From owner-freebsd-current@FreeBSD.ORG Tue Dec 23 21:37:26 2008 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id AC79B1065675 for ; Tue, 23 Dec 2008 21:37:26 +0000 (UTC) (envelope-from rmtodd@servalan.servalan.com) Received: from mx1.synetsystems.com (mx1.synetsystems.com [76.10.206.14]) by mx1.freebsd.org (Postfix) with ESMTP id 7E67C8FC1B for ; Tue, 23 Dec 2008 21:37:26 +0000 (UTC) (envelope-from rmtodd@servalan.servalan.com) Received: by mx1.synetsystems.com (Postfix, from userid 66) id B1911C97; Tue, 23 Dec 2008 16:37:25 -0500 (EST) Received: from rmtodd by servalan.servalan.com with local (Exim 4.69 (FreeBSD)) (envelope-from ) id 1LFE47-0001JS-1f; Tue, 23 Dec 2008 14:41:19 -0600 Date: Tue, 23 Dec 2008 14:41:18 -0600 From: Richard Todd To: Alan Cox Message-ID: <20081223204118.GA4507@ichotolot.servalan.com> References: <20081223011032.GA3228@ichotolot.servalan.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.2.3i Cc: freebsd-current@freebsd.org, rnoland@freebsd.org Subject: Re: Getting 'vm_thread_new: kstack allocation failed' errors after latest drm updates. X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 23 Dec 2008 21:37:26 -0000 On Tue, Dec 23, 2008 at 12:04:49PM -0600, Alan Cox wrote: > Roughly speaking, vm.kvm_* pertain to the kernel's address space and > vm.kmem_* pertain to the kernel's heap within its address space. Your > settings of vm.kmem_* have made the heap so large that there is little or no > free space left in the address space for other things. These other things, > e.g., the buffer cache, are non-trivial in size. > > So, the short answer is that you need to reduce vm.kmem_size. > > Alan Ah, so there's stuff the kernel allocates that's not inside the region controlled by kmem_size. Interesting. Anyway, knocking vm.kmem_size down by 128M seems to fix things. Thanks! Richard