summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* iris: fix some SO overflow query bugs and tidy the code a bitKenneth Graunke2019-02-211-25/+33
|
* iris: add initial transform feedback overflow query paths (V3)Dave Airlie2019-02-211-2/+129
| | | | | v2: fix cpu overflow calc v3: use a struct
* iris: actually flush for storage imagesKenneth Graunke2019-02-213-0/+21
|
* iris: add an extra BT assert from Chris WilsonKenneth Graunke2019-02-211-0/+2
|
* iris: add assertions about binding table startsKenneth Graunke2019-02-211-4/+13
|
* iris: drop pull constant binding table entryKenneth Graunke2019-02-211-4/+0
| | | | nothing uses this
* iris: Use program's num textures not the state tracker's boundKenneth Graunke2019-02-211-1/+1
| | | | the state tracker might bind more textures than the program is using.
* iris: Enable precompilesKenneth Graunke2019-02-214-27/+157
|
* iris: rework program cache interfaceKenneth Graunke2019-02-213-154/+131
| | | | | | | This exposes iris_upload_shader() without having to bind it, which will be useful for precompiles. It also lets us examine the old programs and flag dirty bits at a higher level, rather than cramming all that knowledge into the cache layer.
* iris: Use wrappers for create_xs_state rather than a switch statementKenneth Graunke2019-02-211-46/+78
|
* iris: fix comment locationKenneth Graunke2019-02-211-5/+5
|
* iris: export iris_upload_shaderKenneth Graunke2019-02-212-1/+10
|
* iris: fix prototype warningKenneth Graunke2019-02-211-0/+1
|
* iris: Re-pin even if nothing is dirtyKenneth Graunke2019-02-211-5/+5
|
* iris: Flush for history at various momentsKenneth Graunke2019-02-213-14/+34
| | | | | | | | | | When we blit, transfer, or copy_resource to a buffer, we need to flush to ensure any stale data for that buffer is invalidated in the caches. bind_history will inform us which caches need to be flushed. Also, for any push constant buffers, we need to flag those dirty so that we re-emit 3DSTATE_CONSTANT_*, causing the data to be re-pushed.
* iris: add iris_flush_and_dirty_for_historyKenneth Graunke2019-02-212-0/+45
|
* iris: Track a binding history for buffer resourcesKenneth Graunke2019-02-213-3/+31
| | | | | This will let us know what caches to flush / state to dirty when altering the contents of a buffer.
* iris: drop long dead XXX commentKenneth Graunke2019-02-211-2/+0
|
* iris: Do the 48-bit vertex buffer address invalidation workaroundKenneth Graunke2019-02-213-8/+42
|
* iris: Fix VIEWPORT/LAYER in stream output infoKenneth Graunke2019-02-211-2/+34
| | | | | Fixes glsl-1.50-transform-feedback-builtins and ext_transform_feedback-builtin-varyings gl_PointSize
* iris: Fix buffer -> buffer copy_regionKenneth Graunke2019-02-211-15/+30
| | | | | | | Size can be too large for a surf, blorp_buffer_copy chops things up into segments we can actually handle Fixes map_buffer_range_test and copy_buffer_coherency
* iris: Lie about indirectsKenneth Graunke2019-02-211-4/+5
| | | | fixes interpolateAt tests
* iris: Enable ctx->Const.UseSTD430AsDefaultPackingKenneth Graunke2019-02-211-0/+1
| | | | hooray for obscurely named pipe caps with bizarre descriptions!
* iris: update commentKenneth Graunke2019-02-211-1/+3
|
* iris: RT flush for memorybarrier with texture bitKenneth Graunke2019-02-211-7/+1
| | | | | PIXEL_BUFFER_BARRIER_BIT turns into PIPE_BARRIER_TEXTURE and it ought to trigger an RT flush, according to brw_memory_barrier
* iris: PIPE_CONTROL workarounds for GPGPU modeKenneth Graunke2019-02-211-2/+1
|
* iris: Put batches in an arrayKenneth Graunke2019-02-2113-81/+77
| | | | We keep re-making this array all over the place
* iris: put render batch first in fence codeKenneth Graunke2019-02-211-2/+2
| | | | this shouldn't matter, but it will make the next refactor easier
* iris: flush the compute batch too if border pool is redoneKenneth Graunke2019-02-211-0/+2
|
* iris: leave a TODOKenneth Graunke2019-02-211-0/+4
|
* iris: Add fence support using drm_syncobjChris Wilson2019-02-215-36/+258
|
* iris: Add wait fences to properly sync between render/computeKenneth Graunke2019-02-211-0/+2
| | | | | | When flushing a batch due to a data dependency, we need to not only kick off the other batch's work, but stall our execution until it completes. Just wait on last_syncpt after flushing it.
* iris: Hang on to the last batch's sync-point, so we can wait on itKenneth Graunke2019-02-212-0/+9
|
* iris: Tag each submitted batch with a syncobjChris Wilson2019-02-216-0/+190
| | | | | | (adjusted by Ken to make the signalling sync object immediately on batch reset, rather than batch finish time. this will work better with deferred flushes...)
* iris: Drop vestiges of throttling codeKenneth Graunke2019-02-212-12/+1
|
* iris: Merge two walks of the exec_bos listChris Wilson2019-02-211-5/+2
|
* iris: replace vestiges of fence fds with newer exec_fence APIKenneth Graunke2019-02-212-34/+55
| | | | patch by me and Chris Wilson
* iris: Avoid synchronizing due to the workaround BOKenneth Graunke2019-02-211-0/+8
|
* iris: Avoid cross-batch synchronization on read/readsKenneth Graunke2019-02-211-6/+21
| | | | | This avoids flushing batches just because e.g. both are reading the same dynamic state streaming buffer, or shader assembly buffer.
* iris: Combine iris_use_pinned_bo and add_exec_boKenneth Graunke2019-02-211-35/+34
|
* iris: Use iris_use_pinned_bo rather than add_exec_bo directlyKenneth Graunke2019-02-211-1/+1
| | | | less special this way
* iris: Fix assigning the output handle for exporting for KMSChris Wilson2019-02-211-1/+2
| | | | Fixes gbm_bo_get_handle() used for KMS in glamor.
* iris: Tidy exporting the flink handleChris Wilson2019-02-211-9/+16
|
* iris: Fix SLMKenneth Graunke2019-02-212-3/+2
| | | | | Now that Jason has set up the L3 we can do this. Also, my assert was useless because we hadn't set up the field in the first place. Oops.
* iris: Don't set constant read lengths at upload timeJason Ekstrand2019-02-211-3/+0
| | | | They're set in derived_data as part of store_cs_state
* iris: Configure the L3$ on the compute contextJason Ekstrand2019-02-211-0/+24
|
* iris: properly pin stencil buffersKenneth Graunke2019-02-211-3/+8
|
* iris: Fix TCS/TES slot unificationKenneth Graunke2019-02-211-2/+2
| | | | | | TCS outputs, TES inputs...not TCS inputs Fixes some barrier tests
* iris: more todo notesKenneth Graunke2019-02-211-0/+1
|
* iris: scissored and mirrored blitsKenneth Graunke2019-02-211-10/+206
|