CVE-2024-27435

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> nvme: fix reconnection fail due to reserved tag allocation<br /> <br /> We found a issue on production environment while using NVMe over RDMA,<br /> admin_q reconnect failed forever while remote target and network is ok.<br /> After dig into it, we found it may caused by a ABBA deadlock due to tag<br /> allocation. In my case, the tag was hold by a keep alive request<br /> waiting inside admin_q, as we quiesced admin_q while reset ctrl, so the<br /> request maked as idle and will not process before reset success. As<br /> fabric_q shares tagset with admin_q, while reconnect remote target, we<br /> need a tag for connect command, but the only one reserved tag was held<br /> by keep alive command which waiting inside admin_q. As a result, we<br /> failed to reconnect admin_q forever. In order to fix this issue, I<br /> think we should keep two reserved tags for admin queue.