From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Vrabel Subject: Re: [PATCH 2/3] xen: eliminate scalability issues from initrd handling Date: Thu, 4 Sep 2014 13:52:15 +0100 Message-ID: <5408607F.6020108@citrix.com> References: <1409834326-29287-1-git-send-email-jgross@suse.com> <1409834326-29287-3-git-send-email-jgross@suse.com> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1409834326-29287-3-git-send-email-jgross@suse.com> Sender: linux-kernel-owner@vger.kernel.org To: Juergen Gross , linux-kernel@vger.kernel.org, xen-devel@lists.xensource.com, konrad.wilk@oracle.com, boris.ostrovsky@oracle.com, jbeulich@suse.com List-Id: xen-devel@lists.xenproject.org On 04/09/14 13:38, Juergen Gross wrote: > Size restrictions native kernels wouldn't have resulted from the initrd > getting mapped into the initial mapping. The kernel doesn't really need > the initrd to be mapped, so use infrastructure available in Xen to avoid > the mapping and hence the restriction. [...] > --- a/arch/x86/xen/enlighten.c > +++ b/arch/x86/xen/enlighten.c [...] > @@ -1667,10 +1668,20 @@ asmlinkage __visible void __init xen_start_kernel(void) > new_cpu_data.x86_capability[0] = cpuid_edx(1); > #endif > > + if (xen_start_info->mod_start) > + initrd_start = __pa(xen_start_info->mod_start); > +#ifdef CONFIG_BLK_DEV_INITRD > +#ifdef CONFIG_X86_32 > + BUG_ON(xen_start_info->flags & SIF_MOD_START_PFN); > +#else > + if (xen_start_info->flags & SIF_MOD_START_PFN) > + initrd_start = PFN_PHYS(xen_start_info->mod_start); > +#endif > +#endif Remove these unnecessary #ifdefs and the BUG_ON(). We can trust Xen to not set SIF_MOD_START_PFN if we haven't asked for it. > --- a/arch/x86/xen/xen-head.S > +++ b/arch/x86/xen/xen-head.S > @@ -124,6 +124,9 @@ NEXT_HYPERCALL(arch_6) > ELFNOTE(Xen, XEN_ELFNOTE_L1_MFN_VALID, > .quad _PAGE_PRESENT; .quad _PAGE_PRESENT) > ELFNOTE(Xen, XEN_ELFNOTE_SUSPEND_CANCEL, .long 1) > +#ifdef CONFIG_X86_64 > + ELFNOTE(Xen, XEN_ELFNOTE_MOD_START_PFN, .long 1) > +#endif Why X86_64 only? If there's a good reason the commit message needs to explain why. David