public inbox for openembedded-core@lists.openembedded.org
 help / color / mirror / Atom feed
From: Steve Sakoman <steve@sakoman.com>
To: openembedded-core@lists.openembedded.org
Subject: [OE-core][kirkstone 08/31] wayland: fix CVE-2021-3782
Date: Thu,  3 Nov 2022 17:00:43 -1000	[thread overview]
Message-ID: <2abd43bc1f0d4ea77603920337bc53add4c42394.1667530733.git.steve@sakoman.com> (raw)
In-Reply-To: <cover.1667530733.git.steve@sakoman.com>

From: Narpat Mali <narpat.mali@windriver.com>

An internal reference count is held on the buffer pool,
incremented every time a new buffer is created from the pool.
The reference count is maintained as an int;
on LP64 systems this can cause thereference count to overflow if
the client creates a large number of wl_shm buffer objects,
or if it can coerce the server to create a large number of external references
to the buffer storage. With the reference count overflowing, a use-after-free
can be constructed on the wl_shm_pool tracking structure,
where values may be incremented or decremented;
it may also be possible to construct a limited oracle to leak 4 bytes of
server-side memory to the attacking client at a time.

Reference:
https://nvd.nist.gov/vuln/detail/CVE-2021-3782

Upstream patch:
https://gitlab.freedesktop.org/wayland/wayland/-/commit/b19488c7154b902354cb26a27f11415d7799b0b2

Signed-off-by: Narpat Mali <narpat.mali@windriver.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
---
 .../wayland/wayland/CVE-2021-3782.patch       | 111 ++++++++++++++++++
 .../wayland/wayland_1.20.0.bb                 |   2 +
 2 files changed, 113 insertions(+)
 create mode 100644 meta/recipes-graphics/wayland/wayland/CVE-2021-3782.patch

diff --git a/meta/recipes-graphics/wayland/wayland/CVE-2021-3782.patch b/meta/recipes-graphics/wayland/wayland/CVE-2021-3782.patch
new file mode 100644
index 0000000000..df204508e9
--- /dev/null
+++ b/meta/recipes-graphics/wayland/wayland/CVE-2021-3782.patch
@@ -0,0 +1,111 @@
+From 5eed6609619cc2e4eaa8618d11c15d442abf54be Mon Sep 17 00:00:00 2001
+From: Derek Foreman <derek.foreman@collabora.com>
+Date: Fri, 28 Jan 2022 13:18:37 -0600
+Subject: [PATCH] util: Limit size of wl_map
+
+Since server IDs are basically indistinguishable from really big client
+IDs at many points in the source, it's theoretically possible to overflow
+a map and either overflow server IDs into the client ID space, or grow
+client IDs into the server ID space. This would currently take a massive
+amount of RAM, but the definition of massive changes yearly.
+
+Prevent this by placing a ridiculous but arbitrary upper bound on the
+number of items we can put in a map: 0xF00000, somewhere over 15 million.
+This should satisfy pathological clients without restriction, but stays
+well clear of the 0xFF000000 transition point between server and client
+IDs. It will still take an improbable amount of RAM to hit this, and a
+client could still exhaust all RAM in this way, but our goal is to prevent
+overflow and undefined behaviour.
+
+Fixes #224
+
+Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
+
+Upstream-Status: Backport
+CVE: CVE-2021-3782
+
+Reference to upstream patch:
+https://gitlab.freedesktop.org/wayland/wayland/-/commit/b19488c7154b902354cb26a27f11415d7799b0b2
+
+[DP: adjust context for wayland version 1.20.0]
+Signed-off-by: Dragos-Marian Panait <dragos.panait@windriver.com>
+---
+ src/wayland-private.h |  1 +
+ src/wayland-util.c    | 25 +++++++++++++++++++++++--
+ 2 files changed, 24 insertions(+), 2 deletions(-)
+
+diff --git a/src/wayland-private.h b/src/wayland-private.h
+index 9bf8cb7..35dc40e 100644
+--- a/src/wayland-private.h
++++ b/src/wayland-private.h
+@@ -45,6 +45,7 @@
+ #define WL_MAP_SERVER_SIDE 0
+ #define WL_MAP_CLIENT_SIDE 1
+ #define WL_SERVER_ID_START 0xff000000
++#define WL_MAP_MAX_OBJECTS 0x00f00000
+ #define WL_CLOSURE_MAX_ARGS 20
+ 
+ struct wl_object {
+diff --git a/src/wayland-util.c b/src/wayland-util.c
+index d5973bf..3e45d19 100644
+--- a/src/wayland-util.c
++++ b/src/wayland-util.c
+@@ -195,6 +195,7 @@ wl_map_insert_new(struct wl_map *map, uint32_t flags, void *data)
+ 	union map_entry *start, *entry;
+ 	struct wl_array *entries;
+ 	uint32_t base;
++	uint32_t count;
+ 
+ 	if (map->side == WL_MAP_CLIENT_SIDE) {
+ 		entries = &map->client_entries;
+@@ -215,10 +216,25 @@ wl_map_insert_new(struct wl_map *map, uint32_t flags, void *data)
+ 		start = entries->data;
+ 	}
+ 
++	/* wl_array only grows, so if we have too many objects at
++	 * this point there's no way to clean up. We could be more
++	 * pro-active about trying to avoid this allocation, but
++	 * it doesn't really matter because at this point there is
++	 * nothing to be done but disconnect the client and delete
++	 * the whole array either way.
++	 */
++	count = entry - start;
++	if (count > WL_MAP_MAX_OBJECTS) {
++		/* entry->data is freshly malloced garbage, so we'd
++		 * better make it a NULL so wl_map_for_each doesn't
++		 * dereference it later. */
++		entry->data = NULL;
++		return 0;
++	}
+ 	entry->data = data;
+ 	entry->next |= (flags & 0x1) << 1;
+ 
+-	return (entry - start) + base;
++	return count + base;
+ }
+ 
+ int
+@@ -235,6 +251,9 @@ wl_map_insert_at(struct wl_map *map, uint32_t flags, uint32_t i, void *data)
+ 		i -= WL_SERVER_ID_START;
+ 	}
+ 
++	if (i > WL_MAP_MAX_OBJECTS)
++		return -1;
++
+ 	count = entries->size / sizeof *start;
+ 	if (count < i)
+ 		return -1;
+@@ -269,8 +288,10 @@ wl_map_reserve_new(struct wl_map *map, uint32_t i)
+ 		i -= WL_SERVER_ID_START;
+ 	}
+ 
+-	count = entries->size / sizeof *start;
++	if (i > WL_MAP_MAX_OBJECTS)
++		return -1;
+ 
++	count = entries->size / sizeof *start;
+ 	if (count < i)
+ 		return -1;
+ 
+-- 
+2.37.3
diff --git a/meta/recipes-graphics/wayland/wayland_1.20.0.bb b/meta/recipes-graphics/wayland/wayland_1.20.0.bb
index bd437767b2..9351d2ed6a 100644
--- a/meta/recipes-graphics/wayland/wayland_1.20.0.bb
+++ b/meta/recipes-graphics/wayland/wayland_1.20.0.bb
@@ -16,7 +16,9 @@ SRC_URI = "https://wayland.freedesktop.org/releases/${BPN}-${PV}.tar.xz \
            file://run-ptest \
            file://0002-Do-not-hardcode-the-path-to-wayland-scanner.patch \
            file://0001-build-Fix-strndup-detection-on-MinGW.patch \
+           file://CVE-2021-3782.patch \
            "
+
 SRC_URI[sha256sum] = "b8a034154c7059772e0fdbd27dbfcda6c732df29cae56a82274f6ec5d7cd8725"
 
 UPSTREAM_CHECK_URI = "https://wayland.freedesktop.org/releases.html"
-- 
2.25.1



  parent reply	other threads:[~2022-11-04  3:01 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-04  3:00 [OE-core][kirkstone 00/31] Patch review Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 01/31] openssl: export necessary env vars in SDK Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 02/31] openssl: Fix SSL_CERT_FILE to match ca-certs location Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 03/31] openssl: CVE-2022-3358 Using a Custom Cipher with NID_undef may lead to NULL encryption Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 04/31] openssl: Upgrade 3.0.5 -> 3.0.7 Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 05/31] lighttpd: fix CVE-2022-41556 Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 06/31] tiff: fix CVE-2022-2953 Steve Sakoman
2022-11-07  9:10   ` Shubham Kulkarni
2022-11-07 14:19     ` Steve Sakoman
2022-11-09 10:27       ` Shubham Kulkarni
2022-11-04  3:00 ` [OE-core][kirkstone 07/31] expat: backport the fix for CVE-2022-43680 Steve Sakoman
2022-11-04  3:00 ` Steve Sakoman [this message]
2022-11-04  3:00 ` [OE-core][kirkstone 09/31] cve-update-db-native: add timeout to urlopen() calls Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 10/31] vim: Upgrade 9.0.0598 -> 9.0.0614 Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 11/31] vim: upgrade 9.0.0614 -> 9.0.0820 Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 12/31] ifupdown: upgrade 0.8.37 -> 0.8.39 Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 13/31] scripts/oe-check-sstate: cleanup Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 14/31] scripts/oe-check-sstate: force build to run for all targets, specifically populate_sysroot Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 15/31] psplash: add psplash-default in rdepends Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 16/31] opkg-utils: use a git clone, not a dynamic snapshot Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 17/31] insane.bbclass: Allow hashlib version that only accepts on parameter Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 18/31] oe/packagemanager/rpm: don't leak file objects Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 19/31] u-boot: Remove duplicate inherit of cml1 Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 20/31] bluez5: add dbus to RDEPENDS Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 21/31] glib-2.0: fix rare GFileInfo test case failure Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 22/31] gnutls: Unified package names to lower-case Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 23/31] meson: make wrapper options sub-command specific Steve Sakoman
2022-11-04  3:00 ` [OE-core][kirkstone 24/31] buildtools-tarball: export certificates to python and curl Steve Sakoman
2022-11-04  3:01 ` [OE-core][kirkstone 25/31] qemu-native: Add PACKAGECONFIG option for jack Steve Sakoman
2022-11-04  3:01 ` [OE-core][kirkstone 26/31] runqemu: Do not perturb script environment Steve Sakoman
2022-11-04  3:01 ` [OE-core][kirkstone 27/31] runqemu: Fix gl-es argument from causing other arguments to be ignored Steve Sakoman
2022-11-04  3:01 ` [OE-core][kirkstone 28/31] overlayfs: Allow not used mount points Steve Sakoman
2022-11-04  3:01 ` [OE-core][kirkstone 29/31] cmake-native: Fix host tool contamination (Bug: 14951) Steve Sakoman
2022-11-04  3:01 ` [OE-core][kirkstone 30/31] ltp: backport clock_gettime04 fix from upstream Steve Sakoman
2022-11-04  3:01 ` [OE-core][kirkstone 31/31] perf: Depend on native setuptools3 Steve Sakoman

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=2abd43bc1f0d4ea77603920337bc53add4c42394.1667530733.git.steve@sakoman.com \
    --to=steve@sakoman.com \
    --cc=openembedded-core@lists.openembedded.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