Module Name: src
Committed By: isaki
Date: Fri May 3 03:00:33 UTC 2019
Modified Files:
src/sys/dev/isa [isaki-audio2]: pas.c sb.c sbdsp.c sbdspvar.h
Log Message:
Adapt sbdsp to audio2.
- Use new query_format/set_format interfaces.
The formats are created from sb[pr]modes tables.
- Drop INDEPENDENT property for models proir to SB_16.
To generate a diff of this commit:
cvs rdiff -u -r1.71 -r1.71.2.1 src/sys/dev/isa/pas.c
cvs rdiff -u -r1.90 -r1.90.2.1 src/sys/dev/isa/sb.c
cvs rdiff -u -r1.139.2.1 -r1.139.2.2 src/sys/dev/isa/sbdsp.c
cvs rdiff -u -r1.61 -r1.61.54.1 src/sys/dev/isa/sbdspvar.h
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
Modified files:
Index: src/sys/dev/isa/pas.c
diff -u src/sys/dev/isa/pas.c:1.71 src/sys/dev/isa/pas.c:1.71.2.1
--- src/sys/dev/isa/pas.c:1.71 Sat Mar 16 12:09:58 2019
+++ src/sys/dev/isa/pas.c Fri May 3 03:00:33 2019
@@ -1,4 +1,4 @@
-/* $NetBSD: pas.c,v 1.71 2019/03/16 12:09:58 isaki Exp $ */
+/* $NetBSD: pas.c,v 1.71.2.1 2019/05/03 03:00:33 isaki Exp $ */
/*
* Copyright (c) 1991-1993 Regents of the University of California.
@@ -57,7 +57,7 @@
#include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: pas.c,v 1.71 2019/03/16 12:09:58 isaki Exp $");
+__KERNEL_RCSID(0, "$NetBSD: pas.c,v 1.71.2.1 2019/05/03 03:00:33 isaki Exp $");
#include <sys/param.h>
#include <sys/systm.h>
@@ -125,8 +125,8 @@ void pasconf(int, int, int, int);
const struct audio_hw_if pas_hw_if = {
.open = sbdsp_open,
.close = sbdsp_close,
- .query_encoding = sbdsp_query_encoding,
- .set_params = sbdsp_set_params,
+ .query_format = sbdsp_query_format,
+ .set_format = sbdsp_set_format,
.round_blocksize = sbdsp_round_blocksize,
.halt_output = sbdsp_halt_output,
.halt_input = sbdsp_halt_input,
@@ -138,7 +138,6 @@ const struct audio_hw_if pas_hw_if = {
.allocm = sb_malloc,
.freem = sb_free,
.round_buffersize = sb_round_buffersize,
- .mappage = sb_mappage,
.get_props = sbdsp_get_props,
.trigger_output = sbdsp_trigger_output,
.trigger_input = sbdsp_trigger_input,
Index: src/sys/dev/isa/sb.c
diff -u src/sys/dev/isa/sb.c:1.90 src/sys/dev/isa/sb.c:1.90.2.1
--- src/sys/dev/isa/sb.c:1.90 Sat Mar 16 12:09:58 2019
+++ src/sys/dev/isa/sb.c Fri May 3 03:00:33 2019
@@ -1,4 +1,4 @@
-/* $NetBSD: sb.c,v 1.90 2019/03/16 12:09:58 isaki Exp $ */
+/* $NetBSD: sb.c,v 1.90.2.1 2019/05/03 03:00:33 isaki Exp $ */
/*
* Copyright (c) 1991-1993 Regents of the University of California.
@@ -35,7 +35,7 @@
*/
#include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: sb.c,v 1.90 2019/03/16 12:09:58 isaki Exp $");
+__KERNEL_RCSID(0, "$NetBSD: sb.c,v 1.90.2.1 2019/05/03 03:00:33 isaki Exp $");
#include "midi.h"
@@ -82,8 +82,8 @@ int sb_getdev(void *, struct audio_devic
const struct audio_hw_if sb_hw_if = {
.open = sbdsp_open,
.close = sbdsp_close,
- .query_encoding = sbdsp_query_encoding,
- .set_params = sbdsp_set_params,
+ .query_format = sbdsp_query_format,
+ .set_format = sbdsp_set_format,
.round_blocksize = sbdsp_round_blocksize,
.halt_output = sbdsp_halt_output,
.halt_input = sbdsp_halt_input,
@@ -95,7 +95,6 @@ const struct audio_hw_if sb_hw_if = {
.allocm = sb_malloc,
.freem = sb_free,
.round_buffersize = sb_round_buffersize,
- .mappage = sb_mappage,
.get_props = sbdsp_get_props,
.trigger_output = sbdsp_trigger_output,
.trigger_input = sbdsp_trigger_input,
Index: src/sys/dev/isa/sbdsp.c
diff -u src/sys/dev/isa/sbdsp.c:1.139.2.1 src/sys/dev/isa/sbdsp.c:1.139.2.2
--- src/sys/dev/isa/sbdsp.c:1.139.2.1 Sun Apr 21 06:17:02 2019
+++ src/sys/dev/isa/sbdsp.c Fri May 3 03:00:33 2019
@@ -1,4 +1,4 @@
-/* $NetBSD: sbdsp.c,v 1.139.2.1 2019/04/21 06:17:02 isaki Exp $ */
+/* $NetBSD: sbdsp.c,v 1.139.2.2 2019/05/03 03:00:33 isaki Exp $ */
/*-
* Copyright (c) 1999, 2008 The NetBSD Foundation, Inc.
@@ -74,7 +74,7 @@
*/
#include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: sbdsp.c,v 1.139.2.1 2019/04/21 06:17:02 isaki Exp $");
+__KERNEL_RCSID(0, "$NetBSD: sbdsp.c,v 1.139.2.2 2019/05/03 03:00:33 isaki Exp $");
#include "midi.h"
#include "mpu.h"
@@ -95,6 +95,7 @@ __KERNEL_RCSID(0, "$NetBSD: sbdsp.c,v 1.
#include <sys/audioio.h>
#include <dev/audio_if.h>
+#include <dev/audio/linear.h>
#include <dev/midi_if.h>
#include <dev/isa/isavar.h>
@@ -206,8 +207,34 @@ static struct sbmode sbrmodes[] = {
{ .model = -1 }
};
+/*
+ * We actually can specify any value within the frequency range defined
+ * above. But according to definition of SB_RATE_TO_TC macro, only some
+ * of them are dividable (it's preferable, not mandatory). There are 9
+ * values in the range that satisfy this condition but it's too much.
+ */
+static const int sbdsp_rates[] = {
+ 4000,
+ /* 5000, */
+ /* 6250, */
+ /* 10000, */
+ 12500,
+ /* 15625, */
+ 20000,
+ /* 25000, */
+ 31250,
+};
+
void sbversion(struct sbdsp_softc *);
void sbdsp_jazz16_probe(struct sbdsp_softc *);
+void sbdsp_sbmode2format(struct audio_format *, const struct sbmode *, int);
+int sbdsp_set_format16(struct sbdsp_softc *, int,
+ const audio_params_t *, const audio_params_t *,
+ audio_filter_reg_t *, audio_filter_reg_t *);
+int sbdsp_set_format8(struct sbdsp_softc *, int,
+ const audio_params_t *, const audio_params_t *,
+ audio_filter_reg_t *, audio_filter_reg_t *);
+void sbdsp_init_format(struct sbdsp_softc *);
void sbdsp_set_mixer_gain(struct sbdsp_softc *, int);
void sbdsp_pause(struct sbdsp_softc *);
int sbdsp_set_timeconst(struct sbdsp_softc *, int);
@@ -410,9 +437,6 @@ sbdsp_attach(struct sbdsp_softc *sc)
SBVER_MAJOR(sc->sc_version), SBVER_MINOR(sc->sc_version),
sc->sc_model == SB_JAZZ ? ": <Jazz16>" : "");
- /* XXX It's not true full duplex. */
- sc->sc_fullduplex = 0;
-
if (sc->sc_drq8 != -1) {
sc->sc_drq8_maxsize = isa_dmamaxsize(sc->sc_ic,
sc->sc_drq8);
@@ -438,6 +462,14 @@ sbdsp_attach(struct sbdsp_softc *sc)
}
}
+ /* Construct sc_format from model */
+ sbdsp_init_format(sc);
+ if (sc->sc_nformats == 0) {
+ aprint_error_dev(sc->sc_dev,
+ "No available formats; model mismatch?\n");
+ return;
+ }
+
if (!pmf_device_register(sc->sc_dev, NULL, sbdsp_resume))
aprint_error_dev(sc->sc_dev,
"couldn't establish power handler\n");
@@ -488,119 +520,245 @@ sbdsp_mix_read(struct sbdsp_softc *sc, i
return val;
}
+void
+sbdsp_sbmode2format(struct audio_format *f, const struct sbmode *m, int mode)
+{
+ memset(f, 0, sizeof(*f));
+ f->mode = mode;
+ if (m->precision == 8) {
+ /* ulinear8 is always native endian */
+ f->encoding = AUDIO_ENCODING_ULINEAR_NE;
+ f->validbits = 8;
+ f->precision = 8;
+ } else {
+ f->encoding = AUDIO_ENCODING_SLINEAR_LE;
+ f->validbits = 16;
+ f->precision = 16;
+ }
+ f->channels = m->channels;
+ f->channel_mask = (m->channels == 1) ? AUFMT_MONAURAL : AUFMT_STEREO;
+ f->frequency_type = 0;
+ f->frequency[0] = m->lowrate;
+ f->frequency[1] = m->highrate;
+}
+
+/*
+ * Create sc_formats[] array from sbpmodes[], sbrmodes[].
+ */
+void
+sbdsp_init_format(struct sbdsp_softc *sc)
+{
+ struct audio_format dp[4];
+ struct audio_format dr[4];
+ struct audio_format *dbase;
+ struct audio_format *d;
+ struct audio_format tmp;
+ struct sbmode *sbmodes;
+ struct sbmode *m;
+ int mode;
+ int minrate;
+ int maxrate;
+ int idx;
+ int model;
+ int i;
+ int j;
+ int n;
+
+ /* Later models work like SB16. */
+ model = uimin(sc->sc_model, SB_16);
+
+ memset(&dp, 0, sizeof(dp));
+ memset(&dr, 0, sizeof(dr));
+
+ /*
+ * Step1. Extract elements corresponding to this model.
+ */
+ for (i = 0; i < 2; i++) {
+ if (i == 0) {
+ mode = AUMODE_PLAY;
+ sbmodes = sbpmodes;
+ dbase = dp;
+ } else {
+ mode = AUMODE_RECORD;
+ sbmodes = sbrmodes;
+ dbase = dr;
+ }
+ for (m = sbmodes; m->model != -1; m++) {
+ if (m->model != model)
+ continue;
+
+ sbdsp_sbmode2format(&tmp, m, mode);
+ /*
+ * [0] 8bit mono
+ * [1] 8bit st
+ * [2] 16bit mono
+ * [3] 16bit st
+ */
+ idx = (m->precision / 16) * 2 + (m->channels - 1);
+ d = &dbase[idx];
+ if (d->mode == 0) {
+ /* The first elements of this room */
+ *d = tmp;
+ continue;
+ }
+
+ /* Otherwise merge frequency */
+ /*
+ * Currently the frequency of multiple elements in
+ * the same model are all contiguous.
+ */
+ if (tmp.frequency[0] == d->frequency[1]) {
+ d->frequency[1] = tmp.frequency[1];
+ } else if (tmp.frequency[1] == d->frequency[0]) {
+ d->frequency[0] = tmp.frequency[0];
+ } else {
+ panic("frequency range must be contiguous. "
+ "model=%d\n", model);
+ }
+ DPRINTF(("%s: 1 [%d] mode=%d freq={ %d, %d }\n",
+ __func__, idx, d->mode,
+ d->frequency[0], d->frequency[1]));
+ }
+ }
+
+ /*
+ * Step2. Merge dr into dp.
+ */
+ for (i = 0; i < __arraycount(dp); i++) {
+ if (dp[i].mode == 0 && dr[i].mode == 0)
+ continue;
+ /* Currently all entries in sb[pr]modes are PLAY|REC */
+ if (dp[i].mode == 0 || dr[i].mode == 0)
+ panic("invalid sb[pr]mode table?. model=%d\n", model);
+ dp[i].mode |= dr[i].mode;
+
+ /*
+ * Usually, the recording range is the same or smaller than
+ * the playback range. So extract the common range.
+ */
+ if (dp[i].frequency[0] < dr[i].frequency[0])
+ dp[i].frequency[0] = dr[i].frequency[0];
+ if (dp[i].frequency[1] > dr[i].frequency[1])
+ dp[i].frequency[1] = dr[i].frequency[1];
+
+ DPRINTF(("%s: 2 [%d] mode=%d freq={ %d, %d }\n",
+ __func__, i, dp[i].mode,
+ dp[i].frequency[0], dp[i].frequency[1]));
+ }
+
+ /*
+ * Step3. Prior to SB16, use fixed frequencies rather than raw
+ * frequency range.
+ */
+ if (!ISSB16CLASS(sc)) {
+ for (i = 0; i < __arraycount(dp); i++) {
+ if (dp[i].mode == 0)
+ continue;
+ minrate = dp[i].frequency[0];
+ maxrate = dp[i].frequency[1];
+ n = 0;
+ for (j = 0; j < __arraycount(sbdsp_rates); j++) {
+ if (minrate <= sbdsp_rates[j] &&
+ sbdsp_rates[j] <= maxrate) {
+ dp[i].frequency[n++] = sbdsp_rates[j];
+ }
+ }
+ dp[i].frequency_type = n;
+ if (n == 0) {
+ /* this should not happened */
+ dp[i].frequency[0] = minrate;
+ dp[i].frequency[1] = maxrate;
+ }
+
+ DPRINTF(("%s: 3 [%d] mode=%d freq={ ",
+ __func__, i, dp[i].mode));
+ for (j = 0; j < dp[i].frequency_type; j++) {
+ DPRINTF(("%s%d", (j == 0) ? "" : ", ",
+ dp[i].frequency[j]));
+ }
+ DPRINTF((" }\n"));
+ }
+ }
+
+ /*
+ * Step4. Copy merged dp to sc_formats.
+ */
+ n = 0;
+ for (i = 0; i < __arraycount(dp); i++) {
+ if (dp[i].mode)
+ sc->sc_formats[n++] = dp[i];
+ }
+ sc->sc_nformats = n;
+}
+
/*
* Various routines to interface to higher level audio driver
*/
int
-sbdsp_query_encoding(void *addr, struct audio_encoding *fp)
+sbdsp_query_format(void *addr, audio_format_query_t *afp)
{
struct sbdsp_softc *sc;
- int emul;
sc = addr;
- emul = ISSB16CLASS(sc) ? 0 : AUDIO_ENCODINGFLAG_EMULATED;
+ return audio_query_format(sc->sc_formats, sc->sc_nformats, afp);
+}
- switch (fp->index) {
- case 0:
- strcpy(fp->name, AudioEulinear);
- fp->encoding = AUDIO_ENCODING_ULINEAR;
- fp->precision = 8;
- fp->flags = 0;
- return 0;
- case 1:
- strcpy(fp->name, AudioEmulaw);
- fp->encoding = AUDIO_ENCODING_ULAW;
- fp->precision = 8;
- fp->flags = AUDIO_ENCODINGFLAG_EMULATED;
- return 0;
- case 2:
- strcpy(fp->name, AudioEalaw);
- fp->encoding = AUDIO_ENCODING_ALAW;
- fp->precision = 8;
- fp->flags = AUDIO_ENCODINGFLAG_EMULATED;
- return 0;
- case 3:
- strcpy(fp->name, AudioEslinear);
- fp->encoding = AUDIO_ENCODING_SLINEAR;
- fp->precision = 8;
- fp->flags = emul;
- return 0;
- }
- if (!ISSB16CLASS(sc) && sc->sc_model != SB_JAZZ)
- return EINVAL;
+static struct sbmode *
+sbdsp_find_mode(struct sbmode *sbmodes, int model, const audio_params_t *p)
+{
+ struct sbmode *m;
- switch(fp->index) {
- case 4:
- strcpy(fp->name, AudioEslinear_le);
- fp->encoding = AUDIO_ENCODING_SLINEAR_LE;
- fp->precision = 16;
- fp->flags = 0;
- return 0;
- case 5:
- strcpy(fp->name, AudioEulinear_le);
- fp->encoding = AUDIO_ENCODING_ULINEAR_LE;
- fp->precision = 16;
- fp->flags = emul;
- return 0;
- case 6:
- strcpy(fp->name, AudioEslinear_be);
- fp->encoding = AUDIO_ENCODING_SLINEAR_BE;
- fp->precision = 16;
- fp->flags = AUDIO_ENCODINGFLAG_EMULATED;
- return 0;
- case 7:
- strcpy(fp->name, AudioEulinear_be);
- fp->encoding = AUDIO_ENCODING_ULINEAR_BE;
- fp->precision = 16;
- fp->flags = AUDIO_ENCODINGFLAG_EMULATED;
- return 0;
- default:
- return EINVAL;
+ for (m = sbmodes; m->model != -1; m++) {
+ if (model == m->model &&
+ p->channels == m->channels &&
+ p->precision == m->precision &&
+ p->sample_rate >= m->lowrate &&
+ p->sample_rate <= m->highrate)
+ return m;
}
- return 0;
+ return NULL;
}
int
-sbdsp_set_params(
- void *addr,
- int setmode, int usemode,
- audio_params_t *play, audio_params_t *rec,
- stream_filter_list_t *pfil, stream_filter_list_t *rfil)
+sbdsp_set_format(void *addr, int setmode,
+ const audio_params_t *play, const audio_params_t *rec,
+ audio_filter_reg_t *pfil, audio_filter_reg_t *rfil)
{
struct sbdsp_softc *sc;
- struct sbmode *m;
- u_int rate, tc, bmode;
- int model;
- int chan;
- struct audio_params *p;
- audio_params_t hw __unused;
- int mode;
+ int error;
sc = addr;
if (sc->sc_open == SB_OPEN_MIDI)
return EBUSY;
- /* Later models work like SB16. */
- model = uimin(sc->sc_model, SB_16);
-
- /*
- * Prior to the SB16, we have only one clock, so make the sample
- * rates match.
- */
- if (!ISSB16CLASS(sc) &&
- play->sample_rate != rec->sample_rate &&
- usemode == (AUMODE_PLAY | AUMODE_RECORD)) {
- if (setmode == AUMODE_PLAY) {
- rec->sample_rate = play->sample_rate;
- setmode |= AUMODE_RECORD;
- } else if (setmode == AUMODE_RECORD) {
- play->sample_rate = rec->sample_rate;
- setmode |= AUMODE_PLAY;
- } else
- return EINVAL;
+ if (ISSB16CLASS(sc)) {
+ /* Later models work like SB16. */
+ error = sbdsp_set_format16(sc, setmode, play, rec, pfil, rfil);
+ } else {
+ error = sbdsp_set_format8(sc, setmode, play, rec, pfil, rfil);
}
+ if (error)
+ return error;
+
+ DPRINTF(("%s ichan=%d, ochan=%d\n", __func__,
+ sc->sc_i.dmachan, sc->sc_o.dmachan));
+ return 0;
+}
+
+/* set_format for SB_16 or later */
+int
+sbdsp_set_format16(struct sbdsp_softc *sc, int setmode,
+ const audio_params_t *play, const audio_params_t *rec,
+ audio_filter_reg_t *pfil, audio_filter_reg_t *rfil)
+{
+ struct sbmode *sbmodes;
+ struct sbmode *m;
+ struct sbdsp_state *io;
+ const audio_params_t *p;
+ u_int bmode;
+ int mode;
/* Set first record info, then play info */
for (mode = AUMODE_RECORD; mode != -1;
@@ -608,76 +766,98 @@ sbdsp_set_params(
if ((setmode & mode) == 0)
continue;
- p = mode == AUMODE_PLAY ? play : rec;
- /* Locate proper commands */
- for (m = mode == AUMODE_PLAY ? sbpmodes : sbrmodes;
- m->model != -1; m++) {
- if (model == m->model &&
- p->channels == m->channels &&
- p->precision == m->precision &&
- p->sample_rate >= m->lowrate &&
- p->sample_rate <= m->highrate)
- break;
+ p = NULL; /* XXX shut up gcc */
+ if (mode == AUMODE_PLAY) {
+ p = play;
+ sbmodes = sbpmodes;
+ io = &sc->sc_o;
+ } else {
+ p = rec;
+ sbmodes = sbrmodes;
+ io = &sc->sc_i;
}
- if (m->model == -1)
+ /* Locate proper commands */
+ m = sbdsp_find_mode(sbmodes, SB_16, p);
+ if (m == NULL)
return EINVAL;
- rate = p->sample_rate;
- hw = *p;
- tc = 1;
- bmode = -1;
- if (model == SB_16) {
- switch (p->encoding) {
- /* FALLTHROUGH */
- case AUDIO_ENCODING_SLINEAR:
- case AUDIO_ENCODING_SLINEAR_LE:
- bmode = SB_BMODE_SIGNED;
- break;
- default:
- return EINVAL;
- }
- if (p->channels == 2)
- bmode |= SB_BMODE_STEREO;
- } else if (m->model == SB_JAZZ && m->precision == 16) {
- switch (p->encoding) {
- case AUDIO_ENCODING_SLINEAR:
- case AUDIO_ENCODING_SLINEAR_LE:
- break;
- default:
- return EINVAL;
- }
- tc = SB_RATE_TO_TC(p->sample_rate * p->channels);
- p->sample_rate = SB_TC_TO_RATE(tc) / p->channels;
- hw.sample_rate = p->sample_rate;
+ bmode = SB_BMODE_UNSIGNED;
+ if (p->precision == 16) {
+ /* 16bit is slinear16_le */
+ bmode = SB_BMODE_SIGNED;
} else {
- panic("ulinear8 not supported; use set_format.");
+ /* 8bit is ulinear8_ne */
+ if (mode == AUMODE_PLAY)
+ pfil->codec = audio_internal_to_linear8;
+ else
+ rfil->codec = audio_linear8_to_internal;
}
+ if (p->channels == 2)
+ bmode |= SB_BMODE_STEREO;
- chan = m->precision == 16 ? sc->sc_drq16 : sc->sc_drq8;
- if (mode == AUMODE_PLAY) {
- sc->sc_o.rate = rate;
- sc->sc_o.tc = tc;
- sc->sc_o.modep = m;
- sc->sc_o.bmode = bmode;
- sc->sc_o.dmachan = chan;
- } else {
- sc->sc_i.rate = rate;
- sc->sc_i.tc = tc;
- sc->sc_i.modep = m;
- sc->sc_i.bmode = bmode;
- sc->sc_i.dmachan = chan;
- }
+ io->rate = p->sample_rate;
+ io->tc = 1;
+ io->modep = m;
+ io->bmode = bmode;
+ io->dmachan = m->precision == 16 ? sc->sc_drq16 : sc->sc_drq8;
+
+ DPRINTF(("%s: model=%d, mode=%d, "
+ "rate=%u, prec=%d, chan=%d, enc=%d -> "
+ "cmd=%02x, bmode=%02x, cmdchan=%02x\n",
+ __func__, sc->sc_model, mode,
+ p->sample_rate, p->precision, p->channels, p->encoding,
+ m->cmd, bmode, m->cmdchan));
+ }
+ return 0;
+}
+
+/* set_format for prior to SB_16 */
+int
+sbdsp_set_format8(struct sbdsp_softc *sc, int setmode,
+ const audio_params_t *play, const audio_params_t *rec,
+ audio_filter_reg_t *pfil, audio_filter_reg_t *rfil)
+{
+ struct sbmode *mp;
+ struct sbmode *mr;
+ u_int tc;
+ int chan;
- DPRINTF(("sbdsp_set_params: model=%d, mode=%d, rate=%u, "
- "prec=%d, chan=%d, enc=%d -> tc=%02x, cmd=%02x, "
- "bmode=%02x, cmdchan=%02x\n", sc->sc_model, mode,
- p->sample_rate, p->precision, p->channels,
- p->encoding, tc, m->cmd, bmode, m->cmdchan));
+ /* *play and *rec are the identical because !AUDIO_PROP_INDEPENDENT. */
- }
+ /* Locate proper commands */
+ mp = sbdsp_find_mode(sbpmodes, sc->sc_model, play);
+ if (mp == NULL)
+ return EINVAL;
+ mr = sbdsp_find_mode(sbrmodes, sc->sc_model, rec);
+ if (mr == NULL)
+ return EINVAL;
+
+ tc = SB_RATE_TO_TC(play->sample_rate * play->channels);
+ chan = mp->precision == 16 ? sc->sc_drq16 : sc->sc_drq8;
- DPRINTF(("sbdsp_set_params ichan=%d, ochan=%d\n",
- sc->sc_i.dmachan, sc->sc_o.dmachan));
+ sc->sc_o.rate = play->sample_rate;
+ sc->sc_o.tc = tc;
+ sc->sc_o.modep = mp;
+ sc->sc_o.bmode = -1;
+ sc->sc_o.dmachan = chan;
+
+ sc->sc_i.rate = rec->sample_rate;
+ sc->sc_i.tc = tc;
+ sc->sc_i.modep = mr;
+ sc->sc_i.bmode = -1;
+ sc->sc_i.dmachan = chan;
+
+ if (mp->precision == 8) {
+ pfil->codec = audio_internal_to_linear8;
+ rfil->codec = audio_linear8_to_internal;
+ }
+
+ DPRINTF(("%s: model=%d, "
+ "rate=%u, prec=%d, chan=%d, enc=%d -> "
+ "tc=%02x, cmd=%02x, cmdchan=%02x\n",
+ __func__, sc->sc_model,
+ play->sample_rate, play->precision, play->channels, play->encoding,
+ tc, mp->cmd, mp->cmdchan));
return 0;
}
@@ -2293,21 +2473,20 @@ sb_round_buffersize(void *addr, int dire
return size;
}
-paddr_t
-sb_mappage(void *addr, void *mem, off_t off, int prot)
-{
-
- return isa_mappage(mem, off, prot);
-}
-
int
sbdsp_get_props(void *addr)
{
struct sbdsp_softc *sc;
+ int prop;
sc = addr;
- return AUDIO_PROP_MMAP | AUDIO_PROP_INDEPENDENT |
- (sc->sc_fullduplex ? AUDIO_PROP_FULLDUPLEX : 0);
+ prop = AUDIO_PROP_MMAP;
+
+ /* Prior to the SB16, it has only one clock */
+ if (ISSB16CLASS(sc))
+ prop |= AUDIO_PROP_INDEPENDENT;
+
+ return prop;
}
void
Index: src/sys/dev/isa/sbdspvar.h
diff -u src/sys/dev/isa/sbdspvar.h:1.61 src/sys/dev/isa/sbdspvar.h:1.61.54.1
--- src/sys/dev/isa/sbdspvar.h:1.61 Wed Nov 23 23:07:33 2011
+++ src/sys/dev/isa/sbdspvar.h Fri May 3 03:00:33 2019
@@ -1,4 +1,4 @@
-/* $NetBSD: sbdspvar.h,v 1.61 2011/11/23 23:07:33 jmcneill Exp $ */
+/* $NetBSD: sbdspvar.h,v 1.61.54.1 2019/05/03 03:00:33 isaki Exp $ */
/*
* Copyright (c) 1991-1993 Regents of the University of California.
@@ -116,7 +116,6 @@ struct sbdsp_softc {
#define SB_CLOSED 0
#define SB_OPEN_AUDIO 1
#define SB_OPEN_MIDI 2
- u_char sc_fullduplex; /* can do full duplex */
u_char gain[SB_NDEVS][2]; /* kept in input levels */
#define SB_LEFT 0
@@ -179,6 +178,9 @@ struct sbdsp_softc {
#define SBVER_MAJOR(v) (((v)>>8) & 0xff)
#define SBVER_MINOR(v) ((v)&0xff)
+ struct audio_format sc_formats[4];
+ int sc_nformats;
+
#if NMPU > 0
int sc_hasmpu;
#define SBMPU_EXTERNAL 1
@@ -211,9 +213,10 @@ int sbdsp_set_out_gain_real(void *, u_in
int sbdsp_get_out_gain(void *);
int sbdsp_set_monitor_gain(void *, u_int);
int sbdsp_get_monitor_gain(void *);
-int sbdsp_query_encoding(void *, struct audio_encoding *);
-int sbdsp_set_params(void *, int, int, audio_params_t *, audio_params_t *,
- stream_filter_list_t *, stream_filter_list_t *);
+int sbdsp_query_format(void *, audio_format_query_t *);
+int sbdsp_set_format(void *, int,
+ const audio_params_t *, const audio_params_t *,
+ audio_filter_reg_t *, audio_filter_reg_t *);
int sbdsp_round_blocksize(void *, int, int, const audio_params_t *);
int sbdsp_get_avail_in_ports(void *);
int sbdsp_get_avail_out_ports(void *);
@@ -251,7 +254,6 @@ int sbdsp_mixer_query_devinfo(void *, mi
void *sb_malloc(void *, int, size_t);
void sb_free(void *, void *, size_t);
size_t sb_round_buffersize(void *, int, size_t);
-paddr_t sb_mappage(void *, void *, off_t, int);
int sbdsp_get_props(void *);
void sbdsp_get_locks(void *, kmutex_t **, kmutex_t **);
