kernel/linux.git/drivers/cpufreq/cpufreq.c, branch v7.0.10

cpufreq: Pass the policy to cpufreq_driver->adjust_perf()

2026-05-23T11:08:28+00:00

[ Upstream commit c03791085adcd61fa9b766ab303c7d0941d7378d ] cpufreq_cpu_get() can sleep on PREEMPT_RT in presence of concurrent writer(s), however amd-pstate depends on fetching the cpudata via the policy's driver data which necessitates grabbing the reference. Since schedutil governor can call "cpufreq_driver->update_perf()" during sched_tick/enqueue/dequeue with rq_lock held and IRQs disabled, fetching the policy object using the cpufreq_cpu_get() helper in the scheduler fast-path leads to "BUG: scheduling while atomic" on PREEMPT_RT [1]. Pass the cached cpufreq policy object in sg_policy to the update_perf() instead of just the CPU. The CPU can be inferred using "policy->cpu". The lifetime of cpufreq_policy object outlasts that of the governor and the cpufreq driver (allocated when the CPU is onlined and only reclaimed when the CPU is offlined / the CPU device is removed) which makes it safe to be referenced throughout the governor's lifetime. Closes:https://lore.kernel.org/all/20250731092316.3191-1-spasswolf@web.de/ [1] Fixes: 1d215f0319c2 ("cpufreq: amd-pstate: Add fast switch function for AMD P-State") Reported-by: Bert Karwatzki Acked-by: Viresh Kumar Signed-off-by: K Prateek Nayak Acked-by: Gary Guo # Rust Reviewed-by: Gautham R. Shenoy Reviewed-by: Zhongqiu Han Link: https://lore.kernel.org/r/20260316081849.19368-3-kprateek.nayak@amd.com Signed-off-by: Mario Limonciello (AMD) Signed-off-by: Sasha Levin

cpufreq: Don't skip cpufreq_frequency_table_cpuinfo()

2026-03-23T12:32:57+00:00

The commit 6db0f533d320 ("cpufreq: preserve freq_table_sorted across suspend/hibernate") unintentionally made a change where cpufreq_frequency_table_cpuinfo() isn't getting called anymore for old policies getting re-initialized. This leads to potentially invalid values of policy->max and policy->cpuinfo_max_freq. Fix the issue by reverting the original commit and adding the condition for just the sorting function. Fixes: 6db0f533d320 ("cpufreq: preserve freq_table_sorted across suspend/hibernate") Signed-off-by: Viresh Kumar Cc: 6.19+ # 6.19+ Link: https://patch.msgid.link/65ba5c45749267c82e8a87af3dc788b37a0b3f48.1773998611.git.viresh.kumar@linaro.org Signed-off-by: Rafael J. Wysocki

Convert 'alloc_obj' family to use the new default GFP_KERNEL argument

2026-02-22T01:09:51+00:00

This was done entirely with mindless brute force, using git grep -l '\

treewide: Replace kmalloc with kmalloc_obj for non-scalar types

2026-02-21T09:02:28+00:00

This is the result of running the Coccinelle script from scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to avoid scalar types (which need careful case-by-case checking), and instead replace kmalloc-family calls that allocate struct or union object instances: Single allocations: kmalloc(sizeof(TYPE), ...) are replaced with: kmalloc_obj(TYPE, ...) Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...) are replaced with: kmalloc_objs(TYPE, COUNT, ...) Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...) are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...) (where TYPE may also be *VAR) The resulting allocations no longer return "void *", instead returning "TYPE *". Signed-off-by: Kees Cook

Merge tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux

2026-02-10T04:28:45+00:00

Pull arm64 updates from Will Deacon: "There's a little less than normal, probably due to LPC & Christmas/New Year meaning that a few series weren't quite ready or reviewed in time. It's still useful across the board, despite the only real feature being support for the LS64 feature enabling 64-byte atomic accesses to endpoints that support it. ACPI: - Add interrupt signalling support to the AGDI handler - Add Catalin and myself to the arm64 ACPI MAINTAINERS entry CPU features: - Drop Kconfig options for PAN and LSE (these are detected at runtime) - Add support for 64-byte single-copy atomic instructions (LS64/LS64V) - Reduce MTE overhead when executing in the kernel on Ampere CPUs - Ensure POR_EL0 value exposed via ptrace is up-to-date - Fix error handling on GCS allocation failure CPU frequency: - Add CPU hotplug support to the FIE setup in the AMU driver Entry code: - Minor optimisations and cleanups to the syscall entry path - Preparatory rework for moving to the generic syscall entry code Hardware errata: - Work around Spectre-BHB on TSV110 processors - Work around broken CMO propagation on some systems with the SI-L1 interconnect Miscellaneous: - Disable branch profiling for arch/arm64/ to avoid issues with noinstr - Minor fixes and cleanups (kexec + ubsan, WARN_ONCE() instead of WARN_ON(), reduction of boolean expression) - Fix custom __READ_ONCE() implementation for LTO builds when operating on non-atomic types Perf and PMUs: - Support for CMN-600AE - Be stricter about supported hardware in the CMN driver - Support for DSU-110 and DSU-120 - Support for the cycles event in the DSU driver (alongside the dedicated cycles counter) - Use IRQF_NO_THREAD instead of IRQF_ONESHOT in the cxlpmu driver - Use !bitmap_empty() as a faster alternative to bitmap_weight() - Fix SPE error handling when failing to resume profiling Selftests: - Add support for the FORCE_TARGETS option to the arm64 kselftests - Avoid nolibc-specific my_syscall() function - Add basic test for the LS64 HWCAP - Extend fp-pidbench to cover additional workload patterns" * tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: (43 commits) perf/arm-cmn: Reject unsupported hardware configurations perf: arm_spe: Properly set hw.state on failures arm64/gcs: Fix error handling in arch_set_shadow_stack_status() arm64: Fix non-atomic __READ_ONCE() with CONFIG_LTO=y arm64: poe: fix stale POR_EL0 values for ptrace kselftest/arm64: Raise default number of loops in fp-pidbench kselftest/arm64: Add a no-SVE loop after SVE in fp-pidbench perf/cxlpmu: Replace IRQF_ONESHOT with IRQF_NO_THREAD arm64: mte: Set TCMA1 whenever MTE is present in the kernel arm64/ptrace: Return early for ptrace_report_syscall_entry() error arm64/ptrace: Split report_syscall() arm64: Remove unused _TIF_WORK_MASK kselftest/arm64: Add missing file in .gitignore arm64: errata: Workaround for SI L1 downstream coherency issue kselftest/arm64: Add HWCAP test for FEAT_LS64 arm64: Add support for FEAT_{LS64, LS64_V} KVM: arm64: Enable FEAT_{LS64, LS64_V} in the supported guest arm64: Provide basic EL2 setup for FEAT_{LS64, LS64_V} usage at EL0/1 KVM: arm64: Handle DABT caused by LS64* instructions on unsupported memory KVM: arm64: Add documentation for KVM_EXIT_ARM_LDST64B ...

cpufreq: Add new helper function returning cpufreq policy

2026-01-05T21:11:48+00:00

cpufreq_cpu_get_raw() gets cpufreq policy only if the CPU is in policy->cpus mask, which means the CPU is already online. But in some cases, the policy is needed before the CPU is added to cpus mask. Add a function to get the policy in these cases. Signed-off-by: Lifeng Zheng Reviewed-by: Rafael J. Wysocki (Intel) Reviewed-by: Jie Zhan Acked-by: Beata Michalska Signed-off-by: Will Deacon

cpufreq: cpufreq_boost_trigger_state() optimization

2025-12-16T12:09:48+00:00

Optimize the error handling code in cpufreq_boost_trigger_state(). Signed-off-by: Lifeng Zheng Acked-by: Viresh Kumar Reviewed-by: Jie Zhan [ rjw: Changelog edit ] Link: https://patch.msgid.link/20251202072727.1368285-3-zhenglifeng1@huawei.com Signed-off-by: Rafael J. Wysocki

cpufreq: Return -EOPNOTSUPP if no policy supports boost

2025-12-16T12:09:47+00:00

In cpufreq_boost_trigger_state(), if none of the the policies support boost, policy_set_boost() will not be called and this function will return 0. But it is better to return an error to indicate that the platform doesn't support boost. Signed-off-by: Lifeng Zheng Acked-by: Viresh Kumar Reviewed-by: Jie Zhan [ rjw: Subject and changelog edits ] Link: https://patch.msgid.link/20251202072727.1368285-2-zhenglifeng1@huawei.com Signed-off-by: Rafael J. Wysocki

cpufreq: Replace deprecated strcpy() in cpufreq_unregister_governor()

2025-10-20T19:25:36+00:00

strcpy() is deprecated; assign the NUL terminator directly instead. Link: https://github.com/KSPP/linux/issues/88 Signed-off-by: Thorsten Blum [ rjw: Subject tweaks ] Link: https://patch.msgid.link/20251017153354.82009-2-thorsten.blum@linux.dev Signed-off-by: Rafael J. Wysocki

cpufreq: preserve freq_table_sorted across suspend/hibernate

2025-10-20T19:01:35+00:00

During S3/S4 suspend and resume, cpufreq policies are not freed or recreated; the freq_table and policy structure remain intact. However, set_freq_table_sorted() currently resets policy->freq_table_sorted to UNSORTED unconditionally, which is unnecessary since the table order does not change across suspend/resume. This patch adds a check to skip validation if policy->freq_table_sorted is already ASCENDING or DESCENDING. This avoids unnecessary traversal of the frequency table on S3/S4 resume or repeated online events, reducing overhead while preserving correctness. Signed-off-by: Zihuan Zhang Link: https://patch.msgid.link/20251011072420.11495-1-zhangzihuan@kylinos.cn Signed-off-by: Rafael J. Wysocki