Control: tags 854336 + pending
Dear maintainer,
I've prepared an NMU for spice (versioned as 0.12.8-2.1) and
uploaded it to DELAYED/2. Please feel free to tell me if I
should delay it longer.
I am attaching the debdiff to this bug report.
Regards,
Markus
diff -Nru spice-0.12.8/debian/changelog spice-0.12.8/debian/changelog
--- spice-0.12.8/debian/changelog 2017-01-06 14:50:55.0 +0100
+++ spice-0.12.8/debian/changelog 2017-02-13 21:42:01.0 +0100
@@ -1,3 +1,17 @@
+spice (0.12.8-2.1) unstable; urgency=medium
+
+ * Non-maintainer upload.
+ * Add CVE-2016-9577-and-CVE-2016-9578.patch:
+- CVE-2016-9577: A buffer overflow vulnerability in
+ main_channel_alloc_msg_rcv_buf was found that occurs when reading large
+ messages due to missing buffer size check.
+- CVE-2016-9578: A vulnerability was discovered in the server's
+ protocol handling. An attacker able to connect to the spice server could
+ send crafted messages which would cause the process to crash.
+ (Closes: #854336)
+
+ -- Markus Koschany Mon, 13 Feb 2017 21:42:01 +0100
+
spice (0.12.8-2) unstable; urgency=medium
* Build on all little-endian architectures (Closes: #734218)
diff -Nru spice-0.12.8/debian/patches/CVE-2016-9577-and-CVE-2016-9578.patch spice-0.12.8/debian/patches/CVE-2016-9577-and-CVE-2016-9578.patch
--- spice-0.12.8/debian/patches/CVE-2016-9577-and-CVE-2016-9578.patch 1970-01-01 01:00:00.0 +0100
+++ spice-0.12.8/debian/patches/CVE-2016-9577-and-CVE-2016-9578.patch 2017-02-13 21:42:01.0 +0100
@@ -0,0 +1,54 @@
+From: Markus Koschany
+Date: Mon, 13 Feb 2017 21:38:02 +0100
+Subject: CVE-2016-9577 and CVE-2016-9578
+
+Bug-Debian: https://bugs.debian.org/854336
+Origin: http://pkgs.fedoraproject.org/cgit/rpms/spice.git/commit/?id=d919d639ae5f83a9735a04d843eed675f9357c0d
+---
+ server/main_channel.c | 3 +++
+ server/reds.c | 11 ++-
+ 2 files changed, 13 insertions(+), 1 deletion(-)
+
+diff --git a/server/main_channel.c b/server/main_channel.c
+index 0ecc9df..1fc3915 100644
+--- a/server/main_channel.c
b/server/main_channel.c
+@@ -1026,6 +1026,9 @@ static uint8_t *main_channel_alloc_msg_rcv_buf(RedChannelClient *rcc,
+
+ if (type == SPICE_MSGC_MAIN_AGENT_DATA) {
+ return reds_get_agent_data_buffer(mcc, size);
++} else if (size > sizeof(main_chan->recv_buf)) {
++/* message too large, caller will log a message and close the connection */
++return NULL;
+ } else {
+ return main_chan->recv_buf;
+ }
+diff --git a/server/reds.c b/server/reds.c
+index 61bf735..4c60f58 100644
+--- a/server/reds.c
b/server/reds.c
+@@ -2110,6 +2110,14 @@ static void reds_handle_read_link_done(void *opaque)
+ link_mess->num_channel_caps = GUINT32_FROM_LE(link_mess->num_channel_caps);
+ link_mess->num_common_caps = GUINT32_FROM_LE(link_mess->num_common_caps);
+
++/* Prevent DoS. Currently we defined only 13 capabilities,
++ * I expect 1024 to be valid for quite a lot time */
++if (link_mess->num_channel_caps > 1024 || link_mess->num_common_caps > 1024) {
++reds_send_link_error(link, SPICE_LINK_ERR_INVALID_DATA);
++reds_link_free(link);
++return;
++}
++
+ num_caps = link_mess->num_common_caps + link_mess->num_channel_caps;
+ caps = (uint32_t *)((uint8_t *)link_mess + link_mess->caps_offset);
+
+@@ -2202,7 +2210,8 @@ static void reds_handle_read_header_done(void *opaque)
+
+ reds->peer_minor_version = header->minor_version;
+
+-if (header->size < sizeof(SpiceLinkMess)) {
++/* the check for 4096 is to avoid clients to cause arbitrary big memory allocations */
++if (header->size < sizeof(SpiceLinkMess) || header->size > 4096) {
+ reds_send_link_error(link, SPICE_LINK_ERR_INVALID_DATA);
+ spice_warning("bad size %u", header->size);
+ reds_link_free(link);
diff -Nru spice-0.12.8/debian/patches/series spice-0.12.8/debian/patches/series
--- spice-0.12.8/debian/patches/series 2017-01-06 14:50:42.0 +0100
+++ spice-0.12.8/debian/patches/series 2017-02-13 21:42:01.0 +0100
@@ -1 +1,2 @@
stop-linking-with-libcacard.diff
+CVE-2016-9577-and-CVE-2016-9578.patch
