Date: Fri, 8 Jun 2018 11:16:42 -0700 From: Gleb Smirnoff <glebius@FreeBSD.org> To: Kubilay Kocak <koobs@FreeBSD.org> Cc: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r334819 - head/sys/vm Message-ID: <20180608181641.GN1005@FreeBSD.org> In-Reply-To: <77e67765-4244-e207-ff8d-aa6156c985b8@FreeBSD.org> References: <201806080015.w580F8fM062154@repo.freebsd.org> <77e67765-4244-e207-ff8d-aa6156c985b8@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Jun 08, 2018 at 03:07:07PM +1000, Kubilay Kocak wrote: K> > UMA memory debugging enabled with INVARIANTS consists of two things: K> > trashing freed memory and checking that allocated memory is properly K> > trashed, and also of keeping a bitset of freed items. Trashing/checking K> > creates a lot of CPU cache poisoning, while keeping debugging bitsets K> > consistent creates a lot of contention on UMA zone lock(s). The performance K> > difference between INVARIANTS kernel and normal one is mostly attributed K> > to UMA debugging, rather than to all KASSERT checks in the kernel. K> > K> > Add loader tunable vm.debug.divisor that allows either to turn off UMA K> K> Is 'sample interval' a standard/common enough term for this kind of K> mechanism to name the sysctl with it rather than the implementation? K> K> Or 'sample frequency' Interval definitely doesn't fit here. Frequency is closer, but still not the right term, IMHO. Native speaker required here to judge. I am okay if anyone who is confident changes wording here. -- Gleb Smirnoff
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20180608181641.GN1005>