[PATCH] ACPI: CPPC: Make rmw_lock a raw_spin_lock

[Rafael J. Wysocki]

On Mon, Oct 28, 2024 at 11:52 AM Pierre Gondois <pierre.gondois at arm.com> wrote:
>
> The following BUG was triggered. sugov_update_shared() locks a
> raw_spinlock while cpc_write() locks a spinlock. To have a correct
> wait-type order, update rmw_lock to a raw_spinlock.
>
> Also save irq state.
>
> =============================
> [ BUG: Invalid wait context ]
> 6.12.0-rc2-XXX #406 Not tainted
> -----------------------------
> kworker/1:1/62 is trying to lock:
> ffffff8801593030 (&cpc_ptr->rmw_lock){+.+.}-{3:3}, at: cpc_write+0xcc/0x370
> other info that might help us debug this:
> context-{5:5}
> 2 locks held by kworker/1:1/62:
>   #0: ffffff897ef5ec98 (&rq->__lock){-.-.}-{2:2}, at: raw_spin_rq_lock_nested+0x2c/0x50
>   #1: ffffff880154e238 (&sg_policy->update_lock){....}-{2:2}, at: sugov_update_shared+0x3c/0x280
> stack backtrace:
> CPU: 1 UID: 0 PID: 62 Comm: kworker/1:1 Not tainted 6.12.0-rc2-g9654bd3e8806 #406
> Workqueue:  0x0 (events)
> Call trace:
>   dump_backtrace+0xa4/0x130
>   show_stack+0x20/0x38
>   dump_stack_lvl+0x90/0xd0
>   dump_stack+0x18/0x28
>   __lock_acquire+0x480/0x1ad8
>   lock_acquire+0x114/0x310
>   _raw_spin_lock+0x50/0x70
>   cpc_write+0xcc/0x370
>   cppc_set_perf+0xa0/0x3a8
>   cppc_cpufreq_fast_switch+0x40/0xc0
>   cpufreq_driver_fast_switch+0x4c/0x218
>   sugov_update_shared+0x234/0x280
>   update_load_avg+0x6ec/0x7b8
>   dequeue_entities+0x108/0x830
>   dequeue_task_fair+0x58/0x408
>   __schedule+0x4f0/0x1070
>   schedule+0x54/0x130
>   worker_thread+0xc0/0x2e8
>   kthread+0x130/0x148
>   ret_from_fork+0x10/0x20
>
> Fixes: 60949b7b8054 ("ACPI: CPPC: Fix MASK_VAL() usage")
> Signed-off-by: Pierre Gondois <pierre.gondois at arm.com>
> ---
>  drivers/acpi/cppc_acpi.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c
> index 1a40f0514eaa..e7e4bf932e28 100644
> --- a/drivers/acpi/cppc_acpi.c
> +++ b/drivers/acpi/cppc_acpi.c
> @@ -1127,7 +1127,7 @@ static int cpc_write(int cpu, struct cpc_register_resource *reg_res, u64 val)
>                         return -ENODEV;
>                 }
>
> -               raw_spin_lock_irqsave(&cpc_desc->rmw_lock, flags);
> +               raw_spin_lock(&cpc_desc->rmw_lock);

This won't apply because the current code is

spin_lock(&cpc_desc->rmw_lock);

and analogously below.  I think that it needs to be
raw_spin_lock_irqsave(), though.

Besides, a raw spin lock needs to be declared as such.

>                 switch (size) {
>                 case 8:
>                         prev_val = readb_relaxed(vaddr);
> @@ -1142,7 +1142,7 @@ static int cpc_write(int cpu, struct cpc_register_resource *reg_res, u64 val)
>                         prev_val = readq_relaxed(vaddr);
>                         break;
>                 default:
> -                       raw_spin_unlock_irqrestore(&cpc_desc->rmw_lock, flags);
> +                       raw_spin_unlock(&cpc_desc->rmw_lock);
>                         return -EFAULT;
>                 }
>                 val = MASK_VAL_WRITE(reg, prev_val, val);
> @@ -1175,7 +1175,7 @@ static int cpc_write(int cpu, struct cpc_register_resource *reg_res, u64 val)
>         }
>
>         if (reg->space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY)
> -               raw_spin_unlock_irqrestore(&cpc_desc->rmw_lock, flags);
> +               raw_spin_unlock(&cpc_desc->rmw_lock);
>
>         return ret_val;
>  }
> --
> 2.25.1
>