summaryrefslogtreecommitdiff
path: root/src/gallium/drivers/r600
AgeCommit message (Collapse)Author
2011-02-11r600g: get correct height alignmentDave Airlie
useful for s3tc
2011-02-11r600g: drop two unused | 0 that are actually in word4 anyways.Dave Airlie
these were NOPs anyways. Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-11r600g: handle 16/32 u/s norm formats properlyDave Airlie
add support for the 32-bit types, also fixup the export setting to handle types with channels > 11 bits properly Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-08r600g: fixup assertionMarek Olšák
2011-02-08r600g: add a faster implementation of transfer_inline_writeMarek Olšák
u_default_transfer_inline_write uses util_copy_rect, which is kinda slow.
2011-02-08r600g: slab-allocate buffer and transfer structuresMarek Olšák
2011-02-08r600g: use the same upload buffer for vertices, indices, and constantsMarek Olšák
This should reduce memory consumption.
2011-02-08u_vbuf_mgr: add a way to specify the BIND flag for the upload bufferMarek Olšák
2011-02-08u_vbuf_mgr: make the uploader publicMarek Olšák
2011-02-08Revert "r600g: do not flush the uploader" (with comments)Marek Olšák
This reverts commit 1c2a4f0820ff2272f993e6da28dcf8bcbbc3252a.
2011-02-07r600g: Add support for relative addressing on constant buffers.Henri Verbeet
Relative addressing of constant buffers can't work properly through the kcache, since you can only address within the currently locked kcache window. Instead, this patch binds the constant buffer as a shader resource, and then explicitly fetches the constant using a vertex fetch with fetch type VTX_FETCH_NO_INDEX_OFFSET from the shader. There's probably still some room for improvement, doing the fetch right before the instruction that needs the value may not be quite optimal for example.
2011-02-07r600g: Set the fetch type in r600_bc_vtx_build().Henri Verbeet
2011-02-07r600g: Handle the ADD_INT instruction in r600_bc_get_num_operands().Henri Verbeet
2011-02-07r600g: Generalize the pipe_add_vertex_attrib() functions.Henri Verbeet
This allows them to be used for VS or PS buffer resources as well.
2011-02-07r600g: Remove vs_resource and ps_resource from the pipe context.Henri Verbeet
These are practically unused, only the vs_resource array is being abused for fetch shader resources.
2011-02-07r600g: Split constants in r600_shader_from_tgsi().Henri Verbeet
2011-02-07r600g: Do the tgsi_full_src_register to r600_shader_src conversion in ↵Henri Verbeet
r600_shader_from_tgsi().
2011-02-07r600g: Split r600_bc_alu_src.Henri Verbeet
The r600_bc_alu_src structure is used in two different ways, as a vector and for the individual channels of that same vector. This is somewhat fragile, and probably confusing.
2011-02-07r600g: Store literal values in the r600_bc_alu_src structure.Henri Verbeet
This is much easier to work with, and allows use to get rid of some of the literal handling hacks.
2011-02-07r600g: tgsi_dst() can't fail.Henri Verbeet
2011-02-07r600g: tgsi_src() can't fail.Henri Verbeet
2011-02-07r600g: correctly report supported vertex formatsMarek Olšák
2011-02-07r600g: use the new vertex buffer managerMarek Olšák
2011-02-06r600g: do not flush the uploaderMarek Olšák
2011-02-06r600g: use surface format not underlying texture formatDave Airlie
This uses the surface format to set the CB up not the underlying texture format, since these can and do differ. Fixes piglit fbo-srgb.
2011-02-04r600g: don't set tile_type on evergreen.Dave Airlie
Since we never bind the actual DB to the CB/texture only the flushed one we don't need to track the tile type at the moment.
2011-02-04r600g: fix evergreen sampler view + depth interactionDave Airlie
2011-02-04r600g: get offset for correct texture when setting up CB.Dave Airlie
this fixes the mipmap tests with tiling forced on. Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-04r600g: avoid trying to flush the flushing texture.Dave Airlie
Since these textures still have the depth bit set. Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-04r600g: evergreen CB check for flushed textureDave Airlie
2011-02-04r600g: flushing texture needs all levels.Dave Airlie
For mipmap generation we need all levels in the flushing texture. Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-04r600g: remove unused variablesDave Airlie
2011-02-04r600g: add a flag to just create flushed texture without flushing.Dave Airlie
This just adds a flag to create the texture without doing any flushing to it. Flushing occurs in the draw function. This avoids unnecessary flushes when we end up rebinding a CB/DB/texture due to the blitter just restoring state. Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-04r600g: also check CB bindings for textures to depth flush.Dave Airlie
This checks the color buffer bindings to make sure there is something to flush. Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-04r600g: flush depth texture before a blit from it.Dave Airlie
If we are going to blit from a depth texture we need to flush it before we blit from it. Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-03r600g: Make some more things static.Henri Verbeet
2011-02-03r600g: Get rid of the unused r600_cf_vtx_tc() function.Henri Verbeet
2011-02-03r600g: Make the buffer and texture vbtls static const.Henri Verbeet
2011-02-03r600g: flush differences back to DB copy.Dave Airlie
2011-02-03r600g: fix depth hw resource copies.Dave Airlie
With the previous fixes we can now enabled hw depth copies Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-03r600g: only set depth bit for hw accessible depth buffers.Dave Airlie
If we get a sw accessible buffer like the S8 texture we end up doing depth tracking on it when there is no need since we won't ever bind it to the hardware. This leads to a sw fallback in the transfer destruction which leads to and endless recusion loop of fail in transfer destroy. Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-03r600g: rework dirty / depth texture tracking.Dave Airlie
this adds a flag to keep track of whether the depth texture structure is the flushed texture or not, so we can avoid doing flushes when we do a hw rendering from one to the other. it also renames flushed to dirty_db which tracks if the DB copy has been dirtied by being bound to the hw. Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-02-03r600g: set correct pitch/offset for depth textures in flushed state.Dave Airlie
This fixes zreaddraw in tiling mode
2011-02-02r600g: use burst exports in shadersChristian König
Join multiple exports into just one instruction instead of exporting each register separately.
2011-02-01r600g: just change tile type when buffer is set to depth.Dave Airlie
Not 100% sure on this one, but this is how it should work, the question is whether it will uncover other bugs elsewhere.
2011-02-01r600g: fix evergreen for depth decompress testDave Airlie
2011-02-01r600: only decompress depth when its tile type is wrong.Dave Airlie
If the tile type for the buffer is 1 then its been bound to the DB at some point, we need to decompress it, otherwise its only been bound as texture/cb so don't do anything. This fixes 5 piglit tests here on r600g.
2011-02-01r600g: start looking at evergreen tiling.Dave Airlie
this just adds the ioctl interface and sets the tile type and array mode in the correct place. This seems to bring eg 1D tiling to the same level, and issues as on r600. No idea how to address 2D yet.
2011-02-01r600g: Actually use the info from the flushed depth texture when creating a ↵Henri Verbeet
sampler view on a depth texture. R600/R700 was using incorrect tiling information from the (compressed) depth buffer. Evergreen worked anyway because tiling doesn't work.
2011-01-31r600g: fix invalid ref count handling in r600_set_constant_bufferChristian König
Only decrement ref count if r600_upload_const_buffer really changes the buffer.
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-17 01:18:46 +0000