Hello:

Just got this in my box.

Best,

A.

---

This release contains fixes for the issues reported in today's security
advisory: https://lists.x.org/archives/xorg/2024- … 61525.html

* CVE-2023-6816
* CVE-2024-0229
* CVE-2024-21885
* CVE-2024-21886
* CVE-2024-0408
* CVE-2024-0409

Additionally, it also contains several other fixes for glamor, libEI support,
and FreeBSD.

Jan Beich (2):
       os: Use LOCAL_PEERCRED to determine local client PID on FreeBSD
       os: Use KERN_PROC_ARGS to determine client command on DragonFly and FreeBSD

José Expósito (2):
       Xi: do not keep linked list pointer during recursion
       Bump version to 23.2.4

Michel Dänzer (3):
       glamor: Don't override source alpha to 1.0 if it's used for blending
       glamor: Make glamor_set_alu take a DrawablePtr
       glamor: Fall back for mixed depth 24/32 in glamor_set_alu

Olivier Fourdan (3):
       xwayland: Pass the correct oeffis device types
       glx: Call XACE hooks on the GLX buffer
       ephyr,xwayland: Use the proper private key for cursor

Peter Hutterer (10):
       Xi: require a pointer and keyboard device for XIAttachToMaster
       dix: don't allow for devices with 0 axes
       xwayland: override the XTest sendEventsProc for all devices
       dix: initialize the XTest sendEventsProc for all devices
       dix: allocate enough space for logical button maps
       dix: Allocate sufficient xEvents for our DeviceStateNotify
       dix: fix DeviceStateNotify event calculation
       Xi: when creating a new ButtonClass, set the number of buttons
       Xi: flush hierarchy events after adding/removing master devices
       dix: when disabling a master, float disabled slaved devices too

git tag: xwayland-23.2.4

https://xorg.freedesktop.org/archive/in … 2.4.tar.xz
SHA256: a99e159b6d0d33098b3b6ab22a88bfcece23c8b9d0ca72c535c55dcb0681b46b  xwayland-23.2.4.tar.xz
SHA512: ac3ff208cbef5bbe4637c335cfda226489c93b0a3768f2f4fb0201c588485ede38262fbce77ef1425b3d2a0be61b6580df53341c7b95e6072c8b6371ad29d187  xwayland-23.2.4.tar.xz
PGP:  https://xorg.freedesktop.org/archive/in … tar.xz.sig

---

Best,

A.

Hello:

Just got this in my box.

Good to see X.Org at work.
---

Issues in X.Org X server prior to 21.1.11 and Xwayland prior to 23.2.4
=====================================================
Multiple issues have been found in the X server and Xwayland implementations
published by X.Org for which we are releasing security fixes for in
xorg-server-21.1.11 and xwayland-23.2.4.

1) CVE-2023-6816 can be triggered by passing an invalid array index to DeviceFocusEvent or ProcXIQueryPointer.

2) CVE-2024-0229 can be triggered if a device has both a button and a key class and zero buttons.

3) CVE-2024-21885 can be triggered if a device with a given ID was removed and a new device with the same ID added both in the same operation.

4) CVE-2024-21886 can be triggered by disabling a master device with disabled slave devices.

5) CVE-2024-0409 can be triggered by enabling SELinux xserver_object_manager and running a client.

6) CVE-2024-0408 can be triggered by enabling SELinux xserver_object_manager and creating a GLX PBuffer.

------------------------------------------------------------------------

1) CVE-2023-6816: Heap buffer overflow in DeviceFocusEvent and ProcXIQueryPointer

Introduced in: xorg-server-1.13.0 (2012)
Fixed in: xorg-server-21.1.11 and xwayland-23.2.4
Fix: https://gitlab.freedesktop.org/xorg/xse … 3c58a9e7e3
Found by: Jan-Niklas Sohn working with Trend Micro Zero Day Initiative

Both DeviceFocusEvent and the XIQueryPointer reply contain a bit for
each logical button currently down. Buttons can be arbitrarily mapped to
any value up to 255 but the X.Org Server was only allocating space for the
device's number of buttons, leading to a heap overflow if a bigger value
was used.

xorg-server-21.1.11 and xwayland-23.2.4 have been patched to fix this issue.

2) CVE-2024-0229: Reattaching to different master device may lead to out-of-bounds memory access

Introduced in: xorg-server-1.1.1 (2006)
Fixed in: xorg-server-21.1.11 and xwayland-23.2.4
Fixes:
  - https://gitlab.freedesktop.org/xorg/xse … 636109d6a5
  - https://gitlab.freedesktop.org/xorg/xse … cde53553d5
  - https://gitlab.freedesktop.org/xorg/xse … e0d5981b74
Found by: Jan-Niklas Sohn working with Trend Micro Zero Day Initiative

If a device has both a button class and a key class and numButtons is
zero, we can get an out-of-bounds write due to event under-allocation in
the DeliverStateNotifyEvent function.

xorg-server-21.1.11 and xwayland-23.2.4 have been patched to fix this issue.

3) CVE-2024-21885: Heap buffer overflow in XISendDeviceHierarchyEvent

Introduced in: xorg-server-1.10.0 (2011)
Fixed in: xorg-server-21.1.11 and xwayland-23.2.4
Fix: https://gitlab.freedesktop.org/xorg/xse … dce503cbd1
Found by: Jan-Niklas Sohn working with Trend Micro Zero Day Initiative

The XISendDeviceHierarchyEvent() function allocates space to store up
to MAXDEVICES (256) xXIHierarchyInfo structures in info.

If a device with a given ID was removed and a new device with the same
ID added both in the same operation, the single device ID will lead to
two info structures being written to info.

Since this case can occur for every device ID at once, a total of two
times MAXDEVICES info structures might be written to the allocation,
leading to a heap buffer overflow.

xorg-server-21.1.11 and xwayland-23.2.4 have been patched to fix this issue.

4) CVE-2024-21886: Heap buffer overflow in DisableDevice

Introduced in: xorg-server-1.13.0 (2012)
Fixed in: xorg-server-21.1.11 and xwayland-23.2.4
Fixes:
  - https://gitlab.freedesktop.org/xorg/xse … 54dd0ce36b
  - https://gitlab.freedesktop.org/xorg/xse … 10cc07c3a8
Found by: Jan-Niklas Sohn working with Trend Micro Zero Day Initiative

The DisableDevice() function is called whenever an enabled device
is disabled and it moves the device from the inputInfo.devices linked
list to the inputInfo.off_devices linked list.

However, its link/unlink operation has an issue during the recursive
call to DisableDevice() due to the prev pointer pointing to a
removed device.

This issue leads to a length mismatch between the total number of
devices and the number of device in the list, leading to a heap
overflow and, possibly, to local privilege escalation.

xorg-server-21.1.11 and xwayland-23.2.4 have been patched to fix this issue.

5) CVE-2024-0409: SELinux context corruption

Introduced in: xorg-server-1.16.0 (2014)
Fixed in: xorg-server-21.1.11 and xwayland-23.2.4
Fix: https://gitlab.freedesktop.org/xorg/xse … ea702c94f7
Found by: Olivier Fourdan

The Xserver uses the mechanism of "privates" to store additional data to its
own objects, each private has an associate "type". Each private is allocated
for the relevant size of memory that is declared at creation.

The cursor structure in the Xserver goes as far as having two keys, one for
the cursor itself and another one for the bits that make the cursor shape.

XSELINUX also uses privates but it's a bit of a special case because it uses
the same privates keys for all different objects.

What happens here is that the cursor code in both Xephyr and Xwayland uses the
wrong type of private at creation, using the cursor bits type with the cursor
private and when initiating the cursor, the overwrites the XSELINUX context.

xorg-server-21.1.11 and xwayland-23.2.4 have been patched to fix this issue.

6) CVE-2024-0408: SELinux unlabeled GLX PBuffer

Introduced in: xorg-server-1.10.0 (2011)
Fixed in: xorg-server-21.1.11 and xwayland-23.2.4
Fix: https://gitlab.freedesktop.org/xorg/xse … fe5e15dac3
Found by: Olivier Fourdan and Donn Seeley

The XSELINUX code in the Xserver labels the X resources based on a hook. What
happens here is that the GLX PBuffer code does not call that XACE hook when
creating the buffer, so it remains unlabeled, and when the client issues
another request to access that resource (as here with a GetGeometry) or even
when it creates another resource which needs to access that buffer (such as a
GC), the XSELINUX code will try to use an object that was never labeled and
crash because the SID is NULL.

xorg-server-21.1.11 and xwayland-23.2.4 have been patched to fix this issue.

---

Best,

A.

Hello:

Just got this in my inbox.
Things 'X11' continue to roll along steadily. 8^)

Best,

A.

----

Announce: xterm-389 - 2024/01/01

Files:
  https://invisible-island.net/archives/x … rm-389.tgz
  https://invisible-island.net/archives/x … 89.tgz.asc
  https://invisible-island.net/archives/x … 9.patch.gz
  https://invisible-island.net/archives/x … tch.gz.asc
  https://invisible-island.net/archives/x … rm-389.tgz
  https://invisible-island.net/archives/x … 89.tgz.asc

                            Patch #389 - 2024/01/01

     * interchange  variables  in subparameter parsing, fixing a bug where
       subparameters  after  the  first  parameter  could be misidentified
       (patch by Adam Saponara).
     * correct  popping of icon/window titles in a case where only one was
       pushed from patch #385 changes.
     * add XTQMODKEYS response in DECRQSS, as alternative for vim.
     * correct  DECCIR encoded information on character set size, handle a
       VT525 quirk, and add DECST8C (Windows Terminal #14984).
     * improve  DECRQCRA  (prompted  by  discussion with James Holderness,
       Windows Terminal #14974).
     * add part of VT525 color controls:
          + DECAC,  to  update  default  foreground/background, respond to
            DECRQSS
          + DECATC, to respond with DECRQSS
     * prevent  Unicode  non-characters  from  being  printed (prompted by
       patch by Grady Martin).
     * modify  send_SGR()  to  avoid modifying colors 16 to 255 in printed
       output (patch by Grady Martin).
     * minor cleanup of miscellaneous error-codes with ERROR_MISC.
     * remove legacy CSI 53 for locator status, corrected in patch #294.
     * modify  DECRQUPSS  and  DECAUPSS feature to support VT5xx character
       sets (report by Thomas Wolff).
     * improve configure script:
          + reduce  configure-check compiler warnings (prompted by Florian
            Weimer, Redhat #2251945)
          + improve  usage messages in configure script to make it clearer
            when an option value is optional.
     * improve EWMH handling (report/analysis by Edward Rosten)
          + reset  _NET_WM_STATE_HIDDEN  flag  from  _NET_WM_STATE  before
            mapping the window to deiconify.
          + cache  X  properties  to reduce latency (adapted from patch by
            Edward Rosten).

----