GHSA-7RPW-2P75-RWVF
Vulnerability from github – Published: 2026-08-15 06:32 – Updated: 2026-08-17 06:33In the Linux kernel, the following vulnerability has been resolved:
xprtrdma: Decouple req recycling from RPC completion
rl_kref formerly served two distinct lifetimes through a single refcount: it gated when a Reply could wake its RPC task, and it gated when an rpcrdma_req could return to its free pool. The marshal path took the Send-side reference only when SGEs needed DMA-unmap (sc_unmap_count > 0), which made a Send carrying only pre-registered buffers an exception: the Reply handler dropped rl_kref from 1 to 0 and freed the req while the HCA might still be DMA-reading from its send buffer.
Give rl_kref a narrower job. The RPC layer takes one reference when slot allocation hands a req out. rpcrdma_prepare_send_sges() takes a Send-side reference unconditionally after WR preparation succeeds. xprt_rdma_free_slot() and xprt_rdma_bc_free_rqst() drop the RPC-layer reference; rpcrdma_sendctx_unmap() drops the Send-side reference. The req returns to its free pool only after both owners have signed off.
The existing kref_init(&req->rl_kref) call in rpcrdma_prepare_send_sges() is removed. Initialization moves to the slot-allocation paths (xprt_rdma_alloc_slot and rpcrdma_bc_rqst_get), and the release callback re-arms rl_kref before the req returns to a free pool. A re-init in the marshal path would discard the RPC-layer reference that already exists on entry.
Three invariants follow:
-
Any rpcrdma_req held by an rpc_rqst has rl_kref >= 1. xprt_rdma_alloc_slot(), rpcrdma_bc_rqst_get(), and the backlog-wake branch in xprt_rdma_alloc_slot() each kref_init rl_kref before publishing the req. Without this invariant, an RPC task that aborts between slot allocation and marshal (gss_refresh failure or signal during call_connect, for example) would drive xprt_release() -> xprt_rdma_free_slot() -> kref_put against a refcount of zero, saturating refcount_t and stranding the slot.
-
The Send-side reference is taken only after WR prep succeeds. A mapping failure in rpcrdma_prepare_send_sges() runs rpcrdma_sendctx_cancel(), which DMA-unmaps the sendctx and clears sc_req without touching rl_kref. The sendctx ring walks in rpcrdma_sendctx_put_locked() and rpcrdma_sendctxs_destroy() skip entries with sc_req == NULL, so a burst of -EIO marshal failures cannot hold reqs off rb_send_bufs.
-
The release callback re-arms rl_kref so the next consumer enters with the invariant satisfied.
Replies now complete the RPC directly. rpcrdma_reply_handler() calls rpcrdma_complete_rqst() in place of kref_put on the non-LocalInv branch. The LocalInv branch already completes the RPC from frwr_unmap_async() and is unaffected.
Because Send-side references can now outlive RPC completion, connection teardown drains sendctx entries whose unsignaled Sends never had a later signaled completion to walk the ring. rpcrdma_sendctxs_destroy() walks the active range and runs rpcrdma_sendctx_unmap() on each entry with a non-NULL sc_req before the request buffers are reset, and is moved ahead of rpcrdma_reqs_reset() in rpcrdma_xprt_disconnect() so the reqs are still in their pre-reset state when the Send-side refs are released.
The drain creates a teardown-ordering hazard on the backchannel path. With the new lifetime, releasing a bc_prealloc req from rpcrdma_req_release() re-adds it to bc_pa_list. The disconnect in xprt_rdma_destroy() runs after xprt_destroy_backchannel() has already emptied bc_pa_list, so the drained reqs would otherwise leak. xprt_rdma_destroy() now runs xprt_rdma_bc_destroy(xprt, 0) a second time after the disconnect to reclaim them.
{
"affected": [],
"aliases": [
"CVE-2026-72473"
],
"database_specific": {
"cwe_ids": [],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-08-15T06:22:21Z",
"severity": "CRITICAL"
},
"details": "In the Linux kernel, the following vulnerability has been resolved:\n\nxprtrdma: Decouple req recycling from RPC completion\n\nrl_kref formerly served two distinct lifetimes through a single\nrefcount: it gated when a Reply could wake its RPC task, and it\ngated when an rpcrdma_req could return to its free pool. The\nmarshal path took the Send-side reference only when SGEs needed\nDMA-unmap (sc_unmap_count \u003e 0), which made a Send carrying only\npre-registered buffers an exception: the Reply handler dropped\nrl_kref from 1 to 0 and freed the req while the HCA might still\nbe DMA-reading from its send buffer.\n\nGive rl_kref a narrower job. The RPC layer takes one reference\nwhen slot allocation hands a req out. rpcrdma_prepare_send_sges()\ntakes a Send-side reference unconditionally after WR preparation\nsucceeds. xprt_rdma_free_slot() and xprt_rdma_bc_free_rqst() drop\nthe RPC-layer reference; rpcrdma_sendctx_unmap() drops the\nSend-side reference. The req returns to its free pool only after\nboth owners have signed off.\n\nThe existing kref_init(\u0026req-\u003erl_kref) call in\nrpcrdma_prepare_send_sges() is removed. Initialization moves to\nthe slot-allocation paths (xprt_rdma_alloc_slot and\nrpcrdma_bc_rqst_get), and the release callback re-arms rl_kref\nbefore the req returns to a free pool. A re-init in the marshal\npath would discard the RPC-layer reference that already exists\non entry.\n\nThree invariants follow:\n\n - Any rpcrdma_req held by an rpc_rqst has rl_kref \u003e= 1.\n xprt_rdma_alloc_slot(), rpcrdma_bc_rqst_get(), and the\n backlog-wake branch in xprt_rdma_alloc_slot() each kref_init\n rl_kref before publishing the req. Without this invariant,\n an RPC task that aborts between slot allocation and marshal\n (gss_refresh failure or signal during call_connect, for\n example) would drive xprt_release() -\u003e\n xprt_rdma_free_slot() -\u003e kref_put against a refcount of\n zero, saturating refcount_t and stranding the slot.\n\n - The Send-side reference is taken only after WR prep\n succeeds. A mapping failure in rpcrdma_prepare_send_sges()\n runs rpcrdma_sendctx_cancel(), which DMA-unmaps the sendctx\n and clears sc_req without touching rl_kref. The sendctx\n ring walks in rpcrdma_sendctx_put_locked() and\n rpcrdma_sendctxs_destroy() skip entries with sc_req == NULL,\n so a burst of -EIO marshal failures cannot hold reqs off\n rb_send_bufs.\n\n - The release callback re-arms rl_kref so the next consumer\n enters with the invariant satisfied.\n\nReplies now complete the RPC directly. rpcrdma_reply_handler()\ncalls rpcrdma_complete_rqst() in place of kref_put on the\nnon-LocalInv branch. The LocalInv branch already completes the\nRPC from frwr_unmap_async() and is unaffected.\n\nBecause Send-side references can now outlive RPC completion,\nconnection teardown drains sendctx entries whose unsignaled\nSends never had a later signaled completion to walk the ring.\nrpcrdma_sendctxs_destroy() walks the active range and runs\nrpcrdma_sendctx_unmap() on each entry with a non-NULL sc_req\nbefore the request buffers are reset, and is moved ahead of\nrpcrdma_reqs_reset() in rpcrdma_xprt_disconnect() so the reqs\nare still in their pre-reset state when the Send-side refs are\nreleased.\n\nThe drain creates a teardown-ordering hazard on the backchannel\npath. With the new lifetime, releasing a bc_prealloc req from\nrpcrdma_req_release() re-adds it to bc_pa_list. The disconnect\nin xprt_rdma_destroy() runs after xprt_destroy_backchannel() has\nalready emptied bc_pa_list, so the drained reqs would otherwise\nleak. xprt_rdma_destroy() now runs xprt_rdma_bc_destroy(xprt, 0)\na second time after the disconnect to reclaim them.",
"id": "GHSA-7rpw-2p75-rwvf",
"modified": "2026-08-17T06:33:32Z",
"published": "2026-08-15T06:32:25Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-72473"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/53442c7d0c888e51b8bc3da196970a669cc6b294"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/740975054a1970c0cf15f70ac39724a064f45847"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/8203f760a72bd39a3b66bc4eff0aa272a99fe22b"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/9f3d9b68c1c6c51746e5ecdb52b2e6a2901de37e"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/e7632089523acddcdd8f090ad19e96fb3107b04d"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/e786233d2e0bbff9a82e43f02ae3a46ab4b08ec3"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
]
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.