FreeBE/nvidia/riva_hw.c

/*
 *    This file was taken from the XFree86 distribution.
 *    Minor modifications by Shawn Hargreaves are marked slh:
 */

/***************************************************************************\
|*                                                                           *|
|*       Copyright 1993-1998 NVIDIA, Corporation.  All rights reserved.      *|
|*                                                                           *|
|*     NOTICE TO USER:   The source code  is copyrighted under  U.S. and     *|
|*     international laws.  Users and possessors of this source code are     *|
|*     hereby granted a nonexclusive,  royalty-free copyright license to     *|
|*     use this code in individual and commercial software.                  *|
|*                                                                           *|
|*     Any use of this source code must include,  in the user documenta-     *|
|*     tion and  internal comments to the code,  notices to the end user     *|
|*     as follows:                                                           *|
|*                                                                           *|
|*       Copyright 1993-1998 NVIDIA, Corporation.  All rights reserved.      *|
|*                                                                           *|
|*     NVIDIA, CORPORATION MAKES NO REPRESENTATION ABOUT THE SUITABILITY     *|
|*     OF  THIS SOURCE  CODE  FOR ANY PURPOSE.  IT IS  PROVIDED  "AS IS"     *|
|*     WITHOUT EXPRESS OR IMPLIED WARRANTY OF ANY KIND.  NVIDIA, CORPOR-     *|
|*     ATION DISCLAIMS ALL WARRANTIES  WITH REGARD  TO THIS SOURCE CODE,     *|
|*     INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY, NONINFRINGE-     *|
|*     MENT,  AND FITNESS  FOR A PARTICULAR PURPOSE.   IN NO EVENT SHALL     *|
|*     NVIDIA, CORPORATION  BE LIABLE FOR ANY SPECIAL,  INDIRECT,  INCI-     *|
|*     DENTAL, OR CONSEQUENTIAL DAMAGES,  OR ANY DAMAGES  WHATSOEVER RE-     *|
|*     SULTING FROM LOSS OF USE,  DATA OR PROFITS,  WHETHER IN AN ACTION     *|
|*     OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,  ARISING OUT OF     *|
|*     OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOURCE CODE.     *|
|*                                                                           *|
|*     U.S. Government  End  Users.   This source code  is a "commercial     *|
|*     item,"  as that  term is  defined at  48 C.F.R. 2.101 (OCT 1995),     *|
|*     consisting  of "commercial  computer  software"  and  "commercial     *|
|*     computer  software  documentation,"  as such  terms  are  used in     *|
|*     48 C.F.R. 12.212 (SEPT 1995)  and is provided to the U.S. Govern-     *|
|*     ment only as  a commercial end item.   Consistent with  48 C.F.R.     *|
|*     12.212 and  48 C.F.R. 227.7202-1 through  227.7202-4 (JUNE 1995),     *|
|*     all U.S. Government End Users  acquire the source code  with only     *|
|*     those rights set forth herein.                                        *|
|*                                                                           *|
\***************************************************************************/
/* $XFree86:
 * xc/programs/Xserver/hw/xfree86/vga256/drivers/nv/riva_hw.c,v 1.1.2.3
 * 1998/12/26 00:12:39 dawes Exp $ */

#include <pc.h>

#define inb inportb
#define outb outportb

#include "riva_hw.h"
#include "riva_tbl.h"

/*
 * This file is an OS-agnostic file used to make RIVA 128 and RIVA TNT
 * operate identically (except TNT has more memory and better 3D quality.
 */

static int nv3Busy(RIVA_HW_INST *chip) {
	return ((!(chip->PFIFO[0x00001214 / 4] & 0x10)) |
	        (chip->PGRAPH[0x000006B0 / 4] & 0x01));
}
static int nv4Busy(RIVA_HW_INST *chip) {
	return ((!(chip->PFIFO[0x00001214 / 4] & 0x10)) |
	        (chip->PGRAPH[0x00000700 / 4] & 0x01));
}
static int ShowHideCursor(RIVA_HW_INST *chip, int ShowHide) {
	int current;
	current = chip->CurrentState->cursor1;
	chip->CurrentState->cursor1 =
	    (chip->CurrentState->cursor1 & 0xFE) | (ShowHide & 0x01);
	outb(0x3D4, 0x31);
	outb(0x3D5, chip->CurrentState->cursor1);
	return (current & 0x01);
}

/****************************************************************************\
*                                                                            *
* The video arbitration routines calculate some "magic" numbers.  Fixes      *
* the snow seen when accessing the framebuffer without it.                   *
* It just works (I hope).                                                    *
*                                                                            *
\****************************************************************************/

#define DEFAULT_GR_LWM 100
#define DEFAULT_VID_LWM 100
#define DEFAULT_GR_BURST_SIZE 256
#define DEFAULT_VID_BURST_SIZE 128
#define VIDEO 0
#define GRAPHICS 1
#define MPORT 2
#define ENGINE 3
#define GFIFO_SIZE 320
#define GFIFO_SIZE_128 256
#define MFIFO_SIZE 120
#define VFIFO_SIZE 256
#define ABS(a) (a > 0 ? a : -a)
typedef struct {
	int gdrain_rate;
	int vdrain_rate;
	int mdrain_rate;
	int gburst_size;
	int vburst_size;
	char vid_en;
	char gr_en;
	int wcmocc, wcgocc, wcvocc, wcvlwm, wcglwm;
	int by_gfacc;
	char vid_only_once;
	char gr_only_once;
	char first_vacc;
	char first_gacc;
	char first_macc;
	int vocc;
	int gocc;
	int mocc;
	char cur;
	char engine_en;
	char converged;
	int priority;
} nv3_arb_info;
typedef struct {
	int graphics_lwm;
	int video_lwm;
	int graphics_burst_size;
	int video_burst_size;
	int graphics_hi_priority;
	int media_hi_priority;
	int rtl_values;
	int valid;
} nv3_fifo_info;
typedef struct {
	char pix_bpp;
	char enable_video;
	char gr_during_vid;
	char enable_mp;
	int memory_width;
	int video_scale;
	int pclk_khz;
	int mclk_khz;
	int mem_page_miss;
	int mem_latency;
	char mem_aligned;
} nv3_sim_state;
typedef struct {
	int graphics_lwm;
	int video_lwm;
	int graphics_burst_size;
	int video_burst_size;
	int valid;
} nv4_fifo_info;
typedef struct {
	int pclk_khz;
	int mclk_khz;
	int nvclk_khz;
	char mem_page_miss;
	char mem_latency;
	int memory_width;
	char enable_video;
	char gr_during_vid;
	char pix_bpp;
	char mem_aligned;
	char enable_mp;
} nv4_sim_state;
static int nv3_iterate(nv3_fifo_info *res_info, nv3_sim_state *state,
                       nv3_arb_info *ainfo) {
	int iter = 0;
	int tmp;
	int vfsize, mfsize, gfsize;
	int mburst_size = 32;
	int mmisses, gmisses, vmisses;
	int misses;
	int vlwm, glwm, mlwm;
	int last, next, cur;
	int max_gfsize;
	long ns;

	vlwm = 0;
	glwm = 0;
	mlwm = 0;
	vfsize = 0;
	gfsize = 0;
	cur = ainfo->cur;
	mmisses = 2;
	gmisses = 2;
	vmisses = 2;
	if (ainfo->gburst_size == 128)
		max_gfsize = GFIFO_SIZE_128;
	else
		max_gfsize = GFIFO_SIZE;
	max_gfsize = GFIFO_SIZE;
	while (1) {
		if (ainfo->vid_en) {
			if (ainfo->wcvocc > ainfo->vocc) ainfo->wcvocc = ainfo->vocc;
			if (ainfo->wcvlwm > vlwm) ainfo->wcvlwm = vlwm;
			ns = 1000000 * ainfo->vburst_size / (state->memory_width / 8) /
			     state->mclk_khz;
			vfsize = ns * ainfo->vdrain_rate / 1000000;
			vfsize = ainfo->wcvlwm - ainfo->vburst_size + vfsize;
		}
		if (state->enable_mp) {
			if (ainfo->wcmocc > ainfo->mocc) ainfo->wcmocc = ainfo->mocc;
		}
		if (ainfo->gr_en) {
			if (ainfo->wcglwm > glwm) ainfo->wcglwm = glwm;
			if (ainfo->wcgocc > ainfo->gocc) ainfo->wcgocc = ainfo->gocc;
			ns = 1000000 * (ainfo->gburst_size / (state->memory_width / 8)) /
			     state->mclk_khz;
			gfsize = ns * ainfo->gdrain_rate / 1000000;
			gfsize = ainfo->wcglwm - ainfo->gburst_size + gfsize;
		}
		mfsize = 0;
		if (!state->gr_during_vid && ainfo->vid_en) {
			if (ainfo->vid_en && (ainfo->vocc < 0) && !ainfo->vid_only_once)
				next = VIDEO;
			else if (ainfo->mocc < 0)
				next = MPORT;
			else if (ainfo->gocc < ainfo->by_gfacc)
				next = GRAPHICS;
			else
				return (0);
		} else
			switch (ainfo->priority) {
				case VIDEO:
					if (ainfo->vid_en && ainfo->vocc < 0 &&
					    !ainfo->vid_only_once)
						next = VIDEO;
					else if (ainfo->gr_en && ainfo->gocc < 0 &&
					         !ainfo->gr_only_once)
						next = GRAPHICS;
					else if (ainfo->mocc < 0)
						next = MPORT;
					else
						return (0);
					break;
				case GRAPHICS:
					if (ainfo->gr_en && ainfo->gocc < 0 && !ainfo->gr_only_once)
						next = GRAPHICS;
					else if (ainfo->vid_en && ainfo->vocc < 0 &&
					         !ainfo->vid_only_once)
						next = VIDEO;
					else if (ainfo->mocc < 0)
						next = MPORT;
					else
						return (0);
					break;
				default:
					if (ainfo->mocc < 0)
						next = MPORT;
					else if (ainfo->gr_en && ainfo->gocc < 0 &&
					         !ainfo->gr_only_once)
						next = GRAPHICS;
					else if (ainfo->vid_en && ainfo->vocc < 0 &&
					         !ainfo->vid_only_once)
						next = VIDEO;
					else
						return (0);
					break;
			}
		last = cur;
		cur = next;
		iter++;
		switch (cur) {
			case VIDEO:
				if (last == cur)
					misses = 0;
				else if (ainfo->first_vacc)
					misses = vmisses;
				else
					misses = 1;
				ainfo->first_vacc = 0;
				if (last != cur) {
					ns = 1000000 *
					     (vmisses * state->mem_page_miss + state->mem_latency) /
					     state->mclk_khz;
					vlwm = ns * ainfo->vdrain_rate / 1000000;
					vlwm = ainfo->vocc - vlwm;
				}
				ns = 1000000 *
				     (misses * state->mem_page_miss + ainfo->vburst_size) /
				     (state->memory_width / 8) / state->mclk_khz;
				ainfo->vocc = ainfo->vocc + ainfo->vburst_size -
				              ns * ainfo->vdrain_rate / 1000000;
				ainfo->gocc = ainfo->gocc - ns * ainfo->gdrain_rate / 1000000;
				ainfo->mocc = ainfo->mocc - ns * ainfo->mdrain_rate / 1000000;
				break;
			case GRAPHICS:
				if (last == cur)
					misses = 0;
				else if (ainfo->first_gacc)
					misses = gmisses;
				else
					misses = 1;
				ainfo->first_gacc = 0;
				if (last != cur) {
					ns = 1000000 *
					     (gmisses * state->mem_page_miss + state->mem_latency) /
					     state->mclk_khz;
					glwm = ns * ainfo->gdrain_rate / 1000000;
					glwm = ainfo->gocc - glwm;
				}
				ns = 1000000 *
				     (misses * state->mem_page_miss +
				      ainfo->gburst_size / (state->memory_width / 8)) /
				     state->mclk_khz;
				ainfo->vocc =
				    ainfo->vocc + 0 - ns * ainfo->vdrain_rate / 1000000;
				ainfo->gocc = ainfo->gocc + ainfo->gburst_size -
				              ns * ainfo->gdrain_rate / 1000000;
				ainfo->mocc =
				    ainfo->mocc + 0 - ns * ainfo->mdrain_rate / 1000000;
				break;
			default:
				if (last == cur)
					misses = 0;
				else if (ainfo->first_macc)
					misses = mmisses;
				else
					misses = 1;
				ainfo->first_macc = 0;
				ns = 1000000 *
				     (misses * state->mem_page_miss +
				      mburst_size / (state->memory_width / 8)) /
				     state->mclk_khz;
				ainfo->vocc =
				    ainfo->vocc + 0 - ns * ainfo->vdrain_rate / 1000000;
				ainfo->gocc =
				    ainfo->gocc + 0 - ns * ainfo->gdrain_rate / 1000000;
				ainfo->mocc = ainfo->mocc + mburst_size -
				              ns * ainfo->mdrain_rate / 1000000;
				break;
		}
		if (iter > 100) {
			ainfo->converged = 0;
			return (1);
		}
		ns = 1000000 * ainfo->gburst_size / (state->memory_width / 8) /
		     state->mclk_khz;
		tmp = ns * ainfo->gdrain_rate / 1000000;
		if (ABS(ainfo->gburst_size) + ((ABS(ainfo->wcglwm) + 16) & ~0x7) - tmp >
		    max_gfsize) {
			ainfo->converged = 0;
			return (1);
		}
		ns = 1000000 * ainfo->vburst_size / (state->memory_width / 8) /
		     state->mclk_khz;
		tmp = ns * ainfo->vdrain_rate / 1000000;
		if (ABS(ainfo->vburst_size) + (ABS(ainfo->wcvlwm + 32) & ~0xf) - tmp >
		    VFIFO_SIZE) {
			ainfo->converged = 0;
			return (1);
		}
		if (ABS(ainfo->gocc) > max_gfsize) {
			ainfo->converged = 0;
			return (1);
		}
		if (ABS(ainfo->vocc) > VFIFO_SIZE) {
			ainfo->converged = 0;
			return (1);
		}
		if (ABS(ainfo->mocc) > MFIFO_SIZE) {
			ainfo->converged = 0;
			return (1);
		}
		if (ABS(vfsize) > VFIFO_SIZE) {
			ainfo->converged = 0;
			return (1);
		}
		if (ABS(gfsize) > max_gfsize) {
			ainfo->converged = 0;
			return (1);
		}
		if (ABS(mfsize) > MFIFO_SIZE) {
			ainfo->converged = 0;
			return (1);
		}
	}
}
static char nv3_arb(nv3_fifo_info *res_info, nv3_sim_state *state,
                    nv3_arb_info *ainfo) {
	long ens, vns, mns, gns;
	int mmisses, gmisses, vmisses, eburst_size, mburst_size;
	int refresh_cycle;

	refresh_cycle = 0;
	refresh_cycle = 2 * (state->mclk_khz / state->pclk_khz) + 5;
	mmisses = 2;
	if (state->mem_aligned)
		gmisses = 2;
	else
		gmisses = 3;
	vmisses = 2;
	eburst_size = state->memory_width * 1;
	mburst_size = 32;
	gns = 1000000 * (gmisses * state->mem_page_miss + state->mem_latency) /
	      state->mclk_khz;
	ainfo->by_gfacc = gns * ainfo->gdrain_rate / 1000000;
	ainfo->wcmocc = 0;
	ainfo->wcgocc = 0;
	ainfo->wcvocc = 0;
	ainfo->wcvlwm = 0;
	ainfo->wcglwm = 0;
	ainfo->engine_en = 1;
	ainfo->converged = 1;
	if (ainfo->engine_en) {
		ens = 1000000 *
		      (state->mem_page_miss + eburst_size / (state->memory_width / 8) +
		       refresh_cycle) /
		      state->mclk_khz;
		ainfo->mocc =
		    state->enable_mp ? 0 - ens * ainfo->mdrain_rate / 1000000 : 0;
		ainfo->vocc =
		    ainfo->vid_en ? 0 - ens * ainfo->vdrain_rate / 1000000 : 0;
		ainfo->gocc = ainfo->gr_en ? 0 - ens * ainfo->gdrain_rate / 1000000 : 0;
		ainfo->cur = ENGINE;
		ainfo->first_vacc = 1;
		ainfo->first_gacc = 1;
		ainfo->first_macc = 1;
		nv3_iterate(res_info, state, ainfo);
	}
	if (state->enable_mp) {
		mns = 1000000 *
		      (mmisses * state->mem_page_miss +
		       mburst_size / (state->memory_width / 8) + refresh_cycle) /
		      state->mclk_khz;
		ainfo->mocc = state->enable_mp
		                  ? 0
		                  : mburst_size - mns * ainfo->mdrain_rate / 1000000;
		ainfo->vocc =
		    ainfo->vid_en ? 0 : 0 - mns * ainfo->vdrain_rate / 1000000;
		ainfo->gocc = ainfo->gr_en ? 0 : 0 - mns * ainfo->gdrain_rate / 1000000;
		ainfo->cur = MPORT;
		ainfo->first_vacc = 1;
		ainfo->first_gacc = 1;
		ainfo->first_macc = 0;
		nv3_iterate(res_info, state, ainfo);
	}
	if (ainfo->gr_en) {
		ainfo->first_vacc = 1;
		ainfo->first_gacc = 0;
		ainfo->first_macc = 1;
		gns = 1000000 *
		      (gmisses * state->mem_page_miss +
		       ainfo->gburst_size / (state->memory_width / 8) + refresh_cycle) /
		      state->mclk_khz;
		ainfo->gocc = ainfo->gburst_size - gns * ainfo->gdrain_rate / 1000000;
		ainfo->vocc =
		    ainfo->vid_en ? 0 - gns * ainfo->vdrain_rate / 1000000 : 0;
		ainfo->mocc =
		    state->enable_mp ? 0 - gns * ainfo->mdrain_rate / 1000000 : 0;
		ainfo->cur = GRAPHICS;
		nv3_iterate(res_info, state, ainfo);
	}
	if (ainfo->vid_en) {
		ainfo->first_vacc = 0;
		ainfo->first_gacc = 1;
		ainfo->first_macc = 1;
		vns = 1000000 *
		      (vmisses * state->mem_page_miss +
		       ainfo->vburst_size / (state->memory_width / 8) + refresh_cycle) /
		      state->mclk_khz;
		ainfo->vocc = ainfo->vburst_size - vns * ainfo->vdrain_rate / 1000000;
		ainfo->gocc =
		    ainfo->gr_en ? (0 - vns * ainfo->gdrain_rate / 1000000) : 0;
		ainfo->mocc =
		    state->enable_mp ? 0 - vns * ainfo->mdrain_rate / 1000000 : 0;
		ainfo->cur = VIDEO;
		nv3_iterate(res_info, state, ainfo);
	}
	if (ainfo->converged) {
		res_info->graphics_lwm = (int)ABS(ainfo->wcglwm) + 16;
		res_info->video_lwm = (int)ABS(ainfo->wcvlwm) + 32;
		res_info->graphics_burst_size = ainfo->gburst_size;
		res_info->video_burst_size = ainfo->vburst_size;
		res_info->graphics_hi_priority = (ainfo->priority == GRAPHICS);
		res_info->media_hi_priority = (ainfo->priority == MPORT);
		if (res_info->video_lwm > 160) {
			res_info->graphics_lwm = 256;
			res_info->video_lwm = 128;
			res_info->graphics_burst_size = 64;
			res_info->video_burst_size = 64;
			res_info->graphics_hi_priority = 0;
			res_info->media_hi_priority = 0;
			ainfo->converged = 0;
			return (0);
		}
		if (res_info->video_lwm > 128) { res_info->video_lwm = 128; }
		return (1);
	} else {
		res_info->graphics_lwm = 256;
		res_info->video_lwm = 128;
		res_info->graphics_burst_size = 64;
		res_info->video_burst_size = 64;
		res_info->graphics_hi_priority = 0;
		res_info->media_hi_priority = 0;
		return (0);
	}
}
static char nv3_get_param(nv3_fifo_info *res_info, nv3_sim_state *state,
                          nv3_arb_info *ainfo) {
	int done, g, v, p;

	done = 0;
	if (state->gr_during_vid && ainfo->vid_en)
		ainfo->priority = MPORT;
	else
		ainfo->priority =
		    ainfo->gdrain_rate < ainfo->vdrain_rate ? VIDEO : GRAPHICS;
	for (p = 0; p < 2 && done != 1; p++) {
		for (g = 128; (g > 32) && (done != 1); g = g >> 1) {
			for (v = 128; (v >= 32) && (done != 1); v = v >> 1) {
				ainfo->priority = p;
				ainfo->gburst_size = g;
				ainfo->vburst_size = v;
				done = nv3_arb(res_info, state, ainfo);
				if (g == 128) {
					if ((res_info->graphics_lwm + g) > 256) done = 0;
				}
			}
		}
	}
	if (!done)
		return (0);
	else
		return (1);
}
static void nv3CalcArbitration(nv3_fifo_info *res_info, nv3_sim_state *state) {
	nv3_fifo_info save_info;
	nv3_arb_info ainfo;
	char res_gr, res_vid;

	ainfo.gr_en = 1;
	ainfo.vid_en = state->enable_video;
	ainfo.vid_only_once = 0;
	ainfo.gr_only_once = 0;
	ainfo.gdrain_rate = (int)state->pclk_khz * state->pix_bpp / 8;
	ainfo.vdrain_rate = (int)state->pclk_khz * 2;
	if (state->video_scale != 0)
		ainfo.vdrain_rate = ainfo.vdrain_rate / state->video_scale;
	ainfo.mdrain_rate = 33000;
	res_info->rtl_values = 0;
	if (!state->gr_during_vid && state->enable_video) {
		ainfo.gr_only_once = 1;
		ainfo.gr_en = 1;
		ainfo.gdrain_rate = 0;
		res_vid = nv3_get_param(res_info, state, &ainfo);
		res_vid = ainfo.converged;
		save_info.video_lwm = res_info->video_lwm;
		save_info.video_burst_size = res_info->video_burst_size;
		ainfo.vid_en = 1;
		ainfo.vid_only_once = 1;
		ainfo.gr_en = 1;
		ainfo.gdrain_rate = (int)state->pclk_khz * state->pix_bpp / 8;
		ainfo.vdrain_rate = 0;
		res_gr = nv3_get_param(res_info, state, &ainfo);
		res_gr = ainfo.converged;
		res_info->video_lwm = save_info.video_lwm;
		res_info->video_burst_size = save_info.video_burst_size;
		res_info->valid = res_gr & res_vid;
	} else {
		if (!ainfo.gr_en) ainfo.gdrain_rate = 0;
		if (!ainfo.vid_en) ainfo.vdrain_rate = 0;
		res_gr = nv3_get_param(res_info, state, &ainfo);
		res_info->valid = ainfo.converged;
	}
}
void nv3UpdateArbitrationSettings(unsigned VClk, unsigned pixelDepth,
                                  unsigned *burst, unsigned *lwm,
                                  RIVA_HW_INST *chip) {
	nv3_fifo_info fifo_data;
	nv3_sim_state sim_data;
	unsigned int M, N, P, pll, MClk;

	pll = chip->PRAMDAC[0x00000504 / 4];
	M = (pll >> 0) & 0xFF;
	N = (pll >> 8) & 0xFF;
	P = (pll >> 16) & 0x0F;
	MClk = (N * chip->CrystalFreqKHz / M) >> P;
	sim_data.pix_bpp = (char)pixelDepth;
	sim_data.enable_video = 0;
	sim_data.enable_mp = 0;
	sim_data.video_scale = 1;
	sim_data.memory_width = (chip->PEXTDEV[0x00000000 / 4] & 0x10) ? 128 : 64;
	sim_data.memory_width = 128;
	sim_data.mem_latency = 11;
	sim_data.mem_aligned = 1;
	sim_data.mem_page_miss = 9;
	sim_data.gr_during_vid = 0;
	sim_data.pclk_khz = VClk;
	sim_data.mclk_khz = MClk;
	nv3CalcArbitration(&fifo_data, &sim_data);
	if (fifo_data.valid) {
		int b = fifo_data.graphics_burst_size >> 4;
		*burst = 0;
		while (b >>= 1) (*burst)++;
		*lwm = fifo_data.graphics_lwm >> 3;
	} else {
		*lwm = 0x24;
		*burst = 0x02;
	}
}
static void nv4CalcArbitration(nv4_fifo_info *fifo, nv4_sim_state *arb) {
	int data, pagemiss, cas, width, video_enable, color_key_enable, bpp, align;
	int nvclks, mclks, pclks, vpagemiss, crtpagemiss, vbs;
	int found, mclk_extra, mclk_loop, cbs, m1, p1;
	int mclk_freq, pclk_freq, nvclk_freq, mp_enable;
	int us_m, us_n, us_p, video_drain_rate, crtc_drain_rate;
	int vpm_us, us_video, vlwm, video_fill_us, cpm_us, us_crt, clwm;
	int craw, vraw;

	fifo->valid = 1;
	pclk_freq = arb->pclk_khz;
	mclk_freq = arb->mclk_khz;
	nvclk_freq = arb->nvclk_khz;
	pagemiss = arb->mem_page_miss;
	cas = arb->mem_latency;
	width = arb->memory_width >> 6;
	video_enable = arb->enable_video;
	color_key_enable = arb->gr_during_vid;
	bpp = arb->pix_bpp;
	align = arb->mem_aligned;
	mp_enable = arb->enable_mp;
	clwm = 0;
	vlwm = 0;
	cbs = 128;
	pclks = 2;
	nvclks = 2;
	nvclks += 2;
	nvclks += 1;
	mclks = 5;
	mclks += 3;
	mclks += 1;
	mclks += cas;
	mclks += 1;
	mclks += 1;
	mclks += 1;
	mclks += 1;
	mclk_extra = 3;
	nvclks += 2;
	nvclks += 1;
	nvclks += 1;
	nvclks += 1;
	if (mp_enable) mclks += 4;
	nvclks += 0;
	pclks += 0;
	found = 0;
	vbs = 0;
	while (found != 1) {
		fifo->valid = 1;
		found = 1;
		mclk_loop = mclks + mclk_extra;
		us_m = mclk_loop * 1000 * 1000 / mclk_freq;
		us_n = nvclks * 1000 * 1000 / nvclk_freq;
		us_p = nvclks * 1000 * 1000 / pclk_freq;
		if (video_enable) {
			video_drain_rate = pclk_freq * 2;
			crtc_drain_rate = pclk_freq * bpp / 8;
			vpagemiss = 2;
			vpagemiss += 1;
			crtpagemiss = 2;
			vpm_us = (vpagemiss * pagemiss) * 1000 * 1000 / mclk_freq;
			if (nvclk_freq * 2 > mclk_freq * width)
				video_fill_us = cbs * 1000 * 1000 / 16 / nvclk_freq;
			else
				video_fill_us = cbs * 1000 * 1000 / (8 * width) / mclk_freq;
			us_video = vpm_us + us_m + us_n + us_p + video_fill_us;
			vlwm = us_video * video_drain_rate / (1000 * 1000);
			vlwm++;
			vbs = 128;
			if (vlwm > 128) vbs = 64;
			if (vlwm > (256 - 64)) vbs = 32;
			if (nvclk_freq * 2 > mclk_freq * width)
				video_fill_us = vbs * 1000 * 1000 / 16 / nvclk_freq;
			else
				video_fill_us = vbs * 1000 * 1000 / (8 * width) / mclk_freq;
			cpm_us = crtpagemiss * pagemiss * 1000 * 1000 / mclk_freq;
			us_crt = us_video + video_fill_us + cpm_us + us_m + us_n + us_p;
			clwm = us_crt * crtc_drain_rate / (1000 * 1000);
			clwm++;
		} else {
			crtc_drain_rate = pclk_freq * bpp / 8;
			crtpagemiss = 2;
			crtpagemiss += 1;
			cpm_us = crtpagemiss * pagemiss * 1000 * 1000 / mclk_freq;
			us_crt = cpm_us + us_m + us_n + us_p;
			clwm = us_crt * crtc_drain_rate / (1000 * 1000);
			clwm++;
		}
		m1 = clwm + cbs - 512;
		p1 = m1 * pclk_freq / mclk_freq;
		p1 = p1 * bpp / 8;
		if ((p1 < m1) && (m1 > 0)) {
			fifo->valid = 0;
			found = 0;
			if (mclk_extra == 0) found = 1;
			mclk_extra--;
		} else if (video_enable) {
			if ((clwm > 511) || (vlwm > 255)) {
				fifo->valid = 0;
				found = 0;
				if (mclk_extra == 0) found = 1;
				mclk_extra--;
			}
		} else {
			if (clwm > 519) {
				fifo->valid = 0;
				found = 0;
				if (mclk_extra == 0) found = 1;
				mclk_extra--;
			}
		}
		craw = clwm;
		vraw = vlwm;
		if (clwm < 384) clwm = 384;
		if (vlwm < 128) vlwm = 128;
		data = (int)(clwm);
		fifo->graphics_lwm = data;
		fifo->graphics_burst_size = 128;
		data = (int)((vlwm + 15));
		fifo->video_lwm = data;
		fifo->video_burst_size = vbs;
	}
}
static void nv4UpdateArbitrationSettings(unsigned VClk, unsigned pixelDepth,
                                         unsigned *burst, unsigned *lwm,
                                         RIVA_HW_INST *chip) {
	nv4_fifo_info fifo_data;
	nv4_sim_state sim_data;
	unsigned int M, N, P, pll, MClk, NVClk, cfg1;

	pll = chip->PRAMDAC[0x00000504 / 4];
	M = (pll >> 0) & 0xFF;
	N = (pll >> 8) & 0xFF;
	P = (pll >> 16) & 0x0F;
	MClk = (N * chip->CrystalFreqKHz / M) >> P;
	pll = chip->PRAMDAC[0x00000500 / 4];
	M = (pll >> 0) & 0xFF;
	N = (pll >> 8) & 0xFF;
	P = (pll >> 16) & 0x0F;
	NVClk = (N * chip->CrystalFreqKHz / M) >> P;
	cfg1 = chip->PFB[0x00000204 / 4];
	sim_data.pix_bpp = (char)pixelDepth;
	sim_data.enable_video = 0;
	sim_data.enable_mp = 0;
	sim_data.memory_width = (chip->PEXTDEV[0x00000000 / 4] & 0x10) ? 128 : 64;
	sim_data.mem_latency = (char)cfg1 & 0x0F;
	sim_data.mem_aligned = 1;
	sim_data.mem_page_miss =
	    (char)(((cfg1 >> 4) & 0x0F) + ((cfg1 >> 31) & 0x01));
	sim_data.gr_during_vid = 0;
	sim_data.pclk_khz = VClk;
	sim_data.mclk_khz = MClk;
	sim_data.nvclk_khz = NVClk;
	nv4CalcArbitration(&fifo_data, &sim_data);
	if (fifo_data.valid) {
		int b = fifo_data.graphics_burst_size >> 4;
		*burst = 0;
		while (b >>= 1) (*burst)++;
		*lwm = fifo_data.graphics_lwm >> 3;
	}
}

/****************************************************************************\
*                                                                            *
*                          RIVA Mode State Routines                          *
*                                                                            *
\****************************************************************************/

/*
 * Calculate the Video Clock parameters for the PLL.
 */
static int CalcVClock(int clockIn, int *clockOut, int *mOut, int *nOut,
                      int *pOut, RIVA_HW_INST *chip) {
	unsigned lowM, highM, highP;
	unsigned DeltaNew, DeltaOld;
	unsigned VClk, Freq;
	unsigned M, N, P;

	DeltaOld = 0xFFFFFFFF;
	VClk = (unsigned)clockIn;
	if (chip->CrystalFreqKHz == 14318) {
		lowM = 8;
		highM = 14 - (chip->Architecture == 3);
	} else {
		lowM = 7;
		highM = 13 - (chip->Architecture == 3);
	}
	highP = 4 - (chip->Architecture == 3);
	for (P = 0; P <= highP; P++) {
		Freq = VClk << P;
		if ((Freq >= 128000) && (Freq <= chip->MaxVClockFreqKHz)) {
			for (M = lowM; M <= highM; M++) {
				N = (VClk * M / chip->CrystalFreqKHz) << P;
				Freq = (chip->CrystalFreqKHz * N / M) >> P;
				if (Freq > VClk)
					DeltaNew = Freq - VClk;
				else
					DeltaNew = VClk - Freq;
				if (DeltaNew < DeltaOld) {
					*mOut = M;
					*nOut = N;
					*pOut = P;
					*clockOut = Freq;
					DeltaOld = DeltaNew;
				}
			}
		}
	}
	return (DeltaOld != 0xFFFFFFFF);
}
/*
 * Calculate extended mode parameters (SVGA) and save in a
 * mode state structure.
 */
static void CalcStateExt(RIVA_HW_INST *chip, RIVA_HW_STATE *state, int bpp,
                         int width, int hDisplaySize, int hDisplay, int hStart,
                         int hEnd, int hTotal, int height, int vDisplay,
                         int vStart, int vEnd, int vTotal, int dotClock) {
	int pixelDepth, VClk, m, n, p;
	/*
	 * Save mode parameters.
	 */
	state->bpp = bpp;
	state->width = width;
	state->height = height;
	/*
	 * Extended RIVA registers.
	 */
	pixelDepth = (bpp + 1) / 8;
	CalcVClock(dotClock, &VClk, &m, &n, &p, chip);
	switch (chip->Architecture) {
		case 3:
			nv3UpdateArbitrationSettings(VClk, pixelDepth * 8,
			                             &(state->arbitration0),
			                             &(state->arbitration1), chip);
			state->cursor0 = 0x00;
			state->cursor1 = 0x78;
			state->cursor2 = 0x00000000;
			state->pllsel = 0x10010100;
			state->config = ((width + 31) / 32) |
			                (((pixelDepth > 2) ? 3 : pixelDepth) << 8) | 0x1000;
			state->general = 0x00000100;
			state->repaint1 = hDisplaySize < 1280 ? 0x06 : 0x02;
			break;
		case 4:
			nv4UpdateArbitrationSettings(VClk, pixelDepth * 8,
			                             &(state->arbitration0),
			                             &(state->arbitration1), chip);
			state->cursor0 = 0x00;
			state->cursor1 = 0xFC;
			state->cursor2 = 0x00000000;
			state->pllsel = 0x10000700;
			state->config = 0x00001114;
			state->general = bpp == 16 ? 0x00101100 : 0x00100100;
			state->repaint1 = hDisplaySize < 1280 ? 0x04 : 0x00;
			break;
	}
	state->vpll = (p << 16) | (n << 8) | m;
	state->screen = ((hTotal & 0x040) >> 2) | ((vDisplay & 0x400) >> 7) |
	                ((vStart & 0x400) >> 8) | ((vDisplay & 0x400) >> 9) |
	                ((vTotal & 0x400) >> 10);
	state->repaint0 = (((width / 8) * pixelDepth) & 0x700) >> 3;
	state->horiz = hTotal < 260 ? 0x00 : 0x01;
	state->pixel = pixelDepth > 2 ? 3 : pixelDepth;
	state->offset0 = state->offset1 = state->offset2 = state->offset3 = 0;
	state->pitch0 = state->pitch1 = state->pitch2 = state->pitch3 =
	    pixelDepth * width;
}
/*
 * Load fixed function state and pre-calculated/stored state.
 */
#define LOAD_FIXED_STATE(tbl, dev)                                             \
	for (i = 0; i < sizeof(tbl##Table##dev) / 8; i++)                          \
	chip->dev[tbl##Table##dev[i][0]] = tbl##Table##dev[i][1]
#define LOAD_FIXED_STATE_8BPP(tbl, dev)                                        \
	for (i = 0; i < sizeof(tbl##Table##dev##_8BPP) / 8; i++)                   \
	chip->dev[tbl##Table##dev##_8BPP[i][0]] = tbl##Table##dev##_8BPP[i][1]
#define LOAD_FIXED_STATE_15BPP(tbl, dev)                                       \
	for (i = 0; i < sizeof(tbl##Table##dev##_15BPP) / 8; i++)                  \
	chip->dev[tbl##Table##dev##_15BPP[i][0]] = tbl##Table##dev##_15BPP[i][1]
#define LOAD_FIXED_STATE_16BPP(tbl, dev)                                       \
	for (i = 0; i < sizeof(tbl##Table##dev##_16BPP) / 8; i++)                  \
	chip->dev[tbl##Table##dev##_16BPP[i][0]] = tbl##Table##dev##_16BPP[i][1]
#define LOAD_FIXED_STATE_32BPP(tbl, dev)                                       \
	for (i = 0; i < sizeof(tbl##Table##dev##_32BPP) / 8; i++)                  \
	chip->dev[tbl##Table##dev##_32BPP[i][0]] = tbl##Table##dev##_32BPP[i][1]
static void LoadStateExt(RIVA_HW_INST *chip, RIVA_HW_STATE *state, int all) {
	int i;

	/*
	 * Load HW fixed function state.
	 */
	LOAD_FIXED_STATE(Riva, PMC);
	LOAD_FIXED_STATE(Riva, PTIMER);
	/*
	 * Make sure frame buffer config gets set before loading PRAMIN.
	 */
	chip->PFB[0x00000200 / 4] = state->config;
	switch (chip->Architecture) {
		case 3:
			LOAD_FIXED_STATE(nv3, PFIFO);
			LOAD_FIXED_STATE(nv3, PRAMIN);
			LOAD_FIXED_STATE(nv3, PGRAPH);
			switch (state->bpp) {
				case 15:
				case 16:
					LOAD_FIXED_STATE_15BPP(nv3, PRAMIN);
					LOAD_FIXED_STATE_15BPP(nv3, PGRAPH);
					chip->Tri03 =
					    (RivaTexturedTriangle03 *)&(chip->FIFO[0x0000E000 / 4]);
					break;
				case 24:
				case 32:
					LOAD_FIXED_STATE_32BPP(nv3, PRAMIN);
					LOAD_FIXED_STATE_32BPP(nv3, PGRAPH);
					chip->Tri03 = 0L;
					break;
				case 8:
				default:
					LOAD_FIXED_STATE_8BPP(nv3, PRAMIN);
					LOAD_FIXED_STATE_8BPP(nv3, PGRAPH);
					chip->Tri03 = 0L;
					break;
			}
			for (i = 0x00000; i < 0x00800; i++)
				chip->PRAMIN[0x00000502 + i] = (i << 12) | 0x03;
			chip->PGRAPH[0x00000630 / 4] = state->offset0;
			chip->PGRAPH[0x00000634 / 4] = state->offset1;
			chip->PGRAPH[0x00000638 / 4] = state->offset2;
			chip->PGRAPH[0x0000063C / 4] = state->offset3;
			chip->PGRAPH[0x00000650 / 4] = state->pitch0;
			chip->PGRAPH[0x00000654 / 4] = state->pitch1;
			chip->PGRAPH[0x00000658 / 4] = state->pitch2;
			chip->PGRAPH[0x0000065C / 4] = state->pitch3;
			break;
		case 4:
			LOAD_FIXED_STATE(nv4, PFIFO);
			LOAD_FIXED_STATE(nv4, PRAMIN);
			LOAD_FIXED_STATE(nv4, PGRAPH);
			switch (state->bpp) {
				case 15:
					LOAD_FIXED_STATE_15BPP(nv4, PRAMIN);
					LOAD_FIXED_STATE_15BPP(nv4, PGRAPH);
					chip->Tri03 =
					    (RivaTexturedTriangle03 *)&(chip->FIFO[0x0000E000 / 4]);
					break;
				case 16:
					LOAD_FIXED_STATE_16BPP(nv4, PRAMIN);
					LOAD_FIXED_STATE_16BPP(nv4, PGRAPH);
					chip->Tri03 =
					    (RivaTexturedTriangle03 *)&(chip->FIFO[0x0000E000 / 4]);
					break;
				case 24:
				case 32:
					LOAD_FIXED_STATE_32BPP(nv4, PRAMIN);
					LOAD_FIXED_STATE_32BPP(nv4, PGRAPH);
					chip->Tri03 = 0L;
					break;
				case 8:
				default:
					LOAD_FIXED_STATE_8BPP(nv4, PRAMIN);
					LOAD_FIXED_STATE_8BPP(nv4, PGRAPH);
					chip->Tri03 = 0L;
					break;
			}
			chip->PGRAPH[0x00000640 / 4] = state->offset0;
			chip->PGRAPH[0x00000644 / 4] = state->offset1;
			chip->PGRAPH[0x00000648 / 4] = state->offset2;
			chip->PGRAPH[0x0000064C / 4] = state->offset3;
			chip->PGRAPH[0x00000670 / 4] = state->pitch0;
			chip->PGRAPH[0x00000674 / 4] = state->pitch1;
			chip->PGRAPH[0x00000678 / 4] = state->pitch2;
			chip->PGRAPH[0x0000067C / 4] = state->pitch3;
			break;
	}
	LOAD_FIXED_STATE(Riva, FIFO);
	/*
	 * Load HW mode state.
	 */
	/* slh: added the if (all) checks */
	outb(0x3D4, 0x19);
	outb(0x3D5, state->repaint0);
	outb(0x3D4, 0x1A);
	outb(0x3D5, state->repaint1);
	if (all) {
		outb(0x3D4, 0x25);
		outb(0x3D5, state->screen);
	}
	outb(0x3D4, 0x28);
	outb(0x3D5, state->pixel);
	if (all) {
		outb(0x3D4, 0x2D);
		outb(0x3D5, state->horiz);
		outb(0x3D4, 0x1B);
		outb(0x3D5, state->arbitration0);
		outb(0x3D4, 0x20);
		outb(0x3D5, state->arbitration1);
	}
	outb(0x3D4, 0x30);
	outb(0x3D5, state->cursor0);
	outb(0x3D4, 0x31);
	outb(0x3D5, state->cursor1);
	chip->PRAMDAC[0x00000300 / 4] = state->cursor2;
	if (all) { chip->PRAMDAC[0x00000508 / 4] = state->vpll; }
	chip->PRAMDAC[0x0000050C / 4] = state->pllsel;
	chip->PRAMDAC[0x00000600 / 4] = state->general;

	/*
	 * Turn off VBlank enable and reset.
	 */
	*(chip->VBLANKENABLE) = 0;
	*(chip->VBLANK) = chip->VBlankBit;
	/*
	 * Set interrupt enable.
	 */
	chip->PMC[0x00000140 / 4] = chip->EnableIRQ & 0x01;
	/*
	 * Set current state pointer.
	 */
	chip->CurrentState = state;
	/*
	 * Reset FIFO free count.
	 */
	chip->FifoFreeCount = 0;
}
static void UnloadStateExt(RIVA_HW_INST *chip, RIVA_HW_STATE *state) {
	/*
	 * Save current HW state.
	 */
	outb(0x3D4, 0x19);
	state->repaint0 = inb(0x3D5);
	outb(0x3D4, 0x1A);
	state->repaint1 = inb(0x3D5);
	outb(0x3D4, 0x25);
	state->screen = inb(0x3D5);
	outb(0x3D4, 0x28);
	state->pixel = inb(0x3D5);
	outb(0x3D4, 0x2D);
	state->horiz = inb(0x3D5);
	outb(0x3D4, 0x1B);
	state->arbitration0 = inb(0x3D5);
	outb(0x3D4, 0x20);
	state->arbitration1 = inb(0x3D5);
	outb(0x3D4, 0x30);
	state->cursor0 = inb(0x3D5);
	outb(0x3D4, 0x31);
	state->cursor1 = inb(0x3D5);
	state->cursor2 = chip->PRAMDAC[0x00000300 / 4];
	state->vpll = chip->PRAMDAC[0x00000508 / 4];
	state->pllsel = chip->PRAMDAC[0x0000050C / 4];
	state->general = chip->PRAMDAC[0x00000600 / 4];
	state->config = chip->PFB[0x00000200 / 4];
	switch (chip->Architecture) {
		case 3:
			state->offset0 = chip->PGRAPH[0x00000630 / 4];
			state->offset1 = chip->PGRAPH[0x00000634 / 4];
			state->offset2 = chip->PGRAPH[0x00000638 / 4];
			state->offset3 = chip->PGRAPH[0x0000063C / 4];
			state->pitch0 = chip->PGRAPH[0x00000650 / 4];
			state->pitch1 = chip->PGRAPH[0x00000654 / 4];
			state->pitch2 = chip->PGRAPH[0x00000658 / 4];
			state->pitch3 = chip->PGRAPH[0x0000065C / 4];
			break;
		case 4:
			state->offset0 = chip->PGRAPH[0x00000640 / 4];
			state->offset1 = chip->PGRAPH[0x00000644 / 4];
			state->offset2 = chip->PGRAPH[0x00000648 / 4];
			state->offset3 = chip->PGRAPH[0x0000064C / 4];
			state->pitch0 = chip->PGRAPH[0x00000670 / 4];
			state->pitch1 = chip->PGRAPH[0x00000674 / 4];
			state->pitch2 = chip->PGRAPH[0x00000678 / 4];
			state->pitch3 = chip->PGRAPH[0x0000067C / 4];
			break;
	}
}
static void SetStartAddress(RIVA_HW_INST *chip, unsigned start) {
	int offset = start >> 2;
	int pan = (start & 3) << 1;
	unsigned char tmp;

	/*
	 * Unlock extended registers.
	 */
	outb(chip->LockUnlockIO, chip->LockUnlockIndex);
	outb(chip->LockUnlockIO + 1, 0x57);

	/*
	 * Set start address.
	 */
	outb(0x3D4, 0x0D);
	outb(0x3D5, offset);
	outb(0x3D4, 0x0C);
	outb(0x3D5, offset >> 8);

	/* slh: added the TNT version of this routine. Not sure what this
	 * really should be, but the changed code works on my TNT, where
	 * the original did not.
	 */
	if (chip->Architecture > 3) {
		/* TNT */
		outb(0x3D4, 0x19);
		tmp = inb(0x3D5);
		outb(0x3D5, ((offset >> 16) & 0x1F) | (tmp & 0xE0));

		outb(0x3D4, 0x2D);
		tmp = inb(0x3D5);
		outb(0x3D5, ((offset >> 16) & 0x20) | (tmp & 0xDF));
	} else {
		/* Riva 128 */
		outb(0x3D4, 0x19);
		tmp = inb(0x3D5);
		outb(0x3D5, ((offset >> 16) & 0x0F) | (tmp & 0xF0));
	}

	/*
	 * 4 pixel pan register.
	 */
	offset = inb(chip->IO + 0x0A);
	outb(0x3C0, 0x13);
	outb(0x3C0, pan);
}
static void nv3SetSurfaces2D(RIVA_HW_INST *chip, unsigned surf0,
                             unsigned surf1) {
	while (nv3Busy(chip))
		;
	chip->PGRAPH[0x00000630 / 4] = surf0;
	chip->PGRAPH[0x00000634 / 4] = surf1;
}
static void nv4SetSurfaces2D(RIVA_HW_INST *chip, unsigned surf0,
                             unsigned surf1) {
	while (nv4Busy(chip))
		;
	chip->PGRAPH[0x00000640 / 4] = surf0;
	chip->PGRAPH[0x00000644 / 4] = surf1;
}
static void nv3SetSurfaces3D(RIVA_HW_INST *chip, unsigned surf0,
                             unsigned surf1) {
	while (nv3Busy(chip))
		;
	chip->PGRAPH[0x00000638 / 4] = surf0;
	chip->PGRAPH[0x0000063C / 4] = surf1;
}
static void nv4SetSurfaces3D(RIVA_HW_INST *chip, unsigned surf0,
                             unsigned surf1) {
	while (nv4Busy(chip))
		;
	chip->PGRAPH[0x00000648 / 4] = surf0;
	chip->PGRAPH[0x0000064C / 4] = surf1;
}

/****************************************************************************\
*                                                                            *
*                      Probe RIVA Chip Configuration                         *
*                                                                            *
\****************************************************************************/

void nv3GetConfig(RIVA_HW_INST *chip) {
	/*
	 * Fill in chip configuration.
	 */
	if (chip->PFB[0x00000000 / 4] & 0x00000020) {
		if (((chip->PMC[0x00000000 / 4] & 0xF0) == 0x20) &&
		    ((chip->PMC[0x00000000 / 4] & 0x0F) >= 0x02)) {
			/*
			 * SDRAM 128 ZX.
			 */
			chip->RamBandwidthKBytesPerSec = 800000;
			switch (chip->PFB[0x00000000 / 4] & 0x03) {
				case 2: chip->RamAmountKBytes = 1024 * 4 - 32; break;
				case 1: chip->RamAmountKBytes = 1024 * 2 - 32; break;
				default: chip->RamAmountKBytes = 1024 * 8 - 32; break;
			}
		} else {
			chip->RamBandwidthKBytesPerSec = 1000000;
			chip->RamAmountKBytes = 1024 * 8 - 32;
		}
	} else {
		/*
		 * SGRAM 128.
		 */
		chip->RamBandwidthKBytesPerSec = 1000000;
		switch (chip->PFB[0x00000000 / 4] & 0x00000003) {
			case 0: chip->RamAmountKBytes = 1024 * 8 - 32; break;
			case 2: chip->RamAmountKBytes = 1024 * 4 - 32; break;
			default: chip->RamAmountKBytes = 1024 * 2 - 32; break;
		}
	}
	chip->CrystalFreqKHz =
	    (chip->PEXTDEV[0x00000000 / 4] & 0x00000020) ? 14318 : 13500;
	chip->CURSOR = &(chip->PRAMIN[0x00008000 / 4 - 0x0800 / 4]);
	chip->CURSORPOS = &(chip->PRAMDAC[0x0300 / 4]);
	chip->VBLANKENABLE = &(chip->PGRAPH[0x0140 / 4]);
	chip->VBLANK = &(chip->PGRAPH[0x0100 / 4]);
	chip->VBlankBit = 0x00000100;
	chip->MaxVClockFreqKHz = 230000;
	chip->LockUnlockIO = 0x3C4;
	chip->LockUnlockIndex = 0x06;
	/*
	 * Set chip functions.
	 */
	chip->Busy = nv3Busy;
	chip->ShowHideCursor = ShowHideCursor;
	chip->CalcStateExt = CalcStateExt;
	chip->LoadStateExt = LoadStateExt;
	chip->UnloadStateExt = UnloadStateExt;
	chip->SetStartAddress = SetStartAddress;
	chip->SetSurfaces2D = nv3SetSurfaces2D;
	chip->SetSurfaces3D = nv3SetSurfaces3D;
}
void nv4GetConfig(RIVA_HW_INST *chip) {
	/*
	 * Fill in chip configuration.
	 */
	switch (chip->PFB[0x00000000 / 4] & 0x00000003) {
		case 0: chip->RamAmountKBytes = 1024 * 32 - 128; break;
		case 1: chip->RamAmountKBytes = 1024 * 4 - 128; break;
		case 2: chip->RamAmountKBytes = 1024 * 8 - 128; break;
		case 3:
		default: chip->RamAmountKBytes = 1024 * 16 - 128; break;
	}
	switch ((chip->PFB[0x00000000 / 4] >> 3) & 0x00000003) {
		case 3: chip->RamBandwidthKBytesPerSec = 800000; break;
		default: chip->RamBandwidthKBytesPerSec = 1000000; break;
	}
	chip->CrystalFreqKHz =
	    (chip->PEXTDEV[0x00000000 / 4] & 0x00000040) ? 14318 : 13500;
	chip->CURSOR = &(chip->PRAMIN[0x00010000 / 4 - 0x0800 / 4]);
	chip->CURSORPOS = &(chip->PRAMDAC[0x0300 / 4]);
	chip->VBLANKENABLE = &(chip->PCRTC[0x0140 / 4]);
	chip->VBLANK = &(chip->PCRTC[0x0100 / 4]);
	chip->VBlankBit = 0x00000001;
	chip->MaxVClockFreqKHz = 250000;
	chip->LockUnlockIO = 0x3D4;
	chip->LockUnlockIndex = 0x1F;
	/*
	 * Set chip functions.
	 */
	chip->Busy = nv4Busy;
	chip->ShowHideCursor = ShowHideCursor;
	chip->CalcStateExt = CalcStateExt;
	chip->LoadStateExt = LoadStateExt;
	chip->UnloadStateExt = UnloadStateExt;
	chip->SetStartAddress = SetStartAddress;
	chip->SetSurfaces2D = nv4SetSurfaces2D;
	chip->SetSurfaces3D = nv4SetSurfaces3D;
}
int RivaGetConfig(RIVA_HW_INST *chip) {
	/*
	 * Save this so future SW know whats it's dealing with.
	 */
	chip->Version = RIVA_SW_VERSION;
	/*
	 * Chip specific configuration.
	 */
	switch (chip->Architecture) {
		case 3: nv3GetConfig(chip); break;
		case 4: nv4GetConfig(chip); break;
		default: return (-1);
	}
	/*
	 * Fill in FIFO pointers.
	 */
	chip->Rop = (RivaRop *)&(chip->FIFO[0x00000000 / 4]);
	chip->Clip = (RivaClip *)&(chip->FIFO[0x00002000 / 4]);
	chip->Patt = (RivaPattern *)&(chip->FIFO[0x00004000 / 4]);
	chip->Pixmap = (RivaPixmap *)&(chip->FIFO[0x00006000 / 4]);
	chip->Blt = (RivaScreenBlt *)&(chip->FIFO[0x00008000 / 4]);
	chip->Bitmap = (RivaBitmap *)&(chip->FIFO[0x0000A000 / 4]);
	chip->Tri03 = (RivaTexturedTriangle03 *)&(chip->FIFO[0x0000E000 / 4]);
	return (0);
}