From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:40303) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1b83EL-0008Sw-2g for qemu-devel@nongnu.org; Wed, 01 Jun 2016 06:18:34 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1b83EJ-0007kh-38 for qemu-devel@nongnu.org; Wed, 01 Jun 2016 06:18:28 -0400 References: <201606010952.u519n6ri025307@mx0a-001b2d01.pphosted.com> From: Thomas Huth Message-ID: <574EB66E.1040805@redhat.com> Date: Wed, 1 Jun 2016 12:18:22 +0200 MIME-Version: 1.0 In-Reply-To: <201606010952.u519n6ri025307@mx0a-001b2d01.pphosted.com> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [RFC PATCH v0 0/2] Increase max memslots to 512 for PowerPC List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Bharata B Rao , qemu-devel@nongnu.org Cc: qemu-ppc@nongnu.org, david@gibson.dropbear.id.au, imammedo@redhat.com, pbonzini@redhat.com On 01.06.2016 11:51, Bharata B Rao wrote: > Recently the number of memory slots supported by KVM for PowerPC was ch= anged > from 32 to 512. QEMU was restricting the user specifiable hot-pluggable= memory > slots to 32. This patchset changes that to 512. >=20 > This allows more number of slots to be available for memory hotplugging= . It's certainly a good idea to increase the number of slots for hot-pluggable memory. But should we really increase it to the full maximum of 512 slots? The in-kernel slots are shared with other memory regions, too, e.g. the memory slots for PCI cards. So if you allow the users to plug all slots with DIMMs, they certainly will run into problems there again later. I think x86 is also using 256 DIMM slots only on purpose (see https://lkml.org/lkml/2014/11/14/328 for example), so we should maybe also limit the max. number of DIMM slots on spapr to kvm_get_max_memslots() divided by two ? Thomas