Re: [LEDE-DEV] [PATCH 2/3] ag71xx: Remove ___cacheline_aligned from ring structs.

2018-02-06 Thread Rosen Penev
On Tue, Feb 6, 2018 at 8:27 PM, Florian Fainelli wrote: > On February 6, 2018 2:27:29 PM PST, Rosen Penev wrote: >>Qualcomm's struct members and inner workings of their driver are all >>different. >>While this might make sense for their driver, it seems to hurt here. In >>iperf3, i've seen incons

Re: [LEDE-DEV] [PATCH 2/3] ag71xx: Remove ___cacheline_aligned from ring structs.

2018-02-06 Thread Florian Fainelli
On February 6, 2018 2:27:29 PM PST, Rosen Penev wrote: >Qualcomm's struct members and inner workings of their driver are all >different. >While this might make sense for their driver, it seems to hurt here. In >iperf3, i've seen inconsistent results including a drop of 100mbps on >an >Archer C7v4.

[LEDE-DEV] [PATCH 2/3] ag71xx: Remove ___cacheline_aligned from ring structs.

2018-02-06 Thread Rosen Penev
Qualcomm's struct members and inner workings of their driver are all different. While this might make sense for their driver, it seems to hurt here. In iperf3, i've seen inconsistent results including a drop of 100mbps on an Archer C7v4. This patch keeps the results high and relatively consistent.