summaryrefslogtreecommitdiff
path: root/libpq/postgresql/port/linux/pg_config_os.h
diff options
context:
space:
mode:
Diffstat (limited to 'libpq/postgresql/port/linux/pg_config_os.h')
-rw-r--r--libpq/postgresql/port/linux/pg_config_os.h22
1 files changed, 22 insertions, 0 deletions
diff --git a/libpq/postgresql/port/linux/pg_config_os.h b/libpq/postgresql/port/linux/pg_config_os.h
new file mode 100644
index 0000000..7a6e46c
--- /dev/null
+++ b/libpq/postgresql/port/linux/pg_config_os.h
@@ -0,0 +1,22 @@
+/* src/include/port/linux.h */
+
+/*
+ * As of July 2007, all known versions of the Linux kernel will sometimes
+ * return EIDRM for a shmctl() operation when EINVAL is correct (it happens
+ * when the low-order 15 bits of the supplied shm ID match the slot number
+ * assigned to a newer shmem segment). We deal with this by assuming that
+ * EIDRM means EINVAL in PGSharedMemoryIsInUse(). This is reasonably safe
+ * since in fact Linux has no excuse for ever returning EIDRM; it doesn't
+ * track removed segments in a way that would allow distinguishing them from
+ * private ones. But someday that code might get upgraded, and we'd have
+ * to have a kernel version test here.
+ */
+#define HAVE_LINUX_EIDRM_BUG
+
+/*
+ * Set the default wal_sync_method to fdatasync. With recent Linux versions,
+ * xlogdefs.h's normal rules will prefer open_datasync, which (a) doesn't
+ * perform better and (b) causes outright failures on ext4 data=journal
+ * filesystems, because those don't support O_DIRECT.
+ */
+#define PLATFORM_DEFAULT_SYNC_METHOD SYNC_METHOD_FDATASYNC