Re: [PATCH 1/2] New field stop_pc in tracepoint_hit_ctx

[Pedro Alves <palves@redhat.com>]

On 03/02/2012 01:00 AM, Yao Qi wrote:
> This patch is to add a new field `stop_pc' in tracepoint_hit_ctx, so
> parameter `stop_pc' used here and there can be removed.  This change
> allows us to give a clean interface in next patch.
> 
> Note the name `tracepoint_hit_ctx' is not very accurate, because it
> has been used in collect_data_at_step for "while-stepping" action also.
> It may be renamed to `tracepoint_action_ctx', which is about the
> context of doing tracepoint actions.  I don't rename `tracepoint_hit_ctx'
> to keep this patch as readable as possible.  I can send a follow-up
> patch to rename it if this change is reasonable.
> 
> As `tracepoint_hit_ctx' is about the context of doing tracepoint
> actions, it is natural to add field `stop_pc' to show the pc value when
> to do tracepoint actions.

Fine with me, though honestly I find that the similarity between
tracepoint_action vs tracepoint_action_ctx will increase confusion,
not reduce.

>    tpoint = ust_marker_to_static_tracepoint (mdata);
> +  ctx.base.stop_pc = tpoint->address;

tpoint may be NULL, see just below.

> +
>    if (tpoint == NULL)
>      {

-- 
Pedro Alves