msitranslate is known to cause problems with some device drivers, because it sets the real device in MSI mode while making the guest think is actually in legacy interrupts mode. Some drivers are able to spot this inconsistency and break (Nvidia drivers for example). Disable msitranslate by default. Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> diff --git a/tools/libxl/xl_cmdimpl.c b/tools/libxl/xl_cmdimpl.c index afa0af6..c80b9fb 100644 --- a/tools/libxl/xl_cmdimpl.c +++ b/tools/libxl/xl_cmdimpl.c @@ -562,7 +562,7 @@ static void parse_config_data(const char *config_source, XLU_Config *config; XLU_ConfigList *cpus, *vbds, *nics, *pcis, *cvfbs, *cpuids; int pci_power_mgmt = 0; - int pci_msitranslate = 1; + int pci_msitranslate = 0; int pci_permissive = 0; int e;
On Wed, 2012-06-27 at 14:56 +0100, Stefano Stabellini wrote:> msitranslate is known to cause problems with some device drivers, > because it sets the real device in MSI mode while making the guest think > is actually in legacy interrupts mode. Some drivers are able to spot this > inconsistency and break (Nvidia drivers for example). > > Disable msitranslate by default. > > Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>Acked-by: Ian Campbell <ian.campbell@citrix.com> I saw a qemu (trad) patch and I see I''ve got an xl patch further down my inbox. There is no sequencing requirement here, is there? I wonder if this field shouyld be a defbool instead of just a bool? I don''t recall why I didn''t transition it when I made the defbool stuff.> > diff --git a/tools/libxl/xl_cmdimpl.c b/tools/libxl/xl_cmdimpl.c > index afa0af6..c80b9fb 100644 > --- a/tools/libxl/xl_cmdimpl.c > +++ b/tools/libxl/xl_cmdimpl.c > @@ -562,7 +562,7 @@ static void parse_config_data(const char *config_source, > XLU_Config *config; > XLU_ConfigList *cpus, *vbds, *nics, *pcis, *cvfbs, *cpuids; > int pci_power_mgmt = 0; > - int pci_msitranslate = 1; > + int pci_msitranslate = 0; > int pci_permissive = 0; > int e;
On Wed, 27 Jun 2012, Ian Campbell wrote:> On Wed, 2012-06-27 at 14:56 +0100, Stefano Stabellini wrote: > > msitranslate is known to cause problems with some device drivers, > > because it sets the real device in MSI mode while making the guest think > > is actually in legacy interrupts mode. Some drivers are able to spot this > > inconsistency and break (Nvidia drivers for example). > > > > Disable msitranslate by default. > > > > Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> > > Acked-by: Ian Campbell <ian.campbell@citrix.com> > > I saw a qemu (trad) patch and I see I''ve got an xl patch further down my > inbox. There is no sequencing requirement here, is there?nope> I wonder if this field shouyld be a defbool instead of just a bool? I > don''t recall why I didn''t transition it when I made the defbool stuff.I think it could be a defbool> > > > diff --git a/tools/libxl/xl_cmdimpl.c b/tools/libxl/xl_cmdimpl.c > > index afa0af6..c80b9fb 100644 > > --- a/tools/libxl/xl_cmdimpl.c > > +++ b/tools/libxl/xl_cmdimpl.c > > @@ -562,7 +562,7 @@ static void parse_config_data(const char *config_source, > > XLU_Config *config; > > XLU_ConfigList *cpus, *vbds, *nics, *pcis, *cvfbs, *cpuids; > > int pci_power_mgmt = 0; > > - int pci_msitranslate = 1; > > + int pci_msitranslate = 0; > > int pci_permissive = 0; > > int e; > > >
Stefano Stabellini writes ("Re: [Xen-devel] libxl: disable msitranslate by default"):> On Wed, 27 Jun 2012, Ian Campbell wrote: > > I saw a qemu (trad) patch and I see I''ve got an xl patch further down my > > inbox. There is no sequencing requirement here, is there? > > nopeThanks.> > I wonder if this field shouyld be a defbool instead of just a bool? I > > don''t recall why I didn''t transition it when I made the defbool stuff. > > I think it could be a defboolI''ll apply the patch as-is but perhaps a followup to change it to a defbool would be good ? Ian.
Ian Campbell writes ("Re: [Xen-devel] libxl: disable msitranslate by default"):> On Wed, 2012-06-27 at 14:56 +0100, Stefano Stabellini wrote: > > msitranslate is known to cause problems with some device drivers, > > because it sets the real device in MSI mode while making the guest think > > is actually in legacy interrupts mode. Some drivers are able to spot this > > inconsistency and break (Nvidia drivers for example). > > > > Disable msitranslate by default. > > > > Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> > > Acked-by: Ian Campbell <ian.campbell@citrix.com>Committed-by: Ian Jackson <ian.jackson@eu.citrix.com>