Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | r600g: handle 16/32 u/s norm formats properly | Dave Airlie | 2011-02-11 | 7 | -12/+57 |
| | | | | | | | 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 <[email protected]> | ||||
* | r300g: prevent NULL pointer dereference in r300_buffer_create | Marek Olšák | 2011-02-11 | 1 | -3/+3 |
| | | | | | Should fix: https://bugs.freedesktop.org/show_bug.cgi?id=33185 | ||||
* | r300g: plug a memory leak in winsys | Marek Olšák | 2011-02-11 | 1 | -0/+1 |
| | |||||
* | r300g: remove unneeded code in winsys | Marek Olšák | 2011-02-11 | 1 | -6/+0 |
| | | | | We don't need the read/write flags. | ||||
* | r300g: fix warning | Marek Olšák | 2011-02-11 | 1 | -1/+1 |
| | |||||
* | r300g: import the last bits of libdrm and cleanup the whole thing | Marek Olšák | 2011-02-11 | 27 | -900/+737 |
| | | | | | | | | | | | | | | Based on Dave's branch. The majority of this commit is a cleanup, mainly renaming things. There wasn't much code to import, just ioctl calls. Also done: - implemented unsynchronized bo_map (important optimization!) - radeon_bo_is_referenced_by_cs is no longer a refcount hack - dropped the libdrm_radeon dependency I'm surprised that this has resulted in less code in the end. | ||||
* | u_upload_mgr: Use PIPE_TRANSFER_FLUSH_EXPLICIT. | José Fonseca | 2011-02-10 | 1 | -2/+9 |
| | | | | This can avoid DMAing the whole buffer when just a fraction was changed. | ||||
* | r300g: implement accelerated copy_region for compressed formats | Marek Olšák | 2011-02-10 | 1 | -0/+29 |
| | |||||
* | r300g: add a way to change texture properties arbitrarily | Marek Olšák | 2011-02-10 | 9 | -76/+104 |
| | | | | So that we can implement resource_copy on arbitrary data. | ||||
* | r300g: consolidate buffers and textures to r300_resource | Marek Olšák | 2011-02-10 | 18 | -381/+328 |
| | | | | Transfers and create/destroy are still handled separately. | ||||
* | r300g: simplify WRITE_RELOC API and cleanup | Marek Olšák | 2011-02-10 | 10 | -136/+126 |
| | |||||
* | u_blitter: let the driver check whether there's a recursion | Marek Olšák | 2011-02-10 | 2 | -6/+9 |
| | |||||
* | r300g: use format from pipe_surface instead of pipe_resource | Marek Olšák | 2011-02-10 | 2 | -29/+14 |
| | |||||
* | Revert "r300g: support sRGB colorbuffers" | Marek Olšák | 2011-02-10 | 1 | -20/+0 |
| | | | | | | | | This partially reverts commit 91eba2567eab9409d94efc3c1f07a4a3731d0047. Conflicts: src/gallium/drivers/r300/r300_blit.c | ||||
* | svga: Don't advertise pixel shader addr register support. | José Fonseca | 2011-02-09 | 1 | -3/+8 |
| | | | | It's not fully supported. | ||||
* | nvc0: serialize on PIPE_FLUSH_RENDER_CACHE as well | Christoph Bumiller | 2011-02-09 | 1 | -0/+4 |
| | | | | Effects were easily visible in piglit/fbo-generatemipmap-formats. | ||||
* | nvc0: fix stride of NVC0_3D_RT methods | Christoph Bumiller | 2011-02-09 | 1 | -8/+8 |
| | |||||
* | nvc0: correct storage type for 16 bit surface formats | Christoph Bumiller | 2011-02-09 | 1 | -2/+2 |
| | |||||
* | nvc0: make CSE work for ops with multiple results | Christoph Bumiller | 2011-02-09 | 1 | -13/+72 |
| | |||||
* | nvc0: replace branching with predicated insns where feasible | Christoph Bumiller | 2011-02-09 | 5 | -46/+123 |
| | |||||
* | nvc0: implement local memory load and store ops | Christoph Bumiller | 2011-02-09 | 3 | -12/+38 |
| | |||||
* | nvc0: make sure phi-ops really have one source per in-block | Christoph Bumiller | 2011-02-09 | 2 | -19/+40 |
| | |||||
* | nv50,nvc0: do not forget to apply sign mode to saved TGSI inputs | Christoph Bumiller | 2011-02-09 | 2 | -11/+3 |
| | |||||
* | nvc0: do not generate a backwards jump if a loop ends with BRK | Christoph Bumiller | 2011-02-09 | 1 | -2/+4 |
| | |||||
* | nvc0: store only one value per basic block for TGSI regs | Christoph Bumiller | 2011-02-09 | 1 | -1/+7 |
| | |||||
* | nv50,nvc0: fix condition code change when commuting SET sources | Christoph Bumiller | 2011-02-09 | 2 | -3/+3 |
| | |||||
* | nvc0: set basic block on manual instruction insertion | Christoph Bumiller | 2011-02-09 | 1 | -0/+2 |
| | |||||
* | nvc0: try to fix register conflicts for vector instructions | Christoph Bumiller | 2011-02-09 | 4 | -32/+43 |
| | | | | | | | Vector here means using multiple 32 bit regs which are forced to be consecutive in the register file. This still isn't quite nice. | ||||
* | nvc0: reset texture base address after read transfer | Christoph Bumiller | 2011-02-09 | 1 | -2/+4 |
| | |||||
* | nvc0: don't combine memory loads across block boundaries | Christoph Bumiller | 2011-02-09 | 1 | -0/+5 |
| | |||||
* | nvc0: detect no-op MIN/MAX, do CSE earlier to succeed more often | Christoph Bumiller | 2011-02-09 | 1 | -48/+79 |
| | |||||
* | gallium/docs: fix typo | Marek Olšák | 2011-02-09 | 1 | -1/+1 |
| | |||||
* | st/egl: Fix platform selection | Benjamin Franzke | 2011-02-08 | 1 | -0/+1 |
| | | | | | A break for case _EGL_PLATFORM_X11 is missing. introduced by: 381ea0d67a6d84a34d23571c49bbf4339ffda364 | ||||
* | r600g: fixup assertion | Marek Olšák | 2011-02-08 | 1 | -1/+1 |
| | |||||
* | r600g: add a faster implementation of transfer_inline_write | Marek Olšák | 2011-02-08 | 1 | -1/+25 |
| | | | | u_default_transfer_inline_write uses util_copy_rect, which is kinda slow. | ||||
* | r600g: slab-allocate buffer and transfer structures | Marek Olšák | 2011-02-08 | 3 | -10/+85 |
| | |||||
* | r300g: use the same upload buffer for vertices and indices | Marek Olšák | 2011-02-08 | 4 | -18/+7 |
| | |||||
* | r600g: use the same upload buffer for vertices, indices, and constants | Marek Olšák | 2011-02-08 | 4 | -28/+10 |
| | | | | This should reduce memory consumption. | ||||
* | u_vbuf_mgr: add a way to specify the BIND flag for the upload buffer | Marek Olšák | 2011-02-08 | 4 | -1/+5 |
| | |||||
* | u_vbuf_mgr: remove tabs | Marek Olšák | 2011-02-08 | 1 | -18/+18 |
| | |||||
* | u_vbuf_mgr: make the uploader public | Marek Olšák | 2011-02-08 | 3 | -18/+16 |
| | |||||
* | Revert "r600g: do not flush the uploader" (with comments) | Marek Olšák | 2011-02-08 | 1 | -0/+7 |
| | | | | This reverts commit 1c2a4f0820ff2272f993e6da28dcf8bcbbc3252a. | ||||
* | r600g: silence a few valgrind warnings | Brian Paterni | 2011-02-08 | 3 | -4/+4 |
| | |||||
* | r600g: Add support for relative addressing on constant buffers. | Henri Verbeet | 2011-02-07 | 6 | -22/+152 |
| | | | | | | | | | | 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. | ||||
* | r600g: Set the fetch type in r600_bc_vtx_build(). | Henri Verbeet | 2011-02-07 | 1 | -0/+1 |
| | |||||
* | r600g: Handle the ADD_INT instruction in r600_bc_get_num_operands(). | Henri Verbeet | 2011-02-07 | 1 | -0/+2 |
| | |||||
* | r600g: Generalize the pipe_add_vertex_attrib() functions. | Henri Verbeet | 2011-02-07 | 4 | -28/+20 |
| | | | | This allows them to be used for VS or PS buffer resources as well. | ||||
* | r600g: Remove vs_resource and ps_resource from the pipe context. | Henri Verbeet | 2011-02-07 | 3 | -22/+6 |
| | | | | | These are practically unused, only the vs_resource array is being abused for fetch shader resources. | ||||
* | r600g: Split constants in r600_shader_from_tgsi(). | Henri Verbeet | 2011-02-07 | 1 | -140/+80 |
| | |||||
* | r600g: Do the tgsi_full_src_register to r600_shader_src conversion in ↵ | Henri Verbeet | 2011-02-07 | 1 | -176/+129 |
| | | | | r600_shader_from_tgsi(). |