$ the-wire · showcase
SMB client fixes close kernel-byte leak into F_GETLK
By RepoJournal · Filed · About Linux · Composed from the cited sources · methodology
The 7.4-rc1 fixes wave landed across vfs, smb and io_uring, and the SMB one is worth reading first: the client now validates server-supplied offsets and lengths before use.
Merge tag 'cifs-fixes-7.4-rc1' of https://git.manguebit.org/linux torvalds/linux
Paulo Alcantara's smb client pull validates offsets and lengths from server responses before the client uses them, closing a path where POSIX lock data leaked kernel bytes through F_GETLK, and fixing 32-bit pointer checks that wrapped on NT transaction responses and SMB1 EA lists. A malicious or broken server can no longer make the client read past the received data; mount against untrusted ser...
Merge tag 'io_uring-7.4-20261029' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux torvalds/linux
Jens Axboe's io_uring pull charges registered buffer arrays, rsrc nodes, tctx nodes and personalities to memcg and frees the XArray nodes a failed store leaves behind, and clears bpf_ctx while submitting SQEs from a bpf-ops program so a setsockopt no longer skips the capability check and lock_sock(). Workloads that lean on registered buffers see their memory accounted to the right cgroup instea...
Merge tag 'vfs-7.4-rc1.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs torvalds/linux
The vfs fix set updates dirtied_when rather than dirtied_time_when when a writeback switch puts an inode at the newest end of the target b_dirty list, which had left the list out of order so move_expired_inodes() could stop before older inodes and lazytime expiration was reset. A second fix clears the old wb's dirty-I/O state and bandwidth contribution.
Merge tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux torvalds/linux
Stephen Boyd's clk pull for the merge window is dominated in the diffstat by Qualcomm SoC clock drivers, but the core changes are what matter beyond arm64: clock nexus DT binding support in the framework plus a subsys driver macro in the driver core.
Keep Hyprland off rtkit realtime until its RT watchdog kills are fixed (#17103) basecamp/omarchy
Omarchy keeps Hyprland off rtkit realtime while the compositor's RT watchdog kills remain unfixed. Smaller companion changes: fbdev patches atmel_lcdfb against an empty modelist in probe and replaces the old-style con_blank with consw.con_un for vgacon, and Omarchy's background switch from still to video now fades once instead of showing an old frame for 100 to 400 ms.
