kernel: use gettime(9)/getuptime(9) in lieu of time_second(9)/time_uptime(9)

time_second(9) and time_uptime(9) are widely used in the kernel to
quickly get the system UTC or system uptime as a time_t.  However,
time_t is 64-bit everywhere, so it is not generally safe to use them
on 32-bit platforms: you have a split-read problem if your hardware
cannot perform atomic 64-bit reads.

This patch replaces time_second(9) with gettime(9), a safer successor
interface, throughout the kernel.  Similarly, time_uptime(9) is replaced
with getuptime(9).

There is a performance cost on 32-bit platforms in exchange for
eliminating the split-read problem: instead of two register reads you
now have a lockless read loop to pull the values from the timehands.
This is really not *too* bad in the grand scheme of things, but
compared to what we were doing before it is several times slower.

There is no performance cost on 64-bit (__LP64__) platforms.

With input from visa@, dlg@, and tedu@.

Several bugs squashed by visa@.

ok kettenis@

1 files changed, 4 insertions, 4 deletions

diff --git a/sys/net/pf_norm.c b/sys/net/pf_norm.c
index 570d0523cfc..0a438543e10 100644
--- a/sys/net/pf_norm.c
+++ b/sys/net/pf_norm.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: pf_norm.c,v 1.218 2019/02/28 20:20:47 bluhm Exp $ */
+/*	$OpenBSD: pf_norm.c,v 1.219 2020/06/24 22:03:43 cheloha Exp $ */
 
 /*
  * Copyright 2001 Niels Provos <provos@citi.umich.edu>
@@ -220,7 +220,7 @@ pf_purge_expired_fragments(void)
 
 	PF_ASSERT_UNLOCKED();
 
-	expire = time_uptime - pf_default_rule.timeout[PFTM_FRAG];
+	expire = getuptime() - pf_default_rule.timeout[PFTM_FRAG];
 
 	PF_FRAG_LOCK();
 	while ((frag = TAILQ_LAST(&pf_fragqueue, pf_fragqueue)) != NULL) {
@@ -589,7 +589,7 @@ pf_fillup_fragment(struct pf_frnode *key, u_int32_t id,
 		memset(frag->fr_entries, 0, sizeof(frag->fr_entries));
 		TAILQ_INIT(&frag->fr_queue);
 		frag->fr_id = id;
-		frag->fr_timeout = time_uptime;
+		frag->fr_timeout = getuptime();
 		frag->fr_gen = frnode->fn_gen++;
 		frag->fr_maxlen = frent->fe_len;
 		frag->fr_holes = 1;
@@ -1352,7 +1352,7 @@ pf_normalize_tcp_stateful(struct pf_pdesc *pd, u_short *reason,
 	getmicrouptime(&uptime);
 	if (src->scrub && (src->scrub->pfss_flags & PFSS_PAWS) &&
 	    (uptime.tv_sec - src->scrub->pfss_last.tv_sec > TS_MAX_IDLE ||
-	    time_uptime - state->creation > TS_MAX_CONN))  {
+	    getuptime() - state->creation > TS_MAX_CONN))  {
 		if (pf_status.debug >= LOG_NOTICE) {
 			log(LOG_NOTICE, "pf: src idled out of PAWS ");
 			pf_print_state(state);