June 2011 - Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PATCH] ASoC: Suppress restore of default register values for rbtree cache sync
by Mark Brown 06 Jun '11

06 Jun '11

Currently the rbtree code will write out the entire register map when doing a cache sync which is wasteful and will slow things down. Check to see if the value we're about to write is the default and don't bother restoring it if it is, either the value will have been retained or the device will have been reset and holds the value already. We should really store the defaults in the nodes but this resolves the immediate issue. Signed-off-by: Mark Brown <broonie(a)opensource.wolfsonmicro.com> --- sound/soc/soc-cache.c | 7 ++++++- 1 files changed, 6 insertions(+), 1 deletions(-) diff --git a/sound/soc/soc-cache.c b/sound/soc/soc-cache.c index d8ce34c..d82ab42 100644 --- a/sound/soc/soc-cache.c +++ b/sound/soc/soc-cache.c @@ -607,7 +607,7 @@ static int snd_soc_rbtree_cache_sync(struct snd_soc_codec *codec) struct rb_node *node; struct snd_soc_rbtree_node *rbnode; unsigned int regtmp; - unsigned int val; + unsigned int val, def; int ret; int i; @@ -619,6 +619,11 @@ static int snd_soc_rbtree_cache_sync(struct snd_soc_codec *codec) WARN_ON(codec->writable_register && codec->writable_register(codec, regtmp)); val = snd_soc_rbtree_get_register(rbnode, i); + def = snd_soc_get_cache_val(codec->reg_def_copy, i, + rbnode->word_size); + if (val == def) + continue; + codec->cache_bypass = 1; ret = snd_soc_write(codec, regtmp, val); codec->cache_bypass = 0; -- 1.7.5.3

3 5

[alsa-devel] [PATCH] ASoC: Update omap/ams-delta for TTY API change
by Janusz Krzysztofik 06 Jun '11

06 Jun '11

Following the receive_buf() change to return the number of bytes received, the cx20442 CODEC driver was updated to report that we accepted all the data that was offered to us. Do the same with the corresponding Amstrad Delta sound card level wrapper function, which is actually registered as the tty line discipline callback instead of the the CODEC function, which it then calls. Signed-off-by: Janusz Krzysztofik <jkrzyszt(a)tis.icnet.pl> --- sound/soc/omap/ams-delta.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) --- git/sound/soc/omap/ams-delta.c.orig 2011-04-09 00:39:54.000000000 +0200 +++ git/sound/soc/omap/ams-delta.c 2011-05-12 18:50:33.000000000 +0200 @@ -331,7 +331,7 @@ static int cx81801_hangup(struct tty_str } /* Line discipline .recieve_buf() */ -static void cx81801_receive(struct tty_struct *tty, +static unsigned int cx81801_receive(struct tty_struct *tty, const unsigned char *cp, char *fp, int count) { struct snd_soc_codec *codec = tty->disc_data; @@ -347,7 +347,7 @@ static void cx81801_receive(struct tty_s /* Initialize timer used for config pulse generation */ setup_timer(&cx81801_timer, cx81801_timeout, 0); - v253_ops.receive_buf(tty, cp, fp, count); + ret = v253_ops.receive_buf(tty, cp, fp, count); /* Link hook switch to DAPM pins */ ret = snd_soc_jack_add_pins(&ams_delta_hook_switch, @@ -358,10 +358,10 @@ static void cx81801_receive(struct tty_s "Failed to link hook switch to DAPM pins, " "will continue with hook switch unlinked.\n"); - return; + return count; } - v253_ops.receive_buf(tty, cp, fp, count); + ret = v253_ops.receive_buf(tty, cp, fp, count); for (c = &cp[count - 1]; c >= cp; c--) { if (*c != '\r') @@ -381,6 +381,7 @@ static void cx81801_receive(struct tty_s AMS_DELTA_LATCH2_MODEM_CODEC); break; } + return count; } /* Line discipline .write_wakeup() */

4 10

[alsa-devel] UCM: Possible snd_use_case_get enhancements
by Stephen Warren 06 Jun '11

06 Jun '11

Liam, The recent patches I posted allow querying a modifier/device's supported/ conflicting device list from a modifier/device either within the currently active verb (if any) or an explicitly defined verb. I figured a similar enhancement to snd_use_case_get might be useful; i.e. to allow querying within a specific verb, and also to indicate that one shouldn't fall back to a verb, or overall values. I'd propose something like this: NAME[[=]/<modifier>|<device>[/[=]<verb>]] The "=" would prevent fallback to a "higher" object if present. So for example: Current verb, not from device or modifier, with fallback to verb and ValueDefaults: PlaybackVolume Current verb, from device/modifier with fallback to verb and ValueDefaults: PlaybackVolume/Play Music Current verb, from dev/mod without fallback to verb or ValueDefaults: PlaybackVolume/=Play Music Specify a verb and dev/mod, but allow fallback to verb or ValueDefaults: PlaybackVolume/Play Music/HiFi Specify a verb and dev/mod, but no fallback to verb or ValueDefaults: PlaybackVolume/=Play Music/=HiFi One can also query from just a verb, with/without fallback: PlaybackVolume//HiFi PlaybackVolume//=HiFi Or current verb with no fallback: PlaybackVolume//= Does this seem useful, and a sane query string syntax? An alternative would be to drop the added "=" stuff, and never auto- fallback. However, that'd be a regression relative to current functionality, force clients to implement fallback themselves, and make querying ValueDefaults hard. -- nvpublic

2 1

[alsa-devel] [PATCH 0/8] MFD/ASoC: TWL4030/TWL6040 changes
by Peter Ujfalusi 06 Jun '11

06 Jun '11

Hello, The series will do three major things, and they are in one series, because they pretty much depending on each other, so it is easier to handle them together. 1. Rename the twl4030-codec MFD driver to twl4030-audio Since the ASoC multicomponent introduction (2.6.37) the twl4030-codec MFD dirver had been using twl4030-audio to register (twl4030-codec is the ASoC codec driver). The documentation refers to this part of twl as audio block, so it is better to rename the driver as well to avoid confusion. 2. Introduction of MFD driver for twl6040. The TWL6040 audio IC has codec and vibra functionality. Convert the existing ASoC codec driver to use the new MFD driver's interface. 3. Vibrator driver for TWL6040 (Input/ForceFeedback driver) The driver itself, and support for the vibrators in SDP4430 board. The series has been tested on top of merged omap-next, asoc-next, and mfd-next. It applies on top of linux-next as well. Regards, Peter --- Misael Lopez Cruz (3): mfd: twl6040: Add initial support ASoC: twl6040: Convert into TWL6040 MFD child input: Add initial support for TWL6040 vibrator Peter Ujfalusi (5): MFD: twl4030-codec: Rename internals from codec to audio MFD: twl4030-codec -> twl4030-audio: Rename the driver MFD: twl4030-audio: Rename platform data MFD: twl6040: Change platform data for soc codec driver OMAP4: SDP4430: Add twl6040 vibrator platform support arch/arm/mach-omap2/board-3430sdp.c | 8 +- arch/arm/mach-omap2/board-4430sdp.c | 50 +++- arch/arm/mach-omap2/board-devkit8000.c | 8 +- arch/arm/mach-omap2/board-igep0020.c | 8 +- arch/arm/mach-omap2/board-omap3beagle.c | 8 +- arch/arm/mach-omap2/board-omap3evm.c | 8 +- arch/arm/mach-omap2/board-omap3pandora.c | 8 +- arch/arm/mach-omap2/board-omap3stalker.c | 8 +- arch/arm/mach-omap2/board-omap3touchbook.c | 8 +- arch/arm/mach-omap2/board-overo.c | 8 +- arch/arm/mach-omap2/board-rx51-peripherals.c | 6 +- arch/arm/mach-omap2/board-zoom-peripherals.c | 14 +- arch/arm/plat-omap/include/plat/irqs.h | 12 +- drivers/input/misc/Kconfig | 13 +- drivers/input/misc/Makefile | 1 + drivers/input/misc/twl4030-vibra.c | 12 +- drivers/input/misc/twl6040-vibra.c | 428 +++++++++++++++++++ drivers/mfd/Kconfig | 8 +- drivers/mfd/Makefile | 3 +- drivers/mfd/twl-core.c | 13 +- drivers/mfd/twl4030-audio.c | 278 ++++++++++++ drivers/mfd/twl4030-codec.c | 277 ------------ drivers/mfd/twl6040-core.c | 588 ++++++++++++++++++++++++++ drivers/mfd/twl6040-irq.c | 205 +++++++++ include/linux/i2c/twl.h | 21 +- include/linux/mfd/twl4030-audio.h | 272 ++++++++++++ include/linux/mfd/twl4030-codec.h | 272 ------------ include/linux/mfd/twl6040.h | 260 ++++++++++++ sound/soc/codecs/Kconfig | 3 +- sound/soc/codecs/twl4030.c | 22 +- sound/soc/codecs/twl6040.c | 426 ++++--------------- sound/soc/codecs/twl6040.h | 118 ----- sound/soc/omap/sdp3430.c | 2 +- sound/soc/omap/sdp4430.c | 2 + sound/soc/omap/zoom2.c | 2 +- 35 files changed, 2290 insertions(+), 1090 deletions(-) create mode 100644 drivers/input/misc/twl6040-vibra.c create mode 100644 drivers/mfd/twl4030-audio.c delete mode 100644 drivers/mfd/twl4030-codec.c create mode 100644 drivers/mfd/twl6040-core.c create mode 100644 drivers/mfd/twl6040-irq.c create mode 100644 include/linux/mfd/twl4030-audio.h delete mode 100644 include/linux/mfd/twl4030-codec.h create mode 100644 include/linux/mfd/twl6040.h -- 1.7.5.3

4 20

[alsa-devel] [PATCH 1/4] ASoC: Suppress noop SYSCLK updates in WM8915
by Mark Brown 06 Jun '11

06 Jun '11

Signed-off-by: Mark Brown <broonie(a)opensource.wolfsonmicro.com> --- sound/soc/codecs/wm8915.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-) diff --git a/sound/soc/codecs/wm8915.c b/sound/soc/codecs/wm8915.c index 5a59ef7..c61bc32 100644 --- a/sound/soc/codecs/wm8915.c +++ b/sound/soc/codecs/wm8915.c @@ -55,6 +55,7 @@ struct wm8915_priv { int ldo1ena; int sysclk; + int sysclk_src; int fll_src; int fll_fref; @@ -1834,6 +1835,9 @@ static int wm8915_set_sysclk(struct snd_soc_dai *dai, int src; int old; + if (freq == wm8915->sysclk && clk_id == wm8915->sysclk_src) + return 0; + /* Disable SYSCLK while we reconfigure */ old = snd_soc_read(codec, WM8915_AIF_CLOCKING_1); snd_soc_update_bits(codec, WM8915_AIF_CLOCKING_1, @@ -1885,6 +1889,8 @@ static int wm8915_set_sysclk(struct snd_soc_dai *dai, snd_soc_update_bits(codec, WM8915_AIF_CLOCKING_1, WM8915_SYSCLK_ENA, old); + wm8915->sysclk_src = clk_id; + return 0; } -- 1.7.5.3

2 4

[alsa-devel] [PATCH 1/5] UCM: Deprecate index on Section{Device, Modifier}
by Stephen Warren 06 Jun '11

06 Jun '11

The previous supported "legacy" syntax was: SectionDevice."Speaker".0 { SectionModifier."Capture Voice".0 { This change supports new syntax: SectionDevice."Speaker" { SectionModifier."Capture Voice" { ... but also allows the old syntax, iff the index is exactly "0". If an index is present, but not exactly "0", parsing will appear to succeed, but produce an empty device or modifier. When naming devices and modifiers, even if the legacy format is used, any index is not included in the name; i.e. both sets of syntax above name the device just "Speaker". The SupportedDevice list syntax still also accepts either "x" or "x.0", but internally strips ".0" from the tail of any device name. Any other name including "." is disallowed. Finally, when comparing device or modifier names, a simple exact string compare is now used, since no index data is ever present in device or modifier names. The one functional change introduced here is that a SupportedDevice entry of just "x" will now only ever match a single device. It previously acted as a wildcard for any device named "x.foo". Signed-off-by: Stephen Warren <swarren(a)nvidia.com> --- src/ucm/main.c | 43 ++++----------- src/ucm/parser.c | 156 ++++++++++++++++++++++++++++++++++++------------------ 2 files changed, 114 insertions(+), 85 deletions(-) diff --git a/src/ucm/main.c b/src/ucm/main.c index fb4aa91..5acb5a8 100644 --- a/src/ucm/main.c +++ b/src/ucm/main.c @@ -484,33 +484,18 @@ static int is_modifier_supported(snd_use_case_mgr_t *uc_mgr, struct dev_list *device; struct use_case_device *adev; struct list_head *pos, *pos1; - char *cpos; - int dlen, len; list_for_each(pos, &modifier->dev_list) { device = list_entry(pos, struct dev_list, list); - cpos = strchr(device->name, '.'); - if (cpos) { - if (find(&uc_mgr->active_devices, - struct use_case_device, active_list, - name, device->name)) - return 1; - } else { - dlen = strlen(device->name); - list_for_each(pos1, &uc_mgr->active_devices) { - adev = list_entry(pos1, struct use_case_device, - active_list); - cpos = strchr(adev->name, '.'); - if (cpos) - len = cpos - adev->name; - else - len = strlen(adev->name); - if (len != dlen) - continue; - if (memcmp(adev->name, device->name, len)) - continue; - return 1; - } + + list_for_each(pos1, &uc_mgr->active_devices) { + adev = list_entry(pos1, struct use_case_device, + active_list); + + if (strcmp(adev->name, device->name)) + continue; + + return 1; } } return 0; @@ -528,19 +513,11 @@ static struct use_case_modifier * struct use_case_modifier *modifier; struct use_case_verb *verb = uc_mgr->active_verb; struct list_head *pos; - char name[64], *cpos; list_for_each(pos, &verb->modifier_list) { modifier = list_entry(pos, struct use_case_modifier, list); - strncpy(name, modifier->name, sizeof(name)); - name[sizeof(name)-1] = '\0'; - cpos = strchr(name, '.'); - if (!cpos) - continue; - *cpos= '\0'; - - if (strcmp(name, modifier_name)) + if (strcmp(modifier->name, modifier_name)) continue; if (is_modifier_supported(uc_mgr, modifier)) diff --git a/src/ucm/parser.c b/src/ucm/parser.c index f7de9bd..b0bebe0 100644 --- a/src/ucm/parser.c +++ b/src/ucm/parser.c @@ -59,7 +59,7 @@ int parse_string(snd_config_t *n, char **res) /* * Parse safe ID */ -int parse_is_name_safe(char *name) +int parse_is_name_safe(const char *name) { if (strchr(name, '.')) { uc_error("char '.' not allowed in '%s'", name); @@ -75,7 +75,9 @@ int parse_get_safe_id(snd_config_t *n, const char **id) err = snd_config_get_id(n, id); if (err < 0) return err; - return parse_is_name_safe((char *)(*id)); + if (!parse_is_name_safe((char *)(*id))) + return -EINVAL; + return 0; } /* @@ -151,7 +153,7 @@ static int parse_compound(snd_use_case_mgr_t *uc_mgr, snd_config_t *cfg, n = snd_config_iterator_entry(i); if (snd_config_get_type(cfg) != SND_CONFIG_TYPE_COMPOUND) { - uc_error("compound type expected for %s", id); + uc_error("compound type expected for %s, is %d", id, snd_config_get_type(cfg)); return -EINVAL; } @@ -163,6 +165,19 @@ static int parse_compound(snd_use_case_mgr_t *uc_mgr, snd_config_t *cfg, return 0; } +static int strip_legacy_dev_index(char *name) +{ + char *dot = strchr(name, '.'); + if (!dot) + return 0; + if (dot[1] != '0' || dot[2] != '\0') { + uc_error("device name %s contains a '.'," + " and is not legacy foo.0 format", name); + return -EINVAL; + } + *dot = '\0'; + return 0; +} /* * Parse transition @@ -199,6 +214,12 @@ static int parse_supported_device(snd_use_case_mgr_t *uc_mgr ATTRIBUTE_UNUSED, free(sdev); return err; } + err = strip_legacy_dev_index(sdev->name); + if (err < 0) { + free(sdev->name); + free(sdev); + return err; + } list_add(&sdev->list, dlist); } return 0; @@ -389,12 +410,12 @@ static int parse_value(snd_use_case_mgr_t *uc_mgr ATTRIBUTE_UNUSED, * Parse Modifier Use cases * * # Each modifier is described in new section. N modifiers are allowed - * SectionModifier."Capture Voice".0 { + * SectionModifier."Capture Voice" { * * Comment "Record voice call" * SupportedDevice [ - * "x" # all x device instances - * "y.0" # device y instance 0 only + * "x" + * "y" * ] * * EnableSequence [ @@ -426,14 +447,21 @@ static int parse_modifier(snd_use_case_mgr_t *uc_mgr, { struct use_case_verb *verb = data1; struct use_case_modifier *modifier; - char *name = data2; - const char *id; + const char *name; snd_config_iterator_t i, next; snd_config_t *n; int err; - if (!parse_is_name_safe(name)) - return -EINVAL; + if (data2) { + name = data2; + if (!parse_is_name_safe(name)) + return -EINVAL; + } + else { + if (parse_get_safe_id(cfg, &name) < 0) + return -EINVAL; + } + /* allocate modifier */ modifier = calloc(1, sizeof(*modifier)); if (modifier == NULL) @@ -444,15 +472,7 @@ static int parse_modifier(snd_use_case_mgr_t *uc_mgr, INIT_LIST_HEAD(&modifier->dev_list); INIT_LIST_HEAD(&modifier->value_list); list_add_tail(&modifier->list, &verb->modifier_list); - err = parse_get_safe_id(cfg, &id); - if (err < 0) - return err; - modifier->name = malloc(strlen(name) + strlen(id) + 2); - if (modifier->name == NULL) - return -ENOMEM; - strcpy(modifier->name, name); - strcat(modifier->name, "."); - strcat(modifier->name, id); + modifier->name = strdup(name); snd_config_for_each(i, next, cfg) { const char *id; @@ -530,7 +550,7 @@ static int parse_modifier(snd_use_case_mgr_t *uc_mgr, * Parse Device Use Cases * *# Each device is described in new section. N devices are allowed - *SectionDevice."Headphones".0 { + *SectionDevice."Headphones" { * Comment "Headphones connected to 3.5mm jack" * * EnableSequence [ @@ -551,21 +571,28 @@ static int parse_modifier(snd_use_case_mgr_t *uc_mgr, * } * } */ -static int parse_device_index(snd_use_case_mgr_t *uc_mgr, - snd_config_t *cfg, - void *data1, - void *data2) +static int parse_device(snd_use_case_mgr_t *uc_mgr, + snd_config_t *cfg, + void *data1, + void *data2) { struct use_case_verb *verb = data1; - char *name = data2; + const char *name; struct use_case_device *device; - const char *id; snd_config_iterator_t i, next; snd_config_t *n; int err; - - if (!parse_is_name_safe(name)) - return -EINVAL; + + if (data2) { + name = data2; + if (!parse_is_name_safe(name)) + return -EINVAL; + } + else { + if (parse_get_safe_id(cfg, &name) < 0) + return -EINVAL; + } + device = calloc(1, sizeof(*device)); if (device == NULL) return -ENOMEM; @@ -574,14 +601,7 @@ static int parse_device_index(snd_use_case_mgr_t *uc_mgr, INIT_LIST_HEAD(&device->transition_list); INIT_LIST_HEAD(&device->value_list); list_add_tail(&device->list, &verb->device_list); - if (parse_get_safe_id(cfg, &id) < 0) - return -EINVAL; - device->name = malloc(strlen(name) + strlen(id) + 2); - if (device->name == NULL) - return -ENOMEM; - strcpy(device->name, name); - strcat(device->name, "."); - strcat(device->name, id); + device->name = strdup(name); snd_config_for_each(i, next, cfg) { const char *id; @@ -643,34 +663,66 @@ static int parse_device_index(snd_use_case_mgr_t *uc_mgr, return 0; } -static int parse_device_name(snd_use_case_mgr_t *uc_mgr, - snd_config_t *cfg, - void *data1, - void *data2 ATTRIBUTE_UNUSED) +static int parse_compound_check_legacy(snd_use_case_mgr_t *uc_mgr, + snd_config_t *cfg, + int (*fcn)(snd_use_case_mgr_t *, snd_config_t *, void *, void *), + void *data1) { - const char *id; + const char *id, *idchild; + int child_ctr = 0, legacy_format = 1; + snd_config_iterator_t i, next; + snd_config_t *child; int err; err = snd_config_get_id(cfg, &id); if (err < 0) return err; - return parse_compound(uc_mgr, cfg, parse_device_index, - data1, (void *)id); + + snd_config_for_each(i, next, cfg) { + child_ctr++; + if (child_ctr > 1) { + break; + } + + child = snd_config_iterator_entry(i); + + if (snd_config_get_type(cfg) != SND_CONFIG_TYPE_COMPOUND) { + legacy_format = 0; + break; + } + + if (snd_config_get_id(child, &idchild) < 0) + return -EINVAL; + + if (strcmp(idchild, "0")) { + legacy_format = 0; + break; + } + } + if (child_ctr != 1) { + legacy_format = 0; + } + + if (legacy_format) + return parse_compound(uc_mgr, cfg, fcn, data1, (void *)id); + else + return fcn(uc_mgr, cfg, data1, NULL); } -static int parse_modifier_name(snd_use_case_mgr_t *uc_mgr, +static int parse_device_name(snd_use_case_mgr_t *uc_mgr, snd_config_t *cfg, void *data1, void *data2 ATTRIBUTE_UNUSED) { - const char *id; - int err; + return parse_compound_check_legacy(uc_mgr, cfg, parse_device, data1); +} - err = snd_config_get_id(cfg, &id); - if (err < 0) - return err; - return parse_compound(uc_mgr, cfg, parse_modifier, - data1, (void *)id); +static int parse_modifier_name(snd_use_case_mgr_t *uc_mgr, + snd_config_t *cfg, + void *data1, + void *data2 ATTRIBUTE_UNUSED) +{ + return parse_compound_check_legacy(uc_mgr, cfg, parse_modifier, data1); } /* -- 1.7.0.4

2 5

[alsa-devel] [alsa-lib] A suggestion on file cards.c
by alex dot baldacchino dot alsasub at gmail dot com 05 Jun '11

05 Jun '11

Hello, first of all, let me point out documentation for snd_card_get_index is incorrect: it is stated that zero is returned in case of success, instead, a value greater than or equal to zero is the real success result. In second place, let me suggest a different implementation. There's a number of functions (to get index, name, longname of a card; other ones could be added eventually, as the interface could evolve) resorting to a round of calls and pointer dereferencing to snd_ctl_hw_open() and snd_ctl_card_info()/snd_ctl_hw_card_info() to perform about the same operations as snd_card_load2() (eventually combined with snd_card_load1() ), potentially in a more dangerous fashion and possibly causing snd_card_load2() to be called anyway. Thus, I'm wondering if such calls could be avoided with just a few (not excessively invasive) changes to snd_card_load2() ( + snd_card_load1(), if the case, as I think), and I've made an attempt to implement those functions this way. Let's consider a few things. Function snd_ctl_hw_open() is called with the only purpose to allocate and populate an object of type snd_ctl_t, whose pointer is then passed to snd_ctl_card_info(); the latter one accesses that object fields to end up calling snd_ctl_hw_card_info(), whose purpose is to call the same ioctl as snd_card_load2(), in order to populate an object of type snd_ctl_card_info_t, therefore a call to a (not too much) modified version of snd_card_load2 could be enough for the same goals of successive calls to snd_ctl_hw_open() and snd_ctl_card_info(). This would rise following constraints: 1) Creating an object of type snd_ctl_card_info_t accessible by the caller after the function (snd_card_load2) returns. This is easy to accomplish: the caller uses a local variable of that type, thus let's just add an argument of type snd_ctl_card_info_t * and either use it in place of the internal (automatic variable) one used by snd_card_load2(), or choose whether to use the pointer argument (if not null) or an automatic variable (I've chosen this path). 2) Accessing a device special file corresponding to an integer card number. More reasoning needed here, since snd_ctl_hw_open() does it on its own, whereas snd_card_load2 requires a char* argument. Translating card number to device path can be done in three ways, essentially: - building a proper path name the same way as snd_ctl_hw_open() does it (that is, about the same way as snd_card_load1() does it - see below) in each function calling snd_card_load2(); feasible, yet not very good for maintenance; - doing so inside snd_card_load2() if passed char pointer is null; feasible, but it would still be duplicating (at least part of the) code from snd_card_load1(); - calling snd_card_load2() via snd_card_load1() (as it's done right now, but for one case - within snd_card_get_index - which would remain about the same), by just furnishing snd_card_load1() with a pointer-to-snd_ctl_card_info_t parameter to be passed straightforward to snd_card_load2() (that could be null when specific card infos other than card number aren't needed by the caller). After a few considerations, the latter one seems feasible, and have been my choice. Function snd_ctl_hw_open() builds the path name for a control* device the same way it's done by snd_card_load1() (apparently aload* devices aren't taken into account, but please, read further), then snd_open_device is called, as it's done by snd_card_load2() BUT with an important difference: the latter function opens devices read-only, whereas snd_ctl_hw_open() can (and, as actually called by snd_card_* stuff, it does) open devices read-write, which is both needless and potentially dangerous. In case of failure opening chosen device, snd_ctl_hw_open calls snd_card_load() and retries: this means that snd_card_load1() and snd_card_load2() (via snd_card_load()) can be called anyway, so that some operations are performed more than once needlessly (and there's no need to call snd_card_load() from within snd_card_get_index() anyway), but also that aload stuff, if compiled as supported, is somehow taken into account silently even when it would seem to be ignored (but perhaps shouldn't be?), as in snd_card_get_name/longname. This should make snd_card_load1() suitable to translate card number to device path and call snd_card_load2. 3) Printing out error messages. In some cases both snd_ctl_hw_open() and snd_ctl_hw_card_info() can produce error messages, whereas snd_card_load#() functions are mute in same scenarios. Whenever they're called directly (in current implementation), such a behaviour should remain as is, but when replacing snd_ctl_hw_* stuff same errors/messages should be printed out. This is possible by duplicating those (small) pieces of code from snd_ctl_hw_* functions and executing them conditionally, availing of a further integer argument (used as a boolean flag). To avoid complicating calls to snd_card_load1/2(), a few macros can be defined to both mask out the "printing flag" and give them an alias name appropriate for the context they're being called from (which could add expressiveness to the implementation). Anyway, take this as a simple suggestion, just give it a look. =================================================== src/control/cards.c - attached ===================================================

1 0

[alsa-devel] [PATCH] speaker-test.1: Clarify man page by giving examples
by Daniel Chen 05 Jun '11

05 Jun '11

Hi, Here's a forwarded patch from a Debian bug report for the man page of the speaker-test utility. It gives examples for using speaker-test alongside aplay. Please consider applying. Cheers, -Dan

2 1

[alsa-devel] AudioArduino - an ALSA soundcard driver for FTDI-based Arduinos
by sdaau 04 Jun '11

04 Jun '11

Hi all, (cross-posting to alsa-devel(a)alsa-project.org and gmane.linux.usb.general/linux-usb(a)vger.kernel.org; apologies if it's inappropriate) I'd just like to inform that the AudioArduino project has been posted online: http://imi.aau.dk/~sd/phd/index.php?title=AudioArduino It contains links to open-source Arduino microcontroller code, and corresponding ALSA driver source code - which can demonstrate mono, full-duplex, 8-bit / 44.1 kHz soundcard operation on an FTDI-based Arduino. There is also a related post on the (new) Arduino Forum: http://arduino.cc/forum/index.php/topic,62582.0.html The driver essentially represents an ALSA extension to ftdi-sio (which is why I thought it would be relevant to both alsa-devel, and linux-usb). Many thanks to the community for critical help, like in: ftdi_sio, and usb_serial->interface->dev->driver_data being null http://comments.gmane.org/gmane.linux.usb.general/30455 Help with dummy.c (where/how to write?) http://comments.gmane.org/gmane.linux.alsa.devel/74487 Looking forward to any feedback regarding this project, Cheers!

3 5

[alsa-devel] Need general idea how to architect driver for multi-codec board
by Caleb Crome 04 Jun '11

04 Jun '11

Hello All, I'm building a board with multiple TLV320AIC33 codecs on board, and I don't know how I should architect the interface. The board is set up to use TDM (16 channels on 1 TDM PORT), it has 16 input and 16 output channels, all bonded together -- the hardware does not allow different codecs to run at different speeds. The board also includes several other widgets, like an I2C mux (to address all 8 stereo codecs), a couple of power amplifiers, and a limited number of ways the audio can be routed in or out. My question is this: I'm not sure how to architect the driver: 1) Simply instantiate multiple TLV320IAC33 codec drivers, as the AIC33 driver is already written. This seems like it won't work because I'd end up with multiple user-space 'cards', yet I want the card to appear as a single 16-channel sound card. Also, when you start one PCM interface for one codec, they would all start automatically, which I imagine the linux system wouldn't get. Also, the PCM interface needs to be configured differently for each codec (the TDM offset must be 0, 16, 32, 48, etc. bits). 2) rewrite the AIC33 driver to include all 8 codecs, including all the mixer controls for each codec. This is going to get REALLY big and ugly in the user space mixer control -- there will be a zillion controls, most of which are not really useful for this board. 3) rewrite the AIC33 driver to include all 8 codecs, but pare down the mixer controls to just the ones that I care about. If the card changes later, add in the changes to the driver. I don't have any huge desire to get the driver into the mainstream kernel (though I would not be upset if it was popular enough to warrant that), it's just not a priority. Any thoughts would be most welcome. Thanks, -Caleb

2 10