Ian Campbell
2012-Aug-17 11:46 UTC
[Pkg-xen-devel] Bug#684661: Xen BUG at pci_amd_iommu.c:33
Hi Wei, A Debian user has hit this message and reported it in http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=684661 I think it is the BUG_ON in: struct amd_iommu *find_iommu_for_device(int bdf) { BUG_ON ( bdf >= ivrs_bdf_entries ); return ivrs_mappings[bdf].iommu; } It looks like ivrs_bdf_entries comes from ACPI. Unfortunately the bug report is a bit vague about things like stack traces etc so it's hard to say where BDF came from. Any ideas? I'll also follow up to the submitter to try and get some more details. Ian.
Wei Wang
2012-Aug-17 12:10 UTC
[Pkg-xen-devel] Bug#684661: [Xen-devel] Xen BUG at pci_amd_iommu.c:33
On 08/17/2012 01:46 PM, Ian Campbell wrote:> Hi Wei, > > A Debian user has hit this message and reported it in > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=684661 > > I think it is the BUG_ON in: > struct amd_iommu *find_iommu_for_device(int bdf) > { > BUG_ON ( bdf>= ivrs_bdf_entries ); > return ivrs_mappings[bdf].iommu; > } > > It looks like ivrs_bdf_entries comes from ACPI. Unfortunately the bug > report is a bit vague about things like stack traces etc so it's hard to > say where BDF came from. > > Any ideas? I'll also follow up to the submitter to try and get some more > details.It difficult to identify the issue from the description and I think a full serial boot log would be great. Also using iommu=debug as boot option is very helpful to get more information. Thanks, Wei> Ian. > > > _______________________________________________ > Xen-devel mailing list > Xen-devel at lists.xen.org > http://lists.xen.org/xen-devel >
On Fri, 2012-08-17 at 13:10 +0100, Wei Wang wrote:> On 08/17/2012 01:46 PM, Ian Campbell wrote: > > Hi Wei, > > > > A Debian user has hit this message and reported it in > > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=684661 > > > > I think it is the BUG_ON in: > > struct amd_iommu *find_iommu_for_device(int bdf) > > { > > BUG_ON ( bdf>= ivrs_bdf_entries ); > > return ivrs_mappings[bdf].iommu; > > } > > > > It looks like ivrs_bdf_entries comes from ACPI. Unfortunately the bug > > report is a bit vague about things like stack traces etc so it''s hard to > > say where BDF came from. > > > > Any ideas? I''ll also follow up to the submitter to try and get some more > > details. > > It difficult to identify the issue from the description and I think a > full serial boot log would be great. Also using iommu=debug as boot > option is very helpful to get more information.Thanks, I''ve added that to my request to the submitter. Ian.