Date: Sat, 22 Jul 2017 17:14:47 +0700 From: Eugene Grosbein <eugen@grosbein.net> To: Konstantin Belousov <kostikbel@gmail.com> Cc: Don Lewis <truckman@FreeBSD.org>, pz-freebsd-stable@ziemba.us, freebsd-stable@FreeBSD.org Subject: Re: stable/11 r321349 crashing immediately Message-ID: <59732597.4010603@grosbein.net> In-Reply-To: <20170722095725.GS1935@kib.kiev.ua> References: <201707220542.v6M5ggtP052112@gw.catspoiler.org> <5972E7C5.6070102@grosbein.net> <20170722070529.GP1935@kib.kiev.ua> <5973018B.2050505@grosbein.net> <20170722080012.GR1935@kib.kiev.ua> <59730ECA.7030309@grosbein.net> <20170722095725.GS1935@kib.kiev.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
On 22.07.2017 16:57, Konstantin Belousov wrote: >>From this description, I would be not even surprised if your machine > load fits into the kstacks cache, despite cache' quite conservative > settings. In other words, almost definitely your machine is not > representative for the problematic load. Something that creates a lot of > short- and middle- lived threads would be. I'm having trouble to imagine a real-world task for today's i386 hardware or virtual machine involving heavy usage of many short/middle-lived threads. Perhaps, you have at least a synthetic benchmark so I could try to reproduce kstack/KVA fragmentation-related problem using my i386 hardware? It has 1G RAM, local 16G CompactFlash (13 GB unpartitioned) as ada0, over 200GB free within local IDE HDD (ada1) and mentioned 2TB USB 2.0 HDD. Pretty many resources for 2007 year AMD Geode system, eh? :-) It runs 11.1-PRERELEASE r318642 currently.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?59732597.4010603>