rust/neon
1
0
Fork 0
mirror of https://github.com/neondatabase/neon.git synced 2026-05-28 02:20:42 +00:00
Code Issues Packages Projects Releases Wiki Activity

libpagestore: timeout = max(0, difference), not min(0, difference) (#10274)

Browse Source 
Using `min(0, ...)` causes us to fail to wait in most situations, so a
lack of data would be a hot wait loop, which is bad.

## Problem

We noticed high CPU usage in some situations
This commit is contained in:
Matthias van de Meent
2025-01-07 10:07:38 +01:00
committed by GitHub
parent 02f81b6469
commit 30863c0104
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
pgxn/neon/libpagestore.c
View File

@@ -680,7 +680,7 @@ call_PQgetCopyData(shardno_t shard_no, char **buffer)
* but in the cases that take exceptionally long, it's useful to log the
* exact timestamps.
*/
#define LOG_INTERVAL_US UINT64CONST(10 * 1000000)
#define LOG_INTERVAL_MS INT64CONST(10 * 1000)
INSTR_TIME_SET_CURRENT(now);
start_ts = last_log_ts = now;
@@ -694,7 +694,7 @@ retry:
WaitEvent event;
long timeout;
timeout = Min(0, LOG_INTERVAL_US - INSTR_TIME_GET_MICROSEC(since_last_log));
timeout = Max(0, LOG_INTERVAL_MS - INSTR_TIME_GET_MILLISEC(since_last_log));
/* Sleep until there's something to do */
(void) WaitEventSetWait(shard->wes_read, timeout, &event, 1,
@@ -723,7 +723,7 @@ retry:
INSTR_TIME_SET_CURRENT(now);
since_last_log = now;
INSTR_TIME_SUBTRACT(since_last_log, last_log_ts);
if (INSTR_TIME_GET_MICROSEC(since_last_log) >= LOG_INTERVAL_US)
if (INSTR_TIME_GET_MILLISEC(since_last_log) >= LOG_INTERVAL_MS)
{
since_start = now;
INSTR_TIME_SUBTRACT(since_start, start_ts);