Blackfin arch: Add ability to expend the hardware trace buffer

Add ability to expend the hardware trace buffer via a configurable
software buffer - so you can have lots of history when a crash occurs.

The interesting way we do printk in the traps.c confusese the checking
script

Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>

8 files changed, 213 insertions, 11 deletions

diff --git a/arch/blackfin/Kconfig b/arch/blackfin/Kconfig
index 9ce675e80260..a7a6e0c5827d 100644
--- a/arch/blackfin/Kconfig
+++ b/arch/blackfin/Kconfig
@@ -1024,8 +1024,89 @@ config DEBUG_HUNT_FOR_ZERO
 	  Enabling this option will take up an extra entry in CPLB table.
 	  Otherwise, there is no extra overhead.
 
+config DEBUG_BFIN_HWTRACE_ON
+	bool "Turn on Blackfin's Hardware Trace"
+	default y
+	help
+	  All Blackfins include a Trace Unit which stores a history of the last
+	  16 changes in program flow taken by the program sequencer. The history
+	  allows the user to recreate the program sequencer’s recent path. This
+	  can be handy when an application dies - we print out the execution
+	  path of how it got to the offending instruction.
+
+	  By turning this off, you may save a tiny amount of power.
+
+choice
+	prompt "Omit loop Tracing"
+	default DEBUG_BFIN_HWTRACE_COMPRESSION_OFF
+	depends on DEBUG_BFIN_HWTRACE_ON
+	help
+	  The trace buffer can be configured to omit recording of changes in
+	  program flow that match either the last entry or one of the last
+	  two entries. Omitting one of these entries from the record prevents
+	  the trace buffer from overflowing because of any sort of loop (for, do
+	  while, etc) in the program.
+
+	  Because zero-overhead Hardware loops are not recorded in the trace buffer,
+	  this feature can be used to prevent trace overflow from loops that
+	  are nested four deep.
+
+config DEBUG_BFIN_HWTRACE_COMPRESSION_OFF
+	bool "Trace all Loops"
+	help
+	  The trace buffer records all changes of flow 
+
+config DEBUG_BFIN_HWTRACE_COMPRESSION_ONE
+	bool "Compress single-level loops"
+	help
+	  The trace buffer does not record single loops - helpful if trace 
+	  is spinning on a while or do loop.
+
+config DEBUG_BFIN_HWTRACE_COMPRESSION_TWO
+	bool "Compress two-level loops"
+	help
+	  The trace buffer does not record loops two levels deep. Helpful if
+	  the trace is spinning in a nested loop
+
+endchoice
+
+config DEBUG_BFIN_HWTRACE_COMPRESSION
+	int
+	depends on DEBUG_BFIN_HWTRACE_ON
+	default 0 if DEBUG_BFIN_HWTRACE_COMPRESSION_OFF
+	default 1 if DEBUG_BFIN_HWTRACE_COMPRESSION_ONE
+	default 2 if DEBUG_BFIN_HWTRACE_COMPRESSION_TWO
+
+
+config DEBUG_BFIN_HWTRACE_EXPAND
+	bool "Expand Trace Buffer greater than 16 entries"
+	depends on DEBUG_BFIN_HWTRACE_ON
+	default n
+	help
+	  By selecting this option, every time the 16 hardware entries in
+	  the Blackfin's HW Trace buffer are full, the kernel will move them
+	  into a software buffer, for dumping when there is an issue. This 
+	  has a great impact on performance, (an interrupt every 16 change of 
+	  flows) and should normally be turned off, except in those nasty
+	  debugging sessions
+
+config DEBUG_BFIN_HWTRACE_EXPAND_LEN
+	int "Size of Trace buffer (in power of 2k)"
+	range 0 4
+	depends on DEBUG_BFIN_HWTRACE_EXPAND
+	default 1
+	help
+	  This sets the size of the software buffer that the trace information
+	  is kept in.
+	  0 for (2^0)  1k, or 256 entries,
+	  1 for (2^1)  2k, or 512 entries,
+	  2 for (2^2)  4k, or 1024 entries,
+	  3 for (2^3)  8k, or 2048 entries,
+	  4 for (2^4) 16k, or 4096 entries
+
 config DEBUG_BFIN_NO_KERN_HWTRACE
 	bool "Trace user apps (turn off hwtrace in kernel)"
+	depends on DEBUG_BFIN_HWTRACE_ON
 	default n
 	help
 	  Some pieces of the kernel contain a lot of flow changes which can
diff --git a/arch/blackfin/kernel/irqchip.c b/arch/blackfin/kernel/irqchip.c
index 1fc001c7abda..462ae41144c7 100644
--- a/arch/blackfin/kernel/irqchip.c
+++ b/arch/blackfin/kernel/irqchip.c
@@ -34,6 +34,7 @@
 #include <linux/kallsyms.h>
 #include <linux/interrupt.h>
 #include <linux/irq.h>
+#include <asm/trace.h>
 
 static unsigned long irq_err_count;
 static spinlock_t irq_controller_lock;
@@ -144,4 +145,12 @@ void __init init_IRQ(void)
 	}
 
 	init_arch_irq();
+
+#ifdef CONFIG_DEBUG_BFIN_HWTRACE_EXPAND
+	/* Now that evt_ivhw is set up, turn this on */
+	trace_buff_offset = 0;
+	bfin_write_TBUFCTL(BFIN_TRACE_ON);
+	printk(KERN_INFO "Hardware Trace expanded to %ik\n",
+	  1 << CONFIG_DEBUG_BFIN_HWTRACE_EXPAND_LEN);
+#endif
 }
diff --git a/arch/blackfin/kernel/traps.c b/arch/blackfin/kernel/traps.c
index 792a8416fe10..0ec02fe663e9 100644
--- a/arch/blackfin/kernel/traps.c
+++ b/arch/blackfin/kernel/traps.c
@@ -55,6 +55,7 @@ asmlinkage void trap_c(struct pt_regs *fp);
 
 int kstack_depth_to_print = 48;
 
+#ifdef CONFIG_DEBUG_BFIN_HWTRACE_ON
 static int printk_address(unsigned long address)
 {
 	struct vm_list_struct *vml;
@@ -131,10 +132,14 @@ static int printk_address(unsigned long address)
 	/* we were unable to find this address anywhere */
 	return printk("[<0x%p>]", (void *)address);
 }
+#endif
 
 asmlinkage void trap_c(struct pt_regs *fp)
 {
-	int j, sig = 0;
+#ifdef CONFIG_DEBUG_BFIN_HWTRACE_ON
+	int j;
+#endif
+	int sig = 0;
 	siginfo_t info;
 	unsigned long trapnr = fp->seqstat & SEQSTAT_EXCAUSE;
 
@@ -429,24 +434,56 @@ asmlinkage void trap_c(struct pt_regs *fp)
 
 /* Typical exception handling routines	*/
 
+#define EXPAND_LEN ((1 << CONFIG_DEBUG_BFIN_HWTRACE_EXPAND_LEN) * 256 - 1)
+
 void dump_bfin_trace_buffer(void)
 {
-	int tflags;
+#ifdef CONFIG_DEBUG_BFIN_HWTRACE_ON
+	int tflags, i = 0;
+#ifdef CONFIG_DEBUG_BFIN_HWTRACE_EXPAND
+	int j, index;
+#endif
+
 	trace_buffer_save(tflags);
 
+	printk(KERN_EMERG "Hardware Trace:\n");
+
 	if (likely(bfin_read_TBUFSTAT() & TBUFCNT)) {
-		int i;
-		printk(KERN_EMERG "Hardware Trace:\n");
-		for (i = 0; bfin_read_TBUFSTAT() & TBUFCNT; i++) {
-			printk(KERN_EMERG "%2i Target : ", i);
+		for (; bfin_read_TBUFSTAT() & TBUFCNT; i++) {
+			printk(KERN_EMERG "%4i Target : ", i);
 			printk_address((unsigned long)bfin_read_TBUF());
-			printk("\n" KERN_EMERG "   Source : ");
+			printk("\n" KERN_EMERG "     Source : ");
 			printk_address((unsigned long)bfin_read_TBUF());
 			printk("\n");
 		}
 	}
 
+#ifdef CONFIG_DEBUG_BFIN_HWTRACE_EXPAND
+	if (trace_buff_offset)
+		index = trace_buff_offset/4 - 1;
+	else
+		index = EXPAND_LEN;
+
+	j = (1 << CONFIG_DEBUG_BFIN_HWTRACE_EXPAND_LEN) * 128;
+	while (j) {
+		printk(KERN_EMERG "%4i Target : ", i);
+		printk_address(software_trace_buff[index]);
+		index -= 1;
+		if (index < 0 )
+			index = EXPAND_LEN;
+		printk("\n" KERN_EMERG "     Source : ");
+		printk_address(software_trace_buff[index]);
+		index -= 1;
+		if (index < 0)
+			index = EXPAND_LEN;
+		printk("\n");
+		j--;
+		i++;
+	}
+#endif
+
 	trace_buffer_restore(tflags);
+#endif
 }
 EXPORT_SYMBOL(dump_bfin_trace_buffer);
 
@@ -510,7 +547,9 @@ void show_stack(struct task_struct *task, unsigned long *stack)
 void dump_stack(void)
 {
 	unsigned long stack;
+#ifdef CONFIG_DEBUG_BFIN_HWTRACE_ON
 	int tflags;
+#endif
 	trace_buffer_save(tflags);
 	dump_bfin_trace_buffer();
 	show_stack(current, &stack);
diff --git a/arch/blackfin/mach-bf533/head.S b/arch/blackfin/mach-bf533/head.S
index 9c5378bb8a28..1d5b9dbbbaa7 100644
--- a/arch/blackfin/mach-bf533/head.S
+++ b/arch/blackfin/mach-bf533/head.S
@@ -98,7 +98,7 @@ ENTRY(__start)
 	M2 = r0;
 	M3 = r0;
 
-	trace_buffer_start(p0,r0);
+	trace_buffer_init(p0,r0);
 	P0 = R1;
 	R0 = R1;
 
diff --git a/arch/blackfin/mach-bf537/head.S b/arch/blackfin/mach-bf537/head.S
index 82ea04705917..6dbcb77c8d36 100644
--- a/arch/blackfin/mach-bf537/head.S
+++ b/arch/blackfin/mach-bf537/head.S
@@ -96,7 +96,7 @@ ENTRY(__start)
 	M2 = r0;
 	M3 = r0;
 
-	trace_buffer_start(p0,r0);
+	trace_buffer_init(p0,r0);
 	P0 = R1;
 	R0 = R1;
 
diff --git a/arch/blackfin/mach-bf548/head.S b/arch/blackfin/mach-bf548/head.S
index 72087c2e6a83..e53d74d4c0a2 100644
--- a/arch/blackfin/mach-bf548/head.S
+++ b/arch/blackfin/mach-bf548/head.S
@@ -93,7 +93,7 @@ ENTRY(__stext)
 	M2 = r0;
 	M3 = r0;
 
-	trace_buffer_start(p0,r0);
+	trace_buffer_init(p0,r0);
 	P0 = R1;
 	R0 = R1;
 
diff --git a/arch/blackfin/mach-bf561/head.S b/arch/blackfin/mach-bf561/head.S
index 83cd3f9bbf2d..8c9f73b8e579 100644
--- a/arch/blackfin/mach-bf561/head.S
+++ b/arch/blackfin/mach-bf561/head.S
@@ -96,7 +96,7 @@ ENTRY(__start)
 	M2 = r0;
 	M3 = r0;
 
-	trace_buffer_start(p0,r0);
+	trace_buffer_init(p0,r0);
 	P0 = R1;
 	R0 = R1;
 
diff --git a/arch/blackfin/mach-common/entry.S b/arch/blackfin/mach-common/entry.S
index 207e69786b40..ab278a72f282 100644
--- a/arch/blackfin/mach-common/entry.S
+++ b/arch/blackfin/mach-common/entry.S
@@ -731,6 +731,75 @@ ENTRY(_init_exception_buff)
 	rts;
 ENDPROC(_init_exception_buff)
 
+/* We handle this 100% in exception space - to reduce overhead
+ * Only potiential problem is if the software buffer gets swapped out of the
+ * CPLB table - then double fault. - so we don't let this happen in other places
+ */
+#ifdef CONFIG_DEBUG_BFIN_HWTRACE_EXPAND
+ENTRY(_ex_trace_buff_full)
+	[--sp] = P3;
+	[--sp] = P2;
+	[--sp] = LC0;
+	[--sp] = LT0;
+	[--sp] = LB0;
+	P5.L = _trace_buff_offset;
+	P5.H = _trace_buff_offset;
+	P3 = [P5];              /* trace_buff_offset */
+	P5.L = lo(TBUFSTAT);
+	P5.H = hi(TBUFSTAT);
+	R7 = [P5];
+	R7 <<= 1;               /* double, since we need to read twice */
+	LC0 = R7;
+	R7 <<= 2;               /* need to shift over again,
+				 * to get the number of bytes */
+	P5.L = lo(TBUF);
+	P5.H = hi(TBUF);
+	R6 = ((1 << CONFIG_DEBUG_BFIN_HWTRACE_EXPAND_LEN)*1024) - 1;
+
+	P2 = R7;
+	P3 = P3 + P2;
+	R7 = P3;
+	R7 = R7 & R6;
+	P3 = R7;
+	P2.L = _trace_buff_offset;
+	P2.H = _trace_buff_offset;
+	[P2] = P3;
+
+	P2.L = _software_trace_buff;
+	P2.H = _software_trace_buff;
+
+	LSETUP (.Lstart, .Lend) LC0;
+.Lstart:
+	R7 = [P5];      /* read TBUF */
+	P4 = P3 + P2;
+	[P4] = R7;
+	P3 += -4;
+	R7 = P3;
+	R7 = R7 & R6;
+.Lend:
+	P3 = R7;
+
+	LB0 = [sp++];
+	LT0 = [sp++];
+	LC0 = [sp++];
+	P2 = [sp++];
+	P3 = [sp++];
+	jump _return_from_exception;
+
+#if CONFIG_DEBUG_BFIN_HWTRACE_EXPAND_LEN == 4
+.data
+#else
+.section .l1.data.B
+#endif
+ENTRY(_trace_buff_offset)
+        .long 0;
+ALIGN
+ENTRY(_software_trace_buff)
+	.rept ((1 << CONFIG_DEBUG_BFIN_HWTRACE_EXPAND_LEN)*256);
+	.long 0
+	.endr
+#endif
+
 /*
  * Put these in the kernel data section - that should always be covered by
  * a CPLB. This is needed to ensure we don't get double fault conditions
@@ -764,7 +833,11 @@ _extable:
 	.long _ex_trap_c        /* 0x0E - User Defined */
 	.long _ex_trap_c        /* 0x0F - User Defined */
 	.long _ex_single_step   /* 0x10 - HW Single step */
+#ifdef CONFIG_DEBUG_BFIN_HWTRACE_EXPAND
+	.long _ex_trace_buff_full /* 0x11 - Trace Buffer Full */
+#else
 	.long _ex_trap_c        /* 0x11 - Trace Buffer Full */
+#endif
 	.long _ex_trap_c        /* 0x12 - Reserved */
 	.long _ex_trap_c        /* 0x13 - Reserved */
 	.long _ex_trap_c        /* 0x14 - Reserved */