Merge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
diff --git a/arch/sparc64/kernel/entry.S b/arch/sparc64/kernel/entry.S
index 3e0badb..b483495 100644
--- a/arch/sparc64/kernel/entry.S
+++ b/arch/sparc64/kernel/entry.S
@@ -42,19 +42,15 @@
  * executing (see inherit_locked_prom_mappings() rant).
  */
 sparc64_vpte_nucleus:
-	/* Load 0xf0000000, which is LOW_OBP_ADDRESS.  */
-	mov		0xf, %g5
-	sllx		%g5, 28, %g5
-
-	/* Is addr >= LOW_OBP_ADDRESS?  */
+	/* Note that kvmap below has verified that the address is
+	 * in the range MODULES_VADDR --> VMALLOC_END already.  So
+	 * here we need only check if it is an OBP address or not.
+	 */
+	sethi		%hi(LOW_OBP_ADDRESS), %g5
 	cmp		%g4, %g5
 	blu,pn		%xcc, sparc64_vpte_patchme1
 	 mov		0x1, %g5
-
-	/* Load 0x100000000, which is HI_OBP_ADDRESS.  */
 	sllx		%g5, 32, %g5
-
-	/* Is addr < HI_OBP_ADDRESS?  */
 	cmp		%g4, %g5
 	blu,pn		%xcc, obp_iaddr_patch
 	 nop
@@ -156,26 +152,29 @@
  * rather, use information saved during inherit_prom_mappings() using 8k
  * pagesize.
  */
+	.align		32
 kvmap:
-	/* Load 0xf0000000, which is LOW_OBP_ADDRESS.  */
-	mov		0xf, %g5
-	sllx		%g5, 28, %g5
-
-	/* Is addr >= LOW_OBP_ADDRESS?  */
+	sethi		%hi(MODULES_VADDR), %g5
 	cmp		%g4, %g5
-	blu,pn		%xcc, vmalloc_addr
+	blu,pn		%xcc, longpath
+	 mov		(VMALLOC_END >> 24), %g5
+	sllx		%g5, 24, %g5
+	cmp		%g4, %g5
+	bgeu,pn		%xcc, longpath
+	 nop
+
+kvmap_check_obp:
+	sethi		%hi(LOW_OBP_ADDRESS), %g5
+	cmp		%g4, %g5
+	blu,pn		%xcc, kvmap_vmalloc_addr
 	 mov		0x1, %g5
-
-	/* Load 0x100000000, which is HI_OBP_ADDRESS.  */
 	sllx		%g5, 32, %g5
-
-	/* Is addr < HI_OBP_ADDRESS?  */
 	cmp		%g4, %g5
 	blu,pn		%xcc, obp_daddr_patch
 	 nop
 
-vmalloc_addr:
-	/* If we get here, a vmalloc addr accessed, load kernel VPTE.  */
+kvmap_vmalloc_addr:
+	/* If we get here, a vmalloc addr was accessed, load kernel VPTE.  */
 	ldxa		[%g3 + %g6] ASI_N, %g5
 	brgez,pn	%g5, longpath
 	 nop
diff --git a/arch/sparc64/kernel/ptrace.c b/arch/sparc64/kernel/ptrace.c
index 23ad839..5efbff9 100644
--- a/arch/sparc64/kernel/ptrace.c
+++ b/arch/sparc64/kernel/ptrace.c
@@ -30,6 +30,7 @@
 #include <asm/psrcompat.h>
 #include <asm/visasm.h>
 #include <asm/spitfire.h>
+#include <asm/page.h>
 
 /* Returning from ptrace is a bit tricky because the syscall return
  * low level code assumes any value returned which is negative and
@@ -128,20 +129,20 @@
 	 * is mapped to in the user's address space, we can skip the
 	 * D-cache flush.
 	 */
-	if ((uaddr ^ kaddr) & (1UL << 13)) {
+	if ((uaddr ^ (unsigned long) kaddr) & (1UL << 13)) {
 		unsigned long start = __pa(kaddr);
 		unsigned long end = start + len;
 
 		if (tlb_type == spitfire) {
 			for (; start < end; start += 32)
-				spitfire_put_dcache_tag(va & 0x3fe0, 0x0);
+				spitfire_put_dcache_tag(start & 0x3fe0, 0x0);
 		} else {
 			for (; start < end; start += 32)
 				__asm__ __volatile__(
 					"stxa %%g0, [%0] %1\n\t"
 					"membar #Sync"
 					: /* no outputs */
-					: "r" (va),
+					: "r" (start),
 					"i" (ASI_DCACHE_INVALIDATE));
 		}
 	}
diff --git a/arch/sparc64/kernel/una_asm.S b/arch/sparc64/kernel/una_asm.S
index cbb4058..da48400 100644
--- a/arch/sparc64/kernel/una_asm.S
+++ b/arch/sparc64/kernel/una_asm.S
@@ -17,7 +17,7 @@
 __do_int_store:
 	rd	%asi, %o4
 	wr	%o3, 0, %asi
-	ldx	[%o2], %g3
+	mov	%o2, %g3
 	cmp	%o1, 2
 	be,pn	%icc, 2f
 	 cmp	%o1, 4
diff --git a/arch/sparc64/kernel/unaligned.c b/arch/sparc64/kernel/unaligned.c
index da9739f..42718f6 100644
--- a/arch/sparc64/kernel/unaligned.c
+++ b/arch/sparc64/kernel/unaligned.c
@@ -184,13 +184,14 @@
 			unsigned long *saddr, int is_signed, int asi);
 	
 extern void __do_int_store(unsigned long *dst_addr, int size,
-			   unsigned long *src_val, int asi);
+			   unsigned long src_val, int asi);
 
 static inline void do_int_store(int reg_num, int size, unsigned long *dst_addr,
-				struct pt_regs *regs, int asi)
+				struct pt_regs *regs, int asi, int orig_asi)
 {
 	unsigned long zero = 0;
-	unsigned long *src_val = &zero;
+	unsigned long *src_val_p = &zero;
+	unsigned long src_val;
 
 	if (size == 16) {
 		size = 8;
@@ -198,7 +199,25 @@
 		        (unsigned)fetch_reg(reg_num, regs) : 0)) << 32) |
 			(unsigned)fetch_reg(reg_num + 1, regs);
 	} else if (reg_num) {
-		src_val = fetch_reg_addr(reg_num, regs);
+		src_val_p = fetch_reg_addr(reg_num, regs);
+	}
+	src_val = *src_val_p;
+	if (unlikely(asi != orig_asi)) {
+		switch (size) {
+		case 2:
+			src_val = swab16(src_val);
+			break;
+		case 4:
+			src_val = swab32(src_val);
+			break;
+		case 8:
+			src_val = swab64(src_val);
+			break;
+		case 16:
+		default:
+			BUG();
+			break;
+		};
 	}
 	__do_int_store(dst_addr, size, src_val, asi);
 }
@@ -276,6 +295,7 @@
 		kernel_mna_trap_fault();
 	} else {
 		unsigned long addr;
+		int orig_asi, asi;
 
 		addr = compute_effective_address(regs, insn,
 						 ((insn >> 25) & 0x1f));
@@ -285,18 +305,48 @@
 		       regs->tpc, dirstrings[dir], addr, size,
 		       regs->u_regs[UREG_RETPC]);
 #endif
+		orig_asi = asi = decode_asi(insn, regs);
+		switch (asi) {
+		case ASI_NL:
+		case ASI_AIUPL:
+		case ASI_AIUSL:
+		case ASI_PL:
+		case ASI_SL:
+		case ASI_PNFL:
+		case ASI_SNFL:
+			asi &= ~0x08;
+			break;
+		};
 		switch (dir) {
 		case load:
 			do_int_load(fetch_reg_addr(((insn>>25)&0x1f), regs),
 				    size, (unsigned long *) addr,
-				    decode_signedness(insn),
-				    decode_asi(insn, regs));
+				    decode_signedness(insn), asi);
+			if (unlikely(asi != orig_asi)) {
+				unsigned long val_in = *(unsigned long *) addr;
+				switch (size) {
+				case 2:
+					val_in = swab16(val_in);
+					break;
+				case 4:
+					val_in = swab32(val_in);
+					break;
+				case 8:
+					val_in = swab64(val_in);
+					break;
+				case 16:
+				default:
+					BUG();
+					break;
+				};
+				*(unsigned long *) addr = val_in;
+			}
 			break;
 
 		case store:
 			do_int_store(((insn>>25)&0x1f), size,
 				     (unsigned long *) addr, regs,
-				     decode_asi(insn, regs));
+				     asi, orig_asi);
 			break;
 
 		default:
diff --git a/drivers/video/aty/xlinit.c b/drivers/video/aty/xlinit.c
index 92643af..0bea0d8 100644
--- a/drivers/video/aty/xlinit.c
+++ b/drivers/video/aty/xlinit.c
@@ -174,7 +174,7 @@
 	const struct xl_card_cfg_t * card = &card_cfg[xl_card];
 	struct atyfb_par *par = (struct atyfb_par *) info->par;
 	union aty_pll pll;
-	int i, err;
+	int err;
 	u32 temp;
 	
 	aty_st_8(CONFIG_STAT0, 0x85, par);
@@ -252,9 +252,12 @@
 	aty_st_le32(0xEC, 0x00000000, par);
 	aty_st_le32(0xFC, 0x00000000, par);
 
+#if defined (CONFIG_FB_ATY_GENERIC_LCD)
+	int i;
 	for (i=0; i<sizeof(lcd_tbl)/sizeof(lcd_tbl_t); i++) {
 		aty_st_lcd(lcd_tbl[i].lcd_reg, lcd_tbl[i].val, par);
 	}
+#endif
 
 	aty_st_le16(CONFIG_STAT0, 0x00A4, par);
 	mdelay(10);
diff --git a/include/asm-sparc64/cacheflush.h b/include/asm-sparc64/cacheflush.h
index 51b26e81..ededd26 100644
--- a/include/asm-sparc64/cacheflush.h
+++ b/include/asm-sparc64/cacheflush.h
@@ -4,13 +4,6 @@
 #include <linux/config.h>
 #include <asm/page.h>
 
-/* Flushing for D-cache alias handling is only needed if
- * the page size is smaller than 16K.
- */
-#if PAGE_SHIFT < 14
-#define DCACHE_ALIASING_POSSIBLE
-#endif
-
 #ifndef __ASSEMBLY__
 
 #include <linux/mm.h>
diff --git a/include/asm-sparc64/ide.h b/include/asm-sparc64/ide.h
index 4c10984..c393f81 100644
--- a/include/asm-sparc64/ide.h
+++ b/include/asm-sparc64/ide.h
@@ -15,6 +15,7 @@
 #include <asm/io.h>
 #include <asm/spitfire.h>
 #include <asm/cacheflush.h>
+#include <asm/page.h>
 
 #ifndef MAX_HWIFS
 # ifdef CONFIG_BLK_DEV_IDEPCI
diff --git a/include/asm-sparc64/page.h b/include/asm-sparc64/page.h
index c9f8ef2..7f8d764 100644
--- a/include/asm-sparc64/page.h
+++ b/include/asm-sparc64/page.h
@@ -21,6 +21,13 @@
 #define PAGE_SIZE    (_AC(1,UL) << PAGE_SHIFT)
 #define PAGE_MASK    (~(PAGE_SIZE-1))
 
+/* Flushing for D-cache alias handling is only needed if
+ * the page size is smaller than 16K.
+ */
+#if PAGE_SHIFT < 14
+#define DCACHE_ALIASING_POSSIBLE
+#endif
+
 #ifdef __KERNEL__
 
 #ifndef __ASSEMBLY__
diff --git a/include/asm-sparc64/pgalloc.h b/include/asm-sparc64/pgalloc.h
index b9b1914..a96067c 100644
--- a/include/asm-sparc64/pgalloc.h
+++ b/include/asm-sparc64/pgalloc.h
@@ -10,6 +10,7 @@
 #include <asm/spitfire.h>
 #include <asm/cpudata.h>
 #include <asm/cacheflush.h>
+#include <asm/page.h>
 
 /* Page table allocation/freeing. */
 #ifdef CONFIG_SMP
diff --git a/include/asm-sparc64/pgtable.h b/include/asm-sparc64/pgtable.h
index a2b4f5e..a297f61 100644
--- a/include/asm-sparc64/pgtable.h
+++ b/include/asm-sparc64/pgtable.h
@@ -24,21 +24,23 @@
 #include <asm/processor.h>
 #include <asm/const.h>
 
-/* The kernel image occupies 0x4000000 to 0x1000000 (4MB --> 16MB).
- * The page copy blockops use 0x1000000 to 0x18000000 (16MB --> 24MB).
+/* The kernel image occupies 0x4000000 to 0x1000000 (4MB --> 32MB).
+ * The page copy blockops can use 0x2000000 to 0x10000000.
  * The PROM resides in an area spanning 0xf0000000 to 0x100000000.
- * The vmalloc area spans 0x140000000 to 0x200000000.
+ * The vmalloc area spans 0x100000000 to 0x200000000.
+ * Since modules need to be in the lowest 32-bits of the address space,
+ * we place them right before the OBP area from 0x10000000 to 0xf0000000.
  * There is a single static kernel PMD which maps from 0x0 to address
  * 0x400000000.
  */
-#define	TLBTEMP_BASE		_AC(0x0000000001000000,UL)
-#define MODULES_VADDR		_AC(0x0000000002000000,UL)
-#define MODULES_LEN		_AC(0x000000007e000000,UL)
-#define MODULES_END		_AC(0x0000000080000000,UL)
-#define VMALLOC_START		_AC(0x0000000140000000,UL)
-#define VMALLOC_END		_AC(0x0000000200000000,UL)
+#define	TLBTEMP_BASE		_AC(0x0000000002000000,UL)
+#define MODULES_VADDR		_AC(0x0000000010000000,UL)
+#define MODULES_LEN		_AC(0x00000000e0000000,UL)
+#define MODULES_END		_AC(0x00000000f0000000,UL)
 #define LOW_OBP_ADDRESS		_AC(0x00000000f0000000,UL)
 #define HI_OBP_ADDRESS		_AC(0x0000000100000000,UL)
+#define VMALLOC_START		_AC(0x0000000100000000,UL)
+#define VMALLOC_END		_AC(0x0000000200000000,UL)
 
 /* XXX All of this needs to be rethought so we can take advantage
  * XXX cheetah's full 64-bit virtual address space, ie. no more hole
diff --git a/include/linux/byteorder/generic.h b/include/linux/byteorder/generic.h
index 5fde6f4..04bd756 100644
--- a/include/linux/byteorder/generic.h
+++ b/include/linux/byteorder/generic.h
@@ -5,6 +5,10 @@
  * linux/byteorder_generic.h
  * Generic Byte-reordering support
  *
+ * The "... p" macros, like le64_to_cpup, can be used with pointers
+ * to unaligned data, but there will be a performance penalty on 
+ * some architectures.  Use get_unaligned for unaligned data.
+ *
  * Francois-Rene Rideau <fare@tunes.org> 19970707
  *    gathered all the good ideas from all asm-foo/byteorder.h into one file,
  *    cleaned them up.
diff --git a/sound/sparc/cs4231.c b/sound/sparc/cs4231.c
index 2fb27c4..f4361c5 100644
--- a/sound/sparc/cs4231.c
+++ b/sound/sparc/cs4231.c
@@ -173,7 +173,7 @@
 
 #define CS4231_GLOBALIRQ	0x01	/* IRQ is active */
 
-/* definitions for codec irq status */
+/* definitions for codec irq status - CS4231_IRQ_STATUS	*/
 
 #define CS4231_PLAYBACK_IRQ	0x10
 #define CS4231_RECORD_IRQ	0x20
@@ -402,7 +402,7 @@
 	     	udelay(100);
 #ifdef CONFIG_SND_DEBUG
 	if (__cs4231_readb(chip, CS4231P(chip, REGSEL)) & CS4231_INIT)
-		snd_printk("outm: auto calibration time out - reg = 0x%x, value = 0x%x\n", reg, value);
+		snd_printdd("outm: auto calibration time out - reg = 0x%x, value = 0x%x\n", reg, value);
 #endif
 	if (chip->calibrate_mute) {
 		chip->image[reg] &= mask;
@@ -425,6 +425,10 @@
 	     timeout > 0 && (__cs4231_readb(chip, CS4231P(chip, REGSEL)) & CS4231_INIT);
 	     timeout--)
 	     	udelay(100);
+#ifdef CONFIG_SND_DEBUG
+	if (__cs4231_readb(chip, CS4231P(chip, REGSEL)) & CS4231_INIT)
+		snd_printdd("out: auto calibration time out - reg = 0x%x, value = 0x%x\n", reg, value);
+#endif
 	__cs4231_writeb(chip, chip->mce_bit | reg, CS4231P(chip, REGSEL));
 	__cs4231_writeb(chip, value, CS4231P(chip, REG));
 	mb();
@@ -440,15 +444,12 @@
 	     	udelay(100);
 #ifdef CONFIG_SND_DEBUG
 	if (__cs4231_readb(chip, CS4231P(chip, REGSEL)) & CS4231_INIT)
-		snd_printk("out: auto calibration time out - reg = 0x%x, value = 0x%x\n", reg, value);
+		snd_printdd("out: auto calibration time out - reg = 0x%x, value = 0x%x\n", reg, value);
 #endif
 	__cs4231_writeb(chip, chip->mce_bit | reg, CS4231P(chip, REGSEL));
 	__cs4231_writeb(chip, value, CS4231P(chip, REG));
 	chip->image[reg] = value;
 	mb();
-#if 0
-	printk("codec out - reg 0x%x = 0x%x\n", chip->mce_bit | reg, value);
-#endif
 }
 
 static unsigned char snd_cs4231_in(cs4231_t *chip, unsigned char reg)
@@ -462,61 +463,14 @@
 	     	udelay(100);
 #ifdef CONFIG_SND_DEBUG
 	if (__cs4231_readb(chip, CS4231P(chip, REGSEL)) & CS4231_INIT)
-		snd_printk("in: auto calibration time out - reg = 0x%x\n", reg);
+		snd_printdd("in: auto calibration time out - reg = 0x%x\n", reg);
 #endif
 	__cs4231_writeb(chip, chip->mce_bit | reg, CS4231P(chip, REGSEL));
 	mb();
 	ret = __cs4231_readb(chip, CS4231P(chip, REG));
-#if 0
-	printk("codec in - reg 0x%x = 0x%x\n", chip->mce_bit | reg, ret);
-#endif
 	return ret;
 }
 
-#if 0
-
-static void snd_cs4231_debug(cs4231_t *chip)
-{
-	printk("CS4231 REGS:      INDEX = 0x%02x  ",
-	       __cs4231_readb(chip, CS4231P(chip, REGSEL)));
-	printk("                 STATUS = 0x%02x\n",
-	       __cs4231_readb(chip, CS4231P(chip, STATUS)));
-	printk("  0x00: left input      = 0x%02x  ", snd_cs4231_in(chip, 0x00));
-	printk("  0x10: alt 1 (CFIG 2)  = 0x%02x\n", snd_cs4231_in(chip, 0x10));
-	printk("  0x01: right input     = 0x%02x  ", snd_cs4231_in(chip, 0x01));
-	printk("  0x11: alt 2 (CFIG 3)  = 0x%02x\n", snd_cs4231_in(chip, 0x11));
-	printk("  0x02: GF1 left input  = 0x%02x  ", snd_cs4231_in(chip, 0x02));
-	printk("  0x12: left line in    = 0x%02x\n", snd_cs4231_in(chip, 0x12));
-	printk("  0x03: GF1 right input = 0x%02x  ", snd_cs4231_in(chip, 0x03));
-	printk("  0x13: right line in   = 0x%02x\n", snd_cs4231_in(chip, 0x13));
-	printk("  0x04: CD left input   = 0x%02x  ", snd_cs4231_in(chip, 0x04));
-	printk("  0x14: timer low       = 0x%02x\n", snd_cs4231_in(chip, 0x14));
-	printk("  0x05: CD right input  = 0x%02x  ", snd_cs4231_in(chip, 0x05));
-	printk("  0x15: timer high      = 0x%02x\n", snd_cs4231_in(chip, 0x15));
-	printk("  0x06: left output     = 0x%02x  ", snd_cs4231_in(chip, 0x06));
-	printk("  0x16: left MIC (PnP)  = 0x%02x\n", snd_cs4231_in(chip, 0x16));
-	printk("  0x07: right output    = 0x%02x  ", snd_cs4231_in(chip, 0x07));
-	printk("  0x17: right MIC (PnP) = 0x%02x\n", snd_cs4231_in(chip, 0x17));
-	printk("  0x08: playback format = 0x%02x  ", snd_cs4231_in(chip, 0x08));
-	printk("  0x18: IRQ status      = 0x%02x\n", snd_cs4231_in(chip, 0x18));
-	printk("  0x09: iface (CFIG 1)  = 0x%02x  ", snd_cs4231_in(chip, 0x09));
-	printk("  0x19: left line out   = 0x%02x\n", snd_cs4231_in(chip, 0x19));
-	printk("  0x0a: pin control     = 0x%02x  ", snd_cs4231_in(chip, 0x0a));
-	printk("  0x1a: mono control    = 0x%02x\n", snd_cs4231_in(chip, 0x1a));
-	printk("  0x0b: init & status   = 0x%02x  ", snd_cs4231_in(chip, 0x0b));
-	printk("  0x1b: right line out  = 0x%02x\n", snd_cs4231_in(chip, 0x1b));
-	printk("  0x0c: revision & mode = 0x%02x  ", snd_cs4231_in(chip, 0x0c));
-	printk("  0x1c: record format   = 0x%02x\n", snd_cs4231_in(chip, 0x1c));
-	printk("  0x0d: loopback        = 0x%02x  ", snd_cs4231_in(chip, 0x0d));
-	printk("  0x1d: var freq (PnP)  = 0x%02x\n", snd_cs4231_in(chip, 0x1d));
-	printk("  0x0e: ply upr count   = 0x%02x  ", snd_cs4231_in(chip, 0x0e));
-	printk("  0x1e: rec upr count   = 0x%02x\n", snd_cs4231_in(chip, 0x1e));
-	printk("  0x0f: ply lwr count   = 0x%02x  ", snd_cs4231_in(chip, 0x0f));
-	printk("  0x1f: rec lwr count   = 0x%02x\n", snd_cs4231_in(chip, 0x1f));
-}
-
-#endif
-
 /*
  *  CS4231 detection / MCE routines
  */
@@ -528,11 +482,12 @@
 	/* huh.. looks like this sequence is proper for CS4231A chip (GUS MAX) */
 	for (timeout = 5; timeout > 0; timeout--)
 		__cs4231_readb(chip, CS4231P(chip, REGSEL));
+
 	/* end of cleanup sequence */
-	for (timeout = 250;
+	for (timeout = 500;
 	     timeout > 0 && (__cs4231_readb(chip, CS4231P(chip, REGSEL)) & CS4231_INIT);
 	     timeout--)
-	     	udelay(100);
+	     	udelay(1000);
 }
 
 static void snd_cs4231_mce_up(cs4231_t *chip)
@@ -545,12 +500,12 @@
 		udelay(100);
 #ifdef CONFIG_SND_DEBUG
 	if (__cs4231_readb(chip, CS4231P(chip, REGSEL)) & CS4231_INIT)
-		snd_printk("mce_up - auto calibration time out (0)\n");
+		snd_printdd("mce_up - auto calibration time out (0)\n");
 #endif
 	chip->mce_bit |= CS4231_MCE;
 	timeout = __cs4231_readb(chip, CS4231P(chip, REGSEL));
 	if (timeout == 0x80)
-		snd_printk("mce_up [%p]: serious init problem - codec still busy\n", chip->port);
+		snd_printdd("mce_up [%p]: serious init problem - codec still busy\n", chip->port);
 	if (!(timeout & CS4231_MCE))
 		__cs4231_writeb(chip, chip->mce_bit | (timeout & 0x1f), CS4231P(chip, REGSEL));
 	spin_unlock_irqrestore(&chip->lock, flags);
@@ -563,18 +518,15 @@
 
 	spin_lock_irqsave(&chip->lock, flags);
 	snd_cs4231_busy_wait(chip);
-#if 0
-	printk("(1) timeout = %i\n", timeout);
-#endif
 #ifdef CONFIG_SND_DEBUG
 	if (__cs4231_readb(chip, CS4231P(chip, REGSEL)) & CS4231_INIT)
-		snd_printk("mce_down [%p] - auto calibration time out (0)\n", CS4231P(chip, REGSEL));
+		snd_printdd("mce_down [%p] - auto calibration time out (0)\n", CS4231P(chip, REGSEL));
 #endif
 	chip->mce_bit &= ~CS4231_MCE;
 	timeout = __cs4231_readb(chip, CS4231P(chip, REGSEL));
 	__cs4231_writeb(chip, chip->mce_bit | (timeout & 0x1f), CS4231P(chip, REGSEL));
 	if (timeout == 0x80)
-		snd_printk("mce_down [%p]: serious init problem - codec still busy\n", chip->port);
+		snd_printdd("mce_down [%p]: serious init problem - codec still busy\n", chip->port);
 	if ((timeout & CS4231_MCE) == 0) {
 		spin_unlock_irqrestore(&chip->lock, flags);
 		return;
@@ -590,9 +542,7 @@
 		spin_unlock_irqrestore(&chip->lock, flags);
 		return;
 	}
-#if 0
-	printk("(2) timeout = %i, jiffies = %li\n", timeout, jiffies);
-#endif
+
 	/* in 10ms increments, check condition, up to 250ms */
 	timeout = 25;
 	while (snd_cs4231_in(chip, CS4231_TEST_INIT) & CS4231_CALIB_IN_PROGRESS) {
@@ -604,9 +554,7 @@
 		msleep(10);
 		spin_lock_irqsave(&chip->lock, flags);
 	}
-#if 0
-	printk("(3) jiffies = %li\n", jiffies);
-#endif
+
 	/* in 10ms increments, check condition, up to 100ms */
 	timeout = 10;
 	while (__cs4231_readb(chip, CS4231P(chip, REGSEL)) & CS4231_INIT) {
@@ -619,54 +567,58 @@
 		spin_lock_irqsave(&chip->lock, flags);
 	}
 	spin_unlock_irqrestore(&chip->lock, flags);
-#if 0
-	printk("(4) jiffies = %li\n", jiffies);
-	snd_printk("mce_down - exit = 0x%x\n", __cs4231_readb(chip, CS4231P(chip, REGSEL)));
-#endif
 }
 
-#if 0 /* Unused for now... */
-static unsigned int snd_cs4231_get_count(unsigned char format, unsigned int size)
-{
-	switch (format & 0xe0) {
-	case CS4231_LINEAR_16:
-	case CS4231_LINEAR_16_BIG:
-		size >>= 1;
-		break;
-	case CS4231_ADPCM_16:
-		return size >> 2;
-	}
-	if (format & CS4231_STEREO)
-		size >>= 1;
-	return size;
-}
-#endif
-
 #ifdef EBUS_SUPPORT
 static void snd_cs4231_ebus_advance_dma(struct ebus_dma_info *p, snd_pcm_substream_t *substream, unsigned int *periods_sent)
 {
 	snd_pcm_runtime_t *runtime = substream->runtime;
 
 	while (1) {
-		unsigned int dma_size = snd_pcm_lib_period_bytes(substream);
-		unsigned int offset = dma_size * (*periods_sent);
+		unsigned int period_size = snd_pcm_lib_period_bytes(substream);
+		unsigned int offset = period_size * (*periods_sent);
 
-		if (dma_size >= (1 << 24))
+		if (period_size >= (1 << 24))
 			BUG();
 
-		if (ebus_dma_request(p, runtime->dma_addr + offset, dma_size))
+		if (ebus_dma_request(p, runtime->dma_addr + offset, period_size))
 			return;
-#if 0
-		printk("ebus_advance: Sent period %u (size[%x] offset[%x])\n",
-		       (*periods_sent), dma_size, offset);
-#endif
 		(*periods_sent) = ((*periods_sent) + 1) % runtime->periods;
 	}
 }
 #endif
 
-static void cs4231_dma_trigger(cs4231_t *chip, unsigned int what, int on)
+#ifdef SBUS_SUPPORT
+static void snd_cs4231_sbus_advance_dma(snd_pcm_substream_t *substream, unsigned int *periods_sent)
 {
+	cs4231_t *chip = snd_pcm_substream_chip(substream);
+	snd_pcm_runtime_t *runtime = substream->runtime;
+
+	unsigned int period_size = snd_pcm_lib_period_bytes(substream);
+	unsigned int offset = period_size * (*periods_sent % runtime->periods);
+
+	if (runtime->period_size > 0xffff + 1)
+		BUG();
+
+	switch (substream->stream) {
+	case SNDRV_PCM_STREAM_PLAYBACK:
+		sbus_writel(runtime->dma_addr + offset, chip->port + APCPNVA);
+		sbus_writel(period_size, chip->port + APCPNC);
+		break;
+	case SNDRV_PCM_STREAM_CAPTURE:
+		sbus_writel(runtime->dma_addr + offset, chip->port + APCCNVA);
+		sbus_writel(period_size, chip->port + APCCNC);
+		break;
+	}
+
+	(*periods_sent) = (*periods_sent + 1) % runtime->periods;
+}
+#endif
+
+static void cs4231_dma_trigger(snd_pcm_substream_t *substream, unsigned int what, int on)
+{
+	cs4231_t *chip = snd_pcm_substream_chip(substream);
+
 #ifdef EBUS_SUPPORT
 	if (chip->flags & CS4231_FLAG_EBUS) {
 		if (what & CS4231_PLAYBACK_ENABLE) {
@@ -694,6 +646,60 @@
 	} else {
 #endif
 #ifdef SBUS_SUPPORT
+	u32 csr = sbus_readl(chip->port + APCCSR);
+	/* I don't know why, but on sbus the period counter must
+	 * only start counting after the first period is sent.
+	 * Therefore this dummy thing.
+	 */
+	unsigned int dummy = 0;
+
+	switch (what) {
+	case CS4231_PLAYBACK_ENABLE:
+		if (on) {
+			csr &= ~APC_XINT_PLAY;
+			sbus_writel(csr, chip->port + APCCSR);
+
+			csr &= ~APC_PPAUSE;
+			sbus_writel(csr, chip->port + APCCSR);
+
+			snd_cs4231_sbus_advance_dma(substream, &dummy);
+
+			csr |=  APC_GENL_INT | APC_PLAY_INT | APC_XINT_ENA |
+				APC_XINT_PLAY | APC_XINT_EMPT | APC_XINT_GENL |
+				APC_XINT_PENA | APC_PDMA_READY;
+			sbus_writel(csr, chip->port + APCCSR);
+		} else {
+			csr |= APC_PPAUSE;
+			sbus_writel(csr, chip->port + APCCSR);
+
+			csr &= ~APC_PDMA_READY;
+			sbus_writel(csr, chip->port + APCCSR);
+		}
+		break;
+	case CS4231_RECORD_ENABLE:
+		if (on) {
+			csr &= ~APC_XINT_CAPT;
+			sbus_writel(csr, chip->port + APCCSR);
+
+			csr &= ~APC_CPAUSE;
+			sbus_writel(csr, chip->port + APCCSR);
+
+			snd_cs4231_sbus_advance_dma(substream, &dummy);
+
+			csr |=  APC_GENL_INT | APC_CAPT_INT | APC_XINT_ENA |
+				APC_XINT_CAPT | APC_XINT_CEMP | APC_XINT_GENL |
+				APC_CDMA_READY;
+
+			sbus_writel(csr, chip->port + APCCSR);
+		} else {
+			csr |= APC_CPAUSE;
+			sbus_writel(csr, chip->port + APCCSR);
+
+			csr &= ~APC_CDMA_READY;
+			sbus_writel(csr, chip->port + APCCSR);
+		}
+		break;
+	}
 #endif
 #ifdef EBUS_SUPPORT
 	}
@@ -725,25 +731,12 @@
 			}
 		}
 
-#if 0
-		printk("TRIGGER: what[%x] on(%d)\n",
-		       what, (cmd == SNDRV_PCM_TRIGGER_START));
-#endif
-
 		spin_lock_irqsave(&chip->lock, flags);
 		if (cmd == SNDRV_PCM_TRIGGER_START) {
-			cs4231_dma_trigger(chip, what, 1);
+			cs4231_dma_trigger(substream, what, 1);
 			chip->image[CS4231_IFACE_CTRL] |= what;
-			if (what & CS4231_PLAYBACK_ENABLE) {
-				snd_cs4231_out(chip, CS4231_PLY_LWR_CNT, 0xff);
-				snd_cs4231_out(chip, CS4231_PLY_UPR_CNT, 0xff);
-			}
-			if (what & CS4231_RECORD_ENABLE) {
-				snd_cs4231_out(chip, CS4231_REC_LWR_CNT, 0xff);
-				snd_cs4231_out(chip, CS4231_REC_UPR_CNT, 0xff);
-			}
 		} else {
-			cs4231_dma_trigger(chip, what, 0);
+			cs4231_dma_trigger(substream, what, 0);
 			chip->image[CS4231_IFACE_CTRL] &= ~what;
 		}
 		snd_cs4231_out(chip, CS4231_IFACE_CTRL,
@@ -755,9 +748,7 @@
 		result = -EINVAL;
 		break;
 	}
-#if 0
-	snd_cs4231_debug(chip);
-#endif
+
 	return result;
 }
 
@@ -790,9 +781,6 @@
 	}
 	if (channels > 1)
 		rformat |= CS4231_STEREO;
-#if 0
-	snd_printk("get_format: 0x%x (mode=0x%x)\n", format, mode);
-#endif
 	return rformat;
 }
 
@@ -944,7 +932,7 @@
 	snd_cs4231_mce_down(chip);
 
 #ifdef SNDRV_DEBUG_MCE
-	snd_printk("init: (1)\n");
+	snd_printdd("init: (1)\n");
 #endif
 	snd_cs4231_mce_up(chip);
 	spin_lock_irqsave(&chip->lock, flags);
@@ -957,7 +945,7 @@
 	snd_cs4231_mce_down(chip);
 
 #ifdef SNDRV_DEBUG_MCE
-	snd_printk("init: (2)\n");
+	snd_printdd("init: (2)\n");
 #endif
 
 	snd_cs4231_mce_up(chip);
@@ -967,7 +955,7 @@
 	snd_cs4231_mce_down(chip);
 
 #ifdef SNDRV_DEBUG_MCE
-	snd_printk("init: (3) - afei = 0x%x\n", chip->image[CS4231_ALT_FEATURE_1]);
+	snd_printdd("init: (3) - afei = 0x%x\n", chip->image[CS4231_ALT_FEATURE_1]);
 #endif
 
 	spin_lock_irqsave(&chip->lock, flags);
@@ -981,7 +969,7 @@
 	snd_cs4231_mce_down(chip);
 
 #ifdef SNDRV_DEBUG_MCE
-	snd_printk("init: (4)\n");
+	snd_printdd("init: (4)\n");
 #endif
 
 	snd_cs4231_mce_up(chip);
@@ -991,7 +979,7 @@
 	snd_cs4231_mce_down(chip);
 
 #ifdef SNDRV_DEBUG_MCE
-	snd_printk("init: (5)\n");
+	snd_printdd("init: (5)\n");
 #endif
 }
 
@@ -1022,6 +1010,7 @@
 		       CS4231_RECORD_IRQ |
 		       CS4231_TIMER_IRQ);
 	snd_cs4231_out(chip, CS4231_IRQ_STATUS, 0);
+
 	spin_unlock_irqrestore(&chip->lock, flags);
 
 	chip->mode = mode;
@@ -1136,11 +1125,21 @@
 static int snd_cs4231_playback_prepare(snd_pcm_substream_t *substream)
 {
 	cs4231_t *chip = snd_pcm_substream_chip(substream);
+	snd_pcm_runtime_t *runtime = substream->runtime;
 	unsigned long flags;
 
 	spin_lock_irqsave(&chip->lock, flags);
+
 	chip->image[CS4231_IFACE_CTRL] &= ~(CS4231_PLAYBACK_ENABLE |
 					    CS4231_PLAYBACK_PIO);
+
+	if (runtime->period_size > 0xffff + 1)
+		BUG();
+
+	snd_cs4231_out(chip, CS4231_PLY_LWR_CNT, (runtime->period_size - 1) & 0x00ff);
+	snd_cs4231_out(chip, CS4231_PLY_UPR_CNT, (runtime->period_size - 1) >> 8 & 0x00ff);
+	chip->p_periods_sent = 0;
+
 	spin_unlock_irqrestore(&chip->lock, flags);
 
 	return 0;
@@ -1172,12 +1171,16 @@
 static int snd_cs4231_capture_prepare(snd_pcm_substream_t *substream)
 {
 	cs4231_t *chip = snd_pcm_substream_chip(substream);
+	snd_pcm_runtime_t *runtime = substream->runtime;
 	unsigned long flags;
 
 	spin_lock_irqsave(&chip->lock, flags);
 	chip->image[CS4231_IFACE_CTRL] &= ~(CS4231_RECORD_ENABLE |
 					    CS4231_RECORD_PIO);
 
+	snd_cs4231_out(chip, CS4231_REC_LWR_CNT, (runtime->period_size - 1) & 0x00ff);
+	snd_cs4231_out(chip, CS4231_REC_LWR_CNT, (runtime->period_size - 1) >> 8 & 0x00ff);
+
 	spin_unlock_irqrestore(&chip->lock, flags);
 
 	return 0;
@@ -1196,53 +1199,61 @@
 		chip->capture_substream->runtime->overrange++;
 }
 
-static void snd_cs4231_generic_interrupt(cs4231_t *chip)
+static irqreturn_t snd_cs4231_generic_interrupt(cs4231_t *chip)
 {
 	unsigned long flags;
 	unsigned char status;
 
+	/*This is IRQ is not raised by the cs4231*/
+	if (!(__cs4231_readb(chip, CS4231P(chip, STATUS)) & CS4231_GLOBALIRQ))
+		return IRQ_NONE;
+
 	status = snd_cs4231_in(chip, CS4231_IRQ_STATUS);
-	if (!status)
-		return;
 
 	if (status & CS4231_TIMER_IRQ) {
 		if (chip->timer)
 			snd_timer_interrupt(chip->timer, chip->timer->sticks);
 	}		
-	if (status & CS4231_PLAYBACK_IRQ)
-		snd_pcm_period_elapsed(chip->playback_substream);
-	if (status & CS4231_RECORD_IRQ) {
+
+	if (status & CS4231_RECORD_IRQ)
 		snd_cs4231_overrange(chip);
-		snd_pcm_period_elapsed(chip->capture_substream);
-	}
 
 	/* ACK the CS4231 interrupt. */
 	spin_lock_irqsave(&chip->lock, flags);
 	snd_cs4231_outm(chip, CS4231_IRQ_STATUS, ~CS4231_ALL_IRQS | ~status, 0);
 	spin_unlock_irqrestore(&chip->lock, flags);
+
+	return 0;
 }
 
 #ifdef SBUS_SUPPORT
 static irqreturn_t snd_cs4231_sbus_interrupt(int irq, void *dev_id, struct pt_regs *regs)
 {
 	cs4231_t *chip = dev_id;
-	u32 csr;
-
-	csr = sbus_readl(chip->port + APCCSR);
-	if (!(csr & (APC_INT_PENDING |
-		     APC_PLAY_INT |
-		     APC_CAPT_INT |
-		     APC_GENL_INT |
-		     APC_XINT_PEMP |
-		     APC_XINT_CEMP)))
-		return IRQ_NONE;
 
 	/* ACK the APC interrupt. */
+	u32 csr = sbus_readl(chip->port + APCCSR);
+
 	sbus_writel(csr, chip->port + APCCSR);
 
-	snd_cs4231_generic_interrupt(chip);
+	if ((chip->image[CS4231_IFACE_CTRL] & CS4231_PLAYBACK_ENABLE) &&
+	    (csr & APC_PLAY_INT) &&
+	    (csr & APC_XINT_PNVA) &&
+	    !(csr & APC_XINT_EMPT)) {
+		snd_cs4231_sbus_advance_dma(chip->playback_substream,
+					    &chip->p_periods_sent);
+		snd_pcm_period_elapsed(chip->playback_substream);
+	}
 
-	return IRQ_HANDLED;
+	if ((chip->image[CS4231_IFACE_CTRL] & CS4231_RECORD_ENABLE) &&
+	    (csr & APC_CAPT_INT) &&
+	    (csr & APC_XINT_CNVA)) {
+		snd_cs4231_sbus_advance_dma(chip->capture_substream,
+					    &chip->c_periods_sent);
+		snd_pcm_period_elapsed(chip->capture_substream);
+	}
+
+	return snd_cs4231_generic_interrupt(chip);
 }
 #endif
 
@@ -1290,7 +1301,8 @@
 #ifdef EBUS_SUPPORT
 	}
 #endif
-	ptr += (period_bytes - residue);
+	ptr += period_bytes - residue;
+
 	return bytes_to_frames(substream->runtime, ptr);
 }
 
@@ -1314,7 +1326,7 @@
 #ifdef EBUS_SUPPORT
 	}
 #endif
-	ptr += (period_bytes - residue);
+	ptr += period_bytes - residue;
 	return bytes_to_frames(substream->runtime, ptr);
 }
 
@@ -1328,9 +1340,6 @@
 	int i, id, vers;
 	unsigned char *ptr;
 
-#if 0
-	snd_cs4231_debug(chip);
-#endif
 	id = vers = 0;
 	for (i = 0; i < 50; i++) {
 		mb();
@@ -1985,13 +1994,13 @@
 	chip->port = sbus_ioremap(&sdev->resource[0], 0,
 				  chip->regs_size, "cs4231");
 	if (!chip->port) {
-		snd_printk("cs4231-%d: Unable to map chip registers.\n", dev);
+		snd_printdd("cs4231-%d: Unable to map chip registers.\n", dev);
 		return -EIO;
 	}
 
 	if (request_irq(sdev->irqs[0], snd_cs4231_sbus_interrupt,
 			SA_SHIRQ, "cs4231", chip)) {
-		snd_printk("cs4231-%d: Unable to grab SBUS IRQ %s\n",
+		snd_printdd("cs4231-%d: Unable to grab SBUS IRQ %s\n",
 			   dev,
 			   __irq_itoa(sdev->irqs[0]));
 		snd_cs4231_sbus_free(chip);
@@ -2113,29 +2122,29 @@
 	chip->eb2c.regs = ioremap(edev->resource[2].start, 0x10);
 	if (!chip->port || !chip->eb2p.regs || !chip->eb2c.regs) {
 		snd_cs4231_ebus_free(chip);
-		snd_printk("cs4231-%d: Unable to map chip registers.\n", dev);
+		snd_printdd("cs4231-%d: Unable to map chip registers.\n", dev);
 		return -EIO;
 	}
 
 	if (ebus_dma_register(&chip->eb2c)) {
 		snd_cs4231_ebus_free(chip);
-		snd_printk("cs4231-%d: Unable to register EBUS capture DMA\n", dev);
+		snd_printdd("cs4231-%d: Unable to register EBUS capture DMA\n", dev);
 		return -EBUSY;
 	}
 	if (ebus_dma_irq_enable(&chip->eb2c, 1)) {
 		snd_cs4231_ebus_free(chip);
-		snd_printk("cs4231-%d: Unable to enable EBUS capture IRQ\n", dev);
+		snd_printdd("cs4231-%d: Unable to enable EBUS capture IRQ\n", dev);
 		return -EBUSY;
 	}
 
 	if (ebus_dma_register(&chip->eb2p)) {
 		snd_cs4231_ebus_free(chip);
-		snd_printk("cs4231-%d: Unable to register EBUS play DMA\n", dev);
+		snd_printdd("cs4231-%d: Unable to register EBUS play DMA\n", dev);
 		return -EBUSY;
 	}
 	if (ebus_dma_irq_enable(&chip->eb2p, 1)) {
 		snd_cs4231_ebus_free(chip);
-		snd_printk("cs4231-%d: Unable to enable EBUS play IRQ\n", dev);
+		snd_printdd("cs4231-%d: Unable to enable EBUS play IRQ\n", dev);
 		return -EBUSY;
 	}