The central voice for Linux and Open Source security news.
Secure Boot sits at the point where firmware and operating system trust intersect, and it decides what code is allowed to start the machine. Most systems treat it like background plumbing, but it has a direct influence on Linux security best practices because it defines whether the kernel you think you are running is actually the one that loads. When it works as intended, it gives you a predictable baseline for the rest of the stack. When it doesn't, the failure usually shows up in places that are hard to diagnose and even harder to monitor.
Out-of-bounds reads aren't flashy, but they sit close to the root of a lot of quiet trouble in Linux security. The bug shows up when software pulls data past a buffer's edge and exposes pieces of memory it never meant to share. Most of the time, the leak feels small. Sometimes it hands over the kind of detail an attacker can fold into an ASLR bypass used to execute malicious code or a later privilege move.
Suspicious emails rarely confess in the body. The clues live in headers, MIME parts, and tiny inconsistencies between what a message claims and what it actually delivers. If your team can read those signals quickly''and connect them to the attachment''you'll cut off credential theft, loaders, and ransomware without slowing operations.
It's always been a matter of responding to cybersecurity. Threats happen, defenses are made, attackers adjust their plans, and the cycle starts all over again. But what if we could make that different? What if AI could detect attack patterns before they happen? This would give defenders a head start instead of continually having to catch up.
Update to 2.9.7
Fix seeking and looping of media elements that set the loop property. Fix several crashes and rendering issues. Fix CVE-2025-13947, CVE-2025-43458, CVE-2025-66287
Update to 2.9.7
Loading a manipulated TGA file in krita, an image manipulation program, could result in a heap-based buffer overflow in KisTgaImport.
Update to 143.0.7499.40 * High CVE-2025-13630: Type Confusion in V8 * High CVE-2025-13631: Inappropriate implementation in Google Updater * High CVE-2025-13632: Inappropriate implementation in DevTools * High CVE-2025-13633: Use after free in Digital Credentials
Fix CVE-2025-12744