blob: 663e6272bcb0984b256bf4b61bb2a5640e45a1ed [file] [log] [blame]
Name
ARB_texture_storage
Name Strings
GL_ARB_texture_storage
Contact
Bruce Merry (bmerry 'at' gmail.com)
Ian Romanick, Intel (ian.d.romanick 'at' intel.com)
Contributors
Jeremy Sandmel, Apple
Bruce Merry, ARM
Tom Olson, ARM
Benji Bowman, Imagination Technologies
Ian Romanick, Intel
Jeff Bolz, NVIDIA
Pat Brown, NVIDIA
Maurice Ribble, Qualcomm
Lingjun Chen, Qualcomm
Daniel Koch, Transgaming Inc
Mark Kilgard, NVIDIA
Notice
Copyright (c) 2011-2013 The Khronos Group Inc. Copyright terms at
http://www.khronos.org/registry/speccopyright.html
Specification Update Policy
Khronos-approved extension specifications are updated in response to
issues and bugs prioritized by the Khronos OpenGL Working Group. For
extensions which have been promoted to a core Specification, fixes will
first appear in the latest version of that core Specification, and will
eventually be backported to the extension document. This policy is
described in more detail at
https://www.khronos.org/registry/OpenGL/docs/update_policy.php
Status
Complete. Approved by the ARB on 2011/06/20.
Approved by the Khronos Promoters on 2011/07/29.
Version
Last Modified Date: August 7, 2013
Author Revision: 26
Number
ARB Extension #117
Dependencies
OpenGL ES 1.0, OpenGL ES 2.0 or OpenGL 1.2 is required.
OES_texture_npot, OES_texture_cube_map, OES_texture_3D,
OES_depth_texture, OES_packed_depth_stencil,
OES_compressed_paletted_texture, EXT_texture_type_2_10_10_10_REV,
EXT_texture3D, OES_texture_npot, APPLE_texture_2D_limited_npot,
ARB_texture_cube_map, ARB_texture_cube_map_array,
ARB_texture_rectangle, SGIS_generate_mipmap,
EXT_direct_state_access, OES_EGL_image, WGL_ARB_render_texture,
GLX_EXT_texture_from_pixmap, and core specifications that
incorporate these extensions affect the definition of this
extension.
This extension is written against the OpenGL 3.2 Core Profile
specification.
Overview
The texture image specification commands in OpenGL allow each level
to be separately specified with different sizes, formats, types and
so on, and only imposes consistency checks at draw time. This adds
overhead for implementations.
This extension provides a mechanism for specifying the entire
structure of a texture in a single call, allowing certain
consistency checks and memory allocations to be done up front. Once
specified, the format and dimensions of the image array become
immutable, to simplify completeness checks in the implementation.
When using this extension, it is no longer possible to supply texture
data using TexImage*. Instead, data can be uploaded using TexSubImage*,
or produced by other means (such as render-to-texture, mipmap generation,
or rendering to a sibling EGLImage).
This extension has complicated interactions with other extensions.
The goal of most of these interactions is to ensure that a texture
is always mipmap complete (and cube complete for cubemap textures).
IP Status
No known IP claims
New Procedures and Functions
void TexStorage1D(enum target, sizei levels,
enum internalformat,
sizei width);
void TexStorage2D(enum target, sizei levels,
enum internalformat,
sizei width, sizei height);
void TexStorage3D(enum target, sizei levels,
enum internalformat,
sizei width, sizei height, sizei depth);
When EXT_direct_state_access is present:
void TextureStorage1DEXT(uint texture, enum target, sizei levels,
enum internalformat,
sizei width);
void TextureStorage2DEXT(uint texture, enum target, sizei levels,
enum internalformat,
sizei width, sizei height);
void TextureStorage3DEXT(uint texture, enum target, sizei levels,
enum internalformat,
sizei width, sizei height, sizei depth);
New Types
None
New Tokens
Accepted by the <value> parameter of GetTexParameter{if}v:
TEXTURE_IMMUTABLE_FORMAT 0x912F
Accepted by the <internalformat> parameter of TexStorage* when
implemented on OpenGL ES:
ALPHA8_EXT 0x803C
LUMINANCE8_EXT 0x8040
LUMINANCE8_ALPHA8_EXT 0x8045
Additions to Chapter 2 of the OpenGL 3.2 Core Profile Specification
(OpenGL Operation)
None
Additions to Chapter 3 of the OpenGL 3.2 Core Profile Specification
(Rasterization)
After section 3.8.1 (Texture Image Specification) add a new
subsection called "Immutable-format texture images":
"An alterative set of commands is provided for specifying the
properties of all levels of a texture at once. Once a texture is
specified with such a command, the format and dimensions of all
levels becomes immutable, unless it is a proxy texture (since
otherwise it would no longer be possible to use the proxy). The
contents of the images and the parameters can still be modified.
Such a texture is referred to as an "immutable-format" texture. The
immutability status of a texture can be determined by calling
GetTexParameter with <pname> TEXTURE_IMMUTABLE_FORMAT.
Each of the commands below is described by pseudo-code which
indicates the effect on the dimensions and format of the texture.
For all of the commands, the following apply in addition to the
pseudo-code:
- If the default texture object is bound to <target>, an
INVALID_OPERATION error is generated.
- If executing the pseudo-code would lead to an error, the error is
generated and the command will have no effect.
- Any existing levels that are not replaced are reset to their
initial state.
- If <width>, <height>, <depth> or <levels> is less than 1, the
error INVALID_VALUE is generated.
- The pixel unpack buffer should be considered to be zero i.e.,
the image contents are unspecified.
- Since no pixel data are provided, the <format> and <type> values
used in the pseudo-code are irrelevant; they can be considered to
be any values that are legal to use with <internalformat>.
- If the command is successful, TEXTURE_IMMUTABLE_FORMAT becomes
TRUE.
- If <internalformat> is a specific compressed texture format, then
references to TexImage* should be replaced by CompressedTexImage*,
with <format>, <type> and <data> replaced by any valid <imageSize> and
<data>. If there is no <imageSize> for which this command would have
been valid, an INVALID_OPERATION error is generated [fn: This
condition is not required for OpenGL, but is necessary for OpenGL
ES which does not support on-the-fly compression.]
- If <internalformat> is one of the internal formats listed in table
3.11, an INVALID_ENUM error is generated.
The command
void TexStorage1D(enum target, sizei levels,
enum internalformat,
sizei width);
specifies all the levels of a one-dimensional texture (or proxy) at
the same time. It is described by the pseudo-code below:
for (i = 0; i < levels; i++)
{
TexImage1D(target, i, internalformat, width, 0,
format, type, NULL);
width = max(1, floor(width / 2));
}
If <target> is not TEXTURE_1D or PROXY_TEXTURE_1D then INVALID_ENUM
is generated. If <levels> is greater than floor(log_2(width)) + 1
then INVALID_OPERATION is generated.
The command
void TexStorage2D(enum target, sizei levels,
enum internalformat,
sizei width, sizei height);
specifies all the levels of a two-dimensional, cube-map,
one-dimension array or rectangle texture (or proxy) at the same
time. The pseudo-code depends on the <target>:
[PROXY_]TEXTURE_2D, [PROXY_]TEXTURE_RECTANGLE or
PROXY_TEXTURE_CUBE_MAP:
for (i = 0; i < levels; i++)
{
TexImage2D(target, i, internalformat, width, height, 0,
format, type, NULL);
width = max(1, floor(width / 2));
height = max(1, floor(height / 2));
}
TEXTURE_CUBE_MAP:
for (i = 0; i < levels; i++)
{
for face in (+X, -X, +Y, -Y, +Z, -Z)
{
TexImage2D(face, i, internalformat, width, height, 0,
format, type, NULL);
}
width = max(1, floor(width / 2));
height = max(1, floor(height / 2));
}
[PROXY_]TEXTURE_1D_ARRAY:
for (i = 0; i < levels; i++)
{
TexImage2D(target, i, internalformat, width, height, 0,
format, type, NULL);
width = max(1, floor(width / 2));
}
If <target> is not one of those listed above, the error INVALID_ENUM
is generated.
The error INVALID_OPERATION is generated if any of the following
conditions hold:
- <target> is [PROXY_]TEXTURE_1D_ARRAY and <levels> is greater than
floor(log_2(width)) + 1
- <target> is not [PROXY_]TEXTURE_1D_ARRAY and <levels> is greater
than floor(log_2(max(width, height))) + 1
The command
void TexStorage3D(enum target, sizei levels, enum internalformat,
sizei width, sizei height, sizei depth);
specifies all the levels of a three-dimensional, two-dimensional
array texture, or cube-map array texture (or proxy). The pseudo-code
depends on <target>:
[PROXY_]TEXTURE_3D:
for (i = 0; i < levels; i++)
{
TexImage3D(target, i, internalformat, width, height, depth, 0,
format, type, NULL);
width = max(1, floor(width / 2));
height = max(1, floor(height / 2));
depth = max(1, floor(depth / 2));
}
[PROXY_]TEXTURE_2D_ARRAY, [PROXY_]TEXTURE_CUBE_MAP_ARRAY:
for (i = 0; i < levels; i++)
{
TexImage3D(target, i, internalformat, width, height, depth, 0,
format, type, NULL);
width = max(1, floor(width / 2));
height = max(1, floor(height / 2));
}
If <target> is not one of those listed above, the error INVALID_ENUM
is generated.
The error INVALID_OPERATION is generated if any of the following
conditions hold:
- <target> is [PROXY_]TEXTURE_3D and <levels> is greater than
floor(log_2(max(width, height, depth))) + 1
- <target> is [PROXY_]TEXTURE_2D_ARRAY or
[PROXY_]TEXTURE_CUBE_MAP_ARRAY and <levels> is greater than
floor(log_2(max(width, height))) + 1
After a successful call to any TexStorage* command with a non-proxy
target, the value of TEXTURE_IMMUTABLE_FORMAT for this texture
object is set to TRUE, and no further changes to the dimensions or
format of the texture object may be made. Other commands may only
alter the texel values and texture parameters. Using any of the
following commands with the same texture will result in the error
INVALID_OPERATION being generated, even if it does not affect the
dimensions or format:
- TexImage*
- CompressedTexImage*
- CopyTexImage*
- TexStorage*
The TextureStorage* commands operate identically to the
corresponding command where "Texture" is substituted for "Tex"
except, rather than updating the current bound texture for the
texture unit indicated by the current active texture state and the
target parameter, these "Texture" commands update the texture object
named by the initial texture parameter. The error INVALID_VALUE
is generated if <texture> is zero.
"
In section 3.8.6 (Texture Parameters), after the sentence
"In the remainder of section 3.8, denote by lod_min, lod_max,
level_base, and level_max the values of the texture parameters
TEXTURE_MIN_LOD, TEXTURE_MAX_LOD, TEXTURE_BASE_LEVEL, and
TEXTURE_MAX_LEVEL respectively."
add
"However, if TEXTURE_IMMUTABLE_FORMAT is
TRUE, then level_base is clamped to the range [0, <levels> - 1] and
level_max is then clamped to the range [level_base, <levels> - 1],
where <levels> is the parameter passed the call to TexStorage* for
the texture object.
In section 3.8.9 (Rendering feedback loops) replace all references
to TEXTURE_BASE_LEVEL by level_base.
In section 3.8.9 (Mipmapping), replace the paragraph starting "Each
array in a mipmap is defined..." by
"Each array in a mipmap is defined using TexImage3D, TexImage2D,
CopyTexImage2D, TexImage1D, CopyTexImage1D, or by functions that are
defined in terms of these functions. Level-of-detail numbers proceed
from level_base for the original texel array through the maximum
level p, with each unit increase indicating an array of half the
dimensions of the previous one (rounded down to the next integer if
fractional) as already described. For immutable-format textures,
p is one less than the <levels> parameter passed to TexStorage*;
otherwise p = floor(log_2(maxsize)) + level_base. All arrays from
level_base through q = min(p, level_max) must be defined, as
discussed in section 3.8.12."
In section 3.8.12 (Texture Completeness), modify the last sentence
to avoid refering to level_base and level_max:
"An implementation may allow a texture image array of level 1 or
greater to be created only if a mipmap complete set of image arrays
consistent with the requested array can be supported where the
values of TEXTURE_BASE_LEVEL and TEXTURE_MAX_LEVEL are 0 and 1000
respectively."
Modify section 3.8.13 (Texture State and Proxy State) to add the new
state:
"Each set consists of ..., and a boolean flag indicating whether the
format and dimensions of the texture are immutable."
Add
"The initial value of TEXTURE_IMMUTABLE_FORMAT is FALSE."
Additions to Chapter 4 of the OpenGL 3.2 Core Profile Specification
(Per-Fragment Operations and the Frame Buffer)
None
Additions to Chapter 5 of the OpenGL 3.2 Compatibility Profile Specification
(Special Functions)
In section 5.4.1 (Commands Not Usable in Display Lists), add
TexStorage* to the list of commands that cannot be used.
Additions to Chapter 6 of the OpenGL 3.2 Core Profile Specification
(State and State Requests)
Replace the following statement in 6.1.3 (Enumerated Queries):
"<value> must be one of the symbolic values in table 3.10."
with
"<value> must be TEXTURE_IMMUTABLE_FORMAT or one of the symbolic
values in table 3.22."
Additions to the AGL/EGL/GLX/WGL Specifications
None
Additions to OES_compressed_ETC1_RGB8_texture
Add the following to the additions to Chapter 3:
"Since ETC1 images are easily edited along 4x4 texel boundaries, the
limitations on CompressedTexSubImage2D are relaxed.
CompressedTexSubImage2D will result in an INVALID_OPERATION error
only if one of the following conditions occurs:
* <width> is not a multiple of four, and <width> plus <xoffset> is not
equal to the texture width;
* <height> is not a multiple of four, and <height> plus <yoffset> is
not equal to the texture height; or
* <xoffset> or <yoffset> is not a multiple of four.
Remove CompressedTexSubImage2D from this error:
"INVALID_OPERATION is generated by CompressedTexSubImage2D,
TexSubImage2D, or CopyTexSubImage2D if the texture image <level>
bound to <target> has internal format ETC1_RGB8_OES."
Add the following error:
"INVALID_OPERATION is generated by CompressedTexSubImage2D
if the region to be modified is not aligned to block boundaries
(refer to the extension text for details)."
Additions to AMD_compressed_ATC_texture and AMD_compressed_3DC_texture:
Apply the same changes as for OES_compressed_ETC1_RGB8_texture
above, substituting the appropriate internal format tokens from
these extensions.
Dependencies on EXT_direct_state_access
If EXT_direct_state_access is not present, references to
TextureStorage* should be ignored.
Dependencies on OpenGL ES
On OpenGL ES without extensions introducing TEXTURE_MAX_LEVEL,
mipmapped textures specified with TexStorage are required to have a
full set of mipmaps. If TEXTURE_MAX_LEVEL is not supported, this
extension is modified as follows:
- Where an upper bound is placed on <levels> in this extension (i.e.
the maximum number of mipmap levels for a texture of the given
target and dimensions), an INVALID_OPERATION error is generated if
<levels> is neither 1 nor this upper bound.
- q (the effective maximum number of levels) is redefined to clamp
to the number of levels present in immutable-format textures.
OpenGL ES does not accept sized internal formats (e.g., RGBA8) and
instead derives an internal format from the <format> and <type>
parameters of TexImage2D. Since TexStorage* does not specify texel
data, the API doesn't include <format> and <type> parameters.
On an OpenGL ES implementation, the values in the <internalformat>
column in the tables below are accepted as <internalformat>
parameters, and base internal formats are not accepted. The
TexImage* calls in the TexStorage* pseudocode are modified so that
the <internalformat>, <format> and <type> parameters are
taken from the <format>, <format> and <type> columns (respectively)
in the tables below, according to the <internalformat>
specified in the TexStorage* command.
<internalformat> <format> <type>
---------------- -------- ------
RGB565 RGB UNSIGNED_SHORT_5_6_5
RGBA4 RGBA UNSIGNED_SHORT_4_4_4_4
RGB5_A1 RGBA UNSIGNED_SHORT_5_5_5_1
RGB8_OES RGB UNSIGNED_BYTE
RGBA8_OES RGBA UNSIGNED_BYTE
LUMINANCE8_ALPHA8_EXT LUMINANCE_ALPHA UNSIGNED_BYTE
LUMINANCE8_EXT LUMINANCE UNSIGNED_BYTE
ALPHA8_EXT ALPHA UNSIGNED_BYTE
If OES_depth_texture is supported:
<internalformat> <format> <type>
---------------- -------- ------
DEPTH_COMPONENT16_OES DEPTH_COMPONENT UNSIGNED_SHORT
DEPTH_COMPONENT32_OES DEPTH_COMPONENT UNSIGNED_INT
If OES_packed_depth_stencil is supported:
<internalformat> <format> <type>
---------------- -------- ------
DEPTH24_STENCIL8_OES DEPTH_STENCIL_OES UNSIGNED_INT
If EXT_texture_type_2_10_10_10_REV is supported:
<internalformat> <format> <type>
---------------- -------- ------
RGB10_A2_EXT RGBA UNSIGNED_INT_2_10_10_10_REV_EXT
RGB10_EXT RGB UNSIGNED_INT_2_10_10_10_REV_EXT
Dependencies on texture targets
If a particular texture target is not supported by the
implementation, passing it as a <target> to TexStorage* will
generate an INVALID_ENUM error. If as a result, any of the commands
defined in this extension would no longer have any valid <target>,
all references to the command should be ignored.
Dependencies on OES_texture_npot
If OpenGL ES 2.0 or APPLE_texture_2D_limited_npot is present but
OES_texture_npot is not present, then INVALID_OPERATION is
generated by TexStorage* and TexStorage3D if <levels> is
not one and <width>, <height> or <depth> is not a power of
two.
Dependencies on WGL_ARB_render_texture, GLX_EXT_texture_from_pixmap, EGL
1.4 and GL_OES_EGL_image
The commands eglBindTexImage, wglBindTexImageARB, glXBindTexImageEXT or
EGLImageTargetTexture2DOES are not permitted on an immutable-format
texture.
They will generate the following errors:
- EGLImageTargetTexture2DOES: INVALID_OPERATION
- eglBindTexImage: EGL_BAD_MATCH
- wglBindTexImage: ERROR_INVALID_OPERATION
- glXBindTexImageEXT: BadMatch
Dependencies on OES_compressed_paletted_texture
The compressed texture formats exposed by
OES_compressed_paletted_texture are not supported by TexStorage*.
Passing one of these tokens to TexStorage* will generate an
INVALID_ENUM error.
Errors
Note that dependencies above modify the errors.
If TexStorage* is called with a <width>, <height>, <depth> or
<levels> parameter that is less than one, then the error
INVALID_VALUE is generated.
If the <target> parameter to TexStorage1D is not
[PROXY_]TEXTURE_1D, then the error INVALID_ENUM is generated.
If the <target> parameter to TexStorage2D is not
[PROXY_]TEXTURE_2D, [PROXY_]TEXTURE_CUBE_MAP,
[PROXY_]TEXTURE_RECTANGLE or [PROXY_]TEXTURE_1D_ARRAY, then the
error INVALID_ENUM is generated.
If the <target> parameter to TexStorage3D is not
[PROXY_]TEXTURE_3D, [PROXY_]TEXTURE_2D_ARRAY or
[PROXY_]TEXTURE_CUBE_MAP_ARRAY then the error INVALID_ENUM is
generated.
If the <levels> parameter to TexStorage* is greater than the
<target>-specific value listed below then the error
INVALID_OPERATION is generated:
[PROXY_]TEXTURE_{1D,1D_ARRAY}:
floor(log_2(width)) + 1
[PROXY_]TEXTURE_{2D,2D_ARRAY,CUBE_MAP,CUBE_MAP_ARRAY}:
floor(log_2(max(width, height))) + 1
[PROXY_]TEXTURE_3D:
floor(log_2(max(width, height, depth))) + 1
[PROXY_]TEXTURE_RECTANGLE:
1
If the default texture object is bound to the <target> passed to
TexStorage*, then the error INVALID_OPERATION is generated.
If the <target> parameter to TextureStorage* does not match the
dimensionality of <texture>, then the error INVALID_OPERATION is
generated.
If the <texture> parameter to TextureStorage* is zero, then the
INVALID_VALUE is generated.
If any pseudo-code listed in this extension would generate an error,
then that error is generated.
Calling any of the following functions on a texture for which
TEXTURE_IMMUTABLE_FORMAT is TRUE will generate an
INVALID_OPERATION error:
- TexImage*
- CompressedTexImage*
- CopyTexImage*
New State
Additions to Table 6.8 Textures (state per texture object)
Initial
Get Value Type Get Command Value Description Sec.
--------- ---- ----------- ------- ----------- ----
TEXTURE_IMMUTABLE_FORMAT B GetTexParameter FALSE Size and format immutable 2.6
New Implementation Dependent State
None
Issues
1. What should this extension be called?
RESOLVED: EXT_texture_storage is chosen for consistency with the
glRenderbufferStorage entry point.
2. Should TexStorage* accept a border parameter?
RESOLVED: no.
DISCUSSION: Currently it does not, since borders are a deprecated
feature which is not supported by all hardware. Users of the
compatibility profile can continue to use the existing texture
specification functions, but there is an argument that users of
compatibility profile may also want to use this extension.
3. What is the correct error when <levels> specifies a partial
mipmap pyramid for OpenGL ES?
RESOLVED: INVALID_OPERATION, since it is an interaction between
parameters rather than a single value being invalid. It also makes
sense to relax this condition for desktop GL where it makes sense to
use a truncated pyramid with TEXTURE_MAX_LEVEL.
4. Should use of these entry-points make the metadata (format and
dimensions) immutable?
RESOLVED: Yes.
DISCUSSION: The benefits of knowing metadata can't change will
probably outweigh the extra cost of checking the
TEXTURE_IMMUTABLE_FORMAT flag on each texture specification
call.
5. Should it be legal to completely replace the texture using a new call
to TexStorage*?
RESOLVED. It will not be allowed.
DISCUSSION: This is useful to invalidate all levels of a texture.
Allowing the metadata to be changed here seems easier than trying to
define a portable definition of what it means to change the metadata
(e.g. what if you used an unsized internal format the first time and
the corresponding sized internal format the second time, or vice
versa)?
However, while this is largely similar to deleting the old texture
object and replacing it with a new one, it does lose some of the
advantages of immutability. Specifically, because doing so does not
reset bindings, it doesn't allow a migration path to an API that
validates the texture format at bind time.
6. Should it be legal to use TexImage* after TexStorage* if it doesn't
affect the metadata?
RESOLVED: No.
DISCUSSION: A potential use case is to allow a single level of a
texture to be invalidated using a NULL pointer. However, as noted
above it is non-trivial to determine what constitutes a change.
7. How does this extension interact with APPLE_texture_2D_limited_npot?
RESOLVED. APPLE_texture_2D_limited_npot is equivalent to the NPOT
support in OpenGL ES 2.0.
8. Should this extension be written to work with desktop OpenGL?
RESOLVED: Yes.
DISCUSSION: There has been been interest and it will future-proof it
against further additions to OpenGL ES.
9. Which texture targets should be supported?
RESOLVED. All targets except multisample and buffer textures are
supported.
Initially all targets except TEXTURE_BUFFER were supported. It was
noted that the entrypoints for multisample targets added no useful
functionality, since multisample textures have no completeness
checks beyond being non-empty.
Rectangle textures have completeness checks to prevent filtering of
integer textures. However, since we decided to only force mipmap
completeness, this becomes less useful.
10. Should this extension support proxy textures?
RESOLVED: Yes.
DISCUSSION: It should be orthogonal.
11. Are the <format> and <type> parameters necessary?
RESOLVED. No, they will be removed.
DISCUSSION: For OpenGL ES the type parameter was necessary to
determine the precision of the texture, but this can be solved by
having these functions accept sized internal formats (which are
already accepted by renderbuffers).
12. Should it be legal to make the default texture (id 0)
immutable-format?
RESOLVED: No.
DISCUSSION: This would make it impossible to restore the context to
it's default state, which is deemed undesirable. There is no good
reason not to use named texture objects.
13. Should we try to guarantee that textures made through this path
will always be complete?
RESOLVED: It should be guaranteed that the texture will be mipmap
complete.
DISCUSSION: Future separation between images and samplers will still
allow users to create combinations that are invalid, but
constraining the simple cases will make these APIs easier to use for
beginners.
14. Should these functions use a EXT_direct_state_access approach to
specifying the texture objects?
UNRESOLVED.
DISCUSSION: as a standalone extension, no DSA-like functions will be
added. However, interactions with EXT_direct_state_access and
ARB_direct_state_access need to be resolved.
15. Should these functions accept generic compressed formats?
RESOLVED: Yes. Note that the spec language will need to be modified
to allow this for ES, since the pseudocode is written in terms of
TexImage2D, which does not allow compressed texture formats in ES.
See also issues 23 and 27.
16. How should completeness be forced when TEXTURE_MAX_LEVEL is not
present?
RESOLVED. The maximum level q will be redefined to clamp to the
highest level available.
DISCUSSION: A single-level texture can be made complete either by
making it mipmap complete (by setting TEXTURE_MAX_LEVEL to 0) or by
turning off mipmapping (by choose an appropriate minification
filter).
Some options:
A: Specify that TexStorage* changes the default minification filter
for OpenGL ES. This makes it awkward to add TEXTURE_MAX_LEVEL
support to OpenGL ES later, since switching to match GL would break
compatibility. The two mechanisms also do not give identical
results, since the magnification threshold depends on the
minification filter.
B: Specify that the texture behaves as though TEXTURE_MAX_LEVEL were
zero. To specify this properly probably requires fairly intrusive
changes to the OpenGL ES full specification to add back all the
language relating to the max level. It also does not solve the
similar problem of what to do with NPOT textures; and it may have
hardware impacts due to the change in the min/mag crossover.
C: Specify that TexStorage* changes the default minification filter
for all implementations when a single-level texture is specified.
This may be slightly counter-intuitive to desktop GL users, but will
give consistent behaviour across variants of GL and avoids changing
the functional behaviour of this extension based on the presence or
absence of some other feature.
Currently B is specified. This has potential hardware implications
for OpenGL ES because of the effect of the minification filter on
the min/mag crossover. However, C has potential hardware implications
for OpenGL due to the separation of texture and sampler state.
17. How should completeness be forced when only ES2-style NPOT is
available?
RESOLVED. It is not worth trying to do this, in light of issue 13.
Previous revisions of this extension overrode the minification
filter and wrap modes, but that is no longer the case. Since
OES_texture_npot removes the caveats on NPOT textures anyway, it
might not be worth trying to "fix" this.
18. For OpenGL ES, how do the new sized internal formats interact
with OES_required_internal_format?
RESOLVED.
If OES_required_internal_format is not present, then the
<internalformat> parameter is intended merely to indicate what the
corresponding <format> and <type> would have been, had TexImage*
been used instead. If OES_required_internal_format is present, then
it is intended that the <internalformat> will be interpreted as if
it had been passed directly to TexImage*.
19. Should there be some hinting mechanism to indicate whether data
is coming immediately or later?
RESOLVED. No parameter is needed. An extension can be added to provide
a TexParameter value which is latched at TexStorage time.
DISCUSSION: Some members felt that this would be useful so that they
could defer allocation when suitable, particularly if higher-
resolution images will be streamed in later; or to choose a memory
type or layout appropriate to the usage. However, implementation
experience with BufferData is that developers frequently provide
wrong values and implementations have to guess anyway.
One option suggested was the <usage> parameter currently passed to
BufferData. Another option was to set it with TexParameter.
20. How should this extension interact with
EGLImageTargetTexture2DOES, eglBindTexImage, glXBindTexImage and
wglBindTexImage?
RESOLVED. These functions will not be permitted after glTexStorage*.
Several options are possible:
A) Disallow these functions.
B) Allow them, but have them reset the TEXTURE_IMMUTABLE_FORMAT
flag.
C) Allow them unconditionally.
C would violate the design principle that the dimensions and format
of the mipmap array are immutable. B does not so much modify the
dimension and formats as replace them with an entirely different
set.
21. Should there be a single function for specifying 1D, 2D and 3D
targets?
RESOLVED. No, we will stick with existing precedent.
22. Is it possible to use GenerateMipmap with an incomplete mipmap
pyramid?
RESOLVED. Yes, because the effective max level is limited to the
levels that were specified, and so GenerateMipmap does not generate
any new levels.
However, to make automatic mipmap generation work, it is necessary
to redefine p rather than q, since automatic mipmap generation
ignores the max level.
23. How should this extension interact with
OES_compressed_paletted_texture?
RESOLVED. Paletted textures will not be permitted, and will
generate INVALID_ENUM.
DISCUSSION: OES_compressed_paletted_texture supplies all the mipmaps
in a single function call, with the palette specified once. That's
incompatible with the upload model in this extension.
24. How can ETC1 textures be used with this extension?
RESOLVED. Add language in this extension to allow subregion uploads
for ETC1.
DISCUSSION: GL_OES_compressed_ETC1_RGB8_texture doesn't allow
CompressedTexSubImage*, so it would be impossible to use this
extension with ETC1. This is seen as an oversight in the ETC1
extension. While it cannot be fixed in that extension (since it is
already shipping), this extension can add that capability.
25. Should any other compressed formats be similarly modified?
RESOLVED. Yes, AMD_compressed_ATC_texture and
AMD_compressed_3DC_texture can be modified similarly to ETC1
(Maurice Ribble indicated that both formats use 4x4 blocks). Desktop
OpenGL requires that whole-image replacement is supported for any
compressed texture format, and the OpenGL ES extensions
EXT_texture_compression_dxt1 and IMG_texture_compression_pvrtc
already allow whole-image replacement, so it is not necessary to
modify them to be used with this extension.
26. Should these commands be permitted in display lists?
RESOLVED. No.
DISCUSSION: Display lists are most useful for repeating commands,
and TexStorage* commands cannot be repeated because the first call
makes the format immutable.
27. Should these commands accept unsized internal formats?
RESOLVED: No, for both OpenGL and OpenGL ES.
DISCUSSION: normally the <type> parameter to TexImage* can serve as
a hint to select a sized format (and in OpenGL ES, this is the only
mechanism available); since TexStorage* does not have a <type>
parameter, the implementation has no information on which to base a
decision.
Revision History
Revision 26, 2013/08/07 (mjk)
- Better indicate DSA entrypoints
Revision 25, 2012/01/19 (bmerry)
- Clarify that the pixel unpack buffer has no effect.
Revision 24, 2011/11/10 (bmerry)
- Updated my contact details
Revision 23, 2011/07/25 (bmerry)
- Remove dangling references to MultiTexStorage in Errors section
Revision 22, 2011/07/21 (bmerry)
- Remove dangling reference to <samples> in Errors section
Revision 21, 2011/06/13 (dgkoch)
- restore EXT suffixes on TextureStorage commands since they are
additions to EXT_direct_state_access
- restore EXT suffixes on Alpha/Luminance tokens
Revision 20, 2011/06/13 (dgkoch)
- make ARB extension from EXT_texture_storage (rev 19)
Revision 19, 2011/05/02 (Jon Leech)
- Assign enum value
Revision 18, 2011/01/24 (bmerry)
- Disallow unsized internal formats (oversight in revision 17).
Revision 17, 2011/01/24 (bmerry)
- Added and resolved issue 26.
- Split issue 27 out from issue 15.
- Disallow TexStorage* in display lists.
- Use the term "immutable-format" consistently (bug 7281).
Revision 16, 2010/11/23 (bmerry)
- Disallowed TexStorage on an immutable-format texture
(resolves issue 5).
- Deleted MultiTexStorage* commands (other DSA functions still
unresolved).
- Some minor wording changes suggested by Pat Brown (bug 7002).
Revision 15, 2010/11/09 (bmerry)
- Reopened issue 5.
- Reopened issue 14, pending stabilisation of
ARB_direct_state_access.
- Marked issue 9 resolved, pending any objections.
- Fix references to no object being bound (was meant to refer to
the default object).
- Adding missing pseudocode for TEXTURE_1D_ARRAY.
- Corrected TEXTURE_2D_ARRAY -> TEXTURE_1D_ARRAY in error checks.
- Changed "levels... are removed" to "levels... are reset to their
init state", since desktop GL has per-level state apart from the
texels.
- Miscellaneous wording fixes.
Revision 14, 2010/09/25 (bmerry)
- Add issues 24-25 and alterations to
OES_compressed_ETC1_RGB8_texture, AMD_compressed_ATC_texture and
AMD_compressed_3DC_texture.
Revision 13, 2010/09/19 (bmerry)
- Two typo fixes from Daniel Koch
Revision 12, 2010/09/18 (bmerry)
- Changed resolution to issue 20
- Added and resolved issue 23
- Added explanation of how to upload data (in overview)
- Added spec language to implement resolution to issue 15
Revision 11, 2010/07/21 (bmerry)
- Resolved issue 16
- Reopen issue 20
- Fix some typos
Revision 10, 2010/07/15 (bmerry)
- Update some issues to match core text
- Resolved issue 17
Revision 9, 2010/05/24 (bmerry)
- Marked issue 2 as resolved
- Resolved issue 19 (as no change)
- Resolved issue 20
- Add issues 21-22
- Add in spec language to forbid use on default textures
- Redefine level_base, level_max to be clamped forms of
TEXTURE_BASE_LEVEL/TEXTURE_MAX_LEVEL when using immutable
textures
- Redefine p to also be clamped to the provided levels for
immutable textures, to support automatic mipmap generation
- Removed multisample functions
- Removed language stating that texture parameters were reset to
defaults
Revision 8, 2010/05/18 (bmerry)
- Added issue about EGLimage
- Marked issue 14 as resolved
Revision 7, 2010/05/04 (bmerry)
- Removed some lingering <format>, <type> parameters to the new
functions that should have been removed in revision 4
- Trivial typo fixes
Revision 6, 2010/02/18 (bmerry)
- Resolved issues 5, 6 and 18
- Added MultiTexStorage* functions for DSA interaction
- Added error for texture-target mismatch in DSA
- Allowed TexStorage* to be called again
Revision 5, 2010/01/25 (bmerry)
- Added to contributors list
- Require OpenGL 1.2, to simplify interactions with
TEXTURE_BASE_LEVEL/TEXTURE_MAX_LEVEL and CLAMP_TO_EDGE
- Change default wrap modes to always be CLAMP_TO_EDGE
- Change default filters to always be NEAREST
- Moved language about generating new levels into an interaction,
since it can only happen on OpenGL ES
- Added interaction with EXT_direct_state_access
- Added extra <internalformats> for GL ES when OES_depth_texture,
OES_packed_depth_stencil and EXT_texture_type_2_10_10_10_REV are
present.
- Minor non-functional wording fixes and typos
- Resolved issue 16
- Added issues 17-19
Revision 4, 2010/01/13 (bmerry)
- Changed suffix from ARM to EXT
- Added list of contributors
- Added language to force the texture to always be complete
- Removed <format> and <type> arguments
- Added issues 14-16
- Reopened issue 2
- Reformatted issues to separate resolution and discussion
- Resolved issues 1, 9 and 11-13
- Fixed the max number of levels in a cube map array
Revision 3, 2009/12/17 (bmerry)
- Added missing vendor suffix to TEXTURE_IMMUTABLE_FORMAT_ARM
- Rewritten to against desktop OpenGL
- Added prototypes for 1D and multisample storage functions
- Added issues 8-13
Revision 2, 2009/08/20 (bmerry)
- Resolved issue 2 (no border parameter)
- Resolved issue 4 (metadata becomes immutable)
- Added interaction with OES_texture_cube_map
- Added error if width != height in a cube map
- Added issues 5-7
Revision 1, 2009/05/06 (bmerry)
- First draft