Nouveau - Sep 2015 - [PATCH mesa v2 2/2] nv30: Disable msaa unless requested from the env by NV30_MAX_MSAA

We do not have a generic blitter on nv3x cards, so we must use the
sifm object for color resolving.

This commit divides the sources and dest surfaces in to tiles which
match the constraints of the sifm object, so that color resolving
will work properly on nv3x cards.

Signed-off-by: Hans de Goede <hdegoede at redhat.com>
---
Changes in v2:
-Use 1024x1024 blocks
-Use the sifm on both nv3x and nv4x cards instead of only on nv3x cards
---
 src/gallium/drivers/nouveau/nv30/nv30_miptree.c | 38 ++++++++++++++++++++++++-
 1 file changed, 37 insertions(+), 1 deletion(-)

diff --git a/src/gallium/drivers/nouveau/nv30/nv30_miptree.c
b/src/gallium/drivers/nouveau/nv30/nv30_miptree.c
index 76bb8b8..735c718 100644
--- a/src/gallium/drivers/nouveau/nv30/nv30_miptree.c
+++ b/src/gallium/drivers/nouveau/nv30/nv30_miptree.c
@@ -149,14 +149,50 @@ static void
 nv30_resource_resolve(struct nv30_context *nv30,
                       const struct pipe_blit_info *info)
 {
+   struct nv30_miptree *src_mt = nv30_miptree(info->src.resource);
    struct nv30_rect src, dst;
+   unsigned x, x0, x1, y, y1, w, h;
 
    define_rect(info->src.resource, 0, info->src.box.z,
info->src.box.x,
       info->src.box.y, info->src.box.width, info->src.box.height,
&src);
    define_rect(info->dst.resource, 0, info->dst.box.z,
info->dst.box.x,
       info->dst.box.y, info->dst.box.width, info->dst.box.height,
&dst);
 
-   nv30_transfer_rect(nv30, BILINEAR, &src, &dst);
+   x0 = src.x0;
+   x1 = src.x1;
+   y1 = src.y1;
+
+   /* On nv3x we must use sifm which is restricted to 1024x1024 tiles */
+   for (y = src.y0; y < y1; y += h) {
+      h = y1 - y;
+      if (h > 1024)
+         h = 1024;
+
+      src.y0 = 0;
+      src.y1 = h;
+      src.h = h;
+
+      dst.y1 = dst.y0 + (h >> src_mt->ms_y);
+      dst.h = h >> src_mt->ms_y;
+
+      for (x = x0; x < x1; x += w) {
+         w = x1 - x;
+         if (w > 1024)
+            w = 1024;
+
+         src.offset = y * src.pitch + x * src.cpp;
+         src.x0 = 0;
+         src.x1 = w;
+         src.w = w;
+
+         dst.offset = (y >> src_mt->ms_y) * dst.pitch +
+                      (x >> src_mt->ms_x) * dst.cpp;
+         dst.x1 = dst.x0 + (w >> src_mt->ms_x);
+         dst.w = w >> src_mt->ms_x;
+
+         nv30_transfer_rect(nv30, BILINEAR, &src, &dst);
+      }
+   }
 }
 
 void
-- 
2.4.3

Some modern apps try to use msaa without keeping in mind the
restrictions on videomem of older cards. Resulting in dmesg saying:

 [ 1197.850642] nouveau E[soffice.bin[3785]] fail ttm_validate
 [ 1197.850648] nouveau E[soffice.bin[3785]] validating bo list
 [ 1197.850654] nouveau E[soffice.bin[3785]] validate: -12

Because we are running out of video memory, after which the program
using the msaa visual freezes, and eventually the entire system freezes.

To work around this we do not allow msaa visauls by default and allow
the user to override this via NV30_MAX_MSAA.

Signed-off-by: Hans de Goede <hdegoede at redhat.com>
---
Changes in v2:
-Allow re-enabling msaa by setting NV30_MAX_MSAA in the environment
---
 src/gallium/drivers/nouveau/nv30/nv30_screen.c | 22 +++++++++++++++++++++-
 1 file changed, 21 insertions(+), 1 deletion(-)

diff --git a/src/gallium/drivers/nouveau/nv30/nv30_screen.c
b/src/gallium/drivers/nouveau/nv30/nv30_screen.c
index 7aad26b..4b77f43 100644
--- a/src/gallium/drivers/nouveau/nv30/nv30_screen.c
+++ b/src/gallium/drivers/nouveau/nv30/nv30_screen.c
@@ -319,8 +319,28 @@ nv30_screen_is_format_supported(struct pipe_screen
*pscreen,
                                 unsigned sample_count,
                                 unsigned bindings)
 {
-   if (sample_count > 4)
+   unsigned int max_sample_count;
+
+   /*
+    * Some modern apps try to use msaa without keeping in mind the
+    * restrictions on videomem of older cards. Resulting in dmesg saying:
+    * [ 1197.850642] nouveau E[soffice.bin[3785]] fail ttm_validate
+    * [ 1197.850648] nouveau E[soffice.bin[3785]] validating bo list
+    * [ 1197.850654] nouveau E[soffice.bin[3785]] validate: -12
+    *
+    * Because we are running out of video memory, after which the program
+    * using the msaa visual freezes, and eventually the entire system freezes.
+    *
+    * To work around this we do not allow msaa visauls by default and allow
+    * the user to override this via NV30_MAX_MSAA.
+    */
+   max_sample_count = debug_get_num_option("NV30_MAX_MSAA", 0);
+   if (max_sample_count > 4)
+      max_sample_count = 4;
+
+   if (sample_count > max_sample_count)
       return false;
+
    if (!(0x00000017 & (1 << sample_count)))
       return false;
 
-- 
2.4.3

On Wed, Sep 9, 2015 at 9:52 AM, Hans de Goede <hdegoede at redhat.com>
wrote:
> We do not have a generic blitter on nv3x cards, so we must use the
> sifm object for color resolving.
>
> This commit divides the sources and dest surfaces in to tiles which
> match the constraints of the sifm object, so that color resolving
> will work properly on nv3x cards.
>
> Signed-off-by: Hans de Goede <hdegoede at redhat.com>
> ---
> Changes in v2:
> -Use 1024x1024 blocks
> -Use the sifm on both nv3x and nv4x cards instead of only on nv3x cards
Thanks, pushed.
> ---
>  src/gallium/drivers/nouveau/nv30/nv30_miptree.c | 38
++++++++++++++++++++++++-
>  1 file changed, 37 insertions(+), 1 deletion(-)
>
> diff --git a/src/gallium/drivers/nouveau/nv30/nv30_miptree.c
b/src/gallium/drivers/nouveau/nv30/nv30_miptree.c
> index 76bb8b8..735c718 100644
> --- a/src/gallium/drivers/nouveau/nv30/nv30_miptree.c
> +++ b/src/gallium/drivers/nouveau/nv30/nv30_miptree.c
> @@ -149,14 +149,50 @@ static void
>  nv30_resource_resolve(struct nv30_context *nv30,
>                        const struct pipe_blit_info *info)
>  {
> +   struct nv30_miptree *src_mt = nv30_miptree(info->src.resource);
>     struct nv30_rect src, dst;
> +   unsigned x, x0, x1, y, y1, w, h;
>
>     define_rect(info->src.resource, 0, info->src.box.z,
info->src.box.x,
>        info->src.box.y, info->src.box.width, info->src.box.height,
&src);
>     define_rect(info->dst.resource, 0, info->dst.box.z,
info->dst.box.x,
>        info->dst.box.y, info->dst.box.width, info->dst.box.height,
&dst);
>
> -   nv30_transfer_rect(nv30, BILINEAR, &src, &dst);
> +   x0 = src.x0;
> +   x1 = src.x1;
> +   y1 = src.y1;
> +
> +   /* On nv3x we must use sifm which is restricted to 1024x1024 tiles */
> +   for (y = src.y0; y < y1; y += h) {
> +      h = y1 - y;
> +      if (h > 1024)
> +         h = 1024;
> +
> +      src.y0 = 0;
> +      src.y1 = h;
> +      src.h = h;
> +
> +      dst.y1 = dst.y0 + (h >> src_mt->ms_y);
> +      dst.h = h >> src_mt->ms_y;
> +
> +      for (x = x0; x < x1; x += w) {
> +         w = x1 - x;
> +         if (w > 1024)
> +            w = 1024;
> +
> +         src.offset = y * src.pitch + x * src.cpp;
> +         src.x0 = 0;
> +         src.x1 = w;
> +         src.w = w;
> +
> +         dst.offset = (y >> src_mt->ms_y) * dst.pitch +
> +                      (x >> src_mt->ms_x) * dst.cpp;
> +         dst.x1 = dst.x0 + (w >> src_mt->ms_x);
> +         dst.w = w >> src_mt->ms_x;
> +
> +         nv30_transfer_rect(nv30, BILINEAR, &src, &dst);
> +      }
> +   }
>  }
>
>  void
> --
> 2.4.3
>

On Wed, Sep 9, 2015 at 9:52 AM, Hans de Goede <hdegoede at redhat.com>
wrote:
> Some modern apps try to use msaa without keeping in mind the
> restrictions on videomem of older cards. Resulting in dmesg saying:
>
>  [ 1197.850642] nouveau E[soffice.bin[3785]] fail ttm_validate
>  [ 1197.850648] nouveau E[soffice.bin[3785]] validating bo list
>  [ 1197.850654] nouveau E[soffice.bin[3785]] validate: -12
>
> Because we are running out of video memory, after which the program
> using the msaa visual freezes, and eventually the entire system freezes.
>
> To work around this we do not allow msaa visauls by default and allow
> the user to override this via NV30_MAX_MSAA.
>
> Signed-off-by: Hans de Goede <hdegoede at redhat.com>
> ---
> Changes in v2:
> -Allow re-enabling msaa by setting NV30_MAX_MSAA in the environment
Moved this to screen init and pushed. is_format_supported is called a
lot, not sure how expensive getenv() is, don't want to find out.
> ---
>  src/gallium/drivers/nouveau/nv30/nv30_screen.c | 22 +++++++++++++++++++++-
>  1 file changed, 21 insertions(+), 1 deletion(-)
>
> diff --git a/src/gallium/drivers/nouveau/nv30/nv30_screen.c
b/src/gallium/drivers/nouveau/nv30/nv30_screen.c
> index 7aad26b..4b77f43 100644
> --- a/src/gallium/drivers/nouveau/nv30/nv30_screen.c
> +++ b/src/gallium/drivers/nouveau/nv30/nv30_screen.c
> @@ -319,8 +319,28 @@ nv30_screen_is_format_supported(struct pipe_screen
*pscreen,
>                                  unsigned sample_count,
>                                  unsigned bindings)
>  {
> -   if (sample_count > 4)
> +   unsigned int max_sample_count;
> +
> +   /*
> +    * Some modern apps try to use msaa without keeping in mind the
> +    * restrictions on videomem of older cards. Resulting in dmesg saying:
> +    * [ 1197.850642] nouveau E[soffice.bin[3785]] fail ttm_validate
> +    * [ 1197.850648] nouveau E[soffice.bin[3785]] validating bo list
> +    * [ 1197.850654] nouveau E[soffice.bin[3785]] validate: -12
> +    *
> +    * Because we are running out of video memory, after which the program
> +    * using the msaa visual freezes, and eventually the entire system
freezes.
> +    *
> +    * To work around this we do not allow msaa visauls by default and
allow
> +    * the user to override this via NV30_MAX_MSAA.
> +    */
> +   max_sample_count = debug_get_num_option("NV30_MAX_MSAA", 0);
> +   if (max_sample_count > 4)
> +      max_sample_count = 4;
> +
> +   if (sample_count > max_sample_count)
>        return false;
> +
>     if (!(0x00000017 & (1 << sample_count)))
>        return false;
>
> --
> 2.4.3
>