Bugzilla – Bug 1170828
VUL-0: CVE-2020-12465: kernel-source: remotely triggered kernel memory corruption by malformed IEEE 802.11 packet in mt76 driver
Last modified: 2024-06-25 14:45:45 UTC
This is from upstream but it needs a CVE b102f0c522cf668c8382c56a4f771b37d011cda2 ("mt76: fix array overflow on receiving too many fragments for a packet") The number of fragments is determined by the incoming packet. That means that anybody within radio reception can overwrite a part of the kernel's heap. Strictly speaking only stuff from 4.16 onwards is vulnerable. I will look at older driver code for a similar issue. Takashi has taken the fix for SLE15-SP2, but this needs to be tracked as a CVE.
*** Bug 1170888 has been marked as a duplicate of this bug. ***
CVE-2020-12465 An array overflow was discovered in mt76_add_fragment in drivers/net/wireless/mediatek/mt76/dma.c in the Linux kernel before 5.5.10, aka CID-b102f0c522cf. An oversized packet with too many rx fragments can corrupt memory of adjacent pages.
The mt76 wireless driver first appeared in SLE15-SP2 which is already fixed.
Added CVE and bsc to patch
(In reply to Wolfgang Frisch from comment #4) > The mt76 wireless driver first appeared in SLE15-SP2 which is already fixed. Yes, as far as I can tell, the older code is OK.
done