Re: [alsa-devel] [PATCH 1/1] ALSA: usb: initial USB Audio Device Class 3.0 support

From: Pierre-Louis Bossart
Date: Mon Nov 13 2017 - 11:07:18 EST

Next message: Matthew Wilcox: "RE: [PATCH] radix-tree: replace <linux/spinlock.h> with <linux/spinlock_types.h>"
Previous message: Michal Hocko: "Re: linux-next: Tree for Nov 7"
In reply to: Ruslan Bilovol: "Re: [alsa-devel] [PATCH 1/1] ALSA: usb: initial USB Audio Device Class 3.0 support"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On 11/10/17 8:48 PM, Ruslan Bilovol wrote:

On Wed, Nov 8, 2017 at 9:38 PM, Pierre-Louis Bossart
<pierre-louis.bossart@xxxxxxxxxxxxxxx> wrote:

Nice work, thanks! I double-checked all the descriptors and values and
didn't find anything problematic, the main comment I have is that the clock
source/selection could probably be refactored since the differences are
really minor with UAC2, and there is work to do to select the right Audio
Interface Association.

On 11/6/17 8:01 PM, Ruslan Bilovol wrote:

Recently released USB Audio Class 3.0 specification
introduces many significant changes comparing to
previous versions, like
- new Power Domains, support for LPM/L1
- new Cluster descriptor
- changed layout of all class-specific descriptors
- new High Capability descriptors
- New class-specific String descriptors
- new and removed units
- additional sources for interrupts
- removed Type II Audio Data Formats
- ... and many other things (check spec)

It also provides backward compatibility through
multiple configurations, as well as requires
mandatory support for BADD (Basic Audio Device
Definition) on each ADC3.0 compliant device

This patch adds initial support of UAC3 specification
that is enough for Generic I/O Profile (BAOF, BAIF)
device support from BADD document.

Do you mean to say that the selection of the BADD profile or full-blown
descriptor capabilities will be handled in a follow-up patch?

I have no follow-up patch for configuration switching.
I mean only that it is enough to support BAOF, BAIF profiles.
however, BAIOF profile won't work correctly, becaus it contains mixer unit
which is not supported yet in this patch.

This is because of way I use UAC3. As UAC3 device, I use an UAC3
gadget implementation which I sent before to linux-usb mailing list:
http://www.spinics.net/lists/linux-usb/msg162482.html

The UAC3 gadget implements BAOF+BAIF profile; but not BAIOF profile
which doesn't make sense in this case (there is no reason to mix Audio IN
with Audio Out)

sidetone is the main reason for mixing input on output, very useful too avoid speaking too loudly into the mic or when you use well-insulated headsets. most USB headsets provide this capability.

Thus I can't implement and test UAC3 mixer handling in current patch,
so BAIOF profile isn't supported by it yet.

It's my understanding from Section 3.3 that a UAC3 device is required to
expose
- a backwards-compatible configuration (UAC1 or UAC2),
- a UAC3 BADD profile
- one or more AIA compliant with UAC3 (which may have additional features
not present in the baseline description).

My understanding of Section 3.3 is slightly different. An UAC3 device is
required to expose:
- a backward-compatible first configuration (UAC1 or UAC2)
- an UAC3 BADD profile on another configuration, with only one AIA inside
- and _may_ have one or more configurations with UAC3 support that
provide functionality beyond what is available in BADD

yes, i should have written 0 or more AIA compliant with UAC3

And I am not sure how a driver would make the selection...

In order to test this patch, I created UAC2+UAC3 gadget and manually
switched to UAC3 configuration on Host by:
$ echo 2 > /sys/bus/usb/devices/1-1/bConfigurationValue

I don't think ALSA driver can make decision which configuration to select,
maybe some userspace tool can handle it, like usb_modeswitch does
for networking devices.

The BADD profiles were intended to provide a basic descriptors for hosts that couldn't handle descriptor parsing. That was a very controversial work item, since both the Linux and Windows stacks can parse descriptors the usefulness of BADD profiles remains questionable.

I think at some point we should have the following policy
if one UAC3 is available, use the first one
else
if there is a BADD profile use it
else
fall back to UAC1/2

I don't see how a user will ever choose between multiple UAC3 configurations, should they ever exist, except maybe by trial and error.

In theory though, a UAC3 device should work out of the box even if the host
only supports UAC1 or UAC2.

Yes, this is correct, I verified it as well.

+/*
+ * v1.0, v2.0 and v3.0 of this standard have many things in common. For
the rest
+ * of the definitions, please refer to audio.h and audio-v2.h
+ */
+
+/* All High Capability descriptors have these 2 fields at the beginning
*/
+struct uac3_hc_descriptor_header {
+ __le16 wLength;
+ __u8 bDescriptorType;
+ __u8 bDescriptorSubtype;
+ __le16 wDescriptorID;
+} __attribute__ ((packed));
+
+/* 4.3.1 CLUSTER DESCRIPTOR HEADER */
+struct uac3_cluster_header_descriptor {
+ __le16 wLength;
+ __u8 bDescriptorType;
+ __u8 bDescriptorSubtype;
+ __le16 wDescriptorID;
+ __u8 bNrChannels;
+} __attribute__ ((packed));
+
+/* 4.3.2.1 SEGMENTS */
+struct uac3_cluster_segment_descriptor {
+ __le16 wLength;
+ __u8 bSegmentType;
+ /* __u8[0]; segment-specific data */
+} __attribute__ ((packed));
+
+/* 4.3.2.1.1 END SEGMENT */
+struct uac3_cluster_end_segment_descriptor {
+ __le16 wLength;
+ __u8 bSegmentType; /* Constant END_SEGMENT */
+} __attribute__ ((packed));
+

you didn't include the definitions in 4.3.2.1.2
Cluster Descriptor Segment
Vendor Defined Segment

Correct, they are not used in BADD so I didn't care much, but I can
add it in next patchset

+/* 4.3.2.1.3.1 INFORMATION SEGMENT */
+struct uac3_cluster_information_segment_descriptor {
+ __le16 wLength;
+ __u8 bSegmentType;

this field is a CHANNEL_INFORMATION constant.

+ __u8 bChPurpose;
+ __u8 bChRelationship;
+ __u8 bChGroupID;
+} __attribute__ ((packed));
+
+/* 4.5.2 CLASS-SPECIFIC AC INTERFACE DESCRIPTOR */
+struct uac3_ac_header_descriptor {
+ __u8 bLength; /* 10 */
+ __u8 bDescriptorType; /* CS_INTERFACE descriptor type */
+ __u8 bDescriptorSubtype; /* HEADER descriptor subtype */
+ __u8 bCategory;
+
+ /* includes Clock Source, Unit, Terminal, and Power Domain desc.
*/
+ __le16 wTotalLength;
+
+ __le32 bmControls;
+} __attribute__ ((packed));
+

Missing extended terminal descriptor header from 4.5.2.3.2?

Missing multi-function processing unit descriptor header from 4.5.2.10.3 ?

Yes, I can add them as well.

+
+/* bmAttribute fields */
+#define UAC3_CLOCK_SOURCE_TYPE_EXT 0x0
+#define UAC3_CLOCK_SOURCE_TYPE_INT 0x1
+#define UAC3_CLOCK_SOURCE_ASYNC (0 << 2)
+#define UAC3_CLOCK_SOURCE_SYNCED_TO_SOF (1 << 1)
+
+/* 4.5.2.13 CLOCK SELECTOR DESCRIPTOR */
+struct uac3_clock_selector_descriptor {
+ __u8 bLength;
+ __u8 bDescriptorType;
+ __u8 bDescriptorSubtype;
+ __u8 bClockID;
+ __u8 bNrInPins;
+ __u8 baCSourceID[];
+ /* bmControls and wCSelectorDescrStr omitted */

why is this omitted here and not below in the clock multiplier descriptor?

That's because in this descriptor ther is variable (unknown) field baCSourceID,
so we can't define anythyng else after it as per C standard.
In the clock multiplier descriptor below we don't have such issue.

ah, then maybe explain it, I wasn't sure if there was a technical reason or a filter you applied for some other reason.

+} __attribute__((packed) > +
+/* 4.5.2.14 CLOCK MULTIPLIER DESCRIPTOR */
+struct uac3_clock_multiplier_descriptor {
+ __u8 bLength;
+ __u8 bDescriptorType;
+ __u8 bDescriptorSubtype;
+ __u8 bClockID;
+ __u8 bCSourceID;
+ __le32 bmControls;
+ __le16 wCMultiplierDescrStr;
+} __attribute__((packed));

[snip]

+
+/* 4.5.2.15 POWER DOMAIN DESCRIPTOR */
+struct uac3_power_domain_descriptor {
+ __u8 bLength;
+ __u8 bDescriptorType;
+ __u8 bDescriptorSubtype;
+ __u8 bPowerDomainID;
+ __le16 waRecoveryTime1;
+ __le16 waRecoveryTime2;
+ __u8 bNrEntities;
+ __u8 baEntityID[];
+ /* wPDomainDescrStr omitted */

why?

+} __attribute__((packed));
+
+/* As above, but more useful for defining your own descriptors */
+#define DECLARE_UAC3_POWER_DOMAIN_DESCRIPTOR(n) \
+struct uac3_power_domain_descriptor_##n { \
+ __u8 bLength; \
+ __u8 bDescriptorType; \
+ __u8 bDescriptorSubtype; \
+ __u8 bPowerDomainID; \
+ __le16 waRecoveryTime1; \
+ __le16 waRecoveryTime2; \
+ __u8 bNrEntities; \
+ __u8 baEntityID[n]; \
+ __le16 wPDomainDescrStr; \
+} __attribute__ ((packed))
+

any specific reason why the descriptors are not added in linear order,
following the spec definition? all the descriptors below could be added
earlier.

That's because of way I wrote this header. These descritors are added
in same order as in UAC2 header, then it was more easy to compare
both headers in order to understand if we can reuse anything from UAC2.

If course I should reorganize it, but forgot to do :D

I don't know why this was modified, we should have kept the initial order in the spec when possible.

+/* 4.5.2.1 INPUT TERMINAL DESCRIPTOR */
+struct uac3_input_terminal_descriptor {
+ __u8 bLength;
+ __u8 bDescriptorType;
+ __u8 bDescriptorSubtype;
+ __u8 bTerminalID;
+ __le16 wTerminalType;
+ __u8 bAssocTerminal;
+ __u8 bCSourceID;
+ __le32 bmControls;
+ __le16 wClusterDescrID;
+ __le16 wExTerminalDescrID;
+ __le16 wConnectorsDescrID;
+ __le16 wTerminalDescrStr;
+} __attribute__((packed));
+

+
+/* 4.7.2 CLASS-SPECIFIC AS INTERFACE DESCRIPTOR */
+struct uac3_as_header_descriptor {
+ __u8 bLength;
+ __u8 bDescriptorType;
+ __u8 bDescriptorSubtype;
+ __u8 bTerminalLink;
+ __le32 bmControls;
+ __le16 wClusterDescrID;
+ __le64 bmFormats;
+ __u8 bSubslotSize;
+ __u8 bBitResolution;
+ __le16 bmAuxProtocols;
+ __u8 bControlSize;
+} __attribute__((packed));
+
+#define UAC3_FORMAT_TYPE_I_RAW_DATA (1 << 6)

This seems to come from Table A1 in the formats document, is it inserted
here because of the relationship with the bmFormats? If yes, we should
probably add the full list from Table A1 (as done below for the other codes)

Yes, this comes from Frmts3.0. However, for bits 0..5 we reuse same
values from UAC1 spec, thay are same for UAC2 as well.
See UAC_FORMAT_TYPE_xxx

yes

+
+/* 4.8.1.2 CLASS-SPECIFIC AS ISOCHRONOUS AUDIO DATA ENDPOINT DESCRIPTOR
*/
+struct uac3_iso_endpoint_descriptor {
+ __u8 bLength;
+ __u8 bDescriptorType;
+ __u8 bDescriptorSubtype;
+ __le32 bmControls;
+ __u8 bLockDelayUnits;
+ __le16 wLockDelay;
+} __attribute__((packed));
+
+#define UAC3_CONTROL_PITCH (3 << 0)
+#define UAC3_CONTROL_DATA_OVERRUN (3 << 2)
+#define UAC3_CONTROL_DATA_UNDERRUN (3 << 4)

these are really masks, no?

Yes, same as in UAC2, we can drop them as they are not used yet.

+
+/* 6.1 INTERRUPT DATA MESSAGE */
+#define UAC3_INTERRUPT_DATA_MSG_VENDOR (1 << 0)
+#define UAC3_INTERRUPT_DATA_MSG_EP (1 << 1)

this is the same as in UAC2

+
+struct uac3_interrupt_data_msg {
+ __u8 bInfo;
+ __u8 bSourceType;
+ __le16 wValue;
+ __le16 wIndex;
+} __attribute__((packed));

This seems identical to UAC2, the difference being the bSourceType with
additional values (was bAttribute).

Yes, this is correct. I made separate structure just to make code reading
easier. Same is for UAC3_INTERRUPT_DATA_xxx above.

ok. maybe just mention that this is for readability

+
+/* A.2 AUDIO AUDIO FUNCTION SUBCLASS CODES */
+#define UAC3_FUNCTION_SUBCLASS_UNDEFINED 0x00
+#define UAC3_FUNCTION_SUBCLASS_FULL_ADC_3_0 0x01
+#define UAC3_FUNCTION_SUBCLASS_GENERIC_IO 0x20
+#define UAC3_FUNCTION_SUBCLASS_HEADPHONE 0x21
+#define UAC3_FUNCTION_SUBCLASS_SPEAKER 0x22
+#define UAC3_FUNCTION_SUBCLASS_MICROPHONE 0x23
+#define UAC3_FUNCTION_SUBCLASS_HEADSET 0x24
+#define UAC3_FUNCTION_SUBCLASS_HEADSET_ADAPTER 0x25
+#define UAC3_FUNCTION_SUBCLASS_SPEAKERPHONE 0x26

mention that these are linked to BADD profiles.

OK, will do it

+
+/* A.7 AUDIO FUNCTION CATEGORY CODES */
+#define UAC3_FUNCTION_SUBCLASS_UNDEFINED 0x00
+#define UAC3_FUNCTION_DESKTOP_SPEAKER 0x01
+#define UAC3_FUNCTION_HOME_THEATER 0x02
+#define UAC3_FUNCTION_MICROPHONE 0x03
+#define UAC3_FUNCTION_HEADSET 0x04
+#define UAC3_FUNCTION_TELEPHONE 0x05
+#define UAC3_FUNCTION_CONVERTER 0x06
+#define UAC3_FUNCTION_SOUND_RECORDER 0x07
+#define UAC3_FUNCTION_IO_BOX 0x08
+#define UAC3_FUNCTION_MUSICAL_INSTRUMENT 0x09
+#define UAC3_FUNCTION_PRO_AUDIO 0x0a
+#define UAC3_FUNCTION_AUDIO_VIDEO 0x0b
+#define UAC3_FUNCTION_CONTROL_PANEL 0x0c
+#define UAC3_FUNCTION_HEADPHONE 0x0d
+#define UAC3_FUNCTION_GENERIC_SPEAKER 0x0e
+#define UAC3_FUNCTION_HEADSET_ADAPTER 0x0f
+#define UAC3_FUNCTION_SPEAKERPHONE 0x10
+#define UAC3_FUNCTION_OTHER 0xff
+
+/* A.8 AUDIO CLASS-SPECIFIC DESCRIPTOR TYPES */
+#define UAC3_CS_UNDEFINED 0x20
+#define UAC3_CS_DEVICE 0x21
+#define UAC3_CS_CONFIGURATION 0x22
+#define UAC3_CS_STRING 0x23
+#define UAC3_CS_INTERFACE 0x24
+#define UAC3_CS_ENDPOINT 0x25
+#define UAC3_CS_CLUSTER 0x26
+
+/* A.10 CLUSTER DESCRIPTOR SEGMENT TYPES */
+#define UAC3_SEGMENT_UNDEFINED 0x00
+#define UAC3_CLUSTER_DESCRIPTION 0x01
+#define UAC3_CLUSTER_VENDOR_DEFINED 0x1F
+#define UAC3_CHANNEL_INFORMATION 0x20
+#define UAC3_CHANNEL_AMBISONIC 0x21
+#define UAC3_CHANNEL_DESCRIPTION 0x22
+#define UAC3_CHANNEL_VENDOR_DEFINED 0xFE
+#define UAC3_END_SEGMENT 0xFF
+
+/* A.11 CHANNEL PURPOSE DEFINITIONS */
+#define UAC3_PURPOSE_UNDEFINED 0x00
+#define UAC3_PURPOSE_GENERIC_AUDIO 0x01
+#define UAC3_PURPOSE_VOICE 0x02
+#define UAC3_PURPOSE_SPEECH 0x03
+#define UAC3_PURPOSE_AMBIENT 0x04
+#define UAC3_PURPOSE_REFERENCE 0x05
+#define UAC3_PURPOSE_ULTRASONIC 0x06
+#define UAC3_PURPOSE_VIBROKINETIC 0x07
+#define UAC3_PURPOSE_NON_AUDIO 0xFF
+
+/* A.12 CHANNEL RELATIONSHIP DEFINITIONS */
+/* FIXME: spec is missing these constants. Few found in
BasicAudioDevice3.pdf */

yes, this is a known bug in the released UAC3 document, the values were
removed by accident. If this helps, here's what I have from the last release
candidate, this will hopefully be corrected soon.

My next step was to send a follow-up email to usb.org about this issue,
but since you already have the answer, I'll include these values in the
next version of patch, thanks.

It doesn't hurt to provide feedback to USB since it'll show the need for a spec update.

[snip]

+#define UAC3_CH_RELATIONSHIP_UNDEFINED 0x00
+#define UAC3_CH_MONO 0x01
+#define UAC3_CH_LEFT 0x02
+#define UAC3_CH_RIGHT 0x03
+#define UAC3_CH_FRONT_LEFT 0x80
+#define UAC3_CH_FRONT_RIGHT 0x81
+#define UAC3_CH_FRONT_CENTER 0x82
+#define UAC3_CH_SURROUND_ARRAY_LEFT 0x89
+#define UAC3_CH_SURROUND_ARRAY_RIGHT 0x8A
+#define UAC3_CH_LOW_FREQUENCY_EFFECTS 0xB8
+
+/* A.15 AUDIO CLASS-SPECIFIC AC INTERFACE DESCRIPTOR SUBTYPES */
+/* see audio.h for the rest, which is identical to v1 */
+#define UAC3_EXTENDED_TERMINAL 0x04
+#define UAC3_MIXER_UNIT 0x05
+#define UAC3_SELECTOR_UNIT 0x06
+#define UAC3_FEATURE_UNIT 0x07
+#define UAC3_EFFECT_UNIT 0x08
+#define UAC3_PROCESSING_UNIT 0x09
+#define UAC3_EXTENSION_UNIT 0x0a
+#define UAC3_CLOCK_SOURCE 0x0b
+#define UAC3_CLOCK_SELECTOR 0x0c
+#define UAC3_CLOCK_MULTIPLIER 0x0d
+#define UAC3_SAMPLE_RATE_CONVERTER 0x0e
+#define UAC3_CONNECTORS 0x0f
+#define UAC3_POWER_DOMAIN 0x10
+
+/* A.22 AUDIO CLASS-SPECIFIC REQUEST CODES */
+#define UAC3_CS_REQ_CUR 0x01
+#define UAC3_CS_REQ_RANGE 0x02
+#define UAC3_CS_REQ_MEM 0x03

these first 3 are already in UAC2

+#define UAC3_CS_REQ_INTEN 0x04
+#define UAC3_CS_REQ_STRING 0x05
+#define UAC3_CS_REQ_HIGH_CAPABILITY_DESCRIPTOR 0x06

the last 3 are really UAC3-specific

Tha'ts correct. Do you want to say we can reuse UAC2 values?
This brings an issue I described previously in this mailing list.
It is related to UAC1/2/3 specs has some common values, sometimes
use different values for same things and this makes understanding
of sources quite complex if for example we will use UAC2_xx for
UAC3 in some cases for common things but UAC3_xx for UAC3-specifc.

I see 3 solutions:
1) use UAC2_xxx for common things (hard to read/write the code, you
have to remember what is common and what is not, or leave
a comment near each such place saying it's OK to use it for UAC3
2) rename UAC2_xxx to something like UAC_V2V3_xxx, so it will be
self-explanatory

I vote for this option.

3) create UAC3_xxx for thouse things that are common with UAC2,
so no need to do either 1 or 2, but it can lead to code duplication

In this patch I did some mix of 2 and 3, but additional comments are
welcome, as it variabl naming is the hardest part of programming :)

+
+/* A.23.1 AUDIOCONTROL INTERFACE CONTROL SELECTORS */
+#define UAC3_AC_CONTROL_UNDEFINED 0x00
+#define UAC3_AC_ACTIVE_INTERFACE_CONTROL 0x01
+#define UAC3_AC_POWER_DOMAIN_CONTROL 0x02
+
+#endif /* __LINUX_USB_AUDIO_V3_H */
diff --git a/include/uapi/linux/usb/audio.h
b/include/uapi/linux/usb/audio.h
index a4680a5..66ec2ae 100644
--- a/include/uapi/linux/usb/audio.h
+++ b/include/uapi/linux/usb/audio.h
@@ -26,6 +26,7 @@
/* bInterfaceProtocol values to denote the version of the standard used
*/
#define UAC_VERSION_1 0x00
#define UAC_VERSION_2 0x20
+#define UAC_VERSION_3 0x30
/* A.2 Audio Interface Subclass Codes */
#define USB_SUBCLASS_AUDIOCONTROL 0x01
diff --git a/sound/usb/card.c b/sound/usb/card.c
index 3dc36d9..df3f0ab 100644
--- a/sound/usb/card.c
+++ b/sound/usb/card.c
@@ -7,6 +7,7 @@
* Alan Cox (alan@xxxxxxxxxxxxxxxxxxx)
* Thomas Sailer (sailer@xxxxxxxxxxxxxx)
*
+ * Audio Class 3.0 support by Ruslan Bilovol <ruslan.bilovol@xxxxxxxxx>

does this mean new copyright?

That's a good qestion, I don't know the answer. Just wanted to leave
a comment about work I did for USB audio.

Don't know the answer either.

*
* 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
@@ -44,6 +45,7 @@
#include <linux/mutex.h>
#include <linux/usb/audio.h>
#include <linux/usb/audio-v2.h>
+#include <linux/usb/audio-v3.h>
#include <linux/module.h>
#include <sound/control.h>
@@ -261,7 +263,8 @@ static int snd_usb_create_streams(struct snd_usb_audio
*chip, int ctrlif)
break;
}
- case UAC_VERSION_2: {
+ case UAC_VERSION_2:
+ case UAC_VERSION_3: {
struct usb_interface_assoc_descriptor *assoc =
usb_ifnum_to_if(dev, ctrlif)->intf_assoc;
@@ -281,7 +284,7 @@ static int snd_usb_create_streams(struct
snd_usb_audio *chip, int ctrlif)
}
if (!assoc) {
- dev_err(&dev->dev, "Audio class v2 interfaces need
an interface association\n");
+ dev_err(&dev->dev, "Audio class v2/v3 interfaces
need an interface association\n");

yes, but they can have more than one, so how do you handle the selection
(see 3.3)?

Current implentation of USB audio driver is to create only one ALSA
card per USB device, that works with UAC1/UAC2 and now with UAC3;
so if USB device hase more that one interface association, only first
will be used.

But if you mean configuration switching (between backward
compatible UAC1/UAC2 and UAC3 configuration) that will not limit us,
since after switching to another configuration UAC1/UAC2 interfaces
will be released and UAC3 will be probed, this is done by USB core.

As described above, currently I switching between configs manually.

My hope is that we can have a basic policy to select the configurations, as detailed above.

[snip]

#include <sound/core.h>
#include <sound/pcm.h>
@@ -39,11 +40,11 @@
* @dev: usb device
* @fp: audioformat record
* @format: the format tag (wFormatTag)
- * @fmt: the format type descriptor
+ * @fmt: the format type descriptor (v1/v2) or AudioStreaming descriptor
(v3)
*/
static u64 parse_audio_format_i_type(struct snd_usb_audio *chip,
struct audioformat *fp,
- unsigned int format, void *_fmt)
+ u64 format, void *_fmt)
{
int sample_width, sample_bytes;
u64 pcm_formats = 0;
@@ -69,6 +70,17 @@ static u64 parse_audio_format_i_type(struct
snd_usb_audio *chip,
format <<= 1;
break;
}
+ case UAC_VERSION_3: {
+ struct uac3_as_header_descriptor *as = _fmt;
+ sample_width = as->bBitResolution;
+ sample_bytes = as->bSubslotSize;
+
+ if (format & UAC3_FORMAT_TYPE_I_RAW_DATA)
+ pcm_formats |= SNDRV_PCM_FMTBIT_SPECIAL;

could this be expanded to add at least basic PCM (D0 set).

Basic PCM, and more (D0..D5) are handled, this is needed for special
case of raw data, same as for UAC2, see commit 717bfb5 ("ALSA:
snd-usb: handle raw data format of UAC2 devices")

I don't get this one, you are only handing D6 here? where is D0..D5 handled?

+int snd_usb_parse_audio_format_v3(struct snd_usb_audio *chip,
+ struct audioformat *fp,
+ struct uac3_as_header_descriptor *as,
+ int stream)
+{
+ u64 format = le64_to_cpu(as->bmFormats);
+ int err;
+
+ /* Type I format bits are D0..D6 */
+ if (format & 0x7f)
+ fp->fmt_type = UAC_FORMAT_TYPE_I;
+ else
+ fp->fmt_type = UAC_FORMAT_TYPE_III;

maybe mention that this test only work because type IV is not supported,
otherwise this wouldn't quite right.

Yes, will do it.

+int snd_usb_parse_audio_format_v3(struct snd_usb_audio *chip,
+ struct audioformat *fp,
+ struct uac3_as_header_descriptor *as,
+ int stream);
#endif /* __USBAUDIO_FORMAT_H */
diff --git a/sound/usb/mixer.c b/sound/usb/mixer.c
index 9732edf..5b5dfe4 100644
--- a/sound/usb/mixer.c
+++ b/sound/usb/mixer.c
@@ -51,6 +51,7 @@
#include <linux/usb.h>
#include <linux/usb/audio.h>
#include <linux/usb/audio-v2.h>
+#include <linux/usb/audio-v3.h>
#include <sound/core.h>
#include <sound/control.h>
@@ -189,7 +190,7 @@ static void *find_audio_control_unit(struct
mixer_build *state,
USB_DT_CS_INTERFACE)) != NULL) {
if (hdr->bLength >= 4 &&
hdr->bDescriptorSubtype >= UAC_INPUT_TERMINAL &&
- hdr->bDescriptorSubtype <= UAC2_SAMPLE_RATE_CONVERTER
&&
+ hdr->bDescriptorSubtype <= UAC3_SAMPLE_RATE_CONVERTER
&&
hdr->bUnitID == unit)

this looks like a mistake, the definitions are

#define UAC3_SAMPLE_RATE_CONVERTER 0x0e
#define UAC3_CONNECTORS 0x0f
#define UAC3_POWER_DOMAIN 0x10

You will miss the last two with this test.

That's because connectors descriptor uses High Capability descriptor for
replresentation, thus it can't be compared here.
Power Domain descriptor isn't used yet by this driver, so let's check for
it once this functionality will be added.

Not sure I understand, the goal of this test is to check if the descriptors for the audio function look correct by looking at the range of their subType, which is also supported in High Capability descriptors?

Probably it makes sense to make this comparison UAC version depended
to not break existing UAC1/2 functionality, I'll check it.

- case UAC1_PROCESSING_UNIT:
- case UAC1_EXTENSION_UNIT:
- /* UAC2_PROCESSING_UNIT_V2 */
- /* UAC2_EFFECT_UNIT */
- case UAC2_EXTENSION_UNIT_V2: {
- struct uac_processing_unit_descriptor *d = p1;
-
- if (state->mixer->protocol == UAC_VERSION_2 &&
- hdr[2] == UAC2_EFFECT_UNIT) {
- /* UAC2/UAC1 unit IDs overlap here in an
- * uncompatible way. Ignore this unit for
now.
- */
+
+ /* REVISIT: UAC3 IT doesn't have
channels/cfg */
+ term->channels = 0;
+ term->chconfig = 0;

It does, but you need to get the information from the wClusterDescrID

Correct. Channels are used during mixer parsing which we don't support
yet, chconfig is set in many places but nobody uses it, so we can remove it from
usb_audio_term struct.

Next message: Matthew Wilcox: "RE: [PATCH] radix-tree: replace <linux/spinlock.h> with <linux/spinlock_types.h>"
Previous message: Michal Hocko: "Re: linux-next: Tree for Nov 7"
In reply to: Ruslan Bilovol: "Re: [alsa-devel] [PATCH 1/1] ALSA: usb: initial USB Audio Device Class 3.0 support"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]