Re: perfmon broken

[Stephane Eranian <eranian@hpl.hp.com>]

Alex,

Yes, your patch will likely fix the problem. I need to check
to see if we need to keep TASK_STOPPED.

Thanks.


On Tue, Sep 28, 2004 at 04:58:38PM -0600, Alex Williamson wrote:
> 
>    Looks like this changeset back between 2.6.9-rc1-bk15 & bk16 broke
> perfmon:
> 
> http://linux.bkbits.net:8080/linux-2.5/cset@413f1bdabfaQNzIZpU6bPxNlSxdriQ
> 
>         [PATCH] cleanup ptrace stops and remove notify_parent
>         
>         This adds a new state TASK_TRACED that is used in place of
>         TASK_STOPPED when a thread stops because it is ptraced.  Now
>         ptrace operations are only permitted when the target is in
>         TASK_TRACED state, not in TASK_STOPPED.
>         ...
> 
>  Here's the simple fix, perhaps Stephane has a better approach.  Thanks,
> 
> 	Alex
> 
> -- 
> Signed-off-by: Alex Williamson <alex.williamson@hp.com>
> 
> 
> === arch/ia64/kernel/perfmon.c 1.90 vs edited ==> --- 1.90/arch/ia64/kernel/perfmon.c	2004-09-03 08:54:08 -06:00
> +++ edited/arch/ia64/kernel/perfmon.c	2004-09-28 16:42:05 -06:00
> @@ -2601,7 +2601,7 @@
>  	 */
>  	if (task = current) return 0;
>  
> -	if (task->state != TASK_STOPPED) {
> +	if ((task->state != TASK_STOPPED) && (task->state != TASK_TRACED)){
>  		DPRINT(("cannot attach to non-stopped task [%d] state=%ld\n", task->pid, task->state));
>  		return -EBUSY;
>  	}
> @@ -4755,7 +4755,8 @@
>  	 * the task must be stopped.
>  	 */
>  	if (PFM_CMD_STOPPED(cmd)) {
> -		if (task->state != TASK_STOPPED) {
> +		if ((task->state != TASK_STOPPED) &&
> +		    (task->state != TASK_TRACED)) {
>  			DPRINT(("[%d] task not in stopped state\n", task->pid));
>  			return -EBUSY;
>  		}
> 

-- 

-Stephane