| Message ID | 20201019101523.4145-1-chris@chris-wilson.co.uk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | drm/i915: Force VT'd workarounds when running as a guest OS | expand |
Works for me and makes the fault messages disappear. Thanks. Tested-by: Stefan Fritsch <sf@sfritsch.de> On Mon, 19 Oct 2020, Chris Wilson wrote: > If i915.ko is being used as a passthrough device, it does not know if > the host is using intel_iommu. Mixing the iommu and gfx causing a few > issues (such as scanout overfetch) which we need to workaround inside > the driver, so if we detect we are running under a hypervisor, also > assume the device access is being virtualised. > > Reported-by: Stefan Fritsch <sf@sfritsch.de> > Suggested-by: Stefan Fritsch <sf@sfritsch.de> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Zhenyu Wang <zhenyuw@linux.intel.com> > Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> > Cc: Stefan Fritsch <sf@sfritsch.de> > Cc: stable@vger.kernel.org > --- > drivers/gpu/drm/i915/i915_drv.h | 6 +++++- > 1 file changed, 5 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index 1a5729932c81..bcd8650603d8 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -33,6 +33,8 @@ > #include <uapi/drm/i915_drm.h> > #include <uapi/drm/drm_fourcc.h> > > +#include <asm/hypervisor.h> > + > #include <linux/io-mapping.h> > #include <linux/i2c.h> > #include <linux/i2c-algo-bit.h> > @@ -1760,7 +1762,9 @@ static inline bool intel_vtd_active(void) > if (intel_iommu_gfx_mapped) > return true; > #endif > - return false; > + > + /* Running as a guest, we assume the host is enforcing VT'd */ > + return !hypervisor_is_type(X86_HYPER_NATIVE); > } > > static inline bool intel_scanout_needs_vtd_wa(struct drm_i915_private *dev_priv) >
On 2020.10.19 11:15:23 +0100, Chris Wilson wrote: > If i915.ko is being used as a passthrough device, it does not know if > the host is using intel_iommu. Mixing the iommu and gfx causing a few > issues (such as scanout overfetch) which we need to workaround inside > the driver, so if we detect we are running under a hypervisor, also > assume the device access is being virtualised. > > Reported-by: Stefan Fritsch <sf@sfritsch.de> > Suggested-by: Stefan Fritsch <sf@sfritsch.de> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Zhenyu Wang <zhenyuw@linux.intel.com> > Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> > Cc: Stefan Fritsch <sf@sfritsch.de> > Cc: stable@vger.kernel.org > --- > drivers/gpu/drm/i915/i915_drv.h | 6 +++++- > 1 file changed, 5 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index 1a5729932c81..bcd8650603d8 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -33,6 +33,8 @@ > #include <uapi/drm/i915_drm.h> > #include <uapi/drm/drm_fourcc.h> > > +#include <asm/hypervisor.h> > + > #include <linux/io-mapping.h> > #include <linux/i2c.h> > #include <linux/i2c-algo-bit.h> > @@ -1760,7 +1762,9 @@ static inline bool intel_vtd_active(void) > if (intel_iommu_gfx_mapped) > return true; > #endif > - return false; > + > + /* Running as a guest, we assume the host is enforcing VT'd */ > + return !hypervisor_is_type(X86_HYPER_NATIVE); > } > > static inline bool intel_scanout_needs_vtd_wa(struct drm_i915_private *dev_priv) > -- Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Quoting Stefan Fritsch (2020-10-19 14:08:17) > Works for me and makes the fault messages disappear. Thanks. > > Tested-by: Stefan Fritsch <sf@sfritsch.de> Thanks for the report and testing. In hindsight, this should have been obvious. Pushed, -Chris
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 1a5729932c81..bcd8650603d8 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -33,6 +33,8 @@ #include <uapi/drm/i915_drm.h> #include <uapi/drm/drm_fourcc.h> +#include <asm/hypervisor.h> + #include <linux/io-mapping.h> #include <linux/i2c.h> #include <linux/i2c-algo-bit.h> @@ -1760,7 +1762,9 @@ static inline bool intel_vtd_active(void) if (intel_iommu_gfx_mapped) return true; #endif - return false; + + /* Running as a guest, we assume the host is enforcing VT'd */ + return !hypervisor_is_type(X86_HYPER_NATIVE); } static inline bool intel_scanout_needs_vtd_wa(struct drm_i915_private *dev_priv)
If i915.ko is being used as a passthrough device, it does not know if the host is using intel_iommu. Mixing the iommu and gfx causing a few issues (such as scanout overfetch) which we need to workaround inside the driver, so if we detect we are running under a hypervisor, also assume the device access is being virtualised. Reported-by: Stefan Fritsch <sf@sfritsch.de> Suggested-by: Stefan Fritsch <sf@sfritsch.de> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Zhenyu Wang <zhenyuw@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Stefan Fritsch <sf@sfritsch.de> Cc: stable@vger.kernel.org --- drivers/gpu/drm/i915/i915_drv.h | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)