kernel_optimize_test/sound/usb/line6/pcm.h

200 lines
5.3 KiB
C
Raw Normal View History

/*
* Line 6 Linux USB driver
*
* Copyright (C) 2004-2010 Markus Grabner (grabner@icg.tugraz.at)
*
* 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 the Free Software Foundation, version 2.
*
*/
/*
PCM interface to POD series devices.
*/
#ifndef PCM_H
#define PCM_H
#include <sound/pcm.h>
#include "driver.h"
/*
number of USB frames per URB
The Line 6 Windows driver always transmits two frames per packet, but
the Linux driver performs significantly better (i.e., lower latency)
with only one frame per packet.
*/
#define LINE6_ISO_PACKETS 1
/* in a "full speed" device (such as the PODxt Pro) this means 1ms,
* for "high speed" it's 1/8ms
*/
#define LINE6_ISO_INTERVAL 1
#define LINE6_IMPULSE_DEFAULT_PERIOD 100
/*
Get substream from Line 6 PCM data structure
*/
#define get_substream(line6pcm, stream) \
(line6pcm->pcm->streams[stream].substream)
/*
staging: line6: separate handling of buffer allocation and stream startup There are several features of the Line6 USB driver which require PCM data to be exchanged with the device: *) PCM playback and capture via ALSA *) software monitoring (for devices without hardware monitoring) *) optional impulse response measurement However, from the device's point of view, there is just a single capture and playback stream, which must be shared between these subsystems. It is therefore necessary to maintain the state of the subsystems with respect to PCM usage. We define several constants of the form LINE6_BIT_PCM_<subsystem>_<direction>_<resource> with the following meanings: *) <subsystem> is one of -) ALSA: PCM playback and capture via ALSA -) MONITOR: software monitoring -) IMPULSE: optional impulse response measurement *) <direction> is one of -) PLAYBACK: audio output (from host to device) -) CAPTURE: audio input (from device to host) *) <resource> is one of -) BUFFER: buffer required by PCM data stream -) STREAM: actual PCM data stream The subsystems call line6_pcm_acquire() to acquire the (shared) resources needed for a particular operation (e.g., allocate the buffer for ALSA playback or start the capture stream for software monitoring). When a resource is no longer needed, it is released by calling line6_pcm_release(). Buffer allocation and stream startup are handled separately to allow the ALSA kernel driver to perform them at appropriate places (since the callback which starts a PCM stream is not allowed to sleep). Signed-off-by: Markus Grabner <grabner@icg.tugraz.at> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-01-20 07:09:09 +08:00
PCM mode bits.
There are several features of the Line 6 USB driver which require PCM
staging: line6: separate handling of buffer allocation and stream startup There are several features of the Line6 USB driver which require PCM data to be exchanged with the device: *) PCM playback and capture via ALSA *) software monitoring (for devices without hardware monitoring) *) optional impulse response measurement However, from the device's point of view, there is just a single capture and playback stream, which must be shared between these subsystems. It is therefore necessary to maintain the state of the subsystems with respect to PCM usage. We define several constants of the form LINE6_BIT_PCM_<subsystem>_<direction>_<resource> with the following meanings: *) <subsystem> is one of -) ALSA: PCM playback and capture via ALSA -) MONITOR: software monitoring -) IMPULSE: optional impulse response measurement *) <direction> is one of -) PLAYBACK: audio output (from host to device) -) CAPTURE: audio input (from device to host) *) <resource> is one of -) BUFFER: buffer required by PCM data stream -) STREAM: actual PCM data stream The subsystems call line6_pcm_acquire() to acquire the (shared) resources needed for a particular operation (e.g., allocate the buffer for ALSA playback or start the capture stream for software monitoring). When a resource is no longer needed, it is released by calling line6_pcm_release(). Buffer allocation and stream startup are handled separately to allow the ALSA kernel driver to perform them at appropriate places (since the callback which starts a PCM stream is not allowed to sleep). Signed-off-by: Markus Grabner <grabner@icg.tugraz.at> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-01-20 07:09:09 +08:00
data to be exchanged with the device:
*) PCM playback and capture via ALSA
*) software monitoring (for devices without hardware monitoring)
*) optional impulse response measurement
However, from the device's point of view, there is just a single
capture and playback stream, which must be shared between these
subsystems. It is therefore necessary to maintain the state of the
ALSA: line6: Reorganize PCM stream handling The current code deals with the stream start / stop solely via line6_pcm_acquire() and line6_pcm_release(). This was (supposedly) intended to avoid the races, but it doesn't work as expected. The concurrent acquire and release calls can be performed without proper protections, thus this might result in memory corruption. Furthermore, we can't take a mutex to protect the whole function because it can be called from the PCM trigger callback that is an atomic context. Also spinlock isn't appropriate because the function allocates with kmalloc with GFP_KERNEL. That is, these function just lead to singular problems. This is an attempt to reduce the existing races. First off, separate both the stream buffer management and the stream URB management. The former is protected via a newly introduced state_mutex while the latter is protected via each line6_pcm_stream lock. Secondly, the stream state are now managed in opened and running bit flags of each line6_pcm_stream. Not only this a bit clearer than previous combined bit flags, this also gives a better abstraction. These rewrites allows us to make common hw_params and hw_free callbacks for both playback and capture directions. For the monitor and impulse operations, still line6_pcm_acquire() and line6_pcm_release() are used. They call internally the corresponding functions for both playback and capture streams with proper lock or mutex. Unlike the previous versions, these function don't take the bit masks but the only single type value. Also they are supposed to be applied only as duplex operations. Tested-by: Chris Rorvick <chris@rorvick.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-01-27 22:24:09 +08:00
subsystems with respect to PCM usage.
We define two bit flags, "opened" and "running", for each playback
or capture stream. Both can contain the bit flag corresponding to
LINE6_STREAM_* type,
LINE6_STREAM_PCM = ALSA PCM playback or capture
LINE6_STREAM_MONITOR = software monitoring
IMPULSE = optional impulse response measurement
The opened flag indicates whether the buffer is allocated while
the running flag indicates whether the stream is running.
For monitor or impulse operations, the driver needs to call
line6_pcm_acquire() or line6_pcm_release() with the appropriate
LINE6_STREAM_* flag.
*/
ALSA: line6: Reorganize PCM stream handling The current code deals with the stream start / stop solely via line6_pcm_acquire() and line6_pcm_release(). This was (supposedly) intended to avoid the races, but it doesn't work as expected. The concurrent acquire and release calls can be performed without proper protections, thus this might result in memory corruption. Furthermore, we can't take a mutex to protect the whole function because it can be called from the PCM trigger callback that is an atomic context. Also spinlock isn't appropriate because the function allocates with kmalloc with GFP_KERNEL. That is, these function just lead to singular problems. This is an attempt to reduce the existing races. First off, separate both the stream buffer management and the stream URB management. The former is protected via a newly introduced state_mutex while the latter is protected via each line6_pcm_stream lock. Secondly, the stream state are now managed in opened and running bit flags of each line6_pcm_stream. Not only this a bit clearer than previous combined bit flags, this also gives a better abstraction. These rewrites allows us to make common hw_params and hw_free callbacks for both playback and capture directions. For the monitor and impulse operations, still line6_pcm_acquire() and line6_pcm_release() are used. They call internally the corresponding functions for both playback and capture streams with proper lock or mutex. Unlike the previous versions, these function don't take the bit masks but the only single type value. Also they are supposed to be applied only as duplex operations. Tested-by: Chris Rorvick <chris@rorvick.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-01-27 22:24:09 +08:00
/* stream types */
enum {
LINE6_STREAM_PCM,
LINE6_STREAM_MONITOR,
LINE6_STREAM_IMPULSE,
LINE6_STREAM_CAPTURE_HELPER,
ALSA: line6: Reorganize PCM stream handling The current code deals with the stream start / stop solely via line6_pcm_acquire() and line6_pcm_release(). This was (supposedly) intended to avoid the races, but it doesn't work as expected. The concurrent acquire and release calls can be performed without proper protections, thus this might result in memory corruption. Furthermore, we can't take a mutex to protect the whole function because it can be called from the PCM trigger callback that is an atomic context. Also spinlock isn't appropriate because the function allocates with kmalloc with GFP_KERNEL. That is, these function just lead to singular problems. This is an attempt to reduce the existing races. First off, separate both the stream buffer management and the stream URB management. The former is protected via a newly introduced state_mutex while the latter is protected via each line6_pcm_stream lock. Secondly, the stream state are now managed in opened and running bit flags of each line6_pcm_stream. Not only this a bit clearer than previous combined bit flags, this also gives a better abstraction. These rewrites allows us to make common hw_params and hw_free callbacks for both playback and capture directions. For the monitor and impulse operations, still line6_pcm_acquire() and line6_pcm_release() are used. They call internally the corresponding functions for both playback and capture streams with proper lock or mutex. Unlike the previous versions, these function don't take the bit masks but the only single type value. Also they are supposed to be applied only as duplex operations. Tested-by: Chris Rorvick <chris@rorvick.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-01-27 22:24:09 +08:00
};
/* misc bit flags for PCM operation */
enum {
ALSA: line6: Reorganize PCM stream handling The current code deals with the stream start / stop solely via line6_pcm_acquire() and line6_pcm_release(). This was (supposedly) intended to avoid the races, but it doesn't work as expected. The concurrent acquire and release calls can be performed without proper protections, thus this might result in memory corruption. Furthermore, we can't take a mutex to protect the whole function because it can be called from the PCM trigger callback that is an atomic context. Also spinlock isn't appropriate because the function allocates with kmalloc with GFP_KERNEL. That is, these function just lead to singular problems. This is an attempt to reduce the existing races. First off, separate both the stream buffer management and the stream URB management. The former is protected via a newly introduced state_mutex while the latter is protected via each line6_pcm_stream lock. Secondly, the stream state are now managed in opened and running bit flags of each line6_pcm_stream. Not only this a bit clearer than previous combined bit flags, this also gives a better abstraction. These rewrites allows us to make common hw_params and hw_free callbacks for both playback and capture directions. For the monitor and impulse operations, still line6_pcm_acquire() and line6_pcm_release() are used. They call internally the corresponding functions for both playback and capture streams with proper lock or mutex. Unlike the previous versions, these function don't take the bit masks but the only single type value. Also they are supposed to be applied only as duplex operations. Tested-by: Chris Rorvick <chris@rorvick.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-01-27 22:24:09 +08:00
LINE6_FLAG_PAUSE_PLAYBACK,
LINE6_FLAG_PREPARED,
};
struct line6_pcm_properties {
struct snd_pcm_hardware playback_hw, capture_hw;
struct snd_pcm_hw_constraint_ratdens rates;
int bytes_per_channel;
};
struct line6_pcm_stream {
/* allocated URBs */
struct urb **urbs;
/* Temporary buffer;
* Since the packet size is not known in advance, this buffer is
* large enough to store maximum size packets.
*/
unsigned char *buffer;
/* Free frame position in the buffer. */
snd_pcm_uframes_t pos;
/* Count processed bytes;
* This is modulo period size (to determine when a period is finished).
*/
unsigned bytes;
/* Counter to create desired sample rate */
unsigned count;
/* period size in bytes */
unsigned period;
/* Processed frame position in the buffer;
* The contents of the ring buffer have been consumed by the USB
* subsystem (i.e., sent to the USB device) up to this position.
*/
snd_pcm_uframes_t pos_done;
/* Bit mask of active URBs */
unsigned long active_urbs;
/* Bit mask of URBs currently being unlinked */
unsigned long unlink_urbs;
/* Spin lock to protect updates of the buffer positions (not contents)
*/
spinlock_t lock;
ALSA: line6: Reorganize PCM stream handling The current code deals with the stream start / stop solely via line6_pcm_acquire() and line6_pcm_release(). This was (supposedly) intended to avoid the races, but it doesn't work as expected. The concurrent acquire and release calls can be performed without proper protections, thus this might result in memory corruption. Furthermore, we can't take a mutex to protect the whole function because it can be called from the PCM trigger callback that is an atomic context. Also spinlock isn't appropriate because the function allocates with kmalloc with GFP_KERNEL. That is, these function just lead to singular problems. This is an attempt to reduce the existing races. First off, separate both the stream buffer management and the stream URB management. The former is protected via a newly introduced state_mutex while the latter is protected via each line6_pcm_stream lock. Secondly, the stream state are now managed in opened and running bit flags of each line6_pcm_stream. Not only this a bit clearer than previous combined bit flags, this also gives a better abstraction. These rewrites allows us to make common hw_params and hw_free callbacks for both playback and capture directions. For the monitor and impulse operations, still line6_pcm_acquire() and line6_pcm_release() are used. They call internally the corresponding functions for both playback and capture streams with proper lock or mutex. Unlike the previous versions, these function don't take the bit masks but the only single type value. Also they are supposed to be applied only as duplex operations. Tested-by: Chris Rorvick <chris@rorvick.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-01-27 22:24:09 +08:00
/* Bit flags for operational stream types */
unsigned long opened;
/* Bit flags for running stream types */
unsigned long running;
int last_frame;
};
struct snd_line6_pcm {
/* Pointer back to the Line 6 driver data structure */
struct usb_line6 *line6;
/* Properties. */
struct line6_pcm_properties *properties;
/* ALSA pcm stream */
struct snd_pcm *pcm;
ALSA: line6: Reorganize PCM stream handling The current code deals with the stream start / stop solely via line6_pcm_acquire() and line6_pcm_release(). This was (supposedly) intended to avoid the races, but it doesn't work as expected. The concurrent acquire and release calls can be performed without proper protections, thus this might result in memory corruption. Furthermore, we can't take a mutex to protect the whole function because it can be called from the PCM trigger callback that is an atomic context. Also spinlock isn't appropriate because the function allocates with kmalloc with GFP_KERNEL. That is, these function just lead to singular problems. This is an attempt to reduce the existing races. First off, separate both the stream buffer management and the stream URB management. The former is protected via a newly introduced state_mutex while the latter is protected via each line6_pcm_stream lock. Secondly, the stream state are now managed in opened and running bit flags of each line6_pcm_stream. Not only this a bit clearer than previous combined bit flags, this also gives a better abstraction. These rewrites allows us to make common hw_params and hw_free callbacks for both playback and capture directions. For the monitor and impulse operations, still line6_pcm_acquire() and line6_pcm_release() are used. They call internally the corresponding functions for both playback and capture streams with proper lock or mutex. Unlike the previous versions, these function don't take the bit masks but the only single type value. Also they are supposed to be applied only as duplex operations. Tested-by: Chris Rorvick <chris@rorvick.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-01-27 22:24:09 +08:00
/* protection to state changes of in/out streams */
struct mutex state_mutex;
/* Capture and playback streams */
struct line6_pcm_stream in;
struct line6_pcm_stream out;
/* Previously captured frame (for software monitoring) */
unsigned char *prev_fbuf;
/* Size of previously captured frame (for software monitoring/sync) */
int prev_fsize;
/* Maximum size of USB packet */
int max_packet_size_in;
int max_packet_size_out;
/* PCM playback volume (left and right) */
int volume_playback[2];
/* PCM monitor volume */
int volume_monitor;
/* Volume of impulse response test signal (if zero, test is disabled) */
int impulse_volume;
/* Period of impulse response test signal */
int impulse_period;
/* Counter for impulse response test signal */
int impulse_count;
/* Several status bits (see LINE6_FLAG_*) */
unsigned long flags;
};
extern int line6_init_pcm(struct usb_line6 *line6,
struct line6_pcm_properties *properties);
extern int snd_line6_trigger(struct snd_pcm_substream *substream, int cmd);
extern int snd_line6_prepare(struct snd_pcm_substream *substream);
ALSA: line6: Reorganize PCM stream handling The current code deals with the stream start / stop solely via line6_pcm_acquire() and line6_pcm_release(). This was (supposedly) intended to avoid the races, but it doesn't work as expected. The concurrent acquire and release calls can be performed without proper protections, thus this might result in memory corruption. Furthermore, we can't take a mutex to protect the whole function because it can be called from the PCM trigger callback that is an atomic context. Also spinlock isn't appropriate because the function allocates with kmalloc with GFP_KERNEL. That is, these function just lead to singular problems. This is an attempt to reduce the existing races. First off, separate both the stream buffer management and the stream URB management. The former is protected via a newly introduced state_mutex while the latter is protected via each line6_pcm_stream lock. Secondly, the stream state are now managed in opened and running bit flags of each line6_pcm_stream. Not only this a bit clearer than previous combined bit flags, this also gives a better abstraction. These rewrites allows us to make common hw_params and hw_free callbacks for both playback and capture directions. For the monitor and impulse operations, still line6_pcm_acquire() and line6_pcm_release() are used. They call internally the corresponding functions for both playback and capture streams with proper lock or mutex. Unlike the previous versions, these function don't take the bit masks but the only single type value. Also they are supposed to be applied only as duplex operations. Tested-by: Chris Rorvick <chris@rorvick.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-01-27 22:24:09 +08:00
extern int snd_line6_hw_params(struct snd_pcm_substream *substream,
struct snd_pcm_hw_params *hw_params);
extern int snd_line6_hw_free(struct snd_pcm_substream *substream);
extern snd_pcm_uframes_t snd_line6_pointer(struct snd_pcm_substream *substream);
extern void line6_pcm_disconnect(struct snd_line6_pcm *line6pcm);
extern int line6_pcm_acquire(struct snd_line6_pcm *line6pcm, int type,
bool start);
ALSA: line6: Reorganize PCM stream handling The current code deals with the stream start / stop solely via line6_pcm_acquire() and line6_pcm_release(). This was (supposedly) intended to avoid the races, but it doesn't work as expected. The concurrent acquire and release calls can be performed without proper protections, thus this might result in memory corruption. Furthermore, we can't take a mutex to protect the whole function because it can be called from the PCM trigger callback that is an atomic context. Also spinlock isn't appropriate because the function allocates with kmalloc with GFP_KERNEL. That is, these function just lead to singular problems. This is an attempt to reduce the existing races. First off, separate both the stream buffer management and the stream URB management. The former is protected via a newly introduced state_mutex while the latter is protected via each line6_pcm_stream lock. Secondly, the stream state are now managed in opened and running bit flags of each line6_pcm_stream. Not only this a bit clearer than previous combined bit flags, this also gives a better abstraction. These rewrites allows us to make common hw_params and hw_free callbacks for both playback and capture directions. For the monitor and impulse operations, still line6_pcm_acquire() and line6_pcm_release() are used. They call internally the corresponding functions for both playback and capture streams with proper lock or mutex. Unlike the previous versions, these function don't take the bit masks but the only single type value. Also they are supposed to be applied only as duplex operations. Tested-by: Chris Rorvick <chris@rorvick.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-01-27 22:24:09 +08:00
extern void line6_pcm_release(struct snd_line6_pcm *line6pcm, int type);
#endif