Date: Tue, 9 Aug 2011 11:53:41 +0200 From: Marius Strobl <marius@alchemy.franken.de> To: kib@freebsd.org Cc: Peter Jeremy <peter.jeremy@alcatel-lucent.com>, "freebsd-sparc64@freebsd.org" <freebsd-sparc64@freebsd.org> Subject: Re: 'make -j16 universe' gives SIReset Message-ID: <20110809095341.GP48988@alchemy.franken.de> In-Reply-To: <20110809034004.GK6142@pjdesk.au.alcatel-lucent.com> References: <4E0F6B8D.8000500@rice.edu> <20110704214158.GX14797@alchemy.franken.de> <20110705160709.GA77843@alchemy.franken.de> <4E135420.4080201@rice.edu> <20110705190126.GE14797@alchemy.franken.de> <20110706042634.GP65891@pjdesk.au.alcatel-lucent.com> <20110706103910.GG14797@alchemy.franken.de> <20110706222851.GQ65891@pjdesk.au.alcatel-lucent.com> <20110806180537.GB48988@alchemy.franken.de> <20110809034004.GK6142@pjdesk.au.alcatel-lucent.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Aug 09, 2011 at 01:40:04PM +1000, Peter Jeremy wrote: > On 2011-Aug-07 02:05:37 +0800, Marius Strobl <marius@alchemy.franken.de> wrote: > >Could you please give the below patch a try? This is just a shot in the > >dark though. > > No change. Running pho@'s stress test with INCARNATIONS=150 results > in 'thr1' wedging after about 2 hrs (2 thr1 processes in "urdlck" and > a pile of zombies). > Kib, do you happen to have an idea what could be going on causing no process being able to grab the lock in do_rw_rdlock() (see also 20110706042634.GP65891@pjdesk.au.alcatel-lucent.com)? The whole user mutex implementation (both kernel and userland side) seems weird to me as things like _has_waiters and m_owner sometimes are updated atomically and sometimes not, sometimes using memory barriers and sometimes not ... Marius
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110809095341.GP48988>