path: root/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramnv50.c

/*
 * Copyright 2013 Red Hat Inc.
 *
 * 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 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 HOLDER(S) OR AUTHOR(S) 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: Ben Skeggs
 */

#include <subdev/bios.h>
#include <subdev/bios/bit.h>
#include <subdev/bios/pll.h>
#include <subdev/bios/perf.h>
#include <subdev/bios/timing.h>
#include <subdev/clk/pll.h>
#include <subdev/fb.h>

#include <core/option.h>
#include <core/mm.h>

#include "ramseq.h"

#include "nv50.h"

struct nv50_ramseq {
	struct hwsq base;
	struct hwsq_reg r_0x002504;
	struct hwsq_reg r_0x004008;
	struct hwsq_reg r_0x00400c;
	struct hwsq_reg r_0x00c040;
	struct hwsq_reg r_0x100210;
	struct hwsq_reg r_0x1002d0;
	struct hwsq_reg r_0x1002d4;
	struct hwsq_reg r_0x1002dc;
	struct hwsq_reg r_0x100da0[8];
	struct hwsq_reg r_0x100e20;
	struct hwsq_reg r_0x100e24;
	struct hwsq_reg r_0x611200;
	struct hwsq_reg r_timing[9];
	struct hwsq_reg r_mr[4];
};

struct nv50_ram {
	struct nouveau_ram base;
	struct nv50_ramseq hwsq;
};

#define QFX5800NVA0 1

static int
nv50_ram_calc(struct nouveau_fb *pfb, u32 freq)
{
	struct nouveau_bios *bios = nouveau_bios(pfb);
	struct nv50_ram *ram = (void *)pfb->ram;
	struct nv50_ramseq *hwsq = &ram->hwsq;
	struct nvbios_perfE perfE;
	struct nvbios_pll mpll;
	struct {
		u32 data;
		u8  size;
	} ramcfg, timing;
	u8  ver, hdr, cnt, len, strap;
	int N1, M1, N2, M2, P;
	int ret, i;

	/* lookup closest matching performance table entry for frequency */
	i = 0;
	do {
		ramcfg.data = nvbios_perfEp(bios, i++, &ver, &hdr, &cnt,
					   &ramcfg.size, &perfE);
		if (!ramcfg.data || (ver < 0x25 || ver >= 0x40) ||
		    (ramcfg.size < 2)) {
			nv_error(pfb, "invalid/missing perftab entry\n");
			return -EINVAL;
		}
	} while (perfE.memory < freq);

	/* locate specific data set for the attached memory */
	strap = nvbios_ramcfg_index(nv_subdev(pfb));
	if (strap >= cnt) {
		nv_error(pfb, "invalid ramcfg strap\n");
		return -EINVAL;
	}

	ramcfg.data += hdr + (strap * ramcfg.size);

	/* lookup memory timings, if bios says they're present */
	strap = nv_ro08(bios, ramcfg.data + 0x01);
	if (strap != 0xff) {
		timing.data = nvbios_timingEe(bios, strap, &ver, &hdr,
					     &cnt, &len);
		if (!timing.data || ver != 0x10 || hdr < 0x12) {
			nv_error(pfb, "invalid/missing timing entry "
				 "%02x %04x %02x %02x\n",
				 strap, timing.data, ver, hdr);
			return -EINVAL;
		}
	} else {
		timing.data = 0;
	}

	ret = ram_init(hwsq, nv_subdev(pfb));
	if (ret)
		return ret;

	ram_wait(hwsq, 0x01, 0x00); /* wait for !vblank */
	ram_wait(hwsq, 0x01, 0x01); /* wait for vblank */
	ram_wr32(hwsq, 0x611200, 0x00003300);
	ram_wr32(hwsq, 0x002504, 0x00000001); /* block fifo */
	ram_nsec(hwsq, 8000);
	ram_setf(hwsq, 0x10, 0x00); /* disable fb */
	ram_wait(hwsq, 0x00, 0x01); /* wait for fb disabled */

	ram_wr32(hwsq, 0x1002d4, 0x00000001); /* precharge */
	ram_wr32(hwsq, 0x1002d0, 0x00000001); /* refresh */
	ram_wr32(hwsq, 0x1002d0, 0x00000001); /* refresh */
	ram_wr32(hwsq, 0x100210, 0x00000000); /* disable auto-refresh */
	ram_wr32(hwsq, 0x1002dc, 0x00000001); /* enable self-refresh */

	ret = nvbios_pll_parse(bios, 0x004008, &mpll);
	mpll.vco2.max_freq = 0;
	if (ret == 0) {
		ret = nv04_pll_calc(nv_subdev(pfb), &mpll, freq,
				   &N1, &M1, &N2, &M2, &P);
		if (ret == 0)
			ret = -EINVAL;
	}

	if (ret < 0)
		return ret;

	ram_mask(hwsq, 0x00c040, 0xc000c000, 0x0000c000);
	ram_mask(hwsq, 0x004008, 0x00000200, 0x00000200);
	ram_mask(hwsq, 0x00400c, 0x0000ffff, (N1 << 8) | M1);
	ram_mask(hwsq, 0x004008, 0x81ff0000, 0x80000000 | (mpll.bias_p << 19) |
					     (P << 22) | (P << 16));
#if QFX5800NVA0
	for (i = 0; i < 8; i++)
		ram_mask(hwsq, 0x100da0[i], 0x00000000, 0x00000000); /*XXX*/
#endif
	ram_nsec(hwsq, 96000); /*XXX*/
	ram_mask(hwsq, 0x004008, 0x00002200, 0x00002000);

	ram_wr32(hwsq, 0x1002dc, 0x00000000); /* disable self-refresh */
	ram_wr32(hwsq, 0x100210, 0x80000000); /* enable auto-refresh */

	ram_nsec(hwsq, 12000);

	switch (ram->base.type) {
	case NV_MEM_TYPE_DDR2:
		ram_nuke(hwsq, mr[0]); /* force update */
		ram_mask(hwsq, mr[0], 0x000, 0x000);
		break;
	case NV_MEM_TYPE_GDDR3:
		ram_mask(hwsq, mr[2], 0x000, 0x000);
		ram_nuke(hwsq, mr[0]); /* force update */
		ram_mask(hwsq, mr[0], 0x000, 0x000);
		break;
	default:
		break;
	}

	ram_mask(hwsq, timing[3], 0x00000000, 0x00000000); /*XXX*/
	ram_mask(hwsq, timing[1], 0x00000000, 0x00000000); /*XXX*/
	ram_mask(hwsq, timing[6], 0x00000000, 0x00000000); /*XXX*/
	ram_mask(hwsq, timing[7], 0x00000000, 0x00000000); /*XXX*/
	ram_mask(hwsq, timing[8], 0x00000000, 0x00000000); /*XXX*/
	ram_mask(hwsq, timing[0], 0x00000000, 0x00000000); /*XXX*/
	ram_mask(hwsq, timing[2], 0x00000000, 0x00000000); /*XXX*/
	ram_mask(hwsq, timing[4], 0x00000000, 0x00000000); /*XXX*/
	ram_mask(hwsq, timing[5], 0x00000000, 0x00000000); /*XXX*/

	ram_mask(hwsq, timing[0], 0x00000000, 0x00000000); /*XXX*/

#if QFX5800NVA0
	ram_nuke(hwsq, 0x100e24);
	ram_mask(hwsq, 0x100e24, 0x00000000, 0x00000000);
	ram_nuke(hwsq, 0x100e20);
	ram_mask(hwsq, 0x100e20, 0x00000000, 0x00000000);
#endif

	ram_mask(hwsq, mr[0], 0x100, 0x100);
	ram_mask(hwsq, mr[0], 0x100, 0x000);

	ram_setf(hwsq, 0x10, 0x01); /* enable fb */
	ram_wait(hwsq, 0x00, 0x00); /* wait for fb enabled */
	ram_wr32(hwsq, 0x611200, 0x00003330);
	ram_wr32(hwsq, 0x002504, 0x00000000); /* un-block fifo */
	return 0;
}

static int
nv50_ram_prog(struct nouveau_fb *pfb)
{
	struct nouveau_device *device = nv_device(pfb);
	struct nv50_ram *ram = (void *)pfb->ram;
	struct nv50_ramseq *hwsq = &ram->hwsq;

	ram_exec(hwsq, nouveau_boolopt(device->cfgopt, "NvMemExec", true));
	return 0;
}

static void
nv50_ram_tidy(struct nouveau_fb *pfb)
{
	struct nv50_ram *ram = (void *)pfb->ram;
	struct nv50_ramseq *hwsq = &ram->hwsq;
	ram_exec(hwsq, false);
}

void
__nv50_ram_put(struct nouveau_fb *pfb, struct nouveau_mem *mem)
{
	struct nouveau_mm_node *this;

	while (!list_empty(&mem->regions)) {
		this = list_first_entry(&mem->regions, typeof(*this), rl_entry);

		list_del(&this->rl_entry);
		nouveau_mm_free(&pfb->vram, &this);
	}

	nouveau_mm_free(&pfb->tags, &mem->tag);
}

void
nv50_ram_put(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
{
	struct nouveau_mem *mem = *pmem;

	*pmem = NULL;
	if (unlikely(mem == NULL))
		return;

	mutex_lock(&pfb->base.mutex);
	__nv50_ram_put(pfb, mem);
	mutex_unlock(&pfb->base.mutex);

	kfree(mem);
}

int
nv50_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
	     u32 memtype, struct nouveau_mem **pmem)
{
	struct nouveau_mm *heap = &pfb->vram;
	struct nouveau_mm *tags = &pfb->tags;
	struct nouveau_mm_node *r;
	struct nouveau_mem *mem;
	int comp = (memtype & 0x300) >> 8;
	int type = (memtype & 0x07f);
	int back = (memtype & 0x800);
	int min, max, ret;

	max = (size >> 12);
	min = ncmin ? (ncmin >> 12) : max;
	align >>= 12;

	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
	if (!mem)
		return -ENOMEM;

	mutex_lock(&pfb->base.mutex);
	if (comp) {
		if (align == 16) {
			int n = (max >> 4) * comp;

			ret = nouveau_mm_head(tags, 0, 1, n, n, 1, &mem->tag);
			if (ret)
				mem->tag = NULL;
		}

		if (unlikely(!mem->tag))
			comp = 0;
	}

	INIT_LIST_HEAD(&mem->regions);
	mem->memtype = (comp << 7) | type;
	mem->size = max;

	type = nv50_fb_memtype[type];
	do {
		if (back)
			ret = nouveau_mm_tail(heap, 0, type, max, min, align, &r);
		else
			ret = nouveau_mm_head(heap, 0, type, max, min, align, &r);
		if (ret) {
			mutex_unlock(&pfb->base.mutex);
			pfb->ram->put(pfb, &mem);
			return ret;
		}

		list_add_tail(&r->rl_entry, &mem->regions);
		max -= r->length;
	} while (max);
	mutex_unlock(&pfb->base.mutex);

	r = list_first_entry(&mem->regions, struct nouveau_mm_node, rl_entry);
	mem->offset = (u64)r->offset << 12;
	*pmem = mem;
	return 0;
}

static u32
nv50_fb_vram_rblock(struct nouveau_fb *pfb, struct nouveau_ram *ram)
{
	int colbits, rowbitsa, rowbitsb, banks;
	u64 rowsize, predicted;
	u32 r0, r4, rt, rblock_size;

	r0 = nv_rd32(pfb, 0x100200);
	r4 = nv_rd32(pfb, 0x100204);
	rt = nv_rd32(pfb, 0x100250);
	nv_debug(pfb, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt,
			nv_rd32(pfb, 0x001540));

	colbits  =  (r4 & 0x0000f000) >> 12;
	rowbitsa = ((r4 & 0x000f0000) >> 16) + 8;
	rowbitsb = ((r4 & 0x00f00000) >> 20) + 8;
	banks    = 1 << (((r4 & 0x03000000) >> 24) + 2);

	rowsize = ram->parts * banks * (1 << colbits) * 8;
	predicted = rowsize << rowbitsa;
	if (r0 & 0x00000004)
		predicted += rowsize << rowbitsb;

	if (predicted != ram->size) {
		nv_warn(pfb, "memory controller reports %d MiB VRAM\n",
			(u32)(ram->size >> 20));
	}

	rblock_size = rowsize;
	if (rt & 1)
		rblock_size *= 3;

	nv_debug(pfb, "rblock %d bytes\n", rblock_size);
	return rblock_size;
}

int
nv50_ram_create_(struct nouveau_object *parent, struct nouveau_object *engine,
		 struct nouveau_oclass *oclass, int length, void **pobject)
{
	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
	struct nouveau_bios *bios = nouveau_bios(parent);
	struct nouveau_fb *pfb = nouveau_fb(parent);
	struct nouveau_ram *ram;
	int ret;

	ret = nouveau_ram_create_(parent, engine, oclass, length, pobject);
	ram = *pobject;
	if (ret)
		return ret;

	ram->size = nv_rd32(pfb, 0x10020c);
	ram->size = (ram->size & 0xffffff00) | ((ram->size & 0x000000ff) << 32);

	ram->part_mask = (nv_rd32(pfb, 0x001540) & 0x00ff0000) >> 16;
	ram->parts = hweight8(ram->part_mask);

	switch (nv_rd32(pfb, 0x100714) & 0x00000007) {
	case 0: ram->type = NV_MEM_TYPE_DDR1; break;
	case 1:
		if (nouveau_fb_bios_memtype(bios) == NV_MEM_TYPE_DDR3)
			ram->type = NV_MEM_TYPE_DDR3;
		else
			ram->type = NV_MEM_TYPE_DDR2;
		break;
	case 2: ram->type = NV_MEM_TYPE_GDDR3; break;
	case 3: ram->type = NV_MEM_TYPE_GDDR4; break;
	case 4: ram->type = NV_MEM_TYPE_GDDR5; break;
	default:
		break;
	}

	ret = nouveau_mm_init(&pfb->vram, rsvd_head, (ram->size >> 12) -
			      (rsvd_head + rsvd_tail),
			      nv50_fb_vram_rblock(pfb, ram) >> 12);
	if (ret)
		return ret;

	ram->ranks = (nv_rd32(pfb, 0x100200) & 0x4) ? 2 : 1;
	ram->tags  =  nv_rd32(pfb, 0x100320);
	ram->get = nv50_ram_get;
	ram->put = nv50_ram_put;
	return 0;
}

static int
nv50_ram_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
	      struct nouveau_oclass *oclass, void *data, u32 datasize,
	      struct nouveau_object **pobject)
{
	struct nv50_ram *ram;
	int ret, i;

	ret = nv50_ram_create(parent, engine, oclass, &ram);
	*pobject = nv_object(ram);
	if (ret)
		return ret;

	switch (ram->base.type) {
	case NV_MEM_TYPE_DDR2:
	case NV_MEM_TYPE_GDDR3:
		ram->base.calc = nv50_ram_calc;
		ram->base.prog = nv50_ram_prog;
		ram->base.tidy = nv50_ram_tidy;
		break;
	default:
		nv_warn(ram, "reclocking of this ram type unsupported\n");
		return 0;
	}

	ram->hwsq.r_0x002504 = hwsq_reg(0x002504);
	ram->hwsq.r_0x00c040 = hwsq_reg(0x00c040);
	ram->hwsq.r_0x004008 = hwsq_reg(0x004008);
	ram->hwsq.r_0x00400c = hwsq_reg(0x00400c);
	ram->hwsq.r_0x100210 = hwsq_reg(0x100210);
	ram->hwsq.r_0x1002d0 = hwsq_reg(0x1002d0);
	ram->hwsq.r_0x1002d4 = hwsq_reg(0x1002d4);
	ram->hwsq.r_0x1002dc = hwsq_reg(0x1002dc);
	for (i = 0; i < 8; i++)
		ram->hwsq.r_0x100da0[i] = hwsq_reg(0x100da0 + (i * 0x04));
	ram->hwsq.r_0x100e20 = hwsq_reg(0x100e20);
	ram->hwsq.r_0x100e24 = hwsq_reg(0x100e24);
	ram->hwsq.r_0x611200 = hwsq_reg(0x611200);

	for (i = 0; i < 9; i++)
		ram->hwsq.r_timing[i] = hwsq_reg(0x100220 + (i * 0x04));

	if (ram->base.ranks > 1) {
		ram->hwsq.r_mr[0] = hwsq_reg2(0x1002c0, 0x1002c8);
		ram->hwsq.r_mr[1] = hwsq_reg2(0x1002c4, 0x1002cc);
		ram->hwsq.r_mr[2] = hwsq_reg2(0x1002e0, 0x1002e8);
		ram->hwsq.r_mr[3] = hwsq_reg2(0x1002e4, 0x1002ec);
	} else {
		ram->hwsq.r_mr[0] = hwsq_reg(0x1002c0);
		ram->hwsq.r_mr[1] = hwsq_reg(0x1002c4);
		ram->hwsq.r_mr[2] = hwsq_reg(0x1002e0);
		ram->hwsq.r_mr[3] = hwsq_reg(0x1002e4);
	}

	return 0;
}

struct nouveau_oclass
nv50_ram_oclass = {
	.ofuncs = &(struct nouveau_ofuncs) {
		.ctor = nv50_ram_ctor,
		.dtor = _nouveau_ram_dtor,
		.init = _nouveau_ram_init,
		.fini = _nouveau_ram_fini,
	}
};