llamaCVEVulnerability Intelligence
/
Linux kernel

CVE-2026-99305

Linux kernel use-after-free read in udlfb framebuffer driver

An unprivileged local user can read freed memory by writing to a world-writable sysfs file.

Specific setups onlySin parche publicadofbdev
Descargar PNG JSON
CVSS 3.1
7.8
Alta
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H Fuente: CNA
ExplotaciónSin evidenciaNo figura en CISA KEV
Exploits públicos0Ninguno conocido
ParcheSin parcheEl fabricante no ha publicado corrección
Exposición3 dDesde la divulgación, 16 nov 2026, 4:59 a. m.
Detección8 minDesde la divulgación hasta su ingreso al índice
Timeline

Ciclo de vida de la vulnerabilidad

Hoy
Divulgación
Publicada
Modificada
Explotación
Remediación
Monitoreo
Detectada +2
Verificada
16 nov 26
19 nov 26
20 nov 26
16 nov 26
17 nov 26
18 nov 26
19 nov 26
20 nov 26
FechaFaseEventoDetalle
Divulgación Publicada CVE-2026-99305 divulgada · CVSS 7.8 (CNA)
Monitoreo Detectada Ingresó al índice
Monitoreo Fuente: NIST NVD Registro visto por primera vez en NIST NVD
Monitoreo Fuente: ENISA EUVD Registro visto por primera vez en ENISA EUVD
Divulgación Modificada Última actualización del registro
Monitoreo Verificada Última verificación contra las fuentes
Resumen
Resumen generado por IA · CC-BY 4.0 · 16 nov 2026

The udlfb framebuffer driver in the Linux kernel can read freed memory. An unprivileged user can trigger this by writing to the world-writable edid sysfs attribute at /sys/class/graphics/fb0/edid, which rebuilds the mode list while the console still points into it. The next console switch then reads the freed mode data.

Affects the Linux kernel udlfb framebuffer driver for DisplayLink USB displays while the framebuffer is open on the console. No vendor rating is given. The fix is upstream commit ac9cfa552102b8e80e88c2872e80dc9cb5f07dda.

¿Qué significa?The driver rebuilds the framebuffer mode list while the framebuffer is still in use by the console. This leaves a use-after-free that can corrupt host kernel memory, crash the console worker, or hang console handling.
¿Quién está afectado?The bug needs a udlfb (DisplayLink USB) framebuffer device present with its EDID sysfs attribute reachable, and the framebuffer open by fbcon so the framebuffer use count is nonzero. To check a system, look for the udlfb module with lsmod and whether a udlfb framebuffer is listed under /sys/class/graphics. No affected version range is given in the checked facts.
¿Qué hacer?Apply upstream Linux kernel commit ac9cfa552102b8e80e88c2872e80dc9cb5f07dda. The checked facts name no fixed kernel version.
Descripción original del registro (Linux)

In the Linux kernel, the following vulnerability has been resolved: fbdev: udlfb: don't rebuild the mode list while the framebuffer is in use dlfb_setup_modes() frees info->modelist with fb_destroy_modelist() and rebuilds it, but fbcon holds raw pointers into that list in fb_display[i].mode and nothing un-publishes them the way do_unregister_framebuffer() and store_modes() do with fbcon_delete_modelist(). The function does refuse to install a new mode while the framebuffer has users (dlfb->fb_count != 0), but only after the list is already destroyed. Reached from the 0666 "edid" sysfs attribute, an unprivileged write to /sys/class/graphics/fb0/edid therefore fails with -EINVAL and still leaves every fb_display[i].mode dangling; the next console switch reads the freed fb_videomode in fb_videomode_to_var(). Test fb_count before touching the list. fbcon takes a reference through fbcon_open(), so that condition covers exactly the states in which fb_display[] points into the list, and the errno returned to userspace is unchanged. BUG: KASAN: slab-use-after-free in fb_videomode_to_var (drivers/video/fbdev/core/modedb.c:905) Read of size 4 at addr ffff8880107b449c by task kworker/1:1/47 Workqueue: events console_callback Call Trace: fb_videomode_to_var (drivers/video/fbdev/core/modedb.c:905) display_to_var (drivers/video/fbdev/core/fbcon.c:998) fbcon_switch (drivers/video/fbdev/core/fbcon.c:2182) redraw_screen (drivers/tty/vt/vt.c:994) complete_change_console (drivers/tty/vt/vt_ioctl.c:1141) console_callback (drivers/tty/vt/vt.c:3358) process_one_work (kernel/workqueue.c:3396) worker_thread (kernel/workqueue.c:3479 kernel/workqueue.c:3560) kthread (kernel/kthread.c:436) ret_from_fork (arch/x86/kernel/process.c:158) ret_from_fork_asm (arch/x86/entry/entry_64.S:245) ... The buggy address is located 28 bytes inside of freed 96-byte region [ffff8880107b4480, ffff8880107b44e0)

Contexto

Condiciones y alcance

Condiciones necesarias para que el fallo sea explotable
01
A framebuffer user exists when the bug is hit: fbcon has the udlfb framebuffer open, so fb_count is nonzero.“console defaults to true, so fbcon_open() has already stored fb_display[i].mode into that list and raised fb_count.”
02
A udlfb (DisplayLink USB) framebuffer device must be present and its EDID sysfs attribute reachable.“dlfb_usb_probe() calls the same helper only while fb_count is still 0.”
¿Quién debe revisar?Teams running Linux kernels with the udlfb DisplayLink USB framebuffer driver and fbcon consoles enabled, including Linux kernel stacks.
Vector CVSS 3.1

Desglose de métricas

Opción resaltada = valor del vector · izquierda = peor caso · CNA
Vector de ataqueAV:L
RedAdyacenteLocalFísica
ComplejidadAC:L
BajaAlta
PrivilegiosPR:L
NingunoBajosAltos
InteracciónUI:N
NingunaRequerida
AlcanceS:U
CambiadoSin cambio
ConfidencialidadC:H
AltoBajoNinguno
IntegridadI:H
AltoBajoNinguno
DisponibilidadA:H
AltoBajoNinguno
Versiones

Versiones afectadas

2.6.34
Remediación

Estado del parche por paquete

Paquete / productoEstadoCorregida enDesde divulgaciónÚltima versión
Linux
Sin parche — — —
Explotación

Arsenal público

Sin exploits, módulos ni PoC públicos conocidos a la fecha de verificación.

Registro
EscenarioAV:L - edid_store() is reached by a local write of 128 bytes to the 0666 sysfs attribute /sys/class/graphics/fbN/edid and calls dlfb_setup_modes(), which fb_destroy_modelist()s info->modelist before the fb_count check. Those bytes are not a USB setup packet; dlfb_usb_probe() calls the same helper only while fb_count is still 0. AC:L - console defaults to true, so fbcon_open() has already stored fb_display[i].mode into that list and raised fb_count. One edid write frees the nodes, and the same user reaches the dangling mode in fbcon_switch() by writing CSI 3J (flush_scrollback) or VT_ACTIVATE on their controlling tty. The pointer stays dangling; nothing else has to race. PR:L - The bin attribute is created with mode 0666 and edid_store() never calls capable(), so a normal local uid can free the modelist. The follow-up fbcon_switch() is permitted on that user's own controlling tty without CAP_SYS_TTY_CONFIG or CAP_SYS_ADMIN. UI:N - The attacker writes edid and then writes CSI 3J to a VT they already control, which flush_scrollback() turns into fbcon_switch(). fbcon already holds the framebuffer open. Nobody else has to attach the DisplayLink device or change consoles. S:U - fbcon_switch() uses the freed fb_videomode in the same kernel that registered the udlfb framebuffer. The impact is host-kernel memory corruption, not a VM exit or an IOMMU bypass. C:H - fb_destroy_modelist() kfree()s the kmalloc-96 fb_modelist while fb_display[i].mode still points at the embedded fb_videomode (KASAN: 4-byte read at offset 28, mode->xres). fb_videomode_to_var() copies xres, yres, pixclock and the margins out of that reclaimable object during fbcon_switch(). I:H - A kmalloc-96 spray replaces the freed fb_modelist, and fbcon_switch() feeds the resulting fb_videomode fields through fb_set_var() into dlfb_ops_set_par() and dlfb_set_video_mode(). The use-after-free therefore drives kernel and device updates from attacker-controlled bytes, rather than stopping at a NULL dereference. A:H - console_callback() runs fbcon_switch() under console_lock(), and the KASAN report is a slab-use-after-free in fb_videomode_to_var(). Faulting the freed fb_modelist oopses that worker and abandons console_lock, which hangs later VT switches and console output.Linux
Referencias
Otrahttps://git.kernel.org/stable/c/291010d5de1c9efc38e98c416b6230500bfae438
Otrahttps://git.kernel.org/stable/c/1e13b8957510cfbc7367b3be1f08adc1757bb6d4
Otrahttps://git.kernel.org/stable/c/4bc608af98a3da0fa0b17af67a64d0d11d625a1d
Otrahttps://git.kernel.org/stable/c/ac9cfa552102b8e80e88c2872e80dc9cb5f07dda

Las referencias externas se muestran para copiar: cve.llama.ec no lo redirige a otros sitios.