Logitech USB headset not working in 3.6-rc3

Hi All,

We've had a report[1] that the Logitech USB headset 0003:046D:0A0C isn't working with 3.6-rc3. It seems the last working kernel was based on commit 10c63c9, and it first stopped working with a kernel based on commit 23dcfa6. There are only a few ALSA commits between those revisions, so hopefully this is something that is fairly easy to identify. The only commit to USB audio in that set is:

commit e9ba389c5ffc4dd29dfe17e00e48877302111135

Author: Takashi Iwai <[email protected]>
Date:   Wed Aug 15 12:32:00 2012 +0200

ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream

I've CC'd the reporter and attached the alsa-info is below.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=851619

upload=true&script=true&cardinfo=
!!################################
!!ALSA Information Script v 0.4.60
!!################################

!!Script ran on: Fri Aug 24 18:35:42 UTC 2012

!!Linux Distribution
!!------------------

Fedora release 18 (Rawhide) Fedora release 18 (Rawhide) NAME=Fedora ID=fedora PRETTY_NAME="Fedora 18 (Rawhide)" CPE_NAME="cpe:/o:fedoraproject:fedora:18" Fedora release 18 (Rawhide) Fedora release 18 (Rawhide)

!!DMI Information
!!---------------

Manufacturer:      Dell Inc.
Product Name:      Precision WorkStation 690

Product Version:

!!Kernel Information
!!------------------

Kernel release:    3.6.0-0.rc3.git0.1.fc18.x86_64
Operating System:  GNU/Linux
Architecture:      x86_64
Processor:         x86_64
SMP Enabled:       Yes

!!ALSA Version
!!------------

Driver version:     1.0.25
Library version:    1.0.25
Utilities version:  1.0.25

!!Loaded ALSA modules
!!-------------------

snd_hda_intel
snd_usb_audio

!!Sound Servers on this system
!!----------------------------

Pulseaudio:
Installed - Yes (/usr/bin/pulseaudio)
Running - Yes

aRts:
Installed - Yes (/usr/bin/artsd)
Running - No

Jack:
Installed - Yes (/usr/bin/jackd)
Running - No

!!Soundcards recognised by ALSA
!!-----------------------------

0 [Intel ]: HDA-Intel - HDA Intel

                      HDA Intel at 0xfcffc000 irq 66
 1 [Headset        ]: USB-Audio - Logitech USB Headset
                      Logitech Logitech USB Headset at usb-0000:00:1d.7-6.4, full speed

!!PCI Soundcards installed in the system
!!--------------------------------------

00:1b.0 Audio device: Intel Corporation 631xESB/632xESB High Definition Audio Controller (rev 09)

!!Advanced information - PCI Vendor/Device/Subsystem ID's !!--------------------------------------------------------

00:1b.0 0403: 8086:269a (rev 09)

        Subsystem: 1028:01c0

!!Loaded sound module options
!!--------------------------

!!Module: snd_hda_intel

        align_buffer_size : -1
        bdl_pos_adj : 1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
        beep_mode : N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N,N
        enable : Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y
        enable_msi : -1
        id : (null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null)
        index : -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
        model : (null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null)
        patch : (null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null)
        position_fix : 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
        power_save : 0
        power_save_controller : Y
        probe_mask : -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
        probe_only : 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
        single_cmd : N
        snoop : Y

!!Module: snd_usb_audio

        async_unlink : Y
        device_setup : 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
        enable : Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y,Y
        id : (null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null),(null)
        ignore_ctl_error : N
        index : -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
        nrpacks : 8
        pid : -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
        vid : -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1

!!HDA-Intel Codec information
!!---------------------------
--startcollapse--

Codec: SigmaTel STAC9200
Address: 0
AFG Function Id: 0x1 (unsol 1)
Vendor Id: 0x83847690
Subsystem Id: 0x102801c0
Revision Id: 0x102201
No Modem Function Group found
Default PCM:

    rates [0x7e0]: 44100 48000 88200 96000 176400 192000
    bits [0xe]: 16 20 24
    formats [0x1]: PCM

Default Amp-In caps: N/A
Default Amp-Out caps: ofs=0x1f, nsteps=0x1f, stepsize=0x05, mute=1
GPIO: io=4, o=0, i=0, unsolicited=1, wake=1
IO[0]: enable=0, dir=0, wake=0, sticky=0, data=0, unsol=0 IO[1]: enable=0, dir=0, wake=0, sticky=0, data=0, unsol=0 IO[2]: enable=0, dir=0, wake=0, sticky=0, data=0, unsol=0 IO[3]: enable=0, dir=0, wake=0, sticky=0, data=0, unsol=0
Node 0x02 [Audio Output] wcaps 0xd0401: Stereo
Device: name="STAC92xx Analog", type="Audio", device=0
Converter: stream=8, channel=0
Power states:
Power: setting=D0, actual=D0
Delay: 13 samples
Node 0x03 [Audio Input] wcaps 0x1d0541: Stereo
Device: name="STAC92xx Analog", type="Audio", device=0
Converter: stream=4, channel=0
SDI-Select: 0
Power states:
Power: setting=D0, actual=D0
Delay: 13 samples
Connection: 1

     0x0a

Processing caps: benign=0, ncoeff=0
Node 0x04 [Audio Input] wcaps 0x140311: Stereo Digital
Converter: stream=0, channel=0
SDI-Select: 0
Digital:
Digital category: 0x0
PCM:

    rates [0x160]: 44100 48000 96000
    bits [0xe]: 16 20 24
    formats [0x5]: PCM AC3
  Delay: 4 samples
  Connection: 1
     0x08

Node 0x05 [Audio Output] wcaps 0x40211: Stereo Digital
Converter: stream=0, channel=0
Digital:
Digital category: 0x0
PCM:

    rates [0x1e0]: 44100 48000 88200 96000
    bits [0xe]: 16 20 24
    formats [0x5]: PCM AC3

Delay: 4 samples
Node 0x06 [Vendor Defined Widget] wcaps 0xf30201: Stereo Digital
Delay: 3 samples
Node 0x07 [Audio Selector] wcaps 0x300901: Stereo R/L
Connection: 3

     0x02* 0x08 0x0a

Node 0x08 [Pin Complex] wcaps 0x430681: Stereo Digital
Pincap 0x00010024: IN EAPD Detect
EAPD 0x0:
Pin Default 0x400001f0: [N/A] Line Out at Ext N/A
Conn = Unknown, Color = Unknown DefAssociation = 0xf, Sequence = 0x0
Misc = NO_PRESENCE
Pin-ctls: 0x00:
Unsolicited: tag=00, enabled=0
Power states:
Power: setting=D0, actual=D0
Delay: 3 samples
Node 0x09 [Pin Complex] wcaps 0x400301: Stereo Digital
Pincap 0x00000010: OUT
Pin Default 0x400001f1: [N/A] Line Out at Ext N/A
Conn = Unknown, Color = Unknown DefAssociation = 0xf, Sequence = 0x1
Misc = NO_PRESENCE
Pin-ctls: 0x00:
Connection: 2

     0x05* 0x0a
Node 0x0a [Audio Selector] wcaps 0x30090d: Stereo Amp-Out R/L
  Control: name="Capture Volume", index=0, device=0
    ControlAmp: chs=3, dir=Out, idx=0, ofs=0
  Control: name="Capture Switch", index=0, device=0
    ControlAmp: chs=3, dir=Out, idx=0, ofs=0
  Amp-Out caps: ofs=0x00, nsteps=0x0f, stepsize=0x05, mute=1
  Amp-Out vals:  [0x07 0x07]
  Connection: 1
     0x0c
Node 0x0b [Audio Selector] wcaps 0x300105: Stereo Amp-Out
  Control: name="PCM Playback Volume", index=0, device=0
    ControlAmp: chs=3, dir=Out, idx=0, ofs=0
  Control: name="PCM Playback Switch", index=0, device=0
    ControlAmp: chs=3, dir=Out, idx=0, ofs=0
  Amp-Out caps: N/A
  Amp-Out vals:  [0x9f 0x9f]
  Connection: 1
     0x07
Node 0x0c [Audio Selector] wcaps 0x30010d: Stereo Amp-Out
  Control: name="Mux Capture Volume", index=0, device=0
    ControlAmp: chs=3, dir=Out, idx=0, ofs=0
  Amp-Out caps: ofs=0x00, nsteps=0x04, stepsize=0x27, mute=0
  Amp-Out vals:  [0x00 0x00]
  Connection: 5
     0x10* 0x0f 0x0e 0x0d 0x12

Node 0x0d [Pin Complex] wcaps 0x400181: Stereo
Control: name="Front Headphone Jack", index=0, device=0 Pincap 0x0000003f: IN OUT HP Detect Trigger ImpSense Pin Default 0x0221401f: [Jack] HP Out at Ext Front
Conn = 1/8, Color = Green
DefAssociation = 0x1, Sequence = 0xf
Pin-ctls: 0xc0: OUT HP
Unsolicited: tag=01, enabled=1
Connection: 1

     0x0b

Node 0x0e [Pin Complex] wcaps 0x400181: Stereo
Control: name="Line Out Jack", index=0, device=0 Pincap 0x0000003f: IN OUT HP Detect Trigger ImpSense Pin Default 0x01014010: [Jack] Line Out at Ext Rear
Conn = 1/8, Color = Green
DefAssociation = 0x1, Sequence = 0x0
Pin-ctls: 0x40: OUT
Unsolicited: tag=02, enabled=1
Connection: 1

     0x0b

Node 0x0f [Pin Complex] wcaps 0x400181: Stereo
Control: name="Line Jack", index=0, device=0 Pincap 0x00000037: IN OUT Detect Trigger ImpSense Pin Default 0x01813020: [Jack] Line In at Ext Rear
Conn = 1/8, Color = Blue
DefAssociation = 0x2, Sequence = 0x0
Pin-ctls: 0x20: IN
Unsolicited: tag=04, enabled=1
Connection: 1

     0x0b

Node 0x10 [Pin Complex] wcaps 0x400181: Stereo
Control: name="Mic Jack", index=0, device=0 Pincap 0x00001737: IN OUT Detect Trigger ImpSense
Vref caps: HIZ 50 GRD 80
Pin Default 0x02a19021: [Jack] Mic at Ext Front
Conn = 1/8, Color = Pink
DefAssociation = 0x2, Sequence = 0x1
Pin-ctls: 0x24: IN VREF_80
Unsolicited: tag=03, enabled=1
Connection: 1

     0x0b
Node 0x11 [Pin Complex] wcaps 0x400104: Mono Amp-Out
  Control: name="LFE Playback Volume", index=0, device=0
    ControlAmp: chs=1, dir=Out, idx=0, ofs=0
  Control: name="LFE Playback Switch", index=0, device=0
    ControlAmp: chs=1, dir=Out, idx=0, ofs=0
  Control: name="Speaker Phantom Jack", index=0, device=0
  Amp-Out caps: N/A
  Amp-Out vals:  [0x00]
  Pincap 0x00000010: OUT
  Pin Default 0x90100140: [Fixed] Speaker at Int N/A
    Conn = Unknown, Color = Unknown
    DefAssociation = 0x4, Sequence = 0x0
    Misc = NO_PRESENCE
  Pin-ctls: 0x00:
  Connection: 1
     0x13

Node 0x12 [Pin Complex] wcaps 0x400001: Stereo
Pincap 0x00000020: IN
Pin Default 0x400001f2: [N/A] Line Out at Ext N/A
Conn = Unknown, Color = Unknown DefAssociation = 0xf, Sequence = 0x2 Misc = NO_PRESENCE
Pin-ctls: 0x20: IN
Node 0x13 [Audio Mixer] wcaps 0x200100: Mono
Connection: 1

     0x07

Node 0x14 [Beep Generator Widget] wcaps 0x70000c: Mono Amp-Out
Amp-Out caps: ofs=0x03, nsteps=0x03, stepsize=0x17, mute=1 Amp-Out vals: [0x00]
--endcollapse--

!!USB Mixer information
!!---------------------------
--startcollapse--

USB Mixer: usb_id=0x046d0a0c, ctrlif=0, ctlerr=0
Card: Logitech Logitech USB Headset at usb-0000:00:1d.7-6.4, full speed
Unit: 1

    Control: name="Speaker Playback Volume", index=0
    Info: id=1, control=2, cmask=0x3, channels=2, type="S16"
    Volume: min=-10496, max=768, dBmin=-4100, dBmax=300
  Unit: 1
    Control: name="Speaker Playback Switch", index=0
    Info: id=1, control=1, cmask=0x0, channels=1, type="INV_BOOLEAN"
    Volume: min=0, max=1, dBmin=0, dBmax=0
  Unit: 2
    Control: name="Mic Capture Volume", index=0
    Info: id=2, control=2, cmask=0x0, channels=1, type="S16"
    Volume: min=4096, max=7424, dBmin=1600, dBmax=2900
  Unit: 2
    Control: name="Mic Capture Switch", index=0
    Info: id=2, control=1, cmask=0x0, channels=1, type="INV_BOOLEAN"
    Volume: min=0, max=1, dBmin=0, dBmax=0
  Unit: 6
    Control: name="Mic Playback Volume", index=0
    Info: id=6, control=2, cmask=0x0, channels=1, type="S16"
    Volume: min=-10496, max=-3072, dBmin=-4100, dBmax=-1200
  Unit: 6
    Control: name="Mic Playback Switch", index=0
    Info: id=6, control=1, cmask=0x0, channels=1, type="INV_BOOLEAN"
    Volume: min=0, max=1, dBmin=0, dBmax=0

--endcollapse--

!!ALSA Device nodes
!!-----------------

crw-rw----+ 1 root audio 116,  8 Aug 24 12:18 /dev/snd/controlC0
crw-rw----+ 1 root audio 116,  4 Aug 24 12:18 /dev/snd/controlC1
crw-rw----+ 1 root audio 116,  7 Aug 24 12:18 /dev/snd/hwC0D0
crw-rw----+ 1 root audio 116,  6 Aug 24 12:19 /dev/snd/pcmC0D0c
crw-rw----+ 1 root audio 116,  5 Aug 24 12:19 /dev/snd/pcmC0D0p
crw-rw----+ 1 root audio 116,  3 Aug 24 12:19 /dev/snd/pcmC1D0c
crw-rw----+ 1 root audio 116,  2 Aug 24 12:30 /dev/snd/pcmC1D0p
crw-rw----+ 1 root audio 116,  1 Aug 24 12:18 /dev/snd/seq

crw-rw----+ 1 root audio 116, 33 Aug 24 12:18 /dev/snd/timer

/dev/snd/by-id:
total 0
drwxr-xr-x. 2 root root 60 Aug 24 12:18 .
drwxr-xr-x. 4 root root 260 Aug 24 12:18 ..
lrwxrwxrwx. 1 root root 12 Aug 24 12:18 usb-Logitech_Logitech_USB_Headset-00 -> ../controlC1

/dev/snd/by-path:
total 0
drwxr-xr-x. 2 root root 80 Aug 24 12:18 .
drwxr-xr-x. 4 root root 260 Aug 24 12:18 ..
lrwxrwxrwx. 1 root root 12 Aug 24 12:18 pci-0000:00:1b.0 -> ../controlC0 lrwxrwxrwx. 1 root root 12 Aug 24 12:18 pci-0000:00:1d.7-usb-0:6.4:1.0 -> ../controlC1

!!ALSA configuration files
!!------------------------

!!System wide config file (/etc/asound.conf)

#
# Place your global alsa-lib configuration here... #

!!Aplay/Arecord output
!!------------

APLAY

**** List of PLAYBACK Hardware Devices ****
card 0: Intel [HDA Intel], device 0: STAC92xx Analog [STAC92xx Analog]
Subdevices: 1/1
Subdevice #0: subdevice #0
card 1: Headset [Logitech USB Headset], device 0: USB Audio [USB Audio]
Subdevices: 0/1
Subdevice #0: subdevice #0

ARECORD

**** List of CAPTURE Hardware Devices ****
card 0: Intel [HDA Intel], device 0: STAC92xx Analog [STAC92xx Analog]
Subdevices: 1/1
Subdevice #0: subdevice #0
card 1: Headset [Logitech USB Headset], device 0: USB Audio [USB Audio]
Subdevices: 1/1
Subdevice #0: subdevice #0

!!Amixer output
!!-------------

!!-------Mixer controls for card 0 [Intel]

Card hw:0 'Intel'/'HDA Intel at 0xfcffc000 irq 66'

  Mixer name    : 'SigmaTel STAC9200'
  Components    : 'HDA:83847690,102801c0,00102201'
  Controls      : 15
  Simple ctrls  : 6

Simple mixer control 'Master',0
Capabilities: pvolume pvolume-joined pswitch pswitch-joined penum Playback channels: Mono
Limits: Playback 0 - 31
Mono: Playback 31 [100%] [0.00dB] [off]
Simple mixer control 'PCM',0
Capabilities: pvolume pswitch penum
Playback channels: Front Left - Front Right
Limits: Playback 0 - 31
Mono:
Front Left: Playback 31 [100%] [0.00dB] [off]
Front Right: Playback 31 [100%] [0.00dB] [off]
Simple mixer control 'LFE',0
Capabilities: pvolume pvolume-joined pswitch pswitch-joined penum Playback channels: Mono
Limits: Playback 0 - 31
Mono: Playback 31 [100%] [0.00dB] [off]
Simple mixer control 'Capture',0
Capabilities: cvolume cswitch penum
Capture channels: Front Left - Front Right
Limits: Capture 0 - 15
Front Left: Capture 7 [47%] [10.50dB] [on]
Front Right: Capture 7 [47%] [10.50dB] [on]
Simple mixer control 'Input Source',0
Capabilities: cenum
Items: 'Mic' 'Line'
Item0: 'Mic'
Simple mixer control 'Mux',0
Capabilities: cvolume penum
Capture channels: Front Left - Front Right
Limits: Capture 0 - 4
Front Left: Capture 0 [0%] [0.00dB]
Front Right: Capture 0 [0%] [0.00dB]

!!-------Mixer controls for card 1 [Headset]

Card hw:1 'Headset'/'Logitech Logitech USB Headset at usb-0000:00:1d.7-6.4, full speed'

  Mixer name    : 'USB Mixer'
  Components    : 'USB046d:0a0c'
  Controls      : 6
  Simple ctrls  : 2

Simple mixer control 'Speaker',0
Capabilities: pvolume pswitch pswitch-joined penum
Playback channels: Front Left - Front Right
Limits: Playback 0 - 44
Mono:
Front Left: Playback 35 [80%] [-6.00dB] [on]
Front Right: Playback 35 [80%] [-6.00dB] [on]
Simple mixer control 'Mic',0
Capabilities: pvolume pvolume-joined cvolume cvolume-joined pswitch pswitch-joined cswitch cswitch-joined penum Playback channels: Mono
Capture channels: Mono
Limits: Playback 0 - 29 Capture 0 - 13
Mono: Playback 0 [0%] [-41.00dB] [off] Capture 13 [100%] [29.00dB] [on]

!!Alsactl output
!!-------------

--startcollapse--

state.Intel {
        control.1 {
                iface MIXER
                name 'PCM Playback Volume'
                value.0 31
                value.1 31
                comment {
                        access 'read write'
                        type INTEGER
                        count 2
                        range '0 - 31'
                        dbmin -9999999
                        dbmax 0
                        dbvalue.0 0
                        dbvalue.1 0
                }
        }
        control.2 {
                iface MIXER
                name 'PCM Playback Switch'
                value.0 false
                value.1 false
                comment {
                        access 'read write'
                        type BOOLEAN
                        count 2
                }
        }
        control.3 {
                iface MIXER
                name 'Capture Volume'
                value.0 7
                value.1 7
                comment {
                        access 'read write'
                        type INTEGER
                        count 2
                        range '0 - 15'
                        dbmin 0
                        dbmax 2250
                        dbvalue.0 1050
                        dbvalue.1 1050
                }
        }
        control.4 {
                iface MIXER
                name 'Capture Switch'
                value.0 true
                value.1 true
                comment {
                        access 'read write'
                        type BOOLEAN
                        count 2
                }
        }
        control.5 {
                iface MIXER
                name 'LFE Playback Volume'
                value 31
                comment {
                        access 'read write'
                        type INTEGER
                        count 1
                        range '0 - 31'
                        dbmin -4650
                        dbmax 0
                        dbvalue.0 0
                }
        }
        control.6 {
                iface MIXER
                name 'LFE Playback Switch'
                value false
                comment {
                        access 'read write'
                        type BOOLEAN
                        count 1
                }
        }
        control.7 {
                iface MIXER
                name 'Mux Capture Volume'
                value.0 0
                value.1 0
                comment {
                        access 'read write'
                        type INTEGER
                        count 2
                        range '0 - 4'
                        dbmin 0
                        dbmax 4000
                        dbvalue.0 0
                        dbvalue.1 0
                }
        }
        control.8 {
                iface MIXER
                name 'Input Source'
                value Mic
                comment {
                        access 'read write'
                        type ENUMERATED
                        count 1
                        item.0 Mic
                        item.1 Line
                }
        }
        control.9 {
                iface MIXER
                name 'Master Playback Volume'
                value 31
                comment {
                        access 'read write'
                        type INTEGER
                        count 1
                        range '0 - 31'
                        dbmin -9999999
                        dbmax 0
                        dbvalue.0 0
                }
        }
        control.10 {
                iface MIXER
                name 'Master Playback Switch'
                value false
                comment {
                        access 'read write'
                        type BOOLEAN
                        count 1
                }
        }
        control.11 {
                iface CARD
                name 'Line Out Jack'
                value true
                comment {
                        access read
                        type BOOLEAN
                        count 1
                }
        }
        control.12 {
                iface CARD
                name 'Front Headphone Jack'
                value false
                comment {
                        access read
                        type BOOLEAN
                        count 1
                }
        }
        control.13 {
                iface CARD
                name 'Speaker Phantom Jack'
                value true
                comment {
                        access read
                        type BOOLEAN
                        count 1
                }
        }
        control.14 {
                iface CARD
                name 'Mic Jack'
                value false
                comment {
                        access read
                        type BOOLEAN
                        count 1
                }
        }
        control.15 {
                iface CARD
                name 'Line Jack'
                value false
                comment {
                        access read
                        type BOOLEAN
                        count 1
                }
        }
}
state.Headset {
        control.1 {
                iface MIXER
                name 'Mic Playback Switch'
                value false
                comment {
                        access 'read write'
                        type BOOLEAN
                        count 1
                }
        }
        control.2 {
                iface MIXER
                name 'Mic Playback Volume'
                value 0
                comment {
                        access 'read write'
                        type INTEGER
                        count 1
                        range '0 - 29'
                        dbmin -4100
                        dbmax -1200
                        dbvalue.0 -4100
                }
        }
        control.3 {
                iface MIXER
                name 'Speaker Playback Switch'
                value true
                comment {
                        access 'read write'
                        type BOOLEAN
                        count 1
                }
        }
        control.4 {
                iface MIXER
                name 'Speaker Playback Volume'
                value.0 35
                value.1 35
                comment {
                        access 'read write'
                        type INTEGER
                        count 2
                        range '0 - 44'
                        dbmin -4100
                        dbmax 300
                        dbvalue.0 -600
                        dbvalue.1 -600
                }
        }
        control.5 {
                iface MIXER
                name 'Mic Capture Switch'
                value true
                comment {
                        access 'read write'
                        type BOOLEAN
                        count 1
                }
        }
        control.6 {
                iface MIXER
                name 'Mic Capture Volume'
                value 13
                comment {
                        access 'read write'
                        type INTEGER
                        count 1
                        range '0 - 13'
                        dbmin 1600
                        dbmax 2900
                        dbvalue.0 2900
                }
        }
}

--endcollapse--

!!All Loaded Modules
!!------------------

Module
fuse
nf_conntrack_ipv4
nf_defrag_ipv4
xt_state
nf_conntrack
snd_hda_codec_idt
snd_usb_audio
snd_hda_intel
snd_usbmidi_lib
snd_hda_codec
snd_rawmidi
snd_hwdep
iTCO_wdt
iTCO_vendor_support
lpc_ich
i5000_edac
snd_seq
snd_seq_device
edac_core
snd_pcm
mfd_core
serio_raw
ppdev
snd_page_alloc
snd_timer
coretemp
snd
i5k_amb
soundcore
parport_pc
dcdbas
microcode
shpchp
parport
i2c_i801
vhost_net
tun
macvtap
macvlan
kvm_intel
kvm
uinput
binfmt_misc
xts
gf128mul
dm_crypt
raid1
radeon
i2c_algo_bit
drm_kms_helper
ttm
drm
tg3
mptsas
mptscsih
mptbase
i2c_core
scsi_transport_sas
sunrpc
be2iscsi
bnx2i
cnic
uio
cxgb4i
cxgb4
cxgb3
mdio
libcxgbi
libiscsi_tcp
qla4xxx
iscsi_boot_sysfs
libiscsi
scsi_transport_iscsi

!!Sysfs Files
!!-----------

/sys/class/sound/hwC0D0/init_pin_configs:
0x08 0x40f000f0
0x09 0x40f000f1
0x0d 0x0221101f
0x0e 0x01014010
0x0f 0x0181302e
0x10 0x02a11020
0x11 0x40f000f2
0x12 0x40f000f3

/sys/class/sound/hwC0D0/driver_pin_configs:
0x08 0x400001f0
0x09 0x400001f1
0x0d 0x0221401f
0x0e 0x01014010
0x0f 0x01813020
0x10 0x02a19021
0x11 0x90100140
0x12 0x400001f2

/sys/class/sound/hwC0D0/user_pin_configs:

/sys/class/sound/hwC0D0/init_verbs:

!!ALSA/HDA dmesg
!!------------------

[   38.190306] SELinux: initialized (dev configfs, type configfs), uses genfs_contexts
[   38.229616] snd_hda_intel 0000:00:1b.0: irq 66 for MSI/MSI-X
[   38.270699] ALSA sound/usb/mixer.c:866 6:0: cannot get min/max values for control 2 (id 6)
[   38.274097] ALSA sound/usb/mixer.c:866 1:0: cannot get min/max values for control 2 (id 1)
[   38.276753] ALSA sound/usb/mixer.c:866 2:0: cannot get min/max values for control 2 (id 2)
[   38.279322] ALSA sound/pci/hda/hda_auto_parser.c:322 autoconfig: line_outs=1 (0xe/0x0/0x0/0x0/0x0) type:line
[   38.279326] ALSA sound/pci/hda/hda_auto_parser.c:326    speaker_outs=1 (0x11/0x0/0x0/0x0/0x0)
[   38.279329] ALSA sound/pci/hda/hda_auto_parser.c:330    hp_outs=1 (0xd/0x0/0x0/0x0/0x0)
[   38.279331] ALSA sound/pci/hda/hda_auto_parser.c:331    mono: mono_out=0x0
[   38.279333] ALSA sound/pci/hda/hda_auto_parser.c:335    inputs:
[   38.279336] ALSA sound/pci/hda/hda_auto_parser.c:339      Mic=0x10
[   38.279342] ALSA sound/pci/hda/hda_auto_parser.c:339      Line=0xf
[   38.285983] usbcore: registered new interface driver snd-usb-audio
[   38.290348] input: HDA Intel Line as /devices/pci0000:00/0000:00:1b.0/sound/card0/input8
[   38.294975] input: HDA Intel Mic as /devices/pci0000:00/0000:00:1b.0/sound/card0/input9
[   38.297406] input: HDA Intel Front Headphone as /devices/pci0000:00/0000:00:1b.0/sound/card0/input10
[   38.301369] input: HDA Intel Line Out as /devices/pci0000:00/0000:00:1b.0/sound/card0/input11
[   39.377352] EXT4-fs: Warning: mounting with data=journal disables delayed allocation and O_DIRECT support!

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Quoted 13 lines · Josh Boyer

Hi All,

We've had a report[1] that the Logitech USB headset 0003:046D:0A0C isn't working with 3.6-rc3. It seems the last working kernel was based on commit 10c63c9, and it first stopped working with a kernel based on commit 23dcfa6. There are only a few ALSA commits between those revisions, so hopefully this is something that is fairly easy to identify. The only commit to USB audio in that set is:

commit e9ba389c5ffc4dd29dfe17e00e48877302111135

Author: Takashi Iwai <[email protected]>
Date:   Wed Aug 15 12:32:00 2012 +0200

Quoted 4 lines

ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream

I've CC'd the reporter and attached the alsa-info is below.

Does it work again once you revert that commit?

We've had a report[1] that the Logitech USB headset 0003:046D:0A0C isn't working with 3.6-rc3. It seems the last working kernel was based on commit 10c63c9, and it first stopped working with a kernel based on commit 23dcfa6. There are only a few ALSA commits between those revisions, so hopefully this is something that is fairly easy to identify. The only commit to USB audio in that set is:

!!ALSA/HDA dmesg
!!------------------

[   38.190306] SELinux: initialized (dev configfs, type configfs), uses genfs_contexts
[   38.229616] snd_hda_intel 0000:00:1b.0: irq 66 for MSI/MSI-X
[   38.270699] ALSA sound/usb/mixer.c:866 6:0: cannot get min/max values for control 2 (id 6)
[   38.274097] ALSA sound/usb/mixer.c:866 1:0: cannot get min/max values for control 2 (id 1)
[   38.276753] ALSA sound/usb/mixer.c:866 2:0: cannot get min/max values for control 2 (id 2)
[   38.279322] ALSA sound/pci/hda/hda_auto_parser.c:322 autoconfig: line_outs=1 (0xe/0x0/0x0/0x0/0x0) type:line
[   38.279326] ALSA sound/pci/hda/hda_auto_parser.c:326    speaker_outs=1 (0x11/0x0/0x0/0x0/0x0)
[   38.279329] ALSA sound/pci/hda/hda_auto_parser.c:330    hp_outs=1 (0xd/0x0/0x0/0x0/0x0)
[   38.279331] ALSA sound/pci/hda/hda_auto_parser.c:331    mono: mono_out=0x0
[   38.279333] ALSA sound/pci/hda/hda_auto_parser.c:335    inputs:
[   38.279336] ALSA sound/pci/hda/hda_auto_parser.c:339      Mic=0x10
[   38.279342] ALSA sound/pci/hda/hda_auto_parser.c:339      Line=0xf
[   38.285983] usbcore: registered new interface driver snd-usb-audio

So what does "isn't working anymore" refer to precisely? Are there any more dmesg entries generated once the stream is started?

I see this behavior on three different computers. Rebooting into an old
enough kernel gets things working again. The older kernels also have older initramfs images. I could rebuild those to see if there is some issue there instead of the kernels themselves?

Quoted 5 lines · Bruno Wolff III

I see this behavior on three different computers. Rebooting into an old enough kernel gets things working again. The older kernels also have older initramfs images. I could rebuild those to see if there is some issue there instead of the kernels themselves?

I tested one of the older kernels (where the headset still worked) with a rebuilt initramfs and the headset still worked. So it likely isn't an issue with something in the initramfs images.

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 8 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

I started a scratch build of a kernel with that patch reverted (via patch -R). You'll find it here when it finishes:

http://koji.fedoraproject.org/koji/taskinfo?taskID=4421817

josh

Quoted 13 lines · Josh Boyer

Quoted 8 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

I started a scratch build of a kernel with that patch reverted (via patch -R). You'll find it here when it finishes:

Quoted 3 lines

http://koji.fedoraproject.org/koji/taskinfo?taskID=4421817

josh

I'll test it when it finishes. Thanks. It takes me a long time to build kernels locally, and I wasn't sure how to do a scratch build on koji without committing stuff remotely.

Actually you don't have to rebuild the whole kernel at all just for debugging this usb-audio bug. The workflow below is what I do often for debugging a driver problem on a distro kernel.[...]

[...]

module directory and run depmod again. Pretty convenient, eh?

Hi Takashi,

Am 26.08.2012 14:46 schrieb "Takashi Iwai" <[email protected]>:

Quoted 4 lines

Actually you don't have to rebuild the whole kernel at all just for debugging this usb-audio bug. The workflow below is what I do often for debugging a driver problem on a distro kernel.[...]

Quoted 1 line

[...]

If you want to take back to the original module, just remove updates

Quoted 3 lines

module directory and run depmod again. Pretty convenient, eh?

Could you please put this somewhere on alsa-project.org?

Quoted 6 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

Quoted 2 lines

I can try that, but it takes a long time to build a new kernel on my old hardware.

I started a scratch build of a kernel with that patch reverted (via patch -R). You'll find it here when it finishes:

http://koji.fedoraproject.org/koji/taskinfo?taskID=4421817

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

Quoted 12 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 3 lines

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Btw - thanks a lot for testing -rc kernels, much appreciated!

Quoted 17 lines · Daniel Mack

Quoted 15 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

Quoted 3 lines

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Yep, agreed. If this revert kernel doesn't work, we're likely down to a git bisect, Bruno.

Quoted 13 lines · Daniel Mack

Quoted 12 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 7 lines

Quoted 3 lines

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Yep, agreed. If this revert kernel doesn't work, we're likely down to a git bisect, Bruno.

Quoted 1 line

Btw - thanks a lot for testing -rc kernels, much appreciated!

Indeed!

From 6617bb2463ae3fad21390b59cc2a71f96b9e9ca8 Mon Sep 17 00:00:00 2001

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 22 lines · Josh Boyer

Quoted 13 lines · Daniel Mack

Quoted 12 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 7 lines

Quoted 3 lines

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Quoted 6 lines

Yep, agreed. If this revert kernel doesn't work, we're likely down to a git bisect, Bruno.

Quoted 1 line

Btw - thanks a lot for testing -rc kernels, much appreciated!

Indeed!

Could you please try this patch on top of Takashi's? Thanks again!

Daniel

[2 0001-ALSA-snd-usb-Fix-URB-cancellation-at-stream-start.patch <text/x-patch (7bit)>]

Quoted 1 line

From 6617bb2463ae3fad21390b59cc2a71f96b9e9ca8 Mon Sep 17 00:00:00 2001

From: Daniel Mack <[email protected]>
Date: Wed, 29 Aug 2012 13:17:05 +0200
Subject: [PATCH] ALSA: snd-usb: Fix URB cancellation at stream start

Commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") fixed a scheduling-while-atomic bug that happened when snd_usb_endpoint_start was called from the trigger callback, which is an atmic context. However, the patch breaks the idea of the endpoints reference counting, which is the reason why the driver has been refactored lately.

Revert that commit and let snd_usb_endpoint_start() take care of the URB cancellation again. As this function is called from both atomic and non-atomic context, add a flag to denote whether the function may sleep.

Signed-off-by Daniel Mack <[email protected]>

Cc [email protected] [3.5+]

---
 sound/usb/endpoint.c | 10 ++++++++--
 sound/usb/endpoint.h |  2 +-
 sound/usb/pcm.c      | 13 +++++--------
 3 files changed, 14 insertions(+), 11 deletions(-)

sound/usb/endpoint.c +4 -2

index c411812..678456c 100644
@@ -799,7 +799,9 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
 /**
  * snd_usb_endpoint_start: start an snd_usb_endpoint
  *
- * @ep: the endpoint to start
+ * @ep:		the endpoint to start
+ * @can_sleep:	flag indicating whether the operation is executed in
+ * 		non-atomic context
  *
  * A call to this function will increment the use count of the endpoint.
  * In case it is not already running, the URBs for this endpoint will be

@@ -809,7 +811,7 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
  *
  * Returns an error if the URB submission failed, 0 in all other cases.
  */
-int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
+int snd_usb_endpoint_start(struct snd_usb_endpoint *ep, int can_sleep)
 {
 	int err;
 	unsigned int i;

At Wed, 29 Aug 2012 13:26:25 +0200,

Quoted 90 lines · Daniel Mack

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 22 lines · Josh Boyer

Quoted 13 lines · Daniel Mack

Quoted 12 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 7 lines

Quoted 3 lines

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Quoted 6 lines

Yep, agreed. If this revert kernel doesn't work, we're likely down to a git bisect, Bruno.

Quoted 1 line

Btw - thanks a lot for testing -rc kernels, much appreciated!

Indeed!

Could you please try this patch on top of Takashi's? Thanks again!

Daniel

[2 0001-ALSA-snd-usb-Fix-URB-cancellation-at-stream-start.patch <text/x-patch (7bit)>]

Quoted 1 line

From 6617bb2463ae3fad21390b59cc2a71f96b9e9ca8 Mon Sep 17 00:00:00 2001

From: Daniel Mack <[email protected]>
Date: Wed, 29 Aug 2012 13:17:05 +0200
Subject: [PATCH] ALSA: snd-usb: Fix URB cancellation at stream start

Commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") fixed a scheduling-while-atomic bug that happened when snd_usb_endpoint_start was called from the trigger callback, which is an atmic context. However, the patch breaks the idea of the endpoints reference counting, which is the reason why the driver has been refactored lately.

Revert that commit and let snd_usb_endpoint_start() take care of the URB cancellation again. As this function is called from both atomic and non-atomic context, add a flag to denote whether the function may sleep.

Signed-off-by Daniel Mack <[email protected]>

Cc [email protected] [3.5+]

---
 sound/usb/endpoint.c | 10 ++++++++--
 sound/usb/endpoint.h |  2 +-
 sound/usb/pcm.c      | 13 +++++--------
 3 files changed, 14 insertions(+), 11 deletions(-)

sound/usb/endpoint.c +4 -2

index c411812..678456c 100644
@@ -799,7 +799,9 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
 /**
  * snd_usb_endpoint_start: start an snd_usb_endpoint
  *
- * @ep: the endpoint to start
+ * @ep:		the endpoint to start
+ * @can_sleep:	flag indicating whether the operation is executed in
+ * 		non-atomic context
  *
  * A call to this function will increment the use count of the endpoint.
  * In case it is not already running, the URBs for this endpoint will be

@@ -809,7 +811,7 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
  *
  * Returns an error if the URB submission failed, 0 in all other cases.
  */
-int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
+int snd_usb_endpoint_start(struct snd_usb_endpoint *ep, int can_sleep)
 {
 	int err;
 	unsigned int i;

@@ -821,6 +823,10 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
 	if (++ep->use_count != 1)
 		return 0;
 

sound/usb/endpoint.c int snd_usb_endpoint_start(struct snd_usb_endpoint *ep) · line 828

Quoted 3 lines

sound/usb/endpoint.c +3 -0

+	/* just to be sure */
+	deactivate_urbs(ep, 0, can_sleep);
+	wait_clear_urbs(ep);

It'd be safer to protect the call of wait_clear_urbs() when can_sleep=0.

From 42ff3e2d34427dc4c6faa09052a2531f2e71d7d6 Mon Sep 17 00:00:00 2001

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 92 lines · Takashi Iwai

At Wed, 29 Aug 2012 13:26:25 +0200,

Quoted 90 lines · Daniel Mack

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 22 lines · Josh Boyer

Quoted 13 lines · Daniel Mack

Quoted 12 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 7 lines

Quoted 3 lines

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Quoted 6 lines

Yep, agreed. If this revert kernel doesn't work, we're likely down to a git bisect, Bruno.

Quoted 1 line

Btw - thanks a lot for testing -rc kernels, much appreciated!

Indeed!

Could you please try this patch on top of Takashi's? Thanks again!

Daniel

[2 0001-ALSA-snd-usb-Fix-URB-cancellation-at-stream-start.patch <text/x-patch (7bit)>]

Quoted 1 line

From 6617bb2463ae3fad21390b59cc2a71f96b9e9ca8 Mon Sep 17 00:00:00 2001

From: Daniel Mack <[email protected]>
Date: Wed, 29 Aug 2012 13:17:05 +0200
Subject: [PATCH] ALSA: snd-usb: Fix URB cancellation at stream start

Commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") fixed a scheduling-while-atomic bug that happened when snd_usb_endpoint_start was called from the trigger callback, which is an atmic context. However, the patch breaks the idea of the endpoints reference counting, which is the reason why the driver has been refactored lately.

Revert that commit and let snd_usb_endpoint_start() take care of the URB cancellation again. As this function is called from both atomic and non-atomic context, add a flag to denote whether the function may sleep.

Signed-off-by Daniel Mack <[email protected]>

Cc [email protected] [3.5+]

---
 sound/usb/endpoint.c | 10 ++++++++--
 sound/usb/endpoint.h |  2 +-
 sound/usb/pcm.c      | 13 +++++--------
 3 files changed, 14 insertions(+), 11 deletions(-)

sound/usb/endpoint.c +4 -2

index c411812..678456c 100644
@@ -799,7 +799,9 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
 /**
  * snd_usb_endpoint_start: start an snd_usb_endpoint
  *
- * @ep: the endpoint to start
+ * @ep:		the endpoint to start
+ * @can_sleep:	flag indicating whether the operation is executed in
+ * 		non-atomic context
  *
  * A call to this function will increment the use count of the endpoint.
  * In case it is not already running, the URBs for this endpoint will be

@@ -809,7 +811,7 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
  *
  * Returns an error if the URB submission failed, 0 in all other cases.
  */
-int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
+int snd_usb_endpoint_start(struct snd_usb_endpoint *ep, int can_sleep)
 {
 	int err;
 	unsigned int i;

@@ -821,6 +823,10 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
 	if (++ep->use_count != 1)
 		return 0;
 

Quoted 6 lines

sound/usb/endpoint.c int snd_usb_endpoint_start(struct snd_usb_endpoint *ep) · line 829

Quoted 3 lines

sound/usb/endpoint.c +3 -0

+	/* just to be sure */
+	deactivate_urbs(ep, 0, can_sleep);
+	wait_clear_urbs(ep);

It'd be safer to protect the call of wait_clear_urbs() when can_sleep=0.

Right. New patch attached.

[2 0001-ALSA-snd-usb-Fix-URB-cancellation-at-stream-start.patch <text/x-patch (7bit)>]

Quoted 1 line

From 42ff3e2d34427dc4c6faa09052a2531f2e71d7d6 Mon Sep 17 00:00:00 2001

From: Daniel Mack <[email protected]>
Date: Wed, 29 Aug 2012 13:17:05 +0200
Subject: [PATCH] ALSA: snd-usb: Fix URB cancellation at stream start

Commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") fixed a scheduling-while-atomic bug that happened when snd_usb_endpoint_start was called from the trigger callback, which is an atmic context. However, the patch breaks the idea of the endpoints reference counting, which is the reason why the driver has been refactored lately.

Revert that commit and let snd_usb_endpoint_start() take care of the URB cancellation again. As this function is called from both atomic and non-atomic context, add a flag to denote whether the function may sleep.

Signed-off-by Daniel Mack <[email protected]>

Cc [email protected] [3.5+]

---
 sound/usb/endpoint.c | 11 +++++++++--
 sound/usb/endpoint.h |  2 +-
 sound/usb/pcm.c      | 13 +++++--------
 3 files changed, 15 insertions(+), 11 deletions(-)

sound/usb/endpoint.c +9 -2

index c411812..b896c55 100644
@@ -799,7 +799,9 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
 /**
  * snd_usb_endpoint_start: start an snd_usb_endpoint
  *
- * @ep: the endpoint to start
+ * @ep:		the endpoint to start
+ * @can_sleep:	flag indicating whether the operation is executed in
+ * 		non-atomic context
  *
  * A call to this function will increment the use count of the endpoint.
  * In case it is not already running, the URBs for this endpoint will be

@@ -809,7 +811,7 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
  *
  * Returns an error if the URB submission failed, 0 in all other cases.
  */
-int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
+int snd_usb_endpoint_start(struct snd_usb_endpoint *ep, int can_sleep)
 {
 	int err;
 	unsigned int i;

@@ -821,6 +823,11 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
 	if (++ep->use_count != 1)
 		return 0;
 
+	/* just to be sure */
+	deactivate_urbs(ep, 0, can_sleep);
+	if (can_sleep)
+		wait_clear_urbs(ep);
+
 	ep->active_mask = 0;
 	ep->unlink_mask = 0;
 	ep->phase = 0;

sound/usb/endpoint.h +1 -1

index ee2723f..a8e60c1 100644
@@ -13,7 +13,7 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
 				struct audioformat *fmt,
 				struct snd_usb_endpoint *sync_ep);
 
-int  snd_usb_endpoint_start(struct snd_usb_endpoint *ep);
+int  snd_usb_endpoint_start(struct snd_usb_endpoint *ep, int can_sleep);
 void snd_usb_endpoint_stop(struct snd_usb_endpoint *ep,
 			   int force, int can_sleep, int wait);
 int  snd_usb_endpoint_activate(struct snd_usb_endpoint *ep);

sound/usb/pcm.c +5 -8

index 62ec808..1546577 100644
@@ -212,7 +212,7 @@ int snd_usb_init_pitch(struct snd_usb_audio *chip, int iface,
 	}
 }
 
-static int start_endpoints(struct snd_usb_substream *subs)
+static int start_endpoints(struct snd_usb_substream *subs, int can_sleep)
 {
 	int err;
 

@@ -225,7 +225,7 @@ static int start_endpoints(struct snd_usb_substream *subs)
 		snd_printdd(KERN_DEBUG "Starting data EP @%p\n", ep);
 
 		ep->data_subs = subs;
-		err = snd_usb_endpoint_start(ep);
+		err = snd_usb_endpoint_start(ep, can_sleep);
 		if (err < 0) {
 			clear_bit(SUBSTREAM_FLAG_DATA_EP_STARTED, &subs->flags);
 			return err;

@@ -239,7 +239,7 @@ static int start_endpoints(struct snd_usb_substream *subs)
 		snd_printdd(KERN_DEBUG "Starting sync EP @%p\n", ep);
 
 		ep->sync_slave = subs->data_endpoint;
-		err = snd_usb_endpoint_start(ep);
+		err = snd_usb_endpoint_start(ep, can_sleep);
 		if (err < 0) {
 			clear_bit(SUBSTREAM_FLAG_SYNC_EP_STARTED, &subs->flags);
 			return err;

@@ -544,13 +544,10 @@ static int snd_usb_pcm_prepare(struct snd_pcm_substream *substream)
 	subs->last_frame_number = 0;
 	runtime->delay = 0;
 
-	/* clear the pending deactivation on the target EPs */
-	deactivate_endpoints(subs);
-
 	/* for playback, submit the URBs now; otherwise, the first hwptr_done
 	 * updates for all URBs would happen at the same time when starting */
 	if (subs->direction == SNDRV_PCM_STREAM_PLAYBACK)
-		return start_endpoints(subs);
+		return start_endpoints(subs, 1);
 
 	return 0;
 }

@@ -1175,7 +1172,7 @@ static int snd_usb_substream_capture_trigger(struct snd_pcm_substream *substream
 
 	switch (cmd) {
 	case SNDRV_PCM_TRIGGER_START:
-		err = start_endpoints(subs);
+		err = start_endpoints(subs, 0);
 		if (err < 0)
 			return err;
 

Signature · 1 line

1.7.11.4

At Wed, 29 Aug 2012 15:32:34 +0200,

Quoted 102 lines · Daniel Mack

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 98 lines · Takashi Iwai

At Wed, 29 Aug 2012 13:26:25 +0200,

Quoted 87 lines · Daniel Mack

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 22 lines · Josh Boyer

Quoted 13 lines · Daniel Mack

Quoted 12 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 7 lines

Quoted 3 lines

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Quoted 6 lines

Yep, agreed. If this revert kernel doesn't work, we're likely down to a git bisect, Bruno.

Quoted 1 line

Btw - thanks a lot for testing -rc kernels, much appreciated!

Indeed!

Could you please try this patch on top of Takashi's? Thanks again!

Daniel

[2 0001-ALSA-snd-usb-Fix-URB-cancellation-at-stream-start.patch <text/x-patch (7bit)>]

Quoted 1 line

From 6617bb2463ae3fad21390b59cc2a71f96b9e9ca8 Mon Sep 17 00:00:00 2001

From: Daniel Mack <[email protected]>
Date: Wed, 29 Aug 2012 13:17:05 +0200
Subject: [PATCH] ALSA: snd-usb: Fix URB cancellation at stream start

Commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") fixed a scheduling-while-atomic bug that happened when snd_usb_endpoint_start was called from the trigger callback, which is an atmic context. However, the patch breaks the idea of the endpoints reference counting, which is the reason why the driver has been refactored lately.

Revert that commit and let snd_usb_endpoint_start() take care of the URB cancellation again. As this function is called from both atomic and non-atomic context, add a flag to denote whether the function may sleep.

Signed-off-by Daniel Mack <[email protected]>

Cc [email protected] [3.5+]

---
 sound/usb/endpoint.c | 10 ++++++++--
 sound/usb/endpoint.h |  2 +-
 sound/usb/pcm.c      | 13 +++++--------
 3 files changed, 14 insertions(+), 11 deletions(-)

sound/usb/endpoint.c +4 -2

index c411812..678456c 100644
@@ -799,7 +799,9 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
 /**
  * snd_usb_endpoint_start: start an snd_usb_endpoint
  *
- * @ep: the endpoint to start
+ * @ep:		the endpoint to start
+ * @can_sleep:	flag indicating whether the operation is executed in
+ * 		non-atomic context
  *
  * A call to this function will increment the use count of the endpoint.
  * In case it is not already running, the URBs for this endpoint will be

@@ -809,7 +811,7 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
  *
  * Returns an error if the URB submission failed, 0 in all other cases.
  */
-int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
+int snd_usb_endpoint_start(struct snd_usb_endpoint *ep, int can_sleep)
 {
 	int err;
 	unsigned int i;

sound/usb/endpoint.c int snd_usb_endpoint_start(struct snd_usb_endpoint *ep) · line 828

Quoted 6 lines

sound/usb/endpoint.c +3 -0

@@ -821,6 +823,10 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
 	if (++ep->use_count != 1)
 		return 0;
 
+	/* just to be sure */
+	deactivate_urbs(ep, 0, can_sleep);
+	wait_clear_urbs(ep);

It'd be safer to protect the call of wait_clear_urbs() when can_sleep=0.

Quoted 1 line

Right. New patch attached.

Thanks. This makes me thinking whether we really need to call deactivate_urbs() at snd_usb_endpoint_start(). deactivate_endpoints() is called already in prepare (at the beginning). Which possibility is considered? The comment "just to be sure" implies that my original code before your stream model change was simply optional. Now I'm not quite sure whether we can drop it or not...

Quoted 104 lines · Takashi Iwai

At Wed, 29 Aug 2012 15:32:34 +0200,

Quoted 102 lines · Daniel Mack

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 98 lines · Takashi Iwai

At Wed, 29 Aug 2012 13:26:25 +0200,

Quoted 87 lines · Daniel Mack

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 22 lines · Josh Boyer

Quoted 13 lines · Daniel Mack

Quoted 12 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 7 lines

Quoted 3 lines

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Quoted 6 lines

Yep, agreed. If this revert kernel doesn't work, we're likely down to a git bisect, Bruno.

Quoted 1 line

Btw - thanks a lot for testing -rc kernels, much appreciated!

Indeed!

Could you please try this patch on top of Takashi's? Thanks again!

Daniel

[2 0001-ALSA-snd-usb-Fix-URB-cancellation-at-stream-start.patch <text/x-patch (7bit)>]

Quoted 1 line

From 6617bb2463ae3fad21390b59cc2a71f96b9e9ca8 Mon Sep 17 00:00:00 2001

From: Daniel Mack <[email protected]>
Date: Wed, 29 Aug 2012 13:17:05 +0200
Subject: [PATCH] ALSA: snd-usb: Fix URB cancellation at stream start

Commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") fixed a scheduling-while-atomic bug that happened when snd_usb_endpoint_start was called from the trigger callback, which is an atmic context. However, the patch breaks the idea of the endpoints reference counting, which is the reason why the driver has been refactored lately.

Revert that commit and let snd_usb_endpoint_start() take care of the URB cancellation again. As this function is called from both atomic and non-atomic context, add a flag to denote whether the function may sleep.

Signed-off-by Daniel Mack <[email protected]>

Cc [email protected] [3.5+]

---
 sound/usb/endpoint.c | 10 ++++++++--
 sound/usb/endpoint.h |  2 +-
 sound/usb/pcm.c      | 13 +++++--------
 3 files changed, 14 insertions(+), 11 deletions(-)

sound/usb/endpoint.c +4 -2

index c411812..678456c 100644
@@ -799,7 +799,9 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
 /**
  * snd_usb_endpoint_start: start an snd_usb_endpoint
  *
- * @ep: the endpoint to start
+ * @ep:		the endpoint to start
+ * @can_sleep:	flag indicating whether the operation is executed in
+ * 		non-atomic context
  *
  * A call to this function will increment the use count of the endpoint.
  * In case it is not already running, the URBs for this endpoint will be

@@ -809,7 +811,7 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
  *
  * Returns an error if the URB submission failed, 0 in all other cases.
  */
-int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
+int snd_usb_endpoint_start(struct snd_usb_endpoint *ep, int can_sleep)
 {
 	int err;
 	unsigned int i;

sound/usb/endpoint.c int snd_usb_endpoint_start(struct snd_usb_endpoint *ep) · line 828

Quoted 6 lines

sound/usb/endpoint.c +3 -0

@@ -821,6 +823,10 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
 	if (++ep->use_count != 1)
 		return 0;
 
+	/* just to be sure */
+	deactivate_urbs(ep, 0, can_sleep);
+	wait_clear_urbs(ep);

It'd be safer to protect the call of wait_clear_urbs() when can_sleep=0.

Quoted 8 lines

Quoted 1 line

Right. New patch attached.

Thanks. This makes me thinking whether we really need to call deactivate_urbs() at snd_usb_endpoint_start(). deactivate_endpoints() is called already in prepare (at the beginning). Which possibility is considered? The comment "just to be sure" implies that my original code before your stream model change was simply optional. Now I'm not quite sure whether we can drop it or not...

Yes, we can most probably drop it, but I would clearly do that in another patch for 3.6 - I'll prepare one.

I also found some regressions caused by the recent refactoring, and will send out a patch collection, hopefully later today.

Quoted 92 lines · Takashi Iwai

At Wed, 29 Aug 2012 13:26:25 +0200,

Quoted 90 lines · Daniel Mack

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 22 lines · Josh Boyer

Quoted 13 lines · Daniel Mack

Quoted 12 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 7 lines

Quoted 3 lines

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Quoted 6 lines

Yep, agreed. If this revert kernel doesn't work, we're likely down to a git bisect, Bruno.

Quoted 1 line

Btw - thanks a lot for testing -rc kernels, much appreciated!

Indeed!

Could you please try this patch on top of Takashi's? Thanks again!

Daniel

[2 0001-ALSA-snd-usb-Fix-URB-cancellation-at-stream-start.patch <text/x-patch (7bit)>]

Quoted 1 line

From 6617bb2463ae3fad21390b59cc2a71f96b9e9ca8 Mon Sep 17 00:00:00 2001

From: Daniel Mack <[email protected]>
Date: Wed, 29 Aug 2012 13:17:05 +0200
Subject: [PATCH] ALSA: snd-usb: Fix URB cancellation at stream start

Commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") fixed a scheduling-while-atomic bug that happened when snd_usb_endpoint_start was called from the trigger callback, which is an atmic context. However, the patch breaks the idea of the endpoints reference counting, which is the reason why the driver has been refactored lately.

Revert that commit and let snd_usb_endpoint_start() take care of the URB cancellation again. As this function is called from both atomic and non-atomic context, add a flag to denote whether the function may sleep.

Signed-off-by Daniel Mack <[email protected]>

Cc [email protected] [3.5+]

---
 sound/usb/endpoint.c | 10 ++++++++--
 sound/usb/endpoint.h |  2 +-
 sound/usb/pcm.c      | 13 +++++--------
 3 files changed, 14 insertions(+), 11 deletions(-)

sound/usb/endpoint.c +4 -2

index c411812..678456c 100644
@@ -799,7 +799,9 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
 /**
  * snd_usb_endpoint_start: start an snd_usb_endpoint
  *
- * @ep: the endpoint to start
+ * @ep:		the endpoint to start
+ * @can_sleep:	flag indicating whether the operation is executed in
+ * 		non-atomic context
  *
  * A call to this function will increment the use count of the endpoint.
  * In case it is not already running, the URBs for this endpoint will be

@@ -809,7 +811,7 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
  *
  * Returns an error if the URB submission failed, 0 in all other cases.
  */
-int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
+int snd_usb_endpoint_start(struct snd_usb_endpoint *ep, int can_sleep)
 {
 	int err;
 	unsigned int i;

@@ -821,6 +823,10 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
 	if (++ep->use_count != 1)
 		return 0;
 

Quoted 6 lines

sound/usb/endpoint.c int snd_usb_endpoint_start(struct snd_usb_endpoint *ep) · line 829

Quoted 3 lines

sound/usb/endpoint.c +3 -0

+	/* just to be sure */
+	deactivate_urbs(ep, 0, can_sleep);
+	wait_clear_urbs(ep);

It'd be safer to protect the call of wait_clear_urbs() when can_sleep=0.

Right. New patch attached.

Quoted 1 line

Right. New patch attached.

I'll try and get another kernel built with this later today. I'm at Plumbers so it might have to wait a bit.

Quoted 4 lines · Josh Boyer

Quoted 1 line

Right. New patch attached.

I'll try and get another kernel built with this later today. I'm at Plumbers so it might have to wait a bit.

OK, I got impatient with myself and got the kernel building sooner than I said. Bruno, you can find the kernel here when the build finishes:

http://koji.fedoraproject.org/koji/taskinfo?taskID=4434883

Please test and let us know.

Quoted 92 lines · Takashi Iwai

At Wed, 29 Aug 2012 13:26:25 +0200,

Quoted 90 lines · Daniel Mack

[1 <text/plain; ISO-8859-1 (7bit)>]

Quoted 22 lines · Josh Boyer

Quoted 13 lines · Daniel Mack

Quoted 12 lines · Bruno Wolff III

Quoted 3 lines · Daniel Mack

Can you revert commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") and see if that

I can try that, but it takes a long time to build a new kernel on my old hardware.

Quoted 2 lines

helps? If not, can you summarize again which kernels still work for you and which don't?

Quoted 7 lines

Quoted 3 lines

The latest kernel that works is 3.6.0-0.rc2.git1.2.fc18. The earliest that doesn't work is 3.6.0-0.rc2.git2.1.fc18.

The report you sent doesn't look like it could be caused by e9ba389c5. It fixes a kernel Ooops. But as it is the only relevant patch in that area, it would be interesting if reverting it fixes anything.

Quoted 6 lines

Yep, agreed. If this revert kernel doesn't work, we're likely down to a git bisect, Bruno.

Quoted 1 line

Btw - thanks a lot for testing -rc kernels, much appreciated!

Indeed!

Could you please try this patch on top of Takashi's? Thanks again!

Daniel

[2 0001-ALSA-snd-usb-Fix-URB-cancellation-at-stream-start.patch <text/x-patch (7bit)>]

Quoted 1 line

From 6617bb2463ae3fad21390b59cc2a71f96b9e9ca8 Mon Sep 17 00:00:00 2001

From: Daniel Mack <[email protected]>
Date: Wed, 29 Aug 2012 13:17:05 +0200
Subject: [PATCH] ALSA: snd-usb: Fix URB cancellation at stream start

Commit e9ba389c5 ("ALSA: usb-audio: Fix scheduling-while-atomic bug in PCM capture stream") fixed a scheduling-while-atomic bug that happened when snd_usb_endpoint_start was called from the trigger callback, which is an atmic context. However, the patch breaks the idea of the endpoints reference counting, which is the reason why the driver has been refactored lately.

Revert that commit and let snd_usb_endpoint_start() take care of the URB cancellation again. As this function is called from both atomic and non-atomic context, add a flag to denote whether the function may sleep.

Signed-off-by Daniel Mack <[email protected]>

Cc [email protected] [3.5+]

---
 sound/usb/endpoint.c | 10 ++++++++--
 sound/usb/endpoint.h |  2 +-
 sound/usb/pcm.c      | 13 +++++--------
 3 files changed, 14 insertions(+), 11 deletions(-)

sound/usb/endpoint.c +4 -2

index c411812..678456c 100644
@@ -799,7 +799,9 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
 /**
  * snd_usb_endpoint_start: start an snd_usb_endpoint
  *
- * @ep: the endpoint to start
+ * @ep:		the endpoint to start
+ * @can_sleep:	flag indicating whether the operation is executed in
+ * 		non-atomic context
  *
  * A call to this function will increment the use count of the endpoint.
  * In case it is not already running, the URBs for this endpoint will be

@@ -809,7 +811,7 @@ int snd_usb_endpoint_set_params(struct snd_usb_endpoint *ep,
  *
  * Returns an error if the URB submission failed, 0 in all other cases.
  */
-int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
+int snd_usb_endpoint_start(struct snd_usb_endpoint *ep, int can_sleep)
 {
 	int err;
 	unsigned int i;

@@ -821,6 +823,10 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
 	if (++ep->use_count != 1)
 		return 0;
 

Quoted 6 lines

sound/usb/endpoint.c int snd_usb_endpoint_start(struct snd_usb_endpoint *ep) · line 829

Quoted 3 lines

sound/usb/endpoint.c +3 -0

+	/* just to be sure */
+	deactivate_urbs(ep, 0, can_sleep);
+	wait_clear_urbs(ep);

It'd be safer to protect the call of wait_clear_urbs() when can_sleep=0.

Right. New patch attached.

As the test result looks positive, I applied it for the next pull
request. Thanks.