Re: [net] b58537a1f56: +89.2% netperf.Throughput_Mbps

2014-06-19 Thread Daniel Borkmann
On 06/18/2014 10:41 PM, David Rientjes wrote: On Wed, 18 Jun 2014, Daniel Borkmann wrote: One assumption could be that netperf was setting this knob internally and ignoring the result of it (error/success), but that we might need to look up in netperf source to confirm. Confirmed netperf-2.4.

Re: [net] b58537a1f56: +89.2% netperf.Throughput_Mbps

2014-06-18 Thread David Rientjes
On Wed, 18 Jun 2014, Daniel Borkmann wrote: > One assumption could be that netperf was setting this knob > internally and ignoring the result of it (error/success), but > that we might need to look up in netperf source to confirm. > Confirmed netperf-2.4.5 doesn't do this. -- To unsubscribe from

Re: [net] b58537a1f56: +89.2% netperf.Throughput_Mbps

2014-06-18 Thread Daniel Borkmann
On 06/18/2014 10:29 PM, David Rientjes wrote: On Wed, 18 Jun 2014, Jet Chen wrote: Hi Daniel, FYI, we noticed the below changes on commit b58537a1f5629bdc98a8b9dc2051ce0e952f6b4b ("net: sctp: fix permissions for rto_alpha and rto_beta knobs") There's no way this is legitimate, that commit c

Re: [net] b58537a1f56: +89.2% netperf.Throughput_Mbps

2014-06-18 Thread David Rientjes
On Wed, 18 Jun 2014, Jet Chen wrote: > Hi Daniel, > > FYI, we noticed the below changes on > > commit b58537a1f5629bdc98a8b9dc2051ce0e952f6b4b ("net: sctp: fix permissions > for rto_alpha and rto_beta knobs") > There's no way this is legitimate, that commit cannot possibly affect the performa

[net] b58537a1f56: +89.2% netperf.Throughput_Mbps

2014-06-17 Thread Jet Chen
Hi Daniel, FYI, we noticed the below changes on commit b58537a1f5629bdc98a8b9dc2051ce0e952f6b4b ("net: sctp: fix permissions for rto_alpha and rto_beta knobs") e4f7ae930afafd4 b58537a1f5629bdc98a8b9dc2 --- - 950 ~ 1% +93.7% 1841 ~ 0% lk