In mm_cid_fixup_cpus_to_tasks(), when rq->curr has the target mm and mm_cid.active is set, the CID is checked with cid_in_transit() before setting the transition bit. In per-CPU mode a newly forked or exec'd task can be running with mm_cid.cid == MM_CID_UNSET because CIDs are assigned lazily on schedule-in. With cid_in_transit() the guard passes for MM_CID_UNSET (no transit bit), converts it to MM_CID_UNSET | MM_CID_TRANSIT and stores it back; later mm_cid_schedout() feeds this to clear_bit() with MM_CID_UNSET as the bit number, triggering an out-of-bounds write. Symptoms: this is genuine memory corruption, but a bounded out-of-bounds write, not an arbitrary one. MM_CID_UNSET is the fixed sentinel BIT(31), so once the bad value reaches mm_cid_schedout() the cid_from_transit_cid() strip leaves MM_CID_UNSET, which fails the "cid < max_cids" convergence test and falls into mm_drop_cid() -> clear_bit(MM_CID_UNSET, mm_cidmask(mm)). The cid bitmap is embedded in the mm_struct slab object (after cpu_bitmap and mm_cpus_allowed) and is only num_possible_cpus() bits wide, so clearing bit 31 is a deterministic OOB bit-clear at a fixed offset of 2^31 / 8 == 256 MiB past the bitmap base. The address is not attacker-influenced (fixed sentinel -> fixed offset) and the op only clears a single bit; what sits 256 MiB further along the direct map is whatever kernel object happens to live there, so this corrupts one bit of unpredictable kernel memory -- it is not an arbitrary-address or arbitrary-value write. It triggers only in per-CPU CID mode, when a CPU is running an active task of the target mm whose cid is still MM_CID_UNSET -- the fork()/execve() window before that task's next schedule-in assigns it a real CID -- and a per-CPU -> per-task fixup walks over it (the mode fallback driven by a thread exit, sched_mm_cid_exit(), or by the deferred max_cids recompute in mm_cid_work_fn()). In practice syzkaller surfaced it as a KASAN use-after-free reported in __schedule -> mm_cid_switch_to, where the offending clear_bit() is inlined via mm_cid_schedout() -> mm_drop_cid(). Guard the transition-bit assignment against MM_CID_UNSET, in addition to the existing cid_in_transit() check, so the bit is only set on a genuine task-owned CID. A CPU-owned (MM_CID_ONCPU) CID of a running active task is handled by the cid_on_cpu(pcp->cid) branch above and never reaches this path, so excluding MM_CID_UNSET (and the already-transitioning case) is sufficient.
Unified · Split
| @@ -10909,8 +10909,19 @@ static void mm_cid_fixup_cpus_to_tasks(struct mm_struct *mm) · 10908 unmodified lines | |||
| 10909 | } else if (rq->curr->mm == mm && rq->curr->mm_cid.active) { | 10909 | } else if (rq->curr->mm == mm && rq->curr->mm_cid.active) { |
| 10910 | unsigned int cid = rq->curr->mm_cid.cid; | 10910 | unsigned int cid = rq->curr->mm_cid.cid; |
| 10911 | 10911 | ||
| 10912 | /* Ensure it has the transition bit set */ | 10912 | /* |
| 10913 | if (!cid_in_transit(cid)) { | 10913 | * Set the transition bit only on a genuine task-owned |
| 10914 | * CID. A running active task can legitimately have | ||
| 10915 | * MM_CID_UNSET here: in per-CPU mode CIDs are assigned | ||
| 10916 | * lazily on schedule-in, so the fork()/execve() window | ||
| 10917 | * leaves the task active with no owned CID. Setting the | ||
| 10918 | * transition bit on MM_CID_UNSET would later feed | ||
| 10919 | * clear_bit() an out-of-bounds bit number via | ||
| 10920 | * mm_cid_schedout(), so exclude it. A CPU-owned | ||
| 10921 | * (MM_CID_ONCPU) CID is handled by the cid_on_cpu() | ||
| 10922 | * branch above and never reaches here. | ||
| 10923 | */ | ||
| 10924 | if (cid != MM_CID_UNSET && !cid_in_transit(cid)) { | ||
| 10914 | cid = cid_to_transit_cid(cid); | 10925 | cid = cid_to_transit_cid(cid); |
| 10915 | rq->curr->mm_cid.cid = cid; | 10926 | rq->curr->mm_cid.cid = cid; |
| 10916 | pcp->cid = cid; | 10927 | pcp->cid = cid; |