{"api_version":"1","generated_at":"2026-11-20T17:03:27+00:00","cve":"CVE-2026-99401","urls":{"html":"https://cve.report/CVE-2026-99401","api":"https://cve.report/api/cve/CVE-2026-99401.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-99401","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-99401"},"summary":{"title":"rtnetlink: wait for RCU readers in rtnl_link_unregister()","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nrtnetlink: wait for RCU readers in rtnl_link_unregister()\n\nrtnl_link_ops_get() finds ops on link_ops under rcu_read_lock() and only\nthen takes ops->srcu. rtnl_link_unregister() unlinks ops, then calls\nsynchronize_srcu() and cleanup_srcu_struct() without first waiting for\nRCU readers. A reader that found ops before list_del_rcu() and has not yet\ntaken ops->srcu is not waited for.\n\nSuch a reader then locks a cleaned-up srcu_struct and keeps using ops\nafter the module that owns it is gone. rtnl_newlink() and\nrtnl_dump_ifinfo() both look up ops without RTNL, the latter for any\nuser that dumps links filtered by kind.\n\nWith delays added between the lookup and srcu_read_lock() in\nrtnl_link_ops_get(), and before rtnl_link_ops_put() in\nrtnl_dump_ifinfo(), running \"ip link show type dummy\" alongside\n\"rmmod dummy\" on a KASAN kernel gives:\n\n  BUG: KASAN: vmalloc-out-of-bounds in __srcu_read_unlock+0x57/0x70\n  Read of size 8 at addr ffffffffc02ff318 by task ip/147\n  Call Trace:\n   __srcu_read_unlock+0x57/0x70\n   rtnl_dump_ifinfo.cold+0x30/0x3b\n   rtnl_dumpit+0x7a/0x1b0\n   netlink_dump+0x486/0x11b0\n   __netlink_dump_start+0x661/0x910\n   rtnetlink_rcv_msg+0x611/0x9e0\n   netlink_rcv_skb+0x126/0x380\n   netlink_unicast+0x605/0x830\n   netlink_sendmsg+0x6d0/0xbb0\n   __sys_sendto+0x335/0x3a0\n\nCall synchronize_rcu() before synchronize_srcu(), as\nrtnl_af_unregister() already does. This should be enough because\nrtnl_link_ops_get() is the only RCU walker of link_ops, and it takes\nops->srcu before rcu_read_unlock(), so every reader that found ops is\ncovered by synchronize_srcu() once the RCU grace period ends. The extra\ngrace period only costs time at module unload.\n\nReaders and the writer were serialized by RTNL until link_ops got its\nown mutex, which is why the missing RCU wait did not matter before.","state":"PUBLISHED","assigner":"Linux","published_at":"2026-11-19 15:17:52","updated_at":"2026-11-19 15:17:52"},"problem_types":[],"metrics":[],"references":[{"url":"https://git.kernel.org/stable/c/3b70edea393de445255da6c706eebb823117a955","name":"https://git.kernel.org/stable/c/3b70edea393de445255da6c706eebb823117a955","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/ee636421740bc47e9d9e612ca7c9fa7dc5d17be1","name":"https://git.kernel.org/stable/c/ee636421740bc47e9d9e612ca7c9fa7dc5d17be1","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/9e3a6b707681a6291a410d1b7ce15840a9912a86","name":"https://git.kernel.org/stable/c/9e3a6b707681a6291a410d1b7ce15840a9912a86","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-99401","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-99401","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 6b57ff21a3109b1dba2d286ff415463e6fb1fca3 9e3a6b707681a6291a410d1b7ce15840a9912a86 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 6b57ff21a3109b1dba2d286ff415463e6fb1fca3 ee636421740bc47e9d9e612ca7c9fa7dc5d17be1 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 6b57ff21a3109b1dba2d286ff415463e6fb1fca3 3b70edea393de445255da6c706eebb823117a955 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 6.13","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 6.13 semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 6.18.61 6.18.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.3.2 7.3.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.4-rc1 * original_commit_for_fix","platforms":[]}],"timeline":[],"solutions":[],"workarounds":[],"exploits":[],"credits":[],"nvd_cpes":[],"vendor_comments":[],"enrichments":{"kev":null,"epss":null,"legacy_qids":[]},"source_records":{"cve_program":{"containers":{"cna":{"affected":[{"defaultStatus":"unaffected","product":"Linux","programFiles":["net/core/rtnetlink.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"lessThan":"9e3a6b707681a6291a410d1b7ce15840a9912a86","status":"affected","version":"6b57ff21a3109b1dba2d286ff415463e6fb1fca3","versionType":"git"},{"lessThan":"ee636421740bc47e9d9e612ca7c9fa7dc5d17be1","status":"affected","version":"6b57ff21a3109b1dba2d286ff415463e6fb1fca3","versionType":"git"},{"lessThan":"3b70edea393de445255da6c706eebb823117a955","status":"affected","version":"6b57ff21a3109b1dba2d286ff415463e6fb1fca3","versionType":"git"}]},{"defaultStatus":"affected","product":"Linux","programFiles":["net/core/rtnetlink.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"status":"affected","version":"6.13"},{"lessThan":"6.13","status":"unaffected","version":"0","versionType":"semver"},{"lessThanOrEqual":"6.18.*","status":"unaffected","version":"6.18.61","versionType":"semver"},{"lessThanOrEqual":"7.3.*","status":"unaffected","version":"7.3.2","versionType":"semver"},{"lessThanOrEqual":"*","status":"unaffected","version":"7.4-rc1","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"cpeMatch":[{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"6.18.61","versionStartIncluding":"6.13","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.3.2","versionStartIncluding":"6.13","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.4-rc1","versionStartIncluding":"6.13","vulnerable":true}],"negate":false,"operator":"OR"}]}],"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nrtnetlink: wait for RCU readers in rtnl_link_unregister()\n\nrtnl_link_ops_get() finds ops on link_ops under rcu_read_lock() and only\nthen takes ops->srcu. rtnl_link_unregister() unlinks ops, then calls\nsynchronize_srcu() and cleanup_srcu_struct() without first waiting for\nRCU readers. A reader that found ops before list_del_rcu() and has not yet\ntaken ops->srcu is not waited for.\n\nSuch a reader then locks a cleaned-up srcu_struct and keeps using ops\nafter the module that owns it is gone. rtnl_newlink() and\nrtnl_dump_ifinfo() both look up ops without RTNL, the latter for any\nuser that dumps links filtered by kind.\n\nWith delays added between the lookup and srcu_read_lock() in\nrtnl_link_ops_get(), and before rtnl_link_ops_put() in\nrtnl_dump_ifinfo(), running \"ip link show type dummy\" alongside\n\"rmmod dummy\" on a KASAN kernel gives:\n\n  BUG: KASAN: vmalloc-out-of-bounds in __srcu_read_unlock+0x57/0x70\n  Read of size 8 at addr ffffffffc02ff318 by task ip/147\n  Call Trace:\n   __srcu_read_unlock+0x57/0x70\n   rtnl_dump_ifinfo.cold+0x30/0x3b\n   rtnl_dumpit+0x7a/0x1b0\n   netlink_dump+0x486/0x11b0\n   __netlink_dump_start+0x661/0x910\n   rtnetlink_rcv_msg+0x611/0x9e0\n   netlink_rcv_skb+0x126/0x380\n   netlink_unicast+0x605/0x830\n   netlink_sendmsg+0x6d0/0xbb0\n   __sys_sendto+0x335/0x3a0\n\nCall synchronize_rcu() before synchronize_srcu(), as\nrtnl_af_unregister() already does. This should be enough because\nrtnl_link_ops_get() is the only RCU walker of link_ops, and it takes\nops->srcu before rcu_read_unlock(), so every reader that found ops is\ncovered by synchronize_srcu() once the RCU grace period ends. The extra\ngrace period only costs time at module unload.\n\nReaders and the writer were serialized by RTNL until link_ops got its\nown mutex, which is why the missing RCU wait did not matter before."}],"providerMetadata":{"dateUpdated":"2026-11-19T14:39:48.535Z","orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux"},"references":[{"url":"https://git.kernel.org/stable/c/9e3a6b707681a6291a410d1b7ce15840a9912a86"},{"url":"https://git.kernel.org/stable/c/ee636421740bc47e9d9e612ca7c9fa7dc5d17be1"},{"url":"https://git.kernel.org/stable/c/3b70edea393de445255da6c706eebb823117a955"}],"title":"rtnetlink: wait for RCU readers in rtnl_link_unregister()","x_generator":{"engine":"bippy-1.2.0"}}},"cveMetadata":{"assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","assignerShortName":"Linux","cveId":"CVE-2026-99401","datePublished":"2026-11-19T14:23:38.430Z","dateReserved":"2026-09-25T10:25:14.412Z","dateUpdated":"2026-11-19T14:39:48.535Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-11-19 15:17:52","lastModifiedDate":"2026-11-19 15:17:52","problem_types":[],"metrics":[],"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"99401","Ordinal":"1","Title":"rtnetlink: wait for RCU readers in rtnl_link_unregister()","CVE":"CVE-2026-99401","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"99401","Ordinal":"1","NoteData":"In the Linux kernel, the following vulnerability has been resolved:\n\nrtnetlink: wait for RCU readers in rtnl_link_unregister()\n\nrtnl_link_ops_get() finds ops on link_ops under rcu_read_lock() and only\nthen takes ops->srcu. rtnl_link_unregister() unlinks ops, then calls\nsynchronize_srcu() and cleanup_srcu_struct() without first waiting for\nRCU readers. A reader that found ops before list_del_rcu() and has not yet\ntaken ops->srcu is not waited for.\n\nSuch a reader then locks a cleaned-up srcu_struct and keeps using ops\nafter the module that owns it is gone. rtnl_newlink() and\nrtnl_dump_ifinfo() both look up ops without RTNL, the latter for any\nuser that dumps links filtered by kind.\n\nWith delays added between the lookup and srcu_read_lock() in\nrtnl_link_ops_get(), and before rtnl_link_ops_put() in\nrtnl_dump_ifinfo(), running \"ip link show type dummy\" alongside\n\"rmmod dummy\" on a KASAN kernel gives:\n\n  BUG: KASAN: vmalloc-out-of-bounds in __srcu_read_unlock+0x57/0x70\n  Read of size 8 at addr ffffffffc02ff318 by task ip/147\n  Call Trace:\n   __srcu_read_unlock+0x57/0x70\n   rtnl_dump_ifinfo.cold+0x30/0x3b\n   rtnl_dumpit+0x7a/0x1b0\n   netlink_dump+0x486/0x11b0\n   __netlink_dump_start+0x661/0x910\n   rtnetlink_rcv_msg+0x611/0x9e0\n   netlink_rcv_skb+0x126/0x380\n   netlink_unicast+0x605/0x830\n   netlink_sendmsg+0x6d0/0xbb0\n   __sys_sendto+0x335/0x3a0\n\nCall synchronize_rcu() before synchronize_srcu(), as\nrtnl_af_unregister() already does. This should be enough because\nrtnl_link_ops_get() is the only RCU walker of link_ops, and it takes\nops->srcu before rcu_read_unlock(), so every reader that found ops is\ncovered by synchronize_srcu() once the RCU grace period ends. The extra\ngrace period only costs time at module unload.\n\nReaders and the writer were serialized by RTNL until link_ops got its\nown mutex, which is why the missing RCU wait did not matter before.","Type":"Description","Title":"rtnetlink: wait for RCU readers in rtnl_link_unregister()"}]}}}