nvmet-rdma: handle inline data with a nonzero offset
Linux
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
Linux
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
≥ 0d5ee2b2ab4f6776c361bc975c2323bc8b5cf349 · < 2944113ad5fbcdf5d349d857c03d2a44b6de75b8| No fix yet |
| Linux Linux |
|
|
In the Linux kernel, the following vulnerability has been resolved:
nvmet-rdma: handle inline data with a nonzero offset
nvmet_rdma_use_inline_sg() maps the host-controlled inline data offset into the per-command inline scatterlist. The bounds check admits any offset with off + len <= inline_data_size, but the mapping still assumes the data begins in the first inline page:
sg->offset = off; sg->length = min_t(int, len, PAGE_SIZE - off);
When a port is configured with inline_data_size > PAGE_SIZE (settable up to max(SZ_16K, PAGE_SIZE)), an offset in (PAGE_SIZE, inline_data_size] makes "PAGE_SIZE - off" underflow, so sg->length is set to ~4 GiB and the block backend reads far past the first inline page. num_pages(len) also ignores the offset, so an in-bounds offset whose [off, off+len) span crosses a page boundary under-counts the scatterlist.
Map the offset properly: split it into a page index and an in-page offset, start the scatterlist at that page, and size the page count from page_off + len. Because the request scatterlist may now start at inline_sg[page_idx] rather than inline_sg[0], generalize the inline-SGL identity test in nvmet_rdma_release_rsp() to a range test; otherwise the persistent inline scatterlist is mistaken for an allocated one and nvmet_req_free_sgls() frees an inline page (and warns in free_large_kmalloc()).
From the CVE record · Linux
The Linux kernel’s nvmet-rdma code mishandles host-controlled inline data offsets when a port’s inline_data_size exceeds PAGE_SIZE. A host sending an offset within the configured limit can cause the block backend to read past an inline page; an in-bounds span crossing a page boundary can also be mapped incorrectly. The input does not specify authentication requirements.
Engineers running Linux kernel NVMe over RDMA targets should check their kernel version and inline data configuration; affected versions are not specified.