In FreeRDP before version 2.1.2, there is an out of bounds read in license_read_new_or_upgrade_license_packet. A manipulated license packet can lead to out of bound reads to an internal buffer. This is fixed in version 2.1.2.

In FreeRDP before version 2.1.2, there is an out of bounds read in RLEDECOMPRESS. All FreeRDP based clients with sessions with color depth < 32 are affected. This is fixed in version 2.1.2.

In FreeRDP before version 2.1.2, there is an out-of-bound read in glyph_cache_put. This affects all FreeRDP clients with `+glyph-cache` option enabled This is fixed in version 2.1.2.

In FreeRDP before version 2.1.2, there is a use-after-free in gdi_SelectObject. All FreeRDP clients using compatibility mode with /relax-order-checks are affected. This is fixed in version 2.1.2.

In FreeRDP before version 2.1.2, there is a global OOB read in update_read_cache_bitmap_v3_order. As a workaround, one can disable bitmap cache with -bitmap-cache (default). This is fixed in version 2.1.2.

In FreeRDP before version 2.1.2, there is an integer casting vulnerability in update_recv_secondary_order. All clients with +glyph-cache /relax-order-checks are affected. This is fixed in version 2.1.2.

In FreeRDP before version 2.1.2, an out of bounds read occurs resulting in accessing a memory location that is outside of the boundaries of the static array PRIMARY_DRAWING_ORDER_FIELD_BYTES. This is fixed in version 2.1.2.

In FreeRDP before version 2.1.2, an out of bound reads occurs resulting in accessing a memory location that is outside of the boundaries of the static array PRIMARY_DRAWING_ORDER_FIELD_BYTES. This is fixed in version 2.1.2.

In FreeRDP before version 2.1.2, there is an out of bounds read in TrioParse. Logging might bypass string length checks due to an integer overflow. This is fixed in version 2.1.2.

Mutt before 1.14.4 and NeoMutt before 2020-06-19 have a STARTTLS buffering issue that affects IMAP, SMTP, and POP3. When a server sends a "begin TLS" response, the client reads additional data (e.g., from a man-in-the-middle attacker) and evaluates it in a TLS context, aka "response injection."