From mboxrd@z Thu Jan 1 00:00:00 1970 From: Phil Carmody Date: Fri, 17 Sep 2010 08:44:11 +0000 Subject: Re: [PATCH 0/2] ia64: possible module unwind table optimisation Message-Id: <20100917084411.GC20325@esdhcp04044.research.nokia.com> List-Id: References: <1284115245-23876-1-git-send-email-ext-phil.2.carmody@nokia.com> <20100915124831.GB27658@esdhcp04044.research.nokia.com> <20100915222216.GF27658@esdhcp04044.research.nokia.com> In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ext Tony Luck Cc: "fenghua.yu@intel.com" , "linux-ia64@vger.kernel.org" , "linux-kernel@vger.kernel.org" On 16/09/10 23:56 +0200, ext Tony Luck wrote: > On Wed, Sep 15, 2010 at 3:22 PM, Phil Carmody > wrote: > > Turn on kmemleak, or anything else which repeatedly runs up the stack just > > for the fun of gathering backtraces. Kmemleak's what caused us to notice > > the issue in our ARM-based environment. > > kmemleak needs CONFIG_STACKTRACE - which I haven't got around to putting > into ia64 yet. I dug up an earlier attempt at doing so, and turned on kmemleak, > but it crashes early in boot (before ever getting to my stack tracer) > in possibly > the first call to kmem_cache_alloc(). With SLUB the error is an unaligned > reference to 0xffffffffffffffff, with SLAB I get a NULL deref. > > :-( This is a shame, but not a real problem in the context of my patch - you can't usefully optimise something that's seemingly almost never used. It can be filed in the cylindrical filing cabinet. Phil