在 Linux 内核中,以下漏洞已得到修复: RDMA/erdma:当 destroy 命令失败时,完成对象的彻底清理 当 失败时, 、 、 和 会提前返回,导致队列缓冲区、MTT(内存转换表)、门铃(doorbells)以及 STAG、QPN、CQN 和 AHN 等标识符发生资源泄漏。 由于命令超时会清除 ,并永久禁用命令队列,因此重试无法成功;在销毁器(destructor)失败后,RDMA 核心层仍会保留该对象,而后续的 uverbs 强制清理机制会将相关指针置空,使得这些资源变得不可达(即无法被回收或访问)
| Vendor | Product | Version Range | Status |
|---|---|---|---|
| Linux | Linux | 155055771704f8cbb5c176a4309b7dc30a50450c< 5fcfc988ecf3e2bbe308b95c2c0d2f011d9afabe |
affected |
155055771704f8cbb5c176a4309b7dc30a50450c< ce7d205c264665517c25e8a3231cf2d0c2443e3c |
affected | ||
155055771704f8cbb5c176a4309b7dc30a50450c< 334145d683ad3e31d052247f10807f2ebc950351 |
affected | ||
155055771704f8cbb5c176a4309b7dc30a50450c< 652befcba956ef357f480525ccbe25c59bc81d4d |
affected | ||
6.0 |
affected | ||
< 6.0 |
unaffected | ||
6.12.110≤ 6.12.* |
unaffected | ||
6.18.52≤ 6.18.* |
unaffected | ||
| … +2 more rows | |||
Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.
| # | POC Description | Source Link | Shenlong Link |
|---|
No public POC found.
Login to generate AI POC| CVE-2026-92489 | 9.8 CRITICAL | xfrm: Fix skb double-free in xfrm_dev_direct_output() |
| CVE-2026-90235 | 9.8 CRITICAL | sunrpc: xprtsock: annotate shared socket callbacks with READ_ONCE/WRITE_ONCE |
| CVE-2026-90173 | 9.8 CRITICAL | smb: smbdirect: free completion queues with ib_free_cq() |
| CVE-2026-90104 | 9.8 CRITICAL | NFSv4.1: zero referring call lists before decoding |
| CVE-2026-90151 | 9.8 CRITICAL | NFSv4: remove callback IDR entry on client allocation failure |
| CVE-2026-90110 | 9.4 CRITICAL | inetpeer: randomize RB-tree node comparison using SipHash |
| CVE-2026-90230 | 9.1 CRITICAL | nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate() |
| CVE-2026-90414 | 9.1 CRITICAL | IB/isert: reject PDUs declaring more data than was received |
| CVE-2026-90413 | 9.1 CRITICAL | IB/isert: reject login PDUs declaring more data than was received |
| CVE-2026-90367 | 8.8 HIGH | wifi: mt76: mt7996: hold dev->mt76.mutex while disabling tx worker in SER |
| CVE-2026-90329 | 8.8 HIGH | HID: synchronize input before cleaning up a failed probe |
| CVE-2026-93189 | 8.8 HIGH | HID: core: quiesce input in hid_hw_stop() to prevent use-after-free |
| CVE-2026-90240 | 8.8 HIGH | iommu/vt-d: Flush context cache with correct SID when tearing down aliases |
| CVE-2026-90357 | 8.8 HIGH | wifi: mt76: mt7915: unlink TWT flow if the MCU rejects the agreement |
| CVE-2026-90286 | 8.8 HIGH | drm/amdgpu/gfx6: Use PFP on the compute queues too |
| CVE-2026-93042 | 8.8 HIGH | dmaengine: dw-edma: Terminate all descriptors without callbacks |
| CVE-2026-90256 | 8.8 HIGH | Bluetooth: L2CAP: use proto_lock for l2cap_data to fix l2cap_disconn_ind |
| CVE-2026-90380 | 8.8 HIGH | wifi: mt76: mt792x: fix use-after-free in mt76_rx_poll_complete |
| CVE-2026-90381 | 8.8 HIGH | wifi: mt76: fix handling channel context with different bands in mt76_switch_vif_chanctx() |
| CVE-2026-90425 | 8.8 HIGH | iommu/tegra241-cmdqv: Require exactly one Stream ID for a vSID |
Showing top 20 of 600 CVEs. View all on vendor page → →
No comments yet