)]}' { "log": [ { "commit": "6b5a2b7d9bc156e505f09e698d85d6a1547c1206", "tree": "591d9eef673ca38374127201025640f11f9fc6cf", "parents": [ "c071a4fbb0ec26fd5831c6633762e38720350718", "db956bcf8d681b5a01ebe04c79f6a7b29b9934f9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 17:50:34 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 17:50:34 2026 +0530" }, "message": "Merge tag \u0027trace-tools-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull RTLA tool updates from Steven Rostedt:\n\n - Fix discrepancy in --dump-tasks option\n\n Due to a mistake, rtla-timerlat-hist used the CLI syntax\n \"--dump-task\" instead of the documented \"--dump-tasks\". Change the\n option to match both documentation and the other timerlat tool,\n rtla-timerlat-top.\n\n - Extend coverage of runtime tests\n\n Cover both top and hist tools in all applicable test cases, add tests\n for a few uncovered options, and extend checks for some existing\n tests.\n\n - Add unit tests for actions\n\n rtla\u0027s actions feature is implemented in its source file and contains\n non-trivial parsing logic. Cover it with unit tests.\n\n - Stop record trace on interrupt\n\n Fix a bug where an interval exists after receiving a signal in which\n the main instance is stopped but the record instance is not, leading\n to discrepancies in reported results and sometimes rtla hanging.\n\n - Restore continue flag in actions_perform()\n\n Fix a bug where rtla always continues tracing after hitting a\n threshold even if the continue action was triggered just once, and\n add tests verifying that the flag is reset properly.\n\n - Migrate command line interface to libsubcmd\n\n Replace rtla\u0027s argument parsing using getopt_long() with libsubcmd,\n used by perf and objtool, to reuse existing code and auto-generate\n better help messages. Extensive unit tests are included to detect\n regressions.\n\n - Add -A/--aligned option to timerlat tools\n\n Add an option to align timerlat threads, based on the recently\n introduced TIMERLAT_ALIGN option of the timerlat tracer, together\n with unit tests and documentation.\n\n - Document tests in README\n\n Document how to run unit and runtime tests in rtla\u0027s README.txt,\n including the dependencies needed to run them.\n\n* tag \u0027trace-tools-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (26 commits)\n rtla: Document tests in README\n Documentation/rtla: Add -A/--aligned option\n rtla/tests: Add unit tests for -A/--aligned option\n rtla/timerlat: Add -A/--aligned CLI option\n rtla/tests: Add unit tests for CLI option callbacks\n rtla/tests: Add unit tests for _parse_args() functions\n rtla: Parse cmdline using libsubcmd\n tools subcmd: allow parsing distinct --opt and --no-opt\n tools subcmd: support optarg as separate argument\n rtla: Add libsubcmd dependency\n rtla/tests: Add runtime tests for restoring continue flag\n rtla/tests: Run runtime tests in temporary directory\n rtla/tests: Add unit test for restoring continue flag\n rtla/actions: Restore continue flag in actions_perform()\n rtla: Stop the record trace on interrupt\n rtla/tests: Add unit tests for actions module\n rtla/tests: Add runtime tests for -C/--cgroup\n rtla/tests: Add runtime test for -k and -u options\n rtla/tests: Add runtime test for -H/--house-keeping\n rtla/tests: Cover all hist options in runtime tests\n ...\n" }, { "commit": "c071a4fbb0ec26fd5831c6633762e38720350718", "tree": "fb90dd1cc3e368311e39d23f78c4ec2784475c67", "parents": [ "18ecdd4d0aabc428318bbc2ca6134483a90ea697", "9cb99c598643ba78638dfd668cf020544159cf70" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 17:38:19 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 17:38:19 2026 +0530" }, "message": "Merge tag \u0027trace-latency-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull tracing latency updates from Steven Rostedt:\n\n - Dump the stack to the buffer on timerlat uret threashold event\n\n Record the stack trace in the buffer for THREAD_URET as well as\n THREAD_CONTEXT when the threshold is hit. Otherwise, if the threshold\n was not hit at task wakeup, but was at task return, it will not\n produce a stack trace making it harder to debug.\n\n - Have osnoise trace prints print to all buffers\n\n The osnoise tracer is allowed to print to the main buffer. Add a\n osnoise_print() helper function and use trace_array_vprintk() to\n print osnoise output.\n\n* tag \u0027trace-latency-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n tracing/osnoise: Array printk init and cleanup\n tracing/osnoise: Dump stack on timerlat uret threshold event\n" }, { "commit": "18ecdd4d0aabc428318bbc2ca6134483a90ea697", "tree": "c1f5157dc1680f0eb54d9c9e6672e617589a8aba", "parents": [ "4f7e89065e41fd487e411d35b5caeac2a10af10e", "69efd863a78584b9416ed6be0e1e7349124b4a00" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 17:33:20 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 17:33:20 2026 +0530" }, "message": "Merge tag \u0027probes-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull probes updates from Masami Hiramatsu:\n\n - BTF support for dereferencing pointers\n\n Add syntax to the parsing of eprobes to typecast structure pointer\n trace event fields, enabling BTF-based dereferencing instead of\n relying on manual offsets.\n\n - Improvements and robustness enhancements\n\n - Use flexible array for entry fetch code.\n\n Store probe entry fetch instructions in the probe_entry_arg\n allocation via a flexible array member to simplify memory\n allocation and lifetime management.\n\n - Replace BUG_ON with lockdep_assert_held in uprobe_buffer functions\n\n Replace BUG_ON() calls with lockdep_assert_held() in uprobe buffer\n enable/disable paths to prevent kernel crashes and better verify\n lock ownership.\n\n - Ensure the uprobe buffer size is bigger than event size.\n\n Add a BUILD_BUG_ON() assertion to guarantee that the per-CPU\n uprobe working buffer size is always larger than the maximum probe\n event size.\n\n* tag \u0027probes-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n tracing/eprobes: Allow use of BTF names to dereference pointers\n tracing: Replace BUG_ON with lockdep_assert_held in uprobe_buffer functions\n tracing: Use flexible array for entry fetch code\n tracing/probes: Ensure the uprobe buffer size is bigger than event size\n" }, { "commit": "4f7e89065e41fd487e411d35b5caeac2a10af10e", "tree": "edb15b513de9f2c4bbd13926951adc47af43d8b4", "parents": [ "8b308f96484e37d92d2fc6b72b091f60496c000e", "4135542de6c42bc366a651d0b2d82ea72ac199fc" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 17:29:24 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 17:29:24 2026 +0530" }, "message": "Merge tag \u0027bootconfig-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace\n\nPull bootconfig updates from Masami Hiramatsu:\n\n - bootconfig: move xbc_snprint_cmdline() to lib/bootconfig.c\n\n Move the xbc_snprint_cmdline() function and its buffer from\n main.c to the shared lib/bootconfig.c parser library so it\n can be reused by userspace tools.\n\n - render kernel.* subtree as cmdline string with -C\n\n Add a new -C option to print the kernel.* subtree as a flat\n command-line string at build time, allowing early parameter\n injection without runtime parsing.\n\n* tag \u0027bootconfig-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:\n tools/bootconfig: render kernel.* subtree as cmdline string with -C\n bootconfig: move xbc_snprint_cmdline() to lib/bootconfig.c\n" }, { "commit": "8b308f96484e37d92d2fc6b72b091f60496c000e", "tree": "c3fc5f9c5dbe4814b41d43aad4d9556f9fa786e6", "parents": [ "42eb3a5ef6bc56192bf450c79a3f274e081f8131", "5f4974231fb3656c15a82faece9f2303b8c18a14" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 16:49:07 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 16:49:07 2026 +0530" }, "message": "Merge tag \u0027linux_kselftest-next-7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest\n\nPull kselftest updates from Shuah Khan:\n \"Several fixes and improvements to resctrl tests and a change to\n kselftest document to clarify the use of FORCE_TARGETS build variable\"\n\n* tag \u0027linux_kselftest-next-7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:\n kselftest: fix doc for ksft_test_result_report()\n selftests/resctrl: Reduce L2 impact on CAT test\n selftests/resctrl: Simplify perf usage in CAT test\n selftests/resctrl: Remove requirement on cache miss rate\n selftests/resctrl: Raise threshold at which MBM and PMU values are compared\n selftests/resctrl: Increase size of buffer used in MBM and MBA tests\n selftests/resctrl: Support multiple events associated with iMC\n selftests/resctrl: Prepare for parsing multiple events per iMC\n selftests/resctrl: Do not store iMC counter value in counter config structure\n selftests/resctrl: Reduce interference from L2 occupancy during cache occupancy test\n selftests/resctrl: Improve accuracy of cache occupancy test\n docs: kselftest: Document the FORCE_TARGETS build variable\n" }, { "commit": "42eb3a5ef6bc56192bf450c79a3f274e081f8131", "tree": "cd5e440cd913b4005909eafdd613acad5807783d", "parents": [ "b1cbabe84ca1381a004fb91ee1791a1a53bce44e", "29afed142d64e181749214072315c976f8510bd7" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 16:33:57 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 16:33:57 2026 +0530" }, "message": "Merge tag \u0027linux_kselftest-kunit-7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest\n\nPull kunit updates from Shuah Khan:\n \"Fixes to tool and kunit core and new features to both to support JUnit\n XML (primitive) and backtrace suppression API:\n - Core support for suppressing warning backtraces\n - Parse and print the reason tests are skipped\n - Add (primitive) support for outputting JUnit XML\n - Don\u0027t write to stdout when it should be disabled\n - Add backtrace suppression self-tests\n - Suppress intentional warning backtraces in scaling unit tests\n - Add documentation for warning backtrace suppression API\n - Fix spelling mistakes in comments and messages\n - gen_compile_commands: Ignore libgcc.a\n - qemu_configs: Add or1k / openrisc configuration\"\n\n* tag \u0027linux_kselftest-kunit-7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:\n kunit:tool: Don\u0027t write to stdout when it should be disabled\n kunit: tool: Add (primitive) support for outputting JUnit XML\n kunit: tool: Parse and print the reason tests are skipped\n kunit: Add documentation for warning backtrace suppression API\n drm: Suppress intentional warning backtraces in scaling unit tests\n kunit: Add backtrace suppression self-tests\n bug/kunit: Core support for suppressing warning backtraces\n kunit: Fix spelling mistakes in comments and messages\n kunit: qemu_configs: Add or1k / openrisc configuration\n gen_compile_commands: Ignore libgcc.a\n" }, { "commit": "b1cbabe84ca1381a004fb91ee1791a1a53bce44e", "tree": "827e3d0f4b101c77f4f659c1ea271fa0fb8f5dbf", "parents": [ "ba9c792c824fff732df85119011d399d9b6d9155", "55c99f2d901c783c49dddd98f7c05d98ad834ce9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 13:20:54 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 13:20:54 2026 +0530" }, "message": "Merge tag \u0027for-7.2/dm-changes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm\n\nPull device mapper updates from Mikulas Patocka:\n\n - small cleanups in dm-vdo, dm-raid, dm-cache, dm-zoned-metadata\n\n - rework of dm-ima\n\n - introduce dm-inlinecrypt\n\n - fix wrong return value in dm-ioctl\n\n - fix rcu stall when polling\n\n* tag \u0027for-7.2/dm-changes\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:\n dm-zoned-metadata: Use strscpy() to copy device name\n dm cache: make smq background work limit configurable\n dm-inlinecrypt: add support for hardware-wrapped keys\n dm: limit target bio polling to one shot\n dm-ioctl: report an error if a device has no table\n dm: add documentation for dm-inlinecrypt target\n dm-inlinecrypt: add target for inline block device encryption\n block: export blk-crypto symbols required by dm-inlinecrypt\n dm-ima: use active table\u0027s size if available\n dm-ima: Fail more gracefully in dm_ima_measure_on_*\n dm-ima: Handle race between rename and table swap\n dm-ima: Fix issues with dm_ima_measure_on_device_rename\n dm-ima: remove new_map from dm_ima_measure_on_device_clear\n dm-ima: Fix UAF errors and measuring incorrect context\n dm-ima: don\u0027t copy the active table to the inactive table\n dm-ima: Remove status_flags from dm_ima_measure_on_table_load()\n dm-ima: remove broken last_target_measured logic\n dm-ima: remove dm_ima_reset_data()\n dm-raid: only requeue bios when dm is suspending\n dm vdo: use get_random_u32() where appropriate\n" }, { "commit": "ba9c792c824fff732df85119011d399d9b6d9155", "tree": "a6de1eda726e4e9156366b6fd960b81d09297aa3", "parents": [ "9b40ba14edcdf70240af8114092a76f75f070774", "c7c76f9232bd34835d821f14abdc5fafc17bc938" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 13:02:47 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 13:02:47 2026 +0530" }, "message": "Merge tag \u0027for-7.2/block-20260615\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux\n\nPull block updates from Jens Axboe:\n\n - NVMe pull request via Keith:\n - Per-controller admin and IO timeout sysfs attributes, and\n letting the block layer set request timeouts (Maurizio,\n Maximilian)\n - Multipath passthrough iostats, and PCI P2PDMA enablement for\n multipath devices (Keith, Kiran)\n - A new diag sysfs attribute group exporting per-controller\n counters (retries, multipath failover, error counters, requeue\n and failure counts, reset and reconnect events) (Nilay)\n - FDP configuration validation and bounds check fixes (liuxixin)\n - Various nvmet fixes, including a pre-auth out-of-bounds read in\n the Discovery Get Log Page handler, auth payload bounds\n validation, and tcp error-path leak fixes (Bryam, Tianchu,\n Geliang)\n - nvme-tcp lockdep and workqueue fixes (Shin\u0027ichiro, Kuniyuki,\n Eric)\n - Assorted other fixes and cleanups (John, Yao, Chao, Mateusz,\n Achkinazi, Wentao)\n\n - MD pull request via Yu Kuai:\n - raid1/raid10 fixes for a deadlock in the read error recovery\n path, error-path detection and bio accounting with cloned bios,\n and an nr_pending leak in the REQ_ATOMIC bad-block error path\n (Abd-Alrhman)\n - PCI P2PDMA propagation from member devices to the RAID device\n (Kiran)\n - dm-raid bio requeue fix, and various smaller fixes and cleanups\n (Benjamin, Chen, Li, Thorsten)\n\n - Enable Clang lock context analysis for the block layer, with the\n accompanying annotations across queue limits, the blk_holder_ops\n callbacks, crypto, cgroup, iocost, kyber and mq-deadline (Bart)\n\n - Block status code infrastructure work: a tagged status table, a\n str_to_blk_op() helper, a bio_endio_status() helper, and on top of\n that a new configurable block-layer error injection facility\n (Christoph)\n\n - DRBD netlink rework, replacing the genl_magic machinery with explicit\n netlink serialization and moving the DRBD UAPI headers to\n include/uapi/linux/ (Christoph Böhmwalder)\n\n - bvec improvements: a bvec_folio() helper and making the bvec_iter\n helpers proper inline functions (Willy, Christoph)\n\n - ublk cleanups and a canceling-flag fix for the disk-not-allocated\n case (Caleb, Ming)\n\n - Partition handling fixes: bound the AIX pp_count scan, fix an of_node\n refcount leak, and replace __get_free_page() with kmalloc() (Bryam,\n Wentao, Mike)\n\n - Convert numa_node to int in blk_mq_hw_ctx and -\u003einit_request, and add\n WQ_PERCPU to the block workqueue users (Mateusz, Marco)\n\n - Block statistics and tracing: propagate in-flight to the whole disk\n on partition IO, export passthrough stats, and a new\n block_rq_tag_wait tracepoint (Tang, Keith, Aaron)\n\n - A round of removals, unexports and cleanups across bio, direct-io and\n the bvec helpers (Christoph)\n\n - Various driver fixes (mtip32xx use-after-free, rbd snap_count\n validation and strscpy conversion, nbd socket lockdep reclassify,\n virtio-blk zone report clamp, floppy) and a batch of MAINTAINERS\n email/list updates (Coly, Li, Yu, Christoph Böhmwalder)\n\n - Other little fixes and cleanups all over\n\n* tag \u0027for-7.2/block-20260615\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux: (117 commits)\n MAINTAINERS: Update Coly Li\u0027s email address\n block: check bio split for unaligned bvec\n nbd: Reclassify sockets to avoid lockdep circular dependency\n block: add configurable error injection\n block: add a str_to_blk_op helper\n block: add a \"tag\" for block status codes\n block: add a macro to initialize the status table\n floppy: Drop unused pnp driver data\n block: propagate in_flight to whole disk on partition I/O\n virtio-blk: clamp zone report to the report buffer capacity\n block: optimize I/O merge hot path with unlikely() hints\n drivers/block/rbd: Use strscpy() to copy strings into arrays\n partitions: aix: bound the pp_count scan to the ppe array\n block: Enable lock context analysis\n block/mq-deadline: Make the lock context annotations compatible with Clang\n block/Kyber: Make the lock context annotations compatible with Clang\n block/blk-mq-debugfs: Improve lock context annotations\n block/blk-iocost: Inline iocg_lock() and iocg_unlock()\n block/blk-iocost: Split ioc_rqos_throttle()\n block/crypto: Annotate the crypto functions\n ...\n" }, { "commit": "9b40ba14edcdf70240af8114092a76f75f070774", "tree": "7fd6958aa2c7ab93d86d2afed222b37c67e1c21b", "parents": [ "d29fd593e6836c96c6fd6df2b0cc6a47dda21b74", "d9b710f683dc68b5c0b7dd0c6c64aeb5d27a1ac4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:53:59 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:53:59 2026 +0530" }, "message": "Merge tag \u0027for-7.2/io_uring-20260615\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux\n\nPull io_uring updates from Jens Axboe:\n\n - Rework the task_work infrastructure.\n\n Both the local (DEFER_TASKRUN) and the normal (tctx) task_work lists\n were llist based, which is LIFO ordered, and hence each run had to do\n an O(n) list reversal pass first to restore queue order.\n Additionally, to cap the amount of task_work run, each method needed\n a retry list as well.\n\n Add a lockless MPCS FIFO queue (based on Dmitry Vyukov\u0027s intrusive\n MPSC algorithm) and switch both task_work lists to it. It performs\n better than llists and we can then also ditch the retry lists as well\n as entries are popped one-at-the-time.\n\n On top of those changes, run the tctx fallback task_work directly and\n remove the now-unused per-ctx fallback machinery entirely.\n\n - zcrx user notifications.\n\n Add a mechanism for zcrx to communicate conditions back to userspace\n via a dedicated CQE, with the initial users being notification on\n running out of buffers and on a frag copy fallback, plus\n shared-memory notification statistics.\n\n Alongside that, a series of zcrx reliability and cleanup fixes: more\n reliable scrubbing, poisoning pointers on unregistration, dropping an\n extra ifq close, adding a ctx back-pointer, reordering fd allocation\n in the export path, and killing a dead \u0027sock\u0027 member.\n\n - Allow using io_uring registered buffers for plain SEND and RECV, not\n just for the zero-copy send path.\n\n This enables targets like ublk\u0027s NBD backend to push/pull IO data\n directly to/from a registered buffer over a plain send/recv on a TCP\n socket.\n\n - Registered buffer improvements: account huge pages correctly, bump\n the io_mapped_ubuf length field to size_t, and raise the previous 1GB\n registered buffer size limit.\n\n - Restrict the ctx access exposed to io_uring BPF struct_ops programs\n by handing them an opaque type rather than the full io_ring_ctx, and\n add a separate MAINTAINERS entry for the bpf-ops code.\n\n - Allow opcode filtering on IORING_OP_CONNECT.\n\n - Validate ring-provided buffer addresses with access_ok(), and align\n the legacy buffer add limit with MAX_BIDS_PER_BGID.\n\n - Various other cleanups and minor fixes, including avoiding msghdr\n async data on connect/bind, dropping async_size for OP_LISTEN, making\n the POLL_FIRST receive side checks consistent, re-checking\n IO_WQ_BIT_EXIT for each linked work item, and using\n trace_call__##name() at guarded tracepoint call sites.\n\n* tag \u0027for-7.2/io_uring-20260615\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux: (31 commits)\n io_uring/bpf-ops: add a separate maintainer entry\n io_uring/net: make POLL_FIRST receive side checks consistent\n io_uring: remove the per-ctx fallback task_work machinery\n io_uring: run the tctx task_work fallback directly\n io_uring: switch normal task_work to a mpscq\n io_uring: switch local task_work to a mpscq\n io_uring/mpscq: add lockless multi-producer, single-consumer FIFO queue\n io_uring: grab RCU read lock marking task run\n io_uring/zcrx: kill dead \u0027sock\u0027 member in struct io_zcrx_args\n io_uring/kbuf: validate ring provided buffer addresses with access_ok()\n io_uring/net: support registered buffer for plain send and recv\n io_uring/nop: Drop a wrong comment in struct io_nop\n io_uring/net: Remove async_size for OP_LISTEN\n io_uring/net: Avoid msghdr on op_connect/op_bind async data\n io_uring/bpf-ops: restrict ctx access to BPF\n io_uring/io-wq: re-check IO_WQ_BIT_EXIT for each linked work item\n io_uring/kbuf: align legacy buffer add limit with MAX_BIDS_PER_BGID\n io_uring/zcrx: add shared-memory notification statistics\n io_uring/zcrx: notify user on frag copy fallback\n io_uring/zcrx: notify user when out of buffers\n ...\n" }, { "commit": "d29fd593e6836c96c6fd6df2b0cc6a47dda21b74", "tree": "27755b12b63354ab5955c5d9863fc815adf31b4b", "parents": [ "6f60a6033c7dfd02ba8d94111965189c9a7866d4", "7fde7e806657fbe0d33f489521b488eed94f9b39" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:27:23 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:27:23 2026 +0530" }, "message": "Merge tag \u0027hfs-v7.2-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/hfs\n\nPull hfs/hfsplus updates from Viacheslav Dubeyko:\n \"Several fixes in HFS/HFS+ of syzbot reported issues and HFS//HFS+\n fixes of xfstests failures.\n\n - fix a null-ptr-deref issue reported by syzbot (Edward Adam Davis)\n\n If the attributes file is not loaded during system mount\n hfsplus_create_attributes_file can dereference a NULL pointer.\n\n Also, add a b-tree node size check in hfs_btree_open() with the\n goal to prevent an uninit-value bug reported by syzbot for the case\n of corrupted HFS+ image.\n\n - fix __hfs_bnode_create() by using kzalloc_flex() instead of\n kzalloc() (Rosen Penev)\n\n - fix early return in hfs_bnode_read() (Tristan Madani)\n\n hfs_bnode_read() can return early without writing to the output\n buffer when is_bnode_offset_valid() fails or when\n check_and_correct_requested_ length() corrects the length to zero.\n Callers such as hfs_bnode_read_ u16() and hfs_bnode_read_u8() pass\n stack-allocated buffers and use the result unconditionally, leading\n to KMSAN uninit-value reports.\n\n The rest fix (1) generic/637, generic/729 issue for the case of HFS+\n file system, (2) generic/003, generic/637 for the case of HFS file\n system\"\n\n* tag \u0027hfs-v7.2-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/hfs:\n hfs: rework hfsplus_readdir() logic\n hfs: disable the updating of file access times (atime)\n hfs: fix incorrect inode ID assignment in hfs_new_inode()\n hfsplus: rework hfsplus_readdir() logic\n hfs/hfsplus: zero-initialize buffer in hfs_bnode_read\n hfs/hfsplus: fix u32 overflow in check_and_correct_requested_length\n hfsplus: Add a sanity check for btree node size\n hfsplus: fix issue of direct writes beyond end-of-file\n hfs/hfxplus: use kzalloc_flex()\n hfsplus: Remove the duplicate attr inode dirty marking action\n" }, { "commit": "6f60a6033c7dfd02ba8d94111965189c9a7866d4", "tree": "394c91ae04712e62e16941346bfce8171e3ec9f9", "parents": [ "31b706da2cfd8ee3352391181ccf9696bed3d25d", "e5925f33e4fa9ee313d481557607adce8e30ed2e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:14:20 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:14:20 2026 +0530" }, "message": "Merge tag \u0027nilfs2-v7.2-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/nilfs2\n\nPull nilfs2 updates from Viacheslav Dubeyko:\n \"Fixes of syzbot reported issue and various small fixes in NILFS2\n functionality.\n\n - fix hung task in nilfs_transaction_begin() (Deepanshu Kartikey)\n\n Reported by syzbot. The root cause is that user-supplied segment\n numbers were not validated before nilfs_clean_segments() began\n doing work; the range check on each segnum was performed deep\n inside the call chain by nilfs_sufile_updatev(), which emits a\n nilfs_warn() per invalid entry while still holding the segctor lock\n and the sufile mi_sem.\n\n Fix it by validating the contents of kbufs[4] in\n nilfs_clean_segments() immediately after acquiring ns_segctor_sem\n via nilfs_transaction_lock().\n\n - fix a smatch warning in nilfs_mkdir() warn (Hongling Zeng)\n\n This corrects a semantic issue related to the use of the\n ERR_PTR macro that arose from a recent VFS change.\n\n - fix a backing_dev_info reference leak (Shuangpeng Bai)\n\n setup_bdev_super() initializes sb-\u003es_bdev and takes a reference on\n the block device backing_dev_info when assigning sb-\u003es_bdi.\n\n nilfs_fill_super() takes another reference to the same\n backing_dev_info and stores it in sb-\u003es_bdi again. The extra\n reference is not paired with a matching bdi_put(), since\n generic_shutdown_super() releases sb-\u003es_bdi only once.\n\n Drop the redundant bdi_get() in nilfs_fill_super(). The single\n reference taken by setup_bdev_super() is enough and is released\n during superblock shutdown\"\n\n* tag \u0027nilfs2-v7.2-tag1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vdubeyko/nilfs2:\n nilfs2: Fix return in nilfs_mkdir\n nilfs2: fix backing_dev_info reference leak\n nilfs2: reject CLEAN_SEGMENTS ioctl with out-of-range segment numbers\n" }, { "commit": "31b706da2cfd8ee3352391181ccf9696bed3d25d", "tree": "c76b0a7317b3b58f5638b9c7cabd8f640ce5be56", "parents": [ "477c122f8c1d5d9f57c4f9c1f7a1631beaa38bcc", "ae2eb64bfd9762536f60b690840adcdf622cdcce" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:08:02 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:08:02 2026 +0530" }, "message": "Merge tag \u0027for-7.2-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux\n\nPull btrfs updates from David Sterba:\n \"The most noticeable change is to enable large folios by default, it\u0027s\n been in testing for a few releases. Related to that is huge folio\n support (still under experimental config). Otherwise a few ioctl\n updates, performance improvements and usual fixes and core changes.\n\n User visible changes:\n\n - enable large folios by default, added in 6.17 (under experimental\n build), no feature limitations, a big change internally\n\n - new ioctl to return raw checksums to userspace (a bit tricky given\n compression and tail extents), can be used for mkfs and\n deduplication optimizations\n\n - provide stable UUID for e.g. overlayfs and temp_fsid, also\n reflected in statvfs() field f_fsid, internal dev_t is hashed in to\n allow cloning\n\n - add 32bit compat version of GET_SUBVOL_INFO ioctl\n\n - in experimental build, support huge folios (up to 2M)\n\n Performance related improvements/changes:\n\n - limit bio size to the estimated optimum derived from the queue,\n this prevents build up of too much data for writeback, which could\n cause latency spikes (reported improvement 15% on sequential\n writes)\n\n - don\u0027t force direct IO to be serialized, forgotten change during\n mount API port, brings back +60% of throughput\n\n - lockless calculation of number of shrinkable extent maps, improve\n performance with many memcg allocated objects\n\n Notable fixes:\n\n - in zoned mode, fix a deadlock due to zone reclaim and relocation\n when space needs to be flushed\n\n - don\u0027t trim device which is internally not tracked as writeable\n (e.g. when missing device is being rescanned)\n\n - fix deadlock when cloning inline extent and mounted with\n flushoncommit\n\n - fix false IO failures after direct IO falls back to buffered write\n in some cases\n\n Core:\n\n - remove COW fixup mechanism completely; detect and fix changes to\n pages outside of filesystem tracking, guaranteed since 5.8, grace\n period is over\n\n - remove 2K block size support, experimental to test subpage code on\n x86_64 but now it would block folio changes\n\n - tree-checker improvements of:\n - free-space cache and tree items\n - root reference and backref items\n - extent state exceptions in reloc tree\n\n - subpage mode updates:\n - code optimizations, simplify tracking bitmaps\n - re-enable readahead of compressed extent\n - extend bitmap size to cover huge folios\n\n - add tracepoints related to sync, tree-log and transactions\n\n - device stats item tracking unification, remove item if there are no\n stats recorded, also don\u0027t leave stale stats on replaced device\n\n - allow extent buffer pages to be allocated as movable, to help page\n migration\n\n - added checks for proper extent buffer release\n\n - btrfs.ko code size reduction due to transaction abort call\n simplifications\n\n - several struct size reductions\n\n - more auto free conversions\n\n - more verbose assertions\"\n\n* tag \u0027for-7.2-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (130 commits)\n btrfs: fix use-after-free after relocation failure with concurrent COW\n btrfs: move WARN_ON on unexpected error in __add_tree_block()\n btrfs: move locking into btrfs_get_reloc_bg_bytenr()\n btrfs: lzo: reject compressed segment that overflows the compressed input\n btrfs: retry faulting in the pages after a zero sized short direct write\n btrfs: fix incorrect buffered IO fallback for append direct writes\n btrfs: fix false IO failure after falling back to buffered write\n btrfs: use verbose assertions in backref.c\n btrfs: print a message when a missing device re-appears\n btrfs: do not trim a device which is not writeable\n btrfs: return real error after lookup failure in btrfs_ioctl_default_subvol()\n btrfs: use mapping shared locking for reading super block\n btrfs: use lockless read in nr_cached_objects shrinker callback\n btrfs: switch local indicator variables to bools\n btrfs: send: pass bool for pending_move and refs_processed parameters\n btrfs: use shifts for sectorsize and nodesize\n btrfs: fix deadlock cloning inline extent when using flushoncommit\n btrfs: allocate eb-attached btree pages as movable\n btrfs: add 32-bit compat ioctl for BTRFS_IOC_GET_SUBVOL_INFO\n btrfs: derive f_fsid from on-disk fsid and dev_t\n ...\n" }, { "commit": "477c122f8c1d5d9f57c4f9c1f7a1631beaa38bcc", "tree": "09008189cc45f481fdfd408459d4fe6a5086e302", "parents": [ "974b3dec2bfb4b2726a6194105cb048a9dab0626", "e61113cfcaf71cbdf4c17e3d086d8fb7f92c62bf" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:04:00 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 12:04:00 2026 +0530" }, "message": "Merge tag \u0027dlm-7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm\n\nPull dlm updates from David Teigland:\n \"There are four fixes/cleanups in this series; none are likely to be\n issues in real usage:\n\n - improve debugfs error exit path\n\n - fix sequence number ordering in an artificial test case\n\n - fix usercopy_abort for lvb data\n\n - use hlist_for_each_entry_srcu for srcu lists\"\n\n* tag \u0027dlm-7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/linux-dlm:\n dlm: init per node debugfs before add to node hash\n dlm: fix add msg handle in send_queue ordered\n dlm: add usercopy whitelist to dlm_cb cache\n dlm: use hlist_for_each_entry_srcu for SRCU protected lists\n" }, { "commit": "974b3dec2bfb4b2726a6194105cb048a9dab0626", "tree": "9be0d6ec3791e1d6afc19bda9d601cd0a5bf84a0", "parents": [ "59b1c2aa064fdc4b91a26dce83697fea47cd0a61", "5163e6ee1ea744d412fe516235bfd9cab15141dc" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 11:58:52 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 11:58:52 2026 +0530" }, "message": "Merge tag \u0027fs_for_v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs\n\nPull udf, isofs, ext2, and quota updates from Jan Kara:\n\n - Assorted udf \u0026 isofs fixes for maliciously formatted devices\n\n - Cleanups to use kmalloc() instead of __get_free_page()\n\n - Removal of deprecated DAX code from ext2\n\n* tag \u0027fs_for_v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:\n udf: validate VAT inode size for old VAT format\n udf: validate VAT header length against the VAT inode size\n udf: validate sparing table length as an entry count, not a byte count\n isofs: bound Rock Ridge symlink components to the SL record\n ext2: fix ignored return value of generic_write_sync()\n ext2: Remove deprecated DAX support\n isofs: replace __get_free_page() with kmalloc()\n quota: allocate dquot_hash with kmalloc()\n udf: validate free block extents against the partition length\n" }, { "commit": "59b1c2aa064fdc4b91a26dce83697fea47cd0a61", "tree": "8a3f7f932c4c783970a88f10ea5ca9b506e56eaa", "parents": [ "6271f6ea7f65191762efe17c1a7d33f8922e1eeb", "82c6dd20479bb6a9625e1d63a650c3be8865e2db" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 11:56:43 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 11:56:43 2026 +0530" }, "message": "Merge tag \u0027fsnotify_for_v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs\n\nPull fsnotify updates from Jan Kara:\n\n - fanotify improvements for pidfd reporting\n\n - small cleanup in fanotify_error_event_equal\n\n* tag \u0027fsnotify_for_v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:\n fanotify: allow reporting pidfds for reaped tasks\n fanotify: report thread pidfds for FAN_REPORT_TID\n fanotify: simplify fanotify_error_event_equal\n" }, { "commit": "6271f6ea7f65191762efe17c1a7d33f8922e1eeb", "tree": "85462480581002ac3403b0b2a62314733b16beaa", "parents": [ "481329ec5b31d2c48ac6b8b703c8008133884c7e", "804826eac53cff44f88f42989fcc601c2612c0ed" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 11:50:40 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 11:50:40 2026 +0530" }, "message": "Merge tag \u0027xfs-merge-7.2\u0027 of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux\n\nPull xfs updates from Carlos Maiolino:\n \"The main highlight is the removal of experimental tag of the zone\n allocator feature.\n\n Besides that, this contains a collection of bug fixes and code\n refactoring but no new features have been added\"\n\n* tag \u0027xfs-merge-7.2\u0027 of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (29 commits)\n xfs: shut down the filesystem on a failed mount\n xfs: skip inode inactivation on a shut down mount\n xfs: move XFS_LSN_CMP to xfs_log_format.h\n xfs: shut down zoned file systems on writeback errors\n xfs: cleanup xfs_growfs_compute_deltas\n xfs: pass back updated nb from xfs_growfs_compute_deltas\n xfs: fix pointer arithmetic error on 32-bit systems\n xfs: initialize iomap-\u003eflags earlier in xfs_bmbt_to_iomap\n xfs: only log freed extents for the current RTG in zoned growfs\n xfs: add newly added RTGs to the free pool in growfs\n xfs: factor out a xfs_zone_mark_free helper\n xfs: mark struct xfs_imap as __packed\n xfs: store an agbno in struct xfs_imap\n xfs: massage xfs_imap_to_bp into xfs_read_icluster\n xfs: remove im_len field in struct xfs_imap\n xfs: cleanup xfs_imap\n xfs: remove the call to xfs_buf_reverify in xfs_trans_read_buf_map\n xfs: remove the i_ino field in struct xfs_inode\n xfs: remove xfs_setup_existing_inode\n xfs: convert xchk_inode_xref_set_corrupt to xchk_ip_xref_set_corrupt\n ...\n" }, { "commit": "481329ec5b31d2c48ac6b8b703c8008133884c7e", "tree": "406bb78e41609a198710f5e3d4880a5734a00aba", "parents": [ "673f72944dde6614a56b37a61c6097f584c90ce6", "122b52f0bab007ebeb414c8280c1def17b9ed1f4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 09:21:01 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 09:21:01 2026 +0530" }, "message": "Merge tag \u0027hardening-v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux\n\nPull hardening updates from Kees Cook:\n\n - lkdtm:\n - Add case to provoke a crash in EFI runtime services (Ard Biesheuvel)\n - add PPC_RADIX_TLBIEL test and missed isync (Sayali Patil)\n\n - stddef: Document designated initializer semantics for\n __TRAILING_OVERLAP() (Gustavo A. R. Silva)\n\n - strarray: drop redundant allocation, add __counted_by_ptr (Thorsten\n Blum)\n\n* tag \u0027hardening-v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:\n lkdtm/powerpc: add PPC_RADIX_TLBIEL test for radix MCE validation\n lkdtm/powerpc: add isync after slbmte to enforce SLB update ordering\n lkdtm: Add case to provoke a crash in EFI runtime services\n lib/string_helpers: annotate struct strarray with __counted_by_ptr\n lib/string_helpers: drop redundant allocation in kasprintf_strarray\n MAINTAINERS: add kernel hardening keyword __counted_by_ptr\n stddef: Document designated initializer semantics for __TRAILING_OVERLAP()\n" }, { "commit": "673f72944dde6614a56b37a61c6097f584c90ce6", "tree": "ba501bb9bed98ade9b839ffda6cdd52f74f6f53b", "parents": [ "ef3b7426a63c930c51d60d6c2428663d52a84e2f", "065f978a0e015c4dd9f536f5c08078a37f5509c1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 09:18:53 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 09:18:53 2026 +0530" }, "message": "Merge tag \u0027libcrypto-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux\n\nPull crypto library updates from Eric Biggers:\n\n - Drop the last architecture-specific implementation of MD5\n\n - Mark clmul32() as noinline_for_stack to improve codegen in some cases\n\n* tag \u0027libcrypto-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux:\n lib/crypto: gf128hash: mark clmul32() as noinline_for_stack\n lib/crypto: powerpc/md5: Drop powerpc optimized MD5 code\n" }, { "commit": "ef3b7426a63c930c51d60d6c2428663d52a84e2f", "tree": "99e71c07585109fd24fa2be9f7f7166b623e2981", "parents": [ "0d8c1134936f1fb6678156ab4248ac740d274525", "cbe44c389ae80362e72696ac08f7c55a83f2a050" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 09:16:37 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 09:16:37 2026 +0530" }, "message": "Merge tag \u0027crc-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux\n\nPull CRC updates from Eric Biggers:\n \"Accelerate CRC64-NVME for 32-bit ARM by refactoring the arm64 NEON\n intrinsics implementation to be shared by 32-bit and 64-bit.\n\n Also apply a similar cleanup to the 32-bit ARM NEON implementation of\n xor_gen(), where it now reuses code from the 64-bit implementation\"\n\n* tag \u0027crc-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux:\n crypto: aegis128 - Use neon-intrinsics.h on ARM too\n lib/crc: arm: Enable arm64\u0027s NEON intrinsics implementation of crc64\n lib/crc: Turn NEON intrinsics crc64 implementation into common code\n xor/arm64: Use shared NEON intrinsics implementation from 32-bit ARM\n xor/arm: Replace vectorized implementation with arm64\u0027s intrinsics\n ARM: Add a neon-intrinsics.h header like on arm64\n" }, { "commit": "0d8c1134936f1fb6678156ab4248ac740d274525", "tree": "77ce4235e5512003e636d24bd47946dab23058d3", "parents": [ "f8115f0e8a0585ef1c03d07a68b989023097d16c", "6ea0ce3a19f9c37a014099e2b0a46b27fa164564" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 09:01:23 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 09:01:23 2026 +0530" }, "message": "Merge tag \u0027v7.2-p1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6\n\nPull crypto updates from Herbert Xu:\n \"API:\n - Drop support for off-CPU cryptography in af_alg\n - Document that af_alg is *always* slower\n - Document the deprecation of af_alg\n - Remove zero-copy support from skcipher and aead in af_alg\n - Cap AEAD AD length to 0x80000000 in af_alg\n - Free default RNG on module exit\n\n Algorithms:\n - Fix vli multiplication carry overflow in ecc\n - Drop unused cipher_null crypto_alg\n - Remove unused variants of drbg\n - Use lib/crypto in drbg\n - Use memcpy_from/to_sglist in authencesn\n - Allow authenc(hmac(sha{256,384}),cts(cbc(aes))) in FIPS mode\n - Disallow RSA PKCS#1 SHA-1 sig algs in FIPS mode\n - Filter out async aead implementations at alloc in krb5\n - Fix non-parallel fallback by rstoring callback in pcrypt\n - Validate poly1305 template argument in chacha20poly1305\n\n Drivers:\n - Add sysfs PCI reset support to qat\n - Add KPT support for GEN6 devices to qat\n - Remove unused character device and ioctls from qat\n - Add support for hw access via SMCC to mtk\n - Remove prng support from crypto4xx\n - Remove prng support from hisi-trng\n - Remove prng support from sun4i-ss\n - Remove prng support from xilinx-trng\n - Remove loongson-rng\n - Remove exynos-rng\n\n Others:\n - Remove support for AIO on sockets\"\n\n* tag \u0027v7.2-p1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (196 commits)\n crypto: tegra - fix refcount leak in tegra_se_host1x_submit()\n crypto: rng - Free default RNG on module exit\n crypto: testmgr - allow authenc(hmac(sha{256,384}),cts(cbc(aes))) in FIPS mode\n hwrng: jh7110 - fix refcount leak in starfive_trng_read()\n crypto: atmel-ecc - drop dead code in atmel_ecdh_max_size\n crypto: cavium/cpt - fix DMA cleanup using wrong loop index\n crypto: marvell/octeontx - fix DMA cleanup using wrong loop index\n MAINTAINERS: make myself the maintainer of the Qualcomm QCE driver\n crypto: amcc - convert irq_of_parse_and_map to platform_get_irq\n crypto: sun4i-ss - Remove insecure and unused rng_alg\n hwrng: xilinx - Move xilinx-rng into drivers/char/hw_random/\n crypto: xilinx-trng - Replace crypto_drbg_ctr_df() with HMAC-SHA512\n crypto: xilinx-trng - Fix return value of xtrng_hwrng_trng_read()\n crypto: xilinx-trng - Remove crypto_rng interface\n crypto: exynos-rng - Remove exynos-rng driver\n hwrng: hisi-trng - Move hisi-trng into drivers/char/hw_random/\n crypto: hisi-trng - Remove crypto_rng interface\n crypto: loongson - Remove broken and unused loongson-rng\n crypto: crypto4xx - Remove insecure and unused rng_alg\n crypto: qat - validate RSA CRT component lengths\n ...\n" }, { "commit": "f8115f0e8a0585ef1c03d07a68b989023097d16c", "tree": "737e19f523d2cfd2784f17b8a343ee7457a87e7c", "parents": [ "a87bbc4578fd686d535fbd62e8bc73fc6c7c5415", "dfdfd58cce1c3f5df8733b64595448996c08e424" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:44:43 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:44:43 2026 +0530" }, "message": "Merge tag \u0027slab-for-7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vbabka/slab\n\nPull slab updates from Vlastimil Babka:\n\n - Support for \"allocation tokens\" (currently available in Clang 22+)\n for smarter partitioning of kmalloc caches based on the allocated\n object type, which can be enabled instead of the \"random\"\n per-caller-address-hash partitioning.\n\n It should be able to deterministically separate types containing a\n pointer from those that do not (Marco Elver)\n\n - Improvements and simplification of the kmem_cache_alloc_bulk() and\n mempool_alloc_bulk() API. This includes adaptation of callers\n (Christoph Hellwig)\n\n - Performance improvements and cleanups related mostly to sheaves\n refill (Hao Li, Shengming Hu, Vlastimil Babka)\n\n - Several fixups for the slabinfo tool (Xuewen Wang)\n\n* tag \u0027slab-for-7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vbabka/slab:\n mm/slab: do not limit zeroing to orig_size when only red zoning is enabled\n mm/slub: preserve original size in _kmalloc_nolock_noprof retry path\n mm: simplify the mempool_alloc_bulk API\n mm/slab: improve kmem_cache_alloc_bulk\n mm/slub: detach and reattach partial slabs in batch\n mm/slub: introduce helpers for node partial slab state\n mm/slub: use empty sheaf helpers for oversized sheaves\n tools/mm/slabinfo: remove redundant slab-\u003epartial assignment\n tools/mm/slabinfo: remove dead assignment in get_obj_and_str()\n tools/mm/slabinfo: Fix trace disable logic inversion\n MAINTAINERS: add slab-related scripts and tools to SLAB ALLOCATOR\n mm/slub: fix typo in sheaves comment\n mm, slab: simplify returning slab in __refill_objects_node()\n mm, slab: add an optimistic __slab_try_return_freelist()\n slab: fix kernel-docs for mm-api\n slab: improve KMALLOC_PARTITION_RANDOM randomness\n slab: support for compiler-assisted type-based slab cache partitioning\n mm/slub: defer freelist construction until after bulk allocation from a new slab\n" }, { "commit": "a87bbc4578fd686d535fbd62e8bc73fc6c7c5415", "tree": "0e7140f6c9857a4717bc3cadf371f7cb46d8ef42", "parents": [ "bd77e50c9a70f844d6073499f3d1a6fd193eae73", "fa34b01aa0f59355206b0807f862cced06c2b7a1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:35:59 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:35:59 2026 +0530" }, "message": "Merge tag \u0027docs-7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/docs/linux\n\nPull documentation updates from Jonathan Corbet:\n \"Things have calmed down a bit on the docs front, with no earthshaking\n changes this time around:\n\n - Ongoing work on the Japanese and Portuguese translations\n\n - Better integration of the MAINTAINERS file into the rendered\n documents, including a search interface\n\n - A seemingly infinite supply of fixes for typos, minor grammatical\n issues, and related problems that LLMs find with abandon\"\n\n* tag \u0027docs-7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/docs/linux: (93 commits)\n docs: pt_BR: Translate 3.Early-stage.rst into Portuguese\n docs: pt_BR: update \"Purpose of Defconfigs\" section in maintainer-soc.rst\n Documentation: bug-hunting.rst: fix grammar\n docs/ja_JP: translate submitting-patches.rst (interleaved-replies)\n docs: Fix minor grammatical error\n docs/{it_it,sp_SP,zh_CN,zh_TW}: update references to removed CONFIG_DEBUG_SLAB\n Documentation: process: fix brackets\n Documentation: arch: fix brackets\n docs/dyndbg: explain flags parse 1st\n docs/dyndbg: update examples \\012 to \\n\n docs: kernel-parameters: Fix stale sticore file paths\n docs: real-time: Fix duplicated sched(7) text\n docs: kgdb: Fix stale source file paths\n docs: sonypi: Fix stale header file path\n docs: kernel-parameters: Remove sa1100ir IrDA parameter\n iommu: Documentation: rearrange, update kernel-parameters\n docs: md: fix grammar in speed_limit description\n docs: changes.rst: restore pahole 1.26 minimum (regressed by sort)\n Documentation: Fix syntax of kmalloc_objs example in coding style doc\n docs: pt_BR: update maintainer-handbooks\n ...\n" }, { "commit": "bd77e50c9a70f844d6073499f3d1a6fd193eae73", "tree": "1ef3640af4ec62696fbd7f24b8d848b4f6afec4d", "parents": [ "464e2f089ecb81139433666496ecaeece421b314", "d894c48a57d78206e4df9c90d4acfaf39394806a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:28:41 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:28:41 2026 +0530" }, "message": "Merge tag \u0027fbdev-for-7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/deller/linux-fbdev\n\nPull fbdev updates from Helge Deller:\n \"Beside the removal of the Hercules monochrome ISA graphics driver and\n the corresponding text console driver, there is just the typical\n maintanance with smaller driver fixes and cleanups:\n\n Removal of drivers:\n - Hercules monochrome ISA graphics adapter driver (Ethan Nelson-Moore)\n - Hercules mdacon console driver (Ethan Nelson-Moore)\n\n Changes affecting many drivers at once:\n - possible memory leak fixes in various drivers (Abdun Nihaal)\n - many conversions to use strscpy() (David Laight)\n - Use named initializers in drivers (Uwe Kleine-König)\n\n Code fixes:\n - fbcon: don\u0027t suspend/resume when vc is graphics mode (Lu Yao)\n - modedb: fix a possible UAF in fb_find_mode() (Tuo Li)\n - modedb: Fix entry for 1920x1080-60 mode (Steffen Persvold)\n - arm: Export acorndata_8x8 font symbol for bootloader (Helge Deller)\n - omap2: fix use-after-free in omapfb_mmap (Hongling Zeng)\n\n Cleanups:\n - pxa168fb: use devm_ioremap_resource() (Alberto Arostegui)\n - provice helpers for fb_set_var() and fb_blank() and fbcon updates\n (Thomas Zimmermann)\n - fbcon: Use correct type for vc_resize() return value (Jiacheng Yu)\n - chipsfb: add missing MODULE_DESCRIPTION() macro (Rahman Mahmutović)\n - sunxvr2500: replace printk with device-aware logging functions\n (Rahman Mahmutović)\n - sm712: Fix operator precedence in big_swap macro (Li RongQing)\n - imxfb: Use of_device_get_match_data() (Rosen Penev)\n - atmel_lcdfb: Use of_device_get_match_data() (Rosen Penev)\n\n Documentation fixes:\n - grvga: Fix CLUT register address offset in comment (Eduardo Silva)\n - omap/dss: Fix stale modedb.c path (Costa Shulyupin)\n - correct CONFIG_FB_TILEBLITTING macro name in #endif comment (Ethan\n Nelson-Moore)\"\n\n* tag \u0027fbdev-for-7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/deller/linux-fbdev: (43 commits)\n fbdev: modedb: Fix misaligned fields in the 1920x1080-60 mode\n fbdev: modedb: fix a possible UAF in fb_find_mode()\n fbdev: s3fb: Use strscpy() to copy strings into arrays\n fbdev: sm501fb: Fix buffer errors in OF binding code\n fbcon: correct CONFIG_FB_TILEBLITTING macro name in #endif comment\n fbdev/arm: Export acorndata_8x8 font symbol for bootloader\n fbdev: mmpfb: Use strscpy() to copy device name\n fbdev: sisfb: Replace strlen() strcpy() pair with strscpy()\n fbdev: rivafb: Use strscpy() to copy device name\n fbdev: cyber2000fb: Use strscpy() to copy device name\n fbdev: atmel_lcdfb: Use strscpy() to copy device name\n fbdev: Do not export fbcon from fbdev\n fbdev: Wrap fbcon updates from vga-switcheroo in helper\n fbdev: Wrap user-invoked calls to fb_blank() in helper\n fbdev: Wrap user-invoked calls to fb_set_var() in helper\n fbdev: omap2: fix use-after-free in omapfb_mmap\n docs: omap/dss: Fix stale modedb.c path\n fbdev: pxa168fb: use devm_ioremap_resource() for MMIO\n fbdev: grvga: Fix CLUT register address offset in comment\n fbdev: sunxvr2500: replace printk with device-aware logging functions\n ...\n" }, { "commit": "464e2f089ecb81139433666496ecaeece421b314", "tree": "241ead8a249d1ff1a04dabe6400a8ab4667a5868", "parents": [ "fc2ce3ee106f2d53eb344f5c4963c897bbb21634", "576e40d20e100ba62e21bfda94ad948820c90a23" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:16:55 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:16:55 2026 +0530" }, "message": "Merge tag \u0027mmc-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc\n\nPull MMC updates from Ulf Hansson:\n \"MMC core:\n - Validate host\u0027s max_segs to fail gracefully\n\n MMC host:\n - davinci:\n - Avoid potential NULL dereference in the IRQ handler\n - Call mmc_add_host() in the correct order during probe\n - dw_mmc-exynos:\n - Increase DMA threshold for exynos7870\n - renesas_sdhi:\n - Add support for RZ/G2E, RZ/G2N and R-Car M3Le variants\n - sdhci-msm:\n - Add support for Hawi, Eliza and Shikra variants\n - sdhci-of-k1:\n - Add support for SD UHS-I modes\n - Add support for tuning for eMMC HS200 and SD UHS-I\"\n\n* tag \u0027mmc-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc: (24 commits)\n mmc: dw_mmc: Add desc_num field for clarity\n dt-bindings: mmc: sdhci-msm: Rename the binding to include \u0027qcom\u0027 prefix\n mmc: sdhci-of-dwcmshc: use dev_err_probe() to simplify error paths\n mmc: sdhci-of-dwcmshc: remove redundant IS_ERR() check\n dt-bindings: mmc: sdhci-msm: qcom: Add Hawi compatible\n mmc: renesas_sdhi: Add OF entry for RZ/G2E SoC\n mmc: renesas_sdhi: Add OF entry for RZ/G2N SoC\n dt-bindings: mmc: sdhci-msm: Add Eliza compatible\n mmc: davinci: fix mmc_add_host order in probe\n dt-bindings: mmc: sdhci-msm: Document the Shikra compatible\n mmc: sdhci-of-k1: add comprehensive SDR tuning support\n mmc: sdhci-of-k1: add regulator and pinctrl voltage switching support\n mmc: sdhci-of-k1: enable essential clock infrastructure for SD operation\n dt-bindings: mmc: spacemit,sdhci: add pinctrl support for voltage switching\n mmc: via-sdmmc: Simplify initialisation of pci_device_id array\n mmc: davinci: avoid NULL deref of host-\u003edata in IRQ handler\n memstick: Constify the driver id_table\n mmc: host: Move MODULE_DEVICE_TABLE next to the table itself\n mmc: renesas_sdhi: add R-Car M3Le compatibility string\n dt-bindings: mmc: renesas,sdhi: Document R-Car M3Le support\n ...\n" }, { "commit": "fc2ce3ee106f2d53eb344f5c4963c897bbb21634", "tree": "9477c5967a84f334d7c8a5d7d53ffd8b0679fdac", "parents": [ "7e54cdd0564f5393eb773b88cb49f55859e90474", "9955c92abe72564a49c293b1c15cd3b4f02ea6a0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:10:21 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:10:21 2026 +0530" }, "message": "Merge tag \u0027hwmon-for-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging\n\nPull hwmon updates from Guenter Roeck:\n \"New drivers for the following chips:\n - Analog Devices LTC4283 Swap Controller\n - Analog Devices MAX20830\n - Analog Devices MAX20860A\n - ARCTIC Fan Controller\n - Delta E50SN12051\n - Luxshare LX1308\n - Microchip EMC1812/13/14/15/33\n - Monolithic MP2985\n - Murata D1U74T PSU\n\n New chip support added to existing drivers:\n - asus-ec-sensors: Support for ROG MAXIMUS Z790 EXTREME, ROG STRIX\n B850-E GAMING WIFI, and ROG STRIX B650E-E GAMING WIFI\n - dell-smm: Add Dell Latitude 7530 to fan control whitelist\n - nct6683: Support for ASRock Z890 Pro-A\n - pmbus: Support for Flex BMR316, BMR321, BMR350 and BMR351\n - pmbus/max34440: Support for ADPM12250\n - pmbus/xdp720: Support for Infineon xdp730, and fix driver issues\n reported by Sashiko\n\n New functionality:\n - Add support for update_interval_us chip attribute, and support it\n in ina238 driver\n - Add support for guard() and scoped_guard() for subsystem locks, and\n use it in adt7411, ina2xx, and lm90 drivers\n - emc2305: Support configurable fan PWM at shutdown\n - lm63: Expose PWM frequency and LUT hysteresis as writable\n - lm75: Support active-high alert polarity\n - nct7802: Add time step attributes for tweaking responsiveness\n - pmbus/adm1266: Add rtc debugfs entries for rtc, powerup_counter,\n clear_blackbox, and firmware_revision\n - raspberrypi: Fix delayed-work teardown race, add voltage input\n support as well as voltage domain IDs\n - mcp9982: Add support for reporting external diode faults\n\n Miscellaneous bug fixes, changes and improvements:\n - Use named initializers for platform_device_id arrays and\n i2c_device_data, and remove unused driver data\n - Various drivers: Move MODULE_DEVICE_TABLE next to the table itself\n - ads7871: Convert to hwmon_device_register_with_info(), and use\n DMA-safe buffer for SPI writes\n - adt7411: document supported sysfs attributes\n - adt7462: Add of_match_table to support devicetree\n - adt7475: Add explicit header include\n - coretemp; Fix outdated documentation, coding style issues, and\n replace hardcoded core count with dynamic value\n - cros_ec: Drop unused assignment of platform_device_id driver data\n - emc2305: Fix fan channel index handling\n - gpd-fan: Reject EC PWM value 0 as invalid, fix race condition\n between device removal and sysfs access, upgrade log level from\n warn to err for platform device creation failure, initialize EC\n before registering hwmon device, drop global driver data and use\n per-device allocation\n - htu31: document debugfs serial_number\n - ina238: Add support for samples and update_interval\n - it87: Clamp negative values to zero in set_fan()\n - lm75: Add explicit header include, Add explicit default cases in\n lm75_is_visible(), and add section for sysfs interface to\n documentation\n - pmbus/lm25066: Fix PMBus coefficients for LM5064/5066/5066i\n - tmp102: Use device_property_read_string API\n - tmp401: Read \"ti,n-factor\" as signed\n - Convert zyxel,nsa320-mcu to DT schema\"\n\n* tag \u0027hwmon-for-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging: (81 commits)\n hwmon: tmp401: Read \"ti,n-factor\" as signed\n hwmon: (pmbus/lm25066) Fix PMBus coefficients for LM5064/5066/5066i\n hwmon: (gpd-fan) Reject EC PWM value 0 as invalid\n hwmon: (dell-smm) Add Dell Latitude 7530 to fan control whitelist\n hwmon: temperature: add support for EMC1812\n dt-bindings: hwmon: temperature: add support for EMC1812\n hwmon: (gpd-fan): fix race condition between device removal and sysfs access\n hwmon: (gpd-fan): upgrade log level from warn to err for platform device creation failure\n hwmon: (gpd-fan): Initialize EC before registering hwmon device\n hwmon: (gpd-fan): drop global driver data and use per-device allocation\n hwmon: (pmbus/max34440): add support adpm12250\n hwmon: (ina238) Add update_interval_us attribute\n hwmon: Add update_interval_us chip attribute\n hwmon: (ina238) Add support for samples and update_interval\n gpio: gpio-ltc4283: Add support for the LTC4283 Swap Controller\n hwmon: ltc4283: Add support for the LTC4283 Swap Controller\n dt-bindings: hwmon: Document the LTC4283 Swap Controller\n hwmon: (pmbus/xdp720) Fix driver issues xdp720/730\n hwmon: (pmbus/xdp720) Add support for efuse xdp730\n dt-bindings: hwmon/pmbus: Add Infineon xdp730\n ...\n" }, { "commit": "7e54cdd0564f5393eb773b88cb49f55859e90474", "tree": "e9754c92100b52b973b615984c46cbe99b7558d0", "parents": [ "c41cfae42c75a40783a6fc402f66c3c7852961e2", "54c726fbb53341a2553bf234d519d0cb22a2ce3d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:04:24 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 08:04:24 2026 +0530" }, "message": "Merge tag \u0027watchdog-for-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging\n\nPull watchdog updates and fixes from Guenter Roeck:\n \"Subsystem:\n - Unregister PM notifier on watchdog unregister\n - Various documentation fixes and improvements\n\n Removed drivers:\n - Remove AMD Elan SC520 processor watchdog driver\n - Drop SMARC-sAM67 support\n - Remove driver for integrated WDT of ZFx86 486-based SoC\n\n New drivers:\n - Driver for Andes ATCWDT200\n - Driver for Gunyah Watchdog\n\n Added support to existing drivers:\n - Add \"apple,t8103-wdt\" and \"apple,t8122-wdt\" compatibles to Apple\n watchdog driver\n - Add rockchip,rk3528-wdt and rockchip,rv1103b-wdt to snps,dw-wdt.yaml\n - Document IPQ9650, IPQ5210, Shikra, Nord, and Hawi in qcom-wdt.yaml\n Also document sram property and add support to get the bootstatus\n to qcom wdt driver\n - lenovo_se10_wdt: Fix use-after-rfree and add support for SE10 Gen 2\n platform\n - ti,rti-wdt: Add ti,am62l-rti-wdt compatible\n - renesas: Document RZ/G3L support and rework example for\n renesas,r9a09g057-wdt\n\n Other bug fixes and improvements:\n - Use named initializers (sc1200, ziirave_wdt)\n - Allow pic32-dmt and pic32-wdt to be built with COMPILE_TEST\n - realtek-otto: enable clock before using I/O, and prevent PHASE2 underflows\n - rti_wdt: Add reaction control\n - renesas,rzn1-wdt: Drop interrupt support and other cleanup\n - gpio_wdt: Add ACPI support\n - imx7ulp_wdt: Keep WDOG running until A55 enters WFI on i.MX94\n - sprd_wdt: Remove redundant sprd_wdt_disable() on register failure\n - bcm2835_wdt: Switch to new sys-off handler API\n - sama5d4_wdt: Fix WDDIS detection on SAM9X60 and SAMA7G5\n - hpwdt: Refine hpwdt message for UV platform\n - Convert TS-4800 bindings to DT schema\n - menz069_wdt: drop unneeded MODULE_ALIAS\n - sp5100_tco: Use EFCH MMIO for newer Hygon FCH\"\n\n* tag \u0027watchdog-for-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging: (58 commits)\n watchdog: sc1200: Drop unused assignment of pnp_device_id driver data\n watchdog: unregister PM notifier on watchdog unregister\n dt-bindings: watchdog: qcom-wdt: Document IPQ5210 watchdog\n watchdog: dev: convert to kernel-doc comments\n watchdog: core: clean up some comments\n watchdog: uapi: add comments for what bit masks apply to\n watchdog: linux/watchdog.h: repair kernel-doc comments\n watchdog: add devm_watchdog_register_device() to watchdog-kernel-api\n watchdog: ziirave_wdt: Use named initializers for struct i2c_device_id\n watchdog: realtek-otto: enable clock before using I/O\n watchdog: realtek-otto: prevent PHASE2 underflows\n dt-bindings: watchdog: qcom-wdt: Document IPQ9650 watchdog\n dt-bindings: watchdog: renesas,rzn1-wdt: interrupts are not required\n dt-bindings: watchdog: apple,wdt: Add t8122 compatible\n watchdog: apple: Add \"apple,t8103-wdt\" compatible\n watchdog: rzn1: remove now obsolete interrupt support\n dt-bindings: watchdog: Add watchdog compatible for RK3528\n watchdog: convert the Kconfig dependency on OF_GPIO to OF\n watchdog: Remove AMD Elan SC520 processor watchdog driver\n watchdog: lenovo_se10_wdt: Fix use-after-free and resource leak risk\n ...\n" }, { "commit": "c41cfae42c75a40783a6fc402f66c3c7852961e2", "tree": "80057a78b1ae852d4f867dbeae9c35ad16bb603e", "parents": [ "1e66d46f6d66215ddd7126140e7c1b567400b524", "f846d68992142034b1d34a83200a10cdc713eeda" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:56:47 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:56:47 2026 +0530" }, "message": "Merge tag \u0027spi-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi\n\nPull spi updates from Mark Brown:\n \"This has been quite a busy release, mainly due to the subsystem wide\n work Johan Hovold has done to modernise resource allocation for the\n subsystem on probe, the subsystem did some very clever allocation\n management pre devm which didn\u0027t quite mesh comfortably with managed\n allocations and made it far too easy to introduce error handling and\n removal bugs.\n\n - Cleanup and simplification of controller struct allocation, moving\n everything over to devm and making the devm APIs more robust, from\n Johan Hovold\n\n - Support for spi-mem devices that don\u0027t assert chip select and\n support for a secondary read command for memory mapped flashes,\n some commits for this are shared with mtd.\n\n - Support for SpacemiT K1\"\n\n* tag \u0027spi-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi: (118 commits)\n spi: Fix mismatched DT property access types\n spi: xilinx: use FIFO occupancy register to determine buffer size\n spi: spi-mem: Fix spi_controller_mem_ops kdoc\n spi: xilinx: let transfers timeout in case of no IRQ\n spi: dt-bindings: nuvoton,npcm750-fiu: Convert to DT schema\n spi: meson-spifc: fix runtime PM leak on remove\n spi: Use named initializers for platform_device_id arrays\n spi: rzv2h-rspi: Add suspend/resume support\n spi: dw-pci: remove redundant pci_free_irq_vectors() calls\n spi: ep93xx: fix double-free of zeropage on DMA setup failure\n spi: cadence-xspi: Revert COMPILE_TEST support\n spi: cadence-xspi: Support 32bit and 64bit slave dma interface\n spi: tegra210-quad: Allocate DMA memory for DMA engine\n spi: imx: replace dmaengine_terminate_all() with dmaengine_terminate_sync()\n spi: fsl-lpspi: terminate the RX channel on TX prepare failure path\n spi: fsl-lpspi: replace dmaengine_terminate_all() with dmaengine_terminate_sync()\n spi: atmel: fix DMA channel and bounce buffer leaks\n spi: omap2-mcspi: Use of_device_get_match_data()\n spi: Use named initializers for arrays of i2c_device_data\n spi: aspeed: Replace VLA parameter with flat pointer in calibration helper\n ...\n" }, { "commit": "1e66d46f6d66215ddd7126140e7c1b567400b524", "tree": "e2a7380df312fb64b5a05f9d0c1c4f0043ad07da", "parents": [ "aaee7820e2f7370034c4fd74eb36588eaae65966", "fd964ee0ac9ef14fdc03e30d0ac73459cd60e469" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:50:31 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:50:31 2026 +0530" }, "message": "Merge tag \u0027regulator-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator\n\nPull regulator updates from Mark Brown:\n \"The development of the regulator subsystem continues to be quite\n quiet, we\u0027ve got several new devices, removal of one old device and\n some kernel wide cleanup of platform devices but nothing in the core.\n\n - Cleanups of platform_device_id usage\n\n - Filling out and fixing of the description of the MediaTek MT6359\n\n - Removal of the PCAP regulator driver, the MFD has been removed\n\n - New device support for Qualcomm Nord RPMH, PM8109, PM8150 and\n PMAU0102, and SG Micro SGM3804\"\n\n* tag \u0027regulator-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator: (23 commits)\n regulator: dt-bindings: mt6311: Convert to DT schema\n regulator: qcom_smd-regulator: Add PM8019\n regulator: dt-bindings: qcom,smd-rpm-regulator: Add PM8019\n regulator: mt6359: Fix vbbck default internal supply name\n regulator: bq257xx: drop confusing configuration of_node\n regulator: Unify usage of space and comma in platform_device_id arrays\n regulator: Use named initializers for platform_device_id arrays\n regulator: Drop unused assignment of platform_device_id driver data\n regulator: scmi: fix of_node refcount leak in scmi_regulator_probe()\n regulator: remove used pcap regulator driver\n regulator: add SGM3804 Dual Output driver\n regulator: dt-bindings: document the SGM3804 Dual Output regulator\n regulator: mt6359: Add proper ldo_vcn33_[12] regulators\n regulator: mt6359: Add regulator supply names\n regulator: mt6359: const-ify regulator descriptions\n regulator: dt-bindings: mt6359: Deprecate bogus vcn33_[12]_* split regulators\n regulator: dt-bindings: mt6359: Drop regulator-name pattern restrictions\n regulator: palmas: Move MODULE_DEVICE_TABLE next to the table itself\n regulator: qcom_smd: Add PM8150 regulators\n regulator: dt-bindings: qcom,smd-rpm-regulator: Document PM8150 IC\n ...\n" }, { "commit": "aaee7820e2f7370034c4fd74eb36588eaae65966", "tree": "83172adfde086712895b3b8cde5f8fcf976615a1", "parents": [ "2779759c090ea0e78109a0cad0a81d869adfb459", "34808ac8ddafc3e2c2a59e84eaab0a410e7a0fdc" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:48:30 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:48:30 2026 +0530" }, "message": "Merge tag \u0027regmap-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap\n\nPull regmap update from Mark Brown:\n \"This time around we just have a single fix for a sparse warning\"\n\n* tag \u0027regmap-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:\n regmap-i2c: fix sparse warning in regmap_smbus_word_write_reg16\n" }, { "commit": "2779759c090ea0e78109a0cad0a81d869adfb459", "tree": "db4d89445018c49d313640d347dbe1d375fc4170", "parents": [ "31e6aeafcdde965aa10e10e93ee186520555ec3d", "90220ddfa8e2c93e26af2cd51d6158ca2243c622" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:43:29 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:43:29 2026 +0530" }, "message": "Merge tag \u0027i2c-7.2-part1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/andi.shyti/linux\n\nPull i2c updates from Andi Shyti:\n \"This pull request is mostly made of cleanups and small infrastructure\n improvements across the I2C core, drivers and bindings. It also adds\n support for three drivers and a few new compatibles.\n\n Two major cleanup across drivers and core code:\n - use named initializers in device ID tables\n - replace dev_err() with dev_err_probe()\n\n Drivers:\n - at24: use named initializers for arrays of i2c_device_data\n - at91: add MCHP_LAN966X_PCI dependency\n - cadence: add shutdown callback\n - k1: enable by default on SpacemiT\n - mxs: improve documentation\n - qcom-geni: use pm_runtime_force_suspend/resume for system sleep\n - tegra:\n - disable fair arbitration on non-MCTP buses\n - allocate DMA buffers from the correct DMA device\n - designware:\n - handle active target shutdown cleanly\n - add shutdown callbacks for platform and PCI drivers\n - adopt the new PM_RUNTIME_ACQUIRE() helpers\n\n DT bindings:\n - convert davinci bindings to DT schema\n\n Core and muxes:\n - acpi: report missing I2C resources as -ENOENT\n - gpiolib: add gpiod_is_single_ended() helper for I2C GPIO users\n - i2c-mux-reg: add generic firmware node support\n - a set of 10 patches from Johan Hovold fixing adapter registration\n races, cleanup paths and resource management issues\n\n New support:\n - DesignWare LECA0003 (ACPI ID)\n - Loongson LS2K0300 I2C controller (new driver)\n - Qualcomm CCI Glymur and Shikra compatibles\"\n\n* tag \u0027i2c-7.2-part1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/andi.shyti/linux: (46 commits)\n i2c: mxs: add missing kernel-doc for struct mxs_i2c_dev members\n i2c: qcom-geni: Use pm_runtime_force_{suspend,resume} helpers\n dt-bindings: i2c: qcom-cci: Document Glymur compatible\n dt-bindings: i2c: qcom-cci: Document Shikra compatible\n i2c: mux: reg: use device property accessors\n i2c: acpi: Return -ENOENT when no resources found in i2c_acpi_client_count()\n i2c: at91: Add MCHP_LAN966X_PCI dependency\n i2c: eg20t: Consistently define pci_device_ids using named initializers\n i2c: designware-pcidrv: Consistently define pci_device_ids using named initializers\n i2c: bcm-kona: fix spelling mistake in timeout-check comment\n i2c: cadence: Add shutdown handler\n i2c: tegra: Disable fair arbitration for non-MCTP buses\n i2c: tegra: use dmaengine_get_dma_device() for DMA buffer allocation\n i2c: busses: make K1 driver default for SpacemiT platforms\n i2c: Use named initializers for arrays of i2c_device_data\n i2c: core: clean up adapter registration error label\n i2c: core: clean up bus id allocation\n i2c: core: fix adapter deregistration race\n i2c: core: fix adapter registration race\n i2c: core: disable runtime PM on adapter registration failure\n ...\n" }, { "commit": "31e6aeafcdde965aa10e10e93ee186520555ec3d", "tree": "67c8ac1ad9108695dcb69a152597998cd30c6ef8", "parents": [ "77d084d66b7694e2a912abdd8b9e5a0e7a32d28e", "2abcfdd91e6acc9999c1a38a6a077835a9db380c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:38:04 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:38:04 2026 +0530" }, "message": "Merge tag \u0027pwrseq-updates-for-v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux\n\nPull power sequencing updates from Bartosz Golaszewski:\n \"A set of extensions to the M.2 pwrseq driver allowing it to work with\n more cards than just the one from Qualcomm we supported initially.\n There\u0027s also a tweak to debugfs output and a new function that will be\n used by a bluetooth driver in the next cycle.\n\n Power Sequencing core:\n\n - Add a helper allowing consumers to access the struct device object\n associated with a pwrseq provider\n\n - Print the power sequencing device\u0027s parent in debugfs to add more\n debugging information\n\n Driver updates:\n\n - Extend/rework the M.2 power sequencing driver in order to allow it\n to support more M.2 cards, not just WCN7850\"\n\n* tag \u0027pwrseq-updates-for-v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:\n power: sequencing: pcie-m2: Add PCI ID 0x1103 for WCN6855 Bluetooth\n power: sequencing: Add an API to return the pwrseq device\u0027s \u0027dev\u0027 pointer\n power: sequencing: pcie-m2: Create BT node based on the pci_device_id[] table\n power: sequencing: pcie-m2: Create serdev for PCI devices present before probe\n power: sequencing: pcie-m2: Improve PCI device ID check\n power: sequencing: pcie-m2: Allow creating serdev for multiple PCI devices\n power: sequencing: pcie-m2: Fix inconsistent function prefixes\n power: sequencing: print power sequencing device parent in debugfs\n" }, { "commit": "77d084d66b7694e2a912abdd8b9e5a0e7a32d28e", "tree": "1b21610ec0bfccc6fd4254c66b8ee0641c2c4333", "parents": [ "6b617c82580599994c8dd078ab0dcb0375b42eda", "f953585dafd71ecb0897f9def9c0a3702afc1bf8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:30:52 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:30:52 2026 +0530" }, "message": "Merge tag \u0027gpio-updates-for-v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux\n\nPull gpio updates from Bartosz Golaszewski:\n \"There\u0027s one new driver, one legacy driver removed, a kunit test-suite\n for the GPIO core, support for new models in existing drivers and a\n slew of various changes in many places though I can\u0027t think of\n anything controversial that would stand out - it\u0027s been a relatively\n calm cycle.\n\n GPIO core:\n - Add an initial set of kunit test cases for the GPIO subsystem\n - Use the devres owner as the GPIO chip\u0027s parent in absence of any\n other parent\n - Fix const-correctness of GPIO chip SRCU guards\n - Provide new GPIO consumer interfaces: gpiod_is_single_ended() and\n fwnode_gpiod_get()\n - Quarantine all legacy GPIO APIs in linux/gpio/legacy.h\n - Use __ro_after_init where applicable\n\n New drivers:\n - Add driver for the GPIO controller on Waveshare DSI TOUCH panels\n\n Removed drivers:\n - Remove the obsolete ts5500 GPIO driver\n\n Driver updates:\n - Modernize gpio-timberdale: remove platform data support and use\n generic device property accessors\n - Extend test build coverage by enabling COMPILE_TEST for more GPIO\n drivers\n - Add some missing dependencies in Kconfig\n - Add support for sparse fixed direction to gpio-regmap\n - Remove dead code from gpio-nomadik\n - use BIT() in gpio-mxc\n - use bitmap_complement() in gpio-xilinx and gpio-pca953x\n - Use more appropriate printing functions where applicable\n - Use named initializers for platform_device_id and i2c_device_id\n arrays\n - Convert gpio-altera to using the generic GPIO chip helper library\n - Add support for new models to gpio-dwapb, gpio-zynq, gpio-usbio and\n gpio-tegra186\n - Unify the naming convention for Qualcomm in GPIO drivers\n - Fix interrupt bank mapping to GPIO chips in gpio-mt7621\n - Add support for the lines-initial-states property to gpio-74x164\n - Switch to using dynamic GPIO base in gpio-ixp4xx\n - Move the handling of an OF quirk from ASoC to gpiolib-of.c where\n other such quirks live\n - Use handle_bad_irq() in gpio-ep93xx\n - Some other minor tweaks and refactorings\n\n Devicetree bindings:\n - Document the Waveshare GPIO controller for DSI TOUCH panels\n - Document new models: Tegra238 in gpio-tegra186 and EIO GPIO in\n gpio-zynq\n - Add new properties for gpio-dwapb and fairchild,74hc595\n - Fix whitespace issues\n - Sort compatibles alphabetically in gpio-zynq\n\n Documentation:\n - Fix kerneldoc warnings in gpio-realtek-otto\n\n Misc:\n - Attach software nodes representing GPIO chips to the actual struct\n device objects associated with them in some legacy platforms\n enabling real firmware node lookup instead of string matching\n - Drop unneeded dependencies on OF_GPIO from bus and staging drivers\"\n\n* tag \u0027gpio-updates-for-v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux: (62 commits)\n gpio: nomadik: remove dead DB8540 code from \u003cgpio/gpio-nomadik.h\u003e\n gpio: mt7621: fix interrupt banks mapping on gpio chips\n bus: ts-nbus: drop unneeded dependency on OF_GPIO\n staging: media: max96712: drop unneeded dependency on OF_GPIO\n gpiolib: Replace strcpy() with memcpy()\n gpio: remove obsolete UAF FIXMEs from lookup paths\n gpio: core: fix const-correctness of gpio_chip_guard\n gpio: mxc: use BIT() macro\n gpio: realtek-otto: fix kernel-doc warnings\n gpio: max77620: Unify usage of space and comma in platform_device_id array\n gpio: Use named initializers for platform_device_id arrays\n gpio: cros-ec: Drop unused assignment of platform_device_id driver data\n ARM: omap1: enable real software node lookup of GPIOs on Nokia 770\n ARM: omap1: use platform_device_register_full() for GPIO devices on OMAP 16xx\n ARM: omap1: drop unused variable from omap16xx_gpio_init()\n gpio: gpiolib: use seq_puts() for plain strings\n gpio: ts5500: remove obsolete driver\n gpio: add kunit test cases for the GPIO subsystem\n kunit: provide kunit_platform_device_unregister()\n kunit: provide kunit_platform_device_register_full()\n ...\n" }, { "commit": "6b617c82580599994c8dd078ab0dcb0375b42eda", "tree": "c2a607ac8316e31ff773eefac4e4ffd15183cbcb", "parents": [ "fd1878584db2f0ab3fe8f2b3d8d59316fffa47fa", "6f9b73071c15001530e6697491b6db1bf639f4c7" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:26:47 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:26:47 2026 +0530" }, "message": "Merge tag \u0027pwm/for-7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ukleinek/linux\n\nPull pwm updates from Uwe Kleine-König:\n \"The usual mixture of minor fixes, a few cleanups, a new driver and dt\n updates for the pwm subsystem.\n\n Thanks to Chen Ni, Devi Priya, Manish Baing, Maurice Hieronymus,\n Ronaldo Nunez, Rosen Penev, Shiji Yang and Yixun Lan for the actual\n changes and Bjorn Andersson, Conor Dooley, Frank Li, Michal Wilczynski\n and Rob Herring for reviews and acks\"\n\n* tag \u0027pwm/for-7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ukleinek/linux:\n pwm: th1520: Remove requirement for mul_u64_u64_div_u64_roundup\n dt-bindings: pwm: stmpe: Drop legacy binding\n pwm: pca9685: Use named initializers for struct i2c_device_id\n pwm: pxa: Add optional bus clock\n dt-bindings: pwm: marvell,pxa-pwm: Add SpacemiT K3 PWM support\n pwm: ipq: Add missing module description\n pwm: stm32: Make use of mul_u64_u64_div_u64_roundup()\n pwm: Consistently define pci_device_ids using named initializers\n pwm: Driver for qualcomm ipq6018 pwm block\n pwm: imx27: Fix variable truncation in .apply()\n pwm: mediatek: correct mt7628 clock source setting\n pwm: mediatek: set mt7628 pwm45_fixup flag to false\n pwm: atmel-tcb: Remove unneeded semicolon\n" }, { "commit": "fd1878584db2f0ab3fe8f2b3d8d59316fffa47fa", "tree": "1c1ed6f794fd9d2bb3a15f5192dca26ffa4f00ab", "parents": [ "2b5f5609ae9b218c9ee0cdb7f624ddb03d44cf84", "86e3bbc716332600e9e087f8a4889f4d9714a99c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:21:48 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:21:48 2026 +0530" }, "message": "Merge tag \u0027chrome-platform-firmware-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux\n\nPull chrome-platform firmware updates from Tzung-Bi Shih:\n\n - Add bound checks when iterating the coreboot table\n\n - Skip failing entries only instead of aborting the whole device\n populate from the coreboot table\n\n* tag \u0027chrome-platform-firmware-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux:\n firmware: google: Skip failing entries instead of aborting populate\n firmware: google: Add bounds checks in coreboot_table_populate()\n" }, { "commit": "2b5f5609ae9b218c9ee0cdb7f624ddb03d44cf84", "tree": "995782ec25bc91dc863671c50389fea5dc22ef6e", "parents": [ "b0820861748f43759bbf4c319ed1277cff3c5921", "883f968dcbb08a155101e3a943557530d4ac0463" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:19:53 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 07:19:53 2026 +0530" }, "message": "Merge tag \u0027chrome-platform-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux\n\nPull chrome-platform updates from Tzung-Bi Shih:\n \"Improvements:\n\n - Use named initializers for struct i2c_device_id\n\n Fixes:\n\n - Fix a probe race between cros_ec_sensorhub and cros_ec_sysfs\n\n - Check for the presence of ACPI_COMPANION() for drivers converted\n from acpi_driver to platform_driver to avoid issues where\n device_match_driver_override() might forcibly match the driver to\n the device\n\n - Fix a possible UAF in cros_ec_chardev\n\n - Prevent build for big-endian systems as CHROME_PLATFORMS drivers\n are only running and testing under little-endian systems\n\n Cleanups:\n\n - Drop some redundant bits in cros_kbd_led_backlight and Kconfig\"\n\n* tag \u0027chrome-platform-v7.2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux:\n platform/chrome: Prevent build for big-endian systems\n platform/chrome: Remove superfluous dependencies from CROS_EC\n platform/chrome: cros_ec_chardev: Introduce rwsem for protecting ec_dev\n platform/chrome: cros_ec_chardev: Add event relayer\n platform/chrome: cros_ec_chardev: Move data to chardev_pdata\n platform/chrome: cros_ec_chardev: Introduce chardev_data\n platform/chrome: Use named initializers for struct i2c_device_id\n platform/chrome: wilco_ec: event: Check ACPI_COMPANION()\n platform/chrome: chromeos_tbmc: Check ACPI_COMPANION()\n platform/chrome: chromeos_privacy_screen: Check ACPI_COMPANION()\n platform/chrome: cros_kbd_led_backlight: Drop CONFIG_MFD_CROS_EC_DEV ifdeffery\n platform/chrome: cros_kbd_led_backlight: Pass keyboard_led as parameter\n platform/chrome: cros_kbd_led_backlight: Drop max_brightness from driver data\n platform/chrome: Resolve kb_wake_angle visibility race\n" }, { "commit": "b0820861748f43759bbf4c319ed1277cff3c5921", "tree": "c77f30fc402607316a59656f1b08977e6cf57ee8", "parents": [ "7d36844ae7b0a9ef61dd345c8dfe1d47a5199a53", "2b9ad7a6154e0938b9458691536296dd0224942d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 06:26:12 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 06:26:12 2026 +0530" }, "message": "Merge tag \u0027x86_tdx_for_7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull x86 TDX updates from Dave Hansen:\n \"There are a few cleanups, and some changes that should allow TDX and\n kexec to coexist nicely.\n\n The biggest change, however, is support for updating the TDX module\n after boot, just like CPU microcode. TDX users really want this\n because it lets them do security updates without tearing things down\n and rebooting.\n\n - Add TDX module update support\n\n - Make kexec and TDX finally place nice together\n\n - Put TDX error codes into a single header\"\n\n* tag \u0027x86_tdx_for_7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (30 commits)\n x86/virt/tdx: Document TDX module update\n x86/virt/tdx: Enable TDX module runtime updates\n x86/virt/tdx: Refresh TDX module version after update\n coco/tdx-host: Lock out module updates when reading version\n x86/virt/seamldr: Add module update locking\n x86/virt/tdx: Restore TDX module state\n x86/virt/seamldr: Initialize the newly-installed TDX module\n x86/virt/seamldr: Install a new TDX module\n x86/virt/tdx: Reset software states during TDX module shutdown\n x86/virt/seamldr: Shut down the current TDX module\n x86/virt/seamldr: Abort updates after a failed step\n x86/virt/seamldr: Introduce skeleton for TDX module updates\n x86/virt/seamldr: Allocate and populate a module update request\n coco/tdx-host: Implement firmware upload sysfs ABI for TDX module updates\n coco/tdx-host: Don\u0027t expose P-SEAMLDR information on CPUs with erratum\n coco/tdx-host: Expose P-SEAMLDR information via sysfs\n x86/virt/seamldr: Add a helper to retrieve P-SEAMLDR information\n x86/virt/seamldr: Introduce a wrapper for P-SEAMLDR SEAMCALLs\n coco/tdx-host: Expose TDX module version\n coco/tdx-host: Introduce a \"tdx_host\" device\n ...\n" }, { "commit": "7d36844ae7b0a9ef61dd345c8dfe1d47a5199a53", "tree": "02b05c0c3e8704e6cf3ab472167e4b09a81ecf47", "parents": [ "ee7a8593c0e21bc8877d2b1b53187359c5e7c8fe", "39406c05f8f150f1685839acd38ffdd69ff92031" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 06:17:23 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 06:17:23 2026 +0530" }, "message": "Merge tag \u0027x86_mm_for_7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull x86 mm updates from Dave Hansen:\n \"There\u0027s a small comment fixup, followed by an actual bug fix. The bug\n fix is slightly new territory for us. It fixes a real regression, but\n it is from ~6.18 not a _new_ regression. It was mostly only biting\n folks doing hotplug memory testing. So, it was queued in here and held\n for the merge window instead of going via x86/urgent.\n\n - Fix freeing of PMD-sized vmemmap pages\n\n - Update comment about pgd_list\"\n\n* tag \u0027x86_mm_for_7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n x86/mm: Fix freeing of PMD-sized vmemmap pages\n x86: Update comment about pgd_list\n" }, { "commit": "ee7a8593c0e21bc8877d2b1b53187359c5e7c8fe", "tree": "21de00e94075135e07ea4c12a6c1d8dc196e3cf6", "parents": [ "c61f479852493fd2cf5ca754e42ac272b1e2d2c5", "1d2cf6d5b5994df68e3431f0a1a8f72a191b8aad" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 06:10:22 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 06:10:22 2026 +0530" }, "message": "Merge tag \u0027x86_misc_for_7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull misc x86 updates from Dave Hansen:\n \"These are the usual random pile, with the one exception of moving Rick\n over to be a TDX maintainer. Rick has been doing a great job with TDX\n contributions, especially on the host side of things. It\u0027s time to\n promote him to \"M\".\n\n - Move Rick Edgecombe to TDX maintainer\n\n - Remove unused header\n\n - Ensure printf() validation in all configs\"\n\n* tag \u0027x86_misc_for_7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n MAINTAINERS: Move Rick Edgecombe to TDX maintainer\n x86: Remove unnecessary architecture-specific \u003casm/device.h\u003e\n x86/bug: Put HAVE_ARCH_BUG_FORMAT_ARGS WARN definitions inside __ASSEMBLER__\n x86/bug: Add printf() validation to HAVE_ARCH_BUG_FORMAT_ARGS WARNs\n" }, { "commit": "c61f479852493fd2cf5ca754e42ac272b1e2d2c5", "tree": "8f99e3a4c0fbecc54eb2f0631d0c58e344c5c26c", "parents": [ "97cc7dc16aaee163e15173009c063fc9cd42b5ff", "9d8460a1c7a6b0f2dc6302e5d0f31d4e8c2a7913" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:57:30 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:57:30 2026 +0530" }, "message": "Merge tag \u0027x86_sev_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull x86 SEV updates from Borislav Petkov:\n\n - Remove redundant GHCB initialization guards in the SEV page state and\n SVSM call paths now that the GHCB helpers handle early-boot fallback\n internally\n\n - Skip SNP initialization in the CCP driver immediately when the\n preparation step fails rather than proceeding to an operation that\n will certainly fail\n\n - Abort SNP preparation and return an error when not all CPUs are\n online, since the firmware enforces that every CPU enables SNP and\n will fail init if not\n\n - Simplify the VMM communication exception entry path by replacing\n separate kernel and user mode macros with a single handler that\n dispatches based on the current privilege level\n\n* tag \u0027x86_sev_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n x86/sev: Remove redundant ghcbs_initialized checks around __sev_{get,put}_ghcb()\n crypto/ccp: Skip SNP_INIT if preparation fails\n x86/sev: Do not initialize SNP if missing CPUs\n x86/entry: Zap the #VC entry user and kernel macros\n" }, { "commit": "97cc7dc16aaee163e15173009c063fc9cd42b5ff", "tree": "727f420b3cfbd03a1cd983d9b91177edff191d1f", "parents": [ "454761e12197938b6b362b7ab485e6e025bd4320", "098bcea71b8d257d80b0037b97b66070806600a5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:45:36 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:45:36 2026 +0530" }, "message": "Merge tag \u0027x86_microcode_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull x86 microcode loader updates from Borislav Petkov:\n\n - Move the zero-revision fixup for AMD microcode to the patch level\n retrieval function and restrict it to Zen family processors, ensuring\n patch level arithmetic always operates on a valid revision\n\n - Fix an incorrect comment about which CPUID bit is checked when\n determining whether the microcode loader should be disabled\n\n - Add the latest Intel microcode revision data for a broad range of\n processor models and steppings and add the script which generates the\n header of minimum expected Intel microcode revisions\n\n* tag \u0027x86_microcode_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n x86/microcode/AMD: Move the no-revision fixup to get_patch_level()\n x86/microcode: Fix comment in microcode_loader_disabled()\n scripts/x86/intel: Add a script to update the old microcode list\n x86/microcode/intel: Refresh old_microcode defines with Nov 2025 release\n" }, { "commit": "454761e12197938b6b362b7ab485e6e025bd4320", "tree": "26250b1abe4c690e3cf882ffa82e2665b1c22c2f", "parents": [ "3c26a6bc40fac7051b002411e771a8a5faed028f", "a5f28da54f36f1a8b289d9bdd3e780b2ede0da6f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:41:22 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:41:22 2026 +0530" }, "message": "Merge tag \u0027x86_cleanups_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull x86 cleanups from Borislav Petkov:\n\n - The usual pile of cleanups and fixlets the cat dragged in\n\n* tag \u0027x86_cleanups_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n x86/cpu: Remove obsolete aperfmperf_get_khz() declaration\n x86/pmem: Check for platform_device_alloc() retval\n x86/platform/uv: Use str_enabled_disabled() in uv_nmi_setup_hubless_intr()\n x86/cpu: Keep the PROCESSOR_SELECT menu together\n x86/tlb: Convert copy_from_user() + kstrtouint() to kstrtouint_from_user()\n x86/purgatory: Fix #endif comment\n x86/boot: Get rid of kstrtoull()\n x86/boot/compressed: Use boot_kstrtoul() for hugepages\u003d parsing\n" }, { "commit": "3c26a6bc40fac7051b002411e771a8a5faed028f", "tree": "2062e630a7cad362202f735d591da92ded15cd98", "parents": [ "4b5713ef2f929dd8720fcdb66c063643ef9e3bcb", "1cfa74c683ea82d37156ccd7ab4f4659056dc701" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:35:40 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:35:40 2026 +0530" }, "message": "Merge tag \u0027x86_cache_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull x86 resource control updates from Borislav Petkov:\n \"Preparatory work for MPAM counter assignment:\n\n - Simplify the error handling path when creating monitor group event\n configuration directories\n\n - Make the MBM event filter configurable only on architectures that\n support it and expose this with the respective file modes in the\n event config\n\n - Disallow the MBA software controller on systems where MBM counters\n are assignable, as it requires continuous bandwidth measurement\n that assignable counters do not guarantee\n\n - Replace a compile-time Kconfig option for fixed counter assignment\n with a per-architecture runtime property, and expose whether the\n counter assignment mode is changeable to userspace\n\n - Continue counter allocation across all domains instead of aborting\n at the first failure\n\n - Document that automatic MBM counter assignment is best effort and\n may not assign counters to all domains\n\n - Document the behavior of task ID 0 and idle tasks in the resctrl\n tasks file\"\n\n* tag \u0027x86_cache_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n fs/resctrl: Document tasks file behaviour for task id 0 and idle tasks\n fs/resctrl: Document that automatic counter assignment is best effort\n fs/resctrl: Continue counter allocation after failure\n fs/resctrl: Add monitor property \u0027mbm_cntr_assign_fixed\u0027\n fs/resctrl: Disallow the software controller when MBM counters are assignable\n x86,fs/resctrl: Create \u0027event_filter\u0027 files read only if they\u0027re not configurable\n fs/resctrl: Tidy up the error path in resctrl_mkdir_event_configs()\n" }, { "commit": "4b5713ef2f929dd8720fcdb66c063643ef9e3bcb", "tree": "e4760475adfa9bb2fc8b3e97c7b50e5a7d3bb80c", "parents": [ "80476f22b8b7e193b26f285a7c9f9e4b63abca16", "2556746302b335b3a9b016e315c7a14f272eab12" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:31:01 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:31:01 2026 +0530" }, "message": "Merge tag \u0027edac_updates_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/ras/ras\n\nPull EDAC updates from Borislav Petkov:\n\n - Fix a malformed Kconfig default for the AMD Address Translation\n Library\n\n - Make sure i10nm loads successfully when the ADXL address decoder is\n absent because former has decoding capabilities too\n\n - Ensure error reporting is cleanly disabled on driver teardown and on\n failed initialization for several legacy Intel EDAC drivers\n\n - Fix a grammar issue in a diagnostic warning in the Sandy Bridge\n driver\n\n - Fix a missing resource release callback and incorrect memory topology\n parsing in the igen6 driver, and add support for Intel Panther Lake-H\n and Nova Lake-H SoCs\n\n - Fix an out-of-bounds shift causing undefined behaviour in the Skylake\n driver\n\n - Consolidate memory controller register access helpers into shared\n common code across the Intel Skylake, Ice Lake, and Meteor Lake\n drivers\n\n - Introduce sub-channel awareness and Rank Retry Logic improvements to\n the Intel Skylake and i10nm drivers in preparation for Diamond Rapids\n server support\n\n - Add Rank Retry Logic support for Intel Diamond Rapids server to\n imh_edac\n\n - Make In-Band ECC detection registers configurable per SoC in the\n igen6 driver\n\n - Standardize PCI device ID table definitions across all EDAC drivers\n to use named field initializers and standard PCI helper macros\n\n* tag \u0027edac_updates_for_v7.2_rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/ras/ras: (22 commits)\n EDAC: Consistently define pci_device_ids using named initializers\n EDAC/igen6: Add Intel Nova Lake-H SoC support\n EDAC/igen6: Make registers for detecting IBECC configurable\n EDAC/imh: Add RRL support for Intel Diamond Rapids server\n EDAC/{skx_common,i10nm}: Prepare RRL for sub-channel granularity\n EDAC/skx_common: Add SubChannel support to ADXL decode\n EDAC/{skx_common,i10nm}: Move RRL handling to common code\n EDAC/{skx_common,i10nm}: Introduce rrl_ctrl_mode\n EDAC/{skx_common,i10nm}: Rename rrl_mode to rrl_source_type\n EDAC/{skx_common,skx,i10nm}: Split skx_set_decode()\n EDAC/{skx_common,i10nm,imh}: Move MC register access helpers to skx_common\n EDAC/{skx_common,skx}: Fix UBSAN shift-out-of-bounds in skx_get_dimm_info\n EDAC/igen6: Add one Intel Panther Lake-H SoC support\n EDAC/igen6: Fix memory topology parsing for Panther Lake-H SoCs\n EDAC/igen6: Fix call trace due to missing release()\n EDAC/sb_edac: fix grammar in sb_decode_ddr3 warning\n EDAC/i5400: disable error reporting at teardown and refactor helper\n EDAC/i5100: disable error reporting at teardown and create helper\n EDAC/i5000: disable error reporting at teardown and refactor helper\n EDAC/i7300: disable error reporting if init fails and refactor helper\n ...\n" }, { "commit": "80476f22b8b7e193b26f285a7c9f9e4b63abca16", "tree": "c79143284bc04ce1b0fd8ef23f8acc95cff7c2b2", "parents": [ "25a01b5155d207e72bdd31b138406f37788403cb", "61c19a9feb1d87156e46e38d7759f3ad23710e24" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:18:04 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:18:04 2026 +0530" }, "message": "Merge tag \u0027arm64-upstream\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/arm64/linux\n\nPull arm64 updates from Will Deacon:\n \"It feels like the new world of AI tooling has slowed us down a little\n on the feature side when compared to the fixes side. The extra rounds\n of Sashiko review have also pushed a few things out until next time.\n\n Still, there\u0027s some good foundational stuff here for the fpsimd code\n and hardening work towards removing the predictable linear alias of\n the kernel image.\n\n CPU errata handling:\n - Extend CnP disabling workaround to HiSilicon HIP09 hardware.\n - Work around eternally broken broadcast TLB invalidation on more\n CPUs.\n - Documentation and code cleanups.\n\n CPU features:\n - Add new hwcaps for the 2025 dpISA extensions.\n\n Floating point / SVE / SME:\n - Significant cleanup to the low-level state management code in the\n core architecture code and KVM.\n - Use correct register widths during SVE/SME save/restore assembly.\n - Expose SVE/SME save/restore memory accesses to sanitisers.\n\n Memory management:\n - Preparatory work for unmapping the kernel data and bss sections\n from the linear map.\n\n Miscellaneous:\n - Inline DAIF manipulation helpers so they can be used safely from\n non-instrumentable code.\n - Fix handling of the \u0027nosmp\u0027 cmdline option to avoid marking\n secondary cores as \"possible\".\n\n MPAM:\n - Add support for v0.1 of the MPAM architecture.\n\n Perf:\n - Update HiSilicon PMU MAINTAINERS entry.\n - Fix event encodings for the DVM node in the CMN driver.\n\n Selftests:\n - Extend sigframe tests to cover POE context.\n - Add coverage for the newly added 2025 dpISA hwcaps.\n\n System registers:\n - Add new registers and ESR encodings for the HDBSS feature.\n\n Plus minor fixes and cleanups across the board\"\n\n* tag \u0027arm64-upstream\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/arm64/linux: (73 commits)\n arm64: errata: Mitigate TLBI errata on Microsoft Azure Cobalt 100 CPU\n arm64: errata: Mitigate TLBI errata on NVIDIA Olympus CPU\n arm64: errata: Mitigate TLBI errata on various Arm CPUs\n arm64: cputype: Add C1-Premium definitions\n arm64: cputype: Add C1-Ultra definitions\n Revert \"arm64: mm: Unmap kernel data/bss entirely from the linear map\"\n Revert \"arm64: mm: Defer remap of linear alias of data/bss\"\n arm64: arch_timer: reuse arch_timer_read_cnt{p,v}ct_el0() helpers\n arm64/mm: Rename ptdesc_t\n arm64: mm: Defer remap of linear alias of data/bss\n KVM: arm64: Omit tag sync on stage-2 mappings of the zero page\n arm64: Avoid double evaluation of __ptep_get()\n kasan: Move generic KASAN page tables out of BSS too\n arm64: Rename page table BSS section to .bss..pgtbl\n arm64: patching: replace min_t with min in __text_poke\n perf/arm-cmn: Fix DVM node events\n arm64: fpsimd: Remove \u003casm/fpsimdmacros.h\u003e\n arm64: fpsimd: Move SME save/restore inline\n arm64: fpsimd: Move sve_flush_live() inline\n arm64: fpsimd: Move SVE save/restore inline\n ...\n" }, { "commit": "25a01b5155d207e72bdd31b138406f37788403cb", "tree": "d4b2c03475c65c0ef1ee88acfd671581ea96bbc4", "parents": [ "44308fbe8feb0861053b9173e3fda2849944b355", "37540b8c287fc817bdbd0c62bb75ad6eab0e5d03" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:08:13 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:08:13 2026 +0530" }, "message": "Merge tag \u0027s390-7.2-1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/s390/linux\n\nPull s390 updates from Alexander Gordeev:\n\n - Use CIO device online variable instead of the internal FSM state to\n determine device availability during purge operations\n\n - Remove extra check of task_stack_page() because try_get_task_stack()\n already takes care of that when reading /proc/\u003cpid\u003e/wchan\n\n - Allow user-space to use the new SCLP action qualifier 4 for to\n provide NVMe SMART log data to the platform.\n\n - Send AP CHANGE uevents on successful bind and successful association\n to notify user-space about SE operations on AP queue devices\n\n - Add an s390dbf kernel parameter to configure debug log levels and\n area sizes during early boot\n\n - On arm64 the empty zero page is going to be mapped read-only. Do the\n same for s390 with an explicit set_memory_ro() call\n\n - Improve s390-specific bcr_serialize() and cpu_relax() implementations\n\n - Remove all unused variables to avoid allmodconfig W\u003d1 build fails\n with latest clang-23\n\n - Cleanup default Kconfig values for s390 selftests\n\n - Add a s390-tod trace clock to allow comparing trace timestamps\n between different systems or virtual machines on s390\n\n - Remove the s390 implementation of strlcat() in favor of the generic\n variant\n\n - Make consistent the calling order between\n page_table_check_pte_clear() and secure page conversion across all\n code paths\n\n - Rearrange some fields within AP and zcrypt structs to reduce memory\n consumption and unused holes\n\n - Shorten GR_NUM and VX_NUM macros and move them to a separate header\n\n - Replace __get_free_page() with kmalloc() in few sources\n\n - Introduce an infrastructure for more efficient this_cpu operations.\n Eliminate conditional branches when PREEMPT_NONE is removed\n\n - Enable Rust support\n\n - Use z10 as minimum architecture level, similar to the boot code, to\n enforce a defined architecture level set\n\n - Improve and convert various mem*() helper functions to C. For that\n add .noinstr.text section to avoid orphaned warnings from the linker\n\n - Fix the function pointer type in __ret_from_fork() to correct the\n indirect call to match kernel thread return type of int\n\n - Revert support for DCACHE_WORD_ACCESS to avoid an endless exception\n loop on read from donated Ultravisor pages at unaligned addresses\n\n* tag \u0027s390-7.2-1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/s390/linux: (52 commits)\n s390: Revert support for DCACHE_WORD_ACCESS\n s390/process: Fix kernel thread function pointer type\n s390/tishift: Convert __ashlti3(), __ashrti3(), __lshrti3() to C\n s390/memmove: Optimize backward copy case\n s390/string: Convert memset(16|32|64)() to C\n s390/string: Convert memcpy() to C\n s390/string: Convert memset() to C\n s390/string: Convert memmove() to C\n s390/string: Add -ffreestanding compile option to string.o\n s390: Add .noinstr.text to boot and purgatory linker scripts\n s390/purgatory: Enforce z10 minimum architecture level\n s390: Enable Rust support\n s390/cmpxchg: Fix KASAN stack-out-of-bounds in atomic helpers\n rust: helpers: Add memchr wrapper for string operations\n rust/bindgen_parameters: Mark s390 types as opaque to prevent repr conflicts\n s390/jump_label: Implement ARCH_STATIC_BRANCH_JUMP_ASM and ARCH_STATIC_BRANCH_ASM macros\n s390/bug: Provide ARCH_WARN_ASM for Rust WARN/BUG support\n s390/ap: Fix locking issue in SE bind and associate sysfs functions\n s390/percpu: Provide arch_this_cpu_write() implementation\n s390/percpu: Provide arch_this_cpu_read() implementation\n ...\n" }, { "commit": "44308fbe8feb0861053b9173e3fda2849944b355", "tree": "92593b038b3d8c78297af13d3511489df70c6878", "parents": [ "0e0611827f3349d0a2ac121c023a6d3260dcecdb", "dc5200f6b1ada318463dd141b041ec9f044b2bf5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:01:42 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 05:01:42 2026 +0530" }, "message": "Merge tag \u0027m68k-for-v7.2-tag1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/geert/linux-m68k\n\nPull m68k updates from Geert Uytterhoeven:\n - Replace more deprecated functions by safer counterparts\n - Switch Mac NuBus to a dynamic root device\n - defconfig updates\n - Miscellaneous fixes and improvements\n\n* tag \u0027m68k-for-v7.2-tag1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/geert/linux-m68k:\n m68k: Correct CONFIG_MVME16x macro name in #endif comment\n m68k: hash: Use lower_16_bits() helper\n m68k: defconfig: Update defconfigs for v7.1-rc1\n dio: Update DIO_SCMAX comment\n dio: Use tabs and avoid continuation logging in dio_init\n dio: Replace deprecated strcpy with strscpy in dio_init\n nubus: Switch to dynamic root device\n zorro: sysfs: Replace sprintf() by sysfs_emit()\n" }, { "commit": "0e0611827f3349d0a2ac121c023a6d3260dcecdb", "tree": "7a9b915fdc7435a02ac8ac1bbced08b387b5f399", "parents": [ "ff5ccdb8d5bd242f1064c6f7996603e47e28d095", "462bdd08fbdf41db223c6117d907c8fd68d666ea" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 15:53:57 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 15:53:57 2026 +0530" }, "message": "Merge tag \u0027pull-fixes\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/viro/vfs\n\nPull udf fix from Al Viro:\n \"I just noticed that a udf fix had been sitting in #fixes since\n February; still applicable, Jan\u0027s Acked-by applied. Very belated pull\n request\"\n\n* tag \u0027pull-fixes\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/viro/vfs:\n udf: fix nls leak on udf_fill_super() failure\n" }, { "commit": "ff5ccdb8d5bd242f1064c6f7996603e47e28d095", "tree": "f35bbd7c7c7e9871c33d830efae1718fa581c107", "parents": [ "7561361d7655828d50482cd9e80fa3bf73d9c92e", "2bebd986eddb31f9ff1e02e9245a318036280759" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 15:25:17 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 15:25:17 2026 +0530" }, "message": "Merge tag \u0027x86-cpu-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull x86 cpuid updates from Ingo Molnar:\n\n - CPUID API updates (Ahmed S. Darwish):\n - Introduce a centralized CPUID parser\n - Introduce a centralized CPUID data model\n - Introduce \u003casm/cpuid/leaf_types.h\u003e\n - Rename cpuid_leaf()/cpuid_subleaf() APIs\n - treewide: Explicitly include the x86 CPUID headers\n\n - Update to x86-cpuid-db v3.1 (Maciej Wieczor-Retman)\n\n - Continued removal of pre-i586 support and related simplifications\n (Ingo Molnar)\n\n - Add Intel CPU model number for rugged Panther Lake (Tony Luck)\n\n - Misc fixes, updates and cleanups by Arnd Bergmann, Chao Gao, Lukas\n Bulwahn, Sohil Mehta, Maciej Wieczor-Retman.\n\n* tag \u0027x86-cpu-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (25 commits)\n x86/cpu: Make CONFIG_X86_CX8 unconditional\n x86/cpu: Remove unused !CONFIG_X86_TSC code\n x86/cpuid: Update bitfields to x86-cpuid-db v3.1\n tools/x86/kcpuid: Update bitfields to x86-cpuid-db v3.1\n x86/cpu: Make CONFIG_X86_TSC unconditional\n MAINTAINERS: Drop obsolete FPU EMULATOR section\n x86/cpu: Fix a F00F bug warning and clean up surrounding code\n x86/cpu: Add Intel CPU model number for rugged Panther Lake\n x86/cpuid: Introduce a centralized CPUID parser\n x86/cpu: Introduce a centralized CPUID data model\n x86/cpuid: Introduce \u003casm/cpuid/leaf_types.h\u003e\n x86/cpuid: Rename cpuid_leaf()/cpuid_subleaf() APIs\n x86/cpu: Do not include the CPUID API header in asm/processor.h\n Documentation: core-api/cpu_hotplug: Remove stale cpu0_hotplug docs\n x86/cpu, cpufreq: Remove AMD ELAN support\n x86/fpu: Remove the math-emu/ FPU emulation library\n x86/fpu: Remove the \u0027no387\u0027 boot option\n x86/fpu: Remove MATH_EMULATION and related glue code\n treewide: Explicitly include the x86 CPUID headers\n x86/cpu: Remove the CONFIG_X86_INVD_BUG quirk\n ...\n" }, { "commit": "7561361d7655828d50482cd9e80fa3bf73d9c92e", "tree": "c302cc2596cfd14478ba87e32a993020bdfcc9ae", "parents": [ "2cbf335f8ccc7a6418159858dc03e36df8e3e5cf", "b5884070f9da9ffecd5141b5811cfdbaa274809a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 15:08:14 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 15:08:14 2026 +0530" }, "message": "Merge tag \u0027x86-msr-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull x86/msr updates from Ingo Molnar:\n\n - Large series to reorganize the rdmsr/wrmsr APIs to remove\n 32-bit variants and convert to 64-bit variants (Juergen Gross)\n\n - Fix W\u003d1 warning (HyeongJun An)\n\n* tag \u0027x86-msr-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n x86/msr: Remove wrmsrl()\n x86/msr: Switch wrmsrl() users to wrmsrq()\n x86/msr: Remove rdmsrl()\n x86/msr: Switch rdmsrl() users to rdmsrq()\n x86/msr: Remove wrmsr_safe_on_cpu()\n x86/msr: Switch wrmsr_safe_on_cpu() users to wrmsrq_safe_on_cpu()\n x86/msr: Remove rdmsr_safe_on_cpu()\n x86/msr: Switch rdmsr_safe_on_cpu() users to rdmsrq_safe_on_cpu()\n x86/msr: Don\u0027t use rdmsr_safe_on_cpu() in rdmsrq_safe_on_cpu()\n x86/msr: Remove wrmsr_on_cpu()\n x86/msr: Switch wrmsr_on_cpu() users to wrmsrq_on_cpu()\n x86/msr: Remove rdmsr_on_cpu()\n x86/msr: Switch rdmsr_on_cpu() users to rdmsrq_on_cpu()\n x86/msr: Remove rdmsrl_on_cpu()\n x86/msr: Switch rdmsrl_on_cpu() user to rdmsrq_on_cpu()\n x86/process: Convert rdmsr() to rdmsrq() in arch_post_acpi_subsys_init() to address W\u003d1 warning\n" }, { "commit": "2cbf335f8ccc7a6418159858dc03e36df8e3e5cf", "tree": "1ff36263ce50da08cf0f078d398f3f398ee5a015", "parents": [ "0bcc2dc22f38e57fa97d8238b2e0bcdde5376f33", "c095741713d1bc317b53e2da2b222e7448b6021f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 14:50:18 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 14:50:18 2026 +0530" }, "message": "Merge tag \u0027sched-core-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull scheduler updates from Ingo Molnar:\n \"SMP load-balancing updates:\n\n - A large series to introduce infrastructure for cache-aware load\n balancing, with the goal of co-locating tasks that share data\n within the same Last Level Cache (LLC) domain. By improving cache\n locality, the scheduler can reduce cache bouncing and cache misses,\n ultimately improving data access efficiency.\n\n Implemented by Chen Yu and Tim Chen, based on early prototype work\n by Peter Zijlstra, with fixes by Jianyong Wu, Peter Zijlstra and\n Shrikanth Hegde.\n\n - A series to simplify CONFIG_SCHED_SMT ifdef usage (Shrikanth Hegde)\n\n Fair scheduler updates:\n\n - A series to improve SD_ASYM_CPUCAPACITY scheduling by introducing\n SMT awareness (Andrea Righi, K Prateek Nayak)\n\n - A series to optimize cfs_rq and sched_entity allocation for better\n data locality (Zecheng Li)\n\n - A preparatory series to change fair/cgroup scheduling to a single\n runqueue, without the final change (Peter Zijlstra)\n\n - Auto-manage ext/fair dl_server bandwidth (Andrea Righi)\n\n - Fix cpu_util runnable_avg arithmetic (Hongyan Xia)\n\n - Optimize update_tg_load_avg()\u0027s rate-limiting code (Rik van Riel)\n\n - Allow account_cfs_rq_runtime() to throttle current hierarchy\n (K Prateek Nayak)\n\n - Update util_est after updating util_avg during dequeue, to fix the\n util signal update logic, which reduces signal noise (Vincent\n Guittot)\n\n Scheduler topology updates:\n\n - Allow multiple domains to claim sched_domain_shared (K Prateek\n Nayak)\n\n - Add parameter to split LLC (Peter Zijlstra)\n\n Core scheduler updates:\n\n - Use trace_call__\u003ctp\u003e() to save a static branch (Gabriele Monaco)\n\n Scheduler statistics updates:\n\n - Drop now-stale mul_u64_u64_div_u64() cputime over-approximation\n guard (Nicolas Pitre)\n\n Deadline scheduler updates:\n\n - Reject debugfs dl_server writes for offline CPUs (Andrea Righi)\n\n - Fix replenishment logic for non-deferred servers (Yuri Andriaccio)\n\n RT scheduling updates:\n\n - Turn RT_PUSH_IPI default off for non PREEMPT_RT (Steven Rostedt)\n\n - Update default bandwidth for real-time tasks to 1.0 (Yuri\n Andriaccio)\n\n Proxy scheduling updates:\n\n - A series to implement Optimized Donor Migration for Proxy Execution\n (John Stultz, Peter Zijlstra)\n\n - Various proxy scheduling cleanups and fixes (Peter Zijlstra,\n K Prateek Nayak)\n\n Misc fixes, improvements and cleanups by Aaron Lu, Andrea Righi,\n Zenghui Yu, Chen Yu, Guanyou.Chen, John Stultz, Shrikanth Hegde,\n Peter Zijlstra, Liang Luo and Yiyang Chen\"\n\n* tag \u0027sched-core-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (91 commits)\n sched/fair: Fix newidle vs core-sched\n sched/deadline: Use task_on_rq_migrating() helper\n sched/core: Combine separate \u0027else\u0027 and \u0027if\u0027 statements\n sched/fair: Fix cpu_util runnable_avg arithmetic\n sched/fair: Unify cfs_rq throttling via account_cfs_rq_runtime()\n sched/fair: Move the throttled tasks to a local list in tg_unthrottle_up()\n sched/fair: Call update_curr() before unthrottling the hierarchy\n sched/fair: Use throttled_csd_list for local unthrottle\n sched/fair: Convert cfs bandwidth throttling to use guards\n sched/fair: Allocate cfs_tg_state with percpu allocator\n sched/fair: Remove task_group-\u003ese pointer array\n sched/fair: Co-locate cfs_rq and sched_entity in cfs_tg_state\n sched: restore timer_slack_ns when resetting RT policy on fork\n MAINTAINERS: Fix spelling mistake in Peter\u0027s name\n sched: Simplify ttwu_runnable()\n sched/proxy: Remove superfluous clear_task_blocked_in()\n sched/proxy: Remove PROXY_WAKING\n sched/proxy: Switch proxy to use p-\u003eis_blocked\n sched/proxy: Only return migrate when needed\n sched: Be more strict about p-\u003eis_blocked\n ...\n" }, { "commit": "0bcc2dc22f38e57fa97d8238b2e0bcdde5376f33", "tree": "53d049524fccd4bbcb94f2bae45a60d32ee8e342", "parents": [ "d8d706a27451c015490d23b4f4764de05e553624", "67d27727854def4a7e2b386429941f5c4741ccc4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 14:45:41 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 14:45:41 2026 +0530" }, "message": "Merge tag \u0027perf-core-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull performance events updates from Ingo Molnar:\n \"Core perf code updates:\n\n - Reveal PMU type in fdinfo (Chun-Tse Shao)\n\n Intel CPU PMU driver updates:\n\n - Fix various inaccurate hard-coded event configurations (Dapeng Mi)\n\n Intel uncore PMU driver updates (Zide Chen):\n\n - Fix discovery unit lookup bug for multi-die systems\n\n - Guard against invalid box control address\n\n - Fix PCI device refcount leak in UPI discovery\n\n - Defer ADL global PMON enable to enable_box() to save power\n\n - Fix uncore_die_to_cpu() for offline dies\n\n - Implement global init callback for GNR uncore\n\n AMD CPU PMU driver updates:\n\n - Always use the NMI latency mitigation (Sandipan Das)\n\n AMD uncore PMU driver updates:\n\n - Use Node ID to identify DF and UMC domains (Sandipan Das)\"\n\n* tag \u0027perf-core-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (22 commits)\n perf/x86/amd/uncore: Use Node ID to identify DF and UMC domains\n perf: Reveal PMU type in fdinfo\n perf/x86/intel/uncore: Implement global init callback for GNR uncore\n perf/x86/intel/uncore: Fix uncore_die_to_cpu() for offline dies\n perf/x86/intel/uncore: Move die_to_cpu() to uncore.c\n perf/x86/intel/uncore: Defer ADL global PMON enable to enable_box()\n perf/x86/intel/uncore: Fix PCI device refcount leak in UPI discovery\n perf/x86/intel/uncore: Guard against invalid box control address\n perf/x86/intel/uncore: Fix discovery unit lookup for multi-die systems\n perf/x86/amd/core: Always use the NMI latency mitigation\n perf/x86/intel: Update event constraints and cache_extra_regsfor CWF\n perf/x86/intel: Update event constraints and cache_extra_regsfor SRF\n perf/x86/intel: Update event constraints and cache_extra_regsfor NVL\n perf/x86/intel: Update event constraints for PTL\n perf/x86/intel: Update event constraints and cache_extra_regsfor ARL\n perf/x86/intel: Update event constraints and cache_extra_regsfor LNL\n perf/x86/intel: Update event constraints and cache_extra_regsfor MTL\n perf/x86/intel: Update event constraints and cache_extra_regsfor ADL\n perf/x86/intel: Update event constraints for DMR\n perf/x86/intel: Update event constraints and cache_extra_regsfor SPR\n ...\n" }, { "commit": "d8d706a27451c015490d23b4f4764de05e553624", "tree": "619a8675b8b58982f2b9ea5a4481a89f035cff8e", "parents": [ "764e77d868a5b932c709e20ddb5993f9111a841c", "2d3bb398861ad3ddbf87959cc2e34a7001f0b3ad" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 14:31:58 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 14:31:58 2026 +0530" }, "message": "Merge tag \u0027objtool-core-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull objtool updates from Ingo Molnar:\n\n - A large series of KLP fixes and improvements, in preparation of the\n arm64 port (Josh Poimboeuf)\n\n - Fix a number of bugs and issues on specific distro, LTO, FineIBT and\n kCFI configs (Josh Poimboeuf)\n\n - Misc other fixes by Josh Poimboeuf and Joe Lawrence\n\n* tag \u0027objtool-core-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (53 commits)\n objtool/klp: Cache dont_correlate() result\n objtool: Improve and simplify prefix symbol detection\n objtool/klp: Fix kCFI prefix finding/cloning\n objtool: Grow __cfi_* prefix symbols for all CFI+CALL_PADDING\n objtool/klp: Fix position-dependent checksums for non-relocated jumps/calls\n objtool: Add insn_sym() helper\n objtool/klp: Add correlation debugging output\n objtool/klp: Rewrite symbol correlation algorithm\n objtool/klp: Calculate object checksums\n klp-build: Validate short-circuit prerequisites\n objtool/klp: Remove \"objtool --checksum\"\n klp-build: Use \"objtool klp checksum\" subcommand\n objtool/klp: Add \"objtool klp checksum\" subcommand\n objtool: Consolidate file decoding into decode_file()\n objtool/klp: Extricate checksum calculation from validate_branch()\n objtool: Add is_cold_func() helper\n objtool: Add is_alias_sym() helper\n objtool/klp: Handle Clang .data..Lanon anonymous data sections\n objtool/klp: Create empty checksum sections for function-less object files\n objtool: Include libsubcmd headers directly from source tree\n ...\n" }, { "commit": "764e77d868a5b932c709e20ddb5993f9111a841c", "tree": "e3285615d0abbafba6ca3559d6fbd20bf31c4589", "parents": [ "186d3c4e92242351afc24d9784f31cb4cd08a4b7", "4f070ccb4dc4692e3b6757819fb80655f58b4f12" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 14:21:14 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 14:21:14 2026 +0530" }, "message": "Merge tag \u0027locking-core-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull locking updates from Ingo Molnar:\n \"Futex updates:\n\n - Optimize futex hash bucket access patterns (Peter Zijlstra)\n\n - Large series to address the robust futex unlock race for real, by\n Thomas Gleixner:\n\n \"The robust futex unlock mechanism is racy in respect to the\n clearing of the robust_list_head::list_op_pending pointer because\n unlock and clearing the pointer are not atomic.\n\n The race window is between the unlock and clearing the pending op\n pointer. If the task is forced to exit in this window, exit will\n access a potentially invalid pending op pointer when cleaning up\n the robust list.\n\n That happens if another task manages to unmap the object\n containing the lock before the cleanup, which results in an UAF.\n\n In the worst case this UAF can lead to memory corruption when\n unrelated content has been mapped to the same address by the time\n the access happens.\n\n User space can\u0027t solve this problem without help from the kernel.\n This series provides the kernel side infrastructure to help it\n along:\n\n 1) Combined unlock, pointer clearing, wake-up for the\n contended case\n\n 2) VDSO based unlock and pointer clearing helpers with a\n fix-up function in the kernel when user space was interrupted\n within the critical section.\n\n ... with help by André Almeida:\n\n - Add a note about robust list race condition (André Almeida)\n - Add self-tests for robust release operations (André Almeida)\n\n Context analysis updates:\n\n - Implement context analysis for \u0027struct rt_mutex\u0027. (Bart Van Assche)\n - Bump required Clang version to 23 (Marco Elver)\n\n Guard infrastructure updates:\n\n - Series to remove NULL check from unconditional guards (Dmitry\n Ilvokhin)\n\n Lockdep updates:\n\n - Restore self-test migrate_disable() and sched_rt_mutex state on\n PREEMPT_RT (Karl Mehltretter)\n\n Membarriers updates:\n\n - Use per-CPU mutexes for targeted commands (Aniket Gattani)\n - Modernize membarrier_global_expedited with cleanup guards (Aniket\n Gattani)\n - Add rseq stress test for CFS throttle interactions (Aniket Gattani)\n\n percpu-rwsems updates:\n\n - Extract __percpu_up_read() to optimize inlining overhead (Dmitry\n Ilvokhin)\n\n Seqlocks updates:\n\n - Allow UBSAN_ALIGNMENT to fail optimizing (Heiko Carstens)\n\n Lock tracing:\n\n - Add contended_release tracepoint to sleepable locks such as\n mutexes, percpu-rwsems, rtmutexes, rwsems and semaphores (Dmitry\n Ilvokhin)\n\n MAINTAINERS updates:\n\n - MAINTAINERS: Add RUST [SYNC] entry (Boqun Feng)\n\n Misc updates and fixes by Randy Dunlap, YE WEI-HONG, Fabricio Parra,\n Dmitry Ilvokhin and Peter Zijlstra\"\n\n* tag \u0027locking-core-2026-06-14\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (36 commits)\n locking: Add contended_release tracepoint to sleepable locks\n locking/percpu-rwsem: Extract __percpu_up_read()\n tracing/lock: Remove unnecessary linux/sched.h include\n futex: Optimize futex hash bucket access patterns\n rust: sync: completion: Mark inline complete_all and wait_for_completion\n MAINTAINERS: Add RUST [SYNC] entry\n cleanup: Specify nonnull argument index\n selftests: futex: Add tests for robust release operations\n Documentation: futex: Add a note about robust list race condition\n x86/vdso: Implement __vdso_futex_robust_try_unlock()\n x86/vdso: Prepare for robust futex unlock support\n futex: Provide infrastructure to plug the non contended robust futex unlock race\n futex: Add robust futex unlock IP range\n futex: Add support for unlocking robust futexes\n futex: Cleanup UAPI defines\n x86: Select ARCH_MEMORY_ORDER_TSO\n uaccess: Provide unsafe_atomic_store_release_user()\n futex: Provide UABI defines for robust list entry modifiers\n futex: Move futex related mm_struct data into a struct\n futex: Make futex_mm_init() void\n ...\n" }, { "commit": "186d3c4e92242351afc24d9784f31cb4cd08a4b7", "tree": "bdc76b0f579cd233c77cc14d5c3e2b0853535d4a", "parents": [ "2d6d57f889f3a5e7d19009c560ea2002cdde9fb8", "8d563bd7904734c05a4f2abf4ecca0e4fe764b50" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:57:13 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:57:13 2026 +0530" }, "message": "Merge tag \u0027timers-vdso-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull vdso updates from Thomas Gleixner:\n\n - Remove the redundant CONFIG_GENERIC_TIME_VSYSCALL after converting\n the remaining users over.\n\n - Rework and sanitize the MIPS VDSO handling, so it does not handle the\n time related VDSO if there is no VDSO capable clocksource available.\n Also stop mapping VDSO data pages unconditionally even if there is no\n usage possible.\n\n* tag \u0027timers-vdso-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n MIPS: VDSO: Fold MIPS_CLOCK_VSYSCALL into MIPS_GENERIC_GETTIMEOFDAY\n MIPS: VDSO: Gate microMIPS restriction on GCC version\n MIPS: VDSO: Fold MIPS_DISABLE_VDSO into MIPS_GENERIC_GETTIMEOFDAY\n clocksource/drivers/mips-gic-timer: Only use VDSO_CLOCKMODE_GIC when it is a available\n MIPS: csrc-r4k: Only use VDSO_CLOCKMODE_R4K when it is a available\n MIPS: VDSO: Only map the data pages when the vDSO is used\n MIPS: Introduce Kconfig MIPS_GENERIC_GETTIMEOFDAY\n vdso/datastore: Always provide symbol declarations\n MAINTAINERS: Add include/linux/vdso_datastore.h to vDSO block\n vdso/gettimeofday: Rename __arch_get_vdso_u_timens_data()\n vdso/treewide: Drop GENERIC_TIME_VSYSCALL\n vdso/vsyscall: Gate update_vsyscall() behind CONFIG_GENERIC_GETTIMEOFDAY\n riscv: vdso: Drop CONFIG_GENERIC_TIME_VSYSCALL guard around syscall fallbacks\n" }, { "commit": "2d6d57f889f3a5e7d19009c560ea2002cdde9fb8", "tree": "aa1e373320daae5eaf98d0a58a2ae65858087cd8", "parents": [ "a53fcff8fc7530f59a8171824ed586200df724a0", "bc484a5096732cd858771cccd3164ec985bdc03d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:51:27 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:51:27 2026 +0530" }, "message": "Merge tag \u0027timers-ptp-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull timekeeping updates from Thomas Gleixner:\n \"Updates for NTP/timekeeping and PTP:\n\n - Expand timekeeping snapshot mechanisms\n\n The various snapshot functions are mostly used for PTP to collect\n \"atomic\" snapshots of various involved clocks.\n\n They lack support for the recently introduced AUX clocks and do not\n provide the underlying counter value (e.g. TSC) to user space.\n Exposing the counter value snapshot allows for better control and\n steering.\n\n Convert the hard wired ktime_get_snapshot() to take a clock ID,\n which allows the caller to select the clock ID to be captured along\n with CLOCK_MONONOTONIC_RAW. Additionally capture the underlying\n hardware counter value and the clock source ID of the counter.\n\n Expand the hardware based snapshot capture where devices provide a\n mechanism to snapshot the hardware PTP clock and the system counter\n (usually via PCI/PTM) to support AUX clocks and also provide the\n captured counter value back to the caller and not only the clock\n timestamps derived from it.\n\n - Add a new optional read_snapshot() callback to clocksources\n\n That is required to capture atomic snapshots from clocksources\n which are derived from TSC with a scaling mechanism (e.g. Hyper-V,\n KVMclock).\n\n The value pair is handed back in the snapshot structure to the\n callers, so they can do the necessary correlations in a more\n precise way.\n\n This touches usage sites of the affected functions and data structure\n all over the tree, but stays fully backwards compatible for the\n existing user space exposed interfaces. New PTP IOCTLs will provide\n access to the extended functionality in later kernel versions\"\n\n* tag \u0027timers-ptp-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (28 commits)\n ptp: vmclock: Use hw_cycles from snapshot for precise TSC pairing\n x86/kvmclock: Implement read_snapshot() for kvmclock clocksource\n clocksource/hyperv: Implement read_snapshot() for TSC page clocksource\n timekeeping: Add clocksource read_snapshot() method and hw_cycles to snapshot\n ptp: Switch to ktime_get_snapshot_id() for pre/post timestamps\n timekeeping: Add support for AUX clock cross timestamping\n timekeeping: Remove system_device_crosststamp::sys_realtime\n ALSA: hda/common: Use system_device_crosststamp::sys_systime\n wifi: iwlwifi: Use system_device_crosststamp::sys_systime\n ptp: Use system_device_crosststamp::sys_systime\n timekeeping: Prepare for cross timestamps on arbitrary clock IDs\n timekeeping: Remove ktime_get_snapshot()\n virtio_rtc: Use provided clock ID for history snapshot\n net/mlx5: Use provided clock ID for history snapshot\n igc: Use provided clock ID for history snapshot\n ice/ptp: Use provided clock ID for history snapshot\n wifi: iwlwifi: Adopt PTP cross timestamps to core changes\n timekeeping: Add CLOCK ID to system_device_crosststamp\n timekeeping: Add system_counterval_t to struct system_device_crosststamp\n timekeeping: Add CLOCK_AUX support for ktime_get_snapshot_id()\n ...\n" }, { "commit": "a53fcff8fc7530f59a8171824ed586200df724a0", "tree": "d8ebac39633a9cfd9cebbf8de7f6e0df3aab61ee", "parents": [ "a60ce761d99ff2d9eefe33374c5f20726465a140", "6199f9999a9b62b2b84a1bf5b52a9fd0bb8de5af" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:48:52 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:48:52 2026 +0530" }, "message": "Merge tag \u0027timers-nohz-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull NOHZ updates from Thomas Gleixner:\n\n - Fix a long standing TOCTOU in get_cpu_sleep_time_us()\n\n - Make the CPU offline NOHZ handling more robust by disabling NOHZ on\n the outgoing CPU early instead of creating unneeded state which needs\n to be undone.\n\n - Unify idle CPU time accounting instead of having two different\n accounting mechanisms. These two different mechanisms are not really\n independent, but the different properties can in the worst case cause\n that gloabl idle time can be observed going backwards.\n\n - Consolidate the idle/iowait time retrieval interfaces instead of\n converting back and forth between them.\n\n - Make idle interrupt time accounting more robust. The original code\n assumes that interrupt time accouting is enabled and therefore stops\n elapsing idle time while an interrupt is handled in NOHZ dyntick\n state. That assumption is not correct as interrupt time accounting\n can be disabled at compile and runtime.\n\n - Fix an accounting error between dyntick idle time and dyntick idle\n steal time. The stolen time is not accounted and therefore idle time\n becomes inaccurate. The stolen time is now accounted after the fact\n as there is no way to predict the steal time upfront.\n\n* tag \u0027timers-nohz-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n sched/cputime: Handle dyntick-idle steal time correctly\n sched/cputime: Handle idle irqtime gracefully\n sched/cputime: Provide get_cpu_[idle|iowait]_time_us() off-case\n tick/sched: Consolidate idle time fetching APIs\n tick/sched: Account tickless idle cputime only when tick is stopped\n tick/sched: Remove unused fields\n tick/sched: Move dyntick-idle cputime accounting to cputime code\n tick/sched: Remove nohz disabled special case in cputime fetch\n tick/sched: Unify idle cputime accounting\n s390/time: Prepare to stop elapsing in dynticks-idle\n powerpc/time: Prepare to stop elapsing in dynticks-idle\n sched/cputime: Correctly support generic vtime idle time\n sched/cputime: Remove superfluous and error prone kcpustat_field() parameter\n sched/idle: Handle offlining first in idle loop\n tick/sched: Fix TOCTOU in nohz idle time fetch\n" }, { "commit": "a60ce761d99ff2d9eefe33374c5f20726465a140", "tree": "a7883dcce89453fe59d7cf614620e0b20bea3895", "parents": [ "f20e2fdaaeb74330a6c5d65af22a8c47409a7a91", "87bd2ad568e15b90d5f7d4bcd70342d05dad649c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:39:12 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:39:12 2026 +0530" }, "message": "Merge tag \u0027timers-core-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull timer core updates from Thomas Gleixner:\n \"Updates for the time/timer core subsystem:\n\n - Harden the user space controllable hrtimer interfaces further to\n protect against unpriviledged DoS attempts by arming timers in the\n past.\n\n - Add per-capacity hierarchies to the timer migration code to prevent\n timer migration accross different capacity domains. This code has\n been disabled last minute as there is a pathological problem with\n SoCs which advertise a larger number of capacity domains. The\n problem is under investigation and the code won\u0027t be active before\n v7.3, but that turned out to be less intrusive than a full revert\n as it preserves the preparatory steps and allows people to work on\n the final resolution\n\n - Export time namespace functionality as a recent user can be built\n as a module.\n\n - Initialize the jiffies clocksource before using it. The recent\n hardening against time moving backward requires that the related\n members of struct clocksource have been initialized, otherwise it\n clamps the readout to 0, which makes time stand sill and causes\n boot delays.\n\n - Fix a more than twenty year old PID reference count leak in an\n error path of the POSIX CPU timer code.\n\n - The usual small fixes, improvements and cleanups all over the\n place\"\n\n* tag \u0027timers-core-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (31 commits)\n posix-cpu-timers: Fix pid refcount leak in do_cpu_nanosleep() error path\n time/jiffies: Register jiffies clocksource before usage\n timers/migration: Temporarily disable per capacity hierarchies\n timers/migration: Turn tmigr_hierarchy level_list into a flexible array\n timers/migration: Deactivate per-capacity hierarchies under nohz_full\n timers/migration: Fix hotplug migrator selection target on asymetric capacity machines\n ntsync: Honour caller\u0027s time namespace for absolute MONOTONIC timeouts\n time/namespace: Export init_time_ns and do_timens_ktime_to_host()\n timers/migration: Update stale @online doc to @available\n timers: Fix flseep() typo in kernel-doc comment\n hrtimer: Fix the bogus return type of __hrtimer_start_range_ns()\n hrtimer: Return ktime_t from hrtimer_get_next_event()/hrtimer_next_event_without()\n clocksource: Clean up clocksource_update_freq() functions\n alarmtimer: Remove stale return description from alarm_handle_timer()\n selftests/posix_timers: Use CLOCK_THREAD_CPUTIME_ID for ITIMER_PROF measurements\n scripts/timers: Add timer_migration_tree.py\n timers/migration: Handle capacity in connect tracepoints\n timers/migration: Split per-capacity hierarchies\n timers/migration: Track CPUs in a hierarchy\n timers/migration: Abstract out hierarchy to prepare for CPU capacity awareness\n ...\n" }, { "commit": "f20e2fdaaeb74330a6c5d65af22a8c47409a7a91", "tree": "73497c6bd12008e52c96e54d4e1e79032bdda1b5", "parents": [ "9e94480d81b9eb9bd175499636bf622e5d62176d", "c66494c79ede1af529dbf67f9ed6fdbf42e05ef3" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:34:03 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:34:03 2026 +0530" }, "message": "Merge tag \u0027timers-clocksource-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull clocksource updates from Thomas Gleixner:\n \"Updates for clocksource/clockevent drivers:\n\n - Add devm helpers for clocksources, which allows to simplify driver\n teardown and probe failure handling.\n\n - More module conversion work\n\n - Update the support for the ARM EL2 virtual timer including the\n required ACPI changes.\n\n - Add clockevent and clocksource support for the TI Dual Mode Timer\n\n - Fix the support for multiple watchdog instances in the TEGRA186\n driver\n\n - Add D1 timer support to the SUN5I driver\n\n - The usual devicetree updates, cleanups and small fixes all over the\n place\"\n\n* tag \u0027timers-clocksource-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (24 commits)\n clocksource: move NXP timer selection to drivers/clocksource\n clocksource/drivers/timer-tegra186: Reserve and service a kernel watchdog\n clocksource/drivers/timer-tegra186: Register all accessible watchdog timers\n clocksource/drivers/timer-tegra186: Correct num_wdts for Tegra186 and Tegra234\n clocksource/drivers/timer-tegra186: Fix support for multiple watchdog instances\n clocksource/drivers/timer-ti-dm: Add clockevent support\n clocksource/drivers/timer-ti-dm: Add clocksource support\n clocksource/drivers/timer-ti-dm: Fix property name in comment\n dt-bindings: timer: arm,arch_timer: Fix requirements for interrupt description\n clocksource/drivers/arm_arch_timer: Default to EL2 virtual timer when running VHE\n ACPI: GTDT: Parse information related to the EL2 virtual timer\n ACPI: GTDT: Account for GTDTv3 size when walking the platform timer descriptors\n clocksource: Add devm_clocksource_register_*() helpers\n clocksource/drivers/sun5i: Add D1 hstimer support\n dt-bindings: timer: allwinner,sun5i-a13-hstimer: add H616 and D1\n dt-bindings: timer: Add StarFive JHB100 clint\n dt-bindings: timer: renesas,rz-mtu3: document RZ/{T2H,N2H}\n dt-bindings: timer: renesas,rz-mtu3: Remove TCIU8 interrupt\n dt-bindings: timer: Remove sifive,fine-ctr-bits property\n clocksource/drivers/timer-of: Make the code compatible with modules\n ...\n" }, { "commit": "9e94480d81b9eb9bd175499636bf622e5d62176d", "tree": "3f94c6fd5549be84a5bc3b47669ab95d7200ed72", "parents": [ "8f45c6ce4959edee1ed25131fc14ce8bd261ca35", "9c91efd1d63e995bb0577fc1c5e909613dfe671d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:30:04 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:30:04 2026 +0530" }, "message": "Merge tag \u0027smp-core-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull smp core updates from Thomas Gleixner:\n \"Two small updates to the SMP/hotplug subsystem:\n\n - Add cpuhplock.h to the maintained files\n\n - Provide the missing stubs for lockdep_is_cpus_held() and\n lockdep_is_cpus_write_held() so the usage sites can be simplified\"\n\n* tag \u0027smp-core-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n cpu: Add lockdep_is_cpus_held()/lockdep_is_cpus_write_held() stubs for !CONFIG_HOTPLUG_CPU\n MAINTAINERS: Add include/linux/cpuhplock.h to CPU HOTPLUG area\n" }, { "commit": "8f45c6ce4959edee1ed25131fc14ce8bd261ca35", "tree": "de0d3563552cc2ca7c799fef42f5bc892b91c643", "parents": [ "857ae5a4459c600d70b9ad64c46a730c428770e2", "3661d5f403769b872482b584e658b71bbdb5f55e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:28:24 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:28:24 2026 +0530" }, "message": "Merge tag \u0027irq-msi-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull MSI irq update from Thomas Gleixner:\n \"A trivial update to the MSI interrupt subsystem, which fixes a couple\n of typos\"\n\n* tag \u0027irq-msi-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n genirq/msi: Fix typos in msi_domain_ops comment\n" }, { "commit": "857ae5a4459c600d70b9ad64c46a730c428770e2", "tree": "11efcb1ace235bb7e30e62e8524ef52df2c4dae4", "parents": [ "13e1a6d6a17eb4bca350e5bf59a89a3056c834ca", "a1a35c09241f0577cc40f65d7372fed01138619d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:25:32 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:25:32 2026 +0530" }, "message": "Merge tag \u0027irq-drivers-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull interrupt chip driver updates from Thomas Gleixner:\n\n - Replace the support for the AST2700-A0 early silicon with a proper\n driver for the final A2 production silicon\n\n - Rename and rework the StarFive JH8100 interrupt controller for the\n new JHB100 SoC as JH8100 was discontinued before production.\n\n - Add support for Amlogic A9 SoCs to the meson-gpio interrupt\n controller\n\n - Expand the Econet interrupt controller driver to support MIPS 34Kc\n Vectored External Interrupt Controller mode.\n\n - Prevent a NULL pointer dereference in the GICv4 code as the vLPI code\n blindly assumes that the ITS was populated. Add the missing sanity\n check.\n\n - Add support for software triggered and for error interrupts to the\n Renesas RZ/T2H driver.\n\n - Add interrupt redirection support for the loongarch architecture.\n\n - Add multicore support to the Realtek RTL interrupt driver\n\n - The usual updates, enhancements and fixes all over the place\n\n* tag \u0027irq-drivers-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip: (32 commits)\n irqchip/irq-realtek-rtl: Add multicore support\n irqchip/irq-realtek-rtl: Add/simplify register helpers\n irqchip/loongarch-ir: Add IR (interrupt redirection) irqchip support\n irqchip/loongarch-avec: Return IRQ_SET_MASK_OK_DONE when keep affinity\n irqchip/loongarch-avec: Prepare for interrupt redirection support\n Docs/LoongArch: Add advanced extended IRQ model\n irqchip/qcom-pdc: Use FIELD_GET() to extract bank index and bit position\n irqchip/qcom-pdc: Add PDC_VERSION() macro to describe version register fields\n irqchip/qcom-pdc: Tighten ioremap clamp to single DRV region size\n irqchip/qcom-pdc: Split __pdc_enable_intr() into per-version helpers\n irqchip/exynos-combiner: Remove useless spinlock\n irqchip/renesas-rzt2h: Add error interrupts support\n irqchip/renesas-rzt2h: Add software-triggered interrupts support\n irqchip/gic-v4: Don\u0027t advertise VLPIs if no ITS is probed\n irqchip/gic-v3-its: Use FIELD_MODIFY()\n irqchip/econet-en751221: Support MIPS 34Kc VEIC mode\n dt-bindings: interrupt-controller: econet: Add CPU interrupt mapping\n irqchip/meson-gpio: Add support for Amlogic A9 SoCs\n dt-bindings: interrupt-controller: Add support for Amlogic A9 SoCs\n irqchip/meson-gpio: Use the correct register in meson_s4_gpio_irq_set_type()\n ...\n" }, { "commit": "13e1a6d6a17eb4bca350e5bf59a89a3056c834ca", "tree": "3e7fa2d248c7025e483e9b795a8bebeb5adb7d50", "parents": [ "a04c8472b0bc99963283e379f4ca2c775be4949b", "8f727615134abc6382f0ea07b90270d7bdde578f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:19:41 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:19:41 2026 +0530" }, "message": "Merge tag \u0027irq-core-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull interrupt core updates from Thomas Gleixner:\n\n - Rework of /proc/interrupt handling:\n\n /proc/interrupts was subject to micro optimizations for a long time,\n but most of the low hanging fruit was left on the table. This rework\n addresses the major time consuming issues:\n\n - Printing a long series of zeros one by one via a format string\n instead of counting subsequent zeros and emitting a string\n constant.\n\n - Simplify and cache the conditions whether interrupts should be\n printed\n\n - Use a proper iteration over the interrupt descriptor xarray\n instead of walking and testing one by one.\n\n - Provide helper functions for the architecture code to emit the\n architecture specific counters\n\n - Convert the counter structure in x86 to an array, which\n simplifies the output and add mechanisms to suppress unused\n architecture interrupts, which just occupy space for nothing.\n Adopt the new core mechanisms.\n\n This adjusts the gdb scripts related to interrupt counter statistics\n to work with the new mechanisms.\n\n - Prevent a string overflow in the /proc/irq/$N/ directory name\n creation code.\n\n* tag \u0027irq-core-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n x86/irq: Add missing \u0027s\u0027 back to thermal event printout\n genirq/proc: Speed up /proc/interrupts iteration\n genirq/proc: Runtime size the chip name\n genirq: Expose irq_find_desc_at_or_after() in core code\n genirq: Add rcuref count to struct irq_desc\n genirq/proc: Increase default interrupt number precision to four\n genirq: Calculate precision only when required\n genirq: Cache the condition for /proc/interrupts exposure\n genirq/manage: Make NMI cleanup RT safe\n genirq: Expose nr_irqs in core code\n scripts/gdb: Update x86 interrupts to the array based storage\n x86/irq: Move IOAPIC misrouted and PIC/APIC error counts into irq_stats\n x86/irq: Suppress unlikely interrupt stats by default\n x86/irq: Make irqstats array based\n genirq/proc: Utilize irq_desc::tot_count to avoid evaluation\n genirq/proc: Avoid formatting zero counts in /proc/interrupts\n x86/irq: Optimize interrupts decimals printing\n genirq/proc: Size interrupt directory names for 10-digit interrupt numbers\n" }, { "commit": "a04c8472b0bc99963283e379f4ca2c775be4949b", "tree": "19fc2e685f0d8c5ae725b047175583af0b1aafce", "parents": [ "36808d5e983985bbda87e01059cccc071fe3ec8d", "5a0daaff6ed9c1177d58b7367451ef9714dac81d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:14:36 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 13:14:36 2026 +0530" }, "message": "Merge tag \u0027core-rseq-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip\n\nPull rseq update from Thomas Gleixner:\n \"A trivial update for RSEQ selftests to provide the config fragments\n which contain the config options required to actually run the tests\"\n\n* tag \u0027core-rseq-2026-06-13\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/tip/tip:\n selftests/rseq: Add config fragment\n" }, { "commit": "36808d5e983985bbda87e01059cccc071fe3ec8d", "tree": "fe89e03319a2eeacabced24f1e6b96123ad527f9", "parents": [ "5504ce0317f777dcf102751c3e518284226fc2e1", "fe221742e388bea3f5856b5d9b2cb0a037020ea4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 12:41:17 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 12:41:17 2026 +0530" }, "message": "Merge tag \u0027driver-core-7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/driver-core/driver-core\n\nPull driver core updates from Danilo Krummrich:\n \"Deferred probe:\n - Fix race where deferred probe timeout work could be permanently\n canceled by using mod_delayed_work()\n - Fix missing jiffies conversion in deferred_probe_extend_timeout()\n - Guard timeout extension with delayed_work_pending() to prevent\n premature firing\n - Use system_percpu_wq instead of the deprecated system_wq\n - Update deferred_probe_timeout documentation\n\n device:\n - Replace direct struct device bitfield access (can_match, dma_iommu,\n dma_skip_sync, dma_ops_bypass, state_synced, dma_coherent,\n of_node_reused, offline, offline_disabled) with flag-based\n accessors using bit operations\n - Reject devices with unregistered buses\n - Delete unused DEVICE_ATTR_PREALLOC()\n - Add low-level device attribute macros with const show/store\n callbacks, allowing device attributes to reside in read-only memory\n - Move core device attributes to read-only memory\n - Constify group array pointers in driver_add_groups() /\n driver_remove_groups(), struct bus_type, and struct device_driver\n\n device property:\n - Fix fwnode reference leak in fwnode_graph_get_endpoint_by_id()\n - Initialize all fields of fwnode_handle in fwnode_init()\n - Provide swnode_get()/swnode_put() wrappers around kobject_get/put()\n - Allow passing struct software_node_ref_args pointers directly to\n PROPERTY_ENTRY_REF()\n\n driver_override:\n - Migrate amba, cdx, vmbus, and rpmsg to the generic driver_override\n infrastructure, fixing a UAF from unsynchronized access to\n driver_override in bus match() callbacks\n - Remove the now-unused driver_set_override()\n\n firmware loader:\n - Fix recursive lock deadlock in device_cache_fw_images() when async\n work falls back to synchronous execution\n - Fix device reference leak in firmware_upload_register()\n\n platform:\n - Pass KBUILD_MODNAME through the platform driver registration macro\n to create module symlinks in sysfs for built-in drivers; move\n module_kset initialization to a pure_initcall and tegra cbb\n registration to core_initcall to ensure correct ordering\n - Pass THIS_MODULE implicitly through a coresight_init_driver() macro\n\n sysfs:\n - Upgrade OOB write detection in sysfs_kf_seq_show() from printk to\n WARN\n - Add return value clamping to sysfs_kf_read()\n\n Rust:\n - ACPI:\n\n Fix missing match data for PRP0001 by exporting\n acpi_of_match_device()\n\n - Auxiliary:\n\n Replace drvdata() with dedicated registration data on\n auxiliary_device. drvdata() exposed the driver\u0027s bus device private\n data beyond the driver\u0027s own scope, creating ordering constraints\n and forcing the data to outlive all registrations that access it.\n Registration data is instead scoped structurally to the\n Registration object, making lifecycle ordering enforced by\n construction rather than convention.\n\n - Rust-native device driver lifetimes (HRT):\n\n Allow Rust device drivers to carry a lifetime parameter on their\n bus device private data, tied to the device binding scope -- the\n interval during which a bus device is bound to a driver. Device\n resources like pci::Bar\u003c\u0027a\u003e and IoMem\u003c\u0027a\u003e can be stored directly in\n the driver\u0027s bus device private data with a lifetime bounded by the\n binding scope, so the compiler enforces at build time that they do\n not outlive the binding. This removes Devres indirection from every\n access site and eliminates try_access() failure paths in\n destructors.\n\n Bus driver traits use a Generic Associated Type (GAT) Data\u003c\u0027bound\u003e\n to introduce the lifetime on the private data, rather than\n parameterizing the Driver trait itself. Auxiliary registration\n data, where the lifetime is not introduced by a trait callback but\n must be threaded through Registration, uses the ForLt trait (a\n type-level abstraction for types generic over a lifetime).\n\n Misc:\n - Fix DT overlayed devices not probing by reverting the broken\n treewide overlay fix and re-running fw_devlink consumer pickup when\n an overlay is applied to a bound device\n - Use root_device_register() for faux bus root device; add sanity\n check for failed bus init\n - Fix dev_has_sync_state() data race with READ_ONCE() and move it to\n base.h\n - Avoid spurious device_links warning when removing a device while\n its supplier is unbinding\n - Switch ISA bus to dynamic root device\n - Fix suspicious RCU usage in kernfs_put()\n - Remove devcoredump exit callback\n - Constify devfreq_event_class\"\n\n* tag \u0027driver-core-7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/driver-core/driver-core: (81 commits)\n software node: allow passing reference args to PROPERTY_ENTRY_REF()\n driver core: platform: set mod_name in driver registration\n coresight: pass THIS_MODULE implicitly through a macro\n kernel: param: initialize module_kset in a pure_initcall\n soc/tegra: cbb: Move driver registration from pure_initcall to core_initcall\n firmware_loader: Fix recursive lock in device_cache_fw_images()\n driver core: Use system_percpu_wq instead of system_wq\n driver core: remove driver_set_override()\n rpmsg: use generic driver_override infrastructure\n Drivers: hv: vmbus: use generic driver_override infrastructure\n cdx: use generic driver_override infrastructure\n amba: use generic driver_override infrastructure\n rust: devres: add \u0027static bound to Devres\u003cT\u003e\n samples: rust: rust_driver_auxiliary: showcase lifetime-bound registration data\n rust: auxiliary: generalize Registration over ForLt\n rust: types: add `ForLt` trait for higher-ranked lifetime support\n gpu: nova-core: separate driver type from driver data\n samples: rust: rust_driver_pci: use HRT lifetime for Bar\n rust: io: make IoMem and ExclusiveIoMem lifetime-parameterized\n rust: pci: make Bar lifetime-parameterized\n ...\n" }, { "commit": "5504ce0317f777dcf102751c3e518284226fc2e1", "tree": "51e595aed20eccd86476f8b562e8b197c6b2f904", "parents": [ "f3c4a338b5064effd9ce6d2cfd26c73daa56f1f4", "5ca48f6f2afc0a0c3f0584f95843451042ffc5ca" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 11:37:18 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 11:37:18 2026 +0530" }, "message": "Merge tag \u0027pm-7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/rafael/linux-pm\n\nPull power management updates from Rafael Wysocki:\n \"Over a half of the changes here are cpufreq updates that include core\n modifications, fixes of the old-style governors, new hardware support\n in drivers, assorded driver fixes and cleanups, and the removal of one\n driver (AMD Elan SC4*).\n\n Apart from that, the intel_idle driver will now be able to avoid\n exposing redundant C-states if PC6 is disabled and there are new\n sysctl knobs for device suspend/resume watchdog timeouts, hibernation\n gets built-in LZ4 support for image compression and there is the usual\n collection of assorted fixes and cleanups.\n\n Specifics:\n\n - Fix a race between cpufreq suspend and CPU hotplug during system\n shutdown (Tianxiang Chen)\n\n - Avoid redundant target() calls for unchanged limits and fix a typo\n in a comment in the cpufreq core (Viresh Kumar)\n\n - Fix concurrency issues related to sysfs attributes access that\n affect cpufreq governors using the common governor code (Zhongqiu\n Han)\n\n - Simplify frequency limit handling in the conservative cpufreq\n governor (Lifeng Zheng)\n\n - Fix descriptions of the conservative governor freq_step tunable and\n the ondemand governor sampling_down_factor tunable in the cpufreq\n documentation (Pengjie Zhang)\n\n - Fix use-after-free and double free during _OSC evaluation in the\n PCC cpufreq driver (Yuho Choi)\n\n - Rework the handling of policy min and max frequency values in the\n cpufreq core to allow drivers to specify special initial values for\n the scaling_min_freq and scaling_max_freq sysfs attributes (Pierre\n Gondois)\n\n - Add cpufreq scaling support for Qualcomm Shikra SoC (Taniya Das,\n Imran Shaik).\n\n - Improve the warning message on HWP-disabled hybrid processors\n printed by the intel_pstate driver and sync policy-\u003ecur during CPU\n offline in it (Yohei Kojima, Fushuai Wang)\n\n - Drop cpufreq support for AMD Elan SC4* (Sean Young)\n\n - Minor fixes for cpufreq drivers (Krzysztof Kozlowski, Akashdeep\n Kaur, Hans Zhang, Guangshuo Li, Xueqin Luo)\n\n - Clean up dead dependencies on X86 in the cpufreq Kconfig (Julian\n Braha)\n\n - Allow the intel_idle driver to avoid exposing C-states that are\n redundant when PC6 is disabled (Artem Bityutskiy)\n\n - Fix memory leak and a potential race in the OPP core (Abdun Nihaal,\n Di Shen)\n\n - Mark Rust OPP methods as inline (Nicolás Antinori)\n\n - Fix misc device registration failure path in the PM QoS core (Yuho\n Choi)\n\n - Add sysctl interface for DPM watchdog timeouts (Tzung-Bi Shih)\n\n - Use complete() instead of complete_all() in device_pm_sleep_init()\n to avoid a false-positive warning from lockdep_assert_RT_in_threaded_ctx()\n when CONFIG_PROVE_RAW_LOCK_NESTING is enabled (Jiakai Xu)\n\n - Use a flexible array for CRC uncompressed buffers during\n hibernation image saving (Rosen Penev)\n\n - Make the LZ4 algorithm available for hibernation compression\n (l1rox3)\n\n - Move the preallocate_image() call during hibernation after the\n \"prepare\" phase of the \"freeze\" transition (Matthew Leach)\n\n - Fix a memory leak in rapl_add_package_cpuslocked() in the\n intel_rapl power capping driver and use sysfs_emit() in\n cpumask_show() in that driver (Sumeet Pawnikar, Yury Norov)\n\n - Fix ValueError when parsing incomplete device properties in the\n pm-graph utility (Gongwei Li)\"\n\n* tag \u0027pm-7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/rafael/linux-pm: (40 commits)\n PM: dpm_watchdog: Add sysctl interface for DPM watchdog timeouts\n PM: QoS: Fix misc device registration unwind\n cpufreq: Use policy-\u003emin/max init as QoS request\n cpufreq: Remove driver default policy-\u003emin/max init\n cpufreq: Set default policy-\u003emin/max values for all drivers\n cpufreq: Extract cpufreq_policy_init_qos() function\n cpufreq: Documentation: fix conservative governor freq_step description\n cpufreq: ti: Add EPROBE_DEFER for K3 SoCs\n cpufreq: qcom: Add cpufreq scaling support for Qualcomm Shikra SoC\n dt-bindings: cpufreq: Document Qualcomm Shikra SoC EPSS\n powercap: intel_rapl: Use sysfs_emit() in cpumask_show()\n cpufreq: governor: Fix stale prev_cpu_nice spike when enabling ignore_nice_load\n cpufreq: governor: Fix data races on per-CPU idle/nice baselines\n PM: hibernate: Use flexible array for CRC uncompressed buffers\n powercap: intel_rapl: Fix memory leak in rapl_add_package_cpuslocked()\n PM: hibernate: make LZ4 available for hibernation compression\n PM: sleep: Use complete() in device_pm_sleep_init()\n opp: rust: mark OPP methods as inline\n cpufreq: intel_pstate: Improve warning message on HWP-disabled hybrid CPUs\n cpufreq: elanfreq: Drop support for AMD Elan SC4*\n ...\n" }, { "commit": "f3c4a338b5064effd9ce6d2cfd26c73daa56f1f4", "tree": "98aefc57a16f09d6329d646a991dd6e8847632a4", "parents": [ "eed451031f6ac098ca04ebdd63d34e083ba2d6f6", "bc1c872ce0ab3984f5fcefeb952541a802b891e7" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 11:35:11 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 11:35:11 2026 +0530" }, "message": "Merge tag \u0027thermal-7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/rafael/linux-pm\n\nPull thermal control updates from Rafael Wysocki:\n \"These add new hardware support (i.MX93 TMU, Amlogic T7, Intel Arrow\n Lake, QCom Nord, Shikra and Hawi), fix issues in a number of places in\n the thermal control core and drivers, clean up code and refactor it in\n preparation for future changes:\n\n - Rework the initialization and cleanup of thermal class cooling\n devices to separate DT-based cooling device registration and\n cooling device registration without DT (Daniel Lezcano, Ovidiu\n Panait)\n\n - Update the cooling device DT bindings to support 3-cell cooling\n device representation, where the additional cell holds an ID to\n select a cooling mechanism for devices that offer multiple cooling\n mechanisms, and adjust the cooling device registration code\n accordingly (Gaurav Kohli, Daniel Lezcano)\n\n - Remove dead code from two functions in the thermal core and\n simplify the unregistration of thermal governors (Rafael Wysocki)\n\n - Fix critical temperature attribute removal handling in the generic\n thermal zone hwmon support code and rework that code to register a\n separate hwmon class device for each thermal zone (instead of using\n one hwmon class device for all thermal zones of the same type) to\n address thermal zone removal deadlocks (Rafael Wysocki)\n\n - Use attribute groups for adding temperature attributes to hwmon\n class devices associated with thermal zones (Rafael Wysocki)\n\n - Pass WQ_UNBOUND when allocating the thermal workqueue (Marco\n Crivellari)\n\n - Fix potential shift overflow in ptc_mmio_write() and improve error\n handling in proc_thermal_ptc_add() in the int340x thermal control\n driver (Aravind Anilraj)\n\n - Use sysfs_emit() for cpumask printing in the Intel powerclamp\n thermal driver (Yury Norov)\n\n - Add Arrow Lake CPU models to the intel_tcc_cooling driver (Srinivas\n Pandruvada)\n\n - Add QCom Nord, Shikra and Hawi temperature sensor DT bindings\n (Deepti Jaggi, Gaurav Kohli, Dipa Ramesh Mantre)\n\n - Use devm_add_action_or_reset() for clock disable on the NVidia\n soctherm and switch it to devm cooling device registration version\n (Daniel Lezcano)\n\n - Add the Amlogic T7 thermal sensor along with thermal calibration\n data read from SMC calls (Ronald Claveau)\n\n - Fix atomic temperature read in the QCom tsens driver to comply with\n hardware documentation (Priyansh Jain)\n\n - Add SpacemiT K1 thermal sensor support (Shuwei Wu)\n\n - Add i.MX93 temperature sensor support and filter out the invalid\n temperature (Jacky Bai)\n\n - Enable by default the TMU (Thermal Monitoring Unit) on Exynos\n platform (Krzysztof Kozlowski)\n\n - Rework interrupt initialization in the Tsens driver and add the\n optional wakeup source (Priyansh Jain)\n\n - Fix typo in a comment in the TSens QCom driver (Jinseok Kim)\n\n - Fix trailing whitespace and repeated word in the OF code, remove\n quoted string splitting across lines from the iMX7 driver, and\n remove a stray space from the thermal_trip_of_attr() macro\n definition (Mayur Kumar)\n\n - Update the thermal testing facility code to avoid NULL pointer\n dereferences by rejecting missing command arguments and replace\n sscanf() with kstrtoint() or kstrtoul() in that code (Ovidiu\n Panait, Samuel Moelius)\"\n\n* tag \u0027thermal-7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/rafael/linux-pm: (54 commits)\n thermal: sysfs: Replace sscanf() with kstrtoul()\n thermal: testing: Replace sscanf() with kstrtoint()\n thermal: testing: reject missing command arguments\n thermal: intel: intel_tcc_cooling: Add Arrow Lake CPU models\n thermal/drivers/qcom/tsens: Disable wakeup interrupt setup on automotive targets\n thermal/drivers/qcom/tsens: Switch wake IRQ handling to PM callbacks\n thermal/core: Fix missing stub for devm_thermal_cooling_device_register\n dt-bindings: thermal: cooling-devices: Update support for 3 cells cooling device\n thermal/of: Support cooling device ID in cooling-spec\n thermal/of: Pass cdev_id and introduce devm registration helper\n thermal/of: Add cooling device ID support\n thermal/of: Rename the devm_thermal_of_cooling_device_register() function\n thermal/core: Make cooling device OF node conditional on CONFIG_THERMAL_OF\n thermal/of: Move cooling device OF helpers out of thermal core\n hwmon: Use non-OF thermal cooling device registration API\n thermal/core: Add devm_thermal_cooling_device_register()\n thermal/core: Introduce non-OF thermal_cooling_device_register()\n thermal/drivers/samsung: Enable TMU by default\n thermal/driver/qoriq: Workaround unexpected temperature readings from tmu\n thermal/drivers/qoriq: Add i.MX93 tmu support\n ...\n" }, { "commit": "eed451031f6ac098ca04ebdd63d34e083ba2d6f6", "tree": "69b7102edff2a6fda2731c2ea38e96a72f9c83d0", "parents": [ "966240b9900b33ee7b02a18011bc6109b791ff31", "615f90d3b1b7900d591a9162793f616b051f07b8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 11:32:38 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 11:32:38 2026 +0530" }, "message": "Merge tag \u0027acpi-7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/rafael/linux-pm\n\nPull ACPI support updates from Rafael Wysocki:\n \"These update the ACPICA code in the kernel to upstream version\n 20260408, introduce support for devres-based management of ACPI notify\n handlers and update some core ACPI device drivers on top of that\n (which includes some fixes and cleanups), add _DEP support for PCI/CXL\n roots and Intel CVS devices, fix a couple of assorted issues and clean\n up code:\n\n - Fix multiple issues related to probe, removal and missing NVDIMM\n device notifications in the ACPI NFIT driver (Rafael Wysocki)\n\n - Add support for devres-based management of ACPI notify handlers to\n the ACPI core (Rafael Wysocki)\n\n - Switch multiple core ACPI device drivers (including the ACPI PAD,\n ACPI video bus, ACPI HED, ACPI thermal zone, ACPI AC, ACPI battery,\n and ACPI NFIT drivers) over to using devres-based resource\n management during probe (Rafael Wysocki)\n\n - Replace mutex_lock/unlock() with guard()/scoped_guard() in the ACPI\n PMIC driver (Maxwell Doose)\n\n - Fix message kref handling in the dead device path of the ACPI IPMI\n address space handler (Yuho Choi)\n\n - Use sysfs_emit() in idlecpus_show() in the ACPI processor\n aggregator device (PAD) driver (Yury Norov)\n\n - Clean up device_id_scheme initialization in the ACPI video bus\n driver (Jean-Ralph Aviles)\n\n - Clean up lid handling in the ACPI button driver and\n acpi_button_probe(), reorganize installing and removing event\n handlers in that driver and switch it over to using devres-based\n resource management during probe (Rafael Wysocki)\n\n - Add support for the Legacy Virtual Register (LVR) field in I2C\n serial bus resource descriptors to ACPICA (Akhil R)\n\n - Fix multiple issues related to bounds checks, input validation,\n use-after-free, and integer overflow checks in the AML interpreter\n in ACPICA (ikaros)\n\n - Update the copyright year to 2026 in ACPICA files and make minor\n changes related to ACPI 6.6 support (Pawel Chmielewski)\n\n - Remove spurious precision from format used to dump parse trees in\n ACPICA (David Laight)\n\n - Add modern standby DSM GUIDs to ACPICA header files (Daniel\n Schaefer)\n\n - Fix FADT 32/64X length mismatch warning in ACPICA (Abdelkader\n Boudih)\n\n - Update D3hot/cold device power states definitions in ACPICA header\n files (Aymeric Wibo)\n\n - Fix NULL pointer dereference in acpi_ns_custom_package() (Weiming\n Shi)\n\n - Update ACPICA version to 20260408 (Saket Dumbre)\n\n - Add cpuidle driver check in acpi_processor_register_idle_driver()\n to avoid evaluating _CST unnecessarily (Tony W Wang-oc)\n\n - Suppress UBSAN warning caused by field misuse during PCC-based\n register access in the ACPI CPPC library (Jeremy Linton)\n\n - Add support for CPPC v4 to the ACPI CPPC library (Sumit Gupta)\n\n - Update the ACPI device enumeration code to honor _DEP for ACPI0016\n PCI/CXL host bridges and make the ACPI PCI root driver clear _DEP\n dependencies for PCI roots that have become operational (Chen Pei)\"\n\n* tag \u0027acpi-7.2-rc1\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/rafael/linux-pm: (74 commits)\n ACPI: processor: Add cpuidle driver check in acpi_processor_register_idle_driver()\n ACPI: IPMI: Fix message kref handling on dead device\n ACPI: CPPC: Suppress UBSAN warning caused by field misuse\n ACPI: scan: Honor _DEP for Intel CVS devices\n ACPI: NFIT: core: Fix possible deadlock and missing notifications\n ACPI: NFIT: core: Eliminate redundant local variable\n ACPI: NFIT: core: Fix acpi_nfit_init() error cleanup\n ACPI: NFIT: core: Fix possible NULL pointer dereference\n ACPI: bus: Clean up devm_acpi_install_notify_handler()\n ACPI: button: Switch over to devres-based resource management\n ACPI: button: Reorganize installing and removing event handlers\n ACPI: button: Use string literals for generating netlink messages\n ACPI: button: Clean up adding and removing lid procfs interface\n ACPI: button: Merge two switch () statements in acpi_button_probe()\n ACPI: button: Drop redundant variable from acpi_button_probe()\n ACPI: button: Rework device verification during probe\n ACPI: CPPC: Add support for CPPC v4\n ACPI: PAD: Use sysfs_emit() in idlecpus_show()\n ACPI: scan: Honor _DEP for ACPI0016 PCI/CXL host bridge\n ACPI: PCI: Clear _DEP dependencies after PCI root bridge attach\n ...\n" }, { "commit": "966240b9900b33ee7b02a18011bc6109b791ff31", "tree": "b9def013c4d6c80593b37fb306ffa583aca20d43", "parents": [ "b079329b8691768962aa514b8f8c9077ca352459", "3850c2920a10d5f50f5c2f8acccfb3925002f30b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 11:29:31 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 11:29:31 2026 +0530" }, "message": "Merge tag \u0027nolibc-20260614-for-7.2-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/nolibc/linux-nolibc\n\nPull nolibc updates from Thomas Weißschuh:\n\n - New architectures: OpenRISC and 32-bit parisc\n\n - New library functionality: alloca(), assert(), creat() and\n ftruncate()\n\n - Automatic large file support\n\n - Proper 64-bit system call argument passing on x32 and MIPS N32\n\n - Cleanups of the testmatrix\n\n - Various bugfixes and cleanups\n\n* tag \u0027nolibc-20260614-for-7.2-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/nolibc/linux-nolibc: (37 commits)\n selftests/nolibc: test against -Wwrite-strings\n selftests/nolibc: use mutable buffer for execve() argv string\n tools/nolibc: cast default values of program_invocation_name\n tools/nolibc: add ftruncate()\n tools/nolibc: add a helper to split a 64-bit argument into 32-bit halves\n selftests/nolibc: enable CONFIG_TMPFS for sparc32\n tools/nolibc: stackprotector: Avoid stalling program startup if crng is not init yet\n tools/nolibc: getopt: Fix potential out of bounds access\n selftests/nolibc: test open mode handling\n tools/nolibc: always pass mode to open syscall\n tools/nolibc: split open mode handling into a macro\n tools/nolibc: split implicit open flags into a macro\n tools/nolibc: add support for 32-bit parisc\n selftests/nolibc: avoid function pointer comparisons\n tools/nolibc: add support for OpenRISC / or1k\n selftests/nolibc: use vmlinux for MIPS tests\n selftests/nolibc: trim IMAGE mappings\n selftests/nolibc: trim DEFCONFIG mappings\n selftests/nolibc: trim QEMU_ARCH mappings\n selftests/nolibc: use QEMU_ARCH for QEMU_ARCH_USER\n ...\n" }, { "commit": "b079329b8691768962aa514b8f8c9077ca352459", "tree": "bfa889319f6a8174e3cdf84e339c45f76f93bdd3", "parents": [ "b8b674748fa4b1a384aaa647811109fe9007c0a4", "48b375e482027ba6566107cec40c1b21b453fa4e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 09:25:48 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 09:25:48 2026 +0530" }, "message": "Merge tag \u0027rust-7.2\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/ojeda/linux\n\nPull Rust updates from Miguel Ojeda:\n \"This one is big due to the vendoring of the `zerocopy` library, which\n allows us to replace a bunch of `unsafe` code dealing with conversions\n between byte sequences and other types with safe alternatives. More\n details on that below (and in its merge commit).\n\n Toolchain and infrastructure:\n\n - Introduce support for the \u0027zerocopy\u0027 library [1][2]:\n\n Fast, safe, compile error. Pick two.\n\n Zerocopy makes zero-cost memory manipulation effortless. We write\n `unsafe` so you don\u0027t have to.\n\n It essentially provides derivable traits (e.g. \u0027FromBytes\u0027) and\n macros (e.g. \u0027transmute!\u0027) for safely converting between byte\n sequences and other types. Having such support allows us to remove\n some \u0027unsafe\u0027 code.\n\n It is among the most downloaded Rust crates and it is also used by\n the Rust compiler itself.\n\n It is licensed under \"BSD-2-Clause OR Apache-2.0 OR MIT\".\n\n The crates are imported essentially as-is (only +2/-3 lines needed\n to be adapted), plus SPDX identifiers. Upstream has since added the\n SPDX identifiers as well as one of the tweaks at my request, thus\n reducing our future diffs on updates -- I keep the details in one\n of our usual live lists [3].\n\n In total, it is about ~39k lines added, ~32k without counting\n \u0027benches/\u0027 which are just for documentation purposes.\n\n The series includes a few Kbuild and rust-analyzer improvements and\n an example patch using it in Nova, removing one \u0027unsafe impl\u0027.\n\n I checked that the codegen of an isolated example function (similar\n to the Nova patch on top) is essentially identical. It also turns\n out that (for that particular case) the \u0027zerocopy\u0027 version, even\n with \u0027debug-assertions\u0027 enabled, has no remaining panics, unlike a\n few in the current code (since the compiler can prove the remaining\n \u0027ub_checks\u0027 statically).\n\n So their \"fast, safe\" does indeed check out -- at least in that\n case.\n\n - Support AutoFDO. This allows Rust code to be profiled and optimized\n based on the profile. Tested with Rust Binder: ~13% slower without\n AutoFDO in the binderAddInts benchmark (using an app-launch\n benchmark for the profile).\n\n - Support Software Tag-Based KASAN.\n\n In addition, fix KASAN Kconfig by requiring Clang.\n\n - Add Kconfig options for each existing Rust KUnit test suite, such\n as \u0027CONFIG_RUST_BITMAP_KUNIT_TEST\u0027.\n\n They are placed within a new menu, \u0027CONFIG_RUST_KUNIT_TESTS\u0027, in\n the new \u0027rust/kernel/Kconfig.test\u0027 file.\n\n - Support the upcoming Rust 1.98.0 release (expected 2026-08-20):\n lint cleanups and an unstable flag rename.\n\n - Disable \u0027rustdoc\u0027 documentation inlining for all prelude items,\n which bloats the generated documentation.\n\n - Ignore (in Git) and clean (in Kbuild) the (rarely) \u0027rustc\u0027-generated\n \u0027*.long-type-*.txt\u0027 files.\n\n \u0027kernel\u0027 crate:\n\n - Add new \u0027bitfield\u0027 module with the \u0027bitfield!\u0027 macro (extracted\n from the existing \u0027register!\u0027 one), which declares integer types\n that are split into distinct bit fields of arbitrary length.\n\n Each field is a \u0027Bounded\u0027 of the appropriate bit width (ensuring\n values are properly validated and avoiding implicit data loss) and\n gets several generated getters and setters (infallible, \u0027const\u0027 and\n fallible) as well as associated constants (\u0027_MASK\u0027, \u0027_SHIFT\u0027 and\n \u0027_RANGE\u0027). It also supports fields that can be converted from/to\n custom types, either fallibly (\u0027?\u003d\u003e\u0027) or infallibly (\u0027\u003d\u003e\u0027).\n\n For instance:\n\n bitfield! {\n struct Rgb(u16) {\n 15:11 blue;\n 10:5 green;\n 4:0 red;\n }\n }\n\n // Compile-time checks.\n let color \u003d Rgb::zeroed().with_const_green::\u003c0x1f\u003e();\n\n assert_eq!(color.green(), 0x1f);\n assert_eq!(color.into_raw(), 0x1f \u003c\u003c Rgb::GREEN_SHIFT);\n\n Add as well documentation and a test suite for it, as usual; and\n update the \u0027register!\u0027 macro to use it.\n\n It will be maintained by Alexandre Courbot (with Yury Norov as\n reviewer) under a new \u0027MAINTAINERS\u0027 entry: \u0027RUST [BITFIELD]\u0027.\n\n - \u0027ptr\u0027 module: rework index projection syntax into keyworded syntax\n and introduce panicking variant.\n\n The keyword syntax (\u0027build:\u0027, \u0027try:\u0027, \u0027panic:\u0027) is more explicit\n and paves the way of perhaps adding more flavors in the future,\n e.g. an \u0027unsafe\u0027 index projection.\n\n For instance, projections now look like this:\n\n fn f(p: *const [u8; 32]) -\u003e Result {\n // Ok, within bounds, checked at build time.\n project!(p, [build: 1]);\n\n // Build error.\n project!(p, [build: 128]);\n\n // `OutOfBound` runtime error (convertible to `ERANGE`).\n project!(p, [try: 128]);\n\n // Runtime panic.\n project!(p, [panic: 128]);\n\n Ok(())\n }\n\n Update as well the users, which now look like e.g.\n\n // Pointer to the first entry of the GSP message queue.\n let data \u003d project!(self.0.as_ptr(), .gspq.msgq.data[build: 0]);\n\n - \u0027build_assert\u0027 module: make the module the home of its macros\n instead of rendering them twice.\n\n - \u0027sync\u0027 module: add \u0027UniqueArc::as_ptr()\u0027 associated function.\n\n - \u0027alloc\u0027 module:\n\n - Fix the \u0027Vec::reserve()\u0027 doctest to properly account for the\n existing vector length in the capacity assertion.\n\n - Fix an incorrect operator in the \u0027Vec::extend_with()\u0027 \u0027SAFETY\u0027\n comment; add a doc test demonstrating basic usage and the\n zero-length case.\n\n - Clean imports across several modules to follow the \"kernel\n vertical\" import style in order to minimize conflicts.\n\n \u0027pin-init\u0027 crate:\n\n - User visible changes:\n\n - Do not generate \u0027non_snake_case\u0027 warnings for identifiers that\n are syntactically just users of a field name. This would allow\n all \u0027#[allow(non_snake_case)]\u0027 in nova-core to be removed,\n which Gary will send to the nova tree next cycle.\n\n - Filter non-cfg attributes out properly in derived structs. This\n improves pin-init compatibility with other derive macros.\n\n - Insert projection types\u0027 where clause properly.\n\n - Other changes:\n\n - Bump MSRV to 1.82, plus associated cleanups.\n\n - Overhaul how init slots are projected. The new approach is\n easier to justify with safety comments.\n\n - Mark more functions as inline, which should help mitigate the\n super-long symbol name issue due to lack of inlining.\n\n rust-analyzer:\n\n - Support \u0027--envs\u0027 for passing env vars for crates like \u0027zerocopy\u0027.\n\n \u0027MAINTAINERS\u0027:\n\n - Add the following reviewers to the \u0027RUST\u0027 entry:\n - Daniel Almeida\n - Tamir Duberstein\n - Alexandre Courbot\n - Onur Özkan\n\n They have been involved in the Rust for Linux project for about 7\n collective years and bring expertise across several domains, which\n will be very useful to have around in the future.\n\n Thanks everyone for stepping up!\n\n And some other fixes, cleanups and improvements\"\n\nLink: https://github.com/google/zerocopy [1]\nLink: https://docs.rs/zerocopy [2]\nLink: https://github.com/Rust-for-Linux/linux/issues/1239 [3]\n\n* tag \u0027rust-7.2\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/ojeda/linux: (86 commits)\n MAINTAINERS: add Onur Özkan as Rust reviewer\n MAINTAINERS: add Alexandre Courbot as Rust reviewer\n MAINTAINERS: add Tamir Duberstein as Rust reviewer\n MAINTAINERS: add Daniel Almeida as Rust reviewer\n kbuild: rust: clean `zerocopy-derive` in `mrproper`\n rust: make `build_assert` module the home of related macros\n rust: str: clean unused import for Rust \u003e\u003d 1.98\n rust: str: use the \"kernel vertical\" imports style\n rust: aref: use the \"kernel vertical\" imports style\n rust: page: use the \"kernel vertical\" imports style\n gpu: nova-core: firmware: parse `FalconUCodeDescV2` via `zerocopy`\n rust: prelude: add `zerocopy{,_derive}::FromBytes`\n rust: zerocopy-derive: enable support in kbuild\n rust: zerocopy-derive: add `README.md`\n rust: zerocopy-derive: avoid generating non-ASCII identifiers\n rust: zerocopy-derive: add SPDX License Identifiers\n rust: zerocopy-derive: import crate\n rust: zerocopy: enable support in kbuild\n rust: zerocopy: add `README.md`\n rust: zerocopy: remove float `Display` support\n ...\n" }, { "commit": "b8b674748fa4b1a384aaa647811109fe9007c0a4", "tree": "1489756653bd456f9babab129618b456297802b7", "parents": [ "95e56f0f293ef797123eb032f78f5b5d56a035a6", "e853c1b28580ea93fda3cd729e440a3fc16fa647" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 09:16:00 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 09:16:00 2026 +0530" }, "message": "Merge tag \u0027rcu.release.v7.2\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/rcu/linux\n\nPull RCU updates from Uladzislau Rezki:\n \"Torture test updates:\n\n - Improve kvm-series.sh script by adding examples in its header\n comment\n\n - Lazy RCU is more fully tested now by replacing call_rcu_hurry()\n with call_rcu() and doing rcu_barrier() to motivate lazy callbacks\n during a stutter pause\n\n - Add more synonyms for the \"--do-normal\" group of torture.sh\n command-line arguments\n\n Misc changes:\n\n - Reduce stack usage of nocb_gp_wait() to address frame size warning\n when built with CONFIG_UBSAN_ALIGNMENT\n\n - The synchronize_rcu() call can detect the flood and latches a\n normal/default path temporary switching to wait_rcu_gp() path\n\n - Document using rcu_access_pointer() to fetch the old pointer for\n lockless cmpxchg() updates\n\n - Simplify some RCU code using clamp_val()\n\n - Fix a kerneldoc header comment typo in srcu_down_read_fast()\"\n\n* tag \u0027rcu.release.v7.2\u0027 of gitolite.kernel.org:pub/scm/linux/kernel/git/rcu/linux:\n rcu/nocb: reduce stack usage in nocb_gp_wait()\n rcu-tasks: Fix possible boot-time tests failed for the call_rcu_tasks()\n rcu: Latch normal synchronize_rcu() path on flood\n rcu: Document rcu_access_pointer() feeding into cmpxchg()\n rcu: Simplify param_set_next_fqs_jiffies() by applying clamp_val()\n rcu: Simplify rcu_do_batch() by applying clamp()\n checkpatch: Undeprecate rcu_read_lock_trace() and rcu_read_unlock_trace()\n srcu: Fix kerneldoc header comment typo in srcu_down_read_fast()\n torture: Allow \"norm\" abbreviation for \"normal\"\n torture: Improve kvm-series.sh header comment\n torture: Add torture_sched_set_normal() for user-specified nice values\n rcutorture: Fully test lazy RCU\n" }, { "commit": "95e56f0f293ef797123eb032f78f5b5d56a035a6", "tree": "19372012f5dfbe2ef2aeb4c1a248ed34ff83b876", "parents": [ "2bfc56d9f5e82f6aada2a8d68093aab0b1f00f6a", "07a1a6562ce29e2e0c134a57882d6e52e8758492" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 08:56:31 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 08:56:31 2026 +0530" }, "message": "Merge tag \u0027kcsan-20260612-v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/melver/linux\n\nPull KCSAN update from Marco Elver:\n\n - Silence -Wmaybe-uninitialized when calling __kcsan_check_access()\n\n* tag \u0027kcsan-20260612-v7.2-rc1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/melver/linux:\n kcsan: Silence -Wmaybe-uninitialized when calling __kcsan_check_access()\n" }, { "commit": "462bdd08fbdf41db223c6117d907c8fd68d666ea", "tree": "aee7fc1954511c8f49c82419d77591157d55c530", "parents": [ "2005aabe94eaab8608879d98afb901bc99bc3a31" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Wed Feb 11 15:11:28 2026 -0500" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Sun Jun 14 22:29:21 2026 -0400" }, "message": "udf: fix nls leak on udf_fill_super() failure\n\nOn all failure exits that go to error_out there we have already moved the\nnls reference from uopt-\u003enls_map to sbi-\u003es_nls_map, leaving NULL behind.\n\nFixes: c4e89cc674ac (\"udf: convert to new mount API\")\nAcked-by: Jan Kara \u003cjack@suse.cz\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "2556746302b335b3a9b016e315c7a14f272eab12", "tree": "f4fe77ecfe6a3ee477c66ff53684c68669b826b0", "parents": [ "78d86a71de6ea70c6228e7817f3962c618c1eb54", "2c6a0994e9f15da5aba8cb108df35a257a792030" ], "author": { "name": "Borislav Petkov (AMD)", "email": "bp@alien8.de", "time": "Sun Jun 14 18:39:50 2026 -0700" }, "committer": { "name": "Borislav Petkov (AMD)", "email": "bp@alien8.de", "time": "Sun Jun 14 18:39:50 2026 -0700" }, "message": "Merge remote-tracking branches \u0027ras/edac-drivers\u0027 and \u0027ras/edac-misc\u0027 into edac-updates\n\n* ras/edac-drivers: (21 commits)\n EDAC: Consistently define pci_device_ids using named initializers\n EDAC/igen6: Add Intel Nova Lake-H SoC support\n EDAC/igen6: Make registers for detecting IBECC configurable\n EDAC/imh: Add RRL support for Intel Diamond Rapids server\n EDAC/{skx_common,i10nm}: Prepare RRL for sub-channel granularity\n EDAC/skx_common: Add SubChannel support to ADXL decode\n EDAC/{skx_common,i10nm}: Move RRL handling to common code\n EDAC/{skx_common,i10nm}: Introduce rrl_ctrl_mode\n EDAC/{skx_common,i10nm}: Rename rrl_mode to rrl_source_type\n EDAC/{skx_common,skx,i10nm}: Split skx_set_decode()\n EDAC/{skx_common,i10nm,imh}: Move MC register access helpers to skx_common\n EDAC/{skx_common,skx}: Fix UBSAN shift-out-of-bounds in skx_get_dimm_info\n EDAC/igen6: Add one Intel Panther Lake-H SoC support\n EDAC/igen6: Fix memory topology parsing for Panther Lake-H SoCs\n EDAC/igen6: Fix call trace due to missing release()\n EDAC/sb_edac: fix grammar in sb_decode_ddr3 warning\n EDAC/i5400: disable error reporting at teardown and refactor helper\n EDAC/i5100: disable error reporting at teardown and create helper\n EDAC/i5000: disable error reporting at teardown and refactor helper\n EDAC/i7300: disable error reporting if init fails and refactor helper\n ...\n\n* ras/edac-misc:\n RAS/AMD/ATL: Drop malformed default N from Kconfig\n\nSigned-off-by: Borislav Petkov (AMD) \u003cbp@alien8.de\u003e\n" }, { "commit": "2bfc56d9f5e82f6aada2a8d68093aab0b1f00f6a", "tree": "6441d1db6b098baa5ea17b856d39723952c6f437", "parents": [ "73f399414a84d715bb1794182aaea852b11d0962", "a174910917a8e93cb5334e9dce8bac32bff22c47" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 05:06:02 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 05:06:02 2026 +0530" }, "message": "Merge tag \u0027for-linus-7.2-rc1-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip\n\nPull xen updates from Juergen Gross:\n\n - Several small cleanups of various Xen related drivers\n (xen/platform-pci, xen-balloon, xenbus, xen/mcelog)\n\n - Cleanup for Xen PV-mode related code (includes dropping the Xen\n debugfs code)\n\n - Drop the additional lazy mmu mode tracking done by Xen specific code\n\n* tag \u0027for-linus-7.2-rc1-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:\n xen/xenbus: Replace strcpy() with memcpy()\n x86/xen: Replace generic lazy tracking with cpu specific one\n x86/xen: Get rid of last XEN_LAZY_MMU uses\n mm: Refactor lazy_mmu_mode_pause() and lazy_mmu_mode_resume()\n x86/xen: Change interface of xen_mc_issue()\n x86/xen: Drop lazy mode from trace entries\n x86/xen: Remove Xen debugfs support\n x86/xen: Cleanup Xen related trace points\n x86/xen: Guard PV-only stuff in xen-ops.h with CONFIG_XEN_PV\n xen: balloon: Replace sprintf() with sysfs_emit()\n xen/mcelog: mark g_physinfo, ncpus and xen_mce_chrdev_device as __ro_after_init\n xen: constify xsd_errors array\n xen/platform-pci: Simplify initialization of pci_device_id array\n" }, { "commit": "73f399414a84d715bb1794182aaea852b11d0962", "tree": "91a7d6656d3d93dfb77a40c7b5763048ce9c61c1", "parents": [ "de02909ae81aa4fda213d16915adb5e1b088a7db", "1a1e62a5a48494cdf33e3bfb82fb8f408da7c4cc" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 05:01:15 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 05:01:15 2026 +0530" }, "message": "Merge tag \u0027kbuild-7.2-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kbuild/linux\n\nPull Kbuild / Kconfig updates from Nathan Chancellor:\n \"Kbuild:\n\n - Remove broken module linking exclusion for BTF\n\n - Add documentation around how offset header files work\n\n - Include unstripped vDSO libraries in pacman packages\n\n - Bump minimum version of LLVM for building the kernel to 17.0.1 and\n clean up unnecessary workarounds\n\n - Use a context manager in run-clang-tools\n\n - Add dist macro value if present to release tag for RPM packages\n\n - Detect and report truncated buf_printf() output in modpost\n\n - Add __llvm_covfun and __llvm_covmap to section whitelist in modpost\n\n - Support Clang\u0027s distributed ThinLTO mode\n\n - Remove architecture specific configurations for AutoFDO and\n Propeller to ease individual architecture maintenance\n\n Kconfig:\n\n - Add kconfig-sym-check target to look for dangling Kconfig symbol\n references and invalid tristate literal values\n\n - Harden against potential NULL pointer dereference\n\n - Fix typo in Kconfig test comment\"\n\n* tag \u0027kbuild-7.2-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kbuild/linux: (31 commits)\n kconfig: tests: fix typo in comment\n kconfig: Remove the architecture specific config for Propeller\n kconfig: Remove the architecture specific config for AutoFDO\n modpost: Add __llvm_covfun and __llvm_covmap to section_white_list\n kconfig: add kconfig-sym-check static checker\n kbuild: Remove unnecessary \u0027T\u0027 modifier in cmd_ar_builtin_fixup\n kbuild: distributed build support for Clang ThinLTO\n kbuild: move vmlinux.a build rule to scripts/Makefile.vmlinux_a\n scripts: modpost: detect and report truncated buf_printf() output\n kbuild: rpm-pkg: append %{?dist} macro to Release tag\n run-clang-tools: run multiprocessing.Pool as context manager\n compiler-clang.h: Drop explicit version number from \"all\" diagnostic macro\n compiler-clang.h: Remove __cleanup -Wunused-variable workaround\n kbuild: Remove check for broken scoping with clang \u003c 17 in CC_HAS_ASM_GOTO_OUTPUT\n x86/entry/vdso32: Remove conditional omission of \u0027.cfi_offset eflags\u0027\n x86/module: Revert \"Deal with GOT based stack cookie load on Clang \u003c 17\"\n x86/build: Drop unnecessary \u0027-ffreestanding\u0027 addition to KBUILD_CFLAGS\n scripts/Makefile.warn: Drop -Wformat handling for clang \u003c 16\n riscv: Drop tautological condition from TOOLCHAIN_NEEDS_OLD_ISA_SPEC\n riscv: Remove tautological condition from selection of ARCH_SUPPORTS_CFI\n ...\n" }, { "commit": "de02909ae81aa4fda213d16915adb5e1b088a7db", "tree": "a49a44264dc37e943786d6e074105ad1b458f2b7", "parents": [ "8ab34af09f292ca9620dad8df253fec766729b29", "d53ac61b6120fa596d758ff6e22b5dcb6db21ce8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 04:28:20 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 04:28:20 2026 +0530" }, "message": "Merge tag \u0027pull-configfs-fixed\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs\n\nPull configfs updates from Al Viro:\n \"A couple of fixes (UAF in configfs_lookup() and really old races\n introduced when lseek() on configfs directories stopped locking those\n directories; impact up to and including UAF).\n\n Fixes aside, the main result is that configfs is finally switched to\n tree-in-dcache machinery. It\u0027s *not* making use of recursive removal\n helpers yet, and it still does the bloody awful \"build subtree in full\n sight of userland, with possibility of failure halfway through and\n need to unroll\" that forces the locking model from hell; dealing with\n that is a separate patch series, once this one is out of the way.\n However, it is using DCACHE_PERSISTENT properly now. And apparmorfs is\n the sole remaining user of __simple_{unlink,rmdir}() at that point\"\n\n* tag \u0027pull-configfs-fixed\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:\n create_default_group(): pass parent\u0027s dentry instead of config_group\n configfs_attach_group(): drop the unused parent_item argument\n configs_attach_item(): drop unused parent_item argument\n configfs_create(): lift parent timestamp updates into callers\n kill configfs_drop_dentry()\n configfs: mark pinned dentries persistent\n configfs: dentry refcount needs to be pinned only once\n switch configfs_detach_{group,item}() to passing dentry\n configfs_remove_dir(), detach_attrs(): switch to passing dentry\n populate_attrs(): move cleanup to the sole caller\n populate_group(): move cleanup on failure to the sole caller\n configfs_detach_rollback(): pass configfs_dirent instead of dentry\n configfs_do_depend_item(): pass configfs_dirent instead of dentry\n configfs_depend_prep(): pass configfs_dirent instead of dentry\n configfs_detach_prep(): pass configfs_dirent instead of dentry\n configfs_mkdir(): use take_dentry_name_snapshot()\n configfs: fix lockless traversals of -\u003es_children\n configfs_lookup(): don\u0027t leave -\u003es_dentry dangling on failure\n" }, { "commit": "8ab34af09f292ca9620dad8df253fec766729b29", "tree": "86e19444f0633efc34c2131ded68ed1db87d9bff", "parents": [ "e8a56d6fc828bb569fa2dd33c3e6eb16a165b097", "2555ac7a450bf11f8e461ce6cd0f62613734a796" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 04:21:00 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 04:21:00 2026 +0530" }, "message": "Merge tag \u0027pull-d_add\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs\n\nPull dentry d_add() cleanups from Al Viro:\n \"This converts a bunch of unidiomatic uses of d_add() in -\u003elookup()\n instances to equivalent uses of d_splice_alias(), which is the normal\n mechanism for -\u003elookup()\"\n\n* tag \u0027pull-d_add\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:\n gfs2: use d_splice_alias() for -\u003elookup() return value\n ntfs: use d_splice_alias() for -\u003elookup() return value\n simple_lookup(): use d_splice_alias() for -\u003elookup() return value\n ecryptfs: use d_splice_alias() for -\u003elookup() return value\n configfs_lookup(): switch to d_splice_alias()\n tracefs: use d_splice_alias() in -\u003elookup() instances\n" }, { "commit": "e8a56d6fc828bb569fa2dd33c3e6eb16a165b097", "tree": "1742623129141d47c68509271cf8183133e48849", "parents": [ "79169a1624253363fed3e9a447b77e50bb226206", "3df5153c5f123d6018c82a24341ccd99c79d64a0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 04:15:31 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 04:15:31 2026 +0530" }, "message": "Merge tag \u0027pull-dcache\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs\n\nPull dcache updates from Al Viro:\n\n - d_alloc_parallel() API change (Neil\u0027s with my changes)\n\n - NORCU fixes\n\n - Reorganization and simplification of dentry eviction logic\n\n - Simplifying rcu_read_lock() scopes in fs/dcache.c\n\n - Secondary roots work - getting rid of NFS fake root dentries and\n dealing with remaining shrink_dcache_for_umount() and\n shrink_dentry_list() races\n\n - making cursors NORCU (surprisingly easy)\n\n* tag \u0027pull-dcache\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (22 commits)\n make cursors NORCU\n nfs: get rid of fake root dentries\n wind -\u003es_roots via -\u003ed_sib instead of -\u003ed_hash\n shrink_dentry_tree(): unify the calls of shrink_dentry_list()\n shrinking rcu_read_lock() scope in d_alloc_parallel()\n d_walk(): shrink rcu_read_lock() scope\n document dentry_kill()\n adjust calling conventions of lock_for_kill(), fold __dentry_kill() into dentry_kill()\n Document rcu_read_lock() use in select_collect2()\n Shift rcu_read_{,un}lock() inside fast_dput()\n simplify safety for lock_for_kill() slowpath\n fold lock_for_kill() and __dentry_kill() into common helper\n fold lock_for_kill() into shrink_kill()\n shrink_dentry_list(): start with removing from shrink list\n d_prune_aliases(): make sure to skip NORCU aliases\n kill d_dispose_if_unused()\n make to_shrink_list() return whether it has moved dentry to list\n select_collect(): ignore dentries on shrink lists if they have positive refcounts\n find_acceptable_alias(): skip NORCU aliases with zero refcount\n fix a race between d_find_any_alias() and final dput() of NORCU dentries\n ...\n" }, { "commit": "79169a1624253363fed3e9a447b77e50bb226206", "tree": "e7c5853529da0bd435d04d2445e4d2bed9050021", "parents": [ "7e0e7bd60d4a812b694c477716597fcb038b00cb", "cf30ceccfaec3d2549ff60f7c915625f12dd3a93" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 04:07:58 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 04:07:58 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.procfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull procfs updates from Christian Brauner:\n\n - Revamp fs/filesystems.c\n\n The file was a mess with a hand-rolled linked list in desperate need\n of a cleanup. The filesystems list is now RCU-ified, /proc files can\n be marked permanent from outside fs/proc/, and the string emitted\n when reading /proc/filesystems is pre-generated and cached instead of\n pointer-chasing and printfing entry by entry on every read.\n\n The file is read frequently because libselinux reads it and is linked\n into numerous frequently used programs (even ones you would not\n suspect, like sed!). Scalability also improves since reference\n maintenance on open/close is bypassed.\n\n open+read+close cycle single-threaded (ops/s):\n before: 442732\n after: 1063462 (+140%)\n\n open+read+close cycle with 20 processes (ops/s):\n before: 606177\n after: 3300576 (+444%)\n\n A follow-up patch adds missing unlocks in some corner cases and\n tidies things up.\n\n - Relax the mount visibility check for subset\u003dpid mounts\n\n When procfs is mounted with subset\u003dpid, all static files become\n unavailable and only the dynamic pid information is accessible. In\n that case there is no point in imposing the full mount visibility\n restrictions on the mounter - everything that can be hidden in procfs\n is already inaccessible. These restrictions prevented procfs from\n being mounted inside rootless containers since almost all container\n implementations overmount parts of procfs to hide certain\n directories.\n\n As part of this /proc/self/net is only shown in subset\u003dpid mounts for\n CAP_NET_ADMIN, reconfiguring subset\u003dpid is rejected, the\n SB_I_USERNS_VISIBLE superblock flag is replaced with an\n FS_USERNS_MOUNT_RESTRICTED filesystem flag, fully visible mounts are\n recorded in a list, and the mount restrictions are finally\n documented.\n\n - Protect ptrace_may_access() with exec_update_lock in procfs\n\n Most uses of ptrace_may_access() in procfs should hold\n exec_update_lock to avoid TOCTOU issues with concurrent privileged\n execve() (like setuid binary execution).\n\n This fixes the easy cases - the owner and visibility checks and the\n FD link permission checks - with the gnarlier ones to follow later.\n\n* tag \u0027vfs-7.2-rc1.procfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n fs: fix ups and tidy ups to /proc/filesystems caching\n proc: protect ptrace_may_access() with exec_update_lock (FD links)\n proc: protect ptrace_may_access() with exec_update_lock (part 1)\n docs: proc: add documentation about mount restrictions\n proc: handle subset\u003dpid separately in userns visibility checks\n proc: prevent reconfiguring subset\u003dpid\n proc: subset\u003dpid: Show /proc/self/net only for CAP_NET_ADMIN\n fs: cache the string generated by reading /proc/filesystems\n sysfs: remove trivial sysfs_get_tree() wrapper\n fs: RCU-ify filesystems list\n fs: move SB_I_USERNS_VISIBLE to FS_USERNS_MOUNT_RESTRICTED\n proc: allow to mark /proc files permanent outside of fs/proc/\n namespace: record fully visible mounts in list\n" }, { "commit": "7e0e7bd60d4a812b694c477716597fcb038b00cb", "tree": "4ff61d47485803e7dacab1c8ddef0a4c11b512da", "parents": [ "ff8747aacaff8266dd751b8a8648fb728dcc3b21", "aa5c4fe3ba0cb2af90bbcfa7a8ef4fefcd5c2370" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:59:45 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:59:45 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull misc vfs updates from Christian Brauner:\n \"Features:\n\n - Reduce pipe-\u003emutex contention by pre-allocating pages outside the\n lock in anon_pipe_write().\n\n anon_pipe_write() called alloc_page() once per page while holding\n pipe-\u003emutex. The allocation can sleep doing direct reclaim and runs\n memcg charging, which extends the critical section and stalls any\n concurrent reader on the same mutex. Now up to 8 pages are\n pre-allocated before the mutex is taken, leftovers are recycled\n into the per-pipe tmp_page[] cache before unlock, and any remainder\n is released after unlock, keeping the allocator out of the critical\n section on both sides. On a writers x readers sweep with 64KB\n writes against a 1 MB pipe throughput improves 6-28% and average\n write latency drops 5-22%; under memory pressure - when the cost of\n holding the mutex across reclaim is highest - throughput improves\n 21-48% and latency drops 17-33%. The microbenchmark is added to\n selftests.\n\n - uaccess/sockptr: fix the ignored_trailing logic in\n copy_struct_to_user() to behave as documented and the usize check\n in copy_struct_from_sockptr() for user pointers, and add\n copy_struct_{from,to}_bounce_buffer() and copy_struct_to_sockptr()\n helpers for upcoming users (IPPROTO_SMBDIRECT, IPPROTO_QUIC).\n\n - bpf: add a sleepable bpf_real_inode() kfunc that resolves the real\n inode backing a dentry via d_real_inode(). On overlayfs the inode\n attached to the dentry doesn\u0027t carry the underlying device\n information; this is used by the filesystem restriction BPF program\n that was merged into systemd.\n\n - docs: add guidelines for submitting new filesystems, motivated by\n the maintenance burden abandoned and untestable filesystems impose\n on VFS developers, blocking infrastructure work like folio\n conversions and iomap migration.\n\n Fixes:\n\n - libfs: set SB_I_NOEXEC and SB_I_NODEV by default in init_pseudo()\n and drop the now-redundant assignments in callers. This began as a\n one-line dma-buf fix for a path_noexec() warning; a pseudo\n filesystem has no reason not to set SB_I_NOEXEC. All init_pseudo()\n callers were audited: the only visible effect is on dma-buf where\n SB_I_NOEXEC silences the warning.\n\n - Handle set_blocksize() failures in legacy filesystems (bfs, hpfs,\n qnx4, jfs, befs, affs, isofs, minix, ntfs3, omfs). Mounting a\n device with a sector size \u003e PAGE_SIZE crashed roughly half of them;\n the rest had the same missing error handling pattern. Plus a\n follow-up releasing the superblock buffer_head when setting the\n minix v3 block size fails.\n\n - mount: honour SB_NOUSER in the new mount API.\n\n - fs/fcntl: fix a SOFTIRQ-unsafe lock order in fasync signaling by\n switching the process-group paths of send_sigio() and send_sigurg()\n from read_lock(\u0026tasklist_lock) to RCU, matching the single-PID\n path.\n\n - vfs: add an FS_USERNS_DELEGATABLE flag and set it for NFS, fixing\n delegated NFS mounts (fsopen() in a container with the mount\n performed by a privileged daemon) that broke when non-init\n s_user_ns was tied to FS_USERNS_MOUNT.\n\n - selftests/namespaces: fix a hang in nsid_test where an unreaped\n grandchild kept the TAP pipe write-end open, a waitpid(-1) race in\n listns_efault_test, and a false FAIL on kernels without listns()\n where the tests should SKIP.\n\n - filelock: fix the break_lease() stub signature for\n CONFIG_FILE_LOCKING\u003dn.\n\n - init/initramfs_test: wait for the async initramfs unpacking before\n running; the test and do_populate_rootfs() share the parser state.\n\n - fs/coredump: reduce redundant log noise in\n validate_coredump_safety().\n\n - iomap: pass the correct length to fserror_report_io() in\n __iomap_write_begin().\n\n - backing-file: fix the backing_file_open() kerneldoc.\n\n Cleanups:\n\n - initramfs: refactor the cpio hex header parsing to use hex2bin()\n instead of the hand-rolled simple_strntoul() which is reverted, and\n extend the initramfs KUnit tests to cover header fields with 0x\n prefixes.\n\n - Replace __get_free_pages() and friends with kmalloc()/kzalloc()\n across quota, proc, ocfs2/dlm, nilfs2, nfs, nfsd, libfs, jfs, jbd2,\n isofs, fuse, select, namespace, configfs, binfmt_misc, bfs, and the\n do_mounts init code - part of the larger work of replacing page\n allocator calls with kmalloc().\n\n - Use clear_and_wake_up_bit() in unlock_buffer() and\n journal_end_buffer_io_sync() instead of open-coding the sequence.\n\n - Drop unused VFS exports: unexport drop_super_exclusive(), remove\n start_removing_user_path_at(), and fold __start_removing_path()\n into start_removing_path().\n\n - fs/read_write: narrow the __kernel_write() export with\n EXPORT_SYMBOL_FOR_MODULES().\n\n - vfs: uapi: retire octal and hex constants in favor of (1 \u003c\u003c n) for\n the O_ flags. Finding a free bit for a new flag across the\n architectures was needlessly hard with the mixed bases.\n\n - dcache: add extra sanity checks of dead dentries in dentry_free()\n via a new DENTRY_WARN_ONCE() that also prints d_flags.\n\n - iov_iter: use kmemdup_array() in dup_iter() to harden the\n allocation against multiplication overflow.\n\n - fs/pipe: write to -\u003epoll_usage only once.\n\n - vfs: remove an always-taken if-branch in find_next_fd().\n\n - dcache: use kmalloc_flex() for struct external_name in __d_alloc().\n\n - namei: use QSTR() instead of QSTR_INIT() in path_pts().\n\n - sync_file_range: delete dead S_ISLNK code.\n\n - Comment fixes: retire a stale comment in fget_task_next() and fix\n assorted spelling mistakes\"\n\n* tag \u0027vfs-7.2-rc1.misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (73 commits)\n backing-file: fix backing_file_open() kerneldoc parameter\n iomap: pass the correct len to fserror_report_io in __iomap_write_begin\n vfs: add FS_USERNS_DELEGATABLE flag and set it for NFS\n filelock: fix break_lease() stub signature for CONFIG_FILE_LOCKING\u003dn\n vfs: uapi: retire octal and hex numbers in favor of (1 \u003c\u003c n) for O_ flags\n bpf: add bpf_real_inode() kfunc\n fs/read_write: Do not export __kernel_write() to the entire world\n libfs: drop redundant SB_I_NOEXEC/SB_I_NODEV in init_pseudo() callers\n libfs: set SB_I_NOEXEC and SB_I_NODEV by default in init_pseudo()\n mount: honour SB_NOUSER in the new mount API\n fs/fcntl: fix SOFTIRQ-unsafe lock order in fasync signaling\n selftests/pipe: add pipe_bench microbenchmark\n fs/pipe: pre-allocate pages outside pipe-\u003emutex in anon_pipe_write\n fs: retire stale comment in fget_task_next()\n fs: fix spelling mistakes in comment\n bfs: replace get_zeroed_page() with kzalloc()\n binfmt_misc: replace __get_free_page() with kmalloc()\n configfs: replace __get_free_pages() with kzalloc()\n fs/namespace: use __getname() to allocate mntpath buffer\n fs/select: replace __get_free_page() with kmalloc()\n ...\n" }, { "commit": "ff8747aacaff8266dd751b8a8648fb728dcc3b21", "tree": "2eb6f46a6c2f904de59b37d9edf5c2f1d2386010", "parents": [ "ec5d1ae94e99d8831427d00973da5620c7fb4368", "9722955b54307e9070994f2382ec06af3d7405e0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:54:54 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:54:54 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.xattr\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull simple_xattr updates from Christian Brauner:\n \"This reworks the simple xattr api to make it more efficient and easier\n to use for all consumers.\n\n The simple_xattr hash table moves from the inode into a per-superblock\n cache, removing the per-inode overhead for the common case of few or\n no xattrs. The interface now passes struct simple_xattrs ** so lazy\n allocation is handled internally instead of by every caller, kernfs\n xattr operations on kernfs nodes shared between multiple superblocks\n are properly serialized, and tmpfs constructs \"security.foo\" xattr\n names with kasprintf() instead of kmalloc() plus two memcpy()s.\n\n A follow-up fix links kernfs nodes to their parent before the LSM init\n hook runs: with the per-sb cache kernfs_xattr_set() computes the cache\n via kernfs_root(kn), which faulted on a freshly allocated node when\n selinux_kernfs_init_security() called into it - reproducible as a NULL\n pointer dereference on the first cgroup mkdir on SELinux-enabled\n systems.\n\n On top of this bpffs gains support for trusted.* and security.* xattrs\n so that user space and BPF LSM programs can attach metadata - for\n example a content hash or a security label - to pinned objects and\n directories and inspect it uniformly like on other filesystems. The\n store is in-memory and non-persistent, living only for the lifetime of\n the mount like everything else in bpffs\"\n\n* tag \u0027vfs-7.2-rc1.xattr\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n bpf: Add simple xattr support to bpffs\n kernfs: link kn to its parent before the LSM init hook\n simpe_xattr: use per-sb cache\n simple_xattr: change interface to pass struct simple_xattrs **\n tmpfs: simplify constructing \"security.foo\" xattr names\n kernfs: fix xattr race condition with multiple superblocks\n" }, { "commit": "ec5d1ae94e99d8831427d00973da5620c7fb4368", "tree": "9c2bd793eef39f6a56cd8c8c913818bb4dcc023c", "parents": [ "5b7c3f0fe36d8e867288f8e452ede82d178c757a", "b7bae6880e8de2a5f693c18d87ad5cc26f157eb2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:46:54 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:46:54 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.iomap\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull iomap updates from Christian Brauner:\n\n - Add the vfs infrastructure required to implement fs-verity support\n for XFS with a post-EOF merkle tree: fsverity generates and stores a\n zero-block hash, and iomap learns to verify data on buffered reads,\n to handle fsverity during writeback via the new IOMAP_F_FSVERITY\n flag, and to write fsverity metadata through iomap_fsverity_write().\n\n - Skip the memset of the iomap in iomap_iter() once the iteration is\n done. In high-IOPS scenarios (4k randread NVMe polling via io_uring)\n the pointless memset wasted memory write bandwidth; this improves\n IOPS by about 5% on ext4 and xfs.\n\n - Add balance_dirty_pages_ratelimited() to iomap_zero_iter(), aligning\n it with iomap_write_iter(). This prepares for the exFAT iomap\n conversion where zeroing beyond valid_size can trigger large-scale\n zeroing operations that caused memory pressure without throttling.\n\n - Remove the over-strict inline data boundary check. If a filesystem\n provides a valid inline_data pointer and length there is no reason to\n require that inline data must not cross a page boundary.\n\n - Don\u0027t make REQ_POLLED imply REQ_NOWAIT, matching the earlier\n equivalent block layer fix: there are valid cases to poll for I/O\n completion without REQ_NOWAIT, and REQ_NOWAIT for file system writes\n is currently not supported as writes aren\u0027t idempotent.\n\n - Introduce IOMAP_F_ZERO_TAIL for filesystems that maintain a separate\n valid data length (exFAT, NTFS). For a write starting at or beyond\n valid_size, __iomap_write_begin() now zeroes only the tail portion of\n the block while preserving valid data before it, instead of leaving\n stale data in the page cache. The flag is also added to the iomap\n trace event strings.\n\n* tag \u0027vfs-7.2-rc1.iomap\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n iomap: Add IOMAP_F_ZERO_TAIL flag to trace event strings\n iomap: introduce iomap_fsverity_write() for writing fsverity metadata\n iomap: teach iomap to read files with fsverity\n iomap: introduce IOMAP_F_FSVERITY and teach writeback to handle fsverity\n fsverity: generate and store zero-block hash\n iomap: introduce IOMAP_F_ZERO_TAIL flag\n iomap: don\u0027t make REQ_POLLED imply REQ_NOWAIT\n iomap: remove over-strict inline data boundary check\n iomap: add dirty page control to iomap_zero_iter\n iomap: avoid memset iomap when iter is done\n" }, { "commit": "5b7c3f0fe36d8e867288f8e452ede82d178c757a", "tree": "7fd007b26bd1d2220956471473997a8ad895b56a", "parents": [ "c8ed3a15a749246ddfedb84aab9cf0316c7b9b8a", "a1e9718b406bc4e6d0c63c7b999d06febbdc4091" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:40:54 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:40:54 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.eventpoll\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull eventpoll updates from Christian Brauner:\n\n - eventpoll clarity refactor\n\n The recent eventpoll UAF fixes (a6dc643c6931 and follow-ups) depended\n on invariants in fs/eventpoll.c that were nowhere documented and had\n to be reverse-engineered from the code: the lifetime relationships\n between struct eventpoll, struct epitem, and struct file, the three\n removal paths coordinating via epi_fget() pins and ep-\u003emtx, the\n ovflist sentinel-encoded scan state machine, the POLLFREE\n release/acquire handshake, and the loop / path check globals\n serialized by epnested_mutex. The fixes were correct but the next\n person to touch this code would hit the same learning curve.\n\n This series codifies those invariants in source and tightens the\n surrounding structure. No functional changes intended:\n\n - Documentation: a top-of-file overview with field-protection\n tables for struct eventpoll and struct epitem, a section\n gathering the loop-check / path-check globals next to their\n declarations, labelled comments on the two sides of the POLLFREE\n handshake, refreshed comments on epi_fget() and ep_remove_file(),\n and a docblock on ep_clear_and_put() that names its two-pass\n structure as load-bearing.\n\n - Mechanical renames: ep_refcount_dec_and_test() -\u003e ep_put() to\n pair with ep_get(), attach_epitem() -\u003e ep_attach_file() for\n ep_remove_file() symmetry, the unused depth argument dropped from\n epoll_mutex_lock(), and the CONFIG_KCMP block relocated next to\n CONFIG_COMPAT so the hot-path code is contiguous.\n\n - Helper extraction: ep_insert() splits into ep_alloc_epitem() and\n ep_register_epitem(), ep_clear_and_put()\u0027s two passes become\n ep_drain_pollwaits() and ep_drain_tree() so the ordering\n invariant is enforced by the call sequence rather than\n convention, the per-event delivery loop body becomes\n ep_deliver_event(), and the ep-\u003emtx + epnested_mutex acquisition\n dance lifts out of do_epoll_ctl() into ep_ctl_lock() /\n ep_ctl_unlock().\n\n - Sentinel and predicate cleanup: the EP_UNACTIVE_PTR overload is\n hidden behind named helpers (ep_is_scanning, epi_on_ovflist,\n ...), epi-\u003enext is renamed to epi-\u003eovflist_next, and the boolean\n predicates return bool.\n\n - The per-CTL_ADD scratch state (tfile_check_list, path_count[],\n inserting_into) moves from file-scope globals into a\n stack-allocated struct ep_ctl_ctx plumbed through the loop / path\n check chain.\n\n Two follow-up fixes are included: missing kernel-doc for the new @ctx\n parameters, and restoring the EP_UNACTIVE_PTR sentinel for\n ctx-\u003etfile_check_list - replacing it with NULL termination broke\n ep_remove_file()\u0027s \"never listed\" check for the list tail, causing a\n syzbot-reported use-after-free.\n\n - io_uring related epoll cleanups\n\n One of the nastier things about epoll is how it allows nesting\n contexts inside each other, leading to the necessity of loop\n detection and the issues that have come with that. There is no reason\n to support nesting on the io_uring side, so contain the damage and\n disallow nested contexts from there: eventpoll gains a file based\n control interface and struct epoll_filefd is renamed to epoll_key.\n The io_uring side proper goes on top of this through the block tree.\n\n - Fix epoll_wait() reporting false negatives\n\n ep_events_available() checks ep-\u003erdllist and ep_is_scanning() without\n a lock and can race with a concurrent scan such that neither check\n sees the events, causing epoll_wait() with a zero timeout to wrongly\n report no events even though events are available. A sequence lock\n closes the race and a reproducer is added to the eventpoll selftests.\n\n* tag \u0027vfs-7.2-rc1.eventpoll\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (25 commits)\n eventpoll: restore EP_UNACTIVE_PTR sentinel for ctx-\u003etfile_check_list\n eventpoll: Fix epoll_wait() report false negative\n selftests/eventpoll: Add test for multiple waiters\n eventpoll: add missing kernel-doc for @ctx function parameters\n eventpoll: rename struct epoll_filefd to epoll_key\n eventpoll: add file based control interface\n eventpoll: export is_file_epoll()\n eventpoll: pass struct epoll_filefd through ep_find() and ep_insert()\n eventpoll: hoist CTL_ADD scratch state into struct ep_ctl_ctx\n eventpoll: use bool for predicate helpers\n eventpoll: rename epi-\u003enext and txlist for clarity\n eventpoll: wrap EP_UNACTIVE_PTR in typed sentinel helpers\n eventpoll: extract lock dance from do_epoll_ctl() into ep_ctl_lock()\n eventpoll: extract ep_deliver_event() from ep_send_events()\n eventpoll: split ep_clear_and_put() into drain helpers\n eventpoll: split ep_insert() into alloc + register stages\n eventpoll: relocate KCMP helpers near compat syscalls\n eventpoll: rename attach_epitem() to ep_attach_file()\n eventpoll: drop unused depth argument from epoll_mutex_lock()\n eventpoll: rename ep_refcount_dec_and_test() to ep_put()\n ...\n" }, { "commit": "c8ed3a15a749246ddfedb84aab9cf0316c7b9b8a", "tree": "d4e216fa54148d4e0c34d500a7dc4dc386184258", "parents": [ "c17fdf62aeecbbaf2c2fd5c494e2089c02b0e75b", "f0d857543e4d37464759c338f46ad6c85a618a2e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:36:08 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:36:08 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.bh\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull buffer_head updates from Christian Brauner:\n \"This removes b_end_io from struct buffer_head.\n\n Instead of setting bio-\u003ebi_end_io to end_bio_bh_io_sync() which then\n calls bh-\u003eb_end_io(), the new bh_submit() and __bh_submit() interfaces\n set bio-\u003ebi_end_io to the appropriate completion handler directly,\n replacing two indirect function calls in the completion path with one.\n It is also one fewer function pointer in the middle of a writable data\n structure that can be corrupted, it shrinks struct buffer_head from\n 104 to 96 bytes allowing roughly 7% more buffer_heads to be cached in\n the same amount of memory, and it removes some atomic operations as\n the buffer refcount is no longer incremented before calling the end_io\n handler.\n\n All in-tree users (fs/buffer.c itself, ext4, jbd2, ocfs2, gfs2,\n nilfs2, and md-bitmap) are converted, and submit_bh(),\n mark_buffer_async_write(), and end_buffer_write_sync() are removed\"\n\n* tag \u0027vfs-7.2-rc1.bh\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (34 commits)\n buffer: Remove end_buffer_write_sync()\n buffer: Change calling convention for end_buffer_read_sync()\n buffer: Remove b_end_io\n buffer: Remove submit_bh()\n md-bitmap: Convert read_file_page and write_file_page to bh_submit()\n nilfs2: Convert nilfs_mdt_submit_block to bh_submit()\n nilfs2: Convert nilfs_gccache_submit_read_data to bh_submit()\n nilfs2: Convert nilfs_btnode_submit_block to bh_submit()\n buffer: Remove mark_buffer_async_write()\n gfs2: Convert gfs2_aspace_write_folio to bh_submit()\n gfs2: Remove use of b_end_io in gfs2_meta_read_endio()\n gfs2: Convert gfs2_dir_readahead to bh_submit()\n gfs2: Convert gfs2_metapath_ra to bh_submit()\n ocfs2: Convert ocfs2_write_super_or_backup to bh_submit()\n ocfs2: Convert ocfs2_read_blocks to bh_submit()\n ocfs2: Convert ocfs2_read_block to bh_submit()\n ocfs2: Convert ocfs2_write_block to bh_submit()\n jbd2: Convert jbd2_write_superblock() to bh_submit()\n jbd2: Convert journal commit to bh_submit()\n ext4: Convert ext4_commit_super() to bh_submit()\n ...\n" }, { "commit": "c17fdf62aeecbbaf2c2fd5c494e2089c02b0e75b", "tree": "2cb320b5bc6f1c97da837e8cb43352a72e789267", "parents": [ "0793d39ec8bab2b2255e3a288894c39e88ce5a75", "0275dc184aa007b260374af6d46fb15741c062a8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:30:45 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:30:45 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.writeback\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs writeback updates from Christian Brauner:\n\n - Fix a race between cgroup_writeback_umount() and inode_switch_wbs()\n\n When a container exits, a race between cgroup_writeback_umount() and\n inode_switch_wbs()/cleanup_offline_cgwb() can trigger \"VFS: Busy\n inodes after unmount\" followed by a use-after-free on percpu\n counters.\n\n There is a window between inode_prepare_wbs_switch() returning true\n (having passed the SB_ACTIVE check and grabbed the inode) and the\n subsequent wb_queue_isw() call: if cgroup_writeback_umount() observes\n the global isw_nr_in_flight counter as non-zero but flush_workqueue()\n finds nothing queued yet, it returns early - leaving a held inode\n reference that blocks evict_inodes() and a later iput() that hits\n freed percpu counters.\n\n The race is closed by covering the window from\n inode_prepare_wbs_switch() through wb_queue_isw() with an RCU\n read-side critical section and synchronizing in the umount path.\n\n On top of that the now-dead rcu_barrier() left over from the\n queue_rcu_work() era is removed, and the global\n synchronize_rcu()/flush_workqueue() pair is replaced with a per-sb\n in-flight counter plus pin/unpin/drain helpers so umount no longer\n serializes against switch activity on unrelated superblocks.\n\n Under cgroup writeback churn on a 16 vCPU guest this takes umount\n latency from ~92-138ms p50 down to ~5-8ms p50 and the cumulative cost\n of cgroup_writeback_umount() from ~62ms to ~4us per call.\n\n The initial race fix is kept separate and minimal so it backports\n cleanly to stable trees that still queue switches via\n queue_rcu_work().\n\n - Improve write performance with RWF_DONTCACHE\n\n Dirty DONTCACHE pages are now tracked per bdi_writeback so that the\n writeback flusher can be kicked in a targeted fashion for\n IOCB_DONTCACHE writes instead of relying on global writeback, and the\n PG_dropbehind flag is preserved when a folio is split.\n\n* tag \u0027vfs-7.2-rc1.writeback\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n mm: kick writeback flusher for IOCB_DONTCACHE with targeted dirty tracking\n mm: track DONTCACHE dirty pages per bdi_writeback\n mm: preserve PG_dropbehind flag during folio split\n writeback: use a per-sb counter to drain inode wb switches at umount\n writeback: drop now-unnecessary rcu_barrier() in cgroup_writeback_umount()\n writeback: fix race between cgroup_writeback_umount() and inode_switch_wbs()\n" }, { "commit": "0793d39ec8bab2b2255e3a288894c39e88ce5a75", "tree": "7dfb1928b48c9a4081be709a410cde4bbc3a09a8", "parents": [ "50b900c564b0f0307c126de9f17c21d3a1ba039b", "2c6f0c248a6b49a6fc8c301c84d367860c56ccd8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:25:36 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:25:36 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.super\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs superblock updates from Christian Brauner:\n \"This retires sget().\n\n CIFS plus the two ext4 KUnit tests (extents-test, mballoc-test) were\n the last in-tree callers, and all three convert cleanly to sget_fc().\n\n That lets sget() and its prototype come out, taking ~60 lines that\n only existed to be kept in lockstep with sget_fc() on every\n publish-path change\"\n\n* tag \u0027vfs-7.2-rc1.super\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n fs: retire sget()\n smb: client: convert cifs_smb3_do_mount() to sget_fc()\n ext4: convert mballoc KUnit test to sget_fc()\n ext4: convert extents KUnit test to sget_fc()\n" }, { "commit": "50b900c564b0f0307c126de9f17c21d3a1ba039b", "tree": "59113f35198a8c9b1dede2fd17d0954d798db6e9", "parents": [ "37c405aeaa5c2cbe04c3c727e3989a16a2e9f30f", "318643721de396012da102723f337f35ba7ec1e9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:11:05 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:11:05 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.openat2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull openat2 updates from Christian Brauner:\n \"Features:\n\n - Add O_EMPTYPATH to openat(2)/openat2(2). To get an operable file\n descriptor from an O_PATH file descriptor it is possible to use\n openat(fd, \".\", O_DIRECTORY) for directories, but other file types\n require going through open(\"/proc/\u003cpid\u003e/fd/\u003cnr\u003e\") and thus depend\n on a functioning procfs.\n\n With O_EMPTYPATH an empty path string is accepted and LOOKUP_EMPTY\n is set at path resolution time, allowing to reopen the file behind\n the file descriptor directly. Selftests are included.\n\n - Add an OPENAT2_REGULAR flag for openat2(2) which refuses to open\n anything but regular files with the new EFTYPE error code.\n\n This implements the \"ability to only open regular files\" feature\n requested by userspace via uapi-group.org and protects services\n from being redirected to fifos, device nodes, and friends.\n\n All atomic_open implementations were audited for OPENAT2_REGULAR\n handling. Explicit checks were added to ceph, gfs2, nfs (v4), and\n cifs/smb - these are the filesystems whose atomic_open can\n encounter an existing non-regular file and would otherwise call\n finish_open() on it or return a misleading error code.\n\n The remaining implementations (9p, fuse, vboxsf, nfs v2/v3) only\n call finish_open() on freshly created files and use\n finish_no_open() for lookup hits, letting the VFS catch non-regular\n files via the do_open() safety net.\n\n Cleanups:\n\n - Migrate the openat2 selftests to the kselftest harness and move\n them under selftests/filesystems/. The tests were written in the\n early days of selftests\u0027 TAP support and the modern kselftest\n harness is much easier to follow and maintain. The contents of the\n tests are unchanged and the new emptypath tests are ported on top.\n\n - Make the LAST_XXX last-type constants private to fs/namei.c. The\n only user outside of fs/namei.c was ksmbd which only needs to know\n whether the last component is a regular one, so\n vfs_path_parent_lookup() now performs the LAST_NORM check\n internally. The ints are replaced with a dedicated enum last_type\"\n\n* tag \u0027vfs-7.2-rc1.openat2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n vfs: replace ints with enum last_type for LAST_XXX\n vfs: make LAST_XXX private to fs/namei.c\n selftests: openat2: port emptypath_test to kselftest harness\n kselftest/openat2: test for OPENAT2_REGULAR flag\n openat2: new OPENAT2_REGULAR flag support\n openat2: introduce EFTYPE error code\n selftest: add tests for O_EMPTYPATH\n vfs: add O_EMPTYPATH to openat(2)/openat2(2)\n selftests: openat2: migrate to kselftest harness\n selftests: openat2: switch from custom ARRAY_LEN to ARRAY_SIZE\n selftests: openat2: move helpers to header\n selftests: move openat2 tests to selftests/filesystems/\n" }, { "commit": "37c405aeaa5c2cbe04c3c727e3989a16a2e9f30f", "tree": "88594b10c60e16fd03223136a230ed72e42fce62", "parents": [ "9c9e6bd4cca02f2d183eb260451fb6018f9ee67e", "ee8ab98f831226d69d43ccd93f53c50e6f19b389" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:05:50 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:05:50 2026 +0530" }, "message": "Merge tag \u0027kernel-7.2-rc1.misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull misc kernel updates from Christian Brauner:\n \"Fixes\n\n - rhashtable: give each instance its own lockdep class\n\n syzbot reported a circular locking dependency between ht-\u003emutex and\n fs_reclaim via the simple_xattrs rhashtable being torn down during\n inode eviction.\n\n The predicted deadlock cannot occur: rhashtable_free_and_destroy()\n cancels the deferred worker before taking ht-\u003emutex and\n acquisitions on distinct rhashtables are on distinct mutexes.\n\n Lockdep flags a cycle anyway because every ht-\u003emutex in the kernel\n shared the single static lockdep class from\n rhashtable_init_noprof().\n\n The lockdep key is lifted to a per-call-site static key so every\n rhashtable instance gets its own class.\n\n - selftests/clone3: fix misuse of the libcap library interface in the\n cap_checkpoint_restore test and remove unused variables\n\n - selftests/pid_namespace: compute the pid_max test limits\n dynamically instead of hardcoding values below the kernel-enforced\n minimum of PIDS_PER_CPU_MIN * num_possible_cpus() which made the\n tests fail on machines with many possible CPUs\n\n - selftests: fix the Makefile TARGETS entry for nsfs which wasn\u0027t\n adjusted when the tests moved under filesystems/\n\n Cleanups\n\n - ipc/sem.c: use unsigned int for nsops to match the declaration in\n syscalls.h\"\n\n* tag \u0027kernel-7.2-rc1.misc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n selftests/clone3: remove unused variables\n selftests/clone3: fix libcap interface usage\n ipc/sem.c: use unsigned int for nsops\n selftests: Fix Makefile target for nsfs\n rhashtable: give each instance its own lockdep class\n selftests/pid_namespace: compute pid_max test limits dynamically\n" }, { "commit": "9c9e6bd4cca02f2d183eb260451fb6018f9ee67e", "tree": "5c351e1db3b13ffce9ce02c95173e58d35975d1f", "parents": [ "5d15ab717d503ff10b585a144870648b9a88c616", "38205ecbe6b6dc47968ad4e9c978e2117720969e" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:00:58 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 03:00:58 2026 +0530" }, "message": "Merge tag \u0027kernel-7.2-rc1.task_exec_state\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull task_exec_state updates from Christian Brauner:\n \"This introduces a new per-task task_exec_state structure and relocates\n the dumpable mode and the user namespace captured at execve() from\n mm_struct onto it. It stays attached to the task for its full\n lifetime.\n\n __ptrace_may_access() and several /proc owner and visibility checks\n need to consult two pieces of state for any observable task, including\n zombies that have already gone through exit_mm(): the dumpable mode\n and the user namespace captured at execve(). Both live on mm_struct\n today, which exit_mm() clears from the task long before the task is\n reaped. A reader that races with do_exit() observes task-\u003emm \u003d\u003d NULL\n and either fails the check or falls back to init_user_ns - which\n denies legitimate access to non-dumpable zombies that were running in\n a nested user namespace.\n\n mm_struct loses -\u003euser_ns and the dumpability bits in -\u003eflags.\n MMF_DUMPABLE_BITS is reserved so the MMF_DUMP_FILTER_* layout exposed\n via /proc/\u003cpid\u003e/coredump_filter stays stable. task-\u003euser_dumpable and\n its exit_mm() snapshot are removed.\n\n task_exec_state is the privilege domain established by an execve().\n Within a thread group it is shared via refcount; across thread groups\n each task has its own:\n\n - CLONE_VM siblings (thread-group members, io_uring workers)\n refcount-share the parent\u0027s exec_state.\n\n - Non-CLONE_VM clones (fork(), vfork() without CLONE_VM) allocate a\n fresh exec_state inheriting the parent\u0027s dumpable mode and user_ns.\n\n - execve() in the child allocates a fresh instance and installs it\n under task_lock + exec_update_lock via task_exec_state_replace().\n\n - Credential changes (setresuid, capset, ...) and\n prctl(PR_SET_DUMPABLE) update dumpability on the current task\u0027s\n exec_state, i.e., on the thread group\u0027s shared instance.\n\n On top of this exec_mmap() no longer tears down the old mm while\n holding exec_update_lock for writing and cred_guard_mutex. Neither\n lock is needed for that: exec_update_lock only exists to make the mm\n swap atomic with the later commit_creds() and all its readers operate\n on the new mm; none looks at the detached old mm.\n\n The cost was real: __mmput() runs exit_mmap() over the entire old\n address space and can block in exit_aio() waiting for in-flight AIO,\n so execve() of a large process blocked ptrace_attach() and every\n exec_update_lock reader for the duration of the teardown.\n\n The old mm is now stashed in bprm-\u003eold_mm and released from\n setup_new_exec() after both locks are dropped, with a backstop in\n free_bprm() for the error paths\"\n\n* tag \u0027kernel-7.2-rc1.task_exec_state\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n exec: free the old mm outside the exec locks\n exec_state: relocate dumpable information\n ptrace: add ptracer_access_allowed()\n exec: introduce struct task_exec_state\n sched/coredump: introduce enum task_dumpable\n" }, { "commit": "5d15ab717d503ff10b585a144870648b9a88c616", "tree": "9051bc55df8fd4ecde5736e36a25b2a88518f9fd", "parents": [ "6b5b72ac2c6383e423144beb257f98359b966a3b", "ea3120fd5153c967efb20e6e3330caecbf9d8b0a" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:55:34 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:55:34 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.casefold\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs casefolding updates from Christian Brauner:\n \"This exposes the case folding behavior of local filesystems so that\n file servers - nfsd, ksmbd, and user space file servers - can report\n the actual behavior to clients instead of guessing.\n\n Filesystems report case-insensitive and case-nonpreserving behavior\n via new file_kattr flags in their fileattr_get implementations. fat,\n exfat, ntfs3, hfs, hfsplus, xfs, cifs, nfs, vboxsf, and isofs are\n wired up. Local filesystems that are not explicitly handled default to\n the usual POSIX behavior of case-sensitive and case-preserving.\n\n nfsd uses this to report case folding via NFSv3 PATHCONF and to\n implement the NFSv4 FATTR4_CASE_INSENSITIVE and FATTR4_CASE_PRESERVING\n attributes - both have been part of the NFS protocols for decades to\n support clients on non-POSIX systems - and ksmbd reports it via\n FS_ATTRIBUTE_INFORMATION. Exposing the information through the\n fileattr uapi covers user space file servers.\n\n The immediate motivation is interoperability: Windows NFS clients\n hard-require servers to report case-insensitivity for Win32\n applications to work correctly, and a client that knows the server is\n case-insensitive can avoid issuing multiple LOOKUP/READDIR requests\n searching for case variants.\n\n The Linux NFS client already grew support for case-insensitive shares\n years ago in support of the Hammerspace NFS server - negative dentry\n caching must be disabled (a lookup for \"FILE.TXT\" failing must not\n cache a negative entry when \"file.txt\" exists) and directory change\n invalidation must drop cached case-folded name variants. Such servers\n often operate in multi-protocol environments where a single file\n service instance caters to both NFS and SMB clients, and nfsd needs to\n report case folding properly to participate as a first-class citizen\n there.\n\n A follow-up series brings fixes for the initial work: the nfsd\n case-info probe now uses kernel credentials, maps -ESTALE to\n NFS3ERR_STALE, and has its cost capped across READDIR entries; the nfs\n client avoids transiently zeroed case capability bits during the probe\n and skips the pathconf probe when neither field is consumed; the\n FS_CASEFOLD_FL semantics are clarified in the UAPI header; and the\n tools UAPI headers are synced\"\n\n* tag \u0027vfs-7.2-rc1.casefold\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (22 commits)\n nfsd: Cap case-folding probe cost across READDIR entries\n nfsd: Map -ESTALE from case probe to NFS3ERR_STALE\n nfsd: Use kernel credentials for case-info probe\n fs: Clarify FS_CASEFOLD_FL semantics in UAPI header\n nfs: Skip pathconf probe when neither field is consumed\n nfs: Avoid transient zeroed case capability bits during probe\n tools headers UAPI: Sync case-sensitivity flags from linux/fs.h\n ksmbd: Report filesystem case sensitivity via FS_ATTRIBUTE_INFORMATION\n nfsd: Implement NFSv4 FATTR4_CASE_INSENSITIVE and FATTR4_CASE_PRESERVING\n nfsd: Report export case-folding via NFSv3 PATHCONF\n isofs: Implement fileattr_get for case sensitivity\n vboxsf: Implement fileattr_get for case sensitivity\n nfs: Implement fileattr_get for case sensitivity\n cifs: Implement fileattr_get for case sensitivity\n xfs: Report case sensitivity in fileattr_get\n hfsplus: Report case sensitivity in fileattr_get\n hfs: Implement fileattr_get for case sensitivity\n ntfs3: Implement fileattr_get for case sensitivity\n exfat: Implement fileattr_get for case sensitivity\n fat: Implement fileattr_get for case sensitivity\n ...\n" }, { "commit": "6b5b72ac2c6383e423144beb257f98359b966a3b", "tree": "f9ae46d5db30eb2c3f6db50e08074fcc8258b113", "parents": [ "fac863c887a05d7c3091c5eccf30c89c2116ae11", "246bc86d0fd891273a8502314f158eab23af823c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:50:44 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:50:44 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.directory.delegations\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs directory delegations from Christian Brauner:\n \"This contains the VFS prerequisites for supporting directory\n delegations in nfsd via CB_NOTIFY callbacks.\n\n The filelock core gains support for ignoring delegation breaks for\n directory change events together with an inode_lease_ignore_mask()\n helper, and fsnotify gains fsnotify_modify_mark_mask() and a\n FSNOTIFY_EVENT_RENAME data type.\n\n With this in place nfsd can request delegations on directories and set\n up inotify watches to trigger sending CB_NOTIFY events to clients\n instead of having every directory change break the delegation.\n\n New tracepoints are added to fsnotify() and to the start of\n break_lease(), and trace_break_lease_block() is passed the currently\n blocking lease instead of the new one.\n\n A follow-up fix moves the LEASE_BREAK_* flags out of\n #ifdef CONFIG_FILE_LOCKING to fix the build for CONFIG_FILE_LOCKING\u003dn\n configurations\"\n\n* tag \u0027vfs-7.2-rc1.directory.delegations\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n filelock: move LEASE_BREAK_* flags out of #ifdef CONFIG_FILE_LOCKING\n fsnotify: add FSNOTIFY_EVENT_RENAME data type\n fsnotify: add fsnotify_modify_mark_mask()\n fsnotify: new tracepoint in fsnotify()\n filelock: add an inode_lease_ignore_mask helper\n filelock: add a tracepoint to start of break_lease()\n filelock: add support for ignoring deleg breaks for dir change events\n filelock: pass current blocking lease to trace_break_lease_block() rather than \"new_fl\"\n" }, { "commit": "fac863c887a05d7c3091c5eccf30c89c2116ae11", "tree": "484ba400219e0063b71009b35141d0f6150ba9f9", "parents": [ "4d23bdc4988efc22904c78ca36ebeef5deb950db", "5b451b76c85c8309d2e02caa467b38f5999c986f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:44:23 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:44:23 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.inode\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull vfs inode updates from Christian Brauner:\n \"This extends the lockless -\u003ei_count handling.\n\n iput() could already decrement any value greater than one locklessly\n but acquiring a reference always required taking inode-\u003ei_lock. Now\n acquiring a reference is lockless as long as the count was already at\n least 1, i.e., only the 0-\u003e1 and 1-\u003e0 transitions take the lock.\n\n This avoids the lock for the common cases of nfs calling into the\n inode hash and btrfs using igrab(). Cleanup-wise icount_read_once() is\n added to line up with inode_state_read_once() and the open-coded\n -\u003ei_count loads across the tree are converted, and ihold() is\n relocated and tidied up.\n\n On top of that some stale lock ordering annotations are retired from\n the inode hash code: iunique() no longer takes the hash lock since the\n inode hash became RCU-searchable and s_inode_list_lock is no longer\n taken under the hash lock either\"\n\n* tag \u0027vfs-7.2-rc1.inode\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n fs: retire stale lock ordering annotations from inode hash\n fs: allow lockless -\u003ei_count bumps as long as it does not transition 0-\u003e1\n fs: relocate and tidy up ihold()\n fs: add icount_read_once() and stop open-coding -\u003ei_count loads\n" }, { "commit": "4d23bdc4988efc22904c78ca36ebeef5deb950db", "tree": "5ff6795a547f44257a989f3cb8f2b13172adf65c", "parents": [ "d2cb5e633c6e8f2fb9af57e72fb05899952b1e3c", "79e33ddc62c03cce6c29f0792454e1d618228acf" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:38:54 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:38:54 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.exportfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull exportfs updates from Christian Brauner:\n \"This cleans up the exportfs support for block-style layouts that\n provide direct block device access: the operations for layout-based\n block device access are split out of struct export_operations into a\n separate header, -\u003ecommit_blocks() no longer takes a struct iattr\n argument, and the way support for layout-based block device access is\n detected is reworked.\n\n nfsd\u0027s blocklayout code also stops honoring loca_time_modify. This is\n preparation for supporting export of more than a single device per\n file system\"\n\n* tag \u0027vfs-7.2-rc1.exportfs\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n exportfs,nfsd: rework checking for layout-based block device access support\n exportfs: don\u0027t pass struct iattr to -\u003ecommit_blocks\n exportfs: split out the ops for layout-based block device access\n nfsd/blocklayout: always ignore loca_time_modify\n" }, { "commit": "d2cb5e633c6e8f2fb9af57e72fb05899952b1e3c", "tree": "26c291704709fd0ad60d6898c5a5052d9402dcdf", "parents": [ "8cd9520d35a6c38db6567e97dd93b1f11f185dc6", "07410646f6ff1d23222f105ccab778957d401bbe" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:34:37 2026 +0530" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 15 02:34:37 2026 +0530" }, "message": "Merge tag \u0027vfs-7.2-rc1.kfunc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs\n\nPull bpf filesystem kfunc fix from Christian Brauner:\n \"The bpf_set_dentry_xattr() and bpf_remove_dentry_xattr() kfuncs locked\n the inode of the supplied dentry without checking whether the dentry\n is negative.\n\n Passing a negative dentry (e.g., from security_inode_create) caused a\n NULL pointer dereference. Negative dentries now fail with EINVAL. The\n WARN_ON(!inode) in the bpf xattr permission helpers is dropped as well\n since it could be triggered the same way, amounting to a denial of\n service on systems with panic_on_warn enabled\"\n\n* tag \u0027vfs-7.2-rc1.kfunc\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:\n bpf: fix crash in bpf_[set|remove]_dentry_xattr for negative dentries\n" }, { "commit": "8cd9520d35a6c38db6567e97dd93b1f11f185dc6", "tree": "96667ed2997d7b49f3e9a36c4297b1faca4d87e3", "parents": [ "424280953322cf66314f3ba5e2d1ef345f21c770" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 14 15:58:38 2026 +0100" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 14 15:58:38 2026 +0100" }, "message": "Linux 7.1\n" }, { "commit": "424280953322cf66314f3ba5e2d1ef345f21c770", "tree": "b60272675976be52c6ae25c2eb688f4f24f49b9d", "parents": [ "e21ee273e6fa3879aec9a27251cfce98156e07c4", "009b6c6486b94a3aff566b017256b598dc96bf18" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 14 15:37:39 2026 +0100" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 14 15:37:39 2026 +0100" }, "message": "Merge tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rmk/linux\n\nPull ARM fixes from Russell King:\n\n - Avoid KASAN instrumentation of half-word IO\n\n - Use a byte load for KASAN shadow stack\n\n - Fix kexec and hibernation with PAN\n\n* tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rmk/linux:\n ARM: 9476/1: mm: fix kexec and hibernation with CONFIG_CPU_TTBR0_PAN\n ARM: 9475/1: entry: use byte load for KASAN VMAP stack shadow\n ARM: 9474/1: io: avoid KASAN instrumentation of raw halfword I/O\n" }, { "commit": "61c19a9feb1d87156e46e38d7759f3ad23710e24", "tree": "3464672fcc338240b206d6ee4ca844fb29cdc95a", "parents": [ "68d619cbde1de8c55e11ad8e08a2c418c769ce4c", "72f7be0c2e303be591ace9312c51d22d7c2d6882" ], "author": { "name": "Will Deacon", "email": "will@kernel.org", "time": "Sun Jun 14 12:18:27 2026 +0100" }, "committer": { "name": "Will Deacon", "email": "will@kernel.org", "time": "Sun Jun 14 12:18:27 2026 +0100" }, "message": "Merge branch \u0027for-next/sysregs\u0027 into for-next/core\n\n* for-next/sysregs:\n arm64/sysreg: Add HDBSS related register information\n" }, { "commit": "68d619cbde1de8c55e11ad8e08a2c418c769ce4c", "tree": "27e56af2693a33efbebc5f8650e6648195a3d76a", "parents": [ "79809fca3488a7e86f63d2a705c8082edd15b310", "7da70f41da81ebe624ac4c222792693be7c395d2" ], "author": { "name": "Will Deacon", "email": "will@kernel.org", "time": "Sun Jun 14 12:18:18 2026 +0100" }, "committer": { "name": "Will Deacon", "email": "will@kernel.org", "time": "Sun Jun 14 12:18:18 2026 +0100" }, "message": "Merge branch \u0027for-next/selftests\u0027 into for-next/core\n\n* for-next/selftests:\n kselftest/arm64: Add 2025 dpISA coverage to hwcaps\n kselftest/arm64: Add tests for POR_EL0 save/reset/restore\n kselftest/arm64: Move/add POE helpers to test_signals_utils.h\n kselftest/arm64: Add POE as a feature in the signal tests\n selftests/mm: Fix resv_sz when parsing arm64 signal frame\n" }, { "commit": "79809fca3488a7e86f63d2a705c8082edd15b310", "tree": "f9e52899a9bf7e748b9bc7b76cc458f888224fea", "parents": [ "3594da9b7a7cf690d896310d46bf15cebcf540df", "5936245125f78d896fdb1bbc2ae79213e28a6579" ], "author": { "name": "Will Deacon", "email": "will@kernel.org", "time": "Sun Jun 14 12:18:12 2026 +0100" }, "committer": { "name": "Will Deacon", "email": "will@kernel.org", "time": "Sun Jun 14 12:18:12 2026 +0100" }, "message": "Merge branch \u0027for-next/perf\u0027 into for-next/core\n\n* for-next/perf:\n perf/arm-cmn: Fix DVM node events\n perf: qcom: Unify user-visible \"Qualcomm\" name\n MAINTAINERS: Update HiSilicon PMU driver maintainer to Yushan Wang\n" }, { "commit": "3594da9b7a7cf690d896310d46bf15cebcf540df", "tree": "5e614581fae2b7ba7edc67c06e41d307be297ecf", "parents": [ "df3daf49a8a05751b34ec40c46ac6d9c47e9ec22", "50a42e03cdbd77d93366d301db3d367dce78eda6" ], "author": { "name": "Will Deacon", "email": "will@kernel.org", "time": "Sun Jun 14 12:18:06 2026 +0100" }, "committer": { "name": "Will Deacon", "email": "will@kernel.org", "time": "Sun Jun 14 12:18:06 2026 +0100" }, "message": "Merge branch \u0027for-next/mpam\u0027 into for-next/core\n\n* for-next/mpam:\n arm_mpam: Update architecture version check for MPAM MSC\n arm64: cpufeature: Add support for the MPAM v0.1 architecture version\n" } ], "next": "df3daf49a8a05751b34ec40c46ac6d9c47e9ec22" }