---
title: "CVE-2026-98070: CVE-2026-98070"
description: "In the Linux kernel, the following vulnerability has been resolved: net/rds: acquire RDS_IN_XMIT in rds_tcp_reset_callbacks() rds_tcp_reset_callbacks() quiesces the transmit path by setting the path state to RDS_CONN_RESETTING and then waiting for RDS_IN_XMIT to be sampled clear before swapping the underlying socket and calling rds_send_path_reset(). Sampling the bit clear is not the same as owning it: rds_send_xmit() can re-acquire RDS_IN_XMIT right after the wait_event() returns. Its state recheck after taking the lock is a store-buffering pattern (the resetter writes the state and reads the bit, the sender writes the bit and reads the state) and acquire_in_xmit() is only an acquire operation, so on weakly ordered architectures both sides can miss each other's write and the transmit path then runs concurrently with rds_send_path_reset() rewriting cp_xmit_* state - which is exactly what the comment above rds_send_path_reset() tells its callers to prevent. Take the lock instead, hold it across the socket swap and rds_send_path_reset(), and release it with a wake-up at the end. The lock-ordering constraint documented above the wait still holds: the lock is acquired before lock_sock(), so a sender inside tcp_sendmsg() can never be waited on while we hold the socket lock. Two details of the old code go away with the same change: - t_sock is now read only after the lock is acquired. The old code cached it before waiting; the teardown in rds_conn_shutdown() releases that socket and clears t_sock, so a pointer cached before the wait can be stale by the time the accept path resumes. Reading it under RDS_IN_XMIT is what makes the exclusion complete once the teardown owns the same lock, which the next patch arranges; until then the teardown still only samples the bit, and the two paths remain as exposed to each other as they are today. - The old !osock early path called rds_send_path_reset() with no serialization at all. It now runs under the lock like the normal path. The conditional RDS_CONN_RESETTING transition of the previous patch happens before the socket check either way: a path found without a socket is either still connecting (its reconnect worker blocked on t_conn_path_lock) and legitimately goes RESETTING -> UP on the new socket, or it has been torn down meanwhile and is dropped. The in-function comment describing the old wait-based quiesce is rewritten to describe the lock-based one, and the stale block comment above the function (which still described a return value and an incomplete list of t_sock writers) is refreshed to name all four writers - the connect, accept, teardown and swap paths - and what serializes each of them."
url: https://panop.io/threat-heartbeat/CVE-2026-98070/
---

# CVE-2026-98070 CVE-2026-98070

In the Linux kernel, the following vulnerability has been resolved: net/rds: acquire RDS\_IN\_XMIT in rds\_tcp\_reset\_callbacks() rds\_tcp\_reset\_callbacks() quiesces the transmit path by setting the path state to RDS\_CONN\_RESETTING and then waiting for RDS\_IN\_XMIT to be sampled clear before swapping the underlying socket and calling rds\_send\_path\_reset(). Sampling the bit clear is not the same as owning it: rds\_send\_xmit() can re-acquire RDS\_IN\_XMIT right after the wait\_event() returns. Its state recheck after taking the lock is a store-buffering pattern (the resetter writes the state and reads the bit, the sender writes the bit and reads the state) and acquire\_in\_xmit() is only an acquire operation, so on weakly ordered architectures both sides can miss each other's write and the transmit path then runs concurrently with rds\_send\_path\_reset() rewriting cp\_xmit\_\* state - which is exactly what the comment above rds\_send\_path\_reset() tells its callers to prevent. Take the lock instead, hold it across the socket swap and rds\_send\_path\_reset(), and release it with a wake-up at the end. The lock-ordering constraint documented above the wait still holds: the lock is acquired before lock\_sock(), so a sender inside tcp\_sendmsg() can never be waited on while we hold the socket lock. Two details of the old code go away with the same change: - t\_sock is now read only after the lock is acquired. The old code cached it before waiting; the teardown in rds\_conn\_shutdown() releases that socket and clears t\_sock, so a pointer cached before the wait can be stale by the time the accept path resumes. Reading it under RDS\_IN\_XMIT is what makes the exclusion complete once the teardown owns the same lock, which the next patch arranges; until then the teardown still only samples the bit, and the two paths remain as exposed to each other as they are today. - The old !osock early path called rds\_send\_path\_reset() with no serialization at all. It now runs under the lock like the normal path. The conditional RDS\_CONN\_RESETTING transition of the previous patch happens before the socket check either way: a path found without a socket is either still connecting (its reconnect worker blocked on t\_conn\_path\_lock) and legitimately goes RESETTING -> UP on the new socket, or it has been torn down meanwhile and is dropped. The in-function comment describing the old wait-based quiesce is rewritten to describe the lock-based one, and the stale block comment above the function (which still described a return value and an incomplete list of t\_sock writers) is refreshed to name all four writers - the connect, accept, teardown and swap paths - and what serializes each of them.

CVSS base score 8.1 /10

High

CISA KEV

Not listed

Public exploit

None known

EPSS

0.47% chance of exploitation in 30 days

[View the NVD record ↗](https://nvd.nist.gov/vuln/detail/CVE-2026-98070)

## Details

CVSS vector

CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H

Published

25 September 2026

Updated

15 October 2026

## References

- [https://nvd.nist.gov/vuln/detail/CVE-2026-98070](https://nvd.nist.gov/vuln/detail/CVE-2026-98070)
- [https://git.kernel.org/stable/c/670af4e4a4de5e6bb5daee3838485571a0caa5fa](https://git.kernel.org/stable/c/670af4e4a4de5e6bb5daee3838485571a0caa5fa)
- [https://git.kernel.org/stable/c/830a2e21b200c6b6fdccc63dd3f23932bf7a894b](https://git.kernel.org/stable/c/830a2e21b200c6b6fdccc63dd3f23932bf7a894b)
- [https://git.kernel.org/stable/c/a05790cb4ff03f797d76906990b148d83f63e7a3](https://git.kernel.org/stable/c/a05790cb4ff03f797d76906990b148d83f63e7a3)
- [https://git.kernel.org/stable/c/d184e6dd4b8f3c20c018595fd09795a66b46c8ef](https://git.kernel.org/stable/c/d184e6dd4b8f3c20c018595fd09795a66b46c8ef)
- [https://git.kernel.org/stable/c/02c5f9dc2efd823e061954d564ce00bacd1bebeb](https://git.kernel.org/stable/c/02c5f9dc2efd823e061954d564ce00bacd1bebeb)
- [https://git.kernel.org/stable/c/062d9e008c67289e8e1b221ecdd8f9d60566d012](https://git.kernel.org/stable/c/062d9e008c67289e8e1b221ecdd8f9d60566d012)
- [https://git.kernel.org/stable/c/8e4c3b7844c906c7097b4cfedd9dd1f48c9a6a92](https://git.kernel.org/stable/c/8e4c3b7844c906c7097b4cfedd9dd1f48c9a6a92)
- [https://git.kernel.org/stable/c/d625112564c3e980e02504270222b49b82690cee](https://git.kernel.org/stable/c/d625112564c3e980e02504270222b49b82690cee)

[← All CVEs](https://panop.io/threat-heartbeat/)

## What this record can't tell you

- Whether you run the affected software, and where
- Whether an attacker can actually reach and exploit it
- The misconfigurations that never get a CVE number
- What it would cost the business, ranked into what to fix first

Panop answers these inside your environment, for this CVE and every other finding.

[Book a demo](https://panop.io/demo)