Add the rt linux 4.1.3-rt3 as base
[kvmfornfv.git] / kernel / sound / oss / dmasound / dmasound_paula.c
diff --git a/kernel/sound/oss/dmasound/dmasound_paula.c b/kernel/sound/oss/dmasound/dmasound_paula.c
new file mode 100644 (file)
index 0000000..3f65361
--- /dev/null
@@ -0,0 +1,738 @@
+/*
+ *  linux/sound/oss/dmasound/dmasound_paula.c
+ *
+ *  Amiga `Paula' DMA Sound Driver
+ *
+ *  See linux/sound/oss/dmasound/dmasound_core.c for copyright and credits
+ *  prior to 28/01/2001
+ *
+ *  28/01/2001 [0.1] Iain Sandoe
+ *                  - added versioning
+ *                  - put in and populated the hardware_afmts field.
+ *             [0.2] - put in SNDCTL_DSP_GETCAPS value.
+ *            [0.3] - put in constraint on state buffer usage.
+ *            [0.4] - put in default hard/soft settings
+*/
+
+
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/soundcard.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+
+#include <asm/uaccess.h>
+#include <asm/setup.h>
+#include <asm/amigahw.h>
+#include <asm/amigaints.h>
+#include <asm/machdep.h>
+
+#include "dmasound.h"
+
+#define DMASOUND_PAULA_REVISION 0
+#define DMASOUND_PAULA_EDITION 4
+
+#define custom amiga_custom
+   /*
+    *  The minimum period for audio depends on htotal (for OCS/ECS/AGA)
+    *  (Imported from arch/m68k/amiga/amisound.c)
+    */
+
+extern volatile u_short amiga_audio_min_period;
+
+
+   /*
+    *  amiga_mksound() should be able to restore the period after beeping
+    *  (Imported from arch/m68k/amiga/amisound.c)
+    */
+
+extern u_short amiga_audio_period;
+
+
+   /*
+    *  Audio DMA masks
+    */
+
+#define AMI_AUDIO_OFF  (DMAF_AUD0 | DMAF_AUD1 | DMAF_AUD2 | DMAF_AUD3)
+#define AMI_AUDIO_8    (DMAF_SETCLR | DMAF_MASTER | DMAF_AUD0 | DMAF_AUD1)
+#define AMI_AUDIO_14   (AMI_AUDIO_8 | DMAF_AUD2 | DMAF_AUD3)
+
+
+    /*
+     *  Helper pointers for 16(14)-bit sound
+     */
+
+static int write_sq_block_size_half, write_sq_block_size_quarter;
+
+
+/*** Low level stuff *********************************************************/
+
+
+static void *AmiAlloc(unsigned int size, gfp_t flags);
+static void AmiFree(void *obj, unsigned int size);
+static int AmiIrqInit(void);
+#ifdef MODULE
+static void AmiIrqCleanUp(void);
+#endif
+static void AmiSilence(void);
+static void AmiInit(void);
+static int AmiSetFormat(int format);
+static int AmiSetVolume(int volume);
+static int AmiSetTreble(int treble);
+static void AmiPlayNextFrame(int index);
+static void AmiPlay(void);
+static irqreturn_t AmiInterrupt(int irq, void *dummy);
+
+#ifdef CONFIG_HEARTBEAT
+
+    /*
+     *  Heartbeat interferes with sound since the 7 kHz low-pass filter and the
+     *  power LED are controlled by the same line.
+     */
+
+static void (*saved_heartbeat)(int) = NULL;
+
+static inline void disable_heartbeat(void)
+{
+       if (mach_heartbeat) {
+           saved_heartbeat = mach_heartbeat;
+           mach_heartbeat = NULL;
+       }
+       AmiSetTreble(dmasound.treble);
+}
+
+static inline void enable_heartbeat(void)
+{
+       if (saved_heartbeat)
+           mach_heartbeat = saved_heartbeat;
+}
+#else /* !CONFIG_HEARTBEAT */
+#define disable_heartbeat()    do { } while (0)
+#define enable_heartbeat()     do { } while (0)
+#endif /* !CONFIG_HEARTBEAT */
+
+
+/*** Mid level stuff *********************************************************/
+
+static void AmiMixerInit(void);
+static int AmiMixerIoctl(u_int cmd, u_long arg);
+static int AmiWriteSqSetup(void);
+static int AmiStateInfo(char *buffer, size_t space);
+
+
+/*** Translations ************************************************************/
+
+/* ++TeSche: radically changed for new expanding purposes...
+ *
+ * These two routines now deal with copying/expanding/translating the samples
+ * from user space into our buffer at the right frequency. They take care about
+ * how much data there's actually to read, how much buffer space there is and
+ * to convert samples into the right frequency/encoding. They will only work on
+ * complete samples so it may happen they leave some bytes in the input stream
+ * if the user didn't write a multiple of the current sample size. They both
+ * return the number of bytes they've used from both streams so you may detect
+ * such a situation. Luckily all programs should be able to cope with that.
+ *
+ * I think I've optimized anything as far as one can do in plain C, all
+ * variables should fit in registers and the loops are really short. There's
+ * one loop for every possible situation. Writing a more generalized and thus
+ * parameterized loop would only produce slower code. Feel free to optimize
+ * this in assembler if you like. :)
+ *
+ * I think these routines belong here because they're not yet really hardware
+ * independent, especially the fact that the Falcon can play 16bit samples
+ * only in stereo is hardcoded in both of them!
+ *
+ * ++geert: split in even more functions (one per format)
+ */
+
+
+    /*
+     *  Native format
+     */
+
+static ssize_t ami_ct_s8(const u_char __user *userPtr, size_t userCount,
+                        u_char frame[], ssize_t *frameUsed, ssize_t frameLeft)
+{
+       ssize_t count, used;
+
+       if (!dmasound.soft.stereo) {
+               void *p = &frame[*frameUsed];
+               count = min_t(unsigned long, userCount, frameLeft) & ~1;
+               used = count;
+               if (copy_from_user(p, userPtr, count))
+                       return -EFAULT;
+       } else {
+               u_char *left = &frame[*frameUsed>>1];
+               u_char *right = left+write_sq_block_size_half;
+               count = min_t(unsigned long, userCount, frameLeft)>>1 & ~1;
+               used = count*2;
+               while (count > 0) {
+                       if (get_user(*left++, userPtr++)
+                           || get_user(*right++, userPtr++))
+                               return -EFAULT;
+                       count--;
+               }
+       }
+       *frameUsed += used;
+       return used;
+}
+
+
+    /*
+     *  Copy and convert 8 bit data
+     */
+
+#define GENERATE_AMI_CT8(funcname, convsample)                         \
+static ssize_t funcname(const u_char __user *userPtr, size_t userCount,        \
+                       u_char frame[], ssize_t *frameUsed,             \
+                       ssize_t frameLeft)                              \
+{                                                                      \
+       ssize_t count, used;                                            \
+                                                                       \
+       if (!dmasound.soft.stereo) {                                    \
+               u_char *p = &frame[*frameUsed];                         \
+               count = min_t(size_t, userCount, frameLeft) & ~1;       \
+               used = count;                                           \
+               while (count > 0) {                                     \
+                       u_char data;                                    \
+                       if (get_user(data, userPtr++))                  \
+                               return -EFAULT;                         \
+                       *p++ = convsample(data);                        \
+                       count--;                                        \
+               }                                                       \
+       } else {                                                        \
+               u_char *left = &frame[*frameUsed>>1];                   \
+               u_char *right = left+write_sq_block_size_half;          \
+               count = min_t(size_t, userCount, frameLeft)>>1 & ~1;    \
+               used = count*2;                                         \
+               while (count > 0) {                                     \
+                       u_char data;                                    \
+                       if (get_user(data, userPtr++))                  \
+                               return -EFAULT;                         \
+                       *left++ = convsample(data);                     \
+                       if (get_user(data, userPtr++))                  \
+                               return -EFAULT;                         \
+                       *right++ = convsample(data);                    \
+                       count--;                                        \
+               }                                                       \
+       }                                                               \
+       *frameUsed += used;                                             \
+       return used;                                                    \
+}
+
+#define AMI_CT_ULAW(x) (dmasound_ulaw2dma8[(x)])
+#define AMI_CT_ALAW(x) (dmasound_alaw2dma8[(x)])
+#define AMI_CT_U8(x)   ((x) ^ 0x80)
+
+GENERATE_AMI_CT8(ami_ct_ulaw, AMI_CT_ULAW)
+GENERATE_AMI_CT8(ami_ct_alaw, AMI_CT_ALAW)
+GENERATE_AMI_CT8(ami_ct_u8, AMI_CT_U8)
+
+
+    /*
+     *  Copy and convert 16 bit data
+     */
+
+#define GENERATE_AMI_CT_16(funcname, convsample)                       \
+static ssize_t funcname(const u_char __user *userPtr, size_t userCount,        \
+                       u_char frame[], ssize_t *frameUsed,             \
+                       ssize_t frameLeft)                              \
+{                                                                      \
+       const u_short __user *ptr = (const u_short __user *)userPtr;    \
+       ssize_t count, used;                                            \
+       u_short data;                                                   \
+                                                                       \
+       if (!dmasound.soft.stereo) {                                    \
+               u_char *high = &frame[*frameUsed>>1];                   \
+               u_char *low = high+write_sq_block_size_half;            \
+               count = min_t(size_t, userCount, frameLeft)>>1 & ~1;    \
+               used = count*2;                                         \
+               while (count > 0) {                                     \
+                       if (get_user(data, ptr++))                      \
+                               return -EFAULT;                         \
+                       data = convsample(data);                        \
+                       *high++ = data>>8;                              \
+                       *low++ = (data>>2) & 0x3f;                      \
+                       count--;                                        \
+               }                                                       \
+       } else {                                                        \
+               u_char *lefth = &frame[*frameUsed>>2];                  \
+               u_char *leftl = lefth+write_sq_block_size_quarter;      \
+               u_char *righth = lefth+write_sq_block_size_half;        \
+               u_char *rightl = righth+write_sq_block_size_quarter;    \
+               count = min_t(size_t, userCount, frameLeft)>>2 & ~1;    \
+               used = count*4;                                         \
+               while (count > 0) {                                     \
+                       if (get_user(data, ptr++))                      \
+                               return -EFAULT;                         \
+                       data = convsample(data);                        \
+                       *lefth++ = data>>8;                             \
+                       *leftl++ = (data>>2) & 0x3f;                    \
+                       if (get_user(data, ptr++))                      \
+                               return -EFAULT;                         \
+                       data = convsample(data);                        \
+                       *righth++ = data>>8;                            \
+                       *rightl++ = (data>>2) & 0x3f;                   \
+                       count--;                                        \
+               }                                                       \
+       }                                                               \
+       *frameUsed += used;                                             \
+       return used;                                                    \
+}
+
+#define AMI_CT_S16BE(x)        (x)
+#define AMI_CT_U16BE(x)        ((x) ^ 0x8000)
+#define AMI_CT_S16LE(x)        (le2be16((x)))
+#define AMI_CT_U16LE(x)        (le2be16((x)) ^ 0x8000)
+
+GENERATE_AMI_CT_16(ami_ct_s16be, AMI_CT_S16BE)
+GENERATE_AMI_CT_16(ami_ct_u16be, AMI_CT_U16BE)
+GENERATE_AMI_CT_16(ami_ct_s16le, AMI_CT_S16LE)
+GENERATE_AMI_CT_16(ami_ct_u16le, AMI_CT_U16LE)
+
+
+static TRANS transAmiga = {
+       .ct_ulaw        = ami_ct_ulaw,
+       .ct_alaw        = ami_ct_alaw,
+       .ct_s8          = ami_ct_s8,
+       .ct_u8          = ami_ct_u8,
+       .ct_s16be       = ami_ct_s16be,
+       .ct_u16be       = ami_ct_u16be,
+       .ct_s16le       = ami_ct_s16le,
+       .ct_u16le       = ami_ct_u16le,
+};
+
+/*** Low level stuff *********************************************************/
+
+static inline void StopDMA(void)
+{
+       custom.aud[0].audvol = custom.aud[1].audvol = 0;
+       custom.aud[2].audvol = custom.aud[3].audvol = 0;
+       custom.dmacon = AMI_AUDIO_OFF;
+       enable_heartbeat();
+}
+
+static void *AmiAlloc(unsigned int size, gfp_t flags)
+{
+       return amiga_chip_alloc((long)size, "dmasound [Paula]");
+}
+
+static void AmiFree(void *obj, unsigned int size)
+{
+       amiga_chip_free (obj);
+}
+
+static int __init AmiIrqInit(void)
+{
+       /* turn off DMA for audio channels */
+       StopDMA();
+
+       /* Register interrupt handler. */
+       if (request_irq(IRQ_AMIGA_AUD0, AmiInterrupt, 0, "DMA sound",
+                       AmiInterrupt))
+               return 0;
+       return 1;
+}
+
+#ifdef MODULE
+static void AmiIrqCleanUp(void)
+{
+       /* turn off DMA for audio channels */
+       StopDMA();
+       /* release the interrupt */
+       free_irq(IRQ_AMIGA_AUD0, AmiInterrupt);
+}
+#endif /* MODULE */
+
+static void AmiSilence(void)
+{
+       /* turn off DMA for audio channels */
+       StopDMA();
+}
+
+
+static void AmiInit(void)
+{
+       int period, i;
+
+       AmiSilence();
+
+       if (dmasound.soft.speed)
+               period = amiga_colorclock/dmasound.soft.speed-1;
+       else
+               period = amiga_audio_min_period;
+       dmasound.hard = dmasound.soft;
+       dmasound.trans_write = &transAmiga;
+
+       if (period < amiga_audio_min_period) {
+               /* we would need to squeeze the sound, but we won't do that */
+               period = amiga_audio_min_period;
+       } else if (period > 65535) {
+               period = 65535;
+       }
+       dmasound.hard.speed = amiga_colorclock/(period+1);
+
+       for (i = 0; i < 4; i++)
+               custom.aud[i].audper = period;
+       amiga_audio_period = period;
+}
+
+
+static int AmiSetFormat(int format)
+{
+       int size;
+
+       /* Amiga sound DMA supports 8bit and 16bit (pseudo 14 bit) modes */
+
+       switch (format) {
+       case AFMT_QUERY:
+               return dmasound.soft.format;
+       case AFMT_MU_LAW:
+       case AFMT_A_LAW:
+       case AFMT_U8:
+       case AFMT_S8:
+               size = 8;
+               break;
+       case AFMT_S16_BE:
+       case AFMT_U16_BE:
+       case AFMT_S16_LE:
+       case AFMT_U16_LE:
+               size = 16;
+               break;
+       default: /* :-) */
+               size = 8;
+               format = AFMT_S8;
+       }
+
+       dmasound.soft.format = format;
+       dmasound.soft.size = size;
+       if (dmasound.minDev == SND_DEV_DSP) {
+               dmasound.dsp.format = format;
+               dmasound.dsp.size = dmasound.soft.size;
+       }
+       AmiInit();
+
+       return format;
+}
+
+
+#define VOLUME_VOXWARE_TO_AMI(v) \
+       (((v) < 0) ? 0 : ((v) > 100) ? 64 : ((v) * 64)/100)
+#define VOLUME_AMI_TO_VOXWARE(v) ((v)*100/64)
+
+static int AmiSetVolume(int volume)
+{
+       dmasound.volume_left = VOLUME_VOXWARE_TO_AMI(volume & 0xff);
+       custom.aud[0].audvol = dmasound.volume_left;
+       dmasound.volume_right = VOLUME_VOXWARE_TO_AMI((volume & 0xff00) >> 8);
+       custom.aud[1].audvol = dmasound.volume_right;
+       if (dmasound.hard.size == 16) {
+               if (dmasound.volume_left == 64 && dmasound.volume_right == 64) {
+                       custom.aud[2].audvol = 1;
+                       custom.aud[3].audvol = 1;
+               } else {
+                       custom.aud[2].audvol = 0;
+                       custom.aud[3].audvol = 0;
+               }
+       }
+       return VOLUME_AMI_TO_VOXWARE(dmasound.volume_left) |
+              (VOLUME_AMI_TO_VOXWARE(dmasound.volume_right) << 8);
+}
+
+static int AmiSetTreble(int treble)
+{
+       dmasound.treble = treble;
+       if (treble < 50)
+               ciaa.pra &= ~0x02;
+       else
+               ciaa.pra |= 0x02;
+       return treble;
+}
+
+
+#define AMI_PLAY_LOADED                1
+#define AMI_PLAY_PLAYING       2
+#define AMI_PLAY_MASK          3
+
+
+static void AmiPlayNextFrame(int index)
+{
+       u_char *start, *ch0, *ch1, *ch2, *ch3;
+       u_long size;
+
+       /* used by AmiPlay() if all doubts whether there really is something
+        * to be played are already wiped out.
+        */
+       start = write_sq.buffers[write_sq.front];
+       size = (write_sq.count == index ? write_sq.rear_size
+                                       : write_sq.block_size)>>1;
+
+       if (dmasound.hard.stereo) {
+               ch0 = start;
+               ch1 = start+write_sq_block_size_half;
+               size >>= 1;
+       } else {
+               ch0 = start;
+               ch1 = start;
+       }
+
+       disable_heartbeat();
+       custom.aud[0].audvol = dmasound.volume_left;
+       custom.aud[1].audvol = dmasound.volume_right;
+       if (dmasound.hard.size == 8) {
+               custom.aud[0].audlc = (u_short *)ZTWO_PADDR(ch0);
+               custom.aud[0].audlen = size;
+               custom.aud[1].audlc = (u_short *)ZTWO_PADDR(ch1);
+               custom.aud[1].audlen = size;
+               custom.dmacon = AMI_AUDIO_8;
+       } else {
+               size >>= 1;
+               custom.aud[0].audlc = (u_short *)ZTWO_PADDR(ch0);
+               custom.aud[0].audlen = size;
+               custom.aud[1].audlc = (u_short *)ZTWO_PADDR(ch1);
+               custom.aud[1].audlen = size;
+               if (dmasound.volume_left == 64 && dmasound.volume_right == 64) {
+                       /* We can play pseudo 14-bit only with the maximum volume */
+                       ch3 = ch0+write_sq_block_size_quarter;
+                       ch2 = ch1+write_sq_block_size_quarter;
+                       custom.aud[2].audvol = 1;  /* we are being affected by the beeps */
+                       custom.aud[3].audvol = 1;  /* restoring volume here helps a bit */
+                       custom.aud[2].audlc = (u_short *)ZTWO_PADDR(ch2);
+                       custom.aud[2].audlen = size;
+                       custom.aud[3].audlc = (u_short *)ZTWO_PADDR(ch3);
+                       custom.aud[3].audlen = size;
+                       custom.dmacon = AMI_AUDIO_14;
+               } else {
+                       custom.aud[2].audvol = 0;
+                       custom.aud[3].audvol = 0;
+                       custom.dmacon = AMI_AUDIO_8;
+               }
+       }
+       write_sq.front = (write_sq.front+1) % write_sq.max_count;
+       write_sq.active |= AMI_PLAY_LOADED;
+}
+
+
+static void AmiPlay(void)
+{
+       int minframes = 1;
+
+       custom.intena = IF_AUD0;
+
+       if (write_sq.active & AMI_PLAY_LOADED) {
+               /* There's already a frame loaded */
+               custom.intena = IF_SETCLR | IF_AUD0;
+               return;
+       }
+
+       if (write_sq.active & AMI_PLAY_PLAYING)
+               /* Increase threshold: frame 1 is already being played */
+               minframes = 2;
+
+       if (write_sq.count < minframes) {
+               /* Nothing to do */
+               custom.intena = IF_SETCLR | IF_AUD0;
+               return;
+       }
+
+       if (write_sq.count <= minframes &&
+           write_sq.rear_size < write_sq.block_size && !write_sq.syncing) {
+               /* hmmm, the only existing frame is not
+                * yet filled and we're not syncing?
+                */
+               custom.intena = IF_SETCLR | IF_AUD0;
+               return;
+       }
+
+       AmiPlayNextFrame(minframes);
+
+       custom.intena = IF_SETCLR | IF_AUD0;
+}
+
+
+static irqreturn_t AmiInterrupt(int irq, void *dummy)
+{
+       int minframes = 1;
+
+       custom.intena = IF_AUD0;
+
+       if (!write_sq.active) {
+               /* Playing was interrupted and sq_reset() has already cleared
+                * the sq variables, so better don't do anything here.
+                */
+               WAKE_UP(write_sq.sync_queue);
+               return IRQ_HANDLED;
+       }
+
+       if (write_sq.active & AMI_PLAY_PLAYING) {
+               /* We've just finished a frame */
+               write_sq.count--;
+               WAKE_UP(write_sq.action_queue);
+       }
+
+       if (write_sq.active & AMI_PLAY_LOADED)
+               /* Increase threshold: frame 1 is already being played */
+               minframes = 2;
+
+       /* Shift the flags */
+       write_sq.active = (write_sq.active<<1) & AMI_PLAY_MASK;
+
+       if (!write_sq.active)
+               /* No frame is playing, disable audio DMA */
+               StopDMA();
+
+       custom.intena = IF_SETCLR | IF_AUD0;
+
+       if (write_sq.count >= minframes)
+               /* Try to play the next frame */
+               AmiPlay();
+
+       if (!write_sq.active)
+               /* Nothing to play anymore.
+                  Wake up a process waiting for audio output to drain. */
+               WAKE_UP(write_sq.sync_queue);
+       return IRQ_HANDLED;
+}
+
+/*** Mid level stuff *********************************************************/
+
+
+/*
+ * /dev/mixer abstraction
+ */
+
+static void __init AmiMixerInit(void)
+{
+       dmasound.volume_left = 64;
+       dmasound.volume_right = 64;
+       custom.aud[0].audvol = dmasound.volume_left;
+       custom.aud[3].audvol = 1;       /* For pseudo 14bit */
+       custom.aud[1].audvol = dmasound.volume_right;
+       custom.aud[2].audvol = 1;       /* For pseudo 14bit */
+       dmasound.treble = 50;
+}
+
+static int AmiMixerIoctl(u_int cmd, u_long arg)
+{
+       int data;
+       switch (cmd) {
+           case SOUND_MIXER_READ_DEVMASK:
+                   return IOCTL_OUT(arg, SOUND_MASK_VOLUME | SOUND_MASK_TREBLE);
+           case SOUND_MIXER_READ_RECMASK:
+                   return IOCTL_OUT(arg, 0);
+           case SOUND_MIXER_READ_STEREODEVS:
+                   return IOCTL_OUT(arg, SOUND_MASK_VOLUME);
+           case SOUND_MIXER_READ_VOLUME:
+                   return IOCTL_OUT(arg,
+                           VOLUME_AMI_TO_VOXWARE(dmasound.volume_left) |
+                           VOLUME_AMI_TO_VOXWARE(dmasound.volume_right) << 8);
+           case SOUND_MIXER_WRITE_VOLUME:
+                   IOCTL_IN(arg, data);
+                   return IOCTL_OUT(arg, dmasound_set_volume(data));
+           case SOUND_MIXER_READ_TREBLE:
+                   return IOCTL_OUT(arg, dmasound.treble);
+           case SOUND_MIXER_WRITE_TREBLE:
+                   IOCTL_IN(arg, data);
+                   return IOCTL_OUT(arg, dmasound_set_treble(data));
+       }
+       return -EINVAL;
+}
+
+
+static int AmiWriteSqSetup(void)
+{
+       write_sq_block_size_half = write_sq.block_size>>1;
+       write_sq_block_size_quarter = write_sq_block_size_half>>1;
+       return 0;
+}
+
+
+static int AmiStateInfo(char *buffer, size_t space)
+{
+       int len = 0;
+       len += sprintf(buffer+len, "\tsound.volume_left = %d [0...64]\n",
+                      dmasound.volume_left);
+       len += sprintf(buffer+len, "\tsound.volume_right = %d [0...64]\n",
+                      dmasound.volume_right);
+       if (len >= space) {
+               printk(KERN_ERR "dmasound_paula: overflowed state buffer alloc.\n") ;
+               len = space ;
+       }
+       return len;
+}
+
+
+/*** Machine definitions *****************************************************/
+
+static SETTINGS def_hard = {
+       .format = AFMT_S8,
+       .stereo = 0,
+       .size   = 8,
+       .speed  = 8000
+} ;
+
+static SETTINGS def_soft = {
+       .format = AFMT_U8,
+       .stereo = 0,
+       .size   = 8,
+       .speed  = 8000
+} ;
+
+static MACHINE machAmiga = {
+       .name           = "Amiga",
+       .name2          = "AMIGA",
+       .owner          = THIS_MODULE,
+       .dma_alloc      = AmiAlloc,
+       .dma_free       = AmiFree,
+       .irqinit        = AmiIrqInit,
+#ifdef MODULE
+       .irqcleanup     = AmiIrqCleanUp,
+#endif /* MODULE */
+       .init           = AmiInit,
+       .silence        = AmiSilence,
+       .setFormat      = AmiSetFormat,
+       .setVolume      = AmiSetVolume,
+       .setTreble      = AmiSetTreble,
+       .play           = AmiPlay,
+       .mixer_init     = AmiMixerInit,
+       .mixer_ioctl    = AmiMixerIoctl,
+       .write_sq_setup = AmiWriteSqSetup,
+       .state_info     = AmiStateInfo,
+       .min_dsp_speed  = 8000,
+       .version        = ((DMASOUND_PAULA_REVISION<<8) | DMASOUND_PAULA_EDITION),
+       .hardware_afmts = (AFMT_S8 | AFMT_S16_BE), /* h'ware-supported formats *only* here */
+       .capabilities   = DSP_CAP_BATCH          /* As per SNDCTL_DSP_GETCAPS */
+};
+
+
+/*** Config & Setup **********************************************************/
+
+
+static int __init amiga_audio_probe(struct platform_device *pdev)
+{
+       dmasound.mach = machAmiga;
+       dmasound.mach.default_hard = def_hard ;
+       dmasound.mach.default_soft = def_soft ;
+       return dmasound_init();
+}
+
+static int __exit amiga_audio_remove(struct platform_device *pdev)
+{
+       dmasound_deinit();
+       return 0;
+}
+
+static struct platform_driver amiga_audio_driver = {
+       .remove = __exit_p(amiga_audio_remove),
+       .driver   = {
+               .name   = "amiga-audio",
+       },
+};
+
+module_platform_driver_probe(amiga_audio_driver, amiga_audio_probe);
+
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:amiga-audio");