From mboxrd@z Thu Jan 1 00:00:00 1970 From: Gerd Hoffmann Subject: Re: new domain builder breaks compatibility Date: Thu, 01 Feb 2007 17:33:20 +0100 Message-ID: <45C21650.5020402@suse.de> References: <20070131214843.GB15106@totally.trollied.org.uk> <45C1AD2B.7070002@suse.de> <20070201145517.GA26562@totally.trollied.org.uk> <45C20404.6040109@suse.de> <20070201162259.GA31065@totally.trollied.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20070201162259.GA31065@totally.trollied.org.uk> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: John Levon Cc: xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org John Levon wrote: > On Thu, Feb 01, 2007 at 04:15:16PM +0100, Gerd Hoffmann wrote: > Hmm... am I misunderstanding this code: > 822 if ( (v_end - vstack_end) < (512UL << 10) ) > 823 { > 824 /* Add extra 4MB to get >= 512kB padding. */ > 825 if ( !increment_ulong(&v_end, 1UL << 22) ) > 826 goto error_out; > 827 } No, you are right, it's 4MB unconditionally. Seems to have changed though since I checked last time. > Surely that guarantees we get at least 512Kb and it ends on a 4Mb > boundary? Yep. Is the 2MB alignment ok for Solaris? You can depend on the 512kb only anyway, so the alignment shouldn't matter much ... cheers, Gerd -- Gerd Hoffmann