[PATCH v4 for 4.14 1/3] membarrier: Provide register expedited private command

[Mathieu Desnoyers]

----- On Sep 28, 2017, at 12:16 PM, Nicholas Piggin npiggin at gmail.com wrote:

> On Thu, 28 Sep 2017 15:29:50 +0000 (UTC)
> Mathieu Desnoyers <mathieu.desnoyers at efficios.com> wrote:
> 
>> ----- On Sep 28, 2017, at 11:01 AM, Nicholas Piggin npiggin at gmail.com wrote:
>> 
>> > On Thu, 28 Sep 2017 13:31:36 +0000 (UTC)
>> > Mathieu Desnoyers <mathieu.desnoyers at efficios.com> wrote:
>> >   
>> >> ----- On Sep 27, 2017, at 9:04 AM, Nicholas Piggin npiggin at gmail.com wrote:
>> >>   
> 
> [snip]
> 
>> >> So I don't see much point in trying to remove that registration step.
>> > 
>> > I don't follow you. You are talking about the concept of registering
>> > intention to use a different function? And the registration API is not
>> > merged yet?
>> 
>> Yes, I'm talking about requiring processes to invoke membarrier cmd
>> MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED before they can successfully
>> invoke membarrier cmd MEMBARRIER_CMD_PRIVATE_EXPEDITED.
>> 
>> > Let me say I'm not completely against the idea of a registration API. But
>> > don't think registration for this expedited command is necessary.
>> 
>> Given that we have the powerpc lack-of-full-barrier-on-return-to-userspace
>> case now, and we foresee x86-sysexit, sparc, and alpha also requiring
>> special treatment when we introduce the MEMBARRIER_FLAG_SYNC_CORE behavior
>> in the next release, it seems that we'll have a hard time handling
>> architecture special cases efficiently if we don't expose the registration
>> API right away.
> 
> But SYNC_CORE is a different functionality, right? You can add the
> registration API for it when that goes in.

Sure, I could. However, I was hoping to re-use the same command, with a
"SYNC_CORE" flag, and I would have liked to have consistent behavior
for same commands used with different flags.

> 
>> > But (aside) let's say a tif flag turns out to be a good diea for your
>> > second case, why not just check the flag in the membarrier sys call and
>> > do the registration the first time it uses it?
>> 
>> We also considered that option. It's mainly about guaranteeing that
>> an expedited membarrier command never blocks. If we introduce this
>> "lazy auto-registration" behavior, we end up blocking the process
>> at a random point in its execution so we can issue a synchronize_sched().
>> By exposing an explicit registration, we can control where this delay
>> occurs, and even allow library constructors to invoke the registration
>> while the process is a single threaded, therefore allowing us to completely
>> skip synchronize_sched().
> 
> Okay I guess that could be a good reason. As I said I'm not opposed to
> the concept. I suppose you could even have a registration for expedited
> private even if it's a no-op on all architectures, just in case some new
> ways of implementing it can be done in future.

That's an approach I would be OK with too. Mandating explicit registration
will give us much more flexibility.

> I suppose I'm more objecting to the added complexity for powerpc, and
> more code in the fastpath to make the slowpath faster.

Just to make sure I understand your concern here. The "fastpath" you
refer to is the TIF flag test in membarrier_sched_in() within
finish_task_switch(), and the "slowpath" is switch_mm() which lacks
the required full barrier now, am I correct ?

Would it help if we invoke the membarrier hook from switch_mm()
instead ? We'd therefore only add the TIF flag test in switch_mm(),
rather than for every context switch.

Thanks,

Mathieu


> 
> Thanks,
> Nick

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com