Trouble shooting low rate MCS9 in 802.11ac

[Ben Greear]

On 07/28/2014 04:53 AM, Vu Hai NGUYEN wrote:
>>> My 2 devices have an ARM processor 1200MHz.
>>> If I connect these two devices by an Ethernet cable and doing iperf (in TCP)
>>> I can obtain 880 Mbps, it means that my processor can deal with that speed
>>> of data transfer. But if I use Wi-fi instead of Ethernet to test rate between these
>>> devices, my processor is exhausted, I run command top and see that 0~1%
>>> idle only,  ksoftirqd and kworker take most of my processor resource.
>>>
>>> Last week I discovered by chance that the module loaded in kernel can create
>>> effects to performance of rate, in my case it is "iptables". If I unload all module
>>> that related to iptables I can improve the performance.
>>> So I wonder if I can do some thing with the Wi-fi driver to  improve rate?
>>> Or the only solution is overclock/change my processor?
> 
>> Can you run 'perf top' on your system when it is under load so we can
>> see where the CPU is being used the most?
> 
> I follow a tutorial to cross compile perf so here is my result
> 
> 
> PerfTop:      70 irqs/sec  kernel:98.6%  exact:  0.0% [1000Hz cycles],  (all, 1 CPU)
> -------------------------------------------------------------------------------
> 
>     39.43%  [kernel]       [k] csum_partial_copy_nocheck

Looks like your ethernet driver is not doing checksum offloading?

I am pretty sure ath10k does support this unless you have managed
to disabled it somehow?

Thanks,
Ben

>      5.54%  [kernel]       [k] __kmalloc_track_caller
>      4.93%  [kernel]       [k] kfree
>      4.31%  [mac80211]     [k] ieee80211_tx_status
>      4.31%  [kernel]       [k] memcpy
>      3.49%  [kernel]       [k] __memzero
>      3.29%  [kernel]       [k] mv_eth_poll
>      3.29%  [kernel]       [k] __kfree_skb
>      2.05%  [ath10k_pci]   [k] ath10k_ce_send_nolock
>      1.44%  [mac80211]     [k] $a
>      1.44%  [kernel]       [k] dev_gro_receive
>      1.44%  [kernel]       [k] napi_gro_receive
>      1.44%  [kernel]       [k] eth_type_trans
>      1.44%  [kernel]       [k] inet_gro_receive
>      1.44%  [kernel]       [k] skb_copy_and_csum_bits
>      1.44%  [kernel]       [k] skb_segment
>      1.44%  [ath10k_core]  [k] ath10k_htt_rx_amsdu_pop
>      1.44%  [ath10k_core]  [k] ath10k_txrx_tx_unref
>      1.23%  [kernel]       [k] local_bh_enable                                  
>      1.23%  [kernel]       [k] kmem_cache_alloc                                 
>      1.23%  [kernel]       [k] dev_hard_start_xmit 
> 
> 
> Thank you in advance for your response,
> 
> NGUYEN Vu Hai
> Acita-Sodielec
> Route de Mayres - B.P. 9
> 12100 St GEORGES DE LUZENCON
> FRANCE
> 
> 
> 
> Thanks,
> Ben
> 
>>
>> NGUYEN Vu Hai
>> Acita-Sodielec
>> Route de Mayres - B.P. 9
>> 12100 St GEORGES DE LUZENCON
>> FRANCE
>>
>>
>>
>>
>> _______________________________________________
>> ath10k mailing list
>> ath10k at lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/ath10k
>>
> 
> --
> Ben Greear <greearb at candelatech.com>
> Candela Technologies Inc  http://www.candelatech.com
> 


-- 
Ben Greear <greearb at candelatech.com>
Candela Technologies Inc  http://www.candelatech.com