mm: don't try to NUMA-migrate COW pages that have other uses
Linux
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
| No fix yet |
| Linux Linux |
|
Linux
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
| No fix yet |
| Linux Linux |
|
|
In the Linux kernel, the following vulnerability has been resolved:
mm: don't try to NUMA-migrate COW pages that have other uses
Oded Gabbay reports that enabling NUMA balancing causes corruption with his Gaudi accelerator test load:
"All the details are in the bug, but the bottom line is that somehow, this patch causes corruption when the numa balancing feature is enabled AND we don't use process affinity AND we use GUP to pin pages so our accelerator can DMA to/from system memory.
Either disabling numa balancing, using process affinity to bind to specific numa-node or reverting this patch causes the bug to disappear"
and Oded bisected the issue to commit 09854ba94c6a ("mm: do_wp_page() simplification").
Now, the NUMA balancing shouldn't actually be changing the writability of a page, and as such shouldn't matter for COW. But it appears it does. Suspicious.
However, regardless of that, the condition for enabling NUMA faults in change_pte_range() is nonsensical. It uses "page_mapcount(page)" to decide if a COW page should be NUMA-protected or not, and that makes absolutely no sense.
From the CVE record · Linux
The Linux kernel can corrupt memory when NUMA balancing is enabled, process affinity is not used, and a workload pins pages with GUP for accelerator DMA. The input describes corruption under these workload conditions; it does not establish an attacker exploit path or authentication requirement.
Engineers running Linux kernels with NUMA balancing and workloads that pin pages with GUP for accelerator DMA should check whether their kernel is affected; affected versions were not provided.