Skip to contentSkip to conversation

Cleanup/refactor the usbaudio driver

RFC In review Daniel Mack 3 participants 17 messages Feb 26, 2010 – Mar 4, 2010 1 message not in the archive Sound

lore
Share & tools
lore
#0 lore raw

I've spent some time cleaning up the usbaudio driver. I felt that the existing implementation was quite difficult to follow because it contained code for all sorts of things mixed up in one big file.

So I split the driver into pieces and tried to be as less intrusive as I can, meaning that I didn't touch much real code but rather copied the functions as they were and factored them out. Of course, non-static functions were properly prefixed to avoid name space pollution.

I also moved almost all quirks to a seperate file and introduced two more generic hooks to handle them, so the excecption add less to the actual code of the driver.

Unfortunately it wasn't easily possible to put that big change into smaller pieces, and so 3/3 became a huge chunk.

After all, I feel that it is easier to read now, but I don't know whether you agree. So take this as RFC for now :)

Daniel

#1 lore raw

PATCH 1/3 ALSA: usb: move ua101 driver

As part of the USB audio code cleanup, move the non-standard ua101 driver out of the way.

Cc Clemens Ladisch <[email protected]>
Cc Takashi Iwai <[email protected]>
---
 sound/usb/Makefile            |    5 ++---
 sound/usb/ua101/Makefile      |    2 ++
 sound/usb/{ => ua101}/ua101.c |    2 +-
 3 files changed, 5 insertions(+), 4 deletions(-)
 create mode 100644 sound/usb/ua101/Makefile
 rename sound/usb/{ => ua101}/ua101.c (99%)
sound/usb/Makefile +2 -3
index 5bf64ae..8bc005b 100644
@@ -4,12 +4,11 @@
 
 snd-usb-audio-objs := usbaudio.o usbmixer.o
 snd-usb-lib-objs := usbmidi.o
-snd-ua101-objs := ua101.o
 
 # Toplevel Module Dependency
 obj-$(CONFIG_SND_USB_AUDIO) += snd-usb-audio.o snd-usb-lib.o
-obj-$(CONFIG_SND_USB_UA101) += snd-ua101.o snd-usb-lib.o
+obj-$(CONFIG_SND_USB_UA101) += snd-usb-lib.o
 obj-$(CONFIG_SND_USB_USX2Y) += snd-usb-lib.o
 obj-$(CONFIG_SND_USB_US122L) += snd-usb-lib.o
 
-obj-$(CONFIG_SND) += usx2y/ caiaq/
+obj-$(CONFIG_SND) += ua101/ usx2y/ caiaq/
sound/usb/ua101/Makefile new file +2 -0
new file mode 100644
index 0000000..ccefd81
@@ -0,0 +1,2 @@
+snd-ua101-objs := ua101.o
+obj-$(CONFIG_SND_USB_UA101) += snd-ua101.o
sound/usb/ua101/ua101.c renamed from sound/usb/ua101.c +1 -1
similarity index 99%
rename from sound/usb/ua101.c
rename to sound/usb/ua101/ua101.c
index 4f4ccdf..c7dcfe2 100644
@@ -23,7 +23,7 @@
 #include <sound/initval.h>
 #include <sound/pcm.h>
 #include <sound/pcm_params.h>
-#include "usbaudio.h"
+#include "../usbaudio.h"
 
 MODULE_DESCRIPTION("Edirol UA-101 driver");
 MODULE_AUTHOR("Clemens Ladisch <[email protected]>");
Signature · 1 line
1.6.6.2
#2 lore raw

PATCH 2/3 ALSA: usb: header file cleanups

Rename snd-usb-lib to snd-usbmidi-lib as MIDI functions are the only thing it actually contains. Introduce a new header file to only declare these functions.

Introduced usbmixer.h for all functions exported by usbmixer.c.

Cc Clemens Ladisch <[email protected]>
Cc Takashi Iwai <[email protected]>
---
 sound/usb/Makefile         |   11 +++++----
 sound/usb/ua101/ua101.c    |    1 +
 sound/usb/usbaudio.c       |    3 +-
 sound/usb/usbaudio.h       |   51 --------------------------------------------
 sound/usb/usbmidi.c        |    1 +
 sound/usb/usbmidi.h        |   48 +++++++++++++++++++++++++++++++++++++++++
 sound/usb/usbmixer.c       |    1 +
 sound/usb/usbmixer.h       |   11 +++++++++
 sound/usb/usx2y/us122l.c   |    1 +
 sound/usb/usx2y/usbusx2y.h |    1 +
 10 files changed, 72 insertions(+), 57 deletions(-)
 create mode 100644 sound/usb/usbmidi.h
 create mode 100644 sound/usb/usbmixer.h
sound/usb/Makefile +6 -5
index 8bc005b..dc31306 100644
@@ -3,12 +3,13 @@
 #
 
 snd-usb-audio-objs := usbaudio.o usbmixer.o
-snd-usb-lib-objs := usbmidi.o
+snd-usbmidi-lib-objs := usbmidi.o
 
 # Toplevel Module Dependency
-obj-$(CONFIG_SND_USB_AUDIO) += snd-usb-audio.o snd-usb-lib.o
-obj-$(CONFIG_SND_USB_UA101) += snd-usb-lib.o
-obj-$(CONFIG_SND_USB_USX2Y) += snd-usb-lib.o
-obj-$(CONFIG_SND_USB_US122L) += snd-usb-lib.o
+obj-$(CONFIG_SND_USB_AUDIO) += snd-usb-audio.o snd-usbmidi-lib.o
+
+obj-$(CONFIG_SND_USB_UA101) += snd-usbmidi-lib.o
+obj-$(CONFIG_SND_USB_USX2Y) += snd-usbmidi-lib.o
+obj-$(CONFIG_SND_USB_US122L) += snd-usbmidi-lib.o
 
 obj-$(CONFIG_SND) += ua101/ usx2y/ caiaq/
sound/usb/ua101/ua101.c +1 -0
index c7dcfe2..34a55c3 100644
@@ -24,6 +24,7 @@
 #include <sound/pcm.h>
 #include <sound/pcm_params.h>
 #include "../usbaudio.h"
+#include "../usbmidi.h"
 
 MODULE_DESCRIPTION("Edirol UA-101 driver");
 MODULE_AUTHOR("Clemens Ladisch <[email protected]>");
sound/usb/usbaudio.c +2 -1
index c539f7f..02870d5 100644
@@ -56,7 +56,8 @@
 #include <sound/initval.h>
 
 #include "usbaudio.h"
-
+#include "usbmidi.h"
+#include "usbmixer.h"
 
 MODULE_AUTHOR("Takashi Iwai <[email protected]>");
 MODULE_DESCRIPTION("USB Audio");
sound/usb/usbaudio.h +0 -51

69 lines not on this page · raw message

sound/usb/usbmidi.c +1 -0
index 2c59afd..5915a04 100644
@@ -53,6 +53,7 @@
 #include <sound/rawmidi.h>
 #include <sound/asequencer.h>
 #include "usbaudio.h"
+#include "usbmidi.h"
 
 
 /*
sound/usb/usbmidi.h new file +48 -0

48 lines not on this page · raw message

sound/usb/usbmixer.c +1 -0
index 8e8f871..43d53a3 100644
@@ -41,6 +41,7 @@
 #include <sound/tlv.h>
 
 #include "usbaudio.h"
+#include "usbmixer.h"
 
 /*
  */
sound/usb/usbmixer.h new file +11 -0
new file mode 100644
index 0000000..e199e4b
@@ -0,0 +1,11 @@
+#ifndef __USBMIXER_H
+#define __USBMIXER_H
+
+int snd_usb_create_mixer(struct snd_usb_audio *chip, int ctrlif,
+			 int ignore_error);
+void snd_usb_mixer_disconnect(struct list_head *p);
+
+void snd_emuusb_set_samplerate(struct snd_usb_audio *chip,
+			unsigned char samplerate_id);
+
+#endif /* __USBMIXER_H */
sound/usb/usx2y/us122l.c +1 -0
index 44deb21..4f6518c 100644
@@ -25,6 +25,7 @@
 #define MODNAME "US122L"
 #include "usb_stream.c"
 #include "../usbaudio.h"
+#include "../usbmidi.h"
 #include "us122l.h"
 
 MODULE_AUTHOR("Karsten Wiese <[email protected]>");
sound/usb/usx2y/usbusx2y.h +1 -0
index 1d174ce..9ab97b4 100644
@@ -1,6 +1,7 @@
 #ifndef USBUSX2Y_H
 #define USBUSX2Y_H
 #include "../usbaudio.h"
+#include "../usbmidi.h"
 #include "usbus428ctldefs.h" 
 
 #define NRURBS	        2	
Signature · 1 line
1.6.6.2

A message that is not in the archive (sent off-list or not yet received): [email protected]

#4 lore raw

PATCH 1/3 ALSA: usb-audio: rename substream format field to altset_idx

The snd_usb_substream::format field actually contains the index of the current alternate setting, so rename it to altset_idx to avoid confusion.

sound/usb/card.h +1 -1
@@ -54,7 +54,7 @@ struct snd_usb_substream {
 	struct audioformat *cur_audiofmt;	/* current audioformat pointer (for hw_params callback) */
 	unsigned int cur_rate;		/* current rate (for hw_params callback) */
 	unsigned int period_bytes;	/* current period bytes (for hw_params callback) */
-	unsigned int format;     /* USB data format */
+	unsigned int altset_idx;     /* USB data format: index of alternate setting */
 	unsigned int datapipe;   /* the data i/o pipe */
 	unsigned int syncpipe;   /* 1 - async out or adaptive in */
 	unsigned int datainterval;	/* log_2 of data packet interval */
sound/usb/pcm.c +4 -4
@@ -202,11 +202,11 @@ static int set_format(struct snd_usb_sub
 			return -EIO;
 		}
 		subs->interface = -1;
-		subs->format = 0;
+		subs->altset_idx = 0;
 	}
 
 	/* set interface */
-	if (subs->interface != fmt->iface || subs->format != fmt->altset_idx) {
+	if (subs->interface != fmt->iface || subs->altset_idx != fmt->altset_idx) {
 		if (usb_set_interface(dev, fmt->iface, fmt->altsetting) < 0) {
 			snd_printk(KERN_ERR "%d:%d:%d: usb_set_interface failed\n",
 				   dev->devnum, fmt->iface, fmt->altsetting);
@@ -214,7 +214,7 @@ static int set_format(struct snd_usb_sub
 		}
 		snd_printdd(KERN_INFO "setting usb interface %d:%d\n", fmt->iface, fmt->altsetting);
 		subs->interface = fmt->iface;
-		subs->format = fmt->altset_idx;
+		subs->altset_idx = fmt->altset_idx;
 	}
 
 	/* create a data pipe */
@@ -771,7 +771,7 @@ static int snd_usb_pcm_open(struct snd_p
 	struct snd_usb_substream *subs = &as->substream[direction];
 
 	subs->interface = -1;
-	subs->format = 0;
+	subs->altset_idx = 0;
 	runtime->hw = snd_usb_hardware;
 	runtime->private_data = subs;
 	subs->pcm_substream = substream;
sound/usb/proc.c +1 -1
@@ -116,7 +116,7 @@ static void proc_dump_substream_status(s
 		unsigned int i;
 		snd_iprintf(buffer, "  Status: Running\n");
 		snd_iprintf(buffer, "    Interface = %d\n", subs->interface);
-		snd_iprintf(buffer, "    Altset = %d\n", subs->format);
+		snd_iprintf(buffer, "    Altset = %d\n", subs->altset_idx);
 		snd_iprintf(buffer, "    URBs = %d [ ", subs->nurbs);
 		for (i = 0; i < subs->nurbs; i++)
 			snd_iprintf(buffer, "%d ", subs->dataurb[i].packets);
#5 lore raw

PATCH 2/3 ALSA: usb-audio: use a format bitmask per alternate setting

In reply to a message not in the archive

In preparation for USB audio 2.0 support, change the audioformat structure so that it uses a bitmask to specify possible formats.

sound/usb/card.h +1 -1
@@ -9,7 +9,7 @@
 
 struct audioformat {
 	struct list_head list;
-	snd_pcm_format_t format;	/* format type */
+	u64 formats;			/* ALSA format bits */
 	unsigned int channels;		/* # channels */
 	unsigned int fmt_type;		/* USB audio format type (1-3) */
 	unsigned int frame_size;	/* samples per frame for non-audio */
sound/usb/endpoint.c +2 -2
@@ -94,7 +94,7 @@ int snd_usb_add_audio_endpoint(struct sn
 		if (subs->endpoint == fp->endpoint) {
 			list_add_tail(&fp->list, &subs->fmt_list);
 			subs->num_formats++;
-			subs->formats |= 1ULL << fp->format;
+			subs->formats |= fp->formats;
 			return 0;
 		}
 	}
@@ -268,7 +268,7 @@ int snd_usb_parse_audio_endpoints(struct
 		 */
 		if (fmt[4] == 1 && fmt[5] == 2 && altno == 2 && num == 3 &&
 		    fp && fp->altsetting == 1 && fp->channels == 1 &&
-		    fp->format == SNDRV_PCM_FORMAT_S16_LE &&
+		    fp->formats == SNDRV_PCM_FMTBIT_S16_LE &&
 		    protocol == UAC_VERSION_1 &&
 		    le16_to_cpu(get_endpoint(alts, 0)->wMaxPacketSize) ==
 							fp->maxpacksize * 2)
sound/usb/format.c +5 -5
@@ -319,7 +319,7 @@ static int parse_audio_format_i(struct s
 			return -1;
 	}
 
-	fp->format = pcm_format;
+	fp->formats = 1uLL << pcm_format;
 
 	/* gather possible sample rates */
 	/* audio class v1 reports possible sample rates as part of the
@@ -361,16 +361,16 @@ static int parse_audio_format_ii(struct 
 	switch (format) {
 	case UAC_FORMAT_TYPE_II_AC3:
 		/* FIXME: there is no AC3 format defined yet */
-		// fp->format = SNDRV_PCM_FORMAT_AC3;
-		fp->format = SNDRV_PCM_FORMAT_U8; /* temporarily hack to receive byte streams */
+		// fp->formats = SNDRV_PCM_FMTBIT_AC3;
+		fp->formats = SNDRV_PCM_FMTBIT_U8; /* temporary hack to receive byte streams */
 		break;
 	case UAC_FORMAT_TYPE_II_MPEG:
-		fp->format = SNDRV_PCM_FORMAT_MPEG;
+		fp->formats = SNDRV_PCM_FMTBIT_MPEG;
 		break;
 	default:
 		snd_printd(KERN_INFO "%d:%u:%d : unknown format tag %#x is detected.  processed as MPEG.\n",
 			   chip->dev->devnum, fp->iface, fp->altsetting, format);
-		fp->format = SNDRV_PCM_FORMAT_MPEG;
+		fp->formats = SNDRV_PCM_FMTBIT_MPEG;
 		break;
 	}
 
sound/usb/pcm.c +10 -3
@@ -58,7 +58,9 @@ static struct audioformat *find_format(s
 	list_for_each(p, &subs->fmt_list) {
 		struct audioformat *fp;
 		fp = list_entry(p, struct audioformat, list);
-		if (fp->format != format || fp->channels != channels)
+		if (!(fp->formats & (1uLL << format)))
+			continue;
+		if (fp->channels != channels)
 			continue;
 		if (rate < fp->rate_min || rate > fp->rate_max)
 			continue;
@@ -428,10 +430,15 @@ static int hw_check_valid_format(struct 
 	struct snd_interval *ct = hw_param_interval(params, SNDRV_PCM_HW_PARAM_CHANNELS);
 	struct snd_mask *fmts = hw_param_mask(params, SNDRV_PCM_HW_PARAM_FORMAT);
 	struct snd_interval *pt = hw_param_interval(params, SNDRV_PCM_HW_PARAM_PERIOD_TIME);
+	struct snd_mask check_fmts;
 	unsigned int ptime;
 
 	/* check the format */
-	if (!snd_mask_test(fmts, fp->format)) {
+	snd_mask_none(&check_fmts);
+	check_fmts.bits[0] = (u32)fp->formats;
+	check_fmts.bits[1] = (u32)(fp->formats >> 32);
+	snd_mask_intersect(&check_fmts, fmts);
+	if (snd_mask_empty(&check_fmts)) {
 		hwc_debug("   > check: no supported format %d\n", fp->format);
 		return 0;
 	}
@@ -584,7 +591,7 @@ static int hw_rule_format(struct snd_pcm
 		fp = list_entry(p, struct audioformat, list);
 		if (!hw_check_valid_format(subs, params, fp))
 			continue;
-		fbits |= (1ULL << fp->format);
+		fbits |= fp->formats;
 	}
 
 	oldbits[0] = fmt->bits[0];
sound/usb/proc.c +7 -2
@@ -79,11 +79,16 @@ static void proc_dump_substream_formats(
 
 	list_for_each(p, &subs->fmt_list) {
 		struct audioformat *fp;
+		snd_pcm_format_t fmt;
 		fp = list_entry(p, struct audioformat, list);
 		snd_iprintf(buffer, "  Interface %d\n", fp->iface);
 		snd_iprintf(buffer, "    Altset %d\n", fp->altsetting);
-		snd_iprintf(buffer, "    Format: %s\n",
-			    snd_pcm_format_name(fp->format));
+		snd_iprintf(buffer, "    Format:");
+		for (fmt = 0; fmt <= SNDRV_PCM_FORMAT_LAST; ++fmt)
+			if (fp->formats & (1uLL << fmt))
+				snd_iprintf(buffer, " %s",
+					    snd_pcm_format_name(fmt));
+		snd_iprintf(buffer, "\n");
 		snd_iprintf(buffer, "    Channels: %d\n", fp->channels);
 		snd_iprintf(buffer, "    Endpoint: %d %s (%s)\n",
 			    fp->endpoint & USB_ENDPOINT_NUMBER_MASK,
sound/usb/quirks-table.h +6 -6
@@ -279,7 +279,7 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 				.ifnum = 0,
 				.type = QUIRK_AUDIO_FIXED_ENDPOINT,
 				.data = & (const struct audioformat) {
-					.format = SNDRV_PCM_FORMAT_S16_LE,
+					.formats = SNDRV_PCM_FMTBIT_S16_LE,
 					.channels = 4,
 					.iface = 0,
 					.altsetting = 1,
@@ -296,7 +296,7 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 				.ifnum = 1,
 				.type = QUIRK_AUDIO_FIXED_ENDPOINT,
 				.data = & (const struct audioformat) {
-					.format = SNDRV_PCM_FORMAT_S16_LE,
+					.formats = SNDRV_PCM_FMTBIT_S16_LE,
 					.channels = 2,
 					.iface = 1,
 					.altsetting = 1,
@@ -580,7 +580,7 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 				.ifnum = 0,
 				.type = QUIRK_AUDIO_FIXED_ENDPOINT,
 				.data = & (const struct audioformat) {
-					.format = SNDRV_PCM_FORMAT_S24_3LE,
+					.formats = SNDRV_PCM_FMTBIT_S24_3LE,
 					.channels = 2,
 					.iface = 0,
 					.altsetting = 1,
@@ -597,7 +597,7 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 				.ifnum = 1,
 				.type = QUIRK_AUDIO_FIXED_ENDPOINT,
 				.data = & (const struct audioformat) {
-					.format = SNDRV_PCM_FORMAT_S24_3LE,
+					.formats = SNDRV_PCM_FMTBIT_S24_3LE,
 					.channels = 2,
 					.iface = 1,
 					.altsetting = 1,
@@ -793,7 +793,7 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 				.ifnum = 1,
 				.type = QUIRK_AUDIO_FIXED_ENDPOINT,
 				.data = & (const struct audioformat) {
-					.format = SNDRV_PCM_FORMAT_S24_3LE,
+					.formats = SNDRV_PCM_FMTBIT_S24_3LE,
 					.channels = 2,
 					.iface = 1,
 					.altsetting = 1,
@@ -810,7 +810,7 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 				.ifnum = 2,
 				.type = QUIRK_AUDIO_FIXED_ENDPOINT,
 				.data = & (const struct audioformat) {
-					.format = SNDRV_PCM_FORMAT_S24_3LE,
+					.formats = SNDRV_PCM_FMTBIT_S24_3LE,
 					.channels = 2,
 					.iface = 2,
 					.altsetting = 1,
sound/usb/quirks.c +1 -1
@@ -174,7 +174,7 @@ static int create_uaxx_quirk(struct snd_
 			     const struct snd_usb_audio_quirk *quirk)
 {
 	static const struct audioformat ua_format = {
-		.format = SNDRV_PCM_FORMAT_S24_3LE,
+		.formats = SNDRV_PCM_FMTBIT_S24_3LE,
 		.channels = 2,
 		.fmt_type = UAC_FORMAT_TYPE_I,
 		.altsetting = 1,
sound/usb/urb.c +2 -2
@@ -662,7 +662,7 @@ static int prepare_nodata_playback_urb(s
 	urb->number_of_packets = ctx->packets;
 	urb->transfer_buffer_length = offs * stride;
 	memset(urb->transfer_buffer,
-	       subs->cur_audiofmt->format == SNDRV_PCM_FORMAT_U8 ? 0x80 : 0,
+	       runtime->format == SNDRV_PCM_FORMAT_U8 ? 0x80 : 0,
 	       offs * stride);
 	return 0;
 }
@@ -924,7 +924,7 @@ void snd_usb_init_substream(struct snd_u
 	snd_usb_set_pcm_ops(as->pcm, stream);
 
 	list_add_tail(&fp->list, &subs->fmt_list);
-	subs->formats |= 1ULL << fp->format;
+	subs->formats |= fp->formats;
 	subs->endpoint = fp->endpoint;
 	subs->num_formats++;
 	subs->fmt_type = fp->fmt_type;
#6 lore raw

PATCH 3/3 ALSA: usb-audio: support multiple formats with audio class v2 devices

In reply to a message not in the archive

Change the parser to correctly handle v2 descriptors with multiple format bits set.

sound/usb/format.c +38 -55
@@ -37,19 +37,20 @@
  * @format: the format tag (wFormatTag)
  * @fmt: the format type descriptor
  */
-static int parse_audio_format_i_type(struct snd_usb_audio *chip,
+static u64 parse_audio_format_i_type(struct snd_usb_audio *chip,
 				     struct audioformat *fp,
 				     int format, void *_fmt,
 				     int protocol)
 {
-	int pcm_format, i;
 	int sample_width, sample_bytes;
+	u64 pcm_formats;
 
 	switch (protocol) {
 	case UAC_VERSION_1: {
 		struct uac_format_type_i_discrete_descriptor *fmt = _fmt;
 		sample_width = fmt->bBitResolution;
 		sample_bytes = fmt->bSubframeSize;
+		format = 1 << format;
 		break;
 	}
 
@@ -57,24 +58,7 @@ static int parse_audio_format_i_type(str
 		struct uac_format_type_i_ext_descriptor *fmt = _fmt;
 		sample_width = fmt->bBitResolution;
 		sample_bytes = fmt->bSubslotSize;
-
-		/*
-		 * FIXME
-		 * USB audio class v2 devices specify a bitmap of possible
-		 * audio formats rather than one fix value. For now, we just
-		 * pick one of them and report that as the only possible
-		 * value for this setting.
-		 * The bit allocation map is in fact compatible to the
-		 * wFormatTag of the v1 AS streaming descriptors, which is why
-		 * we can simply map the matrix.
-		 */
-
-		for (i = 0; i < 5; i++)
-			if (format & (1UL << i)) {
-				format = i + 1;
-				break;
-			}
-
+		format <<= 1;
 		break;
 	}
 
@@ -82,15 +66,15 @@ static int parse_audio_format_i_type(str
 		return -EINVAL;
 	}
 
-	/* FIXME: correct endianess and sign? */
-	pcm_format = -1;
+	pcm_formats = 0;
 
-	switch (format) {
-	case UAC_FORMAT_TYPE_I_UNDEFINED: /* some devices don't define this correctly... */
+	if (format == 0 || format == (1 << UAC_FORMAT_TYPE_I_UNDEFINED)) {
+		/* some devices don't define this correctly... */
 		snd_printdd(KERN_INFO "%d:%u:%d : format type 0 is detected, processed as PCM\n",
 			    chip->dev->devnum, fp->iface, fp->altsetting);
-		/* fall-through */
-	case UAC_FORMAT_TYPE_I_PCM:
+		format = 1 << UAC_FORMAT_TYPE_I_PCM;
+	}
+	if (format & (1 << UAC_FORMAT_TYPE_I_PCM)) {
 		if (sample_width > sample_bytes * 8) {
 			snd_printk(KERN_INFO "%d:%u:%d : sample bitwidth %d in over sample bytes %d\n",
 				   chip->dev->devnum, fp->iface, fp->altsetting,
@@ -99,22 +83,22 @@ static int parse_audio_format_i_type(str
 		/* check the format byte size */
 		switch (sample_bytes) {
 		case 1:
-			pcm_format = SNDRV_PCM_FORMAT_S8;
+			pcm_formats |= SNDRV_PCM_FMTBIT_S8;
 			break;
 		case 2:
 			if (snd_usb_is_big_endian_format(chip, fp))
-				pcm_format = SNDRV_PCM_FORMAT_S16_BE; /* grrr, big endian!! */
+				pcm_formats |= SNDRV_PCM_FMTBIT_S16_BE; /* grrr, big endian!! */
 			else
-				pcm_format = SNDRV_PCM_FORMAT_S16_LE;
+				pcm_formats |= SNDRV_PCM_FMTBIT_S16_LE;
 			break;
 		case 3:
 			if (snd_usb_is_big_endian_format(chip, fp))
-				pcm_format = SNDRV_PCM_FORMAT_S24_3BE; /* grrr, big endian!! */
+				pcm_formats |= SNDRV_PCM_FMTBIT_S24_3BE; /* grrr, big endian!! */
 			else
-				pcm_format = SNDRV_PCM_FORMAT_S24_3LE;
+				pcm_formats |= SNDRV_PCM_FMTBIT_S24_3LE;
 			break;
 		case 4:
-			pcm_format = SNDRV_PCM_FORMAT_S32_LE;
+			pcm_formats |= SNDRV_PCM_FMTBIT_S32_LE;
 			break;
 		default:
 			snd_printk(KERN_INFO "%d:%u:%d : unsupported sample bitwidth %d in %d bytes\n",
@@ -122,30 +106,29 @@ static int parse_audio_format_i_type(str
 				   sample_width, sample_bytes);
 			break;
 		}
-		break;
-	case UAC_FORMAT_TYPE_I_PCM8:
-		pcm_format = SNDRV_PCM_FORMAT_U8;
-
+	}
+	if (format & (1 << UAC_FORMAT_TYPE_I_PCM8)) {
 		/* Dallas DS4201 workaround: it advertises U8 format, but really
 		   supports S8. */
 		if (chip->usb_id == USB_ID(0x04fa, 0x4201))
-			pcm_format = SNDRV_PCM_FORMAT_S8;
-		break;
-	case UAC_FORMAT_TYPE_I_IEEE_FLOAT:
-		pcm_format = SNDRV_PCM_FORMAT_FLOAT_LE;
-		break;
-	case UAC_FORMAT_TYPE_I_ALAW:
-		pcm_format = SNDRV_PCM_FORMAT_A_LAW;
-		break;
-	case UAC_FORMAT_TYPE_I_MULAW:
-		pcm_format = SNDRV_PCM_FORMAT_MU_LAW;
-		break;
-	default:
-		snd_printk(KERN_INFO "%d:%u:%d : unsupported format type %d\n",
+			pcm_formats |= SNDRV_PCM_FMTBIT_S8;
+		else
+			pcm_formats |= SNDRV_PCM_FMTBIT_U8;
+	}
+	if (format & (1 << UAC_FORMAT_TYPE_I_IEEE_FLOAT)) {
+		pcm_formats |= SNDRV_PCM_FMTBIT_FLOAT_LE;
+	}
+	if (format & (1 << UAC_FORMAT_TYPE_I_ALAW)) {
+		pcm_formats |= SNDRV_PCM_FMTBIT_A_LAW;
+	}
+	if (format & (1 << UAC_FORMAT_TYPE_I_MULAW)) {
+		pcm_formats |= SNDRV_PCM_FMTBIT_MU_LAW;
+	}
+	if (format & ~0x3f) {
+		snd_printk(KERN_INFO "%d:%u:%d : unsupported format bits %#x\n",
 			   chip->dev->devnum, fp->iface, fp->altsetting, format);
-		break;
 	}
-	return pcm_format;
+	return pcm_formats;
 }
 
 
@@ -313,14 +296,14 @@ static int parse_audio_format_i(struct s
 		default:
 			pcm_format = SNDRV_PCM_FORMAT_S16_LE;
 		}
+		fp->formats = 1uLL << pcm_format;
 	} else {
-		pcm_format = parse_audio_format_i_type(chip, fp, format, fmt, protocol);
-		if (pcm_format < 0)
+		fp->formats = parse_audio_format_i_type(chip, fp, format,
+							fmt, protocol);
+		if (!fp->formats)
 			return -1;
 	}
 
-	fp->formats = 1uLL << pcm_format;
-
 	/* gather possible sample rates */
 	/* audio class v1 reports possible sample rates as part of the
 	 * proprietary class specific descriptor.
#7 lore raw

Re: [PATCH 3/3] ALSA: usb-audio: support multiple formats with audio class v2 devices

Quoted 4 lines · Clemens Ladisch

Change the parser to correctly handle v2 descriptors with multiple format bits set.

I got compile errors with that one, and needed the patch below to fix them. Want to squash that into your commit and resend?

Daniel

Quoted 1 line

From 56ee3facfc825e68ed49c6cefac644d33b82dd71 Mon Sep 17 00:00:00 2001

From: Daniel Mack <[email protected]>
Date: Mon, 1 Mar 2010 15:22:54 +0100
Subject: [PATCH] ALSA: usb-audio: fix quirk formats

'struct audioformat' takes a bit field of supported rates now rather than one fixed value. Adopt that change at two more places.

---
 sound/usb/quirks-table.h |    2 +-
 sound/usb/quirks.c       |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
sound/usb/quirks-table.h +1 -1
index ff2df4e..2107d00 100644
@@ -2233,7 +2233,7 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 				.ifnum = 1,
 				.type = QUIRK_AUDIO_FIXED_ENDPOINT,
 				.data = &(const struct audioformat) {
-					.format = SNDRV_PCM_FORMAT_S24_3BE,
+					.formats = SNDRV_PCM_FMTBIT_S24_3BE,
 					.channels = 2,
 					.iface = 1,
 					.altsetting = 1,
sound/usb/quirks.c +1 -1
index 00e8a97..0119fa0 100644
@@ -89,7 +89,7 @@ static int create_ua1000_quirk(struct snd_usb_audio *chip,
 			       const struct snd_usb_audio_quirk *quirk)
 {
 	static const struct audioformat ua1000_format = {
-		.format = SNDRV_PCM_FORMAT_S32_LE,
+		.formats = SNDRV_PCM_FMTBIT_S32_LE,
 		.fmt_type = UAC_FORMAT_TYPE_I,
 		.altsetting = 1,
 		.altset_idx = 1,
Signature · 1 line
1.6.6.2
#8 lore raw

Re: [PATCH 3/3] ALSA: usb-audio: support multiple formats with audio class v2 devices

Quoted 6 lines · Daniel Mack
Quoted 2 lines · Clemens Ladisch

Change the parser to correctly handle v2 descriptors with multiple format bits set.

I got compile errors with that one, and needed the patch below to fix them.

Sorry, my branch already moved the UA-1000 support into the ua101 driver.

Quoted 1 line

Want to squash that into your commit and resend?

I guess the UA-1000 change will be merged sooner.

Regards,
Clemens

#9 lore raw

Re: [PATCH 1/3] ALSA: usb: move ua101 driver

In reply to Daniel Mack
Quoted 2 lines · Daniel Mack

As part of the USB audio code cleanup, move the non-standard ua101 driver out of the way.

A directory with a single source file in it isn't quite an improvement. Please name it "misc" so that other specialized drivers could be put in there.

Regards,
Clemens

#10 lore raw

Re: [PATCH 1/3] ALSA: usb: move ua101 driver

Quoted 7 lines · Clemens Ladisch
Quoted 2 lines · Daniel Mack

As part of the USB audio code cleanup, move the non-standard ua101 driver out of the way.

A directory with a single source file in it isn't quite an improvement. Please name it "misc" so that other specialized drivers could be put in there.

Ok, will do. However, I wonder which way is best to do that. Given that you posted patches on top of mine, can I take this as a general agreement to my changeset? If so, I will just provide a patch on top to do the ua101 -> misc rename, because otherwise, I would have to rebase all three patches for that.

Daniel

#11 lore raw

Re: [PATCH 1/3] ALSA: usb: move ua101 driver

At Mon, 1 Mar 2010 15:01:29 +0100,

Quoted 10 lines · Daniel Mack
Quoted 7 lines · Clemens Ladisch
Quoted 2 lines · Daniel Mack

As part of the USB audio code cleanup, move the non-standard ua101 driver out of the way.

A directory with a single source file in it isn't quite an improvement. Please name it "misc" so that other specialized drivers could be put in there.

Quoted 5 lines

Ok, will do. However, I wonder which way is best to do that. Given that you posted patches on top of mine, can I take this as a general agreement to my changeset? If so, I will just provide a patch on top to do the ua101 -> misc rename, because otherwise, I would have to rebase all three patches for that.

Well, honestly, I prefer seeing the fixes by Clemens based on the current code, and merge Daniel's refactoring patches after 2.6.34 merge window.

Refactoring is a good thing but it can easily introduce any careless bugs (like copy&paste error), and this split action makes really hard to track the changes. Thus I'd like to keep this well ripened before reaching to the Linus tree, not only for a few days.

thanks,

Takashi

#12 lore raw

Re: [PATCH 1/3] ALSA: usb: move ua101 driver

Quoted 9 lines · Takashi Iwai

At Mon, 1 Mar 2010 15:01:29 +0100,

Quoted 6 lines · Daniel Mack

Ok, will do. However, I wonder which way is best to do that. Given that you posted patches on top of mine, can I take this as a general agreement to my changeset? If so, I will just provide a patch on top to do the ua101 -> misc rename, because otherwise, I would have to rebase all three patches for that.

Quoted 3 lines

Well, honestly, I prefer seeing the fixes by Clemens based on the current code, and merge Daniel's refactoring patches after 2.6.34 merge window.

Clemens fixes only make sense for v2 of the USB audio spec, which isn't fully supported yet. IOW, the code as it will go into 2.6.34 is not able to support that, with or without my refactoring. So rebasing them doesn't make sense IMO.

Quoted 4 lines

Refactoring is a good thing but it can easily introduce any careless bugs (like copy&paste error), and this split action makes really hard to track the changes. Thus I'd like to keep this well ripened before reaching to the Linus tree, not only for a few days.

Yes, certainly. I didn't expect that to be taken for .34.

I would be fine with having it in a branch on your side, and getting it merged to your for-next after any .34 issues are sorted out.

There will be quite some more patches to fully support the most common features, but they won't be as intrusive and big of course.

Daniel

#13 lore raw

Re: [PATCH 1/3] ALSA: usb: move ua101 driver

At Mon, 1 Mar 2010 15:18:48 +0100,

Quoted 11 lines · Daniel Mack
Quoted 9 lines · Takashi Iwai

At Mon, 1 Mar 2010 15:01:29 +0100,

Quoted 6 lines · Daniel Mack

Ok, will do. However, I wonder which way is best to do that. Given that you posted patches on top of mine, can I take this as a general agreement to my changeset? If so, I will just provide a patch on top to do the ua101 -> misc rename, because otherwise, I would have to rebase all three patches for that.

Quoted 8 lines
Quoted 3 lines

Well, honestly, I prefer seeing the fixes by Clemens based on the current code, and merge Daniel's refactoring patches after 2.6.34 merge window.

Clemens fixes only make sense for v2 of the USB audio spec, which isn't fully supported yet. IOW, the code as it will go into 2.6.34 is not able to support that, with or without my refactoring. So rebasing them doesn't make sense IMO.

OK, then we can merge them together later.

Quoted 6 lines
Quoted 4 lines

Refactoring is a good thing but it can easily introduce any careless bugs (like copy&paste error), and this split action makes really hard to track the changes. Thus I'd like to keep this well ripened before reaching to the Linus tree, not only for a few days.

Yes, certainly. I didn't expect that to be taken for .34.

Good that we agree :)

Quoted 5 lines

I would be fine with having it in a branch on your side, and getting it merged to your for-next after any .34 issues are sorted out.

There will be quite some more patches to fully support the most common features, but they won't be as intrusive and big of course.

Fair enough.

thanks,

Takashi

#14 lore raw

Re: [PATCH 1/3] ALSA: usb: move ua101 driver

Quoted 9 lines · Takashi Iwai

At Mon, 1 Mar 2010 15:18:48 +0100,

Quoted 5 lines · Daniel Mack

Clemens fixes only make sense for v2 of the USB audio spec, which isn't fully supported yet. IOW, the code as it will go into 2.6.34 is not able to support that, with or without my refactoring. So rebasing them doesn't make sense IMO.

OK, then we can merge them together later.

Any plan on when to do that? I'm not in a hurry really, but I see the usbaudio driver moving on, so merging would need some manual backporting already. I can do that if you want and prepare a new patchset.

Btw - I'm curious how the commits e61e642c and 864c11080 made it to your tree. Wasn't the alsa-devel list copied for the patch or am I losing email?

Thanks,
Daniel

#15 lore raw

Re: [PATCH 1/3] ALSA: usb: move ua101 driver

At Thu, 4 Mar 2010 16:16:27 +0100,

Quoted 10 lines · Daniel Mack
Quoted 8 lines · Takashi Iwai

At Mon, 1 Mar 2010 15:18:48 +0100,

Quoted 5 lines · Daniel Mack

Clemens fixes only make sense for v2 of the USB audio spec, which isn't fully supported yet. IOW, the code as it will go into 2.6.34 is not able to support that, with or without my refactoring. So rebasing them doesn't make sense IMO.

Quoted 5 lines
Quoted 1 line

OK, then we can merge them together later.

Any plan on when to do that? I'm not in a hurry really, but I see the usbaudio driver moving on, so merging would need some manual backporting already. I can do that if you want and prepare a new patchset.

You can prepare patches based on topic/misc branch of sound git tree,
git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git topic/misc

or base on the latest Linus tree, which already contains USB audio v2.0 patches.

Then I'll create a separate branch (maybe again topic/usb-v2.0) and apply your patches, keep separate from linux-next branch.

Quoted 3 lines

Btw - I'm curious how the commits e61e642c and 864c11080 made it to your tree. Wasn't the alsa-devel list copied for the patch or am I losing email?

The former was a fix by Jaroslav directly done on git, and the latter came from LKML.

Takashi

#16 lore raw

Re: [PATCH 1/3] ALSA: usb: move ua101 driver

Quoted 6 lines · Takashi Iwai
Quoted 4 lines · Daniel Mack

Any plan on when to do that? I'm not in a hurry really, but I see the usbaudio driver moving on, so merging would need some manual backporting already. I can do that if you want and prepare a new patchset.

Quoted 8 lines

You can prepare patches based on topic/misc branch of sound git tree,
git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git topic/misc

or base on the latest Linus tree, which already contains USB audio v2.0 patches.

Then I'll create a separate branch (maybe again topic/usb-v2.0) and apply your patches, keep separate from linux-next branch.

Ok, will do. However, I wonder how new patches to the usbaudio driver will be handled then until the branch is merged?

Thanks,
Daniel

#17 lore raw

Re: [PATCH 1/3] ALSA: usb: move ua101 driver

At Thu, 4 Mar 2010 16:38:34 +0100,

Quoted 11 lines · Daniel Mack
Quoted 9 lines · Takashi Iwai
Quoted 4 lines · Daniel Mack

Any plan on when to do that? I'm not in a hurry really, but I see the usbaudio driver moving on, so merging would need some manual backporting already. I can do that if you want and prepare a new patchset.

You can prepare patches based on topic/misc branch of sound git tree,
git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git topic/misc

Quoted 8 lines
Quoted 5 lines

or base on the latest Linus tree, which already contains USB audio v2.0 patches.

Then I'll create a separate branch (maybe again topic/usb-v2.0) and apply your patches, keep separate from linux-next branch.

Ok, will do. However, I wonder how new patches to the usbaudio driver will be handled then until the branch is merged?

They are merged back to topic/usb-v2.0 branch. Conflicts are solved manually at merges.

Takashi