From: Jakob Koschel <jakobkoschel@xxxxxxxxx> commit 28cb138f559f8c1a1395f5564f86b8bbee83631b upstream. in vt_setactivate an almost identical code path has been patched with array_index_nospec. In the VT_ACTIVATE path the user input is from a system call argument instead of a usercopy. For consistency both code paths should have the same mitigations applied. Kasper Acknowledgements: Jakob Koschel, Brian Johannesmeyer, Kaveh Razavi, Herbert Bos, Cristiano Giuffrida from the VUSec group at VU Amsterdam. Co-developed-by: Brian Johannesmeyer <bjohannesmeyer@xxxxxxxxx> Signed-off-by: Brian Johannesmeyer <bjohannesmeyer@xxxxxxxxx> Signed-off-by: Jakob Koschel <jakobkoschel@xxxxxxxxx> Link: https://lore.kernel.org/r/20220127144406.3589293-2-jakobkoschel@xxxxxxxxx Cc: stable <stable@xxxxxxxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/tty/vt/vt_ioctl.c | 1 + 1 file changed, 1 insertion(+) --- a/drivers/tty/vt/vt_ioctl.c +++ b/drivers/tty/vt/vt_ioctl.c @@ -694,6 +694,7 @@ int vt_ioctl(struct tty_struct *tty, ret = -ENXIO; else { arg--; + arg = array_index_nospec(arg, MAX_NR_CONSOLES); console_lock(); ret = vc_allocate(arg); console_unlock(); Patches currently in stable-queue which might be from jakobkoschel@xxxxxxxxx are queue-4.9/vt_ioctl-add-array_index_nospec-to-vt_activate.patch queue-4.9/vt_ioctl-fix-array_index_nospec-in-vt_setactivate.patch