assigned CVE-2026-46333
Debian trixie 6.12.88-1
Debian bookworm 6.1.172-1

fragnesia-5db89c99566fc
This is a variant of our Fragnesia bug (CVE-2026-46300) that bypasses the merged fix (commit f84eca581739) by exploiting a separate path that remains unpatched in both mainline and the netdev net tree as of 2026-05-15 18:00 UTC.

The bug is in skb_segment() in net/core/skbuff.c. When building GSO segments from an skb that has a frag_list, the function propagates SKBFL_SHARED_FRAG only from the head skb. If a frag_list member carries page-cache-backed frags with the flag set but the head does not, the resulting segment skbs lose the marker. This lets them pass the skip_cow guard in esp_input() and get decrypted in place over page-cache pages, same primitive as the original Dirty Frag and Fragnesia exploits.

Triggering it requires three network namespaces connected by veth pairs. The sender does a normal send() followed by splice() on the same TCP connection. GRO on the forwarding hop coalesces the two into a single skb where the send() segment becomes the head (no flag) and the splice() segment goes into the frag_list (flag set). The forwarder has GSO disabled on its egress veth, so skb_segment() fires and strips the flag. The segments then reach an espintcp receiver that decrypts in place. The GRO coalescing step requires both segments to arrive in the same NAPI poll cycle, which is reliable with back-to-back sends but not fully deterministic, so the exploit retries on failure. The rest of the exploitation is identical to Fragnesia: AES-GCM keystream control gives a deterministic one-byte page-cache write per trigger, and the exploit iterates over a small ELF payload to overwrite a SUID binary.

We have reported this to the relevant parties. There is a pending patch (not currently accepted or merged) on the netdev list that would incidentally help prevent this by propagating the flag earlier in the GRO path, though it was not written to address this bug specifically, and no patch currently proposed fixes the root cause in skb_segment() itself.

https://github.com/v12-security/pocs/tree/main/fragnesia-5db89c99566fc

Pull documentation fixes from Jonathan Corbet:
"This is Willy Tarreau's new document clarifying the definition and
handling of security-related bugs, which we're trying to get out there
quickly on the theory that some of the bug reporters might actually
read and pay attention to it"

* tag 'docs-7.1-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/docs/linux:
docs: threat-model: don't limit root capabilities to CAP_SYS_ADMIN
docs: security-bugs: add a link to the threat-model documentation
Documentation: security-bugs: clarify requirements for AI-assisted reports
Documentation: security-bugs: explain what is and is not a security bug
Documentation: security-bugs: do not systematically Cc the security team

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=36d49bba19f2c19c933d13b25dcf4eb607a030b3

今日为 5 月 17 日，是国际不再恐惧同性恋、跨性别与双性恋日

the react2shell story and what happened next.js
https://sylvie.fyi/posts/react2shell/

只要你发展的够快
戒你的机构就来不及出现：

电子游戏发展得慢
所以有戒游戏机构

互联网发展得较慢
所以有戒网瘾中心

短视频出现得快
没听说成规模的戒短视频中心

AI发展速度更快
以至于戒Agent戒Vibe Coding的概念都没出现
https://fixupx.com/i/status/2056238088826020148

PL Nerd 迎来惊天大变，Vibe Coder 喜提 AI 原生
https://zhuanlan.zhihu.com/p/2039725076204016063

可能有订户觉得咱是 mbt 黑子（咱不否认），但是咱看完这篇文章之后只想给 mbt 郑重道歉。

我要洗眼睛！！！！

📰 Linux AF_ALG Crypto Code Removing Zero-Copy Support Out Of Security Concerns

Given all the recent Linux kernel security concerns and new bugs being discovered, the Linux cryptographic subsystem is proactively dropping zero-copy functionality from AF_ALG due to growing security concerns...

🔗 Source: https://www.phoronix.com/news/Linux-AF-ALF-Zero-Copy-Security

#kernel #linux

可爱捏

#前端 #供应链攻击 #npm #安全 #新动态
想必大家都看麻了x
早用 pnpm 早超生，俩周以内的依赖更新不装（（
Active Supply Chain Attack Compromises @antv Packages on npm：npm 上的 AntV 可视化库包被植入恶意代码，构成活跃的供应链攻击。

AI 摘要：2026 年 5 月，Socket Research Team 发现 @antv 系列 npm 包遭到供应链攻击，攻击者发布恶意版本，通过 postinstall 脚本窃取环境变量和 npm token，影响 @antv/g6、@antv/util 等多个核心包，建议用户立即版本锁定并轮换密钥。

[以下是方便搜索索引的大纲(AI 生成)，请读原文]

1. 攻击事件概览
• 2026 年 5 月 11–19 日期间，大量 @antv 子包集中发布恶意版本，时间戳显示攻击窗口集中。
• Socket 安全团队通过实时包监控（package monitoring）发现异常发布行为并发出警报。

2. 攻击手法与技术细节
• 恶意代码嵌入 postinstall 钩子，安装时自动执行，窃取系统环境变量、npm token 及敏感配置。
• 攻击者利用泄露的维护者凭据或通过社会工程学获取 npm 发布权限。

3. 受影响包范围
• 涉及 @antv/scale、@antv/attr、@antv/component、@antv/g6 等数十个包的多个版本。
• 所有在 2026-05-11 至 2026-05-19 时间窗口内发布的版本均为恶意版本。

4. 紧急缓解措施
• 锁定依赖版本，避免使用上述时间窗口内的版本。
• 运行 npm audit 或使用 Socket 工具扫描项目并移除受影响包。
• 立即轮换所有可能泄露的 npm token、环境变量及 API 密钥。
• 关注 AntV 官方后续安全公告，升级到修复后的版本。

1.18.0-6.1+deb11u6
1.22.1-9+deb12u7
1.26.3-3+deb13u5

haveged 1.9.21 fixes a local root exploit

Fix privilege escalation via command socket (CVE-2026-41054) - the uid check sent a NAK to non-root callers but did not exit the function, allowing unprivileged local users to send commands to the root-running daemon via the abstract UNIX socket.

Problem was found by Gemini Pro, operated by Dirk Mueller of SUSE.
https://bugzilla.suse.com/show_bug.cgi?id=1264086

https://www.openwall.com/lists/oss-security/2026/05/19/3

V12Sec: PinTheft is a Linux local privilege escalation exploit for an RDS zerocopy double-free that can be turned into a page-cache overwrite through io_uring fixed buffers.

The bug lived in the RDS zerocopy send path. rds_message_zcopy_from_user() pins user pages one at a time. If a later page faults, the error path drops the pages it already pinned, and later RDS message cleanup drops them again because the scatterlist entries and entry count remain live after the zcopy notifier is cleared. Each failed zerocopy send can steal one reference from the first page.

The PoC uses io_uring to make that refcount bug useful. It registers an anonymous page as a fixed buffer, giving the page a FOLL_PIN bias of 1024 references. It then steals those references with failing RDS zerocopy sends, frees the page, reclaims it as page cache for a SUID-root binary, and uses the stale io_uring fixed-buffer page pointer to overwrite that page cache with a small ELF payload. Executing the SUID binary drops into a root shell.

Sadly, the RDS kernel module this requires is only default on Arch Linux among the common distributions we tested.

https://github.com/v12-security/pocs/tree/main/pintheft
https://x.com/v12sec/status/2056730095658959061

bullseye's linux-6.1: 6.1.172-1~deb11u1

Google I/O 2026：反正 2026 年的科技公司发布会就是 AI 和模型发布会罢。

- Gemini 3.5 Flash 模型发布。Google 说它很强，Simon 说它很贵 [1]。有多贵？差不多是 Gemini 3.1 Pro 那么贵。
- 所谓全模态但现在是生成视频的 Gemini Omni 模型；YouTube Shorts/YouTube Create app 上可以免费用。
- WebMCP 标准发布；开发者现可使用 chrome://flags 的 enable-webmcp-testing 启用该特性，并使用 Inspector 插件基于网页提供的 MCP 与 LLM 进行对话和操作（需自带 Gemini API key）。
- Chrome DevTools for agents 发布。
- 以及大家都知道的玩意儿：Google AI 搜索、AI 图片暗水印 C2PA、
- 喔对了，Gemini CLI 会在 6/18 停止服务；Google 建议换用 Antigravity CLI。 [2]

blog.google/~

1. https://simonwillison.net/2026/May/19/gemini-35-flash/
2. developers.googleblog.com/~

#Google