once: implement DO_ONCE_LITE for non-fast-path "do once" functionality

Certain uses of "do once" functionality reside outside of fast path,
and so do not require jump label patching via static keys, making
existing DO_ONCE undesirable in such cases.

Replace uses of __section(".data.once") with DO_ONCE_LITE(_IF)?

This patch changes the return values of xfs_printk_once, printk_once,
and printk_deferred_once. Before, they returned whether the print was
performed, but now, they always return true. This is okay because the
return values of the following macros are entirely ignored throughout
the kernel:
- xfs_printk_once
- xfs_warn_once
- xfs_notice_once
- xfs_info_once
- printk_once
- pr_emerg_once
- pr_alert_once
- pr_crit_once
- pr_err_once
- pr_warn_once
- pr_notice_once
- pr_info_once
- pr_devel_once
- pr_debug_once
- printk_deferred_once
- orc_warn

Changes
v3:
  - Expand commit message to explain why changing return values of
    xfs_printk_once, printk_once, printk_deferred_once is benign
v2:
  - Fix i386 build warnings

Signed-off-by: Tanner Love <tannerlove@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Mahesh Bandewar <maheshb@google.com>
Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 3 insertions, 10 deletions

diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index cd80d046c7a5..d5d8c088a55d 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -20,6 +20,7 @@
 #include <linux/irq_work.h>
 #include <linux/workqueue.h>
 #include <linux/ctype.h>
+#include <linux/once_lite.h>
 
 #ifdef CONFIG_FTRACE_SYSCALLS
 #include <asm/unistd.h>		/* For NR_SYSCALLS	     */
@@ -99,16 +100,8 @@ enum trace_type {
 #include "trace_entries.h"
 
 /* Use this for memory failure errors */
-#define MEM_FAIL(condition, fmt, ...) ({			\
-	static bool __section(".data.once") __warned;		\
-	int __ret_warn_once = !!(condition);			\
-								\
-	if (unlikely(__ret_warn_once && !__warned)) {		\
-		__warned = true;				\
-		pr_err("ERROR: " fmt, ##__VA_ARGS__);		\
-	}							\
-	unlikely(__ret_warn_once);				\
-})
+#define MEM_FAIL(condition, fmt, ...)					\
+	DO_ONCE_LITE_IF(condition, pr_err, "ERROR: " fmt, ##__VA_ARGS__)
 
 /*
  * syscalls are special, and need special handling, this is why