Re: [PATCH v2] locking/local_lock, mm: Replace localtry_ helpers with local_trylock_t type

[Vlastimil Babka <vbabka@suse.cz>]

On 4/2/25 23:40, Alexei Starovoitov wrote:
> On Wed, Apr 2, 2025 at 1:56 PM Shakeel Butt <shakeel.butt@linux.dev> wrote:
>>
>> On Tue, Apr 01, 2025 at 01:52:45PM -0700, Alexei Starovoitov wrote:
>> > From: Alexei Starovoitov <ast@kernel.org>
>> >
>> > Partially revert commit 0aaddfb06882 ("locking/local_lock: Introduce localtry_lock_t").
>> > Remove localtry_*() helpers, since localtry_lock() name might
>> > be misinterpreted as "try lock".
>> >
>> > Introduce local_trylock[_irqsave]() helpers that only work
>> > with newly introduced local_trylock_t type.
>> > Note that attempt to use local_trylock[_irqsave]() with local_lock_t
>> > will cause compilation failure.
>> >
>> > Usage and behavior in !PREEMPT_RT:
>> >
>> > local_lock_t lock;                     // sizeof(lock) == 0
>> > local_lock(&lock);                     // preempt disable
>> > local_lock_irqsave(&lock, ...);        // irq save
>> > if (local_trylock_irqsave(&lock, ...)) // compilation error
>> >
>> > local_trylock_t lock;                  // sizeof(lock) == 4
>>
>> Is there a reason for this 'acquired' to be int? Can it be uint8_t? No
>> need to change anything here but I plan to change it later to compact as
>> much as possible within one (or two) cachline for memcg stocks.
> 
> I don't see any issue. I can make it u8 right away.

Are you planning to put the lock near other <64bit sized values in memcg
stock? Otherwise it will be padded anyway?

I hope it won't hurt the performance though, AFAIK at least sub-word atomics
are much slower than using a full word. But we use only read/write once for
acquired so hopefully it's fine?

>> > local_lock(&lock);                     // preempt disable, acquired = 1
>> > local_lock_irqsave(&lock, ...);        // irq save, acquired = 1
>> > if (local_trylock(&lock))              // if (!acquired) preempt disable
>> > if (local_trylock_irqsave(&lock, ...)) // if (!acquired) irq save
>>
>> For above two ", acquired = 1" as well.
> 
> I felt it would be too verbose and not accurate anyway,
> since irq save will be done before the check.
> It's a pseudo code.
> But sure, I can add.
> 
>>
>> >
>> > The existing local_lock_*() macros can be used either with
>> > local_lock_t or local_trylock_t.
>> > With local_trylock_t they set acquired = 1 while local_unlock_*() clears it.
>> >
>> > In !PREEMPT_RT local_lock_irqsave(local_lock_t *) disables interrupts
>> > to protect critical section, but it doesn't prevent NMI, so the fully
>> > reentrant code cannot use local_lock_irqsave(local_lock_t *) for
>> > exclusive access.
>> >
>> > The local_lock_irqsave(local_trylock_t *) helper disables interrupts
>> > and sets acquired=1, so local_trylock_irqsave(local_trylock_t *) from
>> > NMI attempting to acquire the same lock will return false.
>> >
>> > In PREEMPT_RT local_lock_irqsave() maps to preemptible spin_lock().
>> > Map local_trylock_irqsave() to preemptible spin_trylock().
>> > When in hard IRQ or NMI return false right away, since
>> > spin_trylock() is not safe due to explicit locking in the underneath
>> > rt_spin_trylock() implementation. Removing this explicit locking and
>> > attempting only "trylock" is undesired due to PI implications.
>> >
>> > The local_trylock() without _irqsave can be used to avoid the cost of
>> > disabling/enabling interrupts by only disabling preemption, so
>> > local_trylock() in an interrupt attempting to acquire the same
>> > lock will return false.
>> >
>> > Note there is no need to use local_inc for acquired variable,
>> > since it's a percpu variable with strict nesting scopes.
>> >
>> > Acked-by: Vlastimil Babka <vbabka@suse.cz>
>> > Signed-off-by: Alexei Starovoitov <ast@kernel.org>
>>
>> Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>
> 
> Thanks!