net/mlx5e: kTLS, Fix incorrect page refcounting
Linux
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
Linux
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
≥ 84d1bb2b139e0184b1754aa1b5776186b475fce8 · < 2723e8b2cbd486cb96e5a61b22473f7fd62e18df| No fix yet |
| Linux Linux |
|
|
In the Linux kernel, the following vulnerability has been resolved:
net/mlx5e: kTLS, Fix incorrect page refcounting
The kTLS tx handling code is using a mix of get_page() and page_ref_inc() APIs to increment the page reference. But on the release path (mlx5e_ktls_tx_handle_resync_dump_comp()), only put_page() is used.
This is an issue when using pages from large folios: the get_page() references are stored on the folio page while the page_ref_inc() references are stored directly in the given page. On release the folio page will be dereferenced too many times.
This was found while doing kTLS testing with sendfile() + ZC when the served file was read from NFS on a kernel with NFS large folios support (commit 49b29a573da8 ("nfs: add support for large folios")).
From the CVE record · Linux
The net/mlx5e kTLS transmit code can over-release a folio page because it increments and releases page references inconsistently. The issue was found during sendfile() and zero-copy testing with a file read from NFS using large folios; the input does not establish what an attacker can do or whether network access, authentication, or user interaction is required.
Engineers running Linux kernels that use net/mlx5e kTLS with large-folio pages should check their systems; affected versions are not specified.