1/*
2 * Copyright 2011 Joakim Sindholt <opensource@zhasha.com>
3 *
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * on the rights to use, copy, modify, merge, publish, distribute, sub
8 * license, and/or sell copies of the Software, and to permit persons to whom
9 * the Software is furnished to do so, subject to the following conditions:
10 *
11 * The above copyright notice and this permission notice (including the next
12 * paragraph) shall be included in all copies or substantial portions of the
13 * Software.
14 *
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
18 * THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
19 * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
20 * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
21 * USE OR OTHER DEALINGS IN THE SOFTWARE. */
22
23#include "nine_helpers.h"
24#include "nine_shader.h"
25
26#include "vertexdeclaration9.h"
27#include "vertexshader9.h"
28
29#include "device9.h"
30#include "pipe/p_context.h"
31#include "cso_cache/cso_context.h"
32
33#define DBG_CHANNEL DBG_VERTEXSHADER
34
35HRESULT
36NineVertexShader9_ctor( struct NineVertexShader9 *This,
37                        struct NineUnknownParams *pParams,
38                        const DWORD *pFunction, void *cso )
39{
40    struct NineDevice9 *device;
41    struct nine_shader_info info;
42    struct pipe_context *pipe;
43    HRESULT hr;
44    unsigned i;
45
46    DBG("This=%p pParams=%p pFunction=%p cso=%p\n",
47        This, pParams, pFunction, cso);
48
49    hr = NineUnknown_ctor(&This->base, pParams);
50    if (FAILED(hr))
51        return hr;
52
53    if (cso) {
54        This->ff_cso = cso;
55        return D3D_OK;
56    }
57
58    device = This->base.device;
59
60    info.type = PIPE_SHADER_VERTEX;
61    info.byte_code = pFunction;
62    info.const_i_base = NINE_CONST_I_BASE(device->max_vs_const_f) / 16;
63    info.const_b_base = NINE_CONST_B_BASE(device->max_vs_const_f) / 16;
64    info.sampler_mask_shadow = 0x0;
65    info.sampler_ps1xtypes = 0x0;
66    info.fog_enable = 0;
67    info.point_size_min = 0;
68    info.point_size_max = 0;
69    info.add_constants_defs.c_combination = NULL;
70    info.add_constants_defs.int_const_added = NULL;
71    info.add_constants_defs.bool_const_added = NULL;
72    info.swvp_on = !!(device->params.BehaviorFlags & D3DCREATE_SOFTWARE_VERTEXPROCESSING);
73    info.process_vertices = false;
74
75    pipe = nine_context_get_pipe_acquire(device);
76    hr = nine_translate_shader(device, &info, pipe);
77    if (hr == D3DERR_INVALIDCALL &&
78        (device->params.BehaviorFlags & D3DCREATE_MIXED_VERTEXPROCESSING)) {
79        /* Retry with a swvp shader. It will require swvp to be on. */
80        info.swvp_on = true;
81        hr = nine_translate_shader(device, &info, pipe);
82    }
83    nine_context_get_pipe_release(device);
84    if (hr == D3DERR_INVALIDCALL)
85        ERR("Encountered buggy shader\n");
86    if (FAILED(hr))
87        return hr;
88    This->byte_code.version = info.version;
89    This->swvp_only = info.swvp_on;
90
91    This->byte_code.tokens = mem_dup(pFunction, info.byte_size);
92    if (!This->byte_code.tokens)
93        return E_OUTOFMEMORY;
94    This->byte_code.size = info.byte_size;
95
96    This->variant.cso = info.cso;
97    This->variant.const_ranges = info.const_ranges;
98    This->variant.const_used_size = info.const_used_size;
99    This->last_cso = info.cso;
100    This->last_const_ranges = info.const_ranges;
101    This->last_const_used_size = info.const_used_size;
102    This->last_key = (uint32_t) (info.swvp_on << 9);
103
104    This->lconstf = info.lconstf;
105    This->sampler_mask = info.sampler_mask;
106    This->position_t = info.position_t;
107    This->point_size = info.point_size;
108
109    memcpy(This->int_slots_used, info.int_slots_used, sizeof(This->int_slots_used));
110    memcpy(This->bool_slots_used, info.bool_slots_used, sizeof(This->bool_slots_used));
111
112    This->const_int_slots = info.const_int_slots;
113    This->const_bool_slots = info.const_bool_slots;
114
115    This->c_combinations = NULL;
116
117    for (i = 0; i < info.num_inputs && i < ARRAY_SIZE(This->input_map); ++i)
118        This->input_map[i].ndecl = info.input_map[i];
119    This->num_inputs = i;
120
121    return D3D_OK;
122}
123
124void
125NineVertexShader9_dtor( struct NineVertexShader9 *This )
126{
127    DBG("This=%p\n", This);
128
129    if (This->base.device) {
130        struct pipe_context *pipe = nine_context_get_pipe_multithread(This->base.device);
131        struct nine_shader_variant *var = &This->variant;
132        struct nine_shader_variant_so *var_so = &This->variant_so;
133
134        do {
135            if (var->cso) {
136                if (This->base.device->context.cso_shader.vs == var->cso)
137                    pipe->bind_vs_state(pipe, NULL);
138                pipe->delete_vs_state(pipe, var->cso);
139                FREE(var->const_ranges);
140            }
141            var = var->next;
142        } while (var);
143
144        while (var_so && var_so->vdecl) {
145            if (var_so->cso) {
146                cso_delete_vertex_shader(This->base.device->cso_sw, var_so->cso );
147            }
148            var_so = var_so->next;
149        }
150
151        if (This->ff_cso) {
152            if (This->ff_cso == This->base.device->context.cso_shader.vs)
153                pipe->bind_vs_state(pipe, NULL);
154            pipe->delete_vs_state(pipe, This->ff_cso);
155        }
156    }
157    nine_shader_variants_free(&This->variant);
158    nine_shader_variants_so_free(&This->variant_so);
159
160    nine_shader_constant_combination_free(This->c_combinations);
161
162    FREE((void *)This->byte_code.tokens); /* const_cast */
163
164    FREE(This->lconstf.data);
165    FREE(This->lconstf.ranges);
166
167    NineUnknown_dtor(&This->base);
168}
169
170HRESULT NINE_WINAPI
171NineVertexShader9_GetFunction( struct NineVertexShader9 *This,
172                               void *pData,
173                               UINT *pSizeOfData )
174{
175    user_assert(pSizeOfData, D3DERR_INVALIDCALL);
176
177    if (!pData) {
178        *pSizeOfData = This->byte_code.size;
179        return D3D_OK;
180    }
181    user_assert(*pSizeOfData >= This->byte_code.size, D3DERR_INVALIDCALL);
182
183    memcpy(pData, This->byte_code.tokens, This->byte_code.size);
184
185    return D3D_OK;
186}
187
188void *
189NineVertexShader9_GetVariant( struct NineVertexShader9 *This,
190                              unsigned **const_ranges,
191                              unsigned *const_used_size )
192{
193    /* GetVariant is called from nine_context, thus we can
194     * get pipe directly */
195    struct pipe_context *pipe = This->base.device->context.pipe;
196    void *cso;
197    uint64_t key;
198
199    key = This->next_key;
200    if (key == This->last_key) {
201        *const_ranges = This->last_const_ranges;
202        *const_used_size = This->last_const_used_size;
203        return This->last_cso;
204    }
205
206    cso = nine_shader_variant_get(&This->variant, const_ranges, const_used_size, key);
207    if (!cso) {
208        struct NineDevice9 *device = This->base.device;
209        struct nine_shader_info info;
210        HRESULT hr;
211
212        info.type = PIPE_SHADER_VERTEX;
213        info.const_i_base = NINE_CONST_I_BASE(device->max_vs_const_f) / 16;
214        info.const_b_base = NINE_CONST_B_BASE(device->max_vs_const_f) / 16;
215        info.byte_code = This->byte_code.tokens;
216        info.sampler_mask_shadow = key & 0xf;
217        info.fog_enable = device->context.rs[D3DRS_FOGENABLE];
218        info.point_size_min = asfloat(device->context.rs[D3DRS_POINTSIZE_MIN]);
219        info.point_size_max = asfloat(device->context.rs[D3DRS_POINTSIZE_MAX]);
220        info.add_constants_defs.c_combination =
221            nine_shader_constant_combination_get(This->c_combinations, (key >> 16) & 0xff);
222        info.add_constants_defs.int_const_added = &This->int_slots_used;
223        info.add_constants_defs.bool_const_added = &This->bool_slots_used;
224        info.swvp_on = device->context.swvp;
225        info.process_vertices = false;
226
227        hr = nine_translate_shader(This->base.device, &info, pipe);
228        if (FAILED(hr))
229            return NULL;
230        nine_shader_variant_add(&This->variant, key, info.cso,
231                                info.const_ranges, info.const_used_size);
232        cso = info.cso;
233        *const_ranges = info.const_ranges;
234        *const_used_size = info.const_used_size;
235    }
236
237    This->last_key = key;
238    This->last_cso = cso;
239    This->last_const_ranges = *const_ranges;
240    This->last_const_used_size = *const_used_size;
241
242    return cso;
243}
244
245void *
246NineVertexShader9_GetVariantProcessVertices( struct NineVertexShader9 *This,
247                                             struct NineVertexDeclaration9 *vdecl_out,
248                                             struct pipe_stream_output_info *so )
249{
250    struct nine_shader_info info;
251    HRESULT hr;
252    void *cso;
253
254    cso = nine_shader_variant_so_get(&This->variant_so, vdecl_out, so);
255    if (cso)
256        return cso;
257
258    info.type = PIPE_SHADER_VERTEX;
259    info.const_i_base = 0;
260    info.const_b_base = 0;
261    info.byte_code = This->byte_code.tokens;
262    info.sampler_mask_shadow = 0;
263    info.fog_enable = false;
264    info.point_size_min = 0;
265    info.point_size_max = 0;
266    info.add_constants_defs.c_combination = NULL;
267    info.add_constants_defs.int_const_added = NULL;
268    info.add_constants_defs.bool_const_added = NULL;
269    info.swvp_on = true;
270    info.vdecl_out = vdecl_out;
271    info.process_vertices = true;
272    hr = nine_translate_shader(This->base.device, &info, This->base.device->pipe_sw);
273    if (FAILED(hr))
274        return NULL;
275    *so = info.so;
276    nine_shader_variant_so_add(&This->variant_so, vdecl_out, so, info.cso);
277    return info.cso;
278}
279
280IDirect3DVertexShader9Vtbl NineVertexShader9_vtable = {
281    (void *)NineUnknown_QueryInterface,
282    (void *)NineUnknown_AddRef,
283    (void *)NineUnknown_Release,
284    (void *)NineUnknown_GetDevice,
285    (void *)NineVertexShader9_GetFunction
286};
287
288static const GUID *NineVertexShader9_IIDs[] = {
289    &IID_IDirect3DVertexShader9,
290    &IID_IUnknown,
291    NULL
292};
293
294HRESULT
295NineVertexShader9_new( struct NineDevice9 *pDevice,
296                       struct NineVertexShader9 **ppOut,
297                       const DWORD *pFunction, void *cso )
298{
299    if (cso) {
300        NINE_DEVICE_CHILD_BIND_NEW(VertexShader9, ppOut, pDevice, pFunction, cso);
301    } else {
302        NINE_DEVICE_CHILD_NEW(VertexShader9, ppOut, pDevice, pFunction, cso);
303    }
304}
305