kernel-aes67/include/asm-generic
Mathieu Desnoyers fe90f3967b sched: Add missing memory barrier in switch_mm_cid
Many architectures' switch_mm() (e.g. arm64) do not have an smp_mb()
which the core scheduler code has depended upon since commit:

    commit 223baf9d17 ("sched: Fix performance regression introduced by mm_cid")

If switch_mm() doesn't call smp_mb(), sched_mm_cid_remote_clear() can
unset the actively used cid when it fails to observe active task after it
sets lazy_put.

There *is* a memory barrier between storing to rq->curr and _return to
userspace_ (as required by membarrier), but the rseq mm_cid has stricter
requirements: the barrier needs to be issued between store to rq->curr
and switch_mm_cid(), which happens earlier than:

  - spin_unlock(),
  - switch_to().

So it's fine when the architecture switch_mm() happens to have that
barrier already, but less so when the architecture only provides the
full barrier in switch_to() or spin_unlock().

It is a bug in the rseq switch_mm_cid() implementation. All architectures
that don't have memory barriers in switch_mm(), but rather have the full
barrier either in finish_lock_switch() or switch_to() have them too late
for the needs of switch_mm_cid().

Introduce a new smp_mb__after_switch_mm(), defined as smp_mb() in the
generic barrier.h header, and use it in switch_mm_cid() for scheduler
transitions where switch_mm() is expected to provide a memory barrier.

Architectures can override smp_mb__after_switch_mm() if their
switch_mm() implementation provides an implicit memory barrier.
Override it with a no-op on x86 which implicitly provide this memory
barrier by writing to CR3.

Fixes: 223baf9d17 ("sched: Fix performance regression introduced by mm_cid")
Reported-by: levi.yun <yeoreum.yun@arm.com>
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com> # for arm64
Acked-by: Dave Hansen <dave.hansen@linux.intel.com> # for x86
Cc: <stable@vger.kernel.org> # 6.4.x
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/r/20240415152114.59122-2-mathieu.desnoyers@efficios.com
2024-04-16 13:59:45 +02:00
..
bitops riscv: Avoid code duplication with generic bitops implementation 2024-01-24 17:25:36 -08:00
vdso lib/vdso: Avoid highres update if clocksource is not VDSO capable 2020-02-17 20:12:17 +01:00
Kbuild cfi: Flip headers 2023-12-15 16:25:55 -08:00
access_ok.h uaccess: remove CONFIG_SET_FS 2022-02-25 09:36:06 +01:00
agp.h char/agp: introduce asm-generic/agp.h 2023-02-13 22:13:29 +01:00
archrandom.h random: handle archrandom with multiple longs 2022-07-25 13:26:14 +02:00
asm-offsets.h
asm-prototypes.h
atomic.h locking/atomic: make atomic*_{cmp,}xchg optional 2023-06-05 09:57:14 +02:00
atomic64.h locking/atomic: delete !ARCH_ATOMIC remnants 2021-05-26 13:20:52 +02:00
audit_change_attr.h
audit_dir_write.h
audit_read.h
audit_signal.h
audit_write.h
barrier.h sched: Add missing memory barrier in switch_mm_cid 2024-04-16 13:59:45 +02:00
bitops.h include: move find.h from asm_generic to linux 2022-01-15 08:47:31 -08:00
bitsperlong.h lib: extend the scope of small_const_nbits() macro 2021-05-06 19:24:11 -07:00
bug.h bug: Fix no-return-statement warning with !CONFIG_BUG 2024-04-10 22:01:35 +02:00
cache.h
cacheflush.h mm: Introduce flush_cache_vmap_early() 2023-12-14 00:23:17 -08:00
cfi.h cfi: Flip headers 2023-12-15 16:25:55 -08:00
checksum.h asm-generic: Improve csum_fold 2024-01-17 17:52:29 -08:00
cmpxchg-local.h asm-generic: Fix 32 bit __generic_cmpxchg_local 2024-01-05 23:19:14 +01:00
cmpxchg.h asm-generic: avoid __generic_cmpxchg_local warnings 2023-04-04 17:58:11 +02:00
compat.h asm-generic: compat: fix compat_arg_u64() and compat_arg_u64_dual() 2022-11-01 10:20:11 +11:00
current.h asm-generic: current: Don't include thread-info.h if building asm 2023-08-26 22:38:49 +02:00
delay.h
device.h
div64.h ARM: 9117/1: asm-generic: div64: Remove always-true __div64_const32_is_OK() 2021-08-20 11:39:28 +01:00
dma-mapping.h dma-mapping: no need to pass a bus_type into get_arch_dma_ops() 2023-02-15 12:35:20 +01:00
dma.h
early_ioremap.h mm/early_ioremap.c: remove redundant early_ioremap_shutdown() 2021-09-08 11:50:24 -07:00
emergency-restart.h
error-injection.h docs: fault-injection: add requirements of error injectable functions 2023-02-02 22:50:00 -08:00
exec.h
extable.h
fb.h fbdev: Replace fb_pgprotect() with pgprot_framebuffer() 2023-10-12 09:20:46 +02:00
fixmap.h
flat.h
ftrace.h
futex.h futex: Fix additional regressions 2021-12-11 23:31:51 +01:00
getorder.h asm-generic: force inlining of get_order() to work around gcc10 poor decision 2020-12-15 22:46:15 -08:00
hardirq.h irqstat: Move declaration into asm-generic/hardirq.h 2020-11-23 10:31:06 +01:00
hugetlb.h mm: hugetlb: add huge page size param to set_huge_pte_at() 2023-09-29 17:20:47 -07:00
hw_irq.h
hyperv-tlfs.h hyperv-fixes for v6.9-rc4 2024-04-11 16:23:56 -07:00
int-ll64.h
io.h mm: ioremap: remove unneeded ioremap_allowed and iounmap_allowed 2023-08-18 10:12:36 -07:00
ioctl.h
iomap.h asm-generic/iomap.h: remove ARCH_HAS_IOREMAP_xx macros 2023-08-18 10:12:32 -07:00
irq.h
irq_regs.h
irq_work.h
irqflags.h
kdebug.h
kmap_size.h mm/highmem: Provide and use CONFIG_DEBUG_KMAP_LOCAL 2020-11-24 14:42:08 +01:00
kprobes.h treewide: Convert macro and uses of __section(foo) to __section("foo") 2020-10-25 14:51:49 -07:00
kvm_para.h
kvm_types.h KVM: Move x86's version of struct kvm_mmu_memory_cache to common code 2020-07-09 13:29:42 -04:00
linkage.h
local.h locking/generic: Wire up local{,64}_try_cmpxchg() 2023-04-29 09:09:09 +02:00
local64.h locking/generic: Wire up local{,64}_try_cmpxchg() 2023-04-29 09:09:09 +02:00
logic_io.h logic_io instance of iounmap() needs volatile on argument 2021-12-21 21:31:08 +01:00
mcs_spinlock.h
memory_model.h mm, arch: add generic implementation of pfn_valid() for FLATMEM 2023-02-09 16:51:41 -08:00
mm_hooks.h mm: remove arch_bprm_mm_init() hook 2020-01-23 10:41:16 -08:00
mmiowb.h asm-generic/mmiowb: Allow mmiowb_set_pending() when preemptible() 2020-07-17 10:02:03 +01:00
mmiowb_types.h
mmu.h
mmu_context.h asm-generic: add generic MMU versions of mmu context functions 2020-10-26 16:45:03 +01:00
module.h
module.lds.h kbuild: preprocess module linker script 2020-09-25 00:36:41 +09:00
mshyperv.h hyperv-fixes for v6.9-rc4 2024-04-11 16:23:56 -07:00
msi.h genirq: Get rid of GENERIC_MSI_IRQ_DOMAIN 2022-11-17 15:15:20 +01:00
nommu_context.h asm-generic: add generic MMU versions of mmu context functions 2020-10-26 16:45:03 +01:00
numa.h arm64: irq: set the correct node for VMAP stack 2023-12-05 14:26:50 +00:00
page.h asm-generic/page.h: Make pfn accessors static inlines 2023-05-29 11:27:08 +02:00
param.h
parport.h
pci.h asm-generic: Add new pci.h and use it 2022-07-22 17:34:57 -05:00
pci_iomap.h PCI: Stub __pci_ioport_map() for arches that don't support it at all 2022-07-29 12:01:00 -05:00
percpu.h percpu: Fix self-assignment of __old in raw_cpu_generic_try_cmpxchg() 2023-06-08 10:28:39 +02:00
pgalloc.h mm: add statistics for PUD level pagetable 2023-10-06 14:44:10 -07:00
pgtable-nop4d.h mm: rename p4d_page_vaddr to p4d_pgtable and make it return pud_t * 2021-07-08 11:48:22 -07:00
pgtable-nopmd.h mm: recover pud_leaf() definitions in nopmd case 2024-03-13 12:12:21 -07:00
pgtable-nopud.h mm: rename p4d_page_vaddr to p4d_pgtable and make it return pud_t * 2021-07-08 11:48:22 -07:00
pgtable_uffd.h userfaultfd: wp: add pmd_swp_*uffd_wp() helpers 2020-04-07 10:43:39 -07:00
preempt.h riscv: support PREEMPT_DYNAMIC with static keys 2023-08-31 00:18:34 -07:00
qrwlock.h asm-generic changes for 5.19 2022-05-26 10:50:30 -07:00
qrwlock_types.h locking/qrwlock: Change "queue rwlock" to "queued rwlock" 2022-05-11 16:27:04 +02:00
qspinlock.h asm-generic: qspinlock: fix queued_spin_value_unlocked() implementation 2023-11-22 09:32:49 -08:00
qspinlock_types.h locking/qspinlock: Do not include atomic.h from qspinlock_types.h 2020-07-29 16:14:19 +02:00
resource.h
rwonce.h asm/rwonce: Don't pull <asm/barrier.h> into 'asm-generic/rwonce.h' 2020-07-21 10:50:36 +01:00
seccomp.h seccomp: Use -1 marker for end of mode 1 syscall list 2020-07-10 16:01:52 -07:00
sections.h asm-generic: sections: refactor memory_intersects 2022-08-28 14:02:45 -07:00
serial.h
set_memory.h
shmparam.h
signal.h asm-generic: Remove empty #ifdef SA_RESTORER 2022-09-10 09:56:53 +02:00
simd.h
softirq_stack.h asm-generic: Conditionally enable do_softirq_own_stack() via Kconfig. 2022-09-05 17:20:55 +02:00
spinlock.h asm-generic: ticket-lock: Optimize arch_spin_value_unlocked() 2023-09-21 10:17:00 +02:00
spinlock_types.h asm-generic: ticket-lock: New generic ticket-based spinlock 2022-05-11 11:49:38 -07:00
statfs.h
string.h
switch_to.h
syscall.h ptrace: Create ptrace_report_syscall_{entry,exit} in ptrace.h 2022-03-10 13:35:08 -06:00
syscalls.h
timex.h
tlb.h mm/mmu_gather: add __tlb_remove_folio_pages() 2024-02-22 15:27:17 -08:00
tlbflush.h
topology.h mm: replace CONFIG_NEED_MULTIPLE_NODES with CONFIG_NUMA 2021-06-29 10:53:55 -07:00
trace_clock.h
uaccess.h uaccess: remove CONFIG_SET_FS 2022-02-25 09:36:06 +01:00
unaligned.h asm-generic: make sparse happy with odd-sized put_unaligned_*() 2024-01-08 09:48:35 -08:00
user.h
vermagic.h arch: split MODULE_ARCH_VERMAGIC definitions out to <asm/vermagic.h> 2020-04-23 10:50:26 +09:00
vga.h
vmlinux.lds.h treewide: update LLVM Bugzilla links 2024-02-22 15:38:51 -08:00
vtime.h
word-at-a-time.h kernel.h: removed REPEAT_BYTE from kernel.h 2024-02-01 09:47:59 -08:00
xor.h lib/xor: make xor prototypes more friendly to compiler vectorization 2022-02-11 20:39:39 +11:00