Re: [Patch 1/2] ksym_tracer:Fix line-wrapping after removal of ksym tracer entry

[Frederic Weisbecker <fweisbec@gmail.com>]

On Wed, Jun 17, 2009 at 04:35:04AM +0530, K.Prasad wrote:
> Removal of a ksym entry results in missing information and an early return with
> TRACE_TYPE_PARTIAL_LINE code (minus the line return). This patch modifies the
> output function to unconditionally add a line return irrespective of the
> return code. 
> 
> Signed-off-by: K.Prasad <prasad@linux.vnet.ibm.com>
> ---
>  kernel/trace/trace_ksym.c |   27 +++++++++++++++------------
>  1 file changed, 15 insertions(+), 12 deletions(-)
> 
> Index: linux-2.6-tip.hbkpt/kernel/trace/trace_ksym.c
> ===================================================================
> --- linux-2.6-tip.hbkpt.orig/kernel/trace/trace_ksym.c
> +++ linux-2.6-tip.hbkpt/kernel/trace/trace_ksym.c
> @@ -391,29 +391,32 @@ static enum print_line_t ksym_trace_outp
>  
>  	ret = trace_seq_printf(s, "%-15s %-5d %-3d %-20s ", field->p_name,
>  				entry->pid, iter->cpu, field->ksym_name);
> -	if (!ret)
> -		return TRACE_TYPE_PARTIAL_LINE;
> +	if (!ret) {
> +		ret = TRACE_TYPE_PARTIAL_LINE;
> +		goto err_ret;
> +	}
>  
> +	ret = TRACE_TYPE_PARTIAL_LINE;
>  	switch (field->ksym_hbp->info.type) {
>  	case HW_BREAKPOINT_WRITE:
> -		ret = trace_seq_printf(s, " W  ");
> +		if (!trace_seq_printf(s, " W  "))
> +			goto err_ret;
>  		break;
>  	case HW_BREAKPOINT_RW:
> -		ret = trace_seq_printf(s, " RW ");
> +		if (!trace_seq_printf(s, " RW  "))
> +			goto err_ret;
>  		break;
>  	default:
> -		return TRACE_TYPE_PARTIAL_LINE;
> +		goto err_ret;
>  	}
>  
> -	if (!ret)
> -		return TRACE_TYPE_PARTIAL_LINE;
> -
>  	sprint_symbol(str, field->ip);
>  	ret = trace_seq_printf(s, "%-20s\n", str);
> -	if (!ret)
> -		return TRACE_TYPE_PARTIAL_LINE;
> -
> -	return TRACE_TYPE_HANDLED;
> +	if (ret)
> +		return TRACE_TYPE_HANDLED;
> +err_ret:
> +	trace_seq_printf(s, "\n");
> +	return ret;


As told in my email just before, TRACE_TYPE_PARTIAL_LINE
won't print a truncated line or a partial line. Instead, it
will be ignored and entirely retried later.
Then your newline will be ignored.

..unless we have a bug in trace.c

I'm not sure what it the origin of the concatenated printed entries
Ingo has reported.

Did you reproduce it and then this patch fixed it?



>  }
>  
>  struct tracer ksym_tracer __read_mostly =
>