On Fri, 2019-02-01 at 15:40 +, Eads, Gage wrote:
>
> >
> > -Original Message-
> > From: Ola Liljedahl [mailto:ola.liljed...@arm.com]
> > Sent: Wednesday, January 30, 2019 5:36 AM
> > To: Honnappa Nagarahalli ; Richardson,
> > Bruce ; Eads, Gage ;
> > dev@dpdk.org
> > Cc: nd
> > Subje
> -Original Message-
> From: Ola Liljedahl [mailto:ola.liljed...@arm.com]
> Sent: Wednesday, January 30, 2019 5:36 AM
> To: Honnappa Nagarahalli ; Richardson,
> Bruce ; Eads, Gage ;
> dev@dpdk.org
> Cc: nd
> Subject: Re: [RFC] lfring: lock-free ring buffer
>
> On Wed, 2019-01-30 at 05:1
On Wed, 2019-01-30 at 05:17 +, Eads, Gage wrote:
>
> > > 2. On the plus side, the enqueue function design that allows it to use
> > > 1/2 the atomics of mine appears to be independent of reserving ring
> > > slots, and should transfer over fairly cleanly. I'm a little concerned
> > > about th
> On Jan 29, 2019, at 11:17 PM, Eads, Gage wrote:
>
>
>
>> -Original Message-
>> From: Ola Liljedahl [mailto:ola.liljed...@arm.com]
>> Sent: Monday, January 28, 2019 4:26 PM
>> To: Honnappa Nagarahalli ; Richardson,
>> Bruce ; Eads, Gage ;
>> dev@dpdk.org
>> Cc: nd
>> Subject: Re:
> -Original Message-
> From: Ola Liljedahl [mailto:ola.liljed...@arm.com]
> Sent: Monday, January 28, 2019 4:26 PM
> To: Honnappa Nagarahalli ; Richardson,
> Bruce ; Eads, Gage ;
> dev@dpdk.org
> Cc: nd
> Subject: Re: [RFC] lfring: lock-free ring buffer
>
> On Mon, 2019-01-28 at 21:04 +
On Mon, 2019-01-28 at 21:04 +, Eads, Gage wrote:
> Hey Ola,
>
> >
> > -Original Message-
> > From: Ola Liljedahl [mailto:ola.liljed...@arm.com]
> > Sent: Monday, January 28, 2019 6:29 AM
> > To: dev@dpdk.org; Eads, Gage ; Honnappa Nagarahalli
> > ; Richardson, Bruce
> >
> > Cc: nd ;
Hey Ola,
> -Original Message-
> From: Ola Liljedahl [mailto:ola.liljed...@arm.com]
> Sent: Monday, January 28, 2019 6:29 AM
> To: dev@dpdk.org; Eads, Gage ; Honnappa Nagarahalli
> ; Richardson, Bruce
>
> Cc: nd ; Ola Liljedahl
> Subject: [RFC] lfring: lock-free ring buffer
>
> Lock-free
Lock-free MP/MC ring buffer with SP/SC options.
The ring buffer metadata only maintains one set of head and tail
pointers. Tail is updated on enqueue, head is updated on dequeue.
The ring slots between head and tail always contains valid (unconsumed)
slots.
Each ring slot consists of a struct of da
8 matches
Mail list logo
