cx88-core.c

V4l driver for DVB HD

/*
 *
 * device driver for Conexant 2388x based TV cards
 * driver core
 *
 * (c) 2003 Gerd Knorr <[e-mail]> [SuSE Labs]
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, write to the Free Software
 *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */

#include <linux/init.h>
#include <linux/list.h>
#include <linux/module.h>
#include <linux/moduleparam.h>
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/kmod.h>
#include <linux/sound.h>
#include <linux/interrupt.h>
#include <linux/pci.h>
#include <linux/delay.h>
#include "compat.h"
#include <linux/videodev2.h>
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,15)
#include <linux/mutex.h>
#endif

#include "cx88.h"
#include <media/v4l2-common.h>

MODULE_DESCRIPTION("v4l2 driver module for cx2388x based TV cards");
MODULE_AUTHOR("Gerd Knorr <[e-mail]> [SuSE Labs]");
MODULE_LICENSE("GPL");

/* ------------------------------------------------------------------ */

static unsigned int core_debug = 0;
module_param(core_debug,int,0644);
MODULE_PARM_DESC(core_debug,"enable debug messages [core]");

static unsigned int latency = UNSET;
module_param(latency,int,0444);
MODULE_PARM_DESC(latency,"pci latency timer");

static unsigned int tuner[] = {[0 ... (CX88_MAXBOARDS - 1)] = UNSET };
static unsigned int radio[] = {[0 ... (CX88_MAXBOARDS - 1)] = UNSET };
static unsigned int card[]  = {[0 ... (CX88_MAXBOARDS - 1)] = UNSET };

#if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)
MODULE_PARM(tuner,"1-" __stringify(CX88_MAXBOARDS) "i");
MODULE_PARM(radiox,"1-" __stringify(CX88_MAXBOARDS) "i");
MODULE_PARM(card,"1-" __stringify(CX88_MAXBOARDS) "i");
#else
#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
static int dummy;
module_param_array(tuner, int, dummy, 0444);
module_param_array(radio, int, dummy, 0444);
module_param_array(card,  int, dummy, 0444);
#else
module_param_array(tuner, int, NULL, 0444);
module_param_array(radio, int, NULL, 0444);
module_param_array(card,  int, NULL, 0444);
#endif
#endif

MODULE_PARM_DESC(tuner,"tuner type");
MODULE_PARM_DESC(radio,"radio tuner type");
MODULE_PARM_DESC(card,"card type");

static unsigned int nicam = 0;
module_param(nicam,int,0644);
MODULE_PARM_DESC(nicam,"tv audio is nicam");

static unsigned int nocomb = 0;
module_param(nocomb,int,0644);
MODULE_PARM_DESC(nocomb,"disable comb filter");

#define dprintk(level,fmt, arg...)	if (core_debug >= level)	\
	printk(KERN_DEBUG "%s: " fmt, core->name , ## arg)

static unsigned int cx88_devcount;
static LIST_HEAD(cx88_devlist);
static DEFINE_MUTEX(devlist);

#define NO_SYNC_LINE (-1U)

static u32* cx88_risc_field(u32 *rp, struct scatterlist *sglist,
			    unsigned int offset, u32 sync_line,
			    unsigned int bpl, unsigned int padding,
			    unsigned int lines)
{
	struct scatterlist *sg;
	unsigned int line,todo;

	/* sync instruction */
	if (sync_line != NO_SYNC_LINE)
		*(rp++) = cpu_to_le32(RISC_RESYNC | sync_line);

	/* scan lines */
	sg = sglist;
	for (line = 0; line < lines; line++) {
		while (offset && offset >= sg_dma_len(sg)) {
			offset -= sg_dma_len(sg);
			sg++;
		}
		if (bpl <= sg_dma_len(sg)-offset) {
			/* fits into current chunk */
			*(rp++)=cpu_to_le32(RISC_WRITE|RISC_SOL|RISC_EOL|bpl);
			*(rp++)=cpu_to_le32(sg_dma_address(sg)+offset);
			offset+=bpl;
		} else {
			/* scanline needs to be splitted */
			todo = bpl;
			*(rp++)=cpu_to_le32(RISC_WRITE|RISC_SOL|
					    (sg_dma_len(sg)-offset));
			*(rp++)=cpu_to_le32(sg_dma_address(sg)+offset);
			todo -= (sg_dma_len(sg)-offset);
			offset = 0;
			sg++;
			while (todo > sg_dma_len(sg)) {
				*(rp++)=cpu_to_le32(RISC_WRITE|
						    sg_dma_len(sg));
				*(rp++)=cpu_to_le32(sg_dma_address(sg));
				todo -= sg_dma_len(sg);
				sg++;
			}
			*(rp++)=cpu_to_le32(RISC_WRITE|RISC_EOL|todo);
			*(rp++)=cpu_to_le32(sg_dma_address(sg));
			offset += todo;
		}
		offset += padding;
	}

	return rp;
}

int cx88_risc_buffer(struct pci_dev *pci, struct btcx_riscmem *risc,
		     struct scatterlist *sglist,
		     unsigned int top_offset, unsigned int bottom_offset,
		     unsigned int bpl, unsigned int padding, unsigned int lines)
{
	u32 instructions,fields;
	u32 *rp;
	int rc;

	fields = 0;
	if (UNSET != top_offset)
		fields++;
	if (UNSET != bottom_offset)
		fields++;

	/* estimate risc mem: worst case is one write per page border +
	   one write per scan line + syncs + jump (all 2 dwords).  Padding
	   can cause next bpl to start close to a page border.  First DMA
	   region may be smaller than PAGE_SIZE */
	instructions  = fields * (1 + ((bpl + padding) * lines) / PAGE_SIZE + lines);
	instructions += 2;
	if ((rc = btcx_riscmem_alloc(pci,risc,instructions*8)) < 0)
		return rc;

	/* write risc instructions */
	rp = risc->cpu;
	if (UNSET != top_offset)
		rp = cx88_risc_field(rp, sglist, top_offset, 0,
				     bpl, padding, lines);
	if (UNSET != bottom_offset)
		rp = cx88_risc_field(rp, sglist, bottom_offset, 0x200,
				     bpl, padding, lines);

	/* save pointer to jmp instruction address */
	risc->jmp = rp;
	BUG_ON((risc->jmp - risc->cpu + 2) * sizeof (*risc->cpu) > risc->size);
	return 0;
}

int cx88_risc_databuffer(struct pci_dev *pci, struct btcx_riscmem *risc,
			 struct scatterlist *sglist, unsigned int bpl,
			 unsigned int lines)
{
	u32 instructions;
	u32 *rp;
	int rc;

	/* estimate risc mem: worst case is one write per page border +
	   one write per scan line + syncs + jump (all 2 dwords).  Here
	   there is no padding and no sync.  First DMA region may be smaller
	   than PAGE_SIZE */
	instructions  = 1 + (bpl * lines) / PAGE_SIZE + lines;
	instructions += 1;
	if ((rc = btcx_riscmem_alloc(pci,risc,instructions*8)) < 0)
		return rc;

	/* write risc instructions */
	rp = risc->cpu;
	rp = cx88_risc_field(rp, sglist, 0, NO_SYNC_LINE, bpl, 0, lines);

	/* save pointer to jmp instruction address */
	risc->jmp = rp;
	BUG_ON((risc->jmp - risc->cpu + 2) * sizeof (*risc->cpu) > risc->size);
	return 0;
}

int cx88_risc_stopper(struct pci_dev *pci, struct btcx_riscmem *risc,
		      u32 reg, u32 mask, u32 value)
{
	u32 *rp;
	int rc;

	if ((rc = btcx_riscmem_alloc(pci, risc, 4*16)) < 0)
		return rc;

	/* write risc instructions */
	rp = risc->cpu;
	*(rp++) = cpu_to_le32(RISC_WRITECR  | RISC_IRQ2 | RISC_IMM);
	*(rp++) = cpu_to_le32(reg);
	*(rp++) = cpu_to_le32(value);
	*(rp++) = cpu_to_le32(mask);
	*(rp++) = cpu_to_le32(RISC_JUMP);
	*(rp++) = cpu_to_le32(risc->dma);
	return 0;
}

void
cx88_free_buffer(struct videobuf_queue *q, struct cx88_buffer *buf)
{
	BUG_ON(in_interrupt());
	videobuf_waiton(&buf->vb,0,0);
	videobuf_dma_unmap(q, &buf->vb.dma);
	videobuf_dma_free(&buf->vb.dma);
	btcx_riscmem_free((struct pci_dev *)q->dev, &buf->risc);
	buf->vb.state = STATE_NEEDS_INIT;
}

/* ------------------------------------------------------------------ */
/* our SRAM memory layout                                             */

/* we are going to put all thr risc programs into host memory, so we
 * can use the whole SDRAM for the DMA fifos.  To simplify things, we
 * use a static memory layout.  That surely will waste memory in case
 * we don't use all DMA channels at the same time (which will be the
 * case most of the time).  But that still gives us enougth FIFO space
 * to be able to deal with insane long pci latencies ...
 *
 * FIFO space allocations:
 *    channel  21    (y video)  - 10.0k
 *    channel  22    (u video)  -  2.0k
 *    channel  23    (v video)  -  2.0k
 *    channel  24    (vbi)      -  4.0k
 *    channels 25+26 (audio)    -  4.0k
 *    channel  28    (mpeg)     -  4.0k
 *    TOTAL                     = 29.0k
 *
 * Every channel has 160 bytes control data (64 bytes instruction
 * queue and 6 CDT entries), which is close to 2k total.
 *
 * Address layout:
 *    0x0000 - 0x03ff    CMDs / reserved
 *    0x0400 - 0x0bff    instruction queues + CDs
 *    0x0c00 -           FIFOs
 */

struct sram_channel cx88_sram_channels[] = {
	[SRAM_CH21] = {
		.name       = "video y / packed",
		.cmds_start = 0x180040,
		.ctrl_start = 0x180400,
		.cdt        = 0x180400 + 64,
		.fifo_start = 0x180c00,
		.fifo_size  = 0x002800,
		.ptr1_reg   = MO_DMA21_PTR1,
		.ptr2_reg   = MO_DMA21_PTR2,
		.cnt1_reg   = MO_DMA21_CNT1,
		.cnt2_reg   = MO_DMA21_CNT2,
	},
	[SRAM_CH22] = {
		.name       = "video u",
		.cmds_start = 0x180080,
		.ctrl_start = 0x1804a0,
		.cdt        = 0x1804a0 + 64,
		.fifo_start = 0x183400,
		.fifo_size  = 0x000800,
		.ptr1_reg   = MO_DMA22_PTR1,
		.ptr2_reg   = MO_DMA22_PTR2,
		.cnt1_reg   = MO_DMA22_CNT1,
		.cnt2_reg   = MO_DMA22_CNT2,
	},
	[SRAM_CH23] = {
		.name       = "video v",
		.cmds_start = 0x1800c0,
		.ctrl_start = 0x180540,
		.cdt        = 0x180540 + 64,
		.fifo_start = 0x183c00,
		.fifo_size  = 0x000800,
		.ptr1_reg   = MO_DMA23_PTR1,
		.ptr2_reg   = MO_DMA23_PTR2,
		.cnt1_reg   = MO_DMA23_CNT1,
		.cnt2_reg   = MO_DMA23_CNT2,
	},
	[SRAM_CH24] = {
		.name       = "vbi",
		.cmds_start = 0x180100,
		.ctrl_start = 0x1805e0,
		.cdt        = 0x1805e0 + 64,
		.fifo_start = 0x184400,
		.fifo_size  = 0x001000,
		.ptr1_reg   = MO_DMA24_PTR1,
		.ptr2_reg   = MO_DMA24_PTR2,
		.cnt1_reg   = MO_DMA24_CNT1,
		.cnt2_reg   = MO_DMA24_CNT2,
	},
	[SRAM_CH25] = {
		.name       = "audio from",
		.cmds_start = 0x180140,
		.ctrl_start = 0x180680,
		.cdt        = 0x180680 + 64,
		.fifo_start = 0x185400,
		.fifo_size  = 0x001000,
		.ptr1_reg   = MO_DMA25_PTR1,
		.ptr2_reg   = MO_DMA25_PTR2,
		.cnt1_reg   = MO_DMA25_CNT1,
		.cnt2_reg   = MO_DMA25_CNT2,
	},
	[SRAM_CH26] = {
		.name       = "audio to",
		.cmds_start = 0x180180,
		.ctrl_start = 0x180720,
		.cdt        = 0x180680 + 64,  /* same as audio IN */
		.fifo_start = 0x185400,       /* same as audio IN */
		.fifo_size  = 0x001000,       /* same as audio IN */
		.ptr1_reg   = MO_DMA26_PTR1,
		.ptr2_reg   = MO_DMA26_PTR2,
		.cnt1_reg   = MO_DMA26_CNT1,
		.cnt2_reg   = MO_DMA26_CNT2,
	},
	[SRAM_CH28] = {
		.name       = "mpeg",
		.cmds_start = 0x180200,
		.ctrl_start = 0x1807C0,
		.cdt        = 0x1807C0 + 64,
		.fifo_start = 0x186400,
		.fifo_size  = 0x001000,
		.ptr1_reg   = MO_DMA28_PTR1,
		.ptr2_reg   = MO_DMA28_PTR2,
		.cnt1_reg   = MO_DMA28_CNT1,
		.cnt2_reg   = MO_DMA28_CNT2,
	},
};

int cx88_sram_channel_setup(struct cx88_core *core,
			    struct sram_channel *ch,
			    unsigned int bpl, u32 risc)
{
	unsigned int i,lines;
	u32 cdt;

	bpl   = (bpl + 7) & ~7; /* alignment */
	cdt   = ch->cdt;
	lines = ch->fifo_size / bpl;
	if (lines > 6)
		lines = 6;
	BUG_ON(lines < 2);

	/* write CDT */
	for (i = 0; i < lines; i++)
		cx_write(cdt + 16*i, ch->fifo_start + bpl*i);

	/* write CMDS */
	cx_write(ch->cmds_start +  0, risc);
	cx_write(ch->cmds_start +  4, cdt);
	cx_write(ch->cmds_start +  8, (lines*16) >> 3);
	cx_write(ch->cmds_start + 12, ch->ctrl_start);
	cx_write(ch->cmds_start + 16, 64 >> 2);
	for (i = 20; i < 64; i += 4)
		cx_write(ch->cmds_start + i, 0);

	/* fill registers */
	cx_write(ch->ptr1_reg, ch->fifo_start);
	cx_write(ch->ptr2_reg, cdt);
	cx_write(ch->cnt1_reg, (bpl >> 3) -1);
	cx_write(ch->cnt2_reg, (lines*16) >> 3);

	dprintk(2,"sram setup %s: bpl=%d lines=%d\n", ch->name, bpl, lines);
	return 0;
}

/* ------------------------------------------------------------------ */
/* debug helper code                                                  */

static int cx88_risc_decode(u32 risc)
{
	static char *instr[16] = {
		[ RISC_SYNC    >> 28 ] = "sync",
		[ RISC_WRITE   >> 28 ] = "write",
		[ RISC_WRITEC  >> 28 ] = "writec",
		[ RISC_READ    >> 28 ] = "read",
		[ RISC_READC   >> 28 ] = "readc",
		[ RISC_JUMP    >> 28 ] = "jump",
		[ RISC_SKIP    >> 28 ] = "skip",
		[ RISC_WRITERM >> 28 ] = "writerm",
		[ RISC_WRITECM >> 28 ] = "writecm",
		[ RISC_WRITECR >> 28 ] = "writecr",