summaryrefslogtreecommitdiff
path: root/src/mesa/drivers/dri/r300/radeon_span.c
blob: 16f9fb99e6757e66bb230ed81562a29e5cbb2b56 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
/**************************************************************************

Copyright (C) The Weather Channel, Inc.  2002.  All Rights Reserved.
Copyright 2000, 2001 ATI Technologies Inc., Ontario, Canada, and
                     VA Linux Systems Inc., Fremont, California.

The Weather Channel (TM) funded Tungsten Graphics to develop the
initial release of the Radeon 8500 driver under the XFree86 license.
This notice must be preserved.

All Rights Reserved.

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:
 *   Kevin E. Martin <martin@valinux.com>
 *   Gareth Hughes <gareth@valinux.com>
 *   Keith Whitwell <keith@tungstengraphics.com>
 *
 */

#include "main/glheader.h"
#include "swrast/swrast.h"

#include "r300_state.h"
#include "radeon_ioctl.h"
#include "r300_ioctl.h"
#include "radeon_span.h"

#include "drirenderbuffer.h"

#define DBG 0

/*
 * Note that all information needed to access pixels in a renderbuffer
 * should be obtained through the gl_renderbuffer parameter, not per-context
 * information.
 */
#define LOCAL_VARS						\
   driRenderbuffer *drb = (driRenderbuffer *) rb;		\
   const __DRIdrawablePrivate *dPriv = drb->dPriv;		\
   const GLuint bottom = dPriv->h - 1;				\
   GLubyte *buf = (GLubyte *) drb->flippedData			\
      + (dPriv->y * drb->flippedPitch + dPriv->x) * drb->cpp;	\
   GLuint p;							\
   (void) p;

#define LOCAL_DEPTH_VARS				\
   driRenderbuffer *drb = (driRenderbuffer *) rb;	\
   const __DRIdrawablePrivate *dPriv = drb->dPriv;	\
   const GLuint bottom = dPriv->h - 1;			\
   GLuint xo = dPriv->x;				\
   GLuint yo = dPriv->y;				\
   GLubyte *buf = (GLubyte *) drb->Base.Data;

#define LOCAL_STENCIL_VARS LOCAL_DEPTH_VARS

#define Y_FLIP(Y) (bottom - (Y))

#define HW_LOCK()

#define HW_UNLOCK()

/* ================================================================
 * Color buffer
 */

/* 16 bit, RGB565 color spanline and pixel functions
 */
#define SPANTMP_PIXEL_FMT GL_RGB
#define SPANTMP_PIXEL_TYPE GL_UNSIGNED_SHORT_5_6_5

#define TAG(x)    radeon##x##_RGB565
#define TAG2(x,y) radeon##x##_RGB565##y
#define GET_PTR(X,Y) (buf + ((Y) * drb->flippedPitch + (X)) * 2)
#include "spantmp2.h"

/* 32 bit, ARGB8888 color spanline and pixel functions
 */
#define SPANTMP_PIXEL_FMT GL_BGRA
#define SPANTMP_PIXEL_TYPE GL_UNSIGNED_INT_8_8_8_8_REV

#define TAG(x)    radeon##x##_ARGB8888
#define TAG2(x,y) radeon##x##_ARGB8888##y
#define GET_PTR(X,Y) (buf + ((Y) * drb->flippedPitch + (X)) * 4)
#include "spantmp2.h"

/* ================================================================
 * Depth buffer
 */

/* The Radeon family has depth tiling on all the time, so we have to convert
 * the x,y coordinates into the memory bus address (mba) in the same
 * manner as the engine.  In each case, the linear block address (ba)
 * is calculated, and then wired with x and y to produce the final
 * memory address.
 * The chip will do address translation on its own if the surface registers
 * are set up correctly. It is not quite enough to get it working with hyperz
 * too...
 */

static GLuint radeon_mba_z32(const driRenderbuffer * drb, GLint x, GLint y)
{
	GLuint pitch = drb->pitch;
	if (drb->depthHasSurface) {
		return 4 * (x + y * pitch);
	} else {
		GLuint ba, address = 0;	/* a[0..1] = 0           */

#ifdef COMPILE_R300
		ba = (y / 8) * (pitch / 8) + (x / 8);
#else
		ba = (y / 16) * (pitch / 16) + (x / 16);
#endif

		address |= (x & 0x7) << 2;	/* a[2..4] = x[0..2]     */
		address |= (y & 0x3) << 5;	/* a[5..6] = y[0..1]     */
		address |= (((x & 0x10) >> 2) ^ (y & 0x4)) << 5;	/* a[7]    = x[4] ^ y[2] */
		address |= (ba & 0x3) << 8;	/* a[8..9] = ba[0..1]    */

		address |= (y & 0x8) << 7;	/* a[10]   = y[3]        */
		address |= (((x & 0x8) << 1) ^ (y & 0x10)) << 7;	/* a[11]   = x[3] ^ y[4] */
		address |= (ba & ~0x3) << 10;	/* a[12..] = ba[2..]     */

		return address;
	}
}

static INLINE GLuint
radeon_mba_z16(const driRenderbuffer * drb, GLint x, GLint y)
{
	GLuint pitch = drb->pitch;
	if (drb->depthHasSurface) {
		return 2 * (x + y * pitch);
	} else {
		GLuint ba, address = 0;	/* a[0]    = 0           */

		ba = (y / 16) * (pitch / 32) + (x / 32);

		address |= (x & 0x7) << 1;	/* a[1..3] = x[0..2]     */
		address |= (y & 0x7) << 4;	/* a[4..6] = y[0..2]     */
		address |= (x & 0x8) << 4;	/* a[7]    = x[3]        */
		address |= (ba & 0x3) << 8;	/* a[8..9] = ba[0..1]    */
		address |= (y & 0x8) << 7;	/* a[10]   = y[3]        */
		address |= ((x & 0x10) ^ (y & 0x10)) << 7;	/* a[11]   = x[4] ^ y[4] */
		address |= (ba & ~0x3) << 10;	/* a[12..] = ba[2..]     */

		return address;
	}
}

/* 16-bit depth buffer functions
 */
#define VALUE_TYPE GLushort

#define WRITE_DEPTH( _x, _y, d )					\
   *(GLushort *)(buf + radeon_mba_z16( drb, _x + xo, _y + yo )) = d;

#define READ_DEPTH( d, _x, _y )						\
   d = *(GLushort *)(buf + radeon_mba_z16( drb, _x + xo, _y + yo ));

#define TAG(x) radeon##x##_z16
#include "depthtmp.h"

/* 24 bit depth, 8 bit stencil depthbuffer functions
 *
 * Careful: It looks like the R300 uses ZZZS byte order while the R200
 * uses SZZZ for 24 bit depth, 8 bit stencil mode.
 */
#define VALUE_TYPE GLuint

#ifdef COMPILE_R300
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint offset = radeon_mba_z32( drb, _x + xo, _y + yo );		\
   GLuint tmp = *(GLuint *)(buf + offset);				\
   tmp &= 0x000000ff;							\
   tmp |= ((d << 8) & 0xffffff00);					\
   *(GLuint *)(buf + offset) = tmp;					\
} while (0)
#else
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint offset = radeon_mba_z32( drb, _x + xo, _y + yo );		\
   GLuint tmp = *(GLuint *)(buf + offset);				\
   tmp &= 0xff000000;							\
   tmp |= ((d) & 0x00ffffff);						\
   *(GLuint *)(buf + offset) = tmp;					\
} while (0)
#endif

#ifdef COMPILE_R300
#define READ_DEPTH( d, _x, _y )						\
  do { \
    d = (*(GLuint *)(buf + radeon_mba_z32( drb, _x + xo,		\
					 _y + yo )) & 0xffffff00) >> 8; \
  }while(0)
#else
#define READ_DEPTH( d, _x, _y )						\
   d = *(GLuint *)(buf + radeon_mba_z32( drb, _x + xo,			\
					 _y + yo )) & 0x00ffffff;
#endif

#define TAG(x) radeon##x##_z24_s8
#include "depthtmp.h"

/* ================================================================
 * Stencil buffer
 */

/* 24 bit depth, 8 bit stencil depthbuffer functions
 */
#ifdef COMPILE_R300
#define WRITE_STENCIL( _x, _y, d )					\
do {									\
   GLuint offset = radeon_mba_z32( drb, _x + xo, _y + yo );		\
   GLuint tmp = *(GLuint *)(buf + offset);				\
   tmp &= 0xffffff00;							\
   tmp |= (d) & 0xff;							\
   *(GLuint *)(buf + offset) = tmp;					\
} while (0)
#else
#define WRITE_STENCIL( _x, _y, d )					\
do {									\
   GLuint offset = radeon_mba_z32( drb, _x + xo, _y + yo );		\
   GLuint tmp = *(GLuint *)(buf + offset);				\
   tmp &= 0x00ffffff;							\
   tmp |= (((d) & 0xff) << 24);						\
   *(GLuint *)(buf + offset) = tmp;					\
} while (0)
#endif

#ifdef COMPILE_R300
#define READ_STENCIL( d, _x, _y )					\
do {									\
   GLuint offset = radeon_mba_z32( drb, _x + xo, _y + yo );		\
   GLuint tmp = *(GLuint *)(buf + offset);				\
   d = tmp & 0x000000ff;						\
} while (0)
#else
#define READ_STENCIL( d, _x, _y )					\
do {									\
   GLuint offset = radeon_mba_z32( drb, _x + xo, _y + yo );		\
   GLuint tmp = *(GLuint *)(buf + offset);				\
   d = (tmp & 0xff000000) >> 24;					\
} while (0)
#endif

#define TAG(x) radeon##x##_z24_s8
#include "stenciltmp.h"

/* Move locking out to get reasonable span performance (10x better
 * than doing this in HW_LOCK above).  WaitForIdle() is the main
 * culprit.
 */

static void radeonSpanRenderStart(GLcontext * ctx)
{
	radeonContextPtr rmesa = RADEON_CONTEXT(ctx);
#ifdef COMPILE_R300
	r300ContextPtr r300 = (r300ContextPtr) rmesa;
	R300_FIREVERTICES(r300);
#else
	RADEON_FIREVERTICES(rmesa);
#endif
	LOCK_HARDWARE(rmesa);
	radeonWaitForIdleLocked(rmesa);

	/* Read the first pixel in the frame buffer.  This should
	 * be a noop, right?  In fact without this conform fails as reading
	 * from the framebuffer sometimes produces old results -- the
	 * on-card read cache gets mixed up and doesn't notice that the
	 * framebuffer has been updated.
	 *
	 * Note that we should probably be reading some otherwise unused
	 * region of VRAM, otherwise we might get incorrect results when
	 * reading pixels from the top left of the screen.
	 *
	 * I found this problem on an R420 with glean's texCube test.
	 * Note that the R200 span code also *writes* the first pixel in the
	 * framebuffer, but I've found this to be unnecessary.
	 *  -- Nicolai Hähnle, June 2008
	 */
	{
		int p;
		driRenderbuffer *drb =
			(driRenderbuffer *) ctx->WinSysDrawBuffer->_ColorDrawBuffers[0];
		volatile int *buf =
			(volatile int *)(rmesa->dri.screen->pFB + drb->offset);
		p = *buf;
	}
}

static void radeonSpanRenderFinish(GLcontext * ctx)
{
	radeonContextPtr rmesa = RADEON_CONTEXT(ctx);
	_swrast_flush(ctx);
	UNLOCK_HARDWARE(rmesa);
}

void radeonInitSpanFuncs(GLcontext * ctx)
{
	struct swrast_device_driver *swdd =
	    _swrast_GetDeviceDriverReference(ctx);
	swdd->SpanRenderStart = radeonSpanRenderStart;
	swdd->SpanRenderFinish = radeonSpanRenderFinish;
}

/**
 * Plug in the Get/Put routines for the given driRenderbuffer.
 */
void radeonSetSpanFunctions(driRenderbuffer * drb, const GLvisual * vis)
{
	if (drb->Base.InternalFormat == GL_RGBA) {
		if (vis->redBits == 5 && vis->greenBits == 6
		    && vis->blueBits == 5) {
			radeonInitPointers_RGB565(&drb->Base);
		} else {
			radeonInitPointers_ARGB8888(&drb->Base);
		}
	} else if (drb->Base.InternalFormat == GL_DEPTH_COMPONENT16) {
		radeonInitDepthPointers_z16(&drb->Base);
	} else if (drb->Base.InternalFormat == GL_DEPTH_COMPONENT24) {
		radeonInitDepthPointers_z24_s8(&drb->Base);
	} else if (drb->Base.InternalFormat == GL_STENCIL_INDEX8_EXT) {
		radeonInitStencilPointers_z24_s8(&drb->Base);
	}
}