On 2019-02-14, Petr Mladek <pmladek@xxxxxxxx> wrote:
>>> cpu_store looks like an implementation detail. The caller
>>> needs to remember it to handle the nesting properly.
>>>
>>> We could achieve the same with a recursion counter hidden
>>> in struct prb_lock.
>
> The atomic operations are tricky. I feel other lost in them.
> Well, I still think that it might easier to detect nesting
> on the same CPU, see below.
>
> Also there is no need to store irq flags in per-CPU variable.
> Only the first owner of the lock need to store the flags. The others
> are spinning or nested.
>
> struct prb_cpulock {
> atomic_t owner;
> unsigned int flags;
> int nesting; /* intialized to 0 */
> };
>
> void prb_lock(struct prb_cpulock *cpu_lock)
> {
> unsigned int flags;
> int cpu;
I added an explicit preempt_disable here:
cpu = get_cpu();
> /*
> * The next condition might be valid only when
> * we are nested on the same CPU. It means
> * the IRQs are already disabled and no
> * memory barrier is needed.
> */
> if (cpu_lock->owner == smp_processor_id()) {
> cpu_lock->nested++;
> return;
> }
>
> /* Not nested. Take the lock */
> local_irq_save(flags);
> cpu = smp_processor_id();
>
> for (;;) {
With fixups so it builds/runs:
unsigned int prev_cpu = -1;
> if (atomic_try_cmpxchg_acquire(&cpu_lock->owner,
&prev_cpu, cpu)) {
> cpu_lock->flags = flags;
> break;
> }
>
> cpu_relax();
> }
> }
>
> void prb_unlock(struct prb_cpulock *cpu_lock)
> {
> unsigned int flags;
>
> if (cpu_lock->nested)
> cpu_lock->nested--;
And the matching preempt_enable().
goto out;
> }
>
> /* We must be the first lock owner */
> flags = cpu_lock->flags;
> atomic_set_release(&cpu_lock->owner, -1);
> local_irq_restore(flags);
out:
put_cpu();
> }
>
> Or do I miss anything?
It looks great. I've run my stress tests on it and everything is running
well.
Thanks for simplifying this!
John Ogness
