Re: [PATCH] timekeeping: Add a lockdep override in tick_freeze().

[Chris Bainbridge <chris.bainbridge@gmail.com>]

Hi,

I'm getting "WARNING: inconsistent lock state" on resume with this
commit (92e250c624ea37fde64bfd624fd2556f0d846f18):

[   43.747069] ACPI: \_SB_.PEP_: Successfully transitioned to state screen off
[   43.753611] ACPI: \_SB_.PEP_: Successfully transitioned to state lps0 ms entry
[   43.753777] ACPI: \_SB_.PEP_: Successfully transitioned to state lps0 entry
[   43.838479] amd_pmc AMDI0005:00: failed to set RTC: -22
[   43.838489] PM: suspend-to-idle
[   43.838542] amd_pmc: SMU idlemask s0i3: 0xc00e0eb5
[   46.246882] Timekeeping suspended for 2.528 seconds
[   46.249132] PM: Triggering wakeup from IRQ 9
[   46.249292] ACPI: PM: ACPI fixed event wakeup
[   46.249297] PM: resume from suspend-to-idle
[   46.250801] amd_pmc AMDI0005:00: Last suspend didn't reach deepest state
[   46.251096] ACPI: \_SB_.PEP_: Successfully transitioned to state lps0 exit
[   46.257167] ACPI: \_SB_.PEP_: Successfully transitioned to state lps0 ms exit
[   46.258450] ACPI: \_SB_.PEP_: Successfully transitioned to state screen on
[   46.259652] ACPI: EC: interrupt unblocked
[   46.291758] PM: noirq resume of devices complete after 32.885 msecs

[   46.291907] ================================
[   46.291909] WARNING: inconsistent lock state
[   46.291911] 6.15.0-rc1-00002-g92e250c624ea #379 Not tainted
[   46.291914] --------------------------------
[   46.291915] inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage.
[   46.291917] irq/9-acpi/142 [HC0[0]:SC0[0]:HE1:SE1] takes:
[   46.291921] ffffffff88e48d18 (rtc_lock){?...}-{3:3}, at: cmos_interrupt+0x21/0x110
[   46.291939] {IN-HARDIRQ-W} state was registered at:
[   46.291941]   lock_acquire+0xc9/0x2d0
[   46.291946]   _raw_spin_lock+0x30/0x40
[   46.291951]   cmos_interrupt+0x21/0x110
[   46.291956]   __handle_irq_event_percpu+0x81/0x290
[   46.291962]   handle_irq_event+0x38/0x70
[   46.291966]   handle_edge_irq+0x85/0x210
[   46.291970]   __common_interrupt+0x45/0xd0
[   46.291973]   common_interrupt+0x80/0xa0
[   46.291977]   asm_common_interrupt+0x26/0x40
[   46.291981]   cpuidle_enter_state+0x113/0x540
[   46.291985]   cpuidle_enter+0x2d/0x40
[   46.291989]   do_idle+0x1f5/0x250
[   46.291992]   cpu_startup_entry+0x29/0x30
[   46.291994]   start_secondary+0x10c/0x130
[   46.291998]   common_startup_64+0x13e/0x141
[   46.292002] irq event stamp: 877
[   46.292004] hardirqs last  enabled at (877): [<ffffffff88778a90>] _raw_spin_unlock_irqrestore+0x40/0x50
[   46.292010] hardirqs last disabled at (876): [<ffffffff8877884c>] _raw_spin_lock_irqsave+0x5c/0x60
[   46.292014] softirqs last  enabled at (362): [<ffffffff87b0dc3c>] __irq_exit_rcu+0xcc/0xf0
[   46.292019] softirqs last disabled at (353): [<ffffffff87b0dc3c>] __irq_exit_rcu+0xcc/0xf0
[   46.292022] 
               other info that might help us debug this:
[   46.292024]  Possible unsafe locking scenario:

[   46.292025]        CPU0
[   46.292026]        ----
[   46.292027]   lock(rtc_lock);
[   46.292029]   <Interrupt>
[   46.292030]     lock(rtc_lock);
[   46.292032] 
                *** DEADLOCK ***

[   46.292033] no locks held by irq/9-acpi/142.
[   46.292035] 
               stack backtrace:
[   46.292040] CPU: 1 UID: 0 PID: 142 Comm: irq/9-acpi Not tainted 6.15.0-rc1-00002-g92e250c624ea #379 PREEMPT(voluntary) 
[   46.292046] Hardware name: HP HP Pavilion Aero Laptop 13-be0xxx/8916, BIOS F.17 12/18/2024
[   46.292049] Call Trace:
[   46.292052]  <TASK>
[   46.292058]  dump_stack_lvl+0x6e/0x90
[   46.292063]  print_usage_bug.part.0+0x22c/0x2c0
[   46.292067]  mark_lock+0x6f7/0x890
[   46.292072]  ? __lock_acquire+0x449/0x21c0
[   46.292077]  __lock_acquire+0x7e5/0x21c0
[   46.292080]  ? psi_task_switch+0x10a/0x330
[   46.292086]  lock_acquire+0xc9/0x2d0
[   46.292088]  ? cmos_interrupt+0x21/0x110
[   46.292094]  ? acpi_write_bit_register+0x79/0xf0
[   46.292102]  _raw_spin_lock+0x30/0x40
[   46.292106]  ? cmos_interrupt+0x21/0x110
[   46.292110]  cmos_interrupt+0x21/0x110
[   46.292115]  rtc_handler+0x28/0xc0
[   46.292120]  acpi_ev_fixed_event_detect+0xc8/0x140
[   46.292127]  ? irq_thread+0xa4/0x340
[   46.292129]  acpi_ev_sci_xrupt_handler+0x13/0x40
[   46.292132]  acpi_irq+0x16/0x30
[   46.292136]  irq_thread_fn+0x1d/0x50
[   46.292139]  irq_thread+0x1d5/0x340
[   46.292140]  ? irq_copy_pending.isra.0+0x70/0x70
[   46.292144]  ? irq_finalize_oneshot.part.0+0xc0/0xc0
[   46.292147]  ? irq_forced_thread_fn+0x70/0x70
[   46.292149]  kthread+0x10a/0x250
[   46.292154]  ? kthreads_online_cpu+0x130/0x130
[   46.292158]  ret_from_fork+0x31/0x50
[   46.292162]  ? kthreads_online_cpu+0x130/0x130
[   46.292166]  ret_from_fork_asm+0x11/0x20
[   46.292174]  </TASK>
[   46.292918] PM: early resume of devices complete after 0.829 msecs