Han, Weidong
2008-Apr-22 05:56 UTC
[Xen-devel] [PATCH] Fix dom0 boot issue with VT-d enabled
c/s 17487 put iommu_setup() in __initcall(), this makes iommu_setup() be called before dom0 creation. Because iommu_setup() will do something for dom0 such as map 1:1 page table, while dom0 is NULL, reboot happens during dom0 booting. This patch moves iommu_setup() behind dom0 creation. Signed-off-by: Weidong Han <weidong.han@intel.com> _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser
2008-Apr-22 08:56 UTC
Re: [Xen-devel] [PATCH] Fix dom0 boot issue with VT-d enabled
On 22/4/08 06:56, "Han, Weidong" <weidong.han@intel.com> wrote:> c/s 17487 put iommu_setup() in __initcall(), this makes iommu_setup() be > called before dom0 creation. Because iommu_setup() will do something for > dom0 such as map 1:1 page table, while dom0 is NULL, reboot happens > during dom0 booting. This patch moves iommu_setup() behind dom0 > creation.The code for system and domain initialisation of iommu state is a disgraceful mess. The former should be done early, before domains (including dom0) are created. If you have stuff to do for dom0, do it when dom0 is created. You get a call into your code at that point. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Han, Weidong
2008-Apr-22 09:00 UTC
RE: [Xen-devel] [PATCH] Fix dom0 boot issue with VT-d enabled
Okay. I will resend the patch later. Randy (Weidong) Keir Fraser wrote:> On 22/4/08 06:56, "Han, Weidong" <weidong.han@intel.com> wrote: > >> c/s 17487 put iommu_setup() in __initcall(), this makes >> iommu_setup() be called before dom0 creation. Because iommu_setup() >> will do something for dom0 such as map 1:1 page table, while dom0 is >> NULL, reboot happens during dom0 booting. This patch moves >> iommu_setup() behind dom0 creation. > > The code for system and domain initialisation of iommu state is a > disgraceful mess. The former should be done early, before domains > (including dom0) are created. If you have stuff to do for dom0, do it > when dom0 is created. You get a call into your code at that point. > > -- Keir_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Han, Weidong
2008-Apr-22 14:48 UTC
RE: [Xen-devel] [PATCH] Fix dom0 boot issue with VT-d enabled
Resend the patch. This patch split iommu setup and dom0 setup. Randy (Weidong) Keir Fraser wrote:> On 22/4/08 06:56, "Han, Weidong" <weidong.han@intel.com> wrote: > >> c/s 17487 put iommu_setup() in __initcall(), this makes >> iommu_setup() be called before dom0 creation. Because iommu_setup() >> will do something for dom0 such as map 1:1 page table, while dom0 is >> NULL, reboot happens during dom0 booting. This patch moves >> iommu_setup() behind dom0 creation. > > The code for system and domain initialisation of iommu state is a > disgraceful mess. The former should be done early, before domains > (including dom0) are created. If you have stuff to do for dom0, do it > when dom0 is created. You get a call into your code at that point. > > -- Keir_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel