fix: Ensure block io queue segment size is correct (#51)

utcs-scea · Dec 11, 2024 · db4069b · db4069b
1 parent 515a788
commit db4069b
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 5 deletions.
diff --git a/python/kernmlops/data_collection/bpf_instrumentation/bpf/blk_io.bpf.c b/python/kernmlops/data_collection/bpf_instrumentation/bpf/blk_io.bpf.c
@@ -57,6 +57,8 @@ static int block_4k_ios(u32 bytes) {
 
 // copied from linux kernel
 static inline unsigned short blk_rq_nr_phys_segments_dup(struct request* rq) {
+  if ((rq->cmd_flags & REQ_IDLE) && (rq->cmd_flags & REQ_FUA) && (rq->cmd_flags & REQ_META))
+    return 0;
   if (rq->rq_flags & RQF_SPECIAL_PAYLOAD)
     return 1;
   return rq->nr_phys_segments;

diff --git a/python/kernmlops/data_schema/block_io.py b/python/kernmlops/data_schema/block_io.py
@@ -194,15 +194,23 @@ def from_tables(cls, queue_table: BlockIOQueueTable, latency_table: BlockIOLaten
                 "collection_id",
             ],
             how="inner",
-        ).with_columns(
+        ).filter(
+            ((pl.col("finish_ts_uptime_us") - pl.col(UPTIME_TIMESTAMP)) > 0) &
+            ((pl.col("finish_ts_uptime_us") - pl.col(UPTIME_TIMESTAMP)) < pl.col("block_latency_us"))
+        )
+        block_df = block_df.unique([
+            "cpu",
+            "device",
+            "sector",
+            "segments",
+            "block_io_bytes",
+            UPTIME_TIMESTAMP,
+            "block_io_flags",
+        ]).with_columns(
             (pl.col("finish_ts_uptime_us") - pl.col(UPTIME_TIMESTAMP)).alias("measured_latency_us"),
             pl.col("block_io_flags").map_elements(
                 flags_print, return_dtype=pl.String,
             ).alias("block_io_flags_string"),
-        ).filter(
-            pl.col("measured_latency_us") > 0
-        ).filter(
-            (pl.col("measured_latency_us") - 100) < pl.col("block_latency_us")
         ).sort(UPTIME_TIMESTAMP, descending=False)
         return cls.from_df(block_df)