From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:52649) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VBJvO-0003tR-Oj for qemu-devel@nongnu.org; Mon, 19 Aug 2013 03:30:59 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VBJvH-0003jx-Fq for qemu-devel@nongnu.org; Mon, 19 Aug 2013 03:30:50 -0400 Received: from mail-pd0-f179.google.com ([209.85.192.179]:59993) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VBJvH-0003jb-8k for qemu-devel@nongnu.org; Mon, 19 Aug 2013 03:30:43 -0400 Received: by mail-pd0-f179.google.com with SMTP id v10so4803121pde.10 for ; Mon, 19 Aug 2013 00:30:41 -0700 (PDT) Message-ID: <5211C99B.4070904@ozlabs.ru> Date: Mon, 19 Aug 2013 17:30:35 +1000 From: Alexey Kardashevskiy MIME-Version: 1.0 References: <1375862885-12132-1-git-send-email-aik@ozlabs.ru> <520DF5B5.4050805@ozlabs.ru> <5210E6C2.1070208@redhat.com> In-Reply-To: <5210E6C2.1070208@redhat.com> Content-Type: text/plain; charset=KOI8-R Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH] powerpc iommu: enable multiple TCE requests List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Paolo Bonzini Cc: Alexander Graf , qemu-devel@nongnu.org, qemu-ppc@nongnu.org, Paul Mackerras , David Gibson On 08/19/2013 01:22 AM, Paolo Bonzini wrote: > Il 16/08/2013 11:49, Alexey Kardashevskiy ha scritto: >> With KVM, we could fall back to the qemu implementation >>> + * when KVM doesn't support them, but that would be much slower >>> + * than just using the KVM implementations of the single TCE >>> + * hypercalls. */ >>> + if (kvmppc_spapr_use_multitce()) { >>> + _FDT((fdt_property(fdt, "ibm,hypertas-functions", hypertas_propm, >>> + sizeof(hypertas_propm)))); >>> + } else { >>> + _FDT((fdt_property(fdt, "ibm,hypertas-functions", hypertas_prop, >>> + sizeof(hypertas_prop)))); >>> + } > > This prevents migration from newer kernel to older kernel. Can you > ensure that the fallback to the QEMU implementation works, even though > it is not used in practice? How would it break? By having a device tree with "multi-tce" in it and not having KVM PPC capability for that? If this is the case, it will not prevent from migration as the "multi-tce" feature is supported anyway by this patch. The only reason for not advertising it to the guest is that the host kernel already has acceleration for H_PUT_TCE (single page map/unmap) and advertising "multi-tce" without having it in the host kernel (but only in QEMU) would slow things down (but it still will work). -- Alexey