MGLRU rewrite targets the working-set numbers clouds actually use
Kairui Song's frequency-guided promotion claims large reclaim gains and a stable active-memory reading, but Andrew Morton is waiting on more review.
By oopsKairui Song's frequency-guided promotion claims large reclaim gains and a stable active-memory reading, but Andrew Morton is waiting on more review.
By oopsAn unprivileged user can map the kernel image through a perf filter that has existed since Linux 6.3; Steven Rostedt says the permission check belongs in perf, not tracing.
By kexecSBI Supervisor Software Events give perf a path into interrupt-disabled code, with kexec treated as unsafe until firmware state can be cleared.
By kexecUwe Kleine-König wants driver_override opt-in only, after mismatched binds crash drivers that never expected them; VFIO and PCI stubs would have to declare themselves safe.
By kexecA virtual machine can crash the host with UDP tunnel segmentation headers the kernel does not validate, a backport request says.
By kexecLeon Romanovsky's proposal lets boot policy and userspace authorize hardware before drivers attach, without yet tying that decision to IOMMU domains.
By oopsrbrd serializes copies that brd allows to tear, yet its first numbers were mixed and key abstractions were already in flight.
By kexecA Clang attribute now requires every kobj_type to name a release method, exposing leaks and use-after-frees in DLM, UBIFS, AMDGPU, Xen, and others.
By renderChristian König says an unsignaled fence must not complete just because a Rust object went out of scope; the Rust DRM authors say that warning no longer means what it used to.
By renderFailed execve and affinity changes could hand threads a bogus ID, lock up small machines, or panic the kernel.
By kexecChristian Marangi's proposal treats coding sublayers like clocks, so MACs can pick them up from device tree or ACPI even if the PCS driver arrives or leaves later.
By kexecA proposed kernel fix closes a duplicate free during hash-set resize, but a maintainer wants a real name before taking it.
By kexecHe says structured comments next to kernel interfaces would pollute the source, and that any contract belongs outside the tree for tools rather than humans.
By kexecMathieu Desnoyers posted fixes for 33 SCHED_DEADLINE accounting errors, three of which take the machine down.
By kexecEmacs, LLVM, and the kernel are arguing in public whether AI volume justifies bans, tag removals, and pull-request quotas before maintainer capacity gives out.
By renderAlexei Starovoitov proposes growing tables for maps, subprograms and kfuncs, and faster checks so safe programs are no longer turned away for size.
By kexecThe change would let Android userdata survive a move from 4KB to 16KB pages without rewriting the on-disk format.
By oopsA 16-patch series from Xiaomi and Tencent makes Multi-Gen LRU split anon and file scans by refault rates, closing a gap that produced weak swappiness response and premature out-of-memory kills.
By kexecJarkko Sakkinen's series replaces the kernel's only user of assoc_array with a hash table, and fixes lifetime bugs in s390's certificate keyring.
By kexecLoop, NVMe, the SCSI target and similar paths would refuse pseudo filesystem files after a capability bypass let unprivileged tasks trigger kernel actions.
By kexecAndrew Morton queued a change that makes the main Linux list a fallback after duplicated traffic, including AI-generated patches, made the volume unsustainable.
By oopsIan Rogers' series lets users override event encodings and metrics from directories, without rebuilding perf or rescanning every sysfs event.
By oopsIan Rogers replaces the untyped table with a type-safe API after years of NULL-check mistakes and strict-aliasing violations.
By oopsChristian Marangi's phylink changes would let coding-sublayer drivers probe and disappear independently of the MAC.
By oopsA crafted GSO packet of stacked Network Service Headers could overflow the kernel stack when the nsh module was loaded.
By oopsSasha Levin wants allmodconfig to catch userspace breakage that currently slips past bots, but Andrew Morton says module-loading tests still have a separate problem.
By kexecA build-time scan would tag more than 12,000 error-returning functions so fmod_ret programs can skip them, with the authors warning the rule is not fully safe.
By kexecCarlos López proposes skipping secondary-MMU zaps when host permissions only widen, cutting snapshot-time refaults on x86 and s390.
By kexecDavid Hildenbrand's urgent series is already in the mm tree; networking paths that can still multiply page references remain open.
By kexecA peer who knew a session id could collect signatures made with another client's key, and related session errors could stall the connection.
By oopsSasha Levin and Ted Ts'o want coding assistants to report, not guess, while Andrew Morton asks that the same docs force a plain account of who a bug hurts.
By oopsCrafted images can panic the kernel, hang reads of huge files, or expose restricted files to unprivileged users.
By oopsA sixth revision lets physical-function drivers share typed data with virtual functions and roll back a failed enable.
By renderA 29-patch series from Kumar Kartikeya Dwivedi would let bpf_throw() unwind frames that still hold kernel resources, which the verifier currently forbids.
By kexecBroadcom's Zack Rusin wants per-VM backdoor emulation, userspace hypercall exits, and nested routing so unmodified VMware guests, including under Windows VBS, can run on KVM.
By kexecKrzysztof Kozlowski says he may ignore further submissions after an LG G8x series repeated work already on the linux-media list and failed basic style checks.
By kexecChristian Brauner's RFC would stop procfs and sysfs files being used as storage after a capability-free loop setup killed a process.
By kexecShuai Xue's frame-pointer work would let RISC-V apply kernel fixes without a reboot, and it now includes fixes for traces that have been wrong since Linux 6.8.
By oopsPrograms could wait on a queue they build themselves, and spawn kernel threads for reclaim and network work they scale and place.
By oopsA Meta rework of a Linus Torvalds idea would let many CPUs copy the same cached file pages without bouncing folio reference counts.
By oopsNine null-deref, out-of-bounds, and unvalidated-geometry flaws in dirty-log replay turn a crafted volume into a kernel crash at mount time.
By kexecMathieu Desnoyers’s series stops over-admission, lockups, and three kernel crashes in the deadline scheduler, and ships the first dedicated selftests.
By oopsCVE-2026-89788 scored 9.8 and reached later longterm kernels, but 5.15 still lacks the September 2026 tree-connection fix.
By oopsCVE-2026-31705, a CVSS 9.8 out-of-bounds write in SMB extended-attribute handling, was backported to six later stables but omitted from still-maintained 5.15.
By oopsGoogle engineer Ackerley Tng’s 88-patch RFC sketches how confidential guests could back private memory with tmpfs, custom providers, and device BARs.
By oopsUnvalidated outer headers from a guest or tap writer could drive GSO past the start of the skb and crash the host.
By kexecA patch series generalizes PCI's untrusted flag into cross-bus DISABLED, ADVERSARY, and FULL modes that userspace can set before any driver binds.
By kexecA Google-authored series would add a per-namespace capability ceiling and a host-wide sysctl so sandboxes need not expose CAP_NET_ADMIN and friends by default.
By kexecJosef Bacik's 90-patch series seals mounts, network, credentials, and helpers so a container root starts with nothing and host-side escapes are cut off.
By oopsEight accounting bugs could crash Linux or leave deadline scheduling unable to admit tasks or offline CPUs.
By oopsA proposed patch rejects oversized cfg80211 key sequence counters that could overrun a fixed buffer in the Realtek SDIO driver.
By kexecA critical zswap path could return old pages instead of written zeroes, while batched swap reads hid zram and async stalls from pressure and delay accounting.
By kexecUnprivileged-reachable overlayfs bugs found by AI review merge for 7.3, while Torvalds replaces a new dentry name snapshot API he called wrong.
By oopsA production hang lasting minutes on large nohz_full systems prompted kernel fixes so isolated CPUs no longer block LRU drains and cpuset memory moves.
By kexecSasha Levin’s second-round series folds gcov and KCOV together and adds source-based coverage that safety standards require.
By oopsIPv6 extension headers, IPv4 fragments, and ICMP error quotes could bypass hardware encryption intent on the transmit path.
By oopsGoogle's Daeho Jeong posted a 20-patch series so existing 4KB F2FS images can mount on kernels whose page size is larger than the filesystem block size.
By kexecUnprivileged network namespaces could force integer wrap in reassembly accounting and free a live socket.
By oopsLinus Torvalds fixed ordering so the reported distance matches real history instead of lying timestamps.
By segfaultChristian Brauner's series aims to stop local users from stalling network teardown by faulting while a removable /proc entry is pinned.
By render