[PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[Michel Dänzer]

From: Michel Dänzer <michel.daenzer@amd.com>

vram_size is supposed to be the total amount of VRAM that can be used by
userspace, which corresponds to the TTM VRAM manager size (which is
normally the full amount of VRAM, but can be just the visible VRAM when
DMA can't be used for BO migration for some reason).

The above was incorrectly used for vram_visible before, resulting in
generally too large values being reported.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
---
 drivers/gpu/drm/radeon/radeon_drv.c | 3 ++-
 drivers/gpu/drm/radeon/radeon_gem.c | 4 ++--
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index a252bc407aa2..88c41d43ec3d 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -97,9 +97,10 @@
  *   2.46.0 - Add PFP_SYNC_ME support on evergreen
  *   2.47.0 - Add UVD_NO_OP register support
  *   2.48.0 - TA_CS_BC_BASE_ADDR allowed on SI
+ *   2.49.0 - DRM_RADEON_GEM_INFO ioctl returns correct vram_size/visible values
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	48
+#define KMS_DRIVER_MINOR	49
 #define KMS_DRIVER_PATCHLEVEL	0
 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
 int radeon_driver_unload_kms(struct drm_device *dev);
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index 0bcffd8a7bd3..96683f5b2b1b 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -220,8 +220,8 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data,
 
 	man = &rdev->mman.bdev.man[TTM_PL_VRAM];
 
-	args->vram_size = rdev->mc.real_vram_size;
-	args->vram_visible = (u64)man->size << PAGE_SHIFT;
+	args->vram_size = (u64)man->size << PAGE_SHIFT;
+	args->vram_visible = rdev->mc.visible_vram_size;
 	args->vram_visible -= rdev->vram_pin_size;
 	args->gart_size = rdev->mc.gtt_size;
 	args->gart_size -= rdev->gart_pin_size;
-- 
2.11.0

_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev

[PATCH] winsys/radeon: Allow visible VRAM size > 256MB with kernel driver >= 2.49

[Michel Dänzer]

From: Michel Dänzer <michel.daenzer@amd.com>

The kernel driver reports correct values now.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
---
 src/gallium/winsys/radeon/drm/radeon_drm_winsys.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
index a8da62fd36..cacd683879 100644
--- a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
+++ b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
@@ -372,7 +372,12 @@ static bool do_winsys_init(struct radeon_drm_winsys *ws)
     }
     ws->info.gart_size = gem_info.gart_size;
     ws->info.vram_size = gem_info.vram_size;
-    ws->info.vram_vis_size = MIN2(gem_info.vram_visible, 256*1024*1024);
+    ws->info.vram_vis_size = gem_info.vram_visible;
+    /* Older versions of the kernel driver reported incorrect values, and
+     * didn't support more than 256MB of visible VRAM anyway
+     */
+    if (ws->info.drm_minor < 49)
+        ws->info.vram_vis_size = MIN2(ws->info.vram_vis_size, 256*1024*1024);
 
     /* Radeon allocates all buffers as contigous, which makes large allocations
      * unlikely to succeed. */
-- 
2.11.0

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[Christian König]

Am 31.01.2017 um 07:54 schrieb Michel Dänzer:
> From: Michel Dänzer <michel.daenzer@amd.com>
>
> vram_size is supposed to be the total amount of VRAM that can be used by
> userspace, which corresponds to the TTM VRAM manager size (which is
> normally the full amount of VRAM, but can be just the visible VRAM when
> DMA can't be used for BO migration for some reason).
>
> The above was incorrectly used for vram_visible before, resulting in
> generally too large values being reported.
>
> Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>

Good catch, I wonder how long we have that in here.

Reviewed-by: Christian König <christian.koenig@amd.com> for this one as 
well the mesa patch.

Regards,
Christian.

> ---
>   drivers/gpu/drm/radeon/radeon_drv.c | 3 ++-
>   drivers/gpu/drm/radeon/radeon_gem.c | 4 ++--
>   2 files changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
> index a252bc407aa2..88c41d43ec3d 100644
> --- a/drivers/gpu/drm/radeon/radeon_drv.c
> +++ b/drivers/gpu/drm/radeon/radeon_drv.c
> @@ -97,9 +97,10 @@
>    *   2.46.0 - Add PFP_SYNC_ME support on evergreen
>    *   2.47.0 - Add UVD_NO_OP register support
>    *   2.48.0 - TA_CS_BC_BASE_ADDR allowed on SI
> + *   2.49.0 - DRM_RADEON_GEM_INFO ioctl returns correct vram_size/visible values
>    */
>   #define KMS_DRIVER_MAJOR	2
> -#define KMS_DRIVER_MINOR	48
> +#define KMS_DRIVER_MINOR	49
>   #define KMS_DRIVER_PATCHLEVEL	0
>   int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
>   int radeon_driver_unload_kms(struct drm_device *dev);
> diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
> index 0bcffd8a7bd3..96683f5b2b1b 100644
> --- a/drivers/gpu/drm/radeon/radeon_gem.c
> +++ b/drivers/gpu/drm/radeon/radeon_gem.c
> @@ -220,8 +220,8 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data,
>   
>   	man = &rdev->mman.bdev.man[TTM_PL_VRAM];
>   
> -	args->vram_size = rdev->mc.real_vram_size;
> -	args->vram_visible = (u64)man->size << PAGE_SHIFT;
> +	args->vram_size = (u64)man->size << PAGE_SHIFT;
> +	args->vram_visible = rdev->mc.visible_vram_size;
>   	args->vram_visible -= rdev->vram_pin_size;
>   	args->gart_size = rdev->mc.gtt_size;
>   	args->gart_size -= rdev->gart_pin_size;


_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH] winsys/radeon: Allow visible VRAM size > 256MB with kernel driver >= 2.49

[Samuel Pitoiset]

Thanks for fixing this Michel.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>

On 01/31/2017 07:54 AM, Michel Dänzer wrote:
> From: Michel Dänzer <michel.daenzer@amd.com>
>
> The kernel driver reports correct values now.
>
> Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
> ---
>  src/gallium/winsys/radeon/drm/radeon_drm_winsys.c | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> index a8da62fd36..cacd683879 100644
> --- a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> +++ b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> @@ -372,7 +372,12 @@ static bool do_winsys_init(struct radeon_drm_winsys *ws)
>      }
>      ws->info.gart_size = gem_info.gart_size;
>      ws->info.vram_size = gem_info.vram_size;
> -    ws->info.vram_vis_size = MIN2(gem_info.vram_visible, 256*1024*1024);
> +    ws->info.vram_vis_size = gem_info.vram_visible;
> +    /* Older versions of the kernel driver reported incorrect values, and
> +     * didn't support more than 256MB of visible VRAM anyway
> +     */
> +    if (ws->info.drm_minor < 49)
> +        ws->info.vram_vis_size = MIN2(ws->info.vram_vis_size, 256*1024*1024);
>
>      /* Radeon allocates all buffers as contigous, which makes large allocations
>       * unlikely to succeed. */
>
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[Dieter Nützel]

Hello Michel,

as this is for radeon, do you think this could/should fix
the wrong reported VRAM size with Unigine_Heaven/-Valley, too?
Maybe speed things up? ;-)

Unigine_Valley-1.0

GPU: Unknown GPU x1
System memory: 24102 MB
Video memory:  256 MB
Sync threads:  7
Async threads: 8

I'll try patching openSUSE Kernel:stable 4.9.6-2 with this
and maybe this could then go into 4.10-rc7 'cause it is a
bugfix. - Alex?

Greetings,
Dieter

Am 31.01.2017 07:54, schrieb Michel Dänzer:
> From: Michel Dänzer <michel.daenzer@amd.com>
> 
> vram_size is supposed to be the total amount of VRAM that can be used 
> by
> userspace, which corresponds to the TTM VRAM manager size (which is
> normally the full amount of VRAM, but can be just the visible VRAM when
> DMA can't be used for BO migration for some reason).
> 
> The above was incorrectly used for vram_visible before, resulting in
> generally too large values being reported.
> 
> Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
> ---
>  drivers/gpu/drm/radeon/radeon_drv.c | 3 ++-
>  drivers/gpu/drm/radeon/radeon_gem.c | 4 ++--
>  2 files changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/radeon/radeon_drv.c
> b/drivers/gpu/drm/radeon/radeon_drv.c
> index a252bc407aa2..88c41d43ec3d 100644
> --- a/drivers/gpu/drm/radeon/radeon_drv.c
> +++ b/drivers/gpu/drm/radeon/radeon_drv.c
> @@ -97,9 +97,10 @@
>   *   2.46.0 - Add PFP_SYNC_ME support on evergreen
>   *   2.47.0 - Add UVD_NO_OP register support
>   *   2.48.0 - TA_CS_BC_BASE_ADDR allowed on SI
> + *   2.49.0 - DRM_RADEON_GEM_INFO ioctl returns correct
> vram_size/visible values
>   */
>  #define KMS_DRIVER_MAJOR	2
> -#define KMS_DRIVER_MINOR	48
> +#define KMS_DRIVER_MINOR	49
>  #define KMS_DRIVER_PATCHLEVEL	0
>  int radeon_driver_load_kms(struct drm_device *dev, unsigned long 
> flags);
>  int radeon_driver_unload_kms(struct drm_device *dev);
> diff --git a/drivers/gpu/drm/radeon/radeon_gem.c
> b/drivers/gpu/drm/radeon/radeon_gem.c
> index 0bcffd8a7bd3..96683f5b2b1b 100644
> --- a/drivers/gpu/drm/radeon/radeon_gem.c
> +++ b/drivers/gpu/drm/radeon/radeon_gem.c
> @@ -220,8 +220,8 @@ int radeon_gem_info_ioctl(struct drm_device *dev,
> void *data,
> 
>  	man = &rdev->mman.bdev.man[TTM_PL_VRAM];
> 
> -	args->vram_size = rdev->mc.real_vram_size;
> -	args->vram_visible = (u64)man->size << PAGE_SHIFT;
> +	args->vram_size = (u64)man->size << PAGE_SHIFT;
> +	args->vram_visible = rdev->mc.visible_vram_size;
>  	args->vram_visible -= rdev->vram_pin_size;
>  	args->gart_size = rdev->mc.gtt_size;
>  	args->gart_size -= rdev->gart_pin_size;
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev

Re: [PATCH] winsys/radeon: Allow visible VRAM size > 256MB with kernel driver >= 2.49

[Nicolai Hähnle]

Both patches:

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>

On 31.01.2017 07:54, Michel Dänzer wrote:
> From: Michel Dänzer <michel.daenzer@amd.com>
>
> The kernel driver reports correct values now.
>
> Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
> ---
>  src/gallium/winsys/radeon/drm/radeon_drm_winsys.c | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> index a8da62fd36..cacd683879 100644
> --- a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> +++ b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> @@ -372,7 +372,12 @@ static bool do_winsys_init(struct radeon_drm_winsys *ws)
>      }
>      ws->info.gart_size = gem_info.gart_size;
>      ws->info.vram_size = gem_info.vram_size;
> -    ws->info.vram_vis_size = MIN2(gem_info.vram_visible, 256*1024*1024);
> +    ws->info.vram_vis_size = gem_info.vram_visible;
> +    /* Older versions of the kernel driver reported incorrect values, and
> +     * didn't support more than 256MB of visible VRAM anyway
> +     */
> +    if (ws->info.drm_minor < 49)
> +        ws->info.vram_vis_size = MIN2(ws->info.vram_vis_size, 256*1024*1024);
>
>      /* Radeon allocates all buffers as contigous, which makes large allocations
>       * unlikely to succeed. */
>

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[Deucher, Alexander]

> -----Original Message-----
> From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf
> Of Michel Dänzer
> Sent: Tuesday, January 31, 2017 1:54 AM
> To: amd-gfx@lists.freedesktop.org; mesa-dev@lists.freedesktop.org
> Subject: [PATCH] drm/radeon: Fix vram_size/visible values in
> DRM_RADEON_GEM_INFO ioctl
> 
> From: Michel Dänzer <michel.daenzer@amd.com>
> 
> vram_size is supposed to be the total amount of VRAM that can be used by
> userspace, which corresponds to the TTM VRAM manager size (which is
> normally the full amount of VRAM, but can be just the visible VRAM when
> DMA can't be used for BO migration for some reason).
> 
> The above was incorrectly used for vram_visible before, resulting in
> generally too large values being reported.
> 
> Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

> ---
>  drivers/gpu/drm/radeon/radeon_drv.c | 3 ++-
>  drivers/gpu/drm/radeon/radeon_gem.c | 4 ++--
>  2 files changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/radeon/radeon_drv.c
> b/drivers/gpu/drm/radeon/radeon_drv.c
> index a252bc407aa2..88c41d43ec3d 100644
> --- a/drivers/gpu/drm/radeon/radeon_drv.c
> +++ b/drivers/gpu/drm/radeon/radeon_drv.c
> @@ -97,9 +97,10 @@
>   *   2.46.0 - Add PFP_SYNC_ME support on evergreen
>   *   2.47.0 - Add UVD_NO_OP register support
>   *   2.48.0 - TA_CS_BC_BASE_ADDR allowed on SI
> + *   2.49.0 - DRM_RADEON_GEM_INFO ioctl returns correct
> vram_size/visible values
>   */
>  #define KMS_DRIVER_MAJOR	2
> -#define KMS_DRIVER_MINOR	48
> +#define KMS_DRIVER_MINOR	49
>  #define KMS_DRIVER_PATCHLEVEL	0
>  int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
>  int radeon_driver_unload_kms(struct drm_device *dev);
> diff --git a/drivers/gpu/drm/radeon/radeon_gem.c
> b/drivers/gpu/drm/radeon/radeon_gem.c
> index 0bcffd8a7bd3..96683f5b2b1b 100644
> --- a/drivers/gpu/drm/radeon/radeon_gem.c
> +++ b/drivers/gpu/drm/radeon/radeon_gem.c
> @@ -220,8 +220,8 @@ int radeon_gem_info_ioctl(struct drm_device *dev,
> void *data,
> 
>  	man = &rdev->mman.bdev.man[TTM_PL_VRAM];
> 
> -	args->vram_size = rdev->mc.real_vram_size;
> -	args->vram_visible = (u64)man->size << PAGE_SHIFT;
> +	args->vram_size = (u64)man->size << PAGE_SHIFT;
> +	args->vram_visible = rdev->mc.visible_vram_size;
>  	args->vram_visible -= rdev->vram_pin_size;
>  	args->gart_size = rdev->mc.gtt_size;
>  	args->gart_size -= rdev->gart_pin_size;
> --
> 2.11.0
> 
> _______________________________________________
> amd-gfx mailing list
> amd-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [PATCH] winsys/radeon: Allow visible VRAM size > 256MB with kernel driver >= 2.49

[Deucher, Alexander]

> -----Original Message-----
> From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf
> Of Michel Dänzer
> Sent: Tuesday, January 31, 2017 1:54 AM
> To: amd-gfx@lists.freedesktop.org; mesa-dev@lists.freedesktop.org
> Subject: [PATCH] winsys/radeon: Allow visible VRAM size > 256MB with
> kernel driver >= 2.49
> 
> From: Michel Dänzer <michel.daenzer@amd.com>
> 
> The kernel driver reports correct values now.
> 
> Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

> ---
>  src/gallium/winsys/radeon/drm/radeon_drm_winsys.c | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> index a8da62fd36..cacd683879 100644
> --- a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> +++ b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
> @@ -372,7 +372,12 @@ static bool do_winsys_init(struct
> radeon_drm_winsys *ws)
>      }
>      ws->info.gart_size = gem_info.gart_size;
>      ws->info.vram_size = gem_info.vram_size;
> -    ws->info.vram_vis_size = MIN2(gem_info.vram_visible, 256*1024*1024);
> +    ws->info.vram_vis_size = gem_info.vram_visible;
> +    /* Older versions of the kernel driver reported incorrect values, and
> +     * didn't support more than 256MB of visible VRAM anyway
> +     */
> +    if (ws->info.drm_minor < 49)
> +        ws->info.vram_vis_size = MIN2(ws->info.vram_vis_size,
> 256*1024*1024);
> 
>      /* Radeon allocates all buffers as contigous, which makes large allocations
>       * unlikely to succeed. */
> --
> 2.11.0
> 
> _______________________________________________
> amd-gfx mailing list
> amd-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

RE: [Mesa-dev] [PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[Deucher, Alexander]

> -----Original Message-----
> From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf
> Of Dieter Nützel
> Sent: Tuesday, January 31, 2017 6:25 AM
> To: Michel Dänzer
> Cc: Alex Deucher; mesa-dev@lists.freedesktop.org; amd-
> gfx@lists.freedesktop.org
> Subject: Re: [Mesa-dev] [PATCH] drm/radeon: Fix vram_size/visible values in
> DRM_RADEON_GEM_INFO ioctl
> 
> Hello Michel,
> 
> as this is for radeon, do you think this could/should fix
> the wrong reported VRAM size with Unigine_Heaven/-Valley, too?
> Maybe speed things up? ;-)
> 
> Unigine_Valley-1.0
> 
> GPU: Unknown GPU x1
> System memory: 24102 MB
> Video memory:  256 MB
> Sync threads:  7
> Async threads: 8
> 
> I'll try patching openSUSE Kernel:stable 4.9.6-2 with this
> and maybe this could then go into 4.10-rc7 'cause it is a
> bugfix. - Alex?

This patch just fixes the case of the HUD reporting the wrong amount of visible vram.  Most 3D apps just default to 256MB if they don't know how much vram is.  The problem is GL never provided a core way to determine how much vram is available on a GPU so lots of vendor specific extensions and driver specific methods popped up to address this.

Alex

> 
> Greetings,
> Dieter
> 
> Am 31.01.2017 07:54, schrieb Michel Dänzer:
> > From: Michel Dänzer <michel.daenzer@amd.com>
> >
> > vram_size is supposed to be the total amount of VRAM that can be used
> > by
> > userspace, which corresponds to the TTM VRAM manager size (which is
> > normally the full amount of VRAM, but can be just the visible VRAM when
> > DMA can't be used for BO migration for some reason).
> >
> > The above was incorrectly used for vram_visible before, resulting in
> > generally too large values being reported.
> >
> > Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
> > ---
> >  drivers/gpu/drm/radeon/radeon_drv.c | 3 ++-
> >  drivers/gpu/drm/radeon/radeon_gem.c | 4 ++--
> >  2 files changed, 4 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/radeon/radeon_drv.c
> > b/drivers/gpu/drm/radeon/radeon_drv.c
> > index a252bc407aa2..88c41d43ec3d 100644
> > --- a/drivers/gpu/drm/radeon/radeon_drv.c
> > +++ b/drivers/gpu/drm/radeon/radeon_drv.c
> > @@ -97,9 +97,10 @@
> >   *   2.46.0 - Add PFP_SYNC_ME support on evergreen
> >   *   2.47.0 - Add UVD_NO_OP register support
> >   *   2.48.0 - TA_CS_BC_BASE_ADDR allowed on SI
> > + *   2.49.0 - DRM_RADEON_GEM_INFO ioctl returns correct
> > vram_size/visible values
> >   */
> >  #define KMS_DRIVER_MAJOR	2
> > -#define KMS_DRIVER_MINOR	48
> > +#define KMS_DRIVER_MINOR	49
> >  #define KMS_DRIVER_PATCHLEVEL	0
> >  int radeon_driver_load_kms(struct drm_device *dev, unsigned long
> > flags);
> >  int radeon_driver_unload_kms(struct drm_device *dev);
> > diff --git a/drivers/gpu/drm/radeon/radeon_gem.c
> > b/drivers/gpu/drm/radeon/radeon_gem.c
> > index 0bcffd8a7bd3..96683f5b2b1b 100644
> > --- a/drivers/gpu/drm/radeon/radeon_gem.c
> > +++ b/drivers/gpu/drm/radeon/radeon_gem.c
> > @@ -220,8 +220,8 @@ int radeon_gem_info_ioctl(struct drm_device *dev,
> > void *data,
> >
> >  	man = &rdev->mman.bdev.man[TTM_PL_VRAM];
> >
> > -	args->vram_size = rdev->mc.real_vram_size;
> > -	args->vram_visible = (u64)man->size << PAGE_SHIFT;
> > +	args->vram_size = (u64)man->size << PAGE_SHIFT;
> > +	args->vram_visible = rdev->mc.visible_vram_size;
> >  	args->vram_visible -= rdev->vram_pin_size;
> >  	args->gart_size = rdev->mc.gtt_size;
> >  	args->gart_size -= rdev->gart_pin_size;
> _______________________________________________
> amd-gfx mailing list
> amd-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [Mesa-dev] [PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[Dieter Nützel]

Am 31.01.2017 16:43, schrieb Deucher, Alexander:
>> -----Original Message-----
>> From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf
>> Of Dieter Nützel
>> Sent: Tuesday, January 31, 2017 6:25 AM
>> To: Michel Dänzer
>> Cc: Alex Deucher; mesa-dev@lists.freedesktop.org; amd-
>> gfx@lists.freedesktop.org
>> Subject: Re: [Mesa-dev] [PATCH] drm/radeon: Fix vram_size/visible 
>> values in
>> DRM_RADEON_GEM_INFO ioctl
>> 
>> Hello Michel,
>> 
>> as this is for radeon, do you think this could/should fix
>> the wrong reported VRAM size with Unigine_Heaven/-Valley, too?
>> Maybe speed things up? ;-)
>> 
>> Unigine_Valley-1.0
>> 
>> GPU: Unknown GPU x1
>> System memory: 24102 MB
>> Video memory:  256 MB
>> Sync threads:  7
>> Async threads: 8
>> 
>> I'll try patching openSUSE Kernel:stable 4.9.6-2 with this
>> and maybe this could then go into 4.10-rc7 'cause it is a
>> bugfix. - Alex?
> 
> This patch just fixes the case of the HUD reporting the wrong amount
> of visible vram.  Most 3D apps just default to 256MB if they don't
> know how much vram is.  The problem is GL never provided a core way to
> determine how much vram is available on a GPU so lots of vendor
> specific extensions and driver specific methods popped up to address
> this.
> 
> Alex

Ah, no hurry then.

Thanks for clarification.

Dieter

>> Greetings,
>> Dieter
>> 
>> Am 31.01.2017 07:54, schrieb Michel Dänzer:
>> > From: Michel Dänzer <michel.daenzer@amd.com>
>> >
>> > vram_size is supposed to be the total amount of VRAM that can be used
>> > by
>> > userspace, which corresponds to the TTM VRAM manager size (which is
>> > normally the full amount of VRAM, but can be just the visible VRAM when
>> > DMA can't be used for BO migration for some reason).
>> >
>> > The above was incorrectly used for vram_visible before, resulting in
>> > generally too large values being reported.
>> >
>> > Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
>> > ---
>> >  drivers/gpu/drm/radeon/radeon_drv.c | 3 ++-
>> >  drivers/gpu/drm/radeon/radeon_gem.c | 4 ++--
>> >  2 files changed, 4 insertions(+), 3 deletions(-)
>> >
>> > diff --git a/drivers/gpu/drm/radeon/radeon_drv.c
>> > b/drivers/gpu/drm/radeon/radeon_drv.c
>> > index a252bc407aa2..88c41d43ec3d 100644
>> > --- a/drivers/gpu/drm/radeon/radeon_drv.c
>> > +++ b/drivers/gpu/drm/radeon/radeon_drv.c
>> > @@ -97,9 +97,10 @@
>> >   *   2.46.0 - Add PFP_SYNC_ME support on evergreen
>> >   *   2.47.0 - Add UVD_NO_OP register support
>> >   *   2.48.0 - TA_CS_BC_BASE_ADDR allowed on SI
>> > + *   2.49.0 - DRM_RADEON_GEM_INFO ioctl returns correct
>> > vram_size/visible values
>> >   */
>> >  #define KMS_DRIVER_MAJOR	2
>> > -#define KMS_DRIVER_MINOR	48
>> > +#define KMS_DRIVER_MINOR	49
>> >  #define KMS_DRIVER_PATCHLEVEL	0
>> >  int radeon_driver_load_kms(struct drm_device *dev, unsigned long
>> > flags);
>> >  int radeon_driver_unload_kms(struct drm_device *dev);
>> > diff --git a/drivers/gpu/drm/radeon/radeon_gem.c
>> > b/drivers/gpu/drm/radeon/radeon_gem.c
>> > index 0bcffd8a7bd3..96683f5b2b1b 100644
>> > --- a/drivers/gpu/drm/radeon/radeon_gem.c
>> > +++ b/drivers/gpu/drm/radeon/radeon_gem.c
>> > @@ -220,8 +220,8 @@ int radeon_gem_info_ioctl(struct drm_device *dev,
>> > void *data,
>> >
>> >  	man = &rdev->mman.bdev.man[TTM_PL_VRAM];
>> >
>> > -	args->vram_size = rdev->mc.real_vram_size;
>> > -	args->vram_visible = (u64)man->size << PAGE_SHIFT;
>> > +	args->vram_size = (u64)man->size << PAGE_SHIFT;
>> > +	args->vram_visible = rdev->mc.visible_vram_size;
>> >  	args->vram_visible -= rdev->vram_pin_size;
>> >  	args->gart_size = rdev->mc.gtt_size;
>> >  	args->gart_size -= rdev->gart_pin_size;
>> _______________________________________________
>> amd-gfx mailing list
>> amd-gfx@lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[Emil Velikov]

On 31 January 2017 at 15:43, Deucher, Alexander
<Alexander.Deucher@amd.com> wrote:
>> -----Original Message-----
>> From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf
>> Of Dieter Nützel
>> Sent: Tuesday, January 31, 2017 6:25 AM
>> To: Michel Dänzer
>> Cc: Alex Deucher; mesa-dev@lists.freedesktop.org; amd-
>> gfx@lists.freedesktop.org
>> Subject: Re: [Mesa-dev] [PATCH] drm/radeon: Fix vram_size/visible values in
>> DRM_RADEON_GEM_INFO ioctl
>>
>> Hello Michel,
>>
>> as this is for radeon, do you think this could/should fix
>> the wrong reported VRAM size with Unigine_Heaven/-Valley, too?
>> Maybe speed things up? ;-)
>>
>> Unigine_Valley-1.0
>>
>> GPU: Unknown GPU x1
>> System memory: 24102 MB
>> Video memory:  256 MB
>> Sync threads:  7
>> Async threads: 8
>>
>> I'll try patching openSUSE Kernel:stable 4.9.6-2 with this
>> and maybe this could then go into 4.10-rc7 'cause it is a
>> bugfix. - Alex?
>
> This patch just fixes the case of the HUD reporting the wrong amount of visible vram.  Most 3D apps just default to 256MB if they don't know how much vram is.  The problem is GL never provided a core way to determine how much vram is available on a GPU so lots of vendor specific extensions and driver specific methods popped up to address this.
>
vram_size is used for available_texture_mem in st/nine and
GLX_RENDERER_VIDEO_MEMORY_MESA via GLX_MESA_query_renderer.
So maybe we want this in older/stable kernel and mesa releases ? Not
sure how many apps use/honour these though ;-)

-Emil
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev

Re: [Mesa-dev] [PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[Alex Deucher]

On Tue, Jan 31, 2017 at 5:06 PM, Emil Velikov <emil.l.velikov@gmail.com> wrote:
> On 31 January 2017 at 15:43, Deucher, Alexander
> <Alexander.Deucher@amd.com> wrote:
>>> -----Original Message-----
>>> From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf
>>> Of Dieter Nützel
>>> Sent: Tuesday, January 31, 2017 6:25 AM
>>> To: Michel Dänzer
>>> Cc: Alex Deucher; mesa-dev@lists.freedesktop.org; amd-
>>> gfx@lists.freedesktop.org
>>> Subject: Re: [Mesa-dev] [PATCH] drm/radeon: Fix vram_size/visible values in
>>> DRM_RADEON_GEM_INFO ioctl
>>>
>>> Hello Michel,
>>>
>>> as this is for radeon, do you think this could/should fix
>>> the wrong reported VRAM size with Unigine_Heaven/-Valley, too?
>>> Maybe speed things up? ;-)
>>>
>>> Unigine_Valley-1.0
>>>
>>> GPU: Unknown GPU x1
>>> System memory: 24102 MB
>>> Video memory:  256 MB
>>> Sync threads:  7
>>> Async threads: 8
>>>
>>> I'll try patching openSUSE Kernel:stable 4.9.6-2 with this
>>> and maybe this could then go into 4.10-rc7 'cause it is a
>>> bugfix. - Alex?
>>
>> This patch just fixes the case of the HUD reporting the wrong amount of visible vram.  Most 3D apps just default to 256MB if they don't know how much vram is.  The problem is GL never provided a core way to determine how much vram is available on a GPU so lots of vendor specific extensions and driver specific methods popped up to address this.
>>
> vram_size is used for available_texture_mem in st/nine and
> GLX_RENDERER_VIDEO_MEMORY_MESA via GLX_MESA_query_renderer.
> So maybe we want this in older/stable kernel and mesa releases ? Not
> sure how many apps use/honour these though ;-)

vram_size was always correct.  it was just visible vram size that was wrong.

Alex
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [Mesa-dev] [PATCH] drm/radeon: Fix vram_size/visible values in DRM_RADEON_GEM_INFO ioctl

[Marek Olšák]

On Tue, Jan 31, 2017 at 11:23 PM, Alex Deucher <alexdeucher@gmail.com> wrote:
> On Tue, Jan 31, 2017 at 5:06 PM, Emil Velikov <emil.l.velikov@gmail.com> wrote:
>> On 31 January 2017 at 15:43, Deucher, Alexander
>> <Alexander.Deucher@amd.com> wrote:
>>>> -----Original Message-----
>>>> From: amd-gfx [mailto:amd-gfx-bounces@lists.freedesktop.org] On Behalf
>>>> Of Dieter Nützel
>>>> Sent: Tuesday, January 31, 2017 6:25 AM
>>>> To: Michel Dänzer
>>>> Cc: Alex Deucher; mesa-dev@lists.freedesktop.org; amd-
>>>> gfx@lists.freedesktop.org
>>>> Subject: Re: [Mesa-dev] [PATCH] drm/radeon: Fix vram_size/visible values in
>>>> DRM_RADEON_GEM_INFO ioctl
>>>>
>>>> Hello Michel,
>>>>
>>>> as this is for radeon, do you think this could/should fix
>>>> the wrong reported VRAM size with Unigine_Heaven/-Valley, too?
>>>> Maybe speed things up? ;-)
>>>>
>>>> Unigine_Valley-1.0
>>>>
>>>> GPU: Unknown GPU x1
>>>> System memory: 24102 MB
>>>> Video memory:  256 MB
>>>> Sync threads:  7
>>>> Async threads: 8
>>>>
>>>> I'll try patching openSUSE Kernel:stable 4.9.6-2 with this
>>>> and maybe this could then go into 4.10-rc7 'cause it is a
>>>> bugfix. - Alex?
>>>
>>> This patch just fixes the case of the HUD reporting the wrong amount of visible vram.  Most 3D apps just default to 256MB if they don't know how much vram is.  The problem is GL never provided a core way to determine how much vram is available on a GPU so lots of vendor specific extensions and driver specific methods popped up to address this.
>>>
>> vram_size is used for available_texture_mem in st/nine and
>> GLX_RENDERER_VIDEO_MEMORY_MESA via GLX_MESA_query_renderer.
>> So maybe we want this in older/stable kernel and mesa releases ? Not
>> sure how many apps use/honour these though ;-)
>
> vram_size was always correct.  it was just visible vram size that was wrong.

Just to clarify this more: the reported visible vram size currently
has no effect on anything other than scaling the Gallium HUD, which
you can simply suppress if you set the ".d" option to resize it
automatically I think.

Marek
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx