From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:51388) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VFTgS-00019K-Bk for qemu-devel@nongnu.org; Fri, 30 Aug 2013 14:44:44 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VFTgJ-0001XT-Va for qemu-devel@nongnu.org; Fri, 30 Aug 2013 14:44:36 -0400 Received: from mail-ea0-x233.google.com ([2a00:1450:4013:c01::233]:56194) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VFTgJ-0001XE-Nu for qemu-devel@nongnu.org; Fri, 30 Aug 2013 14:44:27 -0400 Received: by mail-ea0-f179.google.com with SMTP id b10so1070557eae.24 for ; Fri, 30 Aug 2013 11:44:26 -0700 (PDT) Sender: Paolo Bonzini Message-ID: <5220E803.8010506@redhat.com> Date: Fri, 30 Aug 2013 20:44:19 +0200 From: Paolo Bonzini MIME-Version: 1.0 References: <20130828103552.GC2038@GaryOffice.site> <521DE2EB.9050404@redhat.com> <521DE3C0.4030603@suse.de> <521DE8C7.1010903@redhat.com> <521DF33E.4090205@suse.de> <20130829082317.GA1852@GaryOffice.site> <521F1CB802000048000E30E0@suse.com> <20130830032828.GA1713@GaryOffice.site> <522067C7.4090606@redhat.com> <52208903.4000706@redhat.com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [edk2] OVMF hung on qemu 1.6.0 with KVM List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: edk2-devel@lists.sourceforge.net Cc: Alexander Graf , Gleb Natapov , "qemu-devel@nongnu.org" , Bruce Rogers , Ludwig Nussel , Jordan Justen , afaerber@suse.de Il 30/08/2013 19:39, Jordan Justen ha scritto: >> However, if you guys can figure out a patch for >> > UefiCpuPkg/ResetVector/Vtf0/Ia32/Flat32ToFlat64.asm so that it builds >> > the page tables in RAM (and removing the page table Python code), that >> > would also work. > That path is not quite how it needs to be fixed, but I'll look into it. Yes, KVM needs to be fixed as well. > I had done some work to entirely eliminate the ROM based tables, but > hit some complications. I'll work on it some more... (It has another > large benefit of saving nearly 28k of ROM space.) I also tried to prototype the change, but I have no idea which parts of low memory are available and/or how to reserve 24k there. Paolo