);
 
 	return 1;
 }
diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index 5d90b9d..acd2877 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -44,6 +44,7 @@
 #include <net/tcp.h>
 #include <linux/lru_cache.h>
 #include <linux/prefetch.h>
+#include <linux/drbd_genl_api.h>
 #include <linux/drbd.h>
 #include "drbd_state.h"
 
@@ -65,7 +66,6 @@
 extern unsigned int minor_count;
 extern int disable_sendpage;
 extern int allow_oos;
-extern unsigned int cn_idx;
 
 #ifdef CONFIG_DRBD_FAULT_INJECTION
 extern int enable_faults;
@@ -865,14 +865,6 @@ struct drbd_md {
 	 */
 };
 
-/* for sync_conf and other types... */
-#define NL_PACKET(name, number, fields) struct name { fields };
-#define NL_INTEGER(pn,pr,member) int member;
-#define NL_INT64(pn,pr,member) __u64 member;
-#define NL_BIT(pn,pr,member)   unsigned member:1;
-#define NL_STRING(pn,pr,member,len) unsigned char member[len]; int member ## _len;
-#include "linux/drbd_nl.h"
-
 struct drbd_backing_dev {
 	struct block_device *backing_bdev;
 	struct block_device *md_bdev;
@@ -1502,7 +1494,7 @@ enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor,
 extern void drbd_free_mdev(struct drbd_conf *mdev);
 extern void drbd_delete_device(unsigned int minor);
 
-struct drbd_tconn *drbd_new_tconn(char *name);
+struct drbd_tconn *drbd_new_tconn(const char *name);
 extern void drbd_free_tconn(struct drbd_tconn *tconn);
 struct drbd_tconn *conn_by_name(const char *name);
 
@@ -1679,16 +1671,22 @@ extern int __drbd_set_out_of_sync(struct drbd_conf *mdev, sector_t sector,
 extern void drbd_al_apply_to_bm(struct drbd_conf *mdev);
 extern void drbd_al_shrink(struct drbd_conf *mdev);
 
-
 /* drbd_nl.c */
-
-void drbd_nl_cleanup(void);
-int __init drbd_nl_init(void);
-void drbd_bcast_state(struct drbd_conf *mdev, union drbd_state);
-void drbd_bcast_sync_progress(struct drbd_conf *mdev);
-void drbd_bcast_ee(struct drbd_conf *, const char *, const int, const char *,
-		   const char *, const struct drbd_peer_request *);
-
+/* state info broadcast */
+struct sib_info {
+	enum drbd_state_info_bcast_reason sib_reason;
+	union {
+		struct {
+			char *helper_name;
+			unsigned helper_exit_code;
+		};
+		struct {
+			union drbd_state os;
+			union drbd_state ns;
+		};
+	};
+};
+void drbd_bcast_event(struct drbd_conf *mdev, const struct sib_info *sib);
 
 /*
  * inline helper functions
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 21914f4..9b41213 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -86,7 +86,6 @@ MODULE_PARM_DESC(allow_oos, "DONT USE!");
 module_param(minor_count, uint, 0444);
 module_param(disable_sendpage, bool, 0644);
 module_param(allow_oos, bool, 0);
-module_param(cn_idx, uint, 0444);
 module_param(proc_details, int, 0644);
 
 #ifdef CONFIG_DRBD_FAULT_INJECTION
@@ -108,7 +107,6 @@ module_param(fault_devs, int, 0644);
 unsigned int minor_count = DRBD_MINOR_COUNT_DEF;
 int disable_sendpage;
 int allow_oos;
-unsigned int cn_idx = CN_IDX_DRBD;
 int proc_details;       /* Detail level in proc drbd*/
 
 /* Module parameter for setting the user mode helper program
@@ -2175,7 +2173,7 @@ static void drbd_cleanup(void)
 	if (drbd_proc)
 		remove_proc_entry("drbd", NULL);
 
-	drbd_nl_cleanup();
+	drbd_genl_unregister();
 
 	idr_for_each_entry(&minors, mdev, i)
 		drbd_delete_device(i);
@@ -2237,6 +2235,9 @@ struct drbd_tconn *conn_by_name(const char *name)
 {
 	struct drbd_tconn *tconn;
 
+	if (!name || !name[0])
+		return NULL;
+
 	write_lock_irq(&global_state_lock);
 	list_for_each_entry(tconn, &drbd_tconns, all_tconn) {
 		if (!strcmp(tconn->name, name))
@@ -2248,7 +2249,7 @@ found:
 	return tconn;
 }
 
-struct drbd_tconn *drbd_new_tconn(char *name)
+struct drbd_tconn *drbd_new_tconn(const char *name)
 {
 	struct drbd_tconn *tconn;
 
@@ -2333,6 +2334,7 @@ enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor,
 
 	mdev->tconn = tconn;
 	mdev->minor = minor;
+	mdev->vnr = vnr;
 
 	drbd_init_set_defaults(mdev);
 
@@ -2462,10 +2464,6 @@ int __init drbd_init(void)
 #endif
 	}
 
-	err = drbd_nl_init();
-	if (err)
-		return err;
-
 	err = register_blkdev(DRBD_MAJOR, "drbd");
 	if (err) {
 		printk(KERN_ERR
@@ -2474,6 +2472,13 @@ int __init drbd_init(void)
 		return err;
 	}
 
+	err = drbd_genl_register();
+	if (err) {
+		pr_err("drbd: unable to register generic netlink family\n");
+		goto fail;
+	}
+
+
 	register_reboot_notifier(&drbd_notifier);
 
 	/*
@@ -2488,12 +2493,12 @@ int __init drbd_init(void)
 
 	err = drbd_create_mempools();
 	if (err)
-		goto Enomem;
+		goto fail;
 
 	drbd_proc = proc_create_data("drbd", S_IFREG | S_IRUGO , NULL, &drbd_proc_fops, NULL);
 	if (!drbd_proc)	{
 		printk(KERN_ERR "drbd: unable to register proc file\n");
-		goto Enomem;
+		goto fail;
 	}
 
 	rwlock_init(&global_state_lock);
@@ -2508,7 +2513,7 @@ int __init drbd_init(void)
 
 	return 0; /* Success! */
 
-Enomem:
+fail:
 	drbd_cleanup();
 	if (err == -ENOMEM)
 		/* currently always the case */
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index f2739fd..a8f27cb 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -29,110 +29,258 @@
 #include <linux/fs.h>
 #include <linux/file.h>
 #include <linux/slab.h>
-#include <linux/connector.h>
 #include <linux/blkpg.h>
 #include <linux/cpumask.h>
 #include "drbd_int.h"
 #include "drbd_req.h"
 #include "drbd_wrappers.h"
 #include <asm/unaligned.h>
-#include <linux/drbd_tag_magic.h>
 #include <linux/drbd_limits.h>
-#include <linux/compiler.h>
 #include <linux/kthread.h>
 
-static unsigned short *tl_add_blob(unsigned short *, enum drbd_tags, const void *, int);
-static unsigned short *tl_add_str(unsigned short *, enum drbd_tags, const char *);
-static unsigned short *tl_add_int(unsigned short *, enum drbd_tags, const void *);
+#include <net/genetlink.h>
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,31)
+/*
+ * copied from more recent kernel source
+ */
+int genl_register_family_with_ops(struct genl_family *family,
+	struct genl_ops *ops, size_t n_ops)
+{
+	int err, i;
 
-/* see get_sb_bdev and bd_claim */
+	err = genl_register_family(family);
+	if (err)
+		return err;
+
+	for (i = 0; i < n_ops; ++i, ++ops) {
+		err = genl_register_ops(family, ops);
+		if (err)
+			goto err_out;
+	}
+	return 0;
+err_out:
+	genl_unregister_family(family);
+	return err;
+}
+#endif
+
+/* .doit */
+// int drbd_adm_create_resource(struct sk_buff *skb, struct genl_info *info);
+// int drbd_adm_delete_resource(struct sk_buff *skb, struct genl_info *info);
+
+int drbd_adm_add_minor(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_delete_minor(struct sk_buff *skb, struct genl_info *info);
+
+int drbd_adm_create_connection(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_delete_connection(struct sk_buff *skb, struct genl_info *info);
+
+int drbd_adm_set_role(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_detach(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_connect(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_resize(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_start_ov(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_new_c_uuid(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_disconnect(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_invalidate(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_invalidate_peer(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_pause_sync(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_resume_sync(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_suspend_io(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_resume_io(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_outdate(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_syncer(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_get_status(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_get_timeout_type(struct sk_buff *skb, struct genl_info *info);
+/* .dumpit */
+int drbd_adm_get_status_all(struct sk_buff *skb, struct netlink_callback *cb);
+
+#include <linux/drbd_genl_api.h>
+#include <linux/genl_magic_func.h>
+
+/* used blkdev_get_by_path, to claim our meta data device(s) */
 static char *drbd_m_holder = "Hands off! this is DRBD's meta data device.";
 
-/* Generate the tag_list to struct functions */
-#define NL_PACKET(name, number, fields) \
-static int name ## _from_tags( \
-	unsigned short *tags, struct name *arg) __attribute__ ((unused)); \
-static int name ## _from_tags( \
-	unsigned short *tags, struct name *arg) \
-{ \
-	int tag; \
-	int dlen; \
-	\
-	while ((tag = get_unaligned(tags++)) != TT_END) {	\
-		dlen = get_unaligned(tags++);			\
-		switch (tag_number(tag)) { \
-		fields \
-		default: \
-			if (tag & T_MANDATORY) { \
-				printk(KERN_ERR "drbd: Unknown tag: %d\n", tag_number(tag)); \
-				return 0; \
-			} \
-		} \
-		tags = (unsigned short *)((char *)tags + dlen); \
-	} \
-	return 1; \
+/* Configuration is strictly serialized, because generic netlink message
+ * processing is strictly serialized by the genl_lock().
+ * Which means we can use one static global drbd_config_context struct.
+ */
+static struct drbd_config_context {
+	/* assigned from drbd_genlmsghdr */
+	unsigned int minor;
+	/* assigned from request attributes, if present */
+	unsigned int volume;
+#define VOLUME_UNSPECIFIED		(-1U)
+	/* pointer into the request skb,
+	 * limited lifetime! */
+	char *conn_name;
+
+	/* reply buffer */
+	struct sk_buff *reply_skb;
+	/* pointer into reply buffer */
+	struct drbd_genlmsghdr *reply_dh;
+	/* resolved from attributes, if possible */
+	struct drbd_conf *mdev;
+	struct drbd_tconn *tconn;
+} adm_ctx;
+
+static void drbd_adm_send_reply(struct sk_buff *skb, struct genl_info *info)
+{
+	genlmsg_end(skb, genlmsg_data(nlmsg_data(nlmsg_hdr(skb))));
+	if (genlmsg_reply(skb, info))
+		printk(KERN_ERR "drbd: error sending genl reply\n");
 }
-#define NL_INTEGER(pn, pr, member) \
-	case pn: /* D_ASSERT( tag_type(tag) == TT_INTEGER ); */ \
-		arg->member = get_unaligned((int *)(tags));	\
-		break;
-#define NL_INT64(pn, pr, member) \
-	case pn: /* D_ASSERT( tag_type(tag) == TT_INT64 ); */ \
-		arg->member = get_unaligned((u64 *)(tags));	\
-		break;
-#define NL_BIT(pn, pr, member) \
-	case pn: /* D_ASSERT( tag_type(tag) == TT_BIT ); */ \
-		arg->member = *(char *)(tags) ? 1 : 0; \
-		break;
-#define NL_STRING(pn, pr, member, len) \
-	case pn: /* D_ASSERT( tag_type(tag) == TT_STRING ); */ \
-		if (dlen > len) { \
-			printk(KERN_ERR "drbd: arg too long: %s (%u wanted, max len: %u bytes)\n", \
-				#member, dlen, (unsigned int)len); \
-			return 0; \
-		} \
-		 arg->member ## _len = dlen; \
-		 memcpy(arg->member, tags, min_t(size_t, dlen, len)); \
-		 break;
-#include "linux/drbd_nl.h"
-
-/* Generate the struct to tag_list functions */
-#define NL_PACKET(name, number, fields) \
-static unsigned short* \
-name ## _to_tags( \
-	struct name *arg, unsigned short *tags) __attribute__ ((unused)); \
-static unsigned short* \
-name ## _to_tags( \
-	struct name *arg, unsigned short *tags) \
-{ \
-	fields \
-	return tags; \
+
+/* Used on a fresh "drbd_adm_prepare"d reply_skb, this cannot fail: The only
+ * reason it could fail was no space in skb, and there are 4k available. */
+static int drbd_msg_put_info(const char *info)
+{
+	struct sk_buff *skb = adm_ctx.reply_skb;
+	struct nlattr *nla;
+	int err = -EMSGSIZE;
+
+	if (!info || !info[0])
+		return 0;
+
+	nla = nla_nest_start(skb, DRBD_NLA_CFG_REPLY);
+	if (!nla)
+		return err;
+
+	err = nla_put_string(skb, T_info_text, info);
+	if (err) {
+		nla_nest_cancel(skb, nla);
+		return err;
+	} else
+		nla_nest_end(skb, nla);
+	return 0;
+}
+
+/* This would be a good candidate for a "pre_doit" hook,
+ * and per-family private info->pointers.
+ * But we need to stay compatible with older kernels.
+ * If it returns successfully, adm_ctx members are valid.
+ */
+#define DRBD_ADM_NEED_MINOR	1
+#define DRBD_ADM_NEED_CONN	2
+static int drbd_adm_prepare(struct sk_buff *skb, struct genl_info *info,
+		unsigned flags)
+{
+	struct drbd_genlmsghdr *d_in = info->userhdr;
+	const u8 cmd = info->genlhdr->cmd;
+	int err;
+
+	memset(&adm_ctx, 0, sizeof(adm_ctx));
+
+	/* genl_rcv_msg only checks for CAP_NET_ADMIN on "GENL_ADMIN_PERM" :( */
+	if (cmd != DRBD_ADM_GET_STATUS
+	&& security_netlink_recv(skb, CAP_SYS_ADMIN))
+	       return -EPERM;
+
+	adm_ctx.reply_skb = genlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL);
+	if (!adm_ctx.reply_skb)
+		goto fail;
+
+	adm_ctx.reply_dh = genlmsg_put_reply(adm_ctx.reply_skb,
+					info, &drbd_genl_family, 0, cmd);
+	/* put of a few bytes into a fresh skb of >= 4k will always succeed.
+	 * but anyways */
+	if (!adm_ctx.reply_dh)
+		goto fail;
+
+	adm_ctx.reply_dh->minor = d_in->minor;
+	adm_ctx.reply_dh->ret_code = NO_ERROR;
+
+	if (info->attrs[DRBD_NLA_CFG_CONTEXT]) {
+		struct nlattr *nla;
+		/* parse and validate only */
+		err = drbd_cfg_context_from_attrs(NULL, info->attrs);
+		if (err)
+			goto fail;
+
+		/* It was present, and valid,
+		 * copy it over to the reply skb. */
+		err = nla_put_nohdr(adm_ctx.reply_skb,
+				info->attrs[DRBD_NLA_CFG_CONTEXT]->nla_len,
+				info->attrs[DRBD_NLA_CFG_CONTEXT]);
+		if (err)
+			goto fail;
+
+		/* and assign stuff to the global adm_ctx */
+		nla = nested_attr_tb[__nla_type(T_ctx_volume)];
+		adm_ctx.volume = nla ? nla_get_u32(nla) : VOLUME_UNSPECIFIED;
+		nla = nested_attr_tb[__nla_type(T_ctx_conn_name)];
+		if (nla)
+			adm_ctx.conn_name = nla_data(nla);
+	} else
+		adm_ctx.volume = VOLUME_UNSPECIFIED;
+
+	adm_ctx.minor = d_in->minor;
+	adm_ctx.mdev = minor_to_mdev(d_in->minor);
+	adm_ctx.tconn = conn_by_name(adm_ctx.conn_name);
+
+	pr_info("adm request: cmd=%u[%s], flags=0x%x, minor=%d, conn=%s\n",
+		cmd, drbd_genl_cmd_to_str(cmd), d_in->flags,
+		d_in->minor, adm_ctx.conn_name ?: "n/a");
+
+	if (!adm_ctx.mdev && (flags & DRBD_ADM_NEED_MINOR)) {
+		drbd_msg_put_info("unknown minor");
+		return ERR_MINOR_INVALID;
+	}
+	if (!adm_ctx.tconn && (flags & DRBD_ADM_NEED_CONN)) {
+		drbd_msg_put_info("unknown connection");
+		return ERR_INVALID_REQUEST;
+	}
+
+	/* some more paranoia, if the request was over-determined */
+	if (adm_ctx.mdev &&
+	    adm_ctx.volume != VOLUME_UNSPECIFIED &&
+	    adm_ctx.volume != adm_ctx.mdev->vnr) {
+		pr_warning("request: minor=%u, volume=%u; but that minor is volume %u in %s\n",
+				adm_ctx.minor, adm_ctx.volume,
+				adm_ctx.mdev->vnr, adm_ctx.mdev->tconn->name);
+		drbd_msg_put_info("over-determined configuration context mismatch");
+		return ERR_INVALID_REQUEST;
+	}
+	if (adm_ctx.mdev && adm_ctx.tconn &&
+	    adm_ctx.mdev->tconn != adm_ctx.tconn) {
+		pr_warning("request: minor=%u, conn=%s; but that minor belongs to connection %s\n",
+				adm_ctx.minor, adm_ctx.conn_name, adm_ctx.mdev->tconn->name);
+		drbd_msg_put_info("over-determined configuration context mismatch");
+		return ERR_INVALID_REQUEST;
+	}
+	return NO_ERROR;
+
+fail:
+	nlmsg_free(adm_ctx.reply_skb);
+	adm_ctx.reply_skb = NULL;
+	return -ENOMEM;
 }
 
-#define NL_INTEGER(pn, pr, member) \
-	put_unaligned(pn | pr | TT_INTEGER, tags++);	\
-	put_unaligned(sizeof(int), tags++);		\
-	put_unaligned(arg->member, (int *)tags);	\
-	tags = (unsigned short *)((char *)tags+sizeof(int));
-#define NL_INT64(pn, pr, member) \
-	put_unaligned(pn | pr | TT_INT64, tags++);	\
-	put_unaligned(sizeof(u64), tags++);		\
-	put_unaligned(arg->member, (u64 *)tags);	\
-	tags = (unsigned short *)((char *)tags+sizeof(u64));
-#define NL_BIT(pn, pr, member) \
-	put_unaligned(pn | pr | TT_BIT, tags++);	\
-	put_unaligned(sizeof(char), tags++);		\
-	*(char *)tags = arg->member; \
-	tags = (unsigned short *)((char *)tags+sizeof(char));
-#define NL_STRING(pn, pr, member, len) \
-	put_unaligned(pn | pr | TT_STRING, tags++);	\
-	put_unaligned(arg->member ## _len, tags++);	\
-	memcpy(tags, arg->member, arg->member ## _len); \
-	tags = (unsigned short *)((char *)tags + arg->member ## _len);
-#include "linux/drbd_nl.h"
-
-void drbd_bcast_ev_helper(struct drbd_conf *mdev, char *helper_name);
-void drbd_nl_send_reply(struct cn_msg *, int);
+static int drbd_adm_finish(struct genl_info *info, int retcode)
+{
+	struct nlattr *nla;
+	const char *conn_name = NULL;
+	const u8 cmd = info->genlhdr->cmd;
+
+	if (!adm_ctx.reply_skb)
+		return -ENOMEM;
+
+	adm_ctx.reply_dh->ret_code = retcode;
+
+	nla = info->attrs[DRBD_NLA_CFG_CONTEXT];
+	if (nla) {
+		nla = nla_find_nested(nla, __nla_type(T_ctx_conn_name));
+		if (nla)
+			conn_name = nla_data(nla);
+	}
+
+	pr_info("adm reply: cmd=%u[%s], retcode=%d, minor=%d, conn=%s\n",
+		cmd, drbd_genl_cmd_to_str(cmd), retcode,
+		adm_ctx.minor, adm_ctx.conn_name ?: "n/a");
+
+	drbd_adm_send_reply(adm_ctx.reply_skb, info);
+	return 0;
+}
 
 int drbd_khelper(struct drbd_conf *mdev, char *cmd)
 {
@@ -142,9 +290,9 @@ int drbd_khelper(struct drbd_conf *mdev, char *cmd)
 			NULL, /* Will be set to address family */
 			NULL, /* Will be set to address */
 			NULL };
-
 	char mb[12], af[20], ad[60], *afs;
 	char *argv[] = {usermode_helper, cmd, mb, NULL };
+	struct sib_info sib;
 	int ret;
 
 	snprintf(mb, 12, "minor-%d", mdev_to_minor(mdev));
@@ -177,8 +325,9 @@ int drbd_khelper(struct drbd_conf *mdev, char *cmd)
 	drbd_md_sync(mdev);
 
 	dev_info(DEV, "helper command: %s %s %s\n", usermode_helper, cmd, mb);
-
-	drbd_bcast_ev_helper(mdev, cmd);
+	sib.sib_reason = SIB_HELPER_PRE;
+	sib.helper_name = cmd;
+	drbd_bcast_event(mdev, &sib);
 	ret = call_usermodehelper(usermode_helper, argv, envp, 1);
 	if (ret)
 		dev_warn(DEV, "helper command: %s %s %s exit code %u (0x%x)\n",
@@ -188,6 +337,9 @@ int drbd_khelper(struct drbd_conf *mdev, char *cmd)
 		dev_info(DEV, "helper command: %s %s %s exit code %u (0x%x)\n",
 				usermode_helper, cmd, mb,
 				(ret >> 8) & 0xff, ret);
+	sib.sib_reason = SIB_HELPER_POST;
+	sib.helper_exit_code = ret;
+	drbd_bcast_event(mdev, &sib);
 
 	if (ret < 0) /* Ignore any ERRNOs we got. */
 		ret = 0;
@@ -362,7 +514,7 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force)
 		}
 
 		if (rv == SS_NOTHING_TO_DO)
-			goto fail;
+			goto out;
 		if (rv == SS_PRIMARY_NOP && mask.pdsk == 0) {
 			nps = drbd_try_outdate_peer(mdev);
 
@@ -388,13 +540,13 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force)
 			rv = _drbd_request_state(mdev, mask, val,
 						CS_VERBOSE + CS_WAIT_COMPLETE);
 			if (rv < SS_SUCCESS)
-				goto fail;
+				goto out;
 		}
 		break;
 	}
 
 	if (rv < SS_SUCCESS)
-		goto fail;
+		goto out;
 
 	if (forced)
 		dev_warn(DEV, "Forced to consider local data as UpToDate!\n");
@@ -438,33 +590,46 @@ drbd_set_role(struct drbd_conf *mdev, enum drbd_role new_role, int force)
 	drbd_md_sync(mdev);
 
 	kobject_uevent(&disk_to_dev(mdev->vdisk)->kobj, KOBJ_CHANGE);
- fail:
+out:
 	mutex_unlock(mdev->state_mutex);
 	return rv;
 }
 
-static int drbd_nl_primary(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			   struct drbd_nl_cfg_reply *reply)
+static const char *from_attrs_err_to_txt(int err)
 {
-	struct primary primary_args;
-
-	memset(&primary_args, 0, sizeof(struct primary));
-	if (!primary_from_tags(nlp->tag_list, &primary_args)) {
-		reply->ret_code = ERR_MANDATORY_TAG;
-		return 0;
-	}
-
-	reply->ret_code =
-		drbd_set_role(mdev, R_PRIMARY, primary_args.primary_force);
-
-	return 0;
+	return	err == -ENOMSG ? "required attribute missing" :
+		err == -EOPNOTSUPP ? "unknown mandatory attribute" :
+		"invalid attribute value";
 }
 
-static int drbd_nl_secondary(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			     struct drbd_nl_cfg_reply *reply)
+int drbd_adm_set_role(struct sk_buff *skb, struct genl_info *info)
 {
-	reply->ret_code = drbd_set_role(mdev, R_SECONDARY, 0);
+	struct set_role_parms parms;
+	int err;
+	enum drbd_ret_code retcode;
 
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
+
+	memset(&parms, 0, sizeof(parms));
+	if (info->attrs[DRBD_NLA_SET_ROLE_PARMS]) {
+		err = set_role_parms_from_attrs(&parms, info->attrs);
+		if (err) {
+			retcode = ERR_MANDATORY_TAG;
+			drbd_msg_put_info(from_attrs_err_to_txt(err));
+			goto out;
+		}
+	}
+
+	if (info->genlhdr->cmd == DRBD_ADM_PRIMARY)
+		retcode = drbd_set_role(adm_ctx.mdev, R_PRIMARY, parms.assume_uptodate);
+	else
+		retcode = drbd_set_role(adm_ctx.mdev, R_SECONDARY, 0);
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
@@ -541,6 +706,12 @@ char *ppsize(char *buf, unsigned long long size)
  *  R_PRIMARY D_INCONSISTENT, and C_SYNC_TARGET:
  *  peer may not initiate a resize.
  */
+/* Note these are not to be confused with
+ * drbd_adm_suspend_io/drbd_adm_resume_io,
+ * which are (sub) state changes triggered by admin (drbdsetup),
+ * and can be long lived.
+ * This changes an mdev->flag, is triggered by drbd internals,
+ * and should be short-lived. */
 void drbd_suspend_io(struct drbd_conf *mdev)
 {
 	set_bit(SUSPEND_IO, &mdev->flags);
@@ -881,11 +1052,10 @@ static void drbd_suspend_al(struct drbd_conf *mdev)
 		dev_info(DEV, "Suspended AL updates\n");
 }
 
-/* does always return 0;
- * interesting return code is in reply->ret_code */
-static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			     struct drbd_nl_cfg_reply *reply)
+int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 {
+	struct drbd_conf *mdev;
+	int err;
 	enum drbd_ret_code retcode;
 	enum determine_dev_size dd;
 	sector_t max_possible_sectors;
@@ -897,6 +1067,13 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 	enum drbd_state_rv rv;
 	int cp_discovered = 0;
 
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto fail;
+
+	mdev = adm_ctx.mdev;
 	conn_reconfig_start(mdev->tconn);
 
 	/* if you want to reconfigure, please tear down first */
@@ -910,7 +1087,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 	 * to realize a "hot spare" feature (not that I'd recommend that) */
 	wait_event(mdev->misc_wait, !atomic_read(&mdev->local_cnt));
 
-	/* allocation not in the IO path, cqueue thread context */
+	/* allocation not in the IO path, drbdsetup context */
 	nbc = kzalloc(sizeof(struct drbd_backing_dev), GFP_KERNEL);
 	if (!nbc) {
 		retcode = ERR_NOMEM;
@@ -922,12 +1099,14 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 	nbc->dc.fencing       = DRBD_FENCING_DEF;
 	nbc->dc.max_bio_bvecs = DRBD_MAX_BIO_BVECS_DEF;
 
-	if (!disk_conf_from_tags(nlp->tag_list, &nbc->dc)) {
+	err = disk_conf_from_attrs(&nbc->dc, info->attrs);
+	if (err) {
 		retcode = ERR_MANDATORY_TAG;
+		drbd_msg_put_info(from_attrs_err_to_txt(err));
 		goto fail;
 	}
 
-	if (nbc->dc.meta_dev_idx < DRBD_MD_INDEX_FLEX_INT) {
+	if ((int)nbc->dc.meta_dev_idx < DRBD_MD_INDEX_FLEX_INT) {
 		retcode = ERR_MD_IDX_INVALID;
 		goto fail;
 	}
@@ -961,7 +1140,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 	 */
 	bdev = blkdev_get_by_path(nbc->dc.meta_dev,
 				  FMODE_READ | FMODE_WRITE | FMODE_EXCL,
-				  (nbc->dc.meta_dev_idx < 0) ?
+				  ((int)nbc->dc.meta_dev_idx < 0) ?
 				  (void *)mdev : (void *)drbd_m_holder);
 	if (IS_ERR(bdev)) {
 		dev_err(DEV, "open(\"%s\") failed with %ld\n", nbc->dc.meta_dev,
@@ -997,7 +1176,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 		goto fail;
 	}
 
-	if (nbc->dc.meta_dev_idx < 0) {
+	if ((int)nbc->dc.meta_dev_idx < 0) {
 		max_possible_sectors = DRBD_MAX_SECTORS_FLEX;
 		/* at least one MB, otherwise it does not make sense */
 		min_md_device_sectors = (2<<10);
@@ -1028,7 +1207,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 		dev_warn(DEV, "==> truncating very big lower level device "
 			"to currently maximum possible %llu sectors <==\n",
 			(unsigned long long) max_possible_sectors);
-		if (nbc->dc.meta_dev_idx >= 0)
+		if ((int)nbc->dc.meta_dev_idx >= 0)
 			dev_warn(DEV, "==>> using internal or flexible "
 				      "meta data may help <<==\n");
 	}
@@ -1242,8 +1421,8 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 
 	kobject_uevent(&disk_to_dev(mdev->vdisk)->kobj, KOBJ_CHANGE);
 	put_ldev(mdev);
-	reply->ret_code = retcode;
 	conn_reconfig_done(mdev->tconn);
+	drbd_adm_finish(info, retcode);
 	return 0;
 
  force_diskless_dec:
@@ -1251,6 +1430,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
  force_diskless:
 	drbd_force_state(mdev, NS(disk, D_FAILED));
 	drbd_md_sync(mdev);
+	conn_reconfig_done(mdev->tconn);
  fail:
 	if (nbc) {
 		if (nbc->backing_bdev)
@@ -1263,8 +1443,7 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 	}
 	lc_destroy(resync_lru);
 
-	reply->ret_code = retcode;
-	conn_reconfig_done(mdev->tconn);
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
@@ -1273,42 +1452,54 @@ static int drbd_nl_disk_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
  * Then we transition to D_DISKLESS, and wait for put_ldev() to return all
  * internal references as well.
  * Only then we have finally detached. */
-static int drbd_nl_detach(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			  struct drbd_nl_cfg_reply *reply)
+int drbd_adm_detach(struct sk_buff *skb, struct genl_info *info)
 {
+	struct drbd_conf *mdev;
 	enum drbd_ret_code retcode;
-	int ret;
+
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
+
+	mdev = adm_ctx.mdev;
 	drbd_suspend_io(mdev); /* so no-one is stuck in drbd_al_begin_io */
-	retcode = drbd_request_state(mdev, NS(disk, D_FAILED));
-	/* D_FAILED will transition to DISKLESS. */
-	ret = wait_event_interruptible(mdev->misc_wait,
-			mdev->state.disk != D_FAILED);
+	retcode = drbd_request_state(mdev, NS(disk, D_DISKLESS));
+	wait_event(mdev->misc_wait,
+			mdev->state.disk != D_DISKLESS ||
+			!atomic_read(&mdev->local_cnt));
 	drbd_resume_io(mdev);
-	if ((int)retcode == (int)SS_IS_DISKLESS)
-		retcode = SS_NOTHING_TO_DO;
-	if (ret)
-		retcode = ERR_INTR;
-	reply->ret_code = retcode;
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-static int drbd_nl_net_conf(struct drbd_tconn *tconn, struct drbd_nl_cfg_req *nlp,
-			    struct drbd_nl_cfg_reply *reply)
+int drbd_adm_connect(struct sk_buff *skb, struct genl_info *info)
 {
-	int i;
-	enum drbd_ret_code retcode;
+	char hmac_name[CRYPTO_MAX_ALG_NAME];
+	struct drbd_conf *mdev;
 	struct net_conf *new_conf = NULL;
 	struct crypto_hash *tfm = NULL;
 	struct crypto_hash *integrity_w_tfm = NULL;
 	struct crypto_hash *integrity_r_tfm = NULL;
-	struct drbd_conf *mdev;
-	char hmac_name[CRYPTO_MAX_ALG_NAME];
 	void *int_dig_out = NULL;
 	void *int_dig_in = NULL;
 	void *int_dig_vv = NULL;
 	struct drbd_tconn *oconn;
+	struct drbd_tconn *tconn;
 	struct sockaddr *new_my_addr, *new_peer_addr, *taken_addr;
+	enum drbd_ret_code retcode;
+	int i;
+	int err;
+
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_CONN);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
+	tconn = adm_ctx.tconn;
 	conn_reconfig_start(tconn);
 
 	if (tconn->cstate > C_STANDALONE) {
@@ -1343,8 +1534,10 @@ static int drbd_nl_net_conf(struct drbd_tconn *tconn, struct drbd_nl_cfg_req *nl
 	new_conf->on_congestion    = DRBD_ON_CONGESTION_DEF;
 	new_conf->cong_extents     = DRBD_CONG_EXTENTS_DEF;
 
-	if (!net_conf_from_tags(nlp->tag_list, new_conf)) {
+	err = net_conf_from_attrs(new_conf, info->attrs);
+	if (err) {
 		retcode = ERR_MANDATORY_TAG;
+		drbd_msg_put_info(from_attrs_err_to_txt(err));
 		goto fail;
 	}
 
@@ -1495,8 +1688,8 @@ static int drbd_nl_net_conf(struct drbd_tconn *tconn, struct drbd_nl_cfg_req *nl
 		mdev->recv_cnt = 0;
 		kobject_uevent(&disk_to_dev(mdev->vdisk)->kobj, KOBJ_CHANGE);
 	}
-	reply->ret_code = retcode;
 	conn_reconfig_done(tconn);
+	drbd_adm_finish(info, retcode);
 	return 0;
 
 fail:
@@ -1508,24 +1701,37 @@ fail:
 	crypto_free_hash(integrity_r_tfm);
 	kfree(new_conf);
 
-	reply->ret_code = retcode;
 	conn_reconfig_done(tconn);
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-static int drbd_nl_disconnect(struct drbd_tconn *tconn, struct drbd_nl_cfg_req *nlp,
-			      struct drbd_nl_cfg_reply *reply)
+int drbd_adm_disconnect(struct sk_buff *skb, struct genl_info *info)
 {
-	int retcode;
-	struct disconnect dc;
+	struct disconnect_parms parms;
+	struct drbd_tconn *tconn;
+	enum drbd_ret_code retcode;
+	int err;
 
-	memset(&dc, 0, sizeof(struct disconnect));
-	if (!disconnect_from_tags(nlp->tag_list, &dc)) {
-		retcode = ERR_MANDATORY_TAG;
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_CONN);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
 		goto fail;
+
+	tconn = adm_ctx.tconn;
+	memset(&parms, 0, sizeof(parms));
+	if (info->attrs[DRBD_NLA_DISCONNECT_PARMS]) {
+		err = disconnect_parms_from_attrs(&parms, info->attrs);
+		if (err) {
+			retcode = ERR_MANDATORY_TAG;
+			drbd_msg_put_info(from_attrs_err_to_txt(err));
+			goto fail;
+		}
 	}
 
-	if (dc.force) {
+	if (parms.force_disconnect) {
 		spin_lock_irq(&tconn->req_lock);
 		if (tconn->cstate >= C_WF_CONNECTION)
 			_conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD);
@@ -1567,7 +1773,7 @@ static int drbd_nl_disconnect(struct drbd_tconn *tconn, struct drbd_nl_cfg_req *
  done:
 	retcode = NO_ERROR;
  fail:
-	reply->ret_code = retcode;
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
@@ -1587,20 +1793,32 @@ void resync_after_online_grow(struct drbd_conf *mdev)
 		_drbd_request_state(mdev, NS(conn, C_WF_SYNC_UUID), CS_VERBOSE + CS_SERIALIZE);
 }
 
-static int drbd_nl_resize(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			  struct drbd_nl_cfg_reply *reply)
+int drbd_adm_resize(struct sk_buff *skb, struct genl_info *info)
 {
-	struct resize rs;
-	int retcode = NO_ERROR;
+	struct resize_parms rs;
+	struct drbd_conf *mdev;
+	enum drbd_ret_code retcode;
 	enum determine_dev_size dd;
 	enum dds_flags ddsf;
+	int err;
 
-	memset(&rs, 0, sizeof(struct resize));
-	if (!resize_from_tags(nlp->tag_list, &rs)) {
-		retcode = ERR_MANDATORY_TAG;
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
 		goto fail;
+
+	memset(&rs, 0, sizeof(struct resize_parms));
+	if (info->attrs[DRBD_NLA_RESIZE_PARMS]) {
+		err = resize_parms_from_attrs(&rs, info->attrs);
+		if (err) {
+			retcode = ERR_MANDATORY_TAG;
+			drbd_msg_put_info(from_attrs_err_to_txt(err));
+			goto fail;
+		}
 	}
 
+	mdev = adm_ctx.mdev;
 	if (mdev->state.conn > C_CONNECTED) {
 		retcode = ERR_RESIZE_RESYNC;
 		goto fail;
@@ -1644,14 +1862,14 @@ static int drbd_nl_resize(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
 	}
 
  fail:
-	reply->ret_code = retcode;
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-static int drbd_nl_syncer_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			       struct drbd_nl_cfg_reply *reply)
+int drbd_adm_syncer(struct sk_buff *skb, struct genl_info *info)
 {
-	int retcode = NO_ERROR;
+	struct drbd_conf *mdev;
+	enum drbd_ret_code retcode;
 	int err;
 	int ovr; /* online verify running */
 	int rsr; /* re-sync running */
@@ -1662,12 +1880,21 @@ static int drbd_nl_syncer_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *n
 	int *rs_plan_s = NULL;
 	int fifo_size;
 
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto fail;
+	mdev = adm_ctx.mdev;
+
 	if (!zalloc_cpumask_var(&new_cpu_mask, GFP_KERNEL)) {
 		retcode = ERR_NOMEM;
+		drbd_msg_put_info("unable to allocate cpumask");
 		goto fail;
 	}
 
-	if (nlp->flags & DRBD_NL_SET_DEFAULTS) {
+	if (((struct drbd_genlmsghdr*)info->userhdr)->flags
+			& DRBD_GENL_F_SET_DEFAULTS) {
 		memset(&sc, 0, sizeof(struct syncer_conf));
 		sc.rate       = DRBD_RATE_DEF;
 		sc.after      = DRBD_AFTER_DEF;
@@ -1681,8 +1908,10 @@ static int drbd_nl_syncer_conf(struct drbd_conf *mdev, struct drbd_nl_cfg_req *n
 	} else
 		memcpy(&sc, &mdev->sync_conf, sizeof(struct syncer_conf));
 
-	if (!syncer_conf_from_tags(nlp->tag_list, &sc)) {
+	err = syncer_conf_from_attrs(&sc, info->attrs);
+	if (err) {
 		retcode = ERR_MANDATORY_TAG;
+		drbd_msg_put_info(from_attrs_err_to_txt(err));
 		goto fail;
 	}
 
@@ -1832,14 +2061,23 @@ fail:
 	free_cpumask_var(new_cpu_mask);
 	crypto_free_hash(csums_tfm);
 	crypto_free_hash(verify_tfm);
-	reply->ret_code = retcode;
+
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-static int drbd_nl_invalidate(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			      struct drbd_nl_cfg_reply *reply)
+int drbd_adm_invalidate(struct sk_buff *skb, struct genl_info *info)
 {
-	int retcode;
+	struct drbd_conf *mdev;
+	enum drbd_ret_code retcode;
+
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
+
+	mdev = adm_ctx.mdev;
 
 	/* If there is still bitmap IO pending, probably because of a previous
 	 * resync just being finished, wait for it before requesting a new resync. */
@@ -1862,7 +2100,8 @@ static int drbd_nl_invalidate(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nl
 		retcode = drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_T));
 	}
 
-	reply->ret_code = retcode;
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
@@ -1875,56 +2114,58 @@ static int drbd_bmio_set_susp_al(struct drbd_conf *mdev)
 	return rv;
 }
 
-static int drbd_nl_invalidate_peer(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-				   struct drbd_nl_cfg_reply *reply)
+static int drbd_adm_simple_request_state(struct sk_buff *skb, struct genl_info *info,
+		union drbd_state mask, union drbd_state val)
 {
-	int retcode;
-
-	/* If there is still bitmap IO pending, probably because of a previous
-	 * resync just being finished, wait for it before requesting a new resync. */
-	wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags));
+	enum drbd_ret_code retcode;
 
-	retcode = _drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_S), CS_ORDERED);
-
-	if (retcode < SS_SUCCESS) {
-		if (retcode == SS_NEED_CONNECTION && mdev->state.role == R_PRIMARY) {
-			/* The peer will get a resync upon connect anyways. Just make that
-			   into a full resync. */
-			retcode = drbd_request_state(mdev, NS(pdsk, D_INCONSISTENT));
-			if (retcode >= SS_SUCCESS) {
-				if (drbd_bitmap_io(mdev, &drbd_bmio_set_susp_al,
-					"set_n_write from invalidate_peer",
-					BM_LOCKED_SET_ALLOWED))
-					retcode = ERR_IO_MD_DISK;
-			}
-		} else
-			retcode = drbd_request_state(mdev, NS(conn, C_STARTING_SYNC_S));
-	}
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
-	reply->ret_code = retcode;
+	retcode = drbd_request_state(adm_ctx.mdev, mask, val);
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-static int drbd_nl_pause_sync(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			      struct drbd_nl_cfg_reply *reply)
+int drbd_adm_invalidate_peer(struct sk_buff *skb, struct genl_info *info)
 {
-	int retcode = NO_ERROR;
+	return drbd_adm_simple_request_state(skb, info, NS(conn, C_STARTING_SYNC_S));
+}
 
-	if (drbd_request_state(mdev, NS(user_isp, 1)) == SS_NOTHING_TO_DO)
-		retcode = ERR_PAUSE_IS_SET;
+int drbd_adm_pause_sync(struct sk_buff *skb, struct genl_info *info)
+{
+	enum drbd_ret_code retcode;
+
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
-	reply->ret_code = retcode;
+	if (drbd_request_state(adm_ctx.mdev, NS(user_isp, 1)) == SS_NOTHING_TO_DO)
+		retcode = ERR_PAUSE_IS_SET;
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-static int drbd_nl_resume_sync(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			       struct drbd_nl_cfg_reply *reply)
+int drbd_adm_resume_sync(struct sk_buff *skb, struct genl_info *info)
 {
-	int retcode = NO_ERROR;
 	union drbd_state s;
+	enum drbd_ret_code retcode;
+
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
-	if (drbd_request_state(mdev, NS(user_isp, 0)) == SS_NOTHING_TO_DO) {
-		s = mdev->state;
+	if (drbd_request_state(adm_ctx.mdev, NS(user_isp, 0)) == SS_NOTHING_TO_DO) {
+		s = adm_ctx.mdev->state;
 		if (s.conn == C_PAUSED_SYNC_S || s.conn == C_PAUSED_SYNC_T) {
 			retcode = s.aftr_isp ? ERR_PIC_AFTER_DEP :
 				  s.peer_isp ? ERR_PIC_PEER_DEP : ERR_PAUSE_IS_CLEAR;
@@ -1933,28 +2174,35 @@ static int drbd_nl_resume_sync(struct drbd_conf *mdev, struct drbd_nl_cfg_req *n
 		}
 	}
 
-	reply->ret_code = retcode;
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-static int drbd_nl_suspend_io(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			      struct drbd_nl_cfg_reply *reply)
+int drbd_adm_suspend_io(struct sk_buff *skb, struct genl_info *info)
 {
-	reply->ret_code = drbd_request_state(mdev, NS(susp, 1));
-
-	return 0;
+	return drbd_adm_simple_request_state(skb, info, NS(susp, 1));
 }
 
-static int drbd_nl_resume_io(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			     struct drbd_nl_cfg_reply *reply)
+int drbd_adm_resume_io(struct sk_buff *skb, struct genl_info *info)
 {
+	struct drbd_conf *mdev;
+	enum drbd_ret_code retcode;
+
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
+
+	mdev = adm_ctx.mdev;
 	if (test_bit(NEW_CUR_UUID, &mdev->flags)) {
 		drbd_uuid_new_current(mdev);
 		clear_bit(NEW_CUR_UUID, &mdev->flags);
 	}
 	drbd_suspend_io(mdev);
-	reply->ret_code = drbd_request_state(mdev, NS3(susp, 0, susp_nod, 0, susp_fen, 0));
-	if (reply->ret_code == SS_SUCCESS) {
+	retcode = drbd_request_state(mdev, NS3(susp, 0, susp_nod, 0, susp_fen, 0));
+	if (retcode == SS_SUCCESS) {
 		if (mdev->state.conn < C_CONNECTED)
 			tl_clear(mdev->tconn);
 		if (mdev->state.disk == D_DISKLESS || mdev->state.disk == D_FAILED)
@@ -1962,138 +2210,261 @@ static int drbd_nl_resume_io(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp
 	}
 	drbd_resume_io(mdev);
 
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-static int drbd_nl_outdate(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			   struct drbd_nl_cfg_reply *reply)
+int drbd_adm_outdate(struct sk_buff *skb, struct genl_info *info)
 {
-	reply->ret_code = drbd_request_state(mdev, NS(disk, D_OUTDATED));
-	return 0;
+	return drbd_adm_simple_request_state(skb, info, NS(disk, D_OUTDATED));
 }
 
-static int drbd_nl_get_config(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			   struct drbd_nl_cfg_reply *reply)
+int nla_put_status_info(struct sk_buff *skb, struct drbd_conf *mdev,
+		const struct sib_info *sib)
 {
-	unsigned short *tl;
-
-	tl = reply->tag_list;
-
-	if (get_ldev(mdev)) {
-		tl = disk_conf_to_tags(&mdev->ldev->dc, tl);
-		put_ldev(mdev);
+	struct state_info *si = NULL; /* for sizeof(si->member); */
+	struct nlattr *nla;
+	int got_ldev;
+	int got_net;
+	int err = 0;
+	int exclude_sensitive;
+
+	/* If sib != NULL, this is drbd_bcast_event, which anyone can listen
+	 * to.  So we better exclude_sensitive information.
+	 *
+	 * If sib == NULL, this is drbd_adm_get_status, executed synchronously
+	 * in the context of the requesting user process. Exclude sensitive
+	 * information, unless current has superuser.
+	 *
+	 * NOTE: for drbd_adm_get_status_all(), this is a netlink dump, and
+	 * relies on the current implementation of netlink_dump(), which
+	 * executes the dump callback successively from netlink_recvmsg(),
+	 * always in the context of the receiving process */
+	exclude_sensitive = sib || !capable(CAP_SYS_ADMIN);
+
+	got_ldev = get_ldev(mdev);
+	got_net = get_net_conf(mdev->tconn);
+
+	/* We need to add connection name and volume number information still.
+	 * Minor number is in drbd_genlmsghdr. */
+	nla = nla_nest_start(skb, DRBD_NLA_CFG_CONTEXT);
+	if (!nla)
+		goto nla_put_failure;
+	NLA_PUT_U32(skb, T_ctx_volume, mdev->vnr);
+	NLA_PUT_STRING(skb, T_ctx_conn_name, mdev->tconn->name);
+	nla_nest_end(skb, nla);
+
+	if (got_ldev)
+		if (disk_conf_to_skb(skb, &mdev->ldev->dc, exclude_sensitive))
+			goto nla_put_failure;
+	if (got_net)
+		if (net_conf_to_skb(skb, mdev->tconn->net_conf, exclude_sensitive))
+			goto nla_put_failure;
+
+	if (syncer_conf_to_skb(skb, &mdev->sync_conf, exclude_sensitive))
+			goto nla_put_failure;
+
+	nla = nla_nest_start(skb, DRBD_NLA_STATE_INFO);
+	if (!nla)
+		goto nla_put_failure;
+	NLA_PUT_U32(skb, T_sib_reason, sib ? sib->sib_reason : SIB_GET_STATUS_REPLY);
+	NLA_PUT_U32(skb, T_current_state, mdev->state.i);
+	NLA_PUT_U64(skb, T_ed_uuid, mdev->ed_uuid);
+	NLA_PUT_U64(skb, T_capacity, drbd_get_capacity(mdev->this_bdev));
+
+	if (got_ldev) {
+		NLA_PUT_U32(skb, T_disk_flags, mdev->ldev->md.flags);
+		NLA_PUT(skb, T_uuids, sizeof(si->uuids), mdev->ldev->md.uuid);
+		NLA_PUT_U64(skb, T_bits_total, drbd_bm_bits(mdev));
+		NLA_PUT_U64(skb, T_bits_oos, drbd_bm_total_weight(mdev));
+		if (C_SYNC_SOURCE <= mdev->state.conn &&
+		    C_PAUSED_SYNC_T >= mdev->state.conn) {
+			NLA_PUT_U64(skb, T_bits_rs_total, mdev->rs_total);
+			NLA_PUT_U64(skb, T_bits_rs_failed, mdev->rs_failed);
+		}
 	}
 
-	if (get_net_conf(mdev->tconn)) {
-		tl = net_conf_to_tags(mdev->tconn->net_conf, tl);
-		put_net_conf(mdev->tconn);
+	if (sib) {
+		switch(sib->sib_reason) {
+		case SIB_SYNC_PROGRESS:
+		case SIB_GET_STATUS_REPLY:
+			break;
+		case SIB_STATE_CHANGE:
+			NLA_PUT_U32(skb, T_prev_state, sib->os.i);
+			NLA_PUT_U32(skb, T_new_state, sib->ns.i);
+			break;
+		case SIB_HELPER_POST:
+			NLA_PUT_U32(skb,
+				T_helper_exit_code, sib->helper_exit_code);
+			/* fall through */
+		case SIB_HELPER_PRE:
+			NLA_PUT_STRING(skb, T_helper, sib->helper_name);
+			break;
+		}
 	}
-	tl = syncer_conf_to_tags(&mdev->sync_conf, tl);
-
-	put_unaligned(TT_END, tl++); /* Close the tag list */
+	nla_nest_end(skb, nla);
 
-	return (int)((char *)tl - (char *)reply->tag_list);
+	if (0)
+nla_put_failure:
+		err = -EMSGSIZE;
+	if (got_ldev)
+		put_ldev(mdev);
+	if (got_net)
+		put_net_conf(mdev->tconn);
+	return err;
 }
 
-static int drbd_nl_get_state(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			     struct drbd_nl_cfg_reply *reply)
+int drbd_adm_get_status(struct sk_buff *skb, struct genl_info *info)
 {
-	unsigned short *tl = reply->tag_list;
-	union drbd_state s = mdev->state;
-	unsigned long rs_left;
-	unsigned int res;
+	enum drbd_ret_code retcode;
+	int err;
 
-	tl = get_state_to_tags((struct get_state *)&s, tl);
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
-	/* no local ref, no bitmap, no syncer progress. */
-	if (s.conn >= C_SYNC_SOURCE && s.conn <= C_PAUSED_SYNC_T) {
-		if (get_ldev(mdev)) {
-			drbd_get_syncer_progress(mdev, &rs_left, &res);
-			tl = tl_add_int(tl, T_sync_progress, &res);
-			put_ldev(mdev);
-		}
+	err = nla_put_status_info(adm_ctx.reply_skb, adm_ctx.mdev, NULL);
+	if (err) {
+		nlmsg_free(adm_ctx.reply_skb);
+		return err;
 	}
-	put_unaligned(TT_END, tl++); /* Close the tag list */
-
-	return (int)((char *)tl - (char *)reply->tag_list);
+out:
+	drbd_adm_finish(info, retcode);
+	return 0;
 }
 
-static int drbd_nl_get_uuids(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			     struct drbd_nl_cfg_reply *reply)
+int drbd_adm_get_status_all(struct sk_buff *skb, struct netlink_callback *cb)
 {
-	unsigned short *tl;
-
-	tl = reply->tag_list;
+	struct drbd_conf *mdev;
+	struct drbd_genlmsghdr *dh;
+	int minor = cb->args[0];
+
+	/* Open coded deferred single idr_for_each_entry iteration.
+	 * This may miss entries inserted after this dump started,
+	 * or entries deleted before they are reached.
+	 * But we need to make sure the mdev won't disappear while
+	 * we are looking at it. */
+
+	rcu_read_lock();
+	mdev = idr_get_next(&minors, &minor);
+	if (mdev) {
+		dh = genlmsg_put(skb, NETLINK_CB(cb->skb).pid,
+				cb->nlh->nlmsg_seq, &drbd_genl_family,
+				NLM_F_MULTI, DRBD_ADM_GET_STATUS);
+		if (!dh)
+			goto errout;
+
+		D_ASSERT(mdev->minor == minor);
+
+		dh->minor = minor;
+		dh->ret_code = NO_ERROR;
+
+		pr_info("dump: minor=%u, conn=%s[%u]\n",
+			dh->minor, mdev->tconn->name, mdev->vnr);
+		if (nla_put_status_info(skb, mdev, NULL)) {
+			genlmsg_cancel(skb, dh);
+			goto errout;
+		}
+		genlmsg_end(skb, dh);
+        }
 
-	if (get_ldev(mdev)) {
-		tl = tl_add_blob(tl, T_uuids, mdev->ldev->md.uuid, UI_SIZE*sizeof(u64));
-		tl = tl_add_int(tl, T_uuids_flags, &mdev->ldev->md.flags);
-		put_ldev(mdev);
-	}
-	put_unaligned(TT_END, tl++); /* Close the tag list */
+errout:
+	rcu_read_unlock();
+	/* where to start idr_get_next with the next iteration */
+        cb->args[0] = minor+1;
 
-	return (int)((char *)tl - (char *)reply->tag_list);
+	/* No more minors found: empty skb. Which will terminate the dump. */
+        return skb->len;
 }
 
-/**
- * drbd_nl_get_timeout_flag() - Used by drbdsetup to find out which timeout value to use
- * @mdev:	DRBD device.
- * @nlp:	Netlink/connector packet from drbdsetup
- * @reply:	Reply packet for drbdsetup
- */
-static int drbd_nl_get_timeout_flag(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-				    struct drbd_nl_cfg_reply *reply)
+int drbd_adm_get_timeout_type(struct sk_buff *skb, struct genl_info *info)
 {
-	unsigned short *tl;
-	char rv;
-
-	tl = reply->tag_list;
+	enum drbd_ret_code retcode;
+	struct timeout_parms tp;
+	int err;
 
-	rv = mdev->state.pdsk == D_OUTDATED        ? UT_PEER_OUTDATED :
-	  test_bit(USE_DEGR_WFC_T, &mdev->flags) ? UT_DEGRADED : UT_DEFAULT;
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
-	tl = tl_add_blob(tl, T_use_degraded, &rv, sizeof(rv));
-	put_unaligned(TT_END, tl++); /* Close the tag list */
+	tp.timeout_type =
+		adm_ctx.mdev->state.pdsk == D_OUTDATED ? UT_PEER_OUTDATED :
+		test_bit(USE_DEGR_WFC_T, &adm_ctx.mdev->flags) ? UT_DEGRADED :
+		UT_DEFAULT;
 
-	return (int)((char *)tl - (char *)reply->tag_list);
+	err = timeout_parms_to_priv_skb(adm_ctx.reply_skb, &tp);
+	if (err) {
+		nlmsg_free(adm_ctx.reply_skb);
+		return err;
+	}
+out:
+	drbd_adm_finish(info, retcode);
+	return 0;
 }
 
-static int drbd_nl_start_ov(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-				    struct drbd_nl_cfg_reply *reply)
+int drbd_adm_start_ov(struct sk_buff *skb, struct genl_info *info)
 {
-	/* default to resume from last known position, if possible */
-	struct start_ov args =
-		{ .start_sector = mdev->ov_start_sector };
+	struct drbd_conf *mdev;
+	enum drbd_ret_code retcode;
 
-	if (!start_ov_from_tags(nlp->tag_list, &args)) {
-		reply->ret_code = ERR_MANDATORY_TAG;
-		return 0;
-	}
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
+	mdev = adm_ctx.mdev;
+	if (info->attrs[DRBD_NLA_START_OV_PARMS]) {
+		/* resume from last known position, if possible */
+		struct start_ov_parms parms =
+			{ .ov_start_sector = mdev->ov_start_sector };
+		int err = start_ov_parms_from_attrs(&parms, info->attrs);
+		if (err) {
+			retcode = ERR_MANDATORY_TAG;
+			drbd_msg_put_info(from_attrs_err_to_txt(err));
+			goto out;
+		}
+		/* w_make_ov_request expects position to be aligned */
+		mdev->ov_start_sector = parms.ov_start_sector & ~BM_SECT_PER_BIT;
+	}
 	/* If there is still bitmap IO pending, e.g. previous resync or verify
 	 * just being finished, wait for it before requesting a new resync. */
 	wait_event(mdev->misc_wait, !test_bit(BITMAP_IO, &mdev->flags));
-
-	/* w_make_ov_request expects position to be aligned */
-	mdev->ov_start_sector = args.start_sector & ~BM_SECT_PER_BIT;
-	reply->ret_code = drbd_request_state(mdev,NS(conn,C_VERIFY_S));
+	retcode = drbd_request_state(mdev,NS(conn,C_VERIFY_S));
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
 
-static int drbd_nl_new_c_uuid(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			      struct drbd_nl_cfg_reply *reply)
+int drbd_adm_new_c_uuid(struct sk_buff *skb, struct genl_info *info)
 {
-	int retcode = NO_ERROR;
+	struct drbd_conf *mdev;
+	enum drbd_ret_code retcode;
 	int skip_initial_sync = 0;
 	int err;
+	struct new_c_uuid_parms args;
 
-	struct new_c_uuid args;
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out_nolock;
 
-	memset(&args, 0, sizeof(struct new_c_uuid));
-	if (!new_c_uuid_from_tags(nlp->tag_list, &args)) {
-		reply->ret_code = ERR_MANDATORY_TAG;
-		return 0;
+	mdev = adm_ctx.mdev;
+	memset(&args, 0, sizeof(args));
+	if (info->attrs[DRBD_NLA_NEW_C_UUID_PARMS]) {
+		err = new_c_uuid_parms_from_attrs(&args, info->attrs);
+		if (err) {
+			retcode = ERR_MANDATORY_TAG;
+			drbd_msg_put_info(from_attrs_err_to_txt(err));
+			goto out_nolock;
+		}
 	}
 
 	mutex_lock(mdev->state_mutex); /* Protects us against serialized state changes. */
@@ -2139,510 +2510,164 @@ out_dec:
 	put_ldev(mdev);
 out:
 	mutex_unlock(mdev->state_mutex);
-
-	reply->ret_code = retcode;
-	return 0;
-}
-
-static int drbd_nl_new_conn(struct drbd_nl_cfg_req *nlp, struct drbd_nl_cfg_reply *reply)
-{
-	struct new_connection args;
-
-	if (!new_connection_from_tags(nlp->tag_list, &args)) {
-		reply->ret_code = ERR_MANDATORY_TAG;
-		return 0;
-	}
-
-	reply->ret_code = NO_ERROR;
-	if (!drbd_new_tconn(args.name))
-		reply->ret_code = ERR_NOMEM;
-
-	return 0;
-}
-
-static int drbd_nl_new_minor(struct drbd_tconn *tconn,
-		      struct drbd_nl_cfg_req *nlp, struct drbd_nl_cfg_reply *reply)
-{
-	struct new_minor args;
-
-	args.vol_nr = 0;
-	args.minor = 0;
-
-	if (!new_minor_from_tags(nlp->tag_list, &args)) {
-		reply->ret_code = ERR_MANDATORY_TAG;
-		return 0;
-	}
-
-	reply->ret_code = conn_new_minor(tconn, args.minor, args.vol_nr);
-
+out_nolock:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-static int drbd_nl_del_minor(struct drbd_conf *mdev, struct drbd_nl_cfg_req *nlp,
-			     struct drbd_nl_cfg_reply *reply)
+static enum drbd_ret_code
+drbd_check_conn_name(const char *name)
 {
-	if (mdev->state.disk == D_DISKLESS &&
-	    mdev->state.conn == C_STANDALONE &&
-	    mdev->state.role == R_SECONDARY) {
-		drbd_delete_device(mdev_to_minor(mdev));
-		reply->ret_code = NO_ERROR;
-	} else {
-		reply->ret_code = ERR_MINOR_CONFIGURED;
+	if (!name || !name[0]) {
+		drbd_msg_put_info("connection name missing");
+		return ERR_MANDATORY_TAG;
 	}
-	return 0;
-}
-
-static int drbd_nl_del_conn(struct drbd_tconn *tconn,
-			    struct drbd_nl_cfg_req *nlp, struct drbd_nl_cfg_reply *reply)
-{
-	if (conn_lowest_minor(tconn) < 0) {
-		drbd_free_tconn(tconn);
-		reply->ret_code = NO_ERROR;
-	} else {
-		reply->ret_code = ERR_CONN_IN_USE;
+	/* if we want to use these in sysfs/configfs/debugfs some day,
+	 * we must not allow slashes */
+	if (strchr(name, '/')) {
+		drbd_msg_put_info("invalid connection name");
+		return ERR_INVALID_REQUEST;
 	}
-
-	return 0;
+	return NO_ERROR;
 }
 
-enum cn_handler_type {
-	CHT_MINOR,
-	CHT_CONN,
-	CHT_CTOR,
-	/* CHT_RES, later */
-};
-struct cn_handler_struct {
-	enum cn_handler_type type;
-	union {
-		int (*minor_based)(struct drbd_conf *,
-				   struct drbd_nl_cfg_req *,
-				   struct drbd_nl_cfg_reply *);
-		int (*conn_based)(struct drbd_tconn *,
-				  struct drbd_nl_cfg_req *,
-				  struct drbd_nl_cfg_reply *);
-		int (*constructor)(struct drbd_nl_cfg_req *,
-				   struct drbd_nl_cfg_reply *);
-	};
-	int reply_body_size;
-};
-
-static struct cn_handler_struct cnd_table[] = {
-	[ P_primary ]		= { CHT_MINOR, { &drbd_nl_primary },	0 },
-	[ P_secondary ]		= { CHT_MINOR, { &drbd_nl_secondary },	0 },
-	[ P_disk_conf ]		= { CHT_MINOR, { &drbd_nl_disk_conf },	0 },
-	[ P_detach ]		= { CHT_MINOR, { &drbd_nl_detach },	0 },
-	[ P_net_conf ]		= { CHT_CONN,  { .conn_based = &drbd_nl_net_conf },	0 },
-	[ P_disconnect ]	= { CHT_CONN,  { .conn_based = &drbd_nl_disconnect },	0 },
-	[ P_resize ]		= { CHT_MINOR, { &drbd_nl_resize },	0 },
-	[ P_syncer_conf ]	= { CHT_MINOR, { &drbd_nl_syncer_conf },0 },
-	[ P_invalidate ]	= { CHT_MINOR, { &drbd_nl_invalidate },	0 },
-	[ P_invalidate_peer ]	= { CHT_MINOR, { &drbd_nl_invalidate_peer },0 },
-	[ P_pause_sync ]	= { CHT_MINOR, { &drbd_nl_pause_sync },	0 },
-	[ P_resume_sync ]	= { CHT_MINOR, { &drbd_nl_resume_sync },0 },
-	[ P_suspend_io ]	= { CHT_MINOR, { &drbd_nl_suspend_io },	0 },
-	[ P_resume_io ]		= { CHT_MINOR, { &drbd_nl_resume_io },	0 },
-	[ P_outdate ]		= { CHT_MINOR, { &drbd_nl_outdate },	0 },
-	[ P_get_config ]	= { CHT_MINOR, { &drbd_nl_get_config },
-				    sizeof(struct syncer_conf_tag_len_struct) +
-				    sizeof(struct disk_conf_tag_len_struct) +
-				    sizeof(struct net_conf_tag_len_struct) },
-	[ P_get_state ]		= { CHT_MINOR, { &drbd_nl_get_state },
-				    sizeof(struct get_state_tag_len_struct) +
-				    sizeof(struct sync_progress_tag_len_struct)	},
-	[ P_get_uuids ]		= { CHT_MINOR, { &drbd_nl_get_uuids },
-				    sizeof(struct get_uuids_tag_len_struct) },
-	[ P_get_timeout_flag ]	= { CHT_MINOR, { &drbd_nl_get_timeout_flag },
-				    sizeof(struct get_timeout_flag_tag_len_struct)},
-	[ P_start_ov ]		= { CHT_MINOR, { &drbd_nl_start_ov },	0 },
-	[ P_new_c_uuid ]	= { CHT_MINOR, { &drbd_nl_new_c_uuid },	0 },
-	[ P_new_connection ]	= { CHT_CTOR,  { .constructor = &drbd_nl_new_conn }, 0 },
-	[ P_new_minor ]		= { CHT_CONN,  { .conn_based = &drbd_nl_new_minor }, 0 },
-	[ P_del_minor ]		= { CHT_MINOR, { &drbd_nl_del_minor },	0 },
-	[ P_del_connection ]    = { CHT_CONN,  { .conn_based = &drbd_nl_del_conn }, 0 },
-};
-
-static void drbd_connector_callback(struct cn_msg *req, struct netlink_skb_parms *nsp)
+int drbd_adm_create_connection(struct sk_buff *skb, struct genl_info *info)
 {
-	struct drbd_nl_cfg_req *nlp = (struct drbd_nl_cfg_req *)req->data;
-	struct cn_handler_struct *cm;
-	struct cn_msg *cn_reply;
-	struct drbd_nl_cfg_reply *reply;
-	struct drbd_conf *mdev;
-	struct drbd_tconn *tconn;
-	int retcode, rr;
-	int reply_size = sizeof(struct cn_msg)
-		+ sizeof(struct drbd_nl_cfg_reply)
-		+ sizeof(short int);
-
-	if (!try_module_get(THIS_MODULE)) {
-		printk(KERN_ERR "drbd: try_module_get() failed!\n");
-		return;
-	}
-
-	if (!cap_raised(current_cap(), CAP_SYS_ADMIN)) {
-		retcode = ERR_PERM;
-		goto fail;
-	}
+	enum drbd_ret_code retcode;
 
-	if (nlp->packet_type >= P_nl_after_last_packet ||
-	    nlp->packet_type == P_return_code_only) {
-		retcode = ERR_PACKET_NR;
-		goto fail;
-	}
+	retcode = drbd_adm_prepare(skb, info, 0);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
-	cm = cnd_table + nlp->packet_type;
+	retcode = drbd_check_conn_name(adm_ctx.conn_name);
+	if (retcode != NO_ERROR)
+		goto out;
 
-	/* This may happen if packet number is 0: */
-	if (cm->minor_based == NULL) {
-		retcode = ERR_PACKET_NR;
-		goto fail;
+	if (adm_ctx.tconn) {
+		retcode = ERR_INVALID_REQUEST;
+		drbd_msg_put_info("connection exists");
+		goto out;
 	}
 
-	reply_size += cm->reply_body_size;
-
-	/* allocation not in the IO path, cqueue thread context */
-	cn_reply = kzalloc(reply_size, GFP_KERNEL);
-	if (!cn_reply) {
+	if (!drbd_new_tconn(adm_ctx.conn_name))
 		retcode = ERR_NOMEM;
-		goto fail;
-	}
-	reply = (struct drbd_nl_cfg_reply *) cn_reply->data;
-
-	reply->packet_type =
-		cm->reply_body_size ? nlp->packet_type : P_return_code_only;
-	reply->minor = nlp->drbd_minor;
-	reply->ret_code = NO_ERROR; /* Might by modified by cm->function. */
-	/* reply->tag_list; might be modified by cm->function. */
-
-	retcode = ERR_MINOR_INVALID;
-	rr = 0;
-	switch (cm->type) {
-	case CHT_MINOR:
-		mdev = minor_to_mdev(nlp->drbd_minor);
-		if (!mdev)
-			goto fail;
-		rr = cm->minor_based(mdev, nlp, reply);
-		break;
-	case CHT_CONN:
-		tconn = conn_by_name(nlp->obj_name);
-		if (!tconn) {
-			retcode = ERR_CONN_NOT_KNOWN;
-			goto fail;
-		}
-		rr = cm->conn_based(tconn, nlp, reply);
-		break;
-	case CHT_CTOR:
-		rr = cm->constructor(nlp, reply);
-		break;
-	/* case CHT_RES: */
-	}
-
-	cn_reply->id = req->id;
-	cn_reply->seq = req->seq;
-	cn_reply->ack = req->ack  + 1;
-	cn_reply->len = sizeof(struct drbd_nl_cfg_reply) + rr;
-	cn_reply->flags = 0;
-
-	rr = cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_KERNEL);
-	if (rr && rr != -ESRCH)
-		printk(KERN_INFO "drbd: cn_netlink_send()=%d\n", rr);
-
-	kfree(cn_reply);
-	module_put(THIS_MODULE);
-	return;
- fail:
-	drbd_nl_send_reply(req, retcode);
-	module_put(THIS_MODULE);
-}
-
-static atomic_t drbd_nl_seq = ATOMIC_INIT(2); /* two. */
-
-static unsigned short *
-__tl_add_blob(unsigned short *tl, enum drbd_tags tag, const void *data,
-	unsigned short len, int nul_terminated)
-{
-	unsigned short l = tag_descriptions[tag_number(tag)].max_len;
-	len = (len < l) ? len :  l;
-	put_unaligned(tag, tl++);
-	put_unaligned(len, tl++);
-	memcpy(tl, data, len);
-	tl = (unsigned short*)((char*)tl + len);
-	if (nul_terminated)
-		*((char*)tl - 1) = 0;
-	return tl;
-}
-
-static unsigned short *
-tl_add_blob(unsigned short *tl, enum drbd_tags tag, const void *data, int len)
-{
-	return __tl_add_blob(tl, tag, data, len, 0);
-}
-
-static unsigned short *
-tl_add_str(unsigned short *tl, enum drbd_tags tag, const char *str)
-{
-	return __tl_add_blob(tl, tag, str, strlen(str)+1, 0);
-}
-
-static unsigned short *
-tl_add_int(unsigned short *tl, enum drbd_tags tag, const void *val)
-{
-	put_unaligned(tag, tl++);
-	switch(tag_type(tag)) {
-	case TT_INTEGER:
-		put_unaligned(sizeof(int), tl++);
-		put_unaligned(*(int *)val, (int *)tl);
-		tl = (unsigned short*)((char*)tl+sizeof(int));
-		break;
-	case TT_INT64:
-		put_unaligned(sizeof(u64), tl++);
-		put_unaligned(*(u64 *)val, (u64 *)tl);
-		tl = (unsigned short*)((char*)tl+sizeof(u64));
-		break;
-	default:
-		/* someone did something stupid. */
-		;
-	}
-	return tl;
-}
-
-void drbd_bcast_state(struct drbd_conf *mdev, union drbd_state state)
-{
-	char buffer[sizeof(struct cn_msg)+
-		    sizeof(struct drbd_nl_cfg_reply)+
-		    sizeof(struct get_state_tag_len_struct)+
-		    sizeof(short int)];
-	struct cn_msg *cn_reply = (struct cn_msg *) buffer;
-	struct drbd_nl_cfg_reply *reply =
-		(struct drbd_nl_cfg_reply *)cn_reply->data;
-	unsigned short *tl = reply->tag_list;
-
-	/* dev_warn(DEV, "drbd_bcast_state() got called\n"); */
-
-	tl = get_state_to_tags((struct get_state *)&state, tl);
-
-	put_unaligned(TT_END, tl++); /* Close the tag list */
-
-	cn_reply->id.idx = CN_IDX_DRBD;
-	cn_reply->id.val = CN_VAL_DRBD;
-
-	cn_reply->seq = atomic_inc_return(&drbd_nl_seq);
-	cn_reply->ack = 0; /* not used here. */
-	cn_reply->len = sizeof(struct drbd_nl_cfg_reply) +
-		(int)((char *)tl - (char *)reply->tag_list);
-	cn_reply->flags = 0;
-
-	reply->packet_type = P_get_state;
-	reply->minor = mdev_to_minor(mdev);
-	reply->ret_code = NO_ERROR;
-
-	cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO);
-}
-
-void drbd_bcast_ev_helper(struct drbd_conf *mdev, char *helper_name)
-{
-	char buffer[sizeof(struct cn_msg)+
-		    sizeof(struct drbd_nl_cfg_reply)+
-		    sizeof(struct call_helper_tag_len_struct)+
-		    sizeof(short int)];
-	struct cn_msg *cn_reply = (struct cn_msg *) buffer;
-	struct drbd_nl_cfg_reply *reply =
-		(struct drbd_nl_cfg_reply *)cn_reply->data;
-	unsigned short *tl = reply->tag_list;
-
-	/* dev_warn(DEV, "drbd_bcast_state() got called\n"); */
-
-	tl = tl_add_str(tl, T_helper, helper_name);
-	put_unaligned(TT_END, tl++); /* Close the tag list */
-
-	cn_reply->id.idx = CN_IDX_DRBD;
-	cn_reply->id.val = CN_VAL_DRBD;
-
-	cn_reply->seq = atomic_inc_return(&drbd_nl_seq);
-	cn_reply->ack = 0; /* not used here. */
-	cn_reply->len = sizeof(struct drbd_nl_cfg_reply) +
-		(int)((char *)tl - (char *)reply->tag_list);
-	cn_reply->flags = 0;
-
-	reply->packet_type = P_call_helper;
-	reply->minor = mdev_to_minor(mdev);
-	reply->ret_code = NO_ERROR;
-
-	cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO);
+out:
+	drbd_adm_finish(info, retcode);
+	return 0;
 }
 
-void drbd_bcast_ee(struct drbd_conf *mdev, const char *reason, const int dgs,
-		   const char *seen_hash, const char *calc_hash,
-			   const struct drbd_peer_request *peer_req)
+int drbd_adm_add_minor(struct sk_buff *skb, struct genl_info *info)
 {
-	struct cn_msg *cn_reply;
-	struct drbd_nl_cfg_reply *reply;
-	unsigned short *tl;
-	struct page *page;
-	unsigned len;
+	struct drbd_genlmsghdr *dh = info->userhdr;
+	enum drbd_ret_code retcode;
 
-	if (!peer_req)
-		return;
-	if (!reason || !reason[0])
-		return;
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_CONN);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
-	/* apparently we have to memcpy twice, first to prepare the data for the
-	 * struct cn_msg, then within cn_netlink_send from the cn_msg to the
-	 * netlink skb. */
-	/* receiver thread context, which is not in the writeout path (of this node),
-	 * but may be in the writeout path of the _other_ node.
-	 * GFP_NOIO to avoid potential "distributed deadlock". */
-	cn_reply = kzalloc(
-		sizeof(struct cn_msg)+
-		sizeof(struct drbd_nl_cfg_reply)+
-		sizeof(struct dump_ee_tag_len_struct)+
-		sizeof(short int),
-		GFP_NOIO);
-
-	if (!cn_reply) {
-		dev_err(DEV, "could not kmalloc buffer for drbd_bcast_ee, "
-			     "sector %llu, size %u\n",
-			(unsigned long long)peer_req->i.sector,
-			peer_req->i.size);
-		return;
+	/* FIXME drop minor_count parameter, limit to MINORMASK */
+	if (dh->minor >= minor_count) {
+		drbd_msg_put_info("requested minor out of range");
+		return ERR_INVALID_REQUEST;
 	}
-
-	reply = (struct drbd_nl_cfg_reply*)cn_reply->data;
-	tl = reply->tag_list;
-
-	tl = tl_add_str(tl, T_dump_ee_reason, reason);
-	tl = tl_add_blob(tl, T_seen_digest, seen_hash, dgs);
-	tl = tl_add_blob(tl, T_calc_digest, calc_hash, dgs);
-	tl = tl_add_int(tl, T_ee_sector, &peer_req->i.sector);
-	tl = tl_add_int(tl, T_ee_block_id, &peer_req->block_id);
-
-	/* dump the first 32k */
-	len = min_t(unsigned, peer_req->i.size, 32 << 10);
-	put_unaligned(T_ee_data, tl++);
-	put_unaligned(len, tl++);
-
-	page = peer_req->pages;
-	page_chain_for_each(page) {
-		void *d = kmap_atomic(page, KM_USER0);
-		unsigned l = min_t(unsigned, len, PAGE_SIZE);
-		memcpy(tl, d, l);
-		kunmap_atomic(d, KM_USER0);
-		tl = (unsigned short*)((char*)tl + l);
-		len -= l;
-		if (len == 0)
-			break;
+	/* FIXME we need a define here */
+	if (adm_ctx.volume >= 256) {
+		drbd_msg_put_info("requested volume id out of range");
+		return ERR_INVALID_REQUEST;
 	}
-	put_unaligned(TT_END, tl++); /* Close the tag list */
-
-	cn_reply->id.idx = CN_IDX_DRBD;
-	cn_reply->id.val = CN_VAL_DRBD;
-
-	cn_reply->seq = atomic_inc_return(&drbd_nl_seq);
-	cn_reply->ack = 0; // not used here.
-	cn_reply->len = sizeof(struct drbd_nl_cfg_reply) +
-		(int)((char*)tl - (char*)reply->tag_list);
-	cn_reply->flags = 0;
 
-	reply->packet_type = P_dump_ee;
-	reply->minor = mdev_to_minor(mdev);
-	reply->ret_code = NO_ERROR;
-
-	cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO);
-	kfree(cn_reply);
+	retcode = conn_new_minor(adm_ctx.tconn, dh->minor, adm_ctx.volume);
+out:
+	drbd_adm_finish(info, retcode);
+	return 0;
 }
 
-void drbd_bcast_sync_progress(struct drbd_conf *mdev)
+int drbd_adm_delete_minor(struct sk_buff *skb, struct genl_info *info)
 {
-	char buffer[sizeof(struct cn_msg)+
-		    sizeof(struct drbd_nl_cfg_reply)+
-		    sizeof(struct sync_progress_tag_len_struct)+
-		    sizeof(short int)];
-	struct cn_msg *cn_reply = (struct cn_msg *) buffer;
-	struct drbd_nl_cfg_reply *reply =
-		(struct drbd_nl_cfg_reply *)cn_reply->data;
-	unsigned short *tl = reply->tag_list;
-	unsigned long rs_left;
-	unsigned int res;
-
-	/* no local ref, no bitmap, no syncer progress, no broadcast. */
-	if (!get_ldev(mdev))
-		return;
-	drbd_get_syncer_progress(mdev, &rs_left, &res);
-	put_ldev(mdev);
-
-	tl = tl_add_int(tl, T_sync_progress, &res);
-	put_unaligned(TT_END, tl++); /* Close the tag list */
-
-	cn_reply->id.idx = CN_IDX_DRBD;
-	cn_reply->id.val = CN_VAL_DRBD;
-
-	cn_reply->seq = atomic_inc_return(&drbd_nl_seq);
-	cn_reply->ack = 0; /* not used here. */
-	cn_reply->len = sizeof(struct drbd_nl_cfg_reply) +
-		(int)((char *)tl - (char *)reply->tag_list);
-	cn_reply->flags = 0;
+	struct drbd_conf *mdev;
+	enum drbd_ret_code retcode;
 
-	reply->packet_type = P_sync_progress;
-	reply->minor = mdev_to_minor(mdev);
-	reply->ret_code = NO_ERROR;
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
 
-	cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO);
+	mdev = adm_ctx.mdev;
+	if (mdev->state.disk == D_DISKLESS &&
+	    mdev->state.conn == C_STANDALONE &&
+	    mdev->state.role == R_SECONDARY) {
+		drbd_delete_device(mdev_to_minor(mdev));
+		retcode = NO_ERROR;
+	} else
+		retcode = ERR_MINOR_CONFIGURED;
+out:
+	drbd_adm_finish(info, retcode);
+	return 0;
 }
 
-int __init drbd_nl_init(void)
+int drbd_adm_delete_connection(struct sk_buff *skb, struct genl_info *info)
 {
-	static struct cb_id cn_id_drbd;
-	int err, try=10;
-
-	cn_id_drbd.val = CN_VAL_DRBD;
-	do {
-		cn_id_drbd.idx = cn_idx;
-		err = cn_add_callback(&cn_id_drbd, "cn_drbd", &drbd_connector_callback);
-		if (!err)
-			break;
-		cn_idx = (cn_idx + CN_IDX_STEP);
-	} while (try--);
+	enum drbd_ret_code retcode;
 
-	if (err) {
-		printk(KERN_ERR "drbd: cn_drbd failed to register\n");
-		return err;
+	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_CONN);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
+
+	if (conn_lowest_minor(adm_ctx.tconn) < 0) {
+		drbd_free_tconn(adm_ctx.tconn);
+		retcode = NO_ERROR;
+	} else {
+		retcode = ERR_CONN_IN_USE;
 	}
 
+out:
+	drbd_adm_finish(info, retcode);
 	return 0;
 }
 
-void drbd_nl_cleanup(void)
+void drbd_bcast_event(struct drbd_conf *mdev, const struct sib_info *sib)
 {
-	static struct cb_id cn_id_drbd;
-
-	cn_id_drbd.idx = cn_idx;
-	cn_id_drbd.val = CN_VAL_DRBD;
+	static atomic_t drbd_genl_seq = ATOMIC_INIT(2); /* two. */
+	struct sk_buff *msg;
+	struct drbd_genlmsghdr *d_out;
+	unsigned seq;
+	int err = -ENOMEM;
+
+	seq = atomic_inc_return(&drbd_genl_seq);
+	msg = genlmsg_new(NLMSG_GOODSIZE, GFP_NOIO);
+	if (!msg)
+		goto failed;
+
+	err = -EMSGSIZE;
+	d_out = genlmsg_put(msg, 0, seq, &drbd_genl_family, 0, DRBD_EVENT);
+	if (!d_out) /* cannot happen, but anyways. */
+		goto nla_put_failure;
+	d_out->minor = mdev_to_minor(mdev);
+	d_out->ret_code = 0;
+
+	pr_info("event: minor=%u, conn=%s\n", d_out->minor, mdev->tconn->name);
+
+	if (nla_put_status_info(msg, mdev, sib))
+		goto nla_put_failure;
+	genlmsg_end(msg, d_out);
+	err = drbd_genl_multicast_events(msg, 0);
+	/* msg has been consumed or freed in netlink_broadcast() */
+	if (err && err != -ESRCH)
+		goto failed;
 
-	cn_del_callback(&cn_id_drbd);
-}
+	return;
 
-void drbd_nl_send_reply(struct cn_msg *req, int ret_code)
-{
-	char buffer[sizeof(struct cn_msg)+sizeof(struct drbd_nl_cfg_reply)];
-	struct cn_msg *cn_reply = (struct cn_msg *) buffer;
-	struct drbd_nl_cfg_reply *reply =
-		(struct drbd_nl_cfg_reply *)cn_reply->data;
-	int rr;
-
-	memset(buffer, 0, sizeof(buffer));
-	cn_reply->id = req->id;
-
-	cn_reply->seq = req->seq;
-	cn_reply->ack = req->ack  + 1;
-	cn_reply->len = sizeof(struct drbd_nl_cfg_reply);
-	cn_reply->flags = 0;
-
-	reply->packet_type = P_return_code_only;
-	reply->minor = ((struct drbd_nl_cfg_req *)req->data)->drbd_minor;
-	reply->ret_code = ret_code;
-
-	rr = cn_netlink_send(cn_reply, CN_IDX_DRBD, GFP_NOIO);
-	if (rr && rr != -ESRCH)
-		printk(KERN_INFO "drbd: cn_netlink_send()=%d\n", rr);
+nla_put_failure:
+	nlmsg_free(msg);
+failed:
+	dev_err(DEV, "Error %d while broadcasting event. "
+			"Event seq:%u sib_reason:%u\n",
+			err, seq, sib->sib_reason);
 }
-
diff --git a/drivers/block/drbd/drbd_state.c b/drivers/block/drbd/drbd_state.c
index 100d48b..34be2ef 100644
--- a/drivers/block/drbd/drbd_state.c
+++ b/drivers/block/drbd/drbd_state.c
@@ -971,6 +971,11 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 	enum drbd_fencing_p fp;
 	enum drbd_req_event what = NOTHING;
 	union drbd_state nsm = (union drbd_state){ .i = -1 };
+	struct sib_info sib;
+
+	sib.sib_reason = SIB_STATE_CHANGE;
+	sib.os = os;
+	sib.ns = ns;
 
 	if (os.conn != C_CONNECTED && ns.conn == C_CONNECTED) {
 		clear_bit(CRASHED_PRIMARY, &mdev->flags);
@@ -985,7 +990,7 @@ static void after_state_ch(struct drbd_conf *mdev, union drbd_state os,
 	}
 
 	/* Inform userspace about the change... */
-	drbd_bcast_state(mdev, ns);
+	drbd_bcast_event(mdev, &sib);
 
 	if (!(os.role == R_PRIMARY && os.disk < D_UP_TO_DATE && os.pdsk < D_UP_TO_DATE) &&
 	    (ns.role == R_PRIMARY && ns.disk < D_UP_TO_DATE && ns.pdsk < D_UP_TO_DATE))
diff --git a/include/linux/drbd.h b/include/linux/drbd.h
index e192167..d28fdd8 100644
--- a/include/linux/drbd.h
+++ b/include/linux/drbd.h
@@ -51,7 +51,6 @@
 
 #endif
 
-
 extern const char *drbd_buildtag(void);
 #define REL_VERSION "8.3.11"
 #define API_VERSION 88
@@ -159,6 +158,7 @@ enum drbd_ret_code {
 	ERR_CONN_IN_USE         = 159,
 	ERR_MINOR_CONFIGURED    = 160,
 	ERR_MINOR_EXISTS	= 161,
+	ERR_INVALID_REQUEST	= 162,
 
 	/* insert new ones above this line */
 	AFTER_LAST_ERR_CODE
@@ -349,37 +349,4 @@ enum drbd_timeout_flag {
 #define DRBD_MD_INDEX_FLEX_EXT -2
 #define DRBD_MD_INDEX_FLEX_INT -3
 
-/* Start of the new netlink/connector stuff */
-
-enum drbd_ncr_flags {
-	DRBD_NL_CREATE_DEVICE = 0x01,
-	DRBD_NL_SET_DEFAULTS =  0x02,
-};
-#define DRBD_NL_OBJ_NAME_LEN 32
-
-
-/* For searching a vacant cn_idx value */
-#define CN_IDX_STEP			6977
-
-struct drbd_nl_cfg_req {
-	int packet_type;
-	union {
-		struct {
-			unsigned int drbd_minor;
-			enum drbd_ncr_flags flags;
-		};
-		struct {
-			char obj_name[DRBD_NL_OBJ_NAME_LEN];
-		};
-	};
-	unsigned short tag_list[];
-};
-
-struct drbd_nl_cfg_reply {
-	int packet_type;
-	unsigned int minor;
-	int ret_code; /* enum ret_code or set_st_err_t */
-	unsigned short tag_list[]; /* only used with get_* calls */
-};
-
 #endif
-- 
1.7.4.1

[PATCH 10/18] drbd: allow holes in minor and volume id allocation

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

s/idr_get_new/idr_get_new_above/

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_int.h  |    1 +
 drivers/block/drbd/drbd_main.c |   33 +++++++++++++++++----------------
 drivers/block/drbd/drbd_nl.c   |    2 +-
 3 files changed, 19 insertions(+), 17 deletions(-)

diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index acd2877..f4c3c71 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -1509,6 +1509,7 @@ extern int is_valid_ar_handle(struct drbd_request *, sector_t);
 
 
 /* drbd_nl.c */
+extern int drbd_msg_put_info(const char *info);
 extern void drbd_suspend_io(struct drbd_conf *mdev);
 extern void drbd_resume_io(struct drbd_conf *mdev);
 extern char *ppsize(char *buf, unsigned long long size);
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 9b41213..64bf9b6 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -2322,6 +2322,7 @@ enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor,
 	struct request_queue *q;
 	int vnr_got = vnr;
 	int minor_got = minor;
+	enum drbd_ret_code err = ERR_NOMEM;
 
 	mdev = minor_to_mdev(minor);
 	if (mdev)
@@ -2389,33 +2390,33 @@ enum drbd_ret_code conn_new_minor(struct drbd_tconn *tconn, unsigned int minor,
 	INIT_LIST_HEAD(&mdev->current_epoch->list);
 	mdev->epochs = 1;
 
-	if (!idr_pre_get(&tconn->volumes, GFP_KERNEL))
-		goto out_no_vol_idr;
-	if (idr_get_new(&tconn->volumes, mdev, &vnr_got))
-		goto out_no_vol_idr;
-	if (vnr_got != vnr) {
-		dev_err(DEV, "vnr_got (%d) != vnr (%d)\n", vnr_got, vnr);
-		goto out_idr_remove_vol;
-	}
-
 	if (!idr_pre_get(&minors, GFP_KERNEL))
 		goto out_idr_remove_vol;
-	if (idr_get_new(&minors, mdev, &minor_got))
+	if (idr_get_new_above(&minors, mdev, minor, &minor_got))
 		goto out_idr_remove_vol;
 	if (minor_got != minor) {
-		/* minor exists, or other idr strangeness? */
-		dev_err(DEV, "available minor (%d) != requested minor (%d)\n",
-				minor_got, minor);
+		err = ERR_MINOR_EXISTS;
+		drbd_msg_put_info("requested minor exists already");
 		goto out_idr_remove_minor;
 	}
+
+	if (!idr_pre_get(&tconn->volumes, GFP_KERNEL))
+		goto out_no_vol_idr;
+	if (idr_get_new_above(&tconn->volumes, mdev, vnr, &vnr_got))
+		goto out_no_vol_idr;
+	if (vnr_got != vnr) {
+		err = ERR_INVALID_REQUEST;
+		drbd_msg_put_info("requested volume exists already");
+		goto out_idr_remove_vol;
+	}
 	add_disk(disk);
 
 	return NO_ERROR;
 
-out_idr_remove_minor:
-	idr_remove(&minors, minor_got);
 out_idr_remove_vol:
 	idr_remove(&tconn->volumes, vnr_got);
+out_idr_remove_minor:
+	idr_remove(&minors, minor_got);
 	synchronize_rcu();
 out_no_vol_idr:
 	kfree(mdev->current_epoch);
@@ -2429,7 +2430,7 @@ out_no_disk:
 	blk_cleanup_queue(q);
 out_no_q:
 	kfree(mdev);
-	return ERR_NOMEM;
+	return err;
 }
 
 /* counterpart of drbd_new_device.
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index a8f27cb..b40c83d 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -133,7 +133,7 @@ static void drbd_adm_send_reply(struct sk_buff *skb, struct genl_info *info)
 
 /* Used on a fresh "drbd_adm_prepare"d reply_skb, this cannot fail: The only
  * reason it could fail was no space in skb, and there are 4k available. */
-static int drbd_msg_put_info(const char *info)
+int drbd_msg_put_info(const char *info)
 {
 	struct sk_buff *skb = adm_ctx.reply_skb;
 	struct nlattr *nla;
-- 
1.7.4.1

[PATCH 11/18] drbd: remove now unused connector related files

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 include/linux/drbd_nl.h        |  172 ----------------------------------------
 include/linux/drbd_tag_magic.h |   84 -------------------
 2 files changed, 0 insertions(+), 256 deletions(-)
 delete mode 100644 include/linux/drbd_nl.h
 delete mode 100644 include/linux/drbd_tag_magic.h

diff --git a/include/linux/drbd_nl.h b/include/linux/drbd_nl.h
deleted file mode 100644
index 1216c7a..0000000
--- a/include/linux/drbd_nl.h
+++ /dev/null
@@ -1,172 +0,0 @@
-/*
-   PAKET( name,
-	  TYPE ( pn, pr, member )
-	  ...
-   )
-
-   You may never reissue one of the pn arguments
-*/
-
-#if !defined(NL_PACKET) || !defined(NL_STRING) || !defined(NL_INTEGER) || !defined(NL_BIT) || !defined(NL_INT64)
-#error "The macros NL_PACKET, NL_STRING, NL_INTEGER, NL_INT64 and NL_BIT needs to be defined"
-#endif
-
-NL_PACKET(primary, 1,
-       NL_BIT(		1,	T_MAY_IGNORE,	primary_force)
-)
-
-NL_PACKET(secondary, 2, )
-
-NL_PACKET(disk_conf, 3,
-	NL_INT64(	2,	T_MAY_IGNORE,	disk_size)
-	NL_STRING(	3,	T_MANDATORY,	backing_dev,	128)
-	NL_STRING(	4,	T_MANDATORY,	meta_dev,	128)
-	NL_INTEGER(	5,	T_MANDATORY,	meta_dev_idx)
-	NL_INTEGER(	6,	T_MAY_IGNORE,	on_io_error)
-	NL_INTEGER(	7,	T_MAY_IGNORE,	fencing)
-	NL_BIT(		37,	T_MAY_IGNORE,	use_bmbv)
-	NL_BIT(		53,	T_MAY_IGNORE,	no_disk_flush)
-	NL_BIT(		54,	T_MAY_IGNORE,	no_md_flush)
-	  /*  55 max_bio_size was available in 8.2.6rc2 */
-	NL_INTEGER(	56,	T_MAY_IGNORE,	max_bio_bvecs)
-	NL_BIT(		57,	T_MAY_IGNORE,	no_disk_barrier)
-	NL_BIT(		58,	T_MAY_IGNORE,	no_disk_drain)
-)
-
-NL_PACKET(detach, 4, )
-
-NL_PACKET(net_conf, 5,
-	NL_STRING(	8,	T_MANDATORY,	my_addr,	128)
-	NL_STRING(	9,	T_MANDATORY,	peer_addr,	128)
-	NL_STRING(	10,	T_MAY_IGNORE,	shared_secret,	SHARED_SECRET_MAX)
-	NL_STRING(	11,	T_MAY_IGNORE,	cram_hmac_alg,	SHARED_SECRET_MAX)
-	NL_STRING(	44,	T_MAY_IGNORE,	integrity_alg,	SHARED_SECRET_MAX)
-	NL_INTEGER(	14,	T_MAY_IGNORE,	timeout)
-	NL_INTEGER(	15,	T_MANDATORY,	wire_protocol)
-	NL_INTEGER(	16,	T_MAY_IGNORE,	try_connect_int)
-	NL_INTEGER(	17,	T_MAY_IGNORE,	ping_int)
-	NL_INTEGER(	18,	T_MAY_IGNORE,	max_epoch_size)
-	NL_INTEGER(	19,	T_MAY_IGNORE,	max_buffers)
-	NL_INTEGER(	20,	T_MAY_IGNORE,	unplug_watermark)
-	NL_INTEGER(	21,	T_MAY_IGNORE,	sndbuf_size)
-	NL_INTEGER(	22,	T_MAY_IGNORE,	ko_count)
-	NL_INTEGER(	24,	T_MAY_IGNORE,	after_sb_0p)
-	NL_INTEGER(	25,	T_MAY_IGNORE,	after_sb_1p)
-	NL_INTEGER(	26,	T_MAY_IGNORE,	after_sb_2p)
-	NL_INTEGER(	39,	T_MAY_IGNORE,	rr_conflict)
-	NL_INTEGER(	40,	T_MAY_IGNORE,	ping_timeo)
-	NL_INTEGER(	67,	T_MAY_IGNORE,	rcvbuf_size)
-	NL_INTEGER(	81,	T_MAY_IGNORE,	on_congestion)
-	NL_INTEGER(	82,	T_MAY_IGNORE,	cong_fill)
-	NL_INTEGER(	83,	T_MAY_IGNORE,	cong_extents)
-	  /* 59 addr_family was available in GIT, never released */
-	NL_BIT(		60,	T_MANDATORY,	mind_af)
-	NL_BIT(		27,	T_MAY_IGNORE,	want_lose)
-	NL_BIT(		28,	T_MAY_IGNORE,	two_primaries)
-	NL_BIT(		41,	T_MAY_IGNORE,	always_asbp)
-	NL_BIT(		61,	T_MAY_IGNORE,	no_cork)
-	NL_BIT(		62,	T_MANDATORY,	auto_sndbuf_size)
-	NL_BIT(		70,	T_MANDATORY,	dry_run)
-)
-
-NL_PACKET(disconnect, 6,
-	NL_BIT(		84,	T_MAY_IGNORE,	force)
-)
-
-NL_PACKET(resize, 7,
-	NL_INT64(		29,	T_MAY_IGNORE,	resize_size)
-	NL_BIT(			68,	T_MAY_IGNORE,	resize_force)
-	NL_BIT(			69,	T_MANDATORY,	no_resync)
-)
-
-NL_PACKET(syncer_conf, 8,
-	NL_INTEGER(	30,	T_MAY_IGNORE,	rate)
-	NL_INTEGER(	31,	T_MAY_IGNORE,	after)
-	NL_INTEGER(	32,	T_MAY_IGNORE,	al_extents)
-/*	NL_INTEGER(     71,	T_MAY_IGNORE,	dp_volume)
- *	NL_INTEGER(     72,	T_MAY_IGNORE,	dp_interval)
- *	NL_INTEGER(     73,	T_MAY_IGNORE,	throttle_th)
- *	NL_INTEGER(     74,	T_MAY_IGNORE,	hold_off_th)
- * feature will be reimplemented differently with 8.3.9 */
-	NL_STRING(      52,     T_MAY_IGNORE,   verify_alg,     SHARED_SECRET_MAX)
-	NL_STRING(      51,     T_MAY_IGNORE,   cpu_mask,       32)
-	NL_STRING(	64,	T_MAY_IGNORE,	csums_alg,	SHARED_SECRET_MAX)
-	NL_BIT(         65,     T_MAY_IGNORE,   use_rle)
-	NL_INTEGER(	75,	T_MAY_IGNORE,	on_no_data)
-	NL_INTEGER(	76,	T_MAY_IGNORE,	c_plan_ahead)
-	NL_INTEGER(     77,	T_MAY_IGNORE,	c_delay_target)
-	NL_INTEGER(     78,	T_MAY_IGNORE,	c_fill_target)
-	NL_INTEGER(     79,	T_MAY_IGNORE,	c_max_rate)
-	NL_INTEGER(     80,	T_MAY_IGNORE,	c_min_rate)
-)
-
-NL_PACKET(invalidate, 9, )
-NL_PACKET(invalidate_peer, 10, )
-NL_PACKET(pause_sync, 11, )
-NL_PACKET(resume_sync, 12, )
-NL_PACKET(suspend_io, 13, )
-NL_PACKET(resume_io, 14, )
-NL_PACKET(outdate, 15, )
-NL_PACKET(get_config, 16, )
-NL_PACKET(get_state, 17,
-	NL_INTEGER(	33,	T_MAY_IGNORE,	state_i)
-)
-
-NL_PACKET(get_uuids, 18,
-	NL_STRING(	34,	T_MAY_IGNORE,	uuids,	(UI_SIZE*sizeof(__u64)))
-	NL_INTEGER(	35,	T_MAY_IGNORE,	uuids_flags)
-)
-
-NL_PACKET(get_timeout_flag, 19,
-	NL_BIT(		36,	T_MAY_IGNORE,	use_degraded)
-)
-
-NL_PACKET(call_helper, 20,
-	NL_STRING(	38,	T_MAY_IGNORE,	helper,		32)
-)
-
-/* Tag nr 42 already allocated in drbd-8.1 development. */
-
-NL_PACKET(sync_progress, 23,
-	NL_INTEGER(	43,	T_MAY_IGNORE,	sync_progress)
-)
-
-NL_PACKET(dump_ee, 24,
-	NL_STRING(	45,	T_MAY_IGNORE,	dump_ee_reason, 32)
-	NL_STRING(	46,	T_MAY_IGNORE,	seen_digest, SHARED_SECRET_MAX)
-	NL_STRING(	47,	T_MAY_IGNORE,	calc_digest, SHARED_SECRET_MAX)
-	NL_INT64(	48,	T_MAY_IGNORE,	ee_sector)
-	NL_INT64(	49,	T_MAY_IGNORE,	ee_block_id)
-	NL_STRING(	50,	T_MAY_IGNORE,	ee_data,	32 << 10)
-)
-
-NL_PACKET(start_ov, 25,
-	NL_INT64(	66,	T_MAY_IGNORE,	start_sector)
-)
-
-NL_PACKET(new_c_uuid, 26,
-       NL_BIT(		63,	T_MANDATORY,	clear_bm)
-)
-
-#ifdef NL_RESPONSE
-NL_RESPONSE(return_code_only, 27)
-#endif
-
-NL_PACKET(new_connection, 28, /* CHT_CTOR */
-	NL_STRING(	85,	T_MANDATORY,	name, DRBD_NL_OBJ_NAME_LEN)
-)
-
-NL_PACKET(new_minor, 29, /* CHT_CONN */
-	NL_INTEGER(	86,	T_MANDATORY,	minor)
-	NL_INTEGER(	87,	T_MANDATORY,	vol_nr)
-)
-
-NL_PACKET(del_minor, 30, ) /* CHT_MINOR */
-NL_PACKET(del_connection, 31, ) /* CHT_CONN */
-
-#undef NL_PACKET
-#undef NL_INTEGER
-#undef NL_INT64
-#undef NL_BIT
-#undef NL_STRING
-#undef NL_RESPONSE
diff --git a/include/linux/drbd_tag_magic.h b/include/linux/drbd_tag_magic.h
deleted file mode 100644
index 0695431..0000000
--- a/include/linux/drbd_tag_magic.h
+++ /dev/null
@@ -1,84 +0,0 @@
-#ifndef DRBD_TAG_MAGIC_H
-#define DRBD_TAG_MAGIC_H
-
-#define TT_END     0
-#define TT_REMOVED 0xE000
-
-/* declare packet_type enums */
-enum packet_types {
-#define NL_PACKET(name, number, fields) P_ ## name = number,
-#define NL_RESPONSE(name, number) P_ ## name = number,
-#define NL_INTEGER(pn, pr, member)
-#define NL_INT64(pn, pr, member)
-#define NL_BIT(pn, pr, member)
-#define NL_STRING(pn, pr, member, len)
-#include "drbd_nl.h"
-	P_nl_after_last_packet,
-};
-
-/* These struct are used to deduce the size of the tag lists: */
-#define NL_PACKET(name, number, fields)	\
-	struct name ## _tag_len_struct { fields };
-#define NL_INTEGER(pn, pr, member)		\
-	int member; int tag_and_len ## member;
-#define NL_INT64(pn, pr, member)		\
-	__u64 member; int tag_and_len ## member;
-#define NL_BIT(pn, pr, member)		\
-	unsigned char member:1; int tag_and_len ## member;
-#define NL_STRING(pn, pr, member, len)	\
-	unsigned char member[len]; int member ## _len; \
-	int tag_and_len ## member;
-#include "linux/drbd_nl.h"
-
-/* declare tag-list-sizes */
-static const int tag_list_sizes[] = {
-#define NL_PACKET(name, number, fields) 2 fields ,
-#define NL_INTEGER(pn, pr, member)      + 4 + 4
-#define NL_INT64(pn, pr, member)        + 4 + 8
-#define NL_BIT(pn, pr, member)          + 4 + 1
-#define NL_STRING(pn, pr, member, len)  + 4 + (len)
-#include "drbd_nl.h"
-};
-
-/* The two highest bits are used for the tag type */
-#define TT_MASK      0xC000
-#define TT_INTEGER   0x0000
-#define TT_INT64     0x4000
-#define TT_BIT       0x8000
-#define TT_STRING    0xC000
-/* The next bit indicates if processing of the tag is mandatory */
-#define T_MANDATORY  0x2000
-#define T_MAY_IGNORE 0x0000
-#define TN_MASK      0x1fff
-/* The remaining 13 bits are used to enumerate the tags */
-
-#define tag_type(T)   ((T) & TT_MASK)
-#define tag_number(T) ((T) & TN_MASK)
-
-/* declare tag enums */
-#define NL_PACKET(name, number, fields) fields
-enum drbd_tags {
-#define NL_INTEGER(pn, pr, member)     T_ ## member = pn | TT_INTEGER | pr ,
-#define NL_INT64(pn, pr, member)       T_ ## member = pn | TT_INT64   | pr ,
-#define NL_BIT(pn, pr, member)         T_ ## member = pn | TT_BIT     | pr ,
-#define NL_STRING(pn, pr, member, len) T_ ## member = pn | TT_STRING  | pr ,
-#include "drbd_nl.h"
-};
-
-struct tag {
-	const char *name;
-	int type_n_flags;
-	int max_len;
-};
-
-/* declare tag names */
-#define NL_PACKET(name, number, fields) fields
-static const struct tag tag_descriptions[] = {
-#define NL_INTEGER(pn, pr, member)     [ pn ] = { #member, TT_INTEGER | pr, sizeof(int)   },
-#define NL_INT64(pn, pr, member)       [ pn ] = { #member, TT_INT64   | pr, sizeof(__u64) },
-#define NL_BIT(pn, pr, member)         [ pn ] = { #member, TT_BIT     | pr, sizeof(int)   },
-#define NL_STRING(pn, pr, member, len) [ pn ] = { #member, TT_STRING  | pr, (len)         },
-#include "drbd_nl.h"
-};
-
-#endif
-- 
1.7.4.1

[PATCH 12/18] drbd: drbd_adm_get_status needs to show some more detail

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

We want to see existing connection objects, even if they do not
currently have volumes attached.

Change the .dumpit variant of drbd_adm_get_status to iterate not over
minor devices, but over connections + volumes.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_int.h  |    3 +-
 drivers/block/drbd/drbd_main.c |   15 +++---
 drivers/block/drbd/drbd_nl.c   |  117 +++++++++++++++++++++++++++++++++-------
 3 files changed, 107 insertions(+), 28 deletions(-)

diff --git a/drivers/block/drbd/drbd_int.h b/drivers/block/drbd/drbd_int.h
index f4c3c71..d84a073 100644
--- a/drivers/block/drbd/drbd_int.h
+++ b/drivers/block/drbd/drbd_int.h
@@ -171,6 +171,7 @@ drbd_insert_fault(struct drbd_conf *mdev, unsigned int type) {
 extern struct ratelimit_state drbd_ratelimit_state;
 extern struct idr minors;
 extern struct list_head drbd_tconns;
+extern struct mutex drbd_cfg_mutex;
 
 /* on the wire */
 enum drbd_packet {
@@ -918,7 +919,7 @@ enum {
 
 struct drbd_tconn {			/* is a resource from the config file */
 	char *name;			/* Resource name */
-	struct list_head all_tconn;	/* List of all drbd_tconn, prot by global_state_lock */
+	struct list_head all_tconn;	/* linked on global drbd_tconns */
 	struct idr volumes;		/* <tconn, vnr> to mdev mapping */
 	enum drbd_conns cstate;		/* Only C_STANDALONE to C_WF_REPORT_PARAMS */
 	struct mutex cstate_mutex;	/* Protects graceful disconnects */
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 64bf9b6..2e79032 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -120,6 +120,7 @@ module_param_string(usermode_helper, usermode_helper, sizeof(usermode_helper), 0
  */
 struct idr minors;
 struct list_head drbd_tconns;  /* list of struct drbd_tconn */
+DEFINE_MUTEX(drbd_cfg_mutex);
 
 struct kmem_cache *drbd_request_cache;
 struct kmem_cache *drbd_ee_cache;	/* peer requests */
@@ -2238,14 +2239,14 @@ struct drbd_tconn *conn_by_name(const char *name)
 	if (!name || !name[0])
 		return NULL;
 
-	write_lock_irq(&global_state_lock);
+	mutex_lock(&drbd_cfg_mutex);
 	list_for_each_entry(tconn, &drbd_tconns, all_tconn) {
 		if (!strcmp(tconn->name, name))
 			goto found;
 	}
 	tconn = NULL;
 found:
-	write_unlock_irq(&global_state_lock);
+	mutex_unlock(&drbd_cfg_mutex);
 	return tconn;
 }
 
@@ -2285,9 +2286,9 @@ struct drbd_tconn *drbd_new_tconn(const char *name)
 	drbd_thread_init(tconn, &tconn->worker, drbd_worker, "worker");
 	drbd_thread_init(tconn, &tconn->asender, drbd_asender, "asender");
 
-	write_lock_irq(&global_state_lock);
-	list_add(&tconn->all_tconn, &drbd_tconns);
-	write_unlock_irq(&global_state_lock);
+	mutex_lock(&drbd_cfg_mutex);
+	list_add_tail(&tconn->all_tconn, &drbd_tconns);
+	mutex_unlock(&drbd_cfg_mutex);
 
 	return tconn;
 
@@ -2302,9 +2303,9 @@ fail:
 
 void drbd_free_tconn(struct drbd_tconn *tconn)
 {
-	write_lock_irq(&global_state_lock);
+	mutex_lock(&drbd_cfg_mutex);
 	list_del(&tconn->all_tconn);
-	write_unlock_irq(&global_state_lock);
+	mutex_unlock(&drbd_cfg_mutex);
 	idr_destroy(&tconn->volumes);
 
 	free_cpumask_var(tconn->cpu_mask);
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index b40c83d..c389995 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -1577,6 +1577,10 @@ int drbd_adm_connect(struct sk_buff *skb, struct genl_info *info)
 
 	new_my_addr = (struct sockaddr *)&new_conf->my_addr;
 	new_peer_addr = (struct sockaddr *)&new_conf->peer_addr;
+
+	/* No need to take drbd_cfg_mutex here.  All reconfiguration is
+	 * strictly serialized on genl_lock(). We are protected against
+	 * concurrent reconfiguration/addition/deletion */
 	list_for_each_entry(oconn, &drbd_tconns, all_tconn) {
 		if (oconn == tconn)
 			continue;
@@ -2220,6 +2224,24 @@ int drbd_adm_outdate(struct sk_buff *skb, struct genl_info *info)
 	return drbd_adm_simple_request_state(skb, info, NS(disk, D_OUTDATED));
 }
 
+int nla_put_drbd_cfg_context(struct sk_buff *skb, const char *conn_name, unsigned vnr)
+{
+	struct nlattr *nla;
+	nla = nla_nest_start(skb, DRBD_NLA_CFG_CONTEXT);
+	if (!nla)
+		goto nla_put_failure;
+	if (vnr != VOLUME_UNSPECIFIED)
+		NLA_PUT_U32(skb, T_ctx_volume, vnr);
+	NLA_PUT_STRING(skb, T_ctx_conn_name, conn_name);
+	nla_nest_end(skb, nla);
+	return 0;
+
+nla_put_failure:
+	if (nla)
+		nla_nest_cancel(skb, nla);
+	return -EMSGSIZE;
+}
+
 int nla_put_status_info(struct sk_buff *skb, struct drbd_conf *mdev,
 		const struct sib_info *sib)
 {
@@ -2248,12 +2270,8 @@ int nla_put_status_info(struct sk_buff *skb, struct drbd_conf *mdev,
 
 	/* We need to add connection name and volume number information still.
 	 * Minor number is in drbd_genlmsghdr. */
-	nla = nla_nest_start(skb, DRBD_NLA_CFG_CONTEXT);
-	if (!nla)
+	if (nla_put_drbd_cfg_context(skb, mdev->tconn->name, mdev->vnr))
 		goto nla_put_failure;
-	NLA_PUT_U32(skb, T_ctx_volume, mdev->vnr);
-	NLA_PUT_STRING(skb, T_ctx_conn_name, mdev->tconn->name);
-	nla_nest_end(skb, nla);
 
 	if (got_ldev)
 		if (disk_conf_to_skb(skb, &mdev->ldev->dc, exclude_sensitive))
@@ -2340,43 +2358,102 @@ int drbd_adm_get_status_all(struct sk_buff *skb, struct netlink_callback *cb)
 {
 	struct drbd_conf *mdev;
 	struct drbd_genlmsghdr *dh;
-	int minor = cb->args[0];
-
-	/* Open coded deferred single idr_for_each_entry iteration.
+	struct drbd_tconn *pos = (struct drbd_tconn*)cb->args[0];
+	struct drbd_tconn *tconn = NULL;
+	struct drbd_tconn *tmp;
+	unsigned volume = cb->args[1];
+
+	/* Open coded, deferred, iteration:
+	 * list_for_each_entry_safe(tconn, tmp, &drbd_tconns, all_tconn) {
+	 *	idr_for_each_entry(&tconn->volumes, mdev, i) {
+	 *	  ...
+	 *	}
+	 * }
+	 * where tconn is cb->args[0];
+	 * and i is cb->args[1];
+	 *
 	 * This may miss entries inserted after this dump started,
 	 * or entries deleted before they are reached.
-	 * But we need to make sure the mdev won't disappear while
-	 * we are looking at it. */
+	 *
+	 * We need to make sure the mdev won't disappear while
+	 * we are looking at it, and revalidate our iterators
+	 * on each iteration.
+	 */
 
+	/* synchronize with drbd_new_tconn/drbd_free_tconn */
+	mutex_lock(&drbd_cfg_mutex);
+	/* synchronize with drbd_delete_device */
 	rcu_read_lock();
-	mdev = idr_get_next(&minors, &minor);
-	if (mdev) {
+next_tconn:
+	/* revalidate iterator position */
+	list_for_each_entry(tmp, &drbd_tconns, all_tconn) {
+		if (pos == NULL) {
+			/* first iteration */
+			pos = tmp;
+			tconn = pos;
+			break;
+		}
+		if (tmp == pos) {
+			tconn = pos;
+			break;
+		}
+	}
+	if (tconn) {
+		mdev = idr_get_next(&tconn->volumes, &volume);
+		if (!mdev) {
+			/* No more volumes to dump on this tconn.
+			 * Advance tconn iterator. */
+			pos = list_entry(tconn->all_tconn.next,
+					struct drbd_tconn, all_tconn);
+			/* But, did we dump any volume on this tconn yet? */
+			if (volume != 0) {
+				tconn = NULL;
+				volume = 0;
+				goto next_tconn;
+			}
+		}
+
 		dh = genlmsg_put(skb, NETLINK_CB(cb->skb).pid,
 				cb->nlh->nlmsg_seq, &drbd_genl_family,
 				NLM_F_MULTI, DRBD_ADM_GET_STATUS);
 		if (!dh)
-			goto errout;
+			goto out;
+
+		if (!mdev) {
+			/* this is a tconn without a single volume */
+			dh->minor = -1U;
+			dh->ret_code = NO_ERROR;
+			if (nla_put_drbd_cfg_context(skb, tconn->name, VOLUME_UNSPECIFIED))
+				genlmsg_cancel(skb, dh);
+			else
+				genlmsg_end(skb, dh);
+			goto out;
+		}
 
-		D_ASSERT(mdev->minor == minor);
+		D_ASSERT(mdev->vnr == volume);
+		D_ASSERT(mdev->tconn == tconn);
 
-		dh->minor = minor;
+		dh->minor = mdev_to_minor(mdev);
 		dh->ret_code = NO_ERROR;
 
 		pr_info("dump: minor=%u, conn=%s[%u]\n",
 			dh->minor, mdev->tconn->name, mdev->vnr);
 		if (nla_put_status_info(skb, mdev, NULL)) {
 			genlmsg_cancel(skb, dh);
-			goto errout;
+			goto out;
 		}
 		genlmsg_end(skb, dh);
         }
 
-errout:
+out:
 	rcu_read_unlock();
-	/* where to start idr_get_next with the next iteration */
-        cb->args[0] = minor+1;
+	mutex_unlock(&drbd_cfg_mutex);
+	/* where to start the next iteration */
+        cb->args[0] = (long)pos;
+        cb->args[1] = (pos == tconn) ? volume + 1 : 0;
 
-	/* No more minors found: empty skb. Which will terminate the dump. */
+	/* No more tconns/volumes/minors found results in an empty skb.
+	 * Which will terminate the dump. */
         return skb->len;
 }
 
-- 
1.7.4.1

[PATCH 13/18] drbd: simplify conn_all_vols_unconf, make it bool

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Get rid of a temporary variable and, funny bitand assignment.
Just short circuit, returning false, once we encounter the first
still configured volume.

FIXME verify call sites for need of rcu_read_lock or stronger.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_state.c |   16 +++++++---------
 drivers/block/drbd/drbd_state.h |    2 +-
 2 files changed, 8 insertions(+), 10 deletions(-)

diff --git a/drivers/block/drbd/drbd_state.c b/drivers/block/drbd/drbd_state.c
index 34be2ef..c6a659b 100644
--- a/drivers/block/drbd/drbd_state.c
+++ b/drivers/block/drbd/drbd_state.c
@@ -48,20 +48,18 @@ static enum drbd_state_rv is_valid_transition(union drbd_state os, union drbd_st
 static union drbd_state sanitize_state(struct drbd_conf *mdev, union drbd_state ns,
 				       const char **warn_sync_abort);
 
-int conn_all_vols_unconf(struct drbd_tconn *tconn)
+bool conn_all_vols_unconf(struct drbd_tconn *tconn)
 {
 	struct drbd_conf *mdev;
-	int minor, uncfg = 1;
+	int minor;
 
 	idr_for_each_entry(&tconn->volumes, mdev, minor) {
-		uncfg &= (mdev->state.disk == D_DISKLESS &&
-			  mdev->state.conn == C_STANDALONE &&
-			  mdev->state.role == R_SECONDARY);
-		if (!uncfg)
-			break;
+		if (mdev->state.disk != D_DISKLESS ||
+		    mdev->state.conn != C_STANDALONE ||
+		    mdev->state.role != R_SECONDARY)
+			return false;
 	}
-
-	return uncfg;
+	return true;
 }
 
 /**
diff --git a/drivers/block/drbd/drbd_state.h b/drivers/block/drbd/drbd_state.h
index d9536cd..55df072 100644
--- a/drivers/block/drbd/drbd_state.h
+++ b/drivers/block/drbd/drbd_state.h
@@ -91,7 +91,7 @@ conn_request_state(struct drbd_tconn *tconn, union drbd_state mask, union drbd_s
 		   enum chg_state_flags flags);
 
 extern void drbd_resume_al(struct drbd_conf *mdev);
-extern int conn_all_vols_unconf(struct drbd_tconn *tconn);
+extern bool conn_all_vols_unconf(struct drbd_tconn *tconn);
 
 /**
  * drbd_request_state() - Reqest a state change
-- 
1.7.4.1

[PATCH 14/18] drbd: Allow a Diskless Secondary volume to be removed

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Even if the connection is still established.
We should be able to reduce a volume from a replication group,
without taking the whole group offline.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_nl.c |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)

diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index c389995..740649b 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -248,6 +248,8 @@ static int drbd_adm_prepare(struct sk_buff *skb, struct genl_info *info,
 		drbd_msg_put_info("over-determined configuration context mismatch");
 		return ERR_INVALID_REQUEST;
 	}
+	if (adm_ctx.mdev && !adm_ctx.tconn)
+		adm_ctx.tconn = adm_ctx.mdev->tconn;
 	return NO_ERROR;
 
 fail:
@@ -2676,10 +2678,15 @@ int drbd_adm_delete_minor(struct sk_buff *skb, struct genl_info *info)
 
 	mdev = adm_ctx.mdev;
 	if (mdev->state.disk == D_DISKLESS &&
-	    mdev->state.conn == C_STANDALONE &&
+	    /* no need to be mdev->state.conn == C_STANDALONE &&
+	     * we may want to delete a minor from a live replication group.
+	     */
 	    mdev->state.role == R_SECONDARY) {
 		drbd_delete_device(mdev_to_minor(mdev));
 		retcode = NO_ERROR;
+		/* if this was the last volume of this connection,
+		 * this will terminate all threads */
+		conn_reconfig_done(adm_ctx.tconn);
 	} else
 		retcode = ERR_MINOR_CONFIGURED;
 out:
-- 
1.7.4.1

[PATCH 15/18] drbd: new-connection and new-minor succeed, if the object already exists

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

Follow O_CREAT semantics when creating connection or minor device/volume
objects.  If we need O_CREAT|O_EXCL semantics some time down the road,
we can add NLM_F_EXCL to the netlink message flags.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_nl.c |   16 ++++++++++++++--
 1 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index 740649b..e89d108 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -2625,8 +2625,11 @@ int drbd_adm_create_connection(struct sk_buff *skb, struct genl_info *info)
 		goto out;
 
 	if (adm_ctx.tconn) {
-		retcode = ERR_INVALID_REQUEST;
-		drbd_msg_put_info("connection exists");
+		if (info->nlhdr->nlmsg_flags & NLM_F_EXCL) {
+			retcode = ERR_INVALID_REQUEST;
+			drbd_msg_put_info("connection exists");
+		}
+		/* else: still NO_ERROR */
 		goto out;
 	}
 
@@ -2659,6 +2662,15 @@ int drbd_adm_add_minor(struct sk_buff *skb, struct genl_info *info)
 		return ERR_INVALID_REQUEST;
 	}
 
+	/* drbd_adm_prepare made sure already
+	 * that mdev->tconn and mdev->vnr match the request. */
+	if (adm_ctx.mdev) {
+		if (info->nlhdr->nlmsg_flags & NLM_F_EXCL)
+			retcode = ERR_MINOR_EXISTS;
+		/* else: still NO_ERROR */
+		goto out;
+	}
+
 	retcode = conn_new_minor(adm_ctx.tconn, dh->minor, adm_ctx.volume);
 out:
 	drbd_adm_finish(info, retcode);
-- 
1.7.4.1

[PATCH 16/18] drbd: bail out if a config requrest is over-determined, and not matching

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

We have resources resp. connections, volumes, and minor numbers.
A config request may specifies all three of them.
If it turns out that the minor belongs to a different connection, or a
different volume number in the same connection, that configuration
request is invalid.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_nl.c |   16 ++++++++--------
 1 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index e89d108..773946d 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -232,20 +232,20 @@ static int drbd_adm_prepare(struct sk_buff *skb, struct genl_info *info,
 	}
 
 	/* some more paranoia, if the request was over-determined */
+	if (adm_ctx.mdev && adm_ctx.tconn &&
+	    adm_ctx.mdev->tconn != adm_ctx.tconn) {
+		pr_warning("request: minor=%u, conn=%s; but that minor belongs to connection %s\n",
+				adm_ctx.minor, adm_ctx.conn_name, adm_ctx.mdev->tconn->name);
+		drbd_msg_put_info("minor exists in different connection");
+		return ERR_INVALID_REQUEST;
+	}
 	if (adm_ctx.mdev &&
 	    adm_ctx.volume != VOLUME_UNSPECIFIED &&
 	    adm_ctx.volume != adm_ctx.mdev->vnr) {
 		pr_warning("request: minor=%u, volume=%u; but that minor is volume %u in %s\n",
 				adm_ctx.minor, adm_ctx.volume,
 				adm_ctx.mdev->vnr, adm_ctx.mdev->tconn->name);
-		drbd_msg_put_info("over-determined configuration context mismatch");
-		return ERR_INVALID_REQUEST;
-	}
-	if (adm_ctx.mdev && adm_ctx.tconn &&
-	    adm_ctx.mdev->tconn != adm_ctx.tconn) {
-		pr_warning("request: minor=%u, conn=%s; but that minor belongs to connection %s\n",
-				adm_ctx.minor, adm_ctx.conn_name, adm_ctx.mdev->tconn->name);
-		drbd_msg_put_info("over-determined configuration context mismatch");
+		drbd_msg_put_info("minor exists as different volume");
 		return ERR_INVALID_REQUEST;
 	}
 	if (adm_ctx.mdev && !adm_ctx.tconn)
-- 
1.7.4.1

[PATCH 17/18] drbd: add forgotten spin_unlock

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

somehow a "goto abort" was introduced with commit
  drbd: Extracted is_valid_transition() out of sanitize_state()
which left drbd_req_state still holding the spin lock.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_state.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)

diff --git a/drivers/block/drbd/drbd_state.c b/drivers/block/drbd/drbd_state.c
index c6a659b..3ca535e 100644
--- a/drivers/block/drbd/drbd_state.c
+++ b/drivers/block/drbd/drbd_state.c
@@ -181,8 +181,10 @@ drbd_req_state(struct drbd_conf *mdev, union drbd_state mask,
 	os = mdev->state;
 	ns = sanitize_state(mdev, apply_mask_val(os, mask, val), NULL);
 	rv = is_valid_transition(os, ns);
-	if (rv < SS_SUCCESS)
+	if (rv < SS_SUCCESS) {
+		spin_unlock_irqrestore(&mdev->tconn->req_lock, flags);
 		goto abort;
+	}
 
 	if (cl_wide_st_chg(mdev, os, ns)) {
 		rv = is_valid_state(mdev, ns);
-- 
1.7.4.1

[PATCH 18/18] drbd: introduce in-kernel "down" command

[Philipp Reisner]

From: Lars Ellenberg <lars.ellenberg@linbit.com>

This greatly simplifies deconfiguration of whole resources.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_main.c |    2 -
 drivers/block/drbd/drbd_nl.c   |  203 ++++++++++++++++++++++++++++++----------
 include/linux/drbd_genl.h      |    2 +
 3 files changed, 154 insertions(+), 53 deletions(-)

diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index 2e79032..de49c8d 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -2303,9 +2303,7 @@ fail:
 
 void drbd_free_tconn(struct drbd_tconn *tconn)
 {
-	mutex_lock(&drbd_cfg_mutex);
 	list_del(&tconn->all_tconn);
-	mutex_unlock(&drbd_cfg_mutex);
 	idr_destroy(&tconn->volumes);
 
 	free_cpumask_var(tconn->cpu_mask);
diff --git a/drivers/block/drbd/drbd_nl.c b/drivers/block/drbd/drbd_nl.c
index 773946d..2970f45 100644
--- a/drivers/block/drbd/drbd_nl.c
+++ b/drivers/block/drbd/drbd_nl.c
@@ -73,6 +73,7 @@ int drbd_adm_delete_minor(struct sk_buff *skb, struct genl_info *info);
 
 int drbd_adm_create_connection(struct sk_buff *skb, struct genl_info *info);
 int drbd_adm_delete_connection(struct sk_buff *skb, struct genl_info *info);
+int drbd_adm_down(struct sk_buff *skb, struct genl_info *info);
 
 int drbd_adm_set_role(struct sk_buff *skb, struct genl_info *info);
 int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info);
@@ -1449,6 +1450,18 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
 	return 0;
 }
 
+static int adm_detach(struct drbd_conf *mdev)
+{
+	enum drbd_ret_code retcode;
+	drbd_suspend_io(mdev); /* so no-one is stuck in drbd_al_begin_io */
+	retcode = drbd_request_state(mdev, NS(disk, D_DISKLESS));
+	wait_event(mdev->misc_wait,
+			mdev->state.disk != D_DISKLESS ||
+			!atomic_read(&mdev->local_cnt));
+	drbd_resume_io(mdev);
+	return retcode;
+}
+
 /* Detaching the disk is a process in multiple stages.  First we need to lock
  * out application IO, in-flight IO, IO stuck in drbd_al_begin_io.
  * Then we transition to D_DISKLESS, and wait for put_ldev() to return all
@@ -1456,7 +1469,6 @@ int drbd_adm_attach(struct sk_buff *skb, struct genl_info *info)
  * Only then we have finally detached. */
 int drbd_adm_detach(struct sk_buff *skb, struct genl_info *info)
 {
-	struct drbd_conf *mdev;
 	enum drbd_ret_code retcode;
 
 	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
@@ -1465,13 +1477,7 @@ int drbd_adm_detach(struct sk_buff *skb, struct genl_info *info)
 	if (retcode != NO_ERROR)
 		goto out;
 
-	mdev = adm_ctx.mdev;
-	drbd_suspend_io(mdev); /* so no-one is stuck in drbd_al_begin_io */
-	retcode = drbd_request_state(mdev, NS(disk, D_DISKLESS));
-	wait_event(mdev->misc_wait,
-			mdev->state.disk != D_DISKLESS ||
-			!atomic_read(&mdev->local_cnt));
-	drbd_resume_io(mdev);
+	retcode = adm_detach(adm_ctx.mdev);
 out:
 	drbd_adm_finish(info, retcode);
 	return 0;
@@ -1713,10 +1719,49 @@ out:
 	return 0;
 }
 
+static enum drbd_state_rv conn_try_disconnect(struct drbd_tconn *tconn, bool force)
+{
+	enum drbd_state_rv rv;
+	if (force) {
+		spin_lock_irq(&tconn->req_lock);
+		if (tconn->cstate >= C_WF_CONNECTION)
+			_conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD);
+		spin_unlock_irq(&tconn->req_lock);
+		return SS_SUCCESS;
+	}
+
+	rv = conn_request_state(tconn, NS(conn, C_DISCONNECTING), 0);
+
+	switch (rv) {
+	case SS_NOTHING_TO_DO:
+	case SS_ALREADY_STANDALONE:
+		return SS_SUCCESS;
+	case SS_PRIMARY_NOP:
+		/* Our state checking code wants to see the peer outdated. */
+		rv = conn_request_state(tconn, NS2(conn, C_DISCONNECTING,
+							pdsk, D_OUTDATED), CS_VERBOSE);
+		break;
+	case SS_CW_FAILED_BY_PEER:
+		/* The peer probably wants to see us outdated. */
+		rv = conn_request_state(tconn, NS2(conn, C_DISCONNECTING,
+							disk, D_OUTDATED), 0);
+		if (rv == SS_IS_DISKLESS || rv == SS_LOWER_THAN_OUTDATED) {
+			conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD);
+			rv = SS_SUCCESS;
+		}
+		break;
+	default:;
+		/* no special handling necessary */
+	}
+
+	return rv;
+}
+
 int drbd_adm_disconnect(struct sk_buff *skb, struct genl_info *info)
 {
 	struct disconnect_parms parms;
 	struct drbd_tconn *tconn;
+	enum drbd_state_rv rv;
 	enum drbd_ret_code retcode;
 	int err;
 
@@ -1737,35 +1782,8 @@ int drbd_adm_disconnect(struct sk_buff *skb, struct genl_info *info)
 		}
 	}
 
-	if (parms.force_disconnect) {
-		spin_lock_irq(&tconn->req_lock);
-		if (tconn->cstate >= C_WF_CONNECTION)
-			_conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD);
-		spin_unlock_irq(&tconn->req_lock);
-		goto done;
-	}
-
-	retcode = conn_request_state(tconn, NS(conn, C_DISCONNECTING), 0);
-
-	if (retcode == SS_NOTHING_TO_DO)
-		goto done;
-	else if (retcode == SS_ALREADY_STANDALONE)
-		goto done;
-	else if (retcode == SS_PRIMARY_NOP) {
-		/* Our state checking code wants to see the peer outdated. */
-		retcode = conn_request_state(tconn, NS2(conn, C_DISCONNECTING,
-							pdsk, D_OUTDATED), CS_VERBOSE);
-	} else if (retcode == SS_CW_FAILED_BY_PEER) {
-		/* The peer probably wants to see us outdated. */
-		retcode = conn_request_state(tconn, NS2(conn, C_DISCONNECTING,
-							disk, D_OUTDATED), 0);
-		if (retcode == SS_IS_DISKLESS || retcode == SS_LOWER_THAN_OUTDATED) {
-			conn_request_state(tconn, NS(conn, C_DISCONNECTING), CS_HARD);
-			retcode = SS_SUCCESS;
-		}
-	}
-
-	if (retcode < SS_SUCCESS)
+	rv = conn_try_disconnect(tconn, parms.force_disconnect);
+	if (rv < SS_SUCCESS)
 		goto fail;
 
 	if (wait_event_interruptible(tconn->ping_wait,
@@ -1776,7 +1794,6 @@ int drbd_adm_disconnect(struct sk_buff *skb, struct genl_info *info)
 		goto fail;
 	}
 
- done:
 	retcode = NO_ERROR;
  fail:
 	drbd_adm_finish(info, retcode);
@@ -2677,9 +2694,21 @@ out:
 	return 0;
 }
 
+static enum drbd_ret_code adm_delete_minor(struct drbd_conf *mdev)
+{
+	if (mdev->state.disk == D_DISKLESS &&
+	    /* no need to be mdev->state.conn == C_STANDALONE &&
+	     * we may want to delete a minor from a live replication group.
+	     */
+	    mdev->state.role == R_SECONDARY) {
+		drbd_delete_device(mdev_to_minor(mdev));
+		return NO_ERROR;
+	} else
+		return ERR_MINOR_CONFIGURED;
+}
+
 int drbd_adm_delete_minor(struct sk_buff *skb, struct genl_info *info)
 {
-	struct drbd_conf *mdev;
 	enum drbd_ret_code retcode;
 
 	retcode = drbd_adm_prepare(skb, info, DRBD_ADM_NEED_MINOR);
@@ -2688,19 +2717,89 @@ int drbd_adm_delete_minor(struct sk_buff *skb, struct genl_info *info)
 	if (retcode != NO_ERROR)
 		goto out;
 
-	mdev = adm_ctx.mdev;
-	if (mdev->state.disk == D_DISKLESS &&
-	    /* no need to be mdev->state.conn == C_STANDALONE &&
-	     * we may want to delete a minor from a live replication group.
-	     */
-	    mdev->state.role == R_SECONDARY) {
-		drbd_delete_device(mdev_to_minor(mdev));
-		retcode = NO_ERROR;
-		/* if this was the last volume of this connection,
-		 * this will terminate all threads */
+	mutex_lock(&drbd_cfg_mutex);
+	retcode = adm_delete_minor(adm_ctx.mdev);
+	mutex_unlock(&drbd_cfg_mutex);
+	/* if this was the last volume of this connection,
+	 * this will terminate all threads */
+	if (retcode == NO_ERROR)
 		conn_reconfig_done(adm_ctx.tconn);
-	} else
-		retcode = ERR_MINOR_CONFIGURED;
+out:
+	drbd_adm_finish(info, retcode);
+	return 0;
+}
+
+int drbd_adm_down(struct sk_buff *skb, struct genl_info *info)
+{
+	enum drbd_ret_code retcode;
+	enum drbd_state_rv rv;
+	struct drbd_conf *mdev;
+	unsigned i;
+
+	retcode = drbd_adm_prepare(skb, info, 0);
+	if (!adm_ctx.reply_skb)
+		return retcode;
+	if (retcode != NO_ERROR)
+		goto out;
+
+	if (!adm_ctx.tconn) {
+		retcode = ERR_CONN_NOT_KNOWN;
+		goto out;
+	}
+
+	mutex_lock(&drbd_cfg_mutex);
+	/* demote */
+	idr_for_each_entry(&adm_ctx.tconn->volumes, mdev, i) {
+		retcode = drbd_set_role(mdev, R_SECONDARY, 0);
+		if (retcode < SS_SUCCESS) {
+			drbd_msg_put_info("failed to demote");
+			goto out_unlock;
+		}
+	}
+
+	/* disconnect */
+	rv = conn_try_disconnect(adm_ctx.tconn, 0);
+	if (rv < SS_SUCCESS) {
+		retcode = rv; /* enum type mismatch! */
+		drbd_msg_put_info("failed to disconnect");
+		goto out_unlock;
+	}
+
+	/* detach */
+	idr_for_each_entry(&adm_ctx.tconn->volumes, mdev, i) {
+		rv = adm_detach(mdev);
+		if (rv < SS_SUCCESS) {
+			retcode = rv; /* enum type mismatch! */
+			drbd_msg_put_info("failed to detach");
+			goto out_unlock;
+		}
+	}
+
+	/* delete volumes */
+	idr_for_each_entry(&adm_ctx.tconn->volumes, mdev, i) {
+		retcode = adm_delete_minor(mdev);
+		if (retcode != NO_ERROR) {
+			/* "can not happen" */
+			drbd_msg_put_info("failed to delete volume");
+			goto out_unlock;
+		}
+	}
+
+	/* stop all threads */
+	conn_reconfig_done(adm_ctx.tconn);
+
+	/* delete connection */
+	if (conn_lowest_minor(adm_ctx.tconn) < 0) {
+		drbd_free_tconn(adm_ctx.tconn);
+		retcode = NO_ERROR;
+	} else {
+		/* "can not happen" */
+		retcode = ERR_CONN_IN_USE;
+		drbd_msg_put_info("failed to delete connection");
+		goto out_unlock;
+	}
+out_unlock:
+	mutex_unlock(&drbd_cfg_mutex);
 out:
 	drbd_adm_finish(info, retcode);
 	return 0;
@@ -2716,12 +2815,14 @@ int drbd_adm_delete_connection(struct sk_buff *skb, struct genl_info *info)
 	if (retcode != NO_ERROR)
 		goto out;
 
+	mutex_lock(&drbd_cfg_mutex);
 	if (conn_lowest_minor(adm_ctx.tconn) < 0) {
 		drbd_free_tconn(adm_ctx.tconn);
 		retcode = NO_ERROR;
 	} else {
 		retcode = ERR_CONN_IN_USE;
 	}
+	mutex_unlock(&drbd_cfg_mutex);
 
 out:
 	drbd_adm_finish(info, retcode);
diff --git a/include/linux/drbd_genl.h b/include/linux/drbd_genl.h
index 84e1684..a07d692 100644
--- a/include/linux/drbd_genl.h
+++ b/include/linux/drbd_genl.h
@@ -347,3 +347,5 @@ GENL_op(DRBD_ADM_OUTDATE,	25, GENL_doit(drbd_adm_outdate),
 	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
 GENL_op(DRBD_ADM_GET_TIMEOUT_TYPE, 26, GENL_doit(drbd_adm_get_timeout_type),
 	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
+GENL_op(DRBD_ADM_DOWN,		27, GENL_doit(drbd_adm_down),
+	GENL_tla_expected(DRBD_NLA_CFG_CONTEXT, GENLA_F_REQUIRED))
-- 
1.7.4.1