Michel:

Considering my last email, how is this patch that I am attaching?  Since you
said hardware only supports a 24 bit depth buffer with stencil, that's what
should be appended to fbconfigs.  It works for me in both situations, with
and without depthbits set to 16.

Regards

2009/4/18 Mario Limonciello <[email protected]>

> Hi Michel:
>
> 2009/4/18 Michel Dänzer <[email protected]>
>
>>
>> Yes, apparently the app / toolkit ends up using the depth 32 visual
>> (maybe because without your patch it doesn't have stencil bits - in
>> which case your patch probably won't help either when running the X
>> server in depth 16 or with Option "DepthBits" "16") but can't really
>> handle it.
>>
>
> Am I just reading the code wrong then?  It looks like both before my patch
> and with my patch, stencil_bits would be set to 8 on the 32 depth.  You had
> also said that the hardware only supports a 24 bit depth buffer with
> stencil.  In my patch, wasn't stencil_bits also set to 8 on 24 bit?
>
>
>>
>>
>> Up until recently, the depth 32 visual was marked as nonconformant to
>> prevent apps from picking it up when they don't really need the real
>> alpha channel. Maybe we should go back to doing that (or mark it as slow
>> instead), but some people would argue that it's up to the apps /
>> toolkits to choose the visual more carefully.
>>
>
> The perplexing thing then is that this is only happening with radeon.  I've
> checked with -intel, -vesa, and -nv and they're all trouble free.  I'd
> assume they're all choosing the default visual as well, but the particular
> problem doesn't come up with mesa 7.4 on them.
>
> This being the case, I seem to think the problem really should lie in the
> implementation of DRI on radeon.
>
> Regards
>
> --
> Mario Limonciello
> [email protected]
> Sent from Austin, Texas, United States
>



-- 
Mario Limonciello
[email protected]
Sent from Austin, Texas, United States
--- a/src/mesa/drivers/dri/radeon/radeon_screen.c	2009-04-18 11:58:34.000000000 -0500
+++ b/src/mesa/drivers/dri/radeon/radeon_screen.c	2009-04-18 15:37:46.000000000 -0500
@@ -1156,6 +1156,8 @@
 static const __DRIconfig **
 radeonInitScreen(__DRIscreenPrivate *psp)
 {
+    const __DRIconfig **current_configs;
+
 #if !RADEON_COMMON
    static const char *driver_name = "Radeon";
    static const __DRIutilversion2 ddx_expected = { 4, 5, 0, 0 };
@@ -1203,11 +1205,16 @@
    if (!radeonInitDriver(psp))
        return NULL;
 
-   return radeonFillInModes( psp,
-			     dri_priv->bpp,
-			     (dri_priv->bpp == 16) ? 16 : 24,
-			     (dri_priv->bpp == 16) ? 0  : 8,
-			     (dri_priv->backOffset != dri_priv->depthOffset) );
+    current_configs = radeonFillInModes( psp,
+                            dri_priv->bpp,
+                            (dri_priv->bpp == 16) ? 16 : 24,
+                            (dri_priv->bpp == 16) ? 0  : 8,
+                            (dri_priv->backOffset != dri_priv->depthOffset) );
+    if (dri_priv->bpp > 24)
+        current_configs = driConcatConfigs(current_configs,
+            radeonFillInModes(psp, 24, 24, 8, (dri_priv->backOffset != dri_priv->depthOffset)));
+
+   return current_configs;
 }
 
 
------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and 
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today. 
Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
Mesa3d-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mesa3d-dev

Reply via email to