#include <u.h>
#include <libc.h>
#include <bio.h>

#include "vga.h"
#include "pci.h"


typedef struct Nvidia	Nvidia;
struct Nvidia {
	ulong	mmio;
	Pcidev*	pci;

	int		arch;
	int		crystalfreq;

	ulong*	pfb;
	ulong*	pramdac;
	ulong*	pextdev;

	ulong	cursor2;
	ulong	vpll;
	ulong	pllsel;
	ulong	general;
	ulong	config;
};

static int extcrts[] = {
	0x19, 0x1A, 0x25, 0x28, 0x2D, 0x30, 0x31, -1
};

static void
snarf(Vga* vga, Ctlr* ctlr)
{
	Nvidia *nv;
	Pcidev *p;
	ulong m, tmp;
	int i;

	if(vga->private == nil){
		vga->private = alloc(sizeof(Nvidia));
		nv = vga->private;
		if((p = pcimatch(0, 0x10DE, 0)) == nil)
			error("%s: not found\n", ctlr->name);
		switch(p->did){
		case 0x0020:		/* Riva TNT */
		case 0x0028:		/* Riva TNT2 */
		case 0x0029:		/* Riva TNT2 (Ultra)*/
		case 0x002C:		/* Riva TNT2 (Vanta) */
		case 0x002D:		/* Riva TNT2 M64 */
		case 0x00A0:		/* Riva TNT2 (Integrated) */
			nv->arch = 4;
			break;
		case 0x0100:		/* GeForce 256 */
		case 0x0101:		/* GeForce DDR */
		case 0x0103:		/* Quadro */
		case 0x0110:		/* GeForce2 MX */
		case 0x0111:		/* GeForce2 MX DDR */
		case 0x0112:		/* GeForce 2 Go */
		case 0x0113:		/* Quadro 2 MXR */
		case 0x0150:		/* GeForce2 GTS */
		case 0x0151:		/* GeForce2 GTS (rev 1) */
		case 0x0152:		/* GeForce2 Ultra */
		case 0x0153:		/* Quadro 2 Pro */
			nv->arch = 10;
			break;
		default:
			error("%s: DID %4.4uX unsupported\n",
				ctlr->name, p->did);
		}

		vgactlw("type", ctlr->name);

		if((m = segattach(0, "nvidiammio", 0, p->mem[0].size)) == -1)
			error("%s: can't attach mmio segment\n", ctlr->name);

		nv->pci = p;
		nv->mmio = m;

		nv->pfb = (ulong*) (nv->mmio + 0x00100000);
		nv->pramdac = (ulong*) (nv->mmio + 0x00680000);
		nv->pextdev = (ulong*) (nv->mmio + 0x00101000);
	}
	nv = vga->private;

	vga->p[1] = 4;
	vga->n[1] = 255;
	vga->f[1] = 350000000;

	/*
	 * Unlock
	 */
	vgaxo(Crtx, 0x1F, 0x57);

	if (nv->pextdev[0x00000000] & 0x00000040){
		nv->crystalfreq = RefFreq;
		vga->m[1] = 14;
	} else {
		nv->crystalfreq = 13500000;
		vga->m[1] = 13;
	}

	if (nv->arch == 4) {
		tmp = nv->pfb[0x00000000];
		if (tmp & 0x0100) {
			vga->vmz = ((tmp >> 12) & 0x0F) * 1024 + 1024 * 2;
		} else {
			tmp &= 0x03;
			if (tmp)
				vga->vmz = (1024*1024*2) << tmp;
			else
				vga->vmz = 1024*1024*32;
		}
	}
	if (nv->arch == 10) {
		tmp = (nv->pfb[0x0000020C/4] >> 20) & 0xFF;
		if (tmp == 0)
			tmp = 16;
		vga->vmz = 1024*1024*tmp;
	}

	for(i = 0; extcrts[i] >= 0; i++)
		vga->crt[extcrts[i]] = vgaxi(Crtx, extcrts[i]);

	nv->cursor2	= nv->pramdac[0x00000300/4];
	nv->vpll		= nv->pramdac[0x00000508/4];
	nv->pllsel		= nv->pramdac[0x0000050C/4];
	nv->general	= nv->pramdac[0x00000600/4];

	nv->config	= nv->pfb[0x00000200/4];

	ctlr->flag |= Fsnarf;
}


static void
options(Vga*, Ctlr* ctlr)
{
	ctlr->flag |= Hlinear|Foptions;
}


static void
clock(Vga* vga, Ctlr* ctlr)
{
	int m, n, p, f, d;
	Nvidia *nv;
	double trouble;

	nv = vga->private;

	if(vga->f[0] == 0)
		vga->f[0] = vga->mode->frequency;

	vga->d[0] = vga->f[0]+1;

	for (p=0; p <= vga->p[1]; p++){
		f = vga->f[0] << p;
		if ((f >= 128000000) && (f <= vga->f[1])) {
			for (m=7; m <= vga->m[1]; m++){
				trouble = (double) nv->crystalfreq / (double) (m << p);
				n = (vga->f[0] / trouble)+0.5;
				f = n*trouble + 0.5;
				d = vga->f[0] - f;
				if (d < 0)
					d = -d;
				if (n & ~0xFF)
					d = vga->d[0] + 1;
				if (d <= vga->d[0]){
					vga->n[0] = n;
					vga->m[0] = m;
					vga->p[0] = p;
					vga->d[0] = d;
				}
			}
		}
	}
	if (vga->d[0] > vga->f[0])
		error("%s: vclk %lud out of range\n", ctlr->name, vga->f[0]);
}


static void
init(Vga* vga, Ctlr* ctlr)
{
	Mode *mode;
	Nvidia *nv;
	int tmp;

	mode = vga->mode;
	/* seems to work at higher depth
	if(mode->z != 8)
		error("%s: only supports 8-bit colour\n", ctlr->name);	
	*/

	nv = vga->private;

	if(vga->linear && (ctlr->flag & Hlinear))
		ctlr->flag |= Ulinear;

	clock(vga, ctlr);

	vga->crt[0x30] = 0x00;
	vga->crt[0x31] = 0xFC;
	nv->cursor2 = 0;
	nv->vpll    = (vga->p[0] << 16) | (vga->n[0] << 8) | vga->m[0];
	nv->pllsel  = 0x10000700;
	if (mode->z == 16)
		nv->general = 0x00001100;
	else
		nv->general = 0x00000100;
	nv->config  = 0x00001114;


	vga->attribute[0x11] = Pblack;
	vga->crt[0x14] = 0x00;


	tmp = vga->crt[0x12];
	vga->crt[0x15] = tmp;
	if(tmp & 0x100)
		vga->crt[0x07] |= 0x08;
	else
		vga->crt[0x07] &= ~0x08;
	if(tmp & 0x200)
		vga->crt[0x09] |= 0x20;
	else
		vga->crt[0x09] &= ~0x20;

	vga->crt[0x16] = vga->crt[0x06] + 1;


	vga->crt[0x02] = vga->crt[0x01];
	tmp = vga->crt[0x00] + 4;
	vga->crt[0x03] = 0x80 | (tmp & 0x1F);
	if (tmp & 0x20)
		vga->crt[0x05] |= 0x80;
	else
		vga->crt[0x05] &= ~0x80;

	/* overflow bits */

	vga->crt[0x1A] = 0x02;
	if (mode->x < 1280)
		vga->crt[0x1A] |= 0x04;

	tmp = (mode->z / 8);
	if (tmp > 3)
		tmp=3;
	vga->crt[0x28] = tmp;

	vga->crt[0x19] = (vga->crt[0x13] & 0x0700) >> 3;

	vga->crt[0x25] = 0x00;
	if (vga->crt[0x06] & 0x400)
		vga->crt[0x25] |= 0x01;
	if (vga->crt[0x12] & 0x400)
		vga->crt[0x25] |= 0x02;
	if (vga->crt[0x10] & 0x400)
		vga->crt[0x25] |= 0x04;
	if (vga->crt[0x15] & 0x400)
		vga->crt[0x25] |= 0x08;
	if ((mode->ehb >> 3) & 0x40)
		vga->crt[0x25] |= 0x10;

	vga->crt[0x2D] = 0x00;
	if (vga->crt[0x00] & 0x100)
		vga->crt[0x2D] |= 0x01;
	if(vga->crt[0x01] & 0x100)
		vga->crt[0x2D] |= 0x02;
	if(vga->crt[0x02] & 0x100)
		vga->crt[0x2D] |= 0x04;
	if(vga->crt[0x04] & 0x100)
		vga->crt[0x2D] |= 0x08;

	ctlr->flag |= Finit;
}

static void
load(Vga* vga, Ctlr* ctlr)
{
	Nvidia *nv;
	int i;

	nv = vga->private;

	/*
	 * Unlock
	 */
	vgaxo(Crtx, 0x1F, 0x57);

	if (nv->arch == 4)
		nv->pfb[0x00000200/4] = nv->config;

	for(i = 0; extcrts[i] >= 0; i++)
		vgaxo(Crtx, extcrts[i], vga->crt[extcrts[i]]);

	nv->pramdac[0x00000300/4] = nv->cursor2;
	nv->pramdac[0x00000508/4] = nv->vpll;
	nv->pramdac[0x0000050C/4] = nv->pllsel;
	nv->pramdac[0x00000600/4] = nv->general;

	ctlr->flag |= Fload;
}


static void
dump(Vga* vga, Ctlr* ctlr)
{
	Nvidia *nv;
	int i, m, n, p, f;
	char buf[100];
	double trouble;

	if((nv = vga->private) == 0)
		return;

	for(i = 0; extcrts[i] >= 0; i++){
		sprint(buf, "Crt%2.2uX", extcrts[i]);
		printitem(ctlr->name, buf);
		printreg(vga->crt[extcrts[i]]);
	}

	p = (nv->vpll >> 16);
	n = (nv->vpll >> 8) & 0xFF;
	m = nv->vpll & 0xFF;
	trouble = nv->crystalfreq;
	trouble = trouble * n / (m<<p);
	f = trouble+0.5;
	printitem(ctlr->name, "dclk m n p");
	Bprint(&stdout, " %d %d - %d %d\n", f, m, n, p);
	printitem(ctlr->name, "CrystalFreq");
	Bprint(&stdout, " %d Hz\n", nv->crystalfreq);
	printitem(ctlr->name, "cursor2");
	Bprint(&stdout, " %lux\n", nv->cursor2);
	printitem(ctlr->name, "vpll");
	Bprint(&stdout, " %lux\n", nv->vpll);
	printitem(ctlr->name, "pllsel");
	Bprint(&stdout, " %lux\n", nv->pllsel);
	printitem(ctlr->name, "general");
	Bprint(&stdout, " %lux\n", nv->general);
	printitem(ctlr->name, "config");
	Bprint(&stdout, " %lux\n", nv->config);
}

Ctlr nvidia = {
	"nvidia",			/* name */
	snarf,				/* snarf */
	options,			/* options */
	init,				/* init */
	load,				/* load */
	dump,				/* dump */
};

Ctlr nvidiahwgc = {
	"nvidiahwgc",			/* name */
	0,				/* snarf */
	0,				/* options */
	0,				/* init */
	0,				/* load */
	0,				/* dump */
};