Date: Tue, 27 Oct 2015 19:40:01 +0700 From: Eugene Grosbein <eugen@grosbein.net> To: Adrian Chadd <adrian.chadd@gmail.com> Cc: "freebsd-mips@freebsd.org" <freebsd-mips@freebsd.org> Subject: Re: Strange results of TP-Link WDR3600 wired ethernet performace test Message-ID: <562F70A1.9010701@grosbein.net> In-Reply-To: <CAJ-VmokyYVPmsY9BU=tRV6zYWweqsECwmYexv_mMpbFob%2ByjRw@mail.gmail.com> References: <562E44E9.8090605@grosbein.net> <562E5D17.7030700@rdtc.ru> <CAJ-VmokyYVPmsY9BU=tRV6zYWweqsECwmYexv_mMpbFob%2ByjRw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 27.10.2015 00:49, Adrian Chadd wrote: > Hi, > > I'm glad that the throughput is higher - I was expecting exactly that. > > The interrupt storm - not sure. I haven't reproduced that at home. > There's some bit(s) in the interrupt status register that don't seem > to be cleared by taking action. I don't know what they are yet. To be precise: first it forwards traffic just fine, without interrupt storm and speed may raise over 22MByte/s but after a minute or less, interrupt storm begins. By default, hw.intr_storm_threshold is 1000 and it does not help increasing it upto 100,000. If I set it to zero (disable throttling), the device locks up solid when storm occurs: no traffic forwarding, no ping replies, even console is dead. Power cycle is required to revive it. This is 100% repeatable. I have PMC (performance monitoring counters) compiled in kernel. Now can I try to debug this?
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?562F70A1.9010701>