At Fri, 3 Jul 2009 12:32:25 +0530, Vinod Koul wrote:
This patch adds the intel_sst.h header file that is places in the include folder for MAD driver (ALSA sound card driver for the platform given in later patches) to use. This file contains the definitions of interfaces exposed by the SST drivers along with the definitions of all the controls for the sound card to be used by MAD driver. intel_sst_ioctl.h - this file exposes the ioctls definitions for player/middleware to use. Many of the definitions are reuse from ALSA
Signed-off-by: Vinod Koul vinod.koul@intel.com Signed-off-by: Harsha Priya priya.harsha@intel.com
new file: include/sound/intel_sst.h new file: include/sound/intel_sst_ioctl.h
include/sound/intel_sst.h | 259 +++++++++++++++++++++++++++++ include/sound/intel_sst_ioctl.h | 351 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 610 insertions(+), 0 deletions(-) create mode 100644 include/sound/intel_sst.h create mode 100644 include/sound/intel_sst_ioctl.h
diff --git a/include/sound/intel_sst.h b/include/sound/intel_sst.h new file mode 100644 index 0000000..c709390 --- /dev/null +++ b/include/sound/intel_sst.h
(snip)
+static inline int sst_sc_reg_access(struct sc_reg_access *sc_access,
int type, int num_val)+{ +#ifndef CONFIG_SST_IPC_NOT_INCLUDED
- int i, retval = 0, j = 0, k = 0, count = 0;
- struct ipc_pmic_reg_data reg_data;
- struct ipc_pmic_mod_reg_data pmic_mod_reg = {0};
This is way too big as an inline function.
diff --git a/include/sound/intel_sst_ioctl.h b/include/sound/intel_sst_ioctl.h new file mode 100644 index 0000000..848ea26 --- /dev/null +++ b/include/sound/intel_sst_ioctl.h
+enum sst_codec_types { +/* AUDIO/MUSIC CODEC Type Definitions */
- SST_CODEC_TYPE_UNKNOWN = 0,
- SST_CODEC_TYPE_PCM, /* Pass through Audio codec */
- SST_CODEC_TYPE_MP3,
- SST_CODEC_TYPE_MP24,
- SST_CODEC_TYPE_AAC,
- SST_CODEC_TYPE_AACP,
- SST_CODEC_TYPE_eAACP,
- SST_CODEC_TYPE_WMA9,
- SST_CODEC_TYPE_WMA10,
- SST_CODEC_TYPE_WMA10P,
- SST_CODEC_TYPE_RA,
- SST_CODEC_TYPE_DDAC3,
- SST_CODEC_TYPE_STEREO_TRUE_HD,
- SST_CODEC_TYPE_STEREO_HD_PLUS,
- /* VOICE CODEC Type Definitions */
- SST_CODEC_TYPE_VOICE_PCM = 0x21, /* Pass through voice codec */
- SRC = 0x64,
- MIXER = 0x65,
- DOWN_MIXER = 0x66,
- VOLUME_CONTROL = 0x67,
- OEM1 = 0xC8,
- OEM2 = 0xC9,
These terms are too generic to be used as public definitions. Better to put a prefix.
+/* Firmware Version info */ +struct snd_sst_fw_version {
- __u8 build; /* build number*/
- __u8 minor; /* minor number*/
- __u8 major; /* major number*/
- __u8 type; /* build type*/
+};
+/* Port info structure */ +struct snd_sst_port_info {
- __u16 port_type;
Just wondering -- is there big-endian support?
+/* Target selection per device structure */ +struct snd_sst_slot_info {
- __u8 mix_enable; /* Mixer enable or disable */
- __u8 device_type;
- __u8 device_instance; /* 0, 1, 2 */
- __u16 slot[2];
- struct snd_sst_pmic_config pcm_params;
- enum snd_sst_target_type type;
Better to use another type with a strict size instead of enum.
+struct snd_sst_vol {
- unsigned int stream_id;
- int volume;
- unsigned long ramp_duration;
Are you sure to use long? Long can be different between 32 and 64bit architectures.
+struct snd_sst_buff_entry {
- union {
void *user;unsigned int offset;- } buffer;
Is it OK? The pointer and int can be different sizes.
Takashi