[PATCH] [xorg-dev][xserver] Full support of sRGB capable fbconfigs.
Ian Romanick
idr at freedesktop.org
Fri Nov 30 09:20:24 PST 2012
On 11/30/2012 07:18 AM, Tomasz Lis wrote:
>
> 2012/11/26 Ian Romanick <idr at freedesktop.org <mailto:idr at freedesktop.org>>
>
> On 11/22/2012 07:08 AM, Tomasz Lis wrote:
>
> From: Tomasz Lis <tomasz.lis at intel.com
> <mailto:tomasz.lis at intel.com>>
>
> Changes to correctly initialize the sRGB capability attribute
> and transfer it between XServer and the client.
> Modifications include extension string, transferring visual
> config attribs and fbconfig attribs.
> Also, attribute is initialized in the modules which do not
> really use it (xquartz and xwin).
>
>
> Please word-wrap commit messages to 72 characters.
>
> Tomasz:
> Ok. will do that.
>
> Signed-off-by: Tomasz Lis <tomasz.lis at intel.com
> <mailto:tomasz.lis at intel.com>>
> ---
> glx/extension_string.c | 1 +
> glx/extension_string.h | 1 +
> glx/glxcmds.c | 7 +++++--
> glx/glxdri2.c | 7 +++++++
> glx/glxdricommon.c | 4 +++-
> glx/glxscreens.c | 2 ++
> glx/glxscreens.h | 3 +++
> hw/xquartz/GL/visualConfigs.c | 3 +++
> hw/xwin/glx/indirect.c | 2 ++
> 9 files changed, 27 insertions(+), 3 deletions(-)
>
> diff --git a/glx/extension_string.c b/glx/extension_string.c
> index 544ca1f..47a9746 100644
> --- a/glx/extension_string.c
> +++ b/glx/extension_string.c
> @@ -74,6 +74,7 @@ static const struct extension_info
> known_glx_extensions[] = {
> { GLX(ARB_multisample), VER(1,4), Y, },
>
> { GLX(EXT_create_context_es2___profile), VER(0,0), N, },
> + { GLX(EXT_framebuffer_sRGB), VER(1,1), N, },
>
>
> Also advertise the ARB string.
>
> Tomasz:
> Agreed. Will change.
>
>
> { GLX(EXT_import_context), VER(0,0), Y, },
> { GLX(EXT_texture_from_pixmap), VER(0,0), Y, },
> { GLX(EXT_visual_info), VER(0,0), Y, },
> diff --git a/glx/extension_string.h b/glx/extension_string.h
> index 7a4a8b1..a91385f 100644
> --- a/glx/extension_string.h
> +++ b/glx/extension_string.h
> @@ -41,6 +41,7 @@ enum {
> ARB_create_context_robustness___bit,
> ARB_multisample_bit,
> EXT_create_context_es2___profile_bit,
> + EXT_framebuffer_sRGB_bit,
> EXT_import_context_bit,
> EXT_texture_from_pixmap_bit,
> EXT_visual_info_bit,
> diff --git a/glx/glxcmds.c b/glx/glxcmds.c
> index c1f4e22..720a1a8 100644
> --- a/glx/glxcmds.c
> +++ b/glx/glxcmds.c
> @@ -913,7 +913,7 @@ __glXDisp_CopyContext(____GLXclientState *
> cl, GLbyte * pc)
>
> enum {
> GLX_VIS_CONFIG_UNPAIRED = 18,
> - GLX_VIS_CONFIG_PAIRED = 20
> + GLX_VIS_CONFIG_PAIRED = 22
> };
>
> enum {
> @@ -1005,6 +1005,8 @@
> __glXDisp_GetVisualConfigs(____GLXclientState * cl, GLbyte * pc)
> buf[p++] = modes->samples;
> buf[p++] = GLX_SAMPLE_BUFFERS_SGIS;
> buf[p++] = modes->sampleBuffers;
> + buf[p++] = GLX_FRAMEBUFFER_SRGB_CAPABLE___EXT;
> + buf[p++] = modes->sRGBCapable;
> buf[p++] = 0; /* copy over
> visualSelectGroup (GLX_VISUAL_SELECT_GROUP_SGIX)__? */
> buf[p++] = 0;
>
> @@ -1017,7 +1019,7 @@
> __glXDisp_GetVisualConfigs(____GLXclientState * cl, GLbyte * pc)
> return Success;
> }
>
> -#define __GLX_TOTAL_FBCONFIG_ATTRIBS (36)
> +#define __GLX_TOTAL_FBCONFIG_ATTRIBS (37)
> #define __GLX_FBCONFIG_ATTRIBS_LENGTH
> (__GLX_TOTAL_FBCONFIG_ATTRIBS * 2)
> /**
> * Send the set of GLXFBConfigs to the client. There is not
> currently
> @@ -1109,6 +1111,7 @@ DoGetFBConfigs(____GLXclientState * cl,
> unsigned screen)
> WRITE_PAIR(GLX_BIND_TO_MIPMAP___TEXTURE_EXT,
> modes->bindToMipmapTexture);
> WRITE_PAIR(GLX_BIND_TO___TEXTURE_TARGETS_EXT,
> modes->bindToTextureTargets);
> + WRITE_PAIR(GLX_FRAMEBUFFER___SRGB_CAPABLE_EXT,
> modes->sRGBCapable);
>
>
> I believe that sRGB visuals and fbconfigs should only be sent to the
> client if the client has said that it supports one of the sRGB
> extensions. Otherwise the client doesn't know what
> GLX_FRAMEBUFFER_SRGB_CAPABLE___ARB means. It will either drop the
> visuals or treat the sRGB and non-sRGB visuals the same.
>
> Tomasz:
> The most popular client - MESA - ignores unknown attributes (in
> __glXInitializeVisualConfigFromTags). This well suits the nature of this
> specific attribute - GLX_FRAMEBUFFER_SRGB_CAPABLE_ARB/EXT informs of a
> capability, and it's up to client to either use this capability, or
> ignore it completely.
> This capability can be safely ignored by the client who does not care
> about it. To use this capability, GLX client would have to call glEnable().
>
> XServer interface documentation should contain information regarding how
> clients ought to react to unknown attributes. Also, fbconfigs and
> specific attributes shouldn't be filtered by the server - it is up to
> client to filter them. But this issue does not concern the sRGB
> capability, and should be discussed separately.
> Please let me know how you'd like this patch to be changed.
>
> Also, please use the _ARB enums instead.
>
> Tomasz:
> Please note that GLX_FRAMEBUFFER_SRGB_CAPABLE_ARB is not defined in
> glxtokens.h, only in glxext.h which requires glx.h - which is quite hard
> to include due to name conflicts with xserver.
Usually this would mean that the missing names would be added to
glxtokens.h. If these tokens were part of a core version (i.e., no
suffix at all), I'd suggest updating glxtokens.h and this patch.
However, it's probably not worth the effort in this case.
> Also, ARB and EXT defines are identical and will always be identical. I
> can't see any superiority of ARB define over EXT. Most of the other
> defines used in that file are not ARBs - but EXTs.
> Do you really think it is necessary to do such change? If you do, we
> need to discuss the correct way to #include a header with the definition
> of ARB.
> Maybe we should define GLX_FRAMEBUFFER_SRGB_CAPABLE_ARB in another
> xserver header, or patch MESA before making such change to XServer?
>
>
>
> if (client->swapped) {
> __GLX_SWAP_INT_ARRAY(buf,
> __GLX_FBCONFIG_ATTRIBS_LENGTH)__;
> diff --git a/glx/glxdri2.c b/glx/glxdri2.c
> index bce1bfa..ef7afb4 100644
> --- a/glx/glxdri2.c
> +++ b/glx/glxdri2.c
> @@ -882,6 +882,13 @@ initializeExtensions(____GLXDRIscreen * screen)
> "AIGLX: enabled GLX_SGI_swap_control and
> GLX_MESA_swap_control\n");
> }
>
> + /* enable EXT_framebuffer_sRGB extension (even if there are
> no sRGB capable fbconfigs) */
> + {
> + __glXEnableExtension(screen->__glx_enable_bits,
> + "GLX_EXT_framebuffer_sRGB");
> + LogMessage(X_INFO, "AIGLX: enabled
> GLX_EXT_framebuffer_sRGB\n");
> + }
> +
> for (i = 0; extensions[i]; i++) {
> #ifdef __DRI_READ_DRAWABLE
> if (strcmp(extensions[i]->name, __DRI_READ_DRAWABLE)
> == 0) {
> diff --git a/glx/glxdricommon.c b/glx/glxdricommon.c
> index c90f380..b027f24 100644
> --- a/glx/glxdricommon.c
> +++ b/glx/glxdricommon.c
> @@ -105,7 +105,9 @@ __ATTRIB(__DRI_ATTRIB_BUFFER___SIZE, rgbBits),
> __ATTRIB(__DRI_ATTRIB_BIND_TO___TEXTURE_RGB,
> bindToTextureRgb),
> __ATTRIB(__DRI_ATTRIB_BIND_TO___TEXTURE_RGBA,
> bindToTextureRgba),
> __ATTRIB(__DRI_ATTRIB_BIND_TO___MIPMAP_TEXTURE,
> bindToMipmapTexture),
> - __ATTRIB(__DRI_ATTRIB___YINVERTED, yInverted),};
> + __ATTRIB(__DRI_ATTRIB___YINVERTED, yInverted),
> + __ATTRIB(__DRI_ATTRIB___FRAMEBUFFER_SRGB_CAPABLE,
> sRGBCapable),
> + };
>
> static void
> setScalar(__GLXconfig * config, unsigned int attrib, unsigned
> int value)
> diff --git a/glx/glxscreens.c b/glx/glxscreens.c
> index 61d590c..e87256d 100644
> --- a/glx/glxscreens.c
> +++ b/glx/glxscreens.c
> @@ -164,7 +164,9 @@ static char GLXServerVendorName[] = "SGI";
> unsigned glxMajorVersion = SERVER_GLX_MAJOR_VERSION;
> unsigned glxMinorVersion = SERVER_GLX_MINOR_VERSION;
> static char GLXServerExtensions[] =
> + "GLX_ARB_framebuffer_sRGB "
> "GLX_ARB_multisample "
> + "GLX_EXT_framebuffer_sRGB "
> "GLX_EXT_visual_info "
> "GLX_EXT_visual_rating "
> "GLX_EXT_import_context "
> diff --git a/glx/glxscreens.h b/glx/glxscreens.h
> index b29df58..0a7b604 100644
> --- a/glx/glxscreens.h
> +++ b/glx/glxscreens.h
> @@ -102,6 +102,9 @@ struct __GLXconfig {
> GLint bindToMipmapTexture;
> GLint bindToTextureTargets;
> GLint yInverted;
> +
> + /* ARB_framebuffer_sRGB */
> + GLint sRGBCapable;
> };
>
> GLint glxConvertToXVisualType(int visualType);
> diff --git a/hw/xquartz/GL/visualConfigs.__c
> b/hw/xquartz/GL/visualConfigs.__c
> index e37eefb..18518af 100644
> --- a/hw/xquartz/GL/visualConfigs.__c
> +++ b/hw/xquartz/GL/visualConfigs.__c
> @@ -326,6 +326,9 @@ __glXAquaCreateVisualConfigs(__int
> *numConfigsPtr, int screenNumber)
>
> c->bindToTextureTargets = 0;
> c->yInverted = 0;
>
> + /* EXT_framebuffer_sRGB */
> + c->sRGBCapable =
> GL_DONT_CARE;
>
>
> This is wrong. sRGBCapable should be either True or False.
> GL_DONT_CARE (~0) is not a valid value.
>
> I can do that change, the xquartz and xwin are not really my area of
> expertise.
> Still, please consider the following scenario:
> The client asks xserver for configs, then compares them with its own
> configs. Only configs which match on client and server are accepted.
> Attribute comparison function checks if the attributes are exact match,
> or if one of them is GLX_DONT_CARE.
GLX_DONT_CARE filtering only happens in glXChooseVisual and friends.
The client / server config handshake for direct rendering is different.
Mesa has at least one or two bugs in this area. I have a patch for
one of them related to sRGB, and I could have sworn that I sent it to
you. I'll try to get it posted to the public list soon.
> This means that if we'll advertise sRGBCapable = false, then all
> sRGB-capable configs will be dropped.
> But if we'll advertise sRGBCapable = GLX_DONT_CARE, then both sRGB
> capable and non-capable configs will be accepted.
GLX_DONT_CARE violates the spec, so the rest of the argument is void.
This also argues against sending the GLX_FRAMEBUFFER_SRGB_CAPABLE
attribute to clients that don't understand it. :) If the client never
sees the attribute, it can't do the wrong thing with it.
> As we don't really support sRGB in the xquartz and xwin, this may be
> desired effect.
>
> Now, MESA works in the exact way described above - it compares the two
> sets of attributes and select matches using such comparison.
> I'm not sure if the situation I described is related to xquartz and
> xwin, so please confirm if the attributes should be set to GL_FALSE.
>
> +
> c = c->next;
> }
> }
> diff --git a/hw/xwin/glx/indirect.c b/hw/xwin/glx/indirect.c
> index c0069a2..5b339af 100644
> --- a/hw/xwin/glx/indirect.c
> +++ b/hw/xwin/glx/indirect.c
> @@ -2017,6 +2017,7 @@ glxWinCreateConfigs(HDC hdc, glxWinScreen *
> screen)
> c->base.bindToMipmapTexture = -1;
> c->base.bindToTextureTargets = -1;
> c->base.yInverted = -1;
> + c->base.sRGBCapable = -1;
>
> n++;
>
> @@ -2411,6 +2412,7 @@ glxWinCreateConfigsExt(HDC hdc, glxWinScreen *
> screen)
> GLX_TEXTURE_1D_BIT_EXT | GLX_TEXTURE_2D_BIT_EXT |
> GLX_TEXTURE_RECTANGLE_BIT_EXT;
> c->base.yInverted = -1;
> + c->base.sRGBCapable = -1;
>
> n++;
More information about the xorg-devel
mailing list