qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Michael Roth <mdroth@linux.vnet.ibm.com>
To: qemu-devel@nongnu.org
Cc: qemu-stable@nongnu.org, Eric Blake <eblake@redhat.com>,
	Paolo Bonzini <pbonzini@redhat.com>
Subject: [Qemu-devel] [PATCH 59/79] nbd: Fix regression on resiliency to port scan
Date: Mon, 28 Aug 2017 19:14:34 -0500	[thread overview]
Message-ID: <1503965694-10794-60-git-send-email-mdroth@linux.vnet.ibm.com> (raw)
In-Reply-To: <1503965694-10794-1-git-send-email-mdroth@linux.vnet.ibm.com>

From: Eric Blake <eblake@redhat.com>

Back in qemu 2.5, qemu-nbd was immune to port probes (a transient
server would not quit, regardless of how many probe connections
came and went, until a connection actually negotiated).  But we
broke that in commit ee7d7aa when removing the return value to
nbd_client_new(), although that patch also introduced a bug causing
an assertion failure on a client that fails negotiation.  We then
made it worse during refactoring in commit 1a6245a (a segfault
before we could even assert); the (masked) assertion was cleaned
up in d3780c2 (still in 2.6), and just recently we finally fixed
the segfault ("nbd: Fully intialize client in case of failed
negotiation").  But that still means that ever since we added
TLS support to qemu-nbd, we have been vulnerable to an ill-timed
port-scan being able to cause a denial of service by taking down
qemu-nbd before a real client has a chance to connect.

Since negotiation is now handled asynchronously via coroutines,
we no longer have a synchronous point of return by re-adding a
return value to nbd_client_new().  So this patch instead wires
things up to pass the negotiation status through the close_fn
callback function.

Simple test across two terminals:
$ qemu-nbd -f raw -p 30001 file
$ nmap 127.0.0.1 -p 30001 && \
  qemu-io -c 'r 0 512' -f raw nbd://localhost:30001

Note that this patch does not change what constitutes successful
negotiation (thus, a client must enter transmission phase before
that client can be considered as a reason to terminate the server
when the connection ends).  Perhaps we may want to tweak things
in a later patch to also treat a client that uses NBD_OPT_ABORT
as being a 'successful' negotiation (the client correctly talked
the NBD protocol, and informed us it was not going to use our
export after all), but that's a discussion for another day.

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1451614

Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20170608222617.20376-1-eblake@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
(cherry picked from commit 0c9390d978cbf61e8f16c9f580fa96b305c43568)
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
---
 blockdev-nbd.c      |  6 +++++-
 include/block/nbd.h |  2 +-
 nbd/server.c        | 24 +++++++++++++++---------
 qemu-nbd.c          |  4 ++--
 4 files changed, 23 insertions(+), 13 deletions(-)

diff --git a/blockdev-nbd.c b/blockdev-nbd.c
index 8a11807..8d7284a 100644
--- a/blockdev-nbd.c
+++ b/blockdev-nbd.c
@@ -27,6 +27,10 @@ typedef struct NBDServerData {
 
 static NBDServerData *nbd_server;
 
+static void nbd_blockdev_client_closed(NBDClient *client, bool ignored)
+{
+    nbd_client_put(client);
+}
 
 static gboolean nbd_accept(QIOChannel *ioc, GIOCondition condition,
                            gpointer opaque)
@@ -46,7 +50,7 @@ static gboolean nbd_accept(QIOChannel *ioc, GIOCondition condition,
     qio_channel_set_name(QIO_CHANNEL(cioc), "nbd-server");
     nbd_client_new(NULL, cioc,
                    nbd_server->tlscreds, NULL,
-                   nbd_client_put);
+                   nbd_blockdev_client_closed);
     object_unref(OBJECT(cioc));
     return TRUE;
 }
diff --git a/include/block/nbd.h b/include/block/nbd.h
index 3e373f0..b69c30d 100644
--- a/include/block/nbd.h
+++ b/include/block/nbd.h
@@ -160,7 +160,7 @@ void nbd_client_new(NBDExport *exp,
                     QIOChannelSocket *sioc,
                     QCryptoTLSCreds *tlscreds,
                     const char *tlsaclname,
-                    void (*close)(NBDClient *));
+                    void (*close_fn)(NBDClient *, bool));
 void nbd_client_get(NBDClient *client);
 void nbd_client_put(NBDClient *client);
 
diff --git a/nbd/server.c b/nbd/server.c
index edfda84..a98bb21 100644
--- a/nbd/server.c
+++ b/nbd/server.c
@@ -81,7 +81,7 @@ static QTAILQ_HEAD(, NBDExport) exports = QTAILQ_HEAD_INITIALIZER(exports);
 
 struct NBDClient {
     int refcount;
-    void (*close)(NBDClient *client);
+    void (*close_fn)(NBDClient *client, bool negotiated);
 
     bool no_zeroes;
     NBDExport *exp;
@@ -796,7 +796,7 @@ void nbd_client_put(NBDClient *client)
     }
 }
 
-static void client_close(NBDClient *client)
+static void client_close(NBDClient *client, bool negotiated)
 {
     if (client->closing) {
         return;
@@ -811,8 +811,8 @@ static void client_close(NBDClient *client)
                          NULL);
 
     /* Also tell the client, so that they release their reference.  */
-    if (client->close) {
-        client->close(client);
+    if (client->close_fn) {
+        client->close_fn(client, negotiated);
     }
 }
 
@@ -993,7 +993,7 @@ void nbd_export_close(NBDExport *exp)
 
     nbd_export_get(exp);
     QTAILQ_FOREACH_SAFE(client, &exp->clients, next, next) {
-        client_close(client);
+        client_close(client, true);
     }
     nbd_export_set_name(exp, NULL);
     nbd_export_set_description(exp, NULL);
@@ -1355,7 +1355,7 @@ done:
 
 out:
     nbd_request_put(req);
-    client_close(client);
+    client_close(client, true);
     nbd_client_put(client);
 }
 
@@ -1381,7 +1381,7 @@ static coroutine_fn void nbd_co_client_start(void *opaque)
     qemu_co_mutex_init(&client->send_lock);
 
     if (nbd_negotiate(data)) {
-        client_close(client);
+        client_close(client, false);
         goto out;
     }
 
@@ -1391,11 +1391,17 @@ out:
     g_free(data);
 }
 
+/*
+ * Create a new client listener on the given export @exp, using the
+ * given channel @sioc.  Begin servicing it in a coroutine.  When the
+ * connection closes, call @close_fn with an indication of whether the
+ * client completed negotiation.
+ */
 void nbd_client_new(NBDExport *exp,
                     QIOChannelSocket *sioc,
                     QCryptoTLSCreds *tlscreds,
                     const char *tlsaclname,
-                    void (*close_fn)(NBDClient *))
+                    void (*close_fn)(NBDClient *, bool))
 {
     NBDClient *client;
     NBDClientNewData *data = g_new(NBDClientNewData, 1);
@@ -1412,7 +1418,7 @@ void nbd_client_new(NBDExport *exp,
     object_ref(OBJECT(client->sioc));
     client->ioc = QIO_CHANNEL(sioc);
     object_ref(OBJECT(client->ioc));
-    client->close = close_fn;
+    client->close_fn = close_fn;
 
     data->client = client;
     data->co = qemu_coroutine_create(nbd_co_client_start, data);
diff --git a/qemu-nbd.c b/qemu-nbd.c
index 14e7947..3b55ffa 100644
--- a/qemu-nbd.c
+++ b/qemu-nbd.c
@@ -335,10 +335,10 @@ static void nbd_export_closed(NBDExport *exp)
 
 static void nbd_update_server_watch(void);
 
-static void nbd_client_closed(NBDClient *client)
+static void nbd_client_closed(NBDClient *client, bool negotiated)
 {
     nb_fds--;
-    if (nb_fds == 0 && !persistent && state == RUNNING) {
+    if (negotiated && nb_fds == 0 && !persistent && state == RUNNING) {
         state = TERMINATE;
     }
     nbd_update_server_watch();
-- 
2.7.4

  parent reply	other threads:[~2017-08-29  0:16 UTC|newest]

Thread overview: 93+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-08-29  0:13 [Qemu-devel] [PATCH 00/79] Patch Round-up for stable 2.9.1, freeze on 2017-09-04 Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 01/79] qga-win: Enable 'can-offline' field in 'guest-get-vcpus' reply Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 02/79] dirty-bitmap: Report BlockDirtyInfo.count in bytes, as documented Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 03/79] qemu-img/convert: Always set ret < 0 on error Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 04/79] qemu-img/convert: Use @opts for one thing only Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 05/79] qemu-img/convert: Move bs_n > 1 && -B check down Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 06/79] block: An empty filename counts as no filename Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 07/79] iotests/051: Add test for empty filename Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 08/79] migration: setup bi-directional I/O channel for exec: protocol Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 09/79] pci: deassert intx when pci device unrealize Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 10/79] block: Do not unref bs->file on error in BD's open Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 11/79] ACPI: don't call acpi_pcihp_device_plug_cb on xen Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 12/79] replication: Make --disable-replication compile again Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 13/79] 9pfs: local: fix unlink of alien files in mapped-file mode Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 14/79] coccinelle: Add script to remove useless QObject casts Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 15/79] qobject: Drop " Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 16/79] qobject: Add helper macros for common scalar insertions Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 17/79] s390x: Drop useless casts Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 18/79] qobject: Use simpler QDict/QList scalar insertion macros Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 19/79] block: Reuse bs as backing hd for drive-backup sync=none Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 20/79] hw/virtio: fix vhost user fails to startup when MQ Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 21/79] aio: add missing aio_notify() to aio_enable_external() Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 22/79] qemu-img: wait for convert coroutines to complete Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 23/79] block/vhdx: Make vhdx_create() always set errp Michael Roth
2017-08-29  0:13 ` [Qemu-devel] [PATCH 24/79] block: Add errp to b{lk, drv}_truncate() Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 25/79] blockdev: use drained_begin/end for qmp_block_resize Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 26/79] target/xtensa: fix mapping direction in read/write simcalls Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 27/79] target/xtensa: fix return value of " Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 28/79] curl: strengthen assertion in curl_clean_state Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 29/79] curl: never invoke callbacks with s->mutex held Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 30/79] curl: avoid recursive locking of BDRVCURLState mutex Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 31/79] stream: fix crash in stream_start() when block_job_create() fails Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 32/79] vvfat: fix qemu-img map and qemu-img convert Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 33/79] virtio: allow broken device to notify guest Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 34/79] virtio-scsi: Unset hotplug handler when unrealize Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 35/79] e1000e: Fix ICR "Other" causes clear logic Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 36/79] s390x/css: catch section mismatch on load Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 37/79] virtio-net: fix wild pointer when remove virtio-net queues Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 38/79] blkdebug: Sanity check block layer guarantees Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 39/79] blkdebug: Refactor error injection Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 40/79] blkdebug: Add pass-through write_zero and discard support Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 41/79] blkdebug: Simplify override logic Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 42/79] blkdebug: Add ability to override unmap geometries Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 43/79] tests: Add coverage for recent block geometry fixes Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 44/79] block: Simplify BDRV_BLOCK_RAW recursion Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 45/79] block: Guarantee that *file is set on bdrv_get_block_status() Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 46/79] mirror: Drop permissions on s->target on completion Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 47/79] virtio-serial-bus: Unset hotplug handler when unrealize Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 48/79] linuxboot_dma: compile for i486 Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 49/79] tests: check-qom-proplist: add checks for cmdline-created objects Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 50/79] monitor: fix object_del for command-line-created objects Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 51/79] pc: Use "min-[x]level" on compat_props Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 52/79] target/ppc: pass const string to kvmppc_is_mem_backend_page_size_ok() Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 53/79] target/ppc: fix memory leak in kvmppc_is_mem_backend_page_size_ok() Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 54/79] spapr: add pre_plug function for memory Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 55/79] spapr: fix memory leak in spapr_memory_pre_plug() Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 56/79] target/xtensa: handle unknown registers in gdbstub Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 57/79] commit: Fix use after free in completion Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 58/79] nbd: Fully initialize client in case of failed negotiation Michael Roth
2017-08-29  0:14 ` Michael Roth [this message]
2017-08-29  0:14 ` [Qemu-devel] [PATCH 60/79] commit: Fix completion with extra reference Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 61/79] blkdebug: Catch bs->exact_filename overflow Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 62/79] blkverify: " Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 63/79] nbd: fix NBD over TLS Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 64/79] block: Do not strcmp() with NULL uri->scheme Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 65/79] 9pfs: local: remove: use correct path component Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 66/79] hid: Reset kbd modifiers on reset Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 67/79] spapr: fix migration to pseries machine < 2.8 Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 68/79] virtio-scsi: finalize IOMMU support Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 69/79] commit: Add NULL check for overlay_bs Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 70/79] spapr: fix memory leak in spapr_core_pre_plug() Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 71/79] virtio-net: fix offload ctrl endian Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 72/79] input: limit kbd queue depth Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 73/79] input: Decrement queue count on kbd delay Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 74/79] qemu-iotests: Test automatic commit job cancel on hot unplug Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 75/79] block: Skip implicit nodes in query-block/blockstats Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 76/79] cpu: don't allow negative core id Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 77/79] hw/i386: allow SHPC for Q35 machine Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 78/79] block/nfs: fix mutex assertion in nfs_file_close() Michael Roth
2017-08-29  0:14 ` [Qemu-devel] [PATCH 79/79] 9pfs: local: fix fchmodat_nofollow() limitations Michael Roth
2017-08-29  0:47 ` [Qemu-devel] [PATCH 00/79] Patch Round-up for stable 2.9.1, freeze on 2017-09-04 Michael Roth
2017-08-29  2:18 ` Thomas Huth
2017-08-29  7:41   ` Cornelia Huck
2017-08-31 16:42   ` Michael Roth
2017-08-31 17:07     ` Peter Maydell
2017-08-31 18:09       ` Michael Roth
2017-08-29 14:04 ` Cole Robinson
2017-08-31 10:21 ` Peter Maydell
2017-08-31 18:00 ` Michael Roth
2017-09-01 23:22   ` Michael Roth
2017-08-31 18:19 ` [Qemu-devel] [Qemu-stable] " Peter Lieven
2017-09-01 19:03 ` Bruce Rogers
2017-09-04 17:33 ` [Qemu-devel] " Peter Maydell

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=1503965694-10794-60-git-send-email-mdroth@linux.vnet.ibm.com \
    --to=mdroth@linux.vnet.ibm.com \
    --cc=eblake@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-stable@nongnu.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;
as well as URLs for NNTP newsgroup(s).