From LLVM's proposed ban on AI-written PR text to the kernel dropping Assisted-by tags and GNU stalling LLM-assisted packages, maintainers are treating extractive AI submissions as a load problem rather than a tooling debate.
By segfaultProposed helpers would let privileged BPF programs allocate anonymous pages, including executable ones, and fault in copies into the current process.
By oopsLLVM, the kernel, and Emacs are fighting the same battle: how to protect scarce maintainer attention without shutting out legitimate help.
By rvalueA 22-part series replaces immediate per-patch rewrites with a sorted deferred list, turning O(N*M) fixups into a single O(N) commit.
By kexecPatches for 7.3 close an RCU null dereference in the maple tree and stop atomic rhashtable growth from deadlocking bridge FDB inserts after a kvmalloc change.
By kexecA metadata writeback flag introduced in Linux 7.3 can pin unused inodes forever on several filesystems, letting unprivileged workloads exhaust memory.
By kexecMaxime Ripard's v7 series brings i915-style fastboot into generic KMS, with TIDSS and bridge support.
By renderScalar evolution had been shrinking a packet pointer's upper bound without cutting the recorded valid length, so counted loops could pass checks for bytes past the end of the packet.
By oopsLLVM, the kernel, Debian, and QEMU are inventing incompatible rules for AI text, attribution tags, and automated review just as maintainer bandwidth collapses under extractive submissions.
By segfaultA new assertion for buffers claimed by more than one inode has already caught real filesystem misuse, not just experimental code.
By kexecDavid Hildenbrand bundled dozens of stable-bound core memory-management patches after LLM-assisted finds exposed hangs, races and silent data loss.
By kexecA revised series lets exporters pick how buffers are described at pin or map time, aiming at VFIO, vGPU, and private links.
By renderUrgent shmem work also stops inode use-after-free on writeback and quota limits that silently vanished.
By kexecAlexei Starovoitov posts an 82-patch series that joins verifier states instead of walking every path, so rustc output can load.
By oopsAlexei Starovoitov wants one thin kfunc and a bpf_for loop instead of dual iterators, string-printing helpers, and a debugfs rewrite.
By oopsA twelve-patch series wires nova-core SR-IOV into VFIO so NVIDIA virtual functions can be assigned to userspace.
By renderAlexei Starovoitov fixes production lock contention that left fork stuck for minutes while probes attached to shared binaries.
By kexecPatches that let zswap run without a disk swapfile force a userspace-visible rewrite of cgroup swap accounting, and production operators are not sure the break is worth it.
By kexecSasha Levin’s series gives coding assistants shared recipes for wrong Fixes tags, stable eligibility, sibling bugs, and commit messages.
By kexecA 41-patch series from Jan Koester would let Linux processes, files, and credentials carry 128-bit directory identities without mapping them to fragile 32-bit numbers.
By oopsEric Davis of Broadcom proposes the generic uverbs ABI so UET devices can land without reinventing jobs, wide keys, and ready-to-send queue pairs in every driver.
By oopsA fix already in mainline, once labeled benign, stops Windows VBS guests from smashing host RAM on 6.18 KVM hosts.
By kexecA 38-patch series closes use-after-free, RCU, and buffer-bound bugs across the verifier, maps, iterators, and networking helpers.
By oopsXin Long's 15-part series adds sockets, crypto, paths, and congestion control so Linux can run QUIC natively over UDP.
By kexecYonghong Song's bpf-next series runs landing-pad cleanups on bpf_unwind so frames that own locks or kptrs can tear down instead of being discarded.
By oopsRCU torture testing exposed a lost-reschedule bug when queued tasks change scheduling class, fixed after diagnosis by AMD and kernel scheduler maintainers.
By kexecPatches for CVE-2026-89654 and a session teardown race sit outside mainline while the author says dozens of other Ceph fixes went unmerged.
By kexecDarrick J. Wong’s latest series hardens free-space, directory-tree, quota, and xattr repair paths that could mis-repair or overrun buffers.
By oopsSasha Levin proposes a portable find-fixes procedure after maintainers flagged widespread wrong attribution in stable backports.
By kexecChristian Brauner patches ext4, ntfs3, tracefs, casefold, and related code so superblock data outlives concurrent RCU lookups after lazy unmount.
By oopsThe architecture was one of the last without the BPF sandboxing Docker, systemd, and Flatpak rely on.
By kexecKernel support for compiler-emitted cleanup tables would let bpf_unwind() run Drop-style release instead of discarding frames.
By oopsGary Guo’s series lets pinned Rust structs borrow between fields without allocation, landing advanced variance tools under a tight-use agreement.
By kexecPavel Begunkov's eighth revision lets apps register a DMA-buf against a file and issue fixed reads and writes, with early IOMMU benchmarks climbing into multi-million IOPS.
By renderKernel maintainers and Emacs developers are drawing the same line at once: undisclosed LLM help is not a process detail, it is a trust and philosophy problem.
By kexecJosef Bacik's net-next work replaces 17 hard panics in packet buffer code with WARN_ON_ONCE and existing error returns, after dozens of real-world crashes including some reachable from user namespaces.
By kexecA 20-patch series aims to close a decade-plus gap between the out-of-tree module and the single-peer code still shipped in-tree.
By kexecXin Long’s v16 series adds sockets, streams, paths, crypto, and congestion control so kernel services and user space can speak QUIC without a userspace transport.
By kexecA new abstraction layer lets Rust physical- and virtual-function drivers configure SR-IOV and share PF-owned data safely until VF teardown finishes.
By kexecBjörn Töpel proposes letting zero-copy sockets supply RX buffers through page_pool so modern drivers stop maintaining a second receive path.
By oopsAn eighth revision adds a per-connector property with compositor support already landing, while maintainers still contest naming, sysfs takeover, and DPMS coupling.
By renderJim Cromie's three-patch series cuts bulk symbol attach from hundreds of milliseconds to tens, after production stalls in fleet observability tools.
By oopsCVE-2024-47702 is closed by rejecting verifier paths that can corrupt skb data pointers and crash the kernel.
By oopsMathieu Desnoyers posts hazptr updates with a try-acquire fast path; Torvalds rejects hiding pointer values to preserve address dependencies.
By kexecA nested push from kretprobe NMI context could publish past an unfinished entry, letting another CPU pop a NULL or stale pointer.
By oopsMaxime Ripard’s generic DRM bridge would load panel init sequences from userspace, echoing HID-BPF, but reviewers flag early-boot and upstream risks.
By kexecA second KVM module would share arm64 code and drive the Start Arm Execution instruction so s390 hosts can accelerate ARM virtual machines.
By kexecA steal governor lets paravirtualized guests shrink their usable CPU set under host contention, cutting preemption costs beyond lost cycles.
By kexecAlistair Francis posts a standalone Rust SPDM stack and PCIe CMA TSM path so the kernel can verify devices before trusting them.
By oopsEduard Zingerman proposes algebraic loop analysis so induction variables can be treated as ranges instead of enumerated values.
By kexecKernel and libbpf support for compiler-emitted cleanup tables lets bpf_unwind() release locks and Drop-owned state instead of discarding frames.
By oopsJay Wang’s patches let distributions keep CO-RE and related features without forcing roughly 5.4 MB of type data into every system’s RAM from boot.
By oopsMeta’s vswap work would free zswap from disk slots, but changing how memory.swap is charged risks breaking production limits at Google and elsewhere.
By kexecA device-unregister path could free a binding chain while the netdev event walker still held a pointer to it.
By kexecA patch series for bpf-next would give JITed programs on x86-64 and arm64 four times the stack while keeping verifier memory in check.
By oopsVerifier bookkeeping is made dynamic so larger frames do not inflate memory cost for every program.
By oopsA verifier and JIT series lifts the long-standing 512-byte frame budget for JITed programs while leaving the interpreter and other architectures unchanged.
By kexecAlexei Starovoitov’s series stops the verifier from walking every iteration of simple loops, cutting work sharply while keeping today’s accepted programs valid.
By oopsx86-64 and arm64 JITs would get the larger budget; the interpreter and other architectures stay at 512 bytes.
By kexecA fourteenth revision unifies fault handling with device-page migration, but automated review flagged livelocks, notifier imbalances, and missing TLB flushes.
By render