Date: Sat, 9 Jan 2016 10:58:37 +0300 From: Chagin Dmitry <dchagin@freebsd.org> To: Gleb Smirnoff <glebius@FreeBSD.org> Cc: Adrian Chadd <adrian.chadd@gmail.com>, FreeBSD Net <freebsd-net@freebsd.org> Subject: Re: long network segment call stack can fail Message-ID: <20160109075837.GA4110@chd.heemeyer.club> In-Reply-To: <20160109033737.GN1906@FreeBSD.org> References: <20160106180008.GA1363@chd.heemeyer.club> <CAJ-Vmon_Va3SqUoSurju3_Wu2cV9pL55o9ENAksPwtkiNSUWyw@mail.gmail.com> <20160109033737.GN1906@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Jan 08, 2016 at 07:37:37PM -0800, Gleb Smirnoff wrote: > On Wed, Jan 06, 2016 at 12:50:52PM -0800, Adrian Chadd wrote: > A> hah, that's very pretty. The downside of a completely direct-dispatch > A> network stack. :) > > The stack is big, but shouldn't eat 4 pages anyway. Back in the netgraph yeha, exctly > hacking times, we used to have stack exhaustion with much longer backtraces. > > I expect that somewhere on the list there something huge on stack. > > I assume Dmitry runs with -O0, this is compile that exposes stack > exhaustion first. Did I guess it right, Dmitry? > no, -O2 here -- Have fun! chd
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20160109075837.GA4110>