Re: Per-CPU Queueing for QoS

2017-11-14 Thread Michael Ma
2017-11-13 20:53 GMT-08:00 Tom Herbert : > On Mon, Nov 13, 2017 at 7:45 PM, John Fastabend > wrote: >> On 11/13/2017 06:18 PM, Michael Ma wrote: >>> 2017-11-13 16:13 GMT-08:00 Alexander Duyck : On Mon, Nov 13, 2017 at 3:08 PM, Eric Dumazet wrote: > On Mon, 2017-11-13 at 14:47 -0800

Re: Per-CPU Queueing for QoS

2017-11-14 Thread Michael Ma
2017-11-13 19:45 GMT-08:00 John Fastabend : > On 11/13/2017 06:18 PM, Michael Ma wrote: >> 2017-11-13 16:13 GMT-08:00 Alexander Duyck : >>> On Mon, Nov 13, 2017 at 3:08 PM, Eric Dumazet >>> wrote: On Mon, 2017-11-13 at 14:47 -0800, Alexander Duyck wrote: > On Mon, Nov 13, 2017 at 10:17 A

Re: Per-CPU Queueing for QoS

2017-11-14 Thread Michael Ma
2017-11-13 18:18 GMT-08:00 Michael Ma : > 2017-11-13 16:13 GMT-08:00 Alexander Duyck : >> On Mon, Nov 13, 2017 at 3:08 PM, Eric Dumazet wrote: >>> On Mon, 2017-11-13 at 14:47 -0800, Alexander Duyck wrote: On Mon, Nov 13, 2017 at 10:17 AM, Michael Ma wrote: > 2017-11-12 16:14 GMT-08:00 S

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Tom Herbert
On Mon, Nov 13, 2017 at 7:45 PM, John Fastabend wrote: > On 11/13/2017 06:18 PM, Michael Ma wrote: >> 2017-11-13 16:13 GMT-08:00 Alexander Duyck : >>> On Mon, Nov 13, 2017 at 3:08 PM, Eric Dumazet >>> wrote: On Mon, 2017-11-13 at 14:47 -0800, Alexander Duyck wrote: > On Mon, Nov 13, 201

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Dave Taht
I have been thinking we'd try to submit sch_cake to mainline on this go-around. It's been out of tree for way too long. I look forward to understanding your patches soon in the tbf case. (I'm only responding because cake uses deficit, rather than a token bucket, scheduler, and is not reliant on t

Re: Per-CPU Queueing for QoS

2017-11-13 Thread John Fastabend
On 11/13/2017 06:18 PM, Michael Ma wrote: > 2017-11-13 16:13 GMT-08:00 Alexander Duyck : >> On Mon, Nov 13, 2017 at 3:08 PM, Eric Dumazet wrote: >>> On Mon, 2017-11-13 at 14:47 -0800, Alexander Duyck wrote: On Mon, Nov 13, 2017 at 10:17 AM, Michael Ma wrote: > 2017-11-12 16:14 GMT-08:00

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Michael Ma
2017-11-13 16:13 GMT-08:00 Alexander Duyck : > On Mon, Nov 13, 2017 at 3:08 PM, Eric Dumazet wrote: >> On Mon, 2017-11-13 at 14:47 -0800, Alexander Duyck wrote: >>> On Mon, Nov 13, 2017 at 10:17 AM, Michael Ma wrote: >>> > 2017-11-12 16:14 GMT-08:00 Stephen Hemminger : >>> >> On Sun, 12 Nov 2017

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Michael Ma
2017-11-13 18:05 GMT-08:00 Michael Ma : > 2017-11-13 15:08 GMT-08:00 Eric Dumazet : >> On Mon, 2017-11-13 at 14:47 -0800, Alexander Duyck wrote: >>> On Mon, Nov 13, 2017 at 10:17 AM, Michael Ma wrote: >>> > 2017-11-12 16:14 GMT-08:00 Stephen Hemminger : >>> >> On Sun, 12 Nov 2017 13:43:13 -0800 >>

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Michael Ma
2017-11-13 15:08 GMT-08:00 Eric Dumazet : > On Mon, 2017-11-13 at 14:47 -0800, Alexander Duyck wrote: >> On Mon, Nov 13, 2017 at 10:17 AM, Michael Ma wrote: >> > 2017-11-12 16:14 GMT-08:00 Stephen Hemminger : >> >> On Sun, 12 Nov 2017 13:43:13 -0800 >> >> Michael Ma wrote: >> >> >> >>> Any commen

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Michael Ma
2017-11-13 14:47 GMT-08:00 Alexander Duyck : > On Mon, Nov 13, 2017 at 10:17 AM, Michael Ma wrote: >> 2017-11-12 16:14 GMT-08:00 Stephen Hemminger : >>> On Sun, 12 Nov 2017 13:43:13 -0800 >>> Michael Ma wrote: >>> Any comments? We plan to implement this as a qdisc and appreciate any ea

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Alexander Duyck
On Mon, Nov 13, 2017 at 3:08 PM, Eric Dumazet wrote: > On Mon, 2017-11-13 at 14:47 -0800, Alexander Duyck wrote: >> On Mon, Nov 13, 2017 at 10:17 AM, Michael Ma wrote: >> > 2017-11-12 16:14 GMT-08:00 Stephen Hemminger : >> >> On Sun, 12 Nov 2017 13:43:13 -0800 >> >> Michael Ma wrote: >> >> >> >>

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Eric Dumazet
On Mon, 2017-11-13 at 14:47 -0800, Alexander Duyck wrote: > On Mon, Nov 13, 2017 at 10:17 AM, Michael Ma wrote: > > 2017-11-12 16:14 GMT-08:00 Stephen Hemminger : > >> On Sun, 12 Nov 2017 13:43:13 -0800 > >> Michael Ma wrote: > >> > >>> Any comments? We plan to implement this as a qdisc and appre

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Alexander Duyck
On Mon, Nov 13, 2017 at 10:17 AM, Michael Ma wrote: > 2017-11-12 16:14 GMT-08:00 Stephen Hemminger : >> On Sun, 12 Nov 2017 13:43:13 -0800 >> Michael Ma wrote: >> >>> Any comments? We plan to implement this as a qdisc and appreciate any early >>> feedback. >>> >>> Thanks, >>> Michael >>> >>> > O

Re: Per-CPU Queueing for QoS

2017-11-13 Thread Michael Ma
2017-11-12 16:14 GMT-08:00 Stephen Hemminger : > On Sun, 12 Nov 2017 13:43:13 -0800 > Michael Ma wrote: > >> Any comments? We plan to implement this as a qdisc and appreciate any early >> feedback. >> >> Thanks, >> Michael >> >> > On Nov 9, 2017, at 5:20 PM, Michael Ma wrote: >> > >> > Currently

Re: Per-CPU Queueing for QoS

2017-11-12 Thread Stephen Hemminger
On Sun, 12 Nov 2017 13:43:13 -0800 Michael Ma wrote: > Any comments? We plan to implement this as a qdisc and appreciate any early > feedback. > > Thanks, > Michael > > > On Nov 9, 2017, at 5:20 PM, Michael Ma wrote: > > > > Currently txq/qdisc selection is based on flow hash so packets from

Re: Per-CPU Queueing for QoS

2017-11-12 Thread Michael Ma
Any comments? We plan to implement this as a qdisc and appreciate any early feedback. Thanks, Michael > On Nov 9, 2017, at 5:20 PM, Michael Ma wrote: > > Currently txq/qdisc selection is based on flow hash so packets from > the same flow will follow the order when they enter qdisc/txq, which >

Per-CPU Queueing for QoS

2017-11-09 Thread Michael Ma
Currently txq/qdisc selection is based on flow hash so packets from the same flow will follow the order when they enter qdisc/txq, which avoids out-of-order problem. To improve the concurrency of QoS algorithm we plan to have multiple per-cpu queues for a single TC class and do busy polling from a