Hi, Hackers,
While auditing src/backend/storage/file, I spotted a small but
unambiguous bug in pg_flush_data()'s sync_file_range() retry path.
Details below in the commit message; the fix is a one-token change.
Thanks,
DaeMyung
---
The retry branch compared the syscall return value (rc) against EINTR,
but sync_file_range() returns -1 on error and reports the actual errno
out-of-band, so rc == EINTR was always false and the retry was dead
code. When sync_file_range() was actually interrupted by a signal,
control fell through to the warning/error path instead of retrying.
Compare errno instead, matching the convention already used elsewhere
in this file (e.g. the pg_pread/pg_pwrite EINTR retries) and a few
lines down for ENOSYS in this same block.
---
src/backend/storage/file/fd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/backend/storage/file/fd.c b/src/backend/storage/file/fd.c
index a8be066afe0..817855e2720 100644
--- a/src/backend/storage/file/fd.c
+++ b/src/backend/storage/file/fd.c
@@ -563,7 +563,7 @@ retry:
{
int elevel;
- if (rc == EINTR)
+ if (errno == EINTR)
goto retry;
/*
