mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Lars Ellenberg <lars.ellenberg@linbit.com>
To: Jens Axboe <axboe@kernel.dk>,
	linux-kernel@vger.kernel.org, linux-block@vger.kernel.org
Cc: drbd-dev@lists.linbit.com
Subject: [PATCH 04/17] drbd: ignore "all zero" peer volume sizes in handshake
Date: Thu, 20 Dec 2018 17:23:31 +0100	[thread overview]
Message-ID: <20181220162344.8430-5-lars.ellenberg@linbit.com> (raw)
In-Reply-To: <20181220162344.8430-1-lars.ellenberg@linbit.com>

During handshake, if we are diskless ourselves, we used to accept any size
presented by the peer.

Which could be zero if that peer was just brought up and connected
to us without having a disk attached first, in which case both
peers would just "flip" their volume sizes.

Now, even a diskless node will ignore "zero" sizes
presented by a diskless peer.

Also a currently Diskless Primary will refuse to shrink during handshake:
it may be frozen, and waiting for a "suitable" local disk or peer to
re-appear (on-no-data-accessible suspend-io). If the peer is smaller
than what we used to be, it is not suitable.

The logic for a diskless node during handshake is now supposed to be:
believe the peer, if
 - I don't have a current size myself
 - we agree on the size anyways
 - I do have a current size, am Secondary, and he has the only disk
 - I do have a current size, am Primary, and he has the only disk,
   which is larger than my current size

Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
---
 drivers/block/drbd/drbd_receiver.c | 33 +++++++++++++++++++++++++++---
 1 file changed, 30 insertions(+), 3 deletions(-)

diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index 1b9822f264d2..fbf30fe45862 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -3981,6 +3981,7 @@ static int receive_sizes(struct drbd_connection *connection, struct packet_info
 	struct o_qlim *o = (connection->agreed_features & DRBD_FF_WSAME) ? p->qlim : NULL;
 	enum determine_dev_size dd = DS_UNCHANGED;
 	sector_t p_size, p_usize, p_csize, my_usize;
+	sector_t new_size, cur_size;
 	int ldsc = 0; /* local disk size changed */
 	enum dds_flags ddsf;
 
@@ -3988,6 +3989,7 @@ static int receive_sizes(struct drbd_connection *connection, struct packet_info
 	if (!peer_device)
 		return config_unknown_volume(connection, pi);
 	device = peer_device->device;
+	cur_size = drbd_get_capacity(device->this_bdev);
 
 	p_size = be64_to_cpu(p->d_size);
 	p_usize = be64_to_cpu(p->u_size);
@@ -3998,7 +4000,6 @@ static int receive_sizes(struct drbd_connection *connection, struct packet_info
 	device->p_size = p_size;
 
 	if (get_ldev(device)) {
-		sector_t new_size, cur_size;
 		rcu_read_lock();
 		my_usize = rcu_dereference(device->ldev->disk_conf)->disk_size;
 		rcu_read_unlock();
@@ -4016,7 +4017,6 @@ static int receive_sizes(struct drbd_connection *connection, struct packet_info
 		/* Never shrink a device with usable data during connect.
 		   But allow online shrinking if we are connected. */
 		new_size = drbd_new_dev_size(device, device->ldev, p_usize, 0);
-		cur_size = drbd_get_capacity(device->this_bdev);
 		if (new_size < cur_size &&
 		    device->state.disk >= D_OUTDATED &&
 		    device->state.conn < C_CONNECTED) {
@@ -4081,9 +4081,36 @@ static int receive_sizes(struct drbd_connection *connection, struct packet_info
 		 *
 		 * However, if he sends a zero current size,
 		 * take his (user-capped or) backing disk size anyways.
+		 *
+		 * Unless of course he does not have a disk himself.
+		 * In which case we ignore this completely.
 		 */
+		sector_t new_size = p_csize ?: p_usize ?: p_size;
 		drbd_reconsider_queue_parameters(device, NULL, o);
-		drbd_set_my_capacity(device, p_csize ?: p_usize ?: p_size);
+		if (new_size == 0) {
+			/* Ignore, peer does not know nothing. */
+		} else if (new_size == cur_size) {
+			/* nothing to do */
+		} else if (cur_size != 0 && p_size == 0) {
+			drbd_warn(device, "Ignored diskless peer device size (peer:%llu != me:%llu sectors)!\n",
+					(unsigned long long)new_size, (unsigned long long)cur_size);
+		} else if (new_size < cur_size && device->state.role == R_PRIMARY) {
+			drbd_err(device, "The peer's device size is too small! (%llu < %llu sectors); demote me first!\n",
+					(unsigned long long)new_size, (unsigned long long)cur_size);
+			conn_request_state(peer_device->connection, NS(conn, C_DISCONNECTING), CS_HARD);
+			return -EIO;
+		} else {
+			/* I believe the peer, if
+			 *  - I don't have a current size myself
+			 *  - we agree on the size anyways
+			 *  - I do have a current size, am Secondary,
+			 *    and he has the only disk
+			 *  - I do have a current size, am Primary,
+			 *    and he has the only disk,
+			 *    which is larger than my current size
+			 */
+			drbd_set_my_capacity(device, new_size);
+		}
 	}
 
 	if (get_ldev(device)) {
-- 
2.17.1


  parent reply	other threads:[~2018-12-20 16:34 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-12-20 16:23 [PATCH 00/17] DRBD updates for 4.21 Lars Ellenberg
2018-12-20 16:23 ` [PATCH 01/17] drbd: narrow rcu_read_lock in drbd_sync_handshake Lars Ellenberg
2018-12-20 16:23 ` [PATCH 02/17] drbd: must not use connection after kref_put(&connection->kref) Lars Ellenberg
2018-12-20 16:23 ` [PATCH 03/17] drbd: centralize printk reporting of new size into drbd_set_my_capacity() Lars Ellenberg
2018-12-20 16:23 ` Lars Ellenberg [this message]
2018-12-20 16:23 ` [PATCH 05/17] drbd: disconnect, if the wrong UUIDs are attached on a connected peer Lars Ellenberg
2018-12-20 16:23 ` [PATCH 06/17] drbd: fix confusing error message during attach Lars Ellenberg
2018-12-20 16:23 ` [PATCH 07/17] drbd: attach on connected diskless peer must not shrink a consistent device Lars Ellenberg
2018-12-20 16:23 ` [PATCH 08/17] drbd: reject attach of unsuitable uuids even if connected Lars Ellenberg
2018-12-20 16:23 ` [PATCH 09/17] drbd: fix comment typos Lars Ellenberg
2018-12-20 16:23 ` [PATCH 10/17] drbd: do not block when adjusting "disk-options" while IO is frozen Lars Ellenberg
2018-12-20 16:23 ` [PATCH 11/17] drbd: avoid spurious self-outdating with concurrent disconnect / down Lars Ellenberg
2018-12-20 16:23 ` [PATCH 12/17] drbd: fix print_st_err()'s prototype to match the definition Lars Ellenberg
2018-12-20 16:23 ` [PATCH 13/17] drbd: don't retry connection if peers do not agree on "authentication" settings Lars Ellenberg
2018-12-20 16:23 ` [PATCH 14/17] drbd: skip spurious timeout (ping-timeo) when failing promote Lars Ellenberg
2018-12-20 16:23 ` [PATCH 15/17] drbd: introduce P_ZEROES (REQ_OP_WRITE_ZEROES on the "wire") Lars Ellenberg
2018-12-20 16:23 ` [PATCH 16/17] drbd: Avoid Clang warning about pointless switch statment Lars Ellenberg
2018-12-20 16:23 ` [PATCH 17/17] drbd: Change drbd_request_detach_interruptible's return type to int Lars Ellenberg
2018-12-20 16:27 ` [PATCH 00/17] DRBD updates for 4.21 Jens Axboe

Reply instructions:

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

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

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

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

  git send-email \
    --in-reply-to=20181220162344.8430-5-lars.ellenberg@linbit.com \
    --to=lars.ellenberg@linbit.com \
    --cc=axboe@kernel.dk \
    --cc=drbd-dev@lists.linbit.com \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

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

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome