[Openchrome-users] Max Texture Size

[Boris Dušek]

Hi Thomas,

I have completed a patch, all getenv's have their driQueryOptionb counterparts 
now. I have tested the patch only by compiling and viewing the options present 
in driconf. The patch includes the previous patch I sent you with 
allow_large_textures.

Btw. driconf has a bug - it uses name "unichrome" for the device, whereas 
"via" is correct, so the config file resulting from it has no effect unless 
one changes the device name to the correct value. I will report that to xorg 
mailing list.

Questions I have:

should I completely remove the env. variable settings below, or leave it as it 
is in the patch, that is a tandem of env. variable and drirc option decides 
and env. variable has higher priotiry?

And now questions specifically for each one:

On Thursday 11 May 2006 22:21, you wrote:
> via_context.c:      if (getenv("ALTERNATE_SCREEN"))
I don't know what it does - could you please provide a description so that 
driconf users get an idea (and me too :)

> via_context.c:    if (getenv("VIA_DEBUG"))
> via_context.c:       VIA_DEBUG = driParseDebugString( getenv( "VIA_DEBUG"
this would benefit from xmlconfig.c extended to handle flag-type variables; if 
you think it is a good idea, I will try to implement it in future if/when I 
have more time;

> ), via_context.c:    if (getenv("VIA_NO_RAST"))
that was the easiest option :)

> via_context.c:    vmesa->vblank_flags = (getenv("VIA_VSYNC") ?
seen some comments about non-functionality of drirc configuration using 
driGetDefaultVBlankFlags, I too don't understand why that would not work 
(maybe the comment is outdated), will try to test it

> via_context.c:    if (getenv("VIA_PAGEFLIP"))
I need a description for it too; I added the option to 
common/xmlpool/t_options.h since other drivers also use PageFlip setting (but 
none queries neither drirc nor env. variable) and may be interested in using 
it too

> via_tex.c:   if (getenv("VIA_NO_SSE"))
What gets resolved with this is toggling the apparently buggy sse_memcpy. Now 
there is "sse_memcpy" option set by default to false. The new thing is that 
one can enable it in drirc with the full "per application settings" 
flexibility. I have reenabled the via_sse_memcpy code, so if one sets
<option name="sse_memcpy" value="true" />, it will have a real effect.

Finally DRI_CONF_FTHROTTLE_MODE is not used anywhere but is declared in 
__driConfigOptions - is it to be used somewhere, or did it accidentally 
result from copying that code from radeon driver, as possibly suggested by the 
comment before the declaration of __driConfigOptions?

Regards,
Boris

>
> These should ideally also be moved to driconf option handling, which is
> only a matter of finishing what others have started.

>
> /Thmas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: via-xmlconfig-2006-05-13.patch
Type: text/x-diff
Size: 4319 bytes
Desc: not available
Url : http://wiki.openchrome.org/pipermail/openchrome-users/attachments/20060513/1216741d/via-xmlconfig-2006-05-13.bin