[rtems commit] timecounter: Merge FreeBSD change r304285

[Sebastian Huber]

Module:    rtems
Branch:    master
Commit:    748871578c3dd0077e1f1de246993c80424ab63b
Changeset: http://git.rtems.org/rtems/commit/?id=748871578c3dd0077e1f1de246993c80424ab63b

Author:    Konstantin Belousov <kib at FreeBSD.org>
Date:      Wed Aug 17 09:52:09 2016 +0000

timecounter: Merge FreeBSD change r304285

Implement userspace gettimeofday(2) with HPET timecounter.

Right now, userspace (fast) gettimeofday(2) on x86 only works for
RDTSC.  For older machines, like Core2, where RDTSC is not C2/C3
invariant, and which fall to HPET hardware, this means that the call
has both the penalty of the syscall and of the uncached hw behind the
QPI or PCIe connection to the sought bridge.  Nothing can me done
against the access latency, but the syscall overhead can be removed.
System already provides mappable /dev/hpetX devices, which gives
straight access to the HPET registers page.

Add yet another algorithm to the x86 'vdso' timehands. Libc is updated
to handle both RDTSC and HPET.  For HPET, the index of the hpet device
to mmap is passed from kernel to userspace, index might be changed and
libc invalidates its mapping as needed.

Remove cpu_fill_vdso_timehands() KPI, instead require that
timecounters which can be used from userspace, to provide
tc_fill_vdso_timehands{,32}() methods.  Merge i386 and amd64
libc/<arch>/sys/__vdso_gettc.c into one source file in the new
libc/x86/sys location.  __vdso_gettc() internal interface is changed
to move timecounter algorithm detection into the MD code.

Measurements show that RDTSC even with the syscall overhead is faster
than userspace HPET access.  But still, userspace HPET is three-four
times faster than syscall HPET on several Core2 and SandyBridge
machines.

Tested by:	Howard Su <howard0su at gmail.com>
Sponsored by:	The FreeBSD Foundation
MFC after:	1 month
Differential revision:	https://reviews.freebsd.org/D7473

Update #3175.

---

 cpukit/score/include/sys/timetc.h | 14 ++++++++++++++
 cpukit/score/src/kern_tc.c        | 19 ++++++++++++++-----
 2 files changed, 28 insertions(+), 5 deletions(-)

diff --git a/cpukit/score/include/sys/timetc.h b/cpukit/score/include/sys/timetc.h
index 73e63f7..361535c 100644
--- a/cpukit/score/include/sys/timetc.h
+++ b/cpukit/score/include/sys/timetc.h
@@ -30,8 +30,18 @@
  */
 
 struct timecounter;
+struct vdso_timehands;
+struct vdso_timehands32;
+#ifndef __rtems__
+typedef u_int timecounter_get_t(struct timecounter *);
+#else /* __rtems__ */
 typedef uint32_t timecounter_get_t(struct timecounter *);
+#endif /* __rtems__ */
 typedef void timecounter_pps_t(struct timecounter *);
+typedef uint32_t timecounter_fill_vdso_timehands_t(struct vdso_timehands *,
+    struct timecounter *);
+typedef uint32_t timecounter_fill_vdso_timehands32_t(struct vdso_timehands32 *,
+    struct timecounter *);
 
 struct timecounter {
 	timecounter_get_t	*tc_get_timecount;
@@ -70,6 +80,10 @@ struct timecounter {
 		/* Pointer to the timecounter's private parts. */
 	struct timecounter	*tc_next;
 		/* Pointer to the next timecounter. */
+#ifndef __rtems__
+	timecounter_fill_vdso_timehands_t *tc_fill_vdso_timehands;
+	timecounter_fill_vdso_timehands32_t *tc_fill_vdso_timehands32;
+#endif /* __rtems__ */
 };
 
 extern struct timecounter *timecounter;
diff --git a/cpukit/score/src/kern_tc.c b/cpukit/score/src/kern_tc.c
index 78e1308..272461d 100644
--- a/cpukit/score/src/kern_tc.c
+++ b/cpukit/score/src/kern_tc.c
@@ -6,11 +6,14 @@
  * this stuff is worth it, you can buy me a beer in return.   Poul-Henning Kamp
  * ----------------------------------------------------------------------------
  *
- * Copyright (c) 2011 The FreeBSD Foundation
+ * Copyright (c) 2011, 2015, 2016 The FreeBSD Foundation
  * All rights reserved.
  *
  * Portions of this software were developed by Julien Ridoux at the University
  * of Melbourne under sponsorship from the FreeBSD Foundation.
+ *
+ * Portions of this software were developed by Konstantin Belousov
+ * under sponsorship from the FreeBSD Foundation.
  */
 
 #ifdef __rtems__
@@ -2388,13 +2391,16 @@ tc_fill_vdso_timehands(struct vdso_timehands *vdso_th)
 	uint32_t enabled;
 
 	th = timehands;
-	vdso_th->th_algo = VDSO_TH_ALGO_1;
 	vdso_th->th_scale = th->th_scale;
 	vdso_th->th_offset_count = th->th_offset_count;
 	vdso_th->th_counter_mask = th->th_counter->tc_counter_mask;
 	vdso_th->th_offset = th->th_offset;
 	vdso_th->th_boottime = th->th_boottime;
-	enabled = cpu_fill_vdso_timehands(vdso_th, th->th_counter);
+	if (th->th_counter->tc_fill_vdso_timehands != NULL) {
+		enabled = th->th_counter->tc_fill_vdso_timehands(vdso_th,
+		    th->th_counter);
+	} else
+		enabled = 0;
 	if (!vdso_th_enable)
 		enabled = 0;
 	return (enabled);
@@ -2409,7 +2415,6 @@ tc_fill_vdso_timehands32(struct vdso_timehands32 *vdso_th32)
 	uint32_t enabled;
 
 	th = timehands;
-	vdso_th32->th_algo = VDSO_TH_ALGO_1;
 	*(uint64_t *)&vdso_th32->th_scale[0] = th->th_scale;
 	vdso_th32->th_offset_count = th->th_offset_count;
 	vdso_th32->th_counter_mask = th->th_counter->tc_counter_mask;
@@ -2417,7 +2422,11 @@ tc_fill_vdso_timehands32(struct vdso_timehands32 *vdso_th32)
 	*(uint64_t *)&vdso_th32->th_offset.frac[0] = th->th_offset.frac;
 	vdso_th32->th_boottime.sec = th->th_boottime.sec;
 	*(uint64_t *)&vdso_th32->th_boottime.frac[0] = th->th_boottime.frac;
-	enabled = cpu_fill_vdso_timehands32(vdso_th32, th->th_counter);
+	if (th->th_counter->tc_fill_vdso_timehands32 != NULL) {
+		enabled = th->th_counter->tc_fill_vdso_timehands32(vdso_th32,
+		    th->th_counter);
+	} else
+		enabled = 0;
 	if (!vdso_th_enable)
 		enabled = 0;
 	return (enabled);