linux_old1/sound/pci/emu10k1/emu10k1_callback.c

551 lines
14 KiB
C

/*
* synth callback routines for Emu10k1
*
* Copyright (C) 2000 Takashi Iwai <tiwai@suse.de>
*
* 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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
#include <linux/export.h>
#include "emu10k1_synth_local.h"
#include <sound/asoundef.h>
/* voice status */
enum {
V_FREE=0, V_OFF, V_RELEASED, V_PLAYING, V_END
};
/* Keeps track of what we are finding */
struct best_voice {
unsigned int time;
int voice;
};
/*
* prototypes
*/
static void lookup_voices(struct snd_emux *emux, struct snd_emu10k1 *hw,
struct best_voice *best, int active_only);
static struct snd_emux_voice *get_voice(struct snd_emux *emux,
struct snd_emux_port *port);
static int start_voice(struct snd_emux_voice *vp);
static void trigger_voice(struct snd_emux_voice *vp);
static void release_voice(struct snd_emux_voice *vp);
static void update_voice(struct snd_emux_voice *vp, int update);
static void terminate_voice(struct snd_emux_voice *vp);
static void free_voice(struct snd_emux_voice *vp);
static void set_fmmod(struct snd_emu10k1 *hw, struct snd_emux_voice *vp);
static void set_fm2frq2(struct snd_emu10k1 *hw, struct snd_emux_voice *vp);
static void set_filterQ(struct snd_emu10k1 *hw, struct snd_emux_voice *vp);
/*
* Ensure a value is between two points
* macro evaluates its args more than once, so changed to upper-case.
*/
#define LIMITVALUE(x, a, b) do { if ((x) < (a)) (x) = (a); else if ((x) > (b)) (x) = (b); } while (0)
#define LIMITMAX(x, a) do {if ((x) > (a)) (x) = (a); } while (0)
/*
* set up operators
*/
static const struct snd_emux_operators emu10k1_ops = {
.owner = THIS_MODULE,
.get_voice = get_voice,
.prepare = start_voice,
.trigger = trigger_voice,
.release = release_voice,
.update = update_voice,
.terminate = terminate_voice,
.free_voice = free_voice,
.sample_new = snd_emu10k1_sample_new,
.sample_free = snd_emu10k1_sample_free,
};
void
snd_emu10k1_ops_setup(struct snd_emux *emux)
{
emux->ops = emu10k1_ops;
}
/*
* get more voice for pcm
*
* terminate most inactive voice and give it as a pcm voice.
*
* voice_lock is already held.
*/
int
snd_emu10k1_synth_get_voice(struct snd_emu10k1 *hw)
{
struct snd_emux *emu;
struct snd_emux_voice *vp;
struct best_voice best[V_END];
int i;
emu = hw->synth;
lookup_voices(emu, hw, best, 1); /* no OFF voices */
for (i = 0; i < V_END; i++) {
if (best[i].voice >= 0) {
int ch;
vp = &emu->voices[best[i].voice];
if ((ch = vp->ch) < 0) {
/*
dev_warn(emu->card->dev,
"synth_get_voice: ch < 0 (%d) ??", i);
*/
continue;
}
vp->emu->num_voices--;
vp->ch = -1;
vp->state = SNDRV_EMUX_ST_OFF;
return ch;
}
}
/* not found */
return -ENOMEM;
}
/*
* turn off the voice (not terminated)
*/
static void
release_voice(struct snd_emux_voice *vp)
{
int dcysusv;
struct snd_emu10k1 *hw;
hw = vp->hw;
dcysusv = 0x8000 | (unsigned char)vp->reg.parm.modrelease;
snd_emu10k1_ptr_write(hw, DCYSUSM, vp->ch, dcysusv);
dcysusv = 0x8000 | (unsigned char)vp->reg.parm.volrelease | DCYSUSV_CHANNELENABLE_MASK;
snd_emu10k1_ptr_write(hw, DCYSUSV, vp->ch, dcysusv);
}
/*
* terminate the voice
*/
static void
terminate_voice(struct snd_emux_voice *vp)
{
struct snd_emu10k1 *hw;
if (snd_BUG_ON(!vp))
return;
hw = vp->hw;
snd_emu10k1_ptr_write(hw, DCYSUSV, vp->ch, 0x807f | DCYSUSV_CHANNELENABLE_MASK);
if (vp->block) {
struct snd_emu10k1_memblk *emem;
emem = (struct snd_emu10k1_memblk *)vp->block;
if (emem->map_locked > 0)
emem->map_locked--;
}
}
/*
* release the voice to system
*/
static void
free_voice(struct snd_emux_voice *vp)
{
struct snd_emu10k1 *hw;
hw = vp->hw;
/* FIXME: emu10k1_synth is broken. */
/* This can get called with hw == 0 */
/* Problem apparent on plug, unplug then plug */
/* on the Audigy 2 ZS Notebook. */
if (hw && (vp->ch >= 0)) {
snd_emu10k1_ptr_write(hw, IFATN, vp->ch, 0xff00);
snd_emu10k1_ptr_write(hw, DCYSUSV, vp->ch, 0x807f | DCYSUSV_CHANNELENABLE_MASK);
// snd_emu10k1_ptr_write(hw, DCYSUSV, vp->ch, 0);
snd_emu10k1_ptr_write(hw, VTFT, vp->ch, 0xffff);
snd_emu10k1_ptr_write(hw, CVCF, vp->ch, 0xffff);
snd_emu10k1_voice_free(hw, &hw->voices[vp->ch]);
vp->emu->num_voices--;
vp->ch = -1;
}
}
/*
* update registers
*/
static void
update_voice(struct snd_emux_voice *vp, int update)
{
struct snd_emu10k1 *hw;
hw = vp->hw;
if (update & SNDRV_EMUX_UPDATE_VOLUME)
snd_emu10k1_ptr_write(hw, IFATN_ATTENUATION, vp->ch, vp->avol);
if (update & SNDRV_EMUX_UPDATE_PITCH)
snd_emu10k1_ptr_write(hw, IP, vp->ch, vp->apitch);
if (update & SNDRV_EMUX_UPDATE_PAN) {
snd_emu10k1_ptr_write(hw, PTRX_FXSENDAMOUNT_A, vp->ch, vp->apan);
snd_emu10k1_ptr_write(hw, PTRX_FXSENDAMOUNT_B, vp->ch, vp->aaux);
}
if (update & SNDRV_EMUX_UPDATE_FMMOD)
set_fmmod(hw, vp);
if (update & SNDRV_EMUX_UPDATE_TREMFREQ)
snd_emu10k1_ptr_write(hw, TREMFRQ, vp->ch, vp->reg.parm.tremfrq);
if (update & SNDRV_EMUX_UPDATE_FM2FRQ2)
set_fm2frq2(hw, vp);
if (update & SNDRV_EMUX_UPDATE_Q)
set_filterQ(hw, vp);
}
/*
* look up voice table - get the best voice in order of preference
*/
/* spinlock held! */
static void
lookup_voices(struct snd_emux *emu, struct snd_emu10k1 *hw,
struct best_voice *best, int active_only)
{
struct snd_emux_voice *vp;
struct best_voice *bp;
int i;
for (i = 0; i < V_END; i++) {
best[i].time = (unsigned int)-1; /* XXX MAX_?INT really */
best[i].voice = -1;
}
/*
* Go through them all and get a best one to use.
* NOTE: could also look at volume and pick the quietest one.
*/
for (i = 0; i < emu->max_voices; i++) {
int state, val;
vp = &emu->voices[i];
state = vp->state;
if (state == SNDRV_EMUX_ST_OFF) {
if (vp->ch < 0) {
if (active_only)
continue;
bp = best + V_FREE;
} else
bp = best + V_OFF;
}
else if (state == SNDRV_EMUX_ST_RELEASED ||
state == SNDRV_EMUX_ST_PENDING) {
bp = best + V_RELEASED;
#if 1
val = snd_emu10k1_ptr_read(hw, CVCF_CURRENTVOL, vp->ch);
if (! val)
bp = best + V_OFF;
#endif
}
else if (state == SNDRV_EMUX_ST_STANDBY)
continue;
else if (state & SNDRV_EMUX_ST_ON)
bp = best + V_PLAYING;
else
continue;
/* check if sample is finished playing (non-looping only) */
if (bp != best + V_OFF && bp != best + V_FREE &&
(vp->reg.sample_mode & SNDRV_SFNT_SAMPLE_SINGLESHOT)) {
val = snd_emu10k1_ptr_read(hw, CCCA_CURRADDR, vp->ch);
if (val >= vp->reg.loopstart)
bp = best + V_OFF;
}
if (vp->time < bp->time) {
bp->time = vp->time;
bp->voice = i;
}
}
}
/*
* get an empty voice
*
* emu->voice_lock is already held.
*/
static struct snd_emux_voice *
get_voice(struct snd_emux *emu, struct snd_emux_port *port)
{
struct snd_emu10k1 *hw;
struct snd_emux_voice *vp;
struct best_voice best[V_END];
int i;
hw = emu->hw;
lookup_voices(emu, hw, best, 0);
for (i = 0; i < V_END; i++) {
if (best[i].voice >= 0) {
vp = &emu->voices[best[i].voice];
if (vp->ch < 0) {
/* allocate a voice */
struct snd_emu10k1_voice *hwvoice;
if (snd_emu10k1_voice_alloc(hw, EMU10K1_SYNTH, 1, &hwvoice) < 0 || hwvoice == NULL)
continue;
vp->ch = hwvoice->number;
emu->num_voices++;
}
return vp;
}
}
/* not found */
return NULL;
}
/*
* prepare envelopes and LFOs
*/
static int
start_voice(struct snd_emux_voice *vp)
{
unsigned int temp;
int ch;
unsigned int addr, mapped_offset;
struct snd_midi_channel *chan;
struct snd_emu10k1 *hw;
struct snd_emu10k1_memblk *emem;
hw = vp->hw;
ch = vp->ch;
if (snd_BUG_ON(ch < 0))
return -EINVAL;
chan = vp->chan;
emem = (struct snd_emu10k1_memblk *)vp->block;
if (emem == NULL)
return -EINVAL;
emem->map_locked++;
if (snd_emu10k1_memblk_map(hw, emem) < 0) {
/* dev_err(hw->card->devK, "emu: cannot map!\n"); */
return -ENOMEM;
}
mapped_offset = snd_emu10k1_memblk_offset(emem) >> 1;
vp->reg.start += mapped_offset;
vp->reg.end += mapped_offset;
vp->reg.loopstart += mapped_offset;
vp->reg.loopend += mapped_offset;
/* set channel routing */
/* A = left(0), B = right(1), C = reverb(c), D = chorus(d) */
if (hw->audigy) {
temp = FXBUS_MIDI_LEFT | (FXBUS_MIDI_RIGHT << 8) |
(FXBUS_MIDI_REVERB << 16) | (FXBUS_MIDI_CHORUS << 24);
snd_emu10k1_ptr_write(hw, A_FXRT1, ch, temp);
} else {
temp = (FXBUS_MIDI_LEFT << 16) | (FXBUS_MIDI_RIGHT << 20) |
(FXBUS_MIDI_REVERB << 24) | (FXBUS_MIDI_CHORUS << 28);
snd_emu10k1_ptr_write(hw, FXRT, ch, temp);
}
/* channel to be silent and idle */
snd_emu10k1_ptr_write(hw, DCYSUSV, ch, 0x0000);
snd_emu10k1_ptr_write(hw, VTFT, ch, 0x0000FFFF);
snd_emu10k1_ptr_write(hw, CVCF, ch, 0x0000FFFF);
snd_emu10k1_ptr_write(hw, PTRX, ch, 0);
snd_emu10k1_ptr_write(hw, CPF, ch, 0);
/* set pitch offset */
snd_emu10k1_ptr_write(hw, IP, vp->ch, vp->apitch);
/* set envelope parameters */
snd_emu10k1_ptr_write(hw, ENVVAL, ch, vp->reg.parm.moddelay);
snd_emu10k1_ptr_write(hw, ATKHLDM, ch, vp->reg.parm.modatkhld);
snd_emu10k1_ptr_write(hw, DCYSUSM, ch, vp->reg.parm.moddcysus);
snd_emu10k1_ptr_write(hw, ENVVOL, ch, vp->reg.parm.voldelay);
snd_emu10k1_ptr_write(hw, ATKHLDV, ch, vp->reg.parm.volatkhld);
/* decay/sustain parameter for volume envelope is used
for triggerg the voice */
/* cutoff and volume */
temp = (unsigned int)vp->acutoff << 8 | (unsigned char)vp->avol;
snd_emu10k1_ptr_write(hw, IFATN, vp->ch, temp);
/* modulation envelope heights */
snd_emu10k1_ptr_write(hw, PEFE, ch, vp->reg.parm.pefe);
/* lfo1/2 delay */
snd_emu10k1_ptr_write(hw, LFOVAL1, ch, vp->reg.parm.lfo1delay);
snd_emu10k1_ptr_write(hw, LFOVAL2, ch, vp->reg.parm.lfo2delay);
/* lfo1 pitch & cutoff shift */
set_fmmod(hw, vp);
/* lfo1 volume & freq */
snd_emu10k1_ptr_write(hw, TREMFRQ, vp->ch, vp->reg.parm.tremfrq);
/* lfo2 pitch & freq */
set_fm2frq2(hw, vp);
/* reverb and loop start (reverb 8bit, MSB) */
temp = vp->reg.parm.reverb;
temp += (int)vp->chan->control[MIDI_CTL_E1_REVERB_DEPTH] * 9 / 10;
LIMITMAX(temp, 255);
addr = vp->reg.loopstart;
snd_emu10k1_ptr_write(hw, PSST, vp->ch, (temp << 24) | addr);
/* chorus & loop end (chorus 8bit, MSB) */
addr = vp->reg.loopend;
temp = vp->reg.parm.chorus;
temp += (int)chan->control[MIDI_CTL_E3_CHORUS_DEPTH] * 9 / 10;
LIMITMAX(temp, 255);
temp = (temp <<24) | addr;
snd_emu10k1_ptr_write(hw, DSL, ch, temp);
/* clear filter delay memory */
snd_emu10k1_ptr_write(hw, Z1, ch, 0);
snd_emu10k1_ptr_write(hw, Z2, ch, 0);
/* invalidate maps */
temp = (hw->silent_page.addr << hw->address_mode) | (hw->address_mode ? MAP_PTI_MASK1 : MAP_PTI_MASK0);
snd_emu10k1_ptr_write(hw, MAPA, ch, temp);
snd_emu10k1_ptr_write(hw, MAPB, ch, temp);
#if 0
/* cache */
{
unsigned int val, sample;
val = 32;
if (vp->reg.sample_mode & SNDRV_SFNT_SAMPLE_8BITS)
sample = 0x80808080;
else {
sample = 0;
val *= 2;
}
/* cache */
snd_emu10k1_ptr_write(hw, CCR, ch, 0x1c << 16);
snd_emu10k1_ptr_write(hw, CDE, ch, sample);
snd_emu10k1_ptr_write(hw, CDF, ch, sample);
/* invalidate maps */
temp = ((unsigned int)hw->silent_page.addr << hw_address_mode) | (hw->address_mode ? MAP_PTI_MASK1 : MAP_PTI_MASK0);
snd_emu10k1_ptr_write(hw, MAPA, ch, temp);
snd_emu10k1_ptr_write(hw, MAPB, ch, temp);
/* fill cache */
val -= 4;
val <<= 25;
val |= 0x1c << 16;
snd_emu10k1_ptr_write(hw, CCR, ch, val);
}
#endif
/* Q & current address (Q 4bit value, MSB) */
addr = vp->reg.start;
temp = vp->reg.parm.filterQ;
temp = (temp<<28) | addr;
if (vp->apitch < 0xe400)
temp |= CCCA_INTERPROM_0;
else {
unsigned int shift = (vp->apitch - 0xe000) >> 10;
temp |= shift << 25;
}
if (vp->reg.sample_mode & SNDRV_SFNT_SAMPLE_8BITS)
temp |= CCCA_8BITSELECT;
snd_emu10k1_ptr_write(hw, CCCA, ch, temp);
/* reset volume */
temp = (unsigned int)vp->vtarget << 16;
snd_emu10k1_ptr_write(hw, VTFT, ch, temp | vp->ftarget);
snd_emu10k1_ptr_write(hw, CVCF, ch, temp | 0xff00);
return 0;
}
/*
* Start envelope
*/
static void
trigger_voice(struct snd_emux_voice *vp)
{
unsigned int temp, ptarget;
struct snd_emu10k1 *hw;
struct snd_emu10k1_memblk *emem;
hw = vp->hw;
emem = (struct snd_emu10k1_memblk *)vp->block;
if (! emem || emem->mapped_page < 0)
return; /* not mapped */
#if 0
ptarget = (unsigned int)vp->ptarget << 16;
#else
ptarget = IP_TO_CP(vp->apitch);
#endif
/* set pitch target and pan (volume) */
temp = ptarget | (vp->apan << 8) | vp->aaux;
snd_emu10k1_ptr_write(hw, PTRX, vp->ch, temp);
/* pitch target */
snd_emu10k1_ptr_write(hw, CPF, vp->ch, ptarget);
/* trigger voice */
snd_emu10k1_ptr_write(hw, DCYSUSV, vp->ch, vp->reg.parm.voldcysus|DCYSUSV_CHANNELENABLE_MASK);
}
#define MOD_SENSE 18
/* set lfo1 modulation height and cutoff */
static void
set_fmmod(struct snd_emu10k1 *hw, struct snd_emux_voice *vp)
{
unsigned short fmmod;
short pitch;
unsigned char cutoff;
int modulation;
pitch = (char)(vp->reg.parm.fmmod>>8);
cutoff = (vp->reg.parm.fmmod & 0xff);
modulation = vp->chan->gm_modulation + vp->chan->midi_pressure;
pitch += (MOD_SENSE * modulation) / 1200;
LIMITVALUE(pitch, -128, 127);
fmmod = ((unsigned char)pitch<<8) | cutoff;
snd_emu10k1_ptr_write(hw, FMMOD, vp->ch, fmmod);
}
/* set lfo2 pitch & frequency */
static void
set_fm2frq2(struct snd_emu10k1 *hw, struct snd_emux_voice *vp)
{
unsigned short fm2frq2;
short pitch;
unsigned char freq;
int modulation;
pitch = (char)(vp->reg.parm.fm2frq2>>8);
freq = vp->reg.parm.fm2frq2 & 0xff;
modulation = vp->chan->gm_modulation + vp->chan->midi_pressure;
pitch += (MOD_SENSE * modulation) / 1200;
LIMITVALUE(pitch, -128, 127);
fm2frq2 = ((unsigned char)pitch<<8) | freq;
snd_emu10k1_ptr_write(hw, FM2FRQ2, vp->ch, fm2frq2);
}
/* set filterQ */
static void
set_filterQ(struct snd_emu10k1 *hw, struct snd_emux_voice *vp)
{
unsigned int val;
val = snd_emu10k1_ptr_read(hw, CCCA, vp->ch) & ~CCCA_RESONANCE;
val |= (vp->reg.parm.filterQ << 28);
snd_emu10k1_ptr_write(hw, CCCA, vp->ch, val);
}