In the Linux kernel, the following vulnerability has been resolved:
svcrdma: Fix pcl_for_each_segment for empty chunks
When a parsed chunk list contains a chunk whose ch_segcount is zero,
pcl_for_each_segment computes its inclusive upper bound as
&chunk->ch_segments[ch_segcount - 1]. ch_segcount is u32, so the
subtraction wraps to 0xFFFFFFFF and the bound lands far past the
ch_segments flex array. The loop body then walks unrelated memory at
sizeof(struct svc_rdma_segment) stride until it faults.
A zero-segcount chunk is reachable from the wire:
xdr_check_write_chunk() only rejects segcount values greater than
rc_maxpages, and pcl_alloc_write() links a freshly allocated chunk
onto rc_write_pcl/rc_reply_pcl before its segment-fill loop runs,
so a Write or Reply chunk advertising zero segments leaves
ch_segcount == 0 on the list. When the transport has negotiated
Send-With-Invalidate, svc_rdma_get_inv_rkey() iterates all four
PCLs with pcl_for_each_segment and dereferences segment->rs_handle
on each iteration, turning the underflow into an out-of-bounds read
and a general protection fault.
<pre>
xdr_check_write_list / xdr_check_reply_chunk
pcl_alloc_write()
chunk = pcl_alloc_chunk(...) /* ch_segcount = 0 */
list_add_tail(&chunk->ch_list, &pcl->cl_chunks)
/* fill loop iterates zero times for wire segcount 0 */
svc_rdma_get_inv_rkey()
pcl_for_each_chunk(rc_write_pcl)
pcl_for_each_segment(segment, chunk)
pos <= &ch_segments[0u - 1u] /* 0xFFFFFFFF */
segment->rs_handle /* OOB read -> GPF */
</pre>
Fix by switching the macro to a half-open upper bound that uses
ch_segcount directly. For ch_segcount == 0 the loop start equals the
loop end and the body is skipped; for ch_segcount > 0 the iteration
range is unchanged. All six existing call sites in
net/sunrpc/xprtrdma/svc_rdma_recvfrom.c and
net/sunrpc/xprtrdma/svc_rdma_rw.c remain correct under the new bound,
so no caller changes are needed.
CVSS Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H
CVSS Score: 9.1
AV:N - svcrdma is the nfsd RPC-over-RDMA server transport and parses Write/Reply chunks from remote peers over iWARP or routable RoCEv2; a remote NFS/RDMA client triggers the bug with a crafted transport header on an established RDMA connection.
AC:L - The attacker sets RPCRDMA_CMP_F_SND_W_INV_OK on connect and sends a Write or Reply chunk with segcount 0, which xdr_check_write_chunk accepts; Send-With-Invalidate is available on typical RDMA devices, so svc_rdma_get_inv_rkey hits the underflow deterministically.
PR:N - svc_rdma_xdr_decode_req and svc_rdma_get_inv_rkey run in svc_rdma_recvfrom before svc_process/svc_authenticate; RDMA CM accept and RPC-over-RDMA header parsing require no NFS credentials or capabilities.
UI:N - Exploitation requires only attacker-generated RDMA connect and RPC-over-RDMA messages to a listening NFS/RDMA service; no victim user action is needed.
S:U - The out-of-bounds read and resulting general protection fault occur in the host kernel NFS/RDMA server and do not cross a VM, IOMMU, or sandbox boundary.
C:H - With ch_segcount 0, pcl_for_each_segment underflows the u32 bound to 0xFFFFFFFF and walks adjacent kernel heap at sizeof(struct svc_rdma_segment) stride, an unbounded out-of-bounds read rather than a few-byte leak.
I:N - svc_rdma_get_inv_rkey only loads segment->rs_handle from the overrun pointer; the defect is a read-only walk with no out-of-bounds write, use-after-free, or control-flow hijack primitive.
A:H - The unbounded walk continues until it hits unmapped memory and raises a general protection fault, oopsing or panicking the kernel, and the attacker can repeat it on each request.
| Attack Vector |
Network |
Scope |
Unchanged |
| Attack Complexity |
Low |
Confidentiality Impact |
High |
| Privileges Required |
None |
Integrity Impact |
None |
| User Interaction |
None |
Availability Impact |
High |
AV:N - svcrdma is the nfsd RPC-over-RDMA server transport and parses Write/Reply chunks from remote peers over iWARP or routable RoCEv2; a remote NFS/RDMA client triggers the bug with a crafted transport header on an established RDMA connection.
AC:L - The attacker sets RPCRDMA_CMP_F_SND_W_INV_OK on connect and sends a Write or Reply chunk with segcount 0, which xdr_check_write_chunk accepts; Send-With-Invalidate is available on typical RDMA devices, so svc_rdma_get_inv_rkey hits the underflow deterministically.
PR:N - svc_rdma_xdr_decode_req and svc_rdma_get_inv_rkey run in svc_rdma_recvfrom before svc_process/svc_authenticate; RDMA CM accept and RPC-over-RDMA header parsing require no NFS credentials or capabilities.
UI:N - Exploitation requires only attacker-generated RDMA connect and RPC-over-RDMA messages to a listening NFS/RDMA service; no victim user action is needed.
S:U - The out-of-bounds read and resulting general protection fault occur in the host kernel NFS/RDMA server and do not cross a VM, IOMMU, or sandbox boundary.
C:H - With ch_segcount 0, pcl_for_each_segment underflows the u32 bound to 0xFFFFFFFF and walks adjacent kernel heap at sizeof(struct svc_rdma_segment) stride, an unbounded out-of-bounds read rather than a few-byte leak.
I:N - svc_rdma_get_inv_rkey only loads segment->rs_handle from the overrun pointer; the defect is a read-only walk with no out-of-bounds write, use-after-free, or control-flow hijack primitive.
A:H - The unbounded walk continues until it hits unmapped memory and raises a general protection fault, oopsing or panicking the kernel, and the attacker can repeat it on each request.
CVSS 3.1