diff options
Diffstat (limited to 'src/gallium/drivers/i965simple/brw_curbe.c')
| -rw-r--r-- | src/gallium/drivers/i965simple/brw_curbe.c | 369 |
1 files changed, 369 insertions, 0 deletions
diff --git a/src/gallium/drivers/i965simple/brw_curbe.c b/src/gallium/drivers/i965simple/brw_curbe.c new file mode 100644 index 00000000000..824ee7fd6df --- /dev/null +++ b/src/gallium/drivers/i965simple/brw_curbe.c @@ -0,0 +1,369 @@ +/* + Copyright (C) Intel Corp. 2006. All Rights Reserved. + Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to + develop this 3D driver. + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + "Software"), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice (including the + next paragraph) shall be included in all copies or substantial + portions of the Software. + + THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. + IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE + LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION + OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION + WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + **********************************************************************/ + /* + * Authors: + * Keith Whitwell <[email protected]> + */ + + + +#include "brw_context.h" +#include "brw_defines.h" +#include "brw_state.h" +#include "brw_batch.h" +#include "brw_util.h" +#include "brw_wm.h" +#include "pipe/p_state.h" +#include "pipe/p_winsys.h" +#include "util/u_math.h" +#include "util/u_memory.h" + +#define FILE_DEBUG_FLAG DEBUG_FALLBACKS + +/* Partition the CURBE between the various users of constant values: + */ +static void calculate_curbe_offsets( struct brw_context *brw ) +{ + /* CACHE_NEW_WM_PROG */ + unsigned nr_fp_regs = align(brw->wm.prog_data->max_const, 16); + + /* BRW_NEW_VERTEX_PROGRAM */ + unsigned nr_vp_regs = align(brw->vs.prog_data->max_const, 16); + unsigned nr_clip_regs = 0; + unsigned total_regs; + +#if 0 + /* BRW_NEW_CLIP ? */ + if (brw->attribs.Transform->ClipPlanesEnabled) { + unsigned nr_planes = 6 + brw_count_bits(brw->attribs.Transform->ClipPlanesEnabled); + nr_clip_regs = align(nr_planes * 4, 16); + } +#endif + + + total_regs = nr_fp_regs + nr_vp_regs + nr_clip_regs; + + /* This can happen - what to do? Probably rather than falling + * back, the best thing to do is emit programs which code the + * constants as immediate values. Could do this either as a static + * cap on WM and VS, or adaptively. + * + * Unfortunately, this is currently dependent on the results of the + * program generation process (in the case of wm), so this would + * introduce the need to re-generate programs in the event of a + * curbe allocation failure. + */ + /* Max size is 32 - just large enough to + * hold the 128 parameters allowed by + * the fragment and vertex program + * api's. It's not clear what happens + * when both VP and FP want to use 128 + * parameters, though. + */ + assert(total_regs <= 32); + + /* Lazy resize: + */ + if (nr_fp_regs > brw->curbe.wm_size || + nr_vp_regs > brw->curbe.vs_size || + nr_clip_regs != brw->curbe.clip_size || + (total_regs < brw->curbe.total_size / 4 && + brw->curbe.total_size > 16)) { + + unsigned reg = 0; + + /* Calculate a new layout: + */ + reg = 0; + brw->curbe.wm_start = reg; + brw->curbe.wm_size = nr_fp_regs; reg += nr_fp_regs; + brw->curbe.clip_start = reg; + brw->curbe.clip_size = nr_clip_regs; reg += nr_clip_regs; + brw->curbe.vs_start = reg; + brw->curbe.vs_size = nr_vp_regs; reg += nr_vp_regs; + brw->curbe.total_size = reg; + +#if 0 + if (0) + DBG("curbe wm %d+%d clip %d+%d vs %d+%d\n", + brw->curbe.wm_start, + brw->curbe.wm_size, + brw->curbe.clip_start, + brw->curbe.clip_size, + brw->curbe.vs_start, + brw->curbe.vs_size ); +#endif + + brw->state.dirty.brw |= BRW_NEW_CURBE_OFFSETS; + } +} + + +const struct brw_tracked_state brw_curbe_offsets = { + .dirty = { + .brw = (BRW_NEW_CLIP | + BRW_NEW_VS), + .cache = CACHE_NEW_WM_PROG + }, + .update = calculate_curbe_offsets +}; + + + +/* Define the number of curbes within CS's urb allocation. Multiple + * urb entries -> multiple curbes. These will be used by + * fixed-function hardware in a double-buffering scheme to avoid a + * pipeline stall each time the contents of the curbe is changed. + */ +void brw_upload_constant_buffer_state(struct brw_context *brw) +{ + struct brw_constant_buffer_state cbs; + memset(&cbs, 0, sizeof(cbs)); + + /* It appears that this is the state packet for the CS unit, ie. the + * urb entries detailed here are housed in the CS range from the + * URB_FENCE command. + */ + cbs.header.opcode = CMD_CONST_BUFFER_STATE; + cbs.header.length = sizeof(cbs)/4 - 2; + + /* BRW_NEW_URB_FENCE */ + cbs.bits0.nr_urb_entries = brw->urb.nr_cs_entries; + cbs.bits0.urb_entry_size = brw->urb.csize - 1; + + assert(brw->urb.nr_cs_entries); + BRW_CACHED_BATCH_STRUCT(brw, &cbs); +} + + +static float fixed_plane[6][4] = { + { 0, 0, -1, 1 }, + { 0, 0, 1, 1 }, + { 0, -1, 0, 1 }, + { 0, 1, 0, 1 }, + {-1, 0, 0, 1 }, + { 1, 0, 0, 1 } +}; + +/* Upload a new set of constants. Too much variability to go into the + * cache mechanism, but maybe would benefit from a comparison against + * the current uploaded set of constants. + */ +static void upload_constant_buffer(struct brw_context *brw) +{ + struct brw_mem_pool *pool = &brw->pool[BRW_GS_POOL]; + unsigned sz = brw->curbe.total_size; + unsigned bufsz = sz * sizeof(float); + float *buf; + unsigned i; + + + if (sz == 0) { + struct brw_constant_buffer cb; + cb.header.opcode = CMD_CONST_BUFFER; + cb.header.length = sizeof(cb)/4 - 2; + cb.header.valid = 0; + cb.bits0.buffer_length = 0; + cb.bits0.buffer_address = 0; + BRW_BATCH_STRUCT(brw, &cb); + + if (brw->curbe.last_buf) { + free(brw->curbe.last_buf); + brw->curbe.last_buf = NULL; + brw->curbe.last_bufsz = 0; + } + + return; + } + + buf = (float *)malloc(bufsz); + + memset(buf, 0, bufsz); + + if (brw->curbe.wm_size) { + unsigned offset = brw->curbe.wm_start * 16; + + /* First the constant buffer constants: + */ + + /* Then any internally generated constants: + */ + for (i = 0; i < brw->wm.prog_data->nr_internal_consts; i++) + buf[offset + i] = brw->wm.prog_data->internal_const[i]; + + assert(brw->wm.prog_data->max_const == + brw->wm.prog_data->nr_internal_consts); + } + + + /* The clipplanes are actually delivered to both CLIP and VS units. + * VS uses them to calculate the outcode bitmasks. + */ + if (brw->curbe.clip_size) { + unsigned offset = brw->curbe.clip_start * 16; + unsigned j; + + /* If any planes are going this way, send them all this way: + */ + for (i = 0; i < 6; i++) { + buf[offset + i * 4 + 0] = fixed_plane[i][0]; + buf[offset + i * 4 + 1] = fixed_plane[i][1]; + buf[offset + i * 4 + 2] = fixed_plane[i][2]; + buf[offset + i * 4 + 3] = fixed_plane[i][3]; + } + + /* Clip planes: BRW_NEW_CLIP: + */ + for (j = 0; j < brw->attribs.Clip.nr; j++) { + buf[offset + i * 4 + 0] = brw->attribs.Clip.ucp[j][0]; + buf[offset + i * 4 + 1] = brw->attribs.Clip.ucp[j][1]; + buf[offset + i * 4 + 2] = brw->attribs.Clip.ucp[j][2]; + buf[offset + i * 4 + 3] = brw->attribs.Clip.ucp[j][3]; + i++; + } + } + + + if (brw->curbe.vs_size) { + unsigned offset = brw->curbe.vs_start * 16; + /*unsigned nr = vp->max_const;*/ + const struct pipe_constant_buffer *cbuffer = brw->attribs.Constants[0]; + struct pipe_winsys *ws = brw->pipe.winsys; + /* FIXME: buffer size is num_consts + num_immediates */ + if (brw->vs.prog_data->num_consts) { + /* map the vertex constant buffer and copy to curbe: */ + void *data = ws->buffer_map(ws, cbuffer->buffer, 0); + /* FIXME: this is wrong. the cbuffer->size currently + * represents size of consts + immediates. so if we'll + * have both we'll copy over the end of the buffer + * with the subsequent memcpy */ + memcpy(&buf[offset], data, cbuffer->size); + ws->buffer_unmap(ws, cbuffer->buffer); + offset += cbuffer->size; + } + /*immediates*/ + if (brw->vs.prog_data->num_imm) { + memcpy(&buf[offset], brw->vs.prog_data->imm_buf, + brw->vs.prog_data->num_imm * 4 * sizeof(float)); + } + } + + if (1) { + for (i = 0; i < sz; i+=4) + debug_printf("curbe %d.%d: %f %f %f %f\n", i/8, i&4, + buf[i+0], buf[i+1], buf[i+2], buf[i+3]); + + debug_printf("last_buf %p buf %p sz %d/%d cmp %d\n", + brw->curbe.last_buf, buf, + bufsz, brw->curbe.last_bufsz, + brw->curbe.last_buf ? memcmp(buf, brw->curbe.last_buf, bufsz) : -1); + } + + if (brw->curbe.last_buf && + bufsz == brw->curbe.last_bufsz && + memcmp(buf, brw->curbe.last_buf, bufsz) == 0) { + free(buf); +/* return; */ + } + else { + if (brw->curbe.last_buf) + free(brw->curbe.last_buf); + brw->curbe.last_buf = buf; + brw->curbe.last_bufsz = bufsz; + + + if (!brw_pool_alloc(pool, + bufsz, + 1 << 6, + &brw->curbe.gs_offset)) { + debug_printf("out of GS memory for curbe\n"); + assert(0); + return; + } + + + /* Copy data to the buffer: + */ + brw->winsys->buffer_subdata_typed(brw->winsys, + pool->buffer, + brw->curbe.gs_offset, + bufsz, + buf, + BRW_CONSTANT_BUFFER ); + } + + /* TODO: only emit the constant_buffer packet when necessary, ie: + - contents have changed + - offset has changed + - hw requirements due to other packets emitted. + */ + { + struct brw_constant_buffer cb; + + memset(&cb, 0, sizeof(cb)); + + cb.header.opcode = CMD_CONST_BUFFER; + cb.header.length = sizeof(cb)/4 - 2; + cb.header.valid = 1; + cb.bits0.buffer_length = sz - 1; + cb.bits0.buffer_address = brw->curbe.gs_offset >> 6; + + /* Because this provokes an action (ie copy the constants into the + * URB), it shouldn't be shortcircuited if identical to the + * previous time - because eg. the urb destination may have + * changed, or the urb contents different to last time. + * + * Note that the data referred to is actually copied internally, + * not just used in place according to passed pointer. + * + * It appears that the CS unit takes care of using each available + * URB entry (Const URB Entry == CURBE) in turn, and issuing + * flushes as necessary when doublebuffering of CURBEs isn't + * possible. + */ + BRW_BATCH_STRUCT(brw, &cb); + } +} + +/* This tracked state is unique in that the state it monitors varies + * dynamically depending on the parameters tracked by the fragment and + * vertex programs. This is the template used as a starting point, + * each context will maintain a copy of this internally and update as + * required. + */ +const struct brw_tracked_state brw_constant_buffer = { + .dirty = { + .brw = (BRW_NEW_CLIP | + BRW_NEW_CONSTANTS | + BRW_NEW_URB_FENCE | /* Implicit - hardware requires this, not used above */ + BRW_NEW_PSP | /* Implicit - hardware requires this, not used above */ + BRW_NEW_CURBE_OFFSETS), + .cache = (CACHE_NEW_WM_PROG) + }, + .update = upload_constant_buffer +}; + |