Re: [PATCH v2 4/7] perf diff: Use hists to manage basic blocks per symbol

[Jin, Yao]

On 6/5/2019 7:44 PM, Jiri Olsa wrote:
On Mon, Jun 03, 2019 at 10:36:14PM +0800, Jin Yao wrote:

SNIP

                data__for_each_file_new(i, d) {
                        pair = get_pair_data(he, d);
                        if (!pair)
@@ -510,6 +683,9 @@ static void hists__precompute(struct hists *hists)
                        case COMPUTE_WEIGHTED_DIFF:
                                compute_wdiff(he, pair);
                                break;
+                       case COMPUTE_CYCLES:
+                               process_block_per_sym(pair, d);
+                               break;
                        default:
                                BUG_ON(1);
                        }
@@ -713,6 +889,14 @@ hist_entry__cmp_wdiff_idx(struct perf_hpp_fmt *fmt 
__maybe_unused,
                                           sort_compute);
  }
  
+static int64_t
+hist_entry__cmp_cycles_idx(struct perf_hpp_fmt *fmt __maybe_unused,
+                          struct hist_entry *left __maybe_unused,
+                          struct hist_entry *right __maybe_unused)
+{
+       return 0;
+}

this is hist_entry__cmp_nop.. please use it instead and
explain in comment why for COMPUTE_CYCLES we need the
default sort

jirka


fmt->sort should be set otherwise since fmt->sort will be called without 
checking valid, the crash happens. Yes, I should use hist_entry__cmp_nop 
instead, and will add some comments for COMPUTE_CYCLES.

Thanks
Jin Yao