linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Dmitry Safonov <dima@arista.com>
To: linux-kernel@vger.kernel.org
Cc: Dmitry Safonov <0x7f454c46@gmail.com>,
	Dmitry Safonov <dima@arista.com>, Adrian Reber <adrian@lisas.de>,
	Andrei Vagin <avagin@openvz.org>,
	Andy Lutomirski <luto@kernel.org>,
	Christian Brauner <christian.brauner@ubuntu.com>,
	Cyrill Gorcunov <gorcunov@openvz.org>,
	"Eric W. Biederman" <ebiederm@xmission.com>,
	"H. Peter Anvin" <hpa@zytor.com>, Ingo Molnar <mingo@redhat.com>,
	Jeff Dike <jdike@addtoit.com>, Oleg Nesterov <oleg@redhat.com>,
	Pavel Emelyanov <xemul@virtuozzo.com>,
	Shuah Khan <shuah@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	containers@lists.linux-foundation.org, criu@openvz.org,
	linux-api@vger.kernel.org, x86@kernel.org
Subject: [RFC 15/20] timens: Optimize zero-offsets
Date: Wed, 19 Sep 2018 21:50:32 +0100	[thread overview]
Message-ID: <20180919205037.9574-16-dima@arista.com> (raw)
In-Reply-To: <20180919205037.9574-1-dima@arista.com>

Fall through on host or in ns without time set.
Add TIMENS_FALLBACK_SYSCALL which might be wired up if timens offsets
should be unknown for userspace (will result in fall-back to syscalls).

Signed-off-by: Dmitry Safonov <dima@arista.com>
---
 arch/x86/entry/vdso/vclock_gettime.c | 17 +++++++++++++----
 include/linux/timens_offsets.h       | 12 ++++++++++--
 kernel/time/posix-timers.c           | 21 ++++++++++++---------
 kernel/time_namespace.c              |  2 +-
 4 files changed, 36 insertions(+), 16 deletions(-)

diff --git a/arch/x86/entry/vdso/vclock_gettime.c b/arch/x86/entry/vdso/vclock_gettime.c
index a265e2737a9a..458cb1992e2e 100644
--- a/arch/x86/entry/vdso/vclock_gettime.c
+++ b/arch/x86/entry/vdso/vclock_gettime.c
@@ -252,17 +252,25 @@ notrace void set_normalized_timespec(struct timespec *ts, time_t sec, s64 nsec)
 	ts->tv_nsec = nsec;
 }
 
-notrace static __always_inline void monotonic_to_ns(struct timespec *ts)
+notrace static __always_inline int monotonic_to_ns(struct timespec *ts)
 {
 #ifdef CONFIG_TIME_NS
 	struct timens_offsets *timens = (struct timens_offsets *) &timens_page;
 	struct timespec offset;
 
+	/* Optimization: time is the same as on host, return right away */
+	if (!(timens->flags & TIMENS_USE_OFFSETS))
+		return 0;
+
+	if (timens->flags & TIMENS_FALLBACK_SYSCALL)
+		return -1;
+
 	offset = timespec64_to_timespec(timens->monotonic_time_offset);
 
 	*ts = timespec_add(*ts, offset);
 
 #endif
+	return 0;
 }
 
 notrace static int __always_inline do_monotonic(struct timespec *ts)
@@ -283,8 +291,6 @@ notrace static int __always_inline do_monotonic(struct timespec *ts)
 	ts->tv_sec += __iter_div_u64_rem(ns, NSEC_PER_SEC, &ns);
 	ts->tv_nsec = ns;
 
-	monotonic_to_ns(ts);
-
 	return mode;
 }
 
@@ -306,7 +312,6 @@ notrace static void do_monotonic_coarse(struct timespec *ts)
 		ts->tv_sec = gtod->monotonic_time_coarse_sec;
 		ts->tv_nsec = gtod->monotonic_time_coarse_nsec;
 	} while (unlikely(gtod_read_retry(gtod, seq)));
-	monotonic_to_ns(ts);
 }
 
 notrace int __vdso_clock_gettime(clockid_t clock, struct timespec *ts)
@@ -319,12 +324,16 @@ notrace int __vdso_clock_gettime(clockid_t clock, struct timespec *ts)
 	case CLOCK_MONOTONIC:
 		if (do_monotonic(ts) == VCLOCK_NONE)
 			goto fallback;
+		if (monotonic_to_ns(ts))
+			goto fallback;
 		break;
 	case CLOCK_REALTIME_COARSE:
 		do_realtime_coarse(ts);
 		break;
 	case CLOCK_MONOTONIC_COARSE:
 		do_monotonic_coarse(ts);
+		if (monotonic_to_ns(ts))
+			goto fallback;
 		break;
 	default:
 		goto fallback;
diff --git a/include/linux/timens_offsets.h b/include/linux/timens_offsets.h
index 92a8ea5601eb..8c43e7c3e632 100644
--- a/include/linux/timens_offsets.h
+++ b/include/linux/timens_offsets.h
@@ -2,6 +2,13 @@
 #ifndef _LINUX_TIME_OFFSETS_H
 #define _LINUX_TIME_OFFSETS_H
 
+enum {
+	/* We're in namespace - add offsets from vvar */
+	TIMENS_USE_OFFSETS	= 1,
+	/* Don't expose host's offsets, fall back to syscall - slow */
+	TIMENS_FALLBACK_SYSCALL	= 2, /* TODO if anyone actually interested */
+};
+
 /*
  * Time offsets need align as they're placed on vvar page,
  * which should have tail paddings on ia32 vdso.
@@ -10,8 +17,9 @@
  * to timespec because of a padding occuring between the fields.
  */
 struct timens_offsets {
-	struct timespec64  monotonic_time_offset __aligned(8);
-	struct timespec64  monotonic_boottime_offset __aligned(8);
+	u64		  flags;
+	struct timespec64 monotonic_time_offset __aligned(8);
+	struct timespec64 monotonic_boottime_offset __aligned(8);
 };
 
 #endif
diff --git a/kernel/time/posix-timers.c b/kernel/time/posix-timers.c
index 701cb0602b7a..576dbd24c498 100644
--- a/kernel/time/posix-timers.c
+++ b/kernel/time/posix-timers.c
@@ -210,7 +210,7 @@ static void common_timens_adjust(clockid_t which_clock, struct timespec64 *tp)
 {
 	struct timens_offsets *ns_offsets = current->nsproxy->time_ns->offsets;
 
-	if (!ns_offsets)
+	if (!ns_offsets || !(ns_offsets->flags & TIMENS_USE_OFFSETS))
 		return;
 
 	switch (which_clock) {
@@ -234,15 +234,16 @@ static int posix_ktime_set_ts(clockid_t which_clock,
 	struct timens_offsets *ns_offsets = current->nsproxy->time_ns->offsets;
 	struct timespec64 ktp;
 
+	if (!ns_offsets)
+		return -EINVAL;
+
 	if (!ns_capable(current->nsproxy->time_ns->user_ns, CAP_SYS_TIME))
 		return -EPERM;
 
 	ktime_get_ts64(&ktp);
 
-	if (ns_offsets)
-		ns_offsets->monotonic_time_offset = timespec64_sub(*tp, ktp);
-	else
-		return -EINVAL;
+	ns_offsets->monotonic_time_offset = timespec64_sub(*tp, ktp);
+	ns_offsets->flags |= TIMENS_USE_OFFSETS;
 
 	return 0;
 }
@@ -296,15 +297,17 @@ static int posix_set_boottime(clockid_t which_clock, const struct timespec64 *tp
 	struct timens_offsets *ns_offsets = current->nsproxy->time_ns->offsets;
 	struct timespec64 ktp;
 
+	if (!ns_offsets)
+		return -EINVAL;
+
 	if (!ns_capable(current->nsproxy->time_ns->user_ns, CAP_SYS_TIME))
 		return -EPERM;
 
 	ktime_get_boottime_ts64(&ktp);
 
-	if (ns_offsets)
-		ns_offsets->monotonic_boottime_offset = timespec64_sub(*tp, ktp);
-	else
-		return -EINVAL;
+	ns_offsets->monotonic_boottime_offset = timespec64_sub(*tp, ktp);
+	ns_offsets->flags |= TIMENS_USE_OFFSETS;
+
 	return 0;
 }
 
diff --git a/kernel/time_namespace.c b/kernel/time_namespace.c
index f88ae0e17d92..4052bdcec110 100644
--- a/kernel/time_namespace.c
+++ b/kernel/time_namespace.c
@@ -171,7 +171,7 @@ static void clock_timens_fixup(int clockid, struct timespec64 *val, bool to_ns)
 	struct timens_offsets *ns_offsets = current->nsproxy->time_ns->offsets;
 	struct timespec64 *offsets = NULL;
 
-	if (!ns_offsets)
+	if (!ns_offsets || !(ns_offsets->flags & TIMENS_USE_OFFSETS))
 		return;
 
 	if (val->tv_sec == 0 && val->tv_nsec == 0)
-- 
2.13.6


  parent reply	other threads:[~2018-09-19 20:51 UTC|newest]

Thread overview: 59+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-09-19 20:50 [RFC 00/20] ns: Introduce Time Namespace Dmitry Safonov
2018-09-19 20:50 ` [RFC 01/20] " Dmitry Safonov
2018-09-28 18:20   ` Laurent Vivier
2018-09-19 20:50 ` [RFC 02/20] timens: Add timens_offsets Dmitry Safonov
2018-09-20 18:45   ` Cyrill Gorcunov
2018-09-20 22:14     ` Cyrill Gorcunov
2018-09-19 20:50 ` [RFC 03/20] timens: Introduce CLOCK_MONOTONIC offsets Dmitry Safonov
2018-09-19 20:50 ` [RFC 04/20] timens: Introduce CLOCK_BOOTTIME offset Dmitry Safonov
2018-09-30  3:18   ` [LKP] [timens] 3cc8de9dcb: RIP:posix_get_boottime kernel test robot
2018-09-19 20:50 ` [RFC 05/20] timerfd/timens: Take into account ns clock offsets Dmitry Safonov
2018-09-19 20:50 ` [RFC 06/20] kernel: Take into account timens clock offsets in clock_nanosleep Dmitry Safonov
2018-09-19 20:50 ` [RFC 07/20] timens: Shift /proc/uptime Dmitry Safonov
2018-09-19 20:50 ` [RFC 08/20] x86/vdso: Restrict splitting vvar vma Dmitry Safonov
2018-09-19 20:50 ` [RFC 09/20] x86/vdso/timens: Add offsets page in vvar Dmitry Safonov
2018-09-19 20:50 ` [RFC 10/20] x86/vdso: Use set_normalized_timespec() to avoid 32 bit overflow Dmitry Safonov
2018-09-19 20:50 ` [RFC 11/20] x86/vdso: Purge timens page on setns()/unshare()/clone() Dmitry Safonov
2018-09-19 20:50 ` [RFC 12/20] x86/vdso: Look for vvar vma to purge timens page Dmitry Safonov
2018-09-19 20:50 ` [RFC 13/20] posix-timers/timens: Take into account clock offsets Dmitry Safonov
2018-09-30  3:11   ` [LKP] [posix] 25217c6e39: BUG:KASAN:null-ptr-deref_in_c kernel test robot
2018-09-19 20:50 ` [RFC 14/20] timens: Add align for timens_offsets Dmitry Safonov
2018-09-19 20:50 ` Dmitry Safonov [this message]
2018-09-19 20:50 ` [RFC 16/20] selftest: Add Time Namespace test for supported clocks Dmitry Safonov
2018-09-24 21:36   ` Shuah Khan
2018-09-19 20:50 ` [RFC 17/20] selftest/timens: Add test for timerfd Dmitry Safonov
2018-09-19 20:50 ` [RFC 18/20] selftest/timens: Add test for clock_nanosleep Dmitry Safonov
2018-09-19 20:50 ` [RFC 19/20] timens/selftest: Add procfs selftest Dmitry Safonov
2018-09-19 20:50 ` [RFC 20/20] timens/selftest: Add timer offsets test Dmitry Safonov
2018-09-21 12:27 ` [RFC 00/20] ns: Introduce Time Namespace Eric W. Biederman
2018-09-24 20:51   ` Andrey Vagin
2018-09-24 22:02     ` Eric W. Biederman
2018-09-25  1:42       ` Andrey Vagin
2018-09-26 17:36         ` Eric W. Biederman
2018-09-26 17:59           ` Dmitry Safonov
2018-09-27 21:30           ` Thomas Gleixner
2018-09-27 21:41             ` Thomas Gleixner
2018-10-01 23:20               ` Andrey Vagin
2018-10-02  6:15                 ` Thomas Gleixner
2018-10-02 21:05                   ` Dmitry Safonov
2018-10-02 21:26                     ` Thomas Gleixner
2018-09-28 17:03             ` Eric W. Biederman
2018-09-28 19:32               ` Thomas Gleixner
2018-10-01  9:05                 ` Eric W. Biederman
2018-10-01  9:15                 ` Setting monotonic time? Eric W. Biederman
2018-10-01 18:52                   ` Thomas Gleixner
2018-10-02 20:00                     ` Arnd Bergmann
2018-10-02 20:06                       ` Thomas Gleixner
2018-10-03  4:50                         ` Eric W. Biederman
2018-10-03  5:25                           ` Thomas Gleixner
2018-10-03  6:14                             ` Eric W. Biederman
2018-10-03  7:02                               ` Arnd Bergmann
2018-10-03  6:14                             ` Thomas Gleixner
2018-10-01 20:51                   ` Andrey Vagin
2018-10-02  6:16                     ` Thomas Gleixner
2018-10-21  1:41               ` [RFC 00/20] ns: Introduce Time Namespace Andrei Vagin
2018-10-21  3:54                 ` Andrei Vagin
2018-10-29 20:33                 ` Thomas Gleixner
2018-10-29 21:21                   ` Eric W. Biederman
2018-10-29 21:36                     ` Thomas Gleixner
2018-10-31 16:26                   ` Andrei Vagin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180919205037.9574-16-dima@arista.com \
    --to=dima@arista.com \
    --cc=0x7f454c46@gmail.com \
    --cc=adrian@lisas.de \
    --cc=avagin@openvz.org \
    --cc=christian.brauner@ubuntu.com \
    --cc=containers@lists.linux-foundation.org \
    --cc=criu@openvz.org \
    --cc=ebiederm@xmission.com \
    --cc=gorcunov@openvz.org \
    --cc=hpa@zytor.com \
    --cc=jdike@addtoit.com \
    --cc=linux-api@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=luto@kernel.org \
    --cc=mingo@redhat.com \
    --cc=oleg@redhat.com \
    --cc=shuah@kernel.org \
    --cc=tglx@linutronix.de \
    --cc=x86@kernel.org \
    --cc=xemul@virtuozzo.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).