When a task is migrated out, there is a probability that the tg->load_avg
value will become abnormal. The reason is as follows:
1. Due to the 1ms update period limitation in update_tg_load_avg(), there
is a possibility that the reduced load_avg is not updated to tg->load_avg
when a task migrates out.
2. Even though __update_blocked_fair() traverses the leaf_cfs_rq_list and
calls update_tg_load_avg() for cfs_rqs that are not fully decayed, the key
function cfs_rq_is_decayed() does not check whether
cfs->tg_load_avg_contrib is null. Consequently, in some cases,
__update_blocked_fair() removes cfs_rqs whose avg.load_avg has not been
updated to tg->load_avg.
Add a check of cfs_rq->tg_load_avg_contrib in cfs_rq_is_decayed(),
which fixes the case (2.) mentioned above.
Fixes: 1528c661c2 ("sched/fair: Ratelimit update to tg->load_avg")
Signed-off-by: xupengbo <xupengbo@oppo.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Aaron Lu <ziqianlu@bytedance.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Tested-by: Aaron Lu <ziqianlu@bytedance.com>
Link: https://patch.msgid.link/20250827022208.14487-1-xupengbo@oppo.com
rt_mutex_setprio() has only one caller: rt_mutex_adjust_prio(). It
expects that task_struct::pi_lock and rt_mutex_base::wait_lock are held.
Both locks are raw_spinlock_t and are acquired with disabled interrupts.
Nevertheless rt_mutex_setprio() disables preemption while invoking
__balance_callbacks() and raw_spin_rq_unlock(). Even if one of the
balance callbacks unlocks the rq then it must not enable interrupts
because rt_mutex_base::wait_lock is still locked.
Therefore interrupts should remain disabled and disabling preemption is
not needed.
Commit 4c9a4bc89a ("sched: Allow balance callbacks for check_class_changed()")
adds a preempt-disable section to rt_mutex_setprio() and
__sched_setscheduler(). In __sched_setscheduler() the preemption is
disabled before rq is unlocked and interrupts enabled but I don't see
why it makes a difference in rt_mutex_setprio().
Remove the preempt_disable() section from rt_mutex_setprio().
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://patch.msgid.link/20251127155529.t_sTatE4@linutronix.de
The sched_class::task_tick() method is called on the donor
sched_class, and sched_tick() hands it rq->donor as argument,
which is consistent.
However, while hrtick() uses the donor sched_class, it then passes
rq->curr, which is inconsistent. Fix it.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: John Stultz <jstultz@google.com>
Link: https://patch.msgid.link/20250918080205.442967033@infradead.org
Pull core irq cleanup from Thomas Gleixner:
"Tree wide cleanup of the remaining users of in_irq() which got
replaced by in_hardirq() and marked deprecated in 2020"
* tag 'core-core-2025-12-03' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
treewide: Remove in_irq()
Pull timer core updates from Thomas Gleixner:
- Prevent a thundering herd problem when the timekeeper CPU is delayed
and a large number of CPUs compete to acquire jiffies_lock to do the
update. Limit it to one CPU with a separate "uncontended" atomic
variable.
- A set of improvements for the timer migration mechanism:
- Support imbalanced NUMA trees correctly
- Support dynamic exclusion of CPUs from the migrator duty to allow
the cpuset/isolation mechanism to exclude them from handling
timers of remote idle CPUs
- The usual small updates, cleanups and enhancements
* tag 'timers-core-2025-11-30' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
timers/migration: Exclude isolated cpus from hierarchy
cpumask: Add initialiser to use cleanup helpers
sched/isolation: Force housekeeping if isolcpus and nohz_full don't leave any
cgroup/cpuset: Rename update_unbound_workqueue_cpumask() to update_isolation_cpumasks()
timers/migration: Use scoped_guard on available flag set/clear
timers/migration: Add mask for CPUs available in the hierarchy
timers/migration: Rename 'online' bit to 'available'
selftests/timers/nanosleep: Add tests for return of remaining time
selftests/timers: Clean up kernel version check in posix_timers
time: Fix a few typos in time[r] related code comments
time: tick-oneshot: Add missing Return and parameter descriptions to kernel-doc
hrtimer: Store time as ktime_t in restart block
timers/migration: Remove dead code handling idle CPU checking for remote timers
timers/migration: Remove unused "cpu" parameter from tmigr_get_group()
timers/migration: Assert that hotplug preparing CPU is part of stable active hierarchy
timers/migration: Fix imbalanced NUMA trees
timers/migration: Remove locking on group connection
timers/migration: Convert "while" loops to use "for"
tick/sched: Limit non-timekeeper CPUs calling jiffies update
Pull MSI updates from Thomas Gleixner:
"Updates for [PCI] MSI related code:
- Remove one variant of PCI/MSI management as all users have been
converted to use per device domains. That reduces the variants to
two:
The modern and the real archaic legacy variant, which keeps the
usual suspects in the museum category alive.
- Rework the platform MSI device ID detection mechanism in the ARM
GIC world to address resource leaks, duplicated code and other
details. This requires a corresponding preparatory step in the
PCI/iproc driver.
- Trivial core code cleanups"
* tag 'irq-msi-2025-11-30' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
irqchip/gic-its: Rework platform MSI deviceID detection
PCI: iproc: Implement MSI controller node detection with of_msi_xlate()
genirq/msi: Slightly simplify msi_domain_alloc()
PCI/MSI: Delete pci_msi_create_irq_domain()
Pull irq core updates from Thomas Gleixner:
"Updates for the interrupt core and treewide cleanups:
- Rework of the Per Processor Interrupt (PPI) management on ARM[64]
PPI support was built under the assumption that the systems are
homogenous so that the same CPU local device types are connected to
them. That's unfortunately wishful thinking and created horrible
workarounds.
This rework provides affinity management for PPIs so that they can
be individually configured in the firmware tables and mops up the
related drivers all over the place.
- Prevent CPUSET/isolation changes to arbitrarily affine interrupt
threads to random CPUs, which ignores user or driver settings.
- Plug a harmless race in the interrupt affinity proc interface,
which allows to see a half updated mask
- Adjust the priority of secondary interrupt threads on RT, so that
the combination of primary and secondary thread emulates the
hardware interrupt plus thread scenario. Having them at the same
priority can cause starvation issues in some drivers"
* tag 'irq-core-2025-11-30' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (33 commits)
genirq: Remove cpumask availability check on kthread affinity setting
genirq: Fix interrupt threads affinity vs. cpuset isolated partitions
genirq: Prevent early spurious wake-ups of interrupt threads
genirq: Use raw_spinlock_irq() in irq_set_affinity_notifier()
genirq/manage: Reduce priority of forced secondary interrupt handler
genirq/proc: Fix race in show_irq_affinity()
genirq: Fix percpu_devid irq affinity documentation
perf: arm_pmu: Kill last use of per-CPU cpu_armpmu pointer
irqdomain: Kill of_node_to_fwnode() helper
genirq: Kill irq_{g,s}et_percpu_devid_partition()
irqchip: Kill irq-partition-percpu
irqchip/apple-aic: Drop support for custom PMU irq partitions
irqchip/gic-v3: Drop support for custom PPI partitions
coresight: trbe: Request specific affinities for per CPU interrupts
perf: arm_spe_pmu: Request specific affinities for per CPU interrupts
perf: arm_pmu: Request specific affinities for per CPU NMIs/interrupts
genirq: Add request_percpu_irq_affinity() helper
genirq: Allow per-cpu interrupt sharing for non-overlapping affinities
genirq: Update request_percpu_nmi() to take an affinity
genirq: Add affinity to percpu_devid interrupt requests
...
Pull rseq updates from Thomas Gleixner:
"A large overhaul of the restartable sequences and CID management:
The recent enablement of RSEQ in glibc resulted in regressions which
are caused by the related overhead. It turned out that the decision to
invoke the exit to user work was not really a decision. More or less
each context switch caused that. There is a long list of small issues
which sums up nicely and results in a 3-4% regression in I/O
benchmarks.
The other detail which caused issues due to extra work in context
switch and task migration is the CID (memory context ID) management.
It also requires to use a task work to consolidate the CID space,
which is executed in the context of an arbitrary task and results in
sporadic uncontrolled exit latencies.
The rewrite addresses this by:
- Removing deprecated and long unsupported functionality
- Moving the related data into dedicated data structures which are
optimized for fast path processing.
- Caching values so actual decisions can be made
- Replacing the current implementation with a optimized inlined
variant.
- Separating fast and slow path for architectures which use the
generic entry code, so that only fault and error handling goes into
the TIF_NOTIFY_RESUME handler.
- Rewriting the CID management so that it becomes mostly invisible in
the context switch path. That moves the work of switching modes
into the fork/exit path, which is a reasonable tradeoff. That work
is only required when a process creates more threads than the
cpuset it is allowed to run on or when enough threads exit after
that. An artificial thread pool benchmarks which triggers this did
not degrade, it actually improved significantly.
The main effect in migration heavy scenarios is that runqueue lock
held time and therefore contention goes down significantly"
* tag 'core-rseq-2025-11-30' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (54 commits)
sched/mmcid: Switch over to the new mechanism
sched/mmcid: Implement deferred mode change
irqwork: Move data struct to a types header
sched/mmcid: Provide CID ownership mode fixup functions
sched/mmcid: Provide new scheduler CID mechanism
sched/mmcid: Introduce per task/CPU ownership infrastructure
sched/mmcid: Serialize sched_mm_cid_fork()/exit() with a mutex
sched/mmcid: Provide precomputed maximal value
sched/mmcid: Move initialization out of line
signal: Move MMCID exit out of sighand lock
sched/mmcid: Convert mm CID mask to a bitmap
cpumask: Cache num_possible_cpus()
sched/mmcid: Use cpumask_weighted_or()
cpumask: Introduce cpumask_weighted_or()
sched/mmcid: Prevent pointless work in mm_update_cpus_allowed()
sched/mmcid: Move scheduler code out of global header
sched: Fixup whitespace damage
sched/mmcid: Cacheline align MM CID storage
sched/mmcid: Use proper data structures
sched/mmcid: Revert the complex CID management
...
Pull scoped user access updates from Thomas Gleixner:
"Scoped user mode access and related changes:
- Implement the missing u64 user access function on ARM when
CONFIG_CPU_SPECTRE=n.
This makes it possible to access a 64bit value in generic code with
[unsafe_]get_user(). All other architectures and ARM variants
provide the relevant accessors already.
- Ensure that ASM GOTO jump label usage in the user mode access
helpers always goes through a local C scope label indirection
inside the helpers.
This is required because compilers are not supporting that a ASM
GOTO target leaves a auto cleanup scope. GCC silently fails to emit
the cleanup invocation and CLANG fails the build.
[ Editor's note: gcc-16 will have fixed the code generation issue
in commit f68fe3ddda4 ("eh: Invoke cleanups/destructors in asm
goto jumps [PR122835]"). But we obviously have to deal with clang
and older versions of gcc, so.. - Linus ]
This provides generic wrapper macros and the conversion of affected
architecture code to use them.
- Scoped user mode access with auto cleanup
Access to user mode memory can be required in hot code paths, but
if it has to be done with user controlled pointers, the access is
shielded with a speculation barrier, so that the CPU cannot
speculate around the address range check. Those speculation
barriers impact performance quite significantly.
This cost can be avoided by "masking" the provided pointer so it is
guaranteed to be in the valid user memory access range and
otherwise to point to a guaranteed unpopulated address space. This
has to be done without branches so it creates an address dependency
for the access, which the CPU cannot speculate ahead.
This results in repeating and error prone programming patterns:
if (can_do_masked_user_access())
from = masked_user_read_access_begin((from));
else if (!user_read_access_begin(from, sizeof(*from)))
return -EFAULT;
unsafe_get_user(val, from, Efault);
user_read_access_end();
return 0;
Efault:
user_read_access_end();
return -EFAULT;
which can be replaced with scopes and automatic cleanup:
scoped_user_read_access(from, Efault)
unsafe_get_user(val, from, Efault);
return 0;
Efault:
return -EFAULT;
- Convert code which implements the above pattern over to
scope_user.*.access(). This also corrects a couple of imbalanced
masked_*_begin() instances which are harmless on most
architectures, but prevent PowerPC from implementing the masking
optimization.
- Add a missing speculation barrier in copy_from_user_iter()"
* tag 'core-uaccess-2025-11-30' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
lib/strn*,uaccess: Use masked_user_{read/write}_access_begin when required
scm: Convert put_cmsg() to scoped user access
iov_iter: Add missing speculation barrier to copy_from_user_iter()
iov_iter: Convert copy_from_user_iter() to masked user access
select: Convert to scoped user access
x86/futex: Convert to scoped user access
futex: Convert to get/put_user_inline()
uaccess: Provide put/get_user_inline()
uaccess: Provide scoped user access regions
arm64: uaccess: Use unsafe wrappers for ASM GOTO
s390/uaccess: Use unsafe wrappers for ASM GOTO
riscv/uaccess: Use unsafe wrappers for ASM GOTO
powerpc/uaccess: Use unsafe wrappers for ASM GOTO
x86/uaccess: Use unsafe wrappers for ASM GOTO
uaccess: Provide ASM GOTO safe wrappers for unsafe_*_user()
ARM: uaccess: Implement missing __get_user_asm_dword()
Pull bug handling infrastructure updates from Ingo Molnar:
"Core updates:
- Improve WARN(), which has vararg printf like arguments, to work
with the x86 #UD based WARN-optimizing infrastructure by hiding the
format in the bug_table and replacing this first argument with the
address of the bug-table entry, while making the actual function
that's called a UD1 instruction (Peter Zijlstra)
- Introduce the CONFIG_DEBUG_BUGVERBOSE_DETAILED Kconfig switch (Ingo
Molnar, s390 support by Heiko Carstens)
Fixes and cleanups:
- bugs/s390: Remove private WARN_ON() implementation (Heiko Carstens)
- <asm/bugs.h>: Make i386 use GENERIC_BUG_RELATIVE_POINTERS (Peter
Zijlstra)"
* tag 'core-bugs-2025-12-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (31 commits)
x86/bugs: Make i386 use GENERIC_BUG_RELATIVE_POINTERS
x86/bug: Fix BUG_FORMAT vs KASLR
x86_64/bug: Inline the UD1
x86/bug: Implement WARN_ONCE()
x86_64/bug: Implement __WARN_printf()
x86/bug: Use BUG_FORMAT for DEBUG_BUGVERBOSE_DETAILED
x86/bug: Add BUG_FORMAT basics
bug: Allow architectures to provide __WARN_printf()
bug: Implement WARN_ON() using __WARN_FLAGS()
bug: Add report_bug_entry()
bug: Add BUG_FORMAT_ARGS infrastructure
bug: Clean up CONFIG_GENERIC_BUG_RELATIVE_POINTERS
bug: Add BUG_FORMAT infrastructure
x86: Rework __bug_table helpers
bugs/s390: Remove private WARN_ON() implementation
bugs/core: Reorganize fields in the first line of WARNING output, add ->comm[] output
bugs/sh: Concatenate 'cond_str' with '__FILE__' in __WARN_FLAGS(), to extend WARN_ON/BUG_ON output
bugs/parisc: Concatenate 'cond_str' with '__FILE__' in __WARN_FLAGS(), to extend WARN_ON/BUG_ON output
bugs/riscv: Concatenate 'cond_str' with '__FILE__' in __BUG_FLAGS(), to extend WARN_ON/BUG_ON output
bugs/riscv: Pass in 'cond_str' to __BUG_FLAGS()
...
Pull scheduler updates from Ingo Molnar:
"Scalability and load-balancing improvements:
- Enable scheduler feature NEXT_BUDDY (Mel Gorman)
- Reimplement NEXT_BUDDY to align with EEVDF goals (Mel Gorman)
- Skip sched_balance_running cmpxchg when balance is not due (Tim
Chen)
- Implement generic code for architecture specific sched domain NUMA
distances (Tim Chen)
- Optimize the NUMA distances of the sched-domains builds of Intel
Granite Rapids (GNR) and Clearwater Forest (CWF) platforms (Tim
Chen)
- Implement proportional newidle balance: a randomized algorithm that
runs newidle balancing proportional to its success rate. (Peter
Zijlstra)
Scheduler infrastructure changes:
- Implement the 'sched_change' scoped_guard() pattern for the entire
scheduler (Peter Zijlstra)
- More broadly utilize the sched_change guard (Peter Zijlstra)
- Add support to pick functions to take runqueue-flags (Joel
Fernandes)
- Provide and use set_need_resched_current() (Peter Zijlstra)
Fair scheduling enhancements:
- Forfeit vruntime on yield (Fernand Sieber)
- Only update stats for allowed CPUs when looking for dst group (Adam
Li)
CPU-core scheduling enhancements:
- Optimize core cookie matching check (Fernand Sieber)
Deadline scheduler fixes:
- Only set free_cpus for online runqueues (Doug Berger)
- Fix dl_server time accounting (Peter Zijlstra)
- Fix dl_server stop condition (Peter Zijlstra)
Proxy scheduling fixes:
- Yield the donor task (Fernand Sieber)
Fixes and cleanups:
- Fix do_set_cpus_allowed() locking (Peter Zijlstra)
- Fix migrate_disable_switch() locking (Peter Zijlstra)
- Remove double update_rq_clock() in __set_cpus_allowed_ptr_locked()
(Hao Jia)
- Increase sched_tick_remote timeout (Phil Auld)
- sched/deadline: Use cpumask_weight_and() in dl_bw_cpus() (Shrikanth
Hegde)
- sched/deadline: Clean up select_task_rq_dl() (Shrikanth Hegde)"
* tag 'sched-core-2025-12-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (44 commits)
sched: Provide and use set_need_resched_current()
sched/fair: Proportional newidle balance
sched/fair: Small cleanup to update_newidle_cost()
sched/fair: Small cleanup to sched_balance_newidle()
sched/fair: Revert max_newidle_lb_cost bump
sched/fair: Reimplement NEXT_BUDDY to align with EEVDF goals
sched/fair: Enable scheduler feature NEXT_BUDDY
sched: Increase sched_tick_remote timeout
sched/fair: Have SD_SERIALIZE affect newidle balancing
sched/fair: Skip sched_balance_running cmpxchg when balance is not due
sched/deadline: Minor cleanup in select_task_rq_dl()
sched/deadline: Use cpumask_weight_and() in dl_bw_cpus
sched/deadline: Document dl_server
sched/deadline: Fix dl_server stop condition
sched/deadline: Fix dl_server time accounting
sched/core: Remove double update_rq_clock() in __set_cpus_allowed_ptr_locked()
sched/eevdf: Fix min_vruntime vs avg_vruntime
sched/core: Add comment explaining force-idle vruntime snapshots
sched/core: Optimize core cookie matching check
sched/proxy: Yield the donor task
...
Pull performance events updates from Ingo Molnar:
"Callchain support:
- Add support for deferred user-space stack unwinding for perf,
enabled on x86. (Peter Zijlstra, Steven Rostedt)
- unwind_user/x86: Enable frame pointer unwinding on x86 (Josh
Poimboeuf)
x86 PMU support and infrastructure:
- x86/insn: Simplify for_each_insn_prefix() (Peter Zijlstra)
- x86/insn,uprobes,alternative: Unify insn_is_nop() (Peter Zijlstra)
Intel PMU driver:
- Large series to prepare for and implement architectural PEBS
support for Intel platforms such as Clearwater Forest (CWF) and
Panther Lake (PTL). (Dapeng Mi, Kan Liang)
- Check dynamic constraints (Kan Liang)
- Optimize PEBS extended config (Peter Zijlstra)
- cstates:
- Remove PC3 support from LunarLake (Zhang Rui)
- Add Pantherlake support (Zhang Rui)
- Clearwater Forest support (Zide Chen)
AMD PMU driver:
- x86/amd: Check event before enable to avoid GPF (George Kennedy)
Fixes and cleanups:
- task_work: Fix NMI race condition (Peter Zijlstra)
- perf/x86: Fix NULL event access and potential PEBS record loss
(Dapeng Mi)
- Misc other fixes and cleanups (Dapeng Mi, Ingo Molnar, Peter
Zijlstra)"
* tag 'perf-core-2025-12-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (38 commits)
perf/x86/intel: Fix and clean up intel_pmu_drain_arch_pebs() type use
perf/x86/intel: Optimize PEBS extended config
perf/x86/intel: Check PEBS dyn_constraints
perf/x86/intel: Add a check for dynamic constraints
perf/x86/intel: Add counter group support for arch-PEBS
perf/x86/intel: Setup PEBS data configuration and enable legacy groups
perf/x86/intel: Update dyn_constraint base on PEBS event precise level
perf/x86/intel: Allocate arch-PEBS buffer and initialize PEBS_BASE MSR
perf/x86/intel: Process arch-PEBS records or record fragments
perf/x86/intel/ds: Factor out PEBS group processing code to functions
perf/x86/intel/ds: Factor out PEBS record processing code to functions
perf/x86/intel: Initialize architectural PEBS
perf/x86/intel: Correct large PEBS flag check
perf/x86/intel: Replace x86_pmu.drain_pebs calling with static call
perf/x86: Fix NULL event access and potential PEBS record loss
perf/x86: Remove redundant is_x86_event() prototype
entry,unwind/deferred: Fix unwind_reset_info() placement
unwind_user/x86: Fix arch=um build
perf: Support deferred user unwind
unwind_user/x86: Teach FP unwind about start of function
...
Pull objtool updates from Ingo Molnar:
- klp-build livepatch module generation (Josh Poimboeuf)
Introduce new objtool features and a klp-build script to generate
livepatch modules using a source .patch as input.
This builds on concepts from the longstanding out-of-tree kpatch
project which began in 2012 and has been used for many years to
generate livepatch modules for production kernels. However, this is a
complete rewrite which incorporates hard-earned lessons from 12+
years of maintaining kpatch.
Key improvements compared to kpatch-build:
- Integrated with objtool: Leverages objtool's existing control-flow
graph analysis to help detect changed functions.
- Works on vmlinux.o: Supports late-linked objects, making it
compatible with LTO, IBT, and similar.
- Simplified code base: ~3k fewer lines of code.
- Upstream: No more out-of-tree #ifdef hacks, far less cruft.
- Cleaner internals: Vastly simplified logic for
symbol/section/reloc inclusion and special section extraction.
- Robust __LINE__ macro handling: Avoids false positive binary diffs
caused by the __LINE__ macro by introducing a fix-patch-lines
script which injects #line directives into the source .patch to
preserve the original line numbers at compile time.
- Disassemble code with libopcodes instead of running objdump
(Alexandre Chartre)
- Disassemble support (-d option to objtool) by Alexandre Chartre,
which supports the decoding of various Linux kernel code generation
specials such as alternatives:
17ef: sched_balance_find_dst_group+0x62f mov 0x34(%r9),%edx
17f3: sched_balance_find_dst_group+0x633 | <alternative.17f3> | X86_FEATURE_POPCNT
17f3: sched_balance_find_dst_group+0x633 | call 0x17f8 <__sw_hweight64> | popcnt %rdi,%rax
17f8: sched_balance_find_dst_group+0x638 cmp %eax,%edx
... jump table alternatives:
1895: sched_use_asym_prio+0x5 test $0x8,%ch
1898: sched_use_asym_prio+0x8 je 0x18a9 <sched_use_asym_prio+0x19>
189a: sched_use_asym_prio+0xa | <jump_table.189a> | JUMP
189a: sched_use_asym_prio+0xa | jmp 0x18ae <sched_use_asym_prio+0x1e> | nop2
189c: sched_use_asym_prio+0xc mov $0x1,%eax
18a1: sched_use_asym_prio+0x11 and $0x80,%ecx
... exception table alternatives:
native_read_msr:
5b80: native_read_msr+0x0 mov %edi,%ecx
5b82: native_read_msr+0x2 | <ex_table.5b82> | EXCEPTION
5b82: native_read_msr+0x2 | rdmsr | resume at 0x5b84 <native_read_msr+0x4>
5b84: native_read_msr+0x4 shl $0x20,%rdx
.... x86 feature flag decoding (also see the X86_FEATURE_POPCNT
example in sched_balance_find_dst_group() above):
2faaf: start_thread_common.constprop.0+0x1f jne 0x2fba4 <start_thread_common.constprop.0+0x114>
2fab5: start_thread_common.constprop.0+0x25 | <alternative.2fab5> | X86_FEATURE_ALWAYS | X86_BUG_NULL_SEG
2fab5: start_thread_common.constprop.0+0x25 | jmp 0x2faba <.altinstr_aux+0x2f4> | jmp 0x4b0 <start_thread_common.constprop.0+0x3f> | nop5
2faba: start_thread_common.constprop.0+0x2a mov $0x2b,%eax
... NOP sequence shortening:
1048e2: snapshot_write_finalize+0xc2 je 0x104917 <snapshot_write_finalize+0xf7>
1048e4: snapshot_write_finalize+0xc4 nop6
1048ea: snapshot_write_finalize+0xca nop11
1048f5: snapshot_write_finalize+0xd5 nop11
104900: snapshot_write_finalize+0xe0 mov %rax,%rcx
104903: snapshot_write_finalize+0xe3 mov 0x10(%rdx),%rax
... and much more.
- Function validation tracing support (Alexandre Chartre)
- Various -ffunction-sections fixes (Josh Poimboeuf)
- Clang AutoFDO (Automated Feedback-Directed Optimizations) support
(Josh Poimboeuf)
- Misc fixes and cleanups (Borislav Petkov, Chen Ni, Dylan Hatch, Ingo
Molnar, John Wang, Josh Poimboeuf, Pankaj Raghav, Peter Zijlstra,
Thorsten Blum)
* tag 'objtool-core-2025-12-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (129 commits)
objtool: Fix segfault on unknown alternatives
objtool: Build with disassembly can fail when including bdf.h
objtool: Trim trailing NOPs in alternative
objtool: Add wide output for disassembly
objtool: Compact output for alternatives with one instruction
objtool: Improve naming of group alternatives
objtool: Add Function to get the name of a CPU feature
objtool: Provide access to feature and flags of group alternatives
objtool: Fix address references in alternatives
objtool: Disassemble jump table alternatives
objtool: Disassemble exception table alternatives
objtool: Print addresses with alternative instructions
objtool: Disassemble group alternatives
objtool: Print headers for alternatives
objtool: Preserve alternatives order
objtool: Add the --disas=<function-pattern> action
objtool: Do not validate IBT for .return_sites and .call_sites
objtool: Improve tracing of alternative instructions
objtool: Add functions to better name alternatives
objtool: Identify the different types of alternatives
...
Pull locking updates from Ingo Molnar:
"Mutexes:
- Redo __mutex_init() to reduce generated code size (Sebastian
Andrzej Siewior)
Seqlocks:
- Introduce scoped_seqlock_read() (Peter Zijlstra)
- Change thread_group_cputime() to use scoped_seqlock_read() (Oleg
Nesterov)
- Change do_task_stat() to use scoped_seqlock_read() (Oleg Nesterov)
- Change do_io_accounting() to use scoped_seqlock_read() (Oleg
Nesterov)
- Fix the incorrect documentation of read_seqbegin_or_lock() /
need_seqretry() (Oleg Nesterov)
- Allow KASAN to fail optimizing (Peter Zijlstra)
Local lock updates:
- Fix all kernel-doc warnings (Randy Dunlap)
- Add the <linux/local_lock*.h> headers to MAINTAINERS (Sebastian
Andrzej Siewior)
- Reduce the risk of shadowing via s/l/__l/ and s/tl/__tl/ (Vincent
Mailhol)
Lock debugging:
- spinlock/debug: Fix data-race in do_raw_write_lock (Alexander
Sverdlin)
Atomic primitives infrastructure:
- atomic: Skip alignment check for try_cmpxchg() old arg (Arnd
Bergmann)
Rust runtime integration:
- sync: atomic: Enable generated Atomic<T> usage (Boqun Feng)
- sync: atomic: Implement Debug for Atomic<Debug> (Boqun Feng)
- debugfs: Remove Rust native atomics and replace them with Linux
versions (Boqun Feng)
- debugfs: Implement Reader for Mutex<T> only when T is Unpin (Boqun
Feng)
- lock: guard: Add T: Unpin bound to DerefMut (Daniel Almeida)
- lock: Pin the inner data (Daniel Almeida)
- lock: Add a Pin<&mut T> accessor (Daniel Almeida)"
* tag 'locking-core-2025-12-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
locking/local_lock: Fix all kernel-doc warnings
locking/local_lock: s/l/__l/ and s/tl/__tl/ to reduce the risk of shadowing
locking/local_lock: Add the <linux/local_lock*.h> headers to MAINTAINERS
locking/mutex: Redo __mutex_init() to reduce generated code size
rust: debugfs: Replace the usage of Rust native atomics
rust: sync: atomic: Implement Debug for Atomic<Debug>
rust: sync: atomic: Make Atomic*Ops pub(crate)
seqlock: Allow KASAN to fail optimizing
rust: debugfs: Implement Reader for Mutex<T> only when T is Unpin
seqlock: Change do_io_accounting() to use scoped_seqlock_read()
seqlock: Change do_task_stat() to use scoped_seqlock_read()
seqlock: Change thread_group_cputime() to use scoped_seqlock_read()
seqlock: Introduce scoped_seqlock_read()
documentation: seqlock: fix the wrong documentation of read_seqbegin_or_lock/need_seqretry
atomic: Skip alignment check for try_cmpxchg() old arg
rust: lock: Add a Pin<&mut T> accessor
rust: lock: Pin the inner data
rust: lock: guard: Add T: Unpin bound to DerefMut
locking/spinlock/debug: Fix data-race in do_raw_write_lock
Pull fd prepare updates from Christian Brauner:
"This adds the FD_ADD() and FD_PREPARE() primitive. They simplify the
common pattern of get_unused_fd_flags() + create file + fd_install()
that is used extensively throughout the kernel and currently requires
cumbersome cleanup paths.
FD_ADD() - For simple cases where a file is installed immediately:
fd = FD_ADD(O_CLOEXEC, vfio_device_open_file(device));
if (fd < 0)
vfio_device_put_registration(device);
return fd;
FD_PREPARE() - For cases requiring access to the fd or file, or
additional work before publishing:
FD_PREPARE(fdf, O_CLOEXEC, sync_file->file);
if (fdf.err) {
fput(sync_file->file);
return fdf.err;
}
data.fence = fd_prepare_fd(fdf);
if (copy_to_user((void __user *)arg, &data, sizeof(data)))
return -EFAULT;
return fd_publish(fdf);
The primitives are centered around struct fd_prepare. FD_PREPARE()
encapsulates all allocation and cleanup logic and must be followed by
a call to fd_publish() which associates the fd with the file and
installs it into the caller's fdtable. If fd_publish() isn't called,
both are deallocated automatically. FD_ADD() is a shorthand that does
fd_publish() immediately and never exposes the struct to the caller.
I've implemented this in a way that it's compatible with the cleanup
infrastructure while also being usable separately. IOW, it's centered
around struct fd_prepare which is aliased to class_fd_prepare_t and so
we can make use of all the basica guard infrastructure"
* tag 'vfs-6.19-rc1.fd_prepare.fs' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (42 commits)
io_uring: convert io_create_mock_file() to FD_PREPARE()
file: convert replace_fd() to FD_PREPARE()
vfio: convert vfio_group_ioctl_get_device_fd() to FD_ADD()
tty: convert ptm_open_peer() to FD_ADD()
ntsync: convert ntsync_obj_get_fd() to FD_PREPARE()
media: convert media_request_alloc() to FD_PREPARE()
hv: convert mshv_ioctl_create_partition() to FD_ADD()
gpio: convert linehandle_create() to FD_PREPARE()
pseries: port papr_rtas_setup_file_interface() to FD_ADD()
pseries: convert papr_platform_dump_create_handle() to FD_ADD()
spufs: convert spufs_gang_open() to FD_PREPARE()
papr-hvpipe: convert papr_hvpipe_dev_create_handle() to FD_PREPARE()
spufs: convert spufs_context_open() to FD_PREPARE()
net/socket: convert __sys_accept4_file() to FD_ADD()
net/socket: convert sock_map_fd() to FD_ADD()
net/kcm: convert kcm_ioctl() to FD_PREPARE()
net/handshake: convert handshake_nl_accept_doit() to FD_PREPARE()
secretmem: convert memfd_secret() to FD_ADD()
memfd: convert memfd_create() to FD_ADD()
bpf: convert bpf_token_create() to FD_PREPARE()
...
Pull cred guard updates from Christian Brauner:
"This contains substantial credential infrastructure improvements
adding guard-based credential management that simplifies code and
eliminates manual reference counting in many subsystems.
Features:
- Kernel Credential Guards
Add with_kernel_creds() and scoped_with_kernel_creds() guards that
allow using the kernel credentials without allocating and copying
them. This was requested by Linus after seeing repeated
prepare_kernel_creds() calls that duplicate the kernel credentials
only to drop them again later.
The new guards completely avoid the allocation and never expose the
temporary variable to hold the kernel credentials anywhere in
callers.
- Generic Credential Guards
Add scoped_with_creds() guards for the common override_creds() and
revert_creds() pattern. This builds on earlier work that made
override_creds()/revert_creds() completely reference count free.
- Prepare Credential Guards
Add prepare credential guards for the more complex pattern of
preparing a new set of credentials and overriding the current
credentials with them:
- prepare_creds()
- modify new creds
- override_creds()
- revert_creds()
- put_cred()
Cleanups:
- Make init_cred static since it should not be directly accessed
- Add kernel_cred() helper to properly access the kernel credentials
- Fix scoped_class() macro that was introduced two cycles ago
- coredump: split out do_coredump() from vfs_coredump() for cleaner
credential handling
- coredump: move revert_cred() before coredump_cleanup()
- coredump: mark struct mm_struct as const
- coredump: pass struct linux_binfmt as const
- sev-dev: use guard for path"
* tag 'kernel-6.19-rc1.cred' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (36 commits)
trace: use override credential guard
trace: use prepare credential guard
coredump: use override credential guard
coredump: use prepare credential guard
coredump: split out do_coredump() from vfs_coredump()
coredump: mark struct mm_struct as const
coredump: pass struct linux_binfmt as const
coredump: move revert_cred() before coredump_cleanup()
sev-dev: use override credential guards
sev-dev: use prepare credential guard
sev-dev: use guard for path
cred: add prepare credential guard
net/dns_resolver: use credential guards in dns_query()
cgroup: use credential guards in cgroup_attach_permissions()
act: use credential guards in acct_write_process()
smb: use credential guards in cifs_get_spnego_key()
nfs: use credential guards in nfs_idmap_get_key()
nfs: use credential guards in nfs_local_call_write()
nfs: use credential guards in nfs_local_call_read()
erofs: use credential guards
...
Pull namespace updates from Christian Brauner:
"This contains substantial namespace infrastructure changes including a new
system call, active reference counting, and extensive header cleanups.
The branch depends on the shared kbuild branch for -fms-extensions support.
Features:
- listns() system call
Add a new listns() system call that allows userspace to iterate
through namespaces in the system. This provides a programmatic
interface to discover and inspect namespaces, addressing
longstanding limitations:
Currently, there is no direct way for userspace to enumerate
namespaces. Applications must resort to scanning /proc/*/ns/ across
all processes, which is:
- Inefficient - requires iterating over all processes
- Incomplete - misses namespaces not attached to any running
process but kept alive by file descriptors, bind mounts, or
parent references
- Permission-heavy - requires access to /proc for many processes
- No ordering or ownership information
- No filtering per namespace type
The listns() system call solves these problems:
ssize_t listns(const struct ns_id_req *req, u64 *ns_ids,
size_t nr_ns_ids, unsigned int flags);
struct ns_id_req {
__u32 size;
__u32 spare;
__u64 ns_id;
struct /* listns */ {
__u32 ns_type;
__u32 spare2;
__u64 user_ns_id;
};
};
Features include:
- Pagination support for large namespace sets
- Filtering by namespace type (MNT_NS, NET_NS, USER_NS, etc.)
- Filtering by owning user namespace
- Permission checks respecting namespace isolation
- Active Reference Counting
Introduce an active reference count that tracks namespace
visibility to userspace. A namespace is visible in the following
cases:
- The namespace is in use by a task
- The namespace is persisted through a VFS object (namespace file
descriptor or bind-mount)
- The namespace is a hierarchical type and is the parent of child
namespaces
The active reference count does not regulate lifetime (that's still
done by the normal reference count) - it only regulates visibility
to namespace file handles and listns().
This prevents resurrection of namespaces that are pinned only for
internal kernel reasons (e.g., user namespaces held by
file->f_cred, lazy TLB references on idle CPUs, etc.) which should
not be accessible via (1)-(3).
- Unified Namespace Tree
Introduce a unified tree structure for all namespaces with:
- Fixed IDs assigned to initial namespaces
- Lookup based solely on inode number
- Maintained list of owned namespaces per user namespace
- Simplified rbtree comparison helpers
Cleanups
- Header Reorganization:
- Move namespace types into separate header (ns_common_types.h)
- Decouple nstree from ns_common header
- Move nstree types into separate header
- Switch to new ns_tree_{node,root} structures with helper functions
- Use guards for ns_tree_lock
- Initial Namespace Reference Count Optimization
- Make all reference counts on initial namespaces a nop to avoid
pointless cacheline ping-pong for namespaces that can never go
away
- Drop custom reference count initialization for initial namespaces
- Add NS_COMMON_INIT() macro and use it for all namespaces
- pid: rely on common reference count behavior
- Miscellaneous Cleanups
- Rename exit_task_namespaces() to exit_nsproxy_namespaces()
- Rename is_initial_namespace() and make argument const
- Use boolean to indicate anonymous mount namespace
- Simplify owner list iteration in nstree
- nsfs: raise SB_I_NODEV, SB_I_NOEXEC, and DCACHE_DONTCACHE explicitly
- nsfs: use inode_just_drop()
- pidfs: raise DCACHE_DONTCACHE explicitly
- pidfs: simplify PIDFD_GET__NAMESPACE ioctls
- libfs: allow to specify s_d_flags
- cgroup: add cgroup namespace to tree after owner is set
- nsproxy: fix free_nsproxy() and simplify create_new_namespaces()
Fixes:
- setns(pidfd, ...) race condition
Fix a subtle race when using pidfds with setns(). When the target
task exits after prepare_nsset() but before commit_nsset(), the
namespace's active reference count might have been dropped. If
setns() then installs the namespaces, it would bump the active
reference count from zero without taking the required reference on
the owner namespace, leading to underflow when later decremented.
The fix resurrects the ownership chain if necessary - if the caller
succeeded in grabbing passive references, the setns() should
succeed even if the target task exits or gets reaped.
- Return EFAULT on put_user() error instead of success
- Make sure references are dropped outside of RCU lock (some
namespaces like mount namespace sleep when putting the last
reference)
- Don't skip active reference count initialization for network
namespace
- Add asserts for active refcount underflow
- Add asserts for initial namespace reference counts (both passive
and active)
- ipc: enable is_ns_init_id() assertions
- Fix kernel-doc comments for internal nstree functions
- Selftests
- 15 active reference count tests
- 9 listns() functionality tests
- 7 listns() permission tests
- 12 inactive namespace resurrection tests
- 3 threaded active reference count tests
- commit_creds() active reference tests
- Pagination and stress tests
- EFAULT handling test
- nsid tests fixes"
* tag 'namespace-6.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (103 commits)
pidfs: simplify PIDFD_GET_<type>_NAMESPACE ioctls
nstree: fix kernel-doc comments for internal functions
nsproxy: fix free_nsproxy() and simplify create_new_namespaces()
selftests/namespaces: fix nsid tests
ns: drop custom reference count initialization for initial namespaces
pid: rely on common reference count behavior
ns: add asserts for initial namespace active reference counts
ns: add asserts for initial namespace reference counts
ns: make all reference counts on initial namespace a nop
ipc: enable is_ns_init_id() assertions
fs: use boolean to indicate anonymous mount namespace
ns: rename is_initial_namespace()
ns: make is_initial_namespace() argument const
nstree: use guards for ns_tree_lock
nstree: simplify owner list iteration
nstree: switch to new structures
nstree: add helper to operate on struct ns_tree_{node,root}
nstree: move nstree types into separate header
nstree: decouple from ns_common header
ns: move namespace types into separate header
...
Pull misc vfs updates from Christian Brauner:
"Features:
- Cheaper MAY_EXEC handling for path lookup. This elides MAY_WRITE
permission checks during path lookup and adds the
IOP_FASTPERM_MAY_EXEC flag so filesystems like btrfs can avoid
expensive permission work.
- Hide dentry_cache behind runtime const machinery.
- Add German Maglione as virtiofs co-maintainer.
Cleanups:
- Tidy up and inline step_into() and walk_component() for improved
code generation.
- Re-enable IOCB_NOWAIT writes to files. This refactors file
timestamp update logic, fixing a layering bypass in btrfs when
updating timestamps on device files and improving FMODE_NOCMTIME
handling in VFS now that nfsd started using it.
- Path lookup optimizations extracting slowpaths into dedicated
routines and adding branch prediction hints for mntput_no_expire(),
fd_install(), lookup_slow(), and various other hot paths.
- Enable clang's -fms-extensions flag, requiring a JFS rename to
avoid conflicts.
- Remove spurious exports in fs/file_attr.c.
- Stop duplicating union pipe_index declaration. This depends on the
shared kbuild branch that brings in -fms-extensions support which
is merged into this branch.
- Use MD5 library instead of crypto_shash in ecryptfs.
- Use largest_zero_folio() in iomap_dio_zero().
- Replace simple_strtol/strtoul with kstrtoint/kstrtouint in init and
initrd code.
- Various typo fixes.
Fixes:
- Fix emergency sync for btrfs. Btrfs requires an explicit sync_fs()
call with wait == 1 to commit super blocks. The emergency sync path
never passed this, leaving btrfs data uncommitted during emergency
sync.
- Use local kmap in watch_queue's post_one_notification().
- Add hint prints in sb_set_blocksize() for LBS dependency on THP"
* tag 'vfs-6.19-rc1.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (35 commits)
MAINTAINERS: add German Maglione as virtiofs co-maintainer
fs: inline step_into() and walk_component()
fs: tidy up step_into() & friends before inlining
orangefs: use inode_update_timestamps directly
btrfs: fix the comment on btrfs_update_time
btrfs: use vfs_utimes to update file timestamps
fs: export vfs_utimes
fs: lift the FMODE_NOCMTIME check into file_update_time_flags
fs: refactor file timestamp update logic
include/linux/fs.h: trivial fix: regualr -> regular
fs/splice.c: trivial fix: pipes -> pipe's
fs: mark lookup_slow() as noinline
fs: add predicts based on nd->depth
fs: move mntput_no_expire() slowpath into a dedicated routine
fs: remove spurious exports in fs/file_attr.c
watch_queue: Use local kmap in post_one_notification()
fs: touch up predicts in path lookup
fs: move fd_install() slowpath into a dedicated routine and provide commentary
fs: hide dentry_cache behind runtime const machinery
fs: touch predicts in do_dentry_open()
...
mutex_init() invokes __mutex_init() providing the name of the lock and
a pointer to a the lock class. With LOCKDEP enabled this information is
useful but without LOCKDEP it not used at all. Passing the pointer
information of the lock class might be considered negligible but the
name of the lock is passed as well and the string is stored. This
information is wasting storage.
Split __mutex_init() into a _genereic() variant doing the initialisation
of the lock and a _lockdep() version which does _genereic() plus the
lockdep bits. Restrict the lockdep version to lockdep enabled builds
allowing the compiler to remove the unused parameter.
This results in the following size reduction:
text data bss dec filename
| 30237599 8161430 1176624 39575653 vmlinux.defconfig
| 30233269 8149142 1176560 39558971 vmlinux.defconfig.patched
-4.2KiB -12KiB
| 32455099 8471098 12934684 53860881 vmlinux.defconfig.lockdep
| 32455100 8471098 12934684 53860882 vmlinux.defconfig.patched.lockdep
| 27152407 7191822 2068040 36412269 vmlinux.defconfig.preempt_rt
| 27145937 7183630 2067976 36397543 vmlinux.defconfig.patched.preempt_rt
-6.3KiB -8KiB
| 29382020 7505742 13784608 50672370 vmlinux.defconfig.preempt_rt.lockdep
| 29376229 7505742 13784544 50666515 vmlinux.defconfig.patched.preempt_rt.lockdep
-5.6KiB
[peterz: folded fix from boqun]
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Waiman Long <longman@redhat.com>
Link: https://lkml.kernel.org/r/20251125145425.68319-1-boqun.feng@gmail.com
Link: https://patch.msgid.link/20251105142350.Tfeevs2N@linutronix.de
Pull timer fix from Borislav Petkov:
- Have timekeeping aux clocks sysfs interface setup function return an
error code on failure instead of success
* tag 'timers_urgent_for_v6.18_rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
timekeeping: Fix error code in tk_aux_sysfs_init()
Pull dma-mapping fixes from Marek Szyprowski:
"Two last minute fixes for the recently modified DMA API infrastructure:
- proper handling of DMA_ATTR_MMIO in dma_iova_unlink() function (me)
- regression fix for the code refactoring related to P2PDMA (Pranjal
Shrivastava)"
* tag 'dma-mapping-6.18-2025-11-27' of git://git.kernel.org/pub/scm/linux/kernel/git/mszyprowski/linux:
dma-direct: Fix missing sg_dma_len assignment in P2PDMA bus mappings
iommu/dma: add missing support for DMA_ATTR_MMIO for dma_iova_unlink()
Pull ring-buffer fix from Steven Rostedt:
- Do not allow mmapped ring buffer to be split
When the ring buffer VMA is split by a partial munmap or a MAP_FIXED,
the kernel calls vm_ops->close() on each portion. This causes the
ring_buffer_unmap() to be called multiple times. This causes
subsequent calls to return -ENODEV and triggers a warning.
There's no reason to allow user space to split up memory mapping of
the ring buffer. Have it return -EINVAL when that happens.
* tag 'trace-ringbuffer-v6.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
tracing: Fix WARN_ON in tracing_buffers_mmap_close for split VMAs
Prior to commit a25e7962db ("PCI/P2PDMA: Refactor the p2pdma mapping
helpers"), P2P segments were mapped using the pci_p2pdma_map_segment()
helper. This helper was responsible for populating sg->dma_address,
marking the bus address, and also setting sg_dma_len(sg).
The refactor[1] removed this helper and moved the mapping logic directly
into the callers. While iommu_dma_map_sg() was correctly updated to set
the length in the new flow, it was missed in dma_direct_map_sg().
Thus, in dma_direct_map_sg(), the PCI_P2PDMA_MAP_BUS_ADDR case sets the
dma_address and marks the segment, but immediately executes 'continue',
which causes the loop to skip the standard assignment logic at the end:
sg_dma_len(sg) = sg->length;
As a result, when CONFIG_NEED_SG_DMA_LENGTH is enabled, the dma_length
field remains uninitialized (zero) for P2P bus address mappings. This
breaks upper-layer drivers (for e.g. RDMA/IB) that rely on sg_dma_len()
to determine the transfer size.
Fix this by explicitly setting the DMA length in the
PCI_P2PDMA_MAP_BUS_ADDR case before continuing to the next scatterlist
entry.
Fixes: a25e7962db ("PCI/P2PDMA: Refactor the p2pdma mapping helpers")
Reported-by: Jacob Moroni <jmoroni@google.com>
Signed-off-by: Pranjal Shrivastava <praan@google.com>
[1]
https://lore.kernel.org/all/ac14a0e94355bf898de65d023ccf8a2ad22a3ece.1746424934.git.leon@kernel.org/
Reviewed-by: Logan Gunthorpe <logang@deltatee.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Reviewed-by: Shivaji Kant <shivajikant@google.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Link: https://lore.kernel.org/r/20251126114112.3694469-1-praan@google.com
Now that all pieces are in place, change the implementations of
sched_mm_cid_fork() and sched_mm_cid_exit() to adhere to the new strict
ownership scheme and switch context_switch() over to use the new
mm_cid_schedin() functionality.
The common case is that there is no mode change required, which makes
fork() and exit() just update the user count and the constraints.
In case that a new user would exceed the CID space limit the fork() context
handles the transition to per CPU mode with mm::mm_cid::mutex held. exit()
handles the transition back to per task mode when the user count drops
below the switch back threshold. fork() might also be forced to handle a
deferred switch back to per task mode, when a affinity change increased the
number of allowed CPUs enough.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://patch.msgid.link/20251119172550.280380631@linutronix.de
When affinity changes cause an increase of the number of CPUs allowed for
tasks which are related to a MM, that might results in a situation where
the ownership mode can go back from per CPU mode to per task mode.
As affinity changes happen with runqueue lock held there is no way to do
the actual mode change and required fixup right there.
Add the infrastructure to defer it to a workqueue. The scheduled work can
race with a fork() or exit(). Whatever happens first takes care of it.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://patch.msgid.link/20251119172550.216484739@linutronix.de
CIDs are either owned by tasks or by CPUs. The ownership mode depends on
the number of tasks related to a MM and the number of CPUs on which these
tasks are theoretically allowed to run on. Theoretically because that
number is the superset of CPU affinities of all tasks which only grows and
never shrinks.
Switching to per CPU mode happens when the user count becomes greater than
the maximum number of CIDs, which is calculated by:
opt_cids = min(mm_cid::nr_cpus_allowed, mm_cid::users);
max_cids = min(1.25 * opt_cids, nr_cpu_ids);
The +25% allowance is useful for tight CPU masks in scenarios where only a
few threads are created and destroyed to avoid frequent mode
switches. Though this allowance shrinks, the closer opt_cids becomes to
nr_cpu_ids, which is the (unfortunate) hard ABI limit.
At the point of switching to per CPU mode the new user is not yet visible
in the system, so the task which initiated the fork() runs the fixup
function: mm_cid_fixup_tasks_to_cpu() walks the thread list and either
transfers each tasks owned CID to the CPU the task runs on or drops it into
the CID pool if a task is not on a CPU at that point in time. Tasks which
schedule in before the task walk reaches them do the handover in
mm_cid_schedin(). When mm_cid_fixup_tasks_to_cpus() completes it's
guaranteed that no task related to that MM owns a CID anymore.
Switching back to task mode happens when the user count goes below the
threshold which was recorded on the per CPU mode switch:
pcpu_thrs = min(opt_cids - (opt_cids / 4), nr_cpu_ids / 2);
This threshold is updated when a affinity change increases the number of
allowed CPUs for the MM, which might cause a switch back to per task mode.
If the switch back was initiated by a exiting task, then that task runs the
fixup function. If it was initiated by a affinity change, then it's run
either in the deferred update function in context of a workqueue or by a
task which forks a new one or by a task which exits. Whatever happens
first. mm_cid_fixup_cpus_to_task() walks through the possible CPUs and
either transfers the CPU owned CIDs to a related task which runs on the CPU
or drops it into the pool. Tasks which schedule in on a CPU which the walk
did not cover yet do the handover themselves.
This transition from CPU to per task ownership happens in two phases:
1) mm:mm_cid.transit contains MM_CID_TRANSIT. This is OR'ed on the task
CID and denotes that the CID is only temporarily owned by the
task. When it schedules out the task drops the CID back into the
pool if this bit is set.
2) The initiating context walks the per CPU space and after completion
clears mm:mm_cid.transit. After that point the CIDs are strictly
task owned again.
This two phase transition is required to prevent CID space exhaustion
during the transition as a direct transfer of ownership would fail if
two tasks are scheduled in on the same CPU before the fixup freed per
CPU CIDs.
When mm_cid_fixup_cpus_to_tasks() completes it's guaranteed that no CID
related to that MM is owned by a CPU anymore.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://patch.msgid.link/20251119172550.088189028@linutronix.de
The MM CID management has two fundamental requirements:
1) It has to guarantee that at no given point in time the same CID is
used by concurrent tasks in userspace.
2) The CID space must not exceed the number of possible CPUs in a
system. While most allocators (glibc, tcmalloc, jemalloc) do not
care about that, there seems to be at least some LTTng library
depending on it.
The CID space compaction itself is not a functional correctness
requirement, it is only a useful optimization mechanism to reduce the
memory foot print in unused user space pools.
The optimal CID space is:
min(nr_tasks, nr_cpus_allowed);
Where @nr_tasks is the number of actual user space threads associated to
the mm and @nr_cpus_allowed is the superset of all task affinities. It is
growth only as it would be insane to take a racy snapshot of all task
affinities when the affinity of one task changes just do redo it 2
milliseconds later when the next task changes it's affinity.
That means that as long as the number of tasks is lower or equal than the
number of CPUs allowed, each task owns a CID. If the number of tasks
exceeds the number of CPUs allowed it switches to per CPU mode, where the
CPUs own the CIDs and the tasks borrow them as long as they are scheduled
in.
For transition periods CIDs can go beyond the optimal space as long as they
don't go beyond the number of possible CPUs.
The current upstream implementation adds overhead into task migration to
keep the CID with the task. It also has to do the CID space consolidation
work from a task work in the exit to user space path. As that work is
assigned to a random task related to a MM this can inflict unwanted exit
latencies.
Implement the context switch parts of a strict ownership mechanism to
address this.
This removes most of the work from the task which schedules out. Only
during transitioning from per CPU to per task ownership it is required to
drop the CID when leaving the CPU to prevent CID space exhaustion. Other
than that scheduling out is just a single check and branch.
The task which schedules in has to check whether:
1) The ownership mode changed
2) The CID is within the optimal CID space
In stable situations this results in zero work. The only short disruption
is when ownership mode changes or when the associated CID is not in the
optimal CID space. The latter only happens when tasks exit and therefore
the optimal CID space shrinks.
That mechanism is strictly optimized for the common case where no change
happens. The only case where it actually causes a temporary one time spike
is on mode changes when and only when a lot of tasks related to a MM
schedule exactly at the same time and have eventually to compete on
allocating a CID from the bitmap.
In the sysbench test case which triggered the spinlock contention in the
initial CID code, __schedule() drops significantly in perf top on a 128
Core (256 threads) machine when running sysbench with 255 threads, which
fits into the task mode limit of 256 together with the parent thread:
Upstream rseq/perf branch +CID rework
0.42% 0.37% 0.32% [k] __schedule
Increasing the number of threads to 256, which puts the test process into
per CPU mode looks about the same.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://patch.msgid.link/20251119172550.023984859@linutronix.de
The MM CID management has two fundamental requirements:
1) It has to guarantee that at no given point in time the same CID is
used by concurrent tasks in userspace.
2) The CID space must not exceed the number of possible CPUs in a
system. While most allocators (glibc, tcmalloc, jemalloc) do not care
about that, there seems to be at least librseq depending on it.
The CID space compaction itself is not a functional correctness
requirement, it is only a useful optimization mechanism to reduce the
memory foot print in unused user space pools.
The optimal CID space is:
min(nr_tasks, nr_cpus_allowed);
Where @nr_tasks is the number of actual user space threads associated to
the mm and @nr_cpus_allowed is the superset of all task affinities. It is
growth only as it would be insane to take a racy snapshot of all task
affinities when the affinity of one task changes just do redo it 2
milliseconds later when the next task changes its affinity.
That means that as long as the number of tasks is lower or equal than the
number of CPUs allowed, each task owns a CID. If the number of tasks
exceeds the number of CPUs allowed it switches to per CPU mode, where the
CPUs own the CIDs and the tasks borrow them as long as they are scheduled
in.
For transition periods CIDs can go beyond the optimal space as long as they
don't go beyond the number of possible CPUs.
The current upstream implementation adds overhead into task migration to
keep the CID with the task. It also has to do the CID space consolidation
work from a task work in the exit to user space path. As that work is
assigned to a random task related to a MM this can inflict unwanted exit
latencies.
This can be done differently by implementing a strict CID ownership
mechanism. Either the CIDs are owned by the tasks or by the CPUs. The
latter provides less locality when tasks are heavily migrating, but there
is no justification to optimize for overcommit scenarios and thereby
penalizing everyone else.
Provide the basic infrastructure to implement this:
- Change the UNSET marker to BIT(31) from ~0U
- Add the ONCPU marker as BIT(30)
- Add the TRANSIT marker as BIT(29)
That allows to check for ownership trivially and provides a simple check for
UNSET as well. The TRANSIT marker is required to prevent CID space
exhaustion when switching from per CPU to per task mode.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://patch.msgid.link/20251119172549.960252358@linutronix.de
Prepare for the new CID management scheme which puts the CID ownership
transition into the fork() and exit() slow path by serializing
sched_mm_cid_fork()/exit() with it, so task list and cpu mask walks can be
done in interruptible and preemptible code.
The contention on it is not worse than on other concurrency controls in the
fork()/exit() machinery.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://patch.msgid.link/20251119172549.895826703@linutronix.de
Reading mm::mm_users and mm:::mm_cid::nr_cpus_allowed every time to compute
the maximal CID value is just wasteful as that value is only changing on
fork(), exit() and eventually when the affinity changes.
So it can be easily precomputed at those points and provided in mm::mm_cid
for consumption in the hot path.
But there is an issue with using mm::mm_users for accounting because that
does not necessarily reflect the number of user space tasks as other kernel
code can take temporary references on the MM which skew the picture.
Solve that by adding a users counter to struct mm_mm_cid, which is modified
by fork() and exit() and used for precomputing under mm_mm_cid::lock.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://patch.msgid.link/20251119172549.832764634@linutronix.de
If kobject_create_and_add() fails on the first iteration, then the error
code is set to -ENOMEM which is correct. But if it fails in subsequent
iterations then "ret" is zero, which means success, but it should be
-ENOMEM.
Set the error code to -ENOMEM correctly.
Fixes: 7b5ab04f03 ("timekeeping: Fix resource leak in tk_aux_sysfs_init() error paths")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Malaya Kumar Rout <mrout@redhat.com>
Link: https://patch.msgid.link/aSW1R8q5zoY_DgQE@stanley.mountain
Pull timer fixes from Ingo Molnar:
- Fix a race in timer->function clearing in timer_shutdown_sync()
- Fix a timekeeper sysfs-setup resource leak in error paths
- Fix the NOHZ report_idle_softirq() syslog rate-limiting
logic to have no side effects on the return value
* tag 'timers-urgent-2025-11-23' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
timers: Fix NULL function pointer race in timer_shutdown_sync()
timekeeping: Fix resource leak in tk_aux_sysfs_init() error paths
tick/sched: Fix bogus condition in report_idle_softirq()
There is a race condition between timer_shutdown_sync() and timer
expiration that can lead to hitting a WARN_ON in expire_timers().
The issue occurs when timer_shutdown_sync() clears the timer function
to NULL while the timer is still running on another CPU. The race
scenario looks like this:
CPU0 CPU1
<SOFTIRQ>
lock_timer_base()
expire_timers()
base->running_timer = timer;
unlock_timer_base()
[call_timer_fn enter]
mod_timer()
...
timer_shutdown_sync()
lock_timer_base()
// For now, will not detach the timer but only clear its function to NULL
if (base->running_timer != timer)
ret = detach_if_pending(timer, base, true);
if (shutdown)
timer->function = NULL;
unlock_timer_base()
[call_timer_fn exit]
lock_timer_base()
base->running_timer = NULL;
unlock_timer_base()
...
// Now timer is pending while its function set to NULL.
// next timer trigger
<SOFTIRQ>
expire_timers()
WARN_ON_ONCE(!fn) // hit
...
lock_timer_base()
// Now timer will detach
if (base->running_timer != timer)
ret = detach_if_pending(timer, base, true);
if (shutdown)
timer->function = NULL;
unlock_timer_base()
The problem is that timer_shutdown_sync() clears the timer function
regardless of whether the timer is currently running. This can leave a
pending timer with a NULL function pointer, which triggers the
WARN_ON_ONCE(!fn) check in expire_timers().
Fix this by only clearing the timer function when actually detaching the
timer. If the timer is running, leave the function pointer intact, which is
safe because the timer will be properly detached when it finishes running.
Fixes: 0cc04e8045 ("timers: Add shutdown mechanism to the internal functions")
Signed-off-by: Yipeng Zou <zouyipeng@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20251122093942.301559-1-zouyipeng@huawei.com
Failing to allocate the affinity mask of an interrupt descriptor fails the
whole descriptor initialization. It is then guaranteed that the cpumask is
always available whenever the related interrupt objects are alive, such as
the kthread handler.
Therefore remove the superfluous check since it is merely a historical
leftover. Get rid also of the comments above it that are obsolete and
useless.
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://patch.msgid.link/20251121143500.42111-4-frederic@kernel.org
When a cpuset isolated partition is created / updated or destroyed, the
interrupt threads are affined blindly to all the non-isolated CPUs. This
happens without taking into account the interrupt threads initial affinity
that becomes ignored.
For example in a system with 8 CPUs, if an interrupt and its kthread are
initially affine to CPU 5, creating an isolated partition with only CPU 2
inside will eventually end up affining the interrupt kthread to all CPUs
but CPU 2 (that is CPUs 0,1,3-7), losing the kthread preference for CPU 5.
Besides the blind re-affining, this doesn't take care of the actual low
level interrupt which isn't migrated. As of today the only way to isolate
non managed interrupts, along with their kthreads, is to overwrite their
affinity separately, for example through /proc/irq/
To avoid doing that manually, future development should focus on updating
the interrupt's affinity whenever cpuset isolated partitions are updated.
In the meantime, cpuset shouldn't fiddle with interrupt threads directly.
To prevent from that, set the PF_NO_SETAFFINITY flag to them.
This is done through kthread_bind_mask() by affining them initially to all
possible CPUs as at that point the interrupt is not started up which means
the affinity of the hard interrupt is not known. The thread will adjust
that once it reaches the handler, which is guaranteed to happen after the
initial affinity of the hard interrupt is established.
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://patch.msgid.link/20251121143500.42111-3-frederic@kernel.org
During initialization, the interrupt thread is created before the interrupt
is enabled. The interrupt enablement happens before the actual kthread wake
up point. Once the interrupt is enabled the hardware can raise an interrupt
and once setup_irq() drops the descriptor lock a interrupt wake-up can
happen.
Even when such an interrupt can be considered premature, this is not a
problem in general because at the point where the descriptor lock is
dropped and the wakeup can happen, the data which is used by the thread is
fully initialized.
Though from the perspective of least surprise, the initial wakeup really
should be performed by the setup code and not randomly by a premature
interrupt.
Prevent this by performing a wake-up only if the target is in state
TASK_INTERRUPTIBLE, which the thread uses in wait_for_interrupt().
If the thread is still in state TASK_UNINTERRUPTIBLE, the wake-up is not
lost because after the setup code completed the initial wake-up the thread
will observe the IRQTF_RUNTHREAD and proceed with the handling.
[ tglx: Simplified the changes and extended the changelog. ]
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://patch.msgid.link/20251121143500.42111-2-frederic@kernel.org
Bring in the UDB and objtool data annotations to avoid conflicts while further extending the bug exceptions.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
The timer migration mechanism allows active CPUs to pull timers from
idle ones to improve the overall idle time. This is however undesired
when CPU intensive workloads run on isolated cores, as the algorithm
would move the timers from housekeeping to isolated cores, negatively
affecting the isolation.
Exclude isolated cores from the timer migration algorithm, extend the
concept of unavailable cores, currently used for offline ones, to
isolated ones:
* A core is unavailable if isolated or offline;
* A core is available if non isolated and online;
A core is considered unavailable as isolated if it belongs to:
* the isolcpus (domain) list
* an isolated cpuset
Except if it is:
* in the nohz_full list (already idle for the hierarchy)
* the nohz timekeeper core (must be available to handle global timers)
CPUs are added to the hierarchy during late boot, excluding isolated
ones, the hierarchy is also adapted when the cpuset isolation changes.
Due to how the timer migration algorithm works, any CPU part of the
hierarchy can have their global timers pulled by remote CPUs and have to
pull remote timers, only skipping pulling remote timers would break the
logic.
For this reason, prevent isolated CPUs from pulling remote global
timers, but also the other way around: any global timer started on an
isolated CPU will run there. This does not break the concept of
isolation (global timers don't come from outside the CPU) and, if
considered inappropriate, can usually be mitigated with other isolation
techniques (e.g. IRQ pinning).
This effect was noticed on a 128 cores machine running oslat on the
isolated cores (1-31,33-63,65-95,97-127). The tool monopolises CPUs,
and the CPU with lowest count in a timer migration hierarchy (here 1
and 65) appears as always active and continuously pulls global timers,
from the housekeeping CPUs. This ends up moving driver work (e.g.
delayed work) to isolated CPUs and causes latency spikes:
before the change:
# oslat -c 1-31,33-63,65-95,97-127 -D 62s
...
Maximum: 1203 10 3 4 ... 5 (us)
after the change:
# oslat -c 1-31,33-63,65-95,97-127 -D 62s
...
Maximum: 10 4 3 4 3 ... 5 (us)
The same behaviour was observed on a machine with as few as 20 cores /
40 threads with isocpus set to: 1-9,11-39 with rtla-osnoise-top.
Signed-off-by: Gabriele Monaco <gmonaco@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: John B. Wyatt IV <jwyatt@redhat.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://patch.msgid.link/20251120145653.296659-8-gmonaco@redhat.com
Currently the user can set up isolcpus and nohz_full in such a way that
leaves no housekeeping CPU (i.e. no CPU that is neither domain isolated
nor nohz full). This can be a problem for other subsystems (e.g. the
timer wheel imgration).
Prevent this configuration by invalidating the last setting in case the
union of isolcpus (domain) and nohz_full covers all CPUs.
Signed-off-by: Gabriele Monaco <gmonaco@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Waiman Long <longman@redhat.com>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://patch.msgid.link/20251120145653.296659-6-gmonaco@redhat.com
