Closed deukyeon closed 1 year ago
There is a bug that causes a segmentation fault when recording histograms again after calling splinterdb_stats_reset(), as shown in the gdb output:
splinterdb_stats_reset()
Thread 30 "ycsbc" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7ffe8bbff640 (LWP 1843516)] trunk_insert (spl=<optimized out>, tuple_key=..., data=...) at src/trunk.c:6188 6188 platform_histo_insert(spl->stats[tid].update_latency_histo,
The issue seems to be that the histogram is zeroed out in trunk_reset_stats(), which is called by splinterdb_stats_reset(), as shown below:
trunk_reset_stats()
void trunk_reset_stats(trunk_handle *spl) { if (spl->cfg.use_stats) { memset(spl->stats, 0, MAX_THREADS * sizeof(trunk_stats)); } }
There is a bug that causes a segmentation fault when recording histograms again after calling
splinterdb_stats_reset()
, as shown in the gdb output:The issue seems to be that the histogram is zeroed out in
trunk_reset_stats()
, which is called bysplinterdb_stats_reset()
, as shown below: