affected products in the CVE record: Linux · Linuxthe product type of Linux kernel in nday's product registry
現在の段階
DIS
公開日時
CVSS 基本値
7.0 HIGH v3.1
# ライフサイクル
横軸は UTC。点=分単位、幅広の帯=日付のみ、白抜き=不明または推定。
日時不明
DIS公開
PAT修正
記録なし
ANA解析
記録なし
POCPoC
記録なし
ITW悪用
記録なし
KEVKEV
記録なし
# 概要
In the Linux kernel, the following vulnerability has been resolved:
RDMA/rxe: insert mcg into mcg_tree only after rxe_mcast_add() succeeds
rxe_get_mcg() publishes a newly allocated multicast group in
rxe->mcg_tree before programming the backing Ethernet multicast address
with rxe_mcast_add(), which runs outside mcg_lock. A local userspace
RDMA client reaches this path with ATTACH_MCAST on a UD QP; if
rxe_mcast_add() then returns an error (for example -ENODEV when the
backing netdev has been removed, or a propagated dev_mc_add() error),
the unwind frees the published group without removing it from the tree.
A later lookup of the same MGID dereferences the freed struct rxe_mcg
from __rxe_lookup_mcg().
Fix this by keeping the new mcg private until rxe_mcast_add() succeeds.
Split the tree publication into __rxe_publish_mcg(), call rxe_mcast_add()
before taking the tree reference, and free the still-private mcg on
failure. Because the group is never visible in mcg_tree until the
multicast address is programmed, no concurrent caller can look it up or
attach a QP to a group that is about to be torn down, so the error path
needs no conditional unwind. If another caller publishes the same MGID
while the address is being programmed, the post-add re-check under
mcg_lock finds the winner; this caller then drops its private object and
balances its own rxe_mcast_add() with rxe_mcast_del() before returning
the winner.
Reproduced by forcing the rxe_mcast_add() error return under KASAN:
without the change the next attach to the same MGID reports a
slab-use-after-free in __rxe_lookup_mcg(); with it the forced failure
returns cleanly. A no-injection attach/detach regression, including a
two-QP shared join/leave and re-attach, stays KASAN- and leak-clean.
Original text of the CVE record (cvelist), shown as published.
# Exploit conditions
What an attack needs, read from the CVSS vector.
Attack vector AV:L
Local: needs a local or SSH session, or a user tricked into acting on the machine
Attack complexity AC:H
High: success depends on conditions outside the attacker's control, such as gathering details of the target or winning a race
Privileges required PR:L
Low: ordinary user privileges
User interaction UI:N
None: no action by anyone other than the attacker
CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H
CVSS 3.1 vector from the CVE record, scored by the CNA Linux.
Your account is suspended: you cannot comment, reply, edit or report. You can still delete your own comments.
formatting
↵↵
empty line: new paragraph
↵
line break: kept as is
*…*
emphasis
**…**
bold
`…`
inline code
```
on its own line: starts and ends a code block
> …
quote
- …
list
1. …
numbered list
[…](https://…)
link (http and https only)
https://…
bare address: link
CVE-2026-1234
links to that vulnerability on nday
\*
shows the character itself
Headings, images, tables, nested lists and HTML are shown as plain text.
Links from new accounts are shown as plain text; CVE IDs are always linked.
reported
# キーボードショートカット
j / k
次の行 / 前の行(一覧の行またはイベント)
Enter
カーソル位置の脆弱性を開く
/
検索(Esc で入力欄を抜ける)
y
CVE 番号をコピー
o
イベントの出典を開く(脆弱性ページ)
u
戻る
?
このヘルプを表示 / 非表示
sign inaccountmodadminAdminProfileFollowingSecurityDeveloperSign outSigned in.Sign-in method linked to your account.Sign-in was cancelled at the provider; nothing changed.That login already belongs to another nday account, or this account already has one from the same provider. Nothing was linked.Sign-in did not complete. Try again.You are still signed in as before: that login belongs to another nday account, or to none yet. Sign out first to use it.Linking was not completed because your last sign-in is no longer recent. Confirm it's you on the account page, then link again.This account is suspended, so new sign-in methods cannot be linked.Dismissworking…Delete this comment? This cannot be undone.Hidden by a moderator: {reason}text (moderators only)earlier versions: {n}replaced at {time}Posting needs the community rules accepted (version {version}).The community rules changed (version {version}); read and accept them to continue.Reported. A moderator will look at it.You have already reported this comment.You are signed out. Sign in again to continue.The request was refused as cross-site. Reload the page and try again.The session token changed. Reload the page and try again.Your account is suspended.Choose a handle on your account page first.Accept the current community rules first.Only moderators can do this.Only the author can do this.You cannot report your own comment.This comment changed in the meantime. Reload the page to see it as it is now.This comment or vulnerability does not exist (any more).Not a valid CVE ID.Only a visible top-level comment can get replies.This thread has reached its reply limit.You posted the same text here a moment ago.A comment needs 1 to {max} characters.Comments cannot be posted from this site language.Unknown purpose, or a purpose on a reply. Reload the page and try again.The reason is missing or too long.The note is too long.Unknown resolution.This position in the list is no longer valid. Reload the page.The request was malformed. Reload the page and try again.The text is too large.Too many requests. Try again in {seconds} s.Too many requests. Try again later.Daily limit reached. Try again after {time}.The server is busy. Try again in a moment.This address does not take comments. Open the page on the main site.Server error. Try again later.Network error: nday could not be reached. Check the connection and try again.The community rules changed while you were reading. Read them again.Write something first.Too long: {n} characters, at most {max}.Tick the box to accept the community rules first.Something went wrong ({code}). Try again.No comments yet.The discussion could not be loaded right now.more commentsfirst commentsmore repliesreplies[deleted account]link to this commenteditedposted on the {language} siteHidden by a moderator.Deleted by its author.technical analysisreproduction reportfix reportdetection methodquestionPlease include the affected version, the environment (OS, configuration), what you did and what happened, and your source.Please include the fixed version or patch, the environment, how you checked the fix and what happened, and your source.残り {n} 件を表示X で見る投稿を読み込めませんでした。リストの研究者が X でこの CVE に触れた投稿は、まだ収集していません。続きを読む閉じるsending…Sent. An editor will look at it; you can follow its status here.titlesummaryaffected products or versionsfix informationtimeline eventsource linksotherwaiting for an editoracceptednot acceptedsent {at}Editors' note: {note}Sign in first.Your account is suspended.Choose a public handle on your account page first.Accept the current community rules on your account page first.The page's session expired; try again.The request was refused; reload the page and try again.Choose what is wrong.Write what is wrong, within the length limit.The evidence link must be a full http:// or https:// address.You already have several reports on this vulnerability waiting for an editor; wait until they are handled.This vulnerability already has many reports waiting for an editor; try again later.You reached the limit of reports for 24 hours; try again later.Too many requests; wait a moment and try again.This vulnerability was not found.The report is too large.Network error; try again.Write what is wrong first.Too long: at most {max} characters.Something went wrong; try again later.