| Commit message (Expand) | Author | Age | Files | Lines |
* | vc4: Refactor qir_get_op_nsrc(enum qop) to qir_get_nsrc(struct qinst *). | Eric Anholt | 2016-11-29 | 1 | -1/+1 |
* | vc4: Re-add R4 to the "any" register class. | Eric Anholt | 2016-11-22 | 1 | -0/+2 |
* | vc4: Fix resource leak in register allocation failure path. | Mun Gwan-gyeong | 2016-11-16 | 1 | -0/+2 |
* | vc4: Use ra14/rb14 as the spilling registers. | Eric Anholt | 2016-11-12 | 1 | -2/+2 |
* | vc4: Add support for register allocation for threaded shaders. | Eric Anholt | 2016-11-12 | 1 | -16/+73 |
* | vc4: Split register class setup for physical files from accumulators. | Eric Anholt | 2016-11-12 | 1 | -17/+19 |
* | vc4: Use register allocator CLASS_BIT_R0_R3 to clean up CLASS_B. | Eric Anholt | 2016-11-12 | 1 | -4/+4 |
* | vc4: Fix register class handling of DDX/DDY arguments. | Eric Anholt | 2016-11-12 | 1 | -1/+1 |
* | vc4: Don't abort when a shader compile fails. | Eric Anholt | 2016-11-09 | 1 | -1/+2 |
* | vc4: Add register allocation support for MUL output rotation. | Eric Anholt | 2016-08-25 | 1 | -0/+13 |
* | vc4: Implement live intervals using a CFG. | Eric Anholt | 2016-07-12 | 1 | -38/+8 |
* | vc4: Add a "qir_for_each_inst_inorder" macro and use it in many places. | Eric Anholt | 2016-07-12 | 1 | -2/+2 |
* | vc4: Switch the unpack ops to being unpack flags on a mov. | Eric Anholt | 2015-10-26 | 1 | -9/+15 |
* | vc4: Fix up the test for whether the unpack can be from r4. | Eric Anholt | 2015-10-26 | 1 | -8/+2 |
* | vc4: Actually allow math results to allocate into r4. | Eric Anholt | 2015-08-21 | 1 | -1/+6 |
* | vc4: Fold the 16-bit integer pack into the instructions generating it. | Eric Anholt | 2015-08-21 | 1 | -6/+7 |
* | vc4: Allow unpack_8[abcd]_f's src to stay in r4. | Eric Anholt | 2015-08-20 | 1 | -1/+15 |
* | vc4: Switch QPU_PACK_SCALED to be two non-SSA instructions. | Eric Anholt | 2015-08-20 | 1 | -1/+2 |
* | vc4: Allow QIR registers to be non-SSA. | Eric Anholt | 2015-08-20 | 1 | -2/+3 |
* | util/ra: Make allocating conflict lists optional | Jason Ekstrand | 2015-08-18 | 1 | -1/+1 |
* | vc4: Make r4-writes implicitly move to a temp, and allocate temps to r4. | Eric Anholt | 2015-08-04 | 1 | -19/+64 |
* | vc4: Add better debug for register allocation failure. | Eric Anholt | 2015-07-14 | 1 | -1/+5 |
* | vc4: Convert from simple_list.h to list.h | Eric Anholt | 2015-05-29 | 1 | -7/+2 |
* | vc4: Move the tests for src needing to be an A register to vc4_qir.c. | Eric Anholt | 2015-01-15 | 1 | -17/+5 |
* | vc4: Avoid the save/restore of r3 for raddr conflicts, just use ra31. | Eric Anholt | 2015-01-11 | 1 | -2/+2 |
* | vc4: Add support for 16-bit signed/unsigned norm/scaled vertex attrs. | Eric Anholt | 2014-12-15 | 1 | -0/+4 |
* | vc4: Add support for 8-bit unnormalized vertex attrs. | Eric Anholt | 2014-12-15 | 1 | -0/+4 |
* | vc4: Rename UNPACK_8* to UNPACK_8*_F. | Eric Anholt | 2014-12-15 | 1 | -4/+4 |
* | vc4: Reserve rb31 instead of r3 for raddr conflict spills. | Eric Anholt | 2014-12-09 | 1 | -3/+3 |
* | vc4: Prioritize allocating accumulators to short-lived values. | Eric Anholt | 2014-12-09 | 1 | -14/+59 |
* | vc4: Interleave register allocation from regfile A and B. | Eric Anholt | 2014-12-08 | 1 | -39/+38 |
* | vc4: Put dead writes into the NOP register when generating code. | Eric Anholt | 2014-09-23 | 1 | -1/+8 |
* | vc4: Add support for 8-bit unorm/snorm vertex inputs. | Eric Anholt | 2014-09-23 | 1 | -0/+8 |
* | vc4: Switch to using Mesa's register allocator. | Eric Anholt | 2014-09-23 | 1 | -105/+105 |
* | vc4: Make a static list of all the registers. | Eric Anholt | 2014-09-23 | 1 | -12/+82 |
* | vc4: Use the same method as for FRAG_Z to handle fragcoord W. | Eric Anholt | 2014-09-19 | 1 | -0/+8 |
* | vc4: Fix memory leaks in register allocation. | Eric Anholt | 2014-09-16 | 1 | -0/+3 |
* | vc4: Move register allocation to a separate file. | Eric Anholt | 2014-09-16 | 1 | -0/+157 |