Alexia Benington
2009-Feb-02 21:16 UTC
[Xen-devel] Option iommu=1 causes hypervisor to start debugger
Hi, In changeset 19130:c8962b24fb50, when using iommu=1, it goes into the hypervisor debugger. Why is this so? Thanks. -Alex _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser
2009-Feb-02 21:41 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
Perhaps the VT-d hardware/BIOS on your system causes Xen to crash and hence drop itself into the debugger environment? -- Keir On 02/02/2009 21:16, "Alexia Benington" <alexbenington@gmail.com> wrote:> Hi, > > In changeset 19130:c8962b24fb50, when using iommu=1, it goes into the > hypervisor debugger. Why is this so? Thanks. > > -Alex > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Alexia Benington
2009-Feb-02 21:48 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
At this moment, I''m not really sure. I''m using Intel DQ45CB with Bios CBQ4510H.86A.0073.2009.0122.1353. Is anybody else using the same board and having trouble with the I/O passthrough? I''ve attached the connected console''s output, can''t really pick up anything abnormal though, except that it just stops as you see it. -Alex On Mon, Feb 2, 2009 at 4:41 PM, Keir Fraser <keir.fraser@eu.citrix.com> wrote:> Perhaps the VT-d hardware/BIOS on your system causes Xen to crash and hence > drop itself into the debugger environment? > > -- Keir_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Neo Jia
2009-Feb-02 21:53 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
I didn''t see any error in that log file. How can you tell it is trapped into debugger? Thanks, Neo On Mon, Feb 2, 2009 at 1:48 PM, Alexia Benington <alexbenington@gmail.com> wrote:> At this moment, I''m not really sure. I''m using Intel DQ45CB with Bios > CBQ4510H.86A.0073.2009.0122.1353. Is anybody else using the same board > and having trouble with the I/O passthrough? > > I''ve attached the connected console''s output, can''t really pick up > anything abnormal though, except that it just stops as you see it. > > -Alex > > On Mon, Feb 2, 2009 at 4:41 PM, Keir Fraser <keir.fraser@eu.citrix.com> wrote: >> Perhaps the VT-d hardware/BIOS on your system causes Xen to crash and hence >> drop itself into the debugger environment? >> >> -- Keir > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel > >-- I would remember that if researchers were not ambitious probably today we haven''t the technology we are using! _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Alexia Benington
2009-Feb-02 22:00 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
Apologies, I should have clarified that it didn''t automatically step into the debugger, but hang at the last line of the output I attached earlier. Pressing ''h'' would show the list of installed handlers. Thanks. -Alex On Mon, Feb 2, 2009 at 4:53 PM, Neo Jia <neojia@gmail.com> wrote:> I didn''t see any error in that log file. How can you tell it is > trapped into debugger? > > Thanks, > Neo > > On Mon, Feb 2, 2009 at 1:48 PM, Alexia Benington > <alexbenington@gmail.com> wrote: >> At this moment, I''m not really sure. I''m using Intel DQ45CB with Bios >> CBQ4510H.86A.0073.2009.0122.1353. Is anybody else using the same board >> and having trouble with the I/O passthrough? >> >> I''ve attached the connected console''s output, can''t really pick up >> anything abnormal though, except that it just stops as you see it. >> >> -Alex >> >> On Mon, Feb 2, 2009 at 4:41 PM, Keir Fraser <keir.fraser@eu.citrix.com> wrote: >>> Perhaps the VT-d hardware/BIOS on your system causes Xen to crash and hence >>> drop itself into the debugger environment? >>> >>> -- Keir >> >> _______________________________________________ >> Xen-devel mailing list >> Xen-devel@lists.xensource.com >> http://lists.xensource.com/xen-devel >> >> > > > > -- > I would remember that if researchers were not ambitious > probably today we haven''t the technology we are using! >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Alexia Benington
2009-Feb-02 23:42 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
If I do not include the iommu option, I/O virtualisation is disabled. I never thought that iommu is a necessary prerequisite for I/O virt, or am I mistaken? -Alex _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser
2009-Feb-03 08:06 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
It''s needed for passing through physical devices to HVM guests. -- Keir On 02/02/2009 23:42, "Alexia Benington" <alexbenington@gmail.com> wrote:> If I do not include the iommu option, I/O virtualisation is disabled. > I never thought that iommu is a necessary prerequisite for I/O virt, > or am I mistaken? > > -Alex > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Alexia Benington
2009-Feb-03 14:25 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
Thanks Keir. Perhaps I should try with Yu Zhao''s patch and see if that fixes the problem. -Alex On Tue, Feb 3, 2009 at 3:06 AM, Keir Fraser <keir.fraser@eu.citrix.com> wrote:> It''s needed for passing through physical devices to HVM guests. > > -- Keir > > On 02/02/2009 23:42, "Alexia Benington" <alexbenington@gmail.com> wrote: > >> If I do not include the iommu option, I/O virtualisation is disabled. >> I never thought that iommu is a necessary prerequisite for I/O virt, >> or am I mistaken? >> >> -Alex >> >> _______________________________________________ >> Xen-devel mailing list >> Xen-devel@lists.xensource.com >> http://lists.xensource.com/xen-devel > > >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Zhao, Yu
2009-Feb-04 01:55 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
Which patch are you talking about? Since we can''t get any clue from your log, can you please send us the output of `lspci -vvv''? Thanks, Yu Alexia Benington wrote:> Thanks Keir. Perhaps I should try with Yu Zhao''s patch and see if that > fixes the problem. > > -Alex > > On Tue, Feb 3, 2009 at 3:06 AM, Keir Fraser <keir.fraser@eu.citrix.com> wrote: >> It''s needed for passing through physical devices to HVM guests. >> >> -- Keir >> >> On 02/02/2009 23:42, "Alexia Benington" <alexbenington@gmail.com> wrote: >> >>> If I do not include the iommu option, I/O virtualisation is disabled. >>> I never thought that iommu is a necessary prerequisite for I/O virt, >>> or am I mistaken? >>> >>> -Alex >>> >>> _______________________________________________ >>> Xen-devel mailing list >>> Xen-devel@lists.xensource.com >>> http://lists.xensource.com/xen-devel >> >> > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Alexia Benington
2009-Feb-04 02:27 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
I was referring to the fix titled "ioemu: fix PCI ROM handling problem in passthrough". Seriously, I don''t even know if the problem is related. Just trying something because I''m totally clueless now. I''ve attached the lspci. Let me know if you need more info. Will be glad to provide. Thanks guys. - Alex On Tue, Feb 3, 2009 at 8:55 PM, Zhao, Yu <yu.zhao@intel.com> wrote:> Which patch are you talking about? Since we can''t get any clue from your > log, can you please send us the output of `lspci -vvv''? > > Thanks, > Yu > > Alexia Benington wrote: >> >> Thanks Keir. Perhaps I should try with Yu Zhao''s patch and see if that >> fixes the problem. >> >> -Alex >> >> On Tue, Feb 3, 2009 at 3:06 AM, Keir Fraser <keir.fraser@eu.citrix.com> >> wrote: >>> >>> It''s needed for passing through physical devices to HVM guests. >>> >>> -- Keir >>> >>> On 02/02/2009 23:42, "Alexia Benington" <alexbenington@gmail.com> wrote: >>> >>>> If I do not include the iommu option, I/O virtualisation is disabled. >>>> I never thought that iommu is a necessary prerequisite for I/O virt, >>>> or am I mistaken? >>>> >>>> -Alex >>>> >>>> _______________________________________________ >>>> Xen-devel mailing list >>>> Xen-devel@lists.xensource.com >>>> http://lists.xensource.com/xen-devel >>> >>> >> >> _______________________________________________ >> Xen-devel mailing list >> Xen-devel@lists.xensource.com >> http://lists.xensource.com/xen-devel > >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Zhao, Yu
2009-Feb-04 02:52 UTC
RE: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
Can you please apply the attached patch and grab the log again? Thanks, Yu>-----Original Message----- >From: Alexia Benington [mailto:alexbenington@gmail.com] >Sent: Wednesday, February 04, 2009 10:27 AM >To: Zhao, Yu >Cc: Keir Fraser; xen-devel@lists.xensource.com >Subject: Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger > >I was referring to the fix titled "ioemu: fix PCI ROM handling problem >in passthrough". Seriously, I don''t even know if the problem is >related. Just trying something because I''m totally clueless now. > >I''ve attached the lspci. Let me know if you need more info. Will be >glad to provide. > >Thanks guys. > >- Alex > >On Tue, Feb 3, 2009 at 8:55 PM, Zhao, Yu <yu.zhao@intel.com> wrote: >> Which patch are you talking about? Since we can''t get any clue from your >> log, can you please send us the output of `lspci -vvv''? >> >> Thanks, >> Yu >> >> Alexia Benington wrote: >>> >>> Thanks Keir. Perhaps I should try with Yu Zhao''s patch and see if that >>> fixes the problem. >>> >>> -Alex >>> >>> On Tue, Feb 3, 2009 at 3:06 AM, Keir Fraser <keir.fraser@eu.citrix.com> >>> wrote: >>>> >>>> It''s needed for passing through physical devices to HVM guests. >>>> >>>> -- Keir >>>> >>>> On 02/02/2009 23:42, "Alexia Benington" <alexbenington@gmail.com> wrote: >>>> >>>>> If I do not include the iommu option, I/O virtualisation is disabled. >>>>> I never thought that iommu is a necessary prerequisite for I/O virt, >>>>> or am I mistaken? >>>>> >>>>> -Alex >>>>> >>>>> _______________________________________________ >>>>> Xen-devel mailing list >>>>> Xen-devel@lists.xensource.com >>>>> http://lists.xensource.com/xen-devel >>>> >>>> >>> >>> _______________________________________________ >>> Xen-devel mailing list >>> Xen-devel@lists.xensource.com >>> http://lists.xensource.com/xen-devel >> >>_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Alexia Benington
2009-Feb-04 20:30 UTC
Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger
Here it is. Thanks. -Alex On Tue, Feb 3, 2009 at 9:52 PM, Zhao, Yu <yu.zhao@intel.com> wrote:> Can you please apply the attached patch and grab the log again? > > Thanks, > Yu > >>-----Original Message----- >>From: Alexia Benington [mailto:alexbenington@gmail.com] >>Sent: Wednesday, February 04, 2009 10:27 AM >>To: Zhao, Yu >>Cc: Keir Fraser; xen-devel@lists.xensource.com >>Subject: Re: [Xen-devel] Option iommu=1 causes hypervisor to start debugger >> >>I was referring to the fix titled "ioemu: fix PCI ROM handling problem >>in passthrough". Seriously, I don''t even know if the problem is >>related. Just trying something because I''m totally clueless now. >> >>I''ve attached the lspci. Let me know if you need more info. Will be >>glad to provide. >> >>Thanks guys. >> >>- Alex >> >>On Tue, Feb 3, 2009 at 8:55 PM, Zhao, Yu <yu.zhao@intel.com> wrote: >>> Which patch are you talking about? Since we can''t get any clue from your >>> log, can you please send us the output of `lspci -vvv''? >>> >>> Thanks, >>> Yu >>> >>> Alexia Benington wrote: >>>> >>>> Thanks Keir. Perhaps I should try with Yu Zhao''s patch and see if that >>>> fixes the problem. >>>> >>>> -Alex >>>> >>>> On Tue, Feb 3, 2009 at 3:06 AM, Keir Fraser <keir.fraser@eu.citrix.com> >>>> wrote: >>>>> >>>>> It''s needed for passing through physical devices to HVM guests. >>>>> >>>>> -- Keir >>>>> >>>>> On 02/02/2009 23:42, "Alexia Benington" <alexbenington@gmail.com> wrote: >>>>> >>>>>> If I do not include the iommu option, I/O virtualisation is disabled. >>>>>> I never thought that iommu is a necessary prerequisite for I/O virt, >>>>>> or am I mistaken? >>>>>> >>>>>> -Alex >>>>>> >>>>>> _______________________________________________ >>>>>> Xen-devel mailing list >>>>>> Xen-devel@lists.xensource.com >>>>>> http://lists.xensource.com/xen-devel >>>>> >>>>> >>>> >>>> _______________________________________________ >>>> Xen-devel mailing list >>>> Xen-devel@lists.xensource.com >>>> http://lists.xensource.com/xen-devel >>> >>> >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel