Bilingual

PostgreSQL and the OOM Killer: Why You Must Use Strict Memory Overcommit
18 days ago
- Strict memory overcommit is used to protect PostgreSQL databases from catastrophic OOM kills by converting destructive failures into graceful errors.
- PostgreSQL's architecture means an OOM kill can corrupt shared memory, leading to connection drops, transaction aborts, and lengthy crash recovery.
- A kernel bug (introduced in Linux 6.5) caused phantom committed memory inflation, leading to unexpected ENOMEM errors under strict overcommit.
- The bug was fixed in Linux 6.8 by correcting a condition in move_vma() that reversed error handling for memory accounting.
- Ubicloud's heuristic for setting the commit limit is: 80% of physical memory (excluding hugepages) plus a fixed 2 GB buffer for sidecar processes.
- Monitoring memory usage and understanding workload characteristics is essential before enabling strict overcommit to avoid frequent ENOMEM errors.
Linux Looking to Retire a Number of Old ARM Platforms in Early 2027
18 days ago
- Linux plans to retire several old ARM platforms and features by early 2027.
- Deprecations include unconverted board files (except OMAP1 and S3C), iWMMXt FPU support, ARMv6/ARM1136r0, and Cortex-M3/M4/M7 microcontroller support.
- Other items for removal: LSI Axxia platform, OABI-related features, and DEPRECATED_PARAM_STRUCT.
- Big endian ARMv7 support is marked as 'BROKEN' but not yet scheduled for removal.
- The proposal, led by ARM maintainer Arnd Bergmann, follows a prior failed effort and targets Linux 7.4 after the LTS release.
Linux DRM Scheduler Patches Yield Improvement for Job Submission Latency
17 days ago
- Linux DRM scheduler patches replace workqueues with kthread_work for asynchronous task execution.
- The change significantly reduces job submission latency when the system is overloaded with runnable CPU processes.
- It helps prevent compositors from missing frames and benefits high-priority DRM clients.
- Currently adapted for Intel Xe and Panthor Arm Mali drivers, with AMDGPU not expected to see major improvements.
- RFC patches are available on the dri-devel list for further review.
C++ Details of Asymmetric Fences
18 days ago
- Asymmetric thread fences optimize concurrency by shifting heavy synchronization costs from a common path to an infrequent slow path, using lightweight compiler barriers on the fast path and stronger fences on the slow path.
- The lightweight fence (asymmetric_thread_fence_light) is implemented as a compiler barrier (e.g., asm volatile("" : : : "memory")), preventing compiler reordering but not hardware reordering.
- The heavy fence (asymmetric_thread_fence_heavy) on Linux uses the membarrier() system call, which issues inter-processor interrupts (IPIs) to enforce memory ordering across cores, involving smp_mb() barriers before and after IPI sending.
- The C++ standard formalizes asymmetric fences (P1202R5) by integrating them into the memory model, with rules affecting the single total order S for memory_order_seq_cst operations and requiring careful handling to avoid transitivity issues.
- Implementation details include fast-path optimizations in membarrier() for single-core or single-mm scenarios, architecture-specific smp_mb() implementations, and considerations for data races and happens-before relationships.
We won $92,337 bug bounty using a single kernel 0-day
14 days ago
- GhostLock (CVE-2026-43499) is a Linux kernel vulnerability found in all major distributions since 2011.
- It allows unprivileged local attackers to get a dangling kernel pointer, write to arbitrary addresses, and hijack function tables for privilege escalation.
- The bug exists in Linux versions 2.6.39 to 7.1 and was fixed in April 2026.
- Exploitation involves creating a PI futex deadlock to trigger a use-after-free on kernel stack memory.
- Attackers can reclaim freed stack memory via PR_SET_MM_MAP and forge fake rt_mutex_waiter structures.
- The exploit overwrites inet6_protos[IPPROTO_UDP] to hijack control flow and gain root access via DirtyMode.
- Mitigations include patching the kernel, enabling RANDOMIZE_KSTACK_OFFSET, and using STATIC_USERMODE_HELPER.
Preempt_none Is Dead; Your Postgres Probably Doesn't Care
15 days ago
- AWS benchmark showed PostgreSQL throughput dropping to 0.51x on Linux 7.0 vs 6.x due to a regression from removing PREEMPT_NONE preemption mode.
- The performance issue is caused by minor page faults inside a spinlock (buffer_strategy_lock) when using 4KB pages with large shared memory (100+ GB), exacerbated by PREEMPT_LAZY.
- Enabling huge pages (huge_pages=on) eliminates the regression by reducing page faults and TLB pressure, making the kernel upgrade safe for most users.
- Self-hosted servers should configure huge pages properly; containerized setups need host-level huge page reservations or reduced shared_buffers.
- Cloud database users rely on vendors to handle kernel and huge page configurations, with no action required on their part.
- The kernel community suggested using rseq for scheduler hints, but this is not a practical immediate fix; PostgreSQL may adopt it in future versions.
- The benchmark artifact highlights the importance of correct configuration (huge pages) for large shared memory workloads, not a fundamental Linux issue.
Realtek RTL8723BS WiFi Linux Driver Hardened Against Malicious WiFi Access Point
8 days ago
- Realtek RTL8723BS WiFi Linux driver has been hardened against malicious WiFi access points.
- Fixes address out-of-bounds memory access conditions when connecting to 'bad' WiFi hosts due to improper validation of over-the-air frames.
- Patches will be merged for Linux 7.2-rc3 and back-ported to stable kernel versions, fixing security issues present since 2017.
- The driver has seen ongoing cleaning efforts since its staging admission, known for its code complexity.
- The RTL8723BS is an 802.11 b/g/n SDIO WiFi chipset with Bluetooth 4.0 support, used in many devices.
Raspberry Pi 5 Iommu Driver Being Worked on for the Mainline Linux Kernel
8 days ago
- Raspberry Pi 5's IOMMU driver is being adapted for the mainline Linux kernel, improving support over two years after release.
- Open-source developer Daniel Drake is leading the effort, adapting the downstream driver with enhancements like generic page table management.
- The driver integrates with the display controller for efficient graphics memory management, using 4KB pages, though 12KB is wasted with the default 16KB kernel PAGE_SIZE.
- Code review is underway, with the goal of upstream inclusion to benefit Raspberry Pi 5 users running the mainline kernel.
New Linux Patches Aim to Better Handle Multiple Swap Devices
7 days ago
- Linux patches from Tencent aim to improve handling of multiple swap devices, especially for swap tiering and layered swap.
- The new design removes the global plist and lock, introduces a percpu reader priority queue, and addresses several long-standing design issues.
- Key improvements include fair distribution of I/O to swap devices, better scalability, and support for future features like swap tiering.
- The patch series eliminates the global percpu cluster cache, which previously conflicted with concepts like swap tiering and caused performance issues.
- It also resolves problems with device rotation, contention on global locks, and unclear rotation rules, leading to a more flexible and robust system.
Linus Torvalds Reaffirms That Linux Is Not "Anti-AI"
6 days ago
- Linus Torvalds reaffirms Linux is not 'anti-AI'; AI/LLMs are tools for developers.
- Linux kernel development doesn't mandate AI use, but opposes arguments against others using it.
- Torvalds stresses AI is useful, despite issues like maintainer workload or finding bugs.
- He emphasizes the kernel project is about technology, not being a 'social warrior' project.
- Decisions are based on technical merit, not fear of new tools like AI.
Support for atomic block writes in Linux 6.13 (2025)
6 days ago
- Linux 6.13 introduces atomic block writes to prevent data tearing in databases.
- Atomic writes require hardware support (NVMe/SCSI) and software stack updates (VFS, filesystem, block layer).
- Databases use double writes for consistency, which atomic writes can replace to improve performance.
- Current implementation supports single-block atomic writes via O_DIRECT and pwritev2() with RWF_ATOMIC flag.
- Filesystem block size must align with device constraints; ext4 limits block size to page size, while XFS supports larger blocks.
- statx() system call queries atomic-write capabilities, including min/max unit sizes and support flags.
- Future work includes multi-block atomic write support and broader filesystem integration.
'Rust makes coding fun again': Why Linux is moving away from C
6 days ago
- Linux stable kernel maintainer Greg Kroah-Hartman announced that Linux and other projects like Git are moving toward Rust.
- Rust is considered safer than C and helps eliminate common vulnerabilities like unchecked pointers and forgotten unlocks.
- Kroah-Hartman estimates that 80% of kernel CVEs would be caught by Rust, allowing maintainers to focus on logic bugs instead.
- Rust is becoming the default for new work in key subsystems, with some drivers only accepted in Rust.
- The Linux kernel will not be rewritten in Rust, but new code in subsystems like graphics will be Rust-only.
- Rust has influenced improvements in C code by introducing safety mechanisms like guards and scoped allocation.
- Kroah-Hartman emphasized that Rust makes coding fun by reducing worries about memory safety and locking issues.
- The Rust experiment in the Linux kernel is now considered official and no longer an experiment.
- Kroah-Hartman highlighted that Rust benefits both maintainers by easing review workload and users by enhancing security.
Linus Torvalds to critics of AI coding in Linux: "Fork it. Or just walk away."
4 days ago
- Linus Torvalds strongly supports integrating AI tools like Sashiko into the Linux kernel development, rejecting anti-AI absolutists.
- Sashiko, an AI-powered code review system, can independently find over 50% of bugs but also generates false positives estimated within a 20% range.
- Torvalds advocates for individual choice in using LLM tools, opposing demands to ban AI-generated code and emphasizing the open-source option to fork the project if needed.
Linux kernel will support $ORIGIN, sort of
11 hours ago
- The goal is to support relocatable binaries in Nix by allowing $ORIGIN in PT_INTERP and shebang.
- Christian Brauner suggested using eBPF for programmable interpreter selection via binfmt_misc, rather than a kernel patch.
- An eBPF program can match ELF files and dynamically determine the interpreter path, supporting $ORIGIN.
- A new binfmt_misc flag 'L' enables loader substitution, making the interpreter override transparent and preserving process identity.
- Future plans include a NixOS module that registers the BPF handler at boot, gated on a new PT_INTERP_NIX segment for opt-in.

Hasty Briefsbeta

#linux kernel

PostgreSQL and the OOM Killer: Why You Must Use Strict Memory Overcommit

Linux Looking to Retire a Number of Old ARM Platforms in Early 2027

Linux DRM Scheduler Patches Yield Improvement for Job Submission Latency

C++ Details of Asymmetric Fences

We won $92,337 bug bounty using a single kernel 0-day

Preempt_none Is Dead; Your Postgres Probably Doesn't Care

Realtek RTL8723BS WiFi Linux Driver Hardened Against Malicious WiFi Access Point

Raspberry Pi 5 Iommu Driver Being Worked on for the Mainline Linux Kernel

New Linux Patches Aim to Better Handle Multiple Swap Devices

Linus Torvalds Reaffirms That Linux Is Not "Anti-AI"

Support for atomic block writes in Linux 6.13 (2025)

'Rust makes coding fun again': Why Linux is moving away from C

Linus Torvalds to critics of AI coding in Linux: "Fork it. Or just walk away."

Linux kernel will support $ORIGIN, sort of