Re: [PATCH 1/2] Adds full stack to critical-section tracing

[Daniel Walker <dwalker@mvista.com>]

On Fri, 2007-08-10 at 08:14 -0600, Gregory Haskins wrote:
> ---
> 
>  include/linux/sched.h  |    7 +++++--
>  kernel/latency_trace.c |   18 +++++++++++-------
>  2 files changed, 16 insertions(+), 9 deletions(-)
> 
> diff --git a/include/linux/sched.h b/include/linux/sched.h
> index 8ebb43c..233d26c 100644
> --- a/include/linux/sched.h
> +++ b/include/linux/sched.h
> @@ -1306,10 +1306,13 @@ struct task_struct {
>  #endif
>  
>  #define MAX_PREEMPT_TRACE 25
> +#define MAX_PREEMPT_TRACE_DEPTH 5
>  
>  #ifdef CONFIG_PREEMPT_TRACE
> -	unsigned long preempt_trace_eip[MAX_PREEMPT_TRACE];
> -	unsigned long preempt_trace_parent_eip[MAX_PREEMPT_TRACE];
> +	struct {
> +		struct stack_trace trace;
> +		unsigned long      data[MAX_PREEMPT_TRACE_DEPTH];
> +	} preempt_trace[MAX_PREEMPT_TRACE];

These changes need some fixing .. The "struct stack_trace"
infrastructure is based on CONFIG_STACKTRACE , so at least you would
want to ifdef your changes on that config options ..

I modified you patch to include the ifdefs,

ftp://source.mvista.com/pub/dwalker/misc/debug-preempt-5-levels-of-stack-frames.patch

The other problem I noticed is that the save_stack_trace(trace); seems
kind of heavy weight to be calling from inside add_preempt_count. If
your only going to a depth of 5 you could potentially use there macros,

#ifdef CONFIG_FRAME_POINTER
# ifndef CONFIG_ARM
#  define CALLER_ADDR0 ((unsigned long)__builtin_return_address(0))
#  define CALLER_ADDR1 ((unsigned long)__builtin_return_address(1))
#  define CALLER_ADDR2 ((unsigned long)__builtin_return_address(2))
#  define CALLER_ADDR3 ((unsigned long)__builtin_return_address(3))
#  define CALLER_ADDR4 ((unsigned long)__builtin_return_address(4))
#  define CALLER_ADDR5 ((unsigned long)__builtin_return_address(5))
# else

Daniel