Re: [Linuxarm] Re: [RFC v2] net: sched: implement TCQ_F_CAN_BYPASS for lockless qdisc

[Yunsheng Lin]

On 2021/3/20 3:03, Jason A. Donenfeld wrote:
> On Thu, Mar 18, 2021 at 1:33 AM Yunsheng Lin <linyunsheng@xxxxxxxxxx> wrote:
>>> That offer definitely still stands. Generalization sounds like a lot of fun.
>>>
>>> Keep in mind though that it's an eventually consistent queue, not an
>>> immediately consistent one, so that might not match all use cases. It
>>> works with wg because we always trigger the reader thread anew when it
>>> finishes, but that doesn't apply to everyone's queueing setup.
>>
>> Thanks for mentioning this.
>>
>> "multi-producer, single-consumer" seems to match the lockless qdisc's
>> paradigm too, for now concurrent enqueuing/dequeuing to the pfifo_fast's
>> queues() is not allowed, it is protected by producer_lock or consumer_lock.
> 
> The other thing is that if you've got memory for a ring buffer rather
> than a list queue, we worked on an MPMC ring structure for WireGuard a
> few years ago that we didn't wind up using in the end, but it lives
> here:
> https://git.zx2c4.com/wireguard-monolithic-historical/tree/src/mpmc_ptr_ring.h?h=tg/mpmc-benchmark

Thanks for mentioning that, It seems that is exactly what the
pfifo_fast qdisc need for locklees multi-producer, because it
only need the memory to store the skb pointer.

Does it have any limitation? More specifically, does it works with
the process or softirq context, if not, how about context with
rcu protection?

> _______________________________________________
> Linuxarm mailing list -- linuxarm@xxxxxxxxxxxxx
> To unsubscribe send an email to linuxarm-leave@xxxxxxxxxxxxx
>